Human CD8+ T cell cross-reactivity across influenza A, B and C viruses.

Influenza A, B and C viruses (IAV, IBV and ICV, respectively) circulate globally and infect humans, with IAV and IBV causing the most severe disease. CD8+ T cells confer cross-protection against IAV strains, however the responses of CD8+ T cells to IBV and ICV are understudied. We investigated the breadth of CD8+ T cell cross-recognition and provide evidence of CD8+ T cell cross-reactivity across IAV, IBV and ICV. We identified immunodominant CD8+ T cell epitopes from IBVs that were protective in mice and found memory CD8+ T cells directed against universal and influenza-virus-type-specific epitopes in the blood and lungs of healthy humans. Lung-derived CD8+ T cells displayed tissue-resident memory phenotypes. Notably, CD38+Ki67+CD8+ effector T cells directed against novel epitopes were readily detected in IAV- or IBV-infected pediatric and adult subjects. Our study introduces a new paradigm whereby CD8+ T cells confer unprecedented cross-reactivity across all influenza viruses, a key finding for the design of universal vaccines.

T cell autoreactivity directed toward CD1c itself rather than toward carried self lipids.

The hallmark function of αß T cell antigen receptors (TCRs) involves the highly specific co-recognition of a major histocompatibility complex molecule and its carried peptide. However, the molecular basis of the interactions of TCRs with the lipid antigen-presenting molecule CD1c is unknown. We identified frequent staining of human T cells with CD1c tetramers across numerous subjects. Whereas TCRs typically show high specificity for antigen, both tetramer binding and autoreactivity occurred with CD1c in complex with numerous, chemically diverse self lipids. Such extreme polyspecificity was attributable to binding of the TCR over the closed surface of CD1c, with the TCR covering the portal where lipids normally protrude. The TCR essentially failed to contact lipids because they were fully seated within CD1c. These data demonstrate the sequestration of lipids within CD1c as a mechanism of autoreactivity and point to small lipid size as a determinant of autoreactive T cell responses.

Drugs and drug-like molecules can modulate the function of mucosal-associated invariant T cells.

The major-histocompatibility-complex-(MHC)-class-I-related molecule MR1 can present activating and non-activating vitamin-B-based ligands to mucosal-associated invariant T cells (MAIT cells). Whether MR1 binds other ligands is unknown. Here we identified a range of small organic molecules, drugs, drug metabolites and drug-like molecules, including salicylates and diclofenac, as MR1-binding ligands. Some of these ligands inhibited MAIT cells ex vivo and in vivo, while others, including diclofenac metabolites, were agonists. Crystal structures of a T cell antigen receptor (TCR) from a MAIT cell in complex with MR1 bound to the non-stimulatory and stimulatory compounds showed distinct ligand orientations and contacts within MR1, which highlighted the versatility of the MR1 binding pocket. The findings demonstrated that MR1 was able to capture chemically diverse structures, spanning mono- and bicyclic compounds, that either inhibited or activated MAIT cells. This indicated that drugs and drug-like molecules can modulate MAIT cell function in mammals.

CD1a on Langerhans cells controls inflammatory skin disease.

CD1a is a lipid-presenting molecule that is abundantly expressed on Langerhans cells. However, the in vivo role of CD1a has remained unclear, principally because CD1a is lacking in mice. Through the use of mice with transgenic expression of CD1a, we found that the plant-derived lipid urushiol triggered CD1a-dependent skin inflammation driven by CD4(+) helper T cells that produced the cytokines IL-17 and IL-22 (TH17 cells). Human subjects with poison-ivy dermatitis had a similar cytokine signature following CD1a-mediated recognition of urushiol. Among various urushiol congeners, we identified diunsaturated pentadecylcatechol (C15:2) as the dominant antigen for CD1a-restricted T cells. We determined the crystal structure of the CD1a-urushiol (C15:2) complex, demonstrating the molecular basis of urushiol interaction with the antigen-binding cleft of CD1a. In a mouse model and in patients with psoriasis, CD1a amplified inflammatory responses that were mediated by TH17 cells that reacted to self lipid antigens. Treatment with blocking antibodies to CD1a alleviated skin inflammation. Thus, we propose CD1a as a potential therapeutic target in inflammatory skin diseases.

T cell receptor reversed polarity recognition of a self-antigen major histocompatibility complex.

Central to adaptive immunity is the interaction between the αß T cell receptor (TCR) and peptide presented by the major histocompatibility complex (MHC) molecule. Presumably reflecting TCR-MHC bias and T cell signaling constraints, the TCR universally adopts a canonical polarity atop the MHC. We report the structures of two TCRs, derived from human induced T regulatory (iT(reg)) cells, complexed to an MHC class II molecule presenting a proinsulin-derived peptide. The ternary complexes revealed a 180° polarity reversal compared to all other TCR-peptide-MHC complex structures. Namely, the iT(reg) TCR α-chain and ß-chain are overlaid with the α-chain and ß-chain of MHC class II, respectively. Nevertheless, this TCR interaction elicited a peptide-reactive, MHC-restricted T cell signal. Thus TCRs are not 'hardwired' to interact with MHC molecules in a stereotypic manner to elicit a T cell signal, a finding that fundamentally challenges our understanding of TCR recognition.

MHCpLogics: an interactive machine learning-based tool for unsupervised data visualization and cluster analysis of immunopeptidomes.

The major histocompatibility complex (MHC) encodes a range of immune response genes, including the human leukocyte antigens (HLAs) in humans. These molecules bind peptide antigens and present them on the cell surface for T cell recognition. The repertoires of peptides presented by HLA molecules are termed immunopeptidomes. The highly polymorphic nature of the genres that encode the HLA molecules confers allotype-specific differences in the sequences of bound ligands. Allotype-specific ligand preferences are often defined by peptide-binding motifs. Individuals express up to six classical class I HLA allotypes, which likely present peptides displaying different binding motifs. Such complex datasets make the deconvolution of immunopeptidomic data into allotype-specific contributions and further dissection of binding-specificities challenging. Herein, we developed MHCpLogics as an interactive machine learning-based tool for mining peptide-binding sequence motifs and visualization of immunopeptidome data across complex datasets. We showcase the functionalities of MHCpLogics by analyzing both in-house and published mono- and multi-allelic immunopeptidomics data. The visualization modalities of MHCpLogics allow users to inspect clustered sequences down to individual peptide components and to examine broader sequence patterns within multiple immunopeptidome datasets. MHCpLogics can deconvolute large immunopeptidome datasets enabling the interrogation of clusters for the segregation of allotype-specific peptide sequence motifs, identification of sub-peptidome motifs, and the exportation of clustered peptide sequence lists. The tool facilitates rapid inspection of immunopeptidomes as a resource for the immunology and vaccine communities. MHCpLogics is a standalone application available via an executable installation at: https://github.com/PurcellLab/MHCpLogics.

Benchmarking Bioinformatics Pipelines in Data-Independent Acquisition Mass Spectrometry for Immunopeptidomics.

Immunopeptidomes are the peptide repertoires bound by the molecules encoded by the major histocompatibility complex [human leukocyte antigen (HLA) in humans]. These HLA-peptide complexes are presented on the cell surface for immune T-cell recognition. Immunopeptidomics denotes the utilization of tandem mass spectrometry to identify and quantify peptides bound to HLA molecules. Data-independent acquisition (DIA) has emerged as a powerful strategy for quantitative proteomics and deep proteome-wide identification; however, DIA application to immunopeptidomics analyses has so far seen limited use. Further, of the many DIA data processing tools currently available, there is no consensus in the immunopeptidomics community on the most appropriate pipeline(s) for in-depth and accurate HLA peptide identification. Herein, we benchmarked four commonly used spectral library-based DIA pipelines developed for proteomics applications (Skyline, Spectronaut, DIA-NN, and PEAKS) for their ability to perform immunopeptidome quantification. We validated and assessed the capability of each tool to identify and quantify HLA-bound peptides. Generally, DIA-NN and PEAKS provided higher immunopeptidome coverage with more reproducible results. Skyline and Spectronaut conferred more accurate peptide identification with lower experimental false-positive rates. All tools demonstrated reasonable correlations in quantifying precursors of HLA-bound peptides. Our benchmarking study suggests a combined strategy of applying at least two complementary DIA software tools to achieve the greatest degree of confidence and in-depth coverage of immunopeptidome data.

Complimentary electrostatics dominate T-cell receptor binding to a psoriasis-associated peptide antigen presented by human leukocyte antigen C∗06:02.

Psoriasis is a chronic skin disease characterized by hyperproliferative epidermal lesions infiltrated by autoreactive T cells. Individuals expressing the human leukocyte antigen (HLA) C∗06:02 allele are at highest risk for developing psoriasis. An autoreactive T cell clone (termed Vα3S1/Vß13S1) isolated from psoriatic plaques is selective for HLA-C∗06:02, presenting a peptide derived from the melanocyte-specific autoantigen ADAMTSL5 (VRSRRCLRL). Here we determine the crystal structure of this psoriatic TCR-HLA-C∗06:02 ADAMTSL5 complex with a stabilized peptide. Docking of the TCR involves an extensive complementary charge network formed between negatively charged TCR residues interleaving with exposed arginine residues from the self-peptide and the HLA-C∗06:02 α1 helix. We probed these interactions through mutagenesis and activation assays. The charged interface spans the polymorphic region of the C1/C2 HLA group. Notably the peptide-binding groove of HLA-C∗06:02 appears exquisitely suited for presenting highly charged Arg-rich epitopes recognized by this acidic psoriatic TCR. Overall, we provide a structural basis for understanding the engagement of melanocyte antigen-presenting cells by a TCR implicated in psoriasis while simultaneously expanding our knowledge of how TCRs engage HLA-C.

PCprophet: a framework for protein complex prediction and differential analysis using proteomic data.

Despite the availability of methods for analyzing protein complexes, systematic analysis of complexes under multiple conditions remains challenging. Approaches based on biochemical fractionation of intact, native complexes and correlation of protein profiles have shown promise. However, most approaches for interpreting cofractionation datasets to yield complex composition and rearrangements between samples depend considerably on protein-protein interaction inference. We introduce PCprophet, a toolkit built on size exclusion chromatography-sequential window acquisition of all theoretical mass spectrometry (SEC-SWATH-MS) data to predict protein complexes and characterize their changes across experimental conditions. We demonstrate improved performance of PCprophet over state-of-the-art approaches and introduce a Bayesian approach to analyze altered protein-protein interactions across conditions. We provide both command-line and graphical interfaces to support the application of PCprophet to any cofractionation MS dataset, independent of separation or quantitative liquid chromatography-MS workflow, for the detection and quantitative tracking of protein complexes and their physiological dynamics.

PCfun: a hybrid computational framework for systematic characterization of protein complex function.

In molecular biology, it is a general assumption that the ensemble of expressed molecules, their activities and interactions determine biological function, cellular states and phenotypes. Stable protein complexes-or macromolecular machines-are, in turn, the key functional entities mediating and modulating most biological processes. Although identifying protein complexes and their subunit composition can now be done inexpensively and at scale, determining their function remains challenging and labor intensive. This study describes Protein Complex Function predictor (PCfun), the first computational framework for the systematic annotation of protein complex functions using Gene Ontology (GO) terms. PCfun is built upon a word embedding using natural language processing techniques based on 1 million open access PubMed Central articles. Specifically, PCfun leverages two approaches for accurately identifying protein complex function, including: (i) an unsupervised approach that obtains the nearest neighbor (NN) GO term word vectors for a protein complex query vector and (ii) a supervised approach using Random Forest (RF) models trained specifically for recovering the GO terms of protein complex queries described in the CORUM protein complex database. PCfun consolidates both approaches by performing a hypergeometric statistical test to enrich the top NN GO terms within the child terms of the GO terms predicted by the RF models. The documentation and implementation of the PCfun package are available at https://github.com/sharmavaruns/PCfun. We anticipate that PCfun will serve as a useful tool and novel paradigm for the large-scale characterization of protein complex function.

PFresGO: an attention mechanism-based deep-learning approach for protein annotation by integrating gene ontology inter-relationships.

MOTIVATION: The rapid accumulation of high-throughput sequence data demands the development of effective and efficient data-driven computational methods to functionally annotate proteins. However, most current approaches used for functional annotation simply focus on the use of protein-level information but ignore inter-relationships among annotations. RESULTS: Here, we established PFresGO, an attention-based deep-learning approach that incorporates hierarchical structures in Gene Ontology (GO) graphs and advances in natural language processing algorithms for the functional annotation of proteins. PFresGO employs a self-attention operation to capture the inter-relationships of GO terms, updates its embedding accordingly and uses a cross-attention operation to project protein representations and GO embedding into a common latent space to identify global protein sequence patterns and local functional residues. We demonstrate that PFresGO consistently achieves superior performance across GO categories when compared with 'state-of-the-art' methods. Importantly, we show that PFresGO can identify functionally important residues in protein sequences by assessing the distribution of attention weightings. PFresGO should serve as an effective tool for the accurate functional annotation of proteins and functional domains within proteins. AVAILABILITY AND IMPLEMENTATION: PFresGO is available for academic purposes at https://github.com/BioColLab/PFresGO. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

HLA-A*11:01-restricted CD8+ T cell immunity against influenza A and influenza B viruses in Indigenous and non-Indigenous people.

HLA-A*11:01 is one of the most prevalent human leukocyte antigens (HLAs), especially in East Asian and Oceanian populations. It is also highly expressed in Indigenous people who are at high risk of severe influenza disease. As CD8+ T cells can provide broadly cross-reactive immunity to distinct influenza strains and subtypes, including influenza A, B and C viruses, understanding CD8+ T cell immunity to influenza viruses across prominent HLA types is needed to rationally design a universal influenza vaccine and generate protective immunity especially for high-risk populations. As only a handful of HLA-A*11:01-restricted CD8+ T cell epitopes have been described for influenza A viruses (IAVs) and epitopes for influenza B viruses (IBVs) were still unknown, we embarked on an epitope discovery study to define a CD8+ T cell landscape for HLA-A*11:01-expressing Indigenous and non-Indigenous Australian people. Using mass-spectrometry, we identified IAV- and IBV-derived peptides presented by HLA-A*11:01 during infection. 79 IAV and 57 IBV peptides were subsequently screened for immunogenicity in vitro with peripheral blood mononuclear cells from HLA-A*11:01-expressing Indigenous and non-Indigenous Australian donors. CD8+ T cell immunogenicity screening revealed two immunogenic IAV epitopes (A11/PB2320-331 and A11/PB2323-331) and the first HLA-A*11:01-restricted IBV epitopes (A11/M41-49, A11/NS1186-195 and A11/NP511-520). The immunogenic IAV- and IBV-derived peptides were >90% conserved among their respective influenza viruses. Identification of novel immunogenic HLA-A*11:01-restricted CD8+ T cell epitopes has implications for understanding how CD8+ T cell immunity is generated towards IAVs and IBVs. These findings can inform the development of rationally designed, broadly cross-reactive influenza vaccines to ensure protection from severe influenza disease in HLA-A*11:01-expressing individuals.

MhcVizPipe: A Quality Control Software for Rapid Assessment of Small- to Large-Scale Immunopeptidome Datasets.

MS-based immunopeptidomics is maturing into an automatized and high-throughput technology, producing small- to large-scale datasets of clinically relevant major histocompatibility complex (MHC) class I-associated and class II-associated peptides. Consequently, the development of quality control (QC) and quality assurance systems capable of detecting sample and/or measurement issues is important for instrument operators and scientists in charge of downstream data interpretation. Here, we created MhcVizPipe (MVP), a semiautomated QC software tool that enables rapid and simultaneous assessment of multiple MHC class I and II immunopeptidomic datasets generated by MS, including datasets generated from large sample cohorts. In essence, MVP provides a rapid and consolidated view of sample quality, composition, and MHC specificity to greatly accelerate the "pass-fail" QC decision-making process toward data interpretation. MVP parallelizes the use of well-established immunopeptidomic algorithms (NetMHCpan, NetMHCIIpan, and GibbsCluster) and rapidly generates organized and easy-to-understand reports in HTML format. The reports are fully portable and can be viewed on any computer with a modern web browser. MVP is intuitive to use and will find utility in any specialized immunopeptidomic laboratory and proteomics core facility that provides immunopeptidomic services to the community.

Quantitative affinity measurement of small molecule ligand binding to major histocompatibility complex class-I-related protein 1 MR1.

The Major Histocompatibility Complex class I-related protein 1 (MR1) presents small molecule metabolites, drugs, and drug-like molecules that are recognized by MR1-reactive T cells. While we have an understanding of how antigens bind to MR1 and upregulate MR1 cell surface expression, a quantitative, cell-free, assessment of MR1 ligand-binding affinity was lacking. Here, we developed a fluorescence polarization-based assay in which fluorescent MR1 ligand was loaded into MR1 protein in vitro and competitively displaced by candidate ligands over a range of concentrations. Using this assay, ligand affinity for MR1 could be differentiated as strong (IC50 < 1 µM), moderate (1 µM < IC50 < 100 µM), and weak (IC50 > 100 µM). We demonstrated a clear correlation between ligand-binding affinity for MR1, the presence of a covalent bond between MR1 and ligand, and the number of salt bridge and hydrogen bonds formed between MR1 and ligand. Using this newly developed fluorescence polarization-based assay to screen for candidate ligands, we identified the dietary molecules vanillin and ethylvanillin as weak bona fide MR1 ligands. Both upregulated MR1 on the surface of C1R.MR1 cells and the crystal structure of a MAIT cell T cell receptor-MR1-ethylvanillin complex revealed that ethylvanillin formed a Schiff base with K43 of MR1 and was buried within the A'-pocket. Collectively, we developed and validated a method to quantitate the binding affinities of ligands for MR1 that will enable an efficient and rapid screening of candidate MR1 ligands.

Need to know which MHCs protect against COVID? There's an App for that!

In this article, we discuss the recent observation by Augusto et al. made using a novel mobile phone application-based COVID-19 Citizen Science Study that an HLA genetic variant, HLA-B*15:01, is associated with asymptomatic SARS-CoV-2 infection. To explain this association, Augusto et al. describe a cross-reactive memory CD8+ T-cell response in HLA-B*15:01+ SARS-CoV-2 unexposed individuals that retains high avidity for two structurally conserved epitopes found in SARS-CoV-2 and seasonal coronavirus strains. These observations provide an insight into potential molecular determinants that facilitate rapid, early clearance of virus.

A tactile approach to introduce the skin autoimmune disease psoriasis to the general public and the vision-impaired community.

Scientific outreach activities play an important role in disseminating knowledge, connecting the general public to research and breaking down scientific skepticism barriers. However, the vision-impaired community is often disadvantaged when the most common audio-visual approach of scientific communication is applied. Here we integrated tactile clues in the scientific communication of immune processes involved in the autoimmune skin disease psoriasis. We fostered the involvement of the vision-impaired community through interactive experiences, including tactile scientific origami art, a haptic poster and wood-carved molecular models. Readily accessible science communication that engages a number of senses is a critical step toward making science more inclusive and engaging for individuals with a wide range of sensory abilities. The approach of the 2023 Monash Sensory Science exhibition aligns with the principles of equity, diversity and inclusion and helps to empower a more informed and scientifically literate public.

Anthem: a user customised tool for fast and accurate prediction of binding between peptides and HLA class I molecules.

Neopeptide-based immunotherapy has been recognised as a promising approach for the treatment of cancers. For neopeptides to be recognised by CD8+ T cells and induce an immune response, their binding to human leukocyte antigen class I (HLA-I) molecules is a necessary first step. Most epitope prediction tools thus rely on the prediction of such binding. With the use of mass spectrometry, the scale of naturally presented HLA ligands that could be used to develop such predictors has been expanded. However, there are rarely efforts that focus on the integration of these experimental data with computational algorithms to efficiently develop up-to-date predictors. Here, we present Anthem for accurate HLA-I binding prediction. In particular, we have developed a user-friendly framework to support the development of customisable HLA-I binding prediction models to meet challenges associated with the rapidly increasing availability of large amounts of immunopeptidomic data. Our extensive evaluation, using both independent and experimental datasets shows that Anthem achieves an overall similar or higher area under curve value compared with other contemporary tools. It is anticipated that Anthem will provide a unique opportunity for the non-expert user to analyse and interpret their own in-house or publicly deposited datasets.

The differentiation state of the Schwann cell progenitor drives phenotypic variation between two contagious cancers.

Contagious cancers are a rare pathogenic phenomenon in which cancer cells gain the ability to spread between genetically distinct hosts. Nine examples have been identified across marine bivalves, dogs and Tasmanian devils, but the Tasmanian devil is the only mammalian species known to have given rise to two distinct lineages of contagious cancer, termed Devil Facial Tumour 1 (DFT1) and 2 (DFT2). Remarkably, DFT1 and DFT2 arose independently from the same cell type, a Schwann cell, and while their ultra-structural features are highly similar they exhibit variation in their mutational signatures and infection dynamics. As such, DFT1 and DFT2 provide a unique framework for investigating how a common progenitor cell can give rise to distinct contagious cancers. Using a proteomics approach, we show that DFT1 and DFT2 are derived from Schwann cells in different differentiation states, with DFT2 carrying a molecular signature of a less well differentiated Schwann cell. Under inflammatory signals DFT1 and DFT2 have different gene expression profiles, most notably involving Schwann cell markers of differentiation, reflecting the influence of their distinct origins. Further, DFT2 cells express immune cell markers typically expressed during nerve repair, consistent with an ability to manipulate their extracellular environment, facilitating the cell's ability to transmit between individuals. The emergence of two contagious cancers in the Tasmanian devil suggests that the inherent plasticity of Schwann cells confers a vulnerability to the formation of contagious cancers.

The allopurinol metabolite, oxypurinol, drives oligoclonal expansions of drug-reactive T cells in resolved hypersensitivity cases and drug-naïve healthy donors.

Allopurinol (ALP) is a successful drug used in the treatment of gout. However, this drug has been implicated in hypersensitivity reactions that can cause severe to life-threatening reactions such as Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN). Individuals who carry the human leukocyte antigen (HLA)-B*58:01 allotype are at higher risk of experiencing a hypersensitivity reaction (odds ratios ranging from 5.62 to 580.3 for mild to severe reactions, respectively). In addition to the parent drug, the metabolite oxypurinol (OXP) is implicated in triggering T cell-mediated immunopathology via a labile interaction with HLA-B*58:01. To date, there has been limited information regarding the T-cell receptor (TCR) repertoire usage of reactive T cells in patients with ALP-induced SJS or TEN and, in particular, there are no reports examining paired αßTCRs. Here, using in vitro drug-treated PBMCs isolated from both resolved ALP-induced SJS/TEN cases and drug-naïve healthy donors, we show that OXP is the driver of CD8+ T cell-mediated responses and that drug-exposed memory T cells can exhibit a proinflammatory immunophenotype similar to T cells described during active disease. Furthermore, this response supported the pharmacological interaction with immune receptors (p-i) concept by showcasing (i) the labile metabolite interaction with peptide/HLA complexes, (ii) immunogenic complex formation at the cell surface, and (iii) lack of requirement for antigen processing to elicit drug-induced T cell responsiveness. Examination of paired OXP-induced αßTCR repertoires highlighted an oligoclonal and private clonotypic profile in both resolved ALP-induced SJS/TEN cases and drug-naïve healthy donors.

Dominant protection from HLA-linked autoimmunity by antigen-specific regulatory T cells.

Susceptibility and protection against human autoimmune diseases, including type I diabetes, multiple sclerosis, and Goodpasture disease, is associated with particular human leukocyte antigen (HLA) alleles. However, the mechanisms underpinning such HLA-mediated effects on self-tolerance remain unclear. Here we investigate the molecular mechanism of Goodpasture disease, an HLA-linked autoimmune renal disorder characterized by an immunodominant CD4+ T-cell self-epitope derived from the α3 chain of type IV collagen (α3135-145). While HLA-DR15 confers a markedly increased disease risk, the protective HLA-DR1 allele is dominantly protective in trans with HLA-DR15 (ref. 2). We show that autoreactive α3135-145-specific T cells expand in patients with Goodpasture disease and, in α3135-145-immunized HLA-DR15 transgenic mice, α3135-145-specific T cells infiltrate the kidney and mice develop Goodpasture disease. HLA-DR15 and HLA-DR1 exhibit distinct peptide repertoires and binding preferences and present the α3135-145 epitope in different binding registers. HLA-DR15-α3135-145 tetramer+ T cells in HLA-DR15 transgenic mice exhibit a conventional T-cell phenotype (Tconv) that secretes pro-inflammatory cytokines. In contrast, HLA-DR1-α3135-145 tetramer+ T cells in HLA-DR1 and HLA-DR15/DR1 transgenic mice are predominantly CD4+Foxp3+ regulatory T cells (Treg cells) expressing tolerogenic cytokines. HLA-DR1-induced Treg cells confer resistance to disease in HLA-DR15/DR1 transgenic mice. HLA-DR15+ and HLA-DR1+ healthy human donors display altered α3135-145-specific T-cell antigen receptor usage, HLA-DR15-α3135-145 tetramer+ Foxp3- Tconv and HLA-DR1-α3135-145 tetramer+ Foxp3+CD25hiCD127lo Treg dominant phenotypes. Moreover, patients with Goodpasture disease display a clonally expanded α3135-145-specific CD4+ T-cell repertoire. Accordingly, we provide a mechanistic basis for the dominantly protective effect of HLA in autoimmune disease, whereby HLA polymorphism shapes the relative abundance of self-epitope specific Treg cells that leads to protection or causation of autoimmunity.