Microarray Analysis of Erythroid Gene Expression In Sickle Cell Anemia Patients Treated with Hydroxyurea.

Although the clinical course of sickle cell anemia (SCA) is highly variable, the level of fetal hemoglobin (HbF) is well-recognized as a critical laboratory parameter; lower HbF is associated with a higher risk of vaso-occlusive complications, organ damage, and early death. Hydroxyurea treatment has been shown to induce HbF, improve laboratory parameters, and ameliorate clinical complications of SCA but its mechanism of stimulating HbF expression remains incompletely defined. We examined the gene expression profiles of early erythroid cells to identify pathways or mechanisms by which hydroxyurea (a) affects red cell development and (b) induces the production of HbF.

Peripheral blood was collected from patients enrolled in the Hydroxyurea Study of Long-term Effects (HUSTLE, ClinicalTrials.gov NCT00305175) and reticulocytes were purified by CD71⁺ positive selection. Total RNA was extracted from each sample and global gene expression was measured by Affymetrix Human HG-U133 Plus2Chip arrays. An average of 3000 transcripts/genes were detected for each patient RNA sample. Analysis of differential gene expression was then performed comparing: (1) paired baseline (pre-treatment) vs. maximum tolerated dose (MTD) hydroxyurea samples; (2) high vs. low HbF induced patient samples.

At MTD all 59 subjects had increased hemoglobin concentration, increased %HbF, along with decreased reticulocyte, neutrophil and WBC counts (all p<0.001). For 32 patients with a matched baseline and MTD RNA sample, we identified 253 significantly differentially expressed genes following hydroxyurea exposure. Gene ontology identified major categories for genes involved with translation elongation, ribosome assembly, chromatin and chromosome organization, response to protein stimulus and vesicle function. There was a large number of ribosomal related proteins downregulated by hydroxyurea (n=30), while genes involved with chromatin organization were mostly upregulated (n=17). Using clinical %HbF data, we identified a group of low responders (n=17, MTD %HbF <20%) and a group of high responders (n=16, MTD %HbF >30%). When the gene expression of these groups were compared, 123 genes were statistically related to optimum HbF induction. Gene ontology showed common themes for genes involved in nucleoplasm function, RNA binding, amino acid ligase activity and histone acetyltransferase activity. Subsequent quantitative PCR experiments confirmed the differential expression of 10 candidate genes; EIF1, ELOVL6, HBBP1, HSP90AA1, MAML3, OSBPL6, RAC2, RPL3 and SNRPN were all statistically different after hydroxyurea exposure; while H3F3A, HBBP1 and SNRPN were all statistically different in the high HbF responders (Table 1).

Using reticulocyte RNA samples from patients with SCA, we have shown that hydroxyurea significantly alters gene expression in early circulating erythroid cells in vivo. Differentially expressed genes involved with translation elongation, ribosome assembly, and chromatin and chromosome organization likely reflect the cytostatic action of hydroxyurea. When genes related to HbF induction were examined, we identified several genes involved with RNA binding (e.g. SNRPN) and histone activity (e.g. H3F3A). Further investigation of these genes should greatly improve our understanding of HbF induction by hydroxyurea and potentially help predict patient clinical benefit prior to commencing hydroxyurea therapy.

Genes were identified with expression differences related to hydroxyurea exposure or to HbF induction. For this subset of 10 genes, the change was seen by microarray and quantitative PCR (qPCR). The fold change is given as negative and positive for downregulated and upregulated genes, respectively.

Off Label Use: The off-label drug use of hydroxyurea to treat the clinical complications of sickle cell anemia in children will be discussed.