Cell division tracking by live cell imaging and image processing

Abstract

Towards the goal of generating specific tissue from stem or progenitor cells for regenerative medicine, it will be necessary to understand the dynamics of stem and progenitor cell development and how environmental cues trigger cell migration, mitosis, apoptosis, and lineage fate. Observing the dynamic process in a continuous manner at the single-cell level will advance our knowledge of these processes. Long-term live cell imaging systems and computational methods to automatically identify and track progenitor cell migration and division will enable this study.

The aims of this thesis were to develop a live cell imaging system with semi-automated software for tracking adherent cell lines and to apply this system to study cardiac stem cell development. The imaging system was benchmarked by tracking NIH3T3 cells in vitro for 4 days. Cardiac stem cells were enriched by fluorescent activated cell sorting (FACS) from the interstitial fraction of the mouse heart. These cells form colonies (c-CFU-F) which were tracked by live cell imaging. Green fluorescent protein (GFP) transgenic mice were used to report cCFU-F cells that express β-actin or platelet-derived growth factor receptor-a (PDGFR-a). These initial studies have focused on characterizing cell motility and cell cycle dynamics of cCFU-F subpopulations.

The growth rate of NIH3T3 (482 cells) tracked by the live cell imaging system was similar to conventional culture methods. Lineage maps of PDGFR-α+ cells (164 maps) and β-actin + cells (352 maps) within passage 3 colonies were constructed by continuous cell tracking over a 5 day culture period. Two distinctive cell morphologies were identified; large flattened-cells with low motility and highly motile spindle-shaped cells. The probabilities of mitosis of flattened- and spindle-shaped cells were estimated for each generation using Kaplan Meier statistics. There were significant differences between the cell cycle distribution and motility for these two subpopulations. Furthermore, Cox regression analysis was used to show that cell cycle progression was related to cell size and colony size. Large flattened-cells infrequently underwent asymmetric division giving birth to a small cell and large cell with a short and long cycle time respectively. These studies have illustrated the value of lineage mapping cCFU-F, leading to a deeper understanding of cCFU-F growth dynamics.