Decoding the link between WWOX and p53 in aggressive breast cancer

ABSTRACT

Basal-like breast cancer (BLBC) and triple-negative breast cancer (TNBC) are aggressive forms of human breast cancer with poor prognosis and limited treatment response. Molecular understanding of BLBC and TNBC biology is instrumental to improve detection and management of these deadly diseases. Tumor suppressors WW domain-containing oxidoreductase (WWOX) and TP53 are altered in BLBC and in TNBC. Nevertheless, the functional interplay between WWOX and p53 is poorly understood. In a recent study by Abdeen and colleagues, it has been demonstrated that WWOX loss drives BLBC formation via deregulating p53 functions. In this review, we highlight important signaling pathways regulated by WWOX and p53 that are related to estrogen receptor signaling, epithelial-to-mesenchymal transition, and genomic instability and how they impact BLBC and TNBC development.

Introduction

Human breast cancers are a group of heterogeneous diseases, harboring different genetic alterations and differentially responding to therapy. The classical molecular classification of breast cancer is mainly based on the expression of estrogen receptor (ER), progesterone receptor (PR) and the receptor tyrosine-protein kinase ErbB-2 (HER2). Breast cancer is hence classified into five subtypes: Luminal A (ER+ PR+, HER2- and Ki67-), Luminal B (ER+ PR±, mostly HER2- and Ki67+), HER2/ErbB2 subtype (ER- PR-, HER2+) where HER2 is usually amplified or overexpressed, normal-like subtype expressing adipose and other non-epithelial genes, basal-like subtype (BLBC) (ER- PR-, HER2- and high expression of basal genes such as cytokeratins CK-5 and CK-14 and EGFR/HER1) [1–3]. Most BLBCs are triple negative (TNBC); however, some ER-positive tumors and HER2-positive tumors display a basal-like gene expression profile. Other tumors are believed to acquire basal-like features through a trans-differentiation process known as EMT (epithelial-to-mesenchymal transition). BLBC is usually associated with increased aggressiveness, invasiveness and metastatic potential hence resulting in poor prognosis [4]. In 2007, Herschkowitz et al. introduced a new breast cancer subtype, called Caludin-low, which is characterized by the low expression of genes involved in tight junctions and cell-cell adhesion, including Claudins 3, 4, 7, Occludin, and E-cadherin [5].

Some of the most known genetic alterations in tumor suppressor genes in BLBC and TNBC are mutations in TP53, as well as the loss of RB1 and CDKN2A [6,7]. In fact, it has been reported previously that up to 80% of BLBC harbor TP53 mutations, which commonly include nonsense and frameshift mutations [8]. Recently and based on RNA-based method, it was reported that the RNA of 99.4% of BLBC cases harbor TP53 mutant-like status [9].

In a recent report, our group has demonstrated that tumor suppressor WWOX (WW domain-containing oxidoreductase) is frequently alerted in breast cancer and its targeted deletion in murine mammary epithelial glands drives the development of BLBC-like tumors via inactivation of Trp53 [10]. Several studies by a number of groups have further demonstrated that the WWOX locus is targeted in BLBC and TNBC tumors. For example, it has been shown that WWOX protein levels are reduced or absent in 96.6% of BLBC tumors [11]. More recently, Chang and colleagues have reported that WWOX protein expression is commonly absent in TNBC and this loss drives tumor metastasis [12]. Lately, our group has also determined that WWOX reduced expression and copy number variation are correlated with advanced stages of TNBC, further highlighting the significance of WWOX function in TNBC development [13]. Beside genomic rearrangements and loss of protein expression, hypermethylation of the regulatory region of WWOX has been documented in neoplastic but not in paired adjacent non-neoplastic tissues [14–16].

WWOX tumor suppressor functions have been proposed to be mediated via its protein adaptor capabilities, through which WWOX’s first WW domain binds with proline-rich motifs of partner proteins hence regulating their localization, stability, and transactivation [17–20]. The consequence of these interactions resulted in regulating cellular pathways including apoptosis [21], DNA repair [22,23], cellular metabolism [24,25] and others [26–30]. Of particular interest, WWOX physical and functional interaction with p53 has been proposed to enhance apoptosis in vitro (reviewed in [31,32]). It was also reported that WWOX enhances the cytotoxic function of tumor necrosis factor (TNF) by down-regulating apoptosis inhibitors Bcl-2 and Bcl-xL and up-regulating apoptotic p53 [33]. Our recent findings indicate that the WWOX-p53 functional interaction is important in BLBC and TNBC development [10] as shown previously in other malignancies including, glioblastoma [34] and osteosarcoma [35]. In this review article, the crosstalk of WWOX-p53 in aggressive breast cancer is discussed in an attempt to better understand the complexity of this fatal disease hoping this would help in improving its detection and management.

Wwox and Trp53 conditional knockout (cKO) in mice

Genetically engineered mouse (GEM) models of cancer have been developed to model tumors with genetic alterations that resemble human cancer. Similar to human BLBC, inactivation of WWOX or p53 in the mammary gland epithelium is associated with BLBC-like in murine models. Previous studies revealed that conditional ablation of Trp53 using MMTV-Cre and WAP-Cre transgenic mice in C57BL/6 strain displays high percentage of tumor incidence, though at a later stage of life; in the MMTV case high percentage of mice (47%-100%) developed mammary tumors with latency of 14.5 months (Average of two different MMTV-Cre models) [36]. These tumors were classified as BLBC [36].

The fact that WWOX DNA sequence is highly conserved suggests its essential role in physiology and explains the pathophysiologies associated with its loss. For this reason, we and others have modeled WWOX loss of expression in different animal models (reviewed in [37]). Our lab has pioneered in studying conventional and conditional deletion of murine Wwox GEM models and contributed several research studies supporting WWOX tumor suppressor function [10,38–44], WWOX’s role in cellular metabolism [25,43] and other functions [45,46].

Using these GEM models, it has been demonstrated that aged germline Wwox-heterozygous mice in mammary susceptible C3H genetic background develop mammary tumors with ~50% penetrance [40]. These mammary tumors were mostly ER-negative and PR-negative, expressing cytokeratin (CK)-14, hence reminiscent of the common WWOX inactivation in BLBC. Conditional knockout of Wwox in the mammary glands (using MMTV-Cre) resulted in mammary tumors (latency of 270 days) [10]. These tumors were also characterized as BLBC-like tumors: ER-negative, PR-negative and show high mRNA expression of the basal markers including Ck-14, Ck-17, Cav1, Cav2 and low expression of Foxa1 and Gata3 [10]. In the same study, conditional knockout of Trp53 in mammary glands (using MMTV-Cre transgenic line) resulted in mammary tumors (latency of 262.5 days) with basal-like features. Tumors of both conditional Wwox and Trp53 knockout mice were indistinguishable and displayed hallmarks of BLBC both at immunohistochemical staining and RNA sequencing [10]. Furthermore, when RNA of Wwox-knockout tumors wascompared to that of new and previous Trp53-knockout models [47], all the analyzed tumors clustered together. These findings indicate the existence of an intimate relationship between WWOX and p53 in BLBC and TNBC development.

WWOX, p53 and estrogen receptor (ER)

ER is considered as a powerful prognostic marker and an efficient target for the treatment of hormone-dependent breast cancer with antiestrogens [48,49]. ERs are ligand-activated transcription factors [50] and are known to have a central role in cell cycle regulation [51,52]. Two forms of ER exist: alpha and beta. Currently, only the ERα subtype is clinically considered for clinical decision-making and treatment in breast cancer [50]. ERα is expressed in the majority of breast tumors (with immunohistochemical staining in approximately two-thirds of breast tumors) [50]. Therefore, it is generally believed that breast tumors depend, at least initially, on the stimulatory effects of estrogens; however, many breast tumors eventually progress to an estrogen-independent growth phenotype [53]. Several factors have been proposed to contribute to this later phenomenon among which are increased expression of estrogen-regulated genes [54], activation of mitogen-activated protein kinase (MAPK) [55], overexpression of the vascular endothelial growth factor (VEGF) [56].

Several lines of evidence showed that there is a close link between WWOX and ER in both normal and cancer contexts. In cancer, WWOX is reduced/absent in BLBC and TNBC tumors (mostly negative for ER and PR) [40]. Absent WWOX expression significantly associates with poor distant disease-free survival when compared with patients that display normal WWOX expression. This association was maintained in the subgroup of ER-negative patients but not in ER-positive patients [57]. An independent study has also reported that there is a strong correlation between WWOX expression and ER; ~80% of ER-negative cases demonstrate loss or reduced expression of WWOX [58]. Moreover, patients with strong expression for WWOX are more sensitive to tamoxifen treatment, while patients with reduced expression of WWOX are tamoxifen-resistant [59]. Remarkably, WWOX depletion in human ER-positive MCF7 breast cancer cell line, using shRNA constructs or by CRISPR/Cas9, results in reduced ER expression and function [10,40]. Along this line, it was shown that in normal mouse mammary gland development WWOX protein levels are induced at the age of 3–4 weeks [42], parallel to ER that is known to be induced at the same time [60]. Additionally, ER conditional knockout mice (using MMTV-Cre) display severe-impaired ductal elongation and side branching [61], akin to Wwox cKO mice [42,62]. How WWOX modulates ER levels and activity is largely unknown, one possibility could be through regulation of ER co-activator WBP2 (WW domain-binding protein 2) that enhances ER function via YAP and/or Wnt signaling and reported to physically interact with WWOX [63–66]. Altogether, these clinical findings suggest that there is a positive correlation between WWOX and ER in normal tissues and imply WWOX is upstream to ER signaling. In ER+ breast cancer, ER signaling somehow progresses to be WWOX-independent, while in ER-cancer it seems that tumor cells initially lose WWOX and then lose ER expression.

On the other hand, it is known that ERα inhibits p53-mediated cell cycle arrest and apoptosis through binding p53 and repressing its transcriptional function (MCF7 cells, as well as in a mouse xenograft model) [67–69]. Moreover, Konduri et al. reported that ER plays an important role in the repression of p53-mediated transcriptional activity [70]. Instead, knockdown of ERα (in MCF7) resulted in decreased expression of p53 and its downstream targets, MDM2 and CDKN1A [71]. It was also demonstrated that ERα activates p53 transcription via binding to estrogen response element within the p53 promoter [71].

In our mammary-specific Wwox-cKO model, we presented evidence that the reduction in p53 levels is due to genomic focal deletion that was validated by PCR of gDNA [10]. Interestingly, when WWOX was knocked out in vitro (in MCF7 cells), p53 dysregulation was not associated with genomic deletion in the TP53 locus but likely was due to transcriptional repression [10]. This observation could imply that WWOX controls p53 transcription through regulation of ER. Further studies would be required to delineate the WWOX-ER-p53 axis to better uncover its dynamics in breast cancer development.

WWOX, p53, and EMT

Epithelial-to-mesenchymal transition (EMT) is defined by enhanced levels of the mesenchymal markers (vimentin, smooth-muscle actin, N-cadherin and cadherin-11), reduced expression of the epithelial marker (E-cadherin) [72], loss of cellular adhesion and changes in polarization of the cell and its cytoskeleton [73]. EMT was described to be associated with malignancy, invasion, and metastasis [73]. As BLBC and TNBC are the most aggressive breast cancer subtypes, they display high metastatic ability and mesenchymal features [73]. Sarrio et al. indeed reported an up-regulation of the EMT markers and reduction of the epithelial markers among basal-like tumors, suggesting that this mesenchymal transition may be related to the high aggressiveness and metastatic spread of the basal-like tumors [72]. Choi et al. also showed that BLBCs have high levels of the EMT markers [74].

Several lines of evidence demonstrate that inactivation of WWOX affects EMT, which might be related to tumor progression [10,13,75]. Gourley et al. reported that loss of WWOX during ovarian cancer development results in increased levels of membranous integrins and increased adhesion and migration of tumor cells on extracellular matrix and hence speculated that this would enhance locoregional peritoneal ovarian tumor spread and metastasis [76]. Consistent with this scenario, loss of WWOX protein expression has previously been correlated with advanced stage disease and poorer survival in ovarian cancer patients [58]. More recently, Khawaled et al. demonstrated that manipulation of WWOX expression in TNBC and BLBC cell lines modulates invasion, metastatic seeding, and colonization [13]. It was demonstrated that WWOX, through modulation of microRNAs, regulates the levels of both epithelial and mesenchymal markers hence antagonizing EMT and invasion of TNBC cells [13]. This has been suggested to be achieved, at least in part, through negative regulation of c-MYC expression and activity, resulting in miR-146a accumulation hence targeting the mesenchymal gene Fibronectin and supports the epithelial phenotype [13]. These findings and others suggest that WWOX tumor suppressor function impacts a plethora of pathways to antagonize invasion and metastasis [12].

p53 has been also described to regulate EMT, often through microRNAs [77–82]. For example, Ohtsuka et al. reported that loss of p53 induces EMT and cellular motility in gastric epithelial cells prior to the development of gastric cancer [83]. Additionally, it was shown that p53 alters ZEB1 and ZEB2 expression, transcription factors known to promote EMT, by upregulating microRNAs, including miR-200 and miR-192 family members [80]. Di Gennaro et al. demonstrated that p53 controls EMT and tumor cell invasion via miR-30a, which probably acts through upregulation of miR-200c [84]. p53 transactivates miR-200c through direct binding to its promoter [77]; when p53 is lost in mammary epithelial cells this results in decreased expression of miR-200c and in EMT activation [77,85]. Altogether, these findings suggest that both WWOX and p53 may play an important role in inhibiting the EMT pathway and subsequently inhibiting the breast cancer progression. Whether WWOX loss mediates p53 regulation of microRNAs in TNBC and BLBC is still unknown and shall be determined in future studies.

WWOX, p53, and genomic instability

Genome instability is considered as an enabling characteristic of almost all cancer types [86]. Genomic instability can be referred to an increased tendency of alterations in the genome during the life cycle of cells. These genomic alterations confer a selective advantage on subclones of cells, enabling their outgrowth and expansion, resulting in cancer [86]. Moreover, genomic instability is thought to play an important role in cancer resistance to therapy [87,88]. Disruption of the DNA damage response (DDR) machinery in human cells leads to genomic instability and an increased risk of cancer progression [86,88].

p53 has been shown to be involved in the DDR. Induction of DSBs activates ATM [89], which in turn phosphorylates p53 directly or indirectly, resulting in p53 accumulation and activation [90–93]. ATM could also phosphorylate MDM2, inhibiting the ability of MDM2 to ubiquitinate p53, thus leading to p53 stabilization [94]. The consequence of p53 accumulation and stabilization results in transcribing its target genes which include those important for cell cycle arrest (to allow repair of the DNA damage) or apoptosis (to eliminate cells with unrepaired DNA). The cell cycle arrest mediated by p53 activation is mainly regulated by p21/WAF [95] while apoptosis is mediated by different pro-apoptotic genes including Puma, Noxa, Bax, and others [95]. In case of p53 deficiency, unrepaired damage and genomic instability are observed [96]. In fact, p53-deficient mammary tumors display increased genomic instability with aneuploidy, amplifications, and deletions [97] and are associated with radioresistance [98].

WWOX has been also shown to play a direct role in the DDR. Upon DNA damage, WWOX levels are induced and accumulate in the cell nucleus, where it interacts with ATM and enhances its activation [22]. WWOX depletion results in reduced ATM signaling and reduced DNA repair, contributing to increased DSBs [10,99]. WWOX loss was also associated with increased chromosomal instability and increased DSBs [23]. Interestingly, the induction of DSBs in WWOX deficient cells is associated with impaired p53 accumulation and signaling [10,22]. More recently, it was shown that WWOX binds BRCA1 and regulates DNA repair [100]. Altogether, we believe that alteration of WWOX expression, which is associated with impaired checkpoint protein signaling, could impair p53 accumulation and activation resulting in improper DDR and genomic instability further contributing to BLBC and TNBC formation and progression.

The interplay between WWOX and p53 in breast carcinogenesis: conclusions and future perspective

Positive and/or negative regulators, mainly at the protein level, do control the p53 protein. Upon stress, kinases including ATM/ATR and Chk1/Chk2 phosphorylate p53, thus stabilize its protein product and promote its DNA binding ability [101]. Another positive regulation mechanism is recognized by increasing the p53 stability as shown by Freeman et al. demonstrating that PTEN stabilizes p53 by increasing its half-life [102]. On the other hand, Mdm2 negatively regulates p53 through the ubiquitin-proteasome system [103–105]. Mdm2 is transcriptionally induced by p53 highlighting a negative feedback loop mechanism in which p53 controls its own degradation. Other reports showed that microRNAs could also regulate p53 levels [106–108]. For example, miR-125b binds to 3′-UTR of the TP53 mRNA resulting in decreased mRNA level [107]. Knockdown of miR-125b induced apoptosis through increasing the levels of p53 in human lung fibroblasts [107]. MicroRNAs can also positively regulate p53 by targeting known negative regulators of p53, such as silent information regulator 1 (SIRT1) gene [109,110].

Our recent findings demonstrate that WWOX, gene product of FRA16D, can also regulate p53 levels and activity. Expression of the WWOX gene is induced upon DNA damage to enhance DNA repair or apoptosis [22,23]. Furthermore, several microRNAs were shown to regulate WWOX expression. One example is overexpression of miR-29 in lung cancer cell lines which has been shown to restore normal patterns of DNA methylation and induce expression of WWOX hence suppressing tumorigenicity both in vitro and in vivo. Interestingly, miR-29 is induced in response to DNA damage and occurs in a p53-dependent manner [111] perhaps contributing to the suppression of breast carcinogenesis [112,113].

The similar behavior of WWOX and p53 suggests that both act in the same way, at least in breast cancer. Moreover, the current evidence suggests that WWOX is an upstream regulator of p53. We showed that WWOX deficiency results in p53 loss/reduction in vitro and in vivo [10,99]. Our recent study suggests new mechanisms of regulation of p53 that are mediated by WWOX action (Figure 1). The first mechanism we propose is p53 regulation through ER signaling; p53 is regulated by ER and WWOX depletion results in reduced ER levels hence it is possible to speculate that WWOX through regulating ER levels controls those of p53. It is possible that the reduction in ER levels in WWOX KO cells is mediated through WBP2 protein. In fact, it was reported that WBP2 is essential for proper activation of PR and ER [63]. Intriguingly, WBP2 contains a PPxY motif that associates with WW domain protein YAP [114], leading to ER and PR transactivation. As WWOX is known to interact with PPxY-containing proteins as well [26], it is possible that WWOX may associate with WBP2 and regulate ER/PR signaling. Consistent with this scenario, it was shown that the WBP2–WWOX interaction attenuates the transactivation functions of ER [115]. It is also possible that WWOX may associate with transcription factors that regulate ER gene transcription. Further studies should shed light on such plausible mechanisms.

Figure 1.

WWOX regulates p53 function, model of action and scenarios. I. WWOX modulates the DNA damage response pathway, antagonizes genomic instability and inhibits loss of p53. II. WWOX modulates ER expression and function, which in turn positively or negatively affects p53. III. WWOX inhibits the epithelial-to-mesenchymal transition (EMT). p53 is known to inhibit both genomic instability and EMT.

Mammary tumors developed in Wwox and Trp53 mouse models display enhanced expression of EMT markers and high EMT scores (compared to other mammary tumor mouse models) [10]. Up till now, the role of both WWOX and p53 in EMT is reported to be mediated by microRNAs. Whether WWOX has an upstream effect on p53 in the EMT program is still to be investigated. Recently, it has been reported that WWOX negatively regulates c-Myc, a master regulator of microRNA expression [13]. Some of these microRNAs may regulate p53 mRNA levels, though this is yet to be shown.

The third possible mechanism of action for WWOX is through maintaining genome stability. The role of WWOX or p53 in the DDR is heavily investigated, but the connection between both of them is poorly studied. The Aqeilan’s lab showed that in response to induction of DSBs or DNA single-strand breaks (SSBs), WWOX accumulates and promotes DNA repair through the activation of ATM or ATR [22,23]. Indeed, WWOX loss is associated with reduced ATM/ATR activation and substrate phosphorylation, such as CHK2 and CHK1 phosphorylation, respectively [22,23]. Reduced ATM signaling could result in reduced p53 phosphorylation [91] and impaired DDR in WWOX-deficient cells. Indeed, we observed that ionizing radiation of WWOX–depleted MCF7 cells results in reduced ATM phosphorylation and reduced nuclear p53 accumulation [10], suggesting that WWOX may affect p53 levels through ATM. It was also reported by Ouchi et al. that BRCA1 coimmunoprecipitates with p53 and acts as a p53 coactivator, which enhances p53-dependent gene expression [116]. Moreover, BRCA1 and/or TP53 alterations are associated with higher homologous recombination deficiency score [117]. The Huebner lab has recently shown that WWOX, via its first WW domain, interacts with BRCA1 and proposed that the BRCA1–WWOX complex supports non-homologous end-joining pathway as the dominant DSB repair pathway in WWOX-sufficient cells [100]. Whether WWOX affects p53’s level/function through BRCA1, is still to be investigated.

Since cloning of the TP53 gene [118], thousands of articles were published describing its critical roles, functions, and partners in cancer and biology. Consequently, p53 is known today as the master regulator of many cellular pathways [119,120]. Here we introduce a new regulator of p53, WWOX. This regulation occurs at three levels: DNA, RNA, and protein.

Funding Statement

This work was supported by the FP7 Ideas: European Research Council [682118]; Israel Science Foundation [1574/15].

Acknowledgments

The Aqeilan lab is supported by European Research Council (ERC)-Consolidator Grant under the European Union’s Horizon 2020 research and innovation program (grant agreement No. 682118) and Israel Science Foundation (grant agreement No 1574/15).

Disclosure statement

No potential conflict of interest was reported by the authors.