Review

ERα is a useful but imperfect predictive marker of endocrine therapy response in breast cancer. The search for additional markers is an extensive research area, and excitement in this regard was heightened with discovery of ERβ. However, its role in breast cancer remains unknown. Data in the literature are contradictory and may be due in part to differential expression of ERβ variant proteins. One particular variant, ERβcx, also known as human ERβ2 was suggested as a dominant-negative inhibitor of ERα [1], and has been identified at both the RNA and protein levels in breast cancer. Due to alternative splicing, ERβcx/2 mRNA encodes a non-ligand binding ERβ variant protein identical to the long form of ERβ1 up to amino acid 468 followed by 27 amino acids that are distinct from those in ERβ1. The current paper uses western blotting and immunohistochemistry (IHC) to investigate expression of ERβcx in normal and neoplastic human breast tissue, and correlates expression with clinical outcome and response to endocrine therapy. Western blotting and IHC used sheep antibodies raised to a peptide specific for ERβcx. ERβcx positivity was defined as presence or absence of an appropriately sized immunoreactive protein by western blotting (although no examples of samples classified as negative on western blotting were indicated). In a subset of tumours (n = 23) western blotting data were correlated with IHC. IHC was scored by addition of staining intensity score and proportion of cells staining. No indication of staining location (nuclear or cytoplasmic or both) was given. Using IHC scores >2 as ERβcx positive, there was >80% correlation of IHC and western blot results.

The key findings are:

• ERβcx status as defined categorically (+/−) by western blotting in invasive breast cancers (n = 73) was not significantly associated with any clinicopathological features of the primary tumours.
• Positive ERβcx status of primary tumours defined categorically as above was not associated with time to progression but was significantly associated with a longer survival rate than those patients whose tumours were ERβcx negative.
• In a small group of patients (n = 23) who received endocrine therapy, positive ERβcx status again defined categorically as above was significantly associated with a better response to endocrine therapy.

The authors suggest, that although the mechanism is unclear, these preliminary data support a link between ERβcx expression and response to endocrine therapy, but a larger study is required to confirm the observation. To date there are three published studies, including the current one, where a specific antibody to ERβcx/2 was used to investigate protein expression in breast tissue in vivo. Two of these studies [2,3] used the same antibody (raised in sheep to ERβcx C-terminal specific peptide) and the third study [4] used a different antibody (raised in rabbits to the same ERβcx C-terminal peptide). Although different criteria for assigning ERβcx positivity and negativity were used similar frequencies of ERβcx positivity (48% [2], 52% [3], 54% [4]) in invasive human breast cancers were obtained. As well in the two studies [2,4] that compared frequency of ERβcx positivity in normal/benign breast tissue, both identified a significantly reduced frequency of ERβcx positivity in normal/benign tissues compared to tumours. These results correlate with data obtained at the RNA level [4,5]. Despite these similarities, in the two studies [2,3] where ERβcx positivity and endocrine therapy response were assessed, opposite conclusions were drawn. However, conclusions from these studies must be underscored with caution since there are distinct differences between them, and both suffer from extremely small sample sizes of n = 23 [2] and n = 18 [3], respectively. The authors argue that differences observed between the two studies, were due (a) to the different techniques employed (immunoblotting vs. IHC) although similar frequencies of ERβcx positivity and negativity in the populations studied could argue against this being a significant issue; and (b) the different populations studied. In study [3] the population was newly diagnosed patients treated neoadjuvantly for 3 months before tumour removal and response of the primary tumour to treatment was assessed, compared to the current study [2] where surgery had generally been undertaken and the patients identified to have locally advanced or metastatic disease. In addition the type of therapy (neoadjuvant vs. neoadjuvant and palliative; numbers per group not given) and length of follow-up (3 months vs. >3 months to years; median times of follow-up not given) were quite different between the two studies.

The findings of the current paper are potentially important. But clearly, larger patient cohorts and standardized methodologies need to be undertaken before robust conclusions can be drawn. If these data are confirmed in a larger study, then a new marker, ERβcx, for determining endocrine therapy response would be established, and its value alone or together with other previously established markers should be evaluated using appropriate clinical trials material together with standardized methodologies. This would in turn support extensive investigation of functional roles of ERβcx in mediating sensitivity to endocrine therapy to identify possible novel targets/approaches for therapy.