Vascular endothelial growth factor overproduced by tumour cells acts predominantly as a potent angiogenic factor contributing to malignant progression

Abstract

To elucidate the role of vascular endothelial growth factor (VEGF), an endothelial cell-specific mitogen, in tumour angiogenesis and malignant progression, an expression vector harboring human VEGF cDNA was stably transfected into three human cancer cell lines with poor VEGF productivity. Though their in vitro growth rate and intrinsic productivity of another angiogenic factor, basic fibroblast growth factor (bFGF), were not changed by transfection, those clones with higher VEGF production were endowed with tumorigenic and angiogenic potentials as follows: firstly, nontumorigenic, lung carcinoma QG90 cells having lower bFGF productivity acquired tumorigenicity as well as significant in vivo angiogenesis-inducing ability, secondly, tumorigenic colorectal carcinoma RPMI4788 cells having higher potency for bFGF production could form more vascularized solid tumour with faster growth rate and thirdly, oestrogen-dependent breast carcinoma MCF-7 cells, which did not produce detectable bFGF, acquired tumorigenicity even in the absence of oestrogen and the solid tumour growth rate was remarkably enhanced, accompanied with increased vascularization, in the presence of oestrogen. These results suggest that tumour progression closely depends on angiogenesis, and VEGF significantly contributes to malignant progression of a variety of tumour cells through its potent angiogenic activity, independent on the bFGF productivity of tumour cells.

Figure 1

(a) VEGF and (b) bFGF productivity of cancer cells transfected with VEGF cDNA. ▪ parental cells; □ transfected clones. The levels of VEGF and bFGF contained in conditioned medium and cell extract, respectively, were determined by sandwich ELISA. Sample values were plotted against a recombinant VEGF or bFGF standard curve (mean ± SE, n = 3). ND: not detectable. QG90, RPMI4788, MCF-7 and their transfected clones were cultured in RPMI1640 containing 10% FCS (normal). Hormones-stripped FCS with (E₂+) or without (E₂−) 10⁻⁸ M 17 β-oestradiol was used for MCF-7 and its transfected clone, M2–24-G10.

Figure 2

HUVE cell growth stimulatory activities in conditioned media of VEGF-transfected clones of (a) ○ QG90; □ Q1-1-C1; • Q1-24-E6; ▪ Q1-24-F6; ▴ Q1-24-H2, (b) ○ RPMI4788; • S0A12; ▪ S2G2 and (c) ○ MCF-7; • M2-24-D10; ▪ M2-24-E11; □ M2-24-G10 cells. HUVE cells were cultured in the presence of various concentrations of conditioned medium from each clone, indicated in Figure 1. After 4 days, HUVE cell growth was determined by MTT assay. The HUVE cell growth induced by 2.5 ng/ml of human recombinant VEGF165 was used as a positive control (♦). Results are presented as the percentage of the control without any growth factor. Each point indicates the mean of duplicate determinations.

Figure 3

Neutralization by anti-VEGF rabbit polyclonal antibodies (pAb) of the HUVE cell growth stimulatory activity secreted by transfected clones. Conditioned medium from each clone was added to HUVE cells at a concentration of 50% in the absence (□) or presence (▪) of anti-VEGF pAb. After the incubation for 4 days, HUVE cell growth was determined by MTT assay. Results are presented as the percentage growth relative to the fresh medium-treated control and are the average of triplicate determinations (mean ± SE). Additional control cells were incubated with 2.5 ng/ml of human recombinant VEGF165 (hr-VEGF165) or 4 ng/ml of human recombinant bFGF (hr-bFGF).

Figure 4

Growth curves of VEGF-transfected clones of (a) ○ QG90; □ Q1-1-C1; • Q1-24-E6; ▪ Q1-24-F6; ▴ Q1-24-H2, (b) ○ RPMI4788; • S0A12; ▪ S2G2 and (c) ○ MCF-7; • M2-24-D10; ▪ M2-24-E11; □ M2-24-G10 cells. The number of cells was counted with a Coulter Counter (mean ± SE, n = 3).

Figure 5

Effects of oestrogen on in vitro growth of the VEGF-overproducing M2–24-G10 clone (□,▪) and parental MCF-7 cells (○,•). The cells grown in the absence (○,□) or presence (•,▪) of 10⁻⁸ M 17β-oestradiol (E₂) were counted with a Coulter Counter (mean ± SE, n = 3).

Figure 6

In vivo angiogenic activity of VEGF-transfected clones of (a) QG90 and (b) RPMI4788 cells. A Millipore chamber containing each clone or HBSS was implanted s.c. into a dorsal air sac of a BALB/c mouse (three to six mice/group). After 4 days, the area of new capillarization induced by each clone was determined with an image analyser. Results (mean ± SE) are presented as percentages of the mean area of the control group treated with a HBSS-containing chamber. * P < 0.01 vs. the control group; ns, not significant (by Student's t-test).

Figure 7

Hematoxylin and eosin-stained sections of tumours 32 days after s.c. injection of (a) RPMI4788 cells and the transfected clones, (b) S0A12 and (c) S2G2 cells into nude mice. RPMI4788 cells yielded small, poorly vascularlized tumours (a), whereas both clones yielded well-vascularlized ones (b, c). Arrow heads indicate blood vessels. Scale bar = 400μm.

Figure 8

a, In vivo tumour growth of M2–24-G10 (•) and the parental MCF-7 (○) β cells in oestrogen-treated mice. Cells were implanted s.c. into groups of four nude mice with weekly s.c. administrations of 17β-oestradiol (E₂). The tumour size was measured with calipers, and the tumour volume was estimated by the formula: (length × width²)/2 (mean ± SE, n = 4). *, P < 0.05 vs. the tumour volume of MCF-7 cells (by Student's t-test). b, tumours excised on day 55 after implantation.