Predictivity of dog co-culture model, primary human hepatocytes and HepG2 cells for the detection of hepatotoxic drugs in humans
Introduction
Along with cardiotoxicity, Drug Induced Liver Injury (DILI) still remains, nowadays, an important hurdle to overcome in drug development (McDonald and Robertson, 2009). Worldwide, the estimated annual incidence rate of DILI is 13.9–24.0 per 100,000 inhabitants (Suk and Kim, 2012). Most cases of DILI are the result of idiosyncratic metabolic responses or unexpected reactions to medication. There is a marked geographic variation in liver injury inducing agents; antibiotics, anticonvulsants, and psychotropic drugs are the most common offending agents in the west, whereas in Asia, herbs and health foods or dietary supplements are more common (Suk and Kim, 2012). Unfortunately, it is well known that the predictive value of animal studies using rodents, dogs, and/or monkeys to detect hepatotoxicity in human is useful but not fully predictive. In an industry-wide study, 55% of drugs known to be hepatotoxic in humans were correctly classified using standard animal models (Olson et al., 2000). Nevertheless, this figure is probably biased as many compounds which have been stopped during preclinical studies should also be classified as hepatotoxic in human. The limited predictivity of animal studies could be due in part to species-specific differences as well as to the use of healthy animals with reduced genetic diversity for the preclinical toxicology studies. It is also acknowledged that patient-related risk factors include underlying disease, age, gender, co-medications, nutritional status, activation of the innate immune system, physical activity, and genetic predispositions (Ulrich, 2007).
For decades, cell-based approaches have been used to predict human hepatotoxic drugs. The main advantages of cellular assays include the relatively low test material requirement, higher experimental throughput, reduction in animal use, and most importantly for human toxicity assessment, the provision of human-specific results via the use of human-derived materials (Li et al., 2012). In the pharmaceutical field, while genotoxicity (Tilmant et al., 2013) and safety pharmacology assays (Rampe and Brown, 2013) are routinely used in screening, the overall value of employing nonspecific cytotoxicity assays remains controversial (Benbow et al., 2010). For instance, on the one hand, in vitro cytotoxicity data have not been applied consistently to the decision making process in early drug development because they have not been highly predictive of in vivo toxicity (McKim, 2010). On the other hand, different studies have shown the utility of in vitro cytotoxicity data to predict in vivo toxicity data. For example, binning compounds into potent (LC50 < 10 μM) and non-potent (LC50 > 100 μM) cytotoxicants in vitro showed that compared to non-potent cytotoxicants the exposure to potent cytotoxicants in vivo resulted in higher overall severity scores at lower exposures (Benbow et al., 2010). Similarly, compounds that have an LC50 ≤ 50 μM are five times more likely to see adverse events at lower Cmax thresholds than those with an LC50 > 50 μM (Greene et al., 2010).
Primary human hepatocytes are considered as the gold standard model for xenobiotic metabolism and cytotoxicity studies (Guillouzo et al., 2007). The main drawbacks of such a model are the scarce availability of fresh cells, complicated isolation procedures, limited life span, inter-individual variability and costs. Immortalized liver-derived cell lines, such as HepG2 cells, a human hepatocellular carcinoma cell line, are also widely used because they are highly differentiated and display many of the genotypic features of normal liver cells (Sassa et al., 1987). Their main limitation is linked to their low metabolic capacities compared to primary hepatocytes and HepaRG cells (Gerets et al., 2012). In terms of predictivity, due to the different approaches used (e.g. number of chemicals, compound classification, presence of serum, duration of exposure, fixed concentrations, multiple of Cmax, cut offs, endpoints), different results have been obtained for primary human hepatocytes (Gerets et al., 2012, Xu et al., 2008) and HepG2 cells (Gerets et al., 2012, O'Brien et al., 2006, Tolosa et al., 2012). For instance, simple cytotoxicity assays using HepG2 cells are relatively insensitive to detect human hepatotoxic drugs (Xu et al., 2004). Other studies using HepG2 cells have also revealed a high sensitivity and specificity (ca 90%) by the means of cell imaging approach (O'Brien et al., 2006, Tolosa et al., 2012). Interestingly, a recent study suggested that from a general screening perspective, hepatic (HepG2), cardiac (H9c2) and kidney (NRK-52E) derived cell lines have relatively equal value in assessing general cytotoxicity and that specific organ toxicity (i.e. liver, kidney and cardiac toxicity) cannot be accurately predicted using such a simple approach (Lin and Will, 2012). This study remarkably used a large number of compounds including 273 hepatotoxic, 191 cardiotoxic and 85 nephrotoxic compounds (Lin and Will, 2012). Specific organ toxicity potentially results from compound accumulation in a particular tissue, cell types within organs, metabolism, and off-target effects (Lin and Will, 2012).
Clearly, there is a need to develop more relevant in vitro model systems to probe and identify pathways that are perturbed following both acute and chronic exposure to chemicals and to help explain species differences in compound biotransformation and bioactivation (LeCluyse et al., 2012). A few recent studies have shown that co-culture models may represent a relevant model for chronic dosing of highly functional hepatocytes (Ukairo et al., 2013) and for hepatotoxicity investigations (Khetani et al., 2013, Li et al., 2012). It is also important to evaluate the metabolic capacities of the emerging models. For instance, in vitro flow systems have been recently developed to better mimic the in vivo situation due to better functionality of the hepatocytes (Chao et al., 2009, Maguire et al., 2009). Flow-based co-culture systems are capable of clearing, with improved resolution and predictive value, compounds with high, medium and low clearance values (Novik et al., 2010). In addition, when co-culture is coupled with flow, higher metabolite production rates are obtained than in static systems (Novik et al., 2010).
For the development of new chemical entities, the actual paradigm in toxicology (supported by regulatory guidelines) is mainly based on in vivo testing using rodent (usually rats) and non-rodent (most often dogs) species to support studies in human because cross-species comparison improves the ability to predict toxicity in human (Olson et al., 2000). Consequently, it is of primary importance to have access to relevant in vitro models using but not limited to rat, dog and human hepatocytes. The in vitro systems should be metabolically competent and functional in order to perform in vitro chronic studies. The main intention is to predict hepatotoxicity using relevant in vitro models from different species to anticipate species differences and ultimately to better select compounds before starting the animal studies. The objective of this present work was to develop a dog co-culture model for long term culture with relevant metabolic capacities and to compare this model with primary human hepatocytes and HepG2 cells in their ability to accurately detect a set of 40 hepatotoxic and 11 non-hepatotoxic drugs in human. The reproducibility of the canine model was also investigated with a set of 14 drugs. In the present manuscript, cytotoxicity endpoints such as cell viability (number of cells or ATP), glutathione (GSH) content and impedance (label free platform) were measured. Multiplexing approaches allow not only to rank and select the most promising compounds but also to gather information that may help to understand some of the toxic events occurring in the cells (Gerets et al., 2009). For instance, GSH is a useful marker to detect potential depletion via reactive metabolites (Xu et al., 2008) and is generally included in multi-parametric testing with liver cells (Khetani et al., 2013). In addition, label-free platforms based on impedance measurement are used to gather information on diverse cellular processes, including proliferation, migration, cytotoxicity and receptor-mediated signaling (Atienzar et al., 2013).
Section snippets
Materials
All compounds and reagents were of analytical grade and were purchased from Sigma-Aldrich (Saint Louis, USA) except bromfenac, entacapone, tolcapone, zileuton, troglitazone, rosiglitazone and cerivastatin which were obtained from Sequoia (Pangbourne, UK).
HepG2 cells
The human hepatocellular carcinoma (HepG2) cell line, purchased from the European Collection of Cell Cultures (ECACC, Salisbury, UK) was maintained as an adherent cell line in Dulbecco's modified Eagle's medium supplemented with 10% fetal
Optimization of the dog co-culture system
The cell viability was between 80 and 85% and 90 and 95% for the dog hepatocytes and non-parenchymal cells, respectively. For the optimization of the dog co-culture model, the influence of parameters including cell morphology, evolution over time, batch selection, media composition, seeding protocol, and thawing was evaluated (data not shown). For instance, Fig. 1 displays the morphology of the canine model after 1, 7 and 14 days of culture with optimal and sub-optimal seeding conditions. Figs. 1
Discussion
Drug-induced Liver Injury (DILI) accounts for approximately 11–13% of acute liver-failure cases in the United States and is the most common cause of death related to this condition (Reuben et al., 2010). It is of extreme importance to detect hepatotoxic candidates as early as possible during the drug development process and before clinical phases. Cell-based approaches have been widely used in the pharmaceutical industry. Some in vitro assays have been developed to measure specific endpoints to
Conflict of interest
Eric I. Novik and Amit Parekh are employees of Hμrel Corporation. James MacDonald is a member of the Hμrel board of directors and Martin L. Yarmush plays the role of Chief Scientific Adviser at Hμrel Corporation.
Acknowledgments
We would like to thank Gaëlle Toussaint, Sylvie Dell'Aiera and Ingrid Van Wallendael for their technical assistance.
References (72)
- et al.
Kinetic cell-based morphological screening: prediction of mechanism of compound action and off-target effects
Chem. Biol.
(2009) - et al.
The use of real-time cell analyzer technology in drug discovery: defining optimal cell culture conditions and assay reproducibility with different adherent cellular models
J. Biomol. Screen.
(2011) - et al.
Drug-induced toxicity on mitochondria and lipid metabolism: mechanistic diversity and deleterious consequences for the liver
J. Hepatol.
(2011) - et al.
Predicting safety toleration of pharmaceutical chemical leads: cytotoxicity correlations to exploratory toxicity studies
Toxicol. Lett.
(2010) - et al.
Evaluation of a microfluidic based cell culture platform with primary human hepatocytes for the prediction of hepatic clearance in human
Biochem. Pharmacol.
(2009) - et al.
FDA-approved drug labeling for the study of drug-induced liver injury
Drug Discov. Today
(2011) - et al.
Selection of cytotoxicity markers for the screening of new chemical entities in a pharmaceutical context: a preliminary study using a multiplexing approach
Toxicol. in Vitro
(2009) - et al.
Using an in vitro cytotoxicity assay to aid in compound selection for in vivo safety studies
Bioorg. Med. Chem. Lett.
(2010) - et al.
The human hepatoma HepaRG cells: a highly differentiated model for studies of liver metabolism and toxicity of xenobiotics
Chem. Biol. Interact.
(2007) Toxicology and safety of COMT inhibitors
Int. Rev. Neurobiol.
(2010)