Vol. 5. 33-39. Jwiuarv 1996 Cancer Epidemiology, Biomarkers & Prevention 33
A Molecular and Epidemiological Study on Bladder Cancer: p.53 Mutations, Tobacco Smoking, and Occupational Exposure to Asbestos1
Annamaria Kannio, Maaret Ridanp#{228}a,2 Heikki Koskinen, ical type of bladder cancer is ICC,3 accounting for >90% of all Timo Partanen, Sisko Anttila, Yrj#{246}Collan, bladder cancer. Tobacco smoking is the most important exter- Eino Hietanen, Harri Vainio, and nal cause of bladder cancer, with an estimated attributable Kirsti Husgafvel-Pursiainen3 fraction of 40% or more in males and 30% in women in Departments of Industrial Hygiene and Toxicology [A. K., M. R., H. V., developed countries (3). K. H-P.[. Occupational Medicine [H. K.. S. Al, and Epidemiology and A number of occupational exposures are known to be Biostatistics [T. P.1. Finnish Institute of Occupational Health, Topeliuksenkatu associated with an elevated risk of bladder cancer. Prominent 41 aA. FIN-(X)25() Helsinki, Finland, and Departments of Pathology IY. C.] excesses have been associated with exposure to aromatic and Clinical Physiology [E. H.]. Turku University Central Hospital. Kiinamyllynkatu 10, FIN-20520 Turku. Finland amines in several industrial processes (3). The role of asbestos exposure in urological cancers has been less well studied and remains unclear. Exposure to asbestos has been related to an Abstract increased risk of kidney cancer (4- 1 1 ). In occupationally ex- posed individuals, the lungs. lymph nodes, and pleura have the In this study, we found an unexpected association (crude highest concentrations of asbestos fibers, but high levels of odds ratio = 2.8; 95% confidence interval = 0.9-8.4) asbestos fibers are also found in the kidneys (12-14). Further- between definite work-related exposure to asbestos and more, asbestos fibers have been demonstrated in the urine of carcinoma of the urinary bladder in a small group of individuals exposed to asbestos (15, 16). A few studies have patients (n = 28) initially recruited as referents for an epidemiological feasibility study on the occupational causes been conducted on the possible association between bladder of lung cancer. We extended the study by using molecular cancer and occupational exposure to asbestos. mostly with negative results (17-19). methods to examine mutations in the p53 tumor suppressor A prospective case-referent feasibility study on occupa- gene in the same cases of bladder cancers. The same tional causes of lung cancer was initiated in Finland in 1988. number of archival samples of transitional cell carcinoma, mainly of grade 3, were added to the analysis. We failed to Consecutive bladder cancer patients attending the urological clinic were recruited as referents for the lung cancer cases. show any association between occupational exposure to Their past occupational exposure to asbestos was assessed from asbestos and p53 mutations among bladder cancer patients. job histories obtained in personal interviews (20). This frame- We observed an increasing occurrence ofp53 mutations in work, and the fact that the identification and collection of nonsmokers (5 of 17, 29%), former smokers (8 of 21, 38%), archival tissue samples from the epidemiological pilot study and current smokers (9 of 16, 56%) in that order; however, was possible, prompted us to conduct a molecular genetic study this was not statistically significant The most prevalent type of mutation was G:C to A:T transition. Tumor grade was using the same subjects. To date, mutations of the p53 gene appear to be among the most common genetic changes in not associated with the frequency of mutations, but the bladder cancer, with the reported frequency varying between 20 higher stage (T3-T4) tumors appeared to have mutations and 60% (21 , 22). As p53 has been suggested as a useful target more frequently than did the less invasive tumors (T1-T2). gene for studies focusing on the role of mutations in carcino- genesis (23, 24), we sought to use the p53 mutation pattern as Introduction a potential biomarker for environmental exposure (in this case The incidence of the cancer of the urinary bladder has been tobacco smoking and asbestos exposure) in bladder cancer. increasing in almost all countries in Europe, including Finland We report here the observations from the epidemiological ( I ). Bladder cancer is the third most common site of primary feasibility study. We examined archival tumor tissue samples cancer among men in Finland (2). The most common histolog- from 55 bladder cancer patients with mainly grade 2-3 TCC. Almost one-half of the patients had been exposed to asbestos at work. Mutations of the p53 gene were analyzed in relation to smoking habits and to interview-based asbestos exposure data. Received 5/23/95: revised 8/3 1/95: accepted 9/5/95. Mutation findings were also compared with the clinical features The costs of publication of this article were defrayed in part by the payment of of the tumors. page charges. This article must therefore be hereby marked advertisetnent in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
I This study was supported in part by the Academy of Finland and the Finnish Subjects and Methods Cancer Foundation. 2 Present address: Department of Medical Genetics. University of Helsinki. P.O. Epidemiological Pilot Study Box 21. FIN-(XX)14 University of Helsinki, Finland. The cases were 28 bladder carcinomas out of 33 patients 3 The abbreviations used are: TCC. transitional cell carcinoma: OR, odds ratio; CI. confidence interval; DGGE. denaturing gradient gel electrophoresis; NSCLC. admitted consecutively for diagnostic or surveillance purposes non-small cell lung cancer; 4-ABP, 4-aminobiphenyl. to the Surgery Clinic, Turku University Central Hospital
Downloaded from cebp.aacrjournals.org on September 30, 2021. © 1996 American Association for Cancer Research. $4 Asbestos, Smoking. and p53 Mutations in Bladder Cancer
Table 1 Asbestos cx posure of bl adder cancer ca ses and refer ents. Data fr om the feasibil ity study (group I).
Cases Referents Asbestos exposure OR 95% Cl Ft (‘:4) t (%)
Definite 16 (57) 9 (32) 2.75 0.90-8.37 Possible I (4) 2 (7) 0.77 0.06-.9.6() Definite plus possible 17 )6l) 11 (39) 2.39 0.82-6.98 None II (39) 17 (61) 1.00
All 28 (100) 28 (100)
(Turku, Finland) between October 1988 and December 1988. Table 2 Occupations of bladder cancer patients and referents who had been There were 23 (82%) men and 5 (1 8%) women; the median age exposed to asbestos. Data from the feasibility study (group I). was 70 years (range, 41-79 years). Nine (32%) of the patients were current smokers, 12 (43%) former smokers, and 7 (25%) Cases Referents Occupation - had never smoked. Twenty-eight population referents (23 men ‘I (‘A’) it ((4.) and 5 women) were selected from 58 persons who were invited Construction worker I I (69) 6 (67) for routine radiographical screening for tuberculosis in three Shipyard worker 2 (13) 1 (II) towns in the health district of Turku University Central Hos- Brake repairer I (6) 2 (22) pital. The median age of the referents was 64 years (range, Foundry worker 1 (6) 52-83 years). Eight (29%) of the referents were current smok- Steam engine maintenance worker I (6) ers, eight (29%) were exsmokers, and 12 (42%) had never All 16 (100) 9 ) I(S)) smoked. Lifelong job histories, comprising job titles, task de- scriptions and duration. and smoking data were obtained from personal interviews with the subjects. completed by a trained interviewer. Patients who had quit smoking 1 year before the was performed once with xylene, followed by two 95% and two interview were classified as exsmokers. The cases were inter- 70% ethanol rinses, and samples were left to dry. Tissues were viewed at the bedside, the referents in connection with the pelleted between each extraction. The samples were incubated in screening procedure. Therefore, the interviewer was not 1 mg/mI protease (Sigma Chemical Co.), 10 mrsi Tris-HC1 (pH blinded to the bladder cancer referent status of the subjects. 7.5), 1 mM EDTA, 0. 1 M NaCl, and I % SDS for 3 days at 50#{176}C, followed by phenol-chloroform extraction and ethanol precipita- p53 Mutation Analysis tion. DNA was diluted into Tris-EDTA buffer, pH 7.4. The quality Tumor Samples. For diagnostic purposes, urinary bladder of DNA was determined on 1% agarose gel. Two .d of genomic cancer samples were obtained during surgery to remove tumors DNA were used in a 50 pi PCR reaction using 2.5 units Taq DNA or by biopsy at the Turku University Central Hospital between Polymerase (Promega) with 1-1.5 mrsi MgCl.,, 25 p concentra- 1975 and 1993. Formalin-fixed paraffin-embedded archival tis- tion of each dNTP (Pharmacia), 50 mrvi KC1, 10 mM Tris-HC1 (pH sue samples, which contained tumor tissue, were obtained for 9.0), 0.1% Triton X-l00, and 0.5-1.0 p.M of each primer (Institute this study from 27 bladder cancer patients recruited to the of Biotechnology, University of Helsinki, Helsinki, Finland). One feasibility study. Representative samples of nontumor tissue of each primer pair contained a 40-bp GC-rich sequence (GC- were not available. In this group (group 1 ), there were nine clamp), which incorporated into the final PCR product. The three- current smokers, ten former smokers, and eight who had never step amplification cycle (1 mm at 96#{176}Cfordenaturation. 1 mm at smoked (23 men and 4 women; mean age, 70 years; range 54-60#{176}C for annealing, and I mm at 74#{176}C)for synthesis was 41-77 years); 46 tissue samples were received. Twenty-three of repeated 30-40 times. The primer sequences for exons 6-8 are the cases were classified as grade I -3 (TCC) of the bladder, one given in Ridanp#{228}#{228}et a!. (27) and for exon 5 in Beck et a!. (28). TCC in situ, one papilloma, one urethra polyp, and one TCC of DGGE Gel Runs and Sequencing. The gels (1-mm thick) the ureter. Group 2 consisted of random cases of TCC, mainly contained 7.5% acrylamide in I X Tris-acetate-EDTA buffer grade 3. Of these 28 patients, 7 were current smokers, 1 1 were (pH 8.0), N,N’-diallyltartaramide as the cross-linker and vary- former smokers, and 9 were nonsmokers, four of whom had ing denaturant concentrations ( I 00% denaturant corresponds to never smoked; for 1 case no data on smoking were available (23 7 M urea and 40% deionized formamide). N,N,N’,N’-tetrameth- men and 5 women; mean, 70 years; range, 43-90 years). Smok- ylethylenediamide and ammonium persulfate were used to ac- ing histories for group 2 were obtained from hospital records tivate the gels polymerization (29). The gels were run in 1 X and based on a patient’s smoking status when admitted to the Tris acetate-EDTA buffer at 60#{176}Casdescribed by Ridanp#{228}#{228}et hospital. Thirty-one tissue samples were received for analysis. a!. (27). The running time for exon 5 was 5 h at 10 W or 150 Tumors were graded histopathologically (25) and staged using V for the parallel gel (40-70% gradient of denaturants). Each the tumor-node-metastasis classification (26). Occupational p53 mutation found in parallel gel runs was also confirmed in histories of the patients in group 2 were not sought. perpendicular gels. The gels were stained with ethidium bro- DNA Extraction and PCR Amplification from the Paraffin- mide and photographed by using a UV transilluminator. For embedded Archival Tissues. Two 25-gm-thick sections were sequencing, PCR amplification was carried out with a biotiny- sliced from each paraffin block and placed in an Eppendorf tube. lated primer. Direct sequencing (Sequenase version 2.0 proto- The microtome blade was carefully cleaned with ethanol between col; United States Biochemical) of the PCR product was per- samples to avoid cross-contamination. Two parallel samples were formed by using magnetic beads (Dynabeads M280 cut from each block. For mutation analysis, DNA was isolated and Streptavidin; Dynal) as solid support. Sequencing gels con- extracted two to three times from each block. Deparaffinization sisted of 4.5% polyacrylamide and 7 M urea (Bio-Rad). The
Downloaded from cebp.aacrjournals.org on September 30, 2021. © 1996 American Association for Cancer Research. Cancer Epidemiology, Biomarkers & Prevention 35
Table 3 Tobacco smoking and cI inical status of the bladder can cer cases who had p53 mutations
Sample no. Sex/age (yr) Smoking” Asbestos exposure’ Graded Stage Exon with mutation
Group I UB 1 I M/56 CS + I ND 6,7 UB26 M/63 CS + 2 T, 7 UBI3 M/70 CS + 3 T4 5 UB24 M/57 ES + I T, 6 UBI M/41 ES + 2 T3 6 UB9 M177 ES + 2 T3 6
UBI9 F/52 CS - I T1 6
UB2O F173 CS - 3 T 6.8
UBI6’ M/64 CS - 3 ND 6
UB7 M/51 ES - 2 T, 6
UB25 F/60 NS - 3 T, 6
UBI5 M178 NS - 2 ND 5
Group 2 UB35 M167 ES ND 2 T3 8 UB32 M/65 CS ND 3 T 7 UBMY’ M/43 CS ND 3 T2 S UB54 M/60 CS ND 3 ND 6 UB4I M/7l ES ND 3 T 5 UB36 F/67 ES ND 3 T, 6 UB52 M/76 ES ND 3 T2 8 UB42 M15l NS ND 3 T3 8 UB56 M170 NS ND 3 T 8 UB47 M/78 NS ND 3 ND 8
“ M. male: F. female. ‘, CS. current smoker: ES. exsmoker; NS, nonsmoker. , +, exposed: -. not exposed: ND, not determined.
‘I Histological grading according to WHO recommendations (25). ,. Pathological staging according to the tumor-node-metastasis classification (26): ND, not determined. , TCC originating from ureter.
S Exon 6 (codon 213) polymorphism also detected.
sequencing primers used were taken from Ridanpaa et a!. (27). with definite exposure to asbestos ranged from 6 to 44 years At least three separate PCR products from each sample were (median, 20 years) among the cases and from 10 to 47 years sequenced to confirm the result. (median, 21 years) among the referents. Of the 28 primary bladder cancer cases in group 1 , 17 (61%) had been exposed to asbestos at work. Of these, 16 Statistical Analysis (94%) were considered to have been definitely exposed. A In bivariate binary analyses, simple logit routines were used to crude OR of 2.8 (95% Cl = 0.9-8.4) was calculated for estimate the OR and their 95% CI for the outcome parameters. definite exposure to asbestos against no asbestos exposure and A small number of multivariate maximum likelihood logistic 2.4 (95% CI = 0.8-7.0) for the pooled category of definite to models were fitted to the mutation data. ORs and 95% CIs for possible exposure against no asbestos exposure (Table 1). the dependent parameter, the presence/absence of p53 muta- p53 Mutations in TCC Patients. Altogether, 77 paraffin- tion, were estimated conditionally on the determinants (previ- embedded tissue samples from 55 patients with urothelial can- ous occupational exposure to asbestos, tobacco smoking. tumor cer were analyzed for mutations in exons 5-8 of the p53 gene differentiation, and stage) and the potential confounders (gen- by using DGGE. In the group of patients recruited for the der and age). The following codings were applied: gender (male feasibility study, a paraffin-embedded tissue sample containing versus female); age at the time of tumor sampling (67-90 years tumor tissue was obtained from 27 cases (Table 3; group 1). A versus younger); tobacco smoking (current or former versus p53 mutation in the tumor tissue was found by DGGE analysis no); asbestos exposure (yes versus no or unknown); missing in 12 (44%) cases. The TCC cases analyzed for mutation in data for asbestos exposure (missing versus known); tumor group I included 15 patients who were classified as having differentiation (grade 3 versus grades 1-2); and tumor stage (T1 been definitely exposed to asbestos; 6 (40%) of them had a versus T,-T4). The missing asbestos exposure variate was mutation, and they were either current or former smokers (Ta- added because data on exposure to asbestos were missing for a ble 3). In addition, we analyzed mutations in 28 random TCC considerable number of subjects. samples, mainly of grade 3 (Table 3; group 2); in this group, 10 (36%) mutations were found. Results From the 22 archival samples, which repeatedly tested Asbestos Exposure in Bladder Cancer Patients. Table 1 positive for mutation using DOGE, we were able to sequence shows the distribution of asbestos exposure among the bladder 14 (Table 4). A representative DGGE gel run of DNA frag- cancer cases and their referents (group 1 ). Table 2 shows the ments amplified from exon 6 showing alterations from the wild latest occupations for those bladder cancer cases and referents type is presented in Fig. 1A. In Fig. 1, a polymorphism in exon who were classified as having been definitely exposed to as- 6 (codon 213) can be distinguished from a mutation in the same bestos at work. The duration of employment in occupations exon by the pattern of migrating DNA fragments. Examples of
Downloaded from cebp.aacrjournals.org on September 30, 2021. © 1996 American Association for Cancer Research. 36 Asbestos, Smoking, and p53 Mutations in Bladder Cancer
8 Table 4 Types of j,53 mutations sequenced from the transitional cell carcinoma sainples
. , Amino acid Sample no. Exon ( odon Base change change 7-. UB4I 5 155 ACC- CCC Thr- Pro UBI3 5 173 GTG- GGG Val-Gly UBI5 5 173 GTG-’GGG Val-’Gly UB36 6 196 CGA- CCA Arg-’ Pro 6 UB2O” 6 213 CGA-*CAA Arg-’Gln UB26 7 234 TAC- TCC Tyr-. 5cr FBI I’ 7 236 TAC-*GAC Tyr-’Asp UB32 7 248 CGG-’TGG Arg-.Trp n=5 5 UB42 S 271 GAG-#{176}AAG Glu-#{176}Lys UB47 S 273 CGT-#{176}TGT Arg-’Cys ri UB35 S 28)) AGA-’ACA Arg-Thr 0 UB52 8 280 AGA-AAA Arg-Lys 4 [ UBSO S 283 CGC-*TGC Arg-Cys UB2O” S 29() CGC-CAC Arg- His E “ Boldface and underlining indicate base change.
,, Patient with two sequenced tiutattons. If) ‘ Patient with one sequenced and one DGGE.detected mutation. c_ C