Survival analysis of Heat Shock Protein on Breast Cancer with focus on molecular subtypes

GUERRERO GIMENEZ, MARTIN EDUARDO; ZOPPINO FELIPE CARLOS MARTIN; CASTRO, GISELA NATALIA; CIOCCA, DANIEL RAMON

Mar del Plata

Congreso; LXI Reunión Científica anual de la Sociedad Argentina de Investigación Clínica; 2016

Sociedad Argentina de Investigación Clínica (SAIC)

Aim: To analyze the mRNA levels of the 95 Heat Shock Proteins (HSP) family members inbreast cancer subtypes (Luminal A, Luminal B, HER2-enriched, Basal-like and Normal-like) and their clinical correlations. Methods: The data used in this study was programmatically extracted from the publiclyavailable data set of mammary adenocarcinoma from The Cancer Genome Atlas Project(TCGA). Standardized and non-standardized gene expression levels from 1097 tumorsamples and 114 normal breast tissue data available in the RNASeqV2 platform wereobtained.Given the expression levels of 50 different specific genes PAM50 classification wasperformed to classify the samples in five intrinsic molecular subtypes. To assess differentialgene expression (DGE) between normal tissue and tumor samples we implementedDESeq2 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 methodfor determining whether changes in gene expression are significantly associated withsurvival. Kaplan Meier curves for each clinically significant HSP were generated to analyzeoverall survival. To assess the effect of several known prognostic predictors (ER, lymphnode status, age, etc.) Cox proportional hazard ratio multivariate analysis was performed.Results: We have found HSP clusters that are specifically down-regulated while othersappeared specifically up-regulated in breast cancer subtypes. After analyzing HSP levelswith survival, we identified novel HSPs that show correlations with the clinical outcome ofthe 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. NovelHSPs not previously related with breast cancer have been associated to this disease. Thecomprehensive map addresses the down-regulation and up-regulation of the HSPs involvedwith these subtypes