Oral Microbiome, Metabolites and Immune Signals Linked in Recurrent Mouth Ulcers

Recurrent aphthous ulceration, the painful mouth condition better known as canker sores, has long been treated as a mostly local nuisance with murky origins. A new study from China argues that the disease is better understood as a body-wide event, one in which the oral bacterial community, the blood’s chemical fingerprint and the immune system appear to move together in a coordinated pattern. The research, published in Frontiers in Microbiology, combined three types of measurement in the same patients and mapped the correlations among them, producing one of the most integrated portraits of the condition to date.

The team, led by researchers at The First Affiliated Hospital of Henan University of Chinese Medicine, enrolled 16 patients with recurrent aphthous ulceration and 16 age- and sex-matched healthy controls between October 2025 and February 2026. Patients had to be within three days of a new ulcer episode and could not have taken antibiotics, probiotics, immunomodulators or anti-inflammatory drugs in the previous four weeks. Controls had no history of the condition or of chronic systemic disease. The groups were comparable in age, sex distribution and body mass index, with no statistically significant baseline differences.

Each participant provided an early-morning fasting saliva sample and a blood sample. Saliva was used for 16S rRNA gene sequencing of the V3–V4 region to profile the oral bacterial community. Blood yielded two streams of data: flow cytometry measured T lymphocyte subsets, including CD3+, CD4+ and CD8+ cells and the CD4+/CD8+ ratio, while enzyme-linked immunosorbent assays quantified the inflammatory cytokines interleukin-1, interleukin-6 and tumor necrosis factor-alpha. Serum was also subjected to untargeted metabolomics using ultra-high-performance liquid chromatography coupled to high-resolution mass spectrometry, run in both positive and negative ion modes.

The microbial results were suggestive rather than definitive. Alpha diversity indices such as Shannon and Simpson showed only a decreasing trend in patients that did not reach statistical significance. Beta diversity, however, told a stronger story: unweighted UniFrac distances separated patients from controls clearly in principal coordinate and non-metric multidimensional scaling analyses. At the genus level, Fusobacterium, Moraxella and Granulicatella showed nominal trends toward higher abundance in patients, while Treponema and Lancefieldella trended lower. Crucially, none of these taxa survived Benjamini-Hochberg false-discovery-rate correction; the authors report raw p-values below 0.05 for Granulicatella and Lancefieldella that did not remain significant after multiple-testing adjustment. Similarity Percentage analysis identified Neisseria, Prevotella and Haemophilus as the major contributors to the dissimilarity between groups.

Because 16S sequencing cannot directly measure what bacteria are doing, the researchers applied PICRUSt2, a computational tool that infers metagenomic function from marker-gene data. The predicted functional profiles separated patients from controls, and the patient group carried 137 unique predicted functions against 11 in controls, with 6,118 functions shared. Predicted pathways related to amino acid metabolism, carbohydrate metabolism and inflammatory signaling differed between groups. The authors are careful to note that these are potential functional characteristics inferred from phylogeny, not experimentally validated microbial activities, a caveat that applies to any PICRUSt2-based analysis.

The metabolomic layer showed the most extensive remodeling. In the negative ion mode, 55 metabolites differed significantly between groups under the study’s screening thresholds — variable importance in projection above 1.0, fold change beyond 1.5 or below 0.667, and adjusted p-value below 0.05 — with 16 up-regulated and 39 down-regulated. The positive ion mode flagged 204 significantly altered metabolites. Among the highlighted compounds were gamma-glutamylcysteine, a precursor of the antioxidant glutathione; pristanic acid and undecanedioic acid, tied to lipid oxidation; eicosapentaenoic acid, the omega-3 fatty acid that feeds the production of inflammation-resolving mediators; crocetin and cysteinyl-containing dipeptides, associated with reactive oxygen species scavenging; and isethionic acid, linked to microbial sulfur metabolism. KEGG pathway enrichment pointed to cysteine and methionine metabolism, purine metabolism, taurine and hypotaurine metabolism, primary bile acid biosynthesis and, notably, caffeine metabolism, which showed the strongest enrichment signal in the negative ion mode.

An important methodological caveat shadows these metabolite findings. Because authentic standards were not systematically used for confirmation, all identities are putative annotations based on accurate mass, fragmentation patterns and database matching against the Human Metabolome Database and PubChem. The authors also concede that diet-influenced metabolites, including those in the caffeine pathway, may reflect unmeasured dietary habits rather than disease biology, since detailed dietary intake, smoking status, oral hygiene and menstrual cycle phase were not systematically recorded.

The immune measurements were the most statistically robust part of the study. Flow cytometry revealed a significant reduction in CD4+ T-cell percentages in patients — 31.71 percent versus 37.89 percent in controls, with a p-value of 0.030 — while CD8+ levels and the CD4+/CD8+ ratio did not differ significantly. All three measured cytokines were elevated in patients: interleukin-1 at 1.56 versus 0.95 picograms per milliliter, interleukin-6 at 0.84 versus 0.54, and tumor necrosis factor-alpha at 2.35 versus 1.28, with p-values of 0.042, 0.015 and 0.006 respectively. Together these results indicate altered adaptive immune characteristics and a measurable systemic inflammatory state accompanying the ulcers.

The integrative step wove the three datasets together using Spearman rank correlations. Granulicatella correlated positively with all three pro-inflammatory cytokines, while Lancefieldella showed negative correlations with cytokines and positive associations with antioxidant-related metabolites. Gamma-glutamylcysteine, eicosapentaenoic acid, crocetin and the cysteinyl dipeptides correlated negatively with inflammatory markers and positively with indicators tied to CD4+ T-cell balance. Metabolites from purine, taurine and cysteine-methionine metabolism sat at connecting points between microbial abundance and immune indicators in the resulting network. The authors present this as a conceptual framework linking oral dysbiosis, oxidative stress, inflammatory lipid metabolism and immune activation — while explicitly stating that the cross-sectional design establishes none of these relationships as causal or directional.

That limitation deserves emphasis. With 32 participants, a single time point, and no adjustment for lifestyle confounders, the study can describe associations but cannot say whether microbial shifts drive the metabolic and immune changes, whether inflammation reshapes the oral ecosystem, or whether both reflect some upstream trigger. The authors call for prospective studies with comprehensive covariate collection, standardized severity scoring and longitudinal designs. What the work does deliver is a template: by measuring bacteria, metabolites and immune cells in the same individuals simultaneously, it shows that recurrent aphthous ulceration leaves a coordinated signature across all three layers, and it nominates specific taxa and metabolites — Granulicatella, Lancefieldella, gamma-glutamylcysteine, eicosapentaenoic acid — as candidates for the mechanistic studies that must now follow.

Subject of Research: Multi-omics analysis of oral microbiota, serum metabolites and immune markers in recurrent aphthous ulceration

Article Title: Multi-omics integration reveals oral microbiota-metabolite-immune interactions in recurrent aphthous ulceration

Article References: Shao, C., Wang, H., Li, J., Guo, H., Gong, X., Liu, R., He, H., & Shao, M. (2026). Multi-omics integration reveals oral microbiota-metabolite-immune interactions in recurrent aphthous ulceration. Frontiers in Microbiology, 17, Article 1889297. https://doi.org/10.3389/fmicb.2026.1889297

Image Credits: AI Generated

DOI: 10.3389/fmicb.2026.1889297

Keywords: recurrent aphthous ulceration, oral microbiota, metabolomics, 16S rRNA sequencing, PICRUSt2, cytokines, CD4 T cells, oxidative stress, dysbiosis, Frontiers in Microbiology, correlation network, cross-sectional study