Preclinical characterization of a potential best-in-class camptothecin-based antibody-drug conjugate targeting ADAM metallopeptidase domain 9

Introduction: ADAM metallopeptidase domain 9 (ADAM9), a member of the disintegrin and metalloproteinase, mediates proteolytic shedding of cytokines and growth factors. ADAM9 dysregulation promotes tumor progression, metastasis, and pathological angiogenesis. ADAM9 is overexpressed in various solid tumors, including pancreatic, lung, gastric, and breast cancers, with low expression in normal adult tissues. Previously, the clinical development of ADAM9-targeted ADC IMGC-936 was terminated due to maytansinoid-related ocular toxicities. Herein, we developed a next-generation ADAM9 ADC using a camptothecin-based linker-payload, which is anticipated to avoid ocular toxicity and possess expanded therapeutic index. Methods: ADC cytotoxicity was evaluated using the CellTiter-Glo (CTG) assay in Jurkat (ADAM9-negative), Calu-3 (intermediate ADAM9 expression), and NCI-H1693 (high ADAM9 expression) cell lines. Bystander activity was assessed in Jurkat cells co-cultured with CHOK1-hADAM9 cells, with Jurkat viability determined by CTG. ADC stability in human and rat plasma was monitored over 14 days, with payload release quantified by LC-MS. Accelerated stability was examined by incubating ADCs at 25 °C for 7 days, followed by drug antibody ratio (DAR) and size exclusion chromatography monomer fraction analysis. In vivo efficacy was investigated in Calu-3 (non-squamous NSCLC, EGFR wildtype) and DLD-1 (CRC, MSI-H, KRAS G13D) cell line-derived xenograft (CDX) mouse models. A PK study was conducted in Sprague-Dawley rats administered intravenously with ADAM9 ADC 5 mg/kg. Results: The parental antibody exhibited nanomolar affinity for human and cynomolgus ADAM9, with no cross-reactivity to rodent orthologs or other ADAM family members. Antibody internalization in MKN-45, MiaPaCa-2, and Calu-3 cells were time-dependent and highly efficient (>50% at 4 h), while the ADC showed comparable or enhanced internalization. The ADAM9-ADC demonstrated potent cytotoxicity and bystander killing effects. In Calu-3 CDX models, a single 3 mg/kg dose elicited 109.3% tumor growth inhibition (benchmark ADC, 91.1%). In DLD-1 CDX models, TGIs reached 90.2% and 66.8% at 3 and 8 mg/kg, respectively, significantly outperforming the benchmark (39.2%) at 3mg/kg. The ADC also displayed excellent plasma stability, with minimal DAR reduction and aggregation in accelerated and freeze-thaw stability assays. The ADAM9 ADC also exhibited a favorable pharmacokinetic profile in rats, characterized by minimal free toxin release and antibody-like half-life. Conclusion: Preclinical studies reveal that the ADAM9-ADC exhibits potent in vitro cytotoxicity and superior in vivo antitumor efficacy relative to benchmarks. Pilot toxicology studies in cynomolgus monkeys are ongoing. Further clinical development is warranted.