WNT4 Balances Development vs Disease in Gynecologic Tissues and Women’s Health

Abstract

The WNT family of proteins is crucial in numerous developmental pathways and tissue homeostasis. WNT4, in particular, is uniquely implicated in the development of the female phenotype in the fetus, and in the maintenance of müllerian and reproductive tissues. WNT4 dysfunction or dysregulation can drive sex-reversal syndromes, highlighting the key role of WNT4 in sex determination. WNT4 is also critical in gynecologic pathologies later in life, including several cancers, uterine fibroids, endometriosis, and infertility. The role of WNT4 in normal decidualization, implantation, and gestation is being increasingly appreciated, while aberrant activation of WNT4 signaling is being linked both to gynecologic and breast cancers. Notably, single-nucleotide polymorphisms (SNPs) at the WNT4 gene locus are strongly associated with these pathologies and may functionally link estrogen and estrogen receptor signaling to upregulation and activation of WNT4 signaling. Importantly, in each of these developmental and disease states, WNT4 gene expression and downstream WNT4 signaling are regulated and executed by myriad tissue-specific pathways. Here, we review the roles of WNT4 in women’s health with a focus on sex development, and gynecologic and breast pathologies, and our understanding of how WNT4 signaling is controlled in these contexts. Defining WNT4 functions provides a unique opportunity to link sex-specific signaling pathways to women’s health and disease.

Wnt ligands are classically paracrine signaling molecules that regulate the proliferation and self-renewal of cells. Wnt signaling can also confer polarity to cells, giving shape to different tissues and playing major roles in organogenesis. In humans, the family of Wnt ligands consists of 19 genes that encode these secreted glycoproteins. The Wnt ligand WNT4 plays a crucial role in kidney, adrenal, lung, and pituitary organogenesis (discussed in later sections), and WNT4 loss-of-function is lethal in utero or immediately postnatal because of multiorgan dysgenesis (1-3) (OMIM 603490). However, beyond these roles in fetal development, WNT4 plays unique roles in the organogenesis and development of female reproductive tissues, including the uterus, ovaries, and mammary glands (3-7). Accordingly, WNT4 dysfunction is associated with differences in sex development, (2, 3, 8), infertility and changes in gestational duration (7, 9), endocrine pathologies including hyperandrogenism (4, 10-12), and gynecologic and breast malignancies (13, 14). However, the mechanisms by which WNT4 signals in developmental and pathologic settings is only crudely understood, as WNT4 signaling pathways are diverse and context dependent, and WNT4 gene expression is regulated by diverse tissue type- and context-specific pathways. We will review the developmental phenotypes and disease processes associated with WNT4 dysfunction in women’s health, as well as discuss the myriad mechanisms of regulation and execution of WNT4 signaling in these settings.

Materials and Methods

NCBI PubMed, Google Scholar, and Web of Knowledge were used to search for the terms “WNT4” or “WNT-4,” alone or in conjunction with “müllerian,” “sex differentiation,” “fertility,” “cancer,” “endometriosis,” “leiomyoma,” “GWAS,” or “cancer.” Searches were primarily conducted from March 2020 to December 2020.

We use human vs murine or other species’ nomenclature throughout this review based on the context of immediately referenced work, and default to human nomenclature (WNT4/WNT4) for broader discussion.

WNT4 Is Critical in Sex Differentiation During Fetal Development

Observations in clinical and laboratory WNT4-null settings, including human and murine fetuses, and zebrafish models, highlight a unique role for WNT4 in the development of müllerian tissues and gonadal differentiation. Notably, complete WNT4 loss of function is lethal in utero for humans and mice, in large part because of impeded development of the metanephric kidney, making WNT4 function during embryogenesis vital for the survival of the fetus on birth. However, examination both of murine and human WNT4-null embryos at fetal autopsy reveals a wide-ranging phenotype of gonadal and genital differences. In humans, SERKAL syndrome (sex reversal, kidneys, adrenal, and lung dysgenesis syndrome) was originally described in one consanguineous family with homozygous WNT4 mutations (2). Three cases of homozygous WNT4-mutant fetuses (identified following miscarriage or terminated pregnancies due to renal agenesis) exhibited phenotypes consistent with XX disorders of sex development. One fetus with 46,XX karyotype had normal testes on pathologic examination, another had ovotestes and hypospadias, and the third had typical female genitalia (no chromosome analysis was reported from the latter 2 fetuses). Importantly, the WNT4 mutation identified (C341T, causing amino acid change A114V) was associated with decreased WNT4 messenger RNA stability and a resulting WNT4-null state. Similarly, homozygous Wnt4-knockout in mice is lethal postnatally because of renal dysgenesis (1), but homozygous Wnt4-mutant XX mice have ectopic androgen production in the ovaries, oocyte depletion, and regression of the müllerian ducts with presence of wolffian ducts (but typical female external genitalia) (3, 15). The critical role of WNT4 in female sex differentiation is also seen in zebrafish; of note, wnt4a-knockout is not lethal because of differences in renal development, allowing evaluation of phenotypic sex in the wnt4a-null setting at maturity. Kossack et al showed that wnt4a loss-of-function mutations caused predominantly male development (≥ 93% male in 3 homozygous mutant strains vs 49% to 76% male in wild-type counterparts) (5). Mirroring the XX sex development differences with WNT4 loss, aberrantly increased WNT4 signaling is also associated with XY sex development differences. Duplication of chromosome 1p36 (containing WNT4) in XY patients is associated with phenotypes ranging from cryptorchidism to disorders of sex development. For example, one patient with confirmed WNT4 duplication presented with hypospadias, fibrous gonads with rudimentary tubules, remnants both of müllerian and wolffian structures, and ambiguous genitalia (16). In mice, insertion of a human WNT4 transgene in C57B/6 mice did not cause complete sex reversal but was sufficient to disrupt testicular vasculature and testosterone synthesis (17, 18). This signaling balance highlights the critical role of appropriate doses of WNT4 in sex differentiation.

Importantly, these studies support that typical female development and the differentiation of female reproductive tissues require active signaling (at least in part via WNT4) and that the female phenotype is not a default state (19). All fetuses initially develop both müllerian and wolffian ducts, typically retaining one pair while the other pair degenerates. Retaining müllerian vs wolffian structures relies on specific factors, including WNT4, to tip the scale toward one developmental pathway over the other (20, 21). In this context, murine models have shown that Wnt4 is initially expressed in the undifferentiated gonads of both sexes then later maintained in ovaries and downregulated in testes (22, 23). In ovaries, Wnt4 (with Foxl2 and Rspo1) downregulates the crucial male-specific signaling molecule Sox9 (reviewed in [8, 24], including signaling diagrams of the pathways described in this section). Conversely, in testes, Sry drives upregulation of Sox9 and Fgf9, which effectively suppresses Wnt4 signaling, and thus suppresses female differentiation of the gonads (25). Notably, WNT4 is required not only for retention of müllerian structures but also for elongation and cell migration in the müllerian duct (26), highlighting that WNT4 is required throughout the differentiation and development of the female reproductive tract. WNT4 signaling in sex differentiation has been primarily associated with “canonical” or β-catenin–dependent Wnt signaling, that is, activation of β-catenin–mediated gene regulation (18, 27, 28). In this context, constitutively active β-catenin is sufficient to drive sex reversal in XY mice (ie, as with increased WNT4 signaling) (29-31). β-catenin can also partially rescue the male-like phenotype of Wnt4-null XX gonads (18), and WNT4 inactivation conversely blunts the effects of constitutively active β-catenin (32), supporting a positive feedback loop with WNT4/β-catenin in gonadal development. The importance of β-catenin in WNT4-driven sex differentiation is further supported by the shared importance of RSPO1 (33-35), as RSPO1 is a critical Wnt coreceptor (8), and homozygous RSPO1-null mutation is similarly associated with XX sex reversal in humans and mice (35, 36). WNT4 signaling via RSPO1/β-catenin is also required for ovarian follicle development as this pathway is critical for meiosis (37), female germ-cell survival (38, 39), and maturation of ovarian follicles. Wnt4-knockout specifically in granulosa cells led to compromised folliculogenesis, severely reduced female fertility, and premature ovarian failure (7, 40).

Though these observations from WNT4-null settings highlight the key role of WNT4 in female sex differentiation, complete WNT4 loss of function is in general incompatible with viability because of broader organ dysgenesis. However, WNT4 mutations causing only partial loss of function are not lethal and are instead associated with sex-reversal phenotypes that may not be diagnosed until puberty, whereas other modes of WNT4 dysfunction are linked to distinct reproductive and gynecologic pathologies. These diverse phenotypic consequences of WNT4 dysfunction show that WNT4 plays highly context-specific and functionally distinct roles in signaling throughout development, and defining these roles is critical to understanding WNT4 biology.

WNT4 Mutations Are Linked to Virilization and Uterine Agenesis

In contrast to the complete loss of function and lethality caused by WNT4 mutations associated with SERKAL syndrome, several other WNT4 mutations have been linked to endocrine and reproductive dysfunction that may be diagnosed at or after puberty. In 2004, a case report identified an 18-year-old female patient with a WNT4 mutation (coding mutation E226G, updated to E216G in current WNT4 alignment) and associated primary amenorrhea secondary to vaginal and uterine agenesis, ectopic ovaries, and elevated serum androgens (androstenedione, dehydroepiandrosterone sulfate, and testosterone) (12). Similar cases were subsequently reported by Biason-Lauber and colleagues, all with female (XX) patients diagnosed with primary amenorrhea, uterine agenesis, ectopic and/or dystrophic ovaries with rare or absent follicles in one or both ovaries, and elevated androgen concentrations (secondary to aberrant expression of 3β-hydroxysteroid dehydrogenase and 17α-hydroxylase in the ovary) (4, 10, 11). Nonsynonymous coding mutations in WNT4 were identified in all 4 patients (L12P, R83C, E216G, A223T). The phenotype associated with these WNT4 mutations resembles Mayer-Rokitansky-Küster-Hauser syndrome (MRKH), but unlike MRKH is uniquely associated with clinical and biochemical hyperandrogenism, and the müllerian phenotypes are restricted to uterine agenesis. This “atypical MRKH” phenotype associated with WNT4 mutation is now termed Biason-Lauber syndrome (11).

Though the müllerian phenotypes associated with WNT4 mutations in Biason-Lauber syndrome parallel those observed in SERKAL syndrome or Wnt4-knockout mice, the phenotype is distinct in that patients are otherwise healthy (unilateral renal agenesis/dysgenesis has occurred in some patients). This suggests that WNT4 mutations in Biason-Lauber syndrome cause only a partial loss of function or create other WNT4 protein dysfunctions that specifically affects WNT4 signaling in female phenotypic development. Biason-Lauber and colleagues expressed each of the identified WNT4 mutant proteins in ovarian cancer cells, and found that each caused dominant-negative suppression of WNT4-associated signaling (measured by β-catenin signaling) (4, 10-12). However, the mechanism of this dominant negative effect is unknown. Notably, the mutations did not have a common effect on WNT4 secretion or palmitoylation (a fatty acid modification required for secretion), but the R83C mutant was found to have aberrant cellular localization in the cytoplasm (localization of other mutants was not reported) (10). The authors suggested that WNT4 mutations caused improper protein folding, causing trafficking defects. However, recent research from our group suggests that these WNT4 mutations could specifically affect a novel mode of intracellular WNT4 signaling (see discussion on WNT4 in breast cancer in later sections) (41). Future work on how WNT4 mutations differentially affect signaling in gynecologic development (ie, Biason-Lauber syndrome) vs lung, adrenal, and kidneys (ie, SERKAL/WNT4-null) may provide novel insight into tissue-specific mechanisms of WNT4 regulation and WNT4 signaling in development and disease.

Polymorphisms in 1p36 at WNT4 Are Associated With Gynecologic Pathologies

Single-nucleotide polymorphisms (SNPs) in the 1p36 locus of the human genome, encompassing the WNT4 gene, have been associated with increased risk of a range of gynecologic pathologies. Genome-wide association studies (GWAS) have linked SNPs in an approximately 100- to 150-kb region of 1p36.12 with endometriosis (with associated infertility), uterine leiomyomas or fibroids, pelvic organ prolapse, and ovarian cancer. Variant alleles are associated with increased risk of endometriosis (odds ratio [OR] = 1.16-1.36) (42-56)), leiomyoma (OR = 1.12-1.19) (42, 57-60), and ovarian cancer (OR = 1.11) (14, 61), but are protective for prolapse (OR = 0.85) (62) and associated with longer gestational duration (~ 2 days longer gestation per variant allele [9]). In most studies, the index SNP occurs in the WNT4 gene locus (Fig. 1A). Nearly all associations observed with this locus have been confirmed in multiple primary studies and in meta-analyses (Fig. 1B) and have been observed across several distinct ethnic/geographic populations.

Figure 1.

Single-nucleotide polymorphisms (SNPs) at the WNT4 locus are associated with gynecologic pathologies. A, Map of index SNPs identified in primary genome-wide association studies and meta-analyses. B, Conditions and associated index SNPs. Studies listed correspond to (9), (14), and (42-62).

The functional 1p36 locus variant that contributes to risk and drives the biology is not clearly defined, but the fact that the pathologies linked to this locus are largely dependent on estrogen exposure implicates SNP rs3820282 as a putative driver. rs3820282 has been identified as the specific index SNP in several of the noted GWAS studies, and fine-mapping studies have implicated it as the functional SNP linked to endometriosis in particular (63). rs3820282 is in an evolutionarily conserved segment of intron 1 of WNT4 (64), approximately 1.5 kb downstream from the transcription start site. At this site, there are 2 half-estrogen response elements (EREs, ie, half the reverse palindrome recognized by the estrogen receptor [ER]). However, the rs3820282 variant allele converts one half-ERE to a consensus full ERE; Zhang et al confirmed that the variant allele increases ER binding affinity to the sequence in vitro (9). Consistent with gene regulatory activity at this site, the rs3820282 locus is marked by epigenomic features of active enhancers across cell types, specifically monomethylation of histone H3 (63). While the rs3820282 locus may be linked to regulation of other nearby genes (53), we previously linked this locus to cell type-specific ER binding and estrogen regulation of WNT4 gene expression (65). This strongly suggests that the rs3820282 variant allele, imparting increased ER binding affinity, mediates aberrant or increased estrogen-mediated WNT4 expression that drives increased risk of gynecologic pathologies.

Notably, rs3820282 is linked to increased risk of any subtype of epithelial ovarian cancer (OR = 1.11) (14), and in particular the clear cell carcinoma subtype (OCCC; OR = 1.24) (66, 67), which is associated with especially poor outcomes among ovarian cancer subtypes (66, 67). Importantly, in 1000Genomes data, the rs3820282 variant frequency (ie, high ER binding) is approximately 20% in White populations, but increased in Latino populations and up to 55% in Asian populations, paralleling increased OCCC incidence in Asian populations (rate ratio 1.65, Asian v White [68]). This suggests a key role for rs3820282 in ovarian cancer etiology and in health disparities associated with OCCC. The role of WNT4 signaling in ovarian cancer models (discussed later) is consistent with a unique role in OCCC etiology, but this link requires further study.

WNT4 Signaling Is Critical During Uterine Development and Pregnancy

Evidence from murine Wnt4-knockout models and human pathologies (SERKAL and Biason-Lauber syndrome) demonstrate the critical role of WNT4 in early uterine development. WNT4 is also necessary for proper functioning of the uterus/endometrium throughout life, in particular in pregnancy (69). Along with the key roles of WNT4 in ovarian folliculogenesis and maturation discussed earlier, proper control of the execution of WNT4 signaling is critical for fertility. In the endometrium, WNT4 expression is tightly controlled by diverse regulators and executes distinct downstream signaling programs throughout the course of uterine decidualization, implantation, and embryogenesis. Endometrial-specific Wnt4 knockout in mice leads to agenesis of endometrial glands, impairing decidualization and ultimately implantation (69, 70). Low WNT4 expression is associated with recurrent implantation failure (71-73). The signaling pathways regulating and executing WNT4 signaling in this context are actively being defined, but several putative networks have been described (Fig. 2) (74).

Figure 2.

WNT4 signaling pathways in uterine development and pregnancy. Details described in text. Transcription factors shown in purple text. Boxes denote key upstream regulatory factors that are “entry points” into the pathway.

A circuit linking Wnt4, Bmp2, and Foxo1 is critical in decidualization (ie, the structural changes in the endometrium before and during pregnancy). Bmp2 is strongly associated with decidualization and implantation failure, as uterine Bmp2-knockout results in infertile mice via impaired decidualization (75). Bmp2 signaling mediates stromal cell survival through upregulation of Wnt4 expression and activation of β-catenin–dependent Wnt signaling (76). Bmp2 and Wnt4 both upregulate the expression of Foxo1 (76). Whereas Foxo1 is known to induce proapoptotic genes, Wnt4 inhibits this function during decidualization by mediating the cytoplasmic shuttling of Foxo1 (69). Furthermore, Foxo1 strongly upregulates Wnt4, creating a positive feedback loop that enables proper control of decidualization (77). Despite this role in stromal cell survival, Wnt4 expression must be appropriately controlled before and after implantation; Wnt4 continuously decreases in mice during days 1 to 4 of pregnancy (preimplantation) and then dramatically increases on day 5 (implantation) (78). This increase appears to be important for proper implantation, as sustained Wnt4 suppression is associated with implantation failure. In humans, the aberrant overexpression of transcription factor EHD1 in endometrial cells is associated with recurrent implantation failure. EHD1 overexpression leads to implantation failure by directly downregulating WNT4 expression in endometrial tissue (73). Along with the direct inhibition of Wnt4 by EHD1, EHD1 may also regulate Wnt4 signaling through inhibiting cytoskeletal reorganization and a Fzd2 signaling pathway (discussed later). EHD1 expression steadily decreases on decidualization, but another inhibitor of Wnt4 signaling, LEFTY2, increases throughout decidualization. Through downregulation of Ca²⁺ channels, LEFTY2 inhibits the expression of Wnt4, Cox2 and Bmp2, but does not affect Hoxa10 expression (see next paragraph), suggesting these are distinct pathways (79).

The homeobox transcription factor Hoxa10 also modulates Wnt4 expression to subsequently upregulate Prl and Igfbp1 to promote trophoblast growth and invasion, prevent immune rejection, and support angiogenesis (71, 80, 81). Notably, the upstream (Hoxa10 > Wnt4) and downstream (Wnt4 > Igfbp1) components of this pathway have been separately examined, but whether this is a “complete” circuit has not been shown in one system. However, cytoskeleton destabilization and inflammatory mediators have also been identified as regulators of this pathway. Injecting the murine uterus with the microtubule destabilizer nocodazole leads to decreased Wnt4 expression and reduced Wnt4 signaling via Wnt receptor Fzd2, blocking Igfbp1 and Prl induction and resulting in implantation failure (72). In this system, induction of Igfbp1/Prl is downstream of Wnt4/Fz2-driven CaMKIIa signaling (ie, “noncanonical” or β-catenin–independent Wnt signaling). Wnt4-driven CaMKIIa signaling also activates nuclear factor-κB transcriptional activity, upregulating production of prostaglandin E₂ (PGE₂). Notably, though PGE₂ production is downstream of Wnt4 signaling, PGE₂ signaling mediates feedback to Wnt signaling and Wnt4 expression through a potential positive feedback loop with Hoxa10 (81, 82). Consistent with this PGE₂/Wnt4 feedback loop, artificial decidualization driven by PGE₂ is blocked in murine uteri deficient in epithelial Porcn (Porcupine O-acyltransferase) expression (83). Porcn is typically required for Wnt protein secretion and paracrine signaling (discussed later), but Porcn ablation in the uterine epithelium leads to dysregulation of Wnt4 expression (84).

Along with the identification of several aspects of the networks regulating Wnt4 expression and signaling, the specific roles of β-catenin–dependent vs β-catenin–independent signaling in these processes are continuing to be defined. Defining the cell types expressing Wnt4 is also important, as stromal-derived Wnt4 may support uterine development despite inhibition of epithelial-derived Wnts via targeted Porcn knockout (69, 84). The role of WNT4 in the uterus also continues beyond implantation. As noted earlier, the WNT4 SNP rs3820282 is linked to gestational duration with the variant allele associated with longer gestation (9), though the mechanisms at play have yet to be elucidated. Together, these observations indicate that WNT4 signaling is tightly regulated and executed by multiple networks throughout pregnancy that are likely specific to the context of the endometrium.

Distinct WNT4 Signaling Pathways Are Dysregulated in Gynecologic Pathologies

Dysregulation of WNT4 expression or signaling has been associated with oncogenic events in several gynecologic pathologies (Fig. 3). Mutations in the gene encoding transcriptional regulator MED12 are common in uterine fibroids (leiomyomas) and are associated with increased WNT4 expression (85). In vitro studies strongly support that mutant MED12 directly induces WNT4 expression, which subsequently drives cell proliferation through β-catenin–dependent Wnt signaling and activation of the mammalian target of rapamycin (mTOR) pathway (86, 87). In contrast, WNT4 has been linked with cervical cancer via the human papillomavirus E6 oncoprotein, which drives increased WNT4 messenger RNA translation. WNT4 subsequently activates β-catenin–independent signaling through the planar cell polarity pathway and JNK signaling (88, 89).

Figure 3.

WNT4 signaling pathways in gynecologic pathologies. Details are described in the text. Transcription factors shown in purple text. Boxes denote key upstream factors that are critical in pathogenesis. Solid vs dashed lines represent transcriptional vs signaling effects, as in Fig. 2. The mechanisms by which WNT4 controls PI3K signaling in ovarian cancer are unknown, denoted by “?”.

Canonical β-catenin–dependent Wnt signaling is linked to therapy resistance in ovarian cancer (89-91). WNT4 signaling in ovarian cancer has not been extensively studied but is associated with unique downstream signaling via the PI3K/mTOR pathway. In murine models of ovarian cancer driven by Pten knockout, Wnt4 is required for transformed fallopian tube epithelium cells to migrate, invade the ovary, and establish a tumor (92), suggesting that WNT4 is a critical partner of the PI3K/mTOR pathway. Supporting this, we found that in The Cancer Genome Atlas, elevated WNT4 gene expression in primary serous ovarian cancer is associated with increased phosphorylation of PI3K/mTOR pathway target proteins (93). Consistent with the key role of WNT4 in mediating cell survival observed in other contexts, we found that WNT4 expression increased in ovarian tumors in response to neoadjuvant chemotherapy (94). PI3K/mTOR signaling may be downstream of β-catenin–independent “noncanonical” Wnt signaling via WNT4, but our research suggests that this pathway may be mediated by a novel cell-intrinsic pathway for WNT4 signaling (41) (discussed later). We also identified this link between WNT4 signaling and the mTOR pathway driving therapy resistance in a subtype of breast cancer, invasive lobular carcinoma (ILC, discussed later) (65, 93, 95).

Paracrine WNT4 Signaling Mediates Luminal-Myoepithelial Crosstalk in Mammary Gland Development

Paralleling the key roles of WNT4 in the uterus during pregnancy, WNT4 is also a critical mediator of pregnancy-related phenotypes in the mammary gland. In murine models, Wnt4 is unique among Wnt-encoding genes as Wnt4 expression is directly induced by progesterone receptor (PR) signaling. As a PR target, Wnt4 is required to mediate progesterone-induced mammary duct development during early pregnancy. Using mammary gland transplant models, Brisken et al showed both PR^–/– and Wnt4^–/– mammary glands display decreased elongation and side branching in response to pregnancy (6). Notably, while Wnt4 is important for early branching, Wnt4 overexpression is not sufficient to induce branching (96), and Wnt4 knockout does not impair branching later in pregnancy (6). More recent work has confirmed the role of progesterone in regulating Wnt4 expression (97), and showed that PR/Wnt4 signaling acts via luminal-myoepithelial crosstalk in the mammary gland (13, 97-99). Progesterone stimulation of luminal PR-expressing “sensor cells” induces Wnt4 expression and secretion, which subsequently signals to neighboring stem cells in the myoepithelium in a paracrine matter (97) (summarized with pathway diagrams in [13, 97]). Wnt4 likely activates canonical β-catenin–dependent signaling in target mammary gland cells (though some studies suggest this is not the case [96]), which mediates stem cell activation, proliferation, and renewal. Studies have implicated regulation of Notch and JAK/STAT signaling in these effects. In this context, Wnt4 is also linked to the integrity of the basement membrane through the regulation of Adamts18 (98, 100).

Taken together, WNT4 has a clear effect on the stability of the mammary progenitor/stem cell niche. The impairment of reconstitution in serial transplantation of Wnt4-null mammary epithelia vs that of Wnt4 wild-type mammary epithelia supports the essential function of Wnt4 in mammary stem cell maintenance (97, 99). This pathway has been implicated in the protective effects of early parity on the risk of breast cancer development. Pregnancy at an early age is a significant modifiable factor for decreasing one’s risk of breast cancer (101). Persistently decreased expression of WNT4 in mammary glands has been demonstrated in mice and women with early pregnancy, which is thought to result from a decrease in the number of PR-positive luminal cells after pregnancy-mediated cell differentiation (102, 103). Despite these putative links between WNT4 and breast tumorigenesis, studies on this are limited but do not suggest that WNT4 is sufficient to drive breast tumorigenesis. However, we identified a unique role for WNT4 in models of ILC of the breast that is functionally distinct from the roles of WNT4 in mammary gland development.

WNT4 Signals via a Novel Cell-Intrinsic Mechanism in Models of Invasive Lobular Carcinoma and Other Cancers

ILC is the second most common histological subtype of breast cancer, accounting for 10% to 15% of new breast cancer diagnoses, and presents with resistance to therapy that is inconsistent with overall favorable tumor biomarkers (104-106). We reported that in ILC cells, contrasting progesterone-regulation of WNT4 in the mammary gland, WNT4 expression is directly controlled by ERα (65, 95). Notably, in ILC cell lines, ER binds the partial EREs present at the site of SNP rs3820282 (65), supporting the idea that this site facilitates ER control of WNT4 expression. ER-driven WNT4 signaling is critical for the proliferation and survival of ILC cells and also mediates antiestrogen resistance (65, 93). The genetic hallmark of ILC is loss of CDH1/E-cadherin, which is associated with loss of β-catenin protein and the inability to activate canonical β-catenin–dependent Wnt signaling (65, 105). Therefore, the mechanisms of WNT4 signaling in ILC cells are likely mediated by noncanonical pathways. However, we found that in ILC cells, WNT4 did not signal via the paracrine signaling that is thought to be central to Wnt function.

In the current paradigm of paracrine Wnt signaling, Wnt secretion starts when nascent Wnt peptides translocate to the smooth endoplasmic reticulum and are palmitoylated by PORCN (107-109). Palmitoylation permits Wnt interaction with the protein Wntless (WLS), which shepherds the Wnt through the endoplasmic reticulum/Golgi and facilitates secretion (110-112). Extensive data support that PORCN-mediated palmitoylation is critical for Wnt activity (113-116), but much of this work is limited in the Wnt proteins and signaling contexts studied (in part for historical reasons [117, 118]). As we and others examine more Wnt proteins in increasingly diverse contexts, a growing body of research challenges the paradigm that the PORCN requirement is absolute (119-124). We found PORCN ablation did not phenocopy WNT4 knockdown in ILC (ie, WNT4-dependent) cells, paralleling PORCN-independent WNT4 secretion in these cells (41). Surprisingly, though WLS was required for WNT4 secretion, WLS knockdown also did not phenocopy WNT4 knockdown. This suggests secreted WNT4 is not the active signaling molecule in all WNT4-dependent cells, and that WNT4 signaling does not require secretion. Supporting this, WNT4 protein in the cell was not palmitoylated or in the secretory pathway, but instead was a soluble protein in an aqueous cell compartment (ie, intracellular WNT4). We found intracellular WNT4 directly activates cell-intrinsic signaling (Fig. 4A), identifying a previously unappreciated mode of Wnt signaling. Intracellular WNT4 signaling functioned in models of breast cancer, ovarian cancer, and fibrosarcoma. In ILC cells, intracellular WNT4 controls mTOR signaling via downstream target S6 kinase, ultimately regulating mitochondrial function and cell metabolism (Fig. 4B) (93). The mechanisms by which intracellular WNT4 engages in these signaling pathways is unknown but is under active investigation. These observations further demonstrate that WNT4 plays highly context-specific roles that must each be specifically defined to better understand the roles of WNT4 in development and disease.

Figure 4.

Intracellular WNT4 signaling. A, Wnt peptides are typically secreted via PORCN/Wntless (WLS) activity, but WNT4 can signal independent of secretion as a soluble, intracellular protein via receptor proteins to be identified. B, In invasive lobular carcinoma (ILC) cells, intracellular WNT4 signaling regulates mammalian target of rapamycin (mTOR) signaling via downstream S6 kinase, and also regulates cellular metabolism via control of mitochondrial dynamics.

Considerations for Studying WNT4 in Development and Disease

WNT4 dysregulation is associated with a spectrum of sex development and women’s health issues that should be a major consideration in research when considering sex as a biological variable and in clinical studies that use inhibitors of Wnt signaling. Historically, the majority of patients in drug trials (and mice in research studies) have been male (125), limiting our understanding of how new therapies may affect the unique biochemical signaling context of females. While WNT4 inhibitors could be an important treatment for gynecologic cancers and other gynecologic pathologies, inhibiting WNT4 signaling may affect fertility, pregnancy, and other hormone- and WNT4-associated phenotypes such as bone mass (126-128). Studying the regulation and execution of WNT4 signaling both in the female contexts discussed earlier and male contexts (eg, the developing prostate [129]) is critical to understand these pathways and develop WNT4-targeted therapies.

Conclusions

WNT4 is critical for embryonic organogenesis, but later plays a unique role among Wnt proteins in the development and maintenance of müllerian and reproductive tissues. In organogenesis vs gynecologic and reproductive health, WNT4 has distinct, separable signaling functions, as highlighted by the differences between loss of function in WNT4 mutations in SERKAL syndrome vs Biason-Lauber syndrome. Dysregulation of WNT4 expression also primarily manifests as gynecologic and reproductive pathologies. These tissue-specific roles of WNT4 also can be observed via the diverse downstream pathways regulated by WNT4, which include known β-catenin–dependent and –independent pathways. However, WNT4 also functions via atypical pathways, including PORCN-independent signaling and novel intracellular signaling roles in ILC and other cancer cells. Defining WNT4 signaling and the underpinning of WNT4 tissue-specific functions can improve the understanding of endocrine organogenesis and reproduction, and will identify new treatment approaches for a broad range of WNT4-related gynecologic pathologies.

Glossary

Abbreviations

ER

estrogen receptor

ERE

estrogen response element

GWAS

genome-wide associations studies

ILC

invasive lobular carcinoma

MRKH

Mayer-Rokitansky-Küster-Hauser syndrome

mTOR

mammalian target of rapamycin

OCCC

ovarian clear cell carcinoma

OR

odds ratio

PGE₂

prostaglandin E₂

Porcn

O-acyltransferase Porcupine

PR

progesterone receptor

SERKAL syndrome

sex reversal kidneys adrenal and lung dysgenesis syndrome

SNP

single-nucleotide polymorphism

WLS

Wntless

Additional Information

Disclosures: The authors have nothing to disclose.

Data Availability

Data sharing is not applicable to this article because no data sets were generated or analyzed during the present study.