Abstract

Phenotypic Machine Learning in High Content Imaging Screening with Organoids

Yu-Chi Chen, Scientist, Bioprinting Group, NCATS

In the last few years, there has been a fast increase in the use of 3D cellular models as physiological relevant assays for drug discovery and development.  The use of U-bottom plates has been widely used for growing 3D spheroids because they facilitate spheroid formation in a scalable and reproducible manner to enable large scale compound screening.  However, the U-bottom shape of the wells in these plates limits the use of high-resolution imaging (>20X) and high-content screening mainly due to light diffraction.  Therefore, the assays used for screening with spheroids have mostly been restricted to simple readouts such as cell viability using standard well-based assays, or high content assays measuring total fluorescence intensity.  As more complex spheroids and organoids models are developed for disease modeling, there is an increased need to be able to quantitate the effects of compounds in different cell types, sub-cellular biomarkers and phenotypes within these 3D systems.  Here we will discuss the development of a 1536-well 3D HCS assay platform that enables the generation of high-resolution sub-cellular images coupled with a Phenotypic machine learning and 3D segmentation analysis within 3D spheroids that enables the implementation of 3D High Resolution Imaging Screening.