A GPR15 homing axis positions regulatory CD8 T cells in the gut, with a 42% reduction in sporadic UC

Background and open questions

Inflammatory bowel disease can be driven jointly by genetics and environment, and early-onset cases are especially suited to the search for monogenic clues. More than 90 genes have been linked to early-onset IBD, but the specific role of GPR15 in human intestinal immune homeostasis has remained unclear; it has been implicated in homing of mouse CD4+ Tregs or effector T cells and described as a receptor responding to GPR15L.

This paper's approach is to connect early-onset IBD families, the function of patient variants, and mouse homing and colitis models. The key question is not to show in general terms that GPR15 is "anti-inflammatory" but to establish which class of T cell it brings to the colonic mucosa and how those cells restrain myeloid inflammation after injury.

Study design

The human discovery work began with 7 individuals with IBD from 4 unrelated families, using WES to look for rare protein-altering variants; most patients had childhood onset. The study then tested GPR15 variant associations in Turkish/Qatari family cohorts, an Iranian cohort and resources such as AoU, and used 293T, HeLa, CHO-K1, Jurkat and patient T cells to measure surface expression, Ca2+ flux and GPR15L migration.

The mechanistic work used patient biopsies, public and in-house scRNA-seq, Gpr15−/− and conditional knockout mice, adoptive transfer and intestinal organoid co-culture to validate homing. Per the Methods, the DSS model used 2.5% DSS in drinking water for 6 days followed by 6–7 days of plain water; sample sizes were not prespecified by statistical calculation, and key between-group differences were tested as stated in the figure legends.

Key results

Variants associated with IBD

The Turkish/Qatari family cohort included 647 people, 74 with IBD; the Iranian cohort included 1,692 people, 109 with IBD. Compared with non-IBD controls, the authors identified 18 rare monoallelic GPR15 variants with MAF<0.001 that were significantly associated with IBD. The discovery cohort also showed clinical manifestations in 3/4 biallelic carriers and 4/10 monoallelic carriers, indicating incomplete penetrance.

Variants impair homing

Functional experiments link the variants to homing defects. WT and F159I had the highest surface expression in CHO-K1, while D306N, Y132S and Y132S/F159I were reduced and Q281X and Y215X were absent; GPR15L-induced Ca2+ influx generally tracked the decline in surface expression, with F159I an exception. Migration of patient T cells toward GPR15L was reduced in WT/M and further reduced in M/M carriers; in Pt 1.1, WT GPR15 restored surface expression and migration, while D306N and Q281X rescued minimally or not at all.

The deficit points to CD8

The core tissue-level readout is not a CD4+ Treg deficit. On the contrary, duodenal and colonic biopsies from 3 patients had a higher ratio of FOXP3+ Tregs to CD4+ T cells; colonic biopsies showed marked reductions in CD8+ T cells in both the epithelium and the lamina propria. The GPR15–GPR15L axis therefore guides a class of colonic CD8+ T cells in humans rather than the CD4+ Treg compartment.

DSS localizes the protective cells

After 2.5% DSS, Gpr15−/− mice lost weight faster and had worse inflammation and more pronounced colon shortening, with Gpr15+/− showing intermediate changes. Conditional knockouts localized this further to CD8+ T cells: Gpr15fl/fl Cd8cre worsened DSS colitis while Gpr15fl/fl Foxp3cre did not. The main mouse groups in Fig. 4 used n=11 per group and were tested by ANOVA, Mann–Whitney or similar.

Macrophages are killed

The immune cell readouts point to macrophages rather than dendritic cell expansion. At day 3 of DSS, F4/80+ macrophages increased in Gpr15−/− colon while CD103+ dendritic cells did not; by day 5 macrophages stained strongly for NOS2 and CD206. In vitro co-culture showed that blocking FasL or TWEAK reduced killing, with combined blockade comparable to anti-FasL alone, indicating that FasL predominates. CD8+ TIGR were reduced 42% in sporadic UC and also significantly reduced in the terminal ileum in Crohn's disease.

Mechanistic interpretation

Demonstrated in the paper: The paper directly shows that several deleterious GPR15 variants reduce receptor surface expression, Ca2+ signalling and GPR15L chemotaxis, with F159I an exception for surface expression and migration; patient T cells agreed with the Jurkat and CHO-K1 results, and reconstitution with WT GPR15 in Pt 1.1 rescued migration while D306N and Q281X did not. Patient biopsies and Gpr15−/− mice both showed loss of colonic CD8+ T cells, with no corresponding deficit in CD4+ FOXP3+ Tregs.

The experiments support each step, but the paper presents the overall claim that "CD8+ TIGR kill inflammatory macrophages through FasL/TWEAK to prevent colitis" as a suggestive model: after 2.5% DSS, colitis worsened in Gpr15−/− mice and CD8-specific but not Foxp3-specific deletion reproduced the phenotype; F4/80+ macrophages accumulated and stained strongly for NOS2/CD206. Activated CD8+ TIGR directly killed LPS-activated macrophages in vitro, and blocking FasL or TWEAK reduced killing, so the death receptor mechanism rests mainly on in vitro blockade.

Author hypotheses: The authors suggest that a reduction in CD8+ TIGR may weaken local restraint on macrophage-driven inflammation in sporadic UC and Crohn's disease, and that CD8+ TIGR TCRs may recognize self, dietary or microbially derived antigens, though those antigens were not identified here. They also propose that enhancing the GPR15 pathway or controlling pathogenic macrophage activity could become a therapeutic direction in IBD.

Limitations and uncertainties

• First, genetic extrapolation is limited by incomplete penetrance. The population frequency of Y132S constrains a monogenic explanation: MAF=0.0012, with 1,916 heterozygotes and 20 homozygotes in gnomAD v.4.1.1, so the authors interpret it as a hypomorphic allele rather than a fully penetrant pathogenic variant.
• Second, the discovery cohort comprised only 7 individuals with IBD from 4 unrelated families; variant carriers were clinically heterogeneous, with penetrance of 3/4 for biallelic and 4/10 for monoallelic carriers. The sporadic UC analysis used HC n=8 and UC n=7 and the Crohn's analysis HC n=8 and CD n=20, which can indicate association but cannot substitute for prospective risk estimates.
• Third, mouse sample sizes were not prespecified by statistical calculation and investigators were not blinded to genotype groups; the DSS model used 2.5% DSS to cause acute epithelial injury with follow-up to day 9 or day 10, which cannot cover chronic relapse or long-term repair. Fourth, the paper does not identify the antigen specificity of CD8+ TIGR, nor does it provide dosing or a safety window for GPR15-targeted therapy.

Clinical and industry implications

If later cohorts confirm the relationship between GPR15 variants, CD8+ TIGR deficits and clinical subtypes, this work could move part of IBD from "generalized mucosal inflammation" toward mechanistic stratification by "impaired homing and uncontrolled macrophages". As a membrane receptor, GPR15 could also become a diagnostic or interventional node.

These conclusions should stay within the paper's evidence, however: patient material supports the genetic association and the tissue deficit, and mouse and in vitro experiments support the causal mechanism; the 42% and the significant reductions in sporadic UC and Crohn's disease remain associations in single-cell datasets and do not show that enhancing CD8+ TIGR has been validated to benefit patients.

Authors, source and verification

Evidence level: Full text; verification record: The full text came from Europe PMC fullTextXML PMC13518229; materials included the Abstract, Results, Methods, figure captions, Discussion and Limitations