Pharmacological Inhibition Of Histone Deacetylase (HDAC) 1, 2 Or 3 Have Distinct Effects On Cellular Viability, Erythroid Differentiation, and Fetal Globin (HbG) Induction

Induction of HbG is an established therapeutic strategy for the treatment of sickle cell disease (SCD), and could also be effective in treating beta-thalassemia (bT). Genetic ablation of HDAC1 or HDAC2, but not HDAC3, results in the induction of HbG expression (Bradner JE, Proc Natl Acad Sci, 2010). Furthermore, we have previously shown that selective chemical inhibitors of HDAC1 and 2 elicit a dose and time dependent induction of HbG mRNA and fetal hemoglobin (HbF) protein in cultured human CD34+ bone marrow cells undergoing erythroid differentiation (Shearstone JS, ASH Annual Meeting Abstracts, 2012). While a variety of selective HDAC inhibitors have been used successfully to induce HbF, further clinical development has been limited by variable efficacy and concerns over off-target side-effects observed in clinical trials, potentially due to inhibition of HDAC3. Additionally, it remains to be determined if HDAC1 or HDAC2 is the preferred therapeutic target. In this work we present data that investigates the effects of selective inhibitors of HDAC1, 2, or 3 on cytotoxicity, erythroid differentiation, and HbG induction in cultured human CD34+ bone marrow cells.

Acetylon Pharmaceuticals has generated a library of structurally distinct compounds with a range of selectivity for each of HDAC1, 2, or 3 (Class I HDAC) as determined in a biochemical assay platform. From our initial chemical series, we identified ACY-822 as a Class I HDAC inhibitor with IC₅₀ values of 5, 5, and 8 nM against HDAC1, 2, and 3, respectively. In contrast, ACY-1112 is 30-fold selective for HDAC1 and 2, with IC₅₀ values of 38, 34, and 1010 nM against HDAC1, 2, and 3, respectively. Treatment of cells for 4 days with ACY-822 (1 μM) resulted in a 20-fold decrease in cell viability, while ACY-1112 (1 μM) treatment resulted in a minimal reduction in viability (1.2-fold) and a 2-fold increase in the percentage of HbG relative to other beta-like globin transcripts. This result suggests that pharmacological inhibition of HDAC3 is cytotoxic and is consistent with the therapeutic rationale for the design selective inhibitors of HDAC1 and 2.

To investigate if HDAC1 or HDAC2 is the preferred therapeutic target, we utilized a second series of structurally distinct compounds. We identified ACY-957 as an HDAC1/2 selective compound biased towards HDAC1 with IC₅₀ values of 4, 15, and 114 nM for HDAC1, 2, and 3, respectively. In contrast ACY-1071 showed balanced HDAC1 and 2 selectivity with IC₅₀ values of 27, 24, and 247 nM for HDAC1, 2, and 3, respectively. Treatment of cells for 6 days with 1 μM of ACY-957 or ACY-1071 resulted in a 3-fold increase in the percentage of HbG relative to other beta-like globin transcripts. However, we found that ACY-957 treatment resulted in an approximately 3-fold decrease in cell viability after 6 days of treatment, while ACY-1071 treatment resulted in a minimal reduction (1.2-fold) in cell viability. Decreased cell viability observed with ACY-957 was associated with a reduction of cells positive for the erythroid differentiation markers CD71 and glycophorinA. This result is consistent with the Mx-Cre mouse model where HDAC1^KO; HDAC2^het had reduced numbers of erythrocytes, thrombocytes, and total bone marrow cells, while the HDAC1^het; HDAC2^KO was unaffected (Wilting RH, EMBO Journal, 2010).

Our results suggest that compounds with a pharmacological profile of increased selectivity towards HDAC2 inhibition versus HDAC1 may be less cytotoxic and minimize effects on differentiation, while still inducing HbG in human CD34+ bone marrow cells.