Large Regions of Uniparental Disomy (UPD) Establish Clonal Hematopoietic Stem Cell Selection in a Subset of Myelodysplastic Syndrome (MDS) Patients with Normal Bone Marrow Cell Karyotypes.

Abstract

Myelodysplastic syndromes (MDS) comprise a large group of hematological diseases sharing clinical features, but the diverse mechanisms underlying their molecular pathogenesis remain largely undefined. In this study, we used high-density 250K SNP arrays to analyze matched constitutional (buccal) and whole bone marrow DNAs from 36 patients with MDS. In 14 cases, we also selected CD34+ stem and progenitor cells for comparison. 23 of the cases were cytogenetically normal. Whereas regions of genomic amplification were rare, we found distinct regions of heterozygous deletion in 9 patients, most frequently targeting chromosomes 5q, 7q, 17p and 20q. In one patient with a large heterozygous 5q deletion, a smaller region in 5q34 was homozygously lost. Based on genotyping we could identify regions of uniparental disomy (UPD) in 4 patients, three of whom had normal bone marrow karyotypes. Two of the regions of UPD targeted chromosome 7q, suggesting that the frequency of 7q deletion is underestimated by conventional cytogenetics, and that a mutation has occurred in these cases within the reduplicated region involved in UPD. Comparison with constitutional DNA was critical for accurate SNP analysis in MDS, allowing us to focus on acquired clonal abnormalities and control for changes due to inherited copy number variation or unusual regions of germline homozygosity. For example, one case had several large blocks of germline homozygosity distributed over various chromosomes, and by comparison with the buccal DNA we were able to discern that these regions did not result from clonal somatic abnormalities associated with MDS. The analysis of CD34+ cellular DNA in 14 cases revealed that all aberrations identified in the stem and progenitor fraction were also identified in whole bone marrow DNA. This finding emphasizes the dominance and stem cell involvement of the abnormal clone in MDS, which makes purification of CD34 cells unnecessary for SNP array analysis in this disease. Most, but not all, clonal deletions evident by cytogenetic analysis were detected as copy number alterations by SNP array, possibly indicating that such abnormalities occurred in subpopulations of the cytogenetically analyzed MDS cells. Thus, SNP array and cytogenetic analysis are complementary methods for the identification of abnormal clones in MDS patients, each detecting overlapping but distinct subsets of patients. Our data indicate that SNP array analysis should be performed in all patients with MDS and normal bone marrow cytogenetics, in order to help discriminate patients with UPD indicating cell autonomous clonal stem cell malignancy from those who might have extrinsically mediated bone marrow failure syndromes.