Biology subjects

Selinger, O.

Publications and source records attributed to Selinger, O..

3 recordsLinked to original sources

bioRxiv2026

Systematic De-Risking of TCR-Mimic Therapeutics Through Proteome-Wide Off-Target Landscaping and a Generalizable Design Rule Framework

TCR-mimic (TCRm) antibodies targeting peptide-human leukocyte antigen (pHLA) complexes enable precision immunotherapy against intracellular antigens, including cancer-testis antigens (CTAs). Achieving high specificity, however, remains challenging because of the vast diversity of the human immunopeptidome and the associated risk of off-target recognition. Here, we introduce ValidaTe, a unified framework for the proteome-scale prediction, validation, and mitigation of off-target liabilities in pHLA-directed therapeutics. ValidaTe integrates rational target prioritization, peptide-centric binder selection, proteome-wide off-target prediction, and therapeutic engineering into a hierarchical de-risking workflow. Using the CTA MAGE-A4 as a proof-of-concept, we identify the TCRm antibodies VR-4 and VR-6 with superior specificity and demonstrate how this workflow enables the discovery of safer pHLA-targeted binders. Furthermore, ValidaTe establishes the basis for the WiFi (Widened Fingerprint) engineering principle, which rationally combines TCRms with complementary off-target fingerprints in trivalent T-cell engagers to minimize unintended interactions while preserving potent target-specific activity. Together, these findings establish a generalizable framework for the rational development of safer and more selective pHLA-targeted therapeutics. We further discuss how orthogonal proteomic characterization may complement this workflow as a final layer of translational safety assessment prior to clinical development. TeaserValidaTe accelerates safe pHLA-targeted immunotherapy through proteome-wide off-target mapping and WiFi design

immunology↗

bioRxiv2025

EpiTox: A Multi-Modular Framework for Population-Aware Off-Target Prediction Highlighting MAGE-A3 Cross-Reactivity

Targeting intracellular tumor antigens presented by peptide human leukocyte antigen (pHLA) complexes offers a promising immunotherapeutic strategy for patients with limited treatment options. However, development of pHLA-targeted drugs such as T-cell receptor (TCR) mimic antibodies (TCRm) and TCR-based therapeutics remains challenging due to severe offtumor and off-target toxicities arising from incomplete pHLA off-target profiling. The fatal neurological toxicities observed in the MAGEA3 TCR-T cell trial, potentially linked to unanticipated cross-reactivity with EPS8L2- and MAGEA12-derived peptides, underscore the urgent need for more comprehensive safety assessment. While current preclinical de-risking methods utilize sequence similarity searches, they lack the layered integration of genetic context, HLA binding profiles, and structured risk assessment needed to comprehensively evaluate peptide cross-reactivity. To address these limitations, we developed EpiTox, a computational multi-modular platform that systematically identifies and evaluates pHLA off-targets. EpiTox integrates proteome-wide sequence similarity, genetic context, HLA binding profiles, and layered risk assessment to provide a holistic evaluation of potential cross-reactive peptides. Using MAGEA3 as a model, EpiTox identified known toxic off-targets and predicted novel epitopes. Notably, an anti-MAGEA3 therapeutic TCRm antibody bound several newly predicted off-targets, including three SNP-derived peptides. By enabling comprehensive and predictive preclinical safety screening, Epi-Tox supports safer TCRm drug development and may help prevent future clinical failures.

bioinformatics↗

bioRxiv2025

Beyond Sequence Similarity: ML-Powered Identification of pHLA Off-Targets for TCR-Mimic Antibodies Using High Throughput Binding Kinetics

T-cell receptor mimic (TCRm) antibodies are an emerging class of tumor-targeting agents used in advanced immunother-apies such as bispecific T-cell engagers and CAR-T cells. Unlike conventional antibodies, TCRms are designed to recognize peptide-human leukocyte antigen (pHLA) complexes that present intracellular tumor-derived peptides on the cell surface. Due to the typically low surface abundance and high sequence similarity of pHLAs, TCRms require high affinity and exceptional specificity to avoid off-target toxicity. Conventional methods for off-target identification such as sequence similarity searches, motif-based screening, and structural modelling focus on the peptide and are limited in detecting cross-reactive peptides with little or no sequence homology to the target. To address this gap, we developed EpiPredict, a TCRm-specific machine learning framework trained on high-throughput kinetic off-target screening data. EpiPredict learns an antibody-specific mapping from peptide sequence to binding strength, enabling prediction of interactions with unmeasured pHLA sequences, including sequence-dissimilar peptides. We applied EpiPredict to two distinct TCRms targeting the cancer-testis antigen MAGE-A4. The model successfully predicted multiple off-targets with minimal sequence similarity to the intended epitope, many of which were experimentally validated via T2 cell binding assays. These findings establish EpiPredict as a valuable tool for lead optimization of TCRms, enabling the identification of antibody-specific off-targets beyond the scope of traditional peptide-centric methods and supporting the preclinical de-risking of TCRm-based therapies.

immunology↗