Integrated proteomics and machine learning reveal AHNAK as a discriminative biomarker of atrial fibrillation in heart failure

Wengen Zhu, Zhenbang Gu, Yuwen Huang, Min Ye, Ruicong Xue, Yugang Dong, Yuzhong Wu, Chen Liu

Atrial fibrillation (AF) is a frequent complication of heart failure (HF), markedly increasing morbidity and mortality. Although atrial fibrosis is increasingly recognized as a central mechanistic substrate linking HF to AF, the molecular determinants driving fibrotic remodeling and enabling early AF discrimination in HF remain poorly defined. We integrated label-free quantitative plasma proteomics with machine learning-based feature selection to analyze samples from 59 age- and sex-matched patients hospitalized with acute decompensated HF. Differentially expressed proteins were subjected to functional and pathway enrichment analyses, and candidate biomarkers were prioritized using an incremental feature selection framework. AHNAK, the top-ranked candidate, was functionally interrogated in vitro using TGF-β1-stimulated primary mouse atrial fibroblasts with siRNA-mediated knockdown and in vivo using a pressure overload-induced HF model in AHNAK heterozygous mice. A two-sample Mendelian randomization analysis was performed to evaluate the association between genetically proxied AHNAK expression and AF risk. Plasma proteomic profiling identified 126 proteins that were significantly dysregulated in HF patients with AF compared with those in sinus rhythm, among which AHNAK was the most upregulated. Machine learning-based feature selection consistently ranked AHNAK as the most discriminative protein, with robust performance for AF discrimination in HF patients (AUC = 0.897, 95% confidence interval: 0.82–0.97). Functionally, AHNAK heterozygous deficiency promoted a profibrotic phenotype, characterized by increased expression of α-smooth muscle actin, fibronectin, and collagen Iα in atrial fibroblasts and exacerbated atrial fibrosis and cardiac dysfunction in pressure-overloaded mice. Mendelian randomization analysis demonstrated a significant protective effect of higher AHNAK expression against AF, with no evidence of horizontal pleiotropy. This study identifies AHNAK as a discriminative plasma biomarker for AF in HF patients and implicates AHNAK deficiency as a contributor to atrial fibrotic remodeling and AF pathogenesis. External validation in larger and prospective HF cohorts will be required to confirm the clinical utility of AHNAK for AF risk stratification and early detection.