Characterization of "in vitro" and "in vivo" models for the investigation of hepatotoxicity

The liver is the primary site of drug metabolism and plays a major role in metabolism,
digestion, detoxification, and elimination of drugs and toxins from the body.
Consequently, drugs affect the liver more frequently than any other organ and place
the liver at increased risk for toxic damage. Drug-induced liver injury (DILI) is a
common cause of acute liver failure and the most frequent reason for the withdrawal
of approved drugs, representing a serious challenge for the pharmaceutical industry.
The risk of developing hepatotoxicity is not only due to the chemical properties of the
drug but also to environmental factors, pre-existing diseases and genetic factors,
leading to the classification into either predictable (high incidence) or unpredictable
(low incidence) hepatotoxicity. Drugs that produce predictable liver injury are
generally a result of direct liver toxicity of the parent drug or its metabolites. However,
the majority of adverse drug-induced hepatic events are unpredictable and the
underlying mechanisms are mostly unknown, but assumed to be either immunemediated
hypersensitivity reactions or idiosyncratic and are able to alter the
susceptibility to adverse events. In recent years mitochondrial dysfunction has been
recognized as β-oxidation of fatty acids, inhibition or uncoupling of the respiratory
chain, or through a primary effect on the mitochondrial genome.
One aim of this thesis was to investigate the juvenile visceral steatosis (jvs) mouse,
which is characterized by microvesicular steatosis of the liver and to impaired renal
reabsorption leading to systemic carnitine deficiency. The main focus was put on the
assessment of the hepatic toxicity of valproate, an antiepileptic drug known to induce
liver injury, and to investigate whether the underlying carnitine deficiency is a risk
factor for valproate-associated hepatotoxicity. Furthermore, in vitro studies using
several hepatic cell lines were performed to estimate the suitability as screening
systems for hepatic metabolism and CYP induction, and one study was conducted to
evaluate the hepatotoxic effect of the plant cimicifuga racemosa.
Initially we assessed the carnitine homeostasis and energy metabolism in carnitinedeficient
(jvs-/-) mice after cessation of carnitine substitution (Chapter 6). It is well
established that sufficient carnitine plasma and tissue levels in jvs mice can be
obtained by carnitine substitution, correcting carnitine deficiency. We studied the
kinetics of carnitine loss from plasma and tissue carnitine stores and markers of
energy metabolism after carnitine deprivation for a maximum of ten days. The total
carnitine concentrations in plasma, liver and skeletal muscle were significantly
decreased, whereas carnitine concentration decreased rapidly in plasma but much
slower in tissue. Deprivation of carnitine was also associated with a further drop in
the plasma β-hydroxybutyrate levels and hepatic fat accumulation.
In a second in vivo experiment (Chapter7) we investigated whether carnitine
deficiency is a risk factor for valproate-associated hepatotoxicity in jvs mice, and we
assessed the effects of valproate on carnitine plasma and tissue stores in these
mice. Therefore, we treated heterozygous jvs+/- and the corresponding wild type
mice with subtoxic oral doses of valproate for two weeks. Our study shows that jvs+/-
mice treated with VPA have impaired hepatic mitochondrial β-oxidation and
increased hepatic fat accumulation, findings associated with increased activities of
serum transaminases and alkaline phosphatase, and hepatocellular damage.
Furthermore, the effect of VPA treatment on the carnitine plasma and tissue stores
was much more dramatic in JVS+/- than in wild type mice, leading to additional and
substantial losses in the plasma and tissue carnitine pools. In conclusion, hepatic
toxicity of VPA was more pronounced in JVS+/- mice than in corresponding wild type
mice, and systemic carnitine deficiency can therefore be considered to be a risk
factor for hepatotoxicity associated with VPA.
In an in vitro study using hepatic cell lines (Chapter 8), drug-induced changes in the
activity of cytochrome P450 isoforms were assessed. Since the activity of most CYPs
can be regulated by induction and/or inhibition by specific drugs, and possibly
affecting the metabolism of other drugs or even their own metabolism, we
investigated the expression and induction of several CYP isozymes and the human
pregnane X receptor in immortalized human hepatocytes for their suitability as
screening systems for hepatic drug metabolism. Our investigations demonstrated that
hHepLT5 cells contain the main human CYP isozymes CYP1A2 and CYP3A4 which
are important for drug metabolism. Summarized, hHepLT5 cells appear therefore to
be a valuable alternative for primary human hepatocytes for studying pharmacological
and toxicological features of new drug entities.
The last described study (Chapter 9) was conducted to assess the hepatotoxicity of
cimicifuga racemosa in experimental animals in vivo, in hepatocyte cultures and in
isolated liver mitochondria. Ethanolic cimicifuga racemosa extract was administered
orally to rats and liver sections were analyzed for microvesicular steatosis by electron
microscopy. Tests for cytotoxicity, mitochondrial toxicity and apoptosis/necrosis were
performed using HepG2 cells, and mitochondrial toxicity was studied using isolated
rat liver mitochondria. The main findings in vivo and in vitro were hepatic
mitochondrial toxicity, as evidenced by microvesicular steatosis and inhibition of β-
oxidation, eventually resulting in apoptotic cell death. These findings suggest that
inhibition of β-oxidation is the initial hepatotoxic event of cimicifuga extract, which
eventually may result in apoptosis of the hepatocytes.