New Deletion in Low-Grade Oligodendroglioma at the Glioblastoma Suppressor Locus on Chromosome 10Q25-26
Total Page:16
File Type:pdf, Size:1020Kb
Oncogene (1997) 15, 997 ± 1000 1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00 SHORT REPORT New deletion in low-grade oligodendroglioma at the glioblastoma suppressor locus on chromosome 10q25-26 Daniel Maier1, Domenico Comparone1, Elisabeth Taylor1, Zuwen Zhang1, Otmar Gratzl1, Erwin G Van Meir2, Rodney J Scott1 and Adrian Merlo1 1Molecular Neuro-Oncology, Departments of Research and Neurosurgery, University Hospital, Schanzenstr.46, CH-4031 Basel; 2Laboratory of Tumor Biology and Genetics, CHUV, Department of Neurosurgery, University Hospital, CH-1011 Lausanne, Switzerland Loss of heterozygosity on chromosome 10 is considered CDKN2 on 9p21 (Kamb et al., 1994; Nobori et al., to be associated with the progression of glioblastomas. 1994), have been shown to be genetically altered in Two closely related regions have recently been proposed glial neoplasms. Cytogenetic (Bigner et al., 1990; to contain the glioblastoma suppressor locus on chromo- Ransom et al., 1992) and molecular evidence some 10q25-26; a 1 cM region between the polymorphic (James et al., 1988; Rasheed et al., 1992; Karlbom et (CA)n-repeat markers D10S587 and D10S216, and an al., 1993; Fults and Pedone, 1993) suggests that area of 5 cM between the markers D10S221 and deletions on chromosome 10 are among the most D10S209. To con®rm and further delineate this region, frequent genetic alterations in glioblastomas. At least we analyzed 51 glioblastomas and 11 intermediate and three separate partial deletions have been described ± low-grade gliomas for loss of heterozygosity on chromo- most consistently loss on chromosome 10q24-26 some 10. 47/62 mostly malignant gliomas displayed (Rasheed et al., 1992; Fults and Pedone, 1993) ± complete loss of chromosome 10 and nine tumors were however, the potential tumor suppressor involved in unaltered, whereas four glioblastomas and two low-grade glioblastoma tumorigenesis has not yet been de®ned. oligodendrogliomas had partial loss on distal 10q. With Functional evidence, based on microcell-mediated these six tumors, we constructed a deletion map with transfer of major portions of chromosome 10 that increased marker density at 10q25-26 which shows two caused regression of the neoplastic phenotype, lends centromeric breakpoints at the markers D10S587 and further support to the presence of a glioblastoma D10S216, thus only con®rming the distal, but not the suppressor gene on 10q24-26 (Steck et al., 1995). It is proximal candidate glioblastoma suppressor locus. Two not known whether this putative suppressor is further out of four low-grade oligodendrogliomas displayed implicated in the development of other glial neoplasms partial deletions on 10q25-26. This suggests that deletion of oligodendrocytic or ependymal origin. on chromosome 10 is not merely a late event in the Recently, two closely related, but distinct regions on progression of glioblastomas, but could play a role chromosome 10q25-26 have been implicated to harbor earlier in the development of gliomas. the glioblastoma suppressor gene (Albarosa et al., 1996; Rasheed et al., 1995). By high density mapping, Keywords: oligodendroglioma; glioblastoma suppressor we de®ned two centromeric breakpoints in a glioblas- gene; deletion mapping; chromosome 10q25-26 toma and in a low-grade oligodendroglioma. Both point to the same 1 cM region which has been proposed to harbor the glioblastoma suppressor gene (Rasheed et al., 1995). Glial neoplasms of the central nervous system (CNS) Gliomas of dierent histopathological subtypes were account for the vast majority of primary brain tumors directly obtained from surgery and classi®ed by two and include astrocytoma, ependymoma, oligodendro- experienced neuropathologists according to the World glioma, glioblastoma, and tumors of mixed cellular Health Organisation (WHO) as summarised in Table 1. composition (Kleihues et al., 1993). Glioblastoma Using polymerase chain reaction (PCR) and the multiforme, the most common solid tumor of the conditions previously described (Merlo et al., 1994), adult CNS, represents the undierentiated end stage of DNA from glial tumor specimens and lymphocytes was glioma tumorigenesis, having a dismal prognosis with a analysed for loss of LOH by ampli®cation of 2 year survival of less than 5% (Walker et al., 1980). dinucleotide and tetranucleotide repeat-containing There is great hope that the discovery of genes sequences. We used 25 polymorphic (CA)n-dinucleo- involved in tumorigenesis will give rise to novel tide repeat markers and two polymorphic (GATA)n therapeutic strategies for solid tumors. Studies of tetranucleotide repeat markers (D10S675 and allelic loss in human tumors have provided a useful D10S676), as shown in Figure 1. Primers for these tool for localising and cloning tumor suppressor genes microsatellite markers were obtained from Research (Cavenee et al., 1985; Friend et al., 1986). Speci®c Genetics (Huntsville, AL). For microsatellite analysis, genes, such as EGFR on chromosome 7p (Libermann the forward primer was labeled with T4-polynucleotide et al., 1985), p53 on 17p (Nigro et al., 1989), p16INK4a/ kinase (New England Biolabs) and [g-32P]ATP. PCR- ampli®ed products were separated using 7% polyacry- lamide-formamide gel electrophoresis in denaturing Correspondence: A Merlo Received 6 February 1997; revised 10 April 1997; accepted 5.6 M urea followed by autoradiography at 7808C 18 April 1997 overnight. For informative cases, allelic loss was scored LOH on 10q25 ± 26 in gliomas DMaieret al 998 Table 1 Partial and complete LOH on chromosome 10 in gliomas if the radiographic signal of one allele was at least 50% Complete Partial No less in tumor DNA as compared to the corresponding Histologya LOH LOH LOH normal allele. Glioblastoma multiforme IV 45 4 2 Initially, we analysed 51 glioblastomas, two anaplas- (n=51) tic astrocytomas, three low-grade astrocytomas, two Anaplastic astrocytoma III (n=2) 2 anaplastic mixed oligoastrocytomas and four low-grade Astrocytoma II (n=3) 3 oligodendrogliomas for loss of heterozygosity on Oligoastrocytoma III (n=2) 2 Oligodendroglioma II (n=4) 2 2 chromosome 10, using 17 polymorphic microsatellite markers. Complete loss of these 17 markers was found Total (n=62) 47 (75.8%) 6 (9,7%) 9 (14.5%) in 45/51 glioblastomas (88%) and in 2/2 anaplastic aAccording to the WHO classi®cation, grades are denoted in Roman mixed oligoastrocytomas (100%) which were therefore ®gures excluded from further analysis. From the remaining 15 gliomas, nine tumors (2/51 glioblastomas, 2/2 anaplas- Figure 1 Deletion map displays partial LOH on chromosome 10q25-26 in six gliomas (four glioblastomas, GBM11, GBM991, GBM6, GBM946; two oligodendrogliomas WHO grade II, OG50, OG40). The map position of the 27 microsatellite marker used is listed according to the integrated map of chromosome 10 on the left (Moschonas, 1996). The bracketed areas denote two published candidate regions on 10q25 (Albarosa et al., 1996; Rasheed et al., 1995); our own minimal area of loss is shown below between markers D10214 and D10S212. Black rectangles, LOH; white rectangles, retention of both informative alleles; SH, shift (allelic expansion or deletion); NI, not informative LOH on 10q25 ± 26 in gliomas DMaieret al 999 tic astrocytomas WHO grade III, 3/3 astrocytomas GBM 946 OG 40 WHO grade II and 2/4 oligodendrogliomas WHO N T N T grade II) did not display any loss on chromosome 10, whereas four glioblastomas and two low-grade oligodendrogliomas had partial LOH on 10q. The distribution of complete and partial losses on chromosome 10 is shown in Table 1. With regard to overall genetic alterations, 49/51 glioblastomas (96%), 2/2 mixed anaplastic oligoastrocytomas (100%) and 2/4 low-grade oligodendrogliomas (50%) displayed either partial or complete loss on chromosome 10. Our initial mapping data, based on four glioblasto- D10S209 D10S209 mas and two low-grade oligodendrogliomas with partial LOH pointed to a large region of approxi- mately 23 cM on 10q25-26 potentially harboring a glioma suppressor locus. To increase mapping resolu- tion, we used 10 additional polymorphic dinucleotide (CA)n-repeat markers from this area (Gyapay et al., D10S587 D10S587 1994). The results obtained from these six tumors are summarized in Figure 1. Mapping data and autoradio- graphs, as shown in Figure 2, demonstrate the centromeric breakpoints in the two most informative gliomas, a glioblastoma (GBM 946) and a low-grade oligodendroglioma (OG40). These two tumors con®rm one of the two closely related regions with a 1 cM area bounded by the markers D10S587 and D10S216 which D10S216 D10S216 has previously been suspected to harbor a glioblastoma suppressor gene (Rasheed et al., 1995). However, it would exclude the other candidate region between the markers D10S221 and D10S209 (Albarosa et al., 1996) that lies more towards the centromere. Taking together our mapping data and the data published (Rasheed et al., 1995), the putative suppressor gene would lie adjacent to the dinucleotide marker D10S216. More- over, the recombination event in glioblastoma GBM946, as shown in Figure 2, might lend further support to this notion. In this tumor, the very distal markers on 10q25-26 display nearly complete allelic D10S169 D10S575 loss while the marker D10S216 at the breakpoint only shows a partial loss of the larger allele whereas more proximal markers clearly retain both copies of chromosome 10. This partial loss of one allele of the marker D10S216 could be accounted for by the presence of a subpopulation of cells within the tumor that