Cytogenetics and Molecular Genetics of Childhood Brain Tumors 1

Neuro-Oncology Cytogenetics and molecular genetics of childhood brain tumors 1 Jaclyn A. Biegel 2 Division of Human Genetics and Molecular Biology, The Children’s Hospital of Philadelphia and the Department of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA 19104 Considerable progress has been made toward improving Introduction survival for children with brain tumors, and yet there is still relatively little known regarding the molecular genetic Combined cytogenetic and molecular genetic approaches, events that contribute to tumor initiation or progression. including preparation of karyotypes, FISH, 3 CGH, and Nonrandom patterns of chromosomal deletions in several loss of heterozygosity studies have led to the identica- types of childhood brain tumors suggest that the loss or tion of regions of the genome that contain a variety of inactivation of tumor suppressor genes are critical events novel tumor suppressor genes and oncogenes. Linkage in tumorigenesis. Deletions of chromosomal regions 10q, analysis in families, which segregate the disease pheno- 11 and 17p, for example, are frequent events in medul- type, and studies of patients with constitutional chromo- loblastoma, whereas loss of a region within 22q11.2, somal abnormalities have resulted in the identication of which contains the INI1 gene, is involved in the develop- many of the disease genes for which affected individuals ment of atypical teratoid and rhabdoid tumors. A review have an inherited predisposition to brain tumors (Table of the cytogenetic and molecular genetic changes identi- 1). The frequency of mutations of these genes in sporadic ed to date in childhood brain tumors will be presented. tumors, however, is still relatively low. Although tumor Neuro-Oncology 1, 139–151, 1999 (Posted to Neuro- development is inuenced both by genetic and environ- Oncology [serial online], Doc. 98-30, April 30, 1999. mental factors, the strongest case for genetic predisposi- URL <neuro-oncology.mc.duke.edu>) tion to the development of malignancies can be made for the youngest patients, particularly for those children diagnosed in the rst year of life. Identifying tumor-specic and tumor-associated genes thus becomes critical on multiple levels. From a genetic standpoint, identifying a germ-line mutation in a family may have great implica- tions for the planning of future pregnancies. In a clinical Received 18 December 1998, accepted 7 January 1999. oncology setting, identifying an acquired chromosomal 1Supported in part by grants CA46274 and NS34514 from the deletion or gene mutation may necessitate very aggres- National Institutes of Health. sive approaches, or may allow the administration of less 2Address correspondence and reprint requests to Jaclyn Biegel, Ph.D., toxic therapy. As we unravel the genetic pathways that Room 1002 Abramson Research Center, The Children’s Hospital of lead to childhood brain tumors, treatments will be tar- Philadelphia, 3516 Civic Center Boulevard, Philadelphia, PA 19104. geted to those genetic alterations that are present in a 3Abbreviations used are as follows: FISH, uorescence in situ given tumor. hybridization; CGH, comparative genomic hybridization; PNET, primitive neuroectodermal tumor; AT/RT, atypical teratoid/rhabdoid tumor; AA, anaplastic astrocytoma; GBM, glioblastoma multiforme. Medulloblastoma 4Feuerstein, B., and Biegel, J.A. (1999) Unpublished data. The most common malignant brain tumor in children is 5Feuerstein, B. (1999) Personal communication. medulloblastoma, the prototype PNET. Primitive neu- 6Biegel, J.A. (1999) Unpublished data. roectodermal tumors that arise in the cerebellum are gen- erally classied as medulloblastoma, whereas similar his- 7James, C.D. (1999) Personal communication. tologic entities that arise in other locations are referred to Neuro-Oncology APRIL 1999 139 J.A. Biegel: Genetics of pediatric brain tumors Table 1. Genetic diseases that predispose to the development of brain tumors Disease Gene Tumor type Li-Fraumeni syndrome TP53 (17p13) Glioma, ependymoma, choroid plexus tumor Neurobromatosis-1 NF1 (17q11) Glioma Neurobromatosis-2 NF2 (22q12) Vestibular schwannoma, ependymoma, meningioma Nevoid basal cell carcinoma syndrome PTC (9q22) Medulloblastoma Tuberous sclerosis TSC1 (9q34) Sub-ependymal giant-cell tumor TSC2 (16p13) Turcot’s syndrome APC (5q21) Medulloblastoma hMLH1 (3p21) Glioblastoma hPMS2 (7p22) Von-Hippel Lindau disease VHL (3p25) Hemangioblastoma as PNETs. For the purposes of this review, the term gests that loss of more than one gene on 17p may be “medulloblastoma” will refer to tumors in the posterior important in tumorigenesis. Furthermore, the identica- fossa and will not distinguish between tumors without tion of tumors with extra copies of the long arm of chro- evidence of differentiation versus those with glial, neu- mosome 17, in the absence of a 17p deletion, suggests ronal, or ependymal differentiation. It should be noted that the duplication of sequences on 17q may confer a that, regardless of location, CNS PNETs are distinct from selective growth advantage to tumor cells. The whole- peripheral PNETs (Ewing’s sarcoma or peripheral neu- arm deletions and duplications of chromosome 17 have roepithelioma), which are characterized by specic made the positional cloning of a chromosome 17 medul- translocations involving the Ewing’s sarcoma region loblastoma gene extremely difcult. Several known (EWS) locus on chromosome 22q12. genes, most notably p53 (Biegel et al., 1992; Raffel et al., Among all of the pediatric brain tumors, the most 1993; Saylors et al., 1991), have been ruled out as the comprehensive cytogenetic studies have been reported chromosome 17 medulloblastoma tumor suppressor for medulloblastoma. Bigner et al. (1997) and Bhat- gene, and a novel candidate has not yet been reported. tacharjee et al. (1997) have recently published large series The frequency of a deleted 17p13 region in medul- on pediatric brain tumors subjected to standard cytoge- loblastoma is estimated to be between 30 and 50% of netic analyses, and detailed ndings of a large number of cases, depending on the individual series analyzed. medulloblastomas as well as pediatric gliomas may be Although i(17q) or 17p deletion is not specic for medul- found in those reports. loblastoma, it is seen in this tumor at a higher frequency Early cytogenetic studies demonstrated that the most than in any other tumor type. Furthermore, the nding of frequent cytogenetic abnormality among medulloblas- an i(17q) as a single structural abnormality in karyotypes tomas was an isochromosome 17q [i(17q] (Biegel et al., suggests that it is a primary cytogenetic event, and not a 1989; Bigner et al., 1988a; Grifn et al., 1988). This cytogenetic change associated with clonal evolution results in loss of most of the short arm of chromosome (Biegel et al., 1989). For this reason, there has been great 17 (17p) as well as duplication of the long arm (17q). interest in determining if deletion of 17p is associated Molecular genetic studies that used restriction length with clinical outcome. Batra et al. (1995) and Cogen and polymorphism analysis and, more recently, polymerase McDonald (1996) suggested that 17p deletion confers a chain reaction–based microsatellite analysis, have con- poor prognosis, whereas Emadian et al. (1996) and rmed the frequent loss of 17p in these tumors (Biegel et Biegel et al. (1997) did not nd a statistically signicant al., 1992; Thomas and Raffel, 1991), as shown in Fig. 1. difference in outcome in patients with or without tumor- Attempts to dene a small common region of deletion in associated 17p deletions. The difference in results regard- 17p, based on the loss of 17p13.3 in a limited number of ing 17p deletion and outcome may be biased by the tumors (Biegel et al., 1992), have been largely unsuccess- patient population included in the cytogenetic and ful. Interphase FISH (Biegel et al., 1995) and microsatel- molecular studies, especially if the treating institution lite analysis have shown that the breakpoints in chromo- sees a higher-risk group of patients. In recent studies some 17 cluster between the centromere and the proxi- reported by Scheurlen et al. (1998), 30 primary medul- mal region of 17 (17p11.2), a region commonly deleted loblastoma—PNET s and 6 metastasis specimens were in patients with the Smith-Magenis syndrome (Scheurlen analyzed for loss of heterozygosity for 17p13, as well as et al., 1998; Wilgenbus et al., 1997). The tendency for amplication of the MYCC oncogene. Loss of 17p13 was chromosome 17 to undergo breakage in this region, with found in 14 of 30 (47%) tumors and 6 of 6 cerebrospinal subsequent isochromosome formation, may be related to uid specimens and was found to be associated with a the repeat sequences present in the region around the poor outcome. However, MYCC amplication was also centromere; therefore, sequences that are interrupted by shown to predict a poor response to therapy, and every the rearrangements may not contain a tumor suppressor case that demonstrated amplication of MYCC also had gene. Deletion of most or all of 17p in medulloblastoma, loss of 17p. It is clear that multivariate statistical analy- instead of mitotic recombination (leading to loss of het- ses on large numbers of patients, preferably in conjunc- erozygosity) or interstitial deletions within 17p13, sug- tion with clinical trials conducted through a cooperative 140 Neuro-Oncology AP RIL 199 9 J.A. Biegel: Genetics of pediatric brain tumors group mechanism, will

Load more