DNA Microarray Analysis of GPI-Negative Hematopoietic Stem Cells in Paroxysmal Nocturnal Hemoglobinuria Patients.

Paroxysmal nocturnal hemoglobinuria (PNH) is an acquired clonal hematological disorder which is manifested by complement-mediated hemolysis, venous thrombosis, and bone marrow failure. Deficiencies of glycosylphosphatidylinositol (GPI)-anchored proteins, due to mutations in the phosphatidylinositol glycan-class A (PIG-A) gene, contribute to complement-mediated hemolysis and affect all hematopoietic lineages in PNH. However, it is unclear how a PNH clone with a PIG-A gene mutation expands in bone marrow. Although some genes, including the Wilms’ tumor gene (Shichishima et al, Blood, 2002), the early growth response gene, anti-apoptosis genes, and the gene localized at breakpoints of chromosome 12, have been reported as candidate genes that may associate with proliferations of a GPI-negative PNH clone, previous studies were not intended for hematopoietic stem cell, indicating that the differences in gene expressions between GPI-negative PNH clones and GPI-positive cells from PNH patients remain unclear at the level of hematopoietic stem cell. To identify genes contributing to the expansion of a PNH clone, here we compared the gene expression profiles between GPI-negative and GPI-positive fractions among AC133-positive hematopoietic stem cells (HSCs). By using the FACSVantage (Becton Dickinson, San Jose, CA) cell sorting system, both of CD59⁺AC133⁺ and CD59⁻ AC133⁺ cells were purified from bone marrow mononuclear cells obtained from 11 individuals with PNH. Total RNA was isolated from each specimen with the use of RNeasy Mini column (Qiagen, Valencia, CA). The mRNA fractions were amplified, and were used to generate biotin-labeled cDNAs by the Ovation Biotin system (NuGEN Technologies, San Carlos, CA). The resultant cDNAs were hybridized with a high-density oligonucleotide microarray (HGU133A; Affymetrix, Santa Clara, CA). A total of >22,000 probe sets (corresponding to >14,000 human genes) were assayed in each experiment, and thier expression intensities were analyzed by GeneSpring 7.0 software (Silicon Genetics, Redwood, CA). Comparison between CD59-negative and CD59-positive HSCs has identified a number of genes, expression level of which was statistically different (t-test, P <0.001) between the two fractions. Interestingly, one of the CD59⁻ -specific genes isolated in our data set turned out to encode a key component of the proteasome complex. On the other hand, a set of transcriptional factors were specifically silenced in the CD59⁻ HSCs. These data indicate that affected CD59-negative stem cells have a specific molecular signature which is distinct from that for the differentiation level-matched normal HSCs. Our data should pave a way toward the molecular understanding of PNH.