LSD1 Inhibition Induces Differentiation and Reduces the Frequency of Leukemia-Initiating Cells in Hoxa9/Meis1-Induced Acute Myeloid Leukemia

The histone demethylating enzyme LSD1 has emerged as a promising epigenetic target in the treatment of acute myeloid leukemia (AML). Inhibition of LSD1 has been shown to induce differentiation and facilitate the responsiveness of AML cells to all-trans retinoic acid (ATRA). As it remains unclear how LSD1 inhibition induces differentiation, we have now used a murine leukemia model based on retroviral overexpression of Hoxa9 and Meis1 to study the effect of of both genetic inactivation as well as pharmacological inhibition of LSD1.

We transformed bone marrow cells from conditional LSD1 knock-out (KO) mice to confirm the responsiveness of Hoxa9/Meis1-induced leukemias to an inhibition of LSD1. Marked granulocytic differentiation was observed in these cells upon induction of the LSD1 KO. The cells also showed a decrease in expression of c-Kit by 53 %. Furthermore the cells acquired the mature granulocytic marker Ly6G. We also found an increased sensitivity to a treatment with ATRA upon induction of the LSD1 KO. We then transplanted Hoxa9/Meis1-transformed cells from conditional LSD1 KO mice into lethally irradiated mice and induced the LSD1 KO 20 days after transplantation by treatment with tamoxifen. This extended the survival of mice from a median of 48 days for control-treated mice (N=5) to a median of 78 days for tamoxifen-treated mice (N=5) (p=0.0021).

We then treated Hoxa9/Meis1-transformed myeloid progenitors with three irreversible (tranylcypromine, trans-N-((2-methoxypyridin-3-yl)methyl)-2-phenylcyclopropan-1-amine (ORY86, AW69), (1S,2R)-2-phenyl-N-(piperidin-4-ylmethyl)cyclopropanamine (AW84)) and three reversible LSD1 inhibitors (MS120 (=Namoline), MS142, SP2509) alone or in combination with ATRA. We observed signs of granulocytic differentiation with the irreversible inhibitors and a reduction in c-Kit expression from 92% to 25% after treatment with AW69. Colony formation of Hoxa9/Meis1-transformed myeloid progenitors was reduced by 70% compared to control cells after treatment with AW69 for 96 hrs. Treatment with AW69 and AW84 also enhanced the response of leukemia cells to ATRA by enhancing the anti-proliferative effect and by reducing colony formation by 99% in Hoxa9/Meis1 cells in the combination. The reversible LSD1 inhibitors exhibited only minor effects on differentiation even when used at effective concentrations.

We also analyzed the effect of LSD1 inhibition on leukemia-initiating cell frequency. We therefore treated Hoxa9/Meis1-transformed myeloid progenitors with AW69 in vitro for a total of 96 hrs and then assessed the LIC frequency in a limiting-dilution assay. In this assay, AW69 induced a striking reduction in LIC frequency from 1/40 (95 % CI 1/10.7 to 1/175) to 1/3200 (95 % CI 1/835.5 to 1/12101) (p < 0.001).

As the differentiating effect of LSD1 inhibition in our model was limited to the irreversible inhibitors we then wanted to investigate if the effect was due to interference with the enzymatic activity of LSD1. We therefore used the combination of Hoxa9 and Meis1 to transform cells from mice carrying a conditional knock-in (KI) of an enzymatically inactive form of LSD1 (LSD1mut) and assessed its consequence both in vitro and in vivo. Similar to the effect seen with the conditional KO we observed granulocytic differentiation upon induction of the conditional LSD1mut KI as well as a downregulation of c-Kit expression. However, we did not find a survival benefit in vivo upon induction of the conditional LSD1mut KI, but observed a more differentiated immunophenotype in the developing leukemias.

In summary, LSD1 depletion or inhibition using tranylcypromine or its derivatives induces granulocytic differentiation, reduces LIC frequency and induces a survival benefit in Hoxa9/Meis1-driven AML. Our data further suggest that these effects of LSD1 inhibition are mediated by interfering with both enzymatic and scaffolding functions of LSD1.