Gender Counts: Defining the Core Genetic Profile of Polycythemia Vera.

Abstract

Intra-abdominal venous thrombosis and exuberant extramedullary hematopoiesis leading to painful splenomegaly and hepatomegaly are two of the most serious consequences of polycythemia vera. Although polycythemia vera is slightly more common in men, both of these complications appear to be more common in women, in whom the disease also presents at an earlier age. We previously reported that polycythemia vera peripheral blood (pb) CD34+ cells could be distinguished from their immunophenotypically similar normal counterparts by gene expression profiling. We, therefore, hypothesized that phenotypic differences in disease behavior between men and women with polycythemia vera were a consequence of gender-based genotypic differences and have employed gene expression profiling to examine this issue. For this purpose, cRNA was prepared from the total RNA of pb CD34+ cells of eleven polycythemia vera patients and six normal controls purified by immunomagnetic bead chromatography. The cRNA was hybridized to an Affymetrix HU133 high-density oligonucleotide microarray chip representing 22,000 genes. Approximately 30–45 % of chip genes were recorded as present in the RNA samples. GC-RMA (Robust Multiarray Analysis) was used for normalization, to adjust for probe effects and for signal estimation. A parametric empirical Bayes statistical modeling method was used for differential gene expression analysis between the patients and controls. A posterior probability of >0.5 was taken to indicate significant differential gene expression. As previously reported, all polycythemia vera patients could be unequivocally distinguished from the controls, indicating that gene expression profiling can be employed as diagnostic test for polycythemia vera. With respect to gender, comparing male patients with male controls, 1106 genes were differentially expressed in the patient group; comparing female patients with female controls, 461 genes were differentially expressed in the patient group. Using unsupervised hierarchical clustering, the control patients segregated as a single group while the polycythemia vera patients segregated into two groups based on exuberant extramedullary hematopoiesis but this segregation was not gender-based despite the marked differences in gender-specific gene expression. Importantly, however, after eliminating gender-specific genes, only 93 genes were concordantly expressed in the male and female patients (59 up regulated and 34 down regulated). This small group appears to represent core genes in polycythemia vera whose behavior may be modified by gender-specific genes but the specific up regulation or repression of which is essential for disease expression. These data indicate that gender-specific effects on gene expression must be addressed in order to determine the basic genetic signature of a hematologic malignancy and they also define a core genetic profile in polycythemia vera.