Diabetes mellitus type 2 (T2D) is associated with accelerated biological aging and the increased risk of onset of other age-related diseases. Epigenetic changes in DNA methylation levels have been found to serve as reliable biomarkers for biological aging. This study explores the relationship between various epigenetic biomarkers of aging and diabetes risk using longitudinal data. Data from the Swedish Adoption/Twin Study of Aging (SATSA) was collected from 1984 to 2014 and included 536 individuals with at least one epigenetic measurement. The following epigenetic biomarkers of aging were employed: DNAm PAI-1, DNAmTL, DunedinPACE, PCHorvath1, PCHorvath2, PCHannum, PCPhenoAge, and PCGrimAge. Firstly, longitudinal analysis of biomarker trajectories was done. Secondly, linear correlations between the biomarkers and time to diabetes were studied within individuals developing diabetes. Thirdly, Cox proportional hazards (PH) models were used to assess the associations between these biomarkers and time of diabetes diagnosis, with adjustments for chronological age, sex, education, smoking, blood glucose, and BMI. The longitudinal trajectories of the biomarkers revealed differences between individuals with and without diabetes. Smoothened average curves for DunedinPACE and DNAm PAI-1 were higher for individuals with diabetes around the age 60–70, compared to controls. Likewise, DunedinPACE and DNAm PAI-1 were higher closer to diabetes onset. However, no significant associations were found between the epigenetic biomarkers of aging and risk of diabetes in Cox PH models. Our findings suggest the potential value of developing epigenetic biomarkers specifically tailored to T2D, should we wish to model and explore the potential for predicting the disease.

2 型糖尿病 (T2D) 与加速生物衰老以及其他与年龄相关疾病的发病风险增加有关。人们发现 DNA 甲基化水平的表观遗传变化可以作为生物衰老的可靠生物标志物。本研究利用纵向数据探讨了衰老的各种表观遗传生物标志物与糖尿病风险之间的关系。瑞典收养/双胞胎衰老研究 (SATSA) 的数据收集自 1984 年至 2014 年，包括 536 名至少进行过一项表观遗传测量的个体。采用以下衰老表观遗传生物标志物：DNAm PAI-1、DNAmTL、DunedinPACE、PCHorvath1、PCHorvath2、PCHannum、PCPhenoAge 和 PCGrimAge。首先，对生物标志物轨迹进行纵向分析。其次，在患有糖尿病的个体中研究了生物标志物与患糖尿病的时间之间的线性相关性。第三，使用 Cox 比例风险 (PH) 模型评估这些生物标志物与糖尿病诊断时间之间的关联，并对实际年龄、性别、教育程度、吸烟、血糖和 BMI 进行调整。生物标志物的纵向轨迹揭示了糖尿病患者和非糖尿病患者之间的差异。与对照组相比，60-70 岁左右的糖尿病患者的 DunedinPACE 和 DNAm PAI-1 的平滑平均曲线更高。同样，DunedinPACE 和 DNAm PAI-1 在糖尿病发病越接近时越高。然而，在 Cox PH 模型中，衰老的表观遗传生物标志物与糖尿病风险之间没有发现显着关联。我们的研究结果表明，如果我们希望建模并探索预测该疾病的潜力，那么开发专门针对 T2D 的表观遗传生物标志物具有潜在价值。