Studying the Mechanobiology of Aortic Endothelial Cells Under Cyclic Stretch Using a Modular 3D Printed System

Here, we describe a motorized cam-driven system for the cyclic stretch of aortic endothelial cells. Our modular design allows for generating customized spatiotemporal stretch profiles by varying the profile and size of 3D printed cam and follower elements. The system is controllable, compact, inexpensive, and amenable for parallelization and long-term experiments. Experiments using human aortic endothelial cells show significant changes in the cytoskeletal structure and morphology of cells following exposure to 5 and 10% cyclic stretch over 9 and 16 h. The system provides upportunities for exploring the complex molecular and cellular processes governing the response of mechanosensitive cells under cyclic stretch.


Supplementary Information 3: Characterisation of the cyclic stretch system
The versatility of our cyclic stretch system enabled us to generate tailored spatiotemporal strain/stress profiles by changing the cam profile, cam size, and follower profile ( Figure   S3a). To showcase this versatility, we designed three cam profiles to generate cyclic trapezoidal, sine, and double-peak displacement profiles ( Figure S3b). The cams were designed in Matlab using cycloidal motion cam dynamics (Charles E. Wilson, 2003;Norton, 2005), as detailed in Supplementary Information 4.
The dimensions of the cams were varied to generate relative vertical displacements (defined as the ratio of axial displacement to the diameter of the cell culture chamber, which is referred to as cyclic stretch) of 5%, 10%, 15%, 20%, and 25% ( Figure S3b). We also designed three followers with conic, rounded, and flat profiles ( Figure S3c).
Numerical simulations were performed using ANSYS static structural module to predict the strain/stress profile of the membrane placed at the bottom surface of the cell culture chamber under various cam/follower combinations. In the first set of simulations, we investigated the dynamic variations of strain/stress at the middle section of the side well membrane ( Figure   S3a). The membrane was displaced using cams capable of generating dynamic trapezoidal, sine, and double-peak displacement profiles when coupled to a flat follower. Our simulations indeed show the ability to generate customized spatiotemporal strain/stress profiles by varying the profile and size of the cam ( Figure S3d).
In the second set of simulations, we investigated the maximum strain/stress induced at the middle section of the side well membrane. The membrane was displaced by coupling cams capable of generating dynamic trapezoidal displacement profiles to conic, rounded and flat followers ( Figure S3e).

Supplementary Information 4: Detailed cam design
The cams were designed using standard design techniques, in which a displacement motion is generated by a series of standardized equations. The cams comprised of a rising segment, where the displacement increases, and a falling segment, where the displacement decreases.
Additional elements were added to the cams to customize them to our needs. These elements were dwells, which keep the displacement constant at the desired segments. The combination of rising, falling and dwell segments allowed us to generate customized motion profiles with varying magnitudes:

A: Trapezoidal displacement design
To design the trapezoidal cam, a combination of standard cam cycloidal displacements and dwells were used where is the maximum displacement (stroke) and is the cam angle of rotation.

B: Sinusoidal displacement design
To design the sinusoidal cam, a standard harmonic displacement was used, where is the maximum displacement (stroke) and is the cam angle of rotation. Figure S5. Sinusoidal displacement design: A rising harmonic motion was used within 0°-180°, followed by a falling harmonic motion within 180°-360°.

C: Double-peak displacement design
To design the double-peak cam, two sets of harmonic displacements were used. On the first half, the maximum displacement was set to while in the second half, the maximum displacement was halved. This arrangement generated two peaks with different magnitudes during each revolution, where is the cam angle of rotation. The histograms evaluated deviation from a set of angles ranging from -89° to 90° relative to the horizontal using a small subregion (20×20 pixels), which was centred around the pixel of interest. The Statistical significance was assessed with a global Watson's U2 test, and statistics were computed using the circular statistic toolbox. For statistical analysis, one-way ANOVA was performed using Prism 8 (GraphPad software), and P < 0.05 was considered significant.
The analysis of nuclear area and circularity was determined by an automated image processing algorithm written in MATLAB. This algorithm uses contrast filters to create a color difference between the nucleus and the structure of the cell. Next, it binarizes the image to separate, label, and measure the number of pixels and perimeter in each nucleus. The number of pixels obtained per cell is used to calculate the area. The circularity is obtained as follows, in which is the cell area and is the cell perimeter: In which a circularity closer to 1 is rounder. The Statistical significance was assessed with a global Watson's U2 test, and statistics were computed using the circular statistic toolbox. For statistical analysis, one-way ANOVA was performed using Prism 8 (GraphPad software), and P < 0.05 was considered significant.