Gut & MicrobiomeResearch PaperOpen Access

Clonal blood mutations linked to higher Crohn's risk in women; Ref-1 inhibition eases colitis in mice

Women with CHIP had higher IBD incidence in UK Biobank and higher odds of IBD in All of Us. In mice with Dnmt3a-mutant marrow, blocking Ref-1 redox signaling eased colitis.

Friday, November 20, 2026 3 views
Published in Blood
a mouse colon specimens of differing lengths laid beside a ruler on a lab bench

Summary

With age, some blood stem cells acquire mutations and produce an expanding share of the body's blood cells, a condition called clonal hematopoiesis of indeterminate potential (CHIP). Among 426,609 UK Biobank participants without IBD at baseline, women with CHIP had higher IBD incidence than women without CHIP. All of Us showed a similar association in women, counting both existing and new cases. Genetic analyses supported a possible causal contribution from CHIP to Crohn's disease. In mice with Dnmt3a-mutant marrow, chemically induced colitis was worse than in controls, in part because of higher expression of the stress-response protein APE1/Ref-1 in marrow and colon. APX3330, which blocks the protein's redox signaling, eased mouse disease. The study did not test treatment in people.

Detailed Summary

Clonal hematopoiesis of indeterminate potential (CHIP) occurs when a mutated blood stem cell expands without detectable blood cancer. Its descendants can show increased inflammatory signaling. This study asked whether CHIP is associated with inflammatory bowel disease (IBD) in people and whether gut inflammation favors mutant blood-cell expansion in mice.

Of 431,279 UK Biobank participants with exome data and no baseline blood cancer, 14,746 (3.4%) had CHIP. After exclusion of existing IBD, 426,609 participants contributed to the IBD analysis; median follow-up was 13.6 years. Women with CHIP had higher IBD incidence (HR 1.29, 95% CI 1.002–1.67); no significant association was found in men. In All of Us (211,028 participants; 8,144 with CHIP), women with CHIP had higher odds of IBD (OR 1.29, 95% CI 1.06–1.58, P=.01). Unlike UK Biobank, this analysis counted both existing and new cases.

For Crohn's disease, the HR in women with CHIP was 1.68 (95% CI 1.19–2.37), driven largely by DNMT3A mutations (HR 1.81, 1.22–2.69); the estimate rose to 2.09 for DNMT3A clones with a variant allele fraction of at least 10%. In men, overall CHIP was not significantly associated with Crohn's, but ASXL1 mutations were (HR 2.40, 1.13–5.07). In a separate analysis by age, the CHIP–Crohn's association was significant in participants over 60 (HR 1.63, 1.19–2.23). Large TET2 clones were linked to ulcerative colitis in people under 45, but the estimate was imprecise (HR 7.84, 1.09–56.22) and likely based on very few cases. No significant CHIP–ulcerative colitis association was seen over age 60.

Two-sample Mendelian randomization supported a possible causal contribution: genetically predicted CHIP was associated with higher odds of Crohn's disease (OR 1.15, 95% CI 1.03–1.27, P=.01), and genetically predicted TET2 mutations with higher odds of ulcerative colitis (OR 1.07, 1.04–1.12, P=.0002). The reverse estimate, Crohn's to CHIP, was close to one and nominally significant (OR 0.98, 0.96–1.00, P=.03).

In a competitive marrow-transplant mouse model, chemically induced colitis favored expansion of Dnmt3a-mutant blood stem and progenitor cells. Mice bearing mutant marrow had worse colitis than controls, attributed in part to higher APE1/Ref-1 expression in marrow and colon. The protein has distinct redox-signaling and DNA-repair functions. APX3330 inhibits its redox signaling, which regulates inflammatory transcription factors including NF-κB, and ameliorated disease in the mutant mice. This identifies a treatment target and a mouse disease-severity benefit, not a demonstrated human treatment or lifespan benefit.

Both heterozygous and homozygous Dnmt3a-deficient mice showed a robust myeloid response to chemically induced colitis; subsequent studies used homozygous deficient marrow, whereas patients typically carry heterozygous mutations. The human cohort estimates are observational, several are subgroup-specific, and the genetic analyses depend on assumptions about the instruments.

Key Findings

  • UK Biobank: 14,746 of 431,279 participants (3.4%) had CHIP. Among 426,609 eligible for the IBD incidence analysis, CHIP was associated with incident IBD in women (HR 1.29, 95% CI 1.002–1.67), not significantly in men.
  • All of Us (n=211,028): women with CHIP had higher odds of IBD (OR 1.29, 95% CI 1.06–1.58, P=.01); the analysis included existing and new cases.
  • Women with DNMT3A CHIP had higher Crohn's incidence (HR 1.81, 95% CI 1.22–2.69), as did men with ASXL1 CHIP (HR 2.40, 1.13–5.07).
  • Mendelian randomization linked genetically predicted CHIP to higher Crohn's odds (OR 1.15, 95% CI 1.03–1.27, P=.01) and predicted TET2 mutations to higher ulcerative colitis odds (OR 1.07, 1.04–1.12, P=.0002). The reverse Crohn's-to-CHIP estimate was close to one (OR 0.98, P=.03).
  • In mice, chemically induced colitis favored expansion of Dnmt3a-mutant blood stem and progenitor cells. Mutant-bearing mice had worse disease than controls, attributed in part to higher APE1/Ref-1 expression in marrow and colon.
  • APX3330 inhibited Ref-1 redox signaling and eased Dnmt3a-driven mouse colitis. The reported benefit concerns disease severity, not lifespan or human treatment efficacy.
  • Large TET2 clones were linked to ulcerative colitis under age 45 (HR 7.84, 95% CI 1.09–56.22), likely on very few cases. No significant CHIP–ulcerative colitis association was seen over age 60.

Methodology

UK Biobank analyses excluded prevalent cases separately for each outcome and estimated incidence over a median 13.6 years. All of Us compared odds using both prevalent and incident cases. Two-sample Mendelian randomization used 26–168 independent, genome-wide significant variants per exposure; the CHIP-to-Crohn's finding was robust in MR-Egger and MR-robust adjusted profile score sensitivity analyses. Mouse experiments compared hematopoietic Dnmt3a deficiency with wild type and used competitive 1:1 marrow transplantation into irradiated recipients followed by dextran sulfate sodium–induced colitis. The supplied excerpt does not give mouse group sizes or the APX3330 regimen.

Study Limitations

The human cohort findings are observational. Some subgroup estimates are imprecise, especially the association between large TET2 clones and ulcerative colitis under age 45. Mendelian randomization depends on instrument assumptions, including that the variants do not affect the outcome through separate pathways. The chemical mouse model does not reproduce all features of human IBD. Both heterozygous and homozygous Dnmt3a-deficient mice were studied initially, but subsequent experiments used homozygous deficient marrow rather than the heterozygous mutations typical of patients. Mouse group sizes and a conflict-of-interest statement are not supplied in the excerpt.

Enjoyed this summary?

Get the latest longevity research delivered to your inbox every week.

Enter your email to subscribe: