A new method to estimate 3D cell parameters from 2D microscopy images

Abstract

Optical microscopy has been a basic and standard technique in cell biology research for decades. Microscopy techniques function well for thin, optically transparent cultures and allow for the imaging of thicker biological specimens. There is no better method of in vitro cell observation and analysis, hence microscopic techniques are extensively used and constitute an optimal tool for cell culture studies. This paper proposes an original methodology of optical microscopy data processing based on the phase contrast technique during cell culture monitoring. By exploiting images recorded during cell proliferation, a surface reconstruction was performed based on assumption, it can be considered that the local brightness of the image depends on the cells' thickness and thus the obtained results can be interpreted in the form of a surface that represents a three-dimensional structure, which allowed for a quantitative description of the cell evolution. The 3D data obtained enabled the investigation of parameters describing the morphology of the cells and the topology of their proliferation. These parameters included cell sizes in plane but also in the direction perpendicular to it, cell volume changes, their spatial distribution, as well as anisotropy and directivity. The method presented provides data carrying information similar to that obtained using a holographic microscope, e.g. A HoloMonitor (Phase Holographic Imaging PHI Inc.), or from confocal scanning microscopy with the “z-stack” mode. The techniques of bright field or phase contrast cell observation are, however, much cheaper, and widely available when compared to holographic microscopy, for instance. Besides, these also enable monitoring of cell activity over time, i.e. the study and quantitative description of dynamic changes in the cells. The proposed approach uses generally available free tools such as ImageJ software with BoneJ and Particle Analyzer plugins. The methodology is suitable for even a basic microscope, it can be easily implemented as a script, and thus data processing can be significantly shortened, the methodology can be automated, and also applied for data processing in real time.