# Gut-Restricted Therapeutics Across GI and Systemic Disease

### A systematic landscape of drugs and biologics engineered to act locally in the gut with minimal systemic exposure

Gut-restriction is not one idea but a family of them. The most useful thing this landscape reveals is that the family is already clinically validated: several distinct strategies have produced approved drugs, and the question for a new program is which strategy fits the target, not whether the concept works. The design goal is constant: concentrate drug action at the diseased intestinal segment while keeping the plasma exposure — and therefore the off-target, systemic toxicity — as low as possible. The motivation is concrete: in the index eosinophilic-esophagitis case that started this project, a patient on systemic dupilumab developed drug-associated conjunctivitis requiring steroid eye drops, whereas a gut-restricted equivalent would have kept the biology where the disease is. That trade-off, efficacy where you want it and exposure where you don't, is the axis this entire field turns on.

This review synthesizes 207 relevant papers (screened from 1,016 retrieved across OpenAlex, PubMed, and citation-graph expansion), a 374-trial ClinicalTrials.gov pipeline, and 22 approved agents from Drugs@FDA. The ceiling is honest: this is an open-API, abstract-level synthesis rather than a full-text read of every paper, and the numeric pipeline counts reflect the specific intervention queries run, not an exhaustive census.

## What is established

**Biological gut-selectivity works when the target itself is gut-restricted.** The clearest validation of the whole concept is vedolizumab, an anti-α4β7 integrin antibody that is dosed systemically but acts only in the gut because its target, the α4β7/MAdCAM-1 lymphocyte-homing axis, is gut-specific. Its selective antagonism of the α4β7-MAdCAM-1 interaction was characterized before the pivotal trials ([Soler 2009](https://doi.org/10.1124/jpet.109.153973)), and it went on to establish efficacy in both Crohn's disease ([Sandborn 2013](https://doi.org/10.1056/nejmoa1215739)) and ulcerative colitis, later beating adalimumab head-to-head in UC ([Sands 2019](https://doi.org/10.1056/nejmoa1905725)). The lesson generalizes: you do not always need to physically confine the drug if the biology confines the effect ([Park 2018](https://doi.org/10.3748/wjg.v24.i17.1868)).

**Physicochemical non-absorption is the oldest and most robust tactic.** A drug that cannot cross the epithelium in quantity acts locally by default. The aminosalicylates (mesalamine, balsalazide, sulfasalazine) have exploited azo-bond bacterial cleavage and pH/matrix formulation to deliver 5-ASA to the colonic mucosa for decades, and the guanylate-cyclase-C agonists linaclotide and plecanatide are approved minimally-absorbed peptides acting on the luminal face of the epithelium. Rifaximin, a non-absorbed rifamycin, is the same principle in antimicrobial form — its clinical benefit in hepatic encephalopathy ([Bass 2010](https://doi.org/10.1056/nejmoa0907893)) comes precisely because it stays in the gut lumen.

**Local delivery of a systemically-toxic drug class rescues its therapeutic index.** Budesonide is the paradigm: a potent corticosteroid engineered for high first-pass hepatic extraction and ileal/colonic release, so the mucosa sees drug and the body largely does not. It was validated in Crohn's disease from the mid-1990s ([Greenberg 1994](https://doi.org/10.1056/nejm199409293311303)). Colon-targeted delivery (pH-triggered, multi-matrix MMX, microbiota-enzyme-triggered) is now a mature formulation science extending from steroids to modern agents.

## What is emerging

**Gut-restricted kinase inhibition is the frontier that most directly addresses systemic toxicity.** Oral JAK inhibitors work in IBD — tofacitinib showed activity in Crohn's disease ([Panés 2017](https://doi.org/10.1136/gutjnl-2016-312735)) — but the class carries a boxed warning for infection and thrombosis driven by systemic exposure. The engineering response is a JAK inhibitor designed for minimal absorption, so mucosal concentration vastly exceeds plasma. Izencitinib (TD-1473) is the lead example, with gut-selective TYK2 approaches close behind. Clinical readouts have been mixed, and this is the strategy where the efficacy-versus-exposure trade-off is being tested most explicitly.

**Sphingosine-1-phosphate modulation confines its benefit to the gut through mechanism, not chemistry.** Oral S1P₁ modulators trap lymphocytes in lymph nodes; ozanimod is approved for UC ([Sandborn 2021](https://doi.org/10.1056/nejmoa2033617)) and etrasimod has followed. They are systemically distributed, but the inflammatory benefit is gut-predominant.

**Oral and protease-stable biologics are the hardest and highest-value emerging modality.** The prize is a macromolecule that survives the GI tract and engages a mucosal target without being absorbed: single-domain antibodies (VHH/nanobodies) engineered for gut delivery, oral α4β7 peptide antagonists (PTG-100, PN-943), and even engineered commensal bacteria that secrete a therapeutic matrix in situ ([Praveschotinunt 2019](https://doi.org/10.1038/s41467-019-13336-6)). The delivery science for oral peptides and nanoparticles across the intestinal barrier is advancing in parallel ([Lundquist 2016](https://doi.org/10.1016/j.addr.2016.07.007)).

**Intestine-restricted metabolic targeting extends the concept beyond inflammation.** Gut-restricted engagement of intestinal nuclear receptors and transporters can treat systemic metabolic disease without systemic receptor activation: an intestine-selective FXR inhibitor improves obesity and insulin resistance ([Jiang 2015](https://doi.org/10.1038/ncomms10166)), and a gut-restricted TGR5 agonist reproduces part of the metabolic benefit of bariatric surgery ([Chaudhari 2020](https://doi.org/10.1038/s41589-020-0604-z)). This is the strategy's expansion frontier into diabetes and obesity.

## What is contested or has failed

The central tension is that gut-restriction and efficacy can trade off: confine a drug too tightly and it may not reach the transmural or submucosal compartment where disease also lives. The most instructive failure is mongersen,, an oral SMAD7 antisense oligonucleotide that produced a strikingly positive Phase 2 signal in Crohn's disease ([Monteleone 2015](https://doi.org/10.1056/nejmoa1407250)) but then failed to replicate in the larger Phase 3 program, a reminder that a locally-acting mechanism with early promise is not de-risked until it survives confirmatory trials. Gut-selective JAK inhibitors face the same open question in the opposite direction: whether restricting exposure preserves enough efficacy to justify the safety gain.

## The pipeline is real but unevenly distributed

The 374-trial pipeline is heavily concentrated in ulcerative colitis (82 trials in the queried set) and Crohn's disease, with meaningful activity in IBS and hepatic encephalopathy. Celiac disease has a genuine early-stage presence (14 trials), but eosinophilic esophagitis is almost absent from the gut-restricted small-molecule/biologic pipeline as queried — a striking gap given that EoE is a mucosal disease of exactly the kind gut-restriction suits, and that the field's current EoE standard (systemic dupilumab) is what motivated this review.

## The EoE and celiac opportunity

Both diseases that anchor this project are mucosal, food-antigen-driven, and (critically) populations that have specific reasons to prefer gut-restricted therapy. The EoE index case's dupilumab-associated conjunctivitis is the concrete argument, since a systemic type-2 biologic carries off-target risk that a locally-acting agent would avoid. The celiac community, as the patient-preference evidence in this project's omics stream showed, is diet-adherent and cautious about systemic immunosuppression, favoring low-burden oral options. Gut-restriction is therefore not a nice-to-have for these two indications; it is the modality the patients are asking for.

The landscape maps cleanly onto the targets this project has already nominated. For **EoE**, the type-2 axis (TSLP, IL-13, IL-4Rα) and the JAK-STAT signaling below it are all candidates for the gut-restricted-kinase or colon/esophageal-targeted-delivery strategies rather than systemic biologics. For **celiac**, this project's two therapeutic streams both fit established gut-restriction templates: the oral TG2 inhibitor concept (ZED1227-class, now in celiac Phase 2 per ClinicalTrials.gov NCT07298343, following an earlier Phase 1 study NCT02679014) is a textbook luminal-enzyme, minimally-absorbed small molecule, and the omics stream's Tier-1 recommendation of a gut-selective JAK/TYK2 inhibitor maps directly onto the izencitinib strategy. The Tier-2 protein designs from this session (the soluble MICA-decoy against NKG2D and the IL-15 binder) belong to the oral/protease-stable-biologic frontier: high-value biology whose translational path runs through exactly the gut-delivery engineering (VHH formats, protease resistance, mucus anchoring) that this landscape identifies as the hardest but most differentiated modality.

## Open gaps

Three gaps are specific enough to act on. First, **EoE is under-served by gut-restricted design** despite being an ideal candidate, since the esophageal-targeted-delivery and locally-acting-kinase strategies are largely unexploited there. Second, **the oral-biologic modality is where unmet need and engineering difficulty coincide**: nanobody and peptide formats that survive the gut and engage mucosal targets would unlock the high-value biologic targets (IL-15, NKG2D-MIC, TSLP) for the tolerability-sensitive populations, but the delivery science is not yet routine. Third, **the efficacy-versus-restriction trade-off lacks a predictive framework** — the mongersen failure and the mixed gut-JAK readouts show the field cannot yet predict, before Phase 3, how much restriction a given target tolerates. A target on the luminal or apical-epithelial face (TG2, α4β7, GC-C) tolerates tight restriction; a submucosal or transmural target may not, and knowing which is which up front is the missing piece.

## Deliverables

| Artifact | Contents |
|----------|----------|
| gut_restricted_master_reference_library.csv | 1,016 screened papers; 207 relevant, tiered, part-tagged |
| gut_restricted_trial_landscape.csv | 374 ClinicalTrials.gov trials, phase × disease |
| gut_restricted_approved_drugs.csv | 22 approved agents (Drugs@FDA) with application numbers |
| gut_restriction_strategy_taxonomy.csv | 9 gut-restriction strategies, mechanism → example |
| gut_restricted_evidence_map.png | Evidence volume, temporal depth, agent coverage |
| gut_restricted_trial_landscape_fig.png | Phase × disease heatmap + approved-agent strategy bar |
| gut_restriction_strategy_taxonomy_fig.png | Strategy × modality × maturity, lead example annotated |