This project utilizes mechanistic pharmacokinetic / pharmacodynamic modeling to guide the development of new
therapeutic strategies, and new therapeutic agents, to improve the safety and efficacy of antibody-drug conjugate
(ADC) therapy for acute myeloid leukemia (AML). The project will generate new monoclonal antibodies and
nanobodies with specificity for three surface proteins overexpressed in AML (CD123, CLL1, TIM3). The new
antibodies and nanobodies will be used as targeting vectors to deliver candidate payload molecules (MMAE,
DM4, SN38, Dxd) to AML cells (Aim #1). Within experiments proposed in Aim #2, the targeting vectors will be
assembled into bispecific constructs to test the hypothesis that bispecific ADCs enable improved efficacy and
decreased off-site on-target toxicity. Aim #3 will investigate a newly developed inverse targeting strategy, where
co-treatment with payload binding antibody fragments (i.e., Payload Binding Selectivity Enhancers, PBSE) is
employed to decrease the delivery of payload molecules to healthy tissues, enabling reduced off-site off-target
toxicity. PBSE with high affinity for MMAE, DM4, SN38, and Dxd will be evaluated for utility in preventing the off-
site, off-target toxicity that results from the exposure of healthy cells to released (i.e., “free”) payload. Due to the
selective binding of PBSE to free payload, with little or no binding of PBSE to intact ADCs, we hypothesize that
our new PBSE agents will allow decreased toxicity of anti-AML ADC therapy, without negatively impacting ADC
efficacy. Aim #4 will perform in vivo experiments in mouse models of human AML to evaluate the effects and
toxicities of monospecific anti-AML ADCs and bispecific ADCs, with or without cotreatment with PBSE. Due to
our development of several targeting vectors, each with a wide range of possible characteristics (e.g., affinity,
modality [nanobody, mAb], bispecificity), and due to our intent to consider several payload molecules, with or
without PBSE co-treatment, many permutations may be considered. Additionally, complete and appropriate
experimental evaluation of relationships between ADC attributes and therapeutic utility is not possible in animal
models, due to the unavailability of models with appropriate co-expression patterns of human antigens on healthy
cells, AML bulk cells, leukemic blasts, and leukemic stem cells. To address these complexities and limitations,
mechanistic PKPD modeling and simulation will be employed throughout the project to predict effects and
toxicities in mouse models and in AML patients, to facilitate engineering efforts (e.g., predicting relationships
between affinity of bispecific binding arms and therapeutic selectivity), and to guide selection of constructs for in
vivo evaluations.