Skip to main content

Generation of an adverse outcome pathway network on cholestatic liver injury for mechanism-based in vitro testing of chemicals.

Periodic Reporting for period 1 - ChOLLATERAL (Generation of an adverse outcome pathway network on cholestatic liver injury for mechanism-based in vitro testing of chemicals.)

Reporting period: 2019-05-15 to 2021-05-14

Cholestasis refers to toxic bile acid accumulation that mainly occurs in the liver and can be induced by many chemical compounds, in particular drugs. At present, drug-induced cholestatic injury (DICI) is difficult to predict and is a prominent cause of drug withdrawal from the market. Cholestasis is of high clinical concern and, therefore, represents a significant economic and societal burden. A solution lies in better understanding the mechanisms of DICI. A pragmatic tool to capture the mechanistic basis of toxic effects is the adverse outcome pathway (AOP, Figure 1), which is a schematic representation of the sequence of events that lead to an adverse outcome.
Current AOP constructs frequently provide an oversimplified reflection of the mechanisms underlying toxicity by underestimating initiating events and biological responses.
The overall objective of this project was to elucidate the molecular mechanisms of cholestatic liver injury based on in vitro experimentation and re-use of in vivo and clinical samples, to reliably predict the cholestatic potential of chemicals. In particular, 2 specific research objectives were outlined:
1. To establish a quantitative representation (AOP network) describing the mechanisms of human DICI by considering multiple molecular initiating events and key events of both adversity and homeostatic adaptation.
2. To accurately establish the human cholestatic potential of new chemical entities, such as cosmetic ingredients, biocides, food additives and nanoparticles, using an in vitro 3D liver spheroid model and a battery of tests mechanistically anchored in the novel AOP network.
This multidisciplinary project addresses a ubiquitous medical need regarding the prediction of cholestatic liver toxicity, and may serve as a generic prototype to reliably predict other types of chemical-induced toxicity based on mechanistic knowledge, avoiding the use of animals. Such human-based, animal-saving, approaches to assess the safety of chemicals are urgently needed because of scientific and ethical reasons. By improving in vitro predictability of DICI, the results of this project aim to contribute towards decreased pre- and post-marketing drug failure, and lessening the clinical burden, with concomitant positive effects at the economical level.
This proposal comprised 4 work packages (WP) describing the experimental part and was complemented with training strategies to enhance the career path of the fellow and defined communication/dissemination/exploitation objectives.
WP1 involved the collection of information from in vivo and human clinical samples. These experiments started immediately after submission of the project proposal as some of the samples were re-used from previous projects (in alignment with the 3Rs) and the fellow was already working with the host group. The work was performed with the help of 2 PhD students of the group and published (Van Campenhout et al. 2019, Gijbels et al. 2020). Some work concerning this WP is currently being concluded by 2 other PhD students and will be published in 2 forthcoming research papers, 1 including the remaining in vivo data and another 1 concerning clinical data.
WP2 concerned the collection of data using the selected human liver-based in vitro model. For that, the fellow did her secondment at the Karolinska Institutet in Sweden where she spent 2.5 months acquiring skills in the generation, maintenance and handling of spheroids of primary human hepatocytes. This knowledge was brought and transferred to the host laboratory. For this project, the spheroids were incubated with selected cholestatic drugs under a 28-day repeated exposure protocol, with assessment of results through total adenosine triphosphate (ATP) quantification. Initially, a dose-response curve was performed for each drug, time-point and donor, to select the concentrations to proceed with the experiment. Then, the selected concentrations were tested in the presence and in the absence of a concentrated bile acid mixture containing the most abundant bile acids in humans. This allowed to calculate the cholestatic index (CIx), a ratio between the viability in the presence and in the absence of bile acid mixture, for each drug, time-point and donor. A CIx below 0.8 was taken as an indication of the cholestatic potential of that drug and selected for transcriptomic analysis, which is currently being finalized with the help of a lab technician. The results collected from WP2 will be included in a research article that is currently in preparation and waiting for the results from the transcriptomic analysis.
WP3 comprised the development of the AOP network, which started with the data collected from an in vivo model of extrahepatic cholestasis (Gijbels et al. 2020). The remaining data that is being collected from WP1 and 2 will also be integrated in the AOP network, which will be updated in the research paper mentioned for WP2.
In WP4, the selected in vitro model was exposed to selected chemicals from distinct categories: paraquat, a biocide; tartrazine, a food additive; and triclosan, a cosmetic ingredient. These chemicals have been associated with cholestatic injury, though information is scarce and not yet fully understood. The literature search included in this WP resulted in 3 review papers addressing the different categories of chemicals and what is known about their ability to cause cholestasis/ hepatotoxicity (Vilas-Boas et al. 2019, Vilas-Boas et al. 2020, Vilas-Boas and Vinken 2020). The review paper addressing the hepatotoxic effects of nanomaterials (Vilas-Boas and Vinken 2020) has been highlighted by the Animal Free Safety Assessment (AFSA) collaboration and has caught the attention of stakeholders with regulatory influence, such as the Scientific Committee on Consumer Safety.
During the timeframe of this project, the fellow had the chance to strengthen her network and build up her CV by attending specific training courses, such as “Safety assessment of cosmetics in the EU – Intensive course 2020”, “Applied In Vitro Toxicology Course” and “Leadership for postdocs”, among others. The participation in conferences was restricted by the COVID-19 outbreak, as the ESTIV conference 2020 and the World Congress on Alternatives and Animal Use in Life Sciences 2020 were postponed.
This project has explored, for the first time, the potential cholestatic effects of chemicals other than drugs in the spheroid model of primary human hepatocytes, therefore expanding the model’s applicability domain. Previous works by other groups have addressed shorter exposure periods (max. 14 days), comparatively to the 28-day repeated exposure scheme used in this project, for which there is no previous record in the literature. This represents a significant logistical challenge, as every run of experiments requires 35 days to be completed and was repeated at least 3 times.
Even though the project was terminated on 31.12.2020 the host laboratory is still running the transcriptomic analyses of the in vitro samples and the experiments on the mouse model of drug-induced cholestasis, whose results will generate 2 publications. These will represent the state-of-the art mechanistic knowledge on chemical-induced cholestasis and its in vitro predictability. The improved predictability of cholestasis in vitro can be used by pharma industry to make more reliable decisions on drug candidate development.
Figure 1 - Simplistic representation of an AOP.