Reevaluating the Role of Progesterone in Ovarian Cancer: Is Progesterone Always Protective?

Abstract

Ovarian cancer (OC) represents a collection of rare but lethal gynecologic cancers where the difficulty of early detection due to an often-subtle range of abdominal symptoms contributes to high fatality rates. With the exception of BRCA1/2 mutation carriers, OC most often manifests as a post-menopausal disease, a time in which the ovaries regress and circulating reproductive hormones diminish. Progesterone is thought to be a “protective” hormone that counters the proliferative actions of estrogen, as can be observed in the uterus or breast. Like other steroid hormone receptor family members, the transcriptional activity of the nuclear progesterone receptor (nPR) may be ligand dependent or independent and is fully integrated with other ubiquitous cell signaling pathways often altered in cancers. Emerging evidence in OC models challenges the singular protective role of progesterone/nPR. Herein, we integrate the historical perspective of progesterone on OC development and progression with exciting new research findings and critical interpretations to help paint a broader picture of the role of progesterone and nPR signaling in OC. We hope to alleviate some of the controversy around the role of progesterone and give insight into the importance of nPR actions in disease progression. A new perspective on the role of progesterone and nPR signaling integration will raise awareness to the complexity of nPRs and nPR-driven gene regulation in OC, help to reveal novel biomarkers, and lend critical knowledge for the development of better therapeutic strategies.

Graphical Abstract

Graphical Abstract.

Essential Points.

• Progesterone has traditionally been perceived as a protective factor for epithelial ovarian cancer (OC) since hormonal contraceptive use and certain reproductive states such as pregnancy correlate with lower cancer risk in premenopausal women.

• High-grade serous ovarian cancer (HGSOC), an aggressive subtype of OC, is often considered a postmenopausal disease because diagnosis typically occurs during this stage of reproductive senescence. We now know that precursor lesions develop much earlier, during a time when ovarian steroid hormones like estrogen and progesterone are prevalent, and that these steroids play significant roles in disease initiation and progression.

• The hormonal millieu is much more complex than progesterone alone. Precursor fallopian tube lesions and advanced tumors synthesize and secrete progesterone and estrogen, express their receptors and those of other steroids (eg. androgens, glucocorticoids), and exhibit active steroid receptor signaling, suggesting the importance of a steroid-rich microenvironment.

• Recent in vivo models show that enhanced nuclear progesterone receptor (nPR) signaling can promote OC development whereas loss of nPR expression can block the initiation and metastasis of HGSOC.

• Progestins acting via nPRs can promote dormancy in p53-mutant fallopian tube cell models through DREAM-mediated G0 arrest and enhance tumor phenotypes such as migration, invasion, and spheroid formation.

• Evidence of dysregulated ovarian steroid biosynthesis in BRCA1/2 mutation carriers and growing genomics analyses in cell and animal models suggest an intriguing interplay between BRCA1/2 functions and nPR signaling that may lead to altered downstream DNA damage response in early and late stages of HGSOC.

• An accurate understanding of progesterone's role in disease progression will require an appreciation of the complexity of nPRs and nPR-driven gene regulation and the further exploration of their interplay with other steroid hormone receptors and DNA damage signaling pathways.

Ovarian cancer (OC) is a gynecologic cancer accounting for ∼3.4% of all cancers diagnosed in women globally. Although rare in comparison to the incidence of breast cancer (∼24%), it remains one of the most lethal with the lowest 5-year survival rate compared to breast, endometrial, and cervical cancers (1). Due to the delayed and often subtle presentation of symptoms and the lack of accurate early detection methodologies, OC is often diagnosed at advanced stages of the disease (2). Overall survival following diagnosis is highly dependent on factors such as the histological subtype, disease spread (eg, stage I–IV), primary surgical cytoreduction, race/ethnicity, patient age, and underlying germline or somatic mutations (3, 4). OC is generally considered to be a postmenopausal disease, since the reported median age at diagnosis is ∼63 years (5). It has often been assumed that ovarian sex steroids, estrogen and progesterone, are unlikely to be important in the initiation and progression of OC since ovarian regression with diminished circulating steroids are a hallmark of aging. Such a perspective ignores the younger age at diagnosis for individuals carrying a germline mutation in BRCA1/2, reported to be as much as 10 to 20 years earlier in these patients (6, 7). Early precursor lesions of OC, such as serous tubal intraepithelial carcinoma (STICs), may be present decades before diagnosis (8-10) at a time when signaling mediated by the estogen receptor (ER) and nuclear progesterone receptor (nPR) isoforms may be highly relevant. In addition, if the evidence suggesting reduced OC risk related to specific reproductive states (eg, parity) and contraceptive use is accepted, then further consideration of the presence of sex steroids and/or their metabolites under these conditions and their role in OC is clearly warranted.

Of the primary ovarian sex steroids, progesterone has long been considered to be “protective” and thought to reduce OC risk. This viewpoint was initially influenced by progesterone's often controversial role in the initiation and progression of cancers in endocrine-regulated tissues, especially breast and endometrial cancers (11, 12). In addition, the past findings of epidemiological studies and the work with in vivo and in vitro mouse and OC cell models (reviewed later) appear to support this view. However, recently published work by multiple researchers provides evidence that challenges this singular protective role of nPR signaling. This review will integrate the historical perspective of natural progesterone and synthetic progestins on OC development and progression with exciting new research and critical interpretations to help paint a broader picture of this steroid's role in OC.

Overview of OC Initiation and Progression

Ovarian carcinomas encompass 3 main groups including the most common epithelial tumors (∼90% of cases) and the nonepithelial tumors, germ cell (∼3%), and sex-cord stromal tumors (∼2%); the remaining cases (∼5%) are unspecified (13). Our understanding of the biology and evolution of these highly heterogeneous neoplasms is complicated by the fact that most of these tumor subtypes have little phenotypic or molecular similarity to the cells found within the mature ovary, leading many to wonder if the “ovarian cancer” classification is accurate (13). The epithelial ovarian carcinoma group includes 5 major subtypes: (1) high-grade serous (HGSOC), (2) low-grade serous (LGSOC), (3) endometrioid (EOC), (4) clear cell (CCOC), and (5) mucinous (MOC). These subtypes and their proposed tissues of origin are illustrated in Fig. 1. HGSOC is the most frequently observed epithelial subtype, accounting for about 50% to 60% of all ovarian malignancies (14). Due to diagnosis at a later stage, this subtype is also the most lethal gynecological malignancy, accounting for about 70% of all OC-related deaths (14). This overview will focus on the HGSOC subtype; there are several reviews describing the details of the other subtypes (15-18). See Table 1 for definitions of abbreviations and terms used throughout this review.

Figure 1.

Proposed tissues of origin for epithelial ovarian cancer subtypes. The epithelial ovarian cancer subtypes originate from a diversity of ovarian and extra-ovarian tissues within the abdominal cavity; the definitive tissue of origin for each subtype remains controversial. Mucinous ovarian cancer (MOC) originates as sloughed cells (and possibly mets) from the stomach, colon, or appendix. Endometrioid (EOC) and clear cell ovarian cancer (CCOC) arise from the retrograde transport of sloughed uterine endometrial cells. High-grade serous ovarian cancer (HGSOC) arises from fallopian epithelial cells lesions, ovarian surface epithelia, and/or transformed epithelial trapped in cortical inclusion cysts. These cells aggregate within the abdominal cavity (dashed arrows), and the spheroids circulate and metastasize to other organs and sites (solid arrows) including the mesothelial lining of the peritoneum. See text for specific subtype references.

Table 1.

Nomenclature and abbreviations used

Abbreviation/term	Name	Definition
DREAM	Dimerization partner, RB-like, E2F and Multi-vulval class B (MuvB)	Transcriptional protein complex that promotes G0 arrest through repression of cell cycle dependent genes
FTE	Fallopian tube epithelia	Columnar epithelia that line the luminal surface of the fallopian tube
HGSOC	High-grade serous ovarian cancer	Common, most aggressive epithelial OC subtype, originating from FTE, OSE, and/or cortical inclusion cysts
mPR	Membrane progesterone receptors	7 transmembrane progesterone receptors with similarity to G-protein coupled receptors; also called PAQR, with 5 members
nPR	Nuclear progesterone receptors	Class of PRs found in the nucleus and/or cytoplasm; involved in transcriptional activation; includes all isoforms
nPRA	Nuclear PR-A receptors	The shorter isoform of nPR encoded by the nuclear PGR (progesterone receptor) gene
nPRB	Nuclear PR-B receptors	The longer isoform of nPR encoded by the nuclear PGR gene
OSE	Ovarian surface epithelia	Single layer of squamous-to-cuboidal epithelia covering the ovarian surface
p53 signature	Distinct cellular expression of the tumor suppressor protein, p53	Subset of benign SCOUT lesions that either express high p53 or no p53, are positive for γH2Ax and negative for Ki67, and restricted to fimbrial edge of FT
PGRMC1/2	PR membrane component receptors	Single transmembrane progesterone receptors; may be adaptor proteins for mPRs
PR agonist	Effect similar to endogenous hormone	Ligand/chemical that activates nuclear progesterone receptor, eg, medroxyprogesterone acetate, megestrol acetate
PR antagonist	Antagonizes effect of endogenous hormone	Ligand/chemical that binds but does not activate and/or inhibits nuclear progesterone receptor, eg, onapristone, mifepristone
Progestins	Progestin	Natural OR synthetic steroid hormone including progesterone and other compounds that elicit the biological effects of this hormone
SCOUT	Secretory cell outgrowths	Earliest precursor lesions; 30 secretory FTE cells; found throughout fallopian tube
STIC	Serous tubal intraepithelial carcinoma	Malignant FTE precursor lesions of high mitotic index. Progressed from p53 signature + SCOUTs; able to shed and lead to metastatic HGSOC

The origin of HGSOC was thought to arise from the ovarian surface epithelium (OSE), a flat, single layer of cuboidal epithelial cells modified from the pelvic mesothelium and covering the surface of the ovary (19). When ovulation occurs, the OSE is broken at the ruptured site, and this damaged OSE may invaginate during the repair process, leading to the formation of structures known as ovarian cortical inclusion cysts [CICs (13, 20)]. There is also evidence that these cysts can contain detached fallopian tube epithelia that invaginate into the ovarian stroma. The resulting epithelium within these cysts is directly exposed to the stromal microenvironment that is composed of growth factors, cytokines, and ovarian steroids, and these epithelial cells can transform into cells that express HGSOC markers like PAX8 and CA-125 (19, 21). This repeated cyclical “wounding” and invagination of the OSE has been hypothesized to promote the initiation of OC, termed the “incessant ovulation hypothesis” (22). Decreased ovulatory cycles are correlated with reduced ovarian cancer risk, and epidemiological data on pregnancy and oral contraceptives that show a decreased OC risk have supported the incessant ovulation hypothesis (23-26).

Recently, evidence has demonstrated that HGSOC may also arise from the fallopian tube epithelium (FTE). Women carrying deleterious BRCA 1 and BRCA 2 mutations are known to be at higher risk for OC. When prophylactic removal of the ovaries and fallopian tubes (ie, salpingo-oophorectomy) was performed in these women, distinct lesions were observed in the fallopian tubes (27-31). These lesions appear to arise from secretory FTE cells, not ciliated FTE, that have acquired initial mutations in genes like the tumor suppressor gene, p53 (27, 32). The ovary itself does not appear to exhibit these lesions (33). Additionally, gene expression and genetic methylation profiles of HGSOC tumors are more similar to normal FTE and not the OSE (34). It is now accepted that the predominant site of origin of HGSOC is the FTE, and studies have been underway to determine how FTE cells transform and colonize the ovary.

Normal secretory epithelial cells of the FT, especially in the distal fimbriae of the tube, can become dysplastic under the influence of hormones and factors secreted from follicular fluid during ovulation (35). Factors from follicular fluid that appear to be involved in the transformation of fallopian tube epithelia include IGF and associated proteins, primarily IGF2, with the IGF-1R/AKT mediated signaling pathway able to promote transformation in the fallopian tube (36). Other factors found in follicular fluid that can also be transformative, enhance proliferation, or support pro-survival mechanisms of the FTE are reactive oxygen species, transferrin, TGF-β, and norepinephrine (37-39). The effects these factors have on FTE transformation are covered in more detail in Bergsten et al (37).

Secretory cell outgrowths (SCOUTs) are proposed to be one of the earliest precursor lesions composed of at least 30 secretory FTE cells and found throughout the fallopian tube (40). Expression of the transcription factor PAX2, expressed in normal fallopian tube cells, is lost in about 90% of SCOUT lesions (40). PAX2-null SCOUTs are more commonly found in postmenopausal women and in women who have HGSOC (40, 41). Loss of PAX2 expression in murine oviductal epithelium leads to an increased expression of ER receptor signaling, which can then lead to a perpetuation of cells that are damaged, as ER is known to be carcinogenic (42). A “p53 signature” refers to a subset of SCOUT lesions that are restricted to the fimbrial edge of the fallopian tube. Such a lesion is composed of benign secretory cells that exhibit either robust p53 staining or lack p53 staining, positive γH2Ax staining, and a lack of Ki67 staining (43). Reduced expression of PAX2 is also found in p53 signatures, as approximately 92% of p53 signatures have a loss of PAX2 (42, 44). The Cancer Genome Atlas performed a study on advanced stage HGSOC samples and found that 96% to 100% of HGSOC patients carry TP53 mutations (45). There seems to be a stepwise progression from PAX2 loss to a p53 signature lesion to serous tubal intraepithelial carcinomas (STICs). STICs are malignant cells that feature a high mitotic index and share the same characteristics of DNA damage and p53 mutations as p53 signatures (43). STIC development is an early event in the pathogenesis of HGSOC, and STIC lesions are precursors to metastatic HGSOC (10).

Based on this evidence, the initiation and progression of HGSOC is thought to progress from early SCOUT lesions in the fallopian tube (Fig. 2, number 1a) and/or CICs within the ovary (Fig. 2, number 1b). Cells shed from these sites can disseminate onto the surface of the ovary, the fimbriae of the fallopian tube, and/or directly into the peritoneal cavity (Fig. 2, number 2). These unattached cells are thought to avoid anoikis and enhance survival by aggregating to form “spheroids”; they undergo cell cycle arrest (quiescence) and actively secreting extracellular matrix (Fig. 2, number 3). The circulation of peritoneal fluids supports the transcoelomic spread or metastasis of spheroids to the omentum and other organs, colonizing the mesothelial cell layers that line tissues within the abdominal cavity (Fig. 2, number 4). The dissociation of cells from these aggregates allows migration along and invasion of the mesothelial cell layers, enhancing disease progression (Fig. 2, number 5).

Figure 2.

The development and progression of high-grade serous ovarian cancer. (1a) Primary lesions of the fallopian tube epithelium such as secretory cell outgrowths and serous tubal intraepithelial carcinoma arise due to mutations in genes such as p53 and cyclin E1 (CCNE1). Cells shed from these lesions can seed onto the surface of the ovary, the fimbriae of the fallopian tube, and/or directly enter the peritoneal cavity. (1b) Transformation of ovarian cortical inclusion cysts along with mutations in p53 can lead to invasive lesions, (2) Dissemination of cells from serous tubal intraepithelial carcinomas and lesions on the ovarian surface into the peritoneal cavity. (3) Aggregation of shed cells into “spheroids” promotes quiescence and cell survival. (4) Transcoelomic spread (metastasis) of spheroids, via circulation of peritoneal fluids, to the omentum and other organs and cell layers lining surfaces (mesothelium). (5) Dissociation of spheroid aggregates followed by adhesion, migration, and invasion of the mesothelial layer occurs at these sites to support tumor growth. Recent in vitro studies and mouse models have implicated progesterone’s promotion of the processes indicated in boxes A and B. See text for further explanation.

Mechanisms of Progesterone's Actions

There are 3 main classes of progesterone receptor proteins that mediate the actions of progesterone and its metabolites, including the nuclear PR receptors (nPR) (46, 47); the membrane PR receptors (mPR or PAQRs) (48); and progesterone receptor membrane component 1/2 (PGRMC1/2) [Fig. 3 (49)]. This review will focus predominantly on the nPR proteins, with some reference to the other 2 receptor classes.

Figure 3.

Classes of the progesterone receptor protein family. Three main classes of progesterone receptor (PR) proteins mediate the actions of progestins including: (1) the nuclear PR (nPR) proteins, which reside in the cytoplasm and/or nucleus, binding with progesterone (P4) and other progestins that have diffused through the plasma membrane and interacting with signaling molecules or activating transcription of specific genes; (2) the membrane PRs (mPR/PAQR), which are 7 transmembrane domain proteins that modulate cAMP/MAPK/PI3K pathways mediating anti-apoptotic, proliferative effects; and (3) the recruited membrane receptors PGRMC1/2, which can bind with Heme receptors and interact with CYP proteins (Cyp P450), activating Ca signaling and mediating pleiotropic effects such as cholesterol and steroid biosynthesis. Evidence supports crosstalk between these classes of PR (dashed blue line).

The nuclear PR receptors nPRA and nPRB are transcribed from the same gene (PGR) possessing 2 different promoters; nPRA, the shorter isoform, excludes 164 amino acids from the N terminus. These receptor proteins contain 3 important functional domains (Fig. 4): a DNA binding domain, a ligand binding domain, and an N-terminal domain. Accessory binding sites (AF1, AF2, and AF3) provide means for complex interactions with numerous accessory proteins and cofactor regulators. AF3 is only present in nPRB, which may explain some of the differences observed between the 2 isoforms (50-52). Unliganded nPRA primarily localizes to the nucleus whereas unliganded nPRB localizes to both the nucleus and the cell membrane or cytoplasm (53). In the canonical steroid receptor signaling pathway, ligand binding induces nPR dimerization and a conformational change, which facilitates nuclear translocation/retention where nPRs interact with DNA promoter and enhancer regions. Morphologically, ligand-bound nPR protein at active sites of transcription can be visualized in the nucleus as dark punctate foci (54). However, unliganded nPR protein can also interact with DNA (55), and gene expression analysis in breast and ovarian cancer cell models indicate that unliganded nPRs regulate numerous genes (56-59). Furthermore, nPRB is involved in rapid signaling events (3-5 minutes) at or near the membrane in association with ERɑ and c-src and can initiate activation of Ras/Raf/MEK/MAPK and P13K/AKT/mTOR/S6 kinase signaling cascades upon ligand binding; these signaling activities are independent but highly integrated with nPRB's action as a nuclear transcription factor (47). Volumes of research have revealed the complexity of nPR signaling by defining multiple layers of regulation, including more than 300 coreceptors and post-translational modifications such as phosphorylation, sumoylation, acetylation, and ubiquitinylation (60). Membrane and cytoplasmic activities of nPR mediate highly localized signaling cascade activation that then informs nPR promoter selection and gene regulation in the nucleus followed by changes in cell biology (61). For example, one model proposes that unliganded nPR binds to a subset of nPR regulated genes and retains the nucleosome inhibitory complex, HP1-LSD1, to keep these genes silenced through interactions with histone H3 (H3K9me3). Upon hormone stimulation, rapid nPR kinase signaling cascades such as ERK1/2 trigger downstream MSK1 phosphorylation, eventual phosphorylation of H3K9me3, and eviction of the nucleosome inhibitory complex. The opened nucleosome then allows access for ligand bound phospho-nPR/coactivator complexes to interact with nPR responsive elements in promoter regions to activate transcription (55).

Figure 4.

Structural properties of the nuclear progesterone receptor (nPR) isoforms. The predominant nPR proteins are transcribed from separate transcriptional start sites present within the single PGR gene. Both isoforms contain key functional domains: a DNA binding domain (DBD; white box) that binds to progesterone (P4) response elements on DNA, a ligand binding domain (LBD; red box) that binds the hormone, and an N-terminal domain (blue, yellow box). The DBD and LBD are connected by a hinge region. The shorter isoform, nPRA, lacks 1–164 amino acids at the N-terminus (yellow box) that is present on the longer nPRB. Accessory binding sites are present in the C-terminus of both isoforms (AF1; pink box) and the N-terminus of both isoforms (AF2; blue box) or exclusively in nPRB (AF3; yellow box). These sites support complex formation with additional accessory proteins and cofactors.

Additional, widely expressed receptor families are known to bind progesterone. The membrane PR receptors, mPRα, mPRβ, mPRγ, mPRδ, and mPRε (also named PAQR7, PAQR5, PAQR8, PAQR6, and PAQR9), are 7 transmembrane proteins with structural similarity to G-protein coupled receptors that demonstrate high progesterone binding affinity (62-64). mPRs have been identified in female mammalian reproductive tissues and are implicated in oocyte maturation (63), follicle development (65), and muscle contraction during parturition (66). While numerous studies in normal reproductive tissues have shown that progesterone can induce G protein signal transduction pathways through mPRα interactions with both Gi and Gβγ, to decrease or increase cAMP, respectively (65), there is still some debate as to whether mPRs act as legitimate G-protein coupled receptors (67). In human ovarian cancer cell lines, SKOV3 and ES2, which express high levels of mPRs but not nPRs (68, 69), progesterone alone did not repress or activate cAMP. However, coadministration of progesterone with isoproteronol, a β1–2 adrenergic receptor agonist did increase cAMP signaling (69).

A third class of progesterone receptors, called progesterone receptor membrane component 1/2 (PGRMC1 and PGRMC2), were isolated from membrane fractions having progesterone binding abilities (70). These proteins also bind to heme and cytochrome P450 proteins, perhaps explaining their localization to the endoplasmic reticulum and functions in chemotherapy resistance. However, the ability of PGRMC1 to bind progesterone has been debated, and interesting studies have suggested PGRMC1 may be more appropriately called an adaptor protein, as overexpression is able to stabilize mPRα at the membrane, and therefore mPR rather than PGRMC1 may be the protein responsible for progesterone binding and signaling in some studies (71). Still, biochemical evidence has shown that PGRMC1 can bind to progesterone (72). Detailed reviews of PGRMC1 and its pleiotropic functions in cancer progression are provided by Cahill et al (73). and others (74) and include discussion of mitosis and cell cycle entry, migration, stemness, apoptosis, and chemotherapy resistance. Separating progesterone-dependent functions (via progesterone/mPRα binding or direct progesterone binding) from progesterone-independent functions requires more study.

Historical Perspective on Progesterone's Actions in OC

Sex Steroid Synthesis

Since the isolation of estrogen, progesterone, and their metabolites in the 1930s, the idea that these ovarian steroids might modulate the development of cancers of endocrine-regulated “reproductive” tissues (such as breast, uterus, cervix, and ovary) has been the focus of basic and clinical research. Numerous studies on ovarian neoplasms over the past 6 decades have shown that epithelial and sex cord-gonadal (nonepithelial) tumors possess the machinery to synthesize and respond to progesterone and other sex steroids. Incubation of tumor homogenates or cystic fluids with labeled substrates like pregnenolone, followed by the detection of progesterone, testosterone, and estradiol, proved the presence of steroidogenic enzyme activity in these ovarian tumors (75, 76). Direct measurement of secreted sex steroids in benign as well as malignant tissues supported these findings (77, 78). Early studies, prior to the discovery of the receptor genes, also revealed the existence of functional high affinity progesterone and estrogen receptors in normal and neoplastic human ovary and fallopian tube tissues, using biochemical receptor binding assays (79-82). Considering the wealth of evidence, it is not surprising that more recent studies have shown the expression of key proteins necessary for progesterone synthesis. In serous and nonserous epithelial OC tumors, the mRNAs encoding the cholesterol transporter, steroidogenic acute regulator protein (StAR), the P450 side-chain cleavage enzyme (P450scc; cholesterol to pregnenolone), and the 3-β hydroxysteroid dehydrogenase enzyme (HSD3B2; pregnenolone to progesterone) are expressed along with their immunoreactive proteins (83). Immunohistochemical studies of low and high grade serous, endometrioid, mucinous, and clear cell tumors revealed the expression of enzymes that convert progesterone to other sex steroids, including CYP17, CYP19 (aromatase), HSD17B1, AKR1C3, observed in both lesions of the epithelium as well as adjacent and distant ovarian stroma (84). Interestingly, benign epithelia of the ovarian surface and those lining cortical inclusion cysts had little to no expression of these markers within their adjacent stroma. The relevance of an activated ovarian stroma as a source and a modulator of steroid production and actions within the epithelia of the tumor is intriguing and suggests an interaction between these cell populations (85, 86). Taken together, these studies support the potential role of the in situ steroid milieu in OC tumor progression.

Receptor Expression

Initial purification of specific nuclear receptor proteins for progesterone (87) as well as estrogen (88) led to the production of antibodies for these steroid receptors (89, 90), now routinely used for the diagnostic determination of their receptor protein expression (91). In addition, the subsequent cloning of the genes encoding nPR (92) and ER (93) was completed. With this knowledge and critical reagents, a plethora of studies have been conducted in attempts to definitively determine the expression of nPR (along with ER) in epithelial OC and establish their value as prognostic biomarkers—such analyses and the associated controversy continues to this day. In general, immunoreactive nPR and ER protein have been observed in all epithelial OC subtypes; varying levels of percentage positivity rates have been reported using single tissue blocks or microarrays of predominantly primary tumors from advanced stage disease. Endometrioid (EOC) and serous (HG/LGSOC) OC consistently show the highest expression of these receptors, ranging from 42% to 90% EOC and 26% to 69% HG/LGSOC for nPR and 39% to 89% EOC and 46% to 77% HG/LGSOC for ER (94-99). In contrast, parallel analyses of MOC and CCOC tumors in these studies reveal much lower expression, with 4% to 16% MOC and 1% to 12% CCOC positive for PR and 3% to 12% MOC and 2% to 6% CCOC positive for ER. The high variability in the observed expression, especially for endometrioid and serous OC, likely reflects biological factors (eg, tumor heterogeneity, patient demographics) and technical issues (eg, fixation, antigen retrieval, antibodies). Few studies have examined the expression of these receptors in the early stages of epithelial OC initiation, partially due to uncertainty as to the tissues of origin for these subtypes as well as the availability of such tissues (100, 101). Interestingly, the p53 mutant fallopian tube epithelial cells within early precursor STIC lesions express nPR and ER, with 100% and 73% positivity for these receptors, respectively (59, 102). It is significant that these cells express the activated nPR, phosphorylated on Ser294 and located in nuclear foci, suggesting functional nPR signaling (59). In addition, STIC lesions appear to express other steroid receptor proteins, including the androgen receptor and the glucocorticoid receptor [GR (102); Mauro and Lange, unpublished observations].

Overall, positive protein expression of nPR in OC has been correlated with a better outcome (103), consistent with progesterone's perceived protective role for these malignancies. Nuclear PR positivity is associated with improved survival for EOC and HGSOC, with no significant association for CCOC, MOC, or LGSOC (98, 103). It has been suggested that nPR expression declines and/or is lost during the evolution of the disease, from high expression in normal OSE or FTE to lower expression in advanced OC. But such comparisons are often made across studies, not based on sampling early to late stages of the disease from single patient cases, and therefore are subject to the variability of collection, processing, and interpretation of these tissues. With regard to expression of nPR isoforms, 3 studies that discerned between nPRA and nPRB isoforms in OC saw a prevalence of nPRB over nPRA (104-106). This is opposite to what has been observed in breast cancer where nPRA is typically the more prevalent isoform, resulting in a high nPRA:nPRB ratio (>1) (107). In breast cancer patients, excessive nPRA has been correlated with lower disease-free survival and tamoxifen resistance (108), and in breast cancer cell models, relative to nPRB, nPRA has been shown to be a more potent driver of cancer cell stemness, as measured by secondary tumorsphere assays (109). Other studies have shown that luminal A or B tumors with higher nPRB expression (eg, nPRA:nPRB ≤0.83) correlates with shorter relapse-free and metastasis-free survival (110). These results in breast cancers contrast with our observations in early OC FTE models, where nPRA promotes cellular proliferation (59), suggesting that the roles of nPRA and nPRB in OC are reversed compared to breast cancer. This skewed expression of the isoforms is not normally detected in healthy tissues where the observed protein expression of nPRA and nPRB is usually equivalent, giving a nPRA:nPRB ratio of 1 (111). It is important to note that such conclusions regarding nPRA vs nPRB expression need to be evaluated in light of questionable antibody specificity; most nPR antibodies cannot distinguish between nPR isoforms when tested in cells overexpressing nPRA or nPRB using Western blot or immunohistochemical analyses. Recent studies have taken a more systemic approach, utilizing an antibody (1294) that detects both isoforms and another that detects predominantly nPRB (250H11), looking at normal human tissues to substantiate differences in nPRA and nPRB expression (112). In the future, additional studies like this, along with the development of more specific antibodies and the use of biomarkers of activated nPR signaling, will be helpful in discerning the prevalence of these isoforms in ovarian tumors.

The potential relevance of other classes of PR receptors (besides nPRs) in cancers has only recently gained attention, and their expression data in OC is still very limited. The membrane PRs (mPRs or PARQs) are expressed, and sometimes overexpressed, in ovarian carcinomas of different subtypes; however, no positive or negative correlation between mPRs and nPRs has been established (113). For the membrane recruited class, PGRMC1 generally appears to be expressed ubiquitously, is overexpressed in ovarian cancer (74), and sometimes presents with intense nuclear localization (114). In advanced ovarian cancers, nPR expression may disappear while PGRMC1 expression increases (115, 116), and high PGRMC1 expression correlates with poor overall survival (115). In a SKOV-3 OC mouse xenograft model, depletion of PGRMC1 resulted in fewer tumors, and those tumors that did develop were smaller and had fewer apoptotic cells, decreased microvascular development, and increased expression of genes associated with hypoxia (HIF1 and ARNT) (116). However, these cancers were also more resistant to cisplatin chemotherapy, a paradox since in vitro studies have suggested that PGRMC1 is at least partially responsible for paclitaxel and cisplatin resistance (114).

Progestins as Cancer Therapy

Clinical reports from the late 1950s discussed the use of compounds like hydroxyprogesterone capronate [Primolut Depot (117)] for the treatment of malignant tumors of the ovary. This early therapeutic use of progestins was justified based on their known antagonistic effects on estrogen-induced proliferation and their low toxicity as well as their inclusion, at that time, in treatment regimens for other hormone-dependent cancers (eg, endometrial and breast). Clinical trials employing the common nPR agonists medroxyprogesterone acetate or megestrol acetate were unconvincing with observable patient response or stabilization of disease averaging 10% to 15%, suggesting that such treatments had no significant benefit [patient cohort for (119) = 75] (118, 119). Both medroxyprogesterone acetate and megestrol acetate were also tested in combination with estrogens [patient cohort n = 65 (120)], anti-estrogens [eg, tamoxifen, patient cohort n = 33; (121)], or chemotherapeutics [eg, melphalan or cisplatin, patient cohort n = 82; (122)], but these regimens still produced low objective patient responses and no improvement in progression-free survival. Overall, many of these progestin treatment studies were plagued by small cohort numbers, poorly defined response parameters, little to no analysis of nPR tumor expression, a lack of biomarkers verifying progestin signaling, and heavily treated patient populations with recurrent or refractory OC. The utilization of hormonal therapy for OC during subsequent decades has been minimal, with standard treatment protocols dominated by the use of established chemotherapeutics (eg, carboplatin, paclitaxel) as well as newly approved drugs targeting DNA damage repair (eg, PARP inhibitors) and emerging immunotherapies. A recently established clinical trial [Memorial Sloan Kettering Cancer (123)] is testing the efficacy of a novel nPR antagonist, onapristone, in patients suffering from nPR-expressing (>1% by immunohistochemistry) low grade serous ovarian/primary peritoneal cancer, examining response rates within 36 weeks of treatment. Preliminary results with a small, recruited cohort of 14 patients diagnosed with granulosa cell OC and 4 patients diagnosed with LGSOC has shown little objective responses as yet (124). It is hoped that this new approach, where tumors are screened for nPR expression, may provide actionable insight into the use of progestin therapies for some forms of epithelial OC.

Reproductive Factors and OC Risk

By the late 1970s to mid-1980s, the possibility that exogenous hormone use and certain reproductive states could modulate OC risk was emerging (125). Reproductive states like pregnancy have been shown in many studies to be associated with reduced OC risk, with multiparous individuals carrying full-term pregnancies showing the greatest risk reduction. The mechanism for this reduction has been attributed to such factors as elevated endogenous progesterone acting as the protective steroid; reduced cumulative ovulations [eg, lifetime ovulatory years (126)]; and/or the milieu of other hormones including estrogen, gonadotropins, and other placental hormones (22, 127, 128). However, several observations diminish the support for these theories, including that (1) pre-term (≤37 weeks) or post-term (≥42 weeks) births are often associated with increased OC risk [eg, ∼2-fold increase (128, 129)], (2) risk reduction is not consistent across all epithelial subtypes (130), and (3) higher circulating levels of progesterone exhibited during pregnancy in some individuals is not associated with shifts in risk for any epithelial subtype (127). Interestingly, a prospective study of more than 1 million UK patients observed that nulliparous women had no significant increase in serous tumors yet exhibited a greater risk for endometrioid and clear-cell tumors, both of which are thought to originate from endometriosis (130).

For reproductive states such as lactation, breastfeeding also appears to reduce OC risk, with up to a 24% reduction for all subtypes and in particular for serous and endometrioid OC (131). The mechanism, like pregnancy, is thought to be due in part to reduced lifetime ovulatory years and/or hormonal milieu. Hormonal profiles during lactation are distinct from that of pregnancy with circulating levels of placental progestins and other hormones having declined rapidly post-partum and remained low during this period (132). Hormones, like prolactin, which are critical for maintenance of galactopoiesis (ie, maintenance of lactation), increase during lactation and the hypothalamic-pituitary-ovary axis is inhibited, resulting in little to no follicular development or subsequent ovulations prior to introduction of solids to the infant. The connection between prolactin and OC risk reduction has been examined, but most studies suggest that enhanced prolactin signaling actually increases tumor aggressiveness and is associated with poorer patient survival (133, 134). In addition, the more significant benefit of breastfeeding appears to occur past the time of anovulation, suggesting that the protective effect could be due to long-term modulation of inflammation, immune response, and metabolism (131). Therefore, in light of these studies, it is unlikely that, in these reproductive states, progesterone is the sole or most significant protective factor that reduces OC risk.

Exogenous Hormones and OC Risk

The use of exogenous hormones for contraception or hormone replacement therapy (HRT) has consistently been associated with the modulation of OC risk. Published studies as early as the late 1970s revealed a 30% to 40% reduction in risk with the use of oral contraceptives [OCP (135)]. Such earlier studies often had small patient cohorts with little power to examine effects of age and OC subtype and were based on the use of older contraceptives with higher doses of the ovarian steroids and different synthetic analogs. More recent prospective and retrospective studies still support that ever use, recent use, or current use of contemporary combined hormonal contraceptives containing reduced estrogens and newer progestins (eg, desogestrel, gestodene, drospirenone) is associated with reduced risk, especially with longer duration of use (23, 136). Risk reduction is strongest for the endometrioid and serous subtypes, with conflicting results on risk for MOC and CCOC (24, 136). Interestingly, the ever use of progestin-only products including oral and intrauterine devices containing second- and third-generation progestins is not always associated with an OC risk reduction (136-138). Likewise, the ability of injectables and implants containing progestins to reduce OC risk is inconclusive (139). Such observations call into question whether it is the progestins in OCPs that are the main factors responsible for reducing OC risk.

In contrast to contraceptives, HRT use is associated with an increased risk of OC. Early HRT regimens used solely estrogens to ameliorate the physiological consequences of menopause until studies revealed the dramatic increase in endometrial cancer risk (140). Following the shift to estrogen-progestin based therapy, an increased risk of cardiovascular disease and breast cancer were subsequently observed during the Women's Health Initiative trials (141). These trials focused on short-term use and short-term effects and had lower power to assess the risk of rare, late endpoints like OC. In the past 2 decades, several investigators have found that increased OC risk is consistently associated with increased duration of HRT use, especially for ≥5-year treatment periods, but less than 5 years current or past use did not increase risk (23, 142). Increased risk with longer use is greater for the serous subtype as compared to the endometrioid, mucinous, and clear cell subtypes (143). Interestingly, in the Million Women Study, no significant difference in risk was observed for estrogen-only vs estrogen-progestin preparations (23). An important question remains: How can similar hormonal preparations have such distinctly opposite effects on OC risk in premenopausal vs postmenopausal women?

Genetic Factors and OC Risk

Germline mutations in genes involved in DNA repair, such as BRCA1 and BRCA2 genes, increase the lifetime risk of epithelial OC in women (144). Possible connections between the sex steroids, BRCA mutations, and disease initiation and progression are poorly understood. BRCA1/2 mutation carriers tend to have more aggressive, HGSOC disease, which has strikingly similar genomic properties to triple-negative breast cancer (145). Interestingly, such carriers exhibited dysregulation of sex steroid synthesis/metabolism during the normal menstrual cycle (146, 147). Elevated serum estrogen and progesterone levels were observed in healthy carriers and were associated with increased follicular uterine endometrial thickness and decreased luteal uterine endometrial thickness relative to noncarriers, supporting a physiological effect of the skewed hormonal secretions. Analysis of fallopian tube tissue has shown that the gene expression profile of normal fallopian tube BRCA1/2 mutant tissue is most similar to HGSOC tumors only if sampled during the luteal phase when circulating progesterone is highest (34). In these tissues, the number of genes regulated between the follicular and luteal phase was much greater in carriers. Such dysregulation has been observed in BRCA1 carrier breast tissue, which showed increased proliferation, loss of nPRB isoform, and altered nPR target gene expression (148). In addition, human breast organoids generated from BRCA1 carriers exhibited distinct gene expression profiles as compared to noncarrier organoids following treatment with the nPR modulator telapristone acetate (149). The intriguing connection between nPR signaling and BRCA1/2 protein function emerging from these studies is supported by molecular investigations showing that BRCA1 can modulate nPR expression and transcriptional activity (150, 151) and that progestins can drive the recruitment of ER/nPR complexes to BRCA1 DNA binding sites (152). Overall, this evidence suggests that sex steroid signaling, in particular nPR signaling, is dysregulated in the context of a BRCA1/2 deficiency and therefore may promote disease initiation and/or progression.

Progestins in Cell Models of OC

The mechanisms responsible for progesterone's putative role as a protective steroid are thought to be related to its inhibition of ovulation, predominantly by negative feedback on the hypothalamic-pituitary axis, and its ability to “clear” or “exfoliate” damaged cells of the OSE that could become cancerous. Once the FTE was discovered as another cell of origin for the HGSOC subtype, it was proposed that PR signaling may also clear this epithelia of cells that are damaged by factors within the follicular fluid during each ovulation (153). To assess how PR signaling would mediate such effects, most in vitro studies have focused on the ability of progestins to modulate cell proliferation and cell death (eg, apoptosis, necroptosis, autophagy), utilizing primary/immortalized cultures of normal OSE cells or OC tumors or established ovarian carcinoma cell lines. Evidence from breast cancer models supports the idea that progesterone, acting through nPRs, serves as a modulator of cell fate, sensing inputs from sex hormone and growth factor-initiated signals via activation of multiple downstream kinases (eg, PI3K/Akt and Ras/Raf/MAPKs) as well as input from other steroid receptors such as estrogen receptors (ERα) (154). Through post-translational modifications of the nPR protein and tight coupling with cell cycle proteins, progesterone can dictate cell cycle progression and DNA damage repair (46, 155, 156). Unfortunately, in many of the OC cell models used to explore such signaling, researchers did not verify that the cells expressed nPRs or other membrane PRs, and they often used progestin treatments at nonphysiological levels or growth conditions that did not eliminate nonspecific steroid effects (eg, phenol red, whole serum). This is critical since members of the nuclear receptor family 3, including nPR, ER, androgen receptor, GR, and mineralocorticoid receptor, have similar molecular properties and can exhibit agonist-antagonist promiscuity (157). This is especially true with nPR and GR, where high concentrations of progesterone may engage GR, thereby reprograming transcriptional activity and downstreaming signaling pathways (158). In addition, progesterone and some synthetic progestins can bind to mPRs and PGRMCs.

In those studies where OC cell models were characterized and well controlled, exposure of OC cell lines to progestins for 3 to 5 days resulted in decreased cell viability and cell numbers supporting the anti-proliferative actions of this hormone in OC cell models (159, 160). Progesterone also appeared to induce apoptosis, with increased caspase-3 activation after 72 hours of treatment and increased p53 and BAX protein and decreased BCL-2 after 24 hours (159). Studies from the Lange lab revealed that signaling through nPRB, in the presence of the nuclear receptor specific progestin R5020 (10 nM, 96 hours), can induce senescence in ES2 and PEO4 advanced OC cell lines via Forkhead-box transcription factor 1 (FOXO1)-dependent upregulation of p21 and p16 (95).

In one of the few cell models of early HGSOC studied, progesterone also promoted necroptosis through the TNFα-RIPK pathway in an immortalized p53-defective FTE model, potentially mediating the clearance of defective cells in the mouse oviduct (161). This cell model exhibited little to no nPR expression, and high concentrations of progesterone (100 μM) were utilized for the experiments. A recent study, utilizing precancerous FTE cell lines and established ovarian carcinoma cells along with intrabursal and intraperitoneal mouse OC tumor models, concluded that progesterone can indirectly mediate cell death via pyroptosis and thereby prevent HGSOC (162). But, again, the expression of nPRs in the cell lines or the primary fibroblasts utilized for these experiments was not verified. Overall, when considering these cellular effects of PR signaling related to the initiation and progression of the disease, it is important to recognize that nPR can also activate unique transcriptional pathways in the absence of progesterone or synthetic progestins; this has been shown by Diep et al in subsequent senescence studies (57) and in early cell models of HGSOC discussed later (59), as well as in breast carcinoma models (163, 164).

The role of PR signaling in cell migration and invasion is also relevant to the processes of OC dissemination within the peritoneal cavity. Progestins have been shown, in multiple breast cancer cell models, to enhance migration and invasion through stabilization of the RhoA complex, modulation of focal adhesions, and transcriptional regulation of key genes in these processes (165-167). Progesterone, allopregnanolone, and mifepristone (RU486) have been shown to increase migration in OC lines (168, 169), but an understanding of the mechanisms involved in OC is limited. The p53 mutations often observed in early SCOUT/STIC lesions can promote enhanced cell adhesion and mesothelial invasion in immortalized “normal” human FTE cell models (170). Interestingly, studies have shown that several p53 mutant species, including the frequent mutation observed in OC, p53 R175H, can disrupt the progesterone-activated nPRA/p53 complexes that regulate p27 expression (171). In addition, an interaction between PR activation, p53 expression, and their subsequent downstream transcriptional effects has been reported (171, 172). It remains to be proven if mutant p53 species in FTE can synergize with progestins to support cell shedding and peritoneal dissemination during the early stages of OC progression. Likewise, progestin-mediated modulation of mesothelial adhesion and invasion requires further study.

Updating Our Perspective on Progesterone's Role

Recent studies using in vitro cell and mouse models provide compelling evidence that progesterone's actions through nPR signaling may promote the initiation and progression of various types of OC. DeMayo and colleagues (173) were interested in exploring the interplay between the nPR isoforms and the importance of a shifted abundance of each in the development of female reproductive cancers. They employed transgenic mice overexpressing the nPRA or nPRB isoforms in tissues that normally express nPRs and observed ovarian neoplasms at 23 weeks of age, with nPRB-overexpressing mice showing 100% penetrance as compared to ∼36% for nPRA + mice. Histopathology and bioassays suggested that the neoplasms occurred within regressing corpora luteum. Transcriptomic analyses revealed that the gene signatures were pro-proliferative with PI3K-AKT pathway enrichment and gene targets involved in cell cycle and DNA recombination. In addition, these signatures were significantly similar to human endometrial and ovarian cancer signatures, in particular clear cell, serous, and mucinous OC subtypes. Although these tumors most likely arose from dysplastic granulosa cells, the resulting ovarian neoplasms progressed without the introduction of genetic mutations, a common approach for the development of genetically engineered mouse models of OC (174). Therefore, these observations strongly suggest that signaling through nPRs can activate pathways that promote tumor development.

Our current insight into the OSE and the FTE as primary cells of origin, and the knowledge that mutations in genes such as p53, BRCA1/2, and PTEN can be critical, has helped drive the creation of multiple useful genetically engineered mouse models of HGSOC. Kim and colleagues originally developed a Dicer1-Pten double knockout model in reproductive tissues, observing 100% penetrance of metastatic HGSOC, arising from the fallopian tube and spreading to ovary, omentum, diaphragm, and peritoneal surfaces, with histopathological and genetic similarities to human disease (175). They utilized this model to address whether ovarian progesterone was a relevant hormonal factor in the development of HGSOC (176). Removal of the ovaries in these mice did not prevent primary fallopian tube tumors but significantly prolonged survival and almost completely abolished peritoneal metastasis, and the cystic, fibrotic primary tumors had little evidence of HGSOC. Treatment with progesterone in these ovariectomized mice reversed the ovariectomized phenotype—100% penetrance of disease with fallopian tube tumors, peritoneal metastasis, and ascites, suggesting that this steroid could drive development of HGSOC. Treatment with estrogen (E2) and progesterone revealed that E2 could block progesterone-induced disease progression, with reduced metastasis and heterogeneous, non-HGSOC tumors, intriguing results that might address the effects of combination OCPs. Mifepristone, the nPR antagonist, was effective at suppressing HGSOC development and progression and enhancing survival, but due to its anti-glucocorticoid activity, genetic inactivation of nPR was employed. Similar to the mifepristone and ovariectomy effects, deletion of the PGR in the original Dicer1-Pten double knockout intact mouse inhibited HGSOC development and reduced peritoneal metastasis. Interestingly, transcriptomic analyses of early HGSOC tumors within the fallopian tube of the progesterone-treated ovariectomized mice revealed dysregulated pathways for cell cycle, DNA damage/repair, and BRCA1 signaling. Overall, this study and the nPR isoform overexpression mouse models described previously support the hypothesis that progesterone is a relevant endogenous hormone that can promote the development and metastasis of HGSOC and that the molecular mechanisms of this activated nPR signaling may be related to its interactions with cell cycle progression, DNA damage/repair, and BRCA1.

The development of cell models of early HGSOC originating from the fallopian tube has been hampered by the decades-old difficulties of creating epithelial cell models that retain ovarian steroid receptor expression. Immortalization of primary cells and 2-dimensional culturing invariably leads to loss of endogenous receptor expression. In an effort to more accurately explore the molecular mechanisms involved in nPR signaling in the initiation and progression of HGSOC, our laboratory developed novel nPRA and nPRB-expressing human FTE models originally genetically modified with hTERT and p53 R175H transgenes (59). Initially, since little is known about the expression of activated phosphospecies of nPRs, immunohistochemical analyses of patient tissues diagnosed with HGSOC was conducted, revealing that activated nPR (phosphorylated S294 protein) is robustly expressed in both STICs and invasive HGSOC tissues, exhibiting punctate nuclear foci—a phenotype indicative of active transcriptional complexes. Utilizing our novel cell models, studies were performed to determine how activated nPRs would influence cell behaviors that support disease initiation and progression. Expression of unliganded nPR led to greater proliferation and migratory behavior for nPRA as compared to nPRB, whereas reversible cell cycle arrest, with predominance in G0/G1, was observed for both isoforms in the presence of hormone. The quiescence state that was induced by expression of either isoform as well as presence of a progestin was essential for spheroid formation, along with the spheroid dissociation and subsequent migration in a Type I collagen invasion assay, a matrix protein thought to be important in mesothelial invasion (177). Transcriptomic analyses revealed that nPR signaling promotes quiescence through the regulation of DREAM [Dimerization partner, RB-like, E2F and Multi-vulval class B], a transcriptional protein complex that promotes arrest through repression of cell cycle dependent genes (59) Activation of nPR by the progestin, R5020, led to increased DREAM protein complex formation and increased recruitment of DREAM proteins (including nPR) to the regulatory elements of DREAM target genes. Nuclear PR recruitment to these genes also resulted in the regulation of the expression of critical DREAM and B-MYB/MMB complex proteins, which, in concert with enhanced DREAM complex formation, led to potent repression of cell cycle gene sets and the observed block in cell cycle progression. Taken together, these studies suggest a model wherein nPR signaling promotes changes in FTE cell fate that could enable early stages of HGSOC progression. The actions of progestins in these published works as well as those discussed in previous sections are summarized in Table 2.

Table 2.

Actions of progestins in ovarian cancer models

Action	Cell modelsa	Mouse modelsb	Descriptionc	Comments	Ref(s)d
Adhesion	hFTE	n/a	Progestins enhance cell adhesion and spheroid formation	hFTE nPRA and nPRB expressing lines	(59)
Apoptosis	hOSE, OC	n/a	Decreased proliferation, enhanced cell death (apoptosis)	nPR expression proven only in ref 161	(159–161)
Dormancy	hFTE	n/a	Progestin drives G0 cell cycle arrest (quiescence) via regulation of DREAM formation, DREAM proteins gene expression	hFTE nPRA and nPRB expressing lines; distinct isoform effects -/+ P4 observed	(59)
Invasion	hFTE	n/a	G0 arrest associated with enhanced invasion of Type I collagen matrix	hFTE nPRA and nPRB expressing lines	(59)
Migration	OC, primary tumors	Inhibin α/Tag (SV40)	Progesterone (P4) and mifepristone drive migration, proliferation, tumor growth via membrane PR, PGRMC1	No nPR expression but high PGRMC1	(169)
Necroptosis	hFTE	p53 knockout (k/o)	P4 promotes necroptosis through the TNFα−RIPK pathway in p53-defective background	Little/no nPR expression in cell lines; high P4 concentrations used	(161)
Pyroptosis	hFTE, OC	Intrabursal/intraperitoneal OC models	Cell death (pyroptosis) progestin-induced via stromal fibroblasts	nPR expression not proven in lines or tumors	(162)
Senescence	OC, primary tumors	n/a	Progestins drive senescence through FOXO1-p21 signaling	ES-2 nPRA and nPRB expressing lines	(57, 95)
Tumor Initiation and Metastasis I	n/a	nPR-A, PR-B knockin	nPR-B overexpression leads to ∼100% penetrance of ovarian neoplasms	Little effect with nPRA; nPRB gene signature similar to OC and endometrial cancers	(173)
Tumor Initiation and Metastasis II	n/a	PTEN, Dicer1 knockout; Pgr knockout	P4 drives FT tumors and HGSOC mets in ovariectomized PTEN/Dicer1 mice; deletion of nPR gene (PGR) blocks HGSOC	Estrogens may antagonize the pro-tumorigenic effects of P4	(175, 176)
Tumor Progression	SKOV-3	Xenografts in athymic nude mice	Observed P4-induced apoptosis and reduced tumor growth depends on PGRMC1 not nPRs	SKOV-3 cells, like many OC cell lines, do not express nPRs	(116)

Abbreviations: hFTE, human fallopian tube epithelium; HGSOC, high-grade serous ovarian cancer; hOSE, human ovarian surface epithelium; nPR, nuclear progesterone receptor; OC, ovarian cancer.

Cell models: hOSE = human ovarian surface epithelial lines; hFTE = human fallopian tube epithelial lines; OC = ovarian carcinoma lines; primary tumors = cultures/explants of OC tumors. See references for specific information on lines/tissues.

Mouse models: description of mouse models used. n/a = not applicable to these studies. intrabursal model = cell injection within sac enclosing ovary/fallopian tube. See references for specific information on these transgenic models.

Description: membrane PRs = includes PGRMC1, progesterone receptor membrane component 1. TNFα-RIPK pathway = tumor necrosis factor-receptor-interacting protein cell death pathway. FOXO1-p21 signaling = signaling involving Forkhead box transcription factor FOXO1 and p21, a cyclin-dependent kinase inhibitor 1.

Ref(s) are article(s) describing these studies.

These recent studies add critical information to the model of HGSOC disease. FTE expressing early mutations as well as activated nPR proteins may enter quiescence, setting the stage for pro-survival mechanisms that would support heightened and unique nPR signaling and potentially promoting DNA damage and further mutations (Fig. 2, box A). This state of dormancy would provide resistance to anoikis following exfoliation from lesions and promote spheroid formation within the peritoneal cavity (Fig. 2, box B). The dissociation and interactions with the mesothelial layers would also be supported by nPR signaling, which modulates adhesion and collagen invasion (Fig. 2, box B). Previous published work and this recent evidence also suggest an intriguing interaction between BRCA1/2 signaling and nPR, which could lead to dysfunctional nPR signaling with BRCA1/2 deficiency.

Conclusions

The actions of the ovarian steroids estrogen and progesterone are highly contextual and complex in both healthy and diseased states—dependent on the target tissue, the cell type, the myriad of receptors expressed, and the hormonal milieu at the cellular and systems levels. To accurately understand the role of progesterone in the initiation and progression of ovarian cancer, in particular HGSOC, we must acknowledge the complexities of nPR signaling itself as well as its interactions with signaling through other steroid receptors, such as ER and GR, that could be relevant to the disease process. First, the recognition that nPR has actions in both an unliganded and liganded state (ie, in the presence and absence of progestins) would broaden the restrictive view that declining levels of progestins during perimenopause/menopause negates any relevant influence of this hormone. Second, the fact that progestins can act through nPR as well as other classes of PRs supports the need to establish additional biomarkers for nPR signaling, ones that more accurately reflect the status of functional signaling rather than a static measure of solely nPR protein. Such biomarkers may alleviate some of the controversy around the true expression of these receptors and give more insight into their importance in disease progression. It is likely that during the lifespan of an HGSOC tumor, for example, from STIC development 10 to 15 years before diagnosis up to the point of tumors becoming aggressive (stage III and IV), the repertoire of PR receptors (nPRA:nPRB, mPRs, and PGRMCs) may change significantly. This shift, along with changes in the expression and function of other steroid receptors, such as ER and especially GR, are predicted to profoundly impact the actions of nPR. Finally, the intriguing molecular interplay between BRCA1/2 and nPR expression and transcriptional activity along with the documented elevation of progesterone in BRCA1/2 mutation carriers begs the question: Is BRCA1/2 deficiency associated with dysfunctional nPR signaling that together may promote DNA damage? Exploring these facets of progesterone should help clarify the ultimate role of this hormone as anti-tumorigenic or pro-tumorigenic during the evolution of the collection of diseases grouped together as OC.

Funding

C.A.L. receives support for this research from the Tickle Family Land Grant Endowed Chair Fund, and C.A.L. and L.J.M. have received support from the Minnesota Ovarian Cancer Alliance and the University of Minnesota Office of the Vice President for Research.

Disclosures

All authors have nothing to disclose.