Matched Exome Sequencing in Mothers and Infants with MLL-Negative Acute Leukemia

Infant leukemia (IL) is an extremely rare, sporadic, but often fatal, form of cancer that is defined as leukemia occurring within the first year of life. Unlike leukemia in older children where survival rates for acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL) are approximately 70% and 85%, respectively, infants have a survival rate of ∼50%. Unfortunately, despite years of research and clinical trials, overall survival for IL hasn't improved substantially since the advent of hematopoietic stem cell transplantation, and those infants that survive are often left with lifelong deficits in cognition, development, end organ function, behavioral milestones and/or other complications due to the intensity of treatment (Pui, NEJM 2003; Chow, J Pediatr 2007). The incidence rate of IL is increasing (Linabery, Cancer 2008) in the US, but cannot be attributed to chromosomal anomalies (Uckun, Blood 1998), environmental exposures (Ross, Epidemiol Rev 1994), or highly penetrant genetic polymorphisms alone. Clearly, a critical component of IL pathophysiology remains undiscovered. Rare variation has been implicated in a host of complex phenotypes and diseases, but the impact of rare germline variants on the etiology and outcome of IL has not been fully explored. The Rare Variant Hypothesis predicts that a population of affected individuals would harbor a diverse collection of functionally significant variants in genes involved in etiologically relevant pathways. Given this model, we hypothesize that the onset of IL requires heritable deleterious germline variants, which act alone or in combination with somatic mutations to induce leukemic transformation.

To explore this possibility, we completed a pilot exome sequencing project on 25 pairs of germline DNA from mothers and their infants with MLL-negative IL collected from the Children's Oncology Group (COG) AE24 “Epidemiology of Infant Leukemia” study. Twelve infants had ALL, 13 had AML and none of the mothers had cancer. We asked if a) infants harbored rare or novel deleterious germline variants in known leukemia genes, b) these variants were inherited from mothers and c) the genes affected by these germline variants fell in common pathways or share common regulatory mechanisms.

We found an average of 16,056 variants per exome with an average of 3,082 (19.2%) being novel. Comparing these results to the COSMIC database (http://www.sanger.ac.uk/genetics/CGP/cosmic/), infants with AML had novel, non-synonymous, deleterious germline variants in 82 genes associated with hematologic malignancies, infants with ALL had similar variants in 64 genes, and 42 additional genes (40%) overlapped between ALL or AML. For infants with ALL, 45% of these variants were inherited from healthy mothers compared to only 23% in AML infants. Presumably, the remaining variants were inherited from fathers, but without paternal DNA, we cannot exclude de novo germline mutation, although such mutations are exceedingly rare and would only account for a few non-maternal variants. We used the g:Profiler (http://biit.cs.ut.ee/gprofiler/) algorithm to determine if any of these genes acted in common leukemia-related pathways or shared regulatory mechanisms. We found that many candidate genes were regulated by microRNAs (MIR) previously implicated in cancer, including MIR10a, MIR29c, MIR291b-3p, MIR369-5p, MIR469, MIR519a, MIR721. Five MIRs have been associated with leukemia, four with acute leukemia. Two MIRs, 291b-3p and 721, are also associated with embryonic stem cell cycle regulation and apoptosis, respectively. Ongoing work is focusing on additional exome sequencing, epidemiologic analysis and in vitro functional studies.

It is clear that the incidence, clinical behavior and outcomes of IL cannot be explained fully through either environmental exposures or somatic mutations alone. We are leveraging the largest epidemiologic study of IL to date to explore the intersection of functional congenital genetic variation, clinical outcomes and maternal prenatal/pregnancy exposures to augment our understanding of IL. A better understanding of the natural history of IL will aid in future recommendations for pre- and post-natal genetic diagnostics, risk stratification of affected infants and ultimately therapeutic decisions.

No relevant conflicts of interest to declare.