Background: Several classification systems are available to assess pathological response to neoadjuvant chemotherapy in breast cancer, but reliable biomarkers to predict the efficiency of primary systemic therapy (PST) are still missing. Deregulation of gap junction channel forming connexins (Cx) has been implicated in carcinogenesis and tumour progression through loss of cell cycle control. In this study we correlated Cx expression and cell proliferation with disease survival and pathological response to neoadjuvant chemotherapy in breast cancers using existing classification systems.Methods: The expression of Cx26, Cx32, Cx43, Cx46 and Ki67 was evaluated in 96 breast cancer patients prior to and after neoadjuvant chemotherapy using duplicate cores in tissue microarrays (TMA). Cx plaques of <1μm were detected with multilayer, multichannel fluorescence digital microscopy. Current classifications to assess residual tumour burden after primary systemic therapy included the EWGBSP, CPS-EG, Miller-Payne, Sataloff and NSABP systems.Results: In our cohort dominated by hormone receptor (ER/PR) positive and HER2 negative cases, only the CPS-EG classification showed prognostic relevance: cases with scores 1–2 had significantly better overall survival (p=0.015) than cases with scores 3–5. Pre-chemotherapy Cx43 expression correlated positively with hormone receptor status both before and after chemotherapy and had a negative correlation with HER2 expression pre-chemotherapy. There was a positive correlation between Cx32 and HER2 expression pre-chemotherapy and between Cx32 and Ki67 expression post-chemotherapy. A negative correlation was found between post-chemotherapy Cx46 and Ki67 expression. Decreased post-chemotherapy Cx26 expression (20%) pre- and post-chemotherapy correlated with significantly better survival in the intermediate prognostic subgroups of EWGBSP TR2b (p_pre-chemo=0.006; Sataloff TB (p_pre-chemo=0.005; p_post-chemo=0.029) and in Miller-Payne G3 (p_pre-chemo=0.002; p_post-chemo=0.012) classifications. Pre-chemotherapy, Cx46 expression was the only marker that correlated with overall survival within these subgroups.Conclusion: Our results suggest that Cx46 and Cx26 expression in breast cancer may improve the assessment of pathological response and refine intermediate prognostic subgroups of residual tumour classifications used after neoadjuvant chemotherapy.

Background: Expression of transcription-factors as Slug and Sox9 was recently described to determine mammary stem-cell state. Sox10 was previously shown to be present also in breast cancer. Protein overexpression of Slug, Sox9 and Sox10 were associated with poor overall survival and with triple-negative phenotype in breast cancer. In this study we tested the stability of Slug, Sox9 and Sox10 expression during chemotherapy and addressed their prognostic role of in neoadjuvant treated primary breast-cancer and their correlation to pathological-response and overall survival. Methods: We analyzed immunohistochemical expression of Slug, Sox9 and Sox10 in tissue microarrays of 96 breast cancers prior to and after neoadjuvant chemotherapy. Expression was evaluated in invasive tumor cells and in tumor stroma and scored as 0, 1+, 2+ 3+. Expression-profile prior to and after chemotherapy was correlated to overall survival (Kaplan Meier) and with established clinico-pathological parameter. Results: Sox9, Sox10 and Slug were expressed in 82–96% of the tumor cells prior to chemotherapy. Slug was expressed in 97% of the cases in tumor stroma before therapy. Change in expression-profile after chemotherapy occurred only in Slug expression in tumor-cells (decreased from 82 to 51%, p = 0.0001, Fisher’s exact test). The other markers showed no significant change after chemotherapy. Stromal Sox9 expression (0 to 2+) correlated to better overall survival after chemotherapy (p = 0.004) and reached almost statistical significance prior to chemotherapy (p = 0.065). There was no correlation between Sox9 and hormone-receptor expression. In multivariate-analysis, the stromal Sox9 expression after chemotherapy proved to be an independent and better prognostic marker than hormone-receptor status. Other clinico-pathological parameter (as HER2-status or pathological-stage) showed no correlation to the analyzed markers. Conclusion: Strong stromal Sox9 expression in breast cancer after chemotherapy was found to bear negative prognostic information and was associated with shortened overall survival. Slug expression was significantly changed (reduced) in samples after neoadjuvant chemotherapy.