28P - Germ cells and mammary gland transcritome comparison generate a powerful gene expression profile that predicts breast cancer survival

Date 07 May 2015
Event IMPAKT 2015
Session Welcome reception and Poster Walk
Topics Breast Cancer
Germ Cell Tumours
Translational Research
Presenter Daniel Tiezzi
Citation Annals of Oncology (2015) 26 (suppl_3): 10-14. 10.1093/annonc/mdv116
Authors D.G. Tiezzi1, H. Noushmehr2, W.A. Silveira1, T.S. Sabedot2, J.M. Andrade1
  • 1Gynecology And Obstetrics, School of Medicine of Ribeirao Preto - USP, 14048-900 - Ribeirao Preto/BR
  • 2Genetics, School of Medicine of Ribeirao Preto - USP, Ribeirao Preto/BR

Abstract

Body

Background: Predict clinical outcome and chemotherapy sensitivity in breast cancer remains a challenging task. Germ cells are the most undifferentiated cells in the human body. We aimed to identify breast cancer genetic signatures in order to predict survival.

Methods: An in silico analysis using publicly available breast cancer datasets (GSE27715, GSE54126, GSE1456 and TCGA) was carried out. Genes down regulated in mouse mammary gland compared to mouse germ line cells were selected using the t test comparison (n= 1096 genes). Using the expression2kinases tool (http://www.maayanlab.net/X2K), the top100 transcription factors (TFs) and kinase encoding genes were selected. After RMA normalization the 200 genes was selected from the GSE1456 dataset and a ROC analysis was performed based on each gene expression to whether the patient status was ‘living or deceased’ as an outcome. Genes with the area under the curves >0.65 were selected (n = 29 genes) and were median centered. The best number of clusters were defined by the consensus clustering analysis. Survival analysis was performed using the Kaplan-Meier curve comparing the clusters defined by the consensus. The expression and methylation level of the 29 genes was analyzed in TCGA database.

Results: The 29 genes signature discriminate into five different tumor clusters. The C2 has the worst prognosis, is characterized by high expression of four TFs (SOX2, THAP11, PPARD, E2F1) and ten kinase encoding genes (AURKA, CHEK1, SRC, MAP2K1, MAPK14, CDK1, CDK2, CDK5, PRKDC, PLK1) and are comprised mostly by basal and luminal B tumors. The C4 patients have a good prognosis with low expression of those TFs and kinase encoding genes. Most of C4 tumors are luminal A or normal like. The PRKDC and PLK1 genes have a lower methylation level in basal tumors.

Conclusion: A small set of genes derived from the comparison of transcriptome analysis between mammary gland and germ cells is able to select breast cancer patients with distinct prognosis. The CHEK1, AURKA, PLK1 and PRKDC expression and PRKDC and PLK1 methylation level in aggressive breast tumors suggest the DNA damage response and repair is reactivate conferring an undifferentiated state.

Disclosure: All authors have declared no conflicts of interest.