Related Experiment Video
Updated: Apr 15, 2026
09:48
Discrimination and Characterization of Heterocellular Populations Using Quantitative Imaging Techniques
Published on: June 30, 2017
8.0K
High-Content Imaging and Machine Learning Classify Phenotypical Change in Coronary Artery Endothelial Cells Caused by
Lavinia Eugenia Ferariu1, Gheorghe Movileanu1, Giulia Gaggi2,3,4
1Department of Automatic Control and Applied Informatics, Gheorghe Asachi Technical University of Iasi, 27 Mageron, 700050 Iasi, Romania.
International Journal of Molecular Sciences
|April 14, 2026
Summary
Bisphenol S (BPS) causes subtle changes in human coronary artery endothelial cells (HCAEC) even at low doses. This study reveals early signs of vascular dysfunction using advanced imaging and machine learning for toxicity assessment.
Area of Science:
- Environmental toxicology
- Cellular and molecular biology
- Biomedical engineering
Background:
- Bisphenol S (BPS) is a common substitute for Bisphenol A.
- Emerging evidence indicates BPS possesses similar endocrine and cardiovascular toxicity concerns.
- Endothelial cells are crucial for vascular health, and their dysfunction is an early indicator of cardiovascular disease.
Purpose of the Study:
- To investigate the subtle phenotypic alterations induced by prolonged low-dose Bisphenol S (BPS) exposure in human coronary artery endothelial cells (HCAEC).
- To establish a sensitive and scalable framework for assessing vascular toxicity of environmental contaminants using high-content imaging and machine learning.
- To identify specific cellular features indicative of early endothelial dysfunction caused by BPS.
Main Methods:
- Human coronary artery endothelial cells (HCAEC) were exposed to 0.1 µM BPS for 96 hours.
- A cell painting assay and high-content microscopy were employed to profile cellular morphology and features.
- Machine learning models, including XGBoost with ReliefF feature selection, were used to classify phenotypic changes from high-dimensional image data.
Main Results:
- BPS exposure resulted in a distinct endothelial cell phenotype, detectable through quantitative image analysis.
- The most significant changes were observed in mitochondrial organization and nuclear chromatin features.
- The XGBoost classifier, utilizing ReliefF-selected features, demonstrated robust performance in identifying BPS-induced alterations.
Conclusions:
- Chronic low-dose BPS exposure induces a specific endothelial phenotype associated with early-stage endothelial dysfunction.
- The integration of high-content imaging and machine learning offers a powerful approach for sensitive vascular toxicity screening.
- This framework can aid in understanding the risks posed by environmental contaminants like BPS to cardiovascular health.
