Inflammatory epicardial adipose tissue–myocardial crosstalk in coronary artery disease: a single-cell analysis

Introduction Coronary artery disease (CAD) is classically viewed as a vascular disorder, yet systemic inflammation may extend beyond the coronary circulation to the myocardium and adjacent adipose depots. Epicardial adipose tissue (EAT) is increasingly recognized as an active modulator of cardiovascular pathology, but the cellular architecture and signaling programs of myocardial-associated EAT (mEAT) in CAD remain poorly defined. We previously identified two inflammatory mEAT cell states enriched in CAD, marked by transcriptional signatures of immune activation, metabolic stress, extracellular-matrix remodeling, and impaired adipogenesis: inflammatory adipogenic cells (IAC) and inflammatory stromal cells (ISC). Purpose We aim to define mEAT–myocardial crosstalk in CAD at single-cell resolution, with a focus on inflammatory mEAT cell populations and their impact on cardiomyocytes and endothelial cells. Methods Single-nucleus RNA sequencing (snRNA-seq) was performed on paired mEAT and right atrial (RA) myocardium from 9 individuals (18 samples): CAD patients undergoing coronary artery bypass grafting (n=4), non-CAD surgical controls with aortic disease (n=3), and forensic controls without cardiovascular disease (n=2). Batch-corrected integration (Harmony) yielded 124,000 nuclei. Putative ligand–receptor interactions were inferred using CellPhoneDB. Findings were complemented by functional assays using conditioned media from mEAT-derived adipocytes (including IAC) applied to human iPSC-derived cardiomyocyte microtissues and primary endothelial cells. Results Integrated analysis of paired mEAT and RA myocardium revealed prominent mEAT-to-myocardium signaling, with cardiomyocytes, stromal cells, endothelial cells, and macrophages receiving THBS1- and semaphorin-mediated inputs, particularly from inflammatory mEAT populations. Interaction analysis prioritized THBS1–CD36 and SEMA4A/SEMA3C–NRP1/PLXNA2/PLXNA4 axes as candidate mediators linking mEAT inflammation to cardiomyocyte stress responses and microvascular dysfunction, suggesting pathogenic adipose–myocardial communication even in the absence of overt myocardial injury. Functionally, conditioned media from CAD-mEAT adipocytes and inflammatory adipogenic fractions induced structural remodeling, altered contractile dynamics, and activated inflammatory gene programs in cardiomyocyte microtissues, while also impairing endothelial tube formation and migration, supporting a paracrine mechanism of myocardial and microvascular impairment. Conclusions We provide a high-resolution map of mEAT–myocardial cellular interactions in CAD and identify candidate disease-driving signaling pathways. Our findings position inflammatory mEAT cell states as key mediators of adipose–myocardial crosstalk and support mEAT as a potential therapeutic target for risk stratification and intervention in CAD.