Elevated 4-Aminobiphenyl and 2,6-Dimethylaniline Hemoglobin Adducts and Increased Risk of Bladder Cancer among Lifelong Nonsmokers—The Shanghai Bladder Cancer Study

Early observational studies show significantly increased incidence of bladder cancer among workers occupationally exposed to arylamines in the workplace, and subsequent studies led to the discovery and confirmation of certain arylamines, including 4-aminobiphenyl (4-ABP), 2-naphthylamine, and benzidine, to be carcinogenic to humans. They are classified as group I carcinogens by the International Agency for Research on Cancer (IARC; ref. 1).

Following governmental regulation of industrial use of arylamines in 1970s, tobacco smoke emerged as the major source of 4-ABP exposure in humans (2). Tobacco use is recognized as an important risk factor for bladder cancer. Smokers exhibit 2- to 3-fold of excess risk for bladder cancer compared with nonsmokers (3). In our case–control study in Los Angeles, California, smokers exhibit more than 3 times higher levels of 4-ABP hemoglobin adducts, a validated biomarker of 4-ABP exposure, than lifelong nonsmokers (4).

Although tobacco use is a known major risk factor for bladder cancer, it only accounts for roughly 50% of disease burden in the United States (5). Relatively little is known about the causes of nonactive smoking related bladder cancer. We have reported environmental tobacco smoke (ETS) exposure as a risk factor for nonsmoking-related bladder cancer (6, 7). The data were based on 2 separate but parallel case–control studies conducted in Los Angeles, California (high-risk region; The Los Angeles Bladder Cancer Study) and Shanghai, China (low-risk region; The Shanghai Bladder Cancer Study). The Los Angeles Study further implicates other diffuse sources of 4-ABP (besides active and passive smoking) as possible causal factors for nonsmoking-related bladder cancer.

More significantly, The Los Angeles Study examined hemoglobin adducts of nine alkylanilines, a subclass of arylamines that was previously unstudied in relation to bladder cancer risk, and found 3 of them, including 2,6-dimethylaniline (2,6-DMA), to be independently and significantly associated with bladder cancer risk among nonsmoking subjects at blood draw. Experimental studies have shown that administration of 2,6-DMA in the diet can induce adenomas and carcinomas as well as several sarcomas in the nasal cavity of rats. 2,6-DMA also produces subcutaneous fibromas and fibrosarcomas and increases the incidence of neoplastic nodules in the livers of rats (8, 9). 2,6-DMA is classified by IARC as a group 2B carcinogen (possibly carcinogenic to humans; refs. 9, 10).

Arylamines require metabolic activation, usually in the liver, to be transformed into fully carcinogenic agents (11). They are oxidized to N-hydroxylamines in vivo and react with hemoglobin to form adducts that may persist for as long as the hemoglobin adducts remains in circulation (4, 12). In the acidic environment of the bladder lumen, a derivative of the glucuronide conjugates of N-hydroxylamines can covalently bind to urothelial DNA and cause malignant transformation of urothelial cells, which may ultimately lead to bladder cancer (13). Hemoglobin adducts of arylamines are validated biomarkers of arylamine exposure in humans (14). These adducts represent both the uptake of arylamines and their genetically controlled metabolism in individual subjects.

Here, we report the findings of the Shanghai Bladder Cancer Study on hemoglobin adducts of 4-ABP and 2,6-DMA in relation to bladder cancer risk. The results are consistent with our prior novel findings in Los Angeles (4, 12) and strongly implicate arylamines as major causal agents of bladder cancer, not only among tobacco users but in lifelong nonsmokers as well.

The present study included participants of the Shanghai Bladder Cancer Study. The study design has been described in detail elsewhere (7). Briefly, patients with bladder cancer were identified through the Shanghai Cancer Registry, a population-based cancer registry covering the approximately 8 million residents of urban area in Shanghai, China, in the 1990s. Of the patients diagnosed with bladder cancer from July 1, 1995, to June 30, 1998, 708 were 25 to 74 year old who met our eligibility criteria for the study. Among the 708 patients, 56 died before we could contact them, 29 refused to be interviewed, and 42 were unable to be located. We interviewed the remaining 581 (82%) eligible patients between July 1996 and June 1999. The diagnosis of bladder cancer for 531 (91%) patients was made based on histopathologic evidence, whereas the remaining 50 (9%) patients diagnoses were based on positive computerized axial tomography scan and/or ultrasonograph with consistent clinical history. Control subjects were randomly selected from the urban population of Shanghai through the Residents Registry of the Shanghai Municipal Government. They were chosen to match the frequency distribution by sex and 5-year age groups of patients with bladder cancer. Among the 750 potential control subjects chosen, 604 (81%) eligible subjects were interviewed during the same time period as the cases. All subjects provided informed consent following procedures approved by the appropriate institutional review boards.

A trained interviewer conducted an in-person interview with each study subject using a structured questionnaire. The questionnaire asked for information on subjects' demographic characteristics, history of tobacco use, history of passive smoking (for nonsmokers only), consumption of beverages, use of hormones (for women only), medical history, usual adult diet, and occupational history. Cigarette smoking was defined as smoking 1 cigarette per day for at least 6 months. Lifelong nonsmokers were defined as subjects who were not smoking on a regular basis in the past 6 months before the reference date (2 years before cancer diagnosis for cases and 2 years before interview for controls) and who did not smoke any cigarettes in the past 6 months prior to the date of blood and urine sample collection. All subjects whose urinary level of cotinine was ≥75 ng/mL were defined as smokers at the time of urine sample collection.

All subjects were asked to donate blood and an overnight urine sample (ending with the first morning void) at the end of the in-person interview. A total of 513 (88% of interviewed) cases and 534 (88% of interviewed) controls provided a blood sample. Blood samples were collected in heparinized (10 mL) and nonheparinized (4 mL) tubes. Heparinized samples were fractioned into plasma, buffy coat, and erythrocytes on the day of the sample collection and were stored at −80°C. When picking up the overnight urine sample, the interviewer asked the subject about use of tobacco products during the past 60 days. Five hundred and thirty-five (92%) of case patients and 543 (90%) of interviewed control subjects donated an overnight urine sample. The urine samples were processed on the day of collection and acidified (400 mg of ascorbic acid per 20 mL of urine) before they were stored at −80°C until analysis.

Erythrocyte fractions, identifiable only by their code numbers, were set on dry ice and sent to the University of Pittsburgh, where they were stored at −30°C until assays for arylamine-Hb adducts were conducted. Hemoglobin adducts of 4-ABP adducts and 2,6-DMA were quantified using the internal standards [methyl-²H₆]-2,6-dimethylaniline and 4-fluoro-4′-aminobiphenyl as described previously (12, 15). Erythrocyte fractions were thawed and 1 mL of each were centrifuged at 1,000 × g for 10 minutes to pellet cell debris. The mixture was then dialyzed for 36 hours against 50 volumes of water using a dialysis bath at 4°C and changing dialysis bath water twice to remove noncovalently bound low–molecular weight compounds. The hemoglobin concentration in each sample was determined by Drabkin's method. Each dialyzed sample was then spiked with the internal standards, the sulfinamide bonds between the electrophilic forms of the arylamines and the sulfur of β93-cysteine residue of hemoglobin were hydrolyzed with aqueous NaOH. The freed arylamines were then extracted into hexane and derivatized with pentafluoropropionic anhydride. The resulting amides were concentrated under a stream of dry nitrogen, redissolved in iso-octane, and analyzed by GC-NICI-SIM-MS using a Carbowax capillary column.

Cotinine is a major proximate metabolite of nicotine but has a longer half-life than nicotine. Therefore, cotinine is a better biomarker for daily use of cigarettes than nicotine itself and other nicotine metabolites. In the present study, urinary cotinine was measured by the standard gas chromatographic-mass spectrometric method (GC-MS; ref. 16).

In the present study, we included 494 (85% of interviewed) case patients and 499 (83% of interviewed) control subjects with available measurement of arylamine hemoglobin adducts. Bladder cancer cases included in this study were comparable with those excluded from the study in terms of age at diagnosis, body mass index, level of education, and smoking status at reference (all P > 0.05). Given the skewed distribution of 4-ABP hemoglobin adducts, formal statistical test was conducted on logarithmically transformed values, and geometric (as opposed to arithmetic) means were presented. The analysis of covariance (ANCOVA) method was applied to identify determinants of 4-ABP in all control subjects. Number of cigarettes smoked per day (0, 1–<10, 10–<20, and 20+) at blood draw and urinary levels of cotinine (0, 1–<100, 100–<500, and 500+ ng/mL) were found to be significantly associated with levels of 4-ABP hemoglobin adducts. Thus, in all statistical analysis that included smokers at the time of blood/urine collection, the latter 2 factors were part of the covariates set in regression models, in addition to gender, age at blood draw, and level of education.

The χ² and nonparametric statistics methods were used to examine the difference in distributions of 2,6-DMA hemoglobin adducts across different exposures (i.e., smoking or urinary total cotinine) among controls given the skewed distribution of the data and small percentage (6%) of controls with detectable level of 2,6-DMA adducts. The same methods were used to examine the difference in distributions of 2,6-DMA adducts between cases and controls.

Unconditional logistic regression models were used to examine the associations between adducts and risk of bladder cancer. A high proportion of study subjects (72% controls and 91% cases) in our study had undetectable level of 2,6-DMA adducts. We classified subjects who had detectable level of 2,6-DMA into low (below median of positive values in controls) and high levels of exposure. The median value in controls was determined to be 115.9 pg/g hemoglobin. Similarly, we grouped subjects by quartiles based on the distribution of 4-ABP adducts in controls. The strength of the association between levels of adducts and bladder cancer risk was measured by ORs and their corresponding 95% confidence intervals (CI) and P values. We assessed the associations in total subjects and in subgroups defined by cigarette smoking status at blood draw and by smoking history over lifetime.

Statistical analyses were conducted using SAS version 9.2 (SAS Institute) statistical software package. All P values are 2-sided. P < 0.05 was considered statistically significant.

The mean age (±SD) of case patients at diagnosis of bladder cancer was 63.6 (±9.9), years whereas the mean age of control subjects at interview was 64.0 (±10.0) years (P = 0.74). Eighty percent of patients with bladder cancer and 77% of control subjects were men (P = 0.29). Case patients had similar distributions as control subjects by levels of education and body mass index (Table 1). More case patients were current smokers and exhibited greater number of pack-years of smoking than controls at the reference date (i.e., 2 years before cancer diagnosis for case patients and 2 years before interview for control subjects; Table 1). However, fewer patients with bladder cancer smoked cigarettes during the past 60 days before blood draw. Case patients and control subjects had comparable levels of total cotinine in urine collected around the same time as the blood draw (Table 1).

2,6-DMA adducts were detected in 136 (26%) of 484 bladder cancer cases and 31 (6%) of 499 control subjects (P < 0.001). Among control subjects, there was no statistically significant difference in 2,6-DMA adducts levels between current smokers at blood draw and lifelong nonsmokers or across different levels of smoking (i.e., number of cigarettes per day at blood draw or urinary total cotinine). Among subjects with positive 2,6-DMA adducts, the difference in 2,6-DMA adducts between cases and controls [median: 233.9 in cases vs. 115.9 pg/g hemoglobin in controls; geometric mean: 232.0 (95% CI, 176.3–307.6) in cases vs. 162.2 (95% CI, 89.6–293.6) pg/g hemoglobin] was not statistically significant (P > 0.13) (data not shown). Compared with subjects who had undetectable adducts, individuals exhibiting low (below median of positive values) and high (above median of positive values) levels had an OR of 4.08 (95% CI, 2.16–7.70) and 7.38 (95% CI, 4.16–13.07), respectively, for bladder cancer after adjustment for potential confounders (P_trend < 0.001; Table 2). The positive association between 2,6-DMA hemoglobin adducts and bladder cancer risk was present separately in smokers and nonsmokers at blood draw, as well as in lifelong nonsmokers.

Among controls, geometric means of 4-ABP adducts increased with increasing number of cigarettes smoked per day at blood draw as well as increasing levels of urinary cotinine (both P_trend < 0.001; Table 3). Levels of 4-ABP adducts were comparable between men and women after differences in number of cigarettes per day, and urinary total cotinine between the 2 sexes were taken into account (data not shown). There was no correlation between 4-ABP and 2,6-DMA adducts among control subjects (correlation coefficient = 0.03, P = 0.764).

Table 4 shows the associations between 4-ABP hemoglobin adducts and risk of bladder cancer. No statistically significant associations were observed in total subjects. However, among lifelong nonsmokers, a dose-dependent, statistically significant association between 4-ABP adducts and bladder cancer was noted (P_trend = 0.009). Subjects possessing the highest quartile of 4-ABP adducts exhibited an OR of 2.3 (95% CI, 1.23–4.24) relative to those in the lowest 2 quartiles (Table 4).

Both hemoglobin adducts of 2,6-DMA and 4-ABP were independently associated with risk of bladder cancer. A positive association between 2,6-DMA hemoglobin adducts and bladder cancer risk was present in both high and low levels of 4-ABP hemoglobin adducts in all subjects as well as in separate groups stratified smoking status at blood draw or over lifetime. Similarly, the level of 2,6-DMA hemoglobin adducts did not modify the positive association between 4-APB hemoglobin adducts and bladder cancer risk (P_interaction = 0.997).

Earlier, on the basis of data of a study in Los Angeles, California, a high-risk region for bladder cancer, we reported a dose-dependent, statistically significant association between 4-ABP hemoglobin adducts, a validated biomarker for 4-ABP exposure in humans, and risk of bladder cancer unrelated to tobacco use (4). We then examined the relationship between hemoglobin adducts of 9 candidate alkylanilines, a subclass of arylamines of which little is known about their potential carcinogenicity to humans, using the same Los Angeles Bladder Cancer Study data. We observed dose-dependent, statistically significant associations between nonsmoking-related bladder cancer and 3 of the 9 alkylaniline adducts, one of which was 2,6-DMA (12). These latter 3 associations were independent of each other. Now, our 2 sets of novel findings are being confirmed in a low-risk population via a case–control study in Shanghai, China, that was initiated in parallel to our Los Angeles study. The remarkable consistency in results between the Los Angeles and the Shanghai studies strongly implicate that the same class of compounds, namely, the ayrlamines, are responsible for most cases of bladder cancer worldwide.

Cigarette smoking is an important source of 4-ABP. Mainstream cigarette smoke was reported to contain 4-ABP at levels of 2.4 to 4.6 ng per cigarette (unfiltered) and 0.2 to 23 ng per cigarette (filtered; refs. 17, 18). We reported earlier that active smokers of non-Asians in Los Angeles had more than 3-fold 4-ABP hemoglobin adducts than lifelong nonsmokers (75.2 vs. 22.1 pg/g hemoglobin) and the relationship was dose-dependent (4). Consistent with our previous findings, the levels of 4-ABP hemoglobin adducts in the present study population also increased with increasing number of cigarettes per day or urinary levels of total cotinine (both P_trend < 0.001). However, the absolute values of 4-ABP hemoglobin adducts in Chinese in Shanghai were much lower in both current smokers (22.43 pg/g hemoglobin) and lifelong nonsmokers (17.73 pg/g hemoglobin) than their non-Asian counterparts in Los Angeles, California. Although we followed the exact protocol for assays and the same standard for 4-ABP in the 2 laboratories that conducted the quantification of 4-ABP hemoglobin adducts, the comparison of absolute values may not be appropriate between the 2 study populations. The smaller difference between smokers and nonsmokers among Shanghai Chinese compared with that among Los Angeles non-Asians might have the following reasons. Cigarette products consumed by Chinese in Shanghai, China, may have lower 4-ABP than those consumed by Non-Hispanic whites and blacks in Los Angeles, California. Although there is lack of specific data on levels of 4-ABP in cigarettes from both markets, the level of tobacco-specific nitrosamines was >20 times higher in U.S. brand cigarettes than in cigarettes domestically manufactured in China (19). Another reason for the different levels of 4-ABP hemoglobin adducts between Shanghai Chinese and Los Angeles non-Asians might be due to the higher prevalence of rapid N-acetylation in Chinese than in non-Asians. N-Acetylation is a major detoxification pathway of tobacco procarcinogens including 4-ABP (20).

Besides tobacco smoke, there are several nonsmoking sources of 4-ABPs that humans are exposed to. 4-Nitrobiphenyl is a product of incomplete combustion that has been identified as component of kerosene heater emission (21) and diesel engine exhaust (22). Exposure to 4-nitrobiphenyl can result in the production of 4-ABP hemoglobin adducts (23). Therefore, exposure to airborne 4-nitrobiphenyl could result in elevated level of 4-ABP-Hb adducts. Fumes from heated cooking oils contain 4-ABP (24). Chinese women usually prepare food for her family and high-temperature cooking with heated oils is a common food preparation method. Finally, 4-ABP has been detected in commercial hair dyes (25). Women using permanent hair dyes were found at increased risk of bladder cancer in our previous study (26). The elevated level of 4-ABP hemoglobin adducts among lifelong nonsmokers, especially in women, could be the results of exposure to these or as-yet-to-be-identified nonsmoking sources of 4-ABP.

In the present study, patients with bladder cancer showed significantly higher hemoglobin adducts of 2,6-DMA than healthy control subjects in both smokers and nonsmokers. Data on the carcinogenicity of 2,6-DMA on humans are limited. Results from toxicologic studies are consistent with a role for 2,6-DMA in human carcinogenesis (9, 27). 2,6-DMA is shown to induce tumors in nasal cavity and in the liver of rats (8, 28) and has been classified by the IARC as a possible human carcinogen (9). A recent mouse model study suggests that metabolites of 2,6-DMA covalently bind to and produce significant levels of DNA adducts in the bladder, ultimately leading to the malignant transformation of urothelial cells (29). Our previous study showed that the 2,6-DMA-Hb adduct was an independent predictor of bladder cancer risk among non-Hispanic whites in Los Angeles, California (12). The consistent findings between our previous study and this study strongly implicate a role of 2,6-DMA in the development of bladder cancer in humans.

2,6-DMA is a metabolite of some drugs (e.g., lidocaine and xylazine; refs. 30, 31). Human exposure to 2,6-DMA in patients receiving lidocaine for local anesthesia or cardiac arrhythmias has been inferred from the resulting increase in 2,6-DMA hemoglobin adduct levels (30). However, the presence of 2,6-DMA hemoglobin adducts in humans not exposed to lidocaine suggests the existence of other sources of 2,6-DMA exposure in humans (14). 2,6-DMA also is a principal metabolite of the veterinary tranquilizer xylazine (32). Xylazine residues have been found in bovine milk and in different tissues (liver, kidney, etc.) collected from treated cattle (32). In the United States, xylazine is not approved by the U.S. Food and Drug Administration (FDA) for use in food-producing animals but it is unclear whether there is widespread practice of non–FDA-approved use of xylazine in China. Finally, 2,6-DMA is used as a chemical intermediate in the manufacture of pesticides, for example, metalaxyl, a systemic fungicide used to control plant diseases. Metalaxyl is widely used to control fungi on a variety of fruit and vegetable crops. Humans are exposed to metalaxyl through ingestion of fruit and vegetables with the pesticide residue (33). The maximum pesticide residue limits in Canada are set at 1.0 ppm for apples and 5 ppm for lettuce and other leafy vegetables (34, 35). There are similar regulatory policy in place in European countries and the United States. However, data regarding the allowable levels of metalaxyl residue on food crops in China are unavailable. Given a strong association between 2,6-DMA hemoglobin adducts and bladder cancer risk, identification of specific sources of 2,6-DMA exposure in the environment should be a high priority.

Strengths of this study included the population-based study design, relatively large sample size, and comprehensively collected data on exposure and genetic determinants of study subjects. The main limitation of the present study was the retrospective nature of the study design. Assessment of 4-ABP and 2,6-DMA hemoglobin adducts was conducted on blood samples taken after cancer diagnosis and/or treatment in cases. It is possible that the postdiagnostic profile in these adducts among the cases does not accurately reflect the group's prediagnostic profile. For example, a large proportion of patients with bladder cancer reduced the amount of cigarettes smoked per day or quit smoking completely following their cancer diagnosis, thus the 4-ABP hemoglobin adducts measured would be lower than expected if blood samples had been collected from patients before their cancer diagnosis. Therefore, the present study might underestimate the effect of the 4-ABP exposure on bladder cancer risk, especially for smokers. Furthermore, the carcinogenic process is believed to take decades to complete, and thus, the relevant exposure periods are far removed from the time of blood draw. It is unknown whether recent exposures in study subjects, as captured by the hemoglobin adducts of 4-ABP and 2,6-DMA, generally reflect exposure levels in decades past. Future studies with a prospective study design can overcome these limitations and establish the temporal relation between exposure to 2,6-DMA and 4-ABP and the risk of developing bladder cancer.

In summary, the findings of the present study show that hemoglobin adducts of 4-ABP and 2,6-DMA are independent risk predictors of bladder cancer for lifelong nonsmokers. These consistent findings in this Chinese population with our previous findings among non-Asians in Los Angeles strongly implicate arylamines as major causal agents of human bladder cancer. Given that cigarette smoking accounts for only approximately 50% of the bladder cancer burden in the United States, identification of nonsmoking-related sources of 4-ABP and 2,6-DMA in the environment should be a high priority.

No potential conflicts of interest were disclosed.

Conception and design: L. Tao, M.C. Stern, V.K. Cortessis, M.C. Yu, J.-M. Yuan

Development of methodology: L. Tao, B.W. Day, M.C. Yu, J.-M. Yuan

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): B.W. Day, Y.-B. Xiang, D. Van Den Berg, Y.-T. Gao, M.C. Yu, J.-M. Yuan

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L. Tao, B.W. Day, R. Wang, D.V. Conti, Y.-T. Gao, M.C. Yu, J.-M. Yuan

Writing, review, and/or revision of the manuscript: L. Tao, B.W. Day, Y.-B. Xiang, M.C. Stern, M. Gago-Dominguez, V.K. Cortessis, Y.-T. Gao, M.C. Yu, J.-M. Yuan

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L. Tao, Y.-B. Xiang, V.K. Cortessis, D. Van Den Berg, M.C. Pike, Y.-T. Gao, M.C. Yu, J.-M. Yuan

Study supervision: B.W. Day, D. Van Den Berg, Y.-T. Gao, J.-M. Yuan

The research study was supported by R01 CA065726 and R01 CA114665, and partially supported by R01 CA080205-10S1 and R01 CA144034, all from the National Cancer Institute, Bethesda, MD.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.