Acute myeloid leukemia (AML) treatment response remains poorly understood. Although multiple studies have focused on understanding the transcriptomic and epigenetic landscape of AML, a genome-wide analysis of SNPs in pediatric AML has not yet been investigated in depth. Thus, we sought to identify genetic variants predictive of AML response, relapse, and survival in pediatric AML patients. For this study, we generated genome-wide SNP data patients (n=160) treated on the multicenter AML02 clinical trial (ClinicalTrials.gov Identifier: NCT00136084) using Infinium Omni 2.5M Exome Beadchip. Standard GWAS QC procedure was followed in order to remove SNPs with call rate < 95%, monomorphic SNPs, SNPs with MAF<5% and samples with call rate<95%. Following QC, a risk-adjusted multi-outcome integrative GWAS was performed to identify SNPs associated with minimal residual disease (MRD) following induction I, relapse-free survival (RFS) and overall survival (OS). We performed a risk-adjusted analysis to identify 21 SNPs mapping to 14 genes at an endpoint-integrative p value <2x10-5. Table 1 provides list of genes with SNPs significantly associated with MRD, RFS, OS as well as in the integrated analysis at <2x10-5. Of interest multiple SNPs in DICER1, which is a key enzyme required for the biogenesis of microRNAs and small interfering RNAs were significantly associated with clinical outcome with promise integrated analysis at p = 0.000011, supported by associations with MRD, RFS and OS at p <0.002 (Figure 1A). DICER1 is over-expressed in AML with its expression under the influence of hematopoietic transcript factor, GATA1. RAI14, a retinoic acid induced 14 is a prognostic marker of poor response in solid tumors and has been associated with development of drug resistance. Multiple SNPs in RAI14 were significantly associated with clinical endpoints. Figure 1B shows RAI14 SNP rs336474 with C allele significantly associated with better RFS (p= 0.027) and OS (p=0.007), with an integrated p= 0.000004. SNP in upstream of RBFOX1, a RNA binding fox-1 homolog 1 and within intron of GRIN2A, glutamate ionotropic receptor NMDA type subunit 2A were significantly associated with MRD, RFS and OS (all p<0.005) and integrated p =0.00001 (Figure 1C). SNPs within genes involved in pyrimidine metabolism such as UPP2, a uridine phosphorylase; tumor suppressor genes such as JPH3, which codes for junctophilin; LILRB4 which encodes for a Leukocyte Immunoglobulin Like Receptor B4, that regulates inflammatory responses and cytotoxicity; HACE1 a potential tumor suppressor involved in the solid tumors pathophysiology; ANK2, an ankyrin family protein with role in cell proliferation and motility; BIRC8, which is implicated in CML disease progression etc. In conclusion, our results demonstrate significance of genome-wide investigation of SNPs to identify novel and clinically relevant SNPs of prognostic significance in childhood AML. We will present the in depth results of our integrated GWAS analysis as well as validation in independent patient cohorts. In summary, our results constitute one of the first integrated GWAS analyses to identify SNPs of prognostic significance in pediatric AML.
Acknowledgments: We are thankful for funding from NIH R01-CA139246 and ALSAC.
No relevant conflicts of interest to declare.