Steroid receptor coactivator 2 is required for female fertility and mammary morphogenesis: insights from the mouse, relevance to the human.

The SRC-1 knockout (KO) mouse was shown to display an attenuated decidual response in the uterus [ Xu et al., 1998 ], suggesting that this coregulator (in concert with others) is required for complete manifestation of this morphological response, which requires initial progesterone stimulation. Although the SRC-3KO exhibits a normal decidual response, a partial block in hormone-induced mammary ductal side-branching and alveologenesis is observed in this animal [ Xu et al., 2000 ]. Collectively, these observations support the proposal that SRC-1 and -3 are required for a subgroup of PR-mediated transcriptional responses in the uterus and mammary gland, respectively. Investigations on the PR activity indicator (PRAI) model provide further support for this supposition [ Han et al., 2006 ; Han et al., 2005 ]. Mouse studies have also underscored important roles for SRC-1 and SRC-3 in areas of normal physiology and disease which are outside the realm of progesterone control, these include: cell growth [ Wang et al., 2000 ; Xu et al., 2000 ], metabolism [ Louet et al., 2006 ; Wang et al., 2006 ], thyroid hormone-based physiology [ Ying et al., 2005 ], bone homeostasis [ Modder et al., 2004 ], prostate biology [ Zhou et al., 2005 ] and B-cell lymphoma developmental progression [ Coste et al., 2006 ]. In contrast to KOs for SRC-1 and -3, the global KO for SRC-2 (termed: Transcriptional Intermediary Factor 2 KO (or TIF2 -/- )) displays striking reproductive abnormalities in both sexes [ Gehin et al., 2002 ]. In the female, abrogation of SRC-2 function triggers placental hypoplasia, which results in a severe hypofertility defect. Subsequent studies found that TIF2 -/- pups (both sexes) are significantly underrepresented in litters from TIF2 +/- crosses (TIF2 -/- females resulting from such crosses are infertile (personal observations)). Similar to KOs for SRC-1 and -3, global ablation of SRC-2 function results in physiological defects not directly linked to reproductive biology, such as a decrease in early postnatal survival [ Mark et al., 2004 ], a breakdown in energy homeostasis [ Jeong et al., 2006 ], as well as elaboration of insidious adrenocortical insufficiency [ Patchev et al., 2007 ]).

The placental defect exhibited by the global KO for SRC-2, in conjunction with the recent observation that a subset of murine cell lineages express both PR and SRC-2 [ Mukherjee et al., 2006b ], suggested that SRC-2 (like SRC-1 and -3) may occupy a critical coregulator role in a subgroup of physiological processes that require PR function. To test this hypothesis, a PR Cre/+ SRC-2 flox/flox bigenic mouse was created [ Mukherjee et al., 2006b ] by crossing a PR Cre/+ knockin mouse [ Soyal et al., 2005 ] with a SRC-2 flox/flox mouse in which exon 11 of the SRC-2 gene was floxed to enable cre-mediated excision [ Gehin et al., 2002 ]; exon 11 encodes the receptor-interacting domain (RID). Therefore, the PR Cre/+ SRC-2 flox/flox bigenic is designed to abrogate SRC-2 function specifically in cell lineages that are PR positive [ Mukherjee et al., 2006b ]. The utility of this genetic strategy is that SRC-2’s role in PR-regulated transcriptional programs can be directly investigated at the whole-animal level without interference from other, unrelated phenotypes resulting from SRC-2’s absence (a key advantage over the global KO for SRC-2).

To date, female and male PR Cre/+ SRC-2 flox/flox mice show normal postnatal development; however, the PR Cre/+ SRC-2 flox/flox female is sterile [ Mukherjee et al., 2006b ]. Unlike the TIF2 -/- mouse, male PR Cre/+ SRC-2 flox/flox mice exhibit normal fertility and neither sex displays phenotypes (outside progestin control) previously described for the TIF2 -/- model [ Gehin et al., 2002 ]. The absence of implantation sites along the uterine horn of the PR Cre/+ SRC-2 flox/flox mouse (5.5 days post-coitum) is the primary underlying cause for the infertility defect displayed by the PR Cre/+ SRC-2 flox/flox female ( Figure 2 A). This result suggests a pivotal role played by SRC-2 in the early cellular changes in the uterus that are required for embryo implantation. Although the SRC-2 flox/flox uterus (a positive control) exhibits a full decidual response to an artificial deciduogenic stimulus ( Figure 2 B and C), the PR Cre/+ SRC-2 flox/flox uterus displays only a partial decidual response ( Figure 2 B and C). These findings support the proposal that a subgroup of PR-mediated transcriptional events are dependent on SRC-2 to launch a complete decidual reaction. The incomplete decidual response phenotype shared by the SRC-1KO [ Xu et al., 1998 ] and PR Cre/+ SRC-2 flox/flox mouse suggests that both SRC coregulators may be required together in PR-mediated signaling cascades that result in a fully decidualized uterus. To test this hypothesis, the SRC-1KO mutation was introduced into the PR Cre/+ SRC-2 flox/flox germline to generate a PR Cre/+ SRC-2 flox/flox SRC-1KO trigenic model. Figure 2 B and C shows that the trigenic uterus fails to mount a decidual response, thereby furnishing essential in vivo support for a cooperative involvement for SRC-1 and SRC-2 in the progesterone-dependent decidual reaction. Note that the PR Cre/+ SRC-2 flox/flox uterine phenotype is not explained by changes in the normal levels of uterine SRC-1 and/or SRC-3 protein ( Figure 2 D). In contrast to the uterus, ovarian and pituitary function is not compromised in the PR Cre/+ SRC-2 flox/flox mouse, suggesting that PR enlists other coregulators in these systems (normal ovarian and pituitary functionality is severely diminished in the PRKO [ Lydon et al., 1995 ]). Furthermore, SRC-2 is not required for progesterone-inhibition of PR expression or suppression of estrogen-induced luminal epithelial proliferation in the uterus. Collectively, these observations suggest that the selective enlistment of SRC-2 by PR in female reproductive tissues may provide one explanation as to why different reproductive tissues display different responses to the same progesterone signal. Having disclosed a central role for uterine SRC-2 in murine peri-implantation biology, future questions to be addressed include: (1) Is uterine SRC-2 expression in the epithelial, stromal, or both cellular compartments required for the development of the receptive uterus? (2) Is SRC-2 expressed in the embryonic-derived trophectoderm? If so, is trophectoderm-derived SRC-2 required for embryo implantation? (3) Does uterine SRC-2 possess coregulator functions necessary for later stages of pregnancy? For example, in the regulation of the onset of parturition; and (4) Does SRC-2 have a role in the etiopathogenesis of such endometrial disorders as uterine hyperplasia and/or endometriosis?

Pioneering in vitro studies by the O’Malley group revealed that the transactivational potency of agonist bound progesterone receptor (PR) can be significantly enhanced by increasing the cellular level of members of the steroid receptor coactivator (SRC/p160) family of coregulators, reviewed in [ McKenna and O'Malley, 2002 ]. The SRC/p160 family consists of three members: SRC-1 (ERAP140/ERAP160/NcoA-1); SRC-2 (TIF-2/GRIP-1/NcoA-2); and SRC-3 (p/CIP/RAC3/AIB1/TRAM-1/ACTR/NcoA-3); reviewed in [ Lonard and O'Malley, 2005 ]. To enhance nuclear receptor (NR)-mediated transactivation, each SRC family member has been shown to directly contact - through discreet LXXLL motifs within their NR interaction domain ( Figure 1 ) - the highly conserved activation 2 domain located in the C-terminal region of NRs. Furthermore, two activation domains (AD1 and AD2) positioned in the C-terminal region of each SRC are responsible for recruiting secondary coregulators (or co-coregulators). For example, AD1 is known to interact with histone acetyltransferases (HATs) p300 and the related cyclic AMP-response element binding protein (CREB)-binding protein (CBP), whereas AD2 is known to recruit arginine methyltransferases such as coactivator-associated arginine methyltransferase 1 (CARM1), reviewed in [ Lonard and O'Malley, 2006 ]. The histone-modifying activities of these secondary coregulators (in addition to the weak intrinsic HAT activity of SRC members) facilitate local chromatin remodeling that allows the general transcriptional machinery open access to promoter regions of NR target genes. Apart from histones, these co-coregulators have been shown to posttranslationally modify other target proteins (i.e., other coregulators and transcription factors) within the transcriptional complex. The N-terminally positioned basic helix loop helix-Per/ARNT/Sim (bHLH-PAS) domain is the most conserved structural motif among SRC members ( Figure 1 ) and is also responsible for co-opting additional coregulators and transcription factors. For SRC-2, these coregulators include coiled-coil coactivator (CoCoA) [ Kim et al., 2003 ], flightless-I (Fli-I) [ Lee and Stallcup, 2006 ], GRIP1-associated coactivator 63 (GAC63) [ Chen et al., 2005 ], as well as the transcription factors myocyte-enhancer factor 2C (MEF-2C) [ Chen et al., 2000b ] and TEF4 [ Belandia and Parker, 2000 ]. In the case of other regulatory proteins, the bHLH-PAS motif has been shown to be involved in both DNA and ligand binding [ Gu et al., 2000 ; Huang et al., 1993 ], indicating that this structural domain feature may be involved in SRC regulatory functions beyond those currently known. Superimposed on the myriad of protein-protein interactions that enable SRC-2 to relay (and control) signaling inputs dispatched from ligand-bound NR to the general transcriptional complex, a multiplicity of interacting signaling inputs (i.e., phosphorylation events triggered by extracellular growth and cell survival factors [ Duong et al., 2006 ; Frigo et al., 2006 ]) are also being transduced by SRC-2 within the multicomponent transcriptional machinery. Although SRC-2 has been primarily considered a coactivator, a subset of investigations provide strong support for a corepressor role for SRC-2 within certain cellular contexts [ Gupta et al., 2007 ; He and Simons, 2007 ; Rogatsky et al., 2002 ; Wang et al., 2007 ]; these studies highlight the versatility and complexity of this multifunctional coregulator. While in vitro experiments disclosed the existence of the SRC family, subsequent experimental mouse genetics would uncover important overlapping and non-overlapping roles for the three SRCs in progestin-initiated signaling events in vivo . Further underscoring their multifunctional properties, mouse studies have also uncovered critical roles for each SRC in signaling processes that reside outside the physiologic area of progestin control.

The detection of SRC-2 protein in mammary epithelial cells that are PR positive [ Mukherjee et al., 2006b ] suggested that mammary SRC-2 may occupy a crucial role in PR-mediated proliferative programs which result in ductal side-branching and alveolar morphogenesis in the mammary gland of the adult. This assumption was supported by the observation that the PR Cre/+ SRC-2 flox/flox mammary gland fails to exhibit the typical morphological changes that occur with combined estrogen and progestin exposure ( Figure 3 A-D). Like the PRKO [ Lydon et al., 1999 ], the underlying cause of the PR Cre/+ SRC-2 flox/flox mammary phenotype is a failure of the mammary epithelium to proliferate in response to hormone ( Figure 3 E). These results support an essential role for SRC-2 in progesterone-induced signaling programs which are required for mammary morphogenesis in the adult. Of note, the PR Cre/+ SRC-2 flox/flox mammary defect was not compensated for by SRC-3 ( Figure 3 E). Although SRC-3 has been shown to be involved in steroid-induced mammary morphogenesis [ Xu et al., 2000 ], as well as tumorigenesis [ Kuang et al., 2005 ; Kuang et al., 2004 ; Torres-Arzayus et al., 2004 ], our data to date suggest that SRC-2 and -3 are operationally distinct in the murine mammary epithelial cell. Irrespective of the functional interrelationships between SRC-2 and other members of the SRC family in this tissue, our studies reveal SRC-2 to be an important coactivator for progestin-initiated signaling in the mammary epithelium. An important question for the future will be to determine whether SRC-2 (like SRC-3/AIB-1 [ Anzick et al., 1997 ]) can act as a mammary oncogene.

Although state-of-the-art genetics demonstrated a critical role for SRC-2 in a subset of progesterone responses in the uterus and mammary gland of the mouse, whether these findings translate to the human is now an important research focus. As previously shown [ Hofman et al., 2002 ], the transactivational potency of the human PR ortholog is significantly enhanced with increasing levels of human SRC-2 ( Figure 4 A). These observations provide strong support for a coregulator involvement for SRC-2 in progestin-dependent physiological processes in the human. As further support for this supposition, immunohistochemistry demonstrates that SRC-2 protein is expressed in a subset of steroid-responsive target tissues in the human ( Figure 4 B-F). In the case of human prostate, ( Figure 4 B), SRC-2 expression is regionally restricted to the epithelial compartment, a known cellular target-site for androgen receptor-mediated signaling and neoplastic transformation [ Berrevoets et al., 2004 ; Culig et al., 2002 ; Ye et al., 2005 ]. In the human endometrium, immunohistochemical studies clearly demonstrate that SRC-2 and PR are expressed in identical cell types within the stromal and epithelial compartments ( Figure 4 C-D); similar findings have been described for the mouse [ Mukherjee et al., 2006b ]. Similar to the murine mammary gland [ Mukherjee et al., 2006b ], immunohistochemical investigations have shown that a subgroup of epithelial cells within the normal human breast express SRC-2 ( Figure 4 E and F). Interestingly, the punctate spatial expression pattern for SRC-2 in the human breast parallels a similar spatial expression pattern reported for PR in the rodent and human breast [ Clarke et al., 1997 ]. However, whether (like the mouse [ Mukherjee et al., 2006b ]) SRC-2 and PR localize to identical cells in the human breast has yet to be demonstrated. Further progress in this area is important, as separation of PR positive mammary epithelial cells from PR negative cells that undergo cell division in response to progesterone is now recognized as an evolutionarily conserved feature that underpins a proposed paracrine mechanism action for PR in the normal breast, reviewed in [ Fernandez-Valdivia et al., 2005 ].

Despite over 200 known coactivators reported [ Lonard and O'Malley, 2006 ], it is significant that PR action is singularly dependent on the coregulator functions of SRC-2 for a subgroup of progesterone-induced physiological responses that are necessary for the maintenance of female fertility and postnatal mammary morphogenesis in the mouse. Overtly distinct from SRC-1 and SRC-3, coregulator properties of which subserve only a selection of progesterone-initiated transcriptional responses either in the uterus or mammary gland, SRC-2 exerts potent coregulator activities in both progesterone target tissues. From a clinical standpoint, the indispensable role of SRC-2 in murine peri-implantation biology demands further study, since recurrent implantation failure is now recognized as a key underlying factor that precludes the establishment of a successful pregnancy [ Norwitz et al., 2001 ]). Importantly, abnormal increases in endometrial SRC-2 levels have also been associated with infertility in women with polycystic ovarian syndrome (PCOS) and with a subset of endometrial cancers [ Gregory et al., 2002 ; Pathirage et al., 2006 ]. Although preliminary, these latter observations suggest a possible connection between perturbation in SRC-2 protein levels and the etiopathogenesis of these uterine disorders. In the case of the murine mammary gland, previous studies demonstrated that PR action is necessary for parity-induced mammary proliferation, which represents a prerequisite developmental step prior to terminal differentiation of this tissue; importantly, the progesterone signal can also influence breast cancer susceptibility, reviewed in [ Fernandez-Valdivia et al., 2005 ]. The finding that SRC-2 ablation results in a mammary phenotype similar to the PRKO mammary defect has prompted three key questions: (1) Can upregulation of SRC-2 expression promote neoplastic transformation in the murine mammary gland? (2) If so, does SRC-2 have an involvement in hormone-responsive breast cancers in the human? and (3) Does the established oncogenic effects of SRC-3 require the presence of SRC-2? Obviously, addressing these questions will extend our current understanding of progesterone’s role in breast cancer promotion and/or progression, and thus may enable the formulation of more powerful diagnostic, prognostic and/or therapeutic approaches in the future clinical containment of this cancer.