The importance of preclinical models in cholangiocarcinoma.

Cholangiocarcinoma (CCA) is an adenocarcinoma of the hepatobiliary system with a grim prognosis. Incidence is rising globally and surgery is currently the only curative treatment, but is only available for patients who are fit and diagnosed in an early-stage of disease progression. Great importance has been placed on developing preclinical models to help further our understanding of CCA and potential treatments to improve therapeutic outcomes. Preclinical models of varying complexity and cost have been established, ranging from more simplistic in vitro 2D CCA cell lines in culture, to more complex in vivo genetically engineered mouse models. Currently there is no single model that faithfully recaptures the complexities of human CCA and the in vivo tumour microenvironment. Instead a multi-model approach should be used when designing preclinical trials to study CCA and potential therapies.

Study design for development of novel safety biomarkers of drug-induced liver injury by the translational safety biomarker pipeline (TransBioLine) consortium: a study protocol for a nested case-control study.

A lack of biomarkers that detect drug-induced liver injury (DILI) accurately continues to hinder early- and late-stage drug development and remains a challenge in clinical practice. The Innovative Medicines Initiative's TransBioLine consortium comprising academic and industry partners is developing a prospective repository of deeply phenotyped cases and controls with biological samples during liver injury progression to facilitate biomarker discovery, evaluation, validation and qualification.In a nested case-control design, patients who meet one of these criteria, alanine transaminase (ALT) ≥ 5 × the upper limit of normal (ULN), alkaline phosphatase ≥ 2 × ULN or ALT ≥ 3 ULN with total bilirubin > 2 × ULN, are enrolled. After completed clinical investigations, Roussel Uclaf Causality Assessment and expert panel review are used to adjudicate episodes as DILI or alternative liver diseases (acute non-DILI controls). Two blood samples are taken: at recruitment and follow-up. Sample size is as follows: 300 cases of DILI and 130 acute non-DILI controls. Additional cross-sectional cohorts (1 visit) are as follows: Healthy volunteers (n = 120), controls with chronic alcohol-related or non-alcoholic fatty liver disease (n = 100 each) and patients with psoriasis or rheumatoid arthritis (n = 100, 50 treated with methotrexate) are enrolled. Candidate biomarkers prioritised for evaluation include osteopontin, glutamate dehydrogenase, cytokeratin-18 (full length and caspase cleaved), macrophage-colony-stimulating factor 1 receptor and high mobility group protein B1 as well as bile acids, sphingolipids and microRNAs. The TransBioLine project is enabling biomarker discovery and validation that could improve detection, diagnostic accuracy and prognostication of DILI in premarketing clinical trials and for clinical healthcare application.

A systems approach reveals species differences in hepatic stress response capacity.

To minimize the occurrence of unexpected toxicities in early phase preclinical studies of new drugs, it is vital to understand fundamental similarities and differences between preclinical species and humans. Species differences in sensitivity to acetaminophen (APAP) liver injury have been related to differences in the fraction of the drug that is bioactivated to the reactive metabolite N-acetyl-p-benzoquinoneimine (NAPQI). We have used physiologically based pharmacokinetic modeling to identify oral doses of APAP (300 and 1000 mg/kg in mice and rats, respectively) yielding similar hepatic burdens of NAPQI to enable the comparison of temporal liver tissue responses under conditions of equivalent chemical insult. Despite pharmacokinetic and biochemical verification of the equivalent NAPQI insult, serum biomarker and tissue histopathology analyses revealed that mice still exhibited a greater degree of liver injury than rats. Transcriptomic and proteomic analyses highlighted the stronger activation of stress response pathways (including the Nrf2 oxidative stress response and autophagy) in the livers of rats, indicative of a more robust transcriptional adaptation to the equivalent insult. Components of these pathways were also found to be expressed at a higher basal level in the livers of rats compared with both mice and humans. Our findings exemplify a systems approach to understanding differential species sensitivity to hepatotoxicity. Multiomics analysis indicated that rats possess a greater basal and adaptive capacity for hepatic stress responses than mice and humans, with important implications for species selection and human translation in the safety testing of new drug candidates associated with reactive metabolite formation.

VEGF Stimulates Activation of ERK5 in the Absence of C-Terminal Phosphorylation Preventing Nuclear Localization and Facilitating AKT Activation in Endothelial Cells.

Extracellular-signal-regulated kinase 5 (ERK5) is critical for normal cardiovascular development. Previous studies have defined a canonical pathway for ERK5 activation, showing that ligand stimulation leads to MEK5 activation resulting in dual phosphorylation of ERK5 on Thr218/Tyr220 residues within the activation loop. ERK5 then undergoes a conformational change, facilitating phosphorylation on residues in the C-terminal domain and translocation to the nucleus where it regulates MEF2 transcriptional activity. Our previous research into the importance of ERK5 in endothelial cells highlighted its role in VEGF-mediated tubular morphogenesis and cell survival, suggesting that ERK5 played a unique role in endothelial cells. Our current data show that in contrast to EGF-stimulated HeLa cells, VEGF-mediated ERK5 activation in human dermal microvascular endothelial cells (HDMECs) does not result in C-terminal phosphorylation of ERK5 and translocation to the nucleus, but instead to a more plasma membrane/cytoplasmic localisation. Furthermore, the use of small-molecule inhibitors to MEK5 and ERK5 shows that instead of regulating MEF2 activity, VEGF-mediated ERK5 is important for regulating AKT activity. Our data define a novel pathway for ERK5 activation in endothelial cells leading to cell survival.

Detection of Hepatic Drug Metabolite-Specific T-Cell Responses Using a Human Hepatocyte, Immune Cell Coculture System.

Drug-responsive T-cells are activated with the parent compound or metabolites, often via different pathways (pharmacological interaction and hapten). An obstacle to the investigation of drug hypersensitivity is the scarcity of reactive metabolites for functional studies and the absence of coculture systems to generate metabolites in situ. Thus, the aim of this study was to utilize dapsone metabolite-responsive T-cells from hypersensitive patients, alongside primary human hepatocytes to drive metabolite formation, and subsequent drug-specific T-cell responses. Nitroso dapsone-responsive T-cell clones were generated from hypersensitive patients and characterized in terms of cross-reactivity and pathways of T-cell activation. Primary human hepatocytes, antigen-presenting cells, and T-cell cocultures were established in various formats with the liver and immune cells separated to avoid cell contact. Cultures were exposed to dapsone, and metabolite formation and T-cell activation were measured by LC-MS and proliferation assessment, respectively. Nitroso dapsone-responsive CD4+ T-cell clones from hypersensitive patients were found to proliferate and secrete cytokines in a dose-dependent manner when exposed to the drug metabolite. Clones were activated with nitroso dapsone-pulsed antigen-presenting cells, while fixation of antigen-presenting cells or omission of antigen-presenting cells from the assay abrogated the nitroso dapsone-specific T-cell response. Importantly, clones displayed no cross-reactivity with the parent drug. Nitroso dapsone glutathione conjugates were detected in the supernatant of hepatocyte immune cell cocultures, indicating that hepatocyte-derived metabolites are formed and transferred to the immune cell compartment. Similarly, nitroso dapsone-responsive clones were stimulated to proliferate with dapsone, when hepatocytes were added to the coculture system. Collectively, our study demonstrates the use of hepatocyte immune cell coculture systems to detect in situ metabolite formation and metabolite-specific T-cell responses. Similar systems should be used in future diagnostic and predictive assays to detect metabolite-specific T-cell responses when synthetic metabolites are not available.

Analysis of serum microRNA-122 in a randomized controlled trial of N-acetylcysteine for treatment of antituberculosis drug-induced liver injury.

AIM: Serum microRNA-122 (miR-122) is a novel biomarker for drug-induced liver injury, with good sensitivity in the early diagnosis of paracetamol-induced liver injury. We describe miR-122 concentrations in participants with antituberculosis drug-induced liver injury (AT-DILI). We explored the relationship between miR-122 and alanine aminotransferase (ALT) concentrations and the effect of N-acetylcysteine (NAC) on miR-122 concentrations. METHODS: We included participants from a randomized placebo-controlled trial of intravenous NAC in AT-DILI. ALT and miR-122 concentrations were quantified before and after infusion of NAC/placebo. We assessed correlations between ALT and miR-122 concentrations and described changes in ALT and miR-122 concentrations between sampling occasions. RESULTS: We included 45 participants; mean age (± standard deviation) 38 (±10) years, 58% female and 91% HIV positive. The median (interquartile range) time between pre- and post-infusion biomarker specimens was 68 h (47-77 h). The median pre-infusion ALT and miR-122 concentrations were 420 U/L (238-580) and 0.58 pM (0.18-1.47), respectively. Pre-infusion ALT and miR-122 concentrations were correlated (Spearman's ρ = .54, P = .0001). Median fold-changes in ALT and miR-122 concentrations between sampling were 0.56 (0.43-0.69) and 0.75 (0.23-1.53), respectively, and were similar in the NAC and placebo groups (P = .40 and P = .68 respectively). CONCLUSIONS: miR-122 concentrations in our participants with AT-DILI were considerably higher than previously reported in healthy volunteers and in patients on antituberculosis therapy without liver injury. We did not detect an effect of NAC on miR-122 concentrations. Further research is needed to determine the utility of miR-122 in the diagnosis and management of AT-DILI.

Interruption of bile acid uptake by hepatocytes after acetaminophen overdose ameliorates hepatotoxicity.

BACKGROUND & AIMS: Acetaminophen (APAP) overdose remains a frequent cause of acute liver failure, which is generally accompanied by increased levels of serum bile acids (BAs). However, the pathophysiological role of BAs remains elusive. Herein, we investigated the role of BAs in APAP-induced hepatotoxicity. METHODS: We performed intravital imaging to investigate BA transport in mice, quantified endogenous BA concentrations in the serum of mice and patients with APAP overdose, analyzed liver tissue and bile by mass spectrometry and MALDI-mass spectrometry imaging, assessed the integrity of the blood-bile barrier and the role of oxidative stress by immunostaining of tight junction proteins and intravital imaging of fluorescent markers, identified the intracellular cytotoxic concentrations of BAs, and performed interventions to block BA uptake from blood into hepatocytes. RESULTS: Prior to the onset of cell death, APAP overdose causes massive oxidative stress in the pericentral lobular zone, which coincided with a breach of the blood-bile barrier. Consequently, BAs leak from the bile canaliculi into the sinusoidal blood, which is then followed by their uptake into hepatocytes via the basolateral membrane, their secretion into canaliculi and repeated cycling. This, what we termed 'futile cycling' of BAs, led to increased intracellular BA concentrations that were high enough to cause hepatocyte death. Importantly, however, the interruption of BA re-uptake by pharmacological NTCP blockage using Myrcludex B and Oatp knockout strongly reduced APAP-induced hepatotoxicity. CONCLUSIONS: APAP overdose induces a breach of the blood-bile barrier which leads to futile BA cycling that causes hepatocyte death. Prevention of BA cycling may represent a therapeutic option after APAP intoxication. LAY SUMMARY: Only one drug, N-acetylcysteine, is approved for the treatment of acetaminophen overdose and it is only effective when given within ∼8 hours after ingestion. We identified a mechanism by which acetaminophen overdose causes an increase in bile acid concentrations (to above toxic thresholds) in hepatocytes. Blocking this mechanism prevented acetaminophen-induced hepatotoxicity in mice and evidence from patients suggests that this therapy may be effective for longer periods after ingestion compared to N-acetylcysteine.

A MicroRNA Next-Generation-Sequencing Discovery Assay (miND) for Genome-Scale Analysis and Absolute Quantitation of Circulating MicroRNA Biomarkers.

The plasma levels of tissue-specific microRNAs can be used as diagnostic, disease severity and prognostic biomarkers for chronic and acute diseases and drug-induced injury. Thereby, the combination of diverse microRNAs into biomarker signatures using multivariate statistics seems especially powerful from the perspective of tissue and condition specific microRNA shedding into the plasma. Although next-generation sequencing (NGS) technology enables one to analyse circulating microRNAs on a genome-scale level, it suffers from potential biases (e.g., adapter ligation bias) and lacks absolute transcript quantitation as well as tailor-made quality controls. In order to develop a robust NGS discovery assay for genome-scale quantitation of circulating microRNAs, we first evaluated the sensitivity, repeatability and ligation bias of four commercially available small RNA library preparation protocols. The protocol from RealSeq Biosciences was selected based on its performance and usability and coupled with a novel panel of exogenous small RNA spike-in controls to enable quality control and absolute quantitation, thus ensuring comparability of data across independent NGS experiments. The established microRNA Next-Generation-Sequencing Discovery Assay (miND) was validated for its relative accuracy, precision, analytical measurement range and sequencing bias and was considered fit-for-purpose for microRNA biomarker discovery. Summarized, all these criteria were met, and thus, our analytical platform is considered fit-for-purpose for microRNA biomarker discovery from biofluids in the setting of any diagnostic, prognostic or patient stratification need. The established miND assay was tested on serum, cerebrospinal fluid (CSF), synovial fluid (SF) and extracellular vesicles (EV) extracted from cell culture medium of primary cells and proved its potential to be used across different sample types.

The application of cytokeratin-18 as a biomarker for drug-induced liver injury.

Drug-induced liver injury (DILI) is a frequent and dangerous adverse effect faced during preclinical and clinical drug therapy. DILI is a leading cause of candidate drug attrition, withdrawal and in clinic, is the primary cause of acute liver failure. Traditional diagnostic markers for DILI include alanine aminotransferase (ALT), aspartate aminotransferase (AST) and alkaline phosphatase (ALP). Yet, these routinely used diagnostic markers have several noteworthy limitations, restricting their sensitivity, specificity and accuracy in diagnosing DILI. Consequently, new biomarkers for DILI need to be identified.A potential biomarker for DILI is cytokeratin-18 (CK18), an intermediate filament protein highly abundant in hepatocytes and cholangiocytes. Extensively researched in a variety of clinical settings, both full length and cleaved forms of CK18 can diagnose early-stage DILI and provide insight into the mechanism of hepatocellular injury compared to traditionally used diagnostic markers. However, relatively little research has been conducted on CK18 in preclinical models of DILI. In particular, CK18 and its relationship with DILI is yet to be characterised in an in vivo rat model. Such characterization of CK18 and ccCK18 responses may enable their use as translational biomarkers for hepatotoxicity and facilitate management of clinical DILI risk in drug development. The aim of this review is to discuss the application of CK18 as a biomarker for DILI. Specifically, this review will highlight the properties of CK18, summarise clinical research that utilised CK18 to diagnose DILI and examine the current challenges preventing the characterisation of CK18 in an in vivo rat model of DILI.

Proteomic profiling of murine biliary-derived hepatic organoids and their capacity for drug disposition, bioactivation and detoxification.

Hepatic organoids are a recent innovation in in vitro modeling. Initial studies suggest that organoids better recapitulate the liver phenotype in vitro compared to pre-existing proliferative cell models. However, their potential for drug metabolism and detoxification remains poorly characterized, and their global proteome has yet to be compared to their tissue of origin. This analysis is urgently needed to determine what gain-of-function this new model may represent for modeling the physiological and toxicological response of the liver to xenobiotics. Global proteomic profiling of undifferentiated and differentiated hepatic murine organoids and donor-matched livers was, therefore, performed to assess both their similarity to liver tissue, and the expression of drug-metabolizing enzymes and transporters. This analysis quantified 4405 proteins across all sample types. Data are available via ProteomeXchange (PXD017986). Differentiation of organoids significantly increased the expression of multiple cytochrome P450, phase II enzymes, liver biomarkers and hepatic transporters. While the final phenotype of differentiated organoids is distinct from liver tissue, the organoids contain multiple drug metabolizing and transporter proteins necessary for liver function and drug metabolism, such as cytochrome P450 3A, glutathione-S-transferase alpha and multidrug resistance protein 1A. Indeed, the differentiated organoids were shown to exhibit increased sensitivity to midazolam (10-1000 µM) and irinotecan (1-100 µM), when compared to the undifferentiated organoids. The predicted reduced activity of HNF4A and a resulting dysregulation of RNA polymerase II may explain the partial differentiation of the organoids. Although further experimentation, optimization and characterization is needed relative to pre-existing models to fully contextualize their use as an in vitro model of drug-induced liver injury, hepatic organoids represent an attractive novel model of the response of the liver to xenobiotics. The current study also highlights the utility of global proteomic analyses for rapid and accurate evaluation of organoid-based test systems.

Pharmacological Activation of Nrf2 Enhances Functional Liver Regeneration.

BACKGROUND AND AIMS: The transcription factor nuclear factor erythroid 2-related factor 2 (Nrf2) regulates an array of cytoprotective genes, yet studies in transgenic mice have led to conflicting reports on its role in liver regeneration. We aimed to test the hypothesis that pharmacological activation of Nrf2 would enhance liver regeneration. APPROACH AND RESULTS: Wild-type and Nrf2 null mice were administered bardoxolone methyl (CDDO-Me), a potent activator of Nrf2 that has entered clinical development, and then subjected to two-thirds partial hepatectomy. Using translational noninvasive imaging techniques, CDDO-Me was shown to enhance the rate of restoration of liver volume (MRI) and improve liver function (multispectral optoacoustic imaging of indocyanine green clearance) in wild-type, but not Nrf2 null, mice following partial hepatectomy. Using immunofluorescence imaging and whole transcriptome analysis, these effects were found to be associated with an increase in hepatocyte hypertrophy and proliferation, the suppression of immune and inflammatory signals, and metabolic adaptation in the remnant liver tissue. Similar processes were modulated following exposure of primary human hepatocytes to CDDO-Me, highlighting the potential relevance of our findings to patients. CONCLUSIONS: Our results indicate that pharmacological activation of Nrf2 is a promising strategy for enhancing functional liver regeneration. Such an approach could therefore aid the recovery of patients undergoing liver surgery and support the treatment of acute and chronic liver disease.

Safety perspectives on presently considered drugs for the treatment of COVID-19.

Intense efforts are underway to evaluate potential therapeutic agents for the treatment of COVID-19. In order to respond quickly to the crisis, the repurposing of existing drugs is the primary pharmacological strategy. Despite the urgent clinical need for these therapies, it is imperative to consider potential safety issues. This is important due to the harm-benefit ratios that may be encountered when treating COVID-19, which can depend on the stage of the disease, when therapy is administered and underlying clinical factors in individual patients. Treatments are currently being trialled for a range of scenarios from prophylaxis (where benefit must greatly exceed risk) to severe life-threatening disease (where a degree of potential risk may be tolerated if it is exceeded by the potential benefit). In this perspective, we have reviewed some of the most widely researched repurposed agents in order to identify potential safety considerations using existing information in the context of COVID-19.

Current Perspective: 3D Spheroid Models Utilizing Human-Based Cells for Investigating Metabolism-Dependent Drug-Induced Liver Injury.

Drug-induced liver injury (DILI) remains a leading cause for the withdrawal of approved drugs. This has significant financial implications for pharmaceutical companies, places increasing strain on global health services, and causes harm to patients. For these reasons, it is essential that in-vitro liver models are capable of detecting DILI-positive compounds and their underlying mechanisms, prior to their approval and administration to patients or volunteers in clinical trials. Metabolism-dependent DILI is an important mechanism of drug-induced toxicity, which often involves the CYP450 family of enzymes, and is associated with the production of a chemically reactive metabolite and/or inefficient removal and accumulation of potentially toxic compounds. Unfortunately, many of the traditional in-vitro liver models fall short of their in-vivo counterparts, failing to recapitulate the mature hepatocyte phenotype, becoming metabolically incompetent, and lacking the longevity to investigate and detect metabolism-dependent DILI and those associated with chronic and repeat dosing regimens. Nevertheless, evidence is gathering to indicate that growing cells in 3D formats can increase the complexity of these models, promoting a more mature-hepatocyte phenotype and increasing their longevity, in vitro. This review will discuss the use of 3D in vitro models, namely spheroids, organoids, and perfusion-based systems to establish suitable liver models to investigate metabolism-dependent DILI.

Critical considerations for targeting colorectal liver metastases with nanotechnology.

Colorectal cancer remains a significant cause of morbidity and mortality worldwide. Half of all patients develop liver metastases, presenting unique challenges for their treatment. The shortcomings of conventional chemotherapy has encouraged the use of nanomedicines; the application of nanotechnology in the diagnosis and treatment of disease. In spite of technological improvements in nanotechnology, the complexity of biological systems hinders the prospect of nanomedicines being applied in cancer therapy at the present time. This review highlights current biological barriers and discusses aspects of tumor biology together with the physicochemical features of the nanocarrier, that need to be considered in order to develop effective nanotherapeutics for colorectal cancer patients with liver metastases. It becomes clear that incorporating an interdisciplinary approach when developing nanomedicines should assure appropriate disease-driven design and that this will form a critical step in improving their clinical translation. This article is characterized under: Therapeutic Approaches and Drug Discovery > Nanomedicine for Oncologic Disease.

Managing the challenge of drug-induced liver injury: a roadmap for the development and deployment of preclinical predictive models.

Drug-induced liver injury (DILI) is a patient-specific, temporal, multifactorial pathophysiological process that cannot yet be recapitulated in a single in vitro model. Current preclinical testing regimes for the detection of human DILI thus remain inadequate. A systematic and concerted research effort is required to address the deficiencies in current models and to present a defined approach towards the development of new or adapted model systems for DILI prediction. This Perspective defines the current status of available models and the mechanistic understanding of DILI, and proposes our vision of a roadmap for the development of predictive preclinical models of human DILI.

Correction: The Nrf2 inhibitor brusatol is a potent antitumour agent in an orthotopic mouse model of colorectal cancer.

[This corrects the article DOI: 10.18632/oncotarget.25497.].

Development of an orthotopic syngeneic murine model of colorectal cancer for use in translational research.

Improving outcomes in colorectal cancer requires more accurate in vivo modelling of the disease in humans, allowing more reliable pre-clinical assessment of potential therapies. Novel imaging techniques are necessary to improve the longitudinal assessment of disease burden in these models, reducing the number of animals required for translational studies. This report describes the development of an immune-competent syngeneic orthotopic murine model of colorectal cancer, utilising caecal implantation of CT26 cells stably transfected with the luciferase gene into immune-competent BALB/c mice, allowing serial bioluminescent imaging of cancer progression. Luminescence in the stably transfected CT26 cell line, after pre-conditioning in the flank of a BALB/c mouse, accurately reflected cell viability and resulted in primary caecal tumours in five of eight (63%) mice in the initial pilot study following caecal injection. Luminescent signal continued to increase throughout the study period with one mouse (20%) developing a liver metastasis. Histopathological assessment confirmed tumours to be consistent with a poorly differentiated adenocarcinoma. We have now performed this technique in 68 immune-competent BALB/c mice. There have been no complications from the procedure or peri-operative deaths, with primary tumours developing in 44 (65%) mice and liver metastases in nine (20%) of these. This technique provides an accurate model of colorectal cancer with tumours developing in the correct microenvironment and metastasising to the liver with a similar frequency to that seen in patients presenting with colorectal cancer, with serial bioluminescent reducing the murine numbers required in studies by removing the need for cull for assessment of disease burden.

Application of porcine gastrointestinal organoid units as a potential in vitro tool for drug discovery and development.

The gastrointestinal tract (GI) is a crucial part of the body for growth and development and its dysregulation can lead to several diseases with detrimental effects. Most of these diseases lack effective treatment, occurring as a result of inappropriate models to develop safe and potent therapies. Organoids are three-dimensional self-organizing and self-renewing structures that are composed of a cluster of different cells in vitro that resemble their organ of origin in architecture and function. Over recent years, organoids have been increasingly used to study developmental biology, disease progression, i.e., cancer, tissue engineering and regenerative medicine and other biological processes. Owing to their complex nature and ability to retain the morphological and molecular patterns of their tissue-of-origin, they have great potential as alternative tools/models for drug screening, development and biomarker discovery. Using a species with similar genetic homology to humans as a source of organoids, such as the porcine model may offer huge translational relevance. This review focuses on the culture and establishment of porcine organoid units and their potential use and application as in vitro models to further the science of drug discovery, by overcoming current limitations of established two- and three-dimensional models. It also highlights the translational application of using porcine organoids as a model of different disease contexts.

Characterisation of the NRF2 transcriptional network and its response to chemical insult in primary human hepatocytes: implications for prediction of drug-induced liver injury.

The transcription factor NRF2, governed by its repressor KEAP1, protects cells against oxidative stress. There is interest in modelling the NRF2 response to improve the prediction of clinical toxicities such as drug-induced liver injury (DILI). However, very little is known about the makeup of the NRF2 transcriptional network and its response to chemical perturbation in primary human hepatocytes (PHH), which are often used as a translational model for investigating DILI. Here, microarray analysis identified 108 transcripts (including several putative novel NRF2-regulated genes) that were both downregulated by siRNA targeting NRF2 and upregulated by siRNA targeting KEAP1 in PHH. Applying weighted gene co-expression network analysis (WGCNA) to transcriptomic data from the Open TG-GATES toxicogenomics repository (representing PHH exposed to 158 compounds) revealed four co-expressed gene sets or 'modules' enriched for these and other NRF2-associated genes. By classifying the 158 TG-GATES compounds based on published evidence, and employing the four modules as network perturbation metrics, we found that the activation of NRF2 is a very good indicator of the intrinsic biochemical reactivity of a compound (i.e. its propensity to cause direct chemical stress), with relatively high sensitivity, specificity, accuracy and positive/negative predictive values. We also found that NRF2 activation has lower sensitivity for the prediction of clinical DILI risk, although relatively high specificity and positive predictive values indicate that false positive detection rates are likely to be low in this setting. Underpinned by our comprehensive analysis, activation of the NRF2 network is one of several mechanism-based components that can be incorporated into holistic systems toxicology models to improve mechanistic understanding and preclinical prediction of DILI in man.