The gene for the major angiogenic factor, vascular endothelial growth factor (VEGF), encodes several spliced isoforms. We reported previously that overexpression of two VEGF isoforms, VEGF121 and VEGF165, by human glioma U87 MG cells induced tumor-associated intracerebral hemorrhage, whereas expression of a third form, VEGF189, did not cause vessel rupture. Here, we test whether these VEGF isoforms have distinct activities for enhancing vascularization and growth of gliomas in mice. U87 MG cells that overexpressed VEGF165 or VEGF189 grew more rapidly than the parental cells in both s.c. and intracranial (i.c.) locations. However, cells that overexpressed VEGF121 only showed enhancement of i.c. tumor growth but had a minimal effect on s.c. glioma progression. At both anatomical sties, VEGF165 and VEGF189 strongly augmented neovascularization, whereas VEGF121 only increased vessel density in brain tumors. In each type of glioma, expression of VEGF receptors -1 and -2 largely phenocopied the tumor vasculature, because increased VEGF/VEGF receptor-activated microvessel densities were strongly correlated with the angiogenicity and tumorigenicity elicited by the VEGF isoforms at both anatomical sites. One notable difference between the sites was the expression of vitronectin, a prototypic ligand of αvβ3 and αvβ5 integrins, detected in i.c. but not in s.c., gliomas. Endothelial cell migration stimulated by VEGF121 was potentiated by vitronectin to a greater extent than that stimulated by VEGF165. This data demonstrates that VEGF isoforms have distinct activities at different anatomical sites and suggest that the microenvironment of different tissues affects the function of VEGF isoforms.

VEGF4 is an important regulator of physiological and pathological angiogenesis. VEGF induces sprouting angiogenesis, increases vessel permeability, and controls vasculature remodeling and integrity (1). VEGF is expressed in many types of tissues and is up-regulated during development, tissue remodeling, wound healing, and in human diseases including cancers. VEGF exerts its cellular functions by interacting with two tyrosine kinase receptors, VEGFR-1 (Flt-1) and VEGFR-2 (Flk-1/KDR). Mice deficient in VEGFR-1 have an increased number of endothelial progenitors that lead to vascular disorganization and embryonic lethality (2). VEGFR-2 is the principal receptor that mediates VEGF stimulation. Targeted disruption of VEGFR-2 results in abrogation of embryonic vascularization/angiogenesis, and suppressed neovascularization and tumor growth (3, 4). Both VEGFR-1 and -2 are up-regulated in primary human tumors (1).

Well-regulated VEGF expression is critical for angiogenesis. Hypoxia stimulates the expression of VEGF (1). Loss of a single allele (1) or a 2-fold increase (5) of VEGF expression in mice results in abnormal blood vessel development. In tumor cells, a 3–5-fold decrease of VEGF expression suppresses their angiogenicity and tumorigenicity in vivo(6, 7). In addition, overexpression of VEGF in mice (8) and tumor cells (9, 10) also leads to vigorous vascularization and increased vessel permeability. In humans, VEGF has six alternatively spliced isoforms: VEGF121, VEGF145, VEGF165, VEGF183, VEGF189, and VEGF206(1, 11). VEGF189 contains exons 6 and 7, and VEGF165 lacks residues encoded by exon 6a, whereas VEGF121 is missing 44 amino acids of exons 6 and 7. VEGF121 is secreted freely from cells and 50% of VEGF165 is retained on the cell surface, whereas VEGF189 is completely sequestered in the ECM (1). Various VEGF isoforms may have distinct functions in angiogenesis. In vitro, VEGF121, VEGF165, and VEGF189 stimulate EC proliferation and migration, but VEGF121 has a 50-fold reduced mitogenic activity (12). In vivo, although the three VEGF isoforms are generally coexpressed, an increase of cell-associated VEGF189 has been observed in lung and colon cancers (13, 14). Overexpression of individual VEGF isoforms increases angiogenicity and tumorigenicity of tumors (9, 10, 15). In MCF-7 breast carcinoma xenografts, VEGF121 has been shown to be more potent than its two larger isoforms (16). In mice that only express VEGF120, a compensatory increase of VEGF120 expression has been observed, and the mice had impaired postnatal myocardial angiogenesis and ischemic cardiomyopathy that led to cardiac failure and death (17). The different VEGF isoforms may form a spatial gradient of patterning information during blood vessel formation, e.g., the longer matrix-bound isoforms (close to the site of VEGF production) provide stronger mitogenic signals than the shorter, more diffusible VEGF isoforms (17). However, the distinct pathophysiological roles of the VEGF isoforms in adult tissues remain largely unknown.

We have reported previously that overexpression of VEGF121 and VEGF165 but not VEGF189, by human glioma U87 MG cells caused tumor-associated intracerebral hemorrhage resulting from the rupture of VEGF-induced neovessels. Although VEGF189 did not promote hemorrhagic development, it was still capable of inducing rapid vessel growth (10). Here, we report that these VEGF isoforms have distinct activities for promoting the vascularization and growth of glioma cells in both orthotopic and hetereotopic sites in vivo. Our data show that the microenvironment at different anatomical sites affects the functions of VEGF isoforms in tumor angiogenesis.

Antibodies and Reagents.

Recombinant human VEGF121 and VEGF165 proteins were from R&D Systems, Minneapolis, MN. Anti-VEGF165 antibody (A-20), anti-NRP-1 antibody (C-19), antihuman VN antibody (C-20) and antihuman FN antibody (C-20) were from Santa Cruz Biotechnology, Inc., Santa Cruz, CA. An antimouse CD31 antibody and its isotype control IgG2a.κ were from PharMingen, Inc., San Diego, CA.

The T014 rabbit polyclonal antibody to VEGFR-2 (18) and the 11B5 monoclonal antibody that preferentially recognizes VEGF complexed with VEGFR-2 or VEGFR-1 (19) were generated and characterized as described. The secondary and tertiary antibodies were from Vector Laboratories, Burlingame, CA, and Jackson ImmunoResearch Laboratories, West Grove, PA. A 3,3′-diaminobenzidine elite kit was from Dako Co., Carpinteria, CA. Aqua block was from East Coast Biologicals, Inc., North Berwick, ME. All of the other chemicals and reagents were from Sigma Chemical Co., St. Louis, MO, or Invitrogen/Life Sciences, Rockville, MD.

The P3H8A9 mouse monoclonal antibody was raised against a recombinant Fc fusion protein containing the entire VEGFR-1 extracellular domain. It reacted with Flt-1-immunoglobulin1–7, but not with Flt-1-immunoglobulin1–3. It also recognized human and rodent full-length VEGFR-1 protein in VEGFR-1 transfected cell lines (data not shown). The anti-NRP rabbit polyclonal antibodies NP1ECD1A and NP2ECD1A were raised against purified peptides sequences located in extracellular domains of NRP-1 and NRP-2, respectively. The antibodies were affinity purified and recognize rodent and human NRP-1 (NP1ECD1A) or NRP-2 (NP2ECD1A) without cross-reactioning each other in Western blot analyses. Neither of the antibodies cross-reacts with VEGFR-2 (data not shown). The rabbit polyclonal antibody against purified recombinant rat NRP-1 protein was a gift from Alex Kolodkin at Johns Hopkins University, Baltimore, MD. This antibody reacts with rodent and human NRP-1.

VEGF Protein and Northern Blot Analyses.

The protein and RNA analyses of VEGF isoforms were performed as described previously (6, 10). To isolate total RNA from U87 MG gliomas, various tumor tissues were homogenized, dissolved, and processed using Trizol reagent (Life Technologies, Inc.) The probe for VEGF expression was a full-length VEGF165 cDNA.

Tumorigenicity, Tissue Processing, and IHC.

Human glioma U87 MG cells, the U87 MG VEGF isoform expressing cells, and their cultures were described previously (6, 10). Intracerebral stereotactic implantation, s.c. inoculation and tumor growth measurements of the various U87 MG glioma cells, preparations of frozen samples, IHC analyses, and vessel quantification were performed as described (6, 10). Specifically, in IHC analyses, the tissues were preblocked with 1% of BSA or Aqua block (1:20 dilution) at room temperature or 37°C for 1 h. VEGF complexed with VEGFR was detected with a 1:25 dilution of the directly biotinylated 11B5 antibody. VN and FN were detected with 1:50 or 1:100 dilutions of their corresponding antibodies, respectively. All of the primary antibodies were incubated with tumor tissues at 37°C for 30 min to 1 h or at 4°C overnight. All of the primary antibody reaction products (except from biotinylated 11B5) were subsequently visualized with their respective biotinylated secondary antibodies. All of the samples were also stained with corresponding mouse, rabbit, or goat IgGs as negative controls, and no detectable signal was found in each sample. The resulting colors were then developed by incubation with the diaminobenzidine chromophore and H2O2 followed by hematoxylin counterstaining. Quantitative analysis of the blood vessel densities of tumor samples was done using the Metamorph Image System for Microsoft Windows (Universal Imaging, West Chester, PA; Ref. 10).

Cell Migration Assays.

EC migration assays were performed as described (10, 20) using Boyden chambers and PAE/KDR. VN (Invitrogen/Life Sciences) was diluted in PBS and total volume of 7.5 μl of VN at various concentrations was placed on the lower surface of a polycarbonate filter (Costar Corp.) and air-dried. In wells of the lower compartment, 28 μl of Hams/F-12 + 0.1% BSA containing VN or recombinant VEGF proteins or nothing was added, and the coated filter was then overlaid on the top of these wells. PAE/KDR cells that were starved overnight in serum-free medium were harvested, counted, and suspended at 1 × 106 cells/ml The cells were preincubated at 37°C in a water bath for 5 min and then stimulated with recombinant VEGF121 (5 ng/ml) or VEGF165 (15 ng/ml; R & D systems, Inc.) at 37°C for 10 min. After removal of unbounded VEGF protein by brief centrifugation, 50 μl of cell suspension was added to the upper compartment. The cells were allowed to migrate at 37°C in a 5% CO2 humidified incubator for 4 h and then the filters were fixed, stained, and mounted as described previously (10). Nonmigrating cells on the upper surface were carefully removed with a cotton swab. In the blocking experiments, the cells were preincubated with 20 μg/ml of c7E3 (anti-β3 antibody, a gift from Dr. Barry S. Coller at Mt. Sinai School of Medicine, New York, NY) at 37°C for 10 min before the recombinant VEGF proteins were added. Migration was quantified by counting the migrated cells in 10 random high-powered fields (400 × total magnification) per filter.

The three principal VEGF isoforms displayed different activities in promoting the tumorigenicity of the glioma U87 MG cells at orthotopic and hetereotopic sites. We have shown previously that overexpression of the VEGF121 and VEGF165 but not VEGF189 in U87 MG cells caused tumor-associated intracerebral hemorrhages on xenografting into immunodeficient mice. In these experiments, VEGF189 enhanced angiogenicity and tumorigenicity of brain gliomas without causing hemorrhage (10).

Here we sought to investigate whether differences existed between the isoforms with respect to their ability to affect the tumorigenicity of U87 MG cells at orthotopic (brain) or hetereotopic (flank) implantation sites. To establish U87 MG glioma that overexpress VEGF121 and VEGF165 in brains, the levels of VEGF expression were adjusted (10) below those that induce i.c. hemorrhage. The levels of VEGF secretion into conditioned medium in cell culture that caused hemorrhage were 150 ng of VEGF per million cells for VEGF165 and 225 ng of VEGF per million cells for VEGF121(10). Because overexpression of VEGF189 in U87 MG cells did not cause rupture of tumor vessels, no adjustment for expression was done before their implantation. The amount of VEGF isoform secretion of the cells into conditioned medium was determined by a VEGF ELISA assay at the time of i.c. injection (10).

To investigate whether similar levels of expression of each VEGF isoform protein convey comparable enhancement of s.c. tumor formation, each type of U87 MG VEGF isoform overexpressing and parental cells were separately implanted into the s.c. site. Intracranial inoculation of 5 × 105 of the parental U87 MG cells consistently resulted in tumors with an average volume of 56.8 ± 4.52 mm3 in 40.2 ± 2.2 days (Table 1, n = 22). Subcutaneous implantation of 1 × 106 of the parental U87 MG cells established gliomas with an average volume of 1.45 ± 0.042 cm3 in 46 days (Fig. 1, n = 9). In contrast, U87 MG VEGF165-expressing cells developed brain tumors of 57.2 ± 2.62 mm3 in 25.1 ± 1.52 days (Table 1; n = 20) and s.c. tumors 1.47 ± 0.45 cm3 by day 32 (Fig. 1). Compared with the parental cell activity these results were significantly different at P < 0.012 for the s.c. site and P < 0.000 for the i.c. site. The VEGF189-expressing cells formed tumors of 58.8 ± 3.12 mm3 in 28.5 ± 2.87 days (n = 18; P < 0.000 compared with parent; NS compared with VEGF165-expressing cells) and s.c. gliomas with a volume of 1.52 ± 0.22 cm3 as early as 29 days (P < 0.000 as compared with parental cells; NS compared with VEGF165-expressing cells). In sharp contrast, whereas the VEGF121-overexpressing cell formed brain tumors of 54.1 ± 4.25 mm3 in 30.5 ± 1.19 days (n = 16; P < 0.0004 compared with parental cells; NS compared with VEGF165- or VEGF189-expressing cells), they developed s.c. tumors of 1.49 ± 0.45 cm3 in 46 days (NS compared with parental cells; P < 0.001 compared with VEGF165- or VEGF189-expressing cells). Thus, expression of the three VEGF isoforms by U87 MG cells caused differing behaviors with respect to intracerebral hemorrhage/angiogenesis (10) and also distinct tumorigenesis at orthotopic (brain) and hetereotopic (flank) sites in vivo.

Whereas VEGF165 and VEGF189 Strongly Augment Neovascularization at Both Anatomical Sites, VEGF121 Only Increases Vessel Growth in i.c. Tumors.

Each of the three VEGF isoforms is known to be potent stimulators of angiogenesis both in vitro and in vivo(1). In other studies, increased tumorigenicity of tumor cell lines caused by overexpression of VEGF165(9) and VEGF121(15, 16) correlated with increased tumor angiogenesis. Therefore, we next determined whether the distinct tumorigenic effects of overexpressing the various VEGF isoforms were accompanied by similar effects on angiogenesis in the two anatomical sites. We immunostained the tumor tissues using an anti-CD31 antibody and determined microvessel densities by computerized image analysis. The vessel densities of the tumors formed at both i.c. and s.c. sites by parental U87 MG cells were similar to previous results (Fig. 2, a and b; Table 2; and Refs. 6, 10). The microvasculature of the VEGF121 s.c. tumors was similar to those in the parental s.c. tumors (Fig. 2, compare c with a; P > 0.35; Table 2). In contrast, vessel density in VEGF121 brain tumors increased 2.9-fold relative to that in the U87 MG parental brain gliomas (Fig. 2, d and b; Table 2). VEGF165-expressing tumors at either site had 5.5-fold increases of vessel density compared with parental cells (Fig. 2, e and f; Table 2; P < 0.001). Increased vessel densities were also present in VEGF189 tumors formed at either anatomical site (Fig. 2, g and h; Table 2; P < 0.000). Interestingly, there were distinct differences in vessel morphologies among tumors formed by cells expressing the three types of VEGF isoform: the vessels were long in VEGF121-expressing and U87 MG parental s.c. tumors (Fig. 2,c), whereas many of the vessels in s.c. or i.c. tumors formed by VEGF165- or VEGF189-expressing cells and i.c. tumors formed by VEGF121-expressing cells were short or dilated (Fig. 2, d and f–h). Tumors formed at both sites by VEGF165-expressing cells displayed greatly increased densities of interconnected and dilated vessels (Fig. 2, e and f). Thus, increased angiogenicity in VEGF165 and VEGF189 gliomas was closely correlated with augmented tumor growth at either orthotopic or hetereotopic sites. The enhanced growth of tumors formed by i.c. implantation of VEGF121-expressing cells was concordant with increased angiogenesis. Conversely, the lack of enhancement of s.c. growth was evident despite a modest increase of angiogenic response.

Inability of VEGF121 to Promote Tumorigenicity at the Hetereotopic Sites Was Not Attributable to the Loss of Expression.

Overexpression of VEGF121 in the U87 MG cells promoted angiogenicity (Fig. 2,d) and tumorigenicity (Table 1) in the brain but failed to enhance glioma vessel growth (Fig. 2,c) and tumor expansion (Fig. 1) at the s.c. sites. To determine whether the expression levels of exogenous VEGF121, VEGF165, and VEGF189 in the established gliomas contributed to the different enhancements of tumor growth, we examined the expression in the various s.c. tumors by Northern blot analyses. As demonstrated in Fig. 3,A, four species of endogenous VEGF mRNA were detected. Their sizes were 4.2, 3.8, 1.4, and 1.2 kb, respectively. Three species of the exogenous VEGF mRNA were found at relatively similar levels in the VEGF isoform-overexpressing tumors, whereas they were absent in the U87 MG parental tumors. The sizes of the exogenous VEGF mRNAs were 0.8 kb for VEGF121, 0.9 kb for VEGF165, and 1.0 kb for VEGF189, respectively. These were consistent with the cDNA lengths of each VEGF isoform that were introduced into the U87 MG cells (6). This data showed that the exogenous VEGF121 mRNAs, as well as the mRNAs of VEGF165 and VEGF189, were overexpressed in these established tumors (Fig. 3 A).

We next examined the expression of VEGF proteins in the tumors by IHC analysis. The parental U87 MG gliomas expressed detectable amounts of VEGF proteins consistent with their endogenous expression (Fig. 3), and increased expression of VEGF isoform proteins were found in the VEGF isoform s.c. tumors (Fig. 3,B, panels b–d). Interestingly, the locations of the VEGF isoform proteins in the various VEGF isoform gliomas were correlated with their biochemical properties described before (20). In both VEGF121 and VEGF165 tumors, cell-bound (arrows) and vessel-associated (arrowheads) proteins were detected, indicating that both isoforms were associated or released from the U87 MG cells (Fig. 3B, panels b and c). As expected, both tumor cell-bound (Fig. 3,B, panel d) and vessel-associated VEGF proteins were also seen in the VEGF189 tumors. The concentrated local expression of VEGF189 proteins was consistent with the fact that ∼50% of VEGF189 proteins were retained in the ECM of the U87 MG cells (1). This contributes to the visual effect of the greatest immunostaining in VEGF189 tumors (Fig. 3,B, panel d). Similarly, markedly augmented expression of VEGF isoform proteins and similar cell-association patterns were also detected in the three types of VEGF isoform brain tumors (data not shown). It should be noted that the anti-VEGF165 antibody used in this assay could not distinguish among the VEGF isoform proteins. However, our data (Ref. 10; Fig. 3 A) has already revealed the identities of various exogenous VEGF isoforms in the U87 MG VEGF-expressing cells and their tumors. These results demonstrated that the inability of VEGF121 to enhance tumor growth and angiogenicity at the s.c. sites was not attributable to the loss of VEGF121 expression in the established gliomas.

VEGFRs Were Expressed in the U87 MG Parental and VEGF Isoform Gliomas.

VEGF exerts its biological functions through interaction with its cognate receptors, VEGFR-1 (Flt-1) and VEGFR-2 (Flk-1/KDR), as well as a VEGF165 specific receptor, NRP-1 (1). Because we could not attribute the differences of potencies of VEGF121 and VEGF189 at the orthotopic and hetereotopic sites to the expression levels of exogenous VEGF, we next performed IHC analysis on expression of VEGFR-1, -2, and NRP-1 and -2 on the various gliomas. We found that VEGFR-1 and -2 were detected in most of the vessels in each of the various tumors. Interestingly, the expression of VEGFR-1 and -2 was well correlated with the CD31 staining of the vessels in both s.c. and i.c. tumors (data not shown). NRP-1 and -2 are VEGF165- and VEGF145-specific receptors (21, 22). NRP-1 and -2 were originally identified as neuronal guidance molecules involved in the development of the nervous system (23). NRP-1 null (24) or NRP-1 transgenic mice (25, 26) developed embryonic vascular abnormality that led to lethality. NRP-1 and -2 have soluble forms that act as negative regulators for angiogenesis (27, 28). By reverse transcription-PCR analyses, we found that NRP-1 or -2 are expressed in the tumors and in isolated ECs of mouse aorta, yolk sac, and brain (data not shown). Using IHC analyses, we found that there were no differences in the expression of NRP-1 among the various types of the U87 MG gliomas established intracerebrally or s.c. (data not shown). We did not detect the expression of NRP-2 in the various U87 MG tumors by IHC analyses using our NRP-2 antibodies, but we detected NRP-2 expression in neurons and nucleus clusters in normal rat brains (data not shown). Thus, the distinct activities of VEGF121 and VEGF189 expression in the U87 MG cells at these two anatomical sites could not be attributed to differences in the expression of VEGFR-1 and -2, or NRP-1 and -2.

VEGF/VEGFR Complexes in the Gliomas at Different Sites Correlated with the Distinct Activities of the VEGF Isoforms.

VEGFRs are critical in mediating VEGF-stimulated tumor angiogenesis. The three VEGF isoforms bind to endothelial VEGFR with different capacities because of their affinities for heparin (1, 11). Emerging evidence has demonstrated that the association between VEGF and VEGFR is critical for angiogenic switch and tumor progression (29). Overexpression of the exogenous VEGF isoforms in the tumors (Fig. 3) and the constitutive expression of VEGFR-1 and -2 in the glioma endothelia (data not shown) prompted us to examine whether association of each VEGF isoform with the VEGFR differed among the various types of U87 MG gliomas. To do this, we used a monoclonal antibody (11B5) that selectively recognizes VEGF that is bound to either VEGFR-1 or -2 (19). As shown in Fig. 4, strong immunoreactivity in tumor vasculature to the 11B5 antibody was apparent in U87 MG i.c. gliomas formed by the cells with overexpression of any of the three VEGF isoforms in brains (Fig. 4, d, f, and h) or from s.c. tumors from the cells with overexpression of VEGF165 and VEGF189 (Fig. 4, e and g). In sharp contrast, VEGF/VEGFR complexes were found at low levels in most of the tumor vasculature in s.c. gliomas from U87 MG VEGF121-overexpressing cells (Fig. 4,c). Similarly, for U87 MG parental tumors established at both anatomical sites, a few of the complexes were found in tumor vessels (Fig. 4, a and b). The low levels of VEGF/VEGFR complexes in both U87 MG parental and VEGF121-overexpressing tumors are attributable to low levels of expression of endogenous VEGF165 and VEGF189, which is 7–12-fold lower than that in the U87 MG VEGF165- or VEGF189-overexpressing gliomas (10). In addition, strong immunoreactivity to the 11B5 antibody was observed in many large vessels in the various established gliomas and also in the normal brain tissues (data not shown). Thus, the association of VEGF with VEGFR-1 or -2 was strongly correlated with VEGF-stimulated angiogenesis and tumor growth. The fact that the complexes were formed in large/mature vessels in both normal and glioma tissues suggest that the association of VEGF to VEGFRs is important for vessel maintenance and readiness for capillary sprouting.

VN Was Not Expressed in the Various U87 MG Gliomas Established at s.c. Sites.

VN is an ECM ligand that binds to integrins αvβ3 and αv β5(30) and has been implicated in glioma progression (31). In one study, VN was expressed in i.c. tumors formed by U251 MG cells but not found in their corresponding s.c. gliomas (32). Therefore, we examined the expression of VN by IHC analysis using an antihuman VN antibody in the established U87 MG tumors to determine whether it correlated with the stimulation of tumor angiogenesis caused by overexpression of VEGF isoforms at the different anatomical sites. VN proteins were detected in the ECM of all of the i.c. U87 MG tumors established at sites (Fig. 5, b, d, f, and h). On the contrary, no expression of VN proteins was found in any of the four types of s.c. U87 MG (Fig. 5, a, c, e, and g). These results contrast with those in which we examined the expression of FN, another ECM ligand to αvβ3 and αvβ5 integrins (30). Using an antihuman FN antibody, we immunostained the same set of tumors analyzed for VN expression and found that FN was strongly expressed in all of the types of U87 MG tumors established both at intracerebral and s.c. sites (data not shown).

Stimulation of EC Migration by VEGF121 Was Greatly Potentiated by VN in Vitro.

Because VN was not expressed in s.c. tumors and because s.c. VEGF121 expression did not have significant impact on tumor growth or angiogenesis, we tested the hypothesis that VN has a differential effect on cells expressing the different VEGF isoforms. We reasoned that lack of stimulation of tumor angiogenesis caused by overexpressing VEGF121 in the U87 MG gliomas might be attributable to an absence of VN expression in these s.c. tumors. To test this possibility, we assessed whether the presence of VN affected EC migration stimulated by VEGF121 and VEGF165in vitro. We used PAE/KDR (10) for our assays. Fig. 6,A shows that the stimulation of PAE/KDR cell migration by recombinant VEGF121 and VEGF165 proteins was dose dependent. The optimal concentrations of the proteins for maximal augmentation of such migrations were 10 ng/ml for VEGF121 and 20 ng/ml for VEGF165, respectively. When variable amounts of VEGF121 or VEGF165 are used to briefly stimulate the migration of PAE/KDR cells (see “Materials and Methods”), the VEGF amounts as high as 10 ng/ml for VEGF121 and 20 ng/ml for VEGF165 only had moderate potency in stimulating the EC migration (Fig. 6,B). Remarkably, in the presence of 200 ng/ml of VN, VEGF121 protein at a suboptimal concentration protein (5 ng/ml; see Fig. 6,A) strongly promoted the EC migration (Fig. 6,B). Similarly, in the presence of the various amounts of VN protein, 5 ng/ml of VEGF121 protein had profound augmentation of EC migrations (Fig. 6,C). In contrast, when similar amounts of VN protein were included in the system, as high as 15 ng/ml of VEGF165 was needed to cause comparable stimulation of such migration (Fig. 6, B and C). When a neutralizing anti-β3 integrin antibody, c7E3 (33), was included in the assays, stimulation of EC migrations toward VN by both VEGF isoforms was inhibited (Fig. 6, C and D). Thus, cells expressing VEGF121 are more sensitive to VN in the extracellular space than cells expressing the other isoforms. This correlates well with the lack of tumor enhancement by VEGF121 expression in the s.c. site where VN is not expressed and with its enhancing abilities in the i.c. site where VN is expressed.

This study demonstrates that the three VEGF isoforms, VEGF121, VEGF165, and VEGF189, display distinct activities in stimulating tumor angiogenesis and growth at the same anatomical sites. Additionally, either VEGF121 or VEGF189 showed different biological activities at the different anatomical sites. Several lines of evidence illustrated here imply that the microenvironment at different anatomical sites may affect the function of VEGF isoforms. First, distinct activities of VEGF121 at different anatomical sites cannot be attributed to the bioavailability of these VEGF isoform proteins in the established U87 MG gliomas. VEGF121 increased tumorigenicity in the brain but failed to significantly enhance tumor expansion in the s.c. space (Fig. 1). This difference was not attributable to low expression levels of exogenous VEGF121 (Fig. 3,A). Second, neovascularization in the established gliomas correlated with the growth rates of VEGF121 and VEGF189 s.c. but was not associated with VEGF121 and VEGF189 i.c. glioma expansion. The slightly increased vessel growth in the s.c.VEGF121 tumors (Fig. 2,c) was correlated with a minimal enhancement of tumor growth (Fig. 1). Similarly, hypervascularized VEGF189 s.c. gliomas (Fig. 2,e) were the fastest expanding tumors (Fig. 1). However, a significant increase of vessel densities in the i.c. VEGF189 gliomas (Fig. 2,h) was not correlated with the less enhanced growth of the tumors (Table 1). Third, consistent with a previous report (33), VN, a prototypic ligand for αvβ3 and αv β5 integrins, was not expressed in various types of established s.c. gliomas and in cultured U87 MG cells, whereas it was expressed in ECM in the i.c. U87 MG gliomas (Fig. 5). Finally, in vitro, VN potentiated VEGF121 (at 5 ng/ml; Fig. 6, B and C) stimulated EC migration at higher efficiency than it did for the stimulation by VEGF165 (at 15 ng/ml; Fig. 6, B and D).

The activities exhibited by VEGF165 at both anatomical sites are consistent with evidence that VEGF165 is a strong angiogenic factor (1). Significant enhancement in s.c. tumor growth displayed by VEGF189 was in agreement with evidence that the ECM-associated VEGF189 in the CEN4 cells was the strongest factor to stimulate proliferation of the bovine EC (34). These observations were also consistent with the hypothesis that the longer matrix-binding VEGF189 and VEGF165 are more potent than the shorter, more diffusible VEGF isoform, VEGF121, (Fig. 1; Ref. 17). The activities of VEGF189 in the U87 MG gliomas at both anatomical sites additionally corroborated the evidence that mice expressing only VEGF188 survived and showed substantial angiogenesis in vivo(35). The incapability of VEGF121 to enhance the hetereotopic (s.c.) U87 MG glioma growth agreed with reports that VEGF120/1 only partially rescued tumor growth in H-ras transformed VEGF null fibroblasts (36) or mutant K-ras knockout colorectal carcinoma cells (7). This finding is additionally supported by evidence that in mice lacking VEGF164 and VEGF188 expression, VEGF120 rescued defective vasculature in heterozygous VEGF-deficient embryos but was not sufficient to sustain a functional vasculature in a homozygous-deficient state (17).

In a recent study, Christofori et al.(37) showed that in RIP1-Tag 2 transgenic mice, both nonangiogenic islets and pancreatic tumors had substantial expression of VEGF, VEGFR-1, and -2, but only the tumors were highly angiogenic. This suggests that expressions of VEGF, VEGFR-1, and -2 are not sufficient indicators for active vessel growth (37). Our analyses of VEGFR-1 and -2 in various gliomas corroborated their findings. The expression profiles of VEGFR-1 and -2 (data not shown) largely phenocopied the tumor vasculature in each type of the U87 MG gliomas, as revealed by CD31 staining (Fig. 2). Moreover, embryos lacking VEGFR-1 displayed an increased outgrowth of ECs and hemoangioblast commitment (2). Mice expressing only truncated flt-1, which lacks the cytoplasmic domain, developed normal vasculature (38). VEGFR-1 has been proposed to be a negative regulator for VEGF function (38). By comparing the expression profiles of VEGFR-1 and -2, we postulate that in addition to proposed inhibitory effect, VEGFR-1 may cooperate with VEGFR-2 and other molecules to optimize active vessel growth. In addition, no differences in the expression of a VEGF165 receptor, NRP-1, were detected in both tumor cells and vessels among the various gliomas (data not shown). Thus, we cannot attribute the expression profiles of VEGFR-1, -2, and NRP-1 to the differences in stimulation of tumor angiogenesis displayed by the VEGF isoforms.

Association of VEGF and VEGFR is a critical step for angiogenic switch and tumor growth. By using a monoclonal antibody that recognizes VEGF complexed with VEGFR-1 and -2, it has been demonstrated that VEGF/VEGFR complexes were only detected in angiogenic islets and tumors, not in the nonangiogenic islets. This result suggested that the association of VEGF and VEGFRs is a prerequisite for angiogenesis (29). This antibody has also been used to show that intense VEGF/VEGFR angiogenic pathway activation is a tumor-specific feature and is associated with poor postoperative outcome (39). Our results of VEGF/VEGFR complexes in various U87 MG gliomas support these two observations. We show that strong immunoreactivities of VEGF/VEGFR activated microvessels in the VEGF165 gliomas at both sites (Fig. 4, e and f) and in the VEGF121 and VEGF189 brain tumors (Fig. 4, d and h) correlated with the augmentation by the VEGF isoforms. In sharp contrast, only few VEGF/VEGFR complexes were detected in vessels in VEGF121 s.c. tumors (Fig. 4,c) and the U87 MG parental gliomas (Fig. 4, a and b). We propose that activation of the VEGF/VEGFR signaling pathway is critical for the stimulation by VEGF isoforms in vivo.

Our data that VN affected the stimulation of EC migration by VEGF121 provides a clue for the mechanism underlying the distinct differences in enhancement of glioma growth by VEGF121 at different anatomical sites. Several studies have shown that overexpression of VEGF121 augmented s.c. tumor growth of MCF-7 breast carcinomas (16), HT 1080 fibrosarcoma (15), and immortalized murine ECs (40). We hypothesize that other molecules that were expressed in the tumors at different anatomical sites modulated the activities of VEGF isoforms. VN is a prototypic ligand for the αvβ3 integrin (30). The αvβ3 plays an important role in angiogenesis (41). VEGF165 activated αvβ3 and several other integrins to augment EC migration toward VN. The activation of integrins by VEGF165 was mediated by VEGFR-2, involved phosphatidylinositol 3′-kinase and Akt, and was negatively regulated by PTEN (33). VN is expressed in primary human glioma tissues (31). In U251 MG xenografted glioma model, VN was only expressed in i.c. but not s.c. tumors (32). More importantly, an αvβ3 integrin antagonist preferentially suppressed i.c. DAOY and U87 MG tumors in mice but had no effect on their hetereotopic (s.c.) tumor growth, perhaps because of a lack of VN expression in those s.c. tumors (42, 43). By reverse transcription-PCR and flow cytometry analysis, we found that αv, β3, β5, and β1 integrins were expressed in cultured U87 MG cells and in the various types of U87 MG parental and VEGF isoform expressing gliomas established at both anatomical sites.5 We also found that FN, another ligand for αvβ3 and αvβ5 integrins, was expressed in the various types of U87 MG gliomas formed at both anatomical sites (data not shown), demonstrating some specificity in this activity. Our data that VN was not expressed in the U87 MG s.c. gliomas (Fig. 5) and that VN potentiated VEGF121-stimulated EC migration at higher efficacy in vitro (Fig. 6) suggests that the expression of VN in U87 MG flank tumors was critical for VEGF121 to augment neovascularization at this hetereotopic site. Whereas we do not yet know the mechanistic basis for expression of VN in i.c. but not s.c. tumors, it has been demonstrated that the microenvironment at different tumor sites might determine certain gene expression (44, 45). Such differentially expressed genes could affect VEGF angiogenic activity. This notion is supported by recent studies that ECM composition determines the transcriptional response to epithelial growth factor activation (46). Also, tumor ECM at different anatomical sites influences diffusion of macromolecules (47).

In a recent report, Grunstein et al.(36) showed that in VEGF-deficient embryonic fibroblasts, which were immortalized and transformed, VEGF isoforms displayed different activities. VEGF164 could fully rescue tumor growth, VEGF120 partially rescued tumor progression, and VEGF188 completely failed to rescue tumor expansion, perhaps because of inadequate recruitment of the host vasculature (36). A possible explanation for the discrepancy between this result and ours relative to which isoform had which activity is that the microenvironments affected gene expression differentially in these two different model systems. In our U87 MG glioma model, differentially expressed VN may affect the VEGF121 activities at different anatomical sites. There is also the possibility that isoform bioavailability differs in distinct anatomical sites.

In summary, our data shows that three VEGF isoforms have distinct activities in stimulating neovascularization at different anatomical sites. The activation of VEGF/VEGFR stimulated pathways is the critical step for each of the VEGF isoforms to enhance tumorigenicity and angiogenicity. The microenvironment in the U87 MG gliomas established at different anatomical sites affected the biological functions of the VEGF isoforms. This data may provide important clues for dissecting the mechanisms of tumor angiogenesis and designing rational targeting of VEGF pathways. For example, when targeting a VEGF pathway in i.c. human gliomas, one should probably aim to disrupt all three of the principal VEGF isoform functions, because they each enhanced i.c. tumor growth. In other tumors such as fibrosarcoma (15) and breast carcinoma (16) such suppression may need to be focused on the smaller VEGF isoforms. The data also suggest that the design of VEGF-directed therapy for a primary tumor may need to differ from its metastatic derivative depending on their two body sites and the microenvironment of each.

Fig. 1.

Effects on s.c. tumorigenicity of U87 MG gliomas by overexpression of VEGF isoforms. U87 MG cells overexpressing various VEGF isoforms were injected into the right flanks of nude mice, whereas the same numbers of the parental U87 MG cells were implanted into the left flanks of the same animals. Tumor volumes were estimated [volume =(a2 × b)/2, a < b; Ref. 6)] at the indicated times after implantation and data are shown as the mean; bars,± SE. Three separate clones from each type of the U87 MG VEGF isoform expressing cells were individually inoculated. In vitro, the growth rates and levels of VEGF overexpression of each of these clones were similar (6). The experiments included four to six mice in each group and were repeated two additional times with similar results.

Fig. 1.

Effects on s.c. tumorigenicity of U87 MG gliomas by overexpression of VEGF isoforms. U87 MG cells overexpressing various VEGF isoforms were injected into the right flanks of nude mice, whereas the same numbers of the parental U87 MG cells were implanted into the left flanks of the same animals. Tumor volumes were estimated [volume =(a2 × b)/2, a < b; Ref. 6)] at the indicated times after implantation and data are shown as the mean; bars,± SE. Three separate clones from each type of the U87 MG VEGF isoform expressing cells were individually inoculated. In vitro, the growth rates and levels of VEGF overexpression of each of these clones were similar (6). The experiments included four to six mice in each group and were repeated two additional times with similar results.

Close modal
Fig. 2.

Angiogenesis of the tumors formed by the U87 MG and VEGF isoform expressing cells. Panels a–h, immunohistochemical stains of tumor vessels with an anti-CD31 monoclonal antibody. Gliomas formed by the U87 MG parental cells (a and b), VEGF121-expressing cells (c and d), VEGF165-expressing cells (e and f), and VEGF189-expressing cells (g and h). Arrows, vessels that were also identified in Figs. 4. Six or more individual tumor samples of each class were analyzed each time, and the experiments were repeated at least two additional times with similar results. Original magnifications: ×200. The photographs were captured into computer files and contrast-enhanced using Adobe Photoshop.

Fig. 2.

Angiogenesis of the tumors formed by the U87 MG and VEGF isoform expressing cells. Panels a–h, immunohistochemical stains of tumor vessels with an anti-CD31 monoclonal antibody. Gliomas formed by the U87 MG parental cells (a and b), VEGF121-expressing cells (c and d), VEGF165-expressing cells (e and f), and VEGF189-expressing cells (g and h). Arrows, vessels that were also identified in Figs. 4. Six or more individual tumor samples of each class were analyzed each time, and the experiments were repeated at least two additional times with similar results. Original magnifications: ×200. The photographs were captured into computer files and contrast-enhanced using Adobe Photoshop.

Close modal
Fig. 3.

Expression of exogenous VEGF isoforms in established U87 MG gliomas. A, Northern blot analysis. Top panel, samples of the U87 MG parental and the three types of VEGF isoform s.c. tumors are shown. In each class, tumors formed by the individual VEGF isoform expressing clones from each type were used. •, endogenous and ▸, exogenous VEGF mRNA species. Bottom panel, methylene blue staining of the membrane after the RNA was transferred into the membrane. ▪, 18 S and 28 S rRNA species. This analysis was repeated two additional times with similar results. B, expressions of VEGF isoform proteins detected by immunohistochemical analysis in the U87 MG parental tumor (a), the VEGF121-expressing tumor (b), the VEGF165-expressing tumor (c), and the VEGF189-expressing tumor (d). Four to six tumors of each class were analyzed, and each sample was stained as least twice with similar results. Arrows, VEGF staining. Arrowheads, vessels that stained positive for VEGF proteins. Original magnification: ×200.

Fig. 3.

Expression of exogenous VEGF isoforms in established U87 MG gliomas. A, Northern blot analysis. Top panel, samples of the U87 MG parental and the three types of VEGF isoform s.c. tumors are shown. In each class, tumors formed by the individual VEGF isoform expressing clones from each type were used. •, endogenous and ▸, exogenous VEGF mRNA species. Bottom panel, methylene blue staining of the membrane after the RNA was transferred into the membrane. ▪, 18 S and 28 S rRNA species. This analysis was repeated two additional times with similar results. B, expressions of VEGF isoform proteins detected by immunohistochemical analysis in the U87 MG parental tumor (a), the VEGF121-expressing tumor (b), the VEGF165-expressing tumor (c), and the VEGF189-expressing tumor (d). Four to six tumors of each class were analyzed, and each sample was stained as least twice with similar results. Arrows, VEGF staining. Arrowheads, vessels that stained positive for VEGF proteins. Original magnification: ×200.

Close modal
Fig. 4.

Association of VEGF with its receptors, VEGFR-1 or -2, in the various types of established U87 MG gliomas. Panels a–h, immunohistochemical stains of identical areas shown in Fig. 5 with a biotinylated anti-VEGF monoclonal antibody (11B5) that selectively recognizes VEGF in the complex with VEGFR-1 or -2. Gliomas formed by the U87 MG parental cells (a and b), the VEGF121-expressing cells (c and d), the VEGF165-expressing cells (e and f), and the VEGF189-expressing cells (g and h). Arrows, vessels that were identified in Figs. 2, and 5, A and B. Arrowheads, vessels that were not stained by 11B5 but were positive for CD31 (Fig. 2) and VEGFR (data not shown). Six or more individual tumor samples of each class were analyzed each time, and the experiments were repeated three times with similar results. Original magnification: ×200.

Fig. 4.

Association of VEGF with its receptors, VEGFR-1 or -2, in the various types of established U87 MG gliomas. Panels a–h, immunohistochemical stains of identical areas shown in Fig. 5 with a biotinylated anti-VEGF monoclonal antibody (11B5) that selectively recognizes VEGF in the complex with VEGFR-1 or -2. Gliomas formed by the U87 MG parental cells (a and b), the VEGF121-expressing cells (c and d), the VEGF165-expressing cells (e and f), and the VEGF189-expressing cells (g and h). Arrows, vessels that were identified in Figs. 2, and 5, A and B. Arrowheads, vessels that were not stained by 11B5 but were positive for CD31 (Fig. 2) and VEGFR (data not shown). Six or more individual tumor samples of each class were analyzed each time, and the experiments were repeated three times with similar results. Original magnification: ×200.

Close modal
Fig. 5.

Expression of VN, the ECM ligand to αvβ3 integrin, in the various types of established U87 MG gliomas. Panels a–h, immunohistochemical analysis of the various types of established U87 MG tumors using an anti-VN antibody. Gliomas formed by the U87 MG parental cells (a and b), the VEGF121-expressing cells (c and d), the VEGF165-expressing cells (e and f), and the VEGF189-expressing cells (g and h). Arrows, VN proteins were expressed and secreted into the ECM of the U87 MG tumors. Six or more individual tumor samples of each class were analyzed each time, and the experiments were repeated two additional times with similar results. Original magnification: ×400.

Fig. 5.

Expression of VN, the ECM ligand to αvβ3 integrin, in the various types of established U87 MG gliomas. Panels a–h, immunohistochemical analysis of the various types of established U87 MG tumors using an anti-VN antibody. Gliomas formed by the U87 MG parental cells (a and b), the VEGF121-expressing cells (c and d), the VEGF165-expressing cells (e and f), and the VEGF189-expressing cells (g and h). Arrows, VN proteins were expressed and secreted into the ECM of the U87 MG tumors. Six or more individual tumor samples of each class were analyzed each time, and the experiments were repeated two additional times with similar results. Original magnification: ×400.

Close modal
Fig. 6.

VN potentiates VEGF stimulated EC migration. A, cell migration as function of concentrations of recombinant VEGF121 and VEGF165 proteins. Various concentrations of VEGF proteins were included in the assays. The EC were stimulated continuously for 4 h. B, cell migration stimulated by various amounts of VEGF121 and VEGF165 in the presence or absence of VN (200 ng/ml). C and D, cell migrations stimulated by VEGF121 (panel C; 5 ng/ml) and VEGF165 (panel D; 15 ng/ml) at various concentrations of VN or without VN. The cell migration assays were performed as described in “Materials and Methods.” The data were shown as means; bars,± SD. The experiments were repeated three independent times with similar results.

Fig. 6.

VN potentiates VEGF stimulated EC migration. A, cell migration as function of concentrations of recombinant VEGF121 and VEGF165 proteins. Various concentrations of VEGF proteins were included in the assays. The EC were stimulated continuously for 4 h. B, cell migration stimulated by various amounts of VEGF121 and VEGF165 in the presence or absence of VN (200 ng/ml). C and D, cell migrations stimulated by VEGF121 (panel C; 5 ng/ml) and VEGF165 (panel D; 15 ng/ml) at various concentrations of VN or without VN. The cell migration assays were performed as described in “Materials and Methods.” The data were shown as means; bars,± SD. The experiments were repeated three independent times with similar results.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Supported in part by a Sidney Kimmel Scholar Award, a Brain Cancer Program Grant of the James F. McDonnell Foundation, a The Brain Tumor Society Research Grant, and start-up funds from the University of Pittsburgh Cancer Institute (to S-Y. C.). M. N. was a fellow of the Japan Brain Foundation.

4

The abbreviations used are: VEGF, vascular endothelial growth factor; VEGFR, VEGF receptor; EC, endothelial cell; NRP, neuropilin; ECM, extracellular matrix; PAE/KDR, porcine aortic ECs expressing exogenous VEGFR-2; IHC, immunohistochemistry; FN, fibronectin; VN, vitronectin; NS, not significant; i.c. intracranial; KDR, kinase insert domain-containing receptor.

5

P. Guo, unpublished observations.

Table 1

Effects on i.c. tumorigenesis of U87MG gliomas by overexpression of three VEGF isoforms

CellsTumor volumes (mm3, mean± SE)Days of postimplantation (Mean± SE)
U87MG 56.8 ± 4.52 (n = 22) 40.2 ± 2.20 
VEGF121 54.1 ± 4.25 (n = 16) 30.5 ± 1.19 (P < 0.0004) 
VEGF165 57.2 ± 2.62 (n = 20) 25.1 ± 1.52 (P < 0.0000) 
VEGF189 58.8 ± 3.12 (n = 18) 28.5 ± 2.87 (P < 0.0000) 
CellsTumor volumes (mm3, mean± SE)Days of postimplantation (Mean± SE)
U87MG 56.8 ± 4.52 (n = 22) 40.2 ± 2.20 
VEGF121 54.1 ± 4.25 (n = 16) 30.5 ± 1.19 (P < 0.0004) 
VEGF165 57.2 ± 2.62 (n = 20) 25.1 ± 1.52 (P < 0.0000) 
VEGF189 58.8 ± 3.12 (n = 18) 28.5 ± 2.87 (P < 0.0000) 
Table 2

Overexpression of three VEGF isoforms elicited angiogenicity in established U87MG gliomas at different anatomic sites

Type of gliomasVessel densities (pixels/mm2,± SE)
s.c.i.c.
U87MG 562.3 ± 38 (1.0) 635.2 ± 48 (1.0) 
VEGF121 728.5 ± 27 (1.3) 1869.7 ± 28 (2.9) 
VEGF165 3081.2 ± 142 (5.5) 3296.4 ± 234 (5.2) 
VEGF189 3468.6 ± 151 (6.2) 3382.7 ± 176 (5.3) 
Type of gliomasVessel densities (pixels/mm2,± SE)
s.c.i.c.
U87MG 562.3 ± 38 (1.0) 635.2 ± 48 (1.0) 
VEGF121 728.5 ± 27 (1.3) 1869.7 ± 28 (2.9) 
VEGF165 3081.2 ± 142 (5.5) 3296.4 ± 234 (5.2) 
VEGF189 3468.6 ± 151 (6.2) 3382.7 ± 176 (5.3) 

We thank Frank Cackowski for editing of this manuscript. We also thank Alex Kolodkin at Johns Hopkins University, Baltimore, Maryland, for the rabbit polyclonal antirat NRP-1 antibody; Barry S. Coller at Mount Sinai School of Medicine, New York, New York, for the anti-β3 antibody (c7E3); T. Byzova and Vickey Byers-Ward at Cleveland Clinic Foundation, Cleveland, Ohio, for advice in the use of modified EC migration assays; Candace Johnson and Ruth Modzelewski at the University of Pittsburgh, Pittsburgh, Pennsylvania, for mouse EC RNAs; Xiao Xiao at the University of Pittsburgh, for use of the cryostat; and Xiang-Dong Ji at PharMingen, San Diego, California, for preparing some of the cryostat sections.

1
Ferrara N. Vascular endothelial growth factor: molecular and biological aspects.
Curr. Top. Microbiol. Immunol.
,
237
:
1
-30,  
1999
.
2
Fong G. H., Zhang L., Bryce D. M., Peng J. Increased hemangioblast commitment, not vascular disorganization, is the primary defect in flt-1 knock-out mice.
Development
,
126
:
3015
-3025,  
1999
.
3
Shalaby F., Rossant J., Yamaguchi T. P., Gertsenstein M., Wu X. F., Breitman M. L., Schuh A. C. Failure of blood-island formation and vasculogenesis in Flk-1-deficient mice.
Nature (Lond.)
,
376
:
62
-66,  
1995
.
4
Laird A. D., Vajkoczy P., Shawver L. K., Thurnher A., Liang C., Mohammadi M., Schlessinger J., Ullrich A., Hubbard S. R., Blake R. A., Fong T. A., Strawn L. M., Sun L., Tang C., Hawtin R., Tang F., Shenoy N., Hirth K. P., McMahon G., Cherrington J. SU6668 is a potent antiangiogenic and antitumor agent that induces regression of established tumors.
Cancer Res.
,
60
:
4152
-4160,  
2000
.
5
Miquerol L., Langille B. L., Nagy A. Embryonic development is disrupted by modest increases in vascular endothelial growth factor gene expression.
Development
,
127
:
3941
-3946,  
2000
.
6
Cheng S. Y., Huang H. J., Nagane M., Ji X. D., Wang D., Shih C. C., Arap W., Huang C. M., Cavenee W. K. Suppression of glioblastoma angiogenicity and tumorigenicity by inhibition of endogenous expression of vascular endothelial growth factor.
Proc. Natl. Acad. Sci. USA
,
93
:
8502
-8507,  
1996
.
7
Okada F., Rak J. W., Croix B. S., Lieubeau B., Kaya M., Roncari L., Shirasawa S., Sasazuki T., Kerbel R. S. Impact of oncogenes in tumor angiogenesis: mutant K-ras up-regulation of vascular endothelial growth factor/vascular permeability factor is necessary, but not sufficient for tumorigenicity of human colorectal carcinoma cells.
Proc. Natl. Acad. Sci. USA
,
95
:
3609
-3614,  
1998
.
8
Thurston G., Suri C., Smith K., McClain J., Sato T. N., Yancopoulos G. D., McDonald D. M. Leakage-resistant blood vessels in mice transgenically overexpressing angiopoietin-1.
Science (Wash. DC)
,
286
:
2511
-2514,  
1999
.
9
Claffey K. P., Brown L. F., del Aguila L. F., Tognazzi K., Yeo K. T., Manseau E. J., Dvorak H. F. Expression of vascular permeability factor/vascular endothelial growth factor by melanoma cells increases tumor growth, angiogenesis, and experimental metastasis.
Cancer Res.
,
56
:
172
-181,  
1996
.
10
Cheng S-Y., Nagane M., Su Huang H-J., Cavenee W. K. Intracerebral tumor-associated hemorrhage caused by overexpression of vascular endothelial growth factor (VEGF) isoforms, VEGF121, and VEGF165 but not VEGF189.
Proc. Natl. Acad. Sci. USA
,
94
:
12081
-12087,  
1997
.
11
Robinson C. J., Stringer S. E. The splice variants of vascular endothelial growth factor (VEGF) and their receptors.
J. Cell Sci.
,
114
:
853
-865,  
2001
.
12
Keyt B. A., Berleau L. T., Nguyen H. V., Chen H., Heinsohn H., Vandlen R., Ferrara N. The carboxyl-terminal domain (111–165) of vascular endothelial growth factor is critical for its mitogenic potency.
J. Biol. Chem.
,
271
:
7788
-7795,  
1996
.
13
Oshika Y., Nakamura M., Tokunaga T., Ozeki Y., Fukushima Y., Hatanaka H., Abe Y., Yamazaki H., Kijima H., Tamaoki N., Ueyama Y. Expression of cell-associated isoform of vascular endothelial growth factor 189 and its prognostic relevance in non-small cell lung cancer.
Int. J. Oncol.
,
12
:
541
-544,  
1998
.
14
Cheung N., Wong M. P., Yuen S. T., Leung S. Y., Chung L. P. Tissue-specific expression pattern of vascular endothelial growth factor isoforms in the malignant transformation of lung and colon.
Hum. Pathol.
,
29
:
910
-914,  
1998
.
15
Mori A., Arii S., Furutani M., Hanaki K., Takeda Y., Moriga T., Kondo Y., Gorrin Rivas M. J., Imamura M. Vascular endothelial growth factor-induced tumor angiogenesis and tumorigenicity in relation to metastasis in a HT1080 human fibrosarcoma cell model.
Int. J. Cancer
,
80
:
738
-743,  
1999
.
16
Zhang H. T., Scott P. A., Morbidelli L., Peak S., Moore J., Turley H., Harris A. L., Ziche M., Bicknell R. The 121 amino acid isoform of vascular endothelial growth factor is more strongly tumorigenic than other splice variants in vivo.
Br. J. Cancer.
,
83
:
63
-68,  
2000
.
17
Carmeliet P., Ng Y. S., Nuyens D., Theilmeier G., Brusselmans K., Cornelissen I., Ehler E., Kakkar V. V., Stalmans I., Mattot V., Perriard J. C., Dewerchin M., Flameng W., Nagy A., Lupu F., Moons L., Collen D., D’Amore P. A., Shima D. T. Impaired myocardial angiogenesis and ischemic cardiomyopathy in mice lacking the vascular endothelial growth factor isoforms VEGF164 and VEGF188.
Nat. Med.
,
5
:
495
-502,  
1999
.
18
Feng D., Nagy J. A., Brekken R. A., Pettersson A., Manseau E. J., Pyne K., Mulligan R., Thorpe P. E., Dvorak H. F., Dvorak A. M. Ultrastructural localization of the vascular permeability factor/vascular endothelial growth factor (VPF/VEGF) receptor-2 (FLK-1, KDR) in normal mouse kidney and in the hyperpermeable vessels induced by VPF/VEGF-expressing tumors and adenoviral vectors.
J. Histochem. Cytochem.
,
48
:
545
-556,  
2000
.
19
Brekken R. A., Huang X., King S. W., Thorpe P. E. Vascular endothelial growth factor as a marker of tumor endothelium.
Cancer Res.
,
58
:
1952
-1959,  
1998
.
20
Houck K. A., Leung D. W., Rowland A. M., Winer J., Ferrara N. Dual regulation of vascular endothelial growth factor bioavailability by genetic and proteolytic mechanisms.
J. Biol. Chem.
,
267
:
26031
-26037,  
1992
.
21
Soker S., Takashima S., Miao H. Q., Neufeld G., Klagsbrun M. Neuropilin-1 is expressed by endothelial and tumor cells as an isoform-specific receptor for vascular endothelial growth factor.
Cell
,
92
:
735
-745,  
1998
.
22
Gluzman-Poltorak Z., Cohen T., Herzog Y., Neufeld G. Neuropilin-2 and neuropilin-1 are receptors for the 165-amino acid form of vascular endothelial growth factor (VEGF) and of placenta growth factor-2, but only neuropilin-2 functions as a receptor for the 145- amino acid form of VEGF.
J. Biol. Chem.
,
275
:
18040
-18045,  
2000
.
23
Fujisawa H., Kitsukawa T., Kawakami A., Takagi S., Shimizu M., Hirata T. Roles of a neuronal cell-surface molecule, neuropilin, in nerve fiber fasciculation and guidance.
Cell Tissue Res.
,
290
:
465
-470,  
1997
.
24
Kawasaki T., Kitsukawa T., Bekku Y., Matsuda Y., Sanbo M., Yagi T., Fujisawa H. A requirement for neuropilin-1 in embryonic vessel formation.
Development
,
126
:
4895
-4902,  
1999
.
25
Kitsukawa T., Shimono A., Kawakami A., Kondoh H., Fujisawa H. Overexpression of a membrane protein, neuropilin, in chimeric mice causes anomalies in the cardiovascular system, nervous system and limbs.
Development
,
121
:
4309
-4318,  
1995
.
26
Kolodkin A. L., Levengood D. V., Rowe E. G., Tai Y. T., Giger R. J., Ginty D. D. Neuropilin is a semaphorin III receptor.
Cell
,
90
:
753
-762,  
1997
.
27
Gagnon M. L., Bielenberg D. R., Gechtman Z., Miao H. Q., Takashima S., Soker S., Klagsbrun M. Identification of a natural soluble neuropilin-1 that binds vascular endothelial growth factor: in vivo expression and antitumor activity.
Proc. Natl. Acad. Sci. USA
,
97
:
2573
-2578,  
2000
.
28
Rossignol M., Gagnon M. L., Klagsbrun M. Genomic organization of human neuropilin-1 and neuropilin-2 genes: identification and distribution of splice variants and soluble isoforms.
Genomics
,
70
:
211
-222,  
2000
.
29
Bergers G., Brekken R., McMahon G., Vu T. H., Itoh T., Tamaki K., Tanzawa K., Thorpe P., Itohara S., Werb Z., Hanahan D. Matrix metalloproteinase-9 triggers the angiogenic switch during carcinogenesis.
Nat. Cell. Biol.
,
2
:
737
-744,  
2000
.
30
Hynes R. O. Integrins: versatility, modulation, and signaling in cell adhesion.
Cell
,
69
:
11
-25,  
1992
.
31
Gladson C. L., Cheresh D. A. Glioblastoma expression of vitronectin and the α v β 3 integrin. Adhesion mechanism for transformed glial cells.
J. Clin. Investig.
,
88
:
1924
-1932,  
1991
.
32
Gladson C. L., Wilcox J. N., Sanders L., Gillespie G. Y., Cheresh D. A. Cerebral microenvironment influences expression of the vitronectin gene in astrocytic tumors.
J. Cell Sci.
,
108
:
947
-956,  
1995
.
33
Byzova V. T., Goldman K. C., Pampori N., Thomas A. K., Bett A., Shattil J. S., Plow F. E. A mechanism for modulation of cellular responses to VEGF: activation of the integrins.
Mol. Cell
,
6
:
851
-860,  
2000
.
34
Park J. E., Keller G. A., Ferrara N. The vascular endothelial growth factor (VEGF) isoforms: differential deposition into the subepithelial extracellular matrix and bioactivity of extracellular matrix-bound VEGF.
Mol. Biol. Cell.
,
4
:
1317
-1326,  
1993
.
35
Carmebet P. VEGF gene therapy: stimulating angiogenesis or angioma-genes?.
Nat. Med.
,
6
:
1102
-1103,  
2000
.
36
Grunstein J., Masbad J. J., Hickey R., Giordano F., Johnson R. S. Isoforms of vascular endothelial growth factor act in a coordinate fashion to recruit and expand tumor vasculature.
Mol. Cell. Biol.
,
20
:
7282
-7291,  
2000
.
37
Christofori G., Naik P., Hanahan D. Vascular endothelial growth factor and its receptors, flt-1 and flk-1, are expressed in normal pancreatic islets and throughout islet cell tumorigenesis.
Mol. Endocrinol.
,
9
:
1760
-1770,  
1995
.
38
Hiratsuka S., Minowa O., Kuno J., Noda T., Shibuya M. Flt-1 lacking the tyrosine kinase domain is sufficient for normal development and angiogenesis in mice.
Proc. Natl. Acad. Sci. USA
,
95
:
9349
-9354,  
1998
.
39
Koukourakis M. I., Giatromanolaki A., Thorpe P. E., Brekken R. A., Sivridis E., Kakolyris S., Georgoulias V., Gatter K. C., Harris A. L. Vascular endothelial growth factor/KDR activated microvessel density versus CD31 standard microvessel density in non-small cell lung cancer.
Cancer Res.
,
60
:
3088
-3095,  
2000
.
40
Arbiser J. L., Larsson H., Claesson-Welsh L., Bai X., LaMontagne K., Weiss S. W., Soker S., Flynn E., Brown L. F. Overexpression of VEGF 121 in immortalized endothelial cells causes conversion to slowly growing angiosarcoma and high level expression of the VEGF receptors VEGFR-1 and VEGFR-2 in vivo.
Am. J. Pathol.
,
156
:
1469
-1476,  
2000
.
41
Eliceiri B. P., Cheresh D. A. The role of αv integrins during angiogenesis: insights into potential mechanisms of action and clinical development.
J. Clin. Investig.
,
103
:
1227
-1230,  
1999
.
42
MacDonald T. J., Taga T., Shimada H., Tabrizi P., Zlokovic B. V., Cheresh D. A., Laug W. E. Preferential susceptibility of brain tumors to the antiangiogenic effects of an α(v) integrin antagonist.
Neurosurgery
,
48
:
151
-157,  
2001
.
43
Chatterjee S., Matsumura A., Schradermeier J., Gillespie G. Y. Human malignant glioma therapy using anti-α(v)β3 integrin agents.
J. Neuro-Oncol.
,
46
:
135
-144,  
2000
.
44
Naito S., von Eschenbach A. C., Giavazzi R., Fidler I. J. Growth and metastasis of tumor cells isolated from a human renal cell carcinoma implanted into different organs of nude mice.
Cancer Res.
,
46
:
4109
-4115,  
1986
.
45
Singh R. K., Bucana C. D., Gutman M., Fan D., Wilson M. R., Fidler I. J. Organ site-dependent expression of basic fibroblast growth factor in human renal cell carcinoma cells.
Am. J. Pathol.
,
145
:
365
-374,  
1994
.
46
Yarwood S. J., Woodgett J. R. Extracellular matrix composition determines the transcriptional response to epidermal growth factor receptor activation.
Proc. Natl. Acad. Sci. USA
,
98
:
4472
-4477,  
2001
.
47
Pluen A., Boucher Y., Ramanujan S., McKee T. D., Gohongi T., di Tomaso E., Brown E. B., Izumi Y., Campbell R. B., Berk D. A., Jain R. K. Role of tumor-host interactions in interstitial diffusion of macromolecules: cranial vs. subcutaneous tumors.
Proc. Natl. Acad. Sci. USA
,
98
:
4628
-4633,  
2001
.