Ras Activation in Astrocytomas and Neurofibromas

Abhijit Guha

Abstract: Oncogenic mutations resulting in activated Ras. Guanosine Triphosphate (GTP) are prevalent in 30% of all human cancers, but not primary nervous system tumors. Several growth factors/receptors are implicated in the pathogenesis of malignant astrocytomas including epidermal growth factor (EGFR) and platelet derived growth factor (PDGF-R) receptors, plus the highly potent and specific angiogenic vascular endothelial growth factor (VEGF). A significant proportion of these tumors also express a truncated EGFR, which is constitutively activated. Our work demonstrates that the mitogenic signals from both the normal PDGF-R and EGFR and the truncated EGFR activate Ras. Inhibition of Ras by genetic or pharmacological strategies leads to decreased astrocytoma tumorgenic growth in vitro and decreased expression of VEGF. This suggests that these agents may be potentially important as novel anti-proliferative and anti-angiogenic therapies for human malignant astrocytomas.

In contrast to astrocytomas, where increased levels of activated Ras.GTP results from transmitted signals from activated growth factor receptors, the loss of neurofibromin is postulated to lead to functional up-regulation of the Ras pathway in neurofibromatosis-1(NF-1). We have demonstrated that NF-1 neurofibromas and neurogenic sarcomas, compared to non-NF-1 Schwannomas, have markedly elevated levels of activated Ras.GTP. Increased Ras.GTP was associated with increased tumor vascularity in the NF-1 neurogenic sarcomas, perhaps related to increased VEGF secretion. The role of Ras inhibitors as potential therapy in this tumor is also under study.

Résumé: Activation de Ras dans les astrocytomes et les neurofibromes (médaille du Collège Royal en Chirurgie, 1997). Les mutations d'oncogènes provoquant une activation de Ras-GTP ont une prévalence de 30% dans tous les cancers humains, mais non dans les tumeurs primitives du système nerveux. Plusieurs facteurs de croissance / récepteurs sont impliqués dans la pathogenèse des astrocytomes malins dont le récepteur du facteur de croissance épidermique (EGF-R) et celui du facteur de croissance plaquettaire (PDGF-R), ainsi que le facteur de croissance endothélial vasculaire (VEGF), un facteur de croissance très puissant et hautement spécifique. Une grande proportion des astrocytomes malins expriment un EGF-R tronqué qui est activé constitutivement. Nos travaux démontrent que les signaux mitogènes du PDGF-R normal et du EGFR normal et du EGFR tronqué activent Ras. L'inhibition de Ras par des stratégies génétiques ou pharmacologiques provoque la diminution de la croissance tumorale astrocytaire in vitro et diminue l'expression du VEGF. Ceci indique que ces agents pourraient être importants comme traitements antiprolifératifs et antiangiogéniques dans les astrocytomes malins humains. Contrairement aux astrocytomes où des niveaux augmentés de Ras.GTP activé résultent de signaux transmis provenant de récepteurs de facteurs de croissance activés, on pense que la perte de la neurofibromine amène une régulation fonctionnelle à la hausse de la voie Ras dans la neurofibromatose I (NFI). Nous avons démontré que les neurofibromes de la NFI et les sarcomes neurogéniques non-NFI ont des niveaux très élevés de Ras.GTP activé comparés aux Schwannomes. Un Ras.GTP augmenté était associé à une vascularité tumorale augmentée dans les sarcomes neurogéniques de la NFI, possiblement en relation avec une augmentation de la sécrétion du VEGF. Le rôle des inhibiteurs de Ras en tant que thérapie dans cette tumeur est aussi à l'étude présentement.

---

The Clinical Problem

Astrocytomas: Astrocytomas are the most prevalent primary intracranial neoplasm, accounting for about 4-5% of all cancer related deaths (approx. 20,000 deaths/year in North America).¹ Astrocytomas are classified by the World Health Organization^2,3 into four increasing grades of malignancy; Grade 1: pilocytic astrocytomas found in children which usually have an indolent course, Grade 2: astrocytoma (low grade), Grade 3: anaplastic astrocytoma (AA) and Grade 4: glioblastoma multiforme (GBM). This histo-pathological classification is somewhat arbritary, as low grade adult astrocytomas (Grade 1, 2) will inevitably progress to a more malignant grade (Grade 3, 4). At the present time, there is no antigenic or molecular marker which distinguishes an astrocytoma cell from a normal astrocyte or between a low and higher grade tumor cell. This often leads to misdiagnosis especially when small specimens are obtained, such as by stereotactic biopsies. Acknowledging the potential difficulties in classification, the W.H.O. grading is nevertheless an important survival prognosticator.

Anaplastic astrocytomas and GBMs are by far the most common type of astrocytomas seen in adults, with fifty percent of the patients dying at 18 (AA) and 12 (GBM) months post-diagnosis.³ This morbid statistic includes patients who have received radiation therapy, which shifts the survival curve favorably only by a few months. Aggressive gross neurosurgical removal, variations in dose and delivery of radiotherapy, chemotherapy, and other adjuvant therapy has had little impact on survival over the last thirty years in management of malignant astrocytomas.

Neurofibromatosis-1: Neurofibromatosis-1(NF-1) and 2, collectively known as von-Recklinghausen's Syndrome⁴ due to their partial clinical overlap, are two distinct diseases based on loss of function of two different genes. The hallmark of NF-1 is the presence of multiple benign peripheral (cutaneous) neurofibromas consisting of a mixture of Schwann cells, fibroblasts, and mast cells.⁵ Other clinical features include café au lait (CAL) spots, freckling of the axilla, groin and other intertriginous areas and pigmented hamartomas of melanocytic origin in the iris known as Lisch nodules. While these are the most commonly noted signs in NF-1, the disease is characterized by a variable number of diverse pathologies throughout the body, including cutaneous, osseous, hematological, developmental, and nervous system abnormalities.⁶ Despite recent advances in our understanding of the molecular basis of NF-1, these clinical criteria continue to be the most reliable means for making the diagnosis.⁷

The two major life-threatening complications of NF-1 are hypertension and a significantly higher rate of malignancy.^8,9 In a 12 year Swedish study, 22 of 70 NF-1 adults being followed died (10-hypertension and 12-malignancies), a rate four times that of the general population. Malignancies noted at a higher frequency in NF-1 patients include pheochromocytomas, astrocytomas (most notably optic gliomas), chronic myeloid leukemias of childhood, and malignant peripheral nerve sheath tumors which arise following malignant transformation of non-cutaneous, more deeply-located plexiform neurofibromas.^10,11 Despite this, the diagnosis of NF-1 is usually not life threatening, with the majority of the patients surviving well into their adulthood, with a mean survival age of 61.2 years in the Swedish study.⁸

Review of Current Molecular Understanding:

Astrocytomas: As schematized in Figure 1, a molecular pathogenic description of the progression of astrocytomas is slowly evolving and involves both tumor suppressor genes (TSG) and oncogenes. Most patients present with a malignant astrocytoma (AA, GBM), suggesting that a normal astrocyte can be directly transformed to a malignant astrocyte. Another plausible explanation is that in these patients the lower grades of astrocytomas were not detected as they were subclinical. Karyotypic analysis of malignant astrocytomas demonstrate that most tumor cells are diploid (2N), although they have gross chromosomal abnormalities including total or partial deletions, duplications, translocations and amplifications.¹² Mutations have been identified in the p53 tumor suppressor gene on chromosome 17p in low grade astrocytomas, suggesting that this may be an early pathogenic event.¹³ Furthermore, there is clonal expansion of the p53 mutated cells (perhaps due to selective growth advantage) as the tumor progress to a more malignant grade.¹⁴ Loss of entire or parts of chromosome 10 is exclusively found in GBMs^15-17 and not in lower grade astrocytomas. Recently a dual specific phosphatase, termed PTEN/MMAC has been identified as a TSG on chromosome 10 associated with GBM's and several other tumors including breast and prostate cancers.^18-20 Other yet to be identified TSG(s) on chromosome 10 lost in GBMs are probably present. The multiple tumor suppressor (MTS1) gene, encoding for the cell cycle inhibitor p16, may also be relevant in progression to a malignant astrocytoma.^21,22 Loss of heterozygosity on 19q may be related to progression of astrocytomas to a higher grade.²³

Epidermal Growth Factor Receptor (EGFR): The proto-oncogene c-erbB which encodes for the EGFR is located on chromosome 7 and amplified in about 50% of GBMs and found at much lower levels in low grade astrocytomas.^24,25 Established human astrocytoma cultures do not retain the DNA amplification, though they still may overexpress EGFR at the protein level. In those GBMs with amplified EGFR, 25-40% express both normal (170kDa) and a truncated EGFR (140^EGFR) which is unable to bind EGF or TGF-a, due to a 801 base pair deletion (exons 2-7; residue 6-273) in its N-terminal extracellular domain.^26-28 There is recent evidence that the 140^EGFR is constitutively activated and enhances in-vivo growth of transfected malignant astrocytoma cells in animal models.²⁹ The signal transduction mechanism(s) that the 140^EGFR utilizes are unknown but may include activation of the Ras pathway.

Platelet Derived Growth Factor Receptor (PDGF-R): The majority of established malignant astrocytoma cell lines overexpress a combination of PDGF ligand and receptor genes which could, in principle, form an autocrine/paracrine loop.³⁰ Amplification or rearrangements of the PDGF subunits (A, B) or PDGF-R (a,b) genes are not found in the majority of astrocytoma specimens.^31,32 However, overexpression of the PDGF-aR as determined by western analysis was demonstrated in 24% of the GBMs examined. Expression of both PDGF-A & B subunits are increased in malignant astrocytomas (AA, GBMs), compared to lower grade astrocytomas. PDGF-aR was overexpressed in all astrocytoma grades compared to non-neoplastic glia.^33,34 This suggests that PDGF-aR expression, which is capable of binding both PDGF subunits, may be an early event in transformation of a normal astrocyte to an astrocytoma cell, and that subsequent progression to a malignant astrocytoma cell may depend on increased expression of PDGF ligands and stimulation of the astrocytoma cell by autocrine/paracrine mechanisms. In another study, we documented that the increased expression of PDGF and PDGF-aR in sporadic malignant astrocytomas, were also prevalent in malignant astrocytomas associated with patients with germline p53 mutations or the Li-Fraumeni syndrome.³⁵ Expression of PDGF and PDGF-Rs is postulated to be functionally relevant in the growth of astrocytomas. To test this hypothesis we created PDGF dominant-negative mutants by site-directed mutagenesis of the mouse PDGF-A cDNA.^36,37 We found that growth of human established astrocytoma cells was decreased both in-vitro and in-vivo, by expression of the PDGF dominant-negative mutants blocking activation of the overexpressed PDGF-Rs.

Vascular Endothelial Growth Factor (VEGF): Angiogenesis is dynamically regulated with both positive and negative endogenous factors, with VEGF being the most potent and specific endothelial cell mitogen.^38-45 VEGF was initially isolated as a 40-46kDa protein from tumor fluid, which was 10,000-50,000 times as potent as histamine in increasing vascular permeability.^46-48 Subsequently, a highly potent and specific endothelial mitogen was identified from bovine pituitary follicular stellate cells,⁴⁹ which was found to be identical to the previously isolated permeability factor.⁵⁰ VEGF is a dimeric growth factor and is highly secreted due to a N-terminal signal peptide.50 Four isoforms of VEGF (121, 165, 189 and 205 amino acids) have been identified as a result of alternate splicing, with VEGF165 being predominant.^51,52 Hypoxia is a strong transcriptional stimulant of VEGF, as are other mitogenic growth factors.^53-57 Recent evidence suggests that both mitogenic signals and angiogenic signals (via induction of VEGF) share a common link by activation of the Ras signaling pathway,^58,59 and hence inhibition of Ras activity may lead to control of both tumor cell and tumor angiogenic growth.

The two primary biological functions of VEGF (endothelial cell mitogen and endothelial cell permeability) are mediated through two high affinity protein receptor tyrosine kinases, Flt-1 and Flk-1 (human counterpart of the latter known as KDR).^60-65 Expression of these receptors only on vascular endothelial cells accounts for the main biological function of VEGF, as it can induce the entire sequence of angiogenesis,^49,50,66 and makes VEGF the most potent of all known angiogenic factors, in contradistinction to other non-specific angiogenic growth factors. In addition to acting as a potent endothelial cell mitogen, VEGF also induces the endothelial cells to express various factors such as proteases, collagenases, urokinase and tissue plasminogen activators, which are all involved in turning on the "angiogenic switch" thereby promoting angiogenesis and metastasis.^67,68

In human astrocytomas, one of the main pathological criteria for grading the degree of malignancy and hence the prognosis is tumor vascularity.^2,3,69 Whether tumor angiogenesis is similarly correlated to malignant potential of peripheral nerve tumors in both NF-1 and non NF-1 patients is presently unknown, although some evidence exists that plexiform neurofibromas and neurogenic sarcomas can be highly angiogenic.^70,71 In contrast to low grade astrocytomas, malignant astrocytoma cells express increased levels of VEGF, especially around the hypoxic peri-necrotic zones, with increased expression of Flk-1/KDR in the hyperproliferative vascular endothelium.^50,54,72-75 In addition to astrocytomas, we have demonstrated a correlationship between tumor vascularity, peri-tumoral edema and VEGF expression in meningiomas,⁷⁶ the second most common adult CNS tumor. Anti-angiogenic therapeutic strategies including VEGF neutralizing antibodies,^74,77-79 antisense constructs,⁸⁰ inhibitory mutants against the ligand or receptor (Flk-1/KDR dominant negative mutant)⁸¹ have demonstrated some efficacy in animal models of astrocytomas. Pharmacologic therapies such as tyrophostins and other small molecules directed at blocking Flk1/KDR receptor activation⁸² are under current study. In the future signaling pathways such as Ras activation which up-regulate VEGF expression, and those pathways involved in transmitting signals from activated VEGF receptors in endothelial cells may be of therapeutic value.

Neurofibromatosis-1: The NF1 gene, located on the pericentromeric region of chromosome 17, is extremely large,⁸³ with a variety of mutations identified in NF-1 patients.^84,85 However, no specific mutational hotspots or significant genotype-phenotype correlationship, where one can predict the clinical presentation based on the location and type of mutation, have been identified.⁸⁴ The lack of mutational hot spots, the large size of the NF1 gene, high spontaneous mutation rate and lack of genotype-phenotype correlationship, have all contributed to the hurdles which still make routine genetic screening of NF-1 patients impractical. The NIH clinical diagnostic criteria,⁸⁶ remains the best method of detecting new patients with NF-1. Homology screening of neurofibromin, the protein product of the NF1 gene, offered clues as to its function. A small region in the central portion of neurofibromin demonstrated 30% homology with the mammalian p120-GAP, and with the Saccharomyces cerevisiae genes Ira1 and Ira2 (inhibitory regulators of the Ras-cAMP pathway in yeast).^87,88 This region has been named the GAP-Related Domain (GRD), and has identified neurofibromin as a member of the Ras-GAP family, of which there are currently four mammalian members, all of which are the key negative regulators of the signal transduction protein Ras.^10,88-94 In NF-1 peripheral nerve tumors it is hypothesized that decreased levels of neurofibromin leads to increased Ras.GTP, with subsequent aberrant mitogenic signals leading to tumor formation. In support of this hypothesis, neurofibrosarcoma cell lines established from NF-1 patients not only lacked neurofibromin expression but have elevated levels of activated Ras.GTP.^95,96 Furthermore, the elevated levels of Ras.GTP directly contributed to mitogenesis, since cellular proliferation could be effectively blocked by microinjecting neutralizing Ras antibody into these cells,^95,96 or by inhibiting Ras activation by farnesyl transferase inhibitors (FTI).⁹⁷ We asked whether aberrant activation of the Ras signaling pathway was also present in peripheral nerve tumor specimens, by adapting an enzymatic assay which allows for quantitative determination of Ras.GTP and Ras.GDP in tissues.^10,98

Ras-mediated Signal Transduction:

The three human Ras genes which code for four 21kDa proteins (Ha,N,K4A,K4B) belong to the small G protein family, and comprise important intracellular signal transduction molecules. Oncogenic activating mutations in Ras (residues 12,13,61), prevent conversion of activated GTP bound Ras to the basal inactive GDP bound Ras by GTP'ase Activating Proteins (i.e., _p120GAP & neurofibromin). The importance of Ras in cellular control is exemplified by the fact that 30% of human cancers, but not those primarily from the nervous system, have oncogenic activating mutations.^92,99,100 Activation of Ras is pivotal in transmitting proliferative, differentiating or transforming signals from a variety of activated growth factor receptor tyrosine kinases (RTKs) such as EGFR and PDGF-R.^101-105 These responses are blocked by micro-injection of neutralizing Ras antibody (Y13-259) into the growth factor stimulated cells, or by over expression of the Ha-Ras-Asn17 dominant inhibitory mutant. This Ha-Ras-Asn17 mutant is unable to effectively bind Mg+2, thereby decreasing its affinity to both GDP and GTP but increasing its affinity towards nucleotide exchange factors (see below), which are required for the activation of Ras. By depleting the availability of these nucleotide exchange factors, normal endogenous Ras cannot be activated to Ras.GTP effectively blocking this signaling pathway, resulting in the dominant inhibitory function of the mutant.^106-112

Activation by receptors: The discovery of protein modules, such as SH2 (src homology-2), SH3 (src homology-3), PTB (phosphotyrosine binding domain) and PH (pleckstrin homology) domains, have led to elucidation of many of the protein-protein interactions involved in signal transduction, including the upstream regulators and downstream substrates of the Ras pathway, Figure 2.^113-118 Ras activation requires post-translational addition of hydrophobic isoprenyl groups to its C-terminal containing CAAX box, mediated by farnesyl transferase, allowing Ras to bind to the inner cell membrane.^119-121 Exchange of GDP for GTP can then occur by nucleotide exchange factors, such as mSos (mammalian homologue of the son of sevenless gene product identified as a Ras activator in D. melanogaster.)¹²² mSos is thought to be always bound to Grb-2, which is an adaptor protein containing a SH2 and two SH3 domains, by its proline rich regions interacting with SH3 binding sites on Grb-2. This Grb-2:Sos complex is brought to the cell membrane near Ras.GDP by binding of the Grb-2 SH2 domain to specific phosphorylated tyrosine(Y) residues on activated RTKs either directly or indirectly. Specificity for SH2 domains is conferred by the 3-4 amino acids C-terminal to the phosphotyrosine binding site (YXXX);118 Indirectly, the Grb-2:Sos complex can bind activated RTKs through another SH2 and PTB containing adapter protein called Shc, which has transforming properties itself, Figure 2.^123-127

Once mSos exchanges Ras bound GDP for GTP, activated Ras.GTP can associate with several downstream effectors. These effectors include neurofibromin (a major negative regulator of Ras activity as described below), but in particular the main mitogenic cascade through Raf, MAPKK (mitogen-activated protein kinase kinase, also known as MEK or ERK kinase), and MAPK (mitogen-activated protein kinase, also known as extracellular signal related kinase or ERK) (Figure 2).

Inactivation: Inactivation of Ras.GTP to Ras.GDP requires binding of the family of enzymes called GAPs (GTP'ase Activating Protein). The two main mammalian Ras-GAPs are _p120GAP and neurofibromin (the gene product of NF-1).^{88-91,94,128-133} Decreased levels of these Ras-GAPs could in theory lead to elevated levels of active Ras.GTP, the presence of which has been documented in neurofibrosarcoma cells lacking neurofibromin.^95,96 Using an enzymatic assay that we have developed, that allows for the first time measurement of levels of GDP and GTP bound to Ras in tissues quantitatively, we have demonstrated that levels of Ras.GTP are increased in NF-1 peripheral nerve tumors compared to other non NF-1 tumors.¹⁰ Which Ras-GAP is predominantly utilized by a particular cell type in regulation of Ras is an area of ongoing research. Recent evidence suggests that neurofibromin (ubiquitously expressed) may be the most important Ras-GAP in basal regulation of Ras.GTP levels, while _p120GAP (containing an SH2 domain capable of interacting with RTKs) may play a dominant role when Ras activation is stimulated by activated RTKs (personal communication -Tony Pawson).

Like other tumor suppressor genes, mutations in the NF-1 gene exist in many sporadic tumors including astrocytomas.¹³⁴ We have examined low and high grade astrocytoma specimens for neurofibromin expression using RT-PCR and Western immunoblot analysis.¹³⁵ Contrary to what one may predict, we have found that neurofibromin levels are elevated in the more malignant astrocytomas. We hypothesized that this may reflect increased Ras.GTP levels in these malignant astrocytomas as suggested by our current work, with neurofibromin levels being increased secondarily by the tumor cells in an attempt to decrease Ras.GTP levels. Elevated levels of neurofibromin (mRNA and protein) in both Ras transformed fibroblasts compared to their normal counterparts, and human malignant astrocytoma cells (high levels of Ras.GTP) compared to those transfected with the Ha-Ras-Asn17 inhibitory mutant (low levels of Ras.GTP), are in agreement with our hypothesis.¹³⁵

Methodology

Inhibition of Ras signaling leads to decreased proliferation and VEGF secretion of human astrocytoma cell lines: Four established human malignant astrocytoma cell lines were transfected with the Ha-Ras-Asn17 dominant inhibitory mutant and also treated with the farnesyl transferase inhibitor (FTI) L-739,749 (Merck Research Laboratories) to inhibit the Ras signaling pathway. The cell lines were characterized for expression and tyrosine phosphorylation of PDGF and EGF receptors, Shc and Grb2 under serum starved and ligand stimulated conditions with Western immunoblot analysis. Ras activity in the cells was measured with the ³²P-Ras loading assay, Figure 3A, 3C and in 20 flash frozen human GBM specimens and two normal human brain specimens using the enzymatic assay described in prior publications.^10,98,136 To determine if inhibition of Ras activation resulted in decreased MAPKinase activity, the major mitogenic signal downstream of activated Ras, the mobility shift assay, an activation-specific MAPKinase antibody (New England Biolabs, U.S.A.), and the myelin basic protein kinase assays were used (data not shown). Colony formation and anchorage dependent and independent proliferation assays, as demonstrated with the U87 cell line, were undertaken on the stably derived clones to determine the effects of inhibiting the Ras signaling pathway in human astrocytoma cells, Figure 4A, 4B. VEGF secretion was evaluated by Western-immunoblot analysis (Figure 5) and ELISA (data not shown) assays, on the conditioned media obtained from the parental astrocytoma cell lines or those with inhibition of Ras activity.

Ras-GTP levels and VEGF expression are elevated in malignant NF-1 neurogenic sarcomas: Ras.GTP levels were measured in the flash frozen specimens of both astrocytomas (Figure 3B) and peripheral nerve tumors (Figure 6), using a novel non-radioactive enzymatic assay for quantitative measurement of Ras.GTP.^10,98,136 Expression of NF-1 mRNA (data not shown) was detected using RT-PCR from total RNA obtained from the NF-1 neurogenic sarcoma, NF-1 neurofibromas and non NF-1 schwannomas specimens using primers and techniques as described in a prior publication.¹⁰ Western-immunoblotblot analysis (Figure 7) and immunohistochemistry (data not shown) with an affinity purified antibody made against a trpE:-human neurofibromin fusion protein injected into rabbits¹³⁰(generously donated by Dr. N. Ratner, Univ. of Cincinnati), was undertaken on the specimens. Embryonic mouse fibroblasts with homozygous knockout of both NF1 alleles (NF1:-/-; gift of Dr. Tyler Jacks, M.I.T., Boston, Massachusetts), and normal mice (NF1:+/+) were used as negative and positive controls, respectively, in the Western blot analysis. VEGF mRNA expression in the NF-1 neurogenic sarcoma cell lines was quantified by Northern blot analysis, Figure 8. Expression of the 4.2Kb VEGF mRNA was determined by hybridization of the Northern blot, with a ³²P radiolabelled VEGF cDNA probe which contains 204 base pairs of the human VEGF sequence, common for all four isoforms (gift from Dr. B. Berse).

Results

Levels of Ras.GTP are elevated in human malignant astrocytoma cell lines and operative specimens which overexpress PDGFRs and EGFRs: Of the four established human malignant astrocytoma cell lines, PDGF-aR was expressed only by the U373 cells, while the other three expressed variable but increased amounts of PDGF-bR compared to NIH/3T3 cells (data not shown). All the cell lines expressed large amounts of EGFR. Upon activation by exogenous PDGF or EGF the corresponding receptors expressed by the astrocytoma cell lines bound to Shc and Grb2, which are signaling proteins involved in activation of Ras (data not shown).

³²P-Ras loading experiments demonstrated that in v-Ha-ras transformed fibroblasts (RT8 cells) approximately 30% of the total Ras was in the Ras.GTP bound state, compared to 5% for non-transformed NIH/3T3 cells, Figure 3A. Although human malignant astrocytomas do not harbor oncogenic Ras mutations,⁹⁹ 20-30% of the total Ras was in the GTP bound state in each of the four astrocytoma cell lines. Thus the level of Ras.GTP in human malignant astrocytoma cell lines is similar to that found in Ras transformed RT8 cells. Using an enzymatic assay (Figure 3B), 20 human GBMs and 2 non-neoplastic brain specimens (head injury) were evaluated for Ras activity. The mean amount of Ras.GTP was 1.62 +/- .36 f.moles/mg of GBM DNA, compared to .08 and .04 f.moles/mg of non-neoplastic brain DNA. When Ras activity was expressed as percent of total Ras normalized to mg tumor protein (%Ras.GTP/Ras.GDP + Ras.GTP), the value for the GBM specimens were similarily elevated at 53.5 +/- 5%, compared to .5 and 2.3% in the non-neoplastic human brain samples (data not shown). In comparison, values for Ras.GTP in v-Ha-ras transformed RT8 and non-transformed NIH/3T3 fibroblasts varied between .8-1.0 and below detection level to .03 Ras.GTP f.moles/mg of DNA respectively (data not shown).

Effect of blocking Ras activation on in-vitro proliferation of human malignant astrocytoma cell lines: To inhibit Ras activation, astrocytoma cells were transfected with the Ha-Ras-Asn17 dominant inhibitory mutant. ³²P-Ras loading assays were used to measure levels of Ras.GTP in the U373 clones transfected with Ha-Ras-Asn17 construct, Figure 3C. The mean percentage of total Ras in the Ras.GTP bound form in the U373 cells transfected with the empty MMTV vector was about 20%, measured in four separate experiments. This value was similar to the v-ras transformed RT8 fibroblasts and parental U373 cells in Figure 3A. In comparison, expression of Ha-Ras-Asn17 significantly decreased levels of activated Ras.GTP in the U373 cells to a mean value of about 10%. Activation of MAP kinase was decreased by expression of Ha-Ras-N17 as detected by the mobility shift, an activation specific antibody, and the myelin basic protein kinase assay (data not shown).

Colony formation and anchorage-dependent proliferation assays, on pooled or stably selected clones of transfected astrocytoma cells, both demonstrated that the proliferation rates of the four astrocytoma cell lines were decreased by blocking Ras activation by the Ha-Ras-Asn17 dominant inhibitory mutant (data not shown). Anchorage independent soft agar assay on U87 human malignant astrocytoma cells transfected with the Ha-Ras-Asn17 inhibitory mutant also demonstrated decreased tumorgenic proliferation, Figure 4A, with the number and average size of the U87 colonies with decreased Ras.GTP levels due to Ha-Ras-Asn17 expression much reduced, compared to MMTV vector only transfected U87 cells.

Treatment with an FTI (12 days of 10mM L-739,749), reduced by 54% and 48% the number of U87 cells treated with L-739,749 compared to parental U87 cells (p = 0.0108 by paired t-test) and vehicle-treated cells (p = 0.0063 by paired t-test) respectively, Figure 4B. Inhibition of proliferation by FTIs has been noted in a wide panel of human astrocytoma cell lines,¹³⁷ within a dose range found to be effective in Ras transformed human cancer cell lines and that well tolerated in animals with minimal toxicity.

VEGF expression and effect of Ras inactivation in human astrocytoma cell lines: The interrelation of the mitogenic Ras-Raf-MAPKinase pathway and VEGF mediated angiogenic pathway as discussed in the introduction, was explored by examining the conditioned media from human astrocytoma cell lines for VEGF secretion. Western immunoblot analysis of the conditioned media demonstrated abundant secretion of all four VEGF isoforms by the parental U118,U87 and U373 astrocytoma cell lines, which were decreased by inhibition of Ras activity by transfection with the Ha-Ras-Asn17 dominant inhibitory mutant or treatment with 10mM L739,749, a FTI, Figure 5. Appropriate controls for the experiments (MMTV vector only, or vehicle (methanol)) are included. Of interest, the constitutively activated and truncated EGFR (p140^EGFR) secretes higher levels of VEGF when expressed in the U118 astrocytoma cells, compared to U118 parentals. Recent experiments have demonstrated that this mutant receptor, which is expressed in 25-40% of GBMs and confers both in-vitro and in-vivo growth advantage, further elevates Ras activity (data not shown). This would be in keeping with our hypothesis that activation of the Ras pathway leads not only to mitogenic signals, but also angiogenic signals via increased VEGF expression, both elements vital to tumorgenic growth. The interrelationship of Ras activation and VEGF expression in the astrocytoma cells was further quantified with ELISA assays (data not shown), demonstrating decreased VEGF secretion with Ras inactivation by Ha-Ras-Asn17 or the FTI (L739,749).

Expression of neurofibromin by peripheral nerve tumors and levels of Ras activity: Western immunoblot (Figure 7) and immunohistochemistry (data not shown) were used to analyze expression of neurofibromin in the peripheral nerve specimens. Lysates from the NF-1 (+/+) embryonic fibroblasts (+'ve control) expressed the large approximately 240kDa neurofibromin protein, while the fibroblasts derived from homozygous NF-1

(-/-) deleted embryos (-'ve control) did not. Lysates from a non-NF-1 sporadic Schwannoma expressed abundant neurofibromin, while both NF-1 neurofibromas and NF-1 neurogenic sarcomas did not, consistent with the RT-PCR data (data not shown).

Ras activity was measured using the enzymatic assay described, with levels of Ras activity expressed as a percentage of total Ras (%Ras.GTP/Ras.GDP + Ras.GTP), or quantitatively as femtomoles of Ras.GTP/mg tumor DNA. The latter measurement gave a quantitative value, and normalized against a nuclear derived parameter (mg DNA) which may reflect more accurately the number of tumor cells in the samples assayed. This would theoretically account for differences in cellular size and amount of extracellular material between the samples, thereby giving a more accurate measurement of the amount of activated Ras in each of the tumor cells. In five NF-1 neurogenic sarcomas, the average percentage of activated Ras was 15.4 +/- 5.2%, or 46.0 +/- 7.0 fmol/mg DNA, Figure 6. The average activated Ras.GTP levels in the four NF-1 benign neurofibroma specimens was 6.0 +/- 2.5% or 13.3 +/- 2.6 fmol GTP/mg DNA, Figure 6. Levels of activated Ras.GTP in the benign neurofibromas, expressed either as a percentage of total Ras or in absolute amounts per mg DNA were therefore approximately one-third the levels detected in the NF-1 neurogenic sarcomas. In the four non NF-1 benign Schwannomas levels of activated Ras.GTP was 1.3 +/- .3% or 3.5 +/- .6 fmol GTP/mg DNA, Figure 6. These levels of activated Ras.GTP in the benign Schwannomas in non NF-1 patients were approximately 8% and 21% the levels found in the NF-1 neurogenic sarcomas and neurofibromas, respectively.

VEGF expression and angiogenesis in NF-1 neurogenic sarcomas: Our recent experiments have started to explore the inter-relationship between Ras activation, VEGF expression and tumor angiogenesis based on the rationale discussed in the introduction in peripheral nerve tumors. Although not as progressed as our work with astrocytomas presented above, some preliminary data do support our hypothesis that increased Ras activity in these tumors would be accompanied by a more angiogenic tumor and increased VEGF expression. In Figure 8, Northern blot analysis demonstrates abundant VEGF mRNA expression in a human NF-1 neurogenic sarcoma cell line (ST88-14), which has been documented to have elevated levels of activated Ras.GTP.^95,96 The levels, when normalized to b-actin, are similar to human malignant astrocytoma cell lines (U87,U373) which also have elevated Ras activity (Figure 3A) and secrete abundant VEGF which is decreased by inhibition of Ras activity by the Ha-Ras-Asn-17 dominant inhibitory mutant or the FTI, L-739,749, Figure 5. In addition, we recently published the clinical, pathological and molecular presentation of an aggressive NF-1 neurogenic sarcoma.⁷⁰ This particular tumor, whose Ras activity was measured and found to be elevated, was highly angiogenic and rapidly metastasized. Invasion of tumor cells into the tumor vascular channels leading to infarction of the tumor and subsequent acute presentation was noted. Expression of VEGF at both the mRNA and protein level in this tumor and other nerve tumor samples whose Ras activity has been measured are underway, to investigate the interrelationship of the mitogenic Ras pathway and VEGF expression.

Discussion and Conclusions

Astrocytomas: Understanding and potentially inhibiting the common signal transduction pathway(s) utilized by the various activated RTKs in human malignant astrocytomas may be of therapeutic advantage compared to inhibition of each RTK separately. For example, the protein kinase-C (PK-C) pathway has been implicated in astrocytoma proliferation, leading to clinical trials with PK-C inhibitors in recurrent GBMs with equivocal results.^138-142 There are several reasons to be interested in Ras-mediated signaling in human malignant astrocytomas expressing activated RTKs. First, the Ras-Raf-MAP kinase pathway is presently the best understood signaling pathway linking RTKs on the cell surface to the nucleus.^{92,100,102,118} Second, activation of Ras is important for proliferative or differentiating signals from a variety of RTKs, including PDGF-Rs and EGFR expressed by malignant astrocytomas.^101,103-106 Third, as demonstrated by our results, activation of Ras not only leads to tumor cell proliferation but also tumor vascularization, by transcriptional upregulation of the extremely potent and specific angiogenic factor, VEGF.^58,59 Hence, activation of Ras may be relevant to the proliferation of astrocytoma cells and to tumor angiogenesis, both of which appear critical for the growth of human malignant astrocytomas, which are one of the most vascularized of human neoplasms. Another reason to examine the role of Ras signaling in astrocytomas is the development of farnesyl-transferase inhibitors to block Ras activation. These agents reportedly have minimal toxicity and appear efficacious in inhibiting the growth of a variety of human tumors in nude mice, though their effect on human astrocytomas has not yet been reported.^{119-121,143-146} Tumors without oncogenic Ras mutations, can also be growth inhibited by farnesyl transferase inhibitors, suggesting that if astrocytomas are dependent on Ras signaling they may be sensitive to these agents, despite the lack of activating Ras mutations in astrocytomas.

Human astrocytomas and derived cell lines, including the ones used in this study, are known to express PDGF-Rs and EGFR.^{25,26,30,34,35,147-149} We demonstrate in this study that these RTKs expressed by the astrocytoma cell lines are capable of activating Ras. The adaptors Shc and Grb-2, are involved in activation of the nucleotide exchange factor mSos by activated RTKs, possibly by bringing mSos in proximity to Ras.GDP at the inner cell membrane, thereby converting it to Ras.GTP,^{118,122-125,127,150-158} or by inducing direct mSos catalytic activation.¹⁵⁹ Upon stimulation, both EGFR and PDGF-R (data not shown) expressed by the malignant astrocytoma cell lines bound Shc proteins and induced tyrosine phosphorylation of the 46 and 52 kDa Shc isoforms. Upon stimulation, the SH2 domain of Grb-2 was able to bind to tyrosine phosphorylated Shc. In summary, these results demonstrate that PDGF-Rs and EGFR, and probably other RTKs expressed by human malignant astrocytomas, are capable of activating the Ras signaling pathway.

Levels of activated Ras.GTP were elevated in all of the four astrocytoma lines, and comparable to oncogenic Ras transformed fibroblasts, Figure 3A. NIH-3T3 cells rather than non-transformed human astrocytes were used as negative controls, as the latter are hard to obtain and sustain in culture to yield sufficient cell numbers to undertake ³²P-Ras loading experiments. To exclude culture induced artifacts, we utilized an enzymatic assay to measure Ras.GTP and Ras.GDP levels in flash frozen tissues.^10,98,136 Similar to the established cell cultures, Ras activity was markedly elevated in the GBM specimens, compared to non-neoplastic brain samples, Figure 3B. We have also recently started to explore whether the Ras pathway is utilized by the p140^EGFR mutant expressed by a large percentage of GBMs, and which has been demonstrated to be constitutively activated and provide growth advantage.^28,29 The p140^EGFR mutant which is not expressed by astrocytoma cell cultures but only in the actual tumors, was transfected into a human malignant astrocytoma cell line (U118) expressing normal EGFR. Although levels of Ras.GTP was elevated in the parental U118 cells similar to other malignant astrocytoma lines (Figure 3A), expression of the p140^EGFR mutant increased basal levels of Ras activity in serum starved conditions even more (data not shown). The U118-p140^EGFR mutant expressing cells are more tumorgenic in anchorage dependent/independent assays and preliminary nude mice experiments compared to the parental U118 cells, the latter in keeping with their increased VEGF secretion, Figure 5. Another piece of evidence in support of the p140^EGFR mutant utilizing the Ras signaling pathway is their increased sensitivity to FTIs, compared to astrocytoma cells expressing only the normal EGFR (data not shown). Further experiments to characterize the Ras mediated signaling of the p140^EGFR mutant receptor are presently ongoing, as is a study investigating whether there is a gradient of Ras activity between Grade 1 to Grade 4 (GBM) specimens using the enzymatic assay and specimens from the University of Toronto Nervous System Tumor Bank that we have established.

Expression of the Ha-Ras-Asn17 dominant inhibitory mutant decreased levels of Ras.GTP in the astrocytoma cells by approximately 50%, as measured in the U373 clone, Figure 3C. Decreased levels of Ras.GTP in the astrocytoma cells correlated with inhibition of the major mitogenic downstream effector of Ras, involving activation of the Ras-Raf-MAP Kinase pathway. Expression of Ha-Ras-Asn17 decreased proliferation in all astrocytoma cell lines, as observed in the pooled proliferation and colony formation assays and those undertaken on the isolated U373 clone. Anchorage independent growth of the U87 cell line (which has the most tumorgenic growth in nude mice of the four astrocytoma lines, unpublished observation), was decreased by blocking Ras activation with the Ha-Ras-Asn17 inhibitory mutant, Figure 4A. These results demonstrate that Ras activity is elevated in human malignant astrocytomas, and activation of this signaling pathway is important for tumor cell proliferation.

Farnesylation is the most critical post-translational modification step of Ras, essential for its localization to the inner cell membrane and subsequent activation by RTKs.^160,161 Farnesyl transferase inhibitors (FTI) have demonstrated minimal toxicity in animal models, while reducing growth of a variety of human tumors, with malignant astrocytomas not yet tested.^{97,119-121,143-146} Surprisingly, growth inhibition was independent of whether or not the tumors harbor activating oncogenic Ras mutations.^162,163 For example, human neurogenic sarcoma cell lines are inhibited by these drugs,97 though elevated levels of Ras.GTP in these cells are secondary to decreased levels of neurofibromin, one of the two major mammalian Ras-GAPs,^95,96 and not due to activating mutations of Ras. Our preliminary data with L-739,749, a FTI developed by Merck Research Laboratories, demonstrates that pharmacological inhibition of the Ras signaling pathway inhibits proliferation of human astrocytoma cells, Figure 4B. In addition, these and other strategies directed at inhibiting the Ras signaling pathway may also block tumor induced angiogenesis. As demonstrated by our results in Figure 5, the abundant VEGF secretion by malignant astrocytomas was decreased by genetic or pharmacological inhibition of the Ras pathway. Hence, these agents may have a greater in-vivo effect in decreasing overall tumor growth compared to their documented in-vitro anti-mitogenic effects, since they may inhibit not only tumor cell proliferation but also tumor induced angiogenesis. Current work in my laboratory is directed to evaluating FTI in animal models of human malignant astrocytomas, to ultimately determine their potential therapeutic role in this presently terminal human cancer.

Neurofibromas: Previously, Ras activity could only semi-quantitatively be measured in cell cultures using the ³²P-Ras loading technique, as utilized for the malignant astrocytoma cell lines, Figure 3A. The enzymatic assay (Figure 3B) developed in collaboration with Dr. Gerry Boss, yields results that are quite comparable to the ³²P-Ras loading technique.^10,98,136 This assay of course has wide applicability in other tumor systems where Ras activity is increased through oncogenic mutations, increased activation (such as in malignant astrocytomas), or decreased inactivation (such as in NF-1 neurofibromas). Some limitations of the assay exists however. 1) Relatively large quantities of tissue are required, as amounts less than approximately 2c.m.3 did not provide reliable measurements. 2) Specimens should be flash frozen (within 30 sec. in these experiments), although the exact time before significant phosphatase and protease activity leads to degradation of the samples is not known. 3) Measurements of Ras bound guanine nucleotides are from the entire tumor, and do not specify the levels specifically found in the tumor cells only, as distinct from infiltrating and surrounding cells. Hence, both tumor and non-tumor cells (a variable and occasionally significant proportion of the cells in some tumors) within the specimens, contribute to the measured values of Ras bound GTP and GDP. This obstacle could be overcome by the development of activation-specific antibodies, capable of recognizing specifically Ras.GTP. 4) Another potential source of variability between tumors is the extent of acellular areas, which may significantly alter the measured levels of Ras bound guanine nucleotides if expressed in terms of protein. This variability is minimized by expressing levels of activated Ras in the specimens as a percentage (%GTP/GDP + GTP), or absolute amounts of Ras.GTP in terms of mg DNA, to standardize for the cellular content between the specimens. Similar conclusions regarding the levels of activated Ras in both the peripheral nerve tumors and astrocytoma experiments were reached using both these methods.

Although NF-1 tumors are predicted to have elevated levels of Ras.GTP due to loss of neurofibromin, a major mammalian Ras-GAP involved in inactivating Ras.GTP to Ras.GDP, this was previously demonstrated in NF-1 neurogenic sarcoma cell lines and not in tissues.^95,96 The results of this study measuring Ras activity in peripheral nerve tumor specimens are in agreement with the cell culture data. Levels of Ras.GTP were elevated approximately 15X in the NF-1 neurogenic sarcomas, compared with non NF-1 Schwannomas, Figure 6. In addition to verification that levels of Ras activity were elevated in the actual NF-1 neurogenic sarcoma specimens, the enzymatic based assay allowed a determination of Ras.GTP levels in benign NF-1 neurofibromas, which do not grow in culture. Compared to non NF-1 Schwannomas, levels of Ras.GTP were increased approximately 4X in these NF-1 neurofibromas, Figure 6. Neurofibromin Western blot analysis (Figure 7) failed to demonstrate any detectable levels of neurofibromin by the NF-1 neurofibroma. Immunohistochemistry although less conclusive was also supportive, since less overall neurofibromin-staining was observed in the NF-1 neurofibromas where many cells did not positively stain compared to the non NF-1 Schwannomas (data not shown). We hypothesize that these neurofibromin negative cells are the actual transformed cells that form the NF-1 neurofibromas and contribute to the elevated levels of activated Ras.GTP. Double labeling experiments with neurofibromin and cell type specific antibodies, to characterize the cellular subpopulations in neurofibromas, are underway to help resolve this issue.

Therefore these experiments demonstrate increased Ras activity in NF-1 peripheral nerve tumors, not due to oncogenic mutations but as a result of loss of neurofibromin, a major inactivator of activated Ras.GTP. In addition to the loss of both NF-1 alleles which lead to the NF-1 neurofibromas as per the "two hit hypothesis" of tumor suppressor genes,^164,165 additional genetic aberrations such as loss of p53 gene has been implicated or are being sought in the malignant neurogenic sarcomatous transformation which occurs in 3-5% of the deeper plexiform neurofibromas in NF-1.¹¹ The increased growth rate, tumor angiogenesis, and metastasizing capability of NF-1 neurogenic sarcomas compared to their benign counterparts are reflected in the increased levels of activated Ras.GTP, Figure 6. We hypothesize and demonstrate in preliminary experiments (Figure 8), that similar to malignant astrocytomas the increased Ras activity in NF-1 neurogenic sarcomas is also accompanied by transcriptional up-regulation of VEGF, the most potent and specific angiogenic factor implicated in tumor angiogenesis and metastasis.

These experiments have demonstrated in cell culture experiments and in operative specimens the involvement of the Ras mediated signaling pathway in two nervous system tumors. Although oncogenic mutations of Ras are not present in astrocytomas and peripheral nerve tumors, there is a functional activation of this critical mitogenic signaling pathway. Aberrant Ras activity leads not only to an increase in cellular proliferation in these two tumor systems, but as our preliminary experiments demonstrate, may play a vital role in regulating VEGF mediated tumor angiogenesis. Survival curves of human malignant astrocytomas and neurogenic sarcomas have not improved despite advances in neurosurgical capabilities, due to limitations posed by the molecular biology of these tumors. The experiments in this report and others that are ongoing in our laboratory, are the first steps in evaluating the potential use of modulators of the Ras signaling pathway in the management of these presently terminal cancers.

Acknowledgements

This work was undertaken by the following members of my laboratory: Matthias Feldkamp, Nelson Lau, Liliana Angelov and Prateek Lala. Active collaborations include Tony Pawson (Toronto), David Gutmann (St.Louis), Gerry Boss (LaJolla) and Charles Stiles (Boston). I would like to acknowledge support grants from MRC-Clinician Scientist Programme, National Cancer Inst. of Canada, Physician and Scientist Inc. and Neurofibromatosis Foundation. Last, I would like to extend my thanks and appreciation to my personal family and my neurosurgical family at The Toronto Hospital.

References

1. Mahaley M, Mettlin, C Natarajan, N Laws E, Peace B. National survey of patterns of care for brain tumor patients. J Neurosurgery 1989; 71: 826-836.
2. Kleihues P, Burger P, Scheithauer, B. Histological Typing of Tumors of the Nervous System, 2nd Edition. Berlin: Springer-Verlag, 1993.
3. Daumas-Duport C, Scheithauer B, O'Fallon J, Kelly P. Grading of astrocytomas; a simple and reproducible method. Cancer 1988; 62: 2152-2165.
4. von Recklinghausen FD. Über die multiplen Fibrome der Haut und ihre Beziehung zu den multiplen Neuromen. Berlin: Hirschwald, 1882.
5. Lott IT, Richardson Jr EP. Neuropathological findings and the biology of neurofibromatosis. Adv Neurol 1981; 29: 23-32.
6. Riccardi V. Neurofibromatosis: phenotype, natural history, and pathogenesis. Baltimore: John Hopkins Univ. Press, 1992.
7. Grifa A, Piemontese MR, Melchionda S, et al. Screening of neurofibromatosis type 1 gene: identification of a large deletion and of an intronic variant. Clin Genet 1995; 47: 281-284.
8. Zoller M, Rembeck B, Akesson HO, Angervall L. Life expectancy, mortality and prognostic factors in neurofibromatosis type 1. A twelve-year follow-up of an epidemiological study in Goteborg, Sweden. Acta Derm Venereol 1995; 75: 136-140.
9. Sorensen S, Mulvhill J, Nielsen A. Longterm followup of von Recklinghausen neurofibromatosis. Survival and malignant neoplasms. N Engl J Med. 1986; 314: 1010-1015.
10. Guha A, Lau N, Gutmann D, et al. Ras-GTP Levels are elevated in human NF1 peripheral nerve tumors. Oncogene 1996; 12: 507-513.
11. Guha A., Bilbao J, Kline DG, Hudson AR. Tumors of the peripheral nervous system. In: Youmans JR, ed., Tumors of the Peripheral Nervous System, pp. Philadelphia: W.B. Saunders, 1996.
12. Bigner SH, Mark J, Burger PC, et al. Specific chromosomal abnormalities in malignant human gliomas. Cancer Research 1988; 48: 405-411.
13. von Deimling A, Eibl R, Ohgaki H, et al. p53 Mutations are associated with 17p allelic loss in grade11 and grade111 astrocytoma. Cancer Res 1992; 52: 2987-2990.
14. Sidransky D, Mikkelsen T, Schwechheimer K, et al. Clonal expansion of p53 mutant cells is associated with brain tumor progression. Nature 1992; 355: 846-847.
15. Pershouse, MA, Stubblefield E, Hadi A, et al. Analysis of the functional role of chromosome 10 loss in human glioblastomas. Cancer Res, 1993. 53: 5043-5050.
16. James C, Collins V. Molecular genetic characterization of CNS tumor oncogenesis. Adv Can Res 1992; 58: 121-142.
17. James CD, Carlbom E, Dumanski JP, et al. Clonal genomic alterations in glioma malignancy stages. Cancer Res 1988; 48: 5546-5551.
18. Li D-M, Sun H. TEP1, Encoded by a candidate tumor suppressor locus, is a novel protein tyrosine phosphatase regulated by transforming growth factor-b. Cancer Res 1997; 57: 2124-2129.
19. Li J, Yen C, Liaw D, et al. PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast and prostate cancer. Science 1997; 275: 1943-1947.
20. Steck P, Pershouse, MA, Jasser, SA, et al. Identification of a candidate tumor suppressor gene, MMAC1, at chromosome 10q23.3 that is mutated in multiple advanced cancers. Nature Genetics 1997; 15: 356-362.
21. Kamb A, Gruis NA, Weaver-Feldhaus, et al. A cell cycle regulator potentially involved in genesis of many tumor types. Science 1994; 264: 436-440.
22. Noborl T, Miura K, Wu DJ, et al. Deletions of the cyclin-dependent kinase-4 inhibitor gene in multiple human cancers. Nature 1994; 368: 753-756.
23. Rubio M-P, Correa KM, Ueki K, et al. Putative glioma tumor suppressor gene on chromosome 19q maps between APOC2 and HRC. Cancer Res 1994; 54: 4760-4763.
24. Libermann T, Razon N. Expression of EGF receptors in human brain tumors. Cancer Res 1984; 44: 753-760.
25. Libermann T, Nusbaum H, Razon N, et al. Amplification, enhanced expression and possible rearrangement of EGF-receptor gene in primary human tumors of glial origin. Nature 1985; 313: 144-147.
26. Steck P, Lee P, Hung M-C, Yung W. Expression of an altered epidermal growth factor receptor by human glioblastoma cells. Cancer Res 1988; 48: 5433-5439.
27. Ekstrand A, Sugawa N, James C, Collins V. Amplified and rearranged epiderma growth factor receptor genes in human glioblastomas reveal deletions of sequences encoding portions of the N- and/or C-terminal tails. Proc Natl Acad Sci USA 1992; 89: 4309-4313.
28. Ekstrand AJ, Longo N, Hamid ML, et al. Functional characterization of a EGF receptor with a truncated extracellular domain expressed in glioblastomas with EGF-R gene amplification. Oncogene 1994; 9: 2313-2320.
29. Nishikawa R, JiX, Harmon RC, et al. A mutant epidermal growth factor receptor common in human gliomas confers enhanced tumorigenicity. Proc Nat Acad Sci USA 1994; 91: 7727-7731.
30. Nistér M, Claesson-Welsh L, Eriksson A, Heldin C-H, Westermark B. Differential expression of platelet derived growth factor receptors in human malignant glioma cell lines. J Biol Chem, 1991; 266: 16755-16763.
31. Fleming T, Saxena A, Clark W, et al. Amplification and/or overexpression of platelet derived growth factor receptors and epidermal growth factor receptor in human glial tumors. Cancer Res 1992; 52: 4550-4553.
32. Fleming TP, Matsui T, Heidran MA, et al. Demonstration of an activated platelet-derived growth factor autocrine pathway and its role in human tumor cell proliferation in vitro 1992; 7: 1355-1359,
33. Guha, A. Platelet derived growth factor: a general review with emphasis on astrocytomas. Pediatr Neurosurg 1992; 92: 14-20.
34. Guha A, Dashner K, McL Black P, Wagner J, Stiles C. Expression of PDGF and PDGF-receptors in human astrocytoma operative specimens. Int J Cancer 1995: 60: 168-173.
35. Guha A, Gowacka D, Carrol R, et al. Expression of platelet derived growth factor and platelet derived growth factor receptor mRNA in a glioblastoma from a patient with Li-Fraumeni syndrome. J Neurol Neurosurg Psychiatry 1995; 58: 711-714.
36. Mercola M, Deininger P, Shamah S, et al. Dominant-negative mutants of a PDGF gene. Genes Development 1990; 4: 2333-2341.
37. Shamah S, Stiles C, Guha A. Dominant-negative mutants of platelet-derived growth factor (PDGF) revert the transformed phenotype of human astrocytoma cells. Mol Cell Biol 1993; 13: 7203-7212.
38. Folkman J, Shing Y. Angiogenesis. J Biol Chem 1992; 267: 10931-10934.
39. Folkman J. What is the evidence that tumors are angiogenesis-dependent? J. Natl Cancer Inst 1990; 82: 4-6.
40. Folkman J, Klagsburn M. Angiogenic factors. Science 1987; 235: 442-447.
41. Klein G, Weinhouse S. Tumor. Angiogenesis. In: Folkman, J., ed. Advances in Cancer Research, New York, Academic Press 1974; 43-52.
42. Folkma, J. How is blood vessel growth regulated in normal and neoplastic tissue? Cancer Res, 1986. 46: 467-473.
43. Fidler IJ, Ellis LM. The implications of angiogenesis for the biology and therapy of cancer metastasis. Cell 1994; 79: 185-188.
44. Bouck N. Tumor angiogenesis: the role of oncogenes and tumor suppressor genes. Cancer Cells 1990; 2: 179-185.
45. O'Reilly MS, Holmgre L, Shing Y, et al. Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 1994; 70: 315-328.
46. Senger DR, Galli SJ, Dvorak AM, et al. Tumor cells secret a vascular permeability factor that promotes accumulation of ascites fluid. Science 1983; 219: 983-985
47. Senger DR, Peruzzi CA, Feder J, Dvorak HF. A highly conserved vascular permeability factor secreted by a variety of human and rodent tumor cell lines. Cancer Res 1986; 46: 5629-5632.
48. Senger DR, Connelly DT, Van De Water L, Feder J, Dvorak H.F. Purification and NH2-terminal amino acid sequence of guinea pig tumor-secreted vascular permeability factor. Cancer Res 1990; 50: 1774-1778.
49. Leung DW, Cachianes G, Kuang W-K, Goeddel DV, Ferrara N. Vascular endothelial growth factor is a secreted angiogenic mitogen. Science 1989; 246: 1306-1312.
50. Ferrara N, Houck K, Jakeman L, Leung D. Molecular and biological properties of the vascular endothelial growth factor family of proteins. Endocrine Rev 1992; 13: 18-32.
51. Houck KA, Ferrara N, Winer J, et al. The vascular endothelial growth factor family: identification of a fourth molecular species and characterization of alternative splicing of RNA. Mol Endocrinol 1991; 5.
52. Tischer E, Mitchell R, Hartman T, et al. The human gene for vascular endothelial growth factor. J Biol Chem 1991; 266: 11947-11954.
53. Tsai J-C, Goldman, CK, Gillespie GY. Vascular endothelial growth factor in human glioma cell lines: induced secretion by EGF, PDGF-BB, and bFGF. J Neurosurg 1995; 82: 864-873.
54. Shweiki D, Itin A, Soffer D, Keshe E. Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 1992; 359: 843-845.
55. Goldman CK, Kim J, Wong W-L, et al. Epidermal growth factor stimulates vascular endothelial growth factor production by human malignant glioma cells: a model of glioblastoma multiforme pathophysiology. Mol Biol Cell 1993; 4: 121-133.
56. Mukhopadhyay D, Tsiokas, L, Zhou X-M, et al. Hypoxic induction of human vascular endothelial growth factor expression through c-src activation. Nature 1995; 375: 577-581.
57. Finkenzeller G, Technau A, Marme D. Hypoxia induced transcription of the vascular endothelial growth factor gene is independent of functional AP1 transcription gactor. Biochem Biophys Res Comm 1995; 208: 432-439.
58. Rak J, Mitsuhashi Y, Bayko L, et al. Mutant ras oncogenes upregulate VEGF/VPF expression: implications for induction and inhibition of tumor angiogenesis. Cancer Res 1995; 55: 4575-4580.
59. Grugel S, Finkenzeller G, Weindel K, Barleon B, Marme D. Both v-Ha-ras and v-raf stimulate expression of the vascular endothelial growth factor in NIH 3T3 cells. J Biol Chem 1995; 270: 25915-25919.
60. Barelon B, Hauser S, Schollmann C, et al. Differential expression of the two VEGF receptors flt and KDR in placenta and vascular endothelial cells. J Cell Biochem 1994; 54: 56-66.
61. De Vries, C, Escobedo, JA, Ueno H, et al. The FMS-like tyrosine kinase, a receptor for vascular endothelial growth factor. Science 1992; 255: 989-991.
62. Terman BI, Dougher-Vermazen D, Carrion ME, et al. Identification of the KDR tyrosine kinase as a receptor for vascular endothelial cell growth factor. Biochem Biophys Res Comm 1992; 87: 1579-1586.
63. Brown LF, Berse B, Jackman RW, et al. Expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in adenocarcinomas of the gastrointestinal tract. Cancer Res 1993; 53: 4727-4735.
64. Breier G, Albrecht U, Sterrer S, Risau W. Expression of vascular endothelial growth factor during embryonic angiogenesis and endothelial cell differentiation. Development 1992; 114: 521-532.
65. Millauer B, Wizigmann-Voos S, SchnGrch H, et al. High affinity VEGF binding and developmental expression suggest Flk-1 as a major regulator of fasculogenesis and angiogenesis. Cell 1993; 72: 835-846.
66. Connelly DT, Heuvelman DM, Nelson R, et al. Tumor vascular permeability factor stimulates endothelial cell growth and angiogenesis. J Clin Invest 1989; 84: 1470-1478.
67. Pepper MS, Ferrara N, Orci L, Montesano R. Vascular endothelial growth factor (VEGF) induces plasminogen activators and plasminogen activator inhibitor-1 in microvascular endothelial cells. Biochem Biophys Res Commun 1991; 181: 902-906.
68. Unemori EN, Ferrara N, Bauer EA, Amenoto EP. Vascular endothelial growth factor induces interstitial collagenase expression in human endothelial cells. J Cell Physiol 1992; 153: 557-562.
69. Burger PC, Scheithauer B, Vogel FS. Surgical pathology of the nervous system and its coverings. Churchill-Livingstone, London. 1991.
70. Feldkamp M, Lau N, Provias J, Guha A. Acute presentation of a neurogenic sarcoma in a patient with neurofibromatosis type 1: a pathologic and molecular explanation. J Neurosurgery, accepted: 1996.
71. Sheela S, Riccardi V, Ratner N. Angiogenic and invasive properties of neurofibroma Schwann cells. J. Cell Biol 1990; 111: 645-653.
72. Windel K, Moringlane JR, Marme D, Weich HA. Detection and quantification of vascular endothelial growth factor/vascular permeability factor in brain tumor tissue and cyst fluid: the key to angiogenesis? Neurosurgery 1994; 35: 439-449.
73. Plate KH, Breier G, Millauer B, Ullrich A, Risau W. Up-regulation of vascular endothelial growth factor and its cognate receptors in a rat glioma model of tumor angiogenesis. Cancer Res 1993; 53: 5822-5827.
74. Kim KJ, Li B, Armanini M, et al. Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumor growth in vivo. Nature 1993; 362: 841-844.
75. Beckmann M, Betsholtz C, Heldin C, et al. Comparison of the biological properties and transforming potential of human PDGF-A and PDGF-B Chains. Science 1988; 241: 1346-1349.
76. Provias JP, Claffey K, delAguila L, et al. Meningiomas: the role of vascular permeability factor/vascular endothelial growth factor (VPF/VEGF) in angiogenesis and peri-tumoral edema. Neurosurgery 1997; 40: 1016-1026.
77. Sioussat TM, Dvorak HF, Brock TA, Senger DR. Inhibition of vascular permeability factor (vascular endothelial growth factor) with antipeptide antibodies. Arch Biochem Biophys 1993; 301: 15-20.
78. Melnyk O, Shuman MA, Kim KJ. Vascular endothelial growth factor promotes tumor dissemination by a mechanism distinct from its effect on primary tumor growth. Cancer Res, 1996. 56: 921-924.
79. Asano M, Yukita A, Matsumoto T, Kondo S, Suzuki H. Inhibition of tumor growth and metastasis by an immuno-neutralizing monoclonal antibody to human vascular endothelial growth gactor/vascular permeability factor121. Cancer Res 1995; 55: 5296-5301.
80. Saleh M, Stacker SA, Wilks AF. Inhibition of growth of C6 glioma cells in-vitro by expression of antisense vascular endothelial growth factor sequence. Cancer Res 1996; 56: 393-401.
81. Millauer B, Shawver LK, Plate KH, Risau W, Ullrich A. Glioblastoma growth inhibited in vivo by a dominant-negative FlK-1 mutant. Nature 1994; 367: 576-578.
82. Strawn L, McMahon G, App H, et al. Flk-1 as a target for tumor growth inhibition. Cancer Res 1996; 56: 3540-3545,.
83. Marchuk DA, Saulino AM, Tavakkol R, et al. cDNA cloning of the type 1 neurofibromatosis gene: complete sequence of the NF1 gene product. Genomics 1991; 11: 931-940.
84. Upadhyaya M, Shen M, Cherryson A, et al. Analysis of mutations at the neurofibromatosis 1 (NF1) locus. Hum Mol Genet 1992; 1: 735-740.
85. Li Y, O'Connell P, Breidenbach HH, et al. Genomic organization of the neurofibromatosis 1 gene (NF1). Genomics 1995; 25: 9-18.
86. Stumpf S, Alksne JF, Annegers JF, et al. Neurofibromatosis. Conference statement. National Institutes of Health Consensus Development Conference. Arch Neurol 1988; 45: 575-578.
87. Buchberg AM, Cleveland LS, Jenkins NA, Copeland NG. Sequence homology shared by neurofibromatosis type-1 gene and IRA-1 and IRA-2 negative regulators of the RAS cyclic AMP pathway. Nature 1990; 347: 291-294.
88. Xu G, O'Conell P, Viskochil D, et al. The neurofibromatosis type 1 gene encodes a protein related to GAP. Cell 1990; 62: 599-608.
89. Bollag G, McCormick F. Differential regulation of rasGAP and neurofibromatosis gene product activities. Nature 1991; 351: 576-579.
90. Downward J. Regulation of p21ras by GTPase activating proteins and guanine nucleotide exchange proteins. Curr Opin Gen Dev 1992; 2: 13-18.
91. Downward JL. Regulatory mechanisms for ras proteins. BioEssays 1992; 14: 177-184.
92. McCormick F. How receptors turn ras on. Nature 1993; 363: 15-16.
93. McCormick F. Ras signaling and NF-1. Curr Opin Gen Dev 1995; 5: 51-55.
94. McCormick F. ras GTPase activating protein: signal transmitter and signal terminator. Cell 1989; 56: 5-8.
95. DeClue J, Papageorge A, Fletcher, et al. Abnormal regulation of mammalian p21ras contributes to malignant tumor growth in von Recklinghausen (type1) neurofibromatosis. Cell 1992; 69: 265-273.
96. Basu T, Gutman D, Fletcher J, et al. Aberrant regulation of ras proteins in malignant tumour cells from type1 neurofibromatosis patients. Nature 1992; 356: 713-715.
97. Yan N, Ricca C, Fletcher J, et al. Farnesyltransferase inhibitors block the neurofibromatosis type 1 (NF-1) malignant phenotype. Cancer Res, 1995. 55: 3569-3575.
98. Scheele JS, Rhee JM, Boss GR. Determination of absolute amounts of ras GDP and GTP bound to ras in mammalian cells: comparison of parental and ras-overproducing NIH 3T3 fibroblasts. Proc Natl Acad Sci USA 1994; 92: 1097-1100.
99. Bos JL. Ras oncogenes in human cancers: a review. Cancer Res 1989; 49: 4682-4689.
100. Cantley LC. Auger KR, Carpenter C, et al. Oncogenes and signal transduction. Cell 1991; 64: 281-302.
101. Satoh T, Endo M, Nakafuku M, et al. Accumulation of p21 ras-GTP in response to stimulation with epidermal growth factor and oncogene products with tyrosine kinase activity. Proc Natl Acad Sci USA 1990; 87: 7926-7929.
102. Satoh T, Kaziro Y. Ras in signal transduction. Cancer Biology 1992; 3: 169-177.
103. Satoh T, Endo M, Nakafuku M, Nakamura S, Kaziro Y. Platelet-derived growth factor stimulates formation of active p21ras.GTP complex in Swiss mouse 3T3 cells. Proc Natl Acad Sci 1990; 87: 5993-5997.
104. Muroya K, Hattori S, Nakamura S. Nerve growth factor induces accumulation of the GTP-bound form of p21ras in rat pheochromocytoma PC12 cells. Oncogene 1992; 7: 277-281.
105. Qui M-S, Green SH. NGF and EGF rapidly activate p21ras in PC12 cells by distinct, convergent pathways involving tyrosine phosphorylation. Neuron 1991; 7: 937-946.
106. Mulcahy LS, Smith MR, Stacey DW. Requirement for ras proto-oncogene function during serum-stimulated growth of NIH 3T3 cells. Nature 1985; 313: 241-243.
107. Kung H-F, Smith MR, Bekesi E, Manne V, Stacey DW. Reversal of transformed phenotype by monoclonal antibodies against Ha-ras p21 proteins. Exp Cell Res 1986; 162: 363-371.
108. Farnsworth C, Feig L. Dominant inhibitory mutations in the Mg+2 binding site of RasH prevent its activation by GTP. Mol Cell Biol 1991; 11: 4822-4829.
109. Farnsworth C, Marshall M, Gibbs J, Stacey D, Feig L. Preferential inhibition of the oncogenic form of RasH by mutations in the gap binding/"effector" domain.

 

Presented in part as the Royal College Medal in Surgery Lecture, 1997, to the Royal College of Physicians and Surgeons of Canada.

Received August 25, 1997. Accepted in final form January 26, 1998.

Reprint requests to: Abhijit Guha, 2-415 McLaughlin Pavilion, The Toronto Hospital, 399 Bathurst Street, Toronto, Ontario, Canada M5T 2S8

Can. J. Neurol. Sci. 1998; 25: 267-281