CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector Developmental Cell Short Article

Nuclear Receptor -6 Attenuates Uterine Estrogen Sensitivity to Permit Embryo Implantation

Jun Kawagoe,1 Qingtian Li,1 Paola Mussi,1 Lan Liao,1 John P. Lydon,1 Francesco J. DeMayo,1 and Jianming Xu1,2,* 1Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX 77030, USA 2Institute for Cancer Medicine, Luzhou Medical College, Sichuan 646000, China *Correspondence: [email protected] http://dx.doi.org/10.1016/j.devcel.2012.09.002

SUMMARY mice have shown that ERa, but not ERb, is essential for preparing endometrium for blastocyst attachment, while both

Uterine receptivity to embryo implantation is coordi- receptors may be dispensable for P4-induced uterine deciduali- nately regulated by 17b-estradiol (E2) and proges- zation (Wang and Dey, 2006). Furthermore, tightly regulated E2 terone (P4). Although increased E2 sensitivity causes concentrations determine the duration of receptive time window infertility, the mechanisms underlying the modulation in the uterus. Lower E2 levels sustain but higher E2 levels shut down this time window (Ma et al., 2003). Although the exact of E2 sensitivity are unknown. We show that nuclear receptor coactivator-6 (NCOA6), a reported coactiva- role of E2 at the preimplantation stage is unclear in human uterus (Wang and Dey, 2006), high serum E levels or E /P ratios tor for estrogen receptor a (ERa), actually attenuates 2 2 4 induced by artificial ovulation induction may generate an adverse E2 sensitivity to determine uterine receptivity to endometrial environment and reduce implantation rate (Val- embryo implantation under normal physiological con- buena et al., 1999). Failure of ERa downregulation at the Ncoa6 ditions. Specifically, conditional knockout of in implantation stage may also increase E2 sensitivity and reduce uterine epithelial and stromal cells does not decrease, the pregnancy rate in women with polycystic ovary syndrome but rather markedly increases E2 sensitivity, which (PCOS) (Gregory et al., 2002). disrupts embryo implantation and inhibits P4-regu- NCOA6 (AIB3, ASC-2, RAP250, NRC, PRIP, or TRBP) is a re- lated and decidual response. NCOA6 enhances ported coactivator for multiple nuclear receptors and has many ERa ubiquitination and accelerates its degradation, biological functions (Mahajan and Samuels, 2005). Ncoa6 KO while loss of NCOA6 causes ERa accumulation in in mice causes embryonic lethality due to developmental defects stromal cells during the preimplantation period. During in placenta, heart, and liver (Kuang et al., 2002; Mahajan and Samuels, 2005). In cultured cells, NCOA6 interacts with ERa the same period, NCOA6 deficiency also caused a and enhances its transcriptional function (Caira et al., 2000; failure in downregulation of steroid receptor coactiva- Ko et al., 2000; Lee et al., 1999; Mahajan and Samuels, 2000). a tor-3 (SRC-3), a potent ER coactivator. Therefore, NCOA6 contains two LXXLL motifs, and ERa mainly interacts NCOA6 controls E2 sensitivity and uterine receptivity with its N-terminal LXXLL motif (Mahajan and Samuels, 2005). by regulating multiple E2-signaling components. Ncoa6 is expressed in many tissues, including uterine ECs and SCs (Zhang et al., 2003). The female heterozygous Ncoa6 KO INTRODUCTION mice exhibit a slightly reduced reproductive function (Mahajan and Samuels, 2005). However, the physiological function and Embryo implantation depends on molecular interactions molecular mechanisms for NCOA6 during early pregnancy are between the -primed uterus and the mature blastocyst unknown. In this study, we defined the function of Ncoa6 in (Wang and Dey, 2006). Implantation failure can be caused by early pregnancy by conditional KO of Ncoa6 in the progesterone defects in embryo or endometrium (Cakmak and Taylor, 2011; receptor (PR)-positive cell lineages in mice, including both ECs Valbuena et al., 1999). The endometrium is receptive to blasto- and SCs in the uterus. We demonstrate that the essential cyst implantation only in a restricted time ‘‘window’’ (Cakmak role of NCOA6 as a ‘‘coactivator’’ in uterine receptivity is actually

and Taylor, 2011; Wang and Dey, 2006), which is determined to attenuate E2 sensitivity through accelerating ERa ubiquitina- by the spatiotemporal regulation of the proliferation and differen- tion and degradation and downregulating SRC-3 in the endome- tiation status of uterine epithelial cells (ECs) and stromal cells trial epithelium at the preimplantation stage.

(SCs) in response to progesterone (P4) and 17b-estradiol (E2) stimuli. Upon P4 priming during early mouse pregnancy, a small RESULTS and temporal E2 increase at 3.5 days postcoitum (dpc) is essen- tial for further induction of uterine SC proliferation and luminal EC Conditional Deletion of the Ncoa6 in the Mouse differentiation to prepare the uterus for blastocyst attachment at Uterus, Corpus Luteum, and Fallopian Tube Causes 4.0 dpc (Matsumoto et al., 2002; Wang and Dey, 2006). a Failure of Embryo Attachment

E2 action is mediated by estrogen receptors a and b (ERa and Since Ncoa6 null mice display embryonic lethality, we generated ERb, respectively). Studies using ERa and ERb knockout (KO) the floxed Ncoa6 (Ncoa6f/f) mice, which were as normal as

858 Developmental Cell 23, 858–865, October 16, 2012 ª2012 Elsevier Inc. Developmental Cell

Ncoa6, E2 Sensitivity, and Embryo Implantation

d/d Figure 1. Ncoa6 Mice Are Infertile and Exhibit Altered Expression Patterns of P4 and E2-Inducible Genes in the Uterus (A) Ncoa6 IHC (in brown) in Ncoa6f/f and Ncoa6d/d uterine sections. Scale bars represent 100 mm. LE, luminal epithelium; GE, glandular epithelium; St, stroma. (B) Embryo ISs in Ncoa6f/f uterus (arrows) but not in Ncoa6d/d uterus at 4.5 dpc. f/f d/d (C) Relative expression levels of P4-regulated genes in Ncoa6 (n = 5) and Ncoa6 (n = 5) uteri at 3.5 dpc. *p < 0.05; **p < 0.01; ***p < 0.0001. f/f d/d (D) Relative expression levels of E2-regulated genes in Ncoa6 (n = 5) and Ncoa6 (n = 5) uteri at 3.5 dpc. f/f d/d (E) Relative expression levels of Muc1, Hbegf, and Vegfa in the uteri of Ncoa6 (n = 5) and Ncoa6 (n = 5) mice treated with sesame oil vehicle (V) or low-dose E2 (6.7 ng per mouse). The data in (C)–(E) are presented as means ± SEM. See also Figure S1.

wild-type (WT) mice (Figures S1A–S1E available online). We next 69.0%, n = 29). At 3.5 dpc, the serum E2 (185 ± 17 and 173 ± Cre/+ crossed these mice to PR mice (Soyal et al., 2005) and 19 ng/ml, respectively) and P4 (39 ± 4 and 35 ± 1 ng/ml, respec- generated Ncoa6f/f and PRCre/+;Ncoa6f/f (hereinafter designated tively) levels in Ncoa6d/d (n = 7) and Ncoa6f/f (n = 6) mice are also as Ncoa6d/d) mice. Total Ncoa6 mRNA and are abundant similar, indicating a normal luteal function in Ncoa6d/d mice. At in Ncoa6f/f uteri but are mostly diminished in Ncoa6d/d uteri 4.5 dpc, we observed normal number of embryo implantation (Figures S1F and S1G). In Ncoa6f/f mice, Ncoa6 protein is de- sites (ISs) (6.8 ± 2.9 per uterus) in Ncoa6f/f uteri (n = 10), but tected in the uterine luminal and glandular ECs, endometrial Ncoa6d/d uteri (n = 10) had no ISs, although 3.3 ± 1.4 unimplanted SCs, and myometrial cells. Low-level Ncoa6 protein is also normal blastocysts per uterus could be recovered from these present in the corpus luteum and fallopian tube. In Ncoa6d/d uteri (Figure 1B). Furthermore, transfer of 101 WT blastocysts mice, Ncoa6 is efficiently knocked out by the expression of into the pseudopregnant Ncoa6f/f uteri (n = 10) generated 41 PR-Cre transgene in the uterine ECs and SCs, corpus luteum, embryo ISs, but transfer of 87 WT blastocysts into the pseudo- and fallopian tubule epithelium, while Ncoa6 protein is still ex- pregnant Ncoa6d/d uteri (n = 8) only produced 1 embryo IS. These pressed in the myometrium where little PR-Cre (Soyal et al., results demonstrate the failure of embryo attachment in 2005) is expressed (Figures 1A and S1H–S1J; data not shown). Ncoa6d/d uteri. Therefore, Ncoa6 plays an essential role in the During breeding with WT male mice for 6 months, 7 Ncoa6f/f uterine receptivity for embryo attachment and implantation. females produced 21 litters with 6.4 ± 2.6 pups per litter, but 9 Ncoa6d/d females did not produce any pups. We next assessed Ncoa6 Deficiency Results in Abnormal Expression the effects of Ncoa6 loss on early pregnant events to pursue the of the Implantation-Related Genes and Increased d/d f/f cause of infertility. Ncoa6 (versus Ncoa6 ) mice showed no E2 Sensitivity in the Uterus significant differences in ovulation (8.4 ± 0.65 oocytes per To pursue the cause of implantation failure, we compared the mouse, n = 8 versus 7.5 ± 0.4 oocytes per mouse, n = 7), oocyte expression levels of implantation-related genes that are regu- fertilization rate (46.4%, n = 69 versus 57.4%, n = 54) at 1.5 dpc, lated by P4 and E2 (Large and Demayo, 2011; Wang and Dey, and blastocyst formation rate at 3.5 dpc (53.6%, n = 28 versus 2006) between Ncoa6f/f and Ncoa6d/d uteri at 3.5 dpc. In

Developmental Cell 23, 858–865, October 16, 2012 ª2012 Elsevier Inc. 859 Developmental Cell

Ncoa6, E2 Sensitivity, and Embryo Implantation

d/d Ncoa6 uteri, P4-regulated genes (1) in the epithelium, such as and pHH3 expression, indicating abnormally sustained EC amphiregulin (Areg)(Ma et al., 2003), lysophosphatidic acid proliferation during the preimplantation period (Figures 2B and receptor-3 (Lpar3)(Hama et al., 2006), and Indian hedgehog Figure S2A). Since the EC proliferation is promoted by E2, the (Ihh)(Lee et al., 2006); (2) in both epithelium and stroma, such proliferative epithelium also indicates the excess E2 signaling as homeobox A10 (Hoxa10)(Cakmak and Taylor, 2011); and (3) caused by Ncoa6 KO. in the stroma, such as heart and neural crest derivatives ex- We further examined ERa in the uterus. In comparison with pressed 2 (Hand2)(Li et al., 2011), are significantly decreased. Ncoa6f/f uteri, Ncoa6d/d uteri exhibit much stronger stromal The Ihh-regulated downstream stromal genes such as bone ERa immunostaining at 3.5 dpc (Figure 2C). Western blot anal- morphogenetic protein-2 (Bmp2) and chicken ovalbumin ysis further confirmed an overall 2-fold ERa increase in upstream promoter factor 2 (COUP-TFII)(Kurihara Ncoa6d/d versus Ncoa6f/f uteri (Figure 2D). Because stromal d/d et al., 2007) are also significantly downregulated in Ncoa6 ERa is responsible for E2-induced EC proliferation in normal uteri (Figure 1C). On the other hand, E2-regulated genes includ- uterus (Cooke et al., 1997; Winuthayanon et al., 2010), the ing mucin-1 (Muc1), lactotransferrin (Ltf), heparin-binding EGF- increased ERa may be involved in epithelial overgrowth of like factor (Hbegf), vascular endothelial growth factor (Vegfa), Ncoa6d/d uteri.

fibroblast growth factors 1/2 (Fgf1/2) and insulin-like growth E2-induced EC proliferation is originated from stroma and factor 1 (Igf1)inNcoa6d/d uteri are expressed at higher levels mediated by the fibroblast growth factors (FGFs)-FGFR-ERK than that in Ncoa6f/f uteri, while the expression of leukemia paracrine signaling pathway (Li et al., 2011). Thus, we investi- inhibitory factor (Lif), a glandular factor that regulates uterine gated the downstream signaling components of ERK in the receptivity, is slightly decreased in Ncoa6d/d uteri (Figure 1D; epithelium. At 3.5 dpc, numerous phospho-ERK1/2-positive and data not shown). These results suggest that Ncoa6 cells were detected in the epithelium of Ncoa6d/d uteri but f/f enhances P4-regulated genes but attenuates E2-regulated undetectable in the epithelium of Ncoa6 uteri (Figure 2E). genes in the uterus during the preimplantation period under Accordingly, the phosphorylated ERa on Ser118, a target site of physiological conditions. phospho-ERK1/2, was not detected in ECs of Ncoa6f/f uteri but d/d To assess the specific role of Ncoa6 in P4 and E2 signaling was evidently detected in ECs of Ncoa6 uteri at 3.5 dpc (Fig- pathways, we treated ovariectomized Ncoa6f/f and Ncoa6d/d ure 2F). Since the Ser118 phosphorylation promotes transcrip- mice with P4 or low-dose E2 (6.7 ng per mouse). P4 treatment tional activity of ERa, we analyzed the expression of an ERa equally induced the expression of P4-regulated genes including target, Muc1, in the uterine epithelium. Muc1 expression was Areg, Ihh, Hoxa10, Bmp2, and COUP-TFII in Ncoa6f/f and markedly increased in Ncoa6d/d uteri versus Ncoa6f/f uteri (Fig- d/d Ncoa6 uteri (Figure S1K). Furthermore, at 2.5 dpc when P4 is ure 2G). Since Muc1 is a barrier on the surface of luminal ECs dominant and E2 is not secreted, the expression of Ihh mRNA that prevents embryo attachment, the increased Muc1 may in Ncoa6f/f and Ncoa6d/d uteri is also similar (Figure S1L). Low- cause blastocyst attachment failure in Ncoa6d/d uteri. dose E2 treatment was unable to induce significant expression of E2-regulated genes such as Muc1, Hbegf, Vegfa, Myc, and Ncoa6-Deleted Uteri Fail to Downregulate SRC-3 Igf1 in Ncoa6f/f uteri and also incapable to stimulate the growth in the Luminal Epithelium of these uteri. On the contrary, the same treatment resulted in Since the SRC family coactivators enhance ERa transcriptional d/d remarkable induction of these E2-regulated genes in Ncoa6 activity (Xu et al., 2009), we compared their expressions in uteri and uterine weight gain (Figures 1E, S1M, and S1N). More- Ncoa6f/f and Ncoa6d/d uteri. In nonpregnant mice, SRC-3 protein over, at 3.5 dpc, the uteri of Ncoa6d/d mice were grossly heavier is equally expressed in ECs of both Ncoa6f/f and Ncoa6d/d uteri than the uteri of Ncoa6f/f mice in response to hormonal stimula- throughout the estrous cycle (Figure S2B). At 3.5 dpc, SRC-3 tion (Figure S1O). These results demonstrate (1) that Ncoa6 is significantly reduced in the epithelium of Ncoa6f/f uteri, as re- does not directly modulate P4 sensitivity but directly suppresses ported previously in normal uteri (Jeong et al., 2007). However, d/d E2 sensitivity and (2) that the E2 supersensitivity that resulted high-level SRC-3 is still retained in ECs of Ncoa6 uteri (Fig- d/d from Ncoa6 KO may inhibit uterine receptivity, as high-dose E2 ure 2H). SRC-3 mRNA is also significantly increased in Ncoa6 treatment makes the uterus refractory to embryo attachment uteri (Figure S2C). The expression levels of SRC-1 and SRC-2 (Ma et al., 2003). are comparable between Ncoa6f/f and Ncoa6d/d uteri at 3.5 dpc (immunohistochemistry [IHC] data not shown). Since Ncoa6 Deficiency Increases Uterine Epithelial SRC-3 is a strong coactivator for nuclear receptors including

Proliferation, Stromal ERa Protein, and Epithelial E2 ERa and other transcription factors such as , AP-1, and Response during the Preimplantation Period PEA3 (Xu et al., 2009), these results suggest that SRC-3 overex-

In normal pregnant uteri, SC growth and cessation of EC prolif- pression may also contribute to the increased E2 sensitivity and eration assemble a closed linear uterine lumen for blastocyst cell proliferation in the epithelium of Ncoa6d/d uteri. attachment. In agreement with this, Ncoa6f/f uteri have closed lumens at 3.5 dpc. However, Ncoa6d/d uteri display intricately Failure of Decidual Response in Ncoa6d/d Uteri Due extended luminal and glandular epithelia that surround the to Reduced Bmp2 Expression curving lumens with many branches on the uterine cross- Excess E2 stimulation suppresses SC differentiation and decidu- sections (Figure 2A), suggesting an epithelial overgrowth. Cell alization in human and rat uteri (Basir et al., 2001; Kennedy, proliferation markers Ki-67 and phospho-histone H3 (pHH3) 1986). To determine the functional impact of E2 supersensitivity f/f d/d are undetectable in the epithelium of Ncoa6 uteri at 3.5 dpc. and its accompanied downregulation of P4 signaling in Ncoa6 However, Ncoa6d/d uterine epithelium exhibits robust Ki-67 uteri, we performed decidual response assay as described in the

860 Developmental Cell 23, 858–865, October 16, 2012 ª2012 Elsevier Inc. Developmental Cell

Ncoa6, E2 Sensitivity, and Embryo Implantation

Figure 2. Ncoa6 Deficiency Causes Epithelial Overproliferation and Elevation of ERa, Phosphorylated ERa, and Muc1 and SRC-3 in the Uterus at 3.5 dpc (A) H&E-stained Ncoa6f/f and Ncoa6d/d uterine sections. (B and C) Ki67 IHC for (B) Ki67 and (C) ERa in Ncoa6f/f and Ncoa6d/d uterine sections. The boxed areas in the left panels of (C) are enlarged in the right panels, which show a stronger ERa immunoreactivity in the stroma of Ncoa6d/d uterus versus Ncoa6f/f uterus. (D) Representative western blot results for ERa in tissue lysates prepared from Ncoa6f/f and Ncoa6d/d uteri. Band intensities were determined by densitometry and normalized to b-actin. The data are presented as means ± SEM. *p = 0.02. (E–H) IHC for (E) pERK1/2, (F) S118-phosphorylated ERa (pERa), (G) Muc1, and (H) SRC-3 in Ncoa6f/f and Ncoa6d/d uterine sections. Arrows indicate cell nuclei with pERa immunoreactivity (F), Muc1 immunoreactivity in the apical membrane of luminal epithelium (G), and SRC-3 immunoreactivity in the nuclei of luminal ECs (H). Scale bars indicate 100 mm. See also Figure S2.

Supplemental Experimental Procedures. The decidual response occurred. At 4.5 dpc, we observed that all Ncoa6f/f uteri (n = 6) in Ncoa6d/d uteri was diminished, and, accordingly, the decidual had ISs (6.2 ± 1.0 per uterus), and so did five out of six response markers, including alkaline phosphatase activity and Ncoa6d/d uteri (3.2 ± 2.1 per uterus) (Figures 3A and 3B). ICI mRNA expression of Bmp2 and follistatin (Fst), were not induced treatment also strikingly reduced Muc1 expression in the luminal (Figures S3A–S3D). Interestingly, this failure of decidual ECs of Ncoa6d/d uteri at 4.5 dpc, making its expression similar to response could be partially restored by administration of exoge- that in the untreated Ncoa6f/f uteri (Figure 3C). These results nous BMP2 protein (Figures S3E and S2F). Therefore, Ncoa6 demonstrate that the Ncoa6 deficiency-caused E2 supersensi- KO-caused downregulation of Bmp2 expression is partially tivity is responsible for the implantation failure in Ncoa6d/d uteri. responsible for the failure of decidualization in Ncoa6d/d uteri. Ncoa6 Accelerates ERa Ubiquitination and Degradation Inhibition of ER Function Rescues Embryo Implantation In pursuing how ERa is increased in Ncoa6d/d uteri, we measured in Ncoa6d/d Uteri its mRNA levels. ERa mRNA is comparably expressed in Ncoa6f/f d/d To test whether suppression of E2 supersensitivity could rescue and Ncoa6 uteri at 3.5 dpc (Figure S4A). We then assessed implantation failure in Ncoa6d/d uteri, we treated Ncoa6f/f and how Ncoa6 deficiency affected ERa protein stability in uterine Ncoa6d/d mice with a low dose of ER antagonist ICI-182780 SCs. SCs isolated from Ncoa6d/d uteri at the preimplantation

(ICI) at 3.0 dpc, the time just before the temporal E2 secretion stage displayed higher levels of ERa protein than those from

Developmental Cell 23, 858–865, October 16, 2012 ª2012 Elsevier Inc. 861 Developmental Cell

Ncoa6, E2 Sensitivity, and Embryo Implantation

NCOA6 proteins failed to support it as an E3 ligase (data not shown). Taken together, these results suggest that Ncoa6 may serve as a molecular modulator to promote ERa ubiquitination

and degradation. This function of Ncoa6 may suppress E2 sensi- tivity through downregulation of ERa protein in the uterine stroma to allow embryo implantation.

DISCUSSION

We have demonstrated that Ncoa6 KO in the uterine SCs and ECs results in embryo implantation failure, indicating that Ncoa6 is absolutely required for appropriate development of uterine endometrial receptivity. Furthermore, Ncoa6 deficiency

robustly increases uterine E2 sensitivity as evidenced by the increased expression of ERa target genes under physiological conditions, the abnormally sustained EC proliferation during the preimplantation period and the supersensitive response to

low dose E2-induced and uterine growth. More- over, suppression of ERa function rescues embryo implantation, Figure 3. Inhibition of E2 Supersensitivity with ICI-182780 Treatment d/d Rescues Embryo Implantation in Ncoa6 Uteri proving that E2 supersensitivity is responsible for the failure of (A) ISs (arrows) in Ncoa6f/f and Ncoa6d/d uteri at 4.5 dpc after ICI-182780 embryo implantation caused by Ncoa6 deficiency. These treatment. discoveries clearly indicate that Ncoa6 actually attenuates (B) A H&E-stained embryo (arrow) attached to the epithelium of Ncoa6d/d E2/ERa physiological function to permit embryo implantation in uterus at 4.5 dpc in a mouse treated with ICI-182780. the uterus, although it was previously reported as an ERa coac- (C) IHC for Muc1 in uterine sections prepared from vehicle (sesame oil)- or ICI-182780-treated Ncoa6d/d mice and untreated Ncoa6f/f mice at 4.5 dpc. tivator based on experiments performed using cultured cells Arrows indicate Muc1 immunoreactivity located at the surface of luminal ECs. (Mahajan and Samuels, 2005). These findings also qualify Scale bars indicate 100 mm. See also Figure S3. Ncoa6 as an essential negative modulator of uterine E2 sensi- tivity during the preimplantation period. f/f Ncoa6 uteri (Figures 4A, 4B, S4B, and S4C). When protein Since E2-induced uterine EC proliferation is indirectly stimu- synthesis was blocked by cyclohexamide (CHX), ERa degraded lated by the ERa-mediated FGFs and IGF1 expression in the slightly faster in Ncoa6f/f SCs versus Ncoa6d/d SCs in the stroma and the subsequent paracrine activation of the receptor absence of E2 (Figures 4A and S4B). More interestingly, E2 treat- tyrosine kinases and downstream protein kinase signaling path- ment induced a rapid ERa degradation in Ncoa6f/f SCs but ways in ECs (Hewitt et al., 2010; Li et al., 2011), the Ncoa6 defi- d/d only slightly accelerated ERa degradation in Ncoa6 SCs ciency-induced uterine E2 supersensitivity should be initiated (Figures 4B and S4C). Moreover, NCOA6 knockdown and from the increased ERa protein in the stroma. Our data, when overexpression in human MCF-7 cells increased and decreased combined with the existing knowledge, suggest that the increase ERa protein, respectively (Figure S4D; data not shown). These in stromal ERa in Ncoa6d/d uteri promotes the expression of results suggest that Ncoa6 levels are inversely correlated with growth factors such as IGF1 and FGFs, which in turn activate ERa levels. their receptors and downstream protein kinase pathways such Next, we examined whether Ncoa6 is involved in ERa ubiqui- as ERK. ERK-mediated phosphorylation of S118 enhances ERa tination-dependent degradation. In Ncoa6f/f uterine SCs, Ncoa6 activity in ECs. The enhanced ERK pathway may also upregulate is associated with ERa in both presence and absence of E2 as SRC-3 through activating the E2F1/SP1–SRC-3 positive feed- evidenced by coimmunoprecipitation (coIP) (Figure 4C). ERa back regulatory loop (Mussi et al., 2006). Since SRC-3 is a potent ubiquitination is detected at low levels in untreated cells, but it ERa coactivator (Xu et al., 2009), the persistent expression of f/f is significantly induced in E2-treated Ncoa6 cells. Ubiquitinated high level SRC-3 may work synergistically with the activated ERa further increases after these cells are treated with protea- ERa in ECs of Ncoa6d/d uteri to support the sustained Muc1 some inhibitor MG132, and reaches the peak level upon expression at 3.5 dpc, preventing embryo attachment. In addi- combined E2 and MG132 treatment (Figure 4C). However, ubiq- tion, since ERK activation also leads to activation of multiple uitinated ERa is barely detectable in un-treated or E2-treated transcription factors such as E2F1, AP-1 and PEA3 that use Ncoa6d/d SCs. Ubiquitinated ERa is also significantly lower in SRC-3 as a coactivator to promote cell proliferation (Xu et al., d/d d/d MG132-treated and MG132+E2-treated Ncoa6 SCs versus 2009), the upregulated SRC-3 in Ncoa6 uteri may be partially the same reagents-treated Ncoa6f/f SCs (Figure 4C). Further- responsible for the sustained EC proliferation at 3.5 dpc, causing more, both human NCOA6a and NCOA6b isoforms (Li and Xu, a blockage of EC differentiation.

2011) ectopically expressed in HeLa cells form complexes with In early pregnancy, counterbalanced E2 and P4 signaling path- ERa and robustly enhance ERa ubiquitination (Figure 4D). ways coordinately regulate uterine receptivity. Overstimulation

Finally, amino acid sequence analysis suggests that Ncoa6 with E2 dosages that inhibit uterine receptivity has been shown does not contain any conserved homologous domain to any to inhibit P4/PR signaling (Tibbetts et al., 1998) and downregulate ubiquitin ligase, and in vitro ERa ubiquitination and protea- PR target genes such as Ihh (Matsumoto et al., 2002), Lpar3 some-mediated degradation assays using purified human (Hama et al., 2006) and Areg (Ma et al., 2003) in the uterine

862 Developmental Cell 23, 858–865, October 16, 2012 ª2012 Elsevier Inc. Developmental Cell

Ncoa6, E2 Sensitivity, and Embryo Implantation

Figure 4. Ncoa6 Enhances ERa Ubiquitina- tion and Degradation (A) The isolated Ncoa6f/f and Ncoa6d/d uterine SCs were cultured in the absence of estrogen and treated with CHX. Cells were collected at different time points for western blot.

(B) The aforementioned cells were treated with E2 and CHX as indicated and assayed by western blot. (C) Ncoa6f/f and Ncoa6d/d uterine SCs were

treated with MG132 and/or E2 as indicated. A small portion of cell lysate was assayed by western blot (upper panel). The larger portion of the cell lysate was subjected to coIP using ERa antibody, followed by western blot analysis of the eluates using antibodies against Ncoa6, ubiquitin, and ERa (lower panels). The amount of sample loaded was adjusted to comparable input of precipitated ERa. The immunoglobulin G (IgG) bands were from the antibody for coIP. (D) HeLa cells were cotransfected with hER- a-expressing plasmid and a plasmid expressing hNCOA6a, hNCOA6b (a splicing isoform without exon 10), or no protein (control). The transfected

cells were treated with MG132 and/or E2 as indicated. A small portion of the cell lysate was assayed by western blot (upper panel). The re- maining cell lysate was subjected to coIP using ERa antibody or control IgG. The eluted precipi- tates were analyzed by western blot for Ncoa6, HA-ubiquitin, and ERa. See also Figure S4.

2003). Interestingly, the current study demonstrates that Ncoa6 interacts with ERa to accelerate ERa ubiquitination

ECs. Thus, the decreased expression of PR target genes and proteasome-mediated degradation regardless of E2-binding including Ihh, Areg and Lpar3 in ECs of Ncoa6d/d uteri may be status. More importantly, this Ncoa6-associated ERa degrada- a result of inhibition from Ncoa6 deficiency-triggered E2 super- tion plays an essential role to avoid abnormal increase in ERa sensitivity. The reduced Ihh expression should explain the protein and thereby prevent the uterus from acquiring E2 super- decreased expression of Ihh-inducible genes Bmp2 and sensitivity for appropriate embryo implantation. Therefore, COUP-TFII in the stroma of Ncoa6d/d uteri (Kurihara et al., 2007; Ncoa6-accelerated ERa ubiquitination and degradation do not Lee et al., 2010). Since both Bmp2 and COUP-TFII are essential functionally couple with transcriptional activation but actually for uterine decidual response (Kurihara et al., 2007; Lee et al., attenuate the expression of ERa target genes in the uterus under 2010), the Ncoa6 deficiency-impaired decidual response should natural conditions. be caused by the decreased expression of these genes, as evi- Increased levels of ERa and SRC-3 expression have been denced by the partial rescue of decidual response in Ncoa6d/d found in the endometrium of women with PCOS associated uteri treated with BMP2 protein. However, the regulatory mecha- with E2 supersensitivity and poor reproductive performance nisms for the decreased expression of Hoxa10 and Hand2 in the (Gregory et al., 2002). Similarly, in Ncoa6d/d mouse uteri, stroma of Ncoa6d/d uteri are currently unclear, which could be Ncoa6 deficiency also increases ERa in the stroma and SRC-3 a consequence of impaired uterine receptivity caused by E2 in the epithelium during the preimplantation period and resulted supersensitivity. in E2 supersensitivity and implantation failure. This suggests that In yeast, the 19S components of the 26S proteasome interact loss of Ncoa6 expression or function might be involved in ERa with transcriptional elongation factors to enhance transcription and SRC-3 upregulation and E2 supersensitivity observed in elongation in an independent manner to their proteolytic roles the uteri of women with PCOS.

(Ferdous et al., 2001). In cultured mammalian cells, E2-bound ERa recruits coactivators and proteasome components to the EXPERIMENTAL PROCEDURES target gene promoters to form a transcriptional activation Mouse Experiments complex, in which the proteasome components accelerate tran- Animal protocols were approved by the Baylor College of Medicine Animal a scription by degrading ER that has finished its transcriptional Care and Use Committee. PR-Cre mice were described previously (Soyal activation task, which may be required to reload the next run et al., 2005). Ncoa6f/f and Ncoa6d/d mice were generated as illustrated in Fig- of transcriptional initiation (Lonard et al., 2000; Reid et al., ure S1 and the Supplemental Experimental Procedures. For collecting uterine

Developmental Cell 23, 858–865, October 16, 2012 ª2012 Elsevier Inc. 863 Developmental Cell

Ncoa6, E2 Sensitivity, and Embryo Implantation

specimens from pregnant or pseudopregnant mice, 8- to 10-week-old females Statistical Analysis were paired with intact or vasectomized WT males. Copulatory plugs were Data are presented as means ± SEM. Statistical analyses were performed checked in the morning, and the morning when vaginal plug was observed using a two-sided unpaired Student’s t test, Mann-Whitney’s U test, one-way was designated as 0.5 dpc. Ovulation and fertilization were examined at analysis of variance, or analysis of covariance. Multiplicity adjustment 1.5 dpc. The early blastocyst development was examined at 3.5 dpc. ISs methods (Bonferroni method) were used when multiple pairwise comparisons were examined at 4.5 or 5.5 dpc. The serum concentrations of E2 and P4 were performed. Chi-square tests were used to evaluate differences of data in were measured using blood samples collected at 3.5 dpc. Detail method is percentages. described in the Supplemental Experimental Procedures. For E2 and P4 treat- ments, 6-week-old female mice were ovariectomized, allowed to rest for SUPPLEMENTAL INFORMATION 2 weeks, and then subcutaneously injected with vehicle (sesame oil), 6.7 ng of E2, or 1 mg of P4 in sesame oil per mouse. Six hours later, uteri were Supplemental information includes four figures and Supplemental Experi- À  collected, frozen in liquid nitrogen, and stored at 80 C for analysis. For ICI mental Procedures and can be found with this article online at http://dx.doi. treatment, a previous protocol (Lee et al., 2010) was slightly modified. Each org/10.1016/j.devcel.2012.09.002. mouse was intraperitoneally injected with 5 ng of ICI in 200 ml of sesame oil at 11 pm on 3.0 dpc. ISs were analyzed at 4.5 dpc. ACKNOWLEDGMENTS

RT-PCR We thank Dr. T. Takahashi for technical advice. This work is partially supported RNA was isolated from uteri using Trizol reagent and converted to cDNA using by National Institutes of Health grants CA119689, DK058242, and CA112403. the Reverse Transcriptase Core Kit (Eurogentec). Relative mRNA levels were measured using matched Universal TaqMan real-time PCR probes (Roche) Received: March 7, 2012 and gene-specific primers. Ncoa6 mRNA was measured using the qPCR Revised: July 26, 2012 Core kit for SYBR Green I (Eurogentec) with matched primers. All probe and Accepted: September 6, 2012 primer sets for quantitative PCR are listed in the Supplemental Experimental Published online: October 15, 2012 Procedures. 18S RNA was measured using the TaqMan Ribosomal RNA Control Reagents (Applied Biosystems) and used to normalize relative REFERENCES mRNA levels in each sample.

Basir, G.S., O, W.S., Ng, E.H., and Ho, P.C. (2001). Morphometric analysis of Primary Culture of Uterine Stromal Cells peri-implantation endometrium in patients having excessively high oestradiol Uterine SCs were isolated from pseudopregnant mice at 3.5 dpc as described concentrations after ovarian stimulation. Hum. Reprod. 16, 435–440. in the Supplemental Experimental Procedures. For analysis of ERa degrada- tion and ubiquitination, cells were cultured in Dulbecco’s modified Eagle’s Caira, F., Antonson, P., Pelto-Huikko, M., Treuter, E., and Gustafsson, J.A. medium/F12 medium with 10% charcoal-stripped FBS for 48 and followed (2000). Cloning and characterization of RAP250, a novel nuclear receptor co- by treatment with reagents as indicated. activator. J. Biol. Chem. 275, 5308–5317. Cakmak, H., and Taylor, H.S. (2011). Implantation failure: molecular mecha- IHC nisms and clinical treatment. Hum. Reprod. Update 17, 242–253. m Uterine sections 5 m thick were prepared from formalin-fixed and paraffin- Cooke, P.S., Buchanan, D.L., Young, P., Setiawan, T., Brody, J., Korach, K.S., embedded tissues and processed for hematoxylin and eosin (H&E) staining Taylor, J., Lubahn, D.B., and Cunha, G.R. (1997). Stromal estrogen receptors and IHC as described (Lee et al., 2006). Antibodies for IHC are listed in Supple- mediate mitogenic effects of estradiol on uterine epithelium. Proc. Natl. Acad. mental Experimental Procedures. Sci. USA 94, 6535–6540. Ferdous, A., Gonzalez, F., Sun, L., Kodadek, T., and Johnston, S.A. (2001). The Transfection 19S regulatory particle of the proteasome is required for efficient transcription Plasmid constructs are described in the Supplemental Experimental Proce- elongation by RNA polymerase II. Mol. Cell 7, 981–991. dures. HeLa cells were transfected with plasmids using lipofectamine 2000 (Invitrogen). MCF-7 cells were transfected with plasmids using lipofectamine Gregory, C.W., Wilson, E.M., Apparao, K.B., Lininger, R.A., Meyer, W.R., LTX (Invitrogen). Total amount of DNA used in each transfection assay was Kowalik, A., Fritz, M.A., and Lessey, B.A. (2002). Steroid receptor coactivator balanced by adding the pcDNA3 parent plasmid. expression throughout the menstrual cycle in normal and abnormal endome- trium. J. Clin. Endocrinol. Metab. 87, 2960–2966. Western Blot and CHX Chase Hama, K., Aoki, J., Bandoh, K., Inoue, A., Endo, T., Amano, T., Suzuki, H., and Protein was extracted from whole uterus or cultured cells using RIPA buffer Arai, H. (2006). Lysophosphatidic receptor, LPA3, is positively and negatively with proteinase inhibitors and phosphatase inhibitors. Western blotting was regulated by progesterone and estrogen in the mouse uterus. Life Sci. 79, performed with 30 mg of protein in each lysate and antibodies listed in the 1736–1740. Supplemental Experimental Procedures. To assay ERa stability, mouse Hewitt, S.C., Li, Y., Li, L., and Korach, K.S. (2010). Estrogen-mediated uterine SCs were treated with 10 mg/ml of CHX. Thirty minutes after adding regulation of Igf1 transcription and uterine growth involves direct binding of CHX, cells were treated with vehicle (ethanol) or 10 nM of E2. Band intensities to estrogen-responsive elements. J. Biol. Chem. of western blots were quantified by densitometry using Image J software. 285, 2676–2685. Jeong, J.W., Lee, K.Y., Han, S.J., Aronow, B.J., Lydon, J.P., O’Malley, B.W., a ER Ubiquitination Assay and DeMayo, F.J. (2007). The p160 steroid receptor coactivator 2, SRC-2, m Mouse uterine SCs were pretreated with vehicle (dimethyl sulfoxide) or 10 M regulates murine endometrial function and regulates progesterone-indepen- of MG132 for 30 min and then treated with vehicle (ethanol) or 10 nM of E2 for dent and -dependent gene expression. Endocrinology 148, 4238–4250. 6 hr. HeLa cells were transfected with mock or expression vectors for hERa, Kennedy, T.G. (1986). Intrauterine infusion of prostaglandins and decidualiza- hNCOA6a, hNCOA6b, and HA-ubiquitin. The transfected cells were treated tion in rats with uteri differentially sensitized for the decidual cell reaction. Biol. with vehicle or 10 mM of MG132 for 6.5 hr and with vehicle or 10 nM of E 2 Reprod. 34, 327–335. for 6 hr (added 0.5 hr later after MG132 addition). The treated cells were lysed in TNE buffer (10 mM Tris-HCl, pH 7.8, 1% NP40, 0.15 M NaCl, 1 mM EDTA) Ko, L., Cardona, G.R., and Chin, W.W. (2000). Thyroid hormone receptor- containing proteinase inhibitors and phosphatase inhibitors. ERa was immu- binding protein, an LXXLL motif-containing protein, functions as a general co- noprecipitated from cell lysate with 500 mg protein using ERa antibody. The activator. Proc. Natl. Acad. Sci. USA 97, 6212–6217. precipitates were assayed by western blots using ERa, ubiquitin, and Ncoa6 Kuang, S.Q., Liao, L., Zhang, H., Pereira, F.A., Yuan, Y., DeMayo, F.J., Ko, L., antibodies. and Xu, J. (2002). Deletion of the cancer-amplified coactivator AIB3 results in

864 Developmental Cell 23, 858–865, October 16, 2012 ª2012 Elsevier Inc. Developmental Cell

Ncoa6, E2 Sensitivity, and Embryo Implantation

defective placentation and embryonic lethality. J. Biol. Chem. 277, 45356– Mahajan, M.A., and Samuels, H.H. (2005). Nuclear hormone receptor coregu- 45360. lator: role in hormone action, metabolism, growth, and development. Endocr. Kurihara, I., Lee, D.K., Petit, F.G., Jeong, J., Lee, K., Lydon, J.P., DeMayo, F.J., Rev. 26, 583–597. Tsai, M.J., and Tsai, S.Y. (2007). COUP-TFII mediates progesterone regulation Matsumoto, H., Zhao, X., Das, S.K., Hogan, B.L., and Dey, S.K. (2002). Indian of uterine implantation by controlling ER activity. PLoS Genet. 3, e102. hedgehog as a progesterone-responsive factor mediating epithelial-mesen- Large, M.J., and Demayo, F.J. (2011). The regulation of embryo implantation chymal interactions in the mouse uterus. Dev. Biol. 245, 280–290. and endometrial decidualization by progesterone receptor signaling. Mol. Mussi, P., Yu, C., O’Malley, B.W., and Xu, J. (2006). Stimulation of steroid Cell. Endocrinol. 358, 155–165 receptor coactivator-3 (SRC-3) gene overexpression by a positive regulatory Lee, D.K., Kurihara, I., Jeong, J.W., Lydon, J.P., DeMayo, F.J., Tsai, M.J., and loop of E2F1 and SRC-3. Mol. Endocrinol. 20, 3105–3119. Tsai, S.Y. (2010). Suppression of ERalpha activity by COUP-TFII is essential for Reid, G., Hu¨ bner, M.R., Me´ tivier, R., Brand, H., Denger, S., Manu, D., successful implantation and decidualization. Mol. Endocrinol. 24, 930–940. Beaudouin, J., Ellenberg, J., and Gannon, F. (2003). Cyclic, proteasome- Lee, K., Jeong, J., Kwak, I., Yu, C.T., Lanske, B., Soegiarto, D.W., Toftgard, R., mediated turnover of unliganded and liganded ERalpha on responsive Tsai, M.J., Tsai, S., Lydon, J.P., and DeMayo, F.J. (2006). Indian hedgehog is promoters is an integral feature of estrogen signaling. Mol. Cell 11, 695–707. a major of progesterone signaling in the mouse uterus. Nat. Genet. Soyal, S.M., Mukherjee, A., Lee, K.Y., Li, J., Li, H., DeMayo, F.J., and Lydon, 38, 1204–1209. J.P. (2005). Cre-mediated recombination in cell lineages that express the Lee, S.K., Anzick, S.L., Choi, J.E., Bubendorf, L., Guan, X.Y., Jung, Y.K., progesterone receptor. Genesis 41, 58–66. Kallioniemi, O.P., Kononen, J., Trent, J.M., Azorsa, D., et al. (1999). A nuclear Tibbetts, T.A., Mendoza-Meneses, M., O’Malley, B.W., and Conneely, O.M. factor, ASC-2, as a cancer-amplified transcriptional coactivator essential for (1998). Mutual and intercompartmental regulation of estrogen receptor and ligand-dependent transactivation by nuclear receptors in vivo. J. Biol. Chem. progesterone receptor expression in the mouse uterus. Biol. Reprod. 59, 274, 34283–34293. 1143–1152. Li, Q., and Xu, J. (2011). Identification and characterization of the alternatively Valbuena, D., Jasper, M., Remohı´, J., Pellicer, A., and Simo´ n, C. (1999). spliced nuclear receptor coactivator-6 isoforms. Int. J. Biol. Sci. 7, 505–516. Ovarian stimulation and endometrial receptivity. Hum. Reprod. 14 (Suppl 2 ), Li, Q., Kannan, A., DeMayo, F.J., Lydon, J.P., Cooke, P.S., Yamagishi, H., 107–111. Srivastava, D., Bagchi, M.K., and Bagchi, I.C. (2011). The antiproliferative Wang, H., and Dey, S.K. (2006). Roadmap to embryo implantation: clues from action of progesterone in uterine epithelium is mediated by Hand2. Science mouse models. Nat. Rev. Genet. 7, 185–199. 331, 912–916. Winuthayanon, W., Hewitt, S.C., Orvis, G.D., Behringer, R.R., and Korach, K.S. Lonard, D.M., Nawaz, Z., Smith, C.L., and O’Malley, B.W. (2000). The 26S pro- (2010). Uterine epithelial estrogen receptor a is dispensable for proliferation teasome is required for estrogen receptor-alpha and coactivator turnover and but essential for complete biological and biochemical responses. Proc. Natl. for efficient estrogen receptor-alpha transactivation. Mol. Cell 5, 939–948. Acad. Sci. USA 107, 19272–19277. Ma, W.G., Song, H., Das, S.K., Paria, B.C., and Dey, S.K. (2003). Estrogen is Xu, J., Wu, R.C., and O’Malley, B.W. (2009). Normal and cancer-related a critical determinant that specifies the duration of the window of uterine functions of the p160 steroid receptor co-activator (SRC) family. Nat. Rev. receptivity for implantation. Proc. Natl. Acad. Sci. USA 100, 2963–2968. Cancer 9, 615–630. Mahajan, M.A., and Samuels, H.H. (2000). A new family of nuclear receptor Zhang, H., Liao, L., Kuang, S.Q., and Xu, J. (2003). Spatial distribution of the coregulators that integrate nuclear receptor signaling through CREB-binding messenger ribonucleic acid and protein of the nuclear receptor coactivator, protein. Mol. Cell. Biol. 20, 5048–5063. amplified in breast cancer-3, in mice. Endocrinology 144, 1435–1443.

Developmental Cell 23, 858–865, October 16, 2012 ª2012 Elsevier Inc. 865