Morphological Responses of Vascular Endothelial Cells Induced by Local Stretch Transmitted Through Intercellular Junctions

Abstract

It has been well established that mechanical stimuli including fluid shear stress and cyclic stretch play a key role in endothelial cell (EC) remodeling. However, in contrast to global remodeling to these mechanical stimuli, little is known of how local mechanical forces are transmitted through cells to induce cell remodeling leading to alteration in cell functions. In this study, we demonstrated that EC remodeling can be exerted by local tension generated in a neighboring EC. In this technique, a glass microneedle was used to apply local stretch in an EC in confluent monolayer and the resulting tension is transmitted to a neighboring EC across intercellular junctions. Local stretch induced reorientation and elongation of ECs parallel to the direction of stretch associated with reorganization of stress fibers. In addition, recruitment of Src homology 2-containing tyrosine phosphatase-2, binding to intercellular adhesion molecules platelet-endothelial cellular adhesion molecules-1, was selectively observed at the force-transmitted intercellular junctions after application of local stretch. These findings suggest that intercellular junctions can not only transmit but also sense local forces, and are potentially involved in EC mechanotransduction pathways.

Appendix. Calculation of SF parameters with image processing

The protocol for image processing to determine SF parameters is shown in Fig. 8. Algorithm based on pixel intensity of fluorescence images of GFP-actin was adapted from previous reports [23, 24]. SFs were extracted from original image [Fig. 8(a)] with convolution filter application, as shown in Fig. 8(b). For giving pixels p _{i ,j} in 8-bit gray-scale, the pixels surrounding p _{i ,j} were selected as a sample matrix [equation (1)]

$${\text{Sample matrix (}}S_{i{\text{,}}j} {\text{) = }}\left( {\begin{array}{*{20}c}{p_{i - 2,j - 2} } & {p_{i - 2,j - 1} } & {p_{i - 2,j} } & {p_{i - 2,j + 1} } & {p_{i - 2,j + 2} } \\{p_{i - 1,j - 2} } & {p_{i - 1,j - 1} } & {p_{i - 1,j} } & {p_{i - 1,j + 1} } & {p_{i - 1,j + 2} } \\{p_{i,j - 2} } & {p_{i,j - 1} } & {p_{i,j} } & {p_{i,j + 1} } & {p_{i,j + 2} } \\{p_{i + 1,j - 2} } & {p_{i + 1,j - 1} } & {p_{i + 1,j} } & {p_{i + 1,j + 1} } & {p_{i + 1,j + 2} } \\{p_{i + 2,j - 2} } & {p_{i + 2,j - 1} } & {p_{i + 2,j} } & {p_{i + 2,j + 1} } & {p_{i + 2,j + 2} } \\\end{array} } \right)$$

(1)

Fig. 8

Protocol for determination of SF parameters based on image processing. (a) Manually extracted single cell image, (b) automatically detected SFs with convolution filters, (c) calculated minimal gradient in pixel intensity with Sobel filter, (d) angle of SF orientation in each pixel

Convolution using horizontal and vertical kernels [equations (2) and (3)], which define weighted sum of neighboring pixels, provided magnitude of brightness gradient of SFs in each pixel [equation (4)].

$${\text{Vertical kernel (}}K_{\text{v}} {\text{) = }}\left( {\begin{array}{*{20}c}{ - {\text{1}}}&{\text{0}}&{\text{2}}&{\text{0}}&{ - {\text{1}}} \\{ - {\text{2}}}&{\text{0}}&{\text{4}}&{\text{0}}&{ - {\text{2}}} \\{ - {\text{4}}}&{\text{0}}&{\text{8}}&{\text{0}}&{ - {\text{4}}} \\{ - {\text{2}}}&{\text{0}}&{\text{4}}&{\text{0}}&{ - {\text{2}}} \\{ - {\text{1}}}&{\text{0}}&{\text{2}}&{\text{0}}&{ - {\text{1}}} \\\end{array} } \right)$$

(2)

$${\text{Horizonal kernel (}}K_{\text{h}} {\text{) = }}K_v^T $$

(3)

$$G = S_{i,j} * K_{\text{h}} + S_{i,j} * K_{\text{v}} $$

(4)

The pixel whose G value was bigger than 5-fold of averaged pixel intensity in whole area of the target cell was assigned as a constituent of SFs. In order to calculate the angle of SF orientation, the angle of minimal gradient in pixel intensity in each pixel [Fig. 8(c)] was calculated using Sobel kernels as previously reported [20]. The angle data allocated to only pixels which have been assigned as a constituent of SFs [Fig. 8(d)]. Angle values were summed up across the cell with vectorial summation and were assumed that each pixel has a vector consisted of the calculated angle and unit length. The angle between the direction of the resultant vector and the tensile direction was defined as the angle of SF orientation. The length of the resultant vector divided by the number of pixels was defined as the uniformity index of SFs, which indicate the degree of alignment of SFs. All calculation processes were executed on Excel 2004 for Mac (Microsoft, USA). Analyses were performed on images obtained at every 5 min after application of local stretch.