(W-004) Development of QSP model of non-small cell lung cancer and its application to support optimal dose selection for acasunlimab, a bispecific antibody targeting 4-1BB and PD-L1

Alexandra Diakonova, N/A – N/A, InSysBio CY; Sergey Smirnov, N/A – N/A, InSysBio CY; Anna Roskoshnaia, N/A – N/A, InSysBio CY; Ivan Borisov, N/A – N/A, InSysBio CY; Patricia Garrido Castro, N/A – N/A, Genmab; Nora Pencheva, N/A – N/A, Genmab; Maria Jure-Kunkel, N/A – N/A, Genmab; Beesan Tan, N/A – N/A, BioNTech US; Tim Maiwald, N/A – N/A, BioNTech US; Manish Gupta, N/A – N/A, Genmab; Oleg Demin, N/A – N/A, InSysBio UK; Craig Thalhauser, N/A – N/A, Genmab

Objectives: Acasunlimab (GEN1046/BNT 311; DuoBody^®-PD-L1×4-1BB) is an investigational bispecific antibody immunotherapy that combines conditional 4-1BB activation with simultaneous and complementary PD-L1 blockade. The aims of this study were to develop a Quantitative Systems Pharmacology (QSP) model of non-small cell lung cancer (NSCLC) that implements mechanisms of action of acasunlimab for dose optimization.

Methods: A QSP model of NSCLC was developed on the basis of a QSP model of HNSCC published in [1]. Acasunlimab was implemented as follows:

1. PK and distribution of acasunlimab is described by a minimal PBPK model;


2. Acasunlimab binds to PD-L1 located on macrophages, dendritic (mDC), cancer, endothelial, B and T cells and/or 4-1BB expressed by T cells and NK cells;


3. Acasunlimab is eliminated via FcRn-dependent consumption by endothelial cells and internalization of its complexes with PD-L1 and 4-1BB;


4. Trans-trimeric complex PD-L1:acasunlimab:4-1BB is formed in immunological synapses between antigen-presenting and T cells, cancer and T cells, and mDC and NK cells;


5. PD-L1:acasunlimab:4-1BB trimer stimulates proliferation of T and NK cells. Parameters describing the effect of trimer on cell dynamics were identified using in vitro data.

A virtual patient (VP) population was generated to simultaneously describe publicly available clinical data (RECIST 1.1 criteria, TtR, DoR, PFS) for patients with different TPS status treated with pembrolizumab or atezolizumab alone and combinations of cisplatin + pemetrexed and cisplatin + pemetrexed + pembrolizumab.

Results: The VP population was validated against acasunlimab PK data and then used for dose selection. We found that across a range of PD-L1 expression levels (i.e. different TPS levels), the PD-L1 receptor occupancy in the immunological synapse between CD8⁺ T cells and tumor cells achieved with atezolizumab 1200 mg Q3W can be achieved at acasunlimab doses of 200 mg Q3W or higher. However, in terms of CD8⁺ T-cell activation in the tumor, lower doses of acasunlimab are required to achieve the same effect as atezolizumab, and a dose of 50 mg Q3W in VPs across different TPS levels is sufficient. The maximum increase in CD8⁺ T cells in the tumor was predicted to occur at a dose of 100 mg in the VP. With respect to PD-L1 receptor occupancy, in which higher doses will lead to higher receptor occupancy, the model predicts fewer T cells in the tumor at doses greater than 100 mg due to decreasing trimers caused by a hook effect.

Conclusions: The developed QSP model can adequately describe the effects of various published drug therapies on progression and outcomes of NSCLC and make predictions on possible PD effects of novel treatments. The model supports efforts to optimize the dose selection for the bispecific antibody acasunlimab.

Citations: [1] Diakonova A, et al. J Immunother Cancer 2023;11:doi: 10.1136/jitc-2023-SITC2023.0874