Regulation of fibroblast growth factor-2 expression in pulmonary arterial smooth muscle cells involves increased reactive oxygen species generation

Abstract

We have previously demonstrated increased fibroblast growth factor-2 (FGF-2) expression in a lamb model of increased pulmonary blood flow secondary to congenital heart disease, which may contribute to the associated increases in pulmonary arterial muscularization. However, the mechanisms underlying these increases in FGF-2 expression remain to be identified. Initially, we found that exogenous FGF-2 increased endogenous FGF-2 promoter activity and protein levels in ovine pulmonary arterial smooth muscle cells (PASMC). Furthermore, we found that these increases in FGF-2 expression were mediated by increases in superoxide levels via NADPH oxidase activation. In addition, FGF-2-mediated increases in FGF-2 expression and PASMC proliferation were attenuated by inhibition of phosphatidylinositol 3-kinase, Akt, and NADPH oxidase. Increases in FGF-2 expression could be stimulated by other factors known to increase reactive oxygen species (ROS) signaling in PASMC (endothelin-1 and transforming growth factor-beta1), whereas antioxidants attenuated these increases. Deletion constructs localized the growth factor- and ROS-sensitive region within the proximal 103 bp of the FGF-2 promoter, and sequence analysis identified a putative hypoxia response element (HRE), a DNA binding site for the ROS-sensitive transcription factor hypoxia-inducible factor-1alpha (HIF-1alpha). Stabilization of HIF-1alpha increased FGF-2 promoter activity, whereas mutation of the putative HRE attenuated FGF-2-induced FGF-2 promoter activity. Furthermore, FGF-2 increased HIF-1alpha protein levels and consensus HRE promoter activity in PASMC via antioxidant-sensitive mechanisms. Thus we conclude that FGF-2 can stimulate its own expression in PASMC via NADPH oxidase-mediated activation of ROS-sensitive transcription factors, including HIF-1alpha. This positive feedback mechanism may contribute to pulmonary vascular remodeling associated with increased pulmonary blood flow.

Fig. 1.

Exogenous fibroblast growth factor-2 (FGF-2) increases FGF-2 expression in pulmonary artery smooth muscle cells (PASMC). Luciferase activity was determined for an 1,800-bp FGF-2 promoter fragment in PASMC exposed to 0–100 ng/ml FGF-2 for 24 h (A) or to 100 ng/ml FGF-2 for 0–48 h (B). Values were normalized to an internal Renilla luciferase control. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells (UTD). C: representative Western blot depicting increased endogenous FGF-2 protein in PASMC exposed to 100 ng/ml FGF-2 for 24 h. C, 10 ng of 17-kDa human recombinant FGF-2 protein control; M, molecular mass markers with sizes indicated at right. D: representative Western blot depicting increases in endogenous FGF-2 protein in PASMC exposed to 100 ng/ml FGF-2 for 24 h. Exogenous FGF-2 was removed following digestion with trypsin. Blots were also probed for β-actin. Migration of molecular mass markers is indicated at right. E: bands from Western blots were quantified using ImageStation software, and FGF-2 signals were normalized to β-actin. Values are means ± SD; n = 6. *P < 0.05 vs. untreated cells. F: representative fluorescent images from immunocytochemistry (ICC) experiments to detect changes in FGF-2 protein in PASMC exposed to 100 ng/ml FGF-2 for 24 h. Images depicting loss of signal when cell permeabilization was omitted with the corresponding 4′,6-diamidino-2-phenylinodole (DAPI) nuclear stains are presented. G: images from ICC experiments were quantified using MetaMorph software and expressed as fold change relative to UTD. Values are means ± SD; n = 5. *P < 0.05 vs. untreated cells.

Fig. 2.

FGF-2 increases superoxide production in PASMC. A: fluorescence microscopy to detect changes in dihydroethidium (DHE) fluorescence. PASMC were treated with 100 ng/ml FGF-2 with or without 10 μM apocynin (Apo) for 24 h before DHE treatment. Cells were also treated with 50 μM diethylthiocarbamate (DETC) or 1 mM paraquat (Para) for 24 h to increase intracellular superoxide. B: DHE fluorescence was quantified using MetaMorph software and expressed as fold UTD. Values are means ± SD; n = 3. *P < 0.05 vs. UTD. †P < 0.05 vs. FGF-2-treated cells. C: dose-dependent changes in DHE fluorescence in PASMC exposed to 0–1,000 ng/ml FGF-2 for 24 h. Values are means ± SD; n = 3. *P < 0.05 vs. untreated cells.

Fig. 3.

Effects of pharmacological inhibition on FGF-2-induced FGF-2 expression and cell proliferation in PASMC. PASMC exposed to 0 or 100 ng/ml FGF-2 for 24 h were coincubated with 10 μM LY-294002 (LY) and 200 nM wortmannin (Wort) to inhibit phosphatidylinositol 3-kinase (PI3 kinase), 10 μM Akt inhibitor (AktI), and 10 μM Apo to inhibit NADPH oxidase. A: FGF-2 promoter activity relative to untreated cells. Values were normalized to an internal Renilla luciferase control. B: data from ICC experiments to detect changes in FGF-2 protein relative to untreated cells. C: serum-starved PASMC were treated as above. After 72 h, the number of actively respiring PASMC was determined using a methylthiazoletetrazolium-based assay. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells. †P < 0.05 vs. FGF-2-treated cells.

Fig. 4.

Endothelin-1 (ET-1) and transforming growth factor-β1 (TGF-β1) increase FGF-2 expression in PASMC. A: luciferase activity was determined for an 1,800-bp FGF-2 promoter fragment in PASMC exposed to 100 nM ET-1 or 1 ng/ml TGF-β1 for 24 h. Values were normalized to an internal Renilla luciferase control. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells. B: representative Western blot depicting changes in FGF-2 protein in PASMC exposed to 100 nM ET-1 or 1 ng/ml TGF-β1 for 24 h. Blots were also probed for β-actin. U, untreated cells. C: bands from Western blots were quantified using ImageStation software, and FGF-2 signals were normalized to β-actin. Migration of molecular weight markers is indicated at right. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells. D: representative fluorescent images from ICC experiments to detect changes in FGF-2 protein in PASMC exposed to ET-1 or TGF-β1 for 24 h. E: cells were treated with 100 nM ET-1 or 1 ng/ml TGF-β1 and 0 or 10 μM Apo for 24 h. Images from ICC experiments were quantified using MetaMorph software and are expressed as fold change relative to untreated cells. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells. †P < 0.05 vs. growth factor-treated cells.

Fig. 5.

Increased intracellular superoxide stimulates FGF-2 expression in PASMC. A: increased promoter activity in PASMC treated with 0–1 mM Para or with 0–50 μM DETC for 24 h. Values were normalized to an internal Renilla luciferase control. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells. B: increased FGF-2 protein in PASMC treated with 1 mM Para or 50 μM DETC for 24 h detected by ICC. Images were quantified using MetaMorph software and are expressed as fold change relative to untreated cells. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells. C: increased FGF-2 promoter activity in response to 100 ng/ml FGF-2 was attenuated by cotreatment with 100 μM manganese(III) tetrakis (4-benzoic acid) porphyrin (MnTBAP). *P < 0.05 vs. untreated cells. †P < 0.05 vs. FGF-2-treated cells. D: increased FGF-2 promoter activity in PASMC treated with 100 ng/ml FGF-2, 100 nM ET-1, or 1 ng/ml TGF-β1 (0 NAC) was attenuated by cotreatment with 1 mM N-acetyl cysteine (NAC). Values are means ± SD; n = 3. *P < 0.05 vs. untreated cells. †P < 0.05 vs. growth factor-treated cells.

Fig. 6.

FGF-2 promoter deletion constructs are responsive to FGF-2 and ROS in PASMC. A: constructs containing promoter DNA from +314 to −973, −555, or −103 bp were transfected into PASMC and treated with 0 or 100 ng/ml FGF-2 for 24 h. Values for FGF-2 treatment are expressed as fold untreated for each deletion construct. B: cells transfected with the −103-bp construct were treated with 50 μM DETC or 1 mM Para for 24 h. C and D: cells were transfected with the wild-type −103-bp construct or with a −103-bp construct containing a 3-bp mutation in the putative hypoxia response element (HRE). Cells were exposed to 100 ng/ml FGF-2 or 100 μM DFO (C) or to 100 nM ET-1 or 1 ng/ml TGF-β1 (D) for 24 h. Luciferase values were normalized to an internal Renilla luciferase control. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells. †P < 0.05 vs. wild type.

Fig. 7.

FGF-2 increases hypoxia-inducible factor-1α (HIF-1α) protein and HRE promoter activity via an antioxidant-sensitive mechanism in PASMC. A: representative Western blot showing increased HIF-1α protein in PASMC treated with 100 ng/ml FGF-2 with or without 1 mM NAC for 24 h. A protein band migrating with a lower molecular mass than HIF-1α was routinely observed using this antibody but was not used for quantification. Blots were also probed for β-actin. Migration of molecular mass markers is indicated at right. B: bands from Western blots were quantified using ImageStation software, and HIF-1α signals were normalized to β-actin. Values are means ± SD; n = 6. *P < 0.05 vs. untreated cells. †P < 0.05 vs. FGF-2-treated cells. C: PASMC were transfected with a plasmid containing a consensus HRE sequence upstream of a luciferase reporter. Cells were treated with 100 mg/ml FGF-2 with or without 1 mM NAC or with 100 μM DFO alone for 24 h. Luciferase values were normalized to an internal Renilla luciferase control. Values are means ± SD; n = 4. *P < 0.05 vs. untreated cells. †P < 0.05 vs. FGF-2-treated cells.