Integrative in silico prioritization of tumor-associated antigen pairs to support bispecific antibody-drug conjugate design in solid tumors.

Abstract

e15036 Background: Antibody–drug conjugates (ADCs) are an established therapeutic modality in solid tumors, although tumor heterogeneity and antigen loss can limit response durability. Bispecific ADCs targeting two tumor-associated antigens (TAAs) represent a promising strategy to improve tumor selectivity and potentially mitigate resistance mechanisms, and integrative in silico analyses may help inform the prioritization of biologically plausible TAA pairs. Methods: TAAs corresponding to ADCs approved by the FDA and/or EMA for solid tumors were selected as reference targets, including HER2, TROP2, NECTIN4, Tissue Factor, FOLR1 and c-MET. Gene expression correlations with an extended panel of clinically relevant TAAs were assessed using TIMER 3.0 across TCGA solid tumors. High-confidence associations were defined using a Spearman correlation coefficient ≥0.7, with tumor types annotated using standard TCGA nomenclature. Results: High-confidence co-expression patterns were identified between selected reference TAAs and an extended panel of TAAs across TCGA solid tumors; correlations observed in more than one tumor type were considered recurrent and are summarized in Table 1. HER2 showed recurrent high-confidence co-expression with ERBB3, together with additional tumor-specific associations. TROP2 and NECTIN4 displayed overlapping recurrent co-expression patterns with multiple partner antigens, alongside additional tumor-specific correlations. In contrast, Tissue Factor and FOLR1 exhibited exclusively tumor-specific high-confidence associations. c-MET demonstrated recurrent co-expression with several partner antigens, together with tumor-specific associations. Conclusions: This integrative in silico analysis identifies high-confidence co-expression patterns between clinically relevant TAAs across solid tumors, enabling the prioritization of biologically plausible TAA pairs. The identification of both recurrent and tumor-specific associations highlights distinct co-expression landscapes that may help inform pan-tumoral and indication-driven bispecific ADC design. TAA co-expression patterns supporting bispecific ADC design (r ≥ 0.7). Reference TAA Recurrent correlated TAAs Tumor-specific correlated TAAs TCGA tumor types HER2 ERBB3 B7-H3, TPBG, PTK7 PAAD, BLCA, THYM, KIRP, THCA TROP2 NECTIN4, ERBB3, FUT3, LYPD3, MUC1, NaPi-2b ITGB6, EGFR, FGFR2/3, Tissue Factor, CEACAM5, B7-H4 ESCA, SKCM, THCA, UCS, THYM, KIRP, HNSC, CESC NECTIN4 TROP2, LYPD3, ERBB3, FUT3 c-MET, EFNA4, EGFR, CDH3, B7-H4, SLC44A4, Tissue Factor ESCA, SKCM, THCA, UCS, PAAD, THYM Tissue Factor – AXL, ITGB6, LYPD3, NECTIN4, ROR1, TROP2 THYM, SKCM, TGCT FOLR1 – NaPi-2b, RORC, B7-H4, ROR2 LUSC, PRAD, UVM c-MET ADAM9, AXL, CD46, EGFR ERBB3, ITGB6, NaPi-2b, FGFR2, TPBG, CD71 LIHC, STAD, THCA, PRAD, UVM, KIRP, KICH, THYM