Heart failure (HF) represents the end-stage phenotype of various cardiac pathologies, including hypertension, ischemic heart disease, diabetes, and cardiomyopathies. In response to cardiac stress, activation of the innate immune system triggers the secretion of pro-inflammatory cytokines, including tumor necrosis factor-alpha (TNF-α). Chronic inflammation contributes to adverse cardiac remodeling and progressive ventricular dysfunction. Animal studies have reported that elevated TNF-α levels are associated with cardiac dilation, fibrosis, and contractile dysfunction (1,2). However, due to species differences, findings from animal models may not always translate directly to humans.

In vitro studies indicate that TNF-α overexpression in human induced pluripotent stem cell–derived cardiomyocytes (hiPSC-CMs) induce apoptosis and impair Ca²⁺ cycling. This oxidative stress–driven dysfunction ultimately leads to reduced contractile amplitude (3,4). While such studies provide valuable insight into the effect of TNF-α on CMs, these models lack physiological relevance. hiPSC-CMs exhibit a neonatal-like phenotype, which may respond differently to TNF-α compared to adult CMs. Moreover, in the human heart, CMs are in interaction with non-cardiomyocytes, including fibroblasts (FBs) and endothelial cells (ECs). These cell–cell interactions likely influence CM responses to TNF-α, yet they are often not considered in traditional in vitro models.

This study aims to address that gap by investigating the effect of TNF-α in tri-culture engineered heart tissues (EHTs) and human living myocardial slices (LMS). Briefly, hiPSC-CMs, cardiac FBs, and human cardiac microvascular endothelial cells (hCMVECs) were used to generate EHTs. These were treated with 100 ng/mL TNF-α for one week. EHTs were electrically paced at 1 Hz, Ca²⁺ kinetics were measured using live Ca²⁺ imaging, and contractility was assessed via video-based motion analysis. For statistical analysis, data were tested for normality, and an unpaired t-test was applied.

Human donor hearts were provided by the NHS Blood and Transplant INOAR program (IRAS project ID: 189069), with approval from the NHS Health Research Authority and in compliance with the Governance Arrangements for Research Ethics Committees. LMS were generated from the left ventricle of donor human hearts (dimensions: 8 mm × 8 mm × 0.3 mm). Slices were cultured in chambers—one as a control, and the other treated with TNF-α and electrically stimulated during culture (1 Hz, 10 V, 10 ms pulse duration). After two days, contractile function was assessed. To determine whether TNF-α induces arrhythmogenic activity, LMS were paced at 2 Hz for 30 seconds, followed by a 2-minute pause to monitor for spontaneous contractions.

Our data show that TNF-α–treated EHTs exhibited a significantly higher spontaneous beating rate and reduced contraction amplitude (n=8-10/ p<0.05), although no changes in Ca²⁺ handling were observed. Additional replicates are required to confirm these findings. Consistently, TNF-α–treated LMS showed a significant reduction in active force, increase in passive force and slices developed arrhythmic events (n= 10-12/ p<0.001).

In summary, TNF-α impairs contractile function and induces arrhythmias in both EHT and LMS. These findings underscore the pathological role of TNF-α in human cardiac tissue and emphasize the importance of physiologically relevant platforms to study inflammation-driven heart failure. Further investigation is needed to elucidate the underlying signaling mechanisms and explore potential receptor-specific therapeutic strategies.