Toxicology and Applied Pharmacology 264 (2012) 335–342

Contents lists available at SciVerse ScienceDirect

Toxicology and Applied Pharmacology

j ourna l homepage: www.e lsev ie r .com/ locate /ytaap

A diet containing the soy phytoestrogen genistein causes infertility in female ratspartially deficient in UDP glucuronyltransferase

Jurgen Seppen ⁎Tytgat Institute for Liver and Intestinal Research, Academic Medical Center, S1-166, Meibergdreef 69, 1105 BK Amsterdam, The Netherlands

⁎ Fax: +31 205669190.E-mail address: j.seppen@amc.uva.nl.

0041-008X/$ – see front matter © 2012 Elsevier Inc. Allhttp://dx.doi.org/10.1016/j.taap.2012.09.013

a b s t r a c t

a r t i c l e i n f o

Article history:Received 24 July 2012Revised 12 September 2012Accepted 14 September 2012Available online 20 September 2012

Keywords:UDP glucuronyltransferaseGenisteinDetoxificationSoyPhytoestrogensFertility

Soy beans contain genistein, a natural compound that has estrogenic effects because it binds the estrogen re-ceptor with relatively high affinity. Genistein is therefore the most important environmental estrogen in thehuman diet.Detoxification of genistein is mediated through conjugation by UDP-glucuronyltransferase 1 and 2 (UGT1and UGT2) isoenzymes.Gunn rats have a genetic deficiency in UGT1 activity, UGT2 activities are not affected. Because our Gunn ratsstopped breeding after the animal chow was changed to a type with much higher soy content, we examinedthe mechanism behind this soy diet induced infertility. Gunn and control rats were fed diets with and with-out genistein. In these rats, plasma levels of genistein and metabolites, fertility and reproductive parameterswere determined. Enzyme assays showed reduced genistein UGT activity in Gunn rats, as compared to wildtype rats. Female Gunn rats were completely infertile on a genistein diet, wild type rats were fertile. Genisteindiet caused a persistent estrus, lowered serum progesterone and inhibited development of corpora lutea inGunn rats.Concentrations of total genistein in Gunn and control rat plasma were identical and within the range ob-served in humans after soy consumption. However, Gunn rat plasma contained 25% unconjugated genistein,compared to 3.6% in control rats. This study shows that, under conditions of reduced glucuronidation, dietarygenistein exhibits a strongly increased estrogenic effect. Because polymorphisms that reduce UGT1 expres-sion are prevalent in the human population, these results suggest a cautionary attitude towards the con-sumption of large amounts of soy or soy supplements.

© 2012 Elsevier Inc. All rights reserved.

Introduction

Legumes, especially soy beans, are rich in isoflavones with rela-tively strong estrogenic activity (Dixon, 2004; Kurzer and Xu, 1997).These phytoestrogens bind estrogen receptors α and β and thus me-diate estrogenic effects (Dixon, 2004; Kurzer and Xu, 1997). Becausesoy contains high concentrations of the phytoestrogens genistin anddaidzin, soy products are the most important source of environmen-tal estrogens in the human diet.

Soy plants synthesize glycosylated genistin and daidzin, these arerapidly deconjugated by gut micro-organisms to the bioactive agly-cones genistein and daidzein (Dixon, 2004; Kurzer and Xu, 1997).The most potent phytoestrogen in soy is genistein, which interactswith the estrogen receptor β, with similar affinity as estradiol and es-trogen receptor α with a lower affinity than estradiol (Kuiper et al.,1998).

Although the health benefits of soy and soy isoflavones are widelyadvertised and soy supplements and high dose genistein tablets are

rights reserved.

readily available, the impact of soy phytoestrogens on human healthis a controversial subject. There is evidence that soy consumption isbeneficial in the prevention of obesity and related metabolic disor-ders but clinical studies have been far from conclusive (Cederrothand Nef, 2009). Soy and phytoestrogen consumption has been linkedto reduced incidence of breast and ovarian cancers but, because of itsestrogenic activity, it has also been suggested that genistein intakemight promote growth of hormonally responsive cancers (Messinaet al., 2006; Velentzis et al., 2008).

Studies showing that high level soy phytoestrogen consumptionleads to adverse effects such as endometrial hyperplasia, bleeding andinfertility have been published (Chandrareddy et al., 2008; Unfer et al.,2004). However, other reports indicated no adverse effects of dietaryphytoestrogens; a meta analysis concluded that phytoestrogens haveno side effects but median study time was only 6.2 months and datafrom different phytoestrogens were pooled (Tempfer et al., 2009).

The effect of dietary estrogens on reproduction was first establishedin the 1940s when it was shown that sheep became infertile after graz-ing on phytoestrogen rich clover pastures (Bennets et al., 1946). It hasalso been shown that injection of high doses of genistein in neonatalmice results in irreversible infertility (Jefferson et al., 2006, 2007).

336 J. Seppen / Toxicology and Applied Pharmacology 264 (2012) 335–342

However, although phytoestrogens in rodent diets can interfere withexperimental models, infertility resulting from dietary phytoestrogenexposure is normally not observed (Jensen and Ritskes-Hoitinga,2007). Indeed, long term multigenerational breeding studies, in whichrats received high dietary doses of genistein, failed to document a signif-icant effect of dietary genistein on reproduction (National ToxicologyProgram, 2008). Thus, animal studies on the reproductive effects of en-vironmental estrogens are conflicting and the relevance of phytoestro-gen consumption for human fertility is controversial (Messina, 2010;Sirtori et al., 2005)

Themaindetoxificationpathwayof phytoestrogens is glucuronidation(Kurzer and Xu, 1997; Zhang et al., 2003). UDP glucuronyltransferases(UGT) are enzymes that catalyze the conjugation of a wide variety of en-vironmental compounds with glucuronic acid (Tukey and Strassburg,2000). After conjugation, biological activity ofmost compounds is great-ly reduced and excretion into bile or urine facilitated. The most impor-tant UGTs for drug detoxification are the UGT1 and UGT2 subfamilies(Guillemette et al., 2010; Tukey and Strassburg, 2000). From the UGT1locus, 13 different UGT isoforms with different and overlapping sub-strate specificities are generated by a mechanism of alternative splicing(Guillemette et al., 2010).Many polymorphisms that affect UGT1 isoen-zyme expression have been described (Strassburg et al., 2008). Com-plete absence of UGT1A1 is the cause of Crigler–Najjar (CN) disease,patients with this disorder have high levels of unconjugated serum bil-irubin which leads to severe brain damage when not treated (Jansen,1999).Whereas CNdisease is rare, amild variant, known as Gilbert syn-drome, is very common. Approximately 10% of the population has thiscondition which is caused by a polymorphism in the UGT1A1 promoterand results in elevated serum bilirubin levels (Bosma et al., 1995). Al-though Gilbert syndrome is considered to be benign, drugs that are me-tabolized by UGT1A1, such as irinotecan, can have severe and lifethreatening adverse effects in people with Gilbert's polymorphism(Strassburg, 2008).

The Gunn rat has no UGT1 activity, due to a frameshift mutation inthe UGT1 gene that is shared by all UGT1 isoenzymes, UGT2 activitiesare not affected (Iyanagi, 1991). Gunn rats are a substrain of Wistarrats (Roy et al., 1987).

Because the Gunn rat has a much milder phenotype than the cor-responding human disease, it has been extensively used in the devel-opment of gene and cell therapy for Crigler–Najjar and as a model forother inherited liver disorders (Chowdhury et al., 1993).

Our homozygous Gunn rats stopped breeding after a change in theanimal chow. An examination of the dietary composition of the twotypes of chow revealed that the new chow, which caused infertility,contained soy protein whereas the protein in the old chowwas main-ly of animal origin. This serendipitous observation prompted us to in-vestigate the effect of soy phytoestrogens on fertility in this model ofreduced UGT activity.

Material and methods

Animal procedures. All animal experiments were approved by theanimal experimental committee of the Academic Medical Center.

Gunn rats are a substrain of Wistar (Roy et al., 1987) and arehoused in a 12 hour light dark cycle. For all experiments, only ratsthat had reached sexual maturity and were at least 10 weeks oldwere used. For determination of plasma parameters, blood was col-lected under isoflurane anesthesia in pediatric heparin tubes fromrats that had been at least two weeks under the indicated dietary reg-imen. Breeding experiments were performed with age matched cou-ples of Wistar and Gunn rats. The breeding pairs were fed genistein orcontrol food for a period of three months and were subsequentlycrossed over to the other type of food. The number of pups was deter-mined the day after the mothers had given birth and the pups wereweaned after 22 days.

AMII food and AMII food with 300 mg/kg genistein (synthetic, LClaboratories, Woburn, USA) was obtained from Arie Blok bv, Woerden,The Netherlands. AMII food is designated control food and AMII withgenistein is designated as genistein food throughout this manuscript.

SDS, rat and mouse breeder and grower food, was obtained fromSpecial Diet Services, Witham, UK and is designated as SDS foodthroughout this manuscript.

Determination of estrous cycle. Seven female Gunn rats were fedcontrol food for three weeks, switched to genistein food for threeweeks and subsequently fed control food for two weeks.

Vaginal lavages were performed every morning with an eye drop-per containing 200 μL distilled water. The cells were spotted on a mi-croscope slide and allowed to air dry overnight. After fixing the cellsfor 10 min in 96% ethanol, the slides were stained for 10 min with0.1% Toluidine blue solution (Merck, Darmstadt, Germany), washedwith water and dried (Goldman et al., 2007). Slides were examinedby three independent observers using a Zeiss Axioskop discussion mi-croscope. Animals were considered to be in estrus when the predom-inant cell type observed was cornified epithelium (Goldman et al.,2007).

Plasma analysis. Bilirubin, aspartate aminotransferase (ASAT), andcreatinine were determined by the routine clinical chemistry labora-tory of our hospital. Progesterone was measured with a commercialradioimmunoassay kit (Orion Diagnostica, Espoo, Finland); functionalsensitivity: 1 nmol/L; intra-assay variation: 6.1% at 7 nmol/L and4.6% at 53 nmol/L. Inter-assay variation: 10.5% at 7 nmol/L and 7.1%at 53 nmol/L. Cross reactivity with corticosterone b1% and with17-α-hydroxyprogesterone and pregnenolone b0.05%. Rat steroid freeplasma was used to dilute the standard.

Genistein in rat plasma was determined by high pressure liquidchromatography (HPLC) as follows. Rat plasma, 40 μL, was acidifiedby the addition of 5 μL 100 mM 4-morpholineethanesulfonic acid(MES) pH 5.0. For the determination of unconjugated genistein, 5 μLPBS was added and protein was precipitated by the addition of100 μL methanol and centrifugation at 8000 g for 2 min. By spikingcontrol rat serum with known amounts of genistein, we determinedthat with this precipitation all genistein remained in the supernatant.

For the determination of total genistein, genistein conjugateswere hydrolyzed by the addition of 2.5 μL sulfatase (Helix pomatia,10 mg/mL in phosphate buffered saline (PBS), Sigma-Aldrich ChemieB.V. Zwijndrecht, Netherlands) and 2.5 μL beta glucuronidase (Bovineliver, 10 mg/mL in PBS Sigma-Aldrich Chemie B.V. Zwijndrecht,Netherlands) and incubation at 37 °C for 20 h. Protein was precipi-tated as described above.

HPLC of deproteinized samples was performed on a Dionex(Amsterdam, The Netherlands) 3000 Ultimate HPLC, equipped witha hypersil C18 column (Thermoscientific, Etten-Leur, The Netherlands),variable wavelength detector set at 262 nM, 20 μL injection and a flowrate of 1 mL/min. The following program was used:

Solvent A: 0.1% formic acid in water. Solvent B: 0.1% formic acid inacetonitrile.Start 20% B; t=2 30% B; t=17 40% B; t=17:30 100% B; t=20:30100% B; t=21:30 20% B; t=24 20% B; end. Linear gradients wereemployed and t is in minutes.

To generate a calibration curve for quantification, genistein wasdissolved in dimethyl sulfoxide (DMSO) and diluted in 66% methanol33% water. Chromeleon (Dionex, Amsterdam, The Netherlands) soft-ware was used to integrate the chromatograms.

Genistein glucuronyltransferase assay. Isolation of microsomes fromGunn andWistar rat liver and UGT assays were performed as describedpreviously (Seppen et al., 1994). Per assay 20–100 μg (determined by

337J. Seppen / Toxicology and Applied Pharmacology 264 (2012) 335–342

BCA protein assay) microsomes were used. To precipitate protein, reac-tion mixtures were diluted at 1:1 with methanol.

Genistein and genistein glucuronides were detected by HPLCwith a p680 pump (Dionex, Amsterdam, The Netherlands), AppliedBiosystems UV detector 759 set at 262 nm and a hypersil C18 col-umn (Thermoscientific, Etten-Leur, The Netherlands), 100 μL injec-tion and a flow rate of 1 mL/min. The following program was used:

Solvent A: 0.1% formic acid in water. Solvent B: 0.1% formic acid inacetonitrile.Start 0% B; t=2 0% B; t=22 50% B; t=25 50% B; t=26 0% B; t=300% B; end. Linear gradients were employed and t is in minutes.

Genistein dilutions were made in DMSO and enzyme activity wascalculated based on the peak areas of unconjugated and conjugatedgenistein.

Identity of genistein glucuronide peakswas confirmed by performingreactionswithout the glucuronic acid donor, UDP glucuronic acid, and byhydrolysis of the reaction mixture with beta glucuronidase as describedabove.

Histochemistry. Tissues from rats were harvested, fixed and stainedwith hematoxylin and phloxin as described (Seppen et al., 2003).

Statistical analysis. Significance of differences between groups wasdetermined by ANOVA analysis using GraphPad Prism software forall plasma analysis and t-test using Excel for growth curves and fertil-ity data.

Results

Initial observation and hypothesis

Our colony of UGT1 deficient Gunn rats stopped breeding afterthey were switched from AMII food, in which the protein source ispredominantly from animal origin, to SDS food, which has a muchhigher soy content. In a pilot experiment, we determined if a soy phy-toestrogen could be the cause of this infertility. Breeding pairs ofGunn rats were fed control AMII food or AMII food with 300 mg/kgof the soy phytoestrogens genistein or daidzein. In a two month fol-low up, breeding pairs on control and daidzein food did producedtwo litters. However, the breeding pairs on genistein food werecompletely sterile. We therefore hypothesized that reduced genisteinglucuronidation increases its estrogenic effects and thus induces in-fertility. This hypothesis was subsequently investigated in moredetail.

Gunn and Wistar rats were fed control food (AMII food containingno soy phytoestrogens) and genistein food (AMII food with the addi-tion of 300 mg/kg genistein). Gunn rats were also fed SDS food whichcontains soy protein. Food consumption in 7 female Gunn rats wasmeasured during three consecutive days. Average intake of foodwas 8.2 g per day corresponding to an average intake of 15 mg genis-tein per kg body weight per day.

Plasma analysis

In the past, we have observed that a change in diet can cause a two-fold increase in the serum bilirubin of Gunn rats (Seppen et al., 2009).Because bilirubin is a strong neurotoxin, an increase of serum bilirubinby a genistein containing diet would be a trivial explanation for the ob-served infertility. We therefore collected plasma from Gunn rats fedcontrol and genistein containing food. Plasma markers for liver (aspar-tate aminotransferase, ASAT) and kidney (creatinine) damage were notdifferent between rats fed genistein or control diets, Figs. 1A and B.

As expected, plasma bilirubin of Gunn rats was elevated, comparedto normalwild type rats. Surprisingly, genistein feeding significantly re-duced plasma bilirubin in Gunn rats, Fig. 1C. Together, these data

demonstrate that genistein food does not cause toxic effects on Gunnrats that might explain the effect on fertility.

Whereas estradiol and estrone were identical in female Gunn ratson control and genistein food (not shown), a genistein diet signifi-cantly reduced serum progesterone levels, Fig. 1D.

Genistein glucuronidation is reduced in Gunn rat liver microsomes

Genistein UDP glucuronyltransferase activity was measured in livermicrosomes from female Gunn andWistar rats. Fig. 2 shows a represen-tative experimentwith genisteinUGT activity at different substrate con-centrations.Wewere unable to calculate kM and Vmax inWistar rats asLineweaver Burke and Eadie Hofstee plots revealed non MichaelisMenten kinetics. This could be explained by the contribution of UGT1and UGT2 isoenzymes to genistein glucuronidation in Wistar rats ornon Michaelis Menten behavior as has been reported for genisteinglucuronidation by UGT1A1 (Tang et al., 2009). In Gunn rats, onlyUGT2 enzymes are present and we were indeed able to determinekinetic parameters in these microsomes. The kM for genisteinglucuronidation in Gunn rats was 147±113 μM and Vmax 50 μMgenistein/mg protein/h. The inset shows the average genistein UGTactivity in Wistar and Gunn rats at 10 μM genistein as determinedfrom three independent microsome isolations (Fig. 2.)

Gunn rats have elevated levels of unconjugated serum genistein comparedto Wistar rats

Concentrations of unconjugated genistein in rat plasma were de-termined by direct injection of plasma into the HPLC system. Total ge-nistein concentrations were determined after hydrolysis of plasmaconjugates with sulfatase and glucuronidase. Plasma from Gunn andWistar rats fed control food had undetectable levels of genistein,Table 1. Because SDS food contains soy, genistein could be detectedin the plasma of Gunn rats fed SDS food.

The concentrations of total plasma genistein were significantly(p=0.0065) lower in Gunn rats fed SDS food, compared to Gunnrats fed genistein food.

However, the levels of unconjugated genistein were not signif-icantly different between Gunn rats fed SDS food or genisteinfood.

The concentrations of total plasma genistein in Gunn and Wistarrats on a genistein diet were identical, indicating similar uptake effi-ciencies. In Gunn and Wistar rats fed genistein food, most plasma ge-nistein was conjugated. However, the concentration of unconjugatedgenistein was significantly higher in Gunn rats (p=0.038), reflectingthe reduced UGT capacity in these rats, Table 1.

Genistein diet inhibits weight gain of female Gunn rats

Gunn and Wistar rats were fed control and genistein food. Weightgain of female Gunn rats was inhibited by a genistein diet, Fig. 3. Malerats were not affected (not shown). The genistein diet only marginal-ly affected the weight gain of female Wistar rats, Fig. 3.

Soy and genistein food causes infertility in female Gunn rats

To determine whether a genistein diet caused female or male in-fertility, a cross over experiments was designed. Breeding pairs ofmale Gunn rats with female Wistar rats and female Gunn rats withmale Wistar rats, were fed control food or genistein food for a periodof three months. All combinations were tested with three breedingpairs. The results are shown in Table 2.

Fertility of breeding pairs was determined during a three monthperiod. All breeding pairs were fertile except female Gunn rats andmale Wistar rats on genistein food. When crossed over from genisteinfood to control food, the breeding pairs composed of female Gunn rats

Fig. 1. Plasma parameters of Gunn rats on control and Genistein diets. Panel A; ASAT (n=6), Panel B; creatinine (n=6), Panel C; bilirubin (n=6) Panel D; progesterone (n=8).Plasma parameters were determined as described. Differences between groups were determined by ANOVA. Where significant, p values are indicated. Median values are shownwith a horizontal bar. NS, not significant.

338 J. Seppen / Toxicology and Applied Pharmacology 264 (2012) 335–342

and male Wistar rat that were infertile on genistein food all produceda litter after the switch. Average litter size was 7.7±0.3.

The breeding pairs of female Gunn rats and male Wistar rats thatwere fertile on control food were switched to genistein containingfood and did not produce any litter in a 2 month follow up. This ex-periment shows that a genistein diet causes female infertility andthat this infertility is reversible.

Fig. 2. Gunn rat microsomes have reduced genistein UDP glucuronyltransferase activityas compared toWistar microsomes. Genistein UGT activity was determined in liver mi-crosomes from female rats as described. A typical kinetic experiment is shown and theinset shows the average genistein UGT activity at 10 μM genistein concentration as de-termined in three different microsome preparations. Bars depict standard deviations.

Genistein diet causes a persistent estrous in Gunn rats

The effects of a genistein diet on the estrus cycle of female Gunnrats were measured in a cross over experiment. The animals werefirst fed control food for three weeks, followed by genistein food forthree weeks. Subsequently they were fed control food for two weeks.

During the whole period of eight weeks daily vaginal lavages ofseven Gunn rats were stained to determine the estrous cycle. An an-imal was considered to be in estrus if the majority of cells obtainedby lavage were cornified epithelial cells.

Fig. 4A shows the daily results from all seven rats. In the period oncontrol food, the animals are cycling normally, in most animals

Table 1Plasma levels of unconjugated and conjugated genistein.

Gunn rats Wistar rats

Unconjugatedgenistein

Totalgenistein

Unconjugatedgenistein

Totalgenistein

Control food b0.005 b0.005 b0.005 b0.005SDS food 0.53±0.09 0.98±0.2 ND NDGenistein food 0.72±0.3 2.9±1.3 0.10±0.05 2.8±0.1

Plasma was collected and analyzed as described. Values shown are averages of 6animals fed SDS food and of 4 animals for control and genistein food, expressed asμM genistein. Averages and standard deviations are shown. Genistein was below thedetection limit of 0.005 μM in all animals fed control food.There was no significant difference in total genistein between Gunn and Wistar rats fedGenistein food. Unconjugated genistein was significantly higher in Gunn rats fedgenistein food as compared to Wistar rats fed genistein food, P=0.038 as determinedby ANOVA.There was no significant difference in unconjugated genistein between Gunn rats fedgenistein food or SDS food. ND = not done.

Fig. 3. Genistein inhibits growth of female Gunn rats. Groups of 4, age matched, femaleGunn and Wistar rats were fed a genistein or control diet. Weight gain was recordedweekly. Values shown are averages with standard deviations. Female Gunn rats on agenistein diet had significantly reduced weight gains as compared to Gunn rats on acontrol diet. At some time points, a genistein diet also significantly reduced weightgains in Wistar rats but the difference was much smaller than that in Gunn rats. Aster-isks indicate p values. *: pb0.05, **: pb0.02, ***: pb0.01 as determined by t test.

339J. Seppen / Toxicology and Applied Pharmacology 264 (2012) 335–342

cornified epithelial cells were only observed during single days.Shortly after initiation of the genistein diet, the vaginal lavages frommost animals revealed cornified epithelial cells during prolonged pe-riods (shown by stretches of + symbols). This is evidence that the ge-nistein diet induces a persistent estrus. After the switch back tocontrol diet, normal cycling is resumed.

Genistein diet reduces corpora lutea in Gunn rats

Histology of ovaria from Gunn rats on control and genistein dietrevealed a significant (p=0.001) difference in the number and sizeof corpora lutea. In ovarian sections from Gunn rats on control diet,an average of 5.9±4 well developed corpora lutea was observed(n=4). In Gunn rats on a genistein diet, only 0.4±0.5 developed cor-pora lutea were observed per section (n=4). Fig. 4B shows the his-tology of two representative sections.

Table 2Breeding mixed pairs of Gunn and Wistar rats on genistein and control diets show thatgenistein causes reversible female infertility.

Control food Genistein food

Averagenumberof litters

Average numberof pups per litter

Averagenumberof litters

Average numberof pups per litter

Female GunnXMale Wistar

3.0±0 8.2±3 0±0 –

Female WistarXMale Gunn

3.0±0 11.2±2 2.3±1 12±2

All results depicted are from three breeding pairs per condition. Values shown areaverages and standard deviations of litters obtained during a three month follow up. Allbreeding pairs composed of a female Gunn rat with a male Wistar rat on genistein food,were infertile. All other breeding pairs were fertile.There was no significant difference between the number of litters in the fertilebreeding pairs. However, female Gunn rats had significantly lower numbers of pups,compared to female Wistar rats, p=0.022 as determined by t test.The breeding pairs of female Gunn rats and male Wistar rats, that were infertile on agenistein diet, all produced litters when switched to a control diet. The breedingpairs of female Gunn rats and male Wistar rats, that were fertile on control diet, didnot produce litters when switched to a Genistein diet.

Discussion

Many studies have addressed the health effects of soy phytoestrogens.Despite this large corpus of information, the benefits and harms of soyphytoestrogens are still debated. Because we show that genisteinglucuronidation capacity is an important determinant for its biological ef-fect, our results might contribute some direction whichmay help to nav-igate this maze of information.

This study provides several lines of evidence for a strong estrogen-ic effect of a genistein diet in rats with reduced UGT activity but not inwild type Wistar rats. The most important observation is that genis-tein only causes infertility in female Gunn rats and fertility of Wistarrats is not affected by a genistein diet. Since Wistar rats were fertileon a genistein diet, our results confirm a previous long term feedingstudy in which a higher dose of genistein feeding did not cause infer-tility in Sprague Dawley rats (National Toxicology Program, 2008).Weight gain was only inhibited in female Gunn rats fed a genisteindiet. Many studies, mainly performed in females, have shown that es-trogen is important for the maintenance of body weight: it can sup-press food intake and is important in the regulation of adiposity(Eckel, 2011). The inhibiting effect of genistein on body weight in fe-male Gunn rats but not in Wistar rats therefore shows that genisteinexhibits increased estrogenicity in female Gunn rats. Body weight ofmale Gunn rats was not changed in response to genistein. Thiscould be explained by more efficient detoxification of genistein inmale rats. Indeed, male Gunn and Wistar rats have similar genisteinUGT capacity (Wang et al., 2009) whereas this study shows that ge-nistein UGT capacity is much lower in female Gunn rats as comparedto female Wistar rats.

Gunn rats fed genistein exhibited a persistent estrus and a de-creased serum progesterone which likely was caused by a reductionin corpora lutea formation. These effects are also consistent with anincreased estrogenic effect of genistein in Gunn rats (Goldman et al.,2007; Jefferson et al., 2007).

Plasma from Gunn rats fed a genistein diet contained significantlyhigher amounts of unconjugated genistein as compared to wild typeWistar rats on the same diet. Importantly, plasma concentrations oftotal genistein were identical in Gunn rats and Wistar rats on a genis-tein diet. Since unconjugated genistein has a higher affinity for the es-trogen receptor than glucuronidated genistein (Zhang et al., 1999),the increased levels of genistein aglycone explain the stronger estro-genic effect of genistein in Gunn rats as compared to control Wistarrats.

Our study therefore shows that plasma levels of unconjugated ge-nistein and not levels of total genistein are important for its biologicaleffect.

The initial observation that led to this study was that soycontaining SDS food caused infertility in rats that were successfullybreeding on AMII food. Indeed, Gunn rats fed SDS food had detectableplasma levels of genistein whereas in Gunn rats fed AMII food, genis-tein could not be detected. Total genistein levels in Gunn rats fed SDSfood were significantly lower than those in Gunn rats fed genisteinfood. However, the levels of unconjugated genistein in SDS fedGunn rats were not significantly different from those of Gunn ratsfed genistein food. Thus, in Gunn rats fed SDS food, a larger percent-age of genistein in plasma is unconjugated than in animals fed genis-tein food. It is important to realize that genistein is administered indifferent forms in these diets. Because the SDS food contains soy, ge-nistein is present as genistin, the glycosidated form synthesized bythe soy plant. In the genistein diet, the phytoestrogen is present asthe synthetic aglycone. It has been shown that the oral availabilityand estrogenic effect of genistin, the natural form present in soy-beans, are higher than that of the synthetic aglycone genistein(Jefferson et al., 2009; Kwon et al., 2007). Our results suggest thatthe difference in the biological effect of genistein and genistin mightbe caused by more efficient metabolism of genistein.

Fig. 4. Effect of genistein diet on estrous cycle and ovarian histology. Panel A depicts the estrous cycle of 7 different Gunn rats on control and genistein diets. The dietary regimen ismarked above the panel. Days in which the animals were in estrus, as evidenced by vaginal lavages that yielded predominantly cornified epithelial cells, are marked with +. Days inwhich the rats were not in estrus are marked with −. Panel B shows representative histology of ovaria from Gunn rats on control and genistein diets. Only in the ovarium obtainedafter a control diet, many well developed corpora lutea can be observed.

340 J. Seppen / Toxicology and Applied Pharmacology 264 (2012) 335–342

Maximal plasma concentrations of genistein in our Gunn andWistarrats were less than 3 μM. A previous study showed that infants fed soyformula had average genistein plasma concentrations of 2.5 μM,where-as genistein plasma concentrations in breastfed or cowmilk formula fedinfants were at least 100 fold lower (Setchell et al., 1997).

Reported plasma genistein concentrations in adults after con-sumption of a soy meal are also in the micromolar range (Rozmanet al., 2006; Watanabe et al., 1998; Xu et al., 1994). Indeed, a recentreview concludes that concentrations of genistein below 5 μM arephysiological (Klein and King, 2007). Thus, the values observed inour experimental animals fall within the physiological range ob-served in humans.

Although plasma unconjugated genistein is higher in Gunn rats on agenistein diet, as compared to Wistar rats on the same diet, in bothstrains the majority of genistein is present in conjugated form. InGunn rats these conjugates likely are glucuronides generated by UGT2isoenzymes and sulfates generated by sulfotransferases. Because wealso show genistein UGT activity in Gunn rat liver microsomes, ourresults confirm a previous study which showed that Gunn rats

have genistein UGT activity (Wang et al., 2009) derived from UGT2isoenzymes. However, we observed a larger difference in genisteinglucuronidation capacity between Gunn rats and Wistar rats. A like-ly explanation for this difference is that this previous study usedmale rats, whereas we used female rats, and gender specific differ-ences in UGT expression levels have been reported (Shelby et al.,2003).

The UGT1A1 and UGT1A9 isoforms are the most important humanUGT1 for genistein glucuronidation (Pritchett et al., 2008), the mostimportant UGT2s for genistein glucuronidation have not been identi-fied. Since large intra individual differences in genistein metabolismhave been reported (Rozman et al., 2006), polymorphisms in UGTgenes involved in genistein metabolism (Guillemette et al., 2010;Strassburg, 2008; Strassburg et al., 2008) could contribute to thisdiversity.

There are several well described examples of polymorphisms inUGT genes that affect drug responses. Patientswith the promoter poly-morphism that leads to the reduced UGT1A1 expression in Gilbert syn-drome can experience severe side effects after irinotecan administration

341J. Seppen / Toxicology and Applied Pharmacology 264 (2012) 335–342

(Strassburg, 2008). Approximately 30% of patients treated with the HIVprotease inhibitor atazanavir develop jaundice and thiswas shown to beassociated with multiple polymorphisms in the UGT1 locus (Strassburg,2008).

Because decreased metabolism of endogenous estrogens might in-crease the risk of breast cancer development, several studies haveaddressed the question whether there is a relation between the devel-opment of breast cancer and UGT activity. Indeed, the Gilbert polymor-phisms in UGT1 were shown to be associated with mammographicbreast density (Haakensen et al., 2010; Yong et al., 2010) and risk ofbreast cancer development (Shatalova et al., 2005).

Collectively, these results demonstrate that glucuronidation ca-pacity is crucial in determining the estrogenic effect of genistein invivo. It therefore seems likely that the estrogenic effects of dietarysoy isoflavones are more pronounced in woman with polymorphismsthat lead to reduced genistein glucuronidation capacity.

Because infants do not have full expression of UGT enzymes, withadult levels for many UGT isoenzymes only reached months afterbirth (de Wildt et al., 1999), newborns might also be more vulnerableto estrogenic effects of phytoestrogen containing products such as soyinfant formula.

Conflict of interest statement

I have no competing interests to declare.

Acknowledgments

I would like to thank, Johan Hiralall, Cindy Kunne, Suzanne Duist,Jung-Chin Chang and Shin-Hao Chang for the help with the experi-ments and Harriet Klinkspoor, Rudi deWaart and Piter Bosma for crit-ically reading the manuscript and the helpful suggestions.

References

Bennets, H.W., Underwood, E.J., Shier, F.L., 1946. A specific breeding problem ofsheep on subterranean clover pastures in Western Australia. Aust. Vet. J. 22,2–12.

Bosma, P.J., Chowdhury, J.R., Bakker, C., Gantla, S., Oostra, B.A., Lindhout, D., Tytgat, G.N.,Jansen, P.L., Oude Elferink, R.P., 1995. The genetic basis of the reduced expressionof bilirubin UDP-glucuronosyltransferase 1 in Gilbert's syndrome. N. Engl. J. Med.333, 1171–1175.

Cederroth, C.R., Nef, S., 2009. Soy, phytoestrogens and metabolism: a review. Mol. Cell.Endocrinol. 304, 30–42.

Chandrareddy, A., Muneyyirci-Delale, O., McFarlane, S.I., Murad, O.M., 2008. Adverseeffects of phytoestrogens on reproductive health: a report of three cases. Comple-ment. Ther. Clin. Pract. 14, 132–135.

Chowdhury, J.R., Kondapalli, R., Chowdhury, N.R., 1993. Gunn rat: a model forinherited deficiency of bilirubin glucuronidation. Adv. Vet. Sci. Comp. Med. 37,149–173.

de Wildt, S.N., Kearns, G.L., Leeder, J.S., van den Anker, J.N., 1999. Glucuronidation inhumans. Pharmacogenetic and developmental aspects. Clin. Pharmacokinet. 36,439–452.

Dixon, R.A., 2004. Phytoestrogens. Annu. Rev. Plant Biol. 55, 225–261.Eckel, L.A., 2011. The ovarian hormone estradiol plays a crucial role in the control of

food intake in females. Physiol. Behav. 104, 517–524.Goldman, J.M., Murr, A.S., Cooper, R.L., 2007. The rodent estrous cycle: characterization

of vaginal cytology and its utility in toxicological studies. Birth Defects Res. B Dev.Reprod. Toxicol. 80, 84–97.

Guillemette, C., Levesque, E., Harvey, M., Bellemare, J., Menard, V., 2010. UGT genomicdiversity: beyond gene duplication. Drug Metab. Rev. 42, 22–42.

Haakensen, V.D., Biong, M., Lingjaerde, O.C., Holmen, M.M., Frantzen, J.O., Chen, Y.,Navjord, D., Romundstad, L., Luders, T., Bukholm, I.K., Solvang, H.K., Kristensen,V.N., Ursin, G., Borresen-Dale, A.L., Helland, A., 2010. Expression levels of uridine 5′-diphospho-glucuronosyltransferase genes in breast tissue from healthy women areassociated with mammographic density. Breast Cancer Res. 12, R65.

Iyanagi, T., 1991. Molecular basis of multiple UDP-glucuronosyltransferase isoenzyme de-ficiencies in the hyperbilirubinemic rat (Gunn rat). J. Biol. Chem. 266, 24048–24052.

Jansen, P.L., 1999. Diagnosis and management of Crigler–Najjar syndrome. Eur. J. Pediatr.158 (Suppl. 2), S89–S94.

Jefferson, W.N., Padilla-Banks, E., Newbold, R.R., 2006. Studies of the effects of neonatalexposure to genistein on the developing female reproductive system. J. AOAC Int.89, 1189–1196.

Jefferson, W.N., Padilla-Banks, E., Newbold, R.R., 2007. Disruption of the female repro-ductive system by the phytoestrogen genistein. Reprod. Toxicol. 23, 308–316.

Jefferson, W.N., Doerge, D., Padilla-Banks, E., Woodling, K.A., Kissling, G.E., Newbold, R.,2009. Oral exposure to genistin, the glycosylated form of genistein, during neona-tal life adversely affects the female reproductive system. Environ. Health Perspect.117, 1883–1889.

Jensen, M.N., Ritskes-Hoitinga, M., 2007. How isoflavone levels in common rodent dietscan interfere with the value of animal models and with experimental results. Lab.Anim. 41, 1–18.

Klein, C.B., King, A.A., 2007. Genistein genotoxicity: critical considerations of in vitroexposure dose. Toxicol. Appl. Pharmacol. 224, 1–11.

Kuiper, G.G., Lemmen, J.G., Carlsson, B., Corton, J.C., Safe, S.H., van der Saag, P.T.,Gustafsson, J.A., 1998. Interaction of estrogenic chemicals and phytoestrogenswith estrogen receptor beta. Endocrinology 139, 4252–4263.

Kurzer, M.S., Xu, X., 1997. Dietary phytoestrogens. Annu. Rev. Nutr. 17, 353–381.Kwon, S.H., Kang, M.J., Huh, J.S., Ha, K.W., Lee, J.R., Lee, S.K., Lee, B.S., Han, I.H., Lee, M.S.,

Lee, M.W., Lee, J., Choi, Y.W., 2007. Comparison of oral bioavailability of genisteinand genistin in rats. Int. J. Pharm. 337, 148–154.

Messina, M., 2010. A brief historical overview of the past two decades of soy and isofla-vone research. J. Nutr. 140, 1350S–1354S.

Messina, M., Caskill-Stevens, W., Lampe, J.W., 2006. Addressing the soy and breast can-cer relationship: review, commentary, and workshop proceedings. J. Natl. CancerInst. 98, 1275–1284.

National Toxicology Program, 2008. NTP toxicology and carcinogenesis studies of ge-nistein (CAS no. 446-72-0) in Sprague–Dawley rats (Feed Study). Natl. Toxicol.Program Tech. Rep. Ser. 1–240.

Pritchett, L.E., Atherton, K.M., Mutch, E., Ford, D., 2008. Glucuronidation of thesoyabean isoflavones genistein and daidzein by human liver is related to levels ofUGT1A1 and UGT1A9 activity and alters isoflavone response in the MCF-7human breast cancer cell line. J. Nutr. Biochem. 19, 739–745.

Roy, C.N., Gross, F., Moscioni, A.D., Kram, M., Arias, I.M., Roy, C.J., 1987. Isolation ofmultiple normal and functionally defective forms of uridine diphosphate-glucuronosyltransferase from inbred Gunn rats. J. Clin. Invest. 79, 327–334.

Rozman, K.K., Bhatia, J., Calafat, A.M., Chambers, C., Culty, M., Etzel, R.A., Flaws, J.A.,Hansen, D.K., Hoyer, P.B., Jeffery, E.H., Kesner, J.S., Marty, S., Thomas, J.A.,Umbach, D., 2006. NTP-CERHR expert panel report on the reproductive and devel-opmental toxicity of genistein. Birth Defects Res. B Dev. Reprod. Toxicol. 77,485–638.

Seppen, J., Bosma, P.J., Goldhoorn, B.G., Bakker, C.T., Chowdhury, J.R., Chowdhury, N.R.,Jansen, P.L., Oude Elferink, R.P., 1994. Discrimination between Crigler–Najjar type Iand II by expression of mutant bilirubin uridine diphosphate-glucuronosyltransferase.J. Clin. Invest. 94, 2385–2391.

Seppen, J., van der Rijt, R., Looije, N., van Til, N.P., Lamers, W.H., Oude Elferink, R.P.,2003. Long-term correction of bilirubin UDP glucuronyltransferase deficiency inrats by in utero lentiviral gene transfer. Mol. Ther. 8, 593–599.

Seppen, J., Filali, E.E., Elferink, R.O., 2009. Small animal models of hepatocyte transplan-tation. Methods Mol. Biol. 481, 75–82.

Setchell, K.D., Zimmer-Nechemias, L., Cai, J., Heubi, J.E., 1997. Exposure of infants tophyto-oestrogens from soy-based infant formula. Lancet 350, 23–27.

Shatalova, E.G., Walther, S.E., Favorova, O.O., Rebbeck, T.R., Blanchard, R.L., 2005.Genetic polymorphisms in human SULT1A1 and UGT1A1 genes associate withbreast tumor characteristics: a case-series study. Breast Cancer Res. 7, R909–R921.

Shelby, M.K., Cherrington, N.J., Vansell, N.R., Klaassen, C.D., 2003. Tissue mRNA expres-sion of the rat UDP-glucuronosyltransferase gene family. Drug Metab. Dispos. 31,326–333.

Sirtori, C.R., Arnoldi, A., Johnson, S.K., 2005. Phytoestrogens: end of a tale? Ann. Med.37, 423–438.

Strassburg, C.P., 2008. Pharmacogenetics of Gilbert's syndrome. Pharmacogenomics 9,703–715.

Strassburg, C.P., Lankisch, T.O., Manns, M.P., Ehmer, U., 2008. Family 1 uridine-5′-diphosphate glucuronosyltransferases (UGT1A): from Gilbert's syndrome togenetic organization and variability. Arch. Toxicol. 82, 415–433.

Tang, L., Singh, R., Liu, Z., Hu, M., 2009. Structure and concentration changes affect char-acterization of UGT isoform-specific metabolism of isoflavones. Mol. Pharm. 6,1466–1482.

Tempfer, C.B., Froese, G., Heinze, G., Bentz, E.K., Hefler, L.A., Huber, J.C., 2009. Side ef-fects of phytoestrogens: a meta-analysis of randomized trials. Am. J. Med. 122,939–946.

Tukey, R.H., Strassburg, C.P., 2000. Human UDP-glucuronosyltransferases: metabolism,expression, and disease. Annu. Rev. Pharmacol. Toxicol. 40, 581–616.

Unfer, V., Casini, M.L., Costabile, L., Mignosa, M., Gerli, S., Di Renzo, G.C., 2004. Endome-trial effects of long-term treatment with phytoestrogens: a randomized, double-blind, placebo-controlled study. Fertil. Steril. 82, 145–148 (quiz).

Velentzis, L.S., Woodside, J.V., Cantwell, M.M., Leathem, A.J., Keshtgar, M.R., 2008. Dophytoestrogens reduce the risk of breast cancer and breast cancer recurrence?What clinicians need to know. Eur. J. Cancer 44, 1799–1806.

Wang, S.W., Kulkarni, K.H., Tang, L., Wang, J.R., Yin, T., Daidoji, T., Yokota, H., Hu, M., 2009.Disposition of flavonoids via enteric recycling: UDP-glucuronosyltransferase (UGT)1As deficiency in Gunn rats is compensated by increases in UGT2Bs activities.J. Pharmacol. Exp. Ther. 329, 1023–1031.

Watanabe, S., Yamaguchi, M., Sobue, T., Takahashi, T., Miura, T., Arai, Y., Mazur, W.,Wahala, K., Adlercreutz, H., 1998. Pharmacokinetics of soybean isoflavones in plas-ma, urine and feces of men after ingestion of 60 g baked soybean powder (kinako).J. Nutr. 128, 1710–1715.

Xu, X., Wang, H.J., Murphy, P.A., Cook, L., Hendrich, S., 1994. Daidzein is a more bio-available soymilk isoflavone than is genistein in adult women. J. Nutr. 124,825–832.

342 J. Seppen / Toxicology and Applied Pharmacology 264 (2012) 335–342

Yong, M., Schwartz, S.M., Atkinson, C., Makar, K.W., Thomas, S.S., Newton, K.M., AielloBowles, E.J., Holt, V.L., Leisenring, W.M., Lampe, J.W., 2010. Associations betweenpolymorphisms in glucuronidation and sulfation enzymes and mammographicbreast density in premenopausal women in the United States. Cancer Epidemiol.Biomarkers Prev. 19, 537–546.

Zhang, Y., Song, T.T., Cunnick, J.E., Murphy, P.A., Hendrich, S., 1999. Daidzein and genis-tein glucuronides in vitro are weakly estrogenic and activate human natural killercells at nutritionally relevant concentrations. J. Nutr. 129, 399–405.

Zhang, Y., Hendrich, S., Murphy, P.A., 2003. Glucuronides are the main isoflavone me-tabolites in women. J. Nutr. 133, 399–404.