Researchers have created a "Liver-Chip" using Organs-on-Chips technology that can predict and characterize the liver toxicity of various drug candidates and compounds in rats, dogs, and humans. The new chips offer a more accurate and translatable platform for assessing the toxicity of potential drugs, which are currently tested in preclinical studies using cells, and rodents or other species. These animal studies cannot always predict which compounds might cause toxicity in people - a recent analysis of 150 drugs showed regulatory testing in animals could only predict 71% of toxicities in humans - and preclinical drug testing is especially inaccurate for predicting liver injury in humans. The FDA also prioritizes understanding the mechanisms of drug toxicity to reduce risks for patients, but how drugs induce liver toxicity is not completely understood. To address these issues, Kyung-Jin Jang and colleagues created the Liver-Chip that can house an array of different types of liver cells, such as endothelial cells, stellate cells, and Kupffer cells, from rat, dog, and human sources. Using the new technology, the scientists were able to detect across species the liver-specific effects of drugs such as bosentan (approved to treat pulmonary hypertension) and determined how high doses of the pain-reliever acetaminophen caused toxic effects on liver cells by generating reactive oxygen species. The chips also showed that compounds that are unsafe in animals may not necessarily be toxic in humans; one experiment demonstrated that the discontinued compound JNJ-2 caused harmful fibrosis in rat chips but did not have negative effects on human liver cells. Finally, the platform could capture the idiosyncratic toxic effects - often missed during early clinical testing - of a failed drug candidate called TAK-875, indicating the chips could help researchers better understand the side effects of drug candidates throughout the testing process.