Distinct Cell-Intrinsic and Extrinsic Consequences of TLR2 Activation Drive Pleiotropic Effects in AML Affecting Overall Survival

Aberrant inflammatory signaling is a hallmark of myeloid malignancies and is generally thought to promote disease. Few studies have systematically dissected the molecular and cellular mechanisms of pathogen-induced inflammation in myeloid disease settings. Using our previously established immune-competent mouse model of DNTM3A/FLT3-mutant acute myeloid leukemia (FD AML), we screened a panel of synthetic Toll-like receptor (TLR) agonists to mimic pathogen-specific responses. Activation of TLR2 heterodimers, but not other TLRs, significantly extended overall survival of leukemia bearing mice. Moreover, TLR2 is the highest expressed TLR in AML patients regardless of mutation profile. These data led us to explore the impact of TLR2 mediated inflammatory signaling on AML disease processes.

To understand AML-cell intrinsic versus extrinsic effects, we conducted a series of reciprocal syngeneic transplants using FD AML and Tlr2^-/-FD AML into wild-type (WT) or Tlr2^-/- recipients. Note, Tlr2 deletion does not significantly alter the development or progression of FD AML. After AML establishment, mice were treated with a single dose of Tlr1/2 specific agonist, Pam₃CSK₄. This revealed Tlr2 expression on AML cells is essential for significantly prolonged survival with Pam₃CSK₄ treatment. Given that TLRs have direct and indirect signaling with functional phenotypic consequences, we employed single-cell RNA sequencing coupled with cell surface proteins (CITE-seq) to unbiasedly investigate the pleiotropic effects of Tlr2 activation. We identified 21 clusters that recapitulate the hierarchical nature of FD AML. Pam₃CSK₄ significantly decreased the fraction of AML cells in stem/progenitor clusters, expanded AML-derived monocyte precursor clusters, and gave rise to a new AML-derived macrophage cluster. There were also decreased AML-derived mature neutrophil clusters despite expansion of neutrophil precursor clusters. We validated these AML cell states using the reciprocal Tlr2 proficient/deficient transplantation models. Our data indicate that Tlr2 activation drove effector macrophages (CD11b⁺F4/80⁺ MHCII⁺ CD80⁺/86⁺) with elevated phagocytic activity from AML. In contrast, Pam₃CSK₄ blocked maturation of AML-derived neutrophils in an AML-non-autonomous manner. Cytokine analyses showed G-CSF production downstream of Tlr2 activation in the microenvironment prevented the cell intrinsic Tlr2-mediated neutrophil maturation of AML cells. By CITE-seq, Cepbewas significantly reduced in AML-derived neutrophil precursors, indicating the block by G-CSF may be happening at a precursor stage. Indeed, FD AML co-treated with anti-G-CSF neutralizing antibodies and Pam₃CSK₄ restored AML-derived neutrophil differentiation, but at the expense of monocyte differentiation, through a bipotential intermediate. Lastly, to understand whether Tlr2 activation on AML cells causes a transient reduction in stem/progenitor numbers or a sustained functional impairment of AML stem cell activity, we transplanted limiting numbers of FD AML stem/progenitor cells from Pam₃CSK₄ and vehicle treated mice into WT recipients. We observed a 5-fold decrease in leukemia initiating cell activity with Pam₃CSK₄ treated donors. In summary, we discovered Tlr2 activation induces pleiotropic effects on a variety of cell types in AML, ultimately impacting overall survival. Our results shed light on the functional consequences of pathogen recognition by AML cells that may be leveraged for novel treatments of this deadly disease.