# Mechanistic Design Rationale — CCL26 and POSTN Binders for EoE *Grounded in retrieved primary literature (PMIDs cited inline). This brief sets the neutralization goal and the target epitope for each binder before structure-based design.* --- ## Target 1 — CCL26 (eotaxin-3): neutralize the eosinophil-recruiting chemokine ### Why CCL26 is the right chemokine node in EoE - CCL26/eotaxin-3 signals **exclusively through CCR3** and is an eosinophil-selective chemoattractant; it is induced by the Th2 cytokines IL-4 and IL-13 (IL-4 ~100× more potent), synergizing with TNF-α, and is suppressed by glucocorticoids (PMID 12061839). - Eotaxin-3 mRNA is the single most discriminating transcript separating EoE from GERD, tracking with CCR3 and eosinophil counts in patient biopsies (PMID 17900656). Our own cross-cohort meta-analysis confirmed CCL26 up in 9/9 EoE cohorts (pooled +4.56 log2FC), localized to suprabasal epithelium at single-cell resolution. - It sits at a **convergent bottleneck**: multiple upstream drivers (IL-4, IL-13, TSLP→Th2) funnel into epithelial CCL26 production, which then recruits eosinophils via CCR3. Neutralizing the secreted ligand blocks the recruitment step regardless of which upstream cytokine drove it — complementary to (not redundant with) the approved anti-IL-4Rα/anti-IL-5 agents. ### The neutralization mechanism and the epitope to block CCL26 chemotaxis requires **two** molecular interactions, and either is a valid design target: 1. **CCR3 engagement** — the receptor-binding surface (chemokine N-terminus + N-loop) that activates the GPCR on eosinophils. 2. **Glycosaminoglycan (GAG) binding** — CCL26 must bind cell-surface/matrix GAGs (heparan sulfate, serglycin) to form the immobilized haptotactic gradient eosinophils crawl along. The GAG-binding site was mapped by alanine scanning to **basic residues in the α-helical region**, and mutants with reduced GAG affinity show correspondingly reduced chemotactic activity; enzymatic removal of heparan sulfate or anti-HS antibody blocks CCL26-driven migration (PMID 35742962). GAGs guiding chemokine gradients is a general and well-established mechanism (PMID 26701132). **Design goal:** a small neutralizing binder / ligand-trap that sequesters mature secreted CCL26 and occludes its functional surface. Targeting the CCR3-binding face directly blocks receptor activation; targeting the basic α-helical GAG-binding face blocks gradient formation. A binder covering either face neutralizes; the ideal exemplar buries part of both. This is the classic **anti-chemokine ligand-trap** strategy — attractive here because CCL26 is secreted (accessible to an extracellular biologic) and its two functional epitopes are biochemically defined. **Modality note:** an anti-ligand binder/trap is preferable to an anti-CCR3 approach for a de novo design exercise — CCR3 is a 7-transmembrane GPCR (hard to target with a soluble mini-binder and broadly expressed), whereas CCL26 is a compact (~71-residue mature) secreted protein with a clean extracellular target surface. --- ## Target 2 — POSTN (periostin): interrupt the remodeling / feed-forward inflammatory loop ### The barrier question, answered carefully You asked whether targeting POSTN could **restore normal esophageal barrier function**, and whether barrier dysfunction persists in remission. The literature supports a nuanced answer — POSTN is a strong anti-remodeling target and part of the barrier-disruption circuit, but the evidence positions it as an **amplifier downstream of the primary barrier lesion (DSG1 loss), not the direct junctional disruptor**. Being precise about this matters for what a POSTN binder can honestly claim. **The causal circuit (as supported):** - The primary epithelial barrier lesion in EoE is **loss of desmoglein-1 (DSG1)**, driven by IL-13. DSG1 silencing alone weakens epithelial integrity and impairs barrier function (IBF) — and, critically, **periostin (POSTN) is the top induced gene when DSG1 is lost.** DSG1 loss thus potentiates allergic inflammation *through* pro-inflammatory mediators including POSTN (PMID 24220297). So the arrow runs IL-13 → DSG1↓ → barrier impairment **and** POSTN↑. - POSTN is a matricellular protein that does not build matrix structure itself but **acts as a ligand for integrins** to regulate matrix composition, myofibroblast differentiation, and collagen production — the mechanism by which it drives fibrosis and remodeling in Th2-high tissues (lung/asthma: PMID 28918442; EoE remodeling: PMID 23019192, where IL-4/IL-13/TGF-β1 all regulate periostin as an integrin ligand). Luminal eotaxin-3 and periostin both correlate with esophageal remodeling severity in EoE patients (PMID 35405206). **Does the abnormality persist in remission?** Yes — and this is the strongest single argument for the target: - A prospective pediatric transcriptomic+proteomic study (247 biopsies) found that a subset of genes and proteins **remain dysregulated in histologic remission**, and its central finding is **persistent DSG1 downregulation and periostin accumulation in histologic remission** (PMID 39343172). The barrier/remodeling axis does not fully normalize when eosinophils clear. - Functional permeability studies agree that a **subset** of histologically-remitted patients retain **unchanged (elevated) esophageal epithelial permeability** after treatment, tracking with residual IL-13/CAPN14 and mast-cell/IgG4 findings (PMID 35888006). (Note the counterpoint: *small-intestinal* permeability by lactulose:mannitol does normalize in remission, PMID 24957264 — i.e. the persistence is an esophageal-epithelial phenomenon, not systemic.) - Barrier function is restorable in principle: IL-13-compromised esophageal epithelium regains TEER and re-expresses DSG1/FLG when the IL-13 insult is countered (SCFA model, PMID 34458999) — showing the junctional machinery can recover if the driving circuit is interrupted. ### The honest design goal for POSTN A POSTN-neutralizing binder is best framed as an **anti-remodeling / loop-breaking** agent rather than a direct tight-junction repair molecule. Mechanistically it would: 1. Block POSTN's integrin-binding surface, interrupting the matricellular signaling that drives subepithelial fibrosis, myofibroblast activation, and the DSG1↓→POSTN↑ feed-forward loop; and 2. Thereby *indirectly* support barrier recovery by removing a persistent pro-inflammatory/ pro-fibrotic signal that itself helps sustain the injured, DSG1-low epithelial state. This is a differentiated mechanism from every agent in the current EoE pipeline (all of which target upstream cytokines or eosinophils), and it directly addresses the **residual disease that persists in remission** — a genuine unmet need. What it should *not* claim is direct re-assembly of desmosomes; barrier restoration would be a downstream consequence of interrupting the remodeling circuit, not a direct binding effect. **Design goal:** a neutralizing binder against POSTN's integrin-interaction surface (the FAS1 domains that engage αv-integrins). Target the mature secreted protein; focus the binder on the integrin-ligand face. --- ## What this sets up for structure-based design | | CCL26 | POSTN | |---|---|---| | Target form | Mature secreted chemokine (~71 aa) | Secreted matricellular protein (FAS1 domains) | | Functional epitope to block | CCR3-binding face and/or basic α-helical GAG-binding site | Integrin (αv) interaction surface on FAS1 domains | | Neutralization logic | Sequester ligand → block CCR3 activation + gradient formation → stop eosinophil recruitment | Block integrin signaling → interrupt remodeling + DSG1↓→POSTN↑ loop | | Modality | Anti-ligand mini-binder / trap | Anti-ligand mini-binder / trap | | Differentiation | Convergent chemokine bottleneck, complements anti-IL4Rα/IL5 | Novel anti-remodeling MoA; targets remission-persistent biology | **Caveats.** These are target-rationale conclusions from published mechanism, not a claim that a binder will succeed — de novo binders require experimental validation. POSTN's barrier benefit is mechanistically indirect (via remodeling/loop interruption), and POSTN has physiological roles in tissue repair, so selectivity and a neutralizing-not-agonist profile would need wet-lab confirmation.