Molecular biology of central nervous system tumors

by David N. Louis, M.D., MGH Division of Neuropathology
and
Webster K. Cavenee, Ph.D., University of California at San Diego

Introduction

Neoplastic transformation appears to be a multi-step process in which the normal controls of cell proliferation and cell-cell interaction are lost, thus transforming a normal cell into a tumor cell. This tumorigenic process involves an interplay between at least two classes of genes: oncogenes and tumor suppressor genes. Oncogenes are abnormally activated versions of cellular genes that promote cell proliferation and growth. Activated oncogenes thereby result in an exaggerated impulse for a cell to grow and divide. Tumor suppressor genes, on the other hand, are normal genes that act to inhibit cell proliferation and growth. The inactivation of these genes results in tumor formation or progression. The most common scenario for inactivation of both copies of a tumor suppressor gene is mutation of one allelic copy, followed by loss of all or part of the chromosome bearing the second allelle. Consequently, the identification of consistent regions of chromosomal loss in specific tumor types suggests a tumor suppressor gene in that chromosomal region. These basic themes of oncogene activation and tumor suppressor gene inactivation coupled with chromosomal homozygosity, underlie the current molecular understanding of human tumor formation. The following chapter reviews the emerging knowledge on the molecular basis of brain tumorigenesis, covering primary tumors of the brain as well as other primary intracranial neoplasms that commonly affect the central nervous system.

Diffuse, fibrillary astrocytomas

Diffuse, fibrillary astrocytomas are the most common type of primary brain tumor in adults. These tumors are divided histopathologically into three grades of malignancy: World Health Organization (WHO) grade II astrocytoma, WHO grade III anaplastic astrocytoma and WHO grade IV glioblastoma multiforme (GBM). WHO grade II astocytomas are the most indolent of the diffuse astrocytoma spectrum. Nonetheless, these low-grade tumors are infiltrative and have a marked potential for malignant progression, and any biological model for astrocytomas must account for these cardinal features of malignant progression and invasion.

The p53 gene, a tumor suppressor gene located on chromosome 17p, has an integral role in a number of cellular processes, including cell cycle arrest, response to DNA damage, apoptosis, angiogenesis and differentiation; as a result, p53 has been dubbed the "guardian of the genome". The p53 gene is involved in the early stages of astrocytoma tumorigenesis 1. For instance, p53 mutations and allelic loss of chromosome 17p are observed in approximately one-third of all three grades of adult astrocytomas, suggesting that inactivation of p53 is important in the formation of the grade II tumors. Moreover, high grade astrocytomas with homogeneous p53 mutations evolve clonally from subpopulations of similarly mutated cells present in initially low grade tumors 2 . Such mutation studies are complemented by functional studies that have recapitulated the role of the p53 inactivation in the early stages of astrocytoma formation. For instance, cortical astrocytes from mice without functional p53 appear immortalized when grown in vitro and rapidly acquire a transformed phenotype. Cortical astrocytes from mice with one functional copy of p53 behave more like wild-type astrocytes and only show signs of immortalization and transformation after they have lost the one functional p53 copy 3,4. Interestingly, those cells without functional p53 become markedly aneuploid 3, confirming prior work showing that p53 loss results in genomic instability 5 and that astrocytomas with mutant p53 are often aneuploid 6. Thus, the abrogation of astrocytic p53 function appears to facilitate some events integral to neoplastic transformation, setting the stage for further malignant progression.

Many growth factors and their receptors are overexpressed in astrocytomas, including platelet-derived growth factor (PDGF), fibroblast growth factors (FGFs), and vascular endothelial growth factor (VEGF). For example, PDGF ligands and receptors are expressed approximately equally in all grades of astrocytoma, suggesting that such overexpression is also important in the initial stages of astrocytoma formation. Tumors often overexpress cognate PDGF ligands and receptors in an autocrine stimulatory fashion 7. The mechanisms for PDGF overexpression in most cases have not been elucidated, although rare astrocytomas display amplification of the PDGF a-receptor gene. Significantly, loss of chromosome 17p in the region of the p53 gene is closely correlated with PDGF a-receptor overexpression, in that 17p loss is most often seen in those astrocytomas that have PDGF a-receptor overexpression (M. Nistr and colleagues, in preparation). These observations may imply that p53 mutations have an oncogenic effect only in the presence of PDGF a-receptor overexpression. This interdependence is highlighted by observations that mouse astrocytes without functional p53 become transformed only in the presence of specific growth factors 4.

Astrocytomas display a remarkable tendency to infiltrate the surrounding brain, confounding therapeutic attempts at local control. These invasive abilities are often apparent in low-grade as well as high-grade tumors, implying that the invasive phenotype is acquired early in tumorigenesis. Investigations into astrocytoma invasion have highlighted the complex nature of cell-cell and cell-extracellular matrix interactions 8. A variety of cell surface molecules such as CD44 glycoproteins, gangliosides and integrins are differentially expressed in astrocytomas. Some, such as the A2B5 ganglioside, are expressed primarily by non-dividing cells that are migrating; others appear somewhat specific for neoplastic astrocytes. Many of the growth factors expressed in astrocytomas, such as FGF, EGF and VEGF, also stimulate migration 8.

Less common molecular changes also occur in grade II astrocytomas. Loss of chromosome 22q, for instance, suggests the presence of a chromosome 22q glioma tumor suppressor gene 9. While the neurofibromatosis 2 (NF2) gene was a likely candidate for this gene, NF2 mutations do not occur in astrocytomas 10 and deletion mapping of chromosome 22q in astrocytomas has suggested a more telomeric locus.

The transition from WHO grade II astrocytoma to WHO grade III anaplastic astrocytoma is accompanied by a marked increase in malignant behavior. While many patients with grade II astrocytomas survive for five or more years, patients with anaplastic astrocytomas often die within two or three years and frequently show transformation to GBM. Histologically, the major differences between grade II and grade III tumors are increased cellularity and the presence of mitotic activity, implying that higher proliferative activity is the hallmark of the progression to anaplastic astrocytoma.

A number of molecular abnormalities have been associated with anaplastic astrocytoma, and recent studies have suggested that most of these abnormalities converge on one critical cell-cycle regulatory complex which includes the p16, cyclin-dependent kinase 4 (cdk4), cyclin D1 and retinoblastoma (Rb) proteins. The simplest schema suggests that p16 inhibits the cdk4/cyclin D1 complex, preventing cdk4 from phosphorylating pRb, and so ensuring that pRb maintains its brake on the cell cycle 11. Individual components in this pathway are altered in up to 50% of anaplastic astrocytomas and in the majority of GBM.

Chromosome 9p loss occurs in approximately 50% of anaplastic astrocytomas and GBMs, with 9p deletions occurring primarily in the region of the CDKN2/p16 (or MTS1) gene, which encodes the p16 protein 12. The frequency of 9p loss increases not only at the transition from astrocytoma to anaplastic astrocytoma, but also at the transition from anaplastic astrocytoma to GBM, implying that the 9p tumor suppressor plays a role in different stages of astrocytoma progression 13. While debate has raged on whether the CDKN2/p16 gene is the primary glioma tumor suppressor gene on chromosome 9p, current evidence does implicate CDKN2/p16. Deletions in primary GBMs almost always involve CDKN2/p16 12,14 and three mutations have been described in primary GBMs with allelic loss of chromosome 9p 15-17. In addition, reduced or absent p16 expression occurs in some malignant gliomas without CDKN2/p16 loss 18, suggesting alternative means, such as hypermethylation 19, of inactivating this gene in GBMs. Moreover, replacement of CDKN2/p16 into GBM cell lines lacking the gene results in growth suppression, but had no effect in cell lines containing the CDKN2/p16 gene 20.

Loss of chromosome 13q occurs in one-third to one-half of high-grade astrocytomas, suggesting the presence of an progression-associated astrocytoma tumor suppressor gene on that chromosome. The 13q14 region containing the RB gene is preferentially targeted by these losses and inactivating mutations of the RB gene occur in primary astrocytomas 21. Overall, analysis of chromosome 13q loss, RB gene mutations and Rb protein expression suggests that the RB gene is inactivated in about 20% of anaplastic astrocytomas and 35% of GBM 21. Interestingly, RB and CDKN2/p16 alterations in primary gliomas are inversely correlated, rarely occurring together in the same tumor 17.

Because amplification of the CDK4 gene and overexpression of cyclin D1 may have similar effects to p16 or pRb inactivation 11, these mechanisms may provide additional alternatives to subvert cell-cycle control and facilitate progression to GBM 14. CDK4, located on chromosome 12q13-14, is amplified in 15% of malignant gliomas 22, although this frequency may be higher among cases without CDKN2/p16 loss, perhaps reaching 50% of GBMs without CDKN2/p16 loss 14. CDK4 amplification and CDKN2/p16 deletions do not occur together in GBM cell lines and some GBM cell lines overexpress cyclin D1 23. On the other hand, in some GBMs and GBM cell lines, CDK4 amplification and cyclin D1 overexpression appear to represent alternative events to CDKN2/p16 deletions, since these genetic changes only rarely occur in the same tumors 14,23. In combination, it is likely that up to 50% of anaplastic astrocytomas and perhaps all GBM have alterations in at least one component of this critical cell-cycle regulatory pathway.

Allelic losses on 19q have been observed in up to 40% of anaplastic astrocytomas and GBMs, indicating a progression-associated glial tumor suppressor gene on chromosome 19q 24. This tumor suppressor gene may be unique to glial tumors 25 and is involved in all three major types of diffuse cerebral gliomas (astrocytomas, oligodendrogliomas, and oligoastrocytomas). This gene maps to a region of chromosome 19q13.3, telomeric to the marker D19S219 and centromeric to the HRC gene 26,27. A number of candidate genes have been isolated from or mapped to this region 28, including the BAX gene, whose product negatively regulates apoptosis with bcl-2, but the tumor suppressor gene remains to be identified.

GBM is the most malignant stage of astrocytoma, with survival times of less than 2 years for most patients. Histologically, these tumors are characterized by dense cellularity, high proliferation indices, endothelial proliferation and focal necrosis. The highly proliferative nature of these lesions is no doubt the result of multiple mitogenic effects. As mentioned above, at least one such effect is deregulation of the p16-cdk4-cyclin D1-pRb pathway of cell-cycle control. The vast majority, if not all, GBM have alterations of this system, whether it be inactivation of p16 or pRb or overexpression of cdk4 or cyclin D1 14,17.

Chromosome 10 loss is a frequent finding in GBM, occurring in 60-95% of GBMs but only rarely in anaplastic astrocytomas 9,29. Attempts to identify this tumor suppressor gene by deletion mapping, however, have been hampered by the observation that, in most cases, the entire chromosome is lost. The gene on the long arm may map to band q25 30-32. On the other hand, there is probably a second tumor suppressor gene on the short arm 29,32, and one study has postulated that a third locus may exist on the long arm, near the centromere 32.

EGFR is a transmembrane receptor tyrosine kinase, whose ligands include EGF and transforming growth factor-a. The EGFR gene is the most frequently amplified oncogene in astrocytic tumors 33, being amplified in approximately 40% of all GBM 29 but in few anaplastic astrocytomas 34. Those GBMs that exhibit EGFR gene amplification have almost always lost genetic material on chromosome 10 29. GBMs with EGFR gene amplification display overexpression of EGFR at both the mRNA and protein levels, suggesting that activation of this growth signal pathway is integral to malignant progression to GBM 34,35. Approximately one-third of those GBM with EGFR gene amplification also have specific EGFR gene rearrangements, which produce truncated molecules similar to the v-erbB oncogene. These truncated receptors are capable of conferring dramatically enhanced tumorigenicity to GBM cells 36. The downstream targets of EGFR activation in GBMs are not well defined, but EGFR is most likely involved in a cascade that facilitates mitogenesis in tumor cells. Less commonly amplified oncogenes include N-myc, gli, PDGF-a receptor, c-myc, myb, K-ras, CDK4 and MDM2, some of which have been discussed above.

As mentioned above, one of the hallmarks of GBM is endothelial proliferation. A host of angiogenic growth factors and their receptors are found in GBMs. For example, VEGF and PDGF are expressed by tumor cells while their receptors, flk-1 and flt-1 for VEGF and the PDGF b-receptor for PDGF, are expressed on endothelial cells. VEGF and its receptors, in particular, appear to play a major role in GBM angiogenesis 37,38. A paracrine mechanism has been suggested in which VEGF is secreted by tumor cells and bound by the VEGF receptors on endothelial cells. Interestingly, VEGF is preferentially upregulated by tumor cells surrounding regions of necrosis, perhaps as a result of necrosis-induced hypoxia, since hypoxia can upregulate VEGF. A link between p53 and tumor angiogenesis has been suggested by the observations that some mutant p53 molecules can enhance VEGF expression 37 and that wild-type p53 regulates the secretion of a glioma-derived angiogenesis inhibitory factor 39.

The assumption that all astrocytomas progress through distinct genetic stages in a linear fashion is most likely an oversimplification. Indeed, it appears as if there are biologic subsets of astrocytomas which may reflect the clinical heterogeneity observed in these tumors (Figure 2). For instance, GBMs can be divided on the basis of molecular genetic analysis. As stated above, loss of chromosome 17p and associated p53 mutation occur in tumors with PDGF-a receptor overexpression (M. NistŽr and colleagues, in preparation), and EGFR gene amplification occurs in tumors with loss of chromosome 10 29. However, EGFR gene amplification almost never occurs in GBMs with loss of chromosome 17p: approximately one-third of GBM have p53/chromosome 17p alterations, one third have EGFR gene amplification, and one third have neither change 40. Significantly, those GBMs with loss of chromosome 17p occur in patients younger than those characterized by EGFR gene amplification 40,41. Recent experimental data also supports this distinction by showing that p53 deficient cells are not transformed when cultured in the presence of EGF, whereas they are transformed in the presence of other growth factors 4. Primary GBMs with p53 mutations may therefore not be expected to acquire EGFR gene amplification, if activation of the EGF-EGFR system does not produce a growth advantage in such cells.

The genetic pathway involving 17p loss may involve progression from a lower-grade astrocytic lesion, since loss of chromosome 17p, p53 mutation and PDGF-a receptor overexpression occur as commonly in lower-grade astrocytomas as in higher-grade astrocytomas (see above). On the other hand, those GBM with EGFR amplification may arise either de novo or rapidly from a pre-existing tumor, without a clinically-evident, preceding lower-grade astrocytoma. Interestingly, younger age at initial diagnosis has been an important prognostic parameter among patients with GBM, with younger patients faring better than older patients 42. The predominance of tumors with 17p loss in a younger population of astrocytoma patients may therefore reflect the age-based difference in prognosis. Indeed, in a few studies, patients with p53 mutations had somewhat better prognoses than those without p53 mutations 6,43. In addition, it is possible that those patients with anaplastic gliomas and a previous history of lower-grade glioma that do better clinically ("dedifferentiated GBM") 44 are the same subset as those GBMs with 17p loss. Finally, those GBMs with EGFR gene amplification appear to recur more quickly than those GBM without EGFR gene amplification 45, further suggesting that tumors with EGFR gene amplification are more rapidly progressing lesions. The data suggest that genetic analysis may begin to explain the clinical observations concerning age differences in astrocytic tumors.

Other gliomas

Other astrocytomas

Brain stem gliomas are a form of pediatric diffuse, fibrillary astrocytoma that often follow a malignant course. Brain stem GBMs share genetic features with those adult GBMs that affect younger patients: frequent p53 gene and chromosome 17p alterations without EGFR gene amplification 46. Since brain stem gliomas are predominantly tumors of childhood, this may suggest a common oncogenic pathway for those diffuse, fibrillary astrocytic tumors that affect younger patients, regardless of their anatomic location. On the other hand, there is considerable variation in the transcription levels of particular growth factors and proto-oncogenes in different normal glial cell populations 47 and certain astrocytes may display differential susceptibilities at different ages to the same genetic alterations.

Pleomorphic xanthoastrocytoma (PXA) is a superficial, low-grade astrocytic tumor that predominantly affects young adults. While these tumors have a bizarre histological appearance, they are typically slow-growing tumors that may be amenable to surgical cure. Some PXAs, however, may recur as GBM. PXAs may have p53 mutations, but the few documented mutations have been somewhat different from those usually found in diffuse, fibrillary astrocytomas 48. EGFR gene amplification does not occur in PXAs, but GBMs that arise from PXAs may display EGFR gene amplfication. On the other hand, allelic losses of chromosomes 9, 10 and 19q are not observed in PXA. Thus, the genetic events that underly PXA formation and progression probably differ from those involved in diffuse astrocytoma tumorigenesis 48.

Pilocytic astrocytoma is the most common astrocytic tumor of childhood and differs clinically and histopathologically from the diffuse, fibrillary astrocytoma that affects adults. Pilocytic astrocytomas do not have the same genomic alterations as diffuse, fibrillary astrocytomas. Because pilocytic astrocytomas frequently affect patients with neurofibromatosis 1 (NF1) and the NF1 gene maps to chromosome 17q11.2, pilocytic astrocytomas may be expected to show allelic loss of chromosome 17q. In fact, allelic loss occurs on chromosome 17q in one-quarter of cases 49. These data suggest the presence of a tumor suppressor gene on 17q that is associated with pilocytic astrocytomas. While the logical candidate for this gene is the NF1 tumor suppressor gene, detailed mutational analysis of the NF1 gene in pilocytic tumors has not yet been performed.

Subependymal giant cell astrocytomas (SEGA) are periventricular, low-grade astrocytic tumors that are usually associated with tuberous sclerosis (TS), and are histologically identical to the so-called "candle-gutterings" that line the ventricles of TS patients. Similar to the other tumorous lesions in TS, these are slowly-growing and may be more akin to hamartomas than true neoplasms. The association of SEGA with TS leads to the prediction that the TS genes may be involved in SEGA formation. Linkage studies have identified the location of two TS genes, one on chromosome 9q and another on chromosome 16p. Loss of heterozygosity studies have shown allelic loss of chromosome 9q and 16p loci in some SEGAs, suggesting that the TS genes act as tumor suppressors 50,51. However, confirmation of the role of the TS genes in glial tumorigenesis must await mutational analysis of these genes in cases of SEGA.

Desmoplastic cerebral astrocytoma of infancy (DCAI) and desmoplastic infantile ganglioglioma (DIGG) are large, superficial, usually cystic, benign astrocytomas that affect children in the first year or two of life. Allelic loss of chromosomes 10 or 17 is not common in these lesions 52. In adult gangliogliomas, another benign form of astrocytic glioma, EGFR gene amplification or allelic loss on chromosomes 10, 13q, 17p, 19q and 22q has not been detected (A von Deimling and DNL, unpublished data).

Oligodendrogliomas and oligo-astrocytomas

Oligodendrogliomas and oligoastrocytomas (mixed gliomas) are diffuse, usually cerebral tumors that are clinically and biologically most closely related to the diffuse, fibrillary astrocytomas. The tumors, however, are far less common than astrocytomas and have generally better prognoses than the diffuse astrocytomas; patients with WHO grade II oligodendrogliomas, for instance, may have mean survival times of 10 years. In addition, oligodendroglial tumors appear to be differentially chemosensitive 53, when compared with the diffuse astrocytomas.

Allelic losses in oligodendrogliomas and oligoastrocytomas occur preferentially on chromosomes 1p and 19q, affecting 40-80% of these tumor types 24,54,55. Because of the frequent loss of these loci in low-grade as well as anaplastic oligodendrogliomas and oligoastrocytomas, the 1p and 19q tumour suppressors are probably important early in oligodendroglial tumorigenesis. Mapping of the chromosome 19q locus has demonstrated that the gene resides in the same vicinity as the astrocytoma gene, between APOC2 and HRC, and is likely the same gene 26,27. Similar mapping of chromosome 1p has implicated the telomeric region of 1p32-36 55. Interestingly, chromosome 1p and 19q losses are closely associated; oligodendroglial tumours with 1p loss typically also have loss of 19q, suggesting that these two putative tumour suppressor genes may be involved in biologically distinct pathways 54,55. In fact, microdissection of the oligodendroglial and astrocytic components in oligoastrocytomas has shown that, despite the histological differences, the molecular changes are identical in these two components 55. Oligoastrocytomas in particular may also suffer allelic losses of chromosome 17p 54, although these losses are not associated with p53 mutations 54,56, perhaps implying a second chromosome 17p glioma gene. Oncogene amplification has only rarely been noted in oligodendroglial tumours 33,54.

Oligodendrogliomas and oligoastrocytomas may progress, either to WHO grade III anaplastic oligodendroglioma or anaplastic oligoastrocytoma, or to WHO grade IV GBM. Thus, the genetic changes that lead to oligodendroglial tumors constitute yet another pathway to GBM (Figure 2). Anaplastic oligodendrogliomas and oligoastrocytomas may display allelic losses of chromosomes 9p and 10 54 and allelic loss of chromosome 10 may be a common finding in high-grade malignant gliomas, whether they are astrocytic or oligodendroglial in original lineage 9. At least one aggressive oligodendroglioma with allelic loss of chromosome 10 rapidly recurred as a GBM 57.

Ependymomas and choroid plexus tumors

Ependymomas are a clinically diverse group of gliomas that vary from aggressive intraventricular tumors of children to benign spinal cord tumors in adults. Chromosome 22q loss is common in ependymomas 58,59. A candidate glioma tumor suppressor gene on chromosome 22q was the neurofibromatosis 2 (NF2) gene 60, since NF2 patients have a higher incidence of gliomas, particularly ependymomas, in addition to schwannomas and meningiomas. Analysis of the NF2 gene in ependymomas, however, has revealed only a single mutation to date, in an ependymoma that had lost the remaining wild-type allele 10. The paucity of NF2 mutations suggests that another, as yet unidentified, chromosome 22q gene will probably be a more integral ependymoma locus. The p53 gene is not mutated in ependymomas 56 or in the malignant transformation of ependymomas to anaplastic ependymoma (S. Cortez and DNL, unpublished data). Transitions of ependymoma to GBM are rare, and have not been studied by molecular genetic techniques.

Choroid plexus tumors are also a varied group of tumors that preferentially occur in the ventricular system, ranging from aggressive supratentorial intraventricular tumors of children to benign cerebellopontine angle tumors of adults. Choroid plexus tumors have been reported occasionally in patients with Li-Fraumeni syndrome and von Hippel-Lindau (VHL) disease (as well as in Aicardi syndrome, which does not predispose to cancer), raising the possibility that the p53 gene on chromosome 17p, responsible for Li-Fraumeni syndrome, or the VHL gene on chromosome 3p is involved in choroid plexus neoplasia. Loss of chromosome 3p sequences have been found in one choroid plexus tumor, suggesting involvement of the VHL gene 61.

Oncogenic viruses may cause human cancer, particularly those viruses whose products interfere with tumor suppressor gene functions, such as the human papillomaviruses implicated in cervical carcinoma. One study has identified sequences similar to SV40 virus, an oncogenic virus that has the ability to inactivate both the Rb and p53 proteins, in human ependymomas and choroid plexus papillomas 62. This observation raised considerable excitement since SV40 has been implicated as an oncogenic factor in transgenic models of choroid plexus neoplasia 63. However, SV40-like sequences have not been found in other choroid plexus papillomas or ependymomas and the role of oncogenic viruses in these tumors remains undefined.

Medulloblastomas

Medulloblastomas are highly malignant, primitive tumors that arise in the posterior fossa, primarily in children. One third to one half of all medulloblastomas have an isochromosome 17q on cytogenetic analysis 64, and corresponding allelic loss of chromosome 17p has been noted on molecular genetic analysis 65,66. Initially, this suggested involvement of the p53 gene, but p53 mutations have proven to be very rare in medulloblastomas 56,67. Allelic losses occur preferentially at regions of chromosome 17p that are telomeric to the p53 locus 66,68, implying the presence of a second, more distal chromosome 17p tumor suppressor gene. Allelic losses of chromosome 6q, 11 and 16q have also been noted frequently in these tumors 65. Deletions of the CDKN2/p16 gene, which are common in many tumors, do not occur in medulloblastomas 69. Oncogene amplification has not been found frequently in medulloblastomas; only c-myc is amplified in significant numbers of cases and this change appears more common in medulloblastoma cell lines than in primary tumors 64.

Meningiomas

Meningiomas are common intracranial tumors that arise in the meninges and compress the underlying brain. Meningiomas are usually benign, but some "atypical" meningiomas may recur locally, and some meningiomas are frankly malignant and may invade the brain or metastasize. Monosomy 22 is common in meningiomas, pointing to a chromosome 22q meningioma suppressor gene. Indeed, the neurofibromatosis 2 (NF2) gene on chromosome 22q is frequently mutated in meningiomas, clearly implicating it in meningothelial tumorigenesis 70-72. In sporadic meningiomas, both chromosome 22q allelic loss and NF2 gene mutations are more common in fibroblastic and transitional subtypes than in meningothelial forms 71. As in schwannomas (see below), NF2 gene alterations result predominantly in immediate truncation, splicing abnormalities or altered reading frames, producing grossly truncated proteins. Interestingly, NF2 gene mutations in meningiomas cluster in the moesin-ezrin-radixin homology domain in the first half of the coding sequence 71,72.

Approximately 40% of meningiomas have neither NF2 gene mutations nor allelic loss of chromosome 22q. For these tumors, it is likely that a second meningioma tumor suppressor gene is involved. This putative second gene is probably not on chromosome 22q, since NF2 gene mutations in meningiomas correlate fairly closely with chromosome 22q loss. Nonetheless, a few meningiomas have been described with loss of portions of chromosome 22q that do not include the NF2 gene, suggesting the possibility of a second meningioma locus on chromosome 22 73. One candidate gene from this second chromosome 22q region is BAM22, a member of the b-adaptin gene family, which may be inactivated in some sporadic meningiomas 74. Another candidate is the MN1 gene, which is disrupted by a translocation in a meningioma 75. Furthermore, a family with multiple meningiomas but without vestibular schwannomas does not show linkage to the NF2 locus on chromosome 22q, suggesting yet another meningioma predisposition gene 76. Allelic losses in meningiomas have been noted on a variety of other chromosomes, including 1p3, 3p, 5p, 5q, 11, 13 and 17p 77.

Atypical and malignant meningiomas are not as common as benign meningiomas. Genetic analyses of malignant meningiomas have revealed preferential losses of chromosomes 1p, 10 and 14q 78,79. In one anaplastic meningioma, molecular genetic analysis of three morphologically distinct regions revealed loss of heterozygosity for chromosomes 1p and 22q in all regions, but for chromosomes 17p and 9q only in the malignant region 78. Chromosome 10 loss, in particular, has been associated with those meningiomas with morphological features of malignancy, rather than those meningiomas which are designated as malignant on the basis of brain invasion alone 79.

Schwannomas

Schwannomas are benign tumors that arise on peripheral nerves. Schwannomas may arise on cranial nerves, particularly the vestibular portion of the eighth cranial nerve (vestibular schwannomas, acoustic neuromas) where they present as cerebellopontine angle masses. NF2 patients are defined by the presence of bilateral vestibular schwannomas 80, although unilateral vestibular schwannomas are common in the general population as well. Therefore, like meningiomas, schwannomas occur frequently in NF2 patients, have frequent loss of chromosome 22q, and harbor NF2 gene mutations in at least 50% of cases, in vestibular tumors as well as schwannomas from other sites 80. A detailed analysis of a cohort of 60 vestibular schwannomas using combined screening methods has revealed mutations in the vast majority of schwannomas (L.B. Jacoby, personal communication). The majority of the somatic changes are small deletions or insertions that create either frameshifts and premature stop codons or altered splicing. Inactivating mutations are relatively evenly distributed across the first 15 exons with no outstanding hot spots. One study has confirmed such inactivation at the protein level, showing loss of merlin expression by immunohistochemistry in schwannomas 81. Thus, inactivation of NF2 is a common feature underlying both inherited and sporadic forms of schwannoma.

Miscellaneous tumors

Hemangioblastomas are tumors of uncertain origin that are composed of endothelial cells, pericytes and so-called stromal cells. These benign tumors most frequently occur in the cerebellum and spinal cord of young adults. Multiple hemangioblastomas are characteristic of von Hippel-Lindau disease (VHL), an inherited tumor syndrome in which patients have a tendency to develop tumors, particularly hemangioblastomas, retinal angiomas, renal cell carcinomas and pheochromocytomas 82. Because of the association of VHL and hemangioblastomas, studies of hemangioblastomas have focussed on the VHL gene.

The VHL gene was originally mapped by linkage analysis to chromosome 3p 83 and allelic loss studies of hemangioblastomas are consistent with this location 84. The gene responsible for VHL has been identified 85 and is mutated in sporadic hemangioblastomas 86, suggesting that the VHL gene acts as a classical tumor suppressor and is involved in both sporadic hemangioblastomas and familial tumors.

Hemangiopericytomas (HPCs) are dural tumors which may display locally aggressive behavior and may metastasize. The histogenesis of dural-based hemangiopericytoma (HPC) has long been debated, with some authors classifying it as a distinct entity and others classifying it as a subtype of meningioma. In constrast, meningiomas contain frequent mutations of the NF2 gene, while HPCs do not, suggesting that HPC is genetically distinct from meningioma 87. In addition, homozygous deletions of the CDKN2/p16 gene are common in HPCs, and suggest that alterations of the p16-mediated cell-cycle regulatory pathway may underlie the malignant potential of some HPCs 88. Rearrangements of chromosome 12q13 are common in peripherally located, soft tissue HPCs and have been reported in meningeal HPCs, implying that an oncogene or tumor suppressor gene at this locus is important in HPC formation. A number of oncogenes reside in this region, including MDM2, CDK4 and CHOP/GADD153. Interestingly, CHOP/GADD153 amplification and constitutive expression has been noted in at least one soft tissue HPC 89.

Neurological tumor syndromes

Hereditary neurological tumor syndromes, in which patients are at risk for developing multiple nervous system tumors, have provided important clues to the genetic basis of brain tumors. These tumor syndromes provide unique insights into tumor suppressor genes. For instance, the hereditary retinoblastoma syndrome, which provided much of the impetus for current tumor suppressor gene research, results from a mutation in both Rb tumor suppressor alleles 90. In the case of familial retinoblastoma, the patient inherits one mutant, inactive copy of the retinoblastoma gene and thus carries a "germline" mutation in every cell 91, which is unveiled when the second copy of the gene is inactivated, either by mutation or by loss of a portion of chromosome 13q. A patient with sporadic retinoblastomas does not carry the germline mutation and must acquire both inactivating "hits" in the same cell during his or her life. Therefore, a familial retinoblastoma patient, in which every cell contains a mutation, is much more likely to develop a second retinoblastoma than a sporadic retinoblastoma patient 92. This same paradigm appears to hold true for most, but not all 93, of the hereditary tumor syndromes.

Neurological tumor syndromes include the so-called neurocutaneous syndromes, such as neurofibromatosis 1 (NF1), neurofibromatosis 2 (NF2), tuberous sclerosis (TS), and von Hippel-Lindau disease (VHL), and other tumor conditions such as Li-Fraumeni, Turcot's, Gorlin's and Cowden's syndromes. Each syndrome is accompanied by a characteristic panoply of tumors, both neurological and non-neurological. A catalogue of only the gliomas would feature optic nerve gliomas and other astrocytomas in NF1; ependymomas and astrocytomas in NF2 ; subependymal giant cell astrocytomas in TS; and various malignant gliomas in Li-Fraumeni syndrome, Turcot's syndrome and the hereditary glioma pedigrees. Linkage studies have provided powerful means for tracking down the genes associated with these tumor syndromes, and have assigned the NF1 gene to chromosome 17q; the NF2 gene to chromosome 22q; the TS genes to chromosome 9q and 16p; Turcot's genes to chromosome 5q (APC gene) and to the DNA mismatch repair genes on various chromosomes. The NF1 gene codes for a guanosine triphosphatase-activating protein (GAP) termed neurofibromin. Neurofibromin interacts with the p21 product of the ras oncogene and is most likely important in growth factor-mediated signal transduction. The NF2 gene codes for a protein, termed merlin, that is homologous with moesin, ezrin and radixin, suggesting that the protein functions to link the cytoskeleton and cell surface. One of the TS genes, TSC2 on chromosome 16p, has also recently been identified. Like NF1, TSC2 encodes a GAP protein. While TSC2 GAP is not homologous with neurofibromin, these proteins may nonetheless be involved in similar cellular pathways . For the Li-Fraumeni syndrome, mutational analyses have implicated the p53 gene on 17p. The other neurological tumor syndromes, their chromosomal locations, known genes and putative functions are catalogued in Table 2 and have been the subjects of recent reviews 80,82,93-95.

Summary

Human brain tumors have molecular alterations that are characteristic of each type of tumor and of most stages of progression. For instance, the formation of grade II astrocytoma involves inactivation of the p53 tumor suppressor gene on chromosome 17p, as well as PDGF overexpression, loss of a putative tumor suppressor gene on chromosome 22q, and the expression of various molecules that facilitate tumor invasion. The transition from astrocytoma to anaplastic astrocytoma is associated with alterations of the critical cell-cycle regulatory pathway that includes p16, cdk4, cyclin D1 and Rb, as well as a putative tumor suppressor gene on chromosome 19q. Finally, progression to GBM involves loss of a at least one putative tumor suppressor gene on chromosome 10, amplification of the EGFR gene and the expression of angiogenic factors such as VEGF. Furthermore, molecular genetic analysis has been used to distinguish subsets of astrocytomas. For instance, one type of GBM, characterized by p53 gene mutations, is more common in younger patients and may be associated with slower progression from lower-grade astrocytoma; another type of GBM, characterized by EGFR gene amplification, is more common in older patients and may be associated with more rapid progression or "de novo" growth.

For the less common gliomas and for other primary tumors such as medulloblastomas, molecular genetic studies have defined only isolated genetic alterations. For meningiomas and schwannomas, the NF2 gene has been clearly implicated, although other genetic alterations must underlie the formation of some meningiomas as well. For those tumors associated with hereditary tumor syndromes, such as the SEGAs in TS and the hemangioblastomas in VHL, the same genes appear responsible for the syndromes when mutated in the germline, and for sporadic tumors when mutated on a somatic basis. At the present time, however, these molecular data are incomplete. Once the molecular pathways are completely understood, such knowledge will no doubt contribute to the development of more effective therapies for many of these tumors .

References

1. Louis DN. The p53 gene and protein in human brain tumors. J Neuropathol Exp Neurol 1994;53:11.

2. Sidransky D, Mikkelsen T, Schwechheimer K, Rosenblum M, Cavenee W, Vogelstein B. Clonal expansion of p53 mutant cells is associated with brain tumor progression. Nature 1992;355:846.

3. Yahanda AM, Bruner JM, Donehower LA, Morrison RS. Astrocytes derived from p53-deficient mice provide a multistep in vitro model for development of malignant gliomas. Molec Cell Biol 1995;15:4249.

4. Bogler O, Huang H-JS, Cavenee WK. Loss of wild-type p53 bestows a growth advantage on primary cortical astrocytes and facilitates their in vitro transformation. Cancer Res 1995;55:2746.

5. Livingstone LR, White A, Sprouse J, Livanos E, Jacks T, Tlsty TD. Altered cell cycle arrest and gene amplification potential accompany loss of wild-type p53. Cell 1992;70:923.

6. van Meyel DJ, Ramsay DA, Casson AG, Keeney M, Chambers AF, Cairncross JG. p53 mutation, expression, and DNA ploidy in evolving gliomas: evidence for two pathways of progression. J Natl Canc Inst 1994;86:1011.

7. Hermanson M, Funa K, Hartman M, Claesson-Welsh L, Heldin C-H, Westermark B, Nister M. Platelet-derived growth factor and its receptors in human glioma tissue: expression of messenger RNA and protein suggests the presence of autocrine and paracrine loops. Cancer Res 1992;52:3213.

8. Pilkington GJ. Tumour cell migration in the central nervous system. Brain Pathol 1994;4:157.

9. James CD, Carlblom E, Dumanski JP, Hansen M, Nordenskjold M, Collins VP, Cavenee WK. Clonal genomic alterations in glioma malignancy stages. Cancer Res 1988;48:5546.

10. Rubio M-P, Correa KM, Ramesh V, MacCollin MM, Jacoby LB, von Deimling A, Gusella JF, Louis DN. Analysis of the neurofibromatosis 2 (NF2) gene in human ependymomas and astrocytomas. Cancer Res 1994;54:45.

11. Serrano M, Hannon GJ, Beach D. A new regulatory motif in cell-cycle control causing specific inhibition of cyclin D/CDK4. Nature 1993;366:704.

12. Jen J, Harper W, Bigner SH, Bigner DD, Papadopoulos N, Markowitz S, Willson JKV, Kinzler KW, Vogelstein B. Deletion of p16 and p15 in brain tumors. Cancer Res 1994;54:6353.

13. von Deimling A, Bender B, Jahnke R, Waha A, Kraus J, Albrecht S, Wellenreuther R, Fassbender F, Nagel J, Menon AG, Louis DN, Lenartz D, Schramm J, Wiestler OD. Loci associated with malignant progression in astrocytomas: a candidate on chromosome 19q. Cancer Res 1994;54:1397.

14. Schmidt EE, Ichimura K, Reifenberger G, Collins VP. CDKN2 (p16/MTS1) gene deletion or CDK4 amplification occurs in the majority of glioblastomas. Cancer Res 1994;54:6321.

15. Ueki K, Rubio M-P, Ramesh V, Correa KM, Rutter JL, von Deimling A, Buckler AJ, Gusella JF, Louis DN. MTS1/CDKN2 gene mutations are rare in primary human astrocytomas with allelic loss of chromosome 9p. Hum Molec Genet 1994;3:1841.

16. Moulton T, Samara G, Chung W, Yuan L, Desai R, Sisti M, Bruce J, Tycko B. MTS1/p16/CDKN2 lesions in primary glioblastoma multiforme. Am J Pathol 1995;146:613.

17. Ueki K, Ono Y, Henson JW, von Deimling A, Louis DN. CDKN2/p16 or RB alterations occur in the majority of glioblastomas and are inversely correlated (submitted for publication).

18. Nishikawa R, Furnari F, Lin H, Arap W, Berger MS, Cavenee WK, Huang H-JS. Loss of p16INK4 expression is frequent in high grade gliomas. Cancer Res 1995;55:1941.

19. Merlo A, Herman JG, Mao L, Lee DJ, Gabrielson E, Burger PC, Baylin SB, Sidransky D. 5'CpG island methylation is associated with transcriptional silencing of the tumor suppressor p16/CDKN2/MTS1. Nature Med 1995;1:686.

20. Arap W, Nishikawa R, Furnari FB, Cavenee WK, Huang H-JS. Replacement of the p16/CDKN2 gene suppresses human glioma cell growth. Cancer Res 1995;55:1351.

21. Henson JW, Schnitker BL, Correa KM, von Deimling A, Fassbender F, Xu H-J, Benedict WF, Yandell DW, Louis DN. The retinoblastoma susceptibility (Rb) gene is involved in the malignant progression of human astrocytomas. Ann Neurol 1994;36:714.

22. Reifenberger G, Reifenberger J, Ichimura K, Melzter PS, Collins VP. Amplification of multiple genes from chromosomal region 12q13-14 in human malignant gliomas: preliminary mapping of the amplicons shows preferential involvement of CDK4, SAS, and MDM2. Cancer Res 1994;54:4299.

23. He J, Allen JR, Collins VP, Allalunis-Turner MJ, Godbout R, Day RSI, James CD. CDK4 amplification is an alternative mechanism to p16 gene homozygous deletion in glioma cell lines. Cancer Res 1994;54:5804.

24. von Deimling A, Louis DN, von Ammon K, Petersen I, Wiestler OD, Seizinger BR. Evidence for a tumor suppressor gene on chromosome 19q associated with human astrocytomas, oligodendrogliomas and mixed gliomas. Cancer Res 1992;52:4277.

25. Seizinger BR, Klinger HP, Junien C, Nakamura Y, Le Beau M, Cavenee W, Emanuel B, Ponder B, Naylor S, Mitelman F, Louis DN, Menon A, Newsham I, Decker J, Kaelbing I, Henry I, von Deimling A. Report of the committee on chromosome and gene loss in human neoplasia. Cytogenet Cell Genet 1991;58:1080.

26. Yong WH, Chou D, Ueki K, Harsh GR, von Deimling A, Gusella JF, Mohrenweiser HW, Louis DN. Chromosome 19q deletions in human gliomas overlap telomeric to D19S219 and may target a 425 kb region centromeric to D19S112. J Neuropathol Exp Neurol 1995;(in press).

27. Rubio M-P, Correa KM, Ueki K, Mohrenweiser HW, Gusella JF, von Deimling A, Louis DN. The putative glioma tumor suppressor gene on chromosome 19q maps between APOC2 and HRC. Cancer Res 1994;54:4760.

28. Yong WH, Ueki K, Chou D, Reeves SA, von Deimling A, Gusella JF, Mohrenweiser HW, Buckler AJ, Louis DN. Cloning of a highly conserved human protein serine threonine phosphatase gene from the glioma candidate region on chromosome 19q13.3. Genomics 1995;(in press).

29. von Deimling A, Louis DN, von Ammon K, Petersen I, Hoell T, Chung RY, Martuza R, Schoenfeld D, Yasargil MG, Wiestler OD, Seizinger BR. Association of epidermal growth factor receptor gene amplification with loss of chromosome 10 in human glioblastoma multiforme. J Neurosurg 1992;77:295.

30. Fults D, Pedone C. Deletion mapping of the long arm of chromosome 10 in glioblastoma multiforme. Genes Chromosom Cancer 1993;7:173.

31. Rasheed BKA, Fuller GN, Friedman AH, Bigner DD, Bigner SH. Loss of heterozygosity for 10q loci in human gliomas. Genes Chromosom Cancer 1992;5:75.

32. Karlbom AE, James CD, Boethius J, Cavenee WK, Collins VP, Nordenskjold M, Larsson C. Loss of heterozygosity in malignant gliomas involves at least three distinct regions on chromosome 10. Hum Genet 1993;92:169.

33. Fuller GN, Bigner SH. Amplified cellular oncogenes in neoplasms of the human central nervous system. Mutat Res 1992;276:299.

34. Ekstrand AJ, James CD, Cavenee WK, Seliger B, Petterson RF, Collins VP. Genes for epidermal growth factor receptor, transforming growth factor a, and epidermal growth factor and their expression in human gliomas in vivo. Cancer Res 1991;51:2164.

35. Wong AJ, Bigner SH, Bigner DD, Kinzler KW, Hamilton SR, Vogelstein B. Increased expression of the epidermal growth factor receptor gene in malignant gliomas is invariably associated with gene amplification. Proc Natl Acad Sci USA 1987;84:6899.

36. Nishikawa R, Ji XD, Harmon RC, Lazar CS, Gill GN, Cavenee WK, Huang H-JS. A mutant epidermal growth factor receptor common in human glioma confers enhanced tumorigenicity. Proc Natl Acad Sci USA 1994;91:7727.

37. Plate KH, Breier G, Risau W. Molecular mechanisms of developmental and tumor angiogenesis. Brain Pathol 1994;4:207.

38. Plate KH, Breier G, Weich HA, Risau W. Vascular endothelial growth factor is a potential tumour angiogenesis factor in human gliomas in vivo. Nature 1992;359:845.

39. Van Meir EG, Polverini PJ, Chazin VR, Huang H-JS, de Tribolet N, Cavenee WK. Release of an inhibitor of angiogenesis upon induction of wild type p53 expression in glioblastoma cells. 1994;8:171.

40. von Deimling A, von Ammon K, Schoenfeld D, Wiestler OD, Seizinger BR, Louis DN. Subsets of glioblastoma multiforme defined by molecular genetic analysis. Brain Pathol 1993;3:19.

41. Rasheed BKA, McLendon RE, Herndon JE, Friedman HS, Friedman AH, Bigner DD, Bigner SH. Alterations of the TP53 gene in human gliomas. Cancer Res 1994;54:1324.

42. Burger PC, Green SB. Patient age, histologic features, and length of survival in patients with glioblastoma multiforme. Cancer 1987;59:1617.

43. Chung RY, Whaley J, Kley N, Anderson K, Louis DN, Menon A, Hettlich C, Freiman R, Hedley-Whyte ET, Martuza R, Jenkins R, Yandell D, Seizinger BR. TP53 mutation and chromosome 17p deletion in human astrocytomas. Genes Chromosom Cancer 1991;3:323.

44. Winger MJ, Macdonald DR, Cairncross JG. Supratentorial anaplastic gliomas in adults. J Neurosurg&127; 1989;71:487.

45. Schlegel J, Merdes A, Stumm G, Albert FK, Forsting M, Hynes N, Kiessling M. Amplification of the epidermal-growth-factor-recptor gene correlates with different growth behaviour in human glioblastoma. Int J Cancer 1994;56:72.

46. Louis DN, Rubio M-P, Correa K, Gusella JF, von Deimling A. Molecular genetic alterations in pediatric brain stem gliomas. Application of PCR techniques to small and archival brain tumor specimens. J Neuropathol Exp Neurol 1993;52:507.

47. Gillaspy GE, Mapstone TB, Samols D, Goldthwait DA. Transcriptional patterns of growth factors and proto-oncogenes in human glioblastomas and normal glial cells. Cancer Lett 1992;65:55.

48. Paulus W, Lisle DK, Tonn JC, Roggendorf W, Wolf HK, Reeves SA, Louis DN. Molecular genetic alterations in pleomorphic xanthoastrocytoma. Acta Neuropathol 1996;

49. von Deimling A, Louis DN, Menon AG, Ellison D, Wiestler OD, Seizinger BR. Allelic loss on the long arm of chromosome 17 in pilocytic astrocytoma. Acta Neuropathol 1993;86:81.

50. Green AJ, Smith M, Yates JRW. Loss of heterozygosity on chromosome 16p13.3 in hamartomas from tuberous sclerosis patients. Nature Genet 1994;6:193.

51. Green AJ, Johnson PH, Yates JR. The tuberous sclerosis gene on chromosome 9q34 acts as a growth suppressor. Human Molec Genet 1994;3:1833.

52. Louis DN, von Deimling A, Dickersin G, Dooling E, Seizinger BR. Desmoplastic cerebral astrocytoma of infancy: a histopathological, immunohistochemical, ultrastructural and molecular genetic study. Hum Pathol 1992;23:1402.

53. Cairncross JG, Macdonald DR, Ramsay DA. Aggressive oligodendroglioma: a chemosensitive tumor. Neurosurg 1992;31:78.

54. Reifenberger J, Reifenberger G, Liu L, James CD, Wechsler W, Collins VP. Molecular genetic analysis of oligodendroglial tumors shows preferential allelic deletions on 19q and 1p. Am J Pathol 1994;145:1175.

55. Kraus JA, Koopman J, Kaskel P, Maintz D, Brandner S, Louis DN, Wiestler OD, von Deimling A. Shared allelic losses on chromosomes 1p and 19q suggest a common origin of oligodendroglioma and oligoastrocytoma. J Neuropathol Exp Neurol 1995;54:91.

56. Ohgaki H, Eibl RH, Wiestler OD, Yasargil MG, Newcomb EW, Kleihues P. p53 mutations in nonastrocytic human brain tumors. Cancer Res 1991;51:6202.

57. Wu JK, Folkerth RD, Ye Z, Darras BT. Aggressive oligodendroglioma predicted by chromosome 10 restriction fragment length polymorphism analysis. Case study. J Neuro-Onc 1993;15:29.

58. Ransom DT, Ritland SR, Kimmel PJ, Moertel CA, Dahl RJ, Scheithauer RB, Kelly PJ, Jenkins RB. Cytogenetic and loss of heterozygosity studies in ependymomas, pilocytic astrocytomas, and oligodendrogliomas. Genes Chromosom Cancer 1992;5:348.

59. James CD, He J, Carlbom E, Mikkelsen T, Ridderheim P-A, Cavenee WK, Collins VP. Loss of genetic information in central nervous system tumors common to children and young adults. Genes, Chromosomes & Cancer 1990;2:94.

60. Trofatter JA, MacCollin MM, Rutter JL, Murrell JR, Duyao MP, Parry DM, Eldridge R, Kley N, Menon AG, Pulaski K, Haase VH, Ambrose CM, Bove C, Haines JL, Martuza RL, MacDonald ME, Seizinger BR, Short MP, Buckler AJ, Gusella JF. The neurofibromatosis 2 tumor suppressor gene encodes a novel moesin-ezrin-radixin-like protein. Cell 1993;72:791.

61. Blamires TL, Maher ER. Choroid plexus papilloma. A new presentation of von Hippel-Lindau (VHL) disease. Eye 1992;

62. Bergsagel DJ, Finegold MJ, Butel JS, Kupsky WJ, Garcea RL. DNA sequences similar to those of simian virus 40 in ependymomas and choroid plexus tumors of childhood. N Engl J Med 1992;326:988.

63. Van Dyke TA. Tumors of the choroid plexus. In: Ed. Molecular Genetics of Nervous System Tumors. New York: Wiley-Liss, 1993:287

64. Bigner SH, Vogelstein B. Cytogenetics and molecular genetics of malignant gliomas and medulloblastoma. Brain Pathol 1990;1:12.

65. Thomas GA, Raffel C. Loss of heterozygosity on 6q, 16q, and 17p in human central nervous system primitive neuroectodermal tumors. Cancer Res 1991;51:639.

66. Cogen PH, Daneshvar L, Metzger AK, Edwards MSB. Deletion mapping of the medulloblastoma locus on chromosome 17p. Genomics 1990;8:279.

67. Saylors RL, Sidransky D, Friedman HS, Bigner SH, Bigner DD, Vogelstein B, Brodeur GM. Infrequent p53 gene mutations in medulloblastomas. Cancer Res 1991;51:4721.

68. Biegel JA, Burk CD, Barr FG, Emanuel BS. Evidence for a 17p tumor related locus distinct from p53 in pediatric primitive neuroectodermal tumors. Cancer Res 1992;52:3391.

69. Raffel C, Ueki K, Harsh GR, Louis DN. The multiple tumor suppressor 1 / cyclin dependent kinase inhibitor 2 gene (MTS1/CDKN2) in human central nervous system primitive neuroectodermal tumor. Neurosurgery 1995;(in press):

70. Ruttledge MH, Sarrazin J, Rangaratnam S, Phelan CM, Twist E, Merel P, Delattre O, Thomas G, Nordenskjold M, Collins VP, Dumanski JP, Rouleau GA. Evidence for the complete inactivation of the NF2 gene in the majority of sporadic meningiomas. Nature Genet 1994;6:180.

71. Wellenreuther R, Kraus JA, Lenartz D, Menon AG, Schramm J, Louis DN, Ramesh V, Gusella JF, Wiestler OD, von Deimling A. Analysis of the neurofibromatosis 2 gene reveals molecular variants of meningioma. Am J Pathol 1995;146:827.

72. Lekanne Deprez RH, Bianchi AB, Groen NA, Seizinger BR, Hagemeijer A, van Drunen E, Bootsma D, Koper JW, Avezaat CJJ, Kley N, Zwarthoff EC. Frequent NF2 gene transcript mutations in sporadic meningiomas and vestibular schwannomas. Am J Hum Genet 1994;54:1022.

73. Ruttledge MH, Xie YG, Han FY, Peyrard M, Collins VP, Nordenskjold M, Dumanski JP. Deletions on chromosome 22 in sporadic meningioma. Genes Chromosom Cancer 1994;10:122.

74. Peyrard M, Fransson I, Xie YG, Han FY, Ruttledge MH, Swahn S, Collins JE, Dunham I, Collins VP, Dumanski JP. Characterization of a new member of the human beta-adaptin gene family from chromosome 22q12, a candidate meningioma gene. Hum Molec Genet 1994;3:1393.

75. Lekanne Deprez RH, Riegman PHJ, Groen NA, Warringa UL, van Biezen NA, Molijn AC, Bootsma D, de Jong PJ, Menon AG, Kley NA, Seizinger BR, Zwarthoff EC. Cloning and characterization of MN1, a gene from chromosome 22q11, which is disrupted by a balanced translocation in a meningioma. Oncogene 1995;10:1521.

76. Pulst S-M, Rouleau GA, Marineau C, Fain P, Sieb JP. Familial meningioma is not allelic to neurofibromatosis 2. Neurology 1993;43:2096.

77. Schneider G, Lutz S, Henn W, Zang KD, Blin N. Search for the putative suppressor genes in meningiomas: significance of chromosome 22. Hum Genet 1992;53:579.

78. Lindblom A, Ruttledge M, Collins VP, Nordenskjold M, Dumanski JP. Chromosomal deletions in anaplastic meningiomas suggest multiple regions outside chromosome 22 as important in tumor progression. Int J Cancer 1994;56:354.

79. Rempel SA, Schwechheimer K, Davis RL, Cavenee WK, Rosenblum ML. Loss of heterozygosity for loci on chromosome 10 is associated with morphologically malignant meningioma progression. Cancer Res 1993;53:2387.

80. Louis DN, Ramesh V, Gusella JF. Neuropathology and molecular genetics of neurofibromatosis 2 and related tumors. Brain Pathol 1995;5:163.

81. Sainz J, Huynh DP, Figueroa K, Ragge NK, Baser ME, Pulst S-M. Mutations of the neurofibromatosis type 2 gene and lack of the gene product in vestibular schwannomas. Hum Molec Genet 1994;3:885.

82. Neumann HPH, Lips CJM, Hsia YE, Zbar B. Von Hippel Lindau syndrome. Brain Pathol 1995;5:181.

83. Seizinger BR, Rouleau GA, Ozelius LJ, Lane AH, Farmer GE, Lamiell JM, Haines J, Yuen JWM, Collins D, Majoor-Krakauer D, Bonner T, Mathew C, Rubenstein A, Halperin J, McConkie-Rosell A, Green JS, Trofatter JA, Ponder BA, Eiermann L, Bowmer MI, Schimke R, Oostra B, Aronin N, Smith DI, Drabkin H, Waziri MH, Hobbs WJ, Martuza RL, Coneally PM, Hsia YE, Gusella JF. Von Hippel Lindau disease maps to the region of chromosome 3 associated with renal cell carcinoma. Nature 1988;332:268.

84. Crossey PA, Foster K, Richards FM, Phipps ME, Latif F, Tory K, Jones MH, Bentley E, Kumar R, Lerman MI. Molecular genetic investigations of the mechanism of tumourigenesis in von Hippel-Lindau disease: analysis of allele loss in VHL tumours. Human Genetics 1994;93:53.

85. Latif F, Tory K, Gnarra J, Yao M, Duh F-M, Orcutt ML, Stackhouse T, Kuzmin I, Modi W, Gell L, Schmidt L, Zhou F, Li H, Ming HW, Chen F, Glenn G, Choyke P, Walther MM, Weng Y, Duan D-SR, Dean M, Glavac D, Richards FM, Crossey PA, Ferguson-Smith MA, Le Paslier D, Chumakov I, Cohen D, Chinault AC, Maher ER, Linehan WM, Zbar B, Lerman MI. Identification of the von Hippel-Lindau disease tumor suppressor gene. Science 1993;260:1317.

86. Kanno H, Kondo K, Ito S, Yamamoto I, Fujii S, Torigoe S, Sakai N, Hosaka M, Shuin T, Yao M. Somatic mutations of the von Hippel-Lindau tumor suppressor gene in sporadic central nervous system hemangioblastomas. Cancer Research 1994;54:4845.

87. Joseph JT, Lisle DK, Jacoby LB, Paulus W, Barone R, Cohen ML, Roggendorf WH, Bruner JM, Gusella JF, Louis DN. NF2 gene analysis distinguishes hemangiopericytoma from meningioma. Am J Pathol 1995 (in press).

88. Ono Y, Ueki K, Joseph JT, Louis DN. Homozygous deletions of the CDKN2/p16 gene in dural hemangiopericytomas (submitted for publication).

89. Forus A, Florenes VA, Maelandsmo GM, Fodstad O, Myklebost O. The protooncogene CHOP/GADD153, involved in growth arrest and DNA damage response, is amplified in a subset of human sarcomas. Cancer Gen Cytogenet 1994;78:165.

90. Cavenee WK, Dryja TP, Phillips RA, Benedict WF, Godbout R, Gallie BL, Murphree AL, Strong LC, White RL. Expression of recessive alleles by chromosomal mechanisms in retinoblastoma. Nature 1983;305:779.

91. Cavenee WK, Murphree AL, Shull MM, Benedict WF, Sparkes RS, Kock E, Nordenskjold M. Prediction of familial predisposition to retinoblastoma. N Eng J Med 1986;314:1201.

92. Friend SH, Dryja TP, Weinberg RA. Oncogenes and tumor suppressor genes. N Eng J Med 1988;318:618.

93. Louis DN, von Deimling A. Hereditary tumor syndromes of the nervous system: overview and rare syndromes. Brain Pathol 1995;5:145.

94. Short MP, Richardson EP, Haines JL, Kwiatkowski DJ. Clinical, neuropathological, and genetic aspects of the tuberous sclerosis complex. Brain Pathol 1995;5:173.

95. von Deimling A, Krone W, Menon AG. Neurofibromatosis type 1: pathology, clinical features, and molecular genetics. Brain Pathol 1995;5:153.