Sixty-two Wistar rats were randomly assigned across two groups: SAL and ELA. In ELA group, rats received porcine pancreatic elastase (2 IU in 0.1 ml of saline solution, Sigma Chemical Co., St. Louis, MO, USA) intratracheally, once weekly for 4 weeks. SAL animals received saline solution only via the same route and dosage schedule. Before each intratracheal instillation, animals were premedicated with intraperitoneal diazepam (10 mg/kg, Compaz®, Cristália, Itapira, SP, Brazil) and anesthetized with 1.5–2.0% isoflurane (Cristália, SP, Brazil) by mask.

Evaluations were performed 1, 3, 5, or 8 weeks after the last intratracheal instillation of saline or porcine pancreatic elastase (Figure S1). Briefly, animals were anesthetized with intraperitoneal diazepam (10 mg/kg, Compaz®, Cristália, Itapira, SP, Brazil), ketamine (50–100 mg/kg, Ketamin-S+®, Cristália, Itapira, SP, Brazil), and midazolam (2 mg/kg, Dormicum®, União Química, São Paulo, SP, Brazil), tracheotomized, paralyzed with pancuronium bromide (2 mg/kg i.v., Cristália, Itapira, SP, Brazil), and mechanically ventilated (Servo-i, MAQUET, Solna, Sweden) in VCV mode, using the following parameters: tidal volume (V_T) = 6 mL/kg, respiratory rate (RR) = 80 breaths/min, fraction of inspired oxygen (FiO₂) = 0.4, and positive end-expiratory pressure (PEEP) = 3 cmH₂O. During ventilation, intravenous fluids were administered to maintain a mean arterial pressure (MAP) ≥70 mmHg and midazolam plus ketamine were administered to maintain sedation. Fifty-six rats were used to evaluate lung mechanics and morphometry, end-expiratory lung volume (EELV), biological markers associated with inflammation (interleukin [IL]-6, cytokine-induced neutrophil chemoattractant [CINC]−1), alveolar stretch (amphiregulin), type II epithelial cell mechanotransduction (surfactant protein [SP]-D), and endothelial cell damage (angiopoietin [Ang]-2 and vascular endothelial growth factor [VEGF]), as well as echocardiographic parameters (n = 7 at each time point in SAL and ELA groups). In six animals (n = 3/group), a computed tomography scan of the lungs was performed at 5 weeks.

After identifying the time point of greatest cardiorespiratory impairment, an additional 32 Wistar rats were randomized into the SAL and ELA groups, as previously described. Animals were premedicated and anesthetized as described for the first protocol. An intravenous catheter (Jelco 24G) was inserted into the tail vein for continuous infusion of midazolam (2 mg/kg/h), ketamine (50 mg/kg/h), and Ringer's lactate (7 mL/kg/h, B. Braun, Crissier, Switzerland). Anesthetized animals were kept in the dorsal recumbent position and tracheotomized via a midline neck incision after subcutaneous injection of 2% lidocaine (Cristália, Itapira, SP, Brazil). The right internal carotid artery was cannulated (18 G, Arrow International, USA) for blood sampling and MAP measurement. Heart rate (HR), MAP, and rectal temperature were continuously recorded (Networked Multiparameter Veterinary Monitor LifeWindow 6000 V, Digicare Animal Health, Florida, USA). Body temperature was maintained at 37.5 ± 1°C using a heating bed. Gelafundin® (B. Braun, São Gonçalo, RJ, Brazil) was administered intravenously in 0.5-mL increments to keep MAP ≥ 70 mmHg. Animals were paralyzed by intravenous administration of pancuronium bromide (2 mg/kg, Cristália, Itapira, SP, Brazil) and mechanically ventilated (Inspira, Harvard Apparatus, Holliston, Massachusetts, USA) in VCV mode with V_T = 6 mL/kg, RR adjusted to maintain arterial pHa in the 7.35–7.45 range, Fio₂ = 0.4, and zero end-expiratory pressure (ZEEP). After hemodynamic stabilization, respiratory system mechanics, and arterial blood gases (Radiometer ABL80 FLEX, Copenhagen NV, Denmark) were measured (Baseline ZEEP). PEEP was then increased to 3 cmH₂O and, after 5 min, respiratory system mechanics, arterial blood gases, and echocardiographic parameters were analyzed (Baseline PEEP). Following this step, SAL and ELA animals were randomized to VV or VCV (n = 8/group). VV was applied on a breath-to-breath basis as a sequence of randomly generated V_T values (Gaussian distribution, n = 1200; mean V_T = 6 mL/kg), with a 30% coefficient of variation (nVentInspira, Dresden, Germany), for 2 h (Huhle et al., 2014). At the end of the experiment, echocardiographic parameters, respiratory system mechanics, and arterial blood gases were analyzed at each ventilator setting. Animals were then euthanized via overdose of intravenous sodium thiopental (50 mg/kg, Cristália, Itapira, SP, Brazil) and their lungs extracted at PEEP = 3 cmH₂O for measurement of EELV, as well as lung histological and molecular biology analyses (Figure S2).

Echocardiographic measurements were obtained at baseline and 1, 3, 5, and 8 weeks after the last intratracheal instillation of SAL or PPE. Right ventricular (RV) end-diastolic area, pulmonary artery acceleration time (PAT), pulmonary artery ejection time (PET), and their ratio (PAT/PET) were used as indirect indexes of pulmonary arterial hypertension.

Echocardiographic parameters were measured before and 2 h after mechanical ventilation (VV or VCV) in the SAL and ELA groups. In addition to the aforementioned parameters associated with right ventricular impairment, left ventricle diastolic area, ejection fraction, and fractional shortening were computed, as mechanical ventilation may also affect left systolic function.

Animals were placed in the dorsal recumbent position and the precordial region was shaved. Transthoracic echocardiography was performed by an expert (NR), using a 10-MHz probe (Esaote model, CarisPlus, Firenze, Italy). Images were obtained from the subcostal and parasternal views. Short-axis two-dimensional views of the left and right ventricles were acquired at the level of the left ventricular papillary muscles to measure left and right ventricular end-diastolic area (LV and RV area, respectively). The left ventricular ejection fraction and fractional shortening were calculated in one-dimensional mode analysis of the left ventricle guided by the parasternal short-axis view. Pulsed-wave Doppler was used to measure PAT, PET, and the PAT/PET ratio. The diameter of the right atrium and inferior vena cava were assessed from the subcostal view. Measurements were obtained in accordance with American Society of Echocardiography Guidelines (Thibault et al., 2010; Lang et al., 2015).

Computed tomography (CT) scans were performed to evaluate the presence of emphysematous areas. CT was performed in SAL and ELA animals at the time point of greatest cardiorespiratory impairment (5 weeks after the last elastase administration) with an Optima 560 PET/CT scanner (GE Healthcare, Boston, USA). The acquisition protocol was based on helical CT with axial slices of 0.625 mm (16 × 0.625 mm) thickness and 48 images, with a beam collimation of 10.0 mm and a DFOV of 10 cm. The X-ray tube was set to 120 kV and 80 mA. The total time for each scan was 12 s. Hounsfield units (HU) were analyzed in both lungs, at the bifurcation of pulmonary arteries, in the Onis 2.5 software environment (DigitalCore, Co. Ltd., Tokyo, Japan).

Airflow, volume, and airway and esophageal pressures were measured (Riva et al., 2008), and lung mechanics were analyzed by the end-inflation occlusion method (Bates et al., 1985). Static lung elastance (Est,L) was calculated by dividing the difference between respiratory system and chest wall elastic pressure (Pel,L) by V_T. Since EELV may change due to emphysema, Est,L was normalized by EELV, which is equal to specific lung elastance (EL,spec).

Airflow, volume, and airway pressure were continuously recorded. Elastance (E,_RS) and resistance (R,_RS) were calculated based on the equation of motion (Uhlig et al., 2014). Volume-independent elastance (E1,_RS) and volume-dependent elastance (E2,_RS) were calculated on a cycle-by-cycle basis (Carvalho et al., 2013). The partition model of respiratory system elastance allows calculation of the E,_RS non-linearity index (%E2), which quantifies the concavity of the dynamic pressure-volume (PV) curve, thus enabling identification of tidal recruitment/overdistension, which can be present in this emphysema model.

All parameters were recorded with a computer running custom software written in LabVIEW® (National Instruments; Austin, Texas, USA) (Silva et al., 2013). All signals were amplified in a three-channel signal conditioner (TAM-D HSE Plugsys Transducers Amplifiers, Module Type 705/2, Harvard Apparatus, Holliston, Massachusetts, USA) and sampled at 200 Hz with a 12-bit analog-to-digital converter (National Instruments; Austin, Texas, USA).

EELV was determined as previously described (Scherle, 1970) to identify the resulting lung volume at end-expiration, as emphysema is characterized by destruction of alveolar septa with hyperinflated areas. Briefly, a jar containing sufficient saline solution with a surplus weight submerged was placed on a common laboratory scale, which was subsequently tared to zero. The lungs were fixed to a laboratory stand by means of a thread with the surplus weight and completely submerged in the saline solution. The liquid displaced by the submerged lungs adds correspondingly to the weight on the scale. Because the specific gravity of saline differs no more than 2–3% from 1 g/cm³, the volume of the organ may be expressed directly by the weight gain registered on the scale.

Lung morphometry was analyzed in SAL and ELA animals to characterize the model of emphysema.

Lung morphometry was evaluated in animals ventilated with VCV and VV, as well as in non-ventilated (NV) animals (SAL and ELA), to analyze the impact of different ventilator strategies on lung parenchyma.

Morphometric analysis was performed in lungs excised at end-expiration (PEEP = 3 cmH₂O). Immediately after removal, the left lung was flash-frozen by immersion in liquid nitrogen, fixed with Carnoy's solution, and paraffin-embedded. Sections (4 μm thick) were cut and stained with hematoxylin-eosin. Investigators (CW and VLC) blinded to the origin of the material performed the microscopic examination. Morphometric analysis was done using an integrating eyepiece with a coherent system made of a 100-point grid consisting of 50 lines of known length, coupled to a conventional light microscope (Axioplan, Zeiss, Oberkochen, Germany). The volume fraction of collapsed and normal pulmonary areas and the fraction of the lung occupied by large-volume gas-exchanging air spaces (hyperinflated structures with a morphology distinct from that of alveoli and wider than 120 μm) were determined by the point-counting technique (Weibel, 1990), at a magnification of × 200 across 10 random, noncoincident microscopic fields. Briefly, points falling on collapsed or normal pulmonary areas or hyperinflated alveoli were counted and divided by the total number of points in each microscopic field. Lung tissue distortion was assessed by measuring the mean linear intercept between alveolar walls (Lm) at a magnification of × 400 (Weibel, 1990). In the first protocol, collagen and elastic fibers (stained with the Picrosirius-polarization method and Weigert's resorcin fuchsin modified with oxidation, respectively) were quantified in alveolar septa at × 400 magnification (Antunes et al., 2014). The area occupied by fibers was determined by digital densitometric recognition in the Image-Pro Plus 7.1 for Windows software environment (Media Cybernetics, Silver Spring, MD, USA) and divided by the area of each studied septum. Results were expressed as the fractional area occupied by elastic and collagen fibers in the alveolar septa. Bronchi and blood vessels were excluded from the measurements.

Quantitative real-time reverse transcription polymerase chain reaction (RT-PCR) was performed to assess IL-6, CINC-1, amphiregulin, SP-D, Ang-2, and VEGF. In the first protocol, these analyses were performed at the time point of greatest cardiorespiratory impairment. Central slices of the right lung were cut, collected in cryotubes, flash-frozen by immersion in liquid nitrogen, and stored at −80°C. Total RNA was extracted from frozen tissues using the RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's recommendations. RNA concentration was measured by spectrophotometry in a Nanodrop ND-1000 system (ThermoScientific, Wilmington, DE, USA). First-strand cDNA was synthesized from total RNA using a Quantitec reverse transcription kit (Qiagen, Hilden, Germany). The primers used are described in the online supplement (Table S1). Relative mRNA levels were measured with a SYBR green detection system in ABI 7500 real-time PCR (Applied Biosystems, Foster City, California, USA). Samples were run in triplicate. For each sample, the expression of each gene was normalized to the acidic ribosomal phosphoprotein P0 (36B4) housekeeping gene and expressed as fold changes relative to SAL animals (first protocol) or to non-ventilated (NV) animals (SAL and ELA—second protocol), using the 2^−ΔΔCt method, where ΔCt = Ct_{reference gene} − Ct_{target gene}.

The fact that VV, but not VCV, decreased dynamic respiratory system elastance and non-linearity index of the dynamic respiratory system elastance over time is likely explained by lung recruitment (Kano et al., 1994; Bersten, 1998). Alveolar collapse and hyperinflation were lower in VV compared to VCV, which may have led to better distribution of regional ventilation and of stress and strain across the lungs. Our observation that VV did not improve gas exchange compared to VCV might be explained by differences in regional perfusion (Pelosi and de Abreu, 2015).

Variable ventilation reduced alveolar collapse and hyperinflation only in ELA thus resulting in decreased dynamic respiratory system elastance and volume-dependent elastance. We hypothesize that VV promoted alveolar recruitment by periodic higher inspiratory pressures, reducing the traction of alveolar walls by collapsed regions and thus mitigating hyperinflation. In short, VV resulted in better distribution of regional ventilation, stress, and strain across the lungs.

In SAL animals, after VV, the fraction area of normal alveoli increased, thus resulting in a reduction of pulmonary vascular resistance, as shown by the increase in the ratio between pulmonary artery systolic acceleration time and pulmonary artery systolic ejection time. This is expected in a normal capillary density scenario (Overholser et al., 1994); however, similar behavior was not observed in ELA animals. After VV, even though the fraction area of normal alveoli increased, pulmonary vascular resistance may have increased as well, thus leading to enlargement of right ventricular end-diastolic area. The relationship between pulmonary vascular resistance and lung volume is “U-shaped,” where the lowest pulmonary vascular resistance values are closer to functional residual capacity (Simmons et al., 1961). In elastase-induced emphysema, arterial remodeling occurs, with increased elastolysis in the vessel wall (Wills et al., 1996). Furthermore, hyperinflated and collapsed alveoli may not receive adequate blood flow, due to vascular compression and hypoxic vasoconstriction, respectively; thus, blood flow could be driven toward normal alveoli. After VV, there was likely a better airflow distribution across the airways, which may have led to improved lung volume distribution, since there was no change in total lung volume as detected by EELV measurements. We hypothesized that, after lung volume redistribution, a reduction in alveolar hyperinflation and recruitment of alveolar collapse occurred, alongside a slight decrease in size of normal alveoli, resulting in increased pulmonary vascular resistance, as indirectly measured by echocardiographic analysis (ratio between pulmonary artery systolic acceleration time and pulmonary artery systolic ejection time, and right ventricular end-diastolic area). This effect can move the point from the lowest pulmonary vascular resistance at functional residual capacity to the right side of the curve, thus increasing the pulmonary vascular resistance by alveolar vessel distortion. Additionally, during VV, the higher frequency of increased tidal volume and plateau pressure in a fixed time course may lead to increased pulmonary vascular resistance, especially in the presence of vessel wall changes.

In SAL animals, VCV, but not VV, led to increased mRNA expression of interleukin-6, amphiregulin, and vascular endothelial growth factor compared to NV. In this line, an in vitro study reported that exposure of normal cultured epithelial cells to increased stress and strain results in cell membrane disruption and death (Tschumperlin and Margulies, 1999). In non-injured lungs, VV reduced stress and strain, preventing inflammation, alveolar stretch, and endothelial damage. In ELA, no difference was observed between VV and VCV regarding interleukin-6 mRNA expression, suggesting that mechanical ventilation per se induces activation of lung inflammation. In the emphysematous lung, which tends to be more structurally fragile (Suki et al., 2013), moderate and high tidal volumes (based on the normal distribution with a 30% coefficient of variation) during VV could increase lung inflammation. On the other hand, VV resulted in beneficial biological effects through an increase in surfactant protein-D expression, which is in agreement with previous experimental findings in experimental ARDS (Thammanomai et al., 2013). We may speculate that application of variable stretch to alveolar epithelial cells led to mechanotransduction toward intracellular signaling and cytoplasmic reorganization in an attempt to restore surfactant production (Roan and Waters, 2011).