Gene Expression Profiles of CD34+ Cells in Myelodysplastic Syndrome

Myelodysplatic syndrome (MDS) represents a heterogeneous group of clonal disorders with ineffective hematopoiesis that is characterized by dysplasia and peripheral cytopenia of one or more cell lineages. We studied gene expression profiles in CD34+ cells of 42 MDS patients and 6 healthy controls using Illumina cDNA microarray technology. Nine patients had RA, 7 patients had RCMD, 17 patients had RAEB and 9 had RAEB-T. CD34+ cells were isolated from bone marrow samples using MACS magnetic columns. The quality of total extracted RNA was confirmed with the Agilent Bioanalyzer 2100. 200ng of total RNA was amplified using Illumina RNA amplification kit. cRNA targets were hybridized on the Sentrix HumanRef-8 BeadChips (> 24 000 probes), which were scanned on the Illumina BeadStation 500. The data were pre-processed and normalized by lumi R package designed to preprocess the Illumina microarray data. Normalized data were filtered by detection p-value <0.01, resulting in total number of 10 091 genes. This gene set was tested for differential expression between clinical groups and control group. For this purpose, statistical testing by ANOVA with correction for multiple testing problem by Bayesian thresholding was performed. Additionally, analysis by random-forests (RAFT) was performed. Significant genes from both analyses were merged resulting in 332 differentially expressed genes detected. Out of these, 79 genes showed ≥2.5 fold changes in gene expression between controls and all MDS groups (22 up-regulated and 57 down-regulated). Our findings were confirmed by real-time quantitative PCR for several genes (TaqMan Gene Expression Assays). We used DAVID database to annotate 79 selected genes: 8 of 22 up-regulated genes in MDS patients were recognized to play a role in regulation of transcription (LEO1, E2F6 and several zing finger proteins). A half of these over-expressed genes could not be annotated due to still unknown biological function. Within the set of the down-regulated genes in MDS patients those biological processes were predominantly detected: cell differentiation (KLF4, FOSL2, STK17B, BCL3, SNF1LK, ID2 etc.), response to stress (CXCL12, SMAD7, CYGB, etc.) and cell proliferation (MXD1, OSM, FTH1, KLF10 etc.). In the set of 31 genes with 5 fold decreased expression, we identified 8 genes involved in B-cell development. (VPREB1, VPREB3, CD79A, EBI2, LEF1, CXCL12, CTGF, GALNAC4S-6ST). RAFT analysis was performed also in the set of 332 statistically differentially expressed genes in order to evaluate accuracy of grouping the patients according their diagnosis. We detected strong heterogeneity in gene expression patterns within the MDS patients, especially in the RAEB group reflecting clinical diversity of MDS. Clustering analysis (Spearman correlation) showed that most of the RAEB-2 patients (7 out of 9) were clustered together with REAB-T whereas RAEB-1 clustered with RCMD or RA. These results underline the need of distinguishing RAEB-1 and RAEB-2 diagnosis according to WHO classification system, since their expression profiles are significantly different.