Abstract
Therapy-related acute myelogenous leukemia (t-AML) is an important late adverse effect of alkylator chemotherapy. Susceptibility to t-AML has a genetic component, yet the specific genes and genetic variations that influence susceptibility are poorly understood. Our lab previously identified mouse strains that are susceptible (SWR/J) or resistant (C57BL/6J and C3H/HeJ) to t-AML induced by the alkylator, ethyl-N-nitrosourea (ENU). To study the genetic basis of these differences, we performed an F2 intercross between susceptible and resistant strains. A single copy of the hCG-PML-RARa (PR) transgene was bred into each mouse. PR is an initiating factor for acute promyelocytic leukemia, but requires cooperating mutations for full leukemic transformation. This provides a platform to detect gene X transgene (PR) and gene X environment (ENU) interactions that promote leukemogenesis. F2 mice were treated (n=141) or untreated (n=141) with ENU and sacrificed and analyzed when moribund. We also analyzed untreated PR+ (n=24) mice from the resistant C57BL/6J X C3H/HeJ (B6C3F1) background. Untreated B6C3F1 PR+ mice developed lethal myeloid leukemia (characterized by splenomegaly > 0.25g, WBC > 30 K/uL, and increased immature myeloid precursors in PB, BM, and spleen) with an incidence of 12.5% and a latency of 234 days. By contrast, 79.4% of untreated PR+ F2 mice developed myeloid leukemia with a latency of 108 days and median survival of 238 days. The earlier onset and increased incidence of leukemia in F2 mice confirm that SWR/J alleles confer increased susceptibility to AML. ENU treatment further increased the leukemia incidence (90.4% vs. 79.4%, p<0.0001 by logrank) and shortened the median survival (168 vs. 238 days) of PR+ F2 mice. F2 mice were genotyped using 357 informative SNPs across the genome to facilitate quantitative trait locus (QTL) mapping. QTL analysis was performed using leukemia-free survival, spleen weight, and WBC as quantitative traits. Because extramedullary hematopoeisis and increased WBC are markers of poor prognosis in human AML, we reasoned that identification of modifier alleles for these traits might also have potential clinical relevance. QTL analysis revealed five peaks associated with survival, 5 with spleen weight, and 3 with WBC. The 1 LPR (likelihood probability ratio) confidence intervals for the QTL range in size from 10 to 50 Mbp. Each region contains between 100 and 750 genes. The most significant peak (LPR=3.94) is a survival QTL on chromosome 1 from 93.4 to 120.5 Mbp that retains significance at the genome wide level. The QTL effect is large in ENU-treated animals but not discernible in untreated animals. Both the additive effect (−0.51, SE=0.17) and the dominant effect (0.62, SE=0.20) are significant at p<0.05. Genes with potential connections to leukemogenesis within this region include a serpin cluster, genes involved in apoptosis (Bcl2, Bok, Pdcd1 Phlpp), and cell cycle genes (Sept2 and Clasp1). Ongoing studies are focused on candidate gene evaluation and fine-mapping the QTL regions to identify the QTL genes and their variants. Validation of these genes in therapy-related leukemogenesis should provide a better understanding of t-AML susceptibility and lead to strategies that moderate t-AML risk in susceptible individuals.
Peridermal fruit skin formation in Actinidia sp. (kiwifruit) is associated with genetic loci controlling russeting and cuticle formation
