Somatic Deletion Mapping on Chromosome 10 and Sequence Analysis of PTEN/MMAC1 Point to the 10Q25-26 Region As the Primary Target in Low-Grade and High-Grade Gliomas

Oncogene (1998) 16, 3331 ± 3335  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc SHORT REPORT Somatic deletion mapping on chromosome 10 and sequence analysis of PTEN/MMAC1 point to the 10q25-26 region as the primary target in low-grade and high-grade gliomas

Daniel Maier1, Zuwen Zhang1, Elisabeth Taylor1, Marie-France Hamou2, Otmar Gratzl1, Erwin G Van Meir2, Rodney J Scott1 and Adrian Merlo1

1Molecular Neuro-Oncology, Neurosurgery and Department of Research, University Hospital, Schanzenstr.46, CH-4031 Basel; 2Laboratory of Tumor Biology and Genetics, Department of Neurosurgery, University Hospital, CH-1011 Lausanne, Switzerland

The 10q25-26 region between the dinucleotide markers dicult to delineate speci®c regions of loss, 10q25-26 D10S587 and D10S216 is deleted in glioblastomas and, has consistently been suspected to harbor a tumor as we have recently shown, in low-grade oligodendro- suppressor gene (Fults et al., 1993; Rasheed et al., gliomas. We further re®ned somatic mapping on 10q23- 1995; Albarosa et al., 1996). Recently, PTEN/MMAC1 tel and simultaneously assessed the role of the candidate on 10q23.3 has been proposed as a candidate tumor tumor suppressor gene PTEN/MMAC1 in glial neo- suppressor gene for gliomas and other neoplasms, plasms by sequence analysis of eight low-grade and 24 clearly de®ning a second locus on 10q (Li et al., 1997; high-grade gliomas. These tumors were selected for Steck et al., 1997). This gene has also been shown to be partial or complete loss of chromosome 10 based on involved in two rare inherited disorders, the Cowden deletion mapping with increased microsatellite marker (Liaw et al., 1997) and the Bannayan-Zonana density at 10q23-tel. Three out of eight (38%) low-grade syndrome (Marsh et al., 1997). PTEN/MMAC1 and 3/24 (13%) high-grade gliomas exclusively target encodes a protein tyrosine phosphatase, shares 10q25-26. We did not ®nd a tumor only targeting extensive homology to the cytoskeletal proteins tensin 10q23.3, and most tumors (23/32, 72%) showed large and auxilin, and has been shown to act either as deletions on 10q including both regions. The sequence positive or negative regulator during signal transduc- analysis of PTEN/MMAC1 revealed nucleotide alteration, cell cycle progression, and cellular transformation tions in 1/8 (12.5%) low-grade gliomas in a tumor with (Tonks et al., 1996), and is regulated by transforming LOH at 10q21-qtel and in 5/21 (24%) high-grade growth factor b (Li and Sun, 1997). gliomas displaying LOH that always included 10q23- The more telomeric region on 10q25-26 between the 26. Our re®ned mapping data point to the 10q25-26 dinucleotide markers D10S587 and D10S216 is region as the primary target on 10q, an area that also involved in glioblastomas (Rasheed et al., 1995; harbors the DMBT1 candidate tumor suppressor gene. Albarosa et al., 1996), and in low-grade oligoden- The fact that we ®nd hemizygous deletions at 10q25-qtel drogliomas, too, as we have recently shown (Maier et in low-grade astrocytomas and oligodendrogliomas ± two al., 1997). The involvement of this very region in histologically distinct entities of gliomas ± suggests the earlier tumor stages has also been described in existence of a putative suppressor gene involved early in endometrial carcinomas (Nagase et al., 1997). We glial tumorigenesis. further re®ned somatic mapping on 10q in another seven low-grade gliomas (®ve astrocytomas and two Keywords: deletion mapping; chromosome 10q23.3 and mixed oligoastrocytomas). From the total number 10q25-26; PTEN/MMAC1 (n=86) of low-grade (n=14) and high-grade (n=72) gliomas which have been used for somatic mapping on chromosome 10, we selected eight low-grade (®ve astrocytomas, one mixed oligoastrocytoma, and two Loss of heterozygosity (LOH) on chromosome 10 has oligodendrogliomas) and 24 high-grade gliomas, which frequently been found in a variety of human cancers, either displayed LOH at the PTEN/MMAC1 locus such as endometrial (Peier et al., 1995), prostate (n=23), selectively targeted 10q25-26 (n=6) or were of (Gray et al., 1995), bladder (Cappellen et al., 1997), low-grade type without LOH on chromosome 10 renal cell carcinoma (Morita et al., 1991), Non- (n=3), to assess the role of PTEN/MMAC1 in glial Hodgkin lymphomas (Speaks et al., 1992), malignant tumorigenesis by analysing the genomic sequence of meningiomas (Rempel et al., 1993), melanomas (Herbst this new candidate suppressor gene on 10q23.3. These et al., 1994) and gliomas (Fults et al., 1990; Fujimoto tumors were selected for partial and complete LOH, et al., 1989; Rasheed et al., 1992; James et al., 1988; since the biallelic inactivation of a tumor suppressor Steck et al., 1995). Eighty to ninety per cent of gene typically involves an intragenic change (nucleotide malignant gliomas display large deletions on chromo- substitution, small insertion, or microdeletion) within some 10. Although somatic mapping has proven one allele, combined with inactivation of the other allele through the loss of a large chromosomal region (Knudson, 1993). Sequence analysis of PTEN/MMAC1 Correspondence: A Merlo revealed nucleotide alterations in 1/8 (12.5%) low- Received 22 October 1997; revised 8 January 1998; accepted 8 grade gliomas and in 5/21 (24%) high-grade gliomas. January 1998 None of the 9/32 (28%) tumors that did not display Two targets on chromosome 10q23-26 in gliomas DMaieret al 3332 LOH at the PTEN/MMAC1 locus showed sequence oligodendrogliomas and ®ve astrocytomas). DNA from alterations of this gene. All mutated tumors displayed tumor specimens and lymphocytes was analysed for LOH at the PTEN/MMAC1 locus. These ®ndings loss of heterozygosity (LOH) by ampli®cation of 31 di- suggest that PTEN/MMAC1 plays a role in glioma and tetranucleotide repeat-containing sequences using tumorigenesis. However, we ®nd partial somatic PCR and the conditions described (Merlo et al., 1994). deletions to be more frequent on 10q25-26 in low- Primers for microsatellite markers were obtained from grade and high-grade gliomas, suggesting that this Research Genetics (Huntsville, AL). Primer sequences region is the primary target in glioma tumorigenesis on for PTENCA, a highly informative dinucleotide repeat chromosome 10q. marker at the 5' end of PTEN/MMAC1, were kindly Tumor tissue and peripheral blood lymphocytes provided by P Cairns and D Sidransky (Wang et al., were obtained from 32 patients diagnosed with 1997). For microsatellite analysis, one primer was primary CNS tumors and were classi®ed according to labeled with T4-polynucleotide kinase (New England the World Health Organization: 24 high-grade astro- Biolabs) and g-32P-ATP. Fifty nanograms of genomic cytomas grade IV (glioblastoma multiforme) and eight DNA (normal and tumor) were subjected to 30 cycles low-grade gliomas grade II (one oligoastrocytoma, two of PCR ampli®cation with annealing temperatures that

Table 1 Mutations and polymorphisms of PTEN/MMAC1 in primary gliomas Tumor type Case no. Codon changea Mutation Exon/Intron Predicted alterationb

All GE31 CAT183→CGT183 missense 3 H61R GBM BS7 ATG594→D[ATG]592 ± 594 frameshift 6 delM198 GBM BS10 T1G intron 4 splicing variant GBM BS11 CGA699→TGA699 nonsense 7 R233Opal BS11 del T intron 3 polymorphism GBM BS54 CGC519→CAC519 missense 6 R173H GBM LS971 TAT264→D[AT]263 ± 264 frameshift 5 Y88fs a Mutated nucleotides are in boldface type. b Positions refer to the deduced PTEN protein sequence; fs, frameshift

b ACGT a NNTT

q 21.1 q 25.2

D10S676 D10S221

q 23.3 q 25.3 PTEN/MMAC1 locus c ACGT

PTENCA D10S587

q 24.2 q 26.1

D10S254 D10S214 Figure 1 LOH analysis on chromosome 10q and mutations of PTEN/MMAC1 in primary gliomas (a) The deletion breakpoint in this low-grade astrocytoma (GE74) lies telomeric to the PTEN/MMAC1 locus (10q23.3) between the markers D10S254 (10q24.2) and D10S221 (10q25.2). Arrowheads denote loss of alleles consistent with LOH. N, normal; T, tumor. (b) Mutation in glioblastoma (GBM7, middle lane; GBM2, left lane; GBM8, right lane). Sequence of nucleotides 571 ± 603 of exon 6 of PTEN/MMAC1 in the antisense orientation. Arrow indicates an in frame 3 bp deletion of nucleotides 592 ± 594. (c) Mutation in a low-grade astrocytoma (GE31). Sequence of nucleotides 173 ± 193 of genomic DNA from blood (left) and primary tumor (right) from the same patient using a sense primer. Arrow indicates a transversion from A to G at nucleotide 182 Two targets on chromosome 10q23-26 in gliomas DMaieret al 3333 ranged from 50 ± 628C. Products were separated using Li et al. in their original reports. The fact that we 7% polyacrylamide-formamide gel electrophoresis detected a mutation in a low-grade astrocytoma followed by autoradiography. For informative cases, (GE31) suggests that this candidate tumor suppressor allelic loss was scored if the radiographic signal of one gene might occasionally also contribute to earlier allele was at least 50% less in tumor DNA as stages of tumorigenesis. Since the mutation rate as compared to the corresponding normal allele. assessed by direct analysis of the genomic sequence of For mutation screening of the candidate tumor PTEN/MMAC1 in primary gliomas was found to be suppressor gene PTEN/MMAC1, 50 ng of genomic much lower than LOH at the PTEN/MMAC1 locus, DNA from primary tumors and blood lymphocytes we searched for alternative mechanisms of gene was ampli®ed by PCR. Cycle sequencing was inactivation, that could be missed by direct sequence performed directly on PCR products using a nested analysis. For example, homozygous deletions are sequencing primer end-labeled with g32P-ATP and the technically dicult to exclude in primary tumor AmpliCycleTM Sequencing Kit (Perkin Elmer, Branch- specimens. It has been shown that an apparently burg, NJ). The primer sequences for exons 1 ± 9 of retained small region within a large hemizygous PTEN/MMAC1 were kindly provided by R Parsons. deletion can be reliably interpreted as a homozygous The products of cycle sequencing were electrophoresed deletion in primary tumors (Cairns et al., 1995). We on standard 6% acrylamide gels followed by auto- therefore used a polymorphic STS marker that lies radiography. All mutations were con®rmed by a within PTEN/MMAC1 (PTENCA) to search for repeated analysis. Because of a relatively high biallelic deletions. However, homozygous deletions admixture of non-neoplastic DNA, specimen GE31 could not be discerned in these tumors using this was recut, and reextracted DNA was used to con®rm approach. In 17/32 gliomas displaying large hemizy- the codon 61 mutation in exon 3. gous deletions always spanning the 10q23-qtel region Our sequencing data of PTEN/MMAC1 combined (not considering the six tumors with mutations of with somatic mapping on chromosome 10q shows two PTEN/MMAC1 and the nine gliomas with retention of distinct regions on 10q to be involved in glioma both alleles at the PTEN/MMAC1 locus), we could not tumorigenesis. One region is de®ned by the candidate observe a case with intragenic retention of both tumor suppressor gene PTEN/MMAC1 on 10q23.3. To PTENCA alleles suspicious for a homozygous dele- study the role of this gene in primary gliomas by tion. PTEN/MMAC1 could also be inactivated by sequence analysis, we selected eight low-grade and 24 transcriptional silencing due to aberrant methylation of high-grade gliomas according to the following criteria: the promoter as shown for the von Hippel-Lindau we included all tumors with partial deletions on (VHL) gene (Herman et al., 1994) and for p16INK4A chromosome 10q (six only targeting 10q25-26, two (Merlo et al., 1995). Whether this mechanism plays a tumors targeting both 10q23.3 and 10q25-26), with role in gene inactivation of PTEN/MMAC1 remains to complete LOH on chromosome 10 (one low-grade, 20 be investigated. high-grade gliomas) and without LOH on chromosome Our re®ned somatic deletion mapping of chromo- 10 (three low-grade gliomas). We detected 5/21 (24%) some 10q which is based on four low-grade and four sequence alterations in high-grade gliomas displaying high-grade gliomas (Figure 2) shows that the 10q25-26 LOH at the PTEN/MMAC1 locus (Table 1 and Figure region is the primary target on 10q in these neoplasms. 1b). In GBM BS10, we detected a splicing junction In addition to 62 previously mapped low-grade (n=7) variant in intron 4. In GBM BS7, we found a 3 bp and high-grade gliomas (n=55), we mapped another deletion in exon 6 (delM198) (Figure 1b). A 2 bp seven low-grade and 17 high-grade gliomas for LOH on deletion at codon 88 in exon 5 (Y88frameshift) was chromosome 10 using the same approach as previously detected in GBM LS971, leading to a truncated version described (Maier et al., 1997). Taken together, complete of PTEN/MMAC1 at the protein level. In GBM BS11, LOH of chromosome 10q was found in 64 gliomas (one we found two sequence alterations: a nonsense WHO grade II, ®ve grade III, and 58 grade IV mutation in exon 7 (R233Opal) and a polymorphism gliomas). To obtain these data, we used on average in intron 3 (del T). In GBM BS54, a missense mutation 16 polymorphic markers (s.d.+6.2, range 7 ± 31) per was detected in exon 6 (R173H). In one (GE31) out of tumor sample; each tumor was covered on average with eight low-grade gliomas (12.5%), we found a missense an informativity of 70% to prove complete LOH. mutation in a tumor that has LOH at 10q21-qtel, Combining the data of these 86 primary gliomas, we leading to an amino acid change from Histidine to obtained the following frequencies of LOH. In low- Arginine at codon 61 in exon 3 (H61R) (Figure 1c). grade gliomas: LOH spanning both areas (10q23.3 and This low-grade glioma has a relatively high admixture 10q25-26) was detected in 2/14 (14%) tumors, LOH of non-neoplastic cells (or an intra-tumoral cell exclusively targeting 10q25-26 was found in 3/14 (22%) population without LOH on 10q), as visible on the tumors, no LOH on chromosome 10 in 9/14 (64%) weak signal of the mutated nucleotide. For this reason, tumors; in high-grade gliomas: LOH spanning both we reassessed DNA from recut tumor tissue of the areas (10q23.3 and 10q25-26) was detected in 64/72 same patient which con®rmed the point mutation. The (89%) tumors, LOH exclusively targeting 10q25-26 was amino acid designated by this codon is conserved in found in 3/72 (4%) tumors, no LOH on chromosome the proteins tensin and auxilin which share extensive 10 in 5/72 (7%) tumors. Not a single tumor was N-terminal sequence homology with PTEN/MMAC1, exclusively targeting the 10q23 region. suggesting that this sequence alteration might be Previously, we demonstrated LOH in a speci®c low- functionally relevant . Mutations in this large region grade glioma subtype (oligodendroglioma) to be of homology (+175 amino acids) might be functionally con®ned to 10q25-26 (Maier et al., 1997). We important to tumor suppression since mutations within extended this analysis now to another low-grade this domain have been described by Steck et al and by glioma subtype (astrocytoma WHO grade II) and to Two targets on chromosome 10q23-26 in gliomas DMaieret al 3334

Figure 2 Deletion map of chromosome 10q in four low-grade (left) and four high-grade (right) gliomas. The chromosomal orientation and the locus of PTEN/MMAC1 is shown on the left. The map position of the 31 microsatellite markers used is listed according to the integrated map of chromosome 10 on the left (Moschonas et al., 1996). The marker PTENCA lies at the 5' end of PTEN/MMAC1 (Wang et al., 1997). * Denotes the locus of DMBT1 (Mollenhauer et al., 1997). Black circles, LOH; white circles, retention of both informative alleles; hatched circles, not informative; S=shift, allelic expansion or deletion. Abbreviations: wt, wild- type; mut, mutated; histological classi®cation according to the WHO grading (A, astrocytoma; OG, oligodendroglioma; GBM, glioblastoma multiforme)

a low-grade glioma of mixed composition (oligoas- least one of the ¯anking loci (AFMA086WG9, trocytoma WHO grade II). In these tumors, we D10S541 and D10S579). Of the 23 gliomas with con®rmed the more telomeric deletion, as shown for LOH at the PTEN/MMAC1 locus (22 gliomas with astrocytoma WHO grade II GE74 (Figures 1a and 2). LOH at PTENCA and tumor GBM LS971 with a non- The site of recombination does not point to PTEN/ informative PTENCA but loss at ¯anking markers), MMAC1 as the target in this tumor. In addition, sequence analysis of this gene revealed altogether six sequence analysis of PTEN/MMAC1 was wild-type for (26%) mutations in these 23 low- and high-grade this low-grade glioma. Wild-type sequence of PTEN/ gliomas with LOH at the PTEN/MMAC1 locus. This MMAC1 was also obtained for the mixed oligoastro- mutation rate in this study is similar to the one cytoma WHO grade II, which displayed complete LOH described in the original reports (Li et al., 1997; Steck on chromosome 10 (OAII 1046). et al., 1997) and does not explain the high rate of LOH The frequency of hemizygous loss at the PTEN/ on chromosome 10q. MMAC1 locus was 25% (2/8) in low-grade and 83% Our re®ned mapping data points to the 10q25-26 (20/24) in high-grade gliomas, as assessed by the most region as the primary target on 10q, surrounded by the informative marker PTENCA (30/32=94% of samples markers D10S209 and D10S587, a 7 cM region that were informative). All cases displaying LOH at the also harbors DMBT1 which has recently been PTENCA locus were also hemizygously deleted in at proposed as a candidate tumor suppressor gene for Two targets on chromosome 10q23-26 in gliomas DMaieret al 3335 medulloblastomas and glioblastomas (Mollenhauer et Acknowledgements al., 1997). Tumors OG BS40 and GBM LS946 point to We would like to thank Dr Pierre-Yves Dietrich and the a region slightly telomeric to the DMBT1 locus neurosurgical sta of the University Hospitals of Basel, (D10S587-D10S216) (Figure 2). The fact that we ®nd Lausanne, and Geneva for providing tumor samples. This hemizygous deletions at 10q25-26 in a signi®cant work was supported by the Swiss National Science Foundation (No. 31-44340.95, AM; No. 31-49194.96, proportion of low-grade astrocytomas and oligoden- EGVM) and the Swiss Cancer League (KFS No. 159-9- drogliomas ± two histologically distinct entities of 1995, AM). gliomas ± suggests the existence of a putative suppressor gene involved early in glial tumorigenesis.

References

Albarosa R, Colombo BM, Roz L, Magnani I, Pollo B, Maier D, Comparone D, Taylor E, Zhang Z, Gratzl O, Van Cirenei N, Giani C, Fuhrman Conti AM, DiDonato S and Meir EG, Scott RJ and Merlo A. (1997). Oncogene, 15, Finocchiaro G. (1996). Am.J.Hum.Genet.,58, 1260 ± 997 ± 1000. 1267. Marsh DJ, Dahia DML, Zheng Z, Liaw D, Parsons R, Cairns P, Polascik TJ, Eby Y, Tokino K, Califano J, Merlo Gorlin RJ and Eng C. (1997). Nature Genet., 16, 333 ± 334. A, Mao L, Herath J, Jenkins R, Westra W, Rutter JL, Merlo A, Gabrielson E, Askin F and Sidransky D. (1994). Buckler A, Gabrielson E, Tockman M, Cho K, Hedrick L, Cancer Res., 54, 640 ± 642. Bova GS, Issacs W, Schwab D and Sidransky D. (1995). Merlo A, Herman JG, Mao L, Lee DJ, Gabrielson E, Burger Nature Genet., 11, 210 ± 212. PC, Baylin SB and Sidransky D. (1995). Nature Med., 1, Cappellen D, Sixtina GDM, Chopin D, Thiery JP and 686 ± 693. Radvanyi F. (1997). Oncogene, 14, 3059 ± 3066. Mollenhauer J, Wiemann S, Scheurlen W, Korn B, Hayashi Fujimoto M, Fults DW, Thomas GA, Nakamura Y, Y, Wilgenbus KK, von Deimling A and Poustka A. (1997). Heilbrun MP, White R, Story JL, Naylor SL, Kagan Nature Genet., 17, 32 ± 39. Hallet KS and Sheridan PJ. (1989). Genomics, 4, 210 ± 214. Morita R, Saito S, Ishikawa J, Ogawa O, Yoshida O, Fults D, Pedone CA, Thomas GA and White R. (1990). Yamakawa K and Nakamura Y. (1991). Cancer Res., 51, Cancer Res., 50, 5784 ± 5789. 5817 ± 5820.. Fults D and Pedone C. (1993). Genes. Chromosom. Cancer, 7, Moschonas NK, Spurr NK and Mao J. (1996). Cytogenet. 173 ± 177. Cell Genet., 72, 99 ± 112. Gray IC, Phillips SMA., Lee SJ, Neoptolemos JP, Nagase S, Yamakawa H, Sato S, Yajima A and Horii A. Weissenbach J and Spurr NK. (1995). Cancer Res., 55, (1997). Cancer Res., 57, 1630 ± 1633. 4800 ± 4803. Peier SL, Herzog TJ, Tribune DJ, Mutch DG, Gersel DJ Herbst RA, Weiss J, Ehnis A, Cavenee WK and Arden KC. and Goodfellow PJ. (1995). Cancer Res., 55, 1922 ± 1926. (1994). Cancer Res., 54, 3111 ± 3114. Rasheed BKA, Fuller GN, Friedman AH, Bigner DD and Herman JG, Latif F, Weng Y, Lerman MI, Zbar B, Liu S, Bigner SH. (1992). Genes. Chromosom. Cancer, 5, 75 ± 82. Samid D, Duan DSR, Gnarra JR, Linehan WM and Rasheed BKA, McLendon RE, Friedman HS, Friedman Baylin SB. (1994). Proc.Natl.Acad.Sci.USA,91, 9700 ± AH, Fuchs HE, Bigner DD and Bigner SH. (1995). 9704. Oncogene, 10, 2243 ± 2246. James CD, Carlbom E, Dumanski JP, Hansen M, Rempel SA, Schwechheimer K, Davis RL, Cavenee WK and Nordenskjold M, Collins VP and Cavenee WK. (1988). Rosenblum ML. (1993). Cancer Res., 53, 2386 ± 2392. Cancer Res., 48, 5546 ± 5551. Speaks SL, Sanger WG, Masih AS, Harrington DS, Hess M Knudson AG. (1993). Proc. Natl. Acad. Sci. USA, 90, and Armitage JO. (1992). Genes. Chromosom. Cancer, 5, 10914 ± 10921. 239 ± 243. Li DM and Sun H. (1997). Cancer Res., 57, 2124 ± 2129. Steck PA, Ligon AH, Cheong P, Yung WKA and Pershouse Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, Puc MA. (1995). Genes. Chromosom. Cancer, 12, 255 ± 261. J, Miliaresis C, Rodgers L, McCombie R, Bigner SH, Steck PA, Pershouse MA, Jasser SA, Yung WKA, Lin H, Giovanella BC, Ittmann M, Tycko B, Hibshoosh H, Ligon AH, Langford LA, Baumgard ML, Hattier T, Davis Wigler MH and Parsons R. (1997). Science, 275, 1943 ± T, Frye C, Hu R, Swedlund B, Teng DHF and Tavtigian 1947. SV. (1997). Nature Genet., 15, 356 ± 362. Liaw D, Marsh DJ, Li J, Dahia PLM, Wang SI, Zheng Z, Tonks NK and Neel BG. (1996). Cell, 87, 365 ± 368. Bose S, Call KM, Tsou HC, Peacocke M, Eng C and Wang SI, Puc J, Li J, Bruce JN, Cairns P, Sidransky D and Parsons R. (1997). Nature Genet., 16, 64 ± 67. Parsons R. (1997). Cancer Res., 57, 4183 ± 4186.