Whole Genome 250K SNP Array Allows for Detection of Previously Unidentified Lesions in Chromosome 5 in Patients with MDS.

Abstract

Chromosome abnormalities in myelodysplastic syndromes (MDS) have important prognostic implications. The frequent occurrence of certain invariant aberrations implies a role of involved genes in the malignant phenotype of the affected clone. The recent therapeutic success of lenalidomide (CC-5013, Revlimid® ) in achieving cytogenetic remissions in a high proportion of patients with the 5q deletion suggests that certain chromosomal lesions may be used to guide targeted molecular therapeutics. Nevertheless, traditional metaphase cytogenetics (MC) has limited resolution and is hampered by frequent lack of growth in vitro. Due to the inherent insensitivity of banding techniques ( > 20% of abnormal metaphases), only large clones can be detected. Significance of FISH-based detection of a few karyotypically abnormal cells remains unclear as smaller clones may be transient and therefore not pathogenic. Whole genome scan using SNP arrays (SNP-A) allows for precise detection of cryptic (due to limited size) defects by simultaneous analysis of LOH and gene copy number. Unlike MC, uniparental disomy (UPD) can be easily detected by SNP-A. The resolution level of this technology is limited only by the density of the arrays; currently 500K SNP-A are readily available. We hypothesized that the application of SNP-A in MDS will allow for the detection of "marker" lesions located within chromosomal regions frequently affected in MDS. In patients with deletion 5q, standard MC analyses have shown that additional chromosome abnormalities impart an incremental adverse prognosis in the absence of marrow blast elevation, suggesting that detection of occult lesions may have immediate prognostic relevance. Here, we applied 250K Affymetrix SNP-A for bone marrow karyotyping. Results were confirmed using microsatellite analysis. Initially, we have analyzed a cohort of 356 MDS patients, 43 patients had losses of genomic material in chromosome 5. While studying a subset of patients (N=112) from this cohort by SNP-A, we confirmed 13/15 previously known 5q31 deletions. In addition, we identified 2 patients with aberrations involving this region that were not detected by MC. One patient had a segmental UPD5q23.3-q31.1, likely due to mitotic recombination, while the second patient with uninformative cytogenetics had a new deletion within 5q. The affected regions involved the cytokine gene cluster at 5q31. The SNP-A-based whole genome scan of patients with previously known 5q aberrations as the sole MC-defined abnormality detected additional deletions present in other chromosomes in 3/7 patients. These lesions were located at 1p21.1, 2q21.2, 4q23, 6p21.33, 7q34, 8q24.23 and 11p15.1. Our findings may point towards rational application of lenalidomide in patients who would otherwise not be primary targets of such therapy. Moreover, as the response rate to lenalidomide is not complete and some patients may relapse, it is possible that detection of certain occult defects using SNP-A may point towards lenalidomide resistance or modify clinical features in individual patients.