Uterine epithelial estrogen receptor α is dispensablefor proliferation but essential for completebiological and biochemical responsesWipawee Winuthayanona, Sylvia C. Hewitta, Grant D. Orvisb,c, Richard R. Behringerc, and Kenneth S. Koracha,1

aLaboratory of Reproductive and Developmental Toxicology, National Institute of Environmental Health Sciences, National Institutes of Health,Department of Health and Human Services, Research Triangle Park, NC 27709; bDevelopmental Biology Program, Sloan–Kettering Institute,York Avenue, New York, NY 10065; and cDepartment of Genetics, University of Texas M.D. Anderson Cancer Center, Houston, TX 77030

Edited by R. Michael Roberts, University of Missouri, Columbia, MO, and approved September 17, 2010 (received for review May 7, 2010)

Female fertility requires estrogen to specifically stimulate estrogenreceptor α (ERα)-dependent growth of the uterine epithelium inadult mice, while immature females show proliferation in bothstroma and epithelium. To address the relative roles of ERα inmediating estrogen action in uterine epithelium versus stroma, auterine epithelial-specific ERα knockout (UtEpiαERKO) mouse linewas generated by crossing Esr mice with Wnt7a-Cre mice. Expres-sion of Wnt7a directed Cre activity generated selective deletion ofERα in uterine epithelium, and female UtEpiαERKO are infertile.Herein, we demonstrate that 17β-estradiol (E2)-induced uterineepithelial proliferation was independent of uterine epithelial ERαbecause DNA synthesis and up-regulation of mitogenic mediatorswere sustained in UtEpiαERKO uteri after E2 treatment. IGF-1 treat-ment resulted in ligand-independent ER activation in both wild-type (WT) and UtEpiαERKO and mimicked the E2 stimulatory effecton DNA synthesis in uterine epithelium. Uterine epithelial ERα wasnecessary to induce lactoferrin, an E2-regulated secretory proteinselectively synthesized in the uterine epithelium. However, lossof uterine epithelial ERα did not alter the E2-dependent progester-one receptor (PR) down-regulation in epithelium. Strikingly, theuterine epithelium of UtEpiαERKO had robust evidence of apopto-sis after 3 d of E2 treatment. Therefore, we surmise that estrogeninduced uterine hyperplasia involves a dispensable role for uterineepithelial ERα in the proliferative response, but ERα is requiredsubsequent to proliferation to prevent uterine epithelial apoptosisassuring the full uterine epithelial response, illustrating the differ-ential cellular roles for ERα in uterine tissue and its contributionduring pregnancy.

conditional knockout ∣ paracrine regulation

Estrogens exert their actions through estrogen receptor (ER)proteins. ER alpha and beta (α and β) are distributed through-

out diverse tissues of the body. ERα is predominantly expressedin the uterus, mammary glands, pituitary, hypothalamus, andovarian theca cells, whereas ERβ is primarily in ovarian granulosacells, lung, and prostate (1). The uterus is a major target tissue forestrogen, and significant levels of ERα are expressed throughoutall the major uterine compartments (luminal and glandularepithelium, stroma, and the myometrium), and the expression le-vel of ERα varies during the estrous cycle and peaks during proes-trous in accordance with the elevation of circulating 17β-estradiol(E2) to elicit a proliferative response (2). Uterine proliferation isshown from a number of studies to be regulated by E2 in an ERα-dependent manner, as demonstrated by the lack of uterine stimu-lation and mitotic growth responses in αERKO uteri (3). Inimmature rodents, E2 stimulates the proliferation of both uterineluminal and glandular epithelium as well as stroma, whereas inadult mice, E2 stimulates the proliferation of only the luminaland glandular epithelium (4). A prior series of studies have pro-posed that estrogen acting on stromal cells initiates responses tomediate epithelial proliferation. Such evidence was generatedfrom neonatal tissue recombination studies in which E2-activated

uterine epithelial DNA synthesis was stromal ERα-dependent viainduction of autocrine or paracrine mitogenic factors in the stro-ma that stimulated the epithelial proliferation (5). Evaluating anepithelial-specific role for ERα in tissue proliferation was re-cently reported in the mammary gland. Use of a MMTV-Crecrossed with floxed ERαmice demonstrated that mammary glandgrowth and ductal epithelial cell proliferation required the mam-mary epithelial ERα (6), which contrasted with previous me-senchymal-epithelial tissue recombination studies that showedmammary epithelial ERα was dispensable for mammary epithe-lial cell proliferation (7). Therefore, we evaluated whether a si-milar discrepancy would occur in another estrogen proliferatingtissue or whether uterine tissue would follow a similar pattern asthe mammary gland regarding an indispensable role of ERα inuterine epithelial proliferation.

Several studies demonstrated that a number of growth factors,such as IGF-1, EGF, or transforming growth factor α (TGFα)(8–10) and their receptors were induced and activated, respec-tively, in the uterus under the influence of E2 treatment (8–12).IGF-1 was shown experimentally to be a key growth factor thatstimulates growth of the nearby epithelial cells as a result of theirestrogen-dependent expression in the uterus (13). Lack of IGF-1stimulation in αERKO mice demonstrated that ERα was re-quired for mediating the uterine IGF-1 action (14–16). The tran-scriptional factor CCAAT enhancer binding protein beta (C/EBPβ) is another estrogen-regulated gene in female reproductivetissues that is essential for uterine functions (17) and regulationof mammary epithelial cell proliferation and differentiation (18).

In addition to mitogenic effects, estrogen also plays a role incontrolling apoptosis (programmed cell death) in several celltypes (19). In the uterus, steroid hormones regulate apoptosis.Circulating levels of E2 are inversely correlated with apoptosisin uterine luminal and glandular epithelium during the estrouscycle, with the highest level of apoptosis in the epithelium atmetestrous as E2 levels decline (20). Estrogen was shown tosuppress apoptosis in the neonatal mouse uterine epitheliumvia an ERα-dependent induction of apoptosis inhibitory proteins(AIPs) including Birc1a (Baculoviral inhibitors of apoptosisrepeat-containing 1) and the inhibition of the activation of theapoptosis executioner, cysteine-aspartic protease-3 (caspase-3)(21). Herein, we evaluated the effect of E2 on adult uterineepithelial apoptosis when the ERα was selectively deleted in theuterine epithelium. The Wnt7a-Cre recombinase and Esr1-loxP

Author contributions: W.W., S.C.H., G.D.O., R.R.B., and K.S.K. designed research;W.W., S.C.H., and G.D.O. performed research; W.W., S.C.H., G.D.O., and K.S.K. analyzeddata; G.D.O. and R.R.B. contributed new reagents/analytic tools; and W.W., S.C.H.,G.D.O., and K.S.K. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: korach@niehs.nih.gov.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1013226107/-/DCSupplemental.

19272–19277 ∣ PNAS ∣ November 9, 2010 ∣ vol. 107 ∣ no. 45 www.pnas.org/cgi/doi/10.1073/pnas.1013226107

expression system was utilized in our study to address the effectof E2 on uterine epithelial responsiveness following selective lossof ERα.

ResultsGeneration of Uterine Epithelial-Specific ERα Knockout Mice. A uter-ine epithelial-specific ERα knockout (UtEpiαERKO) mousemodel was generated by crossing a transgenic mouse line expres-sing Wnt7a-Cre recombinase with our new line of floxed ERαmice containing loxP sites flanking exon 3 of the mouse Esr1 gene(22). The cell type specific expression of Wnt7a-Cre resulted incomplete deletion of ERα specifically in uterine luminal andglandular epithelial cells. ERα protein expression was evaluatedby immunohistochemistry (IHC) and was robustly expressedthroughout WT uteri (epithelium, stroma, and myometrium)(Fig. 1, Left). ERα protein was detected in stromal and myome-trial tissues but absent in the uterine luminal and glandularepithelium of UtEpiαERKO mice (Fig. 1, Right), demonstratingthat the Cre-specific deletion of ERα was effective. Test matingsof UtEpiαERKO females indicated that they are infertile(Table S1). Additional analysis indicates the females exhibitregular estrous cycles, and ovarian histology shows all stages offollicular development and indication of ovulation based onthe presence of corpus lutea (Fig. S1). However, implantationsites were observed in 6 of 9 WT but only 0 of 6 in UtEpiαERKOafter natural mating or embryo transfer into pseudopregnant fe-males, indicating that UtEpiαERKO mice have a compromiseduterine embryo receptivity. Accordingly, the estrogen-regulatedtranscripts Lif (leukemia inhibitory factor) and Ihh (Indianhedgehog), which are essential for embryo attachment (23, 24),remained unchanged in UtEpiαERKO but were up-regulated inWT after E2 treatment (Fig. S2).

Proliferative Effect of E2 on Uterine Epithelium Is Stromal ERα Depen-dent. E2 significantly increased the uterine wet weight in ovariec-tomized (OVEX) WT mice 24 h after treatment (p < 0.001)but not in UtEpiαERKO mice (Fig. 2A). Uterine weight was sig-nificantly increased by E2 in UtEpiαERKO after a 3-d bioassay;however, the increase was significantly lower than WT (Fig. 2B,p < 0.05), and the uterine proliferative response induced by E2

was fully attenuated by ICI 182,780 (ICI) in both WT and UtE-piαERKO mice, indicating the increase is an ER-dependentresponse. Incorporation of the deoxy-thymidine analog EdU(5-ethynyl-2′-deoxyuridine) demonstrated that E2 inducedDNA synthesis in luminal epithelium (Fig. 2C) at 24 h in both

WT and UtEpiαERKO uteri and decreased in both genotypesafter 3 d.

C/EBPβ and IGF-1: The Postulated Mediators of E2-Induced Prolifera-tion in Uterine Epithelium. Based on previous studies (5, 8, 9, 17),we evaluated whether Cebpb and Igf1 could be mediators of E2-induced uterine epithelial proliferation. E2 significantly up-regu-lated Cebpb at 2 h (Fig. 3A) and Igf1 at 6 h (Fig. 3B) in both WTand UtEpiαERKO uteri, and the expression in both genotypesdeclined to the basal level at 24 h. The expression levels of Cebpband Igf1 after E2 treatment in WT and UtEpiαERKO were notsignificantly different. Therefore, the uterine epithelial ERαwas not required for the estrogen induction of the uterine Cebpband Igf1.

We further investigated the effects of growth factor alone onuterine epithelial proliferation. Growth factor, such as IGF-1,was known to be the mitogenic factor that stimulated rodent uter-ine epithelium proliferation (9). To bypass the effect of E2, wetreated the animals with IGF-1 (50 μg∕mouse) in the absenceof E2. IGF-1 induced uterine luminal epithelial DNA synthesisto a lesser extent than with E2 treatment as illustrated by EdUincorporation at 24 h in both WT and UtEpiαERKO (Fig. 2C).Therefore, IGF-1 appears to be a possible downstream effector ofE2 action in the stroma that stimulates uterine epithelial prolif-eration.

Evaluation of E2-Specific Responses in Uterine Epithelium and Stroma.Lactoferrin (Ltf) is a protein regulated by E2 in mammalian uter-ine epithelium (3, 25). After 3 d, E2 significantly induced LtfmRNA in WT uteri (Fig. 4A, p < 0.001) and also increased

Fig. 1. ERα protein expression by IHC analysis in WT (Left) and UtEpiαERKO(Right) uteri with longitudinal sections (Upper; 40×magnification) and cross-ections (Lower; 200× magnification).

Fig. 2. Uterine proliferative response to E2 and IGF-1 inWTand UtEpiαERKOmice. The uterine wet weight increase after (A) 24 h or (B) 3-d (three conse-cutive days) treatment of vehicle, E2 in the presence or absence of ICI treat-ment, and (C) the fluorescent signal of Hoescht 33342 and EdU incorporation/Alexafluor488 staining in the uterine sections frommice treated with vehicle,E2 (24 h or 3 d), or IGF-1 (24 h). (Scale bar: 20 μm.) Results are mean� SEM(n ¼ 7). *, and ***p < 0.05 and 0.001, significantly different from vehiclecontrol of the corresponding genotype, respectively. +, p < 0.05 significantlydifferent from E2-treated WT mice.

Winuthayanon et al. PNAS ∣ November 9, 2010 ∣ vol. 107 ∣ no. 45 ∣ 19273

BIOCH

EMISTR

Y

Ltf protein in the uterine luminal and glandular epithelium(Fig. 4B, arrowheads), and these increases were inhibited byICI (Fig. 4A and B). Loss of ERα in UtEpiαERKO uteri resultedin loss of the E2-stimulated lactoferrin expression (Fig. 4A and B).Furthermore, the progesterone receptor (PR) is essential forestablishment and maintenance of pregnancy (26) and is alsoregulated by E2 in the uterus (27). Therefore, we evaluated thePR protein expression, which showed that PR was down-regu-lated in the epithelium and increased in the stroma followingE2 treatment in WT mice. This PR response was not mediatedvia the uterine epithelial ERα since the response was sustainedin UtEpiαERKO (Fig. 4C). Moreover, the inhibitory effect of ICIin both genotypes further indicated that stromal ERα appearedresponsible for the E2 down-regulation of PR in the uterineepithelium (Fig. 4C). Other examples of uterine estrogen respon-sive transcripts were evaluated as well (Fig. S3) and indicate lossof some responses, such as Aqp8 (aquaporin 8) and Cdkn1a(cyclin dependent kinase inhibitor A or p21), and retention ofothers, such as Dhcr24 (24-dehydrocholesterol reductase) andUbe2c (ubiquitin-conjugating enzyme E2C).

Loss of ERα in Uterine Epithelium Enhances Uterine Apoptosis. Loss ofERα in uterine epithelium in combination with the growth re-sponses following 3 d of E2 treatment altered the appearanceof the epithelia of UtEpiαERKO (Fig. 5A, arrowheads) in a man-ner suggesting increased apoptosis in UtEpiαERKO compared toWT. Consistent with this, E2 treatment significantly induced theapoptosis inhibitor Birc1a expression in WT (p < 0.001) but notin UtEpiαERKO uteri (Fig. 5B). Cleaved caspase-3 (the activeform of caspase-3) was slightly increased by 3 d of treatment withE2 in WT (Fig. 5C, yellow arrowheads) but intensively detected inUtEpiαERKO luminal and glandular epithelium as dense, roundspots on both basolateral and luminal sides of the epithelial cells(Fig. 5C, black arrowheads). However, increased levels of theDNA fragmentation marker, TUNEL (Terminal deoxynucleoti-dyltransferase-mediated dUTP nick end labeling), was foundonly in UtEpiαERKO uteri, which further suggested increasedapoptosis following 3 d of E2 treatment (Fig. 5D, arrowheads).Therefore, loss of the uterine epithelial ERα correlated withincreased uterine apoptosis subsequent to the uterine growthinduced after E2 treatment for 3 d.

DiscussionHerein, we illustrated the physiological relevance of ERα to uter-ine epithelial proliferation and apoptosis under the influence ofE2 regulation. We demonstrated that uterine epithelial ERα wasdispensable for the E2-induced uterine epithelial proliferationin adult OVEX mice in contrast to what has been reported forthe mammary gland epithelial proliferation (6). In the absenceof uterine epithelial ERα, potential mitogenic mediators suchas Cebpb and Igf1 were increased with E2 treatment, and thegrowth factor IGF-1 could bypass the effect of E2 on uterineepithelial proliferation. We confirmed the uterine epithelialE2-specific loss of response by showing that deletion of ERα

in uterine epithelium caused the loss the lactoferrin production.However, the uterine epithelial down-regulation and stromalredistribution of PR by E2 remained intact in UtEpiαERKOmice. Surprisingly, lack of functional ERα in the uterine epithe-lium had a pronounced effect on apoptosis.

E2 is known to be a robust mitogenic ovarian hormone inuterine tissues that induces a pronounced hyperplastic tissueresponse, consistent with endometrial development during the es-trous cycle (28, 29). Staged cell labeling studies showed an effectof E2 on recruiting uterine epithelial cells from a resting state(G0) into the cell cycle and shortened both G1 and S-phasesof the cell cycle to elicit proliferation (4, 30). Administrationof E2 induces the proliferation in both uterine epithelium andstroma of immature mice but only stimulates the proliferationin uterine epithelium in adult mice, and the cause of this lackof stromal cell responsiveness has been linked to progesteroneactivity (4). Previous work using neonatal tissue xenograph mod-els and extrapolation of such a model to adult tissue (5) indicatethat the proliferative effect of E2 on uterine epithelium in adultmice is independent of ERα expression in the luminal epithelium,suggesting the stromal ERα expression is crucial for the mitoticresponse of E2. To evaluate and extend those observations in anatural adult tissue system in vivo, our UtEpiαERKO mousemodel was generated with the specific deletion of ERα in uterineepithelium. Adult UtEpiαERKOmice had a complete deletion ofERα protein in the uterine luminal and glandular epitheliumthroughout the uterine horn. The classical uterine bioassay wasutilized to illustrate the estrogenic effect of E2 on uterine growthresponses at both 24 h and 3 d. Absence of ERα in the uterineepithelium compromised the uterine weight increase induced byE2 after 24 h or 3 d, which suggests the epithelial ERα contributesto a full uterine growth response. Our results suggest that the in-itial (24 h) proliferation response is similar but that the luminaland glandular epithelial cells do not subsequently develop prop-erly into multilayered tissue structures secreting lactoferrin andadditionally show indications of apoptosis. These deficienciesand alterations most likely account for the hampered increasein uterine weight. However, the induction of DNA synthesis after24 h and the decline after 3 d of E2 treatment was seen in bothWTand UtEpiαERKO. Therefore, uterine epithelial ERα in vivowas dispensable for appropriate mitotic proliferative responsefrom E2, consistent with earlier xenograph experiments (5), incontrast to the requirement for epithelial ERα in mammary glanddevelopment and proliferation (6).

Next, we evaluated the possible mitogenic mediators (C/EBPβand IGF-1) downstream from E2. Previous studies showed thatthe transcription factor C/EBPβ is rapidly increased in bothepithelial and stromal cells by E2 in the uterus (17). The essentialrole of C/EBPβ in uterine epithelial proliferation has been de-monstrated by the decreased E2-dependent uterine epithelialproliferation as a result of impaired DNA replication in Cebpbknockout mice (17, 31). IGF-1 has been demonstrated to be aparacrine mitogenic effector secreted from uterine stroma to sti-mulate the proliferation of uterine epithelium (5). E2 increases

Fig. 3. E2-regulated genes in WT and UtEpiαERKO uteri. (A) Cebpb mRNA at 2 h and 24 h, and (B) Igf1 mRNA at 6 h and 24 h after vehicle or E2 treatment(n ¼ 4). Results are mean� SEM. *, **, and ***p < 0.05, 0.01, and 0.001, significantly different from vehicle control of the corresponding genotype, respec-tively. ns; not significantly different in E2 treated for 2 h (A) and 6 h (B) in WT and UtEpiαERKO, analyzed by nonparametric Mann–Whitney test.

19274 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1013226107 Winuthayanon et al.

production of growth factors such as IGF-1 and EGF, and thesegrowth factors induce uterine DNA synthesis (8–10, 16). Igf1knockout mice exhibited impaired uterine response to estrogen,

which implicated the essential role of IGF-1 on uterine growth(32). Overexpression of IGF binding protein-1 (IGFBP-1) in

Fig. 4. Uterine epithelial-specific E2-regulated responses. Ltf expression byusing real-time PCR analysis (A), IHC analysis of lactoferrin (B), and proges-terone receptor (C) protein expression in WT (Left) and UtEpiαERKO (Right)uteri when treated with vehicle, E2 or E2 þ ICI for three consecutive days(n ¼ 7). (Scale bar: 40 μm.) Arrowheads indicate the positive staining of lac-toferrin. Results are mean� SEM. ***p < 0.001, significantly different fromvehicle of the corresponding genotype.

Fig. 5. Apoptosis-related mediator expressions in WTand UtEpiαERKO uteri.After the treatment of vehicle or E2 for three consecutive days, H&E staining(A) indicated apoptotic appearance of epithelial cells in E2-treated UtEpiαER-KO uteri (black arrowheads). Birc1a expression by using real-time PCR analysis(B) increased only in E2-treated WT. Cleaved Caspase-3 IHC (C) increased inboth WT (yellow arrowheads) and UtEpiαERKO (black arrowheads) uteri,and TUNEL IHC (D) increased only in E2-treated UtEpiαERKO uteri (black ar-rowheads) (n ¼ 7). (Scale bar: 20 μm.) Results are mean� SEM. ***p < 0.001,significantly different from vehicle control of the corresponding genotype,respectively.

Winuthayanon et al. PNAS ∣ November 9, 2010 ∣ vol. 107 ∣ no. 45 ∣ 19275

BIOCH

EMISTR

Y

the mouse uterus disrupted DNA synthesis, but the uterine wetweight increased in response to E2 (33). Therefore, the local pro-duction and availability of IGF-1 appears to play a role in uterineproliferation. Moreover, evidence also suggests the proliferativeeffect of E2 on IGF-1 production in the human epitheliumwas mediated in an autocrine fashion in human uterine epithelialcells (34). Further evidence came from the use of a growth factorinhibitor in combination with E2 administration, which partly in-hibited the action of estrogen on the uterine weight increase (13).We previously showed that IGF-1 treatment stimulated the uter-ine epithelial proliferation in an ERα dependent manner (16). Inthis study, the complete deletion of ERα specifically in the epithe-lial target cell of the uterus did not alter DNA synthesis after 24 hof either E2 or IGF-1 treatment. Therefore, in the natural tissueenvironment, ERα residing in the stroma, not uterine epithelium,was the functional ERα that responded to E2 to produce thegrowth factor(s), which was released and acted as paracrine med-iator(s) to promote DNA synthesis in the epithelium. As men-tioned above, these findings in vivo are consistent with previousstudies using neonatal tissue recombination experiments (5).

To substantiate the absence of ERα activity in the uterineepithelium, we evaluated gene regulation of Ltf as an ERα-dependent estrogen-regulated response in uterine epithelium(35). Expression of lactoferrin has been correlated with the cir-culating level of E2 (36). Our result showed that the up-regulationof Ltf mRNA and Ltf protein was a uterine epithelial-specificERα-dependent action, which was consistent with the previousfinding in a tissue recombinant study (37) and supports theabsence of ERα activity in the epithelium. We found that E2

redistributed the PR expression from epithelium to stroma inUtEpiαERKO in a similar manner as in WT. These findings againconcur with the previous studies using neonatal tissue recombi-nation models (38, 39). Therefore, the estrogen-dependent re-sponsiveness for redistribution of PR from uterine epitheliumto stroma upon E2 treatment is intact and appears to be mediatedby stromal ERα.

During the female reproductive cycle, estrogen stimulates theproliferative phase to facilitate implantation and pregnancy. Inrodents, the proliferate epithelial layer is eliminated by inducingapoptosis in situ (40). The proliferation and subsequent elimina-tion of uterine epithelial cells during the estrous cycle is regulatedby the relative mitotic and apoptotic rates under the control ofovarian hormones, especially the circulating level of E2 (40).Withdrawal of E2 treatment from newborn mice results in a highapoptosis index in the uterus (41). AIPs have a protective effectagainst the activation of the essential apoptosis mediator caspase-3, which is a member of a class of cysteine-aspartyl proteases (42).Caspases are also present in normal healthy cells as inactivatedenzymes and are activated by cleavage to initiate apoptosis andinduce cell death (43). In our study, the increase in active caspase-3 in WT may result from the normal regulation between growthand apoptosis after E2-induced proliferation of the uterineepithelium. After diethylstilbestrol treatment, uteri of neonatalERα knockout mice exhibited severe apoptosis, whereas WTmice sustained the apoptotic protection (21), thereby indicatingthat estrogen protection against uterine epithelial apoptosis ismediated via ERα. In this study, we showed that after the growthinduced by E2 treatment, UtEpiαERKO epithelia exhibited

severe apoptosis as shown by increases in active caspase-3 andTUNEL staining. Moreover, in the absence of uterine epithelialERα, E2-mediated increase of the apoptotic inhibitor, Birc1a, isdramatically attenuated. As mentioned earlier, C/EBPβ is a cri-tical mediator of E2-stimulated proliferative response in uterineepithelial cells. C/EBPβ knockout mice undergo uterine epithelialapoptosis as a result of G1-S-phase arrest (31), demonstrating aneed for C/EBPβ in DNA replication and damage response. Inour study the level of C/EBPβ after E2 treatment in UtEpiαER-KO is comparable to that of in WT. However, we have not eval-uated the relative levels of C/EBPβ protein in stromal andepithelial cells, and it is possible that there is a deficiency of thisimportant factor intrinsic to the epithelial cells of UtEpiαERKOfollowing uterine growth. Although the uterine epithelial apop-tosis in UtEpiαERKO following E2 treatment may not be due toan altered C/EBPβ regulation, it may alternatively result from adefect in other mediators of DNA damage response. Therefore,we postulate that during the adult reproductive cycle followingthe E2 stimulated growth, the lack of ERα and E2 action inthe epithelial cells impairs responses initiated by unknownE2-dependent secondary mediators resulting in apoptosis ofthe epithelial cells. Thus, epithelial apoptosis occurs due tothe absence of epithelial ERα. Interestingly, normal regulationof apoptosis has been shown to be important for implantation(44). Therefore, alteration of growth and apoptosis as a resultof the loss of uterine epithelial ERα leads to the infertility phe-notype in UtEpiαERKO. More complete fertility studies will beconducted to address additional details regarding the roles ofuterine epithelial ERα.

Overall, this model indicates that uterine epithelial prolifera-tion is mediated by ERα-dependent mitotic proliferative signalsecreted from the stroma but that ERα is needed followingmitosis within epithelial cells to minimize apoptosis. Therefore,ERα is required in both tissue compartments for a completeuterine growth response. It is surprising that epithelial ERα isrequired for mammary gland proliferation but is dispensable inuterine response. Further studies will be required to determinehow this differential cellular role of ER may change in diseasessuch as endometriosis, endometrial hyperplasia, or cancer.

Materials and MethodsDetailed methods appear in SI Materials and Methods. These include genera-tion of a Wnt7a-Cre transgenic mouse line, generating uterine epithelial-specific ERα knockout mice, chemicals and compounds, uterine bioassay,RNA isolation, and real-time PCR analysis for gene expression, histologicaland IHC analysis of uteri, and statistical analysis.

ACKNOWLEDGMENTS. We thank James Clark, Page Myers, David Gouldingfrom the National Institute of Environmental Health Sciences (NIEHS) AnimalSurgery Group of Comparative Medicine branch for performing the animalsurgeries and embryo transfer analysis, David Monroy for animal care,Geoffrey Hurlburt and Dave Olsen (NIEHS Immunohistochemistry core) forcleaved caspase-3 and TUNEL staining, Casey Reed for mouse genotypingand Drs. April Binder and Diane Klotz for the critical reading of the manu-script and helpful suggestions. This research was supported by the NationalInstitute of Environmental Health Sciences, Division of Intramural Research(funding to W.W., S.C.H., and K.S.K.) project Z01ES70065 as well as NIHHD30284 and CA098258 (SPORE in Uterine Cancer) to R.R.B. G.D.O. was sup-ported by the National Cancer Institute CA09299 Training Program in theMolecular Genetics of Cancer.

1. Couse JF, Korach KS (1999) Estrogen receptor null mice: What have we learned and

where will they lead us? Endocr Rev 20:358–417.

2. Wang H, Eriksson H, Sahlin L (2000) Estrogen receptors alpha and beta in the female

reproductive tract of the rat during the estrous cycle. Biol Reprod 63:1331–1340.

3. Hewitt SC, et al. (2003) Estrogen receptor-dependent genomic responses in the uterus

mirror the biphasic physiological response to estrogen.Mol Endocrinol 17:2070–2083.

4. Quarmby VE, Korach KS (1984) The influence of 17 beta-estradiol on patterns of cell

division in the uterus. Endocrinology 114:694–702.

5. Cooke PS, et al. (1997) Stromal estrogen receptors mediate mitogenic effects of estra-

diol on uterine epithelium. Proc Natl Acad Sci USA 94:6535–6540.

6. Feng Y, Manka D, Wagner KU, Khan SA (2007) Estrogen receptor-alpha expression inthe mammary epithelium is required for ductal and alveolar morphogenesis in mice.Proc Natl Acad Sci USA 104:14718–14723.

7. Cunha GR, et al. (1997) Elucidation of a role for stromal steroid hormone receptors inmammary gland growth and development using tissue recombinants. J MammaryGland Biol 2:393–402.

8. DiAugustine RP, et al. (1988) Influence of estrogens on mouse uterine epidermalgrowth factor precursor protein and messenger ribonucleic acid. Endocrinology122:2355–2363.

9. Murphy LJ, Ghahary A (1990) Uterine insulin-like growth factor-1: Regulation of ex-pression and its role in estrogen-induced uterine proliferation. Endocr Rev 11:443–453.

19276 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1013226107 Winuthayanon et al.

10. Nelson KG, et al. (1992) Transforming growth factor-alpha is a potential mediator ofestrogen action in the mouse uterus. Endocrinology 131:1657–1664.

11. Kahlert S, et al. (2000) Estrogen receptor alpha rapidly activates the IGF-1 receptorpathway. J Biol Chem 275:18447–18453.

12. Hom YK, et al. (1998) Uterine and vaginal organ growth requires epidermal growthfactor receptor signaling from stroma. Endocrinology 139:913–921.

13. Zhu L, Pollard JW (2007) Estradiol-17beta regulates mouse uterine epithelial cellproliferation through insulin-like growth factor 1 signaling. Proc Natl Acad Sci USA104:15847–15851.

14. Hewitt SC, Collins J, Grissom S, Deroo B, Korach KS (2005) Global uterine genomics invivo: Microarray evaluation of the estrogen receptor alpha-growth factor cross-talkmechanism. Mol Endocrinol 19:657–668.

15. Curtis SW, et al. (1996) Physiological coupling of growth factor and steroid receptorsignaling pathways: Estrogen receptor knockout mice lack estrogen-like response toepidermal growth factor. Proc Natl Acad Sci USA 93:12626–12630.

16. Klotz DM, et al. (2002) Requirement of estrogen receptor-alpha in insulin-like growthfactor-1 (IGF-1)-induced uterine responses and in vivo evidence for IGF-1/estrogenreceptor cross-talk. J Biol Chem 277:8531–8537.

17. Mantena SR, et al. (2006) C/EBPbeta is a critical mediator of steroid hormone-regulated cell proliferation and differentiation in the uterine epithelium and stroma.Proc Natl Acad Sci USA 103:1870–1875.

18. Robinson GW, Johnson PF, Hennighausen L, Sterneck E (1998) The C/EBPbeta transcrip-tion factor regulates epithelial cell proliferation and differentiation in the mammarygland. Genes Dev 12:1907–1916.

19. Song RX, Santen RJ (2003) Apoptotic action of estrogen. Apoptosis 8:55–60.20. Wood GA, Fata JE, Watson KL, Khokha R (2007) Circulating hormones and estrous

stage predict cellular and stromal remodeling in murine uterus. Reproduction133:1035–1044.

21. Yin Y, et al. (2008) Estrogen suppresses uterine epithelial apoptosis by inducing birc1expression. Mol Endocrinol 22:113–125.

22. Dupont S, et al. (2000) Effect of single and compound knockouts of estrogen receptorsalpha (ERalpha) and beta (ERbeta) on mouse reproductive phenotypes. Development127:4277–4291.

23. Stewart CL, et al. (1992) Blastocyst implantation depends on maternal expression ofleukaemia inhibitory factor. Nature 359:76–79.

24. Lee K, et al. (2006) Indian hedgehog is a major mediator of progesterone signaling inthe mouse uterus. Nat Genet 38:1204–1209.

25. Teng CT (2002) Lactoferrin gene expression and regulation: An overview. Biochem CellBiol 80:7–16.

26. Conneely OM, Mulac-Jericevic B, DeMayo F, Lydon JP, O'Malley BW (2002) Reproduc-tive functions of progesterone receptors. Recent Prog Horm Res 57:339–355.

27. Tibbetts TA, Mendoza-Meneses M, O'Malley BW, Conneely OM (1998) Mutual andintercompartmental regulation of estrogen receptor and progesterone receptorexpression in the mouse uterus. Biol Reprod 59:1143–1152.

28. Martin L, Finn CA, Trinder G (1973) Hypertrophy and hyperplasia in the mouse uterusafter oestrogen treatment: An autoradiographic study. J Endocrinol 56:133–144.

29. Martin L, Pollard JW, Fagg B (1976) Oestriol, oestradiol-17beta and the proliferationand death of uterine cells. J Endocrinol 69:103–115.

30. Sutherland RL, Reddel RR, Green MD (1983) Effects of oestrogens on cell proliferationand cell cycle kinetics. A hypothesis on the cell cycle effects of antioestrogens. Eur JCancer Clin Oncol 19:307–318.

31. Ramathal C, Bagchi IC, BagchiMK (2010) Lack of CCAATenhancer binding protein beta(C/EBPbeta) in uterine epithelial cells impairs estrogen-induced DNA replication,induces DNA damage response pathways, and promotes apoptosis. Mol Cell Biol30:1607–1619.

32. Baker J, et al. (1996) Effects of an Igf1 gene null mutation onmouse reproduction.MolEndocrinol 10:903–918.

33. Rajkumar K, Dheen T, Krsek M, Murphy LJ (1996) Impaired estrogen action in theuterus of insulin-like growth factor binding protein-1 transgenic mice. Endocrinology137:1258–1264.

34. Kashima H, et al. (2009) Autocrine stimulation of IGF1 in estrogen-induced growth ofendometrial carcinoma cells: involvement of the mitogen-activated protein kinasepathway followed by up-regulation of cyclin D1 and cyclin E. Endocr-Relat Cancer16:113–122.

35. Jefferson WN, Padilla-Banks E, Newbold RR (2000) Lactoferrin is an estrogen respon-sive protein in the uterus of mice and rats. Reprod Toxicol 14:103–110.

36. Walmer DK, Wrona MA, Hughes CL, Nelson KG (1992) Lactoferrin expression in themouse reproductive tract during the natural estrous cycle: Correlation with circulatingestradiol and progesterone. Endocrinology 131:1458–1466.

37. Buchanan DL, et al. (1999) Tissue compartment-specific estrogen receptor-alphaparticipation in the mouse uterine epithelial secretory response. Endocrinology140:484–491.

38. Kurita T, Lee KJ, Cooke PS, Lydon JP, Cunha GR (2000) Paracrine regulation ofepithelial progesterone receptor and lactoferrin by progesterone in themouse uterus.Biol Reprod 62:831–838.

39. Kurita T, et al. (2000) Paracrine regulation of epithelial progesterone receptor byestradiol in the mouse female reproductive tract. Biol Reprod 62:821–830.

40. Sato T, et al. (1997) Apoptotic cell death during the estrous cycle in the rat uterus andvagina. Anat Rec 248:76–83.

41. Jo T, Terada N, Saji F, Tanizawa O (1993) Inhibitory effects of estrogen, progesterone,androgen and glucocorticoid on death of neonatal mouse uterine epithelial cellsinduced to proliferate by estrogen. J Steroid Biochem Mol Biol 46:25–32.

42. Deveraux QL, Takahashi R, Salvesen GS, Reed JC (1997) X-linked IAP is a direct inhibitorof cell-death proteases. Nature 388:300–304.

43. Taylor RC, Cullen SP, Martin SJ (2008) Apoptosis: Controlled demolition at the cellularlevel. Nat Rev Mol Cell Biol 9:231–241.

44. Pampfer S, Donnay I (1999) Apoptosis at the time of embryo implantation in mouseand rat. Cell Death Differ 6:533–545.

Winuthayanon et al. PNAS ∣ November 9, 2010 ∣ vol. 107 ∣ no. 45 ∣ 19277

BIOCH

EMISTR

Y