The prognostic significance of WWOX expression in patients with breast cancer and its association with the basal-like phenotype

Abstract

Purpose

The basal-like phenotype has been found to be an independent poor prognostic factor for breast cancer. The aim of this study was to evaluate the association of WW domain-containing oxidoreductase (WWOX) expression with the basal-like subtype and clinicopathological parameters and to determine the prognostic significance of WWOX expression in patients with breast cancer.

Methods

Immunohistochemical analysis of WWOX expression was performed on 267 breast carcinoma samples, and then the mean value of WWOX expression was correlated to the basal-like status and clinicopathological parameters of the samples. The prognostic value of WWOX in primary breast cancer patients was determined for disease-free survival and overall survival.

Results

Expression of WWOX was negative in 29% of cases, and mean WWOX levels were significantly lower in basal-like breast cancers than in those of the non-basal-like subtype (P = 0.01). WWOX negativity was associated with decreased disease-free survival (DFS) (hazard ratio = 1.83; 95% CI, 1.01 to 3.28), but not with overall survival. Other tumor variables that showed a significant association with patient survival times included node status (hazard ratio = 0.38; 95% CI, 0.17 to 0.85) and breast cancer phenotype (hazard ratio = 0.36; 95% CI, 0.19 to 0.68). Multivariate regression analysis showed that lymph node involvement (hazard ratio = 0.43; 95% CI, 0.19 to 0.97) and basal-like subtype (hazard ratio = 0.33; 95% CI, 0.17 to 0.63) were also significant independent prognostic variables, and WWOX expression was of borderline significance for DFS (hazard ratio = 0.56; 95% CI, 0.31 to 1.03).

Conclusions

Reduced WWOX expression is associated with the basal-like subtype and a poor disease-free survival rate for breast cancer patients. Additional studies are warranted to better understand the role of WWOX expression, to further refine prognosis, and to optimize treatment in patients with basal-like breast cancer.

Background

Recently, Perou et al. (2000) used cDNA microarray profiling studies to categorize breast carcinoma into the following five distinct subtypes: luminal A [estrogen receptor(ER)+ and/or progesterone receptor(PR)+, HER-2/neu−], luminal B (ER+ and/or PR+, HER-2/neu+), HER-2/neu overexpression (ER-, PR-, HER-2/neu+), basal-like (ER-, PR-, HER-2/neu−, CK5/6+), and normal breast tissue. The basal-like subtype is associated with aggressive histology, poor prognosis, unresponsiveness to the usual endocrine therapies, and shorter survival (Sorlie et al. 2003; Livasy et al. 2006). Although this classification system is based on extensive genetic profiling assays, a simplified method of classification (based on ER, PR, and HER-2/neu status) is appealing and more clinically useful. The basal-like phenotype is immunohistochemically characterized as triple-negative (lack of HER-2/neu and ER/PR expression) and positive for basal cytokeratins (CK5/6 and/or CK14) and/or epidermal growth factor receptor (EGFR) expression (Livasy et al. 2006; Nielsen et al. 2004). It should be noted that only approximately 85% of triple-negative breast cancers are deemed to be basal-like when tested by appropriate immunohistochemical means (Livasy et al. 2006).

Due to the absence of specific treatment guidelines for this subgroup, basal-like breast cancers are managed with standard treatments; however, such treatments in this subgroup of patients are associated with a high rate of local and systemic relapse. Parikh et al. (2008a, b) reported that patients with ipsilateral breast tumor recurrence events of the basal-like phenotype had a poor prognosis despite receiving standard chemotherapy. The pathways that drive proliferation of these tumors are still poorly understood; however, once they have been elucidated, targeted agents can be developed, which could result in better outcomes for patients with basal-like tumors.

The WW domain-containing oxidoreductase (WWOX) gene maps to the location of FRA16D, one of the most active common chromosomal fragile sites (Bednarek et al. 2000). The WWOX gene is similar to the fragile histidine triad gene (FHIT) since both encompass very active common fragile sites, both genes show frequent allelic loss and/or homozygous deletion in several human cancers, and both frequently show aberrant transcripts (Iliopoulos et al. 2005; Nunez et al. 2005a, b; Kuroki et al. 2002, 2004; Paige et al. 2001; Yendamuri et al. 2003; Aqeilan et al. 2004; Park et al. 2004). However, until now, there are no data regarding WWOX expression in the basal-like phenotype of breast cancer, and the prognostic value of WWOX expression has not been reported.

The purpose of the current study was twofold. First, in order to better define the role of WWOX in the progression of breast cancer, we performed immunohistochemical staining of WWOX to evaluate the correlation between WWOX status and the basal-like phenotype, as well as prognostic factors. The basal-like phenotype was defined according to the expression of basal cytokeratins, wherein ≥10% of tumor cells expressing cytokeratin 5/6 and/or EGFR were considered basal-like (Cheang et al. 2008). Second, we aimed to determine the prognostic significance of aberrant WWOX expression in patients with breast carcinoma, with attention to both disease-free and overall survival endpoints.

Materials and methods

Patients

The proposal for chart review and archived tumor specimen use was reviewed and approved by the institutional review board at the Shandong University School of Medicine. Patients (267) with primary breast carcinoma underwent modified radical mastectomy, lumpectomy, or breast conserving surgery at Jinan Central Hospital of Shandong University from May 2000 to September 2005. Adjuvant systemic chemotherapy and/or adjuvant hormone therapy was administered as clinically indicated in accordance with standard practices during this time interval. Estrogen receptor-positive patients receiving chemotherapy were also routinely treated with adjuvant hormonal therapy.

Expression of all three molecular markers (ER, PR, and HER-2/neu) of these patients was available. Breast carcinomas were routinely divided into ductal carcinoma, lobular carcinoma, and others (medullary carcinoma, mucinous carcinoma). The tumor grade of invasive carcinomas was classified according to the Scarff-Bloom-Richardson system (Le Doussal et al. 1989). Based on the frequency of cell mitosis, tubule formation, and nuclear pleomorphism, invasive carcinoma was graded as grade 1 (low), grade 2 (moderate), or grade 3 (high). The presence of lymph node metastases was reviewed for each patient. The tumor stage was determined according to the American Joint Committee on Cancer (AJCC) Cancer Staging Manual, sixth edition (Greene et al. 2002). The patients included in this study had no evidence of distant metastases at the time of diagnosis. Patients with evidence of disease within 1 month after primary surgery were excluded. Patients with only carcinoma in situ or with bilateral breast cancer were also excluded from this study.

Basal-like breast carcinoma tissues

ER, PR, and HER-2/neu status were assessed by standard immunohistochemical methods. Immunohistochemical markers were assayed in paraffin-embedded, formation-fixed tissue stained with hematoxylin and eosin using antibodies against ER, PR, and HER-2/neu (Dako, Glostrup, Denmark). For ER and PR, receptor positivity was scored as more than 10% of cells testing positive. HER-2/neu was assessed through immunohistochemistry and was scored on a qualitative scale from 0 to 3+, based on interpretation of staining, with 0 and 1+ classified as negative and 3+ as positive. Tumor specimens that showed an intermediate HER-2/neu staining score of 2+ were further assayed with fluorescent in situ hybridization (FISH) (Vysis, Des Plaines, IL, USA) to determine HER-2 amplification. All samples that had intense, complete, membranous staining in >10% of invasive neoplastic cells determined by IHC or the presence of HER-2 gene amplification determined by FISH were considered HER-2 positive.

The breast cancers were then divided into subtypes included luminal A (positive for ER and/or PR and negative for HER-2/neu), luminal B (positive for ER and/or PR and positive for HER-2/neu), HER-2/neu overexpression (negative for ER and PR and positive for HER-2/neu), and basal-like (negative for ER, PR, and HER-2/neu and positive for CK5/6 and/or EGFR) (Cheang et al. 2008). In the current study, luminal A, luminal B, and HER-2/neu overexpression subtypes were combined into a single ‘non-basal-like’ group and were compared to the basal-like group.

Immunohistochemical staining method

WWOX expression was determined by immunohistochemical staining of tissue sections using a primary rabbit polyclonal antibody against WWOX. Evaluation of WWOX expression was performed via detection of the streptavidin-peroxidase complex. Histopathological assessment was performed on paraffin sections of 5 μm in thickness. Immunohistochemistry was performed on sections after incubating the tissue in 3% H₂O₂ in methanol for 10 min, followed by washing in 0.01 mmol/L phosphate-buffered saline and treatment with 1% normal goat serum for 10 min at room temperature (the latter to block non-specific binding of the antibody). The slides were then incubated with the primary rabbit polyclonal antibody against WWOX (ab14691, 1:100 dilution, Abcam) for 2 h at 4°C. Sites of primary antibody binding were identified by incubating blots with horseradish peroxidase-conjugated anti-rabbit IgG (Santa Cruz Biotechnology, 1:250 dilution) followed by reaction with the 3,3,diaminobenzidine tetrahydrochloride (DAB) substrate. For triple-negative breast tumors, additional immunohistochemistry was performed with antibodies against CK5/6 (clone D5/16B4, 1:50 dilution, DAKO), and EGFR (ab2430, 1:200 dilution, Abcam) using a streptavidin–biotin horseradish peroxidase detection kit (Biogenex, San Ramon, CA, USA). Secondary antibody was applied for 1 h, and staining was visualized according to the manufacturer’s instructions (Santa Cruz Biotechnology). Omission of the primary antibody was used as a negative control. Cases of invasive ductal carcinoma previously known to be positive for WWOX, CK5/6, and EGFR were used as positive controls.

Evaluation of immunohistochemical staining

Immunohistochemical staining was evaluated and scored by two pathologists (H.L., and X.Z.) who worked independently and were blinded to patient clinical information. The expression of WWOX and CK5/6 was cytoplasmic, whereas the expression of EGFR was both cytoplasmic and membranous. Sections with cytoplasmic expression in ≥10% of tumor cells for CK5/6 and membranous staining in ≥10% of tumor cells for EGFR were considered positive. Staining intensity and extent of WWOX were scored according to previous reports (Guler et al. 2004; Iliopoulos et al. 2005). Intensity of staining was graded as following: 1 (no staining), 2 (moderate staining), and 3 (intense staining). The extent of staining was graded according to the percentage of positive cells as follows: 1 (<10%), 2 (11–25%), 3 (26–50%), 4 (51–75%), 5 (76–100%). The staining score was calculated by multiplying the score of intensity and extent of staining. Staining scored >2 fell into the category of positive, while staining scored ≤2 was marked as ‘negative’ for WWOX expression.

Statistical analysis

In order to better understand the role of WWOX, we determined WWOX expression with a quantitative measure and assessed the mean value of WWOX expression against the basal-like subtype and other clinicopathological parameters in the 267 patients. The association between WWOX expression and breast carcinoma phenotype, ER status, PR status, lymph node status, and menopausal status were evaluated using the Student’s t test. One-way ANOVA was used to evaluate the association between the mean score of WWOX and the TNM stage, histological subtype, nodal status, and tumor grade of patients. Spearman rank correlation was used to determine the association between WWOX expression in the primary tumor and various independent factors (age, tumor size, and number of axillary lymph node metastases). All statistical tests were two-sided. A statistical significance of observed difference was set at P < 0.05. All data were analyzed using SPSS statistical software (Version 13.0, SPSS Inc., Chicago, IL).

Despite the short follow-up time, the prognostic value of WWOX in primary breast cancer patients on disease-free survival and overall survival, including all deaths, was determined. Univariate survival curves were generated by the Kaplan–Meier method, and differences in survival among the breast cancer subtypes were assessed by the log-rank test. The Cox regression model was also used to examine several combinations and interactions of different prognostic factors in a multivariate analysis. In the multivariate analysis, only parameters that achieve statistical significance for disease-free survival or overall survival in the log-rank test were included.

Results

Association between WWOX expression and clinicopathological features

The median age of patients was 48.6 (range, 26–72 years), and 124 patients are younger than 50 years of age. Tumor size ranged from 0.3 to 8 cm (mean size, 3.1 cm), and 154 tumors are less than 2 cm. Table 1 summarizes the frequency statistics for the clinicopathologic characteristics of the patient cohort including staging and molecular marker status. Representative immunostaining results for WWOX, CK5/6, and EGFR are shown in Fig. 1. There was a significant difference between mean WWOX expression in the basal-like phenotype and those in non-basal-like subtype (P = 0.01). In addition, there was a significant difference in the mean score of WWOX between patients with negative and positive ER status (P = 0.01), pre- and postmenopausal patients (P = 0.02), and TNM stage (P = 0.03) (Table 1).

Table 1.

Association between WWOX expression and clinicopathological features

Clinicopathological features	n	Score ( ± s)	P value
Breast carcinoma subtype	267	8.24 ± 5.21
Basal-like	40	6.30 ± 4.54
Non-basal-like	227	8.58 ± 5.26	0.01
TNM stage
I	97	7.40 ± 5.04
II	116	8.26 ± 5.25
III	54	9.70 ± 5.22	0.03
Histological subtype
Ductal	171	8.60 ± 5.36
Lobular	59	7.92 ± 4.81
Others	37	7.08 ± 5.09	0.24
Tumor grade
1	89	7.81 ± 5.13
2	123	8.64 ± 5.37
3	55	8.04 ± 5.00	0.49
ER status
Negative	111	7.24 ± 5.27
Positive	156	8.85 ± 5.03	0.01
PR status
Negative	102	7.82 ± 5.40
Positive	165	8.50 ± 5.10	0.31
Menopause status
Premenopausal	138	8.96 ± 5.04
Postmenopausal	129	7.47 ± 5.30	0.02
HER-2/neu status
Negative	188	8.44 ± 5.34
Positive	79	7.76 ± 4.91	0.33
Lymph node metastasis
No	83	8.88 ± 4.54
Yes	184	7.95 ± 5.48	0.18

Fig. 1.

Representative immunohistochemical staining of protein markers in breast cancer. a WWOX-negative expression in basal-like phenotype (WWOX score = 2). b Strong cytoplasmic staining of WWOX protein in non-basal-like phenotype (WWOX score = 12). c Strong cytoplasmic expression of cytokeratins 5 and 6 (CK5/6). d Strong membranous expression of epidermal growth factor (EGFR). (original magnification: a, b, and c ×400, d ×200)

Association between WWOX protein expression and patient survival

As of May 2008, with a median follow-up time of 5.1 years (range 4.4–6.0 years) in this cohort, there have been 19 local relapses, 27 distant recurrences, and 31 deaths. The 5-year DFS was 72% for WWOX-negative tumors and 84% for the WWOX-positive tumors (hazard ratio = 1.83; 95% CI, 1.01 to 3.28). The 5-year OS for the WWOX negative and positive cohorts was 87 and 89%, respectively (hazard ratio = 1.13; 95% CI, 0.53 to 2.40). WWOX expression was not significantly associated with OS. Kaplan–Meier survival curves for 5-year DFS and OS, with respect to WWOX expression, are presented in Fig. 2a and b, respectively. By univariate analysis, significant factors associated with decreased DFS included: axillary lymph nodal metastasis (hazard ratio = 0.38; 95% CI, 0.17 to 0.85), and basal-like phenotype (hazard ratio = 0.36; 95% CI, 0.19 to 0.68). Furthermore, nodal status (hazard ratio = 0.26; 95% CI, 0.08 to 0.88) and breast cancer subtype (hazard ratio = 0.43; 95% CI, 0.19 to 0.97) showed a significant association with overall survival time in this group of patients.

Fig. 2.

Disease-free survival (DFS) a and overall survival (OS) b curves for breast cancer patients whose tumors were WWOX positive [+] (IHC score >2) vs. WWOX negative [×] (IHC score ≤2), as determined by immunohistochemical (IHC) staining (DFS P = 0.04; OS P = 0.74)

WWOX expression, nodal status, and breast cancer subtype were entered into the multivariate model to determine their relationship with 5-year DFS and OS. In the multivariate analysis summarized in Table 2, lymph node involvement (hazard ratio = 0.43; 95% CI, 0.19 to 0.97) and basal-like subtype (hazard ratio = 0.33; 95% CI, 0.17 to 0.63) were also significant independent prognostic variables for DFS and OS. WWOX expression was only of borderline significance for DFS (hazard ratio = 0.56; 95% CI, 0.31 to 1.03).

Table 2.

Univariate and multivariate analysis of disease-free survival (DFS) and overall survival (OS)

Factor	DFS		OS
	HR (95% CI) (univariate)	HR (95% CI) (multivariate)	HR (95% CI) (univariate)	HR (95% CI) (multivariate)
Age (years), <50 vs. ≥50	0.93 (0.52–1.65)		1.03 (0.49–2.03)
Histological grade, (1,2) vs. 3	0.94 (0.45–1.95)		0.89 (0.37–2.17)
Menopausal status, Pre vs. Post	0.79 (0.44–1.41)		0.87 (0.43–1.76)
Tumor size (cm), >2 vs. ≤2	0.73 (0.38–1.41)		0.70 (0.31–1.57)
Nodal status, N+ vs. N0	0.38 (0.17–0.85)	0.43 (0.19–0.97)	0.25 (0.07–0.82)	0.26 (0.08–0.88)
WWOX, Neg vs. Pos	1.83 (1.01–3.28)	0.56 (0.31–1.03)	1.13 (0.53–2.40)	1.02 (0.47–2.19)
Subtype, BLCs vs. non-BLCs	0.36 (0.19–0.68)	0.33 (0.17–0.63)	0.45 (0.21–0.98)	0.43 (0.19–0.97)

AbbreviationsHR hazard ratio, CI confidence interval, BLCs basal-like cancer subtype, Pre premenopausal, Post, postmenopausal, Neg negative, Pos positive

Discussion

Since outcomes of different breast cancer subtypes may be influenced by their underlying biologic features, efforts have been directed at the identification of specific gene expression profiles of these subsets of breast cancer. Analysis of breast cancer tissue with DNA microarrays has enabled identification of five subtypes of breast cancer with distinct gene expression patterns (Perou et al. 2000; Sorlie et al. 2003; Livasy et al. 2006). Currently, a widely used definition for identifying basal-like subtype is by routine immunohistochemical staining (positive for CK5/6 and/or EGFR and negative for ER, PR, and HER-2/neu) (Nielsen et al. 2004). Both the basal-like group and the ER negative, HER2-positive group are associated with the poorest short-term survival rates (Jimenez et al. 2001; Sorlie et al. 2003; Kreike et al. 2007), although there is no evidence that these patients are at higher risk of local relapse after conservative surgery and radiation (Haffty et al. 2006). Basal-like breast cancers are characterized by the presence of markers that are expressed in normal basal and myoepithelial cells, and they form 56–85% of the triple-negative group (Nielsen et al. 2004; Livasy et al. 2006). Recently, in a heterogeneous cohort of conservatively managed breast cancer patients, Parikh et al. (2008a, b) demonstrated that the triple-negative phenotype is a significant independent prediction of distant metastasis and poorer cause-specific survival.

Since the basal-like phenotype is not amenable to any form of endocrine therapy, efforts are underway to develop appropriate targeted therapies. Greater understanding of the pathologic and molecular characteristics of this phenotype may lead to better tailored treatments for these patients. We characterized triple-negative tumors with additional immunohistochemical staining for CK5/6 and EGFR to determine what proportion of the triple-negative tumors had the basal-like phenotype. This simple definition has advantages such as ease of application, reproducibility, and availability of these antibodies in histopathology laboratories. Of 52 triple-negative tumors identified, 40 (77%) were CK5/6 and/or EGFR positive. Our results are similar to other reports and are consistent with estimates that 80 to 90% of triple-negative tumors are basal-like (Kreike et al. 2007; Rakha et al. 2007, 2008).

We have previously demonstrated that the triple-negative phenotype of breast carcinomas is significantly associated with osteopontin (OPN) overexpression (Wang et al. 2008). Additionally, studies have shown that the basal-like/triple-negative subtype is also associated with P53 overexpression (Tuck et al. 1997; Turner et al. 2004; Parikh et al. 2008a, b). These data indicate that growth signals may differ in the aggressive basal-like/triple-negative phenotype of breast cancer and raise questions as to what drives those signals to explain the disparate behavior of the tumors and poor survival of patients with this phenotype. The WWOX (WW domain-containing oxidoreductase) gene is located at 16q23.3-24.1 (Yendamuri et al. 2003), a region that spans the second most common human fragile site, FRA16D (Bednarek et al. 2000). Abnormalities affecting WWOX at the genomic and/or expression level(s) have been reported in numerous neoplasias and cancer-derived cell lines including breast, bladder, lung, gastric, hepatocellular, and prostate malignancies (Iliopoulos et al. 2005; Nunez et al. 2005b; Kuroki et al. 2002, 2004; Paige et al. 2001; Yendamuri et al. 2003; Aqeilan et al. 2004; Park et al. 2004). It has been observed that ectopic WWOX expression is able to inhibit anchorage independent growth and in vivo tumorigenicity of highly aggressive breast carcinoma lines, suggesting a putative tumor suppressor role for this novel protein (Bednarek et al. 2001). Furthermore, WWOX promoter methylation status is associated with lack of expression and is also inactivated in breast cancers by DNA methylation of the regulatory region (promoter and exon 1) (Iliopoulos et al. 2005, 2007). Recently, Iliopoulos et al. reported that restoration of endogenous WWOX protein expression in breast cancer cells by 5-aza-2′-deoxycytidine (5-Aza-CdR), the first demethylation agent with a well-characterized mechanism of action suppressed cancer cell growth (Iliopoulos et al. 2007), suggesting this drug has a promising tumor suppressor-reactivating action and could potentially be used in the clinic as an anticancer treatment, most likely to increase the efficacy of current biological or chemotherapeutic treatments.

In the current study, we demonstrated that the basal-like phenotype of breast cancer was significantly inversely associated with WWOX expression. Mean WWOX levels were significantly lower in the basal-like breast cancers than in the non-basal-like subtype (P = 0.01). Furthermore, we have provided evidence that lack of WWOX expression is not only associated with the basal-like subtype, but its level of expression directly correlates to clinical TNM stages of breast cancer. Thus, aberrant expression of WWOX may contribute to the aggressive behavior and worse prognosis in patients with the basal-like subtype.

Sex steroid-induced activation of WWOX is independent of ER, and this activation positively correlates with cancerous progression of breast tissue to a premetastatic state (Nunez et al. 2005b). Furthermore, a significant association has also been detected between WWOX expression and negative PR status in ovarian carcinoma. These reports suggest that the relationship between WWOX and sex steroid status may be complicated and beyond understanding at present. In the current study, we demonstrated that a significant correlation existed between WWOX expression and ER status in breast carcinomas wherein ER-negative breast carcinomas also lacked WWOX expression. This is also in agreement with observations from other groups (Guler et al. 2004; Nunez et al. 2005b) and strengthens the hypothesis that WWOX plays a role in sex steroid metabolism and likely functions in a pathway or pathways associated with the estrogen–progesterone axis. Furthermore, Guler et al. (2007) recently reported that reduced WWOX expression was better than PR status in prediction of tamoxifen resistance, especially in high-risk patients with breast carcinoma.

Another noteworthy finding of the current study is that there is a relationship between negative WWOX expression and postmenopausal status of patients with breast carcinoma, indicating that WWOX expression may be directly or indirectly associated with the level of expression of steroid hormones. Guler et al. (2004) reported that the extent of WWOX staining in normal tissue adjacent to the cancerous site was related to menopausal status. To our knowledge, this is the first study to demonstrate the association between WWOX expression of breast tumor tissues and menopausal status of patients. It has previously been demonstrated that estradiol concentrations are much higher in breast carcinomas compared to normal breast tissue and also higher than the corresponding circulating level of this hormone in postmenopausal women. Based on the hypothesis that the higher tumor estradiol levels were the result of in situ synthesis (Purohit et al. 2002), the WWOX protein found mostly in ER-positive tumors may be associated with the tumor in situ estradiol synthesis. Further studies are necessary to evaluate the relationship between different steroid hormones and WWOX expression.

It has been reported that subsets of ovarian tumors demonstrate loss of WWOX expression, which is potentially associated with patient outcome. However, the prognostic significance of WWOX expression in patients with breast carcinoma has not been previously reported. In spite of the relatively short follow-up time and limited size of the test group, it is notable that we were able to demonstrate WWOX to be of significant prognostic value for DFS in univariate analysis, and WWOX negativity was of borderline significance for DFS in the multivariate model. The role of WWOX in regulating cancer progression is the subject of intense investigation, and targeting WWOX might be an appropriate therapeutic strategy for the treatment of cancer.

The retrospective nature and relatively small sample size presents potential limitations with respect to the interpretation of this study. However, to our knowledge, we are the first to show the expression of WWOX protein in the basal-like phenotype and the prognostic value of WWOX expression in patients with breast cancers. Additionally, the association of WWOX with ER status and menopausal status strengthens the hypothesis that WWOX plays a role in sex steroid metabolism. The fact that lack of WWOX expression more commonly occurred in patients at advanced stages raised the possibility that WWOX inactivation may play an important role in not only early carcinogenesis but also tumor development. Thus, lack of WWOX expression identifies a subset of patients with a significantly higher risk of local and distant relapse. This could potentially be due to an increased frequency of loss of WWOX in higher stage tumors, which are associated with a poorer outcome, or alternatively due to lack of adjuvant endocrine therapy. Therefore, the association of WWOX with ER-negative expression, postmenopausal status, and advanced TNM stage may contribute to the worse patient outcome. Additional prospective and retrospective studies are warranted to better understand the role of WWOX expression, to further refine prognosis, and to optimize treatment in patients with basal-like breast cancer.