Elevated fetal hemoglobin (HbF, α2γ2) ameliorates the pathophysiology of sickle cell disease (SCD) by diluting HbS and inhibiting its polymerization in patient red blood cells. Hydroxyurea (HU) the only approved drug for SCD acts in part by raising HbF levels. However, up to 50% of adult SCD patients treated with hydroxyurea, do not have a clinically meaningful response, illustrating the need for new improved agents. We identified FOXO3, a potentially "druggable" transcription factor, as a candidate HbF inducer through analysis of rare variants from whole exome sequence data from SCD patients. Specifically, numerous rare variants that potentially impair FOXO3 activity and/or expression were associated with reduced HbF levels. FOXO3 was previously shown to facilitate erythroid precursor survival, maturation and resistance to oxidative stress.
To investigate our findings from genomic studies further, we manipulated FOXO3 expression in normal human CD34+ hematopoietic stem and progenitor cells (HSPCs) induced to undergo erythroid differentiation and examined the effects on γ-globin RNA levels and HbF production. Transfection of CD34+ HSPCs with lentivirus expressing FOXO3 shRNA that reduced levels of the corresponding RNA by 80% reduced γ-globin (HBG1 and HBG2) mRNA by 60% compared to control cultures expressing scrambled shRNA (n= 3 separate experiments, p < 0.0005). A second non overlapping FOXO3 shRNA reduced the corresponding mRNA by 50% and reduced γ-globin by 31% (n=3 separate experiments, p < 0.0001). Reduction of FOXO3 mRNA was accompanied by delay in differentiation (11% GPA+ cells with 80%FOXO3 kd by day 14 compared to 72% GPA+ cells in the scramble kd).
Metformin, an FDA-approved drug used for type 2 diabetes, is known to increase FOXO3 expression and therefore, we tested its effects on HbF expression in CD34+ HSPC-derived erythroblasts by treating with 30, 50 and 100 µM drug, beginning at day 5 of culture. At day 14 of culture, 59% late basophilic erythroblasts were present in control cultures, as evidenced by band3 and α4 integrin expression. In contrast, maturation was accelerated in metformin treated cultures, with 70% late basophilic erythroblasts, consistent with previous studies. Remarkably, metformin treatment also caused a dose dependent increase in HbF-immunostained cells ("F-cells", Figure 1) and in % HbF, as measured by HPLC (Figure 2). HPLC analysis at d14 showed that %HbF (HbF/(HbF + HbA) rose from 10.3% at baseline, to 19.8% and 30% at 50 and 100 µM metformin (n= 3, p=0.04 and 0.005). For dosage comparison patients on standard metformin therapy for diabetes typically exhibit plasma drug levels of 70 µM. When cultures were treated with metformin (100µM) and HU (30µM), %HbF was increased to 45% compared to 26% and 30% in cultures treated with HU or metformin respectively as single agents (n=3, p=<0.0001, p=0.0016, and p=0.005).
Thus, HU and metformin show additive effects, on HbF induction, suggesting that combination therapy in patients may be more effective than standard therapy of HU alone. In addition, metformin may be a superior agent than HU as single therapy because it does not cause myelosuppression, and therefore requires minimal laboratory monitoring, and does not arrest erythropoiesis, making it a potentially effective agent for quantitative beta hemoglobinopathies. Taken together, our results indicate that FOXO3 is a positive regulator of γ-globin expression and a potential therapeutic target for HbF induction. Given our current results and the excellent safety profile of metformin, a pilot clinical trial of the drug (1500-2500 mg/day) alone or combined with hydroxyurea at stable maximum tolerated dose, in patients aged 16-40 years with sickle cell anemia or non-transfusion dependent beta thalassemia has been approved by the Baylor College of Medicine IRB.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.