Generation of a syngeneic mouse model to study the effects of vascular endothelial growth factor in ovarian carcinoma

Abstract

Vascular endothelial growth factor (VEGF) performs multifaceted functions in the tumor microenvironment promoting angiogenesis, suppressing anti-tumor immune response, and possibly exerting autocrine functions on tumor cells. However, appropriate syngeneic animal models for in vivo studies are lacking. Using retroviral transfection and fluorescence-activated cell sorting, we generated a C57BL6 murine ovarian carcinoma cell line that stably overexpresses the murine VEGF164 isoform and the enhanced green fluorescent protein. VEGF164 overexpression dramatically accelerated tumor growth and ascites formation, significantly enhanced tumor angiogenesis, and substantially promoted the survival of tumor cells in vivo. In vitro, VEGF164 overexpression significantly enhanced cell survival after growth factor withdrawal and conferred resistance to apoptosis induced by cis-platin through an autocrine mechanism. VEGF/green fluorescent protein-expressing tumors were not recognized by the adaptive immune system. After vaccination, a specific anti-tumor T-cell response was detected, but tumor growth was not inhibited. This engineered murine carcinoma model should prove useful in the investigation of the role of VEGF in modulating the tumor microenvironment and affecting the complex interactions among angiogenesis mechanisms, anti-tumor immune mechanisms, and tumor cell behavior at the natural state or during therapy in ovarian carcinoma.

Figures

Figure 1.

Stable overexpression of VEGF164 in ID8 cells after retroviral transfection and sorting. ID8 cells were transfected with retrovirus containing GFP or GFP plus VEGF164. A: Top: flow cytometry analysis of GFP expression of cells transfected with VEGF164/GFP-positive retrovirus during sorting. Approximately 7.5% of the cells express low-medium levels of GFP (gate R2, red), whereas ∼7% express high levels of GFP (gate R3, blue). Bottom: Flow cytometry assessment of the purity of cells sorted based on high GFP expression. The purity of the population is >99.7% B: Top: GFP-positive ID8 cells are observed 72 hours after transfection under a fluorescent microscope. Bottom: After sorting, most cells express high levels of GFP. C: Total intracellular VEGF protein level analyzed by Western blotting after 20 passages in vitro. Intracellular VEGF protein level in cells transfected with VEGF/GFP-positive retrovirus is threefold higher compared to control cells. D: Assessment of the stability of GFP expression by flow cytometry. More than 90% of transfected cells stably express GFP after 20 passages (red, control nontransfected cells; blue, VEGF/GFP transfected cells). E: Quantification of VEGF164 expression by real-time quantitative RT-PCR. Cells transfected with GFP/VEGF-positive retrovirus and sorted based on high expression of GFP (GFP/VEGF) express 18-fold higher levels of VEGF164 mRNA (left) and eightfold higher levels of total VEGF mRNA (right) compared to wild-type control ID8 cells (WT) or ID8 cells transfected with GFP-positive retrovirus (GFP).

Figure 2.

VEGF164 overexpression in tumor dramatically accelerates ascites formation and tumor growth in vivo. A: Ascites accumulation is markedly higher in C57BL6 mice inoculated intraperitoneally with 1 × 107 ID8 cells transfected with VEGF/GFP-positive retrovirus (right) compared to mice inoculated with 1 × 107 ID8 cells transfected with GFP-positive retrovirus (left) 8 weeks after tumor inoculation. B: Animals inoculated with ID8 cells transfected with VEGF/GFP-positive cells display a median survival of 8 weeks, whereas control animals injected intraperitoneally with GFP-transfected cells display a median survival of 14.8 weeks (P < 0.05, n = 10/group). C: VEGF/GFP-transfected ID8 cells give rise to markedly larger flank tumors (left flank) compared to control GFP-transfected ID8 cells injected to the contralateral (right) flank. D: Comparison of contralateral tumors resected from three mice. Left: Tumors from GFP-transfected cells growing in the right flanks. Right: Tumors from VEGF/GFP-transfected cells growing in the left flanks. E: Growth of flank tumors after injection of VEGF/GFP-transfected cells (VEGF) and contralateral control GFP-transfected cells (control). Cells were suspended in Matrigel (n = 7/group; *, P < 0.05; **, P < 0.01). F: Stable overexpression of VEGF164 in vivo is demonstrated by RT-PCR. Results from three mice bearing flank tumors are shown. In each mouse, the left RNA sample is from the tumor generated with GFP-transfected, whereas the right sample is from the contralateral tumor generated with VEGF/GFP-transfected tumor. The rightmost lane represents RNA from cultured VEGF/GFP-transfected ID8 cells (positive control). VEGF isoforms 188, 164, and 120 were amplified with isoform-specific primers. Only VEGF164 is overexpressed in vivo. β-actin RNA documents equal amount of RNA used for all samples. G: Real-time quantitative RT-PCR confirms stable overexpression of total VEGF mRNA in vivo. Tumors formed by VEGF/GFP-transfected cells (VEGF) display fourfold higher total VEGF mRNA levels compared to contralateral control tumors formed by GFP-transfected cells (control). Data were normalized with the housekeeping gene GAPDH.

Figure 3.

GFP expression is stable in vivo and provides a sensitive tool to monitor tumor growth and metastasis. Stable expression of enhanced GFP in vivo allows for rapid identification of tumors in both flank and intraperitoneal models. A-F: Flank tumors resected from both sides of the same mouse. The borders between the tumor and normal tissue can be easily observed owing to the distribution of GFP fluorescence. Furthermore, the nonluminous tumor-associated blood vessels are clearly observed against the fluorescence of the GFP-expressing tumors under the fluorescent stereomicroscope. In control tumors from GFP-transfected ID8 cells (A and B), few blood vessels grow into the tumor; whereas in tumors from VEGF/GFP-transfected ID8 cells (C and D), prominent blood vessels growing into the tumor from nearby normal tissue are observed. Tumor from GFP-transfected ID8 cells under light microscope (E) and fluorescence stereomicroscope (F). A large central necrosis area is observed. Note the absence of prominent vessels in comparison with (D). Areas of necrosis were absent in tumors from VEGF/GFP-transfected ID8 cells. G and H: In the intraperitoneal tumor model, microscopic tumor nodules are detected on the spleen by stereomicroscopy. GFP expression allows for accurate detection of tumors. I: Real-time PCR revealed the presence of GFP gene, the genomic tumor marker, in multiple normal tissues of mice bearing VEGF164/GFP flank tumors, whereas in control mice, GFP was only detected in the lung.

Figure 4.

Overexpression of VEGF164 is associated with enhanced angiogenesis in vivo. A and B: Immunohistochemical detection of CD31 (brown) in flank tumors, followed by hematoxylin counterstaining. MVD calculated by the expression of CD31 was dramatically lower in control tumors generated with GFP-transfected cells (A, control) compared to tumors generated with VEGF/GFP-transfected cells (B, VEGF). C: Immunofluorescent staining of CD31 (red) shows prominent vascularization in association with VEGF/GFP-positive tumor cells (green). Nuclei are counterstained by 4,6-diamidiino-2-phenylindole (DAPI) (blue). Most of CD31+ microvasculature is confined within the tumor (GFP+/DAPI+) and does not extend into the GFP−/DAPI+ surrounding stroma. D: Quantitative analysis of CD31 staining density by image analysis. Tumors generated by VEGF/GFP-transfected cells (VEGF) show significantly higher density of CD31+ cells than control tumors (control, P < 0.05).

Figure 5.

VEGF164 overexpression promotes the survival of tumor cells in vivo. A pair of flank tumors from the same mouse is analyzed for detection of apoptosis in vivo with the TUNEL assay. MVD is detected by CD31 immunofluorescent staining. A and B: Tumor generated with GFP-transfected cells analyzed by TUNEL and hematoxylin counterstaining (A) and CD31 (B). A large necrotic area can be seen A, in which decreased MVD is appreciated with CD31 staining (B). Insets F and G are magnified below. C-E: Tumor generated with VEGF/GFP-transfected cells analyzed by CD31 (C) shows dramatically more pronounced MVD in the area occupied by tumor cells, as assessed by detection of GFP (D) and DAPI (E). F and G: Magnification of insets F and G from A. TUNEL assay reveals marked apoptosis in tumor cells in proximity to the necrotic area (F) as well as distal to the necrotic site (G) in control tumor generated with GFP-transfected cells. Tissue is counterstained with hematoxylin. H: Fewer TUNEL-positive apoptotic cells are detected in contralateral tumor generated with GFP/VEGF-transfected cells.

Figure 6.

VEGF receptor flt-1 but not flk-1/KDR is expressed in ID8 cells. Expression of VEGF receptors is assessed by nested RT-PCR in wild-type ID8 cells (WT) and ID8 cells transfected with VEGF/GFP-positive retrovirus (GFP/VEGF). In ID8 cells, VEGFR1/flt-1 is detected at low levels, whereas VEGFR2/flk-1/KDR is not detectable. The VEGF164-specific co-receptor neuropilin-1 is detected at high levels in ID8 cells. Transfection of VEGF164 did not change the expression levels of the above receptors. As positive control, flt-1 and KDR/flk-1 are detected in whole tumor RNA.

Figure 7.

VEGF overexpression promotes survival of ID8 cells in vitro and confers resistance to cis-platin-induced apoptosis in ID8 cells. A: Detection of apoptosis by flow cytometry analysis of annexin-V staining. Top: Control. Markedly fewer ID8 cells transfected with VEGF/GFP-positive retrovirus (GFP/VEGF) exhibit annexin-V staining (apoptosis) after growth factor withdrawal compared to control wild-type ID8 cells (WT) or ID8 cells transfected with GFP-positive retrovirus (GFP). Bottom: cis-platin. Markedly fewer ID8 cells transfected with VEGF/GFP-positive retrovirus exhibit apoptosis after exposure to 50 μmol/L of cis-platin for 24 hours compared to control wild-type ID8 cells or ID8 cells transfected with GFP-positive retrovirus. B: Summary of flow cytometry data from three different experiments. ID8 cells transfected with VEGF/GFP-positive retrovirus exhibit significantly less apoptosis after growth factor withdrawal (left) or exposure to cis-platin (right) compared to control wild-type ID8 cells (WT) or ID8 cells transfected with GFP-positive retrovirus (GFP; *, P < 0.05). C: In situ TUNEL and annexin-V apoptosis assay. Top: Control. In situ TUNEL assay detects apoptosis in control WT or GFP ID8 cells cultured under growth factor withdrawal conditions. Apoptotic TUNEL-positive cells are also detected in VEGF/GFP-positive cells. Middle: cis-platin. In situ TUNEL assay detects apoptosis in control WT or GFP ID8 cells exposed to 50 μmol/L of cis-platin for 24 hours. Note markedly increased prevalence of apoptosis and decreased number of cells after exposure to cis-platin compared to cells grown under growth factor deprivation (above). Apoptosis and cell number reduction is markedly less prominent in VEGF/GFP-positive cells. Bottom: cis-platin. In situ fluorescent annexin-V assay using biotinylated annexin-V followed by streptavidin-Red 670 staining. Apoptotic annexin-V-positive cells (red) are noted among control WT ID8 cells and GFP+ ID8 cells (green) transfected with GFP-positive or VEGF/GFP-positive retrovirus.

Figure 8.

VEGF overexpression reduces cis-platin-induced apoptosis of ID8 cells in vitro. DNA laddering analysis demonstrates that ID8 cells transfected with VEGF/GFP-positive retrovirus exhibit markedly less DNA fragmentation after exposure to cis-platin compared to control wild-type ID8 cells (WT) or ID8 cells transfected with GFP-positive retrovirus (GFP). M1 and M2 are two molecular markers.

Figure 9.

Exogenous VEGF partially reduces cis-platin-induced apoptosis in ID8 cells in vitro. A: Detection of apoptosis by flow cytometry analysis of annexin-V staining. Addition of cis-platin markedly increases apoptosis in wild-type ID8 cells compared to control cells cultured under serum-free, insulin-free conditions. Addition of recombinant murine VEGF partially reduces cis-platin-induced apoptosis. B: Summary of flow cytometry data from three different experiments. Addition of recombinant murine VEGF induces a significant reduction in apoptosis after exposure of cells to cis-platin.

Figure 10.

Genetically engineered tumors may be recognized by the adaptive immune system. A: Flow cytometry detection of major histocompatibility complex class-I (MHC-I) molecule expression in VEGF/GFP-positive ID8 cells. Approximately 50% of cells exhibit expression of surface MHC class-I molecules (red). White: isotype-negative control. B and C: ELISPOT analysis of tumor-reactive T cells in spleens of tumor-naïve nonvaccinated mice (ctrl) as well as mock-vaccinated (nonvaccinated, NV) or tumor-vaccinated (V) mice bearing flank tumors. B: Summary of three different experiments. A threefold increase in the frequency of tumor-reactive T cells secreting IFN-γ is noted in vaccinated animals (V) compared to control mice (*, P < 0.05). C: Depiction of one representative well per condition in one representative experiment. In the absence of tumor vaccination, control animals (NV) exhibit no evidence of tumor-reactive T cells compared to healthy tumor-naïve nonvaccinated C57BL6 female mice of matched age (ctrl). Marked increase in the number of spots staining for IFN-γ is noted, representing clones of antigen-specific (tumor-reactive) T cells recognizing tumor antigen presented by autologous DCs.