While the bulk of the diet is made up of protein, carbohydrate, and lipid as well as significant nondigestible components in crude diets, a number of so-called micronutrients, vitamins, miner- als, and related substituents are necessary for the survival and growth of the multicellular organ- ism (Hunt and Groff, 1990). Table 8.8 lists most of the macro- and micronutrients that serve as exogenous modifiers in carcinogenesis in experimental animals. A number of these in relation to the human are discussed later in the text (Chapter 11). Since the macronutrients in relation to carcinogenesis have already been discussed, this section emphasizes the effect of micronutrients as exogenous modifiers of carcinogenesis.
In the strict sense, vitamins are micronutrients that are essential for the nutrition and via- bility of a multicellular organism and that cannot be synthesized in amounts sufficient for growth and/or normal function of the organism. The effects of the listed vitamins and minerals on the carcinogenic process are primarily to inhibit the process of carcinogenesis when present in sufficient or more than sufficient quantities. However, when the organism is deficient in one or more of these micronutrients, the carcinogenic process in one or more tissues is enhanced. It will also be noted that, as with macronutrients, the stages of carcinogenesis at which micronutri- ents exert their primary effects are, in most cases, initiation and/or promotion.
Retinoids. Over the past two decades, a remarkable effect of excess vitamin A and/or its derivatives on the development of epidermal, pulmonary, and other types of neoplasia has been demonstrated (cf. Lotan, 1980; Sporn and Roberts, 1983; Birt, 1986). In early studies in the hamster given carcinogenic hydrocarbons intranasally (a regimen that induces squamous cell carcinoma of the lung), carcinogenesis could be completely inhibited by the administration of high levels of vitamin A. The feeding of a diet containing high levels of vitamin A inhibited skin carcinogenesis by 7,12-dimethylbenz[a]anthracene in mice (Bollag, 1972). In addition, when vi- tamin A was applied directly to the skin after carcinogenesis was well under way, the morpho- logical type of neoplasm produced was altered. This latter effect was inhibited by actinomycin D (Prutkin, 1971). These earlier studies have now been extended to a variety of other tissues and the development of neoplasms therein. Table 8.9 from Birt (1986) lists many of the more dra- matic effects of retinoids on experimental carcinogenesis in several organs. Some of the most dramatic effects, as noted in this table, can be seen in the inhibition of carcinogenesis in the mouse skin by chemicals through the use of a variety of retinoids (Table 8.9; DeLuca et al.,1996). Retinoic acid was found to enhance or inhibit the development of papillomas in mouse epidermis during the two-stage process of carcinogenesis (Hennings et al., 1982; Verma et al.,1982). The differences in this effect are not clear, although Verma et al. (1982) demonstrated that
when DMBA was given in multiple doses without TPA, retinoic acid failed to inhibit the devel- opment of papillomas. Several different retinoid analogs, administered in the diet after initiation with DMBA, acted to promote the development of papillomas (McCormick et al., 1987). Inter- estingly, mice on a severe vitamin A–deficient diet failed to develop visible tumors in the two- stage carcinogenesis protocol, but supplementation with retinoic acid after 12 weeks on the defi- cient diet resulted in the rapid development of papillomas by week 22 (DeLuca et al., 1989). Thus, the effects of retinoids on epidermal carcinogenesis in the mouse depend on the format of the carcinogenesis experiments and the structure of the retinoid employed. Furthermore, retin- oids given orally inhibit the development of the stage of progression in multistage epidermal carcinogenesis in the mouse (DeLuca et al., 1994).
One of the most striking demonstrations of an effect of vitamin A on carcinogenesis in the rat is the inhibition of N-methyl-N-nitrosourea–induced bladder carcinogenesis by the synthetic retinoid 13-cis-retinoic acid (Sporn et al., 1977). In this case, administration of the retinoic acid derivative inhibited the incidence and extent of bladder cancer development even when its ad- ministration was initiated after cessation of carcinogen dosing. Administration of 13-cis- retinoate to animals given dimethylhydrazine inhibited the production of colonic neoplasms (Newberne and Suphakarn, 1977). Narisawa et al. (1996) have extended these investigations by
demonstrating the inhibition of the development of preneoplastic aberrant crypt foci in the intes- tine of rats following MNU administration. Later studies by Mack et al. (1990) and Sarkar et al. (1995) have further substantiated the inhibition of 3′-methyl-4-dimethylaminoazobenzene–in- duced hepatic neoplasia in the rat. Furthermore, the feeding of β-carotene in a multistage model of hepatocarcinogenesis in the rat significantly inhibited the development of preneoplastic al- tered hepatic foci (Moreno et al., 1995). In addition to the inhibitory effects of retinoids on DMBA- and MNU-induced mammary neoplasia in the rat, dietary retinyl acetate inhibited the development of mammary neoplasms induced by estrogen in female rats (Holtzman, 1988). The inhibition of MNU-induced mammary carcinogenesis by retinyl acetate may be due to preven- tion of the progression of very early neoplastic lesions (Thompson and Becci, 1979). Thus, one may conclude that, in general, retinoids tend to inhibit carcinogenesis of a variety of tissues in rats and mice but to a lesser extent in hamsters. Most inhibitory effects of retinoids occur during the stage of promotion, in part by an inhibition of DNA synthesis induced by the promoting agent (Paulsen, 1984; Gendimenico et al., 1989). However, Dogra et al. (1984) reported that the extent of binding of benzo[a]pyrene to lung DNA in vivo was about 2.5-fold higher in vitamin A-deficient rats than in those receiving a normal amount of vitamin A. Other mechanisms by which retinoids exert their effects may be directed by an alteration of the expression of specific genes important in the stage of promotion, such as ornithine decarboxylase (Chapter 7; Verma et al., 1979). Retinoids are also involved in the induction of transglutaminase in the epidermis (Yuspa, 1982). This enzyme has been implicated in the process of apoptosis in a number of cell populations (Fesus et al., 1996). Since retinoids are known to exert their effect on gene expres- sion via the retinoid receptors (Chapter 3), it is likely that many of the effects of retinoids may be mediated through such a mechanism. Furthermore, there are specific antagonisms between retin- oic acid receptors and other transcriptional activators such as c-jun and c-fos (Yang-Yen et al.,1991). An enhancement of the immune system by retinoids may also serve as an indirect mech- anism for the inhibition of the development of preneoplastic and neoplastic lesions (Watson and Moriguchi, 1985; Malkovsky et al., 1983).
Vitamins C and E. Several vitamins and at least one mineral, selenium, play significant roles in the regulation of the metabolism of active oxygen radicals and related molecules within cells. Vitamins C (ascorbate) and E (α tocopherol) are two of these, while retinoids also exhibit antioxidant effects both in vivo and in vitro (cf. Krinsky, 1993). The structures of these three vitamins and their relatives may be noted in Figure 8.8. Vitamin C is water soluble, while caroti- noids and vitamin E are predominantly lipid-soluble. The action of vitamin E as an antioxidant can be seen in Figure 8.9. In this figure, the interaction between vitamin C (ascorbate) and the reaction of tocopherols with lipid peroxyl radicals can be noted: the tocopheroxyl radical inter- rupts the radical chain reaction, thereby preventing the chain propagation of lipid peroxidation (Sies et al., 1992). Vitamin C is water-soluble and can also react with glutathione to regenerate the vitamin, while reduced glutathione is reformed by its reductase. In this way the antioxidant system exerts its effect through these two vitamins in both an aqueous and a lipid (membrane) environment (Meister, 1994). While vitamin C has been shown to inhibit the development of a number of different histogenetic types of neoplasms—including those of the colon, lung, kidney, and epidermis (Birt, 1986; Chen et al., 1988)—high levels of sodium ascorbate act to enhance the development of colon and bladder neoplasms in rodents (cf. Block and Schwarz, 1994). This latter effect is perhaps not surprising, since ascorbic acid has been known for almost two de- cades to induce mutations in several different life forms (cf. Shamberger, 1984). On the other hand, such investigations have also demonstrated the antimutagenic effects of this vitamin (cf. Shamberger, 1984). The exact mechanisms for these divergent effects of the same molecule are not totally clear at the present time but are probably related to a combination of dose and its effectiveness as both an antioxidant and oxidant.
While administration of vitamin E, either by diet or topically, inhibits the development of skin, forestomach, colon, and mammary neoplasms in rodents (Birt, 1986), Some questions have been raised and discrepancies noted as to the effectiveness of this vitamin. Ura et al. (1987) re- ported an inhibition of the “selective” action of 2-acetylaminofluorene on DEN induction of al- tered hepatic foci in rat liver by dietary vitamin E, but this regimen had no effect on later stages of the process, including progression. In contrast, Lii et al. (1997) were unable to find any effect of vitamin E on the development of preneoplastic hepatic foci resulting from neonatal initiation by DEN with subsequent phenobarbital promotion. Furthermore, Kolaja and Klaunig (1997) re-
Figure 8.8 Structures of vitamins with antioxidant activity. The tocopherol series has the phytyl side chain (R3). The radical function in the tocopheroxyl radical occurs at the hydroxyl of the chromane moiety. (Adapted from Sies and Stahl, 1995, with permission of authors and publisher.)
ported an enhancement of the growth of hepatic focal lesions in mice by both vitamin E supple- mentation and deficiency. In the azaserine induction of preneoplastic foci in rat pancreas, while the growth of acidophilic foci was inhibited in rats administered excess vitamins A and C, vita- min E exerted an inhibitory effect only on basophilic but not acidophilic foci (Woutersen and van Garderen-Hoetmer, 1988a). However, vitamin E as well as β-carotene inhibited the develop- ment of preneoplastic aberrant crypt foci in the colon of rats fed a high-fat, low-fiber diet (Shivapurkar et al., 1995). While the mechanisms of these discrepancies are not clear, both of these vitamins, C and E, may potentially inhibit “endogenous” carcinogenesis by inhibiting nit- rosation of secondary amines ingested in the diet or produced in vivo (cf. Birt, 1986). In addi- tion, both vitamins can inhibit mutagenesis, in vitro carcinogenesis, and the binding of active carcinogenic metabolites to cellular DNA as well as chromosomal breakage induced by chemi- cal carcinogens (cf. Chen et al., 1988).
Figure 8.9 Action of vitamin E as an antioxidant in relation to lipid peroxides and subsequent oxidation by ascorbate. (Adapted from Sies et al., 1992, with permission of authors and publisher.)
Miscellaneous Vitamin Effects in Carcinogenesis. One of the earliest effects of vitamins on carcinogenesis was reported by Morris and Robertson (1943), who observed that develop- ment of spontaneous mammary neoplasms was markedly depressed in C3H mice fed a ribofla- vin-deficient diet. This was in contrast to later studies demonstrating that riboflavin deficiency was accompanied by an increase in the effectiveness of carcinogenic azo dyes to induce hepato- mas. This latter effect was probably due to the markedly diminished azo dye reductase activity seen in livers of riboflavin-deficient rats. Riboflavin is the cofactor of this enzyme, which me- tabolizes azo dyes to noncarcinogenic intermediates (cf. Miller and Miller, 1953). Similarly, folate deficiency, in both animal and human studies, has been associated with an increased risk for the development of neoplasms of the colon, cervix, lung, and brain (cf. Kim et al., 1997), but the role of folate in neoplastic development is not yet entirely clear (Glynn and Albanes, 1994). Cravo et al. (1992) reported the enhancement of the development of colonic neoplasia in rats treated with dimethylhydrazine that were folate deficient as compared with animals on control diets. Other studies have demonstrated that folate deficiency produces significant alterations in DNA and chromosome structure, including breakage, in both animals and humans in vivo (MacGregor et al., 1990; Blount et al., 1997). The active form of vitamin D, 1α,25-dihydroxyvi- tamin D3, inhibits the proliferation of a variety of different neoplastic cells, both in vivo and in
vitro (cf. Reichel et al., 1989). Administration of an analog of 1α,25-dihydroxyvitamin D3 in-
hibited the development of colonic neoplasms induced by azoxymethane in rats when adminis-
tered in the diet (Wali et al., 1995). The effects of this vitamin may be related to its regulation of the metabolism of calcium, whose effectiveness as an exogenous modifier is discussed below.
Minerals Effective as Exogenous Modifiers of Carcinogenesis. Of the minerals known as exogenous modifiers of carcinogenesis, calcium is probably the most widespread in nature and exhibits the most varied functions in this category. The calcium ion is essential for cell repli- cation and survival (Clapham, 1995). Within the cell, there are a large number of different cal- cium-binding proteins (Villereal and Palfrey, 1989) whose functions are modulated by calcium levels within the cell. These functions include muscle contraction, nerve transmission, mem- brane transport, signal transduction, and cell cycle progression (Clapham, 1995; Takuwa et al.,
1995). The reader is referred to the reviews cited for more details on the various functions of calcium and its binding proteins. Since regulation of the cell cycle is very important in carcino- genesis and neoplasia, it is not surprising to find that normal cells, and to a lesser extent neoplas- tic cells, exhibit varying degrees of inhibition of replication (Whitfield, 1992). Calcium added to the diet dramatically inhibited the development of neoplasms in rats administered 1,2-dimethyl- hydrazine in the presence or absence of vitamin E (McIntosh, 1992). In a separate study, Pence et al. (1995) demonstrated that calcium could effectively inhibit cholic acid promotion of azoxymethane-induced colon neoplasms. This finding is also reflected in the apparent inhibition of recurrences of adenomatous polyps in the human after polypectomy and after colorectal sur- gery for colorectal cancer (Duris et al., 1996). Thus, excess calcium may downregulate the stage of promotion in carcinogenesis of the colon and potentially other tissues through its regulatory effects on cell division and signal transduction.
Zinc is another mineral widespread in living systems and involved in numerous functions, including DNA replication, transcription, and a variety of peptidases and dehydrogenases (Wu and Wu, 1987; Coleman, 1992). In concert with these functions, especially those related to DNA synthesis and transcription, zinc deficiency inhibits the growth of a number of transplanted neo- plasms in animals in vivo (cf. Walsh et al., 1994). In contrast, Fong et al. (1984) had previously reported that zinc deficiency enhanced the incidence of esophageal neoplasms induced by methylbenzylnitrosamine. This may be related to a direct effect of the deficiency in damaging the esophageal epithelium, with subsequent enhancement of replication of some epithelial cells. Very high levels of dietary zinc inhibited carcinogenesis by DMBA in hamsters and by azo dyes in livers of rats (cf. Walsh et al., 1994). Zinc deficiency induces thymic atrophy, which may be secondary to glucocorticoid-mediated apoptosis of thymocytes, thus potentially affecting the carcinogenic process indirectly (Fraker et al., 1995).
Another mineral micronutrient that has exhibited significant effects as an exogenous mod- ifier of carcinogenesis is selenium. In living organisms, selenium is almost always associated with proteins which, in animal tissues, are almost exclusively in the form of selenocysteine (Burk and Hill, 1993). Selenocysteine has been termed the 21st amino acid in ribosome-medi- ated protein synthesis in both prokaryotes and eukaryotes. Selenocysteine is formed cotransla- tionally on a seryl tRNA with a UCA anticodon. The complementary mRNA codon is UGA. This codon is also a “stop” codon signifying the end of translation of an mRNA. It appears that the use of the selenocysteine tRNA in preference to the termination of translation depends on adjacent sequences within the mRNA (Stadtman, 1996; Böck et al., 1991). About a half dozen selenoproteins are known in mammals (Birk and Hill, 1993; Stadtman, 1996), but the protein that appears to be most significant in the function of selenium as an exogenous modifier of car- cinogenesis is its presence in glutathione peroxidase. This enzyme participates in the elimination of potentially harmful active oxygen radicals and their products by mechanisms such as depicted in Figure 8.10.
The effects of selenium administration and selenium deficiency on carcinogenesis in several different types of tissues are seen in Table 8.10. Excess dietary inorganic selenium inhib- its colon carcinogenesis induced in rats by bis(2-oxopropyl)nitrosamine (Birt et al., 1982), 1,2-dimethylhydrazine (Jacobs, 1983), and azoxymethane (Reddy et al., 1988). This inhibition can also occur with organic selenium compounds (Reddy et al., 1987), with the effects of this micronutrient occurring primarily during the stage of promotion (Reddy et al., 1988). In con- trast, selenium deficiency either had no effect (Pence and Buddingh, 1985) or inhibited the inci- dence and development of colon tumors in rats after azoxymethane administration (Reddy and Tanaka, 1986).
In this last study, no interaction with vitamin E in the diet was seen. However, Woutersen and van Garderen-Hoetmer (1988b) demonstrated an inhibitory effect of dietary selenium on the
Figure 8.10 The glutathione redox cycle. ROOH/ROH: peroxy radical/alcohol. H2O2: hydrogen perox- ide. GSH-Px: glutathione peroxidase. GSH/GSSG: reduced/oxidized glutathione. GSSG-R: glutathione re- ductase. NADPH/NADP: reduced/oxidized nicotinamide adenine dinucleotide phosphate. G-6-PD: glucose-6-phosphate dehydrogenase. G-6-P: glucose-6-phosphate. HMP: hexose monophosphate.
development of preneoplastic lesions in the pancreas of rats administered azaserine and in ham- sters administered N-nitrosobis(2-oxopropyl)amine when animals were fed diets high in satu- rated fat. In rats initiated with aflatoxin B1, those fed a diet deficient in selenium exhibited a significantly enhanced growth of preneoplastic foci in their livers as compared with livers of rats fed a normal amount of selenium. Rats fed a high-fat diet exhibited an inhibition of initiation by the administration of excess selenium (Baldwin and Parker, 1987). Excess organic and inorganic selenium also inhibited the development of preneoplastic foci in livers of animals administered either azoxymethane or aflatoxin B1 respectively (Sugie et al., 1989; Milks et al., 1985), as well as a necrogenic dose of DEN followed by a partial hepatectomy (LeBoeuf et al., 1985). How- ever, when initiation was carried out with a nonnecrogenic dose of DEN followed by promotion with phenobarbital, the addition of inorganic selenium to the diet actually enhanced the growth of preneoplastic lesions. In contrast, Dorado et al. (1985) were unable to demonstrate any effect of added dietary selenium in the format of the latter experiment. Reportedly, both inorganic and
organic forms of selenium administered in excess inhibited the development of mammary tu- mors in rats induced by either DMBA or MNU (Welsch et al., 1981; Ip and White, 1987), al- though a more recent study by Lane et al. (1990) did not find an inhibition by selenomethionine or selenocysteine. However, excess inorganic selenium did inhibit DMBA-induced mammary cancer development in mice. In the study by Ip and White (1987), vitamin E potentiated the effect of the inorganic but not the organic forms of selenium. In another study, Ip reported that administration of vitamin C in a similar experiment nullified the protective effect of inorganic selenite but had no effect on the inhibitory effect of selenomethionine on mammary carcinogen- esis (Ip, 1986). Deficiencies of selenium together with a deficiency of vitamin E enhanced mam- mary carcinogenesis by MNU, but deficiencies of either single nutrient had no effect in altering the carcinogenic response (Thompson, 1991). In addition, Lijinsky et al. (1989) were unable to find an effect of dietary selenium on the induction of esophageal and bladder neoplasms in rats by two different nitrosamines.
The principal mechanism whereby selenium probably exerts its inhibitory effect on car- cinogenesis is through its role as an essential component of the enzyme glutathione peroxidase, which is an essential part of the antioxidant defense system as described previously (Figures 8.8 and 8.9). The enzymatic activity of glutathione peroxidase is directly related to the concentration of selenium in blood (cf. Hocman, 1988a). Several of the studies noted above found an increased activity of glutathione peroxidase in several tissues in animals fed high levels of selenium in the diet (Birt et al., 1982; Reddy et al., 1987; Lane et al., 1990). In addition, feeding a combination of vitamin E and selenium to rats administered DMBA to induce mammary neoplasms inhibited lipid peroxidation occurring immediately after DMBA administration. This inhibition was di- rectly correlated with an inhibition of mammary carcinogenesis (Takada et al., 1992). In addition to the effect of selenium on oxidant reactions, Medina and Morrison (1988) suggested several other possible mechanisms, including an alteration of carcinogen-DNA interactions, mainte- nance of intracellular glutathione levels, and effects on DNA and protein synthesis as well as gene expression. While a selenium-deficient diet reduced the level of cytochrome P450IA1, it did not affect its inducibility (Gairola and Chow, 1982). Conversely, administration of inorganic selenite to rats administered 1,2-dimethylhydrazine in vivo inhibited DNA alkylation signifi- cantly both in liver and in colon (Harbach and Swenberg, 1981). Dipple et al. (1986) found a similar inhibition of DMBA-deoxyadenosine adducts in fetal mouse cell cultures when adminis- tered 1 hour prior to but not 3 hours after the addition of the carcinogen. This study was some- what at odds with an earlier investigation by Ip and Daniel (1985), who were unable to demonstrate any significant effect of DMBA-DNA binding in vivo in liver and mammary gland when excess inorganic selenite was administered. However, more recent investigations by Liu et al. (1991) showed a marked inhibition of total DMBA-DNA adducts after selenite supplementa- tion of the diet for only 2 weeks. In addition, hepatic DNA of selenium-deficient rats exhibited a significant number of single-strand breaks shortly after injection of 2-acetylaminofluorene, but no effect of dietary selenium on DNA repair could be demonstrated (Wortzman et al., 1980). Thus, while most evidence argues that selenium exerts its inhibitory effect at the promotion stage of carcinogenesis, there is evidence that the initiation stage may also be affected by this mineral.
Other Minerals as Exogenous Modifiers of Carcinogenesis. Another relatively common mineral in living systems is iron and its salts. Feeding an excess level of iron in the diet resulted in increased mammary tumor incidence and lowered natural killer cell activity in rats (Spear and Sherman, 1992). Smith et al. (1993) also reported that injections of iron-dextran producing an iron overload caused a marked enhancement of the development of hepatic nodules in animals administered hexachlorobenzene, a hepatic carcinogen. Iron overload is known to induce tissue damage and increase the formation of reactive oxygen species (Okada, 1996). Iron deficiency enhanced lipid peroxidation in dimethylhydrazine-fed rats. This effect was also accompanied by an increase in liver glutathione peroxidase (Rao and Jagadeesan, 1996). Although these investi- gators did not carry the studies through to neoplasms, their findings, coupled with those cited above, argue that iron may serve as an exogenous modifier of carcinogenesis by altering lipid peroxidation and active oxygen formation. Deficiency of another mineral, copper, enhanced the formation of colonic neoplasms after dimethylhydrazine administration to rats (Greene et al.,
1987; DiSilvestro et al., 1992). The mechanism for this effect is not clear at present.