Pan-cancer analysis of antibody-drug conjugate targets and putative predictors of treatment response
Carlo Bosi,Áron Bartha,Barbara Galbardi,Giulia Notini,Matteo M Naldini,Luca Licata,Giulia Viale,Marco Mariani,Barbara Pistilli,H Raza Ali,Fabrice André,Marta Piras,Maurizio Callari,Marco Barreca,Alberta Locatelli,Lucia Viganò,Carmen Criscitiello,Lajos Pusztai,Giuseppe Curigliano,Balázs Győrffy,Matteo Dugo,Giampaolo Bianchini
DOI: https://doi.org/10.1016/j.ejca.2023.113379
Abstract:Background: Antibody-drug conjugates (ADCs) are a rapidly expanding class of compounds in oncology. Our goal was to assess the expression of ADC targets and potential downstream determining factors of activity across pan-cancer and normal tissues. Materials and methods: ADCs in clinical trials (n = 121) were identified through ClinicalTrials.gov, corresponding to 54 targets. Genes potentially implicated in treatment response were identified in the literature. Gene expression from The Cancer Genome Atlas (9000+ cancers of 31 cancer types), the Genotype-Tissue Expression database (n = 19,000 samples from 31 normal tissue types), and the TNMplot.com (n = 12,494 unmatched primary and metastatic samples) were used in this analysis. To compare relative expression across and within tumour types we used pooled normal tissues as reference. Results: For most ADC targets, mRNA levels correlated with protein expression. Pan-cancer target expression distributions identified appealing cancer types for each ADC development. Co-expression of multiple targets was common and suggested opportunities for ADC combinations. Expression levels of genes potentially implicated in ADC response downstream of the target might provide additional information (e.g. TOP1 was highly expressed in many tumour types, including breast and lung cancers). Metastatic compared to primary tissues overexpressed some ADCs targets. Single sample "targetgram" plots were generated to visualise the expression of potentially competing ADC targets and resistance/sensitivity markers highlighting high inter-patient heterogeneity. Off-cancer target expression only partially explains adverse events, while expression of determinants of payload activity explained more of the observed toxicities. Conclusion: Our findings draw attention to new therapeutic opportunities for ADCs that can be tested in the clinic and our web platform (https://tnmplot.com) can assist in prioritising upcoming ADC targets for clinical development.