Abstract
Optimizing target engagement for bispecific antibodies (bsAbs) is challenging, as bsAbs bind to two different targets in vivo to form a trimolecular complex (trimer), and likely display a bell-shaped concentration-response curve. bsAbs trimer formation is expected to be related to efficacy and safety, and determining the dose level at which it is achieved may support dose selection; however, trimer levels currently cannot be quantified in vivo. GEN1042 (DuoBody®-CD40x4-1BB)-a novel, agonistic bsAb-combines targeting and conditional activation of co-stimulatory molecules CD40 and 4-1BB on immune cells, resulting in enhanced priming and reactivation of tumor-specific immunity. We describe data and models that supported GEN1042 dose selection for expansion, including a minimal physiologically based pharmacokinetic and receptor occupancy (mPBPK/RO) model that can predict levels of GEN1042 trimers crosslinked to CD40 and 4-1BB in tumors and lymph nodes to guide dose selection. The model leverages PK data, physiologic parameters, and in vitro data from the literature for parameterization of GEN1042 kinetics and distribution, its interaction with target receptors CD40 and 4-1BB, and their expression on the surface of various cell types. Trimer level outcomes were further assessed with clinical data for dose selection. An initial expansion dose of GEN1042 100 mg every 3 weeks was chosen based on available clinical safety/tolerability, efficacy, PK/pharmacodynamics data, mPBPK/RO, and exposure-response analysis. This dose level is currently being evaluated in combination with pembrolizumab, with or without chemotherapy, in the expansion phase of the GEN1042 phase 1/2 trial (NCT04083599); other alternative doses are currently being explored.