In vitro and In Vivo FOXO3 dependent induction of fetal hemoglobin Free

Background: Induction of fetal hemoglobin (HbF) is a key therapeutic strategy to treat sickle cell disease (SCD). Unfortunately, the only widely used therapy to induce HbF, hydroxyurea (HU), is clinically effective in only half of adults with SCD. Efforts to develop novel HbF inducing agents have been stymied by an incomplete understanding of gamma-globin regulation. We identified Forkhead Box O3 (FOXO3) and Insulin-like growth factor binding protein 3 (IGFBP3) as positive regulators of gamma–globin in two independent genomic studies. We have shown that metformin can induce HbF in erythroid culture in a FOXO3 dependent manner. A subsequent clinical trial of metformin in SCD failed to show a HbF benefit due to poor adherence secondary to gastrointestinal side effects of metformin. IGFBP3 is produced by the Kupffer cells in the liver; the addition of exogenous IGFBP3 to erythroid culture increases FOXO3 and HbF. Of note, vitamin D supplementation increases plasma levels of IGFBP3. Here, we seek to determine if the HbF induction by IGFBP3 is FOXO3 dependent, and if metformin can induce HbF in an NBSGW mouse model engrafted with human hematopoietic stem cells (HSPCs). Methods: Two phase erythroid culture was initiated using CD34+ cells extracted from red cell exchange bags from two unique SCD patients. On day 5 of culture, the cells were transduced with lentiviral FOXO3 shRNA and scramble shRNA. Purified IGFBP3 (1µg/ml) was added to the culture on day 7 and cells were harvested on day 19 and HbF levels analyzed by HPLC; FOXO3, pFOXO3(Ser 413), Ƴ-Globin, and β-Globin protein levels were determined by western blot analysis. Further we did a pilot study to establish a good xenograft mouse model for studying HbF regulation and to see the impact of metformin on HbF in xenografts. NBSGW mice were engrafted with 1E6 wild type HSCs per mouse by retro orbital injections following busulfan conditioning and fed on control (open std. diets) or 0.5% metformin chow (n=3 in each group) starting post 4 weeks engraftment. One additional mouse in the control group was treated with pomalidomide (10mg/Kg IP daily) to serve as a positive control for HbF induction, chosen as it was deemed less likely to disturb the human CD34+ xenograft than HU. Bone marrow was harvested 15 weeks post engraftment and analyzed for engraftment levels by flow cytometry for human CD45+ and for HbF induction by F-cell flow. A linear mixed model statistical analysis was performed using GraphPad Prism. Results: An 88% knockdown was achieved using FOXO3 shRNA 81 (clone ID: V3LHS_375381) and 71% using FOXO3 shRNA 62 (clone ID: V3LHS_641762) in erythroid precursors derived from HSPCs from two unique SCD donors. IGFBP3 treatment significantly induced HbF (P< 0.001) and increased protein levels of FOXO3 (P< 0.0001), the phosphorylated form of FOXO3 that localizes to the nucleus (pFOXO3 Ser 413) (P< 0.0001); this induction was not seen in erythroid precursors that underwent FOXO3 knockdown. There was no significant difference in the protein levels of HbS from treated or untreated IGFBP3 control or FOXO3 knockdown HSPCs. NBSGW mice were successfully engrafted with wild type HSCs, with more than 90% human CD45+ cells. The median of %F-Cells for the mice fed 0.5% metformin chow was significantly higher 52.50% (34-70%) compared to the mice fed on control chow 6.53% (5-18%). Conclusions: Our study in erythroid culture confirmed the FOXO3 dependence of IGFBP3 mediated HbF induction as FOXO3 knockdown removes IGFBP3's ability to increase HbF levels in erythroid culture. This supports our proposed HbF induction model that IGFBP3 mediated HbF induction occurs by activation of FOXO3. We showed that the NBSGW mouse model engrafted with human HSCs can serve as an excellent model for testing drug induction of HbF in vivo. Metformin treated NBSGW xenograft mice showed an increase in %F-cells supporting HbF induction in vivo, as demonstrated by the induction of HbF observed with both metformin and pomalidomide. Next steps will include determining if manipulation of vitamin D levels in NBSGW mice with human HSCs modulates HbF levels via alternation of IGFBP3.