The FOXO3A genotype is strongly associated with longevity in humans. It encodes a transcription factor that appears to regulate anti-oxidant genes during erythroid differentiation in mice, resulting in a hemolytic anemia. The gene has also been implicated in the regulation of fetal hemoglobin expression in children with sickle cell disease. We performed a gene-wide association study to identify and replicate variants of FOXO3A that might be associated with seven biomarkers in patients with sickle cell anemia.
1198 patients from the Cooperative Study of Sickle Cell Disease (CSSCD) study, 308 patients from the Pulmonary Hypertension and Sickle Cell Disease with Sildenafil Therapy (walk-PHaSST) study, and 220 patients from the Pulmonary Hypertension and the Hypoxic Response in Sickle Cell Disease (PUSH) study were analyzed. Biomarkers included hematocrit, reticulocyte count, fetal hemoglobin (HbF), serum levels of lactate dehydrogenase (LDH), aspartate aminotransferase and bilirubin, and the calculated hemolytic component, each appropriately transformed to achieve normality. A total of 189 single nucleotide polymorphisms (SNP) that were either genotyped or imputed (quality r2 > 0.9) were used. Association between each biomarker and SNP was tested using linear regression assuming an additive genetic model, adjusted for age and sex. None of the patients in the CSSCD were treated with hydroxyurea at the time of measurements of the biomarker variables. In the Walk-PHaSST and PUSH, we first examined whether there was a significant association between the biomarker and treatment effect of hydroxyurea; if there was a significant treatment effect, then we looked at potential SNP-by-treatment interaction. For those with significant interactions, only patients without hydroxyurea treatment were included in the analysis. The genetic analysis results from the three studies were then combined to produce meta-analyzed results. Finally, a survival analysis using Cox regression was performed to model age at death in a subset of 54 patients in the CSSCD.
Among the seven biomarkers, hematocrit showed the most robust enrichment of associations with FOXO3A SNPs. Eight of the 16 published variants had meta-analyzed p-value <0.05. Of those, six had a consistent direction of effects across all three cohorts. Overall, there were 8 loci with 34 SNPs that had meta-analyzed p<0.05 in hematocrit. The most significantly associated SNP (rs6911407; meta-analyzed Beta=-0.0127, meta-analyzed p=0.0013) is one previously associated with human longevity. LDH was most significantly associated with variant rs12206094 (meta-analyzed Beta=0.0256, meta-analyzed p=0.0072), another SNP previously associated with human longevity. Four of the allelic variants associated with LDH were also associated with hematocrit in the appropriate direction. There were also some evidence of enrichment of associations with reticulocyte counts (2 loci with 22 SNPs), HbF (1 locus with 11 SNPs; 10 SNP associations overlapped with reticulocyte count in the appropriate direction), and hemolytic component (2 loci with 8 SNPs); however, the strengths of associations in these biomarkers were marginal (0.01<p<0.05). The survival analysis revealed one locus significantly associated with age at death in the CSSCD patients (rs2802297; p=0.028).
FOXO3A genetic polymorphisms are associated with hematocrit and serum LDH in this meta-analysis of patients with sickle cell anemia, with less robust associations with fetal hemoglobin level and reticulocyte count. Our genetic findings are biologically consistent with published knockout mouse data indicating that FOXO3A regulates red cell antioxidant capacity and hemolytic severity. Our observation of fetal hemoglobin association with FOXO3A helps to validate previously presented results from Sheehan and colleagues. Parallel to well-documented results in the general population, we find preliminarily that a FOXO3A allelic variant predicts longevity in patients with sickle cell anemia. FOXO3A appears to play a significant role in phenotypic variation in sickle cell anemia.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.