Gut microbial and human genetic signatures of inflammatory bowel disease increase risk of comorbid mental disorders

Inflammatory bowel disease (IBD), encompassing Crohn’s disease (CD) and ulcerative colitis (UC), is characterized by relapsing-remitting gastrointestinal inflammation with substantial quality-of-life impact. Beyond established contributions from host genetics, immune dysregulation, gut dysbiosis, and environment, patients with IBD experience markedly elevated rates of comorbid mental disorders (CMDs), particularly anxiety and depression, compared to the general population. Prior epidemiology links IBD activity to higher anxiety/depression prevalence and suggests bidirectional relationships, including depression preceding IBD onset and reduced IBD risk with antidepressant treatment. Molecular evidence and microbiome studies point to shared biomarkers and a gut–brain axis connecting IBD and mental health. However, the specific host genetic variants and gut microbial features, and their interplay, that predispose IBD patients to CMD remain unclear. This study investigates whether IBD-associated gut dysbiosis and human genetic risk contribute to CMD susceptibility in IBD, and whether host variants modulate CMD-relevant microbial taxa.

Epidemiological studies report high CMD prevalence in IBD (pooled anxiety 31.1%, depression 25.2%), far exceeding general population rates (~3–4%). Disease activity correlates with CMD burden, and CMD in IBD associates with worse prognosis and increased surgery risk, particularly in CD. Depression prior to IBD diagnosis raises subsequent IBD risk, and antidepressant treatment may mitigate IBD risk. Molecular and microbiome evidence suggests shared pathophysiology via a gut–brain axis, with multiple IBD- and CMD-associated taxa reported. Genetic studies show a weak but significant genome-wide correlation between depression and IBD; some IBD-risk variants overlap regulatory regions active in brain tissue. Nonetheless, pleiotropic variants explaining joint mechanisms are not well defined. Prior case-only microbiome work in European cohorts related HADS scores to taxa but showed limited replication across ancestries, highlighting population-specific microbiome and genetic architectures.

Study design and cohort: Prospective single-center cohort at Kyung Hee University Hospital (Seoul, Republic of Korea) enrolled April 2018–October 2022. Participants included 507 IBD patients (290 UC, 217 CD) and 75 carefully screened healthy controls with exclusion criteria for recent antibiotics, probiotics, GI-affecting medications, or acute GI symptoms. HADS psychometric data were available for an IBD subset (n=225; 95 UC, 130 CD); CMD was defined as HADS-A≥11 and/or HADS-D≥11 (n=29 CMD-affected, 196 CMD-free). Ethics approval KHUH 2018-03-006; informed consent obtained. Sample collection and sequencing: Stool collected in NORGEN preservation tubes and stored at −80°C. 16S rRNA V3–V4 regions sequenced on Illumina MiSeq after trimming (Trimmomatic) and processed in QIIME2 with ASV denoising (Deblur), chimera removal, and Greengenes v13.8 taxonomy. ASVs collapsed to phylum–genus. Microbiome analyses: Alpha diversity (Shannon, Simpson, richness by observed features, Faith’s PD) computed after rarefaction (QIIME2 core-metrics-phylogenetic). Beta diversity based on Bray–Curtis dissimilarity; PCoA visualization; group differences by PERMANOVA (adonis). Differential abundance testing used DESeq2 on taxa present in ≥10% of samples in comparison groups. Covariates: sex, age, BMI, alcohol, smoking in case–control; plus IBD subtype, disease activity score (Mayo for UC; HBI for CD), and disease duration in case-only (CMD) analysis. Microbial risk score (MRS): Constructed an IBD-specific MRS per individual as a weighted sum of normalized abundances (z) of CMD-associated genera, scaled by the absolute IBD log2 fold-change (|β̂|) and signed by the direction of IBD effect (sgn(β̂)). The MRS captures enrichment of IBD-risk microbial signatures among CMD-associated taxa. Genotyping and imputation: DNA from 225 unrelated IBD patients genotyped on Korea Biobank Array. Post-QC retained 541,449 variants (call rate ≥97%, MAF ≥1%, HWE P≥1×10−7), homogeneous ancestry, per-sample call rate ≥0.95. Phasing (SHAPEIT v2.r904) and imputation (Minimac3) using 1KGP Phase 3 reference; retained imputed variants with R2>0.4 (~10.69M variants). Genetic correlation and PRS: Between-trait genetic correlations among IBD, depression, and anxiety estimated via LDSC using large European GWAS summary statistics and 1KGP3 EUR LD scores. Individual PRSs computed from genome-wide significant, LD-pruned, independent variants: IBD (n=136; UC 87; CD 119), depression (49), anxiety (70). PRS distributions contextualized with out-of-study Korean controls (n=72,179) and tested between groups. mbQTL analysis: Case-only genome-wide association between host variants (dosage) and normalized genus-level abundances for CMD-risk taxa using multivariable linear regression (RVTESTS), adjusting for sex, age, BMI, alcohol, smoking, IBD subtype, disease activity, mental disorder status, and top 5 genetic PCs. Meta-analysis with a large prior mbQTL dataset (n=18,340; 78% European) via METAL (inverse-variance fixed effects). Locus overlap assessed with known IBD and mental disorder risk variants. Statistics: Group comparisons used t-tests for diversity indices and DESeq2 for differential abundance with FDR correction. Trend across control → CMD-free IBD → CMD-affected IBD tested by Jonckheere–Terpstra. Logistic regression for MRS association with CMD adjusted for covariates used in differential abundance models.

- Cohort: 507 IBD patients (290 UC, 217 CD) and 75 healthy controls; 225 IBD patients had HADS and genetic data; 29/225 (12.9%) were CMD-affected (HADS≥11 for anxiety and/or depression). - Alpha diversity: IBD showed reduced diversity versus controls at genus level: Shannon mean 3.98 vs 4.85 (P≈3.93×10−11); Simpson (P=4.32×10−5), richness (P=9.82×10−13), Faith’s PD (P=1.06×10−11) all lower. CMD-affected IBD patients had further reductions versus CMD-free: Shannon 3.59 vs 3.93 (P≈0.047); richness (P=7.77×10−5); Faith’s PD (P=1.04×10−2); Simpson trend P=6.04×10−2. - Beta diversity: Significant dissimilarity in genus-level composition among groups; PCoA PC1 showed a graded trend (healthy controls < CMD-free IBD < CMD-affected IBD) with strong significance (Jonckheere–Terpstra P=2.24×10−11). - Differential abundance: 106 differentially abundant taxa (DATs) in IBD vs controls; 21 DATs between CMD-affected vs CMD-free IBD (FDR<0.05). Of 18 CMD-DATs with full taxonomy, 10 (55.6%) had prior links to anxiety/depression. Thirteen CMD-associated taxa were also IBD-associated in this cohort; 11/13 changed in the same direction in IBD and CMD (binomial test P=6.05×10−10), indicating enhanced deviation from control levels in CMD-affected IBD. - Microbial risk score (MRS): Higher IBD-specific MRS was associated with increased CMD risk among IBD patients (P=7.33×10−3). High-MRS group had an odds ratio ~5.0 (95% CI 1.7–15.4) for CMD development in multivariable logistic regression. - PRS analyses: Versus controls, IBD cases had higher PRSs for IBD, CD, UC, and depression (8.17×10−13 ≤ P ≤ 0.0182), but not anxiety. Within IBD, no evidence that higher IBD genetic burden increased CMD susceptibility (P=0.641); CMD-affected did not have higher anxiety or depression PRSs than CMD-free (P≥0.321). - mbQTL and host–microbe nexus: At the FUT2–FUT1 locus, the T allele of rs35866622 was associated with lower relative abundance of genus Ruminococcus (torques group) at genome-wide significance (P=2.51×10−8; β=−0.061; 95% CI −0.082 to −0.040). Ruminococcus abundance was reduced in IBD vs controls (P=6.83×10−3) and further reduced in CMD-affected vs CMD-free IBD (P=4.13×10−7). rs35866622 is a known IBD/CD risk variant; a nonsense FUT2 variant (rs601338) is in high LD, implicating H-antigen secretion in shaping CMD/IBD-associated taxa. - Example taxa: Butyrate-producing genera (e.g., Ruminococcus; Coprococcus discussed) are implicated in gut–brain pathways and barrier/inflammatory modulation, aligning with observed associations.

The study shows that IBD-specific gut dysbiosis is accentuated in patients with comorbid anxiety and/or depression, addressing the hypothesis that microbiome perturbations linked to IBD contribute to CMD risk. Reduced alpha diversity and distinct beta diversity patterns in CMD-affected IBD, alongside consistent directions of abundance change for taxa overlapping IBD and CMD, support a shared microbial signature that intensifies with CMD. The IBD-specific microbial risk score (MRS) directly ties enrichment of IBD-risk microbial features to elevated CMD susceptibility, quantifying the gut–brain axis contribution. Genetically, although public GWAS suggest a modest genome-wide correlation between IBD and depression, polygenic risk did not differentiate CMD-affected from CMD-free IBD patients, indicating that common variant burden for IBD or mental disorders plays a limited role in CMD development within IBD, relative to microbiome effects. However, host genetics can modulate CMD-relevant taxa: the FUT2–FUT1 IBD-risk variant rs35866622 associates with lower Ruminococcus abundance, a genus showing protective associations against both IBD and CMD. This highlights a partial mediation where host variants influence dysbiosis that, in turn, elevates CMD risk in IBD. Collectively, findings underscore a mechanistic gut–brain axis in IBD, with dysbiosis as a primary driver and host–microbiome interactions as modulators.

In a large, single-center Korean cohort integrating gut microbiome profiles, host genetics, and psychometrics, the study demonstrates that IBD-associated dysbiosis is amplified in patients with comorbid anxiety/depression and substantially elevates CMD risk as captured by an IBD-specific microbial risk score. While overall polygenic burden for IBD or mental disorders did not distinguish CMD status within IBD, an IBD-risk mbQTL variant at FUT2–FUT1 influenced the abundance of a CMD/IBD-associated genus (Ruminococcus), suggesting genetic mediation of dysbiosis contributing to CMD. The identified microbial taxa and genetic markers represent candidate biomarkers and potential therapeutic targets to mitigate CMD in IBD. Future research should validate these findings in larger, multi-ancestry cohorts; incorporate metagenomics/metabolomics and mucosal sampling; evaluate absolute abundances; and test interventions modulating key taxa or FUT2-related pathways to reduce CMD burden in IBD.

- Small number of CMD-affected IBD patients (n=29) limits power to detect modest microbial and PRS effects. - PRS and mbQTL analyses largely relied on European-derived resources; cross-ancestry transferability and LD/allele frequency differences may limit applicability to Koreans. - Stool-based profiling may not fully represent mucosal microbiota relevant to IBD and CMD. - Analyses used relative, not absolute, abundances; observed changes reflect compositional shifts. - HADS data were unavailable for healthy controls; despite clinical screening, undetected elevated HADS in controls cannot be fully excluded.