Dose-dependent response of FGF-2 for lymphangiogenesis

Abstract

Spatio-temporal studies on the growth of capillary blood vessels and capillary lymphatic vessels in tissue remodeling have suggested that lymphangiogenesis is angiogenesis-dependent. We revisited this concept by using fibroblast growth factor 2 (FGF-2) (80 ng) to stimulate the growth of both vessel types in the mouse cornea. When we lowered the dose of FGF-2 in the cornea 6.4-fold (12.5 ng), the primary response was lymphangiogenic. Further investigation revealed that vascular endothelial growth factor-C and -D are required for this apparent lymphangiogenic property of FGF-2, and when the small amount of accompanying angiogenesis was completely suppressed, lymphangiogenesis remained unaffected. Our findings demonstrate that there is a dose-dependent response of FGF-2 for lymphangiogenesis, and lymphangiogenesis can occur in the absence of a preexisting or developing vascular bed, i.e., in the absence of angiogenesis, in the mouse cornea.

Fig. 1.

FGF-2 stimulates corneal lymphangiogenesis. (A) In the traditional corneal assay, 80 ng of FGF-2 (P) stimulates blood vessel growth from the peripheral limbal vasculature (arrowhead). (B) The traditional assay is viewed under fluorescent microscopy after labeling blood vessels yellow–green and lymphatic vessels red (arrowhead). Sucralfate in the FGF-2 pellet (P) autofluoresces green. (C) At the opposite end of the cornea, only lymphatic vessels (arrows) sprout. (Inset) Limbal vessels in the control cornea. (D) Lowering the dose of FGF-2 pellet to 12.5 ng (P) and moving it farther from the limbus results in less angiogenesis, but lymphatic vessels still reach the pellet. (E) Corneal lymphatic vessels were morphologically different from blood vessels. In addition, corneal lymphatic vessels did not express CD34 (F) or Tie2 (G and H, arrowheads) but did express VEGFR-3 (I). (Scale bars, 0.5 mm in B–D, 50 μmin E and I, and 200 μmin F–H.)

Fig. 2.

HMBEC and HMLEC respond similarly to FGF-2 in vitro. (A and B) Monolayers of pure HMBEC and pure HMLEC had cobblestone morphology, but HMLEC had more cytoplasm and less distinct intercellular borders. (C) Commercial HMBEC were shown to be free of HMLEC by FACS using anti-human VEGFR-3 antibody. The left peak represents HMBEC incubated only with anti-mouse-FITC, and the right peak which resides within the area of the left peak represents HMBEC incubated with anti-VEGFR-3 antibody. (D) The purity of isolated HMLEC was confirmed by FACS using anti-VEGFR-3 antibody. The left peak represents cells incubated only with anti-mouse-FITC, and the right peak represents cells positively labeled with anti-VEGFR-3 antibody. (E) Both cell types bound ¹²⁵I-FGF-2 similarly. (F) Neither cell type migrated when stimulated with FGF-2 without serum. (G) Both cell types proliferated similarly in response to FGF-2. (Scale bar, 200 μm in A and B.) HSPG, heparan sulfate proteoglycan-bound fraction; CSR, cell surface receptor-bound fraction. Each point represents the mean ± SD.

Fig. 3.

FGF-2 induces corneal VEGF-A, -C, and -D expression. (A, C, and E) In corneas containing FGF-2 (12.5 ng), in situ hybridization revealed prominent VEGF-D expression in anterior and posterior epithelial cells and stromal cells (A), VEGF-C expression was weak (C), and VEGF-A expression was prominent in the stromal layer (E). (B, D, and F) Corneas containing control pellets had minimal signal for all probes. (Scale bars, 100 μm in A–F.) Tissue separation (asterisk) was due to artifact. E, anterior epithelial layer; S, stromal layer; DM, Descemet's membrane and posterior epithelial layer. (G) Control pellets or pellets with varying amounts of FGF-2 (12.5, 50, and 100 ng) were implanted into corneas of mice. After 3 days, corneal stromas were harvested. The stroma from normal, unmanipulated corneas (0) was also included. Total RNA was transcribed to cDNA, and real-time RT-PCR was performed to quantitate expression levels for VEGF-A, -C, and -D. Arbitrary units represent normalization to β₂-macroglobulin mRNA levels.

Fig. 4.

FGF-2-stimulated lymphangiogenesis and angiogenesis independently require VEGF-A and VEGF-D. (A) VEGFR-3 neutralizing antibody treatment, which blocks VEGF-D binding to VEGFR-3, specifically suppressed lymphangiogenesis, whereas angiogenesis was unaffected. (B) Soluble VEGFR-1, which would deplete local VEGF-A, but not VEGF-D, specifically suppressed angiogenesis, whereas lymphangiogenesis was not affected. (C) Soluble VEGFR-2 completely suppressed both processes. (D) The Fc portion of Ig had no effect on FGF-2-stimulated angiogenesis and lymphangiogenesis. Drawings depict interactions between FGF-2-induced VEGF-A and VEGFR-1 and -2 (present on blood vascular endothelium) and between FGF-2-induced VEGF-D and VEGFR-3 (present on lymphatic endothelium). The interaction between VEGF-D and VEGFR-2 is unclear. Green, activation of angiogenesis; pink, activation of lymphangiogenesis. (Scale bar, 0.5 mm.)