Introduction

Top Abstract Introduction Materials and Methods Results and Discussion References

Watercress (Nasturtium officinale) contains substantial amounts of gluconasturtiin, the precursor to phenethyl isothiocyanate (PEITC¹), a strong inhibitor of carcinogenesis in several animal model systems (Fenwick et al., 1983; Hecht, 1999). When uncooked watercress is chewed or chopped, gluconasturtiin comes into contact with the enzyme myrosinase, which also occurs in the plant. This results in hydrolysis of gluconasturtiin and production of PEITC (Fenwick et al., 1983). In humans, approximately 2 to 6 mg of PEITC is released per ounce of watercress consumed (Chung et al., 1992; Hecht et al., 1995).

PEITC is a particularly effective inhibitor of nitrosamine carcinogenesis in rodents when administered prior to, or concurrent with, carcinogen. It inhibits lung tumor induction by 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) in mice and rats, esophageal tumorigenesis by N-nitrosobenzylmethylamine (NBzMA) in rats, and lung and pancreatic tumor induction by N-nitrosobis-2-oxopropylamine in hamsters (Stoner et al., 1991; Nishikawa et al., 1996; Hecht, 1998). One important mechanism by which PEITC inhibits nitrosamine carcinogenesis in rodents is inhibition of cytochrome P450s (P450s) involved in their metabolic activation, although induction of phase 2 enzymes may also play a role (Hecht, 1998, 1999; Yang et al., 1994; Zhang and Talalay, 1994).

Several studies have investigated the effects of watercress consumption on drug and carcinogen metabolism in humans (Hecht et al., 1995, 1999; Chen et al., 1996; Leclercq et al., 1998). We showed that urinary levels of the NNK metabolites, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) and its glucuronide (NNAL-Gluc), increased in smokers who ate watercress (Hecht et al., 1995). These results were attributed to inhibition of P450-mediated metabolic activation of NNK by PEITC and/or induction of glucuronidation of NNAL (Hecht et al., 1995). P450s involved in NNK metabolism in humans include P450s 1A2, 2A6, and 3A4 (Hecht, 1998). PEITC inhibits human P450 1A2 in vitro (Smith et al., 1996) and the catalysis of NBzMA metabolism by P450s 2E1 and 2A6 (Morse et al., 1999). Chen et al. (1996) observed inhibition of acetaminophen oxidative metabolism following watercress consumption, which was attributed to inhibition of P450 2E1 by PEITC. Studies by Leclercq et al. (1998) support PEITC-mediated inhibition of P450 2E1-catalyzed chlorzoxazone metabolism in vivo, and Moreno and Hollenberg (1999) reported PEITC inhibition of purified P450 2E1.

Although the specific forms have not been identified, P450 2A enzymes appear to play some role in the metabolic activation of NNK and NBzMA in rodents (Hecht, 1998; Patten et al., 1998). P450 2A6 can catalyze the metabolic activation of NNK and NBzMA (Hecht, 1998; von Weymarn et al., 1999). Since PEITC inhibits the metabolic activation of NNK and NBzMA in rodents, and may inhibit the metabolic activation of NNK in humans, we hypothesized that watercress consumption could inhibit P450 2A6 in humans. Metabolism of coumarin to 7-hydroxycoumarin (7OHC) is catalyzed specifically by P450 2A6 in humans (Pelkonen et al., 1997). Therefore, we examined the effects of watercress consumption on the urinary excretion of 7OHC in people who ingested coumarin.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion References

Study Design. The protocol was approved by the University of Minnesota Research Subjects' Protection Programs Institutional Review Board Human Subjects Committee. Subjects were healthy nonsmokers ranging in age from 19 to 30 (mean, 24.7 ± 3.8). There were eight males and seven females. Coumarin was provided by Schaper and Brummer (Salzgitter, Germany) and 5-mg capsules were prepared by the Investigational Drug Pharmacy at the University of Minnesota.

Analysis of Urine. Analysis of urine for 7OHC was carried out by high-performance liquid chromatography essentially as described by Egan and O'Kennedy (1992) except detection was by fluorescence (excitation wavelength, 350 nm and emission wavelength, 453 nm). Analyses of PEITC-NAC in urine and gluconasturtiin in watercress were performed by high-performance liquid chromatography as described previously (Chung et al., 1992; Prestera et al., 1996). Data from baseline and watercress periods were compared using Student's t test.

	Results and Discussion

Top Abstract Introduction Materials and Methods Results and Discussion References

Each 2-oz portion of watercress used in this study contained 112 mg (0.26 mmol) of gluconasturtiin. Therefore, the maximum amount of PEITC that could be released was 43 mg. The urine of our subjects was analyzed for PEITC-NAC, a major metabolite of PEITC (Chung et al., 1992). An average of 12.6 ± 3.6 mg (mean ± S.D., n = 15) of PEITC-NAC, corresponding to 6.3 mg of PEITC, was excreted from 0 to 4 h following consumption of the watercress breakfast. Therefore, the minimum mean uptake of PEITC was 6.3 mg per subject.

Sixty to 75% of orally administered coumarin is converted to the glucuronide of 7OHC (Pelkonen et al., 1997). The glucuronide was hydrolyzed and the released 7OHC quantified. In the baseline period, the mean amount of 7OHC excreted in 4 h was 3.9 ± 0.6 mg, or 70% of the administered dose (Table 1, column 8) and the majority, 80%, was excreted in the first 2 h (Table 1, column 2). The amount of coumarin excreted as 7OHC was determined on baseline visits 2 to 4 for each subject. The mean standard deviation among analyses for all 15 subjects at baseline was 15 ± 0.9%. There was no significant variation in 7OHC excretion on days 2 to 4, and there was no evidence that administration of coumarin affected coumarin metabolism on subsequent days. The effect of watercress consumption on the level of urinary 7OHC was determined on visits 6 and 7 (Table 1). In six subjects, mean levels of 7OHC in urine collected 2 h after coumarin dosing were lower in the watercress period than in the baseline period, while in the other nine subjects there was no change (Table 1, columns 2-4). Overall, the amount of 7OHC in these urine samples at baseline (3.1 ± 0.53 mg) was not significantly different from the amount after watercress consumption (2.8 ± 0.78 mg).

In the urine samples collected in the 2 to 4 h period after dosing with coumarin, levels of 7OHC were higher in five samples after watercress consumption than at baseline, while no change was observed in the other 10 (Table 1, columns 5-7). Three of the increases were seen in samples from individuals (subjects 2, 12, and 15) whose levels of 7OHC were lower in the watercress consumption than in the baseline period, for the 0- to 2-h urine collection. Overall, levels of 7OHC after watercress consumption (1.1 ± 0.50) were significantly higher (P = 0.027) than at baseline (0.77 ± 0.22) in the 2- to 4-h urine samples. Total excretion of 7OHC, 0 to 4 h after taking coumarin, was similar to previous studies (Pelkonen et al., 1997) and was unaffected by watercress consumption (3.9 ± 0.6 versus 3.9 ± 0.84 mg).

In subjects 6 and 15, the excretion of 7OHC was markedly different on the two days of watercress consumption. Subject 6 showed delayed metabolism to 7OHC on the 1st day of watercress consumption (1.2 mg, 0-2 h; 2.3 mg, 2-4 h), but not on the 2nd (2.6 mg, 0-2 h; 0.11 mg, 2-4 h), while subject 15 showed the opposite effect (2.7 mg, 0-2 h; 2.2 mg, 2-4 h on day 1; 0.5 mg, 0-2 h; 3.1 mg, 2-4 h on day 2). This was not due to differing uptake of PEITC on the 2 days (data not shown), nor was it due to analytical variation. Subjects 6 and 15 excreted consistent amounts of 7OHC on each of the 3 baseline days. When the data for subjects 6 and 15 on the watercress consumption days were omitted, the means were 3.0 ± 0.70 (n = 13, 0-2-h period) and 1.0 ± 0.42 (n = 13, 2-4-h period). The results were essentially the same as those obtained when subjects 6 and 15 were included.

Measurement of urinary 7OHC, 0 to 2 and 2 to 4 h after dosing with coumarin, is an effective way to detect changes in coumarin metabolism due to changes in P450 2A6 (Pelkonen et al., 1997). A study that used a protocol similar to ours reported a 25% decrease in 7OHC excretion in 0 to 2 h when subjects were administered grapefruit juice 30 min before coumarin. Therefore, if watercress consumption had inhibited P450 2A6 activity, we would have expected to see a decrease in urinary 7OHC 0 to 2 h after coumarin administration. However, our results indicate that, under the conditions of our study, PEITC and other constituents of watercress had at most a marginal inhibitory effect on P450 2A6 activity.

These results are consistent with those of a recent study in which we analyzed nicotine metabolites in the urine of smokers before, during, and after a period of watercress consumption (Hecht et al., 1999). P450 2A6 appears to play a major role in the metabolism of nicotine to cotinine, as well as in the further metabolism of cotinine to trans-3'-hydroxycotinine and other metabolites (Nakajima et al., 1996a,b; Yamazaki et al., 1999). We found no effect of watercress consumption on these reactions. However, watercress consumption did result in increased glucuronidation of nicotine, cotinine, and trans-3'-hydroxycotinine (Hecht et al., 1999).

In humans, P450s 1A2, 3A4, and 2A6 appear to play a role in NNK metabolic activation (Hecht, 1998). We found that watercress consumption enhanced urinary excretion of the NNK metabolites NNAL and NNAL-Gluc, which was attributed to inhibition of NNK metabolic activation and/or induction of glucuronidation (Hecht et al., 1995). PEITC inhibits NNK metabolic activation by P450 1A2 in vitro, which could account in part for these results (Smith et al., 1996). The results of the present study suggest that inhibition of P450 2A6 by PEITC or other constituents of watercress is probably not involved in the perturbation of NNK metabolism by watercress consumption. Either PEITC is a relatively weak inhibitor of P450 2A6 or this enzyme plays a minor role in NNK metabolism in vivo in humans.