Abstract 4034: Computational ranking-guided discovery of Plexin-B2 as a driver of circulating tumor cell clusters and AI-facilitated neutralization to prevent and inhibit metastasis

Abstract Our previous studies and other works demonstrate that multicellular circulating tumor cell (CTC) clusters are up to 50 times more efficient than single CTCs in mediating viable metastasis (Cancer Discovery, 2018, 2023). Yet, the molecular targets underlying this phenomenon and the CTC cluster interactions with immune cells remain poorly understood. To uncover key mediators of CTC clustering and metastatic progression, we developed a computational ranking pipeline integrating adhesion network analysis and metastasis-free survival modeling. This analysis prioritized Plexin-B2 (PLXNB2) as a top-ranked transmembrane protein associated with poor distant metastasis-free survival and enriched expression in CTC clusters from patients with advanced breast cancer. Functional studies demonstrated that loss of PLXNB2 (Plxnb2) markedly reduces both homotypic tumor-tumor and heterotypic tumor-myeloid (monocyte) cluster formation, thereby suppressing spontaneous metastasis in vivo. Mechanistically, PLXNB2 engages SEMA4C on tumor cells and SEMA4A on myeloid cells to drive intercellular adhesion and metastatic colonization. Integrative proteomic and pathway network analysis further identified downstream PLXNB2 signaling effectors regulating cell adhesion and stemness plasticity, corroborating its importance as a central hub for clustering. To therapeutically block PLXNB2 functions in metastasis, we obtained 106 anti-PLXNB2 monoclonal antibodies (mAbs) for functional screening and optimization, both computationally and experimentally. Using artificial intelligence (AI)-driven structural platforms and language models, we assessed the IgG binding and neutralizing possibilities in clustering and proliferation assays. Experimental screen of top candidates revealed that several antibodies effectively disrupted both homotypic tumor-tumor and heterotypic tumor-monocyte clustering. Furthermore, we implemented an AI-guided structural modeling workflow to predict the PLXNB2-IgG complex, identify active binding residues, and optimize antibody-antigen energetics for enhanced affinity and neutralization. Together, these findings identify PLXNB2 as a key molecular regulator of CTC clustering and metastatic seeding and establish a framework for structure-guided antibody engineering to therapeutically target CTC cluster-driven breast cancer metastasis. Citation Format: Fangjia Tong, Martina Veit Acosta, Yuanfei Sun, Yuzhi Jia, Guofu Hu, Huiping Liu. Computational ranking-guided discovery of Plexin-B2 as a driver of circulating tumor cell clusters and AI-facilitated neutralization to prevent and inhibit metastasis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 4034.