Plasma proteomics identifies insulin resistance signatures predictive of cardiometabolic disease risk
Key Points:
- Researchers used a multiplatform mass spectrometry and affinity proteomics approach (MWP) to profile over 8,000 circulating plasma proteins across individuals with varying insulin sensitivity, including normal glucose tolerance and type 2 diabetes (T2D), validated by hyperinsulinaemic–euglycaemic clamp measurements.
- They identified 488 plasma proteins significantly associated with insulin sensitivity, with tissue-of-origin analysis revealing major contributions from the liver, adipose tissue, skeletal muscle, and other organs; proteins linked to insulin resistance were enriched in inflammatory and lipid metabolism pathways.
- Exercise intervention (8-week high-intensity interval training) partially remodeled the insulin resistance proteomic signature towards improved insulin sensitivity, notably reducing liver-derived inflammatory proteins and increasing extracellular matrix-related proteins.
- A machine learning model combining 13 plasma proteins and clinical variables accurately predicted clamp-derived insulin sensitivity (R2 = 0.73) and was successfully applied to ~20,000 UK Biobank participants, outperforming traditional clinical proxies in predicting cardiometabolic disease risk including T2D, fatty liver, and cardiovascular conditions.
- Validation in an independent cohort (Inter99) showed the predicted insulin sensitivity score also associated with a broad spectrum of diseases and mortality beyond diabetes, suggesting systemic insulin resistance impacts multiple chronic conditions and highlighting plasma proteomics as a scalable tool for metabolic health assessment and precision risk stratification.