ABSTRACT CD33-targeted therapies have long been pursued in acute myeloid leukemia (AML). Longer survival of some patients treated with the antibody-drug conjugate gemtuzumab ozogamicin (GO) validates this approach, but many patients with CD33+ AML do not benefit from GO. This has prompted efforts to develop better CD33- directed therapeutics, including T cell engaging bispecific antibodies (BsAbs). Several agents have recently entered early phase clinical testing, with initial results indicating some efficacy but also substantial toxicities from cytokine-release syndrome associated with T cell activation. There is therefore increasing focus on exploring the utility of other immune cells such as natural killer (NK) cells (e.g. via engagement of CD16), which might circumvent these limitations. How CD33/CD16-directed therapy can be optimized is unknown. As one limitation of CD33-targeted therapy, CD33 antibodies (including GO) typically recognize immune-dominant epitope(s) within the membrane-distal V-set domain. In our preliminary studies with CD33V-set/CD3 BsAbs and artificial CD33 proteins, however, we have observed enhanced T cell-mediated cytotoxicity with membrane-proximal binding of CD33. We have therefore generated a series of antibodies against the membrane-proximal C2-set domain of CD33 that recognize all naturally occurring variants of CD33 (i.e. are “CD33PAN antibodies”) as basis for novel therapeutics. Our previous studies have also shown that the anti-tumor efficacy of T cell-directed BsAbs is abrogated by inhibitory T cell co-receptor signaling, e.g. via the PD-L1/PD-1 axis. We further demonstrated that the use of a paired BsAb which binds and blocks such an inhibitory signal while, in turn, providing co- stimulation to T cells (e.g. via cross-linking CD28) can convert cellular inhibition into activation and dramatically increase the efficacy of T cell engaging BsAbs. So far, the role of activating and inhibitory NK receptors and their corresponding ligands as modulators of BsAb-based NK cell therapy has not been evaluated. Based on our studies with T cell-engaging BsAbs, we hypothesize CD33/CD16-directed NK cell engaging therapy can be optimized by membrane-proximal targeting of CD33. We also predict that activating and inhibitory NK cell ligands modulate the anti-tumor efficacy of CD33/CD16 BsAbs and that, consequently, the resistance to CD33/CD16 BsAbs can be reversed by pairing with a BsAb that binds/blocks an inhibitory NK cell ligand and stimulates an activating NK cell receptor. We will test these hypotheses in well-controlled preclinical studies in vitro and in vivo. Upon completion of the proposed research, it is our expectation that we have gained critical insight into how the anti-AML efficacy of CD33/CD16 NK cell-engaging therapeutics can be maximized. Our work is anticipated to have an important positive impact because it may form the basis for optimized NK cell engaging therapeutics for patients with AML and other CD33+ ...