Multiblock proteomic profiling of the molecular link between prediabetes and preclinical atherosclerosis in apparently healthy adults

Background Preclinical atherosclerosis and prediabetes are key targets of preventive medicine as their prevalence rises. Therefore, it is crucial to identify early processes and limit confounders such as lipid-lowering or antidiabetic therapy and advanced atherosclerosis. Proteomics enables the identification of biomarkers and molecular pathways related to atherogenesis in prediabetes. Purpose To investigate the relationship between prediabetes and preclinical atherosclerosis in apparently healthy individuals using a comprehensive proteomic approach. Methods This cross-sectional, population-based study included 389 participants (mean age 49 ± 10 years; 47% males) from the Białystok PLUS cohort in Poland. Individuals with diabetes, major cardiovascular, inflammatory and malignant diseases, or receiving steroidal or lipid-lowering therapy were excluded. Carotid ultrasound was used to assess preclinical atherosclerosis. Prediabetes was defined as impaired fasting glucose, impaired glucose tolerance, or HbA1c 5.7–6.4%. Proteomic profiling was performed using the Olink® Reveal platform, enabling deep profiling of 1050 proteins with the Proximity Extension Assay and next-generation sequencing readout, yielding log2-scaled NPX (Normalized Protein eXpression) values. Proteomic and clinical data were integrated using multiblock sparse partial least squares discriminant analysis (block-sPLS-DA), with prediabetes status as the outcome. Selected proteins were further assessed for associations with vascular parameters (PWV, AIx, IMT) and carotid plaques using multivariable regression models. Correlations between identified proteins and clinical features were summarized using hierarchical clustering and heatmap visualizations (Figure 1). Results: Block-sPLS-DA distinguished individuals with and without prediabetes, retaining eight clinical variables and 45 discriminative proteins. Vascular and age measures cluster together and share correlation patterns distinct from those of BMI and glycaemic parameters. A protein module including ectodysplasin A2 receptor (EDA2R), leiomodin 1 (LMOD1) and neurofilament light chain (NEFL) correlated positively with age and with vascular parameters and inversely with GFR and HDL-C. Expression of EDA2R (B=0.05), C16orf89 (B=0.05), and LMOD1 (B=0.039) was associated with increased IMT. NEFL expression (OR=3.31) was positively associated with the presence of plaques. Furthermore, functional enrichment indicated an overrepresentation of proteins involved in synapse maturation. Conclusion Integrative proteomic modeling revealed vascular traits and carotid plaque associations for proteins independent of clinical confounders. LMOD1, EDA2R, C16orf89 and NEFL were associated with atherosclerosis-related vascular traits and carotid plaque. Enrichment analyses highlighted proteins known to be involved with synapse maturation and neuronal processes as potential mediators between early glucose disturbances and preclinical atherosclerosis.