Highlight
Immune checkpoint inhibitor (ICI)-associated colitis appears to be driven, at least in part, by epithelial barrier failure rather than by an inseparable extension of antitumour immunity.
In human biopsies and a wild microbiota mouse model, the study links ICI colitis to activation of the MLCK1-dependent tight junction leak pathway, downstream of tumour necrosis factor (TNF) produced by CD4+ and CD8+ T cells.
Genetic deletion of MLCK protected tight junction structure and reduced colitis severity in mice, supporting a causal role for epithelial junctional dysregulation.
The small molecule epicatechin disrupted MLCK1-FKBP8 interaction, limited intestinal barrier loss, and mitigated ICI colitis in preclinical models without reducing antitumour efficacy.
Background and clinical context
Immune checkpoint inhibitors have transformed the treatment landscape across melanoma, lung cancer, renal cell carcinoma, mismatch repair-deficient gastrointestinal cancers, and many other malignancies. Antibodies targeting CTLA-4, PD-1, and PD-L1 can induce durable responses in a subset of patients who previously had few long-term options. This benefit, however, comes with a growing burden of immune-related adverse events (irAEs), which often affect the skin, endocrine organs, liver, lung, and gastrointestinal tract.
Among these toxicities, ICI-associated colitis is especially important clinically. It can present with diarrhoea, abdominal pain, bleeding, urgency, fever, and weight loss; severe cases may lead to dehydration, hospitalisation, colon perforation risk, and treatment interruption. Current management relies primarily on corticosteroids, followed in steroid-refractory cases by biologic immunosuppression such as infliximab or vedolizumab. These approaches are often effective, but they are imperfect. Some patients fail to respond, some experience recurrent symptoms, and broad immunosuppression raises concern about infection, steroid toxicity, and potential attenuation of antitumour immunity, even if this latter issue remains context dependent.
A major unmet need is therefore mechanistically targeted therapy that protects the gut without undermining the anticancer effects of checkpoint blockade. The present study by Xiong and colleagues addresses that problem by focusing on epithelial barrier biology. Specifically, the investigators examine long myosin light chain kinase 1 (MLCK1), a regulator of the perijunctional actomyosin ring and tight junction permeability. This is a biologically plausible target because epithelial barrier dysfunction has long been implicated in inflammatory bowel disease and other inflammatory diarrhoeal states, but its causal role in ICI colitis has remained less clearly defined.
Study design and experimental framework
This was a translational, multimodal study integrating human tissue analysis, animal modeling, organoid systems, and molecular pharmacology.
Human component
Biopsies from patients with ICI colitis underwent single-cell RNA sequencing and spatial transcriptomics to define cell populations, inflammatory programs, and tissue-level localization of relevant pathways. The goal was to identify molecular signatures associated with epithelial injury and immune activation in clinical disease.
Animal models
To model ICI colitis more faithfully than conventional specific pathogen-free systems, the investigators used a wild microbiota mouse model referred to as WildR. This choice is noteworthy because microbiota composition strongly influences susceptibility to intestinal inflammation and responses to checkpoint blockade. The study also included genetically modified mice, tumour-bearing models, and antitumour efficacy experiments in melanoma and MC38 settings. These models allowed assessment of both intestinal toxicity and oncologic outcome.
Mechanistic and pharmacologic studies
Mechanisms were further interrogated using intestinal organoid-immune cell coculture systems, allowing separation of epithelial and immune contributions. The authors also used surface plasmon resonance and microscale thermophoresis to explore protein-small molecule interactions, full-spectrum flow cytometry for immune phenotyping, bulk RNA sequencing, immunostaining, ELISA, and gut permeability assays. Together, these techniques provided a coherent pipeline from clinical observation to pathway validation and candidate therapeutic identification.
Key findings
1. Tight junction integrity is compromised in human and experimental ICI colitis
The study found that intestinal barrier dysfunction is a prominent feature of ICI colitis. In patient biopsies, transcriptomic and spatial analyses indicated epithelial perturbation and tight junction disruption. This observation was reproduced in the WildR mouse model after checkpoint blockade, strengthening the argument that barrier injury is not merely an epiphenomenon of severe inflammation in isolated patients but a reproducible component of disease biology.
Conceptually, this matters because intestinal permeability can amplify mucosal inflammation by allowing greater luminal antigen and microbial product exposure to the mucosal immune system. Once initiated, this can create a feed-forward inflammatory circuit. In that sense, barrier loss may function as both target-organ damage and inflammation amplifier.
2. MLCK1-mediated leak pathway is a central effector of barrier failure
The most important mechanistic result is the implication of MLCK1 in ICI-associated barrier dysfunction. MLCK1 regulates contraction of the perijunctional actomyosin ring, which can increase paracellular permeability without necessarily causing wholesale epithelial denudation. This has been described as a tight junction “leak pathway,” distinct from more catastrophic epithelial injury. The authors present evidence that ICI exposure activates this MLCK1-dependent program, leading to compromised tight junction architecture and increased gut permeability.
The translational significance is high. Tight junction dysregulation is potentially more druggable than broad suppression of systemic T-cell function. If epithelial leak is an essential downstream mediator of colitis symptoms and tissue inflammation, then stabilising the barrier may reduce toxicity while leaving antitumour immune priming intact.
3. T-cell-derived TNF sits upstream of MLCK-dependent epithelial injury
The study identifies TNF produced by both CD8+ and CD4+ T cells as an upstream driver of the MLCK-dependent barrier defect. This is biologically consistent with prior gastrointestinal research showing that TNF can increase epithelial permeability through MLCK activation. Here, the authors extend that paradigm into the irAE setting and show that immune activation induced by checkpoint blockade can injure the epithelium through a defined cytokine-barrier axis.
This finding also helps explain why anti-TNF therapy can be effective in steroid-refractory ICI colitis. At the same time, the current work suggests that TNF blockade may not be the only way to interrupt this axis. Targeting a downstream epithelial effector such as MLCK1 could theoretically offer a more tissue-selective approach.
4. Genetic deletion of MLCK protects against ICI colitis
A particularly persuasive part of the paper is the use of genetic deletion approaches. Loss of MLCK preserved tight junction structure, reduced inflammation, and ameliorated ICI colitis in murine models. This is stronger evidence than association alone because it supports causality: the disease phenotype was materially altered when the pathway was removed.
Importantly, this result positions MLCK not only as a biomarker of epithelial stress but as a mechanistic node required for full expression of colitis in this model system.
5. Epicatechin was identified as a candidate MLCK1-targeted therapy
The pharmacologic screen yielded epicatechin as a small molecule capable of blocking MLCK1-FKBP8 interaction. The study indicates that this interaction is important for recruitment of MLCK1 to the perijunctional actomyosin ring, where it drives barrier dysfunction. By interfering with that recruitment step, epicatechin prevented intestinal barrier loss.
This is an appealing strategy because it does not appear to act as general immunosuppression. Rather, it modifies a spatially defined epithelial process linked to inflammation-induced permeability. The biochemical work, including surface plasmon resonance and microscale thermophoresis, provides support that the observed effects are based on direct pathway engagement rather than a nonspecific anti-inflammatory signal.
6. Epicatechin reduced colitis without compromising antitumour efficacy in mice
The translational headline of the paper is that epicatechin mitigated ICI-induced colitis while preserving antitumour effects in tumour-bearing preclinical models. This “uncoupling” is exactly what clinicians seek in irAE management. In practice, toxicity often forces treatment delay or discontinuation, which may jeopardize cancer control. A barrier-restorative therapy that reduces colitis yet leaves immune-mediated tumour killing intact would represent a meaningful advance.
Although the abstract does not provide numerical effect sizes, confidence intervals, or exact tumour response metrics, the direction of effect is clinically compelling. It suggests that epithelial barrier preservation may be downstream enough to spare the core immunologic mechanisms responsible for tumour response.
Mechanistic interpretation
The study supports a model in which checkpoint blockade activates effector T cells, including TNF-producing CD4+ and CD8+ populations. In susceptible intestinal environments, TNF then induces epithelial MLCK1 activation, promoting contraction at the perijunctional actomyosin ring and opening the tight junction leak pathway. Increased permeability exposes mucosal immune cells to luminal antigens and microbial products, amplifying local inflammation and producing clinically evident colitis.
This framework is attractive for several reasons. First, it links systemic immunotherapy to organ-specific toxicity through a plausible tissue vulnerability mechanism. Second, it distinguishes antitumour immunity from downstream epithelial injury, at least partially, thereby explaining how the two processes might be therapeutically separated. Third, it integrates immune and epithelial biology rather than treating irAEs as purely immune cell-driven phenomena.
The work also aligns with a broader shift in mucosal immunology: disease severity in the gut often depends not only on the magnitude of immune activation but also on epithelial resilience, barrier integrity, and microbial context.
Clinical implications
Potential therapeutic positioning
If confirmed in humans, MLCK1-targeted or barrier-restorative therapy could occupy several possible roles. It might be used prophylactically in patients at high risk of gastrointestinal irAEs, early in the course of mild symptoms to prevent escalation, or alongside standard immunosuppression to shorten disease duration and reduce steroid exposure. It may be especially attractive in situations where clinicians are reluctant to intensify systemic immunosuppression because of infection risk or concern about tumour control.
Why this may differ from conventional immunosuppression
Current colitis treatment mainly suppresses inflammation after it becomes clinically significant. By contrast, epithelial barrier protection aims to interrupt a key effector mechanism before severe tissue injury develops. This could be particularly useful in patients whose symptoms are driven disproportionately by permeability changes rather than by deep ulcerative inflammation.
Relevance to current practice
At present, standard management of moderate to severe ICI colitis still rests on guideline-based immunosuppression, usually corticosteroids followed by infliximab or vedolizumab in refractory cases. The current study does not challenge that standard immediately. Instead, it provides a mechanistic rationale for future adjunctive or alternative approaches that could preserve both quality of life and continuity of oncologic care.
Strengths of the study
The study has several important strengths. It combines human patient data with mechanistic preclinical experimentation rather than relying on only one platform. The use of single-cell and spatial transcriptomic methods improves confidence that the pathway identified is relevant to human disease tissue. The WildR model addresses the known limitations of overly simplified microbiota conditions in gut inflammation research. Genetic deletion experiments support causality, and the organoid-immune coculture approach helps dissect epithelial versus immune contributions. Finally, the inclusion of tumour-bearing models is essential because preserving antitumour efficacy is the central translational claim.
Limitations and unanswered questions
Despite its strengths, the study should be interpreted cautiously.
First, much of the therapeutic evidence remains preclinical. Many agents that improve intestinal inflammation in mice do not translate cleanly to humans, especially in the setting of cancer immunotherapy.
Second, the abstract does not provide detailed quantitative outcomes, such as sample sizes for each experiment, effect sizes, variability estimates, or survival and tumour growth statistics. These details are important for judging robustness and for comparing this strategy with existing treatments.
Third, the relevance across different classes of checkpoint inhibitors remains to be clarified. CTLA-4 blockade and PD-1/PD-L1 blockade produce overlapping but not identical patterns of gastrointestinal toxicity. Whether MLCK1 dependence is uniform across regimens, combination therapy, and tumour types is unknown.
Fourth, epicatechin is a biologically familiar small molecule, but pharmacokinetics, formulation, dosing, target specificity, and off-target effects in humans will need rigorous study. A mechanistically promising compound is not automatically a clinically useful drug.
Fifth, barrier restoration may not be sufficient for all patients. Some cases of ICI colitis are severe, ulcerative, or histologically heterogeneous, with features resembling inflammatory bowel disease, microscopic colitis, or graft-versus-host-like epithelial injury. Different endotypes may require different interventions.
Finally, the interaction with the intestinal microbiome deserves further exploration. Because microbiota shape both ICI efficacy and colitis risk, it will be important to determine whether MLCK1-targeted strategies remain effective across different microbial ecologies and whether they alter microbiome composition in beneficial or harmful ways.
How this fits with existing evidence
The findings build on well-established work showing that epithelial MLCK activation contributes to TNF-induced barrier dysfunction in intestinal inflammation. They also complement clinical observations that anti-TNF therapy can rescue many patients with steroid-refractory ICI colitis. What is new here is the proposition that the critical actionable point may be the epithelial leak pathway itself, not only the upstream cytokine network.
This distinction has practical consequences. If a downstream barrier-specific intervention can reduce colitis without blunting systemic antitumour immune responses, then the therapeutic index could be superior to broad immunosuppression in selected patients. That possibility remains to be proven clinically, but it is a credible translational direction.
Future directions
The next steps should include validation in larger human biopsy cohorts with detailed clinicopathologic phenotyping, including correlation with endoscopic severity, steroid responsiveness, and recurrence risk. Biomarkers of epithelial permeability could help identify patients most likely to benefit from barrier-directed therapy.
Prospective translational trials will be needed to determine whether epicatechin, or a more potent and selective MLCK1-directed agent, can be safely combined with checkpoint inhibitors. Key endpoints should include symptom control, need for corticosteroids, time to ICI resumption, recurrence of colitis, infection risk, and oncologic outcomes such as objective response and progression-free survival.
It will also be important to compare barrier-directed therapy against, or in combination with, established options such as infliximab and vedolizumab. Mechanism-based patient stratification may become feasible if epithelial-dominant and immune-dominant irAE phenotypes can be distinguished.
Conclusion
Xiong and colleagues present compelling evidence that ICI-associated colitis is driven in part by TNF-induced, MLCK1-mediated epithelial tight junction dysfunction. In preclinical models, targeting this pathway preserved barrier integrity, reduced colitis, and, importantly, did not compromise antitumour efficacy. The work reframes ICI colitis not solely as unavoidable collateral immune activation but as a potentially separable toxicity mediated through a druggable epithelial effector pathway. Clinical translation will require careful validation, but the study provides a strong rationale for barrier restoration as a new therapeutic strategy in immuno-oncology.
Funding and trial registration
The abstract provided does not report a ClinicalTrials.gov registration number. Specific funding sources should be confirmed from the full published article in Gut.
Citation
Xiong L, Huang J, Dong Y, Han W, Kuo WT, Xu W, Han Y, An C, Zhu R, Zhu N, Xia H, Rahman A, Tang S, Jiang C, Zhao J, Pei W, Wang J, Wang X, Song J, Wang Z, Wu S, Zhang H, Xu H, Wu B, Huang Q, Bao B, Mei Q, Zhu H, Hou L, Liangpunsakul S, Cao F, Weng H, Tan B, Turner JR, Wang H, Zuo L. Targeting MLCK1 uncouples immune checkpoint inhibitor-induced colitis from antitumour immunity. Gut. 2026-05-12;75(6):1147-1159. PMID: 41494803.
Selected references
Martins F, Sofiya L, Sykiotis GP, et al. Adverse effects of immune-checkpoint inhibitors: epidemiology, management and surveillance. Nat Rev Clin Oncol. 2019;16(9):563-580.
Dougan M. Checkpoint blockade toxicity and immune homeostasis in the gastrointestinal tract. Front Immunol. 2017;8:1547.
Abu-Sbeih H, Ali FS, Wang Y. Immune-checkpoint inhibitors induced diarrhea and colitis: a review of incidence, pathogenesis and management. Curr Opin Gastroenterol. 2020;36(1):25-32.
Turner JR. Intestinal mucosal barrier function in health and disease. Nat Rev Immunol. 2009;9(11):799-809.
Su L, Shen L, Clayburgh DR, et al. Targeted epithelial tight junction dysfunction causes immune activation and contributes to development of experimental colitis. Gastroenterology. 2009;136(2):551-563.
