GENES, CHROMOSOMES & CANCER 14:29S300 (1995)

Frequent Deletions of Material From Chromosome Arm Ip in Oligodendroglial Tumors Revealed by Double-Target Fluorescence In Situ Hybridization and Microsatellite Analysis Naoya Hashimoto, Daisuke Ichikawa, Yoshiki Arakawa, Kousei Date, Satoshi Ueda, Yoshio Nakagawa, Akira Horii, Yusuke Nakamura, Tatsuo Abe, and Johji lnazawa

Departments of Hygiene (N.H., D.I.. Y.A., K.D., T.A., 1.1.) and Neurosurgery (N.H., S.U., Y0.N.). Kyoto Prefectural University of Medicine. Kamigyo-ku. Kyoto, I s t Department of Pathology, Tohoku University, School of Medicine, Aoba-ku. Sendai. Miyagi (AH), and Laboratory of Molecular Medicine, Institute of Medical Science, University of Tokyo, Minato-ku, Tokyo (Yu.N.), Japan

We undertook a cytogenetic analysis of 29 human brain tumors using double-target fluorescence in situ hybridization (FISH) and focusing on chromosome arm I p. One or more tumor suppressor genes in this arm have been suggested to be important in a variety of neuroectodermal tumors. The series included 9 oligodendrogliomas, 4 mixed gliomas, 10 astrocytomas, 4 glio- blastomas, and 2 central neurocytomas. We hybridized pericentromeric ( I q 12) and subtelomeric ( I p36) DNA probes to cell nuclei prepared from pamffin-embedded tissues and observed a strikingly high incidence of deletion of at least part of I p in oligodendrogliomas (lOO%) and mixed gliomas (75%). The results of the FISH analyses were confirmed by demonstration of loss of heterozygosity for a microsatellite polymorphism in 10 of the 29 tumors. As well as supporting the feasibility of FISH for detecting allelic deletions in chromosomes from paraffin-embedded tumor samples, the alteration of I p reported here will contribute to an understanding of the molecular genetic events in oligodendroglial tumor development. Genes Chrornosorn - - Cancer 14:295-300 (1995). 0 1995 Wiley-Liss, Inc.

INTRODUCTION

Glial tumors, the most common primary tumors in the central nervous system, are of three histo- pathological types: astrocytoma, oligodendro- glioma, or ependymoma. Oligodendrogliomas ac- count for 4-18% of glial tumors in humans; they occur predominantly in the cerebral hemispheres and usually arise in the fifth and sixth decades of life (Russell and Rubinstein, 1989). Since 1929, when the first series of patients affected with this tumor was described (Bailey and Bucy, 1929), many studies have focused on prognostic factors and therapeutic approaches. However, even now the standard therapeutic approach, surgical resec- tion followed by radiation, is not curative. T h e most recent review with respect to these tumors reported a 10-year survival rate of 24% (Celli et al., 1 994).

Allelic deletions observed frequently at specific chromosomal loci in tumors of a particular type are usually taken to indicate the locations of patholog- ically important tumor suppressor genes. During the last decade, a number of studies have sug- gested that tumor suppressor genes on chromo- some arms 9p, lOq, 13q, 17p, and 22q may be important in the development of astrocytic tumors, including glioblastomas (Collins and James, 1993). Since losses from l p have been demonstrated in

neuroectodermal tumors such as neuroblastomas (Brodeur and Fong, 1989) and medulloblastomas (Bigner et al., 1988a), structural or numerical ab- errations of this chromosome arm might be in- volved also in glial tumors. Indeed, at least two recent studies (Reifenberger et al., 1994; Kraus et al., 1995) have indicated that this could be the case.

Fluorescence in situ hybridization (FISH) with chromosome-specific DNA probes allows cytoge- netic information to be obtained rapidly from meta- phase and interphase nuclei. Hybridization with fluorescently labeled probes results in compact staining of the chromosomal region targeted by the DNA probe, and permits the copy number and location of the chromosome of interest to be iden- tified rapidly and precisely in numerous cells. In addition, this technique can provide cytogenetic information retrospectively, when cell nuclei have been prepared from paraffin-embedded old tumor samples (Stock et al., 1993).

We have examined 29 paraffin-embedded brain tumors for loss of loci on l p using double-target

Received June 16, 1995; accepted August 15, 1995. Address correspondence to Dr. Johji Inazawa, Department of Hy-

giene, Kyoto Prefectural University of Medicine, Kamigyo-ku, Ky- oto 602, Japan.

0 1995 Wiley-Lis, Inc

296 HASHIMOTO ET AL.

TABLE I. Summary of FISH Results and Microsatellite Analysis in 29 Brain Tumorsa

No. Diagnosis Note Results of FISH: CI I-5335IpUC- I .77 (%) % of deletionsb Microsatellite analysis

01 0 2 0 3 04 0 5 0 6 0 7 0 8 0 9 MI M2 M3 M4

A l A2 A3 A4 A5 A6 A7 A8 A9 AIO

GB I GB2 GB3 GB4 CN I CN2

0 II 0 II 0 II 0 II 0 II 0 II 0 II 0 II 0 II

OA II AOA 111 AOA 111 AOA 111

A II A II A II A II A II

AA 111 A II A II A II

AA 111 GB IV GB IV GB IV GB IV

CN CN

R (MI) irrad.

R. (Al) irrad.

R. (AS) irrad.

R. irrad.

R irrad.

R. (A4) irrad.

I13 (28) 112(5l)

I12 (25) I12 (25)

214 (41) I12 (46) 112 (66) I I2 (46)

213 (26)

112 (43)

2 3 (22)

I12 (43)

212 (33)

212 (39) 112 (31) 2 2 (55) 111 (42) 212 (59) 212 (42) 212 (69) 212 (68)

2/2 (60)

212 (42) 212 (30) 212 (63) 212 (70) 212 (65)

212 (53)

112 (33)

I 14 (26) 111 (15) I l l (26) 111 (24) 1/3(16)

2/3(13) 2 2 (24) I l l (14)

I12 (20)

012(1 I) 012 (20) 1/2(19) 313 (I 9)

313 (31) 212 (24)

212 (24) 0/2 ( 17) 313 (23) 111 (8) 111 (14)

414 (25) 212 (27) 313 (I 5)

112(21)

I l l (10)

313 (20) 111 (9) 2lI (I I) 313 ( 10)

112(12) 98 212 (14) 80

2/2(16) 82 2/3(13) 89

2/2(15) 77

111 (20) 77 113 (10) 94 2/3(12) 73 012 (8) 98

111 (8) 90

113(15) 83 414 ( I 9) 27

414 ( I 0) 8 I l l (12) 66

1/2(21) 71

312 (9) 14 I12 (7) 17 I I2 (6) 25

111 (9) 93

012 (9) 34

l12(10) 33

2l1 (8) 21 I13 (6) I I

414 ( I 0) 20 414 ( I 4) 34

I12 (5) I I

111 (18) 67

313 (7) 27

I12 (9) 15

LOH LOH LOH LOH LOH LOH

LOH (-)

LOH

LOH (-)

LOH (-)

~ ~~

‘Case numbers and diagnoses are given in the first two columns: 0, oligodendroglioma; M, mixed glioma; A. astrocytoma; GB, glioblastoma; CN. central neurocytoma; 0 II, oligodendroglioma WHO grade II; OA 11, oligoastrocytoma grade II; AOA 111. anaplastic oligoastrocytoma grade 111; A II, astrocytoma grade II; AA 111, anaplastic astrocytoma grade 111; GB IV, glioblastoma grade IV. R( ), recurrence (primary tumor); irrad., irradiated. The signal patterns and percentages of nuclei in the three major cell populations are indicated for each tumor. bSummed percentage includes all counted nuclei in any population considered to have deletions of I p36 see also Matsumura et al. (I 992).

FISH with pericentromeric (lq12) and subtelo- nervous system (Kleihues et al., 1993). Histo- meric (lp36) DNA probes. These experiments re- pathological evaluation carried out by neuropathol- vealed a high incidence of deletions in l p among ogists had revealed that all the specimens consisted oligodendrogliomas (100%) and mixed gliomas of at least 80% tumor cells. Twenty-three of the (75%). We confirmed these results using a micro- cases were primary tumors for which the patient satellite polymorphism from l p to detect loss of had received no preoperative irradiation or chemo- heterozygosity (LOH). therapy.

MATERIALS AND METHODS

Tumor Samples

Twenty-nine paraffin-embedded brain tumor samples were derived from 25 patients who had undergone surgery at the Hospital of Kyoto Prefec- tural University of Medicine or other hospitals be- tween 1977 and 1994 (Table 1). All samples were classified according to the World Health Organiza- tion (WHO) classification of tumors of the central

Tumor Cell Processing

To obtain a suspension of single cells, 100 pm sections were cut from paraffin-embedded tissues, dewaxed in xylene for 3 hr, immersed in ethanol for 1 hr, and dehydrated through ethanol series. After incubation at 37°C overnight in 1 x phos- phate-buffered saline (PBS), the tissues were minced with scissors in 1 x PBS and incubated at 37°C for 3-4 hr in a solution containing 200 kg/ml proteinase K. They were mechanically disaggre-

Ip DELETIONS IN OUGODENDROGLIAL TUMORS 291

gated with a homogenizer (Iuchi-Shoeido, Tokyo). T h e homogenized solution was filtered through a 100-pm nylon mesh, and the cell suspensions were smeared on a slide coated with poly-L-lysine using a centrifugal smear machine (Autosmear, Sakura Seiki Co., Japan). FISH

For the detection of deletions of lp , we used the repetitive DNA probe pUC-1.77 (Cooke and Hindley, 1979), specific for the pericentromeric re- gion (lq12), and the cosmid probe CI1-5335 (Ariyama et al., 1995), specific for the subtelomeric region (lp36). CI1-5335 was labeled with digoxi- genin (dig)-1 1 -dUTP (Boehringer, Manheim Gmbh, Manheim) and pUC-1.77 with biotin (bio)- 16-dUTP (Boehringer Manheim Gmbh) by nick translation. Bio- and dig-labeled probe solutions were mixed at a ratio of 7:2 (vh ) (pUC-1.77 and CI1-5335) and 1.0 pl of Cot-1 DNA (5 mg/ml; GIBCO BRL, Gaithersburg, MD) was added to 9.0 pi of the mixed probe solution. Then double- target FISH was performed as previously described (Inazawa et al., 1992; Matsumura et al., 1992). T h e slide preparations were denatured for 2-5 min.

Scoring of Interphase Nuclei

Under a Nikon FXA epifluorescence micro- scope, interphase nuclei were screened through a UV-2A filter (Nikon: exciter, 400-440 nm; barrier, 470 nm); only intact nuclei, no torn or overlapping ones, were evaluated. Signals of both probes were visualized simultaneously through a double band- pass filter (Omega Optical, Brattleboro, VT). Hy- bridization signals were counted in more than 200 interphase nuclei, and the numbers of pericentro- meric signals and subtelomeric signals were re- corded for each nucleus. T h e percentage of dele- tion was defined as the fraction of all nuclei scored having more pericentromeric signals than lp36 sig- nals and included all nuclei having only one peri- centromeric signal and one or no lp36-specific sig- nals, in the same manner as described elsewhere (Matsumura et al., 1992). Control hybridizations were performed simultaneously with paraffin-em- bedded normal brain tissues obtained from areas surrounding tumors from six patients who were not included in this study. DNA Isolation and Microsatellite Repeat Polymorphisms

Ten samples were examined for LOH of a mi- crosatellite repeat polymorphism on lp , for com- parison with FISH results (Table 1). Tumor DNA

was isolated from paraffin-embedded tissues and normal DNA from peripheral blood or from tumor- surrounding normal brain tissues, as previously de- scribed (Horii et al., 1994). T h e microsatellite marker studied was AFM200yf12, at the DlS234 locus (Gyapay et al., 1994). T h e polymerase chain reaction (PCR) was performed in 25-pl volumes of a mixture containing 1 x PCR buffer [6.7 mM Tris (pH 8.8), 16.6 mM (NH,),SO,, 6.7 pM EDTA, 10 mM p-mercaptoethanol], 10 pmol each of unla- beled primer and primer labeled with [gamma- 3zP]ATP (>5,000 Ci/mmol), 50 ng of DNA, 0.5 units of Taq polymerase (Ampli Taq, Perkin- Elmer, Norwalk, CT), 1.5 mM of each dideoxy- nucleotide, and 6.7 mM of MgCI,. T h e reaction was heated to 94°C for 5 min and then cycled 35 times; each cycle consisted of 30 sec at 55"C, 30 sec at 72"C, 30 sec at 94"C, and, for final elongation, 10 min at 72°C. T h e PCR products were denatured, electrophoresed in 6% polyacrylamide gels contain- ing 7.7 M urea and 32% formamide, and autorad- iographed for 1-8 hr on X-ray films.

RESULTS

Although we used old samples made in the 1970s or early 1980s, hybridization signals specific for the probes were delineated discernibly (Fig. 1) be- cause the time of treatment with proteinase K had been elongated (3-4 hr) and the slides were dena- tured for 5 min. Control hybridizations revealed that the average number of nuclei having deletions according to our criteria (see above) was 27% (SD = 2.7%) in populations of normal cells. Based on the mean + 3 SD, we could judge as a signif- icant population any preparation in which more than 35% of signal patterns indicated deletion.

A summary of the results is shown in Table 1, where we indicate the percentages of nuclei show- ing l p deletions in the cell populations from each tumor that showed the three most common signal patterns. In the 9 pure oligodendrogliomas, we found l p loss in all of the samples (loo%), whereas of the 4 mixed gliomas (1 oligoastrocytoma and 3 anaplastic oligoastrocytomas), 3 showed deletion of lp36 (75%). We also observed a tendency toward heterogeneity of signal patterns within individual tumors.

In contrast to oligodendrogliomas, our series of astrocytic tumors showed a low incidence of dele- tions of this region. Except in three samples from two patients (A2, A4, and GB1 in Table l), no deletions were statistically significant. T h e popu- lations of deletion bearing cells from tumors A2, A4, and GB1 accounted for 66%, 71%, and 67% of

298 HASHIMOTO ET AL.

Figure I. Results of double-target FISH using probes specific for I q I 2 (green) and I p36 (red) to interphase nuclei prepared from paraffin- embedded tumor samples. Representative patterns of hybridization sig- nals exhibiting loss of lp36 are seen in 01 (a), 0 6 (b,d,e). 0 7 (f). M3 (8). and A2 (k). The number of signals for I p36 (red)/l q I 2 (green) could be detected unambiguously as 1/3 (a), 2/3 (b), 1/2 (d). 2/4 (f), 2/3 (g). and I12 (k). indicating that the lp36 band is missing from at least one chro-

their respective totals. Two central neurocytomas showed no deletions of lp .

All of the 10 cases tested for heterozygosity at the DlS234 locus were informative. We detected LOH at this locus in 7 samples but not in 3 others, in perfect accord with the FISH results for these tumors (Fig. 2, Table 1).

DISCUSSION

T h e technique of FISH has provided new in- sights into molecular cytogenetics; not only meta-

mosome l in these nuclei. In e, signals specific for l p36 and l q I 2 were detected as single spots, strongly suggesting loss of a whole chromosome I homologue in this nucleus. In two cases of M4 (h,i) and A7 (I). the patterns of I p36/ I q I2 reveal 2/2, 4/4, and 2/2. indicating that neither homologue of chromosomes I has lost I p36 in these nuclei. Panels c and j show DAPI-stained nuclei of b and i, respectively.

phase chromosomes but also interphase nuclei can now be used to detect allelic deletions (Matsumura et al., 1992; Stock et al., 1993). In the present study, we were able to display clear FISH signals even in old samples obtained from formalin-fixed, paraffin-embedded tissues. One of the advantages of FISH is its consistently informative nature, even for individuals who are homozygous at the locus of interest. FISH results can also reflect genetic het- erogeneity of chromosomes within a tumor mass, as indicated in Table 1. Deletion of the subtelomeric

Ip DELETIONS IN OUGODENDROGLIAL TUMORS 299

Figure 2. Examples of LOH at the D I S234 locus. Case numbers are given at the cop of each panel. T. tumor D N A N, normal (constitutional) DNA. Arrows indicate allelic deletions in cases 07, 09, M2. and A2. which are compatible with the results obtained by FISH (Table I).

marker locus on l p had occurred in more than 35% of the counted nuclei (mean: 86%) in all of the 7 cases showing LOH a t DlS234, but in less than 35% (mean: 18%) of the nuclei in each of the 3 tumors that had retained both alleles. These facts suggest that FISH could be used to detect loss of genetic information as reliably as LOH studies, without suffering harm from the inevitable admix- ture of normal cells that may confound L O H in- vestigations of tumors.

We investigated l p deletions in brain tumors be- cause chromosomal aberrations and LOH on chro- mosome arm l p had been demonstrated in a vari- ety of tumor types, including neuroblastomas (Brodeur and Fong, 1989) and medulloblastomas (Bigner et al., 1988a) which arise from neuroecto- dermal tissues. In addition, several cytogenetic studies had revealed structural aberrations involv- ing the short arm of chromosome 1 in glial tumors, e.g., astrocytomas, glioblastomas (Bigner et al., 1988b; Jenkins et a]., 1989; Ransom et al., 1992a),

and oligodendrogliomas (Jenkins et al., 1989; Ran- som et a]., 1992b). However, with regard to non- oligodendrogliomas, previous LOH studies did not show frequent loss of genetic information from chromosome 1 (James et al., 1988; Venter and Thomas, 1991; Ransom et al., 1992b).

Recently, Reifenberger et al. (1994) reported frequent loss of material from the long arm of chro- mosome 19 by extensive allelotyping in a series of oligodendroglial tumors. They also found LOH for loci on l p in 7 of 8 oligodendrogliomas (grade 11) and in 3 of 16 mixed gliomas (grades I1 and 111) using 14 restriction fragment length polymorphism markers spanning the whole arm. One of the oli- godendrogliomas (grade 11) showed LOH only dis- tal to the NGFB gene (lpl3-pter), whereas the six other tumors had larger deletions.

Our data, in comparison, showed higher rates of l p deletion both for oligodendrogliomas (9 of 9 cases) and mixed gliomas (3 of 4). These results strongly suggest that at least one tumor suppressor gene responsible for oligodendroglial tumors, in- cluding mixed gliomas, lies on the short arm of chromosome 1. It is noteworthy that without ex- ception, loss of the lp36 region was involved in the pure oligodendrogliomas. Whether this putative suppressor locus is the same as the one implicated in other neuroectodermal tumors remains to be elu- cidated, but our results lead us to believe that its locus is in lp36. Our data indicating a low inci- dence of deletions at lp36 in astrocytic tumors and central neurocytomas are compatible with previous reports concerning astrocytomas (James et al., 1988; Venter and Thomas, 1991; Ransom et al., 1992b).

ACKNOWLEDGMENTS

We thank Dr. H. Murata for technical advice and our neurosurgical colleagues, Dr. S. Taketomo at Saiseikai Kyoto Hospital, Dr. K. Takemi at Ky- oto 2nd Red Cross Hospital, Dr. K. Kakita at Ky- oto 1st Red Cross Hospital, Dr. M. Fujimoto at Saiseikai Shiga Hospital, and Dr. Y. Horikawa at Kyoto Prefectural Yosanoumi Hospital, for man- agement of patients and provision of tissue sam- ples. We also thank Ms. Chizu Yoshimura for tech- nical assistance.

This work was supported by a Grant-in-Aid from the Ministry of Education, Science and Culture, and the Ministry of Health and Welfare, Japan.

REFERENCES

Ariyama T, Inazawa 1. Ezaki T, Nakamura Y, Horii A, Abe T (1995) High-resolution cytogenetic mapping of the short arm of chromo- some 1 with newly isolated 41 1 cosmid markers by fluorescence

300 HASHIMOT0 ET AL.

in situ hybridization: T h e precise order of 18 markers on lp36.1 on prophase chromosomes and “stretched” DNAs. Genomics 25:

Bailey P, Bucy PC (1929) Oligodendrogliomas of the brain. J Pathol Bacteriol 32735-751.

Bigner SH, Mark J, Friedman HS, Biegel JA, Bigner DD (1988a) Structural chromosomal abnormalities in human medulloblas- toma. Cancer Genet Cytogenet 30:91-101.

Bigner SH, Mark J, Burger PC, Mahaley S, Bullard DE, Muhlbaier LH, Bigner DD (1988b) Specific chromosomal abnormalities in malignant human gliomas. Cancer Res 48:405-411.

Brodeur GM, Fong C T (1989) Molecular biology and genetics of human neuroblastoma. Cancer Genet Cytogenet 41: 153-174.

Celli P, Nofrone I, Palma L, Cantore G, Fortuna A (1994) Cerebral oligodendroglioma: Prognostic factors and life history. Neurosur- gery 351018-1035.

Collins VP, James C D (1993) Gene and chromosomal alterations associated with the development of human gliomas. FASEB J

Cooke HJ, Hindley J (1979) Cloning of human satellite I11 DNA: Different components are on different chromosomes. Nucleic Ac- ids Res 6:3177-3197.

Gyapay G, Morissette J, Vignal A, Dib C, Fizames C (1994) The 1993-94 Genethon human genetic linkage map. Nature Genet

Horii A, Han HJ, Shimada S, Yanagisawa A, Kato Y, Ohta H, Yasui W, Tahara E, Nakamura Y (1994) Frequent patients with multi- ple primary cancers. Cancer Res 54:3373-3375.

Inazawa J, Ariyama T , Abe T (1992) Physical ordering of three polymorphic DNA markers spanning the regions containing a tu- mor suppressor gene of renal cell carcinoma by three-color fluo- rescent in situ hybridization. Jpn J Cancer Res 83:1248-1252.

James CD, Carlbom E, Dumanski JP, Hansen M, Nordenskjold M, Collins VP, Cavenee WK (1988) Clonal genomic alterations in glioma malignancy stages. Cancer Res 485546-5551.

Jenkins RB, Kimmel DW, Moertel CA, Schultz CG, Scheithauer

114-123.

7~926-930.

7~246-339.

BW, Kelly PJ, Dewald GW (1989) A cytogenetic study of 53 human gliomas. Cancer Genet Cytogenet 39:253-279.

Kleihues P, Burger PC, Scheithauer B (1993) Histological Typing of Tumors of the Central Nervous System. Berlin: Springer-Verlag.

Kraus JA, Koopmann J, Kaskel P, Maintz D, Brandner S, Schramm J, Louis DN, Wiestler OD, von Diemling A (1995) Shared allelic losses on chromosome l p and 19q suggest a common origin of oligodendroglioma and oligoastrocytoma. J Neuropathol Exp Neurol 54:91-95.

Matsumura K, Kallioniemi A, Kallioniemi 0, Chen L, Smith HL, Pinkel D, Gray J, Waldman FM (1992) Deletion of chromosome 17p loci in breast cancer cells detected by fluorescence in situ hybridization. Cancer Res 523474-3477.

Ransom DT, Ritland SR, Moertel CA, Dahl RJ, OFallon JR, Scheithauer BW, Kimmel DW, Kelly PJ, Olopade 01, Diaz MO, Jenkins RB (1992a) Correlation of cytogenetic analysis and loss of heterozygosity studies in human diffuse astrocytomas and mixed oligoastrocytomas. Genes Chromosom Cancer 5:357-374.

Ransom DT, Ritland SR, Kimmel DW, Moertel CA, Dahl RJ, Scheithauer BW, Kelly PJ, Jenkins RB (1992b) Cytogenetic and loss of heterozygosity studies in ependymomas, pilocytic astrocy- tomas, and oligodendrogliomas. Genes Chromosom Cancer 5348-356.

Reifenberger J, Reifenberger G, Liu L, James CD, Wechsler W, Collins VP (19%) Molecular genetic analysis of oligodendroglial tumors shows preferential allelic deletions on 19q and lp. Am J Pathol 145: 1175-1 190.

Russell DS, Rubinstein LJ (1989) Tumors of central neuroepithelial origin. In: Pathology of Tumors of the Nervous System. 5th Ed. London: Edward Arnold, pp 172-173.

Stock C, Ambros IM, Mann G, Gadner H, Amann G, Ambros PF (1993) Detection of lp36 deletions in paraffin sections of neuro- blastoma tissues. Genes Chromosom Cancer 6 1-9.

Venter DJ, Thomas DG (1991) Multiple sequential molecular ab- normalities in the evolution of human gliomas. Br J Cancer 63: 753-757.