MCF-7 human breast cancer cell bioassay to determine estrogenic activity in feed
We measured the estrogenic activity in the feed by solvent (methanol) extraction and examination of the degree to which the extract stimulated proliferation of human MCF-7 breast cancer cells; this bioassay provides a highly sensitive method for detecting estrogenic activity in feed (Welshons et al. 1990
). Use of this bioassay to screen feeds for estrogenic activity prior to use is necessary because there can be significant batch-to-batch variability in soy isoflavones in soy-based feeds, as well as variability in other estrogenic contaminants in feeds that do not contain soy. Briefly, approximately 10,000 cells were seeded per well of a 24-well plate on day 0 in phenol red-free (estrogen-free) medium, fed on day 1 with the same medium, and then treated on days 3–6 with the test media, with daily medium changes. Test media for feeds contained methanol extracts from feed. On day 7 the wells were washed twice with 1 mL Hanks’ balanced salt solution, and each well was then assayed for DNA content.
Total estrogenic activity is expressed as parts per million equivalents of a reference, genistein, which is the primary phytoestrogen in soy, even though the total estrogenic activity is due to other estrogenic compounds in addition to genistein. The standard curve was based on triplicate assays, and extracted samples were each added to three wells containing cells. When apparent estrogenic activity was observed, the sample was run a second time to confirm the initial results and to examine whether proliferation occurred with and without addition of an antiestrogen (ICI 182,780) to confirm an estrogenic mechanism for stimulation of proliferation. In all samples tested in this experiment, the antiestrogen inhibited all proliferation observed above estrogen-free controls that did not contain any extract from feed.
We found very low estrogenic activity for the batch of PMI 5K96 feed (expressed as genistein equivalent units; 3.9 ppm) compared with the batches of PMI 5008 (40.0 ppm) and PMI 5001 (25.8 ppm). Because the PMI 5K96 feed is not completely free of estrogenic activity, it is referred to as a low-phytoestrogen feed, although the source of the estrogenic activity was not identified.
Serum estradiol in pregnant females and fetuses
A separate group of 13 females fed PMI 5K96 and 10 females fed PMI 5008 were killed on gestation day (GD) 18 for analysis of serum estradiol levels. These females were paired with stud males for 4 hr each day at the end of the dark phase of the light:dark cycle and then examined for vaginal plugs, which was GD0. Females from each feed group were housed three per cage, and fetuses were delivered on GD18 (1 day before parturition). Maternal and fetal blood samples were collected for measurement of estradiol. Blood from males and from females within each litter was pooled, with each litter yielding one value for male and one value for female fetuses.
On GD18 (1 day before parturition and during differentiation of pre-adipocytes and the reproductive system), serum estradiol levels in fetuses whose mothers were fed PMI 5K96 feed were significantly elevated relative to levels in the fetuses whose mothers were maintained on PMI 5008 feed (). We found no differences in serum estradiol levels in the pregnant females on the PMI 5008 and PMI 5K96 feeds, suggesting that the feed affected estradiol production by the fetuses, not the mothers. The PMI 5K96 feed, with little estrogenic activity and casein rather than soy protein with isoflavones, thus had the paradoxical effect of increasing exposure of mouse fetuses to the potent endogenous estrogen estradiol.
Mean (+ SE) serum estradiol (pg/mL) on GD18 in pregnant females and fetuses. Blood from males and from females within each PMI 5K96C litter (n = 13) and PMI 5008 litter (n = 10) was pooled.
Postnatal body weight, abdominal fat, and serum leptin
To determine the rate of postnatal growth, we weighed one randomly selected male and female from each litter (identified by a toe clip) at birth and weaning and in adulthood; weighing pups every day alters their postnatal growth (Ruhlen RL, unpublished data). Pups produced by PMI 5008-fed females were significantly heavier at birth than pups produced by PMI 5K96-fed females (). There was no effect of feed on body weight on PND20 (weaning), whereas on PND90 (adulthood), the PMI 5K96-fed males and females were significantly heavier than PMI 5008/5001-fed animals of the same sex; adult males fed PMI 5K96 weighed 11% more than males fed PMI 5008/5001, and females fed PMI 5K96 weighed 27% more than females fed PMI 5008/5001. There was thus a reversal of the effect of feed type on prenatal growth, which was significantly increased by high-phytoestrogen–containing feed, and post-weaning growth, which was reduced by high-phytoestrogen–containing feed. Males fed PMI 5008/5001, but not PMI 5K96, were significantly heavier than females on the same feed.
Mean (+ SE) body weight of female pups (A,C,E; n = 18 PMI 5008/5001; n = 19 PMI 5K96C) and male pups (B,D,F; n = 18 PMI 5008/5001; n = 19 PMI 5K96C) at birth (PND1; E,F), weaning (PND20; C,D), and adulthood (PND90; A,B).
On PND90 we examined the fat pads associated with the gonads (epididymal fat pads in males and parametrial fat pads in females) and with the kidneys (renal fat pads) in males and females, which together represent abdominal fat in mice. Mice fed PMI 5K96 throughout life were obese compared with those fed PMI 5008/5001 (). Gonadal and renal fat pads from males fed PMI 5K96 weighed 93% and 115%, respectively, more than those of males fed PMI 5008/5001. Gonadal and renal fat pads from females fed PMI 5K96 weighed 126% and 86%, respectively, more than those of females fed PMI 5008/5001.
Mean (+ SE) weight of gonadal (A) and renal (B) fat pads from 3-month-old males and females (PMI 5008/5001, n = 10; PMI 5K96C, n = 9).
Serum leptin was increased 121% in males and 174% in females fed PMI 5K96 compared with males and females fed PMI 5008/5001 (). Leptin is produced by adipocytes, and serum leptin was significantly correlated (r = 0.79; p < 0.001) with total weight of fat pads.
Mean (+ SE) serum leptin from 3-month-old males (n = 7/group) and females (PMI 5008/5001, n = 10; PMI 5K96C, n = 9).
Adult glucose tolerance
We examined glucose tolerance as previously described (Heine et al. 2000
). The same animals that were examined for body weight, fat-pad weight, and serum leptin were examined in this study, which was conducted 1 week before fat pad collection on PND90. Animals were fasted overnight (12–14 hr), and injected intra-peritoneally with 2 mg glucose per gram body weight in 0.9% saline solution. Blood glucose was measured with an Accu-Chek Glucometer (Roche, Indianapolis, IN) by tail nick. Linearity of the glucometer was verified with the Accu-Chek Linearity Test Kit (Roche).
Basal blood glucose levels and glucose clearance after a glucose injection were not different based on type of feed in females. In contrast, males fed PMI 5K96 had impaired glucose clearance compared with males fed PMI 5008/5001 (). Specifically, compared with males fed PMI 5008/5001, basal blood glucose was significantly elevated in males fed PMI 5K96 (p < 0.001); at both 60 and 90 min postglucose injection, blood glucose was significantly higher in PMI 5K96-fed males than in PMI 5008/5001-fed males.
Figure 5 Mean (± SE) blood glucose before (time 0) and 30, 60, and 90 min after an intraperitoneal injection of 2 mg/g glucose (in saline) in females (A) and males (B; n = 10/group) 1 week before sacrifice for examination of body fat and serum leptin on (more ...) Uterine response to estradiol and serum leptin in prepubertal females
The prepubertal uterotrophic assay has been shown to be influenced by exposure to estrogen during fetal–neonatal development in mice, with developmental exposure to low doses of estrogen increasing uterine responsiveness to estrogen (Newbold et al. 2004
). In addition, the timing of puberty in female mice can be accelerated by females being maintained on a diet with very high estrogenic activity after weaning (Thigpen et al. 2003
). This experiment thus provides a means of determining whether the uterine response to estradiol would be enhanced in females being fed the soy-based PMI 5001 feed containing high levels of phytoestrogens or PMI 5K96 with low-phytoestrogen levels that had resulted in elevated endogenous estradiol levels during fetal life.
Two randomly selected females per litter were weaned on PND17. The females were anesthetized with isoflurane and implanted subcutaneously with a Silastic capsule (0.62 i.d., 1.25 o.d.) that was 5 mm long between the capped ends. One female from each litter received a capsule contained 0.25 μg 17β-estradiol dissolved in 10 μL tocopherol-stripped corn oil (Fisher Scientific, Pittsburgh, PA). This dose of estradiol results in about 80% of the maximum uterine growth response that can be induced with estradiol at this age. The other females from the same litters (controls) were implanted with capsules containing only oil vehicle. The capsules had been preincubated for 24 hr in a physiologic phosphate-buffered saline solution with albumin added to stabilize the rate of estradiol release, which is then constant for the 3 days of this experiment.
When examined on PND20, uterine wet weight in females treated with the oil vehicle did not differ based on diet. However, the uteri of females fed PMI 5K96 were significantly (p < 0.05) heavier in response to estradiol stimulation compared with those of females fed PMI 5008/5001 (). Body weight was slightly, but not significantly, higher in females fed PMI 5K96 (10.50 ± 0.49 g) relative to females fed PMI 5008/5001 (9.59 ± 0.45 g; p > 0.1). Neither feed nor estradiol treatment influenced serum leptin, which was about 2.2 ng/ml.
Mean (+ SE) uterine weight on PND20 in response to a low dose of estradiol or oil vehicle via a subcutaneous Silastic implant (n = 10/group).
Body weight, uterine weight, histology, and serum leptin on PND26
In this experiment, 10 females from each diet group were weaned and weighed on PND19 and housed two or three per cage; as in other experiments, the PMI 5008 feed was replaced with PMI 5001. On PND26 (around the time that puberty would begin) the females were euthanized; blood was collected for analysis of serum leptin; body weight and uterine wet weight were measured; and uteri were fixed for histologic analysis of luminal epithelial cell height, which was measured using Image Pro software (Media Cybernetics, Bethesda, MD). These uterine assays serve as biomarkers of whether females were experiencing an increase in estrogen, indicating that puberty had begun.
Longitudinal sections (7 μm) of the right half of the uterus were mounted on poly-l-lysein–coated glass slides stained with hematoxylin and eosin and examined using an Olympus BH-2 light microscope (Olympus Corp., New Hyde Park, NY) with Image Pro. Uterine luminal epithelial cell height was determined by calculating the average of three measurements of randomly selected cells at three nonadjacent sections in one uterine horn.
At weaning on PND19, the body weight of PMI 5008 females was 10.0 ± 0.2 g and that of PMI 5K96 females was 10.1 ± 0.3 g. Compared with females on the PMI 5008/5001 diets, PND26 females fed PMI 5K96 were significantly heavier and had significantly enlarged uteri, increased epithelial cell height, and significantly elevated serum leptin. Thus, after consuming PMI 5K96 feed for 7 days after weaning, mice fed PMI 5K96 had accelerated body growth and increased serum leptin (suggesting an increase in body fat). The PMI 5K96 feed stimulated an increase in both uterine wet weight and epithelial cell height, indicative of accelerated puberty in the PMI 5K96-fed females compared with females on the high-phytoestrogen PMI 5008/5001 feed ().
Mean (+ SE) body weight (A), uterine weight (B), height of the uterine luminal epithelium (C), and serum leptin (D) in females weaned on PND19 and examined on PND26.
Onset of fertility in females
Two randomly selected females per litter were weaned on PND20 and paired with a sexually experienced male and monitored for the age at parturition. When the female was visibly pregnant, the male was removed. We recorded the age of the dam at delivery and the number, body weights, and sex ratio of pups. This procedure provides a direct measure of the timing of onset of fertility (the first ovulation and mating) as well as reproductive capacity in peripubertal female mice.
The females fed PMI 5K96 produced pups at a significantly younger age (44.7 ± 0.3 days) than the females fed PMI 5008/5001 (46.5 ± 0.6 days; p = 0.01). We found no significant difference in the number of pups produced by females on either feed, although the females fed PMI 5K96 produced slightly more pups per litter, and as a result, the litters tended to weigh more (p = 0.06).
Adult male reproductive organs
On PND90, we collected reproductive organs from male mice fed PMI 5K96 (n = 24) or PMI 5008/5001 diets (n = 22). All male reproductive organs were significantly different based on the type of feed (). The weights of the testes and Wolffian duct–derived organs (epididymides and seminal vesicles) were significantly smaller in males fed PMI 5K96 than in males fed PMI 5008/5001. In contrast, for the prostate, which differentiates from the urogenital sinus, males fed PMI 5K96 had heavier prostates than those fed PMI 5008/5001. In addition to reproductive organs, males on the PMI 5K96 feed had a smaller liver (2.08 ± 0.03 g) and a smaller right kidney (298.9 ± 6.4 mg) than PMI 5008/5001 males (liver: 2.22 ± 0.03 g, p < 0.05; kidney: 354.7 ± 6.7 mg, p < 0.05).
Figure 8 Mean (+ SE) wet weight of testes (A), epidydimides (B), seminal vesicles (C), and prostates (D) collected from 3-month-old male mice fed PMI 5K96C (n = 24) or PMI 5008/5001 (n = 22). With the exception of the seminal vesicles, comparisons are based on (more ...)