SNP Arrays Facilitate Genotyping of Non-Synonymous SNP in MDS To Identify Disease Susceptibility Loci.

Myelodysplastic syndrome (MDS) is a clonal premalignant disease of hematopoietic stem cells characterized by cytopenias and predilection to acute myeloid leukemia (AML). While various exogenous factors (exemplified by chemotherapy-related MDS) constitute known risks for the development of MDS, it is possible that despite long latency, complex genetic traits contribute to MDS susceptibility. Such heritable factors include genes involving DNA repair, apoptosis, senescence, carcinogen catabolism and immune surveillance. Previously, disease association studies were mainly empiric and relied on rational selection of a very limited number of polymorphisms. Recent advances of SNP-array (SNP-A) technology allow for screening of a large number of SNPs. While some SNP-A utilize haplotype tags, custom arrays may specifically target known non-synonymous SNPs. We hypothesized that application of SNP-A genotyping may facilitate identification of potentially pathogenic SNPs. Such a screening approach is a hypothesis forming tool and our study is the first application of this technique to MDS. We have used the 13.9K non-synonymous Genotyping BeadChip (Illumina); DNA from 151 MDS patients (low risk: N=79, advanced: N=51, CMML1/2: N=21) and 99 controls (120 historical controls). In total, ∼2.4 mil genotypes were obtained. Genotype calls were computed and analyzed with Exemplar software. In the initial training Bonferroni correction was not applied. Instead as a hypothesis-forming approach we have ranked all SNPs according to their p value (automated analysis of multiplexed statistical evidence) and case/control ratio. We focused our search on the 100 highest ranking SNPs with a control frequency <5%. Globally, in the whole group 49 SNPs showed a p value of <.001, 75 SNPs present in ≤1% of controls and were found in >5% of cases. For example, the AA (rs3219484) variant of MUTYH, a gene involved in oxidative DNA repair was found in 19% of MDS vs. 2.6% of controls (p=9×10⁻⁸). To limit the impact of heterogeneity, subgroups of MDS were also analyzed separately. Among many interesting SNPs found, the AG genotype of (rs8192297) ANPEP was associated with RARS (35% vs. 15% in controls; p=.007). Similarly, the GA form of rs3730947 in DNA repair gene LIG1 was found in 11% of RARS patients (p=.00045), though it was absent in other MDS subtypes and controls. Similar, “enrichment” was observed in patients with CMML1/2, showing e.g., heterozygosity of WDR35 (rs1060742), ALPK2 (rs3809975) at the frequency of 45% and 30% (13% and 4% of controls; p=.0006, p=.0001). In sum, our study constitutes the first application of SNP-A genotyping to study susceptibility loci in MDS.