PPARγ regulates hypoxia-induced Nox4 expression in human pulmonary artery smooth muscle cells through NF-κB

Abstract

NADPH oxidases are a major source of superoxide production in the vasculature. The constitutively active Nox4 subunit, which is selectively upregulated in the lungs of human subjects and experimental animals with pulmonary hypertension, is highly expressed in vascular wall cells. We demonstrated that rosiglitazone, a synthetic agonist of the peroxisome proliferator-activated receptor-γ (PPARγ), attenuated hypoxia-induced pulmonary hypertension, vascular remodeling, Nox4 induction, and reactive oxygen species generation in the mouse lung. The current study examined the molecular mechanisms involved in PPARγ-regulated, hypoxia-induced Nox4 expression in human pulmonary artery smooth muscle cells (HPASMC). Exposing HPASMC to 1% oxygen for 72 h increased Nox4 gene expression and H₂O₂ production, both of which were reduced by treatment with rosiglitazone during the last 24 h of hypoxia exposure or by treatment with small interfering RNA (siRNA) to Nox4. Hypoxia also increased HPASMC proliferation as well as the activity of a Nox4 promoter luciferase reporter, and these increases were attenuated by rosiglitazone. Chromatin immunoprecipitation assays demonstrated that hypoxia increased binding of the NF-κB subunit, p65, to the Nox4 promoter and that binding was attenuated by rosiglitazone treatment. The role of NF-κB in Nox4 regulation was further supported by demonstrating that overexpression of p65 stimulated Nox4 promoter activity, whereas siRNA to p50 or p65 attenuated hypoxic stimulation of Nox4 promoter activity. These results provide novel evidence for NF-κB-mediated stimulation of Nox4 expression in HPASMC that can be negatively regulated by PPARγ. These data provide new insights into potential mechanisms by which PPARγ activation inhibits Nox4 upregulation and the proliferation of cells in the pulmonary vascular wall to ameliorate pulmonary hypertension and vascular remodeling in response to hypoxia.

nadph oxidases are a major source of superoxide production in the vasculature that contributes to endothelial dysfunction and vascular cell proliferation (4, 19). In nonphagocytic cells, the catalytic moiety of NADPH oxidases is composed of one or more gp91^phox (Nox2) homologs, Nox1, -3, -4, or -5, Duox1, or Duox2 (27). These Nox homologs associate with the membrane-bound p22^phox subunit to generate reactive oxygen species (ROS). Nox4 is highly expressed in vascular wall cells including smooth muscle and endothelial cells (47). In contrast to the other Nox homologs, current evidence indicates that Nox4 is constitutively active (1), and increases in Nox4 mRNA levels increase Nox4 activity (45). Nox4 expression is increased by diverse stimuli (4) including E2F transcription factors, serum starvation (56), urokinase, plasminogen activator, angiotensin II, TGF-β₁, and TNF-α (28). In contrast, oscillatory shear stress (48), PPARγ activation (24), and cyclic strain (18) suppress Nox4 mRNA expression in endothelial cells. PPARγ activation also reduced Nox4 expression in vivo in diabetic vasculature (25) and in the lung following chronic hypoxia (39). The precise molecular mechanisms regulating Nox4 expression continue to be defined.

Mounting evidence demonstrates that NADPH oxidases participate in pulmonary hypertension pathogenesis. Vasoconstriction stimulated by acute hypoxia exposure (53) as well as chronic hypoxia-induced pulmonary arterial superoxide generation, medial wall thickness, and right ventricular pressure were attenuated in gp91^phox knockout mice (31). Nox4 was selectively increased in the pulmonary vasculature and lungs of hypoxia-exposed mice and in pulmonary vascular tissue from patients with pulmonary arterial hypertension (37). Hypoxia also upregulated Nox4 in pulmonary artery adventitial fibroblasts in vitro and in adventitial fibroblasts from patients with idiopathic pulmonary arterial hypertension (29). Increased NADPH oxidase expression and activity has also been reported in: 1) hypoxia-exposed pulmonary resistance arteries in newborn pig (10) as well as porcine (38) and mouse pulmonary artery (15); 2) the Ren2 rat expressing the mouse renin gene in extrarenal tissues, which causes angiotensin II-mediated NADPH oxidase activation (9); and 3) lambs with either shunt-induced increased pulmonary blood flow (20) or ligation of the ductus arteriosus in utero (6). Collectively, these reports establish that NADPH oxidases are associated with numerous models of pulmonary hypertension, suggesting they constitute a fundamental mechanism of pulmonary vascular dysfunction in response to diverse stimuli.

We (39) recently reported that treating mice with rosiglitazone, a synthetic agonist of the peroxisome proliferator-activated receptor-γ (PPARγ), attenuated hypoxia-induced pulmonary hypertension and vascular remodeling in the mouse and reduced hypoxic Nox4 induction and ROS generation in the lung. PPARγ is a ligand-activated transcription factor belonging to the nuclear hormone receptor superfamily. Activation of PPARγ is promoted by naturally occurring fatty acids and their metabolites and by synthetic ligands including the thiazolidinedione (TZD) class of antidiabetic medications (e.g., rosiglitazone, pioglitazone, or troglitazone) (42). Activation of PPARγ promotes heterodimerization with the retinoid receptor, RXR (11), and binding to PPAR response elements (PPRE) in the promoter region of responsive genes to modulate transcriptional activity (42). PPARs can also repress gene expression through transrepression mechanisms that continue to be defined (17).

PPARγ, found normally in pulmonary vascular endothelial and smooth muscle cells (2, 36), was reduced in the complex vascular lesions of patients with advanced primary or secondary pulmonary hypertension or in the lungs of rats with severe pulmonary hypertension (2). Similarly, targeted depletion of PPARγ from smooth muscle cells (22) or from endothelial cells (21) resulted in spontaneous pulmonary hypertension in mice. On the other hand, activation of PPARγ with TZDs attenuated pulmonary hypertension or vascular remodeling caused by monocrotaline (35) or hypobaric hypoxia (8) in rats. TZDs also reduced pulmonary hypertension in ApoE-deficient mice fed high-fat diets (23) and attenuated hypoxia-induced pulmonary hypertension and Nox4 expression and activity in mice (39). Therefore, the current study was designed to more carefully examine the molecular mechanisms by which PPARγ activation modulates Nox4 expression. Our results demonstrate that PPARγ regulates hypoxia-induced Nox4 expression, cell proliferation, and ROS generation in human pulmonary artery smooth muscle cells (HPASMC) by suppressing binding of NF-κB to the Nox4 promoter.

MATERIALS AND METHODS

Cell culture.

HPASMC (Lonza, Basel, Switzerland) monolayers were grown and maintained at 37°C in a 5% CO₂ atmosphere in SmGM-2 media (Lonza) containing 2% fetal calf serum, 10 ng/ml human epidermal growth factor, 1.0 mg/ml hydrocortisone, 12 mg/ml bovine brain extract, 50 mg/ml gentamicin, and 50 ng/ml amphotericin B. In selected studies, because of the ease with which they can be transfected, multipotent mesenchymal stem cells [C3H/10T1/2; American Type Culture Collection (ATCC), Manassas, VA] were maintained in basal modified Eagle's medium with 10% fetal bovine serum. HPASMC or C3H/10T1/2 cells were placed into a hypoxia chamber (1% O₂-5% CO₂; BioSpherix, Lacona, NY) or into a cell culture incubator (21% O₂-5% CO₂) maintained at 37°C for 72 h. During the final 24 h of this exposure, selected cells were treated with either vehicle (0.5% methyl cellulose) or rosiglitazone (10 μM). All manipulations of hypoxic cells were performed in a glove box (BioSpherix) that maintains oxygen (1%) and CO₂ (5%) levels to avoid effects of hypoxia and reoxygenation.

Measurements of H₂O₂ production and cell proliferation.

HPASMC were propagated into 24-well plates (4 × 10⁴ cells/well). After control or hypoxic exposure for 72 h and treatment with or without rosiglitazone, samples of HPASMC media (50 μl) were collected, and H₂O₂ concentrations were analyzed with the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Molecular Probes, Eugene, OR) following the manufacturer's recommendations. H₂O₂ concentrations were determined by measuring the horseradish peroxidase-catalyzed, H₂O₂-dependent oxidation of the nonfluorescent Amplex Red reagent to fluorescent resorufin red. Fluorescence measurements (excitation wavelength = 545 nm; emission wavelength = 590 nm) were recorded with a fluorescence plate reader (PerkinElmer, Waltham, MA) and converted to H₂O₂ concentrations by comparison with standard curves generated with known concentrations of H₂O₂. A quantitative colorimetric assay employing dimethylthiazol (MTT assay; ATCC) was used to measure HPASMC proliferation. Briefly, following exposure to control or hypoxic conditions with or without treatment with rosiglitazone, HPASMC were incubated with the MTT reagent for 4 h. The mitochondrial reductase in living cells reduces MTT to purple formazan, which is detected by spectrophotometry. Samples were analyzed using an ELISA plate reader at a wavelength of 570 nm. The values from treated cells were normalized to values from corresponding control cells.

RNA isolation, reverse transcription, and quantitative PCR.

Total RNA was isolated from cultured HPASMC using RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions, and RNA was quantified by NanoDrop spectrophotometry (Thermo Scientific, Wilmington, DE). cDNA was prepared using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). Quantitative PCR was performed to assess expression of Nox4 and GAPDH using primers that were designed based on human mRNA sequences. Nox4 primers were forward 5′-ggtta aacac ctctg cctgt tc-3′ and reverse 5′-cttgg aacct tctgt gatcc tc-3′, and GAPDH primers were forward 5′-cctgt tcgac agtca gccg-3′ and reverse 5′-cgacc aaatc cgttg actcc-3′. Real-time PCR was performed using iQ SYBR Green Supermix and the iCycler Real-Time PCR Detection System (Bio-Rad). Amplicon expression in each sample was normalized to its GAPDH RNA content. The relative abundance of target mRNA in each sample was calculated using ΔΔC_T methods as described by Applied Biosystems (User Bulletin no. 2).

Western blot analysis.

HPASMC were isolated, and equivalent amounts of protein were resolved by SDS-PAGE and immunoblotted with Nox4 rabbit polyclonal antibody (provided by Dr. David Lambeth, Emory University) or CDK4 rabbit polyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) as previously reported (39). Relative levels of immunoreactive proteins were quantified using the ChemiDoc XRS imaging system and Quantity One software (Bio-Rad).

Nox4 reporter construct, transient transfection techniques, and luciferase assays.

The human Nox4 promoter spans nucleotides −718 to +3 of the Nox4 gene locus relative to the position (+1) of the initiation methionine for the Nox4 open reading frame (PubMed Gene ID 50507). The sequence was amplified from human genomic bacterial artificial chromosome clone RP11-735I13 (BACPAC Resources, Oakland, CA) by PCR using the following 5′ and 3′ primers: 5′-aacct cgagt cccct agagc cccta agaa-3′ and 5′-ggtaa gctta ggacc gaggg tcaaa gact-3′, respectively. The resulting PCR product was digested with Hind III and XhoI, inserted into the pGL4.10-basic luciferase reporter vector (Promega, Madison, WI), and confirmed by automated sequencing. To examine the role of NF-κB in Nox4 expression in HPASMC, expression vectors for the NF-κB subunits p50 and p65 were employed in transient transfection studies as described previously (32). The pcDNA3 cylindromatosis (CYLD) construct, the negative regulator of NF-κB, was a generous gift from Dr. René Bernards (The Netherlands Cancer Institute, Amsterdam, The Netherlands). Because of ease of transfection, C3H/10T1/2 cells were used for selected transfection studies as indicated. The cells were plated in 24-well culture plates (5 × 10⁴ cells/well), incubated for 24 h, and washed with PBS, and then fresh growth medium was added before addition of transfection complexes. Cells were transiently transfected according to the manufacturer's instructions by incubating for 16 h with Effectene (10 μl/well; Qiagen) and with the luciferase reporter constructs for the pGL4.10 human Nox4 promoter. To control for transfection efficiency, cells were cotransfected with 0.1 μg/well pRL-TK Renilla luciferase or pmaxGFP construct (Lonza). After transfection, cells were incubated for 72 h in either control or hypoxic conditions with or without treatment with rosiglitazone for the last 24 h. Cells were then washed twice with PBS and collected into passive lysis buffer (300 μl; Promega). Luciferase activities were measured with the Luciferase Reporter Assay System (Promega) using a luminometer (PerkinElmer). Relative light units were normalized to Renilla or green fluorescent protein (GFP) activity, and all conditions were examined in triplicate.

Gene silencing with siRNA.

Nox4 gene expression was reduced using Nox4 or control small interfering RNA (siRNA; Qiagen), and NF-κB p50 or p65 subunits were reduced using siRNA from Santa Cruz Biotechnology. C3H/10T1/2 cells were incubated for 24 h in antibiotic-free medium containing 10% serum before incubating with the transfection reagent, Oligofectamine (Invitrogen, Carlsbad, CA), for 48 h following the manufacturer's recommendations. For cotransfection studies, Attractene (Qiagen) was employed as the transfection agent following the manufacturer's recommendations, and empty vector pcDNA3 and control siRNA were used as experimental controls. Preliminary studies using these siRNA techniques for Nox4, p50, and p65 confirmed that, compared with control siRNA, target mRNA levels were reduced by at least 50% measured using real-time PCR.

Chromatin immunoprecipitation assays.

Chromatin immunoprecipitation (ChIP) assays were performed using reagents and protocols from Upstate Biotechnology (Lake Placid, NY). Briefly, HPASMC were grown to 90% confluence, and approximately 1–2 × 10⁷ cells were employed for each experiment. After exposure to control or hypoxic conditions and treatment with vehicle or rosiglitazone, cells were treated with 1% formaldehyde for 15 min, harvested, suspended in SDS-lysis buffer, and sonicated. Following centrifugation, supernatants were collected, diluted, and precleared with salmon sperm-saturated protein A (Zymed, San Francisco, CA) to remove nonspecific immunoglobulins. Immunoprecipitation was performed by adding 2 μg of antibodies to p65 (Santa Cruz Biotechnology) or by adding nonimmune IgG (negative control). Immune complexes were then captured with 40 μl of salmon sperm DNA-saturated protein A and washed. A total of 100 μl 10% Chelex (10 g/100 ml H₂O) was added directly to the washed protein A beads and vortexed. After boiling, the Chelex-protein A bead suspension was incubated with proteinase K (100 μg/ml), the suspensions were centrifuged, and the supernatants were collected and used directly as a template in PCR reactions. The recovered DNA was purified with a DNA cleanup kit (Qiagen) and subjected to real-time PCR conditions. The primers used for putative NF-κB p65 binding element on the Nox4 promoter were forward 5′-gcttt agttt gggag tggga-3′ and reverse 5′-gaaat ttgag ccgga aacag-3′. Nox4 primers used to show binding specificity (negative control) were forward 5′-ggtta aacac ctctg cctgt tc-3′ and reverse 5′-cttgg aacct tctgt gatcc tc-3′. The results were corrected by adjusting to equal amounts of starting material (input DNA).

Statistical analysis.

All data were analyzed using ANOVA. Post hoc analysis using the Student-Newman-Keuls test was employed to detect differences between specific groups. In studies comparing only two experimental groups, data were analyzed with Student's t-test to determine the significance of treatment effects. The level of statistical significance was taken as P < 0.05.

RESULTS

Activation of PPARγ with rosiglitazone attenuates hypoxia-induced increases in HPASMC H₂O₂ production, Nox4 expression, and proliferation.

Because rosiglitazone attenuated hypoxia-induced Nox4 expression as well as pulmonary hypertension and muscularization of small pulmonary arterioles in the mouse lung (39), the current study examined rosiglitazone-mediated regulation of hypoxia-induced alterations in Nox4 in HPASMC. HPASMC were exposed to either 21% O₂ or 1% O₂ for 72 h. During the last 24 h of these exposures, cells were treated with 10 μM rosiglitazone or with an equivalent volume of vehicle. Compared with exposure to control conditions, exposure to hypoxia significantly increased Nox4 mRNA levels, and treatment with rosiglitazone decreased Nox4 expression in both control and hypoxia-exposed HPASMC (Fig. 1A). As demonstrated in Fig. 1B, hypoxia caused a sixfold increase in the production of H₂O₂ by HPASMC that was attenuated by treatment with rosiglitazone. As shown in Fig. 1C, hypoxia-induced increases in Nox4 expression and H₂O₂ production were associated with increased HPASMC proliferation, and these alterations were attenuated by treatment with rosiglitazone.

Fig. 1.

Rosiglitazone (Rosi) attenuates hypoxia-induced human pulmonary artery smooth muscle cell (HPASMC) Nox4 expression, H₂O₂ generation, and proliferation. HPASMC were exposed for 72 h to control (21% O₂; C) or hypoxic (1% O₂; H) conditions, and during the last 24 h of exposure, selected cells were treated with rosiglitazone (10 μM) or an equivalent volume of vehicle. In A, HPASMC were collected, and RNA was isolated and reverse-transcribed to analyze mRNA levels using Nox4 and GAPDH primers. Cells treated without and with methylcellulose vehicle (M) were examined. Each bar represents the mean ± SE copies of Nox4 mRNA normalized to copies of GAPDH expressed as fold change relative to control from 2 experiments performed in triplicate. *P < 0.05 vs. C; **P < 0.05 vs. H. In B, HPASMC media were collected and subjected to assays of H₂O₂ concentration. Each bar represents the mean ± SE H₂O₂ concentration from 2 experiments, each performed in triplicate. In C, data from 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell proliferation assays were recorded in absorbance units. Each bar represents HPASMC proliferation as the mean ± SE fold change relative to control. n = 3. *P < 0.05 for H vs. C; **P < 0.05 for H+Rosi vs. H.

Nox4 contributes to hypoxia-induced H₂O₂ production and HPASMC proliferation.

To assess the role of Nox4 in hypoxia-induced ROS generation, selected HPASMC were treated with Nox4 siRNA before exposure to control or hypoxic conditions. As shown in Fig. 2A, Nox4 siRNA caused effective reduction of Nox4 mRNA and protein levels in HPASMC and had little effect on H₂O₂ production in HPASMC exposed to control conditions. In contrast, Nox4 siRNA significantly reduced H₂O₂ production in HPASMC exposed to hypoxic conditions. As shown in Fig. 2B, knockdown of Nox4 significantly attenuated hypoxia-induced HPASMC proliferation.

Fig. 2.

Nox4 RNA interference reduced hypoxia-induced H₂O₂ generation and HPASMC proliferation. HPASMC were transfected with control scrambled (siC) or Nox4 siRNA (siNox4) and subsequently exposed for 72 h to control (21% O₂) or hypoxic (1% O₂) conditions, and during the last 24 h of exposure, selected cells were treated with rosiglitazone (10 μM) or an equivalent volume of vehicle. HPASMC media were thereafter collected and subjected to assays of H₂O₂ concentration. In A, each bar represents the mean ± SE H₂O₂ concentration from 2 experiments, each performed in triplicate. *P < 0.05 for C+siNox4 vs. C+siC; **P < 0.05 for H+siC vs. C+siC; ***P < 0.05 for H+siNox4 vs. H+siC. Representative results from real-time PCR of Nox4 mRNA are presented (inset), demonstrating effective small interfering RNA (siRNA)-mediated knockdown of Nox4. Knockdown of Nox4 in protein levels by siNox4 is also shown in the representative Western blot under the graph. In B, data from MTT cell proliferation assays were recorded in absorbance units. Each bar represents HPASMC proliferation as the mean ± SE fold change relative to control from 2 experiments performed in quadruplicate. *P < 0.05 for H+siC vs. C+siC; **P < 0.05 for H+siNox4 vs. H+siC. CDK4, cyclin-dependent kinase 4.

Analysis of the human Nox4 promoter and its regulation by hypoxia, NF-κB, and PPARγ.

To further examine the regulation of Nox4 expression during hypoxia, a map of homologous transcription factor binding sites was generated using MatInspector (Genomatix, Munich, Germany) to assess potential regulatory sites in the Nox4 promoter. Figure 3 demonstrates that the Nox4 promoter contains putative homologous binding sites for factors known to undergo hypoxic regulation including PPRE, nuclear respiratory factor-1 (NRF-1), forkhead domain factors (FKHD), hypoxia response element (HRE), and NF-κB. To further investigate this promoter, a 959-bp fragment of the proximal Nox4 promoter including a portion of the 5′-untranslated RNA was amplified from a bacterial artificial chromosome (RP11-735I13) containing human genomic Nox4 sequenced for confirmation and cloned into the XhoI and Hind III sites of the pGL4.10-basic luciferase reporter. The effect of hypoxia on Nox4 promoter activity was then measured using the primitive mesenchymal cell line, C3H/10T1/2. As demonstrated in Fig. 4A, hypoxic exposure caused significant upregulation of Nox4 promoter activity in C3H/10T1/2 cells. This finding was confirmed in other cell lines including COS-7 and HEK 293 cells (data not shown). Treatment with rosiglitazone during the last 24 h of hypoxia exposure reduced Nox4 promoter activity to levels comparable with those observed in control cells.

Fig. 3.

Map of the human Nox4 promoter sequence −718 to +241 bp showing predicted binding sites for transcription factors based on homology with known consensus sequences. Homologous binding sequences were determined at >90% probability using MatInspector (Genomatix, Munich, Germany). The base pairs are numbered relative to the Nox4 start site, shown as +1. The translation start site methionine is coded 238 bp downstream (data not shown). Response elements for selected transcription factors that may be activated in hypoxic conditions are labeled in the figure, including peroxisome proliferator-activated receptor (PPAR) response element (PPRE), forkhead domain factor (FKHD), hypoxia response element (HRE), nuclear respiratory factor-1 (NRF-1), and NF-κB p65 and c-rel sites. RXR, retinoid X receptor; HIF, hypoxia-inducible factor.

Fig. 4.

Hypoxia, rosiglitazone, and NF-κB regulate Nox4 promoter activity. C3H/10T1/2 cells were transfected with the human Nox4 promoter luciferase reporter construct along with Renilla as a control for transfection efficiency. In A, cells were exposed for 72 h to control (21% O₂) or hypoxic (1% O₂) conditions, and during the last 24 h of exposure, selected cells were treated with rosiglitazone (10 μM) or an equivalent volume of vehicle. Cells were then harvested and subjected to assays for luciferase activity. Each bar represents the mean ± SE luciferase activity in each sample relative to Renilla expressed in arbitrary units from 3 experiments performed in triplicate. *P < 0.05 vs. C; **P < 0.05 vs. H. In B, C3H/10T1/2 cells were transfected with the human Nox4 promoter luciferase reporter construct along with Renilla and with either pcDNA3 NF-κB-p50 or pcDNA3 NF-κB-p65 to overexpress NF-κB subunits or empty pcDNA3 vector (control). After 48 h, cells were harvested for luciferase activity assays. Each bar represents the mean ± SE luciferase activity relative to Renilla in each sample expressed in arbitrary units from 3 separate experiments. *P < 0.05 vs. C; **P < 0.05 vs. C. In C, C3H/10T1/2 cells were transfected with the human Nox4 promoter luciferase reporter construct along with a Renilla luciferase construct. The cells were also cotransfected with the NF-κB inhibitor, pcDNA3 cylindromatosis (CYLD), or empty pcDNA3 vector (vector control; VC) and exposed to hypoxic (1% O₂) conditions. After 48 h, the cells were harvested for luciferase activity assays. Each bar represents the mean ± SE luciferase activity relative to Renilla in each sample expressed in arbitrary units from 3 separate experiments. *P < 0.05 for H+CYLD vs. H+VC. In D, C3H/10T1/2 cells were transfected with the human Nox4 promoter luciferase reporter construct along with Renilla luciferase construct. The cells were also cotransfected with siRNA to p50 (si-p50) or p65 (si-p65) or with siC and exposed to hypoxic (1% O₂) conditions for 48 h. The cells were then harvested for luciferase activity assays. Each bar represents the mean ± SE luciferase activity relative to Renilla in each sample expressed in arbitrary units from 3 separate experiments. *P < 0.05 vs. H+siC.

Because in silico analysis of the Nox4 promoter identified potential NF-κB binding sites (Fig. 3), and because PPARγ inhibited NF-κB (57), Nox4 promoter activity was examined in C3H/10T1/2 cells following cotransfection with the expression vectors for the NF-κB subunit p50 or p65 along with the Nox4 promoter reporter. As illustrated in Fig. 4B, transfection with p50 or p65 increased Nox4 promoter activity 8- or 15-fold, respectively. To confirm the role of NF-κB in Nox4 promoter activation, C3H/10T1/2 cells were cotransfected with either the CYLD expression vector or empty vector. CYLD, a gene encoding a cytoplasmic protein with three cytoskeletal-associated protein-glycine-conserved (CAP-GLY) domains, functions as a deubiquitinating enzyme that inhibits TNF receptor-associated factor 2 (TRAF2) and NF-κB signaling pathways (34). As shown in Fig. 4C, hypoxia-induced Nox4 promoter activity was reduced to basal levels following inhibition of NF-κB with CYLD. As a complementary approach, introducing either siRNA against p50 (si-p50) or si-p65 to reduce NF-κB signaling also attenuated hypoxia-induced Nox4 promoter activity (Fig. 4D). Collectively, these findings suggested the involvement of NF-κB in hypoxia-induced activation of the Nox4 promoter.

Hypoxia stimulates NF-κB binding to the Nox4 promoter: inhibition by rosiglitazone.

To examine whether NF-κB binds the Nox4 promoter and whether hypoxia or PPARγ activation modulate that binding, ChIP assays were performed in HPASMC following exposure to control or hypoxic conditions and treatment with vehicle or rosiglitazone. Compared with control conditions, hypoxia stimulated binding of p65 to the Nox4 promoter, and treatment with rosiglitazone reduced the binding of p65 to the Nox4 promoter (Fig. 5).

Fig. 5.

NF-κB binding to the Nox4 promoter is stimulated by hypoxia and attenuated by rosiglitazone treatment. HPASMC were exposed for 72 h to control (21% O₂) or hypoxic (1% O₂) conditions, and during the last 24 h of exposure, selected cells were treated with rosiglitazone (10 μM) or an equivalent volume of vehicle. Chromatin immunoprecipitation (ChIP) assays were performed with antibodies against human NF-κB p65 or control IgG. Immunoprecipitated DNA was analyzed by real-time PCR using primers specific for the Nox4 promoter surrounding the putative binding sites. Each bar represents the mean relative NF-κB binding to the Nox4 promoter ± SE from 2 experiments performed in triplicate. *P < 0.05 vs. C; **P < 0.05 vs. H.

DISCUSSION

PPARγ, a member of the nuclear receptor superfamily of ligand-activated transcription factors, has attracted growing interest as a therapeutic target in systemic (13) and pulmonary (8, 23, 35, 39) vascular disease. The current clinical availability of the PPARγ-activating TZD class of insulin-sensitizing drugs, including rosiglitazone and pioglitazone, has stimulated numerous studies investigating the application of these medications to vascular disorders. A meta-analysis suggested potential adverse vascular effects of rosiglitazone in patients with type 2 diabetes (40), although a subsequent unplanned interim analysis of an industry-sponsored trial did not confirm these detrimental effects (51), and the Food and Drug Administration left rosiglitazone on market with additional warnings (41). In contrast, the incidence of cardiovascular events in patients with diabetes was lowered by treatment with pioglitazone (12, 14, 30, 54). Although the mechanisms for these vascular effects of TZD PPARγ ligands are presumed related to improvements in metabolic derangements of diabetic patients, the mechanisms for their effects on vascular disease in nondiabetic subjects continue to be defined.

We recently confirmed that hypoxia-induced pulmonary hypertension in mice was associated with enhanced expression of Nox4 NADPH oxidase in the lung (37) and reported that treatment with rosiglitazone attenuated Nox4 expression, pulmonary hypertension, and pulmonary vascular remodeling (39). The mechanisms underlying hypoxia-induced Nox4 expression in the lung and its contribution to the pathogenesis of pulmonary hypertension have not been defined. Previous reports have demonstrated that Nox4-derived ROS mediate hypoxic-induced pulmonary artery smooth muscle cell proliferation (49), and hypoxia-induced increases in smooth muscle cells surrounding small pulmonary vessels characterize hypoxia-induced pulmonary vascular alterations in the mouse (39). The current study reveals that hypoxia stimulates Nox4 expression, ROS production, and HPASMC proliferation and that rosiglitazone-mediated PPARγ activation attenuates hypoxia-induced Nox4 expression, H₂O₂ production, and HPASMC proliferation. siRNA to Nox4 only partially inhibited hypoxia-induced H₂O₂ production, suggesting less than complete Nox4 knockdown by siRNA as well as potential contribution of additional sources of ROS to hypoxia-induced H₂O₂ production. Nevertheless, Nox4 siRNA more fully inhibited hypoxia-induced HPASMC proliferation, suggesting that Nox4 may specifically contribute to HPASMC proliferation.

Several recent studies suggested that the transcription factor, NF-κB, was actively involved in hypoxic signaling (5, 46). NF-κB constitutes a family of transcription factors that is classically activated following stimulation with proinflammatory ligands, including cytokines, antigens, and bacteria. This complex signaling mechanism involves activation of the IKK complex, leading to IκB phosphorylation and proteasomal degradation. Hypoxic activation of NF-κB is in part due to decreased prolyl hydroxylase-dependent hydroxylation of IκB kinase B in the canonical pathway but may also involve other mechanisms including tyrosine phosphorylation of IκBα (50). In addition, IκBα, through direct interaction with NF-κB and hypoxia-inducible factor-1α (HIF-1α), may play a pivotal role in the cross talk between the molecular events that underlie hypoxic and inflammatory responses (46). Previous studies have also implicated NF-κB in the regulation of other NADPH oxidase subunits, including gp91^phox, p47^phox, and p22^phox (3, 16, 33, 37). Collectively, these reports suggest that NF-κB activation occurs during hypoxia and may contribute to enhanced NADPH oxidase expression. Previous reports have demonstrated that NF-κB stimulates NADPH oxidases in human phagocytes and vascular smooth muscle cells (3, 5, 44). However, little is known about the role of NF-κB in regulation of the Nox4 promoter under normal or hypoxic conditions.

To our knowledge, the present study is the first to report that there are NF-κB binding elements on the Nox4 promoter and that PPARγ regulates Nox4 expression through these binding elements. Our findings demonstrate that the NF-κB subunits p65 and p50 stimulate Nox4 promoter activity, whereas inhibition of NF-κB with siRNA for NF-κB p50 or p65 or with CYLD decreased hypoxia-driven Nox4 promoter activity. It has been established that PPARγ can physically bind to NF-κB p65 and suppress NF-κB (7, 43, 52, 55). More limited evidence suggests that NF-κB p65 and PPARγ are mutually repressive and antagonize the transcriptional activity of each other. This concept is consistent with our (26) recent finding that enhanced NF-κB activity was observed in the aortas of endothelial PPARγ knockout mice. As shown in Fig. 3, in silico analysis of the Nox4 promoter reveals several PPRE in the vicinity of NF-κB binding sites that could potentially regulate Nox4 promoter activity. We speculate that rosiglitazone stimulates PPARγ binding to PPRE sites on the Nox4 promoter, physically preventing NF-κB from binding to adjacent sites and thereby inhibiting Nox4 promoter activity. This postulate is consistent with ChIP assays demonstrating that hypoxia-induced increases in p65 binding to the Nox4 promoter were attenuated by rosiglitazone. Alternatively, PPARγ may interact with p65 before Nox4 promoter interaction (7). The specific mechanisms of these interactions remain areas of active investigation in our laboratories.

Taken together, our findings provide novel evidence for NF-κB-mediated stimulation of Nox4 expression in HPASMC that can be negatively regulated by TZD PPARγ ligands. Based on evidence that Nox4 participates in both hypoxia-induced pulmonary hypertension in mice and in human pulmonary arterial hypertension (37), these findings suggest that PPARγ may represent a new therapeutic target in pulmonary hypertension. Our results provide new insights into potential mechanisms by which PPARγ activation inhibits Nox4 upregulation and the proliferation of cells in the pulmonary vascular wall cell to ameliorate pulmonary hypertension and vascular remodeling in response to hypoxia. Additional in vivo studies will be needed to confirm these findings that may provide a basis for the development of novel treatment strategies for pulmonary hypertension.

GRANTS

This work was supported by funding from the Research Service of the Atlanta Veterans Affairs Medical Center (C. M. Hart and M. S. Nanes) and by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-074518 (C. M. Hart).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).