Clin Res Cardiol (2022).

Nano-Structural Basis for Mechanical Fibroblast-Cardiomyocyte Cross-Talk
S. Schmid1, A. Felekary2, J. Madl1, P. Kohl1, A. Rohrbach2, E. Rog-Zielinska1
1Institut für Experimentelle Kardiovaskuläre Medizin, Universitäts-Herzzentrum Freiburg - Bad Krozingen GmbH, Freiburg im Breisgau; 2Institut für Mikrosystemtechnik, Universität Freiburg, Freiburg;

Following injury, the dynamic communication between cardiomyocytes (CM) and fibroblasts (FB) drives generation of intricate extracellular matrix (ECM), largely composed of collagen fibres. The ECM scaffold ensures structural and mechanical integrity of the injured tissue. Excessive collagen deposition can however lead to a mechanical and electrical impairment of heart function. Knowledge of the mechanisms governing collagen production is therefore of great importance in basic and clinical research. Here we propose that CM and FB communicate with each other directly via thin, FB-borne membrane protrusions - tunnelling nanotubes (TNT) – and that this interaction is important for injury-signal recognition and directed collagen deposition around CM in remodelling hearts.

Primary ventricular FB from human and rabbit, as well as primary CM from rabbit were used as models to study TNT structure, dynamics, and regulation in vitro. Cells were fluorescently labelled with actin- and membrane- targeting fluorescent probes under control conditions and following treatment with the actin disruptor latrunculinB (LatB, 60 nM for 2h) or the pro-fibrotic cytokine TGF-ß1 (transforming growth factor, 10 ng/mL TGF-ß1 for 24h). Confocal and reflection-mode microscopy were used to analyse TNT number, length, actin content, and minute-scale dynamics. Millisecond-scale TNT fluctuations were characterised using rotating coherent scattering microscopy. After fixation by high-pressure freezing, ultrastructure of TNT was examined by
electron tomography. Collagen deposition was visualised using fluorescent CNA-35 collagen probe.

Under control conditions, cardiac FB contained 3.4
±0.4 (SEM) TNT per 10 µm membrane length. These were either ‘free-floating’ or anchored at neighbouring cells. TNT were up to 8 µm long (average length 5.5±0.2 µm) and majority of them (78±3.2%) contained F-actin. TNT in cells treated with LatB were less likely to contain actin (36±7.5%, p<0.001), and were longer (8.1±0.6 µm, p<0.0001) compared to control. Treatment with TGF-ß1 led to a decrease in the number of TNT (2.1±0.3 µm, p<0.05) and an increase in TNT fluctuation width (control 0.29±0.08 µm vs TGF-ß1 0.35±0.01 µm, p<0.05). TNT were frequently structurally associated with collagen strands as visualised by both confocal and electron microscopy.

Our study describes a mode of hetero-cellular communication in the heart with potential relevance for pathological fibrotic remodelling. FB-borne TNT were found to be highly dynamic in vitro, and their properties can be modulated using pharmacological agents such as TGF-ß1 and LatB. Future studies will focus on characterisation of TNT-based coupling dynamics, contributions of TNT to collagen deposition, heterocellular signal exchange, and novel ways of pharmacological, genetic, or environmental manipulation of TNT structure and function.