Hyperbaric Oxygen Therapy Represses the Warburg Effect and Epithelial–Mesenchymal Transition in Hypoxic NSCLC Cells via the HIF-1α/PFKP Axis

Background Tumor cells initiate hypoxia-induced mechanisms to fuel cell proliferation, invasion, and metastasis, largely mediated by low O2-responsive Hypoxia-Inducible Factor 1 Alpha (HIF-1α). Therefore, hyperbaric oxygen therapy (HBO) is now being studied in cancer patients, but its impact upon non-small-cell lung cancer (NSCLC) cell metabolism remains uncharacterized. Methods We employed the NSCLC cell lines A549 and H1299 for in vitro studies. Glucose uptake, pyruvate, lactate, and adenosine triphosphate (ATP) assays were used to assess aerobic glycolysis (Warburg effect). A quantitative glycolytic flux model was used to analyze the flux contributions of HIF-1α-induced glucose metabolism genes. We used a Lewis lung carcinoma (LLC) murine model to measure lung tumorigenesis in C57BL/6J mice. Results HBO suppressed hypoxia-induced HIF-1α expression and downstream HIF-1α signaling in NSCLC cells. One HIF-1α-induced glucose metabolism gene—Phosphofructokinase, Platelet (PFKP)—most profoundly enhanced glycolytic flux under both low- and high-glucose conditions. HBO suppressed hypoxia-induced PFKP transactivation and gene expression via HIF-1α downregulation. HBO’s suppression of the Warburg effect, suppression of hyperproliferation, and suppression of epithelial-to-mesenchymal transition (EMT) in hypoxic NSCLC cell lines is mediated by the HIF-1α/PFKP axis. In vivo, HBO therapy inhibited murine LLC lung tumor growth in a Pfkp-dependent manner. Conclusions HBO’s repression of the Warburg effect, repression of hyperproliferation, and repression of EMT in hypoxic NSCLC cells is dependent upon HIF-1α downregulation. HIF-1α’s target gene PFKP functions as a central mediator of HBO’s effects in hypoxic NSCLC cells and may represent a metabolic vulnerability in NSCLC tumors.


In vitro normobaric hypoxia and HBO therapy
Normobaric hypoxia was administered as previously described . Briefly, cells (1.0 × 10 4 cells/cm 2 ) were seeded under normal conditions and then incubated in a controlled 1% O2 atmosphere in a hypoxia chamber (ASTEC, Japan) for 72 hours. For HBO exposure, hypoxic cells were positioned within a 27-l hyperbaric chamber (OXYCOM 250 ARC, HYPCOMOY, Finland). The hyperbaric chamber was flushed with pure O2 gas at ambient pressure for 15 min. The pressure was gradually increased from 1.0 to 3.0 bar for 10 min and was then kept stable at 3.0 bar for 90 min. The chamber was flushed with pure O2 gas for 5 min every 10 min. After the 90-min HBO period, the chamber was decompressed to 1.0 bar over 15 min.

Glucose uptake
Glucose uptake was measured in cells 72 hours post-transfection. Cells were plated in fresh complete media for 20 minutes before the glucose analog 2-NBDG (5 µM, Thermo Fisher Scientific) was added and left to incubate with the cells for an additional 15 minutes. After incubation, a microplate reader (excitation wavelength: 488 nm) was used to measure the fluorescence intensity of each sample.

Lactate secretion and intracellular pyruvate assays
The Glycolysis Cell Assay Kit (Cayman Chemicals) --a colorimetric method for detecting L-lactate secreted by cultured cells ---was used to measure supernatant lactate levels as per the manufacturer's protocol.
For the intracellular pyruvate assays, cell lysates (5 × 10 5 cells) were incubated in pyruvate assay buffer (100 µl) from the Pyruvate Fluorometric Kit (BioVision) as per the manufacturer's protocol. Following incubation, lysates were centrifuged at 10000 g at 4°C for 10 minutes to remove insoluble debris. Supernatants were collected (50 µl) and loaded into a 96-well plate. Reaction mix was added (50 µl) and samples were left to incubated at room temperature for 30 minutes in the dark before a microplate reader (excitation/emission wavelengths: 535/590 nm) was used to measure the fluorescence intensity of each sample.

Intracellular adenosine triphosphate (ATP) measurements
Cells were isolated, then boiled in a 100°C water bath in pre-boiled ATP buffer (200 µl) for 2 minutes for the preparation of cell lysates. The lysates were then centrifuged for 2 minutes at 4°C before supernatants were collected and transferred to a new tube. BCA protein assays were used to determine protein levels, then the ATP Bioluminescence Assay Kit (Roche) was used as per the manufacturer's protocol. Intracellular ATP levels were detected using a microplate reader. Total protein levels were used to normalize ATP levels. For some experiments, cells were pre-treated with DMSO vehicle or oligomycin A (10 μg/ml in DMSO) for 60 min as previously described (Zhu et al., 2019).
Venn analysis to identify HIF-1α-induced glucose metabolism genes in hypoxic NSCLC cells We performed a Venn analysis to specifically identify HIF-1α-induced, upregulated genes in hypoxic NSCLC cells that participate in glucose metabolism. HIF-1 gene targets were derived from Schödel et al.'s high-resolution genome-wide mapping of HIF-1 binding sites by chromatin immunoprecipitation linked to high throughput sequencing (ChIP-seq) (Schödel et al., 2011). Genes significantly upregulated by hypoxia in both independently cloned into pcDNA4/HisMaxC plasmids as described above. Cells were plated in 12-well plates and transfected with 400 ng of each pcDNA plasmid for 24 hours. To confirm gene overexpression, cells were subjected to immunoblotting at 24 hours post-transfection as described below.
Glucose and lactate measurements were conducted in six-well plates between 24 and 30 hours post-pcDNA plasmid transfection. At 24 hours post-transfection, cells were washed thrice with 1× PBS (2 ml) and then we added 1-ml volume of fresh DMEM containing 10% dialyzed FBS, glucose (5 mM or 25 mM), Lglutamine (0.584 g/l), phenol red (0.015 g/l), and sodium bicarbonate (3.7 g/l). At 30 hours post-transfection, a 300-ml supernatant volume was analyzed for glucose and lactate levels by a biochemical analyzer (HITACHI 7080, Japan) Changes in glucose and lactate were normalized to the calculated mean packed cell volume where f J E is the simple fold-change in lactate secretion (f J ) produced by enzyme (E) overexpression (i.e., lactate secretion in enzyme E-overexpressing cells / lactate secretion in GFP-expressing control cells).

RNA extraction and qPCR
Trizol (Thermo Fisher Scientific) was used to isolated total RNA from cells as per the manufacturer's protocol.
An EasyScript cDNA Synthesis Kit (TransGen, China) was used to generate cDNA from RNA. GoTaq qPCR Mix (Promega) was used to perform qPCR on the CFX96 Detection System (BioRad). All PCR primers were purchased from Origene; the primer sequences are available on the Origene website (www.origene.com). We used human peptidylprolyl isomerase A (PPIA) as the housekeeping gene, as it has been shown to be stable under hypoxic conditions in two human cancer cell lines (Albuquerque et al., 2018). Gene expression data was generated using the 2 -ΔΔCt method.

Western blotting
Cell or tissue lysates were generated using RIPA buffer. BCA Protein Assays (BeyoTime Biotechnology, China) were used to quantify to amount of protein in each sample. Samples were then subjected to SDS-PAGE for separation before being heat transferred into PVDF membranes. Membranes were blocked at room temperature for 1 hour using skim milk (5%) in TBS-T. Next, membranes were washed and incubated at 4°C overnight with primary antibody, before being washed and incubated at room temperature for 1 hour the next

PFKP luciferase reporter assays
The PFKP luciferase reporter (pPFKP-Luc) was used for PFPK transactivation assays as previously reported (Liu et al., 2016). In brief, HEK293T cells were plated into 24 well plates, transfected with pPFKP-Luc and the internal control pTK-Renilla using VigoFect reagent (Vigorous Biotech, Beijing, China), and then cultured for 28 hours. Following hypoxia and/or HBO, luciferase assays were conducted in HEK293T cells using the Dual-Luciferase Reporter Assay System (Promega). Renilla luciferase activity was used to normalize pPFKP-Luc activity.

5-ethynyl-2′-deoxyuridine (EdU) cell proliferation assay
Cells were plated into six-well plates for 12 hours. EdU (20 µM) was added to the growth media and left to incubate for an additional three hours. Following incubation, cells were isolated and paraformaldehyde (4%) fixed for 20 minutes. Then, cells were permeabilized with Triton X-100 (0.5%) at room temperature for 10 minutes. The dye mix solution was added for a final 10 minute incubation at room temperature in the dark. Finally, the samples were ran on the flow cytometer (Accuri, BD Biosciences) to determine the percentage of EdU +ve cells.

Migration and invasion assays
Migration was assessed via Transwell assays (Corning Costar). Cells incubated in either normoxic or hypoxic conditions for 36 hours were subsequently serum-starved for a further 12 hours. Next, cells were placed in the top chamber in serum-free media, with 10% FBS media in the bottom chamber before being incubated for a further 24 hours. Cells on the upper membrane, which had not migrated, were collected. Cells on the bottom side of the membrane were paraformaldehyde (4%) fixed, stained in crystal violet (0.1%), and quantified via microscopy.
Cells were cultured in serum-free media in the top chamber of a Transwell assay covered with Matrigel.
The bottom chamber was filled with 10% FBS media, and Transwells were incubated for 24 hours. Next, EdU was added and the cells were left for an additional 4 hours before membranes were removed and stained with the EdU kit (Invitrogen). Cell counting was then performed via microscopy.

Floating cell detachment assays
Floating cell detachment assays were performed as previously described . Briefly, cells were seeded into a 24-well plate and left to incubate for 12 hours at 37°C. Floating cells that did not attach were isolated by 250 g centrifugation for 5 min at room temperature, resuspended, stained with 0.4% trypan blue, and counted by hemocytometry with a bright-field microscope (200× magnification). Cells that had attached were counted following trypsinization. Detachment was calculated as the ratio of floating cells/attached cells.

In vivo HBO murine model of NSCLC
Male C57BL/6J mice (6-7 weeks of age) were purchased from HuaFuKang Bioscience (China) and housed in plastic cages under a 12-h/12-h light/dark cycle with controlled temperature/humidity for one week prior to experimentation. Mice were fed commercial standard chow and water ad libitum. All efforts were made to minimize animal suffering, and all mice were euthanized by cervical dislocation.
Stably-infected LLC cells (5.0 × 10 5 cells in PBS) were tail vein-injected into male C57BL/6J mice (7-8 weeks of age) to establish the NSCLC model. HBO therapy was performed as previously described (Yttersian Sletta et al., 2017) with minor modifications. Briefly, mice were positioned within a 27-l hyperbaric chamber (OXYCOM 250 ARC, HYPCOMOY, Finland). The hyperbaric chamber was flushed with pure O2 gas at ambient pressure for 15 min. The pressure was gradually increased from 1.0 to 2.5 bar for 10 min and was then kept stable at 2.5 bar for 90 min. The chamber was flushed with pure O2 gas for 5 min every 10 min. After the 90-min HBO period, the chamber was decompressed to 1.0 bar over 15 min. The mice underwent this HBO procedure on days 1, 4, 7, 10 and 13 post-LLC cell injection. Controls were exposed to normoxia at 1.0 bar under otherwise identical conditions. Mice were monitored during the course of the experiment. At day 14 post-LLC cell injection, mice were humanely sacrificed. The lungs were excised for analysis.

Lung tumor analysis
The gross morphology of the excised lungs and the number of lung tumor nodules were assessed. Tumor sizes were measured with calipers; total tumor volumes were calculated as follows: 0.5236 × (d1 × d2 × d3), where the d1-3 parameters represent the three orthogonal diameters. The lungs were inflated and 10% formalinfixed at 25 cm H2O pressure before pathohistological staining. Hematoxylin and eosin (H&E)-stained slides of fixed tumor tissues were scanned with an Aperio Scan Scope XT Slide Scanner.

Immunofluorescence (IF) staining
The fixed lung tumor samples were paraffin-embedded before being sliced into thin sections for IF staining. Sections were heated in citrate buffer (pH 6.0, 0.01 M) in a water bath (95°C) for 10 minutes, then treated with hydrogen peroxide (30%). Sections were then blocked with BSA (10% in PBA) for 60 minutes 8 before incubation with Alexa Fluor® 488-conjugated anti-Ki-67 (#11882, CST) overnight. Hoeschst was used to counterstain nuclei before sections were mounted using mounting media (Thermo Fisher Scientific). The Evos Auto Cell Imager (Thermo Fisher Scientific) was used to visualize Ki-67 +ve cells. ImagePro software was used for quantification.

Immunohistochemistry (IHC) staining
The fixed lung tumor samples were paraffin-embedded and sectioned as described above. Sections were blocked with BSA (10%), then stained overnight with the anti-HIF-1α, anti-PFPK, anti-p-β-catenin(S552) primary antibodies described above. Next, sections were stained at room temperature for 30 minutes with a secondary horseradish peroxidase (HRP)-conjugated antibody. Finally, DAB and hematoxylin were used as a counterstain. The Evos Auto Cell Imager (Thermo Fisher Scientific) was used to visualize staining.     The seven glucose metabolism genes identified in Figure 1 were individually overexpressed in     Figure 4G for p-β-catenin(S552), E-