Targeting Protein Arginine Methyltransferases to Eradicate Acute Myeloid Leukemia

Project Summary Treatment outcomes for patients with acute myeloid leukemia (AML) have continued to lag behind outcomes reported for other hematological malignancies like acute lymphoblastic leukemia (ALL), in part because of the relatively slow development of immunotherapy (including immune checkpoint inhibitors [ICI]) compared with ALL.

However, translation of currently available immunotherapies to AML treatment has been challenging, and novel AML-targeting drugs with immune-stimulating activity are greatly needed. The cGAS-STING signaling is a major pathway that promotes an extrinsic type I Interferon (IFN-I) response, which potently primes T cells in “immune-

cold” cancers, including AML. My laboratory has been active in defining the function of protein arginine methyltransferases (PRMTs) in leukemia. Herein, our preliminary studies on the basis of primary AML cells as well as murine AML models reveal an anti-AML immune activation seen after inhibition of PRMT9, a most

recently defined PRMT. Specifically, PRMT9 inhibition stimulated leukemia-intrinsic cGAS, as evidenced by cGAMP (an immunotransmitter) production, and induced a leukemia-eliminating IFN-I response in murine AML- microenvironment (ME). When combined with an ICI (PD1 inhibitor), the in-house PRMT9 inhibitor LD2

eradicated AML in animal models. Thus, we hypothesize that PRMT9 activity allows AML cells to evade immune surveillance by repressing cGAS-STING activity in the ME, and that LD2, alone or combined with an ICI (PD1 inhibitor), elicits T cell activity to eliminate AML cells. Extracellular cGAMP, which is hydrolyzed by ENPP1, and

STING activation in host immune cells, are reportedly essential for immunity. Thus, in Aim 1, using two AML transplant models, we will determine whether downregulating cGAMP (via either ENPP1-overexpression or cGAS-knockout (KO) in AML cells) or STING-KO in the immune cell compartment of recipient mice would reverse

PRMT9 inhibition-induced AML regression. In Aim 2, we will determine the mechanisms underlying cancer- intrinsic cGAS activation seen after PRMT9 inhibition. Specifically, we will define PRMT9 targets that when unmethylated underlie cGAS activation. In Aim 3, we will assess anti-AML activity of LD2 alone or combined with

a PD1 inhibitor, using AML mouse models. Moreover, to define potential shifts in immune cell types/states after PRMT9 inhibition, we will perform single-cell transcriptome analysis on AML PDX cells from humanized mouse treated with LD2 or the combination. We are the first to identify PRMT9 as a druggable immune activation target

against cancer. If successful, this work would support combining a PRMT9 inhibitor with ICIs as a therapy for AML, in which single ICI therapy has very limited effects.