Phenotypic screening for assessment of risks from environmental chemicals

Alterations in the appearance – morphology – of cells and the distribution of proteins can provide important information on the development of diseases such as cancer and support the identification of potentially harmful environmental chemicals. To detect such phenotypic changes, different staining strategies are performed to visualize cell morphologies, followed by imaging and automated image analysis. During phenotypic screening, thousands of features describing cellular morphologies are extracted from images and further aggregated into phenotypic profiles, which eventually enable the prediction of potential human health risks from environmental chemicals. This process is illustrated in more detail in our External Link:project video.

In the Cell Painting assay, seven fluorescent dyes are utilized to simultaneously stain multiple cellular compartments including the actin cytoskeleton, mitochondria, lysosomes, cell membranes, nuclei, the endoplasmic reticulum, and golgi. From these multi-label images, thousands of morphological features including intensity, size and shape, texture, granularity, and intensity distribution are extracted using the CellProfiler image analysis software. These morphological features are further analyzed using the KNIME data analysis software for clustering and benchmark concentration modelling. This way, chemicals that lead to similar morphological profiles are clustered into the same activity groups.

The 4i (iterative indirect immunofluorescence imaging) method facilitates visualization of more than 20 different proteins in the same sample. It comprises multiple iterations of staining, elution (signal removal) and re-staining steps. Upon imaging and image data analysis, which is similar as described above for Cell Painting, the distribution of proteins labelling organelles, cytoskeletal structures, and cell proliferation markers are projected into a multiplexed protein map. Different combinations of antibodies used result in specific protein maps and can eventually be applied to identify and predict harmful chemical effects.

The described phenotypic screening-methods are applied in the “RISK-HUNT3R” and "MORPHEUS" projects.