Dynamical quorum-threshold model of infection-triggered T cell autoimmunity with explicit regulatory T cells and peptide-dependent quorum thresholds

A 2024 PNAS study (Yin, Melton, Huseby, Kardar, and Chakraborty) showed computationally that collective, quorum-threshold effects in T cell activation allow the repertoire to fight pathogens while suppressing autoimmunity, and that the increasing number of pathogenic peptides presented during persistent or severe viral infection raises the probability of activating weakly self-reactive T cells (molecular mimicry) enough to exceed the quorum threshold and trigger autoimmune disease. The model, however, fixes the quorum threshold $Q$ and folds all regulation by regulatory T cells (Tregs) and suppressor CD8+ T cells into that constant. This produces an artifact — rare, strongly self-reactive peptides should eventually cause autoimmunity in everyone — and leaves the model's key mouse-experiment prediction untested. The open problem is the authors' stated next step: build the dynamical version of the model that explicitly represents Tregs and suppressor CD8+ T cells with a peptide-dependent effective quorum threshold (higher for self-peptides that bind T cells more avidly, because thymic interactions with such peptides promote Treg differentiation), determine whether infection-driven autoimmunity survives this extension while the universal-autoimmunity artifact is eliminated, and subject the mechanism to its proposed experimental test — immunizing C57BL/6 mice with increasing numbers of Mycobacterium tuberculosis peptides homologous to mouse self-peptides.

Background

T cells are lymphocytes whose T cell receptor (TCR) recognizes short peptide fragments bound to major histocompatibility complex (MHC) molecules on the surface of antigen-presenting cells; the recognized complex is called a peptide-MHC (pMHC). Because TCRs are generated randomly, some recognize peptides derived from the body's own proteins (self-peptides). Several tolerance mechanisms normally prevent such autoreactive T cells from causing disease: negative selection in the thymus deletes strongly self-reactive clones, and in the periphery, regulatory T cells (Tregs — a suppressive CD4+ lineage, many induced by strong self-recognition during thymic development) and suppressive CD8+ T cells dampen activation. Statistical-mechanics modeling of thymic selection (Košmrlj, Jha, Huseby, Kardar, and Chakraborty, PNAS 2008) predicts that the number of T cells activated by a self-peptide follows a power-law distribution in TCR-pMHC binding strength, so a few rare self-peptides activate unusually many T cells. Butler, Kardar, and Chakraborty (PNAS 2013) proposed that T cell activation in a tissue functions as a quorum: a functional immune response requires a threshold number $Q$ of concurrently activated T cells, which circulating autoreactive T cells alone cannot reach, while pathogen-specific clones can. Yin, Melton, Huseby, Kardar, and Chakraborty (PNAS 2024) combined these ingredients into a computational model of early T cell activation in a tissue: TCR-pMHC interaction strengths were estimated from amino-acid contact energies (a Miyazawa-Jernigan statistical potential, an empirical pairwise amino-acid contact-energy table, which may overemphasize hydrophobic contacts), thymic selection set the initial repertoire, and the parameter $N_T$ captured the number of self-antigens presented, implicitly absorbing inflammatory effects. The model's central finding is that persistent or severe infection — by presenting many pathogen-derived peptides and rare self-pMHC species — increases the probability that weakly self-reactive T cells (molecular mimicry: pathogen peptides resembling self-peptides) get activated, are re-activated by self-antigen, and jointly exceed the quorum threshold, mounting autoimmunity. The authors flag decisive limitations: $Q$ is held fixed (e.g. $Q=10$) with all Treg and suppressor-CD8+ effects absorbed into it, whereas in reality avidly self-binding peptides should induce more Tregs and thus a higher effective $Q$; the fixed-$Q$ model wrongly predicts that rare highly-reactive self-peptides would eventually trigger autoimmunity in everyone; downstream processes are not modeled — tissue-homing integrins (adhesion molecules that guide activated T cells into target tissues), MHC-I cross-priming of self-antigens (presentation of self-derived peptides on MHC class I molecules by dendritic cells, activating CD8+ T cells), and innate-immune inflammation; and the proposed discriminating experiment — immunizing C57BL/6 mice (which express the I-Ab MHC molecule) with increasing numbers of M. tuberculosis peptides homologous to mouse self-peptides, predicted to raise autoimmune-disease incidence relative to a single-peptide control — has not been performed.

Problem Statement

Develop the dynamical, Treg-explicit version of the quorum-threshold model of infection-triggered T cell autoimmunity, and test its central prediction experimentally. Concretely, as the source paper formulates its own next step: (1) Model — extend the published framework so that regulatory T cells and suppressor CD8+ T cells are represented explicitly rather than absorbed into a fixed quorum threshold $Q$, with the effective quorum threshold varying across self-peptides according to their T cell reactivity (self-peptides that bind T cells more avidly in the thymus induce more Treg differentiation and hence require more concurrently activated conventional T cells before autoimmunity can proceed). Within this dynamical model, determine whether the mechanism survives: whether persistent or severe infection, through the presentation of many pathogen-derived peptides and rare self-pMHC species, can still push weakly self-reactive (molecular-mimicry) T cells past the peptide-dependent quorum threshold, and whether the artifact of the fixed-$Q$ model — the prediction that rare, strongly self-reactive peptides would eventually trigger autoimmunity in everyone over long times — is thereby removed. The source attributes that artifact to the structural assumption of a fixed $Q$ rather than to a particular parameter value, so the artifact should not simply reappear at a re-tuned constant threshold. Either outcome resolves this clause: a dynamical model showing that infection-driven molecular-mimicry autoimmunity persists without the universal-autoimmunity artifact, and a dynamical model showing that the quorum mechanism no longer produces autoimmunity once Tregs and suppressor CD8+ T cells are explicit, both answer the question posed. (2) Experiment — test the model's discriminating prediction: immunizing C57BL/6 mice (I-Ab MHC) with increasing numbers of M. tuberculosis peptides homologous to mouse self-peptides, with a single unique peptide as the acute-infection baseline control, should increase the probability of inducing autoimmune disease, with homologs bearing more hydrophobic residues at TCR-contacting positions — which the source suggests could serve as proxies for rare, strongly activating peptides — predicted to raise that probability further. This component is completed by a concrete experimental design — the immunization protocol (the mouse strain, the M. tuberculosis peptide sets together with how their mouse self-homologs are selected, the single-peptide control, and the autoimmune-disease readouts) together with the model's predicted outcome for each cohort — rather than by wet-lab execution, which is not expected of a solver; whether the predicted rise in disease incidence actually occurs in vivo remains the source's open experimental question. The common scientific objective uniting both components is a dynamically faithful account of how persistent infection overcomes peripheral tolerance: the dynamical model supplies the theory, and the peptide-immunization experiment is the source-designated test of the mechanism's veracity.

For the model component the relevant observable is the source's probability of triggering autoimmunity as a function of the number of pathogen-derived peptides targeted, $P(\mathrm{auto}\mid N_f)$ (its Fig. 5), now computed with the peptide-dependent effective threshold; for the experimental component it is autoimmune-disease incidence in the immunized cohorts relative to the single-peptide control.

References

1. Rose Yin, Samuel Melton, Eric S. Huseby, Mehran Kardar, Arup K. Chakraborty, "How persistent infection overcomes peripheral tolerance mechanisms to cause T cell–mediated autoimmune disease", Proceedings of the National Academy of Sciences USA 121(11):e2318599121 (2024). DOI: 10.1073/pnas.2318599121, https://doi.org/10.1073/pnas.2318599121 (full text: PubMed Central PMCID: PMC10945823)
2. Rose Yin, Samuel Melton, Eric S. Huseby, Mehran Kardar, Arup K. Chakraborty, "How persistent infection overcomes peripheral tolerance mechanisms to cause T cell-mediated autoimmune disease", bioRxiv preprint posted 2023-09-16. DOI: 10.1101/2023.09.13.557414, https://doi.org/10.1101/2023.09.13.557414
3. Thomas Charles Butler, Mehran Kardar, Arup K. Chakraborty, "Quorum sensing allows T cells to discriminate between self and nonself", Proceedings of the National Academy of Sciences USA 110(29):11833-11838 (2013). DOI: 10.1073/pnas.1222467110, https://doi.org/10.1073/pnas.1222467110
4. Andrej Košmrlj, Abhishek K. Jha, Eric S. Huseby, Mehran Kardar, Arup K. Chakraborty, "How the thymus designs antigen-specific and self-tolerant T cell receptor sequences", Proceedings of the National Academy of Sciences USA 105(43):16671-16676 (2008). DOI: 10.1073/pnas.0808081105, https://doi.org/10.1073/pnas.0808081105
5. Kathryn Mizzi, Ruben J. Cauchi, "Epstein-Barr virus in multiple sclerosis pathogenesis: The path towards mechanistically faithful models", Multiple Sclerosis and Related Disorders 107:106952 (2026). DOI: 10.1016/j.msard.2025.106952, https://doi.org/10.1016/j.msard.2025.106952
6. Manuel Ruiz-Pablos, Bruno Paiva, Aintzane Zabaleta, "The origin of autoimmune diseases: is there a role for ancestral HLA-II haplotypes in immune hyperactivity", Frontiers in Immunology 16 (2025). DOI: 10.3389/fimmu.2025.1710571, https://doi.org/10.3389/fimmu.2025.1710571