Oncogene (2000) 19, 4302 ± 4307 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc

SHORT REPORT Identi®cation and characterization of a 500-kb homozygously deleted region at 1p36.2-p36.3 in a neuroblastoma cell line

Miki Ohira1,11, Hajime Kageyama1,11, Motohiro Mihara1, Shigeyuki Furuta1, Taiichi Machida1, Tomotane Shishikura1, Hajime Takayasu1, Ashraful Islam1, Yohko Nakamura1, Masato Takahashi1, Nobumoto Tomioka1, Shigeru Sakiyama1, Yasuhiko Kaneko2, Atsushi Toyoda3, Masahira Hattori3, Yoshiyuki Sakaki4, Misao Ohki5, Akira Horii6, Eiichi Soeda7, Johji Inazawa8, Naohiko Seki9, Hidekazu Kuma10, Iwao Nozawa10 and Akira Nakagawara*,1

1Division of Biochemistry, Chiba Cancer Center Research Institute, Chiba 260-8717, Japan; 2Department of Cancer Chemotherapy, Saitama Cancer Center Hospital, Saitama 362-0806, Japan; 3RIKEN Genomic Sciences Center, Sagamihara, Kanagawa 228-8555, Japan; 4Human Genome Center, Institute of Medical Science, University of Tokyo, Minato-ku, Tokyo 108-8639, Japan; 5Cancer Genomics Division, National Cancer Center Research Institute, Chuoh-ku, Tokyo 104-0045, Japan; 6Department of Molecular Pathology, Tohoku University School of Medicine, Sendai 980-8575, Japan; 7RIKEN Gene Bank, Tsukuba, Ibaraki 305-0074, Japan; 8Department of Molecular Cytogenetics, Division of Genetics, Medical Research Institute, Tokyo Medical and Dental University, Bunkyo-ku, Tokyo 113-8519, Japan; 9Biological Technology Laboratory, Helix Research Institute, Inc., Kisarazu, Chiba 292-0812, Japan; 10New Technology Development Department, Central Research Center, Hisamitsu Pharmaceutical Co., Inc., Tsukuba Research Laboratories, Tsukuba, Ibaraki 305-0856, Japan

Loss of heterozygosity of the distal region of chromo- discovery of additional candidate genes. Oncogene some 1p where tumor suppressor gene(s) might harbor is (2000) 19, 4302 ± 4307. frequently observed in many human cancers including neuroblastoma (NBL) with MYCN ampli®cation and Keywords: neuroblastoma; homozygous deletion; 1p36; poor prognosis. We have identi®ed for the ®rst time a tumor suppressor gene homozygously deleted region at the marker D1S244 within the smallest region of overlap at 1p36.2-p36.3 in two NBL cell lines, NB-1 and NB-C201 (MASS-NB- SCH1), although our genotyping has suggested the Neuroblastoma (NBL) is one of the most common possibility that both lines are derived from the same pediatric tumors and originates from the sympatho- origin. The 800-kb PAC contig covering the entire region adrenal lineage derived from the neural crest (Bolande, of homozygous deletion was made and partially 1974). It shows various clinical behaviors ranging from sequenced (about 60%). The estimated length of the spontaneous regression to malignant progression. deleted region was 500 kb. We have, thus far, identi®ed Genetic aberrations frequently found in NBL are loss six genes within the region which include three known of the distal part of the short arm of chromosome 1 genes (DFF45, PGD, and CORT) as well as three other (1p), ampli®cation of the MYCN oncogene and 17q genes which have been reported during processing our gain, all of which are associated with an unfavorable present project for the last 3‰ years (HDNB1/UFD2, prognosis (Brodeur et al., 1984; Caron, 1995). 1p loss KIAA0591F/KIF1B-b, and PEX14). They include the also occurs frequently in other tumors such as genes related to apoptosis, glucose metabolism, ubiquitin- colorectal cancer, hepatocellular cancer, breast cancer, proteasome pathway, a neuronal microtubule-associated pheochromocytoma, and melanoma (Schwab et al., motor molecule and biogenesis of peroxisome. At least 1996). Fluorescence in situ hybridization (FISH) and three genes (HDNB1/UFD2, KIAA0591F/KIF1B-b, and loss of heterozygosity (LOH) analysis have been PEX14) were di€erentially expressed at high levels in applied to de®ne the deleted region of 1p in NBL in favorable and at low levels in unfavorable subsets of many laboratories, suggesting that at least three primary neuroblastoma. Since the 1p distal region is putative NBL suppressor loci might be present at the reported to be imprinted, those di€erentially expressed distal region of 1p, especially at 1p36.1-pter (Martins- genes could be the new members of the candidate NBL son et al., 1995; White et al., 1995; Amler et al., 1995; suppressor, although RT-PCR-SSCP analysis has de- Ichimiya et al., 1999), 1p36.1-p35 (Cheng et al., 1995) monstrated infrequent mutation of the genes so far and 1p34-p32 (Takeda et al., 1994; Schleiermacher et identi®ed. Full-sequencing and gene prediction for the al., 1994). With respect to 1p36, about 7 cM region region of homozygous deletion would elucidate more between markers D1S214 and D1S244 is a smallest detailed structure of this region and might lead to region of overlap (SRO) of deletion by taking into account all LOH studies (Martinsson et al., 1995; White et al., 1995). To date, several candidate tumor suppressor genes mapped to the distal 1p have been *Correspondence: Akira Nakagawara, Division of Biochemistry, reported. They include CDC2L1 (Lahti et al., 1994), Chiba Cancer Center Research Institute, 666-2 Nitona, Chiba 260- p73 (Kaghad et al., 1997), TNFR2 (White et al., 1993), 8717, Japan 11The ®rst two authors contributed to this work RIZ (He et al., 1998), DAN (Enomoto et al., 1994), Received 11 April 2000; revised 3 July 2000; accepted 6 July 2000 FGR (Dracopoli et al., 1988), EXTL (Wise et al., 1997), Homozygous deletion at 1p36.2 in a human neuroblastoma cell M Ohira et al 4303 and hRAD54 (Carling et al., 1999). However, their and PGD) were also lost in both cell lines (Figure 1b). mutations so far examined are infrequent in NBLs. We then PCR-screened DNA from 180 primary NBLs Detection of a homozygous deletion (HD) in tumor at D1S244. However, it failed to ®nd clear HD cells is the most ecient way to isolate tumor-related probably because of contamination of the stromal cell genes. We screened 29 NBL cell lines for HD by PCR components. using 15 polymorphic markers (D1S80, p73, D1S214, To de®ne the homozygously deleted region, we D1S548, D1S160, D1S508, D1S450, D1S244, TNFR2, assembled a sequence-ready clone contig that spanned D1S228, D1S507, D1S483, D1S2843, D1S458 and the region of deletion. We developed a contig using P1 FGR) and found two cell lines (NB-1 and MASS-NB- arti®cial chromosomes (PACs) containing SGC30343, SCH1) with deletion at a D1S244 marker. MASS-NB- D1S2021, D1S1768, PGD, D1S244, and D1S2736, the SCH1 (abbreviated to NB-C201 in this paper) was a markers that de®ned the HD. The ends of the isolated recently established cell line from the primary tumor PAC clones were sequenced to develop new sequence found by mass screening at the age of 6 months tagged sites (STSs) for further screening of clones and (Hiraiwa et al., 1997). Figure 1a shows the results of for securing the overlaps between the clones (Figure 2). LOH in NB-C201 for the markers of D1S548, D1S244, The insert sizes of the PAC clones constituting the D1S507 and FGR, among which only D1S244 was minimum tiling path of the contig were determined by homozygously deleted. Further analysis using markers using restriction digestion with NotI and/or SalI around the D1S244 locus revealed that additional ®ve followed by pulsed-®eld gel electrophoresis (PFGE). markers (D1S2021, D1S1768, D1S2736, SGC30343, According to the PFGE data, the size of the contig was estimated to be approximately 800 kb. Using the PAC end sequences, both breakpoints in the two cell lines were mapped between dJ1028O13-SP6 and dJ371E1- SP6, and D1S2736 and dJ142A6-T7, as shown in Figure 2. Accordingly, the HD region represented approximately 500 kb or less. The HD in NBL has been uncommon. Only the neuro®bromatosis-1 (NF1) gene on chromosome 17q11.2 (The et al., 1993; Martinsson et al., 1997) and the p16 gene on 9q21 (Diccianni et al., 1996) have been reported to be homozygously lost in NBL. Our precise analysis revealed that the breakpoints of the HD region in two cell lines were identical in cDNA sequences, within the HDNB1/UFD2 and PEX14 genes (data not shown). Both genes were fused to produce a chimeric transcript, suggesting that the deletion occurred in the very limited genomic region in both cell lines and that the deletion hotspot might exist in this region. However, it raised the possibilities that these two cell lines were originated from the same patient or cross-contaminated. The former possibility was excluded by referring the original reports (Im- ashuku et al., 1973; Hiraiwa et al., 1997). To clarify the latter point, we re-obtained the original cell stock and con®rmed that it lost the D1S244 region. In the previous paper, Hiraiwa et al. (1997) performed genotyping and con®rmed establishment of a new cell line. The cell morphology and responsiveness to Figure 1 Identi®cation of homozygous deletion in NBL cell retinoic acid in MASS-NB-SCH1/NB-C201 line were line(s). (a) The results of LOH at D1S548, D1S244, D1S507, and di€erent from those in NB-1 (unpublished data). FGR in case 201. Left: DNA from lymphocyte, right: DNA from Nevertheless, our genotyping in several loci displayed NB-C201 tumor. LOH at the region of chromosome 1p35-p36 was performed using 29 NBL cell lines and 206 primary NBL very similar patterns in both cell lines. These suggested samples (data not shown). During the analysis, D1S244 revealed a possibility that the origin of NB-1 and MASS-NB- to be homozygously deleted in case 201 and NB-1 cell line, SCH1 cells was the same, though that should be although these two cell lines have a possibility to be derived from further pursued. the same origin (see text). The NB-C201 cell line was a gift from Dr M Hamazaki and Dr H Hiraiwa (Department of Pathology, In the next step, gene identi®cation was performed Shizuoka Children's Hospital). It was established from the NBL by cDNA library screening and by rough genome tissue obtained from a right paravertebral region of the sequencing using six minimum PAC clones (dJ587C9, mediastinum of a 7-month-old NBL patient who was found by dJ371E1, dJ694C14, dJ756A17, dJ266I16, and the mass screening program (Hiraiwa et al., 1997). On the other dJ345P21). We identi®ed at least six genes in this hand, NB-1 cell line was derived from a lesion metastatic to cervical lymph node obtained from a 2-year-old patient with NBL region. The EST marker SGC30343 revealed to be a (Imashuku et al., 1973). (b) Detection of the homozygous deletion part of a 4.6 kb novel gene, which we named HDNB1 at six loci. As an internal control, NGFb primers on 1p13 were (a homozygously deleted gene in neuroblastoma 1, used in the same reaction. Template DNAs for PCR reaction are accession no. AB028839) after cloning by screening the SY5Y (lanes 1, 4, 7, 10, 13, 16), NB-1 (lanes 2, 5, 8, 11, 14, 17), fetal brain cDNA library. Recently, it was shown to and NB-C201 (3, 6, 9, 12, 15, 18). The ampli®ed product of the NGFb is shown by an arrow, and the product from each marker correspond to a human homolog of yeast ubiquitin is shown by an arrowhead fusion degradation 2 (UFD2, locus designation:

Oncogene Homozygous deletion at 1p36.2 in a human neuroblastoma cell M Ohira et al 4304

Figure 2 Genome structure of the homozygously deleted region. To cover the HD region, 40 PACs were isolated and aligned in an 800-kb contig. The PAC library consisting of 1440 multiwell microtiter plates was provided by Dr Pieter de Jong (Rosewell Park Institute, Bu€alo, NY, USA) (Ioannou et al., 1994). The end fragments of the PAC clones were sequenced directly from the clone DNA with P6S and T7 vector primers using an ALF ExpressTM DNA sequencer (Amersham Pharmacia Biotech). On the PAC end, the direction of the insert is noted: N; T7 end, S; SP6 end. New STSs for genomic walking were designed from nonrepetitive end sequences. De®nition of the deletion ends was performed by PCR with these primers using genomic DNAs of NB-1, NB-C201, and human placenta. The insert sizes of the PAC clones were determined by restriction digestion followed by pulsed-®eld gel electrophoresis (PFGE). NotI-digested PAC DNA was separated in 1% agarose gel using a CHEF DRII pulsed-®eld gel electrophoresis apparatus (BIORAD). The region deleted in NB-1 and NB-C201 cell lines (approximately 500-kb long) was shown on the top and also shaded in the map. The distal and proximal breakpoints were mapped between dj1028O13-SP6 and dJ371E1- SP6, and between D1S2736 and dJ142A6-T7, respectively. For gene identi®cation, cDNA libraries were screened with PAC DNAs or an EST fragment containing SCG30343. In addition, rough genome sequencing of the PAC clones was also performed as described previously (Hattori et al., 1997). cDNA libraries we used were derived from human fetal brain purchased from Clontech or a human NBL cDNA library we constructed from primary NBLs with favorable prognosis. cDNA clones were extended by RACE PCR using fetal brain Marathon-readyTM cDNA as a template (Clontech). The six genes mapped to the HD region were illustrated with transcription directions. The transcriptional direction of each gene was determined using primers that can work in genomic DNAs. Genomic primer sequences were as follows: PEX14 (5')-genome; (f)GATATGGCCTTCCAGCAGTC; (r)ATATG- AGGTGAGGGGGCTGG, PEX14 (3')-genome; (f)GAGGAGGAGAAGAGGGAGGA; (r)AGAGAGGGAGGCAGGGTATG, PGD(5')-genome; (f)AGCCCAGGGTTGGAGAAC; (r)TTACAGCCATGGGCCACT, PGD(3')-genome; (f)CACGAAGAACG- CATGAGG; (r)TCCCCAGCCATGTCTGTAA, CORT(5')-genome; (f)TGCCATGAAGGGTGAAGG; (r)AAGGAGGCATTTC- CAGCA, and CORT(3')-genome; (f)GGGTGCTTTTTGTGCTGAA; (r)CCATTCGATCACACACCATT

UBE4B) which catalyzes ubiquitin chain assembly in et al., 2000). We recently found that this gene might be conjugation with E1, E2 and E3 in proteolysis a human homolog of a mouse kinesin-superfamily pathways (Johnson et al., 1995; Pukatzki et al., 1998; member Kif1b-b (submitted only in a database, Koegl et al., 1999). The E4 activity is linked to cell accession no. AB023656), whose product had an N- survival under stress conditions in yeast, indicating that terminal motor domain homologous to Kif1b (Nanga- eukaryotes utilize E4-dependent proteolysis pathways ku et al., 1994) and a novel C-terminal cargo-binding for multiple cellular functions. HDNB1 cDNA we domain homologous to Kif1a (Takemura et al., 1996). isolated was likely to be a short form (an alternatively There may be additional alternatively spliced tran- spliced transcript) of the human UFD2 (6.0 kb, scripts in the HD region from the analysis of the accession no. NM004788). genome sequence (unpublished data), so the genome By direct cDNA library screening with PAC DNAs, structure of these kinesin-related genes should be we found that a kinesin-related gene KIAA0591 further elucidated. DFF45/ICAD regulates the DNA (Nagase et al., 1998) resided in PAC clone dJ756A17 fragmentation activity of DFF40/CAD in the apoptotic and that three genes, phosphogluconate dehydrogenase process. The intact structure/conformation of DFF45 (PGD) (Tsui et al., 1996), DNA fragmentation factor is essential for functioning as a speci®c chaperone and 45 (DFF45, locus designation: DFFA)/inhibitor of assisting in the proper synthesis of active DFF40 caspase-activated DNase (ICAD)(Liuet al., 1997), (Enari et al., 1998). Although the signi®cance of DNA and neuron-speci®c preprocortistatin (CORT/CST) fragmentation in apoptosis has not yet been well (Fukusumi et al., 1997; de Lecea et al., 1997) resided clari®ed, distortion of the DNA fragmentation system in dJ266I16 (Figure 2). DFF45/ICAD sequence was may negatively in¯uence on the apoptotic process. also found in the PAC end sequence dJ373D24-SP6. PGD is ubiquitously expressed, and its product is a The KIAA0591 had previously been isolated in a 5'- catalytic enzyme that works in the hexose monopho- truncated form by KAZUSA DNA Research Institute sphate shunt which supplies NADPH and pentoses for (Nagase et al., 1998), and we extended the cDNA to a the nucleic acid metabolism. CORT/CST is a pre- full-length form (7.7 kb) by screening a human sumptive neuropeptide that shares 11 of 14 amino substantia nigra cDNA library (named KIAA0591F as acids with somatostatin. CORT/CST mRNA is ex- full-length KIAA0591, accession no. AB017133) (Nagai pressed in a subset of GABAergic interneurons in the

Oncogene Homozygous deletion at 1p36.2 in a human neuroblastoma cell M Ohira et al 4305 cortex and hippocampus (Fukusumi et al., 1997; de Lecea et al., 1997). This gene could also be one of the good candidates of NBL-related gene. However, Martinsson et al. have recently reported that there is no signi®cant alteration in this gene in NBLs (Fifth International Human Chromosome 1 Workshop '99 at Sanger Centre, UK). Rough genome sequencing followed by homology searching also detected the above ®ve genes, and an additional gene, PEX14 (Will et al., 1999) was identifed in a PAC clone dJ345P21 (Figure 2). The PEX14 gene product is a central component of the peroxisomal protein import machinery (Will et al., 1999). The C- terminus of the protein exposes to the cytoplasm and binds the receptor for the peroxisomal targeting signal 1 (PTS1), and the N-terminus might protrude into the peroxisomal lumen in yeast. Thus, the genes within the HD region could functionally be tumor suppressors. For precise mapping of these six genes on the PAC clones, 5' and 3'-end primers that should work in genomic DNA were designed. Each location and transcriptional direction of the genes were determined by PCR using PAC DNAs as indicated in Figure 2. The Sanger Centre mapped several hundreds of ESTs and genes on chromosome 1 by radiation hybrid Figure 3 The expression status of the genes in the HD region in mapping (Genemap '99). Of these, ESTs corresponding NBL. (a) Expression of the ®ve genes in 7 NBL cell lines and HeLa cell. RT-PCR was performed using cDNA primers and control to HDNB1/UFD2, PGD, DFF45/ICAD and PEX14 GAPDH primers. Primers were designed to work only in the cDNAs were positioned within the interval de®ned by D1S450 but not in genomic DNAs. Primer sequences were as follows: and D1S2667 that agreed with our localization. HDNB1(5'); (f)AGTAGTCTGTGGGGCGACTG; (r)GTTGGC- However, KIAA0591F/KIF1B-b and CORT/CST genes TGAGAGGTCTGTCC, HDNB1(3'); (f)AACGCTGAGACCT- were not localized, suggesting that both genes were new GAAAGGA; (r)GCAGCAAAATGTCCAAGGTT, PEX14(5'); (f)GTCCTCGGAGCAGGCAGAG; (r)ATATGAGGTGAGGG- entries to the map. In addition, three genes KIAA0911, GGCTGG, PEX14(3'); (f)ATCCCAGTCAAGTCACCGTC; PIK3CD, and SCYA5 were positioned in this region (r)ATCCTCCTCATCCTCCTCGT, KIAA0591F; (f)CAGCAA- (between D1S450 and D1S2667, Genemap '99), but we GATGAGATGGAGGA; (r)GCAGTGACTTCTCTAGACTA- localized these genes outside the 500-kb deleted region GG, PGD; (f)TCTTGGCCAATGAGGCAAAG; (r)ATGAAAT- CATCCACAGCTTG, and DFF45/ICAD; (f)CTGCACCAACAA- by PCR using NB-1 and NB-C201 cell lines (data not TATCCC; (r)GTAGAGACAGGGTTTCAC. 3' region of the shown). Thus, HDNB1/UFD2, KIAA0591F/KIF1B-b, HDNB1/UFD2 and 5' region of the PEX14 are not transcribed in PGD, CORT/CST, DFF45/ICAD, and PEX14 were NB-C201 and NB-1. (b) Semi-quantitative RT-PCR in favorable located in the 500 kb HD region. and unfavorable NBLs. cDNAs synthesized from eight favorable We carried out RT-PCR using cDNAs prepared tumors and from eight unfavorable tumors with MYCN ampli®ca- tion were used. Those cDNAs were at ®rst ampli®ed with Cy5- from eight NBL cell lines including NB-1 and NB- labeled GAPDH primers in 25 cycles and the amounts of the PCR C201. Primers were synthesized in at least two products were measured by ALF ExpressTM sequencer and positions in each gene (close to the 5'-and3'-ends). normalized. The band intensities were measured by Intelligent Expression of three central genes (KIAA0591F/KIF1B- Quanti®er (Bio Image). The ratios of the average in favorable and unfavorable tumors were 3.2 : 1 (HDNB1/UFD2), 3.6 : 1 b, PGD, and DFF45/ICAD) were absolutely lost in (KIAA0591F/KIF1Bb), 3.2 : 1 (DFF45/ICAD), 1.6 : 1 (PGD) and both NB-1 and NB-C201, while obvious expression 2.2 : 1 (PEX14) was seen in other control cell lines (Figure 3a). Expression of CORT/CST was also lost in NB-1 and NB-C201 and was generally low in other NBL cell lines (data not shown). The transcripts of HDNB1/UFD2 primary NBLs and 21 NBL cell lines for DFF45/ICAD. and PEX14 were found to be truncated in 3'- and 5'- So far, no base substitution was found in KIAA0591F portion of the genes, respectively, in the two cell lines cDNA (the motor domain was not screened). Three (Figure 3a). The breakpoints in the cDNA sequences base changes were found in DFF45/ICAD, but without were mapped between n.t. 719 and 720 in the former amino acid substitutions as described elsewhere (Fur- (just downstream of a putative translation start site) uta et al., submitted). PEX14 had several mutations, and between n.t. 173 and 174 in the latter. These four silent mutations and a missense mutation (codon results suggested that HDNB1/UFD2 and PEX14 320) from arginine to lysine in two NBL cell lines would not have their normal cellular cellular forms in (OAN, LA-N-6, see Table 1). There was an amino acid NB-1 and NB-C201. substitution in HDNB1/UFD2 from asparagine to According to the Knudson's two-hit theory on the serine (codon 593) in a NBL cell line (LHN cell). tumor suppressor genes (Knudson, 1971), mutation Thus, mutations in the genes within the HD region search for all the genes in this region is necessary. To were infrequent. survey mutations in those genes, RT-PCR-SSCP The epigenetic regulation of the tumor suppressor analyses were performed (see Table 1). We surveyed genes is important (Jones and Laird, 1999). In 23 primary NBLs and 17 NBL cell lines for HDNB1/ addition, 1p36 is the region which is imprinted (Caron UFD2, KIAA0591F/KIF1B-b, PGD and PEX14,and26 et al., 1993). We examined mRNA expression of the six

Oncogene Homozygous deletion at 1p36.2 in a human neuroblastoma cell M Ohira et al 4306 Table 1 Summary of the six genes identi®ed in the HD region Expression between favorable (F) and Expression in unfavorable (UF) Genes Accession no. Mutation analysis Base changes neuroblastoma neuroblastoma

HDNB1/UFD2 AB028839 23 primary NBLs Asn593Ser(AAT?AGT) (LHN cells) + F4UF 17 NBL cell lines KIAA0591F/KIF1B-b AB011163 23 primary NBLs *None + F4UF 17 NBL cell lines PGD U30255 23 primary NBLs Asp246Asn(GAT?AAT) (cases 96, 165, + F=UF NMB, KAN, GAMB) 17 NBL cell lines Ser479Ser(TCG?TCA) (cases 83, 205, SK-N-BE) CORT AF013252 Not determined 7*7/+ DFF45/ICAD U91985 26 primary NBLs Ser118Ser(AGC?AGT) (CHP 134) + F4UF 21 NBL cell lines Glu202Glu(GAG?GAA) (NBTU-1) Gln298Gln(CAG?CAA) (cases 90, 231, 266) PEX14 AB017546 23 primary NBLs Glu26Glu(GAG?GAA) (cases 101, 130, 134) + F4UF 17 NBL cell lines Phe52Phe(TTT?TTC) (case 83) Val256Val(GTG?GTA) (NMB) Arg320Lys(AGG?AAG) (OAN/LA-N-6) Asp338Asp(GAC?GAT) (cases 45, 96, 107 121, 205)

*Mutation search in the motor domain of the KIAA0591F/KIF-1B-b has not been done yet. For the mutation search, RT ± PCR ± SSCP was performed as previously described (Sunahara et al., 1998)

genes in 16 primary NBLs with favorable or unfavor- region was reported to be imprinted, those genes able prognosis by semi-quantitative RT-PCR. Expres- di€erentially expressed could be the candidate NBL sion of all genes except CORT/CST were detectable in suppressor(s) regulated by epigenetic modi®cation. all samples (Table 1). HDNB1/UFD2, DFF45/ICAD Nevertheless, as RT-PCR-SSCP analysis showed in- and KIAA0591F/KIF1B-b were expressed at signi®- frequent mutation for the genes so far identi®ed, we cantly higher levels in favourable stage I tumors than need to search for additional candidate genes by full- in unfavorable tumors with MYCN ampli®cation sequencing for the entire region of the HD. An (Figure 3b). Although the reduction of gene dosage extensive search for additional unknown genes within caused by allelic loss might at least partly a€ect the the region is in progress by gene prediction. expression levels, there might also be an epigenetic modi®cation. Of interest, KIAA0591F/KIF1B-b was expressed 3.6 times higher in favorable NBL than in Acknowledgments unfavorable one. CORT/CST, whose expression in The authors thank Drs S Ichimiya, Y Nimura, and N NBL is very low, could also be one of the candidate Takada for helpful discussions, to Drs M Nagai, M Naka, genes, although it was dicult to compare levels of T Iijima and S Saitoh for experimental support, to Ms A expression between normal and malignant tissues Morohashi, N Sugimitsu, E Kojima for technical assis- because of a diculty in obtaining normal tissue. tance. We are grateful to Drs M Hamazaki and H Hiraiwa Thus, we identi®ed for the ®rst time a 500-kb for providing the cell line and useful discussion and Dr M homozygous deletion that was located within one of Oshimura for giving the hybrid cells. The authors also the SROs at the distal region of 1p in NBL cell line(s). thank the hospitals and institutes collaborating with the Construction of a sequence-ready PAC contig enabled Japanese Neuroblastoma Study Group for providing us to de®ne the 500-kb HD region and to clone and surgical specimens. This work was supported in part by a grant-in-aid from map the genes. cDNA cloning and rough genome the Ministry of Health and Welfare for a New Compre- sequencing of about 60% led to the identi®cation of six hensive 10-year Strategy for Cancer Control, Japan, by a genes in this region. Four genes were di€erentially grant-in-aid from the Ministry of Education, Science, and expressed at high levels in favorable NBLs as Culture of Japan, and by a grant from Uehara Memorial compared to unfavorable ones. Since the 1p distal Foundation.

References

Amler LC, Corvi R, Praml C, Savelyeva L, Le Paslier D and Cheng NC, Van Roy N, Chan A, Beitsma M ,Westerveld A, Schwab M. (1995). Oncogene, 10, 1095 ± 1101. Speleman F and Versteeg R. (1995). Oncogene, 10, 291 ± Bolande RP. (1974). Hum. Pathol., 5, 409 ± 429. 297. Brodeur GM, Seeger RC, Schwab M, Varmus HE and de Lecea L, Ruiz-Lozano P, Danielson PE, Peelle-Kirley J, Bishop JM. (1984). Science, 224, 1121 ± 1124. Foye PE, Frankel WN and Sutcli€e JG. (1997). Genomics, Carling T, Imanishi Y, Gaz RD and Arnold A. (1999). Int. J. 42, 499 ± 506. Cancer, 83, 80 ± 82. DiccianniMB,ChauLS,BatovaA,VuTQandYuAL. CaronH,vanSluisP,vanHoeveM,deKrakerJ,BrasJ, (1996). Cancer Lett., 104, 183 ± 192. Slater R, Mannens M, Voute PA, Westerveld A and Dracopoli NC, Stanger BZ, Lager M and Housman DE. Versteeg R. (1993). Nat. Genet., 4, 187 ± 190. (1988). Genomics, 3, 124 ± 128. Caron H. (1995). Med. Pediatr. Oncol., 24, 215 ± 221.

Oncogene Homozygous deletion at 1p36.2 in a human neuroblastoma cell M Ohira et al 4307 Enari M, Sakahira H, Yokoyama H, Okawa K, Iwamatsu A Martinsson T, Sjoberg RM, Hedborg F and Kogner P. and Nagata S. (1998). Nature, 391, 43 ± 50. (1997). Cancer Genet. Cytogenet., 95, 183 ± 189. Enomoto H, Ozaki T, Takahashi E, Nomura N, Tabata S, Nagai M, Ichimiya S, Ozaki T, Seki N, Mihara M, Furuta S, Takahashi H, Ohnuma N, Tanabe M, Iwai J, Yoshida H, Ohira M, Tomioka N, Nomura N, Sakiyama S, Kubo O, Matsunaga T and Sakiyama S. (1994). Oncogene, 9, 2785 ± Takakura K, Hori T and Nakagawara A. (2000). Int. J. 2791. Oncol. 16, 907 ± 916. Fukusumi S, Kitada C, Takekawa S, Kizawa H, Sakamoto J, Nagase T, Ishikawa K, Miyajima N, Tanaka A, Kotani H, Miyamoto M, Hinuma S, Kitano K and Fujino M. (1997). Nomura N and Ohara O. (1998). DNA Res., 5, 31 ± 39. Biochem. Biophys. Res. Commun., 232, 157 ± 163. Nangaku M, Sato-Yoshitake R, Okada Y, Noda Y, Hattori M, Tsukahara F, Furuhata Y, Tanahashi H, Hirose Takemura R, Yamazaki H and Hirokawa N. (1994). Cell, M, Saito M, Tsukuni S and Sakaki Y. (1997). Nucleic 79, 1209 ± 1220. Acids Res., 25, 1802 ± 1808. Pukatzki S, Tordilla N, Franke J and Kessin RH. (1998). J. He L, Yu JX, Liu L, Buyse IM, Wang MS, Yang QC, Biol. Chem., 273, 24131 ± 24138. Nakagawara A, Brodeur GM, Shi YE and Huang S. Schleiermacher G, Peter M, Michon J, Hugot JP, Vielh P, (1998). Cancer Res., 58, 4238 ± 4244. Zucker JM, Magdelenat H, Thomas G and Delattre O. Hiraiwa H, Hamazaki M, Takata A, Kikuchi H and Hata J. (1994). Genes Chromosomes Cancer, 10, 275 ± 281. (1997). Cancer, 79, 2036 ± 2044. Schwab M, Praml C and Amler LC. (1996). Genes Ichimiya S, Nimura Y, Kageyama H, Takada N, Sunahara Chromosomes Cancer, 16, 211 ± 229. M, Shishikura T, Nakamura Y, Sakiyama S, Seki N, Ohira Sunahara M, Ichimiya S, Nimura Y, Takada N, Sakiyama S, M, Kaneko Y, McKeon F, Caput D and Nakagawara A. Sato Y, Todo S, Adachi W, Amano J and Nakagawara A. (1999). Oncogene, 18, 1061 ± 1066. (1998). Int. J. Oncol., 13, 319 ± 323. Imashuku S, Inui A, Nakamura T, Tanaka J and Miyake S. TakedaO,HommaC,MasekiN,SakuraiM,KandaN, (1973). J. Clin. Endocrinol. Metab., 36, 931 ± 936. Schwab M, Nakamura Y and Kaneko Y. (1994). Genes Ioannou PA, Amemiya CT, Garnes J, Kroisel PM, Shizuya Chromosomes Cancer, 10, 30 ± 39. H, Chen C, Batzer MA and de Jong PJ. (1994). Nat. Takemura R, Nakata T, Okada Y, Yamazaki H, Zhang Z Genet., 6, 84 ± 89. and Hirokawa N. (1996). J. Neurosci., 16, 31 ± 35. Johnson ES, Ma PC, Ota IM and Varshavsky A. (1995). J. The I, Murthy AE, Hannigan GE, Jacoby LB, Menon AG, Biol. Chem., 270, 17442 ± 17456. Gusella JF and Bernards A. (1993). Nat. Genet., 3, 62 ± 66. Jones PA and Laird PW. (1999). Nature Genet., 21, 163 ± 167. Tsui SK, Chan JY, Waye MM, Fung KP and Lee CY. (1996). Kaghad M, Bonnet H, Yang A, Creancier L, Biscan JC, Biochem. Genet., 34, 367 ± 373. Valent A, Minty A, Chalon P, Lelias JM, Dumont X, White PS, Kaufman BA, Marshall HN and Brodeur GM. Ferrara P, McKeon F and Caput D. (1997). Cell, 90, 809 ± (1993). Genes Chromosomes Cancer, 7, 102 ± 108. 819. White PS, Maris JM, Beltinger C, Sulman E, Marshall HN, Knudson AG. (1971). Proc. Natl. Acad. Sci. USA, 68, 820 ± Fujimori M, Kaufman BA, Biegel JA, Allen C, Hilliard C, 823. Valentine MB, Look AT, Enomoto H, Sakiyama S and Koegl M, Hoppe T, Schlenker S, Ulrich HD, Mayer TU and Brodeur GM. (1995). Proc. Natl. Acad. Sci. USA, 92, Jentsch S. (1999). Cell, 96, 635 ± 644. 5520 ± 5524. Lahti JM, Valentine M, Xiang J, Jones B, Amann J, Grenet Will GK, Soukupova M, Hong X, Erdmann KS, Kiel JA, J, Richmond G, Look AT and Kidd VJ. (1994). Nat. Dodt G, Kunau WH and Erdmann R. (1999). Mol. Cell Genet., 7, 370 ± 375. Biol., 19, 2265 ± 2277. Liu X, Zou H, Slaughter C and Wang X. (1997). Cell, 89, Wise CA, Clines GA, Massa H, Trask BJ and Lovett M. 175 ± 184. (1997). Genome Res., 7, 10 ± 16. Martinsson T, Sjoberg RM, Hedborg F and Kogner P. (1995). Cancer Res., 55, 5681 ± 5686.

Oncogene