Taking a SNPshot of t-AML

In this issue of Blood, Ellis and colleagues report on the interaction of SNPs in the p53 tumor suppressor pathway and in the MDM2 309 locus in susceptibility to therapy-related AML.

Genetic variants are associated with disease susceptibility.^1,2  Of the genetic variants, single-nucleotide polymorphisms (SNPs) lend themselves to interrogation in population-based studies. As reported in this issue of Blood, Ellis and colleagues studied 2 separate, large cohorts of therapy-related acute myeloid leukemia (t-AML) patients. Because only a subset of all patients treated with cytotoxic chemotherapy or radiation develop t-AML, it was hypothesized that these individuals may be predisposed due to constitutional genetic variations in DNA damage-response pathways. Although several candidate genes have been previously implicated, Ellis et al focused on the p53 tumor suppressor pathway, as this transcription factor mediates cell-cycle arrest, cell senescence, and apoptosis, and is often lost or mutated in t-AML. They also examined a common SNP (SNP309) of MDM2, a ubiquitin E3 ligase, which negatively affects the stability of p53 and has been examined previously in other series of leukemia patients.³  As reported in their article, an arginine (Arg) at TP53 codon 72 predisposes cells to apoptosis, whereas a proline (Pro) mediates cell-cycle arrest. At the MDM2 SNP309, a G allele indicates high binding ability of SP1 transcription factor, which increases levels of MDM2, thereby decreasing p53 expression. In contrast, a T allele allows increased p53 function.

The cohort of t-AML patients from the University of Chicago was selected because of availability of Epstein Barr virus–transformed lymphoid lines from which DNA could be extracted, whereas peripheral blood DNA was available from the cohort of patients studied from the United Kingdom. A total of 171 cases were studied. It was found that neither p53 nor MDM2 variants by themselves were associated with t-AML risk, but there was an interaction that influenced susceptibility. The figure illustrates the models proposed for these interactive influences. Control cohorts were used to determine that there was not a bias in baseline frequencies of the SNPs, and this SNP interaction was not observed in de novo AML cases. The same interactive influence was noted in those treated with chemotherapy and those who acquired abnormalities of chromosomes 5 or 7. Only TP53 Pro/Pro was associated with increased risk of t-AML in those who received chemotherapy alone. No significant effects on disease latency were noted. The MDM2 TT genotype appeared to offer a protective effect in younger women.

Although this study used 2 relatively large cohorts of t-AML patients, it would have benefitted from the melding of differing methodologies and DNA sources (although these were well-controlled for to reduce bias, and genotype distributions appeared comparable between University of Chicago and United Kingdom control subjects). The 2 series used different means of case selection/identification and different treatment regimens. Also, given that multiple therapeutic regimens were utilized, the effect of SNP interaction as related to exposure to a single agent or combination regimen on development of t-AML could not be determined. Nonetheless, this study demonstrates that interrogating biologically rational interactions between SNPs may be important in determining the risk of susceptibility to disease. Such interactions might also influence the clinical course of disease or define genetic variations that predict different toxicities and efficacies of available treatments.

The 2 SNPs examined in this work are no doubt only a snapshot of the total picture of susceptibility to therapy-related AML, but studies such as this are a beginning to improve our understanding of genetic susceptibilities. If confirmed in prospectively analyzed cohorts or other large retrospective cohorts of t-AML, these markers of therapy-related AML susceptibility might influence the choice of therapeutics to treat malignancies for which alternate therapies are available. They also provide preliminary insights into mechanisms of leukemogenesis, which may facilitate development of targeted therapies for t-AML.