Structure and dynamics of human cardiac fibroblast membrane nanotubes

Introduction: The dynamic communication between fibroblasts (FB) and their environment is important for supporting the continuous cardiac tissue function. Here, we propose that the interaction between FB, other cardiac cells, and the extracellular matrix (ECM) is mediated by thin, dynamic FB-borne membrane nano-tubes (MNT), and that these FB-MNT might play a role during pathological cardiac remodelling. 

 

Methods: Immortalised human atrial FB cultures, as well as primary human atrial FB derived from patients in sinus rhythm (SR) or permanent atrial fibrillation (AF), were used. FB-MNT were studied using a combination of confocal fluorescence, reflection , and electron microscopy. The impact of actin disruption, pro-fibrotic signalling, hypoxia, blockade of integrin-β1 binding to collagen, changes in ECM and collagen density, mechano-sensitive channel (TREK-1) overexpression, as well as mechano-sensitive channel (Piezo1) expression silencing on MNT structure and dynamics were tested, and the interactions between FB-MNT and ECM, and between FB-MNT and human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CM) were probed. Additionally, presence and structure of FB-MNT and ECM were investigated in intact human atrial tissue using confocal and 3D electron microscopy.

 

Results: Cardiac FB-MNT were abundant in vitro. They contained actin fibres, vesicles, and membrane organelles, such as mitochondria. FB-MNT tip ‘scanned’ their environment, and the velocity of tip movement was lower in AF FB-MNT when compared to SR tissue-derived FB. Actin disruption decreased FB-MNT number, length, and velocity. Hypoxia increased the number and length of FB-MNT, as well as their velocity. A pro-fibrotic environment, as well as higher extracellular collagen density, increased FB-MNT tip velocity. Blocking of integrin binding to collagen decreased the number of FB-MNT, but increased their dynamic tip movement. Alterations in mechano-sensitive channel expression did not affect the structure or movement of FB-MNT. FB-MNT in vitro were seen to dynamically interact with fibrillar collagen deposits (e.g. through pulling). In a co-culture with hiPSC-CM, FB-MNT were seen to directly interact with hiPSC-CM, and presence of hiPSC-CM altered the dynamic behaviour of FB-MNT. A structural association of FB-MNT with collagen fibres and cardiomyocytes was observed in intact cardiac tissue. 

In addition, we noted frequent presence of ECM (collagen types I, III, IV, and VI) inside cardiomyocyte surface membrane invaginations (transverse−axial tubular system, TATS) in diseased human tissue; this often coincided with the proximity of FB-MNT to TATS entrances. The presence of ECM was associated with widening of individual TATS elements. 

 

Discussion: We provide a comprehensive characterisation of the structure, dynamics, and regulation of human cardiac FB-MNT. Furthermore, we provide evidence that supports a role of FB-MNT in shaping the cardiac ECM scaffold. Under pathological conditions, FB-MNT activity may play an important role in the process of fibrotic remodelling, and future studies will further explore the impact of FB-MNT activity on cardiac cell-cell communication and ECM in the heart. Our studies additionally indicate that ECM components in fibrotic hearts not only surround cardiomyocytes, but are deposited inside TATS. The exact consequences of extracellular matrix deposition in TATS for cardiomyocyte function are subject of ongoing studies.