Preclinical evaluation of the nectin-4 antibody-drug conjugate MK-3120 (SKB410) in a panel of solid-tumor patient-derived xenograft models

Despite clinical success of some antibody-drug conjugates (ADCs), much remains unknown about what ultimately drives tumor response. Therefore, identification and preclinical testing of selection biomarkers to enrich for responders is critical. MK-3120 (SKB410) is a nectin-4-directed ADC with a unique, bifunctional linker that maximizes payload delivery to tumor cells both through its irreversible connection with a novel anti-nectin-4 monoclonal antibody and pH-sensitive cleavage from a belotecan-derived topo I inhibitor payload in the lysosome. To define molecular determinants of response to nectin-4-targeting ADCs, we evaluated MK-3120 in 48 patient-derived xenograft (PDX) models from indications with documented nectin-4 expression (primarily urothelial, breast, and non-small cell lung cancer). In order to investigate differences in preclinical response biology, we tested enfortumab vedotin in parallel as a comparator ADC that targets the same antigen but delivers MMAE, a microtubule inhibitor. Baseline characterization included nectin-4 immunohistochemistry (IHC; H-score), nectin-4 gene copy number variation (CNV), and RNA sequencing. In a subset of models, nectin-4 CNV and IHC were reassessed after treatment to detect selection-driven changes in surviving tumor tissue. Using a predefined response threshold, 22 of 48 models (46%) responded to MK-3120, while 11 models (23%) responded to enfortumab vedotin; nine models responded to both ADCs. These results indicate tumor-intrinsic heterogeneity in ADC sensitivity and demonstrate that targeting the same antigen does not guarantee cross-response between ADCs with different payloads and linker chemistries. Nectin-4 copy number amplification has been suggested to enrich for responders in clinical trials of nectin-4 targeting agents. Therefore, we investigated the impact of CNV on efficacy of both MK-3120 and enfortumab vedotin. Of the models that responded to MK-3120 and enfortumab vedotin, 15 and 7 models harbored nectin-4 gene amplifications, respectively. In the post-treatment analyses, nectin-4 protein expression (IHC H-score) and CNV showed no consistent, widespread changes after treatment with either ADC. Together, these data demonstrate overlapping but nonidentical responder sets indicating that payload and linker differences shape intrinsic ADC sensitivity, and that biomarker strategies should consider both antigen status and payload-relevant features.