Identification of a prognostic signature based on copy number variations (CNVs) and CNV-modulated gene expression in acute myeloid leukemia

Objectives: Acute myeloid leukemia (AML) is caused by multiple genetic alterations in hematopoietic progenitors, and molecular genetic analyses have provided useful information for AML diagnosis and prognostication. This study aimed to integratively understand the prognostic value of specific copy number variation (CNV) patterns and CNV-modulated gene expression in AML. Methods: We conducted integrative CNV profiling and gene expression analysis using data from the Therapeutically Applicable Research To Generate Effective Treatments (TARGET) and The Cancer Genome Atlas (TCGA) AML cohorts. CNV-related genes associated with survival were identified using the TARGET AML cohort and validated using the TCGA AML cohort. Genes whose CNV-modulated expression was associated with survival were also identified using the TARGET AML cohort and validated using the TCGA AML cohort, and patient bone marrow samples were then used to further validate the effects of CNV-modulated gene expression on survival. CNV and mRNA survival analyses were conducted using proportional hazards regression models (Cox regression) and the "survminer" and "survival" packages of the R Project for Statistical Computing. Genes belonging to the Kyoto Encyclopedia of Genes and Genomes (KEGG) cancer panel were extracted from KEGG cancer-related pathways. Results: One hundred two CNV-related genes (located at 7q31-34, 16q24) associated with patient survival were identified using the TARGET cohort and validated with the TCGA AML cohort. Among these 102 validated genes, three miRNA genes (MIR29A, MIR183. and MIR335) were included in the KEGG cancer panel. Five genes (SEMA4D, CBFB, CHAF1B, SAE1, and DNMT1) whose expression was modulated by CNVs and significantly associated with clinical outcomes were identified, and the deletion of SEMA4D and CBFB was found to potentially exert protective effects against AML. The results of these five genes were also validated using patient marrow samples. Additionally, the distribution of CNVs affecting these five CNV-modulated genes was independent of the risk group (favorable-, intermediate-, and adverse-risk groups). Conclusions: Overall, this study identified 102 CNV-related genes associated with patient survival and identified five genes whose expression was modulated by CNVs and associated with patient survival. Our findings are crucial for the development of new modes of prognosis evaluation and targeted therapy for AML.