Occupational history and genetic N-acetyltransferase polymorphism of urothelial cancer patients in

genetic N-acetyltransferase polymorphism of urothelial cancer patients Leverkusen, Scand Work Objectives The study was designed to realize possible shifts in the ratio of slow to fast acetylators within a group of 196 urothelial cancer patients in an area with earlier benzidine production. Methods The subjects were interviewed for occupational and nonoccupational risk factors. The patients were phenotyped for N-acetyltransferase 2 (NAT2) by Grant's caffeine test. A subgroup of 54 patients was addition-ally genotyped for NAT2. Results The antirnode in the NAT2 phenotyping with the caffeine test (AFMU: I X ratio) was 1 .O, as evidenced by additional genotyping of the subgr'oup of 54 patients. The psevalence of slow acetylators in the entire group of bladder cancer patients was 55%, in accordance with published figures for European populations. In a subgroup of 40 patients with occupational histories as workers in chemical or rubber industries 65% were slow acetylators. In a further subgroup of 28 cases having specifically worlted at che~nical production sites of the local chemical industry, 68% were slow acetylators. study shows no increased prevalence of slow acetylators arnong usothelial cancer patients in cornparison with the normal population. However, in subgroups of cases with a likelihood of past occupational contacts with aromatic amines, there was a trend towards a higher representation of slow acetylators. This finding is in accordance with observations of others that the percentage of slow acetylators in urothelial cancer patients is generally decreasing, possibly because the production of benzidine and benzidine-based dyes ceased in the early 1970s.

The role of N-acetyltransferase (NAT2) polymorphism in the pathogenesis of human bladder cancer has been investigated in several studies using the phenotyping of "slow" and "rapid acetylators. A key study was published by Cartwright et al in 1982 (1). In this hospitalbased case-referent study of 11 1 bladder cancer cases and 95 referents, the proportion of slow acetylators was 57% among the referents but 67% among the cases. In a subgroup of 23 cases with past occupational exposure to benzidine or 2-naphthylamine, as many as 22 slow acetylators were found.
In general, case-referent studies performed in Western countries on this subject have shown an excess of slow acetylators in patients suffering from bladder can-cer, compared with reference populations (2-1 3). This finding has been interpreted as being compatible with a generally increased risk of some 30% for bladder cancer among slow acetylators, compared with rapid acetylators (14).
However, it is important to note that similar studies in non-European populations (15)(16)(17) show results that differ from those of Western studies. Therefore the study of ethnic differences in this specific field is important. In particular, a recent study perfornled i11 China showed a prevalence of rapid acetylators in bladder cancer patients with past occupational contact to benzidi~le (1 7). In this study only 23.3% of the referents were slow acetylators. This finding is in marked contrast to, for example, obser-1 vations of Lewalter & Miksche (vide infra) at the benzidine manufacturing Bayer plant in Leverltusen (18) and points to ethnic dirfesences in the pathways of the metabolic activation of benzidine and related aromatic amines. Quantitative differences among the results of Wesler~l studies (2-13) might be explained by differences in environmental or occupational exposures to aromatic ainines, in addition to possible differences in the particular genetic or ethnic bacltground.
In view of the debated inconsistencies in the previous studies (1 9) further research in this field is needed. In a recent study based on the NAT2 gellotyping of 160 bladder cancer patients and 460 reference subjects in Berlin, Germany, 61.1% of the referents and 65.6% of the bladder cancer cases were slow acetylator genotypes. This very slight difference was not statistically significant (1 3). In general, it appears that the differences in acetylator genotypes or phenotypes between referents and unselected bladder cancer cases in Western populations are decreasing with time, probably indicating a decreasing impact of classical aromatic a~nines (eg, benzidine, 4-aminodiphenyl, 2-naphthylamine) on the general population.
The largest group of bladder cancer patients with a proved previous occupational exposure to benzidine has been studied at Bayer, Leverkusen, Germany, by Lewalter & Miksche (18). Phenotyping of 331 persons having worlted between 1951 and 1967 in the production of benzidine revealed 160 slow acetylators (48%) and 17 1 rapid acetylators (52%). The entire group of workers had been under medical surveilla~lce for 16 years at the time of publication. As many as 92 persons of this group were finally found to suffer from urothelial neoplasms, among them 75 being slow acetylators (82%) and 17 being rapid acetylators (1 8%). This unique local bacltground has now promoted further investigations. The importance of benzidine and related aromatic alnines for the manufacture of dyes, especially in Germany, had been a matter of a previous publicatio~l (20).
Our study was designed to realize shifts in the ratio of slow to fast acetylators within a group of urothelial cancer patients in an area with a specific chemical industry, along with past employment in this local industry, and to cornpare the circutnstances of these patients with the situations described in earlier publications (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13).

Bladder cancer cases
Our study was performed in 1991-1993 with patients with urothelial cancer at the Department of Urology of the Leverltusen Hospital. We used NAT2 phenotyping by the caffeine method (21). Part of the patients were also genotyped by allele-specific polymerase chain reaction (PCR) and restriction fragment length polymorphis~n studies. The genotype analyses were used to calibrate the antimode of the urinary caffeine metabolite ratio (AFMU:lX) in the phenotyping assay. Genotyping was done with the same ~nethodology and in the same laboratory of a preceding study by Cascorbi et a1 (13).
The hospital-based investigation was performed between June 1991 and June 1993 on 129 in-patients and 67 out-patients of the Department of Urology of the Leverkusen Hospital [I60 men, 36 women, total 196, nleall age 69 (SD 11.5) years]. These patients were first histologically diagnosed between 1977 and 1993 to suffer from a tumor of the urotheliuln and were treated using different therapeutic regimens.
Written consent was obtained from all the patients who participated in the study, which included NAT2 phenotyping or genotypi~lg. Turnors of the urogenital tract other (ha11 those of the urothelium (eg, squamouscell carcinomas, adenocarcinornas, sarcomas) were not included in the study. All the persons participating in the study were of European origin. The same physician obtained their history using a questionnaire in a standardized interview. (For details, see reference 22.) The detailed medical histories, smoking habits, drug intake and occupational histories, and, particularly, any type of work in chemical and related industries were recorded.
Memberships of particular health i~lsurailce funds were recorded. Within the group of patients, a first subgroup was formed of all patients who were members of the BKK Bayer (BKK = Betriebsltranltenltasse = health insurance Fund). This subgroup (I) included all those presently employed at Bayer, their dependerlts and the retired ror~ner e~nployees with their dependents (N = 66).
The classification of occupations followed the official list of occupations given by the German Federal Office of Statistics (23).
The interview of the 196 patients with respect to occupational activities led to a total of 379 occupations or occupational functions which lasted for at least one year. Accordi~lg to these occupations, a second subgroup (TI) of patients (39 men and 1 woman) was formed of blue-collar workers in the chemical or rubber iildustry (including classification numbers 141 x "Chemiebetriebswerker" (chemical production worlter); 142x "Chemielaborwerlter" (laboratory worlter); 633x "Chernielaborant" (laboratory assistant); 145x "Gum~nihersteller, -verarbeiter, Vulkaniseur" (rubber manufacturing, processing, curing). Out of this subgroup, the patients who had worlted at chemical production sites (141x "Chemiebetriebswerker") were also separately evaluated as a further subgroup (111) containing 27 men and 1 woman.
Eleven patients had been involved in the production of dyes; among them, six had been compensated for occupationally induced bladder cancer (BK 1301 of the German list of occupational diseases). Dye application was reported by 16 patients; one of them having received compensation. One patient had had contact with dyes in a research laboratory. Two maintenance workers employed at Bayer had been in contact with a large number of chemicals; they had been compensated for bladder cancer induced by aromatic amines. Some patients reported occupational contacts with specific aromatic amines, aniline contact being reported by eight patients (two being compensated), benzidine by six patients (three compensated), Znaphthylamine by two patients (both compensated), I-naphthylamine by one patient, and aromatic amines, in general, by three patients (one compensated). It should be noted that the term "aniline" had also been practically used in former times for aromatic amines in general.

N-acetyltransferase genotyping
Genotyping for "rapid" and "slow" alleles of the NAT2 gene was performed on a subgroup of 54 patients randomly selected from a total of 196 bladder cancer patients, according to the method of Cascorbi et a1 (24).
Genomic DNA (deoxyribonucleic acid) was prepared from leucocytes of anticoagulated blood samples by phenol extraction according to standard methods. Frequent mutations of the human NAT2 gene were detected according to the fundamentals given by Grant (25). The mutation T341C was identified by allele-specific PCR, and the mutations C282T, C481T, G590A, A803G, and G857A by PCR amplification of parts of the NAT2 gene and subsequent digestion with restriction enzymes Fok I, Kpn I, Taq I, Dde I, and Batn HI, respectively. Distinct alleles were defined by the constellation of mutations (table 1). With the use of six known "slow" and one "rapid" allele all combinations of slow alleles only were addressed as slow acetylator genotypes; the remainder were co~lsidered rapid acetylator genotypes according to the new nomenclature of N-acetyltransferases (26).

N-acetyltransferase phenotyping
The caffeine test method of Grant et a1 (21) was used for the NAT2 phenotyping. The subjects drank one to two cups of a caffeine-containing beverage (mostly coffee,  196). No differences were observed in the individual phenotypes, based on these parallel sainples, and both values (AFMU:IX ratios) of the same subject were always very similar. Hence, for persons from whom two urine samples were analyzed, a mean value of the two samples was used in the further evaluation.
The urine was immediately adjusted to pH 3.5 by the addition of small quantities of 1 M hydrochloric acid, and then stored in 10-ml "monovettes" (plastic tubes) at -20°C until processing. The individual storage times did not exceed six weeks. Quantitation of the caffeine metabolites AFMU (5-acetylamino-6-formylamino-3methyluracil; acetylated metabolite) and 1X (l-methylxanthine; non-acetylated metabolite) was done by highpressure liquid chromatography (HPLC), using the standard addition procedure. Standard 1 X was a commercial product of Sigma (Deisenhofen, Germany); AFMU was synthesized by Rohrkasten et a1 (27), according to a modification of the method of Fink et a1 (28). The final product contained 6% AAMU (5-acetylamino-6-amino-3-methyluracil), which was appropriately considered in the calculations based on the standard solution. Stock solutions of AFMU (2.5 mg AFMU in 10 ml of 0.05% acetic acid) were stored frozen in 1-ml fractions and thawed before use. Stock solutions of 1X (1.5 mg 1X in 10 ml of 0.05% acetic acid) were prepared weekly and stored at 4OC until use.
The HPLC procedure followed the method of Grant et al (21). By use of the standard addition method, the molar ratio of AFMU over 1X was calculated for each individual urine sample. The validity of use of the AFMU:lX molar ratio, even for alkaline urines, for NAT2 phenotyping has recently been documented by Butler et a1 (29).
In previous studies using the caffeine test to study acetylator phenotypes, probit tra~lsformations were used to determine the distribution of slow versus rapid acetylators (30). They clearly showed a bimodal distribution of NAT2 phenotypes. Based on probit transfortnation of individual molar AFMU:lX ratios, an antimode of 0.6 between slow and rapid acetylators had been inferred by Butler et a1 (29), both for smokers and nonsmokers. I11 other laboratories, different values for the antiinode had been in use, for example, 0.48 by Grant et a1 (21) and 0.85 by Braz Vieira de Silva Pontes et a1 (31). The numerical value of this antimode is obviously laboratory-dependent (see references 21 and 32); the instability of AFMU standards is probably an important point in this respect. In an early stage of investigation (43 cases which had been both genotyped and phenotyped at that time), an interim assessment of genotypes and corre-sponding phenotypes was made based on an antimode (AFMU: 1X) of 0.48 (21). It appeared at that time that 8 out of 29 rapid acetylator phenotypes did not correspond with the genotype, while only 1 out of 14 slow acetylator phenotypes differed from the corresponding genotype. Hence the critical question of the AFMU: IX antirnode was evaluated in detail. (See the Results section.) Using this evaluation we finally confirmed an antimode of the molar AFMU: 1X ratio of 1.0 to distinguish between slow and rapid acetylator phenotypes.

Coherence between N-acetyltransferase genotypes and phenotypes (caffeine test)
The introductioil of the caffeine test by Grant et a1 (21,30) for NAT2 phenotyping has particularly facilitated studies of diseased patients. However, this test requires careful consideration of the antimode of the urinary caffeine metabolite ratio (AFMU: IX) that distinguishes between slow and rapid acetylators. Fifty-four of the bladder cancer patients studied were also genotyped in a blind manner, in addition to the phenotyping by the caffeine test.
There was no significant difference in the prevalence of NAT2 phenotypes with age or smolcing habits or between in-and out-patients of the hospital. In addition, a careful examination of drugs used or prescribed did not point to i,,f!se:i~es of drug intalce on NAT2 phenotyping. The differentiation of patients according to smolcing histories and histopathological tumor gradings did not indi-  Because of the geographic location of the hospital where the study was performed, close to a chemical enterprise with a major history in the manufacturing and processing of benzidine and other aromatic ainines, it was of interest to concentrate on subgroups of diseased persons with an occupational history related to the local chemical industry. In this context, it is important that the production of benzidine at Bayer had ceased in 1967 (18) and that the production of benzidine-based dyes, in general, had ended in Germany in 1971, according to an When the subjects were exanlined according to their subgroups, 41 persons (62%) were found to be slow mctabolizers.
Although these differences, owing to the number of available cases, were not statistically significant, they followed the expected trend towards a higher representation of slow acetylators in groups of cases with a likelihood of past occupational contact with aromatic amines. (See table 2 on page 335.)

Discussion
Previous hospital-based investigations in different Western countries have uniformly demonstrated a s~nall excess of slow acetylators in groups of bladder cancer patients, in comparison with healthy referents. However, differences appear between studies (table 3) with regard to the percentage of both slow acetylators in "normal" reference populations and in bladder cancer patients. Moreover, owing to ~nethodological factors, the positioning of the "antimode" in NAT2-phenotyping studies using the caffeine test (21) is of paramount methodological importance.
The present data and those of an earlier (13) study, in principle, are in line with the results of others (table 3). However, due to the unique geographic location of the group of patients in our study, close to a inajor chemical enterprise with a special history in the manufacture of aromatic amines, the behavior of the subgroups of patients with an occupational background in the local chemical industry was of particular interest. The results (table 2) suggest that the probability of a past occupational contact with aromatic amines for individuals within the group of bladder cancer patients corresponds with a tendency towards an overrepresentation of slow acetylators. In this respect, the extreme is marked by the group of worlcers from former benzidine production at Bayer (until 1967) who later suffered from bladder cancer, as published by Lewalter & Milcsche (1 8).
Small shifts towards a higher representation of slow acetylators in studies of bladder cancer patients without a special occupational history have mostly been assigned to the effects of aromatic a~nines in tobacco smoke. In this context, environmental tobacco smolce should also be considered (35) in that it contains at least 17 different mono-and bicyclic aromatic amines, including 2-naphthylamine and 4-aminodiphenyl (36). 011 the other hand, recent studies on the polyrnorphisrn of the huinan glutathione transferase GSTM1 (37) and its relevance to human lung (38) and bladder (39) cancer have introduced the concept that the GSTMl gene ought to have a protective role in tobacco-induced bladder cancer, possibly through increased detoxification of biologically reactive i~lternledlates of polycyclic aromatic hydrocarbons.
Hence, further study is required to elucidate the relative roles of aromatic a~nines and of polycyclic aromatic hydrocarbons in the origin of human bladder cancer, both in "normal" populations and in groups with different types of occupational exposures.