Phytoestrogen Content of Purified, Open- and Closed-Formula Laboratory Animal Diets

Julius E. Thigpen,1 Kenneth D. R. Setchell,2 Kathy B. Ahlmark,1 Jacqueline Locklear,1 Tamara Spahr,1 Gordon F. Caviness,1 Mary F. Goelz,1 Joseph K. Haseman,3 Retha R. Newbold,4 and Diane B. Forsythe1

Background and Purpose: Phytoestrogens exert estrogenic effects on the central nervous system, induce estrus, and stimulate growth of the genital tract of female animals. Over 300 plants and plant products, including some used in laboratory animal diets, contain phytoestrogens. Therefore, the source and concentration of phytoestrogens in rodent diets were determined. Methods: Twelve rodent diets and six major dietary ingredients were assayed for phytoestrogens (daidzein, genistein, formononetin, biochanin A, and coumestrol), using high-performance liquid chromatography. Three rodent diets recently formulated to reduce phytoestrogen content also were assayed. Results: Formononetin, biochanin A, and coumestrol were not detected. Soybean meal was the major source of daidzein and genistein; their concentrations were directly correlated to the percentage of soybean meal in each diet. Conclusions: High, variable concentrations of daidzein and genistein are present in some rodent diets, and dietary phytoestrogens have the potential to alter results of studies of estrogenicity. Careful attention should be given to diet phytoestrogen content, and their concentration should be reported. A standardized, open- formula diet in which estrogenic substances have been reduced to levels that do not alter results of studies that are influenced by exogenous estrogens is recommended.

Phytoestrogens are defined as plant compounds that exert phytoestrogens in plants or plant products varies and is in- estrogenic effects on the central nervous system, induce estrus, fluenced by species differences, portion of plant assayed, geo- and stimulate growth of the genital tract of female animals (1). graphic location, time of harvest, and method of processing Over 300 plants and plant products, including some used in (8, 9). In general, the relative potency of the phytoestrogens laboratory animal diets, contain phytoestrogens (2–4). The two is dependent on the type of assay (in vivo or in vitro) used to main classes of phytoestrogens are the lignans and the measure estrogenic potency, animal species and age, dosage, isoflavonoids. The isoflavonoids are further divided into the age at dosing, route, and duration of exposure (6, 12, 25–27). isoflavans, isoflavones, and coumestans (4–6). Natural dietary Previous studies in our laboratory have indicated that ingredients, such as soybeans (7–11), flax (12–14), wheat, bar- commonly used rodent diets differ significantly in estrogenic ley, corn, alfalfa, and oats, are major dietary ingredients that activity, as measured by their ability to increase immature have been reported to contain phytoestrogens (3). The mouse uterine-to-body weight ratio in a 7-day bioassay (27, isoflavones (daidzein, genistein, formononetin, and biochanin 28). The isoflavonoids—daidzein, genistein, coumestrol, A) and the coumestans (coumestrol) are the major biochanin A, and formononetin—have been documented to phytoestrogens found in dietary ingredients used in many labo- be estrogenic in the mouse bioassay, as determined by sig- ratory animal diets (10, 11). Structurally, phytoestrogens re- nificantly increasing mean uterine weight of treated mice semble endogenous steroidal estrogens of humans and animals over that of control mice (20, 29, 30). Also, the in vivo effects and share a common mechanism of action through binding to of phytoestrogens are of principal interest due to their po- the estrogen receptor (6, 12, 15). It is well established that tential health risks and/or beneficial effects on humans and phytoestrogens and the estrogenic mycotoxins can bind to the animals (12, 31–33). Some phytoestrogens (17, 33, 34), when estrogen receptor and induce estrogen-like effects in ani- consumed in large amounts, have been reported to cause re- mals (16–18), humans (19), and cells in culture (6, 18). The productive problems in sheep (35, 36) and liver disease and relative potency of phytoestrogens is significantly lower infertility in captive cheetahs (37). Phytoestrogens have also than that of steroidal estrogens, and animal species differ in been reported to be beneficial by acting as partial estrogen sensitivity to the various phytoestrogens (20–23). The antagonists (12, 31, 33, 38, 39); diets containing isoflavones phytoestrogens are weak estrogens, compared with diethyl- have been documented to reduce the rate of mammary tu- stilbestrol (DES), estradiol-17␤, or estrone, in mouse mor (40–42) and liver, colon, or prostate tumor (43) formation uterotropic assays (20, 24). The concentration of the in animal models of clinical carcinogenesis. Phytoestrogens have also been documented to reduce serum cholesterol con- Comparative Medicine Branch,1 Biostatistics Branch,3 and Laboratory of centration in monkeys (44), rats (45, 46), and humans (47– Toxicology,4 National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina; and Department of Pediatric Gastroenter- 50), and to increase bone density in aged rats (51–53). The ology, Children’s Hospital Medical Center, Cincinnati, Ohio2 phytoestrogens can act as antioxidants (38, 54–56) and sup-

530 Phytoestrogens in Rodent Diets

press angiogenesis (57). Presence of 10% soybean meal in a Purified rodent diets: The following American Institute rodent diet, compared with a soy-free diet, resulted in a 40% of Nutrition (AIN) diets (63)—AIN-76A, a purified (casein) reduction in mammary tumors in two rat models of breast rodent diet; AIN-76A, a purified (soy-protein) rodent diet; cancer (41). The phytoestrogens genistein and daidzein were and AIN-93M, a purified (casein) maintenance diet—were reported to be responsible for this reduction in breast tu- purchased (Research Diets, Inc., New Brunswick, N.J.). For mors by acting as inhibitors of estrogen action or modulators future references, the diets will be identified by their vendor of hepatic metabolism (41). catalog numbers. Soy-based ingredients used in purified or natural-ingredi- Three new test diets: In the second experiment, three ent rodent diets have been reported to vary greatly in new test diets, recently formulated to contain reduced concen- isoflavone content (10, 11). The total isoflavone content in trations of phytoestrogens, were compared for phytoestrogen different varieties of soybeans varied from 116 to 309 mg/g content. Each of the three new test diets came from a differ- and varied from 46 to 195 mg/g within the same variety of ent rodent diet vendor and are identified as test diet A (ven- soybeans grown in different locations (9). The phytoestrogen dor A); test diet B (vendor B); and test diet C (vendor C). content of diets can vary greatly from batch to batch (58). We Dietary ingredients: Six commonly used dietary ingre- reported earlier that some laboratory animal diets contain dients, including whole wheat, whole oats, corn, soybean high concentrations of the phytoestrogens daidzein and meal, wheat middlings, and alfalfa meal (Zeigler Brothers, genistein (59). The biological role that phytoestrogens play Inc., Gardners, Pa.) were also assayed for phytoestrogens. in contributing to the estrogenic activity observed in rodent Phytoestrogen assays: The diets and dietary ingredients diets, and their effect on studies that are influenced by exog- were assayed for the phytoestrogens daidzein, genistein, enous estrogenic substances, is unclear. The objectives of the formononetin, biochanin A, and coumestrol. Soy flakes were as- study reported here were to determine the concentration sayed as a positive control. Dietary phytoestrogens were as- and source of the phytoestrogens daidzein, genistein, sayed, using a high-performance liquid chromatography formononetin, biochanin A, and coumestrol in rodent diets, (HPLC) procedure as described (64). Briefly, dietary estro- and to compare the phytoestrogen content in three new ro- gens were extracted from 5-g samples by prolonged reflux- dent diets recently formulated to reduce phytoestrogen con- ing in 80% ethanol. After extraction, the supernatants were tent. Results of these studies have led to recommendations cooled and filtered. A rotary evaporator was used to remove for the formulation and quality control of diets used when the bulk of ethanol from the extract. Lipids were removed by conducting studies investigating estrogenicity. liquid-liquid partitioning into hexane. The aqueous extract was taken to dryness, and a ␤-glucosidase hydrolysis was Materials and Methods carried out in 0.1 M sodium acetate buffer (pH 5.0) over- In the initial experiment, some commonly used commer- night. The hydrolysate was then passed through Bond Elut- cially available, closed- and open-formula, natural-ingredi- C8 and Bond Elut-C18 cartridges (Varian, Inc., Harbor City, ent rodent diets and some purified rodent diets were Calif.) to extract the dietary estrogens, which were recovered examined for phytoestrogens. Also, lot-to-lot variability in by elution of the cartridges with 5 ml of methanol. The daidzein and genistein concentrations within four different methanolic extract was then subjected to anion exchange chro- milling dates for four selected diets was determined. matography on a lipophilic gel, triethylaminohydroxypropyl Closed-formula, natural-ingredient rodent diets: Sephadex LH-20 (Pharmacia Fine Chemicals, Inc., Piscataway, The following rodent diets—Rodent Laboratory Chow No. N. J.), to separate neutral compounds from phenolics. The di- 5001, a maintenance diet; Certified Rodent Chow No. 5002, etary estrogens were recovered in the phenolic fraction by a maintenance diet; Mouse Chow No. 5015, a breeder diet; elution of the gel with methanol saturated with CO2 gas. and Pico Mouse Chow 20, No. 5058, an irradiated breeder This fraction was evaporated to dryness, and the extract was diet—were purchased (Purina Mills, Inc., St. Louis, Mo.). redissolved in 2 ml of methanol per 0.1 M ammonium ac- The HSD Teklad No. 7012, an autoclavable breeder and etate pH 4.6 (60/40, vol/vol) and was analyzed by use of maintenance diet, was purchased (Harlan Sprague Dawley, HPLC. The diets were analyzed “blind.” The concentration of Teklad Inc., Madison, Wis.). In closed-formula diets, the di- each phytoestrogen was determined from calibration stan- etary ingredients are reported; however, the concentration of dards and is expressed as micrograms per gram (ppm) of the each dietary ingredient is not stated, and may vary from diet. Total daidzein and genistein concentrations were deter- batch to batch or with availability of ingredients. mined for each diet and dietary ingredient. In addition, their Open-formula, natural-ingredient rodent diets: The equivalent DES and estrone activities were calculated on following rodent diets—National Institutes of Health (NIH)- the basis of a previous report (21), which used curve-fitting 31 (60), an autoclavable breeder and maintenance diet; NIH- techniques to estimate the dosages required for each com- 07 (60), a breeder and maintenance diet (used for toxicology pound to produce a 25-mg uterus in immature (19- to 21-day and carcinogenesis studies); National Toxicology Program old) female mice after a treatment period of 4 to 6 days. The (NTP) diet, NTP-88, a maintenance diet; and NTP-2000 (61, quantities for each compound were 0.083 (DES), 1.20 (estrone), 62), a new nonpurified diet (formulated for use in rodent 8,000 (genistein), and 11,000 (daidzein) ␮g. These quantities toxicology and carcinogenesis studies)—were purchased should be considered approximations because a different tar- (Zeigler Brothers, Inc., Gardners, Pa.). In open-formula diets, geted uterine weight would result in different values (21). all dietary ingredients and their concentrations are re- Mycotoxin assays: In addition, the diets were assayed ported, and should not vary from batch to batch. for the mycotoxins—aflatoxins B1, B2, G1, and G2; ochratoxin

531 Vol 49, No 5 Laboratory Animal Science October 1999

A; trichothecene (T-2); zearalenone; and deoxynivalenol—by tected in the three new test diets. Also, the three test diets an independent laboratory (Romer Labs, Inc., Union, Mo.). had less than detectable amounts of mycotoxins, including Pesticide assays: The diets were assayed for >30 pesti- zearalenone, and pesticides, including organochlorine, orga- cide residues, including organochlorine, organophosphates, nophosphates, chlorinated hydrocarbons, and carbamate. carbamates, and chlorinated hydrocarbons, by an independent laboratory (Lancaster Laboratories, Inc., Lancaster, Pa.). Discussion Our results confirmed that high concentrations of the Results phytoestrogens daidzein and genistein were present in sev- Rodent diets: The phytoestrogens detected in purified, eral natural-ingredient rodent diets, and were present in the open- and closed-formula rodent diets are shown in Tables 1 AIN-76A diet if soy-protein is substituted for casein (Table and 4. The lot-to-lot variability in daidzein and genistein 1). Soybean meal was the major source of the daidzein and concentrations within four different milling dates for four genistein, and their concentration was directly correlated selected diets is shown in Table 2. Daidzein and genistein with the percentage of the soybean meal in the diets (Tables were detected in all rodent diets except test diet A and the 1, 3, and 4). These results and those of others (58) also con- AIN-76A (casein) and AIN-93M (casein) diets. The AIN-76A firm that daidzein and genistein are present in some natu- soy-protein diet contained 74 ␮g of daidzein per g and 137 ral-ingredient rodent diets at concentrations that could have ␮g of genistein per g of diet (total = 211 ␮g/g). Comparison of marked effects on in vivo end points of hormone action and daidzein and genistein concentrations, their calculated alter results from studies that investigate the estrogenic, equivalent DES and estrone activities, and the potential reproductive, or carcinogenic activity of different com- daily intake for mice and rats consuming each rodent diet pounds. The effects of the diet on induced or spontaneous are shown in Tables 1 and 4. The no. 5001 diet had the high- carcinogenesis in laboratory animals are well recognized est concentration of the phytoestrogens daidzein and (41, 65). Numerous studies have been devoted to the associa- genistein (total = 491 ␮g/g). This is equivalent to 4.3 ␮g of tion or influence of dietary intake of calories, fat, and fiber DES activity or 62.3 ␮g of estrone activity per kg (Table 1). and the nutritional effects on carcinogenesis in laboratory The phytoestrogens formononetin, biochanin A, and animals (66). Most of these studies involving the coumestrol were not detected in any of the diets. All rodent phytoestrogens have explored their potential health benefits diets contained less than detectable amounts of the myc- for humans. However, less attention has been devoted to the otoxins, including the aflatoxins B1, B2, G1, and G2; ochra- effects of dietary phytoestrogens on carcinogenesis in labo- toxin A; T-2; zearalenone; and deoxynivalenol. Also, all ratory animals. The effects of dietary fat, protein, and fiber rodent diets contained less than detectable amounts of all on mammary tumor incidence and tumor mass were com- pesticides, including organochlorine, organophosphates, pared in the TG.NK transgenic mouse with oncogene c-neu chlorinated hydrocarbons, and carbamate. (erb B2) fed the AIN-76A casein-based diet, and the NIH-07 Dietary ingredients: The concentration of phytoestrogens or NTP-2000 diet for 24 weeks (67). The high fiber content in and their calculated equivalent DES and estrone activities, the NTP-2000 diet was reported to be principally responsible measured in natural individual dietary ingredients used in the for the reduction in mammary tumors, compared with that in preparation of rodent diets, are summarized in Table 3. mice fed the AIN-76A or the NIH-07 diet. The phytoestrogen Daidzein (1,133 ␮g/g) and genistein (954 ␮g/g) were detected content of these diets or the effects of dietary phytoestrogens on at high concentrations in soybean meal. The total concentra- mammary tumors in that study were not reported (67). tion of daidzein and genistein (2,087 ␮g/g) in the soybean Animal studies have indicated reduced incidence of tumor meal is equivalent to 18.4 ␮g of DES activity or 266.7 ␮g of development associated with soybean consumption. The estrone activity per kg (Table 3). In contrast, wheat and oats beneficial effects of the phytoestrogens in animal diets may contained low concentrations of these phytoestrogens. The be due to their ability to act as antiestrogens (31). The total concentration of daidzein and genistein in the wheat mechanism of action is poorly understood but is believed to and oats was only 22.0 and 4.0 ␮g/g, respectively. Daidzein involve hormonal and nonhormonal action of the isoflavones. and genistein were not detected in the corn or alfalfa meal. Feeding a 30% soybean diet to mice protects the liver from The samples of oats and alfalfa meal contained possible es- cancer development induced by the nitrosamine precursors trogenic components that could not be identified (Table 3). dibutylamine and nitrite (43). Rats that consumed a soy- The phytoestrogens formononetin, biochanin A, and based diet developed 40% fewer mammary tumors after ad- coumestrol were not detected in any of the dietary ingredi- ministration of carcinogens, compared with rats fed a soy-free ents examined. diet (41). Also, neonatal rats injected with genistein, allowed to Three new test diets: The phytoestrogen content of mature, then exposed to dimethylbenz[a]anthracene to in- three rodent diets recently formulated to have reduced con- duce mammary carcinogenesis, had significantly (P < 0.01) centrations of phytoestrogens are shown in Table 4. Test diet increased latency in developing tumors, as well as reduced C contained a high concentration of the phytoestrogens incidence of tumors (68). The incidence of induced prostate- daidzein and genistein (294 ␮g/g of diet). In contrast, test related cancer was reduced, and the disease-free period was diet A contained less than detectable amounts of daidzein prolonged by 27% in Lobund Wistar rats fed a high- and genistein, and test diet B (20 ␮g/g) contained low isoflavone (genistein, daidzein)-supplemented soy diet, com- amounts of total daidzein and genistein (Table 4). pared with that in rats fed the same diet but low in Formononetin, biochanin A, and coumestrol were not de- isoflavones (46). The phytoestrogens can also act as antioxi-

532 Phytoestrogens in Rodent Diets

Table 1. Phytoestrogen content of purified, open- and closed-formula rodent diets Potential daily intake of daidzein and genistein Soybean meal Totala daidzein DESb Estronec Diet Rodent No. of or soy-protein Daidzein Genistein and genistein equivalent equivalent Moused Rate type diet batches (%) (␮g/g) (␮g/g) (␮g/g) (␮g/kg) (␮g/kg) (mg) (mg) Closed-formula, natural-ingredient PMI No. 5001 1 * 277 214 491 4.3 62.3 2.5 7.4 PMI No. 5002 2 * 86 73 159 1.4 20.3 0.8 2.4 PMI No. 5015 1 * 130 97 227 2.0 28.7 1.1 3.4 PMI No. 5058 4 * 80 71 151 1.3 19.4 0.8 2.3 HSDTeklad 7012 LM-485 1 * 126 134 260 2.3 33.8 1.3 3.9 Open-formula, natural-ingredient NIH-31 4 5.0 48 46 94 0.8 12.1 0.5 1.4 NIH-07 4 12.0 89 77 166 1.5 21.3 0.8 2.5 NTP-88 1 5.0 38 31 69 0.6 8.8 0.4 1.0 NTP-2000 4 5.0 53 58 111 1.0 14.5 0.6 1.7 Open-formula purified AIN-76A (casein) 2 – ND ND ND ND ND ND ND AIN-76A (soy) 1 20† 74 137 211 2.0 28.6 1.1 3.2 AIN-93M 1 – ND ND ND ND ND ND ND Soy flakes (control) 1 – 697 785 1, 482 13.4 193.8 7.4 22.2 *Closed-formula diet: soybeans present, percentage not reported. †Pure isolated soy protein. ND = not detectable (limit of detection, 5 ␮g/g). aTotal genistein and daidzein (ppm of whole diet); 11 mg of daidzein is equivalent to 8 mg of genistein. b8 mg of genistein is equivalent to 0.083 ␮g of diethylstilbestrol (DES) activity (21). c8 mg of genistein is equivalent to 1.20 ␮g of estrone activity; minimal dosages reported (21) to increase uterine weight in an immature mouse from 9.6 mg to 25 mg in 4 to 6 days. dAssumes an adult mouse consumes 5.0 g of diet/day. Table 3. Phytoestrogen content of the major natural dietary eAssumes an adult rat consumes 15.0 g of diet/day. ingredients used in rodent diets Totala Table 2. Lot-to-lot variability in daidzein and genistein concentrations daidzein within four different milling dates for four selected diets and DESb Estronec Phytoestrogen Dietary Daidzein Genistein genistein equivalent equivalent concentration NIH-07 NIH-31 NTP-2000 PMI No. 5058 ingredient (␮g/g) (␮g/g) (␮g/g) (␮g/kg) (␮g/kg) Daidzein (␮g/g) Soy flakes (control) 697.0 785.0 1,482.0 13.4 193.8 * Mean 89 48 53 80 Soybean meal 1,133.0 954.0 2,087.0 18.4 266.7 SD 30 15 21 23 Whole wheat 4.0 18.0 22.0 0.22 3.1 * Range 55–124 33–65 25–77 54–108 Whole oats 2.0 2.0 4.0 0.04 0.5 Genistein (␮g/g) Wheat middlings 0.5 0.0 0.5 0.0 0.05 † Mean 77 46 58 71 Alfalfa meal 0.0 0.0 0.0 0.0 0.00 SD 30 20 24 9 Whole corn 0.0 0.0 0.0 0.0 0.00 Range 38–104 28–69 31–82 59–80 aTotal genistein and daidzein (ppm of whole diet); 11 mg of daidzein is Total daidzein & equivalent to 8 mg of genistein. genistein (␮g/g) b8 mg of genistein is equivalent to 0.083 ␮g of DES activity (21). Mean 166 94 111 151 cEstrogenic activity of 8 mg of genistein is equivalent to 1.20 ␮g of es- SD 56 34 45 29 trone activity; minimal dosages reported (21) to increase uterine weight Range 115–228 61–134 56–159 113–179 in an immature mouse from 9.6 mg to 25 mg in 4 to 6 days. *Samples contained components that could not be identified. †Sample contained a major compound that was not identified. dants (55). Diethylnitrosamine and phenobarbital induced hepatocarcinogenesis in female F344/N rats. Consumption on the modulation and interpretation of research data from of soybean isoflavone extract at 240 or 280 mg/kg of diet nor- studies that are influenced by exogenous estrogens require malized the activity of hepatic glutathione peroxidase, an careful consideration. The role of rodent diets in carcinoge- antioxidant enzyme suppressed by phenobarbital (55). nicity studies is unclear (66, 67) and needs further evalua- Phytoestrogens have the ability to act as partial estrogen tion. Over the years, substantial improvements have been agonists by increasing the uterine weight of mice (21, 22) or made in the quality of animal research due to the use of dis- rats (34) over that of control animals. It was recently re- ease-free animals and, recently, the use of transgenic mice ported that a rodent diet containing high amounts of (69, 70). Now, appropriate attention should be given to the daidzein (14 mg/kg) and genistein (210 mg/kg) induced a proper selection and use of the most appropriate diet to near-maximal uterotropic response in control and ovariecto- achieve the objective(s) of the particular study. mized, 30-day-old Sprague Dawley rats and altered their The three natural-ingredient test diets recently formu- normal response to administered estradiol (58). It is well lated to reduce phytoestrogen concentration varied greatly known that phytoestrogens, at low or high amounts, may in actual phytoestrogen content (Table 4). Soybean and al- compete for uterine estrogen receptor binding sites (6, 15) falfa meals were omitted from test diet A, and this diet sub- and thus have the potential, even at low concentrations, to sequently had undetectable amounts of daidzein and alter results of studies that are influenced by exogeneous genistein. The results from the three test diets indicate that estrogens. The presence of multiple phytoestrogens in ro- natural-ingredient diets can be formulated to contain less dent diets and the potential effects of these phytoestrogens than detectable amounts (<5 ␮g/g) of daidzein and genistein.

533 Vol 49, No 5 Laboratory Animal Science October 1999

Table 4. Comparative phytoestrogen content of three rodent diets recently formulated to reduce the concentration of phytoestrogens Potential daily intake of daidzein and genistein Soybean Totala daidzein DESb Estronec Test meal Daidzein Genistein and genistein equivalent equivalent Moused Rate diets (%) (␮g/g) (␮g/g) (␮g/g) (␮g/kg) (␮g/kg) (mg) (mg) A – < 5.0 < 5.0 < 5.0 – – < 0.03 < 0.08 B – 10.0 10.0 20.0 0.2 2.6 0.1 0.3 C * 133.56 160.32 293.88 2.7 38.6 1.5 4.4 *Closed-formula diet: soybeans present, percentage not reported. aTotal genistein and daidzein (ppm of whole diet); 11 mg of daidzein is equivalent to 8 mg of genistein. bEstrogenic activity of 8 mg of genistein is equivalent to 0.083 ␮g of DES activity (21). cEstrogenic activity of 8 mg of genistein is equivalent to 1.20 ␮g of estrone activity; minimal dosages reported (21) to increase uterine weight in an immature mouse from 9.6 mg to 25 mg in 4 to 6 days. dAssumes an adult mouse consumes 5.0 g of diet/day. eAssumes an adult rat consumes 15.0 g of diet/day. These data suggest that diets with minimal estrogenic activ- References ity, similar to test diet A, should be considered for use in re- 1. Lieberman, S. 1996. Are the differences between estradiol search and testing studies that might be affected by dietary and other estrogens, naturally occurring or synthetic, merely estrogens. When conducting such studies, it is advisable to semantical? J. Clin. Endocrinol. Metab. 81:850. monitor diets for microbial and chemical contaminants, 2. Colborn, T. 1995. Environmental estrogens: health implica- phytoestrogens, and mycoestrogens, including zearalenone tions for humans and wildlife. EHP 103(7S):135–136. 3. Farnsworth, N. R., A. S. Bingel, G. A. Cordell, et al. 1975. and estrogenic pesticides, to ensure that the diet does not Potential value of plants as sources of new antifertility agents. contain amounts of estrogenic substances that might alter J. Pharmacol. Sci. 64:717–754. research data. If estrogenic substances are detected, their 4. Price, K. R., and G. R. Fenwick. 1985. Naturally occur- concentration(s) should be reported. ring oestrogens in foods—a review. Food Addit. Contam. Additional studies are necessary to more clearly under- 2:73–106. 5. Setchell, K. D. R. 1985. Naturally occurring non-steroidal stand the biological role that the phytoestrogens play in the estrogens of dietary origin, p. 69–85. In J. A. McLachlan (ed.), total estrogenic activity observed in rodent diets. The ad- Estrogens in the environment II. Influences on development. verse or beneficial effects of the dietary phytoestrogens on Elsevier, N.Y. animal reproduction or on hormone-dependent disease 6. Setchell, K. D. R., and H. Adlercreutz. 1988. Mammalian lignans and phyto-oestrogens: recent studies on their forma- states are not clearly understood. It is clear from our studies tion, metabolism and biological role in health and disease, that some rodent diets contain high amounts of daidzein and p. 315–345. In I. Rowland (ed.), Role of the gut flora in toxic- genistein that may significantly alter results of studies that are ity and cancer. Academic Press, Ltd., London. influenced by exogenous estrogens. Although phytoestrogens 7. Coward, L., N. C. Barnes, K. D. R. Setchell, et al. 1993. ␤ are weakly estrogenic, compared with DES or estradiol-17␤ Genistein, daidzein, and their -glycoside conjugates: antitu- mor isoflavones in soybean foods from American and Asian in uterotropic assays (20) and in in vitro assays, such as the diets. J. Agric. Food Chem. 41:1961–1967. estrogen receptor-transfected HeLa cells or the MCF-7 cell 8. Eldridge, A. C. 1982. Determination of isoflavones in soy- assays (6, 18, 25), any diet containing phytoestrogens has bean flours, protein concentrates, and isolates. J. Agric. Food the potential to alter experimental results of studies that Chem. 30:353–355. 9. Eldridge, A. C., and W. F. Kwolek. 1983. Soybean are influenced by exogenous estrogens. isoflavones: effect of environment and variety on composition. In conclusion, all rodent diets tested, except test diet A, J. Agric. Food Chem. 31:394–396. the AIN-76A (casein) and the AIN-93M (casein) diets, con- 10. Murphy, P. A. 1982. Phytoestrogen content of processed soy- tain the phytoestrogens daidzein and genistein at concentra- bean products. Food Technol. 36:60, 62–64. tions that are directly correlated with the soybean meal 11. Murphy, P. A., E. Farmakalidis, and L. D. Johnson. 1982. Isoflavone content of soya-based laboratory animal diets. Food content of the diets. Rodent diets containing high amounts of Chem. Toxic. 20:315–317. daidzein and genistein may alter results of studies involving 12. Adlercreutz, H., B. R. Goldin, S. L. Gorbach, et al. 1995. estrogenicity. We recommend that careful consideration be Soybean phytoestrogen intake and cancer risk. J. Nutr. given to the phytoestrogen content of the diet when conduct- 125:757–770S. 13. Thompson, L. U., and M. Serraino. 1990. Lignans in flaxseed ing studies that are influenced by exogenous estrogens. A and breast and colon carcinogenesis, p. 30–35. In Proceedings of standardized, open-formula diet, in which estrogenic sub- the Flax Institute of the U.S. Fargo, North Dakota. stances have been reduced to minimal amounts or ideally 14. Kurzer, M. S., J. W. Lampe, M. C. Martini, et al. 1995. eliminated, is recommended for use when conducting such Fecal lignan and isoflavonoid excretion in premenopausal studies. This diet should be monitored for estrogenic sub- women consuming flaxseed powder. Cancer Epidemiol. Biomarkers Prev. 4:353–358. stances, and their concentrations should be reported. 15. Kurzer, M. S., and X. Xu. 1997. Dietary phytoestrogens. Annu. Rev. Nutr. 17:353–381. 16. Shutt, D. A., and R. I. Cox. 1972. Steroid and phytoestrogen binding to sheep uterine receptors in vitro. J. Endocrinol. Acknowledgements 52:299–310. We thank Patricia Deese for excellent help in preparing the 17. Welshons, W. V., G. E. Rottinghaus, D. J. Nonneman, et al. manuscript, and Drs. Frank Kari, Michael D. Shelby, and Joe J. 1990. A sensitive bioassay for detection of dietary estrogens Knapka for their time and helpful comments during the review of in animal feeds. J. Vet. Diagn. Invest. 2:268–273. the manuscript.

534 Phytoestrogens in Rodent Diets

18. Tang, B. Y., and N. R. Adams. 1980. Effect of equol on oestro- 41. Barnes, S., C. Grubbs, K. D. R. Setchell, et al. 1990. Soy- gen receptors and on synthesis of DNA and protein in the imma- beans inhibit mammary tumors in models of breast cancer. Clin. ture rat uterus. J. Endocrinol. 85:291–297. Biol. Res. 347:239–253. 19. Cassidy, A., S. Bingham, and K. D. R. Setchell. 1994. Bio- 42. Barnes, S., G. Peterson, C. Grubbs, et al. 1994. Potential role logical effects of a diet of soy protein rich in isoflavones on the of dietary isoflavones in the prevention of cancer. Adv. Exp. Med. menstrual cycle of premenopausal women. Am. J. Clin. Nutr. Biol. 354:135–147. 60:333–340. 43. Fitzsimmons, J. T. R., N. V. Orson, A. A. El-Aaaser. 1989. 20. Bickoff, E. M., A. N. Booth, A. L. Livingston, et al. 1959. Effects of soybeans and ascorbic acid on experimental Determination of estrogenic activity in fresh and dried forage. carcinogenisis. Comp. Biochem. Physiol. 93A:285–290. J. Anim. Sci. 18:1001–1009. 44. Anthony, M. S., T. B. Clarkson, C. L. Hughes, et al. 1996. 21. Bickoff, E. M., A. L. Livingston, A. P. Hendrickson, et al. Soybean isoflavones improve cardiovascular risk factors with- 1962. Relative potencies of several estrogen-like compounds found out affecting the reproductive system of peripubertal rhesus in forages. J. Agric. Food Chem. 10:410–412. monkeys. J. Nutr. 126:43–50. 22. Farmakalidis, E., J. N. Hathcock, and P. A. Murphy. 1985. 45. Lu, Y., and M. Jian. 1997. Effects of soy protein and casein on Oestrogenic potency of genistin and daidzin in mice. Food Chem. lipid metabolism in mature and suckling rats. Nutr. Res. 17: Toxicol. 23:741–745. 1341–1350. 23. Farmakalidis, E., and P. A. Murphy. 1985. Isolation of 6”-o- 46. Pollard, M., and P. Luckert. 1997. Influence of isoflavones in acetylgenistin and 6”-o-acetyldaidzin from toasted defatted soy protein isolates on development of induced prostate-related soyflakes. J. Agric. Food Chem. 33:385–389. cancers in L-W rats. Nutr. Cancer 28:41–45. 24. Newbold, R. 1995. Cellular and molecular effects of develop- 47. Anderson, J. W., B. M. Johnstone, and M. E. Cook-Newell. mental exposure to diethylstilbestrol: implications for other en- 1995. Meta-analysis of the effects of soy protein intake on serum vironmental estrogens. EHP 103(7):83–87. lipids. N. Engl. J. Med. 333:276–282. 25. Safe, S. H. 1995. Environmental and dietary estrogens and hu- 48. Carroll, K. K. 1991. Review of clinical studies on cholesterol- man health: is there a problem? EHP 103(4):346–351. lowering response to soy protein. J. Am. Diet Assoc. 91:820–827. 26. Levy, J. R., K. A. Faber, L. Ayyash, et al. 1995. The effect of 49. Carroll, K. K., and E. M. Kurowska. 1995. Soy consumption prenatal exposure to the phytoestrogen genistein on sexual dif- and cholesterol reduction: review of animal and human studies. ferentiation in rats. Proc. Soc. Exp. Biol. Med. 208:60–66. J. Nutr. 125:594–597S. 27. Thigpen, J. E., L. A. Li, C. B. Richter, et al. 1987. The mouse 50. Sirtori, C. R., R. Evan, and M. R. Lovati. 1993. Soybean pro- bioassay for the detection of estrogenic activity in rodent diets. tein diet and plasma cholesterol: from therapy to molecular II. Comparative estrogenic activity of purified, certified and stan- mechanisms. Ann. N.Y. Acad. Sci. 676:188–201. dard open and closed formula rodent diets. Lab. Anim. Sci. 51. Arjmandi, B. H., L. Alkel, B. W. Hollis, et al. 1996. Dietary 37:602–605. soybean protein prevents bone loss in an ovariectomized rat 28. Thigpen, J. E., E. H. Lebetkin, M. L. Dawes, et al. 1987. The model of osteoporosis. J. Nutr. 126:161–167. mouse bioassay for the detection of estrogenic activity in rodent 52. Ishida, H., T. Uesugi, K. Toda, et al. 1998. Preventive effects diets. III. Stimulation of uterine weight by dextrose, sucrose, and of the plant isoflavones, daidzin and genistin, on bone loss in corn starch. Lab. Anim. Sci. 37: 606–609. ovariectomized rats fed a calcium-deficient diet. Biol. Pharm. Bull. 29. Drane, H. M., D. S. Patterson, B. A. Roberts, et al. 1975. The 21:62–66. chance discovery of oestrogenic activity in laboratory rat cake. 53. Kalu, D. N., E. J. Masoro, B. P. Yu, et al. 1988. Modulation of age Food Cosmet. Toxicol. 13:491–492. related hyperparathyroidism and senile bone loss in Fischer rats by 30. Drane, H. M., D. S. Patterson, B. A. Roberts, et al. 1980. soy protein and food restriction. Endocrinology 122:1847–1854. Oestrogenic activity of soya-bean products. Food Cosmet. Toxicol. 54. Hendrich, S., K. W. Lee, X. Xu, et al. 1994. Defining food com- 18:25–426. ponents as new nutrients. J. Nutr. 124:1789–1792S. 31. Barrett, J. 1996. Phytoestrogens. Friends or foes? EHP 55. Lee, K. W., H. J. Wang, P. A. Murphy, et al. 1995. Soybean 104(5):478–482. isoflavone extract suppresses early but not later promotion of 32. Milligan, S. R., A. V. Balasubramanian, and J. C. Kalita. hepatocarcinogenesis by phenobarbital in female rat liver. Nutr. 1998. Relative potency of xenobiotic estrogens in an acute in vivo Cancer 24:267–278. mammalian assay. EHP 106(1):23–26. 56. Wei, H., R. Bowen, Q. Cai, et al. 1995. Antioxidant and 33. Verdeal, K., R. R. Brown, T. Richardson, et al. 1980. Affinity antipromotional effects of the soybean isoflavone genistein. Proc. of phytoestrogens for estradiol-binding proteins and effect of Soc. Exp. Biol. Med. 208:124–130. coumestrol on growth of 7,12-dimethylbenz(a) anthracene-in- 57. Fostis, T., R. Heikkinen, H. Adlercreutz, et al. 1982. Capil- duced rat mammary tumors. J. Natl. Cancer Inst. 64:285–290. lary gas chromatographic method for the analysis of lignans in 34. Santell, R. C., Y. C. Chang, M. G. Nair, et al. 1997. Dietary human urine. Clin. Chim. Acta 121:361–371. genistein exerts estrogenic effects upon the uterus, mammary 58. Boettger-Tong, H., L. Murthy, C. Chiappetta, et al. 1998. A gland and the hypothalamic/pituitary axis in rats. Am. Soc. Nutr. case of a laboratory animal feed with high estrogenic activity Sci. 263–269. and its impact on in vivo responses to exogenously administered 35. Bennetts, H. W., E. J. Underwood, and F. L. Shier. 1946. A estrogens. EHP 106(7):369–373. specific breeding problem of sheep on subterranean clover pas- 59. Thigpen, J. E., J. Locklear, G. F. Caviness, et al. 1992. Con- tures in western Australia. Aust. J. Agric. Res. 22:131–138. centration, source, and role of plant hormones (phytoestrogens) 36. Schinckel, P. G. 1948. Infertility in ewes grazing subterranean in laboratory animal diets. Contemp. Top. Lab. Anim. Sci. 31:31. clover pastures. Observations on breeding behavior following 60. Knapka, J. J. 1983. Nutrition, p. 51–67. In H. L. Foster, J. D. transfer to ‘sound’ country. Aust. Vet. J. 24:289–294. Small, and J. G. Fox (ed.), The mouse in biomedical research. 37. Setchell, K. D. R., S. J. Gosselin, M. B. Welsh, et al. Academic Press, Inc., New York. 1987. Dietary estrogens—a probable cause of infertility and 61. Rao, G. N. 1996. New diet (NTP-2000) for rats in the National liver disease in captive cheetahs. Gastroenterology 93: Toxicology Programs toxicity and carcinogenicity studies. 225–233. Fundam. Appl. Toxicol. 32:102–108. 38. Wang, H. J., and P. A. Murphy. 1994. Isoflavone content in 62. Rao, G. N. 1997. New nonpurified diet (NTP-2000) for rodents commercial soybean foods. J. Agric. Food Chem. 42:1666–1673. in the National Toxicology Program’s toxicology and carcinogen- 39. Wang, T. T. Y., N. Sathyamoorthy, and J. M. Phang. 1996. esis studies. J. Nutr. 127:842S–846S. Molecular effects of genistein on estrogen receptor mediated 63. Reeves, P. G., F. H. Nielsen, and G. C. Fahey, Jr. 1993. AIN- pathways. Carcinogenesis 17:271–275. 93 purified diets for laboratory rodents: final report of the Ameri- 40. Barnes, S. 1995. Effects of genistein on in vitro and in vivo can Institute of Nutrition ad hoc committee on the reformula- models of cancer. J. Nutr. 125:777S–783S. tion of the AIN-76A rodent diet. J. Nutr. 123:1939–1951.

535 Vol 49, No 5 Laboratory Animal Science October 1999

64. Setchell, K. D. R., and M. B. Welsh. 1987. High-performance 68. Lamartiniere, C. A., J. Moore, M. Holland, et al. 1995. liquid chromatographic analysis of phytoestrogens in soy pro- Neonatal genistein chemoprevents mammary cancer. Proc. Soc. tein preparations with ultraviolet, electrochemical, and Exp. Biol. Med. 208:120–123. thermospray mass spectrometric detection. J. Chromatogr. 69. Tennant, R. W., J. E. French, and J. W. Spalding. 1995. 386:315–323. Identifying chemical carcinogens and assessing potential risk 65. Rogers, A. E. 1997. Diet and breast cancer: studies in labo- in short-term bioassays using transgenic mouse models. EHP ratory animals. J. Nutr. 127:933S–935S. 103(10):942–949. 66. Rogers, A. E., S. H. Zeisel, and J. Groopman. 1993. Diet 70. Barrett, J. C. 1994. Editorial: prevention of environmentally and carcinogenesis. Carcinogenesis 14:2205–2217. related diseases. EHP 102(10):812–813. 67. Rao, G. N., E. Ney, and R. A. Herbert. 1997. Influence of diet on mammary cancer in transgenic mice bearing an oncogene expressed in mammary tissue. Breast Cancer Res. Treat. 45:149–158.

536