Comprehensive self-antigen screening to assess cross-reactivity in promiscuous T-cell receptors

T cell receptor (TCR) therapeutics are an emerging modality of biologic and cell-based medicines with the unique ability to target intracellular antigens and finely discriminate between healthy and infected or mutated cells. An obstacle to the development of new TCR therapeutics is the difficulty in engineering these proteins for enhanced therapeutic efficacy while avoiding introduction of unexpected off-target autoreactivity. In this study, we conduct functional high-throughput screening to profile all possible genome-encoded peptides for ability to trigger response in an engineered candidate therapeutic TCR and assess risk of off-target toxicity. We used multiple approaches for constructing comprehensive self-antigen cell libraries and highlight key considerations for designing toxicity screening campaigns using high-throughput TCR profiling in a live cell context. We then use Tope-seq to screen epitope libraries against a model therapeutic candidate TCR and show that this strategy can be used to detect known cross-reactive epitopes from libraries of >5 × 10⁵ unique peptide-coding sequences at a significance threshold of p < 0.01 in first-pass bulk screening. We also incorporate strategies for iterative biopanning and bioinformatic refinement to improve sensitivity and accuracy and demonstrate here the first proof-of-principle for functional TCR screening on a library of >2 × 10⁷ peptide-coding DNA fragments, further advancing the potential of <i>in vitro</i> approaches to perform unbiased TCR epitope discovery.