Organobodies: A robust and size-controllable system for generating scalable hiPSC-derived liver organoids for drug toxicity screening

Abstract

Background and aimHepatic organoids generated from pluripotent and adult stem cells faithfully mimic the architecture and function of in vivo organs making them powerful tools for preclinical applications such as disease modelling and drug screening. However, developing robust and scalable models capable of preserving liver-specific functions long-term remains technically challenging. MethodsUsing a defined self-assembling peptide, we established a robust and reproducible method to generate hepatic organoids from human induced pluripotent stem cells (hiPSCs), termed hepatic organobodies (OBs), which remained stable in culture for several weeks. The hepatic identity and maturation of OBs were characterized by qPCR, histology, immunofluorescence, and bulk RNA sequencing. Their functional competence was evaluated by ELISA, CYP3A4 induction assays, and assessment of drug biotransformation capacity using HPLC-MS/MS, in parallel with three-dimensional human primary hepatocyte (3D PHH) cultures. Finally, the model was applied to predict the hepatotoxicity of a selected panel of known compounds and benchmarked against HepG2 cells and 3D PHH microtissues. ResultshiPSC-derived hepatocyte-like cells (HLCs) cultured in OBs acquired characteristic hepatic morphology and expressed several key hepatocyte-specific genes, some at levels comparable to those in freshly isolated primary human hepatocytes (PHHs). OBs secreted significantly higher amounts of albumin and 1-antitrypsin (A1AT) compared with parallel 2D-cultured HLCs. Bulk RNA sequencing revealed higher relative expression of major drug metabolism genes, including CYP3A4, CYP2C9, and CYP1A2, as well as enhanced maturation, evidenced by upregulation of the PPAR signalling and fatty acid {beta}-oxidation pathways in OBs relative to 2D HLCs. Moreover, OBs demonstrated CYP450 enzyme activity, with CYP3A4 and CYP2C9 activities comparable to those observed in 3D PHH microtissues. Finally, we show that our 3D model accurately predicted the hepatotoxicity of more than 10 tested chemical compounds. ConclusionHere, we demonstrate a novel, defined, robust and remarkably advanced 3D liver model as a valuable platform for scalable toxicity prediction and drug screening for personalised medicine.