Oncogene (2000) 19, 6317 ± 6323 ã 2000 Nature Publishing Group All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc Chromosome 9 deletions and recurrence of super®cial bladder cancer: identi®cation of four regions of prognostic interest Maryse Simoneau1,He leÁ ne LaRue1, Tahar O Aboulkassim1, FrancËois Meyer1, Lynne Moore1 and Yves Fradet*,1 1Centre de recherche en canceÂrologie de l'Universite Laval, Centre Hospitalier Universitaire de QueÂbec, Pavillon L'HoÃtel-Dieu de QueÂbec, QueÂbec, Canada In a previous study, loss of heterozygosity (LOH) of 28 various grades and stages, however, in low stage chromosome 9 microsatellite markers was assessed on tumors, they are often the only genetic anomaly 139 Ta/T1 bladder tumors. LOH at one or more loci identi®ed. In a previous study of chromosome 9 was detected in 67 tumors, 62 presenting subchromoso- LOH on a series of 139 primary Ta, T1 bladder mal deletions. One hundred and thirty-three of these tumors obtained at initial diagnosis (Simoneau et al., patients have now been followed for up to 8 years. The 1999) we found deletion of at least one chromosome 9 purpose of the present study was to evaluate the potential marker in 48% of tumors. Chromosome 9 monosomy biological signi®cance of chromosome 9 deletions in was a rare event and deletions on 9q were twice as super®cial bladder tumors at initial diagnosis. High frequent as deletions on 9p. Moreover, only 4% of grade was associated with LOH (P=0.004). Large cases had deletions on 9p only. This suggests that tumors carried more frequently 9p deletions (P=0.022). inactivation of one or more tumor suppressor genes on Female patients had more chromosome 9q LOH than the long arm of chromosome 9 may be an early event male patients did (P=0.010). Chromosome 9 LOH at all in the tumorigenic pathway leading to papillary loci was associated with an elevated risk of recurrence tumors. The lower incidence of chromosome 9 but four regions were associated with a particularly high anomalies in higher stage tumors reported in other risk of recurrence. Multivariate analysis taking into studies is consistent with this hypothesis (Tsai et al., account grade, stage, size and number of tumors showed 1990; Dalbagni et al., 1993; Spruck et al., 1994). that tumors deleted in the regions 9ptr-p22, 9q22.3, Recently, Czerniak et al., (1999), in elegant genetic 9q33, and 9q34 recurred signi®cantly more rapidly than mapping studies of cystectomy specimens, found those without deletions (Recurrence rate ratio=2.32, deletions of chromosome 9 in normal appearing 2.53, 2.52 and 2.43 respectively). Log-rank statistics urothelium adjacent to areas of the urinary bladder comparing Kaplan-Meier survival curves for the same mucosa exhibiting early precursor changes. They chromosomal regions con®rmed the correlation observed a clonal relationship among multiple samples (P=0.0002, 0.010, 0.002 and 0.009 respectively). Only from the same patients and suggested several regions four patients progressed to muscle-invasive disease. They on 9q that may contain critical genes for the all had extensive deletions on 9q but none had deletions development of urothelial neoplasia. This hypothesis at 9ptr-p22. This study suggests a link between was further enforced by the detection of chromosome 9 chromosome 9 anomalies and recurrence of super®cial deletions in cells from voided urine samples of patients bladder cancer. Oncogene (2000) 19, 6317 ± 6323. followed up after resection of transitional cell carcinoma (TCC), at a time when no tumor was Keywords: Bladder cancer; LOH; recurrence; prognos- detectable and urine cytologies were negative. tic; chromosome 9 One of the most important features of papillary bladder tumors is the high rate of recurrence after initial treatment. Pre-malignant ®eld changes unidenti- Introduction ®ed at the time of initial resection are likely to be the cause of these tumor re-growths. We previously Bladder cancer is characterized by frequent genetic showed that expression of antigen markers de®ned by alterations of chromosome 9. These include monosomy monoclonal antibodies were associated with indepen- 9 but also more localized deletions of various dent risk of tumor recurrence (Allard et al., 1995). chromosomal regions. Re®ned deletion mapping with Mapping studies of normal appearing urothelium in microsatellite markers has suggested the existence of patients with tumors have shown frequent expression several putative tumor suppressor loci on this chromo- of these antigens in otherwise normal cells (Lee et al., some at 9p22-23, 9p21-22, 9p11-13, 9q12-13, 9q21-22, 1997; Rao et al., 1993). Bartlett et al., (1998) recently 9q31 and 9q33-34 (Habuchi et al., 1995; Simoneau et suggested that loss of chromosome 9 as detected by al., 1996; Czerniak et al., 1999; Simoneau et al., 1999). Fish analysis of the pericentromeric satellite marker at Chromosome 9 deletions have been found in tumors of 9q12 characterized tumors at high risk of recurrence. Moreover they observed a great homogeneity of chromosome 9 copy number between multiple syn- chronous tumors and between the primary and *Correspondence: Y Fradet, Centre de recherche, CHUQ-L'Hoà tel- Dieu de Que bec, 11, coà te du Palais, Que bec, Canada, GIR 2J6 recurrent tumors. Received 9 June 2000; revised 13 October 2000; accepted 18 Thus there is growing evidence that recurrent October 2000 super®cial papillary tumors originate from ®eld Chromosome 9 deletions and recurrence in bladder cancer M Simoneau et al 6318 changes that may be monoclonal in origin. We Table 1 Demographic and biological characteristics of 133 patients hypothesized that identifying genes on chromosome 9 with initial bladder tumors associated with early tumor recurrence may indicate Characteristics Number of patients (%) their implication in early carcinogenesis. We thus Age (years) correlated the frequency of LOH at 28 microsatellite 560 49 (37) markers with clinical outcome in 133 initial low stage 60 ± 70 51 (38) tumors that were part of an epidemiological study 470 33 (25) providing veri®ed clinical information at initial diag- Sex Male 100 (75) nosis (Allard et al., 1995). After up to 8 years of Female 33 (25) follow-up, (5.6 years median) 69% of these patients Smoking status suered at least one recurrence but only 3% progressed Non-smoker 21 (16) to muscle-invasive disease. This low progression rate Ex-smoker 51 (38) Current smoker 61 (46) marks these tumors as typical of the super®cial Stage papillary bladder tumor pathway. In the present study, Ta 105 (81) we used these data to evaluate the potential biological T1 25 (19) signi®cance of chromosome 9 anomalies in super®cial Unknown low stage 3 tumors at initial diagnosis, particularly with regard to Grade 1 44 (33) recurrence. 2 77 (58) 3 12 (9) Number of tumors Results 1 97 (73) 41 36 (27) Tumor diameter (cm) Our initial study, which suggested ®ve candidate tumor 51 18 (14) suppressor loci on chromosome 9, included 139 1 ± 3 55 (44) patients with initial low stage tumors in whom DNA 43 53 (42) was available from the initial tumor and from Unknown 7 Recurrence peripheral blood lymphocytes (Simoneau et al., 1999). No 37 (28) The present study includes 133 of these patients for Yes w/o progression 92 (69) whom we have complete follow-up data for up to 8 Yes w progression 4 (3) years with a median follow-up time of 5.6 years (Table 1). The mean age of these 133 patients was 61.7 years and the male : female ratio was 3 : 1. Most of their bladder tumors were stage Ta (105/133), single (97/133) LOH with sex of patients was even more important in and 3 cm or less in diameter (73/126). Grade 1 those with tumors having deletions only on 9q papillary tumors were found in 44 (33%) of the (P=0.01). To the opposite, no statistically signi®cant patients, while grade 2 tumors were found in 77 dierence was observed in 9p deletions, according to (58%) and grade 3 in 12 (9%). Only 21 patients had sex although all patients with 9p deletions only, were never smoked a cigarette. Recurrence occurred in 96 male. Patients who had a smoking history (past and patients but only four of these suered progression to current smokers) had similar frequencies of deletions as muscle-invasive disease. lifetime non-smokers (P=0.583). Similarly, no signi®- cant dierence was observed according to the age of patients. Chromosome 9 deletions in relation to patient and tumor characteristics Chromosome 9 deletions in relation to recurrence The frequency of allelic deletions (LOH) for each marker analysed along chromosome 9 is illustrated in The most important clinical outcome following initial Figure 1 according to pathological characteristics of TUR in this population was recurrence of super®cial the tumors. There was a great deal of variability in tumors which occurred in 72% (96/133) of patients LOH frequency for the dierent microsatellite markers during follow-up (Table 1). The risk of recurrence analysed but overall, deletions on 9q were twice as during that period was higher for patients with common as deletions on 9p in our tumor population deletions on chromosome 9: 83% of patients with (Simoneau et al., 1999). There was no signi®cant bladder tumors presenting at least one LOH on dierence in LOH frequency according to stage or chromosome 9 suered a recurrence compared to number of tumors (Figure 1a,c). However, for all 61% of patients with tumors without chromosome 9 markers analysed, an association was found between LOH. To determine whether speci®c regions of the the frequency of deletion and grade with grade 2 ± 3 chromosome were associated with a higher risk of tumors having a higher frequency of deletion than recurrence, recurrence rate ratios associated with the grade 1 tumors (P=0.004) (Table 2, Figure 1b).