Olink proteomics profiling platform reveals myocardial metabolism-associated protein biomarkers in heart failure with preserved ejection fraction (HFpEF)

Background
The pathophysiological mechanisms of heart failure with preserved ejection fraction (HFpEF) are complex and highly heterogeneous, with a continued lack of precise diagnostic biomarkers and targeted therapeutic interventions. Recent studies have demonstrated a strong association between disturbances in cardiac metabolism and the pathogenesis of HFpEF. In this study, we employed a novel targeted proteomics technology—Olink Proximity Extension Assay (PEA)—to analyze cardiac metabolism-related proteins, with the aim of identifying high-performance diagnostic biomarkers and establishing a theoretical foundation for developing therapeutic interventions.
Methods
In this study, we employed the Olink PEA Cardiometabolic Panel to quantify the expression levels of 92 myocardial metabolism-related proteins in plasma samples from the discovery cohort, which included 12 healthy controls (control group, CG), 12 HFpEF patients (HFpEF group, PG), and 12 HFrEF patients (HFrEF group, RG). Differential expression analyses were performed comparing PG to CG and RG to CG. Functional mechanisms of the differentially expressed proteins (DEPs) among the groups were explored via enrichment analyses. Subsequently, Venn diagram analysis was used to identify myocardial metabolism-related proteins uniquely expressed in HFpEF. Validation of the candidate proteins was further performed using enzyme-linked immunosorbent assay (ELISA) in an independent cohort, which included 60 healthy controls, 60 HFpEF patients, and 60 HFrEF patients. Finally, a logistic regression model was constructed to predict HFpEF, and the diagnostic performance of HFpEF-specific DEPs was evaluated by receiver operating characteristic (ROC) curve analysis, accompanied by correlation analyses.
Results
In the PG, we identified 28 DEPs, including 18 upregulated and 10 downregulated proteins. These DEPs are strongly associated with key pathways, including cytokine-cytokine receptor interactions, chemokine signaling, TNF signaling, and PI3K-Akt signaling. Venn analysis identified four HFpEF-specific DEPs—DPP4, KIT, SELL, and NCAM1. Subsequent ELISA validation confirmed significant differential expression of DPP4, KIT, and SELL between the PG and CG (all P < 0.0001), whereas NCAM1 showed no intergroup differences. ROC analysis showed that DPP4, KIT, SELL, and their combined model exhibited good diagnostic performance in distinguishing HFpEF patients from healthy controls (AUC > 0.8). Correlation analysis revealed positive pairwise correlations among the three molecules, with significant positive correlations between KIT and BNP, SELL and BNP, as well as between KIT and NYHA classification.
Conclusion
In this study, we identified DPP4, KIT, and SELL as HFpEF-specific biomarkers. These three biomarkers, individually and combined in a predictive model, demonstrated excellent diagnostic performance, providing clinicians with a novel tool for rapid HFpEF diagnosis. Furthermore, our findings offer new insights into the complex pathophysiological mechanisms underlying HFpEF.