Label-free cell tracking enables collective motion phenotyping in epithelial monolayers

Area of Science:

• Cell Biology
• Biophysics
• Computational Biology

Background:

• Collective cell migration is crucial for biological processes, including cancer metastasis.
• Understanding cell-cell interactions and relative motion is key to deciphering collective behaviors.
• Existing methods for analyzing collective cell migration can be limited in scope or require extensive manual effort.

Purpose of the Study:

• To develop an AI-based pipeline for automated segmentation and tracking of cell nuclei in phase-contrast images.
• To enable robust downstream analysis of collective cell motion without requiring new training data.
• To identify quantitative metrics indicative of metastatic potential in cancer cells.

Main Methods:

• Utilized a U-Net convolutional neural network for nuclei segmentation based on nucleus staining.
• Employed the Crocker-Grier algorithm for tracking nuclei movement.
• Applied the developed pipeline to a panel of cell lines with oncogenic mutations.

Main Results:

• The AI pipeline successfully segmented and tracked cell nuclei from phase-contrast microscopy images.
• The collective rearrangement metric, D ² _min, which quantifies non-affine motion, was analyzed.
• This metric demonstrated potential as an indicator of metastatic capacity in the studied cell lines.

Conclusions:

• The AI-driven approach provides a powerful and adaptable tool for analyzing collective cell migration from existing image datasets.
• The D ² _min metric offers a promising, quantitative measure for assessing metastatic potential.
• This work facilitates deeper insights into the mechanisms of collective cell migration and cancer progression.