environmental estrogens differentially engage the histone...

Signaling and Regulation

Environmental Estrogens Differentially Engage the HistoneMethyltransferase EZH2 to Increase Risk of UterineTumorigenesis

K. Leigh Greathouse1, Tiffany Bredfeldt1, Jeffrey I. Everitt2, Kevin Lin1, Tia Berry1,Kurunthachalam Kannan3, Megan L. Mittelstadt4, Shuk-mei Ho5, and Cheryl L. Walker1

AbstractEnvironmental exposures during sensitive windows of development can reprogram normal physiologic responses

and alter disease susceptibility later in life in a process known as developmental reprogramming. For example,exposure to the xenoestrogen diethylstilbestrol during reproductive tract development can reprogram estrogen-responsive gene expression in themyometrium, resulting in hyperresponsiveness to hormone in the adult uterus andpromotion of hormone-dependent uterine leiomyoma.We show here that the environmental estrogens genistein, asoy phytoestrogen, and the plasticizer bisphenol A, differ in their pattern of developmental reprogramming andpromotion of tumorigenesis (leiomyomas) in the uterus. Whereas both genistein and bisphenol A induce genomicestrogen receptor (ER) signaling in the developing uterus, only genistein induced phosphoinositide 3-kinase(PI3K)/AKTnongenomic ER signaling to the histonemethyltransferase enhancer of zeste homolog 2 (EZH2). As aresult, this pregenomic signaling phosphorylates and represses EZH2 and reduces levels of H3K27me3 repressivemark in chromatin. Furthermore, only genistein caused estrogen-responsive genes in the adult myometrium tobecome hyperresponsive to hormone; estrogen-responsive genes were repressed in bisphenol A–exposed uteri.Importantly, this pattern of EZH2 engagement to decrease versus increase H3K27methylation correlated with theeffect of these xenoestrogens on tumorigenesis. Developmental reprogramming by genistein promoted develop-ment of uterine leiomyomas, increasing tumor incidence andmultiplicity, whereas bisphenol A did not. These datashow that environmental estrogens have distinct nongenomic effects in the developing uterus that determines theirability to engage the epigenetic regulator EZH2, decrease levels of the repressive epigenetic histone H3K27methylmark in chromatin during developmental reprogramming, and promote uterine tumorigenesis.Mol Cancer Res;10(4); 546–57. �2012 AACR.

IntroductionThe term developmental reprogramming is used to

describe the effects of early life exposures to adverse stimulithat can alter response to normal physiologic signals and giverise to disease in adulthood. Numerous studies show thatperinatal exposure to xenoestrogens, many found ubiqui-tously in the environment, can developmentally reprogramthe female reproductive tract, causing alterations in mor-phology, hormonalmilieu, and gene expression, and give riseto diseases such as obesity and cancer (1–3). Developmentalreprogramming by in utero exposure to the xenoestrogendiethylstilbestrol is an early example of this phenomenon.Gestational exposure to diethylstilbestrol results in malfor-mations of the uterus, infertility, and vaginal cancers (e.g.,clear cell vaginal adenocarcinoma) in daughters of womenprescribed this drug during pregnancy. New evidence pub-lished in recent epidemiologic reports has also shown acorrelation between in utero diethylstilbestrol exposure andan increased risk for breast cancer (4) as well as uterineleiomyomas (5), though the latter remains somewhat con-troversial (6).

Authors' Affiliations: 1Science Park Research Division, Department ofCarcinogenesis, The University of Texas MD Anderson Cancer Center,Smithville, Texas; 2GlaxoSmithKline, Research Triangle Park, Durham,North Carolina; 3Wadsworth Center, New York State Department of Healthand Department of Environmental Health Sciences, School of PublicHealth, State University of New York at Albany, New York; 4Center forTranslational Cancer Research, Institute of Biosciences and Technology,Texas A&M Health Science Center, Houston, Texas; and 5Department ofEnvironmental Health, Center for Environmental Genetics, University ofCincinnati Medical Center, Cincinnati, Ohio

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Current address for K.L. Greathouse: National Cancer Institute, NationalInstitutes of Health, Laboratory of Human Carcinogenesis, Bethesda, MD.Current address for T. Berry and C.L. Walker: Center for TranslationalCancer Research, Institute of Biosciences and Technology, Texas A&MHealth Science Center, Houston, TX.

Corresponding Author: Cheryl L. Walker, Center for Translational CancerResearch, Texas A&M Health Science Center Institute of Biosciences andTechnology 2121 West Holcombe Blvd. Houston, TX 77030. Phone: 713-677-7450; Fax: 713-677-7725; E-mail: [email protected]

doi: 10.1158/1541-7786.MCR-11-0605

�2012 American Association for Cancer Research.

MolecularCancer

Research

Mol Cancer Res; 10(4) April 2012546

on February 3, 2019. © 2012 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

http://mcr.aacrjournals.org/

Although diethylstilbestrol is no longer in clinical use,other environmental xenoestrogens have the potential toaffect the developing reproductive tract and induce devel-opmental reprogramming. For example, bisphenol A, achemical used in the production of plastics, food canlinings, and dental sealants, can induce morphologicabnormalities of the reproductive tract in rodents exposedneonatally to this xenoestrogen (7–9). Bisphenol A caninduce precocious puberty, persistent vaginal cornifica-tion, absence of corpora lutea, cystic ovaries, cystic endo-metrial hyperplasia, and polyovular follicles in adult ani-mals exposed perinatally to this compound (7, 10, 11). Inthe male reproductive tract, neonatal exposure to environ-mentally relevant micromolar bisphenol A concentrationshas been shown to promote neoplastic transformation,including formation of prostatic intraepithelial neoplasia(12). Genistein, a phytoestrogen in soybeans that is foundin processed foods and soy-based infant formula, can alsoinduce developmental reprogramming of the female repro-ductive tract in animals. Multiple studies report thatneonatal exposure to environmentally relevant doses ofgenistein (e.g., 2.4–6.6 mmol/L in plasma; ref. 13) aber-rantly reprograms reproductive function and morphologyas evidenced by induction of ovarian and uterine mor-phologic abnormalities, persistent estrus, accelerated vag-inal opening, infertility, early reproductive senescence,multioocyte follicles, and uterine adenocarcinomas. In arecent epidemiologic study of more than 19,000 women, acorrelation was found between early-life soy formulaconsumption and increased risk of uterine leiomyomas(5), suggesting a link may also exist between environmen-tal estrogen exposure and development of these tumors inwomen.Uterine leiomyoma, commonly called fibroids, are

benign gynecologic tumors of the uterine myometrium(14). Although these hormone-dependent tumors are themost frequent gynecologic tumor of women, little isknown about how environmental exposures may contrib-ute to the high incidence of this disease (2, 15). Previousstudies from our group have shown in genetically predis-posed Eker rats that susceptibility to the development ofuterine leiomyoma is modulated by developmental expo-sure to diethylstilbestrol via developmental reprogram-ming of the reproductive tract and estrogen-responsivegene expression (16, 17). Exposure of neonatal Eker rats todiethylstilbestrol during postnatal days 3 through 12 wassufficient to increase the penetrance of the tuberoussclerosis complex 2 (Tsc2) tumor suppressor gene defectand increase spontaneous incidence of leiomyomas from65% to 100% (16, 18). In these animals, diethylstilbestrolalso reprogrammed the morphology of the reproductivetract, giving rise to persistent vaginal cornification andovaries that lacked corpora lutea. In addition, microarrayanalyses identified several estrogen-responsive genes devel-opmentally reprogrammed by neonatal diethylstilbestrolexposure, which became hyperresponsive to hormone inthe uteri of adult animals before the onset of these tumors(19).

The mechanism(s) by which xenoestrogens mightinduce developmental reprogramming are not well under-stood. In the female reproductive tract, developmentalreprogramming is mediated by estrogen receptor a (ERa),as shown in ERa knockout mice (ERKO), which areresistant to diethylstilbestrol-induced developmentalreprogramming (20). However, the role of genomic versusnongenomic ER signaling in developmental reprogram-ming has not been defined. The canonical pathway forgenomic ER signaling (i.e., transactivation of gene expres-sion) can be induced by both natural endogenous estro-gens (e.g., 17b-estradiol) as well as by xenoestrogens. Thenongenomic pathway for rapid activation of membraneER signaling is less well understood, though several studiesshow that rapid, nongenomic ER signaling via pathwayssuch as phosphoinositide 3-kinase (PI3K) and mitogen-activated protein kinase (MAPK) are important for bloodvessel vasodilatation, neuron survival, bone loss preven-tion, and reproductive function (21–24). Both endoge-nous and xenoestrogens can trigger nongenomic ER sig-naling through activation of PI3K/AKT, MAPK, andprotein kinase A/protein kinase C (PKA/PKC), albeitwith tissue- and dose-specific effects on pathway activation(25, 26).Both DNA methylation and histone modifications are

thought to be epigenetic targets for developmental repro-gramming by xenoestrogens. Histone modifications,including methylation, create binding sites for severalregulators of gene expression that recognize these site-specific chromatin marks. Histone methylation, which iscatalyzed by histone methyltransferases (HMT), can beepigenetically inherited, repressing or activating geneexpression. Importantly, nongenomic or more aptly prege-nomic ER signaling can change these epigenetic histonemethyl marks. ER activation can modulate pregenomicsignaling through activation of the PI3K/AKT pathway,leading to AKT phosphorylation of the HMT, enhancer ofzeste homolog 2 (EZH2). Phosphorylation of EZH2 byAKT results in decreased EZH2 activity and levels oftrimethylated lysine 27 on histone 3 (H3K27me3; ref. 27).Reduction of H3K27me3 levels, a repressive mark forgene expression, results in increased expression of estro-gen-responsive genes.Here, we show that the environmental estrogen genistein

activates pregenomic PI3K/AKT signaling to modulateEZH2 phosphorylation and decrease the repressiveH3K27me3 methyl mark in the developing uterus, repro-gramming estrogen-responsive genes to enhance responsive-ness to hormone and increase leiomyoma incidence. Incontrast, bisphenol A, which could not increase uterinetumorigenesis, did not induce pregenomic PI3K/AKT sig-naling in the neonatal uterus, increased rather than decreasedH3K27me3 levels, and repressed, rather than enhanced,estrogen-responsive gene expression in the adult myome-trium. Thus, activation of pregenomic ER signaling tomodulate EZH2 activity and reduce H3K27Me3 levels inthe developing uterus distinguishes the genomic and non-genomic activity of xenoestrogens and their ability to induce

Xenoestrogens Engage EZH2 to Increase Tumorigenesis

www.aacrjournals.org Mol Cancer Res; 10(4) April 2012 547



developmental reprogramming of gene expression andtumorigenesis.

Materials and MethodsAnimals and treatmentsEker rats, from a closed colony at The University of Texas

MD Anderson Cancer Center (Smithville, TX), were caredfor in accordance with the guidelines of the MD AndersonCancer Center Animal Care and Use Committee in anALAC-accredited facility. Females were given water andstandard rat chow (Harlan Teklad 22/5 rodent diet) adlibitum and maintained on a 14:10 light–dark cycle, con-ditions historically associated with a 65% leiomyoma inci-dence in this animal model. Eker rats were treated onpostnatal days 10 to 12 with bisphenol A 50 mg/kg (AcrosOrganics), diethylstilbestrol 1mg/kg (SigmaChemical Co.),or genistein 50 mg/kg (Sigma) in sesame oil, using a total of50 mL of this vehicle for each subcutaneous injection. Aseparate group of animals were given vehicle only as controls.Upon weaning, all animals were genotyped for the presenceof the Eker mutation (Tsc2þ/þ vs. Tsc2Ek/þ). For develop-mental reprogramming analysis, 61 animals were sacrificedat 3 months of age.Acute response to xenoestrogen was evaluated by sacrific-

ing postnatal day (PND) 12 animals 1 to 12 hours after asingle injection of genistein, bisphenol A, or vehicle. Upondetermination that the peak of nongenomic signalingappears 6 hours following exposure in uterus, subsequentexposures to evaluate acute responses were 6 hours induration. Low-dose bisphenol A exposure studies wereconducted in both PND12 Eker (females only) and SpragueDawley rats (males and females). Neonatal Eker rats weregiven a single s.c. injection of bisphenol A (50mg/kg, 50 mg/kg, or 50 ng/kg) or diethylstilbestrol (1 mg/kg) on PND 12and sacrificed at 6 hours after exposure. In addition, SpragueDawley rats were given a single s.c. (10 mg/kg) or oral (0.4–50 mg/kg in sesame oil) dose of bisphenol A and groups of 3to 5 animals per time point per dose sacrificed 0.5 or 6 hoursafter exposure. Prostate studies were conducted at PND 3rather than 12, due to the earlier developmental reprogram-ming window of the rat prostate (PND 1–5) versus the ratuterus (PND 10–12).

Tissue collection, histologic studies, andimmunohistochemistryFemale rats were euthanized at 16 months of age, uteri

were fixed before paraffin embedding and staining withhematoxylin–eosin or incubated with antibody directedagainst Calbindin D9k (1:2,000; Swant) as described pre-viously (18). Tumors were also measured and sectioned forpathologic examination, and a portion of each was snapfrozen in liquid nitrogen. In addition, the uninvolved uteruswas sectioned and analyzed for microscopic tumors, whichtogether with quantitation of macroscopic lesions was usedto calculate tumor incidence and multiplicity. Three-month-old animals were euthanized, and their uteri wereremoved and scraped with a sterile scalpel in cold PBS

solution to remove endometrium frommyometrium, whichwere snap frozen in liquid nitrogen and stored at�80�C. Toobtain adequate amounts of tissue for RNA extraction fromanimals euthanized 6 hours after treatment on PND 12, 2 to3 uteri were pooled together.

Histologic categorization of estrusReproductive staging was carried out in accordance with

the procedure described by Cook and colleagues (16). The16-month-old rats were staged according to degree of repro-ductive senescence (pseudopregnant, persistent estrus, oranestrus). Three-month-old rats were categorized as being inproestrus, estrus, metestrus, or diestrus stages of the estruscycle.

Analytic method for bisphenol A determinationsQuantification of bisphenol A serum levels was carried out

as previously described (28). Briefly, serum was obtainedfrom groups of 3 to 5 pooled animals treated with a singledose of bisphenol A (0.4–50 mg/kg), and 5 hg of deuteratedbisphenol A (d16-bisphenol A) was added as an internalstandard. Three mL of ethyl ether was added to the samplesand the mixture was shaken in an orbital shaker for 30minutes. The extract was concentrated, and the solvent wasreconstituted with 0.5 mL of methanol. This fraction con-tained "free" bisphenol A. Another aliquot of sample wasdigested with glucuronidase and then extracted with ethylether; this fraction contained "total" bisphenol A (free þbound). Bisphenol A levels were quantified with a high-performance liquid chromatography (HPLC) coupled withAPI 2000 electrospray triple-quadrupole mass spectrometer(ESI-MS/MS). Data were acquired using multiple reactionmonitoring (MRM) for the transitions of 227 > 212 forbisphenol A and 241 > 223 for d16-bisphenol A. Quanti-fication was based on external calibration curve prepared byinjecting 10 mL of 0.02, 0.05, 0.1, 0.2, 0.5, 1, 5, 10, and 50ng/mL standards.

Real-time PCRFrozen tumors, myometrium or uteri previously exposed

to vehicle, genistein (50 mg/kg), or bisphenol A (50 mg/kg)on PND 10 to 12 were crushed under liquid nitrogen withmortar and pestle and RNA isolated and DNA removed byusing the RiboPure Kit (Ambion Biosystems) according tothe manufacturer's protocol. Following RNA extraction,cDNA was made by reverse transcribing 1 mg of RNAthrough the Invitrogen Superscript First-Strand SynthesisIII System for reverse transcriptase (RT)-PCR (Invitrogen).Real-time PCR was carried out using the 7900T Fast Real-Time detection system from Applied Biosystems (ABI). FastReal-Time Taq-Man assays from ABI were used exclusivelyto analyze expression of Gdf10 (Rn00666937_m1), Cal-bindin D9k (Rn00560940_m1), Rasd2 (Rn00592054_m1), Sfrp2 (Rn01458836_m1), Krt19 (Rn01496867_m1),Nr2f2 (Rn00756178_m1), Gria2 (Rn00568514_m1),Igfbp5 (Rn00563116_m1), Spp1 (Rn00563571_m1), Car8(Rn01473820_m1), Mmp3 (Rn00591740_m1), Tacst1(Rn00684677_m1), Rps6k (Rn00667685_m1), Kcnk2

Greathouse et al.

Mol Cancer Res; 10(4) April 2012 Molecular Cancer Research548



(Rn00572452_m1), Cspg2 (Rn01493763_m1), Aqp3(Rn00581754_m1), Ramp1 (Rn00671666_m1), and Dio2(Rn00581867_m1) genes. Glyceraldehyde-3-phosphate de-hydrogenase (GAPDH) or ribosomal 18s (18s) were used asendogenous controls. 18s was used for comparison to estro-gen-responsive genes in the adults due to significant associ-ation ofGAPDH expression with bisphenol A exposure. Thefollowing set of conditions were used for each fast real-timePCR reaction: 95�C for 10minutes, followed by 40 cycles of1 second at 95�C and 20 seconds at 60�C.Real-time PCR reactions were carried out in triplicate and

quantified by the�DDCt method. A calibrator from each setof samples was chosen from which to subtract individualgenistein and bisphenol A sample DCt values to obtain�DDCt. The fold change for each sample was calculated incomparison to the calibrator by taking 2�DDCt. Transcriptlevels weremeasured at 6 hours after the last injection for 12-day-old animals exposed to vehicle, genistein, or bisphenolA. The calibrator for both 3-month-old and 12-day-old ratswas vehicle-treated myometrium or uteri. For the 3-month-old rats, an estrus-staged matched vehicle myometrium wasused.

Preparation of tissue lysates and immunoblottingFrozen tissue from PND 12 Eker or Sprague Dawley rats

treated with a single injection of vehicle, genistein, orbisphenol A and harvested at 1, 6, or 12 hours after theinjection, as described earlier, was used for the preparation ofprotein lysates for immunoblotting. Uteri or prostates from2 to 3 neonatal animals were pooled together in RSB buffer[10 mmol/L Tris HCl pH 7–7.4, 10 mmol/L NaCl, 3mmol/L MgCl2, 1 mmol/L phenylmethylsulfonylfluoride(PMSF), 1 mmol/L Na3VO4, 1 mmol/L NaF, and Com-plete Protease Inhibitor Cocktail (Roche Applied Science)]with 0.5% NP-40. Tissue was dounce homogenized andincubated for 10minutes on ice. Lysates were centrifuged for15minutes at 15,000 rpm at 4�C and supernatant collected.Protein concentration was determined by the Pierce BCAassay (Pierce Biotechnology). Tissue lysate was separated via10% SDS-PAGE gels (BioRad Laboratories) and transferredto polyvinylidene difluoride. Membranes were incubatedwith primary antibody and washed with TBS plus 0.5%Tween-20 followed by incubation with horseradish perox-idase (HRP)-conjugated secondary antibody. Visualizationof protein abundance was carried out with Pierce ECLsubstrate or ECL plus (Thermo Fischer Scientific). Immu-noblotting was carried out with antibodies against p-AKT(S473) and (T308), AKT, p-S6 (S235/236), S6, and histoneH3 obtained from Cell Signaling Technology; H3K27me3and EZH2 from Active Motif; and p-EZH2 (S21) fromBethyl Incorporated.

ImmunohistochemistryTissue embedded in paraffin were cut into 5-mm sections

and placed on slides for deparaffinization. Slides were heatedat 60�C for 1 hour and then in further deparaffinized inCitrusolv for 5 minutes 3 times. Slides were rinsed twice in100% EtOH for 2 minutes. Sections were progressively

rehydrated in 90%, 80%, and 70% EtOH for 2 minuteseach and rinsed with water for 1 minute. Slides were thenboiled in antigen unmasking solution for 5 minutes andallowed to cool for 30 minutes, then rinsed twice with PBSfor 5 minutes. Endogenous peroxidase activity was blockedin 2%H2O2 for 30 minutes and rinsed in PBS. Nonspecificbinding was blocked with Avidin D and Biotin for 15minutes and washed with PBS. Sections were incubatedovernight at 4� in primary antibody, Calbindin D9k(1:2,000; Swant), in 5% goat serum plus 0.3% Triton-X,and then rinsed in 0.03% TBS/Tween-20 for 10 minutestwice and washed with PBS for 10 minutes. Sections wereincubated with HRP-conjugated secondary antibody for 1hour and visualized by staining with Tablet DAB (SigmaChemical Company).

ImmunoprecipitationTissue lysates were precleared with protein A sepharose

beads (GE Healthcare) and rabbit IgG (Millipore). Pre-cleared lysates were incubated with antibody against p-EZH2 and protein A sepharose beads then washed withCell Signaling Technology lysis buffer (20 mmol/L Tris-HCl pH 7.5, 1 mmol/L EGTA, 1 mmol/L Na2EDTA, 150mmol/L NaCl, 1 mmol/L b-glycerophosphate, 2.5 mmol/Lsodium pyrophosphate, 1% Triton-X, 1 mmol/L PMSF, 1mmol/L NaF, and Roche Complete Protease InhibitorCocktail). Lysate from immunoprecipitation was separatedand immunoblotted as describe earlier.

Acid precipitation of histones from tissue lysateCell pellets obtained from tissue lysates describe earlier

were resuspended in a 1:1 ratio of 5 mmol/L MgCl2 and 0.8mol/L HCl and sonicated for 20 seconds (30% power)followed by 1-hour incubation on ice. Histone proteinswere collected by centrifugation for 10 minutes at 14,000rpm (4�C), supernatant transferred to a new tube andprecipitated with trichloroacetic acid (50%) and ddH2O.Histone precipitants were collected after centrifiguation for20 minutes at 14,000 rpm (4�C). Histone pellets werewashed with cold acetone and allowed to dry before recon-stituting in a solution of 1.5 mol/L Tris-HCl pH 8.8 andddH2O. Histones were quantitated after separation on Tris-tricine gels (10%–20%) and staining with Coomassie. TotalH3 was used to determine relative histone methylationlevels. For determination of rapid changes in H3K27me3levels in prostate and uterus, whole-cell lysates (rather thanacid precipitation of chromatin-associated histones), wereused and levels of H3K27 methylation quantitated asdescribed earlier.

StatisticsFor analysis of real-time PCR data, a linear model analysis

was applied to DCt values to determine the effects of geneand treatment, which were determined to be significant atP value of less than 0.05. To control for multiple compar-isons the Benjamini and Hochberg test controlling for falsediscovery rate was used in determining significance level foreach real-time PCR analysis. Analysis of tumor incidence





was estimated using c2 for the determination of significancebetween treatment groups. Tumormultiplicity was analyzedusing the Poisson regression for comparing multiple tumorsbetween vehicle, genistein, and bisphenol A groups. A valueof P < 0.05 was considered statistically significant.

ResultsXenoestrogen-specific activation of nongenomicsignaling in the neonatal uterusWe previously identified 18 genes, containing known or

putative estrogen response elements, which were differen-tially expressed in hormone-dependent uterine leiomyomas.Six of these genes became developmentally reprogrammedfollowing neonatal exposure to diethylstilbestrol in the adultuterine myometrium to promote tumorigenesis (19). Todetermine whether the activity of diethylstilbestrol could begeneralized to other, more environmentally relevant, xenoes-trogens, we first compared the ability of 2 environmentalestrogens, genistein (50mg/kg) and bisphenol A (50mg/kg),to induce genomic ER signaling (i.e., modulate transcrip-tion) in the developing uterus.Neonatal rats were exposed to vehicle, genistein, and

bisphenol A (vehicle, n¼ 9; genistein, n¼ 6; and bisphenolA, n ¼ 6), and transcript levels were measured at 6 hoursfollowing the final of 3 exposures on PND 10 to 12, theperiod of uterine development most susceptible to develop-mental reprogramming (19). Real-time PCR quantitation oftransactivation/repression of these 18 hormone-responsivegenes revealed that themajority (12 of 18) were responsive toone or both environmental estrogens, being either inducedor repressed by genistein and/or bisphenol A compared with

vehicle controls (Table 1). Genistein modulated the expres-sion of Calbindin D9k, Dio2, Krt19, Gdf10, Car8, Gria2,Mmp3, Igfbp5, Spp1, Sfrp2, Rasd2, Nr2f2, and bisphenol Amodulated the expression of 8 of these same 12 genes(Calbindin D9k, Dio2, Gdf10, Car8, Gria2, Spp1, Sfrp2,and Rasd2). While qualitatively similar in terms of theirability to induce or repress gene expression in the developinguterus, quantitative differences were observed between the 3xenoestrogens evaluated. For example, Calbindin D9Kexpression increased 28- to 52-fold, with genistein > dieth-ylstilbestrol > bisphenol A, Dio2 was induced 7- to 10-foldwith bisphenol A > diethylstilbestrol > genistein, Gria2 wasrepressed 11- to 83-fold with diethylstilbestrol > genistein >bisphenol A, andGdf10 was repressed 5-fold with bisphenolA¼ diethylstilbestrol¼ genistein. Thus, although genisteinand bisphenol A exhibited xenoestrogen-specific patterns ofgene expression in the developing rat uterus, both genisteinand bisphenol A engaged the ER to induce genomic ERsignaling. Transactivation of gene expression in the neonataluterus by genistein and bisphenol A was confirmed at theprotein level by examining induction of Calbindin D9kprotein expression in response to these xenoestrogens byimmunohistochemistry in neonatal uteri. Uteri from ani-mals exposed to either genistein or bisphenol A on PND 10to 12 exhibited an increase in Calbindin D9k protein asshown in Supplementary Fig. S1, confirming that at thedoses selected genistein and bisphenol A both functioned asER agonists, inducing genomic ER signaling and transacti-vating gene expression in the developing rat uterus.We next asked whether genistein and bisphenol A

were also equivalent in their ability to induce nongenomicER signaling in the developing uterus. Activation of

Table 1. Neonatal uterine gene expression in response to xenoestrogen exposure

Fold change � SEM

Gene

Normalestrogenresponse

Neonataldiethylstilbestrolresponsea

Neonatalgenisteinresponse

Neonatalbisphenol Aresponse

Calbindin D9k Induced 39.12 � 1.38 52.30 � 1.96 28.10 � 1.24Dio2 Induced 7.31 � 0.28 6.63 � 0.21 9.51 � 0.6Krt19 Induced 2.05 � 0.13 NSDGdf10 Repressed �4.78 � 0.06 �4.78 � 0.08 �5.00 � 0.06Car8 Repressed �7.14 � 0.05 �20.1 � 0.05 �12.5 � 0.04Gria2 Repressed �83.3 � 0.03 �33.1 � 0.07 �11.1 � 0.01Mmp3 Repressed �3.85 � 0.08 �2.50 � 0.05 NSDIgfbp5 Repressed �3.53 � 0.09 NSDSppl Repressed �2.97 � 0.13 �3.58 � 0.07Sfrp2 Repressed �9.13 � 0.08 �4.35 � 0.13Rasd2 Repressed �5.28 � 0.17 �2.53 � 0.14Nr2f2 Repressed �4.63 � 0.19 NSD

NOTE: Fold change relative to vehicle. Values statistically significant at P < 0.05 using one-way ANOVA. Values in table expressed asfold change � SEM.Abbreviation: NSD, no significant difference.aData previously reported in the work of Greathouse and colleagues (19).

Greathouse et al.




nongenomic PI3K signaling in response to genistein, bisphe-nol A, or vehicle was determined at 1, 6, and 12 hours afterxenoestrogen exposure. Uteri of rats exposed to genisteinexhibited robust activation of PI3K/AKT signaling relativeto vehicle, as evidenced by phosphorylation of AKT at S473and T308 and its downstream effector, S6 at S235/236 (Fig.1). Activation of PI3K/AKT signaling by genistein in thedeveloping uterus was transient, returning to baseline levelswithin 12 hours of exposure, consistent with the rapidkinetics of nongenomic signaling and previous observationsof in vivo nongenomic signaling by diethylstilbestrol (27). Incontrast to genistein, bisphenol A failed to activate PI3K/AKT signaling (Fig. 1). Thus, at doses where both genisteinand bisphenol A induced genomic ER signaling, genisteinbut not bisphenol A activated nongenomic PI3K/AKTsignaling in the developing rat uterus.In some settings, bisphenol A has been shown to exhibit a

U-shaped dose–response curve, exhibiting stimulatoryeffects at low doses not evident at high(er) doses of thiscompound (29). To determine whether doses of bisphenol Alower than those required for genomic ER activity couldinduce nongenomic signaling, we repeated these experi-ments at exposures 3 to 6 orders of magnitude lower. Weobserved that lower bisphenol A doses (50 mg/kg and 50 ng/kg) were similarly ineffective at activating nongenomicPI3K/AKT signaling in the developing uterus relative todiethylstilbestrol (1 mg/kg, positive control) or vehicle

controls (Supplementary Fig. S2). To further investigatewhether the inability of bisphenol A to induce PI3K/AKTsignaling in the developing uterus was caused by the use ofsubcutaneous rather than oral route of exposure, we nextcompared oral administration of bisphenol A over a doserange of 0.4 to 50 mg/kg to an s.c. injection of bisphenol A(10 mg/kg) previously established as able to induce devel-opmental reprogramming and increase susceptibility toprecancerous lesions and tumors in the prostate (12, 30).We found that oral bisphenol A administered at 50 mg/kggave similar levels of circulating total bisphenol A as 10mg/kgs.c. [5.19 and 4.82 ng/mL (ppb) total serum bisphenol A inmales and 15.1 and 12.0 ng/mL (ppb) total serum bisphenolA in females, respectively]. Importantly, over a dose range of0.4, 2, and 50 mg/kg, no activation of PI3K/AKT signalingwas observed in the neonatal uterus, whereas both 2 and 50mg/kg doses of bisphenol A activated PI3K/AKT signaling inthe neonatal prostate (Supplementary Fig. S3A). This con-trasted with what was observed for diethylstilbestrol, whichactivated PI3K/AKT in both the uterus and prostate (Sup-plementary Fig. S3B).

Pregenomic signaling to EZH2 in the developing uterusexhibits xenoestrogen specificityThe HMT EZH2 is a target for AKT, with phosphory-

lation by this kinase at S21 reducing EZH2 activity andlevels of the H3K27me3 repressive methyl mark in

Figure 1. Xenoestrogen-specificmodulation of nongenomic PI3K/AKT signaling. A, Western blotanalysis of AKT phosphorylation(T308 and S473) and S6phosphorylation in pooled (3) uteri ofPND 10 to 12 Eker rats after a singleexposure for 1 to 12 hours to vehicle(VEH), genistein (GEN; 50 mg/kg), orbisphenol A (BPA; 50 mg/kg) via s.c.injection. B, densitometricquantification of Western blotanalysis shows significantdifferences in activation of AKT asrepresented by the ratio p-AKT T308and S473 to total AKT inxenoestrogen-exposed Eker ratsversus vehicle. Mean � SEM isdenoted by error bars on each graph(n¼3). TheStudent t testwasused todetermine statistical significance,which was set at a value of�, P � 0.05.

A

B

n = 3

P ≤ 0.05

n = 3

P ≤ 0.05**

15

10

5

0

25

20

15

10

5

0

p-AKT473

VEHVEH

VEH

1 h G

EN

6 h G

EN

1 h B

PA

6 h B

PA

12 h

GEN

12 h

BPA VEH

VEHVEH

1 h G

EN

6 h G

EN

1 h B

PA

6 h B

PA

12 h

GEN

12 h

BPA

Fo

ld c

ha

ng

e c

om

pa

red

wit

h V

EH

Fo

ld c

ha

ng

e c

om

pa

red

wit

h V

EH

p-AKT308

1 hTime (h)

VEH

p-AKT (T308)

p-AKT (S473)

p-S6 (S235/236)

S6

AKT

GEN BPA VEH GEN BPA VEH GEN BPA

6 h 12 h





chromatin (27, 31). Differences in the ability of genisteinand bisphenol A to activate PI3K/AKT signaling suggestedthat these environmental estrogens might also differentiallymodulate the activity of this epigenetic regulator. Exposureof developing uteri to genistein increased phosphorylationof EZH2 at S21 with kinetics commensurate with activationof PI3K/AKT signaling (Fig. 2A). In contrast, doses ofbisphenol A ranging from 50 mg/kg to 50 ng/kg failed toactivate PI3K/AKT signaling and resulted in no increase inEZH2 phosphorylation relative to vehicle controls (Fig. 2Aand Supplementary Fig. S2). Importantly, as shown in Fig.2B, H3K27me3 levels in chromatin were significantlyreduced in genistein, but not in bisphenol A exposed uteri,concordant with the ability of genistein to induce inhibitoryphosphorylation of EZH2 at S21. In contrast, and consistentwith induction of nongenomic PI3K/AKT signaling in theprostate, the same dose of bisphenol A rapidly (0.5 hour)reduced H3K27me3 levels (Supplementary Fig. S3C).These data indicate that rapid activation of nongenomic,or more appropriately pregenomic, signaling of AKT toEZH2 in the developing uterus reduces H3K27me3 levelsin chromatin and exhibits both tissue and xenoestrogenspecificity.

Developmental reprogramming of gene expression in theadult myometrium by environmental estrogensDevelopmental reprogramming of the female rodent

reproductive tract can result in morphologic and histologicalterations including persistent vaginal cornification, endo-metrial hyperplasia, and polycystic ovaries lacking corporalutea (18, 32, 33). Examination of the uterus, ovary, andvagina of adult female rats exposed neonatally to bisphenol A(n¼ 17), genistein (n¼ 14), or vehicle (n¼ 34) revealed thatthe effect of these xenoestrogens on reproductive tractmorphology differed substantially from that previouslyobserved with diethylstilbestrol (Supplementary Fig. S4 andref. 18). Relative to diethylstilbestrol, adult females exposedneonatally to bisphenol A or genistein had a more normalreproductive tract morphology as illustrated by an absence ofpersistent vaginal cornification, the presence of corpora luteain the ovary, luminal epithelium cell height in the endome-trium, and apparently normal estrus cyclicity (Supplemen-tary Fig. S4C-L). Thus, compared with the potent pharma-cologic xenoestrogen diethylstilbestrol, neonatal exposure tothe environmental estrogens genistein and bisphenol A didnot induce dramatic morphologic alterations of the repro-ductive tract in female rats. In addition, we examined theeffect of neonatal exposure to bisphenol A, genistein, orvehicle on animal weight at 11 and 16 months of age andfound no significant difference for either genistein or bisphe-nol A as compared with vehicle animal weights (data notshown).To determine whether environmental estrogens had

induced developmental reprogramming at the molecularlevel, we examined expression of estrogen-responsive genesin adult uteri of animals exposed neonatally to genistein andbisphenol A. For this analysis, we defined a developmentallyreprogrammed gene as one that displays altered hormone

responsiveness in adult rat myometrium as a result ofdevelopmental xenoestrogen exposure. To control for dif-ferences in estrogen levels associated with different stages ofthe estrus cycle, adult vehicle-, genistein- or bisphenol A–exposed female rats were grouped into the high-estrogen(proliferative) phase, corresponding to animals in proestrusor estrus, or low estrogen (secretory) phase, corresponding toanimals in metestrus and diestrus. Real-time PCR was used

A

B

IP: p-EZH2

6 h

EZH2

EZH2Input

IgG VEH GEN BPA

VEH

H3

H3K27me3

GEN BPA

VEH

2.5

2.0

1.5

1.0

0.5

0.0GEN

sample

*

BPA

Fo

ld c

han

ge r

ela

tive t

o V

EH

n = 3

P ≤ 0.05

Figure 2. Xenoestrogen-specific modulation of EZH2 and H3K27me3.A, neonatal PND 10 to 12 pooled (2, 3) uteri from Eker rats werehomogenized 6hours after a single exposure to vehicle, genistein (50mg/kg), or bisphenol A (50 mg/kg) and immunoprecipitated (IP) with anti-p-EZH2 antibody. Western blot analysis of EZH2 from immunoprecipitantsshow a significant enrichment of p-EZH2 in uteri from genistein-exposedrats as comparedwith bisphenol A and vehicle. B, neonatal PND 10 to 12pooled (2, 3) uteri from Eker rats after a single exposure to vehicle,genistein (50 mg/kg), or bisphenol A (50 mg/kg) for 6 hours were used foracid precipitation of histones. Quantitation fromWestern blot analysis ofhistone proteins after normalizing for total histone H3 levels shows adecrease in levels of H3K27me3 at 6 hours after exposure in genistein-exposed as compared with bisphenol A- or vehicle-exposed animals.Densitometeric analysis of Western blotting data illustrate significantdifferences in phosphorylation and methylation after xenoestrogenexposure in Eker rat uteri. The levels of p-EZH2 and H3K27me3 werenormalized to total levels and were determined to be significantlydifferent from vehicle using the Student t test, which was set at avalue of �, P � 0.05. Error bars represent the mean � SEM in eachgraph (n ¼ 3).

Greathouse et al.




to quantitate gene expression, with average fold change inexpression compared with vehicle controls and normalizedto expression of ribosomal 18s.As shown in Table 2, of the 18 estrogen-responsive

genes examined, 3 were developmentally reprogrammedby both genistein and bisphenol A: Calbindin D9k,Gdf10, and Gria2. Similar to what was seen previouslywith diethylstilbestrol, genistein reprogrammed thesegenes to become hyperresponsive to estrogen (i.e., expres-sion was significantly increased during the proliferativephase of the estrus cycle relative to vehicle controls).In the 3-month-old myometrium, Calbindin D9k, Gdf10,and Gria2 expression became elevated by 4-, 3-, and4-fold, respectively, relative to stage-matched vehiclecontrols. In contrast, bisphenol A had the opposite effecton these genes, reprogramming them to become repressedby estrogen, with expression significantly decreased dur-ing the proliferative phase of the estrus cycle. In 3-month-old rats exposed neonatally to bisphenol A, CalbindinD9k, Gdf10, and Gria2 expression each decreased by3-fold relative to stage-matched vehicle controls.In addition to the 3 genes developmentally repro-

grammed by all xenoestrogens, genistein and bisphenolA each reprogrammed a distinct set of genes, albeit with atrend similar to that observed with Calbindin D9k, Gdf10,and Gria2. As shown in Table 2, neonatal genisteinexposure reprogrammed expression of Dio2, Krt19,Igfbp5, and Spp1 (only one of which, Dio2, was alsoreprogrammed by diethylstilbestrol; ref. 19), causing themajority (but not all) of these genes to become hyperre-sponsive to hormone. Bisphenol A reprogrammed Sfrp2and Rasd2, neither of which was reprogrammed by dieth-ylstilbestrol or genistein. Though this is not an exhaustivelist of all possible genes differentially expressed or devel-opmentally reprogrammed by exposure to genistein andbisphenol A, Table 2 serves as representation of thepatterns of developmental reprogramming in estrogen-responsive genes following neonatal exposure.

Developmental reprogramming by genistein but notbisphenol A promotes uterine tumorigenesisEker rats carrying a Tsc2 tumor suppressor gene defect

(Tsc2Ek/þ) develop uterine leiomyomas by 16 months ofage with a historical tumor incidence of 65% (34). Thedifferential ability of genistein and bisphenol A to inducepregenomic signaling to EZH2 to modulate H3K27me3levels and increase estrogen-responsive gene expression ledus to ask whether this differential engagement of EZH2by these environmental estrogens translated to differencesin their impact on uterine tumorigenesis in this animalmodel. Female Tsc2Ek/þ rats were exposed on PND 10to 12 to genistein, bisphenol A, or vehicle and tumorincidence and multiplicity determined in adult femalesat 16 months. Neonatal genistein exposure significantlyincreased tumor incidence at 16 months to 93% [genistein(n ¼ 14) vs. vehicle controls (n ¼ 17), P < 0.05; Fig. 3A].Tumor multiplicity was also increased by neonatal genis-tein exposure versus vehicle controls (1.6/rat vs. 0.6/rat,respectively; Fig. 3B). In contrast to genistein, neonatalbisphenol A did not significantly increase tumor incidenceor multiplicity relative to vehicle controls, differing sig-nificantly from the effects of both diethylstilbestrol (16)and genistein on tumorigenesis in this animal model[genistein vs. bisphenol A (n ¼ 30), P < 0.01]. Thus,induction of pregenomic signaling, inhibition of EZH2activity, decreased H3K27me3 methylation, increasedestrogen-responsive gene expression, and impact ontumorigenesis were concordant, with genistein but notbisphenol A engaging this pathway to developmentallyreprogram the developing uterus and increase uterinetumorigenesis.

DiscussionER-mediated pregenomic signaling can activate the PI3K/

AKT signaling pathway to phosphorylate the HMT EZH2,modulating its activity and gene expression regulated by

Table 2. Adult myometrial gene expression in animals exposed neonatally to genistein or bisphenol A

Fold change � SEM

GeneNormal estrogenresponse

Adult response(neonatal genistein)

Adult response(neonatal bisphenol A)

Calbindin D9k Induced 3.8 � 0.15b �3.30 � 0.21a

GdflO Repressed 2.9 � 0.19b �2.8 � 0.21a

Gria2 Repressed 4.4 � 0.16b �2.8 � 0.33a

Dio2 Induced �2.1 � 0.09b NSDKrtl9 Induced 2.3 � 0.13b NSDIgfbp5 Repressed �13.9 � 0.12a NSDSppl Repressed �2.2 � 0.13b NSDSfrp2 Repressed NSD �2.0 � 0.19a

Rasd2 Repressed NSD 6.3 � 0.28a

NOTE: Fold change relative to vehicle. Values statistically significant at aP < 0.01 and bP < 0.05 using one-way ANOVA.Abbreviation: NSD, No significant difference; SEM, standard error of the mean.





H3K27me3 methylation (27). We found that environmen-tal estrogens engage this pathway during developmentalreprogramming of gene expression and tumorigenesis in theuterus and that xenoestrogens have distinct pregenomicsignaling profiles and effects on this epigenetic regulatorthat correlate with their ability to induce developmentalreprogramming. Rapid activation of AKT, which repressedthe activity of EZH2 and levels of H3K27me3 in chromatin,distinguished genistein from bisphenol A, despite the factthat both xenoestrogens were able to induce genomic ERsignaling and transactivate the expression of estrogen-responsive genes in the developing uterus. Importantly,these data show that the pathway by which diethylstilbestrolinterferes with epigenetic programming in the developinguterus, that is, via activation of PI3K/AKT, phosphorylationofHMTEZH2, and reduction inH3K27methylation (27),is engaged by genistein but not bisphenol A and thatdisruption of this pathway during development correlateswith the effect of these endocrine disruptors on tumorigen-esis in myometrium.Importantly while genistein and bisphenol A both eli-

cited a genomic ER response (modulation of expression ofestrogen-responsive genes in the neonatal uterus; Table 1),

this genomic response did not correlate with the ability toinduce developmental reprogramming in the adult myo-metrium. Several genes that exhibited a genomic responseto genistein and/or bisphenol A (i.e., were induced orrepressed) in the neonate failed to become reprogrammedin the adult myometrium of exposed animals and/orexhibited differences in estrogen responsiveness, for exam-ple, exhibiting elevated expression in genistein animals anddecreased expression in bisphenol A animals. Similarly,Sfrp2 exhibited a genomic response to genistein in theneonatal uterus but was not reprogrammed in adulthood,and whereas bisphenol A induced Dio2 expression in theneonatal uterus, it did not reprogram this gene in the adultmyometrium. Thus, induction of a genomic response byxenoestrogens in the developing uterus was insufficient toinduce developmental reprogramming in adulthood anddid not predict xenoestrogen specificity for reprogrammingof gene expression.Differences in the ability of xenoestrogens to induce

developmental reprogramming are likely driven by severalintrinsic differences between these xenoestrogens, for exam-ple binding to specific ER subtypes. Diethylstilbestrol,genistein, and bisphenol A have distinct binding efficienciesfor ERa and ERb, with diethylstilbestrol and genisteinbinding with a higher affinity to ERa than bisphenol A(diethylstilbestrol > genistein > bisphenol A), whereasbisphenol A binds with a much higher affinity to ERb thanERa (35). In addition, some actions of bisphenol A may bemediated via a nonclassical G-protein–coupled membranereceptor that activates nongenomic signaling through PKA,phosphorylating targets such as cAMP–responsive ele-ment–binding protein (CREB) in vitro (36). Importantly,in the female reproductive tract, ERa is required for dieth-ylstilbestrol-induced developmental reprogramming, asaERKO mice are resistant to diethylstilbestrol-induceddevelopmental reprogramming (18) and overexpression ofERa exacerbates the effects of diethylstilbestrol on devel-opmental reprogramming. We also have previously showedthat aERKO mice are resistant to the diethylstilbestrol-induced activation of pregenomic PI3K/AKT signaling toEZH2 (27). Xenoestrogen-specific affinity for ERa versusERb has also been shown in prostate tissue from maleaERKO and bERKO mice exposed neonatally to diethyl-stilbestrol (37). In that study, diethylstilbestrol exposureinduced prostate abnormalities in the bERKO mice, whichexpress ERa. Conversely, aERKO mice that lack the ERareceptor were resistant to the effect of diethylstilbestrolexposure. Our findings that diethylstilbestrol and genisteinbut not bisphenol A, developmentally reprogrammed uter-ine tumorigenesis is consistent with ERa being the pre-dominate ER subtype in development of the female repro-ductive tract (38).Histone methylation and cytosine methylation are heri-

table chromatin modifications that regulate gene expression.These epigenetic marks have been shown to influence eachother, although the exact mechanism(s) by which histonemethyl marks direct patterns of DNA methylation and visaversa are just beginning to be understood (39–41).

A

B

VEH

1.0

0.8

0.6

0.4

0.2

0.0

2.0

1.5

1.0

0.5

0.0

16-mo tumor incidence

16-mo tumor multiplicity

Inc

ide

nc

e

GEN BPA

VEH GEN BPA

Mu

ltip

lic

ity

P < 0.01 P < 0.001

P < 0.01P < 0.05

Figure 3. Tumor incidence andmultiplicity of Eker rats exposed neonatallyto genistein, bisphenol A, or vehicle. Incidence and multiplicity of uterineleiomyomas in Eker rats after xenoestrogen exposure. A, tumor incidencein 16-month-old Eker rats (gross andmicroscopic tumors) after genistein(50 mg/kg; n ¼ 14), bisphenol A (50 mg/kg; n ¼ 30), or vehicle (n ¼ 17)exposure for 3 consecutive days, PND 10 to 12. B, tumor multiplicity in16-month-old Eker rats (gross and microscopic tumors). Statisticalsignificance for incidences was determined by c2 for incidence and byPoisson regression for multiplicity analysis.

Greathouse et al.




H3K27me3 methylation of chromatin by EZH2 is requiredfor several key functions during development, including Xinactivation, bivalent chromatin maintenance, and silencingof HOX genes, which have been shown to be critical toproper murine uterine differentiation (42, 43). EZH2 canalso directly silence estrogen-responsive genes, such as pro-gesterone receptor, by inducing H3K27me3 methyl marksthat promote DNA methylation via recruitment of thehistone deacetylase HDAC1, followed by recruitment ofDNA methyltransferases (44). It is not known at presentwhether reduction in H3K27me3 methyl marks by xenoes-trogens in the developing uterus alters gene expression in theadult uterus directly as a result of decreases in histonemethylation or by directing changes in DNA methylationof estrogen-responsive genes.While we have focused on developmental reprogramming

of tumorigenesis by environmental estrogens in the myome-trium, xenoestrogens have been shown to lead to persistentchanges in gene expression that are associatedwith neoplastictransformation in other tissues. In the endometrium,xenoestrogen exposure during critical windows of uterinedevelopment (i.e., PND1–5), permanently alters the expres-sion of the estrogen-responsive genes lactoferrin and c-fos inthe adult endometrium. Altered expression is associated withsite-specific DNA hypomethylation, rather than promoterhypermethylation, which persists in uterine adenocarcino-mas that develop in these animals (45). While the mecha-nism by which xenoestrogens may induce changes in DNAdemethylation is not well understood, it was shown thatparathyroid hormone can engage pregenomic signaling viaPKC, leading to phosphorylation of methyl-binding protein4 (MBD4; ref. 46). MBD4 recruits DNA base excisionrepair machinery, resulting in demethylation of the vitaminD receptor-dependent promoter of the cytochrome p45027B1 (CYP27B1) gene. Thus, chromatin remodeling viapregenomic signaling pathways can result in changes inboth histone methyl marks and DNA methylation andmay function as a general pathway by which activation ofnuclear hormone receptors engage the cell's epigeneticmachinery.Overall, the patterns of developmental reprogramming

that emerged from this analysis revealed that reprogrammingby environmental estrogens exhibits xenoestrogen specific-ity. Genistein developmentally reprogrammed, Gdf10, Cal-bindin D9k,Gria2,Dio2,Krt19, Igfbp5, and Spp1; the first 5of which were also reprogrammed by diethylstilbestrol (19).In contrast, bisphenol A reprogrammed Gdf10, CalbindinD9k, Gria2, Rasd2, and Sfrp2. Unlike diethylstilbestrol andgenistein, bisphenol A reprogramming resulted in decreasedgene expression (with the exception of Rasd). Therefore, ingeneral, the effect of bisphenol A on gene expression was theopposite of diethylstilbestrol and genistein, repressing ratherthan enhancing estrogen responsiveness. These findings arereminiscent of those reported by Dolinoy and colleagues(47), where developmental exposure to genistein and bisphe-nol A had opposite effects on DNA methylation and sub-sequent gene expression of the agouti gene. Furthermore, wehave now shown that the pattern of xenoestrogen-induced

developmental reprogramming in genes targeted by genis-tein, bisphenol A, and diethylstilbestrol correlated with theability of these xenoestrogens to increase tumorigenesis.Both neonatal diethylstilbestrol and genistein causedincreased expression of estrogen-responsive genes in adultanimals relative to vehicle-exposed rats, promoting tumorformation in diethylstilbestrol- and genistein-exposed ani-mals (19). In contrast, neonatal bisphenol A caused a generalrepression of estrogen-responsive genes and did not increasetumor incidence or multiplicity compared with vehicle.Together with our previous report (19) these data indicatethat xenoestrogens such as diethylstilbestrol, bisphenol A, orgenistein have both shared and distinct effects on develop-mental reprogramming of target genes that reflect their effecton uterine tumorigenesis. It is important to note thatbisphenol A tumorigenicity studies were conducted at adose of 50 mg/kg, the minimum required to elicit anequivalent genomic response to genistein and diethylstilbes-trol. Although ourmechanistic studies clearly show that evenat doses 6 orders of magnitude lower (i.e., 50 ng/kg),bisphenol A did not induce pregenomic PI3K/AKT signal-ing to EZH2 in the developing uterus, it is still a formalpossibility that effects on tumorigenicity and reprogram-ming of gene expression not evident at 50mg/kg could occurat lower dose bisphenol A exposures via other mechanisms.Distinctions between the effects of these environmental

estrogens on the developing reproductive tract also wereobserved. Morphologic reprogramming of the reproductivetract following neonatal genistein and bisphenol A exposurewas not observed, unlike what was seen in this model withdiethylstilbestrol (18). However, other studies have reportedreproductive abnormalities as a result of perinatal exposureto genistein and bisphenol A (2, 9, 48, 49). While othermechanisms have been shown for the developmental repro-gramming effects of genistein and bisphenol A, includingdisruption of the hypothalamic-pituitary-gonadal (HPG)axis (50, 51), our previous work (18) from ovariectomized(or sham) Eker rats exposed neonatally to diethylstilbestrol(or vehicle) we observed abolishment of Calbindin D9k andsignificantly reduced progesterone receptor expression com-pared with sham controls, indicating that the genes hadbecome hyperresponsive to endogenous hormones asopposed to constitutively active. Thus, in our model ofuterine tumorigenesis, we believe that ovarian hormones arerequired for the developmental reprogramming effects ofgenistein and bisphenol A to be observed in the adultanimals, though we did not specifically test the effects ofgenistein and bisphenol A on the HPG axis. Several possi-bilities exist for these discordant observations includingdifferences in dose, route of administration, timing ofadministration, andmodel species. Regardless, data obtainedin this rat model system clearly show that reprogramming oftumor susceptibility and estrogen-responsive gene expres-sion at the molecular level can occur at doses of xenoestro-gens that do not cause overt morphologic changes in thefemale reproductive tract. Similar to our observations, Ada-chi and colleagues showed that molecular alterations in thetestis can occur in the absence of morphologic alterations





(52). For example, whereas neonatal diethylstilbestrol expo-sure induced both morphologic and molecular alterations inthe testis, genistein reprogrammed genes in the adult testeswithout inducing morphologic reprogramming. Therefore,the effects of developmental reprogrammingmaymanifest inthe absence of histologic or morphologic alterations, point-ing to the need to develop molecular biomarkers to detectreprogramming even in morphologically normal appearingtissues.In summary, this study shows that neonatal exposure to

genistein, in contrast to bisphenol A, activates pregenomicsignaling to AKT, phosphorylating EZH2, and reducingH3K27me3 levels in the developing uterus. The disparateability of genistein and bisphenol A to activate pregenomicsignaling to EZH2 was reflected in the adult myometriumby different patterns of developmental reprogramming ofestrogen-responsive genes and differences in the abilityof these xenoestrogens to increase tumorigenesis in thistissue. In addition to highlighting what are likely impor-tant intrinsic differences between xenoestrogens, thisstudy also identifies a pathway by which xenoestrogenscan engage the cell's epigenetic machinery during theprocess of developmental reprogramming, an importantfirst step towards understanding how early life xenoestro-gen exposures can increase susceptibility to hormone-dependent tumors in adulthood.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: K.L. Greathouse, S. Ho, C.L. Walker.Development of methodology: K.L. Greathouse, T. Bredfeldt, C.L. Walker.Acquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.): K.L. Greathouse, T. Bredfeldt, T. Berry, K. Kannan, S. Ho, C.L.Walker.Analysis and interpretation of data (e.g., statistical analysis, biostatistics, compu-tational analysis): K.L. Greathouse, T. Bredfeldt, J.I. Everitt, K. Lin, K. Kannan, M.L. Mittelstadt, S. Ho.Writing, review, and/or revision of the manuscript: K.L. Greathouse, T. Bredfeldt,K. Lin, K. Kannan, M. L. Mittelstadt, S. Ho.Administrative, technical, or material support (i.e., reporting or organizing data,constructing databases): K.L. Greathouse, M. L. Mittelstadt.Study supervision: K.L. Greathouse, S. Ho, C.L. Walker.

Grant SupportThe support for this work was funded by grants from the National Institute of

Environmental Health Sciences (P30ES006096, RC2ES018789, P30ES007784,R01ES008263, and 1F32ES016509), National Institute of Child Health and Devel-opment (HD046282), and the National Cancer Institute (P30CA016672). K. L.Greathouse was supported by the Schissler Foundation Fellowship inHumanGeneticsand Cancer from the Graduate School of Biomedical Sciences, University of Texas(Smithville, TX).

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be herebymarked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Received January 3, 2012; revised February 22, 2012; accepted February 22, 2012;published OnlineFirst March 29, 2012.

References1. Kyung-Chul C, Eui-Bae J, Leung PCK. Impact of environmental endo-

crine disruption on the reproductive system for humanhealth. ImmunolEndocrine Metab Agents Med Chem 2006;6:3–13.

2. Newbold RR, Banks EP, Bullock B, Jefferson WN. Uterine adenocar-cinoma inmice treated neonatally with genistein. Cancer Res 2001;61:4325–8.

3. Palmer JR, Wise LA, Hatch EE, Troisi R, Titus-Ernstoff L, StrohsnitterW, et al. Prenatal diethylstilbestrol exposure and risk of breast cancer.Cancer Epidemiol Biomarkers Prev 2006;15:1509–14.

4. Camp EA, Coker AL, Robboy SJ, Noller KL, Goodman KJ, Titus-Ernstoff LT, et al. Breast cancer screening in women exposed in uteroto diethylstilbestrol. J Womens Health (Larchmt) 2009;18:547–52.

5. D'Aloisio AA, Baird DD, DeRoo LA, Sandler DP. Association of intra-uterine and early-life exposures with diagnosis of uterine leiomyomataby 35 years of age in the Sister Study. Environ Health Perspect2010;118:375–81.

6. Wise LA, Palmer JR, Rowlings K, Kaufman RH, Herbst AL, Noller KL,et al. Risk of benign gynecologic tumors in relation to prenatal dieth-ylstilbestrol exposure. Obstet Gynecol 2005;105:167–73.

7. Newbold RR, Jefferson WN, Padilla-Banks E. Long-term adverseeffects of neonatal exposure to bisphenol A on the murine femalereproductive tract. Reprod Toxicol 2007;24:253–8.

8. Howdeshell KL, Hotchkiss AK, Thayer KA, Vandenbergh JG, vomSaalFS. Exposure to bisphenol A advances puberty. Nature 1999;401:763–4.

9. Schonfelder G, Friedrich K, Paul M, Chahoud I. Developmental effectsof prenatal exposure to bisphenol A on the uterus of rat offspring.Neoplasia 2004;6:584.

10. Kato H, Ota T, Furuhashi T, Ohta Y, Iguchi T. Changes in reproductiveorgans of female rats treated with bisphenol A during the neonatalperiod. Reprod Toxicol 2003;17:283.

11. Nikaido Y, Danbara N, Tsujita-Kyutoku M, Yuri T, Uehara N, TsuburaA. Effects of prepubertal exposure to xenoestrogen on developmentof estrogen target organs in female CD-1 mice. In Vivo 2005;19:487–94.

12. Prins GS, Tang WY, Belmonte J, Ho SM. Developmental exposureto bisphenol A increases prostate cancer susceptibility in adultrats: epigenetic mode of action is implicated. Fertil Steril 2008;89:e41.

13. Setchell KD, Zimmer-Nechemias L,Cai J, Heubi JE. Isoflavone contentof infant formulas and the metabolic fate of these phytoestrogens inearly life. Am J Clin Nutr 1998;68:1453S–61S.

14. Walker CL, Stewart EA. Uterine fibroids: the elephant in the room.Science 2005;308:1589–92.

15. Jefferson WN, Padilla-Banks E, Newbold RR. Disruption of the femalereproductive system by the phytoestrogen genistein. Reprod Toxicol2007;23:308.

16. Cook JD,Davis BJ, Cai SL, Barrett JC,Conti CJ,WalkerCL. Interactionbetween genetic susceptibility and early-life environmental exposuredetermines tumor-suppressor-gene penetrance. Proc Natl Acad Sci US A 2005;102:8644–9.

17. Walker CL, Hunter D, Everitt JI. Uterine leiomyoma in the Eker rat: auniquemodel for important diseases of women. GenesChromosomesCancer 2003;38:349–56.

18. Cook JD,Davis BJ,Goewey JA, Berry TD,WalkerCL. Identification of asensitive period for developmental programming that increases risk foruterine leiomyoma in Eker rats. Reprod Sci 2007;14:121–36.

19. GreathouseKL,Cook JD, Lin K, Davis BJ, Berry TD, Bredfeldt TG, et al.Identification of uterine leiomyoma genes developmentally repro-grammed by neonatal exposure to diethylstilbestrol. Reprod Sci2008;15:765–78.

20. Couse JF, Korach KS. Estrogen receptor-alpha mediates the detri-mental effects of neonatal diethylstilbestrol (DES) exposure in themurine reproductive tract. Toxicology 2004;205:55–63.

21. Glidewell-Kenney C, Hurley LA, Pfaff L, Weiss J, Levine JE, JamesonJL. Nonclassical estrogen receptor alpha signaling mediates negativefeedback in the female mouse reproductive axis. Proc Natl Acad Sci US A 2007;104:8173–7.

22. Chen JR, Plotkin LI, Aguirre JI, Han L, Jilka RL, Kousteni S, et al.Transient versus sustained phosphorylation and nuclear accumulation

Greathouse et al.




of ERKs underlie anti-versus pro-apoptotic effects of estrogens. J BiolChem 2005;280:4632–8.

23. Kousteni S, Chen JR, Bellido T, Han L, Ali AA, O'Brien CA, et al.Reversal of bone loss in mice by nongenotropic signaling of sexsteroids. Science 2002;298:843–6.

24. ChenZ,Yuhanna IS,Galcheva-GargovaZ,KarasRH,MendelsohnME,Shaul PW. Estrogen receptor alpha mediates the nongenomic activa-tion of endothelial nitric oxide synthase by estrogen. J Clin Invest1999;103:401–6.

25. Watson CS, Bulayeva NN, Wozniak AL, Alyea RA. Xenoestrogens arepotent activators of nongenomic estrogenic responses. Steroids2007;72:124–34.

26. Li X, Zhang S, Safe S. Activation of kinase pathways in MCF-7 cells by17[beta]-estradiol and structurally diverse estrogenic compounds. JSteroid Biochem Mol Biol 2006;98:122–32.

27. Bredfeldt TG, Greathouse KL, Safe SH, HungMC, BedfordMT,WalkerCL. Xenoestrogen-induced regulation of EZH2 and histone methyla-tion via estrogen receptor signaling to PI3K/AKT. Mol Endocrinol2010;24:993–1006.

28. Prins GS, Ye SH, Birch L, Ho SM, Kannan K. Serum bisphenol Apharmacokinetics and prostate neoplastic responses following oraland subcutaneous exposures in neonatal Sprague-Dawley rats.Reprod Toxicol 2011;31:1–9.

29. Hugo ER, Brandebourg TD, Woo JG, Loftus J, Alexander JW, Ben-Jonathan N. Bisphenol A at environmentally relevant doses inhibitsadiponectin release from human adipose tissue explants and adipo-cytes. Environ Health Perspect 2008;116:1642–7.

30. Ho SM, Tang WY, Belmonte de Frausto J, Prins GS. Developmentalexposure to estradiol and bisphenol A increases susceptibility toprostate carcinogenesis and epigenetically regulates phosphodies-terase type 4 variant 4. Cancer Res 2006;66:5624–32.

31. Cha TL, ZhouBP, XiaW,WuY, YangCC, ChenCT, et al. Akt-mediatedphosphorylation of EZH2 suppresses methylation of lysine 27 inhistone H3. Science 2005;310:306–10.

32. Burroughs CD, Mills KT, Bern HA. Long-term genital tract changes infemale mice treated neonatally with coumestrol. Reprod Toxicol1990;4:127–35.

33. Katsuda S, Yoshida M, Watanabe G, Taya K, Maekawa A. Irreversibleeffects of neonatal exposure to p-tert-octylphenol on the reproductivetract in female rats. Toxicol Appl Pharmacol 2000;165:217–26.

34. Walker CL, Cesen-Cummings K, Houle C, Baird D, Barrett JC, Davis B.Protective effect of pregnancy for development of uterine leiomyoma.Carcinogenesis 2001;22:2049–52.

35. Routledge EJ, White R, Parker MG, Sumpter JP. Differential effects ofxenoestrogens on coactivator recruitment by estrogen receptor (ER)alpha and ERbeta. J Biol Chem 2000;275:35986–93.

36. Bouskine A, Nebout M, Brucker-Davis F, Benahmed M, Fenichel P.Low doses of bisphenol A promote human seminoma cell prolif-eration by activating PKA and PKG via a membrane G-protein-coupled estrogen receptor. Environ Health Perspect 2009;117:1053–8.

37. Prins GS, Birch L, Couse JF, Choi I, Katzenellenbogen B, Korach KS.Estrogen imprinting of the developing prostate gland is mediatedthrough stromal estrogen receptor alpha: studies with alphaERKOand betaERKO mice. Cancer Res 2001;61:6089–97.

38. Mowa CN, Iwanaga T. Developmental changes of the oestrogenreceptor-alpha and -beta mRNAs in the female reproductive organ ofthe rat–an analysis by in situ hybridization. J Endocrinol 2000;167:363–9.

39. Sarraf SA, Stancheva I. Methyl-CpG binding protein MBD1 coupleshistone H3 methylation at lysine 9 by SETDB1 to DNA replication andchromatin assembly. Mol Cell 2004;15:595–605.

40. Fuks F, Hurd PJ, Wolf D, Nan X, Bird AP, Kouzarides T. The methyl-CpG-binding protein MeCP2 links DNA methylation to histone meth-ylation. J Biol Chem 2003;278:4035–40.

41. Esteve PO. Direct interaction between DNMT1 and G9a coordinatesDNA and histone methylation during replication. Genes Dev 2006;20:3089–103.

42. Plath K, Fang J, Mlynarczyk-Evans SK, Cao R, Worringer KA, Wang H,et al. Role of histoneH3 lysine 27methylation in X inactivation. Science2003;300:131–5.

43. Muller J, Hart CM, Francis NJ, Vargas ML, Sengupta A, Wild B, et al.Histone methyltransferase activity of a Drosophila Polycomb grouprepressor complex. Cell 2002;111:197–208.

44. Leu YW, Yan PS, Fan M, Jin VX, Liu JC, Curran EM, et al. Lossof estrogen receptor signaling triggers epigenetic silencing ofdownstream targets in breast cancer. Cancer Res 2004;64:8184–92.

45. Li S, Washburn KA, Moore R, Uno T, Teng C, Newbold RR, et al.Developmental exposure to diethylstilbestrol elicits demethylation ofestrogen-responsive lactoferrin gene in mouse uterus. Cancer Res1997;57:4356–9.

46. KimMS,KondoT, Takada I, YounMY,YamamotoY, Takahashi S, et al.DNA demethylation in hormone-induced transcriptional derepression.Nature 2009;461:1007–12.

47. Dolinoy DC, Huang D, Jirtle RL. Maternal nutrient supplementationcounteracts bisphenol A-inducedDNAhypomethylation in early devel-opment. Proc Natl Acad Sci U S A 2007;104:13056–61.

48. Rubin BS, Murray MK, Damassa DA, King JC, Soto AM. PerinatalExposure to Low Doses of Bisphenol A Affects Body Weight, Patternsof Estrous Cyclicity, and Plasma LH Levels. Environ Health Perspect2001;109:675.

49. Newbold RR, Jefferson WN, Padilla-Banks E. Prenatal exposure tobisphenol a at environmentally relevant doses adversely affects themurine female reproductive tract Later in life. Environ Health Perspect2009;117:879–85.

50. Faber KA, Hughes CL Jr. Dose-response characteristics of neonatalexposure to genistein on pituitary responsiveness to gonadotropinreleasing hormoneand volumeof the sexually dimorphic nucleusof thepreoptic area (SDN-POA) in postpubertal castrated female rats.Reprod Toxicol 1993;7:35–9.

51. FernandezM,BianchiM, Lux-Lantos V, LibertunC. Neonatal exposureto bisphenol a alters reproductive parameters and gonadotropinreleasing hormone signaling in female rats. Environ Health Perspect2009;117:757–62.

52. Adachi T, Ono Y, Koh K-B, Takashima K, Tainaka H, Matsuno Y, et al.Long-term alteration of gene expression without morphologicalchange in testis after neonatal exposure to genistein in mice: toxico-genomic analysis using cDNA microarray. Food Chem Toxicol 2004;42:445.





2012;10:546-557. Mol Cancer Res K. Leigh Greathouse, Tiffany Bredfeldt, Jeffrey I. Everitt, et al. Methyltransferase EZH2 to Increase Risk of Uterine TumorigenesisEnvironmental Estrogens Differentially Engage the Histone

Updated version

http://mcr.aacrjournals.org/content/10/4/546

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2012/03/02/1541-7786.MCR-11-0605.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/10/4/546.full#ref-list-1

This article cites 52 articles, 20 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/10/4/546.full#related-urls

This article has been cited by 4 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mcr.aacrjournals.org/content/10/4/546To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/content/suppl/2012/03/02/1541-7786.MCR-11-0605.DC1

http://mcr.aacrjournals.org/content/10/4/546.full#ref-list-1

http://mcr.aacrjournals.org/content/10/4/546.full#related-urls

http://mcr.aacrjournals.org/cgi/alerts

mailto:[email protected]



environmental estrogens differentially engage the histone...

Documents