Inhibiting Efferocytosis in Acute Myeloid Leukemia Decreases Checkpoint Blockade through Decreased CCL5/STAT6 Signaling and Increases Activation through NF-Kb

Efferocytosis is a tolerogenic wound-healing process carried out by macrophages including phagocytosis of cellular debris and production of T cell suppressive cytokines. MerTK, the prototypic efferocytic receptor, is expressed on macrophages in tissues that harbor high leukemic burden, or "leukemia-associated macrophages" (LAMs), and therefore efferocytosis is co-opted by leukemia for immune evasion. We previously reported that in acute myeloid leukemia (AML) models blocking efferocytosis through inhibition of MerTK led to increased serum [IL-18] and downregulation of checkpoint ligands PD-L1 and PD-L2 on LAMs, which was not attributable to decreased [INF-γ] given that INF-γ was actually increased with MerTK inhibition.

To evaluate a potential mechanism of PD-L1/PD-L2 downregulation when efferocytosis is inhibited, we evaluated for IL-4, IL-10 and IL-6 - all previously associated with checkpoint ligand expression - in serum samples of C57Bl/6 mice inoculated with a syngeneic MLL-ENL AML, and treated with a MerTK efferocytosis inhibitor (MRX-2843, n=15) or vehicle (PBS, n=14). There was no appreciable [IL-4] or [IL-10] detected in either treatment group. There was decreased serum [IL-6] in some MRX-2843-treated samples (mean: 45.4pg/ml) compared to controls (mean: 379.1pg/ml, p=0.3), however nearly all treated and untreated samples were <40pg/ml. We then evaluated several additional serum cytokines from these mice. CCL5/RANTES, known to promote M2 macrophage polarization and modulate PD-L1/PD-L2, was significantly decreased in mice treated with MRX-2843 (12.92pg/ml) compared to vehicle (28.63pg/ml, p<0.001). To validate this as relevant signaling pathway in LAMs, we evaluated STAT6 phosphorylation by western blot analysis in bone marrow derived macrophages (BMDM) from C57Bl/6 mice co-cultured with AML cells and treated with MRX-2843 or vehicle. When MerTK was inhibited, STAT6 phosphorylation was consistently decreased compared to vehicle-treated BMDM. This CCL5/STAT6 axis alteration may be a potential mechanism of PD-L1/PD-L2 downregulation when efferocytosis is inhibited.

We next sought to evaluate whether there was LAM autocrine cytokine signaling within tissues of high leukemic burden, which were not detected by serum cytokine analysis. BMDM co-cultured with AML cells, and treated with MRX-2843 or vehicle were harvested after 12 hours and RT-qPCR was performed. mRNA levels of IFN-β and IL-1β in BMDM treated with MRX-2843 were twice that of controls, suggesting classic macrophage activation. These BMDM did not amplify appreciable levels of IL-4, IL-10, or TGF-β mRNA.

Given the increased IL-1β mRNA and previously reported increased IL-18 associated with MerTK inhibition in LAMs, we evaluated NF-κB activation and subsequent inflammasome assembly. By RT-qPCR we did not detect any consistent alteration in p105 (precursor of p50) or RelA (p65), however when BMDM co-cultured with AML cells and treated with MRX-2843 or vehicle, western blot analysis consistently demonstrated increased phosphorylation of p65 as well as increased NLRP3 when MerTK was inhibited.

In conclusion, blocking efferocytosis through MerTK in LAMs decreased CCL5/STAT6 signaling, which likely contributes to decreased LAM PD-L1 and PD-L2 expression. Furthermore, MerTK inhibition led to activation of NF-κB p65, inflammasome assembly, and production of IL-1β and IL-18. Taken together, these data demonstrate a mechanism by which MerTK inhibition alters antigen presentation through decreased co-inhibition and augmented activating cytokine production. Given that MerTK inhibitors are currently in clinical trials for relapsed/refractory malignancies, these data are relevant, timely, and could provide additional justification for their use in acute leukemias.