Selenomethionine and α-Tocopherol Do Not Inhibit Prostate Carcinogenesis in the Testosterone plus Estradiol–Treated NBL Rat Model Free

Previous studies with selenium and/or vitamin E in prostate carcinogenesis animal models have been negative, but these models may not involve oxidative stress mechanisms. In this study, we examined the potential of selenomethionine and α-tocopherol to modulate prostate cancer development in the testosterone plus estradiol–treated NBL rat, a model that does involve sex hormone–induced oxidative stress mechanisms and prostatic inflammation. One week following the implantation with hormone-filled Silastic implants, rats were fed diets containing l-selenomethionine (1.5 or 3.0 mg/kg), DL-α-tocopherol acetate (2,000 or 4,000 mg/kg), or a natural ingredient control diet (NIH-07). The development of prostate carcinomas was not affected by dietary treatment with either agent. Food intake, body weight, and mortality were also not affected. The high dose of selenomethionine reduced the severity of epithelial dysplasia in the lateral prostate that was not associated with inflammation, and α-tocopherol reduced in a dose-related fashion the incidence of marked inflammation and marked epithelial dysplasia in the lateral prostate, regardless of whether these lesions were associated with inflammation. α-Tocopherol significantly increased the incidence of adenocarcinomas of the mammary glands at both dietary concentrations. Collectively, our findings suggest that selenomethionine and α-tocopherol supplementation does not prevent prostate cancer in rats fed diets with nutritionally adequate levels of selenium and vitamin E. Importantly, the results of the current animal studies and those reported previously were fully predictive of the outcome of the Selenium and Vitamin E Cancer Prevention Trial. Cancer Prev Res; 3(3); 371–80

Prostate cancer is the most common noncutaneous malignancy and the second leading cause of cancer-related deaths in men in Western countries (1). Because the etiology of prostate cancer is not well understood, few risk factors offer opportunities for primary prevention of this malignancy. Therefore, chemoprevention is an attractive and potentially powerful approach to prostate cancer prevention that can be mechanism based (2, 3). Selenium compounds and tocopherols are classes of chemoprevention agents that have shown promise in this respect (4, 5).

A protective effect of selenium and vitamin E is biologically plausible because of their role as antioxidants (6, 7) and selenium has several other properties that may confer anticarcinogenic activity (4). Antioxidants protect cells from oxidative DNA damage, which can cause mutations and subsequent carcinogenesis (8). Inflammatory processes known to involve oxidative damage are suspected to be associated with human prostate cancer (9) and the presence of the oxidized DNA base 8-hydroxy-2'-deoxyguanosine has been observed in human prostate tissue (10).

Secondary analysis of data of the Nutritional Cancer Prevention trial with selenized yeast in men at increased risk for skin cancer yielded suggestive evidence that supplementation with this type of selenium might be protective against prostate cancer (11, 12). A similar secondary analysis of the results of the ATBC trial with α-tocopherol (with or without β-carotene) in male Finnish smokers indicated a potential protective effect of vitamin E (13). These observations formed the basis for the rationale of the Selenium and Vitamin E Cancer Prevention Trial (SELECT) trial in which the potential to prevent prostate cancer was tested of selenomethionine at 200 μg/d and α-tocopherol at 400 IU/d, alone or in combination (14). However, the study was recently terminated after an interim analysis showed that neither selenium nor vitamin E prevented prostate cancer in this study and that there were possible adverse effects; a nonsignificant increase in type II diabetes was observed in the selenium-supplemented groups (15). Interestingly, there was a statistically nonsignificant increased risk of prostate cancer in the vitamin E group, but not in the selenium and the selenium plus vitamin E groups.

Previous studies with selenium (as selenomethionine or selenized yeast) and/or α-tocopherol in animal models of prostate carcinogenesis have been negative, including experiments with tumor induction models (16–19). However, none of these models are known to involve oxidative stress mechanisms, which may explain these null findings. In this study, we examined the potential of selenomethionine and α-tocopherol to modulate prostate cancer development in an animal model that uniquely does involve sex hormone–induced oxidative stress mechanisms and prostatic inflammation, the testosterone plus estradiol–treated NBL rat (20–22).

Male NBL (Noble) rats (NBL/CrCrlBR) were obtained at 7 to 8 wk of age from Charles River. Rats were held in quarantine for 2 wk before the initiation of treatment. They were housed two to a cage in solid bottom cages and were fed the NIH-07 diet (Harlan Teklad) during quarantine. The protocol for these experiments was approved by the Institutional Animal Care and Use Committee in accordance with the U.S. Public Health Service Policy on Humane Care and Use of Laboratory Animals.

After randomization into experimental groups (Table 1), the rats were subjected to the following treatment protocol for induction of prostate cancer (Table 1; ref. 20). All rats received two s.c. Silastic tubing implants (Dow Corning, ID 0.078 inch; absorbance, 0.125 inch) tightly packed over a length of 2 cm with containing crystalline testosterone (Steraloids) and one such implant packed over a length of 1 cm with crystalline 17β-estradiol (Steraloids). The implants were sealed with medical grade silicone sealant (RTV 108, General Electric) placed on top of small glass beads inserted in the tubing after filling with hormone. Immediately before implantation, the implants were rinsed in 70% ethanol and then soaked for 48 h in sterile saline at 37°C, which was replaced twice daily. The Silastic implants were implanted under the skin between the scapulae while the rats were anesthetized and they were replaced once after 26 wk. This treatment increases circulating estradiol levels by 5- to 10-fold in NBL rats while maintaining testosterone at physiologic levels (∼3 ng/ mL; ref. 23).

l-Selenomethionine (Sigma) was mixed into the diet (NIH-07; Harlan Teklad) using a Blend-Master Model B Lab Blender (Patterson-Kelly). A vehicle of D(+)sucrose (Sigma) was used at a dietary concentration of 10 g/kg diet (1% w/w) to achieve final selenium concentrations of 1.5 or 3.0 mg/kg diet. The control diets contained the same amount of sucrose, but no selenomethionine. DL-α-tocopherol acetate (all rac-α-tocopherol acetate; Sigma) was mixed into the diet using corn oil (Sigma) as vehicle at a dietary concentration of 38 g/kg diet (3.8% w/w) to achieve final DL-α-tocopherol acetate concentrations of 2,000 or 4,000 mg/kg diet; the control diet contained the same amount for corn oil, but without tocopherol. The corn oil added 7 to 8 mg α-tocopherol and 35 to 38 mg γ-tocopherol per kilogram to the NIH-07 diet, which contained 22 mg/kg DL-α-tocopheryl acetate and 0.31 mg/kg selenium. The diet concentrations and form of selenium and vitamin E were selected because they were used in previous experiments with another rat model of prostate carcinogenesis (16, 17) and because l-selenomethionine and DL-α-tocopheryl acetate were used in the SELECT trial (14, 15). These dose levels of selenomethionine and α-tocopherol were approximately 40% and 80% of the maximally tolerated doses, as identified previously in Wistar rats, which increased mean plasma selenium levels in the selenomethionine experiment by 9% to 17% from 560 to 625 ng/mL in the control group to 656 to 680 ng/mL in the high selenomethionine dose group, and plasma α-tocopherol levels from ∼5 μg/mL in the control group to 15 μg/mL in either supplemented group.⁴

A first group of 90 NBL rats was randomized into three treatment groups of 30 rats each, receiving no supplement (controls) or diet containing 1.5 or 3.0 mg/kg selenium (Table 1). A second group of 90 NBL rats was also randomized into three groups of 30 rats receiving no supplement or diet containing 2,000 or 4,000 mg/kg α-tocopherol (Table 1). Animals had ad libitum access to food and tap water. The animals were fed the experimental diets starting 1 wk after implantation of the hormone-containing Silastic implants. The experimental diets were prepared every 2 wk and stored at 4°C; fresh food was provided to the animals on Mondays, Wednesdays, and Fridays. Animals were observed daily and weighed weekly for the first 2 mo and monthly thereafter to assess their general health. Food intake per cage over 1 wk was measured after the rats had been fed the experimental diets for 4, 8, 12, 24, and 48 wk. Prostate tumor development was assessed by weekly abdominal palpation starting 26 wk into the hormone treatment period. Rats identified as moribund were euthanized by barbiturate overdose and exsanguination, and these rats and those found dead were necropsied immediately. At 54 wk after the start of hormone treatment, all surviving animals were euthanized and necropsied. At necropsy, the accessory sex glands were excised en bloc together with the urinary bladder. Accessory sex glands, pituitaries, and all grossly observed lesions in other organs were fixed in 10% neutral buffered formalin.

After fixation, the ventral prostate, dorsolateral prostate, and anterior prostate plus seminal vesicles were dissected. From the ventral prostate and grossly observed tumor masses, one section was made, and from all other accessory sex gland tissues, six step sections were prepared at 250 to 300 micrometer intervals, which were stained with H&E (24). All prostate lobes and other accessory sex glands were evaluated histopathologically and the presence, type, and size of all lesions were scored, using previously published criteria (20, 23, 24). Severity and extent of dysplastic and inflammatory lesions were scored semiquantitatively as absent, slight, moderate, or marked by a single pathologist (MCB), using previously published criteria (20, 23). Mammary tumors were classified according to Russo et al. (25).

The lesion incidence data presented reflect the presence of a particular lesion identified in a particular tissue in each animal. Differences in lesion incidence in the accessory sex glands were analyzed using Fisher's exact test (two group comparisons) and χ² analysis (three group comparisons and analysis for linear trend). Comparisons of animal survival were made using log-rank analysis. Analysis of body weight and water consumption data was done using ANOVA, with post hoc comparisons using Dunnett's multiple comparison test.

NBL rats in this study died on average of 41 to 46 weeks after treatment with testosterone plus estradiol had commenced (Tables 2 and 4). This hormonal treatment resulted in a high incidence of multifocal adenocarcinomas originating from the ducts of the dorsolateral and anterior prostate in the periurethral region, confirming previous observations (20). In the selenomethionine experiment, the prostate carcinoma incidence ranged from 93% to 97%, and in the α-tocopherol experiment, it ranged from 83% to 90% (Tables 2 and 4). In the vast majority of animals, these tumors were multifocal with 60% to 80% of tumor-bearing animals having more than two foci of prostate cancer (Tables 2 and 4). Carcinomas of the anterior prostate ducts outside the periurethral area were also frequent (53-73%), but carcinomas originating in the glandular portions of the dorsolateral and anterior prostate lobes and in the seminal vesicle were very rare (Tables 2 and 4). Lesions classified as carcinoma in situ (morphologically identical to carcinomas but without clear invasive growth) were occasionally observed in the ducts and glandular portions of the anterior prostate and seminal vesicles (Tables 2 and 4). All animals had focal dysplastic and inflammatory lesions in the lateral prostate lobes (Tables 2 and 4) and pituitary tumors (Tables 3 and 5). The spectrum and morphology of these neoplastic and nonneoplastic lesions (data not shown) was comparable with that observed previously in this animal model (20–23). Mammary tumors were observed in all treatment groups of both experiments and a few neoplasms were observed in sites other than the accessory sex organs or mammary glands (Tables 3 and 5).

Feeding testosterone plus estradiol–treated NBL rats with selenomethionine mixed into the diet at 1.5 and 3.0 mg/kg did not affect prostate carcinoma incidence and multiplicity (Table 2). Selenomethionine at these doses did not affect body weight (data not shown) or mortality (Table 2). Rats in the high dose group tended to eat 3% to 9% more food than controls, which was statistically significant at 4 weeks but not at 8 and 24 weeks (Table 3).

The incidence of inflammation and inflammation-associated epithelial dysplasia in the lateral prostate caused by the testosterone plus estradiol treatment was also not affected by dietary selenomethionine (Table 2). The high dose of selenomethionine significantly reduced the severity of epithelial dysplasia in the lateral prostate that was not associated with inflammatory lesions: the incidence of dysplastic lesions of moderate to marked severity was reduced from 100% in the control group to 59% in high dose animals, whereas the incidence of slightly dysplastic lesions increased to 41% (Table 2).

Selenomethionine did not affect the incidence of tumors at other sites, including pituitary tumors found in 100% of rats treated with testosterone plus estradiol, which is a known estrogen effect (Table 3).

Feeding testosterone plus estradiol–treated NBL rats with α-tocopherol mixed into diet at 2,000 and 4,000 mg/kg did also not affect prostate cancer incidence and multiplicity (Table 4). Dietary α-tocopherol at these doses did not affect body weight (data not shown), food intake (data not shown), or mortality (Table 4).

The incidence of marked epithelial dysplasia in the lateral prostate caused by the testosterone plus estradiol treatment was significantly reduced by α-tocopherol in a dose-related fashion, regardless of whether the dysplastic lesions were associated with inflammation (Table 4). α-Tocopherol also slightly, but statistically significantly, reduced the incidence of marked inflammation in the lateral prostate in a dose-related fashion (Table 4).

Dietary α-tocopherol significantly increased the incidence of adenocarcinomas of the mammary glands at both dietary concentrations (Table 5). α-Tocopherol did not affect the development of pituitary tumors occurring in 100% of NBL rats treated with testosterone plus estradiol or the incidence of tumors at other sites (Table 5).

In the present study, we showed that feeding of 1.5 and 3.0 ppm selenium in the form of selenomethionine and 2,000 and 4,000 ppm α-tocopherol mixed into a natural ingredient diet did not affect prostate carcinogenesis in an animal model that involves oxidative stress and inflammation, the testosterone and estradiol–treated NBL rat. Importantly, the findings of this study reproduce the result of the SELECT trial that a daily supplement of 200 μg selenium (as l-selenomethionine), 400 IU vitamin E (as DL-α-tocopherol), and their combination do not prevent prostate cancer. These findings are also consistent with our previous observations in the methylnitrosourea plus testosterone-treated WU rat prostate cancer model (16, 17).

The results of this study and those of the SELECT trial do not support the notion that selenium (as selenomethionine) and vitamin E (as α-tocopherol) are inhibitors of prostate carcinogenesis, as was suggested by the results of the Nutritional Cancer Prevention study that used selenized yeast and the ATBC trial with α-tocopherol (11–13). The results of many, but not all, epidemiologic studies suggest a protective role of selenium against prostate cancer (26–30). Prospective studies of toenail selenium levels indicated protection (26, 27), whereas no inverse relation with risk was found in most large cohorts studies of plasma selenium levels and selenium supplement use (26, 28, 29). Our findings are in line with those epidemiologic studies in which a lack of protective activity against prostate cancer was observed for selenium (26, 28–30) and for vitamin E, mostly in nonsmokers (31–33). Protective effects of dietary and/or supplemental vitamin E have only been found in smokers in some, but not all, studies (31–33) and the possible protective effect of selenium may also be modified by smoking status as well as by vitamin E intake (28). There is even some evidence to suggest that in nonsmokers, vitamin E supplements could increase prostate cancer risk (31) and plasma selenium levels were associated with increased risk in a recent study, depending on manganese superoxide dismutase genotype (34). The present study did not assess the possible interaction between selenomethionine and α-tocopherol, but previous studies with the WU rat model did not indicate protective activity of a combination of selenomethionine and α-tocopherol (16, 17). Both agents in combination may reduce the risk of fatal or advanced prostate cancer as suggested by some epidemiologic studies (28, 30), but the design of the present study and previous experiments with the WU rat model do not allow adequate assessment of such effects. The results of a small recent randomized clinical trial indicated that the effects of selenium supplementation on the incidence of skin cancers and internal malignancies only occurred at a selenium dose of 200 μg/d, but not at a 2-fold higher dose (35). A study in dogs suggested a U-shaped dose-response relationship between selenium intake and DNA damage in prostate cells (36). These observations raise the possibility that some of the discrepancies in the literature and the lack of effects of selenium in animal studies are due to such a nonlinear dose-response relationship.

One major issue not addressed in our animal experiments, SELECT, and epidemiologic studies is the influence of baseline selenium and vitamin E status. In the Nutritional Cancer Prevention trial, selenium was only protective in men with blood selenium levels in the lower two tertiles (<123 ng/mL; ref. 12). In SELECT, median baseline serum selenium levels were between 135 and 138 ng/mL with interquartile ranges from approximately 123 to 150 ng/mL (15), well above those of the lower tertiles of the Nutritional Cancer Prevention trial (12). In our animal studies, the control diet contained 0.31 mg/kg selenium, well above the dietary requirement for rats of 0.15 mg/kg diet (37), resulting in a selenium intake of ∼20 μg/kg body weight per day, whereas in the animals fed the supplemented diets, intakes were in the range of 100 and 200 μg/kg body weight per day in the low and high selenium dose groups, respectively. This probably resulted in plasma selenium concentrations in the control group of ∼600 ng/mL and in the high selenium group of ∼670 ng/mL,⁴ considerably higher that those in the SELECT trial. It is conceivable that selenium is only protective against prostate cancer in men with low to marginally deficient selenium levels (36) and that this might also be the case in the rat model we used. In a recent epidemiologic study by Peters et al. (28), the mean serum selenium level was 141 ng/mL (range, 51-253 ng/mL) in a U.S. male population and similar serum selenium levels (means of 119-126 ng/mL) were found in the third National Health and Nutrition Survey study (38). Thus, even if selenium protects against prostate cancer, most U.S. men would not have a selenium status that is sufficiently low to provide benefit from selenium supplementation.

The control diet in our animal studies contained 22 mg/kg α-tocopheryl acetate, which is the approximate dietary requirement for rats of 18 mg α-tocopherol per kilogram diet (37), resulting in a daily α-tocopheryl acetate intake of ∼0.5 mg/kg body weight, whereas intakes in the animals fed the supplemented diets were in the range of 135 and 270 mg/kg body weight per day in the low and high α-tocopherol dose groups, respectively. This likely resulted in plasma α-tocopherol concentrations in the control group of approximately 5 and 15 μg/mL in both supplemented groups.⁴ Baseline plasma α-tocopherol levels in the SELECT trial were ∼12 μg/mL and increased to ∼18 μg/mL in the supplemented groups (15). Thus, the conditions of vitamin E animal study were comparable with those of SELECT.

The natural ingredient NIH-07 diet used in this study contains substantive amounts (∼400 mg/kg) of isoflavones (39) and we have shown that soy isoflavones can inhibit prostate carcinogenesis in another rat model (40) when added to a natural ingredient diet with even higher basal isoflavone levels (39). Because of this observation and the fact that the basal diet in the control and supplemented groups in the present studies was identical, it is unlikely that these dietary isoflavones modulated the effects of selenomethionine and α-tocopherol.

The form of selenium and vitamin E may also be critically important in determining the potential protective effects of these agents. For example, methylseleninic acid and methylselenocysteine, but not selenomethionine, had inhibitory activity on growth of human prostate cancer xenografts in nude mice (41) and methylseleninic acid and methylselenocysteine slowed prostate cancer development in the TRAMP model (42), whereas selenomethionine did not (19). Furthermore, it is conceivable that vitamin E may protect against prostate cancer not as all-rac-α-tocopherol but in its natural form of RRR-α-tocopherol or as γ-tocopherol, the predominant form of vitamin E in the U.S. diet. There is some epidemiologic evidence to suggest that γ-tocopherol is protective against this major male malignancy (33, 43).

It is possible that the NBL rat model and other currently available animal models of prostate carcinogenesis are not suitable for testing the antiprostate cancer activity of antioxidants because they lack significant oxidative stress mechanisms. However, we and others have found evidence in the NBL rat model of formation in the prostate of oxidative DNA damage (8-hydroxy-2′-deoxyguanosine), lipid peroxidation, DNA strand breaks, and reduction of activity of the antioxidant enzymes, specifically linked to the hormone treatment (22, 44–46). These effects are attributed to the hydroxylation of 17β-estradiol to a catechol-estrogen, which can undergo redox cycling that results in the generation of (1) reactive quinones that can adduct to DNA and cause depurination, and (2) reactive oxygen species that can cause oxidative DNA damage and lipid peroxidation (44, 47). We have found evidence to suggest that this redox-cycling mechanism indeed occurs in the testosterone plus estradiol– treated NBL rat (48).

l-Selenomethionine and all-rac-α-tocopherol acetate both reduced the severity of prostatic dysplasia, a lesion comparable with human PIN, and α-tocopherol acetate, but not selenomethionine, slightly reduced the severity of prostatic inflammation. However, neither agent inhibited cancer induction in the testosterone plus estradiol–treated NBL rat. The biological significance of these effects is not clear because the dysplastic lesions in the glandular portion of the dorsal and, particularly, the lateral prostate lobes rarely if ever progress to cancer (20).⁵

However, these findings may suggest that the mechanism of dysplasia and inflammation induction in the NBL rat by testosterone plus estradiol involves oxidative stress and lipid peroxidation, whereas cancer induction may be specifically related to mechanisms other than oxidative DNA damage. One such alternate mechanism may be the formation of the aforementioned depurinating estrogen-quinone DNA adducts (48). Studies are ongoing in our laboratory to determine the effects of both interventions on biomarkers of oxidative stress in the prostatic target tissues of hormone-induced carcinogenesis in the NBL rat.

Interestingly, all-rac-α-tocopherol acetate increased the incidence of adenocarcinomas of the mammary glands from 3% to 23% to 27%. Historic control groups of male NBL rats treated with testosterone plus estradiol in our laboratory had a 0% to 8% incidence of mammary cancer.⁶

However, the mammary cancer incidence in the control group of the selenomethionine experiment was 27% and the incidence was 10% to 27% in the treatment groups of both experiments. These observations suggest that there is a wide variation in the occurrence of mammary carcinomas in NBL rats treated with testosterone plus estradiol. Thus, it is possible that the mammary cancer incidence in control animals of the α-tocopherol experiment was spuriously low, resulting in an apparent increase in incidence in the rats fed the tocopherol-supplemented diets. There is no information about vitamin E and risk of male breast cancer in humans. The observation of mammary cancer in testosterone plus estradiol–treated male NBL rats may be an artifact of this model. In female NBL rats, this hormone treatment results in a high incidence of mammary cancer (49), but the effect of α-tocopherol in this model is not known. There was no evidence to indicate a protective effect of vitamin E against female breast cancer risk in a recent randomized clinical trial (50) and there is no consistent evidence for this from epidemiologic studies (51).

The high incidence of prostate carcinomas in control animals in this study may have reduced the sensitivity of the model to inhibitory effects of chemopreventive agents. However, inhibition of prostatic dysplasia was observed in both studies. Furthermore, others have previously found inhibition of dysplasia in the testosterone plus estradiol–treated NBL rat by 4 months of dietary treatment with 9-cis-retinoic acid and DHEA, agents that were protective in the MNU plus testosterone–treated WU rat model, but not with 4-(hydroxyphenyl)retinamide, which was not inhibitory in the WU rat model (2, 17, 24, 52, 53). Thus, the NBL rat model does seem to be sensitive to the inhibition of prostatic lesion development by chemopreventive agents. Our findings do not exclude the possibility that selenomethionine and α-tocopherol may have the potential to affect the growth of existing metastatic prostate cancers. For example, monomethylated selenium inhibited the growth of LNCaP xenografts in nude mice (38).

Collectively, our current and previous animal model findings suggest that selenomethionine and α-tocopherol do not prevent prostate cancer. Importantly, our results are concordant with and predictive of the results of the SELECT trial and are therefore not supportive of its rationale. Much of the available epidemiologic evidence about these two agents, particularly vitamin E (except in smokers), also suggests the lack of protective activity when given as supplements. Alternatively, the results of our animal studies and SELECT might be interpreted to suggest that oxidative stress and resulting lipid peroxidation are not associated with increased prostate cancer risk. The findings of SELECT and the animal model studies do not exclude the possibility that other forms of vitamin E than α-tocopherol and selenium not in the form of organo-selenium compounds may provide protection against prostate cancer. It is also conceivable that these agents are protective only under conditions of selenium or vitamin E deficiency, but this was not addressed in SELECT and our animal studies.

No potential conflicts of interest were disclosed.

We thank Drs. A.M. Diamond, D.L. McCormick, and M.E. Wright for their helpful comments on this manuscript.

Grant Support: NIH grant no. R01 CA104334 and by NIH Center grants P30 ES00260 and P30 CA16087.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.