Comparison of vascular endothelial growth factor and fibroblast growth factor-2 in a swine model of endothelial dysfunction

Abstract

Objective: Growth factor based angiogenesis, with or without cell therapy, is a promising therapeutic modality for patients with coronary artery disease. We compared the relative efficacies of surgically delivered vascular endothelial growth factor (VEGF) and fibroblast growth factor-2 (FGF-2) in a swine model of hypercholesterolemia-induced endothelial dysfunction which captures many of the pathophysiologic abnormalities of human coronary disease.

Methods: Yucatan mini-swine (20-30 kg), fed a high cholesterol diet (total 20 weeks), underwent circumflex ameroid placement to create chronic myocardial ischemia, followed three weeks later by perivascular administration of VEGF (2 microg; n=6), FGF-2 (100 microg; n=6), or placebo (n=7) in the ischemic territory. Normocholesterolemic animals (n=7) served as controls. Four weeks later, endothelial function, collateral-dependent perfusion, as well as myocardial protein and mRNA levels of angiogenic mediators were assessed.

Results: Endothelial dysfunction was observed in all hypercholesterolemic animals as impaired microvessel relaxation in response to adenosine diphosphate and VEGF. VEGF administration improved baseline-adjusted collateral-dependent perfusion at rest (-0.03+/-0.05 vs -0.12+/-0.04, VEGF vs placebo, p=0.09), but FGF-2 delivery caused a significantly greater improvement in perfusion compared to either group (+0.15+/-0.03, p<0.05 vs HC-placebo and HC-VEGF) at rest. Molecular analysis revealed increased eNOS expression (135%+/-8%, p=0.03 vs placebo) in all growth factor treated animals and increased expression of FGF-2 receptor, FGFR1 (65+/-26%, p=0.04 vs placebo), in FGF-2 treated animals. No significant changes were demonstrated in other angiogenic mediators including Akt, Syndecan-4.

Conclusions: In the setting of hypercholesterolemic endothelial dysfunction, FGF-2 is more effective than VEGF at enhancing collateral-dependent perfusion and thus, may be a better candidate than VEGF for angiogenic therapy in patients with end-stage CAD.

Figure 1

Serum cholesterol levels were significantly elevated in animals fed a hypercholesterolemic diet (HC) compared to a normal diet (NORM) with no significant differences between the hypercholesterolemic groups. * - p < 0.001

Figure 2

Coronary microvessel relaxation responses (depicted as % relaxation of pre-constricted diameter) to (A, B) endothelium dependent (ADP, VEGF) and (C) endothelium independent (SNP) vasodilators in NORM, HC-Placebo, HC-VEGF, and HC-FGF animals. * - p < 0.001, ** - p < 0.01.

Figure 3

Baseline-adjusted circumflex territory perfusion was significantly reduced in HC-Placebo group compared to NORM, both at rest (A) and with pacing (B, * - p < 0.001). VEGF treatment was associated with a small increase in flow compared to placebo († - p = 0.09, ‡ - p = 0.10), but FGF treatment resulted in a significant increase in ischemic territory perfusion (** - p < 0.05 vs. HC-Placebo and HC-VEGF)

Figure 4

(A) VEGF and FGF-2 signaling pathways. (B) Expression of VEGF, its receptors, VEGFR1 and VEGFR2, and downstream mediators Akt and eNOS in placebo and VEGF treated animals. (C) Expression of FGF-2, its receptor, FGFR1, and downstream mediators, syndecan-4 and eNOS, in placebo and FGF-2 treated animals. (* - p < 0.05; VEGF–vascular endothelial growth factor, PI3K–Phosphatidylinositol-3 kinase, eNOS–endothelial nitric oxide synthase, FGF–fibroblast growth factor)