Human FGF-1 gene delivery protects against quinolinate-induced striatal and hippocampal injury in neonatal rats

Fibroblast growth factors (FGFs) are cell mitogens and differentiating factors with neuroprotective properties in the CNS. We have already shown that endothelial cells genetically engineered to secrete human FGF-1 (RBEZ-FGF) survive implantation to neonatal rat brain (Johnston et al. (1996) J. Neurochem. 67, 1643-1652]. In this study, the effects of cell-based FGF-1 gene delivery on quinolinate-induced neurotoxicity in the developing rat brain were examined. Control endothelial cells (RBE4), and RBEZ-FGF cells were implanted into right striatum at post-natal day (PND) 7. On PND 10, quinolinate (150 nmol), an endogenous N-methyl-d-aspartate (NMDA) receptor agonist, or vehicle alone was injected into striatum ipsilateral to cell implantation. Injury was quantified in coronal sections obtained from PND 17 animals by comparing striatal and hippocampal volumes ipsilateral and contralateral to the site of quinolinate injection. Human FGF-1 specific transgene expression in vivo was shown by Northern blot and RT-PCR up to 14 days after cell implantation in control animals, and up to 4 days after quinolinate exposure. Quinolinate reduced the size of ipsilateral striatum by 37% and hippocampus by 38% in animals preimplanted with control endothelial cells. In contrast, quinolinate reduced the size of striatum by only 14% and had no effect on hippocampal size in animals preimplanted with RBEZ-FGF cells. Thus, FGF-1 gene delivery protected the developing striatum and hippocampus from quinolinate-induced volume loss by 62% and 100%, respectively. Intrastriatal quinolinate resulted in a significant decrease in density of NOS+ CA3 hippocampal neurons (-38%) without affecting the density of NOS+ neurons in hippocampal regions CA1, dentate gyrus or striatum. This response of CA3 NOS+ neurons appeared to be only partially reversed by FGF-1 gene delivery. Our results show that intracerebral FGF-1 gene expression within the developing brain can protect striatum and hippocampus from quinolinate-mediated injury.

Squalamine inhibits angiogenesis and solid tumor growth in vivo and perturbs embryonic vasculature

The novel aminosterol, squalamine, inhibits angiogenesis and tumor growth in multiple animal models. This effect is mediated, at least in part, by blocking mitogen-induced proliferation and migration of endothelial cells, thus preventing neovascularization of the tumor. Squalamine has no observable effect on unstimulated endothelial cells, is not directly cytotoxic to tumor cells, does not alter mitogen production by tumor cells, and has no obvious effects on the growth of newborn vertebrates. Squalamine was also found to have remarkable effects on the primitive vascular bed of the chick chorioallantoic membrane, which has striking similarities to tumor capillaries. Squalamine may thus be well suited for treatment of tumors and other diseases characterized by neovascularization in humans.

HGF/SF in angiogenesis

Hepatocyte growth factor/scatter factor (HGF/SF) is a mesenchyme-derived cytokine that stimulates motility and invasiveness of epithelial and cancer cells. These responses are transduced through the c-met proto-oncogene product, a transmembrane tyrosine kinase that functions as the HGF/SF receptor. We have shown that HGF/SF is a potent angiogenic molecule and that its angiogenic activity is mediated primarily through direct actions on vascular endothelial cells. These include stimulation of cell migration, proliferation, protease production, invasion, and organization into capillary-like tubes. We further showed that HGF/SF is overexpressed in invasive human cancers, including breast cancer, relative to non-invasive cancers and benign conditions. In invasive breast cancers, the content of HGF/SF is strongly correlated with that of von Willebrand's factor, a marker of vascular endothelial cells. Furthermore, transfection of breast cancer and glioma cell lines with HGF/SF cDNA greatly enhanced the ability of these cells to grow as tumours in orthotopic sites in syngeneic or immunocompromized host animals. The increased growth rate of the HGF/SF-transfected cells was attributable, in part, to increased tumour angiogenesis. These findings suggest that HGF/SF may function as a tumour progression factor, in part by stimulating tumour cell invasiveness and in part by stimulating angiogenesis.

Scatter factor promotes motility of human glioma and neuromicrovascular endothelial cells

Malignant gliomas are characterized by rapid growth, infiltration of normal brain tissue, and high levels of tumor-associated angiogenesis. The genetic and local environmental tissue factors responsible for the malignant progression from low to high grade gliomas and the highly malignant behavior of glioblastomas are not well understood. In a study of 77 human brain tissue extracts, high grade (III-IV) tumors had significantly greater scatter factor (SF) content than did low grade tumors or non-neoplastic tissue. To investigate the potential significance of SF accumulation in gliomas, we measured the effects of SF on DNA synthesis and motility of cultured human glioma cell lines. SF stimulated DNA synthesis in 7/10 glioma cell lines and in 3/3 neuromicrovascular endothelial cell (NMVEC) lines, consistent with our previous report that SF stimulated cell proliferation of a few human glioma cell lines. SF markedly stimulated the chemotactic migration of 10/10 glioma cell lines as well as 3/3 NMVEC lines. In addition, SF stimulated the 2-dimensional migration of glioma cells on culture surfaces coated with specific extracellular matrix molecules (collagen i.v., laminin, and fibronection). As expected based on these biologic responses to SF, 10/10 glioma lines and 4/4 NMVEC lines expressed mRNA for c-met, the SF receptor. To assess the possible in vivo significance of these migration assays, we compared the chemotactic response of a glioma cell line to human brain cyst fluids and tumor extracts that contained high or low SF concentrations. Fluids and extracts with high SF content tended to induce higher levels of chemotactic migration than did fluids and extracts with low SF content. Addition of anti-SF monoclonal antibody (MAb) inhibited migration induced by fluids and extracts with high SF content by about 30-50%.

Dexamethasone inhibits glioma-induced formation of capillary like structures in vitro and angiogenesis in vivo

Dexamethasone is used frequently in brain tumor therapy of patients. In animal models it is known to inhibit the angiogenesis of solid tumors. We addressed the question, if this is also true in brain tumors. C6 malignant glioma and 9L gliosarcoma cells were implanted into rat-brains. Dexamethasone 3 mg/kg/d intraperitoneal increased the survival compared to saline treated controls. The tumors size and the vascular density were smaller in the dexamethasone groups in both models. In vitro dexamethasone inhibited the growth of the C6 cells but not of 9L cells. Thus the growth inhibition of brain tumors in vivo appeared to be mediated partly by direct growth inhibition of tumor cells in C6 cells but additionally by antiangiogenesis in both tumor models. Several in vitro models were used to address the mechanisms of antiangiogenesis. There was no effect of dexamethasone on the proliferation of central nervous endothelial cells and no effect on the formation of capillary like structures on matrigel. Dexamethasone inhibited, however, the formation of capillary like structures in a coculture model with glioma cells in vitro. Surprisingly, progesterone had the same effect in this model. The in vitro effect was mediated via glucocorticoid receptors since receptor antagonists could inhibit it. The primary target appeared to be the tumor cell because only this cell had the complete set of receptors. These data show, that antiangiogenic therapeutic effects are possible by influencing primarily the tumor cell. This way of targeting might be of value for future developments of new strategies.

Scatter factor/hepatocyte growth factor expression enhances human glioblastoma tumorigenicity and growth

We have shown previously that the multifunctional cytokine scatter factor/hepatocyte growth factor (SF/HGF) is elevated in human malignant gliomas. In this study we investigated how human SF/HGF expression affects the malignancy of the U373 human glioblastoma cell line in vivo and in vitro. Human SF/HGF gene transfer increased U373 glioblastoma tumorigenicity by > or = 20-fold and enhanced the growth rate of intracerebral U373 xenografts by 3- to 8-fold. SF/HGF expression had no effect on the proliferation of glioblastoma cell monolayers but increased their anchorage-independent colony formation in soft agar by 5- to 8-fold. These results are the first to show that SF/HGF expression by human glioblastoma cells enhances their growth dysregulation in vitro and malignancy in vivo.

Scatter factor/hepatocyte growth factor gene transfer enhances glioma growth and angiogenesis in vivo

Scatter factor (SF), also known as hepatocyte growth factor, is angiogenic in systemic tissue, and SF titers correlate with the malignancy and metastatic phenotype of certain systemic cancers. Human gliomas express SF and its receptor c-met, but their role in the malignant progression of these tumors has not been defined. To examine this, 9L glioma cells that express c-met but not SF were transfected with human SF cDNA, and their behavior in vitro and in vivo was examined. SF gene expression was detected in conditioned medium of 9L-SF but not in control 9L-neo-transfected cell lines, by reverse transcriptase-PCR, immunoblot, ELISA, and scatter activity assays. Gliomas derived from 9L-SF and control 9L-neo cell lines implanted in the caudate/putamen of Fisher 344 rats (intracranially) and in the flanks of SCID/Beige mice (subcutaneously) were examined. Extracts from intracranial (i.c.) gliomas contained elevated levels of SF protein as determined by ELISA (1 to 5.5 ng SF/mg protein), whereas no SF was detected in control tumors. Reverse transcriptase-PCR of RNA from i.c. gliomas revealed that only 9L-SF gliomas expressed SF and both 9L-neo and 9L-SF gliomas expressed the c-met SF receptor. By postimplantation Day 14, 9L-SF i.c. gliomas were approximately 5-fold larger than 9L-neo control tumors (p < 0.001). Subcutaneous 9L-SF glioma growth was also greater than that in controls, although the differences were more variable. SF-producing i.c. gliomas contained elevated levels of 48-kd urokinase (3.5-fold) and 92-kd type IV collagenase (2.8-fold), both enzymes that correlate with the malignant progression of human gliomas (p < 0.001). SF-producing and control 9L cell lines did not differ in rates of proliferation, thymidine incorporation, or adhesion-independent growth in vitro. Conditioned medium from 9L-SF cells stimulated thymidine incorporation into microvessel brain endothelial cells 3- to 4-fold higher than did CM from 9L-neo controls (p < 0.001). Intracranial 9L-SF gliomas were more angiogenic than controls based on elevated peak (2.25-fold; p < 0.005) and mean (1.7-fold; p < 0.008) blood vessel densities. These results suggest that SF production by glioma cells enhances glioma malignancy in vivo, in part, by paracrine mechanisms involving glioma-associated angiogenesis.

Scatter factor stimulates tumor growth and tumor angiogenesis in human breast cancers in the mammary fat pads of nude mice

Scatter factor (SF) (also known as hepatocyte growth factor) is a plasminogen-related growth factor that induces tumor cell motility, invasion, and angiogenesis. Its receptor is a tyrosine kinase encoded by c-met, a protooncogene. Human breast cancer cells express SF and c-met in vivo; but human breast cancer cell lines do not produce SF in vitro. To determine whether SF can modulate the in vivo growth of human breast cancers within a natural mammary environment, we studied the orthotopic growth of SF-transfected (SF+) versus control (SF-) clones of MDAMB231 human mammary carcinoma cells in the mammary fat pads of athymic nude mice. SF+ clones expressed SF mRNA and produced very high titers of SF protein, whereas SF- clones did not express SF mRNA or produce detectable SF protein. Two SF+ clones (21 and 29) showed significantly increased tumor growth rates, reaching 3- to 4-fold larger primary tumor volumes and weights by time of killing (p < 0.001), as well as higher rates of axillary lymph node metastasis (p < 0.02), as compared with two SF- clones (32 and 34). In contrast, in vitro proliferation rates, two-dimensional colony formation, and soft agar colony formation were no greater in SF+ than in SF- clones. We performed further studies to investigate the discrepancy between the in vivo and in vitro growth results. Tumor extracts from SF+ clone (21 + 29) tumors had 50-fold higher SF content than did SF- clone (32 + 34) tumors, confirming high-level SF expression in vivo in SF+ tumors. Immunostaining of tumor sections for proliferating cell nuclear antigen revealed only a modest increase in the proportion of cycling cells in SF+ versus SF- tumors (70% versus 60%, respectively). The terminal deoxytransferase-labeling index was equally low (approximately 1%) in SF+ and SF- tumors, suggesting that apoptosis was not responsible for the slower growth of SF- tumors. However, SF+ tumors had significantly higher tumor microvessel densities than SF- tumors (p < 0.001). Moreover, there were much higher titers of chemotactic activity for microvascular endothelial cells in cell-conditioned media and primary tumor extracts from SF+ clones as compared with SF- clones. As demonstrated using the rat cornea assay, there was more angiogenic activity in SF+ tumor extracts than in SF- extracts. The increased chemotactic and angiogenic activities in SF+ tumor extracts were not explained by secondary alterations in the content of the angiogenic mediator, vascular endothelial growth factor, or the antiangiogenic glycoprotein, thrombospondin-1; and those activities were neutralized using an anti-SF monoclonal antibody. These findings suggest that SF promotes the orthotopic growth of human breast cancers, at least in part, by stimulating tumor angiogenesis.

Regulation of angiogenesis in malignant gliomas

Gliomas are highly resistant to conventional therapeutic measures, requiring the development of novel treatments. Since gliomas are particularly vascular tumors, one approach involves treatments directed at inhibiting angiogenic mechanisms. Although multiple factors contribute to the ultimate vascularization of any tumor, some are especially relevant to gliomas. Early experimental work directed at inhibiting angiogenic pathways has shown promise toward achieving control of tumor growth. This article focuses on the evidence that angiogenesis and related vascular cell responses play important roles in glioma biology, and reviews those biochemical pathways known through experimentation to be involved in the vascular response to gliomas. Finally, contemporary vessel-targeted approaches that have been used to inhibit glioma growth are discussed.

Delivery of human fibroblast growth factor-1 gene to brain by modified rat brain endothelial cells

Fibroblast growth factor (FGF) is an endothelial cell mitogen and serves as a mitogen and/or differentiating factor that can be neuroprotective for other cell types within the CNS. We established brain microvascular endothelial cell lines that secrete FGF-1 with the ultimate goal of examining their usefulness as a cellular platform for FGF gene delivery to brain. A chimeric gene consisting of the secretory sequence of FGF-4 linked at the 5' end of human FGF-1 (sp-hst/KS3:FGF-1) was transfected into rat microvascular endothelial cells previously altered to express the lacZ reporter gene (RBEZ), and numerous clones were found to secrete FGF-1 (RBEZ-FGF). Immunoblotting of conditioned medium demonstrated an 18-kDa protein corresponding to FGF-1. Conditioned medium from RBEZ-FGF cells enhanced [3H]thymidine incorporation in BALB/c3T3 fibroblasts by up to sevenfold when compared with conditioned medium of control cell lines, corresponding to as much as 110 ng of active FGF-1/mg of cell protein/24 h. RBEZ-FGF cell lines remained contact-inhibited and proliferated independent of exogenous endothelial mitogens, in contrast to control lines that are mitogen-dependent. Incubation of PC12 cells with RBEZ-FGF cells or their conditioned medium induced neurite outgrowth by PC12 cells. RBEZ-FGF cells survived following implantation to neonatal and adult rat caudate-putamen for at least 21 days based on 5-bromo-4-chloro-3-indolyl beta-D-galactopyranoside (X-gal) histochemistry, and FGF-1 gene expression by these cells in vivo was demonstrated by in situ hybridization and reverse transcriptase-PCR. These findings suggest that endothelial cells may be useful for FGF gene delivery to the CNS.

Endothelial cell-based cytokine gene delivery inhibits 9L glioma growth in vivo

Malignant brain neoplasms present great therapeutic challenges due to their extremely aggressive behavior and relative isolation by the blood-brain and blood-tumor barriers. Endothelial cells may be versatile platforms for delivering genes to solid tumors by virtue of their location at blood-tissue interfaces and their proliferation in response to endothelial mitogens produced by tumors. Immortalized rat brain endothelial cells that express the E. coli lacZ reporter gene and the gene for murine interleukin-2 (RBEZ-IL2) were co-inoculated with 9L glioma cells to Fisher rats to examine the effects of endothelial cell-based cytokine delivery on glioma growth in vivo. 9L glioma growth was not affected by the implantation of control RBEZ cells. The growth of subcutaneous and intracranial 9L gliomas was significantly inhibited by RBEZ-IL2 cells (P < 0.005 and P < 0.01, respectively) when compared to control transfected RBEZ cells. Rats receiving intracranial 9L glioma cells with RBEZ-IL2 cells showed increased survival (P < 0.001). Histologic and immunohistologic analysis showed enhanced activation of microglia/macrophages and CD8-positive T lymphocytes and/or natural killer cells within brain at sites of 9L inoculation with RBEZ-IL2 cells. This report establishes that immortalized endothelial cells can be used for cytokine gene delivery and to activate anti-tumor host responses to experimental gliomas within the central nervous system.

Scatter factor expression and regulation in human glial tumors

Scatter factor (SF) (also known as hepatocyte growth factor [HGF]) is a cytokine that induces cell motility in vitro and angiogenesis in vivo. SF appears to be a determinant of the malignant phenotype in certain systemic cancers. We detected SF in extracts prepared from human gliomas, with the highest levels found in malignant tumors. Human glioblastoma cells expressed both SF and its receptor (c-met protein) in vivo, as demonstrated by immunohistochemistry. Consistent with these observations, we found moderate to high levels of production of immunoreactive and biologically active SF by cultured human glioblastoma cells (3 of 8 lines) and by neural microvascular endothelial cells (NMVEC) (3 of 3 lines). SF stimulated the proliferation of glioblastoma and NMVEC cell lines by paracrine or autocrine mechanisms. Conditioned medium (CM) from both glioblastoma and NMVEC cells contained SF-inducing factor (SF-IF) activity, defined by its ability to stimulate SF production in an indicator cell line (MRC5 human fibroblasts). This activity consisted of a high-molecular-weight (> 30 kDa), heat-sensitive component and a low-molecular weight (< 30 kDa), heat-stable component. Furthermore, glioblastoma CM stimulated NMVEC SF production, and NMVEC CM stimulated glioblastoma cell SF production, by 3- to 6-fold in each case. Our findings demonstrate that SF-dependent interactions between glioma cells, and between glioma cells and endothelium, can contribute to the heterogeneous proliferative and angiogenic phenotypes of malignant gliomas in vivo.

Modulation of serine proteinases and metalloproteinases during morphogenic glial-endothelial interactions

The regulation of microvessel formation and the expression of CNS-specific endothelial properties are attributed to perivascular astroglia. Specific proteolytic pathways mediate processes such as tissue remodeling, differentiation, invasion, and metastasis. We used a co-culture system in which C6 glial cells induce CNS microvascular endothelial cells to form capillary-like structures to examine the role of plasminogen activators and collagenases in CNS microvessel morphogenesis. Fibrin zymography revealed the presence of high-molecular weight urokinase-type plasminogen activator (uPA), low-molecular weight uPA, and uPA/inhibitor complexes within endothelial cultures and cocultures. Gelatin zymography revealed the presence of 92-, 72-, and 62-kDa type IV collagenases within endothelial cultures and cocultures. uPA activity was confirmed by incubating the extracts with amiloride, an inhibitor of uPA. Collagenase activity was confirmed by incubating the gels with EDTA, an inhibitor of metalloproteinases. Quantitative densitometry showed a six- to eightfold decrease in coculture uPA during capillary-like structure formation. Substantially less change in type IV 72-kDa procollagenase activity was seen in cocultures during capillary-like structure formation, but active type IV 62-kDa collagenase activity was significantly increased during capillary-like structure formation. These findings establish that uPA and activated type IV collagenase activity specifically regulates morphogenic endothelial responses to glial interactions and suggest mechanisms by which microvessels respond within the CNS.

Neurovirulent simian immunodeficiency virus replicates productively in endothelial cells of the central nervous system in vivo and in vitro

The perivascular location of human immunodeficiency virus-infected cells suggests that the virus enters the central nervous system (CNS) by traversing the blood-brain barrier (BBB). In this study, the simian immunodeficiency virus (SIV) macaque model was used to determine whether SIV infects CNS endothelial cells. SIV RNA was detected in capillary endothelial cells in brain sections from animals parenterally inoculated with a neurovirulent strain of SIV by double immunohistochemistry and in situ hybridization and by reverse transcriptase-in situ PCR. These in vivo observations were extended by examining whether SIV replicated productively in cultured macaque brain endothelial cells (MBEC). A neurovirulent strain, SIVmac239/17E-Br, replicated productively in MBEC as determined by the presence of viral cytopathic effect (syncytia), viral DNA by PCR, viral RNA by in situ hybridization, and viral antigen by immunohistochemistry and by the production of high titers of cell-free virus. Virus replication was confirmed by electron microscopy. In contrast, a nonneurovirulent strain, SIVmac239, did not infect MBEC. Infection of the endothelial cells was not blocked by soluble CD4. Thus, endothelial cells may provide a CD4-independent pathway of virus entry to the CNS. In addition, damage to the BBB as a result of endothelial cell infection may provide a mechanism for amplification of viral load in the CNS and may contribute to the CNS dysfunction that characterizes AIDS dementia and SIV encephalitis. These data suggest that MBEC may serve a selective role in determining virus entry to the CNS.

Endothelial cell implantation and survival within experimental gliomas

The delivery of therapeutic genes to primary brain neoplasms opens new opportunities for treating these frequently fatal tumors. Efficient gene delivery to tissues remains an important obstacle to therapy, and this problem has unique characteristics in brain tumors due to the blood-brain and blood-tumor barriers. The presence of endothelial mitogens and vessel proliferation within solid tumors suggests that genetically modified endothelial cells might efficiently transplant to brain tumors. Rat brain endothelial cells immortalized with the adenovirus E1A gene and further modified to express the beta-galactosidase reporter were examined for their ability to survive implantation to experimental rat gliomas. Rats received 9L, F98, or C6 glioma cells in combination with endothelial cells intracranially to caudate/putamen or subcutaneously to flank. Implanted endothelial cells were identified by beta-galactosidase histochemistry or by polymerase chain reaction in all tumors up to 35 days postimplantation, the latest time examined. Implanted endothelial cells appeared to cooperate in tumor vessel formation and expressed the brain-specific endothelial glucose transporter type 1 as identified by immunohistochemistry. The proliferation of implanted endothelial cells was supported by their increased number within tumors between postimplantation days 14 and 21 (P = 0.015) and by their expression of the proliferation antigen Ki67. These findings establish that genetically modified endothelial cells can be stably engrafted to growing gliomas and suggest that endothelial cell implantation may provide a means of delivering therapeutic genes to brain neoplasms and other solid tumors. In addition, endothelial implantation to brain may be useful for defining mechanisms of brain-specific endothelial differentiation.

Development of endogenous beta-galactosidase and autofluorescence in rat brain microvessels: implications for cell tracking and gene transfer studies

Cell transplantation is commonly used in studies of CNS development, tumor biology, and gene therapy. Fluorescent dyes and the E. coli lacZ reporter gene are used to identify transplanted cells in host tissues. The usefulness of these methods depends on host autofluorescence and beta-galactosidase (beta-Gal) activity. Our interest in the CNS vasculature led us to examine vascular autofluorescence and beta-Gal activity in postnatal and adult rat brains. Brains were perfusion-fixed (3.7% paraformaldehyde), cryoprotected, and cryostat-sectioned (12 microns). Autofluorescent vessel profiles were quantitated in sections using rhodamine filter sets and beta-Gal-positive vessels were quantitated under bright-field after incubation of sections with X-Gal chromogenic substrate for 1-18 hr at 37 degrees C. Multifocal vessel autofluorescence appeared in postnatal Day (PND) 18 Lewis rats (0.6 +/- 0.4 vessels/field) and increased tenfold in adults (6.8 +/- 0.3/field). The numbers of beta-Gal-positive vessels in PND 18 and adult sections incubated with X-Gal for 18 hr were 21.1 +/- 1.7 and 119 +/- 17, respectively. Host beta-Gal staining was similar to that produced by implanted endothelial cells expressing the bacterial lacZ reporter gene. Reducing incubation times in X-Gal to less than 4 hr eliminated endogenous staining and retained lacZ-specific staining. The presence of vascular autofluorescence and endogenous beta-Gal activity must be considered when either fluorescence- or lacZ-dependent cell markers are used in rat brain.

Endothelial differentiation in intracerebral and subcutaneous experimental gliomas

Blood-brain barrier (BBB) properties of endothelial cells have on impact on brain tumor behavior, diagnosis, and response to therapy. Biochemical BBB properties are expressed by endothelial cells within intracerebral (IC) gliomas but little is known regarding the expression of BBB-associated proteins within gliomas established subcutaneously (SC), a site that is frequently used in experimental glioma models. We compared the expression of two BBB proteins, glucose transporter type-1 (Glut1) and endothelial barrier antigen (EBA), in IC and SC rat 9L and F98 gliomas. The percentage of microvessels with immunohistochemically-detectable Glut1 and EBA in IC 9L tumors (31-98%) contrasted with that found in SC 9L tumors (0-3.9%) (P < 0.0001). Likewise, the percentage of immunohistochemically-positive vessels in IC F98 tumors (35-66%) differed markedly from that in SC F98 tumors (0%) (P < 0.0001). These differences were not explained by effects of tumor location on vessel density or tumor histology. These findings demonstrate that the peritumoral environment influences endothelial differentiation within glial tumors and suggest that glioma cells maintain but do not induce the expression of barrier properties in vessels that infiltrate tumor from surrounding tissue.

Regulation of in vitro glia-induced microvessel morphogenesis by urokinase

Plasminogen activators (PAs) regulate a variety of processes involved in tissue morphogenesis and differentiation. We used a coculture system in which microvascular endothelial cells are induced by glial cells to form capillary-like structures in order to examine the role of urokinase-type PA (uPA) during microvessel morphogenesis within the central nervous system (CNS). Endothelia-derived uPA activity decreased sevenfold within glial-endothelial cocultures when capillary-like structures were formed. Incubation of cocultures with concentrations of phorbol 12-myristate 13-acetate (0.1 and 1.0 nM) that induced endothelial uPA activity (45-210%) inhibited endothelial differentiation (25-70%). Furthermore, incubation of cocultures with proteolytically active low molecular weight uPA (5-500 IU/ml) inhibited endothelial differentiation (37-75%), whereas the amino terminal cell-binding fragment of uPA had minimal effect. Inhibition of plasminogen activation in cocultures with the serine protease/plasmin inhibitors aprotinin and soybean trypsin inhibitor increased glia-induced capillary-like structure formation (96-98%). These findings establish a paracrine/autocrine function for urokinase and its inhibitors in regulating endothelial responses to perivascular glia and provide insight into mechanisms of microvascular reactions to CNS pathology.

Dexamethasone reduces vascular density and plasminogen activator activity in 9L rat brain tumors

Angiogenesis, a process dependent upon perivascular proteolysis, is required for solid tumor growth and is inhibited by certain steroids including glucocorticoids. We examined the relationship between tumor growth and vessel density in experimental rat brain 9L glial tumors following chronic treatment with the glucocorticoid dexamethasone. Tumor growth was inhibited by intraperitoneal administration of 3 mg/kg/day dexamethasone. Maximal cross-sectional areas of post-implantation day 9 tumors were 4.6 +/- 1.0 mm2 in dexamethasone-treated animals and 17.0 +/- 3.4 mm2 in controls (P < 0.01). Microvessel density assessed by laminin immunohistochemistry was 59% lower in dexamethasone-treated tumors (P < 0.01). Plasminogen activator (PA) activity, a proteolytic enzyme related to endothelial migration and vessel growth, was 4.2 +/- 0.9 IU/micrograms protein in dexamethasone-treated tumors and 9.0 +/- 1.0 IU/micrograms protein in control tumors (P < 0.01). Exposure of cultured 9L and central nervous system microvessel endothelial cells to dexamethasone concentrations comparable to those achieved in vivo had no effect on cell growth, but reduced the PA activity of culture supernatant fractions by 78% and 99%, respectively. These findings suggest that inhibition of proteolytic steps involved in vessel growth may underlie, in part, the mechanism by which glucocorticoids decrease brain tumor growth.

Inhibition of astroglia-induced endothelial differentiation by inorganic lead: a role for protein kinase C

Microvascular endothelial function in developing brain is particularly sensitive to lead toxicity, and it has been hypothesized that this results from the modulation of protein kinase C (PKC) by lead. We examined the effects of inorganic lead on an in vitro model of central nervous system endothelial differentiation in which astroglial cells induce central nervous system endothelial cells to form capillary-like structures. Capillary-like structure formation within C6 astroglial-endothelial cocultures was inhibited by lead acetate with 50% maximal inhibition at 0.5 microM total lead. Inhibition was independent of effects on cell viability or growth. Under conditions that inhibited capillary-like structure formation, we found that lead increased membrane-associated PKC in both C6 astroglial and endothelial cells. Prolonged exposure of C6 cells to 5 microM lead for up to 16 h resulted in a time-dependent increase in membranous PKC as determined by immunoblot analysis. Membranous PKC increased after 5-h exposures to as little as 50 nM lead and was maximal at approximately 1 microM. Phorbol esters were used to determine whether PKC modulation was causally related to the inhibition of endothelial differentiation by lead. Phorbol 12-myristate 13-acetate (10 nM) inhibited capillary-like structure formation by 65 +/- 5%, whereas 4 alpha-phorbol 12,13-didecanoate was without effect. These findings suggest that inorganic lead induces cerebral microvessel dysfunction by interfering with PKC modulation in microvascular endothelial or perivascular astroglial cells.

Selective endothelial growth inhibition by tetracyclines that inhibit collagenase

The potential of angiogenesis inhibitors as therapy for human diseases is limited by a lack of clinically available agents. We investigated the mechanism of the anti-angiogenesis effects of minocycline, a commonly used drug, and several derivatives. Endothelial cell proliferation was inhibited by several of these compounds. We found that inhibition was associated with inhibition of collagenase, did not require antibiotic activity, and was not related to cytotoxicity. Other microvessel-associated cells were unaffected. This endothelial antiproliferative effect is a potential mechanism of the anti-angiogenic activity of minocycline.

Vascular differentiation and glucose transporter expression in rat gliomas: effects of steroids

The GLUT1 isoform of the glucose transporter is normally expressed at high levels in differentiated brain vessels that also express a permeability barrier. In contrast, malignant brain neoplasms have relatively undifferentiated vessels that are highly permeable, proliferate to high vascular densities, and often lose GLUT1 expression. Using the rat intracerebral 9L glioma model, we investigated whether dexamethasone-induced changes in permeability are associated with the appearance of other differentiated vascular properties. The percentage of vessels expressing immunohistochemically detectable GLUT1 (74.2 +/- 6.1%) and the tumor vessel density as assessed by laminin immunostaining (282 +/- 37 vessels/mm2) did not vary with control tumor size. Dexamethasone treatment resulted in an 83% reduction of vascular permeability to intravenous Evans blue, an increased percentage of vessels expressing GLUT1 (106.4 +/- 10.5%), lower vascular density (102 +/- 64 vessels/mm2), and smaller tumor size (control cross-sectional area, 17.0 +/- 3.4 mm2; treated, 4.6 +/- 1.0 mm2). Essentially all vessels became GLUT1-positive after dexamethasone treatment. Increased GLUT1 expression by glioma vessels in association with the appearance of other signs of differentiation (low vascular density, slow tumor growth) suggests that immunostaining for GLUT1 may identify neoplasms that are biologically less aggressive.

Steroid Inhibition of Neural Micro vessel Morphogenesis In Vitro: Receptor Mediation and Astroglial Dependence

Steroid hormones alter several aspects of microvascular function within the CNS. Both microvessel formation and blood-brain barrier expression appear to be influenced by interactions between astrocytes and endothelial cells. To determine if steroids alter astrocyte-endothelial interactions, we studied their effects on astroglial-induced microvessel morphogenesis in vitro. C6 astroglial cells induce bovine retinal microvascular endothelial cells to differentiate into capillary-like structures. Dexamethasone, hydrocortisone, and progesterone at 10 nM inhibited C6-induced microvessel morphogenesis by 75, 35, and 30%, respectively. Inhibition by dexamethasone was both time and concentration dependent, reaching 80-100% at 1 microM. Tetrahydrocortisone and 17 alpha-hydroxyprogesterone had only marginal inhibitory effects. Cortexolone, a glucocorticoid receptor antagonist, blocked inhibition by dexamethasone. Progesterone receptors were expressed in C6 but not bovine retinal microvascular endothelial cells, identifying the astroglial cell as the likely effector of progesterone-mediated inhibition. Astroglial cells were further implicated as the effectors of steroid-mediated inhibition because none of the steroids inhibited astroglial-independent capillary-like structure formation in response to a reconstituted extracellular matrix, Matrigel. These findings are evidence that steroids modulate neural microvascular endothelial cell functions indirectly through perivascular astrocytes via a receptor-mediated mechanism.

Vascular expression of glucose transporter in experimental brain neoplasms

Vascular abnormalities in brain neoplasms are important to tumor biology and therapy. Glucose transporter (GLUT1) expression is a differentiated property of normal cerebral microvessels typically associated with expression of the blood-brain barrier. We investigated the relationship of GLUT1 expression to other vascular characteristics in F98, 9L, and C6 gliomas and Walker 256 carcinomas implanted into adult rat brains. The percentages of microvessels with immunohistochemically detectable GLUT1 were 95.5 +/- 3.9 in F98, 60.9 +/- 3.9 in 9L, 45.4 +/- 5.6 in C6, and 1.2 +/- 0.3 in Walker 256 (mean +/- SEM). The percentage of GLUT1-positive vessels in F98 was not statistically different from that in normal brain. GLUT1 expression was not dependent on restricted permeability as all tumors were highly permeable to Evans blue. GLUT1 expression was unrelated to vascular density, vascular morphology, and parenchymal GFAP expression. The expression of GLUT1, a marker of cerebral endothelial differentiation, is a newly described property of glial tumor vessels that may have diagnostic and prognostic significance.