Induction of Fetal Hemoglobin by Inactivation of HDAC1 or HDAC2 without Altering Cellular Proliferation

Abstract 354

Histone deacetylase (HDAC) inhibitors are effective inducers of fetal hemoglobin, and prior studies have shown that selective inactivation of HDAC1 or HDAC2 is sufficient to induce fetal hemoglobin in vitro. In our current work, we demonstrate that HDAC1 and HDAC2 are attractive targets for clinical translation for two reasons: 1) Selective inhibition will decrease off-target effects that currently limit the use of hydroxyurea and pan-HDAC inhibitors, and 2) HDAC inhibitors induce fetal hemoglobin in various preclinical models, and they can be combined with hydroxyurea to achieve further fetal hemoglobin induction. To investigate off-target effects, we selectively inactivated HDAC1, HDAC2 or HDAC3 in human erythroid progenitor cells, and examined the effect of each knockdown on cellular cytotoxicity and cell cycle progression. Although knockdown of HDAC3 negatively influenced growth, selective knockdown of HDAC1 or HDAC2 had no effect on expansion of erythroid progenitors. In addition, knockdown of HDAC2 does not block cell cycle progression. These data support the possibility that an HDAC1- or HDAC2-specific inhibitor may offer a therapeutic advantage by reducing side effects, while maintaining robust HbF induction. Armed with this knockdown data, we are now investigating HDAC inhibitor compounds of various selectivity in in vitro and in vivo models. To perform optimal clinical trials, and ultimately to benefit the most sickle cell disease patients, it would be ideal to combine HDAC inhibitor treatment with hydroxyurea. A combination treatment approach may ameliorate some of the limitations of hydroxyurea use, such as the unpredictable effect on fetal hemoglobin levels, and the lack of benefit in beta thalassemia patients. First, we combined HDAC2 inactivation with hydroxyurea treatment in vitro. Human bone marrow-derived CD34+ cells were infected with lentiviruses containing an shRNA targeting either HDAC2 or a luciferase control gene. The cells were then treated on day 4 of erythroid differentiation with hydroxyurea (10–20 uM dose). Compared to the untreated luciferase control samples, we observed a 7–9-fold increase in gamma-globin expression in the untreated HDAC2-knockdown samples, a 2.5-fold increase in the hydroxyurea-treated luciferase control samples, and a trend toward an additive effect on gamma-globin induction in the cells where HDAC2 knockdown was combined with hydroxyurea treatment. To investigate the effects of HDAC inhibitors in vivo, we administered compounds to BCL11A conditional knockout transgenic mice (by erythroid-selective EpoR-GFP Cre) containing the human beta-globin locus. As reported previously, BCL11A inactivation powerfully de-repressed gamma-globin expression, and administration of an HDAC inhibitor, SAHA, led to a further elevation of gamma-globin mRNA. We now demonstrate that administration of another pan-HDAC inhibitor, panobinostat (LBH589), results in an additional 1.5- to 2.5-fold increase in gamma-globin mRNA relative to pre-treatment baseline. We are currently evaluating the combination of panobinostat and hydroxyurea in these mice to confirm that the compounds have an additive effect in vivo as well as in vitro. Taken together, these experiments indicate that inhibiting HDAC1 or HDAC2 is a promising therapeutic approach to increasing fetal hemoglobin levels in patients with beta-hemoglobinopathies, both alone and in combination with hydroxyurea.