Liu Xingming, Higuchi Yuichiro, Sakamoto Yuta, Yanagida Shota, Uehara Shotaro, Suemizu Hiroshi, Kanda Yasunari, Kawagishi Hiroyuki
Division of Pharmacology, National Institute of Health Sciences, Kawasaki, Japan.
Central Institute for Experimental Medicine and Life Science, Kawasaki, Japan.
J Appl Toxicol. 2025 Jul 21. doi: 10.1002/jat.4863.
Drug-induced liver injury (DILI) is a serious adverse event and a common cause of postmarketing drug withdrawal. Despite nonclinical assessments of DILI risk, which are predominantly conducted in experimental animals, DILI remains a frequent adverse event, highlighting the need to improve nonclinical assessments. Extensive studies have demonstrated that primary human hepatocytes (PHHs) and their three-dimensional (3D) cultures, such as spheroids, exhibit high predictability of drug-induced hepatotoxicity in vitro. However, lot-to-lot variations and inconsistent availability of PHHs remain major limitations. In contrast, HepaSH cells, which are hepatocytes derived from humanized liver chimeric mice, are more consistently available than PHHs and display gene expression levels of drug-metabolizing factors similar to those of PHHs with minimal variation. In this study, we investigated whether HepaSH cell-derived spheroids can be used to assess DILI risk. HepaSH cells were cultured in two-dimensional (2D) and three-dimensional (3D) conditions for up to 21 days. Gene expression and phenotypic analyses revealed that 3D-cultured HepaSH cells exhibited greater functional development of hepatic characteristics than 2D-cultured cells. Cytotoxic assessment using DILI-associated drugs listed in the DILIrank public dataset demonstrated that spheroids cultured in chemically defined William's media exhibited the highest sensitivity to drugs classified as having a high risk of DILI. Thus, our data suggest that human hepatic spheroids generated from HepaSH cells can be used to assess DILI risk in vitro. Further studies involving large-scale compound screening using HepaSH cells would be necessary to investigate the predictability of human DILI in nonclinical assessments.