Genistein is a naturally occurring isoflavone that interacts with estrogen receptors and multiple other molecular targets. Human exposure to genistein is predominantly through consumption of soy products, including soy-based infant formula and dietary supplements. Consumption of soy and genistein has been associated with a variety of beneficial effects in animals and humans, but concerns have also been raised regarding potential adverse effects of genistein, particularly with regard to reproductive toxicity and the induction or potentiation of carcinogenesis, due primarily to its weak estrogenic activity. Because of these concerns, genistein was selected as one of the compounds to be examined using a protocol designed to evaluate the effects of multigenerational and long-term exposures to doses of estrogenic agents that produce subtle reproductive tract lesions in developmentally exposed Sprague-Dawley rat pups. Results from the 2-year study are reported here, and results from the multigenerational reproductive toxicology feed study are reported separately (NTP, 2008a). Data from a preliminary dose range-finding feed study (NTP, 2007) that utilized exposure concentrations up to 1,250 ppm genistein were used to select dietary exposure concentrations of 0, 5, 100, and 500 ppm for the current study. The multigenerational reproductive toxicology study examined F(0) through F(4) generations with F(5) litters terminated at weaning and focused on reproductive endpoints (NTP, 2008a). Animals were exposed from the time that the F(0) generation was 6 weeks old through weaning of the F(3) generation, and animals of the F(0) through F(4) generations were necropsied at 20 weeks of age. The current study was a 2-year dietary study utilizing three exposure arms: continuous exposure from conception through 2 years (designated F(1) continuous, or F(1)C), exposure from conception through 20 weeks followed by control diet to 2 years [designated F(1) truncated at postnatal day (PND) 140, or F(1)T140], and exposure from conception through weaning followed by control diet to 2 years (designated F(3) truncated at PND 21, or F(3)T21). The "F(3)" designation for the F(3)T21 arm indicates that these animals were siblings of the F(3) animals from the multigenerational reproductive toxicology study (NTP, 2008a). The F(1)C and F(1)T140 animals were also siblings but were derived from a separate breeding that was identical to the procedure used to produce the F(1) generation of the multigenerational reproductive toxicology study. The animals in this study were exposed to genistein during various phases of their lives from conception until termination at 2 years, and the ingested doses varied over the course of the study. During pregnancy, the ingested doses of the dams were approximately 0, 0.5, 9, or 45 mg/kg body weight per day. During lactation, the dams' ingested doses were 0, 0.7, 15, or 75 mg/kg per day. Supplementary studies, which are described in the multigenerational reproductive toxicology study, indicated minimal transfer of genistein to pups via the dams' milk. The mean directly ingested genistein doses during the period prior to PND 140 were approximately 0.4, 8, or 44 mg/kg per day for females and 0.4, 7, or 37 mg/kg per day for males. For the period between PND 140 and the end of the study, mean ingested doses were approximately 0.3, 5, or 29 mg/kg per day for females and 0.2, 4, or 20 mg/kg per day for males. For the current study, 50 animals per sex were initially assigned to each exposure group in each arm of the study. In control groups, histopathology data from one to four additional animals that had been assigned as sentinels but that became moribund or died early were also included in the analysis and presentation. Survival was similar in all control and exposed groups and ranged from 62% to 86% for males and 43% to 64% for females. Mean body weights of 500 ppm F(1)C females were less than those of the controls throughout the study. Mean body weights of 500 ppm F(1)T140 rats were less than those of the controls throughout the study. In females of all study arms (F(1)C, F(1)T140, and F(3)T21) an early onset of aberrant estrous cycles, suggesting early reproductive senescence, was observed in the 500 ppm groups. In the F(3)T21 arm, there were also significant effects on the onset of aberrant estrous cycles in the 5 and 100 ppm groups. Pituitary gland weights were significantly increased in females in the 500 ppm groups of the F(1)C and F(1)T140 study arms and in the 100 ppm group of the F(3)T21 arm. In F(1)C females, there was a significant positive trend in the incidences of mammary gland adenoma or adenocarcinoma (combined) regardless of whether an unmodified or natural log-transformed dose scale was used in the analysis, and the incidence in the 500 ppm group was significantly greater than that in the control group. A significant negative trend occurred in the incidences of benign mammary gland fibroadenoma in F(1)C females, and the incidence in the 500 ppm group was significantly less than that in the control group. In 5 and 100 ppm F(1)T140 females, the combined incidences of adenoma and adenocarcinoma were less than those in the control or 500 ppm groups, although these were not statistically significant differences. When the natural log-transformed dose scale was used, a marginally significant positive trend occurred in the incidences of adenoma or adenocarcinoma (combined) in F(3)T21 females. There were positive trends in the incidences of adenoma or carcinoma (combined) in the pars distalis of the pituitary gland of females in the F(1)C and F(1)T140 arms, and the incidence in the 500 ppm group was significantly greater than that in the controls in the F(1)C study arm. In F(1)C males, a significant positive trend (unmodified dose scale only) occurred in the incidences of combined adenoma or carcinoma of the pancreatic islets. While the incidence in the 500 ppm group was elevated relative to that in the control group (6/49 versus 1/49), this was not statistically significant. The fact that transitional lesions (i.e., hyperplasia) were not observed combined with variable control rates in males of this substrain of rats led to the conclusion that this lesion was not likely to be related to genistein treatment.
Conclusions: Under the conditions of this 2-year feed study with continuous exposure to the test compound from conception through termination (F(1)C), there was no evidence of carcinogenic activity of genistein in male Sprague-Dawley rats exposed to 5, 100, or 500 ppm. There was some evidence of carcinogenic activity of genistein in female Sprague-Dawley rats based on increased incidences of mammary gland adenoma or adenocarcinoma (combined) and pituitary gland neoplasms. The incidence of benign mammary gland fibroadenoma in female rats was significantly decreased in the 500 ppm group. Under the conditions of this 2-year feed study with exposure to the test compound from conception through 20 weeks followed by control feed until termination (F(1)T140), there was no evidence of carcinogenic activity of genistein in male Sprague-Dawley rats exposed to 5, 100, or 500 ppm. There was equivocal evidence of carcinogenic activity of genistein in female Sprague-Dawley rats based on marginally increased incidences of pituitary gland neoplasms. Under the conditions of this 2-year feed study where offspring of three prior generations of animals exposed to the test compound were exposed from conception through weaning (PND 21) followed by control feed until termination (F(3)T21), there was no evidence of carcinogenic activity of genistein in male Sprague-Dawley rats exposed to 5, 100, or 500 ppm. There was equivocal evidence of carcinogenic activity of genistein in female Sprague-Dawley rats based on increased incidences of mammary gland adenoma or adenocarcinoma (combined). Exposure to genistein was also shown to accelerate the onset of aberrant estrous cycles in female Sprague-Dawley rats whether exposures were continuous or truncated at PND 140 or at weaning. The effects of genistein on estrous cycling and the incidences of common hormonally related spontaneous neoplasms of female Sprague-Dawley rats are consistent with an estrogenic mechanism of toxicity.