Elucidating the role of survivin on stiffness-dependent VSMC migration using a novel computational pipeline

Introduction: Arterial stiffness is a key contributor to various cardiovascular diseases, including atherosclerosis and coronary artery disease, as it is closely associated with the aberrant migration of vascular smooth muscle cells (VSMCs). Despite this correlation, the molecular mechanisms driving VSMC migration in stiff environments remain unknown. Our lab recently showed that survivin, an inhibitor of apoptosis protein, is highly expressed in human VSMCs cultured on stiff hydrogels and therefore, we hypothesized that survivin may play a crucial role in modulating stiffness-induced VSMC migration. In this study, we investigated how survivin regulates stiffness-dependent VSMC motility using time-lapse video microscopy. Importantly, we developed a novel computational method that allows for the quantification of lamellipodial protrusion, which is a key step in migration. Traditionally, protrusions are measured manually, which is time consuming and prone to bias. Several software, including ADAPT and KymoButler, have attempted to circumvent this problem; however, no method has been developed that accurately detects lamellipodial protrusion dynamics of cells seeded on hydrogels, typically resulting in lower-quality kymograph images compared to cells cultured on glass or plastic dishes. In this study, we elucidated survivin’s role in stiffness-dependent VSMC migration using a computational model we developed to measure lamellipodial protrusion dynamics.

Materials and

Methods: Soft (5.60±2.45kPa) and stiff (16-21.05±3.07kPa) hydrogel fibronectin coated hydrogels were prepared to mimic healthy and diseased arteries, respectively. Human VSMCs were plated on soft and stiff hydrogels overnight; followed by survivin inhibition with varying concentrations of YM155 (survivin inhibitor) or DMSO (vehicle control). Subsequently, they were placed in a Cytation 1 imager for time-lapse imaging for 3 hours at 3-minute intervals. Fiji was then used to determine the velocity from the (x,y) coordinates of the nucleus over time. To detect lamellipodial protrusions, VSMCs were seeded on glass or hydrogels with 2µM YM155 or DMSO and imaged on a DMI8 fluorescent microscope using 20X phase contrast for 20 minutes. Subsequently, kymograph images were generated using Fiji. Using the kymograph images, manual ground truth annotations of protrusions from cells on glass, due to enhanced clarity, were conducted using MATLAB, resulting in the generation of binary masks. The raw kymograph images and resulting binary masks were fed into a U-Net model to train it to automatically segment the raw images and create binary masks. Boundary pixels of protrusion segmentations were then extracted using bwperim, converted to coordinates, interpolated, smoothed, and graphed. MATLAB was then used to identify protrusion peaks and their edges, as well as to measure peak width and distance. From the metrics of each peak, protrusion dynamics was computed, including peak frequency, protrusion distance, protrusion velocity, and retraction velocity (Figure 1). This model was then used to quantify the kymograph images obtained from cells on soft and stiff hydrogels.

Results, Conclusions, and Discussions: Our data showed that VSMCs seeded on soft hydrogels exhibited a decrease in velocity compared to those on stiff hydrogels. Inhibiting survivin in VSMCs cultured on stiff hydrogels resulted in velocity similar to that observed in soft hydrogels, implicating that survivin is crucial for VSMC migration in stiff conditions (Figure 2). Lamellipodial protrusions detected via the computational model under stiff conditions showed an increased number of peaks, protrusion distance, and velocity compared to those observed under soft conditions. Interestingly, survivin inhibition via YM155 significantly reduced the number of peaks, protrusion distance, and velocity, phenocopying the patterns observed under soft conditions (Figure 3). A significantly reduced number of peaks and their velocities when inhibiting survivin would disrupt cell migration by decreasing lamellipodial protrusion, which forms at the leading edge of the migrating cell. This indicates that survivin regulates lamellipodial protrusion activity, influencing VSMC migration in stiff conditions. However, the detailed molecular mechanisms by which survivin regulates stiffness-stimulated cell migration remain unknown and warrant further investigation. In addition to lamellipodial protrusion dynamics, morphology has been implicated to play a role in migration and we will utilize a similar method to segment the cells and identify features descriptive of cell morphology. In conclusion, our novel computational method for quickly and precisely detecting and quantifying lamellipodial protrusion activities in VSMCs on hydrogels was pivotal in elucidating the role of survivin in cell migration. However, further training of the model using VSMCs on hydrogels is needed to increase the accuracy of the model. This method holds promise not only for VSMCs on hydrogels but also for other cell types. We plan to share the code on Github for free access by other researchers upon publication.

Acknowledgements (Optional):