Heme-Regulated Inhibitor (HRI) Depletion Cooperates with Pharmacologic Inducers to Strongly Elevate Fetal Hemoglobin and Reduce Sickle Cell Formation

Increasing fetal hemoglobin (HbF) provides significant clinical benefit for patients with sickle cell disease (SCD) and is a critical goal that is being pursued with a variety of pharmacologic strategies. Hydroxyurea (HU) is currently the only FDA-approved drug aimed at raising HbF but is limited in its efficacy. Recent work in our laboratory identified heme-regulated inhibitor (HRI), an erythroid-specific protein kinase, as a novel HbF regulator (Grevet et al., Science 2018). Depleting HRI in primary human erythroid cells significantly raised HbF levels; however, maximal HbF induction by HRI knockdown required at least 80-85% HRI depletion. It is currently unknown whether this degree of HRI inhibition can be achieved pharmacologically. Besides HU, several compounds have been recently identified as in vitro HbF inducers, including pomalidomide, a third-generation imide, and UNC0638, an EHMT 1/2 histone methyltransferase inhibitor. We therefore set out to test whether combining HRI depletion with pharmacologic HbF inducers would cooperatively increase HbF levels and diminish potential adverse effects on erythroid cell maturation and viability.

We performed three-stage in vitro culture of human CD34+ cells treated with select pharmacologic inducers of HbF, including HU, pomalidomide, or UNC0638, in combination with shRNA knockdown of HRI. HbF levels were assessed by RT-qPCR, Western blot, flow cytometry, and HPLC. We found that treatment with HU displayed comparatively little activity and failed to increase effects of HRI knockdown. However, treatment with UNC0638 combined with HRI depletion showed greater than additive effects on HbF with levels reaching 25-30% HbF, while HRI depletion combined with pomalidomide treatment showed the highest levels of cooperativity, reaching 30-40% HbF. Combination of HRI depletion and HbF pharmacologic inducers showed minimal effects on the erythroid transcriptome by RNA-Seq and did not significantly impair erythroid maturation. Intriguingly, dose-titration experiments indicated that HRI knockdown sensitizes erythroid cells to low doses of pomalidomide, maintaining HbF levels of greater than 40% in HRI-depleted samples despite a ten-fold decrease in pomalidomide concentration. Diminished expression of the HbF repressor BCL11A accounted in large part for HbF induction in HRI knockdown samples, particularly when combined with pomalidomide treatment in which BCL11A depletion exceeded 90%, while other HbF repressors such as LRF were unchanged. We will present RNASeq analyses aimed at elucidating mechanisms of HRI cooperativity. Finally, we found that HRI depletion in SCD patient-derived cells showed significant cooperativity with pomalidomide and UNC0638, achieving 45-50% HbF for UNC0638-treated and 50-60% HbF for pomalidomide-treated drug combinations. Importantly, combination of HRI knockdown and HbF pharmacologic treatments markedly reduced in vitro sickling as measured by low-oxygen sickling assays, suggesting significant amelioration of the sickle cell phenotype.

In sum, we find that dual targeting of HbF induction via HRI inhibition and pharmacologic inducers results in successful cooperative upregulation of HbF levels in both normal and SCD primary human cells without impairing red cell maturation. Furthermore, our data suggest that dose titration of HbF inducers combined with HRI depletion could maximize HbF induction while potentially reducing off-target effects. Moreover, HRI is an attractive target for HbF induction as it is expressed in an erythroid-specific manner. While no specific HRI inhibitors are currently available, our work suggests that future small molecule inhibitors of HRI may be combined with other pharmacotherapies to achieve significant, clinically meaningful HbF induction for the treatment of SCD and other hemoglobinopathies.