Scientists from Singapore’s A*STAR Bioinformatics Institute and the Institute of Bioengineering and Nanotechnology have developed a computerized cell imaging system that can predict with up to 90% accuracy whether a chemical will be toxic to human kidney cells. The system is more accurate than animal tests, and faster than existing in vitro methods for predicting kidney toxicity.

To create the system, the investigators screened over 2 million individual cells that had been treated with more than 40 different compounds – including industrial chemicals, agricultural chemicals, and various pharmaceuticals. They took microscopic images of the cells following each exposure and examined the images for signs of structural changes or damage to the cells. Those cells treated with compounds known to cause kidney toxicity showed a pattern of changes the team used to construct a “toxicity profile.” By training their image analysis software to look for these features in other cell images, they developed a highly accurate, high-speed, automated screening system.

Last summer the team announced a stem cell method for producing a reliable supply of human kidney cells for in vitro screening. However, they noted in the current paper that certain steps in the earlier procedure are difficult to automate.  The new image-based screening system can be completely automated, allowing investigators to screen a much larger number of chemicals in a shorter period of time.

Original article: Su et al. (2015) High-Throughput Imaging-Based Nephrotoxicity Prediction for Xenobiotics with Diverse Chemical Structures. [Open access]