miRNA signatures of prognostic significance in single hormone receptor-positive breast cancer

e12544

Background: Single hormone receptor-positive breast cancer – ER+/PgR– and ER–/PR+, is a distinct entity with vastly undiscovered biology. This study aimed at exploring miRNA profiles of primary tumors of single hormone receptor-positive phenotype and their association with overall survival (OS). Methods: The study group comprised 32 breast cancer patients, collected at 3 Polish centers, with single hormone receptor-positive phenotype, including 14 ER+/PgR– and 18 ER–/PR+ cases, thoroughly characterized for ER (clones: 1D5, EP1 and SP1) and PgR (clone: 636) expression. Expression of 798 miRNAs was profiled using nCounter Human v3 miRNA Expression Assay (NanoString) in each primary tumor sample (FFPE), data available at NCBI GEO (GSE155362). Normalized miRNAs counts were analyzed for the association with OS. Gene targets of miRNAs showing prognostic significance were identified using miRNET 2.0, while their annotation with Gene Ontology Biological Processes was determined with DAVID Bioinformatics Resources 6.8 tool. The results were validated using MINDACT trial dataset [Cardoso et al. 2016]: single hormone receptor-positive (central assessment) cases of ductal histology (including 392 ER+/PgR– and 15 ER–/PR+ cases) where analysed for the prognostic value of gene expression of the identified miRNAs targets. Results: In our cohort, 7 miRNAs showed an association with OS (log-rank test): positive for four: hsa-miR-25-3p (p=0.013), hsa-miR-150-5p (p=0.015), hsa-miR-148a-3p (p=0.038) and hsa-miR-497-5p (p=0.044); and negative for three: hsa-miR-517c-3p (p=0.017), hsa-miR-4284 (p=0.018) and hsa-miR-301a-5p (p=0.029). According to miRNET 2.0, the 4 and 3 prognostic miRNAs have 3801 and 544 gene targets, respectively. The genes targeted by positively associated miRNAs were involved in transcription regulation, vesicle-mediated transport and protein stabilization, while the targets of negatively associated miRNAs were related with mRNA polyadenylation, DNA replication and regulation of DNA-templated transcription. Next, genes targeted by most of miRNAs from either group were identified: 11 targets of positive miRNAs (CANX, CBX5, HIPK1, PDE4DIP, PPM1A, REL, SETD5, SP1, XYLT2, ZMAT3, ZNF460) and 6 targets of negative miRNAs (ARL10, DNAJC28, MTHFD1L, PRPF6, RHOF, TSPAN6). The genes were tested for the association with OS in single hormone receptor-positive subgroup of the MINDACT cohort, with the following significant findings in multivariate analysis including N stage and Ki67 status: PDE4DIP (HR=2.84, p=0.011), CBX5 (HR=2.08, p=0.026) and PRPF6 (HR=0.44, p=0.045). Conclusions: This study identifies 7 miRNAs and their gene targets with a potential prognostic significance in single hormone receptor-positive breast cancer. Thus, several miRNA-mRNA axes merit further investigation both at molecular level and in a separate patient cohort to validate their clinical utility.

Abstract P5-04-01: Expression of epithelial-mesenchymal transition (EMT)-related markers in primary tumors and matched lymph node metastases in breast cancer patients

Background. Malignant tumors may include multiple cancer cell populations with various metastatic potential. More aggressive subpopulations might easier be captured in lymph nodes metastases (LNM) than in primary tumors (PT). We evaluated mRNA and protein levels of master EMT regulators: TWIST1, SNAIL and SLUG, protein levels of EMT-related markers: E-cadherin, vimentin, and expression of classical breast cancer receptors: HER2, ER and PR in PT and corresponding LNM. The results were correlated with clinicopathological data and patient outcomes.

Methods. Formalin-fixed paraffin-embedded (FFPE) samples from PT and matched LNM from 42 stage II-III breast cancer patients were examined. Expression of TWIST1, SNAIL and SLUG was measured by RT-qPCR, with median as cut off for positive result. Protein expression was examined by immunohistochemistry on tissue microarrays. For TWIST1, SNAIL, SLUG, E-cadherin and vimentin >10% of positively stained cells defined positive result. ER and PR were scored according to Allred system, and HER2 - according to HercepTest criteria. Results were considered concordant if PT and LNM were both positive or both negative. Concordance was measured by estimating Cohen's kappa coefficient (κ), with κ value equal 1 indicating perfect agreement. Disease-free survival (DFS) and overall survival (OS) were compared using F-Cox test. Hazard ratios (HRs) with 95% confidence intervals (95% CI) were computed using Cox regression analysis.

Results. On average, expression of TWIST1, SNAIL and SLUG was significantly higher in LNM compared to PT (P < 0.00001 for all). Gene and protein levels of TWIST1, SNAIL and SLUG were highly discordant between PT and matched LNM (Tab 1). Receptor conversion, particularly loss of ER and PgR, occurred in 18% and 29% of cases, respectively. Increased expression of TWIST1 and SNAIL in LNM was associated with shorter OS (p = 0.045 and p = 0.022, respectively) and DFS (p = 0.02 and p = 0.01, respectively), whereas their expression in PT had no prognostic impact. Negative-to-positive switch of TWIST, SNAIL and vimentin correlated with shorter OS and DFS. Conversion of ER and PR had no significant impact on survival.

Conclusions. LNM are enriched in cells with more aggressive phenotype, marked by elevated levels of EMT regulators. High expression of TWIST1 and SNAIL in LNM, as well as negative-to-positive conversion of both TWIST1 and SNAIL confer worse prognosis, confirming the correlation of EMT with aggressive disease behavior.