Increased levels of fetal hemoglobin are associated with decreased symptoms and increased life span in patients with sickle cell disease (SCD). Hydroxyurea, the only drug currently approved for SCD, is not effective in a large fraction of patients and therefore new agents are currently needed. Recent evidence has shown that LSD1, an enzyme that removes monomethyl and dimethyl residues from the lys4 residue of histone H3, is a repressor of γ-globin expression. Tranylcypromine (TCP), an LSD1 inhibitor, was shown to increase γ-globin expression in human βYAC transgenic mice (Shi et al Nat Med 19:291, 2013). Because the arrangement and developmental stage-specific expression pattern of the β-like globin genes is highly conserved between man and other simian primates, the use of an simian primate animal model such as the baboon (P.anubis) is the best predictor of the activity of HbF-inducing agents in man. In this investigation we compared the effect of TCP and the more potent and selective LSD1 inhibitor, RN-1, on HbF expression in anemic baboons (P. anubis). In vitro assays have shown that the LSD1 IC50 of RN-1 is at least 1000 fold less than TCP. Animals were phlebotomized for 14d prior to drug treatment to attain an Hct=20 to induce reticulocytosis and establish baseline HbF levels and were maintained at this Hct by periodic phlebotomies during the course of the experiment. In four baboons treated with varying doses of TCP (2-6mg/kg/; 10-20d; sc) low levels of HbF (4.9-7.9% HbF) were induced that were only slightly higher than those observed at the pretreatment baseline (2.2-4.1% HbF). In contrast, treatment with varying doses of RN-1 (2.5-0.125 mg/kg/d; 5-10d) induced high levels of HbF, F reticulocytes, and F cells in 5 of 6 animals (see Table). At high doses of drug the ratio of 5'Iγ/3'Vγ synthesis was >2 demonstrating a near-complete reversion to the pattern of fetal stage expression. Peak levels of F reticulocytes and γ-globin synthesis were observed 8d and HbF and F cells 11d after the first day of drug administration. Increased γ-globin mRNA levels (γ/γ+β) in reticulocytes measured by RT-PCR showed that increased HbF levels were not due to translational effects. Bisulfite sequence analysis showed that levels of DNA methylation of the γ-globin promoter were similar in pre- and post-treatment BM erythroid precursors from two animals. Flow cytometry analysis using anti-α4-integrin and anti-baboon RBC antibodies showed that RN-1 treatment altered terminal BM erythroid differentiation by increasing the proportion of less differentiated precursors, however no changes in MCHC or total hemoglobin synthesis (α/γ+β) were observed. RN-1 treatment was associated with decreased ANC (290-870 X 103/μl nadir) and increased platelets (1092-1445 X 103/μl peak) and monocytes that were likely caused by effects on hematopoietic differentiation. The ANC nadir and peak platelet and monocyte counts were observed between d14-19 and resolved with 2-3 days. Similar changes in ANC and platelets, although not in monocytes, are observed following treatment with decitabine and can be controlled by modification of dose and schedule of administration (Lavelle et al Blood 119:1240, 2012). We conclude that RN-1, a more potent LSD1 inhibitor than TCP, is a powerful HbF-inducing drug with activity similar to decitabine and predict that LSD1 inhibitors may be useful drugs for the treatment of sickle cell disease.

Table

HbF, F-cells, and F-retics in Baboons Treated with RN-1

Animal RN-1 Dose
(mg/kg/d)
HbF (%) Globin synthesis (γ/γ+β) F-cells (%) F-retics (%)
PrePostPrePostPrePostPrePost
8548 2.5 (4d) 5.8 27.3 ND 0.78 28.5 59.2 34.6 92.3 
8549 0.5 (5d) 2.4 29.5 0.06 0.68 19.1 60.8 33.9 97.5 
8000  0.25 (5d) 4.1 20.6 0.15 0.49 20.3 47.0 45.7 80.7 
8548  0.20 (5d) 3.7 20.5 0.04 0.52 36.9 54.8 36.9 89.2 
8001  0.125 (5d) 1.8 4.8 ND 0.06 13.5 21.8 22.9 31.7 
8549 0.125 (10d) 3.6 16 0.04 0.22 17.6 42.7 21.0 70.0 
Animal RN-1 Dose
(mg/kg/d)
HbF (%) Globin synthesis (γ/γ+β) F-cells (%) F-retics (%)
PrePostPrePostPrePostPrePost
8548 2.5 (4d) 5.8 27.3 ND 0.78 28.5 59.2 34.6 92.3 
8549 0.5 (5d) 2.4 29.5 0.06 0.68 19.1 60.8 33.9 97.5 
8000  0.25 (5d) 4.1 20.6 0.15 0.49 20.3 47.0 45.7 80.7 
8548  0.20 (5d) 3.7 20.5 0.04 0.52 36.9 54.8 36.9 89.2 
8001  0.125 (5d) 1.8 4.8 ND 0.06 13.5 21.8 22.9 31.7 
8549 0.125 (10d) 3.6 16 0.04 0.22 17.6 42.7 21.0 70.0 

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution