Background
Cardiac fibrosis is characterized by excessive accumulation of extracellular matrix (ECM) proteins by activated (myo-)fibroblasts, leading to myocardial stiffness and decline in electromechanical function. Developing an ex-vivo model of fibrosis will enhance our understanding of pathological remodeling and provide a platform for therapeutic investigations. Here, we established such model by exposing beating human and rabbit cardiac slices to inflammatory cytokines (TGFβ, TNFα, IL-1β) known to be upregulated during cardiac remodeling.
Methods
Human cardiac samples were obtained from explanted hearts of patients undergoing heart transplantation or surgery after written informed consent, and from healthy hearts donated within the USA. Rabbit hearts were obtained after lethal sodium pentobarbital injection in deep narcosis. An ex-vivo wound assay was implemented while preparing left-ventricular cardiac slices using a 1 mm biopsy punch needle to create a circular lesion in the center (0.79mm
2), simulating tissue loss and necrotic clearance. Lesion healing was quantified by measuring hole area after cultivation via confocal microscopy (human: n=4/2, rabbit: n=3-8/4; technical/biological replicates). Contractility parameters (contraction force, CF, contraction duration, CD90, time-to-peak, TTP90, time-to-relaxation, TTR90) were determined in human (n=6/2) and rabbit (n=10-23/6) cardiac slices cultured under permanent electrical stimulation (0.5 Hz), while treated with TGFβ, TGFβ+TNFα and TGFβ+TNFα+IL1-β (each 30 ng/ml). Profibrotic markers (α-smooth muscle actin, αSMA, vimentin) and ECM markers (periostin, fibronectin, wheat germ agglutination, WGA) were analyzed with confocal microscopy after 7d (rabbit: 4-10/4) and 18d (human: 4/2) of cultivation.
Results
CF decreased by 38%±50% in the TGFβ group already after 4d of treatment compared to control in rabbit cardiac slices (p<0.05). Furthermore, TGFβ prolonged myocardial relaxation parameters, such as CD90 (1.2-fold; p<0.05) and TTR90 (1.4-fold; p<0.05), while no significant changes were observed in controls. Parameters of contraction velocity (TTP90) remained unchanged. No additional changes were found in rabbit cardiac slices, when TNFα or IL1-β were added. In human cardiac slices, TGFβ treatment also caused a decrease in CF (79%±18%; p<0.05), as well as an increase in CD90 (1.8-fold), TTP90 (2.1-fold) and TTR90 (1.7-hold) after 14d of treatment. These effects were observed in both healthy and failing human myocardium. Fibrotic and ECM markers αSMA (1.5-fold), vimentin (1.1-fold), periostin (4.0-fold), fibronectin (1.4-fold), WGA (1.2-fold) were upregulated in rabbit cardiac slices treated with TGFβ vs control (p<0.05). Cardiac slices undergoing TGFβ treatment showed a smaller hole area (rabbit: 0.32±0.46mm
2, human: 0.34±0.39mm
2) than control (rabbit: 1.48±1.19mm
2, human: 1.54±0.43mm
2, p<0.05).
Conclusion
Exposing beating human and rabbit cardiac slices in long-term culture to pro-inflammatory cytokines, such as TGFβ, induces fibroblast-myofibroblast differentiation and ECM remodeling paralleled by a decline in functional parameters (CF, CD90, TTP90, TTR90). This suggests the feasibility of an ex-vivo slice model that reproduces relevant effects of fibrotic remodeling, including elevated myocardial stiffness and dysfunction
