# PHYSIOLOGICAL AND PATHOLOGICAL RESPONSES TO HYPOXIA AND HIGH ALTITUDE

EDITED BY : Rodrigo Iturriaga, Rodrigo Del Rio and Jean-Paul R-Richalet PUBLISHED IN : Frontiers in Physiology

### Frontiers eBook Copyright Statement

The copyright in the text of individual articles in this eBook is the property of their respective authors or their respective institutions or funders. The copyright in graphics and images within each article may be subject to copyright of other parties. In both cases this is subject to a license granted to Frontiers. The compilation of articles constituting this eBook is the property of Frontiers.

Each article within this eBook, and the eBook itself, are published under the most recent version of the Creative Commons CC-BY licence. The version current at the date of publication of this eBook is CC-BY 4.0. If the CC-BY licence is updated, the licence granted by Frontiers is automatically updated to the new version.

When exercising any right under the CC-BY licence, Frontiers must be attributed as the original publisher of the article or eBook, as applicable.

Authors have the responsibility of ensuring that any graphics or other materials which are the property of others may be included in the CC-BY licence, but this should be checked before relying on the CC-BY licence to reproduce those materials. Any copyright notices relating to those materials must be complied with.

Copyright and source acknowledgement notices may not be removed and must be displayed in any copy, derivative work or partial copy which includes the elements in question.

All copyright, and all rights therein, are protected by national and international copyright laws. The above represents a summary only. For further information please read Frontiers' Conditions for Website Use and Copyright Statement, and the applicable CC-BY licence.

ISSN 1664-8714 ISBN 978-2-88963-800-0 DOI 10.3389/978-2-88963-800-0

### About Frontiers

Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering approach to the world of academia, radically improving the way scholarly research is managed. The grand vision of Frontiers is a world where all people have an equal opportunity to seek, share and generate knowledge. Frontiers provides immediate and permanent online open access to all its publications, but this alone is not enough to realize our grand goals.

### Frontiers Journal Series

The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online journals, promising a paradigm shift from the current review, selection and dissemination processes in academic publishing. All Frontiers journals are driven by researchers for researchers; therefore, they constitute a service to the scholarly community. At the same time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing system, initially addressing specific communities of scholars, and gradually climbing up to broader public understanding, thus serving the interests of the lay society, too.

### Dedication to Quality

Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative interactions between authors and review editors, who include some of the world's best academicians. Research must be certified by peers before entering a stream of knowledge that may eventually reach the public - and shape society; therefore, Frontiers only applies the most rigorous and unbiased reviews.

Frontiers revolutionizes research publishing by freely delivering the most outstanding research, evaluated with no bias from both the academic and social point of view. By applying the most advanced information technologies, Frontiers is catapulting scholarly publishing into a new generation.

### What are Frontiers Research Topics?

Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series: they are collections of at least ten articles, all centered on a particular subject. With their unique mix of varied contributions from Original Research to Review Articles, Frontiers Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author by contacting the Frontiers Editorial Office: researchtopics@frontiersin.org

# PHYSIOLOGICAL AND PATHOLOGICAL RESPONSES TO HYPOXIA AND HIGH ALTITUDE

Topic Editors:

Rodrigo Iturriaga, Pontificia Universidad Católica de chile, Chile Rodrigo Del Rio, Pontifical Catholic University of Chile, Chile Jean-Paul R-Richalet, Université Paris 13, France

The appearance of photosynthetic organisms about 3 billion years ago increased the partial pressure of oxygen (PO2) in the atmosphere and enabled the evolution of organisms that use glucose and oxygen to produce ATP by oxidative phosphorylation. Hypoxia is commonly defined as the reduced availability of oxygen in the tissues produced by different causes, which include reduction of atmospheric PO2 as in high altitude, and secondary to pathological conditions such as sleep breathing and pulmonary disorders, anemia, and cardiovascular alterations leading to inadequate transport, delivery, and exchange of oxygen between capillaries and cells. Nowadays, it has been shown that hypoxia plays an important role in the genesis of several human pathologies including cardiovascular, renal, myocardial and cerebral diseases in fetal, young and adult life.

Several mechanisms have evolved to maintain oxygen homeostasis. Certainly, all cells respond and adapt to hypoxia, but only a few of them can detect hypoxia and initiate a cascade of signals intended to produce a functional systemic response. In mammals, oxygen detection mechanisms have been extensively studied in erythropoietin-producing cells, chromaffin cells, bulbar and cortical neurons, pulmonary neuroepithelial cells, smooth muscle cells of pulmonary arteries, and chemoreceptor cells. While the precise mechanism underpinning oxygen, sensing is not completely known several molecular entities have been proposed as possible oxygen sensors (i.e. Hem proteins, ion channels, NADPH oxidase, mitochondrial cytochrome oxidase). Remarkably, cellular adaptation to hypoxia is mediated by the master oxygen-sensitive transcription factor, hypoxia-inducible factor-1, which can induce up-regulation of different genes to cope the cellular effects related to a decrease in oxygen levels. Short-term responses to hypoxia included mainly chemoreceptor-mediated reflex ventilatory and hemodynamic adaptations to manage the low oxygen concentration while more prolonged exposures to hypoxia can elicit more sustained physiological responses including switch from aerobic to anaerobic metabolism, vascularization, and enhancement of blood O2 carrying capacity.

The focus of this research topic is to provide an up-to-date vision on the current knowledge on oxygen sensing mechanism, physiological responses to acute or chronic hypoxia and cellular/tissue/organ adaptations to hypoxic environment.

Citation: Iturriaga, R., Rio, R. D., R-Richalet, J.-P., eds. (2020). Physiological and Pathological Responses to Hypoxia and High Altitude. Lausanne: Frontiers Media SA. doi: 10.3389/978-2-88963-800-0

# Table of Contents


He Huang, Bao Liu, Gang Wu, Gang Xu, Bing-Da Sun and Yu-Qi Gao


Ryan L. Sheppard, Joshua M. Swift, Aaron Hall and Richard T. Mahon

*44 Arachidonic Acid Metabolism Pathway is Not Only Dominant in Metabolic Modulation but Associated With Phenotypic Variation After Acute Hypoxia Exposure*

Chang Liu, Bao Liu, Lu Liu, Er-Long Zhang, Bind-da Sun, Gang Xu, Jian Chen and Yu-qi Gao


David Morales-Alamo, Borja Guerra, Alfredo Santana, Marcos Martin-Rincon, Miriam Gelabert-Rebato, Cecilia Dorado and José A. L. Calbet


Roberto V. Reyes, Sebastián Castillo-Galán, Ismael Hernandez, Emilio A. Herrera, Germán Ebensperger and Aníbal J. Llanos

*136 Is Carotid Body Physiological O2 Sensitivity Determined by a Unique Mitochondrial Phenotype?*

Andrew P. Holmes, Clare J. Ray, Andrew M. Coney and Prem Kumar


Matiram Pun, Sara E. Hartmann, Michael Furian, Adrienna M. Dyck, Lara Muralt, Mona Lichtblau, Patrick R. Bader, Jean M. Rawling, Silvia Ulrich, Konrad E. Bloch and Marc J. Poulin

*167 Guinea Pig as a Model to Study the Carotid Body Mediated Chronic Intermittent Hypoxia Effects*

Inmaculada Docio, Elena Olea, Jesus Prieto-LLoret, Teresa Gallego-Martin, Ana Obeso, Angela Gomez-Niño and Asuncion Rocher

*179 AMPK-*a*1 or AMPK-*a*2 Deletion in Smooth Muscles Does Not Affect the Hypoxic Ventilatory Response or Systemic Arterial Blood Pressure Regulation During Hypoxia*

Sandy MacMillan and A. Mark Evans

*187 Receptor–Receptor Interactions of G Protein-Coupled Receptors in the Carotid Body: A Working Hypothesis*

Andrea Porzionato, Elena Stocco, Diego Guidolin, Luigi Agnati, Veronica Macchi and Raffaele De Caro

*204 Postural Control in Lowlanders With COPD Traveling to 3100 m: Data From a Randomized Trial Evaluating the Effect of Preventive Dexamethasone Treatment*

Lara Muralt, Michael Furian, Mona Lichtblau, Sayaka S. Aeschbacher, Ross A. Clark, Bermet Estebesova, Ulan Sheraliev, Nuriddin Marazhapov, Batyr Osmonov, Maya Bisang, Stefanie Ulrich, Tsogyal D. Latshang, Silvia Ulrich, Talant M. Sooronbaev and Konrad E. Bloch

*213 Long-Term Chronic Intermittent Hypobaric Hypoxia Induces Glucose Transporter (GLUT4) Translocation Through AMP-Activated Protein Kinase (AMPK) in the Soleus Muscle in Lean Rats*

Patricia Siques, Julio Brito, Karen Flores, Stefany Ordenes, Karem Arriaza, Eduardo Pena, Fabiola León-Velarde, Ángel L. López de Pablo, M. C. Gonzalez and Silvia Arribas

*223 Melatonin Relations With Respiratory Quotient Weaken on Acute Exposure to High Altitude*

Marcelo Tapia, Cristian Wulff-Zottele, Nicole De Gregorio, Morin Lang, Héctor Varela, María Josefa Serón-Ferré, Ennio A. Vivaldi, Oscar F. Araneda, Juan Silva-Urra, Hanns-Christian Gunga and Claus Behn


Ginés Viscor, Joan R. Torrella, Luisa Corral, Antoni Ricart, Casimiro Javierre, Teresa Pages and Josep L. Ventura


Matiram Pun, Veronica Guadagni, Kaitlyn M. Bettauer, Lauren L. Drogos, Julie Aitken, Sara E. Hartmann, Michael Furian, Lara Muralt, Mona Lichtblau, Patrick R. Bader, Jean M. Rawling, Andrea B. Protzner, Silvia Ulrich, Konrad E. Bloch, Barry Giesbrecht and Marc J. Poulin

*294 Carotid Body Type-I Cells Under Chronic Sustained Hypoxia: Focus on Metabolism and Membrane Excitability*

Raúl Pulgar-Sepúlveda, Rodrigo Varas, Rodrigo Iturriaga, Rodrigo Del Rio and Fernando C. Ortiz

*302 Daily Intermittent Normobaric Hypoxia Over 2 Weeks Reduces BDNF Plasma Levels in Young Adults – A Randomized Controlled Feasibility Study*

Andreas Becke, Patrick Müller, Milos Dordevic, Volkmar Lessmann, Tanja Brigadski and Notger G. Müller

*309 Effects of Plyometric Training on Explosive and Endurance Performance at Sea Level and at High Altitude*

David Cristóbal Andrade, Ana Rosa Beltrán, Cristian Labarca-Valenzuela, Oscar Manzo-Botarelli, Erwin Trujillo, Patricio Otero-Farias, Cristian Álvarez, Antonio Garcia-Hermoso, Camilo Toledo, Rodrigo Del Rio, Juan Silva-Urra and Rodrigo Ramírez-Campillo

*318 Ventilatory and Autonomic Regulation in Sleep Apnea Syndrome: A Potential Protective Role for Erythropoietin?*

David C. Andrade, Liasmine Haine, Camilo Toledo, Hugo S. Diaz, Rodrigo A. Quintanilla, Noah J. Marcus, Rodrigo Iturriaga, Jean-Paul Richalet, Nicolas Voituron and Rodrigo Del Rio

*326 Intrapartum Fetal Heart Rate: A Possible Predictor of Neonatal Acidemia and APGAR Score*

Thâmila Kamila de Souza Medeiros, Mirela Dobre, Daniela Monteiro Baptista da Silva, Andrei Brateanu, Ovidiu Constantin Baltatu and Luciana Aparecida Campos


Brina Snyder, Phong Duong, Mavis Tenkorang, E. Nicole Wilson and Rebecca L. Cunningham

*350 Imbalance in Renal Vasoactive Enzymes Induced by Mild Hypoxia: Angiotensin-Converting Enzyme Increases While Neutral Endopeptidase Decreases*

Carlos P. Vio, Daniela Salas, Carlos Cespedes, Jessica Diaz-Elizondo, Natalia Mendez, Julio Alcayaga and Rodrigo Iturriaga

*364 Acute Mountain Sickness is Associated With a High Ratio of Endogenous Testosterone to Estradiol After High-Altitude Exposure at 3,700 m in Young Chinese Men*

Xiao-Han Ding, Yanchun Wang, Bin Cui, Jun Qin, Ji-Hang Zhang, Rong-Sheng Rao, Shi-Yong Yu, Xiao-Hui Zhao and Lan Huang

### *375 Circulating Apoptotic Signals During Acute and Chronic Exposure to High Altitude in Kyrgyz Population*

Djuro Kosanovic, Simon Maximilian Platzek, Aleksandar Petrovic, Akylbek Sydykov, Abdirashit Maripov, Argen Mamazhakypov, Meerim Sartmyrzaeva, Kubatbek Muratali Uulu, Meerim Cholponbaeva, Aidana Toktosunova, Nazgul Omurzakova, Melis Duishobaev, Christina Vroom, Oleg Pak, Norbert Weissmann, Hossein Ardeschir Ghofrani, Akpay Sarybaev and Ralph Theo Schermuly

# Chronic Hypobaric Hypoxia Modulates Primary Cilia Differently in Adult and Fetal Ovine Kidneys

Kiumars Shamloo<sup>1</sup> , Juan Chen<sup>1</sup> , Jasmine Sardar <sup>1</sup> , Rinzhin T. Sherpa<sup>1</sup> , Rajasekharreddy Pala<sup>1</sup> , Kimberly F. Atkinson<sup>1</sup> , William J. Pearce<sup>2</sup> , Lubo Zhang<sup>2</sup> and Surya M. Nauli 1, 3 \*

*<sup>1</sup> Department of Biomedical and Pharmaceutical Sciences, Chapman University, Irvine, CA, United States, <sup>2</sup> Departments of Basic Sciences, Physiology and Pharmacology, Lawrence D. Longo MD Center for Perinatal Biology, Loma Linda University School of Medicine, Loma Linda, CA, United States, <sup>3</sup> Division of Nephrology and Hypertension, Department of Medicine, University of California, Irvine, Irvine, CA, United States*

### Edited by:

*Rodrigo Del Rio, Universidad Autonoma De Chile, Chile*

### Reviewed by:

*Roger Evans, Monash University, Australia Bruno Vogt, University of Bern, Switzerland*

> \*Correspondence: *Surya M. Nauli nauli@chapman.edu; snauli@uci.edu*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *29 June 2017* Accepted: *24 August 2017* Published: *20 September 2017*

### Citation:

*Shamloo K, Chen J, Sardar J, Sherpa RT, Pala R, Atkinson KF, Pearce WJ, Zhang L and Nauli SM (2017) Chronic Hypobaric Hypoxia Modulates Primary Cilia Differently in Adult and Fetal Ovine Kidneys. Front. Physiol. 8:677. doi: 10.3389/fphys.2017.00677* Hypoxic environments at high altitude have significant effects on kidney injury. Following injury, renal primary cilia display length alterations. Primary cilia are mechanosensory organelles that regulate tubular architecture. The effect of hypoxia on cilia length is still controversial in cultured cells, and no corresponding *in vivo* study exists. Using fetal and adult sheep, we here study the effect of chronic hypobaric hypoxia on the renal injury, intracellular calcium signaling and the relationship between cilia length and cilia function. Our results show that although long-term hypoxia induces renal fibrosis in both fetal and adult kidneys, fetal kidneys are more susceptible to hypoxia-induced renal injury. Unlike hypoxic adult kidneys, hypoxic fetal kidneys are characterized by interstitial edema, tubular disparition and atrophy. We also noted that there is an increase in the cilia length as well as an increase in the cilia function in the hypoxic fetal proximal and distal collecting epithelia. Hypoxia, however, has no significant effect on primary cilia in the adult kidneys. Increased cilia length is also associated with greater flow-induced intracellular calcium signaling in renal epithelial cells from hypoxic fetuses. Our studies suggest that while hypoxia causes renal fibrosis in both adult and fetal kidneys, hypoxia-induced alteration in cilia length and function are specific to more severe renal injuries in fetal hypoxic kidneys.

Keywords: cilium, kidney injury, mechanosensor, pO2 , sheep, shear-stress, renal fibrosis

# INTRODUCTION

Hypoxic environments in humans can result in the high-altitude renal syndrome (HARS) (Arestegui et al., 2011). HARS is characterized by reduced renal plasma flow (Becker et al., 1957), microalbuminuria (Becker et al., 1957), proteinuria (Jefferson et al., 2002; Chen et al., 2011), hyperuricemia (Jefferson et al., 2002) and systemic hypertension (Jefferson et al., 2002; Gilbert-Kawai et al., 2016). The kidneys have a fundamental role in adaption to regulate body fluids, electrolyte and acid-base homeostasis during acclimatization to high altitude and in mountain sickness syndromes. Kidneys also respond to hypoxic diuresis and natriuresis through inhibition of renal tubular sodium reabsorption, in addition to erythropoietin production (Haditsch et al., 2015).

Kidneys are very susceptible to lowered oxygen tensions (Mayer, 2011; Fu et al., 2016). In acute hypoxia, kidneys exhibit defense mechanisms centered around the activation of hypoxia-inducible

**7**

factor (HIF) (Minet et al., 2000; Rosenberger et al., 2006). HIF up-regulates pro-angiogenic factors that protect against capillary rarefaction. HIF also increases expression of matrix metalloproteinases that effect repair and protect against fibrosis by degrading extracellular matrix. In chronic hypoxia, however, HIF mRNA is destabilized, shutting-down the kidney's initial protective mechanisms (Uchida et al., 2004). The expression of pro-angiogenic factors is repressed, and the expression of dysangiogenic factors is increased. This may thus result in capillary rarefaction and potential fibrosis.

Primary cilia are sensory organelles that sense extracellular milieu along the tubule of a nephron. Abnormal cilia length and function result in polycystic kidneys (Nauli et al., 2003, 2006; Xu et al., 2007). Renal cilia also display length and function alterations following kidney injury. Primary cilia play a crucial role in cisplatin-induced tubular apoptosis by regulating tubular apoptotic pathway (Wang et al., 2013). Primary cilia also undergo abnormal changes in renal transplant biopsies with acute tubular injury (Hayek et al., 2013). During injury, renal cilia decrease in length in both humans and mice (Verghese et al., 2008, 2011). The shortened cilia length serves to attenuate ciliamediated signaling pathways in response to extracellular stress, which promotes infiltration of neutrophils and macrophages in the kidney (Prodromou et al., 2012). However, the effects of chronic hypobaric hypoxia on primary cilia have yet to be investigated.

Changes in fluid-shear stress generated by urine movement can also contribute to kidney injury if the primary cilia are not functioning normally. Thus, the change in urinary flow associated with nephropathies has been proposed to be a potential insult for tubular cells leading to disorganization of the tubular epithelium (Maggiorani et al., 2015).

Based on the evidence above, it has been proposed that cilia play important roles in sensing environmental cues caused by injury and in the repair process for reestablishing a new epithelial layer of differentiated cells (Bell et al., 2011; Wang and Dong, 2013). By using a series of human renal transplant biopsies, it was found that acute tubular necrosis is associated with more than two-fold longer cilia 1-week after kidney injury, and normalization of cilium length occurred at a later stage. These results indicate that cilia function could be a clinically relevant indicator of kidney injury and repair in patients with kidney transplantation (Wang and Dong, 2013).

Despite the fact that hypoxic environment at high altitude may have significant effects on the kidneys, the histological analysis of renal architecture has never been examined especially in the fetus. Given the background information, our premises indicate that hypoxia causes renal injury/damage and that renal injury often includes altered cilia structure/function. Thus, our hypothesis is that hypoxia causes altered cilia structure/function. Here, we examined the effects of hypoxic high-altitude on the kidneys in adult and fetal sheep. We also took this opportunity to study cilia length-function relationship in response to chronic hypobaric hypoxia, because it has been reported that HIF could alter the structural length of primary cilia in hypoxic in vitro models (Verghese et al., 2011; Ding et al., 2015; Lavagnino et al., 2016; Resnick, 2016).

# METHODS

The sheep were purchased from Nebeker Ranch, Inc. (Lancaster, CA). The normoxic control animals were maintained at sea level throughout the gestation period (300 m). To induce chronic high-altitude hypoxia, the animals were then transported at 30 days of gestation to the Barcroft Laboratory, White Mountain Research Station at Bishop, CA (3,801 m). The hypoxic animals stayed in Bishop for 110 days. Both normoxic control and hypoxic animals were fed Alfalfa hay in a "keyhole" feeder in addition to providing a mineral block ad libitum. The animals were transported from Bishop to the laboratory immediately before the studies (Dasgupta et al., 2012; Thorpe et al., 2013).

Two-years-old pregnant sheep and near-term 146 gestational day fetal lambs were used. Please note that sheep gestation typically lasts ∼150 days. The sheep were anesthetized with thiamylal (10 mg/kg, i.v.) followed by inhalation of 1.5– 2.0% halothane. An incision was made in the abdomen, and kidneys were isolated and immediately placed in either 10% buffer formalin or the phosphate-buffered sucrose solution (30% sucrose, 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 KH2PO4, 1 mM CaCl2•2H2O and 0.5 mM MgCl2•6H2O at pH 7.4). All procedures and protocols were conducted in full compliance with the Animal Welfare Act, followed the guidelines by the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and were approved by the Institutional Animal Care and Use Committee at Loma Linda University, CA.

## Tissue Processing

Kidney cross-sections of the cortex and medulla were dehydrated with isopropyl alcohol and infiltrated with paraffin in a tissue processor (Excelsior AS, Thermo Scientific, Inc.). Tissues were allowed to solidify in a base mold and then snapped into plastic tissue cassettes using a paraffin molding processor (HistoStar, Thermo Scientific, Inc.). Samples were cut to 5 µm thickness sections with a microtome (HM 355 S, Thermo Scientific, Inc.), and continuous "ribbon" of the sections were formed. Sections were carefully transferred onto a 40◦C water dish for about 3–5 min to flatten and avoid wrinkles in the sectioned tissues.

Tissue slices were placed on the charged microscope slides, and deparaffinization was performed by placing the slides in a 60◦C oven for 30 min in order to melt any extra paraffin. Slides were rinsed twice with a xylene solution for 5 min each, followed by hydration through a series of washing for 1 min each in decreasing alcohol concentrations (100, 95, 80, 70%). Slides were then submerged in the distilled water for 3 min for the staining process.

# Tissue Staining

For H&E staining, tissues were stained with hematoxylin for 5 min, rinsed with tap water for 1 min and dipped for 1 sec in acid alcohol (200 ml of 50% alcohol containing 500 µL HCl 5N). Tissues were then stained with Eosin for 2 min and rinsed with distilled water for 1 min followed by dehydration with 95% and 100% alcohol each for 1 min and cleaning with xylene for 2 min. Coverslip was mounted onto a slide with xylene compatible mounting media.

For periodic acid Schiffs (PAS) staining, staining kit (Polysciences, Inc.) was used and experiments were performed based on the Company's protocol. Briefly, tissues were oxidized in 0.5% periodic acid for 5 min. After washing with deionized (DI) water, tissue sections were placed in Schiff's reagent for 15 min then rinsed with 0.55% potassium metabisulfite. Slides were washed in tap water for 10 min to allow the color to develop in the tissues, which were then counterstained with acidified Harris Hematoxylin for 30 s. Slides were washed with tap water until the tissues turned to a blue color. After dehydration with 95 and 100% alcohol each for 1 min and clearing with xylene for 2 min, the coverslips were mounted onto the slides with a xylene compatible mounting media.

For Masson's Trichrome staining, a staining kit (Polysciences, Inc.) was used according to the manufacture's instructions. Briefly, tissues were incubated in Bouin's fixative solution at 60◦C for 1 h and rinsed with warm tap water for 2–3 min until the yellow color disappeared. Fresh Weigert's iron hematoxylin working solution was prepared by mixing Weigert's hematoxylin A and Weigert's hematoxylin B in a 1:1 ratio. Tissues were incubated in the working solution for 20 min and rinsed for 5 min with distilled water. Tissues were then stained with Biebrich Scarlet-Acid Fuchsin solution for 15 min and rinsed with tap water to remove any extra solution-residue. After rinsing with distilled water, the tissues were soaked in Phosphotungstic / Phosphomolibdic acid for 10 min, transferred to Aniline Blue for 7 min and washed with distilled water until the slides were clear. Tissues were then soaked in 1% acetic acid for 1 min and washed with distilled water for 1 min. After dehydration with 95 and 100% alcohol each for 1 min followed by cleaning with xylene for 2 min, the coverslips were mounted onto the slides with a xylene compatible mounting media.

For immunofluorescence staining, heat-induced epitope retrieval was performed using a pressure cooker and Tris-EDTA buffer (10 mM Tris base, 1 mM EDTA solution, 0.05% Tween 20, pH 9.0). Tris-EDTA buffer was first boiled in the pressure cooker to 100◦C. Tissues were then maintained in the pressure cooker at about 121◦C and 15 psi for 8 min. The pressure cooker and slides were then chilled quickly with cold tap water. The slides were rinsed three times with phosphate buffer saline (PBS) and blocked in PBS solution containing 5% bovine serum albumin for 15 min in a humidifying chamber. Using our standard protocol (Loghman-Adham et al., 2003; Nauli et al., 2003, 2006), tissues were stained with the ciliary marker acetylated-α-tubulin (1:1,000 dilution; Sigma-Aldrich, Inc.), distal tubular marker bolichos biflorus agglutinin (DBA; 4:1,000 dilution; Vector Laboratory, Inc.), proximal tubular marker lotus tetragonolobusl lectin (LTL; 4:1,000 dilution; Vector Laboratory, Inc.), and nucleus marker DAPI (Vector Laboratory, Inc.).

# Primary Cell Culture

Isolation of renal epithelia from fresh tissues has been previously described in detail (Loghman-Adham et al., 2003; Nauli et al., 2003, 2006). After extensive washing in phosphate-buffered sucrose solution, the cortex and medullary regions of kidney were dissected into pieces and incubated with collagenase followed by 0.05% trypsin/0.53 mM EDTA at 37◦C for 15–20 min. After vortexing vigorously for 5 min, ice-cold HBSS containing 10% FBS was added to inactivate the trypsin. The cells were further released from the fibrous basement membrane by trituration, washed twice with HBSS. Cells were then sorted based on their surface markers for DBA or LTL and resuspended in fresh culture medium. To allow attachment, the cells was grown in DMEM containing 10% FBS and supplemented with 5 µg/ml insulin, 5 µg/ml transferrin, 5 ng/ml selenium, 36 ng/ml (10−<sup>7</sup> M) hydrocortisone, 10−<sup>8</sup> M triiodothyronine, 10 ng/ml EGF, and 50 ng/ml PGE1, as well as 100 U/ml penicillin, and 100 µg/ml streptomycin in a 37◦C humidified incubator ventilated with 5% CO2—95% O2. The culture medium was changed every 2–3 days until confluency was reached.

### Cilia Function and Length Measurements

Bending primary cilia with fluid-flow activates the cilium and increases cytoplasmic calcium in renal epithelial cells (Praetorius and Spring, 2001; Liu et al., 2005; Jin et al., 2014). Thus, intracellular calcium was measured with Fura2-AM [to study cilia function] as previously described (Nauli et al., 2013). Briefly, after pre-incubation with 5 µM Fura2-AM for 30 min at 37◦C, the tissues were equilibrated for at least a minute. The optimal shear-stress of 0.8 dyne/cm<sup>2</sup> was used to monitor changes in cytosolic calcium as previously described (Nauli et al., 2013). Ionomycin (1 µM) was used at the end of each experiment to confirm Fura-2 loading and determine minimal and maximal ratiometric fluorescence signals. Cells were then analyzed for cilia length by staining with the ciliary marker acetylated-αtubulin by staining with the ciliary marker acetylated-α-tubulin as previously described (Loghman-Adham et al., 2003; Nauli et al., 2003, 2006).

### Data Analysis

All images represent the extracted information from the digital pictures. Images were captured through a Nikon Ti-E microscope. The Nikon NIS Elements for Advanced Research software was used for image capture and analysis, including automatic object recognition, image scanning and color binary segmentation. Scale bars were provided in all figures to indicate the actual image reduction size. All morphometric data were reported as mean ± SEM. Distribution analyses were performed and presented on all data sets to verify normal data distributions. After distribution and variance analyses, data comparisons for more than two groups were performed using ANOVA test followed by a Tukey post-hoc analysis. The correlation between cilia length and cilia function was analyzed with the ordinary least squares (OLS) regression of y on x, because the ordinary least products (OLP) would not allow analysis of covariance as a technique for testing the equality of slopes or intercepts in linear regressions (Ludbrook, 2012). Pearson correlation coefficients were therefore used to analyze the significance of differences in the coefficients correlating primary cilia length and function. Asterisks (<sup>∗</sup> ) denote with statistical significance differences at p < 0.05 relative to corresponding control groups as indicated in the figures.

A total of 12 sheep and 12 lambs was used; for each normoxic or hypoxic group, 6 sheep and 6 lambs were used (N = 6 sheep and N = 6 lambs for normoxia; N = 6 sheep and N = 6 lambs for hypoxia). From each animal, both kidneys were collected for different studies. Each pair of kidneys were randomly selected for an immediate fixation in formalin or trypsinization for cell culture. From each kidney, a minimum of 10 experimental replicates was sampled. All statistical analysis was done with GraphPad Prism, version 5.0b.

### RESULTS

The normoxic control animals had maternal and fetal arterial PO<sup>2</sup> (PaO2) of 102 ± 2 and 25 ± 1 mmHg, respectively. The hypoxic maternal and hypoxic fetal PaO<sup>2</sup> were 60 ± 2 and 19 ± 1 mmHg, respectively.

### Fetal Hypoxic Kidneys Were Characterized by Interstitial Medullary Edema

Representative images of H&E stained renal tissue sections from normoxic and hypoxic kidneys are presented (**Figure 1A**). In all kidney sections, light microscopy analysis indicated that the cortex regions of the kidneys had normal glomeruli. There were no significant differences in glomerulus size between normoxic and hypoxic kidneys (**Figure 1B**). Capillary loops of the glomeruli were also normal, and the number and morphology of endothelial cells were comparable between normoxic and hypoxic tissues.

In normoxic fetal kidneys, all tubules were closely spaced in the interstitium. Tubules were lined with a single layer of epithelia and well-organized nuclei. In hypoxic fetuses, the renal cortex demonstrated some evidence of mesangial and intercapillary cell proliferation. The renal medulla showed apparent atrophic tubules, interstitial fibrosis and edema, and tubular disparition characterized by greater inter-tubular spaces and separation of tubules filled with edema (**Figure 1C**).

While all tubules in adult sheep were closely spaced in the interstitium and lined by a single layer of cells with wellorganized nuclei, some abnormalities were observed in the chronically hypoxic kidney. The hypoxic kidneys showed some proliferation of mesangial cells in the cortex and slight tubular edema in the medullary region. Otherwise, the surrounding tubules appeared normal in both normoxic and hypoxic adult kidneys.

### Hypoxia Modulated Medullary Tubular Basement Membrane Thickness

PAS staining was done to highlight basement membranes of glomerular capillary loops and tubular epithelia. Renal tissue sections from normoxic and hypoxic kidneys are shown in the representative images (**Figure 2A**). There were no significant differences in basement membranes thicknesses of glomerular capillary loops between normoxic and hypoxic adult kidneys.

Basement membrane thicknesses of the tubules were significantly less in hypoxic fetal kidneys compared to normoxic fetal kidneys (**Figure 2B**). Although a reverse trend was

FIGURE 1 | Haemotoxylin and Eosin (H&E) staining to study effects of long-term hypoxia. (A) Kidneys were stained with H&E, and representative images were taken at the cortex and medullary regions of the kidneys. (B) Glomerular diameters were measured and quantified. (C) Representative images show that in fetal hypoxic kidney, tubules are separated by interstitial medullary edema as denoted by asterisks. *N* = 6 animals for each group; statistical analysis was performed with ANOVA followed by the Tukey *post-hoc* test.

observed in adult kidneys, it was not statistically significant in adult normoxic or hypoxic kidneys. To refine the statistical analysis, tubular membrane thicknesses were re-measured by discriminating the tubular origins. In proximal tubules, no significant difference in tubular membrane thickness was observed (**Figure 2C**). In distal collecting tubules, tubular membranes were significantly thicker in hypoxic than normoxic adult kidneys. As expected, tubular membrane thicknesses were significantly less in hypoxic compared to normoxic fetal kidneys.

As observed in the H&E staining, hypoxic fetal kidneys showed interstitial medullary edema, fibrosis, tubular disparition and atrophy characterized by greater inter-tubular space (**Figure 2D**).

### Long-Term Hypoxia Induced Renal Fibrosis

To examine potential fibrosis in the kidneys, Masson's Trichrome staining was performed to highlight deposition of collagen and fibrin fibers. Renal tissue sections from normoxic and hypoxic kidneys are shown in representative images (**Figure 3A**). The percentage of blue-stained area was calculated using a binary threshold program, which indicated significant fibrosis in hypoxic compared to normoxic kidney tissues (**Figure 3B**). In normoxic tissues, there was some collagen fibers that were normally distributed around renal tubules. In hypoxic tissues, normal distributions of collagen fibers were observed, but abnormal depositions of collagen/fibrin fibers were also observed throughout tubular epithelia and interstitial spaces.

As observed in the H&E and PAS staining, hypoxic fetal kidney showed interstitial medullary edema characterized by more inter-tubular space (**Figure 3C**). As also seen in previous staining, hypoxia modulated lumen size of distal tubules in both fetal and adult kidneys (**Figure 3D**). No significant changes in lumen size were observed in proximal tubules.

### Fetal Hypoxic Kidney Was Characterized by Longer Renal Epithelial Primary Cilia

It remains uncertain if hypoxia can maintain and stabilize the primary cilia or it would inhibit primary cilia formation (Verghese et al., 2011; Ding et al., 2015; Lavagnino et al., 2016; Resnick, 2016). To investigate the effect of chronic hypoxia, primary cilia were labeled with the cilia marker acetylated-αtubulin. Representative images of renal tissue sections from normoxic and hypoxic kidneys are shown for staining of cilia and tubular markers. Distal collecting (**Figure 4A**) and proximal (**Figure 4B**) tubular markers were used to identify cilia length in respective tubules. Cilia measurements were the represented in the bar graphs to compare the distributions of the cilia length (**Figure 4C**). While hypoxia did not significantly alter cilia length in adult kidneys, fetal kidneys were very susceptible to hypoxia (**Figure 4D**). Cilia length was significantly longer in hypoxic fetal kidneys than normoxic fetal kidneys.

Of note, that the distal collecting tubules had larger lumen sizes in adult than fetal kidneys. This may be associated with longer cilia in adult than fetal distal collecting tubules. For the first time, our studies show that the length of primary cilia in distal collecting tubules become longer during maturation, while

FIGURE 2 | Periodic acid-Schiff (PAS) staining to study effects of long-term hypoxia. (A) Kidneys were stained with PAS, and representative images were taken at cortex and medullary regions of the kidneys. (B) Tubular basement membrane thickness was measured and quantified. (C) More detail analysis of basement membrane was achieved by separation between distal collecting and proximal tubules. (D) Representative images show that in fetal hypoxic kidney, tubules were separated by interstitial medullary edema as denoted by red asterisks. \*Indicates significant differences between normoxic and hypoxic groups. *N* = 6 animals for each group; statistical analysis was performed with ANOVA followed by the Tukey *post-hoc* test.

FIGURE 3 | Masson's Trichrome staining to study effects of long-term hypoxia. (A) Kidneys were stained with Masson's Trichrome, and representative images were taken at cortex and medullary regions of the kidneys. (B) Interstitial renal fibrosis was measured and quantified. (C) Representative images show that in fetal hypoxic kidney, tubules were separated by interstitial medullary edema as denoted by red asterisks. (D) Tubular areas were measured, quantified and compared. \*Indicates significant differences between normoxic and hypoxic groups. *N* = 6 animals for each group; statistical analysis was performed with ANOVA followed by the Tukey *post-hoc* test.

hypoxia on cilia length *in vivo.* (A) Kidneys were stained with ciliary marker (acetylated-α-tubulin; green), distal collecting marker (DBA; red) and nucleus marker (DAPI; blue). (B) Instead of DBA, kidneys were stained with proximal tubular marker (LTL; red). White boxes show enlargement of the images to *(Continued)*

### FIGURE 4 | Continued

depict the magnified view of primary cilia. (C) The length of primary cilia was measured and represented in the bar graph to depict length distribution within each group. (D) Cilia length was quantified. \*Indicates significant differences between normoxic and hypoxic groups. *N* = 6 animals for each group; statistical analysis was performed with ANOVA followed by the Tukey *post-hoc* test.

there is no change in proximal tubule cilia length from fetuses to adults.

# Fetal Hypoxic Kidney Had Greater Sensitivity in Response to Fluid-Shear Stress

Renal primary cilia are mechanosensory organelles that sense filtrate moving within the tubules. To examine the possibility that hypoxia could alter mechanosensory of cellular responses, renal epithelia were isolated, cultured, stained with a cilia marker and challenged with shear-stress. Representative images of these cells from normoxic and hypoxic kidneys reveal changes in cilia length (**Figure 5A**). On average about 80% of cells had cilia, and there were no apparent differences in cilia formation between normoxic and hypoxic tissues, or between fetal and adult kidneys. Cilia measurements were then tabulated to compare their distributions (**Figure 5B**). Cilia measurements were also tabulated to analyze the impact of hypoxia on cilia lengths (**Figure 5C**). While hypoxia did not significantly alter cilia length in adult cells, cilia length was significantly longer in hypoxic than normoxic fetal cells. Interestingly, trend in cilia length changes was similar for in vitro and in vivo preparations. When the overall cilia length in vivo kidney (**Figure 4**) or in vitro cell culture (**Figure 5**) were averaged, it was apparent that the in vitro cell culture produced much longer cilia than observed in vivo (6.35 ± 0.28 µm vs. 2.84 ± 0.12 µm; p < 0.00005).

To examine the effect of chronic hypoxia on mechanosensory cilia function, cells were challenged with 0.8 dyne/cm<sup>2</sup> shearstress. Changes in cytosolic calcium were averaged and plotted in line graphs (**Figure 6A**). When peaks of cytosolic calcium were examined, hypoxia significantly enhanced mechanosensory function in fetal epithelial cells (**Figure 6B**). In contrast, hypoxia did not significantly alter mechanosensory sensitivity in adult cells.

### Hypoxia-Induced Longer Cilia Were Associated with Greater Cilia Function

It has been hypothesized that longer cilia are more sensitive to fluid-shear stress (Abdul-Majeed et al., 2012; Upadhyay et al., 2014). To examine the effect of hypoxia on cilia length-function relationship, both in vivo (**Figure 4**) and in vitro (**Figure 5**) cilia lengths were used to assess changes in cilia function. When in vivo cilia length was plotted against cilia function, no apparent length-function association was observed (**Figure 7A;** R <sup>2</sup> = 0.42). Because hypoxia did not alter cilia length in adult kidneys, length-function relationship was next analyzed only in

FIGURE 5 | Immunofluorescence staining to study effects of long-term hypoxia on cilia length in *in vitro* cultures. (A) Ovine renal epithelia were isolated and stained with ciliary marker (acetylated-α-tubulin; green) and nucleus marker (DAPI; blue). White boxes show enlargement of the images to depict the presence of primary cilia. (B) The length of primary cilia was measured and tabulated in the bar graph to depict length distribution within each group. (C) Cilia length was quantified. \*Indicates significant differences between normoxic and hypoxic groups. *N* = 6 animals for each group; statistical analysis was performed with ANOVA followed by the Tukey *post-hoc* test.

fetal kidneys. This revealed significant correlation within lengthfunction relationships (**Figure 7B**; R <sup>2</sup> = 0.86). When in vitro cilia length was analyzed, the cilia function was significantly correlated with cilia length following either inclusion (**Figure 7C**; R 2 = 0.79) or exclusion (**Figure 7D**; R <sup>2</sup> = 0.93) of adult kidneys.

### DISCUSSION

Although kidneys are important for regulation of body fluid, pH, electrolyte, hormone and overall metabolism, the effects of chronic hypobaric hypoxia on fetal kidneys have never been examined. We therefore study the effects of chronic hypoxia on fetal and adult sheep kidneys within the cortex and medullary

FIGURE 7 | Correlation between cilia length-function. (A) Cilia length-function relation was plotted in a dot plot from both fetal and adult groups *in vivo*. (B) Because hypoxia affected mainly on fetal renal cilia, the length-function relation showed a greater correlation when plotted only for fetus. (C) Cilia length was next taken from the *in vitro* measurement. Cilia length-function relation was plotted in a dot plot from both fetal and adult groups. (D) Because hypoxia affected mainly on fetal renal cilia, the length-function relation showed a greater correlation when plotted only for fetus. *N* = 6 animals for each group; statistical analysis was performed with the Pearson correlation coefficient test.

regions. Effects of hypoxia on mechanosensory primary cilia are also evaluated. Our studies suggest that: (1) there are no significant differences in glomerulus size between normoxic and hypoxic kidneys; (2) the hypoxic kidneys show proliferation

of mesangial cells, medullary edema and fibrosis; (3) hypoxia modulates changes in tubular basement membranes; and (4) compared to adult kidneys, fetal kidneys are more susceptible to hypoxia-induced tubular disparition and atrophy, characterized by alterations in primary cilia and function.

A previous study has demonstrated enlargement of renal glomeruli in hypoxic children (Naeye, 1965). However, our studies do not support glomerulus enlargement in the hypoxic fetal or adult sheep kidneys. This discrepancy could potentially be due to an age-specific effect. For example, glomerular enlargement only occurs after the first month of life, while the size of glomeruli is normal at birth in those hypoxic children (Naeye, 1965). Since all of these children die as a result of unknown illness, it is also possible that other genetic and environmental factors contributed to the glomerulus enlargement. For example, an oxonic acid diet in rats is known to cause hyperuricemia through elevated plasma renin activity (Eraranta et al., 2008). Glomerular hypertrophy is observed in oxonic acid-induced hyperuricemia and can be prevented by ACE inhibitor therapy in the rats (Nakagawa et al., 2003).

Abnormal mesangial cells, edema and fibrosis are observed in our hypoxic kidneys. This is possibly a result from interstitial renal injury due to chronic hypobaric hypoxia. Of note, the deposition of collagen and fibrin fiber that functions as a byproduct of a reparative process, is commonly used as an index of various renal injury (Ma et al., 1998; Kaissling and Le Hir, 2008; Forbes et al., 2012). Although hypoxia is one of the stimuli that drives chronic kidney disease (Fu et al., 2016), our studies further reveal that hypoxia could alter the thickness of tubular basement membranes. Interestingly, hypoxia only induces changes in basement membranes of distal collecting tubules. This could be due to medullary regions that operate within a relatively lower range of pO2, and it is therefore more susceptible to hypoxic injury than proximal regions (Heyman et al., 1997; El Sabbahy and Vaidya, 2011). Greater susceptibility of medullary regions to hypoxia might thus contribute to an increase in collecting distal tubular basement membrane of adult kidneys. Unlike adult medullary tissues, however, hypoxia actually causes a decrease in medullary basement membrane in fetal kidney. This could possibly be due to tubular disparition and atrophy, which are very apparent in the fetal medulla. The thinning in basement membrane in medullary tissues might therefore be associated with a degeneration of tubular structure, as observed in fetal hypoxic kidneys.

Based on the histological analyses, fetal kidneys are more susceptible to hypoxia, possibly due to a "triplehit hypoxia" phenomenon. In addition to the hypobaric hypoxia in the atmosphere, fetal kidneys are impacted by three additional factors (triple-hit). First, chronic hypoxia induces vasoconstriction in sheep uterine arteries during gestation (Hu et al., 2012; Xiao et al., 2013). This maladaptation of the uteroplacental circulation can result in reduced tissue perfusion and further exacerbate the effects of hypoxia. Second, in hypoxic fetal sheep there is a rapid drop in fetal heart rate and a rise in mean arterial blood pressure that redistributes blood flow to the heart and brain at the expense of the renal circulation. Hypoxia therefore reduces renal blood flow, resulting in renal hypoperfusion (Robillard et al., 1986; Green et al., 1997). Third, the unique fetal circulation system allows oxygen-rich blood from the aorta to mix with oxygen-poor blood before reaching renal circulation (Nakamura et al., 1987). Thus, the fetal circulation system reduces oxygenation support to the kidneys.

Our results also indicate that greater susceptibility to hypoxia in fetal kidneys could be associated with changes in primary cilia length and function throughout the nephrons. Primary cilia are sensory organelles rising from the apical surface of most mammalian cells. Cilia are activated during bending by fluid-flow, which in turn initiates an intracellular calcium response (Praetorius and Spring, 2001; Liu et al., 2005; Nauli et al., 2013). Previous studies demonstrate a link between cilia length and hypoxia-inducible mechanisms. In tendon cells, hypoxia inhibits primary cilia formation and cellular mechano-responsiveness (Lavagnino et al., 2016). However, studies in different hypoxic models using renal epithelial cells indicate that primary cilia are longer and that expression of HIF maintains primary cilia length (Verghese et al., 2011; Ding et al., 2015). Interestingly, the renal epithelial cilia become more flexible during hypoxia (Resnick, 2016), and it thus might alter cilia function. Consistent with this view, a correlation between cilia length and function in response to chronic hypoxia is observed in our studies. Interestingly, greater

cilia length in hypoxic fetal kidney in vivo are maintained as measured in in vitro cell culture. This could be due to epigenetic changes occurring during hypoxia. Of note, long-term hypoxia during gestation has been reported to cause epigenetic adaptation in the sheep (Dasgupta et al., 2012; Chen et al., 2014).

Our studies indicate that chronic hypobaric hypoxia could induce renal injury (**Figure 8**). Renal hypoperfusion as a result from the "triple-hit hypoxia" phenomenon seen in fetus would further induce tubular disparition and atrophy, a process that modulated changes in cilia length and function. This, in turn, induces an increase in intracellular calcium fluxes in response to fluid-shear stress. Our results potentially point to a complex renal injury caused by chronic hypoxia with regards to primary cilia.

It has been reported that kidney compensation induced by a reduction of renal mass results in primary cilia elongation, and this elongation is associated with an increased production of reactive oxygen species (Han et al., 2016). In addition, renal primary cilia lengthens after acute tubular necrosis (Verghese et al., 2009). The contribution of primary cilia from acute to chronic injury is also confirmed when a ciliary protein is inactivated. In this case, renal damages are more pronounced following ischemia/reperfusion, and this induces microcyst formation (Bastos et al., 2009). In addition to the association of human renal carcinoma and primary cilia (Ding et al., 2015), acute injury can induce chronic kidney disease through cyst formation in the absence of normal renal cilia (Patel et al., 2008).

### REFERENCES


Furthermore, mechanosensory function of cilia is abnormal in chronic kidney disease patients (Nauli et al., 2006; Xu et al., 2007). Based on these findings, we speculate that the increases in hypoxia-induced cilia-related calcium signaling in hypoxic fetal kidneys serves a potential mechanism associated with renal injury. Without doubt, future studies on the hypoxia-induced changes in intracellular calcium are warranted.

### AUTHOR CONTRIBUTIONS

KS analyzed data, contributed to drafting the manuscript and oversaw the entire progress. JC and RS performed the calcium studies. JS and RP performed cilia measurements and kidney staining. KA assisted in sample collections. WP and LZ provided the kidney samples. LZ and SMN finalized the manuscript and oversaw the research project.

### FUNDING

The project is funded in part by Congressionally Directed Medical Research Program PR130153 (SMN) and National Institutes of Health Grants HD083132 (LZ).

### ACKNOWLEDGMENTS

Authors would like to acknowledge and thank Alisz Demecs and Maki Takahashi for their editing and technical support.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Shamloo, Chen, Sardar, Sherpa, Pala, Atkinson, Pearce, Zhang and Nauli. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Hematological Risk Factors for High-Altitude Headache in Chinese Men Following Acute Exposure at 3,700 m

He Huang1, 2, 3†, Bao Liu1, 2, 3, 4†, Gang Wu1, 2, 3, Gang Xu1, 2, 3, Bing-Da Sun1, 2, 3 and Yu-Qi Gao1, 2, 3 \*

1 Institute of Medicine and Hygienic Equipment for High Altitude Region, College of High Altitude Military Medicine, Third Military Medical University, Chongqing, China, <sup>2</sup> Key Laboratory of High Altitude Environmental Medicine, Third Military Medical University, Ministry of Education, Chongqing, China, <sup>3</sup> Key Laboratory of High Altitude Medicine, Chinese People's Liberation Army, Chongqing, China, <sup>4</sup> The 12th Hospital of Chinese People's Liberation Army, Kashi Xinjiang, China

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Julio Brito, Arturo Prat University, Chile Antonella Naldini, University of Siena, Italy

> \*Correspondence: Yu-Qi Gao gaoy66@yahoo.com

† These authors have contributed equally to this work.

### Specialty section:

This article was submitted to Vascular Physiology, a section of the journal Frontiers in Physiology

Received: 28 July 2017 Accepted: 29 September 2017 Published: 17 October 2017

### Citation:

Huang H, Liu B, Wu G, Xu G, Sun B-D and Gao Y-Q (2017) Hematological Risk Factors for High-Altitude Headache in Chinese Men Following Acute Exposure at 3,700 m. Front. Physiol. 8:801. doi: 10.3389/fphys.2017.00801 Background: High-altitude headache (HAH) is a notably common disorder affecting the daily life of travelers ascending to high altitude. Hematological parameters are important clinical examinations for various diseases. Today, hematological characteristics of HAH remain unrevealed. Above all, we aimed to ascertain hematological characteristics and independent risk factors/predictors associated with HAH before and after exposure at 3,700 m.

Methods: Forty five healthy men were enrolled in present study. Demographic and clinical data, physiological and hematological parameters were collected 3 days before the ascent and after acute exposure at 3,700 m.

Results: HAH patients featured significantly lower white blood cell count (WBC), neutrophil count (NEU#) and percentage (NEU%), and higher percentage of lymphocyte (LYM%) at 3,700 m and significantly lower NEU#, reticulocyte count (RET#) and percentage (RET%) at sea level (all P < 0.05). HAH severity was significantly and negatively associated with WBC, NEU#, and NEU% at 3,700 m and RET# at sea level, whereas was positively associated with LYM% at 3,700 m (all P < 0.05). Moreover, we have found that RET# at sea level and NEU% at 3,700 m was an independent predictor and risk factor for HAH, respectively.

Conclusion: The present study is the first to examine the hematological characteristics of HAH. Furthermore, lower RET# at sea level and lower NEU% at 3,700 m is a novel independent predictor and risk factor for HAH, respectively.

Keywords: headache, high-altitude headache, acute mountain sickness, hematological parameters, risk factor

**Abbreviations:** AMS, Acute mountain sickness; CCL8, C-C Motif Chemokine Ligand 8; CI, confidence interval; DBP, diastolic blood pressure; HAH, high-altitude headache; HAH+, participants with HAH; HAH−, participants without HAH; HR, heart rate; IL-1β, interleukin-1β; IL-6, interleukin-6; IL-10, interleukin-10; SBP, systolic blood pressure; SpO2, pulse oxygen saturation; TNF-α, tumor necrosis factor-α.

# INTRODUCTION

Headache is the fundamental symptom for a diagnosis of acute mountain sickness (AMS) according to the Lake Louise Scoring System (Imray et al., 2010; Bartsch and Swenson, 2013; Davis and Hackett, 2017). In this context, high-altitude headache (HAH) is a defined and recognized headache disorder, which occurs within 24 h after ascending to an altitude of >2,500 m, based on the International Classification of Headache Disorders 3β criteria (Marmura and Hernandez, 2015; Kim et al., 2016). Approximately 80% of people are thought to experience headache after rapidly ascending to high altitudes, which is associated with disturbances in their daily life and work that create an important public health problem (Carod-Artal, 2014).

During recent decades, numerous researchers have systematically investigated the epidemiology, clinical manifestations, pathophysiological mechanisms, risk factors, prevention, and treatment of HAH (Wilson et al., 2009; Carod-Artal, 2014; Marmura and Hernandez, 2015). Clinical studies have demonstrated that HAH ordinarily presents as a sudden attack and partially mimics the features of migraine (Silber et al., 2003; Broessner et al., 2016). Current studies regarding the pathophysiological etiology of HAH have suggested that HAH can be induced by hypobaric hypoxia causing overperfusion of the microvascular beds, brain swelling, activation of the trigeminal vascular system, and sensitization of intracranial pain receptors (Burtscher et al., 2011). Furthermore, the known independent risk factors for HAH include increased heart rate (HR), posterior cerebral circulation, vertebral artery diastolic velocity, and anxiety; decreased pulse oxygen saturation (SpO2) and vigor; and a history of migraine or microRNA dysregulation (Silber et al., 2003; Lawley, 2011; Alizadeh et al., 2012; Bian et al., 2013, 2015; Liu et al., 2016; Dong et al., 2017; Guo et al., 2017). However, the actual underlying mechanisms, clinical manifestations and risk factors/predictors remain not entirely clear, despite accumulating evidence regarding HAH (Marmura and Hernandez, 2015).

Hematological parameters are important clinical markers and can be analyzed using simple, rapid, and inexpensive techniques (Devasena, 2017). Numerous researchers have suggested that hematological changes are closely associated with various diseases, including myelodysplastic syndrome, laryngeal squamous cell cancer, stable coronary artery disease, stroke, and acute pulmonary embolism (Nayak et al., 2011; List et al., 2012; Zorlu et al., 2012; Ayhan et al., 2013; Sahin et al., 2015; Soderholm et al., 2015; Hsueh et al., 2017). Moreover, recent studies have indicated that migraine conditions may be related to some hematological parameters, especially high hemoglobin levels and broad red blood cell distribution width (Celikbilek et al., 2013; Lippi et al., 2014). Thus, hematological changes are an attractive area for improving our understanding of HAH, although we are not aware of any studies that have examined this issue.

The present study was based on the hypothesis that hematological parameters would be closely related to HAH. Therefore, we evaluated the hematological characteristics of participants before and after exposure to high altitude, which may provide useful information regarding HAH-specific hematological phenotypes, independent risk factors, and predictors.

# METHODS

### Participants

The inclusion criteria were healthy 18–60-year-old Chinese men whose primary residence was at an altitude of ≤1,000 m. The exclusion criteria were cardio-cerebrovascular diseases, respiratory diseases, neuropsychological diseases, kidney disorders, liver disorders, migraine or other headaches, history of travel to an altitude of >2,500 m during the last 2 years, and regularly drinking coffee or tea. Based on these criteria, we successfully recruited and enrolled 45 Chinese men who were healthy and 18–35 years old. The study's protocol was approved by the Third Military Medical University Ethics Committee (China), and complied with the requirements of the Declaration of Helsinki. All participants signed informed consent forms before their enrollment.

### Procedures

During the study, all participants rapidly ascended from Chongqing (sea level, altitude of 200 m) to Lhasa (altitude of 3,700 m) by train within 48 h. Three days before the ascent, the participants provided demographic data and underwent measurements of physiological and hematological parameters at sea level. Within 24 h after their arrival at 3,700 m, the participants underwent assessments of their physiological and hematological parameters, as well as HAH. During the study, the participants maintained the same diet (no coffee, tea, or alcoholic drinks) and avoided strenuous activity to ensure a similar level of physical activity. The participants were monitored by physicians for any signs of high-altitude cerebral or pulmonary edema (Hackett and Roach, 2004; Pennardt, 2013). The possibility of emergent cases of these conditions was addressed by planning for immediate evacuation and treatment using oxygen supplementation and medication.

# Demographic and Clinical Data Collection

Three days before their ascent, the participants completed selfadministered questionnaires to obtain their demographic data (body mass index [BMI], age, smoking history, and drinking history). After the acute exposure at 3,700 m, the physicians scored the participants based on the clinical features of HAH: headache (0 = none, 1 = mild, 2 = moderate, 3 = severe), the location of the headache (bilateral, frontal, frontal-temporal, or other), and factors that worsened the headache (exertion, movement, straining, coughing, or other). The diagnosis of HAH was based on fulfilling at least two of the International Classification of Headache Disorders 3β criteria (Marmura and Hernandez, 2015): mild or moderate severity; bilateral location; and aggravation caused by exertion, movement, straining, coughing, and/or bending.

### Measurement of Physiological Parameters

During the pre-exposure and post-exposure examinations, physicians measured the participants' physiological parameters, including HR, SpO2, diastolic blood pressure (DBP), and systolic blood pressure (SBP). The participants rested for 30 min before being evaluated using a sphygmomanometer (HEM-6200; OMRON, China) and a pulse oximeter (NONIN-9550; Nonin Onyx, USA).

### Blood Collection and Hematological Parameter Measurements

Pre-exposure and post-exposure venous blood samples were collected from the participants by qualified nurses using EDTAcoated tubes and standard procedures. The blood samples were stored at 4◦C until the testing using the AU-2700 analyzers (Olympus, Tokyo, Japan) and commercial reagents. The hematological parameters were white blood cell count (WBC), red blood cell count (RBC), hematocrit (HCT), mean corpuscular volume, red cell distribution width and its coefficient of variation, platelet count, mean platelet volume, plateletcrit, platelet distribution width, platelet-large cell ratio, hemoglobin (HGB), mean corpuscular hemoglobin and concentration, neutrophil count (NEU#), reticulocyte count (RET#), lymphocyte count (LYM#), eosinophil count (EOS#), monocyte count (MON#), neutrophil percentage (NEU%), reticulocyte percentage (RET%), lymphocyte percentage (LYM%), eosinophil percentage (EOS%), and monocyte percentage (MON%) (Supplementary Table 1).

# Statistical Analysis

The parameters' normality was assessed using the Shapiro-Wilk's test. Normally distributed data were presented as mean ± standard deviation, non-normally distributed data were presented as median (interquartile range), and categorical data were presented as number (percentage). Differences between the groups with and without HAH (HAH+/HAH−) were compared using the independent t-test for normally distributed data and the Mann-Whitney U-test for non-normally distributed data. Spearman correlation coefficients were evaluated for the relationships between the parameters and HAH severity scores. Univariate logistic regression was used to identify risk factors and predictors, and significant factors from the aforementioned analyses were included in adjusted logistic regression (**Figure 1**). All analyses were performed using IBM SPSS software (version 19; IBM Corp, Armonk, NY, USA), and differences were

considered statistically significant at a P-value of < 0.05. All statistical methods and results were discussed with and guided by statisticians from the Third Military Medical University.

## RESULTS

### Demographic and Clinical Characteristics

All 45 participants completed the pre-exposure and postexposure assessments, and had complete data regarding their physiological and hematological parameters. The incidence of HAH was 51.1%, although no significant differences were observed between the HAH+ and HAH− groups for age, BMI, drinking, and smoking (all P > 0.05, **Table 1**).

### HAH-Related Parameters at Sea Level

Compared to the HAH− group, the HAH+ group had significantly lower pre-exposure values for RET% (0.76 ± 0.25% vs. 0.89 ± 0.22%), RET# (37.93 ± 12.25 10<sup>9</sup> /L vs. 45.41 ± 9.52 10<sup>9</sup> /L), and NEU# (2.94 ± 0.77 10<sup>9</sup> /L vs. 3.62 ± 1.35 10<sup>9</sup> /L) (all P < 0.05). There were no significant differences in the pre-exposure physiological parameters (all P > 0.05, **Table 1**, Supplementary Table 2A).

TABLE 1 | Differences of each variable between HAH+ and HAH− groups.

### HAH-Related Parameters at 3,700 m

Compared to the HAH− group, the HAH+ group had significantly lower post-exposure values for NEU# (51.04 ± 6.75 10<sup>9</sup> /L vs. 56.70 ± 6.59 10<sup>9</sup> /L), NEU% (3.26 ± 0.77% vs. 4.16 ± 1.31%), and WBC (6.33 ± 1.09 10<sup>9</sup> /L vs. 7.22 ± 1.62 10<sup>9</sup> /L), as well as significantly higher LYM% (37.70 ± 6.45% vs. 32.82 ± 6.13%) (all P < 0.05). Similar to the pre-exposure, there were no significant differences in the post-exposure physiological parameters (all P > 0.05, **Table 1**, Supplementary Table 2A).

### Associations between HAH Severity and the Pre-/Post-exposure Parameters

Among the pre-exposure parameters, only RET# was significantly associated with HAH severity (r = −0.315, P = 0.035). Among the post-exposure parameters, HAH severity was significantly and negatively associated with WBC (r = −0.319, P = 0.032), NEU# (r = −0.315, P = 0.035), and NEU% (r = −0.383, P = 0.009). Furthermore, HAH severity was positively associated with post-exposure LYM% (r = 0.348, P = 0.019). No other significant associations were observed between the remaining demographic or physiological parameters and HAH severity (all P > 0.05, **Table 2**, Supplementary Table 2B).


HAH+, participants with high-altitude headache (HAH); HAH−, participants without HAH; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate; SpO2, pulse oxygen saturation. \*P-value is 0.05 or less; #P-value is 0.01 or less; The bold data represent significant.

### Predictors and Risk Factors for HAH

Among the pre-exposure parameters, the univariate logistic regression analyses revealed that only RET# was significantly associated with HAH (P < 0.05, **Table 3**, Supplementary Table 2C). Among the post-exposure parameters, the univariate analyses revealed that HAH was significantly associated with WBC, NEU#, NEU%, and LYM% (all P < 0.05, **Table 4**, Supplementary Table 2C). The adjusted logistic regression analyses revealed that HAH was independently predicted by preexposure low RET#, and that low post-exposure NEU% was an independent risk factor for HAH (all P < 0.05, **Table 5**).

### DISCUSSION

The present study revealed that some hematological parameters may be novel independent risk factors/predictors for HAH. For example, HAH was associated with low pre-exposure values for RET#, RET%, and NEU#. In addition, the HAH group had low post-exposure values for NEU#, NEU%, and WBC, as well as high values for LYM%. Moreover, HAH was independently predicted by low pre-exposure RET#, while low post-exposure NEU% was an independent risk factor for HAH.


HAH, High-altitude headache; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate; SpO2, pulse oxygen saturation. The bold data represent significant.

### HAH Is Associated with Reticulocyte and Neutrophil Counts at Sea Level

The present study revealed that low pre-exposure RET# was associated with the development of HAH at an altitude of 3,700 m. Furthermore, the severity of HAH was negatively associated with RET#. In this context, reticulocytes are immature RBCs and circulate in the bloodstream for approximately 1– 4 days before developing into young mature RBCs (Lombardi et al., 2013). Interestingly, hypoxia stress can stimulate the production of reticulocytes (MacNutt et al., 2013), and young mature RBCs have greater affinity for oxygen plus more efficient oxygen delivery to the tissues, compared to old RBCs (Samaja et al., 1991; Cohen and Matot, 2013). Moreover, a recent openlabel randomized controlled trial revealed that erythropoietin (which stimulates hematopoiesis) is an effective prophylaxis for AMS (Heo et al., 2014). Therefore, our results may support the prevention of HAH using prophylactic induction of hematopoiesis in case with anticipated acute exposure to a hypoxic environment.

Our results also indicate that the HAH+ group had significantly lower pre-exposure NEU#, compared to the HAH−


CI, Confidence interval; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate; SpO2, pulse oxygen saturation. The bold data represent significant.


CI, Confidence interval; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate; SpO2, pulse oxygen saturation. The bold data represent significant.



CI, Confidence interval.

group. Previous studies have demonstrated that neutrophils undergo cellular adhesion and trans-endothelial migration from the blood stream during inflammation (Hall, 2015). In addition, hypobaric hypoxia could readily amplify pre-existing inflammation, which might lead to cerebral edema and AMS (Eltzschig and Carmeliet, 2011; Song et al., 2016). Furthermore, a recent study demonstrated that inflammation at sea level could predict the development of AMS (Lu et al., 2016). Thus, our results may indicate that inflammation at sea level is a risk factor for HAH.

### Neutrophil Recruitment during HAH

Our results indicate that the HAH+ had significantly lower values for NEU# and NEU%, and higher values for LYM%. Furthermore, HAH severity was negatively associated with NEU# and NEU%, while HAH severity was positively associated with LYM%. Although the relevance of these findings is unclear, there are several potential explanations.

The first potential explanation is that neutrophils undergo trans-endothelial migration from the blood stream during inflammation, which is related to the similar mechanisms underlying stroke and AMS (Jin et al., 2010; Julian et al., 2011; Hall, 2015). Moreover, neutrophils are among the first cells to infiltrate the ischemic brain (within 30 min to a few hours after developing focal cerebral ischemia), where they amplify the inflammatory response with the help of cytokines (e.g., interleukin-1β [IL-1β], IL-6, and tumor necrosis factor-α [TNFα]) that are produced by activated microglial cells (Ziemka-Nalecz et al., 2017). Intriguingly, our previous study revealed that plasma levels of IL-1β, IL-6, and TNF-α were positively correlated with AMS severity at 3,700 m, and another recent study revealed that plasma IL-6 levels were positively associated with AMS severity at 4,450 m and 5,129 m (Boos et al., 2016; Song et al., 2016). Thus, we speculate that hypobaric hypoxia induces brain injury by recruiting circulating neutrophils that aggravate neurogenic inflammation and lead to the development of HAH.

The second potential explanation is that neutrophil recruitment may be due to dampened anti-inflammatory IL-10 and upregulated C-C Motif Chemokine Ligand 8 (CCL8). For example, our recent transcriptome analysis revealed that IL-10 is downregulated in AMS, whereas CCL8 is upregulated (Liu et al., 2017). In addition, IL-10 is a general suppressor of the inflammatory response, while CCL8 promotes neutrophil recruitment (O'Boyle et al., 2007; Ouyang et al., 2011). Thus, we speculate that the lower NEU# and NEU% in patients with HAH are associated with the reduction of IL-10 and upregulation of CCL8, and this topic warrants urgent investigation. The third potential explanation is that the relatively high LYM% in patients with HAH could be the result of lower NEU%, as patients with and without HAH had similar values for LYM#.

### Independent Predictors and Risk Factors for HAH

The adjusted logistic regression analyses revealed that low postexposure NEU% was a risk factor for HAH. This may be because of the inflammatory response that is induced during HAH development. In addition, HAH was independently predicted by low RET#. This may reflect an impaired ability to produce young mature RBCs, which would decrease oxygen delivery to the tissues at high altitudes.

### Limitations

Ours is the first study to examine the hematological characteristics of HAH, and provides the first step toward linking the fields of hematology and HAH. However, the study also has several limitations. First, we only examined a small cohort of young Chinese men, which could have introduced age- or sex-related bias. Second, we did not have access to post-exposure reticulocyte data. Thus, our findings should be validated using larger and more heterogeneous cohorts, as well as more complete experimental data.

# CONCLUSION

The present study is the first to examine the hematological characteristics of HAH, and revealed that HAH was closely associated with WBC, NEU#, NEU%, LYM%, RET#, and RET%. In addition, HAH was independently predicted by low pre-exposure RET#, and low post-exposure NEU% was an independent risk factor for HAH.

### AUTHOR CONTRIBUTIONS

YG conceived and designed the study. BL and GW oversaw laboratory analyses and HH provided the overall supervision of the study. GX and BS did the statistical analysis or contributed the laboratory experiments. Both GW and BL contributed to sample and physiological data collections. HH drafted the report.

### REFERENCES


All authors contributed to the interpretation of results, critical revision of the manuscript and approved the final manuscript. YG is the guarantor.

# ACKNOWLEDGMENTS

This work was supported by the Key Projects in the Military Science & Technology Pillar Program during the 13 5-year Plan Period (AWS14C007), by the Key Research Project of PLA (BWS11J042, AWS16J023).

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphys. 2017.00801/full#supplementary-material


involved in early acclimatization to high altitude in Chinese Han Males. Front. Physiol. 7:601. doi: 10.3389/fphys.2016.00601


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Huang, Liu, Wu, Xu, Sun and Gao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# MiR-29a Suppresses Spermatogenic Cell Apoptosis in Testicular Ischemia-Reperfusion Injury by Targeting TRPV4 Channels

Jin-zhuo Ning1†, Wei Li 2†, Fan Cheng<sup>1</sup> \*, Wei-min Yu<sup>1</sup> , Ting Rao<sup>1</sup> , Yuan Ruan<sup>1</sup> , Run Yuan<sup>1</sup> , Xiao-bin Zhang<sup>1</sup> , Dong Zhuo<sup>3</sup> , Yang Du<sup>1</sup> and Cheng-cheng Xiao<sup>1</sup>

<sup>1</sup> Department of Urology, Renmin Hospital of Wuhan University, Wuhan, China, <sup>2</sup> Department of Anesthesiology, Renmin Hospital of Wuhan University, Wuhan, China, <sup>3</sup> Department of Urology, Wannan Medical College, Wuhu, China

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Xuejun Wang, University of South Dakota, United States Sang-Bing Ong, Duke—NUS Medical School, Singapore Ricardo Daniel Moreno, Pontificia Universidad Católica de Chile, Chile

### \*Correspondence:

Fan Cheng chenfan\_93@126.com

† These authors have contributed equally to this work.

### Specialty section:

This article was submitted to Clinical and Translational Physiology, a section of the journal Frontiers in Physiology

Received: 06 July 2017 Accepted: 14 November 2017 Published: 29 November 2017

### Citation:

Ning J, Li W, Cheng F, Yu W, Rao T, Ruan Y, Yuan R, Zhang X, Zhuo D, Du Y and Xiao C (2017) MiR-29a Suppresses Spermatogenic Cell Apoptosis in Testicular Ischemia-Reperfusion Injury by Targeting TRPV4 Channels. Front. Physiol. 8:966. doi: 10.3389/fphys.2017.00966 Background: MicroRNAs (miRNAs) have emerged as gene expression regulators in the progression of ischemia-reperfusion injury (IRI). Accumulating evidences have indicated miR-29a play roles in myocardial and cerebral IRI. However, the role of miR-29a in testicular IRI has not been elucidated.

Methods: Changes in expression of miR-29a and Transient Receptor Potential Vanilloid 4 (TRPV4) in animal samples and GC-1 spermatogenic cells were examined. The effects of miR-29a on spermatogenic cell apoptosis in testicular IRI were analyzed both in vitro and in vivo.

Results: The expression of MiR-29a was negatively correlated with the expression of TRPV4 and significantly downregulated in animal samples and GC-1 cells as testicular IRI progressed. Further studies revealed TRPV4 as a downstream target of miR-29a. Inhibition of miR-29a expression increased the expression of TRPV4 and promoted spermatogenic cell apoptosis, whereas overexpression of miR-29a downregulated TRPV4 expression and suppressed spermatogenic cell apoptosis caused by testicular IRI in vitro and in vivo.

Conclusion: Our results suggest that miR-29a suppresses apoptosis induced by testicular IRI by directly targeting TRPV4.

Keywords: miR-29a, TRPV4, testicular, IRI, apoptosis

### INTRODUCTION

Testicular torsion is the most common cause of urological emergency in young and adolescent males (Tuglu et al., 2016). Once diagnosed, clinical treatment should be adopted to restore blood flow to the testis within an appropriate time frame (Kim et al., 2016). Testicular torsion/detorsion (T/D) is considered to be the primary pathophysiologic event that induces ischemia-reperfusion injury (IRI), which causes an enhancement of apoptosis and testicular spermatogenesis dysfunction (Meštrovic et al., 2014 ´ ). Therefore, elucidating the mechanism underlying cell apoptosis is critical to helping better understand the progression of testicular IRI and to identification of effective therapeutic targets.

Transient Receptor Potential Vanilloid 4 (TRPV4) is a primary member of the transient receptor potential (TRP) channels family (Fusi et al., 2014). TRPV4 is a non-selective cation channel activated by a wide range of stimuli, including heat, cell swelling, temperature changes, pH, anandamide, arachidonic acid metabolites, and other factors (Vergnolle, 2014). TRPV4 plays a key role in regulating various cellular activities and has been found to be expressed in the kidneys, brain, lungs, skin, heart, liver, and testes (Xu et al., 2009; Wei et al., 2015; Tsuno et al., 2016). Notably, it has been reported that excessive activation of this channel is related to renal, lung, and cerebral IRI (Townsley et al., 2010; Kassmann et al., 2013; Ding et al., 2015), suggesting that TRPV4 may be an important target for mediating reperfusion injury following ischemia in many organs.

MicroRNAs (miRNAs) are endogenous, single-stranded noncoding RNAs with a length of 18–25 nucleotides (Zhang et al., 2012). They are capable of downregulating gene expression by targeting specific mRNAs located mostly in 3′ -untranslatedregions (Berezikov et al., 2005), thereby participating in a wide range of biological processes, including cell proliferation, development, differentiation and apoptosis (Zhang et al., 2010; Bao et al., 2012). Previous studies have reported that miRNAs are tightly associated with apoptosis during the progression of IRI. For example, miR-499 alleviates apoptosis by downregulating PDCD4 in myocardial IRI (Zhu et al., 2016), miR-200c regulates apoptosis of hepatic IRI by targeting ZEB1 (Wu et al., 2016), and miR-30 decreases apoptosis in renal IRI by repressing DRP1 (Gu et al., 2016). Although recent studies have shown that miR-29a plays a role in apoptosis during myocardial and cerebral IRI (Ye et al., 2010; Ouyang et al., 2013; Wang et al., 2015), the biological effect of miR-29a in testicular IRI and the interrelation between miR-29a and TRPV4 have not been described.

In the present study, we found that the expression of miR-29a was obviously reduced while TRPV4 expression was significantly enhanced in testicular IRI. Moreover, we identified miR-29a regulates TRPV4 expression by directly binding to it and demonstrated that miR-29a plays a key role in regulating IR-induced apoptosis by targeting TRPV4 in the testes. These results suggest that miR-29a may be used as a potential therapeutic option to inhibit apoptosis during the progression of testicular IRI.

# MATERIALS AND METHODS

### Animals and Surgical Protocols

All experimental procedures were adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of Wuhan University. Male C57BL/6 mice (20–25 g) were obtained from the Hubei Center for Disease Control. Prior to experiments, all rats were caged in a standard temperaturecontrolled room (22 ± 2 ◦C) and were subjected to alternating 12 h light/dark cycles. They also had free access to food and water. The rats were anesthetized by the intraperitoneal administration of 2% sodium phenobarbital (50 mg kg-1) and were then placed on a homeothermic table to maintain a rectal temperature of 37–38◦C. The left testis was twisted 720◦ clockwise and fixed to the scrotal skin with 5/0 silk. After 1 h of torsion, the testis was allowed to recover to the natural position for 0, 4, 8, 16, or 24 h. In the sham group, the testis was localized via a leftsided scrotal incision. The incision was then sutured with 5/0 silk without additional intervention. Additionally, 1–2 µg of mouse miR-29a agomir and its negative control (RiboBio, Guangzhou, China) was injected into seminiferous tubules using an injection pipette (Liang et al., 2012).

# Oxygen Glucose Deprivation/Reperfusion (OGD/R) in GC-1 Cells

Mouse GC-1 spermatogenic cells were purchased from ATCC (American Type Culture Collection, Manassas, VA, USA) and maintained in Dulbecco's modified Eagle's medium**(**DMEM; GIBCO, MA, USA) supplemented with 10% fetal bovine serum (FBS) at 37◦C under normoxic conditions (5% CO2, 95% O2). Cells were transfected with pri-miR-29a, antimiR-29a, siTRPV4, TRPV4-overexpression(GC-1/TRPV4) and their respective negative controls using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. Then, 48 h after transfection, cells were maintained under hypoxic conditions with glucose-free DMEM for 3 h. After OGD treatment, GC-1 cells were placed in glucose-containing DMEM under normoxic conditions.

## Plasmid Construction and Luciferase Reporter Assays

The putative and mutated miR-29a target binding sequence in TRPV4 were synthesized and cloned into luciferase reporter to generate the wild-type (TRPV4-Wt) or mutated-type (TRPV4- Mut) reporter plasmids. The mutant 3′UTR sequence of TRPV4 was generated using overlap extension PCR, and then both the wild-type and mutant sequences were cloned into a psiCHECK-2 vector (Promega, Madison, WI, USA).

For the luciferase reporter assay, GC-1 cells were seeded on a 24-well plate. The cells were then co-transfected with miR-29a mimics or miR-29a negative control using Lipofectamine 2000 (Invitrogen, USA). Luciferase assay was performed 48 h after transfection using a Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA).

# Western Blot Analysis

Proteins from testicular tissues or cultured GC-1 cells were extracted using RIPA Lysis Buffer (P0013B, Beyotime Institute of Biotechnology), and protein concentration was qualified using a bicinchoninic acid assay kit (BCA; Beyotime, Shanghai, China). Briefly, Equivalent amounts of protein samples were separated by 10% sodium dodecyl sulfate-polyacrylamide (SDS-PAGE) gels electrophoresis, and transferred to (polyvinylidene fluoride) PVDF membranes subsequently. After that, the membrane was blocked with 5% non-fat milk in Tris-buffered saline and 0.1% Tween 20 (TBST) buffer and incubated with the following primary antibodies against TRPV4(ab94868; Abcam, Cambridge, UK), caspase-3 (sc7148; Santa Cruz, CA), Bax (sc493; Santa Cruz, CA) and Bcl-2 (sc7382; Santa Cruz, CA) at 4◦C overnight. GAPDH was used as an internal control. After being rinsed three times with TBST buffer, the membranes were incubated

with secondary antibodies at room temperature for 1 h. Target proteins were visualized using an ECL system kit (Pierce Biotechnology, Beijing, China). Optical densities were detected by enhanced chemiluminescent (ECL) and qualified using ImageJ software (NIH, Bethesda, MD, USA).

### Quantitative Real-Time PCR

Total RNA was isolated from GC-1 cells and testis samples using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA), and the concentration of RNA was determined by a DU800 UV/Vis Spectrophotometer (Beckman Coulter, CA, USA). Total cellular RNAs were reversed transcribed into cDNA using reverse transcription reagent kit (Takara Biotechnology, Dalian China). Real-time quantitative PCR was performed via a Applied Biosystems SYBR Green mix kit and the ABI 7900 Real-Time PCR system (Applied Biosystems Life Technologies, Foster City, CA, USA). Relative miR-29a or TRPV4 mRNA expression were normalized to snRNA U6 (for miRNAs) or GAPDH (for mRNAs), respectively. The quantitative analysis was calculated by using 2−11Ct method (Rao et al., 2013). The primer sequences used are shown in **Table 1**.

### Immunohistochemistry

The expression of TRPV4, Bax and Bcl-2 was detected by immunohistochemical staining. Tissues were fixed in 4% paraformaldehyde, embedded in paraffin and then cut in 4µm thickness. Immunohistochemical staining was performed using rabbit polyclonal anti-TRPV4 (ab94868; Abcam, Cambridge, UK), rabbit polyclonal anti-Bax (sc493; Santa Cruz, CA) and mouse monoclonal anti-Bcl-2 (sc7382; Santa Cruz, CA). After being washed three times with PBS, all sections were incubated in diaminobenzidine (DAB) reagents and counterstained with haematoxylin.

### TUNEL Assays

A TUNEL assay was performed to evaluate spermatogenic cell apoptosis in testicular tissues using a transferase-mediated dUTP nick-end labeling (TUNEL) method with a detection kit (Roche, Mannheim, Germany) following the manufacturer's protocol. The nuclei that stained brown were considered TUNEL-positive cells. Five visual fields were randomly selected in each slice, and the average number of apoptosis cells per 200 cells was counted. The apoptosis index (AI) was determined as follows: AI = (the number of positive cells/the total number of counted cells) × 100%.

### TABLE 1 | RT-PCR primer sequences.


### Cell Apoptosis Analysis

Cell apoptosis was performed using Annexin V-FITC/Propidium Iodide (PI) staining (BD PharMingen, San Jose, CA, USA). GC-1 cells were seeded in 6-well plates at a concentration of 10<sup>6</sup> cells mL-1. The cells were labeled with Annexin V-FITC for 5 min in the dark. Then, 5 mg ml-1 PI was added to each sample for 30 min so that flow cytometry could be performed (BD PharMingen, San Jose, CA, USA).

### Statistical Analysis

All data are presented as the mean ± SD. Differences were assessed by one–way analysis of variance (ANOVA), followed by all pairwise multiple-comparison procedures using the Bonferroni test. A value of P < 0.05 was considered statistically significant. All experiments were performed at least 3 times. Statistical analysis was performed using SPSS 19.0 (SPSS Inc, Chicago, IL, USA).

# RESULTS

### Enhanced Expression of TRPV4 in Testicular IRI Correlates with Downregulation of miR-29a Expression

To investigate if TRPV4 and miR-29a are involved in testicular IRI, we examined their expression levels in animal samples by immunohistochemistry staining, qRT-PCR and western blot. The samples were collected and assayed for expresssion at 0, 4, 8, 16, or 24 h of reperfusion after 1 h ischemia (Aslan Ko¸sar et al., 2015). The results showed that TRPV4 expression increased gradually and peaked at 16 h of reperfusion compared with the sham group (**Figures 1A–D**, n = 5 per group). On the other hand, miR-29a expression decreased significantly during the progression of IRI (**Figure 1E**, n = 5 per group). A two-tailed Pearson's correlation analysis was performed to further investigate the interrelation between miR-29a and TRPV4 expression (**Figure 1F**). Therefore, the expression of miR-29a is negatively correlated with the expression of TRPV4 in vivo.

### miR-29a Is Negatively Correlated with the Expression of TRPV4 in Vitro

To further explore the possibility that miR-29a might negatively correlates with the expression of TRPV4 in vitro, we examined the expression of TRPV4 and miR-29a in GC-1 cells under different reoxygenation conditions (0, 6, 12, 24, and 48 h) after 3 h OGD exposure. Consistent with the in vivo studies, the qRT-PCR and western blot results showed that miR-29a expression was inversely correlated with the expression of TRPV4 at different reoxygenation time intervals (**Figures 2A–D**, n = 6 per group). We next transfected the GC-1 cells with pri-miR-29a and examined TRPV4 expression by western blot and qRT-PCR at 3 h of OGD followed by 24 h of reoxygenation. We found that overexpression of miR-29a led to a significant downregulation of TRPV4 expression. Further, GC-1 cells transfected with a miR-29a inhibitor, displayed a moderate upregulation of TRPV4 expression (**Figures 2E–G**, n = 6 per group).

mRNA expression of miR-29a is inversely correlated with the expression of TRPV4 (p < 0.05).

# Influence of miR-29a and TRPV4 on GC-1 Cell Apoptosis in Vitro

To investigate the effect of miR-29a on OGD/R cell injury, we used pri-miR-29a and anti-miR-29a to change miR-29a levels in GC-1 cells. Indeed, pri-miR-29a and anti-miR-29a markedly increased and decreased miR-29a levels at 3 h of OGD/24 h of reoxygenation, respectively when compared with their negative controls (**Figure 3A**, n = 6 per group). Flow cytometry data also showed that GC-1 cell apoptosis was induced by 3 h of OGD/24 h of reoxygenation. Transfection of pri-miR-29a inhibited cell apoptosis, while transfection of anti-miR-29a promoted GC-1 cell apoptosis induced by 3 h of OGD/24 h of reoxygenation (**Figures 3B,C**, n = 6 per group). In addition, western blot analysis showed that overexpression of miR-29a and knockdown of TRPV4 decreased the expression of Bax and caspase-3 and increased the expression of Bcl-2, respectively. Consistent with this result, inhibition of miR-29a and overexpression of TRPV4 in GC-1 cells resulted in an increase in Bax and

treatments. \*p < 0.05 vs. non-trans (3 h OGD/24 h reoxygenation treatment), n = 6 per group.

caspase-3 levels and a decrease in Bcl-2 expression, respectively (**Figures 3D–K**, n = 6 per group). These results suggest that miR-29a suppresses cell apoptosis and TRPV4 promotes cell apoptosis in vitro.

### miR-29a Directly Targets TRPV4 and Alleviates Apoptosis in Vitro

To further explore the relationship between miR-29a and TRPV4, an in silico prediction was performed using open access software (TargetScan, PicTarget, and miRanda). A putative binding site for miR-29a was identified within the 3 ′UTR of TRPV4. To verify this prediction, we cloned a luciferase reporter sequence in the 3′UTR of TRPV4, which contains the putative miR-29a binding sites. A mutant reporter vector of the 3′UTR of TRPV4 containing luciferase reporter was used as negative control. Data from luciferase reporter assay showed that overexpression of miR-29a significantly decreased reporter vector activity of TRPV4 3′UTR in GC-1 cells but had no effect on the mutated reporter vector (**Figures 4A,B,** n = 6 per group). To further investigate whether

6 per group.

miR-29a regulates apoptosis in testicular IRI via targeting TRPV4, we employed immunoblotting assays to determine apoptosis related proteins after cells were transfected with pri-miR-29a or anti-miR-29a and TRPV4 siRNAs or TRPV4 overexpression (GC-1/TRPV4). We found that overexpression of miR-29a markedly suppresses apoptosis, whereas TRPV4 overexpression abrogated such decrease in apoptosis induced by miR-29a (**Figures 4C–F**, n = 6 per group). On the other hand, knockdown of miR-29a led to an increase in apoptosis, while which could be rescued by TRPV4 inhibition (**Figures 4G–J**, n = 6 per group). Together, these results suggest that miR-29a directly targets TRPV4 and alleviates apoptosis in vitro.

### miR-29a Inhibits Spermatogenic Cell Apoptosis via Downregulation of TRPV4 in Vivo

To further determine the effect of miR-29a expression on spermatogenic cell apoptosis in vivo, miR-29a agomir was injected into the seminiferous tubules to increase miR-29a expression. A TUNEL assay was used to determine whether

miR-29a could affect spermatogenic cell apoptosis in response to a 1 h of ischemia/16 h of reperfusion treatment. The TUNEL assay showed no obvious apoptotic cells in the sham group. Compared with negative control group, the effect of miR-29a overexpression dramatically reduced the number of apoptotic cells (**Figures 5A**,**B**, n = 5 per group). Immunohistochemistry staining revealed that overexpression of miR-29a resulted in lower expression levels of Bax and a higher expression level of Bcl-2 in comparison with negative control group (**Figure 5C,** n = 5 per group). Moreover, western blot data showed that overexpression of miR-29a decreased protein levels of TRPV4 and caspase-3 when compared to the negative control group (**Figures 5D–F**, n = 5 per group). Together, our results suggest that miR-29a suppresses apoptosis in vivo.

### DISCUSSION

The pathophysiological mechanisms of testicular T/D result from direct damage induced by ischemia during torsion and a secondary effect attributed to enhanced blood flow with reperfusion during detorsion (Huang et al., 2012). Testicular IRI triggers apoptosis-related signaling pathways and further negatively affects spermatogenesis, eventually leading to male infertility (Hadziselimovic et al., 1998; Meštrovic et al., ´ 2014). The two major factors affecting testicular damage are the degree and duration of spermatic cord torsion (Cvetkovic et al., 2015). Timely treatment of testis torsion is a crucial issue, and currently both manipulative reduction and orchiectomy treatment are known to cause injury to the bilateral testis (Dursun et al., 2015). Thus, understanding the mechanisms underlying apoptosis in testicular IRI and exploring new molecular interactive targets will help to develop effective therapeutics. Previous studies have demonstrated that miRNA can regulate apoptosis process in organ IRI by targeting mRNAs of specific genes (Berezikov et al., 2005). Based on this information, we tried to identify miRNAs that bind the TRPV4 gene and to determine the regulatory relationship in the progression of testicular IRI.

MicroRNA expression is highly related to apoptosis in the progression of organs IRI. It has been shown that abnormally expressed miRNAs can regulate RNA networks during IRI progression, and a single miRNA may function in pro-apoptosis or anti-apoptosis roles under different contexts (Andreeva et al., 2015). Previous studies have demonstrated that overexpression of miR-29a promotes myocardial apoptosis by directly targeting

IGF-1 and Mcl-1 depending on activation of P13K/Akt signaling pathways (Ye et al., 2010; Wang et al., 2015). Furthermore, a recent study suggested that miR-29a suppresses apoptosis by negatively regulating PUMA in cerebral IRI (Ouyang et al., 2013). In our studies, we analysed animal samples at different time points of reperfusion. Our data demonstrated that miR-29a expression is significantly downregulated in the progression of testicular IRI and that lower expression of miR-29a is negatively correlated with expression of TRPV4 channels.

TRPV4 has been reported to be upregulated during the progression of IRI and to be involved in apoptosis via multiple signaling pathways (Ye et al., 2012; Hong et al., 2016). For example, TRPV4 expression is upregulated during the progression of IRI and involved in apoptosis via multiple signaling pathways (Jie et al., 2015). Inhibition of TRPV4 reduces neurological injury after cerebral infarction, and leads to apoptosis of mouse retinal ganglion cells, which may contribute to the activation of Ca2+-dependent signaling pathways (Ryskamp et al., 2011). In our in vitro studies, we further confirmed the opposite relationship between TRPV4 and miR-29a expression in the GC-1 cells. We found that overexpression of miR-29a suppressed TRPV4 expression and that inhibition of miR-29a promoted TRPV4 expression. More important, we have identified TRPV4 as a direct downstream target of miR-29a using a luciferase reporter assay. These data provide evidence that miR-29a negatively regulates the expression of TRPV4 in the GC-1 cells, which is consistent with findings from in vivo studies.

Apoptosis is a form of cell death based on genetic mechanisms and plays an important role by inducing a series of pathophysiologic changes (Vaux and Korsmeyer, 1999), and it has been generally accepted that the mitochondrial signaling pathway is the major channel for apoptosis, which is regulated by precise gene expression (Desagher and Martinou, 2000). The Bcl-2 family is composed of pro-apoptotic factors (e.g., Bax) and anti-apoptotic factors (e.g., Bcl-2). The ratio of Bcl-2/Bax is a regulator of spermatogenic cell apoptosis and determines the extent of apoptosis in spermatogenic cells exposed to damage (Liang et al., 2013). Moreover, caspase-3 is an inactive zymogen located in the cytoplasm and serves

### REFERENCES


as the convergence point of multiple apoptotic stimuli signals. Its activation signals irreversible commitment to cell apoptosis, leading to changes in cell shrinkage, chromatin condensation and DNA fragmentation (Zheng et al., 2008). In our in vitro studies, we have demonstrated that overexpression of miR-29a and knockdown of TRPV4 reduce cell apoptosis, while inhibition of miR-29a and overexpression of TRPV4 have an opposite effect. Moreover, data from in vivo studies also support that increased expression of miR-29a produces a decrease in apoptosis by downregulating the TRPV4 gene expression. Together, these data suggest that miR-29a-mediated inhibition of TRPV4 is associated with cell apoptosis in testicular IRI.

### CONCLUSIONS

In conclusion, our study revealed a negative correlation between the expression of miR-29a and TRPV4 during the progression of IRI. In addition, we showed that miR-29a alleviated apoptosis by directly targeting TRPV4 both in vitro and in vivo. These findings suggest that miR-29a might serve as an effective therapeutic target to suppress spermatogenic cell apoptosis in testicular IRI.

### AUTHOR CONTRIBUTIONS

JN and WL, Conception and design of the study, data collection and analysis, manuscript writing. FC, WY, TR, and DZ, Design of the study, critical revision, supervised all phases of the study. RY, YR, XZ, YD, and CX, Data collection.

keratinocytes in human non-melanoma skin cancer. J. Invest. Dermatol. 134, 2408–2417. doi: 10.1038/jid.2014.145


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer RDM and handling Editor declared their shared affiliation.

Copyright © 2017 Ning, Li, Cheng, Yu, Rao, Ruan, Yuan, Zhang, Zhuo, Du and Xiao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# The Influence of CO<sup>2</sup> and Exercise on Hypobaric Hypoxia Induced Pulmonary Edema in Rats

### Ryan L. Sheppard1,2 \*, Joshua M. Swift <sup>2</sup> , Aaron Hall <sup>2</sup> and Richard T. Mahon<sup>2</sup>

<sup>1</sup> Department of Submarine Medicine and Survival Systems Groton, Naval Submarine Medical Research Laboratory, Groton, CT, United States, <sup>2</sup> Department of Undersea Medicine, Walter Reed Army Institute of Research and Naval Medical Research Center, Silver Spring, MD, United States

Introduction: Individuals with a known susceptibility to high altitude pulmonary edema (HAPE) demonstrate a reduced ventilation response and increased pulmonary vasoconstriction when exposed to hypoxia. It is unknown whether reduced sensitivity to hypercapnia is correlated with increased incidence and/or severity of HAPE, and while acute exercise at altitude is known to exacerbate symptoms the effect of exercise training on HAPE susceptibility is unclear.

### Edited by:

Rodrigo Del Rio, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Marli Cardoso Martins-Pinge, Universidade Estadual de Londrina, Brazil Beth J. Allison, Hudson Institute of Medical Research, Australia

\*Correspondence:

Ryan L. Sheppard ryan.l.sheppard.mil@mail.mil

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 24 October 2017 Accepted: 08 February 2018 Published: 28 February 2018

### Citation:

Sheppard RL, Swift JM, Hall A and Mahon RT (2018) The Influence of CO2 and Exercise on Hypobaric Hypoxia Induced Pulmonary Edema in Rats. Front. Physiol. 9:130. doi: 10.3389/fphys.2018.00130 Purpose: To determine if chronic intermittent hypercapnia and exercise increases the incidence of HAPE in rats.

Methods: Male Wistar rats were randomized to sedentary (sed-air), CO<sup>2</sup> (sed-CO2,) exercise (ex-air), or exercise + CO<sup>2</sup> (ex-CO2) groups. CO<sup>2</sup> (3.5%) and treadmill exercise (15 m/min, 10% grade) were conducted on a metabolic treadmill, 1 h/day for 4 weeks. Vascular reactivity to CO<sup>2</sup> was assessed after the training period by rheoencephalography (REG). Following the training period, animals were exposed to hypobaric hypoxia (HH) equivalent to 25,000 ft for 24 h. Pulmonary injury was assessed by wet/dry weight ratio, lung vascular permeability, bronchoalveolar lavage (BAL), and histology.

Results: HH increased lung wet/dry ratio (HH 5.51 ± 0.29 vs. sham 4.80 ± 0.11, P < 0.05), lung permeability (556 ± 84 u/L vs. 192 ± 29 u/L, P < 0.001), and BAL protein (221 ± 33µg/ml vs. 114 ± 13µg/ml, P < 0.001), white blood cell (1.16 ± 0.26 vs. 0.66 ± 0.06, P < 0.05), and platelet (16.4 ± 2.3, vs. 6.0 ± 0.5, P < 0.001) counts in comparison to normobaric normoxia. Vascular reactivity was suppressed by exercise (−53% vs. sham, P < 0.05) and exercise+CO<sup>2</sup> (−71% vs. sham, P < 0.05). However, neither exercise nor intermittent hypercapnia altered HH-induced changes in lung wet/dry weight, BAL protein and cellular infiltration, or pulmonary histology.

Conclusion: Exercise training attenuates vascular reactivity to CO<sup>2</sup> in rats but neither exercise training nor chronic intermittent hypercapnia affect HH- induced pulmonary edema.

Keywords: vascular reactivity, HAPE, chemoreflex, hypercapnia, exercise

# INTRODUCTION

Rapid ascent to altitudes greater than 2,500 m by nonacclimatized individuals can induce a form of altitude sickness known as high altitude pulmonary edema (HAPE). Clinical symptoms of HAPE include reduced exercise capacity, tachycardia, dyspnea, cough, frothy sputum, chest tightness (Houston, 1960), crackles on exam and the hallmark bilateral patchy infiltrates observable via chest radiography (Vock et al., 1989). Symptoms typically present and worsen within 2 days of significant altitude exposure in the unacclimatized, particularly when physical exertion is involved. Though HAPE is rare at altitudes below 3,000 m (Maggiorini et al., 1990; Gabry et al., 2003), subclinical HAPE has been estimated to occur in up to 75% of individuals at 4,500 m elevation (Cremona et al., 2002)

HAPE progression is driven by pulmonary vasoconstriction that occurs with prolonged exposure to hypoxia. Pulmonary vascular pressures subsequently increase and can stress the delicate alveolar capillary barrier and induce mechanical stress failure. This pulmonary vascular response likely has multiple driving mechanisms, including increased plasma norepinephrine and sympathetic tone (Duplain et al., 1999), elevated endothelin-1 (Ali et al., 2012; Barker et al., 2016), and decreased levels of exhaled nitric oxide (Duplain et al., 2000; Busch et al., 2001) and nitric oxide metabolite concentrations in both the systemic circulation and bronchoalveolar lavage (BAL) fluid (Swenson et al., 2002; Berger et al., 2005; Bailey et al., 2010). Numerous studies have indicated that HAPE susceptible individuals have a reduced hypoxic ventilation response (HVR) accompanied by lower pO<sup>2</sup> levels and greater hypoxic pulmonary vasoconstriction than non-HAPE sensitives (Hyers et al., 1979; Matsuzawa et al., 1989; Hohenhaus et al., 1995; Bartsch et al., 2002), though a reduced HVR does not appear necessary for HAPE to develop (Selland et al., 1993). The mechanism of reduced HVR is multifactorial, but blunted central and peripheral chemoreceptor sensitivity likely plays a central role (Albert and Swenson, 2014).

Although there is ample evidence of the impact of hypoxia on chemoreceptor reflex, HVR and HAPE, the impact of hypercapnia is less clear. While it has been reported that HAPEsusceptible individuals demonstrate reduced sensitivity to CO<sup>2</sup> (Mathew et al., 1983), the effect of chronic CO<sup>2</sup> exposure on HAPE susceptibility has not been investigated. It is also widely acknowledged that acute exercise exacerbates HAPE symptoms however the role of chronic exercise training in HAPE development is unclear. Exercise training increases cardiac output, which may enhance the heterogeneous pulmonary capillary pressures and leakage induced by hypoxia. Indeed, increased pulmonary fluid and BAL solute concentrations have been confirmed in conditioned endurance athletes after hypoxic exercise (Eldridge et al., 2006), but overall the data is very limited. This study examines the effect of chronic intermittent CO<sup>2</sup> exposure, with and without concomitant exercise training, on the development of hypobaric hypoxia (HH) induced pulmonary edema in a rat model. We hypothesized that chronic CO<sup>2</sup> exposure would result in an increased incidence and severity of HAPE, and that this effect would be exacerbated by exercise training.

# MATERIALS AND METHODS

### Animals

All experiments were reviewed and approved by the Walter Reed Army Institute of Research/Naval Medical Research Center Institutional Animal Care and Use Committee in compliance with DoDI 3216.01 and SECNAVINST 3900.38C and all applicable Federal regulations governing the protection of animals in research. The facility is AAALAC accredited, and all animals were maintained under the surveillance of licensed veterinarians. Eight to ten week old male Wistar rats (Charles River Laboratories) were pair-housed at the animal care facility and maintained on a 12 h light/dark cycle and provided standard rodent chow (Harlan Teklad 8604) and water ad libitum. Animals were quarantined for 1 week (as per institutional policy) before being randomly assigned to one of 5 groups, N = 8 per group (**Table 1**); (1) sham (no HH, no CO2, no exercise), (2) sedentary animals exposed to room air only, (3) sedentary animals exposed to a custom 3.5% CO<sup>2</sup> gas mix, (4) exercised animals exposed to room air, (5) exercised animals exposed to a 3.5% CO<sup>2</sup> mix. A separate cohort of animals, N = 8/group, was utilized for all CO<sup>2</sup> reactivity experiments.

# CO<sup>2</sup> Exposure and Exercise Training

A 3.5% CO<sup>2</sup> level was chosen as the target exposure level with consideration of the following; (1) the US Occupational Safety and Health Administration maximum short-term exposure limit for CO<sup>2</sup> in humans is 3%, (2) 3–4% CO<sup>2</sup> is noticeable by human subjects and induces moderate respiratory stimulation, (3) one treatment group would be receiving CO<sup>2</sup> simultaneously with exercise. For the exercise groups, 3–5 short familiarization runs were performed prior to beginning the 1 month training period in order to identify animals with poor treadmill running compliance. Approximately one in eight animals was removed from the study during the familiarization period due to failure to keep up a steady running pace, repeatedly falling back and touching the end plate (set to deliver a very mild electrical shock), or refusing to move off of the end plate altogether, on two or more occasions. Treadmill exercise training was conducted at 15 m/min, 10% grade, for 1 h/day, 5 days/week for 4 weeks on a metabolic modular treadmill (Columbus Instruments, Columbus


HH, hypobaric hypoxia.

OH). The treadmill lanes were set up to an air-tight configuration and either room air or a custom mixture (3.5% CO2, 20% O2, balance N2) was supplied at a rate of 4 L/min. The 4 L/min flow rate was sufficient to maintain CO<sup>2</sup> and O<sup>2</sup> levels within each treadmill chamber at 3.5 and 20%, respectively, throughout the daily training period (9600-A1BTP O2/CO<sup>2</sup> Analyzer, Alpha Omega Instruments, Lincoln, RI). In non-exercising groups the animals were placed on an inactive treadmill ventilated with either room air or the CO<sup>2</sup> mix for 1 h/day.

# Hypobaric Hypoxia (HH)

Custom-built methylacrylic cylinders measuring 38 cm in length by 20 cm in diameter were constructed in house (**Figure 1**). Evacuation of the chambers was driven by a GAST 6066 vacuum pump (GAST Manufacturing, Benton Harbor MI) through a side port. Fresh air was supplied at a rate of 2–2.5 L/min from compressed air banks; this flow rate was sufficient to prevent any buildup of CO<sup>2</sup> (9600-A1BTP O2/CO<sup>2</sup> Analyzer, Alpha Omega Instruments, Lincoln, RI). Chamber temperatures during operation did not differ significantly from the ambient room temperature (23–26◦C). Chambers were furnished with rodent chow, water and a hide box for each animal to reduce overall stress. An atmospheric pressure equivalent to 25,000 ft of altitude (282.0 torr, CPA2501 digital air indicator, Mensor, TX) was chosen as the target for HH exposure. We arrived at this value after extensive model development testing revealed that this pressure reliably induced a significant increase in mean lung wet/dry weight ratios, whereas greater atmospheric pressures did not reliably increase wet/dry ratios after 24 h (data not shown). Longer duration exposures were not employed due to extreme dehydration (animals refused fluids during the exposures), and further reduced atmospheric pressures dramatically increased mortality. During all experiments the chambers were de-pressurized/re-pressurized at an equivalent rate of 1,000 ft/min. Intake/outtake valves were adjusted to maintain the targeted altitude equivalent of 25,000 ft ± 500 ft at all times. After 24 h, chambers were re-pressurized (1,000 ft/min) and animals were euthanized by overdose of a sodium pentobarbital solution (150 mg/kg, IP) upon return to sea level. Sham group animals were placed in the hypobaric chambers for 24 h with air flow of 2–2.5 L/min, but the chambers were not subjected to vacuum and remained at ambient room pressures.

# Vascular CO<sup>2</sup> Reactivity

Cerebral blood flow (CBF) following acute exposure to CO<sup>2</sup> was assessed after completing the 4-week training program via rheoencephalography (REG) as described (Ainslie and Duffin, 2009). Animals were anesthetized via urethane (1,000 mg/kg, IP) and electrodes were prepared and placed as previously described (Bodo et al., 2005). After electrode positioning baseline REG readings were recorded. A 12% CO2/20% O<sup>2</sup> gas mixture was then supplied for 30 s, followed by a 5 min recovery period on room air. The CO<sup>2</sup> pulse was repeated a second time, and changes in REG amplitude were averaged for each animal.

### Lung Tissue Preparation and Bronchoalveolar Lavage (BAL)

Tracheal cannulas were placed and tied in and lungs were removed en bloc within 10 min of returning to ambient room pressure. A small portion of each right lobe was flash frozen in liquid N<sup>2</sup> for later gene expression analysis, while the majority was weighed, dried at 55◦C for 72 h, and weighed again to determine the wet to dry weight ratio. BAL samples were obtained by gently infusing 1 ml of PBS into the left lung a total of 3 times, pooled, and stored on ice. Portions of each

left lobe were preserved in a 10% formalin solution, sectioned 5µm thick at equidistant points using a microtome and tissue block, hematoxylin/eosin stained, and examined by a board certified veterinary pathologist for evidence of fibrin thrombi, edema, endothelial damage, cellular infiltrates, thickening of the alveolar septum, hemorrhage or necrosis. Tissues were assigned quantitative values for each category on a 0–3 scoring system; 0 = no significant lesions, 1 = ≤ 20% of tissue affected, 2 = 20–50% of tissue affected, 3 = ≥ 50% of tissue affected. To ensure accuracy in scoring, 5 individual sections per animal were examined. Lavage fluid was spun at 1,500 g for 5 min, and the resulting pellets were then re-suspended in 1 ml of ice-cold PBS and immediately analyzed for cell type (Advia 120 Hematology System, Siemens Healthcare, Tarrytown, NJ). The supernatant was analyzed for lactate dehydrogenase activity and total protein content (Pierce BCA Protein Assay, Thermo Scientific, Rockford, IL).

### Lung Vascular Permeability

Lung vascular permeability was examined in vivo after exposure to HH or normobaric conditions for 24 h. Sodium fluorescein dye (5 ml/kg, 2% solution; F6377-500G; Sigma-Aldrich, St. Louis, MO) was injected through the tail vein in a separate cohort of rats immediately after normobaric or hypobaric exposure. Thirty minutes after fluorescein injection rats were anesthetized with a ketamine/xylazine cocktail (80 and 10 mg/kg, respectively, IP) and transcardially perfused with PBS to remove the fluorescent tracer prior to harvesting the lungs. Lungs were rinsed with cold saline and homogenized with 2 ml of PBS. After homogenization, 2 ml of 50% TCA solution was added and the samples were vortexed to mix for 30 s to precipitate proteins in the lung samples. The resultant sample solution was cooled at 4◦C for 30 min and covered to ensure no light exposure to the sample. Upon completion of cooling the samples were centrifuged for 10 min at 3,300 × g in 4◦C. One mL of supernatant from each sample was placed into a 1.5 ml microcentrifuge tube and centrifuge at 16,000 × g at 4◦C for 10 min. 200 ul of supernatant from each sample was added to a 96-well microplate along with 12 standards and read via spectrofluorometer according to manufacturer directions. Results are expressed as relative fluorescence units per liter.

### Analysis

All data are expressed as means ± SEM and analyzed using GraphPad Prism (GraphPad Software, La Jolla, CA). Twoway ANOVA (sedentary/exercised X air/CO2) was used to analyze for main effect and interactions between factors for each dependent variable. Tukey's post-hoc test with multiplicity adjusted P-value was used for pairwise comparisons. The sham group was not included in ANOVA as comparisons of sham vs. HH with CO2/exercise trained groups were not of interest. Instead, unpaired t-tests were used to determine the effects of HH (sham vs. sed-air only) on each dependent variable. Statistical significance was accepted at P ≤ 0.05.

### RESULTS

### Rat HH Induces Pulmonary Edema

HH exposed animals exhibited signs of pulmonary injury consistent with HAPE. Lung wet-to-dry weight ratios in sedair animals increased 15% over sham animals (P = 0.014; **Figure 2**). Pulmonary vascular permeability in the sed-air group, assessed by sodium fluorescein dye leakage, was increased 190% over sham (P < 0.001). Total protein content, white blood cell concentration, and platelet concentration in BAL fluid were significantly increased by HH (**Table 2**), while there was no change in lactate dehydrogenase activity in the BAL fluid, and no red blood cells were detected in either group. Histological examination revealed no evidence of gross pulmonary pathology.

### Exercise Training Attenuates CO2-Induced Increases in CBF

Exposure to a 30 s, 12% CO<sup>2</sup> pulse increased REG amplitude by 343% ± 63.9, 185% ± 33, 161% ± 45, and 100% ± 28, in sedair, sed-CO2, ex-air, and ex-CO<sup>2</sup> groups, respectively (**Figure 3**). Daily CO<sup>2</sup> exposure (sed-CO2) resulted in an apparent reduction


TABLE 2 | Hypobaric hypoxia exposure vs. sham control.

ND, not detected, \*P ≤ 0.05, \*\*P ≤ 0.001, in compares on to <sup>a</sup> sham, <sup>b</sup> sed-air. Hypobaric hypoxia groups are compared to each other; Sham is compared to sed-air group only.

in the CO<sup>2</sup> pulse-induced REG increase but this effect did not reach statistical significance (P = 0.062), while exercise alone (ex-air) and exercise w/CO<sup>2</sup> (ex-CO2) attenuated the CO2-pulse induced REG increase by 53% (P = 0.038) and 71% (P = 0.003) vs. sed-air, respectively. REG values in the sed-CO2, ex-air, and ex-CO<sup>2</sup> groups were not significantly different from each other and there was no interactive effect. One animal in the ex-air group was removed from the study due to a foot injury that precluded treadmill running, and two animals in the sed-air and one animal from the ex-CO<sup>2</sup> group died due to complications from anesthesia and/or REG electrode placement.

## CO<sup>2</sup> and Exercise Training Do Not Affect the Incidence or Severity of HH-Induced Pulmonary Edema

To assess the effects of chronic exercise and CO<sup>2</sup> exposure on the development of HAPE in rodents, we exposed each of the four groups to HH equivalent to 25,000 ft of altitude for 24 h. Though the lung tissues were edematous in comparison to non-altitude exposed tissues, exercise and CO<sup>2</sup> exposure had no effect on between group wet-to-dry weight ratios (**Figure 4**). BAL analysis revealed no difference in total protein content or LDH activity between any of the groups however the ex-air group did have decreased platelet and total white blood cell concentrations in the BAL fluid (**Table 2**). Upon histological examination the tissues from each group were indistinguishable and relatively unremarkable, showing minimal signs of cellular infiltrates, septal thickening, hemorrhage, or necrosis (**Figure 5**).

### DISCUSSION

The primary aim of this study was to determine how chronic intermittent CO<sup>2</sup> exposure may affect HAPE susceptibility in rats. HAPE development is largely dependent upon an inappropriately high level of pulmonary vascular constriction concurrent with exposure to hypoxia. Repeated hypoxic exposures appear to alter the chemoreceptor reflex (Berkenbosch et al., 1992; Mahamed and Duffin, 2001), and we hypothesized that chronic intermittent CO<sup>2</sup> exposure would also affect chemoreceptor sensitivity, exacerbate the pulmonary vasoconstriction effect of hypoxia, and worsen the symptoms and severity of HAPE. We also hypothesized that chronic CO<sup>2</sup> would reduce HVR, further exacerbating HAPE development. As most individuals acutely exposed to altitude are also performing relatively strenuous exercise (mountaineers, extreme sport athletes, military personnel, etc.) an exercise component was included to assess the interaction of chronic intermittent CO<sup>2</sup> and cardiovascular conditioning on HAPE susceptibility. The results presented here do not support our hypotheses, but do include some interesting findings.

Our HAPE model successfully induced pulmonary edema where previously published efforts to develop rat models of HAPE have met with mixed success. The methodology and exposure protocols have varied considerably, and while some studies have reported significant changes in wet-to-dry weight ratios after various altitude exposure profiles (Shukla et al., 2011; Lee et al., 2013), others did not demonstrate any differences in wet/dry ratio (Omura et al., 2000; Berg et al., 2004), used other methods to identify and quantify pulmonary edema (Li et al., 2011; Lin et al., 2011), or used very severe exposure protocols to induce pulmonary edema (Bai et al., 2010). Though our

pressure equivalent to 25,000 ft elevation for 24 h. Sed-air (N = 8), sed-CO<sup>2</sup> (N = 8), ex-air (N = 6), ex-CO<sup>2</sup> (N = 8).

model was generally well tolerated the animals were lethargic, did not consume food or water during the exposure, and lost a significant amount of bodyweight (see **Table 2**). Despite a robust HAPE model however, exercise training and chronic intermittent hypercapnia did not affect HH-induced pulmonary edema. It is possible that simulated altitude of 25,000 ft for 24 h (greater than what most humans are likely to experience) was sufficient to elicit a maximal level of pulmonary vascular stress and that any effects of hypercapnia and exercise were masked or comparatively minor. Lower pressures and/or longer exposure times dramatically increased mortality and morbidity, which lends some support to this hypothesis. A more modest simulated altitude exposure may allow for more differentiation between the treatment groups, but our preliminary tests at higher atmospheric pressures were insufficient to reliably induce an appreciable level of pulmonary edema.

The level of bodyweight loss in HH-exposed animals was quite striking with most animals losing ∼9% of total bodyweight in 24 h. Similar bodyweight losses have been reported in mice after exposure to ∼14,000 ft of altitude for 3 days, however most of this was due to adipose lipolysis (Hannon and Rogers, 1975) which is unlikely to account for much of the losses here, and previous studies have reported reduced fluid intake in rodents exposed to altitude (Jones et al., 1981). The bodyweight reductions seen here were likely mostly due to loss of total body water, however preloading the animals with subcutaneous fluid (25–30 ml normal saline) and humidifying the supply

air to ∼60% during HH also had no discernible effect on bodyweight loss or lung wet/dry weight ratios in a separate cohort of animals. Sham animals that were not exposed to HH averaged 4.3% loss of total bodyweight over the same time period, indicating roughly half of the bodyweight reduction can be attributed to non-HH factors. Nevertheless the additional bodyweight reductions seen in HH-exposed animals may still have been sufficient to appreciably reduce plasma and interstitial fluid volume. As HAPE is driven by heterogeneously high pulmonary vascular pressures, such a reduction in total blood volume may have been sufficient to attenuate HH-induced increases in pulmonary vascular pressures. We did not perform direct measurements of pulmonary artery pressure or compare pre-post-total body water levels, so this possibility cannot be ruled out.

We hypothesized that chronic CO<sup>2</sup> exposure would suppress respiratory drive with HH and the study was designed with the intent to periodically measure respiration throughout the HH exposure period. Despite successful preliminary optimization runs, complications during the collection period made the respiratory data unreliable and it was not feasible to repeat the experiments. This is a significant limitation of the study and without this data we can only speculate, however the central and peripheral chemoreceptors responsible for respiratory regulation are highly sensitive to pCO<sup>2</sup> /H<sup>+</sup> (Nattie, 1999) and a stimulus that was sufficient to alter the CBF response to hypercapnia could be expected to also affect respiratory drive.

While direct measurement of pulmonary artery pressure via catheterization or echocardiogram would have been preferable, we used CBF as a general indicator of a systemic vascular response to CO2, and not as a surrogate of pulmonary vascular reactivity. Exercise markedly suppressed CO2-mediated increases in CBF, indicating that the vasculature was not responding to the increased pCO<sup>2</sup> levels with the degree of dilation seen in control animals. There are several potential mechanisms for this. An increase in circulating H<sup>+</sup> resulting from CO<sup>2</sup> metabolism would normally have a direct effect on peripheral vascular

### REFERENCES


smooth muscle leading to vessel dilation (Kontos et al., 1977), but chronic exercise and CO<sup>2</sup> exposure may have desensitized the vasculature to elevated H<sup>+</sup> and disrupted normal cerebrovascular autoregulation. Alternatively, CO<sup>2</sup> and H<sup>+</sup> can induce formation of reactive oxygen and reactive nitrogen species and may influence brainstem mediated adaptations to hypercapnia and hypoxia (Dean, 2010). Though both cerebral and pulmonary tissues demonstrate a high degree of blood flow autoregulation, local responses to hypoxia are quite different and it is possible this extends to hypercapnia as well. Further investigations comparing vascular autoregulation in the brain vs. lung would be informative. Our findings may also be of interest in the context of High Altitude Cerebral Edema, which was not studied here.

### CONCLUSION

Chronic intermittent CO<sup>2</sup> exposure and exercise training do not affect the incidence of HAPE in rats, but exercise training does attenuate CO<sup>2</sup> induced increases in CBF.

### AUTHOR CONTRIBUTIONS

RS, AH, and RM were responsible for study design and conceptualization. RS and JS collected and analyzed the data. RS compiled the initial manuscript. RS, JS, AH, and RM revised the manuscript and approved the submitted version.

## FUNDING

This work was supported by the Office of Naval Research work unit# 0602236N.0000.000.A1509.

### ACKNOWLEDGMENTS

The authors are grateful to Amber Zhou, Eve Laurent, William Porter, William Hickman, Austin Headley, and Michael Bodo for their technical assistance.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Sheppard, Swift, Hall and Mahon. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Arachidonic Acid Metabolism Pathway Is Not Only Dominant in Metabolic Modulation but Associated With Phenotypic Variation After Acute Hypoxia Exposure

Chang Liu1,2,3, Bao Liu1,2,3,4, Lu Liu1,2,3, Er-Long Zhang1,2,3, Bind-da Sun1,2,3, Gang Xu1,2,3 , Jian Chen1,2,3 \* and Yu-qi Gao1,2,3 \*

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Alla B. Salmina, Krasnoyarsk State Medical University named after Prof. V.F.Voino-Yasenetski, Russia Muhammad Aslam, Justus Liebig Universität Gießen, Germany

### \*Correspondence:

Jian Chen jchenone@163.com Yu-qi Gao gaoy66@yahoo.com

### Specialty section:

This article was submitted to Vascular Physiology, a section of the journal Frontiers in Physiology

Received: 19 September 2017 Accepted: 02 March 2018 Published: 16 March 2018

### Citation:

Liu C, Liu B, Liu L, Zhang E-L, Sun B, Xu G, Chen J and Gao Y (2018) Arachidonic Acid Metabolism Pathway Is Not Only Dominant in Metabolic Modulation but Associated With Phenotypic Variation After Acute Hypoxia Exposure. Front. Physiol. 9:236. doi: 10.3389/fphys.2018.00236 1 Institute of Medicine and Hygienic Equipment for High Altitude Region, College of High Altitude Military Medicine, Army Medical University, Third Military Medical University, Chongqing, China, <sup>2</sup> Key Laboratory of High Altitude Environmental Medicine, Army Medical University, Third Military Medical University, Ministry of Education, Chongqing, China, <sup>3</sup> Key Laboratory of High Altitude Medicine, People's Liberation Army, Chongqing, China, <sup>4</sup> The 12th Hospital of Chinese People's Liberation Army, Kashi, China

### Background: The modulation of arachidonic acid (AA) metabolism pathway is identified in metabolic alterations after hypoxia exposure, but its biological function is controversial. We aimed at integrating plasma metabolomic and transcriptomic approaches to systematically explore the roles of the AA metabolism pathway in response to acute hypoxia using an acute mountain sickness (AMS) model.

Methods: Blood samples were obtained from 53 enrolled subjects before and after exposure to high altitude. Ultra-performance liquid chromatography-quadrupole time-of-flight mass spectrometry and RNA sequencing were separately performed for metabolomic and transcriptomic profiling, respectively. Influential modules comprising essential metabolites and genes were identified by weighted gene co-expression network analysis (WGCNA) after integrating metabolic information with phenotypic and transcriptomic datasets, respectively.

Results: Enrolled subjects exhibited diverse response manners to hypoxia. Combined with obviously altered heart rate, oxygen saturation, hemoglobin, and Lake Louise Score (LLS), metabolomic profiling detected that 36 metabolites were highly related to clinical features in hypoxia responses, out of which 27 were upregulated and nine were downregulated, and could be mapped to AA metabolism pathway significantly. Integrated analysis of metabolomic and transcriptomic data revealed that these dominant molecules showed remarkable association with genes in gas transport incapacitation and disorders of hemoglobin metabolism pathways, such as ALAS2, HEMGN. After detailed description of AA metabolism pathway, we found that the molecules of 15-d-PGJ2, PGA2, PGE2, 12-O-3-OH-LTB4, LTD4, LTE4 were significantly up-regulated after hypoxia stimuli, and increased in those with poor response manner to hypoxia particularly. Further analysis in another cohort

**44**

showed that genes in AA metabolism pathway such as PTGES, PTGS1, GGT1, TBAS1 et al. were excessively elevated in subjects in maladaptation to hypoxia.

Conclusion: This is the first study to construct the map of AA metabolism pathway in response to hypoxia and reveal the crosstalk between phenotypic variation under hypoxia and the AA metabolism pathway. These findings may improve our understanding of the advanced pathophysiological mechanisms in acute hypoxic diseases and provide new insights into critical roles of the AA metabolism pathway in the development and prevention of these diseases.

Keywords: metabolomics, transcriptomics, WGCNA, hypoxia, arachidonic acid

## INTRODUCTION

Hypoxia is a common phenomenon associated with multiple pathophysiological processes, including inflammation, necrosis, and apoptosis, and observed in various clinic conditions such as myocardial ischemia and reperfusion as well as stroke (Michiels, 2004). The arachidonic acid (AA) metabolism pathway was shown to exhibit remarkable metabolic alterations after hypoxia exposure, but its role is controversial (Liao et al., 2016). As a part of the innate immune response, AA and its metabolites are known to induce or suppress inflammation (Bennett and Gilroy, 2016). In addition, the products of the AA cascade are shown to display multidirectional and often conflicting roles in endothelial and nerve cells, such as inhibition or promotion of cell proliferation and aggravation or alleviation of oxidative stress, in response to stimuli (Bogatcheva et al., 2005; Rink and Khanna, 2011). Although the partial function of this pathway was identified, the mechanism underlying the expression of the AA metabolism pathway and its role under acute hypoxic conditions are questionable.

Acute hypobaric hypoxia, one of the most common hypoxia types, occurs in response to the exposure to high altitude and initiates systematic pathophysiological alterations. Acute hypobaric hypoxia may serve as a principal etiological factor for acute mountain sickness (AMS) (Sun et al., 2017). Therefore, AMS is considered as a potent channel and an in vivo model to improve our understanding of the various acute hypoxic diseases with high-quality samples from self-controlled studies (Subudhi et al., 2010). In addition, mal-adaptation to hypoxia is characterized by symptoms such as headache, nausea, and vomiting that may progress into high altitude pulmonary edema or high altitude cerebral edema-the leading lethal diseases for mountaineers (Johnson and Luks, 2016). With ∼15 million people traveling to areas of high altitude (>2,500 m) annually in China alone (Gonggalanzi et al., 2016), it is important to unravel the mechanisms underlying AMS.

The technical platforms of system biology allow researchers to explore systematic alterations under hypoxia exposure. Metabolomic studies have demonstrated that metabolic reprogramming with dramatic turbulence in various metabolic pathways, including linoleic acid metabolism, purinergic signaling, and glycolysis, was implied in response to acute hypoxia (D'Alessandro et al., 2016; Liao et al., 2016). In addition, a recent study on acute hypoxia stimuli based on RNA-seq approach indicated that the disturbance in the inflammation response presented by the reduction in interleukin (IL)-10 was highly related to AMS incidences (Liu et al., 2017a). However, studies based on a single set of omics technology failed to extensively explain the alterations occurring at different levels during physiologic or pathophysiological responses, such as disabilities in interpreting metabolic regulatory processes in metabolomics (Kan et al., 2017). The development of efficient analysis methods and algorithms allows better insights, beyond those available from individual strategies through the integration of diverse omics (Li et al., 2016; Sun and Hu, 2016). In particular, the method of weighted gene co-expression network analysis (WGCNA), which could exhibit multiple network-level interactions, is effective for the identification of gene-to-metabolite relations and construction of the underlying complex regulatory networks (Acharjee et al., 2016; Liu et al., 2017b).

Here, we aimed to integrate clinical features, metabolomic datasets, and transcriptome profiling for the construction of the co-expression network to provide a better understanding of the AA metabolism pathway after acute hypoxia exposure. Our results identified that the AA metabolism pathway, a highly regulated metabolic pathway, was essential in response to acute hypoxia. Furthermore, the AA metabolism pathway may account for variations in response patterns after acute hypoxia stimuli. Our results may improve the understanding of AMS and provide new clues to counteract the effects of hypoxic diseases.

# MATERIALS AND METHODS

### Enrolled Subjects and Experimental Procedures

We enrolled 53 male subjects with intact clinical phenotype data of heart rate (HR), oxygen saturation (SpO2), Lake Louise score (LLS), systolic blood pressure (SBP), diastolic blood pressure (DBP), and hemoglobin (HB) before and after high altitude exposure (5,300 m). These subjects were covered in our previous study (Liao et al., 2016). Physical and physiological characteristics [mean ± standard deviation (SD)] for enrolled subjects were as follows: age, 21.6 ± 2.0 years; height, 172 ± 1 cm; and body weight, 64.9 ± 2.3 kg. Exclusion criteria included prior exposure to high altitude, presence of cardiovascular pathologies or other diseases, and history of drug prescribed for high altitude. These volunteers traveled to an area of 5,300 m from the starting point of 1,400 m within 4 days, as illustrated in our previous article (Liao et al., 2016). Their clinical symptoms of HR, SpO2, LLS, SBP, DBP, HB were monitored before and after hypoxia stimuli. The local ethics committee of the Third Military Medical University approved the experimental design, methods of clinical data collection, and strategies of data evaluation. These procedures were performed in accordance with the approved guidelines. Written informed consents were obtained from all participants.

### Metabolomics Analysis

Metabolic profiling was performed, and metabolic data were obtained as previously described (Liao et al., 2016). Briefly, blood was drawn from a peripheral vein in the morning soon after waking before ascending to high altitudes and after 3 days high-altitude exposures. Blood samples were centrifuged to separate serum at 14,000 × g for 15 min at 4 ◦C. Then, these plasma samples were delivered to Chongqing for further metabolomics analysis in the courier filed with dry ice.

Metabolomics analysis was conducted on an Agilent 1290 Infinity LC system (Agilent, Santa Clara, CA, USA). Chromatographic separations were performed on an ACQUITY UHPLC HSS T3 C18 column (2.1 × 100 mm, 1.8µm; Waters, Milford, Ireland) at 45◦C. The metabolomics detection was performed totally in accordance to the manufacturers' protocols of all devices. Raw LC-MS data were converted to mzData formats via Agilent MassHunter Qualitative software (Agilent, Santa Clara, CA, USA). The program XCMS (version 1.40.0) (https://xcmsonline.scripps.edu/) was used to preprocess the raw data, including peak detection, peak matching, matched filtration, and nonlinear alignment of data, with the default parameters (and snthresh, 5; bw, 10; fwhm, 10).

The resulting matrix from 53 subjects comprising retention time, mass-to-charge ratio (m/z), and normalized ion intensities was introduced into the subsequent analysis based on R platform (https://www.r-project.org). The model of partial squares discriminant analysis (OPLS-DA) was carried out to separate pre- and post-hypoxia in metabolic profiling. Variable importance project (VIP) was used to reveal discriminatory metabolites which were enrolled in the classification effects of hypoxia was calculated. In addition, adjusted p-value from the paired t-test and fold-change value (FC) were executed to discovery dominant metabolites. Pathway analysis was established by MetaboAnalyst platform (http://www. metaboanalyst.ca).

### Transcriptomic Profiling Detection

To improve the reliability of the integrated analysis to uncover interactions at metabolic and transcriptional level, 11 individuals were randomly selected and subjected to metabolomic and RNA-seq detection (Li et al., 2016). RNA sequencing was performed as previously described (Liu et al., 2017a) with minor modifications. Briefly, after extraction, total RNA was quantified by spectrophotometer (NanoDrop 200, Thermo Scientific, Delaware, USA). mRNA-seq libraries were constructed according to the TruSeq RNA Sample Prep Kit v2 (Illumina) and sequenced on an Illumina HiSeq 2000 sequencer, following the manufacturer's instructions. after filtered, the clean reads were aligned to human genome hg19 based on HISAT and assembled by Stringtie (Li et al., 2014). The expression value of each transcript was subsequently calculated based on methods of Fragments Per Kilobase of exon model per Million mapped fragments (FPKM). In addition, GSE52209 datasets of blood samples from individuals suffering from adaptation or mal-adaptation to high altitude exposure were downloaded from Gene Expression Omnibus (GEO) datasets. After annotation, limma package was employed to normalize and detect the significantly differential genes involved in diverse response manner to hypoxia (Ritchie et al., 2015). ClusterProfiler package was used to determine enriched Gene Ontology (GO) terms in differentially expressed gene sets (Yu et al., 2012).

### Weighted Gene Co-expression Network Analysis (WGCNA)

WGCNA, a robust tool for the integrative network analysis, is widely used in complex network construction (Liu et al., 2017b). According to the WGCNA protocol, we constructed clusters of genes firstly based on the matrix of pairwise correlations calculated across the selected samples. Then, we performed the network construction and module detection based on an appropriate soft-thresholding power. The default minimum cluster merge height of 0.25 was retained. These clusters identified by WGCNA are called modules, which represented a group of highly interconnected genes with similar expression profiles across the enrolled subjects. Further, we determined correlations among expression modules and clinical traits for all subjects. Hub genes were identified based on the connectivity among all nodes. Network was illustrated by Cytoscape software (Smoot et al., 2011).

### Statistical Analysis

Differences before and after hypoxia exposure were determined by paired t-test. The unpaired Student's t-test was used to compare the expression between subjects with diverse response patterns to hypoxia. Statistical analysis was performed by GraphPad software (GraphPad Prism, USA). The data in this study were presented as the mean ± SD. A two-tailed p value <0.05 was considered as significant, unless specifically indicated.

# RESULTS

### Baseline Clinical Characteristics of Enrolled Subjects

The response of subjects to acute hypobaric hypoxia was assessed by various physiological parameters. As shown in **Figure 1A**, SpO2, reflective of the oxygen concentration in the blood, significantly declined under hypoxia environment (p < 0.05; Netzer et al., 2017). Among the essential compensatory

measurements, HR (**Figure 1B**), SBP (**Figure 1C**), DBP (**Figure 1D**), and HB (**Figure 1E**) increased along with hypoxia (p < 0.05), indicative of the comprehensive response manner to acute hypoxia stimuli. A significant increase in LLS was observed for all individuals (**Figure 1F**; p < 0.05).

# Arachidonic Acid Metabolism Pathway Shows a Remarkable Correlation With Phenotypic Alterations After Hypoxia Exposure

Metabolomic analysis of blood samples from individuals before and after high altitude exposure revealed a total of 1,720 compounds. As indicated in **Figure 2A**, the method of WGCNA identified that four modules (MEgreen, MEred, MEmagenta, and MEblack) showed significant correlations with all clinical features of SBP, LBP, HR, SpO2, LLS, and HB. To further explore influential metabolites in these modules, a volcano plot for these molecules was obtained (**Figure 2B**). At a cutoff point of p < 0.05, FC ≥ 1.5, and VIP > 1.5, a total of 36 metabolites showed significant alteration and comprised nine downregulated molecules and 27 elevated metabolites (**Table 1**). Moreover, the enrichment analysis demonstrated that AA metabolism pathway was particularly significant in the metabolic alterations in response to acute hypoxia (**Figure 2C**).

FIGURE 2 | Metabolic profiling analysis in response to acute hypoxia exposure. (A) WGCNA analysis identified four connectivity-based modules with high correlation to clinic features of heart rate (HR), oxygen saturation (SpO2), Lake Louise score (LLS), systolic blood pressure (SBP), diastolic blood pressure (DBP), and hemoglobin (HB). High correlation values are indicated by red and negative correlations, in green color. (B) Comparison of all metabolites from plasma of subjects before and after hypoxia exposure. The volcano plot displays the relationship between fold change and significance using a scatter plot view. The green points in the plot represent the differential metabolites with statistical significance. (C) Enrichment pathway analysis for the dominant metabolites identified in the volcano plot before and after hypoxia exposure.



### Integrative Network-Based Analysis Revealed Mechanisms of AA Metabolism Pathway

The physiological characteristics corresponding to SpO2, HR, HB, LLS, SBP, and LBP for subjects evaluated by metabolomics and transcriptomics are listed in Supplemental Figure 1. These subjects chose for both omics detection, posed clinical features similar to those of the whole population after hypoxia stimuli. To provide a comprehensive interaction network between metabolomic and transcriptomic data sets, WGCNA was performed to identify connectivity-based modules comprising genes closely related to the aforementioned impact metabolites. As illustrated in **Figure 3A**, gene modules labeled by a violet frame were significantly correlated with these molecules, especially to metabolites of the AA metabolism pathway. As shown in **Figure 3B**, heme catabolism, gas transport, and porphyrin-related metabolism pathways were dominant in these gene modules after GO analysis (p < 0.05). To further clarify the mechanisms underlying the AA metabolism pathway after hypoxia exposure, genes with the highest connectivity from the co-expression network and a correlation coefficient >0.65 to at least one molecule among 13 metabolites in the AA metabolism pathway are circled in red in **Figure 3C**. These marked genes promoted proliferation and regulation of red blood cell metabolism. The detailed information is listed in **Figure 4**.

# Detailed AA Metabolism Pathway Along With Hypoxia Stimuli

A map of metabolites in AA metabolism pathway was constructed in **Figure 4**. Alterations in these molecules before and after hypoxia exposure are listed in the blue box next to each metabolite. To elucidate the differences in subjects exhibiting diverse response patterns to hypoxia, these individuals were divided into two groups based on their LLS scores. Individuals with an LLS score lower than 4 were enrolled in the control group, while those with an LLS score higher than 9 were included in the group of mal-adaptation to hypoxia. Clinical characteristics of each group are listed in Supplemental Table 1. The metabolites exhibiting significant differences between control and maladjustment group are shown in the red box in **Figure 4**. As observed, the AA metabolism pathway was upregulated after hypoxia exposure and further elevated in those with poor response to hypoxia.

# Validation of AA Metabolism Pathway in Another Dataset

We validated our findings in another population of GSE52209 from GEO database that comprised 14 subjects as the control group and 17 individuals suffering from mal-adaptation to high altitude. In comparison to individuals that adapted to hypoxia, those with poor response to acute hypoxia showed a significant upregulation (p < 0.05, **Figure 5**) in the expression of genes such as GGT1, PTGES, PTGS1, CBR1, and ALOX12 encoding key enzymes of the AA metabolism pathway. The effects of these genes on AA metabolism pathways are labeled with a yellow triangle in **Figure 4**.

# DISCUSSION

Here, we performed an integrated analysis of metabolomic and transcriptomic expression profiling with the whole blood of individuals that experienced acute hypoxia. These approaches revealed pivotal roles of the AA metabolism pathway in the phenomenon of metabolic reprogramming during acute hypoxia exposure. Co-expression network analysis further identified that the AA metabolism pathway showed a significant positive correlation with transcriptional alterations, particularly with genes involved in porphyrin metabolism and gas transport pathways. Taken together, this is the first study to provide

systematic changes in the AA metabolism pathway and reveal its potential mechanisms in response to hypoxia exposure.

We found an increase in the expression of molecules such as AA, prostaglandin A2 (PGA2), and cysteinyl leukotrienes (Cys–LTs) in AA metabolism pathway after hypoxia exposure. These molecules were important sources of pro-inflammatory mediators. The elevation in the inflammatory response and oxidative stress may be important in AMS pathogenesis (Boos et al., 2016; Liu et al., 2017a). AA has been demonstrated to activate p38 mitogen-activated protein kinase (MAPK) and

c-Jun N-terminal kinase (JNK), thereby directly promoting inflammation by increasing levels of tumor necrosis factor (TNF)-α (Saito et al., 2012). AA-derived metabolites such as leukotriene B4 (LTB4) and Cys–LTs from upregulated enzymes, including ALOX5 and GGT1, may operate as potent activators of and chemoattractants for leukocytes, eosinophils, and monocytes to propagate inflammation and oxidative stress (Rink and Khanna, 2011). Elevated PTGES may contribute to increased levels of PGE2 and PGA2, which activate IL-6 and nuclear factor kappa B (NF-κB) to promote vascular inflammation (Gomez et al., 2013). In addition to the biological effects of these metabolites, key enzymes in the AA metabolism pathway may directly trigger inflammation reactions and oxidative stress. Cyclooxygenase-2 (COX-2) contributes to the upregulation of TNF-α after hypoxia stimuli (Xing et al., 2015). The peroxidase activity of PTGS1 may serve as a source of oxygen radicals through the conversion of PGG2 to PGH2 by the removal of oxygen (Rink and Khanna, 2011). Moreover, phospholipase A2 may produce free radicals during the activation of arachidonic acid cascade under hypoxia environment (Tanaka et al., 2003).

With limited evidence focusing on the role of the AA metabolism pathway in the development of hypoxic diseases, our results systematically reveal the expression level of the AA metabolism pathway as well as its possible roles based on the AMS model and provide the basis for further research.

The increase in metabolites indicates the upregulation in sub-pathways activated by COX, lipoxygenase, and PLA2—the key enzymes of the AA metabolism pathway (Wong et al., 2003; Jantan et al., 2014). In line with previous studies, these proteins showed a hypoxia-regulated pattern. A recent study demonstrated that hypoxia-mediated extracellular signalregulated kinase (ERK) activation may induce the activity of PLA2 in the erythrocyte membrane (Wu et al., 2016). In addition, as an essential oxygen-sensing molecule, heme oxygenase-1 was reported to act as an important mediator of transcription and translation of 15-lipoxygenase to promote lipoxins (Nie et al., 2013). Hypoxia-inducible factor-1 alpha (HIF-1a) was reported to directly stimulate COX enzymes after hypoxia exposure (Huang et al., 2016). However, we failed to detect the quantitative expression of the key enzymes in the AA metabolism pathway, owing to the limited blood sample during validation. An intervention experiment with drugs targeting the AA metabolism pathway in larger population samples is now being conducted by our group to verify these findings and search for effective prevention measures.

In our study, we demonstrated a crosstalk between response manners to acute hypoxia and expression level of the AA metabolism pathway. The excessive increase in the AA metabolism pathway after hypoxia stimuli may reflect maladaptation to hypoxia stimuli, as evident from the elevated levels of HR, HB, and LLS as well as the decreased level of SpO2. These observations suggest impaired cardiopulmonary and hemoglobin function (Fuehrer and Huecker, 2017). For cardiopulmonary performance, the addition of AA was shown to aggravate respiration inhibition and decrease oxygen consumption at the mitochondrial level (Egorova et al., 2015), while targeting COX-2 improved hemodynamic parameters of cardiac output, left ventricular pressure, and LVdp/dt (Oshima et al., 2006). AA metabolism was also reported to influence heart rates by alerting histaminergic system (Altinbas et al., 2014). In addition, the elevated level of reactive oxygen species during AA metabolism was involved in the regulation of erythroid differentiation, and the activation of phospholipase A2 was shown to induce protein kinase C to increase erythrogenin level (Luo et al., 2017) (Mason-Garcia and Beckman, 1991). Furthermore, the process of reticulocyte maturation was dependent on phospholipase A2 during the remodeling of the plasma membrane by the removal of specific proteins (Blanc et al., 2007). As an irreversible pathological process with poor clinical outcomes caused by sustained hypoxia, elevated AA metabolism pathway may decrease the apoptosis of endothelial cells, induce the proliferation of pulmonary artery smooth muscle cells, activate angiogenesis, and influence the arterial vascular tonicity via K+ channels (Park et al., 2011; Zhang et al., 2011; Liu et al., 2012; Zhu and Ran, 2012; Pang et al., 2016). An excessive elevation in AA metabolism may result in poor response to acute hypoxia and even worse symptoms under chronic hypoxia environment.

The protective effects of anti-inflammatory molecules are increasingly recognized for the prevention of AMS. Apigenin targeting IL-1β, IL-6, and TNF-α was demonstrated to be effective in reducing the damage caused by acute hypoxia (Du et al., 2015). Recent pilot studies reported that ibuprofen, aspirin, and dexamethasone may be effective for the prevention and treatment of AMS, especially the syndrome of headache, although the underlying mechanisms remain largely unknown (Burtscher et al., 2001; Gertsch et al., 2012; Xiong et al., 2017). In this study, our results theoretically support anti-inflammatory drugs such as non-steroidal anti-inflammatory drugs (NSAIDS) and dexamethasone targeting the AA metabolism pathway for improving clinical symptoms of AMS. In addition, targeted therapies against specific enzymes or metabolites of the AA metabolism pathway may be beneficial in AMS prevention

# REFERENCES


and demand further studies (Atluri et al., 2011). Given its close relationship with hemoglobin metabolism genes, the AA metabolism pathway may be an effective target for the prevention of hypoxia-induced erythrocytosis (Foley et al., 1978; Parise et al., 1991).

In summary, our integrated analysis of metabolomic and transcriptomic profiling revealed that the AA metabolism pathway was one of the most pivotal alterations after acute hypoxia exposure and may account for variations in response patterns to hypoxia stimuli. The detailed description of the AA metabolism pathway may offer the basis for follow-up mechanisms and drug screening.

### AVAILABILITY OF DATA

All data have been submitted to GEO under the accession GSE103940.

# AUTHOR CONTRIBUTIONS

YG and JC: Conceived and designed the study; CL and BL: Oversaw laboratory analyses and contributed the statistical analysis; LL, E-LZ, and BS: Contributed to sample and physical data collections; CL: Drafted the report. All authors reviewed and approved the manuscript.

### ACKNOWLEDGMENTS

This work was supported by the Key Projects in the Military Science & Technology Pillar Program during the Thirteen 5-year Plan Period (AWS14C007&AWS16J023), by National Natural Science Foundation of China (J1310001).

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphys. 2018.00236/full#supplementary-material


different than at its simulation in normobaric hypoxia. Front. Physiol. 8:81. doi: 10.3389/fphys.2017.00081


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Liu, Liu, Liu, Zhang, Sun, Xu, Chen and Gao. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Sensory Processing and Integration at the Carotid Body Tripartite Synapse: Neurotransmitter Functions and Effects of Chronic Hypoxia

Erin M. Leonard, Shaima Salman and Colin A. Nurse\*

Department of Biology, McMaster University, Hamilton, ON, Canada

Maintenance of homeostasis in the respiratory and cardiovascular systems depends on reflexes that are initiated at specialized peripheral chemoreceptors that sense changes in the chemical composition of arterial blood. In mammals, the bilaterally-paired carotid bodies (CBs) are the main peripheral chemoreceptor organs that are richly vascularized and are strategically located at the carotid bifurcation. The CBs contribute to the maintenance of O2, CO2/H+, and glucose homeostasis and have attracted much clinical interest because hyperactivity in these organs is associated with several pathophysiological conditions including sleep apnea, obstructive lung disease, heart failure, hypertension, and diabetes. In response to a decrease in O<sup>2</sup> availability (hypoxia) and elevated CO2/H<sup>+</sup> (acid hypercapnia), CB receptor type I (glomus) cells depolarize and release neurotransmitters that stimulate apposed chemoafferent nerve fibers. The central projections of those fibers in turn activate cardiorespiratory centers in the brainstem, leading to an increase in ventilation and sympathetic drive that helps restore blood PO<sup>2</sup> and protect vital organs, e.g., the brain. Significant progress has been made in understanding how neurochemicals released from type I cells such as ATP, adenosine, dopamine, 5-HT, ACh, and angiotensin II help shape the CB afferent discharge during both normal and pathophysiological conditions. However, type I cells typically occur in clusters and in addition to their sensory innervation are ensheathed by the processes of neighboring glial-like, sustentacular type II cells. This morphological arrangement is reminiscent of a "tripartite synapse" and emerging evidence suggests that paracrine stimulation of type II cells by a variety of CB neurochemicals may trigger the release of "gliotransmitters" such as ATP via pannexin-1 channels. Further, recent data suggest novel mechanisms by which dopamine, acting via D2 receptors (D2R), may inhibit action potential firing at petrosal nerve endings. This review will update current ideas concerning the presynaptic and postsynaptic mechanisms that underlie chemosensory processing in the CB. Paracrine signaling pathways will be highlighted, and particularly those that allow the glial-like type II cells to participate in the integrated sensory response during exposures to chemostimuli, including acute and chronic hypoxia.

Keywords: carotid body, chemoreceptor type I cells, glial-like type II cells, purinergic signaling, neurotransmitters, sensory transmission, petrosal neurons

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Silvia V. Conde, Centro de Estudos de Doenças Crónicas (CEDOC), Portugal Julio Alcayaga, Universidad de Chile, Chile

> \*Correspondence: Colin A. Nurse nursec@mcmaster.ca

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 20 December 2017 Accepted: 28 February 2018 Published: 16 March 2018

### Citation:

Leonard EM, Salman S and Nurse CA (2018) Sensory Processing and Integration at the Carotid Body Tripartite Synapse: Neurotransmitter Functions and Effects of Chronic Hypoxia. Front. Physiol. 9:225. doi: 10.3389/fphys.2018.00225

# INTRODUCTION

Oxygen (O2) is essential to the survival of aerobic organisms that rely on oxidative phosphorylation for the production of ATP as a major energy source. Consequently, both waterand air-breathing vertebrates have evolved mechanisms for monitoring O<sup>2</sup> levels in the environmental water, pulmonary airways, and/or arterial blood. In conditions of O<sup>2</sup> deficiency (hypoxia), strategically-located sensors in the gills of waterbreathers and peripheral chemoreceptor organs of air-breathers initiate compensatory respiratory and cardiovascular reflex responses so as to maintain homeostasis (Gonzalez et al., 1994; Cutz and Jackson, 1999; Milsom and Burleson, 2007; Kumar and Prabhakar, 2012). In mammals, the main peripheral chemoreceptor organs are the carotid bodies (CBs), which contain O2- and CO2/H+-sensitive detectors known as glomus or type I cells (Gonzalez et al., 1994; Peers and Buckler, 1995; López-Barneo et al., 2016). While the chemoreceptive properties of the CB have been well established over much of the last century (Gonzalez et al., 2014), current views consider the CBs as general metabolic sensors capable of detecting not only the respiratory gases and blood acidity, but also blood glucose and circulating insulin levels (López-Barneo, 2003; Conde et al., 2014; Thompson et al., 2016).

The CBs are located bilaterally at the carotid bifurcation where they monitor blood O<sup>2</sup> just before it reaches the brain, an organ that is a critically sensitive to O<sup>2</sup> deprivation. In keeping with their role as circulatory chemical sensors, they are reputed to be the most richly vascularized organs in the body (Gonzalez et al., 1994). The gross morphology of the CB reveals a network of fenestrated capillaries that penetrate clusters of chemoreceptor type I cells which receive sensory innervation from the carotid sinus nerve (CSN) (McDonald, 1981). The type I cells share an intimate association with neighboring glial-like, sustentacular type II cells in a ratio of ∼4:1 (McDonald, 1981; Kondo, 2002), and there is evidence of synapse-like specializations between these two cell types (Platero-Luengo et al., 2014). Thus, the CB chemoreceptor complex, consisting of clusters of receptor type I cells, type II glial cells, and the abutting sensory nerve terminals, displays the features of a "tripartite synapse" that provides the substrate for synaptic integration at many sites in the central nervous system (CNS) (Eroglu and Barres, 2010). During chemoexcitation, the CB output is relayed as an increase in action potential frequency in chemosensory fibers of the CSN whose cell bodies are located in the petrosal ganglia. These fibers terminate centrally in the nucleus tractus solitarius where they help control lung ventilation and sympathetic outflow to the cardiovascular system (Claps and Torrealba, 1988; Gonzalez et al., 1994; Kumar and Prabhakar, 2012).

Over the last ∼30 years there have been numerous studies on the transduction mechanisms by which CB receptor type I cells sense a reduction in PO<sup>2</sup> (hypoxia) and an elevation in PCO2/H<sup>+</sup> and how these signals are relayed to the brain via the CSN. On the presynaptic side, there is a consensus that acute hypoxia causes inhibition of background K<sup>+</sup> channels in type I cells, and this is a key step in chemotransduction leading to voltage-gated Ca2<sup>+</sup> entry and neurotransmitter release

(Buckler, 2015; López-Barneo et al., 2016). However, despite significant recent progress, the identity of the PO<sup>2</sup> sensor is still debatable though there is strong support for a role of the mitochondrial electron transport chain (Buckler, 2015; López-Barneo et al., 2016; Chang, 2017; Prabhakar and Peng, 2017; Rakoczy and Wyatt, 2017). On the postsynaptic side, there has been significant progress in elucidating the roles of several neurochemicals, especially the purines ATP and its metabolite adenosine, in shaping the afferent CSN discharge (Iturriaga and Alcayaga, 2004; Nurse, 2005, 2010; Conde et al., 2012a; Nurse and Piskuric, 2013). However, this task has become more challenging given the increasing number of neuroactive agents present in the CB, and the growing evidence that the gliallike type II cells may not be silent partners during sensory transmission (Tse et al., 2012; Nurse and Piskuric, 2013; Nurse, 2014). An understanding of the roles of these neurochemicals and the mechanisms underlying synapse integration in the CB is important given that alterations in CB sensitivity and CSN discharge are now linked to several cardiorespiratory abnormalities in humans (Kumar and Prabhakar, 2012; Iturriaga, 2017). For example during obstructive sleep apnea, a condition where patients are exposed to bouts of chronic intermittent hypoxia (CIH), CB chemosensitivity becomes exaggerated leading to long-term facilitation (LTF) of the sensory discharge and increased risk of hypertension (Prabhakar et al., 2015). Also, CB chemosensitivity is enhanced in chronic heart failure (CHF), leading to sympathetic hyperactivity and exacerbation of the disease progression (Schultz et al., 2015). In this article, we review the role of several neurotransmitters and neuromodulators in the integrative sensory response at the CB tripartite synapse. In particular, we consider the major transmitters involved in the interactions between type I cells and the sensory nerve ending during chemotransduction, as well as the contribution of paracrine signaling mechanisms to synaptic integration and crosstalk among type I cells, type II cells, and the sensory nerve endings. Finally, we consider how synaptic plasticity mechanisms during exposures to chronic sustained hypoxia might contribute to alterations in neurochemical signaling at the tripartite synapse.

### SYNAPTIC TRANSMISSION BETWEEN CHEMORECEPTOR TYPE I CELLS AND SENSORY NERVE TERMINALS

During chemotransduction type I cells release neurotransmitters resulting in an increase in CSN action potential frequency. As discussed below, the main transmitters responsible for this postsynaptic excitatory response are likely ATP and adenosine; however, contributions from other CB transmitters including ACh, histamine, and 5-HT cannot be ignored and may well depend on developmental age, species under investigation, or the presence or absence of pathophysiological conditions (Zhong et al., 1999; Iturriaga and Alcayaga, 2004; Nurse, 2005, 2010; Lazarov et al., 2006; Del Rio et al., 2008; Conde et al., 2012a; Kumar and Prabhakar, 2012). Further, the magnitude of the CSN excitatory response appears to be blunted by the concurrent action of inhibitory neurotransmitters such as GABA and dopamine (Alcayaga et al., 1999; Iturriaga et al., 2009; Nurse, 2010). In the case of dopamine, it is possible that its release from activated afferent C fibers during chemoexcitation may lead to autocrine-paracrine stimulation of inhibitory D2 receptors (D2R) on nearby type I cells and/or afferent nerve terminals (Benot and Lopez-Barneo, 1990; Almaraz et al., 1993; Iturriaga et al., 2003).

## Role of ATP and Postsynaptic P2X2/3 Receptors

The stimulatory effects of intracarotid administration of ATP on ventilation and CSN discharge have been known for some time (reviewed by Zapata, 2007). However, use of a co-culture model of rat type I cell clusters and juxtaposed petrosal neurons demonstrated that blockers of purinergic P2X2/3 receptors, i.e., suramin and PPADS, inhibited the postsynaptic response elicited by hypoxia and/or hypercapnia (Zhang et al., 2000; Prasad et al., 2001; Zhang and Nurse, 2004; Nurse and Piskuric, 2013). A central role of ATP acting via postsynaptic P2X2/3 receptors during chemoexcitation was confirmed by the following observations: (i) Chemosensitive petrosal neurons in co-culture expressed functional P2X2/3 receptors (Zhang et al., 2000); (ii) Immunoreactive P2X2 and P2X3 subunits were localized to petrosal terminals opposed to type I cells in CB tissue sections in situ (Zhang et al., 2000; Prasad et al., 2001); (iii) P2X2 knockout mice showed a markedly attenuated hypoxic ventilatory response (Rong et al., 2003); (iv) Acute hypoxia induced Ca2+-dependent, vesicular ATP release from isolated whole CB, CB slices, and cultured CB cells (Buttigieg and Nurse, 2004; Conde et al., 2012a); (v) In intact CB-sinus nerve preparations in vitro, P2X2/3 and selective P2X3 receptor antagonists inhibited the CSN discharge evoked by hypoxia (He et al., 2006; Reyes et al., 2007; Niane et al., 2011); and (vi) In the co-culture model, the selective P2X receptor blocker PPADS inhibited the hypoxia-induced postsynaptic response in the petrosal neuron, but not the presynaptic receptor potential in type I cells (**Figures 1A1,A2**). Though most of the preceding data were obtained in rodent preparations, an excitatory postsynaptic role of ATP acting via ligand-gated receptor channels has also been observed in the cat CB (Iturriaga and Alcayaga, 2004; Reyes et al., 2007; Zapata, 2007).

# Role of Adenosine and Postsynaptic A2a Receptors

The excitatory effects of exogenous adenosine on both ventilation and CSN discharge were first confirmed in the 1980's and appear to be species independent (McQueen and Ribeiro, 1983; Monteiro and Ribeiro, 1987; Conde et al., 2006, 2009). While presynaptic adenosine A2a and A2b receptors on type I cells are well described (see later), the excitatory effects of exogenous adenosine appear to be mainly postsynaptic, involving high affinity A2a receptors (A2aR) that are expressed in petrosal chemoafferent neurons (Gauda, 2002; Conde et al., 2006, 2009). During acute hypoxia and hypercapnia a significant fraction of the CSN sensory discharge is dependent on the stimulation of postsynaptic A2aR (Conde et al., 2006; Zhang et al., 2017). The extracellular adenosine level at the CB chemosensory synapse is generated via equilibrative nucleoside transporters (ENTs) in type I cells and/or breakdown of extracellular ATP by ectonucleotidases (Conde and Monteiro, 2004). More recently, molecular characterization and immunolocalization of surfacelocated nucleotidases in the rat CB have revealed the presence of

from a type I cell (A1) and a petrosal neuron (A2) in co-culture. The P2X2/3 receptor blocker PPADS (10µM) reversibly suppressed the "postsynaptic" neuronal response (A2), but had negligible effect on the "presynaptic" type I cell response (A1). (B) The hypoxia (hox)-induced sensory discharge in a co-cultured petrosal neuron (left trace) was reversibly inhibited by the nicotinic cholinergic blocker mecamylamine (mec; 1µM; middle trace). Data adapted from Nurse (2010).

nucleotidase triphosphate diphospho-hydrolase (NTPDase)2,3 and ecto-5′nucleotidase(ecto-5′Nt/CD73) in association with type I cell clusters (Salman et al., 2017). The NTPDase 2,3 pair efficiently hydrolyses ATP to AMP and ADP, while ecto-5 ′Nt hydrolyses AMP to adenosine (Zimmermann, 2000). Interestingly, in a recent study pharmacological inhibition of ecto-5′NT, but not ENTs, caused a reduction in basal and hypoxia-evoked CB sensory discharge, as well as in the hypoxic ventilatory response in adult rats (Holmes et al., 2017). These data suggest that the principal source of adenosine that contributes to the increase in CSN discharge during acute hypoxia is from the catabolism of extracellular ATP. However, in a previous study, pharmacological inhibition of ENT was found to increase adenosine release from rat CBs especially at high-intensity hypoxia, at least when adenosine deaminase was simultaneously inhibited (Conde et al., 2012a). Thus, the relative contributions of ecto-5′Nt and ENT to extracellular adenosine during acute hypoxia may well depend on the PO<sup>2</sup> level. A plausible mechanism by which adenosine, acting via postsynaptic A2aR, increases excitability in petrosal afferent neurons was recently proposed (Zhang et al., 2017). In that study on rat CB co-cultures, adenosine caused a depolarizing shift in the voltage dependence of activation of a hyperpolarization-activated cyclic nucleotide gated cation current Ih in chemosensory neurons. Consistent with the involvement of high affinity A2aR, the effect was mediated by nanomolar concentrations of adenosine and was reversibly inhibited by the selective A2aR blocker, SCH58261. In concert with the known effects of the Ih current in regulating firing frequency in a variety of cell types (Biel et al., 2009), adenosine was also shown to increase membrane excitability and action potential frequency in identified chemosensory petrosal neurons in co-culture (Zhang et al., 2017). Moreover, HCN4 immunoreactive subunits were localized to chemosensory, tyrosine hydroxylase (TH)-positive petrosal neurons in tissue sections of rat petrosal ganglia in situ (Zhang et al., 2017); HCN4 subunits are known to contribute to Ih channel currents and to increases in action potential frequency through elevations in intracellular cAMP in different cell types (Biel et al., 2009).

### Role of ACh and Nicotinic Receptors

The role of ACh as a major excitatory neurotransmitter in the CB has had a controversial history (Eyzaguirre and Zapata, 1984; Nurse and Zhang, 1999; Fitzgerald, 2000; Iturriaga and Alcayaga, 2004). There is no doubt that various nicotinic and muscarinic ACh receptors (AChR) are expressed in the CB of several species, though whether or not type I cells synthesize and store ACh in all cases is controversial (Gauda, 2002; Iturriaga and Alcayaga, 2004; Zhang and Nurse, 2004; Shirahata et al., 2007). However, hypoxia-induced release of ACh from intact CBs has been detected in several species including cats, rabbits, and humans (Fitzgerald, 2000; Shirahata et al., 2007; Zapata, 2007; Kåhlin et al., 2014). Moreover, in the co-culture model of the rat CB a combination of nicotinic and purinergic blockers were required to inhibit most of the hypoxia-induced postsynaptic response (Zhang et al., 2000); often, only partial inhibition was seen when either blocker was present alone (**Figures 1A2,B**). In concert with these results, systemic blockade of both nicotinic and purinergic P2X receptors inhibited ventilation in newborn rats (Niane et al., 2012). Additional evidence supporting ACh as an important CB excitatory neurotransmitter was obtained in the following studies: (i) functional nicotinic AChR were present on >65% of isolated rat petrosal neurons (Zhong and Nurse, 1997) and in the majority of petrosal neurons that were functionally connected to the cat CB (Varas et al., 2003; Iturriaga and Alcayaga, 2004); (ii) in the isolated rat and cat CB-sinus nerve preparations in vitro nicotinic AChR blockers inhibited the hypoxia-induced chemosensory discharge (Iturriaga and Alcayaga, 2004; He et al., 2005; Zapata, 2007; Niane et al., 2009); and (iii) consistent with a postsynaptic role for ACh, α7 nAChR immunoreactivity has been localized to nerve endings surrounding type I clusters in rat and cat CB in situ (Shirahata et al., 2007; Niane et al., 2009), and there is electrophysiological evidence consistent with the presence of functional α7 nAChR in cat petrosal neurons (Alcayaga et al., 2007).

# Role of Histamine and H1-3 Receptors

Histamine is synthesized, stored (in considerably higher amounts than dopamine), and released from the CB during acute hypoxia (Koerner et al., 2004; Del Rio et al., 2008). In addition, histamine receptors (H1, H2, H3) have been localized to both the CB and petrosal ganglion (Lazarov et al., 2006). Application of H1R and H3R agonists to the rat CB had a mild stimulatory effect on ventilatory output (Lazarov et al., 2006). In the cat, H1R antagonists reduced, whereas H3R antagonists enhanced, the excitatory effect of histamine on the sinus nerve discharge (Del Rio et al., 2008). In the latter study, though histamine increased sinus nerve discharge when applied to both the isolated superfused and perfused CB in vitro, it had no effect on the isolated petrosal ganglion discharge. This suggests a direct effect of histamine on the CB parenchymal cells and/or a selective effect on petrosal terminals that is not replicated at the soma. In isolated rat type I cells, H3R agonists inhibited the rise in Ca2<sup>+</sup> elicited by muscarinic agonists (Thompson et al., 2010). Further work is required to clarify the role of histamine and its receptors in the CB though both excitatory and inhibitory pathways appear to be present.

# Role of Inhibitory Neurotransmitters

The preceding sections have focused on postsynaptic interactions that are predominantly excitatory. However, inhibitory mechanisms appear to play an important role in regulating the CB sensory discharge. Dopamine is probably the best-described inhibitory CB neurotransmitter (Benot and Lopez-Barneo, 1990), yet it was only recently that a plausible postsynaptic mechanism was proposed (Zhang et al., 2017). In the latter study on rat CB co-cultures, dopamine caused a hyperpolarizing shift in the voltage dependence of I<sup>h</sup> activation and a decrease in membrane excitability in chemosensory petrosal neurons. This effect was opposite to that discussed above for adenosine, where the voltage dependence of I<sup>h</sup> activation was shifted in the depolarizing direction and resulted in increased membrane excitability. The effect of dopamine on I<sup>h</sup> was prevented during co-incubation with the selective D2 receptor (D2R) antagonist, sulpiride (Zhang et al., 2017). These data suggest that stimulation of D2R on chemosensory petrosal neurons and their terminals causes a decrease in intracellular cAMP leading to inhibition of the cAMP-gated I<sup>h</sup> channels containing HCN4 subunits (Zhang et al., 2017). In this scenario, the source of dopamine is predominantly from nearby type I cells though it may include terminals of excited catecholaminergic C fibers of petrosal CB chemoafferents (Almaraz et al., 1993; Iturriaga et al., 2003).

There is also evidence that GABA, acting via ionotropic GABA(A) receptors on petrosal afferent terminals, can inhibit the chemosensory discharge by a shunting mechanism which diminishes the depolarizing action of excitatory inputs (Zhang et al., 2009). GABA is synthesized and stored in type I cells and its release during chemoexcitation has been inferred since pharmacological blockade of GABA(A) receptors using bicuculline facilitates the hypoxia-induced sensory discharge in cat and rat carotid body (Igarashi et al., 2009; Zhang et al., 2009). Moreover benzodiazepines, which enhance GABA(A) receptor activity, suppress ventilation in cats and inhibit the CB chemosensory discharge evoked by acute hypoxia (Igarashi et al., 2009).

## AUTOCRINE-PARACRINE INTERACTIONS INVOLVING CHEMORECEPTOR TYPE I CELLS AND GLIAL-LIKE TYPE II CELLS

As discussed above, purinergic transmission from CB type I cells to the sensory nerve endings involves fast-acting ionotropic P2X receptors and slow-acting metabotropic adenosine P1 receptors. However, both autocrine and paracrine signaling mechanisms mediated mainly via metabotropic receptors on type I and glial-like type II cells can further fine-tune the synaptic neurotransmitter pool near the sensory nerve endings. In addition to direct synaptic interactions between type I cells and sensory nerve terminal the following cell- cell interactions are possible: (i) type I cell to type I cell; (ii) type I cell to type II cell; (iii) type II cell to type I cell; and (iv) direct communication between type II cells and the sensory nerve ending. Electrical coupling between adjacent type I cells, and between type I cells and the sensory nerve ending, is also possible as reviewed elsewhere (Eyzaguirre, 2007).

# Autocrine-Paracrine Signaling Among Neighboring Type I Cells

These pathways may involve either negative or positive feedback mechanisms. Perhaps the best-studied pathway involves the negative feedback role of dopamine acting on inhibitory D2R on the same or neighboring type I cells. For example, selective blockade of D2R using domperidone or haloperidol caused a concentration dependent enhancement of basal, high K+-, and hypoxia-evoked catecholamine release from intact rat CB (Conde et al., 2008). The effect is likely due to a negative feedback inhibition of intracellular Ca2<sup>+</sup> signaling because dopamine was found to inhibit L-type Ca2<sup>+</sup> currents in isolated type I cells (Benot and Lopez-Barneo, 1990), whereas selective D2R agonists inhibited the hypoxia-evoked rise in intracellular Ca2<sup>+</sup> (Carroll et al., 2005). In addition to DA, there is evidence that ATP and GABA may also inhibit type I cell function via negative feedback mechanisms. For example, ATP was found to inhibit the hypoxia-induced rise in intracellular Ca2<sup>+</sup> in type I cells via metabotropic P2Y1R (Xu et al., 2005) or P2Y12R (Carroll et al., 2012). Because ADP is also an effective ligand for both these receptors, and P2Y12Rs are highly expressed in type I cells (Zhou et al., 2016), it is possible that degradation of ATP by the surface-located nucleotidases NTPDase2,3 (Salman et al., 2017), generates sufficient ADP to further inhibit type I cells. Under normoxic conditions, NTPDase2 mRNA expression in the CB is much higher than NTPDase3 mRNA (Salman et al., 2017). Assuming parallel changes in the corresponding proteins, this expression pattern favors elevated ADP levels since NTPDase2 is relatively ineffective at hydrolyzing ADP to AMP (Zimmermann, 2000). GABA acting on autocrine-paracrine GABA(B) receptors was reported to inhibit the hypoxia-evoked receptor potential in type I cells (Fearon et al., 2003). In the latter study, the signaling mechanism involved activation of a resting TASKlike K<sup>+</sup> conductance via a pertussis toxin- sensitive and PKAdependent pathway.

The type I cell response to chemostimuli such as acute hypoxia may also be modified by positive feedback mechanisms. For example, the ATP metabolite adenosine has been shown to enhance intracellular Ca2<sup>+</sup> signaling in type I cells by triggering membrane depolarization and voltage-gated Ca2<sup>+</sup> entry via adenylate cyclase-PKA dependent inhibition of background TASK channels (Xu et al., 2006). Also, there is evidence that 5-HT acting via 5-HT2a receptors (5-HT2aR) can enhance the hypoxia-evoked receptor potential in type I cells via an autocrineparacrine positive feedback mechanism involving PKC-mediated inhibition of resting and Ca2+-dependent K<sup>+</sup> conductances (Zhang et al., 2003).

# ATP Release From Type 1 Cells During Chemostimulation Can Lead to Intracellular Ca2<sup>+</sup> Elevations in Nearby Type II Cells

The idea that glial-like type II cells may participate in the integrative CB sensory response was first suggested by the observation that the excitatory transmitter ATP can elicit a rise in intracellular Ca2<sup>+</sup> in isolated type II cells (Xu et al., 2003). Though these cells display small voltage-dependent outward currents, they lack significant inward Na<sup>+</sup> or Ca2<sup>+</sup> currents (Duchen et al., 1988; Xu et al., 2003; Zhang et al., 2012), consistent with an intracellular source for the ATP-mediated Ca2<sup>+</sup> rise (Xu et al., 2003). This effect of ATP was mimicked by selective P2Y2 receptor (P2Y2R) agonists such as UTP, and P2Y2R immunoreactivity was detectable in spindle-shaped type II cells in tissue sections of the rat CB in situ (Xu et al., 2003). As exemplified in **Figure 2C**, subsequent studies using the agonist UTP confirmed the P2Y2R-mediated increase in intracellular Ca2<sup>+</sup> in rat type II cells (Piskuric and Nurse, 2012; Zhang et al., 2012). P2Y2Rs typically couple via Gq-proteins to activate the phospholipase C-IP3-PKC signaling pathway

FIGURE 2 | Purinergic activation of the inward current in type II cells is mediated by P2Y2 receptors in association with a rise in intracellular Ca2+. Both (A) ATP (50µM) and (B) UTP (50µM) induced similar inward currents (at −60 mV) in the same type II cell as expected for P2Y2 receptors. (B) Dose–response relation (black curve) for the UTP-evoked currents in a group of type II cells; data are represented as means ± SEM (n = 4) and fitted with the Hill equation with an EC<sup>50</sup> = 37µM. The dose–response curve for ATP (red) is superimposed and is indistinguishable from that for UTP. (C) Intracellular Ca2<sup>+</sup> [Ca2+] i responses monitored simultaneously in type I and type II cells. Both hypercapnia (10% CO2) and high K<sup>+</sup> (30 mM) induced [Ca2+] i responses in the type I cell (upper trace), but not in the two type II cells (lower red and green traces). Conversely, ATP and UTP induced similar [Ca2+] i responses in the type II cells, but the type I cell was unresponsive. Data taken from Zhang et al. (2012).

and mobilize Ca2<sup>+</sup> from internal stores (von Kügelgen, 2006), suggesting this mechanism is likely responsible for the Ca2<sup>+</sup> elevation in type II cells. Because ATP is a key excitatory CB neurotransmitter the question arose whether paracrine stimulation of type II cells by ATP occurs during normal CB chemoexcitation. To address this, isolated rat type I cell clusters containing contiguous type II cells where challenged with chemosensory stimuli such as acute hypoxia and hypercapnia, as well as the depolarizing stimulus, high (30 mM) K<sup>+</sup> (Murali and Nurse, 2016). As expected, all three stimuli evoked rapid intracellular Ca <sup>2</sup><sup>+</sup> elevations in receptor type I cells but, interestingly, nearby type II cells frequently responded with a delayed, secondary increase in intracellular Ca2<sup>+</sup> (Murali and Nurse, 2016), as illustrated in **Figures 3A,B**. The paracrine action of ATP released from type I cells appeared to be responsible for the secondary type II cell Ca2<sup>+</sup> responses because the latter were reversibly inhibited by the P2Y2R antagonist, suramin (**Figure 3A**), as well as by the nucleoside hydrolase, apyrase (**Figure 3B**).

## Several Other CB Neurochemicals Stimulate a Rise in Intracellular Ca2<sup>+</sup> in Type II Cells

In addition to ATP (see above), several other neuroactive chemicals in the CB cause a rise in intracellular Ca2<sup>+</sup> when applied to a significant proportion of isolated type II cells. For example, the small molecule neurotransmitters ACh and 5- HT, acting via metabotropic muscarinic and 5-HT2 receptors respectively, stimulated intracellular Ca2<sup>+</sup> transients in ∼53 and 67% of UTP-sensitive type II cells respectively (Tse et al., 2012; Murali et al., 2015, 2017). In addition, nanomolar concentrations of the neuropeptide angiotensin II stimulated a rise in intracellular Ca2<sup>+</sup> in a significant proportion (∼75%) of type II cells via losartan-sensitive AT<sup>1</sup> receptors (Tse et al., 2012; Murali et al., 2014). The biosynthetic pathway for producing angiotensin II, including angiotensin converting enzyme (ACE) and the precursor angiotensinogen, is expressed in type I cells (Leung et al., 2000; Lam and Leung, 2003), suggesting it could play a paracrine role in the CB. In all cases, the intracellular Ca2<sup>+</sup> response in type II cells persisted in Ca2+-free extracellular solutions and/or following store depletion with thapsigargin or cyclopiazonic acid, suggesting Ca2<sup>+</sup> release from internal stores (Tse et al., 2012; Murali et al., 2014, 2017). Although less well-studied, another neuropeptide, i.e., endothelin-1 (ET-1), that is synthesized and released by type I cells during hypoxia (Chen et al., 2002; Platero-Luengo et al., 2014), is also capable of evoking intracellular Ca2<sup>+</sup> responses in type II cells at nanomolar concentrations (Murali et al., 2015). Whether or not all of these neuroactive agents combine to enhance Ca2<sup>+</sup> signaling in type II cells during chemoexcitation, as described above for ATP, remains to be determined. It is also possible that

FIGURE 3 | Blockade of P2Y2 receptors with suramin or degradation of extracellular ATP with apyrase inhibits crosstalk from type I to type II cells. (A) Example intracellular Ca2<sup>+</sup> traces showing the reversible inhibition of the delayed or indirect Ca2<sup>+</sup> response in a type II cell (blue) by the P2Y2R blocker suramin (100µM) during stimulation of nearby type I cells (red) with high K<sup>+</sup> (30 mM). (B) Application of apyrase (a nucleoside hydrolase) reversibly inhibits the delayed Ca2<sup>+</sup> response in a type II cell during stimulation of nearby type I cells with high (10%) CO2. Data adapted from Murali and Nurse (2016).

their actions may become more relevant in pathophysiological conditions associated with chronic or intermittent hypoxia, when their expression level and/or that of their cognate receptors are elevated (Chen et al., 2002; Lam et al., 2004).

### DOWNSTREAM EFFECTS OF NEUROTRANSMITTER-MEDIATED INTRACELLULAR Ca2<sup>+</sup> SIGNALING IN TYPE II CELLS

There is abundant evidence that glial cells in the CNS contribute to synapse integration by generating intracellular Ca2<sup>+</sup> signals and releasing gliotransmitters such as ATP, GABA, and glutamate (Eroglu and Barres, 2010; Bazargani and Attwell, 2016). As discussed above, several CB neurotransmitters elicited intracellular Ca <sup>2</sup><sup>+</sup> elevations in type II cells raising questions about the downstream consequence, and particularly whether this led to the release of gliotransmitters.

### Carotid Body Neurotransmitters Activate Pannexin-1 (Panx-1) Channels in Type II Cells

In electrophysiological studies on isolated type II cells, the P2Y2R agonists ATP or UTP activated an inward current at the resting membrane potential (Zhang et al., 2012), as illustrated in **Figures 2A,B**. This current reversed direction near 0 mV and, as illustrated in **Figure 4A**, was reversibly inhibited by low concentrations of carbenoxolone (5µM), a putative blocker of pannexin-1 (Panx-1) channels. Panx-1 immunoreactivity also co-localized with isolated GFAP-positive type II cells in dissociated CB cultures (see **Figure 4D**), and stained processes of type II cells in CB tissue sections (Zhang et al., 2012). Panx-1 channels assemble as hexamers in the plasma membrane and though structurally similar to gap junction hemichannels, Panx-1 sequences show closer homology to the invertebrate innexins (Dahl, 2015; Penuela et al., 2015). In addition to ATP, other CB neurotransmitters that evoked a rise in intracellular Ca2<sup>+</sup> in type II cells such as angiotensin II, ACh, and 5- HT also activated an inward Panx-1-like current based on its sensitivity to carbenoxolone (Murali et al., 2014, 2015, 2017). However, carbenoxolone is also a well-known blocker of gap junctional connexins, albeit at typically higher (>10x) concentrations (Lohman and Isakson, 2014), raising questions about the true identity of the channels. In recent studies, the UTP-evoked inward current in type II cells was reversibly inhibited by a more specific Panx-1 mimetic peptide channel blocker <sup>10</sup>Panx (100µM) (see **Figures 4B,C**), but not by similar concentrations of scrambled control peptide scPanx (Murali et al., 2017). Taken together, the combined pharmacological and immunocytochemical data, summarized in **Figure 4**, support the notion that Panx-1 channels mediate the neurotransmitteractivated inward currents in type II cells.

### Activation of Pannexin-1 Channels in Type II Cells Is Dependent on Intracellular Ca2<sup>+</sup>

The observation that several neurotransmitters that elicited a rise in intracellular Ca2<sup>+</sup> also activated Panx-1 currents in type II cells led to an investigation of a possible link between the two events (Murali et al., 2014). Indeed, as illustrated in **Figures 5A–D**, the Panx-1 inward current activated by either angiotensin II or ATP was reversibly inhibited when the membrane-permeable Ca2<sup>+</sup> chelator BAPTA-AM was present in the extracellular solution (Murali et al., 2014). These results are in concert with previous demonstrations that: (i) ATP activated an inward current at negative potentials in oocytes co-expressing Panx-1 channels and P2Y2R (Locovei et al., 2006); (ii) in inside-out patches from these oocytes Panx-1 channels were strongly activated at negative potentials when the cytoplasmic face was exposed to micromolar (but not 0) Ca2+; and (iii) in other glial cell types, e.g., microglia and astrocytes, Panx-1 channel opening is regulated by intracellular Ca2<sup>+</sup> (Dahl, 2015). However, it should be noted that the activation of Panx-1 channels by cytoplasmic Ca2<sup>+</sup>

(2017).

may be dependent on cell type, since in hippocampal neurons Panx-1 channel activation is Ca2+-independent (Thompson, 2015).

### Pannexin-1 Channels as Conduits for Release of the "Gliotransmitter" ATP From Type II Cells

Panx-1 channels have pores large enough to allow passage of molecules <1.5 kDa in various cell types including astrocytes and central neurons (Dahl, 2015; Penuela et al., 2015; Thompson, 2015). These molecules include "gliotransmitters" such as ATP, GABA, and glutamate. Given the central role of ATP in CB neurotransmission (Nurse, 2014), it was of interest to determine whether the Panx-1 channels in type II cells acted as ATP release channels. To test this, P2XR-expressing petrosal neurons served as ATP biosensors in the CB co-cultures containing type 1/type II cell clusters (Zhang et al., 2012). In that study, selective stimulation of P2Y2R on type II cells with UTP led to depolarization and/or increased firing in several nearby petrosal neurons. These petrosal responses were reversibly inhibited by blockers of P2X2/3 receptors (PPADS) or Panx-1 channels (carbenoxolone) suggesting they arose from ATP released through Panx-1 channels (Zhang et al., 2012). Though petrosal neuronsfunctioned as ATP biosensors in the latter study, the results led to the intriguing possibility that simply stimulating type II glial cells alone was sufficient to excite petrosal afferent terminals at the CB tripartite synapse, apparently without type I cell involvement. This occurred in spite of the unfavorably geometry present in this monolayer co-culture model, increasing the likelihood that it may also occur in vivo. If so, the proposed interaction could be facilitated by the close relation between type II cell processes and the afferent terminal (Platero-Luengo et al., 2014), as well as the possibility of physical coupling

between P2X2/3R and Panx-1 channels as recently shown in coimmunoprecipitation studies (Li et al., 2017). This potential for type II cells alone to directly excite petrosal nerve endings via ATP release has further implications. For example, conditions associated with an increase in circulating levels of compounds such as angiotensin II, e.g., CIH-induced hypertension and CHF (Schultz et al., 2015), could lead to enhanced CB excitation solely via stimulation of AT<sup>1</sup> receptors on type II cells. Additional independent support for the posit that type II cells can release ATP through Panx-1 channels was obtained in experiments using fura-2 Ca2<sup>+</sup> imaging to test for possible crosstalk from type II to type I cells (Murali and Nurse, 2016). In that study carbenoxolone reversibly inhibited ATP-dependent crosstalk from type II to type I cells, as discussed further below. The ability of the spindleshaped type II cells to release ATP, coupled with their expression of P2Y2R, also raise the possibility that intercellular Ca2<sup>+</sup> waves may propagate within the network of interconnected type II cells and thereby facilitate delivery of ATP to the synaptic region.

### ATP-Dependent Crosstalk From Type II to Type I Cells

The P2Y2R-[Ca2+]i-Panx-1 pathway discussed in the preceding sections implied a potential role for type II cells as ATP amplifiers due to the mechanism of "ATP-induced ATP release." In addition to causing excitation at the sensory nerve terminal via P2X2/3R, ATP released through Panx-1 channels could in turn inhibit type I cells by negative feedback mechanisms involving P2Y1R or P2Y12R as discussed earlier. However, extracellular ATP at the CB chemosensory synapse can be metabolized by ectonucleotidases to adenosine (Conde and Monteiro, 2004; Holmes et al., 2017; Salman et al., 2017), which is excitatory at both the sensory nerve ending and type I cells (see above). This raised the question whether P2Y2R-mediated activation of Panx-1 channels in type II cells could lead to positive feedback stimulation of neighboring type I cells. Indeed, stimulation of type II cells with UTP often led to a delayed secondary rise in intracellular Ca2<sup>+</sup> in nearby type I cells (Murali and Nurse, 2016). Consistent with a role for adenosine, generated from ATP catabolism after release through Panx-1 channels, the delayed type II cell Ca2<sup>+</sup> response was strongly inhibited by the Panx-1 blocker carbenoxolone, the adenosine A2aR blocker SCH58261, and the ecto-5′ -nucleotidase blocker AOPCP (Murali and Nurse, 2016). Thus, type II cells may participate in the integrated sensory response of CB by contributing to both ATP and adenosine synaptic pools.

### PLASTICITY IN CB NEUROTRANSMITTER FUNCTIONS DURING EXPOSURE TO CHRONIC HYPOXIA

A prominent feature of the CB is its remarkable plasticity following exposure of animals to different patterns of hypoxia including chronic sustained and chronic intermittent hypoxia or CIH (Kumar and Prabhakar, 2012). The reader is referred to other excellent reviews for a discussion of CB plasticity following CIH (Prabhakar et al., 2015; Iturriaga, 2017). Changes in ion channel expression leading to increased membrane excitability of type I cells, as well as alterations in a variety of neurotransmitter systems, are well-described CB plasticity mechanisms that occur during exposures to chronic sustained hypoxia (Prabhakar and Jacono, 2005; Powell, 2007). We focus here mainly on purinergic neurotransmitter mechanisms underlying CB plasticity during chronic hypoxia, as may occur during ascent to altitude. In the latter condition, an adaptive process known as ventilatory acclimatization to hypoxia (VAH) ensures an increase in sensitivity of the CB chemoreflex such that for a given PO<sup>2</sup> there is an augmented CSN discharge frequency over and above that present before acclimatization (Prabhakar and Jacono, 2005; Powell, 2007; Kumar and Prabhakar, 2012). Given the central role of ATP and adenosine in CB chemoexcitation as previously discussed, perhaps it is not surprising that alterations in purinergic signaling pathways have been implicated in VAH. For example, when adult rats were reared under chronic hypoxia (12% O2) for 7 days, infusion of the non-specific adenosine receptor antagonist 8-sulpho-phenyltheophylline (8- SPT) attenuated the increase in respiratory frequency evoked by acute hypoxia (Walsh and Marshall, 2006). Similarly when caffeine, a non-specific adenosine receptor antagonist, was included in the drinking water of chronically-hypoxic rats the CB sensory output was markedly inhibited (Conde et al., 2012b). These data point to a major contribution of adenosine signaling to the mechanisms underlying VAH.

Because surface-located ectonucleotidases control adenosine levels at the CB tripartite synapse (see above), it was of interest to determine whether expression of these enzymes was regulated by chronic hypoxia as occurs in other cell types, e.g., smooth muscle cells (Koszalka et al., 2004; Robson et al., 2005). In a recent study, exposure of rats to chronic hypobaric hypoxia (∼60 kPa, simulating an altitude of ∼4,300 m) for 5–7 days led to an ∼2x increase in expression of NTPDase3 and ecto-5′ -nucleotidase/CD73 mRNAs, concomitant with a significant decrease in expression of NTPDase1 and NTPDase2 mRNAs (Salman et al., 2017). Assuming changes in mRNA expression correlate with parallel changes in protein expression, the modifications seen in chronic hypoxia favor a shift toward adenosine signaling in the CB because NTPDase3 efficiently hydrolyses ATP to AMP, whereas ecto-5′ -nucleotidase catalyzes the conversion of AMP to adenosine in the ratelimiting step (Zimmermann, 2000). This shift could be further aided by the rapid depletion of the ATP and ADP pools because both purines are natural physiological inhibitors of ecto-5′ -nucleotidase (Lecka et al., 2010). Further, the tendency of ecto-5′nucleotidase and adenosine A2aR to co-localize, as previously demonstrated in the striatum (Augusto et al., 2013), could enable the efficiency of adenosine-A2aR interactions at the membranes of both type I cells and sensory nerve terminals. The proposed depletion of extracellular ATP and ADP pools during chronic hypoxia could facilitate type I cell depolarization and increased sensitivity by diminishing their negative feedback influences mediated by interactions with P2Y1R and/or P2Y12R on type I cells (Xu et al., 2006; Carroll et al., 2012). Evidence for an increased role for adenosine signaling during VAH is further supported by the observation that acute hypoxia-evoked adenosine release is markedly potentiated in isolated whole CBs from chronically hypoxic rats (Conde et al., 2012b). It is also possible that autoregulatory "push-pull" mechanisms could prevent adenosine-mediated overexcitation. For example, high levels of adenosine in the micromolar range could activate low affinity A2bR on type I cells and facilitate secretion of dopamine (Conde et al., 2008; Livermore and Nurse, 2013), which inhibits CB function via pre- and postsynaptic D2R on type I cells and petrosal nerve endings (see earlier discussion). Indeed, an increase in both basal and hypoxia-evoked dopamine release has been reported for chronically hypoxic CB chemoreceptor cells (Jackson and Nurse, 1997; Conde et al., 2012b). Taken together, these studies support a major shift toward adenosine-A2R signaling in the CB after exposure to chronic hypoxia in vivo.

# CONCLUSIONS AND FUTURE DIRECTIONS

In this review we considered the evidence that the integrated CB output is determined largely by neurochemical interactions at the tripartite synapse formed by type I chemoreceptor cells, type II glial cells, and sensory nerve endings. Several of these interactions involving fast-acting ionotropic receptors and slower-acting G-protein coupled receptors are summarized in **Figure 6**. There is a growing consensus that the purines ATP and adenosine are key players in mediating CB chemoexcitation and their actions involve P2 and P1 receptors located at preand postsynaptic sites. While the evidence in most species has long favored an inhibitory role for dopamine acting via D2R on type I cells, we considered novel data showing that dopamine-D2R signaling might also inhibit the sensory afferent discharge via modulation of the I<sup>h</sup> current that controls action potential frequency. Moreover, recent evidence suggests that the excitatory effects of P2Y2R-mediated Ca2<sup>+</sup> signaling in varying subpopulations of type II cells might be inhibited by dopamine (Leonard and Nurse, 2017), or histamine (Nurse et al., 2018). The possibility that other CB neurochemicals including small molecules (e.g., histamine and 5-HT) and neuropeptides (e.g., angiotensin II) might act postsynaptically on I<sup>h</sup> so as to regulate firing frequency requires future investigation. We also considered several G-protein coupled pathways by which type II glial cells might contribute to synapse integration; all of these converge on the Ca2+-dependent activation of Panx-1 channels which act as conduits for ATP release. These channels

are known to release other "gliotransmitters" in the CNS, raising the possibility that type II cells may also release other neuroactive chemicals to further shape or fine-tune the CB sensory output. For example, open Panx-1 channels in CNS astrocytes may stimulate release of lactate which serves as an extracellular energy substrate as well as a signaling molecule for glia-neuron communication (Karagiannis et al., 2016). In this regard, the lactate receptor Olfr78 that is highly expressed in type I cells has been proposed as a potential "hypoxia sensor" (Chang, 2017), raising the possibility that this pathway could be activated during acute hypoxia by lactate release from type II cells via Panx-1 channels (see **Figure 6**). Finally, this review considered some of the purinergic mechanisms that are likely to contribute to the increased sensitivity of the CB during exposure to chronic hypoxia. Whether similar or overlapping changes occur during pathophysiological conditions that lead to increased CB chemosensitivity, e.g., exposure to intermittent hypoxia (Kumar and Prabhakar, 2012), requires future investigation.

### AUTHOR CONTRIBUTIONS

EL performed experiments, analyzed data, helped prepare figures, reviewed literature and contributed to first draft. SS performed experiments, analyzed data, helped prepare figures, reviewed literature and contributed to first draft. CN organized review subheadings, helped interpret data, helped prepare figures, reviewed literature and contributed to first draft.

### ACKNOWLEDGMENTS

We thank several colleagues especially Min Zhang, Nikol Piskuric, Sindy Murali, and Cathy Vollmer for their contributions to some of the ideas and to several of the figures presented in this review. Grant support to CAN was provided by the Canadian Institutes of Health Research (MOP 12037 and MOP 142469) and the Natural Sciences and Engineering Research Council of Canada.

### REFERENCES


mediated by A2A and A2B adenosine receptors. J. Neurochem. 98, 616–628. doi: 10.1111/j.1471-4159.2006.03912.x


Health and Disease, Vol. 17, ed T. F. Horbein (New York, NY: Marcel Dekker), 105–320.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Leonard, Salman and Nurse. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Skeletal Muscle Pyruvate Dehydrogenase Phosphorylation and Lactate Accumulation During Sprint Exercise in Normoxia and Severe Acute Hypoxia: Effects of Antioxidants

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Ginés Viscor, Universitat de Barcelona, Spain Christos George Stathis, Victoria University, Australia Laurent André Messonnier, Université Savoie Mont Blanc, France

> \*Correspondence: José A. L. Calbet lopezcalbet@gmail.com

> > Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 10 September 2017 Accepted: 23 February 2018 Published: 19 March 2018

### Citation:

Morales-Alamo D, Guerra B, Santana A, Martin-Rincon M, Gelabert-Rebato M, Dorado C and Calbet JAL (2018) Skeletal Muscle Pyruvate Dehydrogenase Phosphorylation and Lactate Accumulation During Sprint Exercise in Normoxia and Severe Acute Hypoxia: Effects of Antioxidants. Front. Physiol. 9:188. doi: 10.3389/fphys.2018.00188 David Morales-Alamo1,2, Borja Guerra1,2, Alfredo Santana2,3, Marcos Martin-Rincon1,2 , Miriam Gelabert-Rebato1,2, Cecilia Dorado1,2 and José A. L. Calbet 1,2 \*

<sup>1</sup> Department of Physical Education, University of Las Palmas de Gran Canaria, Las Palmas de Gran Canaria, Spain, <sup>2</sup> Research Institute of Biomedical and Health Sciences, Las Palmas de Gran Canaria, Spain, <sup>3</sup> Clinical Genetics Unit, Complejo Hospitalario Universitario Insular-Materno Infantil de Las Palmas de Gran Canaria, Las Palmas de Gran Canaria, Spain

Compared to normoxia, during sprint exercise in severe acute hypoxia the glycolytic rate is increased leading to greater lactate accumulation, acidification, and oxidative stress. To determine the role played by pyruvate dehydrogenase (PDH) activation and reactive nitrogen and oxygen species (RNOS) in muscle lactate accumulation, nine volunteers performed a single 30-s sprint (Wingate test) on four occasions: two after the ingestion of placebo and another two following the intake of antioxidants, while breathing either hypoxic gas (PIO<sup>2</sup> = 75 mmHg) or room air (PIO<sup>2</sup> = 143 mmHg). Vastus lateralis muscle biopsies were obtained before, immediately after, 30 and 120 min post-sprint. Antioxidants reduced the glycolytic rate without altering performance or VO2. Immediately after the sprints, Ser293- and Ser300-PDH-E1α phosphorylations were reduced to similar levels in all conditions (∼66 and 91%, respectively). However, 30 min into recovery Ser293-PDH-E1α phosphorylation reached pre-exercise values while Ser300-PDH-E1α was still reduced by 44%. Thirty minutes after the sprint Ser293-PDH-E1α phosphorylation was greater with antioxidants, resulting in 74% higher muscle lactate concentration. Changes in Ser<sup>293</sup> and Ser300-PDH-E1α phosphorylation from pre to immediately after the sprints were linearly related after placebo (r = 0.74, P < 0.001; n = 18), but not after antioxidants ingestion (r = 0.35, P = 0.15). In summary, lactate accumulation during sprint exercise in severe acute hypoxia is not caused by a reduced activation of the PDH. The ingestion of antioxidants is associated with increased PDH re-phosphorylation and slower elimination of muscle lactate during the recovery period. Ser<sup>293</sup> re-phosphorylates at a faster rate than Ser300-PDH-E1α during the recovery period, suggesting slightly different regulatory mechanisms.

Keywords: sprint exercise, skeletal muscle, hypoxia, human, oxidative stress, pyruvate dehydrogenase, PDH

# INTRODUCTION

During sprint exercise in normoxia lactate accumulates in the recruited skeletal muscles (Jones et al., 1985; Cheetham et al., 1986; Parolin et al., 1999; Morales-Alamo et al., 2012), indicating that the flux through the pyruvate dehydrogenase (PDH) is not sufficient to avoid pyruvate accumulation and its corresponding reduction to lactate (Howlett et al., 1999). Compared to sprint exercise in normoxia, sprint exercise in severe acute hypoxia is accompanied by increased lactate accumulation (McLellan et al., 1990; Morales-Alamo et al., 2012) and greater reactive nitrogen and oxygen species (RNOS) production and more oxidative stress (Cuevas et al., 2005; Morales-Alamo et al., 2012; Morales-Alamo and Calbet, 2014). Animal experiments have shown that excessive RNOS production may reduce PDH activity (Churchill et al., 2005). It remains unknown whether the increased lactate production during sprint exercise in severe acute hypoxia is due, at least in part, to reduced activation of PDH.

Putman et al. (1995) showed that during repeated 30-s isokinetic sprints in normoxia lactate production was not dependent on O<sup>2</sup> delivery. However, we have recently shown that during the last 15 s of a 30-s isokinetic sprint in severe acute hypoxia skeletal muscle VO<sup>2</sup> is indeed limited by O<sup>2</sup> delivery (Calbet et al., 2015). Moreover, during submaximal exercise at 55% of the VO2max in acute hypoxia (FiO2: = 0.11), the level of PDH activation after 1 min cycling was lower than in normoxia (Parolin et al., 2000).

Increased acidification during exercise in hypoxia may reduce nicotinamide adenine dinucleotide phosphate-oxidase 2 (NOX2) activity, which is considered the main source of ion superoxide production during exercise (Simchowitz, 1985; Morgan et al., 2005; Brennan-Minnella et al., 2015), particularly during brief periods (10–15 min) of contractile activity (Jackson, 2015). Superoxide stimulates Ca2<sup>+</sup> efflux through the ryanodine receptors of the sarcoendoplasmic reticulum (Jackson et al., 2007; Dulhunty et al., 2017). The increased (Ca2+) in the sarcoplasm is necessary for the activation of PDH phosphatases, which dephosphorylate and activate PDH (Hucho et al., 1972; Wieland, 1983; Behal et al., 1993) (**Figure 1**). Thus, the increased muscle acidification during sprint exercise in hypoxia may reduce PDH activation by at least four mechanisms: (1) by decreasing Ca2<sup>+</sup> efflux from the sarcoplasmic reticulum (Bailey et al., 2007; Dulhunty et al., 2017), (2) by inducing the activation of pyruvate dehydrogenase kinase (PDK) (Churchill et al., 2005; Wu et al., 2013), (3) by causing oxidative damage to the enzyme (Tabatabaie et al., 1996), and (4) by reducing PDH phosphatase activity (Radak et al., 2013). In the case of an inhibitory influence of exercise-induced RNOS on PDH activation, the ingestion of antioxidants before sprint exercise should enhance PDH dephosphorylation and activation, leading to lower accumulation of lactate.

Pyruvate dehydrogenase is activated by dephosphorylation of serine residues (Ser232, Ser293, Ser300) located in the E1α subunit of PDH (Yeaman et al., 1978) (**Figure 1**). The balance between the activities of pyruvate dehydrogenase kinases (PDKs) and PDH phosphatases (PDP) determines the degree of PDH activation (Hucho et al., 1972; Wieland, 1983; Behal et al., 1993).

Therefore, the main aim of this study was to determine whether PDH activation, as determined by the assessment of its phosphorylation level, is reduced during sprint exercise in hypoxia compared to the same exercise performed in normoxia. Another aim was to ascertain whether antioxidants administration immediately before a sprint exercise in normoxia or hypoxia influences PDH activity during the subsequent recovery period.

### MATERIALS AND METHODS

### Subjects

Initially, 10 healthy male physical education students agreed to participate in this investigation, but completed data for all conditions were obtained in nine of them (age = 25 ± 5 year, height = 176.0 ± 5.1 cm, body mass = 79.4 ± 10.1 kg, body fat = 18.3 ± 6.7%, **Table 1**). The study was performed in accordance with the Declaration of Helsinki and was approved by the Ethical Committee of the University of Las Palmas de Gran Canaria (CEIH-2010-01). All subjects signed a written informed consent before entering the study.

### General Procedures

Body composition was determined by dual-energy x-ray absorptiometry (Hologic QDR-1500, Hologic Corp., software version 7.10, Waltham, MA) as described elsewhere (Calbet et al., 1998). After familiarization, subjects reported to the laboratory to complete different tests on separate days. First their peak VO<sup>2</sup> (VO2peak), maximal heart rate (HRmax) and maximal power output (Wmax) were determined in normoxia (FIO2: 0.21, PIO2: 143 mmHg) and severe acute hypoxia (FIO2: 0.105, PIO2: 75 mmHg) with an incremental exercise test to exhaustion (50 W/min), as previously described (Morales-Alamo et al., 2012, 2013, 2017). Hemoglobin oxygen saturation (SpO2) was determined with a finger pulse oximeter (OEMIII module, 4549- 000, Plymouth, MN). The main experiments consisted on a single 30-s isokinetic sprint at 100 revolutions per minute (Wingate Test) repeated on 4 different days, at least 1 week apart. The Wingate tests were carried out in normoxia (FIO2: 0.21, PIO2: 143 mmHg) and acute hypoxia (FIO2: 0.105, PIO2: 75 mmHg), both preceded by the ingestion of either placebo (200 mg of starch from corn, Sigma-Aldrich, ref. 14302) or antioxidants, with a double-blind design. Both the placebo and the antioxidants were introduced in microcrystalline cellulose capsules of identical characteristics.

Antioxidant supplements were administered in two doses separated by 30 min, with the first dose ingested 2 h before the sprint exercise, as previously reported (Morales-Alamo et al., 2017). The first dose contained 300 mg of α-lipoic acid, 500 mg vitamin C, and 200 IU vitamin E, whereas the second included 300 mg α-lipoic acid, 500 mg vitamin C, and 400 IU vitamin E (water dispersible). This antioxidant cocktail effectively decreases free radicals levels in blood as demonstrated by paramagnetic spin spectroscopy in humans (Richardson et al., 2007). We assume that similar effects were achieved in skeletal muscle as reflected by two facts. Firstly, these antioxidants blunted the RNOS-mediated signaling response elicited by sprint exercise

and, secondly, attenuated the formation of protein carbonyls in this study, as previously reported (Morales-Alamo et al., 2013, 2017).

The experimental days subjects reported to the laboratory early in the morning after a 12 h overnight fast. Then, an antecubital vein was catheterized and after 10 min of rest in the supine position a 20 ml blood sample was withdrawn and used to measure serum glucose and insulin. Right after, a muscle biopsy was obtained from the middle portion of the Vastus lateralis muscle using the Bergstrom's technique with suction (Guerra et al., 2011). After the resting muscle biopsy, the subjects sat on the cycle ergometer for 4 min while breathing either room air or a hypoxic gas mixture from a gas cylinder.

Within 10 s following the sprint, a second muscle biopsy was taken, and then another blood sample was obtained. During the following 2 h, the subjects had free access to water and sat quietly in the laboratory while two additional muscle biopsies and blood samples were obtained at 30 and 120 min post-Wingate. For the last two biopsies, a new incision was performed in the contralateral leg. To avoid injury-triggered activation of signaling cascades, the muscle biopsies were obtained at least 3 cm apart, using the procedures described by Guerra et al. (2011). The muscle specimens were fast cleaned to remove any visible blood, fat, or connective tissue. Then the muscle tissue was immediately frozen in liquid nitrogen and stored at −80◦C for later analysis.

# Oxygen Uptake

Oxygen uptake was measured with a metabolic cart (Vmax N29; Sensormedics, Yorba Linda, California, USA), as previously reported (Morales-Alamo et al., 2012, 2013, 2017). Respiratory variables were analyzed breath-by-breath and averaged every 5 s during the Wingate test and every 20 s during the incremental exercise tests (Calbet et al., 1997). The highest 20-s averaged VO<sup>2</sup> recorded in normoxia and hypoxia was taken as the normoxic and hypoxic VO2peak values, respectively.

# Muscle Metabolites

From each muscle biopsy, 30 mg of wet tissue were treated with 0.5 M HClO<sup>4</sup> and centrifuged at 15,000 g at 4◦C for 15 min. The supernatant was neutralized with KHCO<sup>3</sup> 2.1M. Phosphocreatine (PCr), creatine (Cr), muscle glucose (GLU), glucose-6-phosphate (G6P), glucose-1-phosphate (G1P), fructose-6-phosphate (F6P), pyruvate (Pyr), and lactate (Lac) were enzymatically determined in neutralized extracts by fluorometric analysis (Lowry and Passonneau, 1972; Gorostiaga et al., 2010). Alanine (Ala) muscle accumulation was assumed to be equivalent to 1.43-fold that of pyruvate accumulation (Katz et al., 1986). Muscle metabolite concentrations were adjusted to the individual mean total creatine (PCr + Cr) because this mean should remain constant during the exercise (Harris et al., 1976). The adjustment to total creatine content accounts for variability in solid non-muscle constituents, which may be present in the biopsies (Parra et al., 2000). This correction was applied to all TABLE 1 | Physical characteristics and ergoespirometric variables during incremental and sprint exercise in normoxia and severe acute hypoxia (mean ± SD).


SHRpeak, peak heart rate during the sprints; Wmax, maximal intensity during the incremental exercise test to exhaustion; Wpeak, peak power output during the Wingate test; LLM, lean mass of the lower extremities; Wmean, mean power output during the Wingate test; Accumulated VO2, amount of O<sup>2</sup> consumed during the 30 s Wingate test; SpO2, Hemoglobin saturation (pulse oximetry); PETO2, end tidal O<sup>2</sup> pressure. \*P < 0.05 compared to normoxia.

metabolites including lactate. This assumes that the increase in interstitial lactate and intracellular water was similar in the four trials and that the majority of the lactate produced during the sprints was retained inside the muscle fibers (Calbet et al., 2003). The glycolytic rate (GR) was calculated as GR = 0.5 × (1 Lac + 1 Pyr + 1 Ala), and glycogenolytic rate (GGR) as GGR = 1 G1P + 1 G6P + 1 F6P + 0.5 × (1 Lac + 1 Pyr + 1 Ala) (Chasiotis et al., 1982). The NAD+/NADH<sup>+</sup> and ATP/ADP ratios were calculated as previously reported (Morales-Alamo et al., 2017).

### Total Protein Extraction, Electrophoresis, and Western Blot Analysis

Muscle protein extracts were prepared as described previously (Guerra et al., 2007) with Complete protease inhibitor cocktail and PhosSTOP phosphatase inhibitor cocktail tablets (Roche Diagnostics, Mannheim, Germany). Total protein content was quantified using the bicinchoninic acid assay (Smith et al., 1985). Fifty microgram of each sample were subjected to immunobloting (Guerra et al., 2010) using Immun-BlotTM PVDF membranes from Bio-Rad Laboratories (Hemel Hempstead Hertfordshire, UK). Ser293-PDH-E1α and Ser<sup>300</sup> - PDH-E1α phosphorylation and total PDH were determined by western blot (50 µg in each lane) using specific antibodies (AP1062 and AP1064 from Calbiochem, Darmstadt, Germany; and sc-133898, Santa Cruz Biotechnology, CA, USA). Antibodies were diluted in 4% bovine serum albumin in Tris-buffered saline with 0.1% Tween 20 (TBS-T) (BSA-blocking buffer). To control for differences in loading and transfer efficiency, membranes were incubated with a monoclonal mouse antiα-tubulin antibody (T-5168-ML, Sigma, St. Louis, MO, USA) diluted in TBS-T with 5% blotting grade blocker non-fat dry milk (blotto-blocking buffer). No significant changes were observed in α-tubulin protein levels during the experiments. Antibody-specific labeling was revealed by incubation with a HRP-conjugated goat anti-rabbit antibody (1:20,000; 111-035- 144, Jackson ImmunoResearch, Baltimore, USA) antibody, both diluted in 5% blotto blocking buffer and specific bands were visualized with the Inmmun-StarTM WesternCTM kit, using the ChemiDoc XRS system (Bio-Rad Laboratories, Hercules, CA, USA) and analyzed with the image analysis program Quantity one© (Bio-Rad laboratories, Hercules, CA, USA). The densitometry analysis was carried out immediately before saturation of the immunosignal. Muscle-signaling data were represented as a percentage of immunostaining values obtained for the phosphorylated form of each kinase relative to the respective total form. Samples from each subject trial were run on the same gel. In addition, in all gels a human muscle sample obtained from a healthy young man was used as an internal control, to reduce inter-gel variability (Bass et al., 2017).

### Statistics

Variables were checked for normal distribution by using the Shapiro-Wilks test. A three-way repeated-measures ANOVA was used with three within subject factors: FIO<sup>2</sup> (with two levels: normoxia and hypoxia), antioxidants (with two levels: antioxidants and placebo), and time (with four levels: preexercise, end-exercise, 30 and 120 min recovery). The Mauchly's test of sphericity was run before the ANOVA, and in the case of violation of the sphericity assumption, the degrees of freedom were adjusted according to the Huynh and Feldt test. When a significant main effect or interaction was observed, specific pairwise comparisons were carried out with the least significant difference post-hoc test. The relationship between variables was determined using linear regression analysis. The areas under the curve (AUC) were determined using the trapezoidal rule and compared between conditions with paired Student t-tests. Values are reported as the mean ± standard deviation of the mean (unless otherwise stated). P ≤ 0.05 was considered significant. Statistical analysis was performed using SPSS v.15.0 for Windows (SPSS Inc., Chicago, IL).

# RESULTS

### Performance and Metabolism

Part of the effects on performance (**Table 1**) and muscle metabolites (**Table 2**) have been published previously (Morales-Alamo et al., 2012, 2013, 2017) and are only summarized here. Peak power output was similar between conditions. Mean power output was reduced by 5.3% in hypoxia (P = 0.03) and was not affected by the ingestion of antioxidants (558 ± 62 and 559 ± 61 W, placebo and antioxidants, respectively P = 0.81). The



 or

cP < 0.05 compared with normoxia antioxidants

dP < 0.05 compared with hypoxia placebo at the same time point. Pre, Pre sprint; Post, Post sprint; Post 30 min, 30 min into the recovery period; Antiox, antioxidants;

 at the same time point.

 Lac, lactate; Pyr, Pyruvate. n = 9.

accumulated VO<sup>2</sup> during the Wingate test was reduced by 30% in hypoxia (P = 0.006) and was not altered by the ingestion of antioxidants (1.095 ± 0.242 and 1.098 ± 0.267 L, placebo and antioxidants, respectively P = 0.99). The mean muscle lactate concentration at the end of the sprint was 23% lower after the ingestion of antioxidants (46.8 ± 13.4 and 36.2 ± 12.1 mmol.kg wet−<sup>1</sup> , mean of the two placebo and mean of the two antioxidants conditions, respectively, P = 0.004) and 30% higher in hypoxia than normoxia (46.9 ± 13.9 and 36.1 ± 12.2 mmol.kg wet−<sup>1</sup> , mean of the two hypoxia and the two normoxia conditions, respectively, P = 0.004). The ingestion of antioxidants reduced the glycolytic rate by 25% from 22.3 ± 6.8 to 16.9 ± 6.0 mmol.kg−<sup>1</sup> . (30 s)−<sup>1</sup> (P = 0.002), while the reduction in the glycogenolytic rate did not reach statistical significance (from 27.2 ± 7.0 to 25.4 ± 10.3 mmol.kg−<sup>1</sup> . (30 s)−<sup>1</sup> , for the mean of the two placebo and the two antioxidants conditions, respectively, P = 0.49). Compared to placebo, 30 min after the end of the sprints preceded by antioxidants muscle lactate concentration was 74% greater (4.3 ± 3.0 vs. 8.1 ± 4.5 mmol.kg wet−<sup>1</sup> , mean of the two placebo and the two antioxidants conditions, respectively, P = 0.02). The ratio ATP/ADP was similarly reduced after the sprints and recovered to pre-exercise levels 30 min later, regardless of the FIO<sup>2</sup> and antioxidants ingestion (**Table 2**). Immediately after the sprints, the NAD+/NADH<sup>+</sup> ratio was decreased more in hypoxia than normoxia (from 618 ± 355 to 42.5 ± 14.5 and from 452 ± 201 to 72.8 ± 28.6, mean of the two hypoxia and the two normoxia conditions, respectively, P = 0.003). The NAD+/NADH<sup>+</sup> ratio was lower after the administration of antioxidants (P = 0.03, antioxidants main effect) (**Table 2**).

Serum glucose and insulin concentrations were increased after the sprints (**Figure 2**). Although the changes in serum glucose concentration were similar after the four conditions, the serum insulin concentration tended to be marginally greater (+15%) after the sprints preceded by the ingestion of antioxidants compared to the placebo conditions (ANOVA time effect, P = 0.08). After the sprints performed in hypoxia, the elevation of plasma insulin was marginally greater than in normoxia (FIO<sup>2</sup> × time interaction, P = 0.045).

### PDH Phosphorylation

Immediately after the sprint Ser293-PDH-E1α phosphorylation was reduced to a similar level in all conditions (∼66%, P = 0.002; **Figures 3A–C**). However, 30 min after the sprint preceded by the ingestion of antioxidants Ser293-PDH-E1α phosphorylation reached pre-exercise values (P = 0.86). Ser293-PDH-E1α phosphorylation was 93% greater after the ingestion of antioxidants than placebo (ANOVA antioxidants effect P = 0.017; **Figure 3E**). Ser293-PDH-E1α was dephosphorylated to similar values immediately after the sprints in normoxia and hypoxia (P = 0.68, for the comparison between normoxia vs. hypoxia immediately after the sprints; **Figure 3G**).

Immediately after the sprints, Ser300-PDH-E1α phosphorylation was reduced by 91% (P < 0.001), without significant differences between conditions (ANOVA antioxidants × FIO<sup>2</sup> × time P = 0.40, **Figures 3A,B,D**). However, during the recovery period, the re-phosphorylation of Ser<sup>300</sup> was slower than that of Ser<sup>293</sup> (ANOVA phosphorylation type × time interaction P = 0.012), and 30 min after the end of the sprint Ser300-PDH-E1α phosphorylation was still 44% below pre-exercise values (P = 0.01). Two hours after the end of the sprint Ser300-PDH-E1α phosphorylation reached a value similar to that measured before the sprints (P = 0.80). The Ser300-PDH-E1α phosphorylation values during the experiment preceded by the ingestion of antioxidants were 50% greater than after the intake of placebo. However, this difference was not statistically significant (P = 0.14; **Figure 3F**). Ser300-PDH-E1α phosphorylation changes were similar in the normoxic and hypoxic sprints (ANOVA interaction: P = 0.29; **Figure 3H**).

The area under the curve (AUC) of Ser293-PDH-E1α phosphorylation was 87% greater after the ingestion of antioxidants (P = 0.01), indicating a greater level of rephosphorylation during the recovery period after the ingestion

FIGURE 3 | Representative western blots for Ser293-PDH-1Eα and Ser300-PDH-1Eα phosphorylations and PDH-1Eα total protein expression in response to a single 30 s sprint after placebo (A) or antioxidants (B) intake. Levels of Ser293-PDH-1Eα phosphorylation to PDH-1Eα total protein expression (C,E,G), and Ser300-PDH-1Eα phosphorylation to total PDH-1Eα total protein expression (D,F,H) before, immediately after, 30 and 120 min following the end of a single 30 s all-out sprint (Wingate test). (C,D) Responses to sprints performed in normoxia placebo (black circles), hypoxia placebo (open circles, FIO2: 0.105), normoxia antioxidants (black triangles) and hypoxia antioxidants (open triangles; FIO2: 0.105). (E,F) Represent the responses observed for two placebo conditions (gray circles) averaged compared to the average of the two antioxidants conditions (gray triangles). (G,H) Represent the responses for the average of the normoxic conditions (black squares) compared to the average of the hypoxic conditions (open squares; FIO2:0.105). \*<sup>P</sup> <sup>&</sup>lt; 0.05 in comparison to resting (ANOVA, time main effect). \$<sup>P</sup> <sup>&</sup>lt; 0.05 compared to Ser293-PDH-1Eα phosphorylation recovery. n = 9 for all variables.

of antioxidants. A similar trend was observed for the AUC of the Ser300-PDH-E1α phosphorylation (P = 0.061).

There was an association between the changes in Ser<sup>293</sup> and Ser300-PDH-E1α phosphorylation observed from pre, to immediately after the sprints in the placebo conditions (r = 0.74, P < 0.001; n = 18), that was lost in the sprints preceded by the ingestion of antioxidants (r = 0.35, P = 0.15). There was also a negative association between the level of Ser<sup>300</sup> - PDH-E1α phosphorylation at the end of the sprint and the mean and peak power during the sprint (r = −0.81, and r = −86, respectively, both P < 0.01, n = 9, each point representing the mean value of the four conditions for each subject). Likewise, a negative association was observed between the immediate post-exercise Ser293-PDH-E1α phosphorylation and the corresponding NAD+/NADH<sup>+</sup> ratio (r = −0.80, P = 0.01, n = 9, each point representing the mean value of the four conditions for each subject).

### DISCUSSION

Compared to normoxia, during sprint exercise in severe acute hypoxia the glycolytic rate is increased leading to greater muscle lactate accumulation, acidification and oxidative stress (Morales-Alamo et al., 2012). Here we show that this increased accumulation of lactate is not due to a lower level of PDH activation during the sprint in hypoxia, pointing toward a mitochondrial limitation of the rate of pyruvate decarboxylation to acetyl-CoA and subsequent oxidation of acetyl groups. In addition, by administering a powerful antioxidant cocktail before the sprint we have demonstrated that antioxidants reduce the glycolytic rate and muscle lactate accumulation during the sprints by a mechanism independent of PDH activation, that seems unrelated to FIO2. We have also demonstrated that after the ingestion of antioxidants, PDH is re-phosphorylated to a higher level, which may facilitate a metabolic shift from carbohydrates to fat oxidation (Sjøberg et al., 2017) during the recovery period, at the expense of delaying the removal of muscle lactate. Our results confirm previous in vitro studies showing that the re-phosphorylation of Ser300-PDH-E1α is slower than that of Ser293-PDH-E1α (Yeaman et al., 1978), indicating a slightly different regulation of these two phosphorylation sites in human skeletal muscle or reflecting intrinsic differences between the two phosphorylations. This is also supported by the negative association observed between the level of Ser300-PDH-E1α phosphorylation at the end of the sprint and the mean and peak power during the sprint, which was not observed in the case of Ser293-PDH-E1α. The most plausible link between power output and Ser300-PDH-E1α dephosphorylation is the increase of intracellular Ca2+, which is a fundamental determinant of power output (Bakker et al., 2017).

### PDH Dephosphorylation-Response to Sprint Exercise Is Not Modified by Antioxidants

The activity of PDH is regulated by phosphorylation/ dephosphorylation of three serine residues (Ser232, Ser<sup>293</sup> , Ser300) of the E1α subunit (Kolobova et al., 2001; Korotchkina and Patel, 2001) (**Figure 1**). Phosphorylation at any of the three sites leads to inhibition of the complex in vitro (Korotchkina and Patel, 1995). These phosphorylations are carried out by pyruvate dehydrogenase kinases, of which four isoforms have been identified (PDK1 to 4), while dephosphorylation is performed by PDH phosphatases, of which two isoforms have been identified (PDP1 and 2) (Teague et al., 1982; Huang et al., 1998). PDK1 is the only isoform reported to phosphorylate all three sites, while PDK2, PDK3, and PDK4 act on Ser<sup>293</sup> and Ser<sup>300</sup> in vitro (Kolobova et al., 2001; Korotchkina and Patel, 2001). The most abundant PDK in skeletal muscle is the isoform PDK4, followed by the isoform PDK2 and the less abundant is the isoform PDK1. Nevertheless, knockout mice for PDK2 and PDK4 have a constitutively active PDH in skeletal muscles (Rahimi et al., 2014).

Previous studies have shown that during sprint exercise in normoxia PDH activity is close to its maximum (Parolin et al., 1999), implying that in normoxia lactate accumulation is due to a limitation in the rate at which pyruvate is oxidized (Howlett et al., 1999). As a novelty, the present investigation shows that PDH is dephosphorylated to a similar extent in normoxia and hypoxia with or without antioxidants (**Figure 4**). Assuming that Ser<sup>293</sup> and Ser<sup>300</sup> PDH phosphorylations can be used to assess PDH activity indirectly (Linn et al., 1969; Pilegaard et al., 2006; Kiilerich et al., 2008, 2010), our data indicate similar levels of activation of PDH at the end of the sprint in all conditions. We have reported that during the sprints in severe acute hypoxia our subjects had higher oxidative stress than in normoxia, as shown by a 50% greater intramuscular protein carbonylation and more marked RNOSmediated signaling (Morales-Alamo et al., 2012). Moreover, the NAD+/NADH<sup>+</sup> ratio was decreased more after the sprints in hypoxia than normoxia (Morales-Alamo et al., 2012). NADH<sup>+</sup> is an activator of PDKs (Roche and Hiromasa, 2007), which inhibits PDH through serine phosphorylations (Hucho et al., 1972; Behal et al., 1993). However, PDH dephosphorylation was not greater immediately after the sprint in hypoxia than in normoxia, what is against a potentially greater level of PDK activity during the sprint in hypoxia than in normoxia.

Thus, both in normoxia and hypoxia, the main factor explaining the accumulation of lactate should relate to either an excessive stimulation of glycolysis and/or a limitation of mitochondrial disposal of pyruvate. Given the fact that leg O<sup>2</sup> delivery and leg O<sup>2</sup> uptake are reduced during sprint exercise in hypoxia compared to the same exercise in normoxia (Calbet et al., 2015), the greater accumulation of lactate in hypoxia than in normoxia likely reflects a mitochondrial limitation due to insufficient O<sup>2</sup> supply.

Nevertheless, our data could also indicate that during the sprints in severe acute hypoxia, and perhaps also in normoxia, the stimulation of glycolysis may be excessive in regard with the actual energy demand (Morales-Alamo et al., 2017). This possibility is supported by the fact that during sprint exercise in hypoxia mean power output was not reduced by the ingestion of antioxidants, despite a marked reduction of ATP provision by the glycolysis, that could not be compensated for by increasing

VO2, since during sprint exercise in hypoxia VO<sup>2</sup> was the same in the placebo and the antioxidants conditions. This finding contrasts with the reported reduction of muscle strength and muscle VO<sup>2</sup> after the intravenous administration of vitamin C and tempol in rats (Herspring et al., 2008). Moreover, the fact that mean power output was not reduced after the administration of antioxidants in acute hypoxia is intriguing because the glycolysis provides 50% or more of the overall energy expenditure during the Wingate test in normoxia (Bogdanis et al., 1996; Parolin et al., 1999) and even more during sprint exercise in hypoxia (Morales-Alamo et al., 2012). Therefore, our antioxidant cocktail should have contributed to reduce the energy demand by, for example, improving the P/O ratio and/or by lowering the energy cost of muscle contraction.

Mitochondria produces RNOS mainly at complexes I and III of the electron transport chain (Barja, 1999; Muller et al., 2004). Ascorbate prevents cytochrome c release from the mitochondrial inner membrane and stabilizes the mitochondrial membrane potential in response to hypoxia-reoxygenation (Dhar-Mascareño et al., 2005) and could improve the phosphorylation potential by increasing the proton motive force across the mitochondrial inner membrane (Brand and Nicholls, 2011). The presence of α-lipoic acid in our antioxidant cocktail could have amplified this effect (Cimolai et al., 2014). Moreover, in agreement with an improved mitochondrial efficiency, experiments in humans, have reported enhanced mitochondrial efficiency during exercise in moderate hypoxia (FIO2: 0.13) after nitrate supplementation (Vanhatalo et al., 2014). Antioxidants, like nitrates, increase NO bioavailability (Nyberg et al., 2012; Larsen et al., 2014), which can reduce proton back-leakage across the inner mitochondrial membrane increasing the P/O ratio (Clerc et al., 2007).

Antioxidants may also reduce the energy demand. One of the main contributors to energy expenditure during muscle contraction is the sarcoendoplasmic reticulum calcium pumps (SERCAs), which may account for ∼50% of the overall energy expenditure at rest (Norris et al., 2010) and ∼30–40% during muscle contractions (Barclay et al., 2007). SERCAs-energy expenditure can be lowered by reducing the amount of calcium efflux through the calcium channels [the ryanodine receptor and the inositol 1, 4, 5-trisphosphate receptors (IP3Rs)] and/or by decreasing Ca2<sup>+</sup> leakage from the sarcoendoplasmic reticulum (Chernorudskiy and Zito, 2017). Both processes are increased by RNOS and attenuated by antioxidants (Xu et al., 1998; Kang et al., 2008; Mazurek et al., 2014; Oda et al., 2015). Nevertheless, excessive oxidative stress may also reduce SERCAs activity (Chernorudskiy and Zito, 2017). Calcium transients were likely reduced during the sprints with antioxidants since CaMKII, a kinase that auto-phosphorylates depending on the magnitude of Ca2<sup>+</sup> transients was less phosphorylated after the sprints preceded by the ingestion of antioxidants in our subjects (Morales-Alamo et al., 2013, 2017).

## PDH Re-phosphorylation During the Recovery Period Is Increased by Antioxidants

The recovery of the resting PDH phosphorylation levels requires the activation of PDKs by acetyl-CoA, NADH<sup>+</sup> , or ATP, which re-phosphorylate PDH favoring the oxidation of acetyl groups (Roche and Hiromasa, 2007; Rahimi et al., 2014). Here we show that antioxidants facilitate PDH re-phosphorylation during the recovery period, with faster re-phosphorylation of Ser<sup>293</sup> than Ser<sup>300</sup> (**Figure 4**), as previously shown with experiments in vitro (Yeaman et al., 1978). Cell-culture experiments have shown that antioxidants stimulate PDKs (Churchill et al., 2005). Although not measured here, PDK activity during the recovery period might have been stimulated by the lower NAD+/NADH<sup>+</sup> ratio after the ingestion of antioxidants (**Figure 4**), as previously reported (Morales-Alamo et al., 2013, 2017). Cell-culture experiments have also shown that pyruvate dehydrogenase phosphatase 1 (PDP1) activity is stimulated by lipoic acid (Shan et al., 2014). Nevertheless, an enhanced α-lipoic acid-stimulated PDP1 activity after the ingestion of antioxidants is unlikely inasmuch PDH re-phosphorylation was facilitated during the recovery period after the ingestion of antioxidants. Although PDPs are stimulated by insulin (Thomas and Denton, 1986; Consitt et al., 2016), the insulin response to sprint exercise was not reduced after ingestion of antioxidants in the present study. Thus, the greater PDH re-phosphorylation during the recovery period after antioxidants cannot be attributed to an effect mediated by differences in the plasma insulin concentrations, since our results point toward higher insulin concentrations after the administration of antioxidants, which via PDP activation could delay, but not accelerate, PDH re-phosphorylation.

Another mechanism that could explain a faster rephosphorylation of PDH during the recovery period after the ingestion of antioxidants is a reduction of the exerciseinduced increase in insulin sensitivity by antioxidants (Trewin et al., 2015). Indeed, it has been reported that the increase of insulin sensitivity observed acutely after exercise is augmented in mice lacking the antioxidant enzyme Gpx1, an effect that was abolished by administration of the antioxidant N-acetylcysteine (Loh et al., 2009). This concurs with the lower post-exercise insulin sensitivity observed in humans who received N-acetylcysteine during exercise (Trewin et al., 2015).

As expected from a greater re-phosphorylation of PDH (i.e., inhibition of PDH) following the ingestion of antioxidants, muscle lactate concentration 30 min after the end of the sprints was almost twice as high after the ingestion of antioxidants. This likely reflects a lower rate of lactate oxidation during the recovery after the ingestion of antioxidants. Lactate oxidation must be preceded by the conversion of lactate into pyruvate and subsequently decarboxylated to acetyl-CoA by the PDH, which after the ingestion of antioxidants is less active (Gladden, 2006). Consequently, to maintain the flow of acetyl groups entering the Krebs cycle and the aerobic ATP resynthesis rate during the first 30–60 min after a sprint exercise preceded by the ingestion of antioxidants, fat oxidation should be increased.

### Why Was Performance Not Improved by Antioxidants, Despite the Attenuation of Muscle Lactate Accumulation and Acidification?

Our findings are at odds with the current paradigm of muscle fatigue (Fitts, 1994; Bangsbo and Juel, 2006), but concur with alternative proposals based on human and animal experiments (Nielsen et al., 2001; Lamb and Stephenson, 2006; Morales-Alamo et al., 2015; Torres-Peralta et al., 2016a,b). Muscle fatigue during high-intensity exercise has been traditionally linked to the activation of glycolysis, the accumulation of lactate and the effects caused by the drop in muscle pH (Fitts, 1994). However, in vitro and animal experiments have shown that both intracellular acidification and lactate accumulation contribute to maintain muscle excitability attenuating fatigue (Westerblad et al., 1991; Allen et al., 1995; Pedersen et al., 2004, 2005). Specifically, lactate can inhibit ClC-1 Cl<sup>−</sup> channels and increase the excitability and contractile function of depolarized rat muscles (de Paoli et al., 2010). More recently, the ergogenic effect of lactate accumulation has been confirmed in humans, which recovered partly from fatigue during 1 min of completed bilateral leg occlusion applied at the end of an incremental exercise to exhaustion, despite a 19–22% increase of muscle lactate during the ischemic recovery (Morales-Alamo et al., 2015). In the current investigation, peak power output was not affected by either hypoxia or antioxidants, while mean power output was only reduced by 5.3% in hypoxia (all conditions combined), despite a much greater accumulation of muscle lactate. The fact that muscle lactate could accumulate to higher levels with a minor impact on performance agrees with the idea that lactate and acidification is not as detrimental for muscle contraction as classically thought. However, the potential protective effect on muscle excitability by greater lactate accumulation during exercise in severe acute hypoxia was lost when the sprints were preceded by the intake of antioxidants, which reduced the glycolytic rate to values similar to those observed during the sprints performed in normoxia. This had a minor impact on performance and we think that our findings are compatible with mechanisms of task failure localized outside the active skeletal muscles (Calbet et al., 2015; Morales-Alamo et al., 2015; Torres-Peralta et al., 2016a,b).

In summary, this investigation shows that lactate accumulation during sprint exercise in severe acute hypoxia is not caused by reduced activation of the PDH, which dephosphorylates to similar levels in normoxia and hypoxia. The ingestion of antioxidants before sprint exercise reduces the glycolytic rate in normoxia and hypoxia without a negative impact on performance, implying that either the activation of glycolysis in hypoxia is excessive and/or the energy cost of muscle contraction is reduced by antioxidants. We have also shown that Ser<sup>293</sup> re-phosphorylates a faster rate during the recovery period than Ser300-PDH-E1α, suggesting slightly different regulatory mechanisms that remain to be identified. Finally, the ingestion of antioxidants is associated with increased PDH re-phosphorylation and slower elimination of muscle lactate during the recovery period.

### AUTHOR CONTRIBUTIONS

JAC, DM-A, BG, and CD: Conception and design of the experiments; JAC, DM-A, BG, AS, and CD: Pre-testing, experimental preparation, and data collection; All co-authors:

### REFERENCES


data analysis. The first draft of the manuscript was written by DM-A and JAC. All co-authors edited and proofread the manuscript and approved the final version.

### ACKNOWLEDGMENTS

This study was supported by grants from the Ministerio de Educación y Ciencia (DEP2009-11638, DEP2010-21866, DEP2015-71171-R, and FEDER), FUNCIS (PI/10/07), Programa Innova Canarias 2020 (P.PE03-01-F08), Proyecto Estructurante de la ULPGC: ULPAPD-08/01-4, and Proyecto del Programa Propio de la ULPGC (ULPGC 2009-07 and ULPGC 2015/05), and III Convocatoria de Ayudas a la Investigación Cátedra Real Madrid-UEM (2010/01RM, Universidad Europea de Madrid). Special thanks are given to José Navarro de Tuero for his excellent technical assistance.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Morales-Alamo, Guerra, Santana, Martin-Rincon, Gelabert-Rebato, Dorado and Calbet. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Long-Term Intermittent Work at High Altitude: Right Heart Functional and Morphological Status and Associated Cardiometabolic Factors

Julio Brito<sup>1</sup> \*, Patricia Siques <sup>1</sup> , Rosario López <sup>2</sup> , Raul Romero<sup>1</sup> , Fabiola León-Velarde<sup>3</sup> , Karen Flores <sup>1</sup> , Nicole Lüneburg<sup>4</sup> , Juliane Hannemann<sup>4</sup> and Rainer H. Böger <sup>4</sup>

1 Institute of Health Studies, University Arturo Prat, Iquique, Chile, <sup>2</sup> Department of Preventive Medicine and Public Health, University Autonoma of Madrid, Madrid, Spain, <sup>3</sup> Department of Biological and Physiological Sciences, Facultad de Ciencias y Filosofía/IIA, University Peruana Cayetano Heredia, Lima, Peru, <sup>4</sup> Institute of Clinical Pharmacology and Toxicology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Michael Furian, Universitätsspital Zürich, Switzerland Melissa L. Bates, University of Iowa, United States

> \*Correspondence: Julio Brito jbritor@tie.cl

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 05 December 2017 Accepted: 06 March 2018 Published: 22 March 2018

### Citation:

Brito J, Siques P, López R, Romero R, León-Velarde F, Flores K, Lüneburg N, Hannemann J and Böger RH (2018) Long-Term Intermittent Work at High Altitude: Right Heart Functional and Morphological Status and Associated Cardiometabolic Factors. Front. Physiol. 9:248. doi: 10.3389/fphys.2018.00248 Background:Living at high altitude or with chronic hypoxia implies functional and morphological changes in the right ventricle and pulmonary vasculature with a 10% prevalence of high-altitude pulmonary hypertension (HAPH). The implications of working intermittently (day shifts) at high altitude (hypobaric hypoxia) over the long term are still not well-defined. The aim of this study was to evaluate the right cardiac circuit status along with potentially contributory metabolic variables and distinctive responses after long exposure to the latter condition.

Methods: A cross-sectional study of 120 healthy miners working at an altitude of 4,400–4,800 m for over 5 years in 7-day commuting shifts was designed. Echocardiography was performed on day 2 at sea level. Additionally, biomedical and biochemical variables, Lake Louise scores (LLSs), sleep disturbances and physiological variables were measured at altitude and at sea level.

Results: The population was 41.8 ± 0.7 years old, with an average of 14 ± 0.5 (range 5–29) years spent at altitude. Most subjects still suffered from mild to moderate symptoms of acute mountain sickness (mild was an LLS of 3–5 points, including cephalea; moderate was LLS of 6–10 points) (38.3%) at the end of day 1 of the shift. Echocardiography showed a 23% mean pulmonary artery pressure (mPAP) >25 mmHg, 9% HAPH (≥30 mmHg), 85% mild increase in right ventricle wall thickness (≥5 mm), 64% mild right ventricle dilation, low pulmonary vascular resistance (PVR) and fairly good ventricle performance. Asymmetric dimethylarginine (ADMA) (OR 8.84 (1.18–66.39); p < 0.05) and insulin (OR: 1.11 (1.02–1.20); p < 0.05) were associated with elevated mPAP and were defined as a cut-off. Interestingly, the correspondence analysis identified association patterns of several other variables (metabolic, labor, and biomedical) with higher mPAP.

Conclusions: Working intermittently at high altitude involves a distinctive pattern. The most relevant and novel characteristics are a greater prevalence of elevated mPAP and HAPH than previously reported at chronic intermittent hypobaric hypoxia (CIHH), which is accompanied by subsequent morphological characteristics. These findings are associated with cardiometabolic factors (insulin and ADMA). However, the functional repercussions seem to be minor or negligible. This research contributes to our understanding and surveillance of this unique model of chronic intermittent highaltitude exposure.

Keywords: high-altitude pulmonary hypertension, chronic intermittent hypobaric hypoxia, altitude, right heart, insulin and ADMA

### INTRODUCTION

The right cardiac circuit (rather than the left) of highaltitude populations living in chronic hypobaric hypoxia (CH) undergoes major changes. These changes are characterized by elevated pulmonary artery pressure (PAP), right ventricle hypertrophy, and heart and pulmonary vessel remodeling. Some individuals develop high-altitude pulmonary hypertension (HAPH). Individuals who relocate to live permanently at altitude display the same phenomena (Penaloza and Arias-Stella, 2007). A complex series of pathophysiological and physiological mechanisms are responsible for the responses to hypoxia. The first described mechanism is hypoxic pulmonary vasoconstriction (HPV) (von Euler and Liljestrand, 1946), which is followed by several metabolic and molecular alterations, such as an imbalance between endothelial vasoconstrictors and vasodilators, reactive oxygen species (ROS) (Chen et al., 2012), and some associated factors, including insulin and asymmetric dimethylarginine (ADMA) (Richalet and Pichon, 2014; Lüneburg et al., 2016).

Recently, a new type of exposure to altitude has been of interest: long-term chronic intermittent hypobaric hypoxia (CIHH), which has been acknowledged as a distinct pathophysiological condition, including higher blood pressure at altitude and acute mountain sickness persistence on day 1 (Richalet et al., 2002; Brito et al., 2007). This exposure implies long-term exposure to shifts of 4 to 15 days at an altitude above 3,500 m, which are followed by a resting period of the same number of days at sea level (Richalet et al., 2002; West, 2002). Mining, observatory, army, and frontier control personnel are frequently exposed to these conditions, and their numbers are dramatically increasing. In Chile alone, it has been estimated that there are over 65,000 workers exposed to this condition (Brito et al., 2007). Therefore, many aspects of the underlying molecular mechanisms and clinical consequences are not wellknown. Some studies in humans have demonstrated changes in the right heart circulation that are similar to those occurring in CH, such as a rise in PAP and right ventricular enlargement and/or hypertrophy (Richalet et al., 2002; Sarybaev et al., 2003; Brito et al., 2007). Remodeling of pulmonary vessels has also been demonstrated in animal models (Brito et al., 2015). Notwithstanding, there is scarce research assessing these changes over long durations in larger groups of individuals or looking for new potential associations with other physiological and biochemical variables as associated factors.

The endothelium has been known to play an important role in the regulation of systemic and pulmonary vascular tone, which mainly occurs by secreting the potent vasodilator nitric oxide (NO). NO is synthesized by endothelial NO synthase (NOS), which is competitively inhibited by the endogenous compound asymmetric dimethylarginine ADMA (Böger, 2006). In contrast to ADMA, symmetric dimethylarginine (SDMA) does not directly interfere with NOS activity. Elevated levels of ADMA are a cause of vasoconstriction and high blood pressure and have been associated with a high risk of cardiovascular events and mortality (Zoccali et al., 2001; Böger et al., 2009a,b).

Animal research has already provided some information about molecular or functional interactions in CIHH. Interestingly, NO bioavailability and ROS production have been demonstrated to play a major role in vascular adaptation to altitude hypoxia (Siques et al., 2014b; Lüneburg et al., 2016; Waypa et al., 2016). The morphological, physiological and molecular changes appear to be similar to those occurring in CH, but they may be less pronounced in CIHH (Brito et al., 2008; Siques et al., 2014b; Lüneburg et al., 2016).

Thus, it was expected that altered right heart circuit status and possible associated factors (metabolic, labor, and physiological) would be found in people undergoing long intermittent work at high altitude. Therefore, a cross-sectional study was performed with the aim of determining the morphological and functional status of the right cardiac circuit in miners working intermittently at an altitude between 4,400 and 4,800 m for a period of more than 5 years. Moreover, the study assessed several physiological and metabolic factors to determine whether they were associated with cardiac parameters and whether some of these featured a distinctive response in the evaluated condition.

### MATERIALS AND METHODS

### Subjects and Study Design

A cross-sectional study was performed in a random sample of 120 healthy native Chilean male miners working in a mine settlement in the northern part of Chile at an altitude of 4,400 or 4,800 m (53

**Abbreviations:** BP, Blood pressure; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; SPAP, Systolic pulmonary artery pressure; mPAP, Mean pulmonary artery pressure; RVWT, Right ventricle wall thickness; FCRV, Four-chamber right ventricle; RVOT, Right ventricle outflow track; RAA, Right atrium area; PVR, Pulmonary vascular resistance; LVEF, Left ejection fraction; AD, Aortic diameter; TAPSE, Tricuspid annular plane systolic excursion; RV, Right ventricle; RVH, Right ventricle hypertrophy; ADMA, Asymmetric dimethylarginine; SDMA, Symmetric dimethylarginine; HP, Pulmonary hypertension; HAPH, High-altitude pulmonary hypertension; CIHH, Chronic intermittent hypobaric hypoxia; HPV, Hypoxic pulmonary vasoconstriction; WP, Waist perimeter; SL, Sea level.

and 47% of the study population, respectively) in a shift regimen (7 days at altitude followed by a resting period of 7 days at sea level). The miners slept at 3,800 m and worked a 12-h day shift at highest altitude, with the trip from the dormitories to the pit lasting 30 min. All subjects had undergone a medical examination and laboratory tests to determine altitude fitness. The inclusion criteria were working in shifts (7X7) at high altitude (above 4,000 m) for more than 5 years and a healthy status without serious comorbidities. The exclusion criteria were diabetes, hypertension, diagnosed obstructive sleep apnea, supplementary oxygen in the dormitories and any cardiopulmonary disease.

Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki. The study was approved by the Research Ethics Committee of Universidad Arturo Prat, Iquique, Chile.

### Measured Variables

Measures were taken (1) at altitude (at the mine's health facility) early in the morning 18 h after arrival (after one night's sleep) and (2) at sea level (SL) at an ambulatory medical facility in Iquique, Chile during day 2 of the resting period. An exact definition of the measurement time of each variable is provided below.

### General Data

Age, weight, height, body mass index (BMI), calculated as weight (kg) divided by height squared (m<sup>2</sup> ), waist perimeter (WP), smoking habit, years at altitude, physical activity and medical status were assessed during the basal screening at SL.

### Physiological Parameters

Systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), and hemoglobin oxygen saturation (SaO2) were determined at each study time point (at altitude and SL) in the morning. SBP and DBP were measured in the right arm of each participant while seated and after 5 min of rest using appropriately sized cuffs and calibrated standard mercury sphygmomanometers according to international guidelines (Chobanian et al., 2003; European Society of Hypertension-European Society of Cardiology Guidelines Committee, 2003). Heart rate was measured with an HR-100C Omron device (Omron, Health Care Inc. <sup>R</sup> , Bethesda, Maryland; USA), and SaO2 was determined using a finger pulse oximeter (POX050, Mediaid <sup>R</sup> , Cerritos, CA; USA). The mean of two measurements separated by a 5-min interval was taken as a valid determination of BP, HR, and SaO2. Same measures were obtained at SL for comparison.

### Acute Mountain Sickness (AMS) and Sleep Measurements

The Lake Louise Score-AMS self-assessment test (LLS; Roach et al., 1993), validated in similar Chilean populations (Richalet et al., 2002; Brito et al., 2007), was performed at altitude 18 h after arrival, including the first night's sleep. AMS was diagnosed when headache and at least one other symptom occurred and a LLSs of ≥3 was reached. Severity was assessed according to the following categories: mild (3–4), moderate (5–10), and severe (11–15) (Hackett, 2003). The modified Spiegel questionnaire, validated in similar populations (Richalet et al., 2002; Brito et al., 2007), was recorded to assess sleep status at the same time. Both scores were obtained at SL for comparison.

# Hematological and Biochemical Measurements

Blood samples were taken at SL in the morning after 8 h of fasting through venous puncture without stasis: hematocrit (Htc), hemoglobin (Hb), lipid profile, glycemia, and insulin. Additionally, insulin sensitivity or resistance [homeostatic model assessment (HOMA-IR) index] was calculated by HOMA program V.2.2 (Diabetes trial unit, University of Oxford). The biomarkers asymmetric dimethylarginine (ADMA) and symmetric dimethylarginine (SDMA) were measured using a validated liquid chromatographic–tandem mass spectrometric assay as described previously (Schwedhelm, 2005). The ADMA reference range is ≤0.732 µmol/L, which was derived from a large population-based cohort (Schwedhelm et al., 2009). An SDMA reference range of ≤0.53 µmol/L was later established by the same author (Schwedhelm et al., 2011).

### Echocardiographic Assessment

An echocardiographic assessment was performed by two experienced cardiologists at SL facilities, in the morning after a 1 h rest, using an echocardiograph (GE Vivid-I <sup>R</sup> , GE Healthcare Systems, Tirat Carmel, Israel) with a 1.5–3.6 MHz phased array probe. Left ventricular end diastolic and end systolic measurements and left ventricular septal and posterior wall thicknesses were obtained from parasternal long axis view in Mmode with the ultrasound beam aligned to tips of mitral leaflets. Left ventricular ejection fraction (LVEF) was calculated from the M-mode recordings using Teichholtz formula (Teichholz et al., 1976). Right heart measurements were obtained after aligning the tip of ultrasound beam at true left ventricular apex. Pulmonary ejection time and pulmonary acceleration times were obtained with pulse Doppler recordings of the pulmonary valve from a parasternal short axis view at aortic valve level. Mean pulmonary artery pressure (mPAP) was calculated from pulmonary acceleration time according to Mahan formulas (Dabestani et al., 1987). Tricuspid annular systolic excursion was measured in M-mode from an apical four-chamber view with the ultrasound beam aligned to the lateral aspect of the tricuspid annulus. Right ventricular free wall thickness (RVWT) was obtained from a subcostal four-chamber view. Tricuspid annular plane systolic excursion (TAPSE index) was used for right ventricle (RV) performance (Kaul et al., 1984; López et al., 2012). Pulmonary vascular resistance (PVR) was calculated according to the formula described by Abbas et al. (2003). Similarly, other authors have noted that echocardiography for right heart assessment, including pulmonary hypertension, has shown a high correlation (r 2 ) with invasive right heart catheterization (Kojonazarov et al., 2007; Taleb et al., 2013). All measurements and reference values were acquired according to American Society of Echocardiography Guidelines (Rudski et al., 2010). Two separate cut-off criteria were studied to define pulmonary hypertension (PH): (a) HAPH's consensus (León-Velarde et al., 2005), which is defined at mPAP ≥30 mmHg, and (b) SL PH's

criteria (Rubin and American College of Chest Physicians, 2004), which is defined at mPAP ≥25 mmHg. The reason for including the SL cut-off was two-fold: (1) because the echocardiogram was performed at SL and (2) to have a comparative panorama since the current condition under examination entailed a substantial period of both high-altitude and SL exposure.

### Data Analysis

All data were entered into a database and analyzed using IBM SPSS, V21.0 <sup>R</sup> statistical package (Armonk, NY, USA). For qualitative variables, absolute and relative frequencies were calculated. For quantitative variables, proportions, means, standard deviations, and standard errors were calculated. Normality of the distribution was checked using Kolmogorov– Smirnov test. All variables, except for LLS, were normally distributed. Therefore, a non-parametric Wilcoxon Test was used for LLS. Student's t-test for related or independent samples was used as appropriate. Pearson's chi-square test was used for proportional differences for independent variables. Additionally, Pearson's correlation was performed between quantitative variables. Binary logistic regression models were used to assess the association of all variables measured with the mPAP using two different cut-offs (<30 vs. ≥30 mmHg and <25 vs. ≥25 mmHg). After univariate analyses of all variables, the statistically significant variables were introduced into a multivariable logistic regression model using the forward stepwise method. The results were presented as crude (mutually adjusted) odds ratios (OR) and 95% confidence intervals (CI). The significance level was established at p < 0.05.

To look for association patterns, a multiple correspondence analysis was performed between categorical variables, which might display a more comprehensive panorama. Biomedical, occupational and metabolic variables were chosen and dichotomized at their normal values. The graph and variances of each dimension are provided. Angles <60◦ are considered as association and the longer distance of the variable from its origin the better represented. A Cronbach's alpha of >0.50 was considered appropriate.

### RESULTS

### General Characteristics

The study group was 120 miners with a mean age of 41.8 ± 0.7 years and a mean exposure to CIHH of 14 ± 0.5 years. Most study participants were overweight (BMI: 26.3 ± 0.3 kg/m<sup>2</sup> ) and sedentary (<3 weekly <30 min moderate exercise sessions, according to Chilean criteria) (MINSAL Ministerio de Salud, 2006), and a third of them were current smokers. **Table 1** gives a complete overview of the demographic and anthropometric characteristics of the study group.

### Physiological Parameters

As expected, both SBP and DBP were higher at altitude than at SL, along with a slight increase in HR. Approximately 40% of the subjects had elevated SBP at altitude (≥130 mmHg, p < 0.01), and 18% had elevated DBP (≥90 mmHg, p < 0.05). SaO2 differed between SL and altitude in proportion to the level of altitude.



Values for quantitative variables are means (X) ± standard error (SE) and ranges; values for qualitative variables are proportions (%).

### TABLE 2 | Physiological parameters.


Systolic blood pressure (SBP; mmHg), diastolic blood pressure (DBP; mmHg), heart rate (HR; b/m), hemoglobin oxygen saturation (SaO2; %), Lake Louise (score) and Spiegel test (score), at sea level (SL) and at altitude. Values are means (X)±SE (standard error), and p was obtained from t-test of related samples.

However, 30% of the study population had SaO2 levels ≤88% at altitude (**Table 2**).

### AMS and Sleep Disturbances

Despite long exposure to high altitude, a rather high proportion of individuals showed AMS (38.4%) on the first day, which was mainly moderate, and remarkably many study participants had cephalea (47.5%). Most individuals declared that they had mainly regular non-satisfactory sleep (75%), whereas the sleep disturbances and AMS measured at SL were minimal (20% and

TABLE 3 | Hematological and biochemical measurements at sea level.


Values are means (X) ±standard error (SE) and reference range.

12.5%, respectively). The presence of AMS was significantly associated with mPAP ≥25 and mPAP ≥30 mmHg (p < 0.05).

### Hematocrit and Hemoglobin

Mean Htc and Hb were almost within the normal range (70% below 17 mg/dL). Only two individuals had Hb values of 19 mg/dL, and none had Hb above 21 mg/dL (**Table 3**).

### Lipid Profile

In 48% of the individuals, triglycerides were elevated above 150 mg/dL, and mean triglycerides and VLDL-cholesterol was also elevated. Although mean total cholesterol was within the normal range, 42% of the study participants had values over 200 mg/dL. Mean HDL-cholesterol and LDL-cholesterol were within the respective normal ranges. The Castelli index (total cholesterol and HDL-cholesterol) was within its upper limit, and HDL/LDL index was normal (**Table 3**), although 20% of subjects had low HDL, and 30% had elevated LDL.

### Glycemia, Insulin, and HOMA-IR

Mean glycemia and insulin values were within normal ranges. The mean HOMA-IR index was also normal (**Table 3**). However, for subjects with insulin above 20 IU (11%), the HOMA-IR index was 3.3 ± 0.2. Insulin was found to significantly correlate (p < 0.01) with BMI (r = 0.34); WP (r = 0.39), SBP at altitude (r = 0.25), VLDL and triglycerides (TG) (both r = 0.25) and mPAP (r = 0.22; p < 0.05).

Upon further analysis of the association of mean insulin levels with mPAP values, a distinctive difference in insulin concentrations at each cut-off point of mPAP was found. At mPAP ≥30 mmHg, mean insulin was 16.9 ± 2.14 IU, whereas at mPAP <30 mmHg, a lower mean insulin value was observed (10.6 ± 0.32 IU; p <0.01, **Figure 1A**). Those with high insulin levels had a higher HOMA-IR (**Figure 1B**). Moreover, this association was further corroborated by the correlation of HOMA-IR with higher mPAP (r = 0.24, p <0.001).

### ADMA

Mean ADMA concentration was slightly elevated (0.83 ± 0.2 µmol/L) compared to the reference range (**Table 3**). Half of the subjects had ADMA concentrations ≥0.80 µmol/L. ADMA was positively correlated (p < 0.05) with WP (r = 0.21), Hb (r = 0.20), and SDMA (r = 0.82; p < 0.001). Most importantly, ADMA was correlated with mPAP (r = 0.21; p < 0.05) and RVWT (r = 0.30; p <0.001). Distinctly, mean ADMA concentration was 1.01 ± 0.15 µmol/L in subjects with mPAP ≥30 mmHg, as opposed to 0.81 ± 0.18 µmol/L in subjects with mPAP <30 mmHg (p < 0.001; **Figure 2**).

### Echocardiographic Findings (mPAP and Morphological Status)

Most individuals had mPAP within normal ranges (73.9%) with wide variability. Nevertheless, 26.1% of the subjects had elevated values at a cut-off point of 25 mmHg, but only 9.2% could be categorized as truly HAPH (≥30 mmHg), and none exceeded an mPAP of 36 mmHg (**Table 4**).

Regarding morphological status, it was noted that 85% of the individuals had mild right ventricle hypertrophy (RVH), and over half showed a grade of right ventricular (RV) dilation supported by an increase in FCVR and right ventricle outflow track (RVOT) values. However, a minimal percentage (15%) of right atrium enlargement is seen (**Table 4**). A representative figure of RVH is shown in **Figure 3A**, and pulmonary acceleration curve at outflow tract in **Figure 3B**.

Most subjects display PVR within the normal range, except for 4.8% of the individuals, who have slightly increased PVR. Despite the latter finding, the subjects with mPAP ≥30 mmHg had a value of 1.33 Wood units vs. 1.04 Wood units with mPAP <30 mmHg (p <0.001). The RV performance index (TAPSE) shows no impairment and good performance in this group. For the left ventricle, the left ventricle ejection fraction is within the normal range and the aortic diameter is only mildly enlarged in 22.6% of the individuals (**Table 4**).

Some significantly positive correlations (p < 0.01) were found in all correlations performed between echocardiographic variables: (a) mPAP vs. RVWT r = 0.39, RVOT r = 0.35, and PVR r = 0.40; (b) RVWT vs. RVOT r = 0.30 and PVR r = 0.22; and (c) right atrium area (RAA) vs. four-chamber right ventricle (FCRV) r = 0.44 and PVR r = 0.21.

However, only two associations were shown when both cutoff values for mPAP (25 and 30 mmHg) were used in the univariate logistic regression and in the forward stepwise logistic multivariate regression final model. At the cut-off value of 25 mmHg, ADMA (OR 8.84; CI 1.18–66.39, p < 0.05) and insulin (OR: 1.07, CI 1.01–1.13, p < 0.05) showed an association. At the cut-off value of 30 mmHg, the same associations were observed: ADMA (OR: 10.74, CI 1.16–99.9, p < 0.05) and insulin (OR: 1.11, CI 1.02–1.20, p < 0.05). All OR were adjusted by smoking status, age, and BMI.

### Multiple Correspondence Analysis

Because only two associations were found, a multiple correspondence analysis was performed. This statistical tool allows visualize association patterns or profiles between

different dichotomized variables to be determined: occupational and biomedical (years old and years at altitude), metabolic (ADMA, insulin, triglycerides, BMI, and waist perimeter) and echocardiographic (right ventricle wall thickness and mPAP) were introduced. To interpret the graphical representation, category associations can be detected by their proximity whose reference is the angle formed with the coordinates origin (small angles between two categories indicates greater association) and the relative distance to the origin.

Therefore, on one hand, two different profiles are shown. In one profile, metabolic variables were within normal values, and age < 40 and years at altitude <15 were associated with mPAP < 25. In the other profile, altered metabolic values, higher age, and over 15 years at altitude were found to be associated with mPAP ≥ 25. On the other hand, that mPAP > 25 and > 15HAyears are strongly associated (small angle), mPAP >25 and insulin > 20 are moderately associated (medium angle) and mPAP > 25 and age < 40 (180 degree angle) suggest opposite directions (no association) are depicted (**Figure 4**). Interestingly, when the same procedure was performed with mPAP ≥ 30, the results were similar.

# DISCUSSION

This cross-sectional study, in long-term CIHH working shifts at high altitude, has the following main findings: (1) a distinctive and unique pattern of physiological responses was determined, wherein a third of subjects showed moderate AMS persistence; (2) high proportions of elevated mPAP (26.1%) and HAPH (9.2%) were found by echocardiography; and (3) specific cardiometabolic variables (high triglycerides, insulin resistance, high ADMA, increased waist perimeter and BMI) appeared to be associated factors, with insulin and ADMA clearly associated with elevated mPAP.

### General Aspects

This cross-sectional study with a large number of subjects working intermittently in a long-term CIHH at high altitude contributes three important concepts (mentioned above), which will be individually discussed for methodological reasons; however, these concepts appear to be related and intertwined as a whole.


TABLE 4 | Echocardiographic findings at sea level.

Systolic pulmonary artery pressure (SPAP), mean pulmonary artery pressure by Mahan (mPAP, right ventricle wall thickness (RVWT, four-chamber right ventricle (FCRV), right ventricle outflow track (RVOT), right atrium area (RAA), pulmonary vascular resistance (PVR), left ventricle ejection fraction (LVEF), aortic diameter (AD) and tricuspid annular plane systolic excursion (TAPSE). Values are means (X) ± standard error (SE), range, sea level (SL) cut-off values (reference: American Society of Echocardiography; Rudski et al., 2010) and proportions out of reference value (%).

The first concept is that this population seems to share distinctive and unique features. In fact, there is a remarkable proportion of overweight and sedentarism, as has been previously reported (Esenamanova et al., 2014). Smoking habit proportion is similar to Chilean prevalence. Physiological responses are also coincident with previous reports in that BP rises at altitude but reaches normal values at SL. Moreover, it has been described that BP decreases after day 2 during the shift at high altitude, but it does not reach SL values (Richalet et al., 2002; Brito et al., 2007); nonetheless, a certain proportion of individuals maintain an SBP and DBP within a rather elevated range at altitude according to previous reports (Siques et al., 2009). Conversely, SaO2 drops at altitude and is fully recovered at SL, and the mean SaO2 reached at altitude is similar to acclimatized subjects; however, almost one-third of the subjects have lower SaO2 values, which could be explained by individual variability or a poor response to altitude (day 1's impact) (Brito et al., 2007). Additionally, coincident with previous reports in humans, Htc and Hb do not show pathological values or excessive erythrocytosis (Richalet et al., 2002; Brito et al., 2007; Siques et al., 2007). The latter adds support to the suggestion that under this regimen, high or excessive erythrocytosis is rare. Likewise, consistently high AMS presence (mostly moderate), cephalea and sleep disturbances despite the extensive elapsed time are additional distinctive features. These findings are also coincident with previous studies in which AMS was present during the first 1 or 2 days of a shift at high altitude (Richalet et al., 2002; Brito et al., 2007). However, some cohort studies have described a decline during following exposures (Richalet et al., 2002; Wu et al., 2009). The reasons are beyond the scope of this study, but may be due to an inability to acclimatize properly, the loss of acclimatization during the shift at SL or rapid ascent.

### Metabolic Responses

### Lipid Profile: TG and Total Cholesterol (T-Chol)

Another relevant result of this study are the changes in lipid profiles. It has been assumed by several authors that under CH or CIHH, individuals are prone to good metabolic features (Anderson and Honigman, 2011; Ezzati et al., 2012). However, recent evidence has shown that there is an increasing proportion of individuals with marked metabolic alteration (Mohanna et al., 2006), which would be related to chronic mountain sickness (Miele et al., 2016) and other altitude diseases (San Martin et al., 2017). In fact, our results support this observation; in this model of exposure, a rather high proportion (almost half of the subjects) displayed an altered lipid profile. Previous studies have consistently reported an increase in TG and contradictory results for T-Chol (Li et al., 2007; Siques et al., 2007). The most remarkable changes seen in this study are an increase in TG, VLDL and T-Chol, but the mean Castelli index remains at its upper limit.

Hypoxemia may disturb lipid metabolism by upregulating hepatic SCD-1, leading to de novo TG synthesis, an increase of adipose tissue lipolysis, lipoprotein secretion and decrease of lipoprotein clearance (Li et al., 2007; Drager et al., 2012; Siques et al., 2014a). Thus, it could be surmised that TG alteration could be another distinctive feature of exposure to CIHH and may be a matter of concern. Recognition of TG alteration as a cardiovascular risk factor is a growing issue (Assmann et al., 1996).

FIGURE 3 | Representative echocardiographic images (A) right ventricular hypertrophy and (B) Acceleration curve of pulmonary flow at pulmonary artery outflow tract.

# Insulin, Insulin Resistance (IR)

Complementary to the above metabolic findings, insulin and IR have been noted as tightly related to the development of PH and could be considered associated factors (Zamanian et al., 2009). Their relation to HAPH is still to be demonstrated. However, the finding that the insulin level correlated to anthropometric variables (BMI and WP) and to mPAP—and this population is mostly overweight—might suggest a similar role in HAPH as has been reported either for IR (McLaughlin and Rich, 2004) and/or PH (Taraseviciute and Voelkel, 2006) at SL. The findings of a cut-off point of different insulin values according to mPAP and that higher insulin values are also associated with higher mPAP further support insulin's role. IR may also share other pathophysiological conditions that are present under hypoxia, such as elevation of cytokines (interleukin-6, monocyte chemotactic protein) and ADMA (Zamanian et al., 2009). Recently, an inverse relationship between oxygen hemoglobin saturation and IR has been described in chronic hypoxia. There are many physiopathological explanations (Miele et al., 2016), including the disruption of leptin pathways (Polotsky et al., 2003).

### ADMA as a Marker of Cardiometabolic Risk

ADMA is a competitive NOS inhibitor that has been identified as a regulator of NO production (Böger, 2006). ADMA is endogenously present in the human body and inhibits NO-dependent vasodilation in vitro (Vallance et al., 1992) and in vivo (Böger et al., 1998). ADMA is degraded by dimethylarginine dimethylaminohydrolase (DDAH). Disruption of the ADMA/DDAH pathway causes endothelial dysfunction and elevated blood pressure in the systemic and pulmonary circulation (Leiper et al., 2007). Therefore, the possible role of ADMA in HAPH is a focus of current research. The current results show a significant elevation of ADMA at altitude, and for the first time, they show a correlation of ADMA with both mPAP and right ventricle wall thickness, suggesting a pivotal pathophysiological role of this pathway in the development of pulmonary vascular dysfunction at high altitude. Moreover, a clear cut-off value of ADMA was observed according to mPAP.

Recently published data support the current findings. In rats subjected to long-term CIHH, an impaired NO pathway secondary to elevated ADMA and increased ROS were found (Siques et al., 2014b; Lüneburg et al., 2016). Furthermore, young healthy adults first exposed to CIHH develop elevated ADMA concentrations over the time (Lüneburg et al., 2017). Therefore, ADMA may be a useful biomarker for subjects with HAPH.

# PAP and Right Heart Status

An increase in PAP is a well-known physiological response to hypoxia (von Euler and Liljestrand, 1946). Additionally, permanent residents have changes in their pulmonary vascular circuit that might account for a prevalence up to 10% of HAPH (León-Velarde et al., 2005; Penaloza and Arias-Stella, 2007). Unfortunately, most information comes from acute or chronic exposure and rarely from long-term CIHH; therefore, the described changes might not be entirely applicable to CIHH. Research conducted under different commuting exposure regimes in humans (2-year cohort, 30 × 30 shift; 2-year cohort, 7 × 7 shift) and in a 12-year cross-sectional study (5 × 2 shift) at altitudes between 3,550 and 4,400 m) also described an increase in mPAP, right ventricle enlargement or ventricle hypertrophy, although a small number of subjects was included (Richalet et al., 2002; Sarybaev et al., 2003; Brito et al., 2007). The latter author determined a 4% prevalence of HAPH in CIHH at 3,550 m. Interestingly, experimental studies in rats led to the suggestion that RV changes were achieved to a lesser extent in CIHH than in CH (Corno et al., 2004; Brito et al., 2008).

Despite several limitations (cross-sectional study, echocardiogram performed at SL, only clinical functional capacity assessment and previously selected subjects for highaltitude work) of the current study, the results showed a strikingly high proportion of mPAP over 25 mmHg with a proportion of HAPH greater than previously reported that is similar to the prevalence in CH. Likewise, it could well be inferred that subsequent pulmonary vasculature remodeling triggered by HPV was present, as reported in CH (Penaloza and Arias-Stella, 2007; Sylvester et al., 2012) and reproduced in rats under long-term CIHH (Brito et al., 2015). Similarly, in light of the results and despite the concept of "turn on-turn off " biological responses in this condition (Powell and Garcia, 2000), it seems reasonable to infer that in the current model, individuals undertake a more prolonged pulmonary hypertensive state than expected in this condition at high altitude.

Moreover, the HAPH levels found in this study could be considered mild, which is supported by the fact that no subjects had mPAP values over 36 mmHg and that the subjects had a current healthy status without claims of functional capacity impairment. The lack of functional repercussions (Penaloza et al., 1963) and a mild or moderate HAPH have been described for CH and acute hypoxia (Naeije and Dedobbeleer, 2013); this phenomenon has been noted as the paradox of HAPH (Grover, 2014). Nevertheless, some disputes still exist regarding exercise capacity at high altitude and HAPH. In fact, a recent study of Kyrgyz highlanders with HAPH found a mild reduction in exercise performance and reduced quality of life (Latshang et al., 2017). Accordingly, these rather contradictory findings have led to introduction of the concept of "pulmonary vascular reserve" as a complex mechanism that determines good or poor exercise capacity (Groepenhoff et al., 2012; Naeije and Dedobbeleer, 2013; Pavelescu et al., 2013).

Likewise, this study also shows a striking proportion of RVWT enlargement, suggesting that in the long term, almost all subjects experience RV remodeling and that an mPAP value of 25 mmHg would be sufficient to generate significant changes in the right heart. Whether this is merely an acclimatization response in this model of exposure that causes the RV to increase its performance as a consequence of its homeometric adaptation to afterload increase (Kolár and Ostádal, 1991; Naeije and Dedobbeleer, 2013; Richalet and Pichon, 2014) and/or ROS activation by hypoxia of AMP kinase proteins (Chen et al., 2012; Waypa et al., 2016) or other mechanisms remains to be elucidated. Additionally, a vascular hyperdynamic state triggered by adrenergic activation and autonomic system imbalance must be considered for exerting its influence on the above variables and possibly in RV dilation (Richalet and Pichon, 2014). Conversely, right atrium morphological changes account for a very low proportion. Complementary, several weak correlations between RV morphological parameters showed linearity with RV afterload (mPAP value). These findings support the consistency of the measurements, and they are in line with the PH evolving process (ACCF/AHA, 2009).

Further analysis of the PVR results is limited since the indirect method of measurement could bias their interpretation and is also controversial. In fact, although some studies mentioned above support a high correlation with right heart catheterization, some others have described this method as accurate but with only moderate precision (Naeije, 2003; Rudski et al., 2010; Naeije and Dedobbeleer, 2013). If these results were accurate, it would support a more hyperdynamic state and milder surmised vascular remodeling, most likely as the result of greater NO bioavailability, as shown in CIHH rats (Siques et al., 2014b; Brito et al., 2015).

Hence, the above considerations would better explain the mild HAPH with apparently almost no functional capacity repercussion. In fact, the TAPSE index of ventricle performance is good. The left ventricle does not seem to be affected, except for a low proportion of aortic dilation, which could also be explained by the hyperdynamic state of this model. The latter findings agree with a previous report regarding left ventricle changes under hypoxia (Richalet and Pichon, 2014).

Moreover, as mentioned, the correspondence analysis using both mPAP values depicted two distinctive patterns of associations with metabolic, biomedical and occupational factors. This tool, albeit is a descriptive method, has allowed to demonstrate the association of elevated mPAP with others variables not found in the regression model. In fact, >15 years of working in this model is strongly associated to elevated mPAP, and additionally, the association of insulin >20 with elevated mPAP further and consistently support what was previously found. The findings obtained with this correspondence analysis are of utmost interest, plausible and depict a sort of signature of this exposure model; however, its role in understanding the pathophysiological process of HAPH remains to be determined. Because this analysis only allows association patterns to be determined, its potential usefulness for screening or prediction will require further studies.

Therefore, the current study highlights a rather novel finding, as discussed previously, in which an association of insulin and ADMA with high mPAP and HAPH was found. Also, the years spent at altitude in this CIHH model might be considered. It is worth noting that some of the abovementioned associated variables to elevated mPAP, have been previously found to individually contribute to PH and HAPH (Parameswaran et al., 2006; Wu et al., 2009; Zamanian et al., 2009; Kane et al., 2016; San Martin et al., 2017). Until now, their contribution as a whole to right cardiac circuit parameter values in subjects working in long-term CIHH had not been clearly determined.

Although this study design has some limitations, its use was considered necessary to evaluate right cardiac status during long-term intermittent work at high altitude. First, a crosssectional design allows the determination of only association, status and the absence of changes, but no predictive factors or cause and effect. Moreover, a prospective study might be very difficult.

However, a cross-sectional study has the advantages of the ability to measure several variables, lower cost, and the identification of important points to be studied in a future longitudinal study. Second, the echocardiogram was performed at SL on day 2 of the resting period for logistic reasons, which could produce pulmonary pressures lower than at high altitude. Nevertheless, this study allowed important findings in this condition to be determined. As a whole, while the characteristics found in these subjects could eventually be assumed to result from the study design, the rigorous methodology used, the strict selection and exclusion criteria, the use of more accurate technology and the literature provided support the reasonable validity of the findings of this study in this population. Additionally, altitude differences may explain the elevated proportions of altered morphological and functional right cardiac parameters found in this study compared with those in a previous cross-sectional study (Brito et al., 2007). In fact, the latter study was conducted at 3,550 m, while the subjects in the current study worked at over 4,400 m.

# CONCLUSIONS

In summary, this study contributes to the knowledge of longterm CIHH working conditions at altitude, a rather unique biological situation of hypoxia exposure. Thus, it corroborates the persistence of AMS and lack of excessive erythrocytosis. However, this study highlights some novel findings: a high prevalence of HAPH, which is similar to that reported in CH, and with higher numbers of subjects with elevated mPAP and RVWT. In addition to determining the right circuit morphological and functional status, this study is the first, to our knowledge, to identify an association between increased cardiometabolic variables and others with elevated mPAP in CIHH. Therefore, these findings are likely to have important implications for defining the epidemiological and biological features of this model or prompting actions for public health.

# AUTHOR CONTRIBUTIONS

JB, PS, and RL: conceived of and designed the study, analyzed and interpreted the data, drafted the manuscript, critically revised important intellectual content in the manuscript, and provided overall supervision; RR, KF, NL, JH, and RB: performed data acquisition, analysis, or interpretation and critically revised important intellectual content in the manuscript; FL-V: assisted with critical decisions and revisions; JB, PS, RL, RR, FL-V, KF, NL, JH, and RB: contributed to the interpretation of results and critical revision of the manuscript, approved the final manuscript and agreed to be accountable for all aspects of the work.

### REFERENCES


### FUNDING

This work was supported by grants from AECID # A/030027/10 and FIC-TARAPACA BIP 30477541-0 and from the German Federal Ministry of Research 01DN17046 DECIPHER.

### ACKNOWLEDGMENTS

We would like to thank Jacobo Alarcon (RN) and Gabriela Lamas for their technical assistance.


Powell, F. L., and Garcia, N. (2000). Physiological effects of intermittent hypoxia. High Alt. Med. Biol. 1, 125–136. doi: 10.1089/15270290050074279


Alberta, Canada, eds J. R. Sutton, C. S. Houston, and G. Coates (Burlington, VT: Queen City Printers), 272–274.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Brito, Siques, López, Romero, León-Velarde, Flores, Lüneburg, Hannemann and Böger. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# The Effect of Oxygen Enrichment on Cardiorespiratory and Neuropsychological Responses in Workers With Chronic Intermittent Exposure to High Altitude (ALMA, 5,050 m)

Fernando A. Moraga<sup>1</sup> \*, Iván López <sup>2</sup> , Alicia Morales <sup>3</sup> , Daniel Soza<sup>2</sup> and Jessica Noack <sup>3</sup>

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Frank L. Powell, University of California, San Diego, United States Julio Alcayaga, Universidad de Chile, Chile

> \*Correspondence: Fernando A. Moraga fmoraga@ucn.cl

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 16 November 2017 Accepted: 23 February 2018 Published: 23 March 2018

### Citation:

Moraga FA, López I, Morales A, Soza D and Noack J (2018) The Effect of Oxygen Enrichment on Cardiorespiratory and Neuropsychological Responses in Workers With Chronic Intermittent Exposure to High Altitude (ALMA, 5,050 m). Front. Physiol. 9:187. doi: 10.3389/fphys.2018.00187 <sup>1</sup> Laboratorio de Fisiología, Hipoxia y Función Vascular, Departamento de Ciencias Biomédicas, Facultad de Medicina, Universidad Católica del Norte, Coquimbo, Chile, <sup>2</sup> Safety Group, Atacama Large Millimeter Submillimeter Array, Calama, Chile, <sup>3</sup> Departamento de Ciencias de la Salud, Escuela de Enfermería, Universidad Santo Tomás, Santiago, Chile

It is estimated that labor activity at high altitudes in Chile will increase from 60,000 to 120,000 workers by the year 2020. Oxygenation of spaces improves the quality of life for workers at high geographic altitudes (<5,000 m). The aim of this study was to determine the effect of a mobile oxygen module system on cardiorespiratory and neuropsychological performance in a population of workers from Atacama Large Millimeter/submillimeter Array (ALMA, 5,050 m) radiotelescope in the Chajnantor Valley, Chile. We evaluated pulse oximetry, systolic and diastolic arterial pressure (SAP/DAP), and performed neuropsychological tests (Mini-Mental State examination, Rey-Osterrieth Complex Figure test) at environmental oxygen conditions (5,050 m), and subsequently in a mobile oxygenation module that increases the fraction of oxygen in order to mimic the higher oxygen partial pressure of lower altitudes (2,900 m). The use of module oxygenation at an altitude of 5,050 m, simulating an altitude of 2,900 m, increased oxygen saturation from 84 ± 0.8 to 91 ± 0.8% (p < 0.00001), decreased heart rate from 90 ± 8 to 77 ± 12 bpm (p < 0.01) and DAP from 96 ± 3 to 87 ± 5 mmHg (p < 0.01). In addition, mental cognitive state of workers (Mini-Mental State Examination) shown an increased from 19 to 31 points (p < 0.02). Furthermore, the Rey-Osterrieth Complex Figure test (memory) shown a significant increase from 35 to 70 (p < 0.0001). The results demonstrate that the use of an oxygen module system at 5,050 m, simulating an altitude equivalent to 2,900 m, by increasing FiO<sup>2</sup> at 28%, significantly improves cardiorespiratory response and enhances neuropsychological performance in workers exposed to an altitude of 5,050 m.

Keywords: oxygen enrichment, neuropsychological impairment, oxygen saturation, heart rate, arterial pressure, chronic intermittent hypobaric hypoxia

# INTRODUCTION

Exposure to high altitude has become an increasingly common event. As many as 140 million people live at altitudes over 2,500 m (Moore, 2001). However, mining activities at altitudes over 3,500 m are common in in the north of Chile and Peru. In Chile up until 1995, there were approximately 20,000 workers intermittently exposed to high altitudes for a long period of time (Jiménez, 1995), and this number is expected to increase to over 120,000 by the year 2020 (http://www.ccm.cl/wp-content/ uploads/2016/06/fuerza\_laboral\_de\_la\_gran\_mineria\_chilena\_ 2012\_2020.pdf). In these conditions, chronic intermittent hypobaric hypoxia (CIHH) constitutes a model of hypobaric hypoxic exposure previously described by several authors (Jiménez, 1995; Richalet et al., 2002; Moraga et al., 2014).

The physiological consequences of the initial exposure to high altitude are widely known. For example, the incidence of acute mountain sickness depends on altitude reached, previous experience, velocity of ascent and is self-limited by exposure time (Davis and Hackett, 2017). Early studies showed the effect of acute exposure on neuropsychological performance. For instance, the reduction in oxygen at high altitude impairs mental and physical performance, and general well-being (Barcroft et al., 1923). In addition, studies performed in a population of miners at high altitude showed deterioration in cognition and motor function (Mc Farland, 1937). Exposure to high altitude resulted in impaired sleep quality with increased periodic breathing, increased awakenings and shorter stage 3 and 4 sleep, causing low productivity and altered general well-being (Gerard et al., 2000). All these effects were dependent on the altitude attained.

One way to avoid the consequences of high altitude hypoxemia is to reduce the equivalent altitude (altitude which provides the same PO<sup>2</sup> in moist inspired gas during ambient breathing). The first physiological response is given by an increase in ventilation and second by an artificial increase in oxygen supply. Therefore, we evaluated two different procedures that provide supplemental oxygen at high altitudes: an oxygen concentrator (West, 2016) and the use of liquid oxygen (Moraga et al., 2014). Both procedures mimic the higher oxygen partial pressure of a lower altitude by increasing in the oxygen fraction. The beneficial effects of this procedure enhancing sleep quality, neuropsychological function, and arterial oxygenation has been demonstrated in subjects exposed to conditions of simulated hypoxia at sea level, and subjects that have been acutely exposed to altitude hypoxia (West, 1995; Luks et al., 1998; Gerard et al., 2000; McElroy et al., 2000). Currently the only study performed during exposure to high altitude at 4,200 m, in miners acclimatized to CIHH, evaluated sleep quality and the nocturnal ventilation pattern, finding that they were significantly enhanced by oxygen administration (Moraga et al., 2014). However, no studies have evaluated a CIHH population at altitudes over 4,200 m. In addition, the ministry of health provided a technical guide for high altitude workers in 2014 that included a series of recommendations to reduce malaise during exposure to altitudes over 3,000 m. In regards to environmental oxygenation, the oxygen equivalent must be below 3,000 m with permanent control of temperature, relative humidity and room ventilation (web.minsal.cl/sites/default/files/ guia\_hipobaria\_altitud.pdf). Therefore, our aim was probe a mobile module that reduced the equivalent altitude to values of 2,900 m, in an isolated sector at a very high altitude of 5,050 m in a population of workers from the Atacama Large Millimeter/ submillimeter Array (ALMA) radiotelescope in the Chajnantor valley, (Chile) and evaluate cardiorespiratory and neuropsychological on workers in both conditions.

# SUBJECTS AND METHODS

### Subjects

Thirteen voluntary subjects that lived at a low altitude (<1,000 m) and worked at ALMA (2,900 and 5,050 m) were studied. All subjects worked as antenna operators and had experience with CIHH for more than 4 years. Their shift pattern was 8 days of work at high altitude followed by 6 days rest at sea level. All subjects were free of cardiovascular, pulmonary, hematological, renal or hepatic diseases. All protocols and procedures performed in the present study were performed by following the Helsinki guidelines and have previously been approved by the Ethics Committee of the Facultad de Medicina of Universidad Católica del Norte and the ALMA Safety Department. A consent informed was obtained in written form each one of them previous being monitored.

### Facilities

ALMA is located 30 Km from the town of San Pedro de Atacama. ALMA has two sectors: the first is located 16 kilometers from the town of San Pedro de Atacama and represents the base camp or Operations Support Facility (OSF) at 2,900 m; and the second sector is the Array Operation Site (AOS) where the antennas are located in the Chajnantor Valley. This valley is located 40 km from OSF at an altitude of 5,050 m. The full personnel capacity of ALMA (OSF and AOS) is close to 400 people, including scientists, operation specialists, and services that operate in chronic intermittent hypobaric hypoxia exposure.

### Equipment

This study was performed at the AOS at 5,050 m where we designed a comfortable and mobile module (OxymindTM, INDURA, Chile) with facilities to increase the oxygen concentration (1 to 10%) and enhance the relative humidity and temperature (**Figure 1**). The oxygen concentration of the room air was controlled by a wide range oxygen sensor (0–50%, accuracy ± 2 ppm; Ultima XA, MSA, Pittsburg, USA) and maintained by a precise servo control system that allowed for an increase in oxygen concentration in 7% to reach 28 ± 0.5%, representing an equivalent altitude of 2,900 m with a lower low fire risk (West, 2001, 2003). In addition, precise control of CO2, by using an infrared CO<sup>2</sup> sensor (ppm or % volume) (model Eagle 2, RKI Instruments Inc., Buffalo Grove, USA), was obtained by activating a fan (near 1400 m<sup>3</sup> /h) in order to reduce the accumulation of CO<sup>2</sup> in the room (with a maintenance level of CO<sup>2</sup> < 0.1%). Therefore, ventilation of the room was maintained by following the procedure described by West (1995) and Luks et al. (1998).

### Study Protocol

Cardiorespiratory responses and neuropsychological evaluations were performed in acclimatized workers 48 h after arrival to high altitude. All parameters were measured in two conditions: base camp (OSF, 2,900 m) and AOF (5,050 m) where voluntaries were evaluated breathing ambient air and breathing room air enriched with oxygen (28 ± 0.5%) at 5,050 m in order to reduce equivalent altitude to 2,900 m. The total exposure time in each condition was 20 min (20 min for cardiovascular and neuropsychological evaluations).

### Cardiorespiratory Evaluations

Heart rate and oxygen saturation were continuously recorded in the subjects during the procedures with a pulse oximeter (Wristox 3100, Nonin, Minnesota, USA), all recordings were analyzed by nVision software (Nonin, Minnesota, USA). SpO<sup>2</sup> and heart rate values were extracted and analyzed min-min. Systolic arterial pressure (SAP), diastolic arterial pressure (DAP) and mean arterial pressure (MAP) were taken during the neuropsychological evaluation (model BM3, Bionet).

### Neuropsychological Evaluation

In order to evaluate the effect of oxygen supply on cognitive function, a neuropsychological test was given to each volunteer. This test covered cognitive functions for memory and constructive praxis (Rey-Osterrieth Complex Figure test and Mini-Mental State Examination, MMSE).

# Rey-Osterrieth Complex Figure Test

In our study, we use one of the most widely used test in both clinical and experimental settings to evaluate visuoconstructional abilities and nonverbal memory (Frank and Landeira-Fernandez, 2008). Subjects were given a blank paper to draw the figure best. They don't have of time for the copy or process to memory recall the draw. Afterward, the subjects were evaluated cardiorespiratory variables and MMSE, the time frame for this evaluation was approximately 10 min, and without warning each subject was given another blank sheet to recall the figure. The total time of evaluation, is don't was major than 15–20 min, in order to reduce the effects of fatigue produced by prolonged neurocognitive tests. Furthermore, time over 30 min the performance decay of this time (Loring et al., 1990). Each drawing was evaluated using a traditional scoring unit for the complex figure of 18 particular items of the complex figure (Loring et al., 1988), where each particular item was evaluated in accord to a two-point, with a total score of 36 points (**Table 1**). Considering that in our country we don't have a cut off values. However, empirical experience is accepted to copy values <27 points and memory values <17 points to both condition was considered such as deficiency visuoconstructional abilities and no verbal memory, respectively.

### Mini-Mental State Examination (MMSE)

In our study we use the MMSE version of 35 points (Folstein et al., 1975). Each subject was examined in 5 mental state: TABLE 1 | Traditional evaluation scoring for Rey-Osterrieth complex figure.


Frank and Landeira-Fernandez (2008).

Orientation (10 points), Registration (3 points), Attention and calculation (8 points), Recall (3 points) and language (12 points). The presence of subject with values <24 points were considered cognitive alteration.

All analysis of each drawing of Rey-Osterrith Complex Figure test and MMSE were performed by blind evaluator at sea level. The lower neuropsychological test selection for our study was given by empirical experiences; these tests were significant at high altitude in workers with CIHH at 3,500 and 4,600 m (unpublished data).

### Statistical Analysis

All results were expressed as mean ± standard deviation. Cardiorespiratory variables (SpO2, HR, SAP, DAP) differences on means were analyzed using ANOVA followed by a Newman-Keuls test. Wilcoxon (Non-parametric test) to paired comparison was used to compare the difference of neuropsychological test. All differences were considered statistically significant when p < 0.05. Data analyses were performed using GraphPad Prism version 5.03 (GraphPad Software, Inc.).

### RESULTS

### Cardiorespiratory Variables at 2,900 and 5,050 m

**Table 2** show a summary of the cardiorespiratory response obtained at camp base (OSF at 2,900 m) before ascending to the AOS at 5,050 m and cardiorespiratory values obtained in the AOS at 5,050 m. Pulse oximetry was decreased to values of 84 ± 0.8% (p < 0.05) that represent a fall approximately 9.7% in the arterial oxygenation and heart rate was increased to values of 90 ± 8 bpm (p < 0.05) that represented an increase of 16.8%, systolic arterial pressure was increased to values of 135 ± 18 mmHg that represented an increase of 11.6%, diastolic arterial pressure was increased to values of 96 ± 3 mmHg (p < 0.05) that represented an increase of 18.5% and mean arterial pressure was increased to values of 109 ± 8 mmHg (p < 0.05) that represented an increase of 15.9% was observed in all subjects when they arrived at 5,050 m, respect of values obtained in camp base (OSF, 2,900 m). During the protocol at 5,050 m when the subjects breathed room air, no modifications to the cardiorespiratory variables were observed. However, when subjects were placed in the module with an increase in FiO<sup>2</sup> to 28%, we observed a fast increase in the pulse oximetry to values of 91 ± 0.8 % (p < 0.05) that represented an increase of 8% arterial oxygenation and a fall in the cardiovascular response such as heart rate to values of TABLE 2 | Cardiorespiratory response of workers exposed at 2,900 and 5,050 m.


Values expressed as Mean ± SD. SpO2, pulse oximetry; SAP, systolic arterial pressures; DAP, diastolic arterial pressure; MAP, mean arterial pressure. \*p < 0.001 Room Air 2,900 vs. Room Air 5,050 m and 5,050 m + 8% O2, † p < 0.01 Room air 5,050 m vs. Room Air 5,050 m + 8% O2.

77 ± 12 bpm (p < 0.05) that represented a decreases by 14.4%, systolic arterial pressure to values of 128 ± 10 that represented a decreases 5.2%, diastolic arterial pressure to values of 87 ± 5 mmHg (p < 0.05) that represented a significant fall of 9.4% and mean arterial pressure to values of 101 ± 7 mmHg (p < 0.05) that represented a significant fall of 7.3%. No differences were observed in cardiorespiratory variables between OSF and AOS + O2, reaching values similar to those observed at 2,900 m (OSF). **Figure 2** shows a typical pattern of oxygen saturation and heart rate recordings in a worker during a period of 40 min.

### Neuropsychological Assessment

Neuropsychological evaluations were performed with MMSE and revealed that when the subjects were tested in environmental oxygen conditions, 30% of the population presented disorders. However, upon entering the module with oxygen enrichment, this was reduced to 10% (p < 0.05). **Figure 3** represents the individual score obtained by MMSE.

When asked to evoke the Rey-Osterrieth Complex Figure test (memory), this capacity was significantly reduced in workers exposed to room air, with 50% of the patients meeting the criteria of a memory disorder (**Figures 4A,B**). However, when subjects were tested in the oxygen enriched module, the percentage of subjects below the cut off criteria was 10%, indicating a significant difference compared to subjects tested in environmental oxygen conditions.

### DISCUSSION

This study is the first to assess the effect of environmental oxygen enrichment at 28% in very high altitude conditions (5,050 m) in a population of volunteers acclimatized to chronic and intermittent hypobaric hypoxia exposure for at least 4 years. Our results demonstrate that the mobile module that permits increased oxygen concentration, used to simulate an altitude equivalent to 2,900 m, significantly improved oxygen saturation and reduced heart rate and diastolic blood pressure, and enhanced neuropsychological performance in workers exposed to an altitude of 5,050 m.

Studies performed in CIHH models are scarce. Richalet et al. (2002) published results of a prospective study of workers exposed to CIHH for 2.5 years at an altitude of 4,500 m and described a developing acclimatization pattern along with emphasized health risks; right ventricular dilation and increased blood pressure. In addition, the incidence of acute mountain sickness is elevated on the first day, and is reduced, but does not disappear, with exposure time (Richalet et al., 2002). In addition, Brito et al. (2007) reported tachycardia, high blood pressure and low oxygen saturation after 12 years of CIHH exposure at 3,550 m. In addition, Siqués et al. (2009) found a significant increase in the diastolic pressure associated to lower oxygen saturation in young adults after 12 months of exposure at the same altitude. A study in acclimatized miners (1–14 years of CIHH exposure) showed the existence of periodic apnea breathing with episodes of arterial oxygen saturation reaching 67% (Moraga et al., 2014) at an altitude of 4,200 m. This severe level of hypoxemia triggers an increase in sympathetic tone. Previous studies support the observation that high altitude exposure promotes sympathetic activation and increases arterial pressure (Wolfel et al., 1994; Hansen and Sander, 2003). A preliminary study showed that 63% of people working at an altitude of 4,600 m had increased arterial pressure (>140/>90 mmHg) and oxygen saturation (<80%) (Moraga, unpublished results). In previous studies performed in our laboratory demonstrated that oxygen enrichment by 3–4% in a room at 4,200 m simulates an oxygen equivalent of 3,200 m, enhancing oxygen saturation, reducing heart rate, and inducing a general feeling of well-being in the morning. A better explanation of the effect of a decreased equivalent altitude with increased oxygen room is a decrease in the activation of the peripheral chemoreceptor as a consequence of reduced central neurosystem activation (see Dempsey et al., 2014). Previous studies support the role of carotid bodies in animal models where the carotid afferent

was cut off, resulting in the abolishment of the sympathetic and hypertensive responses (see Prabhakar et al., 2012; Del Rio et al., 2016; Iturriaga et al., 2017).

In regard to the neuropsychological performance evaluated in our study, acute decreased oxygen availability at high altitudes is known to produce impaired mental and physical performance, increasing as the altitude also increases. The first studies performed in miners at 3,800 m, showed deterioration in cognition and motor function (Mc Farland, 1937). The first study that reported modifications in the neuropsychological response in miners exposed to CIHH, showed a decrease by

breathing room air of 5,050 m (open circles) and room air of 5,050 m enriched with oxygen (closed circles). Values represent the score obtained of each subject in both condition. Lines represent the mean ± SD, asterisk p < 0.001.

5% in the general cognitive aptitude on the 5th shift day at 4,500 m. The most affected elements were spatial aptitude (13% decrease) and mathematical reasoning (28% decrease) and neuropsychological attention (11% decrease) compared to evaluations performed at sea level (Jiménez, 1995). Additionally, in controlled hypobaric hypoxia conditions with an over imposed hypoxia of 5,000 m, neuropsychological function after increase of 6% percentage of oxygen on room, in no acclimatized voluntaries, reported an increase in oxygen saturation, quicker reaction times, improved hand-eye coordination, and more positive sense of well-being (Gerard et al., 2000). However, the study showed no significant improvement on neurophysiological function (Gerard et al., 2000). Therefore, our study is the first to assess neuropsychological performance in a population of acclimatized workers exposed to CIHH at a very high altitude (5,050 m) where room air levels were supplemented with oxygen at FiO<sup>2</sup> 28% to obtain an equivalent altitude of 2,900 m. Exposure to high altitude decreased neuropsychological and cognitive functions for memory and constructive praxis (Rey-Osterrieth Complex Figure test, MMSE), however, this decreased response was reverted by increasing environmental oxygen at 28% in the room, supporting the notion that this response is due to low oxygen availability at very high altitudes, and our results suggest that this is a plastic response.

By other way, since the early work of Paul Bert in 1878, it has been assumed that the lower inspiratory pressure of oxygen is the primary stimulus for adaptation to hypoxia (See Conkin and Wessel, 2008). Later, Barcroft in 1925 showed that decreasing FIO<sup>2</sup> could produce a hypoxic condition (See Richard and Koehle, 2012). These studies were the starting point for the discussion about whether the physiological responses are equivalent in normobaric hypoxia and hypobaric hypoxia models (Savourey et al., 2003; Richard and Koehle, 2012) in order to predict susceptibility to suffer from acute mountain sickness (Savourey et al., 2003; Burtscher et al., 2008), optimal training for athletes (Levine and Stray-Gundersen, 2006) and others. In a recent review, evidence supported a difference in ventilation mechanisms between both hypoxia models (Coppel et al., 2015). In addition, the same study showed that the confounding factors such as time spent, temperature, humidity and statistical power, limit the conclusion of these finding (Coppel et al., 2015). However, a series of studies have shown that the hypobaric conditions of high altitude could have physiological effects that are independent of the fall in oxygen pressure (i.e., temperature, air density, fluid balance, the presence of a microbubble in pulmonary circulation, increased dead space, etc.) (Loeppky et al., 2005; Conkin and Wessel, 2008; Richard and Koehle, 2012; Coppel et al., 2015); but it is necessary to indicate that many of these studies were performed in simulated conditions (hypobaric or normobaric chambers). These conditions are very different to field studies, like that performed in our study since this evaluates workers acclimatized to CIHH at very high altitude (5,050 m), where increasing FiO<sup>2</sup> enhances wellbeing during their stay at high altitude. In a previous study, miners acclimatized to CIHH showed a different response to hemoconcentration during maximum exercise performed at sea level and high altitude (3,800 m), given by a mechanism mediated by a negative water balance (low intake, high loss) (Moraga et al., 2017), supporting the idea that the hypobaric condition of high altitude promotes a different response. Another, study evaluated heart rate and oxygen saturation during a hike at high altitude (Manua Kea, 4,205 m) and compared this with a simulate condition (normobaric hypoxia). The results showed that oxygen saturation was lower and heart rate was higher at high altitude compared with normobaric hypoxia (Netzer et al., 2017). More studies are necessary to improve our understanding of the mechanisms involved in tolerance and acclimatization to hypoxia, in all hypoxia exposure models (in our case, in miners exposed to CIHH).

### LIMITATIONS

In our study it was impossible to perform a crossover long term study in the workers as they were only authorized by their supervisors to participate in the first evaluation, due to the high labor demand during the study period.

In conclusion, our study showed that the use of a mobile module oxygen system at 5,050 m, simulating an altitude equivalent to 2,900 m, significantly improves arterial oxygenation and heart rate, reduces diastolic blood pressure, and enhances neuropsychological performance. Our data support the requirements that the health ministry should implement to enhance the environmental work place; preventing and reducing the risk of labor at altitudes over 3,000 m.

### AUTHOR CONTRIBUTIONS

FM conceived and designed the study. IL and AM supervises the overall the study. AM performed the statistical analysis. DS, AM,

### REFERENCES


and JN contributed to sample and data collections. All authors drafted the report. All authors contributed to the interpretation of the results, critical revision of the manuscript and approved the final manuscript. FM is the guarantor.

### FUNDING

This work was supported by the PI 537 I+D CORFO grant, Chile.

### ACKNOWLEDGMENTS

We are grateful to all the volunteers from ALMA that participated in the study, and also for the facility support by the Safety Manager – ALMA.


Wolfel, E. E., Selland, M. A., Mazzeo, R. S., and Reeves, J. T.(1994). Systemic hypertension at 4,300 m is related to sympathoadrenal activity. J. Appl. Physiol. 76, 1643–1650. doi: 10.1152/jappl.1994.76. 4.1643

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Moraga, López, Morales, Soza and Noack. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Leptin Signaling in the Carotid Body Regulates a Hypoxic Ventilatory Response Through Altering TASK Channel Expression

Fang Yuan1,2, Hanqiao Wang<sup>3</sup> , Jiaqi Feng<sup>1</sup> , Ziqian Wei <sup>1</sup> , Hongxiao Yu<sup>1</sup> , Xiangjian Zhang2,4 , Yi Zhang1,2 \* and Sheng Wang1,2 \*

<sup>1</sup> Department of Physiology, Hebei Medical University, Shijiazhuang, China, <sup>2</sup> Hebei Key Laboratory of Vascular Homeostasis and Hebei Collaborative Innovation Center for Cardio-Cerebrovascular Disease, Shijiazhuang, China, <sup>3</sup> Department of Sleep, Third Hospital of Hebei Medical University, Shijiazhuang, China, <sup>4</sup> Department of Neurology, Second Hospital of Hebei Medical University, Shijiazhuang, China

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Andrea Porzionato, Università degli Studi di Padova, Italy Silvia V. Conde, Faculdade de Ciências Médicas, Universidade Nova de Lisboa, Portugal Ana Obeso, University of Valladolid, Spain

### \*Correspondence:

Yi Zhang yizhanghebmu@163.com Sheng Wang wangsheng@hebmu.edu.cn

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 17 January 2018 Accepted: 06 March 2018 Published: 27 March 2018

### Citation:

Yuan F, Wang H, Feng J, Wei Z, Yu H, Zhang X, Zhang Y and Wang S (2018) Leptin Signaling in the Carotid Body Regulates a Hypoxic Ventilatory Response Through Altering TASK Channel Expression. Front. Physiol. 9:249. doi: 10.3389/fphys.2018.00249 Leptin is an adipose-derived hormone that plays an important role in the regulation of breathing. It has been demonstrated that obesity-related hypoventilation or apnea is closely associated with leptin signaling pathways. Perturbations of leptin signaling probably contribute to the reduced sensitivity of respiratory chemoreceptors to hypoxia/hypercapnia. However, the underlying mechanism remains incompletely understood. The present study is to test the hypothesis that leptin signaling contributes to modulating a hypoxic ventilatory response. The respiratory function was assessed in conscious obese Zucker rats or lean littermates treated with an injection of leptin. During exposure to hypoxia, the change in minute ventilation was lower in obese Zucker rats than chow-fed lean littermates or high fat diet-fed littermates. Such a change was abolished in all groups after carotid body denervation. In addition, the expression of phosphorylated signal transducers and activators of transcription 3 (pSTAT3), as well as putative O2-sensitive K<sup>+</sup> channels including TASK-1, TASK-3 and TASK-2 in the carotid body, was significantly reduced in obese Zucker rats compared with the other two phenotype littermates. Chronic administration of leptin in chow-fed lean Zucker rats failed to alter basal ventilation but vigorously increased tidal volume, respiratory frequency, and therefore minute volume during exposure to hypoxia. Likewise, carotid body denervation abolished such an effect. In addition, systemic leptin elicited enhanced expression of pSTAT3 and TASK channels. In conclusion, these data demonstrate that leptin signaling facilitates hypoxic ventilatory responses probably through upregulation of pSTAT3 and TASK channels in the carotid body. These findings may help to better understand the pathogenic mechanism of obesity-related hypoventilation or apnea.

Keywords: leptin, hypoventilation, hypoxic ventilatory response, carotid body, TASK channels, STAT3

# INTRODUCTION

Leptin, a peptide hormone secreted mainly by adipocytes, regulates multiple physiological functions including metabolism, cardiovascular activity, and breathing (Grill et al., 2002; Bassi et al., 2015). Leptin's role in controlling breathing has been implicated in recent studies, with its participation in sleep-related breathing disorders including obesity hypoventilation syndrome

**104**

(OHS) and obstructive sleep apnea (Malhotra and White, 2002). In animal models, leptin deficient mice exhibit impaired ventilatory responses to CO<sup>2</sup> which can be rescued by leptin replacement therapy, in favor of facilitation of breathing by leptin (O'Donnell et al., 2000). Leptin receptors (ob-Rs) are composed of six isoforms termed from ob-Ra to -Rf, with the long form of ob-Rb mediating the majority of leptin's intracellular signal transduction (Tartaglia, 1997). Although the role of leptin signaling pathways in mediating various physiological actions has been investigated intensively, the molecular mechanism underlying its action on breathing remains incompletely understood.

The peripheral respiratory chemoreflex serves as a homeostatic regulatory mechanism by which enough oxygen must be supplied to the organism when challenged by hypoxia through altering respiratory amplitude and frequency. The carotid body (CB) chemoreceptors, located near the fork of the carotid artery, are activated shortly after exposure to hypoxia and then send information to the nucleus tractus solitarius and high integrative centers, with the outcome of adaptive ventilatory responses (Gonzalez-Martin et al., 2011; Ciriello and Caverson, 2014). Accumulated evidence indicates the presence of ob-Rb in the CB cells (Porzionato et al., 2011; Messenger et al., 2013), and that leptin signaling contributes to CB-mediated ventilatory responses (Olea et al., 2015; Ribeiro et al., 2017). However, it remains controversial whether the CB mediates the acute effect of leptin on hypoxic ventilatory response (HVR) because leptin's role may involve the change in gene expression and protein synthesis, requiring hours even days for full effects (Hall et al., 2010). We thereby predicted that the stimulatory effect of leptin on HVR may require chronic activation of CBs, but such a confirmation is yet to be put forward.

In the CBs, leptin signaling pathways involve the downstream signaling proteins of ob-R signal transducers and activators of transcription 3 (STAT3), suppressor of cytokine signaling 3 (SOCS3), and extracellular-signal-regulated kinase 1/2 (ERK1/2) (Messenger et al., 2013; Moreau et al., 2015), reminiscent of modulatory effects of these molecules on carotid chemoreceptor sensitivity. Emerging evidence has shown that the chemosensitivity of glomus cells in CBs requires two-pore K <sup>+</sup> channels including TWIK-related acid-sensitive K (TASK)-1 channels and acid-sensitive ion channels (Trapp et al., 2008; Tan et al., 2010). However, very little is known concerning whether the activation of the ob-R and downstream signaling molecules modulates the sensitization of CB chemoreceptors via affecting these ion channels.

We sought to address herein whether the leptin signaling pathway in the CB contributes to regulating HVR and the possible mechanism involved. We utilized whole body plethysmography (WBP) to assess HVR in obese Zucker rats (ob-R deficiency) or in lean littermate controls treated with injections of leptin. The main findings suggest that chronic application of leptin contributes to facilitation of HVRs probably through upregulation of phosphorylated STAT3 (pSTAT3) and TASK channel expression.

# MATERIALS AND METHODS

### Animals

The experiments were carried out in 12∼20-week-old male obese Zucker rats (OZR) and lean littermates (LZR) obtained from the Charles River Laboratories (USA). Animals, synchronized for a 12:12 h light-dark cycle (lights on at 8 am, lights off at 8 pm), were housed individually and allowed to move freely in standard plastic cages in a climate-controlled room (22 ± 1 ◦C). Food and water were provided ad libitum for LZRs and OZRs. In some cases, a group of LZRs were placed on a highfat diet (HFD, 45% kcal/g fat, Research Diets D12451) for 8 weeks and used as a simple obesity control (LHZR). The LHZR and OZR groups were weight-matched to determine the effect of simple obesity-induced mechanical resistance on ventilation. Body weight was measured once a week (n = 20 for each phenotype). All experiments were performed in accordance to ethical guidelines of the Animal Protection Association and were approved by Animal Care and Ethical Committee of Hebei Medical University. When the animal experiments were completed, an overdose of intraperitoneal injection of sodium pentobarbital (> 200 mg/kg) was carried out for euthanasia.

### Breathing Measurement

Breathing was studied by WBP in conscious, freely moving rats (EMKA Technologies, France) as described previously (Kumar et al., 2015; Fu et al., 2017). In brief, rats were placed in the WBP chamber on the day before the testing protocol (2 h acclimation period). For acute hypoxia, rats were exposed to 10% O<sup>2</sup> (balance N2) for up to ∼7 min by a gas mixture devices (1,500 ml/min, GSM-3, CWE, USA). Ventilatory flow signals were recorded, amplified, digitized and analyzed using IOX 2.7 (EMKA Technologies) to determine breathing parameters over sequential 20 s epochs (∼50 breaths) during periods of behavioral quiescence and regular breathing. Minute volume (VE; ml/min/g) was calculated as the product of the respiratory frequency (fR, breaths/min) and tidal volume (VT; ml/kg), normalized to rat body weight (g). To further confirm the CB-mediated effect of leptin, breathing parameters were also measured in rats with carotid body denervation (CBD). The carotid sinus nerves were sectioned as depicted before (Kumar et al., 2015). Shortly, anesthesia was induced with 4% halothane in 100% O<sup>2</sup> and maintained by reducing the inspired halothane concentration to 1.5∼1.8%. The depth of anesthesia was assessed by an absence of the corneal and hindpaw withdrawal reflex. Body temperature of all mice was maintained at 37◦C using a temperature-controlled heating pad. To prevent any functional regeneration of chemosensory fibers, the carotid sinus nerves were removed completely from the cranial pole of the CB until reaching the branch to the glossopharyngeal nerve. The wound was carefully sutured and disinfected with 10% of polividone iodine. Conscious chemodenervated rats were exposed to ventilatory challenge 5–7 days after recovery. No significant weight loss was observed. All the three groups of rats (LZR, LHZR and OZR) were submitted to surgery.

To examine whether hypoventilation resulted in retention of CO<sup>2</sup> in obese rats, arterial blood gas was measured using an OPTI-CCA blood gas analyzer (OPTI Medical Systems, USA) at a steady state in halothane-anesthetized, paralyzed rats. General anesthesia was induced with 4% halothane in room air as depicted above. Arterial blood (200 µl per sample) was drawn from the femoral artery in the three animal groups. Arterial blood measurements of interest included partial pressure of arterial O<sup>2</sup> (PaO2), partial pressure of CO<sup>2</sup> (PaCO2) and pH.

## Plasma Leptin Levels and Hypodermic Leptin Injections

Measurements of plasma levels of leptin were performed at room air (21% O2) in anesthetized rats. After general anesthesia as described above, whole blood samples were taken through a cardiac puncture. Blood samples were drawn into collection tubes containing the anticoagulant EDTA (Sigma-Aldrich, USA) and kept on ice. After centrifugation, the plasma was stored at−80◦C for leptin analysis by ELISA kit (#ab100773, Abcam, USA), an in vitro enzyme-linked immunosorbent assay for the quantitative measurement as previously described (Panetta et al., 2017). The assay was read using a power wave XS2 plate reader (Biotek Instruments, USA).

To confirm whether the chronic activation of leptin signaling pathways played a part in the HVR, subcutaneous injections of leptin (60 µg/kg) or equal volume of vehicle (saline) were carried out once daily for 7 days in CBI LZRs (n = 8 for each group), and breathing parameters were measured after 7 day injections during exposure to room air or hypoxia. To further confirm the CB's role, subcutaneous injections of leptin or saline were performed 7 days after the carotid sinus nerves were sectioned in each group (n = 8 for both).

### Carotid Body Protein Extracts and Western Blot Analysis

Rats were deeply anesthetized by isoflurane (2–3%) inhalation and then decapitated. The carotid bifurcation was exposed and both CBs were removed and cleaned. The CB (n = 8) were pooled from 4 rats in each group and homogenized in 100 µl of RIPA buffer solution (150 mM NaCl, 1 mM EDTA, 1% Triton-X 100, 50 mM Tris–HCl at pH of 7.5) with a protease inhibitor cocktail (Roche Diagnostics, Canada). The homogenate was centrifuged at 4◦C for 20 min at 13200 rpm. The resultant supernatant was retained as the protein preparation. Equal concentrations of extracted proteins normalized by colorimetric BCA Protein Assay (Pierce Corp., USA). After denaturation, the protein (∼30µg) in each lane was fractionated in 10% polyacrylamide gel and then transferred onto apolyvinylidene fluoride membrane. The membranes were blocked with bovine serum albumin and incubated at 4◦C overnight with primary antibodies anti-ob-R (1:2000, #ab5593, Abcam, USA), anti-TASK-1 (1:200, #APC024, Alomone labs, Israel), anti-TASK-2 (1:200, #APC037, Alomone labs, Israel), anti-TASK-3 (1:200, #APC044, Alomone labs, Israel), anti-pSTAT3 (1:2000, #9145, Cell Signaling Technology, USA), anti-STAT3 (1:1000, #9139, Cell Signaling Technology, USA), and anti-β-actin (1:3,000, #T0022, Affinity Biosciences, USA). The membranes were then incubated with corresponding secondary antibodies for 1 h at room temperature. The reaction was visualized using the enhanced chemiluminescence (ECL) method, and the bands were analyzed by ImageJ software (NIH, USA). The protein contents were normalized to β-actin. See Supplementary Material for original gel images.

# Data Acquisition and Processing

Data are expressed as mean ± SEM. Unless indicated otherwise, two-tailed unpaired t-test, one-way ANOVA with Dunnett's or Tukey's post-hoc test and two-way ANOVA with Bonferroni post-hoc test were used to compare significant difference between different groups. Differences within or between groups with P-values of <0.05 were considered significant.

# RESULTS

## Reduced Basal Ventilation in OZRs

Adult OZRs exhibit many abnormal physiological attributes due to the deficiency of the ob-R, representing a good animal model to study the obesity-related hypoventilation as observed in humans. We thus addressed leptin's role using this phenotype and the littermate control rat. First, we measured the body weight of three groups of rats (n = 20 for each group). The averaged body weight was larger in LHZRs (492 ± 30 g) and OZRs (512 ± 49 g) than LZRs (382 ± 14 g, P < 0.01 vs. LHZR or OZR). In addition, weight gain was accompanied by higher levels of blood plasma leptin level in OZRs (15.28 ± 0.55 µg/L) in relative to LHZRs (7.45 ± 0.50 µg/L, P < 0.01 vs. OZR) or LZRs (7.34 ± 0.38 µg/L, P < 0.01 vs. OZR; P > 0.05, vs. LHZR).

Baseline breathing parameters were measured in the three groups of animals while breathing room air (21% O2). Compared with the LZRs, V<sup>E</sup> and V<sup>T</sup> were considerably lower in OZRs and LHZRs (P < 0.01 for both, vs. LZRs, **Figures 1A,C**), whereas the OZR has a faster f<sup>R</sup> than the other two groups (P < 0.01 for both). Of interest, although no remarkable difference in basal V<sup>E</sup> was observed between OZRs and LHZRs, V<sup>T</sup> and f<sup>R</sup> were comparable (P < 0.01 for both, **Figure 1B**), indicative of a different breathing pattern. To evaluate whether hypoventilation resulted in hypercapnia or respiratory acidosis in obese animals, the arterial blood gas was measured at a steady-state. Apparently, hypercapnia, acidosis and normal PaO<sup>2</sup> were observed in LHZRs and OZRs (**Table 1**). Therefore, both OZRs and LHZRs exhibit basal hypoventilation, overweight and hypercapnia, with the exception of hyperleptinemia in OZRs but not LHZRs.

# Impairment of HVR in OZRs

To address whether the ob-R deficiency yielded a diminished HVR, hypoxia was achieved while inhaling 10% O<sup>2</sup> to activate peripheral respiratory chemoreflex. When acutely challenged with hypoxia, all three phenotypes displayed robust increases in f<sup>R</sup> and V<sup>E</sup> except for VT, with the smallest change in the HVR in OZRs (**Figures 1A–C**). In addition, the hypoxiastimulated increment of V<sup>E</sup> was far less in OZRs compared to the other two groups (P < 0.01, **Figure 1E**). Interestingly, LZRs and LHZRs displayed a similar change in V<sup>E</sup> during hypoxia


LZR, lean Zucker rats; LHZR, HFD-fed LZR; OZR, obese Zucker rats; n = 12 for each group; \*P < 0.05, \*\*P < 0.01 by one-way ANOVA with Tukey's post-hoc test, vs. LZRs.

(P > 0.05, **Figure 1E**). In spite of similar degree of body weight, OZRs exhibited far more severe hypoventilation in response to hypoxia compared to LHZRs (P < 0.01, **Figures 1A–C**). However, exposure to hypoxia caused no significant difference in increments of V<sup>E</sup> in all three groups of rats after sectioning carotid sinus nerves (P > 0.05, **Figures 1D,E**), in favor of the involvement of CBs in such an effect. Collectively, the ob-R deficiency (OZR), instead of simple obesity (LHZR), plays a predominant role in the impaired HVR.

# Downregulation of pSTAT3 and TASK Channels in OZR CBs

The pSTAT3/STAT3 signaling has been implicated in mediating major effects of the ob-R. To determine the expression level of ob-R and STAT3, the quantitative analysis was made using Western blot in the three groups (n = 4 for each group). Compared with LZRs and LHZRs, pSTAT3/STAT3 (**Figures 2C,D**) were remarkably downregulated in the CBs of OZRs (P < 0.01), reliably correlating pSTAT3/STAT3 expression with the ob-R deficiency (**Figures 2A,B**). Several lines of evidence demonstrated that the chemosensitivity is associated with TASK-1 and TASK-3 in CB glomus cells and TASK-2 in retrotrapezoid nucleus neurons (Trapp et al., 2008; Gestreau et al., 2010; Wang et al., 2013). The reduced sensitivity of CBs to hypoxia in OZRs

was probably associated with these ion channels. Evidently, the expression level of TASK-1, TASK-2, TASK-3 in OZR CBs was lower in relative to the other two groups (P < 0.05∼0.01, **Figures 2E–H**). However, no statistical significance in these channel expression was found between LZRs and LHZRs (P > 0.05 for all, **Figures 2E–H**). Hence, the ob-R deficiency contributes to reduced expression of pSTAT3 and TASK channels.

# Facilitation of HVR by Chronic Application of Leptin

To examine the effect of activation of ob-Rs on the HVR, subcutaneous injections of leptin (60 µg/kg) or equal volume of normal saline were carried out once daily for 7 days in LZRs (n = 8 for each group), and breathing parameters were measured at different time points separated by 7 days. As shown in **Table 2**, compared with the vehicle control (8.2 ± 0.4µg/L), the plasma levels of leptin was raised to 13.1 ± 0.6µg/L (P < 0.01) over 7 day treatment and restored to control level after 1 week. Chronic administration of leptin for 7 days produced no significant change in body weight (P > 0.05, data not shown) and basal breathing parameters (VT, f<sup>R</sup> and VE) in relative to the vehicle control (**Figures 3A–C**). Neither PaCO<sup>2</sup> nor blood pH changed significantly in leptin-injected rats (data not shown). During exposure to 10% O2, VT, fR, and V<sup>E</sup> were all increased in either leptin- or saline-injected LZRs (P < 0.05∼0.01) but the change in V<sup>E</sup> was greater in leptin-injected rats in relative to the vehicle controls (**Figure 3E**). The stimulatory effect of leptin on HVR persisted for at least 2 weeks (**Table 3**). After bilaterally sectioning carotid sinus nerves, leptin-induced increase in V<sup>E</sup> was abolished (**Figures 3D,E**). Collectively, chronic administration of leptin potentiated the HVR.

# Effect of Leptin on Expression of pSTAT3 and TASK Channels

To investigate the possible mechanism underlying exogenous application of leptin action on the HVR, we tested the expression of ob-R and the downstream pSTAT3 and TASK-1, TASK-2, TASK-3 channels in CBs after injection of leptin for 7 days. The findings indicated that leptin administration caused greater upregulation of ob-R and of pSTAT3 (P < 0.01, n = 4 for each group, **Figures 4A,B**). Furthermore, leptin also enhanced expression of TASK-1, TASK-2, TASK-3 (P < 0.01, n = 4 for each group, **Figures 4C–E**). The results suggest that the stimulatory effect of leptin on HVR are closely associated with enhanced expression of pSTAT3 and TASK channels, which may contribute to the regulation of CB's chemosensitivity.

# DISCUSSION

We demonstrate herein that the ob-R deficiency, rather than simple obesity, not only reduces baseline ventilation but also inhibits the HVR, with decreased pSTAT3 expression in CBs.



n = 8 for each group; \*\*P < 0.01 by two-tailed unpaired t-test, vs. saline.

Chronic administration of leptin has no marked effects on basal ventilation but considerably enhances the HVR, accompanied by increased expression of pSTAT3. Additionally, either ob-R deficiency or leptin administration is reliably associated with changes in expression of TASK-1, TASK-2, and TASK-3 channels. These findings suggest that leptin signaling in the CB contributes to potentiation of HVR probably through enhancement of pSTAT3 and TASK channels expressions.

The obese Zucker rat represents a good model of ob-R deficiency and manifests relatively early onset obesity (Bray and York, 1971). One line of early evidence shows that respiratory system compliance was significantly lower in the OZR compared with the lean phenotype, and that resting ventilatory parameters (uncorrected for body weight) were similar between obese and lean animals, with similar ventilatory response to hypoxia between two phenotypes (Farkas and Schlenker, 1994). In the present study, we compared the difference between obese and lean phenotypes using the previously described method (Kumar et al., 2015; Fu et al., 2017) to normalize breathing parameters to body weight. Interestingly, with this analysis method, the OZRs exhibited fast fR, reduced V<sup>T</sup> and thus lower V<sup>E</sup> during exposure to room air. This outcome interprets occurrence of hypercapnia and respiratory acidosis observed in OZRs. Based on these attributes, the OZR resembles an animal model of leptin resistance. The LHZR, a simple obesity control carrying genotype of the LZR, displayed hypercapnia, respiratory acidosis, relatively normal serum leptin, basal hypoventilation and moderate response to hypoxia, probably representing a model of simple obesity rather than leptin resistance. In obese patients, a higher level of leptin is found to cause an increase in basal ventilation associated with excess body mass, with OHS patients exhibiting an even higher serum leptin level than eucapnic individuals matched for body mass index (Phipps et al., 2002). In animal models, HFD rats exhibit an unchanged (Olea et al., 2014) or enhanced basal V<sup>E</sup> (Ribeiro et al., 2017). Importantly, Ribeiro et al. found 3 weeks of HFD blunted leptin responses to hypoxia in the CB, probably due to development of CB leptin resistance, suggesting at least 3 weeks required for the establishment of leptin resistance. However, in the present study, hyperleptinaemia and leptin resistance did not occur in the LHZR most likely because of the genetic manipulation which makes the difference and the hypoventilation thus appears to be a restrictive ventilatory pattern. Another explanation may be supported by previous findings suggesting that elevation of leptin levels is a consequence of hypoxia and not of fat accumulation (Tatsumi et al., 2005). Taken together, our findings support that the impaired basal ventilation and HVR in OZRs is ascribed mainly to ob-R deficiency rather than mere obesity. We did not measure chest wall mechanics and compare the difference in chest wall impedance between two obese rats, whereas it would be expected that obesity-induced increase in chest wall impedance must play a relatively small part in such effects. This also helps better understanding of the results that 1V<sup>E</sup> induced by hypoxia is insignificant between LZRs and LHZRs.

Leptin's actions involve fast or slow onset, thus requiring minutes, several hours, even days before major changes occur. Its acute effects on breathing has been implicated in prior studies



each group, \*P < 0.05, \*\*P < 0.01 by two-way ANOVA with Bonferroni post-hoc test, Leptin vs. Saline group.

V<sup>E</sup> , minute ventilation, n = 8 for each group, \*\*P < 0.01 by two-way ANOVA with Bonferroni post-hoc test, vs. Saline.

(Chang et al., 2013; Olea et al., 2015; Pye et al., 2015; Ribeiro et al., 2017). For example, Olea et al. have found that acute application of leptin in anaesthetized animals augmented basal V<sup>E</sup> and potentiated the ischemic hypoxia-induced V<sup>E</sup> in a dosedependent manner (Olea et al., 2015). More recently, Ribeiro et al. reported that leptin increases V<sup>E</sup> both in basal and hypoxic conditions in control rats but such effects were blunted in high fat diet fed rats (Ribeiro et al., 2017). In contrast, acute application of leptin in isolated CB type I cells failed significantly to alter the resting membrane potential and acidification-induced depolarization was unaffected by leptin, thereby suggesting that acute leptin stimulation did not alter CB's chemosensitivity (Pye et al., 2015). In addition to acute effects, chronic treatment with leptin in vivo also has been shown to potentiate respiratory chemoreflex (Bassi et al., 2014). Acute or chronic actions may be mediated by different leptin signaling pathways. Chronic administration of leptin in the present study did not augment basal ventilation but potentiated HVR in conscious rats, an effect persisting for ≥7 days. This appears to play an essential role in reinforcing ventilation to supply more oxygen when challenged by hypoxia. The plasma levels of leptin after 7 days treatment are quite similar to those quantified in plasma for the

OZRs, taken together with the enhancement of HVR with 7 day leptin administration and the impairment of HVR in OZRs rats, indicates that 7 day stimulatory effects of leptin did not result in leptin resistance.

The reason why the basal ventilation was herein not enhanced by the chronic application of leptin would be attributable to animal's state (anesthetized vs. conscious) and the dose of leptin administered. The dose of leptin applied to our animals was chosen based on what was chronically applied previously (Wjidan et al., 2015), and lower than that used in prior reports examining the acute effects of leptin on cardiovascular (Rahmouni et al., 2002; Rahmouni and Morgan, 2007) and respiratory functions (O'Donnell et al., 2000; Bassi et al., 2012). Moreover, in the case of potentially therapeutic utilizations, this dose would be expected to specifically exert a respiratory but not excessive cardiovascular effects. Higher concentrations of leptin would be expected to saturate plasma carrier molecules and some of unbound leptin would be degradated. Furthermore, since the amount of leptin to cross blood-brain barrier is relied on receptor-mediated transport mechanism (Morris and Rui, 2009), access to the brain should be limited somehow.

Leptin's intracellular signal transduction has been extensively investigated, with exception of molecular mechanisms underlying its effect on respiratory chemosensitivity. It remains poor understood how the activation of leptin signaling affects O2 sensitive channels which may determine CB's chemosensitivity. Along with recent studies indicating that the chemosensitivity of CBs is closely associated with TASK-1 and TASK-3 channels (Trapp et al., 2008; Tan et al., 2010), TASK-2 channels has also been evidenced to set central respiratory CO<sup>2</sup> and O<sup>2</sup> sensitivity (Gestreau et al., 2010). In addition, activation of ob-Rs appears to regulate ion channels including ATP-sensitive K<sup>+</sup> channels and voltage-gated K<sup>+</sup> channels (Gavello et al., 2016). Although STAT3, SOCS3, and ERK1/2 may mediate leptin's role in the CBs, the critical information is lacking to data concerning modulatory effects of these molecules on TASK channels. In the present study, we did not directly address how the activation of ob-Rs and downstream signaling proteins regulated TASK-1, TASK-3, and TASK-2 channels, but notably, the altered expression levels of these channels would be expected to be attributable to leptin signaling and contribute to leptin-stimulated facilitation of the HVR. Future work is required for revealing such mechanisms.

In summary, leptin signaling participates in setting CB's O<sup>2</sup> sensitivity probably through the modulation of TASK-1, TASK-3, and TASK-2 channels and thus contributes to the potentiation of HVR. This line of cellular evidence extends our understanding of molecular mechanism of leptin action on breathing, shedding light on the etiology of obesity-related hypoventilation or apnea.

### REFERENCES


# AUTHOR CONTRIBUTIONS

FY, JF, and HW acquired the data; FY, JF, and ZW analyzed and interpreted data; FY, XZ, and HY drafted the manuscript; FY, SW, and YZ were responsible for study concept and design; YZ and SW obtained research funding.

### ACKNOWLEDGMENTS

This study was supported by grants from the National Natural Science Foundation of China (31271223, 31671184) and from the Province Natural Science Foundation of Hebei (H2016206477). The experiments comply with the current laws of the country in which they were performed.

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphys. 2018.00249/full#supplementary-material

the inhibitory effect of leptin on food intake. Endocrinology 143, 239–246. doi: 10.1210/endo.143.1.8589


isoforms in the rat and human carotid body. Brain Res. 1385, 56–67. doi: 10.1016/j.brainres.2011.02.028


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Yuan, Wang, Feng, Wei, Yu, Zhang, Zhang and Wang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Divergent Mitochondrial Antioxidant Activities and Lung Alveolar Architecture in the Lungs of Rats and Mice at High Altitude

Alexandra Jochmans-Lemoine<sup>1</sup> , Susana Revollo<sup>1</sup> , Gabriella Villalpando<sup>2</sup> , Ibana Valverde<sup>2</sup> , Marcelino Gonzales <sup>2</sup> , Sofien Laouafa1,3, Jorge Soliz <sup>1</sup> and Vincent Joseph<sup>1</sup> \*

<sup>1</sup> Centre de Recherche de l'Institut Universitaire de Cardiologie et de Pneumologie de Québec, Université Laval, Quebec City, QC, Canada, <sup>2</sup> Instituto Boliviano de Biologia de Altura, Universidad Mayor de San Andrés, La Paz, Bolivia, <sup>3</sup> Centre National de la Recherche Scientifique, UMR 5023, Université Claude Bernard Lyon 1, Villeurbanne, France

### Edited by:

Rodrigo Del Rio, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Andrew T. Lovering, University of Oregon, United States Beth J. Allison, Hudson Institute of Medical Research, Australia

> \*Correspondence: Vincent Joseph vincent.joseph@fmed.ulaval.ca

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 08 December 2017 Accepted: 14 March 2018 Published: 04 April 2018

### Citation:

Jochmans-Lemoine A, Revollo S, Villalpando G, Valverde I, Gonzales M, Laouafa S, Soliz J and Joseph V (2018) Divergent Mitochondrial Antioxidant Activities and Lung Alveolar Architecture in the Lungs of Rats and Mice at High Altitude. Front. Physiol. 9:311. doi: 10.3389/fphys.2018.00311 Compared with mice, adult rats living at 3,600 m above sea level (SL—La Paz, Bolivia) have high hematocrit, signs of pulmonary hypertension, and low lung volume with reduced alveolar surface area. This phenotype is associated with chronic mountain sickness in humans living at high altitude (HA). We tested the hypothesis that this phenotype is associated with impaired gas exchange and oxidative stress in the lungs. We used rats and mice (3 months old) living at HA (La Paz) and SL (Quebec City, Canada) to measure arterial oxygen saturation under graded levels of hypoxia (by pulse oximetry), the alveolar surface area in lung slices and the activity of pro- (NADPH and xanthine oxidases—NOX and XO) and anti- (superoxide dismutase, and glutathione peroxidase—SOD and GPx) oxidant enzymes in cytosolic and mitochondrial lung protein extracts. HA rats have a lower arterial oxygen saturation and reduced alveolar surface area compared to HA mice and SL rats. Enzymatic activities (NOX, XO, SOD, and GPx) in the cytosol were similar between HA and SL animals, but SOD and GPx activities in the mitochondria were 2–3 times higher in HA vs. SL rats, and only marginally higher in HA mice vs. SL mice. Furthermore, the maximum activity of cytochrome oxidase-c (COX) measured in mitochondrial lung extracts was also 2 times higher in HA rats compared with SL rats, while there was only a small increase in HA mice vs. SL mice. Interestingly, compared with SL controls, alterations in lung morphology are not observed for young rats at HA (15 days after birth), and enzymatic activities are only slightly altered. These results suggest that rats living at HA have a gradual reduction of their alveolar surface area beyond the postnatal period. We can speculate that the elevation of SOD, GPx, and COX activities in the lung mitochondria are not sufficient to compensate for oxidative stress, leading to damage of the lung tissue in rats.

Keywords: oxidative stress, mitochondria, high altitude, lung, postnatal hypoxia

# INTRODUCTION

Mammals living at high altitude (HA) can deploy several strategies to counteract ambient hypoxia, and it is noteworthy that different types of genetic adaptations or physiological acclimatization to low ambient O<sup>2</sup> can be observed. Several studies have focused on rodent (Cheviron et al., 2012; Lau et al., 2017) or lagomorph (Li et al., 2009; Bai et al., 2015) species that are endemic to HA in order to understand the genetic mechanisms and signatures underlying HA adaptation. However, recent migration of lowland native species to HA regions offers unique opportunities to document adaptive strategies that can operate on a much shorter time-frame (Storz et al., 2007; Jochmans-Lemoine et al., 2016). Striking examples of this are found in the HA regions of South-America, where new animal species such as rats or mice have migrated over the past five centuries, following the European conquests (Guenet and Bonhomme, 2003; Storz et al., 2007). We recently described key differences between rats and mice that have been living in La Paz (Bolivia, 3,600 m above sea level—SL) under laboratory conditions for several generations (Jochmans-Lemoine et al., 2015). While mice maintained low hematocrit, high ventilation, and larger lung volumes, rats had high hematocrit, lower ventilation, signs of severe pulmonary hypertension, and lower arterial oxygen saturations when challenged with high or low levels of inspired O2. Overall, the phenotype observed in rats is strikingly similar to the pathological features of chronic mountain sickness, a syndrome of de-adaptation to hypoxia with chronic hypoventilation, high hematocrit levels, pulmonary hypertension, and reduced diffusion capacity within the lungs (Julian et al., 2015; Villafuerte and Corante, 2016). We also reported that it was possible to partially revert this phenotype by a simple exposure to an enriched O<sup>2</sup> environment during the first 2-weeks following birth (Lumbroso et al., 2012). HA rats exposed to inspired SL O<sup>2</sup> pressure during postnatal development had lower hematocrit, signs of reduced pulmonary pressure, and lungs with reduced airspaces suggesting improved alveolar development.

The large surface area of the lungs and continuous contact of the lung epithelium with environmental O<sup>2</sup> predispose to oxidative damage (Santus et al., 2014), and oxidative stress in the lung is a potent cause of injury and inflammation (Salama et al., 2014). Furthermore, it is well acknowledged that chronic hypoxia (Turrens, 2003) and HA (Dosek et al., 2007) lead to excessive reactive oxygen species (ROS) production and oxidative stress. A recent study showed that delayed alveolar formation following exposure to postnatal hyperoxia (used as a model of bronchopulmonary dysplasia) is linked to the mitochondrialdependent generation of ROS (Datta et al., 2015). Accordingly, we performed the present study to assess the hypothesis that the typical low alveolar surface area in the lungs of rats living at HA is associated with impaired gas exchange and oxidative stress in the lungs. We compared arterial oxygen saturation under graded levels of hypoxia, lung histology, and the activity of pro- and anti-oxidant enzymes in homogenized lung tissue from adult rats living at SL or HA. Similar experiments were performed in mice, which represented a control group of animals with successful physiological adaptations to HA. In addition, we tested the hypothesis that altered lung histology and oxidative stress are already present during postnatal development in rats at HA by using newborn rats (2 weeks-old) exposed to HA or SL. Finally, we tested the effects of postnatal hypoxia in rats living at SL, or postnatal O<sup>2</sup> enrichment in rats living at HA during postnatal days 4–14 (P4–P14—the timeframe corresponding to lung alveolar formation in rats—Burri et al., 1974) on arterial oxygen saturation during graded levels of hypoxia, alveolar morphology and oxidative stress balance.

### MATERIALS AND METHODS

### Animals and Experimental Groups Animals

This study was carried out in accordance with the recommendations of the Canadian Council of Animal Care. The protocols were approved by the Committee on Animal Care and Use for Laval University in Canada, or by the scientific committee of the Instituto Boliviano de Biologa de Altura (IBBA) in Bolivia. In both countries, animals were housed under standard conditions, had access to food and water ad-libitum, and were exposed to a 12:12 h light/dark cycle.

### **In Canada**

Adult Sprague Dawley male rats and FVB mice around 2 months old were ordered from Charles-Rivers (Charles-River— St-Constant, Québec), and left undisturbed at least for 7 days before being used. Newborn rats were obtained from 8 Sprague Dawley female rats (Charles-River—St-Constant, Québec) that were housed with males for mating for at least 7 consecutive days. Once pregnancy was confirmed by weight gain, the females were isolated in a separate cage. At birth, all litters were culled to 12 pups with an equal number of males and females.

### **In Bolivia**

All rats were obtained from the Instituto Boliviano de Biologia de Altura (IBBA, La Paz, Bolivia, 3,600 m). These are Sprague-Dawley rats that were originally imported from France (IFFA-CREDO) in 1992, and continuously bred at the IBBA. Males and females were left together until pregnancy was confirmed, and at birth all pups were kept alive (litters generally have less than 12 pups). Adult mice were obtained from the Instituto Nacional de Laboratorios de Salud (INLASA, La Paz, Bolivia). These mice are descended from a lineage of animals that were originally imported from France (IFFA-CREDO) 20–25 years ago and have 73.11% homology with the FVB strain, suggesting that they are either a mix of FVB with 2 other strains or an outbred strain (Jochmans-Lemoine et al., 2015).

### Exposure to Postnatal Hypoxia in Canada

Four days after birth, animals were placed under normobaric hypoxia in a 50 L plexiglass chamber, connected to an Oxycycler (A84XOV, BioSpherix, Redfield, NY, USA). On the 1st day of exposure, the O<sup>2</sup> level in the chamber was progressively decreased to 13.5% (1% O<sup>2</sup> every 20 min) and kept constant at this level for 10 consecutive days (until postnatal day 14). This corresponds to an inspired PO<sup>2</sup> around 100 mmHg, which is the inspired PO<sup>2</sup> found in La Paz. To avoid humidity and CO<sup>2</sup> accumulation, Drierite (anhydrous calcium sulfate; Hammond Drierite, Xenia, OH, USA) and Amsorb Plus (calcium hydroxide; Armstrong Medical, Coleraine, North Ireland) were placed inside the chamber. Control animals were kept in the same room and left undisturbed (except for normal cage cleaning—once a week).

At postnatal day 14 (P14), the O<sup>2</sup> level inside the chamber was progressively returned to 21% (at the rate of 1% O<sup>2</sup> every 15 min), the animals were sacrificed 24 h later for lung tissue sampling.

### Exposure to Postnatal O<sup>2</sup> Enrichment in Bolivia

Four days after birth, animals were placed under hypobaric normoxia in a 50 liter plexiglass chamber or left in ambient room air. The O<sup>2</sup> level was continuously measured in the chamber, and maintained at ≈ 32% O<sup>2</sup> (corresponding to an inspired PO<sup>2</sup> of 160 mmHg, the normal SL inspired PO2) by a continuous flow from a calibrated gas tank. The CO<sup>2</sup> level was controlled twice daily with a dedicated CO<sup>2</sup> sensor and never exceeded 0.3%. The air inside the chamber was continuously mixed with a small fan, and there was no apparent humidity inside the chamber. At P14, the chamber was opened, and the animals were returned to ambient air. These animals were sacrificed 24 h later for lung tissue sampling.

### Pulse Oximetry Recordings

Pulse oximetry recordings were performed in sealed chambers (adapted to the size of the animals and constantly flushed with fresh room air). In Canada, animals were equipped with a neck collar (Mouse OX <sup>R</sup> STARR Life Sciences Corp, USA) and in Bolivia, animals were equipped with a limb sensor (MouseSTAT—Kent Scientific, Torrington, CT, USA) allowing recordings of pulse oximetry capillary O<sup>2</sup> saturation (SpO2). We have verified that at SL, the two systems gave similar SpO<sup>2</sup> values under graded hypoxia. Each animal was placed in the chamber for a period of acclimatization (10–15 min), then the inflowing tube was switched to a nitrogen gas line calibrated to obtain the desired O2%. In Canada, animals were subsequently exposed to 18, 15, 12, and 9% O2, each for 10 min and in Bolivia, animals were exposed to 32% O2–corresponding to the inspired air at SL followed by 18, 15, 12% O<sup>2</sup> for 10 min each. In the present work, we report values that have been collected under comparable pressures of inspired O<sup>2</sup> at 160 mmHg (SL room and 32% O<sup>2</sup> at HA) and 90 mmHg (12% O<sup>2</sup> at SL and 18% O<sup>2</sup> at HA).

### Lung Dissection, Histology, and Morphology

In both countries, we used 2 males from each litter of newborn rats (4 litters per group) and 4 adult males of each species (rats and mice). Animals were deeply anesthetized and perfused through the heart with ice-cold PBS solution. Next, a catheter was fixed in the trachea, the lungs were inflated with 4% PFA for 30 min at a constant pressure of 24 cm H2O, then the trachea was ligated and the lungs dissected. The total volume of the inflated lungs was measured by liquid displacement. Dissected lungs were kept in 4% PFA for 24 h at room temperature. The next day, the lungs were separated into left and right lungs (for newborn rats and adult mice) or into 5 lobes (for adult rats), which were automatically embedded in paraffin (in Canada, using the Tissue-Tek VIP, Miles scientific) or manually (in Bolivia as previously described; Jochmans-Lemoine et al., 2015). Twentyfour hours later, the samples were included in paraffin and stored until they were processed for lung histology as previously described (Jochmans-Lemoine et al., 2015). We have verified the two embedding approaches gave similar morphological results on a sample of SL lung for both species. In another 4 newborn male rats and 6 adult rats and mice, the lungs were immediately removed from the chest after the cardiac PBS perfusion, weighed, frozen on dry ice and stored until they were processed to determine enzymatic activity.

### Lung Morphology

After deparaffinization and coloration (Hematoxylin), images of each lung section were captured (magnification: x100; see Jochmans-Lemoine et al., 2015, for details). We analyzed 3 slides per animal, and randomly selected 3 non-overlapping images from each slide (we used a total of 4 adult male rats and mice in each group, and 8 P15 male rats in each group). The Mean Linear Intercept (Lm) was determined by overlapping a grid of 20 horizontal and vertical lines (189µm each) on each image and by counting the number of intersections with alveolar walls (Hsia et al., 2010). When a line crossed a vessel wall rather than an alveolar wall it was counted as 0.5 intersections. Lm was calculated by using the following equation: Lm = (N.d)/m, with N being the number of lines (20), d the length of each line (189µm), and m the number of intersections with alveolar walls. From Lm values, we calculated the relative alveolar surface area as S (m<sup>2</sup> /cm<sup>3</sup> ) = 4 V/Lm, with V being the volume of one image (Hsia et al., 2010). An estimation of the total alveolar surface area was calculated as the product of the relative alveolar surface area and lung volume.

To compare lung morphology between adult rats vs. mice, and between P15 rats living at SL vs. HA, we used allometric scaling parameters as previously described (Jochmans-Lemoine et al., 2015), with the corresponding units: lung weight (g/g), lung volume (ml/g), and relative alveolar surface area (m<sup>2</sup> /cm<sup>3</sup> /g−0.13).

### Biochemical Analysis

### Protein Extraction

Protein was extracted from frozen lungs (50–100 mg) as follows: after homogenization in 1 ml of PBS-EDTA (0.5 mM), the samples were centrifuged for 4 min at 1,500 g and then for 10 min at 12,000 g (all centrifugations at 4◦C). Four aliquots (200 µl) of the cytosolic fraction were recovered from the second centrifugation and stored until further analysis. The remaining pellet was homogenized in 1.5 ml of mitochondrial isolation buffer (250 mM sucrose, 1 mM EGTA and 20 mM Tris-base pH 7.3) and centrifuged for 10 min at 1,500 g. We collected the supernatant in a new tube and centrifuged it once more at 9,000 g for 11 min. The supernatant was eliminated and the pellet (containing the mitochondrial protein fraction) homogenized with 300 µl of the mitochondrial isolation buffer and stored at −80◦C until further analysis. The concentration of proteins in each fraction was determined using the BCA protein assay kit (ThermoFisher Scientific, catalog #23225) and all subsequent measurements were normalized to protein concentration.

### Xanthine Oxidase (XO) Activity

XO activity was assessed in the cytosolic protein fraction using a cocktail containing nitroblue tetrazolium (NTB−2.2 mM in water), Tris-HCl pH 8 (2.8 mM), NaCN 1 mM (to inhibit the degradation of superoxide anions by cytosolic SODCu−Zn), and diethylene-triamine-penta-acetic acid (DTPA−1.3 mM in Tris-HCl) combined with a fresh solution of hypoxanthine (500µM per well in Tris-HCl). Fifty microliter of sample, 200 µl of cocktail and 30 µl of hypoxanthine solution was added to each well, and the plate was gently shaken for 40 min at room temperature. The wavelength absorbance of the complex (formazan blue) formed by NTB and the superoxide anion produced by XO in the sample was read at 560 nm every 5 min for 1 h; XO activity corresponded to the slope of the formation of formazan blue by time.

### NADPH Oxidase (NOX) Activity

NOX activity was assessed in the cytosolic protein fraction using the same cocktail solution as XO and a fresh solution of NADPH (1 mM). Twenty microliter of sample, 250 µl of cocktail, and 30 µl (100 µM/well) of NADPH were added to each well, the plate was shaken for 2 min at room temperature and absorbance was read at 560 nm every 50 s for 10 min. NOX activity corresponded to the slope of the formation of formazan blue by time.

### Cytosolic and Mitochondrial Superoxide Dismutase (SODCu−Zn and SODMn)

SOD activity was measured in the cytosolic and mitochondrial protein fractions. SODCu−Zn activity was determined in the cytosolic protein fraction by the degree of inhibition of the reaction between the reactive O<sup>2</sup> anion superoxide (O2· −: produced by a hypoxanthine-xanthine oxidase system) and NTB. We used the same cocktail described for XO and NOX activities combined with hypoxanthine (0.19 mM) and prepared a fresh solution of xanthine oxidase (1.02 units/ml). We added 20 µl of sample, 250 µl of cocktail and 20 µl of XO to each well, and mixed the plate for 4–5 s at room temperature. The absorbance was quickly read at 450 nm every 50 s for 5 min. SODMn activity was similarly determined in the mitochondrial protein fraction. For this assay, 4 wells containing 20 µl of PBS 1X (rather than samples) were used as blanks, and 1 mM of NaCN was added in all wells to inhibit SODCu−Zn. SODMn activity corresponded to the difference between the slopes of the formation of formazan blue by time between the blank and each sample.

### Glutathione Peroxidase (GPx)

GPx activity was determined in the mitochondrial protein fraction as the rate of oxidation of NADPH to NADP+ in a cocktail solution containing glutathione reductase, NADPH (1.7 mM) and reduced glutathione (1.6 mM in water) using H2O<sup>2</sup> (0.036% in water) as a substrate (Paglia and Valentine, 1967). We added 20 µl of sample, 200 µl of PBS 1X, 30 µl of cocktail solution and 30 µl of H2O<sup>2</sup> to each well and mixed the plate for 4–5 s at room temperature. The absorbance was quickly read at 340 nm every 50 s for 5 min. GPx activity was measured as the slope of the NADPH extinction by time.

### Cytochrome c Oxidase Assay

Complex IV is the last enzyme of the mitochondrial electron transport chain and catalyzes oxidation of the reduced cytochrome c by O2. We measured the maximum activity of cytochrome c oxidase (COX—complex IV of the mitochondrial respiratory chain) by measuring O<sup>2</sup> consumption in the mitochondrial protein fraction using a high-resolution respirometer system (Oroboros oxygraph-2k). We used 50–150 µl of sample, with 1 µl of Antimycin A, 5 µl of ascorbate and 5 µl of Tétraméthyl-paraphénylènediamine (TMPD) in 2 ml of mitochondrial respiratory buffer (0.5 mM EGTA, 3 mM MgCl2, 60 mM potassium lactobionate, 10 mM KH2PO4, 20 mM Hepes, 110 mM sucrose, 1 g/l BSA). The maximal activity of COX was read when O<sup>2</sup> consumption rate was stable (typically a few minutes after starting the recording).

### Statistical Analysis

All values are reported as mean ± s.e.m., and the significant P-value was set at 0.05. We used GraphPad Prism 6.0c for all analysis. We analyzed arterial oxygen saturation at 160 mmHg and 90 mmHg, lung morphology and the protein assays using a two-way ANOVA with species and altitude as grouping variables. When significant effects or a significant interaction between species and PiO<sup>2</sup> appeared, a post-hoc analysis was performed (Fisher's LSD).

P-values are reported in the figures with the following general pattern: <sup>∗</sup> , ∗∗ , ∗∗∗, and ∗∗∗∗ are used to report P< 0.05, 0.01, 0.001, and 0.0001, respectively. Main ANOVA results are reported in the text whereas Ficher's LSD results are found in the figures.

# RESULTS

### Lung Architecture Is Preserved at HA in Adult Mice but Impaired in Rats

Representative images of lung slices in adult rats and mice living at SL and HA are presented in **Figure 1**, along with the results of the morphometric analysis. Adult mice living at HA have significantly increased (over 3-fold) lung volume compared with their SL counterparts, and this is not observed in rats (P-value for altitude <0.0001, P-value for altitude x species <0.0001). The mean linear intercept (Lm) was significantly lower in mice living at HA compared with SL and tended to be higher in rats living at HA compared with SL (P-value for species x altitude = 0.01—post-hoc P-value for altitude in rats = 0.067). Comparisons between HA and SL revealed that rats had a reduced mass-corrected relative alveolar surface area (in m2 /cm<sup>3</sup> /g−0.13) at HA; but when compared with mice at HA, the mass-corrected relative alveolar surface area was lower in rats (P-value for species x altitude = 0.02). These results suggest that the lung morphology of HA rats does not facilitate improved gas exchange.

values are mean + s.e.m. Number of animals in each group: n = 4. Scale bars: 50µm. \*P < 0.05, and \*\*\*\*P < 0.0001 HA vs. SL. ◦P < 0.05, and ◦◦◦◦P < 0.0001 mice vs. rats.

### HA Mice Maintain a Higher Arterial Oxygen Saturation in Response to Graded Hypoxia

In rats, SpO<sup>2</sup> was lower at HA than at SL for normoxic O<sup>2</sup> levels (160 mmHg), but was similar between the two altitudes for a PiO<sup>2</sup> of 90 mmHg (**Figure 2**). In mice at HA, SpO<sup>2</sup> was higher than HA rats under PiO2's of 160 and 90 mmHg. At 90 mmHg, HA mice had a significantly higher SpO<sup>2</sup> than SL mice.

# Mitochondrial Antioxidant Activities Are Increased in the Lungs of Adult Rats at HA

The activities of NOX, XO, GPx, and SODCu−Zn in the cytosolic protein fraction from the lungs of adult rats and mice are similar between SL and HA (**Figure 3**). In the mitochondrial protein fraction, there was increased SODMn and GPx activity (around 2 fold) in HA adult rats compared with SL. The activity of Complex IV of the mitochondrial respiratory chain (or cytochrome c oxidase—COX) was also increased in rats living at HA compared with SL. In mice, the activity of mitochondrial SODMn and COX were slightly increased at HA compared with SL. Overall this suggests that important changes in mitochondrial oxidative reactions occur in the lungs of rats living at HA compared with SL control animals, while this does not occur in mice.

### The Alveolar Surface Area in Young Rats Is Impaired by Postnatal Hypoxia at SL, but Preserved at HA

Representative images of lung slices along with the results of the morphometric analysis in SL and HA P15 rats exposed to room air, postnatal hypoxia (at SL), or postnatal O<sup>2</sup> enrichment (at HA) are presented in **Figure 4**. At P15, SL rats exposed to

postnatal hypoxia have a similar lung volume to control rats, but a higher value of Lm and reduced mass-corrected relative alveolar surface area. At HA, P15 rats have higher lung volumes compared with SL P15 rats, and a similar Lm and alveolar surface area. At HA, postnatal O<sup>2</sup> enrichment reduced Lm and increased the mass-corrected relative alveolar surface area.

### Arterial Oxygen Saturation in Young Rats Is Impaired by Postnatal Exposure to Hypoxia at SL

At SL, postnatal exposure to hypoxia significantly reduced SpO<sup>2</sup> in young rats during a PiO<sup>2</sup> of 90 mmHg (**Figure 5**). Interestingly, when young control rats maintained at HA were exposed to 90 mmHg, the observed SpO<sup>2</sup> was lower than for control SL rats, but was higher than SL rats exposed to postnatal hypoxia.

### Cytosolic SOD Activity in the Lungs Is Reduced by Postnatal Hypoxia at SL, but Increased by Postnatal O<sup>2</sup> Enrichment at HA

NOX and XO enzymatic activities were lower at HA than at SL (P-value for altitude = 0.0003 and 0.038 respectively—**Figure 6**), but there was no effect resulting from exposure to postnatal hypoxia (at SL), or O<sup>2</sup> enrichment (at HA). At SL, exposure to postnatal hypoxia significantly reduced SOD activity, while at HA postnatal O<sup>2</sup> enrichment had the opposite effect (P-value for altitude × O<sup>2</sup> level = 0.002). These results suggest reduced production of ROS by NADPH, and greater anti-oxidant capacity by GPx in P15 control HA rats compared with SL controls. In the mitochondrial protein fraction, there was no change in enzyme activities (**Figure 6**) due to postnatal O<sup>2</sup> level or altitude.

# DISCUSSION

In the present study, we reported that compared to SL controls, in mice, exposure to HA for several generations led to larger lungs with a higher relative alveolar surface area. The opposite response was observed in rats: at HA, lung volume was not increased compared with SL, and the relative alveolar surface area was reduced. In line with this, mice exposed to low O<sup>2</sup> levels at HA had higher SpO<sup>2</sup> values than rats, but remarkably HA mice also had higher SpO<sup>2</sup> values than SL mice, indicating highly efficient gas exchange functions. Because previous studies demonstrated that chronic hypoxia (Turrens, 2003) and HA (Dosek et al., 2007) lead to excessive ROS production and oxidative stress, a wellknown cause of lung injury and inflammation (Salama et al., 2014), we hypothesized that such damage could occur in rats. We measured the activities of pro- and anti-oxidant enzymes in cytosolic and mitochondrial protein extracts in the lungs. Our data showed that HA had no effect on the activity of pro- or antioxidant enzymes in the cytosol. However, antioxidant enzyme activities in the mitochondrial protein fraction were about 2 times higher in HA rats compared with SL rats, whereas in mice there was a smaller increase only for the SODMn and COX activities (1.25 times higher).

Several studies have shown that postnatal hypoxia reduced lung alveolarization in rats (Blanco et al., 1991; Truog et al., 2008); we thus tested the hypothesis that at HA, young rats (P15) would present early signs of altered lung alveolarization. Surprisingly, this was not the case as our data showed that at HA, young rats had lung volumes, Lm, and relative alveolar surface areas similar to SL young rats. Interestingly, SpO<sup>2</sup> values during a PiO<sup>2</sup> of 90 mmHg were higher in control rats at HA than in SL rats exposed to postnatal hypoxia, therefore indicating more efficient gas exchange function at HA in the former group. Furthermore, our data indicated that O<sup>2</sup> enrichment between postnatal days 4– 14 at HA can further increase the alveolar surface area at P15, but this was not associated with improved SpO<sup>2</sup> values. In previous studies, however, we showed that at HA, O<sup>2</sup> enrichment during postnatal development has long-term effects including improved lung alveolar architecture, reduced pulmonary hypertension and lower hematocrit in adult rats (Lumbroso et al., 2012), likely indicating long-term improvement in arterial O<sup>2</sup> saturation. Young rats at HA had lower activity of NADPH and XO (two

major sources of cytosolic ROS production) compared with SL rats, while O<sup>2</sup> enrichment during postnatal development increased SOD activity in the cytosol, and hypoxic exposure during the same period at SL had the opposite effect.

Overall, these findings suggest that in adult HA rats, the high activities of SOD, GPx, and COX could be a strategy to buffer excessive mitochondrial ROS production. In contrast, antioxidant enzyme activities are increased in young HA rats, and pro-oxidant enzymes are concurrently decreased as a strategy to maintain ROS equilibrium.

# Alterations of Lung Morphology Are Present in Adult Rats at HA Despite High Mitochondrial Anti-oxidant Activity

Hypoxic exposure induces ROS production in the mitochondria (Jezek and Hlavata, 2005), and excessive ROS production induces pro-inflammatory responses and has cytotoxic effects including lipid membrane peroxidation, DNA damage, oxidation of amino acids and oxidative cleavage of peptide bonds in proteins. A previous study showed that iron supplementation induced lung

injuries in rats at HA through excessive production of ROS and induction of inflammatory responses (Salama et al., 2014). Moreover, a role for oxidative stress has been proposed in the development of chronic obstructive lung diseases (van Eeden and Sin, 2013). Since adult rats at HA had elevated activity of mitochondrial antioxidant enzymes, we can hypothesize that there was elevated ROS production in the mitochondria of lung cells in rats, but not mice. Indeed, even if ROS can be produced in cells by diverse mechanisms and in several compartments, up to 90% are produced at complexes I and III of the electron transfer chain in mitochondria (Adam-Vizi, 2005). Due to the fact that ROS play a key role in lung diseases, it is possible that the elevated anti-oxidant activity we observed in the lungs of rats living at HA is not sufficient to completely protect against oxidative stress, and this could in part explain why HA rats had low alveolar surface areas. Using the same line of evidence, this implies that mice at HA were not subjected to excessive mitochondrial oxidative stress in the lung, and while we cannot infer a direct causal link, it is striking to observe that in parallel mice had increased relative alveolar surface area.

Cytochrome-c oxidase (COX—complex IV of the electron transfer chain) catalyzes the final step of the electron transport chain and the production of H2O from O2, thus accounting for mitochondrial O<sup>2</sup> consumption. COX activity is also important for mitochondrial ROS production, and it is well established that low COX activity, which increases mitochondrial ROS production by complexes I and III of the electron transfer chain, is involved in several pathological conditions (Srinivasan and Avadhani, 2012). Our results showed higher maximum COX activity in rats living at HA than at SL, while this effect

of altitude residence was minimal in mice. This is surprising because acute hypoxic exposure reduces COX activity, directly contributing to increasing ROS production in hypoxia (Prabu et al., 2006; Fukuda et al., 2007; Srinivasan and Avadhani, 2012). In the cardiac myocytes of bar-headed geese (a bird species adapted to flight at extreme altitude), maximum COX activity is reduced, but the affinity of COX for reduced cytochrome-c is higher (Scott et al., 2011). A higher affinity for substrates could ensure higher COX activity at a low or normal concentration of reduced cytochrome-c, thus allowing lower ROS production in the electron transport chain (Scott et al., 2011; Cheviron and Brumfield, 2012). Similarly, increased activity of COX was associated with hypoxia tolerance in Tibetan locusts (Zhang et al., 2013). We can, therefore, hypothesize that in HA rats that have been bred at 3,600 m above SL for 25 years, the increased COX activity could contribute to reduced mitochondrial ROS production, which would be in line with the increased SODMn and mitochondrial GPx activities. However, these increased enzymatic activities are apparently not sufficient to buffer the deleterious consequences of oxidative stress, leading to pathological responses such as the altered architecture observed in the lungs of HA rats.

A summary of these findings is presented in **Figure 7**, incorporating a hypothetical mechanism underlying our results. Interestingly, we previously reported that, compared with HA mice, HA rats have lower whole body O<sup>2</sup> consumption and higher glycolytic metabolism as reported by an elevated respiratory exchange ratio (Jochmans-Lemoine et al., 2015). Translated at the cellular level, it is possible to speculate that mitochondrial O<sup>2</sup> consumption and ATP production are lower in HA rats compared with mice. A state of low mitochondrial respiration increases the production of superoxide anion (O2•-) by complexes I and III of the mitochondrial electron transport chain (Murphy, 2009), and in HA rats this can be partially overcome by the high activities of mitochondrial SODMn and GPx that reduce O2•- to H2O<sup>2</sup> and water. High levels of H2O2, however, would lead to the production of hydroxyl radicals (HO•) through the Fenton reaction (Thomas et al., 2009), and this highly reactive ROS could interact with biomolecules to cause oxidative damage (Halliwell, 2001). As alveolar lung morphology is altered in rats at HA, we hypothesize that the elevated antioxidant activity is not sufficient to abolish oxidative stress.

### Alteration of Lung Morphology and Elevated Mitochondrial Anti-oxidant Activity Are Not Present in Young Rats at HA

The fact that HA rats at P15 had a similar Lm and relative alveolar surface area as SL rats was a surprising finding, and was in contrast with the effect of postnatal hypoxia in rats at SL that induced a delay in alveolar formation (the latter results being in line with previous data; Blanco et al., 1991; Truog et al., 2008). In addition, when young rats were exposed to low O2, SpO<sup>2</sup> was higher in HA rats than in SL rats that had been exposed to postnatal hypoxia. Therefore, our results suggest that young HA rats were able to acquire mechanisms to maintain alveolar development in the lungs similar to that of SL rats, and this might have beneficial long-term effects. Indeed, we reported that exposure of HA rats to SL O<sup>2</sup> levels between postnatal days 4– 14 induces a persistent increase in alveolar surface area for adults (Lumbroso et al., 2012), and we have unpublished observations indicating that the alterations induced by postnatal hypoxia at SL persist in adult rats.

We previously reported that the physiological responses observed in rats at HA could be correlated with symptoms reported in chronic mountain sickness (Lumbroso et al., 2012; Jochmans-Lemoine et al., 2015). Interestingly, it has been suggested that chronic mountain sickness is linked to altered perinatal oxygenation (Moore et al., 2007; Julian et al., 2015), and that excessive erythrocytosis (a preclinical state of

chronic mountain sickness) at HA is associated with elevated oxidative stress (Julian et al., 2013). Contrasting with the results for adult rats, in young HA rats the cytosolic activity of the pro-oxidant enzymes (NOX and XO) is decreased compared with SL rats. These findings are interesting, because chronic pulmonary hypertension in newborn rats exposed to

hypoxia has been associated with ROS generated by xanthine oxidase in the lungs (Jankov et al., 2008). Furthermore, in a study where young mice were exposed to hyperoxia for 72 h between postnatal days 0–3, developmental defects in the lungs were observed at P15 and linked to an exaggerated mitochondrial oxidative stress response and amplification of ROS signaling via NOX-1 (a member of the NOX family of NADPH oxidases; Bedard and Krause, 2007) activity (Datta et al., 2015). Thus, it might be possible that the reduced activity of NOX is in part responsible for the maintenance of lung architecture observed in HA young rats compared with those at SL.

### Study Limitations

One of the potential limitations of these studies is the particular nature of the HA species and the fact that they have been genetically isolated for several generations. Aside from the overwhelming influence of constant hypoxic exposure, we cannot completely rule out that these animals have been subjected to genetic drift resulting in a specific phenotype. However, because Charles-River maintains a program to avoid genetic drift between their different stocks of SD rats, we can assume that the original breeders of the HA population and the SD rats used in Quebec are similar. A second limitation is that newborn rats at SL or HA were exposed to room air for 24 h before sacrifice and tissue harvesting. While it is highly unlikely that this short time frame alters lung morphology, it is still possible that it might affect the biochemical analysis, and could contribute to discrepancies in some of our results. As such, the results of the biochemical analysis obtained when animals were exposed to hypoxia at SL or enriched oxygen at HA should be considered cautiously. However, it is striking that there are only limited differences for NOX and XO between newborn rats at SL and HA, while differences are much more substantial in adults. Finally, one should keep in mind that, unfortunately, samples from female rats were not included in these studies. However, despite the fact that ovarian steroids could have potent anti-oxidant effects in mitochondria (Laouafa et al., 2017), we have not detected sexspecific alterations in lung morphology in HA rats (Jochmans-Lemoine et al., 2015).

# CONCLUSION

We conclude from this study that rats living at HA present an unusually elevated mitochondrial anti-oxidant capacity in the lungs, possibly as a compensatory response to high mitochondrial ROS production. Rats also presented morphological alterations in the lung tissue and a low alveolar surface area—a phenotype known to be induced by high oxidative stress in the lungs. Therefore, our data suggest that the high activities of antioxidant enzymes might not be efficient to adequately protect the lungs. Mice at HA had normal anti-oxidant enzymatic activities in the lungs and were able to develop a higher lung volume, alveolar surface area, and high arterial O<sup>2</sup> saturation at HA.

Since mitochondrial anti-oxidant enzyme activities and alveolar surface area were normal in young rats at HA, alteration of lung alveolar morphology and oxidative stress balance occur beyond the second postnatal week in rats living at HA. Finally, the phenotype reported in HA rats is strikingly similar to what is observed in patients affected by chronic mountain sickness (Julian et al., 2015; Villafuerte and Corante, 2016). The present findings might thus have clinical relevance for HA residents and provide evidence that targeting mitochondrial ROS production with selective mitochondrial anti-oxidant agents could be an interesting therapeutic opportunity.

### AUTHOR CONTRIBUTIONS

AJ-L, GV, IV, MG: Contributed to sampling and processing tissue samples in Bolivia. AJ-L, SR, SL: Contributed to sampling and processing tissue samples in Canada. AJ-L, JS, VJ: Contributed to initial study design. AJ-L, VJ: Analyzed data, drafted the figures

### REFERENCES


and manuscript, wrote and edited the manuscript. All authors: Commented early drafts of the paper and data presentation, read and approved the final version.

### ACKNOWLEDGMENTS

This study was funded by the Natural Sciences and Engineering Research Council of Canada (VJ, JS: grant # RGPGP-2014- 00083). AJ-L has been supported by a training grant in respiratory physiology from Réseau en Santé Réspiratoire (Fonds de Recherche du Québec—Santé, and Canadian Institute for Health Research). The authors acknowledge Dr. Tara Adele Janes for careful editorial revision of the paper.

neonatal rats. Am. J. Physiol. Lung Cell. Mol. Physiol. 294, L233–L245. doi: 10.1152/ajplung.00166.2007


by morphometry and directed microarray. Pediatr. Res. 64, 56–62. doi: 10.1203/PDR.0b013e31817289f2


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Jochmans-Lemoine, Revollo, Villalpando, Valverde, Gonzales, Laouafa, Soliz and Joseph. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Revisiting the Role of TRP, Orai, and ASIC Channels in the Pulmonary Arterial Response to Hypoxia

Roberto V. Reyes1,2 \*, Sebastián Castillo-Galán<sup>1</sup> , Ismael Hernandez<sup>1</sup> , Emilio A. Herrera1,2 , Germán Ebensperger1,2 and Aníbal J. Llanos1,2

<sup>1</sup> Unidad de Fisiología y Fisiopatología Perinatal, Programa de Fisiopatología, Instituto de Ciencias Biomédicas, Facultad de Medicina, Universidad de Chile, Santiago, Chile, <sup>2</sup> International Center for Andean Studies, Universidad de Chile, Santiago, Chile

The pulmonary arteries are exquisitely responsive to oxygen changes. They rapidly and proportionally contract as arterial PO<sup>2</sup> decrease, and they relax as arterial PO<sup>2</sup> is re-established. The hypoxic pulmonary vasoconstriction (HPV) is intrinsic since it does not require neural or endocrine factors, as evidenced in isolated vessels. On the other hand, pulmonary arteries also respond to sustained hypoxia with structural and functional remodeling, involving growth of smooth muscle medial layer and later recruitment of adventitial fibroblasts, secreted mitogens from endothelium and changes in the response to vasoconstrictor and vasodilator stimuli. Hypoxic pulmonary arterial vasoconstriction and remodeling are relevant biological responses both under physiological and pathological conditions, to explain matching between ventilation and perfusion, fetal to neonatal transition of pulmonary circulation and pulmonary artery overconstriction and thickening in pulmonary hypertension. Store operated channels (SOC) and receptor operated channels (ROC) are plasma membrane cationic channels that mediate calcium influx in response to depletion of internal calcium stores or receptor activation, respectively. They are involved in both HPV and pathological remodeling since their pharmacological blockade or genetic suppression of several of the Stim, Orai, TRP, or ASIC proteins in SOC or ROC complexes attenuate the calcium increase, the tension development, the pulmonary artery smooth muscle proliferation, and pulmonary arterial hypertension. In this Mini Review, we discussed the evidence obtained in in vivo animal models, at the level of isolated organ or cells of pulmonary arteries, and we identified and discussed the questions for future research needed to validate these signaling complexes as targets against pulmonary hypertension.

Keywords: store operated channels, receptor operated channels, hypoxia, hypoxic pulmonary vasoconstriction, pulmonary arterial remodeling, pulmonary hypertension

### INTRODUCTION

The pulmonary arteries have distinctive properties all along the individual's life, compared to systemic arteries. During gestation, the gas exchange is carried out by the placenta, the pulmonary vascular resistance (PVR) is high and the pulmonary blood flow is low, receiving less than 10–20% of the combined fetal cardiac output (CO). In fact, fetal pulmonary arteries have a narrow lumen and a thick medial layer with immature smooth muscle cells, increased synthesis and

### Edited by:

Rodrigo Del Rio, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Alexander Dietrich, Ludwig-Maximilians-Universität München, Germany Thomas C. Resta, University of New Mexico, United States

\*Correspondence:

Roberto V. Reyes vicreyes@med.uchile.cl; virreyc@gmail.com

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 30 January 2018 Accepted: 16 April 2018 Published: 07 May 2018

### Citation:

Reyes RV, Castillo-Galán S, Hernandez I, Herrera EA, Ebensperger G and Llanos AJ (2018) Revisiting the Role of TRP, Orai, and ASIC Channels in the Pulmonary Arterial Response to Hypoxia. Front. Physiol. 9:486. doi: 10.3389/fphys.2018.00486

**127**

signaling of vasoconstrictors as well as low oxygen tension (PaO2) in the fetal arterial blood, among other factors, which contribute to the pulmonary high resistance, low flow state. At birth, PVR quickly decreases and the pulmonary blood flow increases ∼10 times to accommodate the totality of the CO, allowing the lungs to replace the placenta as gas exchanger at the first minutes of postnatal life (Rudolph, 1979; Suresh and Shimoda, 2016). Increases of PaO<sup>2</sup> and signaling mechanisms favoring the vasodilation contribute to this fetal to neonatal physiological transition. The pulmonary arteries also undergo a quick and dramatic structural transition at birth: the pulmonary artery smooth muscle cells (PASMC) of the medial layer flattens and their cytoskeleton reorganizes deriving in wall thinning and luminal enlargement. These changes finally allow the establishment of postnatal pulmonary circulation with thin pulmonary arterial wall and low mean pulmonary arterial pressure (mPAP ∼8–20 mmHg at rest for humans) compared to systemic circulation (Peñaloza and Arias-Stella, 2007; Gao and Raj, 2010; Gao and Raj, 2011; Nayr and Lakshminrusimha, 2014). A unique feature of the pulmonary arteries is their intrinsic sensitivity to hypoxia, possessing a vasoconstrictor response to acute hypoxia called "hypoxic pulmonary vasoconstriction" (HPV), which is already present in fetal pulmonary arteries and continues to exist in every life stages. HPV is a rapid and reversible contraction of pulmonary arteries, whose intensity is proportional to the degree of hypoxia and it occurs independently of neural or humoral factors (Sylvester et al., 2012). HPV contributes to the physiologically high PVR state during fetal life, while it allows to optimize the ratio between ventilation and perfusion during the postnatal life (Dunham-Snary et al., 2017; Hussain et al., 2017). Pulmonary arteries also mount a maladaptive response to chronic hypoxia, known as pathological pulmonary arterial remodeling, characterized by the hyperplasic and hypertrophic thickening of medial layer and later thickening of adventitia and intima, with increased extracellular matrix deposition. An example of the later is the exposure to chronic hypoxia during pregnancy, early birth or adult life, resulting in pulmonary hypertension, a condition characterized by elevated pulmonary arterial mPAP and PVR, pathological pulmonary artery remodeling, often complicated with right ventricular hypertrophy and cardiac failure (Herrera et al., 2015; Suresh and Shimoda, 2016).

## STORE OPERATED CHANNELS, RECEPTOR OPERATED CHANNELS, STRUCTURE, FUNCTION, AND PHARMACOLOGY

Cytosolic free Ca2<sup>+</sup> [(Ca2+)i] is pivotal for PASMC contraction, differentiation and proliferation, as well as for synthesis of vasoactive compounds from the pulmonary artery endothelial cells (PAEC). The [Ca2+]i may increases through its release from sarcoplasmic reticulum (SR) stores mediated by ryanodine receptors (RyR) or Inositol triphosphate receptors (IP3R), or through calcium influx mediated by voltage operated calcium channels (VOC), receptor operated calcium channels (ROC), or store operated calcium channels (SOC) from plasma membrane (Kuhr et al., 2012).

The SOC is a functional definition for plasma membrane cationic channels that are physiologically activated as calcium stores of SR are gradually depleted by agonists coupled to IP3R or RyR, ensuring a mechanism to refill intracellular Ca2<sup>+</sup> stores and enabling to maintain long term [Ca2+]i signals. The most useful strategy to evaluate SOC function is to measure [Ca2+]i in isolated cells pre-loaded with a fluorescent Ca2+-sensitive dye like FURA-2AM. The cells are superfused in a Ca2+-free medium with nifedipine to block VOC, intracellular calcium stores are passively depleted by inhibition of SR Ca2<sup>+</sup> pump (SERCA) with thapsigargin or cyclopiazonic acid, and [Ca2+]i changes are evaluated after restoration of extracellular Ca2+. This Ca2<sup>+</sup> influx is referred as store operated calcium entry (SOCE). Alternatively, SOCE is also evaluated as the rate of fluorescence quenching by Mn2<sup>+</sup> which enters the cell after store depletion as Ca2<sup>+</sup> surrogate and reduces fluorescence through binding the dye (Bird et al., 2008). Concerning their structure, early studies suggested that SOC were formed by pore-forming subunits of the canonical transient receptor potential (TRPC) proteins. However, posterior discoveries showed a complex formed by Orai1 protein and stromal interacting molecule-1 (Stim1) as basic components able to generate store operated calcium influx, raising a controversy about the real molecular identity of SOCs (Earley and Brayden, 2015). This model suggests that SOC are hexamers of Orai proteins (Orai1, 2 or 3) and/or tetramers of TRPC proteins (TRPC1, 3, 4, 5, 6, or 7) that form the pore, and need the interaction mainly with Stim1 protein as calcium sensor of the SR to be activated. Store depletion of SR calcium results in redistribution of Stim1 from a homogeneous pattern in the bulk SR in resting cells, to a localized pattern into regions of SR called punctae, close to plasma membrane, allowing its interaction with Orai1 or TRPC proteins and their gating. Stim1-Orai complexes generate the Ca2+-release-activated Ca2<sup>+</sup> currents (Icrac) associated with oscillations of Ca2<sup>+</sup> and characterized as a small inwardly rectifying and selective Ca2<sup>+</sup> currents in electrophysiological studies. These oscillations are necessary to further recruit TRPC subunits, mainly TRPC1, to form Stim-TRPC complexes that generate additional but sometimes less selective Ca2<sup>+</sup> currents. The selectivity probably depends on the type of TRPC subunits that tetramerize to form the active complex. The recruitment of both currents results in a sustained Ca2<sup>+</sup> elevation termed the store operated calcium current (Isoc) (Ambudkar et al., 2017; Putney, 2017). This model is consistent with the ability of different sub-types of Stim, TRPC, and Orai proteins to generate SOCE in pulmonary arteries (Fernandez et al., 2012; Earley and Brayden, 2015; Wang et al., 2017). Moreover, other TRP channels such as transient receptor potential vanilloid-4 (TRPV4) and structurally unrelated Na+ and Ca2+-permeable cation channels such as acid-sensing ion channels-1 (ASIC1) may also contribute to SOC signaling in PASMC (Goldenberg et al., 2015; Jernigan, 2015). Whether the mechanism linking store depletion and ASIC1 activation to generate SOCE is related with association of ASIC1 with Stim, Orai, TRPC proteins, or other unknown signal has not been

elucidated. Next to SOC complexes, TRP proteins also work as ROC. After receptor activation by agonist binding, phospholipase C (PLC) isozymes convert phosphatidylinositol-4,5-bisphosphate (PIP2) into inositoltriphosphate (IP3) and diacylglycerol (DAG), and DAG or its synthetic analog OAG, used to identify receptor operated Ca2+ entry (ROCE) are common activators of all TRPC channels (Dietrich et al., 2005; Fernandez et al., 2012; Storch et al., 2017) except TRPC1 whose role as an ion channel or channel regulator is still a matter of debate (Dietrich et al., 2014).

Inhibitors targeting Stim, Orai, and TRP proteins are increasing in number and potency. Nevertheless, for many of these molecules, their mechanism of action, specificity, and toxicity needs further investigation to allow in vivo assays or clinical trials. For instance, two old useful inhibitors are SKF-96365 and 2-APB, are reported to block TRPC3/5, and the Stim/Orai interaction, respectively, at micromolar concentrations, but they also block VOC and IP3R at a similar concentration range (Putney, 2010; Bon and Beech, 2013). Lanthanides, such as La3<sup>+</sup> or Gd3<sup>+</sup> strongly inhibit Orai but their use is limited because their water solubility is poor in the presence of proteins and multivalent anions (Bird et al., 2008). Other blockers such as ML-9, BTP2, some GSK-compounds and RO2959 target other molecules in addition to Stim, Orai, or TRP subunits and/or are poorly soluble in physiological solutions (Prakriya and Lewis, 2015; Tian et al., 2016). Some recently characterized inhibitors show improved potency and selectivity: compound 8009-5364 and larixyl acetate block TRPC6 OAGinduced currents (Urban et al., 2012, 2016), AncoA4 blocks Orai channels and prevents its binding with Stim1 (Sadaghiani et al., 2014) while GSK2193874, GSK2220691, and HC067047 block TRPV4 currents (Everaerts et al., 2010; Thorneloe et al., 2012; Balakrishna et al., 2014). These inhibitors are promising tools to study the role of these channels on pulmonary vascular function (**Table 1**). The development of new agents specific for other TRP or Orai isoforms, combining potency and water-solubility should be helpful to study the composition and stoichiometry of native SOC/ROC complexes in pulmonary arteries and to validate them as potential pharmacological targets for pulmonary hypertension treatment.

## THE ROLE OF TRP, STIM, ORAI, AND ASIC PROTEINS IN THE PULMONARY ARTERIAL RESPONSE TO ACUTE HYPOXIA

The SOCE and different TRP, Orai, Stim, and ASIC proteins are robustly expressed in rodent pulmonary arteries and PASMC (Lin et al., 2004; Wang et al., 2006, 2017; Jernigan et al., 2009; Fernandez et al., 2015). Indeed, in PASMC from distal pulmonary arteries the increase of [Ca2+]i in response to acute hypoxia, SOCE and the expression of Stim1, TRPC1, 4, and 6 are greater than in proximal pulmonary arteries (Lu et al., 2008). Moreover, evidence from knockdown, knockout, or overexpression experiments show that in rodent PASMC, Stim1 and 2, Orai 1, 2 and 3, TRPC1 and 6, and ASIC1 contribute to SOCE (Lin et al., 2004; Jernigan et al., 2009; Nitta et al., 2014; Fernandez et al., 2015; Wang et al., 2017). Both, hypoxiainduced [Ca2+]i increase and HPV responses, are biphasic with a rapid and transient first phase of 5–20 min, followed by a slower and more sustained second phase of more than 180 min (Sommer et al., 2016). The first phase is abolished in TRPC6−/− mice, while knockdown of Stim1 suppresses the second phase (Weissmann et al., 2006; Lu et al., 2009). HPV, SOCE, and ROCE are also attenuated in ASIC1−/− mice compared to ASIC+/+ controls (Nitta et al., 2014). In lung of adult rat, SOC blockade with SKF-96365 inhibits HPV in a concentration-dependent manner (Weigand et al., 2005). The first phase of HPV is also suppressed by TRPC6 blockade with compound 800-5364 and larixyl acetate in mouse isolated lung (Urban et al., 2012, 2016). A significant decrease of HPV is also observed in knockout mice for transient receptor potential vanilloid-4 channel (TRPV4), and there is evidence of "promiscuous" TRPC6/TRPV4 heteromers assembly to generate SOC (Goldenberg et al., 2015). Transcripts of TRPC1, 3, 5, and 6 are detected in PASMC from late gestation ovine fetuses (Resnik et al., 2007) while TRPC1, 3, 4, 5, 6, Orai1, and Stim1 messengers are expressed in lungs from newborn lambs. Moreover the HPV recorded in vivo is significantly suppressed through SOC blockade with 2-APB in lambs (Parrau et al., 2013). Taken together, these data clearly show that at least in neonatal sheep and in adult rodents, SOC/ROC are key for contractile response to acute hypoxia, and that at least Stim1, TRPC6, and TRPV4 form part of the molecular complex involved in HPV. Nevertheless, as active Orai and TRPC complexes are hexamers and tetramers, respectively, the possibility of heteromeric association incorporating other Orai or TRPC subtypes to generate Ca2<sup>+</sup> influx associated to HPV cannot be excluded.

Currently, the mechanism linking hypoxia and SOC/ROC activation is a matter of research. For instance, increase of reactive oxygen species (ROS) during hypoxia is proposed to directly and indirectly activate RyR to deplete SR calcium stores, activate SOC, and increase [Ca2+]i and contraction (Sommer et al., 2016; Suresh and Shimoda, 2016). Hypoxia and ischemia/reperfusion provokes DAG accumulation and TRPC6 activation in PASMC and PAEC, respectively, where H2O<sup>2</sup> directly and indirectly mediates this effect (Weissmann et al., 2006, 2012). Interestingly H2O<sup>2</sup> resulting from increased ROS, also promotes the interaction of Stim1, with Orai1 and TRPC1, and upregulates these proteins to mediate SOCE in PASMC (Chen et al., 2017). H2O<sup>2</sup> also promotes Src family kinase-mediated stimulation of TPRV4 in lung microvascular endothelial cells (Suresh et al., 2015), but it is not elucidated if this mechanism also occurs in PASMC. It also remains to be demonstrated if the rate and the potency of the responses evoked by H2O<sup>2</sup> is consistent with tension development observed in HPV. Indeed, in PASMC, [Ca2+]i evokes contraction through its binding to calmodulin (CaM) and activation of myosin light chain kinase (MLCK), to phosphorylate the 20 kDa myosin light chain (MLC20), and increase the pMLC20/MLC<sup>20</sup> ratio (Ogut and Brozovich, 2008; Kuhr et al., 2012). Despite this obvious link between [Ca2+]i and contraction, the relation between pMLC20/MLC<sup>20</sup> ratio and SOC has been demonstrated only for

TABLE 1 | Current inhibitors of store operated channels and receptor operated channels.


The IC<sup>50</sup> values reported correspond to endogenous Icrac currents or to heterologous expressed TRPC/V, Orai, or Stim currents.

TRPV4 (Goldenberg et al., 2015), while it has not still been demonstrated for other SOC forming subunits. Further, the relation of SOC/ROC with other mechanisms regulating PASMC contraction such as calcium sensitization remains unexplored.

### TRP, STIM, ORAI, AND ASIC PROTEINS: TENSION DEVELOPMENT AND PULMONARY ARTERIAL REMODELING IN RESPONSE TO CHRONIC HYPOXIA

The pathological pulmonary arterial remodeling induced by chronic hypoxia is the result of an imbalance between proliferation and apoptosis of PASMC, changes in the differentiation state of PASMC and fibroblasts, and secretion of vasoactive compounds from PAEC (Suresh and Shimoda, 2016).

The PASMC and pulmonary arterial rings from rats exposed to chronic hypoxia have increased basal [Ca2+]I and arterial tension, which are reduced by SOC blockade. In contrast, SOC blockade have minimal effects on normoxic controls. Interestingly, protein levels of Orai1 and 2, TRPC1 and 6, Stim1 and 2, ASIC1, and SOCE are greater in pulmonary vasculature of hypoxic animals than control rats (Lin et al., 2004; Wang et al., 2006, 2009, 2017; Jernigan et al., 2012; Hou et al., 2013; He et al., 2018). ASIC1−/− mice show decreased SOCE and attenuated hypoxic pulmonary hypertension compared to ASIC+/+ controls (Nitta et al., 2014). Moreover, chronically

hypoxic newborn lambs gestated and born at 3,600 m altitude, have pulmonary hypertension, increased HPV, and upregulated TRPC4 and Stim1 pulmonary transcripts. In these animals, SOC blockade with a single dose of 2-APB evokes a greater attenuation of HPV compared to normoxic controls (Parrau et al., 2013). In another ovine model with partial gestation under chronic hypoxia, the newborn lambs show pulmonary hypertension that persists at sea level and increased TRPC4 and Orai1 expression. Further, an experimental therapy with 2-APB reduces mPAP, PVR, and pathological pulmonary arterial remodeling, both at the level of the medial and adventitial layer in these lambs (Castillo-Galán et al., 2016).

For TRPC1 and 6, as for Orai2, the hypoxic upregulation depends on hypoxia inducible factor-1α (HIF-1α), a cardinal transcription factor involved in the control of oxygenregulated genes (Wang et al., 2006, 2017). An additional increase of TRPC1 transcription is mediated by nuclear factor of activated T-cells c3 (NFATc3), a calcium-sensitive transcription factor, allowing an amplification of the initial hypoxic induction (Wang et al., 2009). Induction of ASIC1 dependent SOCE by chronic hypoxia is due to increased membrane localization mediated by RhoA activation rather than increased transcription (Herbert et al., 2017). In addition, hypoxia may induce the synthesis and secretion of mitogens from PAEC. These mitogens function as paracrine regulators of PASMC proliferation, and SOC may contribute to both syntheses of these mitogens as to their proliferative signaling. For instance, hypoxia upregulates SOCE and TRPC4 in PAEC. This Ca2<sup>+</sup> influx increases DNA binding activity of AP-1, a calcium-sensitive transcription factor, to activate transcription of AP-1 responsive genes like ET-1, PDGF, and VEGF among others. This is a mechanism that may contribute to pulmonary artery obliteration observed in pulmonary hypertension (Fantozzi et al., 2003). In turn, mitogens like platelet-derived growth factor (PDGF) and bone morphogenetic protein-4 (BMP4), a secreted ligand of the TGF-β superfamily, bind to receptors on PASMC. In doing so, PDGF and BMP4 further upregulate TRPC1, 4, 6, Orai1, and Stim1 expression, SOCE, and proliferation. PDGFmediated upregulation of Orai1/Stim1 depends on Akt/mTOR proliferative pathway in both PASMC and pulmonary artery fibroblasts, while BMP4-mediated induction of TRPC1/4/6 depends on p38MAPK-ERK1/2 signaling (Ogawa et al., 2012; Zhang et al., 2014).

Other signaling mechanisms such as extracellular calcium sensing receptor (CaSR) and peroxisome-proliferator activated receptor-γ (PPAR-γ) are also involved in the pathogenesis of pulmonary hypertension and remodeling through SOCE. CaSR are G-coupled protein receptors activated by extracellular Ca2<sup>+</sup> binding, and they are functionally coupled to TRPC6 activation to promote SOCE and ROCE, among other signaling mechanisms. CaSR are overexpressed in proliferating PASMC from hypoxic rodents or from patients with idiopathic pulmonary hypertension (IPAH), while their pharmacological or genetic suppression reduce proliferation and pulmonary hypertension (Yamamura et al., 2012, 2015; Smith et al., 2016; Tang et al., 2016). PPAR-γ is a member of the nuclear receptor hormone super family, and it is downregulated through TGF-β1 signaling in PASMC, whereas TRPC1 and 6 are upregulated in neonatal and adult rodents with hypoxic pulmonary hypertension. Conversely, stimulation of PPAR-γ reverses pulmonary hypertension and remodeling, and down regulates SOC expression (Gong et al., 2011; Yang et al., 2015; Jiang et al., 2016; Du et al., 2017).

To proliferate, PASMC undergo a de-differentiation, from a contractile and quiescent phenotype present in functional and healthy pulmonary vessels, to a synthetic and proliferative type, present in later stages of pathological remodeling. TRPC6, Orai2, and Stim2 contribute to the transition from the contractile to the synthetic phenotype (Fernandez et al., 2015). Theoretically, proliferative increase in response to SOCE may be mediated by direct stimulation of calmodulin kinase (CaMK) and p38-MAPK pathway, and activation of Ca2+-sensitive transcription factors, such as NFATc3, CREB, and NF-Kβ (Kuhr et al., 2012). NFATc3 has been better studied in relation to PASMC proliferation and remodeling. Dephosphorylation of NFATc3 through Ca2<sup>+</sup> and calcineurin promotes its nuclear translocation, to activate responsive genes related to PASMC proliferation, apoptosis resistance, and synthesis of contractile proteins, such as α-actin smooth muscle. Chronic hypoxia stimulates NFATc3 nuclear import. Parallel increase of ET-1 synthesis and activation of the RhoA-Rho kinase (RhoA/ROCK) pathway potentiates calcineurin-NFATc3 import (de Frutos et al., 2007, 2011; Ran et al., 2014). Moreover, the anti-proliferative effect of phosphodiesterase-5 (PDE5) inhibition with sildenafil on PASMC occurs with simultaneous inhibition of NFATc3 translocation, decreased SOCE, and TRPC1 downregulation. Additional SOCE decrease evoked by sildenafil may also be explained by PKG-dependent phosphorylation and inhibition of TRPC6 (Wang et al., 2009; Koitabashi et al., 2010). Conversely, the silencing of Stim1 reduces the proliferation of PASMC together with a decrease of NFATc3 translocation (Hou et al., 2013). Decreased NFATc3 nuclear import is also detected after genetic or pharmacological suppression of ASIC1-mediated SOCE (Gonzalez Bosc et al., 2016). Collectively, these data show that Ca2<sup>+</sup> handling mediated by SOC/ROC stimulate PASMC proliferation through multiple signaling pathways. Several of these transduction pathways need further characterization related to the type of SOC/ROC complexes involved as well as for possible cross talk between them. **Figure 1** depicts the principal links of SOC/ROC with the pulmonary artery response to both acute and chronic hypoxia.

# CONCLUSION AND PERSPECTIVES

Identification of Stim, Orai, and TRP proteins as part of the molecular components of SOCE/ROCE, and description of their direct or indirect regulation by oxygen allowed to gain understanding on their contribution to both hypoxic pulmonary vasoconstriction and remodeling. Nevertheless, some challenging developments are still needed in this field to

FIGURE 1 | SOC and ROC involvement in the pulmonary response to acute and chronic hypoxia. Acute hypoxia generates increase of ROS and H2O2. ROS stimulates RyR opening and depletion of SR-Ca2<sup>+</sup> stores. Both H2O<sup>2</sup> and depletion of SR-Ca2<sup>+</sup> stores stimulates Stim1/Orai1 association to generate the calcium-release activated calcium current (Icrac), which in turn promotes further interaction of Stim1 with TRPC1, TRPC6, and TRPV4 proteins and recruitment to generate the store operated calcium (Isoc) current. The participation of other TRPC proteins in association with TRPC1 or 6 to generate Isoc cannot be discarded. Increase of DAG cell content promoted by H2O<sup>2</sup> activates TRPC6/TRPCx channels. H2O<sup>2</sup> can also directly activate TRPC6 and TRPV4 channels. Finally, ASIC1 also mediates hypoxic increase of SOC through an unknown mechanism. Increase of SOC and ROC results in the final increase of Ca2+, stimulation of the myosin light chain kinase (MLCK), phosphorylation of myosin light chain, and contraction. Chronic hypoxia upregulates hypoxia inducible factor (HIF1) resulting in increased expression the secreted ligand bone morphogenetic protein-4 (BMP-4) from pulmonary artery endothelial cells (PAEC). Hypoxia also upregulates TRPC4 and Ca2<sup>+</sup> increase, stimulates activator protein1 (AP-1) that in turn upregulates platelet-derived growth factor (PDGF). Secreted PDGF upregulates TRPC1, 4, 6, Orai1, and stim1 in pulmonary artery smooth muscle cells (PASMC). Hypoxia also upregulates the calcium sensing receptor (CaSR) coupled to TRPC6 stimulation and downregulates PPAR-γ in PASMC. Chronic hypoxia also stimulates RhoA protein, which stimulates ASIC1 incorporation to de membrane. The net result is an increase of expression and activity of TRPC1, 4, 6, Orai 1, 2, and ASIC1. The contribution of all these proteins to store and receptor operated calcium entry results in sustained Ca2<sup>+</sup> increase to promote PASMC proliferation and remodeling.

fully understand SOC and ROC function in pulmonary arteries.

Despite Stim1, TRPC1, TRPC6, and TRPV4 are identified part of the SOC and ROC signaling complexes involved in HPV in pulmonary arteries, association of these subunits with other TRPC subtypes to form functional tetramers cannot be excluded. The same observation is valid for Orai complexes. The exact stoichiometry of native SOC and ROC complexes, if different homo/hetero multimer variants Orai, TRP, or ASIC are expressed along the development, and if they are associated to different SOCE and ROCE kinetics, is not elucidated.

Another question relates to the mechanism linking SR store depletion to ASIC1-mediated Ca2<sup>+</sup> entry in PASMC. It is not known whether this mechanism is related to Stim as Ca2<sup>+</sup> sensor of the SR, to association of ASIC1 with TRP proteins or to a totally different signal. The possibility to activate ion channels in response to store depletion by mechanisms different to those already discussed here cannot be ruled out.

There is no doubt that continuous chronic hypoxia upregulates Stim, Orai and TRPC proteins, SOCE/ROCE and evokes PASMC proliferation. Nevertheless, if different Stim, Orai, and TRPC proteins are responsive to hypoxia depending on development and duration of the stimulus is an unsolved question. Another form of hypoxia exposure is chronic intermittent hypoxia (CIH), for instance in obstructive sleep apnea (Iturriaga et al., 2014) or intermittent exposition to hypobaria (Herrera et al., 2015). Moreover, rats and mice exposed to CIH develop pulmonary hypertension and pathological remodeling, but the role of Ca2+ influx mediated by Stim, Orai, and ASIC channels on this condition is unexplored (Nisbet et al., 2009; Nara et al., 2015).

SOC and ROC are promising targets against pulmonary hypertension. There is an increasing number of inhibitors targeting Orai, TRP, and ASIC1, but many of them have not been tested in in vivo conditions to revert experimental pulmonary hypertension. SOC/ROC have a widespread distribution in the cardiovascular system, and side effects are likely. For instance, a 2-APB treatment decreases mPAP in hypoxic newborn lambs, but CO also diminishes transiently (Castillo-Galán et al., 2016). Therefore, novel SOC inhibitors targeting the pulmonary circulation need to be develop and tested in whole animals with the possibility to simultaneously record the systemic cardiovascular and hemodynamic responses possible to validate them as pharmacological approaches against hypoxic pulmonary hypertension.

### REFERENCES


# AUTHOR CONTRIBUTIONS

RR drafted the manuscript. SC-G, IH, EH, GE, and AL edited the manuscript. RR, SC-G, IH, EH, GE, and AL approved the final submission.

# FUNDING

This work was supported by grants from the Fondo Nacional de Investigación Científica y Tecnológica FONDECYT 1120605, 1130424, 1140647, and 1151119, and from Vicerrectoría de Investigación y Desarrollo, Universidad de Chile (VID-Enlace, ENL023f16).



and elevated intracellular Ca2<sup>+</sup> in pulmonary arterial smooth muscle cells. Circ. Res. 98, 1528–1537. doi: 10.1161/01.RES.0000227551.68124.98


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The handling Editor declared a shared affiliation, though no other collaboration, with one of the authors SC-G.

Copyright © 2018 Reyes, Castillo-Galán, Hernandez, Herrera, Ebensperger and Llanos. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Is Carotid Body Physiological O<sup>2</sup> Sensitivity Determined by a Unique Mitochondrial Phenotype?

Andrew P. Holmes, Clare J. Ray, Andrew M. Coney and Prem Kumar\*

Institute of Clinical Sciences, College of Medical and Dental Sciences, University of Birmingham, Birmingham, United Kingdom

The mammalian carotid body (CB) is the primary arterial chemoreceptor that responds to acute hypoxia, initiating systemic protective reflex responses that act to maintain O<sup>2</sup> delivery to the brain and vital organs. The CB is unique in that it is stimulated at O<sup>2</sup> levels above those that begin to impact on the metabolism of most other cell types. Whilst a large proportion of the CB chemotransduction cascade is well defined, the identity of the O<sup>2</sup> sensor remains highly controversial. This short review evaluates whether the mitochondria can adequately function as acute O<sup>2</sup> sensors in the CB. We consider the similarities between mitochondrial poisons and hypoxic stimuli in their ability to activate the CB chemotransduction cascade and initiate rapid cardiorespiratory reflexes. We evaluate whether the mitochondria are required for the CB to respond to hypoxia. We also discuss if the CB mitochondria are different to those located in other non-O<sup>2</sup> sensitive cells, and what might cause them to have an unusually low O<sup>2</sup> binding affinity. In particular we look at the potential roles of competitive inhibitors of mitochondrial complex IV such as nitric oxide in establishing mitochondrial and CB O2-sensitivity. Finally, we discuss novel signaling mechanisms proposed to take place within and downstream of mitochondria that link mitochondrial metabolism with cellular depolarization.

Keywords: carotid body, hypoxia, mitochondria, nitric oxide, arterial chemoreceptor, O<sup>2</sup> sensor, metabolism, mitochondrial inhibitors

### INTRODUCTION-THE CAROTID BODY AND HYPOXIA

The mammalian carotid body (CB) is a highly specialized sensory organ derived from the neural crest. The sensory units of the CB are the 'glomus' or 'type I' cells that respond to a variety of stimuli including hypoxia, hypercapnia, acidosis and hormones thereby allowing the CB to function as a polymodal receptor (Kumar and Bin-Jaliah, 2007; Ribeiro et al., 2013; Thompson et al., 2016). Type I cell activation leads to stimulation of adjacent chemoafferent fibers that relay sensory information into the central nervous system. The physiological consequence of CB stimulation is therefore the initiation of series of systemic protective reflexes characterized by increased ventilation, tachycardia, systemic vasoconstriction and adrenaline release from the adrenal medulla (Kumar, 2009).

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Gavin Clive Higgins, Monash University, Australia Fiona D. McBryde, University of Auckland, New Zealand

> \*Correspondence: Prem Kumar p.kumar@bham.ac.uk

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 30 January 2018 Accepted: 30 April 2018 Published: 16 May 2018

### Citation:

Holmes AP, Ray CJ, Coney AM and Kumar P (2018) Is Carotid Body Physiological O<sup>2</sup> Sensitivity Determined by a Unique Mitochondrial Phenotype? Front. Physiol. 9:562. doi: 10.3389/fphys.2018.00562

**136**

There is now a general consensus that a series of key processes contribute to the CB hypoxic chemotransduction cascade. These include attenuation of outward K<sup>+</sup> current, type I cell depolarization, Ca2<sup>+</sup> influx through L-type Ca2<sup>+</sup> channels, neurosecretion and chemoafferent excitation (Kumar and Prabhakar, 2012; Lopez-Barneo et al., 2016). Single type I cells are exquisitely sensitive to O<sup>2</sup> and rapid (within ms) activation of the hypoxic chemotransduction cascade initiates at PO2s of between 20–40 mmHg (Biscoe and Duchen, 1990; Buckler and Vaughan-Jones, 1994; Buckler and Turner, 2013); a level considerably greater than that at which cell metabolism is affected in O2-insensitive cells.

What remains highly controversial is the molecular identity of the specific O<sup>2</sup> sensor within the type I cell. We would argue that the physiological O<sup>2</sup> sensor should exhibit certain key features: (1) expression in the type I cell permitting intrinsic O<sup>2</sup> sensitivity; (2) the ability to bind O2; (3) its binding of O<sup>2</sup> occurs over the physiological range at which the type I cell is stimulated; (4) it is required for the CB to be stimulated by hypoxia; and (5) it is able to activate the CB transduction cascade within milliseconds. Many proposed sensors fit one or two of these criteria but few have been shown to adequately comply with all five.

In mammalian cells, O<sup>2</sup> is the terminal electron acceptor in the mitochondrial respiratory chain. Continuous binding and reduction of O<sup>2</sup> in the CuB/haem a<sup>3</sup> (cytochrome a3) binuclear center of complex IV drives mitochondrial electron transport and promotes activation of the mitochondrial ATP synthase. The long-established, mitochondrial hypothesis for chemoreception proposes that CB excitation induced by hypoxia is initiated by a reduction in O<sup>2</sup> dependent mitochondrial energy respiration. This review will briefly critique the current evidence as to whether the mitochondria can be considered the functionally relevant O<sup>2</sup> sensors within the CB.

### MITOCHONDRIAL INHIBITORS MIMIC ALL ASPECTS OF THE CAROTID BODY HYPOXIC CHEMOTRANSDUCTION CASCADE

All mitochondrial poisons induce chemoafferent excitation (Krylov and Anichkov, 1968; Mulligan et al., 1981; Obeso et al., 1989; Donnelly et al., 2014; Holmes et al., 2016, 2017) leading to rapid increases in ventilation (Owen and Gesell, 1931), heart rate and arterial blood glucose (Alvarez-Buylla and de Alvarez-Buylla, 1988). Chemoafferent responses are rapid, dose dependent and reversible (Donnelly et al., 2014; Holmes et al., 2016) and the magnitude of the rise in chemoafferent frequency caused by saturating concentrations is consistent with those evoked by severe hypoxia or anoxia (Krylov and Anichkov, 1968; Mulligan et al., 1981; Obeso et al., 1989). Furthermore, mitochondrial inhibitors and uncouplers augment neurotransmitter secretion, confirming an action through the type I cell rather than the afferent nerve endings (Obeso et al., 1989; Rocher et al., 1991). Despite the strong consistency between all of the different types of mitochondrial poisons (both in the older and more recent studies), it should be noted that some of the pharmacological agents used may not have acted selectively on the mitochondria and so the conclusions should be viewed with a certain degree of caution.

Mitochondrial poisons cause fast (within ms) type I cell depolarization and increases in [Ca2+]<sup>i</sup> . The size and timing of the [Ca2+]<sup>i</sup> rise observed using many different mitochondrial inhibitors or uncouplers closely resembles that seen in hypoxia (Biscoe et al., 1989; Biscoe and Duchen, 1990; Wyatt and Buckler, 2004; Buckler, 2011). As with hypoxia (Buckler and Vaughan-Jones, 1994; Urena et al., 1994), the increases in [Ca2+]<sup>i</sup> are dependent on cellular depolarization and extracellular Ca2<sup>+</sup> influx through voltage gated Ca2<sup>+</sup> channels (Wyatt and Buckler, 2004). TASK1/3, TREK-1, BKCa, Kv4.1, and Kv4.3, have all been shown to be expressed in the CB and inhibited by hypoxia (Buckler et al., 2000; Sanchez et al., 2002; Williams et al., 2004; Yamamoto and Taniguchi, 2006; Kim et al., 2009; Kreneisz et al., 2009; Turner and Buckler, 2013; Wang et al., 2017b). Of these, TASK1/3 and TASK-like K<sup>+</sup> currents are diminished by a multitude of mitochondrial inhibitors leading to membrane depolarization (Barbe et al., 2002; Wyatt and Buckler, 2004; Buckler, 2011; Turner and Buckler, 2013; Kim D. et al., 2015).

Recent evidence has revealed the presence of TRP and other non-selective Ca2+-activated cation currents in type I cells that are activated by hypoxia (Kumar et al., 2006; Kang et al., 2014; Kim I. et al., 2015; Wang et al., 2017a). Although intriguing, the full functional relevance of these currents in type I cell O2-sensing remains to be further characterized, and in particular whether these currents can be upregulated to preserve O<sup>2</sup> sensing in the absence of TASK channels (Turner and Buckler, 2013). Current evidence suggests that mitochondrial inhibitors are also capable of increasing these inward depolarizing currents (Kim I. et al., 2015; Wang et al., 2017a).

## MITOCHONDRIA ARE NECESSARY FOR THE CAROTID BODY TO RESPOND TO HYPOXIA

The presence of functional mitochondria does appear necessary for the CB to respond fully to hypoxia. For instance, cyanide, rotenone and FCCP all attenuate TASK channel currents in such a way that prevents any further reduction by hypoxia (Wyatt and Buckler, 2004). In intact CB preparations oligomycin, cyanide and azide all reduce or abolish subsequent chemoafferent responses to hypoxia (Mulligan et al., 1981; Donnelly et al., 2014). Some of the attenuation observed in these experiments may have been due to impairment of oxidative phosphorylation in the chemoafferent fibers, limiting their excitability. Any such impairment is not apparent in response to physiological levels of hypoxia, since the PO<sup>2</sup> that activates the type I cells occurs at a much higher level than those which would decrease mitochondrial function in the chemoafferent fibers. As such, chemoafferent responses to sustained hypoxia are better maintained than those in response to sustained high doses of mitochondrial inhibitors (Mulligan et al., 1981).

In a recent study the importance of mitochondrial complex I was tested by developing mice deficient in Ndufs2 (a gene coding for NADH dehydrogenase [ubiquinone] iron-sulfur protein 2- a component of complex I that participates in ubiquinone binding) in tyrosine hydroxylase positive cells (Fernandez-Aguera et al., 2015). Type I cells isolated from these mice were insensitive to hypoxia; they lacked any hypoxia-induced K <sup>+</sup> current attenuation, [Ca2+]<sup>i</sup> elevation or neurotransmitter release. Furthermore, these mice failed to increase respiratory frequency when breathing 10% O2. This work supported a previous study in which type I cell hypoxic chemosensitivity was abolished in the presence of rotenone (Ortega-Saenz et al., 2003). The authors propose a mechanism whereby exposure to hypoxia promotes reverse electron transport and ROS/NADH generation via complex I which is driven by a high rate of succinate oxidation at complex II. Accordingly, they have recently shown that genetic and pharmacological deactivation of complex II completely blocks type I cell hypoxic sensitivity (Gao et al., 2017).

This intriguing and elegant hypothesis does, however, show some discrepancies with evidence from earlier reports. For instance, similar experiments performed on CBs with heterozygous Sdhd knock out displayed an augmented, rather than depressed basal activity and had a completely preserved hypoxic response (Piruat et al., 2004). Furthermore, when rat type I cells were exposed to tetramethyl-p-phenylenediamine (TMPD) and ascorbate in the presence of rotenone, there was still a robust elevation in Ca2<sup>+</sup> upon hypoxic stimulation (Wyatt and Buckler, 2004). This would suggest that feeding electrons into cytochrome c is sufficient to sustain type I cell hypoxic sensitivity even when complex I activity (and ROS generation) is inhibited. Complex IV activity rather than complex I and II may therefore be necessary for hypoxic chemotransduction. The same report also showed that application of H2O<sup>2</sup> was unable to excite the type I cell directly. This observation is consistent with the lack of effect of multiple anti-oxidants used in other CB preparations and animal species (Sanz-Alfayate et al., 2001; Agapito et al., 2009; Gomez-Nino et al., 2009). Interestingly, using novel ex vivo CB culture techniques combined with FRET based ROS sensors, Bernardini et al. (2015) deduced that type I cell ROS actually decreases in hypoxia due to reduction in NADPH oxidase activity (an alternative ROS source). Clearly there is a need for reconciliation between these findings.

# THE CAROTID BODY MITOCHONDRIA ARE UNIQUE AND HAVE A LOW THRESHOLD FOR O<sup>2</sup>

The evidence that mitochondria are required for CB O<sup>2</sup> chemotransduction and that mitochondrial inhibitors can cause chemoexcitation, is not enough to define them as the O2-sensors in the CB. Clearly, mitochondria are able to bind O2. However, the K<sup>m</sup> of the cytochrome a<sup>3</sup> for O<sup>2</sup> is reported to be <1 mmHg in isolated mitochondria and between 1–5 mmHg in dissociated cells and tissue preparations, with little variation existing between different cell types (Wilson et al., 1988; Tamura et al., 1989). This is far lower than the PO<sup>2</sup> at which the CB type I cells begin to be activated and, for this reason, is a common argument against the mitochondrial hypothesis.

However, there is now a substantial body of evidence indicating that the CB type I cell mitochondria are unique. Experiments performed by Mills and Jobsis (1970, 1972), were the first to identify an unusually low affinity cytochrome a<sup>3</sup> within the CB. Using absorbance spectra, they estimated that 43– 67.5% of total cytochrome a<sup>3</sup> within the intact CB preparation had a remarkably low O<sup>2</sup> affinity. This fraction was reported to be almost 100% reduced at PO2s between 7–9 mmHg and 50% reduced at a PO<sup>2</sup> as high as 90 mmHg. In contrast, the remaining fraction was only 50% reduced at a PO<sup>2</sup> of approximately 0.8 mmHg, comparable to cytochrome a<sup>3</sup> found in other tissues (Gnaiger, 2001). Thus, the CB appeared to express both low and high affinity subtypes of cytochrome a3. At that time, the specific cellular location(s) of each was unclear. Later experiments utilized the photolabile binding of CO, to deduce that saturation of cytochrome a<sup>3</sup> with CO prevented any additional chemoafferent excitation during hypoxia, implying that not only was the cytochrome a<sup>3</sup> in the CB unusual, it was also required for O2-sensing (Wilson et al., 1994; Lahiri et al., 1999). It should be pointed out that the concentrations of CO used in these studies could have directly modified the activity of the BKCa channel (Williams et al., 2004, 2008) and the generation of H2S (Yuan et al., 2015) and as such some of the observations could be related to mechanisms independent of the mitochondria.

In dissociated rabbit type I cell clusters, mitochondrial electron transport begins to be inhibited at a high PO<sup>2</sup> value of approximately 40 mmHg (Duchen and Biscoe, 1992a). PO2-NADH response curves demonstrate a significant 'right shift' in type I cells compared to sensory neurons, indicative of a heightened and distinctive O<sup>2</sup> sensitivity. In addition, mitochondrial depolarization occurs at higher PO2s compared to O2-insensitive cells (Duchen and Biscoe, 1992b). More recent work has verified the "right shifts" in both PO2-NADH and PO2-rh-123 response curves in rat type I cells, confirming that the unusually low mitochondrial O<sup>2</sup> affinity is conserved in multiple species (Turner and Buckler, 2013). By isolating complex IV activity with a cocktail of mitochondrial inhibitors plus TMPD and ascorbate, the authors were able to reveal that complex IV activity is a component of the mitochondria with the exceptionally low O<sup>2</sup> affinity. Importantly, type I cell hypoxic response curves for electron transport inhibition, mitochondrial depolarization and complex IV run-down display considerable overlap with the rise in Ca2+, indicating that these processes are intimately linked. Therefore, it does appear that type I cell mitochondria have a highly specialized low affinity for O<sup>2</sup> due to an altered function/subtype of cytochrome a<sup>3</sup> in complex IV that predisposes CB energy metabolism to being impaired at high O<sup>2</sup> tensions. It is likely that the high affinity cytochrome a<sup>3</sup> in the CB described by Mills & Jöbsis is located in the non-O<sup>2</sup> sensing tissue such as the, nerve endings, blood vessels and type II cells.

Understanding the mechanism linking a fall in mitochondrial O<sup>2</sup> consumption with K<sup>+</sup> channel inhibition (or cation channel activation) is contentious. As previously mentioned, there could be a role for elevated mitochondrial ROS generation but this

is still to be validated (Fernandez-Aguera et al., 2015). Another possibility is an alteration in cytosolic nucleotides. Switching from a cell attached to inside-out patch configuration diminishes background K<sup>+</sup> channel activity, suggesting that a basal level of an intracellular factor(s) activates TASK channels in normoxia (Varas et al., 2007). Addition of 5 mM MgATP in the insideout configuration can restore about 50% of this background K <sup>+</sup> channel activity. Both mitochondrial inhibition and hypoxia also significantly elevate free Mg2+, consistent with a decrease in MgATP. Thus, the fall in MgATP during hypoxia is likely to attenuate a significant proportion of TASK-current leading to depolarization. However, the remaining modulators that account for the other 50% of TASK current are still to be identified.

Another proposed mediator of TASK channel activity that is sensitive to changes in cytosolic nucleotide concentrations is AMPK (Wyatt et al., 2007). However, initial favorable studies based on pharmacological evidence have since been challenged by the finding that the AMPK-α1α<sup>2</sup> deficient CB retains complete O2-sensitivity (Mahmoud et al., 2016). Other groups have also shown that pharmacological targeting of AMPK does not impact on the hypoxia-induced K<sup>+</sup> channel inhibition (Kim et al., 2014). Discrepancies may arise from the non-selectivity of the drugs used to evaluate AMPK function and potential redundancy mechanisms known to develop in genetic animal models (Nowak et al., 1997). A final hypothesis is that a buildup in lactate upon mitochondrial inhibition in hypoxia activates the olfactory receptor Olfr78 (Chang et al., 2015). However, the concentration of lactate necessary to elevate Ca2<sup>+</sup> in an intact CB preparation appears to be quite high (30 mM) and whether local levels reach this threshold during hypoxia is uncertain. We await a mechanism demonstrating how activation of Olfr78 (a G-protein-coupled receptor) modulates TASK or cation channel activity. A summary of the proposed O2-sensitive mitochondrial signaling pathways in the CB is presented in **Figure 1**.

# WHAT DETERMINES THE LOW O<sup>2</sup> AFFINITY OF THE CAROTID BODY MITOCHONDRIA?

We propose that there are 2 potential means to account for the extraordinary O<sup>2</sup> sensitivity of type I cell mitochondria. First, there could be a high level of production of a cytosolic factor that is able to freely diffuse into the mitochondria and then compete with O<sup>2</sup> binding in complex IV (**Figure 2**). We predict that this competition would render mitochondrial electron transport more susceptible to subsequent falls in O2. We recently tested this by applying exogenous nitrite to CBs and subsequently measuring hypoxic sensitivity (Holmes et al., 2016). Nitrite is reduced within the mitochondria to generate local NO, a recognized competitive inhibitor of complex IV (Brown and Cooper, 1994; Cleeter et al., 1994; Castello et al.,

2006, 2008; Basu et al., 2008). Moderate basal inhibition of the CB mitochondria by nitrite exaggerated the subsequent chemoafferent excitation during hypoxia signifying an increase in CB O<sup>2</sup> sensitivity. Therefore, we validated the idea that CB hypoxic sensitivity could be adjusted by a factor capable of competing with O<sup>2</sup> in the mitochondria and suggested a physiological role for endogenous NO in establishing type I cell mitochondrial O2-sensitivity. Measurable amounts of NO have been detected in mitochondrial of type I cells (Yamamoto et al., 2006). A possible source is nitric oxide synthase 3 (NOS-3) given its location within the type I cell (Yamamoto et al., 2006). Interestingly, mice with reduced NOS-3 have a dampened hypoxic ventilatory response and a depressed CB function (Kline et al., 2000). One explanation for this is an adaptation to chronic hypoxia brought about by reduced CB blood flow. However, this is unlikely as there is no significant type I cell hyperplasia/hypertrophy (McGregor et al., 1984; Tatsumi et al., 1991). Instead, the blunted CB activity could be due to the lack of NO acting on the CB mitochondria. Consideration of the precise compartmentalization of NO should also be taken into account. Whilst NO in the mitochondria induces chemostimulation, its action in other regions is likely to have opposing effects via modulation of soluble guanylate cyclase and L-type Ca2+/BKCa channels (Summers et al., 1999; Iturriaga et al., 2000; Silva and Lewis, 2002; Valdes et al., 2003). In addition, other diffusible cytosolic factors have been implicated in CB O<sup>2</sup> sensing including H2S and CO (Peng et al., 2010; Yuan et al., 2015). Both of these gasses are capable of inhibiting type I cell mitochondria (Wilson et al., 1994; Lahiri et al., 1999; Buckler, 2011). Future experiments are required to evaluate if these substances act by setting type I cell mitochondrial O<sup>2</sup> sensitivity.

A second explanation for the low O2-affinity of the type I cell mitochondria is that it has a unique gene expression profile (**Figure 2**). Exploring gene expression in the CB is challenging due to its relatively small size and heterogeneity. However, advances in molecular biology techniques now make it possible to perform whole genome analysis using just micrograms of tissue or even single cells. RNA-sequencing analysis has now

electron transport.

revealed a number of mitochondrial related genes that have a particularly high expression in the type I cell (Zhou et al., 2016). Of these Ndufa4l2 and Cox4i2 have been shown to be more highly expressed in the CB compared with tissue from the superior cervical ganglion (Gao et al., 2017). Whether these two genes contribute to the low mitochondrial O<sup>2</sup> affinity remains to be determined but these findings do support the idea that CB mitochondria have a unique genetic signature encoding their mitochondrial complexes. We would expect many more genetic studies to probe this further until the type I cell mitochondria can be accurately modeled to pinpoint the structural conformation underlying its low O<sup>2</sup> affinity. An interesting comparator may be the mitochondria isolated from guinea pig CB which does not appear to have any inherent O<sup>2</sup> sensitivity (Gonzalez-Obeso et al., 2017).

### REFERENCES


# CONCLUSION

On current evidence, it is very hard to disprove the mitochondrial hypothesis of CB O<sup>2</sup> sensing. The mitochondria seem to fulfill all five criteria that we have proposed for adequate O2-sensors. What is less clear is a mechanistic understanding of how the low O<sup>2</sup> sensitivity of the CB mitochondria is achieved and if mitochondria are involved in establishing pathological changes in CB function.

### AUTHOR CONTRIBUTIONS

AH, CR, AC, and PK all contributed to the writing and editing of the manuscript.




**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Holmes, Ray, Coney and Kumar. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Plasmatic Concentrations of ADMA and Homocystein in Llama (Lama glama) and Regulation of Arginase Type II: An Animal Resistent to the Development of Pulmonary Hypertension Induced by Hypoxia

Vasthi López <sup>1</sup> , Fernando A. Moraga<sup>2</sup> \*, Anibal J. Llanos <sup>3</sup> , German Ebensperger <sup>3</sup> , María I. Taborda<sup>1</sup> and Elena Uribe<sup>4</sup>

<sup>1</sup> Laboratorio de Metabolismo de Aminoácidos e Hipoxia, Departamento de Ciencias Biomédicas, Universidad Católica del Norte, Coquimbo, Chile, <sup>2</sup> Laboratorio de Fisiología, Hipoxia y Función Vascular, Departamento de Ciencias Biomedicas, Facultad de Medicina, Universidad Católica del Norte, Coquimbo, Chile, <sup>3</sup> Laboratorio de Fisiología y Fisiopatología del Desarrollo, Departamento de Ciencias Biomédicas, Universidad de Chile, Santiago, Chile, <sup>4</sup> Laboratorio de Enzimología, Departamento de Bioquímica y Biología Molecular, Universidad of Concepción, Concepción, Chile

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Gautham Yepuri, University of Fribourg, Switzerland Zhihong Yang, University of Fribourg, Switzerland

> \*Correspondence: Fernando A. Moraga fmoraga@ucn.cl

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 31 December 2017 Accepted: 04 May 2018 Published: 29 May 2018

### Citation:

López V, Moraga FA, Llanos AJ, Ebensperger G, Taborda MI and Uribe E (2018) Plasmatic Concentrations of ADMA and Homocystein in Llama (Lama glama) and Regulation of Arginase Type II: An Animal Resistent to the Development of Pulmonary Hypertension Induced by Hypoxia. Front. Physiol. 9:606. doi: 10.3389/fphys.2018.00606 There are animal species that have adapted to life at high altitude and hypobaric hypoxia conditions in the Andean highlands. One such species is the llama (Lama glama), which seem to have developed efficient protective mechanisms to avoid maladaptation resulting from chronic hypoxia, such as a resistance to the development of hypoxia -induced pulmonary hypertension. On the other hand, it is widely known that different models of hypertension can arise as a result of changes in endothelial function. The respect, one of the common causes of deregulation in endothelial vasodilator function have been associated with down-regulation of the NO synthesis and an increase in plasma levels of asymmetric dimethylarginine (ADMA) and homocysteine. Additionally, it is also known that NO production can be regulated by plasma levels of L-arginine as a result of the competition between nitric oxide synthase (NOS) and arginase. The objective of this study, was to determine the baseline concentrations of ADMA and homocysteine in llama, and to evaluate their effect on the arginase pathway and their involvement in the resistance to the development of altitude-induced pulmonary hypertension. METHOD: Lowland and highland newborn sheep and llama were investigated near sea level and at high altitude. Blood determinations of arterial blood gases, ADMA and homocysteíne are made and the effect of these on the arginase activity was evaluated. RESULTS: The basal concentrations of ADMA and homocysteine were determined in llama, and they were found to be significantly lower than those found in other species and in addition, the exposure to hypoxia is unable to increase its concentration. On the other hand, it was observed that the llama exhibited 10 times less arginase II activity as compared to sheep, and the expression was not induced by hypoxia.

**144**

Finally, ADMA y Hcy, has no effect on the type II arginase pathway. CONCLUSION: Based on our results, we propose that low concentrations of ADMA and homocysteine found in llamas, the low expression of arginase type II, DDAH-2 and CBS, as well as its insensitivity to activation by homocysteine could constitute an adaptation mechanism of these animals to the hypoxia.

Keywords: llama, ADMA, pulmonary hypertension, homocysteine, arginase

### INTRODUCTION

Hypoxic pulmonary vasoconstriction (HPV) in distal pulmonary arteries is an intrinsic vasoconstrictor response to low oxygen levels. Once the normoxic condition returns, the vasoconstrictor response returns to normal levels. If the hypoxic stimulus becomes chronic it leads to structural and functional maladaptation of the pulmonary arterial bed, characterized by pulmonary artery remodeling and vasoconstriction. This results in sustained pulmonary arterial hypertension (PAH) that is often accompanied by right ventricular hypertrophy, heart failure and eventually death (Giordano, 2005; Pe-aloza, 2012). Species that have adapted to life at high altitude and hypobaric hypoxia conditions in the Andean altiplano have existed for at least 10 million years. One such species is the llama (Lama glama), which developed efficient protective mechanisms to avoid maladaptation resulting from chronic hypoxia, such as a decreased hormonal vasoconstrictor response, an increased blood hemoglobin concentration, decreased cerebral O<sup>2</sup> consumption, and increased Lactate Dehydrogenase activity (Llanos et al., 2003; Ebensperger et al., 2005). The llama possesses an attenuated cardiovascular response to hypoxia, which prevents the development of pulmonary hypertension and arteriolar muscle remodeling (Harris et al., 1982; Moraga et al., 2011). Contrary to llamas, newborn sheep gestated and born at high altitude have marked pulmonary hypertension compared to their counterparts (Llanos et al., 2011a).

Additionally, different models of hypertension can arise as a result of changes in endothelial function (Panza, 1997; Landmesser and Drexler, 2007). One of the most common causes of dysregulated endothelial vasodilator function is an alteration to the biochemical pathway that produces nitric oxide (NO) (Ignarro, 2002). Different models of hypertension are associated with down-regulation of the NO synthetic pathway and an increase in the plasma levels of asymmetric dimethylarginine (ADMA) and homocysteine. ADMA and homocysteine are inhibitors of nitric oxide synthase (NOS) which contributes to the pathogenesis of pulmonary hypertension (Arrigoni et al., 2003; Millatt et al., 2003; Wierzbicki, 2007; Lüneburg et al., 2016). Additionally, it has been reported that dimethyl-aminohydrolase (DDAH-2) (an enzyme that regulates the degradation of ADMA) and cystathionine β-synthase (CBS) (an enzyme that regulated the degradation of homocysteine) are target molecules due to the existence of polymorphisms in their genes that induce the dysregulation of their metabolism and, therefore, increase the risk of developing cardiovascular diseases (Zhang et al., 2014; Xuan et al., 2016).

NO production can be regulated by plasma levels of L-arginine as a result of the competition between NOS and arginase (Böger, 2004; Pernow and Jung, 2013). It has been suggested that arginase activity plays an important role in the pathogenesis of various pulmonary disorders (Maarsingh et al., 2008; Durante, 2013). For instance, an increase in arginase activity has been associated with several pulmonary and systemic hypertension models (Johnson et al., 2005; López et al., 2009; Chu et al., 2016). Nevertheless, there is no information available about the basal levels of ADMA and homocysteine in animals genetically adapted to life in high altitudes and, therefore, resistant to the development of pulmonary hypertension, such as the llama (Lama glama). Also, there is no information regarding the participation of the arginase signaling pathway and its regulation by ADMA and homocysteine in this model. Therefore, the aim of this study was to determine the baseline concentrations of ADMA and homocysteine in an animal model resistant to the development of pulmonary hypertension induced by hypoxia. We also evaluated the effect of ADMA and homocysteine on the arginase pathway and their involvement in the resistance to the development of altitude-induced pulmonary hypertension. We used sheep (Ovis aries) as a non-genetically adapted control group.

### MATERIALS AND METHODS

## Bioethics Committee

All animal care, maintenance, procedures, and experimentation were performed in accordance with the United Kingdom's Animals (Scientific Procedures) Act 1986, and the American Physiological Society's Guiding Principles for Research Involving Animals and Human Being (American Physiologycal Society, 2002) and were reviewed and approved by the Faculty of Medicine Ethics Committee of the University of Chile.

### Animals

Lowland and highland newborn sheep (Ovis aries, n = 6) and llamas (Lama glama, n = 5) were studied near sea level (Santiago, 580 m, 710 mmHg barometric pressure. Lowland) and at high altitude (Putre, 3,600 m, 480 mmHg barometric pressure. Highland).

### Determination of Arterial Blood Gases in Lowland and Highland Newborn Sheep and Llamas

Newborns were submitted to a surgical procedure at 4–5 days old and studied at 7–10 days old. Animals were placed under general anesthesia with ketamine-diazepam association (10 mg kg−<sup>1</sup> i.m. ketostop, Drag Pharma-Invectec, Santiago, Chile: 0.1– 0.5 mg kg−<sup>1</sup> i.m Diazepam, Laboratorio Biosano, Santiago, Chile) and additional local infiltration of 2% lidocaine (Dimecaíne, Laboratorio Beta, Santiago, Chile), polyvinyl catheters (1.2 mm i.d) were placed in the descending aorta and inferior vena cava via hindlimb artery and vein, exteriorized subcutaneously through the animal flank and kept in a pouch sewn onto the skin. Also, a Swan-Ganz catheter (Edwards Swan-Ganz 5 French, Baxter Healthcare Corporation, Irvine, CA, USA) was inserted into the pulmonary artery via an external jugular vein, exteriorized, and placed in a pouch around the neck of the animal. All vascular catheters were filled with a heparinized saline solution (500 IU heparin mL−<sup>1</sup> in 0.9% NaCl) and plugged with a copper pin. Ampicillin 10 mg kg−<sup>1</sup> i.v (Ampicilina, Laboratorio Best-Pharma, Santiago, Chile) was administered every 12 h while the animals were catheterized. The experiments commenced 3 days after surgery.

# Determination of Arterio-Pulmonary Pressure

Pulmonary arterial pressure (PAP), systemic arterial pressure (SAP), and heart rate (HR) were recorded via a data acquisition system (PowerLab/8SP System and LabChart v7.0 Software; ADInstruments) connected to a computer. Cardiac output (CO) was determined with the thermodilution method by the injection of 3 ml of chilled (0◦C) 0.9% NaCl into the pulmonary artery through the Swan-Ganz catheter connected to a CO computer (COM-2 model, Baxter). Mean PAP (mPAP), mean SAP (mSAP), and pulmonary (PVR) vascular resistances were calculated as described previously, Herrera et al. (2007).

### Determination of ADMA and Homocysteine Plasma Concentrations

Plasma levels of ADMA (Enzo Life Sciences, Farmingdale, New York, USA) and homocysteine (Alpco Diagnostics, Salem, New Hampshire, USA) were measured by enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instructions.

### Enzyme Preparations

The lungs were excised and frozen using liquid nitrogen and stored at −80◦C until further use. The heart tissue was homogenized in a solution comprising of 10 mM Tris-HCl (pH 7.5), 0.1 mM PMSF, 1 mM DTT and 5 mM EDTA, and centrifuged at 5,000 g for 10 min at 4◦C. The homogenates were centrifuged for 30 min at 5,000 g and 4◦C and the supernatant was collected. Subsequently, the enzyme solution (supernatant) was separated by chromatography on a DEAEcellulose column equilibrated with 5 mM Tris-HCl pH 7.5. The active fractions eluted with the washings of the column corresponded to arginase I, and the active fractions eluted at 0.10–0.15 M corresponded to arginase II (Venkatakrishnan et al., 2003).

The presence of arginase I and II, in the washing fraction and the eluted fraction, respectively, was confirmed by Western Blot. Finally, the active fractions were pooled and stored at −20◦C.

### Extraction and Quantification of DDAH-2 and Arginase Type II mRNA Level by qRT-PCR

Total RNA was isolated using TRIZOL reagent (Invitrogen Corp., Carlsbad, CA, USA). RNA concentration and purity were assessed by Ultra Violet (UV) spectrophotometry (1.8 < A260/A280 < 2.0). RNA integrity was evaluated using electrophoresis. Reverse transcription reaction was carried out using 4 µg of total RNA and the First Strand complementary DNA (cDNA) Synthesis kit (Thermo Scientific, Boston, MA, USA).

To make standard curves, 1 µL of first-strand cDNA was amplified with AffinityScript cDNA Synthesis Kit and quantification of Polymerase Chain Reaction (PCR) products were used for plotting standard curves.

Quantitative real-time PCR was carried out using a HT-7500 thermo- cycler (Applied Biosystems) and using Brilliant III Ultra-Fast SYBR <sup>R</sup> Green QPCR Master Mix. program PCR was of 95◦C for 3 min, followed by 40 cycles of 95◦C for 30 s and 60◦C for 30 s. GAPDH was used as an internal control.

DDAH-2, CBS and eNOS mRNA level was normalized with GAPDH (Gliceraldehído-3-fosfato deshidrogenasa) mRNA to compensate for variations in initial RNA amounts. Normalization was carried out by dividing the logarithmic value of DDAH- 2, CBS and eNOS by the logarithmic value of GAPDH.

## Arginase Activity Measurements

Specific arginase activity was determined in lung tissues based on urea production from L-arginine using a colorimetric method with α-isonitrosopropiopheonone (Archibald, 1957). Briefly, 50 µl of the fractions obtained from DEAE-cellulose were incubated in 30 mM Tris-HCl (pH 7.5) containing 100 mM L-arginine at 37◦C for 10 min. The reaction was stopped by the addition of 1 ml of an acid mixture containing 9% (v/v) of H3PO<sup>4</sup> and 23% (v/v) of H2PO4. To evaluate the effect of homocysteine and ADMA on arginase activity, a specific arginase activity assay was performed in the presence of physiological concentrations of ADMA and homocysteine (0.5, 1.0, 3.0, 6.0, and 10.0 µM) in a 50 mM Tris-HCl solution (pH 7.5).

TABLE 1 | Arterial blood gases in lowland and highland newborn sheep and llamas (n = 5).


\*Values represent the mean ± SE (n = 5 per group). p < 0.01.

FIGURE 2 | Plasma concentrations of ADMA and homocysteine in newborn llamas and sheep. (A) Plasma concentration of ADMA. A significant difference was found between lowland and highland newborn sheep (\*p < 0.01). (B) Plasma concentration of Homocysteine. A significant difference was found between lowland and highland newborn sheep (\*p < 0.01). No significant differences were observed between ADMA and homocisteine concentrations was found between lowland and highland newborn llama (n.s).

# Data Analysis and Statistics

All statistical analyses were performed using GraphPad Prism 5.0 software. The mean values and standard error (SE) were calculated for each parameter. The normality of quantitative variables was established using the Kolmogorov–Smirnov test. The statistical validity of the difference across all testing conditions was established using analysis of variance (ANOVA) of one factor, and ANOVA for repeated measures was performed for differences within the group. Pearson's correlation was also performed. Linear regression analysis was carried out to establish the association between arginase activity (dependent variable) and the other variables. The significance level was established at p < 0.05.

### RESULTS

A decrease in PaCO2, PaO2, and SaO<sup>2</sup> values was observed at high altitude conditions in both sheep and llamas. However, newborn llama SaO<sup>2</sup> values were higher at high altitude (**Table 1**). The biggest difference observed between the two species was the higher hemoglobin oxygen saturation in newborn llama compared to newborn sheep at high altitude. This difference is explained by the higher hemoglobin oxygen affinity of llamas compared to sheep (Moraga et al., 1996).

### Newborn llamas Are Resistant to the Development of Hypoxia-Induced Pulmonary Hypertension

Newborn llamas do not present significant differences in mPAP (14 ± 1 and 15 ± 0.5 mmHg) between lowland and highland conditions (**Figure 1A**). No modifications were observed in PVR in newborn llamas at lowland vs. highland (**Figure 1B**). On the contrary, newborn sheep had an increased mPAP (12 ± 0.6 mmHg) at lowland compared to highland levels (21 ± 2.0 mmHg) (**Figure 1A**). Additionally, highland newborn sheep exhibited increased PVR values compared to lowland newborn sheep (**Figure 1B**). No significant difference were observed in PVR between both species at lowland and highland (p > 0.05).

PCR of CBS (C) Quantitative PCR of eNOS. Gene expression was measured by real-time PCR; data were normalized to GAPDH levels. (n = 6 per group, \*p < 0.001 group).

observed in arginase activity between these two conditions in llamas (n.s). (The western blots correspond to the eluted fraction of DEAE-cellulose chromatography).

# Llamas Have Low ADMA and Homocysteine Concentrations That do not Increase When Exposed to Hypoxia

To evaluate the role of ADMA and homocysteine in the development of pulmonary hypertension, we measured the basal concentrations of ADMA and homocysteine in the plasma of lowland and highland newborn sheep and llamas. Llamas had significantly lower plasmatic ADMA concentrations compared to sheep, both in lowland and highland newborn animals. In fact, ADMA concentrations in lowland conditions were 1.48µM in sheep and 0.153µM in llamas, approximately 10 times lower (**Figure 2A**).

On the contrary to the expected results, exposure to hypoxia did not increase ADMA concentrations in llamas, whereas hypoxic conditions induced a significant increase in the ADMA concentration in sheep, from 1.4 to 4.45µM (approximately a 4x increase) (**Figure 2A**).

The same phenomenon was observed with the concentration of homocysteine which was significantly lower in llamas compared to sheep, and hypoxia was incapable of producing an increase (**Figure 2B**). In fact, sheep had higher concentrations of homocysteine and hypoxia significantly increased the concentration from 17.32 to 70.4µM (approximately a 4x increase).

# Llamas Have Low DDAH-2 and CBS Expression Levels and Which Are Unaffected by Hypoxia

To evaluate the relationship between the observed concentrations of ADMA and homocysteine and the expression levels DDAH-2 (dimethylarginione dimethyaminohydrolase-1) and CBS (cystathionine-β-synthase), enzymes responsible for the regulation of the endogenous concentrations of ADMA and homocysteine, respectively, the expression levels of DDAH-2 and CBS mRNAs from the lung parenchyma were measured. Low concentrations of DDAH-2 and CBS mRNA from llama lung parenchyma were detected under conditions of normoxia and hypoxia (**Figures 3A,B**), unlike sheep, a significant decrease in expression levels of both genes was observed during hypoxia.

The expression levels of endothelial NOS were also evaluated and it was determined that in the llama there is not significant increase in eNOS expression levels during hypoxia conditions (**Figure 3C**).

### Llamas Have Low Levels of Type II Arginase Activity Which Is Not Up-Regulated During Hypoxia

An increase in the expression and activity of arginase has been related to the development of hypertension in different models. Therefore, we measured arginase type II activity in homogenized lungs from newborn sheep and llamas at lowland and highland (**Figure 4**).

Llamas exhibited 10 times less arginase II activity compared to sheep which was not induced by hypoxia (**Figure 4**). Furthermore, arginase II activity remained at similar levels in llamas at lowland and highland.

Additionally, hypoxia induced a significant increase of arginase type II activity by approximately 5 times in newborn sheep.

# Arginase Type II in Llamas Is Insensitive to Activation by Homocysteine

ADMA and homocysteine have been described as markers for cardiovascular risk and previous results have shown that ADMA and homocysteine have an activation effect on arginase type II from hypoxia-induced hypertensive rats (López et al., unpublished data). Therefore, we evaluated the effect of these metabolites on arginase type II activity from the lung parenchyma of newborn llamas and sheep (**Figures 5**, **6**). ADMA was a poor inhibitor of arginase type II activity in newborn llamas and sheep at lowland (**Figure 5A**) and highland (**Figure 5B**). However, increased inhibition of arginase type II activity by ADMA was observed in hypoxic conditions in llamas, demonstrating an inhibition of approximately 25% at physiological concentrations of ADMA (1µM) and reaching a maximum inhibition level of 40% at higher concentrations (10µM). Additionally, arginase type II activity of newborn llamas in normoxic conditions was not affected by ADMA (**Figures 5A,B**).

FIGURE 5 | Effect of increasing ADMA concentrations on arginase type II activity from lung parenchyma in newborn llamas and sheep. (A) The effect of ADMA on arginase type II activity in lowland conditions. No significant differences were observed in arginase activity between sheep and llamas. (B) The effect of ADMA on arginase type II activity in Highland conditions. A significant difference was found between newborn sheep and llamas. \*Values represent the mean ±SE (n = 5 per group). \*p < 0.01, Llamas vs. sheep.

±SE (n = 5 per group). \*p < 0.01, Llamas vs. sheep.

Experiments with homocysteine demonstrated that arginase type II activity in sheep was not affected at lowland. However, homocysteine (4µM) was an important inhibitor and reduced enzyme activity by approximately 50% (**Figure 6A**) in llamas.

In highland conditions, homocysteine was an important activator of arginase type II in sheep, increasing its activity with increasing concentrations. On the contrary, arginase type II was inhibited in llamas by approximately 50% with 6µM homocysteine (**Figure 6B**).

### DISCUSSION

Llamas are genetically adapted to live in hypobaric hypoxia conditions at high altitudes. Among the physiological adaptations of the adult llama that allow it to live under conditions of oxygen limitation at altitude are: an increased blood hemoglobin concentration, low P50, high muscle myoglobin concentration, a more efficient O<sup>2</sup> extraction at tissue levels, and high lactic dehydrogenase activity. Furthermore, llamas.

the adult llama avoids pulmonary arterial hypertension and cardiac remodeling, among other adverse effects, by having less muscularized pulmonary arterioles than the adult sheep (Llanos et al., 2011b). Together, these adaptations allow the llama to adapt to life at high altitudes. In agreement with the adaptations of the pulmonary circulatory system previously described, ours results showed that newborn llama at lowland or highland do not show hypertension in pulmonary arteries when calculating the PVR using Ohm's law (Herrera et al., 2007). We observed that the lack of PVR modifications in newborn llama is explained by the maintenance in the cardiac output, similar to that described by Herrera et al. (2008). In contrast, in newborn sheep the higher mPAP and PVR observed at high altitude is explained by an increased cardiac output in newborn sheep. The unaltered PVR in newborn llamas can be explained by their lack of hypertrophic pulmonary vascular remodeling and pulmonary hypertension (Llanos et al., 2011b) by a proposed blunting in this response.

Additionally, studies performed by Herrera et al. (2008) showed that there is an increased pulmonary CO production compared to NO production in newborn llamas. CO is an alternative vasodilator synthesized by the HO- carbon monoxide system that is activated during chronic hypoxia because the relative increase in NO production is not sufficient to counteract the increase in PAP at high altitude. Thus, these authors proposed that activation of the HO-carbon monoxide system induces an adaptation mechanism in these animals during birth that would protect their vascular musculature against the effects of chronic hypoxia and would explain their resistance to the development of pulmonary hypertension. In contrast, the newborn sheep present pulmonary vascular remodeling (Llanos et al., 2011b) and produce hypoxic pulmonary constrictor responses. These observations support our results, since we did not observe a significant increase in eNOS levels in llamas (**Figure 3C**), which would indicate an insufficient amount of NO concentrations and, therefore, reinforce the importance of the activation of the HO-carbon monoxide system in these animals.

In this study, we determined the concentrations of ADMA and homocysteine for the first time in an animal genetically adapted to live at high altitudes. Data showed that llamas had a lower ADMA and homocysteine concentration compared to sheep and other described species (Böger et al., 2000; Lüneburg et al., 2016; **Figure 2A**). It is widely known that ADMA and homocysteine are risk factors for cardiovascular diseases (Wierzbicki, 2007; Böger et al., 2009). We have recently described that hypoxia induces an increase in the concentration of homocysteine and ADMA in hypoxia-induced hypertensive rats (López et al., unpublished). Therefore, based on our study, we hypothesize that low concentrations of ADMA and homocysteine ensure a baseline level of NO synthesis, which together with activation of CO production, results in an adaptation system that allows llamas to live in hypobaric hypoxia conditions without developing pulmonary hypertension.

One explanation for the low levels of ADMA and homocysteine found in llamas could be due to the low expression of key enzymes for the regulation of their synthesis. According to currently available literature, key enzymes in the metabolism of methionine, and therefore in the production of ADMA and homocysteine, have been described as "target molecules" due to the presence of polymorphisms that lead to a deregulation in their metabolism and to the development of cardiovascular diseases. Among these enzymes are cystathionine β-synthase (CBS) (Zhang et al., 2014), methylenetetrahydrofolate reductase (MTHFR) (Yang et al., 2014; Heifetz and Birk, 2015) and dimethylaminohydrolase (DDAH-2) (Xuan et al., 2016). The expression and activity levels of these enzymes in llamas and whether they play an important role as therapeutic targets in the hypertensive processes induced by hypoxia remains unknown. Our results show that low expression levels of arginase type II mRNA, associated to the low concentrations of ADMA and Hcy found in llamas, suggest that the NOarginase pathway is not involved in the resistance to the development of altitude- induced pulmonary hypertension (**Figures 2**, **3**).

Additionally, hypoxia did not increase ADMA concentrations which could suggest that, unlike other species, the gene for dimethyl-aminohydrolase-2 (DDAH-2) in llamas does not contain hypoxia response elements such as the hypoxia-inducible factor (HIF-1α) (Pekarova et al., 2015). Contrary to what could be expected in llamas, we found a significant reduction in DDAH-2 expression in rats intolerant to the development of hypoxiainduced hypertension, which could explain the observed increase in ADMA.

Newborn llamas express at least 10 times less arginase type II which is not activated by hypoxia (**Figure 3**). It is widely known that one of the causes of endothelial dysfunction is decreased NO synthesis, which could be explained by a reduction in the concentration of L-arginine required for its synthesis. For this reason, it was shown that an increase in the expression of arginase would be involved in the development of hypertension, since it would compete with eNOS for the use of L- arginine (Durante et al., 2007; Pernow and Jung, 2013). Based on our results, we propose that the lower arginase type II levels leads to greater bioavailability of L-arginine for the synthesis of NO (**Figure 4**) which would explain the resistance to the development of hypertension induced by hypoxia in newborn llamas. In other words, the lower expression of arginase type II in these animals may constitute a mechanism for adaptation at geographical altitudes.

### REFERENCES


Finally, both ADMA and homocysteine are inhibitors of arginase type II in llamas in both lowland and highland conditions (**Figures 5**, **6**). On the contrary, homocysteine from sheep is an activator of arginase (**Figure 6B**). In agreement with previous studies, we found that homocysteine has the same activating effect on arginase II from hypoxia-induced hypertensive rats (López et al., unpublished data). Therefore, the inability of homocysteine and ADMA to activate arginase type II in llamas could represent a protection mechanism in llamas, which would prevent the increased expression of arginase type II in hypoxic conditions in these animals (**Figure 6B**).

Based on our results, we propose that low concentrations of ADMA and homocysteine found in llamas, and also the low arginase type II levels, could constitute an adaptation mechanism that these animals present when faced with hypoxic conditions by impeding the reduction in L-arginine, and thus ensuring baseline synthesis of NO during hypoxia (**Figure 7**).

### AUTHOR CONTRIBUTIONS

EU supervises the overall the study. Contributed to sample and data collections. The authors drafted the report. All authors contributed to the interpretation of the results, critical revision of the manuscript and approved the final manuscript.

### FUNDING

The research leading to these results has been supported by FONDECYT 11075096 and 1140647.


hypertension. Am. J. Physiol. Regul. Integr. Comp. Physiol. 288, R1057–R1062. doi: 10.1152/ajpregu.00758.2004


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 López, Moraga, Llanos, Ebensperger, Taborda and Uribe. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Effect of Acute, Subacute, and Repeated Exposure to High Altitude (5050 m) on Psychomotor Vigilance

Matiram Pun1,2, Sara E. Hartmann1,2, Michael Furian<sup>3</sup> , Adrienna M. Dyck1,2,4 , Lara Muralt<sup>3</sup> , Mona Lichtblau<sup>3</sup> , Patrick R. Bader<sup>3</sup> , Jean M. Rawling<sup>5</sup> , Silvia Ulrich<sup>3</sup> , Konrad E. Bloch<sup>3</sup> and Marc J. Poulin1,2,4,6,7,8 \*

<sup>1</sup> Department of Physiology and Pharmacology, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>2</sup> Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>3</sup> Pulmonary Division, Sleep Disorders Centre and Pulmonary Hypertension Clinic, University Hospital Zurich, Zurich, Switzerland, <sup>4</sup> Faculty of Kinesiology, University of Calgary, Calgary, AB, Canada, <sup>5</sup> Department of Family Medicine, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>6</sup> Libin Cardiovascular Institute of Alberta, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>7</sup> O'Brien Institute for Public Health, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>8</sup> Department of Clinical Neurosciences, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada

Aim: High altitude (HA) hypoxia may affect cognitive performance and sleep quality. Further, vigilance is reduced following sleep deprivation. We investigated the effect on vigilance, actigraphic sleep indices, and their relationships with acute mountain sickness (AMS) during very HA exposure, acclimatization, and re-exposure.

### Edited by:

Jean-Paul Richalet, Université Paris 13, France

### Reviewed by:

Simona Mrakic-Sposta, Istituto di Bioimmagini e Fisiologia Molecolare (IBFM), Italy Alessandro Tonacci, Istituto di Fisiologia Clinica (IFC), Italy

> \*Correspondence: Marc J. Poulin poulin@ucalgary.ca

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 05 February 2018 Accepted: 15 May 2018 Published: 04 June 2018

### Citation:

Pun M, Hartmann SE, Furian M, Dyck AM, Muralt L, Lichtblau M, Bader PR, Rawling JM, Ulrich S, Bloch KE and Poulin MJ (2018) Effect of Acute, Subacute, and Repeated Exposure to High Altitude (5050 m) on Psychomotor Vigilance. Front. Physiol. 9:677. doi: 10.3389/fphys.2018.00677 Methods: A total of 21 healthy altitude-naive individuals (25 ± 4 years; 13 females) completed 2 cycles of altitude exposure separated by 7 days at low altitude (LA, 520 m). Participants slept at 2900 m and spent the day at HA, (5050 m). We report acute altitude exposure on Day 1 (LA vs. HA1) and after 6 days of acclimatization (HA1 vs. HA6). Vigilance was quantified by reaction speed in the 10-min psychomotor vigilance test reaction speed (PVT-RS). AMS was evaluated using the Environmental Symptoms Questionnaire Cerebral Score (AMS-C score). Nocturnal rest/activity was recorded to estimate sleep duration using actigraphy.

Results: In Cycle 1, PVT-RS was slower at HA1 compared to LA (4.1 ± 0.8 vs. 4.5 ± 0.6 s−<sup>1</sup> , respectively, p = 0.029), but not at HA6 (4.6 ± 0.7; p > 0.05). In Cycle 2, PVT-RS at HA1 (4.6 ± 0.7) and HA6 (4.8 ± 0.6) were not different from LA (4.8 ± 0.6, p > 0.05) and significantly greater than corresponding values in Cycle 1. In both cycles, AMS scores were higher at HA1 than at LA and HA6 (p < 0.05). Estimated sleep durations (TST) at LA, 1st and 5th nights were 431.3 ± 28.7, 418.1 ± 48.6, and 379.7 ± 51.4 min, respectively, in Cycle 1 and they were significantly reduced during acclimatization exposures (LA vs. 1st night, p > 0.05; LA vs. 5th night, p = 0.012; and 1st vs. 5th night, p = 0.054). LA, 1st and 5th nights TST in Cycle 2 were 477.5 ± 96.9, 430.9 ± 34, and 341.4 ± 32.2, respectively, and we observed similar deteriorations in TST as in Cycle 1 (LA vs. 1st night, p > 0.05; LA vs. 5th night, p = 0.001; and 1st vs. 5th night, p < 0.0001). At HA1, subjects who reported higher AMS-C scores exhibited slower PVT-RS (r = −0.56; p < 0.01). Subjects with higher AMS-C scores took longer time to react to the stimuli during acute exposure (r = 0.62, p < 0.01) during HA1 of Cycle 1.

Conclusion: Acute exposure to HA reduces the PVT-RS. Altitude acclimatization over 6 days recovers the reaction speed and prevents impairments during subsequent altitude re-exposure after 1 week spent near sea level. However, acclimatization does not lead to improvement in total sleep time during acute and subacute exposures.

Keywords: altitude, psychomotor vigilance task, actigraphy, sleep, hypoxia, brain

### INTRODUCTION

fphys-09-00677 May 31, 2018 Time: 17:23 # 2

High altitude (HA) exposure is associated with low blood oxygen saturation due to the decreased partial pressure of oxygen as a result of reduced barometric pressure (West, 1996). Compared to slow acclimatizing trekking, rapid ascent to HA (via motor vehicle or air travel, for example) leads to profound hypoxemia and related pathophysiological consequences (Murdoch, 1995; West, 2008; Bloch et al., 2009; Strapazzon et al., 2015). The physiological responses to HA exposure include heart rate elevation, increased ventilation, and diuresis and fatigue on exertion (Nussbaumer-Ochsner and Bloch, 2007; Richalet et al., 2012; Luks, 2015). Failure to acclimatize due to rapid ascent leads to acute altitude illness such as acute mountain sickness (AMS) (Hackett and Roach, 2001; Bartsch and Swenson, 2013; Luks et al., 2017). The continuous gain in elevation and prolonged exposure to altitude can lead to adverse neurological consequences (Basnyat et al., 2004; Wilson et al., 2009; Strapazzon et al., 2014; Phillips et al., 2017) and possibly impaired cognitive function although this has not been consistently demonstrated (Latshang et al., 2013; Roach et al., 2014; Beer et al., 2017).

Sleep disturbance is reported commonly during altitude exposure (Johnson et al., 2010; Latshang et al., 2013; Bloch et al., 2015) due to hypoxia and new environmental conditions such as cold, uncomfortable sleeping quarters, and disturbed sleep due to increased frequency of urination (Windsor and Rodway, 2012). Further, exposure to HA is associated with a higher incidence of periodic breathing (Bloch et al., 2010; Ainslie et al., 2013), a disturbance of ventilatory control commonly observed in HA sojourners. Periodic breathing is characterized by waxing and waning ventilation, punctuated with periods of hyperventilation and then central apneas or hypopneas. Affected individuals often suffer from frequent arousals and poor sleep (Nussbaumer-Ochsner et al., 2012).

Poor sleep at altitude may affect daytime performance, especially of those tasks that require sustained attention and delicate fine-motor dexterity. The speed of response to light stimuli in the psychomotor vigilance test (PVT), a validated indicator of alertness, is slower following sleep deprivation (Van Dongen and Dinges, 2005; Lim and Dinges, 2008). Several studies have reported negative effects of sleep-disordered breathing (Kim et al., 2007; Lee et al., 2011; Kitamura et al., 2017), sleep deprivation (Van Dongen and Dinges, 2005; Lim and Dinges, 2008), and HA exposure (Latshang et al., 2013; Roach et al., 2014) on psychomotor vigilance tasks. Hence, these conditions could have deleterious consequences for individuals at HA performing work that demands continuous attention, decisionmaking, delicate handling of tools, and heavy-duty machinery work.

The Atacama Large Millimeter/submillimeter Array (ALMA) Observatory at 5050 m above sea level (asl) in Chile provides a unique pattern of repeated high-altitude exposure. Scientists from all over the world visit the site periodically to conduct scientific observations, experiments, and data collection. Further, lowland-native Chileans work there in various capacities, including equipment maintenance, security operations, scientific data collection, and reporting. A normal commute schedule includes typically, flight to Calama (2900 m asl) from Santiago (520 m asl) and travel about 2 h by bus to the ALMA basecamp (2900 m asl). From there, for a week at a time they ascend to the ALMA Observatory. Workers then spend the day working at HA and return to basecamp by motor vehicle in the evening. These workers suffer from periodic hypoxia (∼6–8 h/day) at very HA, and then sleep in a hypoxic environment at HA. After a week of HA work, the workers descend to their altitude of residence near sea level to rest, before returning to work at HA for another cycle. This pattern repeats itself continuously. How this type of HA exposure affects immediate and long-term health, and consequently work performance, is unknown. Both poor sleep and hypobaric hypoxia at altitude may be related to impaired vigilant attention, thereby affecting daytime task performance.

To investigate the impact of very HA exposure on vigilant attention, we brought a group of young, healthy, altitude-naive individuals to the ALMA Observatory, Chile. We mimicked the pattern of hypoxia exposure that is experienced by workers, in order to study the acute and subacute effects of repeated very HA exposure over two work-shift cycles (i.e., approximately 4 weeks). We had three specific goals in this high-altitude research expedition. First, we wanted to study the effect of acute, subacute, and re-exposure to very HA on PVT reaction speed (PVT-RS) in healthy, young, altitude-naive individuals. Second, we wanted to assess relationships between PVT-RS, AMS-Cerebral (AMS-C) score, blood oxygen saturation and sleep indices as measured by actigraphy. Third, we wanted to explore whether acclimatization obtained in previous exposure is retained during subsequent exposures after 1 week of rest at low altitude (LA). We hypothesized that immediate exposure to very HA would be associated with a slower PVT-RS, worsened sleep indices and increased prevalence of AMS. We further hypothesized that these parameters would improve upon acclimatization, and would be less affected with repeated exposure to altitude.

### MATERIALS AND METHODS

### Study Design and Setting

The study was carried out in Atacama Desert of northern Chile. Baseline measurements were taken in Santiago (520 m asl). After baseline measurements, the study participants traveled

approximately 2 h by air followed by a 2-h bus to the ALMA Operation Support Facility (ASF; 2900 m asl), which serves as ALMA's basecamp. Participants slept at the ASF for 7 nights and ascended by motor vehicle (∼45 min) to the AOS/ALMA to spend ∼7 h (range 4–8 h) per day at very HA. The ascent profile over 3 weeks from Santiago to the AOS has been illustrated in **Figure 1**. Altitude PVT measurements were taken at the ALMA Operation Site (AOS; 5050 m asl) located on the Chajnantor Plateau of the Atacama Desert. Actigraphy was performed over the span of a week cycle at 5050 and 2900 m sleep altitude, respectively. There were two 7-day cycles of HA exposure, separated by a break of 7 days at LA (Santiago, 520 m asl) (see schedule illustrated in **Figure 1**). The study participants' high-altitude exposure schedule was designed to emulate that of workers at the ALMA Observatory.

### Participants

A total of 21 healthy altitude-naive individuals were recruited for the expedition. A total of 18 subjects were residents of Calgary and surrounding area (altitude 1100 m asl) in Canada and were screened for inclusion in the study at the Foothills Medical Center, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada. Three subjects were from Zurich and surrounding area (altitude 490 m asl) in Switzerland and were screened at University Hospital of Zurich. Exclusion criteria included previous altitude intolerance < 3000 m, pregnancy and health impairment, which requires regular treatment. The screening criteria adhered to the ALMA HA Medical Examination guidelines. All participants provided written informed consent. The study was approved by the Conjoint Health Research Ethics Board of The University of Calgary (Ethics ID: REB 15-2709) and the Cantonal Ethics Committee of Zurich (2016-00048) and was registered at ClinicalTrials.gov (NCT02731456).

# Data Collection

The data collection was conducted over nine separate testing sessions before, during, and after two 7-day cycles at HA. The baseline and recovery measurements were taken at Santiago, Chile (520 m asl) while altitude data were collected at ALMA Observatory site (5050 m asl). Participants were familiarized with the measurements upon arrival in Santiago, whereas baseline measurements were performed on the day after. The objective of familiarization session was to make study participants familiar with experimental sessions, scoring sheets, instrumentation, and nature of invasiveness. The sleep data, acquired using actigraphy, were collected at the sleep altitudes (Santiago and the ASF) throughout the cycle, i.e., every night. The participants wore Actiwatches continuously throughout the expedition cycles unless otherwise due to technical and logistical reasons. The experimental sessions or specific data collection time points during HA exposure on the specific days have been illustrated in **Figure 1**.

# Psychomotor Vigilance Test and Trail Making Test

Since speed of psychomotor response to light stimuli is adversely affected following sleep deprivation (Sforza et al., 2004; Dorian et al., 2005; Basner and Dinges, 2011; Basner et al., 2011; Batool-Anwar et al., 2014), we utilized PVT-RS as a tool in assessing daytime function at altitude. Psychomotor vigilance was determined using a standard 10-min duration PVT assessment, previously described (Dorian et al., 2005; Drummond et al., 2005; Basner and Dinges, 2011). Time to react to randomly generated light stimuli was measured using a handheld Multiple Unprepared Reaction Time Test (MURT) device (Latshang et al., 2013). Several responses were obtained within a 10-min period, and the outcome was the mean reciprocal value of the RT (i.e., 1/RT). The PVT is a well-validated tool for vigilant attention deficits as a measure of neurobehavioral reaction speed due to sleep loss or deprivation (Basner and Dinges, 2011). The trail making tests (TMT) were administered in pen and paper version which contained circles containing numbers (Allen and Haderlie, 2010). The participants were instructed to connect circled numbers in sequence as quickly as possible without making an error. Whenever an error was made, the participants were asked to start again from the point where they made the error and continue. The times to complete the TMT were recorded in seconds.

# Sleep Monitoring: Actigraphy

Wrist actigraphy was recorded to estimate nocturnal rest as a surrogate of sleep (Latshang et al., 2016) (Actigraph 2.0, Minimitter; Philips Respironics, Murrysville, PA, United States). All participants wore an actimeter for a total of 9 days, including the baseline and recovery period at LA (520 m asl, Santiago) and 7 days at HA (2900 m asl, ALMA Operation Support Facility). The actigraphic data collected from the field were analyzed by the manufacturer's guidelines and dedicated software (Respironics Actiware, Version 6.0.4). The data were exported and assessed by research team members experienced with Actigraphy. The variables analyzed from the actigraphy included total time in bed (i.e., time from lights-off to lights-on which each participant pressed marker switch in the Actiwatch and also recorded in the sleep diary), total estimated sleep time (TST; i.e., the sum of all epochs with activity below threshold), sleep efficiency (i.e., TST in percentage of time in bed), and sleep latency (i.e., time from lights-off to the beginning of the first three consecutive epochs with activity below the threshold) as described in the previous literature (Latshang et al., 2016). The definitions of sleep parameters such as awakenings and data analysis was carried out using the Actiware scoring algorithm as described previously (Drogos et al., 2016). The participants wore Actiwatches throughout the expedition.

# Acute Mountain Sickness Assessment and Vital Signs

Acute mountain sickness (AMS) was assessed using the Lake Louise Score (LLS) (Roach et al., 1993) and Environmental Symptom Questionnaire (ESQ) – Cerebral (AMS-C subscore)

(Sampson et al., 1983). AMS diagnosis was made based on the LLS score of ≥ 5, and ESQ AMS-C weighted average score ≥ 0.7 (Maggiorini et al., 1998; Dellasanta et al., 2007; Nussbaumer-Ochsner et al., 2011). We have used AMS-C score for the analysis and interpretation. Handgrip (HG) strength was measured in the dominant hand using a strain-gauge dynamometer following manufacturer's guideline (Lafayette, IN, United States). The HG peak strength has been reported in kilograms (kg). The resting arm blood pressure was recorded with automatic blood pressure monitor (Omron Healthcare, Inc., United States) in a comfortable sitting position. Blood oxygen saturation and heart rate were measured with a pulse oximetry.

### Data Analyses and Statistics

The data collected at different time points during the HA expedition have been expressed as the mean ± standard deviation (mean ± SD). The effects of altitude [baseline at sea level vs. HA exposure at Day 1 (i.e., acute exposure; HA1) and HA exposure at Day 6 (i.e., acclimatization; HA6)] were assessed with oneway repeated measures of analysis of variance (ANOVA) for each cycle independently. The sleep parameters were analyzed over different time points of sleep at HA (i.e., number of nights). In the second step of the analysis, three time-points (baseline, acute exposure, and acclimatization effects) were compared with repeated measures ANOVA to tease out acute (1st night sleep at HA) and acclimatization effects (fifth-night sleep) of altitude exposure in each cycle separately. Bivariate correlations were tested between variables of interest (reaction speed or changes in the reaction time with AMS score) using Pearson's correlation coefficient. Multivariate regression analysis was performed to identify independent variables (SpO2, AMS-C score, and TST) associated with PVT-RS at HA1 and HA6 of both cycles to explore independent as well as interaction effects. The results were considered significant when the value of alpha was less than 0.05 (p < 0.05) after Bonferroni's post hoc corrections, as appropriate. The recovery values were compared with baseline and HA values with two-tailed paired t-tests wherever necessary to explore the recovery status of the individuals. All the data analyses were carried out using the Statistical Package for the Social Sciences (SPSS), Version 24, IBM Corporation, United States.

## RESULTS

A total of 21 healthy altitude-naive individuals aged, mean ± SD, 25.3 ± 3.8 years (8 males, 13 females) with body mass index of 22.8 ± 3.0 kg/m<sup>2</sup> were recruited. The baseline characteristics such as weight (65.4 ± 9.3 kg), body mass index (22.7 ± 3.1 kg/m<sup>2</sup> ), heart rate (65.5 ± 9.2 bpm), blood pressure (SBP = 114.2 ± 8.1 mmHg, DBP = 69.9 ± 6.2 mmHg, MAP = 84.7 ± 5.8 mmHg), peak HG strength (39.4 ± 13.1 kg) were taken at Santiago, Chile (520 m asl). The absoulte values from acute (HA1), acclimatization (HA6), and recovery (REC) sessions have been summarized in **Table 1**.

The sustained attention test (i.e., PVT) assessed by the MURT test device has been illustrated in **Figure 2**. The number of mistakes/lapses was not statistically significant during HA exposure although there was a trend to increase with acute

exposure in both cycles. The individuals took a longer time to react (increase in mean and median reaction times) at HA during acute exposure (HA1), and they improved during acclimatization visit (HA6) in Cycle 1, but they were not significantly prolonged in Cycle 2 compared to their respective baselines. The PVT-RS was impaired during acute exposure but improved in subsequent exposure. The PVT-RS was unaffected during Cycle 2 (i.e., individuals were acclimatized in Cycle 1, as shown with values at HA6 and retained it during Cycle 2). The fastest 10% reaction times (i.e., optimum response times) did not significantly change at HA in both cycles but the slowest 10% reaction times (i.e., response times in the lapse domain) increased during acute exposure (HA1) and improved during acclimatization visits (HA6) in both cycles. The PVT parameters from baseline, acute exposure (HA1), acclimatization (HA6), and recovery (REC) have been illustrated in **Figure 2**. We also incorporated the trail-making test (TMT) which did not get worse during acute exposures (HA1) but improved during the acclimatization exposures (HA6) compared to baseline in both Cycles. There was a significantly higher incidence of AMS during the first exposure but not during the acclimatization exposure with both scoring criteria of AMS [i.e., LLS and ESQ (AMS-C)] as shown in **Table 2**.

Actigraphy results are summarized in **Table 3**. Due to technical difficulties, data are given only up to the 5th night for 18 individuals in Cycle 1 and 14 individuals in Cycle 2 for the entire expedition (i.e., 9 nights including baseline, 7 nights at altitude and recovery) (**Table 3**). To explore the acute and acclimatization effects of altitude upon sleep indices, the baseline was compared with acute exposure (HA1) sleep and after acclimatization (HA6) sleep in each cycle as illustrated in **Figure 3**. In Cycle 1, the altitude had an adverse effect on total sleep time at the 5th night (i.e., after the acclimatization days). The individuals tended to spend less time in bed (p = 0.078, 1st vs. 5th nights) while wakefulness after sleep-onset time tended to increase (p = 0.057, baseline vs. 5th night) during the acclimatization night although they were not statistically significant (**Figures 3A** and **B**, Cycle 1). Interestingly, during Cycle 2, participants spent less time in bed (p = 0.001, baseline vs. 5th night; p < 0.001, 1st vs. 5th nights) and had decreased total sleep time (p = 0.001, 1st vs. 5th nights; p < 0.001, 1st vs. 5th nights) during prolonged exposure (5th night) compared to baseline and 1st night at altitude (**Figure 3C**). To find out the break point for the sleep disturbances observed during acclimatization exposure (5th night), we analyzed sleep indices of all the nights (N1, N2, N3, N4, and N5) with baseline (BL) using one-way repeated measures ANOVA (time<sup>∗</sup> 6). The changes in total sleep time (TST) started on the 3rd night at altitude in both cycles – Cycle 1 (BL: 431.3 ± 28.7 vs. N3: 369.7 ± 54.3, p = 0.008 and N1: 418.1 ± 48.6 vs. N3: 369.7 ± 54.3, p = 0.062) and Cycle 2 (BL: 477.5 ± 96.9 vs. N3: 368.9 ± 61.5, p = 0.097 and N1: 430.9 ± 34 vs. N3: 368.9 ± 61.5, p = 0.021). The findings were further supported with a series of repeated measures ANOVA analyses (time<sup>∗</sup> 3) as baseline (BL), acute exposure (N1, 1st night at altitude), and respective nights (N2, N3, N4, and N5). We observed significant changes in TST from 3rd night at altitude onward in both cycles – Cycle 1 (BL: 431.3 ± 28.7 vs. N3: 369.7 ± 54.3, p = 0.002 and N1:418.1 ± 48.6 vs. N3: 369.7 ± 54.3, p = 0.0125) and Cycle 2 (BL: 477.5 ± 96.9 vs. N3: 368.9 ± 61.5, p = 0.004 and N1:430.9 ± 34 vs. N3: 368.9 ± 61.5, p = 0.004).

Finally, the acute exposure to HA is associated with increased incidence of AMS which wears off with acclimatization (**Table 2**). The incidence of AMS decreased by > 50% during Cycle 2 (HA1) exposure compared to Cycle 1 (HA1). Next, the AMS-C score was negatively correlated (r = −0.56, p < 0.01) with PVT-RS. Altitude adversely affected PVT-RS during acute exposure and was associated with increased AMS-C scores as illustrated in **Figure 4A**. The mean reaction time changes (mean reaction at altitude – mean reaction time at baseline) were positively correlated (r = 0.62, p < 0.01) with AMS-C scores i.e., higher AMS-C scoring individuals took a longer time to react at HA during acute exposure as shown in **Figure 4B**. The individuals recovered during the acclimatization phase (i.e., there was no such correlation among these parameters); this recovery was sustained during Cycle 2 even after 1 week of rest at LA. Multiple regression analyses revealed that AMS-C score (p = 0.01) but not the total sleep time and blood oxygen saturation was the independent predictor of PVT-RS during acute HA exposure (HA1) of Cycle 1 (**Table 4**). However, the association was abolished during acclimatization (HA6) and re-exposures



The descriptive characteristics of the 21 altitude-naive healthy individuals, who joined the high altitude expedition, born and raised at an altitude of ≤ 1100 m asl. Values are means ± SD. SL, sea level; HA1, high altitude at day 1; HA6, high altitude day 6; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial blood pressure; HR, heart rate; mmHg, millimeter of mercury; bpm, beats per minute; HG, handgrip; kg, kilogram; SD, standard deviation; kg, kilogram; m, meter. <sup>a</sup>n = 20, Sex: Males (8) and Females (13) and average age: 25.3 ± 3.8 years.

FIGURE 2 | Effect of altitude exposure on PVT parameters during acute and acclimatization exposures. It illustrates changes in different PVT parameters during high altitude exposure. Four important time points of the expedition are illustrated (BL, baseline at LA; HA1, acute high altitude exposure; HA6, acclimatization exposure of both cycles; REC, recovery). The Y-axis depicts sleep parameters with the mean and standard deviation (Mean ± SD) of the PVT parameters. The X-axis depicts different expedition time points (BL1, HA1, HA6, and REC of Cycle 1 and Cycle 2). Solid bars represent Cycle 1 exposure while empty bars represent Cycle 2 exposure. Here, panel A illustrates number of errors/lapses while panel E represents PVT reaction speed. Panels B and F represent mean and median of reaction times respectively. Panels C and G show fastest and slowest 10% reaction times while D and H depict reciprocals of them in respective orders. The significance shown with dashed line indicates baseline comparisons of two cycles. The comparison with solid lines is between different time points of respective cycles. n, number; RT, reaction time; ms, milliseconds; s−<sup>1</sup> , per second; PVT-RS, psychomotor vigilance test reaction speed; p, level of significance; p, ∗ level of significance value with less than 0.05.

TABLE 2 | Oxygen saturation, acute mountain sickness and trail making tests during acute and acclimatization exposures.


Values are means ± SD. SL, sea level; HA1, high altitude at day 1; HA6, high altitude day 6; n, numbers; SpO2, blood oxygen saturation; LLS, Lake Louise Score; AMS, acute mountain sickness; AMS-C, acute mountain sickness – cerebral; ESQ, Environmental Symptom Questionnaire. <sup>a</sup>n = 20; <sup>∗</sup>p < 0.05: Different from respective baseline SL; †p < 0.05: Different from respective HA1.

(Cycle 2). The clinical outcome measures were returned to the baseline during recovery period, i.e., after returning to sea level.

## DISCUSSION

### Major Findings

Several novel findings emerged from this study. First, the PVT-RS was reduced with acute exposure to HA and normalized with acclimatization over 6 days. Further, acclimatization effects were retained upon re-exposure to altitude even after spending a week at sea level. Second, there were 62% and 29% of individuals suffering from AMS in Cycle 1 and Cycle 2, respectively, with an immediate HA exposure. However, after 6 days of acclimatization, the incidence of AMS was significantly reduced to none and 10% in Cycle 1 and Cycle 2, respectively. The incidence of AMS, during acute exposures (HA1), was decreased by more than 50% in Cycle 2 compared to Cycle 1. Third, total estimated sleep time were decreased during Cycle 2. Fourth, the AMS score was associated with worsened PVT performance (decrease in reaction speed and increase in mean reaction time). This appears to be unrelated to estimated sleep duration and blood oxygen saturation. The altitude-naive individuals recovered upon descending to sea level after each cycle.

### The Psychomotor Vigilance Test: Daytime Alertness/Sustained Attention

Sleep-disordered breathing and intermittent hypoxia are associated with excessive daytime sleepiness, loss of focus, memory impairment, and difficulty with tasks that demand sustained attention (Morrell and Twigg, 2006; Dempsey et al., 2010; Beaudin et al., 2017a,b). The PVT, a validated neurocognitive assay for sustained vigilance attention (Dorian et al., 2005; Lim and Dinges, 2008), is impaired in patients with sleep-disordered breathing (Sforza et al., 2004; Kim et al., 2007), sleep-deprived individuals (Lim and Dinges, 2008; Basner and Dinges, 2011), circadian rhythm mismatch (Van Dongen and Dinges, 2005), and in professionals whose work demands continuous focus in challenging operational environments such as aviation pilots, field engineers, and military personnel (Gander et al., 2014, 2016; Shattuck and Matsangas, 2015; Song et al., 2017). The PVT reaction speed, lapses, mean, and median reaction times as well as fastest and slowest 10% reaction times are particularly sensitive in assessing alertness among sleepdeprived individuals (Dorian et al., 2005; Basner and Dinges, 2011). The PVT has also been tested at HA among healthy individuals, in whom no changes in PVT parameters were noted despite a considerable number of periodic breathing episodes and sleep disturbances (Latshang et al., 2013). Seemingly, contrary to this finding, we report impairments in mean/median reaction times, reaction speed, and reciprocal of slowest 10% reaction time during acute exposure (**Figure 2**) while sleeping at moderate altitude and traveling to very HA during the day. The discrepancy in these PVT outcomes might be due to the magnitude and duration of altitude exposure (i.e., overall study design difference). In the previous study (Latshang et al., 2013), the highest altitude gained was 2590 m asl whereas the sleeping altitude in our expedition was 2900 m asl and study participants were exposed to 5050 m asl daily for about 6–8 h per day. The PVT-RS during acute exposure was lower at ALMA (4.1 s−<sup>1</sup> ) vs. 2590 m (5.0 s−<sup>1</sup> ). The PVT changes improved after acclimatization over 6 days – the acclimatization effect was retained during Cycle 2 exposure after spending a week at sea level (except for the reciprocal of slowest 10% reaction time that followed a similar pattern as in Cycle 1) (**Figure 2**). The TMT showed improvement over time at altitude, which might be due to learning effects with acclimatization (Pagani et al., 1998; Buck et al., 2008).

### Sleep at Altitude: Sleep Indices From Actigraphy

We assessed sleep with wrist actigraphy (Sadeh, 2005) which has also been used at HA, having been validated against the gold standard technique of polysomnography (Latshang et al., 2016). The sleep indices from actigraphy in Cycle 1 (baseline and subsequent five nights at altitude) and Cycle 2 (baseline, seven nights at altitude, and recovery at sea level) were not significantly different across all time points (**Table 3**). However, on comparing three time points (i.e., baseline, acute, and acclimatization exposure), the total sleep time was impaired during the 5th night at altitude compared to the baseline and 1st night at altitude in Cycle 1. Similarly, the time in bed


TABLE 3 |Actigraphyoutcome of sleep parameters during

fphys-09-00677 May 31, 2018 Time: 17:23 # 8

FIGURE 3 | Effect of altitude on actigraphy sleep indices during acute and acclimatization exposure. It has two panels as Cycle 1 (Left, A,B) and Cycle 2 (Right, C,D). The Y-axis depicts sleep parameters with the mean ± standard deviation (Mean ± SD) at baseline, acute exposure (1st night altitude), and acclimatization night (5th night at altitude). The X-axis depicts different sleep parameters as reported from actigraphy during the first expedition of Cycle 1 and Cycle 2. Black bars, baseline sleep at Santiago, Chile (520 m asl); gray bars, acute exposure to altitude, i.e., 1st night sleep at high altitude (2900 m asl) and empty bars, sleep after acclimatization at high altitude, i.e., 5th night sleep at high altitude in each cycle. The significance shown with dashed line indicates baseline comparisons of two cycles. The comparison with solid lines is between different time points of respective cycles. TIB, time in bed; TST, total sleep time; SL, sleep latency; WASO, wake after sleep onset; Awak, awakenings; SE, sleep efficiency; SD, standard deviation; min, minute; p, level of significance; n, number; %, percentage; <sup>∗</sup>p, level of significance value with less than 0.05.

tended to decrease on the 5th night compared to the 1st night, and wakefulness after sleep onset tended to increase on the 5th night compared to baseline (**Figure 3**). During Cycle 2, we found both time in bed and total sleep time were decreased on the 5th night compared to baseline and 1st night at altitude (**Figure 3**). It is noteworthy that in both cycles, we observed impaired total estimated sleep time (actigraphy), which is in agreement with previous reports of polysomnography parameters at altitude (Latshang et al., 2016). The sleep disturbance as reflected in TST appears to have significantly affected from 3rd night onward. The inflexion point of sleep disturbance may be related to the time domains of hypoxic ventilatory acclimatization at HA (Pamenter and Powell, 2016). Worsening sleep indices with acclimatization are in contrast to the improvements observed in PVT performances and blood oxygen saturation. The findings on sleep indices are consistent with Latshang et al. (2013) who found disturbance in nocturnal breathing and sleep during first four nights of altitude stay although they did not find similar changes in PVT performance. The PVT performance seems to be influenced by improvements in AMS symptoms and blood oxygen saturation rather than sleep disturbance. The decreased total sleep time index might be related to HA periodic breathing, which is a consequence of an exaggerated hypoxic ventilatory response (Lahiri et al., 1983). Alternatively, individuals might have been sleep deprived and exhausted in the 1st nights after the transfer

FIGURE 4 | Inverse correlation between reaction speed vs. AMS-C score and positive association with changes in reaction time (HA1 – BL) vs. AMS-C score. It has two panels, left panel (A) shows negative correlation between PVT-RS vs. AMS score and the right panel (B) shows positive correlation between changes in reaction time (HA1 – BL) vs. AMS score. The Y-axis depicts reaction speed during the first exposure at altitude in A and the changes in reaction times (HA1, reaction time of high altitude at Day 1 – BL, reaction time at baseline) in B during acute exposure at altitude during Cycle 1. In both panels, the X-axis depicts AMS-C score during the acute exposure Cycle 1. The vertical dashed red lines in both panels separates no AMS subjects from AMS with the cut-off of a ≥ 0.7 AMS-C score. In B, the horizontal dashed line passing through zero of Y-axis is a reference line separating negative (i.e., subjects with decreased reaction time at high altitude) and positive (i.e., subjects with increased reaction time at high altitude) values of changes in reaction times. PVT-RS, psychomotor vigilance test reaction speed; AMS, acute mountain sickness; AMS-C, acute mountain sickness – cerebral; BL, baseline at low altitude; HA1, acute high altitude exposure; s, seconds; ms, milliseconds; r, Pearson's correlation; p, level of significance.

from sea level and therefore might have required recovery sleep time.

### Acute Mountain Sickness and Blood Oxygen Saturation

We found very high AMS incidence (62%) with acute altitude exposure, which decreased with acclimatization (none) in Cycle 1, and the acclimatization benefit was observed in Cycle 2 as well (29% vs. 10%). The acclimatization gained in Cycle 1 reduced AMS incidence by > 50% during acute exposure of Cycle 2. The findings are consistent with previous studies that acclimatization reduces AMS incidence and severity but not entirely protective (Beidleman et al., 2004; Jackson et al., 2010; Muza et al., 2010; Schommer et al., 2010). We used both the LLS (Roach et al., 1993) and the AMS-C (Sampson et al., 1983) scores for the diagnosis of AMS (Maggiorini et al., 1998; Dellasanta et al., 2007), and they were in agreement with each other in most instances (**Table 2**). Similar to Nussbaumer-Ochsner et al. (2012), we found improved oxygen saturation after acclimatization, but we did not observe an improvement in sleep indices. The acute and acclimatization exposures influenced arterial blood pressures in both cycles compared to their respective baselines (**Table 1**). The peak HG strength changed during altitude exposures returned to closer to the baseline during acclimatization exposures (**Table 1**).

Higher AMS-C scores (i.e., more severe symptoms of AMS) were negatively correlated with the PVT-RS (**Figure 4A**) during acute exposure of Cycle 1. Hence, AMS seems to adversely affect alertness and performances that require sustained attention. Similarly, increasing reaction time (i.e., increased the change in reaction time) was positively correlated with the AMS-C score (**Figure 4B**); individuals who had higher AMS scores took longer to react to PVT stimuli. The findings suggest that the individuals suffering from AMS might struggle to perform delicate work tasks such as maintenance, scientific observations, and reporting, engineering and operation of heavy-duty equipment. Furthermore, multiple regression analysis revealed that except AMS-C score, the estimated sleep duration and blood oxygen saturation are not associated with PVT-RS (**Table 4**). The findings further support the previous evidence that hypoxia resulting from HA has global cerebral effects (Kramer et al., 1993; Wilson et al., 2009; Maa, 2010; Roach et al., 2014). The associations were abolished after acclimatization (HA6, i.e., during HA exposure at Day 6). Further, the acclimatization obtained in Cycle 1 was carried over during Cycle 2 even after 1 week of LA rest. These findings support the practice followed by the Chilean HA mining workers (Richalet et al., 2002; Farias et al., 2006). The assessment of AMS along with the incorporation of PVT could provide insight into the status of HA illness and its impact on alertness/fatigue and safety margins during HA work (Vearrier and Greenberg, 2011; Abe et al., 2014).

### Strengths and Limitations

This study was designed to replicate a typical work schedule and ascent profile of the employees at the ALMA Observatory (Chile) with two working cycles separated by a week of rest at LA. The 1-week break at LA is an attempt to minimize the ill effects of long-term sustained chronic hypobaric hypoxia (Richalet et al., 2002; Vearrier and Greenberg, 2011) and at the same time return to work while the acclimatization effects acquired in the previous exposure(s) are still retained (Farias et al., 2006; Muza et al., 2010). However, the extent to which this exposure profile is effective to optimize acclimatization while mitigating AMS symptoms has been explored for the first time. Hence, this study provides insight into the workers' schedule in relation to optimum schedule to maximize their work output while increasing work safety and reducing health


 regression β,sleepcerebral score; SpO2, blood oxygen saturation.

risks. Therefore, the results from this study have implications in reducing health hazards and improving work performance for individuals engaging in HA activities that require high daytime attentiveness such as astronomical observation, engineering, computing, scientific reporting, and operation of heavy-duty equipment.

Our study has a few limitations. First, our sample contains a relatively small group of 21 active young individuals. Due to logistical reasons, the expedition was designed only for two regular work cycles. Workers at the ALMA Observatory are older (ALMA workers' average age is between 35 and 40 years), and they have been working for longer durations (i.e., repeated cycles of HA exposure throughout the year). Second, we could only retrieve complete actigraphy data sets from 18 individuals during Cycle 1 (and up to 5 nights) and 14 individuals during Cycle 2 (complete cycle). Finally, we had relatively heterogeneous population with the mixture of nationalities (Canadians and Swiss) and their background of altitude exposure (Calgary, 1100 m asl; Zurich, 490 m asl).

# CONCLUSION

The present study reports novel findings from repeated exposures to very HA with sleep at HA interspersed with a week of break at low attitude. With acute exposure to HA, individuals suffered from decreased PVT performances, hypoxemia, and increased incidence of AMS. After acclimatization, the individuals revealed slightly less hypoxemia, a lower incidence of AMS and had improved sustained attention functions. The acclimatization effect acquired over a week of HA exposure was retained on a subsequent cycle of exposure even after spending 1 week at near sea level. The individuals with higher AMS-C scores had impaired psychomotor reaction speeds. The sleep indices, especially total estimated sleep time, were worsened during re-exposure, showing no influence of acclimatization on sleep, which is similar to the previous observations that sleep disturbances such as periodic breathing continue to occur (Bloch et al., 2010) despite improved clinical outcome parameters and PVT performances.

# AUTHOR CONTRIBUTIONS

MJP, KB, SU, JR, PB, ML, LM, AD, MF, and SH were involved in the study design, data collection, analysis, manuscript preparation, and submission. MP was involved in the data analysis and took lead in the literature review, manuscript preparation, and submission process.

# FUNDING

MJP was supported by a Discovery Grant from the Natural Sciences and Engineering Research Council (NSERC) of Canada (Principal Investigator, MJP; 2014-05554). MJP was also supported by a CIHR Operating Grant for Intermittent Hypoxia (Principal Investigator: MJP). SH received the support from the Dr. Chen Fong Doctoral Scholarship (Hotchkiss Brain Institute).

fphys-09-00677 May 31, 2018 Time: 17:23 # 11

MJP currently holds The Brenda Strafford Foundation Chair in Alzheimer Research. KB and SU were supported by Lunge Zurich, Swiss National Science Foundation.

### ACKNOWLEDGMENTS

fphys-09-00677 May 31, 2018 Time: 17:23 # 12

We would like to express our sincere thanks to the study participants who took part in the high altitude expedition

### REFERENCES


to ALMA Observatory, Chile. The research would have been impossible without their involvement and cooperation throughout the expedition. We thank Dr. Lauren L. Drogos, Ph.D., for her kind help in the analysis of Actigraphic data. We would also like to acknowledge our Chilean collaborators from the ALMA Head Office in Santiago, Chile, and the staff of the ALMA Observatory (Ivan Lopez, Daniel Soza, Charlotte Pon, and the Health and Safety and Polyclinic teams), who greatly facilitated the study.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Pun, Hartmann, Furian, Dyck, Muralt, Lichtblau, Bader, Rawling, Ulrich, Bloch and Poulin. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

fphys-09-00677 May 31, 2018 Time: 17:23 # 13

# Guinea Pig as a Model to Study the Carotid Body Mediated Chronic Intermittent Hypoxia Effects

Inmaculada Docio1,2,3, Elena Olea2,3,4, Jesus Prieto-LLoret1,2,3, Teresa Gallego-Martin1,2,3 , Ana Obeso1,2,3, Angela Gomez-Niño2,3,5 \* and Asuncion Rocher1,2,3 \*

<sup>1</sup> Departamento de Bioquímica y Biología Molecular y Fisiología, Universidad de Valladolid, Valladolid, Spain, <sup>2</sup> Instituto de Biología y Genética Molecular, Consejo Superior de Investigaciones Científicas, Universidad de Valladolid, Valladolid, Spain, <sup>3</sup> CIBER de Enfermedades Respiratorias, Instituto de Salud Carlos III, Madrid, Spain, <sup>4</sup> Departamento de Enfermería, Universidad de Valladolid, Valladolid, Spain, <sup>5</sup> Departamento de Biología Celular, Histología y Farmacología, Universidad de Valladolid, Valladolid, Spain

### Edited by:

Rodrigo del Rio, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Esteban A. Moya, University of California, San Diego, United States David D. Kline, University of Missouri, United States

### \*Correspondence:

Angela Gomez-Niño angela@biocel.uva.es Asuncion Rocher rocher@ibgm.uva.es

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 01 March 2018 Accepted: 18 May 2018 Published: 05 June 2018

### Citation:

Docio I, Olea E, Prieto-LLoret J, Gallego-Martin T, Obeso A, Gomez-Niño A and Rocher A (2018) Guinea Pig as a Model to Study the Carotid Body Mediated Chronic Intermittent Hypoxia Effects. Front. Physiol. 9:694. doi: 10.3389/fphys.2018.00694 Clinical and experimental evidence indicates a positive correlation between chronic intermittent hypoxia (CIH), increased carotid body (CB) chemosensitivity, enhanced sympatho-respiratory coupling and arterial hypertension and cardiovascular disease. Several groups have reported that both the afferent and efferent arms of the CB chemo-reflex are enhanced in CIH animal models through the oscillatory CB activation by recurrent hypoxia/reoxygenation episodes. Accordingly, CB ablation or denervation results in the reduction of these effects. To date, no studies have determined the effects of CIH treatment in chemo-reflex sensitization in guinea pig, a rodent with a hypofunctional CB and lacking ventilatory responses to hypoxia. We hypothesized that the lack of CB hypoxia response in guinea pig would suppress chemo-reflex sensitization and thereby would attenuate or eliminate respiratory, sympathetic and cardiovascular effects of CIH treatment. The main purpose of this study was to assess if guinea pig CB undergoes overactivation by CIH and to correlate CIH effects on CB chemoreceptors with cardiovascular and respiratory responses to hypoxia. We measured CB secretory activity, ventilatory parameters, systemic arterial pressure and sympathetic activity, basal and in response to acute hypoxia in two groups of animals: control and 30 days CIH exposed male guinea pigs. Our results indicated that CIH guinea pig CB lacks activity elicited by acute hypoxia measured as catecholamine (CA) secretory response or intracellular calcium transients. Plethysmography data showed that only severe hypoxia (7% O2) and hypercapnia (5% CO2) induced a significant increased ventilatory response in CIH animals, together with higher oxygen consumption. Therefore, CIH exposure blunted hyperventilation to hypoxia and hypercapnia normalized to oxygen consumption. Increase in plasma CA and superior cervical ganglion CA content was found, implying a CIH induced sympathetic hyperactivity. CIH promoted cardiovascular adjustments by increasing heart rate and mean arterial blood pressure without cardiac ventricle hypertrophy. In conclusion, CIH does not sensitize CB chemoreceptor response to hypoxia but promotes cardiovascular adjustments probably not mediated by the CB. Guinea pigs could represent an interesting model to elucidate the mechanisms that underlie the long-term effects of CIH exposure to provide evidence for the role of the CB mediating pathological effects in sleep apnea diseases.

Keywords: guinea pig, oxygen sensing, carotid body, chronic intermittent hypoxia, sympathetic activity, ventilation

### INTRODUCTION

fphys-09-00694 June 3, 2018 Time: 11:39 # 2

For the last two decades, chronic intermittent hypoxia (CIH) has been considered a paradigm of detrimental stimulus that leads to a number of comorbidities including autonomic, cardiovascular, metabolic and cognitive dysfunction (Dempsey et al., 2010). CIH is recognized as the main hallmark in obstructive sleep apnea (OSA). It has been reported that the hypoxia/reoxygenation and oxyhemoglobin desaturation associated with apnea produces oxidative stress, inflammation and sympathetic hyperactivity, generating endothelial dysfunction and secondary systemic hypertension (Dempsey et al., 2010; Iturriaga et al., 2017).

Carotid body (CB), the main arterial oxygen sensor, triggers reflex homeostatic adjustments to acute hypoxia and is also responsible for the acclimation to high altitude. Furthermore, recent clinical and experimental evidence suggests a positive correlation between CIH, increased CB responsiveness, enhanced sympatho-respiratory coupling and arterial hypertension and cardiovascular disease (Semenza and Prabhakar, 2015; Iturriaga et al., 2017). CB type I cells are specialized in rapid response to decreased oxygen in blood. Although cellular mechanisms underlying chemo-reflexes activation by hypoxia remain to be elucidated, it has been proposed that the repeated CB type I cells stimulation produced by CIH would induce CB sensitization, increasing secretory response and chemoreceptor input to the brainstem. It originates an exaggerated sympathetic reflex that promotes a rise of circulating catecholamine (CA) and finally, hypertension (Dempsey et al., 2010). Recent evidence points to oxidative stress as the key mediator of both the enhanced CB chemosensory response to hypoxia and the hypertension induced by CIH (Iturriaga et al., 2015; Semenza and Prabhakar, 2015).

Rats and cats exposed to several days of CIH show the phenotype developed in OSA patients such as increased arterial blood pressure, hematocrit and hemoglobin and left ventricular hypertrophy, (Fletcher et al., 1992; Rey et al., 2004). Disrupting the chemo-reflex pathway by bilateral section of the carotid sinus nerve (Fletcher et al., 1992; Lesske et al., 1997) or by selective ablation of the CB while preserving the carotid baroreceptor function (Peng et al., 2014), prevented the increase of plasma CA and hypertension in CIH treated rats (Nanduri et al., 2015; Del Rio et al., 2016; Iturriaga, 2017). These findings suggest that CIH directly distresses CB function, and that other effects on brainstem neurons could be mediated by the altered sensory input from CB. Nevertheless, therapeutic removal of CB in OSA patients can have adverse consequences, such as the loss of adaptation to high altitude, the maintenance of arterial blood gases during exercise, and the cardiorespiratory responses to acute hypoxia (Prabhakar, 2016).

Most hypertension research is conducted with rodents because of blood pressure control and cardiovascular response similarities to humans. Unlike other rodents, guinea pigs show a very poor ventilatory response to hypoxia while maintaining response to hypercapnia (Schwenke et al., 2007). Recently, we have reported that guinea pig CB has a small percentage of tyrosine hydroxylase positive type I cells and they lack O2-sensitive K<sup>+</sup> channels, which would be inhibited by hypoxia, and therefore, lacks hypoxia depolarization capacity and chemo-reflexes activation. It has also been reported that chronic sustained hypoxia (CSH) during 15 days does not modify CB activity (Gonzalez-Obeso et al., 2017). However, to date, the functionality of guinea pig CB exposed to CIH treatment has not been investigated, so we aimed to examine the long-lasting this treatment on guinea pig CB chemosensory activity, sympathetic output and its cardiorespiratory consequences. If guinea pig CB is not activated by hypoxia, CIH would not induce chemo-reflex sensitization and pathological effects derived from the CB hyperactivity would not be observed (Del Rio et al., 2016; Iturriaga, 2017). Adopting an integrative approach combining studies in vivo with experiments in isolated CB, we tested the hypothesis that the systemic arterial hypertensive and sympathetic hyperactivity response to CIH in guinea pig relative to other rodents would be diminished due to the CB hypofunctionality. Accordingly, we have analyzed the CB functionality, sympathetic activity and the cardiopulmonary responses to hypoxia after 30 days of CIH exposure. The ultimate purpose of this study was to determine if guinea pigs could be a model to disclose the CB dependent and non-dependent CIH effects.

### MATERIALS AND METHODS

All experiments were carried out in compliance with the international laws and policies [European Union Directive for Protection of Vertebrates Used for Experimental and Other Scientific Ends (2010/63/EU)], and were reviewed and approved by the University of Valladolid Institutional Committee for Animal Care and Use.

### Animals

Experiments were performed on adult male Hartley guinea pigs (3–6 months old) with free access to standard chow and water and maintained under controlled conditions of temperature, humidity and a stationary 12 h light–dark cycle. Animals were randomly distributed in two groups: control (C) and CIH treated (21% O<sup>2</sup> −80 s/5% O<sup>2</sup> − 40 s; 8 h/day; 30 days), as previously described (Quintero et al., 2016). At the end of experiments, animals were euthanized by the administration of a cardiac lethal dose of sodium pentobarbital.

### Plethysmography

fphys-09-00694 June 3, 2018 Time: 11:39 # 3

Ventilatory parameters as tidal volume (TV; mL/Kg), respiratory frequency (RF; breaths/min), minute ventilation (MV; mL/min/Kg) and O<sup>2</sup> consumption (VO2; mL/min/Kg) were obtained using whole-body, unrestrained plethysmography. Methacrylate-walled chambers (Emka Technologies, Paris, France; BUXCO Research Systems, Wilmington, NC, United States) continuously fluxed (1.5 L/min) with air, hypoxic gas mixtures (12, 10, and 7% O2, reminder nitrogen) and hypercapnic gas mixture (5% CO<sup>2</sup> in air) were used as described in Gonzalez-Obeso et al., 2017. Animals breathed air until achieving a standard resting behavior. Temperature inside the chamber was maintained within the thermo-neutral range (22–24◦C) and animal temperature was constant during the experiment. Pressure modifications inside the chamber reflecting TV were calculated with a high-gain differential pressure transducer. Amplitude of pressure fluctuations is proportionally correlated to TV; a calibration of the system by injections of 5 mL air into the chamber allowed a direct estimation of TV.

The VO<sup>2</sup> was measured using a CO<sup>2</sup> gas analyzer (AUT4499 and BUXCO MAX II preamplifier; Buxco Research Systems). The calibration required one gas without CO<sup>2</sup> and another containing a known concentration of CO<sup>2</sup> (see Lighton, 2008). All parameters were recorded and analyzed with FinePointe software (Buxco Research Systems). For oximetry monitoring during guinea pig CIH exposure, multiple arterial blood samples were analyzed showing that PO<sup>2</sup> nadir level was 27 mmHg (SaO<sup>2</sup> ≈ 50%) in this system.

### Arterial Blood Pressure Measurement

Systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial blood pressure (MABP), and heart rate (HR) were recorded from anesthetized animals (Ketamine 100 mg/Kg and diazepam 2 mg/Kg; i.p.) placed in supine position on a dissection table, tracheostomized and ventilated with room air (CL Palmer) (60 cycles. min−<sup>1</sup> and a positive endexpiratory pressure of 2 cm H2O) or with the selected gas mixture (10% O<sup>2</sup> and 90% N2; 3 min). Arterial blood pressure was continuously monitored by a catheter located in the right common carotid artery and connected to a pressure transducer (Transpac IV; ICU Medical, San Clemente, CA, United States). Signals were stored (BIOPAC Systems, Inc. MP 150, Goleta, CA; Acknowledge 3.9.1) for later analysis.

### Acid-Base Status and Blood Gases

Small arterial blood samples (0.3 ml) taken in heparinized syringes were obtained at 15 min of the baseline and after 3 min of the hypoxic challenge. Arterial pH, PO2, PCO2, HCO<sup>3</sup> <sup>−</sup>, percentage of hemoglobin saturation (SaO2) and hematocrit were measured (ABL 5, Radiometer Medical A/S, Copenhagen, Denmark). Erythropoietin (EPO) was measured by a commercial ELISA kit (MyBiosource, San Diego, CA, United States).

### CB Morphology and Tyrosine Hydroxylase Immunostaining

Control and CIH guinea pig CB were perfused and fixed as described in Gonzalez-Obeso et al. (2017). CB were cryoprotected by immersion in 30% (w/v) sucrose in phosphate buffer, embedded individually in Tissue-Tek <sup>R</sup> (Sakura Finetek, Zoeterwoude, Netherlands) and frozen at −20◦C. Tissue sections of 10 µm (Shandon Cryotome, Thermo, Electron Corporation) were collected in glass slices coated with poly L-lysine. CB sections were washed with PBS at room temperature, hematoxylin and eosin stained (H&E, Sigma-Aldrich, MO, United States), dehydrated and mounted with Eukitt Mounting Medium (Merck). Tyrosine hydroxylase (TH) immunofluorescence staining was performed in slices from control and CIH CB, identifying cell nuclei with DAPI. Sections were examined with microscope (Axioscop 2, Zeiss). Images were captured using a digital camera (CoolSNAP, Photometric, Roper Scientific) coupled to the microscope and analyzed using Metamorph 6.3 software.

Dissociated CB cells (Gomez-Niño et al., 2009) from four different control and four different CIH animals plated on several poly L-lysine-coated coverslips were immunostained for TH and nuclei with DAPI as previously described (Caceres et al., 2007; Gonzalez-Obeso et al., 2017). Cells were imaged using a laser confocal microscope (LEICA TCS, SP5) and confocal micrographs were processed using LAS software.

### Endogenous Catecholamine Content and Catecholamine Outflow From Adrenal Medulla

Endogenous CA content, in CB; superior cervical ganglion, SCG; renal artery, RA; and adrenal medulla, AM, were analyzed after organs were removed from anesthetized animals, glass to glass homogenized (0.1N perchloric acid; PCA and 0.1 mM EDTA), centrifuged and processed for HPLC analysis as described in Gonzalez-Obeso et al. (2017).

For CA outflow measurement, AM were placed in a Lucite chamber containing ice-cold Tyrode solution (in mM: 140 NaCl, 5 KCl, 2 CaCl2, 1.1 MgCl2, and 5 glucose; pH 7.4); tissues were dissected under microscope, transferred to a glass tube with Tyrode-NaHCO<sup>3</sup> solution equilibrated with 20% O2- 5% CO2- remainder N<sup>2</sup> at 37◦C and maintained in a shaker bath during 30 min, to recover from surgical stress. Each AM was transferred to vials containing 500 µl of Tyrode-NaHCO<sup>3</sup> solution equilibrated either with 20% O<sup>2</sup> − 5% CO<sup>2</sup> − N<sup>2</sup> (normoxic solution) or with 2% O<sup>2</sup> − 5% CO<sup>2</sup> − N<sup>2</sup> (hypoxic solution). Superfusion media were collected every 10 min in eppendorf tubes containing 200 µl of 0.4N PCA, 0.1 mM EDTA and frozen at −80◦C until the HPLC-ED analysis of CA was performed.

# Stimuli–Evoked Catecholamine Release From CB and Renal Artery Catecholamine Synthesis

Isolated CB were incubated 2 h with <sup>3</sup>H-tyrosine of high specific activity (40–50 Ci/mmol) and the cofactors for TH and dopamine beta hydroxylase, 100 µM 6-methyltetrahydropterine and 1 mM ascorbic acid, respectively, to evaluate stimuli-evoked secretory response. Afterward, CB were transferred to vials containing Tyrode-bicarbonate solution (in mM: 116 NaCl, 5 KCl, 2 CaCl2, 1.1 MgCl2, 10 HEPES, 5 glucose, 24 NaCO3H) equilibrated with different gas mixtures containing 21 or 2% O<sup>2</sup> and 5% CO<sup>2</sup> (pH 7.40). High K<sup>+</sup> solution was obtained by removing an equimolar amount of NaCl. Incubating solutions were changed every 10 min and their <sup>3</sup>H-CA content was measured by scintillation counter as described before (Chen et al., 1997).

Renal arteries were incubated (37◦C; 2 h) in Tyrode solution, containing 30 µM of 3,5-3H-tyrosine (6 Ci/mmol; Perkin Elmer, Boston, MA, United States), as described above for CB synthesis. After washing the tissues in precursor-free Tyrode solution (4◦C; 5 min), they were homogenized and processed for HPLC-ED analysis. General procedures have been previously described (Olea et al., 2014).

# Chemoreceptor Cell Culture and Intracellular Ca2<sup>+</sup> Recording

Enzymatically dispersed and dissociated CB cells were plated on poly L-lysine-coated coverslips maintained in culture for up to 24 h as formerly described (Gomez-Niño et al., 2009). Coverslips containing chemoreceptor cells were loaded with fura-2 (10 µM; Pluronic F-127; Thermo Fished Molecular Probes; 20◦C, 1 h), placed in a perfusion chamber on the stage of a Nikon Diaphot 300 inverted microscope and superfused with Tyrode solution. Fura-2 is excited at λ = 340 nm and 380 nm and emits fluorescence at 540 nm, collected with a SensiCam digital Camera (PCO CCD imaging, PCO, Kelheim, Germany). The background was removed (MetaFluor program, Molecular Devices, Wokingham, United Kingdom) and the variations in the cytosolic Ca2<sup>+</sup> were presented as the fluorescence emitted after excitation at 340 nm and the fluorescence emitted after excitation at 380 nm (ratio 340 nm/380 nm). The illumination system and camera were driven by Axon Imaging Workbench 4.0 (Molecular Devices, Wokingham, United Kingdom) running on a Pentium computer. Hypoxia (5% CO2, 95% N2) and 35 mM KCl were used as stimuli.

## Data Presentation and Statistical Analysis

Results are presented as mean ± SEM. Statistical analyses were completed by GraphPad Prism version 6.0. Mean value comparisons were performed with unpaired Student'st-test, One-Way Analysis of Variance (ANOVA) with Dunnett's multiple comparison tests, and by Two-way ANOVA with Tukey's or Sidak's multi-comparison test, according to the structure of data. A p-value < 0.05 was considered as statistically significant. All comparisons of experimental data were performed with twotailed tests.

# RESULTS

### Ventilatory Response

Ventilatory response to acute hypoxic tests (12, 10, and 7% O2; 10 min) and hypercapnic test (5% CO2; 10 min) were assessed by continuous recording of respiratory frequency (RF), tidal volume (TV), and minute volume (MV) in the same animals at two different times: initially, before intermittent hypoxia exposure (CIH0), and after 30 days of intermittent hypoxia treatment (CIH30). **Figure 1A** depicts single respiratory recordings obtained from one animal breathing each of the different atmospheres. Ventilatory parameters from CIH0 breathing room air were not modified when exposed to 12% O<sup>2</sup> but frequency significantly increased when guinea pigs breathed 10 and 7% O<sup>2</sup> (p < 0.05 and 0.001, respectively). CIH30 behaved similarly (p < 0.001 in both cases), as shown in **Figure 1B**. TV was not modified at any hypoxic tests in CIH0 or CIH30 (**Figure 1C**). Hypercapnic mixture (5% CO2) produced the same significant increase of RF and TV in CIH0 and CIH30 animals (p < 0.001; **Figures 1B,C**). The effect of 30 days of CIH exposure (CIH0 vs. CIH30) produced a slight (no statistically significant) increase of RF and TV in animals breathing 7% O<sup>2</sup> (**Figures 1B,C**). However, when the MV was calculated (see **Figure 2A**) significant differences were found (p < 0.01 CIH0 vs. CIH30).

**Figure 2A** shows the MV obtained in guinea pigs breathing different gas mixtures. Ventilation in normoxia, 12 and 10% O<sup>2</sup> were unaffected by exposure to CIH, but significantly increased when animals breathed 7% O<sup>2</sup> in CIH0 and CIH30: 397 ± 5 breathing room air and 519 ± 19 ml/min/kg breathing 7% O<sup>2</sup> in CIH0 guinea pigs vs. 418 ± 10 breathing room air and 578 ± 25 ml/min/kg breathing 7% O<sup>2</sup> in CIH30 (p < 0.001). Furthermore, MV increase at 7% O<sup>2</sup> hypoxia test was significantly higher in CIH30 than CIH0 guinea pigs (p < 0.01). Acute hypercapnia test produced a similar significant MV rise (1141 ± 45 in CIH0 vs. 1129 ± 43 ml/min/kg in CIH30; p < 0.001) compared with respective baseline values. Ventilatory response to Dejours test (100% O2; 3 min) showed no ventilatory differences between CIH0 and CIH30 (data not shown). **Figure 2B** represents the oxygen consumption (VO2) under the conditions described above. CIH0 animals breathing 10 and 7% O<sup>2</sup> atmosphere showed significantly metabolism decrease (17.1 ± 1.0 breathing air, 11.8 ± 1.4 breathing 10% O<sup>2</sup> and 12.2 ± 1.1 mL/min/kg breathing 7% O2) compared to the baseline value. Nonetheless, CIH30 guinea pigs significantly increased VO<sup>2</sup> breathing room air, 12 and 10% O2, except when breathing 7% O<sup>2</sup> or 5% CO<sup>2</sup> atmospheres, in which there were no differences between groups. **Figure 2C** depicts the ventilation standardized by metabolic rate (MV/VO2) showing hypoxic hyperventilation at 10 and 7% O<sup>2</sup> (p < 0.01), and hypercapnic hyperventilation (p < 0.001) in CIH0. Conversely, CIH30 group only hyperventilated at 7% O<sup>2</sup> and 5% CO<sup>2</sup> (p < 0.001). Consequently, CIH30 exposure blunted the hyperventilatory response at 10% O<sup>2</sup> (p < 0.05) and at hypercapnia (p < 0.01).

FIGURE 1 | Effect of 30 days of chronic intermittent hypoxia (CIH30) on guinea pig breathing pattern. (A) Sample plethysmography recordings of one control (CIH0) and one CIH30 guinea pig breathing air (21% O2), or 10% O2, 7% O<sup>2</sup> and 5% CO<sup>2</sup> acute tests. In (B), Respiratory frequency (RF) expressed as breaths per minute (bpm) and in (C), Tidal volume (TV) expressed as mL/kg from CIH0 and CIH30 guinea pigs in response to acute hypoxia (12% O2, 10% O2, 7% O2) and hypercapnia (5% CO2) tests. #p < 0.05; ##p < 0.01; ###p < 0.001 vs. basal CIH0; +++p < 0.001 vs. basal CIH30. Data are mean ± SEM; n = 16. Two-way ANOVA with Tukey's multiple comparison test for analysis between CIH0 and CIH30, and one-way ANOVA with Dunnett's multiple comparison test for analysis intra group (vs. basal).

(B) Oxygen consumption (VO2; mL/min/kg) of CIH0 and CIH30 guinea pigs in response to the different gas mixtures. #p < 0.05 ##p < 0.01 vs. basal CIH0; ++p < 0.01 vs. basal CIH30; ∗∗p < 0.01 ∗∗∗p < 0.001 CIH0 vs. CIH30. (C) Normalized ventilation to oxygen consumption (MV/VO2) calculated from guinea pigs breathing under the same conditions described above. #p < 0.05; ###p < 0.001 vs. basal CIH0; +++p < 0.001 vs. basal CIH30; <sup>∗</sup>p < 0.05 ∗∗p < 0.01 CIH0 vs. CIH30. Data are mean ± SEM; n = 16. Two-way ANOVA with Tukey's multiple comparison test for analysis between CIH0 and CIH30 groups, and one-way ANOVA with Dunnett's multiple comparison test for analysis intra group (vs. basal).

### Acid-Base Status and Blood Parameters

**Table 1** shows blood values for pO2, pCO2, pH, HCO<sup>3</sup> −, and SaO<sup>2</sup> obtained from small arterial blood samples taken in normoxia (breathing air) or after acute hypoxia (10% O2; 3 min) in C and CIH guinea pigs. All parameters were comparable when guinea pigs breathed air in C and CIH groups. Neither differences found between groups, except for the lower SaO<sup>2</sup> after the acute hypoxia test in CIH guinea pigs (44 ± 6 vs. 62 ± 5%, respectively; p < 0.05). Whereas hematocrit was identical (41.5 ± 0.8 vs. 41.4 ± 0.8%) in both groups, EPO increased in CIH animals (214 ± 52 vs. 141 ± 17 mU/mL; p < 0.05).

### CB Morphology and Stimuli-Dependent Activation

At day 0 guinea pigs body weight was 637 ± 14 g in C and 610 ± 17 g in CIH0 (p > 0.05; unpaired t-test). After 30 days, body weight was 802 ± 19 g in C and 709 ± 14 g in CIH guinea pigs (n = 16, p < 0.01 unpaired t-test). However, CB weight was no different in C and CIH animals (81 ± 4 µg vs. 87 ± 5 µg, respectively; p > 0.05; n = 8).

**Figure 3A** shows the whole surface of 10 µm thick slices from control (top) and CIH (bottom) guinea pig CB, stained with hematoxylin and eosin. An extended detail of the parenchyma from each slice is also shown. No substantial differences in the structure of the tissue were observed. **Figure 3B** shows TH immunostaining in CB slices from control and CIH animals. The TH positive staining area was lower than that observed in CB from other rodents. **Figure 3C** shows chemoreceptor cells dissociated from CB cultured and immunostained for TH from C and CIH animals. The percentage of TH-positive type I cells was 7% in both cases (by cumulative counting of several cultures), obtained from 4 guinea pigs of each group, a percentage also lower than that observed in CB from other rodents.

**Figure 3D** shows the endogenous CA content in CB from C and CIH animals. NE content was significantly higher in CIH animals (p < 0.05) and although DA content was slightly higher in CIH, the increase was not statistically significant. **Figures 3E,F** show the lack of response to hypoxia from in vitro CB, expressed as evoked CA secretion (% of tissue content) or Ca<sup>i</sup> <sup>2</sup><sup>+</sup> changes,

TABLE 1 | Arterial blood gasometry data, hematocrit and erythropoietin measurements in control and CIH guinea pigs breathing normoxia (21% O2) or acute hypoxia (10% O2).


+++p < 0.001 10% O<sup>2</sup> vs. 21% O<sup>2</sup> in each group; <sup>∗</sup>p < 0.05 CIH vs. C. Data are mean ± SEM (n = 8). Two-Way ANOVA with Sidak's multiple comparison test.

in both groups of animals. Unlike this lack of response to hypoxic stimulus, CB from both groups released comparable amounts of <sup>3</sup>H-CA (7.1 ± 1.1% vs. 9.0 ± 0.9%, respectively; p > 0.05; **Figure 3E**), and increased to comparable rise of Ca<sup>i</sup> <sup>2</sup><sup>+</sup> in response to high external K<sup>+</sup> (**Figure 3F**).

# Endogenous Catecholamine Content

The lack of effect of CIH exposure on the CB activity prompted the analysis of the CA content in sympathetic endings of RA, SCG, AM and plasma CA levels as an index of sympathetic activity. Selective differences were found in tissue CA content from SCG and plasma CA levels. As shown in **Figure 4A**, NE content was 60.8 ± 2.3 in C and 74.8 ± 3.4 pmol/mg SCG in CIH (n = 12; p < 0.001); DA content was 7.7 ± 0.5 in C and 11.8 ± 0.8 pmol/mg SCG in CIH group. There were no changes in RA and AM endogenous CA content from both groups **Figure 4B** shows the similar RA content of NE in C and CIH guinea pigs (7.6 ± 1.3 pmol/mg vs. 8.2 ± 0.7 pmol/mg of tissue, respectively). However, <sup>3</sup>H-NE syntheses was significantly higher in RA from CIH guinea pigs (0.52 ± 0.06 vs. 0.69 ± 0.04 pmol/mg/h of tissue in C and CIH; **Figure 4C**). Adrenal medulla CA content was also similar in both groups (1.13 ± 0.15 vs. 1.11 ± 0.01 nmol/mg of tissue of NE and 19.4 ± 2.3 vs. 20 ± 1.7 nmol/mg of tissue of E in C and CIH, respectively; **Figure 4D**). Plasma NE was significantly higher in CIH than in C guinea pigs (80.7 ± 24.3 vs. 7.2 ± 1.7 pmol/mL). Epinephrine (E) was 11 ± 4 pmol/mL in C and also significantly higher in CIH (64 ± 20 pmol/mL; n = 6–8, p < 0.05; **Figure 4E**). Fasting glucose levels also were significantly higher in CIH than in C guinea pigs (100 ± 5 vs. 86 ± 2 mg/dL; p < 0.05; **Figure 4F**).

### Cardiovascular Effects

Consistent with the increase in plasma CA levels we found that (MABP from anesthetized guinea pigs also increased in the experimental CIH group, although the effect was less noticeable (45 ± 3 vs. 37 ± 2 mmHg in CIH and C, respectively; p < 0.05; **Figure 5A**). In **Figure 5B** is represented the systolic (SBP) and diastolic (DBP) blood pressure obtained from animals breathing air, acute hypoxic test (10% O2) and recovery (air). The profile of SBP (47 ± 1 and 46 ± 2 mmHg in C group vs. 53 ± 3 and 51 ± 4 mmHg in CIH breathing air and 10%O2, respectively) and DBP (31 ± 2 and 22 ± 1 mmHg in C group vs. 36 ± 3 and 24 ± 1 mmHg in CIH, breathing air and 10% O2, respectively) were similar in both groups of animals, but slightly higher in CIH. **Figure 5C** shows the significant HR increase in CIH guinea pigs (245 ± 5 vs. 195 ± 3 bpm; p < 0.001). No differences were observed in whole heart, right ventricle or left ventricle plus septum weight between C and CIH guinea pigs (**Figure 5D**).

# Catecholamine Outflow From the Adrenal Medulla

Data presented in **Figure 4D** have not underscored a clear differential behavior in the adrenal medulla CA content after CIH exposure. However, increase in plasma CA levels (**Figure 4E**) and metabolic response (**Figure 4F**) in CIH guinea pigs prompted checking the possible direct hypoxic sensitivity of

adrenomedullary chromaffin cells. We studied the effect of hypoxia (2% O2) on in vitro adrenal medulla CA outflow from both groups. Results presented in **Figure 6** show that hypoxia does not elicit CA efflux, not altering the ongoing time course of the outflow from AM in either group. In normoxic conditions (21% O2), the amount of E in the incubation solution was 2979 ± 678 pmol/AM and hypoxia (2% O2) induced an outflow of 2583 ± 439 pmol/AM in C guinea pigs. NE levels were 159 ± 36 pmol/AM in normoxia and 151 ± 26 pmol/AM in hypoxia in the same group. Under the same conditions, the amount of E in the AM effluent from CIH guinea pigs was 1527 ± 133 pmol/AM in 21% O<sup>2</sup> and 1628 ± 140 pmol/AM in 2% O2. NE values were 84 ± 6 and 100 ± 9 pmol/AM in normoxia and hypoxia, respectively, from CIH animals. Data showed that CIH does not provide hypoxia-sensitivity to guinea pig AM.

### DISCUSSION

The main results of this work are the lack of CB sensitization after CIH exposure during 30 days in this species and, as a consequence, the lack of respiratory reflex responsiveness to acute hypoxia (except to 7% O2), which has been described in rodents with hypoxia sensitive CB. These findings extend previous observations on the effect of acute and CSH treatment on CB responses in guinea pigs (Gonzalez-Obeso et al., 2017). From an integrative approach, combining functional studies in vivo and in vitro, we have tested the hypothesis that the systemic arterial hypertensive effect of CIH would be removed or ameliorated by the lack of CB responsiveness in guinea pigs. Our data partially support the hypothesis tested, showing that guinea pigs, compared to rats, have a small systemic arterial pressure change after CIH, despite similar sympathetic effects, suggesting a critical role for CIH induced sensitization in oxygen sensing mechanisms in the CB. Conversely to this species, it has been reported that adult rats exposed to 10 days of CIH have augmented hypoxic sensory response (Peng and Prabhakar, 2004) and a similar effect was also observed in cats (Rey et al., 2004) and mice (Peng et al., 2006). Moreover, rats exposed to 30 days of CIH develop hypertension and increased sympathetic nerve activity; these responses are prevented by chronic bilateral sectioning of carotid sinus nerve (Lesske et al., 1997). Our findings constitute the first data of CIH effects on guinea pig ventilation, CB functionality and sympatho-adrenal activation.

Previous work showing that the poor effect of acute hypoxia on ventilation correlates with the lack of CB activity suggested

a lack of action potential generation in the carotid sinus nerve (Schwenke et al., 2007; Gonzalez-Obeso et al., 2017). Although we did not measure the afferent activity in the carotid sinus nerve in our preparation, it is widely accepted that afferent chemosensory activity results from synaptic activation mediated by neurotransmitters released from type I cells (Gonzalez et al., 1994; Prabhakar, 2000; Iturriaga and Alcayaga, 2004). The fact that there was no Ca<sup>i</sup> <sup>2</sup><sup>+</sup> increase or CA secretion from CB in response to hypoxia in CIH exposed animals, suggests that there was no sensory nerve activity in CB from CIH guinea pigs, as is the case for normoxic animals (Schwenke et al., 2007). This would be the cause for the absence of hypoxic, but not hypercapnic, ventilatory response.

glucose plasma levels expressed as mg/dL. <sup>∗</sup>p < 0.05 CIH vs. C; Data are mean ± SEM (n = 16); Unpaired t-test.

Activation of chemoreceptor type I cells by hypoxia depends on the inhibition of O2-sensitive plasma membrane K<sup>+</sup> channels, leading to cell depolarization, Ca2<sup>+</sup> influx and neurotransmitter release. This model of chemoreceptor activation, accepted for several animal models, does not fit the guinea pig CB behavior. Type I cells contain DA allowing to visualize with TH antisera, the rate-limiting CA synthesis enzyme. In a previous study we analyzed the expected presence of TH and the occurrence of endogenous CA (by HPLC). We observed clear species differences between rat and guinea pig CB type I cells with respect to the immunoreactivity to TH antisera. In spite of the similar size to rat CB, a very small number of TH positive cells was found in guinea pig CB and lower levels of CA than in rat CB. We did not find serotonin in control or CIH guinea pig CB as has been reported in rat CB exposed to CIH (Peng et al., 2009). Guinea pig type I cells lack O2-sensitive K<sup>+</sup> channels, and therefore, they would lack hypoxia depolarization capacity and CA secretory response (Gonzalez-Obeso et al., 2017). After CIH we did not observe hypoxia elicited secretory response or intracellular calcium increase, suggesting that CIH exposure does not contribute to developing hypoxia depolarization capacity in guinea pig CB type I cells.

The fact that the guinea pig CB is hypofunctional and does not respond to the full hypoxia range (mild to severe hypoxia) correlates with a blunted ventilatory response in both C and CIH guinea pigs. These breathing responses are typical of mammals and humans adapted to high altitude (Monge and León-Velarde, 1991). The absence of hypoventilation during the Dejours test indicates that, unlike rats and other species, CB is not a contributing drive to normoxic ventilation in guinea pigs. In conscious animals, only severe hypoxia caused a significant MV increase before (30%) and after (38%) CIH exposure, mainly caused by augmented respiratory frequency. Yet, the difference between both responses was significant (25%; p < 0.05), meaning that there was a chemo-reflex sensitization after CIH treatment. Acute hypoxia in CIH0 guinea pigs significantly decreased O<sup>2</sup> consumption (breathing 10 and 7% O2); this would imply that a metabolic depression compensates the blunted ventilation in hypoxic CIH0 guinea pigs. Conversely, after CIH exposure there was an increase (≈50%) in oxygen consumption when breathing air, 12 or 10% O2. Increased O<sup>2</sup> consumption in

CIH animals mismatches their low ventilation. Consequently, normalizing ventilation to O<sup>2</sup> consumption in both groups, we can observe hyperventilation at 10 and 7% O<sup>2</sup> in CIH0 but a blunted ventilation at 10% O<sup>2</sup> and an unaffected response to the most intense hypoxic stimulus (7% O2), after CIH exposure. Increased ventilation and decreased metabolism after CIH, renders similar values for MV/VO<sup>2</sup> ration when breathing 7% O<sup>2</sup> in both groups of animals. Ventilatory response to hypercapnia is similar before and after CIH treatment and comparable to that observed in the rat (Gonzalez-Obeso et al., 2017). However, it can be also observed hypoventilation when normalized to higher oxygen consumption after CIH exposure. Guinea pig CB sensory denervation decreased 28% the hyperventilation induced when breathing a hypercapnic mixture (Schwenke et al., 2007), suggesting that hyperventilation is also mediated by central (≈70%) and arterial (≈30%; Gonzalez et al., 1994) chemoreceptors.

Previously, we have observed that anesthetized guinea pigs have a considerably low (60 mmHg) arterial blood PO<sup>2</sup> (Gonzalez-Obeso et al., 2017). This finding, comparable to other reported PO<sup>2</sup> values of 66 to 57 mmHg in awake and anesthetized guinea pigs (Feuerstein et al., 1985; Mover-Lev et al., 1997; Schwenke and Cragg, 2004), is unaffected after CIH exposure (65 mmHg in C vs. 64 mmHg in CIH). Hypoxic challenge (10% O2) diminished SaO<sup>2</sup> even more in CIH (44% vs. 66% in C) because of a higher metabolism (**Figure 2B**). It is known that guinea pigs have a high hemoglobin affinity (P50≈27 mmHg) compared to rats. This high affinity will secure lung oxygenation at low PO<sup>2</sup> (Turek et al., 1980). Low PO<sup>2</sup> and low SaO<sup>2</sup> would be also compensated by an increase in red blood cells (RBC) and a concomitant oxygen blood transport capacity. In hypoxia, RBC rise to enhance oxygen delivery to tissues by increased erythropoiesis that is mediated by HIF-2, the main regulator of EPO gene transcription (Rankin et al., 2007). However, a rapid change of oxygen tension from hypoxia to normoxia rectifies the hypoxia-induced RBC growth, preferentially eliminating the new RBC or neocytes, a process coined as neocytolysis (Alfrey et al., 1997; Risso et al., 2007). This would be the reason that OSA patients, with significant cycles of severe hypoxia during sleep, do not present polycythaemia (Solmaz et al., 2015; Song et al., 2017). We think this also would be the reason that CIH guinea pigs have identical haematocrit to normoxic control ones, in spite

of a significantly higher level of EPO (214 vs. 141 mU/mL in C); the increased EPO level in CIH group should be considered as a HIF-mediated adaptive response to CIH.

The fact that the guinea pig CB is not sensitized by CIH would strengthen our hypothesis: systemic effects related to CIH in other species should be absent in guinea pigs. However, several paradoxical results would point to an activation of the sympathetic system after CIH: (i) a slight increase in arterial blood pressure (≈10 mmHg); (ii) a slight increase in HR (≈25%); (iii) a large increase in plasma CA levels; and, (iv) an increase in renal artery CA synthesis (≈40%). Guinea pigs are hypotensive animals compared to other rodents (Schwenke and Cragg, 2004; Gonzalez-Obeso et al., 2017) which has been related to the low noradrenergic tone and to a higher capillarity in peripheral tissues decreasing peripheral resistances. The observed increase in plasma NE levels could explain the higher MABP and HR found from guinea pigs after CIH exposure. These results suggest that guinea pigs would possess an O2-sensing mechanism, other than the CB, responsible for the sympathetic cardio-circulatory reflex (Cardenas and Zapata, 1983; Guyenet, 2000). This mechanism remains to be studied.

The CB and AM are sympatho-adrenal tissues with the same developmental origin, functioning as a unit that originates cardiorespiratory reflexes in response to systemic hypoxia. A functional CB-AM axis is essential for hypoxic environment survival, and the most efficient response is the hyperventilation triggered by CB chemoreceptors, nearly absent in guinea pigs (Gonzalez-Obeso et al., 2017). Therefore, we envisioned the possibility that metabolic responses would be over-developed in guinea pigs as a result of the sympathetic activity and the intrinsic sensitivity of AM to hypoxia, mimicking the situation of neonatal animals (Seidler and Slotkin, 1985; Thompson et al., 1997; Nurse et al., 2017). The increased plasma CA and fasting glucose levels, and the blood pressure modifications observed in CIH guinea pigs would arise from the AM sensitization. We also found endogenous CA increase (mainly NE) in SCG, but no changes in AM content. Sympathetic nerve endings are the origin of plasma NE, and E is secreted from AM, so plasma levels of NE and E can represent an index of the general sympathetic tone (Goldstein et al., 2003; Kjaer, 2005). Plasma E/NE ratio might change under stress conditions, such as intermittent hypoxia; therefore, the different source of both CA would allow disclosing the main origin of plasma CA; an intermittent hypoxia activation of the AM would increase the E/NE ratio. However, plasma E/NE ratio was lower in CIH than in C guinea pigs, indicating that AM contribution to plasma CA is lower than sympathetic endings and CIH does not activate AM, but increases sympathetic tone. The increased NE content in SCG would reinforce this finding. We have also studied the effect of CIH on the time course of the CA outflow from isolated AM in both groups of guinea pigs. CIH exposure did not alter hypoxia elicited time course of E and NE adrenal outflow from in vitro AM. The lack of response of adult in vitro AM to hypoxia was expected from previously published findings in adrenomedullary cells from rats (Seidler and Slotkin, 1985; Thompson et al., 1997) and from our own data in guinea pigs (Olea et al., 2018). However, Prabhakar's group reported

that CIH induced oxidative stress and facilitates catecholamine efflux from the AM slices in adult rats (Kumar et al., 2015). The absence of the in vitro AM and CB response to hypoxia in guinea pigs leads to question how the sympatho-adrenal activation is achieved in CIH. Most of the response would be reflex-triggered by the CB in rats (Gonzalez et al., 1994; Marshall, 1994). In guinea pigs, due to the lack of functional CB, there must be additional hypoxia-sensitive structures capable of triggering the sympatho-adrenal activation. Several studies have recognized caudal hypothalamus and rostral ventrolateral medulla areas directly activated by hypoxia in different species, increasing sympathetic discharges, blood pressure and HR (Guyenet, 2000; Neubauer and Sunderram, 2004; Marina et al., 2015).

In summary, CIH exposure has no effect on the guinea pig CB, and none or poor effect on basal ventilation but blunted respiratory responses to acute hypoxic challenges. The lack of response to hypoxia suggests that CIH does not modify the excitability of CB type I cells or their synaptic communication with afferent endings. Moreover, similar time and intensity treatments in other rodents produce profound long-lasting effects in chemosensory reflex and related pathological effects. Further studies are needed to clarify the missing mechanisms that underlie the lack of long-term effects of CIH exposure on the guinea pig CB to provide evidence for the role of the CB in mediating hypertension in OSA. Therefore, guinea pigs could represent an interesting model to study the brainstem sensitivity

### REFERENCES


to hypoxia, the oxygen sensing plasticity and the cardiovascularendocrine responses elicited by chronic intermittent low PO2.

### AUTHOR CONTRIBUTIONS

AG-N and AR designed the study and prepared the manuscript. ID, EO, JP-L, TG-M, AO, AG-N, and AR carried out the experiments and analyzed the data. JP-L and EO designed graphs. All authors read and approved the final manuscript.

# FUNDING

This research was supported by grants: MINECO/FEDER, UE BFU2015-70616R, and ISCiii CIBER CB06/06/0050.

### ACKNOWLEDGMENTS

We thank Ms. Ana Gordillo and Ms. Maria Ll. Bravo for their invaluable assistance in the maintenance and care of the experimental animals and for technical assistance. We acknowledge support of the publication fee by the CSIC Open Access Publication Support Initiative through its Unit of Information Resources for Research (URICI).

chemoreceptor cells of the rat carotid body. Am. J. Physiol. Cell Physiol. 297, C715–C722. doi: 10.1152/ajpcell.00507.2008



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Docio, Olea, Prieto-LLoret, Gallego-Martin, Obeso, Gomez-Niño and Rocher. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# AMPK-α1 or AMPK-α2 Deletion in Smooth Muscles Does Not Affect the Hypoxic Ventilatory Response or Systemic Arterial Blood Pressure Regulation During Hypoxia

### Sandy MacMillan and A. Mark Evans\*

Centre for Discovery Brain Sciences and Centre for Cardiovascular Science, College of Medicine and Veterinary Medicine, University of Edinburgh, Edinburgh, United Kingdom

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Eduardo Colombari, Universidade Estadual Paulista Júlio de Mesquita Filho (UNESP), Brazil Rodrigo Varas, Universidad Autónoma de Chile, Chile

> \*Correspondence: A. Mark Evans mark.evans@ed.ac.uk

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 28 February 2018 Accepted: 14 May 2018 Published: 06 June 2018

### Citation:

MacMillan S and Evans AM (2018) AMPK-α1 or AMPK-α2 Deletion in Smooth Muscles Does Not Affect the Hypoxic Ventilatory Response or Systemic Arterial Blood Pressure Regulation During Hypoxia. Front. Physiol. 9:655. doi: 10.3389/fphys.2018.00655 The hypoxic ventilatory response (HVR) is markedly attenuated by AMPK-α1 deletion conditional on the expression of Cre-recombinase in tyrosine hydroxylase (TH) expressing cells, precipitating marked increases in apnea frequency and duration. It was concluded that ventilatory dysfunction caused by AMPK deficiency was driven by neurogenic mechanisms. However, TH is transiently expressed in other cell types during development, and it is evident that central respiratory depression can also be triggered by myogenic mechanisms that impact blood supply to the brain. We therefore assessed the effect on the HVR and systemic arterial blood pressure of AMPK deletion in vascular smooth muscles. There was no difference in minute ventilation during normoxia. However, increases in minute ventilation during severe hypoxia (8% O2) were, if affected at all, augmented by AMPK-α1 and AMPK-α2 deletion in smooth muscles; despite the fact that hypoxia (8% O2) evoked falls in arterial SpO<sup>2</sup> comparable with controls. Surprisingly, these mice exhibited no difference in systolic, diastolic or mean arterial blood pressure during normoxia or hypoxia. We conclude that neither AMPK-α1 nor AMPK-α2 are required in smooth muscle for the regulation of systemic arterial blood pressure during hypoxia, and that AMPK-α1 deficiency does not impact the HVR by myogenic mechanisms.

### Keywords: AMPK, hypoxia, ventilation, blood pressure, smooth muscle

### INTRODUCTION

The AMP-activated protein kinase (AMPK) is a cellular energy sensor that maintains cellautonomous energy homeostasis. From its 2 α (catalytic), 2 β and 3 γ (regulatory) subunits 12 AMPK heterotrimers may be formed, each harboring different sensitivities to activation by increases in cellular AMP and ADP, and the capacity to directly phosphorylate and thus regulate different targets (Ross et al., 2016). AMPK is coupled to mitochondrial oxidative phosphorylation by two discrete albeit cooperative pathways, involving liver kinase B1 (LKB1) and changes in the cellular AMP:ATP and ADP:ATP ratios. Binding of AMP to the AMPK γ subunit increases activity 10-fold by allosteric activation alone, while AMP or ADP binding delivers increases in LKB1-dependent phosphorylation and reductions in dephosphorylation of Thr172 on the α

subunit that confer 100-fold further activation. All of these effects are inhibited by ATP (Gowans et al., 2013). LKB1 is, therefore, the principal pathway for AMPK activation during metabolic stresses such as hypoxia. However, there are alternative Ca2+ dependent pathways to AMPK activation that are governed by the calmodulin-dependent protein kinase CaMKK2, which delivers increases in Thr172 phosphorylation and thus AMPK activation independent of changes in cellular AM(D)P:ATP ratios.

Classically AMPK regulates cell-autonomous pathways of energy supply by phosphorylating targets that switch off nonessential anabolic processes that consume ATP and switch on catabolic pathways that generate ATP, thereby compensating for deficits in ATP supply or availability (Ross et al., 2016). Recently, however, we demonstrated (Mahmoud et al., 2016) that the role of AMPK in metabolic homeostasis is not limited to such cell autonomous pathways, but extends to the hypoxic ventilatory response (HVR) (Teppema and Dahan, 2010; Wilson and Teppema, 2016) and thus O<sup>2</sup> and energy (ATP) supply to the body as a whole. In doing so AMPK acts to oppose central respiratory depression during hypoxia and thus resists hypoventilation and apnea.

However, the HVR could well be affected by AMPK deficiency in cell systems other than catecholaminergic neurons due to off-target AMPK deletion through "leakage" of Cre beyond those cells targeted by the conditional deletion strategy we employed. This possibility is highlighted by the fact that transient developmental expression of TH occurs in disparate cell groups that do not express TH in the adult (Lindeberg et al., 2004), including, for example, a subset of heart wall cells. Therefore, there remains the possibility that AMPK deficiency might somehow affect ventilatory control mechanisms through myogenic rather than neurogenic mechanisms. This is evident from the fact that systemic arteries dilate in response to tissue hypoxemia in order to match local perfusion to local metabolism (Roy and Sherrington, 1890), and evidence suggests a role for AMPK in this response (Goirand et al., 2007; Schneider et al., 2015). With "leakage" of Cre, off-target deletion of AMPK in arterial myocytes could attenuate arterial dilation and thus impact O<sup>2</sup> supply to the brain during hypoxia and the HVR. This possibility is highlighted by the fact that respiratory failure can be triggered through the Cushing reflex consequent to marked increases in blood pressure (Guyenet, 2000; Paton et al., 2009).

We therefore assessed the impact of AMPK deficiency in smooth muscles on blood pressure control and the HVR. Our data show that AMPK does not contribute to the HVR or blood pressure control during hypoxia through myogenic mechanisms.

### MATERIALS AND METHODS

Experiments were performed in accordance with the regulations of the United Kingdom Animals (Scientific Procedures) Act of 1986. All studies and breeding were approved by the University of Edinburgh and performed under UK Home Office project licenses. Both male and female mice were used, and all were on a C57/Bl6 background. Numbers of mice (≥3 per measure) used are indicated for each experiment. Global, dual knockout of the genes encoding AMPK-α1 (Prkaa1) and AMPK-α2 (Prkaa2) is embryonic lethal. We therefore employed conditional deletion of the genes for the AMPK-α1 or AMPK-α2 subunits, using mice in which the sequence encoding the catalytic site of either or both of the α subunits was flanked by loxP sequences (Lantier et al., 2014). To direct AMPK deletion to cells of the smooth muscle lineage, these were crossed with mice in which Cre recombinase was under the control of the transgelin (smooth muscle protein 22α) promoter (The Jackson Laboratory, Bar Harbor, ME, United States [stock number 017491, Tg(Tagln-cre)1Her/J in C57BL/6:129SJL background]). The presence of wild-type or floxed alleles and CRE recombinase were detected by PCR. We used four primers for Cre (Transgene 5<sup>0</sup> -GCGGTCTGGCAGTAAAAACTATC-3 0 and 5<sup>0</sup> -GTGAAACAGCATTGCTGTCACTT-3<sup>0</sup> ; internal positive control 5<sup>0</sup> -CTAGGCCACAGAATTGAAAGATCT-3<sup>0</sup> and 5<sup>0</sup> -GTAGGTGGAATTTCTAGCATCATCC-3<sup>0</sup> ; expected size WT = 324 bp, Cre = 100 bp). Two primers were used for each AMPK catalytic subunit: α1-forward 5<sup>0</sup> - TATTGCTGCCATTAGGCTAC-3<sup>0</sup> , α1-reverse: 5<sup>0</sup> -GACCTGAC AGAATAGGATATGCCCAACCTC-3<sup>0</sup> (WT = 588 bp, Floxed = 682 bp); α2-forward 5<sup>0</sup> -GCTTAGCACGTTACCC TGGATGG-3<sup>0</sup> , α2-reverse: 5<sup>0</sup> -GTTATCAGCCCAACTAATT ACAC-3<sup>0</sup> (WT = 204 bp, Floxed = 250 bp). 10 µl samples were run on 2% agarose gels with 0.01% v/v SYBR <sup>R</sup> Safe DNA Gel Stain (Invitrogen) in TBE buffer against a 100 bp DNA ladder (GeneRulerTM, Fermentas) using a Model 200/2.0 Power Supply (Bio-Rad). Gels were imaged using a Genius Bio Imaging System and GeneSnap software (Syngene).

## End-Point and Quantitative RT-PCR

For qPCR analysis, 2 µl of cDNA in RNase free water was made up to 20 µl with 10 µl FastStart Universal SYBR Green Master (ROX, Roche), 6.4 µl Ultra Pure Water (SIGMA) and forward and reverse primers for AMPK-α1 and AMPK-α2. The sample was then centrifuged and 20 µl added to a MicroAmpTM Fast Optical 96-Well Reaction Plate (Greiner bio-one), the reaction plate sealed with an optical adhesive cover (Applied Biosystems) and the plate centrifuged. The reaction was then run on a sequence detection system (Applied Biosystems) using AmpliTaq Fast DNA Polymerase, with a 2 min initial step at 50◦C, followed by a 10 min step at 95◦C, then 40x 15 s steps at 95◦C. This was followed by a dissociation stage with 15 s at 95◦C, 20 s at 60◦C, and 15 s at 95◦C. Negative controls included control cell aspirants for which no reverse transcriptase was added, and aspiration of extracellular medium and PCR controls. None of the controls produced any detectable amplicon, ruling out genomic or other contamination.

### Plethysmography

As described previously (Mahmoud et al., 2016), we used a whole-body unrestrained plethysmograph, incorporating a HalcyonTM low noise pneumatochograph (Buxco Research Systems, United Kingdom) coupled to FinePointe acquisition and analysis software (Data Science International, United States). Following acclimation and baseline measurements (awake but quiet, undisturbed periods of breathing) under normoxia (room

air), mice were exposed to hypoxia (8% O2, with 0.05% CO2, balanced with N2) for 10 min. The FinePointe software automatically calculated the respiratory parameters assessed after application of exclusion criteria due to non-ventilatory artifacts (movement, sniffing, etc.). Data were acquired as 2 s averages and 2 to 4 data points of undisturbed breathing were selected for each time point of the HVR. Apneas were defined as a period of cessation of breathing that was greater than the average duration, including interval, of two successive breaths (∼600 ms) of control mice during normoxia with a detection threshold of 0.25 mmHg (SD of noise).

### Blood Pressure Measurements

Blood pressure was recorded using a CODA non-invasive blood pressure system (Kent Scientific, United States). Following five consecutive days of habituation "training," mice were restrained using a pre-warmed mouse holder and placed on a warming platform (Kent Scientific, United States). Resting blood pressure was measured 32 times per session and only stable readings throughout each session were considered for analysis. On separate days hypoxic blood pressure was measured using the following protocol: 13 cycles of normoxia, 16 cycles of hypoxia (8% O2, with 0.05% CO2, balanced with N2; approximately 10 min) and finally eight cycles of recovery (approximately 5 min). Hypoxic gas originating from a pre-mixed gas cylinder (BOC, United Kingdom) was delivered at a flow rate of 2 L/min via a head cone that inserts into the mouse holder.

### Measurement of Arterial Oxygen Saturation and Heart Rate

Arterial SpO<sup>2</sup> measurements of mice were obtained using the MouseSTATTM Pulse Oximeter (Kent Scientific, United States) and an infrared Y-style tail sensor at a sampling rate of 0.5 Hz. As above, mice were restrained using a pre-warmed mouse holder and placed on a warming platform (Kent Scientific, United States). Following 5 min of acclimation, baseline SpO<sup>2</sup> and heart rates were measured for 2 min immediately preceding a 10-min period of hypoxia (8% O2, with 0.05% CO2, balanced with N2), followed by 3–5 min of recovery. Hypoxic gas originating from a pre-mixed gas cylinder (BOC, United Kingdom) was delivered at a flow rate of 2 L/min via a head cone that inserts into the mouse holder.

### Statistical Analysis

Statistical comparison was completed using GraphPad Prism 6 for the following: plethysmography, two-way ANOVA with Bonferroni multiple comparison's test; blood pressure, oneway ANOVA with Bonferroni multiple comparison's test, SpO2, two-way ANOVA with Bonferroni multiple comparison's test. p < 0.05 was considered significant.

# RESULTS

We investigated the possibility that AMPK-α1 or AMPK-α2 subunits might support myogenic responses to hypoxia and thus impact the HVR and blood pressure control. To achieve this goal we developed mice with conditional deletion in smooth muscles of the gene encoding AMPK-α1 (Prkaa1) and AMPKα2 (Prkaa2), respectively. Critical exons of the AMPK-α1 and -α2 subunit genes were flanked by loxP sequences (Lantier et al., 2014), and each floxed mouse line was crossed with mice expressing Cre recombinase under the control of the transgelin promoter (smooth muscle protein 22α) (El-Bizri et al., 2008). Previous studies have shown that transgelin-Cre mice do not exhibit Cre expression in endothelial cells, and therefore provide for selective gene deletion in smooth muscle versus endothelial cells (**Supplementary Figure S1**). It should be noted, however, that transient developmental expression of transgelin has been observed in atrial and ventricular myocytes. Consequently, genomic recombination will also occur in these cells despite the fact that they do not express transgelin in the adult (El-Bizri et al., 2008). Therefore, while imperfect, the use of transgelin-Cre mice provided us with the necessary level of specificity to determine the role of smooth muscle AMPK in myogenic responses to hypoxia that impact the HVR and blood pressure control.

# The Hypoxic Ventilatory Response Is Not Attenuated by AMPK-α1 or AMPK-α2 Deficiency in Smooth Muscles

Mice with conditional deletion of AMPK-α1 or AMPK-α2 subunits in smooth muscles exhibited no obvious phenotype and no ventilatory dysfunction or deficiency was evident during normoxia (not shown). Most significantly in the context of the present investigation, we identified no significant attenuation of the HVR relative to controls (AMPKα1/α2 floxed; **Figure 1** and **Supplementary Figure S2A**). That said, we did observe subtle differences between genotypes.

For smooth muscle AMPK-α1 knockouts alone, there appeared to be an insignificant yet noticeable lowering, relative to controls, of breathing frequency during hypoxia following Roll-Off (≈100 s; −13 ± 4 and 11 ± 4%, respectively) that was maintained for the duration of the Sustained Phase (2– 5 min; −2 ± 7 and 12 ± 4%, respectively) of the HVR (**Figure 1A**). By contrast, in both AMPK-α1 and AMPK-α2 knockouts larger increases in tidal volume were observed when compared to controls (**Figure 1B** and **Table 1**), which reached significance for AMPK-α1 knockouts, but only at the peak of Roll-Off (17 ± 6 versus −1 ± 3%, respectively; p < 0.05). Increases in minute ventilation during hypoxia also showed a tendency to be greater for AMPK-α1 and AMPK-α2 knockouts than for controls (AMPK-α1/α2 floxed). However this only reached significance during the initial Augmenting Phase of the HVR (**Figure 1C**; p < 0.05), which primarily results from increases in carotid body afferent input responses (Day and Wilson, 2007; Teppema and Dahan, 2010; Wilson and Teppema, 2016).

Perhaps most striking was the fact that deletion of neither the AMPK-α1 nor AMPK-α2 catalytic subunits in smooth muscles had any discernible effect on hypoxia-evoked apneas (**Figure 2**, **Table 1**, and **Supplementary Figure S2B**), when compared to controls (AMPK-α1/α2 floxed). This is evident from the fact that there was no difference in either the apnea frequency (**Figure 2A**), duration (**Figure 2B**), or apnea duration index (**Figure 2C**) in mice lacking AMPK-α1 or AMPK-α2 subunits in smooth muscles.

The aforementioned findings are in complete contrast to our observations on mice with AMPK-α1 deletion in catecholaminergic neurons (driven by TH-Cre) which exhibited marked attenuation of the HVR and pronounced increases in apnea frequency, duration and apnea duration index (Mahmoud et al., 2016).

mean ± SEM for the (A) apneic index (per minute), (B) apnea duration, and (C) apnea duration index (frequency × duration) of AMPK-α1/α2 floxed mice (Flx; black, n = 12), and for mice with AMPK-α1 (n = 4; magenta) and AMPK-α2 (n = 4; blue) deletion in smooth muscles (transgelin expressing cells).



A, Augmenting Phase; RO, Roll-Off; SP, Sustained Phase.

FIGURE 3 | Deletion of AMPK-α1 and AMPK-α2 subunits in smooth muscles has no effect on systemic arterial blood pressure during normoxia or hypoxia. (A) Bar charts show (mean ± SEM) the mean, systolic and diastolic blood pressure of control mice [AMPK-α1/α2 floxed (Flx); n = 4; black] and mice with AMPK-α1 (n = 4; magenta) and AMPK-α2 (n = 4; blue) deletion in smooth muscles (transgelin expressing cells). (B) Bar charts show the same measures of blood pressure during normoxia (N), after 10 min to hypoxia (H; 8% O2) and following recovery to normoxia (R). (C) Bar charts show the percentage change in blood pressure during hypoxia.

### Cardiovascular Responses to Hypoxia Remain Unaltered Following Deletion of AMPK-α1 and AMPK-α2 in Smooth Muscle Cells

We next assessed the effect of AMPK deficiency on cardiovascular function during hypoxia by monitoring systemic blood pressure during hypoxia. Deleting either AMPK-α1 or AMPK-α2 subunits in smooth muscle cells had little or no effect on resting blood pressure (**Figure 3A** and **Supplementary Figure S2C**). Furthermore neither smooth muscle AMPK-α1 nor -α2 deficiency affected hypoxia-evoked falls in systemic blood pressure in mice (**Figures 3B,C**, **Table 2**, and **Supplementary Figures S2Cii,iii**), which were comparable to the range of responses reported previously (Campen et al., 2005; Mahmoud et al., 2016).

That comparable levels of arterial hypoxia were achieved in all mice was confirmed by pulse oximetry, which demonstrated significant and similar falls in SpO<sup>2</sup> for all genotypes (**Figure 4** and **Supplementary Figure S2D**, p < 0.0001). Moreover there was no difference with respect to post-hypoxic recovery of arterial SpO<sup>2</sup> upon return to room air.

# DISCUSSION

The present investigation demonstrates that the HVR is, if anything, slightly augmented rather than attenuated by AMPK-α1 or AMPK-α2 deficiency in smooth muscles (driven by transgelin-Cre). This outcome is in marked contrast to our previously reported finding that AMPK-α1 deficiency in catecholaminergic neurons (driven by TH-Cre) markedly attenuates the HVR and precipitates hypoventilation and apnea during hypoxia (Mahmoud et al., 2016). Therefore, outcomes are consistent with our previously proposed model of central oxygen-sensing by catecholaminergic neurons of the brainstem respiratory network. Briefly, we proposed that these neurons deliver increases in respiratory drive during hypoxia in a manner supported by AMPK heterotrimers incorporating the catalytic α1 subunit, the activity of which we hypothesized to be determined by integration of local hypoxic stress at the brainstem with carotid body afferent input responses that provide an index of peripheral hypoxia. Notably, natural selection in high-altitude (Andean) populations has led to single nucleotide polymorphisms in PRKAA1 (Bigham et al., 2014).

FIGURE 4 | Comparable falls in arterial oxygen saturation were evoked in all genotypes during hypoxia. (A) Bar charts show mean ± SEM for the arterial oxygen saturation (SpO2) during normoxia, after 10 min of hypoxia (8% O2) and following recovery to normoxia. (B) Percentage change in SpO2. AMPK-α1/α2 floxed (Flx; black, n = 3), smooth muscle AMPK-α1 KO (magenta, n = 4) and smooth muscle AMPK-α2 KO (blue, n = 4) mice; ∗∗∗∗p < 0.001.


TABLE 2 | Means ± SEM of values obtained by tail cuff blood pressure measurements and percentage changes in arterial SpO<sup>2</sup> before, during, and after exposures to hypoxia (8%O2) across all genotypes.

M, mean blood pressure; Sys, systolic blood pressure; Dias, diastolic blood pressure; N, normoxia (room air); H, hypoxia (8% O2); R, recovery (room air).

The significance of our present findings lies in the fact that reduced cerebral arterial dilation could affect O<sup>2</sup> supply to the brain during hypoxia and thus the HVR through consequent respiratory depression, because the activity of brainstem respiratory networks is ultimately reliant on O<sup>2</sup> delivery via the vasculature that could be modulated systemically and locally through myogenic responses. Respiratory failure may also be triggered through the Cushing reflex consequent to increases in blood pressure (Guyenet, 2000; Paton et al., 2009), which would be exacerbated by block of arterial dilation consequent to AMPK deletion in arterial myocytes. Such systemspecific responses could well be affected by a number of mechanisms through which AMPK has been proposed to regulate myocyte function, rendering outcomes susceptible to off-target AMPK deletion due to "leakage" of Cre beyond those cells targeted by conditional deletion strategies. This was a distinct possibility, given that transient developmental expression of TH occurs in disparate cell groups that do not express TH in the adult (Lindeberg et al., 2004), including, for example, a subset of heart wall cells.

The outcomes of our present investigation indicate that while AMPK has been shown to mediate ex vivo arterial dilation in some circumstances (Goirand et al., 2007; Schneider et al., 2015), neither the expression of AMPK-α1 nor AMPKα2 catalytic subunits in smooth muscles is a pre-requisite for in vivo arterial dilation during hypoxia sufficient enough to impact systemic arterial blood pressures. This is evident because we observed little or no difference in peripheral hypoxic vasodilation between genotypes tested here. Moreover, systemic arterial pressures during normoxia and hypoxia were within the typical range previously reported for wild-type mice (Campen et al., 2004). Our findings do not, however, rule out a role for AMPK in maintaining resting vascular tone locally, or in governing organ-/tissue-specific perfusion and oxygen supply. Indeed, studies on other vascular beds have already revealed that AMPK might adjust local perfusion during hypoxia. For example, AMPK may mediate hypoxic pulmonary vasoconstriction (Evans et al., 2005, 2006; Evans, 2006), and thus assist ventilation-perfusion matching by diverting blood from oxygen deprived to oxygen rich areas of the lung (Bradford and Dean, 1894; von Euler and Liljestrand, 1946). Furthermore,

evidence suggests that AMPK supports dilation of systemic arteries, such as the aorta and mesenteric arteries, at the level of smooth muscles that may counter tissue hypoxemia (Schneider et al., 2015; Moral-Sanz et al., 2016). Accordingly, AMPK has also been implicated in the regulation of uterine artery reactivity during hypoxia (Skeffington et al., 2015), perhaps linking maternal metabolic and cardiovascular responses during pregnancy and governing oxygen and nutrient supply to the fetus.

The mechanisms involved in systemic arterial dilation during hypoxia include AMPK-dependent activation of SERCA and BKCa channels in systemic arterial myocytes (Schneider et al., 2015), while AMPK exerts its effects on pulmonary arterial smooth muscle cells, at least in part, through direct phosphorylation and inhibition of the voltagegated potassium channel KV1.5 (Moral-Sanz et al., 2016). Significant to the context of our study, KV1.5 availability has been proposed to impact cerebral myogenic responses (Koide et al., 2018), which could equally well be affected by loss of capacity for AMPK-dependent activation of SERCA and BKCa channels. Either way, our data argue strongly against the possibility that the HVR is influenced by AMPK-dependent mechanisms within smooth muscles that might affect local myogenic responses, irrespective of the tissue- and circulationspecific function they might impact. Nevertheless, it would be interesting to determine whether AMPK contributes to local autoregulatory mechanisms in the cerebral vasculature during hypoxia, even though we find no evidence of a significant contribution to the HVR or peripheral control of systemic blood pressures, which was the focus of this investigation. That said, we cannot rule out the possibility that the outcome of our previous studies using TH-Cre driven AMPK deletion might have been affected through off-target effects at the level of the vascular endothelium, because the conditional deletion strategy used here does not produce Cre expression in endothelial cells (El-Bizri et al., 2008). However, this seems unlikely given the outcomes of our present investigation. Therefore, hypoxic ventilatory dysfunction precipitated by AMPK-α1 deletion conditional on Cre expression in catecholaminergic neurons does not result, in whole or in part, from AMPK-α1 or AMPK-α2 deficiency in smooth muscles

and consequent changes in systemic arterial blood pressure during hypoxia.

In short, AMPK likely supports the HVR through neurogenic but not myogenic mechanisms as previously proposed, by supporting increased respiratory drive (Mahmoud et al., 2016) and perhaps functional hyperemia (Bucher et al., 2014), each of which may be coordinated by catecholaminergic neurons of the brainstem cardiorespiratory network.

### AUTHOR CONTRIBUTIONS

AME and SM wrote the manuscript. AME developed the conditonal AMPK knockout mice. AME and SM bred and genotyped the mice, performed plethysmography, blood pressure, and arterial oxygen saturation measurements, and analyzed the data.

### FUNDING

This work was funded by the Wellcome Trust (WT081195MA) and the British Heart Foundation (RG/12/14/29885).

### REFERENCES


### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphys. 2018.00655/full#supplementary-material

FIGURE S1 | End-point RT-PCR confirms AMPK deletion. End-point RT-PCR amplicons for transgelin and the catalytic α1 and α2 subunits of AMPK from lungs (right) and primary cultures of pulmonary arterial smooth muscle cells (PASMC, left) obtained from transgelin-Cre, C57BL6 (WT), AMPK-α1 and AMPK-α2 knockout mice.

FIGURE S2 | Deletion of AMPK-α1 or AMPK-α2 subunits in smooth muscles does not attenuate the HVR and has no effect on either apnea duration index or systemic arterial blood pressures during hypoxia. Box plots show the median (solid line), 25th to 75th percentiles (boxes), mean (+) and whiskers according to Tukey's method for: the % change of the ventilatory components of (Ai) breathing frequency, (Aii) tidal volume, and (Aiii) minute ventilation during the initial Augmenting Phase (A), the Roll-Off (RO), and the Sustained Phase (SP) of the hypoxic ventilatory response; (Bi) the apneic index (per minute), (Bii) apnea durations (in msec) and (Biii) apnea duration index (frequency x duration); Mean, Systolic and Diastolic blood pressures (in mmHg) measured (Ci) at rest, (Cii) before, during, and after a 10 min hypoxic challenge, and (Ciii) as % change during hypoxia; arterial oxygen saturation (in %) during (Di) normoxia, 10 min hypoxia, and recovery, and (Dii) as % change during hypoxia and recovery. <sup>∗</sup>p < 0.05; ∗∗∗p < 0.001.

O2-sensing cells? J. Biol. Chem. 280, 41504–41511. doi: 10.1074/jbc.M5100 40200



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 MacMillan and Evans. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Receptor–Receptor Interactions of G Protein-Coupled Receptors in the Carotid Body: A Working Hypothesis

Andrea Porzionato<sup>1</sup> \* † , Elena Stocco<sup>1</sup>† , Diego Guidolin<sup>1</sup> , Luigi Agnati<sup>2</sup> , Veronica Macchi<sup>1</sup> and Raffaele De Caro<sup>1</sup>

<sup>1</sup> Department of Neuroscience, University of Padua, Padua, Italy, <sup>2</sup> Department of Diagnostic, Clinical Medicine and Public Health, University of Modena and Reggio Emilia, Modena, Italy

In the carotid body (CB), a wide series of neurotransmitters and neuromodulators have been identified. They are mainly produced and released by type I cells and act on many different ionotropic and metabotropic receptors located in afferent nerve fibers, type I and II cells. Most metabotropic receptors are G protein-coupled receptors (GPCRs). In other transfected or native cells, GPCRs have been demonstrated to establish physical receptor–receptor interactions (RRIs) with formation of homo/heterocomplexes (dimers or receptor mosaics) in a dynamic monomer/oligomer equilibrium. RRIs modulate ligand binding, signaling, and internalization of GPCR protomers and they are considered of relevance for physiology, pharmacology, and pathology of the nervous system. We hypothesize that RRI may also occur in the different structural elements of the CB (type I cells, type II cells, and afferent fibers), with potential implications in chemoreception, neuromodulation, and tissue plasticity. This 'working hypothesis' is supported by literature data reporting the contemporary expression, in type I cells, type II cells, or afferent terminals, of GPCRs which are able to physically interact with each other to form homo/hetero-complexes. Functional data about cross-talks in the CB between different neurotransmitters/neuromodulators also support the hypothesis. On the basis of the above findings, the most significant homo/hetero-complexes which could be postulated in the CB include receptors for dopamine, adenosine, ATP, opioids, histamine, serotonin, endothelin, galanin, GABA, cannabinoids, angiotensin, neurotensin, and melatonin. From a methodological point of view, future studies should demonstrate the colocalization in close proximity (less than 10 nm) of the above receptors, through biophysical (i.e., bioluminescence/fluorescence resonance energy transfer, protein-fragment complementation assay, total internal reflection fluorescence microscopy, fluorescence correlation spectroscopy and photoactivated localization microscopy, X-ray crystallography) or biochemical (co-immunoprecipitation, in situ proximity ligation assay) methods. Moreover, functional approaches will be able to show if ligand binding to one receptor produces changes in the biochemical characteristics (ligand recognition, decoding, and trafficking processes) of the other(s). Plasticity aspects would be also of interest, as development and environmental stimuli (chronic continuous or intermittent hypoxia) produce changes in the expression of certain receptors which could potentially invest the dynamic monomer/oligomer equilibrium of homo/hetero-complexes and the correlated functional implications.

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Ricardo Fernandez, University of Los Lagos, Chile Joana F. Sacramento, Universidade Nova de Lisboa, Portugal

### \*Correspondence:

Andrea Porzionato andrea.porzionato@unipd.it †These authors have contributed equally to this work.

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 07 April 2018 Accepted: 18 May 2018 Published: 07 June 2018

### Citation:

Porzionato A, Stocco E, Guidolin D, Agnati L, Macchi V and De Caro R (2018) Receptor–Receptor Interactions of G Protein-Coupled Receptors in the Carotid Body: A Working Hypothesis. Front. Physiol. 9:697. doi: 10.3389/fphys.2018.00697

Keywords: carotid body, GPCRs, heteromers, receptor mosaics, hypoxia, adenosine, dopamine, plasticity

# INTRODUCTION

fphys-09-00697 June 5, 2018 Time: 15:3 # 2

The carotid body (CB) is a small polymodal peripheral chemoreceptor standing out for its basic role in case of hypoxia, hypercapnia, acidosis, low glucose; in these circumstances, it promotes an adequate respiratory and cardiovascular response as well as other less studied stimuli (Atanasova and Lazarov, 2014; Ortega-Sáenz et al., 2015). Two cell types can be distinguished in the CB, the 'neuron-like' chemosensitive type I cells and the 'glial-like' supportive type II cells. Type I cells synthetize and release many different neurotransmitters and neuromodulators, such as dopamine, acetylcholine, serotonin, histamine, ATP, glutamate, GABA, and substance P. Neurotransmitters and neuromodulators released by type I cells in great measure exert their activity on the afferent endings of the carotid sinus nerve (Porzionato et al., 2018), conveying the stimulation through the glossopharyngeal nerve and petrosal ganglion. The main excitatory stimuli are considered acetylcholine and ATP, binding to the ionotropic nicotinic and P2X2/P2X<sup>3</sup> receptors, respectively. However, autocrine/paracrine modulation of type I cells is also of particular importance and is mainly performed through a series of metabotropic G protein-coupled receptors (GPCRs), such as A2A, D1/2, H1/2/3, M1/2, 5-HT2A, and others (for comprehensive reviews on neurotransmitters in the CB, see Iturriaga and Alcayaga, 2004; Bairam and Carroll, 2005; Nurse, 2005, 2014). Conversely, metabotropic GPCRs are also present in afferent nerve endings and type II cells, where they play a role in neuromodulation of signaling between type I cells and afferent fibers (Tse et al., 2012; Nurse et al., 2018). Many authors have emphasized the absolute peculiarity of the CB, characterized by "a plethora of neurotransmitters and neuromodulators, as well as an even broader spectrum of their receptors" (Nurse, 2014). Apart from the above sensory innervation, the CB also shows sensory innervation from jugular and nodose ganglia, postganglionic sympathetic nerve fibers from the superior cervical ganglion, and preganglionic parasympathetic and sympathetic fibers reaching ganglion cells in the CB. Efferent parasympathetic and sympathetic innervation of the CB plays a pivotal role in modulation of blood flow (De Caro et al., 2013).

The CB also shows a high level of structural plasticity, undergoing structural and functional changes during development (e.g., Porzionato et al., 2008a,b; De Caro et al., 2013), aging (e.g., Di Giulio et al., 2012; Zara et al., 2013) and as a consequence of environmental stimuli, such as chronic continuous hypoxia, taking place during acclimation to high altitude (e.g., Wang and Bisgard, 2002). In response to chronic hypoxia, type II cells can act in a stem cell-like manner, differentiating into precursors of neural cells that originate mature glomus cells (Pardal et al., 2007; Platero-Luengo et al., 2014).

The GPCR family involves about 800 human receptors that are organized into five subfamilies, namely classes A (the largest group), B, C, frizzled, and adhesion (Foord, 2002; Guidolin et al., 2018). Considering the structure, GPCRs have a similar conformation showing a peculiar seven TM (7TM) α-helical region, an extracellular amino-terminal segment and an intracellular carboxy-terminal tail (Tuteja, 2009). While the domain that is localized in the extracellular portion allows the ligand(s) – receptor interaction, the TM region is subject to a conformational change when the binding with a ligand occurs; this change is then transmitted to the intracellular region of the GPCR responsible for activating the signaling cascade (Schonenbach et al., 2015; Stockert and Devi, 2015).

According to evidences from in vitro and in vivo studies, GPCR monomers can recognize/decode signals (Bayburt et al., 2007; Whorton et al., 2007; Kuszak et al., 2009; Lohse, 2010; Guidolin et al., 2018); in particular, an intrinsic plasticity characterizes GPCR monomers signaling, in fact GPCR activation can determine the onset of different signal transduction patterns, like the G proteins and/or arrestin pathways (Zidar et al., 2009; Guidolin et al., 2018). Furthermore, it is established that GPCRs can functionally interact by sharing signaling pathways or by mechanisms of transactivation (see Köse, 2017; Guidolin et al., 2018). In the 1980s, however, Agnati et al. (2005) provided evidence suggesting that GPCRs could also establish structural receptor–receptor interactions (RRIs; see Guidolin et al., 2015, 2018 for recent reviews), in which they can associate with themselves (homodimerization) or with other proteins of the same family (heterodimerization). This concept was later confirmed in transfected mammalian cell systems but also in many different native tissues (i.e., central nervous system; mammary gland; liver; cancer tissues) (see Ferré et al., 2014; Guidolin et al., 2015). It is a fact that class C GPCRs form constitutive homomers or heteromers (Kniazeff et al., 2011; Guidolin et al., 2018). An example is represented by the metabotropic receptor for gamma-aminobutyric acid b (GABAB), that, in the central nervous system, constitutes the major inhibitory neurotransmitter (Calver et al., 2000; Pontier et al., 2006). The receptor for GABAB, which is formed by the subunits GABAB1 and GABAB2, represents the typical example of constitutive heterodimer (Lohse, 2010). Only GABAB1 can bind GABA; however, GABAB2, even if alone is not functional, is required as GABAB1 as such is considered immature due to the presence of a carboxy-terminal ER retention motif (Galvez et al., 2001; Kamal and Jockers, 2011). Hence, after co-expression, the heterodimerization is the prerequisite to permit the masking of the retention domain and the subsequent translation of the active protein to the plasma-membrane (Terrillon and Bouvier, 2004). The oligomerization process in class A GPCR is a broadly debated topic of interest (see Franco et al., 2016; Guidolin et al., 2018). However, according to the consistent results obtained through multiple approaches, the possibility of class A GPCR complexes is strongly supported (Bouvier and Hébert, 2014; Guidolin et al., 2018). In this respect, studies highlighting the occurrence of a specific and dynamic monomer/oligomer equilibrium in class A homodimers and heterodimers are of particular interest (Kroeger et al., 2003; Palczewski, 2010; Kleinau et al., 2016; Farran, 2017). In fact, according to their half-lives (determined by the association and dissociation rates) class A GPCR dimers are demonstrated to be often transient (Gurevich and Gurevich, 2008). This may contribute in explaining the opposing views regarding the role of class A GPCR monomers versus oligomers (Guidolin et al., 2018).

The amount of data proving the existence of GPCR heteromers showed a huge increase similarly to both development and diffusion of biophysical techniques able to detect the spatial proximity of protein molecules (Guidolin et al., 2015). They involve resonance energy transfer methods including bioluminescence and fluorescence resonance energy transfer (BRET and FRET, respectively), fluorescence complementation, total internal fluorescence microscopy, fluorescence correlation spectroscopy, assays based on bivalent ligands and proximity ligation assays (PLAs).

The basic molecular mechanism leading to the formation of the receptor assemblies are allosteric interactions (Kenakin et al., 2010; Fuxe et al., 2012a). As recently outlined by Changeux and Christopoulos (2016), the cooperativity that emerges in the action of orthosteric and allosteric ligands of the GPCR forming the complex provides the cell decoding apparatus with sophisticated dynamics in terms of recognition and signaling (George et al., 2002; Guidolin et al., 2018). It has been shown, for instance, that GPCR heterodimerization exerts effects by altering or finetuning ligand binding, signaling, as well as internalization of GPCR protomers (Bulenger et al., 2005; Satake et al., 2013; Gomes et al., 2016; Guidolin et al., 2018). Angiotensin II receptor (AT1) – bradykinin receptor (B2) heterodimer (detected in smooth muscle, omental vessel, and platelets) provides an example: angiotensin II triggers inositol triphosphate (IP3) accumulation in a manner that is much more potent and effective than in the cells showing the expression of AT<sup>1</sup> alone; conversely, IP3 accumulation sustained by bradykinin is slightly weaker in cells expressing the AT1–B<sup>2</sup> heterodimer with respect to the cells expressing only B<sup>2</sup> (Satake et al., 2013).

Many evidences highlight that dimers may be formed by different mechanisms (Guidolin et al., 2018). Receptor complexes can arise in plasma membrane under the action of hydrophobic forces (Gahbauer and Böckmann, 2016). Receptor assembly may also occur in the endoplasmic reticulum (ER) before trafficking to the plasma membrane. ER is responsible of the structural quality of the synthetized receptors. Thus, only correctly folded receptors are allowed to exit while degradation occur for unfolded or misfolded proteins (Terrillon and Bouvier, 2004). Finally, recent data showed that GPCRs can be safely transferred by microvesicles from a source to a target cell where they become capable of recognizing and decoding their signal (Guescini et al., 2012).

Since GPCRs are implicated in several human diseases and represent major pharmacological targets, GPCR oligomerization process, determining variations in the processes of agonist recognition, signaling, and trafficking of participating receptors via allosteric mechanisms (Fuxe et al., 2012a), is of potential great importance for physiology and pharmacology, and new therapeutic strategies considering GPCR attitude to underwent toward several levels of receptor organization are under study (Farran, 2017).

Therefore, in a scenario where GPCR proteins are demonstrated to be present as monomer, the assumption that they may multimerize represents an interesting aspect to investigate. The purpose of the present article is to highlight indirect evidence of this with regard to the CB. In this tissue, indeed, the presence of multimeric GPCRs has not yet been consistently investigated. Thus, the main focus of the paper will be on possible homo-/heterodimers or mosaic receptors (already confirmed in other cell types) that could be present in the CB (type I and II cells, carotid sinus nerve afferents), with possible functional implications. The 'working hypothesis' of this paper could represent an interesting field of investigation aiming to clarify some aspects of chemoreception, neuromodulation and plasticity in the CB.

# INVESTIGATIVE APPROACHES TO GPCR COMPLEXES

As briefly discussed before, the term RRI emphasizes the occurrence of an interaction needing a direct physical contact between the involved receptor proteins responsible for the formation of multimeric assemblies of receptors (dimers or highorder oligomers) at the cell membrane level, acting as integrative input units of membrane-associated molecular circuits (Guidolin et al., 2018). Considerable efforts are currently being directed toward the identification of the most adequate strategy to discover such a supramolecular organization of GPCRs, but no single method is free from limitations giving rise to some discrepancy between various studies (Mackie, 2005; Lambert and Javitch, 2014). Thus, evidence obtained through multiple approaches with consistent results is needed to identify receptor complexes (Bouvier and Hébert, 2014).

In particular, an operational definition of RRI (Fuxe et al., 2010) can be exploited to devise experimental procedures fulfilling the following two measurable conditions:

1. The binding of a ligand to one receptor determines a detectable change in the biochemical features of another receptor, that includes ligand recognition, decoding, and trafficking processes;

2. Involved protomers have not only to be colocalized, but also in close proximity to each other (less than 10 nm), as indicated by specific techniques.

To explore the second point, biophysical and biochemical methods are presently available (see Kaczor and Selent, 2011 for a review).

# Biophysical Strategies

Biophysical methods may be divided into three macro-groups that are fluorescence/luminescence-based techniques, specialized microscopic techniques and X-ray crystallography.

As far as fluorescence/luminescence-based techniques are concerned, bioluminescence resonance energy transfer (BRET) and fluorescence resonance energy transfer (FRET) were among the earliest proximity-based assays of this type that were used in studying GPCR heteromerization (Fernández-Dueñas et al., 2012). Both techniques exploit the transfer of energy from a donor to an acceptor in a non-radiative manner (as a result of dipole–dipole coupling). If the donor is a fluorescent molecule with emission characteristics suitable to excite the acceptor, when they are in close proximity (∼10 nm) the result will be light emission from the acceptor fluorophore enabling the

visualization of the subcellular location of specific receptor complexes.

Protein-fragment complementation assays have also been used to analyze GPCR heteromerization in living cells (Gandia et al., 2008). In this approach, cells are transfected with one GPCR fused to a fragment of fluorophore and another GPCR fused to a complementary fragment. Reconstitution of the active fluorophore would occur only if the two receptors were in close proximity. It is noteworthy that a combination of this technique with BRET or FRET can be devised to demonstrate the presence of GPCR trimers or tetramers (Gandia et al., 2008). In this respect, a specialized microscopic technique, namely atomic force microscopy, has also been explored to demonstrate the existence of GPCR complexes (Agnati et al., 2010).

However, all the mentioned techniques show limitations in answering questions concerning the dynamic nature of receptor complexes. Thus, more recently, the details of the spatial and temporal organization of GPCR complexes have been addressed by using specialized microscopic techniques allowing the direct observation of the state and behavior of individual proteins in living cells. They include total internal reflection fluorescence microscopy (Hern et al., 2010; Guidolin et al., 2018), fluorescence correlation spectroscopy (Chen et al., 2003; Guidolin et al., 2018) and photoactivated localization microscopy (PALM; Jonas et al., 2016).

X-ray crystallography represents a further significant experimental approach in the field. During the last years, crystallization techniques have been object of interest (Grisshammer, 2017), determining important consequences for the analysis of GPCR complexes as well as an increase of the number of experimentally assessed structures (Guidolin et al., 2018).

### Biochemical Strategies

Spatial and temporal co-expression of GPCRs in cells and tissues is a necessary condition for the formation of more complex systems. This aspect can be investigated following biochemical approaches. The abundance of a specific mRNA can be detected by Northern-blot analysis, in situ hybridization, RT-PCR or microarray analysis. Immunohistochemistry and co-localization analysis (Agnati et al., 2005) can be performed to detect GPCR coexpression and co-localization at protein level. The practical limit of this technique regards the necessity for high-quality antibodies also due to low GPCRs expression levels.

To identify possible complex formation, it is mandatory for biochemical studies extracting GPCRs from the membrane by use of detergent, although shielding their native conformation in detergent is not a trivial task (Park et al., 2004).

Radiation inactivation technique (or target size method) take advantage of gamma rays or high energy electrons to disrupt polypeptides in order to identify the molecular weight of singular components. It is based on the inverse relationship existing between the size of a macromolecule and its dosedependent inactivation. Hence, the probability to destroy a molecule increases along with its mass: larger is a protein and lower is the energy required to destroy it (Rios et al., 2001). For what it concerns GPCRs, the main limitation of the method is that this variation may be related to interaction with other proteins instead of oligomerization.

To overcome this limitation, co-immunoprecipitation can be used (Skieterska et al., 2013). Interestingly, this method in combination with confocal microscopy and FRET was used to detect homooligomerization of the 5-HT2C receptors as well as the 5-HT<sup>4</sup> homodimerization in combination with BRET (Kaczor and Selent, 2011).

More recently, evidence of heteroreceptor complexes in native tissues has been provided by in situ PLA (Trifilieff et al., 2011). This technique could be adequate to highlight evidences regarding the presence of GPCR heteromers in tissue, giving new insights about basic biological mechanisms, heteroreceptor levels as well as their location. Thus, it allows to study the localization and modulation of heteroreceptor complexes, as formalin fixed tissue is used.

Cross-linking experiments are a further promising approach. They involve the use of assays based on bivalent ligands (agonists, antagonists, heteromer-selective antibodies, crosslinking reagents) allowing a direct targeting of the receptor complex (Yekkirala et al., 2013) in cells and tissues.

# HOMO- OR HETERO-DIMERS BETWEEN A<sup>1</sup> AND/OR A2A RECEPTORS

Adenosine receptor A<sup>1</sup> homodimers have been identified in brain cortex of pigs by radioligand binding experiments (Ciruela et al., 1995). The capability of A<sup>1</sup> receptors to form homodimers has also been demonstrated in Chinese hamster ovary (CHO) cells transfected with human A<sup>1</sup> receptors (Gracia et al., 2008) and in transfected human embryonic kidney (HEK-293T) cells expressing similar levels of A<sup>1</sup> receptor as the native tissue (Gracia et al., 2013). In bovine cortex, the existence of homomers was also shown through PLA (Gracia et al., 2013).

The effects of homodimerization on the receptor function can be inferred from ligand binding assays. A<sup>1</sup> receptor activation inhibits adenylate cyclase and decreases cAMP concentration. Caffeine is a non-selective adenosine receptor antagonist. In case of low caffeine concentrations (when caffeine only binds to one protomer of the empty homodimer), it determines an increase of the agonist affinity for the other protomer in the A<sup>1</sup> receptor homodimer; whereas, at high concentrations (when caffeine highly saturates both protomers of the homodimer) caffeine behaves like an A<sup>1</sup> antagonist with a reduction of the agonist binding to the receptors. Thus, caffeine modulates A<sup>1</sup> agonist-induced cAMP decrease in a biphasic manner. Interestingly, this is a particular behavior for a classical adenosine receptor antagonist like caffeine and is a pharmacological behavior explanation-lacking without taking into account receptor dimers as a minimal quaternary structure. Moreover, this pharmacological ability can also give explanations about the biphasic effects exerted at low and high concentration of caffeine on locomotor activity (Gracia et al., 2013).

As it regards the CB, adenosine is a well-known neurotransmitter produced in response to acute hypoxia. It can be directly released from type I cells, through an equilibrative

nucleoside transporter, or it can be indirectly produced through breakdown of released ATP by ecto-5<sup>0</sup> -nucleotidase (Conde et al., 2012). Adenosine may act pre- and post-synaptically in the modulation of type I and afferents function. It enhances the firing pattern in the carotid sinus afferents through A2A receptors (Conde et al., 2009a,b; Piskuric and Nurse, 2013). Some discrepancies between studies have been reported about A<sup>1</sup> localization into the CB structures (Bairam et al., 2009). Rocher et al. (1999) described the presence of A<sup>1</sup> in rabbit type I cells, through functional studies on cell cultures. Conversely, in the rat, post-synaptic localization has mainly been suggested as A<sup>1</sup> receptors are significantly expressed in the cytoplasm of nodous and petrosal ganglia (Bairam and Carroll, 2005; Gauda et al., 2006) and immunostaining of whole CB did not permit to specifically localize immunoreaction in type I cells (Bae et al., 2005). Possible species-related differences have been suggested and further analyses also on human material would be necessary. However, to reassume, A<sup>1</sup> homodimerization may be hypothesized in nerve fibers and probably in those species expressing A<sup>1</sup> receptors in type I cells.

A2A homodimerization has also been assessed in the plasma membrane of transfected HEK-293T cells by BRET, FRET, time-resolved BRET and immunoblotting and biotinylating experiments. Evidences highlight that more of 90% of A2A receptors are present as homodimers. In intracellular areas, the presence of a certain amount of monomeric species has been highlighted, suggesting that that they assemble into dimers before reaching cell surface (Canals et al., 2004; Lukasiewicz et al., 2007).

In particular, Canals et al. (2004) demonstrated that A2A homodimers are the functional species at the cell surface. Even though A2A homodimerization is quite constitutive, it is further influenced by specific ligands: agonists (i.e., CGS 21680) and antagonists (i.e., SCH 58261; caffeine) increase and decrease it, respectively (Lukasiewicz et al., 2007).

Various authors confirmed the presence of A2A receptor in the rat type I cells through several different techniques (RT-PCR studies, in situ hybridization analysis and immunohistochemistry) (e.g., Gauda, 2000; Kobayashi et al., 2000). Bairam et al. (2009) also demonstrated by Western-blot analysis the tendency of A2A receptor to form homodimer in the CB similarly to the superior cervical ganglion and nucleus tractus solitarius. This is one of the few specific references about receptor dimerization in the CB.

However, there are not studies addressing the issue of how agonist/antagonists may modulate dimerization of adenosine receptors in the nerve fibers and/or type I cells or, conversely, how dimerization may play a role in their pharmacological actions. It must also be considered that environmental conditions (hypoxia, etc.) may modulate receptor expression as well as types of dimerization, potentially modulating pharmacological effects.

Apart from the above homodimers, the ability of A<sup>1</sup> receptors to heterodimerize with A2A receptors was also demonstrated in transfected cells by Ferré et al. (2008). Heterodimerization of these two protomers was assessed in vitro and in vivo by Ciruela et al. (2006), who confirmed the A1/A2A heterodimer formation in rat striatal glutamatergic nerve terminals by immunogold detection and co-immunoprecipitation. Recently it has been shown that at the presynaptic membrane of cortico-thalamic glutamatergic terminals A<sup>1</sup> receptors co-localizes and interacts with A2A receptors, forming functional receptor heterodimer in the striatum (Fernández-Dueñas et al., 2017).

The eventual demonstration of A1/A2A heterodimers in type I cells and/or nerve fibers would be of particular interest in terms of modulation of the glutamatergic transmission. In fact, heterodimerization between the adenosine receptors A<sup>1</sup> and A2A, which are responsible for opposite signaling pathways (inhibitory and excitatory actions, respectively) (Stockwell et al., 2016), has been suggested to exert a function in fine-tuning modulation of striatal glutamatergic neurotransmission by adenosine. Since A<sup>1</sup> receptor shows higher affinity for adenosine than A2A, but A<sup>1</sup> agonist affinity decreases when A2A is activated, the possibility exists that glutamate release may be inhibited or stimulated by a switch mechanism depending on low and high concentrations of adenosine, respectively (Ciruela et al., 2006; Doyle et al., 2012).

### HETERO-DIMERS BETWEEN ADENOSINE (A<sup>1</sup> AND A2A) AND PURINERGIC (P2Y1, P2Y2, P2Y12) RECEPTORS

P2Y<sup>1</sup> (Xu et al., 2005; Tse et al., 2012) and P2Y<sup>12</sup> (Carroll et al., 2012) receptors have been demonstrated in type I cells, where they inhibit the hypoxia-induced rise in intracellular Ca2+. Conversely, in type II cells, P2Y<sup>2</sup> receptors have been demonstrated (Xu et al., 2003), which produces a rise in intracellular Ca2+.

Purinergic receptors have also been reported to form homodimers and hetero-dimers with other purinergic and adenosine receptors (e.g., Schicker et al., 2009). As previously detailed, A2A receptors are expressed in type I cells (Gauda, 2000; Kobayashi et al., 2000) and A1 receptors have been reported in the type I cells of some species (rabbit) (Rocher et al., 1999), although not confirmed in others (rat) (Gauda et al., 2000, 2006). Thus, in type I cells, many RRIs can be hypothesized: P2Y1/P2Y1, P2Y1/P2Y12, P2Y12/P2Y12, P2Y1/A1, P2Y1/A2A, P2Y12/A1, P2Y12/A2A (e.g., Yoshioka et al., 2001; Nakata et al., 2005, 2010; Schicker et al., 2009). Physical interactions between A<sup>1</sup> and P2Y<sup>1</sup> receptors, for instance, produce a conformational change in the A<sup>1</sup> binding pocket with acquisition of P2Y1-like agonistic pharmacology, i.e., a P2Y<sup>1</sup> agonist may bind to the A<sup>1</sup> receptor and produce an inhibition of adenylate cyclase which is prevented by A<sup>1</sup> antagonist (Fuxe et al., 2008). All the above RRIs would be particularly intriguing as they would represent other ways of reciprocal modulation between purinergic and adenosine neurotransmission both in type I and II cells.

# HOMO- AND HETERO-DIMERS BETWEEN DOPAMINE RECEPTORS (D<sup>1</sup> AND D2)

According to co-immunoprecipitation data gathered from both rat brain and cells showing co-expression of the D<sup>1</sup>

and D<sup>2</sup> receptors, occurs the idea they may constitute the same heteromeric protein complex; moreover, according to immunohistochemistry studies, these receptors are convincingly co-expressed and co-localized within neurons of human and rat brain (Lee et al., 2004). For instance, a significant neuronal subpopulation in rat nucleus accumbens co-expresses D<sup>1</sup> and D<sup>2</sup> receptors, which can form a D1/D<sup>2</sup> receptor complex (Hopf et al., 2003; Hasbi et al., 2018). Rashid et al. (2007) also provided evidences concerning their co-expression in human embryonic kidney cells. In parallel, Rashid et al. (2007) and Hasbi et al. (2018) demonstrated the ability of D<sup>1</sup> and D<sup>2</sup> receptors to oligomerize in vivo in rodent and monkey striata, respectively, through in situ PLA, in situ FRET and coimmunoprecipitation.

D<sup>1</sup> and D<sup>2</sup> monomers are coupled to G<sup>s</sup> and G<sup>i</sup> proteins, respectively, and they are usually considered to exert opposite effects at the cellular level (Hopf et al., 2003). Conversely, the heterodimer D1/D<sup>2</sup> is coupled to Gq/11. In the nucleus accumbens, the activation of this heterodimer-specific pathway determines an increase of calcium/calmodulin-dependent protein kinase IIα in contrast with the effect produced by the activation of the Gs-D<sup>1</sup> receptor (Rashid et al., 2007).

Numerous studies documented the presence of D<sup>1</sup> receptor in the CB of various animals through Northern blot analysis, RT-PCR and pharmacological studies, although the precise location has not been verified (Almaraz et al., 1991; Bairam et al., 1998, 2003; Schlenker, 2008). Almaraz et al. (1991) suggested expression in blood vessels. Bairam et al. (1998) suggested that D<sup>1</sup> receptor could be expressed in nerve terminals (sensitive and sympathetic). Conversely, D<sup>1</sup> receptors were identified in hamster type I cells through immunohistochemistry (Schlenker and Schultz, 2011). D<sup>2</sup> has been demonstrated in type I cells and nerve fibers of the rat CB by double immunofluorescence (Wakai et al., 2015) and its expression is higher than D<sup>1</sup> (Bairam et al., 1998). The contemporary presence of D<sup>1</sup> and D<sup>2</sup> in type I cells and nerve fibers would support the hypothesis of heterodimerization in these structures. Exogenous dopamine may also inhibit Ca2<sup>+</sup> responses in type II cells, supporting the presence of corresponding receptors (Leonard and Nurse, 2017; Leonard et al., 2018) and of possible dimerization.

Huey and Powell (2000) explored the consequences of chronic hypoxia on D<sup>2</sup> receptor expression in the arterial chemoreflex pathway. As regards the CB, the authors observed that the regulation of D2-receptor mRNA expression is time-dependent. Briefly, an initial increase of D<sup>2</sup> receptor mRNA levels was observed after 6 and 12 h, with a subsequent decrease (24, 48 h) and a final significant increase after 168 h of hypoxia. In a noteworthy and analog study, Bairam et al. (2003) considered the effects induced by hypoxia on the expression levels of both D<sup>1</sup> and D<sup>2</sup> receptor mRNAs. The authors highlighted that in 1-day-old rabbits hypoxia affects the expression of D2 or D1-receptors mRNA in a manner that is age-dependent. D2- and D1-receptors transcript levels increased and decreased after exposure to moderate (15% O2) and severe (8% O2) hypoxia, respectively. Conversely, D2- and D1-transcript levels decreased independently of hypoxia intensities in adult rabbits. Unlike D<sup>1</sup> receptors, changes in D<sup>2</sup> receptor mRNA levels were independent on exposure time. The D<sup>2</sup> role in the CB has been suggested not to be key under normal conditions, but rather in situations of chronic hypoxia, where D<sup>1</sup> and D<sup>2</sup> receptors density tends to change as previously discussed (Bairam et al., 2003).

The above changes in the expression of the different dopamine receptors after hypoxia suggest the possibility of consequent changes in the amount of D1/D<sup>2</sup> heterodimers. It is intriguing the idea that D1/D<sup>2</sup> heterodimer formation may be under the influence of environmental stimuli, mainly hypoxia but possibly others, in the CB.

# HETERO-DIMERS BETWEEN ADENOSINE (A<sup>1</sup> AND A2A) AND DOPAMINE (D<sup>1</sup> AND D2) RECEPTORS

The co-localization of A<sup>1</sup> and D<sup>2</sup> receptors has been showed by immunofluorescence in rat cerebral cortex neurons (Gines et al., 2000). Moreover, the presence of A1/D<sup>1</sup> complexes was detected in mouse fibroblast Ltk-cells, transfected with human A1, D1, and D2. While the formation of A1/D<sup>1</sup> heteromers was confirmed by coimmunoprecipitation, A<sup>1</sup> did not seem to form heterodimers with D2. A<sup>1</sup> and D<sup>1</sup> receptors, separately, also tend to form homodimers; however, the preferred form of dimerization for these two receptors still needs to be elucidated (Agnati et al., 2003). Additionally, A1/D<sup>1</sup> dimerization was also confirmed by FRET in HEK-293T cells (Shen et al., 2013).

Due to the possible presence of both A<sup>1</sup> and D<sup>1</sup> receptors in type I cells, a role by A1/D<sup>1</sup> dimerization may be also be supposed. In particular, A1/D<sup>1</sup> dimerization could represent a way of integration between agonists and/or antagonists for the two different monomers. For instance, Gines et al. (2000) suggested that A1/D<sup>1</sup> heteromerization may have a role in the antagonistic modulation of D<sup>1</sup> receptor in the brain, thus having an impact on desensitization mechanism of D<sup>1</sup> and receptor trafficking.

A2A/D<sup>2</sup> heterodimers have also been demonstrated both in transfected SH-SY5Y (Hillion et al., 2002; Xie et al., 2010) and HEK-293T cells (Navarro et al., 2014), by immunoprecipitation followed by Western-blotting (SH-SY5Y) and BRET (HEK-293T). In addition, A2A/D<sup>2</sup> heterodimers were identified in neuronal primary cultures of rat striatum by PLA (Navarro et al., 2014) and cAMP accumulation experiments (Hillion et al., 2002). A2A/D<sup>2</sup> heterodimers have also been demonstrated in the mammalian striatum, with particular reference to the striatal enkephalin-containing GABAergic neurons that project to the globus pallidus and comprise the so-called indirect pathway (Fink et al., 1992; Fuxe et al., 1998; Schiffmann et al., 2003; Trifilieff et al., 2011; Doyle et al., 2012; Atack et al., 2014). Thus, A2A/D<sup>2</sup> heterodimers play an important role in the modulation of GABAergic striato-pallidal neuronal function (Bonaventura et al., 2015). In particular, antagonistic A2A-D<sup>2</sup> RRIs occur in the heterodimer, as demonstrated in striatal membrane preparations after incubation with the A2A

agonist CGS21680, leading to a decrease of the affinity of the high affinity D<sup>2</sup> agonist-binding site (Fuxe et al., 1998; Guidolin et al., 2018). Relationships between A2A and D<sup>2</sup> and adenosine/dopamine cross-talks have been proposed as possible new therapeutic approaches for Parkinson's disease, schizophrenia, and drug addiction (Canals et al., 2003; Guidolin et al., 2015).

As previously discussed, both A2A and D<sup>2</sup> are expressed in type I cells of the CB, supporting the hypothesis of A2A/D<sup>2</sup> heterodimer formation also in these cells. The reciprocal influences of the two receptor monomers in the A2A/D<sup>2</sup> complex would be particularly intriguing, as adenosine and dopamine are among the main excitatory and inhibitory neurotransmitters in the CB. Some authors have proposed a functional interaction between A2B and D<sup>2</sup> receptors. In particular, if adenosine levels would become too high, activation of the low affinity A2B would lead to increased dopamine secretion from type I cells (Conde et al., 2006; Livermore and Nurse, 2013) and inhibition of sensory discharge through pre- and post-synaptic D<sup>2</sup> receptors (Conde et al., 2009a,b, 2012; Zhang et al., 2017; Nurse et al., 2018). RRI between A2A and D<sup>2</sup> could represent an opposite way of neuromodulation, increasing signaling efficiency in the presence of low adenosine levels, through allosteric receptor–receptor inhibition of D<sup>2</sup> signaling.

In rat type I cells, Conde et al. (2008, 2009b) also proposed a A2B-D<sup>2</sup> receptor interactions on the basis of in vitro pharmacological experiments, although they did not demonstrate if physical RRI occur but only demonstrated an interaction at the adenylyl cyclase level. In particular, they showed that D<sup>2</sup> agonists decrease catecholamine release and inhibit cAMP production in the CB and that these effects are prevented by A2B agonists; conversely, D<sup>2</sup> antagonists increase the release of catecholamines, this phenomenon being also prevented by A2B antagonists. A2B-D<sup>2</sup> RRI, however, have not yet been demonstrated to date in any cell type.

## HOMO- AND HETERO-DIMERS BETWEEN OPIOID RECEPTORS (µOR AND δOR)

µOR forms homodimers (Lopez and Salome, 2009) with an elevated stability, as demonstrated in transfected HEK-293 cells by BRET and radioligand binding assays (Wang et al., 2005). µOR/µOR homodimerization occurs prior of transportation to the cell membrane. Sarkar et al. (2012) also proved µOR homodimerization in rat NK cells through Western-blot analysis and immunoprecipitation as well as immunohistochemistry and immunofluorescence.

Also δOR homodimerization was confirmed by Westernblotting and co-immunoprecipitation in different transfected cellular systems, such as CHO cells and COS cells (Cvejic and Devi, 1997). Data obtained by cross-linking experiments in CHO cells indicated that homodimerization does not depend on the expression level of the receptor. Subsequently, the occurrence of homodimers was also assessed in HEK-293T cells by BRET (Johnston et al., 2011) and in rat NK cells by Western-blot assays, immunoprecipitation, immunohistochemistry, and immunofluorescence (Sarkar et al., 2012).

δOR/µOR heterodimers have also been demonstrated (George et al., 2000; Gomes et al., 2002; Sarkar et al., 2012). The heteroreceptor complexes exhibit distinct pharmacological properties from that of the monomers. Low non-signaling doses of δ- or µ-ligands potentiate the binding and signaling of the receptors and the affinity of endomorphin-1 and δ-selective agonists is increased in the heteroreceptor complex, when compared to monomers (Kabli et al., 2010). These properties appeared linked to changes in the signaling pathways activated by the heteroreceptor complexes as compared to individual receptors (Gomes et al., 2013). In NK cells, in which δOR and µOR proteins exist as homo- and heterodimers, differences in cell function were observed depending on the prevalent type of receptor dimerization (Sarkar et al., 2012).

As suggested by Cvejic and Devi (1997) the dimerization also plays an important role in the internalization of the δOR receptor which in its monomeric form may not be internalized. This could be related to the mechanism that takes place after long-term exposure to opioid ligands. Dimerization of δOR receptors seems to be agonist dependent.

As it regards the CB, both µOR and δOR are expressed in type I cells (Ichikawa et al., 2005; Ricker et al., 2015). The expression of both receptors in type I cells suggests the occurrence of homo- and hetero-dimerization, with possible pharmacological implications. Perivascular nerve fibers have also been reported to express δOR (Ichikawa et al., 2005) although further studies will be necessary for detection of other opioid receptor types.

## HETERO-DIMERS BETWEEN DOPAMINE (D1) AND OPIOID (µOR) RECEPTORS

Co-immunoprecipitation, BRET and cross-antagonism assays demonstrated the existence of D1/µOR heterodimers in both transfected HEK-293 cells and ventral striatum (Tao et al., 2017). D<sup>1</sup> receptor antagonist can antagonize µOR-mediated signaling and function in a dopamine-independent manner, mostly likely via allosteric interactions through the D1/µOR heteromer. D<sup>1</sup> antagonist (SCH23390) prevented opiate-induced activation of G-protein, inhibition of adenylyl cyclase, phosphorylation of ERK 1/2 and expression of c-fos in transfected cells expressing both receptors and in striatal tissues from wild type, but not DR KO, mice. The ability of an antagonist to inhibit signals originated by stimulation of the partner receptor is a biochemical characteristic that has been described for receptor heterodimers (cross-antagonism) (Tao et al., 2017).

As discussed above, D<sup>1</sup> receptor and µOR were both assessed in the CB, supporting the hypothesis of corresponding RRI. Moreover, changes in the dynamic monomer/oligomer equilibrium may be postulated on the basis of environmental stimuli, as D<sup>1</sup> expression levels, for instance, decrease in hypoxia conditions and along with exposure time (Bairam et al., 2003).

# HETERO-DIMERS BETWEEN DOPAMINE (D<sup>1</sup> AND D2) AND HISTAMINE (H3) RECEPTORS

Koerner et al. (2004) assessed the presence of histamine receptors (H1, H2, H3) in the CB through RT-PCR studies. In particular, H<sup>3</sup> has been detected in rat type I cells by immunohistochemistry (Lazarov et al., 2009; Del Rio et al., 2009; Thompson et al., 2010). H<sup>3</sup> antagonism results in increased chemosensory activity (Del Rio et al., 2009) and H3 activation inhibits the intracellular Ca2<sup>+</sup> signaling mediated by activation of muscarinic receptors in type I cells (Thompson et al., 2010). On the basis of the above and following findings, histamine receptors could also be included in the series of GPCRs with potential oligomerization in the CB.

In fact, Ferrada et al. (2009) demonstrated the D1–H<sup>3</sup> receptor heteromerization in mammalian transfected cells (HEK-293) by BRET and binding assays as well as in the neuronal cell model SK-N-MC by radioligand experiments. The presence of this heteromer in the brain (striatum) was then also proved by Moreno et al. (2011).

An antagonist acting on one of the receptor units that constitute the D1–H<sup>3</sup> receptor heteromer can cause changes in conformation of the other receptor unit, thus, blocking specific signals that originate in the heteromer. This mechanism could be responsible for unsuspected GPCR antagonistsrelated therapeutic potentials (Ferrada et al., 2009). D1/H<sup>3</sup> receptor heteromers work as processors integrating dopamineas well as histamine-related signals involved in controlling the function of striatal neurons of the direct striatal pathway (Moreno et al., 2011).

A strong and selective heteromeric interaction between D<sup>2</sup> and H<sup>3</sup> was also assessed by radioligand binding experiments in striatal membrane preparations and in co-transfected HEK-293 cells by BRET (Ferrada et al., 2008). According to agonist/antagonist competition experiments, a significant decrease of D<sup>2</sup> receptors affinity for the agonist was encountered after stimulation of H<sup>3</sup> receptors with several H<sup>3</sup> agonists (Ferrada et al., 2008).

### HETERO-DIMERS BETWEEN DOPAMINE (D2) AND SEROTONIN (5-HT2A) RECEPTORS

5-HT2A/D<sup>2</sup> heteroreceptor complexes were demonstrated by PLA and immunofluorescence in co-transfected HEK-293T cells (Lukasiewicz et al., 2010; Borroto-Escuela et al., 2014), as well as in discrete regions of the ventral and dorsal striatum and nucleus accumbens (Borroto-Escuela et al., 2014), medial prefrontal cortex as well as in the pars reticulate of the substantia nigra in rat (Lukasiewicz et al., 2010).

The functional properties of this complex are not yet completely clear (Lukasiewicz et al., 2010). However, ligands with high 5-HT2A/D<sup>2</sup> selectivity and partial agonistic activity on D<sup>2</sup> have been recently identified as new antipsychotic drugs which do not significantly induce extrapyramidal side effects (Möller et al., 2015).

Serotonin is released by type I cells and it acts in a autocrine/paracrine manner on 5-HT<sup>2</sup> receptors which are localized in type I cells, as proved by immunofluorescence and pharmacological/functional studies in the rat CB (Zhang et al., 2003; Jacono et al., 2005; Liu et al., 2011; Yokoyama et al., 2015). Possible expression of 5-HT2A in carotid sinus nerve terminals has also been reported (Nurse et al., 2018). Serotonin is also considered to be involved in long-term facilitation of CB sensory discharge due to chronic intermittent hypoxia (Peng et al., 2009; Prabhakar, 2011). The presence of 5-HT2A and D<sup>2</sup> in type I cells suggests the possibility of heterodimer formation, with functional roles to be investigated. 5-HT2A and D<sup>2</sup> mediate excitatory and inhibitory effects on type I cells and the role of dimerization in the modulation of the two response types is particularly intriguing.

### HETERO-DIMERS BETWEEN DOPAMINE (D2) AND NEUROTENSIN (NTS1) RECEPTORS

Among CB receptors possibly interacting with D<sup>2</sup> receptor there is also the neurotensin receptor 1 (NTS1), which has been demonstrated by immunohistochemistry in rat and human type I cells (Porzionato et al., 2009). D2-NTS<sup>1</sup> complexes have also been highlighted in living HEK293T cells by BRET technology (Borroto-Escuela et al., 2013b). Through these RRI neurotensin reduces dopamine binding and function at D<sup>2</sup> receptor, moving dopamine transmission toward D<sup>1</sup> receptor, which is not antagonized by NTS<sup>1</sup> (Borroto-Escuela and Fuxe, 2017).

# HOMO- AND HETERO-DIMERS BETWEEN MELATONIN RECEPTORS (MT<sup>1</sup> AND MT2)

MT<sup>2</sup> receptors have the ability to form homodimers, as first demonstrated in HEK-293T transfected cells through BRET approaches (including single-cell BRET experiments), as well as Western-blotting and co-immunoprecipitation assays (Ayoub et al., 2002). However, it has also been observed that MT<sup>2</sup> receptors tend to preferentially form heterodimers with MT1, as proved by BRET techniques (BRET donor saturation assay) in the same cell model (Ayoub et al., 2004).

The homodimerization of MT<sup>2</sup> receptors is not modulated by ligands as it can exist in a constitutive manner in living cells. However, the speculated role of MT<sup>2</sup> dimers is that they are required for explicating biological functions of cells being considered functional signaling units. Modulation of the signaling and trafficking occurs similarly to the GABA<sup>B</sup> receptors (Ayoub et al., 2002).

As regards the presence of MT receptors in the CB, an in situ hybridization study showed MT<sup>2</sup> expression in the rat type I cells whereas there are not yet data about MT<sup>1</sup> (Tjong et al., 2004).

Thus, to date, we can suppose MT<sup>2</sup> homodimerization in type I cells although further analyses could be of interest for MT<sup>1</sup> detection.

# GALANIN RECEPTOR (Gal1, Gal2) HETEROMERS

Galanin (Mazzatenta et al., 2014) and galanin receptors 1 (Gal1) and 2 (Gal2), but not 3 (Gal3), (Porzionato et al., 2010) have been demonstrated by immunohistochemistry in rat type I cells. Galanin is known to regulate the differentiation of neural stem cells and plasticity responses in the nervous system (e.g., Cordeo-Llana et al., 2014) and it has been suggested that galanin expression in chemoreceptor cells could provide a signal for neurogenesis and chemoreceptor cell differentiation (Di Giulio et al., 2015; Mazzatenta et al., 2016).

As it concerns this paper, galanin receptors are particularly intriguing because evidence is given about the possibility of Gal1- Gal2, D1-Gal1, Gal-NPYY1, Gal1-Gal2-NPYY<sup>2</sup> or Gal1-Gal2-AT<sup>1</sup> heteromers, which can be postulated in type I cells (reviewed in Fuxe et al., 2012b).

Although there are not specific data about the expression of the different NPY receptor types in the CB, NPY immunoreactivity has been identified in nerve fibers as well as type I cells of dog, monkey and rat CB (Oomori et al., 1991, 2002). In the rat CB, a reduction in NPY-immunoreactive type I cells was observed from postnatal week 2 onward; conversely, NPY-immunoreactive fibers mainly increase since week 2 after birth (Oomori et al., 2002).

The local renin-angiotensin system in the CB has been extensively studied in the past years (e.g., Fung et al., 2002; Fung, 2014; Lam et al., 2014) and angiotensin II receptor type 1 (AT1) has been identified in rat type I (Fung et al., 2001; Atanasova et al., 2018) and type II (Murali et al., 2014) cells. In the rat, AT<sup>1</sup> receptors are also up-regulated in response to hypoxia (Fung et al., 2002).

## HOMO- AND HETERO-DIMERS BETWEEN ENDOTHELIN RECEPTORS (ET<sup>A</sup> AND ETB)

Endothelin 1 (ET-1) receptors type A (ETA) and B (ETB) may form homo- and heterodimers, although the functional implications of different dimerization are still largely unknown. Their behavior was studied after expression in transfected HEK-293T cells by immunoprecipitation, immunoblotting and FRET (Evans and Walker, 2008). These results confirmed the previous data by Gregan et al. (2004) who adopted FRET too, co-localizing ET-1 receptors at the plasma membrane. Furthermore, the latter speculated that ET-1-receptors homodimerization (as well as homo-oligomerization) is constitutive and ligand-independent (Evans and Walker, 2008).

Type I cells synthetize and release ET-1, which may act in autocrine/paracrine manner on the same type I cells through both the above receptors. Chronic continuous hypoxia upregulates ET-1 and ET<sup>A</sup> receptor, suggesting a critical role in chronic hypoxia-induced increased chemosensitivity in the rat CB (Chen et al., 2002). Chronic intermittent hypoxia is also known to enhance the CB chemosensory and ventilatory responses to acute hypoxia and produce long-term sensory potentiation of chemosensory discharges (e.g., Peng et al., 2003, 2013; Rey et al., 2004; Pawar et al., 2009). Chronic intermittent hypoxia has been reported to determine an upregulation of ET-1 and ET<sup>B</sup> receptor, but not of ET<sup>A</sup> receptor, in adult cats (Rey et al., 2007); conversely, upregulation of ET-1 and ET<sup>A</sup> receptor, but not ET<sup>B</sup> receptor, have been reported in neonatal (Pawar et al., 2009) and adult (Peng et al., 2013) rats.

Endothelin rise intracellular Ca2<sup>+</sup> in cultured type II cells (Murali et al., 2015). Stimulation of ET<sup>B</sup> receptors by ET-1 has been reported to play a role in proliferation of stem cells derived from type II cells and CB hyperplasia following chronic hypoxia (Platero-Luengo et al., 2014).

Thus, RRIs between ET-1 receptors can be hypothesized in type I and II cells. Hypoxia-induced changes in the expression of two different receptor types, together with the involvement of ET-1 in functional/structural modifications, suggests the idea that changes in the monomer/(homo/hetero)-dimer equilibrium may also play a role.

ET<sup>B</sup> receptor has also been reported to undergo physical interactions with AT<sup>1</sup> receptor located in the cells of renal proximal tubule (Zeng et al., 2005). RRI between the two receptors would be particularly intriguing in the CB, due to their roles in hypoxia responses.

### HETERO-DIMERS BETWEEN GABA (GABAB2) AND MUSCARINIC (M2) RECEPTORS

Boyer et al. (2009) demonstrated by FRET that GABAB<sup>2</sup> is co-localized and directly associates with M<sup>2</sup> receptor in neuronal PC12 cells. In parallel, the authors observed that in another cell model, i.e., HEK-293T, GABAB<sup>1</sup> is also required for GABAB2/M<sup>2</sup> heterodimerization, contrary to what was observed in PC12. Moreover, through co-immunoprecipitation and immunostaining experiments, the authors supported that signaling complexes GABAB2/M<sup>2</sup> exist also in vivo in the brain cortex. The association seems to be specific since GABAB<sup>2</sup> did not associate closely with other related muscarinic receptors (M1) or with a different GPCR (µOR).

The findings that M<sup>2</sup> and GABAB<sup>2</sup> are co-localized and associate in cortical neurons, which overlap with brain regions that receive cholinergic projections, suggest that the heterodimer is involved in a novel mechanism for enhancing cholinergic signaling in the brain. In fact, expression of GABAB<sup>2</sup> in M2 expressing neurons would allow some neurons to maintain muscarinic signaling during elevated or chronic agonist exposure (Boyer et al., 2009).

Type I cells of the CB express both GABAB<sup>1</sup> and GABAB<sup>2</sup> subunits, as confirmed by RT-PCR studies. In addition to molecular biology evidences, localization of GABAB<sup>2</sup>

receptor subunits in sections of the CB was also assayed by immunofluorescence which assessed positive immunoreactivity for the receptor subunit in type I clusters (Fearon et al., 2003). Shirahata et al. (2004) demonstrated the expression of M<sup>2</sup> mRNA and the presence of the receptor protein, by RT-PCR and immunohistochemistry, respectively. In particular, regarding localization, immunohistochemical analysis highlighted M<sup>2</sup> presence in type I cells and petrosal afferent terminals. GABA and acetylcholine (Dasso et al., 1997) show inhibitory and excitatory actions, respectively, on type I cells. The possible occurrence of dimerization may have a role in the reciprocal modulation of actions of the two neurotransmitters.

Muscarinic receptors are also present in type II cells, as muscarinic agonists elicit an increase of intracellular Ca2<sup>+</sup> levels in these cells (Tse et al., 2012; Murali et al., 2015), but there are not data about the possible expression of GABA receptors too.

### CANNABINOID RECEPTOR (CB1) HETEROMERS

The CB<sup>1</sup> receptor was found to be expressed in the CB by techniques such as RT-PCR, immunohistochemistry and in situ hybridization, although its levels were relatively low (McLemore, 2004). These findings suggest that endocannabinoids may have an impact on blood flow regulation in the CB, therefore affecting the oxygen pressure and respiratory control. However, the CB<sup>1</sup> function in CB needs to be further investigated, as data present in another study on CB<sup>1</sup> were somewhat discording (Roy et al., 2012).

The CB<sup>1</sup> can form heterodimers with many other different GPCRs which are also assessed to be present in the CB; among these δOR, µOR, A2A and D<sup>2</sup> receptors. In particular, the ability of CB<sup>1</sup> and D<sup>2</sup> receptors to form heterodimers was demonstrated in co-transfected cells (HEK-293T) through co-immunoprecipitation and FRET techniques (Kearn, 2005; Marcellino et al., 2008). It is noteworthy that the activation of CB1-D<sup>2</sup> receptor heterodimer can have completely opposite effects than activation of the individual receptors.

The above receptors have also been demonstrated to oligomerize in receptor mosaics so that the presence of these complexes may also be postulated for the CB. In particular, the existence of a CB1-D2-A2A receptor mosaic has been demonstrated, where CB<sup>1</sup> receptor activation removes the D<sup>2</sup> inhibition of the A2A receptor signaling (Fuxe et al., 2008; Marcellino et al., 2008). The above receptor mosaicism could represent a further way of subtle reciprocal modulation between different neurotransmitters.

### HETEROCOMPLEXES BETWEEN GPCR AND OTHER RECEPTOR TYPES

G protein-coupled receptors may undergo direct interactions with other membrane receptors, such as ion channel receptors or receptor tyrosine kinases (Borroto-Escuela et al., 2016). Some of these complexes can be postulated in the CB, with possible roles in modulation of neurotransmission and plasticity.

For instance, NMDA receptor subunits 1, 2A and 2B have been detected in CB, through RT-PCR, and type I cells, through immunohistochemistry. Chronic intermittent hypoxia also increases the expression of NMDA<sup>1</sup> and NMDA2B receptors (Liu et al., 2009). It is noteworthy that D<sup>1</sup> receptors can regulate the function of the NMDA receptor by means of direct protein–protein interactions between the carboxyl terminals of D<sup>1</sup> receptor and NMDA<sup>1</sup> and NMDA2A (Lee et al., 2002; Li et al., 2010), stimulating NMDA receptor-mediated long-term potentiation (Nai et al., 2010).

FGF receptor 1, which is a tyrosine kinases receptor, has been reported to undergo direct RRI with A2A receptor (Flajolet et al., 2008; Borroto-Escuela et al., 2013a). In particular, as demonstrated in PC12 cells (showing a series of similar properties with type I cells), contemporary activation of the two receptors, but not the single ones, induces activation of the MAPK/ERK pathway, differentiation and neurite extension (Flajolet et al., 2008). The majority of human CB's type I cells have shown a weak to moderate cytoplasmic immunostaining for FGF<sup>1</sup> receptor (Douwes Dekker et al., 2007). In particular, FGF<sup>1</sup> receptor immunoreactivity was shown in type I cells in postnatal rat CB cultures, in both normoxic and hypoxic conditions, as well as in bFGF presence or absence (Paciga and Nurse, 2001). bFGF increased both inward Na<sup>+</sup> and outward K<sup>+</sup> currents after 2 days of treatment on cultured type I cells from E18-19 rat pups (Zhong and Nurse, 1995). In cultures from rat E17- E19 CB, bFGF increases survival and BrdU incorporation; in postnatal P1-P3 cultures, bFGF still stimulates DNA synthesis but does not affect survival. In fetal rat glomus cells, bFGF stimulates neuronal differentiation, producing neurite outgrowth and inducing neurofilament immunoreactivity; these changes can no longer be detected in postnatal cultures (Nurse and Vollmer, 1997). Given the concomitant expression of A2A receptors in type I cells, an heterocomplex could be hypothesized. Moreover, the above developmental changes in the FGF action could partially derive by changes in the monomer–heteromer equilibrium. It is known, for instance, that the expression of A2A receptors in the rat type I cells decreases by PN14 (Gauda et al., 2000).

It is noteworthy that A2A-D2-FGF receptor mosaic has also been highlighted in other cell types (Fuxe et al., 2010) and could represent a further way of integration between neuromodulation and plasticity mechanisms.

### CONCLUSION AND FUTURE PERSPECTIVES

The CB is characterized by the production and release of many different neurotransmitters and neuromodulators which act on various receptor types identifiable on type I and II cells and nerve fibers. Although the main receptors involved in conveying excitatory stimuli from type I cells to afferent nerve fibers are ionotropic (nicotinic, P2X2/P2X3), a wide series of GPCRs is also expressed in the various structures involved in chemoreception, exerting modulation of neurotransmission.

In other transfected and native cell types, experimental evidence has been provided about the existence of RRIs between GPCRs, with production of homodimers, heterodimers or high-order complexes (receptor mosaics) which may modify ligand binding, signaling and internalization of the protomers. In the present paper, we have reviewed most literature data reporting the contemporary expression in type I cells, type II cells or afferents of GPCRs which are able to physically interact with each other to form homo/hetero-complexes. This is a prerequisite for postulating dimerization and/or oligomerization in the CB. RRIs are particularly intriguing to be hypothesized in the CB, where in vitro/in vivo pharmacological/functional data are available about cross-talks between different neurotransmitters/neuromodulators or 'paradoxical' response changes with different concentrations/doses. At least in some cases, such findings could be interpreted through the existence of direct RRIs. Nevertheless, few authors have considered the possibility of di/oligo-merization in the CB (Conde et al., 2008, 2009b) and to the best of our knowledge there are not studies addressing this aspect through up-to-date methodology (see corresponding paragraph). The literature data reviewed in the present paper support the possibility of RRI in the CB and stress the potential implications of di/oligomerization for chemoreception, neuromodulation and plasticity. It is known, in fact, that the expression of certain receptors varies in response to development or environmental stimuli (hypoxia, hyperoxia, etc.); changes in the dynamic monomer/oligomer equilibrium may be the consequence, with correlated functional implications.

In the present paper, we have collected indirect evidence of our 'working hypothesis,' identifying the most significant homo/hetero-complexes which would be worthwhile to be studied in the CB. The direct identification of homo/heterocomplexes in type I, type II and afferent terminations and the characterization of their functional role and relevance in the chemoreception could represent an exciting field of investigation. As previously stated, demonstration of RRIs in the CB should include (1) assessment of colocalization in "close proximity" of two or more receptors and (2) detectable biochemical/functional change in one receptor induced by the binding of a ligand to another receptor. We have here revised literature data

### REFERENCES


highlighting colocalization of different receptors in the various elements of the CB (i.e., type I and II cells, nerve terminals). In addition, colocalization of other receptors could be preliminarily investigated through double immunofluorescence or in situ hybridization, both in CB tissue samples and cell cultures.

However, colocalization is just a necessary condition to have direct RRIs and the demonstration of close proximity (i.e., a distance lower than 10 nm) between receptor molecules must be provided. To date, many methods of both biochemical and biophysical nature are available to accomplish this task and most of them could be applied to CB cell cultures. In particular, BRET, FRET or atomic force microscopy could be considered the most useful approaches. Interestingly, in vitro models allow to detect changes in cultured CB cells as a response to external pharmacological or environmental (hypoxia, hyperoxia, etc.) stimuli; we have hypothesized that RRIs may increase or decrease in response to external actions. Dynamic modifications of RRIs could also be analyzed in living cells through experimental techniques such as total internal reflection fluorescence microscopy, fluorescence correlation spectroscopy and PALM. Apart from cell cultures, heterocomplexes could be identified in native formalin-fixed CBs through in situ PLA, an approach that would even permit the identification of eventual changes in RRIs with respect to experimental conditions of in vivo models.

Once demonstrated physical RRIs, functional approaches in cell cultures would have to verify how the agonist/antagonist to one receptor may modify the response of the other receptor(s) toward the corresponding agonists/antagonists.

In conclusion, many in vitro and in vivo models are available, for research in CB structure/function, which represent adequate material for the application of consistent methods of analysis of potential RRIs.

### AUTHOR CONTRIBUTIONS

All the authors contributed to the revision of the literature and discussion of the hypothesis proposed. All the authors read and approved the final version of the manuscript.


interaction properties. Mol. Pharmacol. 66, 312–321. doi: 10.1124/mol.104. 000398.reasonable


development and aging. Adv. Exp. Med. Biol. 758, 265–271. doi: 10.1007/978- 94-007-4584-1\_36



in the rat carotid body. J. Appl. Physiol. 106, 259–267. doi: 10.1152/japplphysiol. 90626.2008


allosteric interactions within G protein-coupled receptor heteromers. Chem. Biol. 21, 1546–1566. doi: 10.1016/j.chembiol.2014.10.004


in the rat carotid body. Mol. Med. Rep. 3, 37–41. doi: 10.3892/mmr\_00 000215



neurons of the rat carotid body in co-culture. J. Physiol. doi: 10.1113/JP274743 [Epub ahead of print].


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Porzionato, Stocco, Guidolin, Agnati, Macchi and De Caro. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Postural Control in Lowlanders With COPD Traveling to 3100 m: Data From a Randomized Trial Evaluating the Effect of Preventive Dexamethasone Treatment

Lara Muralt1,2, Michael Furian1,2, Mona Lichtblau1,2, Sayaka S. Aeschbacher1,2 , Ross A. Clark<sup>3</sup> , Bermet Estebesova2,4, Ulan Sheraliev2,4, Nuriddin Marazhapov2,4 , Batyr Osmonov2,4, Maya Bisang1,2, Stefanie Ulrich1,2, Tsogyal D. Latshang1,2 , Silvia Ulrich1,2, Talant M. Sooronbaev2,4 and Konrad E. Bloch1,2,4 \*

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Marli Maria Knorst, Universidade Federal do Rio Grande do Sul, Brazil Wolfgang Schobersberger, Institut für Sport-, Alpinmedizin und Gesundheitstourismus, Austria

> \*Correspondence: Konrad E. Bloch konrad.bloch@usz.ch

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 27 February 2018 Accepted: 29 May 2018 Published: 22 June 2018

### Citation:

Muralt L, Furian M, Lichtblau M, Aeschbacher SS, Clark RA, Estebesova B, Sheraliev U, Marazhapov N, Osmonov B, Bisang M, Ulrich S, Latshang TD, Ulrich S, Sooronbaev TM and Bloch KE (2018) Postural Control in Lowlanders With COPD Traveling to 3100 m: Data From a Randomized Trial Evaluating the Effect of Preventive Dexamethasone Treatment. Front. Physiol. 9:752. doi: 10.3389/fphys.2018.00752 <sup>1</sup> Department of Respiratory Medicine, University Hospital Zurich, Zurich, Switzerland, <sup>2</sup> Kyrgyz-Swiss High Altitude Clinic and Medical Research Center, Tuja-Ashu, Kyrgyzstan, <sup>3</sup> School of Health and Sports Science, University of the Sunshine Coast, Sunshine Coast, QLD, Australia, <sup>4</sup> Department of Respiratory Medicine, National Center for Cardiology and Internal Medicine, Bishkek, Kyrgyzstan

Objective: To evaluate the effects of acute exposure to high altitude and preventive dexamethasone treatment on postural control in patients with chronic obstructive pulmonary disease (COPD).

Methods: In this randomized, double-blind parallel-group trial, 104 lowlanders with COPD GOLD 1-2 age 20–75 years, living near Bishkek (760 m), were randomized to receive either dexamethasone (2 × 4 mg/day p.o.) or placebo on the day before ascent and during a 2-day sojourn at Tuja-Ashu high altitude clinic (3100 m), Kyrgyzstan. Postural control was assessed with a Wii Balance BoardTM at 760 m and 1 day after arrival at 3100 m. Patients were instructed to stand immobile on both legs with eyes open during five tests of 30 s each, while the center of pressure path length (PL) was measured.

Results: With ascent from 760 to 3100 m the PL increased in the placebo group from median (quartiles) 29.2 (25.8; 38.2) to 31.5 (27.3; 39.3) cm (P < 0.05); in the dexamethasone group the corresponding increase from 28.8 (22.8; 34.5) to 29.9 (25.2; 37.0) cm was not significant (P = 0.10). The mean difference (95% CI) between dexamethasone and placebo groups in altitude-induced changes (treatment effect) was −0.3 (−3.2 to 2.5) cm, (P = 0.41). Multivariable regression analysis confirmed a significant increase in PL with higher altitude (coefficient 1.6, 95% CI 0.2 to 3.1, P = 0.031) but no effect of dexamethasone was shown (coefficient −0.2, 95% CI −0.4 to 3.6, P = 0.925), even when controlled for several potential confounders. PL changes were related more to antero-posterior than lateral sway. Twenty-two of 104 patients had an altitude-related increase in the antero-posterior sway velocity of >25%, what has been associated with an increased risk of falls in previous studies.

Conclusion: Lowlanders with COPD travelling from 760 to 3100 m revealed postural instability 24 h after arriving at high altitude, and this was not prevented by dexamethasone.

Trial Registration: clinicaltrials.gov Identifier: NCT02450968.

Keywords: chronic obstructive pulmonary disease, altitude, hypoxia, postural control, dexamethasone, acute mountain sickness

### INTRODUCTION

fphys-09-00752 June 22, 2018 Time: 16:47 # 2

Today many settlements worldwide are located at high altitudes (above 2500 m), with regular working places even above 3000 m. Moreover, mountain tourism and air travel are increasingly popular. Therefore, a large number of people, among them also patients with respiratory conditions, are exposed to hypobaric hypoxia. There are concerns that these patients may suffer from altitude-related adverse health effects including impaired postural control (PC), which may lead to falls or impaired performance during various tasks with consecutive accidents. In healthy individuals we have previously observed an impairment in PC already at altitudes of 1630 and 2590 m (Stadelmann et al., 2015). Other studies in healthy volunteers at a higher altitude (Mount Rosa, 4559 m) and in hypobaric chambers (Holness et al., 1982; Cymerman et al., 2001) have also demonstrated worsening of PC. The underlying mechanisms are poorly understood but it is presumed, that hypoxia affects different sensory functions (visual, somatosensory and vestibular) as well as the central nervous system that controls posture-regulating muscles especially in the lower limbs and trunk within a few minutes of exposure to altitudes of 2438 m (8000 ft) or higher (Wagner et al., 2011; Chiba et al., 2016). During sojourns of more than a few hours at high altitude PC may additionally be disturbed by acute mountain sickness (AMS), which causes headache, ataxia, weakness, dizziness, and decrements of alertness.

Chronic obstructive pulmonary disease (COPD) is associated with chronic inflammation and obstruction of the airways, parenchymal destruction of the lung with impaired gas exchange that promotes hypoxemia and increased pulmonary artery pressure (Vogelmeier et al., 2018). Given the high prevalence of COPD, it is expected that many affected patients are undertaking high altitude or air travel thereby suffering from impaired PC, although this has not been specifically studied. Previous observations suggest that elderly persons and patients with COPD suffer from an impaired PC already at sea level (Muir et al., 2010; Porto et al., 2015). Due to their pulmonary gas exchange impairment COPD patients may have more severe hypoxemia than healthy individuals at corresponding altitude. Thus, we reasoned that COPD patients may experience pronounced PC impairments during altitude or air travel with potentially grave consequences such as dangerous falls, and that measures to prevent or reduce this risk would be desirable.

The purpose of the current study was therefore twofold: (1) to test the hypothesis that lowlanders with COPD would experience impairments in PC during a stay at 3100 m and (2) that these impairments could be prevented by treatment with dexamethasone, a drug with potent glucocorticoid action. We selected dexamethasone for this study because glucocorticoids are used to treat COPD exacerbations and because it is effective in prevention and treatment of AMS in healthy mountaineers.

### MATERIALS AND METHODS

### Study Design and Setting

The current study was performed from June to August 2015 within the scope of a randomized, placebo controlled double blind parallel design trial evaluating effectiveness of dexamethasone in prevention of altitude-related adverse health effects (ARAHE) in lowlanders with COPD traveling to and staying for 2 days at the high altitude (3100 m) clinic of Tuja-Ashu, Kyrgyztan (clinicaltrials.gov Identifier: NCT02450968). The effects of altitude and of dexamethasone on AMS and various other clinical and physiologic outcomes have been reported recently (Furian et al., 2018).

### Participants

Men and women, aged 20–75 years, living in the Bishkek area (Kyrgyz Republic, mean altitude 760 m) diagnosed with COPD according to GOLD guidelines, grades 1–2, FEV1/FVC < 0.7 and FEV<sup>1</sup> > 50% predicted were invited to participate. Exclusion criteria were severe COPD with FEV<sup>1</sup> < 50% predicted, hypoxemia < 92% at 760 m measured by pulse oximetry, COPD exacerbation, reversible airflow obstruction, a history suggesting asthma or other respiratory disease, diabetes, uncontrolled cardiovascular disease (such as systemic arterial hypertension, coronary artery disease, previous stroke), history of obstructive sleep apnea, pneumothorax in the last 2 months, untreated or symptomatic peptic ulcer disease, or glaucoma and other conditions that might have interfered with protocol compliance including current heavy smoking (>20 cigarettes per day). Participants gave written informed consent. The study was approved by the Ethics Committee of the National Center of Cardiology and Internal Medicine, Bishkek, Kyrgyzstan(01- 8/405) and was endorsed by the Cantonal Ethics Committee Zurich, Switzerland.

### Interventions

Participants underwent baseline evaluation in Bishkek (760 m). One to three weeks later, they travelled to the Tuja-Ashu high altitude clinic (3100 m) by minibus within 3–5 h and stayed there for 2 days. On the day before ascent and while staying at 3100 m, participants took 4 mg capsules of oral dexamethasone twice daily or identical looking placebo capsules under the supervision of an investigator. During the study, participants continued

their regular medication and no other treatment was allowed. For safety reasons, participants with clinically relevant AMS defined by an AMSc score ≥0.7 (see below), severe hypoxemia (SpO<sup>2</sup> < 75% for > 30 min, or < 70% for > 15 min), or any other condition requiring an intervention according to the decision of an independent physician were treated with supplemental oxygen and other appropriate means.

### Assessments

A medical history and clinical examination were obtained. AMS was assessed by the environmental symptoms cerebral score (AMSc) comprising 11 questions on AMS symptoms each rated from 0 (not at all) to 5 (extreme) (Sampson et al., 1983). The weighted sum of responses ranges from 0 to 5. Scores ≥0.7 are considered to reflect clinically relevant AMS. Pulse oximetry during rest (SpO2, Konica-Minolta PULSOX-300i) and spirometry (EasyOne; NDD, Zurich, Switzerland) were performed.

PC was assessed by a Wii Balance Board (WBB, Redmond, WA, United States), 30 cm × 50 cm in size, as previously described (Clark et al., 2010; Stadelmann et al., 2015). Examinations took place at 760 m and on the second day of the stay at 3100 m, within 20–26 h after arrival. The subjects stood on the WBB on both legs with eyes open and feet positioned in a 30◦ angle, 20 cm apart. They were instructed to focus on a black dot on the wall, 1.5 m in front of them at eye level, to keep their hands beside the body and stand as still as possible during five tests each lasting 30 s and separated by a resting period of 2–3 min. During the tests, the WBB recorded displacements of the center of gravity of the subject by 4 integrated sensors, using the sampling method and filtering technique described in detail previously (Clark et al., 2017). A customized software (Labview 8.5 National Instruments, Austin, TX, United States) was used to calibrate the WBB with standard weights and to compute the center of gravity path length (COPL), and the mean and SD sway amplitude and velocity in the antero-posterior (AP) and medio-lateral (ML) directions (Clark et al., 2010; Holmes et al., 2013).

### Outcomes and Sample Size Estimation

The main outcome was the center of pressure path length (COPL) and additional outcomes were other variables from PC tests and from clinical and physiological examinations including clinically relevant AMS, severe hypoxemia, and other adverse effects that required an intervention as mentioned above. Since the minimal important difference for COPL and other indices of PC has not been established we based the sample size estimation for this study on previous studies in 51 healthy individuals who showed a significant change in indices of PC measured by the Wii balance board at 2590 m (Stadelmann et al., 2015). In addition, a sample size estimation was performed for the primary outcome of the main trial associated with the current study, the cumulative incidence of ARAHE during the stay at 3100 m. According to these calculations, a minimal number of 100 participants including drop-outs were required to detect a 50% reduction in ARAHE by dexamethasone (Furian et al., 2018).

### Randomization and Blinding

Participants were randomized 1:1 to dexamethasone or placebo treatment by a computer algorithm minimizing for differences in sex, age ≤ or > 50 years, FEV<sup>1</sup> < or ≥ 80% predicted (Pocock and Simon, 1975).

Study drugs were dispensed by an independent pharmacist in sets of capsules labeled with a concealed code. Participants and investigators were blinded to the assigned treatment until the completion of data analysis.

### Data Analysis

The primary data analysis was performed in the per protocol population of participants who had successful evaluations at both altitudes. In addition, an intention to treat analysis of the main outcome, the COPL, was performed including data from all randomized participants with missing data replaced by multiple imputations. A separate analysis restricted to data from patients >40 years of age was also performed excluding occasional younger participants fulfilling spirometric criteria of COPD but may not suffering from the classical form of the disease. To reduce effects of measurement variability, mean results from the five tests at each location are reported. Occasional individual missing data in one of the five tests were replaced by group medians of the corresponding altitude. Since most variables were non-normally distributed, data are summarized by medians and quartiles. Effect sizes were quantified by Cohen's d (i.e., d = the difference between two altitudes divided by the pooled standard); with values of d < 0.2 considered small, 0.5 to 0.8, medium, and > 0.8, strong (Cohen, 1977). The effects of altitude and of the drug were evaluated by computing mean differences and 95% confidence intervals (95% CI). Multivariable regression analysis was performed with the COPL or AP sway velocity as dependent variable and altitude or SpO2, dexamethasone, body height, age, FEV<sup>1</sup> (% predicted) and presence of AMS as independent variables. A p-value < 0.05 was considered statistically significant.

### RESULTS

A total of 118 patients with COPD fulfilled the inclusion criteria and participated in the study (**Figure 1**). In 2 patients in the dexamethasone group, balance tests at 3100 m could not be performed because of AMS; in 12 additional patients, balance tests were not available at both altitudes for various reasons. Datasets from 104 patients who successfully underwent balance tests at both locations could be included into the per protocol analysis (**Figure 1**). **Table 1** shows the demographic data of the study participants. 70 of the 104 participants (67%) suffered from mild airflow obstruction (GOLD grade 1) and 34 (33%) from moderate airflow obstruction (GOLD grade 2); 97 participants (93%) were >40 years old and 6 participants (7%) were <40 years old (31 to 38 years). 27 of the 118 patients (23%), 13 using dexamethasone, 14 using placebo (P = 0.749, Chi-square statistic) suffered from clinically relevant AMS, severe hypoxemia, or other altitude-related adverse health effects.

With ascent from 760 to 3100 m the COPL increased significantly in the placebo group from median (quartiles)

29.2 cm (25.8; 38.2) to 31.5 cm (27.3; 39.3) (P < 0.05) but in the dexamethasone group the corresponding increase from 28.8 cm (22.8; 34.5) to 29.9 cm (25.2; 37.0) was not significant (P = 0.10) (**Table 2**). An example of COPL recordings in a representative individual is shown in **Figure 2**. Mean altitude-induced changes in COPL are shown in **Figure 3**. The mean (95% CI) effect size (Cohen's d) of the altitude-induced COPL change in the placebo group was d = 0.2 (0.02 to 0.39) and in the dexamethasone group 0.17 (−0.02 to 0.37).

Whereas the AP sway velocity increased significantly with ascent to 3100 m in both groups, the medio-lateral sway velocity did not change in any group. Dexamethasone had no significant effect on any index of the PC. The mean difference between dexamethasone and placebo groups in altitude-induced changes in COPL (treatment effect) was −0.3 cm (95% CI −3.2 to 2.5) (P = 0.41) (**Figures 3**, **4**). The intention to treat analysis of the COPL revealed similar results as the per protocol analysis: the mean altitude-induced change in COPL in the placebo group was 2.1 cm (95% CI 0.2 to 3.9, P = 0.028) in the dexamethasone group the corresponding change was 1.8 cm (95% CI 0.01 to 3.7, P = 0.049) and the treatment effect was −0.2 cm (95% CI −2.8 to 2.4, P = 0.869).

Multivariable regression analysis confirmed a significant increase of 1.6 cm (0.2 to 3.1) [mean (95% CI)] in COPL and of 0.048 cm/s (0.009 to 0.087) in AP sway velocity when ascending form 760 m to 3100 m but no decrease with dexamethasone when controlled for several potential confounders (**Tables 3**, **4**). The mean (95% CI) effect size of the altitude-induced change adjusted for potential confounders listed in **Table 3** was d = 0.42 (0.36 to 0.48) in the placebo group and d = 0.49 (0.41 to 0.58) in the dexamethasone group. Taking SpO<sup>2</sup> instead of altitude as the predictor in multivariable regression confirmed that hypoxemia was associated with impaired PC: for each percentage point of reduction in SpO<sup>2</sup> the COPL was elongated by 0.3 cm (95% CI

−0.5 to −0.05, P = 0.017) (Supplementary Table S1). A similar negative correlation was found among SpO<sup>2</sup> and AP sway velocity (Supplementary Table S2).

Dexamethasone had no significant effect in the multivariable regression on the COPL, AP- or ML-velocity (**Tables 3**, **4** and Supplementary Tables S1, S2). However, age and body height were associated with increased COPL and AP-velocity.

The presence or absence of AMS during the whole stay at altitude had no significant influence on the COPL. However, the movements in AP-direction, which are more influenced by changes of altitude, were significantly increased in those participants suffering from AMS (Supplementary Tables S2, S3).

In a regression analysis including the ten consecutive performed tests (five tests at 760 m and the five at 3100 m) no learning effect was observed (Supplementary Table S4).



Values are shown in numbers or median (quartiles); FEV<sup>1</sup> = Forced expiratory volume in 1 second; FVC = Forced vital capacity; SpO<sup>2</sup> = arterial oxygen saturation. GOLD = Global Initiative for Chronic Obstructive Lung Disease.

Restricting the regression analysis to the 97 participants > 40 years old revealed similar results as those in **Tables 3**, **4** including all 104 participants, i.e., there was a significant effect of altitude on COPL and AP sway velocity but no significant effect of dexamethasone (Supplementary Tables S5, S6).

### DISCUSSION

We studied the effects of acute high altitude exposure (3100 m) and of preventive dexamethasone treatment on postural control (PC) in lowlanders with mild to moderate COPD (GOLD grade 1 – 2). Our randomized, placebo controlled, doubleblind trial demonstrates that measures of PC, including AP sway velocity and COPL increased at the higher altitude, consistent with impaired PC during acute exposure to hypobaric hypoxia. The results further revealed that the altitude-induced postural instability was not prevented by treatment with dexamethasone.

The current study is the first evaluating PC in lowlanders with COPD ascending rapidly to high altitude. It is generally assumed that the adverse effects of altitude exposure are related to hypoxia in the nervous system and possibly muscles (Chiba et al., 2016). Consistently, multiple regression analyses taking several independent variables into account (dexamethasone, age, sex, FEV<sup>1</sup> in % predicted, body height) confirmed an independent negative effect of SpO<sup>2</sup> on COPL. Previous studies on PC in various simulated and real altitudes have included small numbers of mainly young, healthy volunteers and were conducted in various settings, with different protocols and devices. In the largest study, a randomized cross-over trial, performed in 51 healthy young men staying for 2 days each at 1630 and 2590 m in the Alps, an altitude-related increase in COPL mainly due to AP sway was observed (Stadelmann et al., 2015). Consistently, in the current study, as well as in other previous studies, postural stability was mainly impaired in the AP direction, while ML sway was unchanged at altitude. Other studies using different protocols and tests evaluated effects of much shorter exposures to normobaric hypoxia (minutes to a few hours) at altitude ranging from 2438 m to 5906 m and generally confirmed impairment of PC.

In one study, sensomotoric, visual, and vestibular components of hypoxia-induced impairment of PC were evaluated in combination and separately by varying experimental conditions such as visual inputs (eyes open/closed, sway-referenced visual

TABLE 2 | Center of pressure path length at 760 m and 3100 m with placebo and dexamethasone treatment.


Values are shown in median (quartiles); AP, antero-posterior; ML, medio-lateral; <sup>∗</sup>p < 0.05 vs. baseline 760 m, paired Wilcoxon rank sum test.

pressure path length (COPL), 21.6 and 39.8 cm; antero-posterior sway velocity, 0.476 and 0.995 cm/s; medio-lateral sway velocity, 0.437 and 0.67 cm/s.

field), resting and moving platform. The results of these investigations confirmed that static PC (eyes open) and reaction time to unexpected movements of the platform were impaired even during very short exposures of less than 1 h to altitudes equivalent to 2438 m and 3048 m (Wagner et al., 2011).

While most previous studies were performed in young volunteers between 21 and 56 years of age, Drum et al. (2016) evaluated the combined mild normobaric hypoxia (altitude equivalent 2600 m) and exercise (a 40 min treadmill walk) in a group of 37 healthy seniors (mean ± SD age 62 ± 4 years).

FIGURE 4 | Effect of the altitude and of dexamethasone on sway velocity in the (A) antero-posterior (AP) direction and (B) medio-lateral (ML) direction. Altitude effect, difference between the center of pressure path length at 760 m and 3100 m; dexamethasone effect, difference of the altitude effect in the dexamethasone group minus the altitude effect in the placebo group. Values are shown as mean (95% CI); <sup>∗</sup>p < 0.05.

TABLE 3 | Effect of high altitude exposure on the center of pressure path length: multivariable regression.


Plc, Placebo; Dex, Dexamethasone; FEV1, % pred., Forced expiratory volume in 1 second in % of the predicted FEV1; AMS, development of acute mountain sickness during altitude exposure assessed by the environmental symptoms score ≥0.7.

TABLE 4 | Effect of high altitude exposure on the antero-posterior sway velocity: multivariable regression.


Plc, Placebo; Dex, Dexamethasone; FEV1, % pred., Forced expiratory volume in 1 second in % of the predicted FEV1; AMS, development of acute mountain sickness during altitude exposure assessed by the environmental symptoms score ≥0.7.

Balance tests did not reveal an effect of simulated altitude at rest but COPL were larger immediately after exercise at near sea level conditions and even more so after exercise at simulated altitude, suggesting a combined effect of muscular fatigue and hypoxia on PC in these elderly individuals. The current study confirms and extends these observations in a different setting in COPD patients by demonstrating an independent negative effect of altitude and older age on PC even at rest without prior exercise (**Tables 3**, **4**). Since COPD results from long-term cumulative noxious exposures it is most commonly observed in older individuals (>40 years). In the current study 7/104 (6%) participants were between 31 and 38 years old, an age that is not typically compatible with the classical form of COPD. We speculate that these individuals might have suffered from infections and exposure to smoke and indoor air pollution since early childhood. Such "disadvantage factors" may have promoted irreversible airflow obstruction at a relatively young age (Vogelmeier et al., 2018). Excluding data from participants <40 years of age from analysis confirmed a robust effect of altitude on PC in the 97 older COPD patients (>40 years old).

Since regression analysis of data in the current study indicates that altitude as well as age and body height adversely affects PC, it is important to take this into account when assessing PC. Moreover, the study by Drum et al. (2016) suggests that prior exercise may also affect PC, which may be particularly relevant for older individuals desiring to hike in the mountains.

Postural imbalance and ataxia are established diagnostic criteria for severe AMS and high altitude cerebral edema (HACE) (West, 1996; Hackett and Roach, 2004). The AMS questionnaire includes items like "dizziness/lightheadedness" or "my coordination is off " and the heel-to-toe walking as elements related to postural balance (Sampson et al., 1983; Sutton et al., 1992). However, studies on the relation among postural instability at altitude and AMS are controversial. Most studies could not show a relation between AMS and PC (Baumgartner et al., 2002; Cymerman et al., 2001). In the current investigation, multiple regression analysis did not reveal consistent results. Whereas AMS was not a significant predictor of the COPL, it was significantly correlated with the AP-velocity and AP-amplitude. We cannot exclude that the lack of a significant association between COPL and AMS was related to the low prevalence of AMS in our study. Further investigations are required to better

define interactions among AMS and PC measured objectively by a Wii balance board and other techniques in order to determine clinically relevant changes in objective indices of PC at high altitude and their relation to altitude-related illnesses including AMS.

To date, data on prevention or treatment of impairment in PC at altitude are scant. Baumgartner and coworkers administered supplemental oxygen to volunteers at 4559 m but failed to detect a beneficial effect on PC although symptoms of AMS were reduced by the intervention (Baumgartner and Bartsch, 2002). These authors therefore suggested that postural ataxia at altitude resulted from different hypoxiadependent mechanisms than AMS. In the current study we used dexamethasone to evaluate whether it might prevent impairments in PC, AMS, and other altitude-related adverse health effects such as insomnia and cognitive deficiencies in COPD patients. The reason for the lack of a protective effect of dexamethasone on PC remains unexplained. For safety reasons, the study protocol required that all participants suffering from clinically relevant AMS, severe hypoxemia or other adverse health effects were treated with supplemental oxygen and other appropriate means. These predefined precautions may have introduced a "survivor effect" preventing the observation of more severe PC impairments in the most susceptible persons and, thus, obscuring a significant effect of dexamethasone in the per protocol analysis. However, the intention to treat analysis with replacement of missing data by multiple imputation revealed similar altitude-induced changes in COPL and confirmed absence of a significant effect of dexamethasone thereby not supporting a relevant "survivor effect".

We cannot exclude that the applied dose of 2 mg × 4 mg dexamethasone/day was insufficient to prevent altitude-induced effects on PC. We selected this dose because it is recommended for AMS prevention in current guidelines and was effective in previous studies (Luks et al., 2014; Zheng et al., 2014). We did not want to use a higher dose of dexamethasone because of potential side effects including hypoglycemia. It is also possible that the only mild impairment of PC we found in the COPD patients reduced the sensitivity of our trial to detect a significant reduction in the altitude-induced impairment. The current results and those from previous studies (Stadelmann et al., 2015) suggest that the Wii balance board is a convenient technique to detect small to moderate altitude-related effect sizes of PC changes under field conditions. However, the accuracy of this simple and inexpensive tool was inferior compared to a laboratory-grade force platform in previous studies (Leach et al., 2014). Therefore, limitations in our measurement technique may have concealed minor effects of dexamethasone on PC even though we tried to minimize instrument-related inaccuracies by repeated calibrations, and by using the same board for all measurements in each individual. Moreover, we computed mean values from five tests in each individual in order to reduce variability in the outcomes.

A limitation of our study is the lack of a healthy, agematched control group of Kyrgyz individuals, which would have allowed to better assess the independent effect of COPD on PC. The clinical relevance of the severity of PC impairment observed in the current study is uncertain as we were unable to correlate these findings directly to a clinical outcome. Kwok and coworkers have shown that elderly individuals, 60–85 years of age, with a greater (i.e., at the 75th percentile) baseline AP sway velocity measured by a Wii balance board had an approximately twofold greater risk of falls in the following year compared to controls with less (i.e., at the 25th percentile) postural sway at baseline. For comparison, 22 of the COPD patients in this current study had an altitude-related increase in the AP sway velocity of >25% – a change that was associated with a twofold risk of falls in the cited study (Kwok et al., 2015). Comparison of the current to the cited study is hampered by differences in protocol and setting. Nevertheless, it is conceivable that any measurable impairment of PC might increase the risk of falls with potentially grave consequences for mountain travelers.

# CONCLUSION

This is the first study investigating PC in non-acclimatized patients with mild to moderate COPD travelling from their altitude of residence of around 760 m to a high altitude of 3100 m. The main findings were an impairment of PC at altitude as demonstrated by an increase in COPL and AP sway velocity. Dexamethasone in a dose of 2 mg × 4 mg per day did not prevent this impairment.

# AUTHOR CONTRIBUTIONS

LM contributed to study design, data collection, analysis, and drafting the manuscript. MF, ML, SA, BE, US, NM, BO, MB, StU, TL, SiU, and TS contributed to study design, data collection, and critically revising the manuscript. RC contributed to analysis of data. KB contributed to study design, data analysis, and revision of the manuscript.

### FUNDING

The study was supported by the Swiss National Science Foundation, the Lunge Zurich, and the Schweizerische Unfallversicherungsanstalt SUVA, Switzerland.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphys. 2018.00752/full#supplementary-material

### REFERENCES

fphys-09-00752 June 22, 2018 Time: 16:47 # 9


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Muralt, Furian, Lichtblau, Aeschbacher, Clark, Estebesova, Sheraliev, Marazhapov, Osmonov, Bisang, Ulrich, Latshang, Ulrich, Sooronbaev and Bloch. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Long-Term Chronic Intermittent Hypobaric Hypoxia Induces Glucose Transporter (GLUT4) Translocation Through AMP-Activated Protein Kinase (AMPK) in the Soleus Muscle in Lean Rats

Patricia Siques<sup>1</sup> \*, Julio Brito<sup>1</sup> , Karen Flores<sup>1</sup> , Stefany Ordenes<sup>1</sup> , Karem Arriaza<sup>1</sup> , Eduardo Pena<sup>1</sup> , Fabiola León-Velarde<sup>2</sup> , Ángel L. López de Pablo<sup>3</sup> , M. C. Gonzalez<sup>3</sup> and Silvia Arribas<sup>3</sup>

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Ginés Viscor, University of Barcelona, Spain Michael Furian, Universitätsspital Zürich, Switzerland

> \*Correspondence: Patricia Siques psiques@tie.cl

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 27 March 2018 Accepted: 07 June 2018 Published: 28 June 2018

### Citation:

Siques P, Brito J, Flores K, Ordenes S, Arriaza K, Pena E, León-Velarde F, López de Pablo ÁL, Gonzalez MC and Arribas S (2018) Long-Term Chronic Intermittent Hypobaric Hypoxia Induces Glucose Transporter (GLUT4) Translocation Through AMP-Activated Protein Kinase (AMPK) in the Soleus Muscle in Lean Rats. Front. Physiol. 9:799. doi: 10.3389/fphys.2018.00799 1 Institute of Health Studies, University Arturo Prat, Iquique, Chile, <sup>2</sup> Department of Biological and Physiological Sciences, Facultad de Ciencias y Filosofía/IIA, Cayetano Heredia University, Lima, Peru, <sup>3</sup> Department of Physiology, Faculty of Medicine, University Autonoma of Madrid, Madrid, Spain

Background: In chronic hypoxia (CH) and short-term chronic intermittent hypoxia (CIH) exposure, glycemia and insulin levels decrease and insulin sensitivity increases, which can be explained by changes in glucose transport at skeletal muscles involving GLUT1, GLUT4, Akt, and AMPK, as well as GLUT4 translocation to cell membranes. However, during long-term CIH, there is no information regarding whether these changes occur similarly or differently than in other types of hypoxia exposure. This study evaluated the levels of AMPK and Akt and the location of GLUT4 in the soleus muscles of lean rats exposed to long-term CIH, CH, and normoxia (NX) and compared the findings.

Methods: Thirty male adult rats were randomly assigned to three groups: a NX (760 Torr) group (n = 10), a CIH group (2 days hypoxia/2 days NX; n = 10) and a CH group (n = 10). Rats were exposed to hypoxia for 30 days in a hypobaric chamber set at 428 Torr (4,600 m). Feeding (10 g daily) and fasting times were accurately controlled. Measurements included food intake (every 4 days), weight, hematocrit, hemoglobin, glycemia, serum insulin (by ELISA), and insulin sensitivity at days 0 and 30. GLUT1, GLUT4, AMPK levels and Akt activation in rat soleus muscles were determined by western blot. GLUT4 translocation was measured with confocal microscopy at day 30.

Results: (1) Weight loss and increases in hematocrit and hemoglobin were found in both hypoxic groups (p < 0.05). (2) A moderate decrease in glycemia and plasma insulin was found. (3) Insulin sensitivity was greater in the CIH group (p < 0.05). (4) There were no changes in GLUT1, GLUT4 levels or in Akt activation. (5) The level of activated AMPK was increased only in the CIH group (p < 0.05). (6) Increased GLUT4 translocation to the plasma membrane of soleus muscle cells was observed in the CIH group (p < 0.05). Conclusion: In lean rats experiencing long-term CIH, glycemia and insulin levels decrease and insulin sensitivity increases. Interestingly, there is no increase of GLUT1 or GLUT4 levels or in Akt activation. Therefore, cellular regulation of glucose seems to primarily involve GLUT4 translocation to the cell membrane in response to hypoxiamediated AMPK activation.

Keywords: altitude, chronic intermittent hypoxia, insulin sensitivity, GLUT4, AMPK and Akt

### INTRODUCTION

fphys-09-00799 June 26, 2018 Time: 19:28 # 2

Glucose metabolism has been suggested to be improved under different types of hypoxic conditions. In fact, several studies have shown that under chronic hypoxia (CH) and short-term chronic intermittent hypoxia (CIH), glycemia and insulin levels decrease and insulin sensitivity increases in humans (Woolcott et al., 2015) and murine models (Chiu et al., 2004; Chen et al., 2010; Gamboa et al., 2011). However, the changes in long-term CIH have rarely been studied. The definitions of short- and long-term CIH vary. Short-term CIH includes minutes or hours of hypoxia during a brief period, while long-term CIH includes several days of hypoxia over a longer period of time (Siques et al., 2006).

Hypoxia exposure induces an adaptation process to replenish ATP in the cells that generate changes in glucose metabolism (Xia et al., 1997; Gamboa et al., 2011). There are several mechanisms involved in glucose regulation under hypoxic conditions, with the most important role played by glucose transporters (GLUTs) (Chiu et al., 2004; Castrejón et al., 2007; Manolescu et al., 2007). Cellular glucose uptake is mainly accomplished via facilitated transport mediated by a family of GLUTs: GLUT1 is responsible for basal glucose uptake, and GLUT4 is an insulinregulated GLUT; both transporters are present in skeletal muscles (Scheepers et al., 2004). Insulin and energetic stress increase GLUT4 membrane translocation in skeletal muscles by different signaling mechanisms. Each stimulus, alone and in combination, results in the translocation of intracellular GLUT4 from vesicles to the cell membrane (Thong et al., 2005; Mackenzie and Watt, 2016). One of the mechanisms involved is the insulin pathway, which activates the PI3K/Akt pathway, acting as a metabolic sensor that responds to insulin stimulation and incorporates GLUT4 into the plasma membrane (Kohn et al., 1996). However, under stressful conditions, such as exercise or hypoxia, glucose transport in skeletal muscles occurs via the insulin-independent AMP-activated protein kinase (AMPK) pathway (Hayashi et al., 2000).

Under hypoxic conditions, the AMPK pathway participates in skeletal muscle glucose uptake. When intracellular ATP levels decrease, AMPK switches off ATP-consuming pathways and switches on alternative pathways for ATP regeneration, altering the AMP/ATP ratio (Hayashi et al., 2000; Carling, 2004). It appears that under hypoxic conditions, AMPK can increase glucose transport into muscle cells but also increases the rate of fat utilization by muscles (Fisher, 2006). In contrast, in mice with CH exposure, the Akt and AMPK pathways are not activated, and GLUT4 levels are unchanged (Gamboa et al., 2011). Studies in murine models under short-term CIH have shown an increase in AMPK and GLUT4 levels in skeletal muscles, though these results are controversial (Chiu et al., 2004; Li et al., 2015; Wang et al., 2015).

To our knowledge, information about the AMPK pathway and GLUT4 translocation under long-term CIH in rats is scarce. This new exposure model involves days at hypobaric hypoxia followed by days at normoxia (sea level) over a long period of time. Therefore, many pathophysiological aspects of this condition remain poorly understood or controversial, such as glucose homeostasis.

It is hypothesized that in lean rats exposed to this model of hypoxia (long-term CIH), which combines normoxic and hypoxic periods, the pathways involved in GLUT4 translocation in the soleus muscle are activated resulting in changes in glucose and insulin levels. Thus, an experimental study was designed to evaluate differences in the levels of AMPK, Akt and GLUT4 and the cellular location of GLUT4 in soleus muscles of lean rats exposed to long-term CIH compared to those in CH and normoxia (NX) rats.

### MATERIALS AND METHODS

### Experimental Model and Study Groups

In this study, 30 male adult Wistar rats (3 months old; body weight 251.6 ± 1.9 g) were obtained from the animal facility of the Institute of Health Studies of Arturo Prat University, Iquique, Chile. The rats were placed in individual cages at a temperature of 22 ± 2 ◦C and a circadian rhythm of 12 h of light and 12 h of dark. Feeding consisted of 10 g/day of food that contained 22.0% crude protein, 5.0% crude fat, 5.0% crude fiber, 9.0% ash, and 12% moisture (5POO <sup>R</sup> , LabDiet <sup>R</sup> , Prolab RMH3000) and water ad libitum. Food intake was measured every 4 days by determining the amount of residual food. Movement inside the cage was not restricted, but no exercise was performed.

The rats were randomly distributed into three experimental groups, as follows: normobaric normoxia (NX), which served as a sea level control (n = 10); chronic intermittent hypobaric hypoxia (CIH), with 2 days of exposure to hypobaric hypoxia alternating with 2 days of exposure to NX (n = 10); and chronic hypobaric hypoxia (CH), which involved permanent exposure to hypoxia (n = 10). The exposure time of each group was 30 days, and the hypobaric hypoxia was simulated in a chamber at 428 Torr, which is equivalent to an altitude of 4,600 m above sea level. Chamber conditions were as follows: internal flow of 3.14 L/min of air and humidity between 21 and 30%. The time of ascension from sea level to 428 Torr was 60 min. NX rats were located in the same room at sea level (760 Torr)

and housed under the same chamber conditions as the groups exposed to hypoxia. At the end of the exposure period (day 30), the rats were euthanized with an overdose of ketamine (0.9 mg/kg of weight), organs were collected, and specific variables were measured. These experiments were performed at Arturo Prat University.

The animal protocol and experimental model were in accordance with Chilean law N◦ 20380 regarding animal experimentation and were approved by the Research Ethics Committee of Arturo Prat University, Iquique, Chile.

### Body Weight, Glucose and Insulin Measurements

Blood extraction (1 mL) for biochemical measurements was performed after 12 h of fasting via cardiac puncture under anesthesia (0.3 mg/kg body weight). Both biochemical and physiological parameters in all the study groups were performed at day 0 (under basal normoxic conditions) and after 30 days (immediately after descending from the chamber). The hematocrit (Hct) and hemoglobin (Hb) values were measured. Serum insulin was measured using a commercial kit (Rat Insulin ELISA Kit <sup>R</sup> , ALPCO, Salem, VT, United States), and glucose was measured using a glucometer (CarenSensN <sup>R</sup> ). The HOMA2 model was used to calculate the sensitivity (HOMA2%S) index with the HOMA2 calculator version 2.2 (Diabetes Trial Unit, University of Oxford), and body weight and residual food were measured every 4 days using an electronic scale (Acculab V-1200 <sup>R</sup> , IL, United States).

### Western Blot Analysis

In this study, 100 mg of skeletal muscle was obtained from each rat. Protein extraction was started by tissue homogenization (Stir-Pak <sup>R</sup> , Barrington, IL, United States) with 1 mL RIPA lysis buffer, which contains a cocktail of phosphatase and protease inhibitors (4 mM PMSF, 10 µM leupeptin, 1 mM EDTA, 1 mM EGTA, 20 mM NaF, 20 mM HEPES, and 1 mM DTT). Then, the homogenates were centrifuged (5804 R Eppendorf AG <sup>R</sup> , Hamburg, Germany) at 12,000 rpm for 20 min at 4◦C, and the supernatant was extracted. For quantification of the total protein extracted, the Bradford reaction was used (Bradford, 1976) with a BioPhotometer (Eppendorf AG <sup>R</sup> , Hamburg, Germany) at 590 nm, and samples were then stored at −80◦C. For western blotting, the samples were previously diluted with Laemmli 2X [0.125 M Tris-HCl, 4% SDS (p/v), 20% glycerol (v/v), 0.004% bromophenol blue, 10% β-mercaptoethanol (pH 6.8)]. The proteins were separated according to their molecular weight (MW) under an electric field via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (30% bis-acrylamide (v/v), 150 mM Tris (pH 6.8 and 8.8), 1.0% TEMED (w/v), H2O). Electrophoretic separation was initiated with the application of direct current to 150 V over 80 min with a power supply (PolySience <sup>R</sup> , EPS-300, Taipei, Taiwan, China), and the proteins were then transferred from the SDS-PAGE gel to a polyvinylidene fluoride (PVDF) membrane at 180 mA for 90 min with a semi-dry electroblotting system (OWLTM Separation systems, Panther

semi-dry Electroblotters, Thomas Scientific <sup>R</sup> , Barrington, IL, United States).

To avoid non-specific antibody binding, the membrane was blocked with bovine serum albumin (BSA) at a concentration range of 3–5% in TBS-T solution containing 10 mM HCl, 150 mM NaCl, 0.05% Tween-20 at pH 7.4. The blocking time was 1 h at room temperature.

Once the PVDF membrane was blocked, it was incubated with the corresponding primary antibody [GLUT4 (sc-1608), GLUT1 (sc-7903) AMPKα1/2 (sc-25792), p-AMPKα1/2 (sc-101630), Akt1/2/3 (sc-8312), p-Akt1/2/3 (sc-33437), and β-actin (sc-130657)] at a dilution of 1:500 (Santa Cruz Biotechnology <sup>R</sup> , CA, United States) and incubated overnight at 4◦C. Finally, the membrane was incubated with secondary antibodies (antigoat and anti-rabbit antibodies, Santa Cruz Biotechnology <sup>R</sup> , CA, United States) at a dilution 1:2000 in 3% BSA for 1 h at room temperature and then washed with TBS-T and imaged in a dark room with a chemiluminescence kit (Chemiluminescence West Pico <sup>R</sup> , Super Signal Substrate, Thermo Scientific <sup>R</sup> , Rockford, IL, United States). The density of the bands was measured with ImageJ and normalized according to β-actin expression.

### Confocal Microscopy

The presence of GLUT4 in the plasma membrane of soleus muscle cells was determined by immunofluorescence using confocal microscopy. After euthanasia, the soleus muscle was detached completely and immersed in 4% paraformaldehyde and embedded in paraffin. Muscles were cut transversally in relation to the direction of the muscle fibers. Slices (3 µm thick) were deparaffinized and then hydrated by incubating them in xylene three times for 5 min and in 100% ethanol and then 95% two times for 10 min. Subsequently, the sections were washed with distilled water two times for 5 min. Antigens were unmasked with citrate in a pascal pot at 95◦C for 20 min and then incubated in a permeabilization buffer (0.4% TRITON X-100 in PBS) for 30 min. Then, the sections were blocked with 5% BSA and incubated in 0.4% TRITON X-100 in PBS for 1.5 h. For the detection of GLUT4, the secondary antibody Alexa Fluor <sup>R</sup> 647 (A-21244) was used, and the nuclei were labeled with 4,6-diamidino-2-phenylindole (DAPI). The plasma membrane was labeled with WGA (L4895, SIGMA <sup>R</sup> , San Luis, MO, United States). The samples were visualized in a mounting medium (Citifluor, Aname, Spain) with a Leica TCS SP2 confocal system (Leica <sup>R</sup> Microsystems, Wetzlar, Germany) at University Autonoma of Madrid, Spain, using an emission wavelength of 405 nm for DAPI and an emission wavelength of 633 nm for Alexa Fluor <sup>R</sup> 647 and 488 nm for WGA. Serial images were 1 µm thick (12 µm in total) and were captured with a 63x objective at a zoom factor of 1–4 in randomly chosen areas under identical conditions of brightness, contrast, and laser power for all of the experimental groups. MetaMorph <sup>R</sup> image analysis software (Universal Imaging Co., United Kingdom) was used for quantification of the total number of cells and the intensity of GLUT4 fluorescence, which was used to calculate the amount of GLUT4 present in the plasma membrane by subtracting the total intensity of the cells from the intensity of the cytoplasm.

### Data Analysis

fphys-09-00799 June 26, 2018 Time: 19:28 # 4

All data recorded were included into a database and analyzed using the SPSS program (IBM SPSS <sup>R</sup> V.21.0 <sup>R</sup> , Armonk, NY, United States). The normality of the variables was established by the Kolmogorov-Smirnov test, and all variables had a normal distribution. The means, standard errors (SEs) and confidence intervals (CIs) were calculated for all variables. To determine differences in the measured variables over time, in each group, a paired-sample Student's T test was performed. To assess the magnitude of change in variables, between days 30 and 0, the means difference and 95% CIs were calculated for each variable. After obtaining these values the means differences between groups were also calculated using paired and independent sample Student's T test, respectively. Equal variances were assumed according to F value. To establish the inter-group differences, one-way analysis of variance (ANOVA) with the least significant difference (LSD) post hoc test was performed. The level of significance was established at the 95% confidence level, with p < 0.05 being considered significant.

### RESULTS

### General Variables

At day 30 under both hypoxic conditions (CIH and CH) rats showed an increase in Hct (p < 0.001) compared to the NX group, with a higher value in the CH group (p < 0.01) than in the CIH group. Likewise, Hb was increased in both hypoxic conditions, with the CH group showing a non-significant trend toward higher levels than in the CIH group. Body weight was lower in both hypoxia-exposed groups (CIH and CH), without a difference between these two groups. It is important to note that food intake was 100% in the CH and NX groups, whereas in the CIH group, 60% of the rats ate only 70% of the food inside the chamber during periods of hypoxia, although during periods of NX, food intake was normal in the CIH group (**Table 1**). Means difference for day 30–0 between groups: for hematocrit NX vs CIH: −18.26 (−24.33, −12.18) and body weight NX vs CIH: −24.75 (−36.03, −13.46) were found.

### Glycemia, Insulin and HOMA2%S

Both hypoxia-exposed groups (CIH and CH) exhibited a decrease in blood glucose levels at day 30 and CH showed levels lower than those in the CIH group (p < 0.05). There was a decrease in serum insulin levels under both hypoxic conditions, but unexpectedly, the level of insulin was lower in the CIH group than in the CH group (p < 0.05). The insulin sensitivity index (HOMA2%S) increased more in the CIH group than in the NX and CH groups (p < 0.05), and the difference was proportional to the insulin level (**Figure 1**).

### Protein Measurements: GLUTs, AMPK and Akt

The protein expression of GLUT1 and 4 was not different among groups (**Figures 2A–C**). AMPK activation, measured as the p-AMPK/total AMPK ratio, surprisingly, showed an


Variables measured in the normoxic group (NX; n = 10), chronic intermittent hypoxia group (CIH; n = 10), and chronic hypoxia group (CH; n = 10). Values are means (X) ± standard error (SEs) for hematocrit, hemoglobin, body weight and food intake; measured at day 0 and 30. <sup>∗</sup>p < 0.001: hypoxia-exposed group vs NX; †p < 0.01: CIH vs CH; #p < 0.001: day 0 vs day 30. Means difference and confidence intervals (95% CI) between days 30 and 0 for each variable are shown. Means difference and confidence intervals (95% CI) of these changes between groups CH vs. CIH (days 30 to 0) are also presented.

increase in the CIH group (p < 0.05), whereas the CH and NX groups showed no differences (**Figures 3A,C**). Conversely, Akt activation, measured as the p-Akt/total Akt ratio, showed no difference among the studied groups (**Figures 3B,C**).

### GLUT4 Translocation

It is worth noting that the results showed a remarkable increase in the translocation of GLUT4 from vesicles to the plasma membrane in rat soleus muscles only in the CIH group (p < 0.05)

(**Figure 4A**). These differences are also shown in representative images (**Figure 4B**).

### DISCUSSION

This research in lean rats exposed to long-term CIH showed the following with respect to the CH and NX groups: (1) different patterns in glucose regulation, (2) lower blood glucose and

plasma insulin levels and an increase in insulin sensitivity, (3) hypoxia-induced AMPK pathway activation, but not insulindependent Akt pathway activation and (4) an increase in GLUT4 translocation to plasma membranes in rat soleus muscle cells.

### General Findings

As expected in this model, Hb and Hct increased in the CIH group but to a lesser extent than in the CH group. These results are in agreement with those of other studies using this model, both in humans (Richalet et al., 2002) and rats (Siques et al., 2006; Brito et al., 2015). The body weight decreases observed in this study are also in agreement with previous reports in rats (Siques et al., 2006; Lüneburg et al., 2016). This latter effect could have ancillary influences on glucose metabolism, since body weight is highly correlated with insulin sensitivity (Evans et al., 1984). Nevertheless, high-altitude-induced CH leads to an inexorable loss of skeletal muscle mass as a consequence of an increase in protein degradation, which could explain the weight loss observed in this study (Chaudhary et al., 2012). Under CIH, this loss could also be attributed to a loss of appetite in rats (Bigard et al., 1996), as rats left some food uneaten during hypoxia exposure in the current study. Therefore, both mechanisms could contribute to the effect.

# Blood Glucose and Serum Insulin Levels

Previous altitude-based studies have shown that glucose homeostasis is influenced by acute or CH exposure in humans (Brooks et al., 1991; Kelly et al., 2010) and by CH exposure in mice (Gamboa et al., 2011). The current study shows that under long-term CIH, rats show improved glucose uptake, lower fasting glucose and insulin levels and increased insulin sensitivity, which also would occur under other types of hypoxia. Thus, the same phenomena have been shown in rats under short-term CIH, normobaric hypoxia and hypoxic regimen lasting hours (Chiu et al., 2004; Chen et al., 2011; Faramoushi et al., 2016), excluding obstructive sleep apnea syndrome studied in humans (Kim et al., 2013) and mice (Thomas et al., 2017). Although, controversially, some reports show no changes in glucose metabolism in rats exposed to CIH (Wang et al., 2015; Li et al., 2016). Likewise, exposure to cyclic hypobaric hypoxia in humans for 10 weeks demonstrated a decrease in glucose but no influence on insulin (Marquez et al., 2013). When drawing conclusions and making comparisons from some of the differences found in the literature regarding these effects, it must be considered that the regimen and the degree of hypoxia will lead to variation in the results (Debevec and Millet, 1985); however, most of these studies tend to agree in showing that under hypoxia, glucose regulation is improved in both rats and humans (Sawhney et al., 1986; Larsen et al., 1997; Woolcott et al., 2015; Sacramento et al., 2016). Moreover, this glucose improvement under different regimens of hypoxia would take some time and reach a plateau in the long term, resulting in rather normal value as observed in human studies (Woolcott et al., 2015; Zebrowska et al., 2018 ˙ ) and in rats (Chen et al., 2016). Therefore, the improved glucose regulation in rats under long-term CIH in this study supports the contribution of hypoxia to these phenomena and, to our knowledge, has not been previously reported.

# GLUT, AMPK and Akt in Soleus Muscle

Skeletal muscle is the most important regulator of glucose homeostasis and uptake (Mas et al., 2006; Gamboa et al., 2011). Glucose uptake in skeletal muscle is normally regulated by insulin-related pathways, where Akt can play a role as an insulinstimulated signal leading to GLUT4 translocation to the cell membrane under physiological conditions (Huang and Czech, 2007). However, under exercise- and hypoxia-induced stress, the AMP/ATP ratio is increased, activating an alternative pathway: insulin-independent AMPK signaling (Hayashi et al., 2000).

It is well known that skeletal muscle has two main GLUT transporters: the constitutive GLUT1 and the insulin-dependent GLUT4, with the latter being the most relevant. In acute exposure to hypoxia GLUT4 levels increase in rats (Dill et al., 2001), and obese rats (Zucker) under short-term CIH and in response to altitude training showed increased GLUT4 levels (Chen et al., 2011), whereas under CH, no increase in GLUT4 occurs in rats and mice (Chiu et al., 2004; Gamboa et al., 2011). According to our results from long-term CIH exposure, both GLUT1 and GLUT4 levels showed no changes, similarly to what has been reported under CH. However, there was an increase in intracellular translocation of GLUT4 to the membrane. This trafficking could be considered as a compensatory mechanism to increase glucose uptake instead the increasing protein levels (Chen et al., 2008; Jørgensen et al., 2009; Gamboa et al., 2011). Interestingly, this translocation has not been previously reported in a long-term CIH model.

Additionally, Gamboa et al. (2011) found that in normobaric CH (during fasting), no activation of the Akt pathway occurs, which is consistent with the results of this study. Likewise,

controversial findings indicating that chronic stress induced by reactive oxygen species (ROS) can also decrease the stimulatory effect of insulin on the Akt pathway and that glucose uptake could be mediated by the intrinsic activity of GLUT4 under hypoxic conditions have been reported (Ding et al., 2016). A role for ROS in the effect of insulin under hypoxia is a promising theory because high levels of ROS have been consistently found in humans with both acute and chronic exposure to hypobaric hypoxia (Jefferson et al., 2004) and in rats exposed to long-term CIH (Siques et al., 2014; Lüneburg et al., 2016).

The increase in GLUT4 translocation seen in this study could be explained on the basis of other pathways involved in the regulation of plasma glucose levels under energetic stress, including the AMPK pathway, as previously reported (Bradley et al., 2015). Activated AMPK acts downstream by phosphorylating and activating AS160 (Akt substrate 160), which is the main regulatory protein of the trafficking of intracellular GLUT4 (Kramer et al., 2006). It has also been shown that activated AMPK can increase the sensitivity to insulin in rats (Fisher et al., 2002). Our results support a role of AMPK but not of Akt, as the activation of AMPK was found to be increased under long-term CIH. Moreover, recent studies suggest that higher levels of insulin downregulate AMPK activity via Ser485/491 phosphorylation of the AMPK-α subunit. In this case, a lower blood insulin concentration might induce AMPK signal activation in rats (Kido et al., 2017), which would be another way of increasing AMPK activation; however, this idea needs further experimental support.

It has been observed that AMPK is activated in mice exposed to acute hypoxia (Viganò et al., 2011). This could explain the greater level of activated AMPK under CIH than under CH because, as described previously, this model involves intermittent and acute episodes of hypoxia, which results in a turn-on–turnoff regime for biological responses (Powell and Garcia, 2000). In this context, under CIH and during hypoxic training in humans, HIF-1α has been observed to be upregulated (Hoppeler and Vogt, 2001), and HIF-1α accumulation has been reported in muscle cell culture (Kubis et al., 2005). Moreover, HIF-1α could have a critical role in maintaining the GLUT4 transporter translocation in skeletal muscle cells (Sakagami et al., 2014). Since AMPK activation is consistent with the increased translocation of GLUT4 observed in this study, it could be surmised that this effect in the soleus muscle is mediated via upregulation of p-AMPK and that p-Akt does not play a role under the conditions studied here. Thus, the increases in translocation of GLUT4 to the cell membrane and AMPK activation in long-term CIH are novel findings.

Interestingly, age could play a role in GLUT regulation. Xia et al. (1997) showed that adult rats under CH exposure show slight increases in GLUT protein expression, whereas immature rats show great increases because immature tissues are more sensitive to oxygen deprivation. This latter report is almost coincident with our results regarding scarce or no GLUT4 protein increase in adult rats, although the current study is in CIH. Thus, this current study might give support to the hypothesis that in adult rats, GLUT regulation would occur not at the protein level but by GLUT translocation, resulting in increased glucose utilization.

This study in lean rats suggests that long-term CIH might have a beneficial effect in improving insulin sensitivity and glucose tolerance as has been suggested for rats exposed to shortterm CIH lasting hours (Tian et al., 2016). However, several considerations must be taken into account: the hypoxia regimen and exposure duration (Xia et al., 1997); the existence of several confounding factors such as vitamin D, pollution, ozone, and diet (Woolcott et al., 2015); the differences in response to hypoxia among rat's strain where Wistar is more intolerant to altitude (Ou and Smith, 1983; Hayward et al., 1999); and whether results from animal models can be fully extrapolated to clinical settings. Additionally, it is important to contrast the present findings with those from another model of CIH, i.e., obstructive sleep apnea syndrome, where the opposite metabolic patterns occur (Kim et al., 2013; Thomas et al., 2017).

This study has some limitations, such as the use of a very specific experimental animal model (lab rats with long-term CIH) that is more sensitive to hypoxia than human beings and with strict diet and environmental control, which increases the difficulty of comparing different CIH regimens and could prevent a direct translation of these findings into clinical or occupational health. However, this rat species was chosen, due to its known hypoxic intolerance, to assess the maximal effects of hypoxia and to perform preliminary molecular studies that would face ethical and logistic difficulties in humans. Another limitation is the difficulty of comparison given the wide variety of regimes, species, models, and the scarce reports on long-term CIH. However, this latter issue makes our results novel. Therefore, this study may contribute to the understanding of glucose metabolism in long-term CIH, which is poorly understood, and may provide directions for future research in animals and humans.

# CONCLUSIONS

Lean rats exposed to long-term CIH show a decrease in glycemia and insulin, along with an increase in insulin sensitivity compared to normoxic exposure. Interestingly, there is no increase in the levels of glucose transporter proteins GLUT1 or GLUT4 nor in the level of activated Akt. Therefore, glucose cell regulation and the relative hypoglycemia observed seem to be primarily a result of increased GLUT4 translocation to the cell membrane elicited by hypoxia-mediated AMPK activation.

# AUTHOR CONTRIBUTIONS

PS, JB, KF, and SO conceived and designed the study, performed the experiments, analyzed and interpreted the data, drafted the manuscript, critically revised important intellectual content in the manuscript, and provided overall supervision. FL-V assisted in critical decisions and revision. KA, EP, FL-V, ÁLdP, MG, and SA contributed to the interpretation of the results and critical revisions of the manuscript. All authors approved the final manuscripts and agreed to be accountable for all aspects of the work.

# FUNDING

fphys-09-00799 June 26, 2018 Time: 19:28 # 9

This study was supported by grants from projects GORE FIC-Tarapaca BIP30434827-0 and FIC-Tarapaca BIP30477541-0.

### REFERENCES


### ACKNOWLEDGMENTS

We would like to thanks to Pilar Rodriguez and Gabriela Lamas for their technical assistance in the laboratory.

insulin binding. Effects of obesity and body fat topography. J. Clin. Invest. 74, 1515–1525. doi: 10.1172/JCI111565



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Siques, Brito, Flores, Ordenes, Arriaza, Pena, León-Velarde, López de Pablo, Gonzalez and Arribas. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Melatonin Relations With Respiratory Quotient Weaken on Acute Exposure to High Altitude

Marcelo Tapia<sup>1</sup> , Cristian Wulff-Zottele<sup>2</sup> , Nicole De Gregorio<sup>3</sup> , Morin Lang<sup>2</sup> , Héctor Varela<sup>4</sup> , María Josefa Serón-Ferré<sup>3</sup> , Ennio A. Vivaldi<sup>3</sup> , Oscar F. Araneda<sup>5</sup> , Juan Silva-Urra<sup>2</sup> , Hanns-Christian Gunga<sup>6</sup> and Claus Behn3,7 \*

<sup>1</sup> Owl Capacitaciones y Asesorías SpA, Antofagasta, Chile, <sup>2</sup> Facultad de Ciencias de la Salud, Universidad de Antofagasta, Antofagasta, Chile, <sup>3</sup> Facultad de Medicina, Instituto de Ciencias Biomédicas, Universidad de Chile, Santiago, Chile, <sup>4</sup> Facultad de Ciencias Básicas, Universidad de Antofagasta, Antofagasta, Chile, <sup>5</sup> Facultad de Medicina, Universidad de los Andes, Santiago, Chile, <sup>6</sup> Center for Space Medicine and Extreme Environments, Institute of Physiology, Charité – Universitätsmedizin Berlin, Berlin, Germany, <sup>7</sup> Facultad de Medicina, Universidad San Sebastián, Santiago, Chile

High altitude (HA) exposure may affect human health and performance by involving the body timing system. Daily variations of melatonin may disrupt by HA exposure, thereby possibly affecting its relations with a metabolic parameter like the respiratory quotient (RQ). Sea level (SL) volunteers (7 women and 7 men, 21.0 ± 2.04 y) were examined for daily changes in salivary melatonin concentration (SMC). Sampling was successively done at SL (Antofagasta, Chile) and, on acute HA exposure, at nearby Caspana (3,270 m asl). Saliva was collected in special vials (Salimetrics Oral Swab, United Kingdom) at sunny noon (SMCD) and in the absence of blue light at midnight (SMCN). The samples were obtained after rinsing the mouth with tap water and were analyzed for SMC by immunoassay (ELISA kit; IBL International, Germany). RQ measurements (n = 12) were realized with a portable breath to breath metabolic system (OxiconTM Mobile, Germany), between 8:00 PM and 10:00 PM, once at either location. At SL, SMCD, and SMC<sup>N</sup> values (mean ± SD) were, respectively, 2.14 ± 1.30 and 11.6 ± 13.9 pg/ml (p < 0.05). Corresponding values at HA were 8.83 ± 12.6 and 13.7 ± 16.7 pg/ml (n.s.). RQ was 0.78 ± 0.07 and 0.89 ± 0.08, respectively, at SL and HA (p < 0.05). Differences between SMC<sup>N</sup> and SMC<sup>D</sup> (SMCN–SMCD) strongly correlate with the corresponding RQ values at SL (r = −0.74) and less tight at HA (r = −0.37). Similarly, mean daily SMC values (SMCx¯ ) tightly correlate with RQ at SL (r = −0.79) and weaker at HA (r = −0.31). SMCN–SMCD, as well as, SMCx¯ values at SL, on the other hand, respectively, correlate with the corresponding values at HA (r = 0.71 and r = 0.85). Acute exposure to HA appears to loosen relations of SMC with RQ. A personal profile in daily SMC variation, on the other hand, tends to be conserved at HA.

Keywords: melatonin, circadian rhythm, high altitude, respiratory quotient, body timekeeping

# INTRODUCTION

Contemporary working conditions, tend to challenge the human body internal timing system. Jet-lag (Coste et al., 2004), and extreme environments (Arendt, 2012; Najjar et al., 2014), affect circadian rhythms. Circadian misalignment sets the basis for metabolic disorders and cell cycle alterations that ultimately implicate risks at work and disease (Archer et al., 2014;

### Edited by:

Jean-Paul R. Richalet, Université Paris 13, France

### Reviewed by:

Paul Kenneth Witting, University of Sydney, Australia Andrew T. Lovering, University of Oregon, United States

> \*Correspondence: Claus Behn clausbehnthiele@gmail.com

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 28 December 2017 Accepted: 07 June 2018 Published: 29 June 2018

### Citation:

Tapia M, Wulff-Zottele C, De Gregorio N, Lang M, Varela H, Serón-Ferré MJ, Vivaldi EA, Araneda OF, Silva-Urra J, Gunga H-C and Behn C (2018) Melatonin Relations With Respiratory Quotient Weaken on Acute Exposure to High Altitude. Front. Physiol. 9:798. doi: 10.3389/fphys.2018.00798

**223**

Smolensky et al., 2016; Swanson et al., 2016). Circadian deregulation on high altitude (HA) exposure (Mortola, 2007, 2017) added to desynchronization by shift-work (Andlauer et al., 1979; Reinberg and Ashkenazi, 2008; Mirick et al., 2013) may well represent a factor involved in the lethal outcome of remote HA mining (Zaldívar Larraín, 2013).

Living beings synchronize with periodic environmental challenges. Various Zeitgebers, among them the daily light/dark cycle synchronize endogenous time keepers, the biological clocks (Reddy and O'Neill, 2010; Thut et al., 2012; Tsang et al., 2013). Rhythms result from a changing balance between activators and repressors in a negative feedback loop or between synthesis and degradation rates of oscillator components (Pulivarthy et al., 2007; see also Li et al., 2017). Intersecting with cellular biochemistry, multiple oscillators finally yield physiological, and behavioral rhythms (Top and Young, 2017). Countless oscillators, with widely differing oscillation periods, constitute the body timing system (Beale et al., 2016; Tran et al., 2016). Interacting among themselves (Schroeder and Lakatos, 2009; Zelano et al., 2016), the oscillators represent temporal reference frames for each other (Thut et al., 2012; Thurley et al., 2017). A complex information handling framework thus results (Rapp, 1987; Lloyd and Rossi, 1993).

Melatonin synchronizes cellular clocks with its own epiphyseal secretion, the latter being driven, via suprachiasmatic nuclei (Coomans et al., 2013), by the daily light/dark cycle determined by Earth rotation (Arendt, 1996; Chakir et al., 2015; Hardeland, 2015). Rhythm synchronization integrates body functions, both by local (Lin et al., 2017), as well as, by systemic means (Pfeffer et al., 2017). Melatonin (N-acetyl-5-methoxytryptamine), an ubiquitous, pleiotropic, and multitasking indoleamine (for recent reviews see Luchetti et al., 2010; Reiter et al., 2010; Hardeland et al., 2012) derives from tryptophan successively being transformed into serotonin and N-acetylserotonin. An N-acetyltransferase, involved in melatonin synthesis, is inhibited by light. Melatonin, thus, acts as a chemical transmitter of darkness (Tan et al., 2010; Hardeland et al., 2011). The non-image-forming vision system entraining body function rhythmicity via melatonin also implicates a subpopulation of retinal ganglion cells (ipRGCs ≈ 1% of the retinal ganglion cell population; Panda et al., 2003). The ipRGCs depolarize in response to photostimulation (Berson et al., 2002). Melanopsin, the photopigment of ipRGCs, absorbs light at aprox. 480 nm, the wavelength most effective in suppressing melatonin secretion (for a recent review see Lucas et al., 2014). Notably, melanopsin is also present in epithelial cells of the lens (Alkozi et al., 2017).

Melatonin involvement in overall circadian regulation relates to energy metabolism (Peschke et al., 2013; Cipolla-Neto et al., 2014) including termoregulation (Gubin et al., 2006; Kräuchi et al., 2006) and redox status (Maciel et al., 2010; Jiménez-Ortega et al., 2012; Tan et al., 2013; Cudney et al., 2014) acting, among others, as a natural antioxidant (Nehela and Killiny, 2018). Melatonin targets genes (Unfried et al., 2010; Hardeland et al., 2011; Torres-Farfán et al., 2011), the epigenome (Korkmaz et al., 2012; Haim and Zubidat, 2015), as well as, mitochondria (Acuña-Castroviejo et al., 2003; Maciel et al., 2010).

High altitude exposure may affect melatonin rhythm by lack of oxygen. Hypoxia, the lack of oxygen as related to aerobic energy requirements (Connett et al., 1990), delays the phase of melatonin rhythm (Coste et al., 2009). Untreated obstructive sleep apnoea syndrome, a clinical condition implicating intermittent hypoxia, leads to an early morning plateau of plasma melatonin concentration. This morning plateau of melatonin is reversed into a night time peak by increasing oxygen supply via CPAP device application in treated obstructive sleep apnoea patiernts (Hernández et al., 2007). Hypoxia applied for two hours in a hypobaric chamber (simulating 8,000 m a.s.l.) increases plasma melatonin concentration in rats (Kaur et al., 2002). This body timing system, thus, may be alterated by an environmental challenge such as a rapid ascent from sea level (SL) up to 3,000 m a.s.l., as usual in Chilean Andes. Respiratory quotient (RQ) elevation on HA exposure indicates an increase of glucose utilization under that condition. Insulin-regulated pathways depend on integrity of biological clocks (McGinnis et al., 2017). We, thus, examined effects of acute exposure at HA on the circadian rhythm of the chronotropic neurohormone melatonin and its relation with a metabolic parameter like RQ, the latter representing, a point of reference for energy metabolism at HA.

# MATERIALS AND METHODS

### Subjects

Fourteen healthy volunteers (**Table 1**), all of them students enrolled in Physical Education Pedagogy at University of Antofagasta, volunteered for the present study in the context of a wider HA research project (FONDECYT 1100161). Having previously been approved by the Ethics Committee of the Faculty of Medicine, University of Chile, the latter project was also endorsed by Bioethical Committee of Faculty of Health Sciences, University of Antofagasta, considering the principles and practices stated in the Declaration of Helsinki for studies of human beings. A written informed consent was obtained from each subject finally participating in the study.

### Study Design

The volunteers were examined for salivary melatonin concentration (SMC) at SL, the site of their usual residence. Cardio-respiratory parameters could be obtained in only 12 of them (**Table 2**). Corresponding measurements at HA were done

TABLE 1 | Body dimensions of the volunteers.

Physical parameters of the volunteers (mean ± SD)


Parameters measured at SL.



The asterisk denotes the difference between SL and HA values being significant (p < 0.05). HR, heart rate; VE, pulmonary ventilation; BR, breathing rate; VO2, oxygen flux; VCO2, carbon dioxide flux; HbO<sup>2</sup> sat, hemoglobin oxygen saturation; RQ, respiratory quotient.

in the context of a pedagogic field trip, on the day after arriving by bus, at Caspana (3,270 m a.s.l.), a small village located in the Andes, 300 km east from Antofagasta.

### Measurements

### Salivary Melatonin Concentration

At SL, as well as, at the HA site, the subjects were required to provide saliva samples for SMC determination, with sun light at midday (SMCD) and dim, ordinary bulb light, at midnight (SMCN). After rinsing the mouth with tap water, samples of saliva (1.5 ml aprox.) were collected into special vials (Salimetrics Oral Swab, United Kingdom), The saliva samples were handled using gloves, coded and stored in liquid nitrogen, to be later on analyzed for SMC with an ELISA kit (IBL International, Germany) in an independent commercial laboratory (Red Lab S.A., Santiago, Chile). SMCN–SMC<sup>D</sup> and SMCx¯ are, respectively, assumed to represent the amplitude of daily SMC change and the average of both day and night SMC value per subject.

### Respiratory Quotient

Cardio-respiratory parameters were determined under resting conditions, after sitting for 5 min. The measurements were done between 8:00 and 10:00 PM, both at SL and HA, once at either location. Evening meals consisted of bread and cheese at SL, as well as at HA. Along a 3 min equilibration period, respiratory CO<sup>2</sup> and O<sup>2</sup> fluxes could be measured in 12 of the 14 subjects with a portable metabolic system, including a breath-to-breath spirometer (OxiconTM Mobile, Germany). RQ was calculated as the ratio between mean CO<sup>2</sup> flux and mean O<sup>2</sup> flux.

### Statistics

Mean values are expressed ± SD. ANOVA for repeated measurements was applied for comparisons between SMC<sup>D</sup> and SMC<sup>N</sup> at SL and HA. Pearson's correlation coefficient and Student's t-test were, respectively, applied for analysis of correlations and for comparison between SL and HA. Calculations were done with the aid of SPSS 22 IBM software package. Statistic significance was established at the p < 0.05 level.

# RESULTS

Age and body mass index were rather similar in women and men volunteering in the present study (**Table 1**). Cardiorespiratory parameters of the volunteers significantly changed on HA exposure as compared to SL (**Table 2**). **Figure 1** shows mean values of SMC at day and night, both at SL (SLD, SLN) and HA (HAD, HAN). Mean SMC values in either those conditions were similar in women and men (data not shown). Daily variations of SMC observed at SL vanish at HA. SMCN–SMC<sup>D</sup> and SMCx¯ , respectively, depict, for the present work, the amplitude of circadian melatonin rhythm and the average value around which the oscillation occurs. These parameters strongly correlate one with the other at SL. At HA, on the contrary, this correlation weakens (**Figure 2**). RQ–SMC<sup>D</sup> relation appears to be strong at SL and weak at HA (**Figure 3A**). Similarly, the RQ–SMC<sup>N</sup> relation appears to be tighter at SL than at HA (**Figure 3B**). Both, SMCN–SMC<sup>D</sup> (**Figure 3C**), as well as, SMCx¯ (**Figure 3D**), also correlate with RQ more strongly at SL than at HA

SMCx¯ values at SL strongly correlate with those at HA (black circles, **Figure 4**). Similarly, SMCN–SMC<sup>D</sup> at SL also tightly correlate with the corresponding values at HA (white circles, **Figure 4**). As also shown in **Figure 4**, the former and the latter relation, respectively, locate mainly above and below the middle line (y = x).

### DISCUSSION

Mean SMC<sup>N</sup> and SMC<sup>D</sup> values differ at SL but not at HA (**Figure 1**). SMC<sup>N</sup> and SMCD, as well as, SMCN–SMC<sup>D</sup> and SMCx¯ , correlate with RQ strongly at SL and much less so at HA (**Figure 3**). Melatonin circadian rhythm, thus, may lose at HA its synchronizing grip on aspects related with energy metabolism. Individual SMCx¯ and SMCN–SMC<sup>D</sup> values at SL, on the other hand, strongly correlate with the corresponding ones at

regression between RQ and SMCx¯ at SL and HA. White circles and gray triangles, respectively, represent SL and HA values (n = 12).

HA (**Figure 4**). Although being distorted at HA (**Figures 1–3**), an individual profile of circadian melatonin rhythmicity, thus, seems to persist under the latter condition (**Figure 4**). Such an individual profile of circadian melatonin rhythmicity may in the future be explored for its potential to predict the capacity for adequately dealing with challenges of the body timing system.

Salivary melatonin has been validated as an adequate marker for phase typing of circadian regulation (Voultsios et al., 1997). Although representing only one third of plasma melatonin concentration (Benloucif et al., 2008), SMC adequately relates to the latter (Voultsios et al., 1997). Hyposalivation and low melatonin levels may limit the reliability of SMC, as measured by radioimmunoassay in the elderly (Gooneratne et al., 2003). Liquid chromatography combined with mass spectrometry, on the other hand, revealed SMC values to exceed free plasma melatonin concentration on average by 36%

(van Faassen et al., 2017). Like in oral mucosa (Chaiyarit et al., 2017), melatonin may be also locally produced in salivary glands (van Faassen et al., 2017). Whether related or not with plasma melatonin, SMC shows in the present work a clear rhythmicity, that may even represent changes occurring at tissue level. An ELISA kit used in the present work yielded SMC values very similar to those reported by others (Lushington et al., 2002; Verheggen et al., 2012).

Mean SMC<sup>N</sup> and SMC<sup>D</sup> values differ at SL but not at HA (**Figure 1**). SMCN–SMC<sup>D</sup> considered, in the present study, as the amplitude of daily melatonin variation, correlates with SMCx¯ (the average value around which the oscillation occurs) more strongly at SL than at HA (**Figure 2**). Distortions of SMC rhythm as shown to occur at HA (**Figures 1**, **2**) may implicate a deregulation of melatonin-dependent periodic processes. High amplitudes in circadian melatonin rhythmicity may prevent and/or delay the development of diabetes (Hardeland, 2017). The amplitude of daily melatonin oscillation, on the other hand, diminishes in the elderly (Gubin et al., 2006; Kim et al., 2014).

Disruption of body timekeeping, implicates deregulation of body functions (Cipolla-Neto et al., 2014; O'Neill and Feeney, 2014). Three weeks of circadian disruption induce a prediabetic condition in otherwise healthy subjects (Buxton et al., 2012). Energy metabolism unbound from circadian pacemakers associates to obesity, diabetes, cardiovascular disease, and cancer (Miller et al., 2010; Blask et al., 2014; Zubidat and Haim, 2017). SMCN–SMCD, as well as, absolute values of SMC<sup>N</sup> and SMC<sup>D</sup> loosening their relation with RQ at HA (**Figure 3**) could mean a decoupling of energy metabolism from circadian control, a possibility that certainly has further to be elucidated. It may be noticed, however, that even acute adequation of energy metabolism to HA exposure is yet far from reaching a consensus (Chicco et al., 2018). It may be provisionally assumed, however, that mistiming of melatonin circadian rhythmicity may represent a metabolic risk factor, particularly under conditions combining shift work with hypoxia as being usual in Chilean Andes.

Deregulation of circadian melatonin rhythmicity may result from changes in oxygen supply. Hypoxia also implicates an increase in sympathetic activity. Sympathetic afferent nerves of the pineal gland activate an N-acetyltransferase, the rate-limiting enzyme for melatonin synthesis. Beta-blockers, older age and a higher body mass, on the other hand, have been found to lower nocturnal urinary 6-sulfatoxymelatonin levels (Davis et al., 2001). Melatonin secretion may, moreover, additionally be altered at HA by hypocapnia prevailing in newcomers at HA. Neurons of suprachiasmatic nucleus are, in fact, particularly sensitive to pH (Chen et al., 2009).

Individual values of SMCN–SMC<sup>D</sup> and SMCx¯ observed at SL, respectively, correlate with the corresponding value at HA (**Figure 4**). Individual patterns in melatonin circadian rhythmicity as observed at SL, thus, appear largely to be conserved at HA. Individual circadian melatonin rhythmicity seems, indeed, to remain relatively stable (Fernández et al., 2017). With exception of sedation and/or artificial ventilation (Olofsson et al., 2004), neither activity, posture, sleep, nor menstrual phase appear to affect individual circadian rhythm of melatonin (Cain et al., 2010). From one subject to another one, nocturnal melatonin concentration can, on the other hand, differ considerably (Zeitzer et al., 1999). Some people seem to be able to rapidly modify their melatonin secretion pattern, as well as, to readily adapt to rotating shift schedules (Quera-Salva et al., 1997). Similarly, physiological adjustments to acute HA exposure vary, indeed, substantially from one subject to another. Individual characteristics of circadian melatonin rhythmicity, yet to be defined, may well relate with the capacity to adequately deal with challenges of the body timing system affecting energy metabolism in health and disease.

To summarize, a rapid ascent to an altitude of about 3,000 m a.s.l., as usual under working conditions in the Andes, tends to override the night-day difference of SMC and to weaken the relations between SMC with RQ, thus, potentially deregulating melatonin-dependent timing of body functions, affecting energy metabolism. Individual SL circadian profile of SMC tends, on the other hand, to be maintained at HA. The SL profile of melatonin circadian rhythm may be further on explored for its potential to predict individual tolerance to challenges of the body timing system at HA.

### AUTHOR CONTRIBUTIONS

CB, CW-Z, JS-U, ML, and MT were mostly implicated in the experimental design, logistics, and development of the experimental work at SL and HA. MT, CW-Z, NDG, ML, HV, MS-F, EV, OA, JS-U, H-CG, and CB substantially contributed to the conception of the work, data analysis and

manuscript revision, approved the final version, and agree to be accountable for the whole work. The M.Sc. thesis of MT at the Faculty of Health Sciences, University of Antofagasta, is mainly based on this work.

### FUNDING

This study was supported by FONDECYT Chile (Project N◦ 1100161) and Bundesministerium für Bildung und Forschung (Project CHLI2Anb, Domeyko-Center) Germany is gratefully acknowledged.

### REFERENCES


### ACKNOWLEDGMENTS

Thanks are due to Mr. Aldo Valdebenito for technical support, as well as, to University of Antofagasta students for their enthusiastic participation as volunteers in the present study. We also thank very much Professor Juan Ahumada, Director of Escuela E-20, for providing access and facilities in Caspana. We acknowledge moreover the critical reading of the manuscript by Dr. Stefan Mendt from Institute of Physiology, Center of Space Medicine and Extreme Environments Berlin, Charité – Universitätsmedizin Berlin, Germany, as well as, the very helpful advice provided by two reviewers from Frontiers.




**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Tapia, Wulff-Zottele, De Gregorio, Lang, Varela, Serón-Ferré, Vivaldi, Araneda, Silva-Urra, Gunga and Behn. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Serotonin and Adenosine G-protein Coupled Receptor Signaling for Ventilatory Acclimatization to Sustained Hypoxia

Esteban A. Moya\* and Frank L. Powell

Division of Physiology, Department of Medicine, University of California, San Diego, La Jolla, CA, United States

Different patterns of hypoxia evoke different forms of plasticity in the neural control of ventilation. For example, acute intermittent hypoxia produces long term facilitation (LTF) of ventilation, while chronic sustained hypoxia (CH) causes ventilatory acclimatization to hypoxia (VAH). In both LTF and VAH, ventilation in normoxia is greater than normal after the hypoxic stimulus is removed and the acute hypoxic ventilatory response can increase. However, the mechanisms of LTF and VAH are thought to be different based on previous results showing serotonin 5HT<sup>2</sup> receptors, which are G protein coupled receptors (GPCR) that activate G<sup>Q</sup> signaling, contribute to LTF but not VAH. Newer results show that a different GPCR, namely adenosine A2A receptors and the G<sup>S</sup> signaling pathway, cause LTF with more severe intermittent hypoxia, i.e., PaO<sup>2</sup> = 25–30 Torr for G<sup>S</sup> versus 35–45 Torr for LTF with the G<sup>Q</sup> signaling pathway. We hypothesized adenosine A2A receptors and G<sup>S</sup> signaling are involved in establishing VAH with longer term moderate CH and tested this in adult male rats by measuring ventilatory responses to O<sup>2</sup> and CO<sup>2</sup> with barometric pressure plethysmography after administering MSX-3 or ketanserin (A2A and 5HT<sup>2</sup> antagonists, respectively, both 1 mg/Kg i.p.) during CH for 7 days. Blocking G<sup>S</sup> or G<sup>Q</sup> signals throughout CH exposure, significantly decreased VAH. After VAH was established, G<sup>Q</sup> blockade did not affect ventilation while G<sup>S</sup> blockade increased VAH. Similar to LTF, data support roles for both G<sup>Q</sup> and G<sup>S</sup> pathways in the development of VAH but after VAH has been established, the G<sup>S</sup> pathway inhibits VAH.

### Edited by:

Jean-Paul R- Richalet, Université Paris 13, France

### Reviewed by:

Ginés Viscor, University of Barcelona, Spain Melissa L. Bates, The University of Iowa, United States

> \*Correspondence: Esteban A. Moya eamoya@ucsd.edu

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 12 April 2018 Accepted: 18 June 2018 Published: 06 July 2018

### Citation:

Moya EA and Powell FL (2018) Serotonin and Adenosine G-protein Coupled Receptor Signaling for Ventilatory Acclimatization to Sustained Hypoxia. Front. Physiol. 9:860. doi: 10.3389/fphys.2018.00860 Keywords: hypoxia, control of breathing, ventilatory acclimatization, serotonin, adenosine, neuroplasticity

# INTRODUCTION

Exposure to chronic sustained hypoxia (CH) produces (1) an increase in ventilation that persists after normoxia is restored and (2) an increase in the acute hypoxic ventilatory response (HVR). This is called ventilatory acclimatization to hypoxia (VAH) and it depends on plasticity in both carotid body chemoreceptors and medullary respiratory control circuits (Pamenter and Powell, 2016). Plasticity in ventilatory control circuits can also be produced by other patterns of hypoxia with similar results. For example, acute intermittent hypoxia produces long term facilitation (LTF) with increases in ventilation and phrenic nerve activity that persist in normoxia after the hypoxia protocol, and increases in the HVR to successive bouts of intermittent hypoxia [reviewed by Dale-Nagle et al. (2010), Pamenter and Powell (2016), and Turner et al. (2017)]. Despite similar physiological changes in LTF and VAH, several lines of evidence have been used to argue that

**231**

the signaling mechanisms for LTF do not explain VAH. For example, LTF is well-known to require activation of serotonin receptors (Bach and Mitchell, 1996; Baker-Herman and Mitchell, 2002) but serotonin receptor blockade does not reverse VAH after 4 h of hypoxia in goats (Herman et al., 1999), and whole-body serotonin depletion in rats does not block VAH after 1 day of sustained hypoxia (Olson, 1987). Also, exposure to sustained hypoxia for 25 min (i.e., a continuous exposure equal to the total duration of hypoxia in an intermittent hypoxia protocol that causes LTF) does not cause LTF (Devinney et al., 2013). Effects of longer exposure to hypoxia, such as 7 days used to show plasticity in CNS respiratory centers (Pamenter and Powell, 2016), have not been studied though.

More recently, a second mechanism for LTF that depends on adenosine receptors and more severe levels of intermittent hypoxia has been described (Nichols et al., 2012; reviewed in Pamenter and Powell, 2013). Exposure to moderate levels of intermittent hypoxia (arterial PO2 = 45–55 mm Hg) induces LTF by a serotonin-dependent pathway but exposure to more severe intermittent hypoxia (arterial PO2 = 25–35 mm Hg) also activated an adenosine-dependent pathway to induce LTF (Nichols et al., 2012). Increased phrenic nerve activity does not strictly depend on intermittent hypoxia per se and can be induced by direct pharmacological activation of serotonin 5-HT<sup>2</sup> receptors (MacFarlane and Mitchell, 2009) or adenosine A2A receptors (Golder et al., 2008). Both of these pathways depend on G-protein coupled receptor (GPCR) signaling but they involve different GPCR pathways. The serotonergic or "Q pathway" depends on activation of G<sup>Q</sup> protein, increased levels of BDNF and phosphorylation of ERK protein to induce phrenic LTF (Satriotomo et al., 2012). The adenosine or "S pathway" depends on activation of G<sup>S</sup> protein, PKA and phosphorylation of AKT (Devinney et al., 2013). Since the blocking of one these pathways can increase LTF, G<sup>S</sup> and G<sup>Q</sup> signaling interact via crosstalk inhibition (Dale-Nagle et al., 2010; Devinney et al., 2013; Navarrete-Opazo and Mitchell, 2014).

The role of adenosine-dependent G<sup>S</sup> mechanisms in VAH, and the contribution serotonin-dependent G<sup>Q</sup> mechanisms to exposures to sustained hypoxia longer than 1 day have not been studied. We hypothesized that longer exposure to moderate hypoxia could activate the adenosine-G<sup>S</sup> pathway described for LTF in severe intermittent hypoxia and contribute to VAH. To test this, we measured the hypoxic and hypercapnic ventilatory response in rats exposed to 7 days of CH with chronic blockade of adenosine A2A receptors during CH. We also tested the effects of chronic serotonin 5-HT<sup>2</sup> receptor blockade during 7 days of CH, and the effects of acute A2A and 5-HT<sup>2</sup> receptor blockade after VAH was established, to compare signaling mechanisms during VAH and LTF.

# MATERIALS AND METHODS

### Animals

Experiments were performed in male Sprague-Dawley rats (Harlan) weighing 250–300 g housed in 12:12 h light dark cycle and fed with standard diet at libitum except during measurements in the plethysmograph. All the experimental procedures were approved by the Institutional Animal Care and Use Committee of the University of California, San Diego.

# Chronic Hypoxia

The rats were exposed to CH in a hypobaric chamber for 7 days (barometric pressure = 380 Torr, PIO2 = 70 Torr, temperature 21◦C and 40% humidity), and normoxic control rats were housed in the same conditions in the room outside the chamber. The chamber was opened every other day for cage cleaning, and replacement of food and water.

### Plethysmography

Ventilatory responses to hypoxia and hypercapnia were measured in unrestrained rats using a whole body barometric plethysmograph (7 L) modified for continuous flow (Reid et al., 2005; Pamenter et al., 2014a). Briefly, flow was maintained constant through the chamber while a pressure transducer (mMP45 with 2 cmH2O diaphragm, Validyne) recorded the changes attributable to warming and expansion of inhaled gasses. On the experimental day, the rats were weighed and sealed into the plethysmograph chamber along with a temperature and humidity probe (Thermalert TH5, Physitemp). A constant gas flow (3 l/min) was delivered with a mass flow controller and gas mixer (MFC-4 Sable Systems) upstream of the chamber. Gasses exited the chamber through a valve and into a vacuum pump (Model 25, Precision Scientific) to isolate pressure changes from breathing in the chamber during constant flow with high input and output impedances. This also allowed us to maintain chamber pressure near-atmospheric pressure and reference pressure measurements in the chamber to atmosphere. Inspired and expired carbon dioxide fractions were measured with an O2/CO2 analyzer (FOXBOX Field Analysis system, Sable Systems) sampling from the chamber.

### Ventilatory Response Measurements

We measured the HVR and the hypercapnic ventilatory response (HCVR) with the following protocols. For normoxic control animals, we put rats in the plethysmograph for 45 min of acclimation to 21% O2, followed by 5 min of exposure to 10% O<sup>2</sup> to further acclimate rats to the experimental conditions, i.e., changes in inspired gasses. Then we returned rats to 21% O<sup>2</sup> for 15 min for the first measure of ventilation in baseline conditions (normoxia in this case). Rats were exposed to 10% O<sup>2</sup> for 15 min, returned to baseline conditions for 15 min, exposed to 7%CO2/21%O<sup>2</sup> for 15 min, and finally returned to baseline conditions for 15 min. The protocol was similar but opposite for CH rats. In this case, the baseline condition was 10% O<sup>2</sup> and they were acclimated to changes in inspired gas by exposure to 21% O<sup>2</sup> for 5 min. The HVR in CH rats was measured by exposing them to 21% O<sup>2</sup> for 15 min after a baseline breathing 10% O2.

All ventilatory parameters were recorded on an analog-digital acquisition system (PowerLab 8SP, AD Instruments) and analyzed with the LabChart 8-Pro Software, sampling at a rate of 1 kHz. We analyzed a minimum of 30 s between 10 and 15 min after changing gas concentrations for respiratory frequency (fR),

tidal volume (VT) and their product, inspired ventilation (V˙ <sup>I</sup>), which was normalized to body mass [ml/(min·kg)] using 0.2-ml calibration pulses (Drorbaugh and Fenn, 1955; Jacky, 1978).

## Serotonin and Adenosine Receptor Antagonist Administration

We designed two studies to determine the effects of serotonin and adenosine on VAH using ketanserin (an antagonist of serotonin 5HT<sup>2</sup> receptors) and MSX-3 (an antagonist adenosine A2A receptors). Firstly, to test the role of 5HT<sup>2</sup> and A2A receptors on VAH during CH exposure, we administered ketanserin tartrate (Tocris, Minneapolis, MN, United States) or MSX-3 hydrate (Sigma–Aldrich, St. Louis, MO, United States) continuously using osmotic pumps (1 mg/Kg/day for both). Rats were initially anesthetized with isofluorane (initially 5% and maintained with 1–2% in 100% O2) and we implanted mini osmotic pumps (Model 2002, Alzet Osmotic Pumps, Cupertino, CA, United States) filled with ketanserin or MSX-3 dissolved in 40% DMSO/Saline subcutaneously 1 day before start the CH exposure. Vehicle control rats were implanted with osmotic pumps filled with 40% DMSO/Saline (Vehicle). After surgery, rats were administrated with bupenophirine (0.03 mg/Kg, i.p.) and enrofloxacin (4 mg/Kg, i.p.).

Secondly, to assess the effect of 5HT<sup>2</sup> (Ketanserin) or A2A (MSX-3) antagonists on VAH after it was established by CH, we studied a different group of rats and measured ventilatory responses in the same individuals before and after exposure to CH (7 days). Then we injected Ketanserin or MSX-3 (1 mg/Kg i.p.) and returned the rats to hypobaric CH for an additional day. The next day, we injected the rats with antagonists again and repeated the ventilatory measurements.

# Statistics

Data was expressed as mean ± SEM. Statistical analysis was performed using two-way ANOVA between drug and CH effect or repeated measurement ANOVA test followed by Bonferroni post hoc analysis (GraphPad Prism, 5.0, United States). p < 0.05 was set as the level of statistical significance.

# RESULTS

# Chronic Serotonin 5HT<sup>2</sup> Receptor Blockade During CH Decreased Ventilatory Acclimatization to Hypoxia

To determine the contribution of serotonin receptors on VAH we studied the effect of the 5HT<sup>2</sup> receptor antagonist ketanserin administrated continuously in rats during exposure to CH. V˙ <sup>I</sup> increased with acute hypoxia (10% O2) and CH as expected for a normal HVR and VAH (**Figure 1**). In rats breathing 21% O2, there was a significant interaction for V˙ <sup>I</sup> between chronic O<sup>2</sup> level and ketanserin (p = 0.0001). Post hoc analysis showed V˙ <sup>I</sup> was significantly decreased by ketanserin after CH (**Figure 1B**) but not in normoxic control conditions (**Figure 1A**). In rats breathing 21% O2, the decrease in V˙ <sup>I</sup> was explained by a significant decrease in VT, which showed a significant interaction between chronic O<sup>2</sup> and drug (p = 0.01) (**Figures 1E,F**) while fR no longer showed a significant increase with drug after CH, which had been observed in normoxic controls (**Figures 1C,D**); the interaction of chronic O<sup>2</sup> level and ketanserin on fR was significant (p = 0.0292).

The effects of ketanserin in rats breathing 10% O<sup>2</sup> were similar to those observed when breathing 21% O<sup>2</sup> (**Figure 1**). V˙ <sup>I</sup> showed a significant interaction between chronic O<sup>2</sup> level and ketanserin (p < 0.0001) with ketanserin significantly decreasing V˙ <sup>I</sup> in 10% O<sup>2</sup> after CH, in contrast to significantly increasing it in normoxic control rats (**Figures 1A,B**). Ketanserin significantly increased fR in normoxic control rats during acute hypoxia while changes in fR with ketanserin were not significant in acute hypoxia after CH (**Figures 1C,D**). The interaction for chronic O<sup>2</sup> level and ketanserin were not significant for VT breathing 10% O<sup>2</sup> although it tended to decrease with ketanserin after CH (**Figures 1E,F**).

**Table 1** shows no significant effect of ketanserin on V˙ <sup>I</sup> in rats breathing 7% CO<sup>2</sup> before or after CH. V˙ <sup>I</sup> increased during hypercapnia after CH in all cases, as expected for acclimatization. Hence, differences observed in V˙ <sup>I</sup> and the HVR with ketanserin after CH are not explained by hypercapnic responses or general changes in ventilatory drive. **Table 2** shows metabolic rates (CO<sup>2</sup> production, V˙ CO2) were not significantly different between normoxic control and CH rats [23.2 ± 0.6 and 26.2 ± 2.0 (ml/(kg min))], with vehicle or ketanserin breathing 21 or 10% O2. Hence, the differences in V˙ <sup>I</sup> found between conditions are not explained by differences in the effects of metabolism on ventilatory drive.

Summarizing, 5HT<sup>2</sup> receptor blockade during CH blunted the increase in V˙ <sup>I</sup> in 10% O<sup>2</sup> that normally occurs with VAH and decreased V˙ <sup>I</sup> in 21% O<sup>2</sup> by a similar amount, i.e., there was a parallel downward shift of the HVR curve (V˙ <sup>I</sup> versus inspired O2). This was mainly due to an effect of ketanserin on VT. Ketanserin increased fR in 21% and 10% O<sup>2</sup> in normoxic control but not CH rats, which had a higher fR that was similar to the elevated level caused by ketanserin in the normoxic controls.

## Chronic Adenosine A2A Receptor Blockade During CH Exposure Decreased Ventilatory Acclimatization to Hypoxia

**Figure 2** shows that the general pattern of changes in ventilation after CH with the A2A antagonist MSX-3 were similar to those described above for effects of 5HT<sup>2</sup> receptor blockade. V˙ <sup>I</sup> increases with acute hypoxia and CH as expected for a normal HVR and VAH (**Figures 2A,B**). V˙ <sup>I</sup> breathing 21% O<sup>2</sup> showed a significant interaction between chronic O<sup>2</sup> level and MSX-3 (p = 0.0005) as MSX-3 tended to increase V˙ <sup>I</sup> in normoxic control rats (**Figure 2A**) and significantly decreased it in CH rats (**Figure 2B**). We observed similar patterns of change in fR and VT with a significant interaction between chronic O<sup>2</sup> level and MSX-3 for VT (p = 0.0002) and fR (p < 0.0001) (**Figures 2C–F**). The effects of MSX-3 during 10% O<sup>2</sup> breathing was similar; there was a significant interaction between chronic O<sup>2</sup> level and MSX-3 for V˙ <sup>I</sup> (p < 0.0011) so V˙ <sup>I</sup> was significantly decreased by MSX-3

ketanserin in Normoxic rats increased ventilation (V˙ I) in acute hypoxia (10% O2) and had no significant effect during normoxia (21% O2). (B) CH significantly increased V˙ <sup>I</sup> with vehicle; ketanserin significantly decreased V˙ <sup>I</sup> in CH rats during acute hypoxia and normoxia, in contrast to normoxic controls. (C–F) V˙ I increased with ketanserin in Normoxic rats because of frequency (fR) but it decreased with ketanserin in CH rats because of tidal volume (VT). <sup>∗</sup>p < 0.05, Bonferroni after two-way ANOVA, n = 6 rats per group.

after CH (p < 0.05) in contrast to a significant increase (p < 0.05) in the normoxic control rats.

**Table 1** shows that these changes in V˙ <sup>I</sup> with MSX-3 treatment were not due to changes in the ventilatory response to CO2, which was not significantly different between vehicle and drug. V˙ <sup>I</sup> increased during hypercapnia after CH in all cases, as expected for acclimatization. Also, V˙ CO<sup>2</sup> was not significantly different between normoxic control and CH rats with vehicle or MSX-3 breathing 21 or 10% (**Table 2**). Hence, the differences in V˙ <sup>I</sup> found between conditions are not explained by differences in the effects of metabolism on ventilatory drive.

Summarizing, A2A receptor blockade during CH blunted the increase in V˙ <sup>I</sup> in 10 and 21% O<sup>2</sup> that normally occur with VAH by a similar amount so there was a parallel downward shift of the HVR curve (V˙ <sup>I</sup> comparing acute O<sup>2</sup> level, 21 versus 10% O2). This was due effects of MSX-3 on both fR and VT.

### Acute Administration of Ketanserin After CH Did Not Reverse VAH

To test if 5HT<sup>2</sup> receptor blockade affects ventilatory responses after VAH is already established, we measured the ventilation in the same rats (1) before and (2) after 7 days of CH, and (3) after 2 more days of CH with two acute doses of ketanserin (1 mg i.p./kg daily). CH caused a significant increase of V˙ <sup>I</sup> during 21 or 10% O<sup>2</sup> breathing (**Figure 3A**). Ketanserin administered for 2 days more

TABLE 1 | Responses to hypercapnia in rats treated with 5HT<sup>2</sup> and A2A receptor antagonists during CH exposure.


Ketanserin or MSX-3 administered during chronic hypoxia (CH) did not affect the ventilation (V˙ I ) during hypercapnia (7% CO2). <sup>∗</sup>p < 0.05 versus Normoxia group breathing 7% CO2. Bonferroni's test after two-way ANOVA comparing the effects of Drug and CH during 7% CO2, n = 5–6 rats for all groups.

TABLE 2 | Metabolic rates in rats treated with 5HT<sup>2</sup> and A2A receptor antagonists during CH exposure.


CO<sup>2</sup> production (V˙ CO2) of rats treated with ketanserin and MSX-3 or vehicle during normoxic or chronic hypoxia (CH). No significant differences between conditions. Bonferroni's test after two-way ANOVA, n = 5–6.

of CH did not cause any significant difference in the HVR versus CH alone (**Figure 3A**). However, V˙ <sup>I</sup> in 21% O<sup>2</sup> breathing after CH + ketanserin resulted from a significantly decreased fR and increased VT (**Figures 3B,C**). Ketanserin did not affect neither V˙ <sup>I</sup> in 7% CO<sup>2</sup> (**Table 3**) nor metabolic rates (**Table 4**) after VAH was established.

### Acute Administration of MSX-3 After CH Increased Ventilation

To test if the A2A receptor blockade had an effect after VAH was established, we measured ventilatory responses using repeated measurements in rats during (1) normoxic control conditions, (2) after 7 days of CH, and (3) after 2 additional days of CH with acute injections of MSX-3 (1 mg/kg i.p. daily). Exposure to CH produced a significant increase of V˙ <sup>I</sup> in rats breathing 21 or 10% O<sup>2</sup> (**Figure 4A**). MSX-3 produced additional increases of V˙ <sup>I</sup> in both, 21 and 10% O<sup>2</sup> breathing (**Figure 4A**). This was due to significant effects of MSX-3 on fR and VT during 21% O<sup>2</sup> breathing and on non-significant increases of fR and VT during 10% O<sup>2</sup> breathing (**Figures 4B,C**). MSX-3 administration after VAH was established by CH did not have significant effects on CH-induced increases in ventilation during 7% CO<sup>2</sup> breathing (**Table 3**) nor produce changes in metabolic rate (**Table 4**).

### DISCUSSION

We found that chronic inhibition of either serotonin 5HT<sup>2</sup> or adenosine A2A receptors during chronic hypoxia decreased VAH: V˙ <sup>I</sup> breathing normoxic or hypoxic gas increased with CH plus drug significantly less than with CH plus vehicle. In contrast, acute blockade of 5HT<sup>2</sup> receptors after VAH had been established by CH had no effect on V˙ <sup>I</sup> and acute blockade of adenosine A2A receptors after VAH had been established actually increased V˙ <sup>I</sup> breathing normoxic or hypoxic gas (relative to CH without drug). We found no changes in metabolic rates between conditions that could cause other changes in ventilatory drive. Also we found no differences in ventilatory responses to CO<sup>2</sup> with drug treatments indicating that the effects of drugs were specific to hypoxic responses and any increases or decreases in V˙ <sup>I</sup> observed with drugs were not generalized increases in ventilatory drive or ventilatory limitations, respectively. Hence our results support a role for G<sup>Q</sup> and G<sup>S</sup> signaling in establishing ventilatory acclimatization to sustained hypoxia for 7 days. Furthermore, the different effects of GPCR antagonists during versus after CH might be explained by changes in cross-talk inhibition between G<sup>Q</sup> and G<sup>S</sup> pathways, activated by 5HT<sup>2</sup> and A2A receptors respectively. Both of these mechanisms have been shown to be important for chemoreflex plasticity with intermittent hypoxia in LTF but this is the first evidence for them with VAH from CH. However, it is important to note that we have not identified the site of action for these drug effects and we have not actually tested elements of the specific signaling pathways being activated by these receptors.

# G<sup>Q</sup> Signaling During Chronic Hypoxia and Ventilatory Acclimatization

In normoxic control rats, we found no effect of chronic 5HT<sup>2</sup> receptor blockade with ketanserin on V˙ <sup>I</sup> during normoxia. In normoxic control rats breathing 10% O<sup>2</sup> or 21% O<sup>2</sup> fR and V˙ <sup>I</sup> significantly increased (**Figure 1**). Similar increases in phrenic nerve responses to acute hypoxia following acute administration of ketanserin to anesthetized normoxic control rats have been reported, although fR increased more than VT in those experiments (Kinkead and Mitchell, 1999).

As expected, CH increased V˙ <sup>I</sup>, in rats breathing 21 or 10% O<sup>2</sup> but chronic ketanserin during CH significantly decreased V˙ <sup>I</sup> at both levels of inspired O2, mainly by effects of V<sup>T</sup> (**Figure 1**). These effects of blocking serotonin receptors and presumably G<sup>Q</sup> signaling on V˙ <sup>I</sup> in CH rats breathing normoxic gas are similar to those found for ventilatory and phrenic long term facilitation (vLTF and pLTF, respectively). Methysergide (a general 5HT receptor antagonist) administered before moderate acute intermittent hypoxia (IH), prevents phrenic LTF and specifically increased amplitude in integrated phrenic activity for 30–90 min after intermittent hypoxia (Bach and Mitchell, 1996). Also, administering ketanserin 20 min before moderate IH prevents the development of phrenic LTF (Kinkead and Mitchell, 1999; Fuller et al., 2001) and ketanserin 1 day before additional to immediately before moderate IH blocks ventilatory LTF (McGuire et al., 2004).

FIGURE 3 | Repeated measurements blocking 5HT<sup>2</sup> receptors after VAH was established. Blocking 5HT<sup>2</sup> receptors with ketanserin after chronic sustained hypoxia (CH) has no effect on ventilation (V˙ <sup>I</sup>, A), respiratory frequency (fR, B) or tidal volume (VT, C) during normoxia (21% O2) or acute hypoxia (10% O2). <sup>∗</sup>p < 0.05, Bonferroni after repeated measurements two-way ANOVA, n = 10 rats.

TABLE 3 | Responses to hypercapnia in rats treated with 5HT<sup>2</sup> and A2A receptor antagonists after CH exposure.


Chronic hypoxia (CH) increased V˙ I in normoxia (0% CO2, 21%O2) and the ventilatory response to hypercapnia (7% CO2); this was not affected by ketanserin or MSX-3 administered after acclimatization was established by CH. <sup>∗</sup>p < 0.05 versus 7% CO<sup>2</sup> Control, Bonferroni's test after repeated measures two-way ANOVA, n = 10 rats for ketanserin and n = 9 for MSX-3 treated rats.

TABLE 4 | Metabolic rates in rats treated with 5HT<sup>2</sup> and A2A receptor antagonists after CH exposure.


CO<sup>2</sup> production (V˙ CO2) was not affected by CH, inspired O<sup>2</sup> or drug with ketanserin or MSX-3 administered after acclimatization to CH. Bonferroni's test after repeated measures two-way ANOVA, n = 9–10 rats per group.

The effects of blocking 5HT<sup>2</sup> receptors and presumably G<sup>Q</sup> signaling on V˙ <sup>I</sup> in rats breathing hypoxic gas appears to be different after IH versus CH. However, we found chronic ketanserin decreased the normal increase in V˙ <sup>I</sup>, during 10% O<sup>2</sup> breathing with CH, thereby blunting the increase in acute HVR that occurs with VAH. This can be compared to the effects of ketanserin on progressive augmentation (PA), which is observed in studies of LTF with IH protocols as increases in ventilation or phrenic activity during successive bouts of hypoxia (Pamenter and Powell, 2016). The effects of IH on ventilatory drive during normoxia and hypoxia have been distinguished as LTF and PA, respectively, in part because LTF and PA seem to involve different mechanisms; PA is not sensitive to serotonin receptor antagonists in contrast to LTF [reviewed by Pamenter and Powell (2016)]. Hence the effects of G<sup>Q</sup> signaling on V˙ <sup>I</sup> in hypoxia appear to differ following CH versus IH. The duration of hypoxic exposure, duration of 5HT<sup>2</sup> receptor blockade, and/or effects of blocking 5HT<sup>2</sup> receptors outside the spinal cord may explain differences in the effects of ketanserin between studies.

### G<sup>S</sup> Signaling During Chronic Hypoxia and Ventilatory Acclimatization

The second mechanism for LTF operative during severe intermittent hypoxia was first described in a pharmacological study designed to test the hypothesis that phrenic motor facilitation (pMF) could be elicited by direct stimulation of a signaling pathways involving A2A receptors, mimicking brain derived neurotrophic factor (BDNF), which was known to be necessary and sufficient for LTF with moderate IH (Golder et al., 2008). The study found activating spinal adenosine A2A receptors produced pMF in anesthetized rats and increased normoxic ventilatory drive in conscious rats. Subsequently, this adenosine mechanism has been shown to work by G<sup>S</sup> signaling and contributes to LTF with severe IH as demonstrated by decreased LTF following spinal (intrathecal) administration of MSX-3 (Nichols et al., 2012). We hypothesized that longer exposure to moderate hypoxia may result in a comparable "dose" of hypoxia and evoke this mechanism to contribute to VAH with days of CH.

In normoxic control rats, chronic systemic blockade of A2A receptors and presumably G<sup>S</sup> signaling with MSX-3 did not affect V˙ <sup>I</sup> breathing 21% O<sup>2</sup> but it did increase V˙ <sup>I</sup> breathing 10% O<sup>2</sup> (**Figure 2A**). Spinal MSX-3 has no effect on phrenic nerve activity during normoxia in anesthetized rats using severe IH for LTF studies (Nichols et al., 2012), which agrees with our results in normoxic controls breathing 21% O2. However, MSX-3 did not affect PA in the LTF study (Nichols et al., 2012),

in contrast to the increase we would predict from effects we observed in acute hypoxia (**Figure 2A**). The duration of hypoxic exposure, duration of A2A receptor blockade, effects of A2A receptor blockade in other parts of the reflex pathway outside the spinal cord, and/or wakefulness may explain differences in the effects of MSX-between studies.

### Effects of Blocking G<sup>S</sup> and G<sup>Q</sup> Signaling After Ventilatory Acclimatization Is Established

Acute administration of ketanserin after VAH was established by CH did not change V˙ <sup>I</sup> in rats breathing 21 or 10% O<sup>2</sup> (**Figure 3**). These results agree with experiments done in goats were administration of ketanserin did not produce significant changes after sustained hypoxia (Herman et al., 2001). In contrast, acute administration of MSX-3 after VAH was established by CH increased V˙ <sup>I</sup> in rats breathing 21 or 10% O<sup>2</sup> (**Figure 4**). Note that this is opposite the effect of MSX-3 to decrease V˙ <sup>I</sup> when administered during CH (compare **Figures 2**, **4**). This might be explained by cross-talk inhibition between G<sup>S</sup> and G<sup>Q</sup> signaling that has been demonstrated recently for pMF (Hoffman et al., 2010; Devinney et al., 2013). In a model proposed for such crosstalk (see Figure 6, Devinney et al., 2016), G<sup>S</sup> and G<sup>Q</sup> signals following moderate acute sustained hypoxia (ASH) cancel each other so pMF is not observed. However, blocking adenosine A2A receptors with MSX-3 disinhibits a 5-HT<sup>2</sup> receptor-dependent mechanism in moderate ASH and reveals pMF (i.e., increases phrenic activity in normoxia after moderate ASH). In contrast, blocking 5-HT<sup>2</sup> receptors in moderate ASH does not increase phrenic activity; the A2A adenosine pathway is disinhibited but it is not sufficiently activated by moderate ASH to cause pMF. Severe ASH does activate the A2A adenosine mechanism, however, and 5HT<sup>2</sup> antagonists do increase pMF following severe ASH.

Hence, our results showing increased V˙ <sup>I</sup> with MSX-3 after CH agree better with plasticity induced in phrenic nerve activity by moderate versus severe ASH. We observed increases in V˙ <sup>I</sup> after MSX-3 but not ketanserin administered after VAH was established. Also, the increases with MSX-3 after CH were greater in 21 than 10% O<sup>2</sup> breathing (**Figure 4**), similar to the significant effect of MSX-3 after moderate ASH on phrenic activity in normoxia (i.e., pMF) but not on the phrenic response to hypoxia (see Figure 2, Devinney et al., 2016). Together, the data do not support our hypothesis that longer durations of moderate sustained hypoxia act more like shorter exposures to severe hypoxia. However, the level of hypoxia in our awake preparations was intermediate to moderate and severe hypoxia in these other studies (e.g., Devinney et al., 2016) with arterial PO2 being measured as 37–42 Torr in other studies of rats breathing 10% O<sup>2</sup> before or after acclimatization (Aaron and Powell, 1993; Popa et al., 2011). A meta-analysis of the effects of different patterns of IH on LTF shows that the physiological consequences (e.g., hypertension versus motor nerve facilitation) depend on both the level of hypoxia and the duration of hypoxia (i.e., total number of IH bouts) (Navarrete-Opazo and Mitchell, 2014). Further study is needed to draw any firm conclusions about the level of PO2 necessary to evoke specific mechanisms of plasticity with short versus long-term hypoxia.

### Limitations

Our results support both similarities and differences in G<sup>Q</sup> and G<sup>S</sup> signaling for plasticity in ventilatory control with CH and intermittent hypoxia. However, as noted above, we did not actually study specific signaling mechanisms but the effects of blocking GPCR that initiate these signaling pathways. Also, it is important that the site of action of drug effects in our experiments cannot be localized because drugs were administered systemically. Specifically, we have not studied the effects of receptor or signaling blockade in phrenic or hypoglossal motor neurons, which are primary sites of plasticity for LTF (Bach and Mitchell, 1996; Fuller et al., 2001; Baker-Herman and Mitchell, 2002). The drug effects we studied could be occurring at phrenic motor neurons but also at the at the carotid bodies, which increase O2-sensitivity with CH (Kumar and Prabhakar, 2012), and/or the nucleus tractus solitarii, which exhibits plasticity contributing to VAH and is the site of the primary synapse from carotid body chemoreceptors is in the CNS (Pamenter and Powell, 2016). While this limits our ability to localize and test precise mechanisms, it also described the physiological consequences of systemic therapeutics acting on these mechanisms as they would occur in a clinical situation without targeted application.

The effects we observed could certainly involve phrenic motor neurons, where G<sup>Q</sup> and G<sup>S</sup> signaling have been described for LTF (see above). However, the changes in ventilation that we observed with ketanserin could be explained also by serotonin effects at 5-HT<sup>2</sup> receptors in the carotid body, which contribute to increased O2-sensitivity with chronic intermittent hypoxia [i.e., "sensory LTF," reviewed by Peng et al. (2006) and Kumar and Prabhakar (2012)] and produced a prolonged chemosensory response to hypoxia (Jacono et al., 2005). Similarly, MSX-3 administration could involve effects on the peripheral chemoreceptors, since adenosine is modulating the activity of the carotid bodies acting through A2A receptors (Nurse and Piskuric, 2013; Conde et al., 2017). Kobayashi et al. (2000) proposed that adenosine inhibits voltage-dependent Ca2<sup>+</sup> currents in the glomus cells, suggesting that activation of A2A receptors have a inhibitory effect on the carotid body activity. On the other hand, a recent publication from Zhang et al. (2017) demonstrated that adenosine A2A receptors have a pre and postsynaptic effects in glomus cell and petrosal ganglion neuron co-cultures, which is consistent with the excitatory effects of adenosine in the carotid bodies observed by other authors (Fitzgerald et al., 2009; Nurse and Piskuric, 2013; Conde et al., 2017).

We have been especially interested in plasticity in the NTS, which is important for VAH as discussed above. Serotonin 5-HT2<sup>A</sup> receptors have been reported in the NTS and to explain in excitatory postsynaptic currents (Austgen and Kline, 2013). Adenosine A2A receptors in the NTS are reported to attenuate sympathetic reflexes (Minic et al., 2015) but their role in the ventilatory control has not been determined. Further investigation into possible roles for serotonin 5-HT2A and adenosine A2A receptors in the NTS on ventilatory

chemoreflexes and testing their potential contribution to plasticity demonstrated for glutamatergic neurotransmission in the NTS with VAH (Pamenter et al., 2014a,b), still need to be investigated. Serotonin acting at central as well as peripheral sites in high altitude adapted pikas modulates the HVR (Bai et al., 2015), so there is a precedent for serotonergic effects in the CNS as well as at the carotid bodies.

### CONCLUSION

The effect of manipulating G<sup>Q</sup> and G<sup>S</sup> signaling by blocking 5HT2 and A2a receptors, respectively, appear more similar for VAH and LTF in rats breathing normoxic versus hypoxic gas, and these effects are different before and after VAH is established. We also found some evidence for cross-talk between G<sup>Q</sup> and G<sup>S</sup> signaling in VAH but it was not reciprocal as demonstrated for LTF. However, all of these ideas need further research. Most

### REFERENCES


LTF studies do not focus on changes in the hypoxic response like VAH studies do, but rather the long-term changes in normoxic ventilatory drive. Most importantly, we need to determine the neuroanatomical substrate for these effects in VAH and the role of these signaling pathways in chemoreceptors, CNS integrative centers and motor neurons.

### AUTHOR CONTRIBUTIONS

EM and FP designed the experiments, interpreted the data, and wrote and reviewed the paper. EM performed the experiments, figures, data analysis, and first draft of the manuscript.

# FUNDING

This work was supported by NIH grant R01 HL-081823.



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Moya and Powell. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Physiological and Biological Responses to Short-Term Intermittent Hypobaric Hypoxia Exposure: From Sports and Mountain Medicine to New Biomedical Applications

Ginés Viscor <sup>1</sup> \*, Joan R. Torrella<sup>1</sup> , Luisa Corral <sup>2</sup> , Antoni Ricart <sup>2</sup> , Casimiro Javierre<sup>2</sup> , Teresa Pages <sup>1</sup> and Josep L. Ventura<sup>2</sup>

### Edited by:

Rodrigo Del Rio, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Cheryl Heesch, University of Missouri, United States Simona Mrakic-Sposta, Istituto di Bioimmagini e Fisiologia Molecolare, Italy

\*Correspondence:

Ginés Viscor gviscor@ub.edu

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 16 February 2018 Accepted: 11 June 2018 Published: 09 July 2018

### Citation:

Viscor G, Torrella JR, Corral L, Ricart A, Javierre C, Pages T and Ventura JL (2018) Physiological and Biological Responses to Short-Term Intermittent Hypobaric Hypoxia Exposure: From Sports and Mountain Medicine to New Biomedical Applications. Front. Physiol. 9:814. doi: 10.3389/fphys.2018.00814 <sup>1</sup> Physiology Section, Department of Cell Biology, Physiology and Immunology, Faculty of Biology, Universitat de Barcelona, Barcelona, Spain, <sup>2</sup> Exercise Physiology Unit, Department of Physiological Sciences, Faculty of Medicine and Health Sciences, Universitat de Barcelona, L'Hospitalet de Llobregat, Barcelona, Spain

In recent years, the altitude acclimatization responses elicited by short-term intermittent exposure to hypoxia have been subject to renewed attention. The main goal of short-term intermittent hypobaric hypoxia exposure programs was originally to improve the aerobic capacity of athletes or to accelerate the altitude acclimatization response in alpinists, since such programs induce an increase in erythrocyte mass. Several model programs of intermittent exposure to hypoxia have presented efficiency with respect to this goal, without any of the inconveniences or negative consequences associated with permanent stays at moderate or high altitudes. Artificial intermittent exposure to normobaric hypoxia systems have seen a rapid rise in popularity among recreational and professional athletes, not only due to their unbeatable cost/efficiency ratio, but also because they help prevent common inconveniences associated with high-altitude stays such as social isolation, nutritional limitations, and other minor health and comfort-related annoyances. Today, intermittent exposure to hypobaric hypoxia is known to elicit other physiological response types in several organs and body systems. These responses range from alterations in the ventilatory pattern to modulation of the mitochondrial function. The central role played by hypoxia-inducible factor (HIF) in activating a signaling molecular cascade after hypoxia exposure is well known. Among these targets, several growth factors that upregulate the capillary bed by inducing angiogenesis and promoting oxidative metabolism merit special attention. Applying intermittent hypobaric hypoxia to promote the action of some molecules, such as angiogenic factors, could improve repair and recovery in many tissue types. This article uses a comprehensive approach to examine data obtained in recent years. We consider evidence collected from different tissues, including myocardial capillarization, skeletal muscle fiber types and fiber size changes induced by intermittent

**241**

hypoxia exposure, and discuss the evidence that points to beneficial interventions in applied fields such as sport science. Short-term intermittent hypoxia may not only be useful for healthy people, but could also be considered a promising tool to be applied, with due caution, to some pathophysiological states.

Keywords: intermittent hypoxia, erythropoiesis, angiogenesis, cardioprotection, bronchial asthma, neuroprotection, circulating stem cells, regenerative medicine

### INTERMITTENT HYPOXIA: CONCEPT AND HISTORICAL BACKGROUND

The term "intermittent hypoxia" is widely used and applies to a wide spectrum of situations that range from alpine expeditions to obstructive sleep apnea (OSA). However, in physiological terms, there are often few similarities between these conditions (Viscor et al., 2014). Although the same physiological responses are elicited by the same sensors and signaling pathways, the varied intensity and duration of the hypoxia switches different mechanisms on and off at different times, thus making the final physiological changes induced in the whole organism highly variable. In addition, a high variability in human tolerance to hypoxia has been reported, and it is now known to vary throughout the lives of individuals (Canouï-Poitrine et al., 2014; Richalet and Lhuissier, 2015). As a consequence, specialists have engaged in an interesting discussion about how to measure and define hypoxic "dosage" (Serebrovskaya et al., 2008; Garvican-Lewis et al., 2016). In general, three different types of intermittent hypoxia can be considered:


The concept of "hypoxic training" was coined during the 1930s in the academic environment of the former Soviet Union, and was considered a useful therapeutic tool after it was shown to have beneficial effects on a number of different pathologies, even though the mechanisms involved in these favorable effects were unclear (Agadzhanyan and Torshin, 1986; Serebrovskaya, 2002). For instance, it was reported to have a beneficial effect on hypertension and cardiovascular diseases (Aleshin et al., 1993). Later, new experimental studies corroborated some of these findings and provided a fresh insight through enhanced plasma lipid profiles (Tin'kov and Aksenov, 2002). Nowadays in former Soviet States, hypoxic training is systematically applied as a nonpharmacological strategy for treating a wide range of alterations, including chronic lung disease, bronchial asthma, hypertension, diabetes mellitus, Parkinson's disease, emotional disorders and radiation toxicity, and for the prophylactic treatment of some occupational diseases (Ge et al., 1994; Xi and Serebrovskaya, 2012). In Western countries, intermittent hypoxia exposure programs were first applied in the field of sports medicine to improve aerobic capacity and for pre-acclimatization to altitude (Richalet et al., 1992; Levine and Stray-Gundersen, 1997; Rodríguez et al., 1999; Stray-Gundersen and Levine, 1999; Casas M. et al., 2000; Ricart et al., 2000).

Our group took an in-depth look at the physiological responses to intermittent exposure to hypobaric hypoxia (IEHH) in hypobaric (low barometric pressure) chambers. A detailed study of the precise mechanisms that underlie these adaptive responses (erythropoiesis, angiogenesis and the release of circulating stem cells) in humans and in an experimental rodent model encouraged us to explore the possibilities of applying IEHH programs with biomedical and therapeutic purposes (Panisello et al., 2007, 2008; Esteva et al., 2009; Viscor et al., 2009; Corral et al., 2014b). Thus, we recently demonstrated the efficacy of applying IEHH programs (passive exposure only or combined with exercise protocols) in the recovery of a range of injuries (Corral et al., 2014a; Núñez-Espinosa et al., 2014; Rizo-Roca et al., 2017a,b). These results are consistent with the new paradigm that proposes biphasic effects in the response to hypoxia (hormesis); that is, its harmful or beneficial effects depend on the frequency and severity of the hypoxic challenge to the organism or tissue in question (Navarrete-Opazo and Mitchell, 2014).

# BIOLOGICAL EFFECTS OF INTERMITTENT HYPOXIA EXPOSURE

The molecular mechanisms involved in the response to hypoxia at the cellular level are relatively well understood (Fábián et al., 2016; Bhattarai et al., 2017; Koyasu et al., 2017). However, the complexity of the interactions between the divergent signaling pathways and the time frame of the various processes on different tissues and organs, together with the significant individual variability in humans' susceptibility to hypoxia (Rathat et al., 1992), pose formidable challenges to researching the potential benefits of regular programs involving exposure to real or simulated altitudes. This paper presents a series of examples of how intermittent exposure to hypoxia can benefit certain patients, although it is not intended to be an exhaustive list. Obviously, these programs must always be applied with due caution and under rigorous clinical controls, as with any pharmacological treatment.

### Erythropoiesis

**Table 1** lists several representative studies of the favorable effects of intermittent hypoxia exposure by increasing erythropoiesis. A detailed report on the rate of erythropoietin (EPO) formation and plasma lifetime in humans in response to acute hypobaric hypoxia exposure in a hypobaric chamber was first provided by the group led by Christian Bauer in Zurich (Eckardt et al., 1989). Later, the pivotal role of hypoxia-inducible factor 1 (HIF-1) in the transcriptional response of EPO to hypoxia was also described (Wang and Semenza, 1993). These seminal works and many subsequent studies sparked an interest in applying IEHH programs to increase erythrocyte mass as a way of improving the aerobic capacity of athletes. Recombinant human erythropoietin (rHuEPO) opened up a new chapter in the correction of uremic anemia due to chronic renal failure (Egrie et al., 1986; London et al., 1989; Najean et al., 1989), but also led to unethical use in sports medicine. The celebration of the 1968 Summer Olympics held in Mexico City (2,250 m above sea level) aroused interest in studying the effects of altitude on human performance. The reduced performance in sporting events with a high aerobic component was evident, while participants in competitions with a clear anaerobic character saw no decline in their performance, and some even beat their records (Di Prampero et al., 1970). As a consequence, stays at high altitude and other artificial hypoxia exposure strategies in athletes were subject to intense study; a wide range of strategies, from permanent stays at moderate or high geographic altitudes (Antezana et al., 1994; Richalet et al., 1994) to different patterns of intermittent exposure (Levine and Stray-Gundersen, 1997; Chapman et al., 1998; Rodríguez et al., 1999; Casas M. et al., 2000; Karlsen et al., 2001; Stray-Gundersen et al., 2001; Ge et al., 2002; Lundby et al., 2007; Richalet and Gore, 2008) were examined. Noticeable differences in protocols and hypoxia exposure methods led to intense debate on the usefulness of intermittent hypoxia exposure for elite athletes, given that hypoxic dose and interindividual variability represent two of the main constraints (Chapman et al., 1998; Casas H. et al., 2000; Julian et al., 2004; Gore et al., 2006; Wilber et al., 2007; Truijens et al., 2008; Rodríguez et al., 2015). In parallel, there was rising interest in understanding the non-erythropoietic effects of EPO. The discovery of multi-tissue erythropoietin receptor expression provided an insight into erythropoietin (EPO) activity that went beyond its role in the regulation of red blood cell production, including a key role in cardioprotection, brain development and neuroprotection, through a coordinated response against tissue oxygen shortage (Noguchi et al., 2007; Arcasoy, 2008; Burger et al., 2009; Chateauvieux et al., 2011; Jia et al., 2012; Zhang Y. et al., 2014b), thus contributing to ischemic preconditioning, an interesting and important property of organ survival that could also prove very useful for new biomedical applications.

### Angiogenesis and Muscle Capillarization

In **Table 2**, some examples of the effects of intermittent hypoxia exposure on angiogenesis, vascular remodeling, muscle capillarization and hypertension are presented. In addition to its erythropoietic role, HIF-1 is the main mediator of angiogenesis in response to hypoxic conditions (Rey and Semenza, 2010) and has been considered a potential therapeutic target in many diseases. Two different strategies have been applied: HIF-1 upregulation for ischemic diseases (Li et al., 2014) and HIF-1 inhibition for cancer and endometriosis (Zhou et al., 2012; Bhattarai et al., 2017). Multiple angiogenic factors have been tested in the past; however, therapies that use only one proangiogenic agent to elicit angiogenesis were shown to be insufficient (Hirota and Semenza, 2006). Therefore, the addition of non-pharmacological treatments based on hypoxia-induced angiogenesis may be a successful strategy (Zimna and Kurpisz, 2015).

The physiological response elicited by hypoxia at moderate altitude exposure (1,800–3,000 m) is low, but increases when combined with exercise, with additional specific responses that are not observed when similar exercise levels are carried out in normoxia (Bärtsch et al., 2008). Moreover, even greater adaptations are obtained when the hypoxic intervention is accompanied by high-intensity training (Faiss et al., 2013; Sanchez and Borrani, 2018).

Although altitude is generally associated with increased health risks in most patients and elderly individuals, several studies have reported therapeutic benefits associated with exercising in mild hypoxia in a variety of alterations (Bailey et al., 2001; Burtscher et al., 2004; Wiesner et al., 2010). As exercising in moderate hypoxia seems to play a valuable role as an additional "therapeutic strategy," albeit one with both benefits and risks, new insights on this paradigm are now increasing. Some critical analysis and guidelines for hypertensive, obese and elderly individuals have recently been proposed (Millet et al., 2016). These concluded that intermittent hypoxic training seems to be well tolerated by most patients, in a similar way to healthy individuals, and that hypoxia and exercise may have additive or synergistic effects, probably mediated by several factors, including nitric oxide, angiogenesis and "altitude anorexia," thereby paving the way for researchers to identify the optimal individual combination of exercise and hypoxia.

### Cardiac Remodeling

Aerobic exercise activities have traditionally been widely recommended for preventing disease and promoting health. Today, resistance training is usually included in physical activity counseling, even for older adults and people with a range of cardiac conditions (Haskell et al., 2007; Nelson et al., 2007; Heuschmann et al., 2010), and there is solid evidence of different echocardiographic repercussions (Kenney et al., 2012). The cardioprotective effects of chronic intermittent hypoxia have been extensively studied, and their positive effect has shown to be related to preservation of mitochondrial function and inhibition


Altitude, geographic altitude; EPO, serum or plasma erythropietin levels; FiO2, Fraction of inspired oxygen; [Hb], Blood hemoglobin concentration; Hbmass, whole body hemoglobin mass; HC, Hypobaric Chamber; Hct, Hematocrit; IHT(5:5), intermittent hypoxic training, alternating 5 min hypoxia with 5 min normoxia along the session; LH-TH, Living High-Training High; LH-TL, Living High-Training Low; NH, normobaric hypoxia; RBC, red blood cell count; RBCmass, whole body erythrocytic mass.

of potassium channels sensitive to ATP (mitoKATP) present in the sarcoplasmic and mitochondrial membranes (Ostádal et al., 1989; Asemu et al., 1999, 2000; Chouabe et al., 2004; Ostadal and Kolar, 2007). Additional myocardial remodeling data were reported by our group using a model of IEHH (Panisello et al., 2007). Both chronic and intermittent exposure models supported the potential beneficial effects of acute exposure in coronary patients reported in pioneering studies conducted by Peruvian cardiologist Emilio Marticorena (Marticorena, 1993; Marticorena et al., 2001; Reynafarje and Marticorena, 2002; Valle et al., 2006). In rodents, it has been demonstrated that endurance exercise training and IEHH modulate cardiac mitochondria to a protective phenotype characterized by decreased induction of mitochondrial permeability transition pore and apoptotic signaling (Magalhães et al., 2013, 2014). However, there are no studies on humans on cardiac remodeling that combine


TABLE 2 | Examples of the effects of intermittent hypoxia exposure on angiogenesis, vascular remodeling, muscle capillarization, and hypertension.

FiO2, Fraction of inspired oxygen; HC, Hypobaric Chamber; HIF, hypoxia inducible factor; NH, Normobaric hypoxia; IHT, Intermittent hypoxic training alternating hypoxia with normoxia along the session; LAD, left anterior descending coronary artery; VO2max, maximal oxygen consumption capacity.

hypoxia and exercise, other than those dealing with OSA models. There have been extensive studies on patients suffering from this syndrome, and several pieces of evidence may be found in the literature: the progression and reversibility of atrial remodeling following stretch release may help prevent atrial fibrillation (Thanigaimani et al., 2017), the important prognostic information of right-sided heart dysfunction (Kusunose et al., 2016) and the evidence that OSA is associated independently with decreasing left ventricular systolic function and reduced right ventricular function (Korcarz et al., 2016). Nevertheless, this cardiac remodeling in OSA patients –individuals characterized by sustained systemic acidosis, hypercapnia and cerebral vasodilation– might not be present during intermittent hypoxia exposure in healthy subjects, in which alkalosis and hypocapnia, both induced by the hyperventilation caused by adrenergic drive, are evident and probably lead to cerebral vascular constriction and reduced effects of hypoxic insult (Viscor et al., 2014). Consequently, some of the changes and/or adaptive responses found in these pathological conditions must be interpreted with caution. In conclusion, there is no solid evidence for pernicious cardiac remodeling, but rather the opposite, after intermittent hypoxia in healthy individuals, whether accompanied by physical exercise or not. **Table 3** contrasts the deteriorated cardiac function in OSA patients in comparison to several studies demonstrating, both in experimental animal models and human coronary patients, the positive effects of intermittent hypoxia exposure on cardiac function.

### Treatment of Bronchial Asthma

Despite modern advances in the treatments, bronchial asthma continues being a potentially severe illness. All treatments focus on the improvement of bronchial obstruction, but nowadays we do not have an etiological definitive treatment. Bronchial asthma generates great dependence on a variety of medications and therefore frequently submits the patient to derived complications (Chiu et al., 1981; Fairfax et al., 1999; van der Woude et al., 2001; Salpeter et al., 2006). The Experts Committee of the United States Food & Drug Administration published a consensus document about the risks of the antiasthma medications, some of them potentially lethal (DeNoon, 2008).

For that reason, every procedure that could diminish pharmacological dependence among asthmatic patients should be considered a benefit. IEHH programs represent a realistic possibility to apply a minimally aggressive nonpharmacological approach that would reduce bronchial obstruction and pharmacological dependence in these patients.

As early as the nineteenth century there was social wisdom and medical knowledge that respiratory illnesses may improve in the mountains. The sanatoriums for respiratory patients, traditionally, were located at moderate altitude in the mountains. Notable examples were the Dutch Asthma Center, Davos, Clavadel, at 1,686 m over the sea level (Switzerland) and the Istituto Pio XII, Misurina, Auronzo at 1,756 m (Italy) devoted to childhood bronchial asthma. The first medical reference we found about bronchial asthma and altitude is an inquiry between the doctors of Davos referring that 133 among their 143 patients with bronchial asthma that spent their holidays in this mountain town, did not present any acute episode of asthma and that 81% reported persistent improvement of the illness (Turban and Spengler, 1906).

Moreover, different epidemiological studies showed the beneficial effects of living at moderate altitude in the prevalence and severity of bronchial asthma (van Velzen et al., 1996; Yangzong et al., 2006; Droma et al., 2007; Kiechl-Kohlendorfer et al., 2007; Sy et al., 2007). However, acute hypoxia exposure, as occurs with acute altitude exposure, as in many other stress situations, can induce an asthmatic episode of bronchoconstriction. On the other hand, when the acclimatization process advances, the asthmatic illness improved or even disappeared (Allegra et al., 1995; Christie


TABLE 3 | Examples of the positive effects of intermittent hypoxia exposure on cardiac pathologies.

Altitude, geographic altitude; DOXO, Doxorubicin treatment; HC, Hypobaric Chamber; NO, nitric oxide; OSA, Obstructive sleep apnea.

et al., 1995; Cogo et al., 1997, 2004; Gourgoulianis et al., 2001; Karagiannidis et al., 2006; Schultze-Werninghaus, 2006, 2008). Regrettably, this improvement vanished upon returning to the sea level.

The triad altitude exposure-hypoxia-acclimatization produces a number of physiological changes, some of which are accepted as related to the improvement of bronchial asthma: (a) a different breathing control pattern (Harrison et al., 2002; Serebrovskaya et al., 2003), (b) mitochondrial changes that optimize oxygen metabolism during the normal acclimatization process (Levett et al., 2012), and (c) the decrease in free radicals and the associated anti-inflammatory and immunosuppressive effects (Meehan, 1987; Simon et al., 1994; Serebrovskaya et al., 2003; Ohta et al., 2011; Oliver et al., 2013).

Since bronchial asthma improves with acclimatization to altitude and IEHH stimulates the acclimatization process (Rodríguez et al., 1999, 2000; Casas H. et al., 2000; Casas M. et al., 2000; Ibáñez et al., 2000; Ricart et al., 2000), it could be hypothesized that IEHH improves bronchial asthma. In fact, some medical studies have shown the usefulness of intermittent hypoxia exposure as a treatment for bronchial asthma (Harrison et al., 2002; Serebrovskaya et al., 2003). However, due to the different techniques and protocols used to produce the hypoxia and the wide dispersion of data, further research is required to design more effective protocols for intermittent hypoxia exposure that could prove useful in treating this disease. The ultimate objective of such treatments must be to reduce bronchial obstruction and the dependence on potentially dangerous drugs. Moreover, if protocols demonstrate a good response in bronchial asthma mitigation, they could also be useful for treating other illnesses with inflammatory backdrop. **Table 4** summarizes some of the results that demonstrate a favorable effect of exposure to hypoxia on the symptoms of bronchial asthma.

### Neurological Impact of Hypoxic Exposure

Hypobaric hypoxic exposure at altitude, usually long-term, results in several pathophysiological and psychological conditions associated with the nervous system. The term high altitude deterioration (HAD) was first used by members of early Mount Everest expeditions to denote the deterioration in mental and physical condition due to prolonged time spent at high altitudes (Ward, 1954). It is well known among climbers that staying at extreme altitudes for long periods is deleterious (Milledge, 2003). Manifestations vary depending on the altitude reached and the individual's hypoxia tolerance, but include acute and chronic mountain sickness, memory loss and high-altitude cerebral edema (Lieberman et al., 1994; Hornbein, 2001; West et al., 2013). Acute mountain sickness generally occurs 6 to 12 h after an unacclimatized person ascends to 2,500 m or higher (Bärtsch and Swenson, 2013). As a result, cognitive function may be impaired under hypoxia (Virués-Ortega et al., 2006), although the physiological changes that occur during acclimatization prevent mountain sickness. Given the acclimatization-like responses triggered by intermittent hypoxic exposure, it offers protection against severe hypoxia exposure damage and has been reported to produce beneficial effects (Kushwah et al., 2016). Our group reported how short-term (3-h sessions on three consecutive days) IEHH with surface muscle electrostimulation increased the concentration of circulating progenitor cells in the peripheral blood of humans (Viscor et al., 2009). However, we were unable to reproduce these results later in healthy patients and those with traumatic brain injuries (Corral et al., 2014a,b), thus raising doubts about the potential role of hypoxia exposure in the release of stem cells to circulation and its involvement in the tissue regeneration process. In any case, the translation of the physiological effects of IEHH to humans is not straightforward in the field of neurology.


Altitude, geographic altitude; BC, Base camp; FiO2, Fraction of inspired oxygen; HC, Hypobaric Chamber; HDM, house dust mite; IHT, Intermittent hypoxic training alternating hypoxia with normoxia along the session; ISSAC, International Study of Asthma and Allergies in Childhood; WHO, World Health Organization.

The brain's protective mechanisms involved in intermittent exposure to hypoxia have been widely studied using experimental animal models, and numerous beneficial effects have been reported. Intermittent hypoxia facilitates the proliferation of neural stem cells in situ in the subventricular zone and dentate gyrus of rat brains (Zhu et al., 2005; Ross et al., 2012). Xu et al. (2007) described a time-dependent migration of neural progenitor cells (NPC), promoted by hypoxia-induced astrocytes, thereby suggesting a role for astrocytes in NPC replacement therapy in the central nervous system. Intermittent hypoxia stimulated hippocampal angiogenesis and neurogenesis and improved short-term memory indices in control mice; and, in brain-injured mice, it reduced injury size and prevented memory impairments (Bouslama et al., 2015). It was recently reported that activation of HIF-1 is involved in hyperglycemia-aggravated blood-brain barrier disruption in an ischemic stroke model (Zhang et al., 2016b). Moreover, glycemic control by insulin abolished HIF-1α upregulation in diabetic animals and reduced blood-brain barrier permeability and brain infarction (Zhang et al., 2016b). Acute intermittent hypoxia can trigger spinal plasticity associated with sustained increases in respiratory, somatic and/or autonomic motor output (Streeter et al., 2017).

In rats, intermittent hypobaric hypoxia preconditioning caused a reduction in the degree of brain injury following ischemia-reperfusion by reducing hippocampal neuronal apoptosis by local upregulation of neuroglobin and Bcl-2 expression (Wu et al., 2015). Neuroglobin is an intracellular monomer hemoprotein that was discovered by Burmester et al. (2000) and is expressed in the central and peripheral nervous system, cerebrospinal fluid, retina and some endocrine areas of the brain (Burmester and Hankeln, 2004). It reversibly binds oxygen with a higher affinity than normal adult hemoglobin, and plays a critical role in brain tissue protection facing a possible oxygen delivery shortage (Ascenzi et al., 2016). Bcl-2 is an anti-apoptotic protein localized in the outer membrane of the mitochondria; overexpression of Bcl-2 in neurons can inhibit neuron apoptosis induced by ischemia-reperfusion injury by maintaining the integrity of mitochondria (Xing et al., 2008; Zhang et al., 2008).

Kushwah et al. (2016) also explored the ameliorating potential of intermittent hypoxia against the detrimental effects of unpredictable chronic mild stress (UCMS) on anxiety and depression-like behavior in rats, through the enhancement of neurogenesis in the hippocampus, a response mediated by brain derived neurotrophic factor (BDNF). In the postischemic rat brain, intermittent hypoxia intervention rescued ischemiainduced spatial learning and memory impairment by inducing hippocampal neurogenesis and functional synaptogenesis via BDNF expression (Tsai et al., 2013).

Nowadays, intermittent hypoxia exposure is known to enhance neurogenesis at multiple stages. Notch1 is a transcription factor in the neuron's membrane that regulates several stages of neurogenesis and promotes differentiation of progenitor cells into astroglia. Notch1 is activated by hypoxia in vivo, and such activation has been shown to be required for hypoxia-induced neurogenesis (Zhang K. et al., 2014a). Chronic IEHH pretreatment can reduce cerebral ischemic injury, which, as similarly reported for myocardium (see above), is mediated through upregulation of the expression and activity of mitochondrial membrane ATP-sensitive potassium channel (mitoKATP) (Zhang et al., 2016a). As is well known, hypoxia inducible factor-1 (HIF-1) is the key transcription factor that controls early adaptive responses to the lack of oxygen in mammalian cells. HIF-1α and HIF-1β expression was measured during acclimatization to hypobaric hypoxia in the rat cerebral cortex, and neurons, astrocytes, ependymal cells and possibly endothelial cells were the cell types that expressed HIF-1α (Chávez et al., 2000). Thus, the vascular remodeling and metabolic changes triggered during prolonged hypoxia may restore normal oxygen delivery levels to brain tissue (Agani et al., 2002; Chavez and LaManna, 2002).

Finally, there is solid evidence of the beneficial effects of intermittent hypoxia exposure on spinal cord neural tissue repair. Complete or incomplete spinal cord injuries are characterized by spared synaptic pathways below the level of the injury. Intermittent hypoxia elicits plasticity in the spinal cord and strengthens these spared synaptic pathways, expressed as respiratory and somatic functional recovery in both experimental animals and humans with traumatic spinal cord injury (Navarrete-Opazo et al., 2015, 2017a,b; Dougherty et al., 2017; Trumbower et al., 2017). **Table 5** lists studies reporting beneficial neurological impact after a wide range of intermittent hypoxia exposure protocols.

### Other Pathological Conditions Where the Use of Intermittent Hypoxia Exposure Has Potential Therapeutic Value

Since the altitude-hypoxia-acclimatization triad is known to have some antioxidant, anti-inflammatory and immunosuppressive effects (Meehan, 1987; Meehan et al., 1988; Ohta et al., 2011; Oliver et al., 2013), a benefit in some other pathologies related to immune response, such as psoriasis, atopy, arthritis or autoimmune pneumonitis can be expected. This is still a controversial field with no extensive clinical studies available, although some medical descriptive studies point to potential future research opportunities (Singh et al., 1977; Vocks et al., 1999; Engst and Vocks, 2000; Steiner, 2009).

Intermittent exposure to both normobaric and hypobaric hypoxia has been related to some protective effects (Cai et al., 2003; Costa et al., 2013) and beneficial outcomes in several pathological conditions, especially in those related to metabolic syndrome (Marquez et al., 2013; Leone and Lalande, 2017; Serebrovska et al., 2017). The possible role of intermittent hypoxia on body weight control has also attracted considerable attention. In addition to improving exercise performance and diet control, intermittent exposure protocols to normobaric and hypobaric hypoxia have been applied in an attempt to potentiate weight loss, showing in some cases positive short-term results (Haufe et al., 2008; Netzer et al., 2008; Lippl et al., 2010; Wiesner et al., 2010; Cabrera-Aguilera et al., in press). However, a long-term study failed to demonstrate permanent body weight reduction after IEHH (Gatterer et al., 2015), suggesting that additional research is needed to clarify the discrepancies reported in this field.

In summary, recent reports call for increased attention to the potential benefits of the application of intermittent hypoxia protocols in several clinical areas (Dale et al., 2014; Mateika et al., 2015). It is likely that future studies will yield important new information regarding potential therapeutic uses of intermittent hypoxia. **Table 6** lists a non-exhaustive but representative sample of studies reporting favorable impact of intermittent hypoxia exposure in other pathological conditions.

# INTERMITTENT HYPOXIA IN SPORT

The benefits of intermittent hypoxia programs in competitive sport are still a subject of scientific debate. **Table 7** shows several examples of intermittent hypoxia exposure effect on the improvement of human physical performance. A prior consideration to bear in mind when dealing with this topic is that two inherent factors justify the diversity of results in the field of elite sport: (a) the narrow margin of improvement detectable in elite athletes; and (b) the limitation in the sample size when performing studies with this population. Both factors contribute to reduced statistical power in most of these studies.

Other sources of variability are the wide range of exposure patterns, differing hypoxic doses (or altitude), and the kind of hypoxia (hypobaric or normobaric), as is discussed below. The

### TABLE 5 | Examples of the effects of intermittent hypoxia exposure with favorable neurological impact.


(Continued)

### TABLE 5 | Continued


In IHT protocols hypoxia was alternated with room air (FiO<sup>2</sup> = 0.209) if nothing else is indicated. AIH, acute intermittent hypoxia; BDNF, brain derived neurotrophic factor; Bcl-2, B cell lymphoma/leukemia-2; BP, barometric pressure; BrdU, 5-Bromo-2-deoxyuridine-5-monophosphate; BWSTT, body weight-supported treadmill training; CAO, carotid artery occlusion; CIHH, chronic intermittent hypobaric hypoxia; CPC, circulating progenitor cells; DG, dentate gyrus; FiO2, Fraction of inspired oxygen; GLUT-1, glucose transporter-1; HC, Hypobaric Chamber; Hct, hematocrit; HIF, Hypoxia inducible factor; Hsp70, heat shock protein70; IGF-1, insulin-like growth factor-1; IH, intermittent hypoxia; IHH, intermittent hypobaric hypoxia; IHT, intervallic hypoxic training alternating hypoxia and normoxia along the session; IMIP, imipramine; I/R, ischemia-reperfusion; KO, knockout mutant; MCA, medium cerebral artery; MCAO, middle cerebral artery occlusion; NH, normobaric hypoxia; NO, nitric oxide; NPC, neural progenitor cells; NX, normoxia (placebo); ER, endoplasmic reticulum; TBI, traumatic brain injury; UCMS, Unpredictable Chronic Mild Stress; VEGF, vascular endothelial growth factor; SES, surface electrical stimulation; SVZ, subventricular zone.

living high-training low (LH-TL) pattern (Levine and Stray-Gundersen, 1997) is the most widespread training schedule, although sometimes the reverse model, living low-training-high (LL-TH), is also applied, especially when artificial hypoxia is used. Moreover, studies evaluating training at altitude during permanent stays are also very usual.

In addition to other reports cited in precedent sections, a high number of studies have consistently found positive effects of IEHH programs to improve exercise performance. Thus, 4 weeks of LH-TL improved sea-level running performance in trained runners (Levine and Stray-Gundersen, 1997). Short-term intermittent hypobaric hypoxia (in a hypobaric chamber) improved the aerobic performance capacity in healthy subjects (Rodríguez et al., 1999). Intermittent hypobaric hypoxia combined with low-intensity exercise induced altitude acclimation, improved lactate threshold and ventilatory anaerobic threshold in healthy subjects (Casas M. et al., 2000). Normobaric hypoxia increased the growth hormone response to maximal resistance exercise in trained men (Filopoulos et al., 2017). Finally, a reduction in the severity of acute mountain sickness was also reported after several intermittent normobaric hypoxia protocols (Schommer et al., 2010; Wille et al., 2012; Dehnert et al., 2014).

In contrast, other studies did not detect significant improvements in exercise performance. Four weeks of IEHH did not improve oxygen transport in trained swimmers and runners (Rodríguez et al., 2007) nor did it change the submaximal economy in a group of well-trained athletes (Truijens et al., 2008). Seven weeks of normobaric hypoxia training in triathletes, caused an improvement in hematological parameters but not in the aerobic performance (Ramos-Campo et al., 2015). Finally, a recent systematic review and meta-analysis did not reveal a significant benefit of resistance training in hypoxia compared to the same training in normoxia (Ramos-Campo et al., 2018).

In general, it is commonly accepted that the application of intermittent hypoxia exposure has beneficial effects for competitive sport in the same way as for the biomedical field. As discussed above, the wide difference in effects that have been reported in the literature can be explained by individual susceptibility and the diversity of intermittent hypoxia patterns applied (Debevec and Mekljavic, 2013). The adaptive or maladaptive responses can be due to differences in the TABLE 6 | Examples of the effects of intermittent hypoxia exposure with favorable impact in other pathological conditions.


(Continued)

TABLE 6 | Continued


3 mmol/L Lac HR, heart rate corresponding to the 3 mmol/L lactate value in the FiO2-specific incremental test; Altitude, geographic altitude; AMS, acute mountain sickness; AUCins, Insulin response (area under curve) to oral glucose tolerance test; BMI, body mass index; BMR, basal metabolic rate; BP, arterial blood pressure; BW; body weight; CVAC, Cyclic Variations in Altitude Conditioning; ECP, eosinophil cationic protein; EET, endurance exercise training; EPO, Erythropietin; FiO2, Fraction of inspired oxygen; [Hb], Blood hemoglobin concentration; HC, Hypobaric Chamber; Hct, Hematocrit; HDL, high density lipoproteins; IHT, Intermittent hypoxic training alternating hypoxia with normoxia along the exposure protocol; HOMA-Index, homeostatic model assessment index of insulin resistance; HR, Heart rate; HVR, hypoxic ventilatory response; IHH, intermittent hypobaric hypoxia; IHT (6:6), intermittent hypoxic training, alternating 6 min hypoxia with 6 min normoxia along the session; LAE, light aerobic exercise; LDL, low density lipoproteins; NH, normobaric hypoxia; TG, plasma triglycerides; VEmax; maximal exercise pulmonary ventilation; VLDL, very low density lipoproteins.

frequency, severity or duration of hypoxic episodes (Serebrovska et al., 2016). Factors such as age, sex or genotypic variability may also contribute to varying results (Almendros et al., 2014). A review of relevant publications between 1980 and 2015 concluded that evidence regarding the effects of altitude training on athletic performance is weak but that the natural stay at altitude combined with a live high-train low training strategy may provide the best protocol for enhancing endurance performance in elite and subelite athletes (Khodaee et al., 2016), thus confirming similar findings in a previous study (Bonetti and Hopkins, 2009). Finally, a recently published meta-analysis on the effect of natural or simulated altitude training in teamsport athletes conclude that hypoxic intervention appears to be a worthwhile training strategy for improvement in teamsport athletes, with enhanced performance over control groups persisting for at least 4 weeks post-intervention (Hamlin et al., 2018). Also, our recent data indicate that contractile activity seems to be necessary to trigger in skeletal muscle the adaptive responses induced by intermittent exposure to hypoxia (Rizo-Roca et al., 2018).

Although most of the expected effects of altitude training and IEHH programs have been mainly aimed at improving aerobic capacity (Stray-Gundersen and Levine, 1999; Wilber, 2004), some reports also have described benefits for strength training at moderate hypoxia levels (Nummela and Rusko, 2000; Manimmanakorn et al., 2013; Álvarez-Herms et al., 2015b, 2016). The underlying mechanism of these responses remain to be clarified, but potential psychological benefits may be derived of the perception of increased effort during hypoxic training (Álvarez-Herms et al., 2016). In addition, training during hypoxia may result in a greater increase in muscular endurance than the same training load performed in normoxia, probably because of increased angiogenesis in skeletal muscle level (Kon et al., 2014), can be involved in those improvements.

### NORMOBARIC VS. HYPOBARIC HYPOXIA: THE SAME STIMULUS?

When artificial methods of producing hypoxia exposure were first developed, no attention was paid to the differences between hypobaric (low barometric pressure) and normobaric (low oxygen content in an inhaled gas mixture) hypoxia. It seems evident that the same physiological effects are expected for TABLE 7 | Examples of the intermittent hypoxia exposure on the improvement of human physical performance.


(Continued)

TABLE 7 | Continued


1RM, one repetition maximum; Altitude, geographical altitude; AMS, acute mountain sickness; CST, circuit strength training; FiO2, Fraction of inspired oxygen; HC, Hypobaric Chamber; HIIE, high intensity intervallic exercise; HT, hypoxic test (at 4,800 m equivalent attitude); IHT, Intermittent hypoxic training alternating hypoxia during exercise with normoxia during recovery between sets; Lact/Vel, Lactate/Velocity curve (up-arrow means right shift); LH-TL, Living High-Training Low; LH-TH, Living High-Training High; MVC3, 3 s maximal voluntary contraction; MVC30, 30 s maximal voluntary contraction; NH, Normobaric hypoxia; TtE, Time to exhaustion; Reps20, maximal number of repetitions at 20% 1RM; TT, time trial; USARIEM, United States Army Research Institute of Environmental Medicine; VO2max, maximal oxygen consumption capacity.

a determined alveolar oxygen partial pressure, regardless of the technical means used. However, data in the literature present consistent discrepancies after applying hypobaric and normobaric hypoxia, which could be attributed to alterations in other environmental parameters that affect alveolar gas composition, such as carbon dioxide partial pressure, humidity and temperature. For instance, even for the same level of oxygen partial pressure, the atmospheric composition in a small hypoxic tent with low air turnover (limited by the flow capacity of the hypoxic device) can be very different from the air composition in a hypobaric chamber. Generally, the use of powerful vacuum pumps in hypobaric chamber systems guarantees sufficient renewal of the air inside the room, thus preventing carbon dioxide accumulation and a rise in air temperature and humidity, factors that could become unbalanced in small volume hypoxic tents, especially with exercising subjects inside. In a comparative study of hypobaric (low pressure chamber) and normobaric hypoxia (hypoxic tent) during a submaximal exercise test in the same subjects, differences were observed in some cardiorespiratory and heart rate variability parameters between the two artificial hypoxia systems used (Basualto-Alarcón et al., 2012).

An interesting and dynamic debate, which is beyond the scope of this review, is currently under way among altitude researchers concerning whether or not hypobaric hypoxia induces different responses from normobaric hypoxia (Girard, 2012; Millet et al., 2012; Hauser et al., 2016; Saugy et al., 2016). Some evidences demonstrate that there are different physiological responses and outcomes between exposure to normobaric and hypobaric hypoxia conditions (Savourey et al., 2003; Fulco et al., 2013; Millet et al., 2013; Beidleman et al., 2014; Debevec and Millet, 2014; Boos et al., 2016; DiPasquale, 2017). For instance, it has been reported that the decrease in air density that accompanies the partial pressure drop of oxygen at geographic altitude affects the way in which explosive actions are executed and increases movement velocity and power during force-velocity bench presses in comparison to normobaric hypoxia (Feriche et al., 2014).

# CONCLUSIONS

A growing body of knowledge supports the beneficial effects of natural or simulated altitude techniques on health outcomes (Navarrete-Opazo and Mitchell, 2014; Millet et al., 2016; Lizamore and Hamlin, 2017). Future research should be oriented to: (1) gain more in-depth knowledge of the subcellular mechanisms involved in the hypoxic response at different tissue levels, (2) standardize hypoxia exposure methods and establish a universal method for measuring, and repeatedly applying, hypoxic dosage, (3) improve predictions of individual hypoxia tolerance to prevent possible negative consequences, (4) apply this new knowledge to the selection and education of altitude workers, (5) improve altitude acclimatization, altitude training camps and altitude competition events to benefit mountaineers, athletes and coaches, and finally (6) cautiously explore the application of IEHH in pathological conditions.

### AUTHOR CONTRIBUTIONS

GV, JV, AR, and LC contributed to the initial draft of the manuscript; GV and JT edited the document after contributions

### REFERENCES


from all authors; All authors reviewed and approved the final version.

# ACKNOWLEDGMENTS

This study was supported in part by research grants DEP2010- 22205-C02-01, DEP2010-2205-C02-02, DEP2013 48334-C2-1P, and DEP2013-48334-C2-2-P from the Plan Nacional de I+D+i (Spain's Ministry of Economy, Industry and Competitiveness). The authors thank Christopher Evans and Paula Weiss (Language Advisory Service, Universitat de Barcelona) for their help editing the manuscript.


resistance with terrestrial altitude alters run sprint performance. J. Appl. Physiol. (Bethesda, MD 1985) 112:1788. doi: 10.1152/japplphysiol.00356.2012


asthma in children: results from a large prospective birth-cohort study. Arch. Dis. Child. 92, 339–342. doi: 10.1136/adc.2006.106278


glucose metabolism in middle-aged men. High Alt. Med. Biol. 14, 263–272. doi: 10.1089/ham.2012.1057


biomarkers by intermittent hypobaric hypoxia and aerobic exercise after eccentric exercise in trained rats. Appl. Physiol. Nutr. Metab. 42, 683–693. doi: 10.1139/apnm-2016-0526


with house-dust mite allergic asthma. Pediatr. Pulmonol. 17, 304–311. doi: 10.1002/ppul.1950170507


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Viscor, Torrella, Corral, Ricart, Javierre, Pages and Ventura. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Contribution of Oxidative Stress and Inflammation to the Neurogenic Hypertension Induced by Intermittent Hypoxia

María P. Oyarce and Rodrigo Iturriaga\*

Laboratorio de Neurobiología, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Santiago, Chile

Chronic intermittent hypoxia (CIH), the hallmark of obstructive sleep apnea, is the main risk factor to develop systemic hypertension. Oxidative stress, inflammation, and sympathetic overflow have been proposed as possible mechanisms underlying the CIH-induced hypertension. CIH potentiates the carotid body (CB) chemosensory discharge leading to sympathetic overflow, autonomic dysfunction, and hypertension. Oxidative stress and pro-inflammatory molecules are involved in neurogenic models of hypertension, acting on brainstem and hypothalamic nuclei related to the cardiorespiratory control, such as the nucleus of the solitary tract, which is the primary site for the afferent inputs from the CB. Oxidative stress and pro-inflammatory molecules contribute to the activation of the CB chemoreflex pathway in CIHinduced hypertension. In this brief review, we will discuss new evidence for a critical role of oxidative stress and neuro-inflammation in development of the CIH-induced hypertension through activation of the CB chemoreflex pathway.

### Edited by:

Valdir Andrade Braga, Federal University of Paraíba, Brazil

### Reviewed by:

Linda F. Hayward, University of Florida, United States Liang-Wu Fu, University of California, Irvine, United States

> \*Correspondence: Rodrigo Iturriaga riturriaga@bio.puc.cl

### Specialty section:

This article was submitted to Autonomic Neuroscience, a section of the journal Frontiers in Physiology

Received: 30 March 2018 Accepted: 21 June 2018 Published: 11 July 2018

### Citation:

Oyarce MP and Iturriaga R (2018) Contribution of Oxidative Stress and Inflammation to the Neurogenic Hypertension Induced by Intermittent Hypoxia. Front. Physiol. 9:893. doi: 10.3389/fphys.2018.00893 Keywords: carotid body, chronic intermittent hypoxia, hypertension, inflammation, nucleus of the solitary tract, pro-inflammatory cytokines

### ROLE OF THE CAROTID BODY IN THE CIH-INDUCED HYPERTENSION

An abnormal heightened carotid body (CB) chemosensory discharge, which elicits sympathetic overflow, has been involved in the cardiovascular and autonomic alterations in preclinical models of human diseases such as obstructive sleep apnea (OSA), systolic heart failure, and neurogenic hypertension (Sun et al., 1999; Peng et al., 2003; Rey et al., 2004; Del Rio et al., 2010; Abdala et al., 2012; McBryde et al., 2013). The OSA syndrome characterized by repeated episodes of chronic intermittent hypoxia (CIH) is considered an independent risk factor for systemic hypertension, and is associated with atrial fibrillation, stroke, and heart failure (Somers et al., 2008; Dempsey et al., 2010). The cardiovascular consequences of OSA has been attributed to oxidative stress, inflammation, and sympathetic overflow induced by CIH, but other factors are influential, such as sleep fragmentation and co-morbid metabolic diseases (Gozal and Kheirandish-Gozal, 2008; Somers et al., 2008; Dempsey et al., 2010; Iturriaga et al., 2016; Iturriaga, 2017). OSA patients show enhanced sympathetic, vasopressor and ventilatory responses to hypoxia, attributed to a

potentiated hypoxic peripheral chemoreflex (Somers et al., 2008; Dempsey et al., 2010). Similarly, rodents exposed to CIH show enhanced cardiorespiratory and sympathetic hypoxic responses, and develop hypertension (Fletcher et al., 1992; Peng et al., 2003; Iturriaga et al., 2009; Del Rio et al., 2010, 2014; Kumar and Prabhakar, 2012). Neural recordings of rat and cat CB chemosensory discharges have shown that CIH selectively increases the baseline discharge in normoxia and enhances the chemosensory responses to hypoxia (Peng et al., 2003; Rey et al., 2004; Del Rio et al., 2010). The enhanced CB chemosensory discharge induced by CIH has been linked with local oxidative stress and increased endothelin-1 (ET-1) levels in the CB (Peng et al., 2003; Rey et al., 2006; Del Rio et al., 2010). The enhanced CB chemosensory discharge plays a crucial role in the onset and progression of the hypertension induced by CIH. Indeed, Fletcher et al. (1992) found that CBs denervation prevents the hypertension in rats exposed to CIH. Furthermore, Del Rio et al. (2016) found that CBs ablation in hypertensive rats exposed to CIH for 21 days, restores the autonomic balance, the cardiac baroreflex sensitivity and reduces the elevated arterial pressure (BP), even when the CIH stimuli was maintained for 7 days and systemic oxidative stress persisted after the elimination of the CBs. Thus, the available evidence supports a crucial role of the CB in the onset and progression of the hypertension induced by CIH.

### CIH AND CB CHEMOREFLEX POTENTIATION

The CB chemoreceptor cells are innervated by sensory petrosal neurons that project to the nucleus of the tractus solitarius (NTS), which is the primary site of integration of gastrointestinal, respiratory and cardiovascular information in the brainstem (Berger, 1980; Finley and Katz, 1992; Grimes et al., 1995). The projections from the petrosal neurons that innervate the CB reach the caudal section of the NTS, specifically the dorsal, medial, and commissural sub-nuclei. In the NTS, second and thirdorder neurons project to the paraventricular nucleus (PVN) and the rostral ventrolateral medulla (RVLM), where are located the pre-sympathetic neurons (Guyenet, 2006). Hypoxia depolarizes chemoreceptor cells releasing excitatory transmitters, which in turn increases the frequency of discharge in the petrosal neurons eliciting reflex hyperventilatory, autonomic and vasopressor responses (Iturriaga and Alcayaga, 2004; Nurse and Piskuric, 2013). CIH enhanced the normoxic CB chemosensory discharge and the neural activity of the cardiorespiratory neurons in the brainstem and hypothalamus (Iturriaga et al., 2017). Indeed, CIH increases the electrical activity of glutamatergic neurons in the NTS (de Paula et al., 2007) and the number of c-fos or FosB positive neurons in the NTS, RVLM, PVN, in the subfornical organ (SFO) and median preoptic nucleus (Knight et al., 2011; Bathina et al., 2013; Sharpe et al., 2013). The activation of NTS and RLVM neurons induced by CIH is associated with local oxidative stress (Peng et al., 2014). Moreover, the increased Fos B in the RVLM induced by CIH (An and En-Shang, 2014) was attenuated by systemic pretreatment with a superoxide dismutase mimetic (Kuo et al., 2011). Thus, it is likely that the CIH-induced activation of the NTS and RVLM neurons is the result of oxidative stress (Daulatzai, 2012). Another plausible explanation is that the activation of the CB chemoreflex neural pathway triggered by the enhanced CB chemosensory discharge may elicit oxidative stress and neuroinflammation in the brainstem. This idea is strongly supported by the finding that CB neurotomy performed before the onset of CIH exposure prevents the oxidative stress in the NTS and RVLM, and the development of the hypertension in rats (Peng et al., 2014).

### OXIDATIVE STRESS AND INFLAMMATION IN THE CB AND THE CHEMOREFLEX NEURAL PATHWAY INDUCED BY HYPOXIA

Reactive oxygen species (ROS) and reactive nitrogen species (RNS) contribute to enhance the CB chemosensory discharge and the progression of the hypertension in rats exposed to CIH (Prabhakar, 2000; Del Rio et al., 2010; Peng et al., 2014). Indeed, the treatment with antioxidants normalized the enhanced CB chemosensory discharge and prevents or reverses the elevated BP in CIH-treated rats (Peng et al., 2003, 2009; Del Rio et al., 2010; Moya et al., 2016).

In addition other molecules downstream the ROS signaling pathway may mediate the CIH-induced excitatory effects on CB chemoreception. Thus, we hypothesized whether proinflammatory molecules may contribute to enhance the CB chemosensory discharge (Iturriaga et al., 2009). Inflammation is part of the response of the immune system to tissue damage or pathogen invasion (Hänsel et al., 2010). The classical clinical signs of inflammation include increased blood flow, capillary permeability, release of inflammatory mediators and the migration of leukocytes (Hänsel et al., 2010). These processes are orchestrated by molecules activated by the nuclear transcription factor κB (NF-κB), which stimulates the release of pro-inflammatory cytokines such as tumor necrosis factor alpha (TNF-α), interleukin 1β (IL-1β) and interleukin 6 (IL-6), chemokines and adhesion molecules (Shih et al., 2015). The combination of cycles of hypoxia followed by re-oxygenation in OSA patients is associated with an increase of plasma levels of TNF-α, IL-6, and C-reactive protein (Meier-Ewert et al., 2004; Irwin et al., 2010). Most of the cellular responses and adaptations to hypoxia are mediated by the hypoxia-inducible factor-1α (HIF-1α) (Prabhakar and Semenza, 2012; Peng et al., 2014). NF-κB is a critical transcriptional activator of HIF-1α and it is necessary for the accumulation of HIF-1α during hypoxia (Hocker et al., 2017). On the other hand, hypoxia may directly activate the NF-κB factor, promoting the transcription of pro-inflammatory cytokines (Eltzschig and Carmeliet, 2011). Moreover, in response to oxidative stress, HIF-1α evokes the translocation of NF-κB to the nucleus increasing the expression of IL-1β, TNF-α, and ET-1 among other pro-inflammatory molecules (Chang et al., 2009). Zhang et al. (2015) studied the serum levels of inflammatory cytokines and the activation of NF-κB and HIF-1α in myocardial

tissues in response to different frequencies of CIH (10–40 times/h for 6 weeks) and the actions of the antioxidant tempol. Intermittent hypoxia increased the serum levels of TNF-α, along with an increase on myocardial expression of NF-κB and HIF-1α in a frequency-dependent manner. Interesting, tempol treatment attenuated this effect (Zhang et al., 2015). Therefore, there is an interplay between oxidative stress, inflammation, and hypoxic induced factors under CIH conditions.

Chronic intermittent hypoxia increases the levels of proinflammatory molecules in the CB (Del Rio et al., 2011, 2012; Lam et al., 2012). Indeed, Rey et al. (2006) found that CIH increased ET-1 in the CB from cats exposed to CIH for 4 days, while bosentan reduced the CB chemosensory response to hypoxia in vitro in CIH-treated cats, but not in sham animals. Lam et al. (2012), reported that 7 days of exposure to CIH increased the mRNA levels of TNF-α, IL-6 and IL-1β, and their receptors in the rat CB. Moreover, Del Rio et al. (2012) found that exposure to CIH for 21 days produced a progressive increase of the immunoreactivity levels of TNF-α, IL-1β and iNOS in the rat CB, while ET-1 showed a transient increase during the first week of CIH. These results suggest that pro-inflammatory molecules may mediate the onset (ET-1) and the maintenance (proinflammatory cytokines) of the CB chemosensory potentiation. The treatment with ascorbic acid abolished the CIH-induced increases of TNF-α and IL-1β immunoreactivity levels in the CB, suggesting that inflammation depends on oxidative stress in the CB (Del Rio et al., 2012). The treatment with ibuprofen administered systemically during CIH did not reduce the enhanced CB chemosensory responses to hypoxia, although reduced the increased chemosensory baseline and the increased levels of pro-inflammatory cytokines in the CB (Del Rio et al., 2012). Nevertheless, the administration of ibuprofen prevents the hypertension induced by CIH exposure and the ventilatory acclimatization in rats, suggesting that ibuprofen may act in other sites of the chemoreflex pathway (Del Rio et al., 2012). Ibuprofen also prevents the increase in the number of c-fos positive neurons in the cNTS of rats subjected to CIH (Del Rio et al., 2012). Therefore, pro-inflammatory molecules may act on other neural structures of the CB chemoreflex pathway, such as the brainstem cardiorespiratory centers. Similarly, Popa et al. (2011), found in rats exposed to sustained hypoxia that the systemic administration of ibuprofen blocked the increase of IL-1β and IL-6 in the NTS and reduced the ventilatory response, indicating that these cytokines were crucial for the onset of the hyperventilatory response elicited by hypoxia (Popa et al., 2011). More recently, De La Zerda et al. (2017) tested the hypothesis that inflammatory signals are necessary to ventilatory acclimatization to sustained hypoxia applied for 11 days once it is established in rats. They found that hyperventilation was not affected by ibuprofen when was administered for the last 2 days of the hypoxic exposure (De La Zerda et al., 2017). In addition, they found that hypoxia (1 h) activated microglia in the NTS, effect that was abolished by ibuprofen administered from the beginning of hypoxic exposure (De La Zerda et al., 2017). The activation of the microglia induced by acute hypoxia lasted for 7 days, and was not altered by ibuprofen administered 2 days after the end of the hypoxia (De La Zerda et al., 2017). Thus, an early increase of pro-inflammatory molecules is required to produce hyperventilation following sustained hypoxia. Snyder et al. (2017) collected tissue punches from brain regions associated with different stages of neurodegenerative diseases in rats exposed CIH and measured oxidative stress and inflammatory markers. They found that CIH for 7 days produces oxidative stress and increases pro-inflammatory cytokines in brain areas associated to early stages of neurodegeneration (substantia nigra and entorhinal cortex) but not in the NTS and RVLM (Snyder et al., 2017). Our results agree with those observations. We found IL-1β, IL-6, and TNF-α mRNA levels were augmented in the NTS of hypertensive rats after 21 days of CIH (Oyarce and Iturriaga, 2018). These findings suggest that pro-inflammatory cytokines in the NTS may contribute to the maintenance of the hypertension, since CIH increases BP in 3–4 days in conscious rats, paralleling the time required to establish the enhanced CB chemosensory discharge (Del Rio et al., 2016).

### NEUROGENIC HYPERTENSION AND INFLAMMATION

Neurogenic hypertension is associated with sympathetic overflow, increased plasma angiotensin II (Ang II) and C-reactive protein, TNF-α, IL-6, monocyte chemotactic protein 1, and adhesion molecules (Zubcevic et al., 2011), highlighting the importance of peripheral inflammation in hypertension. However, the role of central inflammation in neurogenic hypertension is gaining recognition. In the central nervous system, both circulating or released pro-inflammatory cytokines by astrocytes and microglia act on brainstem cardiovascular neurons (Shi et al., 2010). Waki et al. (2007) found that inducing inflammation in the NTS of normotensive rats by increasing the expression of the junctional adhesion molecule (JAM-1), triggers hypertension. They also compared the expression of JAM-1 in the NTS in young and adults spontaneously hypertensive rats (SHRs) and normotensive Wistar–Kyoto rats and found that JAM-1 mRNA was highly expressed in the NTS from SHR rats. Waki et al., 2010 using RT2 Profiler PCR arrays to detect changes on gene expression of cytokines and chemokines in the NTS from SHR, reported an abnormal expression of inflammatory mediators with relevant roles in the cardiovascular homeostasis, suggesting that cytokines may contribute to the hypertension by increasing the neuronal activity in the NTS (Waki et al., 2010). McBryde et al. (2013), reported that CB denervation reduces the number of CD3+ cells in the homogenate of the brainstem of SHR rats, suggesting that an enhanced CB chemosensory drive may induce the infiltration of T cells in brain tissues associated with the BP control (McBryde et al., 2013). The same group showed that systemic inflammation induced by LPS infusion activates the rat RVLM microglia, producing neuroinflammation and oxidative stress in rats, and neurogenic hypertension (Wu et al., 2012). In the PVN, the enhanced expression of pro-inflammatory cytokines elicits hypertension, while the blockade of TNF-α or NF-κB in the PVN attenuates the Ang-II-induced hypertension (Sriramula et al., 2013).

## ROLE OF MICROGLIA ON BRAIN INFLAMMATION AND CIH

Activation of microglia, the brain resident macrophages, played a critical role in neuroplasticity and neuroinflammation (Shi et al., 2010; Bhalala et al., 2014; Kawabori and Yenari, 2015). Although ROS are essential for microglial inflammatory responses (Tschopp and Schroder, 2010), the specific involvement of microglia in CIH-induced neuroinflammation and hypertension is not completely known. Smith et al. (2013) found that 14 days of CIH increases the microglia mRNA expression of IL-1β, IL-6, COX-2 and the innate immune receptor TLR4 in the rat brainstem. Recently, Stokes et al. (2017), studied the role played by glial cells in the rat ventilatory acclimatization to sustained hypoxia. Using minocycline, an inhibitor of microglia activation with anti-inflammatory properties, they blocked both the microglial and astrocyte activation in the NTS and the ventilatory acclimatization of rats submitted to chronic hypoxia. One plausible mediator of the effects of CIH is Ang II, which induces microglial activation in the PVN and hypertension (Paton et al., 2008; Zhang et al., 2010). It is known that Ang II and pro-inflammatory cytokines molecules participates in the communication between neurons and glial cells (Kang et al., 2009).

# ROLE OF THE CIRCUMVENTRICULAR ORGANS IN THE HYPERTENSION INDUCED BY CIH

Circulating Ang II cannot effectively activate AT1 receptors (AT1R) in the NTS and RVLM of healthy subjects, because these receptors are protected by the brain barrier, but Ang II may access the brain through the circumventricular organs (CVOs), regions with weak brain barrier and a high density of AT1R (Banks and Erickson, 2010). Saxena et al. (2015) studied the contribution of Ang II on the sustained BP increase and FosB activation in the median preoptic and the PVN in rats with AT1R knockdown in the SFO. They found that CIH increased BP during the hypoxic exposure in both control and AT1R-knockdown rats. However,

increasing local levels of ROS, Ang II, and pro-inflammatory cytokines.

during the normoxic dark phase, only the controls showed a sustained BP elevation. AT1R-knockdown rats showed a decrease in the FosB mark in the median preoptic nucleus and the PVN. In addition, Kim et al. (2018) found that Ang II may act at the CB level. They found that the acute intermittent hypoxia-induced renal sympathetic overflow (RSO) was prevented by losartan. The CBs denervation and the pharmacological inhibition of the SFO produced a partial reduction of RSO, while combined CB denervation and SFO inhibition eliminated the increased sympathetic overactivity following intermittent hypoxia. Thus, the evidence suggests that SFO mediates the effects of elevated circulating levels of Ang II. Proinflammatory cytokines plays a key role in hypertension, but these molecules do not permeate the blood–brain barrier. Thus, it has been proposed that the CVOs mediate the hypertensive effects of circulating pro-inflammatory cytokines. Wei et al. (2013) found that the increased BP and RSA elicited by the intracarotid injection of TNF-α and IL-1β was attenuated in SFO-lesioned rats. They found that the increased BP and RSO induced by injections of TNF-α or IL-1β into the rat SFO were attenuated by microinjections of losartan and captopril in the SFO (Wei et al., 2015). More recently, Wei et al. (2018) found that the intravenous injection of IL-1β increased mRNA levels of the angiotensin-converting enzyme, AT1R, TNF-α, and IL-1β in the SFO and the PVN. Pretreatment with microinjections of losartan and captopril in the SFO attenuated the expression of these excitatory mediators in the SFO and in the PVN. These results show that proinflammatory cytokines increase renin–angiotensin activity and produce local inflammation in the SFO and PVN.

### PROPOSED MODEL FOR CIH ACTIVATION OF THE CB CHEMOREFLEX PATHWAY

The available evidence indicates that the initial phase of the CIHinduced hypertension relays on the enhanced CB chemosensory

### REFERENCES


drive, which triggers activation of neurons in the NTS and RVLM leading to hypertension (**Figure 1**). The CB is sensible to oxidative stress that contributes to potentiates the chemosensory discharge. As result of the activation of the neural pathway or a direct effect of CIH, the oxidative stress and proinflammatory molecules levels increase in the NTS and RVLM and contribute to the maintenance of hypertension. In addition, it is likely that circulating pro-inflammatory molecules and Ang II levels may enter the central nervous in the SFO and the AP (Simpson, 1981). The increased central activity may enhance the production of both ROS and pro-inflammatory cytokines in the NTS, which may induce microglial activation (Hirooka et al., 2010). At the same time, microglial activation may increase the neuronal expression of NF-ββ, which increase the production of pro-inflammatory cytokines (Shi et al., 2010). In late phases of CIH, the inflammatory state may contribute to increase the sympathetic activity leading to the production of more pro-inflammatory molecules (Fernandez et al., 2014). This positive feedback should result in a hyperactivation of RVLM and PVN neurons. In the NTS and RVLM, a positive feedback between ROS and Ang II may increase the activity of glutamatergic neurons that increase the excitatory sympathetic output to the kidneys, blood vessels, heart and adrenal gland, eliciting a sustained increase in the BP (Crowley, 2014).

### AUTHOR CONTRIBUTIONS

MO and RI contributed equally to the manuscript and approved the final version.

### FUNDING

This work was supported by grant 1150040 from the National Fund for Scientific and Technological Development of Chile (FONDECYT).



by hypoxia-inducible factors 1 and 2. Physiol. Rev. 92, 967–1003. doi: 10.1152/ physrev.00030.2011


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Oyarce and Iturriaga. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Chronic Intermittent Hypoxia-Induced Vascular Dysfunction in Rats is Reverted by N-Acetylcysteine Supplementation and Arginase Inhibition

Bernardo J. Krause<sup>1</sup> \* † , Paola Casanello1,2, Ana C. Dias <sup>3</sup> , Paulina Arias <sup>3</sup> , Victoria Velarde<sup>3</sup> , German A. Arenas <sup>3</sup> , Marcelo D. Preite<sup>4</sup> and Rodrigo Iturriaga3†

<sup>1</sup> Division of Pediatrics, Faculty of Medicine, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>2</sup> Division of Obstetrics & Gynecology, Faculty of Medicine, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>3</sup> Laboratorio de Neurobiología, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>4</sup> Departamento de Química Orgánica, Facultad de Química, Pontificia Universidad Católica de Chile, Santiago, Chile

### Edited by:

Joaquin Garcia-Estañ, Universidad de Murcia, Spain

### Reviewed by:

Rosemary Wangensteen, Universidad de Jaén, Spain James Todd Pearson, National Cerebral and Cardiovascular Center, Japan

> \*Correspondence: Bernardo J. Krause bjkrause@uc.cl

†These authors have contributed equally to this work.

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 27 February 2018 Accepted: 21 June 2018 Published: 24 July 2018

### Citation:

Krause BJ, Casanello P, Dias AC, Arias P, Velarde V, Arenas GA, Preite MD and Iturriaga R (2018) Chronic Intermittent Hypoxia-Induced Vascular Dysfunction in Rats is Reverted by N-Acetylcysteine Supplementation and Arginase Inhibition. Front. Physiol. 9:901. doi: 10.3389/fphys.2018.00901 Chronic intermittent hypoxia (CIH), the main attribute of obstructive sleep apnea (OSA), produces oxidative stress, endothelial dysfunction, and hypertension. Nitric oxide (NO) plays a critical role in controlling the vasomotor tone. The NO level depends on the L-arginine level, which can be reduced by arginase enzymatic activity, and its reaction with the superoxide radical to produce peroxynitrite. Accordingly, we hypothesized whether a combination of an arginase inhibitor and an antioxidant may restore the endothelial function and reduced arterial blood pressure (BP) in CIH-induced hypertensive rats. Male Sprague-Dawley rats 200 g were exposed either to CIH (5% O2, 12 times/h 8 h/day) or sham condition for 35 days. BP was continuously measured by radio-telemetry in conscious animals. After 14 days, rats were treated with 2(S)-amino-6-boronohexanoic acid (ABH 400 µg/kg day, osmotic pump), N-acetylcysteine (NAC 100 mg/kg day, drinking water), or the combination of both drugs until day 35. At the end of the experiments, external carotid and femoral arteries were isolated to determine vasoactive contractile responses induced by KCL and acetylcholine (ACh) with wire-myography. CIH-induced hypertension (∼8 mmHg) was reverted by ABH, NAC, and ABH/NAC administration. Carotid arteries from CIH-treated rats showed higher contraction induced by KCl (3.4 ± 0.4 vs. 2.4 ± 0.2 N/m<sup>2</sup> ) and diminished vasorelaxation elicits by ACh compared to sham rats (12.8 ± 1.5 vs. 30.5 ± 4.6%). ABH reverted the increased contraction (2.5 ± 0.2 N/m<sup>2</sup> ) and the reduced vasorelaxation induced by ACh in carotid arteries from CIH-rats (38.1 ± 4.9%). However, NAC failed to revert the enhanced vasocontraction (3.9 ± 0.6 N/m<sup>2</sup> ) induced by KCl and the diminished ACh-induced vasorelaxation in carotid arteries (10.7 ± 0.8%). Femoral arteries from CIH rats showed an increased contractile response, an effect partially reverted by ABH, but completely reverted by NAC and ABH/NAC. The impaired endothelial-dependent relaxation in

**268**

femoral arteries from CIH rats was reverted by ABH and ABH/NAC. In addition, ABH/NAC at high doses had no effect on liver and kidney gross morphology and biochemical parameters. Thus, although ABH, and NAC alone and the combination of ABH/NAC were able to normalize the elevated BP, only the combined treatment of ABH/NAC normalized the vascular reactivity and the systemic oxidative stress in CIH-treated rats.

Keywords: arginase, chronic intermittent hypoxia, endothelial dysfunction, nitric oxide, oxidative stress, vascular reactivity

### INTRODUCTION

Obstructive sleep apnea (OSA), a growing breathing disorder featured by cyclic episodes of partial or total airflow occlusions during sleep, is considered an independent risk factor for develop systemic hypertension and is linked with stroke, coronary artery disease, and pulmonary hypertension (McNicholas et al., 2007; Dempsey et al., 2010). The airflow occlusion produces hypoxia and hypercapnia, which in turn stimulates the carotid body (CB) causing sympathetic, hypertensive, and ventilatory responses. Among these alterations, chronic intermittent hypoxia (CIH) is considered the main factor to develop systemic hypertension (Lavie, 2003; Kheirandish-Gozal and Gozal, 2008; Dempsey et al., 2010). Pre-clinical models of rodents exposed to CIH, which mimics most of the pathologic features of OSA including hypoxemia and hypertension, are used to study the mechanisms involved in cardiovascular and autonomic alterations induced by OSA (Fletcher, 2000; Del Rio et al., 2010; Prabhakar and Kumar, 2010; Iturriaga et al., 2017). Although the link between OSA and hypertension is well established, the mechanisms underlying the onset and progression of the arterial blood pressure (BP) elevation are not wellknown. It has been proposed that CIH produces oxidative stress, inflammation, and sympathetic overflow, endothelial dysfunction, and hypertension (Lévy et al., 2008; Garvey et al., 2009; Iturriaga et al., 2009). In addition, new evidences suggest that the CB is involved in generation of autonomic and cardiovascular and ventilatory alterations elicited by CIH (Iturriaga et al., 2014, 2017; Prabhakar et al., 2015). The cycles of hypoxia-reoxygenation produce oxidative stress in the CB and enhance its chemosensory responsiveness to hypoxia. The enhanced CB chemosensory drive leads to sympathetic hyperactivation of the sympatho-adrenal axis and the reninangiotensin system (Fletcher, 2000; Iturriaga et al., 2009; Prabhakar et al., 2012).

OSA patients show endothelial dysfunction with reduced vasodilatation to ACh, and vascular remodeling characterized by increased intima-media thickness (Ip et al., 2004; Patt et al., 2010) Similarly, CIH diminish the vasodilatation induced by ACh in rats (Tahawi et al., 2001; Dopp et al., 2011). CIH produces structural changes in the rat skeletal muscle resistance arteries in the first 14 days of CIH (Philippi et al., 2010). We previously found evidence that endothelial dysfunction in CIHinduced hypertension, may result from an imbalance in the ratio of arginase-1 to eNOS expression, vascular remodeling and increased contractile capacity (Krause et al., 2015). Indeed, we found that ex vivo acute arginase inhibition in carotid arteries of CIH-treated rats reverted the impaired ACh-induced relaxation, an effect completely blocked by the NO-synthase inhibitors NGnitro-L-arginine (L-NA). In addition, we found that arginase-1 protein level was increased, whereas eNOS levels decreased in CIH-derived arteries (Krause et al., 2015). Thus, it is plausible that the reduction of the oxidative stress and inhibition of the arginase enzymatic activity and may revert the vascular dysfunction and hypertension associated with CIH. It is wellknown that NO levels play a critical role in vasomotor regulation, depending on L-arginine availability, which can be reduced in conditions where high arginase expression and activity have been evidenced (Demougeot et al., 2005; Bagnost et al., 2010; Krause et al., 2012; Cowburn et al., 2016). In addition, NO may react with the superoxide radical to produce peroxynitrite (ONOO−). Accordingly, we hypothesized whether the administration of an arginase inhibitor and a precursor of the potent antioxidant glutathione, N-acetylcysteine (NAC) (Rushworth and Megson, 2014; Lasram et al., 2015; Schmitt et al., 2015), may reduce the endothelial dysfunction and hypertension induced by CIH. Thus, we assessed the effects of the arginase inhibitor 2(S)-amino-6-boronohexanoic (ABH) and the antioxidant NAC on the elevated BP and endothelial dysfunction in carotid and femoral arteries, from CIH-induced hypertensive rats. Furthermore, since arginase inhibition could interfere with the urea pathway, the effects of high doses of ABH-NAC on renal and hepatic function on rats were assessed. In a separate series, we studied the effect of a large dose of ABH of 400 µg/kg day and NAC 400 mg /kg day on renal and hepatic function and histology in rats. Proteins, creatinine, and urea were determined in urine and creatinine, urea, glutamic-oxalacetic transaminase (GOT), glutamic-pyruvic transaminase (GPT), and lactate dehydrogenase (LDH) were measured in plasma.

### MATERIALS AND METHODS

### Animal Care and Ethics Approval

This study was carried out in accordance with the recommendation of the Guide for the Care and Use of Laboratory Animals of the Bioethics Committee, CONICYT Chile. The experimental protocol for the animal was approved by the Bioethics Committee of the Faculty of Biological Sciences of the Pontificia Universidad Católica de Chile. For human samples the research was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee from the Faculty of Medicine of the Pontificia Universidad Católica de Chile (Protocol #11-247) as well as patient informed consent was obtained. Placentae were collected immediately after delivery from full-term normal normotensive, non-alcohol non-smoking, or drug consuming mothers, without any other obstetrical or medical problem.

# Animals and Exposure to Chronic Intermittent Hypoxia

Experiments were done on male Sprague-Dawley rats, weighing initially ∼200 g. Rats were fed with standard rat chow diet ad libitum, and kept on a 12-h light/dark schedule (8:00 a.m.−8:00 p.m.). Rats were exposed to CIH for 35 days, as previously described (Iturriaga et al., 2010; Del Rio et al., 2016). Briefly, unrestrained, freely moving rats were housed in individual chambers. The CIH protocol consisted of hypoxic cycles of 5–6% inspired O<sup>2</sup> for 20 s, followed by room air for 280 s, applied 12 times/h, 8 h/day. The chambers had a rear N<sup>2</sup> inlet and a front air extractor, which enables the recovery to normoxia. A computerized system controlled the valve inlets and the alternating cycles of the extractors. During hypoxia, the extractors stopped for 30 s, and the rear solenoid valves allowed 100% N<sup>2</sup> flow into the chambers. The O<sup>2</sup> level in the chambers was continuously monitored with an oxygen analyzer (Teledyne AX 380, USA). The CO<sup>2</sup> in the chamber was maintained low by continuous air extraction. In the Sham condition, rats were exposed to compressed air flow into chambers.

# Arterial Blood Pressure Recordings

Arterial blood pressure (BP) was measured in conscious rats using radio-telemetry. Seven days before the beginning of the experiments, rats were anesthetized with 5% isoflurane and maintained with 2% isoflurane in 100% O<sup>2</sup> during the surgical procedure. The tip of a cannula-coupled telemetry device (PC40A-40, Data Science International, USA) was introduced into the femoral artery (Del Rio et al., 2016). After surgery rats received a subcutaneous injection of ketoprofen (1%) and enrofloxacin (1%). The BP recording was started after 5 days of recovery. The raw data from each rat was continuously collected at a sample rate of 1 KHz, and average every 24 h with the ART Dataquest Platform (Data Science International, USA).

### Synthesis and Functional Studies of ABH

The 2(S)-amino-6-boronohexanoic acid (ABH) was synthesized by an optimization of a method previously reported by Vadon-Legoff et al. (2005), which was itself based in the previous work of Collet et al. (2000). The synthesis of ABH consisted of the preparation of the Gly-Ni-BPB complex, which was obtained by reaction of glycine (Gly), nickel(II) chloride, and 2[N-(N′ benzylprolyl)amino]benzophenone (BPB) in the presence of excess KOH in methanol. The Gly-Ni-BPB complex was treated with a base at low temperature and enantioselectively alkylated with the corresponding catechol-protected boronated bromide. The alkylated complex was then hydrolyzed with aqueous acid and purified by ion-exchange chromatography. The effectiveness of synthesized ABH was determined by measuring its ability to inhibit arginase activity, as described previously for endothelial cells (Krause et al., 2012). Briefly, basal arginase activity was determined in the presence of the synthesized inhibitor ABH and, compared with the effect of the commercial inhibitor of arginase, S-(2-boronoethyl)-L-cysteine (BEC. Sigma, USA) in protein extracts obtained from total human placental homogenates (i.e., vascular, syncytium and connective tissue) obtained with lysis buffer [1 mmol/l pepstatine A, 1 mmol/l leupeptine, 200 mmol/l phenylmethylsulfonyl fluoride (PMSF), 50 mmol/l Tris-HCl (pH 7.5), 0.2% Triton X-100] sonicated (20 pulses, 150 W, 3 min). The production of urea was measured using 70µg of placental homogenate protein in 100 µl of reaction incubated at 37◦C for 1 h in the presence of L-arginine (50 mM) and the cofactor Mn2+ (5 Mm), and the presence or absence of BEC (100µM) or, increasing concentrations of synthesized ABH (10–1,000µM). The catalysis was stopped by adding 4 volumes of acid solution (H2SO4: H3PO4: H2O = 1: 3: 7), and the urea formed was determined by adding 25 µl of α-isonitrosopropiophenone (9% ISPF in absolute ethanol) to the assay and heating the mixture (100◦C for 45 min). The concentration of urea was measured by spectrophotometry (OD 450 nm) in an automatic plate reader based on results obtained from a urea calibration curve. We found that synthesized ABH produced a concentrationdependent inhibition of the arginase activity and that it reached levels comparable to those observed in the presence of the commercial inhibitor BEC. The Ki for ABH was 4.78 ± 1.76µM and the solubility in water 125 g/L or 0.7 mol/L.

# Experimental Procedure and Drug Administration

The antioxidant N-acetylcysteine (NAC, Sigma USA) was administered through drinking tap water (400 mg/kg day) from the day 14 of CIH exposure until the end of the experimental protocol. The NAC solution was prepared daily and preserved in dark condition to avoid oxidation. ABH was administrated with osmotic pumps (2ML4, Alzet Scientific Products, USA). Rats were anesthetized with isoflurane in O2, and the osmotic pumps were implanted subcutaneously on the back. Pumps deliver ABH in saline solution at a rate of 400µg/kg/ day. Animals were implanted with the same osmotic pumps (Alzet Scientific Products, USA) filled with saline solution. Rats were out of the hypoxic chambers for 2–3 h for the surgical implant of the osmotic pumps. After surgical procedures, rats were treated with enrofloxacin and ketoprofen as mentioned before.

### Wire Myography

At the end of the experiments, external carotid and femoral arteries were surgically removed from rats anesthetized with sodium pentobarbitone (50 mg/kg ip) and placed in ice-cold PBS. Vessel segments (∼2 mm) of external carotid arteries and second/third-order femoral arteries were mounted on a wire-myograph (model 610A; Danish Myo Technology A/S, Denmark) and contractile responses were studied as previously described (Krause et al., 2015). Animals were euthanized with a higher dose of sodium pentobarbitone (200 mg/kg ip). The internal diameter of vessels was defined by determining the stretch condition at which the maximal contractile response to KCl was obtained (Delaey et al., 2002). Indeed, the internal diameter of the vessels for wire myography was established by determining the opening length (or stretching condition)

at which the vessel presents its maximal contractile response (Mulvany and Aalkjaer, 1990). This ex vivo method has been shown to accurately represent the in vivo internal arterial perimeter in different models (Mulvany and Aalkjaer, 1990; Delaey et al., 2002). Likewise, through this methodology, a direct correlation of the ex vivo contractile response with the biomechanical and structural properties of different blood vessels has been observed (Cañas et al., 2017). Isometric force in response to cumulative concentrations of KCl (16–125 mM) was assayed to determine the maximal contractile force. Similarly, isometric force in response to acetylcholine (ACh, 10−<sup>8</sup> -10−<sup>5</sup> mol/L) in pre-constricted vessels with 37.5 mmol/L KCl were determined. Responses were expressed as percentage of the maximal contractile effect induced by KCl at 40.8 mmol/L (%Kmax). Concentration-response curves were fit with the GraphPad Prism version 5.00 (CA, USA) to obtain the maximal effect and potency (pD<sup>2</sup> = −log EC50).

### Measurement of Systemic Oxidative Stress

Plasmatic oxidative stress was measured with the TBARS assay (Cat N◦ 10009055, Cayman, USA) according to the protocol provided by the supplier. The concentration of thiobarbituric acid-reactive species were expressed as malondialdehyde (MDA) µmol/L. Blood samples were collected from the common carotid artery and placed in heparinized ice-cold microcentrifuge tubes after 1–2 h of the last intermittent hypoxic cycle. Plasma was separated by centrifugation and stored at −80◦C.

### Biosafety of NAC and ABH

Since arginase inhibition may affect the urea pathway, we evaluated whether high doses of ABH and NAC could have a hepatotoxic or nephrotoxic effect. In a separate series, two independent groups were used: one group without treatment (n = 4) and one treated with the combination of ABH and NAC (n = 8). Under isoflurane anesthesia, osmotic pumps were implanted subcutaneously (2ML4, Alzet Scientific Products USA). In the untreated animals, pumps were filled with saline solution. In the rats treated with ABH/NAC, the same osmotic pumps with ABH dissolved in saline solution were installed. Each rat received a dose of ABH of 400 µg/kg day and NAC 400 mg /kg day. Rats were kept for 28 days, and placed in metabolic cages to collect urine. Blood was collected from the ocular orbital sinus under isoflurane 5% in O<sup>2</sup> anesthesia. Proteins, creatinine, and urea were determined in urine and creatinine, urea, lactate dehydrogenase (LDH), glutamic-pyruvic transaminase (GPT), and glutamic-oxalacetic transaminase (GOT) were measured in plasma.

In a separate experimental series of rats (n = 6) without treatment and a group of 6 rats treated with the combination ABH/NAC, creatinine clearance was measured. The collected urine was measured to determine the 24 h volume and then centrifuged at 3,000 rpm for 10 min at 4◦C to discard any precipitate. Under anesthesia with isoflurane, blood was obtained from the left ventricle. Blood was collected in heparinized tubes, centrifuged at 3,000 rpm to obtain plasma. Creatinine was measured in both plasma and urine using the Creatinine Liquicolor kit from human (Wiesbaden, Germany), which evaluates the kinetics of the reaction. Briefly, plasma and urine samples (diluted 1:50) were mixed with a solution of picric acid and alkaline buffer (NaOH/bicarbonate). Immediately the samples were transferred to a reading cell and their absorbance was measured at 492 nm at 30 s and at 2 min after the start of the reaction. With these values and using a creatinine standard, the concentration of creatinine present in each sample was determined.

### Histological Staining and Examination

Anesthetized rats were perfused intracardially with phosphate buffer saline (PBS; pH 7.4) for 10 min, followed by buffered paraformaldehyde (PFA 4%, Sigma, St Louis, USA) for 10 min. The saline and PFA solutions were perfused at constant pressure of 95 mmHg. Pieces of the carotid external artery (3–4 mm length) were collected 2 mm rostral from the carotid sinus and post-fixed by immersion in buffered-PFA 4% for 12 h at 4◦C. Carotid arteries were dehydrated in graded ethanol solutions followed by xylol, included in paraffin, sectioned at 5µm and mounted on silanized slides. The vessels were stained with hematoxylin and eosin and photomicrographs were taken with an Olympus CX 31 microscope with a CCD camera (Olympus Corp, Japan). The internal diameter (ID) was measured from fixed tissues with the ImageJ software (NIH, USA).

Liver and kidney samples obtained from euthanized rats were fixed with 4% paraformaldehyde (Sigma, St Louis, USA), dehydrated and paraffin-embedded. Transverse sections (5µM) were stained with Harris hematoxylin (5 min) and eosin (30 s) and mounted with Entellan (Merck, Whitehouse Station, NJ, USA). Microphotographs were obtained at 40x and 100x with and Olympus CX31 microscope (Olympus Corporation, Tokyo, Japan).

### Statistical Data Analysis

Data were expressed as mean ± SEM. For BP recordings, the average daily MABP for each animal were grouped and compared through one-way ANOVA followed by Dunnet' multiple comparison test among the treatment groups. Comparisons between two groups were performed with the U Mann-Whitney test, and differences between more groups were assessed with one or two-way ANOVA tests, followed by Dunnet post-hoc comparisons. Significance was accepted to p < 0.05.

# RESULTS

# Effects of ABH and NAC on the Arterial Blood Pressure in CIH-Treated Rats

Rats submitted to CIH showed a sustained increase in mean arterial blood pressure (MABP) of ∼8 mmHg after 3–4 days of exposure (**Figure 1** and **Table 1**). To demonstrate the role of arginase activity and oxidative stress in this increased arterial BP, rats exposed to CIH were treated with the arginase inhibitor ABH and/or NAC from day 14 of CIH. ABH and NAC treatment progressively reduced MABP in CIH-rats reaching basal values (**Figures 1B,C**). Similarly, simultaneous administration of ABH and NAC restored the normal MABP levels in CIH-treated rats (**Figure 1D**).


combination of both ABH/NAC (D, n = 8).


days rats were maintained in CIH without treatment (A, n = 8) or receiving the arginase inhibitor ABH (B, n = 8), the glutathione precursor NAC (C, n = 6), or the

Mean ± SEM. \*P < 0.05 vs. baseline. One-way ANOVA followed by Dunnet (n = 6–8 rats per group).

### Effect of ABH and NAC on the Contractile Responses in Arteries From CIH-Treated Rats

To determine the effects of ABH and NAC on vascular reactivity in CIH-rats, the vasoactive response to KCl of carotid and femoral artery segments were assayed using wire myography. Carotid arteries from CIH-rats showed a significant decreased internal diameter measured in histological sections of fixed tissues, whilst the treatment with ABH, NAC and their combination ABH/NAC prevented the reduction in internal diameter (**Figure 2A**). Compared with sham, maximal KClinduced constriction was increased in CIH and CIH NAC groups, while this effect was reverted by ABH, as well as the combination of ABH/NAC (**Figure 2B**). There were no significant changes in the internal diameter of femoral arteries in the groups studied compared with sham rats (**Figure 2C**). However, maximal KCl-induced constriction was increased in CIH. The latter effect was partially reverted by ABH treatment and completely prevented by NAC, as well as ABH/NAC (**Figure 2D**).

# Effect of ABH and NAC on the NOS-Dependent Relaxation in Arteries From CIH-Treated Rats

Carotid arteries from rats exposed to CIH showed a decrease in the concentration-dependent relaxation (**Figure 3A**) and maximal response (**Figure 3B**) to ACh compared to the sham group, and this effect was reverted by the combination of ABH/NAC. Similarly, treatment with NAC did not improve the response to ACh in CIH-rats, whilst ABH treated rats showed a significant increase in the maximal response compared to sham. Sensitivity (pD2) to ACh was comparable between sham, CIH and CIH- NAC groups, but substantially increased (>1 Log unit) in CIH rats treated with ABH and ABH in combination with NAC (**Figure 3C**). Femoral arteries from rats exposed to CIH showed a decrease in the concentration-dependent relaxation (**Figure 4A**) and maximal response (**Figure 4B**) to ACh related to sham, effect reverted by NAC and ABH/NAC. CIH-ABH-treated rats showed a ∼2-fold increase in the maximal response to ACh relative to the sham. Conversely, sensitivity to ACh was comparable between sham, CIH,

FIGURE 2 | Effects of ABH and NAC on the CIH-induced vascular remodeling. Internal diameter measured by histology in fixed tissues (A,C) and maximal contractile response to KCl (B,D) of carotid (A,B) and femoral (C,D) arteries from sham (open bars, n = 10), CIH (solid bars, n = 10), CIH treated with ABH (light gray bars, n = 5), CIH treated with NAC (gray bars, n = 5), and CIH treated with ABH, and NAC (dark gray bars, n = 5) rats. Values expressed as mean ± SEM, \*p < 0.05, \*\*p < 0.01, \*\*\*p < 0.001 vs. sham, ANOVA.

arteries from sham (open circles, n = 6), CIH (solid circles, n = 6), CIH treated with ABH (light gray circles, n = 6), CIH treated with NAC (gray circles, n = 6) and CIH treated with ABH, and NAC (dark gray square, n = 6) rats. Maximal acetylcholine-induced response (B) and pharmacological potency (pD2, i.e., sensitivity) (C) from sham (open bars), CIH (solid bars), CIH treated with ABH (light gray bars), CIH treated with NAC (gray bars), and CIH treated with ABH, and NAC (dark gray bars) rats were derived from (A). Values expressed as mean ± SEM, \*p < 0.05, \*\*p < 0.01, \*\*\*p < 0.001 vs. sham, ANOVA.

CIH-ABH, and CIH-NAC groups, but substantially increased in CIH-rats treated with ABH in combination with NAC (**Figure 4C**).

### Effects of ABH and NAC on Systemic Oxidative Stress in CIH-Rats

Systemic oxidative stress in CIH-treated rats was evaluated by determining plasma levels of MDA as an index of lipids peroxidation after 35 days of CIH or sham condition. CIHtreated rats showed a substantial increase (∼2-fold) in MDA levels compared with the sham group, an effect also observed in CIH-rats treated with ABH (**Figure 5**). On the contrary, CIH-rats treated with NAC, as well as with ABH/NAC, showed plasma levels of TBARS comparable to sham rats (**Figure 5**).

### Biosafety Testing of ABH-NAC Combination

Considering the effect of ABH/NAC combination on normalizing the BP and vascular reactivity in CIH rats, liver and kidney morphology as well as the renal function was evaluated in sham rats treated for 14 days with high doses of ABH (400 µg/kg day) and NAC (400 mg/kg day). Kidney (**Figure 6A**) and liver (**Figure 6B**) morphology were not altered by the treatment. Similarly, we did not found changes in plasma urea concentration and creatinine clearance

FIGURE 4 | Endothelial-dependent relaxation in femoral arteries from CIH-rats. (A) Concentration-dependent relaxation curves in response to acetylcholine of femoral arteries from sham (open circles, n = 6), CIH (solid circles, n = 6), CIH treated with ABH (light gray circles, n = 6), CIH treated with NAC (gray circles, n = 6) and CIH treated with ABH, and NAC (dark gray square, n = 6) rats. Maximal acetylcholine-induced response (B) and pharmacological potency (pD2, i.e., sensitivity) (C) from sham (open bars), CIH (solid bars), CIH treated with ABH (light gray bars), CIH treated with NAC (gray bars), and CIH treated with ABH, and NAC (dark gray bars) rats were derived from (A). Values expressed as mean ± SEM, \*p < 0.05, \*\*p < 0.01, \*\*\*p < 0.001 vs. sham, ANOVA.

between the rats with or without ABH/NAC treatment (**Table 2**).

### DISCUSSION

This study aimed to determine whether the administration of the arginase inhibitor ABH and the glutathione precursor NAC may revert the hypertension and endothelial dysfunction in rats exposed to CIH. The results showed that ABH and NAC, as well as the combination of both drugs, reverted the increase in BP induced by CIH. However, only the combination of ABH and NAC leads to a normalization of BP and endothelial function along with a reversion of

vascular remodeling markers, as suggests the changes in contractile force and internal diameter. In addition, the combination of ABH and NAC at high doses had no


\*P < 0.05 U-Mann-Whitney treated vs. non-treated rats (Untreated Rats n = 4 and ABH-NAC rats = 8).

effect on liver and kidney morphology and biochemical function.

Compelling evidence shows that, in humans, oxidative stress has a prominent role in the cardiovascular dysfunction induced by OSA, a fact that is present also in preclinical models. In this study, similar to previous reports antioxidant treatment reverted the increase in BP in CIH rats. Indeed, Moya et al. (2016) found that rats treated with the antioxidant ebselen and exposed to CIH displayed a significant decreased of the elevated BP, suggesting that CIH-induced hypertension is critically dependent on oxidative stress. In this study, to understand the cardiovascular effects of antioxidants in CIH, we assayed the ex vivo vascular function in two representative arteries (i.e., carotid and femoral). It is well established that NAC antioxidant effect occurs mainly by restoring the potent intracellular reducing agent glutathione, under oxidative stress conditions (Rushworth and Megson, 2014). Notably, in vivo NAC treatment had limited effects on the ex vivo impaired endothelial dependent-relaxation and increased contractile reactivity in carotid arteries, as well as partially reverted the endothelial dysfunction in femoral arteries from animals exposed to CIH. These data suggest that the antihypertensive effect of NAC and other antioxidants occurs partially by improving vascular reactivity and, in a higher degree, by preventing the overactivation of the CB in CIH. Indeed, previous reports have shown that oxidative stress in the CB drives the chemosensory potentiation and hypertensive effects of CIH (Del Rio et al., 2016; Iturriaga et al., 2017). However, considering that our studies on endothelial-dependent relaxation were carried-out after pre-constriction with KCl, a potential effect of NAC enhancing the role of endotheliumderived hyperpolarizing factor (EDHF) (Krummen et al., 2006) or alternative glutathione-dependent vasodilator pathways (Yang and Wang, 2015) cannot be ruled out.

Conversely, endothelial dysfunction has been proposed to play a key role in the cardiovascular risk associated with OSA and CIH. Indeed, OSA patients show a reduced flow-mediated dilation (Ip et al., 2004), and endothelial dysfunction has been directly demonstrated in middle cerebral (Phillips et al., 2004) and carotid arteries from rats exposed to CIH (Krause et al., 2015). In the present study, we add new evidence that vascular dysfunction in CIH is associated with an increased contractile response and impaired NOS-dependent relaxation in femoral arteries. This new data shows the effects of CIH on the vascular function on carotid and femoral arteries. In this context, we previously reported that the reduced relaxation in response to ACh in carotid arteries from CIH-treated rats can be prevented by the ex vivo inhibition of arginase activity, and this effect could be explained by an unbalanced endothelial expression of eNOS and arginase-1 (Krause et al., 2015). Increased arginase expression and its effect on NOS-dependent relaxation have been extensively reported in hypertension and cardiovascular diseases (Caldwell et al., 2015). For instance, a study in humans shows that plasma levels of arginase are increased in subjects with OSA that present a normal cardiovascular function compared with health controls, and these levels are further increased when OSA is associated to vascular dysfunction (Yüksel et al., 2014). In that regard, a higher arginase expression and activity has been reported in diverse models of hypoxia (i.e., intermittent and sustained) in which an increased expression of arginases 1 or 2 occurs in the endothelium leading to a decreased NO synthesis (Krotova et al., 2010; Krause et al., 2013, 2015; Singh et al., 2014; Cowburn et al., 2016). Here we found that chronic in vivo arginase inhibition with ABH reverted the increase in BP induced by CIH, with no effects on circulating markers of oxidative stress. Furthermore, the treatment with ABH potentiated the vasorelaxation response to ACh in carotid and femoral arteries, as well as partially reverted the increased contractile response to KCl. The later effect of ABH treatment on the contractile response could result from the inhibition of the proliferative stimulus by arginases-derived polyamines on vascular smooth muscle cells (Ignarro et al., 2001; Wei et al., 2001; Chen et al., 2009). Similar to our results, previous studies in spontaneously hypertensive rats have demonstrated the potential therapeutic role of arginase inhibitors decreasing arterial blood pressure and restoring endothelial function (Demougeot et al., 2005; Bagnost et al., 2008, 2010). Altogether, the available information reinforces the idea that arginase-mediated impaired endothelial NO synthesis plays a key role in the vascular dysfunction associated with CIH and OSA.

It is worth to note that neither NAC nor ABH treatments separately led to a complete normalization of ex vivo vascular responses in carotid or femoral arteries from CIH rats, despite that they had a clear effect decreasing BP, but a heterogeneous outcome on systemic oxidative stress levels. This confirms the notion that CIH-induced increase in BP (Del Rio et al., 2016; Iturriaga et al., 2017), similarly to other hypertensive models (Godo and Shimokawa, 2017; Handy and Loscalzo, 2017; Mancia et al., 2017), is a multifactorial process in which a CB chemosensory and sympathetic overactivity, oxidative stress and endothelial dysfunction are the cornerstones. Considering this idea, we aimed to determine whether the combined treatment with ABH and NAC could normalize the vascular function in CIH rats. Clearly, the combination of both drugs normalized BP and oxidative stress markers; effects accompanied by the restoration of normal ex vivo contractile and relaxing responses. Compared with ABH treatment, the NAC/ABH combination led to a lower, but normal, maximal NOS-dependent relaxation in carotid and femoral arteries. This counterintuitive finding could result from the buffering effect of NAC, as well as glutathione, on NO levels (Hu et al., 2006; Keszler et al., 2010; Kolesnik et al., 2013), that would be limiting the enhanced levels of NO as a consequence of arginase inhibition. Conversely, ABH/NAC combination reverted the increase in the KCl contractile response induced by CIH. Notably, in a recent study we found that an increased KCl-induced contractile response is directly associated with biomechanical and histological markers of vascular remodeling (Cañas et al., 2017). Altogether, this data strongly suggests that ABH/NAC combination reverted the vascular remodeling observed in rats exposed to CIH (Krause et al., 2015). The mechanisms involved in the combined effect of ABH-NAC, as well as, the heterogeneous changes induced in femoral and carotid arteries need further studies that would include the analysis of alternative vasodilator pathways involving the cysteine and glutathione metabolism (Yang and Wang, 2015). Nonetheless, considering the changes in ACh potency (i.e., sensitivity) and markers of vascular remodeling, this combined effect could result from an increased vascular bioavailability of NO.

Present results show that CIH increased BP in 3–5 days. This fast increase in BP is probably triggered by a potentiated sympathetic vasoconstrictor tone on the arterial blood vessels, that would result from the cyclic hypoxic excitation of the CB (Iturriaga et al., 2017). In agreement to this explanation, we found that external carotid arteries from rats submitted to CIH for 21 days showed moderate enhanced contractile responses to KCl and a diminished vasorelaxation to ACh (Krause et al., 2015). Recently, Del Rio et al. (2016) found that the ablation of both CBs completely reverts the increased BP and sympathetic overflow in hypertensive rats exposed to CIH for 21 days, indicating that the CIH-enhanced CB chemosensory drive mediates the onset and maintenance of neurogenic hypertension. It is known that CIH reduced the ACh-mediated vasodilation (Tahawi et al., 2001; Dopp et al., 2011; Krause et al., 2015). However, there are some reports showing normal endothelial function in hypertensive rats exposed to CIH (Julien et al., 2003; Lefebvre et al., 2006). Indeed, Lefebvre et al. (2006) found that the ACh-induced vasodilation in carotid, aortic and mesenteric vascular beds, as well as the contractile responses to noradrenaline and angiotensin II (Ang II), are similar between CIH-treated and sham rats, while the contraction induced by

### REFERENCES

Bagnost, T., Berthelot, A., Bouhaddi, M., Laurant, P., Andre, C., Guillaume, Y., et al. (2008). Treatment with the arginase inhibitor N(omega)-hydroxynor-L-arginine improves vascular function and lowers blood pressure in adult spontaneously hypertensive rat. J. Hypertens. 26, 1110–1118. doi: 10.1097/HJH.0b013e3282fcc357

exogenous application of ET-1 is augmented in CIH-treated rats. Oxidative stress is associated with the endothelial dysfunction in CIH-treated rats. Indeed, treatment of CIH-exposed rats with allopurinol improves the reduced vasodilatation induced by ACh in gracillis arteries, although neither allopurinol nor CIH affect the vessel morphometry and systemic markers of oxidative stress in rats exposed to CIH for 14 days (Dopp et al., 2011). Similarly, Philippi et al. (2010) report that CIH elicited systemic oxidative stress, although they found that CIH has an inconsistent effect on the oxidative stress marker 3-nitrotyrosine in the vascular wall. Therefore, it is plausible that NAC and ABH may affect CIHinduced hypertension acting at different levels. It is likely that NAC may reduce the enhanced CB chemosensory discharge as all tested antioxidants do (Peng et al., 2003; Del Rio et al., 2010; Moya et al., 2016), while ABH may act at the level of the arteries. In addition, it is plausible that NAC may affect the arginase activity in the blood vessels or ABH may affect the oxidative stress in the blood vessels. If these aspects are different, it may explain the heterogeneous effect of the drugs in the femoral and carotid arteries.

# PERSPECTIVES

Present results support a potential therapeutic application of a combined antihypertensive treatment with an antioxidant and an arginase inhibitor, which not only decrease BP but also normalize endothelial vascular reactivity and revert vascular remodeling, without compromising kidney and liver functions. Further studies are required to demonstrate whether this increased antihypertensive effect is limited to the drugs tested in the present study or can be applied to other combination of antioxidant and antihypertensive agents.

# AUTHOR CONTRIBUTIONS

BK, PC, and RI conceived and designed the experiments. MP synthetized ABH. BK, ACD, PA, RI, VV, and GA performed the experiments. BK, RI, GA, VV, and ACD analyzed the data. RI and BK wrote the draft paper and all authors contributed to the manuscript and approved the final version.

### FUNDING

This work was supported by FONDEF D11I1098 and FONDECYT 1150040.


with obstructive sleep apnea. Am. J. Respir. Crit. Care Med. 182, 1540–1545. doi: 10.1164/rccm.201002-0162OC


**Conflict of Interest Statement:** BK, PC, MP, and RI presented the solicitude of a patent for the pharmaceutic combination for the treatment and prevention of arterial hypertension and vascular dysfunction, number 201602951 INAPI, Chile, and PCT international protection PCT/CL2016/050062 on November 18th, 2016.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Krause, Casanello, Dias, Arias, Velarde, Arenas, Preite and Iturriaga. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Effects on Cognitive Functioning of Acute, Subacute and Repeated Exposures to High Altitude

Matiram Pun1,2, Veronica Guadagni 1,2,3,4, Kaitlyn M. Bettauer 1,2,5, Lauren L. Drogos 1,2 , Julie Aitken<sup>6</sup> , Sara E. Hartmann1,2, Michael Furian<sup>7</sup> , Lara Muralt <sup>7</sup> , Mona Lichtblau<sup>7</sup> , Patrick R. Bader <sup>7</sup> , Jean M. Rawling<sup>8</sup> , Andrea B. Protzner 2,9, Silvia Ulrich<sup>7</sup> , Konrad E. Bloch<sup>7</sup> , Barry Giesbrecht <sup>10</sup> and Marc J. Poulin1,2,3,4,11,12 \*

<sup>1</sup> Department of Physiology & Pharmacology, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>2</sup> Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>3</sup> Department of Clinical Neuroscience, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>4</sup> O'Brien Institute for Public Health, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>5</sup> Department of Psychology, Faculty of Science, University of British Columbia, Vancouver, BC, Canada, <sup>6</sup> Biomedical Engineering Graduate Program, University of Calgary, Calgary, AB, Canada, <sup>7</sup> Pulmonary Division, Sleep Disorders Centre and Pulmonary Hypertension Clinic, University Hospital Zürich, Zurich, Switzerland, <sup>8</sup> Department of Family Medicine, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>9</sup> Department of Psychology, Faculty of Arts, University of Calgary, Calgary, AB, Canada, <sup>10</sup> Department of Psychological and Brain Sciences, Institute for Collaborative Biotechnologies, University of California, Santa Barbara, Santa Barbara, CA, United States, <sup>11</sup> Libin Cardiovascular Institute of Alberta, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada, <sup>12</sup> Faculty of Kinesiology, University of Calgary, Calgary, AB, Canada

### Edited by:

Jean-Paul R. Richalet, Université Paris 13, France

### Reviewed by:

Stephane Perrey, Université de Montpellier, France Alessandro Tonacci, Istituto di Fisiologia Clinica (IFC), Italy

> \*Correspondence: Marc J. Poulin poulin@ucalgary.ca

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 29 March 2018 Accepted: 30 July 2018 Published: 21 August 2018

### Citation:

Pun M, Guadagni V, Bettauer KM, Drogos LL, Aitken J, Hartmann SE, Furian M, Muralt L, Lichtblau M, Bader PR, Rawling JM, Protzner AB, Ulrich S, Bloch KE, Giesbrecht B and Poulin MJ (2018) Effects on Cognitive Functioning of Acute, Subacute and Repeated Exposures to High Altitude. Front. Physiol. 9:1131. doi: 10.3389/fphys.2018.01131 Objective: Neurocognitive functions are affected by high altitude, however the altitude effects of acclimatization and repeated exposures are unclear. We investigated the effects of acute, subacute and repeated exposure to 5,050 m on cognition among altitude-naïve participants compared to control subjects tested at low altitude.

Methods: Twenty-one altitude-naïve individuals (25.3 ± 3.8 years, 13 females) were exposed to 5,050 m for 1 week (Cycle 1) and re-exposed after a week of rest at sea-level (Cycle 2). Baseline (BL, 520 m), acute (Day 1, HA1) and acclimatization (Day 6, HA6, 5,050 m) measurements were taken in both cycles. Seventeen control subjects (24.9 ± 2.6 years, 12 females) were tested over a similar period in Calgary, Canada (1,103 m). The Reaction Time (RTI), Attention Switching Task (AST), Rapid Visual Processing (RVP) and One Touch Stockings of Cambridge (OTS) tasks were administered and outcomes were expressed in milliseconds/frequencies. Lake Louise Score (LLS) and blood oxygen saturation (SpO2) were recorded.

Results: In both cycles, no significant changes were found with acute exposure on the AST total score, mean latency and SD. Significant changes were found upon acclimatization solely in the altitude group, with improved AST Mean Latency [HA1 (588 ± 92) vs. HA6 (526 ± 91), p < 0.001] and Latency SD [HA1 (189 ± 86) vs. HA6 (135 ± 65), p < 0.001] compared to acute exposure, in Cycle 1. No significant differences were present in the control group. When entering Acute SpO<sup>2</sup> (HA1-BL), Acclimatization SpO<sup>2</sup> (HA6-BL) and LLS score as covariates for both cycles, the effects of acclimatization on AST outcomes disappeared indicating that the changes were partially explained by SpO<sup>2</sup>

and LLS. The changes in AST Mean Latency [1BL (−61.2 ± 70.2) vs. 1HA6 (−28.0 ± 58), p = 0.005] and the changes in Latency SD [1BL (−28.4 ± 41.2) vs. 1HA6 (−0.2235 ± 34.8), p = 0.007] across the two cycles were smaller with acclimatization. However, the percent changes did not differ between cycles. These results indicate independent effects of altitude across repeated exposures.

Conclusions: Selective and sustained attention are impaired at altitude and improves with acclimatization.The observed changes are associated, in part, with AMS score and SpO2. The gains in cognition with acclimatization during a first exposure are not carried over to repeated exposures.

Keywords: altitude, cognition, hypoxia, brain, CANTAB, AMS/LLS, SpO2 , ALMA

### INTRODUCTION

Mountains occupy one fifth of the earth's surface and are popular destinations for a variety of activities such as trekking, climbing, pilgrimages, mining, scientific experiments and celestial observations. Further, more than 140 million people worldwide live at altitudes over >2,500 m, (Penaloza and Arias-Stella, 2007) and many high-altitude dwellers sojourn at lower altitudes. The barometric pressure decreases exponentially with altitude gain and this hypobaric hypoxia leads to reduced inspired partial pressure of oxygen (West, 1996). Unacclimatized lowlanders may suffer from cerebral symptoms such as headache, nausea, vomiting and impaired coordination when exposed to high altitudes (>2,500 m) (Hackett and Roach, 2001; Bärtsch and Swenson, 2013). The brain, particularly the hippocampus and other areas within the limbic system, is very sensitive and vulnerable to hypoxia (Hornbein, 2001; Virués-Ortega et al., 2004; Wilson et al., 2009).

High altitude exposure has a detrimental effect on cognitive functions with slower reaction times and reduced psychomotor vigilance i.e., slower reaction times as a measure of reduced sustained attention (high altitude, 1,500**–**3,500 m); impaired learning, spatial and working memory (very high altitude, 3,500– 5,500 m) and impaired memory retrieval (extreme altitude, >5,500 m) (Virués-Ortega et al., 2004; Wilson et al., 2009; Yan, 2014; Taylor et al., 2016; Bickler et al., 2017; McMorris et al., 2017). The effects of hypoxia on cognitive functions have been previously explored among climbers (Kramer et al., 1993), trekkers (Dykiert et al., 2010), military personnel (Shukitt-Hale et al., 1998), flight crews (Nation et al., 2017), and high altitude residents (Virues-Ortega et al., 2011; Wang et al., 2014). Kramer et al. for example, report impairments in learning and memory processes especially when individuals were required to learn new skills while executing perceptual and semantic memory tasks. Similarly, Shukitt-Hale et al. report deterioration of both mood and performance in military personnel exposed to high-altitude. Futher, high altitude exposure has been shown to increase reaction times and impair memory encoding and retention (McMorris et al., 2017; Nation et al., 2017). However, Dykiert et al. suggest more pronounced changes in mean reaction times only above 4,000 m. Different types of hypoxic exposures such as field (Subudhi et al., 2014; Davranche et al., 2016), simulated hypobaric hypoxia (Hornbein et al., 1989; Asmaro et al., 2013; Malle et al., 2013), normobaric hypoxia (Turner et al., 2015) and intermittent hypoxia (Champod et al., 2013) have been investigated. Davranche et al. showed impaired information processing (speed and accuracy) at high altitude while Hornbein et al. reported impaired visual longterm memory during chamber simulation. With very high altitude exposure (5,260 m), Subudhi et al. found impairments in reaction times and in performance on the code substitution tasks (simultaneous and match to sample) which then improved with acclimatization. In the study by Malle et al., an increased rate of error frequency and worsened working memory were reported while Asmaro et al. observed impairments in cognitive flexibility and attention, short-term and working memory and executive functions. Similarly, Turner et al found that acute normobaric hypoxia affected memory, attention and executive functions. Although the aforementioned studies differ in types of hypoxic exposure, duration, modality and severity, the reported neurocognitive impairments are consistent across studies (Virués-Ortega et al., 2004; McMorris et al., 2017). Regardless, the significant differences across study design and the inadequately powered sample sizes limit the currently available studies and futher research is therefore needed.

A large number of high altitude workers, such as the ones involved in the large mining industry in the Chilean mountains or scientists, engineers and staff at the Atacama Large Millimeter/submillimeter Array (ALMA) scientific observatory, travel periodically to high altitudes for work. The workers at ALMA are periodically exposed to "very high altitude" (i.e., 5,050 m) for an entire week [including sleep periods at "high altitude" (i.e., 2,950 m)] followed by a week of rest at near-sea level (i.e., 520 m). Hypoxia associated with high altitude may impair cognitive performance and therefore it may lead to higher rates of errors (Hornbein et al., 1989; Davranche et al., 2016) and elevated risks of occupational injuries while performing their duties as seen among high altitude miners (Vearrier and Greenberg, 2011). However, there is no study examining the effects of this unique ascent profile and work schedule at very high altitude on cognitive functioning.

Hence, here we investigate the effects of acute, acclimatization, and repeated exposure to very high altitude on cognitive functions in altitude-naïve individuals bringing them to ALMA (5,050 m) with the same schedule that the workers would follow over a month. We hypothesize that acute exposure to high altitude will result in slower reaction times, decreased attention and reduced executive functions (reduced flexibility and ability to shift, greater fixation, reduced executive control and planning ability) which will then be restored with acclimatization. Further, we hypothesize that the positive changes in cognitive function due to acclimatization will be carried over to repeated exposure after a week of rest at low altitude. Finally, we aim to explore the role of AMS and blood oxygen saturation (SpO2) on changes in cognitive functions.

### MATERIALS AND METHODS

### Participants

A total of 41 participants (21 altitude-exposed, 20 controls) were recruited. All participants provided written informed consent. Inclusion criteria were currently living at <1,300 m (participants must have been living in Calgary, 1,103 m, for at least 1 year) and no overnight stays at altitudes >1,500 m during the 4 weeks preceding the study. Exclusion criteria included previous history of altitude illnesses at moderate altitude (<3,000 m), current pregnancy, and health impairment that required regular treatment. The screening for the inclusion/exclusion of participants was carried out at the Foothills Medical Center, Cumming School of Medicine, University of Calgary, Calgary, Canada (1,103 m). Twenty-one altitude-naive healthy young adults (age = 25.2 ± 3.7 years, education = 17.1 ± 2.5 years, 13 females, BMI = 24.5 ± 8.1 kg·m−<sup>2</sup> ) took part in the high-altitude expedition, 18 of whom lived in Calgary and three of whom lived in Zurich and surrounding area (Switzerland, altitude 490 m) and were therefore screened at University Hospital of Zurich, Zurich, Switzerland. Twenty altitude-naive healthy young adults completed the testing sessions at the University of Calgary, Canada (altitude, 1,103 m), and formed the control group. Within the control group, three participants did not complete the cognitive sessions for reasons external to the study, and therefore were excluded from the analyses. Hence, the control group included a total of 17 participants in the final analyses (age = 24.9 ± 2.6 years, education = 16.8 ± 1.8 years, 12 females, BMI = 23.4 ± 2.7 kgm−<sup>2</sup> ). The total final sample included in the analyses (from both Altitude and Control) consisted of 38 participants (age = 25.1 ± 3.2 years, education = 17 ± 2.2 years, 25 females). The study was approved by the University of Calgary Conjoint Health Research Ethics Board (CHREB ID: REB15-2709) and registered as a clinical trial in ClinicalTrials.gov (NCT02738307). The flow of study participants through the Altitude and Control protocols is illustrated in **Figure 1**.

### Study Design

The high-altitude exposure schedule spanned over the course of a month with two cycles of high altitude exposure (Cycle 1 and Cycle 2) separated by a week at low altitude (**Figure 2**). The baseline measurements were taken in Santiago, Chile (520 m). The participants then flew to the Calama airport (∼2 h) and took a bus (∼2 h) to the basecamp at The Atacama Large Millimeter/submillimeter Array (ALMA) Operation Support Facility (ALMA ASF; 2,900 m). The participants traveled to the ALMA Observatory (5,050 m) by motor vehicle (about 45 min). Throughout the 6-day high altitude expedition (Cycles 1 and 2), participants spent nights at a support facility (ASF, 2,900 m) and commuted to ALMA Observatory (5,050 m) by motor vehicle to spend 4–8 h each day. According to ALMA policy, the participants were allowed to spend only 4 h at 5,050 m on the first day. Recovery measurements were taken in Santiago, Chile (520 m). The control subjects followed the same testing schedule as the altitude group at the Brain Dynamics Lab, The University of Calgary, Calgary, Canada without changes in altitude (1,103 m). An overview of the cognitive testing schedule is illustrated in **Figure 2** (**Figure 2A** for altitude and **Figure 2B** for control). The first test was a familiarization test in Cycle 1 while the second test was the baseline (BL). The first measurement at altitude (test 3 on day 1 of altitude exposure) was an acute exposure test (HA1) while test 4 at altitude (on day 6 of altitude exposure) was an acclimatization measurement (HA6). A similar schedule was followed in Cycle 2, and the same testing schedule and protocol (i.e., high-altitude protocol) was mirrored in the control group.

### Cognitive Test Battery

Cognitive tests assessed three domains of cognitive function: processing speed, sustained attention and executive functions. We created a custom battery with tests available within the Cambridge Neuropsychological Test Automated Battery (CANTAB <sup>R</sup> Cogntive Assessment Software; Cambridge Cognition, 1994). The CANTAB cognition battery can be administered several times while controlling for learning and repetition effects (Lowe and Rabbitt, 1998; Syväoja et al., 2015). This is achieved by generating random stimuli each time a participant logs in into his/her account for a new testing phase. The tests were administered on iPad Air 1 (model: A1474, dimensions: 9.7 inches retina display, iSO 9.3.1, Apple Inc., Cupertino, CA) and were completed in 30 min. The battery was administered nine times over the course of the expedition and nine times in controls as illustrated in **Figure 2** (**Figure 2A** for altitude and **Figure 2B** for controls). The individual components of the CANTAB battery included the Reaction Time (RTI) task to assess processing speed, the Attention Switching Task (AST) and the Rapid Visual Processing (RVP) task to assess attention and the One Touch Stockings of Cambridge (OTS) to test executive functions**.** The CANTAB cognition outcome measures, along with their abbreviations,and examples of the tasks, have been illustrated in **Figure 3**.

The RTI task is a measure of motor and processing speed. Participants are required to hold a button at the bottom of the screen (starting position). Circles are presented at the top of the screen (either 1 or 5 circles) and at some point, one of the circles will flash yellow. Participants must then tap the

highlighted circle as quickly as possible and then go back to the starting position. The outcome measures that we analyzed for this task are limited to the harder condition including 5 circles. The median reaction time (RTIFMDRT) considers how long it takes for the participants to touch the yellow circle after perceiving the change in color. Movement time (RTIFMMT) instead refers to the interval between the release of the starting position button and contact with the yellow circle. The AST measures individuals' ability to inhibit irrelevant information (selective attention). An arrow appears on the screen either on the left or right portion of the screen, pointing in either direction. Each trial displays a cue prompting the participant to press the left or right button on the screen to evaluate either the position of the arrow in the screen or the direction where the arrow is pointing. The RVP task is a measure of sustained attention. During the test, an array of numbers from 2 to 9 is presented in a pseudo-randomized order in the middle of the screen. The participants are required to press as fast as they can a button at the bottom of the screen when they see a certain array (2-4-6, 4-6-8, and 3-5-7). The OTS task measures executive function and the ability to plan. The participants are shown combinations of three-dimensional (3-D) colored balls and they must indicate in a box at the bottom of the screen, the number of moves (least possible amount) that they think are required to reproduce the same combination from a different starting position. All the CANTAB parameters were measured in milliseconds (ms) except ASTTC and OTSPSFC which represented frequencies (n).

# Lake Louise Score (LLS) and SpO<sup>2</sup>

AMS was assessed using the Lake Louise Score (LLS) (Roach et al., 1993) and diagnosed when the total LLS score was ≥5 (Maggiorini et al., 1998; Dellasanta et al., 2007). The SpO<sup>2</sup> was measured at rest with a finger pulse oximeter placed on the index finger.

### Data Analyses

### **Cognitive outcomes**

Cognitive changes over the course of the high-altitude exposure compared to baseline were analyzed with a series of Repeated Measures Mixed Model Analyses of Variance (RM-ANOVAs). We utilized RM-ANOVA to test our a priori hypothesis of expected changes in cognitive outcomes from baseline (BL, 520 m) to acute exposure to altitude (HA1) and acclimatization period (HA6) during each cycle. The cognitive outcome measures at BL, HA1 and HA6 were entered as within-subjects factors (altitude exposure<sup>∗</sup> 3) while group (Altitude vs. Control) as between-subjects factor in the analysis model. Cognitive outcome measures (RTI, AST, RVP and OTS) for Cycle 1 and Cycle 2 were analyzed separately to test the specific hypothesis of carry-over effects due to re exposure to very high altitude. All the analyses were two-tailed, and statistical significance was set at p < 0.05. Descriptive data were expressed as mean ± standard deviation (mean ± SD) and were assessed for violation of normality. Greenhouse-Geisser (GG) correction was used when sphericity, as measured with the Mauchly's test, was violated. Only significant altitude exposure (BL, HA1, HA6)

∗ group (altitude, controls) interactions were followed up by using pairwise comparisons with Bonferroni corrections to test within group differencies at each altitude exposure. The statistical analyses were carried out using the Statistical Package for the Social Sciences (SPSS, Version 24, IBM Corporation, New York 10504-1722, USA). The graphical illustration (study design and changes in cognition plots) were genereated using SyStat (SigmaPlot 13.0, Systat Software Inc, San Jose, CA, USA).

### **Covariates**

To examine the contribution of SpO<sup>2</sup> changes and total LLS to changes in cognitive measure due to altitude exposure, we used a Repeated Measures Analyses of Covariance (RM-ANCOVAs) with either SpO<sup>2</sup> changes or LLS as covariates. The changes in SpO<sup>2</sup> with altitude exposure were computed as acute change in SpO<sup>2</sup> (Acute SpO<sup>2</sup> = HA1 – BL), acclimatization change in SpO<sup>2</sup> (Acclimatization SpO<sup>2</sup> = HA6 – BL) and change in SpO<sup>2</sup> at

expedition in (A) and data collection in (B).

high altitude (Altitude SpO<sup>2</sup> = HA6 – HA1). The changes were calculated for both cycles separately.

### **Acclimatization carry-over effects over the cycles**

We tested the carry-over effects from Cycle 1 to Cycle 2 by computing differences in scores at baseline (1BL), acute exposure (1HA1) and acclimatization (1HA6) for each cognitive outcome measures i.e., change (1) = Cycle 2 – Cycle 1 at each data point in two cycles. Another RM-ANOVA was used to analyze persisting differences across the two cycles at different altitude exposures. We further extended our analysis to explore the acclimatization carry-over effects by calculating percent change in the cognitive variables and comparing them between the two cycles with paired t-tests. The acute percent change was calculated as [(BL-HA1)/BL<sup>∗</sup> 100] and acclimatization percent change as [(BL-HA6)/BL<sup>∗</sup> 100] whereas percent change at high altitude as [(HA1-HA6)/HA1<sup>∗</sup> 100] for both cycles.

### RESULTS

Details of the analyzed cognitive outcomes (mean ± SD) with RM-ANOVA for the altitude group and control participants are presented in **Table 1**. The more extensive descriptive data for all valid entries across all the sessions for both Cycle 1 (Familiarization, FL; Baseline, BL; Acute exposure, HA1 and Acclimatization exposure, HA6 and Recovery, REC) and Cycle 2 (Baseline, BL; Acute exposure, HA1 and Acclimatization exposure, HA6 and Recovery, REC) in both altitude and control groups have been presented in the **Supplemental Table 1**.

### High Altitude Exposure: Cycle 1 Cognitive Outcomes

### **Reaction Time (RTI)**

There was a significant main effect of altitude exposure (BL = 361.0 ± 37.0, HA1 = 367.1 ± 52.5, HA6 = 353.8 ± 31.8) on the RTIFMTSD [F(1.331,42.590) = 6.01, p = 0.012 GG, η 2 <sup>p</sup> = 0.158). However, there was no group-by-phase interaction (p = 0.181) meaning that there were no differences between the altitude group and controls at different altitude exposures. There was no main effect of altitude or group-by-phase interaction on the RTIFMDRT and the RTIFMMT.

### **Attention Switching Task (AST)**

There was a main effect of altitude exposure [BL = 544.3 ± 94.2, HA1 = 534.7 ± 89.3, HA6 = 503.9 ± 79.9, F(2, 64) = 11.2, p < 0.001, η 2 <sup>p</sup> = 0.259] and a group by altitude exposure interaction (altitude vs. controls) [F(2, 64) = 4.6, p = 0.013, η 2 <sup>p</sup> = 0.126] on the ASTLM scores. This means that there were significant differences between the altitude and control groups


at different altitudes. Follow-up pairwise comparisons revealed indeed that, in the altitude group, the ASTLM score was not impacted by acute exposure to altitude, but decreased with the acclimatization compared to acute exposure [HA1 (587.9 ± 91.9) vs. HA6 (525.8 ± 91.2), t(19) = 5.784, p < 0.001]. No significant differences were present in the control group. We observed similar effects for the ASTLSD with a main effect of altitude exposure [BL= 148.8 ± 63.9, HA1 = 152.0 ± 69.0, HA6 = 123.2 ± 55.9, F(2, 64) = 9.3, p < 0.001, η 2 <sup>p</sup> = 0.226] and a group (altitude vs controls) by altitude exposure interaction [F(2, 64) = 4.8, p = 0.011, η 2 <sup>p</sup> = 0.131]. At follow up comparisons in the altitude group exclusively, the ASTLSD score was not impacted by acute exposure to altitude but decreased with the acclimatization as compared to acute exposure [HA1 (189.4 ± 86.1) vs. HA6 (134.5 ± 64.9); t(19) = 5.427, p < 0.001]. No significant differences were found in the control group. There was no main effect of altitude exposure, nor group (altitude vs. control) by altitude exposure interaction for the ASTTC score.

### **Rapid Visual Processing (RVP)**

There was a significant main effect of altitude exposure on the RVPA score [BL= 0.95 ± 0.05, HA1 = 0.95 ± 0.06, HA6 = 0.97 ± 0.03; F(2, 64) = 3.5, p = 0.037, η 2 <sup>p</sup> = 0.098] and on the RVPMDL score [BL= 414.9 ± 46.7, HA1 = 431.1 ± 90.6, HA6 = 400.7 ± 35.8; F(1.376,44.021) = 3.6, p = 0.03, η 2 <sup>p</sup> = 0.1]. However, the group by altitude exposure interactions showed only a trend (RVPA, p = 0.162, η 2 <sup>p</sup> = 0.055 and RVPMDL, p = 0.064, η 2 <sup>p</sup> = 0.093). The RVPLSD did not change with altitude or between groups.

### **One Touch Stockings of Cambridge (OTS)**

We observed a main effect of altitude exposure on both OTSMLFC [BL= 12.26 ± 1.78, HA1 = 12.20 ± 1.38, HA6 = 12.24 ± 1.74; F(2, 64) = 6.861, p = 0.002, η 2 <sup>p</sup> = 0.177] score and OTSLFCSD [BL= 15646.7 ± 8696.2, HA1 = 14306.3 ± 9106.2, HA6 = 10802.8 ± 8003.5; F(2, 64) = 5.081, p = 0.009, η 2 <sup>p</sup> = 0.137] but no group by altitude exposure interactions. However, there was no main effect of altitude exposure and group by altitude exposure interaction on the OTSPSFC score.

### SpO<sup>2</sup> and Lake Louise Score (LLS)

The effect of altitude exposure on ASTLM reported previously during acute and acclimatization visits disappeared when controlling for Acute SpO<sup>2</sup> and Acclimatization SpO2. However, the effect of altitude exposure on ASTLM persisted when controlling for Altitude SpO<sup>2</sup> [F(2, 30) = 7.6, p = 0.002, η 2 <sup>p</sup> = 0.338]. Similar observations were found in ASTLSD. The differences in altitude exposure on ASTLSD when controlling for both Acute SpO<sup>2</sup> and Acclimatization SpO<sup>2</sup> disappeared but the altitude exposure effect persisted when controlling for Altitude SpO<sup>2</sup> [F(2, 30) = 9.2, p = 0.001, η 2 <sup>p</sup> = 0.38]. During acute exposure (HA1), when total LLS was entered as covariate, the changes in ASTLM due to altitude still persisted [F(2, 30) = 3.6, p = 0.039, η 2 <sup>p</sup> = 0.195]. On the contrary, the effect of altitude exposure on ASTLSD disappeared with LLS as covariate.

 mean \*\*\* <0.001, \*\* <0.01, \* <0.05; The asterisks (\*) indicate statistically significant

 are

group-by-altitude

 exposure interaction.

### High Altitude Exposure: Cycle 2

### Cognitive Outcomes

### **Reaction Time (RTI)**

The RTIFMTSD score had no main effect of altitude exposure, but a group by altitude exposure interaction [F(2, 70) = 5.96, p = 0.004, η 2 <sup>p</sup> = 0.145]. The follow up comparisons showed that the RTIFMTSD score was not affected by the acute altitude exposure. However, it decreased after the acclimatization from acute exposure [HA1 (35.2 ± 21.1) vs. HA6 (23 ± 8), t(19) = 2.590, p = 0.018]. These results, however, did not survive Bonferroni correction for multiple comparisons. The RM-ANOVA on the RTIFMDRT score and on the RTIFMMT score did not show any significant effects of altitude exposure. No significant differences were present in the control group.

### **Attention Switching Task (AST)**

There was a main effect of altitude exposure [BL= 156.8 ± 2.9, HA1 = 154 ± 6, HA6 = 155.7 ± 3.6, F(1.436,50.272) = 7.6, p = 0.004GG, η 2 <sup>p</sup> = 0.178] on the ASTTC and a group by altitude exposure interaction [F(1.436,50.272) = 3.6, p = 0.034GG, η 2 <sup>p</sup> = 0.092]. In the follow up comparisons, the ASTTC decreased from baseline to altitude [BL (156 ± 3) vs. HA1 (152.3 ± 7.2), t(20) = 2.889, p = 0.009] and improved with the acclimatization [HA1 (152.1 ± 7.2) vs. HA6 (154.7 ± 3.6), t(19) = −2.309, p = 0.032] only in the altitude group. However, the changes did not survive multiple comparisons correction. There was a main effect of altitude exposure (BL = 492 ± 75, HA1 = 502 ± 89, HA6 = 478.5 ± 73.6, F(2, 70) = 3.471, p = 0.037, η 2 <sup>p</sup> = 0.09] on the ASTLM score and a group by altitude exposure interaction [F(2, 70) = 3.144, p = 0.049, η 2 <sup>p</sup> = 0.082]. In the followup comparisons, ASTLM score slightly increased with acute exposure although this change was not significant. However, ASTLM score significantly decreased with the acclimatization compared to acute exposure [HA1 (528.4 ± 96.6) vs. HA6 (485.2 ± 82.1), t(19) = 2.879, p = 0.010] in the altitude group and not in the control group. Similarly, we found a main effect of altitude exposure (BL = 123.4 ± 54.8, HA1 = 140.7 ± 71.2, HA6 = 119.3 ± 50.8, F(2, 70) = 3.6, p = 0.032, η 2 <sup>p</sup> = 0.093] on the ASTLSD score, and a group by altitude exposure interaction [F(2, 70) = 4.0, p = 0.023, η 2 <sup>p</sup> = 0.103]. On follow-up comparisons, difference in ASTLSD score was only a trend [BL (135 ± 61) vs. HA1 (161.1 ± 81.5), t(20) = −1.845, p = 0.080]. ASTLSD decreased with the acclimatization compared to acute exposure [HA1 (164.3 ± 82.3) vs. HA6 (122.4 ± 53.3), t(19) = 3.005, p = 0.007]. The AST variables did not change significantly over different time points in the control group as we observed in the altitude group.

### **Rapid Visual Processing (RVP)**

There was no main effect of altitude exposure nor group by altitude exposure interaction on the RVPA. There was no main effect of altitude exposure on the RVPMDL score, but there was a group by altitude exposure interaction [F(2,70) = 4.7, p = 0.012, η 2 <sup>p</sup> = 0.119]. However, when directly compared, there was no difference. There was a trend to decrease with acclimatization compared to acute exposure [HA1 (414.8 ± 34.7) vs. HA6 (392.8 ± 36.5), t(19) = 2.081, p = 0.051] on RVPMDL, however this trend did not survive a correction for multiple comparisons (Bonferroni). The RVPLSD score did not change over time with the exposure to altitude. The RVP variables in control group did not vary significantly over time.

### **One Touch Stockings of Cambridge (OTS)**

There were no significant main effects nor interactions on the OTSPSFC score, OTSMLFC and OTSLFCSD.

### SpO<sup>2</sup> and Lake Louise Score (LLS)

The changes in SpO<sup>2</sup> for Cycle 2 were calculated in a similar manner as in Cycle 1 and entered as covariates in the RM-ANCOVA analysis to investigate the role of 1SpO<sup>2</sup> cognitive changes at altitude. The significant changes due to altitude exposure persisted when controlling for Altitude SpO<sup>2</sup> [F(2, 30) = 3.6, p = 0.037, η 2 <sup>p</sup> = 0.167] on the ASTLM score but disappeared when controlling for Acute 1SpO<sup>2</sup> and Acclimatization SpO2. The significant changes on the ASTLSD score due to altitude exposure persisted even after controlling for Altitude SpO<sup>2</sup> [F(2, 30) = 3.7, p = 0.033, η 2 <sup>p</sup> = 0.172] but there were no significant changes when controlling for Acute SpO<sup>2</sup> and Acclimatization SpO2. The total LLS was entered as covariate in the ANCOVA to investigate the role of AMS symptoms on the ASTLM and ASTLSD scores relative to HA1. The cognitive changes related to different exposures to altitude disappeared with AMS score as covariate.

### Acclimatization Carry-Over Effects Over the Cycles

We explored the carry-over effect across the cycles (Cycle 1 and Cycle 2) with the calculation of changes in cognitive functions i.e. 1CANTAB = Cycle 2 - Cycle 1 at BL, HA1 and HA6 time points. We then ran a RM-ANOVA on the cognitive outcomes change scores at 1BL, 1HA1. and 1HA6. There was a main effect of altitude exposure on ASTLM [F(2, 64) = 5.8, p = 0.005, η 2 <sup>p</sup> = 0.154] with a smaller change in ASTLM over the acclimatization exposure compared to baseline [1BL (−61.2 ± 70.2) vs. HA6 (−28.0 ± 58.0, p = 0.007)]. Similarly, we found a main effect of altitude on ASTLSD score [F(2, 64) = 4.0, p = 0.023, η 2 <sup>p</sup> = 0.112] with a smaller change in ASTLSD over acclimatization exposure compared to baseline [1BL (−28.4 ± 41.2) vs. 1HA6 (−0.2235 ± 34.8), p = 0.032]. For both outcomes (ASTLM and ASTLSD), there was no significant group by altitude exposure interaction. There were no significant changes across the two cycles for AST total correct nor other outcomes for the OTS, RVP, and RTI tests. The changes in cognitive function at altitude over two cycles as analyzed with RM-ANOVA have been illustrated in **Figure 4**.

We did not observe any significant differences in the percent changes of AST parameters during acclimatization indicating that the significant changes observed over the acclimatization period in Cycle 1 are similar in magnitude as the changes observed in Cycle 2. Similarly, there were no significant differences in the percent changes at altitude from the baseline between the two cycles.

FIGURE 4 | Changes in cognition (CANTAB outcome parameters with RM-ANOVA) over two cycles (1CANTAB = Cycle 2 – Cycle 1) at very high altitude during acute, subacute and repeated exposure comparing with controls. Figure has three horizontal box panels. The first panel (A–C) illustrates "processing speed" i.e., changes in Reaction Time (RTI) parameters (1RTIFMDRT, 1RTIFMMT, and 1RTIFMTSD), the second box panel contains "attention" in which first panel within the box (D–F) shows changes in Attention Switching Task (AST) parameters (1ASTTC, 1ASTLM, and 1ASTLSD) while the second panel within the second box (G–I) shows changes in Rapid Visual Processing (RVP) parameters (1RVPA, 1RVPMDL, and 1RVPLSD) and the third box panel (J–L) shows changes in One Touch Stockings of Cambridge (OTS) parameters (1OTSPSFC, 1OTSMLFC, and 1OTSLFCSD). The x-axis depicts different time points of data collection for Altitude exposure and Control groups at Baseline (BL), Acute exposure (HA1) and Acclimatization exposure (HA6). The y-axis depicts changes in cognitive parameters as mean ± SD for Altitude and Control group. Symbols: Black filled bars, Altitude group; empty bars, Control group; 1, Change; ms, milliseconds; n, number. BL, baseline; HA1, acute exposure to altitude (day 1); HA6, acclimatization exposure to altitude (day 6); RTI, Reaction Time; AST, Attention Switching Task; RVP, Rapid Visual Processing; OTS, One Touch Stockings of Cambridge; 1RTIFMDRT, RTI Median Five-choice Reaction Time; 1RTIFMMT, RTI Mean Five-choice Movement Time; 1RTIFMTSD, RTI Five-choice Movement Time Standard Deviation; 1ASTTC, AST Total Correct; 1ASTLM, AST Latency Mean; 1ASTLSD, AST Latency Standard Deviation; 1RVPA, Rapid Visual Processing Accuracy; 1RVPMDL, 1RVP Mean Response Latency; 1RVPLSD, RVP Response Latency Standard Deviation; 1OTSPSFC, OTS Problems Solved on First Choice; 1OTSMLFC, OTS Mean Latency First Choice; 1OTSLFCSD, OTS Latency to First Choice Standard Deviation.

# DISCUSSION

In this study, we investigated the effects of acute exposure, acclimatization and repeated exposure to very high altitude on cognitive functions in altitude-naïve individuals compared to control subjects tested at low altitude. We report four major findings. First, cognitive abilities, particularly sustained attention and inhibition of irrelevant information (selective attention) measured with the AST, significantly improved with acclimatization in both cycles. Second, the improvement gained in cognitive functions during the acclimatization period in Cycle 1 did not carry over to the repeated exposure in Cycle 2. Third, changes in SpO<sup>2</sup> explained changes in ASTLM score and ASTLSD score during acute and acclimatization exposures, but not during altitude stay in both cycles. Finally, the degree of acute mountain sickness, reflected by the LLS, explains in part the changes in AST scores (ASTLSD in Cycle 1, and ASTLM and ASTLSD in Cycle 2). We did not find any significant changes in reaction times, visual processing and executive functions during acute and acclimatization exposures. The novelty and strengths of this study include a strong experimental design, a robust cognitive battery not previously used in altitude studies and the use of a control group tested at low altitude.

Previous studies have used a variety of cognitive tests to separately assess processing speed, attention and executive functions (Harris et al., 2009; Subudhi et al., 2014; Nation et al., 2017; Phillips et al., 2017). However, none has used a battery to test these cognitive domains concurrently. Here, we used a custom cognitive battery built within the CANTAB (Cambridge Cognition, 1994; Strauss et al., 2006) cognitive test collection, which included tests to assess all three domains simultaneously. Custom batteries assembled within CANTAB have been shown to be robust and well suited for repeated measures testing (Syväoja et al., 2015). Within the analyses, we selected outcomes such as mean response latency, standard deviation, and accuracy for each cognitive domain, variables previously used and validated for neuropsychological assessments in other contexts and clinical populations (Sweeney et al., 2000). We included the RTI task because reaction times have been used as a measure of processing speed in both field (Kramer et al., 1993; Ma et al., 2015; Chen et al., 2017) and laboratory (hypoxic chamber) (Hornbein et al., 1989; Turner et al., 2015; Pramsohler et al., 2017) settings. Further, we chose the Attention Switching Task as a measure of attention and ability to inhibit irrelevant information. High altitude exposure significantly decreases test accuracy and increases reaction time in the "Word-Color Stroop Test," a test that also measures inhibition and attention (Asmaro et al., 2013). Similarly, the choice of Rapid Visual Processing Task was based on the previous studies (Kramer et al., 1993; Finn and McDonald, 2012) in which this task was used due to its sensitivity to both neurological damage (Finn and McDonald, 2012) and high altitude exposure (Kramer et al., 1993). Finally, we chose the One Touch Stockings of Cambridge Test to assess executive functions at altitude in line with previous studies (Asmaro et al., 2013). Similarly, we used problems solved on first choice, Mean Latency and SD of first choice which are the outcomes that CANTAB recommends to assess planning and spatial working memory (Naef et al., 2017). To our knowledge, this is the first time that a CANTAB custom cognitive battery has been used to explore the effects of very high altitude and hypoxia exposure. Cognitive assessments in the altitude literature are often confounded by multiple factors such as mode and rate of ascent, absolute altitude gained, physical exertion or exercise, cold, radiation and individual susceptibility to hypobaric hypoxia (Virués-Ortega et al., 2004; Yan, 2014; Taylor et al., 2016; McMorris et al., 2017). In our study, we controlled these confounding factors by using a design that involved rapid ascent to very high altitude (5,050 m), minimal or no physical activity and lack of exposure to environmental stressors because participants remained inside the ALMA facility during the testing sessions.

### Processing Speed

Previous studies have often reported a reduction in reaction times during acute exposure to altitude (Ma et al., 2015; Chen et al., 2017). It is often argued that the reduction in processing speed is a compensatory mechanisms to try to increase test accuracy at the expense of speed (Bahrke and Shukitt-Hale, 1993). Consistently with previous studies (Dykiert et al., 2010; Subudhi et al., 2014), we found that altitude exposure reduced the variability (SD) in the RTI Five-choice Movement Time but that the groupby-altitude interaction was not signficant in Cycle 1. During repeated exposure, in Cycle 2, altitude exposure had no effect on the RTI Five-time Movement time SD score although there was a group by altitude interaction (i.e., altitude and control subjects have different variances in processing speed at different altitudes). With acclimatization in Cycle 2, the RTI Five-time Movement time SD decreased during acclimatization although the significance was lost after post-hoc corrections. The lack of significant effects of acute exposure to altitude on reaction times, contrary to findings of other studies (Sharma et al., 1975; Dykiert et al., 2010), could be due to the small sample size and to the fact that participants were exposed to 5,050 m only 6–8 h/day and slept at lower altitude (2,900 m). Further, participants in our study may have benefitted from the specific exposure pattern with sleeping at lower altitude (Richalet et al., 2002; Farias et al., 2006; Vearrier and Greenberg, 2011) compared to other types of expedition/climbing exposure (Cavaletti et al., 1987; Kramer et al., 1993; Abraini et al., 1998). The mean and median fivechoice reaction time scores did not vary significantly (neither the main effect of altitude nor the group-by-altitude interaction were significant) during acute, subacute and repeated altitude exposures. It is possible that the RTI task was too simple or not sensitive enough to measure the effects of high altitude exposure as reported previously (Roach et al., 2014). Alternatively, it is possible that complex reaction times are not profoundly affected below 6,000 m altitude (Virués-Ortega et al., 2004).

### Attention

In our study, we found that in the AST task both Mean Reaction Latency and Reaction Latency SD were significantly affected by altitude exposure and a group by altitude interaction (i.e., the altitude group and the control group performed differently with different altitude exposures with the only the altitude group showing differences at the three data points). AST Mean Reaction Latency score and AST Reaction Latency SD score improved over acclimatization compared to acute exposure but were not significantly different from baseline. This suggests that acclimatization plays a crucial role in restoring cognitive functions (Subudhi et al., 2014). The Mean Reaction Latency and SD seem to be more sensitive measures to assess the effects of high altitude exposure as compared to AST Total Correct score. With the repeated exposure in Cycle 2, in contrast to Cycle 1, there was a main effect of altitude on the AST Total Correct score and a group by altitude interaction. Particularly, the AST Total Correct score decreased during acute exposure and improved with acclimatization. In Cycle 2, both AST Mean Latency and AST Latency SD showed a main effect of altitude exposure as well as a group by altitude interaction. Both AST Mean Latency and AST Latency SD improved with acclimatization but only trended to increase (i.e., worse performance) with acute exposure. Interestingly, in both cycles, we did not find significant differences in AST Mean Latency and SD due to acute exposure but we found significant improvements with acclimatization. However, the observed changes did not translate to repeated exposures consistent with previous findings on reascent (Subudhi et al., 2014). The intriguing finding, i.e., no effect of acute exposure to altitude, may be due to the passive exposure (the ascent via motorized vehicles) and the lack of physical exertion at altitude. Our study participants were in fact comfortably resting at the ALMA Observatory facility at 5,050 m and measurements were taken a few hours after their arrival. Previous studies either recruited climbers (Cavaletti and Tredici, 1993; Kramer et al., 1993; Bonnon et al., 1999) and trekkers (Harris et al., 2009; Phillips et al., 2017) or simulated altitude by lowering the percentage of Oxygen (FiO2) and/or pressure in an altitude chamber (Hornbein et al., 1989; Pramsohler et al., 2017). This heterogeneity in the study population and absolute altitude reached makes it harder to compare findings from the different studies. The individual variability in cerebral hypoxia susceptibility during acute high altitude exposure (Cavaletti and Tredici, 1993) may also partially explain our findings. The consistent improvement over acclimatization in both cycles could be in fact due to our unique pattern of high altitude exposure (i.e., ∼16 h spent at 2,900 m (sleeping altitude) and ∼8 h spent at very high altitude (5,050 m). The pattern of repeated re-oxygenation (2/3 of the 24-h cycle spent at 2,900 vs. 5,050 m) with restful sleep might have significantly increased the beneficial effects of the acclimatization process and thereby improved AST outcomes. The testing schedule used in our study simulates the schedule of the workers at ALMA and other mining industries in the South American Andes (Richalet et al., 2002; Farias et al., 2006; Vearrier and Greenberg, 2011) and therefore differs slightly from the schedules commonly used in the field (Ma et al., 2015; Chen et al., 2017) or in chamber high altitude simulation (Hornbein et al., 1989; Turner et al., 2015). Overall our findings indicate that sustained attention and the ability to inhibit irrelevant information (selective attention) are impacted by acute high altitude exposure. This suggests that precision tasks that require long-term focus might be affected, and therefore, more difficult to execute during high altitude exposure.

We observed only subtle changes in the Rapid Visual Processing outcomes during altitude exposure in both Cycle 1 and 2. The significant effects observed were lost on post-hoc corrections for multiple comparisons. Our findings regarding RVP are consistent with findings from others who reported that rapid exposure to altitude has little effect on visual and auditory attention as compared to effects on learning and memory (Nation et al., 2017).

### Executive Function

A previous study conducted in an altitude chamber with equivalent simulated altitude of 6,096 m (FiO<sup>2</sup> = 10%) demonstrated impairments in cognitive functions including executive functions (Turner et al., 2015). We found a significant main effect of altitude exposure on both OTS Mean Latency to First Choice and SD scores but no group by altitude interaction. We did not find a significant main effect of altitude exposure on the OTS Problems Solved on First Choice score nor group by altitude interaction. Hence, the executive functions, measured with the OTS task, are not impaired by altitude exposure. Similar results were found for Cycle 2. The executive functions, as measured with the OTS task, do not seem to be affected by exposure to hypoxia as much as other cognitive domains (McMorris et al., 2017).

### Role of SpO<sup>2</sup> and LLS Score on Cognition SpO<sup>2</sup> on Cognitive Changes

The cognitive changes following cerebral impairment due to altitude hypoxia could be related to changes in SpO<sup>2</sup> (Yan et al., 2011; McMorris et al., 2017). In Cycle 1, the significant effects of altitude exposure disappeared when controlling for Acute SpO<sup>2</sup> and Acclimatization SpO<sup>2</sup> for AST Mean Reaction Latency scores and AST reaction Latency SD scores. Hence, the effects seen during acute and acclimatization exposures are primarily driven by hypobaric hypoxia. Interestingly, the significant result persisted when controlling for Altitude SpO<sup>2</sup> in both AST Mean Reaction Latency scores and AST Latency SD scores. During the repeated exposure in Cycle 2, we found a similar pattern as in Cycle 1. Significant cognitive changes in the AST Mean Reaction Latency scores and AST Reaction Latency SD scores due to altitude exposure persisted when controlling for Altitude SpO2, but disappeared when controlling for Acute SpO<sup>2</sup> and Acclimatization SpO2. The altitude effects on AST Mean Latency and AST Latency SD observed in the acclimatization period (HA6-HA1) in both cycles, provide strong evidence of a beneficial effect of acclimatization on cognition. On the other hand, the effects are cycle specific i.e., the effects found in Cycle 1 do not carry over to Cycle 2. Further, our results sheds light on the important role played by Altitude SpO<sup>2</sup> (HA6-HA1) on cognitive functioning strengthening the idea of using oxygen supplementation at very high altitude to improve safety and work performance among scientists and workers (West, 2003, 2015; Moraga et al., 2018).

### Total LLS Score on Cognitive Changes

AMS symptoms are classified as cerebral symptoms (Wilson et al., 2009; Imray et al., 2010) and consequentially, they are expected to be associated with impaired cognitive functions at high altitude (Dykiert et al., 2010) although this relationship is still unclear (Virués-Ortega et al., 2004; Yan, 2014). In Cycle 1, the significant changes in AST mean reaction latency scores due to acute altitude exposure (HA1) persisted when including the LLS score as covariate in the model, indicating that the AMS symptoms are not associated with changes in AST Mean Reaction Latency scores. Conversely, the effect of altitude exposure on AST Reaction Latency SD score disappeared with LLS score entered as covariate, which indicates that AMS symptoms may have played a role in changes in cognitive abilities for this outcome. These findings suggest that AMS score could be associated with certain outcomes (AST Latency SD, an index of variability) but not others (AST Mean Latency) during acute altitude exposure. It is noteworthy that recent studies have found that poor sleep quality is not to be associated with AMS (MacInnis et al., 2013; Hall et al., 2014). It is therefore possible that some of the cognitive functions that are altered by sleep disturbances are not sensitive to high altitude exposure or symptoms of AMS. Consistently, Kramer and colleagues did not find any significant correlations between AMS severity and cognitive performance among climbers (Kramer et al., 1993). The rate of ascent, the absolute altitude gained and the physical activity in the field might be responsible for the discrepancy in the findings. With repeated exposure in Cycle 2, the cognitive changes related to acute altitude exposure (HA1) disappeared when using total LLS score as a covariate. The AMS symptoms seem to play a role in changes in cognitive abilities for these outcomes even during repeated exposures although the LLS score in the repeated exposure was significantly decreased compared to acute exposure (HA1, Cycle 1).

# Acclimatization, Repeated Exposure and Carry-Over Effects

The principal reason behind the implementation of such work schedule in the Chilean workers with sleep periods at lower altitude, repeated exposure to 5,050 m, and a week of rest at sea-level is to try to minimize the adverse effects of very high altitude on workers' health and performance (Richalet et al., 2002; Farias et al., 2006; Vearrier and Greenberg, 2011) and allow them to see their families during the week of rest at low altitude. In our study which reproduces this schedule over two work-week cycles, we found a significant main effect of altitude exposure for the AST Mean Reaction Latency and AST Latency SD with a significant decrease in both AST Mean Reaction Latency and SD over the acclimatization exposure compared to baseline and AST Latency SD change across cycles. This indicates that the changes between baseline and acclimatization in Cycle 1 are different from the changes happening in Cycle 2, confirming that the significant improvement in cognitive functions with acclimatization is not maintained with a week of rest or with repeated exposure. Our findings are consistent with Subudhi and colleagues who reported that cognitive functions improved with acclimatization and the obtained gains are not completely retained upon re-ascent/repeated exposure (Subudhi et al., 2014). It is not entirely clear whether sleep at lower altitude favors this outcome or if this is a consequence of the acclimatization period itself. Further, we did not find any differences when comparing the two cycles in terms of percent change for AST during acute and acclimatization exposures. This means that the magnitude of the changes observed over the acclimatization in Cycle 1 is not different from the magnitude of the changes observed in Cycle 2. Our findings suggest that the acclimatization of cognitive functions at altitude is a dynamic process and it may not reach a plateau within approximately a week, in contrast to other physiological variables such as ventilatory acclimatization (Pamenter and Powell, 2016). One possibility is that cognitive functions may not benefit from high altitude acclimatization (Taylor et al., 2016) unlike other physiological functions such as athletic performance. However, these relationships may be different at higher elevation, with longer durations of stay and increased physical exertion (Shukitt-Hale et al., 1998).

### Strengths and Limitations Strengths

We exposed young healthy altitude-naïve individuals to very high altitude following the same schedule as the ALMA workers. Thus, this study is highly relevant to a significant workforce in Chile and other parts of South America, and could be of interest to governments and policy makers who regulate work at high altitude. We used a custom cognitive battery generated from the Cambridge Neuropsychological Test Automated Battery (CANTAB) test collection, which has not been used previously in high altitude research. The comprehensive cognitive battery tests processing speed, attention and executive function was used to assess the effects of acute exposure, acclimatization exposure and repeated exposure to very high altitude. Participants completed the cognitive assessments at various time points of altitude exposure. Similarly, we recruited a control group at low altitude, to test the between-group differences as well as control for practice effects. The use of this custom testing battery constitutes a strength of the study. The neuropsychological assessment is in fact conducted using portable wireless touch screen tablets and therefore applicable in many remote settings. The data are then stored and can be either wireless transferred or downloaded later. Further we administered the battery in English but it is validated to be administered in multiple languages making it a very flexible assessment tool in a variety of populations and settings. Another strength of the study is the use of a experimental design that controls factors such as environmental stressors and the effects physical exertion. This design allowed us to untangle the effects of hypoxia at high altitude from other confounders.

### Limitations

This study has also some limitations. First, it provides only a snapshot of the high-altitude exposure which is hard to compare to the repeated exposures of the high-altitude workers who have been following this schedule for many years. Second, we did not collect cognitive data at sleeping altitude i.e., mid-altitude (2,900 m) where participants (and workers) spend >15 h/day while they only spent ∼8 h/day at 5,050 m. Sleeping at lower altitude and increased oxygen levels might have had beneficial effects that counteracted the negative effects of acute highaltitude exposure. These should be areas of further investigation. Third, the findings from the study should be interpreted carefully when generalizing to other types of high altitude exposure or paticipants' groups due to the unique ascent profile, relatively small sample size and the weekly shift-work schedule that was used in our study. Finally, the participants in this study were passively exposed to altitude (via travel by plane from 520 to 2,900 m and by motorized vehicle from 2,900 to 5,050 m) and had minimal physical exertion as opposed to previous studies in climbers and trekkers.

### CONCLUSIONS

The findings of our study highlight the importance of acclimatization on restoring cognitive function after acute exposure to very high altitude. However, it is important to consider that the gains in cognitive functions during the acclimatization period in the first exposure are not carried over to repeated exposures. The SpO<sup>2</sup> is associated with cognitive changes during acute and acclimatization exposure and AMS scores might partially explain the cognitive changes. Taken together, our results suggest that the tasks that need sustained focus and high level of precision may be affected during acute exposure and also during repeated exposure or re-ascent. These findings and their implications for the safety and performance of the workers in the mine industry and other high-altitude workers highlight the importance of linking research groups and scientific findings with the organizational strategies of these specialized work sites. The findings would also be helpful for the related organizations and governments in policy formulation aiming at increasing the safety and security of high altitude workers. Future studies should focus on the effects of high altitude on learning and declarative memory, should include data collection at sleeping altitude (2,900 m) and most importantly, should recruit workers who have been working at high altitude for extended periods. The effects of room oxygen enrichment "oxygen conditioning" at high altitude (West, 2016a,b) for newcomers, as well as for high altitude residents, should also be the topic of future investigations.

### ETHICS STATEMENT

The study was approved by the University of Calgary Conjoint Health Research Ethics Board (CHREB ID: REB15-2709).

### AUTHOR CONTRIBUTIONS

MP and VG organized the data, carried out the analyses, drafted the manuscript and took the lead of manuscript finalization and submission process. KMB, LD, and JA were

### REFERENCES

Abraini, J. H., Bouquet, C., Joulia, F., Nicolas, M., and Kriem, B. (1998). Cognitive performance during a simulated climb of Mount Everest: implications for brain function and central adaptive processes under chronic hypoxic stress. Pflugers Arch. 436, 553–559. doi: 10.1007/s0042400 50671

involved in the control data collection, export, organization and preliminary analyses. KEB, BG, JR, and MJP were involved in the conceptualization, design, and planning of the study. SH, MF, SU, KEB, and MJP were involved in the field for the data collection, troubleshooting, and manuscript finalization. All authors went through all the versions of the manuscript and approved them.

# FUNDING

MJP is supported by a Discovery Grant from the Natural Sciences and Engineering Research Council (NSERC) of Canada (Principal Investigator (PI), MJP; 2014-05554), and a CIHR Operating Grant on the Regulation of cerebral blood flow in OSA (PI: MJP). AP is supported by an NSERC Discovery Grant of Canada (PI, ABP; 2013-418454). JA received support from and NSERC CGS-M scholarship. SH received support from the Dr Chen Fong doctoral scholarship (Hotchkiss Brain Institute). LD is supported by an Alberta Innovates Health Services (AIHS) Postgraduate Fellowship. VG is supported by The Brenda Strafford Centre on Aging, within the O'Brien Institute for Public Health, and the Brenda Strafford Foundation Chair for Alzheimer Research (BSFCAR). MJP holds the BSFCAR. KEB and SU are supported by Lunge Zurich, Swiss National Science Foundation. BG is supported by the Institute for Collaborative Biotechnologies through grant W911NF-09-0001 from the U.S. Army Research Office. The content of the information does not necessarily reflect the position or the policy of the Government, and no official endorsement should be inferred.

# ACKNOWLEDGMENTS

We express our sincere thanks to the study participants who took part in the high-altitude expedition to ALMA Observatory, Chile. The research would have been impossible without their involvement and cooperation throughout the expedition. We would also like to acknowledge our Chilean collaborators from the ALMA Head Office in Santiago, Chile and the staff of the ALMA observatory (Ivan Lopez, Daniel Soza, Charlotte Pon, and the health & safety and polyclinic teams), who greatly facilitated the study. Finally, we would like to thank Alicia Morales Soto for helping facilitate the logistics in Santiago.

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphys. 2018.01131/full#supplementary-material

Asmaro, D., Mayall, J., and Ferguson, S. (2013). Cognition at altitude: impairment in executive and memory processes under hypoxic conditions. Aviat. Space Environ. Med. 84, 1159–1165. doi: 10.3357/ASEM.3661.2013

Bahrke, M. S., and Shukitt-Hale, B. (1993). Effects of altitude on mood, behaviour and cognitive functioning. Rev. Sports Med. 16, 97–125.

Bärtsch, P., and Swenson, E. R. (2013). Acute high-altitude illnesses. N. Engl. J. Med. 368, 2294–2302. doi: 10.1056/NEJMcp1214870


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Pun, Guadagni, Bettauer, Drogos, Aitken, Hartmann, Furian, Muralt, Lichtblau, Bader, Rawling, Protzner, Ulrich, Bloch, Giesbrecht and Poulin. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Carotid Body Type-I Cells Under Chronic Sustained Hypoxia: Focus on Metabolism and Membrane Excitability

Raúl Pulgar-Sepúlveda<sup>1</sup> , Rodrigo Varas1,2, Rodrigo Iturriaga<sup>3</sup> , Rodrigo Del Rio4,5,6 \* and Fernando C. Ortiz<sup>1</sup> \*

1 Instituto de Ciencias Biomédicas, Facultad de Ciencias de la Salud, Universidad Autónoma de Chile, Santiago, Chile, <sup>2</sup> Facultad de Ciencias de la Salud, Universidad Autónoma de Chile, Talca, Chile, <sup>3</sup> Laboratorio de Neurobiología, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>4</sup> Laboratory of Cardiorespiratory Control, Department of Physiology, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>5</sup> Centro de Envejecimiento y Regeneración, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>6</sup> Centro de Excelencia en Biomedicina de Magallanes, Universidad de Magallanes, Punta Arenas, Chile

### Edited by:

Eduardo Colombari, Universidade Estadual Paulista Júlio de Mesquita Filho (UNESP), Brazil

### Reviewed by:

Thiago S. Moreira, Universidade de São Paulo, Brazil Davi J. A. Moraes, Universidade de São Paulo, Brazil

\*Correspondence:

Rodrigo Del Rio rdelrio@bio.puc.cl Fernando C. Ortiz fernando.ortiz@uautonoma.cl

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 29 May 2018 Accepted: 24 August 2018 Published: 19 September 2018

### Citation:

Pulgar-Sepúlveda R, Varas R, Iturriaga R, Del Rio R and Ortiz FC (2018) Carotid Body Type-I Cells Under Chronic Sustained Hypoxia: Focus on Metabolism and Membrane Excitability. Front. Physiol. 9:1282. doi: 10.3389/fphys.2018.01282 Chronic sustained hypoxia (CSH) evokes ventilatory acclimatization characterized by a progressive hyperventilation due to a potentiation of the carotid body (CB) chemosensory response to hypoxia. The transduction of the hypoxic stimulus in the CB begins with the inhibition of K+ currents in the chemosensory (type-I) cells, which in turn leads to membrane depolarization, Ca2<sup>+</sup> entry and the subsequent release of oneor more-excitatory neurotransmitters. Several studies have shown that CSH modifies both the level of transmitters and chemoreceptor cell metabolism within the CB. Most of these studies have been focused on the role played by such putative transmitters and modulators of CB chemoreception, but less is known about the effect of CSH on metabolism and membrane excitability of type-I cells. In this mini-review, we will examine the effects of CSH on the ion channels activity and excitability of type-I cell, with a particular focus on the effects of CSH on the TASK-like background K+ channel. We propose that changes on TASK-like channel activity induced by CSH may contribute to explain the potentiation of CB chemosensory activity.

Keywords: carotid body, chronic hypoxia, membrane depolarization, ion channels, TASK-like channel

# INTRODUCTION

The carotid body (CB) is the main peripheral chemoreceptor in mammals. Natural stimuli such as hypoxemia, hypercapnia, and/or acidosis increase the firing rate of the petrosal ganglion (PG) sensory afferent neurons projecting to the cardiovascular and respiratory regions in the brain stem (Gonzalez et al., 1994).

The current model for hypoxic chemoreception states that hypoxia evokes a depolarization of CB type-I (glomus) cells, leading to an increment of intracellular Ca2<sup>+</sup> and the subsequent release of one- or more-excitatory neurotransmitters to the nerve terminals of the PG neurons (Gonzalez et al., 1994; Iturriaga and Alcayaga, 2004; Nurse, 2005). It is well established that a key event in triggering the hypoxic response is the depolarization of the type-I cells (Buckler and Vaughan-Jones, 1994). In type-I cells from several species it has been found that hypoxia produces a fast

**294**

and reversible inhibition of K<sup>+</sup> currents (López-Barneo et al., 1988; López-López and González, 1992; Wyatt and Peers, 1995; Buckler, 1997) leading to a depolarization of type-I cells membrane and the consequence Ca2+entry, mainly throughout L-type Ca2<sup>+</sup> channels (Fieber and McCleskey, 1993; Wyatt et al., 1994). Among several molecules present in type-I cells, acetylcholine (ACh) and adenosine triphosphate (ATP) meets most of the criteria to be consider excitatory transmitters in the pathway (Varas et al., 2003; Iturriaga and Alcayaga, 2004), while dopamine (DA), nitric oxide (NO), and endothelin-1 (ET-1) modulate the chemosensory process (Iturriaga and Alcayaga, 2004).

Mammals exposed to sustained hypoxia (i.e., high altitude) develop ventilatory acclimatization featured by a progressively hyperventilation, due to an augmented CB responsiveness to hypoxia (Bisgard, 2000). In addition to the enhanced CB chemosensory responses, chronic sustained hypoxia (CSH) for weeks or months induces angiogenesis in the CB along with type-I cell hypertrophy and hyperplasia (Heath et al., 1985). Most of the studies on CB hypoxia acclimatization have been focused on putative changes in transmitters or modulators of CB chemoreception. Interestingly, DA, NO, and ET-1 are upregulated in the CB during the first week of chronic hypoxia (Bisgard, 2000). In addition, some reports indicate that CSH increases type-I cell excitability due to changes in K<sup>+</sup> and Na<sup>+</sup> channels expression. In spite of the efforts, the mechanisms by which CSH enhances CB chemosensory responses to hypoxia remains to be established. In this review, we will examine the effects of sustained hypoxia on CB metabolism and ion channels function, focusing on the type-I cell excitability.

### ELECTRICAL PROPERTIES OF TYPE-I CELLS

Type-I cells are small round-shaped cells (diameter ∼10 µm in the rat), with high input resistance of 5–6 G (Duchen et al., 1988; Xu et al., 2005), and resting membrane potential ranging from −50 to −70 mV (Gonzalez et al., 1994). Regarding to the inward currents expressed by type-I cells, in the adult rabbit it has been reported a tetrodotoxin-sensitive, voltage-gated Na<sup>+</sup> current with fast activation and inactivation kinetics (Duchen et al., 1988). However, in both adult and neonatal rat type-I cells, there are contradictory results regarding the expression of functional voltage gated Na<sup>+</sup> channels (Fieber and McCleskey, 1993; Hempleman, 1995; Stea et al., 1995; Cáceres et al., 2007). Voltage-gated Ca2<sup>+</sup> channels have been found in both rat and rabbit type-I cells (Fieber and McCleskey, 1993; Buckler and Vaughan-Jones, 1994; Wyatt et al., 1994). Most of the studies agree that L-type Ca2<sup>+</sup> channel is the most abundant subtype of Ca2<sup>+</sup> channel in type-I cells, however, in rabbit CB type-I cells, pharmacological evidences suggest the additional presence of N-, P/Q-type Ca2<sup>+</sup> channels and a ω-conotoxin-resistent Ca2<sup>+</sup> current (Overholt and Prabhakar, 1997, 1999). In addition, rat type-I cells express anionic currents (mainly Cl−) which are involved in the chemosensory transduction of acidic and hypercapnic stimuli (Carpenter and Peers, 1997; Iturriaga et al., 1998).

Type-I cells express a wide variety of K<sup>+</sup> channels. Rabbit type-I cells express at least two voltage-gated, TEA-sensitive Ca2<sup>+</sup> – independent K<sup>+</sup> currents. One of these conductances correspond to an oxygen-sensitive voltage-gated K<sup>+</sup> channel with a unitary conductance of ∼40 pS and an activation threshold around −40 to −30 mV (hereafter KO2; López-Barneo et al., 1988). The opening of this K<sup>+</sup> channel is reversibly inhibited by hypoxia with an IP<sup>50</sup> (PO<sup>2</sup> at which 50% of maximal inhibition is reached) near to 5–10 mmHg. This inhibition by hypoxia depends on membrane potential: maximal inhibition (40% of control activity) is reached at 0 mV (López-Barneo et al., 1988). The KO<sup>2</sup> channel has at least four closed states (C0–C4), one open state (O) and two inactivated states (I1–I2). Hypoxia induces both, stabilization of the "C0" state and promoting the channel from "O" state to inactivation state "I1" (Ganfornina and Lopez-Barneo, 1992).

Neonatal rat type-I cells express a maxi-K (BK) channel, whose opening depends on the PO2. This maxi-K channel has a large unitary conductance of ∼200 pS, it is blocked by charybdotoxin and activated by both membrane depolarization (threshold ranging from −40 to −20 mV) and by a rise in intracellular [Ca2+] above 100 nM (Peers, 1990; Wyatt and Peers, 1995). Wyatt and Peers (1995) found that hypoxia (PO<sup>2</sup> = 5– 10 mmHg) causes a reversible inhibition of this maxi-K current in cell-attached patch-clamp recordings. Later, Williams et al. (2004) proposed that the oxygen-sensitivity of the rat type-I cells is mediated by a heme oxygenase-2 (HO-2) associated with the maxi-K channel complex. However, this proposition was challenged by Ortega-Sáenz et al. (2006), when they found that hypoxic response of type-I cells remains intact in HO-2 null mice.

Interestingly, Buckler (1997) found that voltage gated K<sup>+</sup> channel blockers (10 mM TEA plus 5 mM 4-aminopyridine and 20 nM charybdotoxin) failed to depolarize type-I cells and to modify intracellular [Ca2+]. However, in the presence of these K<sup>+</sup> channel blockers, a hypoxic stimulus was able to evoke cell depolarization and a rise in intracellular Ca2<sup>+</sup> levels. This background K<sup>+</sup> current was inhibited by hypoxia with an IP<sup>50</sup> ∼10 mmHg, reaching a maximal ∼70% inhibition during anoxia. Hypoxia stabilized the background K<sup>+</sup> channel in its close state, but did not affect the opening/closing kinetics, or the open state duration (Williams and Buckler, 2004). This background K<sup>+</sup> channel correspond to a TWIKrelated acid-sensitive K<sup>+</sup> channel (TASK), member of the two-pore domain K<sup>+</sup> channel superfamily (Buckler et al., 2000). Additional evidences suggest the presence of TASK-1, TASK-2, TASK-3, TRAAK, and TREK (Buckler, 1997; Buckler et al., 2000; Yamamoto et al., 2002; Yamamoto and Taniguchi, 2006).

The discovery of oxygen-dependent voltage-gated K<sup>+</sup> channels promptly leads to the suggestion that hypoxia inhibits K <sup>+</sup> channels, which in turn produces type-I cells depolarization. Nevertheless, the resting membrane potential of type-I cells is quite stable at least 10 mV below the threshold activation of both maxi-K and KO<sup>2</sup> channels. Therefore, it was hard to conciliate a possible role of these K<sup>+</sup> channels in initiating the depolarization in response to hypoxia in type-I cells. Since background K<sup>+</sup> channels (Buckler, 1997; Buckler et al., 2000)

are active at resting conditions, the closure of these channels may explain the initiation of the depolarization evoked by hypoxia. The precise mechanism by which acute hypoxia is sensed remains controversial, but is clear that TASK, maxi-K, and KO<sup>2</sup> channels plays a key role in the depolarization required for the neurotransmitters release from type-I cells in response to hypoxia (**Figure 1**).

# CHRONIC SUSTAINED HYPOXIA (CSH)

Hypoxic-hypoxia (a PO<sup>2</sup> fall), could be classified as acute (seconds to minutes) or chronic (days to years). Acute hypoxia produced CB chemosensory excitation that evokes a reflex hyperventilation. CB increases its size in response to chronic sustained hypoxia (CSH) – to differentiate it from the intermittent paradigm – due to both increased number of cells and enhanced cell bodies diameter (Pequignot et al., 1984; Mills and Nurse, 1993). In this line, Pardal et al. (2007) have reported that in mice exposed to 10% O<sup>2</sup> for 7 days, a subpopulation of type-II cells acts like stem-cells, differentiating into new type-I cells. This mechanism may explain why CB hyperplasia is induced by CSH (Pardal et al., 2007). Likewise, several evidences suggest an important role of type-II cells in CB adaptation to different paradigms of chronic hypoxia (see Leonard et al., 2018 for a recent revision on this subject). The progressively increases in ventilation elicited by CSH, it is known as ventilatory acclimatization (Bisgard, 2000). Ventilatory acclimatization induced by CSH depends on the potentiation of the CB chemosensory response to hypoxia in several species (Bouverot et al., 1973; Olson and Dempsey, 1978; Smith et al., 1986; Vizek et al., 1987; Di Giulio et al., 2003). In this section we summarized some evidences concerning possible mechanisms behind this phenomenon.

### Increased Levels of NO and ET-1 in the CB

Chronic sustained hypoxia increases the levels of NO in the CB (Ye et al., 2002) due to an augmented expression of endothelial and inducible nitric oxide synthase (Ye et al., 2002). An increased NO levels evoked by CSH may change the oxidative state and/or produce protein nitrosylation in type-I cells (Monteiro et al., 2005). It is possible that increased levels of NO induced by CSH may contribute to enhances the oxygen sensitivity in the CB. Iturriaga et al. (2000) reported that during steady chemosensory excitation induced by hypoxia, bolus injections of a NO donor transiently reduced chemosensory discharges. However, during normoxia the same concentration of NO increased chemosensory discharge, suggesting a dual role of NO (see also Iturriaga, 2001). Endothelin-1 (ET-1) is also present in the CB. Chronic hypoxia for 2 weeks increases the ET-1 levels and the expression of the ETA receptor in rat type-I cells (Chen Y. et al., 2002). ET peptides have a proliferative effect in the CB stimulating cellular proliferation in CB primary cultures trough the ETA receptor activation (Paciga et al., 1999). In the rabbit CB superfused in vitro, Chen et al. (2000) found that ET-1 did not modify basal CB chemosensory discharge, but potentiates the response induced by hypoxia. In addition, Chen Y. et al. (2002) found that ET-1 increases Ca2<sup>+</sup> acting on L-type Ca2<sup>+</sup> channels in rabbit type-I cells. Although, ET may enhance the CB response to hypoxia in a CB preparation devoid of vascular effects, the excitatory effect of ET-1 on CB chemoreception is most likely mediated by its vasoconstrictor effect (Rey and Iturriaga, 2004). Chen et al. (2007) found that the enhanced basal and hypoxic evoked chemosensory activity following CSH was significantly reduced by concurrent treatment with the ETA receptors blocker bosentan. Thus, it is plausible that CSH-induced CB acclimatization would be partially mediated by ETA receptors.

# Changes in Levels and Secretion of DA, ACh, ATP, or Their Receptors

Carotid body type-I cells contains high levels of dopamine (DA), thus tyrosine hydroxylase (TH) is a conventional marker for the identification of type-I cells (Chou et al., 1998). CSH enhances both, the expression and activity of TH in CBs (Hanbauer, 1977; Czyzyk-Krzeska et al., 1992); as a consequence, there is an increased DA content. It is worth to mention that even though DA is released in response to acute hypoxia, its role in CB neurotransmission remains controversial (Fidone et al., 1990; Alcayaga et al., 2006). In most species, the evidence strongly suggests an inhibitory role for DA on CB chemoreception. It has been proposed that DA is involved in an auto-regulatory mechanism mediated by D2 dopaminergic receptors expressed in type-I cells (Fidone et al., 1990). Some evidences suggesting that the inhibitory role of DA could be attenuated in CBs exposed to CSH, leading to the increased chemoreactivity (Tatsumi et al., 1995). However, this is in sharp contrast with the increased TH activity during chronic hypoxia. Zhang et al. (2017) proposed that adenosine and DA control CB chemosensory discharge acting on petrosal neurons via their opposing actions on the hyperpolarization-activated, cyclic nucleotide-gated (HCN) cation current (Ih). By using a functional in vitro preparation of rat type-I cells/petrosal neurons co-cultures, they found that adenosine enhanced Ih in petrosal neurons acting on A2a receptors, while DA had the opposite effect via D2 receptors. Indeed, adenosine modified the Ih activation curve and increased firing frequency, while DA caused a hyperpolarizing shift in the curve and decreased the firing frequency (Zhang et al., 2017).

Acetylcholine and ATP have been proposed as excitatory transmitters in the CB, acting on nicotinic and P2X receptors, respectively (Zhang et al., 2000; Varas et al., 2003; Meza et al., 2012). In response to CSH, an increase of nicotinic acetylcholine receptor (nAChR) expression in chemosensory afferents has been reported (He et al., 2005). Paradoxically, nAChRs blockers does not revert the CB hyperreactivity evoked by chronic hypoxia (Bisgard, 2000). Regarding to the role played by ATP, it seems that CSH has no effect on P2X receptors expression in both, type-I cells and postsynaptic terminals of petrosal neurons. However, administration of P2X receptor blockers reduced the enhanced chemosensory response to acute hypoxia induced by CSH (He et al., 2006). Therefore, reports suggest that ACh and ATP may play a role in the hyperreactivity of the CB induced by CSH. Hence, it seems that chronic hypoxia enhances both the excitatory (purinergic and cholinergic) and inhibitory (dopaminergic) chemoreception pathways, in agreement with the proposal stating that the balance between excitatory and inhibitory transmitters is more relevant during chronically hypoxic conditions rather than normoxia (Prabhakar, 2006).

# Increase in the Proportion of Na<sup>+</sup> Currents and a Decrease in K<sup>+</sup> Currents

In type-I cells of neonatal rats subjected to long-lasting chronic hypoxia there is an increase in the proportion of Na<sup>+</sup> currents and a decrease in oxygen-sensitive K<sup>+</sup> current density (Hempleman, 1995). Likewise, Cáceres et al. (2007) found that rat CB hyperreactivity induced by CSH correlated with an increased expression of voltage-gated Na<sup>+</sup> channel Nav1.1 subunit. Carpenter et al. (1998) found that exposing adult rats to 10% O<sup>2</sup> for 3 weeks significantly reduced K<sup>+</sup> current density with no change in voltage-gated K<sup>+</sup> current amplitudes. However, rabbit CBs cultured in 5% O<sup>2</sup> for 2 days shows a decrease in expression of the Kv3.4 subunit of the voltage-gated K<sup>+</sup> channel (Kääb et al., 2005), suggesting a species-specific effect on Kvconductance.

Additionally, neonatal rat type-I cells exposed to hypoxia (9– 14 days, 10% O2) shows a decreased charybdotoxin-sensitive K<sup>+</sup> current, suggesting the downregulation of the maxi-K channel (Wyatt et al., 1995). It is worth to mention that in this paradigm, the rats have "blunted" rather hyperreactive responses to acute hypoxia (Wyatt et al., 1995). In culture, human maxi-K channels show an increased calcium sensitivity after 72 h of hypoxia. This is probably due to the reported threefold increase in maxi-K β-regulatory subunit expression induced by CSH (Hartness et al., 2003). However, cultured type-I cells from neonatal rat subjected to CSH, express normal levels of maxi-K and TASK-1 channels (Nurse and Fearon, 2002). Additionally, cultured CBs cells exposed to chronic hypoxia for 2 weeks showed an increase in Na<sup>+</sup> currents, but no changes in Ca2<sup>+</sup> currents were reported (Stea et al., 1995), findings confirmed later in a paradigm of adult rats exposed to CSH (Peers et al., 1996).

We found an enhanced inhibition of the TASK-like current in neonatal rat type-I cells in cultured exposed to hypoxia for 2 days (Ortiz et al., 2007). While acute hypoxia (1% O2) evokes a ∼70% inhibition of the TASK-like current in control conditions, after 48 h of hypoxia current inhibition reaches ∼90%. Importantly, these results were later reproduced (Ortiz et al., 2013) in an animal model for chronic intermittent hypoxia suggesting a common preserved mechanism of CB adaptation to hypoxic environments.

The evidence suggests that regulation of the expression and/or function of ionic currents are important adaptive mechanisms for the CB subjected to chronic hypoxia. Nevertheless, with the exception of the work of Cáceres et al. (2007), there is no physiological evidence that the CSH protocols used in above mentioned studies (Hempleman, 1995; Stea et al., 1995; Carpenter et al., 1998) really potentiated the CB chemosensory responses to acute hypoxia. On the contrary, most of them used long-term CSH exposition (>2 weeks) which is associated with an attenuation rather than a potentiation of CB chemosensory responses to acute hypoxia (Tatsumi et al., 1991). Hence, no convincing conclusion can be draw.

### Increased Activity of AMP-Dependent Kinase and Protein Kinase C (PKC)

The activities of maxi-K and TASK channels are modulated by cell metabolism. Wyatt et al. (2007) found that AMPdependent kinase (AMPK) is colocalized with both maxi-K and TASK channels in rat type-I cells. They found that AMPK activation inhibits those K<sup>+</sup> currents triggering membrane depolarization, whereas AMPK-antagonist application prevents the depolarization induced by hypoxia (Wyatt et al., 2007).

TASK channel current is activated by intracellular nucleotides, such as ATP (Varas et al., 2007). At physiological concentrations (1–3 mM) ATP strongly promotes channel opening. Accordingly, in absence of ATP, TASK channel activity decreases in ∼50 % in inside-out patches (Williams and Buckler, 2004; Varas et al., 2007). Thus, ATP production seems to be central in regulating the activity of TASK-like and other potassium channels such as maxi-K (Williams and Buckler, 2004; Varas et al., 2007; Holmes et al., 2018). Even though some reports postulate a null or small effect of ATP on these conductances (Xu et al., 2005) one of the earliest proposition, the "metabolic hypothesis", states that CB oxygen sensing is mediated by oxidative phosphorylation, supported by the fact that virtually all metabolic poisons or oxidative phosphorylation inhibitors induce CB type-I cell excitation (Mulligan and Lahiri, 1982; Buerk et al., 1994; Wilson et al., 1994; Buckler and Turner, 2013). This response is normally triggered by potassium conductance(s) inhibition leading to the depolarization of the type-I cell membrane, reinforcing the notion that low PO<sup>2</sup> trigger potassium current blockage (**Figure 1**, see also Holmes et al., 2018). In this regard, it has been postulated that CB-mitochondria express a particular subtype of cytochrome a3 – the functional core of mitochondrion complex IV – which is at least seven times more sensitive to a drop in the PO2, further supporting a pivotal role of mitochondrial metabolism on CB-type-I cell excitation (for a full revision on this matter see Holmes et al., 2018 in this same research topic).

Therefore, in addition to changes in the expression and intrinsic function of ion channels, the fact that cellular metabolism is modified by CSH provides a potential link between CSH and the type-I cell membrane excitability. It is known that CSH increases the cytosolic AMP/ATP ratio in type-I cells. Thus, several enzymatic systems can be activated, including AMPK (Laderoute et al., 2006). Likewise, CSH induces an increase in cAMP levels in CB cells (Nurse et al., 1994). Protein kinase C (PKC) activity is enhanced by chronic hypoxia in several tissues (Zhang et al., 2004; Neckar et al., 2005), whereas the protein kinase A (PKA) activity decreases (Kobayashi et al., 1998). Increased PKC activity in response to acute hypoxia (Summers et al., 2000) was associated to the regulation of L-type Ca2+, maxi-K channels and background K<sup>+</sup> channels in type-I cells (Peers and Carpenter, 1998; Summers et al., 2000; Zhang et al., 2003). Since PKC inhibits TASK channel activity, whereas PKA enhanced it (Zilberberg et al., 2000; Besana et al., 2004) these pathways are well-suited as potential effectors of CSH inducedchanges.

### Activation of Hypoxia-Inducible Transcription Factors (HIFs)

The cellular long-term response to hypoxia is driven by changes in the expression of several genes to cope with the new hypoxic environment. These cellular transcriptional responses are largely dependent on the activation of the so-called hypoxiainducible transcription factors (Shimoda and Semenza, 2011). HIFs are composed by a HIFα subunit (mainly 1α and 2α), regulated by hypoxia and a constitutive β subunit (HIF-1β). Under normoxic conditions, HIFα is continuously degraded through hydroxylation modification catalyzed by a set of oxygen sensitive prolyl-hydroxylases enzymes. During hypoxia, the activity of prolyl-hydroxylases are dramatically reduced, allowing the translocation of HIFα to the cell nuclei where it forms an heterodimer with HIF-1β. Following translocation and dimerization, the HIF-1α/HIF-1β complex bind to hypoxia response elements (HRE) present in the DNA to regulate gene expression (Shimoda and Semenza, 2011). In the CB, the contribution of HIF-1α and HIF-2α has been largely studied in the setting of intermittent hypoxia; much less it is known in the context of CSH. Nevertheless, it has been shown that activation of the HIF-1α pathway promotes the up-regulation of pro-oxidant enzymes (i.e., NADPH) in the CB. On the contrary, intermittent hypoxia reduced the expression of HIF-2α within the CB, resulting in a significant down-regulation of antioxidant enzymes (i.e., SOD) levels being the outcome an increase ROS formation within CB glomus cells, increasing both intracellular Ca2<sup>+</sup> and neurotransmitter release (Prabhakar and Semenza, 2016). The specific contribution of HIF-1α and HIF-2α pathways on the sensitization of the CB to hypoxia under CSH remains to be fully determined. Nevertheless, it is possible to speculate that activation of HIF-1α and inhibition of HIF-2α pathways may occur in the CBs of animals exposed to sustained hypoxia.

We have here discussed some of the mechanisms explaining CB chemosensory hyperreactivity under CSH. This exacerbated activity has systemic consequences inducing the adaptation to CSH. In this regard, a central aspect is the CB-mediated autonomic function and even though the neural mechanism associated to autonomic responses to CSH are not completely elucidated, it has been described that CB type-I cells could be involved. During acute hypoxic exposure, excitatory inputs from the CB activate neural pathways in the brainstem increasing minute ventilation and sympathetic outflow (Barros et al., 2002; Braga et al., 2007). However, during CSH the baseline minute ventilation increases in a time-dependent manner (ventilatory acclimatization to hypoxia), probably related to the changes in oxygen sensitivity of the CB as well as the neural network in the brainstem related to sympathetic response to hypoxia (Powell et al., 2000; Powell, 2007). It has been proposed that hypoxia-mediated sympathetic activation act as a compensatory mechanism to increase oxygen supply to critical organs through regulating cardiac output and vascular conductance (Leuenberger et al., 1991; Calbet, 2003). Indeed, exposure of healthy humans to CSH produces a significant increase in circulating norepinephrine levels reflecting increases in sympathetic outflow during hypoxic exposure (Calbet, 2003; Hansen and Sander, 2003). In addition, it has been shown that rats subjected to 24 h of FiO<sup>2</sup> 10% displayed an increase of pre-sympathetic neurons activation located in the rostral aspect of the ventral medullary surface (RVLM). Furthermore, these set of neurons showed a further increase activation during the late expiratory phase during hypoxia stimulation suggesting a potential role on regulating the neural circuitry related to the respiratory-sympathetic coupling observed during CSH (Moraes et al., 2014). Interestingly, previous studies from our group showed that chronic activation of the CB afferent input to the brainstem using episodic hypoxic stimulation induces the activation of RVLM neurons (Del Rio et al., 2016). Therefore,

it is plausible to hypothesize that during CSH, activation of the CB may contribute to the regulation of sympathetic drive. This hypothesis remains to be determined.

### CONCLUSION

Potentiation of CB chemosensory responses is an absolute requirement for ventilatory acclimatization. How the increased response to acute hypoxia in chronically hypoxic CBs is achieved? There are probably multiple mechanisms underlying such phenomena. In the present review we focused on evidences suggesting that type-I cells excitability and metabolism undergoes substantial changes when exposed to CSH. Interestingly, ion channels, including some involved in the generation of the chemosensory response to acute hypoxia (i.e., maxi-K and TASK channels) are tightly regulated by PKC, PKA, AMPK, and ATP. Therefore, it is possible that throughout changes in cell metabolism, in addition to modifications on ion channels expression, CSH generates changes in type-I cells excitability, ultimate leading to increased depolarization during acute

### REFERENCES


hypoxia. Importantly, increased inhibition of TASK-like currents under chronic – sustained and intermittent – hypoxia provide a novel potential mechanism to explain the CB hyperreactivity (Ortiz et al., 2007, 2013). It seems clear that there is no a single mechanism as full explanation for CB hyperreactivity during CSH. Combined studies covering from cellular/molecular studies to functional approaches will help to understand the physiology of ventilatory acclimatization during chronic sustained hypoxia.

### AUTHOR CONTRIBUTIONS

All authors contributed to writing the article. RV, RI, RDR, and FO designed and edited the paper.

### FUNDING

This work was supported by grants FONDECYT 11160616 to FO, FONDECYT 1180172 to RDR, and FONDECYT 1150040 to RI.




**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Pulgar-Sepúlveda, Varas, Iturriaga, Del Rio and Ortiz. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Daily Intermittent Normobaric Hypoxia Over 2 Weeks Reduces BDNF Plasma Levels in Young Adults – A Randomized Controlled Feasibility Study

Andreas Becke1,2† , Patrick Müller<sup>2</sup> \* † , Milos Dordevic<sup>2</sup> , Volkmar Lessmann3,4 , Tanja Brigadski3,5 and Notger G. Müller2,4

1 Institute of Cognitive Neurology and Dementia Research, Otto-von-Guericke-Universität Magdeburg, Magdeburg, Germany, <sup>2</sup> Neuroprotection Laboratory, German Center for Neurodegenerative Diseases (DZNE), Magdeburg, Germany, 3 Institute of Physiology, Otto-von-Guericke-Universität Magdeburg, Magdeburg, Germany, <sup>4</sup> Center for Behavioral Brain Sciences, Magdeburg, Germany, <sup>5</sup> Informatics and Microsystem Technology, University of Applied Sciences, Kaiserslautern, Kaiserslautern, Germany

### Edited by:

Jean-Paul R-Richalet, Université Paris 13, France

### Reviewed by:

Akira Yoshii, University of Illinois at Chicago, United States Melissa L. Bates, University of Iowa, United States

### \*Correspondence:

Patrick Müller patrick.mueller@dzne.de

†These authors have contributed equally to this work

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 18 June 2018 Accepted: 04 September 2018 Published: 01 October 2018

### Citation:

Becke A, Müller P, Dordevic M, Lessmann V, Brigadski T and Müller NG (2018) Daily Intermittent Normobaric Hypoxia Over 2 Weeks Reduces BDNF Plasma Levels in Young Adults – A Randomized Controlled Feasibility Study. Front. Physiol. 9:1337. doi: 10.3389/fphys.2018.01337 Background: The results from animal and human research indicate that acute intermittent hypoxia can enhance brain-derived neurotrophic factor (BDNF) plasma levels and gene expression. As BDNF is known to promote the differentiation of new neurons and the formation of synapses, it has been proposed to mediate adult neuroplasticity. Thus, the present study aimed to analyze the long-term effects of daily intermittent exposure to normobaric hypoxia (simulating high altitude exposure at approximately 4000–5000 m) over 2 weeks on BDNF levels in young adults.

Methods: Twenty-eight young adults (age: 19–33 years) were randomized into a hypoxic intervention group (N = 14) or the control group (N = 14). Participants in the intervention group breathed intermittent normobaric hypoxic air at resting conditions (5 min intervals, 80–85% SpO<sup>2</sup> measured via a finger pulse oximeter, 12 sessions for 60 min/day for 2 weeks) via a hypoxic generator. BDNF plasma and serum levels were determined at baseline and at 2 weeks after intervention using sandwich ELISAs.

Results: After 2 weeks of daily intermittent hypoxic treatment (IHT), we found a significant group x time interaction effect for BDNF plasma levels based on a significant decrease in BDNF levels in the hypoxia group.

Conclusion: Our results demonstrate that daily intermittent administration of hypoxic air has a significant effect on BDNF regulation in healthy young adults. Contrary to other results reporting an increase in BDNF levels under hypoxic conditions, the present data suggest that hypoxic treatment using intensive IHT can reduce BDNF plasma levels for at least 2 weeks. This finding indicates that the daily application of hypoxic air is too frequent for the aimed physiological response, namely, an increase in BDNF levels.

Keywords: hypoxia, BDNF, neuroplasticity, IHT, adaptation

# INTRODUCTION

fphys-09-01337 September 27, 2018 Time: 16:29 # 2

Hypoxia is defined by a reduced oxygen content in air and can be divided into intermittent and chronic forms. Thereby, intermittent hypoxia applies to a large spectrum of stimuli that range from exercise in high altitude to obstructive sleep apnea (OSA). Intermittent hypoxia treatment (IHT) was first used in sports medicine to enhance human physical performance (erythropoiesis and angiogenesis) (Viscor et al., 2018). During the following years, hypoxic training was increasingly employed for non-pharmacological treatment of several diseases (e.g., bronchial asthma, hypertension, and cardiovascular diseases). IHT can effectively stimulate various metabolic processes (Serebrovskaya et al., 2008) and can have numerous positive health effects similar to cardiovascular physical activity (Enette et al., 2017). IHT may serve as a protective mechanism for the brain by inducing neurogenesis. For instance, histological studies in adult rats have shown that IHT promotes a transient increase in progenitor cell proliferation in the subventricular zone and a long-term increase in the dentate gyrus (Zhu et al., 2005) and has the potential to recover spatial learning deficits after cerebral ischemia by increased hippocampal neurogenesis (Tsai et al., 2011). However, intermittent normobaric hypoxia is not associated with positive effects only per se. For example, the clinical syndrome of OSA leads to intermittent hypoxia as well (Burtscher et al., 2009) and is associated with numerous negative effects such as reduced cognitive performance (Yan, 2014; Malle et al., 2016). Hence, based on different characteristics such as the dose and the duration, we can assume that hypoxia induces both protective and pathological effects. It has been proposed that low-dose intermittent hypoxia (9–16% inspired O2) with short durations can enhance positive physiological processes, whereby high-dose hypoxia (2–8% inspired O2) is associated with progressively pathological mechanisms (Navarrete-Opazo and Mitchell, 2014).

The results from animal and human research indicate that acute intermittent hypoxia (Vermehren-Schmaedick et al., 2012) and physical activity (Enette et al., 2017) can enhance brainderived neurotrophic factor (BDNF) blood levels and BDNF gene expression. Such gene expression is explained by an oxygen deficit recognized by the oxygen sensory system (Sharp and Bernaudin, 2004) changing the oxygen-dependent degradation domain of hypoxia-inducible factor (HIF-1), thereby inducing an increase in HIF-1-alpha levels (Wiener et al., 1996). HIF-1-alpha is known to act as a transcription factor to modulate the expression of several genes, such as BDNF growth factor levels (Helan et al., 2014). The BDNF neurotrophin is a member of the nerve growth factor family and is widely expressed in the human brain, especially in the hippocampus, but it is also expressed in peripheral tissues such as the pulmonary vasculature (Aravamudan et al., 2012; Helan et al., 2014). Current research studies indicate BNDF plasma levels as a potential biomarker for reliable diagnosis of neurocognitive disorders (Levada et al., 2016). The protein is secreted in an activitydependent manner but is also secreted in response to hypoxia (Haubensak et al., 1998; Hartmann et al., 2001; Kohara et al., 2001; Brigadski et al., 2005; Matsuda et al., 2009; Brigadski and Leßmann, 2014; Helan et al., 2014; Edelmann et al., 2015; Hartman et al., 2015). Research results indicate that 75% of the BDNF in the peripheral blood plasma originates from the brain (Krabbe et al., 2007; Rasmussen et al., 2009). Several studies have suggested that BDNF is an important modulator of the CNS and promotes the differentiation of new neurons and synapses (Huang and Reichardt, 2001; Leschik et al., 2013; Park and Poo, 2013; Edelmann et al., 2014). BDNF, therefore, represents one of the major mediators of neuroplasticity (Calabrese et al., 2014). Furthermore, some authors have suggested that BDNF blood levels may serve as a biomarker for the diagnosis of neurodegenerative diseases and psychiatric disorders and can also serve as a surrogate marker for the success of therapies in these disorders (Ruscheweyh et al., 2011). Reduced BDNF blood levels have been reported in Alzheimer's disease (Laske et al., 2007) and mild cognitive impairment (Forlenza et al., 2010).

Regarding the effect of intermittent hypoxia on BDNF blood levels in humans, the status of research is currently unclear. The results from animal and human studies have shown an acute increase in BDNF plasma levels in response to hypoxia. Helan et al. (2014) observed an increase in BDNF levels in 30 healthy volunteers after 72 h of normobaric hypoxia. Schega et al. (2016) reported no effects on BDNF in serum in older adults (N = 34, 66.4 ± 3.3 years) after 4 weeks of intermittent normobaric hypoxia (3× per week for 90 min) in addition to cardiovascular exercise. However, their data indicated that BDNF levels increased in the exercise-intervention group and in the exercise control-group after a compensation period of several weeks. This finding raises the question of whether the delayed effect could have been observed after hypoxic treatment alone, i.e., without concomitant cardiovascular exercise intervention.

Previous studies in animal research indicate an occurrence of neurogenesis in dentate gyrus within 4 weeks subsequent to intermittent hypoxia (Zhu et al., 2005). Based on these results we conducted a feasibility study to test the effects of 2 weeks of daily exposure to hypoxic air, which simulated intermittent hypoxia treatment (IHT), on peripheral BDNF levels. Therefore, we expected an increase in BDNF levels (as a central mediator of neurogenesis).

With respect to previous research on passive IHT methods, a protocol was chosen that has been shown to increase aerobic capacity and exercise tolerance in elderly men (Burtscher et al., 2004). In view of the data from Zhu et al. (2005) and based on recommendations for IHT regimes (Bassovitch and Serebrovskaya, 2009), we estimated the peak long-term effects of IHT to emerge 2 weeks after the intervention. If successful, this process is an easy to administer, low-cost intervention that may have great potential in inducing neuroplasticity and preventing cognitive deficits.

### MATERIALS AND METHODS

The study was designed as a two-week randomized, controlled intervention. The ethics committee at the Otto-von-Guericke-Universität Magdeburg, approved the study, and all of the subjects signed a written informed consent form prior to

participation. The exclusion criteria were acute or chronic cardiovascular, renal, metabolic, orthopedic and/or neurological diseases.

Twenty-eight young adults (age: 19–33 years) were randomized to a hypoxic intervention group [N = 14 (9 female), mean age 27.78, SD = 2.39] or a control group [N = 14 (5 female), mean age 22.85, SD 2.35] using the website www.randomization.com. The participants in the intervention group breathed intermittent normobaric hypoxic air at resting conditions (5 min intervals at a target of 80–85% SpO<sup>2</sup> via a finger pulse oximeter, 12 sessions for 60 min/day for 2 weeks) generated by a hypoxic generator (b-cat and integra ten). The simulated high altitude was continuously manually adjusted between 4000 and 5000 m to reach the target SpO2. The control group received no intervention.

Fasting blood samples were taken in the mornings at baseline and at posttest (2 weeks after the last training session). From the blood samples, the plasma and serum concentrations of BDNF were determined using sandwich ELISAs (BDNF DuoSet; R&D Systems, Wiesbaden, Germany) as previously described (Schega et al., 2016).

For the intervention group, the blood samples for the small blood count were taken 4 times at baseline, 1 week after the intervention, at the end of intervention (consecutive day of last intervention session) and 2 weeks after the intervention. Five missing data sets for the second time point and 3 missing data sets for the third time point were reported (subjects did not show up).

Statistical analysis of BDNF plasma levels, BDNF serum levels and small blood count levels were performed with SPSS (SPSS 22 Inc./IBM). The intervention effects for BDNF were tested using repeated-measures ANOVAs with group (IHT and CG) as the between-subject factor and time (pre and post) as the withinsubject factor. Age and gender were included as covariates. Additionally, post hoc pairwise comparisons were performed to determine the longitudinal changes in the hypoxia and control groups separately. In the case of non-normal distribution of data, we used the Mann-Whitney U-test or the Wilcoxon test instead of t-tests. The effect size was quantified by partial eta squared (η 2 ). For interaction effects, the percentage changes from baseline to post measures were calculated for BDNF and small blood count values and were then correlated with Pearson's formula.

### RESULTS

The plasma and serum levels of BDNF were analyzed in the blood samples before the onset of the intervention as well as after the intervention. A significant group x time interaction effect was observed for the BDNF plasma levels [F(1,26) = 10.742, p = 0.002, η <sup>2</sup> = 0.292]. Post hoc pairwise comparisons showed a significant decrease in BDNF plasma levels only in the hypoxia group from baseline to the posttest period (Wilcoxon-Test, Z = −3.296, p = 0.001). The intraindividual changes in BDNF plasma levels reached a reduction of 66.34% of the pretreatment period. No significant time × group interactions emerged for BDNF serum levels (see **Table 1** and **Figure 1**).

For the intervention group, blood samples for a small blood count were collected 4 times at baseline, 1 week after the intervention, at the end of the intervention (consecutive day of the last intervention session) and 2 weeks after the intervention. Five missing data sets for the second time point and 3 missing data sets for the third time point were reported (subjects did not show up). Using mixed linear effects to model the effect over time, the red blood cell distribution showed a linear decrease over time (p < 0.01; **Table 2**).

Furthermore, an analysis of Pearson correlations between the baseline to post measure changes (%) revealed a close to significant positive correlation (one-tailed) for BDNF plasma and leucocyte counts (RWBC = 0.446; p = 0.055) and a trend for a negative BDNF and lymphocyte interaction (Rlym = −0.374; p = 0.094).

# DISCUSSION

Normobaric hypoxia such as with high-altitude training is generally assumed to have positive effects on physical and cognitive performance. Here, we tested the effect of a daily intermittent normobaric hypoxic training during a period of 2 weeks on the BDNF levels. While we observed the expected effects on blood parameters such as on the mean corpuscular hemoglobin concentration, contrary to our expectation, we found BDNF plasma levels to be significantly reduced 2 weeks after daily intermittent normobaric hypoxia over a period of 2 weeks. Regarding BDNF serum levels, no changes were detected. Research results from Pan et al. (1998) indicate that BDNF can pass the blood brain barrier by a high-capacity, saturable transport system and that 75% of BDNF plasma levels stems from the brain (Krabbe et al., 2007; Rasmussen et al., 2009).

Decreased BDNF levels are typically found in animal research when the animals have previously experienced stress. Various types of stress, including oxidative stress, have been shown to lead to decreased BDNF gene expression in cortical regions, including the hippocampus (Smith, 1996; Smith and Cizza, 1996; Bath et al., 2013; Kwon et al., 2013; Rothman and Mattson, 2013). In humans, a reduction in BDNF levels was seen after muscle damage or with very intensive physical exercise. To avoid such overtraining, successful exercise training is known to require sufficient resting periods (Parra et al., 2000). In rodents, physical activity induces BDNF gene expression in cortical regions, especially in the hippocampus (Neeper et al., 1995; Uysal et al., 2015). Studies on humans have reported an increase in BDNF levels following sportive interventions (Erickson et al., 2012; Müller et al., 2017a,b; Rehfeld et al., 2018). Others, however, failed to show changes in the levels of any of the neurotrophic factors that were assessed (Maass et al., 2016). A current review by Enette et al. (2017) provides a comprehensive analysis of the effects of aerobic training on BDNF plasma and serum levels in older adults. In 11 of the 14 randomized controlled trials that were included, the authors reported significantly increased BDNF plasma and/or serum levels after aerobic intervention.

Together, these findings indicate that our IHT protocol with its daily applications of hypoxic air might have been too intensive

### TABLE 1 | Statistics of rANOVA on BDNF plasma and serum levels.

fphys-09-01337 September 27, 2018 Time: 16:29 # 4


∗∗p < 0.01.

### TABLE 2 | Small blood count for the treatment group.


The mean values and SD for the RBCC, red blood cell count; RDW, red blood cell distribution width; HCT, hematocrit; HBG, hemoglobin; hypEryth, hypochromic erythrocytes; WBC, leukocyte count; MCH, lymphocytes, mean corpuscular/cellular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; Mono, absolute monocyte count; Segs, segmented neutrophil granulocytes; SumGranul, sum granulocytes, thrombocytes. ∗∗p < 0.01 <sup>∗</sup>p < 0.05 <sup>t</sup>p < 0.09.

and, therefore, too stressful for the participants' bodies. In agreement with this finding, we observed a change in blood marker levels that were indicative of inflammation, namely, lymphocytes and granulocytes. Intensive physical exercise also induces inflammatory processes (Brown et al., 2015), and the latter has also been shown to relate to reduced BDNF levels after

acute exercise at higher intensities (Nofuji et al., 2012; Cabral-Santos et al., 2016). Other conditions in which a reduction in BDNF levels was observed in the past include sleep apnea (Wang et al., 2012), birth stress associated with psychiatric disease later in life (Cannon et al., 2008), and stroke with low functional outcome (Lasek-Bal et al., 2015). With respect to the present study, the results of Wang et al. (2017) are of special relevance, as sleep apnea is associated with nocturnal intermittent hypoxia. Again, this finding suggests that "overdosing" hypoxia has detrimental effects on BDNF secretion.

The assumption that our IHT protocol was too intense and therefore decreased BDNF levels leads to the crucial question of whether other less stressful IHT protocols could still have a positive effect. In addition, methodological aspects (sampling time and preanalytical variations) could have an influence on the gained results. Indeed, there is an ongoing discussion of what type of hypoxia treatment is most effective (Serebrovskaya and Xi, 2016). A protocol that increases physical fitness at the same time may have negative effects on BDNF (Enette et al., 2017). Indeed, we had used a protocol that, in a former study, had shown positive effects on aerobic capacity.

### Metabolic and Cardiovascular Response to Hypoxia

Several field experiments in the mountains and environmental studies in chambers report physiological effects of hypoxia (Heinonen et al., 2016). These experiments show that hypoxia can induce cardiovascular stress, can increase sympathetic neural activation and can alter energy metabolism. The complex metabolic response causes a release of various stress hormones (Kayser and Verges, 2013). Regarding cardiovascular response to normobaric hypoxia Heinonen et al. (2014) reported a significantly increased cardiac output, ejection fraction and tachycardia. Additionally, Heinonen et al. (2014, 2017) discuss hypoxia as a potential trigger for the release of brain natriuretic peptide (BNP) and the hormone apelin.

### Limitations and Outlook

This randomized controlled feasibility study has several limitations. First, the sample size was small (N = 28). Second, the

# REFERENCES


blood samples were only analyzed at baseline and after 2 weeks of intervention. Another limiting factor in the BDNF blood analyses is the large variances.

Future studies are needed to evaluate the correct dose of normobaric intermittent hypoxia to increase BDNF plasma levels and examine the underlying neurobiological mechanisms. An intensive assessment (neuropsychology, MRI/PET, cortisol, and IGF-1) would be useful to analyze the physiological adaptations to hypoxia.

In addition, BDNF has been suggested to play a mediating role in schizophrenia (Sokoloff et al., 2004; Guillin et al., 2007). Thus, several studies indicate an increase of the BDNF levels and gene expression in patients with schizophrenia (Laske and Eschweiler, 2006). In conclusion, an intermittent normobaric hypoxia regimen that successfully increases the BDNF levels may offer a non-pharmacological treatment to patients with schizophrenia.

### ETHICS STATEMENT

This study was carried out in accordance with the recommendations of Ethics Committee of the Medical Faculty at the Otto-von-Guericke-Universität Magdeburg with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Ethics Committee of the Medical Faculty at the Otto-von-Guericke-Universität Magdeburg.

### AUTHOR CONTRIBUTIONS

AB contributed to study organization, data analysis, paper writing, and paper reviewing. PM contributed to data analysis, paper writing, and paper reviewing. MD reviewed the paper. VL and TB contributed to data analysis and paper reviewing. NM contributed to study organization, paper writing, and paper reviewing.


high-intensity intermittent exercise: effect the exercise volume. Front. Physiol. 7:509. doi: 10.3389/fphys.2016.00509



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Becke, Müller, Dordevic, Lessmann, Brigadski and Müller. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Effects of Plyometric Training on Explosive and Endurance Performance at Sea Level and at High Altitude

David Cristóbal Andrade1,2, Ana Rosa Beltrán<sup>3</sup> , Cristian Labarca-Valenzuela<sup>4</sup> , Oscar Manzo-Botarelli<sup>4</sup> , Erwin Trujillo<sup>4</sup> , Patricio Otero-Farias<sup>3</sup> , Cristian Álvarez<sup>5</sup> , Antonio Garcia-Hermoso<sup>6</sup> , Camilo Toledo<sup>1</sup> , Rodrigo Del Rio1,7,8, Juan Silva-Urra<sup>4</sup> and Rodrigo Ramírez-Campillo<sup>5</sup> \*

<sup>1</sup> Laboratory of Cardiorespiratory Control, Faculty of Physiological Science, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>2</sup> Centro de Investigación en Fisiología del Ejercicio, Facultad de Ciencias, Universidad Mayor, Santiago, Chile, <sup>3</sup> Departamento de Educación, Facultad de Educación, Universidad de Antofagasta, Antofagasta, Chile, <sup>4</sup> Departamento Biomédico, Centro Investigación en Fisiología y Medicina de Altura, Universidad de Antofagasta, Antofagasta, Chile, <sup>5</sup> Department of Physical Activity Sciences, Research Nucleus in Health, Physical Activity and Sport, Quality of Life and Wellness Research Group, Universidad de Los Lagos, Osorno, Chile, <sup>6</sup> Laboratorio de Ciencias de la Actividad Física, el Deporte y la Salud, Universidad de Santiago de Chile, Santiago, Chile, <sup>7</sup> Centro de Excelencia en Biomedicina de Magallanes, Universidad de Magallanes, Punta Arenas, Chile, <sup>8</sup> Centro de Envejecimiento y Regeneración, Pontificia Universidad Católica de Chile, Santiago, Chile

### Edited by:

Lorenza Pratali, Istituto di Fisiologia Clinica (IFC), Italy

### Reviewed by:

Stephane Perrey, Université de Montpellier, France Simona Mrakic-Sposta, Istituto di Bioimmagini e Fisiologia Molecolare (IBFM), Italy

> \*Correspondence: Rodrigo Ramirez-Campillo

r.ramirez@ulagos.cl

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 01 December 2017 Accepted: 18 September 2018 Published: 09 October 2018

### Citation:

Andrade DC, Beltrán AR, Labarca-Valenzuela C, Manzo-Botarelli O, Trujillo E, Otero-Farias P, Álvarez C, Garcia-Hermoso A, Toledo C, Del Rio R, Silva-Urra J and Ramírez-Campillo R (2018) Effects of Plyometric Training on Explosive and Endurance Performance at Sea Level and at High Altitude. Front. Physiol. 9:1415. doi: 10.3389/fphys.2018.01415 Plyometric training performed at sea level enhance explosive and endurance performance at sea level. However, its effects on explosive and endurance performance at high altitude had not been studied. Therefore, the aim of this study was to determine the effects of a sea level short-term (i.e., 4-week) plyometric training program on explosive and endurance performance at sea level and at high altitude (i.e., 3,270 m above sea level). Participants were randomly assigned to a control group (n = 12) and a plyometric training group (n = 11). Neuromuscular (reactive strength index – RSI) and endurance (2-km time-trial; running economy [RE]; maximal oxygen uptake - VO2max) measurements were performed at sea level before, at sea level after intervention (SL +4 week), and at high altitude 24-h post SL +4 week. The ANOVA revealed that at SL +4 week the VO2max was not significantly changed in any group, although RE, RSI and 2-km time trial were significantly (p < 0.05) improved in the plyometric training group. After training, when both groups were exposed to high altitude, participants from the plyometric training group showed a greater RSI (p < 0.05) and were able to maintain their 2-km time trial (11.3 ± 0.5 min vs. 10.7 ± 0.6 min) compared to their pre-training sea level performance. In contrast, the control group showed no improvement in RSI, with a worse 2-km time trial performance (10.3 ± 0.8 min vs. 9.02 ± 0.64 min; p < 0.05; ES = 0.13). Moreover, after training, both at sea level and at high altitude the plyometric training group demonstrated a greater (p < 0.05) RSI and 2-km time trial performance compared to the control group. The oxygen saturation was significantly decreased after acute exposure to high altitude in the two groups (p < 0.05). These results confirm the

**309**

beneficial effects of sea level short-term plyometric training on explosive and endurance performance at sea level. Moreover, current results indicates that plyometric training may also be of value for endurance athletes performing after an acute exposure to high altitude.

Keywords: reactive strength, jump height, hypoxia, endurance performance, explosive performance, stretchshortening cycle, elastic energy

### INTRODUCTION

fphys-09-01415 October 6, 2018 Time: 15:43 # 2

Endurance performance depends on several aerobic factors (Coyle, 1995), like maximal oxygen uptake (VO2max) and running economy (RE) (Shaw et al., 2014), that is the energy expenditure at different velocities (Saunders et al., 2004a). However, endurance performance may also depend on neuromuscular characteristics like reactive strength index (RSI), muscle strength, stiffness, among others (Noakes, 1988; Sinnett et al., 2001). In fact, some aerobic endurance determinants like RE (Conley and Krahenbuhl, 1980; Saunders et al., 2004a) can be affected by neuromuscular variables (Turner et al., 2003). More so, neuromuscular performance (i.e., jump-related explosive muscle actions) has been related with endurance performance at different distances (Hudgins et al., 2013). Thus, the energy cost of running reflects the sum of both aerobic and anaerobic (neuromuscular factors) metabolism (Daniels, 1985). Hence, training strategies that can increase both aerobic and neuromuscular factors related with endurance performance would be of great value for endurance athletes.

Plyometric training (i.e., a jump-based strength training method) (Markovic and Mikulic, 2010) is a commonly used training strategy to increase neuromuscular strength by means of stretch-shortening cycle muscle actions (Ramirez-Campillo et al., 2013) and may positively affect aerobic-related endurance performance variables (i.e., RE) (Conley and Krahenbuhl, 1980; Saunders et al., 2004a), possible through increasing RSI (Paavolainen et al., 1999). In fact, a previous study reported that after a 6-week period of plyometric training recreational runners improved their RE (Turner et al., 2003) and this increase also can be expected in highly trained middle and long distance runners (Saunders et al., 2006). Moreover, plyometric training can increase RE independently from VO2max changes (Paavolainen et al., 1999). This phenomenon is important in highly trained endurance athletes due to their limited ability to increase VO2max (Midgley et al., 2007). Furthermore, plyometric training has a positive effect on time trial performance (Paavolainen et al., 1999) and endurance athletes with better performance may achieve better adaptive responses to plyometric training (Ramirez-Campillo et al., 2014a). Therefore, plyometric training may increase RE and RSI, which may positively affect endurance performance (Pollock, 1977; Thomas et al., 1999). However, whether the positive effects of plyometric training on aerobic and neuromuscular performance variables are still present after short term exposure at high altitude (HA) remain unexplored.

Acute exposure (<24 h) to high altitude decreases VO2max (Maher et al., 1974; Young et al., 1996; Bassett and Howley, 2000) and endurance performance (Maher et al., 1974; Fulco et al., 1998). Specifically, in well trained subjects, small but significant aerobic performance impairments may occur even at an altitude of ∼540 m, with reductions in aerobic performance up to ∼35% at higher altitudes (Fulco et al., 1998). This phenomenon can be explained by a decrease in partial pressure and arterial saturation in O<sup>2</sup> with a lower barometric pressure in HA (Wagner, 2010), negatively affecting VO2max. Additionally, extrapolation of data showed that VO2max fell 0.9% per every 100 m above altitude ≥1,100 m (Vogt and Hoppeler, 2010). However, neuromuscular variables related to fitness performance (i.e., jump) may not be negatively affected by acute exposure to high altitude (Coudert, 1992; Kayser et al., 1993; Burtscher et al., 2006). In fact, neuromuscular performance variables may even be favorably affected (Wood et al., 2006). For example, in the Olympic Games of 1968, held at 2,300 m above sea level, several world records in maximal-intensity and short-durations events were improved, such as the long jump (Berthelot et al., 2015). Moreover, among elite athletes, jumping performance is usually improved above 1,500 m above sea level (Hamlin et al., 2015). In a recent study (Garcia-Ramos et al., 2018), young male and female swimmers (age, ∼19 years) were acutely exposed to 2,320 m above sea level, and their explosive performance during a loaded jump-task was increased when compared to sealevel performance in terms of maximal velocity (up to 7.6%), maximal power (up to 6.8%) and peak force (up to 3.6%). Similar findings were also reported in physically active subjects after performing repeated jumps under high-hypoxic conditions (Alvarez-Herms et al., 2015). As endurance performance depends on both aerobic and neuromuscular variables (Daniels, 1985), it is possible that the endurance performance can be benefited per neuromuscular variables on high altitude. Therefore, this study was aimed to evaluate if a short-term (i.e., 4 weeks) sea level plyometric training program could affect explosive and endurance performance on sea level and at high altitude (i.e., 3,270 m), through its positive effects on neuromuscular performance variables (i.e., RSI) and aerobic determinants of endurance performance (i.e., 2 km time trial test) at sea level (Turner et al., 2003), independently from changes in VO2max (Paavolainen et al., 1999).

### MATERIALS AND METHODS

### Participants

A group of participants was submitted to neuromuscular (30 cm drop jump reactive strength index [RSI30]) and endurance (2-km time-trial; running economy; VO2max) measurements at sea level before (SL) and after intervention (SL +4 week) and again (RSI30; 2-km time-trial) after 24-h post SL +4 week at high altitude (HA +4 week).

Physically active (i.e., recreational runners) lowlanders participants (sixteen males and seven females; height: 162.6 ± 9.4 cm; body mass: 62.2 ± 12.4 kg; age 21.3 ± 1.3 years) were recruited, and randomly assigned to a control group (n = 12, 8 males, and 4 females) and plyometric training group (n = 11, 8 males, and 3 females). Although not matched for any specific dependent variable, all measurements taken at baseline (at sea level and at high altitude) in both the control and the plyometric training group were homogeneous.

None of the participants had any background (in the 6 month period preceding the study) in regular strength training or competitive sports activity that involved any kind of jumping training exercise used during the treatment. Sample size was computed based on the changes observed in the reactive strength index (1 = 0.33 mm·ms−<sup>1</sup> ; SD = 0.3) after a short-term plyometric training study (Ramirez-Campillo et al., 2014b). Exclusion criteria considered (i) potential medical problems or a history of ankle, knee, or back injury, (ii) any lower extremity reconstructive surgery in the past two years or unresolved musculoskeletal disorders, (iii) history of acute mountain sickness, (iv) previous (≤2 months) pre-acclimation at high altitude exposure (3,000 m above sea level). Additionally, as physical performance at high altitude was a major dependent variable, participants were excluded from the study if they experienced acute mountain sickness during exposure to high altitude.

Institutional review board approval for our study was obtained from the ethical committee of the Universidad de Antofagasta. All participants were carefully informed about the experiment procedures, and about the possible risk and benefits associated with their participation in the study, and an appropriate signed informed consent document has been obtained in accordance with the Declaration of Helsinki. We comply with the human and animal experimentation policy statements guidelines of the American College of Sports Medicine.

### Experimental Design

A schematic representation of the experimental study design is depicted in **Figure 1**.

The participants were carefully familiarized with the tests procedures during several submaximal and maximal exercises before the measurements were taken. The participants also completed several explosive trials to become familiar with the exercises used during training. In addition, several warm-up muscle actions were performed prior to the actual maximal and explosive test actions (Andrade et al., 2015). Tests were always administered in the same order, time of day and by the same investigator, on non-consecutive days before and after (i.e., ≥48 h after last training session) the 4 weeks of intervention. Participants were instructed to (i) have a good night's sleep (≥8 h) before each testing day, (ii) have a meal rich in carbohydrates and to be well hydrated before measurements, (iii) use the same athletic shoes and clothes during the pre- and post-intervention testing, (iv) give their maximum effort during the performance measurements. On day one height, body mass, RSI30 and 2-km time trial test were measured. On day two, RE and VO2max were measured. After the second day of measurements, participants took an autobus during 5 h to ascended at 3,270 m above sea level (Caspana city, Antofagasta Region, Chile), spending 24 h at this elevation before measurements of RSI30 and 2-km time trial tests (HA +4 week). Considering that there are relatively few published studies regarding the acute effects of high altitude on participants that ascent to high altitude after plyometric training (Khodaee et al., 2016), it was deemed of relevance to conduct measurements after an acute exposure of 24 h at 3,270 m above sea level.

Standard warm-up (i.e., 5 min of submaximal running with several displacements, 20 vertical and 10 longitudinal jumps) was executed before each testing day. In addition, height was measured using a wall-mounted stadiometer (HR-200, Tanita, Japan) recorded to the nearest 0.5 cm. Body mass was measured to the nearest 0.1 kg using a digital scale (BF-350, Tanita, Illinois, United States).

Additionally, in order to discard participants suffering acute mountain sickness, at SL +4 week and HA +4 week a self-administered questionnaire from Lake Louise Acute Mountain Sickness Scoring system (Roach, 1993) was applied to participants. A score ≥3 points discarded the participation from the study (Roach, 1993). To verify physiological changes of acute exposure to hypoxia, systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial blood pressure (MABP), heart rate (HR), and oxygen saturation (SpO2) were measured at SL +4 week and at HA +4 week. According to the Lake Louise acute mountain sickness survey, and physiological recordings, none of the participants suffer acute mountain sickness (**Table 1**). Therefore, all subjects were considered eligible for inclusion in this study.

### RSI30 Test

A detailed description of this test is reported elsewhere (Ramirez-Campillo et al., 2013). Briefly, participants performed drop jumps from a 30 cm height platform, using an electronic contact mat system (Globus Tester, Codogne, Italy). The participants were instructed to place their hands on their hip and step off the platform with the leading leg straight to avoid any initial upward propulsion, ensuring a drop height of 30 cm. They were instructed to jump for maximal height and minimal contact time, in order to maximize jump reactive strength. The participants were again instructed to leave the platform with knees and ankles fully extended and to land in a similarly extended position to ensure the validity of the test. Three repetitions were executed, with at least 2 min of rest between them. The best performance trial was used for the subsequent statistical analysis.

### Time Trial Test

This was the only test performed outdoors. At sea level the relative humidity (i.e., between 66 and 68%) and temperature (i.e., between 19 and 21◦C) was similar at baseline and after training intervention. Furthermore, at high altitude, the relative humidity was between 21 and 25% and temperature was 19◦C (Chilean Meteorological Service). Participants ran 28.6 laps in a

FIGURE 1 | Schematic representation of the study experimental design. RSI30, 30 cm drop jump reactive strength index; RE, running economy; VO2max, maximal oxygen uptake.

TABLE 1 | Effect of high altitude (3,270 m above sea level) on physiological parameters and Lake Louise acute mountain sickness survey.


Data are present as mean ± SD. SBP, systolic blood pressure; DBP, diastolic blood pressure; MABP, mean arterial blood pressure; HR, heart rate; SpO2, oxygen saturation. <sup>∗</sup>p < 0.05 vs. Sea level.

70-m outdoor concrete track for a total of 2 km. This track was the same at sea level and high altitude. Aside from the standard warm-up, before the time-trial test, participants completed two submaximal laps around the track and 4 min later, they had one maximal attempt to complete the test.

# VO2max

The VO2max was measured during an incremental test on a treadmill (LifeFitness, model 95Te, United States) to volitional exhaustion, as previously described (Saunders et al., 2004b,c). Briefly, the velocity was increased in 0.5 km.h−<sup>1</sup> every 30 s from 8.0 km.h−<sup>1</sup> to 12.0 km.h−<sup>1</sup> , then the inclination of the treadmill was increased 0.5% every 30 s until volitional fatigue. Gas exchange was recorded continuously with a portable breathto-breath gas analyzer (K4b2, Cosmed, Italy). The analyzer was calibrated according to the manufacturer's instructions prior to each trial run. Pulmonary ventilation (VE), oxygen uptake (VO2), expired carbon dioxide (VCO2), and respiratory exchange ratio (RER) were averaged over 10 s periods, with the highest 30 s value (i.e., three consecutive 10 s periods) used in the analysis. VO2max was determined according to achievement of previously established criteria (Howley et al., 1995): (i) VO<sup>2</sup> plateau (increase <150 ml·min−<sup>1</sup> ), (ii) RER >1.1, and (iii) ≥90% of theoretical maximal heart rate (HRmax). The VO2max was expressed relative to body mass (ml·kg−<sup>1</sup> ·min−<sup>1</sup> ).

### Running Economy

Running economy (RE) was determined by measuring VO<sup>2</sup> at three different sub-maximal velocities (i.e., 10.0, 11.0, and 12.0 km·h −1 ), on a treadmill (LifeFitness, model 95Te, United States), with 0% of inclination. At each sub-maximal velocity, a 4-min-collection period was employed. Running economy was defined as the mean VO<sup>2</sup> attained during the last minute of each running speed. Four minutes was deemed as an adequate time frame to reach steady state (Saunders et al., 2004b,c; Saunders et al., 2006). The RE was expressed as ml·kg−<sup>1</sup> ·min−<sup>1</sup> and relative to VO2max (%) values obtained before and after the intervention to avoid potential confounding factors due to potential pre-post changes in VO2max values.

### Training Program

The plyometric training was completed at sea level during 4 weeks, 3 days per week (i.e., ≥48 h of rest between sessions). Each session lasted ∼25 min. The same standard warm-up described above (i.e., testing procedures) was used prior to the main part of the training session. Participants completed bounce drop jumps drills from 30 cm to 50 cm boxes with the same technique (and instructions) previously described for the RSI30 test, intended to maximize reactive strength. Participants completed a total of 60 foot contacts per session (i.e., three sets of 10 jumps from each box). This volume has

been used in previous studies, obtaining significant benefits (Ramirez-Campillo et al., 2014a). The rest period between repetitions and sets was 15 s (Read and Cisar, 2001) and 2 min, respectively. The same researcher was always present during training sessions, motivating participants to give their maximum effort in each jump. Participants were reminded to maintain their usual physical activity and nutritional habits during the experiment.

During the intervention, participants maintained their habitual running training (i.e., 3–4 sessions per week, 30–60 min per session, at 70–80% of maximum heart rate). During the intervention participants completed a total volume load of 720 jumps.

### Statistical Analyses

All values are reported as mean ± standard deviation (SD). Statistical analyses were performed by GraphPad Prism 6.0 (GraphPad software Inc., San Diego, CA, United States). Normality assumption for all data was checked with Shapiro-Wilk test. ANOVA with 2 factors (group × time of measurement) following Holm-Sidak post hoc analysis was used to determine the effect of intervention on VO2max and RE. For RSI and 2-km time trial, a Kruskal-Wallis test followed by Dunn<sup>0</sup> s multiple comparison test was used to determine the effect of intervention. In addition, the effect size (ES) was calculated for each comparison. The α level used for all statistics was 0.05.

### RESULTS

### After Training, at Sea Level (SL +4 Week)

Regarding to the reactive strength index, compared to the sea level pre-training value (taken as 100% of performance), neuromuscular performance was improved after plyometric training (139.3 ± 11.3% of pre-training performance; p < 0.05; ES = 0.20), while the control group did not modify its neuromuscular performance (**Figure 2B**). Moreover, when the plyometric training group and the control groups were compared, a greater (p < 0.05) reactive strength index was observed in the plyometric training group after training (**Figure 2B**).

In relation to the 2-km time trial performance test, compared to the performance at sea level pre-training (11.3 ± 0.5 min; taken as 100% of performance), after training the plyometric training group improved (reduced) the time needed to complete the test (8.75 ± 0.4 min, 77.4 ± 3.2% of pre-training time; p < 0.05; ES = 0.31), while the control group did not modify its performance (**Figure 2B**). Moreover, when the plyometric training group and the control groups were compared, a greater (p < 0.05) 2-km time trial performance was observed in the plyometric training group after training (**Figure 2A**).

Compared to pre-training VO2max values (44.7 ± 3.4 ml·kg−<sup>1</sup> ·min−<sup>1</sup> ), the VO2max was not significantly affected by plyometric training after 4 weeks (48.6 ± 2.8 ml·kg−<sup>1</sup> ·min−<sup>1</sup> ; ES = 0.13) (**Figure 3A**). However, after plyometric training the RE was improved at 10 km·h −1 (post-training, 31.1 ± 2.8 ml·kg−<sup>1</sup> ·min−<sup>1</sup> ; pre-training, 34.5 ± 4.6 ml·kg−<sup>1</sup> ·min−<sup>1</sup> ; p < 0.05; **Figure 3B**). In addition, the RE normalized to VO2max was significantly (p < 0.05) improved after training at 10 km·h −1 (−13.3%; ES = 0.87) and 12 km·h −1 (−9.4%; ES = 0.76) (**Figure 3C**).

### After Training, at High Altitude (HA +4 Week)

During acute high altitude exposure, SBP, DBP, MABP, and HR were not significant different as compared to sea level (**Table 1**). The SpO<sup>2</sup> was reduced (p < 0.05) in both groups, plyometric (93.1 ± 0.1 vs. 98.2 ± 0.1%, ES = 0.51) and control (89.8 ± 3.2% vs. 99.1 ± 1.1 %, ES = 0.38) after 24 h of high altitude exposure compared to sea level (**Table 1**). Moreover, physiological parameters at sea level and at high altitude were not significant different between plyometric training and control groups (**Table 1**).

Regarding to the reactive strength index, compared to sea level pre-training value (taken as 100% of performance), after training the neuromuscular performance was greater after 24 h of high altitude exposure (140.4 ± 14.5% of pre-training performance; p < 0.05; ES = 0.27) in the plyometric training group, while the control group did not modify its neuromuscular performance (**Figure 2B**). Additionally, when the RSI obtained at high altitude was compared to SL +4 week, no significant differences were observed, nor in the plyometric training group nor in the control group (ES = 0.05; **Figure 2B**). Moreover, when the plyometric training group and the control groups were compared, a greater (p < 0.05) reactive strength index was observed in the plyometric training group at high altitude (**Figure 2B**).

In the 2-km time trial test, compared to sea level pre-training value (10.7 ± 0.6 min; taken as 100% of performance), the plyometric training group showed no performance deterioration at high altitude (11.3 ± 0.5 min; 102.2 ± 3.3% of pre-training time; p > 0.05; ES = 0.03) (**Figure 2A**). On the contrary, the control group suffered a deterioration in the 2-km time trial test at high altitude (10.3 ± 0.8 min; p < 0.05; ES = 0.13) compared to performance at sea level before the intervention (9.02 ± 0.64 min) (**Figure 2A**). Moreover, when the plyometric training group and the control groups were compared, a better (p < 0.05) 2-km time trial performance was observed in the plyometric training group at high altitude (**Figure 2A**).

### DISCUSSION

The aim of this study was to determine the effects of a sea level short-term plyometric training program on explosive and endurance performance at sea level and after an acute exposure to high altitude (i.e., 3,270 m above sea level). Main findings confirm the beneficial effects of sea level short-term plyometric training on RSI30, 2-km time trial and running economy at sea level, without affecting sea level VO2max. Furthermore, main findings expand knowledge about its beneficial effects, reflecting an improved RSI after an acute exposure to high altitude as well as allowing subjects to maintain their pre-intervention sea-level endurance performance when acutely exposed to high altitude. These results are unique, showing that plyometric training,

FIGURE 2 | Effect of 4 weeks plyometric training on a 2-km time-trial running test and 30 cm drop jump reactive strength index (RSIDJ30). SL +4 week: sea level performance after 4 weeks of plyometric training; HA+4 week: performance after 24 h of high-altitude exposure (3,270-m over sea level) after 4 weeks of plyometric training. (A) Note that at SL +4 week the plyometric training group displayed an improvement in 2-km time-trial compared to the control group. At HA +4 week, the control group displayed a reduction in 2-km time-trial performance compared to sea level, and the plyometric training group showed a better performance compared to the control group. (B) At SL +4 week and at HA +4 week the plyometric training group displayed an improvement in RSI compared to the control group. ANOVA with 2 factors, followed by Holm-Sidak post hoc; <sup>∗</sup>p < 0.05. Dotted line reflects baseline performance at sea-level.

4 weeks of plyometric training (SL +4 week, black dots). Note that at 10 km·h −1 the plyometric training group showed a decrease of VO<sup>2</sup> compared to SL pre-training. (C) RE normalized to VO<sup>2</sup> max which showed that at 10 km·h <sup>−</sup><sup>1</sup> and 12 km·h <sup>−</sup><sup>1</sup> plyometric training group displayed an improvement of RE. ANOVA with 2 factors, followed by Holm-Sidak post hoc; <sup>∗</sup>p < 0.05. NS, denotes no significant differences between time periods.

besides optimizing explosive and endurance performance at sea level, may further aid performance at high-altitude.

Current results showed that 4 weeks of plyometric training improves RSI at sea level. Previous studies also observed shortterm improvements in RSI after plyometric training (Adams et al., 1992; Ramirez-Campillo et al., 2013; Ramirez-Campillo et al., 2014a,c). As a novelty, current results showed that the adaptations induced by plyometric training on RSI can be fully transferred when subjects are acutely exposed at high altitude. Short-duration maximal-intensity performance does not depend on aerobic metabolism (Faiss et al., 2013). The RSI measurement from drop jumps usually takes < 1 s of maximal-effort. In this sense, it was not surprising to observe that RSI performance adaptations were equally expressed at sea level and at high altitude. Moreover, variables related to jumping performance might be maintained during acute exposure to high altitude (Coudert, 1992; Kayser et al., 1993; Burtscher et al., 2006) and may even be favorably affected (Wood et al., 2006). In fact,

a previous cross-sectional study (Garcia-Ramos et al., 2018) revealed that an acute ascent to altitude may even induce a significant effect on the force-velocity relationship obtained during a vertical jump task, with greater maximal power and velocity values compared to sea-level. Although our results did not showed an increase in RSI performance after an acute exposure to high altitude, current findings do provide evidence regarding the positive effects of plyometric training conducted at sea level on RSI performance at sea level and at high altitude.

The RE was improved after plyometric training (31.1 ± 2.8 vs. 34.5 ± 4.6 ml·kg−<sup>1</sup> ·min−<sup>1</sup> , SL +4 week vs. SL, respectively, p < 0.05; **Figure 3B**), confirming previous findings (Conley and Krahenbuhl, 1980; Saunders et al., 2004a), including results observed among recreational (Turner et al., 2003) and well trained athletes (Saunders et al., 2006). Additionally, current results indicate this improvement was independent from meaningful changes in VO2max, also confirming previous findings (Paavolainen et al., 1999). Of note, the improved RE was observed in line with improvements in RSI, suggesting that RSI enhancement may underlie RE improvements (Paavolainen et al., 1999). These findings may be particularly relevant among highly trained endurance athletes, due to their limited ability to improve VO2max (Midgley et al., 2007). Moreover, considering that the mechanisms underlying improvements in RE after a short-term plyometric training are of neuromuscular nature (Markovic and Mikulic, 2010; Balsalobre-Fernandez et al., 2016), such neuromuscular variables related to RE mechanisms may not be negatively affected by acute exposure to high altitude (Coudert, 1992; Kayser et al., 1993; Burtscher et al., 2006), and may be even potentiated (Garcia-Ramos et al., 2018), helping toward a better RE at high altitude.

Moreover, 2-km time trial performance improved after plyometric training, in line with previous findings (Paavolainen et al., 1999; Ramirez-Campillo et al., 2014a). As VO2max was not improved after intervention, the time-trial performance improvement may rely on RE and neuromuscular-related (i.e., RSI) adaptations (Coyle, 1995). Different to jumping performance, acute exposure (24 h) to high altitude decreases endurance performance (Maher et al., 1974; Young et al., 1996; Bassett and Howley, 2000). In this context, it is of note that subjects from current study were able to maintain their pre-intervention sea-level endurance performance when acutely exposed to high altitude after short-term plyometric training. This result contrasts with the result observed in the control group that, compared to their pre-intervention sea-level endurance performance, suffered a performance reduction when exposed to high altitude. Thus, it is possible that the improvement of RE and RSI after plyometric training may have enhanced timetrial endurance performance at sea level (Pollock, 1977; Thomas et al., 1999), but also it is possible that these physiological adaptations may have attenuated the negative effects of HA on 2-km time trial performance after acute exposure at high altitude.

As practical application, with only 15 min per day, 3 days per week, 4 weeks of plyometric training may improve endurance and neuromuscular performance at sea level, through adaptations that may be transferred to help subjects perform better after acute exposure to high altitude. This may be of critical importance for endurance athletes competing in hypoxic environments, as aerobic performance decreases after acute exposure to high altitude (Fulco et al., 1998). In this sense, for athletes whose performance relies heavily on endurancerelated variables, such as soccer players, long-distance runners, among others, their training schedules in preparation for high-altitude competitions may consider the inclusion of key plyometric training protocols in order to maximize their performance.

Some potential limitations should be acknowledge. In this sense, our experimental sample size was limited. Moreover, VO<sup>2</sup> max and RE were not assessed at high altitude, limiting the possibility to better understand the mechanisms behind the improved 2-km time trial performance in the plyometric training group compared to the control group after an acute exposure to high altitude. Although the current study did not incorporate measurements of RE at high-altitude, it is reasonable to assume that RE also improved at high-altitude, given the aforementioned relationship between RE and RSI (Paavolainen et al., 1999), the latter being improved at high-altitude after plyometric training. Our findings are also limited to the acute effects of high altitude, with more research needed to assess the effects of plyometric training performed at sea level on endurance and explosive performance during more prolonged periods of high altitude exposure. These potential limitations should be taken into account in the interpretation of current findings.

In conclusion, 4 weeks of sea level short-term plyometric training improves RSI30, 2-km time trial and running economy at sea level, without affecting sea level VO2max. Moreover, after training, both at sea level and at high altitude, the plyometric training group demonstrated a greater RSI and 2-km time trial performance compared to the control group. These results confirm the beneficial effects of sea level short-term plyometric training on explosive and endurance performance at sea level. In addition, current results indicates that plyometric training may also be of value for endurance athletes performing after an acute exposure to high altitude.

### AUTHOR CONTRIBUTIONS

DCA, AB, JS-U, CA, AG-H, RDR, and RR-C designed the work and contributed to analysis and interpretation of the data. DCA, AB, CL-V, OM-B, ET, PO-F, CT, and JS-U acquired the data. DCA, AB, CL-V, OM-B, ET, PO-F, CT, JS-U, DCA, CA, AG-H, RDR, and RR-C drafted the work. DCA, AB, CL-V, OM-B, ET, PO-F, CT, JS-U, CA, AG-H, RDR, and RR-C critically revised the work, approved the final version to be published, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work were appropriately investigated and resolved.

### FUNDING

This work was supported by the "Dirección General Estudiantil" and "Biomedical Department" of the Universidad de Antofagasta. DCA was supported by Proyectos de Iniciación en Investigación, Vicerrectoría de Investigación,

### REFERENCES


Universidad Mayor, Chile (I-2018031), RDR was supported by Fondecyt 1180275 and AB was supported by Semilleros de Investigación, Universidad de Antofagasta (5313). CA and RR-C were supported by Proyectos de Investigación API4, Dirección de Investigación, Universidad de Los Lagos.

stay at high-altitude. Int. J. Sports Med. 14, 244–247. doi: 10.1055/s-2007- 1021171



hypoxia. Eur. J. Sport Sci. 6, 163–172. doi: 10.1080/1746139060057 1005

Young, A. J., Sawka, M. N., Muza, S. R., Boushel, R., Lyons, T., Rock, P. B., et al. (1996). Effects of erythrocyte infusion on VO2max at high altitude. J. Appl. Physiol. 81, 252–259. doi: 10.1152/jappl.1996.81.1.252

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Andrade, Beltrán, Labarca-Valenzuela, Manzo-Botarelli, Trujillo, Otero-Farias, Álvarez, Garcia-Hermoso, Toledo, Del Rio, Silva-Urra and Ramírez-Campillo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Ventilatory and Autonomic Regulation in Sleep Apnea Syndrome: A Potential Protective Role for Erythropoietin?

David C. Andrade1,2, Liasmine Haine<sup>3</sup> , Camilo Toledo1,4, Hugo S. Diaz1,5 , Rodrigo A. Quintanilla<sup>5</sup> , Noah J. Marcus<sup>6</sup> , Rodrigo Iturriaga<sup>7</sup> , Jean-Paul Richalet<sup>3</sup> , Nicolas Voituron<sup>3</sup>† and Rodrigo Del Rio1,4,8 \* †

<sup>1</sup> Laboratory of Cardiorespiratory Control, Department of Physiology, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>2</sup> Centro de Investigación en Fisiología del Ejercicio, Facultad de Ciencias, Universidad Mayor, Santiago, Chile, <sup>3</sup> Laboratoire Hypoxie and Poumon – EA2363, Université Paris 13, Paris, France, <sup>4</sup> Centro de Envejecimiento y Regeneración (CARE), Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>5</sup> Centro de Investigación Biomédica, Universidad Autónoma de Chile, Santiago, Chile, <sup>6</sup> Department of Physiology and Pharmacology, Des Moines University, Des Moines, IA, United States, <sup>7</sup> Laboratorio de Neurobiología, Department of Physiology, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>8</sup> Centro de Excelencia en Biomedicina de Magallanes (CEBIMA), Universidad de Magallanes, Punta Arenas, Chile

### Edited by:

Rohit Ramchandra, University of Auckland, New Zealand

### Reviewed by:

Alessandro Silvani, Università degli Studi di Bologna, Italy Melissa L. Bates, The University of Iowa, United States

> \*Correspondence: Rodrigo Del Rio rdelrio@bio.puc.cl

†These authors have contributed equally to this work as senior authors

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 14 June 2018 Accepted: 21 September 2018 Published: 16 October 2018

Citation:

Andrade DC, Haine L, Toledo C, Diaz HS, Quintanilla RA, Marcus NJ, Iturriaga R, Richalet J-P, Voituron N and Del Rio R (2018) Ventilatory and Autonomic Regulation in Sleep Apnea Syndrome: A Potential Protective Role for Erythropoietin? Front. Physiol. 9:1440. doi: 10.3389/fphys.2018.01440 Obstructive sleep apnea (OSA) is the most common form of sleep disordered breathing and is associated with wide array of cardiovascular morbidities. It has been proposed that during OSA, the respiratory control center (RCC) is affected by exaggerated afferent signals coming from peripheral/central chemoreceptors which leads to ventilatory instability and may perpetuate apnea generation. Treatments focused on decreasing hyperactivity of peripheral/central chemoreceptors may be useful to improving ventilatory instability in OSA patients. Previous studies indicate that oxidative stress and inflammation are key players in the increased peripheral/central chemoreflex drive associated with OSA. Recent data suggest that erythropoietin (Epo) could also be involved in modulating chemoreflex activity as functional Epo receptors are constitutively expressed in peripheral and central chemoreceptors cells. Additionally, there is some evidence that Epo has anti-oxidant/anti-inflammatory effects. Accordingly, we propose that Epo treatment during OSA may reduce enhanced peripheral/central chemoreflex drive and normalize the activity of the RCC which in turn may help to abrogate ventilatory instability. In this perspective article we discuss the potential beneficial effects of Epo administration on ventilatory regulation in the setting of OSA.

Keywords: erythropoietin, peripheral chemoreflex, central chemoreflex, chronic intermittent hypoxia, sleep apnea

# INTRODUCTION

Sleep apnea (SA) syndrome is a pathological condition characterized by recurrent airway obstruction or cessation of breathing during sleep (Gislason et al., 1988; Epstein et al., 2009; Dempsey et al., 2010) resulting in hypercapnia and hypoxemia/oxyhemoglobin desaturation (Badran et al., 2014a). The collapse of the upper airway during obstructive events is likely attributable to both anatomical and non-anatomical determinants (Dempsey et al., 2010) including changes in central and peripheral respiratory drive (Solin et al., 2000; Kara et al., 2003; Dempsey et al., 2012).

**318**

Acutely, activation of peripheral and/or central chemoreflexes during apneic episodes is associated with micro-arousals and sleep fragmentation, as well as increases in ventilation, sympathetic nerve activity, blood pressure, and heart rate. Repetitive apneas as observed in SA syndrome are associated with increased diurnal drowsiness, neurocognitive dysfunction, and cardiovascular morbidity (Young et al., 2002; Gozal, 2013; Konecny et al., 2014). In clinical populations, cardiovascular morbidity in SA patients is often associated with enhanced activity of the sympathetic nervous system, and numerous studies indicate that this increased sympathetic activity stems from a heightened carotid body chemoreflex (CBC) (Narkiewicz et al., 1998; Mansukhani et al., 2015; Iturriaga, 2017). In addition to affecting autonomic outflow, aberrant CBC activity may also have adverse effects on the central respiratory control network.

Central respiratory control is finely regulated by a neural network located mainly in the ponto-medullary region of the brainstem (Richter and Spyer, 2001). In the face of hypoxic or hypercapnic challenge, maintenance of ventilatory and cardiovascular homeostasis is achieved by activation of peripheral and central chemoreceptors and subsequent modulation of this ponto-medullary respiratory control network (Marshall, 1994). It has been proposed that information from these sensory afferents is sufficient to stimulate the central respiratory control system and alter respiratory pattern independent of acidification of cerebrospinal fluid (Guyenet et al., 2017).

Pathological insults associated with central or obstructive apneas can alter chemoreceptor function, change chemoreflex integration in the central nervous system, and/or alter the properties of the central respiratory network (Gozal, 1998; Jokic et al., 2000; Harper et al., 2005; Katz et al., 2009; Carroll et al., 2010). Previous work suggests that enhanced peripheral chemoreflex activation has central effects that contribute to respiratory instability (Levy et al., 2008) and thus may play a role in perpetuating SA via creation of a positive feedback loop. Indeed, several studies have shown that increases in carotid body (CB) and central chemoreflex drive play an important role in the pathophysiology of obstructive sleep apneas (OSA) (For review see Dempsey et al., 2010).

The seminal pathological insult occurring during apneas is the repeated exposure to episodes of hypoxia-reoxygenation. This chronic intermittent hypoxia (CIH) exposure is mechanistically linked to increased peripheral chemoreflex drive, and is associated with oxidative stress and inflammation (Del Rio et al., 2011; Iturriaga et al., 2014; Iturriaga, 2017). Thus many of the major morbidities associated with SA as well as respiratory instability itself may be related to aberrant chemoreflex activation prior to and during exposure to intermittent hypoxia associated with apneic episodes. Therefore, therapies aimed at reducing chemoreflex sensitivity may be beneficial in preventing the pathophysiological sequelae of SA as well as potentially reducing the frequency of apneic episodes generated by respiratory instability.

Accordingly, several rodent models have been developed to study the pathophysiological mechanisms that contribute to enhanced peripheral and central chemoreflex drive, utilizing CIH exposure (Iturriaga et al., 2005, 2014; Rey et al., 2006; Del Rio et al., 2011). The usefulness of these models is confirmed by findings that CIH results in heightened CB activity and ventilatory chemoreflex gain in response to hypoxia (Rey et al., 2004; Del Rio et al., 2010, 2012, 2016), as well as chemoreflexmediated increases in sympathetic activity and blood pressure (Marcus et al., 2010; Del Rio et al., 2016).

### CHEMORECEPTORS, RESPIRATORY CONTROL AND CARDIOVASCULAR REGULATION IN THE SETTING OF SLEEP APNEA

Sleep apnea syndrome is characterized by two types of events, (i) OSA and (ii) central sleep apneas (CSA). OSA is characterized by partial or complete occlusion of the upper airways during sleep; while CSA is characterized by a marked decrease in respiratory motor drive resulting from a reduction in the activity of the central respiratory network (Malhotra and Owens, 2010). It has been proposed that heightened chemoreflex gain may contribute to SA by destabilizing central respiratory network control of airway tone and/or ventilation (Del Rio et al., 2016). The process by which chemoreflex gain affects respiratory stability is often described using a control-systems engineering concept referred to as "loop gain" (Khoo, 2000).

In this application, "loop gain" can be thought of as the ratio of the size of a response (change in ventilation) to the size of a disturbance (change in PaCO2). The components of loop gain include controller gain, plant gain, and feedback gain. Controller gain represents the ventilatory response to PaCO2, plant gain represents the blood gas response to a change in ventilation, and feedback gain represents the delay associated with relaying the feedback signal (PaCO2) to the controller (chemoreceptors). Khoo (2000) explained that there is a chain of events, which is at the origin of ventilatory instability and attendant oscillation of ventilatory drive. Either obstructive or central apneas result in increased PaCO<sup>2</sup> and activation of chemoreceptors. The duration of the oscillatory response and its magnitude are determined by the effect of ventilatory changes on PaCO<sup>2</sup> (i.e., the "plant gain"), as well as by the strength of the chemoreflex response (i.e., the "controller gain"). According to this paradigm, higher loop gain is associated with greater probability of breathing instability as chemoreflex responses to changes in PaCO<sup>2</sup> are likely to be disproportionate and result in ventilatory overshoots that reduce PaCO<sup>2</sup> below the apneic threshold. Conversely, lower loop gains results in a more robust respiratory network which is less prone to instability and development of periodic breathing (Malhotra and Owens, 2010). With respect to the specific topics covered in this perspective article, an increase in the gain of peripheral and central chemoreceptors (controller gain) may trigger ventilatory instability and contribute to higher apnea incidence. In concordance with this notion, it has been shown in experimental low output heart failure, a condition characterized by increased apnea incidence, that peripheral chemoreceptor ablation stabilized ventilation and greatly attenuated apnea incidence (Marcus et al., 2014). In addition, selective elimination of central chemosensory neurons from the ventral medullary surface increases the apneic threshold toward eupneic ventilatory values (Takakura et al., 2008). Taken together, these studies suggest a role of both peripheral and central chemoreceptors in the development of oscillatory breathing patterns and increased apnea incidence.

### Peripheral Chemoreceptors

fphys-09-01440 October 12, 2018 Time: 14:59 # 3

Peripheral chemoreceptors detect changes in arterial blood gases (mainly hypoxemia) and respond by activating the sympathetic nervous system and increasing ventilation to restore blood-gas homeostasis (Kara et al., 2003). Hypoxia-induced hyperventilation is mainly triggered by activation of the CB and to some extent by activation of the aortic body (located on the aortic arch) (Miller and Tenney, 1975; Brophy et al., 1999). The CB chemoreceptors are the main peripheral arterial chemoreceptor and are located in the bifurcation of the carotid artery. They are composed of clusters of chemoreceptor cells (type I cells) surrounded by glial cells (type II cells) (Iturriaga and Alcayaga, 2004). Type I cells are considered polymodal receptors since they respond to a wide variety of stimuli such as changes in arterial levels of pO2, pCO2, pH, blood flow, and temperature (Gonzalez et al., 1994). Upon activation by hypoxia, type I cells release ACh and ATP which interact with receptors on the sensory nerve fibers of the carotid sinus nerve (Gonzalez et al., 1994). The precise biochemical nature of the transmitter released by type I cells during hypoxic stimulation has not been completely identified since more that one molecule has been shown to be released (ACh and ATP) (Iturriaga and Alcayaga, 2004). Moreover, several peptides hormones and gasotransmitters serve as excitatory and inhibitory modulators of CB chemosensitivity (i.e., NO, histamine, and AngII) (Iturriaga and Alcayaga, 2004; Del Rio et al., 2008).

Hypoxic hyperventilation seems synchronous with the increase of the discharge frequency of the sinus nerve fibers (Vizek et al., 1987). The first central integration of sensory information from the peripheral chemoreceptors and the main areas sensory fibers from the sinus nerve project to the commissural and middle divisions of the nucleus of the solitary tract (cNTS and mNTS, respectively) (Claps and Torrealba, 1988; Finley and Katz, 1992). Neurons of the cNTS and mNTS integrate and relay information from peripheral chemoreceptors to other regions of the central nervous system to ultimately orchestrate the hypoxic hyperventilatory response (Ponikowski and Banasiak, 2001; Rosin et al., 2006; Smith et al., 2010 ). In addition, CB stimulation also triggers activation of the sympathetic nervous system to maintain adequate arterial pressure in the face of hypoxic vasodilation (Schultz and Sun, 2000). While normal CB function contributes to maintenance of blood gas homeostasis, pH regulation, and tissue perfusion, mounting evidence indicates that maladaptive changes in CB function contribute to a variety of cardiovascular and metabolic disease states (Schultz et al., 2015).

### Central Chemoreceptors

Central chemoreceptors are located mainly on the ventral surface of the medulla (Nattie and Li, 2012). In response to changes in cerebrospinal fluid CO2/H<sup>+</sup> content, central chemoreceptor neurons send excitatory signals directly to respiratory control centers to increase breathing rate (Guyenet et al., 2005). Importantly, stimulation of central chemoreceptors also elicits an increase in sympathetic outflow mainly by their projections to pre-sympathetic control areas (Moreira et al., 2006). The precise localization of central chemoreceptors within the brain and the circuitry that is activated by CO2/H<sup>+</sup> stimulation is still controversial. However, the retrotrapezoid nucleus (RTN) appears to play a pivotal role in the regulation of the hypercapnic ventilatory response (Guyenet and Bayliss, 2015). The RTN is mostly composed of a group of neurons that are activated by changes in cerebrospinal fluid CO<sup>2</sup> and/or pH that projects to areas related to respiratory control (Lazarenko et al., 2009; Guyenet and Mulkey, 2010; Guyenet et al., 2012; Wang et al., 2013). RTN chemosensitive neurons are rhythmically active and have been shown to be activated by low pH in vivo and in vitro (Lazarenko et al., 2009; Wang et al., 2013). Interestingly, it has been shown that partial elimination of RTN chemosensory neurons (∼70%) in healthy rats increases the apneic threshold (Takakura et al., 2008), meaning apneic events occur at a higher end-tidal CO2. Therefore, it is plausible to hypothesize that RTN chemoreceptor neurons activity/sensitivity may contribute to ventilatory instability. In this regard, a higher ventilatory response following hypercapnia would play a major role in apnea development as enhanced CO<sup>2</sup> "wash-out" would drop PaCO<sup>2</sup> close to or below the apneic threshold (Topor et al., 2001). Indeed, studies in patients with heart failure have shown that apneas result in enhanced chemoreflex responses which result in a resting eupneic PtCO<sup>2</sup> being closer to the apneic threshold (i.e., narrowed CO<sup>2</sup> reserve, Xie et al., 2002). Furthermore, patients with heart failure that have OSA show increased ventilatory responses to hypercapnia (Solin et al., 2000). Thus, alterations in RTN chemoreceptor neuron function may contribute to apnea incidence in OSA patients by altering the apneic threshold itself or the eupneic "proximity" to the apneic threshold.

### Cellular Mechanisms of Enhanced Chemoreceptor Activity in Sleep Apnea/CIH

While the precise mechanisms underlying peripheral and/or central maladaptations to CIH are not completely understood, recent evidence suggests that reconfiguration of the neuronal network involved in sympathetic regulation and breathing stability occurs (Xu et al., 2004; Del Rio et al., 2010, 2012). Numerous studies underscore the role of oxidative stress (Marcus et al., 2010; Badran et al., 2014b; Morgan et al., 2016) and inflammation (Del Rio et al., 2010, 2012) as major drivers of augmented chemoreflex drive observed in the CIH model. Indeed, experimental CIH is associated with elevation of sympathetic outflow which is dependent on ROS production at the level of the peripheral chemoreceptors (Marcus et al., 2010; Braga et al., 2011). Taken together these studies suggest that novel treatments capable of reducing oxidative stress and/or inflammation at the level of the peripheral chemoreceptors may have potential therapeutic value for the treatment of SA-related

autonomic and ventilatory dysregulation and by extension SArelated cardiovascular morbidities (i.e., systemic hypertension).

Despite numerous studies exploring the role of central nervous system ROS in cardiovascular disease, little is known about the role of ROS in the processing of the cardiovascular reflexes within the brainstem (Braga et al., 2011). CIH and Angiotensin II-derived ROS play a crucial role in the modulation of baroreceptor and chemoreceptor function, but also have been shown to play a role in altered neurotransmission in brainstem sympathetic control areas like the NTS and the RVLM (Gao et al., 2005; Nunes et al., 2010; Braga et al., 2011; Del Rio et al., 2016). To date, no studies have addressed the potential role of CIH-derived ROS on central chemoreceptor regions such as the RTN. Additional studies are needed to determine any possible contribution of RTN neurons and the central chemoreflex on the cardiovascular disturbances observed in CIH. Considering that both peripheral and central chemoreceptors potentially contribute to the pathophysiology of OSA and that oxidative stress and inflammation play important roles in abnormal chemoreceptor physiology, it is reasonable to propose that new therapeutic strategies targeting oxidative stress and inflammation may have a positive impact on aberrant peripheral and central chemoreceptor function.

### ERYTHROPOIETIN AND INTERACTION WITH ITS RECEPTOR

Erythropoietin (Epo) is a small signaling molecule produced in the kidney and whose primary known function is the stimulation of erythropoiesis in the bone marrow (Donnelly, 2001; Dzierzak and Philipsen, 2013); however, Epo also has anti-oxidative effects. Epo directly activates intracellular anti-oxidant mechanisms such as heme oxygenase-1 and glutathione peroxidase, and Epo may inhibit iron-dependent oxidative injury indirectly by inducing iron depletion (Katavetin et al., 2007). The Epo receptor (Epo-R) is present on the surface of erythroid progenitors as a homodimer of two identical Epo-R subunits (Livnah et al., 1999), and Epo binds its receptor with very high affinity (Bunn, 2013). However, there is evidence that Epo has non-hematopoietic activity which is mediated by a ß common receptor, a heterodimer with one Epo-R monomer and CD31 (Leist et al., 2004; Chen et al., 2015). In addition to the kidney derived Epo, there is ample evidence to indicate that Epo is also produced outside of the kidney. Indeed, Epo mRNA has been detected in lungs, testis, heart, and brain in rodents (Tan et al., 1992; El Hasnaoui-Saadani et al., 2013; Pichon et al., 2016). Cells from the retina, testes, lungs, and some neurons and glial cells of the central nervous system have been shown to constitutively express several components of the Epo signaling pathway (Digicaylioglu et al., 1995; Gassmann et al., 2003; Grimm et al., 2004; Jelkmann, 2007; Yasuda et al., 2010) and expression of its target receptor (Epo-R) is found in endothelial cells, smooth muscle cells, retinal tissue, testis, and the central nervous system (Masuda et al., 1994; Ammarguellat et al., 1996; Marti et al., 1997; Bernaudin et al., 1999; Gassmann et al., 2003). Taken together, these studies suggest a role for Epo in regulation of physiological functions other than erythropoiesis (Gassmann et al., 2003). Indeed, recent data suggests that Epo regulates the control of breathing via central and peripheral actions (Soliz et al., 2005; Brugniaux et al., 2011; Voituron et al., 2014).

# ERYTHROPOIETIN AND RESPIRATORY REGULATION

A number of studies provide evidence that Epo plays a role in control of breathing. The Epo-R is expressed in the pre-Bötzinger complex, a key region in the brainstem involved in ventilatory rhythmogenesis and regulation (Soliz et al., 2005). Epo increases dopamine release and tyrosine hydroxylase (TH) activity in cells with neural characteristics (Masuda et al., 1994; Koshimura et al., 1999; Yamamoto et al., 2000; Tanaka et al., 2001), and Epo-R is specifically expressed in TH-positive cell groups in the brainstem (Soliz et al., 2005). Furthermore, overexpression of Epo increase brainstem catecholamine turnover in mice (Soliz et al., 2005). Interestingly, the Epo found in the central nervous system does not reach the systemic circulation due to the lack of permeability of the blood–brain barrier (Gassmann et al., 2003). These results strongly suggest that Epo-derived from the central nervous system itself must play a physiological role "in situ" as a local regulator of neuronal function (Jelkmann, 2007).

In addition to altering the metabolism of catecholamines in the brainstem, Epo has been shown to have similar effects in the CB, and has been shown that one injection of human recombinant Epo reduce the tidal volume during hypoxic stimulus in humans and mice (Soliz et al., 2005, 2009; Lifshitz et al., 2009). In support of this notion, recent studies have shown that Epo is released within the RVLM during hypoxic stimulation (Oshima et al., 2018), and that Epo-R is constitutively expressed in peripheral and central structures involved in ventilatory chemoreflex control (Digicaylioglu et al., 1995; Soliz et al., 2005; Lam et al., 2009; Voituron et al., 2014). Epo is known to regulate hypoxic ventilatory response (HVR) in mice by interacting with brainstem and CB (Soliz et al., 2005). The ventilatory response to hypoxia is a sex dependent response, being more pronounced in female sex (Joseph et al., 2000, 2002). Interestingly, Epo has sexually dimorphic effects on the ventilatory response to hypoxia. Indeed, it tends to increase the HVR in female mice and in women via interaction with sex steroid hormones (Soliz et al., 2009). Besides its well-known role in erythropoiesis and influence on control of breathing, Epo also exerts important cytoprotective effects.

### ERYTHROPOIETIN AS A PROTECTIVE MOLECULE DURING EXPOSURE TO INTERMITTENT HYPOXIA

It has been shown that Epo exerts a neuroprotective role in several diseases due to its anti-apoptotic (Sirén and Ehrenreich, 2001), anti-cytotoxic (Morishita et al., 1997), antioxidative (Koshimura et al., 1999), and anti-inflammatory

molecules, which play a pivotal role on the pathophysiology by triggering augmented chemoreflex sensitivity which ultimately lead to altered breathing and cardiovascular events. Epo has been shown to reduce both ROS and inflammation in the brain and at peripheral chemoreceptors located primarily at the carotid body; however, its role on central chemoreceptors sensitivity remains to be studied. Thus, it is plausible to hypothesize that Epo administration and/or its derivates in the setting of sleep apnea may offer an anti-oxidant/inflammatory therapy to control for the augmented chemoreflex drive.

(Villa et al., 2003) properties. Increases in oxidative stress and inflammation are recognized as key mediators affecting control of breathing and cardiovascular function following exposure to CIH in rodents (Del Rio et al., 2010, 2011, 2012, 2016; Marcus et al., 2010; Iturriaga et al., 2014). It has been shown that CIH induces oxidative stress in the CB and potentiation of CB-mediated chemoreflex drive (Del Rio et al., 2010; Marcus et al., 2010). In addition, increased expression of pro-inflammatory cytokines in the CB has been shown following exposure to CIH (Del Rio et al., 2011, 2012). Furthermore, we showed that ibuprofen treatment selectively reduces central inflammation in the NTS in rats exposed to CIH, and that ibuprofen treatment decreases the ventilatory response to hypoxia (Del Rio et al., 2012). Taken together, these results suggest that oxidative stress and inflammation acting predominantly on chemoreflex pathways are involved in the altered chemoreflex function and attendant autonomic dysregulation following CIH. Thus, it is plausible that administration of Epo could have a positive effect on control of breathing and autonomic function during/after exposure to CIH (**Figure 1**).

### CONCLUSION

Sleep apnea syndrome, characterized by cyclic and repeated exposure to brief episodes of hypoxia and hypercapnia, is recognized as a major public health problem worldwide. SA can occur as a result of upper airway obstruction and/or as a result of abnormal respiratory control resulting from aberant

peripheral and central chemoreflex function. Currently, there are no treatments that specifically target abnormal control of breathing in SA. Epo has recently been shown to have novel neuroprotective properties associated with anti-oxidant and anti-inflammatory effects. Increases in oxidative stress and inflammation are both recognized as key mediators in respiratory and cardiovascular disturbances following exposure to CIH. Accordingly, we propose that future studies should address the potential beneficial effect of Epo or Epo-like compounds on cardio-respiratory function during or after exposure to CIH. Epoinduced erythropoiesis could be detrimental in patients with SA therefore, developing new Epo-derived compounds that can bind to the Epo-R with little or no effect on erythropoiesis would be optimal in terms of therapeutic value.

In summary, uncovering a role for Epo in the regulation of the ventilatory response to hypoxia and/or hypercapnia as well as ventilatory instability will open new avenues in the field of control of breathing in the pathological setting of SA. Furthermore,

### REFERENCES


determining the potential therapeutic efficacy of Epo or Epoderived compounds on the enhanced chemoreflex sensitivity observed during CIH will be of potential therapeutic value.

### AUTHOR CONTRIBUTIONS

All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

### FUNDING

This work was supported by FONDECYT 1180172 grants from the National Fund for Scientific and Technological Development of Chile and ECOS-CONICYT CS1603.




by erythropoietin. Am. J. Physiol. Regul. Integr. Comp. Physiol. 296, 1837–1846. doi: 10.1152/ajpregu.90967.2008


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Andrade, Haine, Toledo, Diaz, Quintanilla, Marcus, Iturriaga, Richalet, Voituron and Del Rio. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

fphys-09-01489 October 19, 2018 Time: 16:55 # 1

# Intrapartum Fetal Heart Rate: A Possible Predictor of Neonatal Acidemia and APGAR Score

Thâmila Kamila de Souza Medeiros1,2, Mirela Dobre<sup>3</sup> , Daniela Monteiro Baptista da Silva1,2, Andrei Brateanu<sup>4</sup> , Ovidiu Constantin Baltatu1,2 \* and Luciana Aparecida Campos1,2 \*

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Beth J. Allison, Hudson Institute of Medical Research, Australia Charles Christoph Roehr, University of Oxford, United Kingdom

### \*Correspondence:

Ovidiu Constantin Baltatu ocbaltatu@anhembi.br; ocbaltatu@gmail.com Luciana Aparecida Campos labaltatu@anhembi.br; camposbaltatu@gmail.com

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 08 July 2018 Accepted: 02 October 2018 Published: 22 October 2018

### Citation:

Medeiros TKS, Dobre M, da Silva DMB, Brateanu A, Baltatu OC and Campos LA (2018) Intrapartum Fetal Heart Rate: A Possible Predictor of Neonatal Acidemia and APGAR Score. Front. Physiol. 9:1489. doi: 10.3389/fphys.2018.01489 <sup>1</sup> Center of Innovation, Technology and Education at Anhembi Morumbi University – Laureate International Universities, São José dos Campos, Brazil, <sup>2</sup> School of Health Sciences at Anhembi Morumbi University – Laureate International Universities, São José dos Campos, Brazil, <sup>3</sup> Division of Nephrology and Hypertension, University Hospitals, Cleveland, OH, United States, <sup>4</sup> Medicine Institute, Cleveland Clinic, Cleveland, OH, United States

Background: Predicting perinatal outcomes based on patterns of fetal heart rate (FHR) remains a challenge. The aim of this study was to evaluate intrapartum FHR variability as predictor for neonatal acidemia and APGAR score.

Methods: This was a retrospective observational study of 552 childbirths. Multivariable linear regression models were used to assess the association between FHR variability and each of the following outcomes: arterial cord blood pH and base deficit, Apgar 1, and 5 scores. Variables used for adjustment were maternal age, comorbidities (gestational diabetes, preeclampsia, maternal fever, and hypertension), parity, gravidity, uterine contractions, and newborn gestational age, and weight at birth.

Results: The following factors were associated with an increased risk of metabolic acidosis and low Apgar scores at birth: increased mean and coefficient of variation (CV) of the FHR, type of delivery and decreased parity. Each 10-beat/min increase in the FHR was associated with an increase of 0.43 mEq/L in the base deficit, and a decrease of 0.01 in the pH, 0.2 in the Apgar 1, and 0.14 in the Apgar 5 scores. Each 10% increase in the CV of the FHR was associated with an increase of 4.05 mEq/L in the base deficit and a decrease of 0.13 in the pH, 1.31 in the Apgar 1, and 0.86 in the Apgar 5 scores.

Conclusion: These data suggest the intrapartum FHR variability is physiologically relevant and can be used for predicting the acidemia and Apgar scores at birth of the newborn infants without severe cases of morbidity and from uncomplicated pregnancies.

Keywords: fetal monitoring, heart rate, perinatal outcome, noninvasive prenatal diagnosis, neonatal acidemia, APGAR

# INTRODUCTION

fphys-09-01489 October 19, 2018 Time: 16:55 # 2

Contemporary research aims to identify reliable and early markers for the neonatal acidemia and the physical condition of a newborn infant (Devane et al., 2012, 2017). Intrapartum cardiotocography monitors fetal heart rate (FHR) and uterine contractions and is commonly used for the early detection of fetal distress (Stout and Cahill, 2011).

It is theorized that intrapartum cardiotocography FHR could detect fetal hypoxia and/or acidosis allowing a timely intervention to reduce adverse neonatal outcomes such as postnatal cerebral palsy. This is based on the theory that intrapartum hypoxia may lead to alterations in the fetal central nervous system that directly affects the electrical activity of the fetal heart and could also induce neonatal cerebral palsy (Garabedian et al., 2017). Indeed, cardiotocography FHR parameters including baseline FHR and its variability appear to be independent predictors of fetal acidosis (Silberstein et al., 2017), were associated with a substantial decrease in early neonatal mortality and morbidity (Chen et al., 2011). At present, however, there is no consensus regarding sensitivity and specificity of cardiotocography classifications in predicting acidemia, with three guidelines for cardiotocography interpretation provided by the International Federation of Gynecology and Obstetrics (FIGO), American College of Obstetrics and Gynecology (ACOG), and National Institute for Health and Care Excellence (NICE) (Bhatia et al., 2017; Santo et al., 2017).

The improvement in the accuracy of FHR pattern interpretation through a continuous FHR centralization system is envisaged to be beneficial in reducing the incidence of neonatal acidemia (Michikata et al., 2016). To improve the predictive efficacy of cardiotocography FHR, diagnostic algorithms are currently being investigated to be used in real-time computerized systems for decision support (van Scheepen et al., 2016).

In this study, we aimed at investigating whether the intrapartum baseline and variability parameters of FHR are independently associated with neonatal acidemia and the APGAR scores of the newborn infants without severe cases of morbidity and from uncomplicated pregnancies.

### MATERIALS AND METHODS

### Study Design and Setting

A single-center retrospective observational study of 552 childbirths at the University Hospital Brno (UHB), Czechia (Chudacek et al., 2014). The hospital records were consecutively selected to pair both clinical and cardiotocography signal parameters.

### Clinical Data

Clinical selection criteria (Chudacek et al., 2014) included maternal age (maternal age > 18 years), gravidity (only singleton, uncomplicated pregnancies), gestational week (weeks of gestation ≥ 37), delivery type (the majority of the database consists of vaginal deliveries), fetuses without known congenital defects or known intrauterine growth restriction (IUGR). Additional criteria considered were: sex of the fetus, parity, and risk factors including gestational diabetes, preeclampsia, maternal fever (>37.5◦C), hypertension, and meconium stained fluid. The cohort included neonates without severe cases of neonatal morbidity, hypoxic ischemic encephalopathy, or seizures.

### Cardiotocography Data

Cardiotocography recordings included contained time information and signal of FHR and uterine contractions sampled at 4 Hz were recorded starting 90 min preceding the delivery. Cardiotocography recordings selection criteria (Chudacek et al., 2014) were the signal length: 90 min preceding the delivery with 1st stage was limited to a maximum of 60 min and 2nd stage to 30 min in order to keep recordings easily comparable. Other criteria were: missing signal, noise and artifacts, and type of measurement device. The database is composed as a mixture of recordings acquired by ultrasound doppler probe, direct scalp measurement or combination of both, reflecting the clinical reality at the obstetrics ward. Cardiotocographies were recorded using STAN S21 and S31 (Neoventa Medical, Mölndal, Sweden) and Avalon FM40 and FM50 (Philips Healthcare, Andover, MA, United States) fetal monitors.

The cardiotocography recordings and clinical data were matched using anonymous unique identifiers generated by the hospital information system. A flowchart diagram describing the process of data selection for the final database is presented in the original publication (Chudacek et al., 2014).

The acid-base status (pH and base deficit) of umbilical cord arterial blood was measured at the moment of birth to evaluate the metabolic condition of neonates.

Apgar's scores (Apgar) at 1 and 5 min after birth evaluated the outcome of the delivery based on five categories: breathing effort, heart rate, muscle tone, reflexes, and skin color.

The CARDIOTOCOGRAPHY database is available at PhysioNet<sup>1</sup> ((Chudacek et al., 2014). All Physionet databases have been fully anonymized and may be used without further institutional review board approval.

### Statistical Analysis

The FHR variability was determined using coefficient of variation (CV), which equals the standard deviation divided by the mean (Pereira et al., 2017). The characteristics of study participants were depicted using standard descriptive statistics, overall and stratified by quartiles of the mean of FHR. Chi-square (χ 2 ) test of independence for categorical (nominal) variables and Analysis of variance (ANOVA) test for continuous variables were used to analyze the covariates of interest.

Multivariable linear regression models were used to assess the association between FHRVand each of the following outcomes: pH, base deficit, Apgar 1, and 5 scores. Variables used for adjustment were maternal age, comorbidities, parity, gravidity,

<sup>1</sup>https://physionet.org/physiobank/database/ctu-uhb-ctgdb/

uterine contractions, and newborn gestational age, and weight at birth.

All statistical tests were two sided, and P < 0.05 was considered significant. IBM Corp (2016) SPSS Statistics for MAC, Version 24.0, was used for analyses.

### RESULTS

### Patient Characteristics

fphys-09-01489 October 19, 2018 Time: 16:55 # 3

A total of 552 study participants were included in the study. The age of the mothers was 30 (27–33) median (IQR) years and gestational age was 40 (39–41) median (IQR) weeks. Characteristics of study participants distributed according to quartiles of the mean of FHR are summarized in **Table 1**. Patients in the lowest vs. highest quartile of the mean of FHR had a higher CV of the FHR (14.73 ± 3.56 vs. 11.24 ± 3.85). Mean of FHR was not associated with maternal age, history of diabetes mellitus or hypertension or preeclampsia, delivery type, gravidity, parity, gestational week, and mean of uterine contractions.

## Association Between FHR and Metabolic Acidosis

The median pH values of the cohort were 7.25 (IQR: 7.17– 7.3) [min to max: 6.85–7.47]. In multivariable adjusted models,

TABLE 1 | Characteristics of study participants by quartile of mean fetal heart rate.

the following factors were associated with an increased risk of metabolic acidosis at birth: increased mean and CV of the FHR, type of delivery (Caesarian section over vaginal), decreased parity (**Tables 2**, **3**). The following parameters were not associated with risk of metabolic acidosis at birth: maternal age, history of diabetes mellitus, hypertension, preeclampsia, gravidity, gestational week, uterine contractions, newborn sex or weight.

Each 10-beat/min increase in the FHR was associated with a 0.01 decrease in the pH and a 0.43 mEq/L increase in the base deficit.

Each 10% increase in the CV of the FHR was associated with a 0.13 decrease in the pH and a 4.05 mEq/L increase in the base deficit.

### Association Between FHR and Apgar Scores

In multivariable adjusted models, the following factors were associated with an increased risk of low Apgar scores at birth: increased mean and CV of the FHR, history of preeclampsia, type of delivery (Caesarian section over vaginal), decreased parity (**Tables 4**, **5**). The following parameters were not associated with risk of low Apgar scores at birth: maternal age, history of diabetes mellitus, hypertension, gravidity, uterine contractions, newborn sex or weight.


UC, uterine contractions; FHR, fetal heart rate; CV, coefficient of variation.

fphys-09-01489 October 19, 2018 Time: 16:55 # 4

Each 10-beat/min increase in the FHR was associated with a 0.2 and 0.14 decrease in the Apgar score at 1 and 5 min after birth, respectively.

TABLE 2 | Association between maternal and newborn parameters and pH.


TABLE 3 | Association between maternal and newborn parameters and base deficit.

Each 10% increase in the CV of the FHR was associated with a 1.31 and 0.86 decrease in the Apgar score at 1 and 5 min after birth, respectively.

TABLE 4 | Association between maternal and newborn parameters and Apgar 1 score (1 min after birth).


TABLE 5 | Association between maternal and newborn parameters and Apgar 5 score (5 min after birth).


# DISCUSSION

fphys-09-01489 October 19, 2018 Time: 16:55 # 5

The main findings of this study are that in a cohort of uncomplicated childbirths without severe cases of neonatal morbidity, hypoxic ischemic encephalopathy, or seizures: (1) increased mean and CV of the intrapartum FHR were associated with increased risk of metabolic acidosis and low Apgar scores at birth; (2) FHR was not associated with maternal age, history of diabetes mellitus or hypertension or preeclampsia, delivery type, gravidity, parity, gestational week, mean of uterine contractions. Besides the intrapartum cardiotocography FHR, delivery type, and decreased parity were also associated with neonatal acidemia and the physical condition of a newborn infant.

Intrapartum fetal hypoxia represents an important cause of postnatal cerebral palsy or other neurologic outcomes and in a significant proportion of cases there is evidence of suboptimal care related to fetal surveillance. Umbilical artery metabolic acidosis is commonly used to detect neurological injury (Ross and Gala, 2002).

A three-tiered FHR interpretation system for intrapartum cardiotocography FHR tracing interpretation was proposed (Macones et al., 2008; American College of Obestetricians and Gynecologists, 2009). As our study did not include complicated pregnancies, it supports using the normal cardiotocography FHR (category I), which has a predictive value of 99.7% of an Apgar score more than 7 (Macones et al., 2008; Crovetto et al., 2017; Raghuraman and Cahill, 2017).

Using multivariable models to control confounders, cardiotocography FHR was recently found to be an independent predictor of fetal acidosis (Silberstein et al., 2017), respiratory morbidity in term neonates (Liu et al., 2015), and indicator for preterm cesarean delivery for increased risk of neonatal and childhood morbidity (Mendez-Figueroa et al., 2015). In our multivariable model analysis, metabolic acidosis at birth had an independent relationship with the cardiotocography FHR mean and variability and also with the type of delivery (Caesarian section over vaginal) and parity. Our results are in agreement with the study of Heinonen et al. who also found that parity, but not maternal age, was an independent risk factor for neonatal acidosis (Heinonen and Saarikoski, 2001). Corroborating our results with another study of women with a singleton term pregnancy that found previous cesarean delivery and nulliparity as risk factors for neonatal metabolic acidosis (Westerhuis et al., 2012) may indicate that not only previous but also the current cesarean delivery may represent an actual challenge to the fetus. In two studies of poor neonatal adaptation at birth with severe neonatal acidosis (umbilical artery pH less than 7.10) independent risk factors included abnormal cardiotocography FHR, maternal age 35 years or older, parity, prior neonatal death or cesarean delivery (Maisonneuve et al., 2011; Crovetto et al., 2017). Our data extends a valuable role for cardiotocography FHR in predicting neonatal acidosis in deliveries with Apgar 5 ranging from fairly low to normal without neurological complications.

Heart rate variability (HRV) analysis with search for new algorithms is commonly employed to measure alterations in autonomic tone with predictive value in diseases (Campos et al., 2013). We have identified the CV of the heart rate as a sensitive measure of autonomic dysfunction (Miyabara et al., 2017) and independently associated with vascular atherosclerosis (Pereira et al., 2017). In this study, we found that CV of intrapartum cardiotocography FHR is an independent predictor of neonatal acidemia and Apgar scores.

### Limitations and Strengths of the Study

There are important limitations associated with this study. First, since this was a retrospective study, we could only investigate factors that were available to us in the current dataset, midto-long term outcomes of the infants being not recorded in the database. Also, fetal sleep state was not considered in this study. Second, although robust statistical methods were used to account for differences between groups, the potential for residual confounding cannot be ruled out. This study, as with most retrospective studies, is susceptible to bias. Third, the results of the study cannot be generalizable to all patients, for instance to those with an APGAR score critically low or associated with neurological complications and did not include data from neonates with severe cases of neonatal morbidity, hypoxic ischemic encephalopathy, or seizures. Limitations notwithstanding there are key strengths to this paper. To our knowledge, this is the first study with a relatively large sample size providing sufficient statistical power to explore the relationship between intrapartum FHRV and neonatal acidemia or APGAR score. Also, we adjusted for several confounders including important comorbidities.

# CONCLUSION

In conclusion, the present findings provide evidence that intrapartum cardiotocography FHR could be a predictor for neonatal acidemia and the physical condition of a newborn infant, as determined by arterial cord blood pH and base deficit, Apgar 1, and 5 scores. Further studies are warranted to examine the potential relationship between intrapartum FHR and neonatal pathologies including neurological complications.

# AUTHOR CONTRIBUTIONS

OB, LC, AB, MD, TM, and DdS conceived and designed the study, and analyzed and interpreted the data. OB, LC, MD, and AB drafted the manuscript. TM, MD, DdS, AB, LC, and OB critically revised the manuscript. LC and OB are the guarantors of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

# FUNDING

This research was supported by São Paulo Research Foundation to LC (FAPESP 17/11976-0).

### ACKNOWLEDGMENTS

fphys-09-01489 October 19, 2018 Time: 16:55 # 6

We gratefully acknowledge Prof. Dr. Václav Chudáèek and colleagues for making available the cardiotocography database to

### REFERENCES


PhysioNet. No assistance in the preparation of this article is to be declared. We thank the reviewers for their thoughtful review of the manuscript. They raised important issues and their inputs were very helpful for improving the manuscript.

heart rate: neonatal and neurologic morbidity. Obstet. Gynecol. 125, 636–642. doi: 10.1097/AOG.0000000000000673


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Medeiros, Dobre, da Silva, Brateanu, Baltatu and Campos. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Hemoglobin Changes After Long-Term Intermittent Work at High Altitude

Almaz Akunov1,2† , Akylbek Sydykov1,3† , Turgun Toktash<sup>4</sup> , Anara Doolotova<sup>4</sup> and Akpay Sarybaev1,2 \*

<sup>1</sup> Department of Mountain and Sleep Medicine and Pulmonary Hypertension, National Center of Cardiology and Internal Medicine, Bishkek, Kyrgyzstan, <sup>2</sup> Kyrgyz Indian Mountain Biomedical Research Center, Bishkek, Kyrgyzstan, <sup>3</sup> Excellence Cluster Cardio-Pulmonary System, Universities of Giessen and Marburg Lung Center, Member of the German Center for Lung Research (DZL), Justus Liebig University Giessen, Giessen, Germany, <sup>4</sup> Medical Department, Kumtor Gold Company, Bishkek, Kyrgyzstan

Chronic high altitude hypoxia leads to an increase in red cell numbers and hemoglobin concentration. However, the effects of long-term intermittent hypoxia on hemoglobin concentration have not fully been studied. The aim of this study was to evaluate hemoglobin levels in workers commuting between an elevation of 3,800 m (2-week working shift) and lowland below 1,700 m (2 weeks of holiday). A total of 266 healthy males, aged from 20 to 69 years (mean age 45.9 ± 0.6 years), were included into this study. The duration of intermittent high altitude exposure ranged from 0 to 21 years. Any cardiac or pulmonary disorder was excluded during annual check-ups including clinical examination, clinical lab work (blood cell count, urine analysis, and biochemistry), ECG, echocardiography, and pulmonary function tests. The mean hemoglobin level in workers was 16.2 ± 0.11 g/dL. Univariate linear regression revealed an association of the hemoglobin levels with the years of exposure. Hemoglobin levels increased 0.068 g/dL [95% CI: 0.037 to 0.099, p < 0.001] for every year of intermittent high altitude exposure. Further, after adjusting for other confounding variables (age, living at low or moderate altitude, body mass index, and occupation) using multivariable regression analysis, the magnitude of hemoglobin level changes decreased, but remained statistically significant: 0.046 g/dL [95% CI: 0.005 to 0.086, p < 0.05]. Besides that, a weak linear relationship between hemoglobin levels and body mass index was revealed, which was independent of the years of exposure to high altitude (0.065 g/dL [95% CI: 0.006 to 0.124, p < 0.05]). We concluded that hemoglobin levels have a linear relationship with the exposure years spent in intermittent hypoxia and body mass index.

Keywords: chronic intermittent hypoxia, high altitude, shift workers, hemoglobin, body mass index

# INTRODUCTION

Last decades, the number of people traveling to high altitude is increasing in connection with economic or recreational purposes (West, 2012). High altitude environment poses many challenges, but exposure to alveolar hypoxia is the most prominent among them (Burtscher et al., 2018). The ambient hypoxia triggers a number of physiologic responses including hyperventilation, increased resting heart rate and stimulation of erythrocyte production with the goal

### Edited by:

Jean-Paul R- Richalet, Université Paris 13, France

### Reviewed by:

Michele Samaja, Università degli Studi di Milano, Italy Julio Brito, Arturo Prat University, Chile

> \*Correspondence: Akpay Sarybaev

ak\_sar777@mail.ru

†These authors have contributed equally to this work

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 31 March 2018 Accepted: 16 October 2018 Published: 01 November 2018

### Citation:

Akunov A, Sydykov A, Toktash T, Doolotova A and Sarybaev A (2018) Hemoglobin Changes After Long-Term Intermittent Work at High Altitude. Front. Physiol. 9:1552. doi: 10.3389/fphys.2018.01552

**332**

of maintaining the oxygen content of arterial blood at or above sea level values (West, 2004). In permanent high altitude residents, exposure to chronic hypoxia leads to an increase in erythrocyte numbers and hemoglobin concentration (Leon-Velarde et al., 2000).

A large group of people, such as workers of mining companies, employees of road construction companies or military divisions in high-altitude border areas are exposed repetitively to high altitude over a long time by commuting between high altitude and lowland (Powell and Garcia, 2000; Richalet et al., 2002; West, 2002; Farias et al., 2013). However, the effects of the long-term exposure to intermittent hypoxia on hemoglobin concentration have been less well studied. Only few studies have assessed hematological changes induced by long-term intermittent high altitude exposure. Some studies reported a marked increase in hemoglobin (or hematocrit) levels in response to long-term exposure to intermittent high altitude (Gunga et al., 1996; Heinicke et al., 2003), whereas others did not find a significant effect (Richalet et al., 2002; Brito et al., 2007, 2018). Therefore, the aim of the current study was to determine the association of long-term intermittent high altitude exposure with hemoglobin levels as well as to explore the nature of this possible relationship.

# MATERIALS AND METHODS

### Subjects and Study Design

This is a cross-sectional study of mine employees exposed repetitively to high altitude for long periods of time. All of them work at the gold mine run by the Kumtor Gold Company (Centerra Gold Inc., Canada). Mine employees are transported to the mine site by bus, and the ascent lasts 7 h in total.

Every worker undergoes an annual medical checkup in a specially designated clinic in Bishkek (Kyrgyzstan, 760 m), which includes physical examination, clinical lab work (complete blood count, urine analysis, and biochemistry), ECG, echocardiography, exercise testing and spirometry. All the male employees of the Kumtor Gold Company, who had medical checkup during October–December 2016, were enrolled into the study. The inclusion criteria were male sex, uninterrupted working at high altitude in 14×14 shifts (14 days of working at altitude followed by a resting period of 14 days at lowland) and a healthy status without serious cardiopulmonary comorbidities. The exclusion criteria were chronic obstructive pulmonary disease, moderate or severe hypertension, and coronary heart disease.

In total, 266 healthy males (aged 45.9 ± 0.6 years) were included into the study. Subjects were truck drivers, gold mill operators, camp and kitchen staff, security personnel, and engineers. All subjects commuted between high altitude and living place (760 m or 1,600 m) in a 14×14 shift regimen. The miners slept at 3,800 m and worked a 12-h day shift at 3,800 to 4,000 m, only few subjects (operators of drilling rigs) went to 4,500 m.

Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki. This study was approved by the Research Ethics Committee of the National Center of Cardiology and Internal Medicine, Bishkek, Kyrgyzstan.

# Measured Variables

Measures were taken once at low altitude in a specially designated clinic in Bishkek during the annual medical check-up. Usually, the annual medical check-up is carried out 1 week after descent. Any cardiac or pulmonary disorder was excluded based on the results of clinical examination, clinical lab work (blood cell count, urine analysis, and biochemistry), ECG, echocardiography, and pulmonary function tests.

### General Data

Weight and height measured by means of a digital weight scale and a stadiometer with each participant barefoot and wearing underwear. Body mass index values were calculated as weight in kilograms divided by height in meters squared and rounded to 1 decimal place.

### Hematological Measurements

Complete blood count was performed by an automated hematology analyser Mindray BC-2300 (Guangzhou Shihai Medical Equipment Co., Ltd., China) using 15 µm of EDTA whole venous blood according to the manufacturer's instructions. Hematocrit, concentration of hemoglobin, erythrocytes, leukocytes, and platelets were measured. The analyser uses electrical impedance method for cell counting and cyanide free method for hemoglobin detection. Red blood cell indices included hemoglobin concentration and mean corpuscular volume.

## Data Analysis

Statistical analysis was performed using SPSS version 23.0 for Windows (IBM, Chicago, IL, United States). Data are expressed as mean ± standard error. The Kolmogorov–Smirnov test was used to assess the normality of distribution of quantitative variables. All quantitative variables were distributed normally. Linearity was assessed through examination of various bivariate scatterplots. Linear regression (univariate and multivariable) analysis was used to check the association between hemoglobin levels and other parameters. Differences between different BMI categories were assessed by one-way ANOVA followed by Tukey's multiple comparisons post hoc test. For comparing proportions, we used chi-squared test. For linear trend in quantitative variables we used one-way ANOVA linear contrast method. The values P < 0.05 were considered as statistically significant. Statistical power analysis was performed using G∗Power 3.1.

# RESULTS

Characteristics describing the subjects are provided in **Table 1**. One hundred and eighty-four workers were residents of moderate altitude (1,600 m); the remaining 82 men were residents of low altitude (760 m). The duration of exposure to intermittent high altitude ranged from 0 to 21 years. The average record of service was 10.1 years. The majority of subjects were ethnic Kyrgyz

(92%), but there were some Caucasians too (8%). Prevalence of obesity in the sample was 16.1 ± 2.4%. The mean hemoglobin level in workers was 16.2 ± 0.11 g/dL (**Table 1**). **Table 2** shows higher hemoglobin levels, as well as age and BMI, in subjects with longer duration of high altitude exposure. It is clear that hemoglobin levels are higher when the duration of exposure is longer (p < 0.001). Notably, the differences in age and BMI were also statistically significant; therefore, we cannot ascribe the differences in hemoglobin levels to the effects of high altitude alone. Indeed, hemoglobin values were significantly higher in obese subjects compared to those with normal weight (**Figure 1**).

Univariate linear regression revealed an association of the hemoglobin levels with the years of exposure (**Figure 2**). In a univariate regression model, every consecutive year was associated with an increase in hemoglobin of 0.068 g/dL [95% CI: 0.037 to 0.099, p < 0.001].

Further, after adjusting for other variables (age, living at low or moderate altitude, BMI, and occupation) using multivariable regression analysis, the magnitude of hemoglobin level changes decreased, but remained statistically significant: 0.046 g/dL [95% CI: 0.005 to 0.086, p < 0.05]. Despite a significant association of the hemoglobin levels with age by univariate regression analysis, it failed to prove the significance after adjusting for the rest variables (BMI, years of exposure, altitude of residence (0.006 g/dL [95% CI −0.024 to −0.035]) (**Table 3**). However, BMI and duration of exposure retained to have a weak but significant relationship with hemoglobin levels (0.065 g/dL [95% CI: 0.006

TABLE 1 | Anthropometric hematological characteristics of the subjects (n = 266).


to 0.124] and 0.046 g/dL [95% CI: 0.005 to 0.086], respectively) (**Table 3**).

### DISCUSSION

To our best knowledge, this is the first study performed on a large population of Kyrgyz workers intermittently exposed to high altitude for a very long period up to 21 years. We showed that there is a statistically significant correlation between the hemoglobin levels and the years of exposure. Using multivariable regression analysis, we showed that hemoglobin levels increase by an average of 0.046 g/dL for every consecutive year of intermittent high altitude exposure, after adjusting for other variables (age, living at low or moderate altitude, BMI, and occupation). Interestingly, the residence at low or moderate altitudes did not affect the hemoglobin levels. This may be due to the relatively small difference in the altitude of residence (less than 900 m).

Chronic high altitude hypoxia leads to an increase in red cell numbers and hemoglobin concentration. Previous studies have shown that permanent high altitude residents possess elevated hemoglobin levels and hematocrit values (Leon-Velarde et al., 2000). In sea-level residents, hematocrit and hemoglobin concentration were elevated after exposure to 3,550-m altitude for 8 months; however, none of the parameters reached pathological values (Siques et al., 2007). Only few studies have assessed hematological changes induced by long-term intermittent high altitude exposure in a comprehensive manner. An earlier study in Chilean workers exposed to intermittent high altitude for >5 years revealed hemoglobin values that were comparable to those reported in the literature for high altitude populations (Gunga et al., 1996). Similarly, increased hemoglobin concentration and hematocrit values (15.8 ± 1.2 g/dL and 46.2 ± 3.8%, respectively) were shown in Chilean army officers exposed to intermittent hypoxia for about 22 years after 3 days following descent to sea level (Heinicke et al., 2003). Remarkably, hemoglobin concentration and hematocrit values (16.5 ± 0.9 g/dL and 48.1 ± 2.9%, respectively) measured at high altitude were comparable to those found in permanent high altitude residents (Heinicke et al., 2003). Another study conducted in a group of Chilean army officers exposed to intermittent high altitude for at least 12 years (50% had been >25 years at altitude) reported a smaller increase in hemoglobin

TABLE 2 | Distribution of the hemoglobin levels, age, body mass index, and residence place of the subjects according to the years of high altitude exposure.


Data are presented as mean ± SEM (n = 266). Hb, hemoglobin; BMI, body mass index. p for linearity (hemoglobin, age, BMI) was determined using one-way ANOVA linear contrast analysis; p for trend (residents) was determined using chi<sup>2</sup> -criterion.

FIGURE 2 | Scatter plot of correlation between the hemoglobin (Hb) levels and the years of exposure to intermittent high altitude. The best-fit line is shown, and the shaded area represents the 95% confidence intervals (n = 266).

concentration and hematocrit values (15.1 ± 1.0 g/dL and 45.02 ± 2.7%, respectively) measured on day 1 following descent to sea level (Brito et al., 2007). Similarly, a period of 2.5 year exposure to intermittent hypoxia induced a significant hematocrit increase in Chilean miners, which was, however, lower than what is observed in permanent high altitude residents (Richalet et al., 2002). A recent cross-sectional study in healthy Chilean male miners working at an altitude of 4,400 or 4,800 m for, on average, 14 years reported mean hematocrit and hemoglobin values of 47.6 ± 0.3% and 16.2 ± 0.1 g/dL, respectively, with none of the subjects having pathological values (Brito et al., 2018). In our study, hemoglobin concentration and hematocrit values in shift workers were higher than those observed in sea level residents, but were lower than values reported in Aymara native residents at 3,800–4,065 m (Beall et al., 1998).

Various factors may be responsible for the discrepancy. A meta-analysis and a Monte Carlo simulation on the extracted data showed that red cell volume expansion for a given duration of exposure is dependent on the altitude (Rasmussen et al., 2013). The authors suggested that, at altitudes above 4,000 m, exposure time must exceed 2 weeks to exert a significant effect and that the magnitude of the erythropoietic response depends on the initial red cell volume (Rasmussen et al., 2013). In addition, hemoglobin mass returns to baseline sea level values in 2–3 weeks following descent to sea level (Wachsmuth et al., 2013). Moreover, the mining companies and military divisions in different countries implement various commuting patterns complicating the interpretation of the hematological changes induced by the long-term intermittent high altitude exposure. The most common shift patterns range from 4×3 days to 28×28 days (Gunga et al., 1996; Richalet et al., 2002; Heinicke et al., 2003; Sarybaev et al., 2003; Brito et al., 2007; Vinnikov et al., 2011). Furthermore, ethnic differences have been shown in hematological responses in high altitude residents (Beall, 2006). If ethnicity affects hematological responses to chronic intermittent high altitude exposure is not known.

Another explanation could be the various duration of stay at high altitude and the elevation achieved in different studies Therefore, a term "hypoxic dose" has recently been introduced as a new metric incorporating both altitude and total exposure time to measure an increase in hemoglobin mass because of intermittent hypoxic training or intermittent hypoxic exposure (Garvican-Lewis et al., 2016a). The models suggest that hypoxic exposure in excess of 250 km h is sufficient to produce an increase in hemoglobin mass (Garvican-Lewis et al., 2016b). In our study, the workers were exposed to the hypoxic dose of about 1,300 km h, which is considered to be sufficient to induce an increase in hemoglobin levels.

Univariate analysis showed a relationship of hemoglobin levels with aging; however, after adjusting for other variables, the value ceased to be significant. These results are in line with the findings from other studies. Thus, no significant hematocrit change over 11.6 years was demonstrated in a recent prospective study (Carallo et al., 2011). In another study, while fibrinogen, another blood viscosity parameter, increased with age, the hemoglobin level on the contrary slightly decreased in the elderly subjects (Coppola et al., 2000). Notably, we revealed a significant relation of hemoglobin levels with the BMI in workers exposed to intermittent high altitude for a long period. It is in accordance with other studies that demonstrated a positive association between hemoglobin and BMI in lowland populations (Micozzi et al., 1989; Skjelbakken et al., 2010). In contrast, no correlation between hemoglobin levels and BMI was reported by others (Ghadiri-Anari et al., 2014). Nevertheless, in a recent prospective study of high-altitude mine workers in Peru, male gender, duration of the intermittent high altitude exposure and BMI were independent predictors of hemoglobin level changes (Mejia et al., 2017). The underlying mechanisms remain poorly understood. However, a negative effect of BMI on oxygen saturation was demonstrated in Chinese Han young males during high altitude acclimatization (Peng et al., 2013). Interestingly, a negative association between BMI and blood oxygenation was also found in healthy high altitude residents in Peru (Pereira-Victorio et al., 2014; Miele et al., 2016). These data



BMI, body mass index. <sup>∗</sup>adjusted for the other variables. R<sup>2</sup> = 0.07.

suggest that shift workers with higher BMI will have a greater increase in hemoglobin during chronic intermittent high altitude exposure.

It has been shown that borderline polycythemia (hematocrit above 50%) was associated with increased mortality from coronary heart disease (Kunnas et al., 2009). However, the connection between hemoglobin and cardiovascular diseases is complex and is still not clear (Brown et al., 2001; Frackiewicz et al., 2018). We have to admit that prolonged intermittent high altitude hypoxia can provoke cardiovascular events in patients with clinical or subclinical coronary atherosclerosis because of blood rheology changes. Interestingly, accumulating evidence demonstrates that shortterm daily sessions of hypoxia alternating with equal durations of normoxia for 2–3 weeks exert beneficial effects on various cardiovascular diseases (Serebrovskaya and Xi, 2016), thus potentially opposing deleterious effects of polycythemia. At the same time, the increase in hemoglobin levels due to chronic intermittent high altitude exposure should not lead to increased probability of cardiovascular events per se.

Although the literature on the relation between changes in hemoglobin concentration and cardiovascular disorders in subjects exposed to long-term intermittent high altitude hypoxia is very scarce, we believe that such a small, though statistically significant, increase in hemoglobin concentration represents an adaptive response rather than a risk factor for cardiovascular diseases. An annual increase in hemoglobin level by 0.046 g/dL means that for every 10 years of work at intermittent high altitude hypoxia hemoglobin levels will rise, on average, less than 0.5 g/dL. Indeed, in most of the subjects chronically exposed to intermittent high altitude, hemoglobin concentration and hematocrit reach intermediate values that are higher than those in sea level residents but lower than those in healthy high altitude dwellers (Richalet et al., 2002; Brito et al., 2007). It is unlikely that this rather small increase in hemoglobin levels would have any significant pathological effects (Burtscher, 2014; Corante et al., 2018).

Additionally, another issue related to hemoglobin levels at high-altitude is the change in skeletal muscle mass, since muscles are the main consumer of oxygen in the body. While some studies indicated some decrease in skeletal muscle mass at high-altitude exposure (Hoppeler et al., 1990; MacDougall et al., 1991; Mizuno et al., 2008), the others were not able to show significant loss of muscle mass (Lundby et al., 2004; D'Hulst et al., 2016; Jacobs et al., 2016). Differences in the hypoxic doses may account for this discrepancy. Thus, it has been suggested that a minimum hypoxic exposure of 5,000 km h is required for hypoxia-induced muscle atrophy to develop (D'Hulst and Deldicque, 2017). Although, we have not measured the skeletal muscle fiber cross sectional area changes in the workers, it seems unlikely that the relatively low hypoxic dose in our study can significantly affect the skeletal muscle.

Another kind of intermittent hypoxia which deserves further commentary is obstructive sleep apnea syndrome (OSAS). OSAS is a pathological condition characterized by recurrent or cyclic short periods of isobaric hypoxia and asphyxia during sleep (coupled with oxygen desaturation), often more than 60 times per hour (Neubauer, 2001). Such frequent fluctuations of oxygen saturation lead to sympathetic overactivity, increased oxidative stress, and activation of inflammatory response pathways (Mansukhani et al., 2014; Passali et al., 2015; de Lima et al., 2016). In addition to hypoxemia, these events are associated with significant hypercapnia. Consequently, OSAS is an independent and well-known major risk factor for various cardiovascular diseases, including hypertension, stroke, myocardial infarction, and congestive heart failure (Kendzerska et al., 2014; Drager et al., 2015; Bauters et al., 2016). In contrast, chronic intermittent hypoxic exposure in miners involves prolonged cycles of hypobaric hypoxia alternating with normobaric normoxia. The frequency of high altitude intermittent hypoxia is usually one to four times a month, which does not have such unfavorable effects on the body. Thus, as previously pointed out (Richalet et al., 2002), OSAS is quite different from the chronic intermittent hypoxic exposure in miners.

One of the limitations of this study is its cross-sectional nature. Further, most of the subjects were permanent residents of a moderate altitude, so the difference in altitude between the residence place and working altitude was relatively small. Nevertheless, the strengths of our study include high number of subjects, which provided the high statistical power, and larger duration of intermittent high altitude exposure than in most other studies.

### CONCLUSION

In summary, this study adds to the growing body of knowledge concerning the physiology of long-term intermittent

high altitude exposure. We defined for the first time hemoglobin levels in Kyrgyz shift workers commuting between high altitude and lowland. Further, our study provides evidence that hemoglobin levels have a linear relationship with years of intermittent high altitude exposure and BMI. Apparently, in chronic intermittent hypoxia exposure even over longer periods, excessive erythrocytosis does not represent a major problem for healthy shift workers. Our findings might have important implications for occupational medical surveillance to monitor health status of the workers exposed to chronic intermittent high altitude. We hope that our study would encourage further research to explore the long-term consequences of this unique biological condition.

### REFERENCES


### AUTHOR CONTRIBUTIONS

AA, AbS, and ApS conceived and designed the study, drafted the manuscript, and provided overall supervision. AA, AbS, TT, and AD performed the data acquisition, analysis, or interpretation. AA, AbS, TT, AD, and ApS critically revised important intellectual content in the manuscript and approved the final version of the manuscript.

### FUNDING

This work was supported by Ministry of Education and Science of the Kyrgyz Republic (Grant No. 0005823).


living at 4,355, 4,660, and 5,500 meters above sea level. High Alt. Med. Biol. 1, 97–104. doi: 10.1089/15270290050074233


**Conflict of Interest Statement:** TT and AD were employed by the Kumtor Gold Company.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Akunov, Sydykov, Toktash, Doolotova and Sarybaev. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Rat Strain and Housing Conditions Alter Oxidative Stress and Hormone Responses to Chronic Intermittent Hypoxia

Brina Snyder, Phong Duong, Mavis Tenkorang, E. Nicole Wilson and Rebecca L. Cunningham\*

Department of Physiology and Anatomy, University of North Texas Health Science Center, Fort Worth, TX, United States

Sleep apnea has been associated with elevated risk for metabolic, cognitive, and cardiovascular disorders. Further, the role of hypothalamic–pituitary–adrenal (HPA) activation in sleep apnea has been controversial in human studies. Chronic intermittent hypoxia (CIH) is a rodent model, which mimics the hypoxemia experienced by patients with sleep apnea. Most studies of CIH in rats have been conducted in the Sprague Dawley rat strain. Previously published literature suggests different strains of rats exhibit various responses to disease models, and these effects can be further modulated by the housing conditions experienced by each strain. This variability in response is similar to what has been observed in clinical populations, especially with respect to the HPA system. To investigate if strain or housing (individual or pair-housed) can affect the results of CIH (AHI 8 or 10) treatment, we exposed individual and pair-housed Sprague Dawley and Long-Evans male rats to 7 days of CIH treatment. This was followed by biochemical analysis of circulating hormones, oxidative stress, and neurodegenerative markers. Both strain and housing conditions altered oxidative stress generation, hyperphosphorylated tau protein (tau tangles), circulating corticosterone and adrenocorticotropic hormone (ACTH), and weight metrics. Specifically, pair-housed Long-Evans rats were the most sensitive to CIH, which showed a significant association between oxidative stress generation and HPA activation under conditions of AHI of 8. These results suggest both strain and housing conditions can affect the outcomes of CIH.

Keywords: hypothalamic–pituitary–adrenal axis, reproducibility, oxidative stress, corticosterone, ACTH, hyperphosphorylated tau

### INTRODUCTION

There is a lack of consensus in the literature related to the basic scientific model of sleep apnea, chronic intermittent hypoxia (CIH). CIH has been reported to be both protective and damaging in subsequent stroke outcomes. Elevated mean arterial pressure, inflammation, oxidative stress, and cognitive impairment have been described in models of CIH (Cai et al., 2010; Smith et al., 2013; Rosenzweig et al., 2014; Briancon-Marjollet et al., 2016; Shell et al., 2016; Snyder et al., 2017), while other studies have reported lower oxidative stress and pre-conditioning effects of CIH (Zhen et al., 2014; Yuan et al., 2015). These divergent reports further complicate interpretation of the

### Edited by:

Rodrigo Del Rio, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Angela J. Grippo, Northern Illinois University, United States Noah J. Marcus, Des Moines University, United States

### \*Correspondence:

Rebecca L. Cunningham rebecca.cunningham@unthsc.edu

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 08 July 2018 Accepted: 17 October 2018 Published: 06 November 2018

### Citation:

Snyder B, Duong P, Tenkorang M, Wilson EN and Cunningham RL (2018) Rat Strain and Housing Conditions Alter Oxidative Stress and Hormone Responses to Chronic Intermittent Hypoxia. Front. Physiol. 9:1554. doi: 10.3389/fphys.2018.01554

**339**

CIH animal model and consequently our understanding of sleep apnea (Navarrete-Opazo and Mitchell, 2014). The pivotal factors contributing to these dichotomous observations of CIH are most likely the frequency and severity of the intermittent hypoxia used in each study (Navarrete-Opazo and Mitchell, 2014). There appears to be a threshold in which studies using a more frequent normal room air to low oxygen cycle per hour, modeling the apnea/hypopnea index (AHI), result in damaging effects, while models with very slow air changes per hour report protective mechanisms.

In addition to differences in CIH protocols, prior studies have been conducted on various rat strains under different housing conditions. Our laboratory has exposed single-housed Sprague Dawley (Snyder et al., 2017), pair-housed Brown Norway (Wilson et al., 2018), and pair-housed Long-Evans (Snyder et al., 2018) rat strains to CIH with varying oxidative stress responses. Most CIH protocols have been performed on either Sprague Dawley or Wistar rat strains, but not the Long-Evans or Brown Norway rat strains. Generally, housing conditions are not reported (Ramanathan et al., 2005; Cai et al., 2010; Williams et al., 2015; Luneburg et al., 2016). Of the studies that include housing conditions, it is interesting to note that publications using singlehoused Sprague Dawley rats report increased oxidative stress in response to CIH (Shell et al., 2016; Snyder et al., 2017; Li et al., 2018). Therefore, rat strain and housing conditions may be important variables.

The scientific community has recognized a number of factors exist which contribute to contradictory reports in many basic science models. In recognition of this, the National Institutes of Health (NIH) updated their mission of Rigor and Responsibility to improve reproducibility (Hewitt et al., 2017). Current variables recognized as having an impact on experimental outcomes are sex, age, weight, and current health status. Here, we present results supporting model strain and housing conditions as key components affecting the observed outcomes in our study using CIH. These results provide evidence that the variable, housing conditions, should be included in future studies.

# MATERIALS AND METHODS

## Animals

Two outbred rat strains were used in this study – Sprague Dawley and Long-Evans. Adult male Sprague Dawley and Long-Evans rats (both strains 58–64 days old, 250–275 g body weight, Charles River) were housed in a temperature-controlled environment with the lights on a 12:12 h cycle. The Sprague Dawley rat strain is a non-aggressive rat strain, whereas the Long-Evans rat strain is more aggressive and active than Sprague Dawley rats (Adams and Blizard, 1987; Henry et al., 1993; Turner and Burne, 2014). Thus, we examined single-housing versus pairhousing in both rat strains. Upon arrival, animals were either housed individually or pair-housed with an unfamiliar rat of similar size and weight for the remainder of the experiment. Food and water were provided ad libitum. Animals were weighed each week during cage cleaning and at the end of testing. Final body weights were used for reporting purposes. All experiments were conducted according to National Institute of Health guidelines on laboratory animals and approved by the Institutional Care and Use Committee at UNT Health Science Center.

# Chronic Intermittent Hypoxia (CIH)

Hypoxic A-Chambers and OxyCycler A84XOV controllers were purchased from BioSpherix, Ltd. (Parrish, NY, United States). One week after arrival, rats were separated into either normoxic or CIH treatment groups with at least 10 animals per group. This resulted in eight treatment groups: single-housed Sprague Dawley normoxic (n = 13), single-housed Sprague Dawley CIH (n = 12), pair-housed Sprague Dawley normoxic (n = 17), pairhoused Sprague Dawley CIH (n = 16), single-housed Long-Evans normoxic (n = 12), single-housed Long-Evans CIH (n = 12), pairhoused Long-Evans normoxic (n = 12), and pair-housed Long-Evans CIH (n = 12). Home cages, each containing either single or pair-housed animals, were placed into each A-chamber for acclimation to the apparatus for 1 week at normoxic conditions (21% oxygen). Acclimation to the chambers was followed by CIH exposure for 7 days from 8 am to 4 pm during the light (sleep) phase. Our CIH protocol utilized 8-min cycles of low oxygen (10%) followed by reoxygenation (21%) for 8 h during the light phase to model an AHI 8 (Epstein et al., 2009; Snyder et al., 2018; Wilson et al., 2018). Specifically, nitrogen was injected into the chamber over a period of 5 min to reach a low oxygen concentration of 10%, followed by injection of oxygen over 3 min to return to and maintain normal room air concentrations (21%). For the remaining 16 h, animals were exposed to room air. To control for sleep deprivation, due to noises from the CIH apparatus, normoxic controls were housed under similar conditions but not administered hypoxia.

An additional group of single-housed Sprague Dawley normoxic (n = 3), single-housed Sprague Dawley CIH (n = 8), pair-housed Long-Evans normoxic (n = 4), and pair-housed Long-Evans CIH (n = 8) were exposed to a CIH protocol injecting nitrogen over a period of 3 min to reach an oxygen concentration of 10% and reoxygenation for 3 min to mimic an AHI 10 to compare results to previously published observations. Both AHI of 8 and 10 protocols result in oxygen nadirs of 10% lasting 75 s. One limitation of this CIH animal model for sleep apnea is that the nadir is longer than a typical apnea observed in patients with obstructive sleep apnea. However, the nadir duration is similar to the 15–60 s of hypoxemia experienced by obstructive sleep apnea patients (Dewan et al., 2015). This difference in nadir duration is primarily due to the size of the chambers that house the rats in their home cages and the amount of time needed for gas exchanges.

### Sample Collection

Between 08:00 and 10:00 on the morning following the final CIH exposure, which was during the first 2 h of the light phase, animals were anesthetized with isoflurane (2–3%) and sacrificed by decapitation, as previously described (Snyder et al., 2017). Trunk blood was collected in 7 mL EDTA tubes. The samples were then centrifuged at 2,000 × g for 10 min at 4◦C. Plasma

was removed and aliquoted for storage in microcentrifuge tubes at −80◦C until assayed. Whole brains were immediately removed and flash frozen in PBS and coronally sliced to reveal brain nuclei of interest: the entorhinal cortex, and dentate gyrus (DG), CA1, and CA3 regions of the hippocampus. These nuclei were selected based on oxidative stress damage due to CIH that was observed in a previous publication (Snyder et al., 2017). Nuclei were dissected by micropunch as previously described (Snyder et al., 2017) and stored in microcentrifuge tubes at −80◦C until homogenized for protein analysis.

### Advanced Oxidative Protein Products (AOPP) Assay

Plasma oxidative stress was assayed using Cell Biolabs, Inc. OxiSelect Advanced Oxidative Protein Products assay kit, according to our previously published protocol (Cunningham et al., 2011; Snyder et al., 2017, 2018). This kit measures the amount (uM) of all oxidized proteins in the sample relative to a known standard. Chloramine in the kit reacts with oxidized proteins to produce a color change which can be read at 340 nm. Assay results were reported as percent of control [individual value/(average of normoxic control values) × 100].

# Hormone Measurements

# Corticosterone

Circulating nadir corticosterone was assayed using a commercially available competitive immunoassay (Corticosterone Mouse/Rat ELISA kit, RTC002R, BioVendor, Brno, Czechia), according to manufacturer's instructions. Sensitivity of the assay was 6.1 ng/ml at the 2 SD confidence limit. The intra-assay coefficient of variation was 7.37% and the inter-assay coefficient of variation was 7.63%. Specificity of this assay is as follows: corticosterone (100%), cortisol (2.3%), aldosterone (0.3%), testosterone (<0.1%), progesterone (6.2%), and androsterone (<0.1%). Results are expressed as ng/ml.

### Adrenocorticotropic Hormone

Plasma adrenocorticotropic hormone (ACTH) was assayed by double-antibody radioimmunoassay using <sup>125</sup>I (hACTH Double Antibody RIA Kit, 07-106102, MP Biomedicals, Solon, OH, United States), according to manufacturer's protocol. Samples were performed in duplicates and the assay was measured using a gamma counter (Cobra Auto-Gamma, LPS Biomedical Instrument Services, Redmond, WA, United States), with counting time of 3 min per sample at 80% efficiency. The intraassay coefficient of variation was 5.45% and the inter-assay coefficient of variation was 7.30%. The specificity of the assay is as follows: ACTH1−<sup>39</sup> (100%), ACTH1−<sup>24</sup> (100%), hβ lipotropin (0.8%), hα lipotropin (0.1%), hβ endorphin (<0.1%), hα MSH (<0.1%), and hβ MSH (<0.1%). The following formula was used to determine the %B/B0:

[(CPMSanple − CPMNSB)/(CPM0Standard − CPMNSB)] × 100

in which CPM = counts per minute, NSB = non-specific binding (blank), 0 standard = total binding (B0). The %B/B0 for the unknowns were then plotted against the standard %B/B0 using a 4 parameter-log function to determine ACTH concentration (analysis). Results are expressed as ng/ml.

### Tissue Homogenization

Frozen tissue samples were thawed in 50 uL RIPA solution (Amresco) containing 3 uM phosphatase inhibitor (Sigma-Aldrich), 1 uM EDTA (Sigma Aldrich), and 1 uM dithiothreitol (Sigma-Aldrich) as previously described (Snyder et al., 2017). Protein quantification was assessed using a commercially available Pierce BCA Protein Assay Kit (Thermo Fisher) and absorbance read at 562 nm to determine sample volumes for further analysis. Samples were stored at −80◦C.

### Immunoblotting

Equal volumes of tissue samples containing 20 uL protein were loaded into a Bio-Rad 4–20% polyacrylamide gel for electrophoresis at 25 mA, followed by overnight transfer onto a PDVF membrane at 60 mA. Transfer was verified by Ponceau S staining, the washed for 30 min in TBST. Membranes were blocked for 60 min with 5% non-fat milk in TBS-Tween (TBST) at room temperature. Membranes were then transferred to a 1% non-fat milk TBST solution containing specific primary antibody for tau phosphorylated at S202 (Abcam ab108387, 1:10,000) and incubated overnight at 4◦C. Afterward, membranes were washed in 10 min increments for 30 min, and then incubated in 1% milk TBST secondary antibody solutions (goat anti-rabbit 1:10,000) at room temperature for 1 h. Protein bands were visualized using West Pico enhanced chemiluminescence detection assay (Thermo scientific) on an Syngene G:Box system using FlourChem HD2 AIC software. Membranes were then incubated in 1% non-fat milk TBST solutions containing primary antibody for GAP-DH (D16H11) XP (HRP conjugate) (Cell Signaling Technology #8884, 1:10,000) for 2 h at room temperature, followed by chemiluminescent visualization. Although prior studies have shown GAP-DH expression can be affected by hypoxic exposure, no significant differences due to CIH were observed in any of the brain regions examined. NIH ImageJ software (version 1.50i) was used to quantify band densitometry and values were normalized to GAP-DH values using the equation (mean gray value for protein/mean gray value of GAP-DH)<sup>∗</sup> 100.

### Statistical Analysis

IBM SPSS (SPSS v. 23, IBM, 2015) was used for statistical analysis. Three-way ANOVA was used to test for significant interactions between strain, housing condition, and hypoxic exposure. Fisher's LSD was used for post hoc analysis. Oneway ANOVAs were used to analyze the effect of hypoxia within individual brain regions. Results are shown as mean ± SEM. Statistical significance for all measurements was at p ≤ 0.05.

# RESULTS

fphys-09-01554 November 2, 2018 Time: 17:6 # 4

# Strain Differences in Chronic Intermittent Hypoxia-Induced Oxidative Stress

Previously, we published that exposure to CIH (AHI 10) induced oxidative stress in plasma and brain regions associated with neurodegeneration in single-housed Sprague Dawley male rats (Snyder et al., 2017). To investigate if this observation is maintained across strain and housing conditions, both Sprague Dawley and Long-Evans male rats were housed singly or in pairs and then exposed to 7 days mild CIH (AHI 8) or normoxic conditions. Pair-housed Sprague Dawley rats and single-housed Long-Evans rats did not experience an increase in oxidative stress following 7 days of CIH (**Figure 1**). However, a significant interaction between strain, housing, and hypoxia (F1,<sup>89</sup> = 5.842; p < 0.05) was observed. A significant elevation of oxidative stress was observed in pair-housed Long-Evans rats (139.84 ± 32.87%) (**Figure 1**) following CIH exposure. Unlike what is observed at AHI 10 (**Figure 2**), single-housed Sprague Dawley rats did not have a significant increase in oxidative stress due to CIH at AHI 8. At AHI 10, there were no significant differences between strains in the magnitude of the oxidative stress response to CIH (F1,<sup>14</sup> = 2.53, p = 0.13). Interestingly, under normoxic conditions Long-Evans rats exhibit higher circulating basal oxidative stress (312.3 ± 96.41 uM) than Sprague Dawley rats (219.69 ± 83.63 uM), regardless of housing condition (F1,<sup>51</sup> = 14.00, p < 0.05). These observations suggest Long-Evans rats are more sensitive to hypoxic insults than Sprague Dawley rats, which may be due to higher circulating basal oxidative stress in Long-Evans rats. These results indicate housing conditions can impact susceptibility to oxidative stress differently between strains.

### Strain Differences in Phosphorylated Tau Induced by Chronic Intermittent Hypoxia

Chronic intermittent hypoxia has been associated with cognitive impairments in rodent models of sleep apnea (Xu et al., 2004; Cai et al., 2010; Smith et al., 2013; Sapin et al., 2015; Snyder et al., 2018). To determine if strain or housing impact brain regions involved in memory pathways, we assessed the presence of tau hyperphosphorylated at serine 202 (p-tau), an indicator of tau tangle accumulation. The presence of mitochondrial oxidative damage has been reported concurrently with tau tangle accumulation in post-mortem temporal lobes of persons with preclinical stages of Alzheimer's disease (Terni et al., 2010).

Because we were most interested in processes that may be impacted by oxidative stress generation, we only used singlehoused Sprague Dawley rats and pair-housed Long-Evans rats, which exhibited CIH-induced plasma oxidative stress. Although we did not observe any differences in p-tau protein expression in the entorhinal cortex (ETC) or hippocampal structures at AHI 8 in any of our groups (data not shown), strain differences in the p-tau protein expression were observed under CIH condition of AHI 10. Significantly more p-tau in the ETC (F1,<sup>5</sup> = 12.457; p < 0.05) and CA1 (F1,<sup>7</sup> = 8.543; p < 0.05) hippocampal brain regions was observed in single-housed Sprague Dawley rats exposed to CIH at AHI 10 than their normoxic counterparts (**Figure 3A**). Similar to our observations with AOPP, pair-housed Long-Evans rats appear to be more sensitive to CIH exposure, with more p-tau protein expression in the dentate gyrus (DG) and CA1 (F1,<sup>8</sup> = 8.739; p < 0.05) of the hippocampus (F1,<sup>9</sup> = 5.340; p < 0.05), as well as in the ETC (F1,<sup>8</sup> = 8.800; p < 0.05), following CIH exposure at AHI 10 (**Figure 3B**).

### Strain Differences in Hypothalamus–Pituitary–Adrenal (HPA) Axis Activation HPA Hormone Differences

Corticosterone and its releasing hormone, ACTH, can be affected by housing conditions, and corticosterone may contribute to oxidative stress burden (Zafir and Banu, 2009; Spiers et al., 2015; Solanki et al., 2017). To determine if strain type or housing conditions produces a differential response to CIH

FIGURE 3 | Evidence of tau tangles are present in memory-associated brain nuclei following 7 days of chronic intermittent hypoxia (CIH) at apnea/hypopnea index (AHI) = 10. Tissue homogenate from each brain region investigated was run on separate gels for western blot analysis. Representative images from each membrane are grouped together and displayed in (A,B) for each strain of rat. Western blot analysis of hyperphosphorylated tau (p-tau) was performed on tissue homogenate from the entorhinal cortex (ETC), and dentate gyrus (DG), CA1, and CA3 regions of the hippocampus of normoxic (N) and CIH (C) exposed male rats. Evidence of CIH contribution to tau tangle accumulation was observed in a strain dependent manner. (A) Single-housed Sprague Dawley rats exhibit significantly elevated p-tau in the ETC and CA1 after CIH exposure. (B) Pair-housed Long-Evans rats exhibit significantly elevated p-tau in the ETC, as well as in the DG, and CA1 regions of the hippocampus. Results are reported as mean ± SEM (p-tau/GAP-DH expression), <sup>∗</sup> compared to normoxic control.


TABLE 1 | Hypothalamic–pituitary–adrenal axis hormone concentrations are affected by strain and hypoxia at AHI 8, with an interaction between housing and hypoxia.

Male Long-Evans rats have significantly higher circulating nadir corticosterone (CORT) and ACTH than Sprague Dawley rats. Results are reported as mean ± SD (ng/ml), <sup>∗</sup>compared to normoxic control; <sup>+</sup>compared to AHI 10; # interaction with housing; statistical significance was set at p ≤ 0.05.

at AHI 8, circulating nadir corticosterone and ACTH were assayed. A significant interaction between strain and CIH at an AHI 8 (F1,<sup>63</sup> = 5.988, p < 0.05) was observed in ACTH levels, with a main effect of both strain (F1,<sup>63</sup> = 75.735, p < 0.05) and CIH (F1,<sup>63</sup> = 13.475, p < 0.05) on plasma ACTH levels (**Table 1**). Under normoxic conditions, Long-Evans rats had significantly higher ACTH than Sprague Dawley rats. In pair-housed Long-Evans rats, CIH significantly decreased ACTH compared to normoxic ACTH levels. Following CIH at AHI 10, both single-housed Sprague Dawley rats (F1,<sup>8</sup> = 4.96, p < 0.05) and pair-housed Long-Evans rats (F1,<sup>14</sup> = 29.33, p < 0.05) had significantly lower ACTH than normoxic controls (**Table 1**). Further, ACTH following AHI 10 was significantly lower than after AHI 8 in both strains (Sprague Dawley single-housed: F1,<sup>14</sup> = 11.85, p < 0.05; Long-Evans pair-housed: F1,<sup>12</sup> = 13.33, p < 0.05; **Table 1**).

An interaction between housing and CIH (AHI 8) on corticosterone was also observed (F1,<sup>68</sup> = 6.015, p < 0.05). Similar to ACTH, Long-Evans pair-housed rats exposed to CIH had significantly lower corticosterone than their normoxic counterparts (**Table 1**). Since prior studies have shown that corticosterone can increase oxidative stress (Zafir and Banu, 2009; Spiers et al., 2015), we wanted to determine if there was a relationship between corticosterone and CIH-induced oxidative stress. Therefore, we investigated this association on our treatment groups (single-housed Sprague Dawley and pair-housed Long-Evans rats) that showed an elevation of oxidative stress in response to CIH. We observed a positive association between corticosterone and oxidative stress only in Long-Evans pair-housed rats exposed to CIH (**Figure 4A**). Neither the Long-Evans single-housed rats nor either of the Sprague Dawley rat cohorts exhibited an association between oxidative stress and nadir corticosterone following CIH exposure (**Figures 4A,B**).

Following CIH at AHI 10, Long-Evans pair-housed rats exhibited lower corticosterone than normoxic controls (F1,<sup>14</sup> = 54.53, p < 0.05; **Table 1**), similar to what was observed with ACTH. Additionally, exposure to the higher AHI = 10 resulted in further suppression of corticosterone than AHI 8 did (F1,<sup>12</sup> = 4.90, p < 0.05; **Table 1**). Unlike what was observed with ACTH, there were no significant differences in corticosterone levels in the Sprague Dawley single-housed rats following AHI 10 (F1,<sup>8</sup> = 0.377, p < 0.05; **Table 1**). Additionally, an association between oxidative stress and nadir corticosterone was not present in either strain exposed to AHI 10 (data not shown). This indicates a strain difference in sensitivity to hypoxia that is dependent on social housing environment and frequency of air exchanges.

### Body Weight Changes due to Strain and Housing

Activation of the HPA system can influence body weight (Faraday et al., 2005; Bhatnagar et al., 2006; Page et al., 2016). In our study, a statistically significant interaction between strain and housing conditions on final weight was observed at an AHI 8 (F1,<sup>93</sup> = 7.597; p < 0.05). Analysis of the main effects revealed a significant difference in weight between the two strains of rats (F1,<sup>93</sup> = 41.746, p < 0.05), with Long-Evans rats weighing more than Sprague Dawley rats (**Figure 5A**). This strain difference was maintained between Sprague Dawley and Long-Evans rats when they were exposed to an AHI 10 (F1,<sup>17</sup> = 55.673, p < 0.05; **Figure 5B**). A significant difference in housing (F1,<sup>93</sup> = 5.075, p < 0.05) was observed in Long-Evans rats in which pair-housed Long-Evans rats weighed less than the single-housed Long-Evans rats (**Figure 5A**). The housing conditions of Sprague Dawley rats did not affect weight in this study. Additionally, exposure to CIH at either AHI 8 or 10 did not affect the final weight of any of the treatment groups or correlate with oxidative stress measurements (**Table 2**).

### DISCUSSION

Our current experiment utilized a protocol modeling an AHI 8 or 10 to examine the effects of 7-day CIH treatment on two different strains of rats, Sprague Dawley and Long-Evans, housed either singly or in pairs. Prior studies have indicated substantial physiological changes occur at this early time point which contributes to the overall detrimental effects of CIH. For example, there is an elevation of mean arterial pressure

and heart rate (Knight et al., 2011), activation of the HPA axis (Ma et al., 2008), and an elevation of oxidative stress and inflammation in both plasma and brain regions susceptible to neurodegenerative processes (Snyder et al., 2017). Behavioral studies have provided evidence that strain differences and social interaction via housing conditions can affect the outcome of many studies (Bielajew et al., 2002; Faraday, 2002; Faraday et al., 2005; Tan et al., 2009; Konkle et al., 2010; Turner and Burne, 2014). Consideration of these factors allows for a more robust understanding of the mechanisms of disease and improves therapeutic outcomes. To investigate how strain and housing conditions can affect reported outcomes of CIH, we elected to quantify markers of HPA activation, plasma oxidative stress, and hyperphosphorylated tau (p-tau), an indicator of neurodegenerative processes.

Interestingly, only one study has examined the differences between Sprague Dawley and Long-Evans rats in response to oxygen exposure. Unlike our study using intermittent low oxygen levels, Chrysostomou et al. (2010) examined hyperoxia (75% oxygen) exposure for 14 days and found Long-Evans rats were more sensitive to oxygen than Sprague Dawley rats, resulting in increased cell death and astrocyte upregulation. Consistent with these observations, significant differences in oxidative stress and tau tangles in a strain dependent manner were observed in this study (**Figures 1**–**3**). Pair-housed Long-Evans rats exhibited an increase in oxidative stress similar to what was previously observed in the single-housed Sprague Dawley rats at a slightly higher AHI (Snyder et al., 2017). The single-housed Sprague Dawley rats used in this study did not show significant increase in oxidative stress under CIH conditions at an AHI 8, but did exhibit elevated oxidative stress at the higher hypoxic frequency of AHI 10, confirming our previous studies at AHI 10 (Snyder et al., 2017). Thus, Long-Evans rats in group housing conditions may be more sensitive to oxidative insults than Sprague Dawley rats. Interestingly, neither the pair-housed Sprague Dawley nor the single-housed Long-Evans rats exhibited an oxidative stress response to early CIH at AHI 8. These results suggest an interaction between genetic differences in the rat strains and housing conditions influences oxidative stress activation. These parameters should be considered when investigating mechanisms contributing to oxidative stress.

In our previous study (Snyder et al., 2017), 7 days CIH at an AHI 10 elevated plasma oxidative stress and oxidative stress within brain regions susceptible to early neurodegenerative mechanisms. Evidence exists in the literature supporting the involvement of elevated oxidative stress on p-tau generation (Terni et al., 2010; Schulz et al., 2012). To investigate how elevated plasma oxidative stress due to CIH might reflect neuropathological changes, p-tau was quantified in the ETC and subregions of the hippocampus. Although we saw no effects on p-tau at AHI 8, p-tau was found to be significantly elevated in the ETC and CA1 subregion of the hippocampus in both pair-housed Long-Evans rats and single-housed Sprague Dawley rats exposed to CIH at AHI 10. Additionally, pairhoused Long-Evans rats were observed to express p-tau in the dentate gyrus subregion of the hippocampus, whereas singlehoused Sprague Dawley rats did not. This is consistent with data supporting Long-Evans rats may be more susceptible to changes in oxygen than Sprague Dawley rats. Therefore, the p-tau observed in our study may be associated with the oxidative stress response to CIH. Of note, the increase in p-tau

protein expression at AHI 10 coincided with significant HPA dysregulation, suggesting the HPA axis may also be involved in tau tangles. Indeed, the study by Carroll et al. (2011), demonstrates that acute administration of corticosterone reduced p-tau, whereas chronic stress for 1 month resulted in elevated p-tau. Our results are consistent with changes due to chronic HPA activation.

In both humans and rodents, social interaction and environmental stress impact disease risk (Saxton et al., 2011). Activation of the HPA axis occurs under stressful scenarios.



Under acute stress conditions, elevated peak corticosterone or ACTH is indicative of an HPA activation (Bhatnagar et al., 2006). However, under chronic stress, decreased HPA hormones have been observed (Bielajew et al., 2002; Rich and Romero, 2005). Maintaining homeostatic HPA hormones is necessary in maintaining sleep architecture (Karatsoreos et al., 2011; Machado et al., 2013). Corticosterone and ACTH fluctuate in diurnal patterns that mirror each other, with low concentrations occurring at the beginning of the sleep phase (Chacon et al., 2005; Caride et al., 2010). Although our samples were collected during the nadir phase, the further suppression of HPA hormones as the AHI increased indicates CIH was a chronic stressor resulting in dysregulation of the HPA axis. The floor effect of corticosterone and elevated oxidative stress at AHI 10, highlights the potential sensitivity of Long-Evans rats to intermittent hypoxia. Similar to our results, prior studies found basal ACTH was not altered in single-housed Sprague Dawley rats exposed to 7 days CIH (AHI 10), but was more reactive when those same rats were subjected to a subsequent stressor (30 min immobilization) (Ma et al., 2008). Sprague Dawley rats housed individually consistently present significantly greater HPA responses than Sprague Dawley rats in group housing, suggesting socialization desensitizes Sprague Dawley rats to stress (Sharp et al., 2002).

The observed strain difference in ACTH, in which Long-Evans rats have higher ACTH than Sprague Dawley rats, suggests our measurements of the HPA axis are in agreement with existing publications (Sarrieau et al., 1998; Faraday et al., 2005; Ma et al., 2008). Although previous publications reported a difference in corticosterone between Sprague Dawley and Long-Evans rats, there were no basal differences in corticosterone due to strain or housing in this study. We did observe differences in nadir ACTH and corticosterone at AHI 8 and 10 between strains that are consistent with an HPA axis response to chronic stress (**Table 1**). The suppression in ACTH and corticosterone observed in the single-housed Sprague Dawley rats and the pair-housed Long-Evans rats may be indicative of HPA activation resulting in either elevated peak hormone levels or dysregulation of the diurnal cycle. Future studies which collect samples at peak times (at the beginning of the active phase) will help determine which scenario is accurate. Regardless, nadir corticosterone

levels were positively correlated with oxidative stress only in the Long-Evans pair-housed rats at AHI 8, which was the only group to show CIH-induced oxidative stress at this AHI. These results suggest a relationship between activation of the HPA axis and oxidative stress generation. However, this relationship is no longer evident upon more intense AHI. Therefore, it possible that an AHI 8 induces an incomplete suppression of the HPA axis, suggesting AHI 8 is a milder chronic stressor than AHI 10.

The 2011 recommendation by the National Research Council for the care and use of laboratory animals is that social animals, such as rats, are to be housed in pairs or as a group (National Research Council (US) Committee for the Update of the Guide for the Care and Use of Laboratory Animals, 2011). This recommendation was based on prior studies using Sprague Dawley and Wistar rat strains that found decreased stress responses in rats housed in groups of three to four male rats/cage (Klir et al., 1984; Sharp et al., 2002). Similarly, our results show that housing conditions did not adversely affect Sprague Dawley male rats, under normoxic conditions. No differences in HPA hormones or final body weights were found in either individually or pair-housed rats, consistent with prior reports in Sprague Dawley male rats (Turner and Burne, 2014) (**Figures 4**, **5**). However, not all rat strains respond the same to housing conditions. Group housing in Long-Evans rats has been associated with increased anxiety and reduced body weight (Scheufele et al., 2000; Bean et al., 2008). Similarly, our results showed male Long-Evans rats were adversely affected by pair-housing, as evidenced by decreased final body weights, regardless of hypoxic exposure (**Figure 5**). It has been proposed that this strain difference could be due to the level of aggression displayed by the different rat strains. For example, the Sprague Dawley rat strain is a non-aggressive rat strain, whereas the Long-Evans rat strain is more aggressive (Adams and Blizard, 1987; Henry et al., 1993). Henry (1992) and Henry et al. (1993) found that group housing of unfamiliar adult Long-Evans rats, and not Sprague Dawley rats, resulted in a prolonged activation of the stress-response and the inability to establish a stable dominance hierarchy. Indeed, we observed increased aggressive behaviors [attacks, threats, aggressive mounts, boxing, and dominance postures (Cunningham and McGinnis, 2006)] by Long-Evans males and not in Sprague Dawley males (data not shown). Therefore, stress, as evidenced by body weight loss, may underlie the observed differences in Sprague Dawley and Long-Evans rat strains to CIH.

These results suggest that mechanisms which render an organism susceptible to oxidative stress insults may also impair the HPA axis. Elevated oxidative stress as well as HPA dysregulation may, in turn, contribute to mechanisms which initiate neurodegenerative processes, such as tau tangles (a key neuropathology characteristic of Alzheimer's disease) (**Figure 6**). The results may not be immediately observable under non-stressful conditions, but manifest with the addition of a psychological or physiological stressor. A similar phenomenon is observed in clinical populations who experience chronic life stressors or illness and are subsequently exposed to an additional injury or infection (Rao, 2009; Seib et al., 2014; Duric et al., 2016; Jacob et al., 2018). They often succumb more rapidly and have lingering health concerns compared to individuals with less stress-response activation. Therefore, sleep apnea mechanisms may be additive and pose the highest risk to individuals with additional physiological or psychological stress.

Differences in strain response to CIH were observed in oxidative stress and corticosterone/ACTH measurements under different housing conditions and at different AHI levels. Although the differences between AHI 8 and AHI 10 were small, they were significant between strains and housing conditions. These results underscore the need for housing conditions to be included with strain reporting, especially in the investigations of any stressful stimuli, such as CIH. Factors which affect the HPA axis may influence the outcome of these types of studies. This study may shed light on discrepancies found between labs that use different animal strains and housing conditions, as well as guide future experimental design choices when selecting an animal model.

### AUTHOR CONTRIBUTIONS

BS wrote the paper, collected the data, and performed the analysis. PD edited the methods, collected the data, and performed the analysis. MT and EW collected the data and

### REFERENCES


performed the analysis. RC conceived and designed the analysis, edited the paper, and is the primary investigator.

### FUNDING

This work was supported by the NIH under grant R01 NS091359 and The Alzheimer's Association New Investigator Research Grant NIRG-14-321722 to RC and by a NIH training grant T32 AG 020494 to BS.

### ACKNOWLEDGMENTS

We would like to thank Drs. J. Thomas Cunningham, Steve Mifflin, Ann Schreihofer, and Robert Luedtke for their advice and provision of experimental equipment. In addition, we would like to acknowledge the technical assistance of Joel Little and Jessica Proulx.


physiologic responses of Sprague-Dawley and Long Evans rats. J. Am. Assoc. Lab. Anim. Sci. 49, 427–436.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Snyder, Duong, Tenkorang, Wilson and Cunningham. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Imbalance in Renal Vasoactive Enzymes Induced by Mild Hypoxia: Angiotensin-Converting Enzyme Increases While Neutral Endopeptidase Decreases

Carlos P. Vio1,2 \*, Daniela Salas<sup>1</sup> , Carlos Cespedes<sup>1</sup> , Jessica Diaz-Elizondo<sup>1</sup> , Natalia Mendez1,3, Julio Alcayaga<sup>4</sup> and Rodrigo Iturriaga<sup>5</sup>

### Edited by:

Ovidiu Constantin Baltatu, Anhembi Morumbi University, Brazil

### Reviewed by:

Zaid A. Abassi, Technion – Israel Institute of Technology, Israel Manuel Ramírez-Sánchez, Universidad de Jaén, Spain Beth J. Allison, Hudson Institute of Medical Research, Australia

> \*Correspondence: Carlos P. Vio cvio@uc.cl

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 28 June 2018 Accepted: 28 November 2018 Published: 11 December 2018

### Citation:

Vio CP, Salas D, Cespedes C, Diaz-Elizondo J, Mendez N, Alcayaga J and Iturriaga R (2018) Imbalance in Renal Vasoactive Enzymes Induced by Mild Hypoxia: Angiotensin-Converting Enzyme Increases While Neutral Endopeptidase Decreases. Front. Physiol. 9:1791. doi: 10.3389/fphys.2018.01791 <sup>1</sup> Department of Physiology, Center for Aging and Regeneration CARE UC, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Santiago, Chile, <sup>2</sup> Facultad de Medicina y Ciencia, Universidad San Sebastián, Santiago, Chile, <sup>3</sup> Institute of Anatomy, Histology, and Pathology, Facultad de Medicina, Universidad Austral de Chile, Valdivia, Chile, <sup>4</sup> Laboratorio de Fisiología Celular, Departamento de Biología, Facultad de Ciencias, Universidad de Chile, Santiago, Chile, <sup>5</sup> Laboratorio de Neurobiología, Department of Physiology, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Santiago, Chile

Chronic hypoxia has been postulated as one of the mechanisms involved in saltsensitive hypertension and chronic kidney disease (CKD). Kidneys have a critical role in the regulation of arterial blood pressure through vasoactive systems, such as the renin-angiotensin and the kallikrein–kinin systems, with the angiotensin-converting enzyme (ACE) and kallikrein being two of the main enzymes that produce angiotensin II and bradykinin, respectively. Neutral endopeptidase 24.11 or neprilysin is another enzyme that among its functions degrade vasoactive peptides including angiotensin II and bradykinin, and generate angiotensin 1–7. On the other hand, the kidneys are vulnerable to hypoxic injury due to the active electrolyte transportation that requires a high oxygen consumption; however, the oxygen supply is limited in the medullary regions for anatomical reasons. With the hypothesis that the chronic reduction of oxygen under normobaric conditions would impact renal vasoactive enzyme components and, therefore; alter the normal balance of the vasoactive systems, we exposed male Sprague-Dawley rats to normobaric hypoxia (10% O2) for 2 weeks. We then processed renal tissue to identify the expression and distribution of kallikrein, ACE and neutral endopeptidase 24.11 as well as markers of kidney damage. We found that chronic hypoxia produced focal damage in the kidney, mainly in the cortico-medullary region, and increased the expression of osteopontin. Moreover, we observed an increase of ACE protein in the brush border of proximal tubules at the outer medullary region, with increased mRNA levels. Kallikrein abundance did not change significantly with hypoxia, but a tendency toward reduction was observed at protein and mRNA levels. Neutral endopeptidase 24.11 was localized in proximal tubules, and was abundantly expressed under normoxic conditions, which markedly decreased both at protein and mRNA levels

**350**

with chronic hypoxia. Taken together, our results suggest that chronic hypoxia produces focal kidney damage along with an imbalance of key components of the renal vasoactive system, which could be the initial steps for a long-term contribution to salt-sensitive hypertension and CKD.

Keywords: renal hypoxia, kallikrein, angiotensin-converting enzyme, neprilysin, neutral endopeptidase, subtle renal injury

## INTRODUCTION

fphys-09-01791 December 7, 2018 Time: 16:23 # 2

Hypertension and chronic kidney disease (CKD) are major public health problems worldwide, with chronic hypoxia being one of the suggested mechanisms involved in the genesis of salt-sensitive hypertension and to the progression of CKD (Suga et al., 2001; Nangaku and Fujita, 2008). Several factors, such as angiotensin II, cyclosporine A, phenylephrine, and hypokalemia have been postulated to induce local vasoconstriction and renal hypoxia, which subsequently may lead to CKD (Johnson et al., 2002; Vio and Jeanneret, 2003).

Renal function and arterial blood pressure regulation are under physiological control by renal vasoactive substances, through the activation of different systems such as the following: nitric oxide synthases, the endothelin system, the reninangiotensin system (RAS), the kallikrein–kinin system (KKS), and renal eicosanoids produced by cyclooxygenases 1 and 2. Among nitric oxide synthases, neuronal nitric oxide synthase (nNOS, NOS-1) plays an important role in the kidneys by regulating renal hemodynamics in the immature kidney during pre and postnatal stages (Rodebaugh et al., 2012). Additionally, NOS-1 and NOS-3 are upregulated by acute hypoxia (Gess et al., 1997). On the other hand, endothelin-1, which is a multifunctional peptide, has potent vasoconstrictor and profibrotic effects on the systemic vasculature and kidneys. The endothelin system seems to play a fundamental role in diabetes, proteinuric renal disease, hypertension, and renovascular disease (Hargrove et al., 2000; Kelsen et al., 2011; Meyers and Sethna, 2013). Our laboratory mainly studied the following two systems: the RAS, the KKS, and renal eicosanoids, which are related to both systems. We have focused on enzymes in both systems, since they are currently known as important targets of inhibitors in antihypertensive treatment, and are used in CKD. The RAS leads to vasoconstriction and sodium retention, and its main active peptide is angiotensin II. The latter is produced by the angiotensin-converting enzyme (ACE), and is metabolized by aminopeptidase A to angiotensin III (Mizutani et al., 2008) and by neutral endopeptidase 24.11 (NEP) or neprilysin to inactive metabolites (Rice et al., 2004; Campbell, 2017). The KKS produces vasorelaxation and sodium excretion through bradykinin produced by kallikrein, which is the key enzyme of the system. The renal KKS participates in renal and extrarenal events such as regulation of blood pressure and control of sodium and water excretion. Kallikrein originates in the connecting tubule (CNT) and generates bradykinin, which is the effector hormone in the kidney regulating sodium excretion and glomerular hemodynamics, among other effects (Jaffa et al., 1992; Vio et al., 1992; Madeddu et al., 2007; Zhang et al., 2018). Bradykinin levels are regulated by ACE and NEP. There is another family of natriuretic peptides that participates in the regulation of arterial blood pressure, with atrial natriuretic peptide being the main one that involves renal action (Chen, 2005). Furthermore, the availability of vasoactive hormones is regulated by peptidases that degrade them. In kidneys, NEP has the critical function of degrading bradykinin, angiotensin II and natriuretic peptides, and transforming angiotensin I into angiotensin 1–7 (Rice et al., 2004; Judge et al., 2015; Campbell, 2017).

Renal vasoactive systems are involved in a delicate balance of opposite effects that lead to the production of either vasoconstrictor or vasodilator hormones, which activate sodiumretaining or excretory mechanisms, and have profibrotic or antifibrotic consequences (Bader and Ganten, 2008). For example, ACE and NEP have opposite actions in the regulation of vasoactive peptides: ACE transforms angiotensin I into angiotensin II (with hypertensive effects) and NEP transforms angiotensin I into angiotensin 1–7 (with antihypertensive effects). Angiotensin 1–7 can be further degraded by ACE into angiotensin 1–5 (Rice et al., 2004; Campbell, 2017).

Alterations of vasoactive enzymes have been described during the progression of CKD and hypertension. In fact, most drugs used to treat hypertension have been designed to inhibit members of the RAS family. Furthermore, new studies suggest that the use of the angiotensin II receptor antagonist in addition to the NEP inhibitor is a promising strategy to treat heart failure associated to hypertension (Volpe et al., 2016).

It is known that the kidney is very sensitive to changes in oxygen supply. Partial pressure of oxygen (PO2) is carefully balanced between the cortex and the outer medulla, where under normoxia, the PO<sup>2</sup> gradient within the kidney has been found to reach about 70 mmHg in the cortex and outer medulla, and up to 10 mmHg in the inner medulla and papilla (Leichtweiss et al., 1969; Baumgärtl et al., 1972; Günther et al., 1974; Lübbers and Baumgärtl, 1997; O'Neill et al., 2015). Adequate kidney oxygenation is crucial to fuel active transportation processes of electrolytes and water in the nephron. Although kidneys receive a very high blood flow, oxygen extraction is relatively low. Consequently, kidneys are particularly susceptible to hypoxic injury because small changes in flow, which can be generated by vasoconstriction for example, are translated into local hypoxia (Epstein, 1997). This can then generate focal lesions, and lead to manifestations of kidney damage when the condition becomes chronic.

Based on the current knowledge exposed above, the present study was guided by the hypothesis that chronic reduction of oxygen, under normobaric conditions, affects renal vasoactive enzyme components, altering the normal balance of vasoactive systems, and favoring the vasoconstrictor profibrotic RAS.

# MATERIALS AND METHODS

fphys-09-01791 December 7, 2018 Time: 16:23 # 3

### Animals and Experimental Procedures

Experiments were performed in 12 adult male Sprague-Dawley rats (180–200 g, n = 6 for each group, normoxia and hypoxia). This study was carried out in accordance with recommendations in the "Manual de Normas de Bioseguridad" (Biosafety Norms Manual, 2nd Ed., 2008, FONDECYT-CONICYT). The experimental protocol for animals was approved by the Bio-Ethics Committee of the School of Science at Universidad de Chile. Animals were housed in a 12 h light/dark cycle with free

TABLE 1 | Body weight and hematocrit values in normoxic and hypoxic animals.


<sup>∗</sup>P < 0.001.

access to food (Prolab RMH 3000, Purina LabDiet) and water, and were randomly assigned to either a control group (normoxia) or to chronic normobaric hypoxia for 2 weeks.

## Chronic Normobaric Hypoxia Exposure

Animals (n = 12) were exposed to normoxia (n = 6), serving as controls for normobaric hypoxia (n = 6) for 2 weeks as previously described (Icekson et al., 2013). Briefly, three rats were allocated to a cage (D × W × H; 48 cm × 26 cm × 15 cm) and two cages were placed in a 300 L (60 cm × 50 cm × 100 cm) acrylic chamber with a hermetic lid. The oxygen content (FiO2) in the chamber was continuously monitored through an oxygen sensor (AX300, Teledyne Analytical Instruments, CA, United States), whose output was fed to an automatic programmable controller (Zelio SR2B121BD, Schneider Electric, France) that controlled two solenoid valves (2026BV172, Jefferson Solenoid Valves, FL, United States). A relief output valve opened simultaneously with the admission valves, and remained opened for 40 s after the closure of the latter ones, while two mechanic relief valves opened whenever the pressure inside the chamber exceeded approximately 17 mmHg. Thus, the mean pressure in the chamber was slightly larger (1P = 1.3 ± 0.1 mmHg) than the atmospheric one (717.2 ± 0.3 mmHg). The atmosphere within

FIGURE 1 | Conventional staining with H-E in renal tissue from normoxic and hypoxic animals. Representative images of kidneys from animals in control (A,B) and hypoxic (C,D) conditions. (A,B) Normoxic tissue has tubules and tubulointerstitial space of normal aspect, both in the medullary zone and in the cortex. Blood vessels present usual appearance. (C,D) In hypoxic kidneys, it is possible to observe tubular dilation in some areas of the field, affecting certain blood vessels that irrigate such areas. Dotted areas correspond to higher magnification images on the right column. G = Glomerulus. Scale bar = 100 µm.

the chamber was continuously homogenized by four internal fans. After closure, chamber air was purged for approximately 5 min with pure N2, until attaining approximately 9.5% of FiO2. From then on, the system automatically regulated FiO<sup>2</sup> levels in the chamber by admitting N<sup>2</sup> or compressed air into the chamber if FiO<sup>2</sup> values were over 9.8% or below 8.7%, respectively. Mean FiO<sup>2</sup> within the chamber was 9.33 ± 0.12% (mean ± SD). CO<sup>2</sup> produced by animal ventilation was trapped by using CaCO<sup>3</sup> (250 g), and urinary NH<sup>3</sup> with H3BO<sup>3</sup> (60 g). The chamber remained open for about 5 min every other day to clean the internal cages and replenish the water containers, whereas CO<sup>2</sup> and NH<sup>3</sup> traps were changed every four days. The cages were rotated within the chamber to maximize homogeneity. The pressure inside the chamber was continuously measured with a gauge transducer (Statham P20), and the FiO<sup>2</sup> signal was recorded at 1 Hz with a computerized analog to digital acquisition system (DI-158U, DATAQ Instruments Inc., OH, United States). The temperature in the chamber was recorded at 5 min intervals throughout the 14 days of hypoxia with a data logger (EL-USB-2, Lascar Electronics Inc., PA, United States).

At the end of the treatment, rats were deeply anesthetized with isoflurane (isoflurane in O2, induction 4–5%) and euthanized by exsanguination. Blood was obtained from Vena Cava in heparintubes, and both kidneys were removed. Then, isoflurane was increased until breathing stopped and death of the animals was confirmed. Pneumothorax was performed as a secondary physical method of euthanasia.

Kidneys were rapidly decapsulated and approximately 100 mg slices were cut with a transverse cross section in the middle of the kidney (discarding both poles) to obtain samples for quantitative polymerase chain reaction (qRT-PCR). Samples for qRT-PCR were stored at –80◦C until they were processed.

The hematocrit was determined by microcentrifugation using two uncoated hematocrit glass tubes (length 75 mm) per sample, filled and centrifuged for 5 min in a micro-hematocrit centrifuge (Hermle Z200). The hematocrit was established using a microhematocrit tube reader, and the mean value was calculated.

### Source of Antisera and Chemicals

ACE antiserum (SC-12187) was purchased from Santa Cruz Biotechnology, United States. Osteopontin (OPN) monoclonal antibody (MPIIIB10) was obtained from Developmental Studies Hybridoma Bank, IA, United States, and NEP antiserum (AB-5458) from EMD Millipore. Kallikrein antibody was used

FIGURE 2 | Conventional staining with PAS in renal tissue from normoxic and hypoxic animals. (A,C) show a general view of cortical and medullary zones for each group. (A,B) correspond to the normoxia group. The color fuchsia is present mainly in the basal membrane and the brush border of proximal tubules. No signs of injury are observed in this group. (C,D) correspond to the hypoxia group, showing an area that presents focal damage. Only a few tubules present dilation, while the area that surrounds it presents normal appearance. Dotted areas indicate higher magnification images presented in the right column. G = Glomerulus. Scale bar = 100 µm.

as reported previously (Salas et al., 2007). Random primer dNTPs and FAST SYBR Green Master Mix were acquired from Thermo Fisher Scientific. Secondary antibody and corresponding peroxidase-anti-peroxidase (PAP) complex were purchased from MP Biomedicals, Inc., Germany.

# Tissue Processing and Immunohistochemical Analysis

fphys-09-01791 December 7, 2018 Time: 16:23 # 5

Renal tissue samples (3 mm thick) were fixed by immersion in Bouin's solution for 24 h at room temperature. Tissue was then dehydrated, embedded in Paraplast Plus, serially sectioned at 5 µm thick with a Leica rotatory microtome, mounted on glass slides, and stored.

Conventional staining of tissue sections was performed with Hematoxylin-Eosin (H-E), Periodic Acid-Schiff (PAS) for histological analysis, and Picrosirius red for collagen staining (Villanueva et al., 2006, 2014).

Immunostaining was carried out using an indirect immunoperoxidase technique to localize kallikrein, ACE, OPN, and NEP in rat kidneys. Briefly, tissue sections were dewaxed, rehydrated, rinsed in 0.05 M Tris-phosphate-saline (TPS) buffer, pH 7.6, and incubated overnight at 22◦C with a primary antiserum raised against kallikrein, ACE, OPN, or NEP. The secondary antibody and corresponding PAP complex were applied for 30 min each at 22◦C. The immunoperoxidase reaction was visualized after incubation of sections in 0.1% (wt/vol) diaminobenzidine and 0.03% hydrogen peroxide. Sections were rinsed with TPS buffer between incubations, counterstained with hematoxylin, dehydrated, cleared with xylene and then coverslipped. Controls for the immunostaining procedure were prepared by omission of the first antibody. Images were examined with conventional light microscopy and acquired using a Nikon Eclipse E600 microscope and Nikon DS-Ri1 digital camera.

### Quantitative RT-PCR

Total RNA was extracted from whole kidney tissue using TRIzol, according to the manufacturer's instructions. RNA integrity was determined by 1% agarose gel electrophoresis and its concentration, by absorbance at 260/280 nm using a 2.5 µg aliquot of total RNA to cDNA synthesis. Samples were treated with DNAse I using MMLV, dNTPs and random primers to obtain cDNA. The housekeeping gene used was glyceraldehyde-3-phosphate dehydrogenase (GAPDH)

FIGURE 3 | Osteopontin in renal tissue from normoxic and hypoxic animals. The arrows indicate the localization of OPN that appears stained in brown. (A,B) correspond to the normoxia group. In (A) it is possible to observe a little OPN stain that indicates low expression levels of this protein. Brown stains are totally absent in (B), and the same happens in others fields of the same renal section, representing normal conditions. (C,D) correspond to the hypoxia group. It is possible to observe an increase of OPN staining, expressed mainly in tubular cells. The left panel shows a cortical area where a glomerulus is observed (G). The right panel corresponds to the outer stripe of the outer medulla. All images have the same magnification. Scale bar = 100 µm.

(forward primer: 5<sup>0</sup> -CACGGCAAGTTCAACGGC-3<sup>0</sup> , reverse primer 5<sup>0</sup> -GGTGGTGAAGACGCCAGTA-3<sup>0</sup> ). The primer sequences used were the following: kallikrein, forward primer: 5 0 -GCATCACACCTGACGGATTG-3<sup>0</sup> , reverse primer: 5<sup>0</sup> - GGCCTCCTGAGTCACCCTTG -3<sup>0</sup> ; ACE, forward primer: 5 0 -AACACGGCTCGTGCAGAAG-3<sup>0</sup> , reverse primer: 5<sup>0</sup> - CCTGCTGTGGTTCCAGGTACA-3<sup>0</sup> ; and NEP, forward primer: 5 0 -TCAGCCTTTCTGTGCTCGTC-3<sup>0</sup> , reverse primer: 5<sup>0</sup> - ATTGCGTTTCAACCAGCCTC-3<sup>0</sup> . Quantitative PCR was performed in duplicate in a StepOnePlus Real-Time PCR System (Applied Biosystems) using FAST SYBR Green Master Mix for amplification. Results were normalized by GAPDH. Mathematical quantification was made using the 2−11CT method (Livak and Schmittgen, 2001).

## Statistics

All data are presented as mean ± SEM. Statistical analyses were performed using an unpaired Student's t-test and GraphPad Prism software (version 5.0c for Windows, GraphPad Software, CA, United States). Differences with P < 0.05 were considered statistically significant.

# RESULTS

# General Features and Morphological Traits

An increase in hematocrit and a decrease in body weight were observed in hypoxic animals compared to controls, over the 2-week period of hypoxia, which is characteristic of this experimental condition (**Table 1**).

For the morphological study we selected conventional stains with different properties to address tissue structure at the light microscopy level. They were H-E, PAS and Picrosirius Red. They provide information about general structure (H-E), more detailed information of structural shapes of tubules, glomeruli and vessels, provided by the glycoprotein and basement membrane staining (PAS), and collagen staining as an index of tissue damage (Picrosirius red).

The immunohistochemistry technique was used to evaluate de novo expression of tubular OPN, which is a key macrophage chemokine that is not normally expressed in adult kidneys, and is induced by tubular damage of different origins (Lombardi et al., 2001).

FIGURE 4 | Collagen staining with Picrosirius red stain in renal tissue from normoxic and hypoxic animals. (A,B) correspond to the normoxia group. (C,D) correspond to the hypoxia group. The left panel shows a panoramic view of the cortical and medullary zones of both groups. The dotted area indicates images with higher magnification on the right column. The arrows indicate the location of collagen stained in red. Note that peritubular collagen is scarce in all images. However, in the hypoxia group it is possible to observe small areas of fibrosis, characterized by an increase in local collagen. In (A,B) an arterial blood vessel is observed with perivascular collagen, which does not constitute fibrosis. All structures other than collagen are stained in yellow. G corresponds to the glomerulus. Scale bar = 100 µm.

Also, immunohistochemistry was used to characterize the protein expression at tissue level and the distribution in the kidney of the vasoactive enzymes kallikrein, ACE, and NEP.

Renal tissue specimens were examined in coded samples by two of the researchers (C.Cespedes and C.Vio), and systematic analysis of kidney tissue samples was done focusing on the overall morphological aspect, and on cortical and medullary tubules, glomeruli, blood vessels, interstitial space, and intratubular spaces.

An overall and detailed examination of renal tissue with conventional PAS and H-E staining of kidneys from normoxic animals showed no signs of pathological alterations in the cortex or the medulla. No vascular changes were observed, tubules were shaped normally in terms of diameter and cell size, and no signs of cell infiltration and inflammation were observed in the tubulointerstitial space. Glomeruli from cortical or juxtamedullary nephrons had normal aspect (**Figures 1A,B**, **2A,B**).

A panoramic examination of kidney tissue from hypoxic animals over the cortex and medulla revealed a mild and focalized morphological alteration, characterized by focal lesions in the outer medullary and inner cortical zone, as evidenced with H-E staining (**Figure 1**). More detailed information was obtained with PAS staining, showing the focal alteration at this outer medullary region (**Figure 2**).

Focal areas with signs of alterations consisted of "spotty" areas mainly located in the outer medulla and inner cortical zone, close to juxtamedullary nephrons (**Figures 1**, **2**). They contained dilated tubules with atrophic tubular epithelia, and basophilic aspect. Such atrophic and dilated tubules had a mild cellular infiltration in the corresponding tubulointerstitial space (**Figures 2C,D**).

In control animals, the presence of renal OPN was observed with a very scarce distribution in tubular cells of the cortex, as expected for normal tissue (**Figures 3A,B**). In contrast to normal tissue, kidneys from hypoxic animals showed an increased expression of OPN in cortical tubules with a focal distribution, as shown in **Figures 3C,D**, which corresponds to the typical pattern of de novo expression of OPN in renal injury.

In normal tissue, staining with Picrosirius red revealed typical collagen staining around arteries, and very faint staining in peritubular locations (**Figures 4A,B**). On the other hand, focal staining was observed in tissue from hypoxic animals in discrete areas corresponding to focal areas of local injury, as presented in

FIGURE 5 | Kallikrein immunostaining in renal tissue from normoxic and hypoxic animals. Green arrows indicate the localization of kallikrein (brown immunostaining) in both groups. It is expressed only in the cortical zone, specifically in the apical pole and perinuclear area of connecting tubule cells. (A,B) correspond to the normoxia group. Greater expression of kallikrein stains are observed in this group compared to the hypoxia group (C,D), which presents less staining in the same cell type. The immunostained mark is not present in the collecting duct or another tubule type, in the glomerulus (G), or in blood vessels. All images of the cortical zone have the same magnification. Scale bar = 100 µm.

**Figures 4C,D**, where peritubular signs of fibrosis around dilated and injured tubules were observed.

Thus, the histological evidence presented above is consistent with mild manifestations of lesions, with focal distribution patterns located in the outer medulla and inner cortical zone of the kidneys. No similar lesions were observed in the inner medulla or in the outer cortical zone of normoxic animals.

It is relevant to note that this outer medullary/inner cortical area is the zone where juxtamedullary nephrons reside. Juxtamedullary nephrons are fewer than cortical or superficial nephrons, but are very important for sodium management and blood pressure regulation because they generate medullary circulation.

# Vasoactive Enzyme Expression and Distribution

### Kallikrein

In normal and hypoxic experimental groups, kallikrein is restricted to connecting tubular cells (CNTc) of CNT of distal nephrons. In normal tissue, CNTc display abundant immunoreactive kallikrein in the apical pole and in the perinuclear area, where the Golgi complex resides. CNTc are intermingled with kallikrein negative cells, which correspond to intercalated cells. CNT are always in close anatomical relation with afferent arterioles close to glomeruli (**Figures 5A,B**). Kallikrein distribution in renal tissue from hypoxic animals was similar to control animals in the CNTc. However, decreased immunostaining intensity and stained focal areas were observed, and there was faint staining in focal tubules (**Figures 5C,D**).

Kallikrein gene expression measured by qRT-PCR was not significantly modified (P > 0.05), although it tended to decrease (**Figure 6**).

### Angiotensin-Converting Enzyme

In normal animals, ACE was located exclusively in the S3 segment of proximal tubules, and was more concentrated in the inner cortical zone and outer medulla. The enzyme was present in the brush border of the apical cellular pole. No ACE was observed in other tubular cells or in the tubulointerstitial space (**Figures 7A,B**). In hypoxic animals, the same pattern of localization was observed over the proximal tubules, although with more intensity. Furthermore, strong staining in more disarranged proximal tubules was observed in focal areas, and focal ACE staining on the peritubular side (**Figures 7C,D**).

ACE gene expression, as measured by qRT-PCR, was significantly higher (P < 0.001) in hypoxic animals compared to controls (**Figure 8**).

### Neutral Endopeptidase 24.11 or Neprylisin

This was observed in control animals in the brush border of proximal tubules, and more concentrated in the outer medulla than in the cortex. Heavy staining was observed in control renal tissue (**Figures 9A,B**), whereas less area was stained in hypoxic kidneys compared to controls. Furthermore, the presence of NEP in focal areas was observed in rather dilated and atrophied tubules (**Figures 9C,D**). Moreover, NEP gene expression, as measured by qRT-PCR, significantly decreased (P < 0.001) by almost 50% in hypoxic animals compared to controls (**Figure 10**).

### DISCUSSION

Hypoxia due to renal vasoconstriction has been proposed to contribute to early stages of CKD and salt-sensitive hypertension (Suga et al., 2001; Johnson et al., 2002). The main findings of the present study are that normobaric hypoxia (10% O2) for 2 weeks induced subtle renal injury, with discrete focal lesions, and local changes in components of renal vasoactive systems, such as increases in kallikrein and ACE, and less NEP expression.

Normobaric hypoxia may be induced by a variety of conditions that reduce the supply of oxygen to organs, such as chronic obstructive pulmonary disease (Kent et al., 2011), obstructive sleep apnea syndrome (Chiang, 2006), CKD (Shoji et al., 2014), and certain conditions that induce local renal hypoxia like cyclosporine A, hypokalemia, angiotensin II or contrast media (Lombardi et al., 2001; Suga et al., 2001; Zhou and Duan, 2014; Fähling et al., 2017). Normobaric hypoxia studies have an advantage over experiments using hypobaric conditions in that hypoxia effects are not mixed with those of low pressure. In fact, arterial blood pressure has been reported to remain unchanged in chronic hypobaric hypoxia (Rabinovitch et al., 1979; Hsieh et al., 2015), but to increase in chronic normobaric hypoxia (Schwenke et al., 2008; Moraes et al., 2014; Flor et al., 2018). The latter highlights the importance of choosing the right type of hypoxia depending on the pathological conditions to be studied, since both conditions may evoke distinct physiological responses (Savourey et al., 2003; Richard and Koehle, 2012). Hence, we used the normobaric hypoxia model in order to mimic the low oxygen supply that kidneys face under pathological conditions, and to distinguish the effects of hypoxia from hypobaria.

# Morphological Alterations

fphys-09-01791 December 7, 2018 Time: 16:23 # 9

Hypoxia is a key pathogenic event that can activate a vicious cycle of destructive processes. In our model, we observed a discrete focal lesion of tubular damage with conventional H-E, and Picrosirius red staining without glomerular damage, indicating the mild and patchy nature of the lesion. The expression of tubular OPN, which is a key macrophage chemokine, without infiltrative cells was somehow unexpected. However, it suggests that longer periods or more severe hypoxia are required to induce infiltration. When compared to other models of mild damage (angiotensin II, hypokalemia, phenylephrine and cyclosporine A) (Johnson et al., 2002), the present model displays a similar damage pattern, but with less intensity. A possible explanation for the low intensity of the damage is that this degree of hypoxia is counteracted in kidneys by regulatory mechanisms that supply oxygen to kidneys as the increased hematocrit (**Table 1**). Similar results have been shown in proximal and distal tubular cells in renal ischemia, tubulointerstitial nephritis, glomerulonephritis, and acute hypoxia (Hampel et al., 2003; Mazzali et al., 2003). These effects are in agreement with upregulation of inflammatory and profibrotic genes in response to hypoxia (Fu et al., 2016; Ow et al., 2018). Fibrosis and subsequent tubular damage exacerbate hypoxia, and induce the activation of genes that favor the expression of vasoconstrictor mediators such as endothelin-1, which could reduce oxygen delivery even more (Hampel et al., 2003). Considering that the outer medulla of the kidney is particularly vulnerable to reduced oxygen supply due to an imbalance between oxygen requirements and medullary blood flow (Fine et al., 2000; Zhou and Duan, 2014), we and others have proposed that a common pathway is perhaps local hypoxia, which creates a vicious circle that initiates and maintains focal renal injury (Vio and Jeanneret, 2003).

A larger inflammatory response was described in rats subjected to chronic hypoxia (81.9 mmHg) (Mazzali et al., 2003). The latter study achieved low PO<sup>2</sup> by reducing barometric pressure to an equivalent altitude of 5550 m, although PO<sup>2</sup> reduction was more severe in the present study (67.1 mmHg). This reduction in barometric pressure, however, may induce

FIGURE 7 | Angiotensin-converting enzyme (ACE) immunostaining in renal tissue from normoxic and hypoxic animals. ACE localization (immunostained in brown) is indicated with green arrows. It is mainly expressed in the medullary zone, at the brush border of the proximal tubule; more specifically in segment S3. (A,B) correspond to the normoxia group. The ACE stain is less in the control group compared to the hypoxia group (C,D). Additionally, in image (D), the immunostained mark appears in the tubulointerstitial space, where there is also a higher number of infiltrative cells. All images correspond to the medullary zone with the same magnification. Scale bar = 100 µm.

physiological responses that are independent from PO<sup>2</sup> reduction (Savourey et al., 2003, 2007; Richard and Koehle, 2012). Moreover, the exposure period was 10 days shorter in the experiments reported here. In chronic intermittent hypoxic experiments, changes in arterial pressure did not develop until 3 weeks of exposure (Del Rio et al., 2010). Thus, differences may arise from the different exposure period and barometric pressure.

### Alterations in Vasoactive Systems

Changes in renal vasoactive components have been proposed to contribute to early stages of CKD and salt-sensitive hypertension. In fact, Johnson et al. (2002) proposed that changes in vasoactive systems constitute a common base for a variety of stimuli (e.g., hypokalemia, angiotensin II, phenylephrine, cyclosporine A) that induce vasoconstriction, leading to ischemia and subsequent subtle renal injury. In such conditions, rats are more sensitive to high-salt induced hypertension, even though arterial blood pressure augmentation is reversed after removing the stimuli. This suggests that renal damage is at the base of this mechanism.

### Decreased Kallikrein Levels

Kallikrein is a serine protease with a key function for kidneys, since it produces bradykinin, which is one of the main vasorelaxing peptides that also inhibits sodium reabsorption through the activation of bradykinin receptor type 2 (Mukai et al., 1996; Sivritas et al., 2008). Our histology and immunohistochemistry data showed a local decrease of kallikrein in CNTc. It is worth noting that kallikrein-containing cells are always in close contact with arterioles underlying the anatomical link between KKS and RAS (Vio et al., 1988). Even though changes in kallikrein gene expression were not significant, there was a tendency to decrease. We have previously reported that hypokalemia, ischemia and phenylephrine (stimuli that produce renal hypoxia and kidney damage) also reduced kallikrein expression, which could contribute to the imbalance of vasoactive enzymes that favor vasoconstriction and sodium retention. The latter may, in turn, lead to the onset of renal injury and hypertension (Martinez et al., 1990; Suga et al., 2001; Johnson et al., 2002; Basile et al., 2005). Further studies are required to determine whether kallikrein reduction is a cause or a consequence of hypoxia. Thongboonkerd et al. (2002) reported a dual role of kallikrein, where episodic hypoxia (which mimics obstructive sleep apnea syndrome) reduced its activity associated with hypertension, while sustained hypoxia increased kallikrein levels. The difference in kallikrein expression observed under the experimental conditions reported here may have been due to the reoxygenation produced between the end of hypoxia and the extraction of renal tissue. This could be explained by the oxidative stress generated during reoxygenation, which may activate different pathways, and could influence kallikrein expression (Milano et al., 2004; Bianciardi et al., 2006). Reduced kallikrein expression favors vasoconstriction, which may worsen hypoxia and lead to kidney damage, since overexpression of human kallikrein in kidneys has been observed to protect rats from hypoxia-induced hypertension (Thongboonkerd et al., 2002).

### Increased ACE Levels

ACE plays a critical role in RAS signaling: It produces angiotensin II (the main vasoconstrictor peptide), which activates pathways leading to hypertension and fibrosis, while also degrading bradykinin. Results from immunohistochemistry in this study showed a local increase of ACE in proximal tubules at the outer medullary region. The observed change in protein expression is consistent with greater gene expression, as measured by qRT-PCR. For many years, much effort has been placed on creating drugs to inhibit ACE, but very little is known about its regulation. It has been reported that phenylephrine, angiotensin II infusion and hypokalemia induce the ACE expression, which has been associated with kidney hypoxia and damage (Johnson et al., 1999; Lombardi et al., 2001; Suga et al., 2001; Vio and Jeanneret, 2003).

It seems that ACE regulation differs depending on the tissue analyzed (Jackson et al., 1986; Oparil et al., 1988; King et al., 1989). For instance, Oparil et al. (1988) reported that hypoxia reduced ACE activity in lungs (associated to angiotensin II reduction), but increased its activity in renal tissue, which was reversed to normal levels when rats were returned to normoxia. Reasons behind organ-specific regulation of ACE are not known, but lungs and kidneys seem to present opposite ACE regulations. Vascular endothelial growth factor (VEGF) is postulated to be the major growth factor modified in response to hypoxia, which increases local production of ACE (Enholm et al., 1997; Saijonmaa et al., 2001), and generates synergy between the RAS and VEGF, hence contributing to angiogenesis and vascular remodeling in response to hypoxia.

nephrons) at the medullary ray and outer strip of the medulla, respectively. (A,B) correspond to the normoxia group. NEP expression is higher in the control group compared to the hypoxia group (C,D). All images correspond to the medullary zone, and they have the same magnification. Scale bar = 100 µm.

NEP gene expression. Arbitrary units were defined as fold change between control versus hypoxic group. Results were averaged and mean values were compiled for statistical analysis. There is a significant decrease in NEP gene expression in the hypoxia group compared to the normoxia group ( ∗∗∗P < 0.001). GAPDH was used as a housekeeping gene.

Increases in ACE may be acting in the following two ways: (1) to increase the amount of angiotensin II that leads to greater vascular resistance and fibrosis, and (2) to decrease the amount of kallikrein by degradation, contributing to an imbalance of vasoactive enzymes and thus to the onset of renal injury observed during hypoxia.

# Decreased NEP Levels

Together with ACE, NEP is the most important kinin-degrading enzyme in the cardiovascular system. Renal NEP is responsible for processing a range of substrates as vasoactive peptides, including bradykinin, endothelins, angiotensin I and angiotensin II, among others (Rice et al., 2004; Judge et al., 2015; Campbell, 2017). In our model, we observed that hypoxia decreased NEP expression both at protein and gene expression levels. Immunohistochemical staining showed that NEP is abundantly expressed in proximal tubules at the outer medullary region under normoxic conditions, but hypoxia reduced its expression. One of the consequences of NEP reduction is a lower generation of angiotensin 1–7, an antihypertensive peptide that counteracts the effects of angiotensin II, which is consistent with a hypertensive phenotype.

There are few studies that report on NEP localization, although the enzyme exhibits vast tissue distribution, with greater

abundance in kidneys and lungs (Li et al., 1995). Previous studies by Carpenter and Stenmark (2001) showed that normobaric hypoxia, although for a much shorter time (72 h), caused a decline in NEP expression at the pulmonary level, which caused increases in vascular permeability. The latter corresponds to a phenotype that is reversed through recombinant NEP administration or a bradykinin receptor antagonist. Downregulation of NEP by hypoxia has been reported in mouse primary cortical and hippocampal neurons as well as in prostate cancer cell lines, where NEP decrease is associated with a loss of beneficial effects mediated by the specific degradation of the substrate (Wang et al., 2011; Nalivaeva et al., 2012; Mitra et al., 2013). It has been shown that the hypoxia inducible factor negatively regulates NEP expression, due to the presence of hypoxia-responsive elements in its promoter (Mitra et al., 2013), which could be responsible for NEP reduction.

Taken together, the changes in the enzymes that regulate the production/degradation of the vasoactive peptides observed in our experimental conditions suggest that hypoxia induces a hypertensive phenotype due to an imbalance of the vasoactive peptides, leading to an increase in the vasoconstrictors and a decrease in vasodilators, such as: decreased bradykinin (antihypertensive peptide) due to decreased kallikrein and increased ACE; increased angiotensin II (hypertensive peptide) due to increased ACE; and reduction of angiotensin 1–7 (antihypertensive peptides) due to decreased NEP.

In summary, we describe local and subtle renal injuries and changes in renal vasoactive components that could damage the kidney, making it more sensitive to high-salt induced hypertension under conditions such as high salt load. Future experiments could involve challenging renal function with high sodium diets to increase oxygen consumption and induce saltsensitive hypertension.

### STUDY LIMITATIONS

Although this study attempted to limit oxygenation time before collecting the samples (2 h maximum) and during feeding/cleaning the animals (5 min every 48 h), changes occurring during reoxygenation (e.g., changes in redox status) could not be ruled out, and may have altered our results.

### REFERENCES


Additionally, renal function could not be measured, since the metabolic cages could not be placed under hypoxic conditions.

Due to space limitations in the hypoxic chambers, the experiments were performed using only male rats. Considering the important differences between males and females in terms of cardiovascular and renal function, it is important to perform these experiments on females in the future to evaluate the possible differences that may exist in the response to hypoxia.

Finally, changes in the expression of key enzymes of the RAS and the KKS were measured, but it cannot be assumed that this will translate into changes in peptide levels. Local levels of angiotensin II and bradykinin will be measured and evaluated in future studies.

### AUTHOR CONTRIBUTIONS

CV, JA, RI, and CC conceived and designed the study, analyzed and interpreted the data, drafted the manuscript, critically revised important intellectual content in the manuscript, and provided overall supervision. DS, CC, JA, and JD-E performed the experiments and analyzed the data and drafted the manuscript, and contributed to intellectual content in the manuscript. NM analyzed and interpreted the data and contributed to intellectual content in the manuscript. All authors approved the final manuscripts and agreed to be accountable for all aspects of the work.

### FUNDING

This work was supported by a grant from CONICYT awarded to the CARE UC Center, grant AFB 170005, FONDECYT grant 11170245 to NM, FONDECYT grant 1090157 to JA, and FONDECYT grant 1150040 to RI.

### ACKNOWLEDGMENTS

The authors are grateful to Maria Alcoholado for technical assistance in tissue sections. DS, JD-E, and NM are postdoctoral fellows from the CARE UC Center and CONICYT.


by serum, growth factors, oncoproteins and hypoxia. Oncogene 14, 2475–2483. doi: 10.1038/sj.onc.1201090


fphys-09-01791 December 7, 2018 Time: 16:23 # 13


Physiol. Regul. Integ. Comp. Phys. 290, R861–R870. doi: 10.1152/ajpregu.00384. 2005


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Vio, Salas, Cespedes, Diaz-Elizondo, Mendez, Alcayaga and Iturriaga. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Acute Mountain Sickness Is Associated With a High Ratio of Endogenous Testosterone to Estradiol After High-Altitude Exposure at 3,700 m in Young Chinese Men

Xiao-Han Ding1,2, Yanchun Wang<sup>1</sup> , Bin Cui<sup>3</sup> , Jun Qin<sup>3</sup> , Ji-Hang Zhang<sup>3</sup> , Rong-Sheng Rao<sup>4</sup> , Shi-Yong Yu<sup>3</sup> , Xiao-Hui Zhao<sup>3</sup> and Lan Huang<sup>3</sup> \*

<sup>1</sup> Department of Health Care and Geriatrics, Lanzhou General Hospital of Lanzhou Military Region, Lanzhou, China, <sup>2</sup> Department of Geriatric Cardiology, Chinese PLA General Hospital, Beijing, China, <sup>3</sup> Department of Cardiology, Xinqiao Hospital, Army Medical University (Third Military Medical University), Chongqing, China, <sup>4</sup> Department of Ultrasonography, Xinqiao Hospital, Army Medical University (Third Military Medical University), Chongqing, China

### Edited by:

Rodrigo Del Rio, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Nicolas Voituron, Université Paris 13, France Julio Alcayaga, Universidad de Chile, Chile

> \*Correspondence: Lan Huang huanglan260@126.com

### Specialty section:

This article was submitted to Clinical and Translational Physiology, a section of the journal Frontiers in Physiology

> Received: 01 April 2018 Accepted: 22 December 2018 Published: 25 January 2019

### Citation:

Ding X-H, Wang Y, Cui B, Qin J, Zhang J-H, Rao R-S, Yu S-Y, Zhao X-H and Huang L (2019) Acute Mountain Sickness Is Associated With a High Ratio of Endogenous Testosterone to Estradiol After High-Altitude Exposure at 3,700 m in Young Chinese Men. Front. Physiol. 9:1949. doi: 10.3389/fphys.2018.01949 Background: A large proportion of populations suffer from acute mountain sickness (AMS) after exposure at high altitude. AMS is closely related with age and gender implying that the sex hormones may play critical roles in AMS. Our observational study aimed to identify the association between the endogenous testosterone (T), estradiol (E2) and AMS.

Methods: A total of 113 subjects were recruited in 2012. The participants were evaluated at 500 m and after acute (1 day) and short-term (7 days) high-altitude exposure at 3,700 m. The subjects also completed a case report form questionnaire and underwent blood pressure measurements and an echocardiography examination. The red blood cell (RBC) count, Hb concentration ([Hb]), hematocrit (HCT), E2, T, and erythropoietin (EPO) were measured.

Results: Upon acute high-altitude exposure, E2 and EPO were significantly lower in AMS<sup>+</sup> group, and T/E2 and stroke volume were higher. On the 1st day, AMS score correlated positively with the T/E2 ratio while it negatively correlated with E2. After 7 days at 3,700 m, the AMS<sup>+</sup> subjects had higher erythropoietic parameters: EPO, T, and T/E2 were significantly higher in the AMS<sup>+</sup> group. [Hb], RBC count, HCT, EPO, T and T/E2 were also correlated with AMS score. EPO, HCT, and the RBC count were also correlated with T/E2. Regression analyses indicated that T/E2 significantly correlated to AMS score and T/E2 on the 1st day was an independent predictor for AMS on the 7th day.

Conclusion: AMS was correlated with T/E2 ratio and EPO. After short-term exposure, higher T/E2 may contribute to AMS together with EPO via erythropoiesis. Furthermore, T/E2 level at high altitude in the early stage was an independent predictor for AMS in the latter stage.

Keywords: testosterone, estradiol, acute mountain sickness, erythropoiesis, high-altitude exposure

### Ding et al. T/E2 Predicts AMS

## INTRODUCTION

fphys-09-01949 January 24, 2019 Time: 16:46 # 2

Acute mountain sickness (AMS) has been defined as a syndrome involving headache, dizziness, gastrointestinal symptoms, insomnia and fatigue after arrival at high-altitude (>2500 m) (Imray et al., 2010; Bartsch and Swenson, 2013). The underlying mechanisms of AMS, which may involve the ventilation response, circulatory alteration and disequilibrium in the endocrine system induced by hypobaric hypoxia, have not been fully elucidated (Hackett and Roach, 2001; Imray et al., 2010; Davis et al., 2011; Bartsch and Swenson, 2013). The respiratory and cardiovascular aspects of AMS have been studied for many years; however, attention should be given to the endocrine aspects, particularly the effects of sex hormones and hematopoiesis on AMS (Naeije, 2010; Oliver et al., 2012). The endocrine system has been proposed to be of critical importance for the high-altitude adaptation process, which involves regulation of the hemoglobin (Hb) level (Gonzales, 2011). Previous studies have identified that several genes are associated with high-altitude adaptation and three of them are associated with [Hb] indicating that high-altitude adaptation and diseases may be associated with hematopoiesis and genetic factors (Simonson et al., 2010; Lorenzo et al., 2014). Testosterone (T) and E2 are the two most active sex hormones and are suggested to participate in AMS, based on the gender differences in the incidence of AMS (Harris et al., 1966; Wu et al., 2012).

Erythropoiesis is a hormonally regulated process that involves at least two hormones, EPO and testosterone (Richmond et al., 2005; Bachman et al., 2013). The crucial regulator of erythrocyte production, EPO, is elevated after short-term highaltitude exposure and hypoxia treatment. However, EPO does not increase proportionally with exposure time, and it even shows a tendency to decrease after reaching a peak (Koury, 2005; Gunga et al., 2007). Testosterone also enhances hemopoiesis and participates in many physiological functions, including erythropoiesis (Congote et al., 1977; Coviello et al., 2008; Rochira et al., 2009; Gonzales et al., 2011; Oskui et al., 2013; Shin et al., 2016). Though it has been suggested that high T concentrations, could compromise adaptation to high altitude, an association between T and EE in CMS has also been demonstrated (Gonzales et al., 2009, 2011), which indicates that T may participate in mountain sickness through its effects on erythropoiesis. The possibility that the variation in the normal physiological range of T concentration modulates men's adaptation to hypobaric high-altitude hypoxia by stimulating Hb production and/or causing respiratory disturbances and exacerbating hypoxemia has been suggested (Gonzales, 2011, 2013; Gonzales et al., 2011b). E2 exhibits opposite effects that limit EPO and RBC production, and testosterone is converted to E2 by aromatase directly or after exerting its action. Both pathways of the conversion will reduce T biological actions. Thus, the T/E2 ratio may be used as a indicator of sex hormones balance in many diseases.

Numerous investigations have been performed on the acute responses in respiratory and cardiovascular systems and on EE in CMS (Ou et al., 1994; Cornolo et al., 2004; Iturriaga et al., 2005; Imray et al., 2010); however, fewer studies have given proper attention to erythropoiesis, which is balanced by sex hormones and EPO upon acute high-altitude exposure, and their roles in AMS. We postulate that the indicator of the sex hormone balance, ratio of T to E2, may participate the pathophysiological process of AMS. Furthermore, T/E2 represent the balance of T and E2 which may indicate diseases. The current study aimed to identify the relationship between endogenous T, E2, T/E2, and AMS upon acute and a short-term of high-altitude exposures.

### MATERIALS AND METHODS

### Participants and Procedures Participants

Altogether, 113 subjects (6 were lost to follow-up) participated in the study according to the inclusion and exclusion criteria. The inclusion criteria were that the subjects be healthy males who were between 18 and 60 years old. The exclusion criteria were as follows: respiratory diseases, cardiovascular diseases, neuropsychosis, cerebrovascular diseases, malignant tumor, diseases of the liver, and diseases of the kidneys.

Subjects who participated in the study voluntarily were fully familiar with the purpose and procedures of this study and signed informed consent before the field trials. The study was reviewed and approved by the Ethics Committee of Xinqiao Hospital, the Third Military Medical University. The trials have been performed in accordance with the ethical standards laid down in an appropriate version of the Declaration of Helsinki.

### Procedures

The participants were recruited in June 2012. The field trials were performed in June 2012 within 18–24 h after the participants arrived at 3,700 m by plane, which occurred within a 2-h window, and on the 7th day after their arrival. The baseline data were recorded at 500 m within 1 week prior to departure for the high-altitude trials.

Structured CRF questionnaires were used to record demographic data (age, BMI, smoking and alcohol consumption) and the symptoms of AMS including headache (0 = no headache; 1 = mild headache; 2 = moderate headache; 3 = severe headache), dizziness (0 = no dizziness; 1 = mild dizziness; 2 = moderate dizziness; 3 = severe dizziness), gastrointestinal symptoms (0 = without and 1 = with any gastrointestinal symptom), insomnia (0 = normal sleep pattern; 1 = worse than usual; 2 = wake up multiple times during the night and 3 = extreme difficulty sleeping) and fatigue (0 = without fatigue and 1 = with fatigue). AMS was diagnosed by the LLS. Subjects were diagnosed with AMS if they had a headache upon arriving at high altitude and recorded an LLS score > 3. The SpO<sup>2</sup> and

**Abbreviations:** AMS+, with AMS; AMS−, without AMS; AMS, acute mountain sickness; BMI, body mass index; CIn, cardiac index; CMS, chronic mountain sickness; CO, cardiac output; CRF, case report form; DBP, diastolic blood pressure; E2, estradiol; EE, excessive erythrocytosis; EF, ejection fraction; EPO, erythropoietin; [Hb], Hb concentration; HCT, hematocrit; HIF, hypoxia-inducible factor; HR, heart rate; LLS, Lake Louise self-assessment scoring system; PCT, plateletocrit; PLT, platelet count; RBC, red blood cell; SBP, systolic blood pressure; SpO2, pulse oxygen saturation; SV, stroke volume; T, testosterone.

HR were measured by a pulse oximeter (NONIN-9550, Nonin Onyx, United States). The SV, EF, and CO were measured by ultrasonography using the S5-1 cardiac probe (CX50, Philips, United States). CIn were calculated using CO divided by body surface area.

The venous blood samples were obtained from all participants between 8 and 10 am after an overnight fast (at least 12 h). The measurements of RBC count, [Hb], HCT, PLT, PCT, E2, T, and EPO were as follow: Blood samples were obtained after a rest of 30 min by vein puncture and collected in tubes containing EDTA. Routine blood tests (RBC count, [Hb] and HCT) were performed using an automated hematology corpuscle analyzer (AU400; Olympus Optical, Co., Tokyo, Japan) at 3,700 m in the Laboratory Department in Army Brigade Hospital immediately after the blood samples were obtained. Routine blood tests [RBC count, [Hb] and HCT] were performed using an automated hematology corpuscle analyzer (AU400; Olympus Optical, Co., Tokyo, Japan). The venous blood samples were allowed to clot for 30–60 min and were then centrifuged at 1,000 g for 10 min at room temperature to obtain serum, which was immediately stored at −20◦C until the assays for hormonal analysis. The E2, T, and EPO were measured in duplicate using commercially available ELISA kits (human E2, T and EPO kits, BlueGene Ltd., China). To measure the exact concentrations of E2, T, and EPO, each sample was assayed in duplicate. All of the biochemical variables (except blood routine examinations) were measured from the blood specimens in the Clinical Laboratory Department, Xinqiao Hospital, Army Medical University, following the criteria of the World Health Organization Lipid Reference Laboratories.

## Statistical Analysis

The CRFs were excluded if the required items were not completed. The mean concentrations of E2, T, and EPO were calculated from duplicate wells in the ELISA plates. Because all variables (age, BMI, [Hb], RBC counts, HCT, E2, T, EPO, T/E2, HR, SV, EF, CO, and CIn) were normally distributed, they are expressed as mean ± standard deviation (SD). The changes in [Hb], RBC count, HCT, E2, T, EPO, T/E2, HR, SV, EF, CO, and CIn from sea level to 3,700 m (acute and short-term exposures) were evaluated by analyses of variance for repeated measurements. The variables mentioned above were compared between the AMS<sup>+</sup> and AMS<sup>−</sup> groups by the independent-samples Student's t-test on the 1st and 7th days at 3,700 m, respectively. The enumerated data (smokers and AMS) are expressed as the rate of occurrence (%). Additionally, the relationships between the AMS score and the above-mentioned parameters at 3,700 m were obtained by Spearman's correlation analyses. Multiple regression analyses were also performed after the AMS scores were normalized by natural logarithm transformation. Variable with p less than 0.3 was selected for further adjusted analyses after each of them was analyzed by univariate Logistic regression. The statistical analyses were performed using SPSS 19.0 software for Windows. P ≤ 0.05 was considered statistically significant. The detailed statistical methods and flow chart are summarized in **Figure 1**.

# RESULTS

The age of the subjects in this study was 23.08 ± 4.17 years, and the BMI was 21.56 ± 2.19 kg/m<sup>2</sup> . Overall, the incidences of AMS upon acute exposure and upon 1-week exposure at 3,700 m were 57.9%b (62/107) and 19.6% (21/107), respectively.

The alterations of SpO2, HR, [Hb], RBC count and HCT after exposures to 3,700 m were revealed in **Figures 2A–E**. The erythrocytosis-regulated hormone EPO increased significantly after acute exposure (3.31 ± 1.20 and 3.85 ± 1.42 mIU/ml, p = 0.045) and increased even further after short-term exposure to 3,700 m (4.46 ± 1.44 mIU/ml, **Figure 2F**). Furthermore, the PLT and PCT were decreased on the 1st day and recovered on the 7th day (**Figures 2G,H**). E2 had a slight increase (72.11 ± 14.78 vs. 74.07 ± 16.00 pg/ml, but not significantly) and T (1713.16 ± 468.06 vs. 1992.29 ± 423.96 pg/ml) had a significant increase from baseline to the acute time point. Both fell to a lower level than baseline at 7 days (64.26 ± 15.41 and 1592.82 ± 330.55 pg/ml, respectively) (**Figures 2I,J**). Though T decreased after a sharp elevation and E decreased after 7 days, the ratio of T to E2 increased significantly (24.76 ± 8.91 vs. 28.96 ± 8.93) from baseline to 24 h, and it remained at a relatively high level on the 7th day (25.48 ± 6.34) (**Figure 2K**). We also found that EF and SV increased after arrival at 3,700 m remained significantly increased after 7 days (**Figures 2L,M**). Both CO and CIn increased at 24 h and remained significantly increased after 7 days (**Figures 2N,O**).

Upon acute high-altitude exposure, E2 and EPOwere significantly lower in the AMS<sup>+</sup> group than in the AMS<sup>−</sup> group, while the T/E2 ratio was higher in the AMS<sup>+</sup> group without significant changes in T. SV was significantly lower in the AMS<sup>+</sup> group than the AMS<sup>−</sup> group. However, SpO2, HR, CO, [Hb], RBC count and HCT showed no significant differences between AMS<sup>+</sup> and AMS<sup>−</sup> individuals, as indicated in **Table 1**.

After short-term exposure to 3,700 m, the AMS<sup>+</sup> group had higher erythrocytosis parameters. Specifically, the AMS<sup>+</sup> subjects showed significantly higher [Hb], RBC count and HCT (all p-values < 0.05). We also observed that T and T/E2 were higher in AMS<sup>+</sup> individuals than in AMS<sup>−</sup> individuals, whereas E2 showed no differences between the two groups. In contrast to the data obtained after 24 h at high altitude, EPO was higher in the AMS<sup>+</sup> group than in the AMS<sup>−</sup> group after 7 days. The hemodynamic parameters showed no differences between the AMS<sup>+</sup> and AMS<sup>−</sup> groups (all p-values > 0.05, **Table 1**).

After a 24 h-exposure to 3,700 m, both T (p < 0.05) and T/E2 (p < 0.05) were highly correlated with the AMS score, and E2 (p < 0.05) was negatively correlated with the AMS score (**Table 2**). Additionally, the AMS score on the 1st day at 3,700 m was correlated with SV (p < 0.05) and HR (p < 0.05).

On the 7th day at 3,700 m, the AMS score was closely correlated with [Hb] (r = 0.206, p = 0.033), RBC (r = 0.307, p = 0.001) and HCT (r = 0.275, p = 0.004). Furthermore, T/E2 had a positive relationship with the AMS score (r = 0.241, p = 0.012). More importantly, the AMS score was correlated with EPO (r = 0.210, p = 0.030) (**Table 2**).

Among the erythropoietic and hemodynamic parameters, SV was significantly correlated with EPO (r = 0.199, p = 0.039) and T/E2 (r = −0.236, p = 0.015) on the 1st day at 3,700 m. The RBC count had significant positive relationships with EPO (r = 0.293) and T/E2 (r = 0.281, p = 0.003) after 7 days of exposure to high altitude. A significant relationship between EPO and T/E2 (r = 0.193, p = 0.046) was also observed (**Table 3**).

Regression analyses demonstrated that EPO and T/E2 (p < 0.001) correlated with the AMS score after 24 h of exposure to 3,700 m, while HCT and T/E2 were related to the AMS after 7 days (**Table 4**). Univariate regressions showed that E2 and SV at sea level were associated with AMS on the 1st day whereas SpO2, RBC and [Hb] can be included in the adjusted regression to identify the predictor for AMS on the 7th day (**Table 5**). Regarding to variables within 24 h at 3,700 m, only SV and T (T/E2) entered adjusted Logistic analyses. The adjusted Logistic regressions showed that baseline RBC count can predict AMS on the 7th day. Furthermore, T/E2 on the 1st day at 3,700 m is an independent predictor for AMS after 7 days' exposure (**Table 6**).


TABLE 1 | Differences of sex hormones, EPO, cardiovascular, hematopoietic and erythropoiesis between AMS<sup>+</sup> and AMS<sup>−</sup> groups.

SpO2, %; HR, beat per min; E2, pg/ml; EPO, mIU/ml; T, pg/ml; [Hb], g/L; RBC, 1012/L; HCT, l/L; SV, ml; EF, %; CO, L/min and CIn, L/(min · m<sup>2</sup> ); PCT, l/L; PLT, 10<sup>9</sup> /L.


T/E2 was the only variable significantly related to AMS score after 24 h and 7 days high-altitude exposure.

# DISCUSSION

We have found that the AMS was closely associated with T/E2 and the T/E2 may predict AMS in a short term of time in this study.

# Associations Between AMS and Sex Hormones, EPO, and Hematopoiesis

Few studies have examined the relationships among sex hormones, EPO, hematopoiesis, and AMS. Most reports have demonstrated that CMS is characterized by high T concentrations and erythropoiesis, mainly due to the relationships among the T level, erythropoiesis and the symptoms of headache and sleep disturbance (Ou et al., 1994; Gonzales et al., 2009, 2011; Gonzales, 2011, 2013; Ekart et al., 2013). This evidence suggests that T and erythropoiesis may also participate in the pathogenesis of AMS. Hence, we observed that AMS was characterized by higher T, EPO, HCT, RBC count and [Hb] after 7 days at high altitude, though the T, HCT, RBC count and [Hb] were not significantly different between the AMS<sup>+</sup> and AMS<sup>−</sup> groups following acute exposure.

These results partly agree with the previous findings that high serum T and Hb are adequate for acclimatization due to the improvement of oxygen transport, while even higher levels of T and Hb are associated with EE (Gonzales, 2011; Ekart et al., 2013), which may account for the positive relationships between T/E2 ratio and AMS score on the 1st day as well as association of after 7 days of exposure. However, E2 was higher in the AMS<sup>−</sup> group than in the AMS<sup>+</sup> group, which may suggest that E2 may play a protective role in AMS upon acute exposure. EPO was also higher in the AMS<sup>−</sup> group upon acute exposure, which suggests that in the acute phase, higher E2 and EPO may be the primary protective factors against AMS, perhaps due to the respiratory action of E2 and the non-hematopoietic effect of EPO (see below). Furthermore, E2 may protect individuals from AMS via the relief of pulmonary artery contraction through its depressive effect on pulmonary endothelin-1 induced by hypoxia (Earley and Resta, 2002). Though both EPO and E2 appear to protect against AMS, E2 downregulates HIF response genes, including EPO (Earley and Resta, 2002). It is possible that the depression of EPO by E2 is attenuated by acute hypoxia. Thus, the preventive effect of E2 may also be ascribed to its indirect role in polycythemia and right ventricular functions in the acute phase.

High T and T/E2 are correlated with hypoventilation, sleep disturbances, EE and, in turn, the etiopathogenesis of CMS in natives (Gonzales et al., 2011b). E2 prevents CMS via its protective effects on cardiovascular and respiratory functions

TABLE 3 | Associations among erythropoiesis, EPO, echocardiography, and sex hormones.


TABLE 4 | Multiple linear regressions at 3,700 m.


#Normalized by natural logarithm transformation.

95%CI lower bound and 95%CI upper bound are the upper and lower limits of the 95% confidence interval.

β, regression coefficient.

t, statistic of T-test for the regression.

(Zabka et al., 2006; Mattingly et al., 2008). Thus, it is reasonable to postulate that T and E2 play opposing roles in AMS, including the hematopoietic and respiratory effects.

Though T/E2 at sea level cannot be used as independent predictors for AMS neither on the 1st day nor the 7th day, T/E2 values upon acute exposure at 3,700 m was shown to be an independent predictor for AMS after a short-time exposure.

## Sex Hormones May Participate in AMS via Hematopoietic or Non-hematopoietic Effect of EPO

The production of erythrocytes by erythropoiesis requires time, which is in accordance with our result that the RBC count was unchanged in the acute phase followed by a rise at 7 days. Previous study showed that the plasma volume and blood volume were decreased about 10–12% after acute high altitude exposure. However, they will be recovered in few days to 1 or 2 weeks at high altitude or after return to sea level. There was an interesting result that the HCT and PCT were decreased on the 1st day following by PCT remained significantly reduced after 7 days while HCT increased, which may be caused by the water-sodium retention in acute phase. This have been also reported in few studies which were not completely coincident with other researches. Also, in the days at high altitude, the HCT were significantly increased combined with the increased of [Hb] and RBC, may revealed that the decreases of plasma volume and blood volume have emerged. However, on the 7th day, the mean increase percentages of [Hb] and RBC were 24 and 12%, whilst the HCT evaluated only 9% compared with sea level. This mismatching of [Hb] or RBC with HCT may not be explained by the decrease of PV completely. Which may indicate the erythropoiesis process have started by EPO (peak level appeared on 2–4 days at high altitude). Indeed, it is important to measure the plasma volume and blood volume directly to evaluate the changes of them. However, we searched the published papers and note that the simplest and easily get methods as follow: PV = BV − RCV; BV = RCV<sup>∗</sup> 100/Hct; RCV = tHb/MCHC<sup>∗</sup> 100. In our study, it is difficult for us to measure the total tHb due the difficult of flied study at high altitude. Thus, we cannot directly evaluate PV in this study. Therefore, the following explanation of our results dependent on the increased amplitudes in [Hb], RBC count and HCT indirectly. That is, in the short terms of high altitude exposure, both of PV decrease and erythropoiesis have emerged. Further analysis showed that the erythropoiesis may be preponderant in our study, which needs further precise studies.

Acute mountain sickness was characterized by higher erythropoiesis after 7 days of exposure, but it was not associated with [Hb] or RBC count upon acute exposure to high altitude. What are the underlying mechanisms by which EPO participates in AMS during acute exposure, if not by promoting erythropoiesis? The non-hematopoietic effect of EPO has been revealed by studies that have reported its inotropic effect on the myocardium, which results in elevated EF (Riksen et al., 2008). This observation may be interpreted as a protective effect of EPO in the acute exposure phase, before erythropoiesis plays a major role, as indicated by the relationship between SV and EPO in the present study.

E2 per se has a protective effect on the cardiovascular system and a stimulatory effect on the respiratory system (Mattingly et al., 2008; Smith et al., 2014). Combined with the enhanced respiration effects of EPO, we can suggest that the regulation of the respiratory and cardiovascular systems by E2 and EPO may account for their roles in preventing AMS in the case of acute exposure.

However, after a short time, the modifications of RBC, HCT, and [Hb] induced by the synergism of T and EPO were of critical importance in AMS. The RBC count had a positive correlation with EPO, which was also positively correlated with T. T also has a direct erythropoiesis effect (Gonzales et al.,



The variable with p-value less than 0.3 was selected for the adjusted Logistic regression.

β, regression coefficient.

OR, odds ratio.


RBC at sea level predicts AMS independently while T/E2 at 3,700 m on the 1st day is the independent predictor for AMS after 7 days.

2009; Gonzales, 2011), which was revealed here by the positive associations of the RBC count and HCT with the T level. Furthermore, our novel observation that EPO was higher in the AMS<sup>+</sup> group after 7 days of exposure but that it was lower in AMS<sup>+</sup> individuals upon acute exposure may further indicate that EPO has a diphasic role in AMS, depending on the duration of hypoxic exposure. Thus, in the short-term condition, T on its own and in combination with EPO induced erythropoiesis, perhaps enhancing the viscosity of the blood, which is harmful for acclimating to high-altitude hypoxia (Heinicke et al., 2003; Bachman et al., 2013; Ekart et al., 2013). Though previous studies indicate that EPO does not play a further important role in EE (Koury, 2005; Gunga et al., 2007), it may stimulate erythrocytosis when its effects are combined with the effects of T. T induces erythrocytosis via direct effects on bone marrow and indirect effects on EPO (Gonzales et al., 2009; Maggio et al., 2013), which finally results in the augmentation of [Hb], HCT, and RBC count that aggravate AMS.

We may conclude that E2 and EPO prevent AMS via the non-hematopoietic effect (inotropic effect on the myocardium, alterations in hemodynamics) in the acute phase, while T and EPO aggravated AMS through erythropoiesis (alterations in the hematological system) after short-term exposure (**Figure 3**). Additionally, E2 was still beneficial to individuals through the stimulation of ventilation in the latter phase.

pathophysiological processes in a time dependent way.

# Disequilibrium of Endogenous E2 and T May Be Involved in AMS

T and E2 are the most active sex hormones in the human body. They play opposite roles in respiratory regulation and erythropoiesis (Ou et al., 1994). T is converted to E2 by aromatase after acting on erythrocytosis (Jiang et al., 2010). Thus, the ratio of T/E2, which is determined by the levels of T and E2 as well as aromatase, is more important than the absolute value of each sex hormone.

Though the effects of T/E2 on SpO<sup>2</sup> have been reported, which our results do not corroborate the relationship between SpO<sup>2</sup> and T/E2 (Gonzales and Villena, 2000) (which our results do not corroborate), as have the effects of T/E2 on respiratory regulation, erythrocytosis, and even CMS, few studies have focused on AMS. Our observations reveal that T/E2 was closely associated with the AMS score, which indicates the importance of T/E2 in the pathophysiological process of AMS. The erythropoietic effect of T occurred before aromatase could convert it into E2, which had no effect on erythropoiesis (even with a reverse action) (Zabka et al., 2006). On the other hand, T enhanced EPO production and erythropoiesis and induced hypoventilation, while E2 enhanced ventilation, increased SpO<sup>2</sup> and inhibited EPO and RBC production. However, T may reduce these effects by down regulating the estradiol and progesterone receptors in response to ventilation, favoring hypoventilation and stimulating erythropoiesis (Ou et al., 1994; Zabka et al., 2006; Jiang et al., 2010).

Based on the evidence described above for CMS and our observations, it may be possible to regulate the T/E2 ratio by increasing the aromatization of T into E2, increasing aromatase expression or even applying E2 exogenously to prevent and treat AMS. Furthermore, T/E2 may participate in AMS via erythropoiesis, as indicated by the associations among T/E2, EPO, and RBC. The precise mechanisms by which T/E2 participates in AMS warrant further basic research.

# The Predictive Roles of T/E2 in AMS at High Altitude

The T/E2 at sea level was on an independent predictor for AMS on the 1st day or the 7th day at 3,700 m indicated by both univariate or adjusted regressions. However, the T/E2 on the 1st day at 3,700 m can predict AMS after 7 days indicating that the role of T/E2 was in progressing pathophysiological progress. Though, the sea level T/E2 failed to predict AMS, its association between AMS has been revealed in our present study. In addition, the predictive of T/E2 may be dependent on other risk factors that warrant further study.

### CONCLUSION

fphys-09-01949 January 24, 2019 Time: 16:46 # 10

Acute mountain sickness was correlated with the ratio of T to E2 and the level of EPO. Our observations suggest that higher T/E2 may promote AMS by acting through or combining with EPO to affect erythropoiesis, however, the mechanisms by which this may occur warrant further study. Furthermore, though T/E2 at sea level was not an independent predictor for AMS, the T/E2 on the 1st day at 3,700 m was an independent predictor for AMS on the 7th day.

### Limitations

The subjects of this study were all young male workers, mostly aged between 16 and 40 years, which could perhaps generate a bias due to age or gender. The current study is an observational study with a small sample size, and larger sample sizes and mechanistic studies are needed.

## AUTHOR CONTRIBUTIONS

LH and X-HD participated in the design of this research. X-HD, YW, and LH drafted the manuscript and performed the statistical

### REFERENCES


analyses. LH, J-HZ, and JQ reviewed and revised this manuscript critically for important intellectual content. X-HD, J-HZ, BC, and X-HZ completed the collection of clinical data and performed the measurements of BP, HR, and the biomarkers (T, E2, and blood routine tests). R-SR and S-YY performed the echocardiography examinations.

### FUNDING

This work was supported by the Ministry of Health of the People's Republic of China (Grant No. 201002012) and People's Liberation Army China (Grant No. BWS14J040).

### ACKNOWLEDGMENTS

We would like to thank all of the individuals who participated in this study for their support. We are also grateful to Mr. Can Chen, Bai-Da Xu, Guo-Zhu Chen, Xu-Gang Tang, and Cheng-Rong Zheng and Miss. Wen-Yun Guo, Shuang-Fei Li, and Yang Liu for their helps in the data collection.


The results of a clinical trial in older men. Andrology 1, 24–28. doi: 10.1111/j. 2047-2927.2012.00009.x


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Ding, Wang, Cui, Qin, Zhang, Rao, Yu, Zhao and Huang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

fphys-09-01949 January 24, 2019 Time: 16:46 # 11

# Circulating Apoptotic Signals During Acute and Chronic Exposure to High Altitude in Kyrgyz Population

Djuro Kosanovic1,2, Simon Maximilian Platzek<sup>1</sup> , Aleksandar Petrovic<sup>1</sup> , Akylbek Sydykov<sup>1</sup> , Abdirashit Maripov<sup>3</sup> , Argen Mamazhakypov<sup>1</sup> , Meerim Sartmyrzaeva<sup>3</sup> , Kubatbek Muratali Uulu<sup>3</sup> , Meerim Cholponbaeva<sup>3</sup> , Aidana Toktosunova<sup>3</sup> , Nazgul Omurzakova<sup>3</sup> , Melis Duishobaev<sup>3</sup> , Christina Vroom<sup>1</sup> , Oleg Pak<sup>1</sup> , Norbert Weissmann<sup>1</sup> , Hossein Ardeschir Ghofrani<sup>1</sup> , Akpay Sarybaev<sup>3</sup> and Ralph Theo Schermuly<sup>1</sup> \*

### Edited by:

Rodrigo Iturriaga, Pontificia Universidad Católica de Chile, Chile

### Reviewed by:

Vincent Joseph, Laval University, Canada Soumya Pati, Shiv Nadar University, India

### \*Correspondence:

Ralph Theo Schermuly ralph.schermuly@ innere.med.uni-giessen.de

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 29 June 2018 Accepted: 17 January 2019 Published: 05 February 2019

### Citation:

Kosanovic D, Platzek SM, Petrovic A, Sydykov A, Maripov A, Mamazhakypov A, Sartmyrzaeva M, Muratali Uulu K, Cholponbaeva M, Toktosunova A, Omurzakova N, Duishobaev M, Vroom C, Pak O, Weissmann N, Ghofrani HA, Sarybaev A and Schermuly RT (2019) Circulating Apoptotic Signals During Acute and Chronic Exposure to High Altitude in Kyrgyz Population. Front. Physiol. 10:54. doi: 10.3389/fphys.2019.00054 <sup>1</sup> Chair for Pulmonary Pharmacotherapy, Member of the German Center for Lung Research, Universities of Giessen and Marburg Lung Center, Giessen, Germany, <sup>2</sup> Sechenov First Moscow State Medical University (Sechenov University), Moscow, Russia, <sup>3</sup> Kyrgyz National Centre for Cardiology and Internal Medicine, named after Academician Mirsaid Mirrakhimov, Bishkek, Kyrgyzstan

Background: Circulating apoptotic signals (CASs) have been described in the pathologies associated with dysregulated apoptosis, such as cancer, heart diseases, and pulmonary hypertension (PH). However, nothing is known about the expression profiles of these markers in the circulation of humans exposed to acute and chronic effects of high altitude (HA).

Methods: Gene expression levels of different apoptotic signals (ASs) were analyzed in human pulmonary artery smooth muscle cells (PASMCs) upon hypoxia incubation. In addition, we measured the plasma values of relevant CAS in Kyrgyz volunteers during acute and chronic exposure to HA. Finally, we analyzed the effects of pro-apoptotic mediator Fas ligand (FasL) on apoptosis and proliferation of human PASMCs.

Results: Several cellular AS were increased in PASMCs exposed to hypoxia, in comparison to normoxia condition. Among analyzed CAS, there was a prominent reduction of FasL in lowlanders exposed to HA environment. Furthermore, decreased circulatory levels of FasL were found in highlanders with HA-induced PH (HAPH), as compared to the lowland controls. Furthermore, FasL concentration in plasma negatively correlated with tricuspid regurgitant gradient values. Finally, FasL exerted pro-apoptotic and anti-proliferative effects on PASMCs.

Conclusion: Our data demonstrated that circulating levels of FasL are reduced during acute and chronic exposure to HA environment. In addition, dysregulated FasL may play a role in the context of HAPH due to its relevant functions on apoptosis and proliferation of PASMCs.

Keywords: high altitude, circulating apoptotic markers, Fas ligand, hypoxic pulmonary hypertension, pulmonary artery smooth muscle cells

# INTRODUCTION

fphys-10-00054 February 1, 2019 Time: 17:51 # 2

High altitude is a well-known extreme environment characterized by hypoxia, among other abiotic factors, and may exert prominent acute and chronic effects on respiratory and cardiovascular systems (Maggiorini and Leon-Velarde, 2003; West, 2012; Mirrakhimov and Strohl, 2016). An important part of the human population inhabits high altitudes of our planet and a large number of people visit these elevations periodically due to several reasons (Mirrakhimov and Strohl, 2016; Azad et al., 2017). Short-term exposure of non-acclimatized people to high altitude may provoke the appearance of several acute mountain disorders, including acute mountain sickness and high-altitude pulmonary edema (West, 2012). Long-term exposure to this challenging external surrounding may lead to development of high altitude-induced pulmonary hypertension (HAPH), which is a pathological condition currently classified in the group 3 of pulmonary hypertension (PH) (Maggiorini and Leon-Velarde, 2003; Simonneau et al., 2013; Mirrakhimov and Strohl, 2016).

In general, pulmonary vascular remodeling as the main attribute of PH pathology appears due to significant dysregulation of normal processes in the pulmonary vascular cells, such as proliferation and apoptosis (Schermuly et al., 2011; Malczyk et al., 2016; Stenmark et al., 2018; Thenappan et al., 2018). Abnormal regulation of apoptosis and occurrence of "apoptotic-resistant" phenotype in pulmonary vascular cells, such as pulmonary artery smooth muscle cells (PASMCs), are important hallmarks of hypoxia-induced PH pathogenesis (Yang et al., 2001; Chen et al., 2014; Li et al., 2017; You et al., 2018). Due to the existence of these cellular abnormalities, e.g., increased proliferation and resistance to apoptosis, PH is often described as a "tumor-like disease" in recent years (Chan and Rubin, 2017; Marshall et al., 2018). However, despite clear advancement in understanding of the role of apoptosis in hypoxia-associated PH and PH in general, many questions remain unresolved and insufficiently investigated.

A disbalance between pro- and anti-apoptotic molecular pathways has been described as an important phenomenon in the cancer field (Holdenrieder and Stieber, 2004). In addition to the general pathobiology of cellular apoptosis, several circulating apoptotic signals (CASs) have been evaluated as promising biomarkers relevant to the tumor pathologies (Holdenrieder and Stieber, 2010). In the field of cardiovascular diseases, some of the apoptotic markers have been also identified in the blood circulation (Parissis et al., 2004; Niessner et al., 2009). With regard to the pulmonary vascular disease, it has been demonstrated that the levels of CASs, namely Fas ligand (FasL) and tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL), are changed in response to therapy or altered in patients with PH, respectively (Akagi et al., 2013; Liu et al., 2015). However, almost nothing is known about the profiles of circulating apoptotic markers in the context of human individuals exposed to high altitude hypoxic condition.

Therefore, we investigated for the first time the potential changes in the plasma levels of different apoptotic signals, including FasL, TRAIL and apolipoprotein C1 (ApoC1), in human individuals who live permanently in high altitude regions and lowland subjects who spent a short period of time in such environment. Finally, the potential alteration in gene expression profile of various apoptotic players, for example caspase (Casp) 1 and 3, survivin, Fas-associated death domain protein (FADD)-like ICE inhibitory protein (FLIP), ApoC1, TRAIL and FasL, was analyzed in human PASMCs exposed to hypoxia for different time durations.

### MATERIALS AND METHODS

### Study Design

In order to analyze the circulating profiles of various apoptotic markers during acute and chronic exposure to high altitude hypoxia, we have performed two studies among the Kyrgyz population.

First, for the purpose of investigating the people permanently living at high altitude environment, we have selected the human community of Sary-Mogol and Achyk-Suu villages, which are located in Alay and Chon-Alay districts of the province of Osh in the southern part of Kyrgyzstan. These villages are settled at altitudes of 3000–3100 m, but the most of the residents usually spend 3–4 months per year at even higher elevations (3200– 3600 m). Importantly, we ensured that all subjects enrolled in the study were ethnic Kyrgyz and were born and permanently living at high altitudes. As the lowland control, we have included the volunteers from Bishkek, Kyrgyzstan (approximately 760 m above the sea level). All participants of the study underwent the general anthropometric and echocardiographic measurements (please see below), followed by collection of the peripheral blood and separation of the citrated platelet free plasma (PFP).

Secondly, for the purpose of investigating the people acutely exposed to high altitude hypoxia, we had a group of healthy Kyrgyz males. Initially, the subjects underwent examination in Bishkek (low altitude (LA 1) at 760 m, with the outside temperature ranging from 28 to 35◦C during the study execution). After that, the participants were transported by road to an altitude of 3200 m (Tuya-Ashuu pass, Kyrgyzstan) (**Figure 1**), where the outside temperature was in the range 5–20◦C, while inside the rooms in the research station the temperature was maintained at 22 ± 2 ◦C. The first 2 days after arrival to high altitude environment the subjects took

FIGURE 1 | High altitude research station near to the Tuya-Ashuu pass (3200 m), Kyrgyzstan. The photograph is original work of one of the authors.

complete rest. During the next 2 days they were allowed to walk at plains and downhill. Thereafter, the participants were involved in common activities, such as indoor games, watching TV, playing table tennis or billiard, along with morning drill, walking and other galley duties. Importantly, all subjects were free of cardiac or neurological problems and were not consuming any kind of medications. In addition to the basal examination in Bishkek (LA 1) at low altitude, all individuals underwent echocardiographic measurements, followed by the collection of the peripheral blood and separation of the EDTA plasma on days 2 (HA 2), 7 (HA 7), and 20 (HA 20) of high altitude exposure, and on the second day after descent to Bishkek at low altitude again (LA 2). In addition, anthropometry was performed in all participants.

Both studies protocols were approved by the Ethics Committees of the National Centre for Cardiology and Internal Medicine, Bishkek, Kyrgyzstan (01-1/08 and 01-1/07) and the faculty of Medicine at Justus-Liebig University, Giessen, Germany (AZ: 236/16). The study was performed in agreement to the principles outlined in the Declaration of Helsinki of the World Medical Association. Finally, written informed consent was obtained from all participants.

### Anthropometric and Echocardiographic Measurements

Several general parameters were obtained, such as age, gender ratio, body mass index (BMI), and body surface area (BSA). BMI (kg/m<sup>2</sup> ) was calculated using the formula: BMI = weight (kg)/(height (m)<sup>2</sup> ). BSA (m<sup>2</sup> ) was calculated using the Du Bois formula as follows: BSA = 0.007184 × [(height (m) × 100)0.725] × (weight (kg)0.425). The right ventricular to right atrial pressure gradient was used as a surrogate of the estimated systolic pulmonary artery pressure. Continuouswave Doppler echocardiography was employed to estimate the tricuspid regurgitant gradient (TRG, in mmHg) from the peak flow tricuspid regurgitation velocity by means of the simplified Bernoulli equation measured using continuous-wave Doppler, as previously described (Yock and Popp, 1984). All the above mentioned parameters are presented in the **Table 1** (acute high altitude exposure) and **Table 2** (chronic exposure to high altitude). In the case of acute high altitude study, all general parameters, including age, gender, BMI, and BSA were comparable among all participants (**Table 1**). TRG values were initially increased on the second day (HA 2) of high altitude

TABLE 2 | Chronic exposure to high altitude – human subjects' anthropometric and echocardiographic data.


Adult human subjects permanently living at lowland regions (LA) (n = 10) and high altitude (HA) locations were included for the study. Individuals settled at high altitude regions were separated into two groups: highlanders without pulmonary hypertension (PH) (HA) (n = 10) and highlanders with PH (HA-PH) (n = 12). Various anthropometric and echocardiographic parameters were presented. Results are shown as Mean ± SD (n = 10–12). ∗∗∗∗p < 0.0001 LA compared to the HA-PH; §§§§ p < 0.0001 HA compared to the HA-PH. one-way ANOVA with Tukey's multiple comparisons test was performed for statistical analysis. m, male; f, female; BMI, body mass index (in kg/m<sup>2</sup> ); BSA, body surface area (in m<sup>2</sup> ); TRG, tricuspid regurgitant gradient (in mmHg).

exposure in comparison to the lowlands (LA 1), and later gradually decreased until return to the low altitude (LA 2) (**Table 1**). There were statistically significant elevations of TRG on the days 2 and 7 at high altitude, as compared to the lowland conditions (LA 1) (**Table 1**). Also, there was a significant increase of TRG values on the day 2 of high altitude exposure, in comparison to the low altitude upon return (LA 2) (**Table 1**).

In the case of chronic exposure to high altitude (**Table 2**), BMI and BSA were comparable among all 3 groups: lowland control subjects (LA), highlanders without (HA; TRG ≤ 23 mmHg) and with PH (HA-PH; TRG ≥ 40 mmHg). However, there were noticeable differences in gender ratio and age (lowland control group consisted only of younger males). This fact is a potential limitation of our study. TRG values were significantly increased in highlanders with PH, as compared to the lowland controls and highlanders without PH (**Table 2**).

## Cell Culture and RT-qPCR

Primary human PASMCs were purchased from Lonza and cultured in Smooth Muscle Growth Medium-2 containing supplement-mix. Human PASMCs passage 7 were incubated at 37◦C in a humidified atmosphere of 5% CO<sup>2</sup> in either normoxic (21% O2) or hypoxic conditions (1% O2) for a period of 24, 48, and 72 h. Upon the end of the incubation period cells were lysed in RLT buffer (Qiagen) and total cell RNA was extracted using the RNeasy Mini Kit (Qiagen). Complementary DNA (cDNA)

TABLE 1 | Acute exposure to high altitude – human subjects' anthropometric and echocardiographic data.


Adult human subjects who initially came from the lowlands (LA 1) (n = 8) were exposed to high altitude (HA) for 2 (HA 2) (n = 8), 7 (HA 7) (n = 8), and 20 (HA 20) (n = 8) days. After 20 days, they returned to the lowlands again (LA 2) (n = 8). Various anthropometric and echocardiographic parameters at different time points were presented in the table. Results are shown as Mean ± SD (n = 8). <sup>∗</sup>p < 0.05; ∗∗∗p < 0.001 compared to the LA 1. §§ p < 0.01 compared to the LA 2. Friedman test with Dunn's multiple comparisons test was performed for statistical analysis. m, male, f, female; BMI, body mass index (in kg/m<sup>2</sup> ); BSA, body surface area (in m<sup>2</sup> ); TRG, tricuspid regurgitant gradient (in mmHg).




FP, forward primer; RP, reverse primer.

was produced by reverse transcriptase polymerase chain reaction via iScript cDNA Synthesis Kit (Bio-Rad). Real-time PCR was performed in Mx3000P qPCR system (Stratagene) using the iTaq Universal SYBR Green Supermix (Bio-Rad). Primer sequences are given from (5<sup>0</sup> to 3<sup>0</sup> ) and presented in the **Table 3**. Human porphobilinogen deaminase (PBGD) served as a housekeeping gene. In order to confirm specific amplification of the expected PCR product, gel electrophoresis and melting curve analysis were performed.

### Apoptosis and Proliferation Assays

For assessment of apoptosis, human PASMCs (3000 cells per well) were seeded in 96-well white-walled plate in Smooth Muscle Growth Medium-2. After 48 h of recovery time cells were stimulated with different concentrations of SuperFasLigand (Enzo Life Sciences) and incubated at 37◦C in water saturated incubators for 24 h under normoxic conditions (21% O2). The apoptosis Caspase-Glo 3/7 assay (Promega) was performed following the manufacturer's instructions. The luminescence of the caspase cleaved substrate reaction was measured after 30 min of incubation at room temperature.

Proliferation of human PASMCs exposed to normoxic (21% O2) or hypoxic conditions (1% O2) was assessed by Cell Proliferation ELISA, 5-bromo-2<sup>0</sup> -deoxyuridine (BrdU) assay (Sigma Aldrich). In more details, 7500 cells per well were seeded into 24-well plates and following the 24 h starvation period PASMCs were incubated for 24 or 48 h in Smooth Muscle Growth Medium-2. After that the cells were treated with different concentrations of SuperFasLigand or 500 ng/ml of neutralizing anti-Fas antibody ZB4 (Millipore). In addition, PASMCs were incubated for 48 h in Smooth Muscle Basal Medium and stimulated with 50 ng/ml of platelet-derived growth factor (PDGF) (R&D Systems) and/or 500 ng/ml of neutralizing anti-Fas antibody ZB4. The absorbance of the substrate reaction was measured at 370 nm (reference wavelength 492 nm).

### Enzyme-Linked Immunosorbent Assay (ELISA)

Citrated platelet free plasma (PFP) or EDTA plasma samples were obtained from the peripheral blood from all participants of the study. In order to analyze the circulating levels of different markers, several ELISA measurements were performed for the following targets (ELISA kits): ApoC1 (Abnova), TRAIL (R&D Systems), FasL (LSBio/R&D Systems) and B-type natriuretic peptide (BNP) (Abnova). The concentrations of the markers were expressed as µg or pg per mL of plasma.

### Data Analysis

Results are presented as Mean ± SD. Unpaired t-test with Welch's correction or ordinary one-way ANOVA with Dunnett's or Tukey's multiple comparisons test were used to compare data derived from the cell culture. Data based on ELISA and echocardiographic measurements were analyzed using Friedman test with Dunn's multiple comparisons test, RM one-way ANOVA with Tukey's multiple comparisons test or ordinary one-way ANOVA with Tukey's multiple comparisons test. Spearman or Pearson tests were used for the correlation analysis. Statistical significance was considered when p-value was < 0.05, < 0.01, < 0.001, and < 0.0001.

### RESULTS

### Alteration in the Gene Expression Levels of Various Apoptotic Markers in Human PASMCs After Hypoxic Incubation

As already described in Section "Materials and Methods," human PASMCs were exposed to hypoxia condition for different time durations (24, 48, and 72 h), followed by quantification of the gene expression levels of various apoptotic markers, such as Casp1 and 3, survivin, FLIP, ApoC1, TRAIL, and FasL (**Figures 2**, **3**). In general, all investigated apoptotic signals had increased their expression profiles upon hypoxic stimuli in human PASMCs (**Figures 2**, **3**), as compared to the normoxia condition. However, there were no clear and uniform time-dependent changes in the expression profiles of investigated apoptotic markers. The most prominent and statistically significant upregulation was noticed in the case of survivin and FasL, during all time points of hypoxia exposure, in comparison to the respective normoxic controls (**Figures 2C**, **3C**).

### Circulating Profiles of Apoptotic Markers in Human Subjects Exposed to Acute High Altitude Conditions

We have already mentioned in the Section "Materials and Methods" that circulating levels of different apoptotic markers, such as ApoC1, TRAIL and FasL, were measured by ELISA in the plasma samples of human volunteers taken at different time points: in the lowland conditions (LA 1), on days 2 (HA 2), 7 (HA 7), and 20 (HA 20) of high altitude exposure, and

FIGURE 2 | Gene expression profiles of different apoptotic markers in human pulmonary arterial smooth muscle cells (PASMCs) exposed to hypoxia I. Human PASMCs were exposed to hypoxia (Hox) for different time durations (24, 48, and 72 h), followed by RT-qPCR for measurement of various apoptotic markers, such as: (A) caspase 1 (Casp1) and (B) 3 (Casp3), (C) survivin, and (D) Fas-associated death domain protein (FADD)-like ICE inhibitory protein (FLIP). Normoxia (Nox) exposure served as a control. Results are presented as a fold change normalized to the respective normoxic controls. Results are expressed as Mean ± SD (n = 4). <sup>∗</sup>p < 0.05, ∗∗p < 0.01; ∗∗∗p < 0.001 Nox versus Hox. Unpaired t-test with Welch's correction was performed for statistical analyses.

with Welch's correction was performed for statistical analyses.

after return to the low altitude (LA 2). In addition, ELISA was performed in the plasma samples during the same time points in order to analyze the level of circulating BNP. Investigated circulating apoptotic markers had different expression profiles and due to the technical reasons not all values for all enrolled subjects are available. Apo C1 circulating levels (in µg/mL) were comparable among different time points, except on day 7 (HA 7) of high altitude exposure, when the levels of this marker were significantly decreased in comparison to the low altitude upon descent from the high altitude (**Figure 4A**). Further, circulating

levels of TRAIL (in pg/mL) were gradually increasing due to high altitude exposure, with being significantly enhanced on days 7 (HA 7) and 20 (HA 20), in comparison to the lowland conditions (LA 1) (**Figure 4B**). After returning to the lowlands again, there was a significant decrease of circulating TRAIL levels (**Figure 4B**). In the case of FasL (in pg/mL), there was clear, but not timedependent reduction of the circulating levels of this marker in high altitude conditions, as compared to the lowlands (LA 1) (**Figure 4C**). Surprisingly, decreased levels of circulating FasL were kept upon return to the low altitude (**Figure 4C**). Finally, there was a noticeable trend of increased levels of circulating BNP (in pg/mL) during different time points at high altitude, with prominent reduction upon return to the lowlands (**Figure 4D**). However, the significant difference was present only between the circulating levels of this marker on day 7 (HA 7) and the second low altitude condition (LA 2) (**Figure 4D**).

### Circulating Profiles of Apoptotic Markers in Kyrgyz Highlanders and Lowlanders

As already indicated in the Section "Materials and Methods," circulating levels of different apoptotic markers, such as ApoC1, TRAIL and FasL, were measured by ELISA in the plasma samples of human subjects permanently living at high altitudes,

FIGURE 5 | Circulating apoptotic markers in human subjects permanently living at high altitude. Human subjects permanently living at high altitude regions of Kyrgyzstan (highlanders) were separated into two groups: individuals without developed pulmonary hypertension (Non-PH) (n = 9–10) and individuals with this pulmonary vascular disease (PH) (n = 10). People living at the low altitude served as a control (n = 9–10). Echocardiographic measurements and collection of the peripheral blood were performed for all volunteers. Plasma was separated and enzyme-linked immunosorbent assay (ELISA) was performed for the detection and quantification of the following circulating apoptotic markers: (A) apolipoprotein C1 (ApoC1), (B) TNF-related apoptosis-inducing ligand (TRAIL), and (C) Fas ligand (FasL). In addition, the circulating profile of B-type natriuretic peptide (BNP) was analyzed by ELISA. (D) Results are expressed as concentrations of the above mentioned markers (in µg or pg per mL of plasma) and presented as Mean ± SD (n = 9–10). ∗∗p < 0.01 compared to the lowland control. One-way ANOVA with Tukey's multiple comparisons test was performed for statistical analysis.

in comparison to the people settled in the lowland locations (Lowland Control). Highlanders were further divided into two groups, those who developed PH (PH) and those who did not develop this pulmonary vascular disease (Non-PH). In addition, ELISA was performed in the plasma samples of these three groups, in order to analyze the level of circulating BNP. Due to the technical reasons not all values for all enrolled subjects are available. ApoC1 circulating levels (in µg/mL) were increased in both highlander groups, with being statistically significant in the case of highlanders without PH, in comparison to the lowland controls (**Figure 5A**). TRAIL circulating profile (in pg/mL) did not reveal significant changes among groups, however, there was a trend of reduction in the level of this marker in highlanders with PH, as compared to the people living at low altitude (**Figure 5B**). Further, there was a visible decrease in the circulating levels of FasL (in pg/mL) in both highlander groups, with statistically significant alteration in highlanders with PH, in comparison to the lowland control (**Figure 5C**). Finally, there were no significant changes in the context of BNP (in pg/mL) among different groups (**Figure 5D**). Surprisingly, there was a trend of elevated levels of circulating BNP in highlanders without PH, as compared to other two groups (**Figure 5D**).

# Circulating FasL Levels Negatively Correlate With TRG Values in Kyrgyz Population

Furthermore, we have investigated the correlation between the circulating levels of FasL (pg/mL) in both acute and chronic exposure of Kyrgyz volunteers with TRG values (**Figure 6**).

FIGURE 6 | Correlations between the circulating levels of Fas ligand and tricuspid regurgitant gradient (TRG) during acute and chronic exposure to high altitude. Correlations between the concentrations of circulating Fas ligand (FasL, in pg/mL) and TRG (in mmHg) values in human subjects exposed to acute (A) and chronic (B) effects of high altitude environment are shown. Spearman or Pearson tests were performed for statistical analyses. <sup>∗</sup>p < 0.05; ns, not significant.

using apoptosis assay (A) and 5-bromo-2<sup>0</sup> -deoxyuridine (BrdU) incorporation assays (B–E), respectively. Results are presented as Mean ± SD (n = 4). <sup>∗</sup>p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 compared to the respective controls. \$\$\$\$p < 0.0001 basal/normoxic control compared to the PDGF/hypoxic control, respectively. Ordinary one-way ANOVA with Dunnett's or Tukey's multiple comparisons test was performed for statistical analysis. PDGF, platelet-derived growth factor.

Although there was a slight trend, there was no significant correlation between these two parameters in human subjects exposed acutely to high altitude environment (**Figure 6A**). However, there was a significant negative correlation between circulating levels of FasL and TRG in permanent residents of lowlands and highlands in investigated Kyrgyz population (**Figure 6B**).

# FasL Exerts Pro-apoptotic and Anti-proliferative Effects in Human PASMCs

Finally, we have analyzed the relevant cellular functions of FasL in human PASMCs, such as apoptosis and proliferation (**Figure 7**). We have found that different concentrations

(5 and 25 ng/ml) of FasL significantly increased apoptosis of PASMCs (**Figure 7A**). Furthermore, the same concentrations of this ligand significantly reduced the proliferation of PASMCs under normoxic conditions (**Figure 7B**). In addition, anti-Fas enhanced the cellular proliferation under the basal conditions as well as during the stimulation with PDGF (**Figure 7C**). PASMCs were also exposed to hypoxic conditions (**Figures 7D,E**) and FasL (25 ng/ml) significantly decreased the proliferation of these cells (**Figure 7D**). Finally, anti-Fas resulted in augmentation of the PASMCs proliferation (**Figure 7E**). Clearly, FasL demonstrates pro-apoptotic and anti-proliferative properties in human PASMCs.

### DISCUSSION

In general, the findings of our study revealed that:


Cellular apoptosis is an important, complicate and complex process (Sartorius et al., 2001). Under various stimuli, caspases' may activate and perform the apoptosis (Sartorius et al., 2001). Results of our study demonstrated an augmentation in gene expression of Casp 1 and 3 after the hypoxia incubation in PASMCs. Despite the fact that PH in general is characterized by "apoptosis-resistant" events, the increase in Casp3 expression was found in the lungs of monocrotaline model and the interpretation has to be taken with caution, since there are also some evidences that caspases may even exert anti-apoptotic features (Sartorius et al., 2001; Yang and Widmann, 2001; Le Goff et al., 2012; Hong et al., 2014).

With regard to other apoptotic players, our study indicated an increase in gene expression profiles of ApoC1, FLIP, and survivin in PASMCs exposed to hypoxia, in comparison to the respective normoxic controls. The literature suggested that ApoC1, FLIP, and survivin are generally considered as anti-apoptotic signals in different conditions (Sartorius et al., 2001; Takano et al., 2008; Portt et al., 2011; Fan et al., 2015; Zhang et al., 2015, 2016). Importantly, survivin is well-investigated in the context of hypoxia-induced PH and our data are in line with the literature sources (Fan et al., 2015; Zhang et al., 2015, 2016).

In the case of TRAIL and FasL, we have shown a noticeable increase in the gene expression for both targets during the hypoxic exposure of PASMCs, as compared to their respective normoxic controls. Limited evidences from the literature implicated these two signals in the context of pulmonary vascular disease (Hameed et al., 2012; Saito et al., 2015). Interestingly, although TRAIL is usually considered as proapoptotic mediator, its role in experimental PH pathogenesis has been described (Hameed et al., 2012). Taken together, it is clear that there is a complex play of different apoptotic mediators and this cellular process is indeed dysregulated in the pulmonary vascular cells exposed to hypoxia, however, it goes beyond the simplistic explanation based only on the expression patterns (increased/decreased) or historically known pro- or antiapoptotic features of these signals.

In the cancer field, abnormally regulated apoptosis causes the release of apoptotic products into circulation, and they may represent potential biomarkers (Holdenrieder and Stieber, 2010). Up to date, there is no evidence about the expression profile of circulating apoptotic markers in humans exposed to acute and chronic effects of high altitude. In our study, we have investigated the concentration of ApoC1, TRAIL, and FasL in the plasma of Kyrgyz lowlanders who spent a short period of time in high altitude environment. The same signals were explored in the circulation of the permanent dwellers of high altitude who did or did not develop the HAPH, as well as in the blood of lowlanders. In addition, we have analyzed the circulating BNP, which is usually considered as an important cardiac biomarker, relevant to the PH field (Mellor et al., 2014; Anwar et al., 2016). Briefly, BNP has shown a trend to initially increase with days spent at high altitude in the circulation of Kyrgyz volunteers, as compared to the lowland conditions, and after day 7 there was a tendency to decrease, with a visible reduction upon descent to the lowlands again. Similarly, there was no clear change in the circulating profile of this cardiac marker in highlanders compared to the lowlanders, except non-significant augmentation in highlanders without PH.

ApoC1 circulatory levels were mostly comparable during the acute exposure of subjects to high altitude conditions, while there was some signal of increased values of this marker in the circulation of highlanders, in comparison to lowlanders. Yet, the data do not convincingly suggest the significant alteration of ApoC1 neither during acute nor chronic exposure to high altitude conditions.

Interestingly, the circulating concentrations of TRAIL showed the opposite manner in acute versus chronic exposure to high altitude in human volunteers. On one hand, there was a clear augmentation of this marker with days spent at high altitude in the blood of lowlanders, with significant reduction upon descent to low altitude again. On the other hand, there was noticeable, yet insignificant reduction of the circulating levels of TRAIL in highlanders (particularly those who developed HAPH) in comparison to the lowlanders. Furthermore, the literature suggested the elevated values of this marker in PH patients (Liu et al., 2015). Therefore, further studies are crucially needed to reveal the precise role of this apoptotic signal in both acute and chronic exposure to high altitude.

Finally, clearer and more informative results were obtained with focus on FasL in the circulation of humans in both acute and chronic exposure to high altitude geographic locations. There was a significant reduction of circulating FasL values in lowlanders exposed to high altitude environment for a short

period of time. In addition, the circulating concentrations of this apoptotic marker were significantly decreased in highlanders with PH, as compared to the lowlanders. Interestingly, there were attenuated values of FasL in highlanders without PH also, in comparison to the lowland control, however, the change was not statistically significant. Lastly, there was a negative correlation between the circulatory levels of FasL and TRG, clearly indicated the potential biomarker properties of this apoptotic signal in the case of chronic effects of high altitude and HAPH. In accordance with our findings, FasL serum values were shown to be relevant to the field of pulmonary vascular diseases, as evident in the literature (Akagi et al., 2013; Saito et al., 2015). Importantly, FasL expression was increased in PASMCs derived from patients with idiopathic pulmonary arterial hypertension upon the treatment with prostaglandin I2, which was associated with induction of apoptosis (Akagi et al., 2013). Following this line of thinking, we have also demonstrated in this study that FasL enhanced the apoptotic process in PASMCs. Consistently, pro-apoptotic features of FasL in PASMCs were described in the literature (Zhang et al., 2003; Wang et al., 2010). Furthermore, we have shown that the same concentrations of FasL used for the assessment of apoptosis significantly reduced the proliferation of these pulmonary vascular cells under normoxic conditions. In addition, the blockage of FasL resulted in increased proliferation of PASMCs under the basal conditions as well as during the stimulation with PDGF. Of note, PDGF signaling is a very wellknown player associated with hypoxia and pulmonary vascular remodeling (Schermuly et al., 2005; ten Freyhaus et al., 2011; Veith et al., 2014). Finally, we have demonstrated that FasL exerted anti-proliferative effects on PASMCs exposed to hypoxic conditions, and blockage of FasL resulted in enhancement of the proliferation process.

### REFERENCES


Overall, our study identified for the first time that circulating levels of apoptotic signal FasL are reduced during acute and chronic exposure to high altitude environment. In addition, due to its pro-apoptotic and anti-proliferative functions in the relevant cells of the pulmonary vasculature, dysregulated FasL signal may play an important role in the context of HAPH.

### AUTHOR CONTRIBUTIONS

DK, HG, ASa, and RS conceived, designed, and created the study. DK, SP, AP, ASy, AbM, ArM, MS, KMU, MC, AT, NO, MD, and CV organized the expeditions and performed the field and laboratory measurements. DK, ASa, AbM, HG, ASy, and RS drafted and wrote the manuscript. OP and NW contributed with significant intellectual content. All authors read and approved the manuscript.

### FUNDING

This study was supported by the Cardiovascular Medical Research and Education Fund (CMREF), Universities of Giessen and Marburg Lung Center (UGMLC), Excellence Cluster Cardio-Pulmonary System (ECCPS), and Ministry of Education and Science of the Kyrgyz Republic (Grant No. 0005823).

### ACKNOWLEDGMENTS

We are eternally thankful to all Kyrgyz volunteers for participation in this study.

apoptosis-inducing ligand (TRAIL) reverses experimental pulmonary hypertension. J. Exp. Med. 209, 1919–1935. doi: 10.1084/jem.20112716



**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Kosanovic, Platzek, Petrovic, Sydykov, Maripov, Mamazhakypov, Sartmyrzaeva, Muratali Uulu, Cholponbaeva, Toktosunova, Omurzakova, Duishobaev, Vroom, Pak, Weissmann, Ghofrani, Sarybaev and Schermuly. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

digital media

of impactful research

article's readership