Occurrence of indole compounds in some vegetables : toxicological implications of nitrosation with emphasis on mutagenicity

From the introduction to the chemistry, occurrence and formation of N-nitroso compounds (NOC) (Chapter 1), it can be concluded that human exposure to these potent carcinogenic compounds is mainly through the endogenous nitrosation of dietary precursors. Vegetables are the major source of nitrosating compounds, while also nitrosatable substrates can occur in vegetables. Therefore the first aim of the present study was to screen Dutch vegetables for the presence of nitrosatable compounds by measuring their potential to form directly mutagenic NOC upon nitrite treatment. The second aim was to study the identity and mutagenic properties of the NOC formed. Since it was not feasible to cover all vegetables in connection to the second aim of the study, the efforts were concentrated on two of them, brassicas and fava beans. The latter were chosen because they were already known to contain precursors of directly mutagenic NOC and consequently the study was divided into two parts: Part 1 dealing with brassicas and Part 2 with fava beans.<strong>Part 1</strong>In Chapter 4 experiments are described in which extracts of 31 Dutch vegetables were screened for their ability to form directly mutagenic NOC upon nitrite treatment, irrespective of their nitrate content. All vegetables tested formed NOC upon nitrosation and induced <em>Salmonella</em> ( <em>S.</em> ) <em>typhimurium</em> revertants; Brassica vegetables were high responders on both of these parameters. Moreover, a significant correlation was found between their glucosinolate content (both alkyl/aryland indolylglucosinolate) and the amounts of NOC formed in extracts of these vegetables upon nitrosation. This suggests that glucosinolates are involved in the formation of NOC. Therefore purified glucosinolates were tested for their potential to form NOC (Chapter 6). Only indolylglucosinolates and their hydrolysis products formed NOC upon nitrosation. Mutagenicity was restricted to the nitrosated hydrolysis products. Since upon hydrolysis of indolylglucosinolates indole compounds are formed, which are nitrosatable substrates, the kinetics of the formation of NOC from indole compounds were investigated, as well as the stability of the nitrosated products formed (Chapter 5). Indole-3-acetonitrile (I _{<font size="-1">3</font>} A), indole-3-carbinol (I _{<font size="-1">3</font>} C) and indole, the hydrolysis products of the most commonly indoyglucosinolate, glucobrassicin, immediately reacted with nitrite to form directly mutagenic NOC and after an incubation time of about 15 min. maximal amounts of NOC were formed. The formed NOC were stable at both pH 2 and 8, but only when nitrite was present.In order to determine the contribution of indole compounds to the total mutagenicity of nitrite treated brassicas, the presence of several known indole compounds in green cabbage was investigated (Chapter 6). Only indole-3-carboxaldehyde and I _{<font size="-1">3</font>} A could be detected. Both were not found to be important precursors of directly mutagenic NOC in green cabbage. The former did not form NOC at all and the second, although it occurred in considerable amounts (12 mg/kg fresh weight), only contributed for 2% to the total mutagenicity of nitrite treated green cabbage.From these results it can be concluded that:1. The correlation found between the amounts of glucosinolates in brassicas and the levels of NOC formed in extracts of these vegetables upon nitrite treatment is not based on a causal relationship.2. In brassicas both glucosinolates and indole compounds should not be considered as important precursors of NOC.3. By testing compounds out of their normal matrix, results can be obtained which differ considerably from those obtained under more realistic situations; (I _{<font size="-1">3</font>} A), when tested pure, was found to be a potent precursor of NOC, while in green cabbage its contribution to the total mutagenicity was marginal.Although (I _{<font size="-1">3</font>} A), was not considered an important precursor of NOC and contributed only marginally to the mutagenicity of nitrite treated brassicas, it was found to be a potential genotoxic and tumour promoting agent after nitrosation (Chapter 9). Recently it was shown in a Japanese study that nitrosated I _{<font size="-1">3</font>} A is able to induce tumorous lesions in the fore stomach of rats. Therefore the possible risk involved in the consumption of brassicas due to the endogenous formation of nitrosated I _{<font size="-1">3</font>} A was assessed (Chapter 7). For different reasons it seems most unlikely that the endogenous formation of nitrosated I _{<font size="-1">3</font>} A out of brassicas should be considered as a realistic threat to human health: (1) The nitrosation rate of I _{<font size="-1">3</font>} A <em>in vivo</em> is expected to be low. (2) In animal studies nitrosated I _{<font size="-1">3</font>} A was administered pure, in very high concentrations. (3) Nitrosated I _{<font size="-1">3</font>} A will only be stable in the presence of nitrite.Since in brassicas other precursors of directly mutagenic NOC occur, further studies are recommended to elucidate their identity and the implications of their potential endogenous nitrosation.<strong>Part 2</strong>4-Chloro-6-methoxyindole (4C6MI), the naturally occurring indole compound in fava beans, was evaluated for its genotoxic and tumour promoting potential after nitrite treatment (Chapter g). Remarkably, nitrosated 4C6MI appeared to have both genotoxic and tumour promoting potentials. The initiation effects were measured in bacteria and mammalian cells, the tumour promoting effects were measured by inhibition of gap junctional intercellular communication of V79 Chinese hamster cells. Both effects were observed at concentrations, which were in the same order of magnitude as the estimated daily intake of 4C6MI in a Colombian population. Hence, the results support the model for gastric cancer etiology, proposed by Correa et al. (1976, 1983). In this model the formation of NOC out of fava beans was supposed to be a causative factor of gastric cancer of the intestinal type, endemic in Colombia.In the study described in Chapter 10 the occurrence of 4C6MI in Dutch fava beans was investigated. An improved purification method was developed because the procedures described in the literature proved to be inadequate. Although the new purification method requires further improvement, a reasonable estimate of the levels of 4C6MI could be made. The levels ranged from about 3 to 7.5 mg/kg dry weight, which is one to two orders of magnitude higher than those reported for Colombian beans. However, in Chapter 12 It was assumed that the levels of 4C6MI in Colombian beans may be much higher than those reported. Further studies are required to investigate this aspect. From the results described in Chapter 10 it was concluded that almost all mutagenicity of nitrite treated Dutch fava beans can be attributed to 4C6MI.Although in previous studies a difference was observed in the mutagenicity of nitrite treated white and brown cooking cultivars after nitrosation, this was not found in the present study (Chapter 10 & 11). Moreover, all cultivars induced much more <em>S. typhimurium</em> revertants after nitrosation than in previous studies. No explanation can be given for these equivocal results and therefore studies are recommended to investigate the influence of environmental factors on the levels of 4C6MI in fava beans (Chapter 10).The mutagenicity of nitrite treated fava beans could be inhibited for 80-100% by addition of casein, indicating binding of nitrosated 4C6MI to casein (Chapter 11). This binding is independent of pH, in a range of pH 2-6 and appeared to be reversible, since mutagens could be released from the casein. It was estimated that about 25% of nitrosated 4C6MI formed in nitrite treated fava beans binds to proteins present in fava beans. This binding is reversible too. Fava bean mutagens also bind to wheat bran. Although the binding efficiency to wheat bran is less than to casein and the reversibility of this binding has not been studied, the binding of fava bean mutagens to wheat bran can be an important observation. Wheat bran fibres will not be digested and therefore might serve as vehicles for mutagens to leave the body without harmful effects. Further investigation of this is recommended.The results of the present study are further support for the hypothesis that the consumption of fava beans is causally related to the etiology of gastric cancer in Colombia. But the risks involved in the consumption of fava beans due to the endogenous nitrosation of 4C6MI for the population in The Netherlands is considered to be low, because of: (1) The low consumption of fava beans. (2) The lower levels of nitrate in Dutch drinking water. (3) Differences in food patterns between the Colombian and the Dutch population (the levels of consumption of vegetables and dairy products in particular). However, it can not be excluded that there are groups in The Netherlands with consumption habits deviating from that of the average population, who can in principle be assigned as persons at risk.It is recommended to investigate the <em>in vivo</em> nitrosation of 4C6MI and the carcinogenicity of its nitrosated product in further detail by using an appropriate animal study.