Five-Lineage Human Liver Organoid Screen with Indirect-Toxicity Follow-Up
Compare multi-lineage and hepatocyte-only constructs, then use component cultures, transferred vesicles and pathway perturbations to investigate a differential drug-injury signal.
Detect and dissect cell-interaction-dependent drug-induced liver injury.
Model & Design
H1 human embryonic stem-cell derivatives: 60% hepatocytes and 10% each stellate, endothelial, Kupffer and cholangiocyte populations; 30,000 cells per initial organoid.
Experimental Format
58-drug, 14-day ATP-viability comparison, referenced to human clinical maximum plasma concentration (Cmax). Results Sec5 states 1× Cmax for each compound; Methods Sec28 describes concentrations generally at or above Cmax. Exact per-drug exposures are listed in source Supplementary Data 1. Below 80% of vehicle plus P < 0.05 defines injury. Separate imipramine 0.87 μM mechanistic experiments include three-day stellate-cell vesicle conditioning.
Model Components
Component
Models and panel
Description
Five-lineage organoids, hepatocyte-only spheroids/monolayers and component cultures; 34 drugs associated with clinical drug-induced liver injury (DILI) and 24 negative comparators.
Component
Mechanistic comparisons
Description
Conditioned medium, isolated exosome fractions, TRKB/p53 perturbations, CD9 uptake inhibition, GW4869 production inhibition and miR-34/XIAP reporter/response measurements.
Measurements & Endpoints
ATP viability and panel classification
Cell-type-specific apoptosis
Vesicle number/diameter/uptake
miR-34 cargo and XIAP/caspase response
Separate mouse injury markers
Analysis
Keep panel classification, differential model sensitivity, causal component experiments and animal corroboration separate. Vesicle-cargo changes are distinguished from unchanged vesicle quantity/size.
Published Evidence
Study Design
Compare 58 drug responses, select imipramine for component/vesicle experiments, and assess the proposed pathway in a separate mouse study.
Study Sample
One H1 human embryonic stem-cell background; 58 drugs. Screen plot n = 6 technical replicates; functional/mechanistic experiments have separately reported biological repeats. Mouse apoptosis n = 6/group and liver enzymes n = 5/group.
Comparison
Hepatocyte-only models, vehicle, control vesicles, TRKB/p53 perturbations and vesicle/cargo interventions; separate saline mouse controls.
Finding
28/34 positive drugs detected (82.4% sensitivity); 6/24 negative comparators falsely flagged (75% specificity). Imipramine altered stellate-derived miR-34 cargo and recipient XIAP/apoptosis without increasing vesicle number or size; several interventions attenuated injury.
Interpretation & Limits
Single stem-cell genetic background and endpoint ATP screening.
Six false-negative and six false-positive panel results; some hepatocyte-only toxicities absent from organoids.
Static culture, absent adaptive immunity and incompletely characterized vascular/biliary function.
Mechanistic follow-up uses distinct exposure conditions from the screening panel; mouse results do not measure human clinical incidence.
A 58-drug comparison in human stem-cell-derived liver organoids detected imipramine toxicity that hepatocyte-only spheroids missed. Follow-up experiments linked stellate-cell injury to altered exosome cargo, transfer of miR-34 to hepatocytes and reduced expression of the apoptosis inhibitor XIAP.