Correlation between RhoB and Estrogen Receptor α (ERα) Expressions: From Experimental Data to Human Breast Tumors.

Abstract

Abstract Introduction: Hormonotherapy (HT) of breast cancers (BC) is strongly limitated by the resistances. Because Rho proteins are key elements in the cross-talks between ERα and growth factors signaling, they may be involved in the response to HT. RhoB could play a negative role in oncogenesis but there is no valid data regarding RhoB expression in BC. Experimental results: Our data suggest that RhoB participates in the balance of expression of ERα. The inhibition of the expression of RhoB using two sequences of interfering RNA in hormonodependent cell lines (MCF-7, T47D, ZR75) and in anti-estrogen resistant cell lines (LCC2, LCC9), decreases the expression of ERα both at the protein and mRNA levels. Moreover, in Mouse Embryonic Fibroblasts and uteri collected from RhoB knock-out mice, a dramatic decrease of ERα expression is observed. We therefore investigated the expression level of RhoB and ERα in BC tumors. Patients and methods: A tissue microarray (TMA) was constructed from a cohort of 113 patients (pts) enrolled in a randomized trial for adjuvant tamoxifen (median follow-up: 249.9 months). RhoB, ERα, and PR expressions were measured by immunochemistry. Cut-off used for ERα and PR + was 10% of stained cells. The expression of RhoB was calculated with the IRS score. Correlation of RhoB expression score with clinically diagnostic and prognostic variables was assessed using Man Whitney and Spearman's rank tests. Univariate survival analysis was performed for disease free survival (DFS) by applying the log-rank test to RhoB expression levels stratified by median value. TMA results: 65 (58.6%) were grade I-II; 74(66.1%) were ERα+ and 59(52.7%) PR+; 23 (22.1%) presented lymphovascular (LV) invasion. 39 pts (34.8%) had lymph nodes (LN) +. Pts under tamoxifen (n=62) had a less favorable pathological profile regarding + LN (p=0.0039), pathological tumour size (p=0.0486) and number of mitoses (p=0.0556). Age, ERα/PR status, histological grade or type and LV invasion status were similar in the two groups. We found less RhoB IRS expression in pts with tumor grade III (median 8 [1;12]) than in grade I-II (median 9 [3;12], p=0.0142). RhoB IRS expression was much higher in ERa + tumors (median:10.5 [3;12]) than in ERa negative tumors (median=8 [1;12], p=0.0023). RhoB IRS score was not correlated with the presence or not of LV invasion (p=0.26), neither with the presence of LN invasion (p= 0.74). Furthermore, RhoB expression is i) strongly correlated with the % of expression of ERα (Spearman'sρ=0.3659; p = 0.001) and PR (Spearman's ρ=0.2544 ; p=0.0076) ii) inversely correlated with histologic size (Spearman's ρ= -0.2344 ; p = 0.0166) and with number of mitose (Spearman's ρ=-0.2009 ; p=0.0362). The DFS of pts with ERα + tumors under tamoxifen or not was not affected by the level of RhoB expression. Conclusion: The analysis of 113 human breast tumors allowed to confirm experimental results, demonstrating that RhoB expression is strongly correlated with ERα expression. The role of RhoB as a potential suppressor gene of tumours in BC and its role in the response to HT have then to be investigated further. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5146.