SNP Arrays in Patients with Myelodysplastic Syndromes and Ringed Sideroblasts: A Karyotypic and Disease Association Analysis

Introduction: The myelodysplastic syndromes (MDS) are a heterogeneous group of clonal hematopoietic stem cell disorders. Ringed sideroblasts (RS) are found in the following subclasses of MDS: refractory anemia with ringed sideroblasts (RARS), refractory cytopenia with multilineage dysplasia and ringed sideroblasts (RCMD-RS), and refractory anemia with ringed sideroblasts associated with marked thrombocytosis (RARS-T). The objective of this study was to evaluate the use of single nucleotide polymorphism (SNP) arrays (SNP-A) in patients with MDS and RS and specifically

1. to compare chromosomal abnormalities detected by metaphase karyotyping (MC) with those detected using high-resolution SNP based karyotyping (which can detect unbalanced genomic lesions in addition to copy-neutral loss of heterozygozity) and

2. to conduct a disease association analysis using the SNP-A.

Methods: We reviewed the electronic records of patients with MDS and RS seen at our institution between 2002 and 2008. DNA was extracted using the Puregene DNA Purification Kit. Gene Chip Mapping 250K Assay Kit (Affymetrix) was used. Signal intensity and genotype calls were analyzed using CNAG v.3.0. For the disease association analysis, the Fisher’s p-value was used to compare SNPs found in patients with MDS and RS versus 150 normal controls.

Results: 83 patients with MDS who have RS were identified. 40 (48%) had RARS, 25 (30%) had RCMD-RS, and 18 (22%) had RARS-T. The mean age of these patients was 70.7 years, 53 patients (64%) were males, and 70 (84%) were Caucasian. Of those 83 patients, 45 had available DNA for SNP analysis, 23 (51%) of whom had RARS, 11 (24%) had RCMD-RS, and 11 (24%) had RARS-T. The mean age of these 45 patients was 69.9 years, 29 (64%) were males, and 39 (87%) were Caucasian. By MC, 20/45 (44.5%) patients had abnormal karyotypes and 25/45 (55.5%) patients had normal karyotypes. Using SNP-A, chromosomal abnormalities including UPD were identified in 29/45 (64.5%) of patients. Of the 25 pts who had normal karyotypes by MC, 11 (44%) had abnormal karyotypes by SNP-A. The chromosomal distributions of the lesions detected by MC were as follows: chromosome 5 (18.4%), chromosome 7 (15.8%), chromosome 8 (13.1%), chromosome 17, 18, 19, 20, 21 (5.2% in each), and others (26.3 % in total). The distribution of chromosomal lesions detected by SNP-array analysis (excluding UPD) was as follows: chromosome 8 (18.7 %), chromosome 5 (14.6%), chromosome 7 (12.5%), chromosome 17 (10.4%), chromosome 20 (8.3%), chromosome 4 (6.2%), chromosomes 2, 3, 13, 22 (4.1% each), and others (12.5% in total). UPD was found in 12/45 (26.7%) patients mostly affecting chromosome 1 (27.8%). A large number of SNPs were found to be significantly more prevalent in patients with MDS with RS than in controls (with p-value < 0.0001). Genes within 50 kb from these SNPs were scrutinized. At least 11 of those genes (RP1, LIMD1, CHL1, ATP6V1F, TEAD2, SPTLC2, CDH13, DIAPH2, DLEU2, FAM10A4, TRPM8) are known to be related to cancer in the literature. Given that karyotypic abnormalities were more prevalent in chromosomes 8, 5, and 7, we looked specifically at the SNPs in those chromosomes which were significantly associated with disease (rs 409429, rs 446153, rs 453186 and rs 509273 in chromosome 8; rs6891109 in chromosome 5; and rs6970371 in chromosome 7). The genes within 50 kb of these SNPs that are known to be associated with cancer are: RP1 in chromosome 8 (colon cancer), and ATP6V1F in chromosome 7 (prostate cancer).

Conclusion: This study shows that SNP-A based karyotyping is a useful tool for karyotyping and can detect more chromosomal abnormalities than MC (64.5 versus 44.5%, odds ratio 1.45). We also found that about half of the patients who had normal karyotypes by MC were found to have karyotypic abnormalities by SNP-A. In addition, we show multiple candidate genes that could be important in the pathogenesis of MDS with RS.