Oncogene (1998) 16, 481 ± 487  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00

Identi®cation of two distinct deleted regions on chromosome 13 in prostate cancer

Chunde Li1, Catharina Larsson2, Andrew Futreal3, Jonathan Lancaster3, Catherine Phelan2, Ulla Aspenblad4, Birgitta Sundelin4, Yie Liu5, Peter Ekman1, Gert Auer4 and Ulf SR Bergerheim1

Departments of 1Urology, 2Molecular Medicine, 4Pathology and 5Oncology, Karolinska Hospital, 171 76 Stockholm, Sweden; 3Division of Gynecologic Oncology, Department of Surgery, Duke University Medical Center, Durham, North Carolina 27710, USA

Aberrations of 13q occur frequently in prostate cancer In prostate cancer, somatic deletions of chromosome and this chromosome contains two known tumor 13q have been frequently observed both by compara- suppressor genes, BRCA2 and Rb1. This study analysed tive genomic hybridization (CGH) and loss of 13q LOH, DNA ploidy, BRCA2 mutation and pRb heterozygosity (LOH) analysis (Cher et al., 1996; expression in prostate cancers. In total, 13q deletions Cooney et al., 1996; Visakorpi et al., 1995). The Rb1 were found in 18 of 36 tumors but did not correlate with locus is frequently deleted, and in one prostate cancer histological grade, stage or DNA ploidy. Two smallest cell line a 103 bp deletion of the promoter region was regions of overlapping deletions were de®ned: one ¯anked observed with concomitant loss of the wild type allele, by D13S218 and D13S153; the other ¯anked by D13S31 suggesting inactivation of this TSG by a two-hit and D13S137. BRCA2 was less frequently deleted mechanism (Bookstein et al., 1990; Riley et al., 1994). whereas Rb1 did have a high frequency of deletion. Furthermore, immunostaining studies have revealed None of the two genes was located in any of these two that the Rb1 expression is frequently decreased or regions. Furthermore, BRCA2 mutation was not found in absent in a signi®cant proportion of sporadic prostate the ®ve tumors where deletions had involved the BRCA2 cancers (Phillips et al., 1994). However, screening for locus. Neither did the Rb1 deletion correlate with absent somatic mutations in Rb1 have given negative results pRb expression. In addition, tetraploidy was found in 14 also in the studies where larger series of tumors have out of 25 tumors analysed and correlated with aberrant been examined (Kubota et al., 1995; Sarkar et al., pRb expression. Our results indicate that 13q deletion is 1992; Tricoli et al., 1996). On the other hand, in one an early non-random event. Tumor suppressor genes study on families with hereditary breast cancer other than BRCA2 or Rb1 may be the target of 13q attributable to BRCA2, the male carriers have a 3.3- deletions. Aberrant pRb expression may not re¯ect the fold increased risk of developing prostate cancer as two-hit Rb1 inactivation but may be involved in the compared to non carriers (Wooster et al., 1995). Thus, tetraploidization of prostate cancer cells. the importance of the BRCA2 gene in sporadic prostate cancer can also be anticipated. Keywords: prostate cancer; chromosome 13; BRCA2; Two approaches can be used to identify a candidate Rb1; pRb; DBM gene to be a TSG in sporadic cancers. One is to identify the somatic mutation coupled with loss of the normal chromosome homologue and the other is to observe the gene deletion in combination with absent Introduction or aberrant gene expression. These two approaches were used in the present study to investigate the Aberrations of the long arm of chromosome 13 occur involvement of BRCA2 and Rb1 in sporadic prostate frequently in many types of tumors (Cletonjansen et cancer. The results indicate that deletion of chromo- al., 1995; Cooney et al., 1996; Li et al., 1994; Maestro some 13q is an early non-random event, and that et al., 1996; Pearce et al., 1996; Radford et al., 1995; neither BRCA2 nor Rb1 was the targets but instead Yang et al., 1992; Zhang et al., 1994). So far two other putative TSGs might be located nearby. tumor suppressor genes (TSG) have been identi®ed on this chromosome: the retinoblastoma gene (Rb1)on 13q14.2 and the breast cancer susceptibility locus 2 Results (BRCA2) on 13q12.1 (Lee et al., 1987; MacGee et al., 1989; Wooster et al., 1995). Rb1 is a cell cycle regulator Overview of 13q losses in prostate cancer which is involved in both hereditary and sporadic retinoblastoma, while BRCA2 mutations have been The losses observed (Figure 1) either resulted in described in a subset of families with site speci®c breast complete absence of one of the constitutional alleles cancer only (Oddoux et al., 1996; Phelan et al., 1996; in the tumor (pure LOH) or a signi®cantly decreased Thorlacius et al., 1996; Wooster et al., 1995; Yandell et intensity, but not complete loss of one constitutional al., 1989). allele (allelic imbalance). The decrease of the intensity was not always related directly to the tumor cell content of the tissue sample from which the tumor DNA was derived. In both types of LOH, the retained Correspondence: C Li allele had usually stronger signal intensity as compared Received 23 June 1997; revised 8 September 1997; accepted 9 to the corresponding allele in the normal DNA (Figure September 1997 1). Tumor suppressor genes on 13q in prostate cancer CLiet al 482 Case D4 tumors B8, D9 de®ned a distal limitation with marker D13S153 and LOH of tumors D4, D7 and E5 a N T N T proximal limitation with marker D13S218. Likewise for the distal SRO (SRO2), the LOH of tumor D3 de®ned a distal ¯anking marker D13S137 and the LOH of tumors D1 and E5 a proximal ¯anking marker D13S31ca (Figure 2). As DNA ploidy may re¯ect the total cancer genome instability, the two LOH regions were analysed relating to DNA ploidy, however, neither of the regions was D13S260 D13S171 associated with a special pattern of DNA ploidy (Tables 1 and 2). No correlation was found between N T these two regions and histological grade, stage, or N T origin of tumor material (Table 2).

No somatic BRCA2 mutation in sporadic prostate cancer The LOH frequency of the BRCA2 gene was low (Figure 2): only two (B1, D1) out of 27 (7%) D13S218 informative tumors showed LOH at D13S171 (within D13S267 the BRCA2 gene) and only ®ve tumors (B1, B6, C7, D1, D9) showed a deletion which may involve the N T N T BRCA2 locus (D13S260-D13S171-D13S267). The BRCA2 locus was not included in any of the two SROs observed (Figure 2). Five tumors for which the LOH might have involved the BRCA2 locus and a further two tumors (B7, C9) which retained both copies of BRCA2 locus were analysed for mutation in D13S153 D13S319 the BRCA2 coding region. A base substitution, G10338 to A, was found in tumor B1 but this substitution did not alter the amino acid Argentine. No other mutations were detected in these seven tumors. N T N T pRb expression patterns in prostate cancer The expression of pRb in tumor cells demonstrated D13S31ca three patterns: pRb expression limited to the nucleus (nuclear pRb), pRb expression in both nucleus and cytoplasm (ectopic cytoplasmic pRb), and complete absence of pRb. In nuclear pRb positive cases, tumor D13S137 cells showed a heterogeneous expression pattern with only a few cells being positively stained. In Figure 1 LOH analysis of prostate cancer case D4. Retention of heterozygosity at markers D13S260, D13S171, D13S267 and cytoplasmic pRb positive tumors, tumor cells always D13S218; LOH at markers D13S153 (pure LOH), D13S319 maintained nuclear pRb expression and were usually (allelic imbalance), D13S31ca (allelic imbalance) and D13S137 homogeneously stained (Figure 3). In all cases with (pure LOH). Arrows point out the lost allele in the tumor DNA. absent pRb expression, almost no tumor cells had N: normal. T: tumor positive staining at all. However, disregarding the expression pattern of pRb in the tumor cells, the surrounding lymphocytes as well as the interstitial stromal cells always demonstrated only nuclear pRb In total, 18 out of 36 (50%) prostate cancers staining. displayed LOH with at least one out of the eight In total, 34 of the 36 prostate tumors were analysed markers analysed (Figure 2). The most frequently for pRb expression and 22 showed aberrant expression: deleted locus was D13S153 (within Rb1) with nine 11 with ectopic cytoplasmic expression and 11 with out of 26 (35%) informative tumors showing LOH. absent expression (Table 3). The expression of ectopic From this central point of 13q loss, the frequency of cytoplasmic pRb was mainly observed in well LOH gradually decreased in both proximal and distal di€erentiated primary tumors, in contrast to the directions (Figure 2). Tumor B1 had a break point absent expression of nuclear pRb in poorly differ- proximal to D13S31ca and lost all the ®ve proximal entiated and metastatic prostate cancers (Table 3). loci. In contrast, tumor D4 had a break point distal to D13S218 and all the four distal markers showed No correlation of Rb1 LOH and pRb expression deletion (Figure 1). However, the LOH region of tumors B8 and D9 did not overlap with the LOH Table 1 shows the data of Rb1 LOH and pRb region of tumors C2, D3 and E6 (Figure 2), thus, two expression in each tumor. Of nine tumors with LOH smallest regions of overlapping deletions (SRO) were at D13S153 (Rb1), four (44%) maintained and ®ve observed. For the proximal SRO (SRO1), LOH of (56%) lost nuclear pRb expression (Table 3). Tumor suppressor genes on 13q in prostate cancer CLiet al 483

Figure 2 Prostate cancers in this study showing LOH on chromosome 13. The eight microsatellite markers used for the deletion mapping are given to the right of an ideogram of the chromosome. Tumor numbers refer to Table 1. Empty circles symbolize retained heterozygosity and ®lled circles indicate LOH. The frequency of LOH at each locus is given as number of cases with LOH/ number of informative cases. Two smallest regions of overlapping deletions are pointed out by SRO1 and SRO2

Table 1 Rb1 LOH, pRb expression and DNA ploidy in prostate cancer Tumor Material Stage Grade Rb1 pRb Nucl pRb Cyto DNA Ploidy C7 brain C-D M LOH ± ± B3 brain C-D P NI + ± B1 brain C-D PU LOH ± ± B9 brain C-D U NI ± ± F1 LN C-D WM RT + + T C3 LN C-D M NI ± ± T D4 LN C-D MP LOH ± ± T E5 LN C-D MP LOH ± ± T E9 LN C-D MP RT B4 LN C-D MP RT + ± C8 LN C-D MP RT + ± D3 primary A-B W RT + + T D6 primary A-B W LOH + ± D D7 primary A-B W LOH + + T D8 primary A-B W NI D9 primary A-B W RT + ± D E3 primary A-B W LOH + + T E7 primary A-B W RT + + D B2 primary A-B WM RT ± ± B5 primary A-B WM LOH + ± A B6 primary A-B WM LOH ± ± T C1 primary C-D WM NI + ± D C9 primary C-D WM NI + ± D D1 primary A-B WM NI ± ± T D2 primary A-B WM RT + + T E6 primary A-B WM RT + ± T E8 primary A-B WM NI + + T B8 primary C-D M RT + ± C2 primary A-B M NI ± ± T C5 primary C-D M RT + + D E1 primary A-B M RT + + A C4 primary C-D MP RT + + D D5 primary A-B MP RT ± ± T E4 primary A-B MP RT + ± D E2 primary A-B P RT + ± B7 primary C-D U NI + + D Primary: primary prostate cancer. LN: regional lymph node metastasis from prostate cancer. Brain: brain metastasis from prostate cancer. The letter in WHO grading system stands: W for well di€erentiated, M for moderately di€erentiated, P for poorly di€erentiated and U for undi€erentiated. LOH: loss of heterozygosity. The Rb 1 LOH is presented by the LOH at D13S153 which is within the gene. RT: retain of both alleles. NI: non-informative. pRb Nucl: pRb nuclear expression. pRb Cyto: ectopic pRb cytoplasmic expression. +: positive. 7: negative. In DNA ploidy, the letter stands: D for diploidy, T for tetraploidy and A for aneuploidy. The cell in the table without data means no data in this item Tumor suppressor genes on 13q in prostate cancer CLiet al 484 Table 2 Correlation of 13q LOH with histological grade, stage or Table 3 Correlation of pRb expression with histopathologic and DNA ploidy genomic characteristics Proximal Distal No 13q Nucl+/Cyto± Nucl+/Cyto+ Nucl±/Cyto± 13q LOH 13q LOH LOH Total Material Materials Primary 9 10 5 Primary 11 6 11 25 LN 2 1 3 LN 2 2 5 7 Brain 1 0 3 Brain 2 1 2 4 Grade Grade W-M 7 9 6 W-M 11 6 9 23 MP-U 5 2 5 MP-U 4 3 9 13 Stage Stage A-B 6 7 5 A-B 7 4 9 19 C-D 6 4 6 C-D 8 5 9 17 DNA ploidy DNA ploidy D 5 4 0 D 3 2 4 9 T 1 6 7* T 6 6 5 14 A 1 1 0 A 1 0 1 2 Rb1 Proximal 13q LOH: LOH includes D13S218-D13S153. Distal 13q LOH 2 2 5 LOH: LOH includes D13S31ca-D13S137. The two LOH regions refer RT 7 7 2 to Figure 2. Other items refer to Table 1 Each item refers to Table 1. *P<0.05

E3 NT

D13S153

Figure 3 Prostate cancer E3 with LOH (allelic imbalance) at D13S153 (Rb1), tetraploidy and intensi®ed ectopic cytoplasmic pRb expression Tumor suppressor genes on 13q in prostate cancer CLiet al 485 Table 4 Correlation between DNA ploidy pattern and pathological (Figure 1). We also observed a lack of concordance features between LOH with D13S153 (Rb1) and aberrant pRb Diploidy Tetraploidy Aneuploidy expression (Tables 1 and 3) as reported in a previous study (Cooney et al., 1996). In particular, tumors with Material Primary 9 10 2 Rb1 LOH presented also normal nuclear pRb LN 0 4 0 expression (case B5, D6, D7 and E3), indicating the Brain 0 0 0 existence of an intact copy of the Rb1 gene (Table 1) Grade and contradicting the two-hit inactivation mechanism W-M 6 11 2 MP-U 3 3 0 (Tables 1 and 3). This may further disqualify the Rb1 Stage gene to be the target of 13q deletions. A-B 4 10 2 Other candidate genes in the proximal critical region C-D 5 4 0 include the serotonin receptor subtype 2 (HTR2) gene and Each item refers to Table 1 the gene for an orphan G-protein coupled receptor (U16) which may need to be investigated. Both genes encode G- protein coupled receptors which have wide biological activities (Hsieh et al., 1990; Herzog et al., 1996). HTR2 pRb expression in tetraploid tumors is mapped centromeric to Rb1 on 13q14.2, while U16 is DNA ploidy was analysed in 25 of the 36 tumors. located in the intron 17 or Rb1 (Hsieh et al., 1990; Fourteen (56%) were tetraploid. Most of the tetraploid Herzog et al., 1996; Schuler et al., 1996). In the distal cases were primary tumors with well to moderate critical region, one TSG candidate may be the DBM gene di€erentiation (Table 4). Interestingly, tetraploidy was which has been found to be homozygously deleted in B correlated to pRb expression. Tetraploidy was seen in cell malignancy (Liu et al., 1995). However, the exact 14% of tumors with normal nuclear pRb expression, physical distance of the marker D13S319 from the DBM 55% of tumors with ectopic cytoplasmic pRb gene is unknown (Figure 2). expression, and 100% of tumors with absent pRb A previous study has shown that mouse embryonic expression (Table 3, P50.05 in Pearson's test). ®broblast (MEF) cell line with null pRb function when exposed to mitotic spindle inhibitors has a strong tendency of polyploidization via whole genome duplica- Discussion tion (Leonardo et al., 1997). Our results further showed an association between pRb aberration and tetraploidy LOH presents the imbalance ratio between the in a subset of tumor materials from prostate cancer numbers of the homologous chromosomes. In spite of patients (Table 3). The clinical importance of this indicating the secondary event to speci®c TSG observation needs further investigation. mutation, LOH may also re¯ect the general instability In summary, this study has shown the importance of of the cancer cell's genome. This type of LOH occurs 13q deletion in prostate cancer. The alteration of randomly and is related to parameters which roughly BRCA2 gene is rare and does not play an important re¯ect the whole genomic alteration, such as DNA role in this tumor. Aberrant expression of pRb may ploidy. We have found 50% of prostate tumors not re¯ect a two-hit inactivation of the Rb1 gene but showing LOH at one or more 13q markers and no may cause the tetraploidization of a subset of prostate correlation between 13q LOH and DNA ploidy, cancers. The result suggests the existence and indicating 13q deletion to be an early, non-random, importance of other yet unidenti®ed genes on and independent chromosomal alteration. chromosome 13q in prostate cancer. The frequently deleted region on 13q in these 36 prostate cancers was in agreement with the lost region detected by previous CGH analyses (Cher et al., 1994; Materials and methods Visakorpi et al., 1995). However, in the LOH mapping (Figure 2), two independent smallest regions of Prostate cancer patients overlapping deletions were further de®ned. The distal region was ¯anked by marker D13S31ca and D13S137 Diagnoses of prostate cancer in 36 patients were con®rmed (Figure 2) which overlaps the region described in a by histopathological examination of the surgical specimen. previous LOH analysis (Cooney et al., 1996). The The mean age at diagnosis was 63 years (ranges 52 ± 72). The disease extent (Table 1) was determined according to proximal region ¯anked by markers D13S218 and Whitemore and Jewitt staging system (Jewitt, 1975; D13S153 (Figure 2) has not been described before Whitemore, 1956). None of the 36 patients was subjected (Cooney et al., 1996). These two regions may suggest to any hormonal or radiation therapies prior to surgery. the existence of more than one TSG. The BRCA2 gene has been suggested to be a Prostate cancer tissue candidate but it was not located within these minimal regions (Figure 2). Lack of mutation in the coding Immediately after the specimen was removed, a horizontal region in tumors with LOH at the BRCA2 locus section across the tumor nodular lesion was made. Tumor further exclude the BRCA2 gene as being involved in tissue was sampled inside the nodule to avoid contamina- sporadic prostate cancer. tion of surrounding normal tissue. One piece of the sample was paran embedded and used for pRb immunostaining, The Rb1 gene spans approximately 200 kb DNA of and DNA ploidy measurement. The other piece was snap 13q14 (Lee et al., 1987; McGee et al., 1989). Marker frozen in 7708C and used for DNA extraction. The D13S153 which is located within the gene had the representativity of the tissue was estimated by histopatho- highest frequency of LOH (Figure 2). However, this logic examination and only samples which contained 50% marker seems not within the two SROs observed or more tumor cells were included in the study. Tumor suppressor genes on 13q in prostate cancer CLiet al 486 Of 36 tumor samples, 25 were from primary tumors, eosin staining was used to distinguish the tumor cells seven were from lymph node metastases, and four were from from surrounding stroma as well as in®ltrating brain metastates. The di€erentiation of prostate cancer lymphocytes. (Table 1) was estimated according to the World Health If the cell nucleus alone or together with cytoplasm was Organization's prostate cancer grading system (Mosto® et positively stained it was regarded as a pRb positive cell. A al., 1980). tumor with pRb negative staining was de®ned as in a ®eld under a 10610 times magni®cation less than 1% of over all tumor cells was pRb positive. Three investigators (C Li, CL DNA isolation and USRB) independently evaluated all the immunostaining High molecular weight DNA was isolated from matched slides and the ®nal judgement was reviewed by the pairs of peripheral leukocytes and tumor samples using pathologist (GA). standard methods. DNA ploidy analysis LOH analysis The detailed procedure was described before (Steinbeck Eight markers on 13q were used: D13S260-D13S171 (with- et al., 1994). Brie¯y, paralleled with pRb immunostain- in BRCA2)-D13S267-D13S218-D13S153 (within Rb1)- ing, a 4 mm thick section was prepared for Feulgen D13S319-D13S31ca-D13S137 (Figure 1). The order of the staining. The DNA content of individual tumor cell was markers is according to the Human Genome Data Base measured in a TV-based microscopic image analysis (GDB). D13S319 is located within the homozygously system. In each slide, at least 100 tumor cell nuclei were deleted DBM (disrupted in B cell malignancy) region. counted (Figure 3). The DNA content pro®le was D13S31ca was synthesized by the Nordic Human Genome recorded and compared with an internal diploid Mapping Program at the Uppsala University Hospital. standard obtained from the measurement of at least 20 Other markers were supplied by Research Genetic Inc. normal cells. (Huntsville, AL). The sequences and location of all the markersusedcanbefoundintheGDB. Statistical analysis The technical aspects of the PCR and microsatellite analysis were described previously (Liu et al., 1995). LOH The 36 tumors were divided into three subgroups according was de®ned as a complete absence of or signi®cantly to the origin of the tumor material analysed: primary decreased signal intensity of one of the constitutional alleles tumors, lymph node metastases and brain metastases in the tumor DNA and can be unambiguously detected by (Tables 1 ± 4). To facilitate the statistical analysis, the the naked eye (Figure 2). The result was reviewed by three WHO tumor grading system was integrated into two investigators (C. Li, CL and USRB). groups (Tables 2 ± 4): well to moderately di€erentiated (W-M) and moderately poor to undi€erentiated (MP-U). For the same reason, the prostate cancer stage was The BRCA2 gene mutation analysis integrated into two groups (Tables 1 ± 4): stage A-B and The details for the single strand conformational analysis stage C-D. Fisher's and Pearson's tests were used to (SSCA) and direct DNA sequencing of the BRCA2 gene analyse these categories each in association with 13q LOH, have been published before (Phelan et al., 1996). In short, DNA ploidy pattern and pRb expression pattern. The same 26 coding exons of the gene were ampli®ed using 55 tests were also used to analyse the association between pRb di€erent fragments of 200 ± 300 bp each, and the PCR expression pattern, the Rb1 LOH, and DNA ploidy products were run on MDE gels. When a SSCA shift was pattern. seen, the shifted band was excised from the gel, puri®ed and sequenced using a 377 Automated Fluorescent Sequencer (Applied BioSystems, CA).

pRb immunostaining The immunostaining was performed according to a Acknowledgements standard avidin-biotin complex (ABC) procedure on Dr Lena SpaÊ nberg at the Department of Clinic Genetics of 3 mm thick slide of paran embedded tissue selection. Uppsala University Hospital is acknowledged for kind and The pRb monoclonal antibody, PMG3-245 (PharMin- free supplement of microsatellite markers. Thanks should gen), speci®cally recognizes the pRb epitope of amino also go to colleagues at the Department of Urology in acid residue 300 ± 380 (DeCaprio et al., 1988). The Karolinska Hospital for their collaboration in collecting dilution of the primer was 1 : 200. One normal brain the special prostate cancer materials. This study was tissue and 10 tissues of histologically con®rmed benign supported in part by research grants from the Swedish prostatic hyperplasia were used as controls in each Cancer Foundation, Swedish Cancer Society and NCI/ experiment. The corresponding slide of hematoxylin and Duke University SPORE in breast cancer, P50-CA68438.

References

Bookstein R, Rio P, Madreperla SA, Hong F, Allred C, Cletonjansen AM, Collins N, Lakhani SR, Weissenbach J, Grizzle WE and Lee W-H. (1990). Proc. Natl. Acad. Sci. Devilee P, Cornelisse CJ and Stratton MR. (1995). Br. J. USA, 87, 7762 ± 7766. Cancer, 72, 1241 ± 1244. Cher ML, Bova GS, Moore DH, Small EJ, Carroll PR, Pin Cooney KA, Wetzel JC, Merajver SD, Macoska JA, SS, Epstein JI, Isaacs WB and Jensen RH. (1996). Cancer Singleton TP and Wojno KJ. (1996). Cancer Res., 56, Res., 56, 3091 ± 3102. 1142 ± 1145. Cher ML, MacGrogan D, Bookstein R, Brown JA, Jenkins DeCaprio JA, Ludlow JW, Figge J, Shew J-Y, Huang C-M, RB and Jensen RH. (1994). Genes Chrom. Cancer, 11, Lee W-H, Marsilio E, Paucha E and Livingston DM. 153 ± 162. (1988). Cell, 54, 275 ± 283. Tumor suppressor genes on 13q in prostate cancer CLiet al 487 Herzog H, Darby K, Hort YJ and Shine J. (1996). Genome Schuler CD, Boguski MS, Stewart EA, Stein LD, Gyapay G, Res., 6, 858 ± 861. Rice K, White RE, Rodriguez-Tome P, Aggarwal A, Hsieh CL, Bowcock AM, Farrer LA, Huang KN, Cavalli- Bajorek E, Bentolila S, Birren BB, Butler A, Castle AB, Sforza LL, Julius D and Francke U. (1990). 16, 567 ± Chiannikulchai N, Chu A, Clee C, Cowles S, Day PJR, 574. DiblingT,DrouotN,DunhamI,DupratS,EastC, Jewitt HJ. (1975). Urol. Clin. North. Am., 2, 105. Edwards C, Fan J-B, Fang N, Fizames C, Garrett C, Kubota Y, Fujinami K, Uemura H, Dobashi Y, Miyamoto Green L, Hadley D, Harris M, Harrison P, Brady S, Hicks H, Iwasaki Y, Kitamura H and Shuin T. (1995). Prostate, A, Holloway E, Hui L, Hussain S, Louis-Dit-Sully C, Ma 27, 314 ± 320. J, MacGilvery A, Mader C, Maratukulam A, Matise TC, Leonardo AD, Khan SH, Linke SP, Creco V, Seidita G and McKusickKB,MorissetteJ,MugallA,MuseletD, GMW. (1997). Cancer Res., 57, 1013 ± 1019. Nusbaum HC, Page DC, Peck A, Perkins S, Piercy M, Lee WH, Bookstein R, Hong F, Young LJ, Shew JY and Lee QinF,QuackenbushJ,RanbyS,ReifT,RozenS,Sanders EY. (1987). Science, 235, 1394 ± 1399. C, She X, Silva J, Slonim DK, Soderlund C, Sun W-L, Li X, Lee NK, Ye YW, Waber PG, Schweitzer C, Cheng QC Tabar P, Thangarajah T, Vega-Czarny N, Vollrath D, and Nisen PD. (1994). J. Natl. Cancer Inst., 86, 1524 ± Voyticky S, Wilmer T, Wu X, Adams MD, Ao€ray C, 1529. Walter NAR, Brandon R, Dhejia A, Goodfellow PN, Liu Y, Hermanson M, Grander D, Merup M, Wu XS, Houlgatte R, Hudson Jr JR, Ide SE, Iorio KR, Lee WY, Heyman M, Rasool O, Juliusson G, Gahrton G, Seki N, Nagase T, Ishikawa K, Nomura N, Phillips C, Detlofsson R, Nikiforova N, Buys C, Soderhall S, Polymeropoulos MH, Sandusky M, Schmitt K, Berry R, Yankovsky N, Zabarovsky E and Einhorn S. (1995). Swanson K, Torres R, Venter JC, Sikela JM, Beckmann Blood, 86, 1911 ± 1915. JS, Weissenbach J, Myers RM, Cox DR, James MR, Maestro R, Piccinin S, Doglioni C, Gasparotto D, Bentley D, Deloukas P, Lander ES and Hudson TJ. (1996). Science, 274, 540 ± 546. Vukosavljevic T, Sulfaro S, Barzan L and Boiocchi M. Steinbeck RG, Heselmeyer KM and Auer GU. (1994). Anal. (1996). Cancer Res., 56, 1146 ± 1150. Quant. Cyto., 16, 196 ± 202. McGee TL, Yandell DW and Dryja TP. (1989). Gene, 80, Thorlacius S, Olafsdottir G, Tryggvadottir L, Neuhausen S, 119 ± 128. Jonasson JG, Tavtigian SV, Tulinius H, Ogmundsdottir Mosto® FK, Sesterhenn I and Sobin LH. (1980). Interna- HM and Eyfjord JE. (1996). Nature Genet., 13, 117 ± 119. tional Histological Classi®cation Of Tumors. World Health Tricoli JV, Gumerlock PH, Yao JL, Chi SG, D'Souza SA, Organization. No. 10. Nestok BR and de Vere WR. (1996). Genes Chrom. Oddoux C, Struewing JP, Clayton CM, Neuhausen S, Brody Cancer, 15, 108 ± 114. LC, Kaback M, Haas B, Norton L, Borgen P, Jhanwar S, Visakorpi T, Kallioniemi AH, Syvanen AC, Hyytinen ER, Goldgar D, Ostrer H and Ot K. (1996). Nature Genet., Karhu R, Tammela T, Isola JJ and Kallioniemi OP. 14, 188 ± 190. (1995). Cancer Res., 55, 342 ± 347. Pearce SH, Trump D, Wooding C, Sheppard MN, Clayton Whitemore WFJ. (1956). Am. J. Med., 21, 697. RN and Thakker RV. (1996). Clin. Endocrinol., 45, 195 ± Wooster R, Bignell G, Lancaster J, Swift S, Seal SJM, 200. Collins N, Gregory S, Gumbs C, Micklem G, Barfoot R, Phelan CM, Lancaster J, Tonin P, Gumbs C, Carter R, Hamoudi R, Patel SRC, Biggs P, Hashim Y, Smith A, Ghadirian P, Peter C, Moslehi R, Dion F, Faucher M-C, Connor F, Arason AJG, Ficenec D, Kelsell D, Ford D, Dole K, Karimi S, Lounis H, Warner E, Goss P, Anderson Tonin P, Bishop DT, Spurr NK, Ponder BAJ, Eels R, Peto D, Larsson C, Narod SA and Futreal PA. (1996). Nature J, Devilee P, Cornelisse C, Lynch H, Narod S, Lenoir G, Genet., 13, 120 ± 122. Egilsson V, Barkadottir RS, Easton DF, Bentley DR, Phillips SM, Barton CM, Lee SJ, Morton DG, Wallace DM, Futreal PA, Ashworth A and Stratton MR. (1995). Lemoine NR and Neoptolemos JP. (1994). Br. J. Cancer, Nature, 378, 789 ± 792. 70, 1252 ± 1257. Yandell DW, Campbell TA, Dayton SH, Petersen R, Walton Radford DM, Fair KL, Phillips NJ, Ritter JH, Steinbrueck D, Little JB, McConikie-Rosell A, Buckley EG and Dryja T, Holt MS and Doniskeller H. (1995). Cancer Res., 55, TP. (1989). N. Engl. J. Med., 321, 1689 ± 1695. 3399 ± 3405. Yang FT, Li S, Han H and Schwartz PE. (1992). Int. J. Riley DJ, Lee EY-HP and Lee W-H. (1994). Annu. Rev. Cell Cancer, 52, 575 ± 580. Biol., 10, 1±29. ZhangX,XuHJ,MurakamiY,SachseR,YashimaK, SarkarFH,SakrW,LiYW,MacoskaJ,BallDEand Hirohashi S, Hu SX, Benedict WF and Sekiya T. (1994). Crissman JD. (1992). Prostate, 21, 145 ± 152. Cancer Res., 54, 4177 ± 4182.