Profiling extracellular matrix-driven heterogeneity of single cell migration and morphology.

Cell migration and morphology are fundamental to various biological processes, including tissue development, wound healing, and cancer progression. In this study, we investigated the migration and morphological characteristics of HeLa cells cultured on different extracellular matrix (ECM) components-laminin (Ln), collagen 1 (Col1), and collagen 3 (Col3)-using machine learning-based image analysis. Cells on Ln exhibited significantly higher motility, increased turning angles, and reduced directional persistence compared to those on Col1 and Col3. Morphologically, these cells showed larger areas and higher solidity, suggesting a distinct attachment pattern compared to those on collagens. In addition, principal component analysis (PCA) effectively classified cells based on motility parameters, with cells on Ln forming a clearly separate cluster in PCA space. Further biological assays demonstrated that cells on Ln exhibited increased cell-cell interactions and smaller but more numerous focal adhesions, indicating that Ln promotes enhanced migratory plasticity and intercellular interactions. These biological factors were identified as key contributors to the distinctions observed in PCA classification. This study highlights the differential effects of ECM components on cellular behavior and demonstrates the value of integrating machine learning-based quantitative analysis with biological interpretation to advance our understanding of cell-microenvironment interactions.