开发用于高通量抗体-药物偶联物癌症筛选的3D细胞模型

Antibody-drug conjugates (ADCs) are a rapidly evolving class of biotherapeutics designed to deliver potent cytotoxic agents directly to cancer cells while minimizing systemic toxicity. ADCs offer this highly targeted approach by combining the selectivity of antibodies with the lethality of small-molecule drugs. However, their preclinical development still faces substantial challenges, including species-specific differences in antigen expression and antibody binding that complicate translational studies, often leading to discrepancies between preclinical efficacy and clinical outcomes. As a result, there is growing interest in developing more predictive preclinical models, including those that incorporate extracellular matrix (ECM) components into three-dimensional (3D) culture models which can help dissect how microenvironmental factors influence ADC target engagement. Such models provide a more physiologically relevant environment to assess ADC diffusion and payload delivery. Using Inventia Life Science's RASTRUM TM Allegro platform, we developed and validated the utility of high-throughput 3D bioprinted patient-derived colorectal cancer (CRC) models for the testing of a CEACAM5-targeted ADC in two CRC subtypes. These engineered hydrogel-based tumoroid models, using the Gibco TM OncoPro TM CRC Tumoroid Cell Lines (ThermoFisher Scientific), were tuned to mimic the stiffness and ECM composition of CRC tumors and maintained subtype-specific molecular and phenotypic features. This allowed comparisons of ADC efficacy versus standard-of-care chemotherapies and here we show clinically-relevant differential drug responses between subtypes, with the ADC demonstrating enhanced cytotoxicity over chemotherapy alone. We also provide evidence for the ease of assessing target expression in situ by immunofluorescence imaging of models. Furthermore, the ability to test on-target cytotoxicity dependence was demonstrated by blocking target binding with anti-CEACAM5, which reduced ADC cytotoxicity in the sensitive tumoroid models. This scalable and reproducible workflow is applicable across many cancer types, and bridges the gap between oversimplified in vitro assays and complex in vivo systems, offering more predictive insights into ADC performance by using models that incorporate physical and biochemical features relevant in patient tumors.