Expression of vascular endothelial growth factor (VEGF) and platelet ...

3 downloads 0 Views 197KB Size Report
Journal of Molecular Neuroscience. 177. Volume 13, 1999. Expression of Vascular Endothelial Growth Factor. (VEGF) and Platelet-Derived Growth Factor.
Journal of Molecular Neuroscience Copyright © 1999 Humana Press Inc. All rights of any nature whatsoever reserved. ISSN0895-8696/99/13:177–185/$12.25

Expression of Vascular Endothelial Growth Factor (VEGF) and Platelet-Derived Growth Factor Receptor-β (PDGFR-β) in Human Gliomas J. V. Lafuente,*,1 B. Adán,1 K. Alkiza,1 J. M. Garibi,2 M. Rossi,3 and F. F. Cruz-Sánchez 4 1

Department of Neurosciences, Basque Country University, Leioa, Spain; 2Department of Neurosurgery, Cruces Hospital, Barakaldo, Spain; 3Department of Neuropathology, Walton Centre for Neurology & Neurosurgery, Liverpool, UK; and 4Institute of Neurological and Gerontological Sciences, International University of Catalonie, Barcelona, Spain

Abstract The growth of solid tumors is highly dependent on vascular proliferation. Vascular endothelial growth factor (VEGF), the main mediator of angiogenesis, and platelet-derived growth factor receptor-β (PDGFR-β), receptor for the potent mitogen PDGF, are two indicators of the angiogenic potential of human gliomas. We studied a series of 57 surgical biopsies of astrocytic neoplasms by immunohistochemistry to elucidate the relationship between tumor proliferation, quantified as Ki67-LI, and the expression of these two proteins. Ki67-LI increases throughout histological malignancy, although staining in endothelial cells has rarely been recorded. Elevated amounts of VEGF-positive tumor cells (VEGF-LI) were found in anaplastic astrocytomas and glioblastomas, mainly around areas of necrosis, cysts, or edema. Endothelium of blood vessels was consistently stained. PDGFR-β positivity was found in glomeruloid formations and in tumor cells, excluding pilocytic astrocytomas. Multinucleated giant cells and perivascular tumor cells were positive in glioblastomas. In addition, peritumoral microglia-like cells were also stained in some cases. Statistical correlation was only found between PDGFR-β and Ki67 LIs. In conclusion, VEGF as permeability factor is involved in the development of secondary neoplastic changes, whereas PDGFR-β is directly correlated to proliferation indexes. Strong expression of VEGF and PDGFR-β found in endothelium and tumor cells would seem to support a combined role in tumoral neoangiogenesis Index Entries: Angiogenesis; brain tumors; proliferation; VEGF; PDGFR-β.

*Author to whom all correspondence and reprint requests should be addressed.

Journal of Molecular Neuroscience

177

Volume 13, 1999

178

Introduction The growth of solid tumors is largely dependent on their ability to induce angiogenesis. Vascular proliferation in gliomas has traditionally been considered a feature of diagnostic value (Zülch, 1979). Several cytokines which modulate angiogenesis, including vascular endothelial growth factor (VEGF) (Ferrara and Henzel, 1989) and plateletderived growth factor (PDGF) (Hermansson et al., 1988), are present in human gliomas. Most evidence points to VEGF as the main mediator of angiogenesis (Plate et al., 1992b; Schweiki et al., 1992). VEGF is an extensively characterized 46-kDa dimeric glycoprotein inducing endothelial proliferation and capillary permeability (Conolly et al., 1989). Biochemical homology exists between VEGF and PDGF, but different cell types respond differently (Tischer et al., 1991). PDGF is a dimeric molecule of around 30 kDa that has been identified as a potent mitogen for mesenchymal and neuroectodermal cells, whereas their receptors are structurally related to transmembrane glycoproteins constituting a subfamily within the superfamily of tyrosine kinases (Hanks et al., 1988). Plate et al. (1992a) demonstrated that a malignant phenotype in gliomas is associated with upregulation of plateletderived growth factor receptor-β (PDGFR-β) expression by endothelial cells. These findings suggest that PDGFR-β and VEGF are markers for evaluating angiogenic potential in human gliomas. Although angiogenesis and tumor proliferation are two closely interconnected events, few studies have integrated these two key aspects (Nakamura et al., 1993; Plate et al., 1994). The aim of the present study is to elucidate the relationships among histological features, proliferation (cell-cycle-related protein Ki67-LI [Gerdes et al., 1984, 1991; Verheijen et al., 1989]), and angiogenic potential (immunohistochemical expression of VEGF and PDGFR-β) in gliomas.

Materials and Methods Fifty-seven neoplasms grouped according to the WHO criteria (Kleihues et al., 1993) (5 pilocytic astrocytomas, 10 low-grade astrocytomas, 9 anaplastic astrocytomas, and 33 glioblastoma multiforme) were studied. Clinical data are shown in Table 1.

Journal of Molecular Neuroscience

Lafuente et al. All tissue samples were fixed in 4% paraformaldehyde and routinely embedded in paraffin. Tissue blocks were cut at 5-µm thick sections. Prior to the immunohistochemical procedure, slices were deparaffinized in two xylene baths for 10 min each, followed by two 5-min washes in acetone. In order to block endogenous peroxidase, slices were immersed in methanol with 0.4% hydrogen peroxide for 30 min and then washed and rehydrated in 0.1 M phosphate-buffered saline (PBS) (pH 7.4).

Immunohistochemistry for VEGF and PDGFR-β Histological sections were preincubated in a moist chamber with a solution of 0.1 M PBS (pH 7.4) containing 0.03% casein for 30 min at room temperature to reduce undesirable background staining occurring during immunohistochemical processing of paraffin-embedded samples. Slices were then incubated at 4°C overnight with two commercially available polyclonal antisera: 1) antiVEGF165, against the amino-terminal 20 amino acids of human VEGF (Santa Cruz Biotechnology Inc., Santa Cruz, CA) and 2) anti-PDGFR-β, raised against a GST fusion protein construct containing PFGFR-β sequences corresponding to amino acid residues 958-1106 mapping at the carboxy-terminus of the type B PDGFR of human origin (Santa Cruz Biotechnology). Both antibodies were used at a concentration of 2 µg/mL.

Ki67 Immunohistochemistry For Ki67, specific antigen retrieval was needed after normal goat serum incubation by immersing sections in 10 mM citrate buffer (pH 6.0) and heating them in a microwave oven for 5-7 min. Samples were then incubated overnight using a rabbit antihuman Ki67 antiserum (Dako) at 1:100 dilution. This antiserum has a reactivity similar to that obtained with the monoclonal anti-Ki-67 antibody (clone MIB-1) (Key et al., 1993). In all immunohistochemical procedures, control sections were processed in the same way, except for the primary antiserum, which was replaced with PBS. After washing in PBS, slices were incubated for 30 min at room temperature with corresponding biotinylated goat antirabbit immunoglobulin and peroxidase-conjugated avidin–biotin complex (ABC kit Elite, Vector Laboratories, Burlingame,

Volume 13, 1999

Expression of VEGF and PDGFR-β

179

Table 1 Distribution of Patients According to Histological Typesa Sex Age, mean,

Ki-67 LI, %

VEGF LI, %

Mean

Mean

SE

SE

Histological type

Range, yr

M

F

Range

Range

Pilocytic astroc. (no = 5) Low-grade astroc. (no = 10) Anaplastic astroc. (no = 9) Glioblast multif. (no = 33)

10 4 – 18 42 27 – 61 51 28 – 72 54 28 – 75

2

3

4

6

2

7

23

10

2.63 ± 0.16 0.25 – 5.8 0.45 ± 0.06 0.25 – 0.75 6.29 ± 0.24 3.37 – 9.75 17.04 ± 0.37 2.96 – 45.19

30.61 ± 0.46 21.52 – 38.72 25.00 ± 0.43 11.98 – 44.89 23.53 ± 0.42 10.00 – 38.15 33.09 ± 0.47 12.28 – 64.92

PDGFR-β LI, % Mean

SE

Range — 19.92 ± 040 12.17 – 33.96 16.61 ± 0.37 6.01 – 21.45 26.88 ± 0.44 10.31 – 53.77

a

Age, sex, and mean percentages of positive tumor cells for Ki67, VEGF, and PDGFR-β are shown. SE = standard error.

CA). The final reaction product was developed by immersing slices in a solution containing 0.4 mg/mL 3,3-diaminobenzidine and 0.1% hydrogen peroxide in 0.1 M PBS (pH = 7.4). Specific positivity was identified as brown cytoplasm/membrane stain (VEGF and PDGFR-β) and nuclear staining (Ki67). Sections were counterstained with hematoxylin for further microscopical observation. To estimate proliferative activity, Ki67 labeling index (% of nuclei expressing the antigen) was obtained by counting 400 tumor cell nuclei from each case. Tumor cells expressing the antigen to any extent (light to dark brown staining) were considered positive. Areas with the highest density of positive cells were selected for quantitative evaluation performed by the same researcher in order to avoid interobserver variability. Immunoreactivity for VEGF and PDGFR-β was recorded in both tumor cells and vascular endothelia, but the labeling index (LI) for VEGF and PDGFR-β was only evaluated by counting 400 tumor cells in each case.

Statistical Analysis Spearman coefficient of correlation was used to correlate single VEGF and PDGFR-β LIs, and also to correlate each of them to their respective Ki-67 LI. Statistical significance was determined by Student’s t-test (p = 0.05). The H-test (Kruskal-Wallis)

Journal of Molecular Neuroscience

was also used to determine the ability of these markers to separate out histological groupings.

Results Ki67 LI values increased with the histological degree of anaplasia, and the highest number of positive tumor nuclei was found in glioblastomas (Table 1). Blood vessel proliferation was conspicuous in glioblastomas, but endothelial cell nuclei were positive for Ki67 in only two cases, which also showed strong immunoreactivity for VEGF and PDGFR-β in both vessels and tumor cells.

VEGF Immunoreactivity In normal brain tissue, VEGF immunoreaction was restricted to a mild stain in neurons and vascular endothelium, remainig negative the glial cells. All tumors showed positive endothelial and tumor cells, the pattern of positivity being finely granular with diffuse cytoplasmic distribution. In lowgrade astrocytomas and occasionally in peritumor tissue, endothelial staining highlighted the vascular wall for some length. In peritumoral areas, intensely VEGF-positive hypertrophic astrocytes were found (Fig. 1A). The mean percentage of positive cells (VEGF LI) is similar for the different histological groups (Table 1). Within tumor proper,

Volume 13, 1999

180

Lafuente et al.

Fig. 1. VEGF expression in astrocytic glioma. (A) Positivity in reactive hypertrophic astrocytes and in the vessel wall in peritumoral area (200x). (B) Immunostaining of vascular formations surrounding cysts in a pilocytic astrocytoma and moderate staining of tumor cells (165x). (C) Cytoplasmic staining in perivascular tumor cells and slight positivity of endothelial cells in a glioblastoma multiforme (210x).

scattered, lightly stained tumor cells were observed. Pilocytic astrocytomas showed positive immunoreactivity in vessels, mainly next to microcysts (Fig. 1B). In glioblastomas, a higher number of cells completely covered by the brown precipitate is remarkable, mainly around vessels (Fig. 1C). Cells expressing VEGF were often also found in the proximity of necrosis. Tumor areas with an edematous appearance showed consistent staining for VEGF

Journal of Molecular Neuroscience

in endothelium, in some perivascular tumor cells and even in the extracellular matrix. There was no statistical correlation between VEGF- and Ki67-LIs.

PDGFR-β Immunoreactivity Low-grade astrocytomas showed variable endothelial staining for PDGFR-β and strong cytoplasmic positivity in clustered tumor cells. No pos-

Volume 13, 1999

Expression of VEGF and PDGFR-β

181

Fig. 2. PDGFR-β expression in astrocytic glioma. (A) Strong immunostaining of cubicated vessel wall cells in a glomeruloid vessel (220x). (B) Highly undifferentiated tumor cells in glioblastoma showing granular staining in cytoplasm (455x). (C) PDGFR-β staining of microglia located in peritumoral tissue of an anaplastic astrocytoma (235x).

itivity was observed in pilocytic astrocitomas. In anaplastic astrocytomas and glioblastomas, immunopositivity was observed in both endothelium and tumor cells (Table 1). Heterogeneous distribution of staining was frequently found in

Journal of Molecular Neuroscience

endothelium of different areas of glioblastomas. An intense granular pattern of staining was seen in glomeruloid vessels between peritumor and tumor areas. Vessels with several concentric layers of cuboidal cells showed strong positivity (Fig. 2A).

Volume 13, 1999

182 Positive giant-tumor cells in glioblastoma, including gemistocytes, large binucleated cells, and bizarre tumor cells (Fig. 2B), were frequently PDGFR-β-positive with an irregular, granular cytoplasmic pattern. In highly cellular glioblastomas, strongly positive cells were mainly distributed around weakly stained proliferating vessels. Peritumor tissue in glioblastomas showed positive elongated bipolar cells with hyperchromatic bar-like nucleus resembling microglia (Fig. 2C). Both endothelia and glial cells were consistently immunonegative in nonneoplastic brain tissue. The PDGFR-β-labeling index is similar for several histological groups (Table 1). Correlation between PDGFR-β and Ki67 LIs was statistically significant (r = 0.4; p = 0.03), i.e., an increase in PDGFR-β expression corresponded to an increase in proliferation. Statistical analysis showed significant correlation between VEGF and PDGFR-β LIs (r = 1; p = 0.0001). Kruskal Wallis analysis showed that Ki67 LIs were significantly different between histological groups (p = 0.0001), whereas angiogenicrelated proteins did not show differences statistically significant between histological groups.

Discussion Although VEGF and PDGFR-β mRNAs have been demonstrated in gliomas by in situ hybridization (Plate et al., 1992a,b), few immunohistochemical studies have been carried out to detect the corresponding proteins. The present study provides data on the distribution of VEGF and PDGFRβ proteins in human astrocytic gliomas and their relationship with tumor proliferation. Of the four different VEGF isoforms, only three (VEGF121, VEGF165, VEGF189) are present in human brain tumors. VEGF 165 is the most frequent (Berkman et al., 1993) and may be found bound to cell membranes or secreted in soluble form. When tumor cells switch to the angiogenic phenotype, they induce the formation of new capillary blood vessels, which ensure that tumor growth is sustained (Folkman, 1995). Vascular proliferation could be found in every degree of gliomas, as we have seen using immunohistological markers, but is particularly evident in glioblastomas (Russell and Rubinstein, 1989; Hatva et al., 1995). However, staining

Journal of Molecular Neuroscience

Lafuente et al. intensity and the number of positive tumor cells increase with increasing anaplasia. Immunohistochemistry provides evidence of the strong expression of VEGF by tumor cells of malignant astrocytomas and also by the highly permeable proliferating microvessels (Plate et al., 1992b). However, these microvessels do not seem to express VEGF mRNA as judged by in situ hybridization (Dvorak et al., 1995). Therefore, the vascular staining observed in our study could be caused by the binding of VEGF secreted by adjacent tumor cells, which would explain the higher density of strongly positive cells seen around VEGF-positive vessels. We found that VEGF positivity in both endothelium and tumor cells in astrocytomas was not directly correlated with proliferation, i.e., proliferation rates (Ki67-LI) were similar in tumors with a low or elevated percentage of VEGF-positive cells. These findings agree with the published data on VEGF mRNA (Plate et al., 1994). VEGF has been proposed as the major inducer of endothelial proliferation (Pepper et al., 1991; Unemori et al., 1992); however, few positive endothelial cells for VEGF actually showed positivity for Ki67 in the gliomas studied. On the other hand, VEGF overexpression is also related to other associated effects, such as edema, cyst formation, and necrosis (Weindel et al., 1994). We found strong immunoreactivity for VEGF in endothelium and even in extracellular matrix (ECM) in very edematous areas. The fact that the secreted forms of VEGF (initially described as vascular permeability factor [VPF]) bind heparin-rich ECM proteins should be noted. Our results support the contention that VEGF, acting as permeability factor, is involved in peritumoral edema formation (Tsai et al., 1995). In addition, VEGF has been proposed as the common pathophysiological link between peritumoral edema and tumor microcyst formations (Schweiki et al., 1992; Strugar et al., 1995). VEGF can increase von Willebrand factor release from human endothelial cells (Brock et al., 1991). Consistent decrease of von Willebrand factor expression in glioblastoma vessels has been described recently (Lafuente et al., 1995). VEGFstimulated endothelial cells cannot store von Willebrand factor and release it into plasma. The subsequent local hypercoagulability leads to microthrombi formation contributing to the pro-

Volume 13, 1999

Expression of VEGF and PDGFR-β duction of foci of necrosis in glioblastoma (Tsai et al., 1995). This is consistent with the observation of numerous vessels close to necrotic foci, which are strongly positive for VEGF and do not express von Willebrand factor (Adan, personal communication). PDGF is a potent mitogen, but its effect on endothelial cells has not sufficiently been elucidated. However, upregulation of β-receptor expression in vascular endothelial cells is associated with a malignant glioma phenotype, thus suggesting its involvement in angiogenesis (Hermansson et al., 1988). However, pericytes and vascular smooth muscle cells also express PDGF receptors-β in microvascular proliferations, supporting the contribution of these cells in tumor angiogenesis (Wesseling et al., 1995). Our results show a direct correlation between expression of PDGFR-β and proliferative potential (Ki67-LI) of astrocytomas. PDGFR-β expression by vessel wall cells has been thought to be induced by glioma cells (Plate et al., 1992a) and could therefore be interpreted as a feature pointing to invasiveness. Although PDGF is not an endothelial mitogen by itself, it may be considered as an indirect mediator of angiogenic response. PDGF originates from platelet lysates or through secretion by glioma cells and stimulates PDGFR-β positive astrocytic tumor cell proliferation with induction of VEGF production (Tsai et al., 1995). Our results show a linear correlation between expression of PDGFR-β and VEGF proteins, especially in malignant astrocytomas. Expression of PDGFR-β protein by microglia has not previously been described. According to a previous study (Lafuente et al., 1995), Ricinus communis agglutinin (RCA) lectin-positivity would confirm the microglial nature of positive cells. Weller et al. (1990) reported on the possible role of PDGF secreted by local microglia in proliferative vitreoretinopathy where PDGF expression could be explained through the mesenchymal origin of microglial cells. The presence of PDGFR-β in human monocyte-derived macrophages has been reported recently, suggesting that PDGF influences the function of macrophages and smooth muscle cells in the vascular wall in atherosclerosis (Inaba et al., 1993). Our findings indicate that microglia express PDGFR-β in peritumoral tissue where events, such as vascular proliferation and vessel wall repair, take place. It is possible that microglial cells in glial neo-

Journal of Molecular Neuroscience

183 plasms are also susceptible to PDGF-mediated physiological responses. To sum up, although Ki67 LI is a useful marker to segregate histological groups, the studied related angiogenic proteins did not show any value for identifying glioma histological subtypes. However, strong expression of VEGF and PDGFR-β found in endothelium and tumor cells would seem to support their combined role in neoangiogenesis. VEGF also plays a prominent role as permeability factor linked to secondary neoplastic changes and PDGFR-β expression correlated with proliferation. Further studies of angiogenesis-linked proteins are needed to elucidate their relation to the proliferative and invasive potential of gliomas.

Acknowledgments This work was supported by projects FIS 92/ 0754 and UPV 102/92. B. Adán and K. Alkiza have fellowships from the Basque government (BFI/93 and BFI/94).

References Berkman R. A., Merrill M. J., Reinhold W. C., Monacci W. T., Saxena A., Craig C. W., et al. (1993) Expression of the vascular permeability factor/vascular endothelial growth factor gene in central nervous system neoplasms. J. Clin. Invest. 91, 153–159. Brock T. A., Dvorak H. F., and Senger D. R. (1991) Tumour-secreted vascular permeability factor increases cytosolic Ca2+ and von Willebrand factor release in human endothelial cells. Am. J. Pathol. 138, 213–221. Conolly D. T., Heuvelman D. M., Nelson R., Olander J. V., Eppley B. L., Delfino J. J., et al. (1989) Tumour vascular permeability factor stimulates endothelial cell growth and angiogenesis. J. Clin. Invest. 84, 1470–1478. Dvorak H. F., Brown L. F., Detmar M., and Dvorak A. M. (1995) Vascular permeability factor/vascular endothelial growth factor, microvascular permeability and angiogenesis. Am. J. Pathol. 46, 1029–1039. Ferrara N. and Henzel W. J. (1989) Pituitary follicular cells secrete a novel heparin-binding growth

Volume 13, 1999

184 factor specific for vascular endothelial cells. Biochem. Biophys. Res. Commun. 161, 851–859. Folkman J. (1995) Angiogenesis in cancer, vascular, rheumathoid and other disease. Nature Med. 1, 27–31. Gerdes J., Lemke H., Baisch H., Wacker H. H., Schwab U., and Stein H. (1984) Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J. Immunol. 33, 1710–1715. Gerdes J., Schlueter C., and Duchrow M. (1991) Immunobiochemical and molecular biologic characterization of the cell proliferation-associated nuclear antigen that is defined by monoclonal antibody Ki-67. Am. J. Pathol. 138, 867–873. Hanks S. K., Quinn A. M., and Hunter T. (1988) The protein kinase family: conserved features and deduced phylogeny of the catalytic domains. Science 241, 42–52. Hatva E., Kaipainen A., Mentula P., Jääskeläinen J., Paetau A., Haltia M., and Alitalo K. (1995) Expression of endothelial cell-specific receptor tyrosine kinases and growth factors in human brain tumors. Am. J. Pathol. 146, 368–378. Hermansson M., Nister M., Betsholz C., Heldin C. H., Westermark B., and Funa K. (1988) Endothelial cell hyperplasia in human glioblastoma: coexpression of mRNAfor platelet-derived growth factor (PDGF) B chain and PDGF receptor suggests autocrine growth stimulation. Proc. Natl. Acad. Sci. USA 5, 7748–7752. Inaba T., Shimano H., Gotoda T., Harada K., Shimada M., Ohsuga J., et al. (1993) Expression of plateletderived growth factor beta receptor on human monocyte-derived macrophagues and effects of platelet-derived growth factor BB dimer on cellular function. J. Biol. Chem. 268, 24,353–24,360. Key G., Petersen J. I., Becker M. H., Duchrow M., Schluter C., Askaa J., et al. (1993) New antiserum against Ki67 antigen suitable for double immunostaining of paraffin wax sections. J. Clin. Pathol. 46, 1080–1084. Kleihues P., Burger P. C., and Scheithauer B. W. (1993) The new WHO classification of brain tumours. Brain Pathol. 3, 255–268. Lafuente J. V., Adán B., Alkiza K., Rossi M., and CruzSánchez F. F. (1995) Vascular proliferation and related changes in the vessel wall of brain tumors. J. Neuropathol. Exp. Neurol. 54(S), S80. Nakamura T., Takeshita I., Inamura T., Tashima T., and Fukui M. (1993) Expression of the B-chain of platelet-derived growth factor and proliferative

Journal of Molecular Neuroscience

Lafuente et al. activity of human brain tumors. Neurol. Med. Chir. Tokyo 33, 205–211. Pepper M. S., Ferrara N., Orci L., and Montesano R. (1991) Vascular endothelial growth factor (VEGF) induces plasminogen activators and plasminogen activator inhibitor-1 in microvascular endothelial cells. Biochem. Biophys. Res. Comm. 181, 902–906. Plate K. H., Breier G., Farrell C. L., and Risau W. (1992) Platelet-derived growth factor receptor-b is induced during tumor development and upregulated during tumor progression in endothelial cells in human gliomas. Lab. Invest. 67, 529–534. Plate K. H., Breier G., Weich H. A., and Risau W. (1992) Vascular endothelial growth factor is a potential tumor angiogenesis factor in human gliomas in vivo. Nature 359, 845–848. Plate K. H., Breier G., Weich H. A., Mennel H. D., and Risau W. (1994) VEGF and glioma angiogenesis: coordinate induction of VEGF receptors, distribution of VEGF protein and possible in vivo regulatory mecahisms. Int. J. Cancer 59, 520–529. Russell D. C. and Rubinstein L. J. (1989). Pathology of Tumors of the Nervous System, 5th ed. Edward Arnold, London. Schweiki D., Itin A., Soffer D., and Keshet E. (1992) Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 359, 843–845. Strugar J. G., Criscuolo G. R., Rothbart D., and Harrington W. N. (1995) Vascular endothelial growth/permeability factor expression in human glioma specimens: correlation with vasogenic brain edema and tumor-associated cysts. J. Neurosurg. 83, 682–689. Tischer E., Mitchell R., Hartman T., Silva M., Gospodarowicz D., Fiddes J. C., et al. (1991) The human gene for vascular endothelial growth factor. J. Biol. Chem. 266, 11,947–11,954. Tsai J. C., Goldman C. K., and Gillespie G. Y. (1995) Vascular endothelial growth factor in glioma cell lines: induced secretion by EGF, PDGF-BB and bFGF. J. Neurosurg. 82, 864–873. Unemori E. N., Ferrara N., Bauer E. A., and Amento E. P. (1992) Vascular endothelial growth factor induces interstitial collagenase expression in human endothelial cells. J. Cell. Physiol. 153, 557–562. Verheijen R., Kuipers H. J. H., and Schlingemann R. O. (1989) Ki-67 detects a nuclear matrix-associated

Volume 13, 1999

Expression of VEGF and PDGFR-β proliferation-related antigen. J. Cell. Sci. 92, 123–130. Weindel K., Moringlane J. R., Marme D., and Weich H. A. (1994) Detection and quantification of vascular endothelial growth factor/vascular permeability factor in brain tumor tissue and cyst fluid: the key to angiogenesis? Neurosurgery 35, 439–449. Weller M., Heimann K., and Wiedemann P. (1990) The pathogenesis of vitreoretinal proliferation and traction: a working hypothesis. Med. Hypotheses 31, 157–159.

Journal of Molecular Neuroscience

185 Wesseling P., Schlingemann R. O., Rietveld F. J. R., Link M., Burger P. C., and Ruiter D. J. (1995) Early and extensive contribution of pericytes/vascular smooth muscle cells to microvascular proliferation in glioblastoma multiforme: an immuno-light and immuno-electron microscopic study. J. Neuropathol. Exp. Neurol. 54, 304–310. Zülch K. J. (1979) Histological typing of tumours of the central nervous system, in International Histological Classification of Tumors, vol. 21. World Health Organization, Geneva.

Volume 13, 1999