Abstract
Conventional cytogenetics and FISH reveal chromosomal defects in approximately 50% of MDS patients. These mostly consist of gross gains and losses of specific chromosomal regions or entire chromosomes like 5q-, monosomy 7 and trisomy 8. Currently, the genes that are critical for MDS development remain largely unknown, which hampers both a proper diagnosis of clonal disease as well as development of targeted therapy. To identify the affected genetic loci and to map the critical regions and genes in MDS, we performed high-resolution (250k) SNP-based CGH. So far, 231 controls and 87 MDS patients from various subclasses were analyzed. In all patients and controls, loss of heterozygosity (LOH) without copy number changes was observed at multiple loci across the entire genome. Although large areas of LOH encompassing the main part of the p- or q-arm of chromosomes were only seen in MDS patients, no genomic regions were identified that were statistically more often affected in patients compared to control DNA. Copy number changes (excluding known regions of normal variation) were seen in 53% of patients with a normal karyotype (n=54). In 231 controls and in non-malignant T cells of a subset of patients, these areas were not affected, indicating that they were disease-specific. The number of affected regions per patient ranged from 0–7. The majority (82%) of karyotypic aberrations were confirmed using SNP-arrays. Only balanced translocations and some subclonal aberrations could not be detected. Importantly, SNP-array analysis revealed additional copy number changes in 70% of patients with an abnormal karyotype. Copy number changes that were observed in only one patient might reflect general genomic instability in the tumor cells and may not represent genes that are implicated in the pathogenesis of MDS. Therefore, we selected areas that were affected in at least two patients. In total, we found 51 different recurrent genomic loci. This indicates that MDS is genetically diverse, which is in agreement with its diverse clinical and morphological presentation. Among the 51 recurrent loci, 15 contained only a single gene (Table). Among these genes, there were several known to be implicated in MDS (e.g. ETV6 and RUNX1), whereas others represent novel genes that are potentially implicated in the pathogenesis of MDS. For several of these, a biological function has been described that may be linked to control of differentiation and proliferation, like the transcription- and proliferation-regulating gene JARID2 and the transcription factor DMTF1. Currently, we are performing a high thoughput mutation- and expression-analysis of these genes in a larger group of patients.
Single gene copy number changes in MDS Chr Cytoband Loss/Gain Cases Size (Mb) Gene 1 p35.1 loss 2 0.01 CSMD2 3 p24.2 loss 2 0.07 LRRC3B 6 p22.3 loss 3 0.02 JARID2 8 p23.2-1 gain 2 0.14 MCPH1 9 p13.2 gain 2 0.23 MELK 9 p24.3 gain 2 1.14 SMARCA2 11 q22.3 gain 2 0.05 SLC35F2 12 p12.1 loss 3 0.08 ST8SIA1 12 p13.2 loss 4 0.08 ETV6 12 q23.2 loss 2 0.03 IGF1 16 q23.3 loss 2 0.06 MPHOSPH6 21 q22.12 loss 3 0.07 RUNX1 21 q22.2 gain 2 0.62 DSCAM 22 q12.2 gain 2 0.00 PES1 X q13.1 loss 2 0.17 EDA
Detection of gene copy number changes in CGH microarrays using a spatially correlated mixture model
