面向平台化PBPK-QSP模型用于GenSci139的首次人体剂量选择——GenSci139是一种潜在同类最佳的EGFR×HER2双特异性ADC，用于癌症患者

Background: GenSci139 is a novel bispecific ADC (BsADC) targeting both EGFR and HER2 by carrying a cleavable linker conjugated to a topoisomerase I inhibitor as payload. Preclinical studies have showed broad-spectrum efficacy in different tumors models. To support first-in-human (FIH) study (NCT07230977) dose selection, a multiscale whole body physiologically based pharmacokinetics (PBPK) - quantitative systems pharmacology (QSP) model was established to predict the effective dose range in advanced solid tumors pts. Methods: The multiscale PBPK-QSP model was developed using PK-Sim, MoBi and R softwares, integrating the systemic pharmacokinetics (PK) of the BsADC, its payload and naked antibody (nAb) with the key steps of FcRn-mediated recycle, EGFR/HER2 dual-target binding, crosslinking internalization, intracellular trafficking, payload release, and the subsequent tumor growth inhibition (TGI). The model also integrated the synthesis and degradation of EGFR/HER2 targets, and the competitive binding dynamics of the BsADC and nAb for EGFR/HER2 targets across tumor and various healthy tissues. Physiological parameters were utilized to enable cross-species translation. Monkey PBPK model of GenSci139 and clinical PBPK models of trastuzumab (mAb targeting HER2) and cetuximab (mAb targeting EGFR) were developed to enable the estimation of human EGFR/HER2 expression across tissues. To acquire the TGI parameter for human model, we developed the mouse PBPK-QSP model of GenSci139 based on plasma PK and tumor PK of BsADC, total antibody (Tab) and payload, and TGI data from two bladder cancer cell line derived xenograft (CDX) mice models (RT112/84 and HT-1376). These multiscale parameters were then integrated to construct a whole-body PBPK-QSP model for bladder cancer pts. Virtual clinical trial simulations were conducted to predict the PK and efficacy of GenSci139. Results: The PBPK model in monkeys accurately captured the plasma PK of GenSci139 BsADC, Tab and payload across dosing regimens. Similarly, PBPK models of trastuzumab and cetuximab successfully simulate the plasma PK in cancer pts at multiple dose levels. The mouse PBPK-QSP model of GenSci139 effectively predicted plasma PK, tumor PK and TGI data in CDX models. Using proposed human PBPK-QSP model, the efficacious does range was predicted in cancer pts. The efficacious dose was defined as dose achieving 100% TGI relative to baseline tumor size. Conclusions: The whole-body PBPK-QSP model was developed to mechanistically elucidate the dose-exposure-efficacy relationship of GenSci139. This model will incorporate toxicity predictions for a comprehensive dose-exposure-efficacy/toxicity profile as clinical data emerging. It represents a step toward an integrated platform PBPK-QSP model capable of facilitating ADC design, lead candidate selection, and clinical dose optimization.