Dual roles of endothelial FGF-2-FGFR1-PDGF-BB and perivascular FGF-2-FGFR2-PDGFRβ signaling pathways in tumor vascular remodeling

Abstract

Perivascular cells are important cellular components in the tumor microenvironment (TME) and they modulate vascular integrity, remodeling, stability, and functions. Here we show using mice models that FGF-2 is a potent pericyte-stimulating factor in tumors. Mechanistically, FGF-2 binds to FGFR2 to stimulate pericyte proliferation and orchestrates the PDGFRβ signaling for vascular recruitment. FGF-2 sensitizes the PDGFRβ signaling through increasing PDGFRβ levels in pericytes. To ensure activation of PDGFRβ, the FGF-2-FGFR1-siganling induces PDGF-BB and PDGF-DD, two ligands for PDGFRβ, in angiogenic endothelial cells. Thus, FGF-2 directly and indirectly stimulates pericyte proliferation and recruitment by modulating the PDGF-PDGFRβ signaling. Our study identifies a novel mechanism by which the FGF-2 and PDGF-BB collaboratively modulate perivascular cell coverage in tumor vessels, thus providing mechanistic insights of pericyte-endothelial cell interactions in TME and conceptual implications for treatment of cancers and other diseases by targeting the FGF-2-FGFR-pericyte axis.

Fig. 1. FGF-2-induced angiogenesis, pericyte recruitment, and tumor growth in vivo.

a Tumor microvessel and pericyte contents. CD31⁺ endothelial cell (red) and NG2⁺ pericyte (green) signals in FGF-2⁺ and FGF-2⁻ tumors. Bar = 50 μm. b Quantification of microvessel density, vascular coverage by pericytes, and NG2⁺ pericyte area (n = 7 random fields; n = 4 mice for each group). c CD31⁺ endothelial (red) and NG2⁺ pericyte (green) signals in matrigels containing FGF-2^- tumors with and without FGF-2 protein. Bar = 50 μm. d Quantification of vascular coverage by pericytes and NG2⁺ pericyte area in matrigels with and without FGF-2 protein (n = 7 random fields; n = 4 mice for each group). e Tumor growth rates of scrambled-shRNA and Fgf2-shRNA-transfected FGF-2⁺ tumors (n = 4–6 animals/group). f CD31⁺ endothelial (red) and NG2⁺ pericyte (green) contents in scrambled-shRNA and Fgf2-shRNA-transfected FGF-2⁺ tumors. Bar = 50 μm. g Quantification of microvascular density, vascular coverage by pericytes, and NG2⁺ pericyte area in scrambled-shRNA and Fgf2-shRNA-transfected FGF-2⁺ tumors (n = 7 random fields; n = 4 mice for each group). h Tumor growth rates of T241-vector and T241-FGF-2 fibrosarcomas (n = 6 animals/group). i CD31⁺ endothelial (red) and NG2⁺ pericyte (green) contents in T241-vector and -FGF-2 fibrosarcomas. Bar = 50 μm. j Quantification of microvessel density, vascular coverage by pericytes, and NG2⁺ pericyte area in T241-vector and T241–FGF-2 fibrosarcomas (n = 7 random fields; n = 4 mice for each group). Vessels and pericytes were visualized using whole mount staining. All data as means ± S.E.M. *P < 0.05, **P < 0.01 and ***P < 0.001

Fig. 2. Vascular perfusion and permeability in FGF-2⁺ and FGF-2^- tumors in vivo.

a, c, e CD31⁺ tumor vasculature (red) and perfusion of 2000-kDa dextran (green) in FGF-2⁺ and FGF⁻, scrambled-shRNA and Fgf2-shRNA-transfected FGF-2⁺, and T241-vector and T241-FGF-2 tumors. Yellow color indicates double positive signals and perfused vessels. Bar = 50 μm. b, d, f Quantification of blood perfusion in in FGF-2⁺ and FGF⁻, scrambled-shRNA and Fgf2-shRNA-transfected FGF-2⁺, and T241-vector and T241-FGF-2 tumors (n = 10 random fields; n = 3 mice for each group). g, i, k CD31⁺ tumor vasculature (red) and leakiness of 70-kDa dextran (green) in FGF-2⁺ and FGF⁻, scrambled-shRNA and Fgf2-shRNA-transfected FGF-2⁺, and T241-vector and T241-FGF-2 tumors. Arrowheads indicate extravagated dextran (green). Intravascular dextran molecules are in yellow color. Bar = 50 μm. h, j, l Quantification of vascular permeability of 70-kDa dextran in FGF-2⁺ and FGF⁻, scrambled-shRNA and Fgf2-shRNA-transfected FGF-2⁺, and T241-vector and T241-FGF-2 tumors (n = 10 random fields; n = 3 mice for each group). Images are shown using whole mount staining. All data as means ± S.E.M.; Student’s t test, *P < 0.05, **P < 0.01 and ***P < 0.001

Fig. 3. FGFRs and PDGFRs in pericyte recruitment.

a RT-PCR analysis of mRNA expression levels of FGF receptors in pericytes freshly isolated from T241-vector and T241–FGF-2 tumors using magnetic bead separation. Beta-actin serves as a control. b CD31⁺ endothelial (red) and NG2⁺ pericyte (green) signals in vehicle (VT)-, anti-FGFR1 neutralizing antibody-, anti-FGFR2 neutralizing antibody treated-, anti-FGFR3 neutralizing antibody-, BGJ398-, anti-PDGFRα neutralizing antibody-, anti-PDGFRβ neutralizing antibody-, and imatinib-treated FGF-2⁺ tumors. Arrowheads indicate pericyte-associated vessels. Images are presented using whole mount staining. Bar = 50 μm. c, d Quantification of NG2⁺ pericyte area versus the total CD31⁺ microvessels and vascular coverage. (n = 7 random fields; n = 4 mice for each group). All data as means ± S.E.M.; Student’s t test, *P < 0.05, **P < 0.01 and ***P < 0.001

Fig. 4. Interplay between the FGF-2–FGFR2 and PDGF-B–PDGFR signaling in perivascular functions.

a Ki67⁺ (red) and NG2⁺ pericyte (green) in T241-vector and T241–FGF-2 tumor tissue. Cell nuclei were counterstained with DAPI (blue). Allows point to Ki67⁺-NG2⁺ double positive pericytes. Bar = 50 μm. Images are obtained using immunohistochemistry. Quantification of percentages of Ki67⁺-NG2⁺ double positive pericytes tumors (n = 15 random fields; n = 4 mice for each group). b Ki67⁺ (red) and NG2⁺ pericyte (green) in vehicle (VT)-, BGJ398-, and imatinib-treated FGF-2⁺ tumor tissue. Allows point to Ki67⁺-NG2⁺ double positive pericytes. Withdrawal experimental settings were performed at day 6 after drug cessation. Images are acquired using immunohistochemistry. Bar = 50 μm. c Quantification of percentages of Ki67 ⁺-NG2⁺ double positive pericytes in VT-, BGJ398-, or imatinib-treated FGF-2⁺ tumors (n = 15 random fields; n = 4 mice for each group). d Pericyte in vitro proliferation after FGF-2 stimulation (n = 6 samples/group). e Pericyte in vitro migration after FGF-2 stimulation (n = 6 samples/group) P = 0.43. f Pericyte in vitro proliferation after PDGF-BB stimulation with or without FGF-2 pretreatment (n = 6 samples/group). g Pericyte in vitro migration after PDGF-BB stimulation with or without FGF-2 pretreatment (n = 6 samples/group). h FGF-2-induced proliferation of scrambled-siRNA-, Fgfr1-siRNA-, or Fgfr2-siRNA-transfected pericytes in vitro (n = 6 samples/group). i PDGF-BB-induced proliferation of scrambled-siRNA-, Fgfr1-siRNA-, or Fgfr2-siRNA-transfected pericytes with or without FGF-2 pretreatment in vitro (n = 6 samples/group). j PDGF-BB-induced migration of scrambled-siRNA-, Fgfr1-siRNA-, or Fgfr2-siRNA-transfected pericytes with or without FGF-2 pretreatment in vitro (n = 6 samples/group). All data as means ± S.E.M.; Student’s t test, *P < 0.05, **P < 0.01 and ***P < 0.001. All in vitro experiments were repeated at least twice

Fig. 5. FGF-2-induced sensitization of PDGF-BB–PDGFR signaling by enhancing PDGFRβ recycling.

a Total and phosphorylated PDGFRβ protein levels in PDGF-BB-stimulated pericytes with or without FGF-2 pretreatment. β-actin levels were used as standard loading controls. Data were presented as means ± S.E.M. (n = 3). b Time-course analysis of PDGFRβ protein levels in FGF-2-, PDGF-BB-, or FGF-2 plus PDGF-BB-stimulated pericytes. β-actin levels were used as standard loading controls; NT no treatment. c Time-course analysis of phosphorylated PDGFRβ protein levels in FGF-2, PDGF-BB-, or FGF-2 plus PDGF-BB-stimulated pericytes. β-actin levels were used as standard loading controls. NT no treatment. d PDGF-BB-stimulated and non-stimulated pericytes were stained with specific anti-PDGFRβ, Rab7, and Rab11 antibodies with or without FGF-2 pretreatment. Nuclei were counterstained with DAPI. Bar = 25 μm (left) and 15 μm (right). All data as standardized values. All experiments were repeated three times

Fig. 6. FGF-2-stimulated endothelial Pdgfb and Pdgfd expression in vivo and in vitro.

a Pdgfb and Pdgfd mRNA levels in CD31⁺ ECs freshly isolated from T241-vector or T241–FGF-2 tumors (n = 3 samples; n = 3 mice for each group). b Pdgfb and Pdgfd mRNA levels in CD31⁺ ECs freshly isolated from VT- or BGJ398-treated T241-vector and -FGF-2 tumors (n = 3 samples; n = 3 mice for each group). c Pdgfb and Pdgfd mRNA levels in cultivated CD31⁺ ECs in response to FGF-2 stimulation (n = 3 samples/group). d Pdgfb and Pdgfd mRNA levels in FGF-2-stimulated cultivated CD31⁺ ECs transfected with scrambled-RNA, siFgfr1-RNA, or siFgfr2-RNA (n = 3 samples/group). All data as means ± S.E.M.; Student’s t test, *P < 0.05, **P < 0.01 and ***P < 0.001. All in vitro experiments were repeated at least twice

Fig. 7. Mechanisms of FGF-2-induced dual effects on endothelial cells and pericytes on vascular remodeling.

a In healthy vasculatures, angiogenic endothelial cell-derived PDGF-BB promotes pericyte recruitment. b In FGF-2⁺ tumors, tumor-derived FGF-2 stimulates endothelial PDGF-BB and PDGF-DD production through the FGFR1 signaling pathway. In addition to endothelial function, FGF-2 stimulates pericyte proliferation and migration through the direct pericyte effect on the FGFR-2 signaling and indirect mechanism by sensitizing PDGFRβ signaling on pericytes. c The FGF-2-FGFR1 endothelial and FGF-2-FGFR2 pericyte signaling pathways in vascular remodeling. EC, endothelial cell; PC, pericyte; TC, tumor cell