Biology subjects

Rossjohn, J.

Publications and source records attributed to Rossjohn, J..

At least 19 recordsLinked to original sources

bioRxiv2026

Adaptive-like features of the γδ TCR couple chronic BTNL recognition to NK-like tissue immunity

Intestinal V{gamma}4 {gamma}{delta} intraepithelial lymphocytes (IELs) persistently bind the constitutively expressed epithelial ligand BTNL3/8 through germline-encoded T cell receptor (TCR) determinants. While they resemble innate-like T cells such as NKT cells, unlike these populations, V{gamma}4 IELs encounter ligand only after thymic development. Moreover, unlike conventional {beta} T cells, V{gamma}4 IELs sustain persistent physiological ligand engagement without becoming exhausted. Here, using biophysical, functional, and multimodal single-cell approaches, we show how V{gamma}4 IELs address this challenge. While germline-encoded TCR regions broadly mediate BTNL3 recognition, productive activation requires additional non-germline TCR features that license responsiveness to BTNL3/8 and enable local selection in the gut. Rather than driving exhaustion, BTNL3/8 reactivity directly promotes expression of an NK-like program marked by adaptor molecules that license innate-like signaling in healthy tissue. These findings reveal how combined innate and adaptive features of the {gamma}{delta}TCR enable durable tissue specialization under conditions of persistent physiological ligand engagement.

immunology↗

bioRxiv2026

Persistent circulating autoreactive PD1⁺TIGIT⁺ peripheral helper T cells reflect synovial lymphoid activity and poor response to conventional disease-modifying anti-rheumatic drugs in early rheumatoid arthritis

ObjectiveIn the first year after onset of the autoimmune disease RA (RA), 40-60% do not achieve remission on conventional synthetic disease-modifying anti-rheumatic drugs (csDMARDs). To understand how autoreactive T cells may contribute to unstable or non-remission, we studied CD4+ T cells, including those recognising citrullinated (Cit) vimentin in participants with RA. MethodsTwo cohorts of drug-naive new-onset participants with RA were treated with csDMARDs. Disease activity score (DAS28-CRP) and peripheral blood (PB) mononuclear cells were collected longitudinally. In HLA-DR-shared epitope+ cohort 1 (n=21), T cells were assessed with a 17-marker spectral flow panel, incorporating HLA-DRB1*04:01/01:01- VimentinCit6459-71 or HLA-DRB1*04:04-VimentinCit7166-78 tetramers. Changes in T cell subsets over time were assessed in remitting and non-remitting participants using a generalized linear mixed model with a negative binomial distribution. In cohort 2 (n=26), the transcriptome of disaggregated synovial tissue (ST) and PB CD4+ T cells were analysed at baseline, and ST biopsy spatial proteomics at baseline and 6 months. ResultsCD4CXCR5-PD1+ peripheral helper T cells (Tph), including Cit-vimentin-reactive Tph, CD4CCR7+CXCR5-PD1+ stem-like Tph and TIGIT+PD1+ Tph were increased with moderate/high DAS28-CRP at any time point. Remission outcome was associated with low CD4 follicular helper T cell (Tfh) and Cit-vimentin-reactive Tfh. In non-remitting participants, Tph/fh infiltrated germinal-centre-like ST aggregates. This decreased in remission. Circulating TIGIT+ Tph genes reflected B lymphoid activation and lymph node egress, while in ST they reflected local differentiation. ConclusionPersistently high circulating TIGIT+ Tph and Tfh, including Cit-vimentin specificities, reflecting antigen-presenting B-cell interactions, are associated with reduced response to csDMARDs in recent-onset RA. Key messagesO_ST_ABSWhat is already known on this topicC_ST_ABSO_LITph and Tfh cells interacting with B cells are implicated in active ACPA+ RA C_LIO_LICitrullinated (Cit)-vimentin is an important neutrophil extracellular trap-derived target of ACPA C_LI What this study addsO_LIHigh circulating TIGIT+ Tph, Cit-vimentin-autoreactive Tph and Tfh associate with failure to reach remission on conventional synthetic DMARDs within the first year in early HLA-DR shared-epitope+ RA C_LIO_LITph/Tfh and adjacent regulatory T cells surround follicular B-cells in lymphoid aggregates in synovial tissue in non-remission C_LIO_LICirculating TIGIT+ Tph bear a transcriptional signature of lymphoid tissue expansion and lymph node egress as compared to functional differentiation and antigen experience in synovial tissue C_LI How this study might affect research, practice or policyO_LIWhen a remission target is not achieved in HLA-DR shared-epitope+ RA patients during the first year of treatment, high circulating TIGIT+ Tph implicate autoreactive T-B-lymphoid expansion in synovial tissue. C_LI

immunology↗

bioRxiv2026

Unconventional presentation of an immunodominant HLA-DR15-restricted nephritogenic proteinase 3 epitope implicated in vasculitis

Proteinase 3 (PR3) anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis is associated with HLA-DR15, implicating HLA-DR15-restricted CD4 T cell autoimmunity in the disease. However, the PR3 epitopes presented by HLA-DR15 and the autoreactive CD4 T cell responses they elicit are undefined. Using humanized hPR3.DR15+ mice, we identified PR3216-231 as an immunodominant human PR3-derived CD4 T cell epitope presented by HLA-DR15. Immunopeptidomic profiling confirmed the natural processing and HLA-DR15-restricted presentation of PR3216-231. Structural analyses of HLA-DR15-PR3216-231 complexes revealed an unconventional mode of HLA class II antigen presentation, in which a 10 amino acid peptide core occupies the 9 amino acid binding groove, causing a central kink in the bound peptide. PR3216-231-specific CD4 T cells from both hPR3.DR15+ mice and patients with PR3-ANCA-associated vasculitis exhibited convergent T cell receptor features, and patients with active vasculitis displayed a clonally expanded, PR3216-231-specific CD4 TCR repertoire. Immunization of hPR3.DR15 mice with PR3216-231 induced cell-mediated glomerulonephritis characterized by increased renal infiltration of CD4 T cells and macrophages together with segmental glomerular necrosis. These findings identify a nephritogenic PR3-derived CD4 T cell epitope presented by HLA-DR15 in ANCA-associated vasculitis and define a new extended core binding pattern, broadening our understanding of HLA class II peptide presentation.

immunology↗

bioRxiv2026

Characterising AlphaFold 3s ability to predict T cellantigen specificity

T cells are a key part of the adaptive immune system. Using their surface-bound T cell antigen receptors (TCRs), these cells scan peptides and other antigens presented to them by major histocompatibility complex molecules (MHCs) on the surface of cells, searching for abnormalities. Although determining the map between TCRs and their target antigens is of vital importance for the design of safe and effective T cell-based vaccines and therapeutics, decoding these interactions is challenging. Experimental methods are not scalable, and sequence-based computational methods have issues generalising to new antigens. The IMMREP25 benchmark of methods for predicting T cell antigen specificity showed that AlphaFold-based methods promise improved generalisation to novel antigens. However, the ability of structure prediction models to predict T cell antigen specificity has not been robustly evaluated previously. In this work, we characterise AlphaFolds ability to predict T cell antigen specificity. We created a pipeline for high-throughput prediction of TCR:peptide-MHC (pMHC) structures using AlphaFold that is > 100 fold faster than the default implementation and used it to benchmark AlphaFold 3 (AF3) and similar models at predicting T cell antigen specificity. We investigated the underlying correlates of AlphaFold-derived binding scores and found that the models predictive power is related to the positioning of TCRs over the pMHC and not chemical interactions. Furthermore, we refine the AlphaFold-derived binding scores by training a machine learning model we call the PAE Aggregator. We then investigate AF3s ability to uncover the clustering rules of TCR repertoires and recapitulate mutational scanning experiments. These analyses show that AlphaFold3 clusters sequence-similar TCRs according to their binding mode and detects disrupting point mutations accurately. Our results highlight both the promise and the current limitations of structure-based approaches for predicting TCR specificity, guiding the development of more reliable immunological prediction methods.

systems biology↗

bioRxiv2026

Conserved T cell receptor usage underpins recognition of CD1c presenting a mycobacterial lipid

The mechanism by which T cell receptors (TCR) recognise mycobacterial lipids presented by CD1 family members not well understood. We used CD1c tetramers loaded with the mycobacterial phosphomycoketide (PM) or mannosyl-PM (MPM) to isolate T cells ex vivo in healthy blood donors and in individuals from a tuberculosis (TB)-endemic region of South Africa, with higher frequencies observed in individuals from the TB-endemic region. High throughput analysis of >200 paired CD1c-mycoketide tetramer+ {beta}TCRs identified a conserved TCR motif, encoded by TRBV4-1 or TRBV7-9 variable region genes in greater than half of TCRs sequenced. A cryo-EM structure of a TRBV7-9+ TCR in complex with CD1c-PM demonstrated that the TCR bound to the F side of CD1c, directly contacting the phospholipid antigen and F-portal residues. Analysis of multiple T cell clones interacting with CD1c mutants suggested that this TCR docking mode is representative of the larger TRBV7-9+ T cell population. Collectively, this study provides insight into mycobacterial lipid-antigen recognition by CD1c-restricted T cells. SummaryCao and Soliman et al study CD1c-restricted human T cells that recognize mycobacterial phosphomycoketide lipids, where they are shown to be more frequent in blood a TB-endemic region. The authors describe a strong TCR-{beta} repertoire bias and use cryoEM to provide a high-resolution structure of an archetypal TCR engaging CD1c in complex with a mycobacterial lipid.

immunology↗

bioRxiv2026

Generative design of antigen-specific T-cell receptor sequences with a conditional diffusion model

T cell receptor (TCR)-based immunotherapy holds immense potential for treating cancers, autoimmunity, and infectious diseases, where antigen-specific TCR recognition is crucial for adaptive immune responses. Engineering or de novo generation of the complementarity-determining region 3 (CDR3) loops of TCRs using artificial intelligence offers a powerful alternative to designing antigen-specific TCRs rather than laborious experimental screening. However, current in silico approaches are constrained by weak conditional guidance, limited flexibility, and a lack of rigorous functional validation. To address these limitations, we introduce TCRDiff, a generative diffusion framework for designing antigen-specific TCRs conditioned on peptide-MHC (pMHC) targets and germline-encoded TCR variable genes. By leveraging pre-trained knowledge from massive T-cell repertoires and TCR-pMHC recognition data, TCRDiff generates CDR3{beta} sequences that closely resemble native-binding TCRs via a denoising diffusion process. Furthermore, incorporating interface geometry features generated TCR-pMHC complexes with superior structural plausibility than models relying solely on sequence-based diffusion or structure-based modeling. As a proof of concept, we deployed TCRDiff in a systematic pipeline to design candidate TCRs against a clinically validated cancer antigen. In vitro activation assays validated that TCRDiff-generated TCRs efficiently recognize the MAGE-A3 epitope with minimal off-target reactivity. Thus, TCRDiff establishes a powerful, validated computational paradigm to accelerate the development of TCR-based immunotherapies.

bioinformatics↗

bioRxiv2026

CD1a-Mediated Presentation of Canonical Microbial Peptides to T Cells

Langerhans cells express the nonpolymorphic antigen-presenting molecule CD1a, positioning them as contributors to host immunity against Mycobacterium leprae in human leprosy. CD1a was originally shown to present non-canonical lipopeptide antigens such as dideoxymycobactin and chemically diverse hydrophobic ligands. Here, we generated CD4 T cell lines from leprosy lesions that recognized M. leprae in a CD1a-restricted manner. Unexpectedly, antigen recognition was protease-sensitive, prompting biochemical purification that identified two microbial protein antigens: LppX, a 25-kDa lipoglycoprotein, and Ag85A, a 30-kDa secreted protein with no known lipid modification. Recombinant proteins activated the corresponding T cell lines in a CD1a-dependent manner. Epitope mapping identified 12-mer peptides that fully reconstituted antigenicity, were conserved between M. leprae and M. tuberculosis, and elicited robust, dose-dependent IFN-{gamma} production and T cell proliferation, establishing that DNA-encoded, ribosomally translated peptides serve as CD1a-restricted cognate antigens. Biochemical analyses showed peptide binding to CD1a, supported by isoelectric focusing and surface plasmon resonance (KD [~]75 M for Ag85A). CD1a-peptide tetramers specifically stained cognate T cells, soluble CD1a was sufficient to present peptide antigen, and transfer of the LppX-specific TCR into naive T cells restored antigen responsiveness. Using CD1a-peptide tetramers, we identified antigen-specific T cells enriched in patients undergoing reversal reactions compared with patients with lepromatous leprosy and healthy donors. The CD1a-restricted T cell lines secreted IFN-{gamma} and IL-26, cytokines with established antimicrobial activity. Together, these findings demonstrate that CD1a can present canonical microbial peptides as part of a cell-mediated immune response in leprosy, extending the known spectrum of CD1a ligands. Because CD1a is nonpolymorphic and presents antigens to antimicrobial T cells, CD1a-peptide complexes may provide a broadly applicable platform for studying, detecting, and potentially targeting mycobacterial immunity.

immunology↗

bioRxiv2026

Specific targeting of MR1-antigen complexes using nanobodies

T cell receptor mimic (TCRm) antibodies and nanobodies that specifically bind peptide-HLA complexes have great therapeutic potential, as they can target polymorphic HLA on tumour cells furnishing peptides derived from tumour-associated antigens. MR1 is an MHC class-I-like molecule that exhibits limited polymorphism that binds and presents conserved metabolites, such as 5-OP-RU, derived from microbial riboflavin biosynthesis. Whether antibodies targeting such MR1-5-OP-RU complexes can be generated remains unclear. Using yeast display technology and in vitro affinity maturation, a nanobody with high affinity and fine specificity toward MR1-5-OP-RU complex was generated. These nanobodies bind both mouse and human MR1-5-OP-RU and inhibited MAIT cell responses to 5-OP-RU in vitro and in vivo demonstrating therapeutic potential. Moreover, we provide a molecular basis underpinning the fine specificity of these nanobodies, solving the crystal structures of MR1 in complex with either 5-OP-RU or Ac-6-FP. Here, the nanobody co-bound MR1 and 5-OP-RU, akin to a TCRm antibody. Moreover, we engineer bispecific antibodies targeting both MR1-5-OP-RU and CD3, that drive broad T cell killing of bacterially-infected cells as well as tumour cells treated with 5-OP-RU, thereby providing proof-of-principle for targeting the MR1 molecule with with TCRm-based nanobodies. One Sentence SummaryWe report the development of a nanobody targeting MR1-5-OP-RU complex and demonstrate its utility to modulate MAIT cells responses, and as a bispecific engager.

immunology↗

bioRxiv2025

Sideways lipid presentation by the antigen-presenting molecule CD1c

Here we report mass spectrometry analyses of endogenous lipids captured by CD1c when bound to an autoreactive {beta}TCR. CD1c bound twenty-six lipids with bulky headgroups that could not fit within the tight TCR-CD1c interface. We determined the crystal structures of CD1c presenting several gangliosides, revealing a general mechanism whereby two lipids, rather than one, are bound in the CD1c cleft. Bulky lipids were orientated sideways so that their polar headgroups protruded laterally through a side portal of the CD1c molecule - an evolutionarily conserved structural feature. The sideways presented ganglioside headgroups did not hinder TCR binding and so represent a mechanism that allows autoreactive TCR recognition of CD1c. In addition, ex vivo studies showed sideways presented gangliosides could also represent TCR recognition determinants. These findings reveal a general mechanism whereby CD1c simultaneously presents two lipid antigens from the top and side of its cleft that differs markedly from other antigen-presenting molecules.

immunology↗

bioRxiv2025

The antigen presenting molecule MR1 binds riboflavin catabolites

Major histocompatibility-complex (MHC) class I-related (MR1) protein presents vitamin B based antigens to Mucosal-Associated Invariant T (MAIT) cells. While microbial riboflavin precursors are well documented MR1 ligands, it is unclear whether host-generated riboflavin catabolites influence MR1-mediated immunity. Here, we report that riboflavin catabolites, including 10-formylmethylflavin (FMF), lumichrome, lumiflavin and alloxazine bind to MR1 with moderate affinity, while riboflavin itself binds weakly. In contrast to the microbial riboflavin antigens which increase MR1 cell surface expression, the riboflavin catabolites moderately reduced cell surface levels of MR1 by stabilizing and retaining MR1 in the endoplasmic reticulum (ER). The riboflavin catabolites appeared to bind to the intracellular MR1 and inhibit MR1 exit from the ER. These riboflavin catabolites also weakly competed with Vit B based Ags for MR1 binding, thereby inhibiting MAIT cell activation. The crystal structures of MR1 complexed with riboflavin, FMF, lumichrome and lumiflavin, show binding of these three-ringed ligands in the A-pocket of MR1. The crystal structure of MR1-lumichrome revealed that lumichrome formed a covalent "flavin bond" with MR1-Lys43 differing from the typical Schiff-base bond of MR1-Lys43-Ag complexes. Collectively, we identified three ring isoalloxazines that can bind MR1 and downregulate cell surface expression levels, suggesting a potential role in dampening MAIT cell immunity.

biochemistry↗

bioRxiv2025

A transgenic mouse allows to capture the HLA-C*06:02 immunopeptidome in a model of psoriasis

Psoriasis vulgaris is a T cell-mediated autoimmune skin condition affecting around one in fifty people worldwide. Whilst advanced immunomodulatory therapeutic options have become available in recent years, ongoing disease suppression is still required with no curative treatment available to date. The human leukocyte antigen class I allele HLA-C*06:02 is the main genetic risk determinant of psoriasis. Its function is to present peptide antigens to CD8+ T cells and natural killer cells which in turn elicit and perpetuate the immune response, yet little is known about ligands presented by HLA-C*06:02. To gain an understanding which HLA-C*06:02-restricted peptides are presented by epidermal cell populations and might be initiators of the autoimmune response in psoriasis, we have conducted an in depth immunopeptidomics analysis of HLA-C*06:02+ keratinocyte and melanocyte cell lines. Furthermore, we introduce a HLA-C*06:02 transgenic mouse which, in conjunction with the imiquimod model of psoriasis, allowed us to assess the ex vivo immunopeptidome of HLA-C*06:02 in psoriasiform skin. Overall, we detected 20,812 high confidence HLA-C*06:02 bound peptide ligands. Thus, we present a comprehensive coverage HLA-C in vitro immunopeptidomics dataset and the first HLA-C*06:02 ex vivo dataset of psoriasis-relevant peptide antigens that may inform the development of antigen-specific, novel curative therapeutic approaches in psoriasis.

immunology↗

bioRxiv2025

Human Neonatal MR1T Cells Have Diverse TCR Usage, are Less Cytotoxic and are Unable to Respond to Many Common Childhood Pathogens.

Neonatal sepsis is a leading cause of childhood mortality. Understanding immune cell development can inform strategies to combat this. MR1-restricted T (MR1T) cells can be defined by their recognition of small molecules derived from microbes, self, and drug and drug-like molecules, presented by the MHC class 1-related molecule (MR1). In healthy adults, the majority of MR1T cells express an invariant -chain; TRAV1-2/TRAJ33/12/20 and are referred to as mucosal-associated invariant T (MAIT) cells. Neonatal MR1T cells isolated from cord blood (CB) demonstrate more diversity in MR1T TCR usage, with the majority of MR1-5-OP-RU-tetramer(+) cells being TRAV1-2(-). To better understand this diversity, we performed single-cell-RNA-seq/TCR-seq (scRNA-seq/scTCR-seq) on MR1-5-OP-RU-tetramer(+) cells from CB (n=5) and adult participants (n=5). CB-derived MR1T cells demonstrate a less cytotoxic/pro-inflammatory phenotype, and a more diverse TCR repertoire. A panel of CB and adult MAIT and TRAV1-2(-) MR1T cell clones were generated, and CB-derived clones were unable to recognize several common riboflavin-producing childhood pathogens (S. aureus, S. pneumoniae, M. tuberculosis). Biochemical and structural investigation of one CB MAIT TCR (CB964 A2; TRAV1-2/TRBV6-2) showed a reduction in binding affinity toward the canonical MR1-antigen, 5-OP-RU, compared to adult MAIT TCRs that correlated with differences in {beta}-chain contribution in the TCR-MR1 interface. Overall, this data shows that CB MAIT and TRAV1-2(-) MR1T cells, express a diverse TCR repertoire, a more restricted childhood pathogen recognition profile and diminished cytotoxic and pro-inflammatory capacity. Understanding this diversity, along with the functional ability of TRAV1-2(-) MR1T cells, could provide insight into increased neonatal susceptibility to infections.

immunology↗

bioRxiv2025

Self-iterative multiple instance learning enables the prediction of CD4+ T cell immunogenic epitopes

Accurately predicting the antigen presentation to CD4+ T cells and subsequent induction of immune response is fundamentally important for vaccine development, autoimmune disease treatments, and cancer neoepitope identification. In immunopeptidomics, single-allelic data are highly specific but limited in allele scope, while multi-allelic data contain broader coverage at the cost of weakly labeling. Existing computational approaches either overlook the massive multi-allelic data or introduce label ambiguity due to inadequate modeling strategies. Here, we introduce ImmuScope, a weakly supervised deep-learning framework integrating precise MHC-II antigen presentation, CD4+ T cell epitopes, and immunogenicity predictions. ImmuScope leverages self-iterative multiple-instance learning with positive-anchor triplet loss to explore peptide-MHC-II (pMHC-II) binding from weakly labeled multi-allelic data and single-allelic data, comprising over 600,000 ligands across 142 alleles. Moreover, ImmuScope can also interpret the MHC-II binding specificity and motif deconvolution of immunopeptidomics data. We successfully applied ImmuScope to discover melanoma neoantigens, revealing variations in pMHC-II binding and immunogenicity upon epitope mutations. We further employed ImmuScope to assess the effects of SARS-CoV-2 epitope mutations on immune escape, with its predictions aligned well with experimentally determined immune escape dynamics. Overall, ImmuScope provides a comprehensive solution for CD4+ T cell antigen recognition and immunogenicity assessment, with broad potential for advancing vaccine design and personalized immunotherapy.

bioinformatics↗

bioRxiv2025

Fast and accurate modeling of TCR-peptide-MHC complexes using tFold-TCR

Alpha-beta T cell receptor ({beta}TCR) recognition of peptide-major histocompatibility complexes (pMHCs) is a corner- stone of the adaptive immune system. Fast and accurate modeling of TCR-pMHC structures is crucial for understanding TCR recognition of pMHCs at the molecular level, which is essential for the development of TCR-based therapeutics and vaccines. Despite significant interest, this challenge remains unresolved due to the diversity of TCR-pMHC interactions and limited structural data. Here, we present tFold-TCR, a high-throughput, end-to-end universal model for predicting three-dimensional (3D) atomic-level structures of TCR-pMHC complexes, capable of predicting TCRs of different classes and MHC structures from diverse systems. tFold-TCR leverages a specially trained, protein-protein interaction-sensitive large protein language model to extract intra- and inter-chain residue contact information and evolutionary relationships, bypassing the need for multiple sequence alignment (MSA) searches. It also features innovative structure prediction and flexible docking modules to enhance accuracy, particularly for interacting contacts. Compared to existing methods, including AlphaFold-3, tFold-TCR demonstrates a 30.7% increase in prediction success rate evaluated by DockQ and is over 25 times faster. These advancements enable large-scale structural characterization of TCRs and their interactions with pMHCs. Utilizing this capability, we constructed TCRStructDB, the largest database of TCR-pMHC structures to date, encompassing 2.2 million TCRs, 0.8 million pMHCs, and 45,000 TCR-pMHC complexes. TCRStructDB provides unprecedented insights into one of the most diverse receptor-ligand interactions in biology.

bioinformatics↗

bioRxiv2024

Quantifying conformational changes in the TCR:pMHC-I binding interface

T cells form one of the key pillars of adaptive immunity. Using their surface bound T cell antigen receptors (TCRs), these cells screen millions of antigens presented by major histocompatability complex (MHC) or MHC-like molecules. In other protein families, the dynamics of protein-protein interactions have important implications for protein function. Case studies of TCR:class I peptide-MHCs (pMHC-Is) structures have reported mixed results on whether the binding interfaces undergo conformational change during engagement and no robust statistical quantification has been done to generalise these results. It thus remains an open question if movement occurs in the binding interface that enables recognition and activation of T cells. In this work, we quantify the conformational changes in the TCR:pMHC-I binding interface by creating a dataset of 358 structures, comprising 25 TCRs, 20 MHC alleles, and 58 peptide structures in both unbound (apo) and bound (holo) conformations. In support of some case studies, we demonstrate that all complimentary determining region (CDR) loops move to a certain extent but only CDR3 and CDR3{beta} loops modify their shape when binding pMHC-Is. We also map out the contacts between TCRs and pMHC-Is, generating a novel fingerprint of TCRs on MHC molecules and show that the CDR3 tends to bind the N-terminus of the peptide and the CDR3{beta} tends to bind the C-terminus of the peptide. Finally, we show that the presented peptides can undergo conformational changes when engaged by TCRs, as has been reported in past literature, but novelly show these changes depend on how the peptides are anchored in the MHC binding groove. Our work has implications in understanding the behaviour of TCR:pMHC-I interactions and providing insights that can be used for modelling T cell antigen specificity, an ongoing grand challenge in immunology.

immunology↗

bioRxiv2024

Prediction of KIR3DL1/Human Leukocyte Antigen binding

KIR3DL1 is a polymorphic inhibitory Natural Killer (NK) cell receptor that recognizes Human Leukocyte Antigen (HLA) class I allotypes that contain the Bw4 motif. Structural analyses have shown that in addition to residues 77-83 that span the Bw4 motif, polymorphism at other sites throughout the HLA molecule can influence the interaction with KIR3DL1. Given the extensive polymorphism of both KIR3DL1 and HLA class I, we built a machine learning prediction model to describe the influence of allotypic variation on the binding of KIR3DL1 to HLA class I. Nine KIR3DL1 tetramers were screened for reactivity against a panel of HLA class I molecules which revealed different patterns of specificity for each KIR3DL1 allotype. Separate models were trained for each of KIR3DL1 allotypes based on the full amino sequence of exons 2 and 3 encoding the 1 and 2 domains of the class I HLA allotypes, the set of polymorphic positions that span the Bw4 motif, or the positions that encode 1 and 2 but exclude the connecting loops. The Multi-Label-Vector-Optimization (MLVO) model trained on all alpha helix positions performed best with AUC scores ranging from 0.74 to 0.974 for the 9 KIR3DL1 allotype models. We show that a binary division into binder and non-binder is not precise, and that intermediate levels exist. Using the same models, within the binder group, high- and low-binder categories can also be predicted, the regions in HLA affecting the high vs low binder being completely distinct from the classical Bw4 motif. We further show that these positions affect binding affinity in a nonadditive way and induce deviations from linear models used to predict interaction strength. We propose that this approach should be used in lieu of simpler binding models based on a single HLA motif.

immunology↗

bioRxiv2024

Epitope-anchored contrastive transfer learning for paired CD8+ T cell receptor-antigen recognition

Understanding the mechanisms of T-cell antigen recognition that underpin adaptive immune responses is critical for the development of vaccines, immunotherapies, and treatments against autoimmune diseases. Despite extensive research efforts, the accurate identification of T cell receptor (TCR)-antigen binding pairs remains a significant challenge due to the vast diversity and cross-reactivity of TCRs. Here, we propose a deep-learning framework termed Epitope-anchored Contrastive Transfer Learning (EPACT) tailored to paired human CD8+ TCRs from single-cell sequencing data. Harnessing the pre-trained representations and the contrastive co-embedding space, EPACT demonstrates state-of-the-art model generalizability in predicting TCR binding specificity for unseen epitopes and distinct TCR repertoires, offering potential values for practical outcomes in real-world scenarios. The contrastive learning paradigm achieves highly precise predictions for immunodominant epitopes and facilitates interpretable analysis of epitope-specific T cells. The TCR binding strength predicted by EPACT aligns well with the surge in spike-specific immune responses targeting SARS-CoV-2 epitopes after vaccination. We further fine-tune EPACT on TCR-epitope structural data to decipher the residue-level interactions involved in T-cell antigen recognition. EPACT not only exhibits superior capabilities in quantifying inter-chain distance matrices and identifying contact residue pairs but also corroborates the presence of molecular mimicry across multiple tumor-associated antigens. Together, EPACT can serve as a useful AI approach with significant potential in practical applications and contribute toward the development of TCR-based diagnostics and immunotherapies.

bioinformatics↗

bioRxiv2023

ConBoTNet: supervised contrastive learning enhances MHC-II peptide binding affinity prediction

Accurate prediction of major histocompatibility complex (MHC)-peptide binding affinity can improve our understanding of cellular immune responses and guide personalized immunotherapies. Nevertheless, the existing deep learning-based approaches for predicting MHC-II peptide interactions fall short of satisfactory performance and offer restricted model interpretability. In this study, we propose a novel deep neural network, termed ConBoTNet, to address the above issues by introducing the designed supervised contrastive learning and bottleneck transformer extractors. Specifically, the supervised contrastive learning pre-training enhances the models representative and generalizable capabilities on MHC-II peptides by pulling positive pairs closer and pushing negative pairs further in the feature space, while the bottleneck transformer module focuses on MHC-II peptide interactions to precisely identify binding cores and anchor positions in an unsupervised manner. Extensive experiments on benchmark datasets under 5-fold cross-validation, leave-one-molecule-out validation, independent testing, and binding core prediction settings highlighted the superiority of our proposed ConBoTNet over current state-of-the-art methods. Data distribution analysis in the latent feature space demonstrated that supervised contrastive learning can aggregate MHC-II-peptide samples with similar affinity labels and learn common features of similar affinity. Additionally, we interpreted the trained neural network by associating the attention weights with peptides and innovatively find both well-established and potential peptide motifs. This work not only introduces an innovative tool for accurately predicting MHC-II peptide affinity, but also provides new insights into a new paradigm for modeling essential biological interactions, advancing data-driven discovery in biomedicine.

bioinformatics↗