ABSTRACT: Cell-cell communication proteins Eph and ephrin constitute the largest family of receptor tyrosine kinases (RTKs). They are distinguished by the fact that both receptors and ligands are membrane-bound, and both can drive intracellular signaling in their respective cells. Ever since these RTKs have been found to be involved in cancer development, strategies to target them therapeutically have been actively pursued. However, before this goal can be rationally achieved, the contributions of either Eph receptors or their ephrin ligands to cancer development and progression should be scrutinized in depth. To assess the clinical pertinence of this concern, we performed a systematic review and meta-analysis of the prognostic/predictive value of EphB2 and its multiple cognate ephrin ligands in breast cancer. We found that EphB2 has prognostic value, as indicated by the association of higher EphB2 expression levels with lower distant metastasis-free survival (DMFS), and the association of lower EphB2 expression levels with poorer relapse-free survival (RFS). We also found that higher EphB2 expression could be a prognostic factor for distant metastasis, specifically in the luminal subtypes of breast cancer. EFNB2 showed a marked correlation between higher expression levels and shorter DMFS. EFNA5 or EFNB1 overexpression is correlated with longer RFS. Increased EFNB1 expression is correlated with longer OS in lymph node (LN)-negative patients and the luminal B subtype. Higher levels of EFNB2 or EFNA5 are significantly correlated with shorter RFS, regardless of LN status. However, while this correlation with shorter RFS is true for EFNB2 in all subtypes except basal, it is also true for EFNA5 in all subtypes except HER2+. The analysis also points to possible predictive value for EphB2. In systemically treated patients who have undergone either endocrine therapy or chemotherapy, we found that higher expression of EphB2 is correlated with better rates of RFS. Bearing in mind the limitations inherent to any mRNA-based profiling method, we complemented our analysis with an immunohistochemical assessment of expression levels of both the EphB2 receptor and cognate ephrin ligands. We found that the latter are significantly more expressed in cancers than in normal tissues, and even more so in invasive and metastatic samples than in ductal carcinoma in situ (DCIS). Finally, in an in vitro cellular model of breast cancer progression, based on H-Ras-transformation of the MCF10A benign mammary cell line, we observed dramatic increases in the mRNA expression of EphB2 receptor and EFNB1 and EFNB2 ligands in transformed and invasive cells in comparison with their benign counterparts. Taken together, these data show the clinical validity of a model whereby EphB2, along with its cognate ephrin ligands, have dual anti- and pro-tumor progression effects. In so doing, they reinforce the necessity of further biological investigations into Ephs and ephrins, prior to using them in targeted therapies.
