Gene Expression Profiling of the Bone Marrow Mononuclear Cells from Patients with Melodysplastic Syndrome.

Myelodysplastic syndrome (MDS), a clonal disorder of hematopoietic stem cells, is characterized by cytopenia in at least one lineage of peripheral blood and dysplastic features in BM cells. It has been considered as a preleukemic condition for a high risk of transformation to AML. The FAB classification has divided MDS into five clinical categories: RA, RAS, RAEB, RAEB-t and CMML. To date, the key genes involved in the pathogenesis and its progression to acute leukemia has not been clearly identified. This was at least partially caused by a narrow scale of genes which could only been studied once a time by using traditional methods. The present study was to investigate the molecular mechanism of MDS by using cDNA microarray techniques. We have first applied total RNA of bone marrow monoculear cells (BMNCs) from 2 MDS patients (one RA and one RAS) to a BioStar H141 microarray ( United Gene Holdings Co. LTD, Shanghai, China) containing 13484 gene cDNA clones and ets. Based on the preliminary results of these assay, 500 genes which were shown most remarkably differentially expressed in MDS compare to normal control, and with known functions and potential involvement in hematopoiesis regulation, were selected to design and compose 10 cDNA chips. These arrays were then used to analyze the gene expression pattern of BMNCs from 10 patients with MDS, including 4 RA, 1 RTC, 4 RAEB, and 1 RAEBt. To confirm the microarray results and to evaluate the disease relevance of some selected genes from array results, a semiquantitative RT-PCR was performed to analyze gene expression in fifty addi-tional patients with MDS (28 RA, 15 RAEB, 7 RAEBt), 38 acute myeloid leukemia (7 M1, 12 M2, 4 M3, 4 M4, 5 M5, 6 M6), 15 atypical anaplastic anemia (AA), and 12 hypercellular anemia. Our results revealed that 95 genes were abnormally expressed in at least five MDS patients compared to normal controls, involving cell growth and differentiation regulation, cell cycle control, signaling, redox, such as thrombospondin 1 (THBS1), phosphatase and tensin homolog (PTEN), MAD, DNA-damage-inducible transcript 3 (DDIT3), ets variant gene 1 (ETV1), and G1 to S phase transition 1 (GSPT1). These MDS patients in different risk groups could be clustered into two groups overall by hierarchical clustering, wherein a case with isolated thrombocytopenia and other RA patients were clustered into two subgroups 5 genes (RNAHP, DDIT3, FOXO3A, GSPT1, and ETV1) which displayed a most marked differential expression pattern in most MDS patients, were selected as "candidate genes". Consistent expression patterns of 3 (60%) in 5 genes were confirmed by semiquantitative RT-PCR. Further analysis showed different transcript levels of RNAHP, DDIT3 among patients with MDS in different risk group, AML, and normal controls. Meanwhile, the transcript levels of five genes were also compared in the patients with RA, AA and other hypercelluler anemia patients. There was significant difference in RNAHP levels between RA and CAA, or other hypercelluler anemia (P<0.05), similar diversity also seen in ETV1 levels between RA and AA (P<0.01). However, significant differences in DDIT3, FOXO3A, and GSPT1 levels were not observed. Our study suggested that gene expression profiling of MDS patients may reveal a specific transcription patterns for BMNCs in MDS. The abnormal expression of RNAHP, DDIT3 and ETV1 may play roles in the patho-genesis of MDS and may provide useful biomarkers for the molecular diagnosis of MDS.