Vitamin A has long been associated with decreased cancer incidence in animals, especially of epithelial cancers. This concept was promoted by early insights into the effects of vitamin A depletion on animals. Wolbach16 published studies in 1925 demonstrating that vitamin A deficiency resulted in squamous metaplasia and correlated with increased respiratory and upper alimentary canal cancer. Studies with laboratory animals, prompted by these early observations, revealed the conversion of mucous secreting and/or ciliated epithelium into squamous keratinizing epithelial as the most conspicuous effects of vitamin A deficiency. These observations focused attention of vitamin A function on epithelia, and gave birth to the conclusion that the primary systemic function of vitamin A was to control epithelial differentiation.1 These observations also prompted multiple intervention studies with laboratory animals, which concluded that dosing with various retinoids (retinyl palmitate, retinyl acetate, RA) reduced the incidence of chemically induced (7,12-dimethyl-1,2-benzanthracene, benzo[a]pyrine, 3-meth-ylcholanthrene, N-methyl-N-nitrosourea) lesions in several epithelial targets, including the respiratory tract, the stomach, the vagina, and skin.17 Most likely because of poorly controlled protocols and/or different protocols that really were not comparable, the earliest animal studies reported mixed results of dietary vitamin A supplementation on the incidence of chemically caused carcinogenesis. Many studies, however, showed amazing reductions in cancer incidence using larger amounts of vitamin A in rodents exposed to chemical carcinogens. This stimulated great interest to pursue vitamin A and retinoid supplementation in cancer chemoprevention.
For example, as reviewed by De Luca,17 a study with golden hamsters painted a 5% solution of DMBA twice weekly for 13 weeks on check pouches and either fed a vitamin A-depleted diet (but not necessarily deficient) or gave a pair-fed group (400 nmol) retinyl palmitate per week. Oddly, the death rate of the vitamin A-supplemented animals was twice as high as the depleted animals, but the cancer rate in the surviving vitamin A-dosed animals was fivefold lower.18 A more clearly promising outcome was noted in a study in which hamsters were dosed with DMBA (10 mg/week) or BP (10 mg/week) with or without co-dosing with retinyl palmitate (100 mg/week = 190,000 nmol). The incidence of stomach cancer was 22 and 62% in the DMBA- and BP-dosed hamsters, respectively, and 0% in both carcinogen plus retinyl palmitate-dosed groups.19 Note, however, that the recommended amount of vitamin A in the AIN93 rodent diet is 4 nmol retinol/g diet (1 nmol = 1 IU). Sequential dosing studies, perhaps, provided a more realistic model, at least of chemoprevention "after the fact." Retinyl palmitate (19,000 nmol/week for life) decreased respiratory tumors 66% and squamous tumors 90% in hamsters dosed with BP intratracheally. Higher concentrations of vitamin A (7700 nmol) started 1 week after BP exposure increased the incidence of respiratory tumors 40%, but the tumor incidence decreased 20% when the hamsters were placed in laminar flow hoods.20
These studies were characterized by high doses of carcinogens applied topically or dosed intragastrically to rodents, higher doses of vitamin A or its esters (not synthetic retinoids), and were done using stock diets, with few exceptions. They examined a variety of tissues. These variables made it difficult to draw clear conclusions, but the studies did demonstrate that under certain circumstances pharmacological doses of vitamin A can prevent chemically induced epithelial cancer in some rodent tissues. This is an intriguing conclusion, but left unresolved whether a lifetime of enhanced vitamin A nutrition would help prevent cancer and whether diet has a confounding influence on vitamin A effects.
As more animal studies were published and studies were broadened to include RA and synthetic retinoids, clearer patterns emerged in laboratory animals.21 Perhaps the earliest, most consistent effects of vitamin A and other retinoids were obtained with the initiation/promotion model of skin carcinoma. Painting mice skin with a single dose of a carcinogen, such as DMBA, followed by one or more paintings with a promoter, such as croton oil or TPA, invariably produces many papillomas. Within 5 to 8 months, a few percent of the papillomas develop into carcinomas. Bollag22 showed that intraperitoneally dosing with high amounts of RA (100 mg/kg = >300,000 nmol) reduced papilloma volume >80% relative to controls. In a subsequent study, Bollag23 showed that oral dosing with RA (200 mg/kg every 2 weeks) during promotion also reduced the incidence of carcinomas by ~67%. These results presaged data, which consistently showed that retinol, retinyl palmitate, RA, and certain synthetic retinoids delayed appearance of skin papillomas, decreased their number and volume, and decreased their rate of progression into carcinomas in rabbits, mice, and rats. Mechanistic work on this phenomenon demonstrated that in this model retinoids function as antipromoters, specifically inhibiting induction of ornithine decarboxylase, an enzyme necessary for replication.24 These insights were extended by the demonstration that skin basal cell carcinomas in humans regressed upon topical RA treatment.23,25
Retinyl acetate, fed in rather high amounts (250 to 750 nmol/g food) to rats also has a chemopreventive effect with respect to mammary tumors induced by chemical carcinogens including BP, DMBA, and MNU.22 However, retinyl acetate causes hepatoxicity at effective doses. Fortunately, the synthetic retinoid N-(4-hydroxyphenyl)retinamide (fenretinide) also is effective, and tends to concentrate in the breast.26 Curiously, dietary 13-ds-RA is not an effective chemopreventive of MNU-induced breast cancer. These results illustrate two issues concerning retinoids as chemopreventives: (1) doses/compounds effective against cancer in one tissue may have toxic effects in other tissues; (2) chemopreventive activity in test systems in vitro often does not predict actions in the intact mammal. Obviously, the pharmacology of a retinoid (tissue distribution, metabolism rate, etc.) has great impact on its chemopreventive action and toxicity in vivo.
The conclusion that specific retinoids act on cancer of specific tissues in experimental animals is extended by the observation of the impact of retinoids on bladder cancer.27 For example, 13-cis-RA reduces the incidence of transitional cell carcinoma in rats initiated by N-butyl-N-(4-hydroxybutyl)nitrosamine, a bladder-specific carcinogen. Recall that dietary 13-cis-RA does not prevent MNU-induced breast cancer in rats. It is also important to note that transitional cell carcinoma accounts for the major type of human bladder cancer, rather than squamous cell carcinoma.
Lasnitski28 developed organ culture of mouse prostate and demonstrated with it that RA and synthetic retinoids could inhibit or reverse MC-induced neoplastic changes. This, of course, generated a great deal of interest in the potential use of retinoids to prevent and/or treat prostate cancer. A subsequent study demonstrated that fenretinide, a retinoid active against carcinogen-induced breast cancer, caused a 76% decrease in prostate cancer in rats treated with the carcinogen MNU
and testosterone.29 In addition, RA and other retinoids at higher doses inhibited the growth of established prostate cancer cell lines and induced apoptosis.30 Perhaps not surprisingly, given previous experience with retinoids and with established cell lines, retinoid effects on cultured prostate cancer cells are retinoid-and cell line-dependent. RA, for example, is effective with LNCaP cells, WPEI-NB11, and WEI-NB14 cells, but not with DU-145 or PC-3 cells. Fenretinide, however, was more effective than RA in the WPEI cell lines. Alas, these are established cancer cell lines. Regardless of these reports of success, and others like them,27 a very recent review of the literature concluded that "no evidence" exists that vitamin A or synthetic retinoids can be used as prostate cancer chemo-preventive agents, because of either toxicity or lack of efficacy.31 Consistent with this conclusion, a review of clinical studies with RA concluded that only modest effects were obtained in inducing apoptosis of prostate cancer.30
To summarize: (1) there tends to be a degree of species, retinoid, carcinogen, and tissue specificity associated with the effects of retinoids on cancer chemo-prevention in experimental animals; (2) in vitro models, especially cancer cells in culture, are not necessarily good predictors of retinoid chemopreventive effects in intact animals; (3) doses of retinoids that act chemopreventively at one site in intact animals may be toxic to other tissues. Nevertheless, retinoids still hold tremendous potential as cancer chemopreventive or therapeutic agents. Use of synthetic retinoids, as exemplified by fenretinide, may circumvent toxicity associated with naturally occurring retinoids.
Was this article helpful?