Development of Glypican-3-Targeting Antibody-Drug Conjugates for Hepatocellular Carcinoma Therapy

Abstract Hepatocellular carcinoma, a leading cause of global cancer mortality, is characterized by intratumoral heterogeneity, chemoresistance, and limited treatments. Glypican-3 (GPC3), an oncofetal glycoprotein overexpressed in ~70% of HCC cases but absent in normal adult tissues, represents an attractive target for antibody-drug conjugates (ADCs). To address this unmet need, three GPC3-specific antibodies were developed: a fully human Mab-A with moderate affinity (Kd=45.7 nM), recognizing a unique epitope distal to the GPC3’s membrane region; a humanized Mab-B with high affinity (Kd=8.84 nM), targeting an epitope similar to those of several known therapeutic antibodies; and a biparatopic BpAb-AB with the highest affinity (Kd=0.383 nM). Three payloads were used for conjugation: duocarmycin SA (DUBA), pyrrolobenzodiazepine (PBD) dimer and deruxtecan (Dxd). In HCC cell-derived xenograft (CDX) models, DUBA conjugates demonstrated satisfactory efficacy but less potent than other conjugates, whereas PBD dimer conjugates showed potent tumor suppression with narrow therapeutic windows due to dose-limiting toxicity. Mab-B conjugates exhibited comparatively lower efficacy and suboptimal pharmacokinetics. In contrast, Mab-A-Dxd and BpAb-AB-Dxd ADCs produced sustained tumor regression and acceptable safety. Despite lower binding affinity and in vitro activities, Mab-A-Dxd outperformed BpAb-AB-Dxd in vivo, driven by superior tumor penetration, prolonged exposure, and reduced target-mediated drug disposition (TMDD). Efficacy was further validated in patient-derived xenograft models. Collectively, the optimized Mab-A-Dxd ADC, engineered with high drug-to-antibody ratio (DAR~10) and Fc-silencing mutations, achieves a favorable balance between tumor penetration, minimized TMDD, and effective GPC3-mediated delivery, positioning it as a promising therapeutic candidate for refractory HCC and other GPC3-expressing malignancies.