Polymorphisms of Exon 10 in Stem Cell Factor Receptor KIT Are Associated with a Significantly Improved Relapse-Free Survival in Childhood AML.

Molecular markers like KIT, FLT3, CEBPA and NPM1 have been identified in AML as functional and prognostic relevant factors with an impact on therapy and outcome of AML. KIT is encoded by twenty-one exons. Exons 1 to 9 encode the extracellular IG-like domain, exon 10 encodes the transmembrane (TMD), exon 11 the juxtamembrane (JMD), exon 17 the kinase insert (KI) and the remaining exons encode the kinase domain (KD). Certain exons like exon 10, 11 and 17 are highly conserved between species. Several mutations lead to a constitutively aberrant activation of KIT with relevance in cancerogenesis like in GIST and AML. Overall approximately 18% of childhood AML carry KIT mutations mostly in exons 8, 11 and 17 which are of prognostic and therapeutic relevance as potential targets for RTK inhibitors. Besides mutations several single nucleotide polymorphisms (SNP) are identified for KIT such as 1642 A/C (exon 10), 1659 A/G (exon 10), 2415 C/T (exon 17), 2607 G/C (exon 18) and 3196 G/A (intron). Little is know about the potential impact of KIT polymorphisms, aberrant expression and epigenetic effects on the pathogenesis and prognosis of childhood AML.

Genomic DNA from 188 newly diagnosed pediatric patients with AML who were treated according to the AML-BFM 98 and AML-BFM 04 protocols was purified and exons 8, 10, 11 and 17 were screened for mutation detection and SNP discovery with a WAVE nucleic fragment analysis system (DHPLC based detection of heteroduplex formation). Samples with abnormal WAVE pattern were further analyzed via direct sequencing for confirmation. SNP vs WT distribution was analyzed statistically in a Kaplan-Meier survival analysis model for relapse and event-free survival (EFS) in all pts and the standard risk (SR) and high-RISK (HR) subpopulation. In AML-BFM HR is defined by FAB (M0, M1/2 without Auer rods, M4 to M7, but not M4EO, FLT3-ITD and all pts >5% blast in the bone marrow at day 15 of treatment)

The median age of pts was 9.8 years (6 months to 17.9 years). In 149 pts (79%) the exon 10 sequence was WT and in 40 pts (21%) exon 10 SNPs were detected: 12 (30%) at position 1659, 24 (60%) at position 1642 and 4 (10%) at both positions. In the WT group were 9% infants (<1yrs), 45.5% children (1-9yrs) and 45.5% >10yrs. In the SNP group were 4.5% infants, 29% children and 66.5% >10yrs. The WT group had 49% female and 51% male and in the SNP group 56% male and 44% female. In WT pts 71% and in SNP pts 85% had <100000 white blood cells at diagnosis. The FAB distribution was comparable in WT and SNP pts. Distribution of cytogenetic abnormalities (t(8;21), t(15;17), inv(16) and others) between both groups were comparable. Treatment response (<5% blasts) at treatment day 15 was similar with 71% in WT vs 67% in SNP. Distribution in the SR and HR arm was also similar in WT and SNP pts. There was no significant difference within the SNP variants neither in patient distribution nor outcome. The 5-yrs pEFS was 64% (SE 8%) in SNP and 41% (SE 5%) in WT (p=0.08). The pEFS in the SR group was 72% (SE 7%) in SNP and 73% (SE 10%) in WT pts. The pEFS in the HR group was 57% (SE 10%) in SNP and 28% (SE 7%) in WT (p=0.09). The incidence of relapses at 5 years was 40%, SE 5% in WT and 23% (SE 7%) in SNP pts (p=0.06). The incidence of relapse was similar in the SR-Group (SNP 27%, SE 11%; WT 23%, SE 6%; p=0.65). In the HR group the relapse rate at 5 years was 51% (SE 7%) in WT and 19% (SE 9%) in pts with SNP (p=0.02). In a Cox regression analysis of the relapse rate including favorable cytogenetics and bone marrow count day 15 > 5% as co-variables the mutation remained significant (RR 0.45, 95% CI 0.21 – 0.97). Also for EFS the effect was significant (RR 0.51, 95% CI 0.27 – 0.99)

It seems that the KIT Exon 10 polymorphisms are associated with a significantly reduced relapse rate and an improved EFS in patients with HR childhood AML. The published incidence of exon 10 SNPs in healthy subjects is approximately 20% and is reflected in our population. The underlying pathophysiology is unclear. Inasmuch post-transcriptional protein modifications are affected by these subtle nucleotide substitutions is unclear. The effect on relapse and EFS of the different exon 10 variants will be analyzed prospectively in the upcoming AML-BFM trial. Knowledge of the predictive value of the exon 10 SNP variants could help to improve stratification if these data remain significant in multivariate analyses of a prospective trial.

No relevant conflicts of interest to declare.