Association of a cytarabine chemosensitivity related gene expression signature with survival in cytogenetically normal acute myeloid leukemia

// Han Yan 1, 2, 4 , Lu Wen 3 , Dan Tan 1, 2, 4 , Pan Xie 1, 2, 4 , Feng-mei Pang 1, 2, 4 , Hong-hao Zhou 1, 2, 4 , Wei Zhang 1, 2, 4 , Zhao-qian Liu 1, 2, 4 , Jie Tang 1, 2, 4 , Xi Li 1, 2, 4 , Xiao-ping Chen 1, 2, 4 1 Department of Clinical Pharmacology, Xiangya Hospital, Central South University, Changsha 410008, P. R. China 2 Institute of Clinical Pharmacology, Central South University, Hunan Key Laboratory of Pharmacogenetics, Changsha 410078, P. R. China 3 Department of Diagnostic Radiology, Hunan Cancer Hospital, Changsha 410013, Hunan, P. R. China 4 Hunan Province Cooperation Innovation Center for Molecular Target New Drug Study, Hengyang 421001, P. R. China Correspondence to: Xiao-Ping Chen, email: chenxp74@hotmail.com Xi Li, email: lixi6931430@126.com Keywords: cytarabine, chemosensitivity, acute myeloid leukemia, prognosis, signature Received: August 08, 2016      Accepted: November 15, 2016      Published: November 26, 2016 ABSTRACT The prognosis of cytogenetically normal acute myeloid leukemia (CN-AML) varies greatly among patients. Achievement of complete remission (CR) after chemotherapy is indispensable for a better prognosis. To develop a gene signature predicting overall survival (OS) in CN-AML, we performed data mining procedure based on whole genome expression data of both blood cancer cell lines and AML patients from open access database. A gene expression signature including 42 probes was derived. These probes were significantly associated with both cytarabine half maximal inhibitory concentration values in blood cancer cell lines and OS in CN-AML patients. By using cox regression analysis and linear regression analysis, a chemo-sensitive score calculated algorithm based on mRNA expression levels of the 42 probes was established. The scores were associated with OS in both the training sample ( p =5.13 × 10 -4 , HR=2.040, 95% CI: 1.364-3.051) and the validation sample ( p =0.002, HR=2.528, 95% CI: 1.393-4.591) of the GSE12417 dataset from Gene Expression Omnibus. In The Cancer Genome Atlas (TCGA) CN-AML patients, higher scores were found to be associated with both worse OS ( p =0.013, HR=2.442, 95% CI: 1.205-4.950) and DFS ( p =0.015, HR=2.376, 95% CI: 1.181-4.779). Results of gene ontology (GO) analysis showed that all the significant GO Terms were correlated with cellular component of mitochondrion. In summary, a novel gene set that could predict prognosis of CN-AML was identified presently, which provided a new way to identify genes impacting AML chemo-sensitivity and prognosis.