Browse
Log in / Register
Share
May 10, 2026 — BCMA-directed CAR-T therapy can trigger a severe form of enterocolitis that goes beyond the loss of mucosal B cells and plasma cells. A new study using intestinal biopsies from multiple myeloma patients shows that persistent mucosal CAR-T cells, inflammatory stromal remodeling, and interferon-driven JAK-STAT signaling are central to this complication. The findings also offer a practical lead: in two patients, the JAK1 inhibitor upadacitinib produced improvements at the clinical, endoscopic, and tissue level.
Ciltacabtagene autoleucel, marketed as Carvykti, is a chimeric antigen receptor (CAR) T-cell therapy that targets B cell maturation antigen (BCMA). It has become a key treatment for relapsed or refractory multiple myeloma, a cancer of plasma cells. But like all CAR-T products, it can cause immune-mediated side effects. Cytokine release syndrome and neurotoxicity are the most discussed. Gastrointestinal toxicity often gets less attention, yet it can be serious. Some patients develop enterocolitis, an inflammatory injury of the colon and/or small intestine that may require endoscopy, hospitalization, and intensive immunosuppression.
The researchers call this complication CAR-T-associated enterocolitis, or CAR-TEC. The goal of the new work was to define what drives it, how the gut tissue changes, and whether those changes point to a specific treatment.
Investigators collected intestinal biopsies from 10 multiple myeloma patients who developed CAR-TEC after BCMA CAR-T therapy. For comparison, they studied 7 myeloma patients who received the same type of treatment but did not develop enterocolitis, as well as 26 healthy volunteers.
Each biopsy was examined with single-cell transcriptomics, flow cytometry, and tissue imaging. That combination allowed the team to profile nearly every cell type in the mucosa, including immune cells, structural cells, and epithelial cells. They also modeled cell-to-cell communication to identify the signaling pathways that were active during the disease.
Both groups of treated patients had pronounced losses of mucosal B cells and plasma cells, which reflects the intended on-target activity of a BCMA-directed therapy. The controls tolerated this depletion. The patients with enterocolitis did not.
In the CAR-TEC group, the mucosa contained an expanded population of cytotoxic CAR-T cells with active effector profiles. Inflammatory myeloid cells were more abundant. Stromal cells shifted toward a less supportive, more inflammatory state. Enteric glial cells, a population often overlooked in gut inflammation, also changed in ways consistent with injury.
The key difference, then, was not whether B cells were depleted. It was whether the rest of the tissue could compensate. In patients without enterocolitis, it could. In patients with enterocolitis, the mucosal system decompensated.
Communication analyses pointed to telocytes, a specialized group of gut stromal cells that help maintain the tissue niche. In CAR-TEC, telocyte-related signals that normally support repair appeared to be lost or distorted. The dysfunction extended beyond the stroma: endothelial cells, epithelial cells, and immune cells all showed signs of an interferon-driven inflammatory program.
That program centered on JAK-STAT signaling. This pathway is downstream of multiple inflammatory cytokines, including interferons. It is also druggable. JAK inhibitors are already approved for immune-mediated conditions, which made them a logical candidate for CAR-TEC.
On the basis of the molecular findings, the researchers treated two patients with CAR-TEC using upadacitinib, an oral selective JAK1 inhibitor sold as Rinvoq. Both patients improved on clinical examination, endoscopy, and follow-up biopsy, with histologic evidence that the mucosal injury was healing.
The numbers are small, and the observations are not from a randomized trial. Still, the response fits the biology. Blocking JAK1 reduces the signaling that the study identified as a core driver of tissue damage. That makes upadacitinib a mechanism-based option worth testing in a larger group of patients.
First, diarrhea after BCMA CAR-T therapy should not automatically be attributed to infection or antibiotics. CAR-TEC deserves consideration, and endoscopy with biopsy may be needed to make the diagnosis. Second, the condition is not simply a B-cell depletion effect. That means treatments aimed only at replacing antibodies or controlling plasma cell-related changes may miss the actual inflammatory process. Third, JAK inhibition is a reasonable avenue to explore when standard immunosuppression fails.
The FDA-approved label for Carvykti already lists gastrointestinal adverse reactions, including colitis. These findings add a mechanistic explanation for what clinicians have observed at the bedside and may eventually lead to more structured management pathways.
CAR-T therapies are no longer confined to oncology. BCMA-directed and CD19-directed products are being investigated in autoimmune diseases, where mucosal immunity often plays a central role. The gut is a likely site of toxicity in those settings, too.
If CAR-TEC turns out to be a general pattern for B-cell-targeted CAR-T therapies, the lessons from this study—focusing on persistent mucosal CAR-T cells, JAK-STAT activation, and the need to support the stromal niche—could be used to predict, monitor, and treat gastrointestinal complications before they become severe.
This is a rare complication, and the study's sample size reflects that. Ten cases of CAR-TEC is a small number, but the depth of analysis is what gives the study its weight. The two patient responses to upadacitinib are encouraging but not conclusive.
What is needed now is prospective data. That means tracking the incidence of CAR-TEC across treatment sites, standardizing the use of endoscopy in patients with persistent diarrhea, and testing JAK inhibitors in a controlled fashion. The molecular map from this study provides a framework for designing those next steps.
BCMA CAR-T enterocolitis is not a simple on-target, off-tumor complication. It involves persistent mucosal CAR-T cells, a disrupted stromal niche, and diffuse JAK-STAT-driven inflammatory remodeling across the intestinal wall. The early signal with upadacitinib is consistent with that biology and provides a concrete clinical path forward.
