The visualization and analysis of cell polarity based on dynamic fluorescence images

Abstract

Biosensors fused with fluorescent proteins can now track the dynamics of proteins and protein-protein interactions in living cells. The live-cell imaging with these biosensors enables monitoring cellular processes in real-time, opening possibilities to reveal hitherto hidden information about cellular processes if processing the images effectively. This thesis develops methods to analyze dynamic fluorescence images for visualizing the cell polarity dynamics upon shear stress by pipelining the analysis to the preprocessing and the formal polarity analysis. The preprocessing section utilizes a two-step segmentation combining Otsu’s method with a high-pass filter to rapidly obtain refined cell edge. It proposes a new algorithm based on k-means clustering to identify focal adhesions stably, which extracts the potential polarity regions accurately and reliably and reduces the defects due to human factors in the earlier methods. The polarity analysis section applies to different forms of polarity. It quantifies the analysis by normalizing the cell with zones to analyze the polarity distributed in the whole cell and defining upstream/downstream based on the incident angle between the cell edge and the external forcing to detect the polarity occurring on subcellular structures, enlarging the application range of cell polarity analysis. The analysis results show a pathway to sense and transduce the extracellular mechanical signals to intracellular force and transmit the force to the remote sites through membrane fluidity and cytoskeleton to activate signaling proteins and establish cell polarity. In summary, the thesis considers polarity analysis based on dynamic fluorescence images as an independent issue for the first time, establishes an integrated pipeline to obtain polarity information at both cellular and subcellular scales, improves the accuracy and reliability of polarity analysis, and further provides an efficient tool to analyze different forms of polarity. It promotes the application of fluorescent proteins related techniques in exploring cellular processes, offers new insights to explain the establishment of cell polarity, and further advances how mechanotransduction occurs in cells.

Keywords: fluorescent protein, live-cell images, cell polarity, mechanotransduction, image processing