A computational analysis of human tissue samples has identified a group of aging-related gene signals shared by periodontitis and rheumatoid arthritis. The study also found similar estimated increases in plasma cells and gamma-delta T cells in samples from both conditions, providing a shared immune signal for further investigation.
The analysis was not designed to establish cause and effect. Because it used secondary analysis of public gene-expression datasets, its findings are hypothesis-generating, and the functional roles of the candidate genes still require experimental validation.
A large overlap, narrowed by an aging-gene screen
Researchers compared gene activity in periodontitis and rheumatoid arthritis tissue with normal-control samples. The primary periodontitis dataset contained 241 patient samples and 69 normal controls. The rheumatoid arthritis analysis combined datasets containing 51 patient samples and 36 normal controls.
Using a statistical method called limma, the researchers looked for genes whose activity differed between disease and control samples after batch adjustment. They counted a gene as differentially expressed when it met both of the study's stated thresholds: a p-value below 0.05 and an absolute base-2 log fold change above 0.6.
The screen produced 726 differentially expressed genes in periodontitis and 1,536 in rheumatoid arthritis. Of these, 220 were shared between the two diseases. The researchers then compared the shared set with 17,531 aging-related genes listed in GeneCards, yielding 191 aging-related crosstalk genes.
The 191-gene total depends in part on how the study defined aging-related genes. It used a median relevance-score threshold in GeneCards, and the authors caution that this operational definition may reflect bias in the source data.
Five genes and an immune signature
After protein-interaction mapping and machine-learning screening, the analysis reported five confirmed hub genes with higher expression in both periodontitis and rheumatoid arthritis tissues: RAC2, CSF2RB, IL2RG, ENTPD1 and IL10RA. An earlier overlap analysis also listed WIPF1, but the report did not explain why that gene was absent from the later five-gene set.
Pathway analysis highlighted cytokine and cytokine-receptor interaction, chemokine signaling, viral protein interaction with cytokines and their receptors, and phagosome pathways. These were enrichment results rather than functional tests of what the genes or pathways do in either disease.
The immune-cell analysis used CIBERSORT, a computational method that estimates the mix of immune cells represented in a tissue-expression profile. It found a similar increase in plasma cells and gamma-delta T cells in periodontitis and rheumatoid arthritis samples. In periodontitis, IL10RA expression was positively associated with the estimated plasma-cell signal, with a correlation coefficient of 0.61 and a p-value below 0.0001.
Promising model results, limited clinical meaning
The researchers evaluated gene-based nomograms using ROC/AUC analysis, a statistical approach for measuring how well a model separates disease samples from controls. The reported area under the curve was 0.93 for the periodontitis model, with a confidence interval of 0.89 to 0.96, and 0.95 for the rheumatoid arthritis model, with a confidence interval of 0.90 to 0.99.
Those figures indicate strong separation between disease and control samples in the datasets analyzed. They do not establish that the models would work in routine care or in new patients. The study used external datasets for validation, but prospective validation and calibration details were not reported.
The authors note that important information was missing or unevenly recorded, including smoking, genetic factors, autoantibody status and uniform age data. That prevented full adjustment for potential confounding and left open the possibility that some apparent gene differences reflect differences between the datasets or participants rather than a shared disease process.
The authors say larger, better-annotated rheumatoid arthritis tissue cohorts, cross-checks against other aging-gene databases or age-stratified datasets, and laboratory experiments are needed. Such work would test whether the five genes and the highlighted immune pathways have the functional roles suggested by this analysis.
The analysis drew on seven publicly accessible Gene Expression Omnibus datasets. The paper's front matter reports that it was received on 6 March 2026 and accepted on 22 July 2026.
Paper data and sources
Original title: Bioinformatics analysis reveals shared gene signatures and molecular mechanisms between periodontitis and rheumatoid arthritis in the context of aging.
Authors: Hanlu Xiao, Yajun Zhao, Yishu Wang et al.
Journal/Repository: Acta odontologica Scandinavica
Status: Peer-reviewed
First online: 2026-08-21
DOI: 10.2340/aos.v85.46622
Original paper · Full text