Distinct role of fibroblast growth factor-2 and vascular endothelial growth factor on tumor growth and angiogenesis

Abstract

Tumors express more than a single angiogenic growth factor. To investigate the relative impact of fibroblast growth factor-2 (FGF-2) and vascular endothelial growth factor (VEGF) on tumor growth and neovascularization, we generated tumor cell transfectants differing for VEGF and/or FGF-2 expression. Human endometrial adenocarcinoma HEC-1-B-derived Tet-FGF-2 cells that express FGF-2 under the control of the tetracycline-responsive promoter (Tet-off system) were further transfected with a VEGF(121) anti-sense (AS-VEGF) cDNA. Next, Tet-FGF-2 and AS-VEGF/Tet-FGF-2 cells were transplanted subcutaneously in nude mice that received tetracycline or not in the drinking water. Simultaneous expression of FGF-2 and VEGF in Tet-FGF-2 cells resulted in fast-growing lesions characterized by high blood vessel density, patency and permeability, and limited necrosis. Blood vessels were highly heterogeneous in size and frequently associated with pericytes. Inhibition of FGF-2 production by tetracycline caused a significant decrease in tumor burden paralleled by a decrease in blood vessel density and size. AS-VEGF expression resulted in a similar reduction in blood vessel density associated with a significant decrease in pericyte organization, vascular patency, and permeability. The consequent decrease in tumor burden was paralleled by increased tumor hypoxia and necrosis. A limited additional inhibitory effect was exerted by simultaneous down-regulation of FGF-2 and VEGF expression. These findings demonstrate that FGF-2 and VEGF stimulate vascularization synergistically but with distinctive effects on vessel functionality and tumor survival. Blockade of either one of the two growth factors results in a decrease in blood vessel density and, consequently, in tumor burden. However, inhibition of the expression of VEGF, but not of FGF-2, affects also vessel maturation and functionality, leading to tumor hypoxia and necrosis. Our experimental model represents an unique tool to investigate anti-neoplastic therapies in different angiogenic environments.

Figure 1.

VEGF and FGF-2 expression by Tet-FGF-2 and AS-VEGF/Tet-FGF-2 transfectants. Tet-FGF-2 15H cells were transfected with the pcDNA-3 expression vector harboring the VEGF₁₂₁ anti-sense cDNA. A: Stable transfectants (AS-VEGF/Tet-FGF-2 A5 and C1 clones) and mock-transfected cells (Tet-FGF-2 A4 and C7 clones) were analyzed for AS-VEGF mRNA steady-state levels by Northern blotting of total RNA (20 μg/lane) using a digoxigenin-UTP-labeled AS-VEGF riboprobe. B: Uniform loading of the gel was assessed by methylene blue staining of the filter. C: Serum-free conditioned media from the different clones (60 μg of protein) were probed with anti-VEGF antibodies by Western blotting showing decreased levels of the secreted forms of VEGF in AS-VEGF/Tet-FGF-2 cells. D: Lysates (500 μg) from cells grown in the absence (−) or in the presence (+) of tetracycline were loaded onto 0.1-ml heparin-Sepharose columns and bound material was probed with anti-FGF-2 antibodies in a Western blot. All of the clones express the 24-, 22-, and 18-kd FGF-2 isoforms whose expression is hampered by tetracycline treatment. E: Endothelial GM 7373 cells were incubated for 24 hours with increasing concentrations of the conditioned media from untreated and tetracycline-treated (+tet) Tet-FGF-2 A4 and AS-VEGF/Tet-FGF-2 C1 clones. At the end of incubation cells were trypsinized and counted. Data are the mean ± SE of four determinations.

Figure 2.

In vitro and in vivo growth properties of Tet-FGF-2 and AS-VEGF/Tet-FGF-2 transfectants. A: AS-VEGF/Tet-FGF-2 A5 and C1 clones (▵, ▴) and mock Tet-FGF-2 C7 and A4 clones (○, •) were seeded at 25,000 cells/well in 24-well plates in complete cell culture medium. At the indicated time points, cells were trypsinized and counted. B: Cells were transplanted subcutaneously in nude mice (1 × 10⁶ cells/animal) and tumor growth was reported. Each point is the mean ± SE of five tumors. Statistical analysis of tumor volume by orthogonal comparison analysis of variance at day 45 showed a significant difference (P < 0.001) between Tet-FGF-2 and AS-VEGF/Tet-FGF-2 clones (**). At sacrifice, three Tet-FGF-2 A4 tumors and three AS-VEGF/Tet-FGF-2 C1 tumors were analyzed for AS-VEGF mRNA expression (arrow) by Northern blotting (inset). Uniform loading of the gel was assessed by methylene blue staining of the filter.

Figure 3.

Effect of tetracycline on Tet-FGF-2 and AS-VEGF/Tet-FGF-2 tumor growth. Nude mice were transplanted subcutaneously with 1 × 10⁶ Tet-FGF-2 A4 cells (○, •) or AS-VEGF/Tet-FGF-2 C1 cells (▵, ▴). Animals were left untreated (•, ▴) or received 2 mg/ml of tetracycline in the drinking water throughout the whole experimental period starting 4 days before cell transplantation (○, ▵). Each point is the mean ± SE of five tumors. The results are representative of three independent experiments. Statistical analysis of tumor volumes by analysis of variance followed by Tukey-Kramer posthoc test showed a significant difference (P < 0.01) between the untreated Tet-FGF-2 group and all of the other groups (day 43, **) and between tetracycline-treated Tet-FGF-2 group and the two AS-VEGF/Tet-FGF-2 groups (day 43 and day 64, *). No significant difference was observed between untreated and tetracycline-treated AS-VEGF/Tet-FGF-2 groups at any time point.

Figure 4.

Tumor necrosis and mononuclear cell infiltrate in Tet-FGF-2 and AS-VEGF/Tet-FGF-2 xenografts. At the end of the experimentation (shown in Figure 3 ▶ ), animals were sacrificed and tumor sections were stained with H&E (A, B). Note the large necrotic area at the center of AS-VEGF/Tet-FGF-2 lesions (arrows in B) compared to the limited necrosis observed in Tet-FGF-2 tumors (A). In C, the percentage of necrotic tumor parenchyma was quantified by computerized image analysis of tissue sections (two slides per tumor, five animals per group) from Tet-FGF-2 and AS-VEGF/Tet-FGF-2 xenografts treated (+) or not (−) with tetracycline. **, Statistically different from AS-VEGF (P < 0.05 or better). D and E: CD11b immunostaining of Tet-FGF-2 and AS-VEGF/Tet-FGF-2 tumors. Large areas of CD11b⁺ monocyte/macrophage infiltrates are observed in Tet-FGF-2 xenografts (D) but not in AS-VEGF/Tet-FGF-2 lesions (E). A similar immunostaining was observed in all tumors examined (five tumors per group). Original magnifications: ×2.5 (A, B); ×20 (D, E); ×63 (inset in D).

Figure 5.

Vascularization of Tet-FGF-2 and AS-VEGF/Tet-FGF-2 tumors. Tet-FGF-2 and AS-VEGF/Tet-FGF-2 tumor sections from control (−) and tetracycline-treated (+) nude mice (see Figure 3 ▶ ) were processed for CD31 immunostaining. For each section, the most vascularized area within the tumor parenchyma at the periphery of the lesion was selected and microvessels in a ×400 field were counted (A). Total CD31⁺ blood vessel area (B) and total pericyte/myofibroblast α-SMA⁺ area (C) in 0.4-mm² fields of viable tumor parenchyma (three fields per tumor) were quantified by computerized image analysis. D: Semiquantitative scoring of HIF-1α immunostaining was performed on the whole tumor sections using an arbitrary scale from 0 (negative) to 4+ (strongly positive). n = 5 mice per group. **, P < 0.05 or better.

Figure 6.

Immunohistochemical characterization of Tet-FGF-2 and AS-VEGF/Tet-FGF-2 tumors from tetracycline-untreated mice. Representative images from Tet-FGF-2 (A, C, E, G) and AS-VEGF/Tet-FGF-2 (B, D, F, H) lesions are shown. See Figure 5 ▶ for quantitative analysis performed on both tumor types in the absence and in the presence of tetracycline treatment. A and B: Tumor sections were processed for CD31 immunostaining. Note the large-sized vessels in Tet-FGF-2 tumors (A) when compared to small-caliber microvessels in AS-VEGF/Tet-FGF-2 lesions (B). Small-caliber microvessels were observed also in all of the Tet-FGF-2 and AS-VEGF/Tet-FGF-2 tumors from tetracycline-treated animals (not shown). C and D: Tumor sections were double-labeling immunostained to visualize CD31⁺ endothelial cells (in red) and α-SMA⁺ cells (in green) by confocal microscopy. Note the numerous α-SMA⁺ cells frequently associated with CD31⁺ blood vessels in Tet-FGF-2 lesions (C) when compared to AS-VEGF/Tet-FGF-2 tumors (D). E and F: Animals were injected intravenously 2 minutes before sacrifice with fluorescent Hoechst 33342. A remarkable extravasation of the dye (in blue) from patent CD31⁺ vessels (in red) was observed in Tet-FGF-2 lesions (E) when compared to AS-VEGF/Tet-FGF-2 tumors (F). Note the nonpatent Hoechst 33342-negative/CD31⁺ vessels frequently detected in AS-VEGF/Tet-FGF-2 lesions (arrows in F). G and H: HIF-1α immunostaining of Tet-FGF-2 (G) and AS-VEGF/Tet-FGF-2 (H) tumors showing a strong, mainly nuclear (inset), HIF-1α expression in AS-VEGF-transfected lesions. Tetracycline administration had no significant effect on α-SMA and HIF-1α immunostaining and on Hoechst 33342 extravasation in both Tet-FGF-2 and AS-VEGF/Tet-FGF-2 tumors when compared to untreated lesions (data not shown). Original magnifications: ×20 (A–H); ×63 (inset in H).