Objective:

To identify blood-based DNA methylation signatures that distinguish high-risk prediabetes clusters, emphasizing the significance of these clusters in predicting diabetes risk.

Key Findings:

• More than 120,000 differentially methylated CpG sites were identified across clusters, highlighting the complexity of prediabetes.
• 1,557 CpG sites were selected as stable predictors of cluster membership with over 95% classification accuracy in the discovery cohort, indicating strong predictive power.
• In the independent replication cohort, methylation-based partitioning reproduced high-risk cluster identities with 92% accuracy, confirming the reliability of the findings.
• Cluster-specific markers were enriched in genes related to TGF-β receptor, MAPK signaling, and Wnt pathways, suggesting potential biological mechanisms.
• 1,512 of the 1,557 CpG sites correlated with anthropometric or metabolic traits, particularly insulin sensitivity and glucose levels, underscoring their clinical relevance.

Interpretation:

The findings suggest that peripheral blood epigenetic profiling can effectively stratify metabolic risk in prediabetes, potentially replacing more invasive testing methods and improving accessibility to risk assessment.

Limitations:

• The study population was limited to individuals of Central European ancestry, which may affect generalizability.
• Longitudinal validation is needed to determine predictive value for incident type 2 diabetes or complications, and potential biases or confounding factors should be considered.

Conclusion:

The identification of blood-borne biomarkers for prediabetes could facilitate broader population screening and risk stratification without extensive clinical testing, significantly impacting public health strategies.

Sources:

• Biomarker Research