CD8+ and CD4+ T cells recognize peptides presented on cell-surface major histocompatibility complex (MHC) class I and class II molecules, respectively, to survey the intracellular and extracellular landscape for pathogens and assess cellular health. The mechanism by which T cells recognize specific peptide–MHC (pMHC) combinations is through their membrane-bound T cell receptor (TCR). TCR recognition of the cognate pMHC results in T cell activation and induction of various effector functions that are critical for mounting an adaptive immune response.
Knowledge of the antigens that T cells recognize through their TCR is foundational for our understanding of how and why they are involved in the pathology of human diseases such as cancer and autoimmunity. Moreover, harnessing their exquisite specificity holds therapeutic potential and is the basis for successful vaccines and adoptive T cell therapies. However, the inherent diversity of the TCR repertoire, MHC alleles and the peptides that can be presented on any given MHC molecule make the task of mapping T cell antigens a complex problem. In addition, TCRs can be polyspecific, recognizing multiple pMHC combinations, and typically have lower affinity for their antigens (micromolar) compared to antibody–ligand interactions (picomolar to nanomolar). Over the past decade, we and others have developed strategies to determine TCR specificities against tumors, self-antigens, pathogens and allergens but few methods have demonstrated success in assessing TCR reactivities at the proteome scale, hindering high-throughput and unbiased antigen discovery efforts. Furthermore, antigen discovery technologies have largely focused on class I-restricted human TCRs and equivalent methods to interrogate class II-restricted CD4+ T cells or assess mouse TCR reactivities have not kept pace. Ultimately, a universal method that has utility for both human T cell antigen discovery efforts and other frequently used preclinical model organisms, such as mice, and that can be applied for class I-restricted and II-restricted TCRs is highly desirable.
Our lab previously developed T-Scan, which can perform genome-wide analysis of TCR specificities across both viral and human proteomes. T-Scan works by using patient or donor T cells programmed with the TCR of interest in a screen against a library that expresses protein fragments presented on MHC molecules in target cells containing a granzyme reporter. Upon T cell recognition of target cells expressing the cognate antigen, granzyme B is cytosolically delivered to target cells, which activates the fluorescent reporter for harvesting by cell sorting.
A limitation of this system is the need to obtain fresh primary T cells for each screen. In addition, the assay kills the target cells, limiting the opportunity for further enrichment and subsequent rescreening for signal amplification. Thus, we were motivated to devise an improved screening method that did not rely directly on patient T cells and killing of target cells as part of the recognition assay.
Here we describe a new, cell-based T cell antigen discovery method called TCR mapping of antigenic peptides (TCR-MAP). TCR-MAP enables TCRs with unknown specificity to be queried against a large, peptide tiling library of antigen-presenting cells (APCs) expressing processed peptides on patient-specific or mouse-specific MHC alleles. This system relies on a synthetic circuit expressed in Jurkat cell lines transduced with the TCRs of interest. Upon T cell recognition of the cognate pMHC, the bacterial transpeptidase sortase A (SrtA) is induced and expressed on the cell surface of Jurkats and covalently biotinylates the reciprocal cognate APC. This new method is high throughput and can capture unbiased reactivities against the complete human, mouse or viral proteome or any genetically encoded peptide library of choice. Moreover, it is highly sensitive and can reproducibly discover both high-affinity and low-affinity TCR antigens. We demonstrate the utility of TCR-MAP for antigen discovery efforts for both CD8+ and CD4+ T cells and for TCRs derived from humans or mice. Application of this technology has the potential to enhance T cell antigen discovery efforts in the context of cancer, infectious disease and autoimmunity.
To establish a highly sensitive and specific reporter system that would capture cognate TCR–pMHC interactions, we selected a previously reported proximity labeling strategy using the Staphylococcus aureus transpeptidase SrtA, which covalently transfers substrates containing the polypeptide motif LPXTG to nearby N-terminal oligoglycine residues. We designed a two-cell method consisting of immortalized Jurkat T cells expressing a genetically fused mouse CD40 ligand–SrtA (mCD40L–SrtA) construct under an inducible nuclear factor of activated T cells (NFAT) promoter to serve as the donor (SrtA-Jurkats) and human leukocyte antigen (HLA) class I-null HEK-293T APCs transduced with an N-terminal oligoglycine-tagged mouse CD40 receptor (G 5-mCD40) as the SrtA substrate acceptor (G 5-targets) (Fig. 1a). Jurkat cells can be easily engineered to express TCRs of interest and additional TCR signaling components such as the CD8 coreceptor. Target cells can be further transduced with the desired MHC alleles and antigens encoded as peptide fragments or full-length proteins (FL-ORF (open reading frame)) for presentation. Upon T cell activation, mCD40L–SrtA is induced on the Jurkat cell surface and catalyzes the transfer of exogenously added LPETG–biotin substrates onto cognate target cells expressing the G 5-mCD40 acceptor (Fig. 1a and Extended Data Fig. 1b). This method, which we call TCR-MAP, relies on a TCR-stimulated circuit in immortalized T cells to activate sortase-mediated biotinylation of cognate APCs.
a, Schematic of the TCR-MAP antigen discovery method. Target APCs expressing HLA alleles of interest, the TCR-MAP biotin acceptor (G 5-mCD40) and peptides or proteins of interest were cocultured with Jurkat cells transduced with TCRs of interest, coreceptors and the NFAT-inducible SrtA reporter (mCD40L–SrtA). Upon TCR activation, mCD40L–SrtA was expressed on the cell surface of Jurkats and covalently transferred LPETG–biotin substrates to cognate target cell acceptors. b, Representative flow cytometry plots of HLA-A2+ G 5-target cells pulsed with or without the NLVPMVATV peptide cocultured with NLV3 TCR+ SrtA-Jurkat cells. Antigen recognition was assessed using streptavidin fluorophores to label biotinylated target cells. c, Quantification of overall target cell biotinylation when G 5-targets expressed either the restricted (HLA-A*02:01) or nonrestricted (HLA-A*01:01) HLA allele and were transduced with either a 56-aa polypeptide containing the pp65 antigen NLVPMVATV or the full-length protein (FL-ORF) and cocultured with NLV3 TCR+ SrtA-Jurkat cells. ****P< 0.0001 for each group relative to the nonrestricted HLA allele control, determined by one-way ANOVA with a Tukey–Kramer multiple-comparison test. d, Quantification of target cell biotinylation when cocultured with the indicated antigen-positive target cell and TCR+ SrtA-Jurkat. ****P< 0.0001 for each group relative to the no-antigen control, determined by a two-tailed t-test. e, Schematic of TCR-MAP for class II-restricted TCRs. f, Quantification of target cell biotinylation when cocultured with the indicated antigen-positive target cell and TCR+ SrtA-Jurkat for class II-restricted T cells. Each dot in c, d and f represents a different biological replicate, where error bars indicate the mean and s.d. The data in b–d and f are representative of n = 3 independent biological replicates.
