Quantitative analysis of angiogenesis in microfluidic devices

• Main Author: Wang, Mengmeng
• Other Authors: Justin Dauwels
• Format: Theses and Dissertations
• Language: English
• Published: 2017
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/72455
• Institution: Nanyang Technological University

Angiogenesis is the growth of new blood vessels from existing vessels. It is a critical process in fetal development, wound healing, and growth and development. However, it also plays an important role in cancer growth and spread, which is one of the leading causes of deaths worldwide. Better understanding of angiogenic mechanism is required to develop effective antiangiogenic therapies for cancer treatment. During angiogenesis, endothelial cells (ECs) migrate in a coordinated manner by specializing into two distinct phenotypes: tip cells and stalk cells. Tip cells extend dynamic filopodia to sense and respond to angiogenic stimuli, and stalk cells trail behind tip cells to form solid vessels. Both cell phenotypes can dynamically switch positions and functions during the sprouting process. The inter-transition of the cell phenotypes plays an important role for ECs sprouting out from monolayer, extending and creating new branches, as well as reconnecting in a later stage. Automated image analysis tools to analyze experimental microscopy images are useful for a wide range of biomedical investigations, since they can provide quantitative understanding of the biological processes for biologists and generate numerical data to build computational models. In this thesis, we focus on ECs migration and angiogenic vascular formation for a better understanding of angiogenesis. Therefore, our objectives include culturing angiogenic vessels with branching in 3D in vitro environments, developing automated image analysis systems to track multiple ECs migration and to track the angiogenic vessel formation. In the following, we will discuss each objective in detail. Firstly, we conduct angiogenic experiments in the 3D microfluidic devices (MFDs), which mimic the in vivo system. By providing tumor angiogenic growth factors (TAFs) such as Vascular Endothelial Growth Factor (VEGF) and Sphingosine-1-phosphate (S1P) in the 3D MFDs, angiogenic vessels with branching are formed. The sprouting processes are observed with a phase contrast microscope daily and with a confocal microscope at the end time. To investigate the influence of S1P on angiogenic vessel morphologies, we consider three different S1P conditions. The experimental observations suggest that the positive S1P gradient increases the average length of the angiogenic vessels. Secondly, we develop an automated multi-cell tracking system (AMCTS) to track the migrating ECs within the angiogenic vessels from the time-lapse phase contrast images. The proposed system consists of preprocessing to eliminate the non-gel region and obtain binary angiogenic vessel shapes, cell detection to _nd and label the EC candidates, and multiple hypothesis Kalman _ltering to associate and track the detected EC candidates over the image sequences. Biological knowledge is incorporated when estimating the track probability during cell association. Numerical results indicate that the proposed system is able to track the cell migration trajectories accurately. Cell lineage plots, showing the history of the cell proliferation and cell migration into the gel, with timestamps of when it is in-focus and out-of-focus, are also generated automatically from this system. Thirdly, we propose an automated vessel formation tracking system (AVFTS) to track the vessel formation and extract quantitative vessel information from the timelapse phase contrast images. The proposed system consists of preprocessing, skeletonization, and branch tracking. After obtaining the binary vessel shapes through preprocessing, we apply a distance transform (DT) and an augmented fast marching method (AFMM) to extract the vessel skeletons, which are a group of segments connected by nodes. Next, we train a support vector machine (SVM) classifier to distinguish and remove the specious segments which are not part of the vessels. Lastly, we identify the route for each branch by joining the linked skeleton segments and associate the branches over image sequences by the Hungarian method. This system is applicable to distinguish tip and stalk cells based on their relative cell positions within a branch. Moreover, it also helps biologists to investigate the influence of different angiogenic factors by automatically extracting quantitative vessel information including vessel length, vessel width, and the number of branches from experimental microscopy images. Specifically, we quantify the experimental results under different S1P conditions and find that the positive S1P gradient increases ECs migration and vessel elongation, and leads to higher possibilities for branching to occur. In conclusion, we culture angiogenic vessels over long period of time in 3D MFDs and develop two automated image analysis systems to quantitatively analyze these angiogenic vessels. These systems would bene_t biologists by providing quantitative and accurate comparison of the inuence of di_erent growth factors on angiogenic vessel morphologies. They also provide numerical data to build computational models for angiogenesis prediction.