Abstract
Background: Isothiocyanates, compounds found primarily in cruciferous vegetables, have been shown in laboratory studies to possess anticarcinogenic activity. Glutathione S-transferases (GSTs) are involved in the metabolism and elimination of isothiocyanates; thus, genetic variations in these enzymes may affect in vivo bioavailability and the activity of isothiocyanates.
Objective: The objective was to prospectively evaluate the association between urinary isothiocyanate concentrations and colorectal cancer risk as well as the potential modifying effect of GST genotypes on the association.
Design: A nested case-control study of 322 cases and 1251 controls identified from the Shanghai Women's Health Study was conducted.
Results: Urinary isothiocyanate concentrations were inversely associated with colorectal cancer risk; the inverse association was statistically significant or nearly significant in the GSTM1-null (P for trend = 0.04) and the GSTT1-null (P for trend = 0.07) genotype groups. The strongest inverse association was found among individuals with both the GSTM1-null and the GSTT1-null genotypes, with an adjusted odds ratio of 0.51 (95% CI: 0.27, 0.95), in a comparison of the highest with the lowest tertile of urinary isothiocyanates. No apparent associations between isothiocyanate concentration and colorectal cancer risk were found among individuals who carried either the GSTM1 or GSTT1 gene (P for interaction < 0.05).
Conclusion: This study suggests that isothiocyanate exposure may reduce the risk of colorectal cancer, and this protective effect may be modified by the GSTM1 and GSTT1 genes.
INTRODUCTION
Cruciferous vegetables, including cabbage, broccoli, bok choy, Brussels sprouts, kale, and cauliflower, are unique dietary sources of glucosinolates, which are the parent compounds of isothiocyanates and indoles (1). These glucosinolate break-down products are capable of inhibiting carcinogen-activating enzymes and regulating apoptosis and cell proliferation in cultured tumor cells (2, 3). Administration of crucifers or isothiocyanates to experimental animals has been shown to inhibit the development of colonic aberrant crypt foci (4, 5) and to reduce the incidence and multiplicity of chemical-induced tumors, including tumors of the colon and forestomach (6–8). Isothiocyanates are also potent inducers of phase II enzymes, which are involved in detoxifying potential endogenous and exogenous carcinogens (9, 10). It has been shown that intake of cruciferous vegetables effectively increases the urinary excretion of potential carcinogens such as the heterocyclic amines found in well-done meat (11, 12). This suggests that cruciferous vegetables or their constituent isothiocyanates may confer cancer chemopreventive effects in humans.
Exposure to isothiocyanates in vivo depends not only on dietary intake and absorption of isothiocyanates but on inherent capacity in isothiocyanate metabolism and excretion as well. Glutathione S-transferases (GSTs), the primary enzymes involved in isothiocyanate metabolism, catalyze the conjugation of isothiocyanates with glutathione (13–15), giving rise, in most cases, to less reactive metabolites that are more readily excreted with urine. It has been hypothesized that individuals that are homozygous for deletion of either the GSTM1 or GSTT1 gene may metabolize and eliminate isothiocyanates at a slower rate and therefore may be more intensely exposed to isothiocyanates after consumption of cruciferous vegetables (16). A few epidemiologic studies have recently evaluated this hypothesis and suggest that the anticancer effect of isothiocyanates may differ by GST genotype (17–19).
Urine is the principal disposal route for isothiocyanates and their metabolites. Because there are no endogenous sources of urinary isothiocyanates in humans (20), urinary isothiocyanate concentrations are considered to be an aggregate measure of the level of isothiocyanate intake, absorption, and metabolism and, thus, to reflect the cumulative internalized dose biologically available from multiple sources of dietary exposures. Total urinary isothiocyanates and their metabolites can be quantified with high sensitivity and accuracy by using an HPLC-based method by cyclocondensation reaction (9, 21).
In this report we describe a comprehensive evaluation of the association of colorectal cancer risk with isothiocyanate exposure, as assessed by both dietary crucifer intake and by prediagnostic measurements of urinary isothiocyanates, in a case-control study nested within the Shanghai Women's Health Study—a large cohort study of Chinese women who are known to habitually consume large amounts of crucifers and have a low incidence of colorectal cancer (22). We also evaluated whether GST genotypes interact with isothiocyanates to modify colorectal cancer risk.
SUBJECTS AND METHODS
Cohort of the Shanghai Womenrsquos Health Study
The design and methods of the Shanghai Women's Health Study were described in detail elsewhere (23). Briefly, the cohort includes 74,942 women who were recruited between 1996 and 2000 from 7 urban communities of Shanghai and were 40–70 y of age at study enrollment. The participation rate was 92.7%. All women completed a detailed baseline survey that collected information on demographic characteristics, lifestyle and dietary habits, medical history, family history of cancer, and other exposures. Anthropometric measurements, including weight, height, and circumferences of the waist and hips, were also taken.
Usual dietary intake over the 12 mo before the interview was assessed at baseline for all cohort members and was reassessed 2–3 y after the baseline survey for ≈91% of cohort members using a comprehensive, quantitative, food-frequency questionnaire (FFQ). Five cruciferous vegetables commonly consumed in this population were listed as separate items on the questionnaire, including Chinese greens (bok choy), green cabbage, Chinese cabbage (nappa), cauliflower, and white turnip/radish. Nutrient intakes were calculated by multiplying the amount of each food consumed by the nutrient content of the specific food derived from the Chinese food composition tables (24).
At enrollment, most cohort members donated a urine sample (n = 65,755; 88%) and a blood sample (n = 56,832; 76%) (23). Urine samples were collected into a sterilized cup containing 125 mg ascorbic acid to prevent oxidation of labile metabolites. A 10-mL blood sample was drawn into an EDTA-containing evacuated tube. For those who did not donate a blood sample at baseline, a sample of exfoliated buccal cells (n = 8,934) was collected during the first follow-up survey by using a modified mouthwash method (23). After collection, the samples were kept in a portable styrofoam box with ice packs (at ≈0–4°C) and processed within 6 h for long-term storage at –70°C. A biospecimen collection form was completed for each woman, which included information on the date and time of sample collection, time of last meal, and day of last menstruation (for premenopausal women) as well as intake of selected foods, cigarette smoking, and medication use over the previous 24 h and during the previous week. The study was approved by the relevant Institutional Review Boards for human research in both China and the United States, and written informed consent was obtained from all study participants.
Outcome ascertainment
The cohort was followed for occurrence of cancer and other chronic diseases by a combination of biennial home visits and annual record linkage to the Shanghai Cancer Registry and Shanghai Vital Statistics database. Nearly all cohort members were successfully followed; the response rates for the first (2000–2002), second (2002–2004), and third (2004–2007) in-person follow-up surveys were 99.8%, 98.7%, and 96.7%, respectively. All possible incident cancer cases were verified by home visits. Medical charts from the diagnostic hospital were reviewed to verify the diagnosis.
Nested case-control design
The nested case-control study described in this report included 322 incident colorectal cancer cases who provided a urine sample at baseline and in whom cancer was diagnosed before 31 December 2005. We included only participants who donated a urine sample before any cancer diagnosis. The incidence-density method was used for case-control matching. Controls were selected from women who donated a urine sample at baseline and were free of any cancer at the time of cancer diagnosis for the index case. Cases and controls were individually matched for age at baseline (±2 y), date (≤30 d) and time (morning or afternoon) of urine collection, interval since last meal (≤2 h), menopausal status (before or after), and antibiotic use (yes or no) in the week before sample collection. For most cases (n = 302), we identified 4 controls for each case. For the remaining cases, for whom 4 matched controls could not be found, fewer controls were included. Three cases had a case-to-control ratio of 1:1, 4 cases had a ratio of 1:2, and 13 cases had a ratio of 1:3. A total of 1258 controls were selected.
Urinary isothiocyanate measurement and GST genotyping
Total urinary isothiocyanates and their metabolites were assayed by using an HPLC-based method by the cyclocondensation reaction (9, 21). Urine samples and standards were assayed in triplicate. Three representative standards and a reagent blank were included in all analytic runs. Samples of N-acetyl-l- cysteine conjugates of phenethyl isothiocyanate (0.2–25 mmol/L) in urine from subjects on a controlled diet were analyzed weekly for a standard curve. To control for batch-to-batch variability, samples for each case-control set were analyzed in the same laboratory run. All laboratory assays were performed in 2007–2008. Laboratory staff were blinded to the case-control status of the urine samples and the identity of the quality control samples. The limit of detection for the urinary isothiocyanates was 0.1 μmol/L. This assay showed a high degree of interday and intraday precision, with CVs of 1.4% and 0.5%, respectively (21). Urinary creatinine was measured by using a test kit from Sigma Company (St Louis, MO), and isothiocyanate measurements were reported as nmol/mg creatinine. The average of 3 measurements for each participant was used in the analysis.
After DNA was extracted from blood (86.4%) or exfoliated buccal cells (13.6%), the copy number for the GSTM1 and GSTT1 genes was determined by a duplex real-time quantitative polymerase chain reaction (PCR)–based assay according to the method described in the NCI SNP500 project, with modifications (25). The assay was designed to detect whether an individual has 0, 1, or 2 copies of the GSTM1 and GSTT1 genes. All sequences used in the assay design were obtained from GenBank (GSTM1, NM_000561 and GSTT1, NM_000853). Real-time PCR was performed in a 384-well plate with ABI PRISM 7900 Sequence Detection Systems (Applied Biosystems, Foster City, CA). Laboratory staff were blinded to the case-control status of the samples. Coriell DNA samples containing 0, 1, or 2 copies of the GSTM1 and GSTT1 genes were included to serve as internal quality controls. The concordance rate for quality control samples, including water, Coriell DNA, and blinded DNA samples was 100%. There were no differences in genotyping success rate and genotype distribution of these 2 genes between DNA extracted from blood and buccal cells.
Statistical analyses
We excluded women with a prior history of familial adenomatous polyposis (n = 3) and missing data on urinary isothiocyanate concentrations (n = 4). The final analytic data set included 322 colorectal cancer cases and 1251 individually matched controls.
For samples with undetectable isothiocyanate values (n = 140; 8.9%), we estimated values by dividing the lowest detectable value of the assay by the square root of 2 (26). To better estimate usual dietary intake, we used the average intake of the first FFQ at baseline and the second FFQ conducted 2–3 y after the baseline survey. Pairwise comparisons for urinary isothiocyanate concentrations and other continuous variables were conducted by using Wilcoxon's signed-rank test. The Wald test from a conditional logistic regression model was used to compare the frequency of categorical variables between cases and controls. Wilcoxon's signed-rank test or the Kruskal-Wallis test was also used to examine the difference in urinary isothiocyanate concentrations across GST genotype categories.
Urinary isothiocyanate concentrations and dietary crucifer intakes were grouped based on tertile distributions in the controls; the lowest tertile served as the reference. Conditional logistic regression modeling was used to estimate the odds ratios (ORs) of developing colorectal cancer and their 95% CIs associated with urinary isothiocyanate concentration or dietary crucifer intake and to adjust for potential confounders. Potential confounders adjusted for in multivariable models included age at enrollment (continuous), education (4 categories), household income (4 categories), body mass index (calculated as weight in kilograms divided by the square of height in meters, continuous), physical activity level [measured by metabolic equivalent (MET)-hours per week per year, continuous], colorectal cancer in first-degree relatives (yes or no), and intakes (continuous) of total energy, calcium, fruit, noncruciferous vegetables, and red meat. Tests for trend were performed by entering categorical variables as continuous variables in the model.
Potential effect modification by GST genotypes was evaluated in stratified analyses by breaking the matching and using unconditional logistic regression models. In addition to all of the covariates listed above, all matching variables were included in the model. Multiplicative diet-gene interactions were determined based on the likelihood ratio test comparing models that included only the main effect terms with models that included both the main effect and the interaction terms. Statistical analyses were carried out by using SAS version 9.1 (SAS Institute, Cary, NC). All statistical tests were based on 2-sided probability.
RESULTS
The distribution of baseline characteristics in this study population is presented in Table 1. Cases and controls were well matched for age and menopausal status at enrollment. Cases and controls did not differ significantly with regard to socioeconomic status, most lifestyle characteristics, or potential risk factors for colorectal cancer. However, compared with controls, cases appeared to have a lower household income, consumed slightly less red meat, and were less likely to be a cigarette smoker.
TABLE 1.
Baseline characteristics of colorectal cancer cases and controls: Shanghai Women's Health Study (1997–2005)1
| Cases (n = 322) | Controls (n = 1251) | P | |
| Education, college and above (%) | 9.3 | 12.1 | 0.142 |
| Household income, >30,000 Yuan/y (%) | 10.6 | 15.7 | 0.012 |
| Cigarette smoking (%) | 1.9 | 4.4 | 0.042 |
| Alcohol consumption (%) | 2.2 | 2.4 | 0.722 |
| Physical activity, >100 MET-h/wk per year (%) | 58.1 | 55.2 | 0.342 |
| Family history of colorectal cancer (%) | 2.8 | 2.3 | 0.572 |
| Postmenopausal at enrollment (%) | 78.9 | 79.0 | 0.972 |
| GSTM1 (%) | |||
| Null | 58.7 | 58.4 | 0.612 |
| 1 copy | 36.3 | 34.8 | — |
| 2 copies | 5.0 | 6.9 | — |
| GSTT1 (%) | |||
| Null | 50.9 | 48.9 | 0.722 |
| 1 copy | 38.8 | 41.4 | — |
| 2 copies | 10.3 | 9.7 | — |
| Age at enrollment (y) | 62 (54, 66)3 | 62 (54, 66) | 0.214 |
| BMI (kg/m2) | 24.6 (22.4, 26.6) | 24.5 (22.2, 26.9) | 0.794 |
| Total energy intake (kcal/d) | 1646 (1392, 1889) | 1618 (1372, 1893) | 0.674 |
| Cruciferous vegetable intake (g/d) | 84.9 (48.8, 137.5) | 84.0 (49.5, 135.7) | 0.344 |
| Intake of all vegetables (g/d) | 261.9 (175.4, 371.2) | 250.7 (170.6, 361.5) | 0.624 |
| Red meat intake (g/d) | 38.3 (21.9, 62.2) | 39.9 (24.0, 61.1) | 0.054 |
| Urinary ITCs (nmol/mg creatinine)5 | 1.57 (0.58, 3.70) | 1.76 (0.72, 4.04) | <0.00014 |
MET-h, metabolic equivalent hours; Yuan, Chinese currency (1 US dollar = ≈8 Yuan at recruitment); ITCs, isothiocyanates.
Derived from Wald test from conditional logistic regression.
Median; interquartile range (between the 25th and the 75th percentiles) in parentheses (all such values).
Derived from Wilcoxon's signed-rank test for case-control differences within each case-control set.
The median (interquartile range) difference between cases and controls (urinary ITCs of cases – mean ITCs of controls in the case-control set) was −0.81 (−2.60, 0.96).
The GSTM1 and GSTT1 genotypes were within Hardy-Weinberg equilibrium for both cases and controls. There were no significant case-control differences in the copy number of the GSTM1 and GSTT1 genes (Table 1). The proportion of individuals with the null genotype for both genes was comparable with that of other Asian populations (27). Neither the GSTM1 nor GSTT1 genotypes were associated with colorectal cancer risk (data not shown).
Urinary isothiocyanate concentrations by GST genotype are shown in Table 2. Because only a small number of women carried both copies of the GSTM1 or GSTT1 gene and because there was no significant allelic dosage effect (1 compared with 2 copies) of the GSTM1 or GSTT1 gene on urinary isothiocyanate concentrations (data not shown), women with 1 and 2 copies of these genes were combined into a single group in this analysis. Compared with women homozygous for GSTM1 deletion, women with at least one copy of the GSTM1 gene had a higher level of urinary isothiocyanate excretion (P = 0.04). Women who carried both the GSTM1 and GSTT1 genes had the highest level of urinary isothiocyanate excretion, whereas women with the double null genotype had the lowest level of urinary isothiocyanate excretion (P = 0.02). This association persisted in a model mutually adjusted for both GST genotypes and dietary intake (P = 0.01) and was similar when cases and controls were analyzed separately. No significant differences were observed in dietary intake of cruciferous vegetables according to GSTM1 or combined GSTM1 and GSTT1 genotypes.
TABLE 2.
Baseline urinary isothiocyanate (ITC) excretion and crucifer intake by glutathione S-transferase genotype: Shanghai Women's Health Study (1997–2005)1
| Urinary ITCs |
Crucifer intake |
|||
| Median (IQR) | P2 | Median (IQR) | P2 | |
| nmol/mg creatinine | g/d | |||
| GSTM1 | ||||
| Null (n = 845) | 1.84 (0.85, 4.18) | 0.04 | 83.7 (47.2, 135.3) | 0.25 |
| Present (n = 588) | 2.16 (1.01, 4.26) | — | 84.9 (53.5, 137.5) | — |
| GSTT1 | ||||
| Null (n = 709) | 1.82 (0.87, 4.03) | 0.08 | 87.6 (54.2, 138.8) | 0.03 |
| Present (n = 724) | 2.09 (0.92, 4.65) | — | 80.4 (45.3, 133.7) | — |
| GSTM1 and GSTT1 | ||||
| Double null (n = 387) | 1.60 (0.88, 3.84) | 0.02 | 86.2 (50.5, 135.9) | 0.61 |
| Either one null (n = 780) | 1.99 (0.84, 4.38) | — | 84.9 (49.0, 137.4) | — |
| Double present (n = 266) | 2.25 (1.13, 4.68) | — | 77.0 (48.5, 133.0) | — |
Subjects with undetectable urinary ITC concentrations (n = 140) or missing data on GSTM1 (n = 2) or GSTT1 (n = 4) were excluded from the analyses. IQR, interquartile range (between the 25th percentile and the 75th percentile).
Derived from Wilcoxon's signed-rank test or the Kruskal-Wallis test.
Baseline urinary isothiocyanate concentrations in colorectal cancer cases were 50.1% (median) lower than in controls in a pairwise comparison, with a median case-control difference of −0.81 nmol/mg creatinine (Table 1). When analyzed as categorical variables in conditional logistic regression models, urinary isothiocyanate concentrations were inversely associated with overall colorectal cancer risk, but the results were not statistically significant (Table 3). The inverse association, however, was statistically significant among women with the GSTM1-null genotype (P for trend = 0.04) and was nearly significant among women with the GSTT1-null genotype (P for trend = 0.07). The strongest inverse association was found in women null for both the GSTM1 and GSTT1 genes (OR for the comparison of extreme tertiles: 0.51; 95% CI: 0.27, 0.95; P for trend = 0.03). No associations were found for women who carried either the GSTM1 or GSTT1 gene.
TABLE 3.
Association of urinary isothiocyanate (ITC) concentrations with colorectal cancer risk by glutathione S-transferase genotype: Shanghai Women's Health Study (1997–2005)
| Tertile of urinary ITCs (nmol/mg creatinine) |
|||||
| 1 (<0.95) | 2 (0.95–2.98) | 3 (>2.98) | P for trend | P for interaction | |
| Overall | |||||
| No. of cases/controls | 117/413 | 105/414 | 100/427 | — | — |
| Crude OR (95% CI) | 1.00 | 0.88 (0.65, 1.20) | 0.81 (0.58, 1.12) | 0.20 | — |
| Multivariable OR (95% CI)1 | 1.00 | 0.87 (0.63, 1.18) | 0.77 (0.55, 1.07) | 0.12 | — |
| GSTM1 null | |||||
| No. of cases/controls | 77/242 | 58/242 | 54/245 | — | — |
| Multivariable OR (95% CI)2 | 1.00 | 0.75 (0.50, 1.11) | 0.66 (0.44, 1.00) | 0.04 | 0.10 |
| GSTM1 present | |||||
| No. of cases/controls | 40/168 | 47/170 | 46/180 | — | — |
| Multivariable OR (95% CI)2 | 1.00 | 1.22 (0.75, 1.99) | 1.12 (0.68, 1.83) | 0.68 | — |
| GSTT1 null | |||||
| No. of cases/controls | 66/199 | 50/215 | 48/196 | — | — |
| Multivariable OR (95% CI)2 | 1.00 | 0.66 (0.42, 0.99) | 0.67 (0.44, 1.05) | 0.07 | 0.33 |
| GSTT1 present | |||||
| No. of cases/controls | 51/211 | 55/197 | 52/229 | — | — |
| Multivariable OR (95% CI)2 | 1.00 | 1.13 (0.73, 1.75) | 0.92 (0.59, 1.44) | 0.71 | — |
| GSTM1 and GSTT1 null | |||||
| No. of cases/controls | 40/105 | 32/119 | 24/102 | — | — |
| Multivariable OR (95% CI)2 | 1.00 | 0.61 (0.35, 1.08) | 0.51 (0.27, 0.95) | 0.03 | <0.05 |
| GSTM1 or GSTT1 null | |||||
| No. of cases/controls | 63/231 | 44/219 | 54/237 | — | — |
| Multivariable OR (95% CI)2 | 1.00 | 0.75 (0.48, 1.16) | 0.84 (0.55, 1.28) | 0.40 | — |
| GSTM1 and GSTT1 present | |||||
| No. of cases/controls | 14/74 | 29/74 | 22/86 | — | — |
| Multivariable OR (95% CI)2 | 1.00 | 2.16 (0.98, 4.78) | 1.25 (0.56, 2.78) | 0.73 | — |
Odds ratios (ORs) were estimated by using multivariable conditional logistic regression models, adjusted for age, education, household income, physical activity, cigarette smoking, alcohol consumption, BMI, family history of colorectal cancer, and intakes of total energy, fruit, noncruciferous vegetables, red meat, and calcium.
ORs were estimated by using multivariable unconditional logistic regression model, adjusted for all covariates included in the model above and all matching variables including menopausal status at sample collection, time of sample collection, time interval (h) between last meal and sample collection, antibiotic use in the past week, and time interval (y) between sample collection and ITC measurement.
The associations between urinary isothiocyanates and colorectal cancer did not vary by lifestyle characteristics, including body mass index, physical activity, or consumption of red meat, fruit, or noncruciferous vegetables (data not shown); nor did the associations vary appreciably by anatomic subsite (colon compared with rectum) or clinical stage of tumors. Very few women in this cohort ever smoked cigarettes (2.7%) (23), and the results were similar between nonsmokers and all subjects included in the study (data not shown). Furthermore, results from analyses that excluded cases diagnosed in the first year of follow-up were similar to the results presented in Table 3. For example, among individuals null for both the GSTM1 and GSTT1 genes, the multivariable OR for the highest compared with the lowest tertile of urinary isothiocyanates was 0.52 (95% CI: 0.28, 0.995; P for trend = 0.04).
No apparent associations were found between dietary crucifer intake and colorectal cancer risk, either overall or by GST genotype (Table 4). Analyses of the correlation between urinary isothiocyanate excretion and dietary intake of crucifers showed that the mean concentrations of urinary isothiocyanates increased from 1.48 to 1.80 and 1.90 nmol/mg creatinine with increasing tertiles of dietary crucifer intake (P for trend = 0.003). In a mutually adjusted model, both GST genotypes and dietary intake were significantly associated with urinary isothiocyanate concentrations. The correlation at the individual level, however, was very weak, either overall (Spearman correlation coefficient = 0.08, P = 0.001) or by GST genotype (Spearman correlation coefficients ranged from 0.05 to 0.11).
TABLE 4.
Association of cruciferous vegetable intake with colorectal cancer risk by glutathione S-transferase genotype: Shanghai Women's Health Study (1997–2005)
| Tertile of cruciferous intake |
|||||
| 1 | 2 | 3 | P for trend | P for interaction | |
| Overall | |||||
| No. of cases/controls | 113/414 | 104/412 | 105/425 | — | — |
| Multivariable OR (95% CI)1 | 1.00 | 0.94 (0.69, 1.28) | 0.93 (0.66, 1.31) | 0.66 | — |
| Multivariable OR (95% CI) by 2 | |||||
| Null | 1.00 | 0.95 (0.63, 1.44) | 0.93 (0.59, 1.49) | 0.77 | 0.96 |
| Present | 1.00 | 0.96 (0.58, 1.58) | 1.00 (0.58, 1.72) | 0.98 | — |
| Multivariable OR (95% CI) by 2 | |||||
| Null | 1.00 | 0.97 (0.62, 1.51) | 0.92 (0.56, 1.52) | 0.75 | 0.72 |
| Present | 1.00 | 0.99 (0.63, 1.56) | 0.96 (0.58, 1.59) | 0.88 | — |
| Multivariable OR (95% CI) by GSTM1 and 2 | |||||
| Double null | 1.00 | 0.87 (0.48, 1.57) | 0.70 (0.35, 1.39) | 0.31 | 0.82 |
| Either one null | 1.00 | 0.99 (0.63, 1.56) | 1.15 (0.70, 1.88) | 0.61 | — |
| Double present | 1.00 | 1.14 (0.55, 2.36) | 0.86 (0.38, 1.96) | 0.76 | — |
Odds ratios (ORs) were estimated by using multivariable conditional logistic regression models, adjusted for age, education, household income, physical activity, cigarette smoking, alcohol consumption, BMI, family history of colorectal cancer, and intakes of total energy, fruit, noncruciferous vegetables, red meat, and calcium.
ORs were estimated by using multivariable unconditional logistic regression model, adjusted for all covariates included in the model above and all matching variables including menopausal status at sample collection, time of sample collection, time interval (h) between last meal and sample collection, antibiotic use in the past week, and time interval (y) between sample collection and isothiocyanate measurement.
DISCUSSION
In this nested case-control study, the largest urine-based biomarker study to date of isothiocyanate exposure and colorectal cancer, we found an inverse association between urinary isothiocyanates and colorectal cancer risk. The protective effect of isothiocyanates was seen among individuals with homozygous deletion of GSTM1 and particularly with deletion of both GSTM1 and GSTT1. However, no apparent association between isothiocyanate concentration and colorectal cancer risk was found among individuals who carried either the GSTM1 or GSTT1 gene. We also found that urinary isothiocyanate concentrations were lower in women who had a homozygous deletion of either the GSTM1 or GSTT1 gene than in women who carried these GST genes and dietary intake was likely independent of GST genotypes, suggesting that the isothiocyanate metabolic clearance rate differs significantly by GST genotypes. These findings support the notion that individuals with GST deletion may metabolize isothiocyanates less efficiently and, therefore, may have a higher exposure to isothiocyanates, which allows them to benefit more from consumption of cruciferous vegetables in terms of reduction of cancer risk.
Our findings on the interactive effect of isothiocyanate exposure and GST genotype on colorectal cancer risk are supported by several previous epidemiologic studies of other cancer types (17, 28). For example, London et al (17) reported in a cohort study of men in Shanghai that high urinary isothiocyanate excretion levels were associated with a reduced risk of lung cancer, predominantly among men with the GSTM1-null and/or GSTT1-null genotypes. In general, studies of dietary isothiocyanate/crucifer intake and colorectal cancer, mostly with a case-control study design (19, 29–31), have reported an inverse relation between isothiocyanate exposure and risk of colorectal adenoma (19) or cancer (18, 29, 30) among individuals with null/low-activity GST genotypes, although the results are not entirely consistent (31). To date, only 2 nested case-control studies have prospectively evaluated the association between urinary isothiocyanate concentrations and colorectal cancer. Moy et al (32) reported recently in the Shanghai Men's Cohort that urinary isothiocyanate concentrations were inversely associated with the risk of colorectal cancer; however, this finding was observed only among men whose urine samples were collected ≥5 y before disease diagnosis, and the interaction with related genes was not evaluated. A similar protective effect of isothiocyanates was also reported in the Multiethnic Cohort Study (33), with a reduction in colorectal cancer risk of ≈40% among individuals with detectable concentrations of urinary isothiocyanates. The effect was more pronounced among individuals with the G allele (encoding a lower-activity allozyme) for the GSTP1 Ile105Val polymorphism (P for interaction = 0.09). As acknowledged by the authors (33), that study, however, was limited by insufficient statistical power, particularly in stratified analyses and a short follow-up time (the median interval between the time of collection of urine samples and disease diagnosis was 1.4 y). The association between dietary intake of cruciferous vegetables and colorectal cancer risk was also evaluated in the Multiethnic Cohort Study, with no evidence of a protective effect of cruciferous vegetable intake as assessed by FFQ (34), similar to the results reported in the present study.
Several case-control studies, including our previous studies (35), have evaluated the association of colorectal cancer with dietary crucifer intake (18, 19, 29–31, 36). Results have been mixed but generally point to an inverse association (18, 19, 29, 30, 36). However, no such preventive effect was found in any of the recent cohort studies conducted in the United States, including the Iowa Women's Health Study (37), the Nurses' Health Study (38), the Health Professionals Follow-Up Study (38), the Multiethnic Cohort Study (34), and the Cancer Prevention Study II Nutrition Cohort (39) or in cohort studies conducted in Japan (40) and Finland (41). One exception was a case-control study nested within the Singapore Chinese Health Study (18). Higher intakes of cruciferous vegetables were associated with a reduced risk of colorectal cancer among individuals null for both the GSTM1 and GSTT1 genes. In a recent pooled analysis of 14 cohort studies (42), the pooled multivariable relative risk of the highest compared with the lowest tertile of crucifer intake was 0.99 (95% CI: 0.93, 1.06), similar to the findings of our study.
These studies may have been limited by possible misclassification errors in dietary assessment. The level of glucosinolates varies considerably in vegetables depending on their species, growing conditions, and maturity at the time of harvest (1). The formation of isothiocyanates is catalyzed by myrosinase, a thermosensitive enzyme in plant cells and in gastrointestinal microflora. Boiling cruciferous vegetables results in a 30–60% loss of intact glucosinolates due to thermal degradation and leaching (20). The bioavailability of isothiocyanates from fully cooked broccoli is 3-fold lower than in lightly cooked broccoli as a result of the inactivation of myrosinase by heat (43). Therefore, exposure levels of isothiocyanates can be further affected by many other factors, such as food-preparation methods and the status of colonic bacteria (44–46). Crucifer intake assessed by a food questionnaire is not an ideal measure of dietary exposure to isothiocyanates and related phytochemicals. In the present study, the mean concentration of urinary isothiocyanates increased with increasing intake of cruciferous vegetables, similar to a previous observation in Singapore Chinese (15); however, the correlation at the individual level was very weak. Such a weak correlation between urinary isothiocyanate concentrations and crucifer intakes was also reported previously (33, 47), which suggested that dietary crucifer intake assessed by FFQ may not adequately reflect the exposure level of isothiocyanates in vivo.
Although urinary isothiocyanate concentrations reflect crucifer intakes over the prior 24–48-h period (48), the current study was conducted in a population that is among the most frequent consumers of cruciferous vegetables in the world. This population consumes 84 g cruciferous vegetables/d on average, which is ≈6-fold higher than that in Western populations (≈0.2 serving/d) (39). We showed previously that the within-person variation in urinary isothiocyanate measurements is relatively small in participants of the Shanghai Women's Health Study. The intraclass correlation coefficient was 0.50 for isothiocyanates measured in 4 urine samples collected over a 12-mo period, indicating that a single measurement in a spot urine sample reflects the level of isothiocyanate exposure over a longer period reasonably well in our study population. In addition, we matched cases and controls on the date of specimen collection to minimize potential seasonal and storage effects on concentrations of urinary isothiocyanates. Moreover, because recent antibiotic use can alter the status of gastrointestinal microflora that converts glucosinolates to isothiocyanates in the gut, we further matched cases and controls on antibiotic use. We also performed analyses of urinary isothiocyanates after excluding antibiotic users. The results did not differ appreciably. Other strengths of the study included the use of a prediagnostic urine sample, high participation rates at baseline, and a virtually complete follow-up of the cohort, which minimized potential biases inherent in case-control studies.
In summary, using dietary intake data and urine specimens from a large, population-based, prospective cohort study, we found a strong inverse association between urinary isothiocyanate concentrations and subsequent risk of developing colorectal cancer among women with the null genotype of the GSTM1 and GSTT1 genes. Given the long latency of cancer, continued follow-up is needed to provide the data necessary to further characterize the effect of isothiocyanates or crucifer intakes on colorectal cancer incidence.
Acknowledgments
We are grateful to the participants and research staff of the Shanghai Women's Health Study for their contributions to the study. We also thank Wanqing Wen for statistical consultation in data analysis and Bethanie Hull and Rod Jones for assistance in preparing the manuscript.
The authors' responsibilities were as follows—GY, X-OS, and WZ: study design, data collection, statistical analyses, and writing of manuscript; Y-TG, H-LL, and Y-BX: data collection and management; B-TJ, NR, and W-HC: data collection and revision of the manuscript; and QC, G-LL, MD, and F-LC: laboratory assays and revision of the manuscript. None of the authors had any financial conflicts of interest to declare.
REFERENCES
- 1.Fenwick GR, Heaney RK, Mullin WJ. Glucosinolates and their breakdown products in food and food plants. Crit Rev Food Sci Nutr 1983;18:123–201 [DOI] [PubMed] [Google Scholar]
- 2.Bonnesen C, Eggleston IM, Hayes JD. Dietary indoles and isothiocyanates that are generated from cruciferous vegetables can both stimulate apoptosis and confer protection against DNA damage in human colon cell lines. Cancer Res 2001;61:6120–30 [PubMed] [Google Scholar]
- 3.Zhang Y, Yao S, Li J. Vegetable-derived isothiocyanates: anti-proliferative activity and mechanism of action. Proc Nutr Soc 2006;65:68–75 [DOI] [PubMed] [Google Scholar]
- 4.Smith TK, Mithen R, Johnson IT. Effects of Brassica vegetable juice on the induction of apoptosis and aberrant crypt foci in rat colonic mucosal crypts in vivo. Carcinogenesis 2003;24:491–5 [DOI] [PubMed] [Google Scholar]
- 5.Chung FL, Conaway CC, Rao CV, Reddy BS. Chemoprevention of colonic aberrant crypt foci in Fischer rats by sulforaphane and phenethyl isothiocyanate. Carcinogenesis 2000;21:2287–91 [DOI] [PubMed] [Google Scholar]
- 6.Barrett JE, Klopfenstein CF, Leipold HW. Protective effects of cruciferous seed meals and hulls against colon cancer in mice. Cancer Lett 1998;127:83–8 [DOI] [PubMed] [Google Scholar]
- 7.Arikawa AY, Gallaher DD. Cruciferous vegetables reduce morphological markers of colon cancer risk in dimethylhydrazine-treated rats. J Nutr 2008;138:526–32 [DOI] [PubMed] [Google Scholar]
- 8.Shen G, Khor TO, Hu R, et al. Chemoprevention of familial adenomatous polyposis by natural dietary compounds sulforaphane and dibenzoylmethane alone and in combination in ApcMin/+ mouse. Cancer Res 2007;67:9937–44 [DOI] [PubMed] [Google Scholar]
- 9.Zhang Y, Talalay P. Anticarcinogenic activities of organic isothiocyanates: chemistry and mechanisms. Cancer Res 1994;54:1976s–81s [PubMed] [Google Scholar]
- 10.Fahey JW, Zhang Y, Talalay P. Broccoli sprouts: an exceptionally rich source of inducers of enzymes that protect against chemical carcinogens. Proc Natl Acad Sci USA 1997;94:10367–72 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Murray S, Lake BG, Gray S, et al. Effect of cruciferous vegetable consumption on heterocyclic aromatic amine metabolism in man. Carcinogenesis 2001;22:1413–20 [DOI] [PubMed] [Google Scholar]
- 12.Walters DG, Young PJ, Agus C, et al. Cruciferous vegetable consumption alters the metabolism of the dietary carcinogen 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) in humans. Carcinogenesis 2004;25:1659–69 [DOI] [PubMed] [Google Scholar]
- 13.Zhang Y, Kolm RH, Mannervik B, Talalay P. Reversible conjugation of isothiocyanates with glutathione catalyzed by human glutathione transferases. Biochem Biophys Res Commun 1995;206:748–55 [DOI] [PubMed] [Google Scholar]
- 14.Kolm RH, Danielson UH, Zhang Y, Talalay P, Mannervik B. Isothiocyanates as substrates for human glutathione transferases: structure-activity studies. Biochem J 1995;311:453–9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Seow A, Shi CY, Chung FL, et al. Urinary total isothiocyanate (ITC) in a population-based sample of middle-aged and older Chinese in Singapore: relationship with dietary total ITC and glutathione S-transferase M1/T1/P1 genotypes. Cancer Epidemiol Biomarkers Prev 1998;7:775–81 [PubMed] [Google Scholar]
- 16.Lam TK, Gallicchio L, Lindsley K, et al. Cruciferous vegetable consumption and lung cancer risk: a systematic review. Cancer Epidemiol Biomarkers Prev 2009;18:184–95 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.London SJ, Yuan JM, Chung FL, et al. Isothiocyanates, glutathione S-transferase M1 and T1 polymorphisms, and lung-cancer risk: a prospective study of men in Shanghai, China. Lancet 2000;356:724–9 [DOI] [PubMed] [Google Scholar]
- 18.Seow A, Yuan JM, Sun CL, Van Den BD, Lee HP, Yu MC. Dietary isothiocyanates, glutathione S-transferase polymorphisms and colorectal cancer risk in the Singapore Chinese Health Study. Carcinogenesis 2002;23:2055–61 [DOI] [PubMed] [Google Scholar]
- 19.Lin HJ, Probst-Hensch NM, Louie AD, et al. Glutathione transferase null genotype, broccoli, and lower prevalence of colorectal adenomas. Cancer Epidemiol Biomarkers Prev 1998;7:647–52 [PubMed] [Google Scholar]
- 20.Zhang Y. Cancer-preventive isothiocyanates: measurement of human exposure and mechanism of action. Mutat Res 2004;555:173–90 [DOI] [PubMed] [Google Scholar]
- 21.Chung FL, Jiao D, Getahun SM, Yu MC. A urinary biomarker for uptake of dietary isothiocyanates in humans. Cancer Epidemiol Biomarkers Prev 1998;7:103–8 [PubMed] [Google Scholar]
- 22.Jin F, Devesa SS, Chow WH, et al. Cancer incidence trends in urban Shanghai, 1972-1994: an update. Int J Cancer 1999;83:435–40 [DOI] [PubMed] [Google Scholar]
- 23.Zheng W, Chow WH, Yang G, et al. The Shanghai Women's Health Study: rationale, study design, and baseline characteristics. Am J Epidemiol 2005;162:1123–31 [DOI] [PubMed] [Google Scholar]
- 24.Yang YX, Wang GY, Pan XC, . China food composition tables 2002. Beijing, China: Beijing University Medical Press, 2002 [Google Scholar]
- 25.Moore LE, Huang WY, Chatterjee N, et al. GSTM1, GSTT1, and GSTP1 polymorphisms and risk of advanced colorectal adenoma. Cancer Epidemiol Biomarkers Prev 2005;14:1823–7 [DOI] [PubMed] [Google Scholar]
- 26.Hornung EWRL. Estimation of average concentration in the presence of nondetectable values. Appl Occup Environ Hyg 1990;5:46–51 [Google Scholar]
- 27.Cotton SC, Sharp L, Little J, Brockton N. Glutathione S-transferase polymorphisms and colorectal cancer: a HuGE review. Am J Epidemiol 2000;151:7–32 [DOI] [PubMed] [Google Scholar]
- 28.Moore LE, Brennan P, Karami S, et al. Glutathione S-transferase polymorphisms, cruciferous vegetable intake and cancer risk in the Central and Eastern European Kidney Cancer Study. Carcinogenesis 2007;28:1960–4 [DOI] [PubMed] [Google Scholar]
- 29.Slattery ML, Kampman E, Samowitz W, Caan BJ, Potter JD. Interplay between dietary inducers of GST and the GSTM-1 genotype in colon cancer. Int J Cancer 2000;87:728–33 [PubMed] [Google Scholar]
- 30.Turner F, Smith G, Sachse C, et al. Vegetable, fruit and meat consumption and potential risk modifying genes in relation to colorectal cancer. Int J Cancer 2004;112:259–64 [DOI] [PubMed] [Google Scholar]
- 31.Tijhuis MJ, Wark PA, Aarts JM, et al. GSTP1 and GSTA1 polymorphisms interact with cruciferous vegetable intake in colorectal adenoma risk. Cancer Epidemiol Biomarkers Prev 2005;14:2943–51 [DOI] [PubMed] [Google Scholar]
- 32.Moy KA, Yuan JM, Chung FL, et al. Urinary total isothiocyanates and colorectal cancer: a prospective study of men in Shanghai, China. Cancer Epidemiol Biomarkers Prev 2008;17:1354–9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Epplein M, Wilkens LR, Tiirikainen M, et al. Urinary isothiocyanates; glutathione S-transferase M1, T1, and P1 polymorphisms; and risk of colorectal cancer: the Multiethnic Cohort Study. Cancer Epidemiol Biomarkers Prev 2009;18:314–20 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Nomura AM, Wilkens LR, Murphy SP, et al. Association of vegetable, fruit, and grain intakes with colorectal cancer: the Multiethnic Cohort Study. Am J Clin Nutr 2008;88:730–7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Chiu BC, Ji BT, Dai Q, et al. Dietary factors and risk of colon cancer in Shanghai, China. Cancer Epidemiol Biomarkers Prev 2003;12:201–8 [PubMed] [Google Scholar]
- 36.Lynn A, Collins A, Fuller Z, Hillman K, Ratcliffe B. Cruciferous vegetables and colo-rectal cancer. Proc Nutr Soc 2006;65:135–44 [DOI] [PubMed] [Google Scholar]
- 37.Steinmetz KA, Kushi LH, Bostick RM, Folsom AR, Potter JD. Vegetables, fruit, and colon cancer in the Iowa Women's Health Study. Am J Epidemiol 1994;139:1–15 [DOI] [PubMed] [Google Scholar]
- 38.Michels KB, Edward G, Joshipura KJ, et al. Prospective study of fruit and vegetable consumption and incidence of colon and rectal cancers. J Natl Cancer Inst 2000;92:1740–52 [DOI] [PubMed] [Google Scholar]
- 39.McCullough ML, Robertson AS, Chao A, et al. A prospective study of whole grains, fruits, vegetables and colon cancer risk. Cancer Causes Control 2003;14:959–70 [DOI] [PubMed] [Google Scholar]
- 40.Kojima M, Wakai K, Tamakoshi K, et al. Diet and colorectal cancer mortality: results from the Japan Collaborative Cohort Study. Nutr Cancer 2004;50:23–32 [DOI] [PubMed] [Google Scholar]
- 41.Pietinen P, Malila N, Virtanen M, et al. Diet and risk of colorectal cancer in a cohort of Finnish men. Cancer Causes Control 1999;10:387–96 [DOI] [PubMed] [Google Scholar]
- 42.Koushik A, Hunter DJ, Spiegelman D, et al. Fruits, vegetables, and colon cancer risk in a pooled analysis of 14 cohort studies. J Natl Cancer Inst 2007;99:1471–83 [DOI] [PubMed] [Google Scholar]
- 43.Rungapamestry V, Duncan AJ, Fuller Z, Ratcliffe B. Effect of meal composition and cooking duration on the fate of sulforaphane following consumption of broccoli by healthy human subjects. Br J Nutr 2007;97:644–52 [DOI] [PubMed] [Google Scholar]
- 44.Getahun SM, Chung FL. Conversion of glucosinolates to isothiocyanates in humans after ingestion of cooked watercress. Cancer Epidemiol Biomarkers Prev 1999;8:447–51 [PubMed] [Google Scholar]
- 45.Rouzaud G, Young SA, Duncan AJ. Hydrolysis of glucosinolates to isothiocyanates after ingestion of raw or microwaved cabbage by human volunteers. Cancer Epidemiol Biomarkers Prev 2004;13:125–31 [DOI] [PubMed] [Google Scholar]
- 46.Rungapamestry V, Rabot S, Fuller Z, Ratcliffe B, Duncan AJ. Influence of cooking duration of cabbage and presence of colonic microbiota on the excretion of N-acetylcysteine conjugates of allyl isothiocyanate and bioactivity of phase 2 enzymes in F344 rats. Br J Nutr 2008;99:773–81 [DOI] [PubMed] [Google Scholar]
- 47.Fowke JH, Fahey JW, Stephenson KK, Hebert JR. Using isothiocyanate excretion as a biological marker of Brassica vegetable consumption in epidemiological studies: evaluating the sources of variability. Public Health Nutr 2001;4:837–46 [DOI] [PubMed] [Google Scholar]
- 48.Shapiro TA, Fahey JW, Wade KL, Stephenson KK, Talalay P. Chemoprotective glucosinolates and isothiocyanates of broccoli sprouts: metabolism and excretion in humans. Cancer Epidemiol Biomarkers Prev 2001;10:501–8 [PubMed] [Google Scholar]