Pluripotent hematopoietic stem cells augment α-adrenergic receptor-mediated contraction of pulmonary artery and contribute to the pathogenesis of pulmonary hypertension

Abstract

Pulmonary hypertension (PH) is a multicellular and progressive disease with a high mortality rate. Among many cell types, hematopoietic stem cells (HSCs) are incriminated in the pathogenesis of PH. However, our understanding of the mechanisms that increase HSCs in blood and lungs of hypertensive animals or patients and the role played by HSCs in the pathogenesis of PH remains elusive. Studies suggest that glycolysis is critical for the survival and growth of HSCs. In various cell types from hypertensive lungs of animals and patients, glycolysis and the glucose-6-phosphate dehydrogenase (G6PD) activity are increased. Herein, we demonstrated in mice that chronic hypoxia increased HSCs (CD34⁺, CD117⁺, CD133⁺, CD34⁺/CD117⁺, and CD34⁺/CD133⁺) in bone marrow and blood and around hypertensive pulmonary arteries in a time-dependent manner. Intriguingly, we found fewer CD133⁺ cells in the bone marrow of C57BL/6 mice compared with Sv129J mice, and C57BL mice developed less severe chronic hypoxia-elicited PH and heart failure than Sv129J mice. Similarly, the numbers of CD34⁺ and CD117⁺ cells in blood of patients with pulmonary arterial hypertension (PAH) were higher (>3-fold) compared with healthy individuals. By allogeneic bone marrow transplantation, we found that GFP⁺ bone marrow cells infiltrated the lungs and accumulated around the pulmonary arteries in lungs of hypoxic mice, and these cells contributed to increased α-adrenergic receptor-mediated contraction of the pulmonary artery cultured in hypoxia. Inhibition of G6PD activity with (3β,5α)-3,21-dihydroxypregnan-20-one, a novel and potent G6PD inhibitor, decreased HSCs in bone marrow, blood, and lungs of hypoxic mice and reduced α-agonist-induced contraction of the pulmonary artery and established hypoxia-induced PH. We did not observe CD133⁺ cells around the pulmonary arteries in the lungs of chronically hypoxic G6PD-deficient mice. Furthermore, knockdown of G6PD and inhibition of G6PD activity: 1) downregulated canonical and noncanonical Wnt and Fzd receptors genes; 2) upregulated Bmpr1a; 3) decreased Cxcl12, and 4) reduced HSC (CD117⁺ and CD133⁺) numbers. In all, our findings demonstrate unexpected function for bone marrow-derived HSCs in augmenting α-adrenergic receptor-mediated contraction of pulmonary arteries and remodeling of pulmonary arteries that contribute to increase pulmonary vascular resistance in PAH patients and hypoxic mice and suggest that G6PD, by regulating expression of genes in the WNT and BMPR signaling, contributed to increase and release of HSCs from the bone marrow in response to hypoxic stimuli.

INTRODUCTION

Pulmonary hypertension (PH) is a multicellular and progressive disease and is a significant cause of morbidity and mortality in patients with several different clinical conditions. Severe PH remains debilitating and deadly in women and men (8). Smooth muscle cells (SMCs), endothelial cells, fibroblasts, macrophages, and stem cells are incriminated in the pathogenesis of chronic hypoxia-induced PH (HPH; group III) and idiopathic (iPAH) or heritable pulmonary arterial hypertension (PAH; group I) (23, 59, 64, 65).

About 30% of systemic sclerosis patients who show PH have increased CD34⁺CD133⁺ and CD133⁺/vascular endothelial growth factor receptor-2 (VEGFR-2)⁺ progenitor cells, aka hematopoietic stem cells (HSCs), in peripheral blood (21, 45). Similarly, the number of CD34⁺CD133⁺ cells are higher in blood (5, 44, 47) and in the bone marrow of PAH patients than in healthy controls (24). CD133⁺ and CD133⁺VEGFR-2⁺ cells are also increased in the blood of commonly used animal models of PH: 1) beagle dogs after treatment with monocrotaline (70), 2) Sugen/hypoxia/normoxia rats (15), and 3) hypoxic mice (32). Interestingly, increased CD34⁺CD133⁺ cells in the blood of PAH patients correlate with elevated pulmonary arterial pressure (5) and cardiac dysfunction (40). In addition, CD34⁺CD133⁺, CD133⁺VEGFR-2⁺, CD133⁺, and CD31⁺ cells are increased in the lungs of PH patients (24, 41, 64, 72). CD133⁺ cells potentially contribute to the remodeling of the pulmonary artery in iPAH patients (5) and are found around pulmonary arteries of rats with hypoxia- and Sugen/hypoxia/normoxia-induced PH (15) and of mice with bone morphogenetic protein receptor 2 (BMPR2) mutation (68). Bone marrow cells transplanted from BMPR2(R899X) mutant to control recipient mice are reported to contribute to the pathogenesis of PAH with remodeling and inflammation in recipient mice (71). Xenograft of CD133⁺ cells of iPAH patients leads to lung endothelial injury, angio-proliferative pulmonary vascular remodeling, and widespread in situ thrombosis in immunodeficient NOD-SCID (nonobese diabetic-severe combined immunodeficiency) mice (6). Engraftment of CD133⁺ cells from iPAH patients in mice also results in hypertrophy of the right ventricle (RV) and death (6). HSCs are progenitors of proinflammatory macrophages and immunogenic (T and B) cells that contribute to the pathogenesis of pulmonary artery remodeling (23, 38). Although these studies suggest that increased circulating HSCs in PAH patients play a potential role in the pathogenesis of hypertension, others suggest that HSCs in the blood of PAH patients either do not differ from control subjects (3, 56, 57) or are lower in patients with PAH than in those without PAH (54). Moreover, our understanding of the mechanisms that increase HSCs in blood and lungs of animals or patients with PH and how these cells modify the function of pulmonary arteries remains elusive. Therefore, one purpose of this study is to determine whether bone marrow-derived HSCs increased in chronic HPH and PAH patients, and whether these cells modulate function of the pulmonary artery.

Glycolysis is critical for the survival and growth of HSCs (15, 46, 51). In various cell types from hypertensive lungs of animals and patients, glycolysis and the glucose-6-phosphate dehydrogenase (G6PD) activity are increased (18). G6PD activity increased by ectopic expression of G6PD and hypoxia promotes proliferation of CD133⁺ progenitor cells and influences pulmonary artery SMC-CD133⁺ interactions in cell culture experiments (15). Additionally, we recently found that Sv129J mice lacking cytochrome P-450 2C44 enzyme, which produce eicosanoids in a NADPH-dependent manner, develop severe chronic HPH (35) and have high levels of CD133⁺ and occlusive lesions (32). Therefore, another purpose of this study is to test the hypothesis that G6PD plays a role in increasing HSCs in mice exposed to chronic hypoxia, and that inhibition of G6PD mitigates chronic hypoxia-induced maladaptive HSC biology and PH in mice.

MATERIALS AND METHODS

All experiments were performed following an Institutional Animal Care and Use Committee (IACUC) approved protocol in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals Internal Review Board. All protocols were approved by IACUC and Institutional Review Board. Male and female Sv129J mice [25–30 g; wild type (WT); purchased from Jackson Laboratory], G6PD^def mice (29–35 g, gift from Dr. Leopold, Harvard Medical School, Boston, MA) and its eugenic control, and C57BL/6-Tg mice (CAG-EGFP; 25–30 g; purchased from Jackson Laboratory) (GFP mice) were used in this study. All chemical reagents were purchased from Sigma Chemical Co. (MO), BD Biosciences (CA), or Thermo-Fisher Scientific (NJ). Antibodies are as follows.

Adoptive bone marrow transfer.

Methods for bone marrow transplantation were described previously (10). Briefly, isolation of bone marrow of GFP donor mice and injection to C57BL/6 recipient mice are described below. GFP donor mice were anesthetized with 2% isoflurane. Killed mice were immediately immersed in 70% ethanol for several minutes, the skin was removed from the lower extremities, and the hind legs were cut off. The skeletal muscles were removed from the tibias and femurs, which were placed in tubes containing ice-cold DMEM (Lonza, Walkersville, MD) containing 10% FBS (Atlanta Biologicals, Lawrenceville, GA) and 1% antibiotics (penicillin-streptomycin-amphotericin B; MP Biomedicals, Santa Ana, CA). The bone marrow was exposed by cutting of the bone caps under sterile conditions and were collected into sterile centrifuge tubes by flushing the bones with DMEM containing 10% FBS and heparin (10 U/mL; APP Pharmaceuticals, Schaumburg, IL). The cells were collected by centrifugation at 300 g for 5 min at 4°C. The supernatant was aspirated, and the cell pellet was resuspended in 10 mL of red cell lysis buffer (0.15 M NH₄Cl; 0.01 M KHCO₃; 0.1M Na₂EDTA, pH 7.2–7.4) and incubated on ice for 15 min. Cells were then collected by centrifugation, washed twice with PBS containing 1% albumin, and resuspended to a single cell suspension in PBS containing 1% albumin. Trypan blue was used to visualize dead cells. Overall, viability >95% was observed. Cells were collected by centrifugation and resuspended in DMEM containing 1% albumin and heparin (5 U/mL) at a concentration of 5 × 10⁶ cells/100 μL. Recipient mice received a lethal dose of γ-irradiation (130 rad/min, 900 rad), followed by injection of 10 × 10⁶ bone marrow cells in 200 μL into the tail vein under anesthesia with 2% isoflurane. One week before and 4 wk after irradiation, recipient mice were placed on acidified water (pH 3.0) supplemented with 0.17 mg/mL enrofloxacin (Baytril; Bayer Healthcare, Shawnee Mission, KS).

Isolation of bone marrow cells and blood cells and flow cytometry.

Bone marrow cells were collected from the tibia and femur, and blood samples were collected from the left ventricle (LV) of Sv129J and C57BL/6 male and female mice (25–30 g). In some experiments, bone marrow cells were cultured in DMEM media (15%) for 24 h. Cells (10⁶) suspended in 90 µL of buffer were treated with 10 µL of FcR blocking reagent (Miltenyi Biotec, CA) for 10 min at 4°C and stained with 10 µL of fluorescent antibodies for 15 min at 4°C. We used phycoerythrin (PE)-conjugated anti-CD117 antibody, PE-anti-CD11b antibody, fluorescein (FITC)-conjugated anti-CD34 antibody, FITC-anti-F4/80 antibody, and allophycocyanin (APC)-conjugated-anti-CD133 antibody. All antibodies were purchased from Miltenyi Biotec. After red blood cell lysis using lysing buffer (BD Biosciences), cells were analyzed by MoFlo XDP (Beckman Coulter, CA), and FCM analysis software are Kaluza version 1.3 (Beckman Coulter) and FlowJo version 10 (FlowJo, LLC). Negative control (without) primary antibody-treated cells were used each time as validation of antibodies.

Bone marrow cell culture.

Bone marrow cells isolated from bone marrow of Sv129J and C57BL/6 male and female mice (25–30 g), as explained earlier, were cultured in hematopoietic progenitor cell expansion medium (PromoCell) supplemented with cytokine mix E (PromoCell) containing thrombopoietin, stem cell factor, Flt-3 ligand, and IL-3. Cells were expanded for 1 wk in 75-cm² flasks (Corning, Corning, NY) at 37°C in a humidified CO₂ incubator before use.

Isolation of small IPA and measurement of IPA tone.

Mice (25–30 g) were euthanized by cervical dislocation and placed in 70% (vol/vol) ethanol for 2 min. The left lung and caudal lobe of the right lung where dissected and placed in 10% FBS containing 1% penicillin-streptomycin. Small intrapulmonary arteries (IPA) of third order (100–150 μm in diameter) were taken from the lung, dissected free of connective tissue, and placed in FBS. IPA was added to a cell culture dish with bone marrow cells fraction and coincubated for 48 h at 37°C in a hypoxic chamber (InvivO₂300, Ruskin Technology) under 3% O₂ and 5% CO₂, or under normoxia at 21% O₂ and 5% CO₂. After 48 h, IPA were replaced in Krebs bicarbonate buffer solution (pH 7.4) containing the following (in mM): 118 NaCl, 4.7 KCl, 1.5 CaCl₂ × 2H₂O, 25 NaHCO₃, 1.1 MgSO₄, 1.2 KH₂PO₄, 5.6 glucose, and 10 HEPES. Then the vessels were mounted on a wire myograph (Danish Myo Technology A/S, Aarhus, Denmark) and bathed in Krebs buffer solution at 37°C and an optimal passive tension of 3 mN. After 30 min of incubation, arterial preparations viability and equilibration were assessed by the stimulation of the vessels with repeated 10-min exposures to 60 mM KCl (60K). A brief SMC membrane depolarization of vascular preparations with 60K increases the reproducibility of subsequent vascular responses. For registration of vascular ring contractile activity and its following analysis, Chart 5.5.4 and LabChart Reader 8.1.9 (ADInstruments, Inc.) software were used. Vascular tension is presented as a percentage of the constriction level obtained from the exposure to 60K. Acetylcholine-induced vasodilatation is presented as a percentage of the constriction level obtained from the exposure to 10 μM phenylephrine (PhE).

Induction of pulmonary hypertension.

Mice were exposed to normobaric hypoxia (10% O₂) in a ventilated chamber for 5 wk, as described recently (32, 35). The normoxic control mice were kept in room air for all 5 wk. At the end of the experiments, mice were euthanized and lungs were harvested for biochemical and histological analyses.

Echocardiography.

Echocardiography was performed in 2% isoflurane anesthetized mice using a Vevo 770 imaging system (Visual Sonics, Toronto, ON, Canada). Briefly, in the beginning of the experiment (week 0) and at the end of the experiment (week 3 and 5), two-dimensional parasternal short-axis view was obtained, M-mode assessment of LV function was performed, and LV parameters were measured, as described previously (27). Two-dimensional parasternal short-axis view at the level of the aortic valve was obtained, and pulsed-wave Doppler recording of the pulmonary artery blood flow was recorded, as described previously (62). The ratio of pulmonary artery acceleration time (PAAT; time taken from start of flow to maximal velocity) and ejection time (ET; time taken from start of flow to the end of flow) were determined, as described previously (17).

Hemodynamic.

Hemodynamic measurements were performed as described previously (35). Briefly, at the end of the experiment protocol, adult mice were anesthetized by 4% isoflurane, and 1–2% isoflurane was used to maintain anesthesia for the entire duration of the surgery and data acquisition. Body temperature of the animal during the surgery was maintained using a heating pad. About a 3-cm² area of skin over the ventral neck region was exposed to locate right common carotid artery, which was then carefully isolated, and a 1.4 F Millar Mikro-Tip pressure catheter was inserted and pushed to the LV, and the LV hemodynamic parameter was measured.

RNA-Seq analysis.

After collecting smooth muscle and lungs from mice, total RNA was isolated from tissue using the Qiagen All Prep DNA/RNA/miRNA Universal kit, according to manufacturer’s instructions. RNA was quantified using the NanoDrop (ThermoFisher), and quality control was assessed using the Agilent Bioanalyzer 2100. RNA-seq library construction was performed using the TruSeq Stranded Total RNA Preparation kit (Illumina) with 200 ng of RNA as input, according to the manufacturer’s instructions. Libraries were sequenced on the HiSeq2500 with single-end reads of 100 nt at the University of Rochester Genomics Research Center. Single-end sequencing was done at a depth of 10 million reads per replicate (n = 3). Quantitative analysis, including statistical analysis of differentially expressed genes, was done using Cufflinks 2.0.2 and Cuffdiff2 (http://cufflinks.cbcb.umd.edu). The Benjamini-Hochberg method was applied for multiple test correction (false discovery rate < 0.05).

RNA expression by quantitative PCR.

mRNA analysis was performed by quantitative PCR, as previously described (13, 16, 34). Briefly, total RNA was extracted from lungs and HSCs using a Qiagen miRNEasy kit (catalog no. 217004). The input RNA quality and concentration were measured on the Synergy HT Take3 Microplate Reader (BioTek, Winooski, VT), and cDNA was prepared using miR-specific TaqMan miR assays (Applied Biosystems, Foster City, CA) or using SuperScript IV VILO Master Mix (catalog no. 11756500, Invitrogen) for mRNA. Quantitative PCR was performed in triplicates using TaqMan Universal PCR Master Mix (catalog no. 4324018) for miRNAs or using TaqManTM Fast Advanced Master Mix (catalog no. 44–445–57) for mRNA using a Mx3000p Real-Time PCR System (Stratagene, Santa Clara, CA). The primers for the quantitative PCR were purchased from Thermo Fisher Scientific/TaqMan. Results for mRNA expression were normalized to internal control Gapdh, and relative mRNA expression was determined using the ΔCt method, as described previously (39).

Histology.

Mice were euthanized, and lungs were harvested for histological analyses. The left lung lobe was inflated with 0.5% agarose in 1% neutral buffered formalin at 20-cmH₂O pressure and fixed in 10% neutral buffered formalin overnight (2). Formalin-fixed lung lobes were blocked and embedded in paraffin. Formalin-fixed paraffin embedded sections were cut at 5-μm thickness for the immunohistology in the core histology laboratory at New York Medical College.

Immunofluorescent staining.

Paraffin-embedded lung sections (5 μm) were deparaffinized and heated with 1 × citrate buffer. The sections were permeabilized with 0.2% Triton X-100 for 45 min and blocked with 10% goat serum for 1 h at room temperature. The sections were then incubated with primary antibodies, anti-CD117 rabbit antibody (1:50, Santa Cruz, CA), anti-CD133 rabbit antibody (1:300, Santa Cruz), anti-von Willebrand factor (vWF) rabbit antibody (1:200, Dako, Glostrup, Denmark), anti-GFP mouse antibody (1:100, GeneTex), smooth muscle myosin heavy chain rabbit antibody (1:100, Sigma), and anti-α-smooth muscle actin mouse antibody (1:400, Sigma) over night at 4°C. We and others have validated these antibodies for histology in the previous studies (15, 36). For immunofluorescence staining, sections were incubated with fluorescent secondary antibodies, Alexa Fluor 594-conjugated goat anti-rabbit IgG (1:400, Thermo Fisher Scientific) and Alexa Fluor 488-conjugated goat anti-mouse IgG (1:200, Thermo Fisher Scientific) for 1 h at room temperature. The nucleus was stained with DAPI (1 μg/mL). Imaging was done via Axio imager M1 microscope, Axiocam MRm camera, and AxioVision microscopy software (Carl Zeiss, Jena, Germany) or via FV1000-D confocal laser scanning microscope (OLYMPUS, Tokyo, Japan), and data were analyzed using image J software (52, 53).

Statistical analysis.

Values are means ± SE from different animals (n). Statistical analyses were performed with a two-way ANOVA with Bonferroni correction for comparing multiple groups and unpaired Student’s t test. P < 0.05 was used to establish statistical significance between two groups.

RESULTS

Hypoxia-induced PH.

We exposed Sv129J mice to hypoxia for 5 wk, and, as expected in hypoxic mice, the PAAT-to-ET ratio decreased (Fig. 1A) and RV hypertrophy (RV/LV + septum) increased (Fig. 1B), and cardiac output (Fig. 1C) and stroke work (Fig. 1D) decreased progressively at 3 and 5 wk. Interestingly, these mice developed LV stiffness (calculated from the ratio of maximum change in pressure over time to maximum change in volume over time, Fig. 1E) and large systemic artery stiffness (Fig. 1F) between 3 and 5 wk, compared with their normoxic controls.

Fig. 1.

Measurement of cardiac function in normoxic and hypoxic Sv129J mice. Mice (male = 3 and female = 2; 25–30 g) were exposed to normoxia for 5 wk and to hypoxia for 3 and 5 wk, after which echocardiography and cardiac catheterization were performed. Summary data of the ratio of pulmonary artery acceleration time (PAAT) to ejection time (ET) (A), right ventricle (RV) hypertrophy (B), cardiac output (C), stroke work (D), left ventricular (LV) stiffness (E), and arterial stiffness (F) are shown. Values are means ± SE; n = 5 mice in each group. dP/dt_max, maximum change in pressure over time; dV/dt_max, maximum change in volume over time; Ea, arterial elastance. *P < 0.05 vs. normoxia (0 wk hypoxia) and #P < 0.05 vs. hypoxia (3 wk).

Hematopoietic stem cells were increased in bone marrow and blood of hypoxic mice.

To understand the role of bone marrow-derived HSCs in the pathobiology of HPH, we first examined the time course of HSC increase in the bone marrow, blood, and lungs of mice exposed to hypoxia (10% O₂) for up to 5 wk. In the bone marrow of these mice, the levels of HSCs (CD34⁺, CD117⁺, and CD133⁺; Fig. 2, A–E) and differentiated HSCs (CD34⁺CD117⁺; Fig. 2, A–E) increased significantly in the 5th wk of hypoxia compared with their respective normoxia controls. Interestingly, HSCs (CD34⁺, CD117⁺, and CD133⁺; Fig. 2F) and differentiated HSCs (CD34⁺CD117⁺; Fig. 2F) in the blood decreased significantly in the first 2 wk of hypoxia, and subsequently their numbers began to increase between the 2nd and 3rd wk and significantly elevated at the end of the 5th wk of hypoxia compared with their respective normoxia controls. CD34⁺CD133⁺ cell numbers did not change in bone marrow and blood (Fig. 2, E and F). During the same time period, the expression of pulmonary artery SMC markers (CNN1 and MYH11) decreased in the pulmonary arteries of hypoxic mice (Fig. 3A), and CD133⁺ cells began to accumulate in and around pulmonary arteries by 2 wk of hypoxia and surrounded the pulmonary arteries by 5 wk of hypoxia (Fig. 3B). Additionally, wVF⁺ cells were seen on the luminal side and in the lumen of the pulmonary artery at 2 and 5 wk of hypoxia, respectively (Fig. 3C). Moreover, the expression of CD133 (Fig. 3D), determined by antibody previously used and validated by us and others (4, 15, 19), increased in the lungs of hypoxic compared with normoxic mice. Consistently, the expression of mRNA of stem cells (Cd164, Flt3) or T cell (Cd7, Cd19) markers (Fig. 3E and Table 1) is altered in the lungs of hypoxic compared with normoxic mice.

Fig. 2.

Hematopoietic stem cell (HSC) numbers in bone marrow (BM) and blood of normoxic and hypoxic Sv129J mice. Mice (male and female; 25–30 g) were exposed to hypoxia up to 5 wk. The leukocyte-enriched cell fraction was isolated from BM and blood of normoxic mice after 5 wk and blood of hypoxic mice after 2, 3, and 5 wk. HSC numbers were analyzed by flow cytometry. A–D: typical histograms showing bone marrow cells stained with HSC markers: CD34 (A), CD117 (B), CD133 (C), and CD34 and CD117 (D). E and F: percentages of cells positive for each HSC marker in the bone marrow (E) and blood (F). Values are means ± SE; n = 9 (normoxia; male = 5 and female = 4), n = 3 (2-wk hypoxia; male = 1 and female = 2), n = 5 (3-wk hypoxia; male = 3 and female = 2), and n = 6 (5-wk hypoxia; male = 3 and female = 3) mice. APC, allophycocyanin; BMC, bone marrow cells; PE, phycoerythrin; WBCs, white blood cells. *P < 0.05 vs. normoxia (0-wk hypoxia).

Fig. 3.

CD133⁺ and vWF⁺ cells increased in lungs around occluded pulmonary arteries of Sv129J mice. Mice (male and female; 25–30 g) were exposed to hypoxia (Hx) for 5 wk. Lungs from these mice were removed for immunohistochemistry. Lung sections (5 μm thickness, n = 4 in each group; male = 2 and female = 2) of normoxic (Nx) and Hx mice were prepared and stained for smooth muscle myosin heavy chain 11 (MYH11), calponin 1 (CNN1), prominin (CD133), actin (ACTA), von Willebrand factor (vWF), and 4′,6-diamidino-2-phenylindole (DAPI). A–C: representative images show markers for CNN1 and MYH11 (A), CD133 and ACTA (B), and endothelial cells, vWF, and ACTA (C). Nucleus is stained in blue with DAPI. D: representative Western blot and summary data (n = 5; male = 3 and female = 2) showing CD133 expression in lungs of normoxic and hypoxic mice. E: expression of genes that encode stem cell markers in lungs of normoxic (n = 4; male = 2 and female = 2) vs. hypoxic (n = 4; male = 2 and female = 2) mice was determined by whole-genome RNAseq analysis (also see Table 1). WT, wild type. Values are means ± SE. *P < 0.05 vs. Normoxia. Scale bar is 20 µm.

Table 1.

Expression of stem cell and T-cell marker genes in lungs of hypoxic versus normoxic Sv129J mice

Gene	Log2 Fold Change	P Value	Adjusted P Value
Cd164	2.45812452	0.00079069	0.04305923
Flt1	−1.2002586	0.01713251	0.18027399
Flt3	1.95576036	0.13795471	NA
Cd7	1.45924762	0.28575461	NA
Cd19	2.43984267	0.01981061	NA
Cxcr4	2.43921605	0.01857532	NA
Cxcl12	1.38203347	0.0009729	0.04608311

NA, not applicable.

GFP-labeled bone marrow cells accumulated around pulmonary arteries in lungs of hypoxic mice.

Although we found CD133⁺ cells around the pulmonary arteries of mice exposed to hypoxia, it was unclear whether the HSCs resident in the lungs or the HSCs that infiltrated from the circulation encircled the pulmonary arteries of hypoxic mice. Therefore, to determine whether infiltrated HSCs accumulated around the arteries, we performed allogeneic bone marrow transplantation from GFP (donor) to C57BL6 (recipient) mice (Fig. 4A). C57BL6 recipient mice transplanted with bone marrow cells of GFP mice showed GFP⁺ cells in the bone marrow and blood (Fig. 4, B and C). We observed GFP⁺ cells in the lungs of normoxic recipient mice (Fig. 4D). Interestingly, the number of GFP⁺ cells in and around remodeled pulmonary arteries was increased by hypoxia in recipient mice (Fig. 4E).

Fig. 4.

Transplanted bone marrow (BM) cells accumulate in lungs of hypoxic mice. A: C57BL/C mice (male and female; 25–30 g) were irradiated and transplanted with bone marrow cells from green fluorescent protein (GFP) mice, after which recipient mice were exposed to hypoxia for 3 wk. B and C: the leukocyte-enriched cell fraction was isolated from the bone marrow and blood of nonrecipient mice (B) and recipient mice (C), and flow-cytometry was performed. Typical histograms of the BM (left) and blood (right) showing GFP⁺ cells were observed. D and E: representative images of lung sections (5 μm thickness) of recipient mice in normoxia (D) and hypoxia (E) showing GFP (green), smooth muscle actin (ACTA, red), and DAPI (blue nucleus). GFP⁺ cells (transplanted-GFP⁺-descendent cells) were observed in the bone marrow, blood, and lung tissues. n = 5 Mice (male = 3 and female = 2). Scale bar is 20 µm.

CD133⁺ cells are more abundant in the bone marrow of Sv129J than in C57BL/6 mice, and chronic hypoxia-elicited heart failure was more severe in Sv129J than C57BL/6 mice.

C57BL/6 and Sv129 mice are commonly used in laboratory research, including in the study of PH. However, Sv129J and C57BL/6 mice show significant variability in immunity, phenotypic variance, and genetics and viability (50). Therefore, keeping in mind that results should be extrapolated cautiously from single strain to human (patho)physiology, we compared HSCs numbers and hemodynamics in Sv129J and C57BL/6 mice. As demonstrated in Fig. 5, A–C, there are significantly more CD133⁺ and CD133⁺CD11b⁺ cells in the bone marrow of Sv129J than in C57BL/6 mice. Along those lines, expression of genes for HSC markers (Cd164 and Flt3), T- and B-cell markers (Cd7 and Cd19), and CXCL12 receptors (Cxcr4) was more increased in the lungs of hypoxic compared with normoxic Sv129J mice (Fig. 5D and Table 1), but not as much in lungs of hypoxic compared with normoxic C57BL/6 mice (Fig. 5D and Table 2). Concomitantly, the hypoxia-induced 1) decrease of PAAT-to-ET ratios, 2) RV hypertrophy, and 3) large arterial and LV stiffness, and a decrease in stroke work and cardiac output, were more severe in Sv129J than in C57BL/6 mice (Fig. 5, E–J).

Fig. 5.

Comparison of hematopoietic stem cell (HSC) levels, expression of genes encoding stem cell markers, and hypoxia-induced pulmonary hypertension in Sv129J and C57BL mice. A–C: CD34⁺, CD133⁺, and CD133⁺CD11b⁺ cells in the bone marrow isolated from normoxic (Nx) Sv129J mice (male = 2–3 and female = 2–3; 25–30 g) and C57BL/6 mice (male = 2 and female = 2; 25–30 g) were quantitated. Sv129J and C57BL/6 mice were then exposed to normoxia or hypoxia (Hx) for 5 wk. D: expression of genes encoding stem cell markers in lungs of normoxic (n = 4; male = 2 and female = 2) vs. hypoxic (n = 4; male = 2 and female = 2) Sv129J and C57BL/6 mice were quantitated by RNAseq (also see Tables 1 and 2). E–J: summary data show the ratio of pulmonary artery acceleration time (PAAT)-to-ejection time (ET) (E), right ventricle (RV) hypertrophy (F), arterial stiffness (G), left ventricular (LV) stiffness (H), stroke work (I), and cardiac output (J) in Sv129J (n = 6–9; male = 3–5 and female = 3–4) vs. C57BL/6 (n = 7–10; male = 4–5 and female = 4–5) mice. P < 0.05 in Sv129J-hypoxia (Hx) as compared with C57BL/6-hypoxia (Hx). Values are means ± SE. BMCs, bone marrow cells; dP/dt_max, maximum change in pressure over time; dV/dt_max, maximum change in volume over time; Ea, arterial elastance; NS, nonsignificant; S, septum.

Table 2.

Expression of stem cell and T-cell marker genes in lungs of hypoxic versus normoxic C57BL/6 mice

Gene	Log2 Fold Change	P Value	Adjusted P Value
Cd164	0.34272723	0.52604988	0.86028228
Flt3	−1.3236639	0.07872876	0.38601569
Cd7	−0.2592401	0.79400187	NA
Cd19	−0.5359321	0.45229887	0.81194312
Cxcr4	−0.5820116	0.23906503	0.61172618
Cxcl12	1.99678799	0.00027356	0.06232805

NA, not applicable.

CD34⁺ and CD117⁺ cells are increased in blood of PAH patients.

To determine whether stem cells increase and contribute to pulmonary artery disease in PAH patients, we performed FACS analysis and found that pluripotent CD34⁺ (Fig. 6A) and CD117⁺ (Fig. 6B) cell numbers are between two- and threefold higher in blood of PAH (five scleroderma and one idiopathic) patients than those found in healthy (non-PAH) individuals. Along these lines, in PAH patients, CD34⁺, but not CD117⁺ (data not shown), cells positively correlated with pulmonary vascular resistance (PVR; Fig. 6C). Patient’s demographic and hemodynamic results are presented in Table 3. Furthermore, by immunohistology, we found that CD133⁺ (Fig. 6, D–G) and CD117⁺ (Fig. 6, H–K) cells are present in lungs of PAH patients compared with individuals without PAH (control). CD133 expression is seen in plexiform lesion (Fig. 6, F and G), possibly including endothelial cells (Fig. 6F) or SMCs (Fig. 6G). CD117 expression is seen in plexiform lesions (Fig. 6, J and K, arrows), possibly in inflammatory mononuclear cells.

Fig. 6.

Hematopoietic stem cell (HSC) numbers increased in blood and lungs of pulmonary arterial hypertension (PAH) patients. The leukocyte-enriched cell fraction was isolated from the blood and HSC numbers were analyzed by flow cytometry. A and B: percentages of CD34⁺ and CD117⁺ cells of the total leukocyte fraction are shown. Values are means ± SE. *P < 0.05, PAH (n = 8; female = 6) vs. healthy subjects (Ctrl, control; n = 5; male = 2 and female = 3). C: increased CD34⁺ cells positively correlated with pulmonary vascular resistance (PVR). D–K: normal (D, E, H, and I) and IPAH (F, G, J, and K) lungs stained for CD133 (D–G) and CD117 (H–K). CD133 expression is seen in plexiform (plx) lesion (arrow, F and G), possibly including endothelial cells (F) or smooth muscle cells (G). Note lack of expression in and around normal pulmonary arteries (v). CD117 expression is seen in plexiform lesions (J and K, arrows), possibly in inflammatory/mononuclear cells. WBCs, white blood cells.

Table 3.

Pulmonary hypertensive patient demographics and hemodynamics

ID	Age, yr	Sex	PH Type	Medication	RA, mmHg	PAP Sys, mmHg	PAP Diast, mmHg	PAP Mean, mmHg	PCWP, mmHg	TPG, mmHg	PVR, Woods unit	CO, L/min	CI, L·min−1·m−2
A1	75	F	Scleroderma	IV remodulin 20 mg; riociguat 1 mg TID; ambrisentan 10 mg daily; plaquenil 200 mg BID; medrol 12 mg daily	14	58	21	38	21	17	3.2	6.7	3.2
A2	56	F	Scleroderma	IV remodulin 40 mg, riociguat 2.5 mg TID	12	88	33	56	9	47	7.7	8.6	4.1
A3	59	F	Scleroderma	IV remodulin 40 mg, ambrisentan 10 mg daily, tadalafil 40 mg daily	8	55	13	34	11	23	6.7	3.4	1.9
A4	60	F	Scleroderma	Tadalafil 40 mg daily, macitentan 10 mg daily, selexipag 400 µg BID	5	60	21	36	11	25	7.4	3.4	2.3
A5	75	F	Scleroderma	IV remodulin 20 mg, riociguat 2.5 mg TID, ambrisentan 10 mg daily	2	86	28	48	2	49	17.6	2.8	1.7
B1	54	F	Idiopathic	IV remodulin 72 mg, riociguat 2.5 mg TID, macitentan 10 mg daily	3	76	24	44	6	38	8.6	4.4	2.5

BID, twice a day; CO, cardiac output; CI, cardiac index; F, female; ID; patient identification code; IV, intravenous; M, male; PH type, pulmonary hypertension type; PAP diast, pulmonary artery pressure in diastole; PAP mean, pulmonary artery pressure mean; PAP sys, pulmonary artery pressure in systole; PCWP, pulmonary capillary wedge pressure; PVR, pulmonary vascular resistance; RA, right atrial pressure; TID, three times a day; TPG, transpulmonary gradient.

Bone marrow-derived stem cells augmented α-adrenergic receptor-mediated contraction of pulmonary arteries cultured in hypoxia.

As CD34⁺ positively correlated with PVR in PAH patients, we postulated that the stem cells potentially contribute to constrict pulmonary arteries. To test this hypothesis, we determined whether these cells influence the function of the pulmonary artery. To do so, we isolated intralobar pulmonary artery and bone marrow cells and cultured the pulmonary artery in the absence or presence of bone marrow-derived stem cells (BMDSCs) in 21% (normoxia) or 3% (hypoxia) O₂ (Fig. 7A). After 48 h, we measured pulmonary artery contraction. There was no difference between KCl (60 mM)-induced contraction of pulmonary arteries cultured with or without BMDSCs in normoxia or hypoxia (Fig. 7B). However, α-adrenergic receptor-mediated contraction of the pulmonary arteries significantly increased when arteries were cultured with BMDSCs, compared without BMDSCs, in hypoxia, but not in normoxia (Fig. 7, C and D). The inhibitor of Rho kinase, Y-27632 (10 μM), abolished α-agonist-induced contraction of pulmonary artery cultured with and without BMDSCs in hypoxia and normoxia.

Fig. 7.

Bone marrow cells augmented α-agonist-induced contraction of the isolated pulmonary artery. A: intralobar pulmonary artery isolated from normoxic (NX) Sv129J mice were incubated with and without bone marrow-derived stem cells (BMDSCs) from the same mice for 48 h in normoxia (21% O₂) or hypoxia (HX; 3% O₂). B: force generated by the pulmonary artery in response to KCl (30 mM) is shown. C and D: pulmonary artery incubated with BMDSCs (+BMDSCs) in hypoxia (C), but not in normoxia (D), contracted more to α-agonist, phenylephrine (10⁻⁹ to 10⁻⁵ M), than the arteries incubated without BMDSCs (−BMDSCs). Rho kinase inhibitor, Y27632 (10 μM), pretreatment inhibited phenylephrine (10⁻⁹ to 10⁻⁵ M)-induced contraction of arteries cultured with and without BMDSCs. Values are means ± SE; n = 6 in +BMDSCs (male = 3 and female = 3) and n = 15 in −BMDSCs (male = 8 and female = 7) groups. DHP20, (3β,5α)-3,21-dihydroxypregnan-20-one; ROCKi, Rho kinase inhibitor. *P < 0.05 vs. −BMDSCs.

G6PD inhibition decreased the expression of genes in the WNT signaling pathway in hypoxic mice.

Our understanding of the mechanisms that increase HSCs in peripheral blood and lungs of PH animals or patients remains enigmatic. Since G6PD interacts with WTN (14) and influences the biology of CD133⁺ cells (15), we sought to determine whether the expression levels of the genes related to WNT signaling pathways were altered in the lungs of hypoxic versus normoxic (control) and hypoxic + (3β,5α)-3,21-dihydroxypregnan-20-one (DHP20), a novel and potent G6PD inhibitor [IC₅₀ = 0.9 μM (31)], versus hypoxic mice. WNT-Frizzled (FZD) and BMPR signaling are well known to regulate the biology of HSCs, including CD133⁺ cells (48, 49, 60, 63, 69). By RNA-seq analysis, we found that the expression of genes that encode for canonical and noncanonical WNT proteins, FZD, and coreceptors of FZD, flamingo (CELSR2) and low-density lipoprotein receptor-related protein (LRP3), were decreased in the lungs of hypoxic mice treated with DHP20 compared with hypoxic mice (Fig. 8A and Table 4). Whereas these genes were increased more than twofold in hypoxic compared with normoxic lungs.

Fig. 8.

Glucose-6-phosphate dehydrogenase (G6PD) inhibition decreased expression of genes encoding proteins involved in the WNT-Fzd pathway and stromal cell-derived factor 1 (Cxcl12) and increases expression of bone morphogenetic protein receptor type-1A (Bmpr1a). Mice (male and female; 25–30 g) were exposed either to hypoxia (Hx) for 5 wk or to hypoxia + G6PD inhibitor applied in the last week of 5 wk of hypoxia. Lungs from hypoxic and normoxic (Nx) mice were isolated and gene expression was determined by whole-genome RNAseq analysis. A: bar graph showing the expression of genes that encode WNT-FZD pathways, Cxcl12 and Cxcr4, determined by RNAseq (also see Table 4), in lungs of hypoxic mice (n = 4; male = 2 and female = 2) vs. normoxic mice (n = 4; male = 2 and female = 2) and hypoxic+G6PD inhibitor [(3β,5α)-3,21-dihydroxypregnan-20-one (DHP20)] vs. hypoxic mice (n = 4; male = 2 and female = 2). B: the expression of Cxcl12, determined by RT-quantitative PCR, in lungs of hypoxic mice compared with hypoxic+DHP20 and normoxic mice is shown. C: bone marrow cells (BMCs) were isolated from normoxic mice and cultured in vitro with or without DHP20 for 24 h. D–I: in these cultured cells: the expression of Cxcl12 (D) and Bmpr1a (E); the number of CD34⁺ (F), CD117⁺ (G), and CD133⁺ (H) cells was determined by RT-quantitative PCR and flow cytometry, respectively; and CD133⁺ live and dead BMCs by methylene blue methods (I). ND, not detected. Values are means ± SE; n = 5 (male = 3 and female = 2) in each group. *P < 0.05 vs. without DHP20 group and scrambled-shRNA (control).

Table 4.

Expression of Wnt and Wnt receptor genes in lungs of hypoxic versus normoxic Sv129J mice and hypoxic versus hypoxic + DHP20 Sv129J mice

	Hypoxia vs. Normoxia		Hypoxia + DHP20 vs. Hypoxia
Gene	Log2 fold change	Adjusted P value	Log2 fold change	Adjusted P value
Lrp3	1.41696126	0.30426605	−5.228284	6.01E-07
Celsr2	2.56674022	NA	−0.8495691	0.03853362
Fzd1	−0.709158	NA	1.02193666	0.5674096
Fzd2	0.53100523	NA	−0.9511394	NA
Fzd4	−0.5060281	0.73072356	0.04122632	0.97359647
Fzd9	0.06427941	NA	−0.4634255	NA
Fzd10	0.54481509	NA	−1.4135922	NA
Fzd3	0.06788867	NA	−0.3593854	NA
Fzd5	−1.3490523	NA	2.16034937	0.2302355
Fzd6	2.42805298	NA	−0.123328	NA
Fzd7	−1.287704	NA	0.81290232	NA
Fzd8	−0.5284	NA	0.10526828	NA
Wnt2b	−0.557537	NA	−0.0004857	NA
Wnt2	−0.3649784	NA	0.64711843	0.74595818
Wnt3a	1.10615623	NA	−0.5531579	NA
Wnt3	0.25285309	NA	−0.4889798	NA
Wnt7b	0.3190979	NA	−1.17298	NA
Wnt7a	0.29985851	NA	−1.1425085	NA
Wnt9b	0.24943129	NA	−0.507423	NA
Wnt9a	−0.6305631	NA	0.64888628	NA
Wnt10b	0.22573042	NA	−1.4140793	NA
Wnt4	−0.4031157	NA	−0.110538	0.97032824
Wnt5b	1.02501424	NA	−0.768839	NA
Wnt5a	0.71650779	NA	0.01612577	0.98002449
Wnt11	1.87875103	NA	0.38844879	0.9090996
Cxcr4	2.43921605	NA	−1.320835	0.40323384
Cxcl12	1.38203347	0.04608311	−2.8434381	2.42E-07

DHP20, (3β,5α)-3,21-dihydroxypregnan-20-one; NA, not applicable.

G6PD inhibition decreased Cxcl12 and increased Bmpr1a expression and reduced HSC numbers in bone marrow cell culture.

In addition to decreasing the expression of genes involved in WNT signaling, we also found that G6PD inhibition with DHP20 in mice decreased expression of Cxcl12, a gene for stromal cell-derived factor 1 (SDF-1) (Fig. 8, A and B, and Table 4), and Cxcr4 gene, which encodes the CXCL12 receptor (Fig. 8A and Table 4). Cxcl12 and Cxcr4 are expressed in CD133⁺ cells (25, 33, 64). To determine whether G6PD inhibition regulated expression of these genes in isolated bone marrow cells, we extracted the bone marrow and cultured the bone marrow cells (Fig. 8C), as described previously (20). Application of DHP20 to the bone marrow cell culture for 24 h 1) decreased expression of Cxcl12 (Fig. 8D); 2) increased expression of Bmpr1a (Fig. 8E), a gene encoding BMPR1a that is known to regulate phenotype of stem cells in the bone marrow niche (37); and 3) decreased CD117⁺ and CD133⁺ cell numbers (Fig. 8, G and H). Notably, knockdown of G6PD decreased the live-to-dead cell ratio of CD133⁺ (Fig. 8I). DHP20 had no effect on CD34⁺ cell numbers (Fig. 8F).

G6PD inhibition decreased HSCs in the bone marrow and blood of hypoxic mice.

Since the application of DHP20 and shRNA to bone marrow cell culture decreased CD133⁺ and CD117⁺ cells (Fig. 8, G–I), we determined whether DHP20 treatment to hypoxic mice decreases HSCs in the bone marrow and blood. DHP20 (1.5 mg·kg⁻¹·day⁻¹) for 1 wk after we had exposed mice to 4 wk of hypoxia decreased CD34⁺, CD117⁺, CD34⁺CD117⁺, and CD34⁺CD133⁺ cells in bone marrow (Fig. 9A) and blood (Fig. 9B), and CD133⁺ cells in blood but not in bone marrow (Fig. 9, A and B).

Fig. 9.

Glucose-6-phosphate dehydrogenase (G6PD) inhibition and knockdown decreased hematopoietic stem cell (HSC) numbers in hypoxic (Hx) mice. A specific and potent G6PD inhibitor, (3β,5α)-3,21-dihydroxypregnan-20-one (DHP20), was applied to Sv129j mice (male and female; 25–30 g) for the last week of 5 wk of hypoxia. The leukocyte-enriched cell fraction was isolated from the bone marrow (BM) and blood of mice exposed to normoxia (Nx), hypoxia, and hypoxia+DHP20. HSC numbers were analyzed by flow cytometry. A and B: graphs show percentages of cells positive for each HSC marker in BM and blood, respectively, in each experimental group. n = 9 (normoxia; male = 4 and female = 5), n = 6 (hypoxia; male = 3 and female = 3), and n = 6 (hypoxia-DHP20; male = 3 and female = 3). C: lung sections (5 μm thickness, n = 4 in each group) of normoxic and hypoxic wild-type (WT), WT-DHP20-treated, and G6PD^def mice (male and female; 29–35 g) were prepared and stained by immunohistochemistry. WT mice and G6PD^def mice (n = 4; male = 2 and female = 2) were exposed to Hx for 5 wk. Representative images show actin (ACTA; green), DAPI (blue), and CD133 (red). D: summary data are the percentage of pulmonary arteries (PAs) expressing CD133⁺ cells in the perivascular region to ACTA⁺ pulmonary arteries. BMCs, bone marrow cells; ND, not detected; WBCs, white blood cells. Values are means ± SE. *P < 0.05 vs. normoxia and #P < 0.05 vs. hypoxia.

G6PD inhibition decreased accumulation of CD133⁺ on the abluminal surface of pulmonary arteries in hypoxic mice.

We previously found that increased G6PD activity by hypoxia influences CD133⁺ cell biology and their interaction with pulmonary artery SMCs in cell culture (in vitro) experiments (15). Therefore, we sought to determine whether the inhibition of G6PD with DHP20 decreased accumulation of CD133⁺ cells on the abluminal surface of pulmonary arteries in the lungs of hypoxic mice. We treated mice with DHP20 (1.5 mg·kg⁻¹·day⁻¹) for 1 wk after we had exposed them to 4 wk of hypoxia, at which time HSCs increased in lung, and PH was established. Subcutaneous injection of DHP20 to mice decreased CD133⁺ cells in lungs of hypoxic mice (Fig. 9, C and D). Consistently, we did not observe CD133⁺ cells in the lungs of G6PD^def compared with WT mice exposed to hypoxia for 5 wk (Fig. 9, C and D).

G6PD inhibition attenuated hypoxia-induced PH in mice.

We previously reported that knockdown of G6PD decreases the CD133⁺ cells cultured in hypoxia (15) and pulmonary artery pressure (14); therefore, in this study, we sought to determine whether application of DHP20 reduces α-adrenergic receptor-mediated contraction of pulmonary arteries cultured with and without BMDSCs in hypoxia. We found that application of DHP20 to pulmonary arteries cultured with BMDSCs in hypoxia, but not in normoxia, for 48 h significantly decreased the PhE-induced contraction of the pulmonary arteries (Fig. 10, A and B). BMDSCs did not affect acetylcholine-induced relaxation of pulmonary arteries precontracted with PhE (10 μM; Fig. 10, C and D). Although DHP20 slightly but significantly decreased the RV hypertrophy (Fig. 10E), it dramatically reduced LV (Fig. 10F) and large artery stiffness (Fig. 10G) and increased cardiac output (Fig. 10H) in comparison to placebo-treated hypoxic mice.

Fig. 10.

Inhibition of glucose-6-phosphate dehydrogenase (G6PD) activity with a novel inhibitor decreased bone marrow (BM) cells, augmented α-agonist-induced contraction of the isolated pulmonary artery, and improved cardiovascular function in hypoxic Sv129J mice. A and B: intralobar pulmonary artery isolated from mice (male and female; 25–30 g) were incubated with (+BMDSCs) and without BM-derived stem cells (−BMDSCs) for 48 h in normoxia (21% O₂) or hypoxia (3% O₂). A G6PD inhibitor, (3β,5α)-3,21-dihydroxypregnan-20-one (DHP20), blocked contraction augmented by BMDSCs in hypoxia but not in normoxia. Control (Ctrl) data are from Fig. 7, which demonstrate pulmonary artery incubated with BMDSCs in hypoxia (A), but not in normoxia (B), contracted more to α-agonist, phenylephrine (10⁻⁹ to 10⁻⁵ M), than the arteries incubated without BDMSCs. C and D: DHP20 did not affect acetylcholine induction of pulmonary arteries precontracted with phenylephrine (10⁻⁵ M) in hypoxia or normoxia. n = 6 (male = 3 and female = 3) in +BMDSCs and n = 15 (male = 8 and female = 7) in the −BMDSCs group. n = 6 (male = 3 and female = 3) in the ±BMDSC groups treated with DHP20. Sv129J mice were exposed to hypoxia (10% O₂) for 5 wk, and DHP20 was applied to mice for the last week of hypoxia. E–H: summary data of right ventricular (RV) hypertrophy (E), left ventricular (LV) stiffness (F), arterial stiffness (G), and cardiac output (H) are shown. Values are means ± SE; n = 8–10 (male = 4 to 5 and female = 4 to 5) in each group. *P < 0.05 vs. normoxia (control) and #P < 0.05 vs. hypoxia (pulmonary hypertension, PH). I: schematic illustrating potential contribution of the bone marrow stems cells (CD133⁺ and c-kit⁺, aka CD117⁺) to the progressive remodeling of pulmonary arteries in hypertensive mice. Ea, arterial elastance; HSC, hematopoietic stem cells; dP/dt_max, maximum change in pressure over time; dV/dt_max, maximum change in volume over time; S, septum.

DISCUSSION

PAH is a multicellular disease. Herein, we demonstrated that pluripotent HSCs (CD34⁺, CD117⁺, CD133⁺, CD34⁺CD117⁺, CD34⁺CD133⁺, and CD117⁺CD133⁺) increased in the bone marrow and blood of hypoxic mice and in the blood of idiopathic and scleroderma-associated PAH patients. Furthermore, our results demonstrated, to the best of our knowledge for the first time, an unexpected role of HSCs in augmenting the α-adrenergic receptor-mediated pulmonary artery contraction, and that HSCs contributed, at least partly, to the remodeling of pulmonary arteries in hypoxic mice. Intriguingly, inhibition of G6PD activity decreased the hypoxia-induced increase of HSC numbers and remodeling of pulmonary arteries, and BMDSCs augmented α-adrenergic receptor-mediated contraction of pulmonary arteries.

HSCs differentiate to hematopoietic/immunogenic cells and to endothelial cell lineage. Hence, we propose that the pluripotent HSCs increased in mice exposed to chronic hypoxia and in PAH patients, potentially to produce more red blood cells (to increase hematocrit), leukocytes (to immunogenic reaction), and endothelial progenitor cells (to mitigate endothelial cell damage or loss of endothelial cell function). However, these cells become maladaptive and contribute to the pathogenesis of PH. Plexiform and neointimal lesions are pathognomonic of PAH and are formed, at least partly, by pressure overload-mediated perivascular inflammation (1). Consistently, CD133⁺ increased in the blood of hypoxic Cyp2c44^−/− mice that develop very severe hypoxia-induced occlusive remodeling of pulmonary arteries and heart failure (32), and CD133⁺VEGFR-2⁺ increased in the blood of beagle dogs after treatment with monocrotaline that induces PH (70). In humans, CD34⁺CD133⁺, CD133⁺CD117⁺, CD133⁺CD45⁺, CD133⁺VEGFR-2⁺, and CD133⁺CD34⁺VEGFR-2⁺ cells are increased more so in the blood of severe iPAH patients than in healthy control subjects (5, 21, 24, 26, 44, 45, 64); CD133⁺CD45⁺CD34⁺ cells are increased in chronic obstructive pulmonary disease patients with PH, but not in those without PH (47). In contrast, CD133⁺CD34⁺VEGFR-2⁻, CD133⁺CXCR2⁺, CD133⁺CD45⁺CD34⁺, CD34⁺CD45⁺, and CD133⁺CD31⁺ cells either remain unchanged or are decreased in blood of iPAH and BMPR2 mutation-associated PAH patients, but not in chronic thromboembolic PH patients compared with the control subjects (26, 28, 64). These studies and our present findings taken together imply that changes in lineage-specific (i.e., leukocyte versus angiogenic) HSC levels depend on the etiology of the disease.

HSCs first decreased (at 2 wk) and then increased (between 3 and 5 wk) in the blood of mice exposed to hypoxia. Since these cells increased in the bone marrow between 3 and 5 wk of hypoxia, it is reasonable to imply that their decrease from the circulation triggered a feed-forward mechanism to produce and release them from the bone marrow to compensate for their loss in blood. In beagle dogs, CD133⁺VEGFR-2⁺ progenitor cells increase in blood with 48 h of monocrotaline injection (70). Since monocrotaline elicits endothelial cell damage and death within 48 h of monocrotaline application in vivo and in vitro settings (43, 67), it is, therefore, logical to suggest that HSCs are increased in blood by hypoxic and monocrotaline stimuli to repair the damaged endothelial cells or to replace the dead endothelial cells.

Herein, we demonstrated that there were significantly more CD133⁺ and CD133⁺CD11b⁺ cells in the bone marrow of Sv129J compared with C57BL/6 normoxic-control mice. Additionally, we found that 1) the expression of hematopoietic cell markers was less in the lungs of hypoxic C57BL/6 than in Sv129J mice compared with their respective normoxic control; 2) hypoxia-induced systemic arterial and LV stiffness was more severe in Sv129J than in C57BL/6 mice; and 3) more CD133⁺ cells in bone marrow, blood, and lungs correlated to an increase in arterial stiffness of hypoxic Sv129J mice. Consistently, others have found significant variability in immunity and phenotypic variance between Sv129J and C57BL/6 mice (50), and a hypobaric hypoxia-induced increase of RV pressure and hypertrophy is more severe in Sv129J than in C57BL/6 mice (61). Similarly, the increase of HSCs in the blood of PAH patients correlated with PVR (Fig. 6) and elevated pulmonary arterial pressure (5) and cardiac dysfunction (40). Therefore, these findings suggest that there is a causal relationship between the increase of HSCs and deterioration of cardiovascular function.

Allogeneic bone marrow transplantation experiments suggest GPF⁺ cells infiltrated the lungs and increased around the ACTA⁺ pulmonary arteries in hypoxic mice engrafted with GFP-bone marrow. CD133⁺ cells appear to encircle ACTA⁺ pulmonary arteries in the lungs of hypoxic mice. In addition, expression of CD133 and genes of HSC markers (Cd164 and Flt3) are increased in the lungs of hypoxic mice. Immunofluorescence findings, taken together with gene expression results, indicated that HSCs increased more in the lungs of hypoxic than in normoxic mice. Consistently, Yan and colleagues (71) showed that bone marrow-derived circulating cells contribute to remodeling of the pulmonary artery and inflammation in control recipient mice transplanted with bone marrow from BMPR2(R899X) mutant mice. CD133⁺ cells are increased in the abluminal area of plexiform lesions in Sugen/hypoxia/normoxia rats (15). In iPAH patients, CD133 expression is seen in plexiform lesion, possibly including endothelial cells or SMCs (Fig. 6, D–G). In the human lungs, CD133 is detected in type 2 pneumocytes (in the alveolar/septal junction), monocytes, and undifferentiated cells (in the adventitia of vessels) in iPAH and control subjects (26). Furthermore, CD133⁺ cells are found in the endothelium of plexiform lesions in individuals with familial-, idiopathic-, and congenital heart disease-associated PAH (64), and xenograft of CD133⁺ cells derived from iPAH patients in NOD-SCID mice elicits angio-proliferative pulmonary vascular remodeling and widespread in situ thrombosis (6). Nonetheless, whether circulating CD133⁺ cells contribute to the remodeling of pulmonary arteries remains an open question (26). There are CD34⁻CD133⁺ and CD34⁺CD133⁺ types of pluripotent HSCs (19, 30, 73). CD133⁺ cells, also known as cancer stem cells, were originally implicated in tumorigenesis in the brain and other organs (9, 55). Later studies found that CD133⁺ cells are precursors to immunogenic cells (4) and endothelial progenitor cells (29), which further differentiate to mature CD31⁺ or vWF⁺ endothelial cells. Our results suggest that increased bone marrow-derived CD133⁺ cells in blood and adventitia potentially contribute to the remodeling of pulmonary artery in hypoxic mice.

Pulmonary artery remodeling and constriction together increase PVR and pressure in various groups of PH (59). Interestingly, our findings suggest that pulmonary arteries cocultured with BMDSCs generated more force to α-adrenergic receptor agonist, but not KCl, than the arteries that were cocultured without BMDSCs, while pulmonary arteries cocultured without BMDSCs contracted less to α-adrenergic agonist in hypoxia than normoxia. This observation is not unprecedented, because a study has previously reported that pulmonary artery from hypoxic rat contract less to a α-adrenergic agonist due to a reduction in a postreceptor contraction pathway in pulmonary artery SMCs (42). Furthermore, in arteries cocultured with and without BMDSCs, we identified that α-agonist-induced force was mediated by Rho kinase-dependent signaling pathway, which is a major signaling pathway activated by α-agonist-induced G protein-coupled receptors (66). Therefore, our results suggest that BMDSCs prevented hypoxia-induced impairment of postreceptor signaling potentially by augmenting Rho kinase signaling in hypoxic pulmonary arteries. These findings, to the best of our knowledge for the first time, demonstrate that BMDSCs modulate the pulmonary artery contraction-relaxation function. Therefore, we propose that stem cell-derived constricting factor selectively augmented α-adrenergic receptor-mediated contraction of SMCs, which are the epicenter of vascular function. These findings will have wider implications beyond the pulmonary vascular bed, because BMDSCs will undoubtedly modulate the function of α-adrenergic receptors present on the vascular beds in systemic circulation and the heart, in which the sympathetic nervous system-α-adrenergic system critically regulates heart rate and contractility.

We have previously reported that SMCs chemotactically interact with CD133⁺ cells in ex vivo coculture experiments (15). CD133⁺-SMC interaction facilitates the transition of SMCs from differentiated to dedifferentiated phenotype, and dedifferentiated (hyperproliferative, proinflammatory, and extracellular matrix producing) SMCs contribute to remodeling of pulmonary arteries (14, 58). We and others previously suggested that HSCs or bone marrow-derived proangiogenic cells contribute to remodeling and increasing large-artery and arteriole stiffening in PH (11, 32). Our present results suggest that HSCs augmented α-adrenergic receptor-mediated contraction of pulmonary arteries and contributed to remodeling of pulmonary arteries in lungs of mice exposed to chronic hypoxia. Therefore, the contribution of pluripotent HSCs to the pathogenesis of different groups of PH diseases should not be ignored while developing new pharmacotherapeutics to treat PH and PAH.

Glycolysis plays a significant role in regulating the phenotypic fate of HSCs. Inhibition of G6PD by subcutaneous injection of DHP20 to hypoxic mice decreased CD133⁺ cells in bone marrow, blood, and lungs. Moreover, deficiency of G6PD prevented the hypoxia-induced accumulation of CD133⁺ cells around ACTA⁺ large (>100 μm) pulmonary arteries in G6PD^def mice. G6PD inhibition also alleviated hypoxia-elicited large-artery stiffness and decreased the α-agonist-induced contraction of pulmonary arteries amplified by BMDSCs in hypoxia. Inhibition of G6PD activity did not reduce the acetylcholine-induced relaxation, and, therefore, it is safe to suggest that it did not decrease the endothelial cell function or nitric oxide-mediated relaxation. Therefore, our findings suggest that increased G6PD activity is critical to increasing and releasing CD133⁺ cells from the bone marrow to blood and the homing of these HSCs to the hypoxic lungs where they potentially elicit contraction and remodeling of pulmonary arteries.

Consequentially, our obvious question was, how does G6PD control HSC biology? Inhibition of G6PD with DHP20 decreased expression of various Wnt and WNT-receptor genes (Lrp3, Ceslr2, and Fzd6/10) in hypoxic vascular tissues and lungs. WNT-FZD/LRP3/CELSR2 signaling is a critical regulator of HSC self-renewal and proliferation (48, 49, 60, 63, 69), and, therefore, WNT signaling protein expression decreased by G6PD inhibition would explain why HSCs decreased. In addition, inhibition of G6PD with DHP20 also decreased the Cxcl12 gene, which encodes SDF-1, in the lungs of hypoxic mice and bone marrow cells cultured in vitro. It is well known that CXCL12/SDF-1 modulates migration and fate of HSCs and is a potent chemotactic for lymphocytes, mesenchymal stem cells, and endothelial progenitor cell. In addition, application of DHP20 to the bone marrow cell culture increased the expression of the Bmpr1a gene. BMPR1A regulates HSC proliferation and differentiation by influencing the osteoblasts in the stem cell niche, and conditional knockdown of Bmpr1a gene increases the number of HSC-supporting osteoblastic cells lining the bone surface that lead to an increase in the numbers of HSCs (12, 74). Since expression of BMPR1A is decreased in most forms of PH (7, 22), it is reasonable to assume that this could be one of the causes for increasing HSC levels observed in PAH patients and hypoxic mice. Our results suggest that inhibition of G6PD activity, at least partly, reduced the hypoxia-induced increase of HSCs via regulating WNT-FZD/LRP3/CELSR2-, BMPR1A-, and CXCL12-dependent signaling pathways that control proliferation and differentiation of HSCs, respectively.

In summary, the salient findings in this study demonstrated unexpected function for bone marrow-derived HSCs in augmenting α-adrenergic receptor-mediated constriction of pulmonary arteries that contribute to increase PVR in PAH patients and hypoxic mice. Furthermore, these results demonstrated that G6PD inhibition and deficiency are beneficial in alleviating HSCs’ numbers, α-adrenergic receptor-mediated contraction of pulmonary arteries amplified by BMDSCs, CD133⁺ cells-mediated pulmonary artery remodeling, and large-artery and LV stiffness in hypoxic mice.

GRANTS

This study was supported by the SENSHIN Medical Research Foundation, Osaka, Japan (R. Hashimoto), National Heart, Lung, and Blood Institute Grant RO1 HL132574 (S. A. Gupte), American Heart Association Grant-in-Aid 17GRNT33670454 (S. A. Gupte), Intramural Funding from Cardiovascular Translational Institute (S. A. Gupte, G. M. Lanier, and M. S. Wolin). I. Waddell and A. Jordon were wholly funded by Cancer Research UK (Grants C480/A1141 and C5759/A17098).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.H., G.M.L., I.F.M., and S.A.G. conceived and designed research; R.H., V.D., S.R.J., and R.M.T. performed experiments; R.H., V.D., S.R.J., R.M.T., and S.A.G. analyzed data; R.H., S.R.J., A.J., I.W., R.M.T., K.R.S., M.S.W., I.F.M., and S.A.G. interpreted results of experiments; R.H., V.D., S.R.J., and S.A.G. prepared figures; R.H., A.J., and S.A.G. drafted manuscript; R.H., G.M.L., A.J., I.W., K.R.S., M.S.W., I.F.M., and S.A.G. edited and revised manuscript; R.H., G.M.L., V.D., S.R.J., I.W., R.M.T., K.R.S., M.S.W., I.F.M., and S.A.G. approved final version of manuscript.