Loss of heterozygosity at 4p16.3 and mutation of FGFR3 in ... - Nature

ã

Oncogene (2001) 20, 686 ± 691 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Loss of heterozygosity at 4p16.3 and mutation of FGFR3 in transitional cell carcinoma Kathryn Sibley1, Darren Cuthbert-Heavens2 and Margaret A Knowles*,1 1

ICRF Clinical Centre, St James's University Hospital, Leeds, UK; 2ICRF Mutation Detection Facility, St James's University Hospital, Leeds, UK

4p16.3 has previously been identi®ed as a region of nonrandom LOH in transitional cell carcinoma, suggesting the presence of a tumour suppressor gene. One candidate within this region is ®broblast growth factor receptor 3 (FGFR3). Germline mutations in FGFR3 are known to cause several autosomal dominant skeletal dysplasias, the severity of which depends on the position and nature of the mutation in the protein. We investigated the frequency and nature of FGFR3 mutations in a panel of transitional cell carcinomas and cell lines and studied the possible link between mutation and loss of heterozygosity (LOH) on 4p16.3. FGFR3 coding sequence from 63 transitional cell carcinomas (TCC) of various stages and grades, and 18 cell lines was analysed by ¯uorescent SSCP. Samples with abnormal migration patterns were sequenced to identify the mutation or polymorphism. Thirty-one of the 63 tumours had previously been assessed to have LOH at 4p16.3. Twenty-six of the 63 tumours (41%) and 4/18 (22%) of the cell lines had missense mutations in FGFR3. All mutations detected in our panel have been reported in the germline where all apart from one cause lethal conditions. One tumour contained K652Q which has recently been identi®ed in less severe cases of skeletal dysplasia. Tumours with and without LOH at 4p16.3 had mutations in FGFR3 suggesting that these two events are not causally linked. The frequency of FGFR3 mutation indicates that this protein plays an important role in TCC. Oncogene (2001) 20, 686 ± 691. Keywords: bladder; cancer; FGFR3; mutation Introduction Bladder cancer is the fourth most common cancer in males in the US and UK (Landis et al., 1999; HMSO, 1994). Tumour development is believed to require multiple genetic events which result in both the activation of oncogenes and the inactivation of tumour suppressor genes. One way to identify tumour suppressor genes involved in bladder cancer has been

*Correspondence: MA Knowles, ICRF Clinical Cancer Centre, St. James's University Hospital, Beckett Street, Leeds LS9 7TF, UK Received 26 September 2000; revised 31 October 2000; accepted 13 November 2000

®rstly to identify common regions of chromosomal loss followed by mutation analysis of candidate genes on the retained allele. LOH of chromosome 4 has been identi®ed in bladder (Elder et al., 1994; Knowles et al., 1994; Polascik et al., 1995) and several other cancer types including colorectal (Vogelstein et al., 1989), ovarian (Sato et al., 1991), hepatocellular (Buetow et al., 1989), head and neck squamous cell carcinoma (Nawroz et al., 1994), neuroblastoma (Caron et al., 1996), gastric adenocarcinoma (Yustein et al., 1999) and breast carcinoma (Shivapurkar et al., 1999). A subset of these studies has pinpointed 4p16.3 as a likely tumour suppressor gene locus. 4p16.3 is a gene rich region which has been extensively mapped due to its proximity to the Huntington's disease gene (Hadano et al., 1998; McCombie et al., 1992; Pribill et al., 1997; Zuo et al., 1992). One candidate gene located in this region is FGFR3. Fgfr3 is a member of the highly conserved ®broblast growth factor receptor (Fgfr) family of proteins that share a common structure comprising an extracellular ligand binding domain, a transmembrane domain and an intracellular tyrosine kinase domain. Currently there are four members of the FGFR family, each encoded by separate genes on dierent chromosomes. The gene encoding FGFR3 at 4p16.3 comprises 19 exons spanning 16.5 kb (Perez-Castro et al., 1997). Mutations in the coding region have been identi®ed in several autosomal dominant skeletal dysplasias such as achondroplasia (ACH), thanatophoric dysplasia (TD), and hypochrondroplasia (HCH). These mutations are invariably the result of a single base change which causes an amino acid substitution and constitutively activates the receptor in a ligand independent manner (Webster and Donoghue, 1996, 1997). In bone, the constitutively activated Fgfr3 has increased tyrosine kinase activity, ultimately activating several STAT proteins which in turn upregulate transcription of speci®c cell cycle inhibitors (Li et al., 1999; Su et al., 1997). In FGFR3 mutant mice, activation of cell cycle inhibitors results in a decrease in proliferation and an increase in the number of quiescent chondrocytes at the developing bone growth plate (Li et al., 1999). Assuming a similar pathway in humans, activating mutations in FGFR3 could negatively regulate bone growth and account for the phenotypic eect of short limbs in individuals with a germline mutation.

LOH at 4p16.3 and mutation of FGFR3 in bladder cancer K Sibley et al

In sporadic tumours, mutations that activate FGFR3 have recently been identi®ed. In multiple myeloma, Y373C has been identi®ed in the KMS11 cell line (Chesi et al., 1997; Richelda et al., 1997) and K650E/ K650M mutations have been detected in the OPM2 cell line and one primary tumour (Chesi et al., 1998). These mutations are always accompanied by a t(4;14) (p16.3;q32.3) translocation, the breakpoints being approximately 50 kb centromeric to FGFR3 within the MMSET gene on 4p16.3 and within the IgH locus at 14q32.3. The translocation of FGFR3 to 14q32.3 brings the gene under the in¯uence of strong enhancer elements which upregulate Fgfr3 expression (Chesi et al., 1998). Therefore, 25% of multiple myeloma tumours that possess this translocation have upregulated expression of Fgfr3, and a subset of these tumours also have a mutation in FGFR3 which constitutively activates the receptor. Studies on an IL6-dependent mouse cell line expressing constitutively active FGFR3 have shown that STAT3 is a downstream target and is constitutively phosphorylated. Activation of STAT3 and the subsequent eects of downstream targets appears to protect cells from apoptosis following removal of IL6 (Plowright et al., 2000). These ®ndings suggest how mutations in FGFR3 can contribute to the malignant phenotype, and provide an explanation for the opposite regulatory eects that these mutations have in chondrocytes. Recently, an analysis of FGFR3 cDNA from bladder and cervical carcinomas revealed that 35% of bladder tumours and 25% of cervical carcinomas possess a previously identi®ed missense mutation in FGFR3 (Cappellen et al., 1999). We have shown that 4p16.3 has loss of heterozygosity (LOH) in transitional cell carcinomas and is therefore likely to harbour a tumour suppressor gene (Elder et al., 1994). This study set out to examine the link between LOH at 4p16.3 and FGFR3 mutations by examining the coding sequence of FGFR3 in a panel of 63 tumours (31 with LOH and 32 without LOH) and 18 cell lines. We now report the frequency and nature of FGFR3 mutations in transitional cell carcinoma and our ®ndings on the link between mutation and loss of heterozygosity. Results We examined the coding sequence of FGFR3 from 63 tumours and 18 cell lines using ¯uorescent SSCP. In total, we found that 26/63 (41%) tumours and 4/18 (22%) cell lines possessed mutations in FGFR3 (Table 1). FGFR3 mutations in TCC tumours did not correlate with a particular stage or grade. Thirty of 31 mutations have been previously described in thanatophoric dysplasia whereas one mutation (K652Q) has only been found in milder cases of hypochondroplasia. Of the 31/63 tumours previously found to have LOH at 4p16.3 (Elder et al., 1994), 11/ 31 (35%) tumours had the S249C mutation in exon 7 but no other type of mutation. We previously described

687

Table 1 FGFR3 mutations in bladder tumours and cell lines Tumour

Stage/Gradea Mutation Exon

Mutations in tumours with loss of heterozygosity

3 5 6 9 10 11 17 20 21 23 27

NI NI NI NI G3pT3 G1pT1 G3pT3 G2pT2 G1pTa NI G3pT1

S249C S249C S249C S249C S249C S249C S249C S249C S249C S249C S249C

7 7 7 7 7 7 7 7 7 7 7

Mutations in tumours without loss of heterozygosity

38 41 43 44 49 50 52 59 60 61 37 39 56 62 53

G2pT2 G1 G1pTa/T1 G1 G2pT1 G1pT1 G2 G2 G2pTa G2pTa G2pT2 G1 pT1 G2pT1 G2pT1

S249C S249C S249C S249C S249C S249C S249C S249C S249C S249C Y375C G372C Y375C Y375C K652Q

7 7 7 7 7 7 7 7 7 7 10 10 10 10 15

Mutations in cell lines

JO'N HT1197 J82 609BC/CR

± ± ± ±

S249C S249C K652E K652E

7 7 15 15

a

Tumours were graded according to WHO recommendations (Mosto® et al., 1973). Staging was according to the TNM classi®cation (UICC, 1978). NI indicates no information

subsets of tumours with LOH con®ned to dierent regions of 4p (Elder et al., 1994). Mutations were identi®ed in all these groups. 10/32 (31%) tumours had the S249C mutation, 3/32 (9%) had Y375C in exon 10, 1/32 (3%) had G372C in exon 10, and 1/32 (3%) had a K652Q novel mutation in exon 15. Two of 18 (11%) cell lines were found to have the S249C mutation and 2/18 (11%) cell lines contained K652E. Mutations at amino acids 372, 375 and 652 in epithelial cells which express the Fgfr3b isoform correspond to amino acids 370, 373 and 650 of the Fgfr3c isoform in chondrocytes, the dierence resulting from alternative splicing earlier in the coding sequence which gives epithelial Fgfr3 an additional two amino acids. Examples of the missense mutations are shown in Figure 1. Mutant and wild type sequences were frequently seen in tumours with LOH at 4p16.3, probably re¯ecting the fact that some normal tissue was present when DNA was extracted from the tumour. Two tumours with LOH (3 and 11) were homozygous at the position of the mutation suggesting a lack of contaminating normal tissue in these tumours. None of the samples tested had more than one mutation in FGFR3. SSCP also identi®ed abnormally migrating bands in exons 9, 14, 15, and introns 6, 9, and 17. Sequencing revealed that these band shifts were caused by single base changes which did not alter the amino acid nor Oncogene


688

Figure 1 Examples of FGFR3 mutations in TCC tumours and cell lines. (a) Tumour 43 (heterozygote) has a G to C transversion shown here in the reverse strand which alters the sequence of codon 249 from TCC (Ser) to TGC (Cys). (b) Tumour 11 has the same mutation as tumour 43 but due to LOH at 4p16.3 does not have any wild-type sequence present. (c) Tumour 53 (heterozygote) has an A to C transversion which alters the sequence of codon 652 from AAG (Lys) to CAG (Gln). (d) 609BC has an A to G transition at codon 652 which alters the sequence to codon 652 from AAG (Lys) to GAG (Glu)

did they aect intron-exon splice boundaries (Table 2 and Figure 2). In tumour 21 the CÁ4T base change was also found in the matched normal sample (not shown) suggesting that this sequence variation is a polymorphism, but a larger study will be necessary to assess the frequency. Normal DNA was not available for tumour 31. The C/T polymorphism within exon 7 was detected at the same frequency as reported (Tartaglia et al., 1998). Discussion Our results show that FGFR3 is frequently mutated in TCC, irrespective of whether the tumour has LOH at Oncogene

4p16.3. Tumours without LOH did have a slightly higher frequency of mutation than tumours with LOH (47% compared to 35%), and more variation in mutation type although S249C was the most common mutation in both groups. This may re¯ect the relatively small sample size of tumours analysed (31 tumours with LOH and 32 tumours without LOH). Alternatively, it may re¯ect a real dierence between the groups, suggestive of a relationship between a potential tumour suppressor gene on 4p and signalling via Fgfr3. A larger study will be required to ascertain whether the dierences between tumours with and without LOH at 4p16.3 are signi®cant. Our ®nding of both normal and mutant sequence in many of the tumours with recorded 4p LOH is likely to re¯ect the presence of contaminat-


ing normal DNA in the samples. However, as FGFR3 lies just telomeric to the most telomeric marker used in our LOH analysis (D4S43) it is also possible that some tumours with LOH at D4S43 were heterozygous at FGFR3 and this must now be tested. Activating mutations in FGFR3 have now been identi®ed in dierent cancer types and it seems likely that these mutations contribute to the malignant phenotype (Cappellen et al., 1999; Chesi et al., 1998; Richelda et al., 1997). The same missense mutations are also responsible for causing lethal forms of autosomal dominant skeletal dysplasias (Passos-Bueno et al., 1999). It is noteworthy that only one mutation associated with the less severe skeletal disorders such as achondroplasia (ACH) or hypochondroplasia (HCH) has been detected in tumours in this study or from the work of others. The K652/650Q mutation identi®ed in

Table 2 Polymorphisms in FGFR3 Sample

Intron/Exon

Positiona

Effect

Tumour 31

Intron 6

CGTC?CGTT/C

Tumour 21

Exon 9

HT1197 SW1710 HT1376 EJ

Intron 9

IVS6-15C>T (L78725) +1013C>T (NM000142) IVS9+30G>C (L78728)

Exon 14

SCaBER

Exon 15

UM-UC3

Intron 17

a

+1646G>T (NM000142) +1952>G (NM000142) IVS17-33G>A (Y08100)

ACC?ACC/T Thr?Thr CCGC?CCC/GCb GGG?GGG/T Gly?Gly ACA?ACA/G Thr?Thr GCGG to GCG/AG

Nomenclature according to previously published guidelines (Antonarakis, 1998). Genbank accession numbers of reference sequences are in parentheses. bPolymorphism con®rmed by AciI restriction enzyme digest (not shown)

one tumour, was recently reported in a study of patients with suspected hypochrondoplasia where it was associated with a less severe phenotype than patients with K650M and K650E (Bellus et al., 2000). This mutation was found to result in constitutive activation of the FGFR3 tyrosine kinase but to a lesser degree than the K650E or K650M mutations. Patients with heterozygous mutations in FGFR3 causing ACH or HCH are not reported to have a higher incidence of cancer, so it is possible that the non lethal FGFR3 mutations do not suciently activate the protein to signi®cantly contribute to tumour development. Mutations in FGFR3 have been shown to be dominant and result in constitutive activation of the receptor (Webster and Donoghue, 1996, 1997), making this gene a good candidate oncogene. The fact that FGFR3 is located at 4p16.3, a region known to have LOH in 22% of transitional cell carcinomas, and therefore a candidate tumour suppressor gene, is contradictory. Based on the fact that mutations are not associated with LOH in TCC, we believe that FGFR3 is not a tumour suppressor gene. FGFR3 is frequently mutated in bladder cancer suggesting that aberrant FGF receptor signalling may contribute to tumorigenesis. Fgfr3 belongs to a family of receptor tyrosine kinases which share a common structure and which have at least 19 ligands. There is accumulating evidence that perturbation of FGF signalling pathways through distinct ligand-receptor interactions may be involved in tumorigenesis. Previous work has shown that FGF1, a ligand for Fgfr3, could transform a mouse cell line and substantially increase its tumorigenicity (Forough et al., 1993). More recent studies using a rat bladder carcinoma model have shown that cells transfected with FGF1 are more tumorigenic and cells transfected with either FGF1 or FGF2 produced more vascularized tumours (Jouan-

689

Figure 2 Identi®cation of polymorphisms in FGFR3. (a) Reverse strand sequencing of intron 6 from tumour 31 reveals an A/G (T/C in sense strand) polymorphism at position 715. (b) SCaBER has an A/G polymorphism which alters codon 653 in exon 15 from ACA to ACG. Both of these codons encode threonine resulting in no change to the protein Oncogene


690

neau et al., 1997). A signi®cant proportion of breast cancers have also been shown to overexpress FGF8, a ligand for Fgfr3 (Marsch et al., 1999). Several tumourderived cell lines have been shown to have high FGFR3 expression (Chandler et al., 1999) and overexpression of FGFR3 has been shown to promote overgrowth of con¯uent thyroid carcinoma cell lines (Onose et al., 1999). In addition, chimeric receptors possessing a highly activated form of the Fgfr3 tyrosine kinase domain can morphologically transform ®broblasts (Webster and Donoghue, 1997). Taken together, these studies suggest that FGFs and their receptors are potentially involved in several key stages of tumorigenesis. In conclusion, this study has identi®ed four cell lines containing two dierent mutations in FGFR3, which will be useful tools to investigate the downstream targets of Fgfr3 in bladder cancer. In addition, K562Q has been found in a tumour for the ®rst time, which may indicate that other mutations in FGFR3 are yet to be described. Mutations of FGFR3 in transitional cell carcinoma have been found at a high frequency (41%), suggesting that FGFR3 plays an important role in the normal regulation of cellular activity and that constitutive activation of this receptor is likely to contribute to the malignant phenotype.

(Kyriazis et al., 1984), 5637 (G Cannon, unpublished), SHIM (a newly established TCC cell line, unpublished), SD (Pauli et al., 1983), and A1698 (Santos et al., 1984). 609BC and 609CR are two sublines of HU609T which was derived from the normal human urothelium of a patient with renal cancer (Vilien et al., 1983). The remaining three cell lines included one EBV immortalized lymphoblastoid cell line (J82-EBV, kindly provided by Dr C O'Toole) which is derived from the same patient as J82, one bladder squamous cell carcinoma cell line (SCaBER; O'Toole et al., 1976), and one cervical carcinoma cell line (HeLa; Gey et al., 1952). All cell lines were grown under standard conditions. Genomic DNA was extracted using a DNeasy kit (Qiagen, Crawley, West Sussex, UK). Polymerase chain reaction Intronic oligonucleotide primer pairs spanning FGFR3 exons 2 ± 19 inclusive were synthesized and labelled with FAM (sense primer) or TET (antisense primer) phosphoramidite dyes (ICRF laboratories, Clare Hall, London, UK). PCR reactions were carried out under standard conditions with the addition of 5% DMSO (Sigma, Dorset, UK) to all reactions. Primer sequences will be provided on request. Normal human genomic DNA and no template were included as controls in all experiments. The cycle times were as follows: 10 min at 958C followed by 45 cycles of 958C for 30 s, 55 ± 658C for 30 s, 728C for 40 s and one cycle at 728C for 5 min. SSCP analysis

Materials and methods Patient samples Matched tumour and peripheral lymphocyte DNA was extracted from a cohort of patients which have been described elsewhere (Elder et al., 1994). Cell lines Nineteen cell lines were used in this study. Fifteen were derived from transitional cell carcinomas including 253J (Elliott et al., 1974), J82 (O'Toole et al., 1978), VM-CUB2 (Williams, 1980), UMUC3 (Grossman et al., 1986), JO'N (Human Tumour Cell Laboratory, Memorial Sloan-Kettering Cancer Centre, New York, US), RT4 (Rigby and Franks, 1970), EJ (Bubenik et al., 1973), HT1197, HT1376 (Rasheed et al., 1977), RT112 (Hastings and Franks, 1981), SW1710

Products were combined with formamide and TAMRA-500 standards (PE Biosystems, Warrington, UK), denatured at 958C for 2 min and `snap cooled' in iced water. Samples were loaded onto 36 cm well to read, 6.5% (49 : 1) acrylamide, 5% glycerol gels on an ABI 377 DNA sequencer linked to an external circulating cooling water bath set at 188C. Run conditions were set at 50 W power limiting and gels run for up to 16 h. Data was collected using ABI Genescan 3.1 software and analysed using ABI Genotyper 2.5 software (PE Biosystems). DNA sequencing Samples which showed abnormal migration patterns on the SSCP gels were re-ampli®ed using the corresponding nonlabelled primer pair and sequenced (Big Dye Terminator kit, PE Biosystems). Base pair changes were detected using sequence analysis software (PE Biosystems).

References Antonarakis SE. (1998). Hum. Mutat., 11, 1 ± 3. Bellus GA, Spector EB, Speiser PW, Weaver CA, Garber AT, Bryke CR, Israel J, Rosengren SS, Webster MK, Donoghue DJ and Francomano CA. (2000). Am. J. Hum. Genet. (In press). Bubenik J, Baresova M, Viklicky V, Jakoubkova J, Sainerova H and Donner J. (1973). Int. J. Cancer, 11, 765 ± 773. Buetow KH, Murray JC, Israel JL, London WT, Smith M, Kew M, Blanquet V, Brechot C, Redeker A and Govindarajah S. (1989). Proc. Natl. Acad. Sci. USA, 86, 8852 ± 8856.

Oncogene

Cappellen D, De Oliveira C, Ricol D, de Medina S, Bourdin J, Sastre-Garau X, Chopin D, Thiery JP and Radvanyi F. (1999). Nat. Genet., 23, 18 ± 20. Caron H, van Sluis P, Buschman R, Pereira do Tanque R, Maes P, Beks L, de Kraker J, Voute PA, Vergnaud G, Westerveld A, Slater R and Versteeg R. (1996). Hum. Genet., 97, 834 ± 837. Chandler LA, Sosnowski BA, Greenlees L, Aukerman SL, Baird A and Pierce GF. (1999). Int. J. Cancer, 81, 451 ± 458. Chesi M, Nardini E, Brents LA, Schrock E, Ried T, Kuehl WM and Bergsagel PL. (1997). Nat. Genet., 16, 260 ± 264.


Chesi M, Nardini E, Lim RS, Smith KD, Kuehl WM and Bergsagel PL. (1998). Blood, 92, 3025 ± 3034. Elder PA, Bell SM and Knowles MA. (1994). Oncogene, 9, 3433 ± 3436. Elliott AY, Cleveland P, Cervenka J, Castro AE, Stein N, Hakala TR and Fraley EE. (1974). J. Natl. Cancer Inst., 53, 1341 ± 1349. Forough R, Xi Z, MacPhee M, Friedman S, Engleka KA, Sayers T, Wiltrout RH and Maciag T. (1993). J. Biol. Chem., 268, 2960 ± 2968. Gey GO, Coman WO and Kubicek MT. (1952). Cancer Res., 12, 264 ± 265. Grossman HB, Wedemeyer G, Ren L, Wilson GN and Cox B. (1986). J. Urol., 136, 953 ± 959. Hadano S, Ishida Y and Ikeda JE. (1998). DNA Res., 5, 177 ± 186. Hastings RJ and Franks LM. (1981). Int. J. Cancer, 27, 15 ± 21. HMSO. (1994). Cancer registration statistics. England and Wales. Jouanneau J, Plouet J, Moens G and Thiery JP. (1997). Oncogene, 14, 671 ± 676. Knowles MA, Elder PA, Williamson M, Cairns JP, Shaw ME and Law MG. (1994). Cancer Res., 54, 531 ± 538. Kyriazis AA, Kyriazis AP, McCombs WB and Peterson Jr WD. (1984). Cancer Res., 44, 3997 ± 4005. Landis SH, Murray T, Bolden S and Wingo PA. (1999). CA Cancer J. Clin., 49, 8 ± 31. Li C, Chen L, Iwata T, Kitagawa M, Fu XY and Deng CX. (1999). Hum. Mol. Genet., 8, 35 ± 44. Marsh SK, Bansal GS, Zammit C, Barnard R, Coope R, Roberts-Clarke D, Gomm JJ, Coombes RC and Johnston CL. (1999). Oncogene, 18, 1053 ± 1060. McCombie WR, Martin-Gallardo A, Gocayne JD, FitzGerald M, Dubnick M, Kelley JM, Castilla L, Liu LI, Wallace S, Trapp S et al. (1992). Nat. Genet., 1, 348 ± 353. Mosto® FK, Sobin LH and Humberto T. (1973). Histological Typing of Urinary Bladder Tumors. WHO, Geneva. Nawroz H, van der Riet P, Hruban RH, Koch W, Ruppert JM and Sidransky D. (1994). Cancer Res., 54, 1152 ± 1155. O'Toole C, Nayak S, Price Z, Gilbert WH and Waisman J. (1976). Int. J. Cancer, 17, 707 ± 714. O'Toole C, Price ZH, Ohnuki Y and Unsgaard B. (1978). Br. J. Cancer, 38, 64 ± 76. Onose H, Emoto N, Sugihara H, Shimizu K and Wakabayashi I. (1999). Eur. J. Endocrinol., 140, 169 ± 173. Passos-Bueno MR, Wilcox WR, Jabs EW, Sertie AL, Alonso LG and Kitoh H. (1999). Hum. Mutat., 14, 115 ± 125. Pauli BU, Cohen SM and Weinstein RS. (1983). J. Urol., 129, 646 ± 652.

Perez-Castro AV, Wilson J and Altherr MR. (1997). Genomics, 41, 10 ± 16. Plowright EE, Li Z, Bergsagel PL, Chesi M, Barber DL, Branch DR, Hawley RG and Stewart AK. (2000). Blood, 95, 992 ± 998. Polascik TJ, Cairns P, Chang WY, Schoenberg MP and Sidransky D. (1995). Cancer Res., 55, 5396 ± 5399. Pribill I, Barnes GT, Chen J, Church D, Buckler A, Baxendale S, Bates GP, Lehrach H, Gusella MJ, Duyao MP, Ambrose CM, Gusella JF and MacDonald ME. (1997). Somat. Cell. Mol. Genet., 23, 413 ± 427. Rasheed S, Gardner MB, Rongey RW, Nelson-Rees WA and Arnstein P. (1977). J. Natl. Cancer Inst., 58, 881 ± 890. Richelda R, Ronchetti D, Baldini L, Cro L, Viggiano L, Marzella R, Rocchi M, Otsuki T, Lombardi I, Maiolo AT and Neri A. (1997). Blood, 90, 4062 ± 4070. Rigby CC and Franks LM. (1970). Br. J. Cancer, 24, 746 ± 754. Santos E, Martin-Zanca D, Reddy EP, Pierotti MA, Della Porta G and Barbacid M. (1984). Science, 223, 661 ± 664. Sato T, Saito H, Morita R, Koi S, Lee JH and Nakamura Y. (1991). Cancer Res., 51, 5118 ± 5122. Shivapurkar N, Sood S, Wistuba II, Virmani AK, Maitra A, Milchgrub S, Minna JD and Gazdar AF. (1999). Cancer Res., 59, 3576 ± 3580. Su WC, Kitagawa M, Xue N, Xie B, Garofalo S, Cho J, Deng C, Horton WA and Fu XY. (1997). Nature, 386, 288 ± 292. Tartaglia M, Saulle E, Bordoni V and Battaglia PA. (1998). Mol. Cell Probes, 12, 335 ± 337. UICC. (1978). Classi®cation of Malignant Tumors, Bladder, edition 3. International Union Against Cancer: Geneva. Vilien M, Christensen B, Wolf H, Rasmussen F, Hou-Jensen C and Povlsen CO. (1983). Eur. J. Cancer. Clin. Oncol., 19, 775 ± 789. Vogelstein B, Fearon ER, Kern SE, Hamilton SR, Preisinger AC, Nakamura Y and White R. (1989). Science, 244, 207 ± 211. Webster MK and Donoghue DJ. (1996). EMBO J., 15, 520 ± 527. Webster MK and Donoghue DJ. (1997). Mol. Cell. Biol., 17, 5739 ± 5747. Williams RD. (1980). Invest. Urol., 17, 359 ± 363. Yustein AS, Harper JC, Petroni GR, Cummings OW, Moskaluk CA and Powell SM. (1999). Cancer Res., 59, 1437 ± 1441. Zuo J, Robbins C, Taillon-Miller P, Cox DR and Myers RM. (1992). Hum. Mol. Genet., 1, 149 ± 159.

691

Oncogene