Dietary Constituents and Carcinogenesis
The caloric content of the diet is made up of the caloric values of its various constituents. The constituents of the diet consist of protein, fat, and carbohydrate as the major components and vitamins, minerals, and related factors as the minor components. As shown above in relation to the overall carcinogenic process, the caloric content of the diet plays a major role. However, individual constituents of the diet also modify the carcinogenic process.
Dietary Protein. Early studies with rats demonstrated that the incidences of certain spontaneous neoplasms are very dependent on the level of the protein in the diet (Ross and Bras,1973). In these animals the spontaneous incidence of thymoma increased more than threefold in animals on a low-protein diet (10% casein), while increasing the dietary protein content by a factor of five decreased the incidence of thymoma by a factor of eight. In most other examples,
Figure 8.6 Effect of a 95-day period of caloric restriction on the volume of spontaneously occurring altered hepatic foci (PPF) and on the formation of liver tumors in animals administered nafenopin, a perox- isome proliferator and tumor promoting agent. Animals were fed ad libitum except for a group that were calorically-restricted to 60% of the calories of the control diet from day 285 to day 380 at the end of which time the volume of altered hepatic foci had decreased considerably (open circle, a). The insert at the right shows the number of liver tumors per rat in control (Co) and calorically restricted (FR) animals at the end of the experiment (day 900). The total column represents the number of grossly detectable lesions per liver, the hatched and solid part of the bar represents the numbers of adenomas and carcinomas. (Adapted from Grasl-Kraupp et al., 1994, with permission of authors and publisher.)
feeding either a 10% or a 50% casein diet lowered the spontaneous incidence of epithelial neo- plasms compared with that of animals on a 22% casein diet. While total tumor risk was found to be correlated with caloric consumption, the diet containing a higher amount of protein for a given caloric intake resulted in a higher risk of malignancy. Furthermore, a high absolute protein consumption shortly after weaning and that relative to body weight during the early adult period were correlated with a higher probability of developing spontaneous neoplasms (cf. Mohr and Lewkowski, 1989).
The chemical induction of neoplasia in various rodent tissues is likewise affected by the protein content of the diet. Early studies in rats showed that when animals were fed a 5% (casein) protein diet, they exhibited significantly fewer aflatoxin-induced hepatomas than rats fed four to six times higher levels of casein (Madhaven and Gopalan, 1968; Wells et al., 1976). More recent investigations by Campbell and associates (O’Connor and Campbell, 1987; Schul- singer et al., 1989) demonstrated that lowering the dose of casein dramatically inhibits the growth of altered hepatic foci induced by aflatoxin B1 in vivo. Furthermore, if the casein in the diet were replaced by wheat gluten, a low-quality protein, the development of altered hepatic foci was significantly inhibited compared with animals fed a casein diet at the same level for the same period. Lysine supplementation of the wheat gluten in the diet during the stage of promo- tion returned the growth of altered hepatic foci to normal (Schulsinger et al., 1989).
The protein content of the diet does not influence the growth of skin tumors, but a high- protein diet enhances DMBA-induced lung tumors in mice and DMH-induced neoplasms in rats (cf. Kritchevsky, 1988). In his discussion, the author points out that most studies have been car- ried out with casein as the protein source, emphasizing the significance of the study by Schul-singer et al. (1989). An increase in intake of protein prior to the administration of DMBA to induce mammary neoplasms decreased the number of tumors induced (Clinton et al., 1979). However, increasing the protein content of the diet after initiation, during the promotion phase, resulted in an increased incidence of mammary neoplasms. These results can be explained on the basis of an increased rate of metabolism of the carcinogen to noncarcinogenic metabolites in the liver of animals on a high-protein diet when DMBA is administered, resulting in a decreased level of potentially active metabolites of DMBA reaching the mammary gland (Hawrylewicz et al., 1982). The composition of the diet also appears to be important in determining the inci- dence of mouse hepatomas. In early studies, Tannenbaum and Silverstone (1949) showed that the incidence of mouse hepatomas increased with the rising content of casein in a partially puri- fied diet until 26% of the diet consisted of protein. At a higher level of dietary casein, the inci- dence decreased significantly. Thus, the effect of the high protein content of the diet on carcinogenesis may occur at the stage of initiation by stimulating the metabolism of the chemi- cal carcinogen to noncarcinogenic metabolites, with subsequent inhibition of this stage. The ef- fect of high-protein diets on the stage of promotion may be either an enhancement or an inhibition, as noted above. Unfortunately, our knowledge of the mechanism of this effect is very limited.
Of the amino acid constituents of protein, tryptophan has been most widely studied for its effects on carcinogenesis in several different species in both bladder and liver (Sidransky, 1998). The initial studies in this area were described by Dunning et al. (1950), wherein Fischer female rats fed a purified diet, that had been supplemented with DL-tryptophan and contained 0.06% 2-AAF, exhibited a high incidence of bladder carcinomas, but no carcinomas developed when 2-AAF was administered in the absence of dietary tryptophan. Later Radomski and associates (1977) reported that female beagle dogs fed a diet containing 4-aminobiphenyl with supple- mental DL-tryptophan developed a low incidence of bladder tumors, while animals receiving 4-aminobiphenyl in the absence of supplemental dietary tryptophan developed no such neo- plasms. A number of studies on hepatocarcinogenesis in the rat gave variable results when tryp- tophan was added to the diet, in some cases an enhancing effect and in other cases no effect (cf. Sidransky, 1998). L-Tryptophan has many of the characteristics of a promoting agent, including an apparent interaction with a nuclear receptor (cf. Sidransky, 1998). Thus, some of the effects of increased dietary protein on carcinogenesis may be related to the tryptophan content of the diet causing a promoting action in the carcinogenic process.
Dietary Lipids and Carbohydrates. As suggested earlier in this chapter, the effect of lip- ids in the diet on carcinogenesis may be controversial in that the effects noted are the result primarily of the caloric content of the diet rather than specifically of the lipid content and of individual lipids themselves (see above). This is emphasized by the data depicted in Table 8.5. However, the fact still remains that dietary lipids do have significant effects, primarily in en- hancing the development of a number of neoplasms induced by a variety of chemical and physi- cal factors, as noted by the listing depicted in Table 8.7 from Birt (1987). Almost all of the effects of dietary fat listed in Table 8.7 are to enhance the development of neoplasms of a variety of tissues as a result of spontaneous and induced carcinogenesis.
Notably absent from the table is the effect of dietary lipids on mammary carcinogenesis, which has been the subject of considerable investigation, both experimentally (Pariza et al.,1986) and in the human (Prentice et al., 1989; Hunter et al., 1996). A variety of studies in exper- imental animals have indicated that the caloric and fat contents of the diet independently en- hanced mammary tumor incidence in rats and mice (Freedman et al., 1990). However, Welsch et al. (1990) indicated that the increase in mammary neoplasia following DMBA administration in rats occurred only in animals fed ad libitum. Mammary carcinogenesis induced by the direct- acting carcinogen N-nitrosomethylurea (NMU) was primarily enhanced by the level of calories
Table 8.7 Effects of Dietary Fat on Carcinogenesis at Sites Other Than the Mammary Gland in Several Rodent Species
in the diet (Beth et al., 1987) although the tumor-promoting effects of dietary fat manifested a threshold rather than a linear dose-response effect (Cohen et al., 1986). Furthermore, these ef- fects of dietary fat were reportedly not related in a significant way to changes in endogenous levels of estrogen or prolactin (Clinton et al., 1995). However, early studies (Gammal et al.,1967) suggested that the type of dietary fat—i.e., corn oil versus coconut oil—significantly af- fected the development of mammary neoplasms in that animals receiving the corn oil diet devel- oped a greater number of tumors. Subsequent studies demonstrated that fish oils rich in ω3 fatty acids, particularly eicosapentaenoic and docosahexaenoic acids, consistently inhibited mam- mary carcinogenesis in mice and rats when administered during the promotion stage (cf. Welsch,1992). In contrast, safflower and corn oil, which contain predominantly ω6 fatty acids such as linoleic acid, exhibited a significant tumor-promoting effect in mouse mammary carcinogenesis induced by DMBA (Cameron et al., 1989). The structures of representative ω3 and ω6 fatty ac- ids are depicted in Figure 8.7. The mechanism of this difference in effect on tumor promotion of ω3 and ω6 fatty acids is not known but may be related to their differential effect on the synthesis of prostaglandins, a series of hormones derived from these molecules (Lands, 1992).
As indicated from Table 8.7, increases in dietary fat enhance the development of neo- plasms in a number of other tissues. The early studies of Miller and Rusch demonstrated the importance of the type of fatty acids in the diet on the incidence of azo dye–induced liver neo- plasms (cf. Miller et al., 1944). When animals were fed unsaturated fat together with the carcin-
Figure 8.7 Structures of representative ω3 and ω6 polyunsaturated fatty acids.
ogen, the tumor incidence was much greater than when animals were fed an isocaloric amount of saturated lipids. Birt et al. (1989) reported that in Syrian hamsters initiated with N-nitrosobis- (2-oxopropyl)amine (BOP) at 8 weeks of age, a three- to fourfold enhancement of pancreatic carcinogenesis occurred in those hamsters fed high-fat diets ad libitum or fed equivalent caloric intakes. Rats initiated with azaserine, which induces acinar cell adenomas and carcinomas—in contrast to the BOP-induction of ductular carcinomas of the pancreas in hamsters—were studied during the promotion stage of carcinogenesis. Azaserine induces acinar foci of acidophilic and basophilic appearance. A diet containing 20% unsaturated fat induced an increase in the number of acidophilic foci as well as an increase in the thymidine labeling index of their nuclei com- pared with rats fed a 20% saturated fat diet or a control diet of 5% unsaturated fat (Roebuck,1986). Later studies by O’Connor et al. (1989) demonstrated an inhibition of the development of these preneoplastic acinar cell foci and nodules by ω3 fatty acids in the diet, whereas ω6 fatty acids induced a significant increase in the growth of these preneoplastic lesions. Furthermore, dietary linoleic acid acts to promote pancreatic lesions in rats and hamsters (Appel et al., 1994). In the rat, an apparent threshold for this effect of linoleic acid is seen at concentrations between 4% and 8% in the diet, above which both the number and size of acidophilic foci increase (cf. Roebuck, 1992).
Dietary lipid content also enhances the development of preneoplasia and neoplasia in the intestine of mice and rats. In the mutant mouse strain min, which is genetically prone to develop polyps in the intestine, increasing the dietary fat content increased the number of polyps in the large and small bowel as well as increasing the size of the polyps in the small bowel (Wasan et al., 1997). In a review of some 14 studies of such experiments, a positive relationship was found between the incidence of colon carcinoma and fat intake (Zhao et al., 1991). This association was seen with the inbred Fischer 344 strain, but no such association was apparent for Sprague- Dawley rats. In contrast, while a high-fat diet enhanced the early development of colon neo- plasms in rats after azoxymethane administration for 2 weeks, alteration in the fat content of the diet after neoplasms had developed did not change the neoplastic outcome, suggesting a lack of effect of dietary fat on the stage of progression in this tissue (Bird et al., 1996). As with mam- mary and pancreatic carcinogenesis, the type of lipid in the diet has significant effects on the carcinogenic process in the intestine. Dietary cholesterol enhances both the development of pre- neoplastic aberrant crypts and ultimately the number of colon neoplasms in mice and rats (Ken-dall et al., 1992; Hiramatsu et al., 1983). In mice administered 1,2-dimethylhydrazine (DMH), corn oil containing a large percentage of unsaturated fats markedly enhanced both the number and volume of neoplasms in the intestine as compared with mice fed an equivalent amount of beef tallow, which is composed primarily of saturated fats (Nutter et al., 1990). Also, as in the case of pancreatic and mammary carcinogenesis, dietary ω3 fatty acids inhibited while ω6 fatty acids enhanced the development of intestinal neoplasms induced by azoxymethane (Reddy and Sugie, 1988; Minoura et al., 1988).
The effect of carbohydrate administration on carcinogenesis seems to be of less impor- tance and to exert less influence than protein and lipid components of the diet. A high-carbohy- drate diet inhibited the development of hyperplastic nodules and neoplasms of the liver induced by 3′-methyl-4-dimethylaminoazobenzene (DEN) (Sato et al., 1984), whereas a high-sucrose diet—in comparison with an equivalent (65%) level of glucose—significantly enhanced the growth of preneoplastic hepatic foci initiated with DEN (Hei and Sudilovsky, 1985). Treatment of Sprague-Dawley rats with azoxymethane for 8 weeks resulted in a lower growth rate and less dysplasia of preneoplastic aberrant crypts when animals were administered cornstarch rather than sucrose in the diet (Caderni et al., 1994, 1997). In general, however, neither sucrose feeding nor systemic administration of sucrose enhanced spontaneous or induced carcinogenesis (cf. Kritchevsky, 1988).
Mechanisms of the Effects of Protein, Lipid, and Carbohydrate on Carcinogenesis. Mechanisms of the effects of the major dietary constituents on carcinogenesis differ where evi- dence for a mechanism exists. As indicated earlier (see above), inhibition of carcinogenesis by dietary protein may be related to its effect on xenobiotic metabolism to enhance the latter, caus- ing the formation of noncarcinogenic metabolites of the agent involved (cf. Mohr and Lewkowski, 1989). It has also been suggested that enhanced dietary protein levels may produce effects on the immune system that could enhance or inhibit the development of carcinogenesis, although how this might directly relate to effects of dietary protein on carcinogenesis is not clear (Mohr and Lewkowski, 1989). The enhancement effect of dietary sucrose, especially in hepato- carcinogenesis, may be related to the fact that dietary fructose significantly enhanced the growth of altered hepatic foci in livers of rats initiated with N-nitrosomorpholine (Enzmann et al.,1989). Since fructose is one of the two sugars in the sucrose molecule and glucose does not appear to have a similar enhancing effect (Hei and Sudilovsky, 1985), the effects of sucrose may be that of the fructose resulting from hydrolysis in the intestine.
In contrast to our lack of knowledge of the mechanisms involved in the effects of protein and carbohydrate on carcinogenesis, there is significant information available on potential mechanisms of the effect of dietary lipids on carcinogenesis. The differential effect of ω3 and ω6 fatty acids on mammary carcinogenesis was not reflected in the number of mammary neo- plasms induced by NMU in rats that exhibited an Ha-ras codon 12 mutation (Ronai et al., 1991). However, Lu et al. (1995) did find that a much larger proportion of mammary neoplasms in rats from this protocol exhibited a wild-type Ha-ras genotype when fed a high-fat diet in comparison with a low-fat diet. In colon carcinogenesis induced by azoxymethane, Singh et al. (1997) dem- onstrated that diets high in ω6 fatty acids exhibited an enhanced expression of the small ras-p21 proto-oncogene as well as an increased accumulation of the protein product in the cytoplasm in comparison with animals fed an ω3 fatty acid–rich diet. These studies suggest mechanism(s) by which high-fat diets of different composition may affect the process of initiation through altering the expression of wild-type and mutant ras proto-oncogenes. In this vein, dietary fat was also found to promote spontaneous carcinogenesis of the mammary gland in MMTV/v-Ha-ras trans- genic mice (DeWille et al., 1993), and studies by others (Etkind et al., 1995) have indicated a more rapid development of mammary neoplasms in mice containing an endogenous MMTV proviral DNA in their genome. That ω6 fatty acids act as promoting agents in the development of colon neoplasia in rats was seen from the demonstration of a significantly lower number of apoptotic cells/crypts in the proximal colon of rats compared with those receiving ω3 fatty acids in the diet (Chang et al., 1997). This further indicates that the promoting action of ω6 fatty acids is in part to inhibit apoptosis, presumably selectively in altered cells within the colonic mucosa. The fact that dietary cholesterol may enhance the development of intestinal neoplasms is proba- bly related to its metabolic conversion to bile salts, a number of which are known promoting agents for intestinal cancer in the rodent (Klurfeld et al., 1983). Corn oil diets, high in ω6 fatty acids, also enhance the metabolism of chemical carcinogens through a direct effect on xenobi- otic metabolism, leading, in this case, to an increased production of the proximate and ultimate forms of such agents (Newberne et al., 1979; Wade et al., 1982). These studies on the mecha- nisms of the effects of dietary protein, lipid, and carbohydrate on the stages of carcinogenesis further substantiate their effectiveness in influencing the stages of initiation and promotion, pri- marily the latter.