Pulmonary artery smooth muscle cell hyperproliferation and metabolic shift triggered by pulmonary overcirculation

Lamb model.

As described in detail previously (38), an 8.0-mm Gore-tex vascular graft, ∼2 mm in length, was anastomosed between the ascending aorta and main pulmonary artery in anesthetized late-gestation fetal lambs (137–141 days gestation, term = 145 days) from three mixed-breed Western ewes. Four weeks after spontaneous delivery, normal, age-matched controls (n = 3) or shunt lambs (n = 3) were anesthetized, mechanically ventilated, and instrumented to continuously measure hemodynamics and harvest tissue. Animals' vital signs, including core temperature, were monitored throughout the study, and they were given intravenous fluids and prophylactic antibiotics per protocol. At the end of each protocol, all lambs were euthanized with a lethal injection of pentobarbital sodium followed by bilateral thoracotomy as described in the NIH Guidelines for the Care and Use of Laboratory Animals. All protocols and procedures were approved by the Committees on Animal Research of the University of California, San Francisco and University of California, Davis.

Cell isolation and culture.

Primary smooth muscle cells from shunt (n = 3) and control (n = 3) lambs were isolated from sections of the proximal (main, left, and right) pulmonary arteries by explant method as previously published (19). Identity was confirmed as PASMC by immunostaining (>99% positive) with established smooth muscle markers (40) SMC actin, smooth muscle myosin heavy chain, and calponin as described previously (26). The cells were maintained in culture with a medium of low-glucose DMEM (1 g/l) with sodium pyruvate (110 mg/l) and glutamine supplemented with 10% fetal bovine serum (Lonza), penicillin/streptomycin, and fungizone. They were grown in an incubator at 5% CO₂ and 37°C and passaged between 70–90% confluence. All experiments were performed on passage-matched (within ± 1 passage) primary cells between passages 4–8.

Immunofluorescence.

PASMCs from shunt and control animals were grown on cover glass to 80% confluence, fixed with 4% paraformaldehyde (Fisher) for 30 min, and permeabilized with 0.1% Triton X-100 (Merck KGaA) for 5 min. After that, cells were incubated with anti-calponin IgG (1:400 rabbit, Abcam) primary antibody for 30 min, conjugated with a goat anti-rabbit IgG TRITC (1:200, ProteinTech Group) secondary antibody, and stained with 4′,6-diamidino-2-phenylindole (DAPI, 1:10.000, Invitrogen). As negative controls, the primary antibody was replaced by bovine serum albumin (BSA, Serva Electrophoresis GmbH). Finally, cells were mounted on histological slices using mounting medium (Cell Signaling Technologies) and imaged by a Leica SP5 inverted confocal microscope (Zeiss).

Baseline proliferation and drug treatment assays.

SMCs from shunt and control animals were trypsinized with 0.25% trypsin solution, washed with PBS and counted using a desktop coulter based cell counter (Moxi Z, Orflo). Uniform cell counts were then seeded into 24-well cell culture plates with standard cell culture medium (as above). Three distinct clonal cell lines from shunts and three from controls were plated in replicates of 8 per line for each experimental growth time point. At sequential 24-h time points after seeding, cells from each line were trypsinized and counted (as per above) out to 72 h. This procedure was repeated in triplicate. Total cell counts were averaged for all shunt cell lines and all control cell lines and plotted along with error bars corresponding to 1 SD in both the positive and negative direction. Total cell counts for all control and all shunt cell lines at each experimental time point were compared with two-tailed Student's t-test. To further control for differences in starting cell number due to natural variability and any differences in efficiency or time of adherence after trypsinization, all cell counts were log transformed and linear regression analysis was performed during the time period of 24–72 h representing the log growth phase. The slopes of the lines, corresponding to the actual rate of growth, were then compared using multiple regression/correlation analysis as outlined previously (10). Cell doubling times were calculated by use of least-squares fitting exponential.

For treatment with dichloroacetate (DCA) and dimethyl malate (DMM), and VAS 3947, cells were counted as above and seeded with uniform density into 24 well culture plates using standard growth media. At 24 h after cells had been allowed adequate time to adhere, media was changed to standard growth media with the drug and dose of interest (experimental) or standard media with an equivalent volume of drug solvent (control). For dichloroacetate experiments, 0.5, 1, or 5 mM DCA (Sigma Aldrich 634522) were used for experimental groups. For dimethyl malate experiments, 5 mM DMM (Sigma Aldrich 374318) was used for experimental groups. For VAS 3947, 10 μM (CalBiochem 532336) was used for experimental groups. At 72 h cells from all groups were counted in replicates of 8. All counts in treatment groups were normalized to control 72 h growth averages for comparison. Comparison between shunt and control growth was performed using unpaired, two-tailed Student's t-test.

Extracellular flux analysis.

For our experiments we used the Seahorse XF-24 Extracellular Flux Analyzer (Seahorse Bioscience, Copenhagen, Denmark) to compare shunt and control PASMCs. Per manufacturer instructions, initial optimization assays were conducted using both shunt and control cell lines to establish optimum cellular density for the assay. Experimental assays were then performed at optimal cell density (∼10,000–30,000 cells/well). Basal metabolism comparisons were performed as per manufacturer recommendations. Media for all extracellular flux assays was DMEM with sodium pyruvate (110 mg/l), glutamine, and glucose (1 g/l) without NaHCO₃ and without phenol red (Sigma). After reconstitution from powder, media was pH adjusted to 7.40 and sterile filtered. Cells from each line were plated in replicates of 5. Immediately following basal metabolism assays, cells from all wells were counted and averaged per experimental group. OCR and ECAR data for each experimental group were normalized to average cell number for analysis. Given that there is variation between assay plates, to perform comparison across multiple experiments, all values were normalized to a single cell line included in all assays.

OCR measurements with DMM addition were performed using a modification of the basal metabolism measurements. Cells and media were prepared as above. Cells from each line were plated in replicates of 5 for 2 shunt cell lines and 2 control cell lines. Following measurement of basal OCR, DMM was injected into the assay medium to reach a total final concentration of 5 mM, and measurements of OCR were repeated in triplicate following a 15 min mixing cycle. OCR with DMM was subsequently normalized to OCR under basal conditions within each group for comparison.

Protein isolation.

For whole cell protein extraction, PASMCs were seeded and grown on 10-cm culture plates to 70–80% confluence. Whole cell lysates were prepared on ice using urea lysis buffer [8 M urea, 10% glycerol, 5 mM DTT, 10 mM Tris-HCl pH 6.8, 1% SDS, and 1 × Proteinase Inhibitor Cocktail (Roche Diagnostics)]. Cells were scraped and sonicated and protein lysates were subsequently stored at −80°C. Protein was quantified using the Nanodrop microvolume spectrophotometer (Thermo Scientific).

Western blots.

For Western blot analysis of HIF-1α and PDHK, 30 μg of total PASMC protein lysate was loaded for each sample onto an 8–16% SDS-PAGE Gel (Lonza) and transferred to a PVDF membrane using the semi-dry transfer method. The membranes were then blocked using Li-COR blocking buffer (LI-COR Biosciences) with 0.1% Tween and probed with primary antibodies directed against HIF-1α (Cayman Chemical, 10006421), PDHK 1 (Cell Signaling 3820), and α-Tubulin (Sigma Aldrich T9026), which served as a loading control. This was followed by incubation with IRDye 800CW goat anti-rabbit and IRDye 680 goat anti-mouse (LI-COR Biosciences) at a dilution of 1:5,000. Blots were visualized using the Odyssey Imaging System (LI-COR) and quantitated with Image Studio software in accordance with manufacturer's protocol.

For Western Blot analysis of Nox 2, Nox 4, p47 phox, and Rac-1, 20 μg of total PASMC protein lysate was loaded for each sample onto a 10% SDS-PAGE and transferred to a PVDF membrane using the wet transfer method. The membranes were then blocked with 5% nonfat dried milk in 130 mM NaCl and 25 mM Tris (TBS, pH 7.5) and probed with primary antibodies against Gp91-phox (H-60) (NOX2) (Santa Cruz Biotechnology, Sc-20782), NOX4 (Abcam, ab109225), p47-phox (Cell Signaling, 4301), RAC-1 (Sigma, GW22212F), and β-actin (Abcam), which served as a loading control. This was followed by incubation with the appropriate horseradish peroxidase-conjugated secondary antibodies. Chemiluminescence was then used to detect bands (SuperSignal West Pico Chemiluminescent Substrate kit, Pierce Biotechnology, Rockford, IL). Densitometry was performed using a public domain Java image processing program, ImageJ (NIH Image).

Metabolomics.

For metabolite quantification, 1 line each of shunt and control cells were selected and grown in 10-cm culture plates in replicates of five each (n = 5). Once cells reached 75–80% confluence, media was aspirated and cells were briefly rinsed with ∼10 ml deionized water before metabolism was quenched by direct addition of ∼10 ml of liquid N₂ to the plates. The plates were then stored at −80°C. Quantification was performed by the Michigan Regional Comprehensive Metabolomics Resource Core, supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-097153 to the University of Michigan. For extraction, plates were transferred to a 4°C cold room, and 1.5 ml of ice-cold 90% 9:1 MeOH:CHCl₃ was immediately added to each plate and cells scraped/suspended with a cell lifter. The extraction solvent was spiked with ¹³C-labeled internal standards. Extracts were transferred to 1.5 ml microcentrifuge tubes and pelleted at 4°C. Supernatants were then assayed by HPLC/mass spectrometry. Samples were separated on an Agilent 1200 HPLC on a Phenomenex Luna NH₂ column, then analyzed either in positive or negative ESI by a coupled Agilent 6530 QTOF mass spectrometer (24). Compounds were identified based on retention time and mass/charge ratio match to injections of authentic standards. Quantitation was performed using Agilent Technologies MassHunter Quantitative software. Peak areas were measured from extracted ion chromatograms of metabolite ions with detection windows centered on the theoretical mass. Peak areas from internal standards were measured using an identical procedure. Data processing was performed using Agilent Technologies MassHunter Qualitative software for peak picking and MassProfiler Professional for data alignment, statistical analysis, and visualization as described (24). Total protein content of each sample was measured concurrently and all metabolite measurements for each sample were normalized to protein. Normalized levels were compared between shunt and control lines by unpaired, two-tailed Student's t-test.

Cellular ROS measurement by flow cytometry.

To compare total cellular ROS, three shunt and three control cell lines were grown to 70–90% confluency in 6-cm plates. Prior to analysis by flow cytometry, cells were incubated per manufacturer's recommendations in media with CellROX Deep Red at a concentration of 5 μM at 37°C for 30 min. Media with Deep Red was washed off with PBS × 3 and the cells were then trypsinized with 0.25% Trypsin EDTA solution and resuspended in PBS for analysis. After preparation, cells were immediately analyzed on a BD LSR II flow cytometer using FACSDiva software. The dye was excited using a 640-nm red laser, and emission from CellROX was measured in the 650–670 nm spectra. Data were analyzed using FlowJo V10 software. Cells were gated according to forward and sidescatter parameters to exclude acellular debris and isolate singlet events. Mean and median cellular fluorescence intensity were calculated along with parameters of variance for each sample. Median fluorescence intensity between groups was compared using two-tailed unpaired Student's t-test. To generate overlay graphic, all gated cells from shunt and control samples, respectively, were concatenated into histograms based on fluorescence in the 650–670 nm spectra and plotted using FlowJo software.

Determination of mitochondrial superoxide levels and mitochondrial membrane potential.

MitoSOX Red mitochondrial superoxide indicator (Molecular Probes, Grand Island, NY), a fluorogenic dye for selective detection of superoxide in the mitochondria of live cells, was used. Briefly, cells (1 × 10⁵) of the same passage (p8) from three control and three shunt PASMC lines were plated on into six-well plates. Twenty-four hours later, cells were washed with fresh media, incubated in media containing MitoSOX Red (5 μM) or TMRM (50 nM), for 30 min at 37°C in dark conditions, then subjected to fluorescence microscopy using an excitation of 510 nm and an emission at 580 nm (for MitoSOX) or an excitation of 548 nm and an emission at 575 nm (for TMRM). An Olympus IX51 microscope equipped with a CCD camera (Hamamatsu Photonics) was used for acquisition of fluorescent images. The average fluorescent intensities (to correct for differences in cell number) were quantified using ImagePro Plus version 5.0 imaging software (Media Cybernetics).

Measurement of NADPH oxidase activity.

Measurement of NADPH oxidase activity was performed as described previously (25). Briefly, cells were trypsinized, pelleted, then homogenized with Tris-sucrose buffer [10 mM Tris base (Fisher, Pittsburgh, PA), 340 mM sucrose (Mallinkrodt Baker, Philipsburg, NJ), 1 mM EDTA (Mallinkrodt Baker), 10 μg/ml protease inhibitor mixture (Sigma)]. The homogenate protein concentration was measured for each sample and 100 μg of homogenate protein + 100 μM NADPH substrate (Sigma) was added to an 8-mm test tube with 500 μl reaction buffer (5 μM lucigenin, 1 mM EGTA, and 50 mM phosphate buffer, pH 7.0) and then incubated at 37°C for 5 min. NADPH oxidase activity was measured by luminescence with emission measured at 15 s in a luminometer (Model TD-20/20; Turner Designs, Sunnyvale, CA).

Measurement of superoxide levels using electron paramagnetic resonance (EPR) spectroscopy.

To detect superoxide generation in intact cells, we performed electron paramagnetic resonance (EPR) measurements as described previously (49). Briefly, we used the spin probe 1-hydroxy-3-methoxycarbonyl-2,2,5,5-tetramethylpyrrolidine·HCl (CMH; Alexis Biochemicals, San Diego, CA). Twenty microliters of a spin-trap stock solution consisting of CMH [20 μM in Dulbecco's PBS (DPBS) plus 25 μM desferrioxamine (Calbiochem) and 5 μM diethyldithiocarbamate (Alexis Biochemicals, Lausen, Switzerland)] plus 2 μl of DMSO were added to each well before a 2 h apocynin (100 μm) treatment. Adherent cells were trypsinized and pelleted at 500 g after a 45-min incubation at 37°C following the treatment to allow entrapment of superoxide by the spin trap. Cell pellet was washed and suspended in a final volume of 35 μl of DPBS (plus desferrioxamine and diethyldithiocarbamate), loaded into a 50-μl capillary tube, and analyzed with a MiniScope MS200 EPR (Magnettech, Berlin, Germany) at a microwave power of 40 mW, modulation amplitude of 3,000 mG, and modulation frequency of 100 kHz EPR spectra. EPR signal amplitudes were analyzed using ANALYSIS 2.0 software (Magnettech). To quantitate the amount of superoxide per milligram of protein, we performed a standard reaction of the superoxide-generating enzyme xanthine oxidase as described previously (37). After the experiment, protein concentrations in mitochondria samples were measured. Final superoxide generation rate was expressed as picomoles per 40 min per milligram of protein.

PASMC lines from shunt lambs proliferate faster than those from controls.

Using previously described explant methods (41) primary PASMCs were derived from the proximal pulmonary arteries of shunt lambs (exposed to pulmonary overcirculation) and healthy control lambs with normal cardiac anatomy. Smooth muscle lineage was confirmed by immunohistochemical analysis of each cell line with antibodies directed against established smooth muscle lineage markers (40) including isoforms SM1 and SM2 (Sigma Aldrich M7786), and calponin (Fig. 1). Quantification of cell proliferation indicated that when plated at identical density and counted at sequential 24-h time points, PASMCs derived from shunt animals consistently exhibited significantly greater proliferation rates at all time points (Fig. 2).

PASMCs from shunt lambs exhibit metabolic alterations distinct from advanced PAH.

Given that shunt PASMCs proliferate much more rapidly than their control counterparts, we asked whether they would also display metabolic alterations similar to what has been described in transformed cancer cells or in the PASMCs of humans with advanced PAH. To this end, we performed extracellular flux analyses using the Seahorse XF-24 Extracellular Flux Analyzer (Seahorse Bioscience, Copenhagen, Denmark). This assay provides a multiplexed platform to obtain measurements of the cellular microenvironment in culture and provides simultaneous readings of cellular oxygen consumption and extracellular acidification. The oxygen consumption rate (OCR) represents a direct measurement of mitochondrial respiration while the extracellular acidification rate (ECAR) is an indirect measure of glycolytic lactate production (12). However, other acidification mechanisms may also contribute to ECAR, including carbonic acid produced by conversion of TCA cycle-derived CO₂ by carbonic anhydrases (29). Figure 3A shows the basal OCR readings at sequential time points within a single representative assay for one shunt and one control cell line. Figure 3B shows the composite average of 3 independently derived shunt and control cell lines under basal conditions normalized across multiple experiments. Fitting with our initial hypothesis, shunt PASMCs showed a 66% decrease in the rate of oxygen consumption compared with control cells (P < 0.001). Since the standard oxygen consumption measurements were performed in specially prepared media lacking the supplemental BSA in which the cells were routinely cultured, we were concerned that the assay may be missing a significant fraction of oxygen consumption related to the oxidation of exogenous fatty acids. We therefore repeated basal measurements with the addition of conjugated palmitate:albumin and found that although OCR did indeed increase by approximately 25% and 35% in controls and shunts, respectively (data not shown), the shunt PASMCs continued to have significantly lower O₂ consumption rates than control PASMCs, and the relative rate of decrease was maintained. Surprisingly, however, concurrent measurements of ECAR indicated that PASMCs from the shunted animals also exhibited substantially lower rates of basal extracellular acidification. As seen in Fig. 3, C and D, PASMCs derived from shunt animals consistently exhibited a 76% decrease in ECAR (P < 0.001) compared with controls. While some of the decrease in extracellular acidification may be related to decreased CO₂ production from the suppression of mitochondrial oxygen consumption, there is clearly no countervailing increase from glycolytic lactate production. Notably, the relative decrease in ECAR in these cells is even greater than the decrease in oxygen consumption. This finding represents a major deviation from the expected Warburg phenotype, which would couple a decrease in aerobic metabolism with a substantial increase in glycolytic activity, presumably to help supply biosynthetic precursor production required for enhanced cell division.

Despite the evidence that both mitochondrial oxidation and glycolytic flux are decreased in shunt compared with control cells, these cells do not exhibit any evidence of bioenergetic compromise. In addition to generating sufficient energy to carry out the biosynthetic processes required to facilitate their increased rate of proliferation, the shunt cells showed no deficit in the levels of the energetic intermediates ATP or NADH and no increase in the respective monophosphate or oxidized forms of these cofactors (Fig. 4). In fact, the level of AMP is slightly decreased in the shunt cells, resulting in an increased ATP-to-AMP ratio in these cells, typically an indicator of abundant energy supplies.

Glycolytic intermediates are decreased while pentose phosphate pathway intermediates are increased in shunt PASMCs.

Given that extracellular flux analyses can be limited, particularly when attempting to investigate glycolytic flux, we performed HPLC-mass spectrometry-based quantitative metabolite analyses in shunt and control PASMCs. We screened targeted metabolites from three of the central carbohydrate catabolic pathways: glycolysis, the pentose phosphate shunt, and the TCA cycle. Figure 5A shows results for specific intermediates of the glycolytic pathway in shunt and control cells. Fructose-1,6-bisphosphate is the first energetically committed metabolite of glycolysis, with preceding intermediates able to flow freely into other catabolic and anabolic pathways. Phosphoenolpyruvate (PEP) is the final pathway intermediate before the glycolytic production of pyruvate. Both were found at significantly lower levels in the shunt cells compared with controls (Fig. 5A). When interpreted in the context of the overall decrease in extracellular acidification rates, these results are consistent with an overall reduction in glycolytic flux in shunt PASMCs. While the ratio of fructose-6P to glucose-6-P does not reach statistical significance due to large variance in the control cells, the increased ratio in control cells is similarly consistent with this model. In contrast, Fig. 5B shows that specific intermediates of the pentose phosphate pathway (PPP) are increased in shunt PASMCs. d-Sedoheptulose-7-P is a unique seven-carbon sugar of the nonoxidative branch of the PPP, and erythritol is a metabolic product/by-product of this pathway. There is also an increased ratio of ribose-5-P relative to xylulose-5-P, suggesting a directionality of flow within the pathway towards production of ribose, a necessary precursor for DNA synthesis. Taken together, our results indicate an overall shift in glucose catabolism away from glycolysis and towards the PPP. We additionally measured pertinent energetic cofactors including NAD⁺ and NADH, ATP/AMP, and GTP/GMP (Fig. 4). Importantly, these results indicate that overall cellular bioenergetic status is not compromised in shunt PASMCs.

Shunt PASMCs exhibit accumulation of acetyl CoA and alterations in the TCA metabolite malate.

From the same targeted metabolic analysis, we found that shunt PASMCs have increased levels of acetyl-CoA (Fig. 5C). While levels of acetyl-CoA are below the detection threshold of the assay (<0.1 nanomoles per plate) in all of the control cell lines analyzed, the shunt cells exhibit a significant accumulation. This finding suggests that control PASMCs, with their relatively high oxygen consumption, are more rapidly utilizing acetyl-CoA as a fuel source for mitochondrial respiration while the shunt cells, and their corresponding lower oxygen consumption rates, are accumulating acetyl-CoA due to downstream suppression of the TCA cycle and/or mitochondrial electron transport chain inhibition. Of particular note, this elevation in acetyl-CoA levels is inconsistent with the observed decrease in this metabolite in PASMCs from advanced disease due to HIF-1α mediated suppression of mitochondrial respiration through regulation of pyruvate dehydrogenase (20, 32), which occurs upstream of acetyl-CoA production. It is thus quite interesting that there are distinct differences found within the TCA cycle itself. While the ratio of citrate to isocitrate, and succinate levels are not significantly different between control and shunt PASMC lines, the late cycle intermediate malate is significantly decreased in shunt cells (Fig. 5D).

Shunt PASMCs do not exhibit increased basal levels of HIF-1α or PDHK and are insensitive to treatment with dichloroacetate.

While shunt PASMCs show evidence of suppressed mitochondrial respiration consistent with the Warburg phenotype described in advanced pulmonary hypertension, the finding of elevated acetyl-CoA led us to question a similar role of HIF-1α mediated regulation in our model. Toward that end, we evaluated levels of HIF-1α and PDHK proteins in shunt and control cells in culture at atmospheric oxygen tension. The results, shown in Fig. 6, A and B, reveal that HIF-1α levels are nearly undetectable under these conditions in both shunt and control cells. Consistent with this, PDHK levels are also not increased in shunt PASMCs. To test whether pharmacological PDHK inhibition would affect PASMC proliferation, we cultured shunt and control PASMCs with the PDHK inhibitor DCA. As shown in Fig. 6C, at standard dosing concentrations described for cell culture (0.5 mM to 1 mM), there is no growth inhibitory effect of DCA on either shunt or control PASMCs. At higher doses, however (5 mM), cell line-independent repression of cell proliferation was observed, likely due to toxic effects.

Supplementation with cell-permeable dimethyl malate (DMM) does not alter the oxygen consumption or growth phenotype of shunt PASMCs.

These findings support the implication that the decrease in mitochondrial respiration we observe in the shunt PASMCs is not driven by the HIF-1α/PDHK upregulation found in advanced disease, but rather due to some limitation of the TCA cycle or intrinsic to shunt mitochondria themselves. In this context, the finding of decreased malate in these cells becomes an intriguing finding that may point to TCA cycle insufficiency as an underlying cause of the observed mitochondrial suppression. Malate and its metabolism have also been implicated in abnormal cell growth. Hereditary deficiency of fumarase, the enzyme that converts fumarate to malate, results in a cancer syndrome characterized by renal cell carcinoma and benign smooth muscle tumors (23). Conversely, in non-small cell lung cancers, overexpression of an abnormal mitochondrial malic enzyme diverts malate away from the TCA cycle and facilitates rapid cell growth. Malic enzyme knockdown or supplementation with cell permeable dimethyl malate (DMM) enhances O₂ consumption and slows proliferation (39). To test the hypothesis that malate deficiency may underlie mitochondrial repression in our model, we supplemented shunt and control cells with 5 mM DMM and evaluated proliferation and O₂ consumption. Figure 7A shows that the addition of DMM to standard assay media did not significantly alter O₂ consumption, by either shunt or control cells. Similarly, supplementation of growth media with DMM resulted in no significant change in proliferation of either shunt or control cells (Fig. 7B).

Shunt PASMCs exhibit increased cellular ROS and oxidative stress responses.

The oxidative arm of the PPP, controlled primarily by activity of the rate-limiting enzyme glucose-6-phosphate dehydrogenase (G6PDH), serves as a major source of reduced NADPH within the cell (Fig. 8A). Despite evidence that metabolic intermediates of the pentose phosphate pathway are increased, the level of the oxidized cofactor NADP⁺ is markedly elevated and the ratio of NADPH/NADP⁺ is significantly decreased in the shunt PASMCs (Fig. 8B). It initially seems paradoxical that the cofactor is skewed so heavily towards the oxidized form if overall PPP activity is increased; however, the shift towards NADP⁺ at the expense of NADPH has been shown to be a potent activator of G6PDH and PPP activity (7). Furthermore, other authors have speculated that this ratio is an important determinant of G6PDH activity in the pulmonary arteries (15), and in rodent models of PAH, G6PDH has been shown to play a role in the regulation of PASMC growth and contractile phenotypes (8). In its reduced form, NADPH is an important enzymatic cofactor in a number of essential cellular processes including the reductive biosynthesis of fatty acids, nucleotides, amino acids and cholesterols, generation of superoxides via NADPH oxidases, and protection of cells from free radical damage through the regeneration of reduced glutathione (7). It has been shown previously in this animal model that lung tissue shows an increase in NADPH dependent superoxide generation (44), and other investigators have established that cyclic stretch increases levels of cytosolic ROS in PASMCs (48). We therefore hypothesized that the alteration in the NADPH/NADP⁺ in our cells may represent a state of increased NADPH consumption secondary to cellular oxidative stress. To evaluate the oxidative state of the cells we treated them with CellROX Deep Red, an agent that fluoresces in proportion to the amount of total cellular ROS, and analyzed the cellular fluorescence intensity by flow cytometry. As shown in Fig. 8C, the mean fluorescence intensity of the shunt PASMCs was indeed elevated by 24.7% compared with that of controls (P < 0.05). This is consistent with other models of pulmonary hypertension which have described increased cellular ROS originating from a variety of sources (33, 48).

Cellular ROS may originate from various sources, but in the lungs and pulmonary vasculature they primarily arise from the NADPH oxidase (Nox) complexes, uncoupled nitric oxide synthase (NOS) and the mitochondrial electron transport chain (2). Both mitochondria and Nox complexes have been implicated as sources of increased smooth muscle cell ROS in other models of PAH (2, 4). Given our findings of altered mitochondrial respiration, we first sought to determine whether the mitochondria accounted for the increase in cellular ROS levels. We specifically measured mitochondrial ROS with the selective dye MitoSox Red, and found that the shunt PASMCs exhibited reduced mitochondrial ROS levels (Fig. 9A). Consistent with their diminished activity, mitochondria in shunt PASMCs also had lower membrane potentials (Fig. 9B). Thus the increase in cellular ROS observed in shunt PASMCs does not appear to originate from their mitochondria.

Shunt PASMCs exhibit increased NADPH oxidase expression and activity which contributes to their hyperproliferative phenotype.

We then turned our attention to the Nox system as a potential source of increased ROS in shunt PASMCs. The Nox complexes are transmembrane multimers that catalyze the formation of superoxide (O₂⁻) by electron transfer from NADPH to molecular O₂. There are several catalytic isoforms, but Nox 1, 2, 4 and 5 have been described in the pulmonary vasculature. Some isoforms, such as 4, are constitutively active, while others require stabilization and activation by other subunits to function (21). We first looked at expression of the Nox complexes and found increased protein levels of both the Nox 4 and Nox 2 catalytic domains, along with the Rac1 and p47^phox subunits of the Nox 2 complex (Fig. 10A). We then evaluated total NADPH oxidase activity in cellular homogenates using an established chemiluminescence assay (23) and demonstrated a fivefold increase in total Nox activity in shunt compared with control PASMCs (Fig. 10B). To validate this result in intact cells, we performed electron paramagnetic resonance (EPR) measurement as previously described (49) to measure Nox-dependent superoxide generation. As seen in Fig. 10C, shunt PASMCs exhibited a 115% increase in Nox-dependent superoxide production compared with control cells. We then sought to determine what effect, if any, this increase in cellular Nox expression and activity was exerting on the cellular proliferation phenotype. To this effect we utilized the small molecule VAS 3947 to pharmacologically inhibit Nox activity and evaluate the impact on PASMC proliferation. VAS 3947 is a commercially available triazolo-pyrimidine compound that inhibits Nox activity and is notable for its lack of off-target effects (notably including NOS and xanthine oxidase) and the absence of intrinsic antioxidant activity (3). VAS 3947 (10 μM) preferentially slowed the growth of shunt PASMCs in vitro (17.3% slower than standard growth conditions) with no significant effect on the growth of control PASMCs (Fig. 10D). We therefore conclude that increased Nox activity may be a driver of PASMC hyperproliferation and metabolic shift in the setting of pulmonary overcirculation.