Role of estrogen receptor � in uterine stromaand epithelium: Insights from estrogenreceptor ��/� miceOsamu Wada-Hiraike*, Haruko Hiraike*, Hiroko Okinaga†, Otabek Imamov*, Rodrigo P. A. Barros*, Andrea Morani*,Yoko Omoto*, Margaret Warner*, and Jan-Åke Gustafsson*‡

*Department of Biosciences and Nutrition, Karolinska Institutet, S-141 86 Novum, Sweden; and †Department of Internal Medicine, Teikyo University Schoolof Medicine, 2-11-1 Kaga, Itabashi-ku, Tokyo 173-8605, Japan

Contributed by Jan-Åke Gustafsson, October 6, 2006 (sent for review September 19, 2006)

In this study, we compared the uterine tissue of estrogen receptor(ER)��/� mice and their WT littermates for differences in morphol-ogy, proliferation [the percentage of labeled cells 2 h after BrdUrdinjection and EGF receptor (EGFR) expression], and differentiation(expression of progesterone receptor, E-cadherin, and cytokera-tins). In ovariectomized mice, progesterone receptor expression inthe uterine epithelium was similar in WT and ER��/� mice, butE-cadherin and cytokeratin 18 expression was lower in ER��/�

mice. The percentage of cells in S phase was 1.5% in WT mice and8% in ER��/� mice. Sixteen hours after injection of 17�-estradiol(E2), the number of BrdUrd-labeled cells increased 20-fold in WTmice and 80-fold in ER��/� mice. Although ER� was abundant inintact mice, after ovariectomy, ER� could not be detected in theluminal epithelium of either WT or ER��/� mice. In both untreatedand E2-treated mice, ER� and ER� were colocalized in the nuclei ofmany stromal and glandular epithelial cells. However, upon E2 �progesterone treatment, ER� and ER� were not coexpressed in anycells. In WT mice, EGFR was located on the membranes and in thecytoplasm of luminal epithelium, but not in the stroma. In ER��/�

mice, there was a marked expression of EGFR in the nuclei ofepithelial and stromal cells. Upon E2 treatment, EGFR on cellmembranes was down-regulated in WT but not in ER��/� mice.These findings reveal an important role for ER� in response to E2

and in the organization, growth, and differentiation of the uterineepithelium.

differentiation � proliferation � uterus

Proliferation of epithelial cells (1, 2), uterine hyperemia, f luiduptake (termed water imbibition), recruitment of inflam-

matory leukocytes from the bloodstream into the stromal com-partment (3, 4), and the induction of the progesterone receptor(PR) are caused by 17�-estradiol (E2). Progesterone (P4) inhibitsestrogen-induced cell proliferation of the luminal and glandularepithelial compartment (5) and stimulates epithelial differenti-ation in preparation for embryo implantation. Physiologicallyand pharmacologically, P4 is important for prevention of endo-metrial hyperplasia (6). After the cessation of ovarian function,estrogen replacement is needed for preservation of the skeletal,cardiovascular, and central nervous systems. To oppose theproliferative effects of estrogen on the uterus and reduce the riskof endometrial cancer, progesterone is used together with es-trogen in hormone replacement therapy after menopause (7, 8).

Most of the known actions of estrogen are mediated by theestrogen receptors (ERs) ER� and ER�, both of which bind E2and modulate transcription of E2-responsive genes (9). A singleinjection of E2 results in a synchronized wave of cell prolifera-tion, with DNA synthesis in epithelial cells beginning 6–9 h afterE2 injection and peaking at 12–15 h. DNA synthesis is followedby a wave of cell division (1, 2, 10). P4 elicits its function throughbinding to the PR.

During the secretory phase of the estrus cycle, endometrialstroma becomes predecidual in preparation for decidualization

if pregnancy occurs. Decidualization can be induced by P4 andE2 and involves EGF action on the stromal EGF receptor(EGFR). E2-induced epithelial cell proliferation is mediated bystromal ER�. It is currently thought that in the uterus, stromalER� induces secretion of EGF. Although ER� is the predom-inant ER in the adult rodent uterus, ER� mRNA has also beendetected in WT and ER��/� mouse uteri (11). Targeted dis-ruption of ER� in mice (12) has revealed roles for ER� indifferentiation of various epithelial and nonepithelial cell typesin many tissues and organs (13–19). In the uteri of ER��/� mice,there is an exaggerated responsiveness to E2, resulting in en-largement of the lumen, increase in volume and protein contentof uterine secretion, and induction of aberrant epithelial expres-sion of PR after E2 injection (13). ER� is expressed in all celltypes of uterine tissue and modulates estrogenic function byinhibiting ER� function (13, 20). ER� has been reported to playa role in decidualization of rat uterus (21) and in human cervicalripening, which is essential for parturition (22).

To better understand the role of ER� in the regulation of uterinetissue development, we have compared the uteri of WT andER��/� mice by observing markers of proliferation and differen-tiation before and after treatments with E2 alone or E2 � P4.

ResultsE-Cadherin Expression and ER��/� Uterus Epithelium. Uterus sec-tions from ER��/� mice and their WT littermates were stainedwith antibodies specific for E-cadherin. Expression of E-cadherin on the lateral surfaces of the epithelial cells wasmarkedly lower in untreated ER��/� mice than in WT mice (Fig.1 A and D). Upon treatment with E2 or with E2 � P4, expressionof E-cadherin increased in both WT and ER��/� mice (Fig. 1 Band C and E and F, respectively). Visualization of E-cadherin onthe epithelial membrane and cell borders permitted a clearpicture of the shape of the epithelial cells and revealed that inER��/� mice, the lateral surfaces of the epithelial cells weredeformed.

Hyperresponsiveness of ER��/� Uterus to Proliferative Effects of E2.BrdUrd was injected 14 h after injection of E2. DNA synthesisin the epithelial cells in the uterus commences �6 h after E2injection and peaks at 12 to 15 h (1, 10). E2 treatment ofovariectomized adult mice resulted in uterine luminal epithelialcell proliferation in both WT (Fig. 2 A and B) and ER��/� (Fig.

Author contributions: O.W.-H., H.O., O.I., A.M., and M.W. designed research; O.W.-H., H.H.,O.I., R.P.A.B., and Y.O. performed research; O.W.-H., M.W., and J.-Å.G. analyzed data; andO.W.-H., M.W., and J.-Å.G. wrote the paper.

Conflict of interest statement: J.-Å.G. is a cofounder, consultant, deputy board member,and shareholder of Karo Bio AB.

Abbreviations: CK, cytokeratin; E2, 17�-estradiol; ER, estrogen receptor; EGFR, epidermalgrowth factor receptor; P4, progesterone; PR, progesterone receptor.

‡To whom correspondence should be addressed. E-mail: jan-ake.gustafsson@mednut.ki.se.

© 2006 by The National Academy of Sciences of the USA

18350–18355 � PNAS � November 28, 2006 � vol. 103 � no. 48 www.pnas.org�cgi�doi�10.1073�pnas.0608861103

Dow

nloa

ded

by g

uest

on

Aug

ust 2

9, 2

020

2 C and D) littermates. No stromal BrdUrd-labeled cells wereobserved. Consistent with the earlier report (13), there weremore BrdUrd-stained cells in the epithelium of ER��/� than inWT mice. After administration of E2, almost all of the luminalepithelial cells of ER��/� mice were BrdUrd-labeled (Fig. 2 Dand E).

Differentiation of Cells in the Uterine Endometrium and Cervix inER��/� and WT Littermates. In the absence of ovarian hormones,expression of cytokeratin (CK) 18 in the luminal and glandularepithelium was similar in WT and ER��/� mice (Fig. 3 A and D).Upon hormonal treatment, expression of CK18 in uteri showeda slight reduction in WT mice but was significantly decreased inER��/� mice (Fig. 3 B and C and E and F, respectively). Incontrast, expression of CK13, which is preferentially expressedin cells at the uterine cervix, was similar in WT and ER��/�

mice, and hormonal treatment resulted in no remarkablechanges as determined by using immunohistochemical staining(Fig. 3 G–L).

Colocalization of ER� and ER� in Uterine Tissue. Cellular localizationof ER� and ER� was studied in the uteri of WT mice treatedwith vehicle (Fig. 4 A and D), E2 (Fig. 4 B and E), or E2 � P4(Fig. 4 C and F). In Fig. 4, Cy3-conjugated anti-chicken IgG (red)detects ER�, and FITC-conjugated anti-rabbit IgG (green)detects ER�. Cells in which ER� and ER� are colocalized arestained yellow or orange (arrows). ER� was abundantly ex-pressed in intact WT mice (Fig. 4M), but in vehicle-treatedovariectomized mice, ER� (but not ER�) was expressed in theluminal epithelium (Fig. 4A). Both receptors were expressed inthe stroma and glandular epithelium. In the glandular epithe-lium, the two receptors colocalized in the same nuclei, but in thestroma, they were in separate cells (Fig. 4 A and D). Upon E2treatment, ER� was lost from the luminal epithelium, whereasthe two receptors were colocalized in nuclei of the stroma andglandular epithelium (Fig. 4 B and E). When mice were treated

Fig. 1. Immunohistochemical detection of E-cadherin. Nine-week-old micereceived a single s.c. injection of 150 �l of olive oil containing no hormone (Aand D), 100 ng of E2 (B and E), or 100 ng of E2 � 1 mg of P4 (C and F), and 16 hlater, expression of E-cadherin in the uterus was examined. As expected,E-cadherin immunoreactivity was confined to the epithelial compartment. Itsexpression was markedly lower in vehicle-treated ER��/� than in WT mice (Aand D). (Insets) Outlined areas at 4 times greater magnification. Treatmentwith E2 restored E-cadherin levels in ER��/� mice.

Fig. 2. Detection of S-phase cells in uteri of WT and ER��/� littermates.Sixteen hours before being killed, mice received a single injection of vehicle (Aand C) or 100 ng of E2 (B and D). Two hours before being killed, mice wereinjected i.p. with BrdUrd at 30 mg�kg. In vehicle-treated WT mice, 1.5% ofepithelial cells were labeled, whereas 8% were labeled in ER��/� mice. Afteradministration of E2, 23% of luminal epithelial cells were labeled in WT mice,and 82% were labeled in ER��/� mice (B, D, and E).

Fig. 3. Immunohistochemical detection of CK18 (epithelial marker) andCK13 (uterine cervical marker) in the uteri of WT and ER��/� littermates. Micereceived a single injection with vehicle (A, D, G, and J), 100 ng of E2 (B, E, H, andK), or 100 ng of E2 � 1 mg of P4 (C, F, I, and L). The differentiation marker CK18is confined to luminal and glandular cells in the uteri of WT (A–C) and ER��/�

mice (D–F). In vehicle-treated mice, CK18 expression was lower in ER��/� thanin WT mice (A and D). Treatment with E2 or E2 � P4 resulted in reduction ofCK18 expression in WT mice (B and C) but complete loss CK18 in ER��/� mice(E and F). Expression of CK13 in the uterine cervix was similar in WT (G–I) andER��/� mice (J–L).

Wada-Hiraike et al. PNAS � November 28, 2006 � vol. 103 � no. 48 � 18351

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 2

9, 2

020

with E2 � P4, only ER� was detected in the glandular epithelium(Fig. 4F). In the stroma, many cells were ER�-positive, and a fewwere ER�-positive, but the two receptors were not coexpressedin the same cells (Fig. 4C). Expression of ER� was studied in theuteri of WT and ER��/� mice treated with vehicle (Fig. 4 G andJ), E2 (Fig. 4 H and K), or E2 � P4 (Fig. 4 I and L). Expressionof ER� showed similar pattern in both genotypes but tended tobe more intense in ER��/� mice. Surprisingly, after ovariectomy,ER� could not be detected in the luminal epithelium of eitherWT or ER��/� mice (Fig. 4 G–L).

PR as a Stromal Differentiation Marker in ER��/� and WT Littermates.In vehicle-treated mice there was intense staining for PR in themajority of luminal and glandular epithelial cells in both WT andER��/� mice. There was faint staining in the stromal andmuscular compartments (Fig. 5 A, D, and G). E2 and E2 � P4treatment resulted in a significant increase in expression of PRin the stroma (Fig. 5 B, C, and E–G), with only a small differencein expression of PR between WT and ER��/� mice (Fig. 5G).The intensity of PR staining in the luminal epithelium wasreduced by treatment with E2 or E2 � P4 in mice of bothgenotypes (Fig. 5 B, C, E, and F).

Stromal Expression of EGFR in ER��/� Mice. EGFR is up-regulatedby estrogens (23) and plays an important role in uterine andvaginal organ growth in response to estrogens (24). Immuno-histochemical staining in uteri from ovariectomized mice re-vealed specific EGFR staining in the cytoplasm of luminal

epithelium in WT mice and much weaker staining in ER��/�

mice (Fig. 6 A and D). E2 treatment enhanced cytoplasmicstaining in epithelium of WT but resulted in patchy staining inthe nuclei of epithelial cells in ER��/� mice (Fig. 6 B and E).Upon treating ER��/� mice with E2 � P4, there was a strikingincrease in nuclear staining of EGFR in the stromal cells (Fig.6F). In WT mice, stromal nuclear staining was rarely detected(Fig. 6C).

DiscussionOur studies show that deficiency of ER� leads to hyperprolif-eration and loss of differentiation in the uterine epithelium. Thisfinding is similar to what has been observed in the ventralprostate (17) and colon (18). Thus, in the ventral prostate, colon,and uterus, ER� seems to be essential for driving cellulardifferentiation.

Previous studies have shown aberrant expression of adhesionmolecules in the mammary gland and colon of ER��/� mice.There is decreased expression of E-cadherin, connexin 32,occludin, and integrin �2 in mammary glands (15) and of�-catenin and plectin in the colon (18). These changes areaccompanied by alterations in cytoarchitecture as revealed byEM (15, 18, 19). In the present study, the shape of the luminalepithelial cells was distorted in ER��/� mice. This distortion wasmost evident as a lack of straight lateral walls. This change inshape may be attributed to the decreased expression of E-cadherin and CK18 exhibited in ER��/� mice. E-cadherin isrequired at adherence junctions (25) and is essential for anchor-

Fig. 4. Colocalization of ER� and ER� in uteri of WT mice. In intact WT mice, ER�, as expected, was abundantly expressed in both epithelium and stroma (M).However, after ovariectomy, ER� could not be detected in the luminal epithelium of either WT or ER��/� mice (G–L). Ovariectomized mice received a singleinjection of vehicle (A, D, G, and J), 100 ng of E2 (B, E, H, and K), or 100 ng of E2 � 1 mg of P4 (C, F, I, and L). Colocalization of ER� (green) and ER� (red) in thestroma was seldom seen in mice treated with vehicle or E2 � P4 (A–C) but was seen in mice treated with E2 alone. In the glandular epithelium (D–F) receptorswere colocalized (yellow and arrows) in vehicle-treated (D) and E2-treated (E, arrows) mice. Upon administration of E2 � P4 (F), expression of ER� wasextinguished, whereas ER� expression remained.

18352 � www.pnas.org�cgi�doi�10.1073�pnas.0608861103 Wada-Hiraike et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

9, 2

020

ing of epithelial cells. Loss of E-cadherin expression can disruptthese adherence junctional complexes and elicit phenotypicchanges, including the acquisition of invasive growth (26). Inaddition to physical attachment, adhesion molecules transmitregulatory information through cytoplasmic adaptors (27) suchas �-catenin, an adherence-junction protein that attaches toE-cadherin and EGFR (28). E-cadherin expression is estrogen-dependently regulated by ER� (29). The present data show thatER� plays a regulatory role in E-cadherin expression. AlthoughE-cadherin expression was reduced in the absence of ER�, afteradministration of E2, E-cadherin expression levels returned tonormal in ER��/� mice. Thus, both ERs are involved in expres-sion of E-cadherin.

Tissue recombinant experiments have clearly shown thatuterine stroma is necessary for the proliferative response of theepithelium to E2 (30). ER�-positive stroma behaves as aninterpreter of the estrogen signal, translating it into a growth-stimulating message that is sent to the epithelium. Loss of growthcontrol is a characteristic of ER��/� mouse uterus (13), bonemarrow (16), ventral prostate (14, 17), and colon (18). Thepresent study shows that in the uterine epithelium of ovariec-tomized ER��/� mice, both in the absence and in the presenceof E2, there were markedly more cells in S phase of the cell cyclethan in WT littermates. Several peptide growth factors whoseexpression is hormonally responsive have been identified in theuterus. These factors include EGF, insulin-like growth factor-I,TGF-�, and TGF-�. The present studies show that dysregulationof the EGF system may be part of the reason for the hyperpro-liferative response of the ER��/� uterus to E2. The role of theEGF family in estrogenic effects in the female genital tract is welldocumented (31). Estrogens have effects on both EGF andEGFR in the uterus (32, 33). The present study revealed thatEGFR is found not only in association with the cell membranebut also in the cytoplasm and in nuclei of stromal cells in ER��/�

mice treated with E2 � P4. As is already known, P4 inducesproliferation of stromal cells and primes the uterus to respondto decidualizing stimuli (34–36). The nuclear localization ofEGFR is of special interest. Recently, several studies have

reported that growth factors and their receptors are not confinedto the membrane compartment of cells (37–39). It has beensuggested that nuclear EGFR is strongly correlated with highlyproliferating activity (39) and that nuclear EGFR might functionas a transcription factor to activate genes required for prolifer-ation. Because the epithelial and stromal compartments com-municate with each other through growth factors, the substantialexpression of EGFR in stromal nuclei of ER��/� mice might berelated to the enhanced proliferation of the epithelium.

The pattern of differentiation of the uterine epithelium in theabsence of ER� signaling was studied by immunohistochemicalstaining. In epithelial cells, CKs are intermediate filaments thatdetermine cell shape. Over 20 distinct proteins constitute the CKfamily, which includes types I (CK9–CK20) and II (CK1–CK8). Allepithelial cells express one or more type I and type II CKs. Changesin expression of CKs are seen during epithelial cell differentiation.CK18 is expressed in luminal and glandular epithelial cells of bothhuman and rabbit endometrium at all developmental stages (40).CK13 exhibits stage-specific expression in the secretory-phaseluminal epithelium of humans and the peri-implantation-stageendometrium of rabbits (40). We report here that CK18 is found inall luminal and glandular epithelial cells, and CK13 is found in theuterine cervix. After E2 treatment, there was a marked reduction inCK18 expression in epithelial cells of ovariectomized ER��/� butnot WT mice. This reduction in CK18 may simply reflect thatalthough most of the luminal epithelial cells are proliferating, veryfew are fully differentiated.

In uterine tissue, both ER� and ER� are expressed. In uterinestroma, ER� and ER� are colocalized in nuclei in E2-treated mice.We interpret this finding to indicate that as ER� mediates EGFsecretion from the stroma, ER� plays a modulatory role onER�-mediated growth factor secretion. In the uterine glandularepithelium of WT mice there are very few cells where the tworeceptors are colocalized. In rodents, uterine secretory products ofthe endometrial glands are required for establishment of uterinereceptivity and conceptus implantation. When ER� is absent,uterine glands are hyperresponsive to E2, with exaggerated uterinesecretion of proteins, including growth factors (13). ER��/� micehave hypoplastic uteri that contain all characteristic cell types, butthe number of uterine glands is reduced (41). We hypothesize thatthe adenogenesis and development of the uterine gland requiresER�, whereas ER� mainly plays a role in differentiation.

Fig. 5. Immunohistochemical detection of PR. PR was abundant in theluminal and glandular epithelium of both WT and ER��/� mice. Stromal PRimmunoreactivity was faint in vehicle-treated WT and ER��/� mice (A, D, andG). It was markedly induced in both WT and ER��/� mice by the administrationof 100 ng of E2 (B, E, and G), or 100 ng of E2 � 1 mg of P4 (C, F, and G).

Fig. 6. Immunohistochemical analysis of EGFR protein expression. Micereceived a single injection of vehicle (A and D), 100 ng of E2 (B and E), or 100ng of E2 � 1 mg of P4 (C and F). In vehicle-treated mice, there was strongmembrane EGFR immunoreactivity in the luminal epithelium of WT mice (A)but no membrane or cytoplasmic staining in ER��/� mice (D). E2 treatmentresulted in enhanced EGFR expression in the cytoplasm of WT mice (B) andincreased nuclear staining in ER��/� mice. In ER��/� mice treated with E2 � P4,there was a marked increase in nuclear localization of EGFR in the stroma (F),whereas in WT mice, this treatment resulted in nuclear localization of EGFR inthe epithelium but not in the stroma. (Insets) Outlined areas at 3 times greatermagnification.

Wada-Hiraike et al. PNAS � November 28, 2006 � vol. 103 � no. 48 � 18353

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 2

9, 2

020

Consistent with the studies in refs. 13 and 20, ER� has a rolein regulating PR expression in the rodent uterus. The 5� regionof the PR gene contains clusters of estrogen-responsive ele-ment half-sites that are essential for trans-activation of the PRgene by liganded ER. However, the mouse PR gene is not atypical estrogen-responsive element-regulated gene, and its 5�region contains many other putative binding sites for otherregulatory transcription factors. Because administration of E2decreases PR in the uterine epithelium of immature WT micebut increases PR in ER��/� mice, it has been suggested thatthe induction of PR is an ER�-mediated event, whereasrepression of epithelial PR is ER�-mediated (13). This studyshows that ER� is not required for an E2-induced decrease ofuterine epithelial PR in adult mice. Interestingly, the E2induction of stromal PR in ER��/� mice suggests that ER�might be responsible for stromal PR induction.

When ER��/� mice are treated with E2 � P4, there is strongstimulation of growth in the stroma. Hom et al. (24) found thatEGFR is required for the stromal response to estrogen. Thehyperproliferation of the epithelium of ER��/� mice afterovariectomy suggests that when E2 and ER� are absent, there isgrowth-factor-induced proliferation of the uterus. Thus, loss ofER� predisposes the uterus to aberrant endometrial prolifera-tion. In view of the increased risk for endometrial cancer thataccompanies estrogen treatment of postmenopausal women, wesuggest that these women should be routinely checked forexpression of ER� in the endometrium. Loss of ER� wouldidentify women who are at higher risk for proliferative diseasesof the endometrium.

Materials and MethodsAnimals. WT and ER��/� mice were bred from heterozygousmice. Genotyping by using PCR was performed on DNA isolatedfrom tails of 2-week-old mice as described in ref. 12. Mice werefed a soy-free standard chow diet (Special Diet Services,Whitham, Essex, U.K.) and tap water ad libitum. All animalswere housed in Huddinge University Hospital Animal Facilityunder controlled lightning (12-h light, 12-h dark cycle), temper-ature (21–22°C), and relative humidity (50%) conditions. Ani-mals were maintained in accordance with the Sodra Djurfor-soksetiska Namnd Guidelines for Care and Use of ExperimentalAnimals.

Hormone Treatments. Twelve WT and ER��/� mice were ovari-ectomized at 8 weeks of age. Seven days after surgery, mice wereinjected once s.c. with 150 �l of olive oil containing no hormone,100 ng of E2, or 100 ng of E2 � 1 mg of P4. Mice were killed bycervical dislocation 16 h after the injection, and the uteri wereisolated.

Chemicals and Antibodies. BrdUrd was purchased from RocheDiagnositcs (Mannheim, Germany). E2 and P4 were purchasedfrom Sigma–Aldrich (St. Louis, MO). Chicken polyclonalantibody anti-ER� 503 IgY was produced in our laboratory(18). Rabbit antibodies were anti-ER� (catalog no. sc-542;Santa Cruz Biotechnology, Santa Cruz, CA), anti-PR (catalogno. A0098; DAKO, Carpinteria, CA), and anti-CK18 (catalogno. 10830-1-AP; Protein Tech Group, Chicago, IL). Goatpolyclonal antibody anti-EGFR was purchased from SantaCruz Biotechnology (catalog no. sc-03-G). Mouse monoclonalantibodies were anti-BrdUrd (catalog no. 3D4; BD BiosciencesPharMingen, San Jose, CA) and anti-CK13 (catalog no.ab16112; Abcam, Cambridge, U.K.). Biotinylated goat anti-rabbit IgG, anti-mouse IgG, horse anti-goat IgG, and avidin-biotin complex kits were from Vector Laboratories (Peterbor-ough, U.K.). Cy-3-conjugated donkey anti-chicken IgG(catalog no. 703-165-155) and FITC-conjugated donkey anti-

rabbit (catalog no. 711-095-152) were purchased from JacksonImmunoResearch (West Grove, PA).

Treatment with BrdUrd. For measurement of proliferation, micewere treated i.p. with BrdUrd at 30 mg�kg 2 h before being killed.

Immunohistochemistry. For immunohistochemical studies, uteriwere removed and fixed f lat on wet filter papers overnight in4% paraformaldehyde. They were routinely embedded inparaffin wax, and 4-�m sections were mounted on organosi-lane-coated slides. Four-micrometer paraffin sections weredewaxed in xylene and rehydrated through graded ethanol towater. Antigens were retrieved by boiling in a 10 mM citratebuffer (pH 6.0) for 20 min except for the immunostaining ofEGFR. The cooled sections were incubated in methanolcontaining 0.3% H2O2 for 30 min to quench endogenousperoxidase. To block the nonspecific binding, sections wereincubated in PBS containing 3% BSA and 0.1% Nonidet P-40for 10 min at room temperature. Sections were then incubatedwith the antisera anti-BrdUrd (1:800), anti-CK13 (1:200),anti-CK18 (1:200), anti-PR (1:100), anti-EGFR (1:50), andanti-E-cadherin (1:200) in PBS containing 3% BSA and 0.1%Nonidet P-40 overnight at 4°C. Negative controls were incu-bated with PBS containing 3% BSA and 0.1% Nonidet P-40without a primary antibody. The avidin–biotin complexmethod was used to visualize the signal, according to themanufacturer’s manual (Vector Laboratories). The sectionswere incubated in appropriate biotinylated IgG solution(1:200) for 1 h at room temperature, followed by washing withPBS and incubation in avidin-biotin-horseradish peroxidasefor 1 h. After washing in PBS, sections were developed with3,3�-diaminobenzidine tetrahydrochloride substrate (DAKO)for 1–3 min, producing a brown stain. The sections were lightlycounterstained with Mayer’s hematoxylin, dehydrated througha graded ethanol series and xylene, and mounted.

Immunofluorescence Labeling of ER� and ER�. ER� and ER� weredetected in tissue sections by standard immunofluorescence asdescribed in ref. 42. For immunofluorescent detection, 4-�mparaffin sections were dewaxed in xylene and rehydrated throughgraded ethanol to water. Antigens were retrieved by boiling in a10 mM citrate buffer (pH 6.0) in a microwave oven for 20 min.Sections were incubated with PBS containing 0.5% Triton X-100for 30 min, then with PBS containing 3% BSA for 15 min. Fordouble staining of ER��ER�, sections were incubated sequen-tially with anti-ER� (1:200) and anti-ER� (1:100) in PBScontaining 3% BSA and 0.1% Nonidet P-40 overnight at 4°C.Negative controls were incubated with PBS containing 3% BSAand 0.1% Nonidet P-40 without a primary antibody. The sec-ondary antibodies were Cy-3-conjugated anti-chicken IgG(1:200), and FITC-conjugated anti-rabbit IgG (1:200). Afterincubation for 1 h, the sections were washed and dipped in DAPIsolution (0.1 �g�ml in PBS) for 20 sec at room temperature todelineate nuclei and mounted in Vectashield antifading medium(Vector Laboratories). The sections were examined under aZeiss (Gottingen, Germany) Axioplan 2 fluorescence micro-scope by using suitable filters for selectively detecting thefluorescence of FITC (green) and Cy3 (red). Colocalization wasindicated by yellow or orange in cells in which both FITC- andCy3-conjugated secondary antibodies were sequestered.

We thank Jose Inzunza and AnnMarie Witte for managing ER��/�

mice. This study was supported by the Ministry of Education, Science,and Culture, Japan, JMS Schering, the Japan Menopausal Society, Japan(to O.W.-H.), the European Union CASCADE Network of Excellence,the Swedish Cancer Fund, Konung Gustav V och Drottning VictoriasStiftelse, and Karo Bio AB.

18354 � www.pnas.org�cgi�doi�10.1073�pnas.0608861103 Wada-Hiraike et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

9, 2

020

1. Martin L, Finn CA, Trinder G (1973) J Endocrinol 56:133–144.2. Quarmby VE, Korach KS (1984) Endocrinology 114:694–702.3. Kachkache M, Acker GM, Chaouat G, Noun A, Garabedian M (1991) Biol

Reprod 45:860–868.4. De M, Wood GW (1990) J Endocrinol 126:417–424.5. Martin L, Das RM, Finn CA (1973) J Endocrinol 57:549–554.6. Persson I (1996) Maturitas 23 (Suppl): S37–S45.7. Cohen CJ, Bruckner HW, Deppe G, Blessing JA, Homesley H, Lee JH,

Watring W (1984) Obstet Gynecol 63:719–726.8. Henderson BE, Ross RK, Pike MC (1993) Science 259:633–638.9. Paech K, Webb P, Kuiper GGJM, Nilsson S, Gustafsson J-Å, Kushner PJ,

Scanlan TS (1997) Science 277:1508–1510.10. Martin L, Pollard JW, Fagg B (1976) J Endocrinol 69:103–115.11. Couse JF, Lindzey J, Grandien K, Gustafsson J-Å, Korach KS (1997) Endo-

crinology 138:4613–4621.12. Krege JH, Hodgin JB, Couse JF, Enmark E, Warner M, Mahler JF, Sar M,

Korach KS, Gustafsson J-Å, Smithies O (1998) Proc Natl Acad Sci USA95:15677–15682.

13. Weihua Z, Saji S, Makinen S, Cheng G, Jensen EV, Warner M, Gustafsson J-Å(2000) Proc Natl Acad Sci USA 97:5936–5941.

14. Weihua Z, Makela S, Andersson LC, Salmi S, Saji S, Webster JI, Jensen EV,Nilsson S, Warner M, Gustafsson J-Å (2001) Proc Natl Acad Sci USA98:6330–6335.

15. Forster C, Makela S, Warri A, Kietz S, Becker D, Hultenby K, Warner M,Gustafsson J-Å (2002) Proc Natl Acad Sci USA 99:15578–15583.

16. Shim GJ, Wang L, Andersson S, Nagy N, Kis LL, Zhang Q, Makela S, WarnerM, Gustafsson J-Å (2003) Proc Natl Acad Sci USA 100:6694–6699.

17. Imamov O, Morani A, Shim GJ, Omoto Y, Thulin-Andersson C, Warner M,Gustafsson J-Å (2004) Proc Natl Acad Sci USA 101:9375–9380.

18. Wada-Hiraike O, Imamov O, Hiraike H, Hultenby K, Schwend T, Omoto Y,Warner M, Gustafsson J-Å (2006) Proc Natl Acad Sci USA 103:2959–2964.

19. Morani A, Barros RPA, Imamov O, Hultenby K, Arner A, Warner M,Gustafsson J-Å (2006) Proc Natl Acad Sci USA 103:7165–7169.

20. Kurita T, Lee K, Saunders PT, Cooke PS, Taylor JA, Lubahn DB, Zhao C,Makela S, Gustafsson J-Å, Dahiya R, Cunha GR (2001) Biol Reprod 64:272–283.

21. Tessier C, Deb S, Prigent-Tessier A, Ferguson-Gottschall S, Gibori GB, ShiuRP, Gibori G (2000) Endocrinology 141:3842–3851.

22. Wu JJ, Geimonen E, Andersen J (2000) Eur J Endocrinol 142:92–99.23. Das SK, Tsukamura H, Paria BC, Andrews GK, Dey SK (1994) Endocrinology

134:971–981.24. Hom YK, Young P, Wiesen JF, Miettinen PJ, Derynck R, Werb Z, Cunha GR

(1998) Endocrinology 139:913–921.25. Adams CL, Nelson WJ (1998) Curr Opin Cell Biol 10:572–577.26. Behrens J, Mareel MM, Van Roy FM, Birchmeier W (1989) J Cell Biol

108:2435–2447.27. Thiery JP (2002) Nat Rev Cancer 2:442–454.28. Hoschuetzky H, Aberle H, Kemler R (1994) J Cell Biol 127:1375–1380.29. Fujita N, Jaye DL, Kajita M, Geigerman C, Moreno CS, Wade PA (2003) Cell

113:207–219.30. Cooke PS, Buchanan DL, Young P, Setiawan T, Brody J, Korach KS, Taylor

J, Lubahn DB, Cunha GR (1997) Proc Natl Acad Sci USA 94:6535–6540.31. Mukku VR, Stancel GM (1985) J Biol Chem 260:9820–9824.32. DiAugustine RP, Petrusz P, Bell GI, Brown CF, Korach KS, McLachlan JA,

Teng CT (1988) Endocrinology 122:2355–2363.33. Sakai Y, Nelson KG, Snedeker S, Bossert NL, Walker MP, McLachlan J,

DiAugustine RP (1994) Cell Growth Differ 5:527–535.34. Finn CA, Martin L (1970) J Endocrinol 47:431–438.35. Lydon JP, DeMayo FJ, Funk CR, Mani SK, Hughes AR, Montgomery CA, Jr,

Shyamala G, Conneely OM, O’Malley BW (1995) Genes Dev 9:2266–2278.36. Lydon JP, DeMayo FJ, Conneely OM, O’Malley BW (1996) J Steroid Biochem

Mol Biol 56:67–77.37. Podlecki DA, Smith RM, Kao M, Tsai P, Huecksteadt T, Brandenburg D,

Lasher RS, Jarett L, Olefsky JM (1987) J Biol Chem 262:3362–3368.38. Antoine M, Reimers K, Dickson C, Kiefer P (1997) J Biol Chem 272:29475–

29481.39. Marti U, Ruchti C, Kampf J, Thomas GA, Williams ED, Peter HJ, Gerber H,

Burgi U (2001) Thyroid 11:137–145.40. Olson GE, Winfrey VP, Blaeuer GL, Palisano JR, NagDas SK (2002) Biol

Reprod 66:1006–1015.41. Curtis SW, Clark J, Myers P, Korach KS (1999) Proc Natl Acad Sci USA

96:3646–3651.42. Barros RPA, Machado UF, Warner M, Gustafsson J-Å (2006) Proc Natl Acad

Sci USA 103:1605–1608.

Wada-Hiraike et al. PNAS � November 28, 2006 � vol. 103 � no. 48 � 18355

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Aug

ust 2

9, 2

020