University of Groningen

Mechanisms and prevention of hepatocyte cell deathVrenken, Titia Eveline

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2008

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Vrenken, T. E. (2008). Mechanisms and prevention of hepatocyte cell death. [S.n.].

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 07-05-2021

. - . ..

Mechanisms and prevention of hepatocyte cell death

Titia Woudenberg-Vrenken

r

�

Research The research described in this thesis was performed at the Department of Gastroenterology and Hepatology, University Medical Center Groningen, University of Groningen, Groningen The Netherlands. The research group participates in the Graduate School for Drug Exploration (GUIDE).

The work described in this thesis was funded by the Dutch Digestive Foundation (Maag Lever Darm Stichting, Nieuwegein) through grant WS 02-13. Part of the research was supported by the Netherlands Organisation for Scientific Research (NWO), the Van Walree Fund of the Royal Netherlands Academy of Arts and Sciences (KNAW), and the J.K. de Cock foundation.

Financial support for publication of this thesis was kindly provided by - AstraZeneca B.V., Zoetermeer, The Netherlands

Graduate School for Drug Exploration (GUIDE), The Netherlands J.E. Jurriaanse Stichting, Rotterdam, The Netherlands Maag Lever Darm Stichting (MLDS), Nieuwegein, The Netherlands Nederlandse Vereniging voor Hepatologie (NVH), Haarlem, The Netherlands Novartis Pharma B.V., Arnhem, The Netherlands Rijksuniversiteit Groningen, Groningen, The Netherlands Sanofi-Aventis Netherlands B.V., Gouda, The Netherlands

Their contribution is gratefully acknowledged!

Paranimfen Manon Buist-Homan Hugo Vrenken

© 2008 T .E. Woudenberg-Vrenken All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means without written permission of the author and the publisher holding the copyright of the published articles.

Cover and page design: Titia Woudenberg-Vrenken Printed by Print Partners lpskamp, Enschede, The Netherlands

Stellingen behorend bij het proefschrift

Mechanisms and prevention of hepatocyte cell death

1. De bescherming van hepatocyten tegen apoptose ge'induceerd door het toxische galzout glycochenodeoxycholzuur door het anti-diabetische geneesmiddel metformine is afhankelijk van een functionerend fosfoinositide-3 kinase overlevingsroute. Dit proefschrift

2. De geneesmiddelen metformine en leflunomide bieden effectieve bescherming tegen galzout-ge'induceerde apoptose en necrose in hepatocyten en zijn daarom therapeutische opties die nader onderzocht dienen te warden. Dit proefschrift

3. Toediening van anti-oxidanten als therapie voor galzout-gemedieerde leverschade blijkt geen gunstig effect te hebben. Dit proefschrift

4. Hepatocyten afkomstig uit cholestatische en cirrotische levers zijn minder gevoelig voor apoptose veroorzaakt door verschillende stimuli. Georgiev et al. Gut 2007;56:121-128, Osawa et al. Hepatology 2005;42:1320-1328, Black et al. Journal of Hepatology 2004;40:942-951, Lieser et al. Gastroenterology 1998;115:693-701 en dit proefschrift

5. Het onderzoek naar mechanismen van celdood en de bescherming hiertegen zou meer moeten warden uitgevoerd in 'zieke' hepatocyten, omdat deze anders reageren op schade veroorzakende klinisch relevante factoren. Dit proefschrift

6. In een arm land zijn mensenrechten een onbetaalbare luxe. Absolute macht doodt absoluut- Per Ah/mark

7. lk zie het zo: wil je een regenboog, dan zul je de regen voor lief moeten nemen. Dolly Parton

8. Niet op het verstand komt het aan, maar op dat wat het leidt: hart en karakter. Fiodor Dostojevski

9. We are all travelers in the wilderness of this world, and the best that we find in our travels is an honest friend. Robert Louis Stevenson

10. Leven is reizen naar de grens van kennis en dan de sprang wagen. D.H. Lawrence

Ccnlralc U

Metlischc M

Bibi int heck C

Gruningcu G

CJ/n 0.

RIJKSUNIVERSITEIT GRONINGEN

Mechanisms and prevention of hepatocyte cell death

Proefschrift

ter verkrijging van het doctoraat in de Medische Wetenschappen

aan de Rijksuniversiteit Groningen op gezag van de

Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op

woensdag 11 juni 2008 om 14.45 uur

door

Titia Eveline Vrenken

geboren op 21 november 1978 te Harlingen

Ccntr.tlc

Mcdisc:he

Bibliothcck

Groningcn

u M C

G

Promotor

Prof. Dr. A.J. Moshage

Beoordelingscommissie

Prof. Dr. U.H.W. Beuers Prof. Dr. F. Kuipers Prof. Dr. G. Molema

Table of Contents

Chapter 1 General Introduction 7

Chapter 2 Metformin protects rat hepatocytes against bile acid-induced apoptosis 23

Chapter 3 The active metabolite of leflunomide, A77 1726, protects rat 37 hepatocytes against bile acid-induced apoptosis

Chapter 4 Reactive oxygen species are not essential for bile acid-induced cell death 51

Chapter 5 Carbon monoxide blocks oxidative stress-induced hepatocyte 63 apoptosis via inhibition of the p54 JNK isoform

Chapter 6 Isolation, culture and characterization of cholestatic hepatocytes: an in vitro approach to study cholestasis

Chapter 7 Summary and General Discussion

Chapter 8 Nederlandse Samenvatting

Curriculum Vitae and List of Publications

Acknowledgements/Dankwoord

77

87

97

103

107

Chapter 1

General Introduction

Chapter 1

The liver The liver is the second largest organ of the body after the skin, and is divided in two lobes, which are surrounded by connective tissue, the Glisson's capsule. The liver contains a variety of cell types. The majority of liver cells are parenchymal cells or hepatocytes. Hepatocytes require mutual contact to coordinate their physiological functions. Therefore, hepatocytes are positioned in plates, forming a radius around a hepatic venule (Fig 1 ). The

plate of hepatocytes connects at least at two sides to sinusoids via the space of Disse. A sinusoid consists of endothelial cells, Kupffer cells and stellate or fat­storing cells, and transports blood from the periphery of the liver lobule to the hepatic vein. Sinusoidal endothelial cells are a specific type of fenestrated

�'-----··" endothelial cells, which line the sinusoids of the liver. In addition, these endothelial cells are antigen­presenting cells involved in the immune response. Kupffer cells are part of the host defense system, and have a strong phagocytotic activity, removing large particles in particular. Furthermore, Kupffer cells are an important source of inflammatory mediators during infection and inflammation. Hepatic stellate cells are situated in the space of Oisse. Stellate cells have two functions: 1) storage of vitamin A in cytoplasmic lipid droplets, 2) production of extracellular matrix, in particular collagen type I (1, 2).

Figure 1. Schematic representation of the liver

Bile canaliculi are located between adjacent hepatocytes, forming a network within the hepatocyte plates. Bile is transported via the canals of Hering and interlobular ducts into the extrahepatic bile duct and common bile duct system. This network of branching ducts is called the biliary tree. Cholangiocytes, the epithelial cells of the bile duct, modify bile by secreting a hydrogen carbonate-rich fluid. They are also involved in bile acid recycling in the cholehepatic shunt pathway (1-3).

The various functions of the liver, executed mostly by hepatocytes, include metabolism of (toxic) compounds and drugs, bile production and secretion, (plasma) protein and lipid synthesis, metabolic function e.g. gluconeogenesis and subsequent storage of metabolites (2).

Bile formation and the enterohepatic circulation Bile is composed mainly of water (80%). Bile contains bile acids, phospholipids, cholesterol, glutathione (conjugates) and metabolites like bilirubin. Bile acids are synthesized in the liver and actively secreted into the bile canaliculi. Bile acids are conjugated to glycine or taurine in the liver for efficient transmembrane transport in the enterohepatic circulation. After passage through the small intestine, the majority of bile acids (90%) are efficiently reabsorbed at the terminal ileum and the remainder of bile acids is lost via the feces. This loss of bile acids is compensated by the de novo synthesis from cholesterol in hepatocytes producing primary bile acids. Two pathways for bile acid synthesis exist: 1. the neutral or classical pathway, 2. the acidic or alternative pathway, both yielding cholic acid (CA) and chenodeoxycholic acid (COCA). The majority of the de novo synthesized bile acids are produced via the neutral pathway. Both pathways require specific members of the cytochrome P450 (Cyp) family of enzymes for bile acid synthesis. In humans, the bile acid pool consists for almost one half of CA, and for one third of COCA. The primary bile acids are converted by bacteria in the intestines into the secondary bile acids deoxycholic acid (OCA) and lithocholic acid (LCA; 1, 2, 4, 5).

8

General Introduction

The enterohepatic circulation is the permanent shuttling of bile acids between the liver and the intestine (Fig 2). Hepatocytes take up bile acids (BA) and organic anions (OA) from the portal blood at the basolateral (or sinusoidal) membrane via members of two transporter families: 1. Na•ttaurocholate cotransporting polypeptide (NTCP), a member of the solute carrier family (SLC10), 2. organic anion transporting polypeptides (OATPs), members of the solute carrier organic anion transporter family (SLCO). NTCP is a sodium-dependent transporter, requiring a sodium gradient for the exchange of one bile acid for two sodium ions, and is highly specific for conjugated primary bile acids. OATPs function as sodium­independent transporters of unconjugated bile acids and organic anions. Inside the hepatocyte, bile acids are transferred from the basolateral to the canalicular membrane via either intracellular trafficking or vesicle-mediated trafficking. Bile acid secretion into bile at the canalicular (or apical) membrane requires ATP-binding cassette (ABC) transporters. The bile salt export pump (BSEP) is the major bile acid exporter, which transports univalent bile salts. The multidrug resistance-associated protein (MRP) type 2 is the primary transporter for organic anions like bilirubin glucuronides and glutathione conjugates, but is also able to excrete sulphated and glucuronidated bile acids. The transport of bile acids into bile is very critical for the enterohepatic circulation, being the rate-limiting step in bile formation and excretion by hepatocytes (4-8).

11A

"' "' "'

Figure 2. Enterohepatic circulation

Bile acid secretion into bile influences the vesicular transport of free cholesterol and phosphatidylcholine. Multidrug-resistance (MDR) protein type 3 (Mdr2 in rodents) flips phospholipids from the inner to the outer leaflet of the canalicular membrane, and is responsible for the delivery of phosphatidylcholine to bile. Cholesterol is excreted into bile by the heterodimeric ABC transporter ABCG5 and ABCG8. In this way, mixed micelles are formed consisting of bile acids, phosphatidylcholine and cholesterol, which are stored in the gallbladder (4-9).

Bile passes the biliary tree, where exchange of organic anions and proteins by cholangiocytes takes place, thus altering the composition of bile. The biliary tree is also the first place of bile acid absorption. Cholangiocytes in the bile duct actively absorb bile acids and shuttle them back to hepatocytes for re-excretion. This route is known as the cholehepatic shunt pathway. Therefore, cholangiocytes express the SLC10 member, apical sodium-dependent bile acid transporter (ASST) at the canalicular membrane, and MDR3, t­ASBT and the heteromeric organic solute transporters (Ost) a and [3 at the basolateral membrane. Inside cholangiocytes, transcellular bile acid transport is achieved by the ileal bile acid binding protein, IBABP (4, 5, 8, 9).

After passage through the biliary tree, bile enters the gallbladder. In response to food intake, the gallbladder contracts and bile is released into the small intestine. The absorption of bile acids in the ileum is accomplished by ASST in the enterocytes. Bile acids are then transferred intracellularly to the basolateral area by IBABP for secretion into the portal circulation by Osta/[3. In the portal blood, bile acids are bound to albumin and lipoproteins. Due to a conformational change of albumin in the proximity of the basolateral membrane of hepatocytes are bile acids released from albumin. Finally, hepatocytes extract the bile acids from the portal blood via NTCP and OATPs. The majority of conjugated bile acids are absorbed during one passage through the liver. Additional bile acid transporters exist that come into play when the enterohepatic circulation is not functioning properly. MRP3, like NTCP and OATPs, is located in the basolateral membrane of hepatocytes. However, MRP3 is involved in the reverse transport of bile acids, thus secreting bile acids and organic anions

9

Chapter 1

from the hepatocyte into the sinusoids (4, 5). The homeostasis of bile acids, cholesterol and triglycerides is regulated by members of the nuclear hormone receptor family such as farnesoid X receptor (FXR), liver X receptor (LXR) and pregnane X receptor (PXR). Bile acids directly bind to these nuclear receptors, which regulate the activity of hepatic transporter proteins and Gyp-enzymes involved in the de novo bile acid synthesis and metabolism (sulphation, glucuronidation) of bile acids. FXR is known to regulate bile acid synthesis and transport, LXR is mainly involved in cholesterol homeostasis and PXR regulates enzymes involved in the conjugation of bile acids. Hence, bile acid binding to FXR results in down-regulation of NTCP and ASBT, while BSEP, MRP2, IBABP and Osta/� are up-regulated (4, 5, 8).

Bile has multiple functions. Cholesterol is removed through the de novo synthesis of bile acids in the liver, and via the incorporation in mixed micelles. In the intestine bile is, also via the mixed micelles, essential for the absorption of lipids, cholesterol and lipid-soluble vitamins. In addition, bile acids activate signaling pathways involved in the electrolyte homeostasis and apoptosis in the epithelium of intestine and colon. Thus, bile and bile constituents are key players in bile formation and in the homeostasis of bile, lipids and cholesterol (1, 2, 4-9).

Cell death Disturbances of bile homeostasis can lead to cell injury and cell death, culminating in loss of organ (liver) function and eventually organ (liver) failure. Cell death can occur in at least two forms: apoptosis and necrosis, although intermediary forms (aponecrosis) and additional types of cell death (autophagy, entosis) have been reported.

Apoptosis, or programmed cell death, is an energy-dependent process, involving a specific family of cysteine proteases, the caspases. Apoptotic cell death is characterized by chromatin condensation, nuclear fragmentation and cell shrinkage, resulting in the fragmentation of the cell into apoptotic bodies. Apoptotic bodies are cleared by surrounding phagocytosing cells, thus minimizing inflammation. The caspase-family comprises fourteen members, which participate mainly in apoptosis and inflammation. Caspases are present as inactive precursor enzymes, the pro-caspases, that become activated via an interaction with caspase recruitment domains (CARD), death domains (DD) or death effector domains (OED) on various apoptosis-related proteins and death receptors. The pro-domain is cleaved from the pro-caspase, resulting in the active form consisting of a small and a large subunit. Based on their function in the apoptotic cascade, caspases are divided in two groups, initiator and effector caspases. Initiator caspases (caspases 8, 9 and 10) are involved in the initiation of apoptotic signaling. The initiator caspases are responsible for the activation of the effector caspases 3, 6 and 7, operating at the final stage of apoptosis. Effector caspases cleave essential cellular proteins involved in signal transduction, RNA synthesis, DNA-repair and structure, which ultimately results in the morphological modifications that characterize apoptotic cells. Apoptosis can be induced by a variety of agents, such as cytokines, oxidative stress, growth factor withdrawal or DNA damage, all resulting in the activation of caspases (10-15).

A distinction is made between the intrinsic pathway, resulting from an apoptotic trigger inside the cell, and the extrinsic pathway mediated by death receptor activation at the cell surface.

Intrinsic pathway The intrinsic pathway of apoptosis is also known as a stress pathway (Fig 3). A stress event can trigger the mitochondrial permeability transition (MPT), which is characterized by the loss of potential between the mitochondrial inner and outer membrane (intermembrane space), mitochondrial swelling and outer mitochondrial membrane rupture. As a consequence, the damaged mitochondria release a variety of pro-apoptotic proteins from the intermembrane space, like cytochrome c, second mitochondria-derived activator of caspases 10

General Introduction

(Smac)/direct inhibitor of apoptosis protein (IAP) binding protein with low pl (Diablo), Omi/high temperature requirement A2 (HtrA2) and apoptosis inducing factor (AIF). Cytochrome c binds to apoptotic protease activating factor 1 (Apaf-1 ), thereby facilitating the binding of {d)ATP to Apaf-1. Then, pro-caspase-9 is recruited by the CARD motif of Apaf-1, and simultaneously Apaf-1 assembles into a complex called the apoptosome, which consists of Apaf-1, cytochrome c, (d)ATP and procaspase-9. Inside the apoptosome, caspase-9 is activated via Apaf-1-regulated homodimerization, and cytochrome c is a critical cofactor in this process. Subsequently, caspase-9 activates the effector caspases pro-caspase-3 and -7. Although the apoptosome activates caspases, studies have shown that the apoptosome is an amplifier, rather than a crucial initiator, of apoptosis (11, 13, 16-22).

Extrinsic pathway The extrinsic or death receptor pathway of apoptosis is initiated by the ligation of death receptors (DR) at the cell membrane (Fig 3). Death receptors belong to the tumor necrosis factor/nerve growth factor (TNF/NGF) receptor super family, including TNF-R1, Fas and TNF-related apoptosis inducing ligand {TRAIL)-R1/DR4 and R2/DR5 In general, upon ligand binding, the death receptor monomers oligomerize at the cell surface, providing a docking platform at the cytoplasmic side, which facilitates the recruitment of adaptor proteins like Fas-associated death domain (FADD). An interaction is established via the death domains (DD) of the death receptors and the adaptor proteins, and subsequently, pro-caspase-8 and/or -10 are recruited to the death effector domain (OED) of FADD. The resulting complex, consisting of death receptors, FADD and pro-caspases is called the death inducing signaling complex {DISC). DISC formation results in the enzymatic activation of procaspase-8, and apoptosis is induced either via direct activation of caspase-3 (type I cells) or via a pathway involving the mitochondria (type II cells). In type II cells, like hepatocytes, active caspase-8 cleaves Bid, a pro-apoptotic member of the Bcl-2 family. Truncated-Bid (t-Bid) translocates to the mitochondria inducing the formation of pores in the mitochondrial outer membrane, hence stimulating the release of cytochrome c, the formation of the apoptosome and subsequently the induction of apoptosis (10, 23-25).

Intrinsic pathway

hepatocyte

Figure 3. Intrinsic and extrinsic signaling of apoptosis

Extrinsic pathway

apoptosis , ••• - . apoptosome

An exception to this common pathway is TNF-R1, which recruits the TNF receptor 1-associated protein, TRADD. TRADD performs a twofold function; apoptosis is induced by an interaction with FADD via the ODs, but TRADD also represses apoptosis by recruiting secondary adaptor proteins like receptor interacting protein (RIP) and TNF receptor­associated factor (TRAF) 2 and 5. The binding of RIP to TRADD activates kinases that eventually cause degradation of IKB resulting in the release from, and activation of the transcription factor nuclear factor (NF)-KB that regulates the expression of many anti­apoptotic genes. In hepatocytes, the net effect caused by the ligand TNFa favors cell

11

Chapter 1

survival rather than cell death. In contrast, the death receptors Fas and TRAIL-R1 and R2 which only recruit FADD, but not TRADD, do not or hardly activate NF-KB and thus always induce cell death. Some cell types also express a second TNF receptor, TNF-R2, which lacks the death domain and is therefore unable to bind adaptor proteins and induce signaling pathways. The liver is known to highly express the death receptors Fas, TRAIL-R1 and R2 and TNF-R1. Hepatocytes, cholangiocytes, activated stellate cells and endothelial cells express Fas, hepatocytes and endothelial cells express TNF-R1 and hepatocytes also express TRAIL-R1 and R2 ( 10, 23, 26).

In contrast to apoptotic cell death, necrotic cell death is a passive process, which via mitochondrial and cellular swelling results in plasma membrane rupture and subsequent release of the cellular content into the extracellular environment and circulation. Thus, an inflammatory response is triggered. Compared to apoptosis this mode of cell death is rather disorganized, and necrotic cells are not eliminated by phagocytosis. Usually, liver injury occurs as a combination of apoptosis and necrosis since these two mechanisms of cell death share common pathways involving the mitochondria (21, 22, 27).

Reactive oxygen species Reactive oxygen species (ROS) comprise free radicals (HO·), peroxides (e.g. H202) and oxygen ions such as superoxide anions (02 ·-). Under normal conditions, the majority of ROS in a cell are formed during oxidative phosphorylation in the mitochondria. Upon cellular stress, the levels of ROS can be enhanced dramatically, exceeding the capacity of the cell to detoxify ROS. This situation is known as oxidative stress and results in cellular damage. A variety of conditions can lead to increased ROS generation and oxidative stress, like ageing, compromised antioxidant status, metabolism of xenobiotics (e.g. acetaminophen), alcohol metabolism and inflammation. Hepatocytes contain various defense mechanisms against oxidative stress (Fig 4), including superoxide dismutases (SODs). The SOD family consists of mitochondrial manganese (Mn)-SOD, cytoplasmic copper, zinc (Cu, Zn)-SOD and extracellular (EC)-SOD. SOD family members catalyze the reduction of superoxide anions to hydrogen peroxide, which is subsequently converted into water by either catalase or glutathione peroxidase. In addition to water, catalase releases an oxygen molecule during the conversion of hydrogen peroxide. Glutathione peroxidase oxidizes glutathione, while reducing hydrogen peroxide into water. In contrast to the conversion into water, hydrogen peroxide can also react with ferrous ions, resulting in the production of the highly reactive

xanthme oxidase

l oxidative dmnage

DNA ,p1ds prole ns

Figure 4. Generation and metabolism of reactive oxygen species

hydroxyl radicals. Cells contain many other antioxidant systems, such as reduced glutathione (GSH), vitamins (particularly vitamin E) and uric acid. Pro-oxidant systems involve nicotinamide adenine di nucleotide phosphate (NADPH) oxidase, which generates superoxide anions while oxidizing NADPH and xanthine oxidase.

Since reactive oxygen species are generated in both acute and chronic liver diseases like (non)-alcoholic steatohepatitls ((N)ASH) and chronic cholestasis, ROS have been implicated to be key players in hepatocyte cell death in these disorders. One of the aims of this thesis was to determine the involvement of ROS in bile acid-induced cell death (28-36).

12

General Introduction

Signal transduction pathways Signal transduction pathways, including phosphoinositide-3 kinase (Pl3K), mitogen-activated protein kinases (MAPK), Src tyrosine kinases, NF-KB activating pathways, Bcl-2 family members and calcium signaling regulate cell death and survival in hepatocytes. This thesis analyzed the role of selected signal transduction pathways in hepatocyte cell death, which are summarized below.

Phosphoinositide-3 kinase family The Pl3K family comprises a large group of structurally related enzymes that phosphorylate inositol lipids at specific sites. Members of the Pl3K family are involved in a broad range of cellular functions such as cell survival, cell growth and proliferation, insulin-induced responses and cellular trafficking. The Pl3K enzymes are arranged in three classes (I, I I and Ill) based on their substrate specificity, and each class is involved in particular functions. Class I is sub-divided into IA and 18. Members of the class IA family play a role in cell survival and proliferation, and 18 subtypes participate in allergy and inflammation. Class II enzymes are responsible for lipid binding and protein interactions, and class Ill members are involved in intracellular trafficking. Protein kinases, phospholipases and G-proteins are downstream targets of Pl3K enzymes. An important Pl3K class I-effector in cell survival is Akt or protein kinase B (PKB), a serine/threonine kinase. Akt activation requires phosphorylation at Thr308 and Ser473, which occurs at the plasma membrane. Activated Akt protects from apoptosis via phosphorylation and inhibition of the pro-apoptotic Bcl-2 family member Bad, and the initiator caspase-9, and by enhanced degradation of IKB kinase (IKK), and subsequent activation of NF-KB. Another Akt target is the mammalian target of rapamycin (mTOR), which is a key player in transcription, translation, cytoskeletal arrangement and protein degradation. The important role of Akt in cell survival has been demonstrated: a constitutively active form of Akt inhibited apoptosis induced by various agents, like UV radiation, DNA damage and cytokines, whereas an inactive form of Akt promoted cell death (27, 37-47).

Mitogen-activated protein kinases MAPK members are involved in cell differentiation, proliferation, apoptosis and cell survival. All members become activated by a sequence of phosphorylation events, starting with a MAPK kinase kinase (MAPKKK), followed by a MAPK kinase (MAPKK) and finally the MAPK is activated via twofold phosphorylation at a threonine and tyrosine residue. The MAPK family consists of five groups: a) extracellular signal regulated kinase (ERK) 1 and 2; b) Jun N-terminal kinase (JNK) 1, 2 and 3; c) p38 a, B, y, 6; d) ERK3 and 4; e) ERK5 (27, 48, 50-51 ). ERK1/2 is activated by growth factors, cytokines and phorbol esters, and is an important regulator of cell proliferation, while only a minor role of ERK1/2 is implicated in apoptosis. Main targets of activated ERK1/2 are other kinases, transcription factors, e.g. NF-KB, cytoskeletal proteins and the anti-apoptotic Bcl-2 family member Bcl-2. The exact outcome of the activation of ERK1/2 depends on the length and strength of ERK1/2 activation (27, 48-51 ). JNK activity is induced by various stress stimuli such as UV irradiation, cytokines, oxidative stress or growth factors. JNK participates in the stress response, inflammation and apoptosis by stimulating c-Jun-dependent transcription and phosphorylation of downstream targets, including other kinases and apoptosis-modulating proteins like Itch (52, 53). Most stimuli like environmental stress and cytokines, which activate JNK, also activate p38. The activity of p38 is crucial for the regulation of the immune system and inflammation, thus explaining the involvement of p38 in asthma and autoimmune diseases. Targets of p38 are transcription factors, genes coding for inflammatory cytokines and downstream kinases (27, 48-51 ). ERK3 and 4 are so-called atypical MAPK members, they lack the threonine and tyrosine residues which become phosphorylated by upstream kinases. In contrast, ERK3 and 4 are activated via auto-phosphorylation (48).

1 3

Chapter 1

Growth factors activate ERK5, which is essential for angiogenesis (48).

Src tyrosine kinase family The Src family consists of 9 members, all of them containing specific Src homology (SH) domains and a tyrosine kinase domain. Src kinases phosphorylate specific target proteins (e.g. Pl3-kinase) at a tyrosine residue. Tyrosine phosphorylation of proteins is involved in numerous signal transduction mechanisms. Therefore, Src kinases are implicated in many cellular processes, including gene transcription, apoptosis and differentiation. Recently, Src family members were shown to be important in cytotoxicity induced by oxidative stress, Fas ligand (Fasl) and bile acids, since specific inhibition of Src kinases reduced hepatocyte cell death induced by these stimuli (32, 54-60).

NF-KB signaling The NF-KB subunit family is a key player in cell survival, inflammation and stress, and contains five members; p65 (RelA), RelB, c-Rel, p50/p105 (NF-KB1 ), and p52/p100 (NF­KB2). NF-KB members are present in the cytoplasm of cells as inactive homo- or heterodimers. The most common heterodimer is composed of a p65 and p50 subunit. NF-KB proteins can be activated by various signal transduction pathways of which the classical or canonical pathway and the alternative or non-canonical route are best described. The classical pathway can be triggered by a variety of cell surface receptors such as TNF­receptors and Toll-like receptors, while the alternative pathway only becomes activated by specific receptors, for instance lymphotoxin B (11 , 27, 61 -63). Since TNF-R1 activates the classical pathway, this signal transduction pathway will be described in more detail. In the classical pathway, inactive NF-KB is bound to one of the inhibitory IKB isoforms. Upon receptor activation, IKB becomes phosphorylated via sequential activation of adapter proteins like IKK�, and as a consequence IKB is released from the p65 subunit of NF-KB. On the p65 subunit, a nuclear localization sequence is exposed and NF-KB translocates to the nucleus. In the nucleus, NF-KB induces transcription via the interaction with KB binding sites in the regulatory sequences of target genes (62, 63). Many signal transduction molecules, including JNK, Pl3K/Akt and protein kinase C (PKG) modulate NF-KB activation. NF-KB effector genes include cytokines, superoxide dismutase (SOD), inducible nitric oxide synthase (iNOS) and anti-apoptotic members of the Bcl-2 family. The Bcl-2 family consists of pro-apoptotic members (e.g. Bad, Bax and Bak), as well as anti-apoptotic members (e.g. Bcl-2, Bel-xi and Bfl-1 or A 1 ). NF-KB specifically induces the expression of the anti-apoptotic proteins Bcl-2, Bel-xi and A 1/Bfl-1 (1 1 , 27, 61-68).

Liver diseases Bile acids, reactive oxygen species and cytokines are key players in liver diseases. Bile acids are involved in acute and chronic cholestatic liver disorders. Cholestasis is described as impaired bile flow, and symptoms are jaundice, pruritus and fatigue. The disease is initiated by a genetic, drug-induced, immunologic or mechanical insult to the bile ducts or hepatocytes. Progression of cholestasis occurs via hepatocytes, inflammatory cells (Kupffer cells, neutrophils) and stellate cells, due to the release of cytokines and ROS, and deposition of extracellular matrix. Cholestasis develops via fibrosis into cirrhosis and eventually end-stage liver failure (34, 69, 70). The most common genetic liver disorders associated with cholestasis are progressive familial intrahepatic cholestasis (PFIC) type 1, 2 and 3. These hereditary diseases result in impaired bile flow due to mutations in transporter proteins (putatively) involved in bile formation. PFIC type 1 or Byler's disease is caused by mutations in the FIC1 gene. FIC1 is an ATP-dependent protein pump at the canalicular membrane of hepatocytes and cholangiocytes. FIC1 is direct or indirect involved in the bile formation, since defects in FIC1 result in impaired transport of bile acids in the liver and the intestine. Mutations in the BSEP gene cause PFIC type 2 or Byler syndrome. Different BSEP mutations lead to different outcomes, varying from disturbed BSEP translocation to the canalicular membrane, abrogated bile acid transport by BSEP or no effect on either transport or trafficking. PFIC type 3 or MDR3 deficiency is characterized by the absence of MDR3, 14

General Introduction

and thus the lack of phospholipids in bile. Bile acids are no longer captured in mixed micelles, and the toxic bile acids can induce severe injury (4, 5, 69, 70).

Drug-induced cholestasis can also be related to BSEP. Certain drugs, such as cyclosporine A and rifampicin, inhibit BSEP, which results in bile acid accumulation and liver injury. Other drugs known to induce cholestasis include oral contraceptives, penicillin and anabolic steroids. Cholestasis related to the immune system can be caused by bacterial or viral infections, and is very often associated with sepsis or allograft rejection. Inflammation is also the cause of bile duct destruction in the auto-immune disorder primary biliary cirrhosis (PBC). Mechanical obstruction of the bile ducts may be due to gall stones or tumors. However, in some cholestatic disorders like biliary atresia and primary sclerosing cholangitis (PSC) the bile ducts are damaged due to a yet unidentified reason (4, 5, 69-72).

In cholestatic patients and animal models of cholestasis like bile duct ligation (BDL) or cholate-feeding, the bile acid retention in hepatocytes affects several transporters involved in the enterohepatic circulation. NTCP and OATP1 at the basolateral membrane and MRP2 at the canalicular membrane are down-regulated, while MRP3 is up-regulated , and BSEP is unaffected. Furthermore, the glycine-conjugates of hydrophobic bile acids trigger hepatocyte apoptosis and necrosis in cholestasis. Increasing bile acid levels also induce proliferation of cholangiocytes, which via the release of pro-inflammatory cytokines results in the death of the cholangiocytes (4, 5, 9, 73-75).

Ursodeoxycholic acid (UDCA) is a dihydroxy bile salt which is used in the treatment of chronic cholestatic liver diseases. The hepatoprotective mechanism of UDCA is accomplished via upregulation of the basolateral bile acid transporters and stimulation of bile acid secretion of hepatocytes. In addition, activation of anti-apoptotic pathways by UDCA or its taurine conjugate TUDCA has been described. In cholangiocytes, UDCA and TUDCA inhibit the bile acid-induced proliferation and loss of cholangiocytes. However, cholestatic disorders do not always respond to UDCA-treatment, and often surgery is required to drain or eliminate the obstruction. As a final option, when complications occur, a liver transplantation is required (75-81 ).

Reactive oxygen species are generated in both non-alcoholic fatty liver disease (NAFLD) and alcoholic liver diseases. NAFLD comprises a group of diseases varying from triglyceride accumulation in hepatocytes (steatosis), via hepatic steatosis accompanied by inflammation (steatohepatitis), into excessive accumulation of connective tissue (fibrosis) and finally to cirrhosis and liver failure. Steatosis or fatty liver is associated with the metabolic syndrome, insulin resistance, drugs and nutritional factors. Non-alcoholic steatohepatitis (NASH) is becoming a serious problem in countries with a Western life style, due to obesity, excessive caloric intake and lack of exercise. NASH is believed to develop in two sequential steps. First, accumulation of fat in hepatocytes leads to increased mitochondrial and peroxisomal B­oxidation, resulting in ROS generation and oxidative stress. Second, fat accumulation results in inflammation, increased production of cytokines like TNFa and cell death. NASH progresses into fibrosis in approximately 20% of the patients. The most effective treatment in fatty liver diseases is weight loss, thus removing the condition that causes the disease. Medications with a beneficial effect are the anti-diabetic drug metformin and in some cases UDCA. Alcoholic liver diseases resemble NASH, although the type and pattern of fat accumulation is different. The injury in alcoholic liver diseases is caused by the metabolism of alcohol (ethanol), which generates ROS and reactive metabolites like acetaldehyde. Acetaldehyde is a pro-fibrogenic factor. Apoptosis appears to be the dominant mode of cell death. Indeed, in patients with alcoholic hepatitis, marked numbers of apoptotic hepatocytes were detected (82-84 ).

Cytokines participate in all liver diseases, but play a predominant role in acute liver diseases like acute hepatitis and acute liver failure (ALF). Acute liver injury induces the rapid

1 5

Chapter 1

generation of pro-inflammatory cytokines and reactive oxygen species which results in hepatocyte cell death, mainly apoptosis. This inflammation-induced apoptosis of hepatocytes and loss of functional liver cells plays an important role in the augmentation of liver injury and ultimately liver failure. Hepatitis is most often the result of an infection with the hepatitis B, C, D or E virus, while an important trigger for acute liver failure is drug-induced toxicity (e.g. acetaminophen overdose) or mushroom poisoning (85-86).

From this overview it becomes clear that the majority of chronic liver diseases progress into cirrhosis and end-stage liver failure, demanding a liver transplantation as only possible treatment. Since chronic liver diseases are characterized by the progressive loss of hepatocytes, the functional liver cells, it is important to investigate the signaling pathways involved in cell survival and death in these conditions. This will allow the identification of potential targets for intervention. In addition, the effects of existing drugs on signal transduction pathways involved in hepatocyte survival needs to be explored. In this thesis, two of these drugs, metformin and leflunomide were investigated in detail.

Metformin Metformin is a drug used in the treatment of Diabetes Mellitus type I I , where it improves insulin sensitivity. Metformin reduced aminotransferase levels and decreased liver size in patients with NASH and in animal models for steatosis. Moreover, metformin treatment improved hepatocyte viability in fatty livers. However, the protective mechanism of metformin in patients or experimental animals with diabetes or fatty livers is unknown. Metformin is known to stimulate the activity of AMP-activated protein kinase (AMPK) family members a1 and a2. AMPK is involved in energy homeostasis, and becomes activated upon a rise in cellular AMP or changes in the AMP/ATP ratio. Furthermore, AMPK can be activated by stimuli not changing the AMP/ATP ratio, like oxidative stress, hypoxia, hyperosmotic stress or pharmacological compounds. AMPK activity represses signaling via mTOR, a downstream target of Akt and Pl3K. Based on studies describing protective effects of metformin in NAFLD, we investigated the hepatoprotective effect of metformin in this thesis using two different in vitro models for liver damage: 1) the toxic, hydrophobic bile acid glycochenodeoxycholic acid (GCDCA), as a model for acute bile acid toxicity, resulting in both apoptosis and necrosis; 2) the cytokine TNFa in combination with the transcriptional inhibitor actinomycin D (to inhibit NF-KB activation), a model for acute liver failure, which is characterized by apoptosis (87-98).

Leflunomide Leflunomide (N-[(4-trifluoromethyl)phenyl]-5-methylisoxazole-4-carboxamide) is a well­known anti-inflammatory agent, as therapy for autoimmune diseases, particularly rheumatoid arthritis. Furthermore, leflunomide is used in the prevention of rejection after organ transplantation, and in the treatment of Crohn's disease and cancer. After administration, the prodrug leflunomide, is almost completely metabolized (>95%) to its active compound A77 1726 or teriflunomide (2-cyano-3-oxo-N-[(4-trifluoromethyl)phenyl]butyramide). Proposed molecular mechanisms explaining the anti-inflammatory, anti-proliferative and immunosuppressive effects of leflunomide are inhibition of protein tyrosine kinases, e.g. kinases of the Src family, blocking dihydroorotate dehydrogenase (DHODH) involved in pyrimidine synthesis and progression through the cell cycle, inhibition of MAP kinases, including JNK and reduced production of pro-inflammatory cytokines via inhibition of NF-KB. In experimental models, leflunomide was protecting liver cells against injury induced by acetaminophen, concanavalin A (model for hepatitis), bile duct ligation and ischaemia/reperfusion. The protective mechanisms observed were inhibition of cytochrome c release from the mitochondria and caspase activities, reduced plasma cytokine concentrations and aminotransferase levels. In this thesis we studied the potential protective mechanisms of leflunomide in three different in vitro models of cell death: 1) GCDCA, 2) TN Fa and actinomycin D, 3) the superoxide anion donor menadione as a model of oxidative stress-induced apoptosis (99-110). 1 6

General Introduction

Scope of the thesis The aim of this thesis was to unravel the mechanisms of hepatocyte cell death induced by factors present in liver diseases: bile acids, inflammatory mediators and oxidative stress. For this purpose, we used three different in vitro models of hepatocyte injury, representing various inducers of cell death relevant for liver diseases: bile acid-induced toxicity, as model for cholestatic diseases; 2. cytokine induced injury, representing acute hepatitis and acute liver failure; 3. reactive oxygen species, representative for many chronic liver diseases accompanied by oxidative stress, such as alcoholic and non-alcoholic steatohepatitis. Furthermore, we investigated the effect and mechanism of action of various potential drugs against liver diseases. In chapter 2, we examined the protective role of the anti-diabetic drug metformin against hepatocyte cell death. Metformin was shown to protect primary rat hepatocytes against bile acid-induced apoptosis. The protective effect was dependent on the Pl3K/Akt-signaling pathway. In chapter 3, the cytoprotective effects of leflunomide, an anti­inflammatory compound, are described. Leflunomide protects hepatocytes against bile acid­induced apoptosis and necrosis, but the drug does not protect against cytokine or oxidative stress-induced cell death. In addition, we studied the mechanism involved in the protection of leflunomide. In Chapter 4 we studied the involvement of reactive oxygen species in bile acid-induced hepatocyte cell death. We report that reactive oxygen species are not required for bile acid-induced hepatocyte apoptosis, but Src-kinase family members are critically involved in bile acid-induced cell death. Chapter 5 describes the role of the stress­responsive protein H0-1 in in vivo and in vitro models of liver injury. The protective effect of H0-1 and carbon monoxide, one of its products, against superoxide anion-induced hepatocyte death were examined. Chapter 6 describes an in vitro approach to study cholestasis. We examined the isolation and culture of cholestatic hepatocytes in detail and demonstrate that cholestatic hepatocytes retain their in vivo phenotype during in vitro experiments. Thus, we provide a tool to study the signaling pathways occurring in hepatocytes during cholestatic liver diseases in detail. In conclusion, this thesis reports about the prevention and mechanisms of hepatocyte cell death in liver diseases.

17

Chapter 1

Reference List

(1 )

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(1 1 )

(12)

(13)

(14)

( 15)

(16)

(17)

( 18)

(19)

(20)

(21 )

(22)

(23)

(24)

(25)

(26)

18

Young B, Stevens A, Heath JW, Lowe JS, Lowe J. Wheater's Functional Histology: A text and colour

atlas. Elsevier Health Sciences. Fourth Edition. 2000:274-282.

Junqueira LC, Carneiro J, Wisse E. Functlonele Histologie. Elsevier Gezondheidszorg. Tweede Druk.

1 982:376-393.

Boyer JL. Bile duct epithelium: frontiers in transport physiology. Am J Physiol 1 996 Jan:270(1 Pt 1 ):G1-

G5.

Kullak-Ublick GA, Stieger B, Meier PJ. Enterohepatic bile salt transporters in normal physiology and liver

disease. Gastroenterology 2004 Jan;1 26(1 ):322-342.

Alrefai WA, Gill RK. Bile acid transporters: structure, function, regulation and pathophysiological

implications. Pharm Res 2007 Oct:24(1 0):1 803-1823.

Mil ls CO, Milkiewicz P, Muller M, Roma MG, Havinga R, Coleman R, et a l . Different pathways of

canalicular secretion of sulfated and non-sulfated fluorescent bile acids: a study ln isolated hepatocyte

couplets and TR- rats. J Hepatol 1 999 Oct:31 (4):678-684

Kouzuki H, Suzuki H, Ito K, Ohashi R, Sugiyama Y. Contribution of sodium taurocholate co-transporting

polypeptide to the uptake of its possible substrates into rat hepatocytes. J Pharmacol Exp Ther 1 998

Aug:286(2):1043-1050.

Takikawa H. Hepatobiliary transport of bile acids and organic anions. J Hepatob1liary Pancreat Surg

2002;9(4 ):443-447.

Rippin SJ, Hagenbuch B, Meier PJ, Stieger B. Cholestatic expression pattern of sinusoidal and

canalicular organic anion transport systems in primary cultured rat hepatocytes. Hepatology 2001

Apr;33(4 ):776-782.

Reed JC. Mechanisms of apoptosis. Am J Pathol 2000 Nov;1 57(5):1 4 1 5-1430.

Adams JM. Ways of dying: multiple pathways to apoptosis. Genes Dev 2003 Oct 1 5;1 7(20):2481 -2495.

Mehmet H. Caspases find a new place to hide. Nature 2000 Jan 6,403(6765):29-30.

Adrain C, Martin SJ. The mitochondrial apoptosome: a killer unleashed by the cytochrome seas. Trends

Biochem Sci 2001 Jun;26(6):390-397.

Cohen GM. Caspases: the executioners of apoptosis. Biochem J 1 997 Aug 1 5;326(Pt 1 ):1 - 16.

Gewies A. Aporeview: Introduction to apoptosis. 2003:1-26.

Slee EA, Harte MT, Kluck RM, Wolf BB, Casiano CA, Newmeyer DD, et al. Ordering the cytochrome c­

initiated caspase cascade: hierarchical activation of caspases-2, -3, -6, -7, -8, and - 1 0 in a caspase-9-

dependent manner. J Cell Biol 1 999 Jan 25;144(2):281 -292.

Adams JM, Cory S. Apoptosomes. engines for caspase activation. Curr Opin Cell Biol 2002

Dec:14(6):71 5-720.

Kim JS, He L, Lemasters JJ. Mitochondrial permeability transition: a common pathway to necrosis and

apoptosis. Biochem Biophys Res Commun 2003 May 9;304(3):463-470.

Zamzami N, Kroemer G. Apoptosis: mitochondrial membrane permeabilizatlon--the (w)hole story? Curr

Biol 2003 Jan 2 1 ; 1 3(2):R71-R73.

van Gurp M, Festjens N, van Loo G, Saelens X, Vandenabeele P . Mitochondrial intermembrane proteins

in cell death. Biochem Biophys Res Commun 2003 May 9;304(3):487-497.

Lemasters JJ. Mechanisms of hepatic toxicity - V. Necrapoptosis and the mitochondrial permeability

transition: shared pathways to necrosis and apoptosis. American Journal of Physiology-Gastrointestinal

and Liver Physiology 1 999 Jan:276(1 ):G1-G6.

Lemasters JJ, Qian T, He LH, Kim JS, Elmore SP, Cascio WE, et al . Role of mitochondrial inner

membrane permeabilization in necrotic cell death, apoptosis, and autophagy. Antioxidants & Redox

Signaling 2002 Oct:4(5):769-781 .

Yoon JH, Gores GJ. Death receptor-mediated apoptosis and the liver. J Hepatol 2002 Sep;37(3):400-

410.

Peter ME, Krammer PH. The CD95(APO-1/Fas) DISC and beyond. Cel l Death Differ 2003 Jan : 10(1 ):26-

35.

Higuchi H, Gores GJ. Bile acid regulation of hepatic physiology: IV. BIie acids and death receptors. Am J

Physiol Gastrointest Liver Physiol 2003 May;284(5):G734-G738.

Varfolomeev EE, Ashkenazi A. Tumor necrosis factor: an apoptosis JuNKie? Cell 2004 Feb

20;1 1 6(4):491-497.

General Introduction

(27) Schoemaker MH, Moshage H. Defying death: the hepatocyte's survival kit. Clin Sci (Land) 2004

Jul;107(1 ):13-25.

(28) Chu SH, Lee-Kang J, Lee KH, Lee K. Roles of reactive oxygen species, NF-kappaB, and peroxiredoxins

in glycochenodeoxycholic acid-induced rat hepatocytes death. Pharmacology 2003 Sep;69(1 ):12-19.

(29) Park SY, Chang I , Kim JY, Kang SW, Park SH, Singh K, et al. Resistance of mitochondrial DNA­

depleted cells against cell death: role of mitochondrial superoxide dismutase. J Biol Chem 2004 Feb

27;279(9):7512-7520.

(30) Giannoni E, Buricchi F, Raugei G, Ramponi G, Chiarugi P. Intracellular reactive oxygen species activate

Src tyrosine kinase during cell adhesion and anchorage-dependent cell growth. Molecular and Cellular

Biology 2005 Aug;25(15):6391-6403.

(31) Reinehr R, Becker S, Braun J, Eberle A, Grether-Beck S, Haussinger D. Endosomal acidification and

activation of NADPH oxidase isoforms are upstream events in hyperosmolarity-induced hepatocyte

apoptosis. Journal of Biological Chemistry 2006 Aug 11 ;281 (32):23150-23166.

(32) Conde de la Rosa L, Schoemaker MH, Vrenken TE, Buist-Homan M, Havinga R, Jansen PL, et al.

Superoxide anions and hydrogen peroxide induce hepatocy1e death by different mechanisms:

involvement of JNK and ERK MAP kinases. J Hepatol 2006 May;44(5):918-929.

(33) Czaja MJ. Induction and regulation of hepatocyte apoptosis by oxidative stress. Antioxidants & Redox

Signaling 2002 Oct:4(5):759-767.

(34) Jaeschke H, Gores GJ, Cederbaum Al, Hinson JA, Pessayre D, Lemasters JJ. Forum - Mechanisms of

hepatotoxicity. Toxicological Sciences 2002 Feb;65(2):166-176.

(35) Li JM, Shah AM. ROS generation by nonphagocytic NADPH oxidase: potential relevance in diabetic

nephropathy. J Am Soc Nephrol 2003 Aug;14(8 Suppl 3):S221-S226.

(36) Manni ML, Oury TD. Commentary on "Copper chaperone for Cu,Zn-SOD supplement potentiates the

Cu,Zn-SOD function of neuroprotective effects against ischemic neuronal damage in the gerbil

hippocampus". Free Radie Biol Med 2007 Sep 15;43(6):899-900.

(37) Foster FM, Traer CJ, Abraham SM, Fry MJ. The phosphoinositide (Pl) 3-kinase family. J Cell Sci 2003

Aug 1 ;116(Pt 15):3037-3040.

(38) Taker A. Protein kinases as mediators of phosphoinositide 3-kinase signaling. Molecular Pharmacology

2000 Apr;57(4 ):652-658.

(39) Tee AR, Blenis J. mTOR, translational control and human disease. Seminars in Cell & Developmental

Biology 2005 Feb;16(1):29-37.

(40) Hay N, Sonenberg N. Upstream and downstream of mTOR. Genes & Development 2004 Aug

15;18(16):1926-1945.

(41) Datta SR, Brunet A, Greenberg ME. Cellular survival: a play in three Akts. Genes Dev 1999 Nov

15;13(22):2905-2927.

(42) Carrera AC. TOR signaling in mammals. Journal of Cell Science 2004 Sep 15;117(20):4615-4616.

(43) Proud CG. The multifaceted role of mTOR in cellular stress responses. Ona Repair 2004 Aug;3(8-

9 ):927-934.

(44) Hahn-Windgassen A, Nogueira V, Chen CC, Skeen JE, Sonenberg N, Hay N. Aki activates the

mammalian target of rapamycin by regulating cellular ATP level and AMPK activity. J Biol Chem 2005

Sep 16;280(37):32081-32089.

(45) Cantley LC. The phosphoinositide 3-kinase pathway. Science 2002 May 31 ;296(5573):1655-1657.

(46) Brazil DP, Yang ZZ, Hemmings BA. Advances in protein kinase B signalling: AKTion on multiple fronts.

Trends in Biochemical Sciences 2004 May;29(5):233-242.

(47) Huang SL, Bjornsti MA, Houghton PJ. Rapamycins - Mechanism of action and cellular resistance.

Cancer Biology & Therapy 2003 May;2(3):222-232.

(48) Qi MS, Elion EA. MAP kinase pathways. Journal of Cell Science 2005 Aug 15;118(16):3569-3572.

(49) Roux PP, Blenis J. ERK and p38 MAPK-activated protein kinases: a family of protein kinases with

diverse biological functions. Microbial Mal Biol Rev 2004 Jun:68(2):320-344. (50) Seger R, Krebs EG. The MAPK signaling cascade. FASEB J 1995 Jun;9(9):726-735.

(51) Schaeffer HJ, Weber MJ. Mitogen-activated protein kinases: specific messages from ubiquitous

messengers. Mal Cell Biol 1999 Apr:19(4):2435-2444.

(52) Liedtke C, Trautwein C. The role of JNK2 in toxic liver injury. J Hepatol 2006 Nov;45(5):762-764.

19

Chapter 1

20

(53) Chang L, Kamala H, Solinas G, Luo JL, Maeda S, Venuprasad K, et al. The E3 ubiquitin llgase Itch

couples JNK activation to TNFalpha-induced cell death by inducing c-FLIP(L) turnover. Cell 2006 Feb

1 0;124(3):601-613 .

(54) Reinehr R, Becker S, Hengen A, Haussinger D. The Src family kinase Yes triggers hyperosmotic

activation of the epidermal growth factor receptor and CD95. J Biol Chem 2004 Jun 4;279(23):23977-

23987.

(55) Reinehr R, Becker S, Wettstein M, Haussinger D. Involvement of the Src family kinase yes in bile salt­

induced apoptosis. Gastroenterology 2004 Nov;127(5) : 1540-1 557.

(56) Reinehr R, Becker S, Eberle A, Grether-Beck S, Haussinger D. Involvement of NADPH oxidase

isoforms and Src family kinases In CD95-dependent hepatocyte apoptosis. Journal of Biological

Chemistry 2005 Jul 22;280(29):271 79-271 94.

(57) Yoshizumi M, Abe J, Haendeler J , Huang Q, Berk BC. Src and Cas mediate JNK activation but not

ERK1/2 and p38 kinases by reactive oxygen species. J Biol Chem 2000 Apr 21 ;275(1 6):1 1 706-1 1 71 2.

(58) Abe J, Takahashi M, Ishida M, Lee JD, Berk BC. c-Src is required for oxidative stress-mediated

activation of big m1togen-activated protein kinase 1 . J Biol Chem 1 997 Aug 1 5;272(33):20389-20394.

(59) Pandey A, Peri S, Thacker C, Whipple CA, Collins JJ, Mann M. Computational and experimental

analysis reveals a novel Src family kinase in the C. elegans genome. Biolnformatics 2003 Jan

22;1 9(2): 1 69-1 72.

(60) Tatosyan AG, Mizenina OA. Kinases of the Src family: structure and !unctions. Biochemistry (Mose )

2000 Jan;65(1 ):49-58.

(61 ) Schoemaker MH, Moshage H Die another day. J Hepatol 2003 Jun;38(6):873-875.

(62) Hayden MS, Ghosh S. Signaling to NF-kappaB. Genes Dev 2004 Sep 15 ; 18(1 8):21 95-2224.

(63) Gilmore TD. Introduction to NF-kappaB: players, pathways, perspectives. Oncogene 2006 Oct

30;25(51 ):6680-6684.

(64) Duckett CS. Apoptosis and NF-kappa B: the FADD connection. J Clin Invest 2002 Mar;1 09(5):579-580.

(65) Kaltschmidt B, Kaltschmidt C, Hofmann TG, Hehner SP, Drage W, Schmitz ML. The pro- or anti­

apoptotic function of NF-kappa B Is determined by the nature of the apoptotic stimulus. European

Journal of Biochemistry 2000 Jun,267(12):3828-3835.

(66) Luedde T, Trautwein C. Intracellular survival pathways m the liver. Liver Int 2006 Dec;26(1 0): 1 1 63-1 1 74.

(67) Tamatan i M, Che YH, Matsuzaki H, Ogawa S, Okada H. Miyake S, et al. Tumor necrosis factor induces

Bcl-2 and Bcl-x expression through NFkappaB activation in primary hlppocampal neurons. J Biol Chem

1 999 Mar 26;274(13):8531 -8538.

(68) Chen C, Edelstein LC, Gelinas C. The Rel/NF-kappaB family directly activates expression of the

apoptosis inhibitor Bcl-x(L). Mal Cell Biol 2000 Apr;20(8):2687-2695.

(69) Sokol RJ, Devereaux M, Dahl R, Gumpricht E. "Let there be bile " - Understanding hepatic injury in

cholestasis. Journal of Pediatric Gastroenterology and Nutrition 2006 Jul;43:S4-S9.

(70) Paumgartner G. Medical treatment of cholestatic liver diseases: From pathobiology to pharmacological

targets. World Journal of Gastroenterology 2006 Jul 28;1 2(28):4445-4451 .

(71 ) Miyoshi H, Rust C, Guicciardi ME. Gores GJ. NF-kappaB Is activated in cholestasis and functions to

reduce liver injury. Am J Pathol 2001 Mar;1 58(3):967-975.

(72) Lieser MJ, Park J, Natori S, Jones BA, Bronk SF, Gores GJ. Cholestasis confers resistance to the rat

liver mitochondrial permeability transition. Gastroenterology 1 998 Sep; 1 1 5(3):693-701 .

(73) Geier A, Dietrich CG, Voigt S, Kim SK. Gerloff T , Kullak-Ublick GA, et al. Effects of proinflammatory

cytokines on rat organic anion transporters during toxic liver injury and cholestasis. Hepatology 2003

Aug;38(2 ):345-354.

(74) Li ZD, Mizuno S, Nakamura T. Anlinecrotic and antiapoptotic effects of hepatocyte growth factor on

cholestatic hepatitis in a mouse model of blle-obstructive diseases. American Journal of Physiology­

Gastrointestinal and Liver Physiology 2007 Feb;292(2):G639-G646.

(75) Rost D, Herrmann T, Sauer P, Schmidts HL, Stieger B, Meier PJ, et al. Regulation of rat organic anion

transporters in bile salt-induced cholestatic hepatitis: effect of ursodeoxycholate. Hepatology 2003

Jul;38(1 ):1 87-1 95.

(76) Schoemaker MH, Conde de la Rosa, Buist-Homan M, Vrenken TE, Havinga R, Poelstra K, et al.

Tauroursodeoxycholic acid protects rat hepatocytes from bile acid-induced apoptosis via activation of

survival pathways. Hepatology 2004 Jun;39(6):1 563-1 573.

General Introduction

(77) Oyama K, Shiota G, Ito H, Murawaki Y, Kawasaki H. Reduction of hepatocarcinogenesis by

ursodeoxycholic acid in rats. Carcinogenesis 2002 May:23(5):885-892.

(78) Beuers U, Nathanson MH, Isales CM, Boyer JL. Tauroursodeoxycholic acid stimulates hepatocellular exocytosis and mobilizes extracellular Ca++ mechanisms defective in cholestasis. J Clin Invest 1993

Dec:92(6):2984-2993.

(79) Beuers U, Bilzer M, Chittattu A, Kullak-Ublick GA, Keppler D, Paumgartner G, et al.

Tauroursodeoxycholic acid inserts the apical conjugate export pump, Mrp2, into canalicular membranes

and stimulates organic anion secretion by protein kinase C-dependent mechanisms in cholestatic rat

liver. Hepatology 2001 May;33(5):1206-1216.

(80) Alpini G. Ursodeoxycholate and tauroursodeoxycholate inhibit cholangiocyte growth and secretion of

BDL rats through activation of PKG alpha (vol 35, pg 1 041 , 2002). Hepatology 2002 Jul;36(1 ):265.

(81) Rodrigues CM, Ma X, Linehan-Stieers C, Fan G, Kren BT, Steer CJ. Ursodeoxycholic acid prevents

cytochrome c release in apoptosis by inhibiting mitochondrial membrane depolarization and channel

formation. Cell Death Differ 1999 Sep;6(9):842-854.

(82) Browning JD, Horton JD. Molecular mediators of hepatic steatosis and liver injury. J Clin Invest 2004

Jul;114(2):147-152.

(83) Adachi M, Ishii H. Role of mitochondria in alcoholic liver injury. Free Radical Biology and Medicine 2002

Mar 1 5;32(6):487-491.

(84) Moshage H, Casini A, Lieber CS. Acetaldehyde selectively stimulates collagen production in cultured rat

liver fat-storing cells but not in hepatocytes. Hepatology 1 990 Sep;12(3 Pt 1 ):511 -518.

(85) Georgiev P, Navarini AA, Eloranta JJ, Lang KS, Kullak-Ublick GA, Nocito A, et al. Cholestasis protects

the liver from ischaemic injury and post-ischaemic inflammation in the mouse. Gut 2007 Jan;56(1):1 21-1 28.

(86) Bantel H, Lugering A, Poremba C, Lugering N, Held J, Domschke W, et al. Caspase activation

correlates with the degree of inflammatory liver injury in chronic hepatitis C virus infection. Hepatology

2001 Oct;34(4 Pt 1):758-767.

(87) Lin HZ, Yang SQ, Chuckaree C, Kuhajda F, Ronne! G, Diehl AM. Metformin reverses fatty liver disease

in obese, leptin-deficient mice. Nature Medicine 2000 Sep:6(9):998-1003.

(88) Hardie DG. The AMP-activated protein kinase pathway-new players upstream and downstream. J Cell

Sci 2004 Nov 1 ;11 7(Pt 23):5479-5487.

(89) Reiter AK, Bolster DR, Crozier SJ, Kimball SR, Jefferson LS. Repression of protein synthesis and

mTOR signaling in rat liver mediated by the AMPK activator aminoimidazole carboxamide

ribonucleoside. Am J Physiol Endocrinol Metab 2005 May;288(5):E980-E988.

(90) Kimura N, Tokunaga C, Dalal S, Richardson C, Yoshino K, Hara K, et al. A possible linkage between

AMP-activated protein kinase (AMPK) and mammalian target of rapamycin (mTOR) signalling pathway.

Genes Cells 2003 Jan;8(1 ):65-79.

(91) Yin W, Mu J, Birnbaum MJ. Role of AMP-activated protein kinase in cyclic AMP-dependent l ipolysis in

3T3-L 1 adipocytes. Journal of Biological Chemistry 2003 Oct 31 ;278(44):43074-43080.

(92) Angelico F, Burattin M, Alessandri C, Del Ben M, Lirussi F. Drugs improving insulin resistance for non­

alcoholic fatty liver disease and/or non-alcoholic steatohepatitis. Cochrane Database of Systematic

Reviews 2007;(1 ).

(93) Bugianesi E, Marzocchi R, Villanova N, Marchesini G. Non-alcoholic fatty liver disease/non-alcoholic

steatohepatitis (NAFLD/NASH): treatment. Best Practice & Research in Clinical Gastroenterology 2004

Dec;1 8(6):1105-1116.

(94) Angulo P, Lindor KO. Treatment of non-alcoholic steatohepatitis. Best Practice & Research in Clinical

Gastroenterology 2002 Oct;1 6(5):797-810.

(95) Hawley SA, Gadalla AE, Olsen GS, Hardie DG. The antidiabetic drug metformin activates the AMP­

activated protein kinase cascade via an adenine nucleotide-independent mechanism. Diabetes 2002

Aug;51 (8):2420-2425.

(96) Musi N, Hirshman MF, Nygren J, Svanfeldt M, Bavenholm P, Rooyackers 0, et al. Metformin increases

AMP-activated protein kinase activity in skeletal muscle of subjects with type 2 d iabetes. Diabetes 2002 Ju1;51 (7):2074-2081 .

(97) Fryer LGD, Parbu-Patel A, Carling D . The anti-diabetic drugs rosiglitazone and metformin stimulate

AMP-activated protein kinase through distinct signaling pathways. Journal of Biological Chemistry 2002

Jul 12;277(28):25226-25232.

21

Chapter 1

(98) Zhou GC, Myers R, Li Y, Chen YL, Shen XL, Fenyk-Melody J, et a l . Role of AMP-activated protein

kinase in mechanism of metformin action. Journal of Clinical Investigation 2001 Oct;1 08(8): 1 1 67-1 1 74.

(99) Latchoumycandane C, Seah QM, Tan RCH , Sattabongkot J, Beerhelde W, Boelsterli UA. Leflunomide

or A77 1726 protect from acetaminophen-induced cell injury through inhibition of JNK-mediated

mitochondrial permeability transition in Im mortalized human hepatocytes. Toxicology and Applied

Pharmacology 2006 Nov 1 5;217(1 ):125-133.

( 100) Manna SK, Mukhopadhyay A. Aggarwal BB. Leflunomide suppresses TNF-induced cellular responses:

Effects on NF-kappa B, activator protein-1 , c-Jun N-terminal protein kinase, and apoptosis. Journal of

Immunology 2000 Nov 1 5; 165(1 0):5962-5969.

( 1 0 1 ) lmose M, Nagaki M, Kimura K, Takai S, lmao M, Naiki T, et al. Leflunomide protects from T-cell­

mediated liver injury in mice through inhibition of nuclear factor kappaB. Hepatology 2004

Nov;40(5):1 1 60-1 1 69.

(1 02) Karaman A, lraz M, Klrimlloglu H, Karadag N , Tas E, Fadillioglu E. Hepatic damage in biliary-obstructed

rats is ameliorated by leflunomide treatment. Pediatr Surg Int 2006 Sep;22(9):701 -708.

(1 03) Manna SK, Aggarwal BB. lmmunosuppressive leflunomide metabolite (A77 1 726) blocks TNF­

dependent nuclear factor-kappa B activation and gene expression. Journal of Immunology 1999 Feb

1 5; 162(4 ):2095-21 02.

(1 04) Latchoumycandane C, Goh CW, Ong MM, Boelsterli UA. Mitochondrial protection by the JNK inhibitor

leflunomide rescues mice from acetaminophen-induced liver injury. Hepatology 2007 Feb;45(2):41 2-

421 .

(1 05) Karaman A, Fadillioglu E , Turkmen E, Tas E , Yilmaz Z . Protective effects of leflunomide against

ischemia-reperfusion injury of the rat liver Pediatr Surg Int 2006 May:22(5):428-434.

(1 06) Yao HW, Li J, Chen JQ, Xu SY. Leflunomide attenuates hepatocyte injury by inhibiting Kupffer cells

World J Gastroenterol 2004 Jun 1 ; 10( 1 1 ): 1 608-1 61 1 .

(1 07) Alexander D, Friedrich B, Abruzzese T, Gondolph-Zink B, Wulker N , Aicher WK. The active form of

leflunomide, HMR1 726, facilitates TNF-alpha and IL-1 7 induced MMP-1 and MMP-3 expression. Cell

Physiol Biochem 2006;1 7(1 -2):69-78.

(1 08) Vergne-Salle P, Leger DY, Bertin P, Treves R, Beneytout JL, Liagre B. Effects of the active metabolite

of leflunomide, A77 1 726, on cytokine release and the MAPK signalling pathway in human rheumatoid

arthritis synoviocytes. Cytokine 2005 Sep 7;31 (5):335-348.

(1 09) Layseca-Espinosa E, Pedraza-Alva G, Montiel JL, del RR, Fierro NA, Gonzalez-Amaro R, et al. T cell

aggregation induced through CD43: intracellular signals and inhibition by the immunomodulatory drug

leflunomide. J Leukoc Biol 2003 Dec;74(6):1 083-1 093.

(1 10) Xu X, Williams JW, Bremer EG, Finnegan A, Chong AS. Inhibition of protein tyrosine phosphorylation in

T cells by a novel immunosuppressive agent, ieflunomide. J Biol Chem 1 995 May 26;270(21 ): 1 2398-

1 2403.

22

Chapter 2

Metformin protects rat hepatocytes against bile acid-induced apoptosis

Titia E. Vrenken, Laura Conde de la Rosa, Manon Buist-Homan, Klaas Nico Faber, Han Moshage

Department of Gastroenterology and Hepatology; University Medical Center Groningen, University of Groningen, Gron ingen , the Netherlands

In preparation

Chapter 2

2.1 Abstract

Background: Metformin is used in the treatment of Diabetes Mellitus type II and improves liver function in patients with steatosis. Metformin activates AMP-activated protein kinase (AMPK), the cellular energy sensor that is sensitive to changes in the AMP/ATP-ratio. AMPK is an inhibitor of mammalian target of rapamycin (mTOR). Both AMPK and mTOR are able to modulate cell death. Aim: to evaluate the effects of metformin on hepatocyte cell death. Methods: Apoptotic cell death was induced in primary rat hepatocytes using either the bile acid glycochenodeoxycholic acid (GCDCA) or tumor necrosis factor (TNF) a in combination with actinomycin D (ActD). Metformin (0.1-2 mmol/L) was added 10 minutes prior to the apoptotic stimuli, unless stated otherwise. AMPK, mTOR and phosphoinositide-3 kinase (Pl3K)/Akt were inhibited using pharmacological inhibitors. Apoptosis and necrosis were quantified by caspase activation, acridine orange staining and Sytox green staining respectively. Results: Metformin dose-dependently reduces GCDCA-induced apoptosis, even when added 2 hours after GCDCA, without increasing necrotic cell death. Metformin does not protect against TNFa/ActD-induced apoptosis. The protective effect of metformin is dependent on an intact Pl3-kinase/Akt pathway, but does not require AMPK/mTOR­signaling. Metformin does not inhibit NF-KB activation. Conclusion: Metformin protects against bile acid-induced apoptosis, but not TNFa/ActD­induced apoptosis. If these results can be confirmed in in vivo models, metformin could be considered in the treatment of chronic liver diseases accompanied by inflammation.

24

Metformin protects rat hepatocytes against bile acid-induced apoptosis

2.2 Introduction

In chronic liver diseases, like chronic cholestasis, hepatocytes are exposed to increased levels of bile acids, toxic metabolites and oxidative stress, causing cell death. In acute liver diseases, such as acute viral hepatitis, cell death is caused by pro-inflammatory and pro­apoptotic cytokines and reactive oxygen species. Cell death can occur as apoptosis or necrosis. Apoptosis or programmed cell death is an energy dependent process, involving the caspase-family of cysteine proteases. Apoptosis is characterized by cell shrinkage, nuclear condensation and formation of apoptotic bodies. On the contrary, necrosis is a passive process, involving plasma membrane rupture and release of cellular content into the extracellular environment and systemic circulation, causing inflammation (1-8).

Metformin is a drug used in the treatment of Diabetes Mellitus type II, where it improves insulin sensitivity. Recently, metformin was shown to have beneficial effects in fatty liver diseases: metformin was found to reduce leptin secretion (9). Furthermore, metformin reduced aminotransferase levels and decreased liver size in patients with non-alcoholic steatohepatitis (NASH) and animal models for steatosis. Moreover, metformin treatment improved hepatocyte viability in fatty livers (10-13). Metformin is known to stimulate AMP­activated protein kinase (AMPK) activity in whole liver, in human and rat primary hepatocytes, and in a rat hepatoma cell line (14-16). Among the 5 members of the AMPK family, AMPK-a1 and -a2 are activated by metformin (14, 17). AMPK consists of a catalytic a subunit and two regulatory subunits (P, y; 14, 18). AMPK is involved in energy homeostasis, and becomes activated upon a rise in the cellular AMP concentration or changes in the AMP/ATP-ratio. Furthermore, AMPK can be activated by stimuli not changing the AMP/ATP­ratio, like hyperosmotic stress, hypoxia, oxidative stress or pharmacological compounds. AMPK activity is dependent on the phosphorylation of Thr172 in the a subunit (16, 18-24). Activation of AMPK using the cell permeable adenosine analogue 5-aminoimidazole-4-carboxamide 1-P-D-ribofuranoside (AICAR) was shown to be pro-apoptotic in liver cells, via activation of Jun N-terminal kinase (JNK) and caspase-3 (25). In a rat hepatoma cell line AMPK activity stimulated apoptosis, and in pancreatic f3-cells both metformin and AICAR induced apoptosis. On the contrary, in astrocytes and endothelial cells, activation of AMPK reduced apoptosis (26). Moreover, in DLD-1 cells, Ark5, another AMPK family member, was protective against Fas-mediated cell death. Ark5 directly inhibited one of the effector caspases, caspase-6, and Ark5 activity was shown to be controlled by Akt, a key regulator in survival signaling (15, 19, 27-30). In whole liver, AMPK activity represses signaling via the mammalian target of rapamycin (mTOR), a downstream target of Akt and phosphoinositide-3 kinase (Pl3K). mTOR is a key player in transcription, translation, cytoskeletal arrangement, and protein degradation (18, 20, 29, 31-36). Akt was found to suppress apoptosis in various cell types, including liver cells. In a rat hepatoma cell line, constitutive activation of Pl3K blocks glycochenodeoxycholic acid (GCDCA)-induced apoptosis. In primary rat hepatocytes, the protection of tauroursodeoxycholic acid (TUDCA) against GCDCA-induced apoptosis was abolished when the Pl3K/Akt survival pathway was inhibited (5, 37-41). In addition to Pl3K/Akt, there are several important survival pathways in hepatocytes, such as the transcription factor nuclear factor-KB (NF-KB) and the mitogen-activated protein kinases (MAPK; 42). Activation of NF-KB leads to the induction of survival genes and subsequently inhibition of apoptosis. In cholestatic livers, NF-KB is activated, and reduces liver injury (43). GCDCA-induced apoptosis was also reduced by NF-KB activation in primary rat hepatocytes (44). Furthermore, inhibition of both the Pl3K pathway and one of the anti-apoptotic MAPK members p38 or extracellular signal regulated kinase (ERK) 1/2 enhanced bile acid induced cell death in rat hepatocytes (45).

Considering the controversy about the role of metformin in apoptosis and the apparent beneficial effect in the treatment of steatohepatitis, we investigated the effect of metformin in 2 different in vitro models of hepatocyte cell death. We studied the effects of metformin on hepatocyte survival pathways, and examined if downstream targets of metformin modulate hepatocyte cell death. We describe a hepatoprotective action of metformin against bile acid-

25

Chapter 2

induced apoptosis which is independent of AMPK activation, but dependent on an intact Pl3K/Akt signaling pathway.

2.3 Materials and Methods

Animals Specified pathogen-free male Wistar rats (200-250 g) were purchased from Charles River Laboratories Inc (Wilmington, MA, USA). Rats were housed under standard laboratory conditions with free access to standard laboratory chow and water. Prior to isolation, rats were fasted overnight. Experiments were performed following the guidelines of the local Committee for Care and Use of laboratory animals.

Rat hepatocyte isolation Hepatocytes were isolated from rats by a two step collagenase perfusion as described before (46). Cell viability was greater than 85% as determined by Trypan Blue exclusion. After isolation, 112,500 cells per cm2 were plated on Vitrogen® (Cohesion Technologies Inc., Palo Alto, CA, USA) coated plates in William's E medium (lnvitrogen, Breda, The Netherlands) supplemented with 50 µg/mL gentamycin (lnvitrogen) and penicillin­streptomycin-fungizone (Lonza, Verviers, Belgium). During the attachment period (4 hrs) 50 nmol/L dexamethasone (Dept. of Pharmacy, UMCG, Groningen, The Netherlands) and 5% fetal calf serum (lnvitrogen) were added to the medium. Cells were cultured in a humidified incubator at 37°C and 5% CO2.

Experimental design Experiments were started 4 hrs after isolation of hepatocytes. Cell death was induced by bile acids or cytokines. Monolayer cultures were exposed to 50 µmol/L GCDCA (Calbiochem, La Jolla, CA, USA) for 4 hrs or the indicated time period , or hepatocytes were exposed for 16 hrs to 20 ng/ml recombinant murine tumor necrosis factor a (mTNFa, R&D Systems, Abingdon, United Kingdom) in combination with 200 ng/ml actinomycin D (ActD, Roche Diagnostics, Almere, The Netherlands) or a cytokine mixture (CM) consisting of 20 ng/ml mTNFa, 10 ng/ml recombinant human interleukin-113 (hlL-113, R&D Systems), 10 ng/ml recombinant rat interferon-y (rlFNy, R&D Systems) and 10 µg/ml lipopolysaccharide (LPS, Escherichia coli, serotype 0127:BB, Sigma-Aldrich, St. Louis, MO, USA). 1, 1-dimethylbiguanide hydrochloride (metformin, Sigma-Aldrich) was used at a concentration range from 0.1-2 mmol/L. Signal transduction pathways were specifically blocked using 0.1 µmol/L of the AMPK inhibitor 5'-iodotubercidin (Calbiochem), 0.5 µmol/L of the mTOR inhibitor rapamycin (Calbiochem) or 50 µmol/L of the Pl3K inhibitor L Y294002 (Calbiochem). All inhibitors were added 30 minutes prior to bile acids or cytokines. Each experimental condition was performed in triplicate wells. Each experiment was performed at least three times, using hepatocytes from different isolations. Cells were harvested at the indicated time points after the addition of the apoptotic stimuli, and washed three times with ice cold phosphate-buffered saline (PBS) before the addition of hypotonic cell lysis buffer (protein analysis, caspase-3 assay), 2x concentrated sample buffer (Western blot analysis) or Tri-reagent (RNA isolation, Sigma-Aldrich) as described previously ( 45 ).

Caspase enzyme activity assays Caspase-3 like activity was assayed as described previously (45). The arbitrary fluorescence unit (AFU) was corrected for the amount of protein in the cell lysate. Caspase-6 like activity was assayed according to the manufacturer's instructions (Biovision, Mountain View, CA, USA).

Sytox green and acridine orange nuclear staining To determine necrotic cell death, hepatocytes were incubated for 15 minutes with Sytox green (1 :40,000, lnvitrogen) nucleic acid stain. Sytox can only enter cells with compromised plasma membranes. Hepatocytes exposed to 5 mmol/L hydrogen peroxide (H202, Merck 26

Metformin protects rat hepatocytes against bile acid-induced apoptosis

Chemicals Ltd, Nottingham, UK) for 6 hrs served as positive control for necrosis. Apoptotic nuclei were visualized with acridine orange (1 :5,000, Sigma-Aldrich) as described previously (46). Fluorescent nuclei were visualized using an Olympus CKX41 microscope at 450-490 nm. Necrosis and apoptosis were quantified by counting fluorescent nuclei (necrotic or apoptotic cells) and the total number of cells in 3 randomly chosen high power fields.

Western blot analysis Western blot analysis of cell lysates was performed by SOS-PAGE followed by semi dry­blotting to transfer the proteins to Hybond ECL n itrocellulose membrane (Amersham Biosciences, Piscataway, NJ , USA). Ponceau S 0.1 % w/v (Sigma-Aldrich) staining was used to confirm electrophoretic transfer. Activation of AMPK was detected using the polyclonal antibody against phosphorylated AMPKa (Thr 172, lnvitrogen) at a dilution of 1 :500. Hepatocytes exposed to 250 µmol/L AICAR (Biomol Research Laboratories Inc, Plymouth Meeting, PA, USA) for 60 minutes served as positive control for phosphorylation and activation of AMPK. After Western blot analysis, blots were stripped as described previously (46) and incubated with a monoclonal antibody against �-actin (Sigma-Aldrich, dilution 1: 10000). The blots were analyzed in a ChemiDoc XRS system (Bio-Rad, Hercules, CA, USA). Protein band intensities were quantified by Quantity One software (Bio-Rad).

RNA isolation and quantitative polymerase chain reaction (PCR) RNA isolation was performed as described previously (47). RNA concentration was determined with the Ribogreen RNA quantitation reagent and kit (lnvitrogen). Reverse transcription (RT) PCR was carried out on 2.5 µg of total RNA using random primers in a total volume of 50 µL using Moloney murine leukemia virus (M-MLV) reverse transcriptase system according to the manufacturer's instruction (Sigma-Aldrich). Quantitative PCR was performed on 4 µL 20-times diluted complementary DNA in a final volume of 20 µL (47). Fluorescence was measured using the ABI PRISM 7700 Sequence Detector version 1.7 software (Applied Biosystems, Nieuwerkerk a/d IJssel, The Netherlands) starting with 10 minutes at 95°C, followed by 40 cycles of 15 seconds at 95°C and 1 minute at 60°C. Details of primers and probes are listed in Table 1. Each sample was analyzed in duplicate. 18S mRNA levels were used as endogenous control. Table 1. Sequences of primers and probes used for quantitative PCR analysis

18S Sense 5'-CGG CTA CCA CAT CCA AGG A-3'

rat Antisense 5'-CCA ATT ACA GGG CCT CGA AA-3'

Probe 5'FAM-CGCGCAAA TT ACCCACTCCCGA-TAMRA3'

iNOS Sense 5'-GTG CTA ATG CGG AAG GTC ATG-3'

rat Antisense 5'-CGA CTT TCC TGT CTC AGT AGC AAA-3'

Probe 5'FAM-CCC GCG TCA GAG CCA CAG TCC T-TAMRA3'

Bcl-x, Sense 5'-TTG TGG ATC TCT ACG GGA ACA AT-3'

rat Antisense 5'-GTC AGG AAC CAG CGG TTG AA-3'

Probe 5'FAM-CTC CTG GCC TTT CCG GCT CTC G-TAMRA3'

Statistical analysis All numerical results are reported as the mean of at least 3 independent experiments ± standard error of the mean. For each experiment, the results were analyzed using the Kruskal-Wallis test to verify the significance. If P<0.05 for the Kruskal-Wallis test, a Mann­Whitney U test was used to determine the significance of differences between experimental groups. A P-value smaller than 0.05 was considered to be statistically significant.

2.4 Results

Metformin dose-dependently reduces bile acid induced apoptosis To examine whether metformin is hepatoprotective, two different models for apoptosis were used: bile acid (GCDCA)-induced apoptosis and TNFa/ActD-induced apoptosis. GCDCA­induced caspase-3 activity peaks at 4 hrs (45), while for TNFa/ActD-induced apoptosis,

27

Chapter 2

caspase-3 activity is maximal between 12-16 hrs (data not shown). Therefore, these time points were chosen to study the effects of metformin. Metformin inhibited GCDCA-induced caspase-3 activity dose-dependently from 24 % inhibition at 0.1 mmol/L to an almost complete block (84% inhibition) at a concentration of 2 mmol/L (Fig 1A). In contrast, 1 mmol/L metformin, a concentration which reduces GCDCA-induced caspase-3 activity for 63%, had no effect on TNFa/ActD-induced caspase-3 activity (Fig 18). Metformin (0.1-2 mmol/L) alone did not induce caspase-3 activity. However, higher concentrations (5 mmol/L and higher) were toxic to hepatocytes (data not shown). Acridine orange staining confirmed the results obtained with the caspase-3 assay. Condensed nuclei are visible in 47% of GCDCA-treated hepatocytes (Fig 1 C), whereas in combination with metformin this number of apoptotic cells is reduced by 50% to 23%. However, metformin hardly affects TNFa/ActD­induced nuclear condensation (Fig 1 C). Apoptotic cells were not observed in control or metformin-exposed hepatocytes. Next, the effects of metformin on GCDCA-induced caspase-6 activity were studied. Metformin (1 mmol/L) decreased GCDCA-induced caspase-6 activity by 46 % (Fig 1 D).

{'I 1! �'-=�1 ____ 1 __ iL---a·---· ----- _ll ____ , ....

Figure 1. Metformln dose-dependently reduces bile acid- nduced caspase-3 and caspase-6 act,vity, but not TNFaiActD-induced apoptosis. Pr•mary rat hepatocytes were exposed to metformin prior to (A, D) GCDCA or (B) TNFa/ActD. Caspase-3 l'ke activity ,s shown as fold induction compared to control, control values were set at one. (C) Metform1n prevents

---==--- GCDCA (4 hrs) induced nuclear condensat,on as demonstrated by acridlne orange staining. but has no effect on TNFalActD ( 16 hrs) induced nuclear condensation. Percentages represent condensed nuclei. Magnification 20X. (D) Caspase-6 like activity is measured in pr mary rat hepatocytes and is shown as fo d induction compared to control; contr� values were set at one Stat1stica1 analysis: • p<0.05 or + p<O 01 compared to control. # p<0.05 compared to GCDCA-trealed cells

To demonstrate that metformin inhibits and does not delay bile acid induced apoptosis, a time course study was performed. At the indicated time points, from four to nine hrs after addition of GCDCA, metformin (added 10 min before GCDCA) inhibited GCDCA-induced caspase-3 activity (Fig 2A). After 24 hrs, levels of caspase-3 activity of GCDCA and GCDCA plus metformin treated cells were both comparable to control cells. Acridine orange staining was performed to ensure that the peak of caspase-3 activity of hepatocytes treated with GCDCA plus metformin was not between any of the time points studied. Both after 4 and 24 hr exposure, metformin reduced the GCDCA-induced nuclear condensation by 72 % (Fig 2B).

i1 : 1�1 ------· ---'---·---·�.:. _____ _ Figure 2. Metform:n does not delay bile ackt-induced apoplosis. Primary rat hepatocytes were exposed to GCDCA without or in combjnat,on with metformin for 4, 6, 9 and 24 hrs MetformIn was added 10 minutes before GCDCA Control cells are rat hepatocytes not treated with GCDCA or metformm At the indicated time points, cells were analyzed for (A) caspase-3 1ke actlvty or (B) nuclear condensation as ndicated by acndine orange staining and expressed as % of condensed nuc el Magnilicahon 20X Statlstical analysis + p<0.01 compared to control. # p<0.05 or • p<0.01 compared to GCDCA-treated cells (same t me-point).

28

Conlrol , .,.

C4,.rol u,.

GCDCA 50 � .. ,.

GCDCA501,1M .,,.

GCDCA501,1M• matformln 1 mM ,.,.

GCDCA50 JIN • metfonnln 1 mM

23%

Metformin protects rat hepatocytes against bile acid-induced apoptosis

Metformin does not block the uptake of GCDCA at the cell membrane Next, we examined whether metformin only protects cells when added before GCDCA, or also after bile acid exposure. Metformin was added either 10 minutes before, simultaneously with, or 1, 2 or 3 hrs after addition of the bile acid. Metformin significantly decreased GCDCA induced caspase-3 activity when added up to 2 hrs after GCDCA (Fig 3). No difference in reduction of caspase-3 activity between a pre-incubation of 10 minutes with metformin or addition at the same time as GCDCA was detected (data not shown).

40

control

Figure 3. Mettormin protects primary rat hepatocytes against apoptosis when added after GCDCA Hepatocytes were treated without (control) or with GCDCA for 4 hrs. Metformin was added with GCDCA or 1 . 2 and 3 hrs after addition of GCDCA. Caspase-3 like activity was measured and is shown as fold induction compared to control; control values were set at one.

----.-•• -,---,-,.-,.-.,.....:....c......_

2-

,..-.-•• -,---3-,.,

-.-•• -,- Statistical analysis: + p<0.01

GCOCA compared to control. # p<0.05 or • -----------==:::....---------- p<0.01 compared to GCDCA· treated cells

Furthermore, equal uptake of fluorescent bile acids in hepatocytes treated with or without metformin was observed (data not shown). These results indicate that the protective effect of metformin is not due to reduced uptake of GCDCA into hepatocytes.

Metformin reduces apoptosis without increasing necrotic cell death Subsequently, we studied whether the reduction in apoptotic cell death by metformin results in a shift towards necrotic cell death. At the peak of caspase-3 activity, 4 hrs after exposure to GCDCA the percentage of necrotic cells was determined. GCDCA caused necrosis in 11 % of the cells, whereas 7% of the hepatocytes exposed to GCDCA and metformin were necrotic (Fig 4). Also after 16 hrs, there was no significant difference between cultures treated with GCDCA or GCDCA plus metformin (1 mmol/L, data not shown). No significant necrosis was observed in control cultures or cultures treated with metformin (Fig 4 ).

Figure 4. Metformln reduces apoptosis without increasing necrotic cell death. Cells w11<e exposed to GCOCA for 4-16 hrs. Metformin was added before addition of GCDCA Cells st,mu.ated with H202 (5 mmolll. 6 hrs) were used as positive control Sytox green nuclear staining was used to determine necrotic cell death Representative images 4 hrs after stimulat1on are shown. percentages indk:ate necrotic ce�s Upper panel phase contrast and fluorescence. Lower panel fluorescence. Magnification 20X

Control <1%

GCDCA 11%

GCDCA + metformln

7%

Metformin 5%

positive control >95%

The anti-apoptotic effects of metformin are dependent on Pl3K/Akt signaling, but not on AMPK/mTOR signaling Previous results showed the involvement of the Pl3K survival pathway in the protective effect of TUDCA against GCDCA-induced apoptosis (45). Thus, we investigated whether the Pl3K pathway is also involved in the anti-apoptotic effects of metformin using the specific Pl3K inhibitor, L Y294002. Metformin reduced the GCDCA-induced caspase-3 like activity with 69% (Fig 5A). This protective effect of metformin against apoptosis was completely abolished when the Pl3K pathway was blocked (Fig 5A). A significant increase in GCDCA­induced apoptosis was observed when the Pl3K pathway was inhibited in hepatocytes (data

29

Chapter 2

not shown) as described previously (45). Next, the involvement of the AMPK/mTOR pathway was examined. AMPK was inhibited using a pharmacological inhibitor (5'-iodotubercidin) and mTOR was blocked by the specific inhibitor rapamycin (Fig 58). As shown in figure 58, blocking either AMPK or mTOR had no effect on the anti-apoptotic actions of metformin. Both inhibitors did not change GCDCA-induced caspase-3 activity (data not shown). Hepatocytes treated with the inhibitors alone or in combination with metformin showed caspase-3 values similar to control cells (data not shown). Interestingly, Western blot analysis demonstrated that hepatocytes treated for 1 hr with either GCDCA or metformin caused a slight induction (2-fold) of phosphorylated AMPKa compared to control cells (Fig SC). Cells exposed to both GCDCA and metformin showed a further 2-fold increase of AMPKa phosphorylation compared to control hepatocytes, while cells treated with AICAR induced a 30-fold increase of AMPKa phosphorylation.

A .. .. J t ,

U " ul

C

t

�- ' -·- ' · +LY29400260il.t

+INCtiim*I t ""'

CCllCA

<--I ----Jl_-_,_I _•__,1-N.1PKa

1--1-1- -

"'"'"" OCOCA .........

.. . n

u �

'"""" +5�011611 .,...,.,,_osi1.t +�1 ....

CCllCA

Figure 5. The protective effect of metformtn involves the Pl3K/Akt pathway, bul not lhe AMPK/mTOR palhway Primary rat hepalocyles were exposed lo GCDCA and melformln for 4 hrs wilh or wilhoul (A) LY294002 or (B) 5"-lodoluberc·d n or rapamycm. Caspase-3 aclivily is presented as fold induction compared to control; control values were set at one Statistical anarysis: • p<0 05 or + p<0.01 compared to control. # p<0 05 compared lo GCDCA-lrealed cells. • p<0 05 compared lo GCDCA plus melformin-lrealed eel s. (C) Western b'ol analysis for AMPK phosphorylation on cell lysales (45 µI). Lane 1 control hepalocyles; lane 2: GCDCA treated cells (50 µmoVL, 1 hr); lane 3: GCDCA plus melformln (1 mmol/L) lane 4: melformin (1 mmoVL); lane 5: positive control: hepalocyles exposed lo AICAR (250 GCOCA+ - µmolll, 1 hr). Western blots were quanbfied and band intensities of

phaspho'Y ated AMPKa were expressed as ratio of the 1ntens1ty of f!-act1n prote111 levels Data is presented as mean of at least three independent expenments +/. S.E M. Control values were set at one A representative Western blot Is shown.

Metformin induces Bel-xi expression Members of the Bel-2 family are key regulators of apoptosis. Previous studies revealed a role of the anti-apoptotic Bcl-2 family member, Bel-xi, in the protection of rat hepatocytes against superoxide anion induced apoptosis (unpublished data). Thus, the involvement of Bel-xi in GCDCA-induced apoptosis was studied. GCDCA did not change Bel-xi gene expression compared to control cells (Fig 6). Metformin caused a slight increase of Bel-xi gene expression, which was further enhanced in hepatocytes exposed to both GCDCA and metformin (Fig 6).

comrol

30

+ metfonnin 1 mM metformin 1 mM

GCDCA

Figure 6. Metformin increases Bc.l·xl mRNA expression, which is further enhanced In combination wilh GCDCA qPCR analysis for Bel­xi using cDNA of primary rat hepatocytes stimulated for 4 hrs with GCDCA (50 µmol/L) and/or melformln ( 1 mmoVL).

Metformin protects rat hepatocytes against bile acid-induced apoptosis

Metformin is not an inhibitor of NF-KB Since inhibition of NF-KB sensitizes hepatocytes to TNFa-induced cell death (39), it is important to determine the effect of metformin on NF-KB activation, prior to its use in inflammatory (TNFa-mediated) liver diseases. Therefore, we studied the involvement of NF­KB signaling in the protective effect of metformin. Previous studies demonstrated strong NF­KB-dependent cytokine-induced expression of inducible nitric oxide synthase (iNOS) in hepatocytes (39). Cytokine mixture induces a 200-fold relative induction of iNOS expression compared to control cells (Fig 7 A), which was not affected by metformin. As expected, the transcriptional inhibitor actinomycin D completely blocked cytokine-induced iNOS expression (Fig 7A). Metformin alone did not induce iNOS expression (data not shown). The qPCR results were confirmed using the caspase-3 assay. TNFa alone, or in combination with 1 mmol/L metformin did not induce caspase-3 activation (Fig 7B). In contrast, caspase-3 activity was strongly induced in hepatocytes exposed to TNFa in combination with actinomycin D.

• '""*'"'*' 1 � + � O --

B

11 �._I -------�I C-""" + ........ 1 mM +�D -

Figura 7. Metformin does not inhibit NF-KB activalion (A) Metform1n does not reduce cytokine mixture-induced iNOS mRNA expression qPCR analysis for iNOS using cDNA of primary rat hepatocytes exposed for 4 hrs to cytokine mixture with or without 1 mmol/L metform n The transcriptional inhibitor actinomycin D (200 ng/ml) served as positive control for NF-KB inhibition. (B) Metformin does not sensitize hepatocytes to TNFo-lnduced apoptosis Caspase-3 like activity was measured in primary rat hepatocytes exposed to TNFo (20 ng/ml, 16 hrs) with or without metformln (1 mmol/L) or actinomycin D (200 ng/ml, positive control for induction of TNFa-induced caspase-3 activity). Caspase-3 activity is shown as fold induction compared to control: control values were set at one. Statistical analysis: Fig 7A: • p<0.05 or + p<0 01 compared to control # p<0.01 compared to cytokine mixture-treated cells. • p<0.01 compared to cytokine mixture plus metformin-treated cells. Fig 78 • p<0.05 compared to control. # p<0.05 compared to TNFa-treated cells. • p<0.05 compared to TNFa plus mettormin-treated cells.

2.5 Discussion

In this study we investigated the hepatoprotective actions of metformin in models of TNFa­and bile acid-induced apoptosis in primary rat hepatocytes. Other studies described that metformin improves the viability of hepatocytes in fatty livers and reduced serum levels of ALT and AST in NASH (1 1-1 3). We show that metformin is protective against bile acid­induced cell death, but metformin has no effect on TNFa-induced apoptosis. Metformin reduces, but does not delay, GCDCA-induced caspase activity and nuclear condensation. The opposite effects of metformin in these two models of hepatocyte apoptosis is most likely due to the involvement of different signaling pathways in cytokine and bile acid induced apoptosis. Cytokines like TNFa exert their effects via death receptors at the cell membrane, while bile acids need to be taken up by the bile acid transporter, Na

+/taurocholate cotransporting polypeptide (Ntcp; 1 , 4, 7). Furthermore, bile acid-induced apoptosis has been shown to involve activation of the Fas pathway via epidermal growth factor (EGF)­receptor phosphorylation (48). TNFa is known to activate the transcription factor NF-KB, while we have previously demonstrated that GCDCA has no effect on NF-KB signaling (44). To elucidate the mechanism of the protective effect of metformin, we first excluded the possibility that metformin is protective by inhibiting GCDCA uptake at the cell membrane. The results obtained with the fluorescent bile acids clearly demonstrated that metformin has no effect on bile acid uptake. Furthermore, metformin was able to reduce apoptosis when added up to 2 hrs after bile acid exposure. These data suggest that metformin does not interfere with bile acid uptake, but rather activates survival pathways very rapidly. In this respect, the protective effect of metformin resembles the protective effect of TUDCA. TUDCA also protected against GCDCA-induced cell death when added up to 2 hrs after the bile acid (45). We confirm previous findings (45), that caspase-3 activation is still minimal 2

31

Chapter 2

hours after GCDCA exposure, and apoptosis can be prevented , while beyond this time point the apoptotic machinery cannot be reversed anymore.

Previously, we demonstrated that under particular conditions the mode of cell death can shift from apoptosis to necrosis (46). In this study, metformin does not cause a switch from apoptotic to necrotic cell death. The percentage of necrotic cells is similar for GCDCA and GCDCA plus metformin treated hepatocytes. Furthermore, metformin does not significantly reduce GCDCA-induced necrosis. These results demonstrate that metformin only protects against apoptotic cell death.

Since the AMPK/mTOR pathway is an important downstream target of metformin, we investigated the involvement of these pathways in the protective effect of metformin. Our results demonstrate that the AMPK/mTOR signaling pathway is not involved in the anti­apoptotic actions of metformin (20, 21, 29, 49). Both known inhibitors for AMPK and mTOR, 5'-iodotubercidin and rapamycin, had no effect on the reduction in GCDCA-induced apoptosis caused by metformin. Interestingly, GCDCA slightly increased the phosphorylation and activation of AMPK, although AMPK does not appear to be involved in the protective effect of metformin. The mechanism of this effect and its relevance remains to be elucidated. Since AMPK is not involved in the protective actions of metformin, the drug must exert its anti-apoptotic actions via other pathways. Our results demonstrate that the protective effect of metformin against GCDCA-induced apoptosis is dependent on an intact Pl3K pathway. These data are in accordance with other reports emphasizing the importance of Pl3K/Akt in the protection against bile acid-induced apoptosis (5, 37, 38, 40, 41, 45). However, our data imply that members of the MAP kinase family are not involved in the anti-apoptotic signaling of metformin, since inhibitors of the anti-apoptotic MAP kinases ERK1/2 and p38 do not affect the protective effect of metformin (data not shown).

Our results demonstrate the induction of Bel-xi mRNA by metformin. Surprisingly, Bel-xi mRNA expression was further enhanced in combination with GCDCA, while Bel-xi expression levels were unaffected by the bile acid alone. Several reports have described NF-KB-controlled Bel-xi expression in different cell types (50, 51 ). However, in primary rat hepatocytes cytokine induction of the anti-apoptotic protein Bel-xi was not observed, while the anti-apoptotic Bcl-2 family member A1/Bfl-1 was induced by cytokines (39). Studies have reported both induction and down-regulation of NF-KB activity upon bile acid treatment (52-55). In human fetal hepatocytes, chenodeoxycholic acid (COCA) and its taurine, but not glycine, conjugate induced NF-KB activation (55). Another report describes the dose­dependent activation of NF-KB by GCDCA in hepatocytes. However, in this study GCDCA concentrations up to 400 µmol/L were used (52). In contrast, GCDCA at 500 µmol/L inhibited NF-KB activity in primary rat hepatocytes (53). We and others have previously reported that GCDCA at a concentration of 50 µmo!/L does not activate NF-KB in primary rat hepatocytes (44, 55). Thus, we can only speculate why metformin in combination with GCDCA further induces Bel-xi expression. The possibility that GCDCA sensitizes hepatocytes to NF-KB activation can be excluded, since metformin, either alone or in combination with GCDCA, did not induce the expression of the NF-KB-dependent gene iNOS. One study reported caspase­regulated Bel-xi expression, via caspases downstream of ERK1/2. Using the specific ERK1/2 inhibitor U0126, Bel-xi expression was reduced and apoptosis was induced. This U0126-induced down-regulation of Bel-xi was reversed by the pancaspase inhibitor Z-VAD-FMK (56). Although a different cell type was used, caspase-dependent Bel-xi expression could explain our results: in our model GCDCA-induced caspase activity is reduced by metformin and this could result in an additional increase in Bel-xi expression.

Our data clearly show that metformin does not inhibit NF-KB activity. Hepatocytes are not sensitized to TNFa-induced apoptosis by metformin, and metformin does not affect cytokine induced NF-KB-dependent gene transcription. Our data are in contrast with several reports describing inhibition of NF-KB by metformin via AMPK (57, 58). However, these studies are not performed in hepatocytes but in endothelial cells, which could explain the discrepancy. 32

Metformin protects rat hepatocytes against bile acid-induced apoptosis

Furthermore, these reports showed the involvement of AMPK in metformin signaling, while in our study AMPK is not involved in the anti-apoptotic actions of metformin.

In summary, we have shown that metformin, at concentrations similar to the clinical situation, is protecting primary rat hepatocytes against GCDCA-induced apoptosis, while metformin has no effect on TNFa/ActD-induced apoptosis. The Pl3K/Akt survival pathway is necessary for the anti-apoptotic effects of metformin. The protection is independent of the AMPK/mTOR signaling pathway. Furthermore, metformin induces the expression of the anti-apoptotic Bcl-2 family member Bel-xi. In addition, we have demonstrated that metformin does not inhibit NF-KB activity and hence does not sensitize hepatocytes to TNFa-induced apoptosis. Therefore, if our data can be confirmed in in vivo models, metformin could be useful in the treatment of chronic liver diseases, in which an inflammatory component is present.

2.6 Acknowledgements

This study was supported by the Dutch Digestive Foundation (WS 02-13) and the J .K. de Cock Foundation.

2. 7 Reference List

(1) Reed JC. Mechanisms of apoptosis. Am J Pathol 2000 Nov;1 57(5):141 5-1430.

(2) Varghese J, Khandre NS, Sarin A. Caspase-3 activation is an early event and initiates apoptotic

damage in a human leukemia cell line. Apoptosis 2003 Aug;8(4):363-370.

(3) Adams JM, Cory S. Apoptosomes: engines for caspase activation. Curr Opin Cell Biol 2002

Dec;14(6):715-720.

(4) Higuchi H, Gores GJ. Bile acid regulation of hepatic physiology: IV. Bile acids and death receptors.

Am J Physiol Gastrointest Liver Physiol 2003 May;284(5):G734-G738.

(5) Rust C, Karnitz LM, Paya CV, Mosca! J, Simari RD, Gores GJ. The bile acid taurochenodeoxycholate

activates a phosphatidylinositol 3-kinase-dependent survival signaling cascade. J Biol Chem 2000 Jun

30;275(26):20210-2021 6.

(6) Kim JS, He L, Lemasters JJ. Mitochondrial permeability transition: a common pathway to necrosis and

apoptosis. Biochem Biophys Res Commun 2003 May 9;304(3):463-470.

(7) Cohen GM. Caspases: the executioners of apoptosis. Biochem J 1997 Aug 15;326(Pt 1 ):1-16 .

(8) Lemasters JJ. Mechanisms of hepatic toxicity - V. Necrapoptosis and the mitochondrial permeability

transition: shared pathways to necrosis and apoptosis. American Journal of Physiology­Gastrointestinal and Liver Physiology 1999 Jan;276(1 ):G1-G6.

(9) Klein J, Westphal S, Kraus D, Meier B, Perwitz N, Ott V, et al. Metformin inhibits leptin secretion via a

mitogen-activated protein kinase signalling pathway in brown adipocytes. J Endocrinol 2004

Nov; 183(2 ):299-307.

(10) Lin HZ, Yang SQ, Chuckaree C, Kuhajda F, Ronne! G, Diehl AM. Metformin reverses fatty liver disease in obese, leptin-deficient mice. Nature Medicine 2000 Sep;6(9):998-1003.

(11 ) Angelico F , Burattin M , Alessandri C, Del Ben M , Lirussi F . Drugs improving insulin resistance for non­

alcoholic fatty liver disease and/or non-alcoholic steatohepatitis. Cochrane Database of Systematic Reviews 2007;(1 ).

(12) Bugianesi E, Marzocchi R, Villanova N, Marchesini G. Non-alcoholic fatty liver disease/non-alcoholic

steatohepatitis (NAFLD/NASH): treatment. Best Practice & Research in Clinical Gastroenterology

2004 Dec;18(6):11 05-1 1 1 6.

(13) Angulo P, Linder KD. Treatment of non-alcoholic steatohepatitis. Best Practice & Research in Clinical Gastroenterology 2002 Oct:16(5):797-810.

(14) Zang M, Zuccollo A, Hou X, Nagata D, Walsh K, Herscovitz H, et al. AMP-activated protein kinase is

required for the lipid-lowering effect of metformin in insulin-resistant human HepG2 cells. J Biol Chem

2004 Nov 12;279(46):47898-47905.

(15) Yin W, Mu J, Birnbaum MJ. Role of AMP-activated protein kinase in cyclic AMP-dependent lipolysis in

3T3-L 1 adipocytes. Journal of Biological Chemistry 2003 Oct 31 ;278(44 ):4307 4-43080.

33

Chapter 2

( 16)

(17)

(18)

(19)

(20)

(21 )

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31 )

(32)

(33)

(34)

(35)

(36)

(37)

34

Zhou GC, Myers R, Li Y, Chen YL, Shen XL, Fenyk-Melody J, et al. Role of AMP-activated protein

kinase in mechanism of metformin action . Journal of Clinical Investigation 2001 Oct; 1 08(8): 1 1 67-

1 1 74.

Dranoff JA, Nathanson MH. Regulation of bile acid transport· beyond molecular cloning. Hepatology

1999 Jun;29(6): 1912-1 913 .

Hardie DG. The AMP-activated protein kinase pathway-new players upstream and downstream. J

Cell Sci 2004 Nov 1 ; 1 1 7(Pt 23):5479-5487.

Suzuki A, Kusakai G, Kishimoto A, Lu J, Ogura T, Esumi H. ARK5 suppresses the cell death induced

by nutnent starvation and death receptors via inhibition of caspase 8 activation, but not by

chemotherapeutic agents or UV irradiation. Oncogene 2003 Sep 1 8;22(40):6177-6182.

Kimura N, Tokunaga C, Dalal S, Richardson C, Yoshino K, Hara K, et al. A possible linkage between

AMP-activated protein kinase (AMPK) and mammalian target of rapamycin (mTOR) signalling

pathway. Genes Cells 2003 Jan;8(1 ):65-79.

Rattan R, Giri S, Singh AK, Singh I . 5-aminoimidazole-4-Carboxamide-1 -beta-4-Ribofuranoside

inhibits cancer cell proliferation in vitro and in vivo via AMP-activated protein kinase (AMPK). J Biol

Chem 2005 Sep 2 1 .

Hawley SA, Gadalla AE, Olsen GS, Hardie DG. The antidiabetic drug metformin activates the AMP­

activated protein kinase cascade via an adenine nucleotide-independent mechanism. Diabetes 2002

Aug;51 (8):2420-2425.

Musi N, Hirshman MF, Nygren J, Svanfeldt M, Bavenholm P, Rooyackers 0, et al Metformin

increases AMP-activated protein kinase activity in skeletal muscle of subjects with type 2 diabetes.

Diabetes 2002 Jul;51 (7):2074-2081 .

Fryer LGD, Parbu-Patel A, Carling D. The anti-diabetic drugs rosiglitazone and metformin stimulate

AMP-activated protein kinase through distinct signaling pathways. Journal of Biological Chemistry

2002 Jul 1 2;277(28):25226-25232.

Me1sse D, Van de Casteele M, Beauloye C, Hainault I , Kefas BA, Rider MH, et al. Sustained

activation of AMP-activated protein kinase induces c-Jun N-terminal kinase activation and apoptosis

1n liver cells. FEBS Lett 2002 Aug 28;526(1 -3):38-42.

Luo Z, Saha AK, Xiang X, Ruderman NB. AMPK, the metabolic syndrome and cancer. Trends

Pharmacol Sci 2005 Feb;26(2):69-76.

Campas C, Lopez JM, Santidrian AF, Barragan M, Bellosillo B, Colomer D, et al. Acadesine activates

AMPK and induces apoptosis in B-cell chronic lymphocytic leukemia cells but not in T lymphocytes.

Blood 2003 May 1 ; 101 (9):3674-3680.

Abbud W, Habinowski S, Zhang JZ, Kendrew J, Elkairi FS, Kemp BE, et al. Stimulation of AMP­

activated protein kinase (AMPK) is associated with enhancement of G/ut1 -mediated glucose

transport. Arch Biochem Biophys 2000 Aug 1 5;380(2):347-352.

Reiter AK, Bolster DR, Crozier SJ, Kimball SR, Jefferson LS. Repression of protein synthesis and

mTOR signaling in rat liver mediated by the AMPK activator aminoimidazole carboxamide

ribonucleoside. Am J Physiol Endocrinol Metab 2005 May;288(5):E980-E988.

Suzuki A, Kusakai G, Kishimoto A, Shimojo Y, Miyamoto S, Ogura T, et al. Regulation of caspase-6

and FLIP by the AMPK family member ARK5. Oncogene 2004 Sep 1 6;23(42):7067-7075.

Tee AR, Blenis J. mTOR, translational control and human disease. Seminars in Cell & Developmental

Biology 2005 Feb;16(1 ):29-37.

Hay N, Sonenberg N. Upstream and downstream of mTOR. Genes & Development 2004 Aug

15 ; 18(16):1 926-1 945.

Carrera AC. TOR signaling in mammals. Journal of Cell Science 2004 Sep 1 5;1 17(20):461 5-461 6

Proud CG. The multifaceted role of mTOR i n cellular stress responses. Ona Repair 2004 Aug:3(8-

9):927-934.

Hahn-Windgassen A, Nogueira V, Chen CC, Skeen JE, Sonenberg N, Hay N. Akt activates the

mammalian target of rapamycin by regulating cellular ATP level and AMPK activity. J Biol Chem 2005

Sep 1 6;280(37):32081 -32089.

Huang SL, Bjomsti MA, Houghton PJ. Rapamycins - Mechanism of action and cellular resistance.

Cancer Biology & Therapy 2003 May;2(3):222-232.

Foster FM, Traer CJ, Abraham SM, Fry MJ. The phosphomosit1de (Pl) 3-kinase family. J Cell Sci 2003

Aug 1 ;1 16(Pt 1 5):3037-3040.

Metformin protects rat hepatocytes against bile acid-induced apoptosis

(38) Qiao L, Yacoub A, Studer E, Gupta S, Pei XY, Grant S, et al. Inhibition of the MAPK and Pl3K

pathways enhances UDCA-induced apoptosis in primary rodent hepatocytes. Hepatology 2002

Apr;35(4 ):779-789.

(39) Schoemaker MH, Ros JE, Homan M, Trautwein C, Liston P, Poelstra K, et al. Cytokine regulation of

pro- and anti-apoptotic genes in rat hepatocytes: NF-kappaB-regulated inhibitor of apoptosis protein 2

(clAP2) prevents apoptosis. J Hepatol 2002 Jun;36(6):742-750. (40) Hatano E, Brenner DA. Akt protects mouse hepatocytes from TNF-alpha- and Fas-mediated

apoptosis through NK-kappa B activation. Am J Physiol Gastrointest Liver Physiol 2001

Dec;281 (6 ):G 1357-G 1 368.

(41) Webster CR, Anwer MS. Phosphoinositide 3-kinase, but not mitogen-activated protein kinase,

pathway is involved in hepatocyte growth factor-mediated protection against bile acid-induced

apoptosis in cultured rat hepatocytes. Hepatology 2001 Mar;33(3):608-615.

(42) Schoemaker MH, Moshage H. Defying death: the hepatocyte's survival kit. Clin Sci (Lond) 2004

Jul;107(1 ):1 3-25.

(43) Miyoshi H, Rust C, Guicciardi ME, Gores GJ. NF-kappaB is activated in cholestasis and functions to

reduce liver injury. Am J Pathol 2001 Mar;158(3):967-975.

(44) Schoemaker MH, Gommans WM, Conde de la Rosa L, Homan M, Klok P, Trautwein C, et al.

Resistance of rat hepatocytes against bile acid-induced apoptosis in cholestatic liver injury is due to

nuclear factor-kappa B activation. J Hepatol 2003 Aug;39(2):153-161.

(45) Schoemaker MH, Conde de la Rosa L, Buist-Homan M, Vrenken TE, Havinga R, Poelstra K, et al.

Tauroursodeoxycholic acid protects rat hepatocytes from bile acid-induced apoptosis via activation of

survival pathways. Hepatology 2004 Jun;39(6):1563-1 573. (46) Conde de la Rosa L, Schoemaker MH, Vrenken TE, Buist-Homan M, Havinga R, Jansen PL, et al.

Superoxide anions and hydrogen peroxide induce hepatocyte death by different mechanisms:

involvement of JNK and ERK MAP kinases. J Hepatol 2006 May;44(5):918-929. (47) Blokzijl H, Vander Borghi S, Bok LI , Libbrecht L, Geuken M, van den Heuvel FA, et al. Decreased P­

glycoprotein (P-gp/MDR1) expression in inflamed human intestinal epithelium is independent of PXR

protein levels. lnflamm Bowel Dis 2007 Jun;13(6):710-720. (48) Reinehr R, Graf D, Haussinger D. Bile salt-induced hepatocyte apoptosis involves epidermal growth

factor receptor-dependent CD95 tyrosine phosphorylation. Gastroenterology 2003 Sep;125(3):839-

853.

(49) Kefas BA, Cai Y, Kerckhofs K, Ling ZD, Martens G, Heimberg H, et al. Metformin-induced stimulation

of AMP-activated protein kinase in beta-cells impairs their glucose responsiveness and can lead to

apoptosis. Biochemical Pharmacology 2004 Aug 1 ;68(3 ):409-416.

(50) Tamatani M, Che YH, Matsuzaki H, Ogawa S, Okado H, Miyake S, et al. Tumor necrosis factor

induces Bcl-2 and Bcl-x expression through N FkappaB activation in primary hippocampal neurons. J

Biol Chem 1999 Mar 26;274(13):8531-8538.

(51) Chen C, Edelstein LC, Gelinas C. The Rel/NF-kappaB family directly activates expression of the

apoptosis inhibitor Bcl-x(L). Mol Cell Biol 2000 Apr;20(8):2687-2695.

(52) Chu SH, Lee-Kang J, Lee KH, Lee K. Roles of reactive oxygen species, NF-kappaB, and

peroxiredoxins in glycochenodeoxycholic acid-induced rat hepatocytes death. Pharmacology 2003

Sep;69(1 ):12-19.

(53) Bucher BT, Feng XS, Jeyabalan G, Zhang BC, Shao LF, Guo Z, et al. Glycochenodeoxycholate

(GCDC) inhibits cytokine induced NOS expression in rat hepatocytes. Journal of Surgical Research

2007 Mar;138(1 ):15-21.

(54) Turner DJ, Alaish SM, Zou T, Rao JN, Wang JY, Strauch ED. Bile salts induce resistance to apoptosis

through NF-kappaB-mediated XIAP expression. Ann Surg 2007 Mar:245(3):415-425.

(55) Hirano F, Haneda M, Makino I. Chenodeoxycholic acid and taurochenodexycholic acid induce anti­

apoptotic clAP-1 expression in human hepatocytes. J Gastroenterol Hepatol 2006 Dec;2 1 (1 2):1807-

1813.

(56) Mori M, Uchida M, Watanabe T, Kirito K, Hatake K, Ozawa K, et al. Activation of extracellular signal­

regulated kinases ERK1 and ERK2 induces Bcl-xL up-regulation via inhibition of caspase activities in

erythropoietin signaling. J Cell Physiol 2003 May;195(2):29-297.

(57) Hattori Y, Suzuki K, Hattori S, Kasai K. Metformin inhibits cytokine-induced nuclear factor kappaB

activation via AMP-activated protein kinase activation in vascular endothelial cells. Hypertension 2006

Jun;4 7(6):11 83-1188.

35

Chapter 2

36

(58) lsoda K, Young JL, Zir1ik A, MacFarlane LA, Tsuboi N, Gerdes N, et al . Metformin inhibits

proinflammatory responses and nuclear factor-kappaB in human vascular wall cells. Arterioscler

Thromb Vase Biol 2006 Mar;26(3):61 1 -617.

Chapter 3

The active metabolite of leflunomide, A77 1 726, protects rat hepatocytes against bile acid-induced apoptosis

Titia E. Vrenken, Manon Buist-Homan, Allard Jan Kalsbeek, Klaas Nico Faber, Han Moshage

Department of Gastroenterology and Hepatology; University Medical Center Groningen, University of Groningen, Groningen, the Netherlands

Submitted

Chapter 3

3.1 Abstract

Background: Leflunomide is used in the treatment of autoimmune diseases as anti­inflammatory agent. Leflunomide and its active metabolite A77 1726 modulate mitogen­activated protein kinases (MAPK), Src kinases, the phosphoinositide-3 kinase (Pl3K)/Akt­pathway and nuclear factor (NF)-KB activation. Both cell protective and cytotoxic effects of leflunomide have been described. Aim: Since leflunomide affects pathways involved in hepatocyte cell survival, we examined the effects of A77 1726 on hepatocyte cell death. Methods: Primary rat hepatocytes were exposed to the bile acid glycochenodeoxycholic acid (GCDCA), the superoxide anion donor menadione, or tumor necrosis factor (TNF) a in combination with A77 1 726. Activation of MAPK members was determined by Western blot analysis. Apoptosis and necrosis were analyzed by acridine orange staining and caspase activity, and Sytox Green staining, respectively. Results: A77 1726 dose-dependently reduces GCDCA-induced apoptosis and necrosis, but not menadione- or TNFa/ActD-induced apoptosis. The hepatoprotective effect of A77 1726 does not involve ERK1/2, p38 or Pl3K/Akt activation and is mimicked by Src inhibition. A77 1726 does not inhibit NF-KB activation. Conclusion: A77 1726 inhibits bile acid-induced cell death and does not sensitize primary rat hepatocytes to TNFa. In vivo studies are necessary to confirm the observed protective effects of A77 1726 in animal models for chronic liver injury, since our in vitro experiments suggest that A77 1726 might be useful as therapy for chronic liver diseases accompanied by elevated bile acid levels and inflammation.

38

The active metabolite of leflunomide, A 77 1 726, protects rat hepatocytes against bile acid-induced apoptosis

3.2 Introduction

Chronic cholestatic liver diseases result in cell death of hepatocytes via elevated levels of toxic metabolites, bile acids and reactive oxygen species. In acute liver injury, like acute viral hepatitis, hepatocyte cell death is mainly a consequence of pro-inflammatory cytokines and oxidative stress. Cell death often occurs as a combination of apoptosis and necrosis. Apoptosis or programmed cell death is an energy consuming process, which involves caspases, a family of cysteine proteases. Apoptosis is characterized by chromatin condensation, nuclear fragmentation and cell shrinkage, leading to the formation of apoptotic bodies. In contrast, necrotic cell death is a passive process, which via mitochondrial and cellular swelling results in plasma membrane leakage and subsequent release of cellular content into the extracellular environment and circulation, thus triggering inflammation ( 1 -8).

Leflunomide (N-[(4-trifluoromethyl)phenyl]-5-methylisoxazole-4-carboxamide) or Arava® is a well-known anti-inflammatory drug, used in the treatment of autoimmune diseases, particularly rheumatoid arthritis. Furthermore, leflunomide is used in the protection against rejection after organ transplantation, and in the treatment of Crohn's disease and cancer. After administration, the prodrug leflunomide, is almost completely metabolized (>95%) to its active compound A77 1726 or teriflunomide (2-cyano-3-oxo-N-[(4-trifluoromethyl)phenyl]butyramide; 9-14). Proposed molecular mechanisms explaining the anti-inflammatory, anti-proliferative and immunosuppressive effects of leflunomide are inhibition of protein tyrosine kinases, e.g kinases of the Src family, blocking dihydroorotate dehydrogenase (DHODH) involved in pyrimidine synthesis and progression through the cell cycle, inhibition of mitogen activated protein kinase (MAPK) members, including the pro­apoptotic Jun N-terminal kinase (JNK), and reduced production of pro-inflammatory cytokines via inhibitory effects on the transcription factor nuclear factor- B (NF- B; 9, 1 4-1 8). Leflunomide was shown to protect against acetaminophen (APAP)-induced cell death in human hepatocytes (9) and diminished APAP-mediated liver injury in mice (19). The protective mechanisms included inhibition of cytochrome c release from the mitochondria, and reduced activation of both the pro-apoptotic MAPK family member JNK1/2 and the pro­apoptotic protein, Bcl-2 (9, 19). In primary human hepatocytes (20) and erythroleukemia cells (21) A77 1 726 reduced activation of p38 and JNK1/2. The Aki survival pathway and extracellular signal regulated kinase (ERK) 1/2 were activated upon leflunomide treatment in these cells (21 ). Variable effects of leflunomide on NF-KB activation have been reported: 1 ) leflunomide did not affect NF-KB signalling in primary human hepatocytes and erythroleukemia cells (20, 21 ); 2) leflunomide suppressed NF-KB activation in Jurkat, myeloid, cervical epithelial and glioma cells (15, 16), and in an in vivo study of concanavalin A (Con A)-induced hepatitis in mice (1 0). The cytoprotective effect of leflunomide is illustrated by the reduction of Con A-induced aminotransferase levels in the serum, caspase activities, plasma cytokine concentrations, and p38 activity in mice (10), a significant improvement of bile duct ligation (BDL)-induced liver injury (22) and reduced hepatocyte damage after ischemia/reperfusion (1/R) injury (11 ). In contrast, in rat thymocytes apoptosis was induced by A77 1726 (23). The synthetic analogue of A77 1726, called LFM-A12 (a­cyano-l3-hydroxy-l3-methyl-N-[4-(trifluoromethoxy)phenyl]-propenamide), induced apoptosis in breast cancer cells in vitro (24). Thus, both cell protective as well as cytotoxic effects of leflunomide have been described.

Considering this discrepancy in actions of leflunomide, and the variable effects on signal transduction pathways that are also involved in hepatocyte survival, including MAPK members, Src-family members, NF-KB signalling and the phosphoinositide-3 kinase (Pl3K)/Akt survival pathway, we investigated the effects of the prodrug and its active metabolite A77 1726 on hepatocyte cell death. The effect of leflunomide in 3 different in vitro models was examined. We report a cytoprotective effect of A77 1726 against bile acid­induced hepatocyte cell death, which is independent of the anti-apoptotic MAPK members, p38 and ERK1/2 and the Pl3K/Akt signalling pathway. Furthermore, no effect of A77 1 726 on NF-KB activation in hepatocytes was observed.

39

Chapter 3

3.3 Materials and Methods

Animals Specified pathogen-free male Wistar rats (200-250 g) were purchased from Charles River Laboratories Inc. (Wilmington, MA, USA). Rats were housed under standard laboratory conditions with free access to standard laboratory chow and water. Prior to isolation, rats were fasted overnight. Experiments were performed following the guidelines of the local Committee for Care and Use of laboratory animals.

Rat hepatocyte isolation Hepatocytes were isolated from rats by a two step collagenase perfusion as described before (25). Cell viability was greater than 85% as determined by Trypan Blue exclusion. After isolation, 112,500 cells per cm2 were plated on Vitrogen® (Cohesion Technologies Inc. , Palo Alto, CA, USA) coated plates in William's E medium (lnvitrogen, Breda, The Netherlands) supplemented with 50 µg/mL gentamycin (lnvitrogen) and penicillin­streptomycin-fungizone (Lonza, Verviers, Belgium). During the attachment period (4 hrs) 50 nmol/L dexamethasone (Dept. of Pharmacy UMCG, Groningen, The Netherlands) and 5% fetal calf serum (lnvitrogen) were added to the medium. Cells were cultured in a humidified incubator at 3rc and 5% CO2.

Experimental design Experiments were started immediately after the 4 hrs hepatocyte attachment period. Monolayer cultures were exposed to 50 µmol/L glycochenodeoxycholic acid (GCDCA, Calbiochem, La Jolla, CA, USA) for 4 hrs. Alternatively, hepatocytes were treated for 12 hrs with 50 µmol/L of the superoxide anion donor, menadione (2-methyl-1,4-naphthoquinone, Sigma-Aldrich, St. Louis, MO, USA) or for 16 hrs with 20 ng/ml recombinant murine tumor necrosis factor a (mTNFa, R&D Systems, Abingdon, United Kingdom) in combination with 200 ng/ml of the transcriptional inhibitor actinomycin D (ActD, Roche Diagnostics, Almere, The Netherlands) or for 4 hrs with a cytokine mixture (CM) consisting of 20 ng/ml mTNFa, 10 ng/ml recombinant human interleukin-113 (hlL-113, R&D Systems), 10 ng/ml recombinant rat interferon-y (rlFN y, R&D Systems) and 10 µg/ml lipopolysaccharide (LPS, Escherichia coli, serotype 0127:BB, Sigma-Aldrich). Leflunomide (Sigma-Aldrich) and its active metabolite, A77 1726 (also known as teriflunomide, Calbiochem) were used at concentrations ranging from 1-50 µmol/L. MAPK members were specifically blocked using 10 µmol/L of the cell permeable p38 kinase inhibitor SB203580 (Calbiochem), 10 µmol/L of U0126 (Promega, Madison, WI, USA), the inhibitor of the ERK1/2 signalling cascade or 5 µmol/L of the JNK1/2-inhibitor SP600125 (lnvitrogen). The Src-family of tyrosine kinases were inhibited using 10 µmol/L SU6656 (Calbiochem) The Pl3K pathway was inhibited using L Y294002 (50 µmol/L, Calbiochem). All inhibitors were added 30 minutes prior to GCDCA. Each experimental condition was performed in triplicate. Cells were harvested at the indicated time points after the addition of the apoptotic stimuli, and washed three times with ice cold phosphate-buffered saline (PBS) before the addition of hypotonic cell lysis buffer (protein analysis, caspase-3/-6 assay), Tri-reagent (RNA isolation, Sigma-Aldrich) or 2x concentrated sample buffer (Western blot analysis) as described previously (26).

Each experiment was performed at least three times, using hepatocytes from different isolations.

Caspase enzyme activity assays Caspase-3 like activity was assayed as described previously (26). The arbitrary fluorescence unit (AFU) was corrected for the amount of protein in the cell lysate. Caspase-6 like activity was assayed according to the manufacturer's instructions (Biovision, Mountain View, CA, USA)

40

The active metabolite of leflunomide, A 77 1 726, protects rat hepatocytes against bile acid-induced apoptosis

Sytox green and acridine orange nuclear stainings To determine necrotic cell death, hepatocytes were incubated for 1 5 minutes with Sytox green (lnvitrogen) nucleic acid stain. Sytox binds DNA, but can only enter cells with leaky plasma membranes, characteristic of necrotic cell death. Sytox cannot cross the membranes of viable cells or apoptotic bodies. Hepatocytes exposed to 5 mmolll hydrogen peroxide (H2O2, Merck Chemicals Ltd, Nottingham, UK) for 6 hrs served as positive control for necrosis. Apoptotic nuclei were visualized with acridine orange (Sigma-Aldrich) as described previously (25). Fluorescent nuclei were visualized using an Olympus CKX41 microscope at 450-490 nm.

Western blot analysis Western blot analysis of cell lysates treated for 1 hr with GCDCA and/or A77 1726 was performed by SOS-PAGE followed by semi dry-blotting to transfer the proteins to Hybond ECL nitrocellulose membrane (Amersham Biosciences, Piscataway, NJ, USA). Ponceau S (Sigma-Aldrich) staining (0.1 % w/v) was used to monitor electrophoretic transfer. Activation of the MAP kinases JNK, p38 or ERK1 /2 was detected using polyclonal antibodies against phosphorylated JNK (lnvitrogen), phosphorylated p38 (Cell Signaling Technology, Inc, Danvers, MA, USA) or the monoclonal antibody against phosphorylated ERK1/2 (New England Biolabs, Leusden, The Netherlands) at a dilution of 1 :500. After Western blot analysis, blots were stripped as described previously (25) and incubated with polyclonal antibodies against total JNK, p38 (Cell Signaling Technology, Inc) or ERK1/2 (Santa Cruz Biotechnology, Inc, Heidelberg, Germany) respectively (dilution 1 : 1 000). Western blots were quantified and band intensities of phosphorylated protein were expressed as ratio of the intensity of the corresponding total protein.

RNA isolation and quantitative polymerase chain reaction (PCR) RNA isolation was performed as described previously (27). RNA concentration was determined with the Ribogreen RNA quantitation reagent and kit (lnvitrogen). Reverse transcription PCR (RT-PCR) was carried out on 2.5 µg of total RNA using random primers in a total volume of 50 µL using the moloney murine leukemia virus (M-MLV) reverse transcriptase system (Sigma-Aldrich) according to the manufacturer's instruction. Quantitative PCR was performed on 4 µL of 20-times diluted complementary DNA in a final volume of 20 µL (27). Fluorescence was measured using the ABI PRISM 7700 Sequence Detector version 1 .7 software (Applied Biosystems, Nieuwerkerk aid IJssel, The Netherlands) starting with 10 minutes at 95°C, followed by 40 cycles of 1 5 seconds at 95°C and 1 minute at 60°C. Details of primers and probes are listed in Table 1. Each sample was analyzed in duplicate. 18S mRNA levels were used as endogenous control. Table 1. Sequences of primers and probes used for quantitative PCR analysis

18S Sense 5'-CGG CTA CCA CAT CCA AGG A-3'

rat Antisense 5'-CCA ATT ACA GGG CCT CGA AA-3'

Probe 5'FAM-CGCGCAAA TT ACCCACTCCCGA-TAMRA3'

!Nos Sense 5'-GTG CTA ATG CGG AAG GTC ATG-3'

rat Antisense 5'-CGA CTT TCC TGT CTC AGT AGC AAA-3'

Probe 5'FAM-CCC GCG TCA GAG CCA CAG TCC T-TAMRA3'

Statistical analysis All numerical results are reported as the mean of at least 3 independent experiments ± standard error of the mean. A Mann-Whitney test was used to determine the significance of differences between experimental groups. A P-value smaller than 0.05 was considered to be statistically significant.

41

Chapter 3

3.4 Results

A77 1 726, the active metabolite of leflunomide, dose-dependently reduces bile acid­induced apoptosis The effects of the drug leflunomide and its active metabolite, A77 1726, were examined in an in vitro model of acute bile acid toxicity: GCDCA-induced apoptosis. Leflunomide, did not change GCDCA-induced caspase-3 activity at concentrations up to 50 µmol/L (Fig 1A). In contrast, A77 1726 dose-dependently reduced GCDCA-induced caspase-3 activity in the concentration range of 1 µmol/L (27.5% reduction) to 50 µmol/L (79.2% inhibition; Fig 1 B). A77 1726 alone did not induce caspase-3 activity (data not shown). Since A77 1726 at a concentration of 25 µmol/L inhibited GCDCA-induced caspase-3 activity comparable to 50 µmol/L A77 1726, 73.7% and 79.2% respectively, we decided to use the lower concentration (25 µmol/L) of A77 1726 to study the protective effects in more detail. At this concentration, A77 1726 (25 µmol/L) also decreased GCDCA-induced caspase-6 activity (51.1 %; Fig 1 C). These results were confirmed by acridine orange staining. Hepatocytes exposed to GCDCA for 4 hrs demonstrate apoptotic nuclei in 52.6% of the cells (Fig 1 D), whereas in combination with A77 1726 the percentage of condensed nuclei is reduced to 33.7 %. In control or A77 1726-treated hepatocytes apoptotic cells were nearly absent.

A ,.

c:::::J control + 10 1,!M

al\unomida GCDCA

C ..

B

+ SO µM aftunomida

+ATTH28 2S µM GCDCA

D

'"1

,o I

control

Control, 4hra '8%

1 µM

GCDCA50.,iM 52.IY,

+

#

•µM 1 0 µM 25 µM SO µM

+An•m

GCOCA

-GCOCA !IO l,III + An 172925µ11 ATT 1T292S 1,111 31%

33.7%

Figure 1. The active metabol te of leflunomlde. A77 1 726 reduces bile acid induced apoptosls. Primary rat hepatocytes were exposed to (A) leflunom,de or 1B) A77 1726 added 30 minutes prior to GCDCA (50 µmol/L. 4 hrs) Caspase-3 like activity was measured, and is shown as fold induction compared to controt values· control values were set at one. (C)

A77 1 726 also reduces GCOCA induced caspase-6 act v1ty. Caspase-6 hke act1v1ty was measured 1n primary rat hepatocytes exposed to GCDCA with or without A77 1726. Caspase-6 like act1V1ty 1s shown as fo d induction compared to control, control values were set at one. (0) A77 1726 prevents GCDCA (4 hrs) nduced nuclear condensation as indicated by acnd1ne orange staining and expressed as ,.o of condensed nuclei. Magnification 20X. Statistical analysis • p<0 05 or + p<0 01 compared to control cells # p<0.05 or p<0.01 compared to GCDCA treated ce Is.

A77 1 726 does not protect against cytokine or oxidative stress induced apoptosis Since these results indicate that A77 1726 protects hepatocytes against bile acid-induced apoptosis, we decided to study the effects of A77 1726 also in other models of liver cell toxicity. Two models were examined, an in vitro model for acute liver failure induced by TNFa in combination with ActD, and receptor-independent oxidative stress-induced cell death, elicited by the superoxide anion donor menadione. A77 1726 (25 µmol/L) did not change TNFa/ActD-induced caspase-3 activity (Fig 2A). In contrast, menadione-induced caspase-3 activity was even further increased, although not significantly, by A77 1726 (Fig 28).

42

The active metabolite of /eflunomide, A 77 1 726, protects rat hepatocytes against bile acid-induced apoptosis

40

control

TNFJActD

+

'""""

+

+An 1 728 25 �

n1enmUone

Figure 2. A77 1726 does not protect hepatocytes against cytokine- or oxidative stress-induced apoptosis Primary rat hepatocytes were exposed to A77 1 726 added 30 minutes prior to (A) TNFa/ActD (20 nglml, 200 nglml, 16 hrs) or (B) menadione (50 µmol/L, 12 hrs) Caspase-3 like activity was measured, and is shown as fold induction compared to control values, control values were set at one Statistical analysis • p<0.05 or + p<0.01 compared to control

A77 1 726 blocks apoptotic cell death without inducing necrotic cell death Next, we studied whether the reduction in apoptosis by A 77 1726 resulted in a shift to necrosis. The percentages of necrotic cells were determined both at the peak of caspase-3 activity (4 hrs) as well as after overnight exposure. Hepatocytes exposed to GCDCA for 4 hrs resulted in necrosis in 12% of the cells, while 6% of the cells treated with GCDCA and A77 1726 were necrotic (Fig 3). After overnight exposure, A77 1726 also significantly lowered GCDCA-induced necrosis {data not shown). In control cultures or hepatocytes exposed to A77 1726 no significant number of necrotic cells were observed (Fig 3) .

Control 1.7%

GCDCA 12.3'4

GCDCA + A77 1726 '

6.4%

A77 1726 2.8%

•

Figure 3. A77 1 726 reduces necrotic cell death Cells were treated without (control) or with GCDCA (50 µmol/L) for 4-16 hrs A 77 1 726 was added 30 minutes before GCDCA. Sytox green nuclear staining was used to determine

positive control >95%

necrotic cell death Cells stimulated with H202 (5 mmoUL, 6 hrs) were used as positive control Representative images 4 hrs after stimulation are shown Upper panel: combined phase contrast and fluorescence. Lower panel fluorescence. Magnification 20X Statistical analysis # p<0.05 compared to GCDCA-treated cells.

A77 1 726 is protective after addition of bile acids Subsequently, we examined whether A77 1726 also protects hepatocytes when added after GCDCA exposure. A 77 1726 was added either 30 minutes prior to, simultaneously with or 1, 2 or 3 hrs after addition of the bile acid. A77 1726 significantly reduced GCDCA-induced

,. caspase-3 activity when added up to 1 + hr after GCDCA (Fig 4). At all three

l 2 1n •'* , .......

time points of A77 1726 addition, pre­incubation, simultaneously with or one hr after bile acid, around 65% reduction in caspase-3 activity was observed. When added 2 or 3 hrs after GCDCA, A77 1726 still decreased GCDCA-induced caspase-3 activity, although not significantly (Fig 4).

Figure 4. A77 1726 protects against apoptosis when added after GCDCA. Rat hepatocytes were treated without (control) or with GCDCA (50 µmol/L) for 4 hrs. A77 1726 (25 µmol/L) was added 30 minutes prior to GCDCA, simultaneously with the bile acid or 1, 2 and 3 hrs after addition of GCDCA. Caspase·3 like activity was measured, and is shown as fold induction compared to control values, control values were set at one Statistical analysis: + p<0.01 compared to control # p<0.05 compared to GCDCA treated cells.

43

Chapter 3

Our results suggest that A77 1726 does not have a major inhibitory effect on GCDCA-uptake by bile acid transporters like ntcp.

The hepatoprotective effect of A77 1 726 does not require ERK1/2 and p38 MAPK activation, is independent of the PIK/Akt survival pathway and is mimicked by Src­inhibition The involvement of the anti-apoptotic members of the MAPK family, ERK1/2 and p38, in the protection of A77 1726 was examined using the specific inhibitors U0126 and SB203580, respectively. The reduction in GCDCA-induced caspase-3 like activity caused by A77 1726 was partially reversed by treatment with U0126 or SB203580 (Fig 5A), but this reversal was not statistically significant. Both inhibitors alone had no effect on GCDCA-induced caspase-3 activity (data not shown) as described previously (26). Subsequently, we studied the involvement of ERK1/2 and p38 activation in the protective mechanism of A77 1726 by Western blot analysis. Similar levels of phosphorylated ERK1/2 (Fig 5B) were detected in GCDCA and/or A77 1726-treated hepatocytes compared to control cells. Phosphorylated p38 (Fig 5C) appeared to be slightly increased in hepatocytes treated with GCDCA and/or A77 1726 compared to control cells, but no significant difference between exposure to GCDCA or GCDCA in combination with A77 1726 was observed. Together, these results suggest that both ERK1/2 and p38 MAP kinase are not required for the protective effect of A77 1726. A '] ,1

ii ' U J

C oon1ro1 GCOCA

I ----1-,. , _____ , .,,

! '

i ..

I o: ...,.._.-.__."""' __ ....,•,•, ,,'-n A.11112'

........ ·-­+An 1728 25t"!

GCDCA

B - GCOCA GCOCA• A.77 1728 An 1720

1-�� � --.....,I EA11;112 ----�.__.

D GCOCA • ....., GCOCA An1ne An1ne

I ,-• i ( ,.

+ .. ..

h ii ; I: � j 30

h u ll

- 0 '"""" + Src inhibm OCDCA

Figure 5. The protection of A77 1726 does not involve ERK1/2 and p38 MAPK members ncr the Pl3K/Akt survival pathway. Pr mary rat hepalocytes were exposed to GCDCA and A77 1726 for 4 hrs with (A) the ERK1/2 inhibitor U0126 (10 µmol/L), the p38 inhibitor 5B203580 (10 µmol/L), the JNK inhibitor SP600125 (5 µmoVL) or (E) the Pl3-kinase inhibitor L Y294002 (50 µmol/L) and (F) the Src inhibitor SU6656 (10 µmol/L). Caspase-3 like activity was measured, and 1s shown as fold induction compared to control; control values were set at one. Statishcar analysis: • p<0 05 or + p<0 01 compared to control. # p<0.05 compared to GCDCA treated eel s Western blot analysis for (B) phosphorylated and total ERK1/2, (C) phosphorylated and total p38 and (D) phosphorylated and total JNK1/2 on hepatocytes treated without (control) or with GCDCA (50 µmol/L, 1 hr) in combination wilh A77 1726 (25 µmol/L). Western blots were quantified and band intensities of phosphorylated protein were expressed as ratio of the intensity of the corresponding total protein. Data is presented as mean of at least three independent expe,,ments +/. S E.M. Control values were set at one. A representative Western blot is shown.

44

The active metabolite of leflunomide, A 77 1 726, protects rat hepatocytes against bile acid-induced apoptosis

The MAPK member J NK is pro-apoptotic in bile acid-induced apoptosis, and inhibition of JNK phosphorylation and activation using the inhibitor SP600125 reduced GCDCA-induced caspase activity and apoptosis (Fig 5A). Since A77 1726 mimicked the effect of the J NK1/2 inhibitor SP600125 on GCDCA-induced apoptosis, we next determined whether A 77 1726 acts as JNK-inhibitor. The levels of phosphorylated JNK1 and 2 were equal in hepatocytes treated with GCDCA, A77 1726 or both (Fig 5D). Next, we investigated whether the Pl3K/Akt survival pathway is involved in the anti-apoptotic actions of A77 1726. As shown in figure 5E, inhibition of the Pl3K pathway using LY294002 had no effect on the protection by A77 1726 against GCDCA-induced apoptosis. As described previously (26), GCDCA-induced caspase-3 activity was significantly increased when the Pl3K pathway was inhibited (data not shown). Finally, we showed that the effect of A77 1726 was mimicked by the Src inhibitor SU6656 (Fig. 5F).

A77 1 726 does not inhibit NF-KB activation Since it has been reported that A77 1726 inhibits TNFa-mediated NF-KB activation in inflammatory and immune cells (15, 16), we examined the role of NF-KB signalling in the anti-apoptotic actions of A77 1726. First, the effects of A77 1726 on the NF-KB-dependent cytokine-induced inducible nitric oxide synthase (iNOS) expression were studied. Hepatocytes exposed to cytokine mixture showed a 270-fold relative induction of iNOS expression compared with control cells (Fig 6A). In contrast to actinomycin D, which completely prevents cytokine-induced iNOS expression, A77 1726 did not affect iNOS expression (Fig 6A). Next, we confirmed the qPCR results by determining apoptosis. Both hepatocytes treated with TNFa or TNFa plus A77 1726 show only base-line caspase-3 activity levels similar to control cells (Fig 68), indicating that A77 1726 does not sensitize hepatocytes to TNFa-induced cell death. In contrast, actinomycin D in combination with TN Fa strongly induced caspase-3 activity.

A

! �I,____ I ____ I +AT11711ZI� •�D --

B

1 1 �I . l I ,-1-�,= ..... =, -----�-...... n-,•,,.•,.-,...�-.Aclnomyc., o T,.,

Figure 6. A77 1726 does nol Inhibit TNFo-induced NF-KB activation. qPCR analysis far iNOS using cDNA al primary rat hepalacytes exposed far 4 hrs lo cytokine mixlure with or withaul A77 1726. The transcriptional inhiMor actinornycin D (200 nglml) served as posnive control for NF-KB inhibition (A) Statistical analysis: + p<0.01 compared ta control. • p<0 01 compared to cells stimulated with cytokine mixture or # p<0.01 compared to cells treated with cytokine mixture + actinomycin D. (B} A77 1 726 does not sensitize hepatocytes to TNFa-induced apoptosis Caspase-3 like actJvIty was measured in primary rat hepatacytes exposed ta TNFo (16 hrs) with or without A77 1726 or aclinamycin D (positive control for TNFa-sens1tizaUon). Caspasa.3 activity is shown as fold induction compared to control values, control values were set at one Statistical analysis • p<0.05 compared to control • p<0.05 compared ta TNFo stimulated cells or # p<0.05 compared ta TNFo/ActD treated cells.

3.5 Discussion

Leflunomide (Arava®) is a drug that is used in the treatment of autoimmune disorders and rheumatoid arthritis. It has anti-inflammatory and anti-proliferative actions and its reported mechanisms of action involve signal transduction pathways that are also relevant in hepatocyte survival and apoptosis. Therefore, we examined the effects of leflunomide and its active metabolite, A77 1726, in models of bile acid, oxidative stress and TNFa-induced liver injury in primary rat hepatocytes. Previous studies reported that leflunomide ameliorated liver damage in hepatocytes and animal models of cholestasis including Con A-induced hepatitis, drug-induced cytotoxicity and I/R injury (9-11, 19, 22). In the present study, we show that leflunomide does not affect GCDCA-induced apoptosis, while its active metabolite A77 1726 is protective against cell death induced by GCDCA. We examined the effects of both compounds at similar concentrations (µmolar range) confirming previous reports that A77 1726 has a 100-fold higher cytoprotective activity than the pro-drug leflunomide (9).

45

Chapter 3

Leflunomide is metabolized in the gastrointestinal wall, the liver and plasma. The metabolism of leflunomide to A77 1726 was analyzed in in vitro experiments using rat plasma, whole blood and liver microsomes. In plasma or whole blood a t112 for leflunomide of 36 and 59 minutes respectively was observed, while in liver microsomes only 5 to 6 % was metabolized in 30 minutes. However, in the presence of NADPH the t1 12 i ncreased to 40 minutes (28). Thus, apart from the variation in cytoprotective potency between leflunomide and A77 1726, the lack of protection by the prodrug leflunomide may be caused by insufficient metabolism into A77 1 726 within the time frame of our experiments.

A77 1726 reduced several parameters of apoptosis, including caspase-3 and caspase-6 activity, nuclear fragmentation, as well as the number of necrotic cells. A77 1 726 only prevented bile acid-induced cell death, and did not protect hepatocytes against TNFa/ActD or menadione-induced apoptosis, although in these models of cell death caspase activation and mitochondria-derived signals are also of major i mportance in the execution of apoptosis. A likely explanation is that these inducers of cell death activate different intracellular signal transduction pathways involved in cell death and/or cell survival. TNFa is known to activate the transcription factor NF-KB, while GCDCA and menadione do not affect NF-KB activity (25, 29). Bile acids require uptake by specific transporters, e.g. ntcp and Fas/EGF-receptor mediated events (30, 31) to exert their effects (26), while TN Fa signalling is initiated by activation of the TNF-receptor at the cell membrane and apoptosis induced by reactive oxygen species (superoxide anions) is receptor-independent. Differences in the mechanism of cell death are also demonstrated by the fact that the anti-apoptotic bile acid tauroursodeoxycholic acid (TUDCA) inhibits GCDCA-induced apoptosis, but not superoxide anion or TNFa-induced apoptosis (25). In fact, both A77 1 726 and TUDCA increase menadione-induced caspase-3 activity, although not significantly. This slight increase in superoxide anion-induced hepatocyte apoptosis was not further explored.

A77 1726 was still effective when added 1 hour after GCDCA exposure and apoptosis tended to be lower when A77 1726 was added between 1 and 3 hours after GCDCA exposure. This implies that the protective effect of A77 1726 is probably not due to reduced bile acid uptake. Surprisingly, the protection of A77 1 726 was equal when added before, simultaneous with or 1 hr after the bile acid. Since bile acid induced caspase-3 activity peaks at 3-4 hours (26) our results indicate that A77 1 726 can prevent bile acid induced apoptosis at a relatively late stage and that A77 1 726 rapidly activates survival pathways. Numerous studies have implicated the involvement of Pl3K/Akt, Src kinases and MAPK members in the regulation of apoptosis and cell survival. Since these kinases can be activated very rapidly, their role in the protective effect of A77 1 726 was examined. Our results do not support an important role for p38 and ERK1/2 or Pl3K/Akt in the anti-apoptotic actions of A77 1726, since inhibition of these kinases did not reverse the protective effect of A77 1726. Furthermore, A77 1 726 did not change the phosphorylation and hence the activity of these kinases, either in the absence or presence of bile acids. These results are in contrast to other studies, describing inhibitory effects of A77 1 726 on p38 activation (10, 20, 21), and stimulation of ERK1/2 and Akt activity (21). The effects on p38 have been observed in various cell types. Using a mouse model of hepatitis (10) p38 activation in whole liver sections was determined. In this setting, the effects of other liver cell types, like Kupffer cells that may regulate protein kinase phosphorylation are included, while in our in vitro experiments one cell type, primary rat hepatocytes, was studied. The d ifferent findings using primary human (20) or rat hepatocytes compared to whole liver sections can be explained by the involvement of cytokines that are released by non-parenchymal cell types, notably Kupffer cells. A77 1726 is able to inhibit p38-activation induced by IL-1 , however, in our model, p38 was not significantly induced by GCDCA compared to control cells and A77 1726 has no effect on p38. In accordance with a previous study in primary human hepatocytes, A77 1726 alone had no effect on the phosphorylation of the MAPK family members ERK1/2 and p38. In contrast, in two human erythroleukemia cell lines leflunomide not only inhibited p38 activity, but stimulated ERK1 /2 activation (21 ). Furthermore, at variance with our observations, in these cell lines phosphorylation of Akt was induced by leflunomide. These 46

The active metabolite of leflunom1de, All 1 726, protects rat hepatocytes against bile acid-induced apoptosis

differences can be attributed to the use of different cell types and/or the difference between primary cells and (transformed) cell lines.

In contrast to other studies (9, 14, 15, 19-21) A77 1726 did not change the activation of the pro-apoptotic MAPK member JNK in our study. In synoviocytes, A77 1726 slightly increased cytokine-induced J NK activation (14), while in Jurkat cells (15), hepatocytes (9, 20) and liver sections (19) phosphorylated JNK was reduced by A77 1726 or leflunomide.

Our results clearly show that A77 1726 does not inhibit NF-KB activity. Rat hepatocytes are not sensitized to TNFa-induced apoptosis by A77 1726, and A77 1726 does not influence cytokine-induced NF-KB-dependent gene transcription. These data are in line with other studies reporting no effect of A77 1726 or leflunomide on NF-KB activity (20, 21 ), but are in contrast with several reports describing inhibition of NF-KB by leflunomide or its metabolite (10, 16). In both primary human (20) and rat hepatocytes, A77 1726 does not affect NF-KB activation, while NF-KB-activity induced by a variety of stimuli is inhibited in Jurkat, myeloid, epithelial and glioma cells (15, 16). Also, in a mouse model of autoimmune hepatitis, leflunomide inhibited NF-KB in whole liver (10). These contradictory observations appear to be cell type specific. Our observation that leflunomide nor its metabolite A77 1726 inhibit NF-

B activity is very important, since it implies that leflunomide does not sensitize hepatocytes to TNF-induced apoptosis and hence, could be considered for the treatment of cholestatic liver diseases in the presence of inflammation.

None of the investigated signal transduction pathways appeared to be responsible for the hepatoprotective effect of A77 1726 against GCDCA-induced cell death. Therefore, we can only speculate about the signalling pathways that are involved in this protective effect. One possibility comprises the Src-family of tyrosine kinases. Leflunomide was shown to block the activity of the Src members Lek and Fyn in T-cells (16-18). Recent papers reported the involvement of the Src kinase family members, Yes, Fyn and Lek in oxidative stress­mediated cytotoxicity and Yes in bile acid-induced cell death (32, 33). Specific inhibition of Yes by SU6656 blocked tauro-lithocholic acid-3 sulfate (TLCS)-induced apoptosis (33). In our study, we observed a similar effect of SU6656 on GCDCA induced apoptosis. Since inhibition of Yes using SU6656 and exposure to A77 1726 result in similar hepatoprotective effects, and since A 77 1726 does not affect the MAPK family members, nor the Pl3K pathway or NF-KB pathway, we speculate that the Src-kinase Yes could be a potential target for A77 1726.

In summary, we have shown that A77 1726 protecs hepatocytes against bile acid-induced cell death, while A77 1726 has no effect on TNFa/ActD- or superoxide anion-induced apoptosis. The anti-apoptotic MAPK family members, p38 and ERK1/2, and the Pl3K/Akt survival pathway are not involved in the protective effect of A77 1726. Activation of the pro­apoptotic MAPK member JNK is not modulated by A77 1726. Furthermore, in primary rat hepatocytes A77 1726 does not sensitize hepatocytes to TNFa-induced apoptosis. In vivo experiments are necessary to confirm our in vitro observations, and to support or reject our data which indicate that A 77 1726 could be useful as therapy for chronic cholestatic liver diseases accompanied by inflammation.

3.6 Acknowledgements

This study was supported by the Dutch Digestive Foundation (WS 02-13) and the J .K. de Cock Foundation.

3.7 Reference List

(1) Reed JC. Mechanisms of apoptosis. Am J Pathol 2000 Nov;1 57(5):141 5-1430.

(2) Varghese J , Khandre NS, Sarin A. Caspase-3 activation is an early event and initiates apoptotic damage in a human leukemia cell line. Apoptosis 2003 Aug;8(4):363-370.

47

Chapter 3

48

(3) Adams JM, Cory S. Apoptosomes· engines for caspase activation. Curr Opin Cell Biol 2002

Dec;14(6):715-720.

(4) Higuchi H, Gores GJ. Bile acid regulation of hepatic physiology: IV. Bile acids and death receptors. Am J

Physiol Gastromtest Liver Physiol 2003 May;284(5):G734-G738.

(5) Rust C, Karnitz LM, Paya CV, Moscat J, Simari RD, Gores GJ. The bile acid taurochenodeoxycholate

activates a phosphatidylinositol 3-kinase-dependent survival signaling cascade. J Biol Chem 2000 Jun

30;275(26):2021 0-20216.

(6) Kim JS, He L. Lemasters JJ. Mitochondrial permeability transition: a common pathway to necrosis and

apoptosis. Biochem Biophys Res Commun 2003 May 9 ;304(3):463-470.

(7) Cohen GM. Caspases: the executioners of apoptosis. Biochem J 1997 Aug 15;326(Pt 1): 1 -16.

(8) Lemasters JJ. Mechanisms of hepatic toxicity - V. Necrapoptosis and the mitochondrial permeability

transition: shared pathways to necrosis and apoptosis. American Journal of Physiology-Gastrointestinal

and Liver Physiology 1999 Jan;276(1 ):G1-G6.

(9) Latchoumycandane C, Seah QM, Tan RCH, Sattabongkot J, Beerheide W, Boelsterli UA. Leflunomide

or A77 1726 protect from acetaminophen-induced cell injury through inhibition of JNK-mediated

mitochondrial permeability transition in immortalized human hepatocytes. Toxicology and Applied

Pharmacology 2006 Nov 15;217(1 ):1 25-133.

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

lmose M, Nagaki M, Kimura K, Takai S, lmao M, Naiki T, et al. Leflunomide protects from T-cell­

mediated liver inJury in mice through inhibition of nuclear factor kappaB. Hepatology 2004

Nov;40(5): 1 160-1169.

Karaman A. Fadillloglu E, Turkmen E. Tas E, Yilmaz Z. Protective effects of leflunomide against

ischemia-reperfusion inJury of the rat liver. Pediatr Surg Int 2006 May;22(5):428-434. Yao HW, Li J, Chen JQ, Xu SY. Leflunomide attenuates hepatocyte injury by inhibiting Kupffer cells.

World J Gastroenterol 2004 Jun 1 ;10(11 ):1608-1 611 .

Alexander D, Friedrich B. Abruzzese T, Gondolph-Zink B, Wulker N, Aicher WK. The active form of

leflunomide, HMR1726, facilitates TNF-alpha and IL-17 induced MMP-1 and MMP-3 expression. Cell

Physiol Biochem 2006;17(1-2):69-78.

Vergne-Salle P, Leger DY, Bertin P, Treves R, Beneytout JL, Liagre B. Effects of the active metabolite

of leflunomide, A77 1 726. on cytokine release and the MAPK signalling pathway in human rheumatoid

arthritis synoviocytes. Cytokine 2005 Sep 7;31 (5):335-348.

Manna SK, Mukhopadhyay A, Aggarwal BB. Leflunomide suppresses TNF-induced cellular responses:

Effects on NF-kappa B, activator proteln-1, c-Jun N-terminal protein kinase, and apoptosis. Journal of

Immunology 2000 Nov 1 5;165(10):5962-5969. Manna S K, Aggarwal BB. lmmunosuppressive leflunomide metabolite (A77 1726) blocks TNF­

dependent nuclear factor-kappa B activation and gene expression. Journal of Immunology 1999 Feb

15;162(4 ):2095-21 02.

Layseca-Espinosa E, Pedraza-Alva G. Montiel J L, del RR, Fierro NA, Gonzalez-Amaro R, et al. T cell

aggregation induced through CD43: intracellular signals and inhibition by the immunomodulatory drug

leflunomide J Leukoc Biol 2003 Dec;74(6):1083-1093.

Xu X, Williams JW, Bremer EG, Finnegan A, Chong AS. Inhibition of protein tyrosine phosphorylation in

T cells by a novel 1mmunosuppress1ve agent, leflunomide. J Biol Chem 1995 May 26;270(21 ):12398-

1 2403.

Latchoumycandane C, Goh CW, Ong MM, Boelsterli UA. Mitochondrial protection by the JNK inhibitor leflunom1de rescues mice from acetaminophen-Induced liver injury. Hepatology 2007 Feb;45(2):412-

421.

Mig,ta K, Miyashita T, Maeda Y, Nakamura M, Yatsuhashi H , Ishibashi H, et al. An active metabolite of

leflunomlde, A77 1726, inhibits the production of serum amyloid A protein in human hepatocytes

Rheumatology 2005 Apr;44(4 ):443-448.

Leger DY, Liagre B, Beneytout JL. Low dose leflunomide activates P13K/Akt signalling in

erythroleukem,a cells and reduces apoptosis induced by anticancer agents Apoptosis 2006

Oct:11(10):1747-1760.

Karaman A, lraz M, Kirimlioglu H, Karadag N, Tas E, Fadillioglu E. Hepatic damage in biliary-obstructed

rats is ameliorated by leflunomide treatment. Pediatr Surg Int 2006 Sep;22(9):701-708.

Colic M, Popovic P, Vucevic D, Dimitrijevic M. Leflunomide induces apoptosis of thymocytes and T-cell

hybridoma: differences in sensitivity and signaling pathways. Transplant Proc 2001 May;33(3):2344-

2346.

The active metabolite of /eflunomide, A 77 1 726, protects rat hepatocytes against bile acid-induced apoptosis

(24) Ghosh S, Zheng Y, Jun X, Narla RK, Mahajan S, Navara C, et al. Alpha-cyano-beta-hydroxy-beta­

methyl-N-(4-(trifluoromethoxy)phenyl) propenamide: an inhibitor of the epidermal growth factor receptor

tyrosine kinase with potent cytotoxic activity against breast cancer cells. Clin Cancer Res 1998

Nov;4(11 ):2657-2668.

(25) Conde de la Rosa L, Schoemaker MH, Vrenken TE, Buist-Homan M, Havinga R, Jansen PL, et al.

Superoxide anions and hydrogen peroxide induce hepatocyte death by different mechanisms:

involvement of JNK and ERK MAP kinases. J Hepatol 2006 May;44(5):918-929.

(26) Schoemaker MH, Conde de la Rosa L, Buist-Homan M, Vrenken TE, Havinga R, Poelstra K, et al.

Tauroursodeoxycholic acid protects rat hepatocytes from bile acid-induced apoptosis via activation of

survival pathways. Hepatology 2004 Jun;39(6):1563-1573.

(27) Blokzijl H, Vander Borghi S, Bok LI, Libbrecht L, Geuken M, van den Heuvel FA, et al. Decreased P­

glycoprotein (P-gp/MDR1 ) expression in inflamed human intestinal epithelium is independent of PXR

protein levels. lnflamm Bowel Dis 2007 Jun;13(6):710-720. (28) Kalgutkar AS, Nguyen HT, Vaz AD, Doan A, Dalvie DK, McLeod DG, et al. In vitro metabolism studies

on the isoxazole ring scission in the anti-inflammatory agent lefluonomide to its active alpha-cyanoenol

metabolite A771726: mechanistic similarities with the cytochrome P450-catalyzed dehydration of

aldoximes. Drug Metab Dispos 2003 Oct:31 (10):1240-1 250.

(29) Schoemaker MH, Gommans WM, de la Rosa LC, Homan M, Klok P, Trautwein C, et al. Resistance of

rat hepatocytes against bile acid-induced apoptosis in cholestatic liver injury is due to nuclear factor­

kappa B activation. J Hepatol 2003 Aug:39(2):153-161.

(30) Reinehr R, Graf D, Haussinger D. Bile salt-induced hepatocyte apoptosis involves epidermal growth

factor receptor-dependent CD95 tyrosine phosphorylation. Gastroenterology 2003 Sep:125(3):839-853.

(31) Reinehr R, Haussinger D. Inhibition of bile salt-induced apoptosis by cyclic AMP involves

serine/threonine phosphorylation of CD95. Gastroenterology 2004 Jan;126(1 ):249-262.

(32) Reinehr R, Becker S, Hongen A, Haussinger D. The Src family kinase Yes triggers hyperosmotic

activation of the epidermal growth factor receptor and CD95. J Biol Chem 2004 Jun 4;279(23):23977-

23987.

(33) Reinehr R, Becker S, Wettstein M, Haussinger D. Involvement of the Src family kinase yes in bile salt­

induced apoptosis. Gastroenterology 2004 Nov;127(5): 1540-1557.

49

Chapter 4

Reactive oxygen species are not essential for bile acid-induced cell death

Titia E. Vrenken, Manon Buist-Homan , Klaas Nico Faber, Han Moshage

Department of Gastroenterology and Hepatology; University Medical Center Groningen, University of Groningen, Groningen, the Netherlands

Submitted

Chapter 4

4.1 Abstract

Background: Bile acids, reactive oxygen species (ROS) and inflammatory cytokines are crucial regulators of cell death in liver diseases. It has been hypothesized that bile acid­induced generation of ROS contributes to bile acid-induced hepatocyte death. Aim: to investigate the role of ROS in bile acid-induced cell death. Methods: Primary rat hepatocytes were exposed to the bile acids glycochenodeoxycholic acid (GCDCA) or taurolithocholic acid-3 sulfate (TLCS) in the absence or presence of ROS scavengers or antioxidants. Heme oxygenase (HO)-1 mRNA levels were analyzed by qPCR. Total glutathione levels were determined, and apoptosis and necrosis were quantified by acridine orange and Sytox green staining, respectively. Results: Bile acid-induced cell death is not reduced by ROS-scavengers or antioxidants. Bile acids do not induce the oxidative stress responsive gene HO-1 and do not deplete glutathione levels. The Src-kinase inhibitor, SU6656, reduces GCDCA-induced apoptosis and necrosis. Inhibition of NADPH oxidase reduces bile acid-induced apoptosis, but increases necrosis. Conclusion: Bile acid induced toxicity is not mediated via generation of ROS and bile acids do not modulate indicators of oxidative stress. Src-kinases play an important role in bile acid­induced cell death. NADPH oxidase inhibition does not reduce cell death but induces a shift from apoptosis to necrosis.

52

Reactive oxygen species are not essential for bile acid-induced cell death

4.2 Introduction

A common feature of both acute and chronic liver diseases is the excessive death of hepatocytes, leading to loss of liver function and eventually liver failure. Several mediators of hepatocyte death have been proposed, including pro-inflammatory and pro-apoptotic cytokines like tumor necrosis factor (TNF) a and Fas ligand (Fasl}, reactive oxygen species (ROS), bile acids and toxic metabolites (1 -3). The exact contribution of each of these mediators to hepatocyte cell death in the various liver diseases remains to be elucidated. Cell death often occurs as a mixture of apoptosis and necrosis. Apoptosis or programmed cell death is a tightly regulated process, involving the caspase-family of cysteine proteases. The energy-dependent apoptotic pathway results in cell shrinkage and formation of apoptotic bodies, which are cleared very efficiently by surrounding cells (4-7). Necrosis, a passive mode of cell death, leads to plasma membrane rupture, and subsequent release of cellular content in the extracellular environment and circulation and an increased inflammatory response (7, 8).

Reactive oxygen species comprise a variety of free radicals and peroxides. Under normal conditions, ROS are formed during oxygen-dependent metabolism. Upon cellular stress or during the metabolism of xenobiotics, e.g. alcohol, the endogenous generation of ROS can be significantly enhanced (9- 1 1 ). In addition, in inflammatory conditions there is exposure to exogenous ROS, generated by inflammatory cells. Cells contain many other pro-oxidant systems, such as nicotinamide adenine dinucleotide phosphate (NADPH) oxidases ( 1 1 -1 3) and xanthine oxidase. NADPH-oxidase generates superoxide anions while oxidizing NADPH (1 1 , 1 4). Hepatocytes contain a variety of defence mechanisms against oxidative stress. The enzyme superoxide dismutase (SOD) catalyzes the reduction of superoxide anions to hydrogen peroxide (H202), which is subsequently converted into water and oxygen by either catalase or glutathione peroxidases (1 2, 1 3, 1 5, 1 6). Glutathione peroxidase oxidizes the antioxidant glutathione, while reducing hydrogen peroxide into water. In contrast to its conversion into water, hydrogen peroxide can also react with ferrous ions, resulting in the production of highly reactive hydroxyl radicals ( 13). In addition to enzymatic antioxidants, cells contain many other antioxidants such as vitamins, glutathione and bilirubin ( 12, 1 3, 1 7).

ROS are generated in both acute and chronic liver injury, and increased ROS generation has been implicated in hepatocyte cell death. Indeed, antioxidant therapy has been shown to be effective in liver injury known to be associated with oxidative stress, e.g. alcoholic liver injury (1 8). It has been hypothesized that cholestasis and exposure to bile acids also cause generation of ROS and oxidative stress, resulting in hepatocyte death and liver injury. The bile acid glycochenodeoxycholic acid (GCDCA) at concentrations of 50-100 µmol/L was shown to generate ROS in isolated mitochondria and hepatocytes (1 9-23). In contrast, only one paper described that GCDCA (50 µmol/L) did not increase ROS generation, while higher concentrations of the bile acid (200 and 400 µmol/L) clearly induced ROS in primary rat hepatocytes (9). Additional studies have described the generation of ROS after bile acid exposure: tauro-lithocholic acid-3 sulfate (TLCS), taurochenodeoxycholate (TCDC; 1 9, 24-26), taurodeoxycholic acid (TOCA; 27) and deoxycholic acid (DCA; 27, 28) promoted oxidative stress in either freshly isolated (25, 27, 28) or 24-hour cultured ( 19, 24, 26) rat hepatocytes. The ROS production could be reduced by a variety of antioxidants (1 9, 25-27), which also decreased bile acid-induced caspase activation (26) or cell death (1 9, 25). The anti-apoptotic bile acid ursodeoxycholic acid (UDCA) lowered bile acid-induced ROS generation and apoptosis in primary rat hepatocytes (22, 28). Likewise, GCDCA-induced ROS generation was reduced by antioxidants (1 9-21 ) and antioxidants lowered the levels of apoptosis (21 ). However, some studies described no effect of antioxidants on lipid peroxidation (29), or even a complete reduction in lipid peroxidation by antioxidants without a reduction in liver damage (25, 30). Recently, the importance of members of the Src tyrosine kinase family and NADPH oxidases in oxidative stress, Fas and bile acid-induced cytotoxicity was reported ( 14, 26, 31 -35). The Src kinase family consists of 9 members, all of them containing the Src-specific homology domains SH2 and SH3 and a tyrosine kinase

53

Chapter 4

domain (36). Oxidative stress results in activation of the Src family members Fyn, Yes and Lek, while hydrophobic bile acids only activate Yes in cultured hepatocytes (31, 32). The Src inhibitor SU6656, which specifically inhibits the Src kinases Fyn , Yes, Lek and Src (37), blocked bile acid-induced Yes activation (32). In addition, antioxidants that reduced bile acid­induced ROS generation, also inhibited Yes phosphorylation (32). Membrane-associated NADPH oxidase isoforms have been postulated to act up-stream of Yes activation: bile acids activate NADPH oxidases, resulting in ROS generation, Yes activation and ultimately apoptosis (26).

In this study, the involvement of ROS in bile acid-induced cell death was examined using 2 different bile acids. We demonstrate that neither hydrogen peroxide nor superoxide anions are involved in GCDCA- or TLCS-induced cell death, while the Src-family of kinases is implied in bile acid-induced apoptotic and necrotic signaling. In addition, our results suggest that inhibition of NAPDH oxidase shifts GCDCA-induced cell death from apoptosis to necrosis.

4.3 Materials and Methods

Animals Specified pathogen-free male Wistar rats (200-250 g) were purchased from Charles River Laboratories Inc (Wilmington, MA, USA). Rats were housed under standard laboratory conditions with free access to standard laboratory chow and water. Prior to isolation, rats were fasted overnight. Experiments were performed following the guidelines of the local Committee for Care and Use of laboratory animals.

Rat hepatocyte isolation Hepatocytes were isolated from rats by a two step collagenase perfusion as described before (10). Cell viability was greater than 85% as determined by trypan blue exclusion. After isolation, 112,500 cells per cm2 were plated on Vitrogen® (Cohesion Technologies Inc., Palo Alto, CA, USA) coated plates in William's E medium (lnvitrogen , Breda, The Netherlands) supplemented with 50 µg/mL gentamycin ( lnvitrogen) and penicillin-streptomycin-fungizone (Lonza, Verviers, Belgium). During the attachment period (4 hrs) 50 nmol/L dexamethasone (Dept. of Pharmacy, UMCG, Groningen, The Netherlands) and 5% fetal calf serum (lnvitrogen) were added to the medium. Cells were cultured in a humidified incubator at 37°C and 5% CO2.

Experimental design Experiments were started after the attachment period of 4 hrs. Monolayer cultures were exposed to 50 µmol/L GCDCA (Calbiochem, La Jolla, CA, USA) for 4 hrs or to 100 µmol/L TLCS (Sigma-Aldrich , St. Louis, MO, USA) for 6 hrs. In some experiments, hepatocytes were treated with the superoxide anion donor, menadione (2-methyl-1,4-naphthoquinone, 50 µmol/L, Sigma-Aldrich) or a cytokine mixture (CM) consisting of 20 ng/ml recombinant murine TNFa (R&D Systems, Abingdon, United Kingdom), 10 ng/ml recombinant human interleukin-113 (hlL-113, R&D Systems), 10 ng/ml recombinant rat interferon-y (rlFN y, R&D Systems) and 10 µg/ml lipopolysaccharide (LPS, Escherichia coli, serotype 0127:88, Sigma­Aldrich). The involvement of reactive oxygen species was studied using 100 U/ml superoxide dismutase-polyethylene glycol from bovine erythrocytes (PEG-SOD, Sigma-Aldrich), 500 U/ml catalase-polyethylene glycol (PEG-CAT, Sigma-Aldrich), 5 mmol/L of the antioxidant and glutathione precursor N-acetyl-L-cysteine (NAC, Sigma-Aldrich) or 5 mmol/L glutathione monoethyl ester (GSH-MEE, Merck Chemicals Ltd, Nottingham, UK), the cell permeable derivative of glutathione. Diphenylene iodonium (DPI, Biomol Research Laboratories Inc, Plymouth Meeting, PA, USA) was used at 2.5 µmol/L to block NADPH oxidase and nitric oxide synthase (NOS, ref 306). 10 µmol/L SU6656 (Merck) was used to inhibit the Src kinase family members Fyn, Yes, Lek and Src (37), and 1 µmol/L herbimycin A Streptomyces Sp. (Merck) was used to block a variety of protein tyrosine kinases including Src, heat shock proteins, and nuclear factor-KB (NF-KB) activation (38). All inhibitors were 54

Reactive oxygen species are not essential for bile acid-induced cell death

added 30 minutes prior to bile acids. Each experimental condition was performed in triplicate. Cells were harvested at the indicated time points after the addition of the apoptotic stimuli, and washed three times with ice cold phosphate-buffered saline (PBS) before the addition of hypotonic cell lysis buffer (protein analysis, caspase-3 assay), sample buffer 2x (Western blot analysis) or Tri-reagent (RNA isolation, Sigma-Aldrich) as described previously (39). Each experiment was performed at least three times, using hepatocytes from different isolations.

Caspase enzyme activity assay Caspase-3 like activity was assayed as described previously (39). The arbitrary fluorescence unit (AFU) was corrected for the amount of protein in the cell lysate.

Sytox green and acridine orange nuclear staining To determine necrotic cell death, hepatocytes were incubated with Sytox green (lnvitrogen, 1 :40000) nucleic acid stain. Sytox can only enter cells with compromised plasma membranes, and cannot cross the membranes of viable cells or apoptotic bodies. Apoptotic nuclei were visualized with acridine orange (Sigma-Aldrich, 1 :5000) as described previously (10). Fluorescent nuclei were visualized using an Olympus CKX41 microscope at 450-490 nm. Necrosis and apoptosis were quantified by counting fluorescent nuclei (necrotic or apoptotic cells) and the total number of cells in 3 randomly chosen high power fields.

Western Blot analysis Western blot analysis of cell lysates was performed by SOS-PAGE followed by semi dry­blotting to transfer the proteins to Hybond ECL nitrocellulose membrane (Amersham Biosciences, Piscataway, NJ, USA). Ponceau S 0.1 % w/v (Sigma-Aldrich) staining was used to confirm electrophoretic transfer. Protein expression of heme-oxygenase-1 (HO-1) was detected using a monoclonal antibody against HO-1 (Stressgen, Sanbio BV, Uden, The Netherlands) at a dilution of 1 :500. After Western blot analysis, blots were stripped as described previously (10) and incubated with a monoclonal antibody against GAPDH (Calbiochem, dilution 1 :5000). The blots were analyzed in a ChemiDoc XRS system (Bio­Rad, Hercules, CA, USA). Protein band intensities were quantified by Quantity One software (Bio-Rad).

RNA isolation and quantitative polymerase chain reaction (PCR) RNA isolation, reverse transcription PCR and quantitative PCR (qPCR) were performed as described previously (40). Details of primers and probes are listed in Table 1. Each sample was analyzed in duplicate. 1 BS mRNA levels were used as endogenous control. Table 1. Sequences of primers and probes used for quantitative PCR analysis

185 Sense 5'-CGG CTA CCA CAT CCA AGG A-3'

rat Antisense 5'-CCA ATT ACA GGG CCT CGA AA-3'

Probe S'FAM-CGCGCAAA TT ACCCACTCCCGA-TAMRA3'

INOS Sense 5'-GTG CTA ATG CGG AAG GTC ATG-3'

rat Antisense 5'-CGA CTT TCC TGT CTC AGT AGC AAA-3'

Probe S'FAM-CCC GCG TCA GAG CCA CAG TCC T-TAMRA3'

H0-1 Sense 5'-CAC AGG GTG ACA GAA GAG GCT AA-3'

rat Anti sense 5'-CTG GTC TTT GTG TTC CTC TGT CAG-3'

Probe S'FAM-CAG CTC CTC AAA CAG CTC AAT GTT GAG C-TAMRA3'

Glutathione determination Total glutathione (reduced glutathione and glutathione disulfide, GSSG) was measured in cell lysates by the enzymatic recycling procedure as described by Griffith (41) using glutathione reductase from yeast (Roche Diagnostics Nederland BV, A/mere, The Netherlands), GSSG (Sigma-Aldrich), NADPH (Roche Diagnostics) and 5,5'-dithiobis-(2-

55

Chapter 4

nitrobenzoic acid) (DTNB, Roche Diagnostics). Samples were analyzed on a 30°C Uvikon spectrophotometer for 1 O minutes at 412 nm. Each sample was analyzed in duplicate. Values were normalized over the protein amount in the lysates.

Statistical analysis All numerical results are reported as the mean of at least 3 independent experiments ± standard error of the mean. For each experiment, the results were analysed using the Kruskal-Wallis test to verify the significance. If P<0.05 for the Kruskal-Wallis test, a Mann­Whitney U test was used to determine the significance of differences between experimental groups. A P-value smaller than 0.05 was considered to be statistically significant.

4.4 Results

Reactive oxygen species are not involved in bile acid induced cell death Caspase-3 activity peaked at 3-4 hours after addition of GCDCA as shown previously (42). GCDCA-induced caspase-3 activity was not affected by either the cell permeable enzyme superoxide dismutase (SOD) or catalase (Fig 1 A), nor did the cell permeable glutathione derivative glutathione monoethyl ester (GSH-MEE) change GCDCA-induced caspase-3 activity. Caspase-3 activity was slightly reduced by the antioxidant N-acetyl-L-cysteine, however this effect was not significant (Fig 1A). The antioxidants alone did not induce caspase-3 activity (data not shown). Reactive oxygen species have been implicated in TLCS-induced cell death ( 19, 24, 26). In our hands, all three compounds examined, superoxide dismutase, catalase and glutathione-monoethylester, slightly reduced TLCS­induced caspase-3 acitivity (Fig 1B). However, this decrease was not statistically significant for any of the antioxidants tested. A

80

,1

I l! �_ jj _U_ ..... + PEO-aOD • P£Q.CAT ·-= ..........

OCOCAOOIM

•••• ....... ,. GC1)CA,0,.. '"' GCDCA•nG-IOO GCOCA • PEG-CAT .. "' ,tf" Figure 1. Reactive oxygen species are not Involved in bile ac,d induced cell death. Antioxidants were added 30 minutes prior to bile acids. Primary rat hepalocyles were lfeated w�h (A) GCDCA (50 µmoVL, 4 hrs) or (B) TLCS (100 µmoVL, 6 hrs). Caspase-3 like acl.vIty was measured, and is

B

l ii

(

....... ,.

·-I

� • P£G-t00 + PEG-CAT ,,w;

TlC11 t•tH

QCIDCA60... GCDCA+KG.tOO GCDCA+KG-CAT ,... 111' 1n.

shown as fold induction compared to control, control values were set at one. Superoxide dismutase and catalase do not affect GCDCA-induced nuclear condensation (C) or tD) necrotic cell death (C) Apoptotic nuclei were determined by acndtne orange staining 4 hrs after GCOCA· exposure and expressed as 0o of total nuclei Magnification 10X. (0) Sytox green nuclear stairnng was used to monitor necrosis Upper panel combined phase contrast and fluorescence Lower panel fluorescence Magnification 10X Statistical analysis· • p<0 01 compared 10 control

The results from the caspase-3 activity assay (Fig 1 A) were confirmed by acridine orange staining, visualizing condensed, apoptotic nuclei . Hepatocytes exposed for 4 hrs to GCDCA demonstrate apoptotic nuclei in 52% of the cells (Fig 1 C). In combination with PEG-SOD or PEG-catalase the percentage of condensed nuclei remained 48% and 47% respectively. In control conditions, apoptotic cells were hardly present (Fig 1 C). Next, the effects of PEG­catalase and PEG-SOD on GCDCA-induced necrosis were analyzed. Necrosis was observed in 16% of the cells treated with GCDCA for 4 hrs (Fig 1 D) and this percentage was not significantly changed by either PEG-SOD or PEG-catalase: 16% and 12% necrotic cells, respectively. Control hepatocytes or hepatocytes exposed to PEG-SOD or PEG-catalase did not display necrosis (Fig 1 D, data not shown). 56

Reactive oxygen species are not essential for bile acid-induced cell death

To test the effectiveness of the antioxidants PEG-SOD and PEG-catalase, we determined the effect of PEG-SOD on menadione-induced apoptosis and PEG-catalase on hydrogen peroxide-induced necrosis. PEG-SOD completely abolished menadione-induced apoptosis (data not shown) as described previously (10) and PEG-catalase completely prevented hydrogen peroxide induced necrosis (data not shown).

Bile acid exposure does not exert oxidative stress on hepatocytes To determine whether GCDCA exerts oxidative stress on hepatocytes, we measured expression of the oxidative stress responsive gene HO-1 and total glutathione levels. As shown in figure 2A, exposure to GCDCA or TLCS did not change HO-1 mRNA expression compared to control hepatocytes. In contrast, exposure to the superoxide anion donor, menadione results in a 20-fold induction of HO-1 expression compared to control cells (Fig 2A). Western blot analysis confirmed that HO-1 is not induced by GCDCA (Fig 2B) and quantification of HO-1 protein levels, standardized on GAPDH intensities, revealed relative amounts of 1.00 and 0.91 for control and GCDCA-treated hepatocytes respectively (Fig 2B). The total gluthathione level of hepatocytes exposed to GCDCA was not significantly lower compared to control hepatocytes (Fig 2C).

A B """"" GCDCA

F---1 �· I I GAPOH

C

Figure 2. The oxidative stress responsive gene H0-1 is not induced by bile acids. qPCR analysis for H0-1 using cDNA of primary rat hepatocytes exposed for 4 hrs to 50 µmol/L GCDCA or 6 hrs to 100 µmol/L TLCS (AJ. The superoxide anion donor, menadione (SO µmol/L, 4 hrs) served as positive control for H0-1 induction (B) Western blot analysis of H0-1 prote n expression 1n cell lysates (45 µI) of primary rat hepatocyles treated without (control) or with GCDCA (SO µmol/L, 4 hrs). Western blots were quantified and band intensities of H0-1 protein were expressed as ratio of the intensity of GAPDH protein Data is presented as mean of at least three independent experiments +/- S E M. Control values were set at one. A representative Western blot is shown. (C) GCDCA does not affect total glutath,one 'eve s Total glutathione (mmol/mg protein) was measured and is shown as percentage of control; control values were set at 100%.

Bile acid induced apoptosis involves NADPH oxidase and the Src family of kinases, but inhibiting NAPDH oxidase shifts GCDCA-induced cell death from apoptosis into necrosis Subsequently, the involvement of the Src kinase and NADPH oxidase signaling pathways in bile acid-induced cell death were studied. Exposure of hepatocytes to the NADPH-oxidase inhibitor DPI prior to GCDCA, strongly reduced caspase-3 activity (Fig 3A). Similarly, treatment with the specific Src-kinase inhibitor SU6656 decreases GCDCA-induced caspase-3 activity by about 67% (Fig 3A). In contrast, the protein tyrosine kinase inhibitor herbimycin A only slightly lowered caspase-3 activity upon GCDCA exposure, but this effect was not significant (Fig 3A). The inhibitors alone did not induce caspase-3 activity (data not shown). The effects of DPI, SU6656 and herbimycin A were also investigated on TLCS­induced apoptosis. DPI almost completely inhibited TLCS-induced caspase-3 activity (Fig 3B), while SU6656 caused a small but non-significant reduction. Herbimycin A even further enhanced TLCS-induced caspase-3 activity, although not statistically significant (Fig 3B). Again, acridine orange staining was performed to verify the results obtained with the caspase-3 activity assay (Fig 3A). GCDCA exposure for 4 hrs resulted in nuclear condensation in 52% of the hepatocytes (Fig 3C). DPI and SU6656 reduced the percentage of condensed nuclei to 22% and 15% respectively, whereas herbimycin A (50%) had no effect on GCDCA-induced nuclear condensation. Apoptosis was barely present in control hepatocytes (Fig 3C) or cells treated with only the inhibitors (data not shown). Next, we determined the effects of these inhibitors on necrotic cell death. GCDCA-treatment induced necrosis in 16% of the hepatocytes after 4 hrs ( Fig 3D), which was further increased after overnight exposure to 37% (Fig 3E). At 4 hrs, GCDCA-induced necrosis was not affected by

57

Chapter 4

DPI (18% necrotic hepatocytes, Fig 3D). In contrast, overnight exposure of hepatocytes to DPI and GCDCA significantly induced necrotic cell death to 83% (Fig 3E). SU6656 reduced GCDCA-induced necrosis to less than 10% at both time points (Fig 3D, data not shown). Exposure to herbimycin A slightly but not significantly increased GCDCA-induced necrosis al 4 hrs (Fig 3D), and percentages were similar to GCDCA after overnight exposure (data not shown). A

D

E

90

control 3%

GCOCA SOµM 371jii

,op,

GCOCA 50pM 52%

• GCDCA + DPI •

83'11

B IO

+ Sl.111658 +hart.nycinA con..,

CCDCA 50 !1.t

GCOCA + OPI ' 22'11

GCDCA • 01'1 18%

• GCOCA + SUIS058 • GCOCA + herblmycln A

15% 50%

GCDCA + SUH51 10%

GCDCA + h.rblmycln A 3'%

" #

+OP! + SU6S56 TLCS 100 �

+ h•rb1myc1n A

Figure 3. 81 e ac,d�,nduced apoptosis involves the Src-kinase family and inhibition of NADPH ox1dase shifts GCDCA-1nduced cell death from apoptosis nto necrosis. lnh biters were added prior to (A) GCDCA or (B) TLCS Caspase-3 like activity was measured, and is shown as fold mduclion compared to contra . control values were set at one. (C) Nuclear condensation upon GCDCA exposure was also reduced in hepatocyles pre-treated with DPI or SU6656. Apoptotic nuclei were delerm·ned by acridine orange staining 4 hrs after GCDCA-exposure and expressed as % of tolal nuclei Magnification 10X (D, E) The decrease n GCDCA-induced apoptosis by DPI is accompanied by an increase in necrosis Sytox green nuclear staining was used to visuahze necrotic cell death after 4 hrs (D) and after overnight exposure (E) to GCDCA. Upper panel: combined phase contrast and fluorescence Lower panel fluorescence. Magnification 20X. Statist cal analysis. • p<0.05 or • p<0.01 compared to contra # p<0.05 or p<0.01 compared to GCDCA or TLCS stimu•ated cells

We observed that DPI protected against apoptosis, but strongly increased necrotic cell death. Since DPI is also known to inhibit NOS, in addition to NADPH oxidase, we examined the possible role of inducible NOS, iNOS, in bile acid induced cell death. The relative mRNA

f 200

i I ,00

� t --- ---, ...... GCDCA

58

' #

TLCS cy1olllne mlwre

expression level of iNOS was not changed in hepatocytes exposed to GCDCA or TLCS, whereas a 200-fold induction of iNOS expression was observed upon cytokine mixture-treatment (Fig 4).

Figure 4. Bile acids do not affect NOS gene expression qPCR ana ysis for iNOS using cDNA of primary rat hepatocytes exposed for 4 hrs to GCDCA (50 µmol/L) or 6 hrs to TLCS (100 µmoVL). The cytokone mixture-·nduced iNOS mRNA expression served as positive control. Statistical analysis • p<0.05 compared to control. # p<0 05 compared to GCDCA treated cells

Reactive oxygen species are not essential for bile acid-induced cell death

4.5 Discussion

We investigated the involvement of reactive oxygen species in bile acid-induced injury in primary rat hepatocytes since previous studies described contradictory findings (9, 19, 21, 23-27, 43). In our studies we show that ROS are not essential for bile acid induced cell death, since various scavengers of ROS and antioxidants do not reduce GCDCA or TLCS induced apoptosis or necrosis. In addition, parameters of exposure to oxidative stress, such as heme oxygenase-1 induction and glutathione depletion, did not change in bile acid exposed hepatocytes. These results are in accordance with a previous study (9) showing that GCDCA, used at the same concentration as in our experiments, does not induce ROS generation. GCDCA did generate ROS at higher concentrations (� 200 µmol/L), but these concentrations result predominantly in necrotic rather than apoptotic cell death. This study (9) and our results are in contrast to many other reports describing a rapid and sustained ROS production in response to GCDCA exposure (19-23). These conflicting data cannot be explained by the use of different cell types. The induction of ROS has been shown in both freshly isolated rat hepatocytes (21, 23), cultured primary rat hepatocytes (19) and isolated mitochondria from human (22) and rat (20). Although a few studies (20, 23) used 2-fold higher concentrations of GCDCA as used in our study, ROS generation induced by 50 µmol/L of bile acid has been described (19, 2 1). An interesting previous observation in this regard is that inhibition of caspase-8 completely abrogated GCDCA-induced apoptosis, but only partially reduced ROS generation (21 ), indicating no or only a weak correlation between ROS generation and apoptosis.

A possible explanation for these contradictory results is that bile acids at concentrations below 100 µmol/L generate low levels of ROS, as has been shown in many studies, that are insufficient to induce oxidative stress accompanied by HO-1 induction and glutathione depletion and are not causally involved in apoptotic cell death. At these concentrations, bile acids activate apoptotic signaling pathways involving Src kinase family members like Yes. At very high concentrations of bile acids (above 200 µmol/L), high levels of ROS may be generated, accompanied by oxidative stress and necrotic rather than apoptotic cell death. Our findings are supported by patient studies (29, 30) in which antioxidant treatment did not reduce lipid peroxidation or reduce liver injury.

The Src-kinase signal transduction pathway has been implicated in bile acid-induced cell death and it has been hypothesized that bile acid-induced generation of ROS activate Src­kinase family members. The inhibitor SU6656, which predominantly inhibits the Yes family member, significantly reduced GCDCA-induced apoptosis, whereas the protein tyrosine kinase inhibitor herbimycin A had no significant inhibitory effect on GCDCA-induced apoptosis. In contrast, both compounds had no significant effect on TLCS-induced apoptosis. This distinction is in line with previous papers describing the involvement of different signaling pathways in GCDCA- and TLCS-induced cell death. An inhibitory effect of SU6656 on TLCS-induced apoptosis has been described (32). In our study SU6656 tended to decrease TLCS-induced apoptosis, but this reduction did not reach significance.

It has been hypothesized that bile acids activate the membrane-associated enzyme NADPH­oxidase. Activated NADPH-oxidase generates ROS and this ROS, via Yes activation, induces apoptosis (26). In our study, inhibition of NADPH oxidase using DPI significantly reduced GCDCA and TLCS-induced apoptosis, as described before (26). However, although necrotic cell death was not affected by DPI 4 hours after exposure to GCDCA, it was significantly increased after overnight treatment of hepatocytes to GCDCA in combination with DPI, demonstrating a shift in cell death from apoptosis towards necrosis. A shift from apoptotic to necrotic cell death has been observed before using a model system involving superoxide anions and the nitric oxide donor SNAP (10). The DPl-induced increase in necrotic cell death has not been noted before and leads to the conclusion that inhibition of NADPH oxidase does not protect against bile acid induced apoptosis. Together with our

59

Chapter 4

observations on the effect of ROS scavengers and parameters of oxidative stress, we can conclude that ROS generation is not responsible for bile acid induced apoptotic cell death.

4.6 Acknowledgements

This study was supported by the Dutch Digestive Foundation (WS 02-13) and the J .K. de Cock Foundation.

4.7 Reference List

( 1 )

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(1 0)

(1 1 )

(12)

( 1 3)

( 14)

(15)

( 1 6)

(1 7)

( 18)

( 19)

(20)

60

Yoon JH, Gores GJ. Death receptor-mediated apoptosis and the liver J Hepatol 2002 Sep;37(3):400-

410.

Adams JM. Ways of dying: multiple pathways to apoptosis. Genes Dev 2003 Oct 1 5;1 7(20):2481 -

2495.

Higuchi H, Gores GJ. Bile acid regulation of hepatic physiology: IV. Bile acids and death receptors.

Am J Physiol Gastrointest Liver Physiol 2003 May;284(5):G734-G738.

Reed JC. Mechanisms of apoptosis. Am J Pathol 2000 Nov;l 57(5): 1 4 1 5-1 430.

Adams JM, Cory S. Apoptosomes: engines for caspase activation. Curr Opin Cell Biol 2002

Dec;l 4(6):71 5-720.

Cohen GM. Caspases: the executioners of apoptosls. Biochem J 1 997 Aug 1 5;326(Pt 1 ) : 1 -16 .

Kim JS, He L, Lemasters JJ. Mitochondrial permeability transition: a common pathway to necrosis and

apoptosls. Biochem Biophys Res Commun 2003 May 9;304(3):463-470.

Lemasters JJ. Mechanisms of hepatic toxicity - V. Necrapoptos1s and the mitochondrial permeabil ity

transition: shared pathways to necrosis and apoptosis. American Journal of Physiology­

Gastrointestinal and Liver Physiology 1 999 Jan;276(1 ):G1 -G6.

Chu SH, Lee-Kang J, Lee KH, Lee K Roles of reactive oxygen species, NF-kappaB, and

peroxiredoxins in glycochenodeoxycholic acid-induced rat hepatocytes death. Pharmacology 2003

Sep;69(1 ):12-1 9.

Conde de la Rosa L, Schoemaker MH, Vrenken TE, Buist-Homan M, Havinga R, Jansen PL, et a l.

Superoxide anions and hydrogen peroxide induce hepatocyte death by different mechanisms:

involvement of JNK and ERK MAP kinases. J Hepatol 2006 May;44(5):91 8-929.

Li JM. Shah AM. ROS generation by nonphagocytic NADPH oxidase: potential relevance In diabetic

nephropathy. J Am Soc Nephrol 2003 Aug;14(8 Suppl 3):S221 -S226.

Czaja MJ. Induction and regulation of hepatocyte apoptosis by oxidative stress. Antioxidants & Redox

Signaling 2002 Oct;4(5):759-767.

Jaeschke H, Gores GJ, Cederbaum Al, Hinson JA, Pessayre D, Lemasters JJ Forum - Mechanisms

of hepatotoxicity. Toxicological Sciences 2002 Feb;65(2):166-176.

Reinehr R, Becker S, Braun J, Eberle A, Grether-Beck S, Haussinger D. Endosomal acidification and

activation of NADPH oxidase isoforms are upstream events in hyperosmolarity-induced hepatocyte

apoptosis. Journal of B iological Chemistry 2006 Aug 1 1 ;281 (32)·231 50-23166.

Park SY, Chang I, Kim JY, Kang SW, Park SH, Singh K, et al. Resistance of mitochondrial DNA­

depleted cells against cell death: role of mitochondrial superoxide dismutase. J Biol Chem 2004 Feb

27;279(9):751 2-7520.

Giannoni E, Buricchi F, Raugei G, Ramponi G, Chiarugi P. Intracellular reactive oxygen species

activate Src tyrosine kinase during cell adhesion and anchorage-dependent cell growth. Molecular

and Cellular Biology 2005 Aug;25(1 5):6391 -6403.

Granato A, Gores G, Vilei MT, Tolando R, Ferraresso C, Muraca M. Bilirubin inhibits bile acid induced

apoptosis in rat hepatocytes. Gut 2003 Dec 1 ;52(1 2) :1 774-1778.

Arteel GE. Oxidants and antioxidants in alcohol-induced liver disease. Gastroenterology 2003

Mar: 1 24(3):778-790.

Remehr R, Graf D, Haussinger D. Bile salt-induced hepatocyte apoptosis involves epidermal growth

factor receptor-dependent CD95 tyrosine phosphorylation. Gastroenterology 2003 Sep;1 25(3):839-

853

Sokol RJ, Straka MS, Dahl R, Devereaux MW, Yerushalmi B, Gumpricht E, et al. Role of oxidant

stress in the permeability transition induced in rat hepatic mitochondria by hydrophobic bile acids.

Pediatr Res 2001 Apr;49(4 ):51 9-531 .

Reactive oxygen species are not essential for bile acid-induced cell death

(21 ) Yerushalmi B, Dahl R, Devereaux MW, Gumpricht E, Sokol RJ. Bile acid-induced rat hepatocyte

apoptosis is inhibited by antioxidants and blockers of the mitochondrial permeability transition.

Hepatology 2001 Mar;33(3):61 6-626.

(22) Sokol TJ, Dahl R, Devereaux MW, Yerushalmi B, Kobak GE, Gumpricht E. Human hepatic

mitochondria generate reactive oxygen species and undergo the permeability transition in response to

hydrophobic bile acids. Journal of Pediatric Gastroenterology and Nutrition 2005 Aug;41 (2):235-243.

(23) Gumpricht E, Dahl R, Devereaux MW, Sokol RJ. beta-carotene prevents bile acid-induced cytotoxicity

in the rat hepatocyte: Evidence for an antioxidant and anti-apoptotic role of beta-carotene in vitro.

Pediatric Research 2004 May;55(5):814-821.

(24) Reinehr R, Haussinger D. Inhibition of bile salt-induced apoptosis by cyclic AMP involves

serine/threonine phosphorylation of CD95. Gastroenterology 2004 Jan;1 26(1 ):249-262.

(25) Borgognone M, Perez LM, Basiglio CL, Ochoa JE, Roma MG. Signaling modulation of bile salt­

induced necrosis in isolated rat hepatocytes. Toxicological Sciences 2005 Jan;83(1 ):114-125.

(26) Reinehr R, Becker S, Keitel V, Eberle A, Grether-Beck S, Haussinger D. Bile salt-induced apoptosis

involves NADPH oxidase isoform activation. Gastroenterology 2005 Dec;129(6):2009-2031 .

(27) Fang Y, Han SI , Mitchell C, Gupta S, Studer E , Grant S, e t al. Bile acids induce mitochondrial ROS,

which promote activation of receptor tyrosine kinases and signaling pathways in rat hepatocytes.

Hepatology 2004 Oct;40(4):961-971.

(28) Rodrigues CM, Fan G, Ma X, Kren BT, Steer CJ. A novel role for ursodeoxycholic acid in inhibiting

apoptosis by modulating mitochondrial membrane perturbation. J Clin Invest 1 998 Jun

15;101 (12):2790-2799.

(29) Socha P, Koletzko B, Pawlowska J, Proszynska K, Socha J. Treatment of cholestatic children with

water-soluble vitamin E (alpha-tocopheryl polyethylene glycol succinate): effects on serum vitamin E,

lipid peroxides, and polyunsaturated fatty acids. J Pediatr Gastroenterol Nutr 1997 Feb;24(2):189-

193.

(30) Watson JP, Jones DE, James OF, Cann PA, Bramble MG. Case report: oral antioxidant therapy for

the treatment of primary biliary cirrhosis: a pilot study. J Gastroenterol Hepatol 1 999 Oct;14(10):1034-

1 040.

(31) Reinehr R, Becker S, Hengen A, Haussinger D. The Src family kinase Yes triggers hyperosmotic

activation of the epidermal growth factor receptor and CD95. J Biol Chem 2004 Jun 4;279(23):23977-

23987.

(32) Reinehr R, Becker S, Wettstein M, Haussinger D. Involvement of the Src family kinase yes in bile salt­

induced apoptosis. Gastroenterology 2004 Nov;1 27(5):1540-1 557.

(33) Reinehr R, Becker S, Eberle A, Grether-Beck S, Haussinger D. Involvement of NADPH oxidase

isoforms and Src family kinases in CD95-dependent hepatocyte apoptosis. Journal of Biological

Chemistry 2005 Jul 22;280(29):27179-27194.

(34) Yoshizumi M, Abe J, Haendeler J , Huang Q, Berk BC. Src and Gas mediate JNK activation but not

ERK1/2 and p38 kinases by reactive oxygen species. J Biol Chem 2000 Apr 21;275(16):11706-

1 1 712.

(35) Abe J , Takahashi M, Ishida M, Lee JD, Berk BC. c-Src is required for oxidative stress-mediated

activation of big mitogen-activated protein kinase 1. J Biol Chem 1 997 Aug 15;272(33):20389-20394.

(36) Pandey A, Peri S, Thacker C, Whipple CA, Collins JJ, Mann M. Computational and experimental

analysis reveals a novel Src family kinase in the C. elegans genome. Bioinformatics 2003 Jan

22;19(2):169-172.

(37) Blake RA, Broome MA, Liu XD, Wu JM, Gishizky M, Sun L, et al. SU6656, a selective Src family

kinase inhibitor, used to probe growth factor signaling. Molecular and Cellular Biology 2000

Dec;20(23):9018-9027.

(38) Fukazawa H, Li PM, Yamamoto C, Murakami Y, Mizuno S, Uehara Y. Specific inhibition of

cytoplasmic protein tyrosine kinases by herbimycin A in vitro. Biochem Pharmacol 1991 Oct

9;42(9):1661-1671.

(39) Schoemaker MH, Conde de la Rosa L, Buist-Homan M, Vrenken TE, Havinga R, Poelstra K, et al.

Tauroursodeoxycholic acid protects rat hepatocytes from bile acid-induced apoptosis via activation of

survival pathways. Hepatology 2004 Jun;39(6):1563-1573.

(40) Blokzijl H, Vander Borghi S, Bok LI, Libbrecht L, Geuken M, van den Heuvel FA, et al. Decreased P­

glycoprotein (P-gp/MDR1) expression in inflamed human intestinal epithelium is independent of PXR

protein levels. lnflamm Bowel Dis 2007 Jun;13(6):710-720.

61

Chapter 4

62

(41 ) Griffith OW. Determination of glutathione and glutathione disulfide using glutathione reductase and 2-

vinylpyridine. Anal Biochem 1 980 Jul 1 5; 106(1 ):207-212.

(42) Schoemaker MH, Gommans WM, Conde de la Rosa L, Homan M, Kick P, Trautwein C, et al

Resistance of rat hepatocytes against bile acid-induced apoptosis in cholestatic liver injury is due to

nuclear factor-kappa B activation. J Hepatol 2003 Aug;39(2): 1 53-1 61 .

(43) Graf D, Kurz AK, Reinehr R, Fischer R, Kircheis G, Haussinger D. Prevention of bile acid-induced

apoptosis by betaine in rat l iver. Hepatology 2002 Oct;36(4 Pt 1 ):829-839.

(44) Higuchi H, Adachi M, Miura S, Gores GJ, Ishii H. The mitochondrial permeability transition contributes

to acute ethanol-induced apoptosIs in rat hepatocytes. Hepatology 2001 Aug;34(2):320-328.

Chapter 5

Carbon monoxide blocks oxidative stress-induced hepatocyte apoptosis via inhibition of the p54 J N K isoform

Laura Conde de la Rosa 1 , Titia E. Vrenken 1 , Rebekka A. Hannivoort1 , Manon Buist­Homan 1 , Rick Havinga 1 , Dirk-Jan Slebos2

, Henk F . Kauffman2, Klaas N ico Faber 1 ,

Peter L .M. Jansen3, Han Moshage1

1Center for Liver, Digestive and Metabolic Diseases, and 2Dept. of Pulmonary Diseases and Lung Transplantation, Universi� Medical Center Groningen, University of Gron ingen, Groningen, the Netherlands. Department of Gastroenterology and Hepatology, Academic Medical Center Amsterdam, The Netherlands

Free Radie Biol Med 2008;44(7): 1 323-33

Chapter 5

5.1 Abstract

Background: Most chronic liver diseases are accompanied by oxidative stress, which may induce apoptosis in hepatocytes and liver injury. Oxidative stress induces heme oxygenase-1 (HO-1) expression. This stress-responsive cytoprotective protein is responsible for heme degradation into carbon monoxide (CO), free iron and biliverdin. CO is an important intracellular messenger; however, the exact mechanisms responsible for its cytoprotective effect are not elucidated yet. Aim: To investigate whether HO-1 and CO protect primary hepatocytes against oxidative stress-induced apoptosis. Methods: In vivo, bile duct ligation was used as model of chronic liver disease. In vitro, primary hepatocytes were exposed to the superoxide anion-donor menadione in a normal and a CO containing atmosphere. Apoptosis was determined by measuring caspase-9,-6,-3 activity and PARP cleavage, and necrosis by Sytox Green staining. Results: HO-1 is induced in chronic cholestatic liver diseases. Superoxide anions induce HO-1 activity time and dose-dependently. HO-1 overexpression inhibits superoxide anion­induced apoptosis. CO blocks superoxide anion-induced JNK-phosphorylation, caspase-9, -6, -3 activation and abolishes apoptosis, but does not increase necrosis. Conclusion: HO-1 and CO protect primary hepatocytes against superoxide anion-induced apoptosis partially via inhibition of JNK-activity. CO could represent an important candidate for treatment of liver diseases.

64

Carbon monoxide blocks oxidative stress-induced apoptosis via inhibition of the p54 JNK isoform

5.2 Introduction

Chronic liver diseases, such as non-alcoholic steatohepatitis (NASH), chronic cholestasis and alcoholic and chronic viral hepatitis, are almost invariably accompanied by exposure to reactive oxygen species (ROS; 1-3). ROS can contribute to hepatocyte cell death and liver injury by either apoptosis or necrosis.

Caspases (4, 5) represent the central executioners of apoptosis (6, 7). Apoptosis is an active process characterized by cell shrinkage, chromatin condensation and formation of apoptotic bodies. In contrast, necrosis is passive and associated with ATP depletion, rupture of the plasma membrane and spilling of the cellular content eliciting inflammation (6).

Heme oxygenase-1 (HO-1) catalyzes the oxidation of heme to form equimolar amounts of ferrous iron, biliverdin and carbon monoxide (CO). Biliverdin is rapidly converted into bilirubin by NAD(P)H:biliverdin reductase. HO-1 is induced by a variety of stimuli, most of them associated with oxidative stress (8). Many studies have suggested that HO-1 acts as an inducible defense against oxidative stress, e.g. in models of inflammation, ischemia/reperfusion, hypoxia and hyperoxia-mediated injury (9). In the liver, HO-1 induction protected against ischemia/reperfusion injury (10, 11) and endotoxemia (12, 13). In addition, overexpression of HO-1 by gene transfer has been shown to protect against hyperoxia induced lung injury (14), immune-mediated apoptotic liver damage in mice (15) and against CYP2E1-dependent toxicity in HepG2 cells (16). However, the mechanisms by which HO-1 mediates cytoprotection are not elucidated yet. Protective effects of both biliverdin and CO have been reported and several papers suggest that biliverdin protects against oxidative stress by acting as an antioxidant in different models of liver injury (17, 18). Less is known about the biological actions of CO. Although CO is known to be toxic and lethal at high doses, interest has recently emerged for the role of CO as a signaling and regulatory molecule in cellular processes (19, 20). A cytoprotective effect of CO has been demonstrated in several conditions associated with apoptosis.

In the present study, our aim was to investigate the regulation of HO-1 in vivo (model of acute inflammation and chronic cholestasis) and in vitro in primary cultures of rat hepatocytes. In addition, we investigate whether CO contributes to the protective effect of HO-1 against oxidative stress-induced apoptosis in primary hepatocytes, and the mechanisms involved in the process.

5.3 Materials and Methods

Animals Specified pathogen-free male Wistar rats (220-250 g) were purchased from Harlan (Zeist, The Netherlands). They were housed under standard laboratory conditions with free access to standard laboratory chow and water. Experiments were performed following the guidelines of the local Committee for Care and Use of laboratory animals.

Animal models Bile Duct ligation Male Wistar rats were anaesthetized and subjected to bile duct ligation (BDL) as a model of chronic cholestasis (21). At the indicated time points after BDL, rats (n=4 per group) were sacrificed, livers were perfused with saline and removed. Control rats (n=4) received a sham operation (SHAM) or not (control). Model of acute inflammation Rats were injected intraperitoneally (IP) with 5 mg/kg body weight lipopolysaccharide (LPS serotype 0127:88; Sigma-Aldrich, St. Louis, MO) or the same volume of phosphate-buffered saline (PBS, control group, n=4 for each experimental group), with or without diethylmaleate (DEM: 4 mmol/kg bodyweight, Sigma-Aldrich) 30 minutes prior to LPS or PBS injection. LPS

65

Chapter s

and DEM were dissolved in sterile PBS. Rats were sacrificed 6 hrs after LPS injection. Livers were perfused with saline and snap-frozen in liquid nitrogen until further use.

Rat hepatocyte isolation Hepatocytes were isolated as described previously (22) and cultured in William's E medium (lnvitrogen; Breda, The Netherlands) supplemented with 50 µg/ml gentamycin (lnvitrogen) without the addition of hormones or growth factors. During the attachment period (4 hrs), 50 nmol/L dexamethasone (Sigma-Aldrich) and 5% fetal calf serum (lnvitrogen) were added to the medium. Cells were cultured in a humidified incubator at 37°C and 5% CO2. Hepatocyte viability was always more than 90% and purity more than 95% as determined by trypan blue staining.

Experimental design Experiments were started 24 hrs after the isolation of hepatocytes. Cells were exposed to the intracellular superoxide anion-donor menadione (2-methyl-1,4-naphthoquinone, 50 µmolll, Sigma-Aldrich; 23), hydrogen peroxide (H2O2, 1 mmol/L, MERCK, Haarlem, The Netherlands) or a cytokine mixture (24) composed of 20 ng/ml murine TNFa (R&D Systems, Abingdon, United Kingdom), 10 ng/ml human interleukin-113 (R&D Systems) and 10 ng/ml rat interferon-y (R&D Systems) for the indicated time. Activation of ERK1/2 mitogen-activated protein kinase (MAPK) was inhibited using the MEK inhibitor U0126 at 10 µmol/L (Promega, Madison, WI, USA). 1 H-[1,2,4]oxadiazolo-[4,3-a]quinoxalin-1-one (ODQ, 50 µmol/L, Sigma­Aldrich) was used to inhibit soluble guanylate cyclase (sGC). Depletion of intracellular reduced glutathione was achieved using L-buthionine-(S, R)-sulfoximine (BSO, 200 µmol/L, Sigma-Aldrich) and 1 mmolll DEM (Sigma-Aldrich). In some experiments, caspase activity was inhibited using a cocktail of caspase inhibitors, composed of 5 µmol/L Z-DEVD-FMK (caspase-3 inhibitor; R&D Systems), 5 µmol/L Z-IETD-FMK (caspase-8 inhibitor; R&D Systems), 5 µmol/L Z-VEID-FMK (caspase-6 inhibitor; R&D Systems) and 5 µmol/L Z-LEHD­FMK (caspase-9 inhibitor; R&D Systems). Hepatocytes were exposed to the adenoviruses Ad5HO-1, Ad51KBAA or Ad5LacZ (multiplicity of infection of 10) 15 hours before exposure to the superoxide anion-donor menadione or cytokine mixture. Parallel experiments were performed in a normal and a carbon monoxide containing culture atmosphere. Hepatocytes were placed in the carbon monoxide containing atmosphere after attachment, approximately 15 hrs before the start of the experiments. Each experimental condition was performed in triplicate wells. Each experiment was performed, at least three times, using hepatocytes from different isolations.

Carbon monoxide exposure CO at a concentration of 1 % (10000 parts per million, ppm) in compressed air was mixed with compressed air containing 5% CO2 before being delivered into the culture incubator, yielding a final concentration of 400 ppm CO. The incubator was humidified and maintained at 37°C. A CO analyzer was used to determine CO levels in the chamber. After the chamber had stabilized, no oscillations were measured in the CO concentration.

Caspase enzyme activity assay Caspase-3 enzyme activity was assayed as described previously (25).

Sytox green nuclear staining Rupture of the plasma membrane distinguishes necrotic from apoptotic cell death (26). To determine necrosis, hepatocytes were incubated 1 5 minutes with Sytox green (lnvitrogen) nucleic acid stain, which only penetrates cells with compromised plasma membranes but does not cross the membranes of viable cells or apoptotic bodies. Fluorescent nuclei were visualized using an Olympus CKX41 microscope at 450-490 nm.

66

Carbon monoxide blocks oxidative stress-induced apoptosis via inhibition of the p54 JNK isoform

RNA isolation and reverse transcription polymerase chain reaction (RT-PCR) RNA was isolated using Tri-reagent (Sigma-Aldrich) according to the manufacturer's instructions. Reverse transcription was performed on 5µg of total RNA using random primers in a final volume of 75µ1 according to the manufacturer's instruction (Sigma-Aldrich).

Quantitative PCR Reverse transcription was performed on 2.5µg of total RNA using random primers in a final volume of 50µ1. Real time detection was performed on the ABI PRISM 7700 (Applied Biosystems, Nieuwerkerk aid IJssel, The Netherlands). Samples were heated for 10 min at 95°C, followed by 40 cycles of 15 sec at 95°C, and 1 min at 60°C. Each sample was analyzed in duplicate. 18S mRNA levels were used as an endogenous control. Primers and probes are listed in table 1. Table 1 Sequences of primers and probes used for quantitative PCR analysis

18S Sense 5'-CGG CTA CCA CAT CCA AGG A-3'

rat Antisense 5'-CCA ATT ACA GGG CCT CGA AA-3'

Probe 5'FAM-CGCGCAAA TT ACCCACTCCCGA-TAMRA3'

H0-1 Sense 5'-CAC AGG GTG ACA GAA GAG GCT AA-3'

rat Antisense 5'-CTG GTC TTT GTG TTC CTC TGT CAG-3'

Probe 5'FAM-CAG CTC CTC AAA GAG CTC AAT GTT GAG C-TAMRA3'

Western blot analysis Western blot analysis of cell lysates was performed by SOS-PAGE followed by transfer to Hybond ECL nitrocellulose membrane (Amersham Biosciences, Piscataway, NJ, USA). An antibody against GAPDH (glyceraldehyde 3-phosphate dehydrogenase, Calbiochem, La Jolla, CA, USA) and Ponceau S (Sigma-Aldrich) staining were used to ensure equal protein loading and electrophoretic transfer. Caspase cleavage was detected using polyclonal rabbit antibodies recognizing only cleaved caspase-9, -6, and -3. A monoclonal antibody was used to detect HO-1 (Stressgen, Sanbio BV, Uden, The Netherlands). Poly{ADP-ribose) polymerase (PARP) cleavage was detected using a rabbit anti-PARP polyclonal antibody. PARP (116 kDa) is a substrate of caspase-3 yielding a product of 89 kDa and it is considered a late marker for apoptosis. After Western blot analysis for phosphorylated ERK1/2 and -JNK using monoclonal antibodies (Santa Cruz Biotechnology, Inc, Heidelberg, Germany), blots were stripped using 0.1 % SDS in PBS/Tween-20 at 65°C for 30 minutes and incubated with antibodies against, total-ERK1/2 or total-JNK. All antibodies were obtained from Cell Signaling Technology, Inc (Danvers, MA, USA), unless indicated otherwise, and used at 1: 1000 dilution. The blots were analyzed in a ChemiDoc XRS system (Bio-Rad, Hercules, CA, USA). Protein band intensities were quantified by Quantity One software (Bio-Rad).

Adenoviral constructs Recombinant, replication-deficient adenovirus Ad5IKBAA was used to inhibit NF-KB activation as previously described (27). Functionality of the Ad5IKBAA virus was demonstrated by its ability to sensitize hepatocytes to cytokine-induced hepatocyte apoptosis. Ad5LacZ containing the E.coli 13-galactosidase gene, was used as a control virus. The Ad5HO-1 adenovirus was a kind gift of prof. Augustine Choi, University of Pittsburg, Pittsburg, USA and has been described before (28, 29).

Statistical analysis All data are expressed as the mean of at least 3 independent experiments ± standard deviation. Statistical significance was determined by the Mann-Whitney U test; p<0.05 was considered statistically significant.

67

Chapter s

5.4 Results

Regulation of H0-1 in vivo To investigate the regulation of HO-1 in vivo, models of chronic cholestatic liver disease (BDL) and acute inflammation were used. Chronic cholestasis increased HO-1 mRNA level about 6-fold compared to controls in a time dependent manner (Fig 1 ).

A 1 week 2 weeks 4 weeks Figure 1. HO-1 expression is nduced in

Con Sham BOl Sham SOL Con Sham BOl chronic cholestasis. Rats were subjected to

ilill!!!a- I H0-1 bile duct ligat-on as a model of cholestasis. - -At Indicated t me points after BDL, rats (n=4

B per group) were sacrificed, bvers were perfused with saline and removed Control

! rats received a sham operation (SHAM) or

I 5 not (Con). H0-1 expression is induced in • BDL at prote·n leve (A) and at mRNA level

� (B) GAPDH was used as a protein oadlng J control ,n Western blots and 1 BS mRNA E

;; 2 levels were used as an endogenous control :z: � ,n qPCR. Statistical analysis p<0 05 1 compared to control in each group #

p<O 05 compared to BDL 1 week group. a

Control Shem BOl Sham BDL Conlrol Sham BDL 1 week 2 weeks 4 weeks

In contrast, acute inflammation does not increase HO-1 expression, although iNOS expression was highly elevated in the same livers at protein and at mRNA level as

00

CON lPS • OEM OEM

previously demonstrated (30). DEM-induced oxidative stress, superposed on acute inflammation, strongly increases HO-1 mRNA level in acute inflammation in vivo. Induction of oxidative stress with DEM alone also increases HO-1 mRNA in the liver (Fig 2). The induction of HO-1 by DEM alone or in combination with LPS correlates with results obtained in rat intestine in the same model (31 ). Figure 2. Acute inflammation does not induce HO-1 expression ,n lhe liver.

However OEM-induced oxtdal ve stress, superposed on acute inflammation, strongly increases HO-1 mRNA evel in acute inflammation ,n vivo. qPCR analys s of cONA from rats ( ,ver) exposed to LPS or the same volume of PBS (contro group Con) with or without DEM. In the figure, a representative set of a group (n-4) is shown Statistica analysis: · p<0.05 compared to control. # p<0.05 compared to LPS group.

Regulation of H0-1 in vitro To confirm our in vivo data we investigated the regulation of HO-1 expression in vitro in primary rat hepatocytes. The superoxide anion-donor menadione induces HO-1 expression in a time (Fig 3A) and concentration dependent manner (Fig 38).

Figure 3. The 1r1tracellular superoxide an on­donor menad1one induces HO-1 expression in a lime- and concentration-dependent manner in vitro. Cells were exposed to menadione (Men, 50 µmol/L) and harvested at different time points (A) or exposed to different concentrations of menadione for 9 hrs (B) HO-1 expression was detenmined at A) protein level by Western b ot. GAPDH was used as a protein loading control and at B) mRNA level by qPCR (9 hrs). 1BS mRNA levels were used as an endogenous control. Statistical analysis • p<0 05 compared to control.

68

A CON

I-B C

i 25 : 20

i 1 5

E 10

"'

Men 2hr

Control

Men 4hr

Men SµM

Mon 6hr

Mon lltu

Men 10µM

Men 12hr

Men 25µM

Men 24hr kOa

1 32 • H0-1

: 3B - GAPDH

Men 50µM

Carbon monoxide blocks oxidative stress-induced apoptosis via 1nh1b1t1on of the p54 JNK isoform

Exposure of hepatocytes to H2O2 does not induce HO-1 expression in hepatocytes (Fig 4A). In contrast, cytokine mixture only slightly induces HO-1 expression in vitro (Fig 5), although iNOS expression was highly induced under the same circumstances as shown previously (30). Depletion of glutathione using BSO, an inhibitor of glutamate cysteine ligase the rate limiting step in GSH biosynthesis, or DEM, further enhances the induction of HO-1 mRNA and HO-1 protein level in vivo (Fig 2) and in vitro (Fig 4, 5) in all tested conditions.

e Con Men Men •SSO eso kDa

1-.- ;;;;;;;;--- ;;;;;;;;;1 32 - H0-1

Figure 4. H0-1 expression is not increased by hydrogen peroxkfe in vitro. Cells were exposed to hydrogen peroxtde or I .... _ 35 - GAPDH

• I menadione with or without BSD or OEM. H0-1 mRNA level was

- - _ increased by menadione, BSD and OEM but not by hydrogen HPi H • .O: • BSO Hi'J: • OE.M ..., Meri • BSO Men • DEM peroxide as determined by qPCR (A) and H0-1 expression is CON

induced by menadione and BSO as determined by Western blot (B). Statistical analysis: • p<0.05 compared lo control. # p<0.05 compared to H202 group. ,:, P< 0.05 compared to menadione treated cells.

Figure 5. Cylokines slightly induce H0-1 expression in primary hepatocytes in vitro. Cells were exposed to cytoklne mixture (10 hrs) with or without BSD or OEM. H0-1 mRNA level was increased by B50 and DEM but only a minor increase was observed after exposure Jo CM as determined by qPCR. Slalislical analysis: • p<0.05 compared lo control group. # p<0.05 compared lo CM treated group.

Heme Oxygenase-1 overexpression protects against oxidative stress To investigate whether HO-1 plays a role in the protection against oxidative stress, HO-1 was overexpressed in primary rat hepatocytes by transfection with Ad5HO-1 15 hrs before exposure to menadione. Although metalloporphyrins like hemin or chromium mesoporphyrin are widely used to induce or reduce expression of HO-1 respectively, in our hands it showed serious toxic side-effects. This confirms recent reports that metalloporphyrins induce oxidative stress themselves and therefore cannot be used in combination with menadione (32). In addition, it has recently been reported that metalloporphyrins interfere with caspase activity, independent of its HO-1 inducing effect (33). To avoid these confounding side effects, we opted to use an adenoviral construct to achieve overexpression of HO-1. As shown in Fig 6A, Ad5HO-1 infection dose dependently increased HO-1 expression at mRNA and protein level in hepatocytes at the time of menadione exposure. Overexpression of HO-1 protected against menadione-induced caspase-3 activity and processing compared with control and Ad5LacZ infected hepatocytes (Fig 6B). Hepatocytes infected with only the Ad5LacZ virus or Ad5HO-1 virus and not exposed to menadione did not display elevated caspase-3 activity (Fig 6B).

A Ad5l.ocZ .Ad!!H0-1 Ad5LacZ AdSH0-1 � � � �kDa

I � --- 1 32 -H0,1

1..-- � -135 -GAPOH

rn I Conln>I Ad5La'1 Ad5H().1

11 -1

Con Men +

AdSH0-1 ..... B I � ...... 1 11 ... �,po

�I!' !!!!!!!!!!��!!!!!· !!!!!!!!!!�!!!!!!!!!!!!.�!!!!llt!-�I 3" - GAPDH

i�J.___. ------ . __ I ___ I ---· ---· Con Men + Men + Men AdSlacZ Ad9-10-\

Ad5H0-1 Ad5La:Z

Figure 6. Overexpression of HO-1 , using the adenovirus AdSHO-1, leads to protectk>n of primary hepatocytes against oxidative stress-induced apoptosis Cells were exposed to Ad5H0-1 or Ad5LacZ (virus control) 15 hrs before the addition of menadione (50 µmoVL, 9 hrs) A) Ad5H0-1 highly Induces H0-1 expression dose dependent y at protein level (upper panels) and at mRNA level (up to 70-fold as delermined by qPCR, lower graph) compared with Ad5LacZ and conlrol B) Overexpression of H0-1 prevenls capase-3 aclivily (graph) and processing (Western biol, upper panel) Statistical analys s • p<O 05 compared to control. # p<0.05 compared to menadione treated cells.

69

Chapter s

Carbon monoxide protects primary rat hepatocytes against oxidative stress-induced apoptosis Since HO-1 overexpression protects primary rat hepatocytes against superoxide anion­induced apoptosis, we investigated whether CO contributes to the protective effect of HO-1 expression. CO suppresses menadione-induced PARP-cleavage and apoptosis in primary hepatocytes, indicating that CO has a protective role against superoxide anion-induced apoptosis (Fig 7A, B).

A B -Con Men Con Men kOa �

1� -�- 7 ":o } PARP f - � ·- 38 - GAPDH $

:!. Nanna! atmosphere CO atmosphere [j

50000

40000

30000

20000

10000

Nonna! almosphe•e CO atmos:inllre

Figure 7. Carbon mono,dde prevents O)(idative stress-induced apoptosis in hepatocytes. Cells were incubated with menadione or not (control) in the presence and absence of carbon monoxide A) Western blot analysis: CO Inhibits PARP-cleavage induced by menadlone. GAPOH was used as a protein loading control. B) Caspase-3 activity assay: Carbon monoxide blocks

menadtone-induced apoptos s (expressed as capase 3 aclivaty) Statistic-a[ analysis. • p<0.01 compared to control. # p<0.01 compared to menadione treated cells in a normal atmosphere.

In addition, CO also inhibits menadione-induced caspase-9, -6 and -3 activation (Fig 8), which is in accordance with our previous finding that menadione-induced apoptosis is

con �� con � •o• dependent on caspase activation in primary hepatocytes (27). CO did not induce necrosis in primary rat hepatocytes (data not shown), suggesting that CO is really protective against oxidative stress-induced apoptosis and does not switch the balance between apoptosis and necrosis.

I = I � I: =�:::::::::

:--=:::::::w!::�:::::::::l'iiii::::;_: : ::: ··· .... 3 ====-=====

Norma. a1'n05phere CO alm05phere

Figure 8 Carbon monoxide prevents menad10ne induced caspase activation in hepatocytes Celis were incubated w th menadione (50 µmoJ/L, 9 hrs) and without (control) in the presence and absence of carbon monoxide Western blot CO inhibits caspase-9, -6 and -3 cleavage induced by menadione. GAPDH was used as a protein fOading control.

The protective effect of CO is not dependent on soluble guanylate cyclase To investigate whether soluble guanylate cyclase (sGC) is involved in the protective effect of CO against menadione-induced apoptosis in primary hepatocytes, cells were exposed to the sGC inhibitor ODQ with or without menadione in the presence of carbon monoxide. The ability of CO to inhibit menadione-induced apoptosis was not reversed by ODQ (Fig 9), indicating that the anti-apoptotic effect of CO is not dependent on the activation of sGC or the generation of cGMP. In the absence of CO, ODQ had no effect on primary hepatocytes.

35000

25000

f 20000

l 15000

10000

5000

Con Men SOµM

Men + ODO

CO atmosphere

ODO Con Men SO.,M

Men + ODO ODO

Normal atmosphere

Figure 9. The protective effect of carbon monoxide is not dependent on the act1VaUon of sGC. Cells were exposed to ODO with or without menad10ne (50 µmol/L, 9hrs) in the presence of carbon monoxide Caspase�3 activ ty assay The abi'.ity of CO to nhibit menadione-induced apoptos1s was not reversed by ODO Statistical analysis • p<0 05 compared to control. # p<0.05 compared to menadione treated cells in a normal atmosphere

70

Carbon monoxide blocks oxidative stress-induced apoptosis via inhibition of the p54 JNK isoform

NF-KB pathway is not involved in the protective effect of CO against menadione­induced apoptosis In our study, inhibition of the NF-KB pathway by recombinant adenovirus expressing dominant negative IKB did not reverse the protective effect of CO against superoxide anion­induced apoptosis (Fig1 0A). Functionality of the virus was demonstrated by sensitizing hepatocytes to cytokine-induced apoptosis (Fig 108). A B

Men Men • AdShi:SAA

s-IL � � > "fl .. "I "' .. a. "' .. u

40000

30000

20000

1 0000

CM CM +Adh,BAA

CO atmosphere Normal atmosphere

Figura 10. NF-KB pathway is not involved in the protective effect of CO against oxidative stress-induced apoptoslS A) Caspase-3 activity assay inhibition of the NF-KB pathway, using the AdSdnlKB 1 5 hrs before menad1one addition, did not reverse the protective effect of carbon monoxide against menadione-induced apoptosis. B) Functionality of the Ad5IKBAA virus was demonstrated by its abi ity to sensitize hepatocytes to cytokine-induced hepatocyte apoptosis. Statistical analysis • p<0 05 compared to control. # p<0.05 compared to menadione treated cells in a normal atmosphere.

Carbon monoxide blocks phosphorylation of the pro-apoptotic MAP-kinase JNK Menadione-induced caspase activation and apoptosis are dependent on J NK activity in primary rat hepatocytes (27). Since CO prevents caspase-9, -6 and -3 activation, we also investigated the effect of CO on MAPK-phosphorylation. As shown in figure 11A, CO induces ERK-phosphorylation at 2hrs. To investigate the role of ERK in the protective effect of CO against menadione-induced apoptosis, hepatocytes were exposed to CO in the presence of the ERK1/2 inhibitor U0126. Inhibition of ERK1/2 did not reverse the protective effect of CO (Fig 11 B). The functionality of U0126 was shown by its ability to reduce ERK1/2 phosphorylation as shown by Western blot (Fig 11 C). On the other hand, CO inhibits predominantly the menadione-induced phosphorylation of the p54 JNK isoform (Fig 11A). Since inhibition of JNK activity, using the inhibitor SP6001 25, prevents menadione-induced caspase activation and apoptosis (27), our results indicate that the protective effect of CO against superoxide anion-induced apoptosis is at least partly due to inhibition of JNK activation. To distinguish between the possibility that JNK activation is the cause of apoptosis and the possibility that J NK activation is the result of apoptosis, we determined JNK phosphorylation in response to menadione in the presence of a cocktail of caspase inhibitors (caspase-3,6,8,9 inhibitors). Inhibition of apoptosis did not change menadione-induced JNK phosphorylation (data not shown).

CO + CO + "- Control Con:rol � Me kOa - I .. } ... ·· - - - - •e == � � totlhJNK (1h!)

; � � ��.::,�) , :I I l ;1--.....-------.....-1--.

Con Con ""*"' Men •CO •CO

.

1= 1 I ·: .

ea,

I 1=

- • Ii -.... •U0129

=-... ..... ..... U)l,0 -- ---U)l,0

l �J ·--1 :;} t6'K

- I --Figure 1 1 . Carbon monoxide Induces ERK1/2- and inhibits JNK-phosphorylalion. Cells were exposed to menadione (1·2 hrs, A) or 9hrs

(B) in the presence and absence of carbon monoxide. A) Western blot analysis against phosphorylaled and total JNK, phosphorylated and total ERK1/2 CO induces ERK112 phosphorylalton and prevents menadione-induced p54 JNK phosphorylation. Western blots were quantified and band intensities of phosphorylated protein were expressed as ratio of the intensity of the corresponding total protein (lower graphs) B) Caspase-3 activity assay Cells were exposed lo U0126 before and during Iha incubation with menadione in the presence of CO. Inhibition of ERK1/2 activation does nol prevent the protective effect of CO. C) Inhibition of ERK1/2 aclivation by lhe specific inhibitor U0126. Statistical analysis • p<0 05 compared to control # p<0 05 compared to menadione treated cells in a normal atmosphere.

71

Chapter 5

5.5 Discussion

In this study, we demonstrate that HO-1 expression is induced in different models of oxidative stress in the liver both in vitro and in vivo, including chronic cholestasis. Interestingly, acute inflammation, induced by endotoxin administration, is a very weak inducer of H O-1, although in this model NF- KB-regulated genes like iNOS are highly induced as shown previously (30). A similar reciprocal regulation of AP-1 responsive genes like HO-1 and NF-KB responsive genes like iNOS was observed in intestinal epithelial cells (31) and is in accordance with other reports (39). We have demonstrated that this switch from an NF-KB regulated stress response to an AP-1 regulated stress response in hepatic inflammation is controlled by the extent of oxidative stress. An important regulator of this switch might be the content of reduced glutathione (GSH) in the cell: GSH depletion inhibits NF-KB-mediated cytokine-inducible iNOS expression and at the same time increases oxidative stress and the expression of oxidative stress inducible genes like HO-1 (30, 31 ). These results suggest that TNFa or CM (in vitro) or LPS exposure (in vivo) in the presence of sufficient GSH-levels do not induce important oxidative stress and hence no AP-1 activation and HO-1 induction. This switch might be an adaptation to the type of stress. NF­KB regulated protective genes include antl-apoptotic genes, for example iNOS, which protect against pro-apoptotic inflammatory cytokines like TNFa, whereas AP-1 regulated genes, such as HO-1, are involved in the protection against various forms of oxidative stress. In vitro, in primary cultures of rat hepatocytes, HO-1 expression is induced by the superoxide anion donor menadione, but not by H2O2• Hepatocytes have a high capacity to detoxify H2O2 (27), and it is likely that this molecule is hardly sensed by hepatocytes. This also correlates with our previous finding that 1 mmol/L H2O2 does not induce cell death. The induction of HO-1, both in vivo and in vitro, was much more pronounced after depletion of the antioxidant glutathione. This correlates with our in vivo model of acute inflammation demonstrating that acute inflammation is not accompanied by extensive oxidative stress and depletion of antioxidants.

Numerous studies have shown the cytoprotective and anti-apoptotic properties of HO-1 against oxidative stress (34). Although, the superoxide anion-donor menadione induces HO-1 expression in primary hepatocytes, this induction does not prevent apoptosis, probably because HO-1 induction is not high and/or rapid enough. In support of this we observed that overexpression of HO-1 prior to menadione protects against oxidative stress-induced apoptosis. Metalloporphyrins like hemin or chromium mesoporphyrin are widely used as inducers or inhibitors of HO-1, respectively. In our hands these compounds showed serious side-effects, confirming recent reports that describe aspecific effects of these compounds, in particular when studying phenomena related to oxidative stress and apoptosis (32, 33).

In this study we provide evidence that carbon monoxide (CO), one of the products of HO-1, protects against superoxide anion-induced apoptosis. Previously, we showed that superoxide anion-induced apoptosis is dependent on caspase activation (27). In accordance with these findings, we demonstrate that CO is a potent inhibitor of caspase activation, PARP cleavage and apoptosis in primary hepatocytes exposed to the superoxide anion­donor menadione. Since CO inhibits caspase-9 activation, we speculate that CO prevents disruption of mitochondria and subsequent apoptosis. Indeed, it has been shown that CO blocks mitochondrial cytochrome c release (35, 36). We also show that CO does not shift the balance from apoptotic cell death to necrotic cell death. This is not a trivial finding as we have shown previously (27). Therefore, CO really protects hepatocytes against superoxide anion-induced cell death. Our findings are in accordance with studies using different inducers of hepatocyte cell death, including glucose-deprivation in BNL CL.2 hepatoma cells or by Fas (37) and TNFa in hepatocytes (38). In addition, a protective effect of CO has been demonstrated in a wide range of different cell types (35, 39, 40). In contrast, a pro-apoptotic effect of CO has been observed for murine thymocytes and bovine pulmonary arterial endothelial cells (41, 42). These data indicate that CO may regulate apoptosis and cytotoxicity in a cell-specific manner. 72

Carbon monoxide blocks oxidative stress-induced apoptosis via inhibition of the p54 JNK isoforrn

It has been suggested that the cytoprotective effect of CO is due to activation of soluble guanylate cyclase, however we saw no effect of sGC inhibition on the protective effect of CO. Previous studies are contradictory concerning the role of sGC in the anti-apoptotic effect of CO, both confirming (40) and contradicting (35) our findings. Again, different signaling pathways may be involved in the protective effect of CO in different cell types. NF-KB inhibition did not influence the protective effect of CO against superoxide anion-induced apoptosis in hepatocytes, although it was reported that the protective effect of CO was dependent on NF-KB activation in hepatocytes exposed to TNFa (38).

In this study we demonstrate that CO inhibits JNK activation, in particular the p54 isoform of JNK. A similar inhibitory effect of CO on JNK-phosphorylation has been described in a model of sepsis in vivo (43). In a previous study, we demonstrated that superoxide anion-induced apoptosis is dependent on J NK activity, since inhibition of J NK activity prevents menadione­induced caspase activation and apoptosis in primary hepatocytes (27). However, we cannot conclude beyond any doubt that the inhibitory effect of CO on J NK activity is responsible for the protective effect of CO, since CO appears to inhibit mainly the phosphorylation of the p54 isoform of J NK, whereas the pharmacological inhibitor SP600125 inhibits both the p54 and p46 isoforms of JNK.

Menadione-induced JNK activation was not changed in the presence of a cocktail of caspase inhibitors. Together with our previous finding (27) that inhibition of JNK activity using SP600125 abolishes menadione-induced caspase activation and apoptosis and that JNK activation precedes caspase activation and apoptosis by at least 6-8 hours, these results indicate that reduced JNK activation is the cause of reduced apoptosis and not the result of reduced apoptosis.

The mechanisms by which J NK exerts its pro-apoptotic properties are probably diverse and not completely elucidated yet. Many studies suggest that JNK triggers the mitochondrial death pathway, including phosphorylation and activation of pro-apoptotic bcl-2 family members (44-50). Recently, a role for JNK in the degradation of the anti-apoptotic protein c­FLIP has been reported (51 ).

Another finding of the present study is that CO induces ERK 1/2 MAPK-phosphorylation. ERK1/2-activity is anti-apoptotic in hepatocytes and its activation attenuates cell death. However, although significant, the magnitude of the protective effect of ERK1/2 activity against superoxide anion-induced apoptosis is moderate (27). Apparently, the anti-apoptotic action of CO is mainly due to its inhibitory effect on J NK activation and not to its stimulatory action on ERK1/2 activation. Regulation of MAPK signaling pathways by CO in a variety of cell types has been reported in other studies (36, 52). In addition, it has been reported that CO mediates protection against CYP2E1-dependent cytotoxicity in part via inhibition of CYP2E1 activity (16).

In this study, we focus on the mechanism of the protective action of CO. It has been described that biliverdin, another product of HO-1, is also protective against oxidative stress­induce injury in part due to its antioxidant properties (17). Indeed, a cooperative protective effect of biliverdin and carbon monoxide has been shown in immune-mediated liver injury in mice (18).

In conclusion, HO-1 is induced by oxidative stress in primary hepatocytes in vivo and in vitro. HO-1 protects primary hepatocytes against superoxide anion-induced apoptosis, and this protection may be mediated in part via CO production. CO inhibits the activity of the pro­apoptotic JNK MAPK and, in addition, induces ERK1/2 MAPK anti-apoptotic activity. Modulation of HO-1 expression, or the levels of its product CO, may become an important therapeutic target in liver disorders caused by exposure to excessive oxidative stress.

73

Chapter s

5.6 Acknowledgements

This study was supported by the Dutch Digestive Foundation and the J .K. de Cock Foundation.

5. 7 Reference List

74

(1) Reid AE. Nonalcoholic steatohepatitis. Gastroenterology 2001 Sep;121(3):710-723.

(2) Bomzon A, Holt S, Moore K. Bile acids, oxidative stress, and renal function in biliary obstruction. Semin

Nephrol 1997 Nov:17(6):549-562.

(3) Loguercio C, Federico A. Oxidative stress in viral and alcoholic hepatitis. Free Radle Biol Med 2003 Jan

1 ;34(1): 1 -10.

(4) Thornberry NA, Lazebnlk Y. Caspases: enemies within. Science 1998 Aug 28;281 (5381 ): 1 31 2-1316.

(5) Riedl SJ, Shi Y. Molecular mechanisms of caspase regulation during apoptosis. Nat Rev Mal Cell Biol

2004 Nov:5(11 ):897-907.

(6) Nieminen AL. Apoptosis and necrosis in health and disease· role of mitochondria. Int Rev Cytol

2003;224:29-55.

(7) Granville DJ, earthy CM, Hunt DW, McManus BM. Apoptosis: molecular aspects of cell death and

disease. Lab Invest 1998 Aug;78(8):893-913.

(8) Bauer M , Bauer I . Heme oxygenase-1: redox regulation and role in the hepatic response to oxidative

stress. Antioxid Redox Signal 2002 Oct;4(5):749-758.

(9) Ryter SW, Otterbein LE, Morse D, Choi AM. Heme oxygenase/carbon monoxide signaling pathways:

regulation and functional significance. Mal Cell Biochem 2002 May:234-235(1-2):249-263.

(10) Rensing H, Bauer I , Dalene V, Palau C, Pannen BH, Bauer M. Differential expression pattern of heme

oxygenase-1 /heat shock protein 32 and nitric oxide synthase-II and their impact on liver injury in a rat

model of hemorrhage and resuscitation. Grit Care Med 1999 Dec:27(12):2766-2775.

(11 ) Amersi F, Buelow R, Kato H, Ke B, Caito AJ, Shen XD, et al. Upregulation of heme oxygenase-1

protects genetically fat Zucker rat livers from ischemia/reperfusion injury. J Clin Invest 1999

Dec:104(11 ):1631-1639.

(12) Kyokane T, Norimlzu S, Taniai H, Yamaguchi T, Takeoka S, Tsuchida E, et al. Carbon monoxide from

heme catabolism protects against hepatobiliary dysfunction in endotoxin-treated rat liver. Gastroenterology 2001 Apr; 120(5 ): 1 227-1240.

(13) Dorman RB, Bajt ML, Farhood A, Mayes J, Jaeschke H. Heme oxygenase-1 induction in hepatocytes

and non-parenchymal cells protects against liver injury during endotoxemia. Comp Hepatol 2004 Jan

14;3 Suppl 1:S42.

(14) Otterbein LE, Kolls JK, Mantell LL, Cook JL, Alam J, Choi AM. Exogenous administration of heme

oxygenase-1 by gene transfer provides protection against hyperoxia-induced lung Injury. J Clin Invest

1999 Apr;103(7):1047-1054.

(15) Sass G, Soares MC, Yamashita K, Seyfried S, Zimmermann WH, Eschenhagen T, et al. Heme

oxygenase-1 and its reaction product, carbon monoxide, prevent inflammation-related apoptotic liver

damage in mice. Hepatology 2003 Oct:38(4):909-918.

(16) Gong P, Cederbaum Al, Nieto N. Heme oxygenase-1 protects HepG2 cells against cytochrome P450

2E1-dependent toxicity. Free Radie Biol Med 2004 Feb 1 ;36(3):307-318.

(17) Fondevila C, Shen XD, Tsuchiyashi S, Yamashita K, Csizmadia E, Lassman C, et al. Biliverdin therapy

protects rat livers from ischemia and reperfusion injury. Hepatology 2004 Dec:40(6):1333-1341. (18) Sass G, Seyfried S, Parreira Soares M, Yamashita K, Kaczmarek E, Neuhuber WL, et al. Cooperative

effect of biliverdin and carbon monoxide on survival of mice in immune-mediated liver Injury. Hepatology

2004 Nov;40(5):1128-1135.

(19) Maines MD. The heme oxygenase system: a regulator of second messenger gases. Annu Rev

Pharmacol Toxlcol 1 997;37:51 7-554.

(20) Verma A, Hirsch DJ, Glatt CE, Ronnett GV, Snyder SH . Carbon monoxide: a putative neural messenger. Science 1993 Jan 15;259(5093):381-384.

(21) Kountouras J, Billing BH, Scheuer PJ Prolonged bile duct obstruction: a new experimental model for cirrhosis in the rat. Br J Exp Pathol 1 984 Jun:65(3):305-311.

(22) Moshage H, Casini A, Lieber CS. Acetaldehyde selectively stimulates collagen production in cultured rat liver fat-storing cells but not in hepatocytes. Hepatology 1990 Sep;12(3 Pt 1 ):511-51 8.

Carbon monoxide blocks oxidative stress-induced apoptosis via inhibition of the p54 JNK isoform

(23) Thor H, Smith MT, Hartzell P, Bellomo G, Jewell SA, Orrenius S. The metabolism of menadione (2-

methyl-1,4-naphthoquinone) by isolated hepatocytes. A study of the implications of oxidative stress in

intact cells. J Biol Chem 1982 Oct 25;257(20):12419-12425.

(24) Schoemaker MH, Gommans WM, Conde de la Rosa L, Homan M, Klok P, Trautwein C, et al.

Resistance of rat hepatocytes against bile acid-induced apoptosis in cholestatic liver injury is due to

nuclear factor-kappa B activation. J Hepatol 2003 Aug;39(2): 1 53-161 .

(25) Schoemaker MH, Conde de la Rosa L, Buist-Homan M , Vrenken TE, Havinga R, Poelstra K , e t al.

Tauroursodeoxycholic acid protects rat hepatocytes from bile acid-induced apoptosis via activation of

survival pathways. Hepatology 2004 Jun;39(6):1563-1 573.

(26) Nagai H, Matsumaru K, Feng G, Kaplowitz N. Reduced glutathione depletion causes necrosis and

sensitization to tumor necrosis factor-alpha-induced apoptosis in cultured mouse hepatocytes.

Hepatology 2002 Jul;36(1 ):55-64.

(27) Conde de la Rosa L, Schoemaker MH, Vrenken TE, Buist-Homan M, Havinga R, Jansen PL, et al.

Superoxide anions and hydrogen peroxide induce hepatocyte death by different mechanisms:

involvement of JNK and ERK MAP kinases. J Hepatol 2006 May;44(5):918-929.

(28) Fallaux FJ, Kranenburg 0, Cramer SJ, Houweling A, Van Ormond! H, Hoeben RC, et al.

Characterization of 911: a new helper cell line for the titration and propagation of early region 1-deleted

adenoviral vectors. Hum Gene Ther 1996 Jan 20;7(2):215-222.

(29) McGrory WJ, Bautista DS, Graham FL. A simple technique for the rescue of early region I mutations into

infectious human adenovirus type 5. Virology 1 988 Apr;163(2):614-617 .

(30) Vos TA, van Goor H, Tuyt L, Jager-Krikken A, Leuvenink R, Kuipers F, et a l . Expression of inducible

nitric oxide synthase in endotoxemic rat hepatocytes is dependent on the cellular glutathione status.

Hepatology 1999 Feb;29(2):421-426.

(31) Dijkstra G, Blokzijl H, Bok L, Homan M, van Goor H, Faber KN, et al. Opposite effect of oxidative stress

on inducible nitric oxide synthase and haem oxygenase-1 expression in intestinal inflammation: anti­

inflammatory effect of carbon monoxide. J Pathol 2004 Nov;204(3):296-303.

(32) Erario MA, Gonzales S, Noriega GO, Tomaro ML. Bilirubin and ferritin as protectors against hemin­

induced oxidative stress in rat liver. Cell Mol Biol (Noisy -le-grand) 2002 Dec;48(8):877-884.

(33) Blumenthal SB, Kiemer AK, Tiegs G, Seyfried S, Holtje M, Brandt B, et al. Metalloporphyrins inactivate

caspase-3 and -8. FASEB J 2005 Aug;19(10):1 272-1279.

(34) Farombi EO, Surh YJ . Heme oxygenase-1 as a potential therapeutic target for hepatoprotection. J

Biochem Mol Biol 2006 Sep 30;39(5):479-491.

(35) Liu XM, Chapman GB, Peyton KJ, Schafer Al, Durante W. Carbon monoxide inhibits apoptosis in

vascular smooth muscle cells. Cardiovasc Res 2002 Aug 1 ;55(2):396-405.

(36) Zhang X, Shan P, Alam J, Davis RJ, Flavell RA, Lee PJ. Carbon monoxide modulates Fas/Fas ligand,

caspases, and Bcl-2 family proteins via the p38alpha mitogen-activated protein kinase pathway during

ischemia-reperfusion lung injury. J Biol Chem 2003 Jun 1 3;278(24):22061-22070.

(37) Choi BM, Pae HO, Kim YM, Chung HT. Nitric oxide-mediated cytoprotection of hepatocytes from glucose deprivation-induced cytotoxicity: involvement of heme oxygenase-1. Hepatology 2003

Apr;37(4):810-823.

(38) Zuckerbraun BS, Billiar TR, Otterbein SL, Kim PK, Liu F, Choi AM, et al. Carbon monoxide protects

against liver failure through nitric oxide-induced heme oxygenase 1. J Exp Med 2003 Dec

1 ;198(11 ):1 707-1716.

(39) Zhang X, Shan P, Otterbein LE, Alam J, Flavell RA, Davis RJ, et al. Carbon monoxide inhibition of

apoptosis during ischemia-reperfusion lung injury is dependent on the p38 mitogen-activated protein

kinase pathway and involves caspase 3. J Biol Chem 2003 Jan 10;278(2):1248-1258.

(40) Brouard S, Otterbein LE, Anrather J, Tobiasch E, Bach FH, Choi AM, et al. Carbon monoxide generated by

heme oxygenase 1 suppresses endothelial cell apoptosis. J Exp Med 2000 Oct 2;192(7):1 015-1026.

(41) Thom SR, Fisher D, Xu YA, Notarfrancesco K, lschiropoulos H. Adaptive responses and apoptosis in

endothelial cells exposed to carbon monoxide. Proc Natl Acad Sci U S A 2000 Feb 1 ;97(3):1 305-1 310 .

(42) Turcanu V, Dhouib M , Gendrault JL, Poindron P. Carbon monoxide induces murine thymocyte apoptosis

by a free radical-mediated mechanism. Cell Biol Toxicol 1998 Feb;14(1 ):47-54.

(43) Morse D, Pischke SE, Zhou Z, Davis RJ, Flavell RA, Loop T, et al. Suppression of inflammatory cytokine

production by carbon monoxide involves the JNK pathway and AP-1. J Biol Chem 2003 Sep

26;278(39):36993-36998.

75

Chapter s

(44) CzaJa MJ, Liu H, Wang Y. Oxidant-induced hepatocyte injury from menadione is regulated by ERK and

AP-1 signaling. Hepatology 2003 Jun:37(6):1 405-1 413.

(45) Davis RJ . Signal transduction by the JNK group of MAP kinases. Cell 2000 Oct 1 3;1 03(2):239-252.

(46) Toumier C, Hess P, Yang DD, Xu J, Turner TK, N imnual A, et al. Requirement of JNK for stress-induced

activation of the cytochrome c-mediated death pathway. Science 2000 May 5;288(5467):870-874.

(47) Okuno S, Saito A, H ayashi T. Chan PH. The c-Jun N-terminal protein kinase signaling pathway mediates

bax activation and subsequent neuronal apoptosis through interaction with bim after transient focal

cerebral ischemia. J Neurosci 2004 Sep 8;24(36):7879-7887.

76

(48) Lei K, Nlmnual A, Zang WX, Kennedy NJ, Flavell RA, Thompson CB. et al. The Bax subfamily of Bcl2-

related proteins is essential for apoptotic signal transduction by c-Jun NH(2)-terminal kinase. Mal Cell

Biol 2002 Jul;22(1 3):4929-4942.

(49) Schroeter H. Boyd CS, Ahmed R, Spencer J P, Duncan RF, Rice-Evans C, et al. c-Jun N-terminal kinase

(JNK)-mediated modulation of brain mitochondria function: new target proteins for JNK signalling in

mitochondrion-dependent apoptos1s. B1ochem J 2003 Jun 1 ;372(PI 2):359-369.

(50) Chauhan D, Li G, Hideshima T, Podar K, Mitsiades C, Mitsiades N, et al. JNK-dependent release of

mitochondrial protein, Smac, during apoptosis in multiple myeloma (MM) cells. J Biol Chem 2003 May

1 6;278(20):1 7593-17596.

(51 ) Chang L, Kamala H, Solinas G, Luo JL, Maeda S, Venuprasad K, et al. The E3 ubiquitin ligase itch

couples JNK activation to TNFalpha-mduced cell death by inducing c-FLIP(L) turnover. Cell 2006 Feb

1 0;124(3):601-6 1 3.

(52) Tsui TY, Siu YT, Schlitt HJ. Fan ST. Heme oxygenase-1 -derived carbon monoxide stimulates adenosine

triphosphate generation in human hepatocyte. Biochem Biophys Res Commun 2005 Oct 28;336(3):898-

902.

Chapter 6

Isolation, culture and characterization of cholestatic hepatocytes: an in vitro approach to study cholestasis

Titia E. Vrenken , Manon Buist-Homan, Jannes Woudenberg, Fiona A.J. van den Heuvel , Klaas Nico Faber, Han Moshage

Department of Gastroenterology and Hepatology; University Medical Center Groningen, Un iversity of Groningen, Groningen, the Netherlands

In preparation

Chapter 6

6.1 Abstract

Background: Cholestasis or impaired bile flow is characterized by increased serum levels of bile acids, hepatocyte injury and loss of hepatocytes. Prolonged cholestasis may induce adaptive changes in hepatocytes which might affect the hepatocellular response to toxic compounds and to drugs. Therefore, detailed knowledge of these adaptive changes and altered responsiveness are required, but are limited by the lack of a suitable in vitro model to study cholestatic hepatocytes. Aim: to develop and characterize a cell culture model to study cholestatic hepatocytes in vitro. Methods: Primary rat hepatocytes were isolated from 1 week bile duct ligated (BDL) and sham operated rats. mRNA levels of bile acid transporters were analyzed by qPCR. Cholestatic and sham hepatocytes were exposed to bile acids. Apoptosis and necrosis were quantified by caspase-3 activity assay and acridine orange staining and Sytox green staining, respectively. Results: Compared to sham hepatocytes, cholestatic hepatocytes show decreased mRNA levels of Ntcp and Mrp2, and increased mRNA levels of Mrp3 immediately after isolation and during culture. The toxic bile acid glycochenodeoxycholic acid (GCDCA) induces apoptosis in sham hepatocytes, but not in cholestatic hepatocytes. The anti-apoptotic bile acid tauroursodeoxycholic acid (TUDCA) is only effective in sham hepatocytes. Conclusion: We have developed a representative in vitro model for cholestatic hepatocytes, which can be used to study signaling pathways in these hepatocytes. Our preliminary data show that cholestatic hepatocytes maintain their in vivo phenotype during culture, and respond differently to hydrophobic bile acids than sham hepatocytes.

78

Isolation, culture and characterization of cholestatic hepatocytes: an in vitro approach to study cholestasis

6.2 Introduction

Cholestasis or impaired bile flow occurs in many acute and chronic liver diseases. Cholestasis can be caused by a genetic defect (e.g. progressive familial intrahepatic cholestasis (PFIC) type 1 to 3), drugs (such as cyclosporine A, oral contraceptives, anabolic steroids) or an immunologic (e.g. viral infection, sepsis) or mechanical (gall stones, tumors) insult to the bile ducts (1-5). In some cholestatic disorders, the cause of bile duct injury remains unknown. Many of these diseases, like biliary atresia, occur in children (5). Cholestasis progresses via hepatocyte damage and activation of inflammatory cells and hepatic stellate cells. Cholestatic disorders develop via fibrosis, the increased deposition of extracellular matrix (1, 6-8), into cirrhosis and eventually end-stage liver failure (1 ).

In cholestatic patients and animal models of cholestasis like bile duct ligation (BDL), endotoxemia (lipopolysaccharide (LPS) administration) or chelate-feeding, the bile acid retention in hepatocytes affects several transporters involved in the enterohepatic circulation (2-4, 9, 10). The transporters involved in the uptake of bile acids and organic anions at the basolateral membrane, Na•/taurocholate cotransporting polypeptide (Ntcp) and organic anion transporting polypeptide (Oatp), are down-regulated. Multidrug resistance-associated protein (Mrp) 2, responsible for the secretion of bile acids and glutathione conjugates across the canalicular membrane into bile, is also down-regulated (3, 4, 9, 10), while Mrp3, which secretes bile acids and organic anions across the basolateral membrane of hepatocytes into the sinusoids (reverse transport) is up-regulated (3).

The glycine-conjugates of hydrophobic primary and secondary bile acids trigger hepatocyte apoptosis and necrosis in cholestasis (1 , 2, 1 1 , 1 2). Apoptosis or programmed cell death is a tightly regulated process, involving the caspase-family of cysteine proteases (1 3, 14). The energy-dependent apoptotic pathway results in cell shrinkage and formation of apoptotic bodies, which are cleared very efficiently by surrounding phagocytosing cells (13, 15, 16). Necrosis, a passive mode of cell death, leads to plasma membrane rupture, and subsequent release of cellular content in the extracellular environment and circulation and hence an increased inflammatory response (17-1 9).

Ursodeoxycholic acid (UDCA) is a dihydroxy bile salt which is used in the treatment of chronic cholestatic liver diseases (2, 20). In patients with primary biliary cirrhosis (PBC), primary sclerosing cholangitis (PSC) and PFIC, UDCA therapy improved liver function and histology (20). The progression of PBC was also delayed by U DCA treatment (20). The hepatoprotective mechanism of UDCA is accomplished via upregulation of the bile acid transporters and hence stimulation of bile acid secretion of hepatocytes, and via the protection of hepatocytes against bile acid induced cell death (2, 10, 20, 21). Inhibitory effects of UDCA on the mitochondrial permeability transition (MPT) and cytochrome c release from the mitochondria in in vivo and in vitro animal studies have been described (20). In addition, activation of anti-apoptotic signaling pathways including phosphoinositide-3 kinase (Pl3K) and the mitogen-activated protein kinase (MAPK) member extracellular signal regulated kinase (ERK) 1/2 by UDCA have been described (20, 22-25). Cholestatic disorders do not always respond to UDCA therapy, and surgery is often required to drain or eliminate the obstruction. Often, a liver transplantation is indicated when complications occur (1 ). Because of the limited therapeutic options for cholestatic disorders, there is an immediate need for novel therapies. Therefore, signaling pathways involved in cholestatic liver injury and the identification of potential drug targets are the focus of many studies.

In recent years, some of these pathways have been identified. Hepatocyte apoptosis, stellate cell activation and fibrogenesis were reduced in Fas-deficient (6) and cathepsin B­deficient (26) mice subjected to BDL for 3 days compared to wild type mice. Livers of bile duct ligated mice were resistant to tumor necrosis factor (TNF) a-induced apoptosis via the Pl3K-effector, Akt. (27). Furthermore, cholestatic livers were more resistant to ischemia/reperfusion injury (28), and mitochondria isolated from cholestatic livers were less

79

Chapter 6

susceptible to the MPT induced by bile acids (11 ). Other signaling molecules that have been reported to exert anti-cholestatic effects are calcium ions (25, 29), protein kinase C (PKC; 30) and hepatocyte growth factor (HGF; 31 ). In addition to cholestatic livers, cirrhotic livers were resistant to apoptosis induced by various stimuli (cytokines, transforming growth factor (TGF) 13 and UV irradiation). This resistance was dependent on an increased generation of reactive oxygen species (ROS), since these cirrhotic hepatocytes were again susceptible to apoptosis after exposure to antioxidants (32).

The different response of cholestatic livers to toxic compounds is clinically relevant and prompted us to examine in detail the underlying mechanism(s) of this altered responsiveness. For this purpose, an in vitro system of cholestatic hepatocytes that maintains its in vivo phenotype in culture is obliged. In this study, we describe a method to isolate, culture and analyze cholestatic hepatocytes in vitro. Hepatocytes were isolated from bile duct ligated rats and compared with hepatocytes from sham operated rats, immediately after isolation (to hepatocytes) and in primary culture. Both cholestatic and sham hepatocytes maintain in culture their in vivo phenotype. Our preliminary results show that the toxic bile acid glycochenodeoxycholic acid (GCDCA) induces cell death in sham hepatocytes, but not in cholestatic hepatocytes. In addition, the anti-apoptotic bile acid tauroursodeoxycholic acid (TUDCA) is only protective against apoptosis in sham hepatocytes. In conclusion, we report a functional in vitro approach to study hepatocytes in cholestatic liver diseases, and the preliminary results indicate that cholestatic hepatocytes are resistant to GCDCA-induced cell death.

6.3 Materials and Methods

Animals Specified pathogen-free male Wistar rats (200-250 g) were purchased from Charles River Laboratories Inc (Wilmington, MA, USA). Rats were housed under standard laboratory conditions with free access to standard laboratory chow and water. Experiments were performed following the guidelines of the local Committee for Care and Use of laboratory animals.

Animal model of cholestasis Rats were anaesthetised with 02/N20 and subjected to common bile duct ligation or sham operation (SHAM) as described perviously (33). Hepatocytes were isolated from BDL (cholestatic hepatocytes) and SHAM (sham hepatocytes) rats at day 7 and 14 after surgery. Rats were weighed before and after surgery (at least three times a week) as indicator of their welfare. Blood was taken from the tail vein immediately prior to the hepatocyte isolation. Blood samples were analyzed for markers of liver damage and cholestasis.

Rat hepatocyte isolation Hepatocytes were isolated from overnight fasted rats by a two step coflagenase perfusion as described before (34). Cell viability as determined by trypan blue exclusion was greater than 84% for sham hepatocytes and greater than 68% for cholestatic hepatocytes. Aliquots of approximately 1 million cells were snap-frozen in liquid nitrogen immediately after isolation (to hepatocytes) and 112,500 cells per cm2 were plated on Vitrogen® (Cohesion Technologies Inc., Palo Alto, CA, USA) coated plates in William's E medium (lnvitrogen, Breda, The Netherlands) supplemented with 50 µg/mL gentamycin (lnvitrogen) and penicillin-streptomycin-fungizone (Lonza, Verviers, Belgium). During the attachment period (4 hrs) 50 nmol/L dexamethasone (Dept. of Pharmacy, UMCG, Groningen, The Netherlands) and 5% fetal calf serum (lnvitrogen) were added to the medium. Cells were cultured in a humidified incubator at 37°C and 5% CO2•

Experimental design Experiments were started after the attachment period of 4 hrs. Monolayer cultures of hepatocytes were exposed to 50 µmol/L GCDCA (Calbiochem, La Jolla, CA, USA) or 50

80

Isolation, culture and characterization of cholestatic hepatocytes: an in vitro approach to study cholestasis

µmol/L TUDCA (Calbiochem) for 4 hrs. Each experimental condition was performed in triplicate. Cells were harvested at the indicated time points, and washed three times with ice cold phosphate-buffered saline (PBS) before the addition of hypotonic cell lysis buffer (protein analysis, caspase-3 assay) or Tri-reagent (RNA isolation, Sigma-Aldrich) as described previously (24 ). Hepatocytes were isolated from different SHAM (n=4) and BDL (n-=2) animals.

Caspase enzyme activity assay Caspase-3 like activity was assayed as described previously (24). The arbitrary fluorescence unit (AFU) was corrected for the amount of protein in the cell lysate.

Sytox Green and acridine orange nuclear staining To determine necrotic cell death, hepatocytes were incubated with Sytox green (lnvitrogen, 1 :40,000) nucleic acid stain. Apoptotic nuclei were visualized using acridine orange (Sigma­Aldrich, 1 :5,000). Fluorescent nuclei were visualized using an Olympus CKX41 microscope at 450-490 nm as described previously (34 ). Necrosis and apoptosis were quantified by counting fluorescent nuclei (necrotic or apoptotic cells) and the total number of cells in 3 randomly chosen high power fields.

RNA isolation and quantitative Polymerase Chain Reaction (qPCR) RNA isolation, reverse transcription (RT) PCR and qPCR were performed as described previously (35). Details of primers and probes are listed in Table 1. Each sample was analyzed in duplicate. 18S mRNA levels were used as endogenous control. Table 1. Sequences of primers and probes used for quantitative PCR analysis

Sense 5'-CGG CTA CCA CAT CCA AGG A-3' 1BS

Antisense 5'-CCA ATT ACA GGG CCT CGA AA-3' rat

Probe 5'FAM-CGCGCAAA TT ACCCACTCCCGA-TAMRA3'

Sense 5'-GAC GAC GAT GAT GGG CTG AT-3' Mrp2

Antisense 5'-CTT CTC ATG GCC AAG GAA GCT-3' rat

Probe S'FAM-CCC ACC ATG GAG GAA ATC CCT GAG G-TAMRA3'

Mrp3 Sense 5'-TCC CAC TTC TCG GAG ACA GTA ACT-3'

rat Antisense 5'-CTT AGC ATC ACT GAG GAC CTT GAA-3'

Probe 5'FAM-CAG TGT CAT TCG GGC CTA CGG CC-TAMRA3'

Ntcp Sense 5'-ATG ACC ACC TGC TCC AGC TT-3'

rat Antisense 5'-GCC TTT GTA GGG ACC TTG T-3'

Probe 5'FAM-CCT TGG GCA TGA TGC CAC TCC TC-TAMRA3'

6.4 Results

Characterization of cholestatic hepatocytes isolated from bile duct ligated rats Rats were weighed before and after surgery to monitor their welfare. Sham-operated animals lost weight in the first day after surgery, but within three days body weight was back to pre-surgery levels (Fig 1A). BDL resulted in a weight loss that continued for 6 days post­surgery. Thereafter, body weight increased again with a slope comparable to the sham group (Fig 1A). Aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT) and alkaline phosphatase (AP) levels were increased 1 week after bile duct ligation, and these enzymes were further enhanced during the second week of BDL (Fig 1 B). Gamma glutamyltransferase (GGT) levels were comparable in all three groups (Fig 1 B). As expected, total bilirubin was elevated after bile duct ligation (Fig 1 B).

81

Chapter 6

A '""' B 800 SHAM

-- BDl. 1 wll - B DL 2 wk 600

"" _ .. --i:

100'% ----- _________ ., ____ _.._-....--\ _.,--- �-----

200

.,,.,__ __________ _ e e 10 12 1-1

post•pentlon time (dlys)

200 "" 40 200 OSHAM •BDL 1 wk tBOl.2..W

Figure 1 . Body weight and plasma analysis or rats after sham operation (SHAM) or bile duct ligation (BDL). (A) Up to 1 week after BDL, body weigh! is still decreased compared to pre-surgery level, while lhe weight loss caused by the surgery itself (SHAM) is recovered w·thin 3 days. (B) Plasma levels ol markers for liver damage (ASAT, ALAT, alkal ine phosphatase and gamma-GT) and cholestasis (total b lirubin).

Hepatocytes were isolated from sham and bile duct ligated-rats. After one week BDL, hepatocytes were isolated that were less viable then hepatocytes from sham-operated rats (average 70.0% and 86.9% resp.), and also fewer in number (69 vs 353 million cells). We were unable to isolate viable hepatocytes after 2 weeks BDL (data not shown), therefore we only studied hepatocytes isolated 1 week after BDL.

Expression of bile acid transporters in cholestatic hepatocytes First, we examined the effects of bile duct ligation on bile acid transporters. BDL caused down-regulation of Mrp2 and Ntcp mRNA expression both in t0 hepatocytes and primary cultures, while Mrp3 mRNA levels were increased compared to sham (Fig 2A). Next, we studied the effect of culturing (4 hrs attachment followed by 4 hrs experiment) on the expression of these transporters. In culture, Mrp2 and Mrp3 mRNA levels increased compared to t0 hepatocytes: 5.4 and 3.0 fold respectively for sham hepatocytes and 2.9 and 2.6 times for cholestatic hepatocytes (Fig 2B). Ntcp mRNA expression was enhanced in culture 1.5 times in sham hepatocytes, while in cholestatic hepatocytes the Ntcp mRNA level was not significantly changed compared to t0 hepatocytes (Fig 2B) A

i ..,,, j

. . L • t 15 t � i ,1 � I i 2

...... B0t.1a : 0

B

..,,,

IIDLh.ti

j E t I i : Q .l ·-

.. ,.

IIDL 1 M

Figure 2 . Expression pattern o f b,le acid transporters after bile duct ligat10n. qPCR analysis for Mrp2. Mrp3 and Ntcp usmg cDNA of hepatocytes from BDL or sham operated rats immediately after isolation (to) or m primary cu tures at the end of the experiments (8 hrs aller plating). (A) One week BDL reduces hepatocyte Mrp2 and Ntcp

i mRNA levels, whlle Mrp3 is

E Nlcp O SHAM increased compared to sham J

2

U:I

aaDL 1 w11. hepatocytes (8) Effects of culturing

i u on transporter expression. mRNA � levels of Mrp2 and Mrp3 are

j 0 _ increased similar in sham and EE , hep«oq1es pmaryculurH cha estatic hepatocytes during

culture

Cholestatic hepatocytes are resistant to GCDCA-induced apoptosis The toxic bile acid GCDCA induced caspase-3 activity in primary sham hepatocytes, which is reduced by the anti-apoptotic bile acid TUDCA by 66% (Fig 3A) as described previously (24). In contrast, cholestatic hepatocytes exposed to GCDCA or GCDCA in combination with TUDCA resulted in caspase-3 activities similar to control cholestatic hepatocytes (Fig 3A). TUDCA alone did not induce caspase-3 activity in sham or cholestatic hepatocytes (Fig 3A). The results from the caspase-3 activity assay (Fig 3A) were confirmed by acridine orange staining, visualizing condensed apoptotic nuclei. Sham hepatocytes exposed to GCDCA for 4 hrs demonstrate apoptotic nuclei in 36% of the cells (Fig 3B), and in combination with TUDCA the percentage of condensed nuclei was reduced to 13%. In control hepatocytes or cells exposed to TUDCA only, apoptotic cells were hardly present (Fig 3B). Cholestatic 82

Isolation, culture and characterization of cholesta tic hepatocytes: an in vi tro approach to study cholestasts

hepatocytes exposed to GCDCA displayed apoptotic nuclei in 41 % of the cells (Fig 3B), while in combination with TUDCA the percentage of condensed nuclei remained at 41 %. Apoptosis was observed in 32% of the cells treated with only TUDCA, while 23% of cholestatic control hepatocytes showed nuclear condensation (Fig 3B).

A 30 B SHAM hepatocytes OSHAM

--.. • BDL 1 wk

f! ,. ? jj

h 10

" J control GCDCA501,,1M GCDCA + TUOCA TUDCA !S01,,1M "' 38% 13% 0%

BOL hopatocytes

'""'"" GCDCA GCDCA+ TUDCA TUOCA

Figure 3. GCDCA does not induce apoptosis in primary cholestatic hepatocytes. Primary rat hepatocytes from BDL or sham operated animals were exposed for 4 hrs to GCDCA (50 µmoUL), TUDCA (50 µmol/L) or GCDCA control GCDCA &01,,1M GCDCA + TUOCA TUDCA 50µM in combination with TUDCA. (A) Caspase-3 like activity 23% '1% '1% 32%

was measured, and is shown as fold induction compared to control values of the corresponding group, control values were set at one. (B) Nuclear condensation upon GCDCA exposure was reduced by TUDCA in sham hepatocytes, but not in cholestatic hepatocytes. Primary rat hepatocytes were exposed to GCDCA, TUDCA, both bile acids or left untreated (control) Apoptot1c nuclei were determined by acridine orange staining 4 hrs after bile acid-exposure and expressed as % of total nuclei. Representative images are shown. Magnification 20X

Since control hepatocytes isolated from BDL rats displayed a marked number of apoptotic nuclei (Fig 3B), we decided to express caspase-3 activity in cholestatic hepatocytes relative to sham control hepatocytes (Fig 4). Caspase-3 activity of cholestatic control hepatocytes was 8-times higher than sham control hepatocytes, and GCDCA only slightly enhanced caspase-3 activity in cholestatic hepatocytes (Fig 4). Again, TUDCA alone resulted in caspase-3 activities similar to cholestatic control hepatocytes.

I I

CSHAM • BDL 1 wlc

I

Figure 4. Cholestatic control hepatocytes display high levels of apoptosis compared to control hepatocytes from sham animals Primary hepatocytes from BDL or sham operated rats were exposed for 4 hrs to GCDCA (50 µmol/L), TUDCA (50 µmoUL) or the combination of GCDCA and TUDCA Caspase-3 like activity was measured, and is shown as fold induction compared to control hepatocytes from sham animals; sham control values were set at one

control GCDCA GCDCA + TUDCA TUOCA

Bile acid-induced necrosis in cholestatic hepatocytes sHAM hopa_... Lastly, we analyzed necrotic cell death upon

control 311

BDL hepatoc:ytes

GCDCA50pM Ill

GCDCA• TUDCA 811

T\JOCA 511

control GCDCA 15D 1.1M GCDCA • TUOCA. TUDCA 12% 23% 14% 15%

Figure 5. GCDCA-induced necrosis is not significantly affected by TUDCA. Primary rat hepatocytes were treated without (control) or with GCDCA, TUDCA or the combination of GCDCA and TUDCA. Sytox green nuclear staining was used to visualize necrotic cell death after 4 hrs exposure. Representative images are shown. Upper panel combined phase contrast and fluorescence. Lower panel: fluorescence. Magnification 20X.

bile acid exposure in hepatocytes isolated from BDL and sham operated rats. In sham hepatocytes, necrosis was observed in 9% of the cells treated for 4 hrs with GCDCA (Fig 5) and this percentage was not significantly changed by TUDCA (8% necrotic cells). Control hepatocytes or hepatocytes exposed to TUDCA displayed low levels (<5%) of necrosis (Fig 5). In cholestatic hepatocytes, control cells showed relative high levels of necrosis (12%) compared to control sham hepatocytes (3%). Treatment with GCDCA resulted in 23% of necrotic cells, while the combination of GCDCA and TUDCA or TUDCA alone resulted in a slightly higher percentage of necrotic cells than control conditions (Fig 5).

83

Chapter 6

6.5 Discussion

The aim of this study was to develop an in vitro model for the study of cholestatic hepatocytes. Although some studies described the use of primary cholestatic hepatocytes (11, 29), a comprehensive comparison of sham and bile duct ligated rat hepatocytes in culture was never reported. We demonstrate the functionality of primary cholestatic hepatocytes both immediately after isolation and after an 8 hr culture period (4 hr attachment and 4 hr experiment). Our preliminary results suggest that cholestatic hepatocytes maintain their in vivo phenotype during the in vitro experiments and that, in contrast to sham hepatocytes, cholestatic hepatocytes are resistant to bile acid induced cell death. However, it should be noted that control hepatocytes from BDL rats already exhibit elevated levels of apoptosis and necrosis. This new approach to study the signaling pathways in hepatocyte cell death in cholestatic livers will provide more insight in possible targets for intervention.

The increases in serum ALAT, ASAT, alkaline phosphatase and bilirubin after bile duct ligation are in accordance with previous results (33) and indicate hepatocyte damage, damage to the biliary tract and cholestasis. Weight loss of bile duct ligated animals was more severe and prolonged than of sham operated rats, but after 1 week weight gain was similar in both groups. Reduced mRNA expression of Mrp2 and Ntcp and increased expression of Mrp3 mRNA has been reported in bile duct ligated rat livers in vivo (3, 4, 9, 10). These changes in mRNA expression can be explained by adaptation of the liver to increased serum bile acid levels and decreased bile flow. The down-regulation of Ntcp and Mrp2 results in reduced bile acid uptake and excretion into the canalicular lumen, while increased Mrp3 expression stimulates bile acid excretion across the basolateral membrane into the blood. In addition, the shift in mRNA expression of transporters during culture is similar in sham hepatocytes and cholestatic hepatocytes: mRNA expression of Mrp2 and Mrp3 tend to increase during culture in both sham and cholestatic hepatocytes, whereas Ntcp expression is not affected or increases slightly during culture. In this study, we demonstrate that the observed changes in transporter mRNA levels are maintained immediately after isolation of hepatocytes and during the experiments. These observations validate our in vitro model, since we demonstrates that the in vivo phenotype is preserved in vitro, and phenotypic alterations due to the culturing process do not differ between normal and cholestatic hepatocytes.

As described previously (24), GCDCA strongly induced caspase-3 activity and nuclear condensation in sham hepatocytes, which were partially inhibited by TUDCA. In contrast, caspase-3 activity in cholestatic hepatocytes did not respond to GCDCA, even though nuclear condensation upon GCDCA exposure was equal in sham and cholestatic hepatocytes (Fig 3). This apparent paradox is due to the high amount of apoptotic cells present in cholestatic control hepatocytes: 22% of these cells already display nuclear condensation. In fact, when we express caspase-3 activity in cholestatic hepatocytes relative to sham hepatocytes, the cholestatic hepatocytes already show significantly induced levels of caspase-3 activity (Fig 4 ). Surprisingly, despite the lower starting viability of cholestatic hepatocytes the exposure to the hydrophobic bile acid GCDCA did not induce additional caspase-3 activity. The percentage of necrotic cells was also elevated in cholestatic control hepatocytes and GCDCA treatment did not significantly change necrosis (Fig 5). In contrast, bile acid exposure to sham hepatocytes increased necrosis as reported previously (unpublished data), however TUDCA is not protective against GCDCA-induced necrosis.

In summary, our preliminary data show that in cholestatic hepatocytes the response to GCDCA is very limited relative to control hepatocytes. We can only speculate about possible explanations. One important reason could be that conjugated bile acids are no longer able to enter the hepatocytes. The down-regulation of Ntcp, the transporter involved in the basolateral uptake of conjugated bile acids, by 82% implies that Ntcp is almost absent at the plasma membrane and therefore, a minimal bile acid import into the cholestatic hepatocyte compared to sham hepatocytes. In addition to the altered expression of bile acid 84

Isolation. culture and charactenzation of cholestatic hepatocytes: an in vitro approach to study cholestasis

transporters, other mechanisms may be operative in the altered responsiveness of cholestatic hepatocytes to bile acids. Recently, increased resistance of cirrhotic hepatocytes to TGFl3 was reported. This increased resistance was due to sustained generation of ROS in cirrhotic hepatocytes. The addition of antioxidants to cirrhotic hepatocytes completely abolished the increased resistance to TGFl3. ROS are known to activate numerous signal transduction pathways and genes such as heme oxygenase-1 (HO-1) involved in the regulation of cell survival. Our in vitro model provides the opportunity to study these mechanisms in cholestasis in detail.

Therapy of chronic liver diseases usually starts when the disease is already advanced. Therefore, therapy should be effective in 'chronically diseased' hepatocytes, e.g. cholestatic or cirrhotic hepatocytes. Hence, an in vitro model which represents 'diseased' livers is obliged. Our study provides the first steps towards a representative in vitro model for cholestatic hepatocytes, which can be used to study adaptive changes in these hepatocytes and to examine the effects of drugs in cholestatic liver diseases. Our preliminary data imply that cholestatic hepatocytes respond different to hydrophobic bile acids than sham hepatocytes.

6.6 Acknowledgements

This study was supported by the Dutch Digestive Foundation (WS 02-13) and the J .K. de Cock Foundation.

6. 7 Reference List

(1) Sokol RJ, Devereaux M, Dahl R, Gumpricht E. "Let there be bile " - Understanding hepatic injury in

cholestasis. Journal of Pediatric Gastroenterology and Nutrition 2006 Jul;43:S4-S9.

(2) Paumgartner G. Medical treatment of cholestatic liver diseases: From pathobiology to pharmacological

targets. World Journal of Gastroenterology 2006 Jul 28;12(28):4445-4451 .

(3) Kullak-Ublick GA, Stieger B, Meier PJ. Enterohepatic bile salt transporters in normal physiology and liver

disease. Gastroenterology 2004 Jan;126(1 ):322-342.

(4) Alrefai WA, Gill RK. Bile acid transporters: structure, function, regulation and pathophysiological implications. Phann Res 2007 Oct;24(10):1 803-1823.

(5) Rutherford AE, Pratt DS. Cholestasis and cholestatic syndromes. Curr Opin Gastroenterol 2006

May;22(3 ):209-214.

(6) Canbay A, Higuchi H, Bronk SF, Taniai M, Sebo TJ, Gores GJ. Fas enhances fibrogenesis in the bile

duct ligated mouse: a link between apoptosis and fibrosis. Gastroenterology 2002 Oct:123(4):1323-1330.

(7) Jaeschke H, Gores GJ, Cederbaum Al, Hinson JA, Pessayre D, Lemasters JJ. Forum - Mechanisms of

hepatotoxicity. Toxicological Sciences 2002 Feb;65(2):166-176.

(8) Kolios G, Valatas V, Kouroumalis E. Role of Kupffer cells in the pathogenesis of liver disease. World J

Gastroenterol 2006 Dec 14;1 2(46):7413-7420.

(9) Geier A, Dietrich CG, Voigt S, Kim SK, Gerloff T, Kullak-Ublick GA, et al. Effects of proinflammatory

cytokines on rat organic anion transporters during toxic liver injury and cholestasis. Hepatology 2003

Aug;38(2):345-354.

(10) Rost D, Herrmann T, Sauer P, Schmidts HL, Stieger B, Meier PJ, et al. Regulation of rat organic anion

transporters in bile salt-induced cholestatic hepatitis: effect of ursodeoxycholate. Hepatology 2003

Jul;38(1 ):187-195.

(1 1 ) Lieser MJ, Park J, Natori S, Jones BA, Bronk SF, Gores GJ. Cholestasis confers resistance to the rat

liver mitochondrial permeability transition. Gastroenterology 1 998 Sep:115(3):693-701.

(12) Greim H, Trulzsch D, Czygan P, Rudick J, Hutterer F, Schaffner F, et al. Mechanism of cholestasis. 6 .

Bile acids in human livers with or without biliary obstruction. Gastroenterology 1972 Nov;63(5):846-850. (13) Reed JC. Mechanisms of apoptosis. Am J Pathol 2000 Nov:157(5):1415-1430.

(14) Adams JM. Ways of dying: multiple pathways to apoptosis. Genes Dev 2003 Oct 15;17(20):2481-2495.

(15) Adams JM, Cory S. Apoptosomes: engines for caspase activation. Curr Opin Cell Biol 2002 Dec; 14(6):715-720.

85

Chapter 6

86

(16) Cohen GM. Caspases: the executioners of apoptosis. Biochem J 1 997 Aug 15;326(Pt 1 ) : 1-16 .

(17) Kim JS, He L, Lemasters JJ . Mitochondrial permeability transition: a common pathway to necrosis and

apoptosis. Biochem Blophys Res Commun 2003 May 9;304(3):463-470.

(18) Lemasters JJ. Mechanisms of hepatic toxicity - V. Necrapoptosis and the mitochondrial permeabil ity

transition: shared pathways to necrosis and apoptosis. American Journal of Physiology-Gastrointestinal

and Liver Physiology 1 999 Jan;276(1 ):G1-G6.

( 19) Lemasters JJ, Qian T, He LH, Kim JS, Elmore SP, Cascio WE, et al. Role of mitochondrial inner

membrane permeabilizat1on in necrotic cell death, apoptosis, and autophagy. Antioxidants & Redox

Signaling 2002 Oct;4(5):769-781 .

(20) Paumgartner G, Beuers U . Ursodeoxycholic acid I n cholestatic liver disease: mechanisms of action and

therapeutic use revisited . Hepatology 2002 Sep:36(3):525-531 .

(21 ) Beuers U, Bilzer M, Chittattu A, Kullak-Ublick GA, Keppler D, Paumgartner G, et al.

Tauroursodeoxycholic acid inserts the apical conjugate export pump, Mrp2, into canalicular membranes

and stimulates organic anion secretion by protein kinase C-dependent mechanisms in cholestatic rat

liver. Hepatology 2001 May:33(5): 1 206-1 216.

(22) Qiao L, Yacoub A, Studer E, Gupta S, Pei XY, Grant S, et al. Inhibition of the MAPK and Pl3K pathways

enhances UDCA-induced apoptosis in primary rodent hepatocytes. Hepatology 2002 Apr;35(4):779-789.

(23) Rodrigues CM, Fan G, Ma X. Kren BT, Steer CJ. A novel role for ursodeoxychollc acid in inhibiting

apoptosis by modulating mitochondrial membrane perturbation. J Clin Invest 1 998 Jun 1 5; 1 01(12):2790-

2799.

(24) Schoemaker MH, Conde de la Rosa L, Buist-Homan M, Vrenken TE, Havinga R, Poelstra K, et al.

Tauroursodeoxycholic acid protects rat hepatocytes from bile acid-Induced apoptosis via activation of

survival pathways. Hepatology 2004 Jun;39(6):1 563-1 573.

(25) Marzioni M, Francis H, Benedetti A, Ueno Y, Fava G, Venter J, et al. Ca2+-dependent cytoprotective

effects of ursodeoxycholic and tauroursodeoxycholic acid on the biliary epithelium in a rat model of

cholestasis and loss of bile ducts American Journal of Pathology 2006 Feb; 1 68(2):398-409.

(26) Canbay A, Gu1cciardi ME, Higuchi H, Feldstein A, Bronk SF, Rydzewski R, et al. Cathepsin B

inactivation attenuates hepatic injury and fibrosis during cholestasis. J Clin Invest 2003 Jul: 1 1 2(2):1 52-

1 59.

(27) Osawa Y, Hannun YA, Proia RL, Brenner DA. Roles of AKT and sphingosine kinase In the antiapoptotic

effects of bile duct ligation in mouse liver. Hepatology 2005 Dec;42(6): 1 320-1328.

(28) Georgiev P, Navanni AA, Eloranta JJ , Lang KS, Kullak-Ublick GA, Nocito A, et al. Cholestasis protects

the liver from ischaemic injury and post-ischaemic inflammation in the mouse. Gut 2007 Jan:56(1) : 121-

1 28.

(29) Beuers U, Nathanson MH, Isales CM, Boyer JL. Tauroursodeoxycholic acid stimulates hepatocellular

exocytosis and mobilizes extracellular Ca++ mechanisms defective In cholestasis. J Chn Invest 1 993

Dec:92(6):2984-2993.

(30) Alpini G. Ursodeoxycholate and tauroursodeoxycholate inhibit cholang1ocyte growth and secretion of

BDL rats through activation of PKC alpha (vol 35, pg 1 041 , 2002). Hepatology 2002 Jul;36(1 ):265.

(31 ) Li ZD, Mizuno S, Nakamura T. Antinecrotic and antiapoptotic effects of hepatocyte growth factor on

cholestatic hepatitis in a mouse model of bile-obstructive diseases. American Journal of Physiology­

Gastrointestinal and Liver Physiology 2007 Feb;292(2):G639-G646.

(32) Black D, Bird MA, Samson CA, Lyman S, Lange PA, Schrum LW, et al. Primary cirrhotic hepatocytes

resist TGF beta-induced apoptosis through a ROS-dependent mechanism. Journal of Hepatology 2004

Jun:40(6):942-951 .

(33) Schoemaker MH, Gommans WM, Conde d e la Rosa L , Homan M, Kick P , Trautwein C, et al.

Resistance of rat hepatocytes against bile acid-induced apoptosis 1n cholestatic liver inJury is due to

nuclear factor-kappa B activation. J Hepatol 2003 Aug,39(2) :1 53-161 .

(34) Conde de la Rosa L, Schoemaker MH, Vrenken TE, Buist-Homan M, Havinga R, Jansen PL, et al .

Superoxide anions and hydrogen peroxide induce hepatocyte death by different mechanisms:

involvement of JNK and E RK MAP kinases J Hepatol 2006 May;44(5):91 8-929.

(35) Blokzijl H, Vander Borghi S, Bok LI, Libbrecht L, Geuken M, van den Heuvel FA, et a l . Decreased P­

glycoprotein (P-gp/MDR1 ) expression in inflamed human intestinal epithelium is independent of PXR

protein levels. lnflamm Bowel Dis 2007 Jun ;13(6):71 0-720.

Chapter 7

Summary and General D iscussion

Chapter 7

In many liver diseases, hepatocyte damage occurs upon exposure to toxic bile acids, inflammatory cytokines and increased levels of reactive oxygen species (ROS). Detailed information of the signaling pathways involved in hepatocyte damage will facilitate the discovery of new targets for intervention. Therefore, in this thesis the modulation of signaling pathways involved in three models of hepatocyte cell death was studied: 1) glycochenodeoxycholic acid (GCDCA), as a model for acute bile acid toxicity, 2) the cytokine tumor necrosis factor (TNF) a in combination with the transcriptional inhibitor actinomycin D (ActD, to inhibit nuclear factor-KB (NF-KB) activation), as a model for acute liver failure and 3) the superoxide anion donor menadione as a model for oxidative stress-induced apoptosis. These models are clinically relevant, because bile acids, cytokines and reactive oxygen species are present at increased levels in most liver diseases. In addition to a detailed investigation of the mechanisms of cell death in these models, the effects of potential drugs for the treatment of liver diseases were studied.

Chapter 2 describes the hepatoprotective mechanism of metformin against bile acid-induced apoptosis. Metformin, a drug used in the treatment of Diabetes Mellitus type II disease, was shown to improve liver function (lower plasma levels for alanine aminotransferase {ALAT) and aspartate aminotransferase (ASAT)) in patients and animal models with non-alcoholic fatty liver disease (NAFLD; 1-5). Furthermore, metformin and its downstream target AMP­activated protein kinase (AMPK) have been implicated in both anti-apoptotic (6, 7) and pro­apoptotic (8, 9) actions. Considering this controversy, we examined the effects of metformin in bile acid- and cytokine-induced cell death in primary rat hepatocytes. We demonstrate a protective effect of metformin against bile acid-induced apoptosis (caspase-3 activation and nuclear fragmentation). Our results are in contrast to previous reports describing pro­apoptotic effects of activated AMPK in liver cells In contrast, anti-apoptotic effects of activated AMPK were observed in astrocytes and endothelial cells (10). Apparently, the protective effect of metformin is restricted to specific death stimuli and cell types. Previously, we have demonstrated that metformin protects against superoxide anion-induced apoptosis (Conde de la Rosa L et al: unpublished data), while metformin does not protect against TNFa/ActD-mediated apoptosis. Ark5, another AMPK family member, did protect against Fas-induced apoptosis in colon carcinoma cells (6). This selectivity of metformin is probably due to the use of different signaling pathways by these death stimuli. Similar to Ark5 (6), the hepatoprotective effect of metformin requires an intact phosphoinositide-3 kinase (Pl3K)/Akt signaling pathway. The important role of Pl3K/Akt signaling In cell survival has been demonstrated before in models of apoptosis induced by UV radiation, DNA damage and cytokines (11 ). In addition, inhibition of the Pl3K/Akt pathway completely abolished the protective effect of tauroursodeoxycholic acid (TUDCA) against GCDCA-induced apoptosis (12). The metformin-mediated increased expression of Bel-xi might also be regulated via the Pl3K/Akt pathway. Collectively, these data identify the Pl3K/Akt pathway as a key player in the regulation of cell survival and hence, an important target for therapeutic intervention.

Although many papers (13-18) reported downstream effects of metformin on the AMPK/mammalian target of rapamycin (mTOR) pathway, in our study, this signaling pathway appears to be not involved in the anti-apoptotic actions of metformin. In line with this, we observed that metformin only moderately increased AMPK activity, compared to 5-aminoimidazole-4-carboxamide 1-13-D-ribofuranoside (AICAR), a known potent activator of AMPK.

Important in our study is the observation that metformin does not inhibit NF-KB activity. Again, our data are in contrast to several reports describing inhibition of NF-KB by metformin via AMPK activation (19, 20). However, these studies also demonstrated the involvement of AMPK in metformin signaling, while in our experiments the anti-apoptotic actions of metformin are independent of AMPK activity. Since metformin does not affect NF-KB activity, and does not sensitize hepatocytes to TNFa-induced cell death, metformin might be a promising drug for the treatment of chronic liver diseases accompanied by inflammation 88

Summary and General Discussion

(TNFa). In vivo studies, validating the observed in vitro effects of metformin in models without insulin-sensitivity, e.g. bile duct ligation (BDL) would help to elucidate the hepatoprotective mechanism of metformin and its possible therapeutic applications.

In chapter 3 of this thesis the protective effect of leflunomide and its active metabolite (A77 1 726) against bile acid-induced cell death of hepatocytes is described, and we speculate about the mechanisms involved in this hepatoprotective effect. Leflunomide is used in the treatment of autoimmune diseases, particularly rheumatoid arthritis, because of its anti­inflammatory, anti-proliferative and immunosuppressive properties (21 -26). In vivo and in vitro studies have demonstrated beneficial effects of leflunomide in several models of liver damage, including acetaminophen (21 , 27), concanavalin A (22), bile duct ligation (28) and ischemia/reperfusion (1/R) injury (23). In contrast, pro-apoptotic effects of leflunomide have been reported in rat thymocytes (29) and breast cancer cells (30). We describe the reduction in bile acid-induced apoptosis (caspase-3 and -6 activation, nuclear condensation) and necrosis by A77 1 726, while the prodrug leflunomide - used at similar concentrations - had no effect. Thus, we confirm previous observations (21 ) demonstrating a 1 00-fold higher activity of A 77 1 726 than leflunomide. Although hepatocytes can convert the inactive prodrug into the active metabolite, the short duration of the experiments probably precluded the generation of sufficient amounts of the active metabolite. A77 1 726 was even protective when added after the bile acid, a mechanism described before for TUDCA ( 1 2) and indicating a rapid activation of survival pathways by leflunomide. However, our results do not support a key role for the anti-apoptotic mitogen-activated protein kinases (MAPK) p38 and extracellular signal regulated kinase (ERK) 1 /2 or the Pl3K/Akt survival pathway in the hepatoprotective actions of A77 1726. These observations are in contrast to other studies, describing inhibitory effects of A77 1 726 on p38 activation (22, 3 1 , 32) and stimulation of ERK1 /2 and Akt activity (32). Furthermore, we did not observe an inhibition of the pro­apoptotic MAPK member, JNK, by A77 1 726 as has been reported by others (21 , 26, 27, 3 1 , 32).

Finally, we show that NF-KB activity is not affected by A77 1 726, which is in line with some studies (31 , 32), but in contrast to others (22, 33, 34). These conflicting results appear to be cell type specific. Because A77 1 726 does not inhibit NF-KB activity and hepatocytes are not sensitized to TNFa-induced apoptosis by A77 1 726, our results imply that A77 1726 could be valuable as therapy for chronic cholestatic liver diseases in the presence of inflammation. Since none of the major signaling pathways involved in hepatocyte survival and death is implicated in the protective mechanism of leflunomide, we can only speculate about possible targets of leflunomide. Recently, members of the Src family of kinases were identified as important proteins in hepatocyte death (35, 36). Moreover, leflunomide was shown to block the activity of some specific Src family members in T-cells (34, 37, 38). In chapter 4, we describe a similar effect of a specific Src-inhibitor on bile acid-induced cell death as observed with A77 1 726 in this study. Therefore, we speculate that the Src kinases can be a potential target for A77 1 726. Additional experiments are required to study this hypothesis in more detail.

Although reports have described effects of leflunomide in acute and chronic liver injury (21 -23, 27, 28), in our study A77 1 726 was ineffective in the protection against menadione or TNFa/ActD-induced apoptosis of hepatocytes. Again, this could be related to the activation of different signaling pathways by these death stimuli.

Chapter 4 examines the role of reactive oxygen species in bile acid-induced cell death. Previous studies have reported contradictory information concerning the role of reactive oxygen species (ROS) in bile acid-induced hepatocyte death. The majority of studies have described increased ROS production in response to bile acid exposure (39-43). However, it is not clear whether bile acid-induced ROS generation contributes to hepatocyte cell death. Although ROS generation is diminished by antioxidants, the correlation with bile acid-

89

Chapter 7

induced cell injury is not always observed (39-41, 44-46). According to Reinehr et al (36, 4 7), bile acids activate the membrane-associated enzyme nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase, which generates ROS. Subsequently, ROS activate the Src kinase family member Yes. Activated Yes facilitates the activation of Fas, death inducing signaling complex (DISC) formation and apoptosis.

We demonstrate that ROS are not critical for bile acid induced cell death. Bile acid exposure neither depletes glutathione nor induces the oxidative stress responsive gene heme­oxygenase-1 (HO-1 ). Secondly, ROS scavengers and various antioxidants do not reduce bile acid-induced apoptosis. Finally, although inhibition of NADPH-oxidase reduces GCDCA­induced apoptosis, necrosis is strongly induced. In addition, the involvement of the Src kinase family member Yes in bile acid-induced cell death was confirmed using the Src­inhibitor SU6656.

Our results strongly argue against a role for reactive oxygen species in bile acid-induced hepatocyte apoptosis. Our observations are in agreement with a recent study (41) describing complete abrogation of GCDCA-induced apoptosis by caspase-8 inhibition, while ROS generation was only partially reduced. This indicates a weak or no correlation between ROS generation and cell death. A possible explanation for these contradictory results is that bile acids at concentrations below 100 µmol/L generate low levels of ROS, as has been shown in many studies, which are insufficient to induce oxidative stress accompanied by HO-1 induction and glutathione depletion. Thus ROS are not causally involved in apoptotic cell death. Higher concentrations of bile acids (above 200 µmol/L) might generate high levels of ROS, which are accompanied by oxidative stress and necrotic rather than apoptotic cell death.

Hence, future experiments should focus on the role of ROS in the toxicity of high concentrations of bile acids and the involvement of ROS generation in cholestatic liver diseases. These studies are also important to establish the value of antioxidant therapy in cholestatic liver diseases.

In chapter 5 we study the protective effects of HO-1 and carbon monoxide (CO) against oxidative stress-induced apoptosis. ROS are key players in many liver diseases like non­alcoholic steatohepatitis (NASH). We demonstrate the induction of HO-1 in an in vivo model of cholestatic liver injury, but not in acute inflammation. The lack of HO-1 induction in acute inflammation might be the result of sufficient levels of reduced glutathione. Indeed, the glutathione level controls the balance between the NF-KB (inflammation) and AP-1 (oxidative stress) regulated stress-response (48-50). Glutathione depletion in the model of acute inflammation leads to a shift from the NF-KB response (iNOS induction) to the AP-1 response (HO-1 induction). Superoxide anions, but not hydrogen peroxide, induce HO-1 induction and overexpression of HO-1 protects hepatocytes against superoxide anion­induced apoptosis. The anti-apoptotic action of HO-1 is, at least in part, due to one of the products of HO-1, CO. These observations are in line with previous studies describing cytoprotective effects of CO both in hepatocytes (51, 52) and in other cell types (50, 53, 54), although pro-apoptotic actions of CO have also been observed (55, 56). In our study, the anti-apoptotic effects of CO are the result of direct inhibition of the pro-apoptotic MAPK member JNK, similar as described in a model of sepsis (57). CO did not modulate NF-KB activation, as has been described before (52). Our data demonstrate protection of HO-1 and one of its products, CO, against superoxide anion-induced apoptosis. These compounds, as well as other HO-1 products (e.g. biliverdin), might be potential therapeutic targets in oxidative stress-related liver injury.

In chapter 6 a novel approach to study cholestasis in vitro is described. Most patients with chronic liver injury, including cholestatic liver diseases, start treatment when the disease becomes clinically manifest and is already well advanced. Therefore any therapeutic 90

Summary and General Discussion

intervention should be aimed at 'diseased' (e.g. cholestatic) hepatocytes and not normal 'healthy' hepatocytes. Likewise, mechanisms of cell death should be studied in 'diseased' hepatocytes. Thus, studying the signaling pathways involved in hepatocyte death in cholestatic liver injury is crucial. Only a few papers described the use of primary hepatocytes isolated from cholestatic rats (58, 59). In this study, we compare the phenotype of primary hepatocytes isolated from bile duct ligated rats and sham-operated rats. The cholestatic phenotype observed in vivo after BDL, i.e. decreased expression of Na+/taurocholate cotransporting polypeptide (Ntcp) the basolateral transporter involved in the uptake of bile acids and multidrug resistance-associated protein (Mrp) 2, responsible for the bile acid secretion into bile at the canalicular membrane and increased expression of Mrp3, which secretes bile acids and organic anions from the hepatocytes into the sinusoids (reverse transport), is maintained in hepatocytes both directly after isolation and during the time of culture. The effects on transporters in our primary cholestatic hepatocytes are in agreement with previous reports (60-63) describing similar transporter regulation in patients and animals with cholestatic disorders. Hence, we demonstrate that hepatocytes retain the cholestatic phenotype during the timeframe of our experiments. Furthermore, our preliminary results suggest that cholestatic hepatocytes are less sensitive to bile acid-induced apoptosis. Although cholestatic control hepatocytes demonstrate increased levels of apoptosis and necrosis, no additional induction of cell death by bile acids was observed. Increased resistance of cholestatic and cirrhotic livers to apoptosis induced by different stimuli, including 1/R injury, TNFa and transforming growth factor (TGF) 13 has been described (64, 65) and is in accordance with our observations. A possible explanation for the observed resistance to bile acid-induced apoptosis is the diminished uptake of conjugated bile acids into the cholestatic hepatocyte due to the absence of Ntcp at the plasma membrane.

Our preliminary data demonstrate the validity of our in vitro system to study cholestasis in primary hepatocytes. Future experiments will focus on the signaling mechanisms in cholestatic hepatocytes and the response of these cells to therapeutic drugs.

Conclusions New therapies for liver diseases can only be discovered via firm knowledge of the signaling mechanisms contributing to the death and survival of hepatocytes during liver damage. This thesis describes the protective mechanism of metformin and leflunomide against bile acid­induced cell death. Our analysis and modulation of pathways involved in bile acid signaling has provided more insight in the role of oxidative stress in bile acid-induced apoptosis. In addition, we have examined a novel approach to study signaling pathways in cholestatic hepatocytes in vitro. In the future, this new technique will provide a better understanding of hepatocyte behavior during liver injury.

91

Chapter ?

Reference List

( 1 )

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

( 10)

(1 1 )

( 12)

( 1 3)

(14)

(1 5)

(16)

(17)

( 1 8)

(19)

(20)

92

Lin HZ, Yang SQ, Chuckaree C, Kuhajda F, Ronne! G, Diehl AM. Metforrnin reverses fatty liver disease

in obese, leptin-deficient mice. Nature Medicine 2000 Sep:6(9):998-1003.

Klein J, Westphal S, Kraus D, Meier B, Perwitz N, Ott V, et al. Metformin inhibits leptin secretion via a

mitogen-activated protein kinase signalling pathway in brown adipocytes. J Endocrinol 2004

Nov;1 83(2)·299-307.

Angelico F, Burattln M, Alessandri C, Del Ben M, Lirussi F. Drugs improving insulin resistance for non­

alcoholic fatty liver disease and/or non-alcoholic steatohepalitis. Cochrane Database of Systematic

Reviews 2007;(1 ).

Bugianesi E, Marzocchi R, Villanova N, Marchesinl G. Non-alcoholic fatty liver disease/non-alcoholic

steatohepatitis (NAFLD/NASH): treatment. Best Practice & Research in C linical Gastroenterology 2004

Dec;1 8(6):1 1 05-1 1 1 6.

Angulo P, Linder KD. Treatment of non-alcoholic steatohepatitis. Best Practice & Research in Clinical

Gastroenterology 2002 Oct; 1 6(5):797-810.

Suzuki A, Kusakai G, Kishimoto A, Shimojo Y, Miyamoto S, Ogura T, et a l . Regulation of caspase-6 and

FLIP by the AMPK family member ARK5. Oncogene 2004 Sep 1 6;23(42):7067-7075.

Suzuki A, Kusakai G, Kishimoto A, Lu J, Ogura T, Esumi H. ARK5 suppresses the cell death induced by

nutnent starvation and death receptors via inhibition of caspase 8 activation, but not by

chemotherapeutic agents or UV irradiation. Oncogene 2003 Sep 1 8;22(40):61 77-6182.

Campas C, Lopez JM, Santidrian AF, Barragan M, Bellosillo B, Colomer D, et al. Acadeslne activates

AMPK and induces apoptosis in 8-cell chronic lymphocytic leukemia cells but not in T lymphocytes.

Blood 2003 May 1 ; 1 01 (9):3674-3680.

Meisse D, Van de Casteele M, Beauloye C. Hainault I, Kefas BA, Rider MH, et al. Sustained activation

of AMP-activated protein kinase induces c-Jun N-terrninal kinase activation and apoptosis in liver cells.

FEBS Lett 2002 Aug 28;526(1 -3):38-42.

Luo Z, Saha AK, Xiang X, Ruderman NB. AMPK, the metabolic syndrome and cancer. Trends

Pharmacol Sci 2005 Feb;26(2):69-76.

Datta SR, Brunet A, Greenberg ME. Cellular survival. a play in three Akts. Genes Dev 1 999 Nov

1 5;1 3(22):2905-2927.

Schoemaker MH, Conde de la Rosa L, Buist-Homan M, Vrenken TE, Havinga R, Poelstra K, et al.

Tauroursodeoxycholic acid protects rat hepatocytes from bile acid-induced apoptosis via activation of

survival pathways. Hepatology 2004 Jun:39(6). 1 563-1573.

Zang M, Zuccollo A, Hou X, Nagata D, Walsh K, Herscovitz H, et al . AMP-activated protein kinase is

required for the l ipid-lowering effect of metformin in insulin-resistant human HepG2 cells J Biol Chem

2004 Nov 1 2 ;279(46):47898-47905.

Reiter AK, Bolster DR, Crozier SJ, Kimball SR, Jefferson LS. Repression of protein synthesis and

mTOR signaling in rat hver mediated by the AMPK activator aminoimidazo!e carboxamide

ribonucleoside. Am J Physiol Endocrinol Metab 2005 May;288(5):E980-E988.

Kimura N, Tokunaga C, Dalal S, Richardson C, Yoshino K, Hara K, et al. A possible linkage between

AMP-activated protein kinase (AMPK) and mammalian target of rapamycin (mTOR) signalling pathway.

Genes Cells 2003 Jan;8(1 ):65-79.

Yin W, Mu J, Birnbaum MJ. Role of AMP-activated protein kinase In cyclic AMP-dependent l ipoiysis 1n

3T3-L 1 adipocytes. Journal of Biological Chemistry 2003 Oct 31 ,278(44):43074-43080.

Hahn-Windgassen A, Nogueira V, Chen CC, Skeen JE. Sonenberg N, Hay N. Aki activates the

mammalian target of rapamycin by regulating cellular ATP level and AMPK activity. Journal of B iological

Chemistry 2005 Sep 1 6;280(37):32081 -32089.

Zhou GC, Myers R, Li Y, Chen YL, Shen XL, Fenyk-Melody J, et al. Role of AMP-activated protein

kinase in mechanism of metformin action. Journal of Clinical Investigation 2001 Oct;108(8): 1 1 67-1 1 74.

Hattori Y , Suzuki K, Hattori S, Kasai K. Metformln inhiblts cytokine-induced nuclear factor kappaB

activation via AMP-activated protein kinase activation in vascular endothelial cells. Hypertension 2006

Jun,47(6): 1 1 83-1 1 88.

isoda K, Young JL, Zirlik A, MacFarlane LA, Tsuboi N, Gerdes N, et al. Metformin inhibits

proinOammatory responses and nuclear factor-kappaB in human vascular wall cells. Arterioscler Thromb

Vase Biol 2006 Mar;26(3):61 1 -617.

Summary and General Discussion

(21) Latchoumycandane C, Seah QM, Tan RCH, Sattabongkot J, Beerheide W, Boelsterli UA. Leflunomide

or A77 1726 protect from acetaminophen-induced cell injury through inhibition of JNK-mediated

mitochondrial permeability transition in immortalized human hepatocytes. Toxicology and Applied

Pharmacology 2006 Nov 15;217(1 ): 1 25-133.

(22) lmose M, Nagaki M, Kimura K, Takai S, lmao M, Naiki T, et al. Leflunomide protects from T-cell­

mediated liver injury in mice through inhibition of nuclear factor kappaB. Hepatology 2004

Nov;40(5): 1160-1169.

(23) Karaman A, Fadillioglu E, Turkmen E, Tas E, Yilmaz Z. Protective effects of leflunomide against

ischemia-reperfusion injury of the rat liver. Pediatr Surg Int 2006 May:22(5):428-434.

(24) Yao HW, Li J, Chen JQ, Xu SY. Leflunomide attenuates hepatocyte injury by inhibiting Kupffer cells.

World J Gastroenterol 2004 Jun 1 ;1 0(1 1 ):1608-1611.

(25) Alexander D, Friedrich B, Abruzzese T, Gondolph-Zink B, Wulker N, Aicher WK. The active form of

leflunomide, HMR1726, facilitates TNF-alpha and IL-17 induced MMP-1 and MMP-3 expression. Cell Physiol Biochem 2006;1 7(1 -2):69-78.

(26) Vergne-Salle P, Leger DY, Bertin P, Treves R, Beneytout JL, Liagre B. Effects of the active metabolite

of leflunomide, A77 1726, on cytokine release and the MAPK signalling pathway in human rheumatoid

arthritis synoviocytes. Cytokine 2005 Sep 7;31 (5):335-348.

(27) Latchoumycandane C, Goh CW, Ong MM, Boelsterli UA. Mitochondrial protection by the JNK inhibitor

leflunomide rescues mice from acetaminophen-induced liver injury. Hepatology 2007 Feb;45(2):412-

421.

(28) Karaman A, lraz M, Kirimlioglu H, Karadag N, Tas E, Fadillioglu E. Hepatic damage in biliary-obstructed

rats is ameliorated by leflunomide treatment. Pediatr Surg Int 2006 Sep:22(9):701 -708.

(29) Colic M, Popovic P, Vucevic D, Dimitrijevic M. Leflunomide induces apoptosis of thymocytes and T-cell

hybridoma: differences in sensitivity and signaling pathways. Transplant Proc 2001 May;33(3):2344-2346.

(30) Ghosh S, Zheng Y, Jun X, Narla RK, Mahajan S, Navara C, et al. Alpha-cyano-beta-hydroxy-beta­

methyl-N-(4-(trifluoromethoxy)phenyl) propenamide: an inhibitor of the epidermal growth factor receptor

tyrosine kinase with potent cytotoxic activity against breast cancer cells. Clin Cancer Res 1 998

Nov;4(1 1 ):2657-2668.

(31) Migita K, Miyashita T, Maeda Y, Nakamura M, Yatsuhashi H, Ishibashi H, et al. An active metabolite of

leflunomide, A77 1726, inhibits the production of serum amyloid A protein in human hepatocytes.

Rheumatology 2005 Apr;44(4):443-448.

(32) Leger DY, Liagre B, Beneytout JL. Low dose leflunomide activates PI3K/Akt signalling in

erythroleukemia cells and reduces apoptosis induced by anticancer agents. Apoptosis 2006

Oct;11(10):1 747-1760.

(33) Manna SK, Mukhopadhyay A, Aggarwal BB. Leflunomide suppresses TNF-induced cellular responses:

Effects on NF-kappa B, activator protein-1 , c-Jun N-terminal protein kinase, and apoptosis. Journal of

Immunology 2000 Nov 1 5;165(10):5962-5969.

(34) Manna SK, Aggarwal BB. lmmunosuppressive leflunomide metabolite (A77 1726) blocks TNF­

dependent nuclear factor-kappa B activation and gene expression. Journal of Immunology 1 999 Feb

15;162(4 ):2095-21 02.

(35) Reinehr R, Becker S, Hengen A, Haussinger D. The Src family kinase Yes triggers hyperosmotic

activation of the epidermal growth factor receptor and CD95. J Biol Chem 2004 Jun 4;279(23):23977-

23987.

(36) Reinehr R, Becker S, Wettstein M, Haussinger D. Involvement of the Src family kinase yes in bile salt­

induced apoptosis. Gastroenterology 2004 Nov;127(5):1 540-1557.

(37) Layseca-Espinosa E, Pedraza-Alva G, Montiel JL, del RR, Fierro NA, Gonzalez-Amaro R, et al. T cell

aggregation induced through CD43: intracellular signals and inhibition by the immunomodulatory drug

leflunomide. J Leukoc Biol 2003 Dec:74(6):1 083-1093.

(38) Xu X, Williams JW, Bremer EG, Finnegan A, Chong AS. Inhibition of protein tyrosine phosphorylation in

T cells by a novel immunosuppressive agent, leflunomide. J Biol Chem 1 995 May 26;270(21 ):12398-12403.

(39) Reinehr R, Graf D, Haussinger D. Bile salt-induced hepatocyte apoptosis involves epidermal growth

factor receptor-dependent CD95 tyrosine phosphorylation. Gastroenterology 2003 Sep;125(3):839-853.

93

Chapter ?

94

(40) Sokol RJ, Straka MS, Dahl R, Devereaux MW. Yerushalmi B, Gumpricht E. et al . Role of oxidant stress

in the permeability transition induced in rat hepatic mitochondria by hydrophobic bile acids. Pedlatr Res

2001 Apr;49(4):51 9-531 .

(41 ) Yerushalmi B . Dahl R , Devereaux MW, Gumpricht E , Sokol RJ. Bile ac1d-1nduced rat hepatocyte

(42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

(50)

(51 )

(52)

(53)

(54)

(55)

(56)

(57)

(58)

(59)

(60)

apoptosis is inhibited by antioxidants and blockers of the mitochondrial permeability transition.

Hepatology 2001 Mar;33(3):61 6-626.

Sokol T J, Dahl R, Devereaux MW, Yerushalmi B, Kobak GE, Gumpricht E. Human hepatic mitochondria

generate reactive oxygen species and undergo the permeability transition in response to hydrophobic

bile acids. Journal of Pediatric Gastroenterology and Nutrition 2005 Aug;41 (2):235-243.

Gumpricht E, Dahl R, Devereaux MW, Sokol RJ. beta-carotene prevents bile acid-induced cytotoxicity in

the rat hepatocyte: Evidence for an antioxidant and antl-apoptotic role of beta-carotene In vitro. Pediatric

Research 2004 May;55(5):81 4-82 1 .

Borgognone M , Perez LM, Baslglio CL, Ochoa J E , Roma MG. Signal ing modulation of bile salt-induced

necrosis in isolated rat hepatocytes. Toxicological Sciences 2005 Jan;83(1 ):1 14-1 25.

Socha P, Koletzko B, Pawlowska J, Proszynska K, Socha J . Treatment of cholestatlc children with

water-soluble vitamin E (alpha-tocopheryl polyethylene glycol succlnate): effects on serum vitamin E ,

lipid peroxides, and polyunsaturated fatty acids. J Pediatr Gastroenterol Nutr 1997 Feb;24(2) :189-193.

Watson JP, Jones DE, James OF, Cann PA, Bramble MG. Case report: oral antioxidant therapy for the

treatment of primary biliary cirrhosis: a pilot study. J Gastroenterol Hepatol 1 999 Oct:14(1 0):1 034-1040.

Reinehr R. Becker S, Keitel V, Eberle A, Grether-Beck S, Haussinger D. Bile salt-induced apoptosis

involves NADPH oxidase isoform activation. Gastroenterology 2005 Dec:1 29(6):2009-2031 .

Vos TA, Van GH, Tuyt L, De Jager-Krikken A, Leuvenink R, Kuipers F, et al. Expression of inducible

nitric oxide synthase in endotoxemic rat hepatocytes is dependent on the cellular g lutathione status.

Hepatology 1 999 Feb;29(2):421-426.

D ijkstra G, Blokzijl H , Bok L, Homan M. Van GH, Faber KN, et al. Opposite effect of oxidative stress on

inducible nitric oxide synthase and haem oxygenase-1 expression in intestinal inflammation: anti­

inflammatory effect of carbon monoxide. J Pathol 2004 Nov:204(3):296-303.

Zhang X, Shan P, Otterbein LE, Alam J, Flavell RA, Davis RJ, et al. Carbon monoxide inhibition of

apoptosis during ischemia-reperfusion lung injury is dependent on the p38 mltogen-activated protein

kinase pathway and involves caspase 3. J Biol Chem 2003 Jan 1 0;278(2):1 248-1 258.

Choi BM, Pae HO, Kim YM, Chung HT. Nitric oxide-mediated cytoprotection of hepatocytes from

glucose deprivation-induced cytotoxicity: involvement of heme oxygenase-1 . Hepatology 2003

Apr;37(4):81 0-823.

Zuckerbraun BS, Billiar TR, Otterbein SL, Kim PK, Liu F, Choi AM, et al . Carbon monoxide protects

against liver failure through nitric oxide-induced heme oxygenase 1 . J Exp Med 2003 Dec

1 ; 1 98(1 1 ):1 707-1 716.

Liu XM, Chapman GB, Peyton KJ, Schafer Al , Durante W. Carbon monoxide inhibits apoptosis in

vascular smooth muscle cells. Cardiovasc Res 2002 Aug 1 ;55(2):396-405.

Brouard S, Otterbein LE, Anrather J, Tobiasch E, Bach FH, Choi AM, et al. Carbon monoxide generated

by heme oxygenase 1 suppresses endothelial cell apoptosis. J Exp Med 2000 Oct 2,1 92(7): 1 0 1 5-1 026.

Thom SR, Fisher D, Xu YA, Notarfrancesco K, lschiropoulos H. Adaptive responses and apoptosis in

endothelial cells exposed to carbon monoxide. Proc Natl Acad Sci U S A 2000 Feb 1 ;97(3 ):1 305-1310.

Turcanu V, Dhouib M, Gendrault JL, Poindron P. Carbon monoxide induces murine thymocyte apoptosis

by a free radical-mediated mechanism. Ceil Biol Toxicol 1 998 Feb ; 14(1 ):47-54.

Morse D, Pischke SE, Zhou Z, Davis RJ, Flavell RA, Loop T, et al. Suppression of inflammatory cytokine

production by carbon monoxide involves the JNK pathway and AP-1 . J Biol Chem 2003 Sep

26;278(39):36993-36998.

Lieser MJ, Park J, Natori S, Jones BA, Bronk SF, Gores GJ. Cholestasis confers resistance to the rat

liver mitochondrial permeability transition. Gastroenterology 1 998 Sep; 1 1 5(3):693-701 .

Beuers U , Nathanson MH, Isales CM, Boyer JL. Tauroursodeoxycholic acid stimulates hepatocellular

exocytosis and mobilizes extracellular Ca++ mechanisms defective in cholestasis. J Clin Invest 1 993

Dec;92(6):2984-2993.

Geier A, Dietrich CG, Voigt S, Kim SK, Gerloff T, Kullak-Ublick GA, et al. Effects of proinflammatory

cytokines on rat organic anion transporters during toxic liver injury and cholestasis. Hepatology 2003

Aug;38(2):345-354.

Summary and General Discussion

(61 ) Rost D, Herrmann T, Sauer P, Schmidts HL, Stieger B, Meier PJ, et al. Regulation of rat organic anion

transporters in bile salt-induced cholestatic hepatitis: effect of ursodeoxycholate. Hepatology 2003

Jul;38(1 ):187-195.

(62) Kullak-Ublick GA, Stieger B, Meier PJ. Enterohepatic bile salt transporters in normal physiology and liver

disease. Gastroenterology 2004 Jan;126(1 ):322-342. (63) Alrefai WA, Gill RK. Bile acid transporters: structure, function, regulation and pathophysiological

implications. Pharm Res 2007 Oct;24(10):1803-1 823.

(64) Black D, Bird MA, Samson CA, Lyman S, Lange PA, Schrum LW, et al. Primary cirrhotic hepatocytes

resist TGF beta-induced apoptosis through a ROS-dependent mechanism. Journal of Hepatology 2004

Jun;40(6):942-951. (65) Osawa Y, Hannun YA, Proia RL, Brenner DA. Roles of AKT and sphingosine kinase in the antiapoptotic

effects of bile duct ligation in mouse liver. Hepatology 2005 Dec;42(6):1 320-1328.

95

Chapter 8

Nederlandse Samenvatting

Chapter B

lntroductie De lever is verantwoordelijk voor verschillende functies, waaronder het metabolisme van (toxische) stoffen en geneesmiddelen, de productie en secretie van gal, de synthese van eiwitten en vetten, het metabolisme van koolhydraten (b.v. gluconeogenese) en de opslag van metabolieten. De meeste functies warden uitgevoerd door de parenchymcellen of hepatocyten. Hepatocyten hebben onderling contact nodig voor het optimaal vervullen van hun taken, en zijn daarom gerangschikt in zogeheten platen. Naast de hepatocyten vinden we in de lever endotheelcellen, die de sinusorden vormen, Kupffer cellen en stellaatcellen. Endotheelcellen zijn betrokken bij de immuunresponse en de uitwisseling van stoffen tussen bloed en lever. De Kupffer cellen zijn leverspecifieke macrofagen, die deel uitmaken van het afweersysteem en de stellaatcellen zorgen voor de opslag van vitamine A en voor de productie van extracel lulaire matrix. Tussen de hepatocyten bevinden zich de galkanaaltjes. Deze vertakken zich via de kanalen van Hering en eindigen in de galgangen. Dit netwerk van vertakkende galkanalen heet het galgangstelsel of galwegsysteem.

Gal bestaat voor het grootste deel uit water. Daarnaast bevat gal galzouten, fosfolipiden, cholesterol, glutathion (conjugaten) en metabolieten. Galzouten warden gesynthetiseerd in de lever, waarna actieve secretie van galzouten in de galkanaaltjes plaatsvindt. In de lever warden galzouten geconjugeerd met glycine of taurine, zodat ze efficient getransporteerd kunnen warden in de enterohepatische kringloop. De meerderheid van galzouten (90%) wordt na passage in de dunne darm gereabsorbeerd in het terminale ileum. Het resterende deel gaat verloren via de faeces, hetgeen gecompenseerd wordt middels de novo synthese van galzouten (primaire galzouten) uit cholesterol in hepatocyten. Twee routes voor galzout synthese zijn bekend, de 'neutrale' of klassieke route (meest voorkomend) en de 'zure' of alternatieve weg. Beide routes resulteren in de productie van cholzuur en chenodeoxycholzuur en specifieke leden van de cytochrome P450 enzym familie zijn nodig voor de galzoutsynthese. In de darm warden de primaire galzouten omgezet in de secundaire galzouten deoxycholzuur en lithocholzuur door bacterien.

De enterohepatische kringloop is het permanente transport van galzouten tussen de lever en de darm. Galzouten hebben meerdere functies, cholesterol wordt verwijderd uit de lever via de de novo synthese van galzouten in hepatocyten en via de inbouw in micellen. In de darm zorgt gal voor de absorptie van vetten, cholesterol en vet-oplosbare vitamines. Daarnaast activeren galzouten verschillende signaal transductie routes in de darmen (en lever). De enterohepatische kringloop is dan oak een belangrijk proces voor de vorming van gal en de homeostase van gal, vetten en cholesterol. De enterohepatische kringloop begint bij de hepatocyten. Aan het basolaterale membraaan nemen hepatocyten galzouten op uit het bloed via de transport eiwitten Ntcp en Oatp's. Na intracellulair transport warden de galzouten aan het canaliculaire membraan in de gal uitgescheiden via de transport eiwitten Bsep en Mrp2. Andere transport eiwitten zorgen ervoor dat cholesterol en fosfolipiden in de gal terechtkomen, waarna micellen uit galzouten, cholesterol en fosfolipiden warden gevormd. Na passage door het galgangstelsel, waar een deel van de galzouten geabsorbeerd wordt, vindt opslag van gal plaats in de galblaas. De galblaas contraheert in respons op voedsel, waardoor gal in de dunne darm wordt gepompt. In het ileum warden eveneens galzouten geabsorbeerd, die via verschillende transporters naar het bloed warden vervoerd. Via het bloed komen de galzouten terug in de lever waar ze weer over het basolaterale membraan van de hepatocyten getransporteerd warden door Ntcp en Oatp's.

Veel leveraandoeningen warden veroorzaakt door verstoringen in de enterohepatische kringloop, welke resulteren in celdood en leverschade. Twee vormen van celdood warden beschreven in dit proefschrift, apoptose en necrose. Apoptose of geprogrammeerde celdood is een energie afhankelijk proces, waarbij bepaalde enzymen, de caspases, betrokken zijn. Apoptotische cellen warden gekenmerkt door condenserende kernen en verschrompeling en fragmentatie van de eel tot apoptotische lichamen, waarbij de plasmamembranen intact blijven. Deze celfragmenten warden efficient verwijderd door fagocyterende cellen. Hierdoor treedt minimale ontsteking op. Caspases warden verdeeld in twee categorieen: caspases

98

Nederlandse Samenvattmg

die een rol spelen bij het 'starten' van de apoptotische route (caspases 8, 9 en 1 0) en caspases die betrokken zijn bij het eind van het apoptotische proces (caspases 3, 6 en 7). Er bestaan twee apoptotische routes, de extrensieke en de intrinsieke route. De extrensieke weg wordt ge"initieerd door de binding van een ligand aan receptoren op de celmembraan (b.v. tumor necrose factor, TNF) en de intrinsieke route wordt ge"induceerd door een stressprikkel in de eel (b.v. oxidatieve stress) en begint in een celorganel (b.v. mitochondrien, kern of endoplasmatisch reticulum). Seide routes leiden tot de activatie van caspases. Verschillende factoren zoals cytokines, oxidatieve stress en DNA schade kunnen apoptose induceren.

Necrose is een passieve vorm van celdood. Mitochondriele en cellulaire zwelling veroorzaakt lekkage van het plasma membraan, waarna de cellulaire inhoud vrijkomt in de circulatie. Hierdoor wordt een ontstekingsrespons uitgelokt. In tegenstelling tot apoptose is necrose een veel minder gestructureerde vorm van celdood. Bij de meeste leveraandoeningen wordt celschade gekenmerkt door een combinatie van apoptose en necrose.

Tijdens cellulaire stress neemt de productie van vrije zuurstof radicalen (o.a. waterstof peroxide en superoxide anionen) dusdanig toe, dat de eel niet meer in staat is om deze vrije zuurstof radicalen te ontgiften. Dit verschijnsel wordt oxidatieve stress genoemd. Condities die gepaard gaan met oxidatieve stress zijn o.a. veroudering, alcohol metabolisme en ontsteking. Hepatocyten beschikken over verschillende afweermechanismen tegen oxidatieve stress: de superoxide dismutase enzym familie, catalase, glutathion-peroxidases, gereduceerd glutathion en vitamines. Ook pro-oxidatie systemen zijn aanwezig zoals NADPH oxidase (generatie van superoxide anionen) en xanthine oxidase. Vrije zuurstof radicalen worden gegenereerd in zowel acute als chronische leveraandoeningen. Een voorbeeld van een leverziekte die gepaard gaat met oxidatieve stress is niet-alcoholische steatohepatitis (NASH).

Verschillende signaal transductie routes die betrokken zijn bij celdood en eel overleving zijn in dit proefschrift bestudeerd, waaronder fosfoinositide-3 kinase (Pl3K), mitogeen­geactiveerde proteine kinases (MAPK), de Src tyrosine kinase familie en de transcriptie factor NF-KB. De Pl3K route is een bekende overlevingsroute in cellen vooral via effecten op het prote"ine kinase B (PKB)/Akt. De bescherming van het anti-apoptotische galzout tauroursodeoxycholzuur (TUDCA) tegen glycochenodeoxycholzuur (GCDCA)-ge"induceerde apoptose bleek veroorzaakt te zijn door effecten van TUDCA op Pl3K. De MAPK familie is onderverdeeld in pro- en anti-apoptotische eiwitten. Het pro-apoptotische JNK is belangrijk voor de inductie van apoptose, terwijl ERK1 /2 en p38 bekend staan als anti-apoptotische eiwitten. Remming van JNK resulteert in vermindering van apoptose door galzouten en oxidatieve stress. Terwijl remming van ERK1 /2 en p38 zorgde voor een afname van het beschermend effect van het anti-apoptotische galzout TUDCA. Recent is aangetoond dat specifieke leden van de Src familie een belangrijke rol spelen bij celdood veroorzaakt door cytokines en galzouten. In dit proefschrift is de rol van het Src kinase familielid Yes beschreven in galzout-ge"induceerde celdood. De transcriptie factor NF-KB reguleert zowel pro- als anti-apoptotische genen. NF-KB-gereguleerde genen zijn onder andere cytokines, superoxide dismutases en anti-apoptotische leden van de Bcl-2 familie. NF-KB activatie wordt bepaald door veel factoren waaronder JNK en Pl3K/Akt.

Dit proefschrift beschrijft de mechanismen van celdood tijdens leverziekten en de mogelijkheden voor preventie van deze celdood. In leveraandoeningen spelen galzouten, oxidatieve stress en cytokines een belangrijke rol bij celschade en celdood van hepatocyten. Galzouten spelen een cruciale rol bij acute en chronische cholestase, een aandoening die beschreven wordt als een verstoorde enterohepatische kringloop waardoor accumulatie van galzouten in bloed en lever optreedt. Dit kan veroorzaakt worden door een genetisch defect, geneesmiddelen, immunologische of mechanische schade (obstructie a.g.v. tumor of galstenen). Vrije zuurstof radicalen zijn belangrijk bij alcohol-gerelateerde leverschade en

99

Chapter B

NASH. Cytokines spelen een rol bij vrijwel elke vorm van leverschade, maar zijn specifieke veroorzakers van acuut leverfalen en acute hepatitis. De meeste chronische leverziekten verlopen van fibrose via cirrose en eindigen in leverfalen. De enige effectieve remedie bij leverfalen is een levertransplantatie. Geneesmiddelen die warden toegepast bij leveraandoeningen zijn onder andere ursodeoxycholzuur (UDCA, cholestase) en insuline­gevoeligheid verhogende middelen als metformine (NASH). Helaas zijn deze middelen niet altijd effectief, waardoor de ziekte zich verder kan ontwikkelen en uiteindelijk een transplantatie nodig is. Het is dan oak van zeer groat belang om kennis te hebben van de mechanismen die de leveraandoening veroorzaken, zodat nieuwe mogelijkheden voor ingrijpen en potentiele geneesmiddelen bestudeerd kunnen warden. In dit proefschrift is onderzoek gedaan naar drie verschillende in vitro modellen van leverschade. Ten eerste het toxische galzout GCDCA als model voor acute galzout schade. GCDCA induceert zowel apoptose als necrose in hepatocyten. Ten tweede het cytokine TNF, een model voor acuut leverfalen gekenmerkt door apoptose. Als derde is de superoxide anion donor menadione gebruikt als model voor oxidatieve stress-ge"induceerde apoptose.

In hoofdstuk 2 wordt de beschermende werking van het anti-diabetische geneesmiddel metformine in galzout-ge"induceerde celdood beschreven. Naast de positieve effecten van metformine op de insuline-gevoeligheid in patienten met Diabetes type I I , heeft metformine oak gunstige effecten op merkers voor leverschade bij patienten en diermodellen voor steatotische leverziekten. Een bekende effector van metformine is AMP-geactiveerd prote"ine kinase (AMPK), de energie sensor van de cellen. AMPK wordt geactiveerd door veranderingen in de energie huishouding (AMP:ATP ratio), maar oak door bijvoorbeeld farmacologische middelen of tijdens condities van cellulaire stress. AMPK veroorzaakt zijn effecten a.a. via mTOR (mammalian target of rapamycin), een kinase dat betrokken is bij transcriptie, translatie en eiwit synthese.

Onze studie toont aan dat metformine dosis-afhankelijk beschermt tegen GCDCA· ge"induceerde apoptose in ge"isoleerde rat hepatocyten, zonder een effect te hebben op necrotische celdood. Metformine biedt geen bescherming in ons model voor acuut leverfalen (TNFa in combinatie met de transcriptie remmer actinomycine D). Metformine is in staat om apoptose le remmen tot 2 uur na toevoeging van het galzout. De bescherming van metformine is niet het gevolg van een verminderde galzout opname in de hepatocyten door transporter eiwitten. De bescherming van metformine moet daarom afhankelijk zijn van een snelle activatie van signaaltransductie routes, zoals de Pl3K/Akt overlevingsroute: het beschermend effect van metformine is afhankelijk van een functionerende PI3K/Akt overlevingsroute. De bescherming is onafhankelijk van de bekende effector route van metformine, A MPK/mTOR. De verschillende leden van de MAPK familie, zowel pro- als anti­apoptotisch, zijn eveneens niet betrokken bij het beschermend effect van metformine. Metformine heeft oak geen remmend effect op de transcriptie factor NF-KB. Dit laatste resultaat is erg belangrijk, omdat metformine hierdoor mogelijk toepasbaar is in cholestatische leverziekten die gepaard gaan met ontsteking.

Hoofdstuk 3 beschrijft de bescherming van leflunomide tegen galzout-ge"induceerde celdood in ge"isoleerde rat hepatocyten. Leflunomide is een geneesmiddel voor de behandeling van auto-immuunziekten, o.a. rheumato"ide artritis. Daarnaast wordt leflunomide toegepast als therapie bij de ziekte van Crohn en kanker. Leflunomide is een zogeheten 'prodrug': in het lichaam wordt leflunomide voor ongeveer 95% omgezet in de actieve metaboliet A77 1726. Verscheidene effecten van leflunomide en de actieve metaboliet zijn beschreven in de literatuur waaronder remmende effecten op tyrosine kinases, pro­apoptotische MAPK leden en cytokine productie via effecten op NF-KB. Bovendien is van beide stoffen bekend dat ze in experimentele modellen levercellen beschermen tegen schade veroorzaakt door onder andere paracetamol, concanavalin A (model voor immuun­gemedieerde hepatitis) en galgang ligatie (model voor cholestase, BDL). Er zijn echter oak celdood stimulerende effecten voor beide stoffen beschreven, echter wel in andere celtypes. 1 00

Nederlandse Samenvatting

Ons onderzoek toont aan dat de actieve metaboliet, A77 1726, primaire rat hepatocyten beschermt tegen GCDCA-ge"induceerde apoptose en necrose, maar geen bescherming biedt tegen apoptose ge"induceerd door cytokines of oxidatieve stress. Het mechanisme achter de anti-apoptotische effecten van leflunomide blijft onbekend. Onze resultaten laten duidelijk zien dat de bescherming niet veroorzaakt wordt via de anti-apoptotische MAP kinases p38 en ERK1/2 of via de Pl3K/Akt signaal transductie route. Aangezien A77 1726 eenzelfde resultaat veroorzaakt op galzout-ge"induceerde celdood als een farmacologische remmer van de Src-kinases, impliceren onze resultaten dat A77 1726 mogelijk een Src­kinase remmer is. Vervolgonderzoeken zijn nodig om deze suggestie te kunnen onderbouwen. Het feit dat leflunomide geen remmend effect heeft op NF-KB impliceert dat deze stof toepasbaar is bij leverziekten waarbij ontsteking een rol speelt. In vivo studies zijn nodig om onze in vitro resultaten te onderbouwen.

In hoofdstuk 4 van dit proefschrift word! de rot van vrije zuurstof radicalen in galzout­ge"induceerde celdood van hepatocyten nader onderzocht. Verschillende studies beschrijven een rot voor oxidatieve stress in galzout-ge"induceerde celdood en de bescherming hiertegen middels anti-oxidanten. Andere publicaties tonen echter aan dat er een gering of zelfs geen verband bestaat tussen galzout-ge"induceerde productie van vrije zuurstof radicalen en celdood. In ons onderzoek zijn de mechanismen van twee verschillende toxische galzouten, GCDCA en taurolithocholzuur 3-sulfaat {TLCS), bestudeerd. Verschillende anti-oxidanten en 'wegvangers' van vrije zuurstof radicalen veroorzaakten geen afname in celdood ge"induceerd door deze galzouten. GCDCA en TLCS zorgden ook niet voor een inductie van het oxidatieve stress gevoelige gen HO-1 of een reductie van de totale glutathion spiegels in de hepatocyten. Tezamen suggereren deze bevindingen dat oxidatieve stress niet noodzakelijk is voor galzout-ge"induceerde celdood. De specifieke Src­remmer SU6656 bood wel bescherming tegen GCDCA-ge"induceerde celdood, zowel tegen apoptose als necrose. Oak remming van het enzym NADPH oxidase, welke vrije zuurstof radicalen vormt, resulteerde in een sterke afname van galzout-ge"induceerde apoptose. Deze reductie ging echter gepaard met een significante toename van necrose in hepatocyten. Een vergelijkbaar fenomeen -de verschuiving van apoptose naar necrose- is eerder waargenomen met de stikstofmonoxide donor SNAP in combinatie met de superoxide anion donor menadione. Onze resultaten suggereren dat de generatie van vrije zuurstof radicalen door galzouten geen essentiele rot speelt in galzout-ge"induceerde apoptose, maar dat de Src-kinases, met name Yes, hierbij wel een cruciale rot spelen.

Hoofdstuk 5 beschrijft de beschermende rol van het enzym HO- 1 en een van de producten van dit enzym, koolstofmonoxide (CO), in verschillende in vivo en in vitro modellen van leverschade. Het oxidatieve stress-gevoelige gen HO-1 is ge"induceerd na galgang ligatie, een in vivo model voor chronische cholestase, terwijl er geen effect op HO-1 waarneembaar is in een in vivo model voor acute ontsteking. In in vitro experimenten induceerde de superoxide anion donor menadione HO-1, terwijl waterstof peroxide en cytokines niet resulteerden in verhoogde HO-1 expressie in hepatocyten. Overexpressie van HO-1 vermindert menadione-ge"induceerde apoptose, hetgeen deels veroorzaakt wordt door koolstofmonoxide. De anti-apoptotische werking van koolstofmonoxide is het gevolg van een directe remming van de pro-apoptotische MAP kinase JNK. Een vergelijkbaar effect van koolstofmonoxide op JNK activatie is beschreven in een model voor sepsis. Uit dit onderzoek kunnen we concluderen dat HO-1 en koolstofmonoxide, maar mogelijk ook andere HO-1 producten zoals biliverdine, potentiele therapeutische opties zijn voor de behandeling van oxidatieve stress-gerelateerde leverschade.

Patienten met leveraandoeningen beginnen meestal met een therapie als de symptomen merkbaar worden, terwijl de ziekte dan vaak al langere tijd aanwezig is. Het is dan ook belangrijk om onderzoek te doen naar reeds beschadigde levercellen en de werkingsmechanismen in deze cellen te bestuderen. Aangezien nog veel onderzoek naar leverschade wordt uitgevoerd met normale 'gezonde' cellen beschrijft hoofdstuk 6 van dit proefschrift de toepassing van cholestatische hepatocyten in in vitro experimenten.

1 01

Chapter B

Hepatocyten werden geTsoleerd uit proefdieren die een galgang ligatie (1 week) hadden ondergaan. Ten opzichte van hepatocyten ge"isoleerd uit controle proefdieren hadden de cholestatische hepatocyten een verminderde vitaliteit en was de opbrengst van de cellen verlaagd. Direct na isolatie en tijdens experimenten werd het fenotype van de cholestatische hepatocyten vergeleken met de normale hepatocyten. Zoals beschreven in de literatuur, is de expressie van de transporter eiwitten Ntcp en Mrp2 verlaagd in cholestatische hepatocyten, terwijl de expressie van Mrp3 juist verhoogd is. Deze expressie patronen blijven gehandhaafd gedurende de experimenten (8 uur). De cholestatische hepatocyten behouden dus hun fenotype. Preliminaire resultaten suggereren dat cholestatische hepatocyten minder gevoelig zijn voor galzout-ge"induceerde celdood. GCDCA induceert apoptose en necrose in normale hepatocyten, maar in cholestatische hepatocyten is, ondanks de lagere vitaliteit van de cellen, geen verschil met de ongestimuleerde cellen waar te nemen. Een verklaring h iervoor zou kunnen zijn dat cholestatische hepatocyten -door de verlaagde expressie van Ntcp- niet meer in staat zijn om galzouten op te nemen. Uitgebreid onderzoek is vereist om deze ongevoeligheid en het mechanisme hiervan in cholestatische hepatocyten in detail te bestuderen. Het hier beschreven en gevalideerde in vitro model voor cholestase is een belangrijk middel voor deze toekomstige studies.

Conclusie Het ophelderen van de rol van de verschillende signaal transductie routes die betrokken zijn bij de dood van hepatocyten bij verschillende leverziekten is noodzakelijk voor de ontwikkeling van nieuwe therapieen voor de behandeling van deze ziekten. In dit proefschrift worden de beschermende effecten van metformine en leflunomide tegen galzout­ge"induceerde celdood in vitro beschreven. Met deze onderzoeken is een begin gemaakt aan de mogelijke toepassing van metformine en leflunomide bij cholestatische leverziekten. Daarnaast heeft ons onderzoek naar de mechanismen van galzout-ge"induceerde celdood meer kennis opgeleverd over de rol van oxidatieve stress in dit proces. Als laatste hebben we een nieuwe opzet voor het bestuderen van cholestatische hepatocyten in vitro beschreven. Deze techniek zal in de toekomst bijdragen aan de kennis over de mechanismen en preventie van celdood van hepatocyten tijdens cholestatische leveraandoeningen.

1 02

Curriculum Vitae

List of Publications

Curriculum Vitae

Curriculum Vitae

Titia (Eveline) Vrenken was born on November 21st 1978 in Harlingen, The Netherlands. She completed the VWO in 1997 at the Rijksscholengemeenschap Simon Vestdijk in Harlingen, and started with the study Pharmacy at the Rijksuniversiteit Groningen. She performed her research project at the Department of Pharmaceutical Biology of Prof. Dr. W.J. Quax. Supervised by Dr. M.J. Drage, she investigated the improvement of the enantioselectivity of LipaseA from Bacillus subitilis. She obtained her M.Sc. in Pharmacy in January 2003. In May 2003, she started to work on her PhD at the Department of Hepatology and Gastroenterology at the U niversity Medical Center Groningen, The Netherlands. Under supervision of Prof. Dr. A.J. Moshage, she studied the mechanisms and prevention of hepatocyte cell death, as described in this thesis. This project was financed by the Dutch Digestive Foundation (MLDS, grant WS 02-13). During her PhD she did an internship of three months in 2006 at the Medical U niversity of South Carolina in Charleston, the United States of America. At the Department of Pharmaceutical Sciences of Prof. Dr. J.J. Lemasters she learned specific mitochondrial stainings using confocal laser scanning microscopy. This internship was funded by grants from the Netherlands Organisation for Scientific Research (NWO), the Van Walree Fund of the Royal Netherlands Academy of Arts and Sciences (KNAW) and the Dutch Digestive Foundation. She is currently working as a post-doctoral researcher at the Department of Physiology at the Radboud University Nijmegen Medical Centre, The Netherlands under supervision of Dr. J.G.J. Hoenderop en Prof. Dr. R.J.M. Bindels.

1 04

List of Publications

List of Publications

Vrenken TE, Buist-Homan M, Kalsbeek AJ , Faber KN, Moshage AJ. The active metabolite of leflunomide, A77 1726, protects rat hepatocytes against bile acid-induced apoptosis. Submitted

Vrenken TE, Buist-Homan M, Faber KN, Moshage AJ. Reactive oxygen species are not essential for bile acid-induced cell death. Submitted

Vrenken TE, Conde de la Rosa L, Buist-Homan M, Faber KN, Moshage AJ. Metformin protects rat hepatocytes against bile acid-induced apoptosis. In preparation

Vrenken TE, Buist-Homan M, Woudenberg J , van den Heuvel FAJ, Faber KN, Moshage AJ. Isolation, culture and characterization of cholestatic hepatocytes: an in vitro approach to study cholestatis. In preparation

Conde de la Rosa L, Vrenken TE, Hannivoort RA, Buist-Homan M, Havinga R, Slebos DJ, Kauffman HF, Faber KN, Jansen PLM and Moshage H. Heme oxygenase-1 derived carbon monoxide protects primary hepatocytes against oxidative stress-induced apoptosis via inhibition of JNK activation. (2008) Free Radie Biol Med 44(7): 1323-33

Conde de la Rosa L, Schoemaker MH, Vrenken TE, Buist-Homan M, Havinga R, Jansen PLM and Moshage H. Superoxide anions and hydrogen peroxide induce hepatocyte death by different mechanisms: involvement of J NK an ERK MAP kinases. (2006) Journal of Hepatology 44(5): 918-929

Droge MJ , Boersma YL, van Pouderoyen G, Vrenken TE, Ruggeberg CJ, Reetz MT, Dijkstra BW and Quax WJ. Directed evolution of Bacillus subtilis lipase A by use of enantiomeric phosphonate inhibitors: crystal structures and phage display selection. (2006) Chembiochem 7(1): 149-57

Schoemaker MH, Conde de la Rosa L, Buist-Homan M, Vrenken TE, Havinga R, Poelstra K, Haisma HJ , Jansen PLM and Moshage H. Tauroursodeoxycholic acid protects rat hepatocytes from bile acid-induced apoptosis via activation of survival pathways. (2004) Hepatology 39(6): 1563-73

105

Acknowledgements/Dankwoord

Acknowledgements/Dankwoord

Het is al weer even geleden dat ik vanuit het bacterie onderzoek bij Farmaceutische Biologie de overstap maakte naar de lever. In mei 2003 had ik nooit kunnen bedenken wat voor gevolgen dit promotie onderzoek allemaal voor mijn persoonfijke leven zou hebben. In deze jaren heb ik veel geleerd, mezelf overtroffen en vriendschappen gelegd die ook buiten Groningen stand houden. Daarnaast is dit proefschrift het bewijs van de succesvolle jaren bij het lab Maag Darm en Leverziekten (MDL), hetgeen zonder mijn collega's niet gelukt was.

Han, eerst als begeleider en later als promotor ben je altijd erg betrokken geweest bij mijn onderzoek. Jouw kennis over de lever, hepatocyten en galzouten hebben voor dit mooie proefschrift gezorgd. lk heb de afgelopen jaren veel van je kunnen leren, en waardeer je wetenschappelijke inzichten enorm. Ook jouw 'maffe' feitenkennis en bizarre verhalen, vooral tijdens de hepatocyten isolaties op het CDL, staan mij na. Ondanks dat we het niet altijd eens waren-gelukkig hoeft dat ook niet-wil ik je heel erg bedanken dat je mijn promotor bent. Manon, zonder jou was dit maar een half proefschrift geworden. En dan heb ik het niet alleen over de technische ondersteuning bij de experimenten, maar ook de vele gezellige uren samen in de celkweek, het CDL (koekjes) en op het lab, je bent voor mij een enorme steun geweest de afgelopen jaren; heel erg bedankt dat je mijn paranimf wilt zijn; je taak eindigt op 11 juni! En niet te vergeten de gezellige avonden en vele wijntjes (jij soms cola) die we met Marcus, en naderhand met Jannes erbij, gedronken hebben. En natuurlijk een dikke knuffel voor Anna. Klaas Nico, jou wil ik graag bedanken voor je steun en advies de afgelopen jaren, je deur stand altijd voor me open.

Graag wil ik de leescommissie bestaande uit Ulrich Beuers, Folkert Kuipers en Ingrid Molema bedanken voor het kritisch beoordelen van mijn manuscript en hun nuttige suggesties.

De andere (ex-)MDLers: Antonella, Axel, Elise, Fiona (bedankt voor de gezellige samenwerking met name op het CDL), Gerard, Golnar, Hans, Jacqueline, Janette, Krzysztof (Friday beers, cheese and chats in town), Kyrjon (heel erg bedankt voor alle hulp aan mijn project), Laura, Lisette, Lu, Marieke, Mariska, Martijn, Rebekka, Sandra en Tjasso allemaal bedankt voor de gezellige tijd op het lab en de leuke dingen daarbuiten! Dank ook aan de oud-studenten Sytze, Paul, Hugo en Allard Jan voor hun bljdrage aan dit onderzoek. En natuurlijk kan een MDL-lab niet bestaan zonder de kliniek, daarom wil ik mijn collega's uit de kliniek bij deze heel hartelijk bedanken voor hun suggesties tijdens de research meetings.

Verder zijn veel goeie herinneringen van de afgelopen jaren verbonden met de uren die ik heb doorgebracht in onze A IO-kamer samen met mijn (ex)collega's van MDL en kindergeneeskunde. Leonie, Jelske (veel geluk in Edmonton), Maaike, Esther, Hans, Martijn, Niels, Miriam, Anniek W, Tineke en Renate bedankt voor alle gezellige momenten, de morele ondersteuning (veel chocola) en het heerlijke meiden geklets. Daarnaast natuurlijk de barrels en etentjes buiten het werk om. Vooral de laatste maanden-voordat we langzaam maar zeker allemaal klaar waren (Maaike, nag even volhouden, jij bent de volgende!)-het vieren van afgeronde proefschriften, het eten bij Leo's oom Ton aan board met zijn enorme hoeveelheden lekker eten en vele liters wijn; in een woord super.

En de overige (ex)collega's van kindergeneeskunde, waarmee MDL het lab deelt, Aldo, Anja, Anke, Annelies, Anniek K, Baukje, Christian, Dirk-Jan, Edmond, Ekkehard, Frans C. , Frans S. , Feike, Gea, Gemma, Harmen, Henk, Henkjan, Hester, Hilde H. , Hilde R. , Jaap, Jan Freark, Janine, Janny, Juul, Karin, Margot, Marijke, Mark, Maxi, Nicolette, Renze, Rick (bedankt voor de hepatocyten isolaties), Robert, Sabina, Terry, Thierry, Thomas, Torsten, Uwe, Vincent, Wietze, Yan en iedereen die ik nog vergeten ben heel erg bedankt voor de gezellige jaren op het lab.

108

Acknow/edgements/Dankwoord

Voor een MDL-AIO heb ik relatief veel tijd doorgebracht op het dierenlab. Annemieke, Annette, Arie, Flip, Hester, Natasha, Marcia, Ragnhild en Yvonne, bedankt voor alle hulp en ondersteuning bij de proeven en voor de gezellige praatjes.

lk wil in het bijzonder Peter Braun en Riekje Banus (GUIDE) bedanken voor hun belangstelling in mij en mijn onderzoek.

I'd like to thank John Lemasters for his kind invitation to visit his lab in Charleston, South Carolina to learn specific microscopy techniques. Thanks to 'my colleagues' in Charleston, Akira (thank you very much for learning me the mouse hepatocyte isolations) , lnsil, Silvia, Tom, Venkat and Zhi. Special thanks to my true friends Silvia&Angel. Thanks for showing me the good things of Charleston; I really enjoyed all the dinners, drinks, dancing and in particular watching the matches during the world championship soccer! Silvia, your flowers will bloom forever. Also thanks to my roomies in Charleston, Ricardo and Elba, without whom I wouldn't have survived.

Naast de collega's wil ik graag mijn vrienden en familie bedanken. Eefje&Axel, onze vriendschap gaat al lang terug, en wat een lief en leed hebben we al gedeeld in die jaren. Heel erg bedankt voor alle gezellige dingen (barrels, etentjes, toneel, cabaret, musical) die we samen gedaan hebben, zelfs aan de Struuntocht-hoe ellendig die ook was-heb ik goeie herinneringen over gehouden. Spannend waar jullie terecht zullen komen, hopelijk wonen we straks weer iets dichter bij elkaar dan nu! Leonie, van een collega ben jij een vriendin geworden en meer. lk mis je vreselijk, al weet ik dat het super voor je is om in Australie te zijn, toch had ik je graag als mijn paranimf willen hebben om mij bij de laatste loodjes bij te staan. Tot snel in Nijmegen of Australie! Andrea&Arjan, heel erg bedankt voor alle gezellige etentjes, bij jullie was ik altijd even helemaal weg uit de werksfeer; en knuffelen met Naut is het leukste wat er is! Heel veel geluk in 't Harde, we komen snel kijken in jullie paleis. Eva& Tobin, van buuf via buuf naar ver weg verhuizing. De zachte g zit er bij mij nog niet echt in, maar wie weet . . . Bedankt voor de vele gezellige avonden in de van Heemskerckstraat, de Hyacintstraat en de Nieuwe Kijk (ben benieuwd hoe het uiteindelijk wordt). Dit jaar weer een feestje in Plasmolen? Martina, als oudste huisgenoot hebben wij samen veel meegemaakt. Hilarische studenten-stap-avonturen en vele uithuilmomenten door de jaren heen, bedankt voor gezelligheid. Dorianne en Lodewijk, vrienden uit de AEGEE-/middelbare school-tijd en nog steeds gaan we regelmatig eten en spelletjes doen tot diep in de nacht. Gelukkig zijn we overgestapt naar de zaterdagavonden! Bedankt voor de vele uren gezelligheid, snel weer een keer afspreken. Annette&Peter, Annette wij zijn al zo lang vriendinnen en ondanks de afstand blijft dat. Onze vakantie vorig jaar op St. Maarten was super, eindelijk kon ik zien hoe jouw leven samen met Peter daar is (vreselijk, wonen aan het strand!). Super dat jullie bij onze trouwerij waren, op naar die van jullie op 12 juli. Margot&Lieuwe, lieve Margot, zelfs voor een telefoon-date zijn we vaak al weken voor uit aan het plannen, en we zien elkaar ook niet zo vaak als ik wel zou willen. Hopelijk komt daar-nu dit proefschrift eindelijk af is-verandering in; op naar Den Haag! Hans&Judith, Hans als collega heb ik je enorm gewaardeerd (en erg gemist toen je naar Zwolle ging), en ik heb het erg leuk gevonden om Judith te leren kennen en met z'n vieren gezellige avonden door te brengen; wat een prachtige zoon hebben jullie. Tot snel in Nijmegen. Hilda&Rienk, ons goede voornemen om vaker dan eens per jaar at te spreken is volgens mij nu al weer mislukt . . . dan maar weer in september naar Heerenveen (in jullie nieuwe huis). Bedankt voor de gezellige eetafspraken door de jaren heen!

Mijn familie. Zander mijn familie had ik dit niet kunnen doen. Lieve Caroline&Hans en Jock, jullie huis in A'dam is altijd een fijne plek om te zijn, heerlijk kletsen en spelletjes doen in de tuin, voetballen in het park of logeren op Schier (deze zomer wel weer!). Mag 2008 niets dan goeds brengen. Tante Myra&Dolf, bedankt voor de leuke kaarten, mailtjes en gesprekken. Onze logeerpartij bij jullie inclusief de warmte van de familie heeft gezorgd voor het meest relaxte congres ooit afgelopen oktober! Marianne, als rokers waren we altijd een slag apart bij de familie-uitjes. Bedankt voor de vele sigaretjes samen en goeie gesprekken door de

1 09

Acknowledgements/Dankwoord

jaren heen. Mijn schoonfamilie, wat klinkt dat officieel, wil ik oak graag bedanken. Rikje&Evert, jullie boerderij is altijd een warm, gastvrij huis voor mij; kalfjes knuffelen blijft een van mijn favoriete bezigheden bij jullie. Henrique&Erik, Lil& Theo, Evert&Merel en Bram, Erik&Rob bedankt voor de gezelligheid en interesse, gelukkig wonen we nu weer iets dichter bij.

Hugo, jij bent als oudere broer altijd mijn grate voorbeeld geweest. Vandaar waarschijnlijk oak dat ik alles doe wat jij hebt gedaan, studeren in Groningen, promoveren; behalve dat ik getrouwd ben en jij en Roos een geregistreerd partnerschap hebben. Allebei heel erg bedankt voor alle gezellige momenten en goeie gesprekken die we samen hebben gehad in Amsterdam, Groningen en Harlingen. Super dat jij mijn paranimf wilt zijn. Kleine zusjes warden groat, maar na 11 juni is het volgens mij wel klaar. Bedankt dat je mijn broer bent. Pap en mam, bedankt voor de steun en warmte die ik altijd van jullie heb gekregen. Jullie thuis was en is altijd een veilige haven. Jullie zijn mijn basis en daar ben ik trots op. Dan rest mij nag om Pim&Pom te noemen, onze 2 katten. De belangrijkheid van deze monsters bleek al wel uit onze trouwkaart. De liefde en troost die ik van hun krijg heeft mij door menig moeilijk moment heen geholpen, net zoals ik evenveel plezier heb beleefd aan hun grappige, soms irritante en speelse gedrag. En als allerlaatste mijn lieve Jannes. Jij bent het allerleukste wat ik aan dit AID-project heb overgehouden! Volgens mij heb ik met onze trouwerij al wel laten weten wat jij voor mij betekent. Heel erg bedankt voor je onvoorwaardelijke steun de afgelopen jaren (en vooral de laatste tijd), jouw advies en relativeringsvermogen hebben mij er doorheen getrokken tijdens de moeilijke momenten. lk weet niet of ik het zonder jou oak had volgehouden ... Op naar Nijmegen!

Liefs Titia

1 10