# Section 7B: Regulatory Pathways and Precedent

An antigen-specific or pMHC-based vaccine for eosinophilic esophagitis (EoE) will not be evaluated in a regulatory vacuum. It follows a trail already blazed by Tzield (teplizumab) in type 1 diabetes, KAN-101 in celiac disease, and dupilumab in EoE itself — and it will need to navigate the same categorical tension that has shaped each of those programs: how to credential a disease-modifying, immune-mediated mechanism to a regulator whose institutional muscle memory is built around symptom relief. The strategic lessons are as much about endpoint architecture and diagnostic infrastructure as they are about designation-seeking.

## Expedited-Program Options

The FDA's toolkit for immune-mediated and antigen-directed therapies offers several non-mutually-exclusive routes, each with different evidentiary triggers:

- **Breakthrough Therapy Designation (BTD)** requires preliminary clinical evidence that the drug may offer substantial improvement over available therapy on a clinically significant endpoint. Tzield's BTD was anchored in delay-of-onset data — a hard, objective endpoint — and the designation bought it intensive FDA interaction and rolling review. An EoE pMHC vaccine would need analogous early signal, most plausibly a biomarker or histologic effect size that credibly predicts downstream symptomatic and structural benefit.
- **Fast Track**, already granted to KAN-101 in celiac disease, is the lower-bar entry point: it requires only that the drug address an unmet need in a serious condition, with nonclinical or mechanistic rationale sufficient to open rolling submission and more frequent FDA touchpoints. This is the more realistic near-term target for an EoE antigen-specific program prior to human proof-of-concept.
- **Orphan Drug Designation** is available if the sponsor's target EoE population (e.g., a specific pMHC/allele-restricted subgroup) is defined narrowly enough to fall under prevalence thresholds — a strategic choice with downstream consequences for label breadth and market exclusivity.
- **Accelerated Approval**, built around a surrogate reasonably likely to predict clinical benefit, is the pathway most structurally suited to an antigen-specific vaccine, since immunologic desensitization or antigen-specific tolerance biomarkers are unlikely to constitute a direct measure of clinical benefit on first pass.
- **RMAT** designation, though built for cell and gene therapies, extends to some regenerative and reprogramming immunotherapies; whether a pMHC-based tolerizing vaccine qualifies depends on how the mechanism is characterized (durable immune reprogramming vs. repeat-dosed biologic), and is worth testing with FDA early rather than assuming.

None of these designations substitutes for endpoint strategy; they accelerate interaction and review timelines but do not lower the evidentiary bar for approval itself.

## Endpoint Selection Lessons

Tzield's approval rested on a single, objective, disease-modification endpoint — time to Stage 3 (clinical) diabetes in an at-risk, autoantibody-positive population. That endpoint worked because it was unambiguous, hard, and directly tied to an accepted natural history model of disease staging. It did not require the FDA to accept a subjective patient-reported measure as sufficient on its own.

Celiac disease illustrates the opposite failure mode. There is no approved drug for celiac disease and no FDA-endorsed single primary endpoint. Nexvax2's reliance on protection from symptoms induced by a gluten challenge proved to be a fragile foundation — symptom response to acute challenge is noisy, driven by heterogeneous mechanisms (some non-immune), and does not cleanly map onto the chronic, cumulative mucosal injury that defines the disease. The lesson for EoE is that a symptom-only or challenge-only endpoint, chosen for trial convenience, invites exactly the kind of endpoint controversy that has stalled celiac drug development for over a decade.

EoE sits closer to Tzield than to Nexvax2 in one respect (an objective histologic correlate exists) but is bound by its own disease-specific guidance in a way neither precedent fully anticipated.

## The Co-Primary Challenge for EoE

FDA's EoE-specific guidance requires co-primary endpoints: a validated patient-reported dysphagia/symptom instrument (such as a DSQ-type measure) together with histologic response, defined as reduction in peak eosinophil count on esophageal biopsy. Dupilumab's 2022 EoE approval demonstrates that this dual bar is clearable, but it also demonstrates the operational burden — the pivotal program had to show statistically robust separation on both a subjective, patient-experienced axis and an objective, biopsy-dependent axis simultaneously, with neither endpoint permitted to compensate for a shortfall in the other.

For an antigen-specific pMHC vaccine, this is a materially harder co-primary problem than for a broad anti-inflammatory biologic like dupilumab. An antigen-specific mechanism may plausibly produce durable histologic and immunologic change on a delayed timeline relative to symptom improvement — tolerization is not degranulation blockade, and eosinophil count normalization may lag or lead symptom resolution depending on mechanism of action. Trial designs will need to prespecify a timepoint (or multiple timepoints) at which both axes are assessed, and sponsors should engage FDA early on whether histologic remission alone, in a population with prior objective diagnosis, can serve as a registration-enabling surrogate under Accelerated Approval with dysphagia-symptom confirmation as a post-marketing requirement — an approach that would borrow Tzield's disease-modification logic while still respecting EoE's guidance-mandated symptom axis.

## Companion Diagnostic Strategy

A pMHC-restricted vaccine is inherently genotype-dependent: efficacy is only interpretable, and possibly only present, in patients carrying the relevant HLA allele(s). This makes a companion diagnostic not an optional commercial add-on but a regulatory precondition — analogous to HLA-restricted therapies in other indications where label and diagnostic are co-reviewed. Three strategic issues follow. First, the diagnostic development timeline must run in parallel with, not behind, clinical development, since late-stage trials will need prospective HLA-based enrollment or stratification to generate a clean efficacy signal in the labeled population. Second, sponsors must decide early whether to pursue a single-allele label (narrower population, cleaner effect size, easier orphan-drug qualification) or a multi-allele panel (broader reach, more complex diagnostic and statistical burden). Third, the commercial and access implications of a mandatory genetic screen — testing infrastructure, payer coverage of the companion diagnostic, and equitable access across populations with variable HLA allele frequency — should be modeled well before Phase 3, since diagnostic bottlenecks have historically slowed launch uptake for genotype-restricted therapies more than the therapeutic approval itself.

## Proposed Regulatory Roadmap for an EoE pMHC Vaccine

A phase-appropriate strategy, informed by the precedents above, would proceed as follows:

- **Phase 1 (safety/pharmacodynamics):** Establish safety and tolerability in HLA-genotyped participants; use this stage to generate the preliminary mechanistic and biomarker evidence needed to support a Fast Track (and potentially Breakthrough) designation request, and to open early dialogue with FDA on endpoint architecture and companion diagnostic co-development.
- **Phase 2 (biomarker plus symptom signal):** Enroll genotype-stratified patients; assess peak eosinophil count change alongside a validated dysphagia/symptom instrument to generate the dose-response and effect-size data needed to negotiate the Phase 3 co-primary or Accelerated Approval structure, and to test timepoint assumptions about the relative kinetics of histologic versus symptomatic response.
- **Phase 3 (progression/co-primary confirmation):** Design around FDA's mandated co-primary framework (histology plus patient-reported symptom outcome), while incorporating a longer-term disease-modification or progression endpoint — analogous in spirit to Tzield's delay-of-onset model — to differentiate an antigen-specific tolerizing mechanism from symptom-suppressing standard of care, with companion diagnostic validation completed in parallel to support simultaneous drug/diagnostic filing.

Engagement with EMA under PRIME should track this same staged logic; Tzield's progression through EMA's PRIME scheme (with EU marketing authorization anticipated around 2026) suggests that disease-modification and delay-of-onset framing can be credentialed under EMA's parallel expedited scheme, making transatlantic alignment on endpoint definitions — particularly around what constitutes an acceptable co-primary or surrogate structure for EoE — a priority for sponsor engagement well before pivotal trial design is locked.