整合的类器官筛选方法可实现对蛋白降解剂疗效和靶点结合的临床前评估

Introduction Proteolysis-targeting chimeras (PROTACs) and molecular glues are novel therapeutics designed for selective protein degradation, offering new avenues for targeted cancer treatment. Two cereblon (CRBN)-mediated protein degrader candidates, CC-885 (GSPT1 degrader; molecular glue) and ARV-471 (estrogen receptor (ER) degrader; PROTAC), were evaluated using patient-derived organoids to address the unmet need for physiologically relevant platforms in preclinical protein degrader development. Methods: CC-885 was screened using the OrganoidXplore™ CellTiterGlo assay at 6 doses across 98 NGS characterized organoids derived from six tissue types (colorectal, cervical, gastric, head and neck, lung, and pancreatic cancer), including non-diseased models. ARV-471 efficacy was assessed in ten organoid models of ER-sensitive indications (breast, ovarian) in high content imaging (HCI) assays. CC-885 sensitivity was validated in HCI assay for a selection of models, and GSPT1 and ER degradation were assessed in immuno-fluorescent (IF) assays. CC-885-sensitive models were selected to generate CRISPR/Cas9-mediated CRBN knockout organoid models. (Patient-matched) PDX models were subcutaneously engrafted in immunodeficient mice, and in vivo efficacy was evaluated by tumor growth inhibition (TGI) measurement. Results: Large panel organoid screening of CC-885 revealed heterogeneous tumor responses (IC50s ranging from ~0.1nM to 50nM) that did not correlate to GSPT1 or CRBN mRNA expression levels. In isogenic KO organoid models, loss of CRBN abrogated CC-885 responses, confirming E3-ligase dependency. HCI studies validated CC-885 responses and quantified target engagement and degradation through IF staining. ARV-471 efficacy testing in ER positive organoids by HCI demonstrated both robust activity and ER degradation. The integration of efficacy and target engagement through HCI provided a mechanistic link between observed responses and molecular mode-of-action. Lastly, findings from organoid screens were compared to responses in matched patient-derived xenograft (PDX) models. Conclusion: Organoids offer genomically stable, patient-derived models that capture patient diversity and tumor heterogeneity, enabling clinically translatable insights for precision medicine. Our results establish organoid-based screening platforms (OrganoidXploreTM), combining CTG and HCI modalities, as robust tools for preclinical protein degrader evaluation. These platforms uniquely enable identification of tumor subtypes likely to respond to targeted degradation and will potentially facilitate translational biomarker discovery through simultaneous efficacy and mechanistic readouts. PDXO to PDX follow-up opportunities further underscore the utility of these platforms for guiding early therapeutic development and target selection in oncology.