Polycythemia vera (PV) is a myeloproliferative disease characterized by accumulation of erythrocytes and cells of the myeloid and megakaryocyte lineages. Although genes like PRV-1 and PTP-MEG2 have been implicated in the pathology of PV, there is no consensus on their importance in the disease process. Progenitor cells from PV patients can grow in the absence of erythropoietin, and are hypersensitive to a variety of other growth factors. This suggests that polycythemic hematopoietic progenitor cells possess a significantly different genetic program. We tested this idea by molecular profiling hematopoietic progenitor cells (CD34+) from PV specimens and normal donors. We purified CD34+ cells from the marrow of 10 PV patients and harvested total RNA. Biotinylated cRNA was made through two rounds of linear amplification, and hybridized to Affymetrix HGU133A genechips. CD34+ cells from marrow mononuclear cells of 5 normal controls were processed similarly. The resulting datasets were normalized to the median across chips and across genes. Unsupervised hierarchical clustering showed that PV samples had a distinct gene expression profile from the controls. We then performed supervised clustering using a non-parametric t-test (Wilcoxon rank sum test) using the Benjamini and Hochberg multiple testing correction held to a p-value of 0.01 to determine genes that were significantly different between disease and normal samples. Using these stringent criteria, there were 331 genes that reached significance. Strikingly most of these were decreased in expression compared with control CD34+ cells and only 34 genes were upregulated in PV. A 35 gene predictor set was discovered through the use of a k-nearest neighbor metric. This set was 100% accurate for the prediction of PV in a leave one out cross-validation approach. Among these genes are EVI1, a known oncogene and one of only two genes upregulated in PV on this list, and the putative tumor suppressor genes TUSC4 (NPR2), NDRG1 and KLF4. Also among the predictor genes is BAALC, a gene expressed in normal CD34+ cells and known to be a prognostic indicator gene for acute myeloid leukemia.

Author notes

Corresponding author

Sign in via your Institution