# Part III — pMHC Diagnostics and Therapeutics: Antigen-Specific Approaches for EoE *A review of the emerging field of peptide-MHC (pMHC)-based diagnostics and antigen-specific therapeutics, spanning the EoE-specific primary literature and the broader precedents in allergy and autoimmunity that this new area borrows from. Because the field is young, this part cites both primary papers and reviews, and is explicit about what is EoE-specific versus adapted from adjacent diseases. Part III of a four-part systematic review. All citations DOI-verified.* --- ## 1. Why antigen-specificity is the right frame for EoE Parts I–II established that EoE is a food-antigen-driven, CD4⁺/Th2-mediated disease in which a definable set of dietary proteins (milk, wheat, soy, egg being the most common) triggers a class-II-restricted T-cell response. This has a direct corollary: the disease-causing agent is a *specific peptide presented on a specific MHC-II molecule to a specific T-cell receptor (TCR)*. That pMHC–TCR triad is simultaneously (a) a **diagnostic** handle — detect the food-reactive T cell, name the trigger — and (b) a **therapeutic** handle — silence or delete that clone to induce tolerance. Both are more precise than the current standard (empiric elimination + serial endoscopy, Part I) and more disease-modifying than blocking downstream cytokines. This part reviews the tools for each. ## 2. Detecting antigen-specific CD4⁺ T cells Three complementary technologies underpin antigen-specific T-cell detection, all transferable to EoE food antigens: - **pMHC class-II tetramers/multimers.** Fluorescent MHC-II molecules loaded with a defined peptide directly stain TCRs of that specificity, enabling enumeration, phenotyping, and isolation of antigen-specific CD4 T cells (methods: Vollers & Stern, *MHC Class II Tetramers*, doi:10.4049/jimmunol.1102398). The landmark proof that food/antigen-specific CD4 cells can be tracked in a human GI immune disease came from **celiac disease**, where gut-homing gluten-specific T cells were visualized with HLA-DQ2–gliadin tetramers (Ráki 2007, doi:10.1073/pnas.0608610104). Celiac is the direct methodological template for EoE: an HLA-class-II-restricted, dietary-protein-driven T-cell response detectable by tetramer. - **Activation-induced marker (AIM) assays.** Antigen-specific CD4 T cells upregulate surface markers (OX40/CD134, CD25, CD137/4-1BB, CD154/CD40L) after short peptide stimulation, allowing TCR-agnostic, HLA-agnostic detection without pre-built tetramers — a practical advantage when the restricting allele or exact epitope is not yet known (the inclusive, HLA-agnostic path this project favored). - **Paired single-cell TCR-seq + transcriptomics.** Sequencing the TCR alongside the transcriptome of individual T cells identifies clonally expanded, pathogenic-effector-Th2 (peTh2) cells and recovers their exact TCR sequences — the route to defining a disease-causing clonotype. ## 3. The EoE-specific breakthrough: food-specific TCRs (Hill–Spergel platform) The field's defining EoE-specific advance is the identification and validation of a food-antigen-specific TCR in EoE. Dilollo, Spergel, Cerosaletti, Hill and colleagues combined blood single-cell RNA-seq with paired TCR-seq to isolate pathogenic effector-Th2 clonotypes, and validated **eoeTCR-4, which recognizes β-casein (milk) residues ~59–78 restricted to HLA-DRB1\*07:01**, confirmed by tetramer (Dilollo/Hill 2025, doi:10.1016/j.jaci.2025.01.008). This is the first molecularly defined, tetramer-validated EoE food-specific TCR, and it establishes the antigen-specific paradigm as real for EoE, not just borrowed from celiac. It builds on single-cell dissection of the EoE tissue T-cell compartment (Wen 2019, doi:10.1172/jci125917). **Prior-art and attribution (stated plainly).** The Hill–Spergel group's platform — using functional/antigen-specific T-cell assays to identify EoE-causal foods — is **foundational prior art and is patent-protected**. Any diagnostic this project describes operates downstream of, and must be positioned relative to, that work. This project's narrower potential contributions are (i) a *personalized, HLA-typing-guided peptide-pool front end* that pre-computes which food epitopes a given patient can present (to focus the functional assay), and (ii) an *aeroallergen comparator* arm to distinguish food-driven from environmentally-driven responses. These are refinements to the front end of an existing, patented paradigm — not an independent invention — and the commercialization discussion (Part IV / preprint) is bounded by that patent. ## 4. From detection to diagnosis: naming the trigger The diagnostic logic is: identify which food's peptides elicit a *pathogenic-Th2* (not merely detectable) T-cell response in a given patient, and name that food for targeted elimination. The critical scientific subtlety — validated in this project's index-case work — is that **antigen presentation is not pathology**: a patient's MHC-II can present peptides from a food the patient tolerates (e.g. the index case presented wheat/gluten on celiac-risk DQ2.2 but tolerated wheat, consistent with negative tTG-IgA). The presentation prediction is only a *prior*; the **functional Th2/peTh2 readout is the specificity-defining evidence**. This is why a binding-prediction panel alone cannot be a diagnostic, and why the functional assay (AIM/tetramer on patient cells) is the mandatory confirmatory step — the boundary this project's calibration made explicit. ## 5. From detection to therapy: antigen-specific tolerance and anergy If a food-reactive pathogenic clone can be identified, it can in principle be silenced. The therapeutic strategies, ordered by specificity: - **The teplizumab principle (proof that clone-directed intervention modifies disease).** In type-1 diabetes, anti-CD3 (teplizumab) delivers a partial-agonist TCR signal that drives autoreactive effector T cells toward exhaustion/anergy, delaying clinical disease onset in at-risk individuals (Herold 2019, doi:10.1056/nejmoa1902226; earlier Protégé trial in new-onset T1D, Sherry 2011, doi:10.1016/s0140-6736(11)60931-8). Teplizumab is *not* antigen-specific — it modulates all T cells — which is both its power (no need to know the antigen) and its liability (global immunomodulation). It is the template that motivates antigen-specific analogues. - **pMHC-directed, antigen-specific tolerance.** Peptide-loaded pMHC reagents, tolerogenic nanoparticles, and related constructs aim to anergize or delete *only* the disease-specific clone, sparing the rest of the repertoire (reviewed for PLGA/biocompatible-nanoparticle delivery, doi:10.3390/ijms20010204). This is the mechanism this project prioritized for EoE: a food-peptide-loaded reagent to tolerize the food-reactive peTh2 clone — teplizumab's disease-modification with antigen-specific safety. - **Restoring oral tolerance / Treg balance.** The physiological target state is the Treg-dominated tolerance that normally governs responses to fed protein. Tregs suppress effector responses (Thornton & Shevach 1998, doi:10.1084/jem.188.2.287), and oral-tolerance mechanisms can be co-opted therapeutically — though notably the tolerance program can itself be subverted toward a Th2-like phenotype in allergic settings (Noval Rivas 2015, doi:10.1016/j.immuni.2015.02.004), a caution for tolerance-induction strategies. ## 6. Precedents in adjacent diseases (what EoE is borrowing from) The pMHC field is most mature outside EoE, and honest synthesis requires naming the source: - **Celiac disease** — the closest analogue: HLA-DQ2/DQ8-restricted, gluten-peptide-driven, tetramer-trackable gluten-specific T cells (Ráki 2007, doi:10.1073/pnas.0608610104). Celiac provides the diagnostic-tetramer and HLA-restriction template directly. - **Type-1 diabetes** — the therapeutic-modification template (teplizumab; antigen-specific tolerance trials with islet peptides). - **Multiple sclerosis and other autoimmunity** — tolerogenic-nanoparticle and peptide-tolerance approaches in active development. These fields supply methods and mechanistic logic; what they do *not* supply is EoE-specific validation, which (outside the Hill–Spergel TCR work) remains largely to be established. ## 7. Computational epitope and TCR prediction (supporting tools) Antigen-specific programs are increasingly supported by in-silico prediction: MHC-II binding predictors (netMHCIIpan, mhcnuggets — used in this project to pre-rank food epitopes per patient HLA) and emerging TCR–peptide specificity predictors (NetTCR-2.0, doi:10.1038/s42003-021-02610-3; and related models, doi:10.1371/journal.pcbi.1008814). These are hypothesis-generating priors that must be confirmed functionally — the same presentation-≠-pathology caveat from §4 applies. This project's epitope panels are explicitly computational and framed as priors for functional testing, not standalone results. ## 8. Boundaries and honest limits - **The area is genuinely new.** Outside the single validated EoE TCR (eoeTCR-4) and the celiac tetramer lineage, most pMHC-therapeutic evidence is preclinical or from other diseases. Claims for EoE pMHC *therapeutics* are at the hypothesis/rationale stage. - **Prior art constrains scope.** The functional-assay-to-name-food paradigm is patented (Hill–Spergel); this project's role is a downstream refinement, not a foundational claim. - **Computational predictions are priors, not diagnoses.** DQ predictions in particular are less benchmarked than DR, and presentation capacity overpredicts pathology. - **HLA-agnostic by design.** For equitable deployment this project favored AIM-style functional readouts and per-patient personalized panels over HLA-restricted tetramers as an enrollment gate — a deliberate inclusivity choice. --- ### Synthesis EoE is unusually well-suited to antigen-specific medicine: a known class of triggering antigens, a class-II-restricted CD4 mechanism, and now a validated food-specific TCR. The diagnostic opportunity (name the trigger from a blood-based functional T-cell assay) and the therapeutic opportunity (tolerize the food-reactive clone, teplizumab-style but antigen-specific) both follow directly from the mechanism in Part II. The field's maturity is asymmetric — detection methods and the celiac/T1D precedents are solid, the single EoE TCR is a real anchor, but EoE-specific pMHC *therapeutics* remain to be demonstrated — and the foundational diagnostic paradigm is patent-protected. Part IV turns to the genetic/HLA basis, the prevention extension, and the approved-plus-pipeline therapeutic landscape. *Evidence base: pMHC/tolerance Tier-1/2 references in the master library plus the manually-verified Hill 2025 anchor; primary sources and key reviews cited inline, with adjacent-field precedents explicitly labeled as such.*