Comparative Clinical and Gene Expression Analysis of Sickle Cell Anemia and Hemoglobin SC Disease

Homozygous hemoglobin (Hb) SS is the most common form of sickle cell disease (SCD), while HbSC represents a third of patients and has been less studied. Despite similarities in pathogenesis, there are significant clinical differences between genotypes. To elucidate any molecular differences, we compared gene expression profiles for HbSS and HbSC using microarrays. The Paxgene (PAX) blood RNA system was used to extract RNA from whole blood, which was then globin reduced. We also isolated peripheral blood mononuclear cell (PBMC) RNA to assess transcript patterns in cell populations lacking erythrocytes and granulocytes.

To demonstrate clinical differences, we compared HbSS and HbSC subjects from the Bethesda Sickle Cell Cohort Study (ClinicalTrials.gov identifier: NCT00011648). Hematologic parameters were significantly different, with HbSS having higher leukocytes, lower hematocrit, and higher platelets, reticulocytes and fetal hemoglobin. Significant differences were also observed for transaminases, LDH, bilirubin, creatinine kinase, ESR and ferritin. On the other hand, HbSC patients had higher diastolic blood pressure (P=0.0009), higher SpO₂ (P<0.0001), less frequent leg ulcers (Odds Ratio 2.90, 95% CI 1.15-7.35, P=0.008) and more prevalent tricuspid jet velocity above 2.9 m/s (Odds Ratio 5.07, 95% CI 1.75-14.71, P=0.0006). These associations led us to hypothesize that these are distinct forms of SCD whose biological differences might be identified with gene expression profiling.

RNA was isolated from ten subjects each with HbSS, HbSC or normal volunteers (NVs) using whole blood, PBMCs and Ficoll purified PBMCs (devoid of platelets). Microarray profiling showed that SCD subjects, regardless of genotype, have significant enrichment in glucose metabolism, cellular respiration, and mitochondrial energy generation pathways compared to NVs. Next, we used hierarchical clustering to distinguish SCD expression profiles from NVs, but we were unable to further distinguish HbSS from HbSC samples using this method. We attribute the above differences between SCD and NV samples to the erythroid cell RNA expression in whole blood. There is significant overlap between the 100 most abundantly expressed transcripts and a published signature in purified reticulocytes. However, no significant differences were observed when comparing SCD and NV samples using PBMCs and Ficoll purified PBMCs.

We only observed expression differences comparing HbSS and HbSC at the gene level for ATP1B1 (1.3 fold change, P=1.04E-4), WSB1 (1.3 FC, P=3.64E-3) and IL6ST (1.3 FC, P=1.13E-4), despite the evident hematological and clinical differences. ATP1B1 encodes a small ATPase for Na+/K+ transport and has hardly been explored in hemoglobinopathies. Furthermore, ATP1B1 was highly up-regulated in HbSS patients (against controls) but showed less change in HbSC patients. Our results provide an important first step towards elucidating the molecular pathophysiology that might explain the differences in clinical symptoms and survival between these 2 common forms of SCD.