Rat C6 glioma cells (Cell Biology Research Institute of Shanghai, Shanghai, China) and C6 cells with expression vectors containing antisense (C6/VEGF⁻) VEGF164 cDNA or an empty vector (C6/vec) which were confirmed by assays for VEGF protein in cell culture supernatants and saved in our laboratory (7,8) were cultured in RPMI-1640 medium (1640M) (Invitrogen, Carlsbad, CA, USA) supplemented with fetal calf serum (10%).

For cell proliferation assay, 2×10⁴ cells were placed in a 6-well plate and were counted after 24, 48, 72, 96, 120 and 144 h culture by hemocytometer.

To measure VEGF secretion in vitro, 5×10⁵ cells were placed in 6-well plates and treated with serum-free 1640M. Medium was collected after 48 h of culturing. Debris was removed by centrifugation at 2000 × g for 5 min and supernatant was collected for enzyme-linked immunosorbent assay (ELISA). A commercially available ELISA kit (R&D Systems, Minneapolis, MN, USA) was used to detect mouse VEGF according to the manufacturer’s recommendations. Each experiment was performed a minimum of three times.

Confluence (80–90%) of C6 cells, C6/vec cells and C6/VEGF⁻ cells were treated with serum-free 1640M for 48 h. Media were collected and VEGF concentration were measured by ELISA. Human umbilical vein endothelial cells (HUVEC) (2×10⁴, Cell Biology Research Institute of Shanghai) were placed in a 6-well plate and treated with 100 ng of VEGF secreted from three C6 cell lines. The HUVEC cell growth curve was monitored by cell count.

Male 4–6-week-old BALB/c (nu/nu) mice (SLAC, Shanghai, China) (n=33) were randomized into three groups (n=11). Two individual clones of stable transfected C6 cells in the logarithmic growth phase were used for each construct (vector only, VEGF antisense), and the parental cell line was also studied. Two hundred microliters of cells (in serum-free medium with a final concentration of 7.5×10⁶ cells/ml) were injected subcutaneously into the right inguinal area of the mice. Anesthetized mice were sacrificed with decapitation and tumors were removed from the athymic (nu/nu) mice at 20 days post-implantation. The animals were sacrificed and the tumors were removed and then quickly frozen in liquid nitrogen for further analysis. All procedures met the national guidelines for the care and use of laboratory animals and were approved by the Institutional Animal Care and Use Committee of the Fujian Medical University, Fujian, China.

Total RNA was isolated and cDNA was synthesized as has been described (7,8). The sequences of primer sets were VEGF, forward: 5′-CCCAAGCTTATGAACTTTCTGCTCTCTTG-3′, reverse: 5′-CGCGGATCCTCACCGCCTTGGCTTGTC-3′; AQP4, forward: 5′-GCATGAATCCAGCTCGATCCTTTGG-3′ revese: 5′-AATGGGTGGCAGGAAATCTGAGGC-3′; β-actin, forward: 5′-GAGGCATCCTGACCCTGAAG-3′, reverse: 5′-CATCACAATGCCAGTGGTACG-3′. The calculation of expression levels of VEGF and AQP4 was normalized by β-actin.

To detect VEGF and AQP4 expression in vitro, cells were fixed in 4% paraformaldehyde and blocked with 3% normal goat serum for 2 h at room temperature for immunocytochemistry analysis. To determine VEGF and AQP4 expression in vivo, anesthetized mice were decapitated, and tumor tissues were removed and quickly fixed in 10% formalin. For immunohistochemistry, 4-μm-thick sections were cut and rehydrated, treated with 0.3% hydrogen peroxide in methanol for 30 min to inactivate endogenous peroxidase, rinsed with 0.1 M phosphate buffer (PB) for 10 min, and exposed to blocking serum (3% normal goat serum) for 2 h at room temperature.

Immunoreactions were performed as previously described (14). After incubation with anti-VEGF (1:150 dilution, United States Biological, Swampscott, MA, USA), and anti-AQP4 (1:150 dilution, Oncogene, Cambridge, MA, USA), the slices were rinsed with 0.1 M PB and exposed to anti-rabbit IgG HRP (1:500, Maixin, Fuzhou, China). After an additional 10-min rinse, the slices were treated with Vectastain^® Elite ABC reagent (Maixin) for 30 min and developed with DAB detection kit (Maixin). The slices were counterstained by hematoxylin and mounted by Permount (Maixin).

To measure VEGF secretion in vitro, 5×10⁵ cells were placed in 6-well plates and treated with serum-free 1640M. Medium was collected after 48 h of culture. Debris was removed by centrifugation at 2,000 × g for 5 min and the supernatant collected for ELISA assay.

To measure VEGF levels in vivo, tumor tissues (0.1 g) were homogenized in Tris-HCl buffer (25 mM, pH 7.6) containing 100 mM NaCl, 1 mM EDTA, and 1 mM phenylmethanesulfonyl fluoride (PMSF). Debris was removed by centrifugation at 2,000 × g for 5 min, followed by centrifugation at 20,000 × g for 20 min, and supernatants were collected for ELISA assay. Protein concentrations were measured using Protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA), and was normalized to a concentration of 1 mg/ml. Series dilutions of samples with the highest and the lowest expected values were performed to determine VEGF expression level using commercial VEGF ELISA kit (R&D Systems, Minneapolis, MN, USA) following the manufacturer’s instructions. VEGF expression levels were calculated by a standard curve available from R&D Systems. All experiments were performed in triplicate.

Referring to previously described studies (7,8), the water contents of the tumor samples were measured and taken to represent the degree of edema. Tumor tissues from the same sample which was also sampled for assays of VEGF expression were immediately weighed on an electronic analytical balance to obtain the wet weight (WW). The samples were then dried in a gravity oven at 100°C for 24 h to obtain the dry weight (DW). Water content was expressed as a percentage of wet weight; the formula for calculation was (WW-DW)/WW × 100%.

Tumor-bearing mice received a 0.1-ml/g i.v. injection of Evans blue dye (1% in saline; Sigma-Aldrich, St. Louis, MO, USA). After 6 h the animals were sacrificed and Evans blue was extracted from tumor as described (7). Briefly, tumors were removed and homogenized with 3 ml of N,N-dimethylformamide (Sigma-Aldrich), and incubated at 57°C for 12 h. The solutions were vortexed, then 2 ml of 1 N hydrochloric acid (HCl) were added, and the solutions were vortexed again, and then centrifuged at 2,500 rpm for 15 min. Supernatant was collected and measured at 620 nm with a spectrophotometer (Beckman Coulter, Fullerton, CA, USA). Concentrations were calculated by using a standard curve for Evans blue dye.

To analyze AQP4 expression in vivo, tumor tissues were homogenized in Tris-HCl buffer (50 mM, pH 8.0) containing the protease inhibitor cocktail V (Calbiochem, San Diego, CA, USA). Homogenate (20 μg) proteins were separated by electrophoresis on 4–20% SDS-PAGE gel and transferred onto Immobilon membranes (Millipore, Billerica, MA, USA). Western blot analyses were conducted using antibodies against AQP4 (1:100), and β-actin (1:2000, Neomarker, Fremont, CA, USA). Bands were visualized using an electrochemiluminescence (ECL) kit (Amersham Biosciences, Piscataway, NJ, USA).

SPSS 12.0 software (SPSS, Chicago, IL, USA) was applied for statistical data analysis. Data were analyzed by using one-way analysis of variance (ANOVA) followed by Dunnett’s post hoc test for multiple comparisons to the control groups. Differences were considered significant at P<0.05.

Immunostaining with VEGF antibody showed that endogenous VEGF level was lower in C6/VEGF⁻ cells than in C6/vec and C6 cells after 48 h of serum deprivation (Fig. 1A-C). Similarly, the level of VEGF protein in the medium from C6/VEGF⁻ cells was significantly lower than that from C6/vec and C6 cells (Fig. 1D). To investigate the effect of antisense VEGF on tumor cell proliferation, the number of C6/VEGF⁻, C6/vec, and C6 cells, respectively, placed into 6-well plates (2×10⁴ cells/well) were counted after 24, 48, 72, 96, 120 and 144 h in culture. Proliferation of C6/VEGF⁻ cells was found to be slower than that of the two control cells (C6/vec and C6; Fig. 1E), the morphology of all three cell lines was similar and remained unchanged (data not shown). However, the biological activity of VEGF released from different cell lines was not altered. Monitoring the growth of HUVEC cells treated with the same amount of VEGF secreted from our three C6 cell lines revealed similar levels of HUVEC cell growth, regardless of the source of VEGF, and slower growth in the absence of VEGF (Fig. 1F).

AQP4 expression in the three cell lines was assayed by RT-PCR and immunostaining. Similar AQP4 mRNA levels (Fig. 2A) were found in all three cell lines. In addition, the intensity of AQP4 immunoreactivity was similar in all lines after 48 h of serum deprivation (Fig. 2B-D). Thus, VEGF does not appear to have a direct role in AQP4 expression in glioma tumor cells.

To evaluate the possible role of VEGF in tumorigenesis, mice were injected with glioma cells (C6, C6/vec, or C6/VEGF⁻ cells) directly into the right inguinal area, and tumor size was measured. At 20 days after inoculation, tumor size (Fig. 3A) and tumor weight (Fig. 3B) were notably smaller in C6/VEGF⁻ mice.

To confirm the expression of VEGF in vivo as well as in vitro, VEGF levels were determined in genetically modified C6 cells by RT-PCR and ELISA. ELISA analysis (Fig. 4A) and RT-PCR (Fig. 4B) showed markedly lower level of VEGF mRNA and protein, respectively, in tumors from C6/VEGF⁻mice than tumors from C6/vec and C6 mice. A common feature of malignant brain tumors is increased capillary permeability leading to edema. Assay of tumor water content showed that C6/VEGF tumors had a lower water content than either of the two control tumors (C6/vec and C6) (Fig. 4C). To confirm that the edema was attributable to vascular hyperpermeability, vascular extravasation was examined using a dye tracer. Vascular leakage was markedly reduced in the C6/VEGF⁻ tumors (Fig. 4D). Also, Pearson’s correlation analysis found a correlation between water content and VEGF expression (Pearson’s correlation, r=0.946 P=0.00).

Levels of AQP-4 appeared to be lower in C6/VEGF⁻ tumors compared with C6G, C6/vecG tumors by immunohistochemistry (Fig. 5A-C), and western blot analysis (Fig. 5D and E), and Pearson’s correlation analysis show that AQP4 expression paralleled the level of VEGF expression (r=0.883, P=0.00) and the water content (r=0.912, P=0.00) of glioma tissue. Thus, aquaporin-4 expression in glioma tissue is suggested to be a reaction to glioma-associated edema induced by VEGF.