CONICET | Buscador de Institutos y Recursos Humanos

Aim: To analyze the clinical correlations of mRNA levels of the 95 Heat Shock Proteins (HSP) family members in breast cancer subtypes (Luminal A, Luminal B, HER2-enriched, Basal-like and Normal-like).Methods: The data used in this study was programmatically extracted from the publicly available data set of mammary adenocarcinoma from The Cancer Genome Atlas Project (TCGA). Standardized and non-standardized gene expression levels from 1097 tumor samples and 114 normal breast tissue data available in the RNASeqV2 platform were obtained.Given the expression levels of 50 different specific genes PAM50 classification was performed to classify the samples in five intrinsic molecular subtypes. To assess differential gene expression (DGE) between normal tissue and tumor samples we implemented DESeq2 analysis were log2 Fold change values were obtained associated with exact p-values and False Discovery Rate values (FDR). SAMseq was used as a screening method for determining whether changes in gene expression are significantly associated with survival. Kaplan Meier curves for each clinically significant HSP were generated to analyze overall survival. To assess the effect of several known prognostic predictors (ER, lymph node status, age, etc.) Cox proportional hazard ratio multivariate analysis was performed.Results: We have found HSP clusters that are specifically down-regulated while others appeared specifically up-regulated in breast cancer subtypes. After analyzing survival depend on HSP levels, we identified novel HSPs that show correlations with the clinical outcome of the cancer patients.Conclusions: We report the complexity of the expression of the HSPs in breast cancer, reporting in a large dataset their expression levels according to the genetic subtypes. Novel HSPs not previously related with breast cancer have been associated to this disease. The expression level of certain HSPs are in relation with the survival of patients.