A blood signature may predict immunotherapy colitis before it strikes

Anyone who has watched a family member start immunotherapy for melanoma or lung cancer knows the fine print: these drugs can save lives, and they can also turn the immune system against the gut. Immune checkpoint inhibitor colitis (ICI-colitis) — severe colon inflammation triggered by cancer drugs that release the brakes on immune cells — sends some patients to the hospital and forces others off treatment entirely. A new study in mice and a small group of melanoma patients, published in Cell Reports, traces the trouble to a surprising source: misbehaving B cells (antibody-producing immune cells better known for fighting infections) that migrate toward the gut before symptoms appear, guided partly by shifts in resident gut bacteria.

Key takeaways:

• Mice and a small group of melanoma patients who later developed severe checkpoint-inhibitor colitis both showed a distinct rise in circulating B cells carrying gut-homing markers weeks before any symptoms began.
• Depleting B cells before starting checkpoint inhibitors prevented colitis in mice, and a fecal microbiota transplant (FMT, replacing gut bacteria with a healthy donor's) calmed the same inflammatory chain reaction.
• In a pilot of 10 patients, the 4 who developed grade 3 colitis already carried a distinct blood signature beforehand, hinting at a possible early-warning test.
• The human findings come from just one small pilot cohort at a single center, so this blood signature needs validation in larger, multi-site studies before it could actually guide patient care.

Rogue Gut-Homing B Cells Appear Before Colitis Symptoms

The core discovery is that B cells start assembling gut-targeting equipment before checkpoint inhibitor colitis produces a single symptom. Researchers gave mice a brief chemical insult to the colon (dextran sodium sulfate, or DSS, which mimics the low-grade gut injury and bacterial imbalance already seen in some cancer patients), then treated them with checkpoint inhibitors — cancer drugs including anti-PD-1 and anti-CTLA-4 antibodies that work by releasing the brakes immune cells normally keep on themselves. Before any symptoms appeared, these "colitis-susceptible" mice showed a surge of circulating B cells equipped with gut-homing receptors (CCR9 and LPAM-1), costimulatory molecules (CD80, CD86, CD40), and inflammatory cytokines including IL-17A and IL-6. Mice given both anti-PD-1 and anti-CTLA-4 together — the combination regimen used in many melanoma patients — developed the most severe colitis of any group, with the heaviest colon damage and the largest influx of colonic T and B cells.

Colitis Is a Major Barrier to Completing Cancer Immunotherapy

The real-world stakes are straightforward: colitis is one of the most common reasons cancer patients have to stop a treatment that's otherwise working. Checkpoint inhibitors have transformed outcomes in melanoma and lung cancer, but they carry a distinct side-effect category called immune-related adverse events, which a comprehensive clinical review of checkpoint inhibitor colitis describes as capable of inflaming the colon, skin, thyroid, liver, or lungs when the unleashed immune system attacks healthy tissue instead of tumor cells. Colitis is among the most disruptive, causing diarrhea, abdominal pain, and sometimes bleeding or bowel perforation — often serious enough to require hospitalization, steroids, or permanently stopping immunotherapy. Combination regimens, like the anti-PD-1 plus anti-CTLA-4 pairing modeled here, carry substantially higher colitis risk than single-drug treatment, exactly the pattern seen in both the mice and the human patients.

A Mouse Model Paired With a Real Patient Cohort

What makes this study more convincing than a typical mouse experiment is that its central finding was tested twice, in two different settings. In mice, the team intervened at the pre-symptomatic stage: depleting B cells with antibodies (targeting CD19, B220, and CD22) before checkpoint inhibitor treatment substantially reduced colon damage and cut pathogenic, cytokine-producing T cells in the colon. Adoptive transfer experiments, moving immune cells from colitis-susceptible mice into healthy mice, showed the effect traveled with the cells, reinforcing that B cells were driving the process rather than just marking it. The team then analyzed baseline blood from a pilot cohort of 10 melanoma patients, mostly stage III or IV, before any started checkpoint inhibitor therapy. Six went on to receive combination immunotherapy and four received single-agent anti-PD-1. Of those 10, 4 patients (all in the combination-therapy group) developed grade 3 colitis by the standard Common Terminology Criteria for Adverse Events (CTCAE) scale, while the other six did not. The four who later developed colitis already had more circulating B cells carrying the same gut-homing markers (CCR9, LPAM-1) and inflammatory signature (IL-17A, TNF-α, IL-6) that had flagged susceptible mice — a cross-species echo hard to dismiss as coincidence, even in such a small sample.

What Patients Considering Immunotherapy Should Know

For a reader wondering what this means for someone about to start, or already on, checkpoint inhibitors, the honest answer is: not yet a clinical test, but a promising lead worth watching. This study doesn't establish that doctors can currently order a blood test to predict colitis risk — the signature described here has been observed, not validated as a diagnostic tool. It does suggest that a patient's baseline gut and immune status may influence colitis risk independent of drug combination. If you or someone you love is starting combination checkpoint inhibitor therapy, that's a reasonable moment to ask an oncologist how colitis risk will be monitored and which symptoms (persistent diarrhea, blood in stool, abdominal pain) warrant an immediate call. This research doesn't support probiotics, restrictive diets, or any gut intervention as prevention — none was tested here, and the study's own microbiome intervention was done under controlled lab conditions in mice, not validated as a home remedy.

A Small Pilot Cohort Leaves the Human Picture Unfinished

The most important limitation is scale: the human portion includes just 10 patients from a single center, with only 4 colitis cases to analyze — far too few to turn a promising correlation into a validated biomarker. The mouse model is also a proxy rather than natural disease: DSS-priming chemically induces gut injury and dysbiosis (microbial imbalance) resembling susceptible patients' guts, but isn't identical to how colitis risk builds in a person over years. The researchers are also still working out exactly how latent dysbiosis feeds into B cell dysfunction — supported by the depletion and transplant experiments, but the precise molecular signals bacteria use remain uncharacterized.

Gut Bacteria Shifts Tracked Closely With Colitis Risk

A parallel thread ties the B cell story back to the gut microbiome itself, echoing broader research on how gut bacteria shape checkpoint inhibitor colitis risk. Colitis-susceptible mice showed a shift in gut bacteria before treatment even began: a drop in Firmicutes, a rise in Verrucomicrobiota (largely the bacterium Akkermansia), and depletion of Prevotellaceae and Roseburia, groups often linked to a healthier gut lining. When researchers gave susceptible mice a fecal microbiota transplant before checkpoint inhibitor treatment, disease activity scores, colon length, and tissue damage all improved, and the transplant reduced the same inflammatory B and T cell signatures (IL-17A, IFN-γ, TNF-α) described above — evidence that dysbiosis and B cell dysfunction are mechanistically linked, not just co-occurring.

Next Steps: Validating a Biomarker and Testing Interventions

The researchers frame their findings as a foundation for two follow-ups rather than a finished product. The first is validating the circulating B cell signature as a pre-treatment biomarker in larger, multi-center cohorts, turning a 10-patient pilot into something statistically robust. The second is testing whether interventions that worked in mice, such as B cell-targeted therapy or microbiome restoration, could move into human trials for patients flagged as high-risk before starting checkpoint inhibitors. No timeline was given, and both need prospective trials that track blood B cell profiles and follow real colitis outcomes. That path has precedent: fecal microbiota transplants have already helped some patients with treatment-resistant checkpoint inhibitor colitis in separate clinical work, suggesting the microbiome piece of this story has a plausible route to the clinic.

Common Questions

Does this mean doctors can now test for immunotherapy colitis risk before treatment? Not yet. The blood signature identified here was observed in a 10-patient pilot study and would need validation in much larger trials before any clinical test could be built on it.

Is checkpoint inhibitor colitis the same as inflammatory bowel disease (IBD)? No, though they can look similar. ICI-colitis is triggered specifically by cancer immunotherapy releasing the immune system's brakes, whereas IBD (Crohn's disease and ulcerative colitis) has different triggers, even though both involve gut inflammation and some overlapping bacterial patterns.

Should someone on immunotherapy try to change their gut bacteria to prevent colitis? This study doesn't support that. Its fecal microbiota transplant experiments were done in mice under controlled conditions, not tested as human prevention, so no specific probiotic, diet, or supplement should be considered proven protection.

Why did combination checkpoint inhibitors cause worse colitis than single-drug treatment? Combining anti-PD-1 and anti-CTLA-4 removes two separate immune brakes at once, which in this study produced the most severe colon damage and the largest B and T cell inflammatory response of any treatment group — consistent with the higher real-world colitis rates seen with combination regimens.

References

1. Naimah Turner, Synat Keam, Noha Abdel-Wahab, et al. (2026). B cell-guided inflammatory cascade perpetuates ICI-mediated colitis. Cell Reports. https://doi.org/10.1016/j.celrep.2026.117969
2. Liansha Tang, Jialing Wang, Nan Lin, et al. (2021). Immune Checkpoint Inhibitor-Associated Colitis: From Mechanism to Management. Frontiers in Immunology. https://doi.org/10.3389/fimmu.2021.800879
3. Guanzhou Zhou, Nana Zhang, Ke Meng, Fei Pan (2022). Interaction between gut microbiota and immune checkpoint inhibitor-related colitis. Frontiers in Immunology. https://doi.org/10.3389/fimmu.2022.1001623
4. Fecal microbiota transplantation for refractory immune checkpoint inhibitor-associated colitis. Nature. https://www.nature.com/articles/s41591-018-0238-9