INTRODUCTION 147 2 MATERIALS AND METHODS

2.1. HepaRG cell culture ... 149 2.1.1. Differentiated 3D cultures (3D Diff) ... 149 2.1.2. Under-Differentiation 3D cultures (3D U-Diff) ... 150 2.2. Structural arrangement of the cell spheroids ... 151 2.3. Hepatocyte function assays. ... 151 2.3.1 Clinical chemistry indicators ... 151 2.3.2. Hepatocyte CYP450 enzymes activity. ... 152 2.3.2.1. LC-MS Analysis. ... 152 2.3.3. Uridine diphosphate glucuronoltransferase (UGT) activity. ... 154 2.4. Biocompetency assessment - Acetaminophen (APAP) Toxicity. ... 154 2.5. Cell based PBPK model ... 154 2.6. Calculations and statistical analysis ... 155 3. RESULTS ... 156 3.1 Hepatocytes function assays – clinical chemistry indicators (3Diff) ... 156 3.2. Structural arrangement of the cell spheroids (3D Diff) ... 156 3.3. Phase I and Phase II hepatocyte enzyme activities (3D Diff) ... 158 3.3.1 Functional activity of CYP450 enzymes ... 158 3.3.2 Long term characterization of biotransformation- determination of CYP3A4 and UGT activity ... 159 3.4. Biocompetency: APAP toxicity (3D Diff) ... 161 3.4.1. Simulated results ... 161 3.5. Decreasing differentiation time in 3D (3D U-Diff) ... 162 4. DISCUSSION ... 162 5. ACKNOWLEDGMENTS ... 168 6. REFERENCES ... 168

C h ap te r V 1. INTRODUCTION

It has become increasingly clear that the kinetic and metabolic fate of a compound has an important influence on its toxic potential, disposition in the body and eventual excretion (Coecke et al. 2006). Although in vitro and in silico human metabolic competent test systems are considered essential parts of integrated test strategies for systemic toxicities in general, metabolism is still deemed a bottleneck in in vitro toxicological test development. Therefore, in order to predictably screen the toxic compounds, it is essential to develop reliable and relevant human-based in vitro test systems that are metabolically competent and will model the hepatocytes process of biotransformation. The in vitro results obtained from such a system can be then further integrated by in silico computer models (such as physiologically-based toxicokinetic (PBTK) models) and converted into dose-response information for the entire organism enabling to assess the safety profile of compounds. This opens the possibility of using in silico strategies on kinetic data to bridge the in vitro and in vivo paradigm. However, more sophisticated in vitro models that maintain the liver function over a long time course for assessing drug metabolism and toxicity are currently needed to improve the predictive power of this approach.

To prove the metabolic competence of any metabolic competent in vitro system such as cell lines, liver slices, primary cells or stem cells, the presence and activity of phase I and phase II biotransformation enzymes should be evaluated (Coecke et al. 2006). Furthermore, it is important that the metabolic machinery demonstrates to be functional and the polar arrangement of the hepatocytes in vivo, with apical and biliary sides characterised by transporters distribution, is present.

Human hepatocyte cultures are considered to be the gold standard for testing liver toxicity since they better reflect what happens in the human liver. However, these cells are rather difficult to obtain and besides undergoing to spontaneously dedifferentiation, their profile is highly dependent on the donor and thus presents a high inter-donor variability. Moreover, they can keep their functionality in culture for a relatively short time.

A good alternative to human hepatocytes is represented by the HepaRG cells developed by Gripon et. al (Gripon et al. 2002). This is a human hepatoma cell line, derived from hepatocellular carcinoma that had shown liver-specific functions comparable to the human hepatocytes (Lubberstedt et al. 2011). After treatment with 2% (v/v) DMSO, the cells fully differentiate in a co-culture of biliar and hepatocyte- like cells expressing major cytochrome P450 (CYP450) enzymes, nuclear receptors, transporter proteins and transcription factors (Aninat et al. 2006; Kanebratt and Andersson 2008). These liver-specific functions can be maintained at high levels in for longer when compared to primary hepatic cultures. Since liver is a 3D organ, in which cells are maintaining cell-cell contacts important for their function and specific polarity, a lot of effort is undertaken to mimic the same conditions in vitro. It was already proven that human hepatocytes cultured in stirred tank bioreactors (spinner vessels) as 3D cultures produce tissue-like structures and maintain the liver-specific functions for longer and at higher level (Miranda et al. 2009; Leite et al. 2011; Tostões 2011; Tostões et al. 2011). Since the major application of toxicological models is in expensive screening of compounds, it is an added value that the developed method is compatible with existing standard laboratory equipment and high-throughput techniques (Schutte et al. 2011). Moreover, it takes into consideration factors important for toxicological approaches, eg. commercial availability of the system, easiness for routine approach, cost-effectiveness and avoiding the necessity to use matrixes or tubing that could lead to non-specific binding.

The aim of this study was to apply the 3D culture techniques in the stirred tank bioreactor to HepaRG cells, a high standard hepatic cell line. Besides characterization of phase I and II enzymes capabilities, the biocompetency of the developed system was assessed by testing the bioactivation of acetominophen (APAP). This assay also intended to test the applicability of the developed approach in toxicological assays, since APAP was activated to the toxic metabolite. Being one of the major cause of death by liver failure in US (Hawkins et al. 2007), APAP is taken as reference hepatotoxic compound for in vitro assays.

C h ap te r V

Experimental data have been analyzed using a process based modelling approach similar to PBTK modelling, but adapted to our cell-based assays with the aim of elucidating the main factors responsible of the in vitro behaviour.

Since the HepaRG differentiation is a quite long process (approx. 14 days) and not always complies with the fast drug screening tests, an approach tested previously for neuronal cells was evaluated to decrease the time of differentiation was evaluated (Serra et al. 2007).

2. MATERIALS AND METHODS