LSD1 Inhibition Leads to Differentiation in Hoxa9/Meis1- but Not in MN1-Induced Acute Myeloid Leukemia

Lysine specific demethylase 1 (LSD1) represents a promising epigenetic target in the treatment of acute myeloid leukemia (AML) and inhibition of LSD1 has been demonstrated to facilitate a response of AML cells to all- trans retinoic acid (ATRA). We have previously shown that LSD1 depletion or inhibition using tranylcypromine or its derivatives induces granulocytic differentiation, reduces leukemia-initiating cell (LIC) frequency and leads to a survival benefit in Hoxa9/Meis1 (H/M)-driven AML. We have now modeled the effect of pharmacological or genetic inactivation of LSD1 in another murine leukemia model based on the retroviral overexpression of MN1.

We transformed lineage-depleted bone marrow cells from conditional LSD1 knock-out (KO) mice with MN1 and induced the LSD1 KO by treatment with 4-hydroxy-tamoxifen (4-OHT). As a control we compared the results observed in this system with results from cells transformed with the combination of Hoxa9 and Meis1 (H/M). We analyzed short-term proliferation based on XTT assays and differentiation based on cell morphology as well as surface marker expression. We observed a 61% reduction in proliferation in H/M cells, while we only found 28% reduction in proliferation in MN1 cells. LSD1 KO induced marked granulocytic differentiation in H/M cells, but we did not observe any signs of granulocytic differentiation in MN1 cells. While LSD1 KO led to a reduction in the expression of c-Kit and upregulation of Ly-6G in H/M cells, c-Kit expression remained unchanged in MN1 cells. There were also no changes in Gr1 and Mac1, but MN1 cells acquired Sca1 and FcεR1. LSD1 inhibition has been shown to enhance the sensitivity of leukemic cells towards a treatment with ATRA. We therefore now compared the influence of LSD1 KO on the effect of ATRA treatment in H/M and MN1 cells. We saw a 95 % reduction in proliferation in H/M cells with LSD1 KO upon treatment with 0.1 µM of ATRA. In MN1 cells with LSD1 KO we only observed a reduction in proliferation of 28 % at this concentration.

We previously found a differentiating effect of irreversible LSD1 inhibitors in H/M transformed cells. We therefore also tested different potent LSD1 inhibitors (derivatives of tranylcypromine) in MN1 cells (trans-N-((2-methoxypyridin-3-yl)methyl)-2-phenylcyclopropan-1-amine (ORY86, AW69), (1S,2R)-2-phenyl-N-(piperidin-4-ylmethyl) cyclopropanamine (AW84), GSK-LSD1). Again we observed no signs of granulocytic differentiation, but an acquisition of the surface markers Sca1 and FcεR1 without concomitant changes in the expression of c-Kit, Gr1 and Mac1. Treatment with AW69 leads to a strong reduction in LIC frequency in the H/M model system. We now investigated if pre-treatment with AW69 in vitro for 96 h also affected the ability of MN1 cells to induce leukemia in mice. In this experiment we observed no change in leukemia development upon transplantation of 20.000 cells per mouse. However while 2 of 3 mice transplanted with control treated cells developed leukemia upon transplantation of 2.000 cells per mouse, no leukemia development was seen at this cell dose with AW69-treated cells. We therefore cannot exclude a moderate effect on LIC frequency in the MN1 model.

In order to test if our findings can be transferred to human AML we characterized the expression of MN1, HOXA9 and MEIS1 in 6 primary human AML cell cultures. We then treated these cells with GSK-LSD1 and assessed the expression of CD86, a known differentiation marker induced by LSD1 inhibition. In these experiments we observed a particularly strong upregulation of CD86 in samples expressing higher levels of MEIS1.

In summary, our results show that while pharmacological or genetic inactivation of LSD1 potently induces differentiation in H/M-induced AML, it is not sufficient to induce differentiation in MN1-driven AML. This finding could be relevant for selecting patients for treatment with LSD1 inhibitors and warrants further molecular investigations.