Hepatotoxicity in Drug Development: Detecting DILI Early with Zebrafish

Hepatotoxicity in Drug Development

Drug-induced liver injury (DILI) is liver damage resulting from an adverse reaction to a drug or increasingly to herbal and dietary supplements, with a presentation that varies by compound, dose, exposure, and individual susceptibility. It is one of the most persistent liabilities in drug development, accounting for roughly half of all acute liver failure cases in the US.

 

DILI in the Pharmaceutical Pipeline: Why Liver Toxicity Keeps Derailing Drug Development

DILI is a major safety concern and a leading reason for denial of approval, market withdrawal, or "black box" warnings. It drives most post-marketing withdrawals and affects about 1 in 10,000 people each year, likely an underestimate.

Part of what makes it so hard to design out is that it wears two faces. Some DILI is dose-dependent and broadly predictable; acetaminophen overdose, also known as paracetamol, is the textbook example. It accounts for ~50% of acute liver failure cases in the USA and some European countries. On the other hand, some cases are idiosyncratic. Although it occurs more frequently with doses of >50-100 mg per day, this dose varies among individuals and surfaces only late or after market release, making it very difficult to predict and a recurring reason promising candidates fail. 

 

Mechanisms of Drug-Induced Hepatotoxicity: Intrinsic, Idiosyncratic, and Reactive Pathways

DILI is generally classified into three categories:

  • Intrinsic DILI is predictable and dose-dependent, occurring in most individuals above a given dose. 
  • Idiosyncratic DILI is largely dose-independent, although it seems there is a necessary minimum dose to trigger the cellular cascade of events leading to damage. It is highly unpredictable and dependent on individual susceptibility factors.
  • Indirect DILI results from the biological action of the drug rather than the drug itself, a category brought into focus by immune checkpoint inhibitors.

Injury then converges on three pathways: organelle stress (mitochondrial and endoplasmic reticulum stress), cholestasis (inhibition of bile efflux transporters such as BSEP), and immune responses (driven by damage-associated molecular patterns and, in idiosyncratic cases, recognition of drug-protein adducts). Acetaminophen illustrates the reactive-metabolite route: it is converted to NAPQI, normally neutralized by glutathione, but in overdose, glutathione is depleted, NAPQI forms protein adducts, and mitochondrial oxidative stress drives necrosis.

 

Why Idiosyncratic DILI Is the Hardest to Predict

Idiosyncratic DILI is typically identified only in late development or after market release, occurring in fewer than one in 10,000 to 100,000 people at therapeutic doses. Because cases are so rare, the pathogenesis is poorly understood and hard to anticipate early.

It depends on individual susceptibility, largely a person's metabolic or immune response. The most commonly reported risk factor is the HLA genotype, but HLA status alone cannot predict who will develop DILI. This makes it hard to reproduce in preclinical models, and it is as rare in animals as in patients. The field increasingly uses genome-wide association studies (GWAS) to identify genetic risk factors, which created a perfect opportunity for in vivo models that can validate those candidates efficiently.

 

The Limits of Current Preclinical Models: Where the Detection Gap Lives

In vitro systems (liver slices, cultured primary hepatocytes, and immortalized lines such as HepG2 and HepaRG) are attractive because they support high-throughput screening. Primary hepatocytes, the long-standing gold standard, are difficult to obtain and short-lived. Even in 3D configurations where their phenotype and functions are maintained long enough, they offer an inherently reductionist approach. Much of DILI emerges from whole-organism processes (ADME, cell-tissue interactions, and immune involvement) that simple cell cultures cannot reproduce. 

Moving in vivo solves some problems and creates others. Mammalian studies allow dose-dependent toxicity to be examined within the physiology of a whole organism, with greater sensitivity than cell-based assays. Yet, they are slow, costly, and hampered by species differences. A seminal study by the industry found the positive concordance with human liver toxicities was only 33% (rats), 27% (dogs), and 50% (monkeys) (Monticello et al., 2017). 

The gap sits between the two: high-throughput but reductionist in vitro systems, and slow, species-divergent mammalian ones.

 

Zebrafish Hepatotoxicity Screening: How It Closes the DILI Prediction Gap

Zebrafish deliver whole-organism, in vivo liver data while retaining much of the throughput and cost profile of in vitro screening.

The zebrafish liver is detectable at 3 days post-fertilization (dpf) and fully functional by 5 dpf, with structure and function highly conserved relative to the human liver. Hepatocytes, stellate cells, biliary epithelial cells, and bile ducts are present in the zebrafish liver, although their structural organization is different than that of humans.

Drug metabolism is also conserved. The zebrafish genome contains 94 CYP genes spanning all 18 families present in humans, with both phase I and phase II metabolism functional —cyp3a65 shares about 54% sequence identity with human CYP3A4. Zebrafish metabolize drugs through reactions similar to those in humans, generating the same reactive metabolite (NAPQI) from paracetamol, and CYP induction and inhibition behave as they do in mammals. 

 

Zebrafish Hepatotoxicity Screening

Developmental stages of zebrafish and liver. A. Two-cell stage zebrafish embryo is visible at 1 hpf. The chorion provides protection to the embryo while the yolk supplies all the required nutrients; B. the embryo develops in the chorion until the end of day 2. The 36 hpf embryo is still in the chorion; C. larval stages begin after hatching from the chorion. Liver can be seen as a small tissue at 3 dpf, while gut is visible as of 4 dpf; D. transgenic reporter line fabp10a:mCherry is used to label hepatocytes. Liver tissue is visible as of 3 dpf with this line. Scale bar: 500 μm. Source: Cakan-Akdogan et al. 2023.

Zebrafish are built for screening: optically transparent embryos that develop ex utero, high fecundity (around 200 eggs per week), small larvae suited to multiwell plates, simple drug delivery in water, low-cost husbandry, and alignment with the 3Rs.

Predictivity holds up against the data. In ZeClinics' own ZeGlobalTox validation, we evaluated 24 compounds for their known human toxicity. Hepatotoxicity assessment in zebrafish larvae reached 80% sensitivity, 77% specificity, and 82% accuracy when benchmarked against known human liver toxicity (Cornet et al. 2017). 

The model also offers multiple quantifiable readouts. Using a liver-specific fluorescent reporter line, we can detect changes in liver area that flag hepatomegaly or necrosis, while Oil Red O staining detects drug-induced steatosis (lipid accumulation), which is a fundamental consequence of liver malfunction. These studies can be complemented with histopathology (steatosis, apoptosis, necrosis) and circulating biomarkers such as ALT and microRNA-122. 

Their ease of genetic manipulation allows the creation of zebrafish humanized models that can help understand the mechanisms of CYP involvement in DILI. Transgenic lines expressing human CYP3A4 in hepatocytes (Tg(fabp10a:hCYP3A4-mCherry)) shift metabolism toward the human profile. The immune system is also similar to that of mammals, relevant for the immune-mediated component of DILI. 

Beyond ready-to-use strains, zebrafish can be engineered into custom models to validate candidate DILI genes emerging from human GWAS. CRISPR/Cas9 enables stable gene knockouts; Crispants (F0 crispants) allow rapid functional screening of candidate genes without waiting for stable lines; and Tol2-based transgenesis supports stable reporter or humanized lines. The same humanization approach behind CYP3A4 livers can be extended to other DILI-relevant genes, and gene-edited cohorts can then be challenged with a drug to test whether a variant actually confers susceptibility. 

 

Detect liver liabilities earlier with ZeClinics

At ZeClinics, we have developed a hepatotoxicity assay to evaluate hepatic morphological and functional parameters in 5 dpf zebrafish larvae, which provides quantitative readouts on liver necrosis, hepatomegaly, and hepatic steatosis. 

 

 

Detect liver liabilities earlier with ZeClinics

ZeClinics evaluation of hepatic steatosis. A. Proportion of zebrafish larvae (5 dpf) that show increased liver fat after treatment with a candidate molecule compared to negative and positive controls. B. Representative images of healthy zebrafish larvae (image on the left) and steatotic zebrafish livers (middle and right images) after specific fat staining.

If your question is genetic, our capabilities and experience in CRISPR/Cas9, Crispants generation, and Tol2-based transgenesis allow us to generate from rapid validation models to stable transporter or humanized lines to validate DILI candidate genes in a living vertebrate according to your needs. 

Contact us for more information.

 

Sources

Cakan-Akdogan G, et al. Zebrafish as a model for drug induced liver injury: state of the art and beyond. Explor Dig Dis. 2023;2:44–55. doi: 10.37349/edd.2023.00017

Cornet C, et al. ZeGlobalTox: An Innovative Approach to Address Organ Drug Toxicity Using Zebrafish. Int J Mol Sci. 2017 Apr 19;18(4):864. doi: 10.3390/ijms18040864.

Monticello TM, et al. Current nonclinical testing paradigm enables safe entry to first-in-human clinical trials: the IQ consortium nonclinical to clinical translational database. Toxicol Appl Pharmacol. 2017;334:100-9. doi:10.1016/j.taap.2017.09.006

Skat-Rørdam J, et al. Mechanisms of drug induced liver injury. Cell Mol Life Sci. 2025;82:213. doi: 10.1007/s00018-025-05744-3.

Vliegenthart ADB, et al. Zebrafish as model organisms for studying drug-induced liver injury. Br J Clin Pharmacol. 2014;78(6):1217–1227. doi: 10.1111/bcp.12408

Javier Terriente ZeCardioTX By Javier Terriente

Javier is the co-founder of ZeClinics and ZeCardio Therapeutics, two biotech firms specializing in zebrafish-based preclinical drug discovery for cardiovascular, neural, and toxicology applications. He combines scientific leadership with business acumen, having successfully driven fundraising efforts and strategic partnerships.

Currently leading scientific efforts at ZeCardioTx (and formerly CSO at ZeClinics), Javier also serves on the Board of Directors of AseBio, where he advocates for industry collaboration. His academic background includes a PhD in Molecular Biology and a Marie Curie Fellowship. Recognized as an expert in zebrafish models, he has published extensively and has supervised five industrial PhD theses.

drug developmenthepatotoxicity