点击显示 收起
1 From the Phytochemicals and Health Programme, Institute of Food Research, Norwich Research Park, Colney, Norwich, United Kingdom (AVG, JRB, and RFM) and the Center for Analytical Bioscience, School Pharmacy (DAB, AA, and AVG) and Biomedical Sciences (MAT and AVG) and the Wolfson Digestive Diseases Centre (JAS, CA, PF, and CJH), University of Nottingham, Nottingham, United Kingodm
2 Supported by the Biotechnology and Biological Sciences Research Council, the University of Nottingham, Nottingham, United Kingdom, and Seminis Seeds Inc. 3 Reprints not available. Address correspondence to RF Mithen, Phytochemicals and Health, Institute of Food Research, Norwich Research Park, Colney, Norwich, United Kingdom. E-mail: richard.mithen{at}bbsrc.ac.uk.
ABSTRACT
Background: Broccoli consumption is associated with a reduction in the risk of cancer, particularly in persons with a functional glutathione S-transferase M1 allele, as opposed rotrose whose GSTM1 gene has been deleted. Sulforaphane, the major isothiocyanate derived from 4-methylsulfinylbutyl glucosinolate, is thought to be the main agent conferring protection.
Objective: We compared sulforaphane metabolism in GSTM1-null and GSTM1-positive subjects after they consumed standard broccoli and high-glucosinolate broccoli (super broccoli).
Design: Sixteen subjects were recruited into a randomized, 3-phase crossover dietary trial of standard broccoli, super broccoli, and water. Liquid chromatography linked to tandem mass spectrometry was used to quantify sulforaphane and its thiol conjugates in plasma and urine.
Results:GSTM1-null subjects had slightly higher, but statistically significant, areas under the curve for sulforaphane metabolite concentrations in plasma, a greater rate of urinary excretion of sulforaphane metabolites during the first 6 h after broccoli consumption, and a higher percentage of sulforaphane excretion 24 h after ingestion than did GSTM1-positive subjects. Consumption of high-glucosinolate broccoli led to a 3-fold greater increase in the areas under the curve and maximum concentrations of sulforaphane metabolites in plasma, a greater rate of urinary excretion of sulforaphane metabolites during the first 6 h after consumption, and a lower percentage of sulforaphane excretion after its ingestion than did the consumption of standard broccoli.
Conclusions:GSTM1 genotypes have a significant effect on the metabolism of sulforaphane derived from standard or high-glucosinolate broccoli. It is possible that the difference in metabolism may explain the greater protection that GSTM1-positive persons gain from consuming broccoli. The potential consequences of consuming glucosinolate-enriched broccoli for GSTM1-null and -positive persons are discussed.
Key Words: Sulforaphane • glucosinolates • broccoli • super broccoli • GSTM1 • chemoprotection
INTRODUCTION
Of all the fruit and vegetables associated with a potential reduction in cancer risk, the evidence is strongest for cruciferous vegetables. Crucifers belong to the family Brassicaceae and include Brassica oleracea (broccoli, cabbage, cauliflower, and Brussels sprouts), B. rapa (Chinese cabbage and turnips), and several salad crops, such as Rorippa nasturtium-aquaticum (watercress) and Eruca sativa (rocket). Evidence for the protective effect of crucifers comes from epidemiologic studies conducted in the United States (1-5), Shanghai (6, 7), and Singapore (8, 9), in which associations were reported between the consumption of cruciferous vegetables and the reduction in cancer risk at several sites, including lung (4, 5, 7, 9), breast (1, 6), colon and rectum (3, 8), and prostate (2, 10, 11). The epidemiologic evidence is supported by studies of animal and cell models (12). The main mechanism proposed for the protective effect of crucifers is the activity of isothiocyanates derived from the metabolism of glucosinolates that accumulate within these vegetables. Isothiocyanates are generated from glucosinolates either by the action of plant thioglucosidases or, if the plant enzymes have been denatured by cooking, by the action of microbial enzymes in the colon.
Sulforaphane (1-isothiocyanato-4-methylsulfinyl butane, Figure 1), the major isothiocyanate derived from the 4-methylsulfinylbutyl glucosinolate that accumulates in broccoli florets, is a potent inducer of phase II detoxification enzymes (13), inhibits phase I enzymes (14, 15), and can induce cell cycle arrest and apoptosis (16, 17). When sulforaphane is absorbed, it is conjugated with glutathione and metabolized via the mercapturic acid pathway to be excreted predominantly as N-acetylcysteine conjugates (Figure 1) (18). Although the conjugation between sulforaphane and glutathione can occur nonenzymically, it is likely that this reaction occurs via the activity of glutathione S-transferases (GSTs) in vivo (19).
FIGURE 1.. The chemical structure of glucosinolate, its conversion to isothiocyanate (ITC) and thiol conjugates, and the glucosinolate side chain structures of sulforaphane and alkenyl isothiocyanates (inset).
Mammalian GSTs were recently reviewed by Hayes et al (20). Briefly, 3 mammalian GST gene families exist—cytosolic, mitochondrial, and microsomal. The cytosolic GSTs are all dimeric and are derived from 17 subunits within 7 classes: (GSTA1-A5), µ (GSTM1-M5), (GSTO1 and O2), (GSTP1), (GSTT1 and T2), (GSTS1), and (GSTZ1). The subunits from the and µ classes can form heterodimers. A single class of mitochondrial GST occurs in humans, GST . The microsomal GSTs have been designated as membrane-associated proteins in eicosanoid and glutathione metabolism and comprise 4 subgroups. They are structurally unrelated to cytosolic and mitochrondial GSTs. The catalytic activity of relatively few GSTs in isothiocyanates and glutathione conjugation has been studied, but these studies have concluded that, in general, GSTM1-1 and GSTP1-1 have the greatest activity on sulforaphane, although conjugation between glutathione and sulforaphane tends to be slower than with other dietary isothiocyanates (19, 21). The high expression of GSTM1 in the liver combined with its activity suggests that it may be of particular importance in the metabolism of dietary isothiocyanates. It is also important to note that GSTs will catalyze the disassociation of isothicyanate-glutathione conjugates, although the decomposition of the glutathione conjugates is slow compared with the rate of conjugate formation (19, 22). Polymorphisms have been described in many genes of this family (20, 23), including null mutations in GSTM1 and GSTT1, which result in the absence of a functional gene product. The frequency of the homozygous null genotype of GSTM1 varies between 39% and 63%. The frequency of the homozygous null GSTT1 genotype is between 10% and 21% for whites, but can be as high as 64% in some Asian populations (24).
Epidemiologic studies conducted in the United States have concluded that GSTM1-positive persons gain greater cancer protection from either broccoli consumption or total cruciferous vegetable consumption than do GSTM1-null persons (2, 4, 5). Additional support for this theory comes from the study by Lin et al (3) in which a significant reduction in colorectal adenomas was observed only in GSTM1-positive subjects who consumed one portion of broccoli/wk, although both GSTM1-positive and -null subjects had a reduced incidence of adenomas with higher broccoli intakes. In contrast, studies conducted in Asia that estimated crucifer consumption either by food-frequency questionnaires (9) or by quantification of isothicyanates in urine (6, 7), which has been correlated with crucifer intake (25, 26), concluded that GSTM1- and GSTT1-null persons may gain greater protection from crucifer consumption than do GSTM1- and GSTT1-positive persons. The major cruciferous component of the diets in these Asian-based studies would have been Chinese cabbage, as opposed to broccoli. In contrast with the GSTM1 and GSTT1 polymorphisms, no statistically significant association was observed between the GSTP1 313AG polymorphism, urinary isothiocyanate concentrations, and breast cancer risk in a study conducted in Shanghai (6). The possible role of other GST polymorphisms in the association between crucifer intake, isothiocyanate excretion, and cancer risk remains to be explored.
Broccoli has become a popular vegetable and consumers associate its consumption with health benefits (27). Epidemiologic studies conclude that consuming one portion per week can reduce the risk of prostate and lung cancer by 50% in certain parts of the population (2, 4). It is difficult to reconcile this reduction in risk with current knowledge of sulforaphane metabolism, in which sulforaphane, after absorbption, is found at relatively low concentrations in the plasma and rapidly excreted.
In the present study, we investigated the consequences of GSTM1 genotypes on sulforaphane metabolism and excretion after the consumption of standard broccoli and a novel, conventionally bred broccoli (super broccoli), which was selected for its enhanced glucosinolate content, with the expectation that it could deliver higher systemic concentrations of sulforaphane.
SUBJECTS AND METHODS
Subjects and study protocol
Seven men and 9 women aged 18–46 y (Table 1) were recruited by research nurses at the Wolfson Digestive Diseases Center, Queen's Medical Center, Nottingham University Hospital National Health Service Trust. Ethical approval for the trial was obtained from the University of Nottingham Medical School Ethics Committee (reference E/10/2003). All subjects gave written informed consent and were screened for full blood count, liver function tests, urea electrolytes, and fasting glucose concentrations. Subjects were excluded if they were pregnant, smokers, had a diagnosis of a long-term medical condition, or were taking dietary supplements. The subjects were allocated into a randomized, 3-phase (broccoli, super broccoli, and water), crossover design with a 21-d washout between each phase. The trial was conducted from 8 January to 19 May 2004.
View this table:
TABLE 1. Sex, age, and BMI for glutathione S-transferase (GST) M1 positive and null genotypes
The subjects avoided eating foods that were known to contain glucosinolates or isothiocyanates, spicy foods, and alcohol for 24 h before the start of the study. Baseline samples of blood and urine were obtained from the subjects after they had fasted. On the following day, the subjects returned to the Center and consumed a 150-mL test meal of broccoli soup (at room temperature, see below), super broccoli soup, or water. A subsample of the soup was taken immediately before consumption for analysis of sulforaphane and thiol conjugates. This enabled the precise amount of sulforaphane and metabolites that was consumed by each subject to be quantified. Blood samples were collected in tubes that contained lithium heparin at 0, 20, 40, 60, 100, 120, 150, 180, 210, 240, 300, 360, 420, and 480 min after consumption of the test meal. An additional blood sample was obtained the next day (24 h after intervention). Plasma was immediately prepared by centrifugation of the blood samples at 2000 x g for 10 min at 4 °C and subsequently stored at –40 °C until analyzed. Urine was collected between 0–2, 2–4, 4–6, and 6–24 h after consumption, acidified with hydrochloric acid, and stored at –40 °C until analyzed after the volume was recorded. Biopsy samples of gastric mucosal tissue were obtained by endoscopy at baseline and 6 and 24 h after consumption of the soup for gene and protein expression studies that will be reported separately, but the samples were also used for GST genotyping.
Broccoli cultivation and preparation
Standard broccoli (cultivar Marathon) and high-glucosinolate broccoli (super broccoli) were grown at the Agricultural Development and Advisory Service experimental research station, Terrington, United Kingdom. The development of super broccoli was described elsewhere (28). Individual soup portions of the 2 cultivars were prepared by cooking 100 g florets with 150 mL water for 90 s on high power in a 700-W microwave oven followed by homogenization. Preliminary studies showed that plant thioglucosidases remain active and glucosinolates are converted to isothiocyanates without nitrile formation with this cooking method. The broccoli soup was stored frozen at –18 °C. Portions were thawed at room temperature 4 h before consumption with no additional processing.
Analytic standards
The cysteine, cysteine-glycine, glutathione, and N-acetylcys-teine conjugates of sulforaphane and N-acetyl(N-butylthiocar-bamoyl)-L-cysteine (internal standard) were synthesized according to published methods (29).
Laboratory techniques
Analysts were blinded to the intervention meal and to the results of genotyping. Sulforaphane and its cysteine, cysteine-glycine, glutathione, and N-acetylcysteine conjugates were quantified in soups, plasma, and urine by gradient liquid chromatography linked to a tandem electrospray ionization mass spectrometer (Micromass Quattro Ultima; Waters, Manchester, United Kingdom) with multiple reaction monitoring mass spectrometry and internal standard quantification. A detailed description of the method is available (30). Briefly, broccoli soup samples were extracted with water (x3) and centrifuged at 11 600 x g for 5 min at room temperature. Plasma samples (0.5 mL) were prepared by adding 100 µL precooled (4 °C) trifluoroacetic acid followed by centrifugation at 11 600 x g at 4 °C for 5 min and filtration of the supernatant fluid through a 0.2 µm membrane. Urine samples (1.0 mL) were diluted to 10 mL with precooled (4 °C) ammonium acetate buffer. The injection volume was 10–50 µL, the HPLC column was a Zorbax SB-Aq (3.5 µm particle size, 100 x 2.1 mm, Agilent Technologies, Waldbronn, Germany), and the mobile phase consisted of ammonium acetate buffer (10 mmol/L, pH4; solvent A) and acetonitrile plus 0.01% acetic acid (solvent B) in a linear gradient from 5% B to 30% B over 5 min with 6 min reequilibration time. Mass spectrometry analysis was performed with the use of electrospray ionization in positive ion mode with source and desolvation temperatures set at 125 and 350 °C, respectively. The interday precision values (percentage of relative SD) for quality-control samples for the plasma and urine methods were 9.1% and 14.5%, respectively. The limits of quantitation for all analytes (0.5 mL) were in the ranges 10–100 nmol/L for plasma and 15–150 nmol/L for urine (1.0 mL).
Gastric tissue samples were obtained from the subjects by endoscopy. Genomic DNA was isolated from the tissue with a Genelute Genomic DNA mini prep kit (Sigma-Aldrich, Poole, United Kingdom). GSTM1 and GSTT1 genotypes were identified by real time polymerase chain reaction (PCR) (Taqman; Applied Biosystems, Warrington, United Kingdom) on a 7500 Real Time PCR System (Applied Biosystems). DNA was amplified in 25-µL reactions that contained Taqman Universal Master Mix (Applied Biosystems), 50 ng genomic DNA, 500 nM forward and reverse primers, and 200 nM Taqman probe dye-labeled with 5'-FAM (6-carboxyfluorescein; Sigma Genosys, Haverhill, United Kingdom) and 3'-TAMRA (6-carboxytetramethylrhodamine; Sigma Genosys) (Table 2). Reaction conditions consisted of Amplitaq Gold activation at 95 °C for 10 min followed by 40 PCR cycles at 95 °C for 15 s and then 60 °C for 1 min. The genotyping was repeated twice, with identical results.
View this table:
TABLE 2. Primer sequences for assessing the glutathione S-transferase (GST) M1 and T1 genotypes
Pharmacokinetic analysis
The plasma areas under the curve (AUCs) were calculated with the use of linear trapezoidal approximation from the plasma concentration–time data, and the elimination half-life was estimated by nonlinear least squares regression analysis of the elimination phase of the plasma concentration–time curve (user-defined functions in Microsoft Office Excel, 2003 edition; Microsoft Corporation, Seattle, WA). The rate of urinary excretion of sulforaphane metabolites between 0 and 6 h was estimated with the use of a linear regression model.
Statistical analysis
A statistical model was created in R version 2.1.1 (31) to examine the effect of broccoli type, GSTM1 genotype, and the interaction between these 2 factors on AUC, maximum concentration (Cmax), 24 h concentrations, the rate of urinary excretion, the percentage of ingested sulforaphane recovered in urine, and the percentage of sulforaphane and the different sulforaphane metabolites in plasma and urine. Because the data consist of repeated measures on each subject (each of the 16 subjects consumed broccoli and super broccoli), repeated-measures models were fitted alongside the standard regression models. In most cases, the repeated-measures models were a significant improvement over the standard models. Regression diagnostics were checked and, in some cases (plasma AUC, plasma Cmax and % metabolites excreted in urine), a log transformation of the response variable was required.
RESULTS
Of the 16 subjects, 9 were GSTM1 positive and 7 were GSTM1 null (Table 1). One subject in each of these groups was GSTT1 null. One subject, who was GSTM1 positive, was unable to complete one phase of the study because of illness.
The super broccoli soups contained 3.4-fold greater amounts of sulforaphane and sulforaphane-derived thiol conjugates than did the standard broccoli soups (Table 3). In each case, >80% of metabolites were free sulforaphane, with lower concentrations of the sulforaphane-glutathione, sulforaphane-cysteine-glycine, and sulforaphane-cysteine conjugates (Figure 2). After consumption of water, no sulforaphane or sulforaphane-derived thiol conjugates were detected in plasma or urine. In contrast, after consumption of both soups, a rapid increase in the plasma concentration of free sulforaphane and sulforaphane-derived metabolites was observed, which reached a maximum concentration after 1.5 h for standard broccoli and after 2 h for super broccoli (Figure 3). After this point, the concentration fell rapidly so that only trace concentrations of sulforaphane and its metabolites could be detected in the plasma after 24 h (Figure 3). The AUC, as a measure of exposure, and Cmax after consumption of the super broccoli were 3 times those after consumption of the standard broccoli (Figure 3 and Table 3).
View this table:
TABLE 3. Summary of metabolic data after consumption of either broccoli or super broccoli by glutathione S-transferase (GST) M1 positive and null subjects1
FIGURE 2.. The mean (±SD) percentage (n = 31) of free sulforaphane and sulforaphane-thiol conjugates in the soup (broccoli and super broccoli), plasma (area under the curve 0–24 h), and urine (total excreted 0–24 h). SF, sulforaphane; SF-GSH, sulforaphane-glutathione; SF-Cys-Gly, sulforaphane-cysteine-glycine; SF-Cys, sulforaphane-cysteine; SF-NAC, sulforaphane-N-acetylcysteine.
FIGURE 3.. Mean (±SD) change in concentration of free sulforaphane (SF) and its thiol conjugates in plasma with time (n = 15 for broccoli and 16 for super broccoli).
When stratified by genotype, a small but significant increase in plasma AUC was observed for the sulforaphane metabolites (ie, the sum of free sulforaphane, sulforaphane-glutathione, sulforaphane-cysteine-glycine, sulforaphane-cysteine, and sulforaphane-N-acetylcysteine) in the GSTM1-null subjects compared with the GSTM1-positive subjects (Table 3). The GSTM1-null subjects had Cmax values that were marginally, but not significantly, greater than those of the GSTM1-positive subjects (P = 0.059). No significant differences were observed in the 24 h concentrations, by which time low concentrations of sulforaphane metabolites were detected in the plasma.
The excretion of sulforaphane metabolites by each subject for the first 6 h after the consumption of either broccoli or super broccoli fitted a linear model (
FIGURE 4.. The mean (±SD) percentage of free sulforaphane and its thiol conjugates excreted in urine (n = 9 for GSTM1-positive subjects and 7 for GSTM1-null subjects). = GSTM1-null subjects, • = GSTM1-positive subjects.
The percentage of free sulforaphane and the different thiol conjugates in plasma were not significantly different between the 2 GSTM1 genotypes or after consumption of the different types of broccoli, with 2 exceptions. First, significantly greater amounts of sulforaphane-glutathione conjugates were observed in the plasma after consumption of super broccoli than after consumption of standard broccoli (P = 0.002; Table 4). Second, the GSTM1-positive subjects had a higher percentage of N-acetylcysteine conjugates in the plasma than did the GSTM1-null subjects (P = 0.038; Table 4). However, the absolute differences were relatively small in both cases. The ratio of the different thiol conjugates in urine was in contrast with that of plasma (Figure 3, Table 4, and Table 5). However, as with plasma, only small effects of GSTM1 genotype or type of broccoli on the percentage of the different sulforaphane metabolites were observed. The GSTM1-positive subjects had a higher percentage of sulforaphane-N-acetylcysteine conjugates in the urine than did the GSTM1-null subjects (P = 0.02, Table 5), which was consistent with the results from the plasma analyses. A greater percentage of sulforaphane-cysteine conjugates was observed in the urine after consumption of super broccoli than after consumption of standard broccoli (P = 0.004, Table 5). No evidence was seen for any interaction between genotype and broccoli type for the percentage of the different sulforaphane metabolites in plasma or urine, with the exception of sulforaphane-cysteine-glycine conjugates in urine, but this was only marginally significant (P = 0.05)
View this table:
TABLE 4. The percentage of free sulforaphane and thiol conjugates in plasma in the 24 h after consumption of either broccoli or super broccoli1
View this table:
TABLE 5. The percentage of free sulforaphane and thiol conjugates in the urine in the 24 h after consumption of either broccoli or super broccoli1
DISCUSSION
The concentration of sulforaphane in soups made from super broccoli was 3-fold that in soups made from standard broccoli (Table 3). Likewise, after consumption of the super broccoli soup, peak concentrations of sulforaphane-derived metabolites in the plasma and AUC were 3-fold those after standard broccoli soup was consumed (Table 3 and Figure 3). This suggests that increasing the concentration of sulforaphane in broccoli is directly related to an increased exposure after consumption. After 24 h, only trace concentrations of sulforaphane metabolites were detected in the plasma after consumption of either meal. A relatively high proportion of sulforaphane metabolites was recovered in the urine—82% for broccoli and 55% for super broccoli (without stratifying by genotype)—24 h after consumption. These values are higher than those previously reported in a study of the recovery of isothiocyanate metabolites from broccoli (32%) (18). The high level of recovery of sulforaphane metabolites in our study may be due to our method of preparation, which maximizes the availability of sulforaphane for absorption. Furthermore, a study of watercress reported a recovery of up to 77% of ingested isothiocyanate (32), similar to the recovery we report.
The GSTM1 genotype had significant, but relatively small, effects on the plasma AUC values for sulforaphane metabolites. The GSTM1-null subjects had higher AUC values for total sulforaphane metabolites than did the GSTM1-positive subjects, but the absolute difference in the AUCs was relatively small (Table 3). No significant difference in Cmax was observed between the GSTM1 genotypes. However, strikingly, highly significant differences in both the rate of sulforaphane metabolite excretion and the total amount excreted were observed between the GSTM1 genotypes (Table 3 and Figure 4). In contrast with previous assumptions (6, 7, 9), the GSTM1-null subjects excreted sulforaphane metabolites at a faster rate during the first 6 h after consumption and excreted significantly more sulforaphane metabolites during the 24 h after consumption of either standard broccoli or super broccoli than did the GSTM1-positive subjects (Table 3 and Figure 4). The rate of sulforaphane metabolite excretion was also greater for both GSTM1 genotypes after consumption of super broccoli than after consumption of standard broccoli. However, the percentage of excretion of total ingested sulforaphane metabolites was significantly less after consumption of super broccoli than after consumption of standard broccoli (Table 3 and Figure 4).
Because of our novel method of liquid chromatography linked to tandem mass spectrometers, we were also able to quantify the percentage of free sulforaphane and the individual thiol conjugates in plasma and urine, as opposed to quantifying the cyclocondensation product between isothiocyanate-thiol conjugates and 1,2-benzenedithiol, as in previous studies (18). The only effect of genotype was the lower percentage of sulforaphane-N-acetylcysteine conjugates found in the plasma and urine of the GSTM1-null subjects than in the GSTM1-positive subjects (Table 4). Likewise, the type of broccoli had little effect, with a small but significant increase in the percentage of sulforaphane-glutathione conjugates in plasma and sulforaphane-cysteine conjugates in urine after consumption of super broccoli than after consumption of standard broccoli. Thus, we conclude that neither the genotype nor the type of broccoli has a major effect on the relative proportions of sulforaphane conjugates in plasma or urine.
Three hypotheses can explain the lower concentrations of sulforaphane metabolites that occur in the plasma and the lower proportion of ingested sulforaphane metabolites excreted in the urine of GSTM1-positive persons than in GSTM1-null persons. First, GSTM1-positive persons may absorb less sulforaphane from the gastrointestinal tract lumen than do GSTM1-null persons, with the remaining sulforaphane being excreted with the feces. Second, GSTM1-positive persons may absorb the same amount of sulforaphane that GSTM1-null persons do, but they may metabolize and excrete sulforaphane and sulforaphane metabolites by an alternative pathway. Third, GSTM1-positive persons may retain sulforaphane and sulforaphane metabolites in certain tissues that are then metabolized and excreted at a later time, either via the mecapturic acid pathway or by an alternative route. Currently, no experimental evidence exists in humans to support any of these 3 hypotheses. However, because the epidemiologic evidence suggests that GSTM1-positive persons gain a greater protection than do GSTM1-null persons (Table 6), it may be surprising if they actually absorbed less sulforaphane.
View this table:
TABLE 6. Summary of US-based epidemiologic studies that correlate broccoli or crucifer consumption with the risk of cancer, stratified by glutathione S-transferase (GST) M1 genotype1
The study by Conaway et al (33) provides evidence of isothiocyanate metabolism through nonmercapturic acid pathways. [14C] labeled phenylethyl isothiocyanates and phenylhexyl isothiocyanates (PHITCs) were administered to rats, the latter being 1–2 orders more potent as an anticarcinogenic agent. Although phenylethyl isothiocyanate was rapidly excreted in urine, PHITC was retained in several organs, notably in the liver and lungs. Moreover, there was evidence of exhalation of labeled CO2 derived from PHITC (33), which was indicative of alternative metabolic routes. Although this difference may have been due to differences in lipophilicity between phenylethyl isothiocyanates and PHITC, it does indicate that isothiocyanate has alternative metabolic pathways.
The hypothetical scenario in which a proportion of sulforaphane is retained and not rapidly excreted in the urine may explain the apparent discrepancy between the protective effect of a relatively modest consumption of broccoli (approximately one portion per week) in GSTM1-positive persons (Table 6) and the concentration of sulforaphane required to induce protective mechanisms in laboratory studies. In our study, the maximum concentration of sulforaphane and metabolites in the plasma was 2.3 µmol/L for standard broccoli and 7.3 µmol/L for super broccoli (Table 3). Studies conducted on cell cultures and animal models have routinely used concentrations considerably greater than this, up to 40 µmol/L, for several hours to induce gene and enzyme expression (16, 34-36). The only tissues in which these concentrations are likely to be found, in the absence of any accumulation in tissues, are in the gastrointestinal tract before absorption and the bladder where these compounds accumulate before excretion. Despite this, epidemiologic evidence provides support for a protective effect at other sites, including the prostate, lung, and breast. This discrepancy may be resolved if sulforaphane, in a similar manner to PHITC, accumulates in certain tissues and results in higher local concentrations.
Thus, in our study, 2 factors were important in determining the exposure to sulforaphane after broccoli consumption—GSTM1 genotype and the amount of sulforaphane consumed. The GSTM1-null subjects excreted essentially all (
Although our data are consistent with epidemiologic studies that were conducted in the United States, they are inconsistent with those conducted in Singapore and Shanghai. This may be due to contrasting patterns of crucifer consumption within these different locations. In the United States, 40% of crucifer consumption is broccoli, whereas consumption of B. rapa, either as turnips or Chinese cabbage, is very low (1, 3, 5, 39). Conversely, in Singapore and Shanghai, most Brassica vegetable consumption is Chinese cabbage and other forms of B. rapa (6, 8). These differences may be critical in interpreting the effect of GSTM1 genotype. Within the United States, the most prevalent isothiocyanate in the diet is sulforaphane, which is obtained from broccoli and some other forms of B. oleracea, whereas in Shanghai and Singapore the most prevalent isothiocyanates are 3-butenyl and 4-pentenyl isothiocyanates, which are obtained from B. rapa (Figure 1). The enzymology of sulforaphane and 2-propenyl isothiocyanates with GST isozymes are different (19). Assuming that the biochemistry of 2-propenyl isothiocyanate is similar to that of 3-butenyl and 4-pentenyl isothiocyanates, we expect that a GSTM1 deletion will have contrasting effects on the metabolism of sulforaphane and alkenyl isothiocyanates. This may explain the apparent paradoxical diet-gene interactions observed in the United States and Asia.
In conclusion, we speculate that a proportion of sulforaphane may be retained within the body, rather than rapidly excreted, and that this may mediate the anticarcinogenic activity of broccoli. The amount retained depends on GSTM1 genotype and the amount of sulforaphane initially ingested. If this hypothesis is correct, GSTM1-null persons could ameliorate the effect of their genotype by consuming a higher dose of sulforaphane, either by eating larger servings of standard broccoli or by eating the same amount of a high-glucosinolate broccoli.
ACKNOWLEDGMENTS
We thank the participating doctors and technicians for clinical advice and technical support. We also thank Y-P Bao for advice about GST genotyping, R Foxall for designing the statistical model, and the volunteers.
RFM conceived and coordinated the study. RFM designed the final protocol in collaboration with DAB, AVG, CJH, JAS, and MAT. AVG, JAS (study scientists), CA (research nurse), and PF (endoscopist) conducted the study. JRB developed the genotype assay. AA and AVG undertook the analyses, under supervision of DAB and RFM. AVG and RFM wrote the manuscript, and all authors were involved in the critical revision of the manuscript. None of the authors had any conflicts of interest.
REFERENCES