Mesoscopic structural analysis via deep learning processing, with a special reference to in vitro alteration in collagen fibre induced by a gap junction inhibitor

Abstract

Abstract Dense connective tissue, including the ligament, tendon, fascia and cornea, is formed by regularly arranged collagen fibres synthesized by fibroblasts (Fbs). The mechanism by which fibre orientation is determined remains unclear. Periodontal ligament Fbs consistently communicate with their surroundings via gap junctions (GJs), leading to the formation of a wide cellular network. A method to culture Fb–synthesized collagen fibres was previously reported by Schafer et al. (‘Ascorbic acid deficiency in cultured human fibroblasts’. J. Cell Biol. 34: 83–95, 1967). This method has been applied to investigate the ability and activity of Fb collagen synthesis/phagocytosis using conventional electron microscopy (EM). However, the three-dimensional mesoscopic architecture of collagen fibres and the influence of GJ inhibitors on collagen fibre formation in vitro are poorly understood. In this study, three-dimensional mesoscopic analysis was used to elucidate the mechanism of directional fibre formation. We investigated the influence of GJ inhibitors on collagen formation driven by periodontal ligament Fbs in vitro, histomorphometrically, and the structural properties of in vitro collagen fibre on a mesoscale quantitatively, using correlative light and EM optimized for picrosirius red staining and focused ion beam-scanning EM tomography. Our results indicate that under culture conditions, in the presence of a GJ inhibitor, the orientation of collagen fibres becomes more disordered than that in the control group. This suggests that the GJ might be involved in determining fibre orientation during collagen fibre formation. Elucidation of this mechanism may help develop novel treatment strategies for connective tissue orientation disorders. Graphical Abstract