Sickle cell disease (SCD) affects >100,000 Americans and millions more worldwide. Symptoms and sequelae of SCD can be ameliorated by increasing fetal hemoglobin (HbF, α2γ2) levels. Unfortunately, up to 50% of adult SCD patients treated with hydroxyurea, the only FDA-approved and widely used HbF inducer, do not have a clinically meaningful response to the drug. Additional oral HbF inducing agents, especially those that require less intense laboratory monitoring, are urgently needed. Development of such drugs has been stymied by an incomplete understanding of γ-globin regulation. We hypothesized that natural human genetic variation can be used to identify genes that may be drug targets for HbF induction. To test this hypothesis, we performed whole exome sequencing on 171 pediatric SCD patients to identify variants associated with endogenous HbF levels. Gene-based analysis identified seven unique non-synonymous variations in a Forkhead box O transcription factor, FOXO3, as significantly associated with lower HbF (p=5.6x10-4, β-value of log transformed (ln) HbF= -0.66). Two variants in the α2 subunit of AMPK, a FOXO3 activator, were also associated with reduced HbF (p=1.56x10-4, β-value ln%HbF=-1.5).

We then performed functional studies to verify the association between FOXO3 and endogenous HbF levels in an ex vivo model of erythroid differentiation from CD34+ cells isolated from peripheral blood of normal human blood donors. Lentiviral short hairpin RNA (shRNA) knockdown of FOXO3 reduced γ-globin expression from 1 to 0.4, p= 0.0005. While γ-globin expression and protein levels were reduced by FOXO3 knockdown, β-globin levels remained unchanged. These results suggest that FOXO3 is a positive regulator of γ-globin. Morphologic and flow cytometry analysis of primary erythroid culture with and without FOXO3 knockdown indicates that knockdown of FOXO3 delays erythroid maturation, while reducing γ-globin production. We therefore conclude that FOXO3 appears to regulate γ-globin through a specific mechanism rather than through alteration of erythroid maturation kinetics. FOXO3 is a viable therapeutic target for the treatment of individuals with SCD as well as those with quantitative hemoglobinopathies like β-thalassemia, who do not benefit from hydroxyurea due to its delay of erythropoiesis.

FOXO3 expression is known to be increased by three drugs, metformin, phenformin, and resveratrol. We found that these drugs also cause FOXO3 to accumulate in the nucleus, where it is active, rather than in the cytoplasm, where FOXO3 is degraded. We have investigated the effects of these agents on FOXO3 and γ-globin expression in K562 cells. Metformin, phenformin and resveratrol increased FOXO3 and γ-globin transcription levels in a dose-dependent manner. We then treated primary erythroid culture cells with a range of metformin doses (20-200µM), with and without a stable dose of 30 µM hydroxyurea. Alone, metformin had a modest effect (1.5 fold) on γ-globin induction at all concentrations. In combination with hydroxyurea, 50 µM metformin increased γ-globin expression 3.7-fold compared to 2.5-fold with hydroxyurea alone when analyzed by RT-qPCR. β-globin levels were unchanged by hydroxyurea or metformin. γ-globin induction persisted through terminal maturation of the culture when measured serially every 5 days. Taken together, our results indicate that not only is FOXO3 is a positive regulator of γ-globin expression, but it is an excellent therapeutic target for HbF induction. Metformin, a well-studied, well tolerated oral agent, will be investigated in combination with hydroxyurea in a phase II trial as an adjunctive agent to increase HbF induction.


Off Label Use: This presentation will discuss off-label use of metformin as a possible HbF inducing agent..

Author notes


Asterisk with author names denotes non-ASH members.