Gut Gene RHBDF2 Emerges as Molecular Bridge Between Colitis and Brain Barrier Damage

Ulcerative colitis has long been viewed as a disease of the gut, but a growing body of epidemiological evidence suggests that its consequences may reach far beyond the intestinal wall. Patients with inflammatory bowel disease carry a substantially elevated risk of dementia, with a nationwide Taiwanese cohort study finding that they developed dementia more than twice as often as matched controls and roughly seven years earlier. A meta-analysis of seven studies confirmed an increased risk of dementia in inflammatory bowel disease overall, with an even stronger association for Alzheimer’s disease. Now, a new integrative study published in Frontiers in Synaptic Neuroscience offers what may be the most detailed molecular map to date of how chronic intestinal inflammation could compromise the brain, and it points to a single gene, RHBDF2, as a plausible mechanistic bridge between the two conditions.

The research team, led by Jiongjiong Cheng and Aman Xu of the First Affiliated Hospital of Anhui Medical University, began with a large-scale transcriptomic sweep. They performed random-effects meta-analyses across five independent human ulcerative colitis cohorts, encompassing 1,130 patient samples and 508 healthy controls, and three separate mouse cohorts modeling blood–brain barrier disruption caused by cerebrovascular injury with viral infection, cryptococcal meningoencephalitis, and lipopolysaccharide-induced systemic inflammation. Using stringent criteria that required genes to be measured in every cohort, show consistent directionality, and exceed defined effect-size and statistical thresholds, the analysis yielded 245 robust differentially expressed genes in ulcerative colitis and 141 in blood–brain barrier disruption. Notably, all of the barrier-disruption genes were up-regulated, indicating a stereotyped pro-inflammatory transcriptional response of the injured neurovascular unit.

To connect these two disease signatures across species, the researchers mapped the mouse barrier-disruption genes to their human orthologues and intersected the result with the ulcerative colitis signature, applying a three-layer direction-consistency filter. Only four genes survived: TNF, RHBDF2, PILRA, and CXCL16. All four were significantly and concordantly up-regulated in both conditions. TNF showed the largest effect in the barrier-disruption cohorts, while RHBDF2 displayed one of the strongest and most consistent signals in ulcerative colitis, with a meta-analytic effect size of 1.51 log2 fold change and an adjusted p value of 2.1 × 10⁻¹⁵. Protein–protein association network analysis revealed that these four genes form a tightly interconnected module organized around ADAM17, a protease known as the principal sheddase of TNFα, with RHBDF2 serving as its essential activating chaperone.

RHBDF2, also known as inactive rhomboid protein 2 or iRhom2, is a catalytically inactive pseudoprotease that nonetheless plays a decisive regulatory role. It is required for the maturation and substrate-selective activation of ADAM17, particularly in immune cells, and prior work has shown that iRhom2 is up-regulated in the colonic mucosa of patients with inflammatory bowel disease, where it promotes TNFα release and disease severity. To resolve the cellular context of RHBDF2 in the gut, the team analyzed single-cell RNA-sequencing data from colonic mucosa of six ulcerative colitis patients and six healthy controls. The diseased mucosa showed a dramatic loss of epithelial cells, which fell from 41.1 percent of all cells in healthy tissue to 8.4 percent in ulcerative colitis, accompanied by an expansion of plasma cells, myeloid cells, B cells, and CD4 T cells. Within this landscape, RHBDF2 was most robustly induced in the epithelial compartment, with the fraction of RHBDF2-positive epithelial cells rising from 1.4 percent in healthy controls to 6.1 percent in patients.

To probe the functional consequences of epithelial RHBDF2, the researchers employed scTenifoldKnk, a computational method that constructs cell-type-specific gene-regulatory networks and simulates the deletion of a target gene. Virtual knockout of RHBDF2 in colonic epithelial cells significantly perturbed 23 genes, all of which were up-regulated upon deletion. The most strongly de-repressed genes included RPL41, SH2D6, LGALS3, S100A10, B2M, IL6R, and MCL1, and enrichment analysis pointed to nuclear receptor signaling, apoptotic pathways, cellular stress responses, HIF-1 signaling, and neutrophil degranulation. These findings suggest that epithelial RHBDF2 normally acts as a gatekeeper restraining stress and inflammatory programs, and that its dysregulation in ulcerative colitis may couple epithelial injury to innate immune activation.

The team then turned to the blood–brain barrier itself, analyzing a paired single-cell RNA-sequencing dataset from six recurrent glioblastoma patients undergoing focused ultrasound-mediated barrier opening, in which each patient contributed both a barrier-opened and a barrier-intact brain sample. After annotating 86,403 cells into eleven major brain cell types, the researchers performed virtual knockout of RHBDF2 separately in the three core constituents of the neurovascular unit: endothelial cells, pericytes, and astrocytes. The results were striking. Endothelial RHBDF2 perturbed by far the largest transcriptional program, affecting 78 genes, compared with only 14 in pericytes and 2 in astrocytes. The endothelial module included ADAM17 and CXCL16 themselves, complement components C1QA, C1QB, C1QC, and C3, the microglial TYROBP–TREM2 signaling machinery, and vascular regulators such as CAV1, CAV2, S1PR1, and the VEGFA–VEGFR2 pathway, positioning endothelial RHBDF2 as a candidate regulatory hub of the neurovascular unit.

To assess whether these transcriptomic findings have causal relevance to human disease, the researchers performed summary-data-based Mendelian randomization, integrating whole-blood cis-eQTL data from the eQTLGen consortium with genome-wide association statistics from the FinnGen study for four dementia endpoints. Genetically predicted higher expression of RHBDF2 and PILRA was significantly associated with increased risk of all-cause dementia and Alzheimer’s dementia after multiple-testing correction. The association between PILRA and Alzheimer’s dementia was the most robust, being the only one consistent with a shared causal variant in the heterogeneity-dependent instruments test. The authors caution, however, that the RHBDF2 associations showed evidence of heterogeneity and rest on single instrumental variables, making them hypothesis-generating rather than definitive causal evidence.

The most direct experimental support came from an in vivo validation experiment. The team induced colitis in mice using dextran sulfate sodium and treated a subset of animals preventively with GW280264X, a pharmacological inhibitor of ADAM17. DSS-treated mice lost weight progressively, showed elevated disease activity scores, and exhibited marked colonic shortening, all of which were substantially attenuated by the inhibitor. Critically, when the researchers assessed blood–brain barrier permeability using Evans blue dye extravasation, they found that colitic mice accumulated significantly more dye in their brains than healthy controls, directly evidencing barrier disruption, and that ADAM17 inhibition significantly reduced this leakage. Western blot analysis of brain tissue confirmed that cerebral Rhbdf2, Adam17, and TNFα protein levels were all elevated in colitic mice and lowered by the inhibitor, while colonic Rhbdf2 mRNA was likewise up-regulated and partially suppressed by treatment.

These convergent findings support a model in which colitis-associated induction of the RHBDF2–ADAM17–TNFα axis contributes to blood–brain barrier injury, providing a molecular candidate for the emerging concept of a gut–brain vascular axis and the proposed entity of enterogenic dementia. The authors note several important caveats. The mouse barrier-disruption cohorts modeled disparate non-colitic insults, so the shared signature may partly reflect generic neuroinflammation rather than ulcerative colitis-specific biology. The virtual knockout analyses are computational inferences, and the human brain single-cell dataset derives from peritumoral tissue. The in vivo experiment used an acute, preventive, pharmacological regimen without behavioral or cognitive endpoints, and definitive proof of RHBDF2’s function at the barrier will require cell-type-specific genetic knockout models. The researchers plan to quantify key molecules of the axis in the blood of ulcerative colitis patients with different barrier statuses in follow-up work.

From a translational perspective, the identification of a druggable node linking intestinal inflammation to brain barrier dysfunction carries significant implications. Tissue-restricted targeting of iRhom2 has been proposed as a strategy to inhibit ADAM17 in immune cells while sparing its homeostatic functions, which could in principle circumvent the toxicity concerns that have hindered systemic protease inhibition. If the RHBDF2–ADAM17–TNFα axis indeed couples intestinal inflammation to barrier disruption, it may represent a candidate therapeutic target for the central nervous system complications of ulcerative colitis, though the authors emphasize that much work remains to establish causality definitively, including chronic colitis models, cognitive readouts, and validation in human specimens.

Subject of Research: Molecular mechanisms linking ulcerative colitis to blood–brain barrier disruption and dementia risk

Article Title: Multi-cohort meta-analysis, single-cell and Mendelian randomization analyses with experimental validation identify RHBDF2 as a shared mediator of ulcerative colitis and blood–brain barrier disruption

Article References: Multi-cohort meta-analysis, single-cell and Mendelian randomization analyses with experimental validation identify RHBDF2 as a shared mediator of ulcerative colitis and blood–brain barrier disruption. (n.d.). https://doi.org/10.3389/fnsyn.2026.1970578

Image Credits: AI Generated

DOI: 10.3389/fnsyn.2026.1970578

Keywords: ulcerative colitis, blood-brain barrier, RHBDF2, ADAM17, TNF-alpha, dementia, Mendelian randomization, single-cell RNA sequencing, gut-brain axis, inflammatory bowel disease, GW280264X, neuroinflammation