The steroid hormone, progesterone, plays a central coordinate role in diverse events associated with female reproduction. In humans and other vertebrates, the biological activity of progesterone is mediated by modulation of the transcriptional activity of two progesterone receptors, PR-A and PR-B. These receptors arise from the same gene and exhibit both overlapping and distinct transcriptional activities in vitro. To delineate the individual roles of PR-A and PR-B in vivo, we have generated mouse models in which expression of a single PR isoform has been ablated. Analysis of the reproductive phenotypes of these mice has indicated that PR-A and PR-B mediate mostly distinct but partially overlapping reproductive responses to progesterone. While selective ablation of the PR-A protein (PR-A knockout mice, PRAKO mice) shows normal mammary gland response to progesterone but severe uterine hyperplasia and ovarian abnormalities, ablation of PR-B protein (PRBKO mice) does not affect biological responses of the ovary or uterus to progesterone but results in reduced pregnancy-associated mammary gland morphogenesis. The distinct tissue-specific reproductive responses to progesterone exhibited by these isoforms are due to regulation of distinct subsets of progesterone-dependent target genes by the individual PR isoforms. This review will summarize our current understanding of the selective contribution of PR isoforms to the cellular and molecular actions of progesterone in reproductive tissues.
The steroid hormone, progesterone, is a key modulator of normal reproductive functions. These include ovulation, uterine and mammary gland development and the neurobehavioral expression associated with sexual responsiveness (Clarke & Sutherland 1990, Lydon et al. 1995). The physiological effects of progesterone are mediated by interaction of the hormone with two specific intracellular progesterone receptors (PRs) termed PR-A and PR-B. Progesterone receptors are members of a large family of structurally related gene products known as the nuclear receptor (NR) superfamily (Evans 1988, O’Malley & Conneely 1992, Tsai & O’Malley 1994a, Mangelsdorf et al. 1995) of transcription factors. NRs regulate gene transcription by discriminative binding to DNA regulatory sequences as well as by specific interactions with co-activator and/or co-repressor proteins to regulate the activity of the RNA polymerase complex (McKenna & O’Malley 2001).
PRs are expressed from a single gene as a result of transcription from two alternative promoters (Kastner et al. 1990, Kraus & Katzenellenbogen 1993) and translation initiation at two alternative AUG initiation codons (Conneely et al. 1989). The production of these two isoforms is conserved in a number of vertebrate species including humans and rodents (Lessey et al. 1983, Conneely et al. 1989, Shyamala et al. 1990, Giangrande & McDonnell 1999) and the ratios of the individual isoforms vary in reproductive tissues as a consequence of developmental (Shyamala et al. 1990) and hormonal (Duffy et al. 1997) status and during carcinogenesis (Graham et al. 1996).
Binding of progesterone to PRs induces a significant conformational change on the receptor proteins (Allan et al. 1992a,b) that results in dimerization of two ligand receptor complexes (Tsai et al. 1988, Tsai & O’Malley 1991), increased receptor phosphorylation (Weigel et al. 1995), binding of receptor dimers to specific hormone responsive DNA elements located in the promoter regions of target genes (Gronemeyer 1991, Tsai & O’Malley 1994b), and interaction of the receptor complex with specific coactivator proteins and general transcripton factors (Onate et al. 1995, Kamei et al. 1996) to form a productive transcription initiation complex on specific target gene promoters. The overall structural features of the PRs that are responsible for these activities are well defined (Fig. 1). The amino terminal or A/B region of PRs is the most hypervariable region in terms of both size and amino acid sequence among members of the superfamily. This region contains transactivation domains (AF-1 and AF-3) that recruit coactivator proteins to the receptor to modulate the level and promoter specificity of target gene activation (Tora et al. 1988, Dobson et al. 1989, Meyer et al. 1992, Sartorius et al. 1994) as well an inhibitory domain (ID) responsible for recruitment of transcriptional inhibitory corepressor proteins (Giangrande et al. 1997). The most conserved region (C) is the DNA binding domain (DBD) that, in the case of PRs, is centrally located. This domain consists of approximately 66–68 amino acids and is composed of two type II zinc finger structures that facilitate binding of the receptor to specific cis-acting DNA sequences and are the hallmark of the nuclear receptor superfamily (Luisi et al. 1991, Freedman 1992). A highly conserved ligand binding domain (LBD) is located on the carboxy terminal side of the DBD. In addition to its progesterone binding function, this region contains an additional transactivation domain (AF-2) required for hormone-dependent coactivator recruitment, sequences important for interaction of inactive receptors with heat shock proteins and for receptor dimerization (DIM) (Pratt et al. 1988, Webster et al. 1988, Fawell et al. 1990, Vegeto et al. 1992).
In the early 1990s, in vitro biochemical studies revealed that the two PR isoforms displayed quite distinct transactivational properties that are specific to both the cell type and the target gene promoter used (Tora et al. 1988, Meyer et al. 1992, Vegeto et al. 1993, Hovland et al. 1998). Specifically, PR-B has been shown to function as a strong activator of transcription of several PR-dependent promoters and in a variety of cell types in which PR-A is inactive. In addition, when the PR-A and PR-B proteins are co-expressed in cultured cells, the PR-A can repress the activity of PR-B as well as the activity of other nuclear receptors (Giangrande & McDonnell 1999). Further, the PR-A and PR-B proteins also respond differently to progesterone antagonists (reviewed in Giangrande & McDonnell 1999). While antagonist bound PR-A is inactive, antagonist bound PR-B can be converted to a strongly active transcription factor. When expressed in equimolar ratios in cells, the PR-A and PR-B proteins can dimerize and bind DNA as three distinct species: A:A or B:B homo-dimers or A:B heterodimers. The differential transactivation properties contributed to these complexes by the presence or absence of the PR-B-specific AF-3 domain is likely to contribute to the complete repertoire of physiological responses to progesterone.
While it is well established that PRs mediate the transcription regulatory effects of progesterone, the signal transduction properties of PRs are not restricted to respond to their steroidal ligand. Studies using tissue culture systems have revealed that PRs as well as some other nuclear receptors can be activated in a ligand-independent manner by stimuli that modulate intracellular kinase activity (Denner et al. 1990, Aronica & Katzenellenbogen 1991, 1993, Beck et al. 1993, Sartorius et al. 1993). The physiological relevance of a progesterone-independent mechanism of receptor activation has been substantiated in vivo: PR-mediated lordosis behavior exhibited by rats (Mani et al. 2001) and mice (Mani et al. 1996) can be stimulated either in response to progesterone or in the absence of progesterone by dopamine-activated intracellular signaling pathways.
In addition to its PR-mediated genomic actions, progesterone also stimulates cellular responses that are independent of the transcription regulatory function of PRs by at least two distinct mechanisms. The first is PR dependent and involves PR-mediated activation of intracellular phosphorylation cascades, at least one of which is the Src/Ras/MAP kinase pathway, to modulate cellular responses to the hormone (Boonyaratanakornkit et al. 2001, Ballare et al. 2003). The second mechanism is independent of PRs and appears to be mediated by interactions with specific membrane receptors for progesterone. Plasma membrane-associated progesterone-specific receptors have been isolated and cloned from a range of tissues and from a number of species (Bramley 2003) and are identified as G protein-coupled receptors distinct from classical intra-cellular PRs (Zhu et al. 2003). Rapid non-genomic effects of progesterone have been implicated in modulation of the sperm acrosome reaction (Bronson et al. 1999, Patrat et al. 2000), modulation of neuronal excitability (Genazzani et al. 2000, Lambert et al. 2003), prevention of preterm labor (Grazzini et al. 1998), and Xenopus oocyte maturation (Bayaa et al. 2000). While several of these responses appear to be regulated in a PR-independent manner, the contribution of PR-mediated non-genomic signaling to the physiological activities of progesterone remains poorly understood.
Progesterone receptor knockout (PRKO) models demonstrate tissue specific roles of PRs in reproductive tissues
Phenotypic studies of the PRKO mouse model in which the expression of both isoforms was inhibited by null mutation of the PR gene provided compelling evidence of an essential role of PRs in all aspects of female reproduction. Specifically, female mice lacking both PRs exhibit impaired sexual behavior and neuroendocrine gonadotropin regulation, anovulation, uterine dysfunction and impaired pregnancy-associated mammary gland morphogenesis (Lydon et al. 1995, Mani et al. 1996, Chappell et al. 1999, Tibbetts et al. 1999). Furthermore, studies of the PRKO mouse reveal that PRs also play an essential role in regulation of thymic involution during pregnancy (Tibbetts et al. 1999) and in the cardiovascular system through regulation of endothelial and vascular smooth muscle cell proliferation and response to vascular injury (Vazquez et al. 1999).
The differences in transcriptional activities and coregulator interactions between the PR-A and PR-B observed in vitro predicted that these proteins may mediate different physiological responses to progesterone. In addition, the selective ability of PR-A to inhibit transcriptional responses induced by both PR-B and the estrogen receptors suggested that PR-A has the capacity to diminish overall progesterone responsiveness in certain tissues as well as contribute to the antiestrogenic activities of progesterone previously observed in the uterus (Lydon et al. 1995). However, physiological validation of the functional differences between the PR-A and PR-B isoforms has been hampered due to a lack of information on the specific cell types that express each isoform in vivo and a lack of appropriate animal models in which to dissect their selective functions.
The observation that PR-A and PR-B are produced by translation at two distinct AUG signals encoded by a single gene predicted that mutation of either ATG codon in the PR gene would result in selective ablation of expression of a single isoform in vivo (Conneely et al. 1989, Kastner et al. 1990). Thus, the CRE-loxP gene targeting approach in embryonic stem cells was used to introduce a point mutation into the PR gene at the ATG codon encoding Met 1 (M1L) in order to specifically ablate expression of the PR-B protein, and at the ATG encoding Met 166 (M166A) in order to ablate expression of PR-A (Mulac-Jericevic et al. 2000, 2003). This strategy has provided a powerful model system to examine the selective expression of each isoform in situ as well as to assess the selective contributions of PR-A and PR-B in their normal cellular context to the physiological functions of progesterone.
PRs and ovarian function
Luteinizing hormone (LH), the primary signal for rupture of preovulatory ovarian follicles leading to ovulation, stimulates transient expression of PR mRNA and proteins in granulosa cells (Natraj & Richards 1993, Park-Sarge & Mayo 1994) and the antiprogestin, RU486, inhibits ovulation (Loutradis et al. 1991). Definitive proof that PRs are essential mediators of ovulation has been provided by analysis of the ovarian phenotype of the PRKO mouse. Analysis of this model revealed that PRs are required specifically for LH-dependent follicular rupture leading to ovulation but not for differentiation of granulosa cells to form a corpus luteum (luteinization) (Lydon et al. 1995). Follicular rupture requires induction of a prostaglandin-mediated inflammatory response to LH as well as tissue degradation at the apex of the preovulatory follicle, an event that is mediated by matrix proteinases (Espey 1994). Analysis of the expression of potential mediators of ovulation in PRKO mice has demonstrated that LH-induced regulation of Cox-2, an enzyme that catalyses the production of prostaglandins and is essential for ovulation, is unaffected (Robker et al. 2000). In contrast, the expression of two metalloproteinases, ADAMTS-1 (a desintegrin and metalloproteinase with thrombospondin motifs) and cathepsin-L (a lysosomal cysteine protease) is inhibited in mural granulosa cells of the mature follicle (Tetel et al. 1999, Robker et al. 2000) in PRKO mice. One of these proteases, ADAMTS-1, is essential for ovulation (Robker et al. 2000) suggesting that this protein may represent a critical mediator of the progesterone-induced ovulatory event.
Both the PR-A and PR-B proteins are induced in preovulatory follicles in response to LH stimulation (Natraj & Richards 1993). Analysis of PRAKO mice, in which PR-A is ablated, showed that ovulation is severely impaired but unlike in PRKO mice, is not completely absent (Mulac-Jericevic et al. 2000). Histological analysis of the ovaries of PRAKO mice showed numerous mature anovulatory follicles that contained an intact oocyte and were arrested at a similar stage to that previously observed in PRKO mice. In contrast, ovulation is unaffected in PRBKO mice indicating that PR-A expression is both necessary and sufficient to mediate the ovulatory response to progesterone (Mulac-Jericevic et al. 2003). The ovulatory defects in PRAKO mice must therefore be due to an inability of PR-B to regulate signaling pathways necessary for follicular rupture that are as yet unidentified. The observation that the PR-A and PR-B proteins are not functionally redundant in the ovary provides physiological validation of previous studies in tissue culture demonstrating that these transcription factors have different functional activities. From a mechanistic standpoint, the observation that PR-A alone is sufficient to support normal ovulation indicates that heterodimeric interactions between the PR-A and PR-B proteins are not required for regulation of essential progestin-responsive target genes associated with ovulation. With the exception of the above-mentioned proteases, the PR-dependent signaling pathways that mediate follicular rupture have not yet been elucidated. Because the PRKO mouse has a specific defect in follicular rupture while luteinization is maintained, differential array analysis using this model provides an excellent system to delineate the signaling pathways regulated by PRs that are specific to follicular rupture. The identification of these genes should facilitate the identification of PR isoform selective target genes that are essential for ovulation, in addition to providing important new information on the molecular mechanisms of progesterone-induced follicular rupture.
PRs and uterine function
Progesterone plays a critical role during early pregnancy in preparation of the uterine epithelium for receptivity and in differentiation of endometrial stromal cells to a decidual phenotype that supports the development of the implanting embryo. Furthermore, progesterone is a potent antagonist of estrogen-induced proliferation in the uterine epithelium.
The uterus of PRKO mouse fails to support implantation after embryo transfer and is unresponsive to a decidual stimulus (Lydon et al. 1995). In PRKO mice, epithelial cells in the uterus become hyperplastic as a result of unopposed proliferative estrogen action. PRs are expressed in the epithelial, stromal and myometrial compartments of the uterus and their spatiotemporal expression within these compartments is regulated by both estrogen and progesterone (Tibbetts et al. 1998) and undergoes dynamic changes during the estrous cycle and early pregnancy (Tan et al. 1999). This intercompartmental regulation of PR expression is essential for the appropriate regulation of uterine gene expression (Kurita et al. 1998). Recent analysis of the effects of PR ablation on the expression of several genes previously implicated in progesterone-dependent uterine implantation have shown that the defects observed in PRKO mice are associated with inhibition of expression of several epithelial markers of uterine receptivity (Mulac-Jericevic et al. 2000) and at least one essential stromal mediator of decidualization, hoxa-10 (Lim et al. 1999).
Analysis of progesterone-dependent uterine function in PRAKO and PRBKO mice has revealed that PR-A and PR-B have distinct functions in the uterus (Mulac-Jericevic et al. 2000, 2003). In PRAKO mice, progesterone-induced differentiation of endometrial stromal cells to a decidual phenotype is inhibited, suggesting that PR-A plays a crucial role in decidualization of the stroma prior to implantation (Mulac-Jericevic et al. 2000). In contrast, analysis of uterine function in PRBKO mice has shown that expression of PR-A is both necessary and sufficient to mediate both the antiproliferative- and implantation-associated responses to progesterone (Mulac-Jericevic et al. 2003). The uterine defects observed in PRAKO mice are due to an inability of PR-B to regulate a subset of PR-dependent target genes rather than to differences in the spatiotemporal expression of PR-A relative to PR-B in the uterus (Mulac-Jericevic et al. 2000).
Surprisingly, selective activation of PR-B in the uterus of PRAKO mice resulted in an abnormal progesterone-dependent induction of epithelial cell proliferation in contrast to its ability to inhibit estrogen-induced proliferation in the wild-type uterus (Mulac-Jericevic et al. 2000). This gain of PR-B-dependent proliferative activity upon removal of PR-A indicates that PR-A is required not only to inhibit estrogen-induced hyperplasia of the uterus but also to limit potentially adverse proliferative effects of the PR-B protein.
In the human endometrium, the levels of PR-A and PR-B are differentially regulated during the reproductive cycle (Mangal et al. 1997, Mote et al. 2000). The physiological importance of maintaining the correct relative expression levels of PR isoforms in the uterus is indicated by the identification of aberrant ratios of PR isoforms in human endometrial cancers and the recent identification of a functional polymorphism in the human PR promoter that results in increased expression of the human PR-B isoform and is associated with increased risk of endometrial cancer (Arnett-Mansfield et al. 2001, De Vivo et al. 2002). Given the opposing effects of the PR isoforms in the uterus, one could predict that aberrant changes in the relative spatiotemporal expression patterns of PR isoforms in the uterus could play an important role in determining appropriate responsiveness to progestin therapy in the treatment of uterine epithelial hyperplasias.
PRs and mammary gland morphogenesis
Mammary gland development is predominantly postnatal and is controlled by a complex interplay of endocrine hormones, in particular estrogen, progesterone and prolactin, acting together with locally acting growth factors (Anderson 2002, Soyal et al. 2002). Postnatal mammary gland development involves two distinct growth phases that are initiated at the onset of puberty and pregnancy respectively. At puberty, estrogen promotes ductal elongation and dichotomous branching to the limits of the mammary fat pad. At adulthood, the virgin gland becomes relatively quiescent with the exception of minimal side branching and alveolar budding that occur over time as a result of the cyclic rise of ovarian steroids during the estrous cycle. At pregnancy, exposure to progesterone and prolactin (PRL) results in extensive epithelial proliferation, increased dichotomous side branching and differentiation of milk-filled alveolar lobules. At weaning, removal of the suckling stimulus elicits involution of the lobular alveolar system through apoptosis and matrix degrading proteinase-mediated remodeling. At the end of the involution process, the postnatal developmental cycle of mammary gland development is completed and the mammary gland resembles the general architecture of the pre-pregnant mammary gland (Anderson 2002, Soyal et al. 2002).
Null mutation of both PR isoforms in PRKO mice has demonstrated that PRs are specifically required for pregnancy-associated ductal proliferation and lobuloalveolar differentiation of the mammary epithelium. Mammary glands of PRKO mice failed to develop the pregnancy-associated side-branching of the ductal epithelium with attendant lobular alveolar differentiation despite normal postpubertal mammary gland morphogenesis of the virgin mice (Lydon et al. 1995, Seagroves et al. 2000). Thus, in contrast to its antiproliferative role in the uterus, progester-one is an essential pregnancy-associated proliferative stimulus in the mammary gland.
PRs are expressed exclusively in the mammary epithelium (Seagroves et al. 2000, Sivaraman et al. 2001, Ismail et al. 2002) and development of the mammary gland from juvenile to adult state is associated with a change in the pattern of expression of PRs from a uniform to a scattered pattern of expression in a subset of epithelial cells (Seagroves et al. 2000, Grimm et al. 2002). Recent studies have shown that PR-expressing cells are segregated from proliferating cells in the normal mammary glands of both rodents and humans (Clarke et al. 1997, Seagroves et al. 2000, Ismail et al. 2002). Consistent with this finding, the proliferative responses of the ductal and alveolar epithelium to progesterone are associated with local induction of PR-dependent growth factors that act in a paracrine manner on PR-negative cells to control their proliferation (Brisken et al. 1998, 2000). In contrast to the normal mammary gland, segregation of the steroid receptor expressing cells from proliferating cells is lost in mammary epithelial cells that have been exposed to carcinogen (Sivaraman et al. 2001) and in cells of breast tumors (Graham & Clarke 2002). This aberrant change in pattern of receptor expression is likely to contribute to abnormal growth of breast cancer cells.
Both isoforms of PR are expressed in the mammary gland of the virgin mouse (Shyamala et al. 1990) and during pregnancy (Fantl et al. 1999) and the levels of PR-A protein exceed those of the PR-B isoform. Progesterone-activated PR-B in PRAKO mice elicits side branching and lobular alveolar development in the mammary gland comparable to that of wild-type mammary gland (Mulac-Jericevic et al. 2000). Thus, PR-B is sufficient to elicit normal proliferation and differentiation of the mammary epithelium in response to progesterone and neither process appears to require functional expression of the PR-A protein. In contrast, recent analysis of the mammary glands of PRBKO mice has shown reduced pregnancy-associated ductal side branching and lobuloalveolar development as a consequence of decreased ductal and alveolar epithelial cell proliferation and increased apoptosis of alveolar epithelium (Mulac-Jericevic et al. 2003). Despite these defects, PR-A retains its normal segregated spatiotemporal pattern relative to proliferating cells in PRBKO mice and is expressed at a higher level than that observed for PR-B in PRAKO mice. Examination of the molecular genetic signaling pathways that are differentially regulated by PRs in the mammary gland showed that the defects observed in PRBKO mice are associated with a PR-B isoform-selective regulation of the receptor activator of nuclear factor kappa B (NFκB) ligand (RANKL) signaling pathway (Mulac-Jericevic et al. 2003) which is essential for alveologenesis (Fata et al. 2000). In contrast, progesterone-dependent activation of secreted growth factor, wnt4, a key mediator of branching morphogenesis (Brisken et al. 2000), is normally regulated by either PR-A or PR-B (Mulac-Jericevic et al. 2003). Thus PR-A and PR-B appear to regulate both overlapping and distinct subsets of progesterone-dependent signaling pathways required for mammary gland development.
Analysis of the phenotypic consequences of selective ablation of the PR-A or PR-B isoforms in mice has facilitated a comprehensive analysis of the selective spatiotemporal expression of these receptors in the reproductive tract and mammary gland as well as elucidation of their individual contributions to the established reproductive activities of progesterone. These studies have revealed that PR-A and PR-B exhibit mostly distinct but partially overlapping hormonal responses (Table 1). PR-A activation is both necessary and sufficient for establishment and maintenance of pregnancy but elicits reduced mammary gland morphogenic responses to hormonal stimulation relative to PR-B. In contrast, PR-B activation is insufficient to support female fertility but is a potent proliferative mediator in the mammary gland and, most surprisingly, in the uterus. The distinct tissue-selective activities of PR-A and PR-B observed in vivo in these studies support the conclusion that modulation of PR activity using isoform-selective progestins may contribute towards elimination of adverse tissue-specific side effects of progestins.
Components of female reproductive system and phenotype identified in progesterone receptor knockout mouse models.
|Ovaries||PRKO||Inability to ovulate|
|PRAKO||Severely impaired ovulation|
|Mammary gland||PRKO||Impaired mammary gland development|
|PRAKO||Normal progesterone-induced mammary gland morphogenesis|
|PRBKO||Reduced pregnancy-associated side-branching and lobuloalveolar development|
Aronica SM & Katzenellenbogen BS1991 Progesterone receptor regulation in uterine cells: stimulation by estrogen cyclic adenosine 3′5′-monophosphate and insulin-like growth factor I and suppression by antiestrogens and protein kinase inhibitors. Endocrinology1282045–2052.
Aronica SM & Katzenellenbogen BS1993 Stimulation of estrogen receptor-mediated transcription and alteration in the phosphorylation state of the rat uterine estrogen receptor by estrogen cyclic adenosine monophosphate and insulin-like growth factor-I. Molecular Endocrinology7743–752.
Ballare CUhrig MBechtold TSancho EDi Domenico MMigliaccio AAuricchio F & Beato M2003 Two domains of the progesterone receptor interact with the estrogen receptor and are required for progesterone activation of the c-Src/Erk pathway in mammalian cells. Molecular and Cellular Biology231994–2008.
Grimm SLSeagroves TNKabotyanski EBHovey RCVonderhaar BKLydon JPMiyoshi KHennighausen LOrmandy CJ & Lee AVet al.2002 Disruption of steroid and prolactin receptor patterning in the mammary gland correlates with a block in lobuloalveolar development. Molecular Endocrinology162675–2691.
Pratt WBJolly DJPratt DVHollenberg SMGiguere VCadepond FSchweizer-Groyer GCatelli MGEvans RM & Baulieu EE1988 A region in the steroid binding domain determines formation of the non-DNA binding 9S glucocorticoid receptor complex. Journal of Biological Chemistry263267–273.
Tsai M-J & O’Malley BW1991 Mechanisms of regulation of gene transcription by steroid receptors. In The Hormonal Control of Gene Expression pp 101–116. Eds P Cohen & JG Foulkes. Amsterdam: Elsevier Press.
Tsai M-J & O’Malley BW1994a Mechanism of Steroid Hormone Regulation of Gene Transcription. Molecular Biology intelligence unit Austin TX: RG Landes Co.