Inhibition of Rat Mammary Gland Chemical Carcinogenesis by Dietary ...

(CANCER RESEARCH 51. 481-486. January 15, 1991]

Inhibition of Rat Mammary Gland Chemical Carcinogenesis by Dietary Dehydroepiandrosterone or a Fluorinated Analogue of Dehydroepiandrosterone1 Thomas A. Ratko,2 Carol J. Detrisac, Rajendra G. Mehta, Gary J. Kelloff, and Richard C. Moon3 Laboratory of Paihophysiology, Life Sciences Division, I1T Research Institute, Chicago, Illinois 60616 [T. A. R., R. G. M., C. J. D., and R. C. M.], and National Cancer Institute, Chemoprevention Branch, Bethesda, Maryland 20892 [G. J. K.J

ABSTRACT The chemopreventive efficacy of p.o. administered dehydroepiandrosterone (DHEA), DHEA plus Ar-(4-hydroxyphenyl)retinamide (4-HPR), or 16a-fluoro-5-androsten-17-one (DHEA analogue 8354) was examined in rats treated with A'-methyl-A'-nitrosourea (MNU; 50 mg/kg bodyweight, i.v.) at 50 days of age. Semipurified diet (AIN-76A) containing each steroid alone, or DHEA plus 4-HPR, was administered during initiation (—1week to +1 week post-MNU), promotion/progression (+1 week post-MNU to termination), or both phases (—1week post-MNU to termination) of the carcinogenic process. Neither DHEA nor DHEA analogue 8354 (0.2%, w/w) significantly affected the initiation of mam mary cancer when administered alone; however, DHEA (0.2%, w/w) plus 4-HPR (1 mmol/kg diet) significantly reduced cancer multiplicity (26%) when given during initiation. All three treatments were strongly effective when given during promotion/progression, significantly reducing mam mary cancer multiplicity by 77% (DHEA), 84% (DHEA/4-HPR), and 66% (DHEA analogue 8354), relative to carcinogen controls. Cancer incidence was significantly inhibited by DHEA (33% inhibition) and DHEA/4-HPR (24% reduction) during promotion/progression. However, the most effective chemopreventive treatment encompassed both phases of carcinogenesis. Thus, under these conditions, DHEA (0.2% or 0.1%, w/w) reduced cancer incidence (52% and 32% reductions, respectively) and multiplicity (91% and 86% reductions, respectively). Further reduc tion in mammary cancer incidence was observed in animals that received DHEA (both doses) plus 4-HPR (1 and 0.5 mmol/kg diet, respectively). DHEA analogue 8354 (0.2% or 0.1%, w/w) given for the duration of the study reduced only cancer multiplicity (61% and 56% reductions, respec tively). Tumor-related mortality was significantly lower in rats that received long-term treatment with DHEA or DHEA/4-HPR, when com pared with carcinogen controls. Except for a slight, but significant, postcarcinogen decrease in the mean body weights of rats treated concom itant 1\ with DHEA (plus or minus 4-HPR) and MNU, additional gross manifestations of steroid-induced toxicity were not observed.

levels or excretion of this steroid or its metabolites and a number of disease syndromes including cancer (2). In women, plasma concentrations of DHEA (present almost exclusively as the sulfate conjugate, DHEA sulfate) show a continual decline from peak levels that occur in the second decade of life (3, 4). Conversely, the incidence of breast cancer increases with age (5). An early report by Bulbrook et al. (6) showed that preoperative urinary concentrations of 11-deoxy17-ketosteroids were subnormal in women afflicted with pri mary breast cancer, and suggested that this alteration may be a predictor of risk for the subsequent appearance of the disease. Subsequent to that report, a prospective study showed a definite inverse correlation between the urinary excretion of etiocholanolone and androsterone, the two main urinary metabolites of adrenal androgens, and the development of breast cancer in women (7). More recently, it was shown that 24-h mean plasma levels of DHEA and DHEA sulfate were subnormal in women with premenopausal breast cancer, whereas postmenopausal patients had supranormal plasma levels of these compounds (8), thus providing a potential biochemical marker for the dichotomization of this disease relative to menopausal status (5). The in vivo relevance of DHEA to the genesis of neoplastic disease has been experimentally explored in a number of animal models. Schwartz (9) has shown that p.o. administration of DHEA inhibited the appearance of spontaneous mammary cancer in female C3H(A'J) mice and reduced the incidence and

multiplicity of lung tumors induced in A/J mice by either DMBA or urethan (10). In CD-I mice, topical application of either DHEA or the synthetic derivative 3/3-methylandrost-5en-17-one significantly reduced the number and incidence of DMBA-induced skin papillomas and carcinomas (11). Dietary INTRODUCTION administration of DHEA also resulted in significant reductions DHEA4 is an abundantly secreted adrenal steroid that is an in the rate of appearance and frequency of all stages (atypical hyperplasia, carcinoma in situ, invasive carcinoma) of colonie intermediate in the biosynthesis of other hormones, including lesions induced in female BALB/c mice with 1,2-dimethylhytestosterone and estradici-170 (1). Although a physiological drazine (12). Finally, postcarcinogen dietary administration of role for DHEA has yet to be defined, a growing body of DHEA significantly inhibited the development of thyroid tu evidence, both epidemiológica! and experimental, suggests a mors and the frequency of putative neoplastic precursor lesions strong inverse relationship between alteration in the serum in the livers of male F344 rats that were treated with dihydroxydi-n-propylnitrosamine (13). Received 3/1/90; accepted 10/24/90. The costs of publication of this article were defrayed in part by the payment The results described above indicate that DHEA has signifi of page charges. This article must therefore be hereby marked advertisement in cant chemopreventive efficacy against experimental carcinogen accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1Supported in part by contract N01-CN-45912-14 and Grant CA-34664 (R. esis at a diverse group of target sites. However, with the G. M.) from the National Cancer Institute. exception of breast cancer, there is no clear precedent relating 2 Present address: American Medical Association, 515 North State Street, Chicago. IL 60610. specific cancer risk in humans to alterations in circulating levels 3To whom requests for reprints should be addressed, at IIT Research Institute, of this steroid or the production and excretion of its metabolites 10 West 35th Street, Chicago, IL 60616. 4 The abbreviations used are: DHEA, dehydroepiandrosterone; DHEA ana (2). For this reason, and given the lack of previous experimental logue 8354, 16a-fluoro-5-androsten-17-one; MNU, N-methyl-N-nitrosourea; 4data regarding chemopreventive efficacy of DHEA in mammary HPR, /V-(4-hydroxyphenyl)retinamide; DMBA, 7,12-dimethylbenzanthracene; carcinogenesis, we have undertaken to explore the temporal ANOVA. analysis of variance; NCW, net carcass weight; G6PDH, glucose-6phosphate dehydrogenase; TPA, 12-O-tetradecanoylphorbol-13-acetate. specificity of DHEA as an inhibitor of MNU-induced mammary 481

DEHYDROEPIANDROSTERONE

ANTICARCINOGENESIS

were prepared by spiking 5 g of basal diet with known amounts of material, extracting and quantitating as described above. Recoveries of DHEA were generally >90%; recoveries of DHEA analogue 8354 were generally >84%. As determined by the described method, both steroids were stable for 4 days at room temperature when mixed in the basal diet at the levels used in the studies. All diet materials were replaced twice weekly. DHEA/4-HPR Combination Chemoprevention Study. Two hundred sixty rats were randomized by weight into 12 groups at 43 days of age. With the exception of the carcinogen control group, which received a diet containing 4-HPR vehicle (30 rats), the other carcinogen-treated experimental groups (8 additional) each contained 25 animals. Groups (3 total) that received MNU vehicle (0.85% NaCl solution) consisted of 10 rats each. Provision of diets supplemented with steroid plus retinoid was also begun at this time according to 1 of 3 schedules, to encompass the periods of initiation (1 week before through 1 week after MNU treatment, i.e., -1 week to +1 week), promotion/progression (1 week after MNU treatment, continued to the end of the study, i.e., +1 week to the end), or both (1 week before MNU treatment, continued to the end of the study, i.e., —1week to the end). As determined in preliminary feeding studies, DHEA was well tolerated when fed alone at 0.2% or 0.1% (w/w) of diet (6.93 and 3.47 mmol/kg diet, respec tively); combination diets contained the same doses of DHEA plus 4HPR at 1 or 0.5 mmol (391 or 195.5 mg/kg diet, respectively). At 50 days of age, all carcinogen-treated rats received a single i.v. injection of freshly prepared MNU solution (50 mg/kg body weight) via the jugular vein as described previously (18). Crystalline MNU (Ash-Stevens, Detroit, MI) was dissolved to a concentration of 12.5 MATERIALS AND METHODS mg/ml in 0.85% NaCl solution acidified to pH ~5.0 with glacial acetic Experimental Animals. Virgin female Sprague-Dawley [Hsd: (SD) acid. Control animals received an i.v. injection of the NaCl solution BR] rats were received from Harlan/Sprague-Dawley (Indianapolis, IN) only. at 28 days of age and maintained in isolation for 2 weeks. A total of Commencing 4 weeks after receiving MNU, animals were palpated 390 rats were used in the reported studies. Animals were housed in weekly to monitor mammary tumor appearance. The date of appearance groups of 2 to 3 in polycarbonate cages containing hardwood bedding. and location of every palpable tumor were recorded. The body weights Cage materials were replaced twice weekly. The animal rooms were of all rats were recorded once a week for the duration of each study. illuminated for 12 h each day and maintained at a temperature of 22 ± All rats were observed twice daily for any indications of agent-induced 1°C(SE) and 50% relative humidity. Animals were allowed free access toxicity. At no time during the experiments were the estrous cycles of to food and water throughout each study. any rats monitored. Diets and Chemopreventive Agents. The basal diet for each study was After 180 days of chemopreventive treatment, surviving rats in all modified AIN-76A semipurified diet (adjusted corn starch) TD 85449 groups were sacrificed by CO2 asphyxiation. During each study, animals (Teklad, Madison, WI). 4-HPR was obtained from Cilag AG, Schaffthat appeared moribund were killed by CO2 asphyxiation. All rats that house, Switzerland. The required vehicle for 4-HPR consisted of 12.5 were killed or found dead were promptly given thorough postmortem g of ethanol, 36.9 g of corn oil, and 0.6 g of Tenox 20 (20% tertiary examination. Mammary tumors were coded by location, removed, butylhydroquinone, 10% citric acid, 70% propylene glycol; Eastman measured, and weighed. All tumors and any other grossly abnormalChemicals, Kingsport, TN) per kg diet. Control animals (carcinogen appearing tissues were removed and fixed in 10% buffered formalin. and carcinogen vehicle groups) in the study with DHEA received a Sections were stained with hematoxylin and eosin and classified histobasal diet containing the 4-HPR vehicle. DHEA was purchased from pathologically. Mammary tumor pathology was defined according to Sigma Chemical Co. (St. Louis, MO). The DHEA analogue 8354 was the criteria of Young and Hallowes (24). a gift from Fort Washington Resources (Hatboro, PA). Its synthesis DHEA Analogue 8354 Chemoprevention Study. One hundred thirty and chemical properties were described previously (21). Fresh batches rats were randomized by weight into 7 groups at 43 days of age. Each of diet were prepared weekly (high doses) or every 3 weeks (low doses), as more animals received the high-dose diets, and stored at — 20°C group that received MNU (5 total) consisted of 20 animals; in addition, a control group that received steroid-supplemented diet and carcinogen before use. vehicle also contained 20 rats. Another vehicle-treated group that As analyzed by high-pressure liquid chromatography (23), 4-HPR received AIN-76A basal diet contained 10 animals. Dose schedules for was completely stable under the storage conditions used and when diet chemopreventive treatment were identical to those described above for supplemented with 4-HPR was left at room temperature for 4 days. the DHEA/4-HPR combination study. The DHEA analogue 8354 was The stability and content of DHEA and DHEA analogue 8354 in the fed to treated rats at 0.2% or 0.1% (w/w) of diet (6.89 and 3.44 mmol/ basal diet were assessed by high-pressure liquid chromatography using kg diet, respectively). All procedures and observations were performed extraction and Chromatographie methodology developed in our labo as described above for the DHEA/4-HPR combination study. ratory.5 Each steroid was extracted from the diet with 100% methanol Statistical Analysis. Tumor incidence curves were generated by the (50 ml/5 g of diet) by occasionally swirling the mixture over a period life table method and compared by log rank analysis (25). The statistical of 2 h at ambient temperature. After gravity filtration to remove significance of differences between mean tumor multiplicities was as paniculate material, the volume of each filtrate was adjusted as neces sessed using one-way ANOVA. Individual tumor numbers from the sary to 50 ml, and 10-^1aliquots were injected directly onto a Whatman DHEA/4-HPR combination study were transformed by the square root Partisi! ODS-2 column. The steroids were eluted from the column method to normalize the data before comparison by ANOVA (26). using 100% methanol at a flow rate of 1.2 ml/min. The DHEA peak Differences in group total mean body weights at termination of the was visualized by UV absorption spectroscopy at a wavelength of 210 DHEA analogue 8354 study were tested for statistical significance by nm; DHEA analogue 8354 was detected at 215 nm. Standard curves ANOVA, using untransformed individual weights. Due to significant *T. A. HuitÃ-n,unpublished observations. unresolved heterogeneity among the group variances, total mean body 482

gland carcinogenesis in female rats. The potential clinical sig nificance in using this breast cancer model is its similarity to the human disease in terms of histopathology (14) and response to hormonal manipulations (15-17). In addition, the use of MNU, a direct acting mammary gland carcinogen (18), to induce cancer effectively rules out complications related to DHEA-mediated effects on carcinogen metabolism and dispo sition (2, 10). Previous work demonstrated that the chemopreventive efficacy of retinoids (including 4-HPR) in experimental breast cancer is enhanced when they are combined with various hormonal treatment modalities (19, 20). The present study was designed to evaluate the potential steroid modulating effects of DHEA, administered alone and in combination with the syn thetic retinoid 4-HPR in rat mammary carcinogenesis. Finally, the chemopreventive activity of a fluorinated derivative of DHEA, denoted DHEA analogue 8354 (21), was examined in a separate study. This analogue is a more potent inhibitor than DHEA of several biochemical processes considered central to carcinogenesis. The analogue apparently does not elicit the androgenic and estrogenic effects associated with the parent steroid (21, 22). Thus, comparison of these agents in chemoprevention studies may lead to insights as to the mechanism(s) by which these compounds suppress carcinogenesis.

DEHYDROEPIANDROSTERONE

ANTICARCINOGENESIS

Table 1 Effect of dietary DHEA on MNV-induced mammary gland cardnogenesis infernale Sprague-Dawley rats Virgin, female Sprague-Dawley rats (50 days old) received single injections of MNU (50 mg/kg body weight) dissolved in a vehicle of 0.85% NaCI solution (pH ~5.0), or vehicle alone. Administration of AIN-76A semipurified diet (basal diet) supplemented with 4-HPR vehicle or DHEA at the indicated doses was begun as described below. The study was terminated 180 days after rats received MNU or NaCI solution. Different numbers of asterisks in each column of data denote statistically significant intergroup differences. No. of rats MNU24 +25 +25 +25 +30 +10 Vehicle10

diet, w/w)0.2'o.y0.2'0.1'—(%)100*67"48"68"100*Cancers/rat"'*8.8 1.0*2.3 ± ±0.9*2.4 0.6"0.9 ± 0.6"1.0 ± 0.2"1.4 ± ±0.2"1.4 ±0.2"10.1 ±0.2**10.7 *0.2*ACancerincidence ±0.8*Tumors/rat"''9.1±0.8*Terminal

VehicleDHEA(% " Group mean average ±SEM. * Includes all histologically confirmed mammary carcinomas found at necropsy. ' Includes all palpable tumors. d NCW represents terminal weight of tumor-bearing rat minus weight of excised tumors. ' Diet given from —1week to +1 week post-MNU. ^Diet given from +1 week post-MNU to end of study. * Diet given from -1 week post-MNU or vehicle to end of study. * Diet contained 4-HPR vehicle (see text for details). ' P < 0.05 versus respective control group.

weights for the DHEA/4-HPR combination study were compared using (as appropriate) a i test for means with significant heterogeneity of variances, as suggested by Sokal and Rohlf (26). In the latter case, significance was attributed to P < 0.01. As shown in the "Results." we also have calculated a NCW for each animal that survived to the end of each study. The NCW represents the total (gross) terminal weight of each tumor-bearing rat minus the total weight of tumor tissue (i.e., the tumor burden). This parameter (NCW) was also evaluated by ANOVA as described above to detect differences between the various means. An appropriate discrimination test for making unplanned com parisons (least significant difference method) was used to define statis tically significant intergroup differences in means (26). Differences in percentage survival and cancer incidences were tested for significance by Fisher's exact test (2-tailed). In all cases, with the exception of multiple / tests, statistical significance was ascribed to a comparison only when a P < 0.05 was attained.

RESULTS DHEA/4-HPR Combination Study. The chemopreventive ac tivity of DHEA alone, and DHEA plus 4-HPR, were examined in a combined study using common MNU and MNU vehicle controls. However, to facilitate their presentation, the data have been separated into two Tables, with the common controls shown in each. The data summarized in Table 1 indicate that dietary DHEA significantly inhibited MNU-induced tumorigenesis in female rats when given during the +1 week to end schedule. Thus, rats fed DHEA at 0.2% (w/w) during this period developed cancer at an incidence of 67% and mean (±SEM) multiplicity of 2.3 ±0.6 cancers/rat, representing significant (P< 0.05) reductions in these parameters relative to the controls. In contrast, treat ment with 0.2% DHEA during the -1 week to +1 week period had no significant effect on mammary cancer incidence (100%) or mean (±SEM) multiplicity (8.8 ±1.0 cancers/rat) when compared with controls that received a diet containing 4-HPR vehicle only. This control group had a cancer incidence of 100% and mean (±SEM) multiplicity of 10.1 ±0.8 cancers/rat. Treatment with DHEA over the in-life phase of the study (—1 week to end) significantly reduced mammary cancer incidence to 48% (0.2% dose) and 68% (0.1 % dose). Mean (±SEM)cancer multiplicity was 0.9 ±0.2 (0.2% dose) and 1.4 ±0.2 (0.1% dose) cancers/rat. However, statistical comparison of the tumor data for the two groups that received DHEA from —1week to 483

survival63*8096"92**57*100100Terminal total body NCW(g)*-'254 (g)a269 wt 1265±1 ±6255 5"261± ±5287 12*265 ± 7'294±

±9261251257245265294655

±12Terminal

end at 0.1% or 0.2% showed no significant intergroup differ ences in cancer incidence or mean multiplicity. Finally, log rank analysis of cancer incidence curves showed significant length ening of cancer latency in groups treated with DHEA during the +1 week to end or —1week to end periods, relative to those treated during the —1week to +1 week period or the controls. As shown in Table 2, combination treatment with DHEA plus 4-HPR was most effective when given from —1week to the end of the study. Cancer incidence and multiplicity also were significantly lower in the groups treated with the combi nation during the +1 week to end period relative to the controls. Furthermore, cancer multiplicity in the rats that received DHEA plus 4-HPR during the —1week to +1 week period was significantly lower than that of the carcinogen controls. Finally, comparison of the data in Tables 1 and 2 indicates that com bined treatment with DHEA plus 4-HPR during the -1 week to end period was significantly more effective in reducing cancer multiplicity and incidence than was treatment with the steroid alone. When administered alone to rats at 1 mmol/kg AIN76A diet, 4-HPR has no effect on MNU-induced tumorigene sis.6 The survival data in Tables 1 and 2 show that 57% of the carcinogen control rats lived until the end of the experiment (187 days); survival of rats that received DHEA alone ranged from 63% (-1 week to +1 week) to 96% (-1 week to end, 0.2%) (Table 1). As shown in Table 2, survival of carcinogentreated rats that received DHEA plus 4-HPR ranged from 72100%. During the experiment, only 4 rats died in a cancer-free state. In the control groups that received carcinogen vehicle and diet containing DHEA plus or minus 4-HPR, grossly visible evidence of steroid-induced pathology was not observed at necropsy. Fig. 1 shows that the mean body weight of rats treated with dietary DHEA at 0.2% (w/w) was approximately 18% lower (P < 0.05) than that of placebo-fed controls within 6 days of receiving MNU injections. However, the rate of weight gain in that group recovered within the subsequent 7-14 days after carcinogen exposure, and was comparable with that of the carcinogen controls for at least 60 additional days. At termi nation of the study, although the gross mean body weight of the MNU-treated group on the diet containing 0.2% DHEA 'T. A. Ratko. C. J. Detrisac, R. G. Mehla, G. J. Kelloff. and R. C. Moon, unpublished observations.

DEHYDROEPIANDROSTERONE

ANTICARCINOGENESIS

Table 2 Effect of dietary DHEA plus 4-HPR on MNU-induced mammary gland curcinogenests infernale Sprague-Dawley rats Virgin, female Sprague-Dawley rats (50 days old) received single injections of MNU (50 mg/kg body weight) dissolved in a vehicle of 0.85% NaCI solution (pH ~5.0), or vehicle alone. Administration of A1N-76A semipurified diet (basal diet) supplemented with 4-HPR vehicle or DHEA plus 4-HPR at the indicated doses was begun as described below. The study was terminated 180 days after rats received MNU or NaCI solution. Different numbers of asterisks in each column of data denote statistically significant intergroup differences. No. of rats MNU25

incidence100*76**29*"36*"100*Cancers/rat""''7.5 survival72***92"92**100"57*100100Terminal total NCW body (g)°278 wt. (g)"-'2552672572622452756446106 diet, w/w)0.2'o.y0.2*0.1*A0.2*A4-HPR(mmol/kg diet)1.01.01.00.51.0Cancer ±0.6***1.6 ±0.6***1.7 8268+ +25 0.3"0.7 + 0.3"0.7 + +24 ±4258 0.3"0.6 ± 0.3**0.6 ± 4"263± +25 0.2**10.1 + 0.2**10.7 + 6287+ +30 ±0.8*Tumors/rat°'c7.8 ±0.8*Terminal 12*275 ± +10 6294+ Vehicle10 VehicleDHEA(% 12 ±12Terminal °Group mean average ±SEM * Includes all histologically confirmed mammary carcinomas found at necropsy. ' Includes all palpable tumors. NCW represents terminal weight of tumor-bearing rat minus weight of excised tumors. e Diet given from —1 week to +1 week post-MNU. 'Diet given from +1 week post-MNU to end of study. * Diet given from —1week post-MNU or vehicle to end of study. * Diet contained 4-HPR vehicle (see text for details).

of incidence curves showed that cancer latency was significantly longer in that group of rats when compared with the controls or those that received steroid for the —1week to +1 week period (data not shown). The percentage of rats surviving to the end of the study was not related to treatment with analogue 8354, when compared with controls that received the basal diet (Table 3). The DHEA analogue also had no significant effect on either terminal mean body weight or terminal mean NCW of MNUtreated rats, when compared with the carcinogen controls. Treatment with the DHEA analogue plus MNU also caused no early, postcarcinogen transient effects on body weights. In contrast, the analogue did have an adverse impact on the body weight of rats that received MNU vehicle. However, no other overt signs of agent-mediated toxicity were observed in animals that received the compound.

220 200 O—MNU vehicle, plecebo diet •— MNU. plecebo diet •A MNU. 0.2X B354 (-1 Hk to end) •4—MNU vehicle. 0.2X 8354 (-1 Hk to O—MNU vehicle. 0.2X DHEA (-1 wk to •—MNU. 0.2X DHEA (-1 Hk to end)

160

120

-10

0

10 20 Interval

40 after

60

80

MNU or

100 vehicle

120 injection

140

160

end) end)

180

DISCUSSION

(days)

Fig. 1. Influence of dietary DHEA (0.2%, w/w) or DHEA analogue 8354 (0.2%, w/w) on mean body weight of female Sprague-Dawley rats that received i.v. injections of MNU (50 mg/kg body weight) or MNU vehicle (0.85% NaCI solution, pH ~5.0) at day 0.

for —1week to end was significantly (though not substantially) lower than that of the carcinogen controls, the mean NCW (body weight of tumor-bearing rat minus weight of gross tu mors) of each group was not significantly different (Tables 1 and 2). In contrast to MNU-treated rats, those that received carcinogen vehicle and DHEA at 0.2% of the diet throughout the study showed no transient weight loss after vehicle injection; however, the mean weight of steroid-treated rats was signifi cantly lower than that of placebo-fed controls at termination of the study. The effect of combination treatment with DHEA and 4-HPR on body weights of rats paralleled that observed in animals that received steroid alone. DHEA Analogue 8354 Study. The data summarized in Table 3 show that DHEA analogue 8354 significantly inhibited MNU-induced mammary gland tumorigenesis when given dur ing the +1 week to end of study (0.2%, w/w) or -1 week to end of study periods (both doses), as evidenced by significant (5666%) reductions in cancer multiplicity relative to the carcinogen controls. Although the percentage incidence of mammary can cer was unaffected by dietary treatment with the analogue at 0.2% (w/w) during the —1week to end period, log rank analysis 484

The data herein reported are, to our knowledge, the first to show significant inhibitory activity of the adrenal-derived ste roid DHEA, or its fluorinated analogue 8354, toward the in duction of mammary carcinomas in the female rat. In addition, we have shown a positive chemopreventive interaction between DHEA, and the synthetic retinoid, 4-HPR, when they were concomitantly administered to rats via the diet. The results further indicate that DHEA, or its analogue 8354, are primarily active against the promotion/progression phase of MNU-in duced mammary carcinogenesis. This relationship can be easily deduced by comparing the tumor data of the appropriate car cinogen controls with those obtained in the 3 groups that received chemopreventive treatment via different feeding pro tocols. A major problem in validly interpreting the results of any cancer chemoprevention study is agent-induced toxicity in the experimental animals. Among any number of toxic manifesta tions that may accompany treatment with various agents, weight loss, or inhibition of the rate of weight gain relative to untreated controls, particularly as a function of reduced caloric intake, can be especially confounding (27). For example, the growth and development of DMBA-induced mammary tumors in rats can be modulated by restriction of caloric intake during the presumed period of initiation, and for a short period there after (28), possibly as a result of alteration in serum levels of

DEHYDROEPIANDROSTERONE

ANTICARCINOGENESIS

Table 3 Effect of dietary DHEA analogue 8354 on MNU-induced mammary gland carcinogenesis infernale Sprague-Dawley rats Virgin, female Sprague-Dawley rats (50 days old) received single injections of MNU (50 mg/kg body weight) dissolved in a vehicle of 0.85% NaCI solution (pH ~5.0), or vehicle alone. Administration of AIN-76 semipurified diet (basal diet) supplemented with DHEA analogue 8354 at the indicated doses was begun as described below. The study was terminated 180 days after rats received MNU or NaCI solution. Different numbers of asterisks in each column of data denote statistically significant intergroup differences. survival6555805070100100Terminal NCW(g)"'"258 8354 incidence10095100100100Cancers/rat"'*11.2 total body (g)fl276 wt (% diet, w/w)0.2'o.y0.2'0.1"— 1.0*4.1 ± 1.0*4.2 ± 7254± +20 ±7239 ±0.6**4.8 0.6**4.8 ± +20 12272 ± 14259 ± 1.1**5.4 ± 1.0"5.8 ± ±5265 ±6255 +20 0.8"12.2± ± 0.8"12.6 ± ±5266 ±6240 +20 *0.2"ACancer 1.0*Tumors/rat°'r11.5 + 0.9*Terminal ±7266 ±6266 +20 5*305 ± Vehicle10 5305± ±10"Terminal VehicleDHEA ±10 a Group mean average ±SEM * Includes all histologically confirmed mammary carcinomas found at necropsy. 1 Includes all palpable tumors. d NCW represents terminal weight of tumor-bearing rat minus weight of excised tumors. ' Diet given from -1 week to +1 week post-MNU. ^Diet given from +1 week post-MNU to end of study. * Diet given from —1week post-MNU or vehicle to end of study. * Basal diet (AIN-76A). No. of rats MNU20

prolactin and gonadotropins (29). In the present study, we have no data showing an effect of dietary DHEA on food consump tion, therefore, it is not possible to identify alteration of caloric intake as a specific factor that could be causally related to the early transient weight loss observed in DHEA-treated rats immediately after their exposure to MNU (Fig. 1). Since tran sient weight loss did not occur in the groups that received DHEA and MNU vehicle, it appears that toxicity of the steroid per se was not a sole factor in that phenomenon. Interestingly, a similar effect was documented by Schwartz et al. (30) in female CD-I mice, which were fed DHEA at 0.4% or analogue 8354 at 0.2% and 0.4% in the diet, concomitant with their exposure to DMBA via topical application. Thus, our data suggest that the observed anticarcinogenicity of DHEA was specific to some unknown action of the steroid, rather than to inhibition of body weight gain. This interpreta tion is also supported by the NCW data, which indicate similar body weights in all rats, regardless of the DHEA treatment protocol. However, it is possible, though difficult to quantify, that in the carcinogen control rats, the sizeable mean tumor burden induced significant cachexia, which could have reduced their mean NCW, thus giving the appearance that treatment with DHEA had no effect on this parameter. This may indeed have occurred, as evidenced by the lower mean NCW of the carcinogen controls, although data are not available to support this contention. The present data show that DHEA analogue 8354 did not significantly affect weight gain of rats given the steroid at 0.1 % of 0.2% (w/w) via the diet. This result contrasts with data reported by Schwartz et al. (21 ), who showed a significant effect of analogue 8354 on weight gain of male BALB/cJ mice that received the agent at 0.25% in the diet, despite their consuming quantities of food similar to those ingested by controls. How ever, due to the differences in species and sex of animals, it is difficult to directly compare our results with those of Schwartz et al. (21). Examination of the findings from animals receiving DHEA alone or in combination with 4-HPR reveals two beneficial effects associated with the addition of 4-HPR to the DHEA diets. At both doses evaluated, mammary cancer incidence was further reduced when the combination diets were given through out all phases of carcinogenesis. Furthermore, concomitant administration of these agents significantly lowered cancer mul 485

tiplicity during the initiation phase. The results of the 4-HPR/ DHEA combination study are not without precedent in showing a positive chemopreventive interaction between a retinoid and hormonal manipulation in the MNU-induced mammary carci nogenesis model. For example, in rats that received dietary 4HPR, s.c. injections of tamoxifen, or 4-HPR plus tamoxifen injections, the incidence and multiplicity of subsequent tumors after surgical excision of the first lesion were reduced to a greater extent by the combination of agents when compared with either dietary 4-HPR alone or injections of tamoxifen alone (19). A previous study showed a positive interaction between tamoxifen and 4-HPR in the inhibition of MNUinduced tumorigenesis (20). Also, when compared with either treatment mode alone, an enhanced inhibitory effect on mam mary carcinogenesis was observed in ovariectomized rats that also received dietary 4-HPR (31). Thus, the significant inter action between DHEA and 4-HPR, as was previously docu mented between the retinoid and various other hormonal ma nipulations, suggests that further investigation of this combi nation in cancer chemoprevention is warranted. Both DHEA and its analogue 8354 significantly modify a number of processes that are considered to be central to tumor promotion. These effects include the inhibition by DHEA of DNA synthesis in breast epithelium of C3H mice and phorbol ester-stimulated skin of ICR mice (32), reduction of TPA stimulation of Superoxide production in human neutrophils (33), and inhibition of TPA-induced increases of prostaglandin E2 levels in mouse skin (22). In addition, both DHEA and DHEA analogue 8354 are potent inhibitors of mammalian G6PDH (21), the first enzyme in the hexose-monophosphate shunt (34). This metabolic pathway is the primary extramitochondrial source of NADPH, which is required in a number of biochemical processes including xenobiotic catabolism by mixed function oxidase, and the biosynthesis of fatty acids and steroid hormones (34). It is also the primary pathway by which hexose phosphates are converted to the ribose and deoxyribose phosphates required for de novo biosynthesis of nucleotides and nucleic acids (34). Thus, it is apparent that treatment of rats with either DHEA or its analogue could have profound effects on a metabolically active population of target cells of the young rat mammary gland, with several possible loci at which tumor promotion could be blocked or suppressed, perhaps as a func tion of the inhibition of G6PDH (35).

DEHYDROEPIANDROSTERONE

Chemical modification of the basic DHEA structure has been used as a strategy by other investigators to reduce the incidence and severity of side effects engendered in experimental animals by the parent hormone. Schwartz et al. (21) recently showed that contrary to the effects seen with DHEA, the DHEA 8354 analogue was nonestrogenic in a rat uterine weight assay, and nonandrogenic when assayed using the castrated rat seminal vesicle model. This analogue is also more potent than the parent steroid in suppressing TPA-induced increases in epidermal DNA synthesis, Superoxide formation, and prostaglandin E2 synthesis (22). Our findings demonstrate no increased efficacy of the 8354 analogue over DHEA in the promotion and pro gression phases of mammary carcinogenesis. This suggests that the previously measured parameters that differentiate these compounds in vivo are not responsible for these observed chemopreventive effects. The similarity of the data for DHEA and the DHEA 8354 analogue in the MNU rat mammary model system thus may be related to the effects on G6PDH activity, which are similar for both compounds (21). The effects of chronic p.o. administration of DHEA or DHEA analogue 8354 on elements of rat reproductive physiology are presently un known. A study to evaluate blood steroid hormone levels during such chronic conditions is presently underway. Such data will provide additional avenues for exploration of the chemopreventive mechanism of action of these compounds in this model of human breast cancer. ACKNOWLEDGMENTS The authors wish to thank our technical staff for their expert per formance. Additional thanks are due to Fort Washington Resources for providing the DHEA analogue 8354, Dr. A. G. Schwartz for helpful discussion regarding this project, and P. Moser for excellent secretarial assistance in preparing this manuscript.

REFERENCES 1. Vande Wiele, R., and Lieberman, S. The metabolism of dehydroisoandrosterone. In: G. Pincus and E. Vollmer, (eds.). Biological Activity of Steroids in Relation to Cancer, pp. 93-110. New York: Academic Press. Inc., 1960. 2. Gordon, G. B., Shantz, L. M., and Talalay, P. Modulalion of growth, differentiation and carcinogenesis by dehydroepiandrosterone. Adv. Enzyme Regul., 26: 355-382, 1987. 3. Orentreich, N., Brind, J. L., Rizer, R. L., and Vogelman, J. H. Age changes and sex differences in serum dehydroepiandrosterone sulfate concentrations throughout adulthood. J. Clin. Endocrinol. Metab., 59: 551-555, 1984. 4. Carlstrom, K., Brody, S., Lunnell, N-O.. Lagrelius, A., Mollerstrom. G.. Pousette, A., Rannevik, G., Stege, R., and von Schoultz, B. Dehydroepian drosterone sulphate and dehydroepiandrosterone in serum: differences re lated to age and sex. Maturitas, 10: 297-306. 1988. 5. DeWaard. F.. DeLaive, J. W. J., and Baanders-Van Halewijn. E. A. On the bimodal age distribution of mammary carcinoma. Br. J. Cancer, 14: 437448, 1960. 6. Bulbrook, R. D.. Hayward, J. L., Spicer, C. C, and Thomas, B. S. Abnormal excretion of urinary steroids by women with early breast cancer. Lancet, 2: 1238-1240, 1962. 7. Bulbrook, R. D., Hayward, J. L., and Spicer. C. C. Relation between urinary androgen and corticoid excretion and subsequent breast cancer. Lancet. 2: 395-398, 1971. 8. Zumorff, B., Levin, J., Rosenfeld, R. S., Markham, M., Strain, G. W., and Fukushima, D. K. Abnormal 24-hr mean plasma concentrations of dehydroisoandrosterone and dehydroisoandrosterone sulfate in women with primary operable breast cancer. Cancer Res., 41: 3360-3363, 1981. 9. Schwartz, A. G. Inhibition of spontaneous breast cancer formation in female C3H (A'Va) mice by long-term treatment with dehydroepiandrosterone. Cancer Res., 39: 1129-1132, 1979. 10. Schwartz, A. G., and Tannen, R. H. Inhibition of 7,12-dimethylbenz[a] anthracene- and urethan-induced lung tumor formation in A/J mice by longterm treatment with dehydroepiandrosterone. Carcinogenesis (Lond.), 2: 1335-1337. 1981.

486

ANTICARCINOGENESIS

11. Pashko, L. L., Hard, G. C, Rovito. R. J.. Williams, J. R., Sobel, E. L.. and Schwartz. A. G. Inhibition of 7,12-dimethylbcnz(a)anthracene-induced skin papillomas and carcinomas by dehydroepiandrosterone and 3-fi-methyl-androst-5-en-l7-one in mice. Cancer Res.. 45: 164-166. 1985. 12. Nyce, J. W.. Magee. P. N.. Hard, G. C., and Schwartz. A. G. Inhibition of 1,2-dimethylhydrazine-induced colon tumorigenesis in BALB/c mice by de hydroepiandrosterone. Carcinogenesis (Lond.), 5: 57-62. 1984. 13. Moore, M. A, Thamavit, W., Tsuda, H., Sato, K.. Ichihara. A., and Ito, N. Modifying influence of dehydroepiandrosterone on the development of dihydroxy-di-n-propylnitrosamine-initiated lesions in the thyroid, lung and liver of F344 rats. Carcinogenesis (Lond.), 7: 311-316, 1986. 14. Tseng, M. T. Ultrastructure of the hormone-dependent jV-nitrosomethylureainduced mammary carcinoma of (he rat. Cancer Res., 40: 3112-3115. 1980. 15. Turcot-Lemay. L., and Kelly, P. A. Characterization of estradiol, progester one, and prolactin receptors in nitrosomelhylurea-induced mammary tumors and effect of antiestrogen treatment on the development and growth of these tumors. Cancer Res., 40: 3232-3240. 1980. 16. Arafah, B. M., Fullino. P. M., Manni. A., and Pearson. O. H. Effect of ovariectomy on hormone receptors and growth of A'-nitrosomethylureainduced mammary tumors in the rat. Cancer Res., 40:4628-4630, 1980. 17. Gandilhon, P., Melancon, R., Djiane, J., and Kelly. P. A. 7V-Nitroso-Afmethylurea-induced mammary tumors in the rat: role of prolactin and a prolactin-lowering drug. J. Nati. Cancer Inst., 70: 105-109, 1983. 18. McCormick, D. L.. Adamowski. C. B.. Fiks, A., and Moon, R. C. Lifetime dose-response relationships for mammary tumor induction by a single ad ministration of A'-methyl-A'-nitrosourea. Cancer Res., 41: 1690-1694. 1981. 19. Ratko. T. A,. Detrisac, C. J., Dinger, N. M., Thomas, C. F., Kelloff, G. J., and Moon, R. C. Chemopreventive efficacy of combined retinoid and tamoxifen treatment following surgical excision of a primary mammary cancer in female rats. Cancer Res., 49:4472-4476, 1989. 20. McCormick, D. L.. and Moon. R. C. Retinoid-tamoxifen interaction in mammary cancer chemoprevention. Carcinogenesis (Lond.), 7: 193-196, 1986. 21. Schwartz, A. G., Lewbart, M. L., and Pashko, L. L. Novel dehydroepian drosterone analogues with enhanced biological activity and reduced side effects in mice and rats. Cancer Res., 48:4817-4822, 1988. 22. Hastings. L. A., Pashko. L. L., Lewbart, M. L., and Schwartz, A. G. Dehydroepiandrosterone and two structural analogs inhibit 12-O-tetradecanoylphorbol-13-acetale stimulation of prostaglandin E2 content in mouse skin. Carcinogenesis (Lond.). 9: 1099-1102, 1988. 23. HuitÃ-n,T. A., Mehta, R. G., and Moon, R. C. A simple high-pressure liquid Chromatographie method for the separation of retinoids including \ (4 hydroxyphenyl)-all-rrans-retinamide. J. Chromatogr., 341: 187-192, 1985. 24. Young, S.. and Hallowes, R. C. Tumours of the mammary gland. In: V. S. Turusov (ed.). Pathology of Tumours in Laboratory Animals, Vol. I, Part I, pp. 31-73. Lyon, France: International Agency for Research on Cancer, 1973. 25. Peto, R., Pike, M. C., Armitage. P.. Breslow. N. E., Cox, D. R., Howard, S. V., Mantel, N., McPherson. K.. Peto. J., and Smith, P. G. Design and analysis of randomized clinical trials requiring prolonged observations of each patient. II. Analysis and examples. Br. J. Cancer, 35: 1-39, 1977. 26. Sokal, R. R., and Rohlf, F. J. Biometry. San Francisco: W. H. Freeman and Co., 1981. 27. Tannenbaum. A. Genesis and growth of tumors. II. Effects of caloric restric tion per se. Cancer Res., 2: 460-467, 1942. 28. Sylvester, P. W., Aylsworth. C. F., and Meites, J. Relationship of hormones to inhibition of mammary tumor development by underfeeding during the "critical period" after carcinogen administration. Cancer Res., 41: 13841388, 1981. 29. Campbell, G. A.. Kurcz, M., Marshall, S., and Meites. J. Effects of starvation in rats on serum levels of follicle stimulating hormone, leuteinizing hormone, thyrotropin, growth hormone and prolactin; response to LH-releasing hor mone and thyrotropin-releasing hormone. Endocrinology, 100: 580-587, 1977. 30. Schwartz, A. G., Fairman, D. K.. Polansky, M., Lewbart, M. L., and Pashko, L. L. Inhibition of 7,12-dimethylben/.|a]-anthracene-initiatedand 12-O-tetradecanoylphorbol-13-acetate-promoted skin papilloma formation in mice by dehydroepiandrosterone and two synthetic analogs. Carcinogenesis (Lond.), 10: 1809-1813. 1989. 31. McCormick, D. L., Mehta, R. G., Thompson, C. A., Dinger, N., Caldwell, J. A., and Moon, R. C. Enhanced inhibition of mammary carcinogenesis by combined treatment with A'-(4-hydroxyphenyl)retinamide and ovariectomy. Cancer Res., 42: 508-512, 1982. 32. Pashko, L. L., Schwartz. A. G., AbouGharbia, M.. and Swern, D. Inhibition of DNA synthesis in mouse epidermis and breast epithelium by dehydroepi androsterone and related steroids. Carcinogenesis (Lond.). 2:717-721. 1981. 33. Whitcomb, J. M., and Schwartz. A. G. Dehydroepiandrosterone and 16«Br-epiandrosterone inhibit 12-O-tetradecanylphorbol-13-acetate stimulation of Superoxide radical production by human polymorphonuclear leukocytes. Carcinogenesis (Lond.), 6: 333-335, 1985. 34. Lehninger, A. L. Biochemistry'- New York: Worth Publishers. Inc., 1975. 35. Raineri, R., and Levy, H. R. On the specificity of steroid interaction with mammary gland glucose-6-phosphate dehydrogenase. Biochemistry. 9:22332243, 1970.