\:L,:\

n.5,. ,0 gm
“n. .

J: ‘ ,
.6.” \umxc.

«my

. 7
“$3.
t.” .2. “twin.
)l . on”
.

1 .M. 5.1;.” Ms
arxw‘hl .‘ «mkflwm
u guuua
Y .

;
)

€01! r. ,,
.uflfldmairm
70‘s.: .

I
7..
“awn”:

"i?! 1

It!

I .X . I
(£153.)
L512.

 

 

l ‘l u .
19154.5

9%:

{j :vf'fy/(f

This is to certify that the
dissertation entitled

Inflammation and idiosyncratic reactions: Ranitidine as a
model

presented by
James Parker Luyendyk

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Pharmacblogy and Toxicology_

 

 

va/L WED

Major Professor’s Signature
l '2" _, Cl - C K]

Date

 

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DateDue.p65-p.15

 

INFLAMMATION AND IDIOSYNCRATIC DRUG REACTIONS:
RANITIDINE AS A MODEL

By

James Parker Luyendyk

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Pharmacology and Toxicology

2004

ABSTRACT
INFLAMMATION AND IDIOSYNCRATIC DRUG REACTIONS:
RANITIDINE As A MODEL
By

James Parker Luyendyk

Idiosyncratic reactions occur in a small fraction of people taking a drug. The liver
is a fiequent target. For the vast majority of drugs associated with these reactions,
including the histamine2 (HZ) receptor antagonist ranitidine, the mechanism of toxicity is
unknown. Inasmuch as a modest inflammatory response can render rats susceptible to
hepatotoxic effects of several xenobiotics, inflammation might also be a determinant of
sensitivity to drug toxicity. This dissertation tested the hypothesis that an inflammatory
response caused by a nonhepatotoxic dose of bacterial lipopolysaccharide (LPS) could
precipitate idiosyncrasy-like RAN hepatotoxicity in rats. Administration of a small dose
of LPS (44 X 106 EU/kg, iv) two hours before a nonhepatotoxic dose of RAN (30 mg/kg,
iv) caused significant hepatotoxicity characterized by midzonal hepatocellular oncotic
necrosis, the onset of which occurred between 2-3 h after drug treatment. These
histopathologic findings and changes in clinical chemistry resembled idiosyncratic RAN
hepatotoxicity in people. The H2-antagonist famotidine (F AM), which is not associated
with idiosyncratic hepatotoxicity, was not rendered toxic by LPS cotreatment. Hepatic
gene expression was evaluated at a time before injury in rats treated with LPS and/or
RAN, and hierarchical clustering of active genes segregated rats to their respective
treatment groups. Several genes related to hypoxia, inflammation and cell death were

expressed to a greater degree in LPS/RAN-treated rats compared to either agent given

alone. This pattern of expression for plasminogen activator inhibitor-1 (PAI-l) was
confirmed by real-time PCR and was mirrored by the PAL] concentration in the plasma
of LPS/RAN-treated rats. Consistent with the antifibrinolytic function of PAH, hepatic
fibrin deposition was observed only in livers of LPS/RAN-treated rats. To determine if
hepatic fibrin deposition was a consequence of impaired fibrinolysis, coagulation system
activation, or both, several biomarkers of coagulation were evaluated. Changes in serum
hyaluronic acid, an indicator of altered sinusoidal endothelial cell (SEC) homeostasis,
thrombin-antithrombin dimers, and fibrinogen suggested coagulation system activation in
LPS/RAN-treated rats. Liver injury and hepatic fibrin deposition were attenuated by both
the fibrinolytic agent streptokinase and the anticoagulant heparin, indicating that the
hemostatic system is involved in LPS/RAN-induced liver injury. Liver hypoxia, one
consequence of fibrin deposition, was observed only in livers of LPS/RAN-treated rats
and was reduced by heparin coadministration. Neutrophils (PMNs) accumulated in livers
of LPS/RAN-treated rats, and killing of primary rat hepatocytes by PMN elastase in vitro
was augmented under hypoxic conditions. The results suggest that the hemostatic system
is important for liver injury in LPS/RAN-treated rats and are consistent with resultant
hypoxia interacting with inflammatory mediators to cause hepatocellular injury.
Furthermore, these studies support the possibility of predicting and understanding the
mechanisms of some idiosyncratic drug reactions by utilization of a drug/LPS-

cotreatrnent model.

ACKNOWLEDGMENTS

Of all the sections in this dissertation, this is perhaps the most difficult to write. A
very large number of people fi'om Michigan State University from the Department of
Pharmacology and Toxicology, Department of Biochemistry and Molecular Biology, and
the Center for Integrative Toxicology deserve tremendous thanks. In addition, my
interactions with other collaborators and colleagues fi'om Pharmacia Corporation, Bristol-
Myers Squibb, University of Louisville, the Society of Toxicology, and the Michigan
Regional Chapter of the Society of Toxicology have dramatically increased my
excitement for Toxicology. In addition to laboratory funding I would like to thank the
organizations which have supported my research through awards and fellowships. To
paraphrase these thanks would be insufficient. Accordingly, grab your 64 oz. soda and
get ready to read.

Let’s start with the MSU folks, shall we. First and foremost I would like to thank
my mentor and fi'iend, Dr. Robert Roth. Deciding to join his laboratory was perhaps the
most influential and best decision I have made. I hope I can reflect Bob’s enthusiasm and
perseverance in toxicological research as I move onward in my career. The advice Bob
has given me will not be forgotten. Dr. Patricia Ganey also deserves my thanks. Patti
played a very significant role in my graduate career by serving as a member of my
graduate committee and as another mentor in the lab, and I am proud to have coauthored
several manuscripts with her. I am perhaps more of a pest at times to Dr. Jane Maddox.
However, this speaks strongly to Jane’s involvement in the lab’s activities. Jane’s door is

always open and she seems to know the answers to all my questions. I would like to

iv

thank Jane for the significant amount of time she put in to training me in the lab as well
as the teamwork on this dissertation work. I have enjoyed working with Jane and will
miss greatly my daily interactions with her. I would also like to thank Dr. James Pestka
for serving on my graduate committee and for his continued interest in my research and
career goals.

I have had the pleasure of working with so many other incredible people in the
laboratory during my time at MSU and all of them deserve my thanks. Several post-docs
in the lab have helped me at times including Dr. Umesh Hanumegowda, Dr. Bryan
Copple, and Dr. Francis Tukov. In addition to these guys, I would like to thank all of the
laboratory technicians, graduate students, and undergraduates I have had the opportunity
to work with including Sandy Newport, Cathy Rondelli, John Phipps, John Buchweitz,
Mary Kinser, Steve Yee, Xiaomin Deng, Chris Green, Steve Bezdecny, Shawn Kinser,
Pat Shaw, Colin North, Kate Shores, Rohan Pradhan, Theresa Eagle, Peer Karmaus,
Natasha Snyder, Brook Wooley, and Danielle Feldpausch. In addition to folks in the lab,
I would like to thank Dr. Tim Zacharewski for his assistance with our microarray work.
Along this line, I need to thank Lyle Burgoon for his patience and significant help with
rnicroarray analysis. I would also like to thank Dr. Larry Fischer and the administrative
staff of the Center for Integrative Toxicology (Institute for Environmental Toxicology). I
also thank Dr. Stephanie Watts for her continued leadership of the Graduate Program in
Pharmacology and Toxicology and for her much needed advice on several occasions.

In addition to colleagues on campus I have offered the unique opportunity to
collaborate with others in industry and academia on the work presented herein as well as

other projects. Dr. Greg Cosma has served as another mentor to me and as a member of

my committee. Greg has been helpful in all aspects of my career fi'om experimental
design to encouragement. I met Greg through our collaboration with Pharmacia where I
also had the chance to work with several other awesome folks including Dr. Gary
Cockerell, Dave Blakeman, John Wijsman, and Dr. Glenn Cantor. Through the assistance
of Glenn and Greg we were able to forge an incredible collaboration with Lois Lehman-
McKeeman at Bristol-Myers Squibb. I would like to thank Lois, the members of her
group including David Nelson, Vasanthi Bhaskaran, as well as Dr. Bruce Car, Dr. Glenn
Cantor, and Dr. Bill Foster for the opportunity to visit EMS and analyze gene array data
in the real world. I would also like to thank Glenn and Inge Cantor for a fantastic 3-day
stay at the Cantor homestead in Princeton. In addition to industrial collaborators, I would
like to thank Dr. Gavin Arteel for opening up both his lab and home to me for a fun 4-day
visit to Louisville. Dr. Arteel’s encouragement is also greatly appreciated.

I owe a big thank you to the Society of Toxicology for all that it has offered me
during the past 4 years. The list of folks is too big, but I’ll just say thanks. SOT has been,
and will be a big part of my career. One person who deserves a big thank you is Betty
Eidemiller. More or less, I should just thank Betty for everything. I would also like to
thank the Committee and organizers of the Michigan Regional Chapter of the Society of
Toxicology. Interactions at the Regional Chapter level have had an enormous impact on
me personally and professionally.

I should take a moment to thank those who made my research possible. First off,
Bob kept my studies going by seeking fimding from multiple sources and deserves all the
credit. In addition, I would like to thank the Center for Integrative Toxicology (N IEHS

training grant), Michigan State University, the Society of Toxicology and Novartis

vi

Corporation for their assistance with paying my stipend. Also, I would like to thank the
MSU College of Veterinary Medicine, the Midwest Regional Chapter of the SOT, the
Michigan Regional Chapter of the SOT, and the ASPET Toxicology Division for
sponsoring awards enabling me to travel to several professional meetings.

I would also like to thank 7-11 for keeping the soda fountain running and Bryan
and Eric at University Foreign Car for keeping the Volvo running. Finally, I would like to
thank my wife Elizabeth for her unconditional support and love despite how crazy I get.
Also, Liz deserves all the credit for my new and improved fashion sense. Finally, without
my parents I would have missed out on the opportunities I have had and would not be
looking forward to those in the future. Nor would I have such a cool car. So thanks for

everything Mom and Dad...and yes, I checked the oil.

vii

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. xi
LIST OF FIGURES ............................................................................... xii
LIST OF ABBREVIATIONS .................................................................... xiv
CHAPTER 1
General Introduction .............................................................................. 1
1.1 Idiosyncratic hepatotoxicity ...................................................... 2
1.1.1 Background and proposed mechanisms .............................. 2
1.1.2 Examples: halothane and diclofenac ................................. 6
1.1.3 The “multiple determinant hypothesis” .............................. 16
1.1.4 Ranitidine and idiosyncratic hepatotoxicity ......................... 20
1.2 Inflammation .......................................................................... 25
1.2.1 Incidence and triggers .................................................. 25
1.2.2 Lipopolysaccharide-induced liver injury ............................ 28
1.2.3 Lipopolysaccharide—induced sensitivity to hepatotoxicity. . . . .....35
1.3 Inflammation-induced idiosyncrasy: a paradigm for hepatotoxicity ........ 44
1.4 Overview of dissertation .......................................................... 51
CHAPTER 2
Ranitidine treatment during a modest inflammatory response precipitates
idiosyncrasy-like liver injury in rats ............................................................ 52
2.1 Abstract .............................................................................. 53
2.2 Introduction ......................................................................... 54
2.3 Methods ............................................................................. 54
2.3.1 Materials ................................................................. 54
2.3.2 Animals .................................................................. 55
2.3.3 Experimental protocol .................................................. 55
2.3.4 Assessment of hepatotoxicity ......................................... 55
2.3.5 Histopathology .......................................................... 56
2.3.6 Hepatocyte Isolation ................................................... 56
2.3.7 Polyrnorphonuclear leukocyte (PMN) isolation and
conditioned medium (CM) preparation .............................. 57
2.3.8 Effect of RAN and FAM on PMN-CM induced cytotoxicity.....57
2.3.9 Statistical analyses ...................................................... 58
2.4 Results ............................................................................... 58
2.4.1 Dose-ranging studies ................................................... 58
2.4.2 Development of hepatotoxicity after LPS/RAN cotreatment. . 59
2.4.3 Histopathological examination ....................................... 59
2.4.4 Comparison of RAN and F AM ....................................... 66
2.4.5 Effect of RAN on killing of hepatocytes by PMN-CM ............ 71
2.5 Discussion ........................................................................... 71

viii

CHAPTER 3

Gene expression analysis points to hemostasis in livers of rats cotreated with

lipopolysaccharide and ranitidine ............................................................... 80
3.1 Abstract .............................................................................. 81
3.2 Introduction ......................................................................... 82
3.3 Methods ............................................................................. 84

3.3.1 Materials ................................................................. 84
3.3.2 Animals .................................................................. 84
3.3.3 Experimental protocol .................................................. 84
3.3.4 Hepatotoxicity Assessment ............................................ 85
3.3.5 RNA isolation ........................................................... 85
3.3.6 Affymetrix gene chip analysis ........................................ 85
3.3.7 Data analysis and determination of gene activity .................. 86
3.3.8 Reverse transcription and real-time PCR ........................... 87
3.3.9 Immunohistochemistry ................................................ 89
3.3.10 Quantification of fibrin staining ...................................... 90
3.3.11 Evaluation of serum PAI-l concentration ........................... 90
3.3.12 Statistical analysis ...................................................... 90
3.4 Results ............................................................................... 91
3.4.1 Development of hepatic parenchymal cell injury .................. 91
3.4.2 Cluster analysis ......................................................... 91
3.4.3 Gene expression changes afier LPS, RAN, or LPS/RAN
treatment ................................................................. 96
3.4.4 Real-time PCRIOO
3.4.5 Evaluation of serum PAI-l ............................................ 105
3.4.6 LPS/RAN treatment and sinusoidal endothelial cells .............. 105
3.4.7 Hepatic fibrin deposition ............................................... 108
3.5 Discussion ........................................................................... 109
3.6 Supplemental data .................................................................. 119

CHAPTER 4

Augmentation of lipopolysaccharide-induced fibrin deposition by ranitidine

and its connection to idiosyncrasy—like liver injury in rats ................................... 121
4.1 Abstract .............................................................................. 122
4.2 Introduction ......................................................................... 123
4.3 Methods ............................................................................. 125

4.3.1 Materials ................................................................. 125
4.3.2 Animals. . . .' .............................................................. 125
4.3.3 Experimental protocol .................................................. 125
4.3.4 Hepatotoxicity assessment ............................................ 128
4.3.5 Determination of senun PAI-l, plasma fibrinogen and plasma
thrombin-antithrombin dimer (TAT) concentrations ............... 128
4.3.6 Fibrin and RECA-l immunohistochemistry ........................ 129
4.3.7 Morphometric evaluation .............................................. 130
4.3.8 Irnmunohisochemical evaluation of liver hypoxia .................. 130

ix

4.3.9 Statistical analysis ...................................................... 132
4.4 Results ............................................................................... 132
4.4.1 Coagulation system activation after LPS/RAN-treatment. . . . . 132
4.4.2 Altered sinusoidal endothelial cell function after LPS/RAN
treatment ................................................................ 133
4.4.3 Effect of LPS/RAN treatment on hepatic fibrin deposition ....... 137
4.4.4 Effect of LPS/RAN treatment on serum PAI-l concentration. . ..137
4.4.5 Effect of streptokinase on LPS/RAN-induced liver injury ........ 142
4.4.6 Effect of heparin on LPS/RAN-induced liver injury ............... 142
4.4.7 LPS/RAN-treatment and liver hypoxia .............................. 142
4.5 Discussion ........................................................................... 148
CHAPTER 5
Role of coagulation system activation and liver PMN accumulation in
LPS/RAN-treated rats ............................................................................. 162
5.1 Abstract ............................................................................. 163
5.2 Introduction ......................................................................... 164
5.3 Methods .............................................................................. 165
5.3.1 Materials ................................................................. 166
5.3.2 Animals .................................................................. 166
5.3.3 Experimental protocol .................................................. 166
5.3.4 Hepatotoxicity assessment ............................................. 167
5.3.5 Fibrin immunhistochemistry and quantification .................... 167
5.3.6 Evaluation of liver hypoxia ............................................ 167
5.3.7 Evaluation of hepatic PMN accumulation and ClNC-l
and elastase concentration ............................................. 167
5.3.8 Hepatocyte isolation .................................................... 167
5.3.9 Effect of hypoxia on elastase-induced cytotoxicity ................ 168
5.3.10 Statistical analysis ...................................................... 168
5.4 Results ................................................................................. 169
5.4.1 Effect of heparin on hepatotoxicity afier LPS/RAN treatment. . .169
5.4.2 Effect of heparin on liver hypoxia after LPS/RAN treatment. . .. 169
5.4.3 Serum CINC-l concentration and PMN accumulation in
livers of LPS/RAN-treated rats ....................................... 172
5.4.4 Effect of heparin on serum CINC-l concentration and PMN
accumulation in livers of LPS/RAN-treated rats ................... 177
5.4.5 Effect of hypoxia on PMN-E induced killing of rat hepatocytes. 1 77
5.5 Discussion ........................................................................... 182
CHAPTER 6
Summary and Conclusions ....................................................................... 187
6.1 Summary of conducted research ........................................ 188
6.2 Proposed mechanism ..................................................... 192
6.3 Significance of research, knowledge gaps, and future studies ...... 195
BIBLIOGRAPHY ................................................................................. 201

LIST OF TABLES
Table 1.1 Clinical features associated with reported cases of ranitidine
hepatotoxicity ........................................................................ 24

Table 1.2 Some conditions for which inflammation is a consequence
or contributor .......................................................................... 26

Table 1.3 Examples of inflammatory mediators important for LPS-induced

liver injury ............................................................................. 34
Table 1.4 Xenobiotics for which hepatotoxicity is augmented by LPS ................... 37
Table 1.5 The inflammation hypothesis of idiosyncratic drug reactions and its

relationship to typical features of idiosyncratic drug reactions ................ 45
Table 2.1 Midzonal hepatic necrosis after LPS/RAN cotreatment ........................ 67
Table 3.1 Expression pattern groups and gene product classification ..................... 100
Table 3.2 Genes with greater signal in livers from LPS/RAN-treated rats

compared to other treatments ....................................................... 102
Table 4.1 Serum HA concentration after LPS/RAN treatment ............................. 136

LIST OF FIGURES
Images in this dissertation are presented in color

Figure 1.1 Reactive metabolite, hapten, danger and multiple determinant
hypotheses of idiosyncratic hepatotoxicity from halothane (HAL)

and diclofenac (DCF) ............................................................. 19
Figure 1.2 Noninjurious endotoxemia increases sensitivity of rats to AFBl-induced

hepatotoxicity ....................................................................... 39
Figure 1.3 Factors influencing sensitivity of hepatocytes to TNF-induced

cytotoxicity ......................................................................... 43
Figure 1.4 Integration of the “inflammation hypothesis” with previously proposed

hypotheses for mechanisms of idiosyncratic hepatotoxicity ................. 49
Figure 2.1 Dose-ranging studies with RAN and LPS ....................................... 61
Figure 2.2 Hepatotoxicity from LPS/RAN cotreatment .................................... 63
Figure 2.3 Representative photomicrographs of liver after LPS/RAN

cotreatment .......................................................................... 65
Figure 2.4 Comparison of LPS/RAN and LPS/F AM cotreatrnents ....................... 70
Figure 2.5 Effect of RAN and FAM on killing of hepatocytes by PMN-CM ........... 73

Figure 3.1 Evaluation of hepatic parenchymal cell injury after
LPS/RAN-cotreatment ............................................................. 93

Figure 3.2 Hierarchical clustering of hepatic gene expression after LPS/RAN-
cotreatment ......................................................................... 95

Figure 3.3 Venn diagram depiction of probeset activity relative to

Veh/Veh-treated rats ............................................................... 98
Figure 3.4 Real-time PCR confirmation of gene array data ............................... 104
Figure 3.5 Serum plasminogen activator inhibitor-1 (PAI-l) concentration after

LPS/RAN treatment ............................................................... 107
Figure 3.6 Effect of LPS/RAN cotreatment on serum hyaluronic acid (HA)

concentration ....................................................................... 111
Figure 3.7 Effect of LPS/RAN cotreatment on hepatic fibrin deposition ................ 113

xii

Figure 4.1 Balancing of coagulation and fibrinolysis by the coagulation system ...... 127
Figure 4.2 Coagulation system activation afier LPS/RAN treatment ..................... 135
Figure 4.3 Quantification of RECA-l staining in livers afier LPS/RAN treatment. . .. 137
Figure 4.4 Liver fibrin deposition in LPS/RAN—treated rats .............................. 141
Figure 4.5 Serum concentration of PAH after treatment with LPS/RAN ............... 143
Figure 4.6 Effect of streptokinase (SK) on LPS/RAN-induced liver injury ............. 146
Figure 4.7 Effect of heparin on LPS/RAN-induced liver injury ........................... 151
Figure 4.8 Effect of LPS/RAN treatment on hepatic PIM-adduct staining .............. 153
Figure 4.9 Effect of LPS/RAN treatment on hepatic HIF-lor staining ................... 155
Figure 4.10 Effect of RAN on LPS-induced fibrin deposition and the connection

to liver injury ........................................................................ 161
Figure 5.1 Effect of heparin on hepatotoxicity after LPS/RAN treatment ............... 171
Figure 5.2 Effect of heparin on liver hypoxia in LPS/RAN-treated rats ................. 174
Figure 5.3 Serum CINC-l concentration and PMN accumulation in livers

of LPS/RAN-treated rats ........................................................... 176
Figure 5.4 Efl‘ect of heparin on serum CINC-l concentration and hepatic PMN

accumulation in LPS/RAN-treated rats ......................................... 179
Figure 5.5 Effect of hypoxia on Plvm-E-induced killing of rat hepatocytes ............ 181
Figure 6.1 Proposed mechanism of hepatotoxicity caused by LPS/RAN-cotreatment

in rats ................................................................................. 195

xiii

APAP
ADR
AFB.
AA
ANOVA
APC
ALT
ALP
AST
Btg-2
CL
CIM
CxcllO
COX-2
CYP
CINC-l
DCF
Egr-l
EU
ELISA
FDR
FAM
fMLP
GC
Gal
GGT
GI
GLUTl
GSH
H2
HAL
HA
HIF- 1 or
HPC
HPF
Igfbp-l
IL-l

Iv

1p
JNK
KC
LAL
LP

LIST OF ABBREVIATIONS

acetaminophen

adverse drug reaction
aflatoxin Bl

allyl alcohol

analysis of variance

antigen presenting cell
alanine aminotransferase
alkaline phosphatase
aspartate aminotransferase
b—cell translocation gene 2
centrilobular

cirnetidine

CXC chemokine ligand 10
cyclooxygenase-Z
cytochrome P450
cytokine-induced chemoattractant-l
diclofenac

early growth response-1
endotoxin unit
enzyme-linked irnmunosorbent assay
false discovery rate
famotidine

f-Met-Leu-Phe

gadolinium chloride
galactosamine
gamma-glutamyl transferase
gastrointestinal

glucose transporter l
glutathione

histamine2-

halothane

hyaluronic acid

hypoxia inducible factor-1a
hepatic parenchymal cell
high power field

insulin-like growth factor binding protein-1
interleukin 1

intravenous

intraperitoneal

c-Jun NH(2)-terminal kinase
Kupffer cell

Limulus arnoebocyte lysate
lipid peroxidation

xiv

LPS

LR
MCT

NF-KB
NGFI-B
NSA
LOX-l
PBS
PDE48
PIM
PAI-l
PAR
PMN
PMN-CM
PMN-E
PP

RECA- 1
SEC

SK
TAT

TFA
TGZ
TLR4

lipopolysaccharide
lipopolysaccharide/vehicle
vehicle/ranitidine
lipopolysaccharide/ranitidine
monocrotaline

mitochondrial permeability transition
nuclear factor kappaB

nerve growth factor-induced B
neoantigen-specific antibodies
lectinlike oxidized low-density lipoprotein receptor-1
phosphate-buffered saline
phosphodiesterase 4B
pimonidazole

plasminogen activator inhibitor-1
protease activated receptor
neutrophil
neutrophil-conditioned medium
neutrophil elastase

pefiportal

ranitidine

rat endothelial cell antigen-l
sinusoidal endothelial cell
streptokinase
thrombin-antithrombin dimmer
tissue factor

trifluoroacetyl

troglitazone

toll-like receptor 4

tumor necrosis factor-alpha

XV

Chapter 1

General Introduction

1.1 Idiosyncratic hepatotoxicity

1.1.1 Background and proposed mechanisms

Despite continued advances in predicting and understanding adverse drug
reactions (ADRs), they remain a large problem for patients and the pharmaceutical
industry alike. Untoward drug reactions contribute significantly to hospitalization and
mortality in the United States (Lazarou et al., 1998). In addition, the incidence of ADRs
is likely under-reported, with one study suggesting an incidence about l6-fold higher
than suggested by current reporting methods (Bagheri et al., 2000; Sgro et al., 2002). The
impact of ADRs goes beyond the obvious relationship to human health. Removal of a
drug from the market can cause a significant deficit in successful treatment of a disease
or condition. One example of this situation is the anti-epileptic felbamate (Pellock, 1999).
Felbamate is effective in treating severe cases of epilepsy, but its utilization in treatment
of epilepsy was reduced as a consequence of rare hepatotoxicity, and it is left now only
for those patients for which the therapeutic benefit outweighs the risk of toxicity (for
review, see Dieckhaus et al., 2002). Furthermore, ADRs affect the pharmaceutical
industry financially through loss of investment capital, future profits, and potential
financial loss from lawsuits. Overall, the consequences of ADRs only emphasize the
importance of studies aimed at understanding mechanisms and facilitating prediction.

ADRs affect numerous organ systems as well as circulating cells, with a frequent
target organ being the liver. In fact, one report estimated more than half of acute liver
failure cases occur as a result of ADRs (Gill and Sterling, 2001). Acute, drug-induced

liver injury is associated with several therapeutic/pharrnacologic classes of drugs and

may manifest as an array of histopathologic and clinical features including acute
hepatocellular necrosis, biliary injury, or a combination of the two (Zimmerman, 1978).
More than 25 years ago, Zimmerman described agents that caused liver toxicity as either
“intrinsic hepatotoxins” or as “nonpredictable” hepatotoxins (Zimmerman, 197 8),
alternatively defined as type A or type B ADRs, respectively (Pirmohamed et al., 1998).
Type B reactions are unpredicted reactions occurring in a small percentage of people
taking a drug (typically < 5%) and are therefore frequently termed “idiosyncratic.”
Contrasting intrinsically hepatotoxic drugs (type A ADR), which cause dose-dependent
hepatotoxicity over a predictable timecourse, the dose-response relationship for drugs
that cause idiosyncratic hepatotoxicity is obscured or absent, and reactions occur at
various times after the start of drug therapy (Zimmerman, 1993). Furthermore, unlike
type A reactions, for which hepatotoxicity might be due to an exaggerated pharmacologic
effect of the drug, type B reactions are seemingly unrelated to the drug’s pharmacology.
Overall, the features of idiosyncratic reactions make these reactions more dangerous,
more difficult to understand, and unfortunately, more difficult to predict.

A recent example of a drug associated with type B drug—induced liver toxicity is
that of troglitazone (Rezulin®), one of many thiazolidinediones marketed for treatment of
type 2 diabetes. The fiaction of patients that developed hepatotoxicity from troglitazone
treatment was relatively small (about 120,000) (Faich and Moseley, 2001). Nevertheless,
several cases of hepatotoxicity were documented in which the injtu'y was severe enough
to necessitate liver transplant or cause death of the patient (Kohlroser et al., 2000;
Murphy et al., 2000). Interestingly, the mechanism of liver injury appears to be

independent of the drug’s pharmacologic action, as other PPAR-gamma agonists used to

date lack the same propensity to cause idiosyncratic hepatotoxicity (Tolman and
Chandramouli, 2003). Another well—studied drug is acetaminOphen (APAP), a component
of numerous over-the-counter medications and a leading cause of drug-induced liver
failure. APAP typically produces dose-dependent hepatotoxicity, and several mechanisms
for the toxicity have been proposed (Kaplowitz, 2004b). However, the exact mechanism
of APAP-induced liver injury is still incompletely understood, and in some cases even
patients taking large doses of APAP do not develop liver injury (Tredger et al., 1995;
Shayiq et al., 1999). This has led to the proposal that even APAP hepatotoxicity has
characteristics of idiosyncratic hepatotoxicity (Kaplowitz, 2004b). Thus, it appears that
idiosyncratic reactions still occur even for drugs for which the mechanisms of
hepatotoxicity are relatively well understood.

Currently, there is no model for the accurate preclinical prediction of idiosyncratic
drug reactions. Drugs continue to carry “black box” warnings or suffer curtailed use after
being approved for market as a result of unpredicted human toxicity (Lasser et al., 2002).
Inasmuch as these responses are not typically observed until a large population of
patients has taken a drug, it is not surprising that they are not observed in small groups of
animals in laboratory experiments. Lack of statistical power and/or lack of necessary
susceptibility factors might both contribute to the inability to detect the potential for
toxicity. Interestingly, a recent paper by Alden et 01 examined the preclinical toxicology
studies of three drugs associated with idiosyncratic hepatotoxicity: bromfenac, zileuton,
and troglitazone (Alden, 2003). The authors questioned the rationale for the assumption
that idiosyncratic reactions cannot be predicted preclinically and offered evidence in the

case of each drug that subtle, but important changes in liver biomarkers might have

predicted the toxicity. Ultimately, the ability to predict these reactions during preclinical
development would allow the pharmaceutical industry to bring their safest candidates to
market.

Better prediction of idiosyncratic liver injury will require a better understanding
of the mechanisms underlying toxicity. Uetrecht and colleagues appropriately note that
“No one model fits the characteristics of all idiosyncratic drug reactions” (Seguin and
Uetrecht, 2003). Properties of the drug, the genetic background of the patient, and other
factors including disease and cell stress probably influence the incidence of idiosyncratic
reactions (Boelsterli, 2003a; Kaplowitz, 2001). Several hypotheses have been proposed to
explain the underlying basis of idiosyncratic reactions. For example, genetic
polymorphisms, resulting in altered expression of drug metabolizing enzymes might be
important for causing drug idiosyncrasy (Pirmohamed et al., 1996). Ultimately, the
“reactive metabolite hypothesis” suggests that changes in drug metabolism could result in
formation of a reactive metabolite or reduced detoxification of reactive intermediates. An
extension of the reactive metabolite hypothesis is the hapten hypothesis, in which a drug
or reactive drug metabolite covalently binds to proteins such that an antibody-driven
response is directed against “self” by the immune system, resulting in liver toxicity (Ju
and Uetrecht, 2002). Furthermore, Uetrecht and others have applied Matzinger’s Danger
Hypothesis to idiosyncratic drug reactions (Pirmohamed et al., 2002; Seguin and
Uetrecht, 2003). This hypothesis suggests that a second “danger signal,” in addition to the
presence of autoantibodies, is necessary to mount a specific immune response and the
consequent hepatotoxicity. This signal might be any number of factors including some

form of cellular stress, underlying disease, or other environmental factors. Although for

some drugs experimental evidence supporting these mechanisms is available, in some
cases the exact etiology remains controversial. Furthermore, the basis of idiosyncratic
reactions for the vast majority of drugs remains unclear. The following section will
critically review the supporting evidence, shortcomings, and some alternative thinking
surrounding underyling mechanisms of idiosyncratic hepatotoxicity for two different

drugs, halothane (HAL) and diclofenac (DCF).

1.1.2 Examples: HAL and DCF

HAL. Hepatotoxicity of the inhalation anesthetic HAL was identified more than
40 years ago as two different manifestations of hepatotoxicity. A mild elevation in serum
enzymes occurs in about 20% of exposed patients. However, a much more severe “HAL
hepatitis” occurs in a much smaller fraction of exposed patients (for review, see Gut et
al., 1993). Numerous animal models have been evaluated with the purpose of trying to
understand the mechanism of idiosyncratic HAL hepatotoxicity. Interestingly, animal
models have reproduced the variability in response observed in people, but not
necessarily all the histopathologic features or severity of lesions in people.
Complementing the body of work in laboratory animals, many epidemiological studies
examining HAL hepatotoxicity in exposed populations have provided mechanistic clues
about HAL idiosyncrasy. Despite this rather large body of literature, the exact mechanism
of idiosyncratic HAL hepatotoxicity remains elusive.

HAL undergoes oxidative metabolism by cytochromes P450 (CYPs) to form the
metabolite triflouroacetyl chloride (TFA; Gut et al., 1993). Hydrolysis of TFA-chloride

yields triflouroacetic acid, the primary HAL metabolite identified in mine of HAL-

exposed people (Stier, 1964). In addition, TFA can form adducts with several types of
cellular macromolecules, including proteins (Gut et al., 1993). CYP2E1 is probably the
primary isoforrn responsible for HAL metabolism in experimental animals and in people
(Kenna et al., 1990; Spracklin et al., 1997; Kharasch et al., 1996). Treatment of rat and
human hepatocytes with HAL results in the formation of TFA-modified proteins (Ilyin et
al., 1994; Van Pelt and Kenna, 1994), and increased metabolism of HAL by induction of
CYP activity is associated with increased formation of TFA-modified proteins (Kenna et
al., 1990). Furthermore, inhibition of CYP activity in people and experimental animals
results in a reduction in the formation of TFA-modified proteins (Kharasch et al., 1996;
Spracan et al., 2003). Overall, the data suggest that the formation of TFA-modified
proteins is probably a consequence of CYP-mediated generation of the reactive HAL
metabolite TFA. Interestingly, some studies have shown that P450 induction in animals
worsens HAL hepatotoxicity (Rice et al., 1987). Furthermore, experiments in outbred
guinea pigs, which manifest a range of hepatotoxic responses to HAL, have suggested
that TFA-modified proteins are a determinant of sensitivity to HAL-induced liver injury
(Bourdi et al., 2001).

As mentioned above, one possible consequence of protein modification by
reactive drug metabolites is recognition of a metabolite/protein hapten by the specific
immune system. In the case of HAL, the haptens formed are frequently referred to as
TFA-neoanti gens. The hapten hypothesis has a strong foothold in theories about HAL
idiosyncrasy, as several animal models have been utilized to study the specific immune
response after exposure to HAL, and numerous epidemiological studies have suggested

an antibody response occurs in people exposed to HAL.

Studies in animals have demonstrated that the liver is a target for the formation of
TFA-modified proteins (Satoh et al., 1985). Inasmuch as the liver is the primary site of
HAL metabolism, this is perhaps not surprising. How might the specific immune system
recognize TFA-neoantigens in the liver? In addition to their participation in the innate
immune response, several cell types in the liver can contribute to antigen presentation and
control of a specific immune response (Knolle and Gerken, 2000). One possibility is that
resident liver macrophages, otherwise known as Kupffer cells (KCs), process and present
TFA-neoanti gens to other immune cells. Indeed, studies in guinea pigs have detected
TFA-neoanti gens in KCs of HAL-treated guinea pigs but not in circulating monocytes or
in the spleen or lymph nodes (F urst et al., 1997). In addition, KC adhesion to
lymphocytes is increased in the presence of TFA-modified proteins or homogenatc from
livers of HAL-treated animals, and it is decreased in the presence of an anti-MHCII
antibody (Furst and Gandolfi, 1997). These data are consistent with the generation of an
immune response through processing and presentation of TFA-antigens by resident liver
antigen-presenting cells (APCs). However, the exact role of KC’s in HAL-induced
hepatotoxicity is unclear, as in rats, a species only sensitive to HAL hepatitis after CYP
induction and hypoxia exposure, TFA-neoantigens were observed in KCs in some, but
not all studies (Amouzadeh and Pohl, 1995; Christen et al., 1991; McLain et al., 1979).
Furthermore, studies examining the effect of KC inactivation on HAL hepatotoxicity in
animal models are lacking.

Other clinical features of idiosyncratic HAL hepatotoxicity are somewhat
suggestive of an immunologically based mechanism of hepatotoxicity. For example, there

is usually a delay between exposure to HAL and development of liver injury.

Furthermore, patients frequently respond more rapidly and robustly upon rechallenge,
suggesting immunologic memory. Indeed, TFA-neoantigen-specific antibodies (N SA)
have been identified in patients who developed HAL hepatitis (Kenna et al., 1987; Martin
et al., 1993b; Martin et al., 1993a; Pumford et al., 1993a; Vergani et al., 1980; Bird and
Williams, 1989; Njoku et al., 2002), that are not always found in sera fi'om unexposed
donors (Kitteringham et al., 1995; Bird and Williams, 1989), suggesting a causal role
between these antibodies and idiosyncratic HAL hepatitis. Overall, the observation that
circulating TFA-NSA are frequently associated with HAL hepatitis in patients probably
has led to the dogmatic association of a specific immune response with HAL
hepatotoxicity. Although these data are supportive, downstream pathways of
hepatotoxicity are not yet known, and controversy exists as to the role of a humoral
immune response in HAL-induced liver injury.

Some experiments in laboratory animals have not placed the same emphasis on
the humoral immune response as a mediator of HAL hepatitis. In one experiment, guinea
pigs given TFA-modified proteins prior to exposure to HAL had an enhanced antibody
response when exposed to HAL. However, this response was not accompanied by HAL
hepatitis, leading the authors to suggest that enhancing the humoral immune response is
not sufficient to cause HAL hepatotoxicity (Hastings et al., 1995). In addition, the
connection between circulating TFA-NSA and HAL hepatotoxicity in people is
controversial. In one study, Njoku et al. 2002 examined the serum of over 150
anesthesiologists, 20 HAL hepatitis patients, and 20 unexposed individuals for the
presence of these antibodies. Although autoantibodies were identified in the sera of

anesthesiologists, the vast majority did not develop HAL hepatitis (Njoku et al., 2002).

Another study by Kitteringham et al. also found autoantibodies in 3 of 6 patients exposed
to HAL that did not deveIOp HAL hepatitis (Kitteringham et al., 1995). In addition,
TFA-NSA are not always identified in patients who develop of HAL hepatitis (Bird and
Williams, 1989). The latter observation might be due to lack of assay sensitivity or
selection of antigen for detection. Nevertheless, these studies call into question the
relationship between autoantibodies resulting from HAL exposure and the manifestation
of HAL-induced idiosyncrasy. Further study is warranted to determine the exact role of
autoantibodies in HAL hepatotoxicity.

Although the focus on mechanisms of HAL hepatotoxicity has remained on
autoantibodies, several other hypotheses have received relatively little attention. For
example, the role of hypoxia in causing HAL hepatotoxicity has been explored and
debated. Oxygen tension seems to be important for determining the relevant role of
various CYP 450 isoforms in metabolism of HAL. As mentioned above, CYP 2E1 seems
to be important in oxidative metabolism of HAL to TFA. However, under low oxygen
tension, other CYP 450 isoforms might become important for reductive metabolism of
HAL to a free radical form (Spracklin et al., 1996). Studies in animal models and human
liver microsomes have demonstrated that lipid peroxidation (LP) occurs after HAL
exposure (Minoda and Kharasch, 2001; Yamazoe et al., 1998; Sato et al., 1990; Akita et
al., 1989; Younes et al., 1988; Wood et al., 1976). Interestingly, LP is enhanced under
conditions of hypoxia (Yamazoe et al., 1998; Younes et al., 1988; El-Bassiouni et al.,
1998; de and N011, 1984) and is further enhanced by pretreatment with the P450-inducer
phenobarbital (Awad et al., 1996). Furthermore, hypoxic rats exposed to HAL developed

liver injury afier phenobarbital treatment, but not after pretreatment with the CYP2E1

10

inducer izoniazid, calling into question the sole importance of CYP 2E1 in the toxicity
(Awad et al., 1996). Interestingly, glutathione (GSH) depletion has also been associated
with LP, both of which precede hepatic injury in guinea pigs exposed to HAL (Akita et
al., 1988). Moreover, GSH depletion enhances LP (Younes et al., 1988) and the hepatic
toxicity of HAL (Lind et al., 1994; Lind et al., 1992), suggesting that GSH status is an
important determinant of HAL hepatotoxicity. Other compounds, with antioxidant
effects, have provided protection seemingly unrelated to their effects on HAL metabolism
(Lind and Gandolfi, 1997). Overall, evidence in these models suggests that LP and GSH
depletion might be important for HAL-induced liver injury.

In summary, evidence supporting the involvement of the humoral immune
response in HAL-induced liver injury includes correlation of autoantibody formation with
the incidence of hepatotoxicity. However, this relationship has been called into question
by some studies, and other mechanisms of HAL-induced injury have also been proposed.
Observations including HAL-induced LP or GSH depletion might represent the missing
link in connecting autoantibodies to the development of HAL-induced injury, especially
if considered as danger signals in the context of the “danger hypothesis.” For example, in
the report by Kitteringham et al. 1995 in which autoantibodies were identified in patients
with HAL-induced liver injury, a second but necessary signal such as decreased
antioxidant capacity could be important. Thus, even for HAL, one of the “classic”
examples of immune-mediated drug toxicity, current hypotheses are controversial, and
the exact mechanism is not yet known.

DCF. Nonsteroidal anti-inflammatory drugs are used for the treatment of a variety

of inflammatory conditions including arthritis. Several drugs in this class, including DCF,

ll

are associated with infrequent hepatotoxicity in patients (Boelsterli et al., 1995). Small
increases in liver enzymes in DCF patients were noted in early clinical studies, but these
increases do not frequently progress to frank liver injury (Helfgott et al., 1990). The
incidence of more severe DCF hepatotoxicity in people has been estimated to be as low
as 0.001 % (Purcell et al., 1991; Walker, 1997), but this is probably an underestimate if
one considers the potential for underreporting mentioned earlier. The severity of DCF-
induced hepatotoxicity outcome is illustrated by a review of 250 published DCF case
reports in which the mortality rate was found to be about 10% (Lewis, 2003). Severe
DCF-induced liver injury is associated primarily with pronounced elevations in markers
of hepatocellular injury (Lewis, 2003), and the time to onset of severe DCF
hepatotoxicity ranges from 1 month to one year after the start of therapy (Banks et al.,
1995). Overall, DCF is an important example of idiosyncratic toxicity.

DCF metabolism by CYP 450s including CYP2C9 and CYP3A4, and by
glucuronidation results in formation of metabolites containing moieties capable of
binding proteins (for review, see Boelsterli, 2003b; Tang, 2003). Indeed, rat and human
liver microsomal systems are capable of metabolizing DCF to reactive metabolites, and
DCF protein adducts have been identified in cultured hepatocytes and in livers of rodents
treated with DCF (Kretz-Rommel and Boelsterli, 1994; Pumford et al., 1993b).
Therefore, in the case of DCF, not unlike HAL, reactive metabolites are produced, and
their adducts can be identified in vitro and in animals. However, the relationship of these
adducts to the toxicity of DCF is debated. For example, large concentrations of DCF can
induce cell death in cultured hepatocytes (Kretz-Rommel and Boelsterli, 1993) and

inhibition of CYP-mediated metabolism reduces DCF cytotoxicity in hepatocytes (Bort et

12

al., 1999; Kretz-Rommel and Boelsterli, 1993), but in one study inhibition of metabolism
was. without effect on DCF-adduct formation (Kretz-Rommel and Boelsterli, 1993).
Inhibition of UDP-glucuronosyltransferase activity decreased DCF-adducts but
significantly increased DCF-toxicity. Furthermore, identified polymorphisms in the
CYP2C9 gene do not appear to cause difl‘erences in DCF metabolism, and some studies
have failed to find an association between CYP2C9 polymorphism and DCF
hepatotoxicity (Yasar et al., 2001; Aithal et al., 2000). Overall, the relationship of DCF-
metabolites and their protein adducts to DCF-induced liver injury is unclear.

As described previously for HAL, one possibility is that DCF-modified proteins
are immunogenic and that an antibody-mediated response is responsible for DCF
hepatitis. Indeed, some case reports of DCF-induced liver injury are consistent with an
immune-mediated response, but not all (Boelsterli, 2003b). Antibodies directed against
red blood cells and platelets have been identified in some patients taking DCF (Bougie et
al., 1997; Sachs et al., 2004; Meyer et al., 2003), and antibodies against DCF-modified
liver proteins were identified in 7 patients with DCF-induced liver injury (Aithal et al.,
2004). Nevertheless, in this same study, antibodies were also identified in more than 50%
of patients taking DCF that did not have hepatotoxicity, suggesting the presence of the
antibodies was not sufficient to cause liver injury. In one interesting study, mice treated
with DCF conjugated to the immunogenic carrier, keyhole limpet hemocyanin (KLH),
developed antibodies against the DCF-modified protein. Splenocytes isolated fi'om these
animals were able to kill DCF-treated hepatocytes in culture. Furthermore, although
supematants from culttued Splenocytes fi'om KLH-DCF-treated mice were not sufficient

to injure hepatocytes, addition of naive Splenocytes to hepatocytes cultures given

13

supernatant resulted in cytotoxicity (Kretz-Rommel and Boelsterli, 1995). Despite these
intriguing results, antibody-mediated DCF hepatotoxicity has not been demonstrated in
animal models, and a clear-cut, immune-mediated mechanism for DCF-induced liver
injury in people has not been established.

Other effects of DCF on the liver and hepatocytes in culttue might reveal novel
pathways or danger signals for DCF-induced hepatotoxicity. For example, ATP content is
decreased in hepatocytes treated with DCF and is associated with opening of the
mitochondrial permeability transition (MPT) pore (Masubuchi etal., 2002). Indwd, DCF
can induce MPT pore opening in isolated mitochondria at relatively small concentrations.
Furthermore, antioxidants and inhibitors of MPT pore opening inhibit DCF-induced
apoptosis in cultured hepatocytes (Gomez-Lechon et al., 2003a; Gomez-Lechon et al.,
2003b). In and of themselves, these results provide insight into how DCF can kill cells in
vitro. However, administration of large doses of DCF to mice does not cause acute liver
injury, suggesting that if these mechanisms operate in viva, they are not sufficient to
cause liver injury. In this regard, other susceptibility factors must accompany
mitochondrial effects of DCF on hepatocytes.

Another possible mechanism of DCF hepatotoxicity that might merit attention is
related to accumulation of DCF metabolites in the bile canaliculi. Boelsterli and
colleagues demonstrated that transport of DCF glucuronides into the bile of rats was
dependent on activity of the canalicular pump mrp2 (Seitz et al., 1998). Moreover, DCF-
modified proteins were not identified in canalicular membranes of rats lacking this
transporter. The consequences of DCF-modified proteins in bile canaliculi is not clear,

but a decrease in activity of several resident canalicular enzymes was found in DCF-

14

treated rats (Sallustio and Holbrook, 2001). Canalicular transporters are important
regulators of drug and metabolite elimination from hepatocytes, and altered export of
reactive DCF-metabolites might be one determinant of DCF hepatotoxicity. In this
regard, it is interesting to note that genetic polymorphisms have been identified in the
mrp2 gene (Suzuki and Sugiyama, 2002; Kerb et al., 2001). Given the importance of
mrp2 in the elimination of DCF metabolites, characterization of these polymorphisms in
patients with DCF-induced liver injury is of interest. Although little is known about the
consequences of mrp2-mediated export of DCF metabolites on DCF-induced
hepatotoxicity, it is interesting that the mechanism of idiosyncratic hepatotoxicity of
other compounds might also relate to alterations in canalicular transport (Funk et al.,
2001).

In summary, mechanisms of DCF-induced hepatotoxicity are not fully
understood. Several studies have demonstrated that DCF treatment results in the
formation of reactive metabolites capable of binding to proteins, and antibodies to these
proteins have been identified in some patients with DCF hepatitis. However, these results
still cannot account for individual susceptibility to DCF hepatotoxicity. Induction of
oxidative stress and mitochondrial dysfunction by DCF are also of interest, but secondary
susceptibility factors are likely required for manifestation of hepatotoxicity. Interestingly,
for both DCF and HAL, about 20% of the people taking the drug experienced a mild
elevation in liver enzymes, but full-blown hepatitis develops only in a small fraction of
the patients. In light of the observations by Alden (2003) that subtle changes in serum
enzymes might be predictive of idiosyncratic reactions, these minor elevations might in

fact be predictive of more severe hepatotoxicity and perhaps are indicators of altered

15

hepatocellular homeostasis triggered by drug-induced oxidative stress, LP, or other

factors.

1.1.3 The “multiple determinant hypothesis.”

The basis for all idiosyncratic reactions is probably not a single mechanism and
therefore cannot be described by a single hypothesis or model. Figure 1.1 shows how this
may be the case for both DCF and HAL. Each agent is associated with the formation of
reactive metabolites, consistent with the reactive metabolite hypothesis of idiosyncrasy.
Autoantibodies have been identified in people taking each drug, suggesting that the
formation of drug-haptens is associated with the injury. Furthermore, consistent with the
danger hypothesis of immune-mediated idiosyncrasy, cell stressors such as LP and
mitochondrial dysfunction also occur after exposure to each of these agents. Gaps in each
of these individual hypotheses have been identified in this brief review: taken together,
this evidence paints a picture of multiple operative mechanisms.

Recently, a hypothesis has been proposed to explain the low occurrence of
idiosyncratic reactions in people. The “multiple determinant hypothesis” proposed by Li
states that the probability of an idiosyncratic reaction occurring for a drug is controlled by
the product of the probabilities of a plethora of factors including, but not limited to drug
exposure, environmental factors, genetic polymorphism, altered metabolism, formation of
antigens, and inadequate liver repair (Li, 2002). Thus, the primary emphasis of this
hypothesis is that one mechanism can’t describe all idiosyncratic reactions, and instead it
incorporates numerous possibilities (e.g., reactive metabolites and specific immune

response) into a model describing the probability of idiosyncratic reactions. Li suggests

l6

that both genetic and environmental factors also influence the overall probability of an
idiosyncratic reaction occurring. Identification of individual susceptibility factors such as
concurrent disease or environmental factors and development of models that can account

for such factors might provide insight into mechanisms of idiosyncratic drug reactions.

17

Figure 1.1: Reactive metaboliteJapten. da_nger. and multiple determinant hypotheses of
idiosyncratitflpatotoxicity from halogane (HAL) and diclofenac (DCF). DCF and HAL
are associated with several cell stressors that might be caused by either the drug or a
reactive drug metabolite. Reactive HAL and DCF metabolites are also capable of reacting
with proteins, and circulating autoantibodies have also been identified in patients with
hepatotoxicity from either agent. Cell stress induced by HAL or DCF might serve as
danger signals which, in the presence of neoantigen specific antibodies (N SA), may cause
hepatotoxicity. Cell stress might also cause hepatotoxicity independent of other factors.
Furthermore, numerous other susceptibility factors might be necessary triggers for
hepatotoxicity caused by cell stress or a specific immune response. Overall, this paradigm
suggests that several mechanisms might be operative in idiosyncratic hepatotoxicity from

DCF and HAL.

l8

 

Drug
(HAL, DCF)

I

Reactive
Metabolite

 

 

 

 

 

 

 

 

Cell Stressors
HAL DCF

 

Neoantigen-specific

 

antibodies

 

LP MitOA
JvGSH Apoptosis
Hypoxia Etc.

 

Etc. . C

 

 

Danger Signal

 

Other Determinants
{-9 Genetic polymorphism
H Environmental exposures
lOrgan repair
Drug coexposures

 

 

 

  

 

  

Drug
rechallenge

 

 

 

  
   

@

 

Hepatotoxicity

 

 

l9

1.1.4 Ranitidine and idiosyncratic hepatotoxicity

Ranitidine (Zantac®, RAN) is one of several histamine 2 (H2) receptor-
antagonists commonly used for the treatment of duodenal ulcers, gastric hypersecretory
diseases and gastroesophageal reflux disease. Side-effects with RAN as well as other H2-
antagonists including cimetidine (CIM, Tagamet®), famotidine (FAM, Pepcid®) and
nizatidine (Axid®) are rare, allowing transition of these drugs to over-the-counter
availability. Nonetheless, each carries an association with infrequent adverse effects in
people. RAN is associated with several adverse effects, all of which occur in a very small
fraction of treated patients (Vial et al., 1991). One of these is idiosyncratic hepatotoxicity.
The incidence of idiosyncratic liver injury for RAN seems to be less than that of CIM,
whereas reports of liver injury associated with F AM are few and suspect (Garcia et al.,
1997; Luyendyk et al., 2003b). Mechanisms underlying RAN -induced liver injury in
people remain unclear.

Although one H2-antagonist, oxmetidine, is hepatotoxic and produces
concentration- and time-dependent cytotoxicity in isolated hepatocytes (Oldharn et al.,
1985; Rush et al., 1985), millimolar concentrations of RAN, CIM, and FAM did not
produce cytotoxicity in isolated hepatocytes (Francavilla et al., 1989; Zimmerman et al.,
1986). Additionally, RAN is not hepatotoxic when given to rats at large doses.
Accordingly, RAN does not fit the classical paradigm (Le, a type A hepatotoxicant) of
producing liver injury in a dose and time-dependent manner.

RAN is, however, as mentioned above, associated with idiosyncratic
hepatotoxicity in people. Elevations in liver enzymes are occasionally described during

RAN treatment, and the incidence of RAN-associated liver injury in people has been

20

estimated to be less than 0.1% of those taking the drug (Vial et al., 1991; Mills et al.,
1997), and has not always been apparent, even in large clinical trials. To characterize
better the features of RAN ~induced liver injury in people, a literature search was
conducted using National Library of Medicine PubMed. This search identified 36 case
reports of RAN-induced liver injury in patients or subjects in clinical trials. The results
from this evaluation are displayed in Table 1.1. The age range identified for RAN-
induced liver injury was from 19-82, with 19 males, 12 females, and 5 cases for which
the patient’s gender was not reported. The average time to onset (or report) of liver injury
relative to start of RAN therapy was 4 weeks, with some responses occurring as early as
1 week and others 8 months after the patient started RAN therapy. Accordingly, no
obvious pattern of age, gender, or time to onset was identified for RAN-induced
idiosyncratic hepatotoxicity.

Injury to hepatic parenchymal cells and cholestasis were described in the majority
of cases, as estimated by changes in clinical chemistry and in some cases,
histopathological observations. Elevations in serum alanine aminotransferase (ALT) and
aspartate aminotransferase (AST) activities were typically more robust as compared to
markers of cholestasis (GGT, gamma-glutarnyl transferase; ALP, alkaline phosphatase;
bile acids). Nonetheless, RAN has been associated with cholestatic injury without
elevations in markers of parenchymal cell injury (Lee et al., 1986; Rarnrakhiani et al.,
1998; Coutellier et al., 1993). Liver biopsy samples were taken in some cases for
evaluation of liver histopathology. Hepatocellular necrosis and inflammatory cell
infiltration were frequently noted and accompanied occasionally by cholestasis, bile

pigment in KCs and hepatocytes, and intralobular eosinophil infiltration. It is important to

21

recognize that evaluation of clinical chemistry and liver histopathology in these cases
might not represent the time of onset or peak of toxicity since patients likely present with
symptoms at various times. For example, in one case a liver biopsy was taken 2 weeks
after elevations in clinical chemistry markers were observed (Rarnrakhiani et al., 1998).
The authors note that eosinophilic infiltration of the liver might have occurred
secondarily to hepatotoxicity, bringing into question a causal role for these cells in the
toxicity. Overall, hepatocellular oncotic necrosis and infiltration of mixed inflammatory
cells seem to be primary features of RAN-induced liver lesions, with some evidence for a
biliary component.

An interesting feature of idiosyncratic reactions is that rechallenge with the drug
does not always lead to production of the idiosyncratic reaction. Rechallenge is
performed only infrequently for obvious reasons, but in some cases it can be revealing.
As shown in Table 1.1, 7 case reports describing the results of a RAN rechallenge were
identified. Upon rechallenge with RAN, 4 patients experienced the idiosyncratic reaction
a second time. However, this was not true of all the cases. For example, Graham et a1,
(1985) rechallenged a patient with RAN at a dose “sufficient to reproduce the changes”
without any change in ALT or AST. In addition, RAN therapy was not discontinued
despite elevation in serum enzymes (see Drug not stopped, Table 1.1) in two cases. In the
report by Barr and Piper (1981), elevations in serum ALT resolved despite continued
therapy, while GGT activity did not normalize until RAN treatment was discontinued.
Another report by Colin-Jones (1984) described another case of suspected RAN

idiosyncrasy where continued RAN therapy was not associated with worsening of

symptoms.

22

One potential commonality among many cases of RAN idiosyncrasy was revealed
by conducting this literature search. Each case report was examined for prodromal signs
of inflammation, including fever, abdominal pain, nausea, vomiting, diarrhea,
illness/infection, or others factors (e.g., ethanol consumption or recent surgery) known to
increase the circulating concentration of inflammagens. One or more of these indicators
was reported in greater than 60% of the cases. This result on its own does not prove a
definitive connection between inflammation and RAN hepatotoxicity, and detailed
epidemiological investigation would be required to establish such an association.

However, it is consistent with inflammation as a potential susceptibility factor.

23

 

Hugo : 0:38: moweaom 88938 «39 8 non 858 o». W32 :0 Roxana . > 50388 8263 28 83:38 55m ZES
3388 8 50st 88.4. o». E EmOmvfiano rouaofiofiom? wanna c8330? Sue—88 o». rovaSSxmomQ. 3a 099. 355.5
amine.— casum 208 :28. 008390 83.258 85 3 mafia 5 9a emu—mega? Emu—n mumoom SE88 93 m @765 Sommfio 28 no”

manage. E8889 SEQ 80858.

24

1.2 Inflammation

1.2.1 Incidence and triggers

Individual susceptibility factors to drug-induced toxicity are probably not
represented in animal models of drug toxicity (e.g., underlying disease). This section will
provide evidence for the possibility that inflammation is one such factor. Our
understanding of inflammation and the effects of inflammatory mediators on tissues has
stretched beyond “a localized protective reaction of tissue to irritation, injury, or
infection, characterized by pain, redness, swelling. . .” (The American Heritage®
Dictionary of the English Language, Fourth Edition) to include inflammatory cells, the
mediators they produce, as well as altered cellular signaling and gene expression.
Although inflammation is important for warding off infection, it probably has other
protective and sometimes deleterious effects as well. Indeed, as we learn more about
inflammatory processes, it has become clear that inflammation, both acute and chronic,
plays a role in the pathogenesis of many diseases, can cause tissue injury, and can
increase sensitivity of tissues to the toxic effects of xenobiotics.

Inflammation is a feature of numerous diseases, and may be identified as either an
innocent bystander or as a contributor to disease outcome. Several obvious examples for
which inflammation is involved in disease are conditions for which treatment focuses on
inhibition of inflammatory mediators (e. g, arthritis, asthma, inflammatory bowel disease)
However, inflammation is now recognized as a component of several other conditions,
including those affecting the cardiovascular system to cancer. Some of these are listed in

Table 1.2.

25

Table 1.2:

Some conditiong for which inflammation is a consmuence or contributor

 

 

 

 

 

 

 

 

 

 

 

 

 

Condition For review, see:

viral hepatitis (N akamoto and Kaneko, 2003)

Parkinson’s disease (Barcia et al., 2003)

diabetes (Tracy, 2003)

acute stroke (Price et al., 2003)

obesity ‘ (Cottarn et al., 200$

multiple sclerosis (Martino et al., 2002)

pulmonary hypertension (Tuder and Voelkel, 1998)

cancer (Platz and De Marzo,
2004;Whitcomb, 2004)

cardiovascular disease Q’illerson and Ridker, 2004)

HIV (Dalgleish and O'Byme, 2002)

atherosclerosis (Lind, 2003;Shishehbor and
Hazen, 2004)

periodontitis (Van Dyke and Serhan, 2003)

 

 

osteoporosis

 

(Pfeilschifter et al., 2002)

 

 

As an example, inflammation is a recognized component of obesity. Adipocytes
actively produce several factors, including pro-inflammatory mediators such as cytokines
and lipid-derived products (for review, see Cottam et al., 2004). Correlations of the acute
phase protein, C-reactive protein, and obesity have been observed in numerous
epidemiological studies (V isser et al., 2001; Visser et al., 1999). In addition, weight loss
is associated with a decrease in circulating concentrations of pro-inflammatory mediators
(Cottam et al., 2004). Interestingly, obese animals are also more sensitive to
inflammatory liver injury (Yang et al., 1997). Accordingly, obesity is associated with
inflammation, and might be a condition sufficient to render tissues sensitive to injury
from inflammatory mediators. Indeed, the inflammatory component of obesity might be a
determinant of obesity-related disease (e.g, diabetes, heart disease) (Cottam et al., 2004).
Overall, inflammation accompanies several different diseases, either as a consequence or
a key player.

Bacterial lipopolysaccharide (LPS, endotoxin) is an outer cell wall component of
gram-negative bacteria and a potent inflammagen in people. Exposure to large amounts
of LPS, such as during Gram-negative bacterial sepsis, can result in damage to several
organs, including the liver, as well as mortality (Hewett and Roth, 1993). LPS activates
toll-like receptors, resulting in a cascade of inflammatory events. This makes it an
effective and commonly used model inflammagen in animal studies. Human exposure to
small amounts of LPS is commonplace and can occur by multiple modes, including
environmental exposure and translocation of LPS released by dividing or dying resident

gut flora across the gastrointestinal (GI) mucosa into the circulation. Exposure to small

27

amounts of LPS might be beneficial to the host, but it might also increase susceptibility to
other toxicities (see section 1.2.3).

Inhalation of environmental LPS occurs by occupational exposure to grains or
dusts and from contaminated household air (Lane et al., 2004; Gereda et al., 2001).
Contamination of household air with LPS can vary with season and is dependent on
several factors including household pets and central air conditioning (Gereda et al., 2001;
Park et al., 2000; Heinrich et al., 2001). Under several conditions, enhanced exposure to
LPS in humans also occurs by translocation of endogenous GI tract LPS into the portal
circulation (for review, see Roth et al., 1997). For example, liver and GI diseases are
associated with increased circulating LPS (Gardiner et al., 1995), as are surgeries in
which the GI tract is disrupted. Changes in diet, alcohol consumption, and antibiotic
treatment are also associated with increased LPS translocation (Roth et al., 1997; Lepper
et al., 2002). Interestingly, even strenuous exercise has been shown to cause mild
endotoxemia in humans and other species (Ashton et al., 2003; Jeukendrup et al., 2000;
Camus et al., 1998; Camus et al., 1997). Overall, exposure of people to LPS is episodic
and commonplace, and can occur under numerous conditions in the absence of any acute

tissue injury.

1.2.2 LPS-induced liver injury.

Inasmuch as episodes of inflammation occur commonly in people, either as a
component of a disease or in response to inflammagens like LPS, it is likely that
simultaneous exposure to xenobiotics, including drugs, occurs in some people. The

following section briefly discusses the ability of LPS to incite an inflammatory response

28

and the current understanding of mediators important for liver injury from large doses of
LPS. Furthermore, studies evaluating the effect of xenobiotic/inflammagen coexposure
on liver injury will be reviewed, and the role of inflammatory mediators in this
interaction discussed.

LPS activation of toll-like receptors (TLRs) on several cell types in the liver leads
to changes in gene expression, altered sinusoidal endothelial cell homeostasis, activation
of the coagulation system, accumulation and/or activation of inflammatory cells such as
macrophages (KCs resident in liver), neutrophils and platelets, as well as production of
cytokines such as IL-1 and TNF and other soluble factors including lipid mediators. This
inflammatory response is in turn governed by both positive and negative regulation.
Among this laundry list of inflammatory factors, numerous cell types and pro-
inflammatory mediators have been identified as critical mediators of LPS-induced liver
injury. The summary below pertains primarily to hepatic events that occur after a large,
hepatotoxic dose of LPS:

_K_Cs:_ KCs might be classified as “first responders” to LPS in the liver. Activation
of Toll-like receptor-4 on KCs is thought to be one of the earliest events triggered by LPS
in the liver. This results in production of inflammatory mediators, including reactive
oxygen species, cytokines, etc. Inactivation of KCs with gadolinium chloride (GC)
attenuates liver injury in rats from a hepatotoxic dose of LPS (Pearson et al., 1996a;
Vollmar et al., 1996) but is without effect on circulating fibrinogen (Pearson et al.,
1996a), suggesting that thrombin generation occurs by KC-independent mechanisms.
However, KC inactivation does result in a decrease in hepatic accumulation of platelets.

Stimulation of KCs with LPS results in production of TNF (Su et al., 2000), and

29

inhibition of KCs with GC reduced plasma TNF concentration in some (V ollmar et al.,
1996), but not all studies (Pearson et al., 1996a).

_TN_E TNF production is triggered in several cell types by exposure to LPS.
Hepatic mRNA and plasma TNF concentration are increased early after exposure to LPS
(Hewett and Roth, 1993). TNF-neutralizing antibodies and inhibition of TNF synthesis
significantly attenuate LPS-induced injury in rodents (Hewett et al., 1993). After an
hepatotoxic dose of LPS, inhibition of TNF biosynthesis or neutralization of TNF does
not influence hepatic PMN accumulation, suggesting either that TNF injures the liver
after exposure to LPS by a neutrophil-independent mechanism or that TNF affects a
critical firnction of PMNs (e.g., activation and release of cytotoxic proteases) (Hewett et
al., 1993).

Platelets: Platelet depletion prior to LPS exposure significantly reduces LPS-
induced activation of the coagulation system and liver injury (Pearson et al., 1995). A
reduction in the number of circulating platelets was not associated with a change in
hepatic PMN accumulation (Pearson et al., 1995).

Ms; PMN accumulation occurs in liver rapidly after treatment with LPS.
Hepatic gene expression of the chemokines, cytokine-inducible neutrophil
chemoattractant-l (CINC-l) and macrophage inflammatory protein-2 (MIP-2), is
increased after LPS exposure, and this is mediated by effects of cytokines such as IL-1
(Calkins et al., 2002). Activated PMNs release a variety of cytotoxic factors that can kill
hepatocytes in vitro including ROS and the lysosomal proteases, cathepsin G and elastase
(Ho et al., 1996; Ganey et al., 1994; Mavier et al., 1988). Neutrophil depletion or

inhibition of hepatic PMN accumulation by neutralization of these chemokines

30

significantly attenuates liver injury (Li et al., 2004; Hewett et al., 1992). Furthermore,
inhibitors of PMN elastase reduce hepatocellular damage in models of LPS-induced liver
injury (Ishii et al., 2002).

Despite these results, the role of PMN-derived proteases has been debated. Liver
injury deve10ps in LPS-treated rodents within 6 h, whereas PMN protease killing of
hepatic parenchymal cells (HPCs) takes longer to manifest in vitro (Ho et al., 1996).
Accordingly, other PMN-derived factors might act in concert with proteases to kill
parenchymal cells. For example, an inhibitor of PMN NADPH oxidase activity
significantly reduced liver injury in galactosamine-sensitized rats given LPS. This
inhibition was associated with a decrease in chlorotyrosine staining (indicator of
myeloperoxidase activity), suggesting that PMN-derived reactive oxygen species (ROS)
are important for injury in galactosamine (gal)/LPS-cotreated rats (Gujral et al., 2004).
However, PMNs stimulated with MLP plus PMA to release proteases and ROS,
respectively, do not kill I-IPCs with a timecourse different from that of proteases alone
(Ganey et al., 1994). This result suggests that an interaction of proteases with other
inflammatory mediators, with other liver cell types, or with some change in
hepatocellular homeostasis not represented in vitro is necessary for full manifestation of
LPS-induced liver injury.

Hemostatic system: Treatment of rodents with LPS causes activation of the
coagulation system (Hewett and Roth, 1995; Akahane et al., 2001), but the exact
mechanism of LPS-mediated coagulation activation is not yet known. Several studies
have suggested a role of tissue factor (TF) in activation of the coagulation system afier

LPS exposure. LPS exposure induced TF (for review, see Mackman, 1996) and it is

31

stored in circulating platelets (Engelmann et al., 2003). When expressed on the cell
surface, it can activate the coagulation system through interactions with coagulation
factor VIIa. Mice deficient in tissue factor expression were less susceptible to LPS-
induced coagulation activation (Pawlinski et al., 2004). Furthermore, expression of
human TF pathway inhibitor in endothelial cells, platelets and monocytes dramatically
reduced coagulation system activation and fibrin deposition in a mouse model of
endotoxemia (Chen et al., 2004). Overall, the expression of TF after LPS exposure
appears to be important for LPS-induced thrombin generation.

Inhibition of thrombin activation significantly attenuates LPS-induced liver
injury (Pearson et al., 1996b; Moulin et al., 2001), but this protection is independent of
circulating fibrinogen (Hewett and Roth, 1995). Although fibrin clots are observed in
liver after treatment of rodents with LPS, this result suggests that thrombin activation is
important for liver injury fi'om LPS, but that the formation of fibrin clots is not required
for the injury. Thrombin has activities outside the traditional scope of the hemostatic
system. For example, thrombin-mediated activation of its receptor, protease activated
receptor-l (PAR-1), influences inflammatory cell accumulation and/or activation in the
liver. Although inhibition of thrombin activation did not attenuate accumulation of PMNs
in liver after LPS-treatment, it reduced PMN activation as estimated by increased plasma
elastase concentration (Pearson et al., 1996b; Copple et al., 2003). Furthermore, perfusion
of livers from LPS-treated rats with thrombin or the PAR-1 agonist TFLLR caused
hepatocellular injury (Moulin et al., 2001;Copple et al., 2003), an effect eliminated by

prior depletion of PMNs (Moulin et al., 2001). Overall, these results suggest that

32

thrombin-mediated activation of PAR-l is important for PMN activation and that these
two events are sufficient for injuring livers of LPS-treated rats.

Overall, several inflammatory mediators are associated with LPS-induced liver
injury including KCs, TNF, PMNs and their chemokines, platelets, and thrombin. These
are summarized in Table 1.3. Much remains to be learned about how these mediators
network to cause liver injury from LPS. Moreover, though not discussed in detail here,
changes in the activation state of other cell types such as SECS and stellate cells might be
important. Interestingly, even though thrombin activation is paramount for LPS
hepatotoxicity, its action is seemingly unrelated to coagulation and fibrin deposition;
rather, thrombin is important for receptor-mediated activation of PMNs and perhaps other

inflammatory cells.

33

Table 1.3: Examples of inflammatory mediators important for LPS-induced liver injury.
Several inflammatory cells, the mediators they produce, and components of the
hemostatic system cause events critical to LPS-induced liver injury. For example, KCs,
neutrophils, chemokines, platelets, TNF, and the coagulation system are all involved in

LPS-induced injury.

 

 

 

 

 

 

 

 

 

 

. Reference demonstratin
Mediator contribution to liver injurg'y

TNF Q—lewett et al., 1993)

PMN chemokines (Li et al., 2004)

PMNs (Hewett et al., 1992;Gujral et
al., 2004;Ishii et al., 2002)

Platelets (Pearson et al., 1995)

Thrombin (Hewett and Roth,
1995;Moulin et al., 1996)

KCs (Pearson et al., 1996a;Vollmar
et al., 1996)

 

34

1.2.3 LPS-induced sensitivity to hepatotoxicity

A robust inflammatory response, such as that observed after exposure to a large,
hepatotoxic dose of LPS, can result in frank injury to organs including the liver. By
contrast, smaller doses of LPS produce a modest, but noninjurious inflammatory
response. Inasmuch as inflammatory mediators can alter tissue homeostasis, our group
and others tested the hypothesis that a modest underlying inflammatory response can
render the liver sensitive to hepatotoxic effects of xenobiotics. One way to evaluate this
hypothesis is to examine the sensitivity of animals to xenobiotic-induced liver injury in
the face of normally noninjurious inflammation. Accordingly, several animal models
have been developed in which rats are treated with a small dose of inflammagen
(typically LPS) to elicit a nonhgmtotoxic inflammatory response and then are cotreated
with a nonhepatotoxic dose of a xenobiotic agent. LPS coexposure augments the
hepatotoxicity of numerous organic chemicals, heavy metals, and some drugs (Table 1.4).
Although this discussion focuses on models of LPS-augmented hepatotoxicity, LPS has
also been shown to enhance extrahepatic toxicity of some chemicals (Wagner et al.,
2001; Rumbeiha et al., 2000; Zhou et al., 2000). Overall, these studies demonstrate that
an underyling inflammatory response has the potential to lower the threshold for
xenobiotic toxicity. Furthermore, the results suggest that inflammation might be a
determinant of sensitivity to chemically induced liver injury in people.

The effect of inflammation on hepatotoxicity has been investigated in several
experiments by examining the dose-response relationship in the presence or absence of
inflammation. In one model, rats are cotreated with a small, nonhepatotoxic dose of the

fungal toxin aflatoxin B1 (AF B1) and a small dose of LPS. Whereas liver injury does not

35

develop in rats treated with either agent alone, AF Bl/LPS-cotreated rats develop marked
hepatotoxicity characterized by periportal hepatocellular oncotic necrosis and injury to
intrahepatic bile ducts (Barton et al., 2000b). At the timepoint of maximal injury in
AFBlfLPS-treated rats, sensitivity to AF Bl-induced liver injury increases 10-20 fold in
rats coexposed to a small dose of LPS (Luyendyk et al., 2002; Fig 1.3). A leftward shift
in the xenobiotic dose-response curve has also been observed in other models of LPS-
augmented hepatotoxicity (Sneed et al., 1997; Yee et al., 2000). Overall, these
experiments suggest that an underlying inflammatory response has the potential to
decrease the threshold for xenobiotic-induced hepatotoxicity in rats.

Animal models of LPS-potentiated liver injury have not only suggested that
inflammation is a determinant of sensitivity for liver injury but have provided models in
which the specific mediators of toxicity can be evaluated. Regarding mechanisms of
toxicity in these models, one possibility is that the xenobiotic agent has the ability to
directly render hepatocytes more sensitive to a normally nontoxic inflammatory response.
For example, the cytotoxicity of allyl alcohol (AA) in cultured rat hepatocytes is
augmented by cotreatment with a prostaglandin released by activated KCs (Maddox et
al., 2004). Another possibility is that the xenobiotic can boost the LPS-induced
inflammatory response such that it exceeds the threshold for inflammation-induced
hepatotoxicity. One example of this is the ability of the plant toxin monocrotaline (MCT)
to activate PMNs (Yee et al., 2003c). In all likelihood, both of these mechanisms are

operative in models of LPS-potentiated hepatotoxicity.

36

Table 1.4: Xenobiotics for which h_epatotoxicity_i§_rrirgrnented b1LPS.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Xenobiotic Reference
Aflatoxin B1 (Barton et al., 2000b)
Monocrotaline (Yee et al., 2000)

Allyl alcohol (Sneed et al., 1997)
Carbon tetrachloride (Charnulitrat et al., 1995)
Galactosarnine (Bahrami et al., 1994)
Ethanol (Hansen et al., 1994)
T2-toxin (Tai and Pestka, 1988)
Cadmium (Cook et al., 1974)

HAL (Lind et al., 1984)
Cocaine (Labib et al., 2002)
Chlorpromazine (Buchweitz et al., 2002a)
Lead (Honchel et al., 1991)
PCBs (Brown et al., 1996)
TCDD (Patterson et al., 2003)

 

 

37

 

Figure 1.2: Noninjurious endotoxemia increa_ses sensitivity of rat_s to AF Bl-induced
hefltotoxicig. Rats given a small, nonhepatotoxic dose of LPS are rendered lO-20-fold
more sensitive to AF Bl-induced hepatotoxicity (Luyendyk et al., 2002). Hepatotoxicity
was estimated by increased serum alanine aminotransferase (ALT) activity.

*Significantly different from respective group not given AF B1.

38

a.
O
O
O
1

 

 

—e— LPS
+ Veh *

1000 ' *
800 r
600 1
400 ' *
200 ‘

l
o ”I. V V U 'I'U'l V

0.0 0.1 1.0 10.0
Aflatoxin B1 Dose (mg/kg)

..aa
233
GO

 

 

 

 

 

 

Serum ALT Activity (UIL)

39

Identification of inflammatory mediators involved in models of LPS-potentiated
hepatotoxicity has revealed a great deal of commonality with mechanisms involved in
liver injury from large, hepatotoxic doses of LPS. For example, KCs are important for
liver injury in rats cotreated with LPS and monocrotaline (Yee et al., 2003a). TNF has
also been identified as a mediator of hepatotoxicity in some (Barton et al., 2001 ;Yee et
al., 2003a), but not all models (Sneed et al., 2000) of LPS-potentiated hepatotoxicity.
Similarly, cyclooxygenase 2 (COX-2) is important for the toxicity from AA/LPS-
cotreatrnent (Ganey et al., 2001), but despite its augmented expression in AF B1/LPS-
treated rats, is not important for hepatotoxicity (Barton et al., 2001). Thus, augmented
expression of an inflammatory mediator does not prove its involvement in liver injury.
Interestingly, some common ground can be found across models of LPS-augmented
hepatotoxicity despite the differences described above. For example, PMNs are important
for liver injury in AFBl/LPS, MCT/LPS, and AA/LPS cotreated rats (Barton et al.,
2000a; Yee et al., 2003c; Kinser et al., 2004). Furthermore, activation of the coagulation
system occurs in each of these models, and anticoagulation provides marked if not
complete protection against the liver injury (Luyendyk et al., 2003a; Yee et al., 2003d;
Kinser et al., 2002). Overall, mediators important for LPS/xenobiotic-induced liver injury
likely depend on the xenobiotic agent.

One question at the core of these experiments is what critical change in
hepatocellular homeostasis (e.g., signal transduction, gene expression) is initiated by
coexposure of the cell to a particular inflammatory factor and xenobiotic? That is, what
effects does a particular xenobiotic have on cells (e.g., hepatocytes) that render them

sensitive to deleterious effects of inflammatory mediators? One example of how altered

40

HPC homeostasis can influence the outcome of exposure to inflammatory mediators is
TNF-induced cytotoxicity in hepatocytes. Hepatocytes are not sensitive to TNF-induced
cell death in vitro unless sensitized by one or more changes in signaling. For example,
inhibition of NF-KB function is associated with an increase in the activity of c- Jun NH-
terminal kinase (JNK) that renders hepatocytes sensitive to TNF-induced cytotoxicity
(Jones et al., 2000; Liu et al., 2002). This increase in JNK activity may be due to lack of
upregulation of negative INK regulators by NF-KB. Consistent with this observation,
JNK and c-Jun activity are required for TNF-induced cytotoxicity under conditions of
NF-xB inhibition (Schwabe et al., 2004; Liu et al., 2002). Depletion of GSH also renders
hepatocytes sensitive to TNF-induced cytotoxicity and is associated with decreased NF-
1cB activation and increased JNK activity (Nagai et al., 2002; Matsumaru et al., 2003).
These studies illustrate how alterations in signaling pathways can influence the sensitivity
of hepatocytes to TNF-induced cytotoxicity (Fig. 1.3). Similarly, xenobiotics capable of
1) depleting GSH, 2) inhibiting NF-tcB activation, and/or 3) activating INK pathways
might be expected to sensitize hepatocytes to TNF-induced cytotoxicity. The example of
TNF is only one of several possible ways that cellular alteration could interact with LPS-
generated inflammatory mediators to result in toxicity. Another mode of xenobiotic/LPS

interaction is the focus of the thesis research.

41

Figure 1.3: Factog influencing sensitivity of hepgtocytes to TNF-induced cytotoxicity.

Activation of cell death pathways involving JNK is important for TNF-induced killing of
hepatocytes. Activation of TNFRI receptors by TNF results in activation of NF-kB-
mediated transcription and prevention of JNK activation. Inhibition of NF-kB
translocation (B) leads to an increase in JNK activity and sensitizes the hepatocytes to
TNF-induced cell death. GSH depletion (A) is associated with decreased NF-kB
activation, an increase in JNK activity, and sensitization of hepatocytes to TNF-induced
killing. Agents modulating JNK activity (C) can also modulate TNF-mediated cell death
signaling. Xenobiotics that alter cell signaling at points A, B, or C might increase

sensitivity to TNF-induced cytotoxicity.

42

 

 

 

 

 

 

 

 

 

 

 

 

 

 

43

1.3 Inflammation-induced idiosyncrasy: a paradigm for hepatotoxicity

The previous section presented data supporting the hypothesis that inflammation
increases the sensitivity of the liver to hepatotoxic effects of xenobiotics. Furthermore,
potential molecular effects of inflammatory mediators were discussed in the context of
increasing sensitivity of target cells to xenobiotic toxicity. We have hypothesized that
underlying inflammatory responses might be a determinant of sensitivity to ADRs and
could precipitate some idiosyncratic responses to drugs (Roth et al., 2003). Modest,
episodic exposures to inflammagens (e.g. LPS), although noninjurious by themselves,
might bring about a leftward shift in the dose-response curve for drug-induced
hepatotoxicity. If this leftward shift is as dramatic as in the case of AFB. (Fig. 3), it is
conceivable that a normally therapeutic (and noninjurious) dose of a drug could be
rendered hepatotoxic in the face of an inflammatory response.

An interesting point about the inflammation hypothesis is that it can adequately
describe the features classically associated with idiosyncratic reactions. This is
summarized in Table 1.5. For example, in idiosyncratic reactions the affected organ is
typically not the pharmacologic or therapeutic target. Why might the liver be a common
target for idiosyncratic reactions under the inflammation hypothesis? Not only is the liver
the site of first pass for orally administered drugs, it also receives the greatest exposure to
gastrointestinal LPS via the portal vein (Jacob et al., 1977; Nolan, 1989). Furthermore,
the response of resident KCs to LPS via TLR4 signaling (Su et al., 2000) and elaboration
of inflammatory mediators by stellate cells, endothelial cells, and hepatocytes makes the

liver both a first responder to and target for LPS (Hewett and Roth, 1993).

 

Table 1.5: The inflammation hypothesis of idiosyncratic drug reactions and its

relationship to apical features of idiosvncrafi drug reactions.

 

Idiosyncratic drug reactions

“Inflammation hypothesis”

 

Liver is a frequent target organ

Exposure to GI-derived LPS
Resident liver cells (KCs, SECS) can respond to
LPS

 

Seemingly unrelated to dose

Episodic leftward shift in dose-response curve

 

Inconsistent time to onset,
Rechallenge does not always cause
the response

Exposure to inflammagens is episodic

 

Relatively rare

 

Intersection of drug exposure and inflammagen
exposure is infiequent

 

45

 

Accordingly, localized LPS-induced inflammatory responses might render the liver
sensitive to hepatotoxicity. Idiosyncratic reactions are also frequently described as dose-
independent or as not having a simple dose-response relationship for toxicity.
Idiosyncratic reactions may indeed be dependent on dose, but this relationship could be
concealed by a varying threshold for toxicity caused by episodic inflammation (Roth et
al., 2003). For this reason, drugs associated with idiosyncratic reactions might behave as
Paracelsus postulated (ie., the dose makes the poison), with an additional factor (ie.,
inflammation) required to reveal hepatotoxicity at a therapeutic dose in people. In this
regard, it is interesting that Uetrecht suggested that drugs for which the dose is below 10
mg/kg have less propensity to cause idiosyncratic reactions (Uetrecht, 2001).
Accordingly, the therapeutic plasma concentration of more potent drugs might lie below
the new threshold for hepatotoxicity established by inflammation, even after a leftward
shift in the dose-response curve. This observation is consistent with the much smaller
propensity of FAM to cause idiosyncratic hepatotoxicity compared with RAN, as the
therapeutic dose is considerably smaller than that of RAN. Idiosyncratic hepatotoxicity
occurs at variable times after the start of drug therapy (i.e., for RAN, 1 week to 9
months). This variability can be explained by the episodic nature of exposure to
inflammagens (Roth et al., 2003). Finally, the relatively small occurrence of
idiosyncratic reactions might be due to numerous factors related to the genesis of the
inflammatory response (e.g., timing of drug exposure, magnitude of inflammation,
tolerance to inflammatory mediators, etc) or to polymorphisms in genes encoding

important regulators or members of the inflammatory response. Overall, the hypothesis

that inflammation might precipitate some idiosyncratic drug reactions is consistent with
features associated with this type of drug toxicity.

The inflammation hypothesis of idiosyncratic reactions need not be distinct from
other hypotheses of mechanisms underlying idiosyncratic drug reactions (Figure 1.4),
including formation of reactive metabolites. For example, exposure to LPS can decrease
hepatic cytochrome P450 content, and could therefore result in drug accumulation.
Interestingly, activated leukocytes, such as PMNs, and the products they release can
metabolize drugs to reactive metabolites (Uetrecht, 1991). Accordingly, exposure to LPS
can potentially increase drug concentration or reactive drug metabolite generation by
altered hepatic metabolism and by activation of inflammatory cells. Inflammation might
also be an important factor in antibody-mediated hepatotoxicity. For example, in one
model of allergic hepatitis, cotreatment with LPS markedly increased the hepatotoxicity
(Mizoguchi et al., 1990). Inflammation might also be a sufficient danger signal, such that
in the presence of autoantibodies hepatotoxicity occurs. Proinflarnmatory cytokines such
as TNF have been hypothesized to provide a necessary danger signal, and there is
certainly evidence supporting a role for both inflammation and autoantibodies in HAL
idiosyncrasy. Finally, some change in liver homeostasis caused by either the drug or
inflammatory mediators might be a necessary factor for idiosyncratic hepatotoxicity.

Experimental evidence of reactive metabolites or an immune-mediated response
as mediators of idiosyncratic hepatotoxicity from DCF and HAL is available. Is there

evidence for inflammation as a component of the toxicity from either of these two drugs?

47

Figure 1.4: Integration of the “inflammation hypgthesis” with previously msed

 

 

hymtheses for mechanisms of idiosmcrgtic hepatotoxicity. Inflammation can influence a
drug’s propensity to cause idiosyncratic toxicity by several modes. For example,
inflammation might increase the concentration of a drug by inhibiting P450 metabolism.
Activated leukocytes have also been implicated in metabolism of drugs to reactive
species. Modification of proteins by these metabolites could result in formation of
autoantibodies, which during concurrent inflammation (i.e., a danger signal) can cause
hepatotoxicity. Furthermore, the ability of the drug or inflammatory mediators to alter
hepatocellular homeostasis might render the liver sensitive to injury from normally

noninj urious activation of inflammatory mediators.

48

Inflammation

./

- T or I CYP450 expression
- Metabolism by leukocytes u - u
i Danger Signal
® TNF-a
PMNS ‘/

Hemostasis

 

 

 

 

 

 

 

 

 

 

COX-2, etc.

 

 

TlDrug] or
T[Reactive metabolite]

 

 

 

p Antibodies

 

 

 

 

 

 

 

 

_—+
® 199,!“ it... firm-“‘3 ;

Idiosyncratic
Hepatotoxicity

49

For example, some evidence supports the possibility that inflammation plays a role in
idiosyncratic HAL hepatotoxicity. Cotreatrnent with LPS causes hepatotoxicity in
hypoxic rats exposed to HAL (Lind et al., 1984). The case of DCF is perhaps a bit
confirsing, as DCF is an anti- inflammatory agent. Nonetheless, gastrointestinal damage
caused by DCF might cause an increase in circulating LPS (Boelsterli, 2003b), and COX-
2 is not a critical mediator of injury in some models of LPS-augmented hepatotoxicity
(Barton et al., 2001). Accordingly, the anti-inflammatory action of DCF might be masked
by its propensity to cause release of GI LPS, which could prompt expression of non-
prostanoid inflammatory mediators. In addition, recent studies have suggested a
relationship between cytokine polymorphisms and DCF hepatotoxicity (Aithal et al.,
2004). Accordingly, some evidence supports inflammation as a mediator of
hepatotoxicity for two drugs for which the preponderance of studies have focused on
other hypotheses.

Unfortunately, epidemiological studies demonstrating a connection between
inflammation and idiosyncratic responses have not been done. However, as described
earlier, signs of inflammation frequently accompany idiosyncratic reactions to RAN (e.g.,
see prodromal inflammatory signs, Table 1.1), and a similar observation has been made
for Chlorpromazine (CPZ), another agent associated with human idiosyncratic reactions
that can be mimicked in rats cotreated with LPS (Buchweitz et al., 2002). In addition,
underlying diseases with inflammatory components might be susceptibility factors for
idiosyncratic drug reactions (Boelsterli, 2003b; Ganey et al, 2004). Overall, retrospective
evaluation of case reports and some experiments in animals have suggested a relationship

between inflammation and idiosyncratic hepatotoxicity.

50

1.4 Overview of dissertation

Inasmuch as modest concurrent inflammation can render rats sensitive to
hepatotoxicity from several xenobiotics, we have hypothesized that some idiosyncratic
responses to drugs might be precipitated by inflammatory responses that occur during
drug therapy (Roth et al., 2003; Buchweitz et al., 2002). Accordingly, the primary goal of
this dissertation is to characterize the dose-response relationship and timecourse of
hepatotoxicity in LPS/RAN-cotreated rats. Furthermore, experiments aimed at
identification of mechanisms of toxicity in this RAN idiosyncrasy model are described.
In the chapters that follow, the hypothesis to be tested is that RAN is rendered
hepatotoxic by concurrent, noninjurious inflammation in rats. The features of this model
are discussed in the context of RAN idiosyncrasy in people. Hepatic gene expression is
evaluated in this model, and the possible relationship of gene expression changes to
LPS/RAN-induced liver injury is discussed. The effects of LPS/RAN on activation of the
hemostatic system and generation of liver hypoxia as critical toxicologic consequences
are emphasized. Finally, preliminary evidence supporting a role for PMNs and a

PMN/hypoxia interaction in the LPS/RAN-model is presented.

51

CHAPTER TWO

Luyendyk, J .P, Maddox, J .F ., Cosma, G.N., Ganey, P.E., Cockerell, G.L., and Roth, RA.
(2003). Ranitidine treatment during a modest inflammatory response precipitates

idiosyncrasy-like liver injury in rats. J Pharmacol Exp Ther. 307(1):9-l6.

52

2.1 Abstract:

Drug idiosyncrasy is an adverse event of unknown etiology that occurs in a small
fraction of people taking a drug. Some idiosyncratic drug reactions may occur fi'om
episodic decreases in the threshold for drug hepatotoxicity. Previous studies in rats have
shown that modest underlying inflammation triggered by bacterial LPS can decrease the
threshold for xenobiotic hepatotoxicity. The histamine2 (1-12)-receptor antagonist RAN
causes idiosyncratic reactions in people, with liver as a usual target. We tested the
hypothesis that RAN could be rendered hepatotoxic in animals undergoing a modest
inflammatory response. Male rats were treated with a nonhepatotoxic dose of LPS (44 x
106 EU/kg, iv) or its vehicle, then 2 hours later with a nonhepatotoxic dose of RAN (30
mg/kg, iv) or its vehicle. Liver injury was evident only in animals treated with both RAN
and LPS as estimated by increases in serum alanine aminotransferase, aspartate
aminotransferase and y-glutamyl transferase activities within 6 h after RAN
administration. LPS/RAN cotreatment resulted in midzonal liver lesions characterized by
acute necrosuppurative hepatitis. FAM is an H2-antagonist for which the propensity for
idiosyncratic reactions is far less than RAN. Rats given LPS and F AM at a dose
pharmacologically equipotent to that of RAN did not develop liver injury. In vitro, RAN
sensitized hepatocytes to killing by cytotoxic products fi'om activated PMNs, whereas
FAM lacked this ability. The results indicate that a response resembling human RAN
idiosyncrasy can be reproduced in animals by RAN exposure during modest

inflammation.

53

2.2 Introduction

As described in section 1.3, the ability of modest inflammation to potentiate the
toxicity of numerous xenobiotic agents led us to hypothesize that some drug
idiosyncrasies result from episodes of mild inflammation occurring during drug therapy
(Roth et al., 2003). The hepatotoxicity of two drugs associated with idiosyncratic
reactions, CPZ and HAL, is augmented by coexposure to a small, noninjurious dose of
LPS. The purpose of this study was to test the hypothesis that underlying inflammation
triggered by a nonhepatotoxic dose of LPS renders RAN hepatotoxic in rats, revealing a
response resembling human RAN idiosyncrasy. Additionally, the hypothesis was tested
that the H2 antagonist famotidine (F AM), for which the propensity to cause idiosyncratic
reactions is markedly smaller, would not be rendered hepatotoxic by LPS. Finally, the
ability of RAN to influence the killing of rat hepatocytes by cytotoxic products released

by activated inflammatory cells was explored in vitro.

2.3 Materials and Methods

2.3.1 Materials

Unless otherwise noted, all chemicals were purchased fi'om Sigma Chemical Co.
(St. Louis, Mo). LPS derived fiom E. coli serotype 055:35 with an activity of 6.6 x 106
EU/mg was used for these studies. This activity was determined using a colorometric,
kinetic Limulus Amebocyte Lysate (LAL) assay (Kit #50-650U) purchased from

Biowhittaker (Walkersville, MD).

54

2.3.2 Animals

Male, Sprague-Dawley rats (Crl:CD (SD)IGS BR; Charles River, Portage, MI)
weighing 250-350 grams (in vivo studies) or 90-150 grams (in vitro studies) were used
for these studies. Animals were fed standard chow (Rodent chow/Tek 8640, Harlan
Teklad, Madison, WI) and allowed access to water ad libitum. They were allowed to

acclimate for 1 week in a 12-h light/dark cycle prior to use.

2.3.3 Experimental Protocol

Rats fasted for 24 hours were given 44.4 X 106 EU/kg LPS or its saline vehicle, iv
Two hours later 30 mg/kg RAN, 6mg/kg F AM or sterile phosphate-buffered saline (PBS)
vehicle was administered iv RAN solution was administered at 2 ml/kg at a rate of
approximately 0.15 ml/min. Three, 6, 12 and 24 hours later, separate groups of rats were
anesthetized with sodium pentobarbital (50 mg/kg, ip) and killed by exsanguination.
Blood was allowed to clot at room temperature, and serum was collected and stored at -
20° C until use. Three 100 mg sections of liver were taken from the interior portion of the
right media lobe, flash frozen individually in liquid nitrogen, and stored at -—80 degrees C
for RNA isolation. Representative (3-4 mm) slices of the ventral half of the left lateral

liver lobe were collected and fixed in 10% neutral buffered formalin.

2.3.4 Assessment of hepatotoxicity
Hepatic parenchymal cell injury was estimated as an increase in serum alanine
aminotransferase (ALT) activity was determined spectrophotometrically using Infinity-

ALT reagent from Sigma Chemical Co. (St. Louis, MO). Hepatic parenchymal cell injury

55

was estimated as increases in serum alanine aminotransferase (ALT) and aspartate
aminotransferase (AST) activities. Biliary injury was estimated from increases in gamma-

glutamyl transferase (GGT) activity.

2.3.5 Histopathology

F ormalin-fixed liver samples (3-4 samples/rat) were embedded in paraffin,
sectioned at 5 pm, stained for hematoxylin and eosin (H&E), and examined by light
microscopy. All tissue sections were examined by the pathologist without knowledge of
treatment (i.e., performed in blinded fashion). All lesions were assigned a score of 0 to 5,
with 0 representing no significant lesion and increasingly higher numbers representing

progressively more severe lesions.

2.3.6 Hepatocyte isolation

Hepatocytes were isolated using the Gibco Hepatocyte Product Line (Invitrogen,
Carlsbad, CA) including liver perfusion medium, liver digest medium, and hepatocyte
wash buffer (Cat. Nos 17701, 17703, 17704). All reagents were warmed to 37 °C prior to
perfusion. Rats were anesthetized with sodium pentobarbital (50 mg/kg, ip), and the
portal vein was cannulated. The liver was initially perfused with 150 ml perfirsion
medium at a rate of 12 ml/min with excess medium draining from the severed vena cava.
This was followed immediately by perfusion with 100 ml of liver digestion medium at a
rate of 12 ml/min. The liver was transferred carefully to a culture dish containing
hepatocyte wash medium and gently scraped to separate cells. The resulting liver digest

was passed through sterile gauze and spun briefly at 50xg to pellet hepatocytes. The

56

resulting pellet was washed two additional times with 50 ml volumes of hepatocyte wash
medium. Hepatocytes were then resuspended at a density of 2.5 X 105 cells/ml in
Williams’ Medium E (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum
(F BS, Atlanta Biologicals, Norcross, GA) and plated in 12-well cell culture plates

(Corning-Costar) at 0.80 ml/well. Cells were allowed to attach for 4-5 h before treatment.

2.3.7 Polymorphonuclear leukocyte (PMN) isolation and conditioned medium (CM)
preparation

Rat PMNs were isolated by glycogen elicitation, as described previously (Ganey
et al., 1994). Washed PMNs were suspended at a concentration of 2.5 X 106 cells/ml in
Williams’ Medium B. PMN—conditioned medium (PMN-CM) was prepared by treating
PMNs with cytochalasin B (final concentration, 5 rig/ml), then 1 minute later with the
PMN activator f-Met-Leu-Phe (fMLP) at a concentration of lOOnM. PMNs were then

incubated at 37°C for 30 min. They were removed by centrifugation and the supernatant

(PMN-CM) collected. Aliquots were stored at -80° C until use.

2.3.8 Effect of RAN and FAM on PMN-CM induced cytotoxicity

Serum free Williarns’ Medium B containing various concentrations of PMN-CM
(0, 25 or 50%) and either RAN (0, 175, 526, or 877 jig/ml) or FAM (0, 35, 105.2, or
175.4 rig/ml) was added after hepatocytes were attached (4-5h). After 16 h, the medium
was collected, and the remaining attached cells were lysed with 1% triton X-lOO followed
by brief sonication. Media and lysates were centrifuged at 600xg for 5 min. Cytotoxicity

was assessed by measuring ALT release into the medium using Sigma Diagnostics

57

Infinity ALT reagent (Sigma, St. Iouis, MO). ALT activity in the medium was expressed

as a percent of total ALT activity (i.e., medium activity plus lysate activity).

2.3.9 Statistical Analyses

Results are presented as mean :1: SEM. For studies in viva, one-way or two-way
analysis of variance (ANOVA) was utilized as appropriate. Histopathology scores were
compared using a rank sum test. For hepatocyte studies in vitra, one- or two-way
repeated measures AN OVA was applied after appropriate data transformation. All
individual group comparisons were made using Tukey’s test. The criterion for

significance was p<0.05 for all studies.

2.4 Results

2.4.] Dose-ranging studies.

Rats were given LPS (44.4 x 106 EU/kg) 2 h before various doses ofRAN (0, 10,
20, 25 or 30 mg/kg). Serum ALT activity was evaluated 24 h after RAN administration.
A statistically significant increase in ALT activity was not observed in rats cotreated with
LPS and RAN at doses of less than or equal to 25 mg/kg, whereas ALT activity was
significantly increased in animals cotreated with LPS and 30 mg/kg RAN (Figure 2.1A).
No dose of RAN alone caused a significant increase in ALT activity, and doses above 30
mg/kg resulted in significant mortality (data not shown). A similar dose-response study
was performed for LPS at 30 mg/kg RAN (Figure 2.13). Rats were given various doses

LPS (0, 7.4, 14.8, 22.2 or 44.4 x 106 EU/kg) 2 h before 30 mg/kg RAN and ALT activity

58

was evaluated 24 h after RAN. A significant increase in ALT activity was observed in
animals cotreated with RAN and an LPS dose as low as 14.8 X 106 EU/kg. Based on
these results, doses of RAN and LPS were selected to be 30 mg/kg and 44.4 X 106

EU/kg, respectively, for the remaining studies.

2.4.2 Development of hepatotoxicity after LPS/RAN cotreatment.

Rats were given LPS or its vehicle 2 h before RAN or its vehicle. RAN or LPS
given alone was without significant effect on ALT (Figure 2.2A) activity compared to
control at any time evaluated. Treatment with RAN caused a slight but statistically
significant increase in AST activity at 12 h but was without effect at other times (Figure
2.28). Since AST is not specific for liver injury and RAN caused no change in serum
ALT activity or liver histopathologic change (see below), the small increase in AST
activity likely arose from an extrahepatic source. Cotreatrnent of rats with LPS/RAN
resulted in a 6-10-fold increase in ALT (Figure 2.2A) and a 7-14-fold increase in AST
(Figure 2.28) activities that were significant as early as 6 h after RAN treatment and
remained elevated through 24 h. Biliary injury, as reflected in serum GGT activity,
increased by 6 h after administration of LPS/RAN and remained elevated by at least 1.5-

fold through 24 h. RAN or LPS treatment alone had no effect (Figure 2.2C).

2.4.3 Histopathological examination.
Acute, multifocal, midzonal hepatic necrosis developed in LPS/RAN-cotreated rats as
early as 3 h and increasing in severity and incidence through 24 h (Figure 2.3). Necrotic

foci varied in size and number and were characterized by hepatocellular cytoplasmic

59

Figure 2.1: Dose-ranging studies with RAN jand Lfi (A) Rats were treated with 44.4 X
106 13ng LPS, (iv), then two hours later with various doses ofRAN (0, 10, 20, 25 or 30
mg/kg) (iv). n= 3-10 animals per group. (B) Rats were treated with various doses of LPS
(0, 7.4, 14.8, 22.2 or 44.4 X 106 EU/kg), (iv), then two hours later with 30 mg/kg RAN,
(iv). n= 4-7 animals per group. Hepatic parenchymal cell injury was estimated 24 h after
RAN administration from increases in serum ALT activity. Data are expressed as mean :1:

SEM. *Significantly different from respective control group (p<0.05).

60

i

c»
8

a:
8

 

ALT Activity (UIL)
§

 

 

0 5 10 15 20 25 30 35

 

 

 

 

RAN (ma/k9)
2
a
E
.2
‘6
<
5
<
0 10 20 30 40 50
LPS (x 10‘ EUIkg)

61

Figure 2.2: flgpgotoxicitv from LPS/RAN comm Rats were treated with 44.4 X
106 EU/kg LPS or its vehicle (iv), then two hours later with 30 mg/kg RAN or its vehicle
(iv). Hepatic parenchymal cell injury was estimated 3, 6, 12 or 24 h after RAN
administration by increases in serum (A) ALT and (B) AST activities. Cholestatic injury
was estimated from increases in serum (C) GGT activity. n= 6-17 rats per group. Data are
expressed as mean :1: SEM. *Significantly different fiom all other groups at the same

time. ”Significantly different from Veh/Veh-treated rats at that time (p<0.05).

62

+ VehNeh
+ LPSNeh *

+ VehIRAN
+ LPS/RAN

;-
8

N
8

d

 

 

ALT Activity (UIL)
‘é

 

 

ty_(U_iL
i i

-—1000‘

AST Activ
§

 

 

 

16":
14*
12‘

10‘

 

4

GGT Activity (UIL)

 

 

0 s 1.2 is 24 .10
Time after RAN (Hours)

63

Figure 2.3: Representative photomicrographs of liver after L_PS/RAN cotreatment, Rats
were treated with 44.4 X 106 EU/kg LPS or its vehicle (iv), then two hours later with 30
mg/kg RAN or its vehicle (iv). Livers were removed 3 (A), 6 (B), 12 (C) or 24 h (D) after
RAN administration, fixed in 10% neutral-buffered formalin, stained for H&E, and
examined by light microscopy. Acute, multifocal, necropurulent hepatitis (arrows) was
present at each time period, and increased in severity and fiequency from 3 to 24 h. The
inset in panel D shows infiltrating PMNs, many of which are necrotic themselves. CV

indicates central vein. Images kindly provided by Dr. Gary Cockerell.

. . u... .
...f:. .. . .Lhnr 1.5..
, . . . .. . “.449“...
.r n ,. . . uh-a .
rein... . ,

 

l
. r n h a .
.c. are... .....
r. , u.ar...:. . \
$061” ‘- 4 a .

1.4.. .

(

 

.. ..m.
when

00..

 

 

 

65

eosinophilia and nuclear pyknosis. Variable numbers of infiltrating PMNs, many of
which were necrotic, were consistently present within hepatocellular necrotic foci.
Qualitatively similar lesions developed in the same time fiarne in LPS/Veh-treated rats,
but with much less severity and frequency as compared to LPS/RAN-cotreated rats. This
lesion was not present in any other treatment group. Table 2.1 presents a summary of the
severity of the liver lesion in rats treated with LPS and or RAN over 24 h.

Additional histopathological changes included vasculitis of mild to moderate
severity in livers of all rats treated with LPS, irrespective of drug treatment. This began
by 3 h in LPS/RAN-treated rats and somewhat later (6 h) in rats treated only with LPS.
Diffuse sinusoidal hypercellularity occurred within 3 h to a similar degree in all 3 groups
treated with RAN and/or LPS. This comprised hypertrophy of Kupffer cells and

increased numbers of sinusoidal PMNs.

2.4.4 Comparison of RAN and FAM.

The anti-secretory potency of F AM is at least 5 times greater than RAN in rats
(Scarpignato et al, 1987). For these studies, doses of RAN and FAM that inhibit gastric
acid secretion to a similar degree in rats (30 mg/kg RAN and 6 mg/kg FAM) were used.
This dose of F AM was not hepatotoxic when administered alone (data not shown).
Significant increases in markers of hepatic parenchymal cell injury were not observed in
animals cotreated with LPS/FAM after 24 h, whereas marked elevations were observed in
animals given LPS/RAN (Figure 2.4A and 2.4B). Similar results were observed for GGT
activity (Figure 2.4C). Histopathologic evaluation revealed midzonal hepatocellular

necrosis in livers of rats treated with LPS/RAN, but this was absent in LPS/FAM-treated

66

Table 2.1: Midzonal hepatic necrosis after LPS/RAN cotreatment. Rats were treated with
44.4 X 10‘5 EU/kg LPS or its vehicle (iv), then two hours later with 30 mg/kg RAN or its
vehicle (iv). Livers were removed 3, 6, 12 or 24 h after RAN treatment, fixed in 10%
neutral buffered formalin, and evaluated by light microscopy. Lesions characterized by
midzonal hepatocellular necrosis were assigned a histopathology score based on the
following scale: 0, no significant lesion; 1, minimal; 2, mild; 3, moderate; 4, marked; 5,
severe. n= 617 rats per group. Data are expressed as mean and range of scores for each
group. *Significantly different from Veh/Veh-treated rats at that time (p<0.05). Livers

were evaluated and scored by Dr. Gary Cockerell.

67

 

Time after RAN (h)

 

 

7mm" 3 6 12 24
VehNeh 0 0 0 0
LPSNeh 0 0.2 (0-1) 0.3 (0-2) 0.5 (0-2)
Veh/RAN 0 0 0 0
LPS/RAN 0.2 (04) 0.7 (03) 1.6 (0-3)* 2.0 (04)*

 

68

 

Figure 2.4: Comparison of LPS/RAN and L_PS/F AM cotreatments. Rats were treated with
44.4 X 106 EU/kg LPS or its vehicle (iv), then two hours later with either 30 mg/kg RAN,
a pharmacologically equipotent dose of FAM (6 mg/kg) or vehicle (iv). Hepatic
parenchymal cell injury was estimated 24 h after H2 antagonist administration as
increases in serum (A) ALT and (B) AST activities. Cholestatic injtuy was estimated as
increases in serum (C) GGT activity. Data are expressed as mean :1: SEM. n=5-21 rats per

group. *Significantly different fi'om LPS/Veh-treated rats (p<0.05).

69

 

 

 

 

 

E m

a m
A.... 31.. .- C.
mmmmommmmmmounmser.

3:: 33:2 H2 :5. 33:2 5a. 3:: assoc. pee

RAN FAM
LPS

70

 

Veh

rats. Compared to rats treated with LPS alone, no significant increase in ALT activity was
observed in rats cotreated with LPS and a dose of FAM that was equimolar to that of
RAN (data not shown).

2.4.5 Effect of RAN on killing of hepatocytes by PMN-CM.

Cytotoxicity was evaluated by release of ALT into culture medium after 16 h of
incubation with RAN/PMN-CM. PMN-CM alone caused a concentration-dependent
increase in ALT release (Figure 2.5), as reported previously (Ho et al, 1996). By itself,
RAN did not cause significant ALT release at any concentration used (Figure 2.5A).
However, it enhanced the hepatocellular killing activity of PMN-CM in a concentration-
dependent manner. Pharmacologically equipotent concentrations of FAM were tested for
comparison. FAM alone at the largest concentration used caused a very slight, but
statistically significant increase in ALT release (Figure 2.53). The cytotoxicity of PMN-

CM was unaffected by FAM at any concentration of the drug tested.

2.5 Discussion

It is widely accepted that idiosyncratic drug reactions arise from production of reactive
drug metabolites capable of causing tissue damage or fiom a specific immune response to
drug or drug metabolite haptens. These modes of action have been proposed for RAN
idiosyncrasy (Vial et al., 1991), but supporting evidence is lacking, and neither appears to
explain easily all features observed in clinical cases of RAN hepatotoxicity. For example,
time of onset of hepatotoxicity relative to the initiation of RAN therapy varies greatly:

some episodes appear as early as one week, whereas others do not occur until months

71

Figure 2.5: Effect of RAN and FAM on kilipg of hepatocytes by PMN-CM. Rat
hepatocytes were cultured at a cell density of 2.5 X 105 cells/ml in Williams’ Medium B
containing 10% F BS. Four h later the medium was changed to serum-free medium
containing conditioned medium from activated PMNs (PMN-CM) at a concentration of 0,
25 or 50%, and either (A) RAN or (B) FAM at the concentrations indicated. Cytotoxicity
was evaluated as ALT released into culture medium 16 h later. Data are expressed as
mean 1 SEM. For RAN treatment, n=6 separate hepatocyte isolations. For FAM
treatment, n=3 separate hepatocyte isolations. ‘Significantly different from respective
treatment in the absence of PMN-CM. " Significantly different fiom the value at the same

% PMN-CM in absence of drug. (p<0.05).

72

% ALT Release

% ALT Release

-O- Control
60 , + RAN (175 pglml) #*
+ RAN (526 uglml)
50 . -I— RAN (877 rig/ml)

#*

  
   

40 '
30 ‘

4 M*
20

10

 

 

0 2'0 411
% PMN CM

-0— Control
50 1 + FAM (35 rig/ml)
-'V'- FAM (105 uglml)
50 y + FAM (175 pglml)

 

 

 

°/o PMN CM

73

after the start of maintenance therapy (Ramrakhiani et al., 1998;Hiesse et al., 1985;
Halparin, 1984). Furthermore, elevations in markers of hepatocellular damage resolve
despite continued RAN therapy (Barr and Piper, 1981), which seems unlikely to occur if
accumulation of reactive metabolites is necessary. A role for a specific immune response
in RAN idiosyncrasy is equally unsupported. Clinical signs of allergic responses such as
eosinophilia have been observed in some cases (Souza Lima, 1984; Devuyst et al., 1993)
but is not a consistent finding (Hiesse et al., 1985; Barr and Piper, 1981). Anti-RAN
antibodies have not been identified. Additionally, an autoimmune component of RAN
idiosyncrasy has not been identified conclusively. Of 14 cases for which the presence of
autoantibodies was assessed, only one described the presence of anti-smooth muscle
antibodies, albeit at a very low concentration (Barr and Piper, 1981). Additionally,
rechallenge with RAN does not always result in a recurrence of toxicity (Graham et al.,
1985), as might be expected with drug allergy. Thus, current hypotheses regarding
mechanisms of RAN-induced liver injury are not consistent with all of the clinical
features of these reactions.

Interestingly, prodromal indicators consistent with inflammation and endotoxemia
are observed in many cases of RAN idiosyncrasy. Evaluation of 34 cases of RAN-related
liver injury revealed accounts of diarrhea, fever, nausea/vomiting and/or abdominal pain
in nearly 60% of the cases. Exposure of people to inflammagens such as LPS is episodic
and commonplace (Roth et al., 1997). Indeed, health care providers have noted
associations between transient illness characterized by signs consistent with
inflammation/endotoxemia and increases in liver enzymes in serum (Barr and Piper,

1981; Halparin, 1984). In addition, factors known to cause translocation of endogenous

74

LPS across the gastrointestinal lumen such as excessive alcohol consumption and surgery
have been noted in some cases of RAN idiosyncrasy (Halparin, 1984; Hiesse et al.,
1985). Thus, although no definitive studies in humans have been reported, it appears that
hepatotoxic responses to RAN are often accompanied by signs consistent with
inflammation.

Modest inflammation can markedly increase sensitivity to hepatotoxic effects of
xenobiotic agents (Ganey and Roth, 2001; Luyendyk et al., 2002). Thus, it is likely that
the liver may emerge as a target organ if an inflammatory episode occurs during RAN
treatment. The studies presented herein showed that a normally nonhepatotoxic dose of
RAN is rendered hepatotoxic when administered to rats undergoing acute inflammation
triggered by LPS. Rats cotreated with LPS/RAN showed a larger change in serum
markers of parenchymal cell injury (e.g., ALT, AST) as compared to cholestasis (e.g.,
GGT). These data are consistent with observations made in clinical cases of human RAN
idiosyncrasy, in which increases in serum markers of hepatocellular injury were usually
greater than markers of cholestatic injury. Thus, the nature of alterations in serum
markers of liver damage after treatment of rats with LPS/RAN is similar to RAN
idiosyncrasy in people. It should be borne in mind, however, that this model involves
acute administration of RAN and may represent one of several mechanisms of
idiosyncratic liver injury. Case reports of severe RAN idiosyncrasy describe marked
acute inflammatory changes such as intra-acinar accumulation of plasma cells,
macrophages, eosinophils and PMNs, acwmpmw by bridging hepatocellular necrosis
(Lauritsen et al., 1984; Ribeiro et al., 2000). Consistent with the elevated serum ALT and

AST activities, lesions in LPS/RAN-treated rats were characterized by acute midzonal

75

hepatocellular necrosis accompanied by large numbers of infiltrating PMNs. The
elevation in GGT activity did not have an obvious histological correlate. Although
several confounding factors impinge on comparing lesions observed in severe cases of
RAN idiosyncrasy in humans and LPS/RAN-treated rats (e.g., time of liver sampling),
similar features such as marked inflammatory cell infiltrates and severe hepatocellular
necrosis are found in both.

To our knowledge, there have only been 3 published reports linking F AM
administration to hepatotoxicity as compared to the 34 published reports for RAN. One
report described a hepatotoxic response that occurred greater than 2 months after FAM
therapy was discontinued, leading the authors to question whether FAM was responsible
(J imenez-Saenz et al., 2000). The link between F AM and idiosyncratic hepatotoxicity in
another case was confounded by earlier RAN treatment, and the authors noted that a
RAN contribution could not be ruled out (Ament et al., 1994). F AM-associated
hepatotoxicity has been observed in a third patient who also developed hepatotoxicity
after treatment with the H2-antagonist (CIM) cimetidine (Hashirnoto et al., 1994),
suggesting a general sensitivity to HZ-receptor antagonists. Accordingly, FAM has been
associated with few published reports of idiosyncratic hepatotoxicity, and in those cases
the contribution of FAM to hepatotoxic responses described was not clear. Thus, unlike
RAN, FAM has little propensity to cause idiosyncratic hepatotoxicity.

We tested the hypothesis that FAM would not have the same hepatotoxic
interaction with LPS that occurs with RAN. For people the recommended dose of FAM is
less than that for RAN, since FAM is a more potent H2 receptor blocker (Lin, 1991).

Accordingly, pharmacologically equipotent doses of the two drugs were selected based

76

on their relative antisecretory effect and pharmacologic potencies in rats (Scarpignato et
al., 1987). Cotreatrnent of rats with LPS and RAN resulted in the expected hepatocellular
damage as marked by increases in sermn ALT and AST activities, whereas cotreatment
of rats with FAM and LPS was without significant effect. Furthermore, LPS/F AM did not
cause an elevation in GGT activity. These results suggest that the ability of inflammation
to cause a drug in this class to produce liver injury may be selective for those drugs that
have a propensity to cause idiosyncratic reactions in humans.

Rats treated with large doses of LPS develop acute liver injury characterized by
midzonal necrosis accompanied by Kupffer cell swelling and marked PMN accumulation
(Hewett et al., 1992). Liver lesions resulting from chemical-LPS synergy can resemble
those produced by hepatotoxic doses of the chemical or LPS or both (Barton et al.,
2000b; Yee et al., 2000). LPS/RAN treatment caused an acute, midzonal, suppurative,
necrotizing hepatitis that resembled lesions in animals treated with a hepatotoxic dose of
LPS. This result suggests that RAN may increase hepatic parenchymal cell sensitivity to
an LPS-like hepatotoxic response. In livers of rats treated with LPS/RAN, inflammatory
infiltrates comprised predominately PMNs, suggesting the possibility of a role for these
cells in LPS/RAN liver injury. In other models of interaction between xenobiotic agents
and LPS, PWs are present in the liver lesions and contribute to the hepatotoxic response
(Barton et al., 2000a; Yee et al., 2003c). PMNs are also critically involved in the
hepatotoxic response to large, toxic doses of LPS and probably act through the release of
cytotoxic factors when these cells are activated (Hewett et al., 1992; Ho et al., 1996).

The exact role of PMNs in LPS/RAN liver injury has not been evaluated.

Interestingly, RAN attenuates liver injury after ischemia-reperfusion, probably by

77

inhibiting release of cytotoxic factors by PMNs (Okajirna et al., 2002). Previous studies
demonstrated that RAN was nontoxic to hepatocytes even at high (e.g., SmM)
concentrations (Zimmerman et al., 1986), and our results confirmed these previous
findings (Figure 2.5). However, hepatocytes treated with RAN were rendered more
sensitive to killing by cytotoxic factors released by activated PMNs. In the context of
observations in LPS/RAN-treated rats, these results suggest that RAN may act by
increasing hepatocellular sensitivity to PMN-derived factors. In contrast, FAM did not
increase the sensitivity of hepatocytes to killing by PMN-derived cytotoxic factors. Since -
pharmacologically equipotent concentrations were used, this suggests that the sensitizing
effect of RAN on hepatocytes is independent of H2 receptor blockade. Further studies are
needed to understand the mechanism by which RAN alters hepatocyte sensitivity to
PMN-derived products.

In summary, RAN was rendered hepatotoxic in rats undergoing a mild
inflammatory response triggered by LPS. LPS/RAN-cotreated animals developed
midzonal necrosuppurative hepatitis, and a liver-related clinical chemistry pattern
resembling human cases of RAN idiosyncrasy. In contrast, animals cotreated with LPS
and a pharmacologically equipotent or equimolar dose of F AM did not develop liver
damage, a result consistent with the far lesser (and debatable) propensity of F AM to
cause idiosyncrasy in people. Treatment of hepatocytes in vitra with RAN, but not with
F AM, increased hepatocellular sensitivity to cytotoxicity fi'om PMN-derived factors.
Overall, our demonstration that modest inflammation causes the emergence of liver as a
target for RAN toxicity in rats suggests a role for inflammation in idiosyncratic reactions

to this H2-antagonist. In addition, the results raise the possibility of developing animal

78

and cell-based models for predicting which drug candidates are more or less likely to
cause idiosyncratic reactions in people and for studying the underlying mechanisms by

which these reactions occur.

79

CHAPTER 3

Luyendyk, J .P., Mattes, W.B., Burgoon, L.D., Zacharewski, T.R., Maddox, J.F., Cosma,

G.N., Ganey, PE, and Roth, RA. (2004). Gene expression analysis points to hemostasis

in livers of rats cotreated with lipopolysaccharide and ranitidine. Toxicol Sci. 80:203-13.

80

3.1 Abstract

Studies in rats have demonstrated that modest underlying inflammation can
precipitate idiosyncrasy-like liver injury from the histamine 2-receptor antagonist
ranitidine (RAN). Coadministration to rats of nonhepatotoxic doses of RAN and the
inflammagen, bacterial lipopolysaccharide (LPS), results in hepatocellular injury. We
tested the hypothesis that hepatic gene expression changes could distinguish Vehicle,
LPS-, RAN- and LPS/RAN-treated rats before the onset of significant liver injury in
LPS/RAN-treated rats (i.e., 3 h post-treatment). Rats were treated with LPS (44 x 106
EU/kg, iv) or its vehicle, then two hours later with RAN (30 mg/kg, iv) or its vehicle.
They were killed 3 h after RAN treatment, and liver samples were taken for evaluation of
liver injury and RNA isolation. Hepatic parenchymal cell injury, as estimated by
increases in serum alanine aminotransferase (ALT) activity, was not significant at this
time. Hierarchical clustering of gene expression data fi'om Affymetrix U34A rat genome
arrays grouped animals according to treatment. Relative to treatment with vehicle alone,
treatment with RAN and/or LPS altered hepatic expression of numerous genes, including
ones encoding for products involved in inflammation, hypoxia, and cell death. Some of
them were enhanced synergistically by LPS/RAN cotreatment. Real-time PCR confirmed
robust changes in expression of B-cell translocation gene 2, early grth response-1 , and
plasminogen activator inhibitor-1 (PAI-l) in cotreated rats. The increase in PAI-l mRNA
was reflected in an increase in serum PAI-l protein concentration in LPS/RAN-treated
rats. Consistent with the antifibrinolytic activity of PAH, significant fibrin deposition

occurred only in livers of LPS/RAN-treated rats. The results suggest the possibility that

81

expression of PAH promotes fibrin deposition in liver sinusoids of LPS/RAN-treated

rats and are consistent with the development of local ischemia and consequent tissue

hypoxia.

3.2 Introduction

LPS recognition by toll-like receptors on Kupffer cells and other inflammatory
cells activates signal transduction pathways, leading to cell activation and elaboration of
inflammatory mediators (Beutler, 2002). An important component of LPS activity is
transcriptional activation of numerous genes (Gao et al., 2002) in macrophages and other
inflammatory cells. Many of these gene products, such as tumor necrosis factor-a (TNF),
can fruther activate transcription of other cytokines, adhesion molecules, and neutrophil
(PMN) chemokines in other cell types, such as endothelial cells (Zhao et al., 2003).
Increased TNF-0t mRNA can be detected in livers of rats treated with LPS, and serum
TNF-or concentration is markedly increased after LPS exposure (Barton et al., 2001)
(Hewett et al., 1993). Interestingly, TNF-or is important for liver injury from large doses
of LPS in a mechanism dependent on PMNs (Hewett et al., 1993). Thus, enhanced
expression of certain genes after LPS exposure is important for liver injury, and
understanding these changes could help to elucidate mechanisms of inflammatory tissue
injury.

Investigation of gene expression patterns might also identify mechanisms of
pathogenesis in models in which modest, noninjurious inflammation potentiates

xenobiotic-induced liver injury. For example, cotreatment of rats with a nonhepatotoxic

82

dose of LPS potentiates allyl alcohol (AA)—induced liver injury and results in greater
expression of hepatic cyclooxygenase-2 (COX-2) compared to treatment with either
agent alone (Ganey et al., 2001). In this model, COX-2 inhibition afforded partial
protection fi'om liver injury, suggesting that augmented COX-2 gene expression is
important for AA/LPS-induced liver injury (Ganey et al., 2001). In rats cotreated with
nonhepatotoxic doses of aflatoxin 31 (AF B1) and LPS, TNF-0t mRNA is increased in
liver to a level similar to rats treated with LPS alone. However, the serum concentration
of TNF-or is significantly greater in AF B1/LPS-treated rats, and this cytokine is causally
involved in the potentiation of hepatocellular injtny (Barton et al., 2001). Thus, in other
models of LPS-potentiation, a difference in magnitude of gene expression in LPS and
LPS/xenobiotic-treated rats may or may not be sufficient to cause liver injury, and post-
transcriptional increases in protein or interaction between two genes expressed at
otherwise noninjurious levels might contribute to liver injury. The use of gene array
technology can facilitate investigation of such interactions. For example, in rats treated
with galactosamine and LPS, gene arrays were used to identify changes in gene
expression related to inflammation and oxidative stress, among others (Li et al., 2003).
The purpose of this study was to test the hypothesis that hepatic gene expression
changes could distinguish rats treated with LPS, RAN or LPS/RAN before the onset of
significant liver injury in LPS/RAN-treated rats. Genes were segregated based on their
patterns of expression and classified based on the nature of their respective gene
products. Real-time PCR and ELISA were used to confirm enhanced expression of one of

these genes, plasminogen activator inhibitor-1 (PAI-l ), and hepatic fibrin deposition was

83

evaluated to determine if the enhanced PAI-l expression was associated with tissue fibrin

deposition as a functional consequence.

3.3 Materials and Methods

3.3.1. Materials

For information on this topic please refer to Chapter 2 Materials and Methods.

3.3.2 Animals

For information on this topic please refer to Chapter 2 Materials and Methods.

3.3.3 Experimental protocol

Rats fasted for 24 hours were given 44.4 X 106 EU/kg LPS or its saline vehicle, iv
Two hours later 30 mg/kg RAN or sterile phosphate-buffered saline (PBS) vehicle was
administered iv. RAN solution was administered at 2 mng at a rate of approximately
0.15 ml/min. Accordingly, rats were treated with either saline and PBS (V eh group), LPS
and PBS (LPS group), saline and RAN (RAN group), or LPS and RAN (LPS/RAN
group). Three hours later, rats were anesthetized with sodium pentobarbital (50 mg/kg,
ip) and killed by exsanguination. Blood was allowed to clot at room temperature, and
serum was collected and stored at -80‘ C until use. Slices (3-4 mm) of the left lateral
liver lobe were collected and fixed in 10% neutral buffered formalin. Three 100 mg

midlobe pieces of the right medial lobe were flash frozen in liquid nitrogen for RNA

84

isolation. Groups treated with either Veh, LPS, or RAN contained 3 rats, whereas 4 rats

comprised the LPS/RAN group.

3.3.4 Hepatotoxicity assessment

Sinusoidal endothelial cell (SEC) function was estimated using a commercially
available, enzyme-linked immunosorbent assay (ELISA) for hyaluronic acid (Corgenix
Medical Corporation, Westminster, CO). Otherwise, for more information on this section

please refer to Chapter 2 Materials and Methods.

3.3.5 RNA isolation

Total RNA was isolated from snap-frozen liver samples (approximately 100 mg)
in accordance with protocols recommended by Affymetrix Inc. (Santa Clara, CA) for
GeneChip experiments. Total RNA was isolated with Trizol reagent (Invitrogen Corp,
Carlsbad, CA) according to manufacturer’s instructions. Samples were subsequently
passed over a Qiagen RNeasy column (Qiagen Corp., Valencia, CA) for further
purification. RNA quality and concentration were assessed by absorbance at 280 and 260

nm and by analysis with a 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA).

3.3.6 Affymetrix GeneChip® analysis

RNA isolated from each rat was processed and analyzed as described below using
separate Affymetrix RG_ U34A arrays. Synthesis of double-stranded cDNA from total
RNA, synthesis of biotin-labeled cRNA, fragmentation of the cRNA for target

preparation, eukaryotic target hybridization, washing, staining and scanning of the

85

RG_U34A arrays were carried out according to the Affymetrix GeneChip® Expression
Analysis Technical Manual (701021 rev 1). Scan analysis was carried out with both the
scaling and normalization factors set to 1. For data normalization, the array was treated
as a XYZ-dimensional vector, and normalized by dividing each data point by the
cartesian length (magnitude) of the vector, then multiplied by the average of the
magnitudes of the XX arrays in the data set. XYZ=number of data points on the array
XX=number of arrays. Normalized signal data were imported into the JMP® (Release

5.0.1.2, SAS Institute Inc.) software for principal component analysis.

3.3.7 Data analysis and determination of gene activity

To determine which probesets changed after treatment compared to Veh-treated
rats, the following steps were performed. Preliminary filtering was performed using the
transcript detection call as described in the Affymetrix Statistical Reference Guide.
Probesets that did not have at least 2 samples in any treatment group with “present” or
“marginal” detection calls were eliminated from further analysis. Assessment of gene
activity was made by the Student’s t-test using R software (version 1.7.0., www.r-
project.org). Adjustment for multiple comparisons was made using a false discovery rate
(FDR) criterion (Benjamini, 1995). The FDR provided an approach capable of
decreasing, to a user-detemrined level, the likelihood of committing a type I error, at the
same time as providing a manageable number of probesets for continued analysis. For
this study, the 1000 most active probesets compared to the Veh-treated group were
identified for each treatment (i.e., LPS, RAN, or LPS/RAN) by a FDR criterion with

a=0.05. Each probeset that changed relative to its expression in Veh-treated rats was then

86

assigned to a set defined by the treatment or treatments that produced a change in its
activity (i.e., LPS/RAN [LR], L, R, LnR, LRnL, LRnR, LRnLnR, where “n”
indicates intersection of sets). The resulting sets were visualized using a Venn diagram
(Figure 3.3). Student’s t-test was used to compare the expression of FDR-active probesets
after LPS/RAN treatment with their expression after treatment with either agent alone.
Genes with greater expression in LPS/RAN-treated rats compared with either agent given
alone were identified in the LR, LRnL, LRnR and LRnInR sets (see below). Genes
with similar expression in LPS/RAN- and LPS-treated rats, or in LPS/RAN- and RAN-
treated rats were identified in the LRnL and LRnLnR or in the LRnR and LRnLnR
sets, respectively. Genes expressed to a greater degree in rats treated with LPS alone as
compared to rats cotreated with LPS/RAN were identified in the L and LRnL sets.
Hierarchical clustering was performed using Spotfire Decision Site for
Functional Genomics (Spotfire Inc., Somerville, MA) on all unique probesets showing a
significant treatment effect. Two-way agglomerative hierarchical clustering was
performed using an unweighted average and Euclidean distance as the similarity measure.
Probeset annotation was completed as described previously (Mattes, 2004). A cosine
correlation similarity measure in the profile search tool in Spotfire Decision Site for
Functional Genomics (Spotfire Inc., Somerville, MA) was used to identify genes with

patterns of expression similar to increases in serum HA concentration.

3.3.8 Reverse transcription and real-time PCR
RNA quantification was performed on the same samples fi'om the gene array

experiment using a Spectrarnax Microplate Spectrophotometer (Molecular Devices,

87

Sunnyvale, CA). Random priming was performed in a final volume of 12.5ml using
500ng of total RNA, 7.5 mM Random Hexarners (Amersham Biosciences, Piscataway,
NJ), and lmM dNTPs (Invitrogen, Carlsbad, California). RNA was denatured at 65°C
for 5 minutes and chilled on ice. Reverse Transcription (RT) master mix was prepared in
a final volume of 12.5 ml with a final concentration of 20 U/ml of Superscript II Rnase
H- Reverse Transcriptase, 4 Units/ml RNAseOut, 2mM dithiothreitol, and 1X lst Strand
RT buffer (Invitrogen, Carlsbad California). Denatured RNA and RT master mix were
combined in total volume of 25 pl. The reverse transcriptase reaction was performed at
room temperature for 10 minutes, at 42°C for 60 minutes, and then at 70°C for 15
minutes in a MJ Research Thermocycler (MJ Research Inc., Reno, NV).

The following oligonucleotide primers, designed using Oligo 6 program software
(Molecular Biology Insights, Cascade, CO), were used to quantify mRN A levels of the
following genes by real-time PCR analyses. Early growth response-1 (egr-l): upper
primer- 5’ TGA ACG CAA GAG GCA TAC CA 3’; lower primer- 5’ GAG CCC GGA
GAG GAG TAA GAG 3’. B-cell translocation gene-2 (btg-2): upper primer- 5’ CCA
GCC AGT CAC CCT TAG TG 3’; lower primer- 5’ CGG GCA GAG TGT TTG GTA
AGT 3’. Plasminogen activator inhibitor-1 (PAI-l): upper primer- 5’ AAC CCA GGC
CGA CIT CA 3’; lower primer- 5’ CAT GCG GGC TGA GAC TAG AAT 3’.
Ribosomal protein L19 (Rpll9): upper primer- 5’ CTC GAT GCC GGA AGA ACA C
3’; lower primer- 5’ CGA GCG T'TG GCA GTA CCC 3’. A 2100 Bioanalyzer (Agilent
Technologies, Palo Alto, CA) was used to analyze purities of RNA and PCR products.

Real-time PCR reactions using SYBR Green dye methodology were prepared in a

final volume of 25 pl per reaction with 1 ng of cDNA and 1X SYBR Green PCR Master

88

Mix (PE Applied Biosystems, Foster City, CA). Primer mixture was prepared in 5 pl per
reaction with a final concentration of 0.3 uM per primer. SYBR Green real-time PCR
was performed using an ABI 7900 (PE Applied Biosystems, Foster City, CA). Relative
amounts of target were calculated using the comparative CT method with ribosomal
protein L19 (RPL19, RefSeq Accession # NM_031103) as an endogenous reference and

a calibrator consisting of RNA pooled fiom all livers of Veh-treated rats.

3.3.9 Immunohistochemistry

Liver samples evaluated for fibrin immunohistochemistry were fi'om the 3 h post-
treatment timepoint in a previous study (Luyendyk et al., 2003b). A 1cm3 block of liver
cut from the lefi medial lobe was flow for 8 minutes in liquid nitrogen-chilled
isopentane then stored at —80°C until processing. 8 urn-thick sections of frozen liver
were fixed in 10% buffered formalin containing 2% acetic acid for 30 minutes at room
temperature. This fixation protocol solubilizes all fibrinogen and fibrin except for cross-
linked fibrin; therefore, only cross-linked fibrin remains in sections of liver (Schnitt et al.,
1993). Sections were blocked with PBS containing 10% horse serum (i.e., blocking
solution; Vector Laboratories) for 30 minutes, and this was followed by incubation
overnight at 4°C with goat anti-rat fibrinogen antibody diluted (1:1000, ICN
Pharmaceuticals, Aurora, OH) in blocking solution. Next, sections were incubated for
three hours with donkey anti-goat secondary antibody conjugated to Alexa 594 (1:1000,
Molecular Probes, Eugene, OR) in blocking solution for 3 hours. Sections were washed
three times, 5 minutes each, with PBS and visualized using a fluorescent microscope. No

staining was observed in controls for which the primary or secondary antibody was

89

eliminated from the staining protocol. Liver sections from all treatment groups that were

compared morphometrically were stained at the same time.

3.3.10 Quantification of fibrin staining

Fluorescent staining in sections of liver was visualized with an Olympus AX-80T
microscope (Olympus, Lake Success, NY). Ten randomly chosen digital images (100X
magnification) were captured using 3 SPOT II camera and SPOT advanced soflware
(Diagnostic Instruments, Sterling Heights, MI). Samples were coded such that the
evaluator was not aware of treatment. Each digital image encompassed a total area of 1.4
mm2 and contained several centrilobular and periportal regions. Quantification of
immunostaining was performed with Scion Image Beta 4.0.2 (Scion Corporation,
Frederick, MD) using the method described by (Copple et al., 2002a). Ten random fields

analyzed for each liver section were averaged and counted as a replicate, i.e., each

replicate represents a different rat.
3.3.11 Evaluation of serum PAI-l concentration

Serum total PAI-l concentration (i.e., inactive, active, and bound to plasminogen
activator) was evaluated using a commercially available ELISA purchased from

American Diagnostica Inc. (Greenwich, CT.).

3.3.12 Statistical analysis

90

Two-way analysis of variance with Tukey’s test for multiple comparisons was
used for analysis of clinical chemistry, immunohistochemistry, and ELISA. The criterion

for significance was p<0.05.

3.4 Results

3.4.] Development of hepatic parenchymal cell injury.

Given alone, the doses of LPS and RAN are not hepatotoxic up to 24 h after
administration @uyendyk et al., 2003b). Confirming earlier results (Luyendyk et al.,
2003b), no significant change in ALT was observed in rats given RAN or LPS (Figure
3.1), and LPS/RAN cotreatment did not cause a statistically significant increase in ALT
by 3 h (Figure 3.1). However, one of the LPS/RAN-cotreated animals had a serum ALT

activity (454 U/L) that was considerably greater than the others (103, 119, 167 U/L).

3.4.2 Cluster analysis.

Affymetrix U34A probesets defined as active after treatment with LPS and/or
RAN (see Materials and Methods) were subjected to hierarchical clustering. The
resulting dendrogram is displayed in Figure 3.2. Four clusters resolved fiom this analysis,

segregating animals by their specific treatment (V eh, LPS, RAN, or LPS/RAN).

91

Figure 3.1: Eyaliation of hepa_tic parenchymal cell inju_ry after LPS/RAN-cotreatment.
Rats were treated with 44.4 X 106 EU/kg LPS or its vehicle (iv), then two hours later with
30 mg/kg RAN or its vehicle (iv). Hepatic parenchymal cell injury was estimated 3 h
after RAN administration by increases in serum ALT activity. n=3 for rats given
Veh/Veh, LPSN eh, or Veh/RAN. nM for rats given LPS/RAN. Data are expressed as

mean d: SEM. No treatment was found to be significantly different from Veh-treated rats.

92

Dmflfl

 

 

o o
o o
3 2

3:: 33:2 5<

100 i

400i

LPS Veh LPS

Veh

 

RAN

Veh

93

Figure 3.2: Hierarchical clustering of hepatic gene expression after BPS/RAN-
cotreatment. Rats were treated with 44.4 x 10° EU/kg LPS or its vehicle (iv), then two
hours later with 30 mg/kg RAN or its vehicle (iv). Three hours after RAN administration,
RNA was isolated from liver, and gene expression was evaluated by Affymetrix U34A
Rat Genome Array. RNA from each rat was analyzed using a separate array. Affymetrix
probesets passing a false discovery rate statistical filter were subjected to hierarchical
clustering using an unweighted average and Euclidean distance as the similarity measure.
As the dendrogram (top of figure) indicates, the cluster analysis segregated rats by

treatment.

94

 

 

 

 

 

95

Additionally, animals treated with LPS, RAN or LPS/RAN clustered separately from

Veh-treated animals.

3.4.3 Gene expression changes after LPS, RAN, or LPS/RAN treatment.

From the population of probesets examined, those for which gene expression was
altered by LPS and/or RAN treatment relative to Veh control were selected. Each
probeset in this group was assigned to a set defined by its change after treatment with
LPS, RAN, or LPS/RAN. Sets were also identified for those probesets altered by more
than one treatment. For example, the set defined by the intersection of LPS/RAN and
LPS sets (i.e., LRnL) contains probesets changed after LPS/RAN treatment and after
LPS treatment. The resulting sets are summarized as a Venn diagram in Figure 3.3. The
genes represented by probesets defining each set were identified and are shown along
with gene symbol, Unigene identification (Rn build 117), locuslink identification, signal
intensities relative to Veh treatment, and standard deviations in Supplemental Tables 1-7.
Several probesets were either increased (27) or decreased (381) only in LPS/RAN-treated
animals. This set of probesets is of obvious interest since liver injury results only fiom
this treatment (Luyendyk et al., 2003b). Several probesets were also changed only after
LPS treatment (163 increased; 46 decreased) or only afier RAN treatment (71 increased;
20 decreased). Changes in these probesets are likely not sufficient to cause liver injury by
themselves since rats treated with LPS or RAN alone at these doses do not develop liver
injury (Luyendyk et al., 2003b); however, the potential exists for interaction of one of
these gene products with another, resulting in liver injury. Probesets changing afier

LPS/RAN treatment and after either agent given alone (i.e., LRnL, LRnR) are

96

Figure 3.3: Venn diaggam depiction of probeset activity relative to Veh/Veh—treated Lag,
Rats were treated with 44.4 X 10° EU/kg LPS or its vehicle (iv), then two hours later with
30 mg/kg RAN or its vehicle (iv). Three hours after RAN administration, RNA was
isolated from liver and gene expression evaluated by Affymetrix U34A Rat Genome
Array. The number of Affymetrix probesets increased (1‘) or decreased (Jr) relative to
Veh-treated rats in a given treatment set is shown. Probesets with activities altered by
more than one treatment are indicated by an intersection symbol (0). LR: probesets
changed only afier treatment with LPS/RAN (Supplemental Table l); L: probesets
changed only after treatment with LPS (Supplemental Table 2); R: probesets changed
only after treatment with RAN (Supplemental Table 3); InR: probesets changed after
treatment both with LPS alone and with RAN alone (Supplemental Table 4); LRnL:
probesets changed after treatment both with LPS/RAN and with LPS alone
(Supplemental Table 5); LRnR: probesets changed after treatment both with LPS/RAN
and with RAN alone (Supplemental Table 6); LRnLnR: probesets changed after

treatment with LPS/RAN, LPS alone, and RAN alone (Supplemental Table 7).

97

potentially important, but injury would likely require a different magnitude of expression
in LPS/RAN-treated rats, since liver injury only occurs in this group.

Genes with specific treatment-related expression patterns were identified in the
sets described above and further categorized into groups using a secondary statistical
filter as described in Materials and Methods. Genes with greater expression in
LPS/RAN-treated rats compared with either agent given alone were identified in the LR,
LRnL, LRnR and LRnLnR sets (Table 3.1). Overexpression of one or more genes in
this group might be a determinant of liver injury in rats treated with LPS/RAN. In
addition, groups of genes with similar expression in LPS/RAN- and LPS-treated rats
(from LRnL and LRnLnR sets) or in LPS/RAN- and RAN-treated rats (from LRnR
and LRnInR sets) were identified. Since liver injury does not occur after treatment of
rats with LPS or RAN alone, expression of these genes might be important if an
interaction occurs between two or more gene products. Lastly, a group of genes
expressed to a greater degree in rats treated with LPS alone as compared to rats cotreated
with LPS/RAN was identified in the L and LRnL sets. The importance of this group lies
in the possibility that RAN might interfere with the upregulation by LPS of a gene that
protects against liver injury.

Genes were grouped based on these 4 expression patterns, and their gene products
were classified into one or more categories, including inflammation, acute phase,
hypoxia-inducible, oxidative stress, cell death signaling, cell cycle control, and repair
(Table 3.1). This classification revealed that several genes with greater expression in

LPS/RAN-treated rats compared to other treatments were related to inflammation and/or

99

 

0000 3.0002 0.0005008:

 

 

 

 

 

 

 

. . >050 I xi- Oxamnze 00.. 0005 00.. o 0.0 . c883
manage: 383 583388 38¢ sages 93.8 188:? near. ween: mm;
..nmagz #2... ea vm>z o m a a a N a m
remix/z a cum 8 m a a a m u an
$220.2 a m>z a a N a a a a a
:85; a :8 B a o m B N a m;

 

 

 

 

 

 

 

 

 

.328 P: mx08mmm0El~maoB 8: m 88 go 395 oflmmmmmommoaboaom 23. mos. 95.88.: 96888: 388m $88
580388 #08 n8 8% 0m «38 .802: 8 398 w 8 88108 E Kwanzaa E:— 78.50%. m9: was? $88 89:38
8888 3. 938 96888 8 m «8&2. 8888 8 rwmfi>260qg8¢ 88 000688 8 £92 £2: £80 £008
elem; v Cum 8.8 v E. min—E. 968880 39. 80083 «<5. Cum; on Cow mehfiwZthmv. 38.2.
96888: 8.8—. 80883 25. 50mg 0a E meng. Ea mgam 88888 8 m 8me aam8a “59.
Sam; 008388 008388 8 H8882: 2:: Cum E08 PvaZDrwmv. Q88 8 :88 mace? :88 08858
3088 00 88 888338 0m :8: mono .0883 8 008 on 808 55303. > mama 88. 8 088an 8 308 800 008 macaw.
€qung 828 m.: mm” m8 meoommo moan 08988 25:: one: 9.0%. 0808 $8 5:802: magmas on can: 5:802:

8030: m8 E8 80%?

100

were hypoxia-inducible (Table 3.1). For example, hypoxia-inducible genes, including
early growth response-1 (egr-l), glucose transporter-1 (GLUT-l), insulin-like growth
factor binding protein-1 (igtbp-l), and plasminogen activator inhibitor-1 (PAI-l ), had
greater expression afier LPS/RAN treatment as compared to expression afier treatment
with LPS or RAN alone (Table 3.2). Furthermore, genes involved in inflammation such
as the chemokine (C-X-C motif) ligand 10 (CxCllO) and the cell cycle regulator B-cell
translocation gene-2 (btg-2) segregated into this group (Table 3.2).

Table 3.1 shows that groups of genes expressed to a similar degree afier treatment
with LPS or LPS/RAN had gene products largely related to inflammation. Likewise,
several genes expressed similarly after LPS/RAN or RAN treatment fell into this group.
A group of genes with attenuated expression in LPS/RAN-treated rats compared to rats
treated with LPS alone was also identified. Gene products in this group were related to
inflammation, cell death signaling, and oxidative stress (Table 3.1). Specific genes

comprising each of these groups are available online in Supplemental Tables 8-11.

3.4.4 Real-time PCR.

Genes were selected for real-time PCR analysis based on the FDR filter results in
addition to treatment comparisons using Spotfire Decision Site for F untional Genomics.
Real-time PCR was performed for three of the genes with increased expression in
LPS/RAN-treated rats: PAI-l, egr-l, and btg-2. PAI-l expression was significantly
increased in LPS-treated and RAN-treated rats by 102- and 10-fold, respectively, but by
nearly 700-fold in LPS/RAN-treated rats (bars, Figure 3.4A). This pattern of expression

was consistent with signal intensities for the corresponding Affymetrix probeset

101

 

90:0 3.0000. 0.805080:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. . . 00..
Q80 2980 58338.0: >0c8 509.8- 900820 00: 00.8: 96.0 80m:
9: 0:80. 50:06.0 98¢ 982.50 0 0 :8.
w.00__ "88.00020: 00:0
M 503.. x x x x x
0303058 8-x.0 305
E 0 52.30» x x
001K 905:: ”800:8 : E» x x x
E. a 5).. x
83008: 5).» x x x
3006880 w C..»Dr x
.:mc_.:-_=8 9923 808a
030.8 @083 ._ E» x x
E: m 068.8% 5).» x x x x x x x
30030800 008088
2.00020008088330
E0 80:: E» x
3088388.: rmDrDm x x x x x x
08088308800880
m383 END.» x
8» <_rwo 08303 E» x
083580: 02282
33083 52.).» x x x
4.8580: 5).» x x

 

 

.0020 uh” Q88 88

meg—08088 .0920 m.

 

E 8 :88 W08 hwmgéoaoa 88 008 0.80 8 088 80888. Q88 <18
«888 0008880 8 fi0m>~>2A8080 88 008088 «$8 0068880 038 88888800 0m Cum on E 0800
208 808080 8 .80 Cw. 50)? EN)? 80 ENDED” 88 0». 080808 9808 8 8 0801000 8 7.88188 88
30808. mun: mono 8 88 m800 £8 088500 0800 00 80 88888.0 0». m8 @000 08003 8 000 on 808
800.008. Q88 08 .880 80800805.. Q88 8 88 @000 88 00:5 :0" 00 088500 0a 208 Mm? 08 :88 8

102

Figure 3.4: Real-time PCR confirmation of gne mam Rats were treated with 44.4
X 106 EU/kg LPS or its vehicle (iv), then two hours later with 30 mg/kg RAN or its
vehicle (iv). Three hours afier RAN administration, RNA was isolated from whole liver,
and SYBR Green real-time PCR was performed for (A) plasminogen activator inhibitor-1
(PAI-l), (B) early growth response-l (egr-l), and (C) B-cell translocation gene-2 (btg-Z).
Ribosomal protein L19 was used as a housekeeping gene. Results are shown as fold
change relative to average expression in Veh-treated rats as determined by the
comparative Ct (AACt) method (bars). Normalized signal intensities for corresponding
Affymetlix probesets are graphed for comparison (circles). Although a single probeset for
btg-2 (M6092 ] _at) is shown for comparison, all 3 probesets for this gene determined to
be active by the FDR filter had a similar expression pattern (M60921_at, M60921_g_at,
rc_AA944] 56_s_at). Data are expressed as mean :4.- SEM. n=3-4. *Significantly difiemnt

from Veh/Veh-treated rats (p<0.05). #Significantly different fiom all other treatments

(p<0.05).

103

.0005 03:05:02 .0005 03:00:02 .0005 03:00:02

i .

-2500
000
500
000

      

 

    

 

 

 

 

 

s
l .0. a. UN
m m m... M...
m. mm. m... w
.l '7. .l s
M”. .0 mm M”. w."
1 40 0... 0
mm m.» mm 0v
no no a. v
A. . . w. B. . . -4-. Cw. . . ..
mmmmo 000005.. M0000

00:20 0.0... 00:20 so“. 00:20 0.0“.

1M

(circles, Figure 3.4A) identified as active by the FDR filter. For egr-l (Figure 3.4B), a
significant increase in mRNA was observed in livers from LPS/RAN-cotreated rats, but
not in livers of rats treated with LPS or RAN alone. This was consistent with signal
intensities for the associated Affymetrix probeset (circles) identified by the FDR as active
only in LPS/RAN-cotreated rats (Figure 3.48). Btg-2 expression was significantly
increased in rats given LPS or RAN alone by 25- and 4-fold, respectively, but a much
greater increase (67-fold) occurred in rats cotreated with LPS/RAN (Figure 3.4C).
Although a single probeset for btg-2 (M60921_at) is shown for comparison (circles,
Figure 3.4C), all 3 probesets for this gene determined to be active by the FDR filter had a

similar expression pattern (M60921_at, M60921 _g_at, rc_AA944156_s_at).

3.4.5 Evaluation of serum PAI-l.

To determine if the change in hepatic gene expression of PAH resulted in altered
serum concentration of PAH protein, total PAI-l protein was measured. Serum PAI-l
was significantly increased after either l..PS or RAN treatment by 450- and 70-fold,
respectively. Serum PAI-l in LPS/RAN-treated rats was significantly greater than that of

rats treated with either agent given alone (Figure 3.5).

3.4.6 LPS/RAN treatment and sinusoidal endothelial cells (SECs).

In addition to expression in hepatic parenchymal cells, PAI-l is expressed by
endothelial cells exposed to factors such as LPS and inflammatory cytokines (Colman,
1994). To investigate alteration of sinusoidal endothelial cell (SEC) function in livers of

LPS/RAN—treated rats, serum hyaluronic acid (HA) was measured. Ordinarily, 90% of

105

Figure 3.5: Serum plasminogen activator inhibitor-1 (EAI-l) concentration after
LPS/RAN treatment. Rats were treated with 44.4 X 106 EU/kg LPS or its vehicle (iv),

then two hours later with 30 mg/kg RAN or its vehicle (iv). Serum concentration of PAI-
1 in its active, latent and complexed forms was evaluated three hours later using an
ELISA. n= 3-4 rats per group. Data are expressed as mean 0 SEM. *Significantly
different fiom VehNeh-treated rats (p<0.05). ”Significantly different from all other

treatments (p<0.05).

106

 

 

 

 

e11
LPS Veh LPS

.0

 

Veh

 

 

‘ ‘ ‘ J

86 2

000 t

1

:55... 720 5.000

 

RAN

Veh

107

HA in the blood is cleared by SECS in the liver (Kobayashi et al., 1999). Accordingly,
increased plasma HA concentration suggests altered SEC function, and this has been used
as a biomarker afier toxic insult (Copple et al., 2002a; Deaciuc et al., 1993). A modest,
but significant elevation in serum HA concentration was observed in rats treated with
either LPS or RAN alone, whereas serum HA was elevated more than 8-fold in rats
cotreated with LPS/RAN (Figure 3.6). PAI-l expression, as determined by either gene
array or real-time PCR, correlated significantly (r2=0.93) with changes in serum HA

concentration (data not shown).

3.4.7 Hepatic fibrin deposition.

Enhanced PAI-l expression and serum HA concentration in LPS/RAN-cotreated
rats suggested altered SEC function consistent with a procoagulant environment in the
liver, raising the possibility of enhanced fibrin deposition. Accordingly, livers were
removed 3 h after RAN treatment, as in the gene array experiment, and processed for
fibrin immunohistochemistry. Figure 3.7 shows representative images of hepatic fibrin
staining in livers. Fibrin staining in the intima of the larger vessels of Veh-treated rats
(Figure 3.7A) occurs post-mortem (i.e., artifactually) and can be prevented by perfusing
the liver with heparin prior to organ removal (data not shown). Minimal staining was
observed in the sinusoids of Veh-treated rats. Similarly, no sinusoidal staining was
observed in rats given RAN alone (Figure 3.7C). Slight fibrin staining was observed in
livers of rats treated with LPS alone (Figure 3.73). In LPS/RAN-treated rats (Figure

3.7D), a much more pronounced, panlobular fibrin staining occurred in sinusoids.

108

Morphometry revealed a statistically significant increase in fibrin staining only in livers

from animals cotreated with LPS/RAN (Figure 3.7E).

3.5 Discussion

The work presented approached the study of drug-inflammation interaction by examining
gene expression in an animal model of RAN idiosyncratic liver injury. In animals given a
nonhepatotoxic dose of LPS followed two hours later by a nonhepatotoxic dose of RAN,
we showed previously that livers were normal at 3 h post RAN treatment but became
injured by 6 h. We chose to examine gene expression changes in liver at a time (3 h) just
before the onset of liver injury in LPS/RAN-treated animals (Luyendyk et al., 2003b). At
this time, hierarchical cluster analysis of hepatic gene expression changes was able to
segregate rats by treatment group (Figure 3.2). To identify changes in gene expression
related to initiation of liver injury in this model, the activity of genes after treatment with
LPS and/or RAN was compared to activity in Veh-treated rats (Figure 3.3). Increases in
pro-inflammatory gene products or altered hepatocellular homeostasis might precipitate
liver injury in LPS/RAN-treated rats, therefore, we identified genes within these sets that
followed four expression profiles consistent with potential involvement in the
pathogenesis (Table 3.1). The most obvious genes to examine were those with greater or
attenuated expression in LPS/RAN-treated rats compared to treatment with LPS or RAN

alone, since only rats in this group develop liver injury.

109

Figure 3.6: Effect of LPS/RAN cotreatment on serum hyaluronic acid (HA)
concentraftion. Rats were treated with 44.4 X 106 EU/kg LPS or its vehicle (iv), then two
hours later with 30 mg/kg RAN or its vehicle (iv). Altered sinusoidal endothelial cell
function was estimated 3 h afier RAN administration by increases in serum HA activity.
n= 3-4 rats per group. Data are expressed as mean d: SEM. ‘Significantly different from

Veh/Veh-treated rats (p<0.05). “Significantly different fiom all other treatments (p<0.05).

110

LPS Veh LPS

Veh

 

 

o 0 0 0 0
w o o o
3 2 1

2.59.. 5.. E390:

 

Veh

111

Figure 3.7: Effect of LPS/RAN cotreagnent on hepatic fibrin deposition. Rats were
treated with 44.4 X 106 EU/kg LPS or its vehicle (iv), then two hours later with 30 mg/kg
RAN or its vehicle (iv). Livers were removed 3 h after RAN treatment and processed for
immunohistochemistry as described in Materials and Methods. Representative images of
fibrin staining in livers from (A) Veh- and (C) RAN- treated rats showing minimal
staining (black). (B) Representative image from rat treated with LPS showing slight
panlobular fibrin staining. (D) Representative image fi'om LPS/RAN-cotreated rat
showing marked panlobular fibrin deposition. PP, periportal. CV, central vein. For (E),
the area of positive fibrin staining in 10 randomly chosen, IOOX fields per tissue was
determined morphometrically as described in Materials and Methods. Data are expressed

as mean i SEM. n=3. “Significantly different from all other treatments (p<0.05).

112

 

PP

 

 

 

 

 

 

 

 

P".

PP'

CV

 

 

 

 

 

 

 

 

 

 

 

C
, PF
9'
CV.“

E

“i 0.030 ' g

g 0,025.
a -
2 3 0.020‘
g 2 0015‘
O: c
.0 .9 0.0104 I
E a

g 0005‘ ,

5 0.000 [ j i 3

Veh LPS Veh LPS
Veh RAN

 

A less obvious possibility is that genes expressed to a similar degree after LPS or
LPS/RAN treatment or to a similar degree after RAN or LPS/RAN treatment are
important for liver injury in LPS/RAN- treated rats. Since liver injury does not develop in
rats treated with LPS or RAN alone, induction of such a gene is probably not sufficient to
cause liver injury by itself but may be involved in the pathogenesis of liver injury if it
interacts with one or more gene products. Several genes were identified with similar
expression in livers after treatment with LPS compared to treatment with LPS/RAN (see
LPS/RANzLPS in Table 3.1, Supplemental Table 9). Genes in this group were related to
inflammation or could be identified as LPS-inducible (Supplemental Table 9). For
example, genes encoding inflammatory cytokines, including TNF-a and interleukin-1|} as
well as inducible nitric oxide synthase, showed similar expression after LPS-treatment or
LPS/RAN-cotreatment. Furthermore, cell surface molecules including the adhesion
molecules lectin-like oxidized low-density lipoprotein receptor-l (LOX-l), CD14 and
CD3 8, the transcription factors CCAAT/enhancer binding protein (C/EBP)-delta and —
beta, and products involved in signal transduction such as Janus kinase 2 (Jak2) and
phosphodiesterase 4B (PDE43) fit this pattern of expression.

PDE4B is an important regulator of inflammatory responses, including expression
of cytokines and activation of inflammatory cells such as neutrophils (PMNs; J in and
Conti, 2002; Essayan, 1999). In this regard, it is of interest that hepatic lesions which
develop in LPS/RAN-treated rats are laden with PMNs (Luyendyk et al., 2003b).
Regulation of inflammation by PDE48 or other gene products may be essential for
LPS/RAN-induced liver injury but not sufficient to produce liver injury in the absence of

RAN cotreatment. Indeed, RAN can sensitize hepatocytes to the cytotoxic effects of

114

PMN-derived factors (Luyendyk et al., 2003b). It therefore seems possible that RAN
sensitizes hepatocytes to become injured from otherwise noninj urious upregulation of
pro-inflammatory genes, leading to idiosyncratic hepatotoxicity.

Whereas some genes were induced after LPS- and LPS/RAN-treatment to a
similar degree, the expression of others was increased after LPS treatment but showed an
attenuated response after cotreatment with RAN (see LPSIRAN>LPS in Table 3.1,
Supplemental Table 11). This pattern is of interest because RAN might prevent
expression of gene products that downregulate inflammation or cell death signals, thereby
enhancing inflammation to tissue-damaging levels and/or activating cell death pathways.
Indeed, within this group, several members of signal transduction pathways were
identified, including protein kinase C (PKC)-epsilon. Interestingly, PKC—epsilon and
another gene with this expression pattern, oxygen regulated protein (1 50kD) (ORPl 50),
have been implicated in protection against ischemic stress (Gray et al., 1997; Ozawa et
al., 1999). This suggests the possibility that livers fi'om LPS/RAN-cotreated rats are more
sensitive to harmful effects of local ischemia due to loss of protective gene products. In
addition, expression of the proteasome subunits LMP2 and LMP7 was observed in LPS-
treated rats, but this increase was attenuated after LPS/RAN treatment. Cells lacking
LMP2 and LMP7 have defective NF-kB translocation and are sensitive to TNF-a-
induced apoptosis (Hayashi and Faustman, 2000). Overall, these and other genes
expressed to a greater degree after LPS treatment (heat shock 70kD protein 1A, heme
oxygenase-2, dnaJ homolog subfamily b member 9) compared to LPS/RAN cotreatment
may confer some degree of cytoprotection, perhaps by decreasing sensitivity to hypoxia,

inflammatory mediators, or oxidative stress.

115

Genes expressed to a greater degree in LPS/RAN-cotreated rats compared to
treatment with either agent alone could represent a population with mechanistic
importance if the gene product became expressed at or above the level required to
participate in liver injury. Our analysis revealed that genes characterized by this
expression pattern were primarily hypoxia-inducible or involved in inflammation (see
LPS/RAN>LPS>RAN in Table 3.1, Supplemental Table 8, Table 3.2). Several of the
hypoxia-inducible genes in this group can also participate in inflammatory responses. For
example, CxCllO (interferon-inducible cytokine IP-l O) is induced under hypoxic
conditions and modulates recruitment and retention of inflammatory cells (Neville et al.,
1997). The hypoxia-inducible transcription factor egr-l is involved in cell death signaling
(Thiel and Cibelli, 2002), but it can also influence inflammatory responses by altering
cytokine expression (Yan et al., 1999; Shi et al., 2002). The observation that numerous
hypoxia-inducible genes are expressed in LPS/RAN-cotreated rats suggests the
possibility that hypoxia is involved in liver injury in this idiosyncrasy model. However,
further studies are required to confirm tissue hypoxia in livers of LPS/RAN-cotreated rats
and its role in pathogenesis.

Another gene product expressed to a larger degree in LPS/RAN—cotreated rats
compared to rats treated with LPS or RAN alone is PAI-l. PAI-l is induced by various
stimuli, including LPS, inflammatory cytokines, and hypoxia and is expressed by several
cells in the liver, including parenchymal and endothelial cells (Binder et al., 2002;
Kietzmann et al., 1999; Hamaguchi et al., 2003). Although the cell source and

mechanism of enhanced PAI-l expression in livers of LPS/RAN-cotreated rats is not

known, one possibility is that RAN may indirectly augment expression by increasing

116

levels of cytokines known to induce PAI-l, such as TNF-a or IL-1. Interestingly,
although hepatocellular liver injury was not observed at this early time after LPS/RAN
cotreatment, serum HA concentration was significantly increased, suggesting altered SEC
homeostasis. This elevation in serum HA concentration supports SECS as a potential
source of PAH. Perturbation of SECS by hypoxia (Kietzmann et al., 1999) or altered
signal transduction may be responsible for augmented PAI-l expression in endothelial
cells in LPS/RAN-treated rats. P38 mitogen-activated protein kinase is involved in
induction of PAH during hypoxia but does not appear to be important in induction of
PAH by TNF-a (Kietzmann et al., 2003; Hamaguchi et al., 2003). Another possibility is
that cotreatment with RAN influences PAI-l expression at the transcriptional level.
(Gruber et al., 2003) demonstrated that the PAL] promoter contains a response element
for the orphan receptor Nur7 7 (N GF I-B, TR3). Furthermore, Nur77 overexpression in
human umbilical vein endothelial cells activates a luciferase reporter gene controlled by a
PAL] promoter, and Nur77 is necessary for induction of PAH expression by TNF-oz
(Gruber et al., 2003). Since Nur77 is also a hypoxia-inducible gene (Choi et al., 2004), its
expression might be expected to be enhanced in LPS/RAN-treated rats, since numerous
other hypoxia-inducible genes, including PAI-l , were identified in this group (Table 3.2).
Although probesets for Nur77 did not emerge as active after LPS/RAN-cotreatment, the
expression of these was more than 10-fold greater than Veh-control in 3 of 4 LPS/RAN-
treated rats, whereas less than 2-fold changes occurred in livers of rats given LPS or RAN
alone. This result raises the possibility that Nur77 is important for enhanced expression of
PAH in LPS/RAN-treated rats. However, additional experiments are necessary to prove

such a connection.

117

The greater expression of the PAL] gene in livers of LPS/RAN-treated rats
(Figure 3.4A) was reflected in enhanced concentration of PAH protein in serum (Figure
3.5), suggesting a firnctional consequence to its induction. PAI-l has many physiological
roles including inhibition of fibrinolysis and modulation of inflammatory cell migration
(Binder et al., 2002; Marshall et al., 2003). Consistent with its antifibrinolytic activity and
pattern of expression, significant fibrin deposits occur only in livers of rats given
LPS/RAN. Concurrently elevated serum concentrations of both HA and PAL] suggest
the possibility that altered SEC homeostasis might favor activation of the hemostatic
system and fibrin deposition in liver. Fibrin deposition could cause local ischemia, and
resultant hypoxia might contribute to the development of liver inj tn'y. This hypothesis is
consistent with the observation that numerous hypoxia-inducible genes were expressed in
livers of LPS/RAN-cotreated rats (Table 3.3). Inasmuch as the PAH gene is hypoxia-
inducible, the presence of hypoxia might further enhance its expression, triggering a
cascade of hemostatic dysregulation. Additionally, hypoxic upregulation of egr-l could
facilitate coagulation system activation by upregulating tissue factor on liver cell
membranes (Pawlinski et al., 2003). Overall, the data suggest that enhanced PAI-l
expression in LPS/RAN-cotreated rats encourages formation of fibrin clots, possibly
resulting in disrupted liver blood flow and hepatocellular hypoxia that could contribute to
the development of necrosis. Preliminary studies showing protection fi'om liver injury by
fibrinolytic or anticoagulant drugs in LPS/RAN-treated rats support this hypothesis
(Luyendyk, 2004a).

In summary, rats were treated with either a nonhepatotoxic dose of LPS or its Veh

and with either RAN or its Veh. Of the four treatments, only LPS/RAN treatment results

118

in liver injury (Luyendyk et al., 2003b). At a time before the onset of significant
hepatotoxicity in LPS/RAN-treated rats, hierarchical clustering of hepatic gene
expression segregated animals by treatment. Hypoxia-inducible genes, including PAL]
and egr-l , were expressed to a greater degree in livers after LPS/RAN treatment
compared to either agent given alone. The enhanced PAI-l gene expression was reflected
in increased PAI-l protein in serum. Significant fibrin deposits in liver sinusoids were
observed only in LPS/RAN-cotreated rats, consistent with the antifibrinolytic activity of
PAH . Overall, the results suggest that altered expression of genes promoting hemostasis
might contribute to liver injury in LPS/RAN-cotreated rats. The studies presented
represent a snapshot in time at one dose; examination of gene expression at other times
after treatment and other doses of LPS and RAN will illuminate this connection further.
The association of hypoxia-inducible gene expression with hepatic fibrin deposition in
LPS/RAN-cotreated rats is consistent with the development of tissue ischemia/hypoxia as

a contributing factor to liver pathogenesis in this model of RAN idiosyncrasy.

3.6 Supplemental data

The results summarized as a Venn diagram in Figure 3 are available for download
in Microsoft Excel format. The genes represented by probesets defining each set were
identified and are shown along with gene symbol, Unigene identification (Rn build 117),
locus-link identification (hyperlink available), signal intensities relative to vehicle
treatment, and standard deviations in Supplemental Tables 1—7. Supplemental text
describing the Venn diagram is also available for download in Microsoft Word format.

Genes expressed with the specific patterns summarized in Table 1 are available as Tables

119

8—11 in Microsofi Word format. Following each table are selected references for each
gene product. In some cases, probesets for the same gene followed the same expression
pattern but were segregated to different patterns by statistical analysis. Supplemental data

are available at www.toxsci.oupjoumals.org.

120

CHAPTER 4

Luyendyk, J.P., Maddox, J.F., Green C.D., Ganey, P.E., and Roth, RA. (2004).

Augmentation of lipopolysaccharide-induced fibrin deposition by ranitidine and its

connection to idiosyncrasy-like liver injury in rats. Hepatology. In press.

121

4.1 Abstract

Coadministration of nonhepatotoxic doses of the histamine 2-receptor antagonist
ranitidine (RAN) and bacterial lipopolysaccharide (LPS) results in hepatocellular injury
in rats, the onset of which occurs in 3-6 h. This reaction resembles RAN idiosyncratic
hepatotoxicity in humans. Early fibrin deposition occurs in livers of rats cotreated with
LPS/RAN. Accordingly, we tested the hypothesis that the hemostatic system contributes
to liver injury in LPS/RAN-treated rats. Rats were given either LPS (44.4 x 106 EU/kg)
or its vehicle, then RAN (30 mg/kg) or its vehicle 2 h later. They were killed 2, 3, 6, 12,
or 24 h after RAN treatment, and liver injury was estimated from serum alanine
aminotransferase (ALT) activity and liver histopathology. A modest elevation in serum
hyaluronic acid, which was most pronounced in LPS/RAN-cotreated rats, suggested
altered sinusoidal endothelial cell function. A decrease in plasma fibrinogen and
increases in thrombin-antithrombin (TAT) dimers and in serum concentration of
plasminogen activator inhibitor-1 (PAI-l) occurred before the onset of liver injury.
Hepatic fibrin deposition was observed in livers from LPS/RAN-cotreated rats 3 and 6 h
after RAN. Liver injury was abolished by the anticoagulant, heparin, and was
significantly attenuated by the fibrinolytic agent, streptokinase. Hypoxia, one potential
consequence of sinusoidal fibrin deposition, was observed in livers of LPS/RAN-treated
rats. Taken together, the results suggest that the hemostatic system is activated after
LPS/RAN cotreatment and that fibrin deposition in liver is important for the genesis of

hepatic parenchymal cell injury in this model.

122

4.2 Introduction

Exposure of rats to large doses of LPS results in midzonal hepatic necrosis that
requires several factors including cytokines, inflammatory cells, an activated hemostatic
system, and platelets (Hewett and Roth, 1995; Hewett et al., 1993; Hewett et al., 1992;
Pearson et al., 1995). The involvement of similar factors in LPS/RAN-induced
hepatotoxicity has not been investigated. The hemostatic system is composed of two
branches (Figure 3.1), the coagulation and fibrinolytic pathways. In the liver, SECS play a
critical role in regulation of coagulation and fibrinolysis. Under resting (normal)
conditions, SECS express factors including thrombomodulin and protein C that inhibit
coagulation, thereby suppressing fibrin formation (Schultze and Roth, 1998). However,
when activated or damaged the anticoagulant properties of these cells are lost, and
expression of procoagulant factors including tissue factor (TF) can facilitate coagulation
system activation. In a healthy vasculature, the accumulation of fibrin is also prevented
by the process of fibrinolysis. SECS produce both profibrinolytic and antifibrinolytic
factors. Accordingly, altered SEC homeostasis is a potential predictor of disturbances in
the hemostatic system.

Coagulation system activation occurs through both the intrinsic and extrinsic
pathways, ultimately resulting in conversion of prothrombin to thrombin. Thrombin can
enzymatically cleave circulating fibrinogen to fibrin monomers, which are then cross-
linked by factor XIIIa, resulting in formation of insoluble fibrin clots (Figure 4.1).
Thrombin is rapidly inactivated by antithrombin III, resulting in formation of thrombin-
antithrombin III dimers (TAT). The formation of fibrin clots is counteracted via their

lysis by the enzyme plasrnin, a key member of the fibrinolytic system. Conversion of

123

plasminogen to plasmin is controlled by the plasminogen activators, urokinase
plasrrrinogen activator (uPA) and tissue-type plasminogen activator (tPA). In ttu‘n,
plasmin activity is controlled by a2-antiplasmin, and tPA and uPA activities are inhibited
by plasminogen activator inhibitor-1 (PAI-l ), and important physiological regulator of
fibrinolysis (Figure 4.1).

Several biomarkers can be measured to evaluate coagulation system activation
and fibrinolysis. An increase in plasma TAT concentration, decrease in plasma fibrinogen
concentration, and/or tissue fibrin deposition are indicative of an activated coagulation
system. Pharmacologic inhibition of coagulation system activation (see Figure 4.1) is
accomplished by inhibition of thrombin activation (e.g., heparin) and by agents (e.g.,
warfarin) that interfere with synthesis of vitamin K-dependent coagulation factors such as
fibrinogen. Profibrinolytic agents include streptokinase (SK), tPA, and urokinase, which
increase plasmin-mediated fibrinolysis. Elevated plasma concentrations of PAH can
impair fibrinolysis.

Recent results suggested the possibility that the hemostatic system is important in
this model (Luyendyk et al., 2004c). For example, Sinusoidal endothelial cell (SEC)
homeostasis is altered before the onset of significant hepatic parenchymal cell injury in
LPS/RAN-treated rats (Luyendyk et al., 2004c). Additionally, significant hepatic fibrin
deposition resulted from LPS/RAN-cotreatment but not fi'om treatment with either agent
alone (Luyendyk et al., 2004c). Also, consistent with impaired fibrinolysis, hepatic
expression of PAH was augmented in LPS/RAN-treated rats (Luyendyk et al., 2004c).

Such suppression would be expected to enhance fibrin deposition. However, whether or

124

not the coagulation system is activated afier LPS/RAN-treatment and the consequences
of hemostatic dysregulation in LPS/RAN-induced liver injury have not been determined.

The studies presented here tested the hypothesis that the hemostatic system is activated
before the onset of liver injury and is critical for HPC injury in LPS/RAN-treated rats.
Toward this end, biomarkers of thrombin activation, hepatic fibrin deposition, and serum
concentration of PAH were evaluated at times before and afier the development of
parenchymal cell injury. Rats were treated with the anticoagulant heparin or the
fibrinolytic agent streptokinase (SK) to investigate the importance of the hemostatic
system and fibrin deposition in LPS/RAN-induced parenchymal cell injury. In addition,
we examined whether LPS/RAN cotreatment caused liver hypoxia as one consequence of

fibrin deposition that could promote hepatocellular injury.

4.3 Materials and Methods

4.3.1 Materials

For information on this topic please refer to Chapter 2 Materials and Methods.

4.3.2 Animals

For information on this topic please refer to Chapter 2 Materials and Methods.

4.3.3 Experimental Protocol

Plasma was collected by drawing blood from the vena cava into a syringe

containing sodium citrate (final concentration, 0.38%), and rats were then killed

125

Figure 4.1 Balancing of coagulation a_nd fibrinolysis by the cofltion system.
Formation of fibrin clots is regulated by both the procoagulant and fibrinolytic arm of the
hemostatic system. Coagulation activation via either the intrinsic or extrinsic pathways
results in conversion of prothrombin to thrombin, conversion of soluble fibrinogen to
insoluble fibrin monomers, ultimately producing fibrin clots. The process of fibrinolysis
balances this system through degredation of fibrin clots by the enzyme plasmin.
Plasminogen activators (i.e., uPA and tPA) are important for activation of plasmin and
can be negatively regulated by plasminogen activator inhibitor-1. Figure developed by

Robert Roth.

126

Intrinsic System Extrinsic System

 

 

 

(surface contact) (tla no damage)

; C \ /

[.9

:3 x
3» xm

1 8 Protirrombin—->Thrombir.

go

: 3 Fibrinogen Fibrin -—v Fibrin

’;::; (monomer) (polymer)
51’ Clot Dissolution —— Fibrin :
3,. Degradation Plasmin
O
E Plasminogen
E @Activators (tPA, uPA)

 

 

 

P Al- 1 Plasminogen

127

by exsanguination fiom the dorsal aorta. Blood was allowed to clot at room temperature,
and serum was collected and stored at 20° C until use. Representative (34 mm) Slices of
the ventral portion of the left lateral liver lobe were collected and fixed in 10% neutral
buffered formalin. For more information on this section please refer to Chapter 2
Materials and Methods.

Inhibition of coagulation system activation was achieved by administration of
heparin. Rats were treated with LPS/RAN as above, but 1 h before RAN treatment,
heparin (3000 U/kg, so.) or sterile saline was administered. Rats were killed 6 h after
RAN administration, and serum and liver samples were taken. For studies with SK, rats
were treated with LPS and RAN, then two h later they were given SK (25,000 U/kg) or
sterile saline (i.p.). Three h later, a second administration of SK (20,000 U/kg) was given.
Rats were killed 6 h after treatment with RAN, and serum and liver samples were

collected as described above.

4.3.4 Hepatotoxicity assessment

For information on this topic please refer to Chapter 3 Materials and Methods.

4.3.5 Determination of serum PAI-l, plasma fibrinogen and plasma thrombin-
antithrombin dimer (TAT) concentrations

Total serum PAI-l concentration was evaluated using a commercially available
ELISA purchased fi'om American Diagnostica Inc. (Greenwich, CT.). The concentration
of functionally active PAI-l in serum was assessed using a commercially available

ELISA purchased from Molecular Innovations Inc. (Southfield, MI). Plasma fibrinogen

128

was determined fi'om thrombin clotting time of diluted samples using a fibrometer and a
commercially available kit (B4233) fi'om Dade-Behring Inc. (Deerfield, IL). Plasma TAT

concentration was determined using kit #OWMGI 5 from Dade-Behring.

4.3.6 Fibrin and RECA-l immunohistochemistry

A 1cm3 block of liver cut from the left medial lobe was frozen for 8 rrrinutes in
liquid nitrogen-chilled isopentane. For liver endothelial cell immunostaining, 8 rim-thick
sections of flow liver were fixed in acetone (4°C) for 5 minutes. Sections were
incubated in a blocking solution consisting of PBS with 5% goat serum (Vector
Laboratories, Burlingame, CA) for 30 minutes, then overnight at 4°C in blocking solution
containing diluted (1:20) mouse anti-rat RECA-l (rat endothelial cell antigen-1) antibody
(Serotec, Inc., Raleigh, NC). The RECA-l antibody binds to rat endothelium but not to
other cell types (Duijvestijn et al., 1992). In the liver, this antibody stains both SECS and
endothelial cells of larger vessels (Copple et al., 2002a). After incubation with the
RECA-l antibody, sections were washed three times, 5 minutes each, with PBS then
incubated for 3 hours at room temperature with goat anti-mouse secondary antibody
conjugated to Alexa 594 (1:1000, Molecular Probes, Eugene, OR) in blocking solution
containing 2% rat serum. Sections were washed three times, 5 minutes each, with PBS
and visualized using a fluorescent microscope. For more information on fibrin
immunohistochemistry, please refer to Chapter 3 Materials and Methods. For both
protocols, no staining was observed in controls in which the primary or secondary
antibody was eliminated fi'om the staining protocol. All treatment groups that were

compared morphometrically were stained irnmunohistochemically at the same time.

129

4.3.7 Morphometric evaluation

For more information on this section please refer to Chapter 3 Materials and

Methods.

4.3.8 Evaluation of liver hypoxia.

Liver hypoxia was evaluated by two methods. First, hypoxic areas of liver were
identified by injection of pimonidazole (PM) and immunostaining for PIM-modified
proteins. PIM is a 2-nitroimidazole marker of hypoxia and has been used to identify
regions of hypoxia in liver (Arteel et al., 1998;Arteel et al., 1995). Rats were given 120
mg/kg Hypoxyprobem-l (PIM hydrochloride; Chemicon International Inc., Temecula,
CA) i.p. two hours before they were killed. PIM-adduct immunostaining was performed
on formalin-fixed liver samples sectioned at 5 urn. Tissues were deparaffinized at room
temperature by 3 X 5 minute incubations in xylene, 2 X 5 min incubations in 100%
ethanol, 1 X 5 min incubation in 95% ethanol, 1 X 5 min incubation in 75% ethanol, 1 X
5 min incubation in distilled water, 1 X 2 min incubation in distilled water + 0.2% Brij 35
(Fisher Scientific, Pittsburgh, PA), and l X 2 min incubation in PBS + 0.2% Brij 35.
Sections were washed 3 times with PBS then incubated in BioMeda pronase reagent
(Biomeda Corp., Foster City, CA) at 40° C for 40 min. They were then washed with PBS
+ 0.2% Brij 35 for 2 min and blocked for 5 min at room temperature in undiluted DAKO
serum-flee protein block solution (DakoCytomation, Carpinteria, CA). Sections were
incubated for 40 min at room temperature with the monoclonal hypoxyprobe antibody

(1 :50, Cherrricon International Inc., Temecula, CA) in PBS containing 0.2% Brij 35 and l

130

drop DAKO block solution/ml. Sections were washed 3X with PBS + 0.2% Brij 35, then
incubated for 3 hours at room temperature with rabbit anti-mouse secondary antibody
conjugated to Alexa 594 (1:500, Molecular Probes, Eugene, OR) in PBS containing 0.2%
Brij 35 and 1 drop DAKO block solution/ml. Sections were washed 3X with PBS
containing 0.2% Brij 35 and 3X with PBS and visualized using a fluorescent microscope.
Quantification of PIM immunostaining was performed using Scion Image Beta 4.0.2 as
for fibrin staining (above). Background was estimated to be the average pixel intensity
identified in periportal regions of Veh/Veh-treated livers (i.e., an area where no hypoxia
occurs (Arteel et al., 1995). An increase in positive immunostaining for PIM-modified
proteins indicates hypoxia in the liver tissue.

Second, immunostaining for hypoxia-inducible factor-1a (HIP-1a) was
performed. HIF-la is a key regulator of responses to hypoxia (Semenza, 1999) and
stabilization of HIF-l or protein can be detected immunohistochernically in hepatocyte
nuclei in hypoxic liver (Stroka et al., 2001). For HIF-lor immunostaining, 8 rim-thick
sections of frozen liver were fixed in 4% neutral-buffered formalin at room temperature
for 10 minutes. Sections were blocked with PBS containing 5% goat serum (Vector
Laboratories, Burlingame, CA) for 30 minutes, and this was followed by incubation
overnight at 4°C with mouse anti- HIF-lot antibody (NB 100-123, Novus Biologicals,
Littleton, CO) diluted (1:100) in PBS containing 5% goat serum. After incubation with
the HIP-1a antibody, sections were washed three times, 5 minutes each, with PBS then
incubated for 3 hours at room temperature with goat anti-mouse secondary antibody
conjugated to Alexa 594 (1:500, Molecular Probes, Eugene, OR) in PBS containing 5%

goat serum and 2% rat serum. Sections were then washed three times, 5 minutes each,

131

with PBS and visualized using a fluorescent microscope. Quantification of liver HIF-la
staining was performed using Scion Image Beta 4.0.2 as with fibrin staining. An increase
in nuclear staining of HIF-la indicates liver hypoxia. For each of these procedures all

slides were stained and visualized on the same day.

4.3.9 Statistical Analysis

Two-way analysis of variance with Tukey’s test for multiple comparisons was
used for analysis of clinical chemistry in the streptokinase study, immunohistochemistry,
ELISA, and fibrinogen measurements. Student’s t-test was used to compare PAI-l
ELISA data in the same treatment group between times. One-way analysis of variance
with Tukey’s test for multiple comparisons was used in the heparin study. The criterion

for significance for all studies was p<0.05.

4.4 Results

4.4.1 Coagulation system activation after LPS/RAN treatment

Coagulation system activation was evaluated in rats treated with LPS/RAN at a
time prior to the onset of significant liver injury (i.e., 2 h). Consistent with previous
results (Luyendyk et al., 2003b), serum ALT activity was not changed in any treatment
group by 2 h (Figure 4.2A). Relative to Veh/Veh-treated rats, plasma fibrinogen
concentration was not significantly changed by LPSNeh-treatment, but it was decreased
slightly in rats treated with Veh/RAN (Figure 4.28). In contrast, a marked decrease
(~85%) was observed in LPS/RAN-treated rats (Figure 4.2B). Thrombin was estimated

by measuring the plasma concentration of TAT. TAT concentration was not significantly

132

changed in Veh/RAN-treated rats. However, treatment with LPSNeh resulted in a
significant increase in TAT concentration (~5-fold), whereas a more pronounced increase

(~14-fold) was observed after LPS/RAN treatment (Figure 4.2C).

4.4.2 Altered sinusoidal endothelial cell (SEC) function after treatment with
LPS/RAN

Since hepatic SECS remove circulating HA, an increase in serum HA
concentration has been used as a biomarker of altered SEC function. We reported
previously that HA concentration was slightly increased in Veh/RAN-treated rats 3h after
treatment (Luyendyk et al., 2004c). Results presented in Table 4.1 confirm these findings
and demonstrate that this increase is transient: HA concentration in Veh/RAN-treated rats
was not different fiom VehNeh-treated rats from 6—24 h. In rats treated with LPSNeh,
serum HA concentration increased by 3 h and continued to increase until 12 h, then
leveled off thereafter. (Table 4.1). Rats treated with LPS/RAN had a serum HA
concentration significantly greater than rats treated with either LPSNeh or Veh/RAN at
3, 6, and 24 h.

To evaluate whether overt SEC injury occurred after LPS/RAN treatment, livers
were stained irnmunohistochemically for RECA-l . Decreased RECA-l staining intensity
is associated with endothelial cell loss in other models of hepatotoxicity.(Yee et al.,
2003b) No treatment caused a significant change in RECA-l staining intensity at either 3

or 6 b (Figure 4.3).

133

Figure 4.2: Coaggation system activation after LPS/RAN treatment. Rats were given
LPS (44.4 X 106 EU/kg, iv) or its Veh 2 h prior to administration of RAN (30 mg/kg) or
its Veh. Hepatic parenchymal cell injury was estimated 2 h after RAN administration by
increases in serum ALT activity (A). Coagulation system activation was evaluated by
measuring plasma fibrinogen concentration (B) and plasma TAT concentration (C). n=4-
6. Data are expressed as mean :1: SEM. *Significantly different from Veh/Veh-treated

rats. “Significantly different from all other treatments. (p<0.05)

134

 

 

 

 

 

 

t . a

L N

n m

m r.

S

I .. ._

n.

m r. v

, V

A oi r o r o B m ..u ..u o c o. ..u r r
m m w s m m m m m m
:5. 33:2 5< 32.95 585.9... :33 FE

135

 

 

 

 

 

 

 

53¢ .52. ”>2 :5
5.2-3.2: m a 5 NA
5:29 Se H as an H ea .5.» H .5 a a H «a
Size: 83 H 3.. 5... H 3.9. 83 H 3.2 we .u H N3...
5.532 e... H 3a. 8.» H :3 2; H we Sb H .3
Swat? :2 H N91 83 H 3% and.” H 8.1 88 H seat

 

 

 

 

 

 

Hugo Pr E E 833530: when CemEZ £33.03. mag 2,» £8 SammEda mm
m: 838mg: 0». mmo $53330: 3 Com 82:9. géamnaa 8.8 w. a. 5. Ba no : wane
E @853". 509.3 88 v9 «3%. 9% m8 axvammaa 8 Ba»: H mmz. imam—853,
@5295 Won. <o~<<a~rndm$a Ba. $555833 95.2.2: ~88 w: 099 #8538 m" 9m:

man. cage.

136

4.4.3 Effect of LPS/RAN treatment on hepatic fibrin deposition.

Minimal fibrin staining was associated with the larger vessels but not in the
sinusoids in livers from Veh/Veh-treated rats (Figure 4.4A). This staining occurs post
mortem and can be prevented by perfusing the liver with heparin before its removal (data
not shown). Marked panlobular staining occurred in livers of LPS/RAN-treated rats
(Figure 4.4B and C). Morphometric analysis revealed no significant increase in hepatic
fibrin deposits in RAN-treated rats at either 3 or 6 h (Figure 4.4D), confirming earlier
results (Luyendyk et al., 2004c). A slight increase in fibrin staining occurred in LPSNeh-
treated rats that became significant at 6 h (Figure 4.4D). Staining in LPS/RAN-treated

rats was significantly greater than in all other treatment groups at 3 and 6 h (Figure 4.4D).

4.4.4 Efl'ect of LPS/RAN treatment on serum PAI-l concentration

The concentrations of total PAI-l protein (Figure 4.5A) and active PAI-l (Figure
4.53) were evaluated in rats 3 and 6 h after RAN treatment. Treatment with Veh/RAN
caused a slight increase in total serum PAI-l concentration at 3 and 6 h but was without
effect on the concentration of active PAI-l . The serum concentrations of both total and
active PAI-l were significantly increased in LPSNeh-treated rats at 3 h, but these
increases waned by 6 h. The serum concentrations of both total and active PAI-l were
increased in LPS/RAN-treated rats at 3 and 6 h to a greater degree than after treatment
with LPS/Veh or Veh/RAN. In contrast to rats treated with LPSNeh, PAI-l
concentration in LPS/RAN-treated rats did not decrease significantly between these

times.

137

Figure 4.3: Q_ua_ntification of RECA-l staining in livers after L_PS/RAN treatment. Rats
were given LPS (44.4 x 106 EU/kg, iv) or its Veh 2 h prior to administration ofRAN (30
mg/kg) or its Veh. Three or 6 h after RAN treatment, livers were removed and stained
immunohistochemically for RECA-l as described in Materials and Methods. The total
area of RECA—l staining was evaluated in 10 randomly chosen fields (IOOX) per liver
section and analyzed morphometrically as described in Materials and Methods. n=3-7
rats per group at each time. None of the treatments caused a significant change in RECA-

1 staining intensity relative to Veh/Veh-treated rats at either 3 or 6 h. (p<0.05)

138

 

 

:22 LPSNeh

0,4 . — Veh/RAN
C: LPS/RAN

 

 

3210.
0,000

$95 .32 eczema:
33mm

5 . — VehNeh

Time after RAN (hours)

139

Figure 4.4: L_iver fibrin deposition in LPS/RAN-trezied rats Rats were given LPS (44.4
X 106 EU/kg, iv) or its Veh 2 h prior to administration of RAN (30 mg/kg) or its Veh.
Three or 6 h after RAN treatment, livers were removed and stained
immunohistochemically for fibrin as described in Materials and Methods. (A)
Representative photomicrograph (IOOX magnification) of fibrin staining in liver of a
VehNeh-treated rat showing minimal staining (black) in the intima of larger vessels. (B)
Representative photomicrograph showing panlobular fibrin staining characteristic of
livers taken from LPS/RAN-treated rats 3 h after RAN. (C) Representative
photomicrograph showing panlobular sinusoidal fibrin staining characteristic of livers
taken from LPS/RAN-treated rats 6 h after RAN. PP, periportal region. CL, centrilobular
region. (D) Quantification of liver fibrin staining in rats treated with LPS and/or RAN.
n=3-6. Data are expressed as mean 1: SEM. *Significantly different from Veh/Veh-

treated rats at that time. ”Significantly different from all other treatments at that time.

(p<0.05)

140

 

PP
CL , ’

PP

CL

 

 

 

 

.PP'.

cL

 

 

Fibrin
(fraction total
9 9 9
3 S 8

 

3 h e h
Time after RAN (hours)

141

 

4.4.5 Effect of SK on LPS/RAN-induced liver injury

To investigate the role of fibrin clots in LPS/RAN-induced liver injury, the
fibrinolytic agent SK was used. SK treatment caused a slight but statistically significant
decrease in liver fibrin staining in Veh/Veh-treated rats (Figure 4.6A). Fibrin deposition
was elevated by LPS/RAN treatment, and this effect was markedly reduced (60%) by SK
(Figure 4.6A), confirming the effectiveness of the SK treatment. The effect of SK
treatment on hepatic parenchymal cell injury was estimated 6 h after LPS/RAN treatment
by changes in serum activities of ALT (Figure 4.63) and AST (Figure 4.6C). SK was
without effect in Veh/Veh-treated rats but significantly reduced the serum elevations in

ALT (~50%) and AST (~35%) in LPS/RAN-treated rats.

4.4.6 Effect of heparin on LPS/RAN-induced liver injury

The importance of an activated coagulation system in LPS/RAN-induced liver
injury was evaluated by treating rats with the anticoagulant, heparin. Activation of the
coagulation system was evaluated by changes in plasma fibrinogen concentration. The
concentration of plasma fibrinogen was significantly decreased in rats treated with
LPS/RAN, and this decrease was prevented by coadministration of heparin (Figure 4.7A).
LPS/RAN treatment significantly increased serum ALT and AST activities, and this

increase was prevented by coadministration of heparin (Figure 4.7B, C).

4.4.7 LPS/RAN treatment and liver hypoxia

142

 

Figure 4.5: Serum concentration of PAH after treatment with LPS/RAN. The
concentrations of (A) total PAH and (B) active PAI-l were evaluated at 3 h and 6 h in
serum taken from rats treated with LPS and/or RAN. n=6-7. Data are expressed as mean
:t SEM. *Significantly different from Veh/Veh-treated rats at that time. “Significantly
different from all other treatments at that time. “Significantly different from the same

treatment at 3 h. (p<0.05)

143

Total PAl-1 (nglml)

>

.5
g

A;

§

N
:8

 

 

 

 

 

Active PAI-1 (nglml)

 

1'
I
H; U;
3 h a It
Time after RAN (hours)

144

d

§

A

 

 

 

 

 

w — Veh/V0“
* r22: LPSNeh
L — Veh/RAN
:2: LPS/RAN

I

a!
- inl—nij-

3 h s h

Time after RAN (hours)

Figure 4.6: Effect of streptokinase (SK) on L_PS/RAN-induced liver iniurv. Rats were
given SK or saline 1 h and 4 h after treatment with either VehNeh or LPS/RAN. Livers
were removed at 6 h and stained immunohistochemically for fibrin, which was quantified
as described in Materials and Methods (A). Hepatic parenchymal cell injury was
estimated at 6 h by increases in serum ALT (B) and AST (C) activities. n=6-8. Data are
expressed as mean i- SEM. *Significantly different from VehNeh-treated rats.

”Significantly different from LPS/RANNeh and VehNeh/SK-treated rats. (p<0.05).

145

 

 

 

0 030
0.025
0 020
0 015 1
0 010
0.005 '

$25 .32 eczema
ESE

0.000 -

v

 

 

ooooowo
05050
332211

3:: 33:2 5<

C

SK

VehNeh

 

Veh

 

000000
owooo
5 321

600 1

:5. E52 5<

LPSIRAN

146

Rats were treated with LPS and/or RAN and immunostaining for PIM-adducts
and HIF-l or protein was evaluated at a time near the onset of hepatotoxicity in
LPS/RAN-treated rats (i.e., 3 h after RAN administration). In contrast to previous
experiments (Luyendyk et al., 2003b), serum ALT activity was significantly increased at
3 h in LPS/RAN-treated rats compared to rats given only LPS or RAN (data not shown),
suggesting that 3 h marks the approximate time of onset of liver injury in this model.
LPS/RAN hepatotoxicity was not influenced by PIM administration (data not shown).
Little PIM-adduct staining was observed in livers of Veh/Veh-treated rats (Figure 4.8A).
PIM-adduct staining increased slightly in livers of rats treated with LPSNeh (Figure
4.8B) or Veh/RAN (Figure 4.8C). By contrast, a dramatic increase in the area and
intensity of positive PIM-adduct staining was observed in livers of LPS/RAN-treated rats
(Figure 4.8D). No zonal specificity was observed, although staining appeared darker in
midzonal and centrilobular regions. Quantification of PIM-adduct staining revealed
statistically significant increases in PIM-adduct staining in livers of rats treated with
LPSN eh or Veh/RAN (Figure 4.8E). PIM-adduct staining in livers of LPS/RAN-treated
rats was markedly greater (~10 times) than staining in livers of rats treated with LPS or
RAN alone (Figure 4.8E).

Immunostaining for I-IIF-la protein revealed mild and scattered nuclear staining
in livers of VehNeh-treated rats (Figure 4.9A). Although livers were removed fi'om the
animals and fiozen rapidly, this staining might have resulted from stabilization of HIF-la
during tissue removal. HIF-la staining in livers of LPSNeh-treated (Figure 4.9B) and
Veh/RAN-treated (Figure 4.9C) rats also appeared as mild nuclear staining that was not

significantly different from VehNeh-treated rats (Figure 4.9B). In livers from LPS/RAN-

147

treated rats, HIF-la nuclear staining was significantly greater than in livers from rats

treated with either agent alone.

4.5 Discussion

Previous studies demonstrated that LPS/RAN-treated rats develop liver injury
characterized by midzonal hepatocellular necrosis and elevations in ALT and AST
activities by 6 h (Luyendyk et al., 2003b). The studies presented here tested the
hypothesis that the coagulation system is activated after LPS/RAN treatment. At a time
before the onset of liver injury in LPS/RAN-treated rats (i.e., 2 h, Figure 4.2A), a
pronounced decrease in plasma fibrinogen concentration (Figure 4.2B) was associated
with a significant increase in plasma TAT concentration (Figure 4.2C). This result
suggests that thrombin activation (i.e., activation of the coagulation system) occurred
prior to liver injury. Treatment with LPSN eh, which at this dose was not hepatotoxic
within 24 h (Luyendyk et al., 2003b), caused a significant increase in plasma TAT
concentration (Figure 4.2C), suggesting activation of thrombin; however, the degree of
activation was insufficient to cause a decrease in plasma fibrinogen (Figure 4.2B).
Interestingly, a slight but statistically significant decrease in plasma fibrinogen occurred
without an increase in TAT concentration in Veh/RAN-treated rats (Figure 4.2B and C),
suggesting consumption of fibrinogen that was not related to the action of thrombin.
Other proteases, such as matrix metalloproteinases, can also degrade fibrinogen without
coagulation system activation (Bini et al., 1996), but the cause for this decrease in

Veh/RAN-treated rats is not understood. Overall, the data indicate that the coagulation

148

system is markedly activated after treatment with LPS/RAN before the onset of hepatic
parenchymal cell injury.

One contributor to coagulation system activation in LPS/RAN-treated rats could
be endothelial cell activation (Hewett and Roth, 1993; Colman, 1994). Confirming earlier
results (Luyendyk et al., 2004c), the concentration of serum HA was elevated (~2.5-fold)
in LPSNeh-treated rats (Table 4.1) at 3 h. Veh/RAN treatment also increased serum HA.
In LPS/RAN-cotreated rats, the effects of the two agents appeared to be additive at this
time. No additional effect of Veh/RAN treatment occurred after 3 h. The rate of increase
from 3-12 h was similar in both LPS-treated groups irrespective of RAN cotreatment
(LPSNeh, 19.1 ng/ml/h, 12:0.99; LPS/RAN, 17.7 ng/ml/h, r2=0.99). Taken together,
these results suggest the occurrence of an early, transient effect of RAN and a sustained
LPS effect on SEC fimction. The early changes in serum HA concentration were not
accompanied by altered RECA-l staining, suggesting the absence of SEC destruction.
Between 12-24 h, the concentration of HA increased markedly in the cotreated rats. The
latter change may be a consequence of overt liver injury (Luyendyk et al., 2003b).
Overall, these results suggest that SEC dysfimction in LPS/RAN-treated rats occurred at
a time prior to hepatocellular injury and to a greater degree than in LPSNeh-treated rats.
A modest increase in HA concentration has been reported previously at doses of LPS that
do not cause hepatocellular injury (Y ee et al., 2003b;Luyendyk et al., 2003a). Although
its contribution is not fully understood, this perturbation of SEC function might be
important for liver injury afier LPS/RAN treatment. For example, HA has been reported
to enhance expression of PAH (Horton et al., 2000), increasing the likelihood of

sustained fibrin clotting.

149

Figure 4.7: Effect of hepg’n on LPS/RAN-induced liver injurv. LPS/RAN-treated rats
were given heparin or saline 1 h after treatment with LPS. Plasma fibrinogen (A) and
hepatic parenchymal cell injury were evaluated 6 h afier RAN treatment. Hepatic
parenchymal cell injury was estimated by increases in serum ALT (B) and AST (C)
activities. n=3-9. Data are expressed as mean :h SEM. *Significantly different from

Veh/Veh/Veh-treated rats. # Significantly different from LPSNeh/RAN-treated rats.

(p<0.05)

150

W

 

 

 

 

N 0 0 0 0
wwmm
fl

:5. 3.28 S<

Aeeeeeosm
2211

3395 semester.

5004

 

 

mmmm

:3. £28

000
000
321

Heparin
LPSIRAN

Veh

VehNeh

151

Figure 4.8: Effect of LPS/RAN treatment onMic PIM-adduct star_ning._Rats were
treated with 44.4 x 106 EU/kg LPS or its Veh (iv), then two h later with 30 mg/kg RAN
or its Veh (iv). PIM (120 mg/kg, ip) was injected 2 h after RAN treatment. Livers were
removed 3 h after RAN treatment and processed for PIM-adduct immunohistochemistry
as described in Materials and Methods. (A) Representative photomicrograph of PIM-
adduct staining in liver from a VehNeh-treated rat showing minimal staining (black).
Representative photomicrographs from rats treated with LPSNeh (B) and Veh/RAN (C)
show a slight increase in PIM staining. (D) Representative image from LPS/RAN-
cotreated rat showing marked, panlobular PIM-adduct staining. For (E), the area of
positive PIM staining in 10 randomly chosen, IOOX fields per tissue was determined
morphometrically as described in Materials and Methods. Data are expressed as mean i
SEM. n=5-8 rats. *Significantly different from the respective group not given LPS;

”Sigrificantly different fi'om respective group not given RAN (p<0.05)

152

 

 

 

 

 

 

 

 

 

s
m.
. a
.yV
m
V
7. 6. 5. 4. 3 2. 1. 0.
o 0 o o o o o 0
.8: .38 3.8.2:
2.6%:

Veh

153

Figure 4.9: Effect of L_PS/RAN treatment on hepatic HIF-la staining. Rats were treated
with 44.4 x 106 EU/kg LPS or its Veh (iv), then two h later with 30 mg/kg RAN or its
Veh (iv). Livers were removed 3 h after RAN treatment and processed for HIF -1a
immunohistochemistry as described in Materials and Methods. (A) Representative
photomicrograph of HIP-1a staining in liver from a VehNeh-treated rat showing modest
nuclear staining (black). Representative photomicrographs from rats treated with
LPS/V eh (B) and Veh/RAN (C) show HIF-lcr staining similar to Veh/V eh. (D)
Representative image from LPS/RAN-cotreated rat showing marked panlobular HIF -1 or
staining. For (E), the area of positive HIP-la staining in 10 randomly chosen, IOOX
fields per tissue was determined morphometrically as described in Materials and
Methods. Data are expressed as mean :1: SEM. n=5-8 rats ‘Significantly different from the
respective group not given LPS; “Significantly different from respective group not given

RAN (p<0.05)

154

 

 

 

 

 

 

 

 

 

LPS Veh LPS

RAN

155

One interpretation of the results is that activation of the hemostatic system by LPS
is magnified by RAN cotreatment. Treatment with LPSNeh caused slight thrombin
activation and the appearance of fibrin in the liver; both of these were more pronounced
in LPS/RAN-treated rats (Figure 4.2, 4.3). Thus, activation of the coagulation system is
associated with increased fibrin deposition in livers of LPS/RAN-treated rats. An
impaired fibrinolytic system could also contribute to microvascular fibrin deposits by
decreasing plasmin’s capacity to degrade fibrin (Colman, 1994). PAL] is an important
downregulator of plasmin activation and is expressed in animals and in cells after
exposure to numerous agents, including cytokines and LPS (Binder et al., 2002;
Hamaguchi et al., 2003; Sawdey and Loskutoff, 1991). In animal models of endotoxemia,
antibody-mediated inhibition or genetic knockout of PAH significantly attenuated fibrin
deposition in tissues, suggesting that PAI-l activity is important for this effect of LPS
exposure (Montes et al., 2000; Abrahamsson et al., 1996; Savov et al., 2003).
Interestingly, hepatic expression of the gene encoding PAI-l is enhanced in LPS/RAN-
treated rats (Luyendyk et al., 20040), and this expression is mirrored by an augmented
concentration of PAH protein in serum (Figure 4.5). Moreover, the increase in serum
PAI-l persisted in LPS/RAN-treated rats, whereas it waned by 6 h in rats exposed only to
LPS. Accordingly, persistent PAI-l overexpression might contribute to stabilizing fibrin
clots in livers of LPS/RAN-treated rats in the face of ongoing coagulation system
activation.

Hepatic fibrin deposition occurs in numerous models of hepatotoxicity, and
anticoagulants afford protection against hepatocellular injury in some of them (Luyendyk

et al., 2003a; Fujiwara et al., 1988; Copple et al., 2002b; Yee et al., 2003d; Kinser et al.,

156

2002; Pearson et al., 1996b). In LPS/RAN-treated rats, anti-coagulation by heparin
prevented the development of hepatocellular injury (Figure 4.7). In addition, SK
treatment significantly reduced hepatic fibrin deposition as well as hepatic parenchymal
cell injury (Figure 4.6). These results suggest that fibrin deposition is important for
LPS/RAN-induced hepatotoxicity.

The mechanism by which fibrin clots cause toxicity in this model is not
understood. One potential consequence of fibrin deposition is disruption of sinusoidal
hepatic blood flow leading to hypoxia. For example, fibrin deposition and hypoxia
precede centrilobular oncotic necrosis in rats treated with monocrotaline (MCT) (Copple
eta1., 2004a). The anticoagulant warfarin significantly reduces fibrin deposition, hypoxia,
and hepatocellular injury after MCT exposure, suggesting a causal role for fibrin and
hypoxia in the injury (Copple et al., 2002b; Copple et al., 2004a). Exposure to hypoxia
alone is sufficient to cause hepatic parenchymal cell injury in isolated, perfused livers
(Marotto et al., 1988; Lemasters et al., 1981). Interestingly, the severity of lesions caused
by LPS is enhanced by exposing rats to a hypoxic atmosphere (Shibayama, 1987).
Indeed, enhanced hepatic expression of hypoxia-regulated genes occurred in LPS/RAN-
treated rats (Luyendyk et al., 2004c), and two markers of hypoxia were observed in livers
of LPS/RAN-treated rats (Figure 4.8 and Figure 4.9,). Taken together, these results
suggest that LPS/RAN-treatment results in liver hypoxia with a timeframe similar to that
of hepatic fibrin deposition and hepatocellular injury.

Appreciation of mechanisms by which the hemostatic system causes
hepatotoxicity in the LPS/RAN rat model might yield insight into sensitivity of people to

idiosyncratic RAN hepatotoxicity. A well-defined association between idiosyncratic

157

hepatotoxicity and hemostatic dysregulation has not been previously proposed, and
epidemiological studies examining this relationship are lacking. As in the rat, RAN
administration alone does not appear to enhance coagulation in people (Stadnicki, 1984).
The results presented herein suggest, however, that RAN augments activation of the
hemostatic system caused by exposure to inflammagens (e. g., LPS). Interestingly, a
recent case report described RAN hepatotoxicity in a patient with a deficiency in the
vitamin-K dependent anticoagulant factor, Protein 8 (V alois et al., 2003 ), suggesting that
genetic predisposition favoring coagulation might be a susceptibility factor for RAN
idiosyncrasy. Polymorphisms in several components of the hemostatic system, including
PAI-l, have been identified (for review, see Lane and Grant, 2000) and could represent a
potential interaction between genetic and environmental (i.e., LPS exposure) factors in
causing idiosyncratic reactions. Furthermore, it seems reasonable to speculate that
consequences of fibrin deposition such as tissue hypoxia might be important in drug
idiosyncrasy. In this regard, LPS treatment results in liver injury in hypoxic rats exposed
to halothane, an inhalation anesthetic associated with idiosyncratic hepatotoxicity (Lind
et al., 1984). Another drug associated with infrequent hepatotoxicity during its clinical
trials is the quinoxalinone anxiolytic, panadiplon (Ulrich et al., 2001), and interestingly,
treatment of hepatocytes with panadiplon rendered them more sensitive to hypoxia-
induced cell death (Bacon et al., 1996). Despite these associations, much remains to be
understood about the interplay of inflammation, hypoxia, and the hemostatic system in
idiosyncratic hepatotoxicity.

Figure 4.10 summarizes RAN’s influence on LPS-induced activation of both the

procoagulant and fibrinolytic arms of the hemostatic system and the relationship of this

158

action to liver injury. At a time prior to liver injury, RAN enhanced LPS-induced SEC
dysfunction. Furthermore, a decrease in plasma fibrinogen, increase in TAT
concentration and increase in hepatic fibrin deposition all occurred in livers of LPS/RAN-
treated rats before the onset of liver injury. RAN cotreatment also caused a persistent
increase in serum PAI-l triggered by LPS beginning at a time before liver injury,
suggesting that RAN impairs activity of the fibrinolytic arm of the hemostatic system.
Liver injury from LPS/RAN treatment was significantly attenuated by activation of
fibrinolysis with SK and abolished by inhibition of coagulation system activation by
heparin. Overall, the results show that RAN can augment LPS-induced hepatic fibrin
deposition in rats, leading to marked fibrin accumulation and tissue hypoxia as a
fimctional consequence. Furthermore, the data suggest that this procoagulant state is

critical for liver injury by a mechanism dependent on hepatic fibrin deposition.

159

Figure 4.10: Effect of RAN on LPS-induced fibrin deposition Me connection to liver
m Nonhepatotoxic doses of LPS alter SEC homeostasis leading to a procoagulant
state and modest fibrin deposition (see Figure 1, Table l, and Fig 3). Additionally, LPS
exposure increases expression of PAH (see Figure 4). This response alone is not
sufficient to cause parenchymal cell injury. RAN magnifies the effect of LPS on both
activation of the coagulation arm of the hemostatic system and on inhibition of
fibrinolysis by PAI-l, resulting in marked hepatic fibrin deposition (see Figure 4).
Anticoagulation or activation of fibrinolysis decreases hepatocellular injury fi'om

LPS/RAN, indicating that fibrin clots are critical factors in the injury (see Figure 5 and

Figure 6).

160

       
   
   

" li‘i.A.‘~

Altered SEC
Homeostasis System

  
  

Sinusoidal
Fibrin

161

Chapter 5

Role of coagulation system activation and liver PMN accumulation in LPS/RAN-treated

rats

162

5.1 Abstract

Rats cotreated with normally noninjurious doses of bacterial lipopolysaccharide
(LPS) and the histamine-2 (H2)-receptor antagonist ranitidine (RAN) develop
idiosyncrasy-like hepatocellular injury characterized by midzonal necrosis and neutrophil
(PMN) infiltration. Activation of the coagulation system and liver hypoxia occur in
LPS/RAN-treated rats with a timecourse sinrilar to the development of liver injury, and
the anticoagulant, heparin, significantly reduces LPS/RAN—induced liver injury. We
tested the hypothesis that an activated coagulation system is required for liver hypoxia in
LPS/RAN-treated rats. Furthermore, the effect of anticoagulation on liver PMN
accumulation was evaluated. Rats were given LPS (44.4 X 106 EU/kg) or its Veh, then
two h later they were given RAN (30 mg/kg) or its Veh. LPS treatment caused an
increase in the serum concentration of cytokine-induced chemoattractant-l (CINC-l) and
accumulation of PMNs in liver, neither of which were affected by RAN cotreatment.
Confirming previous results, heparin treatment significantly attenuated hepatocellular
injury in LPS/RAN-treated rats, as estimated by serum alanine aminotransferase (ALT)
activity 3 and 6 h afier drug administration. Heparin also reduced liver hypoxia in
LPS/RAN-treated rats but was without effect on serum CINC-l or liver PMN
accumulation. In vitro, exposure to hypoxia rendered hepatocytes sensitive to killing by
PMN elastase (PMN-E). Overall, the results suggest that the coagulation system is
important for the generation of liver hypoxia after LPS/RAN-cotreatment and that

hypoxia increases the sensitivity of HPCs to PMN-mediated killing.

163

5.2 Introduction

One possible consequence of hepatic fibrin deposition is disruption of blood flow
and liver hypoxia. For example, rats treated with a hepatotoxic dose of monocrotaline
(MCT) develop centrilobular liver injury that is preceded by hepatic fibrin deposition and
hypoxia (Copple et al., 2002a; Copple et al., 2004b). The anticoagulant warfarin
significantly reduced fibrin clot formation and liver hypoxia, and this was associated with
decreased parenchymal cell injury (Copple et al., 2004a). Liver hypoxia was also
observed in LPS/RAN-treated rats (Figure 4.8) at a time near the onset of hepatotoxicity
(Luyendyk, 2004b), but the role of thrombin generated by LPS/RAN-cotreatrnent in
causing hypoxia is not known.

Hypoxia/anoxia is sufficient to cause injury to isolated-perfused livers and
cultured hepatocytes (Lemasters et al., 1981; Marotto et al., 1988; Khan and O’Brien,
1997). Hypoxia can also increase susceptibility to hepatotoxicity from certain agents. For
example, LPS-induced liver injury is magnified in hypoxic rats (Shibayama, 1987),
suggesting an increased susceptibility to liver injury from inflammatory mediators. The
role of hypoxia and interactions between hypoxia and inflammation in LPS/RAN-treated
rats is not known. It is possible that hypoxia caused by LPS/RAN magnifies the ability of
inflammatory mediators to cause liver injury. As mentioned in chapter 1, several
inflammatory mediators contribute to liver injury from large doses of LPS, including
neutrophils (PMNs) and the mediators (e.g., ROS, elastase, cathepsin G) they release. In
accordance, PMNs accumulated in liver lesions caused by LPS/RAN-treatment in rats
(Luyendyk et al., 2003b; See chapter 2). Inasmuch as hypoxia and PMN accumulation

occur concurrently, the potential exists for the two to interact. Although the release of

164

inflammatory mediators by isolated PMNs was blunted under hypoxic conditions in vitro
(Derevianko et al., 1996), PMNs harvested fiom blood of people exposed to modest
hypoxia showed enhanced release of cytotoxic mediators and delayed apoptosis in vitro
(T amura et al., 2002), suggesting an indirect priming of PMN function under conditions
of hypoxia in vivo. In addition, hypoxia might alter sensitivity of hepatocytes to the toxic
effects of PMN-derived cytotoxic factors. In one study, hypoxia/reoxygenation rendered
cardiac myocytes sensitive to killing by PMN elastase during reoxygention (Buerke et al.,
1994). However, the effect of hypoxia on PMN accumulation/activation has not been
investigated in the LPS/RAN model, and the effect of hypoxia on sensitivity of
hepatocytes to the toxic effects of PMN-derived products is not understood.

The purpose of this study was to test the hypothesis that LPS/RAN-treatment
causes liver hypoxia by a mechanism dependent on thrombin activation. Furthermore, the
effect of anticoagulation on hepatic PMN accumulation was evaluated. To this end, the
effect of heparin treatment on markers of hepatocellular injury, fibrin deposition, and
liver hypoxia was evaluated after LPS/RAN treatment. The plasma concentration of the
PMN chemokine cytokine-induced chemoattractant-l (ClNC-l) and hepatic PMN
accumulation were evaluated at a time near the onset of injury in LPS/RAN-treated rats,
and the effect of heparin coadmirristration on each was assessed. In addition, an in vitro
system was developed to test the hypothesis that hypoxia renders hepatocytes sensitive to
killing by PMN elastase. The implications of these results and ongoing experiments are

discussed.

5.3 Materials and Methods

165

5.3.1 Materials

For information on this topic please refer to Chapter 2 Materials and Methods.

5.3.2 Animals

For information on this topic please refer to Chapter 2 Materials and Methods.

5.3.3 Experimental Protocol

Rats fasted for 24 hours were given 44.4 X 106 EU/kg LPS or its saline vehicle
(V eh) iv and food was then returned. Two hours later, 30 mg/kg RAN or sterile
phosphate-buffered saline (PBS) Veh was administered iv. RAN solution was
administered at 2 ml/kg at a rate of approximately 0.15 ml/min. To simplify treatment
nomenclature for the four groups, the following designations will be applied: Saline/PBS
(V eh/Veh), LPS/PBS (LPS/V eh), Saline/RAN (V eh/RAN), and LPS/RAN. Three or 6 h
after RAN treatment, rats were anesthetized with sodium pentobarbital (50 mg/kg, i.p.).
Plasma was collected by drawing blood from the vena cava into a syringe containing
sodium citrate (final concentration, 0.38%), and rats were then killed by exsanguination
from the dorsal aorta. This blood was allowed to clot at room temperature, and serum was
collected and stored at -20° C until use. Representative (3-4 mm) slices of the ventral
portion of the left lateral liver lobe were collected and fixed in 10% neutral buffered
formalin.

Inhibition of coagulation system activation was achieved by administration of

heparin. Rats were treated with LPS/RAN as above, but 1 h before RAN treatment,

166

heparin (3000 U/kg, 3.0.) or sterile saline was administered. Rats were killed 3 or 6 h after

RAN administration, and serum and liver samples were taken.

5.3.4 Hepatotoxicity Assessment

For information on this topic please refer to Chapter 4 Materials and Methods.

5.3.5 Fibrin immunohistochemistry and quantification

For information on this topic please refer to Chapter 3 Materials and Methods.

5.3.6 Evaluation of liver hypoxia

For information on this topic please refer to Chapter 4 Materials and Methods.

5.3.7 Evaluation of hepatic PMN accumulation and serum ClNC-l concentration
PMN immunohistochemistry was performed on formalin-fixed liver sections as
described previously (Y ee et al., 2003c). Hepatic PMN accumulation was evaluated by
identifying the average number of PMNs counted in 20 randomly selected, high-powered
fields (HPF, 400X). Slides were coded and the evaluator was unaware of treatment. The
serum concentration of CINC-l was determined using a commercially available, enzyme-

linked immunosorbent assay (ELISA) purchased fiom Assay Designs, Inc. (Ann Arbor,

MI).

5.3.8 Hepatocyte isolation

167

Hepatocytes were allowed to attach to plates in 5% Cosmic Calf Serum (Hyclone,
Logan, UT). Otherwise, for information on this topic please refer to Chapter 3 Materials

and Methods.

5.3.9 Efl'ect of hypoxia on elastase-induced cytotoxicity

Serum fi'ee Williams’ Medium B containing various concentrations (0, 0.7, 1.3,
2.2, 3.3, 4.4, 6.6, 8.8 units (U) of activity/ml) of human PMN elastase (PMN-E,
Molecular Innovations, Southfield, MI) was added to hepatocytes. PMN-E activity was
determined using a colorometric PMN-E substrate MeOSuc-Ala—Ala-Pro—Val-pNA
(Calbiochem, San Diego, CA). One unit of PMN-E activity was defined as the amount of
PMN-E (i.e. ug) of enzyme required to cause a change of 1.0 absorbance unit at 410 nm
in 10 minutes at 37 degrees C. PMN-E-treated cells were immediately transferred to
incubators containing either 20% or 5% oxygen (02) (balanced with nitrogen (N2)), with
carbon dioxide (C02) controlled at 5%. Two or 8 h later, the medium was collected, and
the remaining attached cells were lysed immediately with 1% triton X-lOO followed by
brief sonication. Media and lysates were centrifuged at 600xg for 5 min. Cytotoxicity was
assessed by measuring ALT release into the medium using Infinity-ALT reagent from
Thermo Electron Corp. (Louisville, CO). ALT activity in the medium was expressed as a

percent of total ALT activity (i.e., medium activity plus lysate activity).

5.3.10 Statistical Analysis

Two-way analysis of variance (ANOVA) with Tukey’s test for multiple

comparisons was used for the heparin study and for CINC-l and PMN accumulation

168

studies. Repeated measures two-way AN OVA was applied for in vitro experiments. The

criterion for significance for all studies was p<0.05.

5.4 Results

5.4.] Effect of heparin on hepatotoxicity after LPS/RAN treatment

Given alone, the doses of LPS or RAN are not hepatotoxic up to 24 h after
administration (Luyendyk et al., 2003b). Previous studies have shown that the
coagulation system is a critical mediator of LPS/RAN-induced liver injury (Luyendyk,
2004b). Confirming these results, LPS/RAN-cotreatment caused a significant increase in
serum ALT activity, and this was greatly attenuated by heparin at 6 h (Figure 5.1A), a
time associated with maximal injury in LPS/RAN-treated rats (Luyendyk et al., 2003b).
Complementing this result, heparin completely prevented the slight increase in ALT
activity observed at 3 h in this study (Figure 5.1A), a time near the onset of LPS/RAN-
induced liver injury (Luyendyk et al., 2004b; Luyendyk et al., 2003b). Heparin efficacy
was confirmed by immunohistochemical staining for hepatic fibrin deposition. Heparin

prevented or significantly attenuated LPS/RAN-induced hepatic fibrin deposition at 3 and

6 h, respectively (Figure 5.1B).

5.4.2 Effect of heparin on liver hypoxia after LPS/RAN treatment

Previous studies demonstrated that hepatic fibrin deposition is important for
LPS/RAN-induced liver injury, and that one consequence of fibrin deposition, i.e., liver
hypoxia, occurs at time near the onset of liver injury @uyendyk, 2004b). Accordingly,

the effect of the anticoagulant heparin on liver hypoxia was evaluated using PIM

169

Figure 5.1: Effect of hepa‘rin on hepatotoxicity afier LPS/RAN treatment, Rats were
treated with 44.4 x 10‘5 EU/kg LPS or its Veh (iv), then one h later with 3000 U/kg
heparin or its Veh (s.c.). One h later, 30 mg/kg RAN or its Veh was administered (iv).
For (A), hepatic parenchymal cell injury was estimated 3 or 6 h after RAN administration
by increases in serum ALT activity. For (B), livers were removed 3 or 6 h after RAN
treatment and processed for fibrin immunohistochemistry, and the area of positive fibrin
staining in 10 randomly chosen, IOOX fields per tissue was determined
morphometrically. Data are expressed as mean i SEM. n=6-9 in each group.
‘Significantly different from VehNeh/Veh-treated rats at that time. #Significantly
different fiom LPSNeh/RAN-treated rats at that time. 1‘Significantly different fiom the

same treatment at 3 h. (p<0.05).

170

>
'8!!!

ALT Activity (UIL)
§ § § §

 

 

 

 

171

 

 

— VehNOhNeh . a 3 ° ‘ — VehNehNeh
==== LPSNeh/RAN E «:22 LPSNehIRAN * a
— LPSIl-leparlanAN - — LPSIHeparlanAN l
5 0.0+
.g '2 a 1."
.n c
, E o 0.02-
. =-
a ‘4 g
._I.i:i.l_.lrULj_ .3 m. L
3 h e h 3 h a b
Time after RAN (hours) Time after RAN (hours)

immunohistochemical staining. Little PIM-adduct staining was noted in livers of
VehNeh-treated rats (Figure 5.2A). As described previously (Luyendyk, 2004b), PIM-
adduct staining was significantly increased in livers of LPS/RAN-treated rats at 3 h, and
this staining persisted to 6 h (Figure 5.2D). Heparin significantly attenuated the increase

in PIM-adduct staining at both 3 and 6 h (Figure 5.2D).

5.4.3 Serum CINC-l concentration and PMN accumulation in livers of LPS/RAN—
treated rats

Accumulation of PMNs in liver and serum CINC-l concentration were evaluated
at a time near the onset of injury in LPS/RAN-treated rats (i.e., 3 h). In this study, serum
ALT activity was not significantly increased in LPS/RAN-treated rats compared to
Veh/Veh- treated rats at this time (data not shown). ClNC-l, the rat homologue of human
interleukin 8, is important for PMN accumulation in liver and lungs after exposure to
LPS (erang et al., 1995;Yamasawa et al., 1999). Treatment with RAN alone caused a 10-
fold increase in serum ClNC-l concentration (Figure 5.3A). CINC-l was increased much
more dramatically (~900-fold) in rats treated with LPS alone, and was increased similarly
in LPS/RAN-cotreated rats. PMNs accumulated in livers of LPS-treated rats and to a
lesser degree in LPS/RAN-treated rats at 3 b (Figure 5.3B). Treatment with RAN alone
appeared to cause a small increase in PMN accumulation that did not reach statistical
significance compared to VehNeh-treated rats (Figure 5.3B). Histological evaluation
revealed that PMNs were distributed evenly across the liver lobule in LPSNeh-treated
rats. A similar distribution was observed in LPS/RAN-treated rats, although infrequent

midzonal and subserosal clusters of PMNs were noted in some rats.

172

A:

T

.\

vo~

Figure 5.2: Effect of hemrin on liver hypoxia in LPS/RAN-treated 12% Rats were
treated with 44.4 x 106 EU/kg LPS or its Veh (iv), then one h later with 3000 U/kg
heparin or its Veh (s.c.). One h later, 30 mg/kg RAN or its Veh was administered (iv).
Livers were removed 3 or 6 h after RAN treatment and processed for PIM-adduct
immunohistochemistry as described in Materials and Methods. (A) Representative
photomicrograph (l 00X) showing little PIM-adduct staining in a liver from a
VehNeh/Veh-treated rats. (B) Representative photomicrograph (l 00X) showing marked
panlobular PIM-adduct staining in a liver from a LPSNeh/RAN-treated rat. (C)
Representative photomicrograph (l 00X) showing reduced staining intensity compared to
LPSNeh/RAN as a result of heparin administration. The area of positive PIM-adduct
staining in 10 randomly chosen, 100X fields per tissue was determined morphometrically
as described in Materials and Methods. Data are expressed as mean i SEM. n=6-9 rats
per group. *Significantly different from VehNeh/Veh-treated rats at that time.

“Significantly different fi'om LPSNeh/RAN-treated rats.

173

 

 

 

 

 

 

 

174

— VohNohNeI't
13‘ 0.30 1:: LPSNethAN

    
   

g 025 — LPsn-leparlnRAN
,2 - 0.20
g
gl- 0.15
>. c
:r: _9 010
g 0.05
n- 0.00 . '
3h 6h
Time after RAN (hours)

 

‘6

 

Figure 5.3: Serum ClNC-l concentration and PMN accumulation in livers of L_PS/RAN-
treated rats. Rats were treated with 44.4 x 106 EU/kg LPS or its Veh (iv), then two h
later with 30 mg/kg RAN or its Veh (iv). The serum concentration of ClNC-l (A) and
hepatic PMN accumulation (B) were evaluated 3 h afier drug treatment. Data are
expressed as mean :1: SEM. n=3-4 rats in each group. ’Significantly different from
respective group not given LPS ”Significantly different from respective group not given

RAN (p<0.05).

175

A

 

LPS

LPS Veh

Veh

 

5 o 5
2 2 1

o 5 o
1

 

 

 

 

 

u_n=._ X23322."—
.. ._l n
I—
t 1* m
m
L

 

 

 

m a.
:83... s

o 1 1.
1 0

02.0 anm

RAN

Veh

RAN

Veh

176

5.4.4 Effect of heparin on serum CINC—l concentration and PMN accumulation in
livers of LPS/RAN-treated rats

LPSNeh/RAN-treatment caused a significant increase in serum CINC-l at 3 h
that persisted at 6 h (Figure 5.4A), and this increase was unaffected by heparin
cotreatment at either time. Consistent with the change in serum CINC-l, PMNs
accumulated in livers of LPS/RAN-treated rats at 3 h and their presence persisted at 6 h
(Figure 5.4B). PMNs were distributed across the hepatic lobule in LPS/RAN-treated rats,
but more rrridzonal and subserosal clusters were observed in lesioned areas at 6 h. PMN
accumulation in LPS/RAN-treated rats was unaffected by heparin cotreatment at both
times. However, the reduction in hepatocellular injury by heparin was associated with a
more panlobular distribution of PMNs as opposed to clustering of PMNs in lesioned

areas.

5.4.5 Effect of hypoxia on PMN-E induced killing of rat hepatocytes

Preliminary experiments were conducted to identify a reduced 02 exposure not
associated with significant cytotoxicity, as estimated by release of ALT into culture
medium 8 h after exposure. Incubation of hepatocytes for 8 h in 2% 02 but not in 5% 02
caused significant ALT release relative to exposure to an oxygen replete atmosphere (i.e.,
20% 02) (data not shown). Accordingly, 5% 02 was chosen as an hypoxic chamber
concentration that did not by itself cause cytotoxicity. At 8 h, 8.8 U/ml elastase caused
significant ALT release from hepatocytes cultured in 20% 02, consistent with the
timecourse of PMN-CM-mediated cytotoxicity (Ganey et al., 1994). However, exposure

to 5% 02 resulted in a leftward shift in PMN-E-induced cytotoxicity at this time

177

Figure 5.4: Effect of heparin on serum CINC-l concentration and hepatic PMN
agcumalaktion in LPS/RAN-treated rag Rats were treated with 44.4 x 106 EU/kg LPS or
its Veh (iv), then one h later with 3000 U/kg heparin or its Veh (s.c.). One h later, 30
mg/kg RAN or its Veh was administered (iv). Serum ClNC-l concentration (A) and
hepatic PMN accumulation (B) were evaluated 3 or 6 h after drug administration. Data
are expressed as mean :1: SEM. n=4-9 rats in each group. *Significantly different from
Veh/Veh/Veh-treated rats at that time. ”Significantly different from LPSNeh/RAN-

treeted rats. “Significantly different from the same treatment at 3 h. (p<0.05).

178

A

d
§ §

ClNC-1 (nglml)

9
_s

 

— VehNehNoh
:23 LPSNethAN
— LPSIl-leparlanAN

 

 

 

 

 

 

W * *8 ea
FT
3 h 0 h
Time after RAN (hours)

179

PMN8I400X HPF

.r. .s N
0 0| O
A l I

 

— VohNehNeh
m LPSNehIRAN
— LPSll-leparinIRAN

*#

    

 

3 6 b
Time after RAN (hours)

Figure 5.5: Effect of hypoxia on PMN-E-induced MM gt hepatocytes. Rat
hepatocytes were cultured at a cell density of 2.5 X 105 cells/ml in Williams’ Medium B
containing 5% CCS. Two h later, the medium was changed to serum-free medium
containing human PMN elastase (PMN-E) at a concentration of 0, 0.7, 1.3, 2.2, 3.3, 4.4,
6.6, or 8.8 U/ml, and incubated in either 5% or 20°/o Oz. Cytotoxicity was evaluated as
ALT released into culture medium 2 h (A) or 8 h (B) later. Data are expressed as mean 2i:
SEM. n=3 separate hepatocyte isolations. ‘Significantly different from respective
treatment with 0 U/ml PMN-E. " Significantly different from the respective treatment

incubated in 20% 02 (p<0.05).

180

% ALT Release

 

 

 

 

 

 

o 00

8 50‘

% «a

E ,,.

2‘ 20

32 10
. . - - . o . , . a a ,
2 4 0 a 10 o 2 4 0 a 10

Elastase (UnitslmL) Elastase (UnitsImL)

These data were generated by Patrick Shaw

181

(Figure 5.5B). Contrasting its ability to kill hepatocytes at 8 h, PMN-E did not cause
significant ALT release at any concentration used when hepatocytes were cultured in
20% 02 for 2 b (Figure 5.5A). However, PMN-E at 6.6 and 8.8 U/ml caused significant

ALT release from hepatocytes cultured in 5% 02 at this time (Figure 5.5A).

5.5 Discussion

Previous experiments demonstrated that LPS/RAN-cotreatment caused
hepatotoxicity in rats beginning near 3 h (Luyendyk et al., 2003b). Liver injury in this
model is dependent upon an activated coagulation system and the formation of fibrin
clots (Luyendyk, 2004b). In addition, liver hypoxia, one possible consequence of hepatic
fibrin deposition, occurred in livers of LPS/RAN-cotreated rats. The studies presented
here tested the hypothesis that an activated coagulation system causes liver hypoxia in
LPS/RAN-treated rats. Heparin administration prevented the slight increase in serum
ALT activity and hepatic fibrin deposition observed at 3 h, and confirming previous
results (Luyendyk, 2004b), it also attenuated this increase at 6 h (Figure 5.1A). Heparin
administration significantly decreased liver PIM-adduct staining by nearly 70% (Figure
5.2), suggesting that coagulation system activation is responsible for the majority of liver
hypoxia in LPS/RAN-treated rats. Overall, these results support the hypothesis that
hepatic fibrin deposition causes liver hypoxia and hepatotoxicity in LPS/RAN-treated
rats. However, whether or not the protective effect of heparin in this model is mediated
entirely through its ability to reduce liver hypoxia is not known.

If severe enough, hypoxia alone is sufficient to cause hepatocellular injury in rats

(Fassoulaki et al., 1984) and in isolated perfused livers (Lemasters et al., 1981), and it is

182

capable of inciting cell death in hepatocytes (Khan and O’Brien, 1997). Furthermore,
exposure to hypoxia can render the liver and/or isolated hepatocytes sensitive to
secondary insult fi'om hepatotoxicants and some drugs (Bacon et al., 1996; Khan and
O'Brien, 1995; Silva et al., 1992; McGirr et al., 1990; Khan and O’Brien, 1997; Shen et
al., 1982) Overall, these results suggest that otherwise noninjurious degrees of hypoxia
can lower the threshold for hepatotoxicity. Lesions in livers of rats coexposed to hypoxia
and LPS were LPS-like (Shibayama, 1987), suggesting that hypoxia increases either the
degree of LPS-induced inflammation or the sensitivity of the liver to the cytotoxic effects
of inflammatory mediators. Indeed, interplay between hypoxia and some inflammatory
factors, including macrophages, COX-2, and PMNs, has been reported (Hannah et al.,
1995; Zhong et al., 2004; Lahat et al., 2003; Tamura et al., 2002).

PMN accumulation was noted in lesions within livers taken fi'om LPS/RAN-
treated rats (Luyendyk et al., 2003b), but their role in this model remains unclear.
Treatment with LPS alone caused quantitatively similar increases in serum ClNC-l and
hepatic accumulation of PMNs as compared to LPS/RAN at a time near the onset of liver
injury in LPS/RAN-treated rats (Figure 5.3). Inasmuch as liver injury does not occur in
rats treated with this dose of LPS given alone, this suggests that PMN accumulation is not
sufficient to cause hepatotoxicity. A similar observation has been made in livers of rats
cotreated with aflatoxin B1 (AF B.)/LPS, where PMN accumulation is nearly identical in
rats treated with LPS alone; however, PMNs are required for hepatotoxicity in
AF Bt/LPS-cotreated rats. Indeed, PMN extravasation rather than PMN accumulation is
likely a required event for PMN-mediated hepatocellular damage in models of

endotoxemia (Chosay et al., 1997; Jaeschke and Smith, 1997).

183

Accumulation of PMNs in hepatic sinusoids is independent of adhesion molecules
(i.e., VCAM-l, [CAM-1, CD11a/b/CD18) in models of endotoxemia; however, they are
required for extravasation of PMNs in the liver and hepatocellular injury (Jaeschke et al.,
1996; Jaeschke et al., 1991; Essani et al., 1997; Essani et al., 1995). In addition,
chemokine production by hepatocytes can trigger PMN extravastation and tissue injury
(Maher et al., 1997; Li et al., 2004). Overall, in light of these results, the possibility that
PMNs are important for injury in the LPS/RAN model should not be diminished by the
observation that PMN accumulation occurs to a similar degree in both LPS-treated
groups. Experiments are underway to evaluate activation of PMNs and the relationship of
PMN activation to hepatocellular injury in LPS/RAN-treated rats.

PMN accumulation or activation might occur as a consequence of thrombin
generation in LPS/RAN-treated rats. Thrombin and PNst are sufficient extrahepatic
factors to cause hepatocellular injury in isolated, perfused livers from rats treated with
large doses of LPS (Moulin et al., 2001), and inhibition of thrombin attenuates PMN
activation in rats treated with a large dose of LPS (Copple et al., 2003). Heparin
administration did not significantly affect serum ClNC-l or accumulation of PMNs in
livers of LPS/RAN-treated rats at any timepoint examined (Figure 5.4). This result
suggests that PMN accumulation and increased serum CINC-l after LPS/RAN-
cotreatment occur by a thrombin-independent mechanism. Experiments are underway to
test the hypothesis that thrombin activation precedes and is required for PMN activation
in LPS/RAN-treated rats. Thrombin can indirectly cause activation of PMNs through a
mechanism requiring activation of its receptor, protease activated receptor-1, which is

expressed by KCs, SECS, and hepatic stellate cells, but not on PMNs (Copple et al., 2003;

184

Marra et al., 1998). Another possibility is that thrombin-mediated liver hypoxia modifies
PMN accumulation or activation. However, the decrease in liver hypoxia accompanying
heparin administration was not associated with a decrease in PMN accumulation in
LPS/RAN-treated rats. Nevertheless, hypoxia might promote PMN activation directly, or
indirectly through other mechanisms, including altered expression of adhesion molecules
on SECS (T amura et al., 2002; Amould et al., 1995; Amould et al., 1994).

PMNs release several cytotoxic factors including the proteases, cathepsin G and
elastase, which are sufficient to kill rat hepatocytes in vitro (Ho et al., 1996). In addition
to possibly enhancing activation of PMNs and release of these proteases, hypoxia might
also increase the sensitivity of hepatic parenchymal cells to their cytotoxic effects. Few
studies have examined this relationship. In one cell-based model of
hypoxia/reoxygenation, cardiac myocytes were rendered more sensitive to elastase
cytototoxicity during the reoxygenation phase (Buerke et al., 1994). However, the
relative role of hypoxia and reoxygenation phases in increasing sensitivity to elastase
cannot be separated. We show here that hepatocytes cultured under hypoxic (5% 02)
conditions are rendered sensitive to killing by normally nontoxic concentrations of PMN-
E (Figure 5.5). Interestingly, the timecourse over which hepatocyte killing occurs
resembles the timecourse of injury (i.e., beginning near 2-3 h) in the LPS/RAN model
(Luyendyk et al., 2003b), whereas killing of hepatocytes by PMN-E in an oxygen-replete
atmosphere (20% 02) requires 8-16 h to reach statistical significance (Ganey et al., 1994).
This result suggests that hypoxia can increase the sensitivity to elastase cytotoxicity, as

well as accelerate signaling required for PMN-E-induced cell death.

185

In summary, rats cotreated with LPS/RAN develop liver hypoxia with a
timecourse similar to that of hepatocellular injury (Luyendyk et al., 2003b).
Anticoagulation with heparin significantly attenuated hepatic fibrin deposition, hypoxia,
and hepatocellular injury, suggesting a connection between these 3 events. PMN
accumulation and the serum concentration of ClNC-l in LPS/RAN-treated rats were not
different from rats treated with LPS alone, and neither of these was affected by heparin
administration. Overall, the results are consistent with reduction of LPS/RAN-induced
liver injury by heparin through a mechanism related to prevention of fibrin deposition
and liver hypoxia, but independent of PMN accumulation. Inasmuch as hypoxia rendered
hepatocytes sensitive to killing by PMN-E, PMN activation might be important for injury
in livers of LPS/RAN-treated rats. Further studies are required to evaluate the modulating
effects of anticoagulation and hypoxia on PMN activation and their connection to injury

in this model.

186

CHAPTER 6

Summary and Conclusions

187

6.1 Summary of conducted research

Initially, the hypothesis was tested that modest underlying inflammation renders
ranitidine hepatotoxic in rats. Dose-response experiments were conducted to identify
doses of LPS and RAN that when given alone to rats did not cause hepatotoxicity. In the
case of RAN, a dose that would cause hepatotoxicity without significant mortality was
not identified. Preliminary dose-ranging studies were performed to identify doses of LPS
and RAN that when coadmirristered produced maximal hepatotoxicity by 24 h (Figure
2.1). Rats treated with a nonhepatotoxic dose of LPS 2 h before a noninjurious dose of
RAN developed hepatic parenchymal cell injury (i.e., increased serum ALT and AST
activities, Figure 2.2) by 6 h after drug treatment. The primary histopathologic finding in
livers of LPS/RAN-treated rats was midzonal oncotic necrosis with neutrophilic infiltrate
(Figure 2.3). This lesion increased in frequency and severity between 3 and 6 h in
LPS/RAN-treated rats (Table 2.1). As noted earlier, some similarities can be identified
between injury in the LPS/RAN model and case reports of RAN hepatotoxicity in people
(Table 1.1). For example, increased serum activities of markers of hepatic parenchymal
cell injury occured more frequently and were often more dramatic than changes in
biomarkers of cholestatic injury. Moreover, histopathologic changes in livers of
LPS/RAN-treated rats resemble some changes observed in severe cases of RAN
hepatotoxicity in people, including hepatocellular necrosis with inflammatory infiltrate.
Overall, these results suggest that RAN is rendered hepatotoxic in rats undergoing a
modest inflammatory response.

In a subsequent study, LPS/RAN treatment caused significant hepatotoxicity,

whereas cotreatment of rats with the same dose of LPS and FAM at either an equi-

188

efficacious (EE) or equimolar dose as RAN did not result in significant injury (Figure
2.4). Accordingly, the propensity of each H2-antagonist to cause idiosyncratic
hepatotoxicity was predicted by the potential of each to be rendered hepatotoxic by
inflammation. Furthermore, this result suggests that H2-antagonism is not sufficient to
cause hepatotoxicity in this model. Similar results were observed in vitro when rat
hepatocytes were exposed to PMN-conditioned medium. RAN treatment did not cause
significant cytotoxicity at any concentration tested, but it did render hepatocytes sensitive
to the cytotoxic effects of PMN-CM (Figure 2.5). Matching the in vivo result, F AM
treatment did not cause this effect. This suggests that, at least in part, RAN might alter
hepatocellular sensitivity to cytotoxic factors released by activated PMNs. However, the
millimolar concentrations required to cause this effect are likely not reached in RAN-
treated rats, so that the relevance of this observation to the mechanism of LPS/RAN-
hepatotoxicity is therefore unclear.

Inasmuch as changes in gene expression are important for manifestation of
inflammatory responses, hepatic gene expression was evaluated by Affymetrix
microarray analysis in rats treated with LPS and/or RAN at a time just before the
development of significant hepatic parenchymal cell injury (i.e., 3 h, Figure 3.1). At this
time, hierarchical clustering of active genes was sufficient to segregate rats to their
respective treatments (Figure 3.2). Several probesets were changed only after treatment
with LPS, RAN, or LPS/RAN, and others were changed after more than one treatment
(Figure 3.3). Within the population of active genes, 4 patterns of potential mechanistic
interest were identified. Genes changing similarly after treatment with LPS alone or RAN

alone and in LPS/RAN-cotreated rats (LPS 5 LPS/RAN, RAN s LPS/RAN) were

189

involved in inflammation and/or associated with cell death signaling (Table 3.1).
However, treatment with RAN or LPS alone did not cause hepatotoxicity, suggesting that
the products of genes with this expression pattern are likely not sufficient to cause injury.
They might, however, interact with other changes to cause toxicity.

RAN cotreatment attenuated the LPS-induced change in several genes (LPS/RAN
< LPS, Table 3.1). This group of genes might point to an ability of RAN to prevent
upregulation of cytoprotective or anti-inflammatory genes. Other genes were increased to
a greater degree in LPS/RAN-treated rats as compared to any other treatment group
(LPS/RAN > LPS and > RAN, Table 3.1). Gene products in this group were related to
inflammation and many were hypoxia-inducible, including PAI-l, Egr-l, Btg-2, and
others (Table 3.2). The change in expression for three of these (PAI-l, Egr-l, Btg-2) was
confirmed by real-time PCR (Figure 3.4). Inasmuch as injury only occurs in the
LPS/RAN group, augmented expression of one or more of these gene products might be
involved in the injury. Further investigation and data-mining may reveal interactions and
relationships between these genes as they relate to LPS/RAN-induced liver injury.

Alterations in hepatic gene expression elucidated by nricroarray analysis
suggested altered liver homeostasis (i.e., hypoxia) and involvement of the hemostatic
system (i.e., increased PAl-l mRNA) in LPS/RAN-induced liver injury. Indeed, the
change in PAI-l mRNA in liver was reflected by a change in serum PAI-l concentration
at 3 h (Figure 3.5). Further examination revealed a persistently augmented serum PAl-l
concentration in LPS/RAN-treated rats, and a less robust increase in rats treated with LPS
that faded by 6 h (Figure 4.5). Consistent with the change in serum PAI-l concentration,

hepatic fibrin deposition was observed in livers of LPS/RAN-treated rats both before

190

(Figure 3.7 and 4.3) and after (Figure 4.4) the development of liver injury. SK treatment
significantly reduced hepatic fibrin deposition and parenchymal cell injury in LPS/RAN-
treated rats (Figure 4.6). The results suggest that hepatic fibrin deposition is a mediator of
liver injury in LPS/RAN-treated rats and are consistent with impaired fibrinolysis as a
consequence of augmented PAI-l expression.

Hepatic fibrin deposition might result from activation of the coagulation system,
impaired fibrinolysis, or both. Serum HA was increased in LPS/RAN-treated rats in the
absence of a change in RECA-l staining in liver, suggesting perturbation in the function
of SECS in the absence of overt cell destruction (Figure 3.6 and Table 4.1). In addition to
causing increased PAI-l production, altered SEC firnction can result in activation of the
coagulation system. Indeed, LPS/RAN-treatment caused coagulation system activation
prior to hepatocellular injury. Inhibition of coagulation with heparin significantly reduced
LPS/RAN-induced fibrin deposition and hepatic parenchymal cell injury (Figure 4.7 and
5.1). Overall, the results suggest that RAN augments LPS-mediated activation of the
hemostatic system (Figure 4.20). In light of the protection provided by SK in this model,
the results suggest that the protection provided by heparin is likely mediated by its
prevention of hepatic fibrin deposition in LPS/RAN-treated rats.

Possible consequences of hepatic fibrin deposition include disruption of liver
blood flow and hypoxia. Indeed, LPS/RAN-treatment caused liver hypoxia at a time near
the onset of liver injury (Figure 4.8 and 4.8), a response that was attenuated by heparin
administration. Although hypoxia alone might be sufficient to cause injury in LPS/RAN-
treated rats, it is likely that it influences the action of other inflammatory cells/mediators.

Since PMNS accumulate in lesions after LPS/RAN-treated rats, preliminary studies were

191

conducted to evaluate a possible effect of hypoxia on their contribution to liver injury.
Interestingly, serum CINC-l and hepatic PMN accumulation were similar in LPSN eh
and LPS/RAN—treated rats, suggesting that if PMNS are involved in the injury, their
accumulation in liver is not sufficient and that secondary activation Signals are involved.
Heparin affected neither ClNC-l nor PMN accumulation in LPS/RAN-treated rats. This
result suggests that activation of the coagulation system is not required for PMN
accumulation and that a reduction in liver hypoxia does not influence PMN
accumulation. Nevertheless, hypoxia rendered hepatocytes sensitive to killing by PMN-E,
suggesting that if PMN activation occurs in LPS/RAN-treated rats, the liver might be a
sensitive target. Furthermore, the timecourse of hypoxia-enhanced HPC killing by PMN-

E was also consistent with the development of liver injury in LPS/RAN-treated rats.

6.2 Proposed mechanism

Figure 6.1 illustrates a proposed mechanism for LPS/RAN-induced liver injury.
This mechanism is based on the results presented in this dissertation but will most
certainly evolve as we learn more about this model. RAN can augment the effects of LPS
on the hemostatic system, including activation of thrombin (Figure 4.2) and PAI-l
expression (Figure 4.4). Inasmuch as fibrin clots are required for injury in this model

(Figure 4.6), impaired fibrinolysis by PAI-l likely contributes to injury. One contribution

192

Figure 6.1: Proposed mechanism of hepatotoxicity cased by LPS/RAN-comtrnerlga

Lag See section 6. 2 for details.

193

 

 

 

 

 

Pmcoagulant S
Coagulation inusoidal
C9 Altered SEC“ System —'—’ Fibrin

Ho stasis Activation Deposition

 

 

 

 

 

    
   

1.

PMN accumulation Protease

194

of fibrin clots to LPS/RAN-induced liver injury might be mediated through ischemia and
hypoxia. Indeed, LPS/RAN—treatment caused liver hypoxia (Figure 4.8 and 4.9), and the
protective effect of heparin on HPC injury was associated with a reduction in liver
hypoxia (Figure 5.1 and 5.2). Hypoxia might cause hepatocellular injury directly or
indirectly by affecting other inflammatory factors or hepatocellular homeostasis. One
possibility is that activation of the coagulation system and dramatic liver hypoxia
contribute to hepatocellular injury in LPS/RAN-treated rats by causing PMN activation
and release of toxic proteases (e.g., elastase). Furthermore, hypoxia might increase
sensitivity of hepatic parenchymal cells to killing by PMN-derived proteases (Figure 5.5).
Not illustrated in Figure 6.1, but of potential importance are direct effects of RAN

(Figure 2.5) on hepatocellular sensitivity to these mediators.

6.3 Significance of research, knowledge gaps, and future studies

Several important findings were identified by these studies. First, the observation
that idiosyncrasy-like hepatotoxicity develops in rats cotreated with LPS and RAN, but
not FAM, suggests an ability of this model to predict the propensity of these two H2-
antagonists to cause idiosyncratic liver injury. Additional studies evaluating other drugs
associated with idiosyncratic hepatotoxicity within this pharmacologic class (e.g.,
cimetidine and nizatidine) and others are needed to validate further its use as preclinical

predictor of some idiosyncratic drug reactions. Given the potential impact of losing

195

a drug during development, the inclusion of “negative comparators” in validation of the
LPS/drug-cotreatment model would assist in identification of the apparent “false-
positive” rate. Overall, these studies add another layer of promise to a growing body of
evidence suggesting the predictive potential of this model (Buchweitz et al., 2002).

In addition to its use as a “negative comparator” in these studies, treatment of rats
with F AM and LPS provided useful insight into the mechanism of injury in the LPS/RAN
model. LPS-treated rats given a pharmacologically equiefficacious dose of FAM (relative
to RAN) did not develop liver injury, suggesting that the pharmacologic effect of the
drug is not sufficient to cause toxicity. Consistent with this observation is the differing
propensity of marketed H2-antagonists to cause liver injury (see Chapter 1).
Nevertheless, H2-receptor blockade might be a required factor for LPS/RAN-induced
liver injury. For example, histamine attenuates LPS-induced TNF synthesis by an H2-
receptor-dependent mechanism in monocytes (V annier et al., 1991). Furthermore,
histamine affords protection against P. acnes/LPS- and ethanol-induced liver injury by a
H2-receptor dependent mechanism (Homyak et al., 2003; Yokoyama et al., 2004).
Accordingly, a hypothesis to be tested is that H2-antagonism is a permissive event for
liver injury in LPS/RAN-treated rats, but alone it is insufficient to cause injury.
Administration of histamine dihydrochloride would be expected to compete with RAN
for H2-receptor binding and to attenuate LPS/RAN-induced liver injury. Although the
role of endogenous histamine antagonism in LPS/RAN-induced liver injury remains
unknown, it is clear that H2-independent properties of these drugs determine their ability

to interact with LPS to cause hepatotoxicity.

196

Evaluation of liver gene expression using microarray technology provided a
wealth of information about effects of RAN alone and in combination with LPS.
Selection of specific expression patterns revealed useful information regarding the
function/association of gene products changing in each treatment group. However,
important changes in expression likely remain undiscovered within the dataset presented
herein. Future data mining efforts should build on statistical and filtering strategies
outlined here to include the use of updated rat gene nomenclature and ontologies, as well
as tools designed to investigate interactions among gene products (e. g., in cell death
signaling). In addition, analytical strategies designed for identification of synergistic
changes in gene expression could be applied to this experimental design. AS noted in
Chapter 3, the addition of timecourse studies evaluating gene expression would most
certainly be useful in understanding the relationship of gene expression to liver toxicity in
LPS/RAN-treated rats. One advantage of the time selected for the studies presented here
is that it provided a snapshot of gene expression before the onset of liver injury, thereby
allowing for potential mechanistic association of a change in expression with the injury.
However, as interesting hypotheses emerge, it is important to recognize that a change in
gene expression is not always causal to injury, making post-hoe mechanistic analysis an
absolute requirement for elucidating mechanisms.

In the case of PAI-l, real-time PCR and ELISA were used to confirm augmented
mRN A and protein in LPS/RAN-treated rats. However, it is not known whether PAI-l is
a mediator of liver injury in this model. Studies to test this hypothesis using an inhibitor
of PAH activity are in the planning stage. However, the potential for impaired

fibrinolysis as a result of increased PAI-l expression prompted us to test, and ultimately

197

prove true, the hypothesis that fibrin clots are important for liver injury in this model.
Analysis of liver gene expression in LPS/RAN-treated rats revealed a large set of genes
that were expressed under hypoxic conditions (Table 3.1 and 3.2). Follow-up
investigation of this observation using two other markers (PIM-adducts and HIF-lalpha)
demonstrated that LPS/RAN-treatment does cause liver hypoxia. By and large, analysis
of gene expression in LPS/RAN-treated yielded several hypotheses that proved useful in
identification of potential mediators of liver injury.

A unique challenge presented by this model is the abrupt and somewhat variable
time to onset and severity of liver injury in LPS/RAN-treated rats. For example, whereas
some LPS/RAN-treated rats never develop liver injury, the maximal increase in ALT
observed in LPS/RAN-treated rats is variable and can reach values as high as 1000 U/L
(unpublished observation, JPL). In addition, the time to onset of hepatotoxicity is
variable, occurring sometimes as early as 2 h. Therefore, it is somewhat challenging to
use the temporal relationship between exposure and injury onset to support a causal
association between a particular mediator and liver injury. For example, the experiments
in Chapter 4 showed that liver hypoxia in LPS/RAN-treated rats accompanied
hepatocellular injury but they did not distinguish the temporal relationship between these
two outcomes (i.e., Figure 2.2 and 3.1). Accordingly, identification of liver hypoxia in
LPS/RAN-treated rats before the onset of injury Should be a focus of future experiments.
The hypothesis that liver hypoxia causes LPS/RAN—induced liver injury has not yet been
tested. However, the result that heparin decreases both liver hypoxia and injury is
consistent with a protective effect mediated through reduction of hypoxia, especially

considering the importance of fibrin clots in this model. Further experiments are required

198

to identify hypoxia as a cause rather than a consequence of liver injury in this model. For
example, one hypothesis to test is that exposure to hypoxia renders rats susceptible to
RAN hepatotoxicity (i.e., hypoxia can replace LPS). Furthermore, evaluation of gene
expression data for hypoxia-inducible, cell death-related factors in LPS/RAN-treated rats
might provide further insight.

AS eluded to in Chapter 5, hypoxia might increase PMN activation or increase the
sensitivity to killing by PMN-derived products in livers of LPS/RAN-treated rats. Indeed,
several clues suggesting a role for PMNS in LPS/RAN hepatotoxicity have been
identified. PMN accumulation occurs in livers of LPS/RAN-treated rats (Figure 2.3 and
5.3), RAN renders hepatocytes sensitive to PMN-CM (Figure 2.5), and hypoxia renders
hepatocytes sensitive to PMN-E (Figure 5.5). An important hypothesis to test before
proceeding with further investigation of this hypothesis is that PMNS are critical
mediators in LPS/RAN-induced liver injury. If a hypoxia/PMN interaction is required for
LPS/RAN-induced liver injury, prevention of either PMN accumulation or hypoxia
Should result in a reduction in liver injury. Current efforts are focused on development of
assays for assessment of PMN activation in LPS/RAN-treated rats. In addition, plans for
studies evaluating the role of PMNS (antibody-mediated PMN depletion) and PMN
elastase (elastase inhibitor GW 311616) in LPS/RAN-induced liver injury are currently
underway.

The results presented in this dissertation demonstrate that a modest inflammatory
response renders RAN hepatotoxic in rats. The results provide a rationale for further
evaluation of the hypothesis that inflammation is involved in some idiosyncratic drug

reactions. Further validation of the LPS/drug cotreatment model may lead to a predictive

199

preclinical tool. Studies comparing RAN and FAM support this, but clearly testing more
drugs and their “negative comparators” is required for complete validation. In addition,
the results of these studies point to the hemostatic system as a critical mediator of
LPS/RAN-induced liver injury and contribute to our understanding of the consequences
of drug/inflammation interaction associated with hemostatic system activation in
hepatotoxicity (i.e, fibrin deposition, hypoxia). Finally, results of in vitro studies suggest
that hypoxia can increase sensitivity to cytotoxic effects of PMN-E. This relationship has
implications for our understanding of mechanisms in other models of PMN-dependent

hepatotoxicity.

200

Bibliography

Abrahamsson T, Nerme V, Stromqvist M, Akerblom B, Legnehed A, Pettersson K and
Westin EA (1996) Anti-thrombotic effect of a PAI-l inhibitor in rats given endotoxin.

Thromb Haemost 75:] 18-126.

Aithal GP, Day CP, Leathart JB and Daly AK (2000) Relationship of polymorphism in
CYP2C9 to genetic susceptibility to diclofenac-induced hepatitis. Pharmacogenetics

10:51 1-518.

Aithal GP, Ramsay L, Daly AK, Sonchit N, Leathart JB, Alexander G, Kenna JG,
Caldwell J and Day CP (2004) Hepatic adducts, circulating antibodies, and cytokine
polymorphisms in patients with diclofenac hepatotoxicity. Hepatology 39: 1430-1440.

Akahane K, Okamoto K, Kikuchi M, Todoroki H, Higure A, Ohuchida T, Kitahara K,
Takeda S, Itoh H and Ohsato K (2001) Inhibition of factor Xa suppresses the expression
of tissue factor in human monocytes and lipopolysaccharide-induced endotoxemia in rats.

Surgery 130:809-818.

Akita S, Kawahara M, Takeshita T, Morio M and Fujii K (1989) Halothane—induced
hepatic microsomal lipid peroxidation in guinea pigs and rats. J Appl Toxicol 9:9-14.

Akita S, Morio M, Kawahara M, Takeshita T, Fujii K and Yamarnoto M (1988)
Halothane—induced liver injury as a consequence of enhanced microsomal lipid
peroxidation in guinea pigs. Res Commun Chem Pathol Pharmacol 61:227-243.

Alden C, Lin J, and Smith P (2003) Predictive Toxicology Technology for Avoiding
Idiosyncratic Liver Injury. Preclinica May/June: 27-35.

Ament PW, Roth JD and Fox CJ (1994) F amotidine-induced mixed hepatocellular
Jaundice. Ann Pharmacother 28:40-42.

AIIfouzadeh HR and Pohl LR (1995) Processing of endoplasmic reticulum luminal
antigens associated with halothane hepatitis in rat hepatocytes. Hepatology 22:936-943.

AI'llould T, Michiels C, Janssens D, Delaive E and Remacle J (1995) Hypoxia induces
PMN adherence to umbilical vein endothelium. Cardiovasc Res 30: 1009-1016.

Amould T, Michiels C and Remacle J (1994) Hypoxic human umbilical vein endothelial
cells induce activation of adherent polymorphonuclear leukocytes. Blood 83:3705—3716.

Arteel GE, Thurman RG and Raleigh JA (1998) Reductive metabolism of the hypoxia
tnal‘ker pimonidazole is regulated by oxygen tension independent of the pyridine
nucleotide redox state. Eur J Biochem 253:743-750.

201

Arteel GE, Thurman RG, Yates JM and Raleigh JA (1995) Evidence that hypoxia
markers detect oxygen gradients in liver: pimonidazole and retrograde perfusion of rat
liver. Br J Cancer 72:889-895.

Ashton T, Young IS, Davison GW, Rowlands CC, McEneny J, Van BC, Jones E, Peters
JR and Jackson SK (2003) Exercise-induced endotoxemia: the effect of ascorbic acid
supplementation. Free Radic Biol Med 35:284-291.

Awad JA, Horn J L, Roberts LJ and Franks JJ (1996) Demonstration of halothane-induced
hepatic lipid peroxidation in rats by quantification of F2-isoprostanes. Anesthesiology
84:910-916.

Bacon JA, Cramer CT, Petrella DK, Sun EL and Ulrich RG (1996) Potentiation of
hypoxic injury in cultured rabbit hepatocytes by the quinoxalinone anxiolytic,
panadiplon. Toxicology 108:9-16.

Bagheri H, Michel F, Lapeyre-Mestre M, Lagier E, Cambus JP, Valdiguie P and
Montastruc JL (2000) Detection and incidence of drug-induced liver injuries in hospital:
a prospective analysis from laboratory signals. Br J Clin Pharmacol 50:479-484.

Bahrami S, Red] H, Leichtfried G, Yu Y and Schlag G (1994) Similar cytokine but
different coagulation responses to lipopolysaccharide injection in D-galactosarnine-
sensitized versus nonsensitized rats. Infect Immun 62:99-105.

Banks AT, Zimmerman HJ, Ishak KG and Harter JG (1995) Diclofenac-associated
hepatotoxicity: analysis of 180 cases reported to the Food and Drug Administration as
adverse reactions. Hepatology 22:820-827.

Barcia C, Fernandez BA, Poza M and Herrero MT (2003) Parkinson's disease and
inflammatory changes. Neurotox Res 5:411-418.

Barr GD and Piper DW (1981) Possible ranitidine hepatitis. Med J Aust 2:421.

Barton CC, Barton EX, Ganey PE, Kunkel SL and Roth RA (2001) Bacterial
lipopolysaccharide enhances aflatoxin B1 hepatotoxicity in rats by a mechanism that
depends on tumor necrosis factor alpha. Hepatology 33:66-73.

Barton CC, Ganey PE and Roth RA (2000a) Lipopolysaccharide augments aflatoxin
B(1)-induced liver injury through neutrophil-dependent and -independent mechanisms.
T oxicol Sci 58:208-2 l 5 .

Barton CC, Hill DA, Yee SB, Barton EX, Ganey PE and Roth RA (2000b) Bacterial
lipopolysaccharide exposure augments aflatoxin B(1)-induced liver injury. T oxicol Sci
55:444-452.

Benjamini Y and Hochberg Y (1995) Controlling the false discovery rate: a practical and
powerfirl approach to multiple testing. J R Statist Soc 57 : 289-300.

202

Beutler B (2002) TLR4 as the mammalian endotoxin sensor. Curr Top Microbiol
Immunol 270:109-120.

Binder BR, Christ G, Gruber F, Grubic N, Hufnagl P, Krebs M, Mihaly J and Prager GW
(2002) Plasminogen activator inhibitor 1: physiological and pathophysiological roles.
News Physiol Sci 17:56-61.

Bini A, Itoh Y, Kudryk BJ and Nagase H (1996) Degradation of cross-linked fibrin by
matrix metalloproteinase 3 (stromelysin 1): hydrolysis of the gamma Gly 404—Ala 405
peptide bond. Biochemistry 35: 1 3056-1 3063.

Bird GL and Williams R (1989) Detection of antibodies to a halothane metabolite hapten
in sera from patients with halothane-associated hepatitis. J Hepatol 9:366-3 73.

Black M, Scott WE, Jr. and Kanter R (1984) Possible ranitidine hepatotoxicity. Ann
Intern Med 101:208-210.

Boelsterli UA Disease-related determinants of susceptibility to drug-induced
idiosyncratic hepatotoxicity. (2003a) Curr Opin Drug Discov Devel 6:81-91.

Boelsterli UA (2003b) Diclofenac-induced liver injury: a paradigm of idiosyncratic drug
toxicity. T oxicol Appl Pharmacol 192:307-322.

Boelsterli UA, Zimmerman HI and Kretz-Rommel A (1995) Idiosyncratic liver toxicity
of nonsteroidal antiinflammatory drugs: molecular mechanisms and pathology. Crit Rev
Toxic0125z207-235.

Bort R, Mace K, Boobis A, Gomez-Lechon MJ, Pfeifer A and Castell J (1999) Hepatic
metabolism of diclofenac: role of human CYP in the minor oxidative pathways. Biochem
Pharmacol 58:787-796.

Bougie D, Johnson ST, Weitekamp LA and Aster RH (1997) Sensitivity to a metabolite
of diclofenac as a cause of acute immune hemolytic anemia. Blood 90:407-413.

Bourdi M, Amouzadeh HR, Rushmore TH, Martin JL and Pohl LR (2001) Halothane-
induced liver injury in outbred guinea pigs: role of trifluoroacetylated protein adducts in
animal susceptibility. Chem Res Toxicol 14:362-370.

Bredfeldt JE and von HC (1984) Ranitidine, acetaminophen, and hepatotoxicity. Ann
Intern Med 101:719.

Brown AP, Schultze AE, Holdan WL, Buchweitz JP, Roth RA and Ganey PE (1996)
Lipopolysaccharide-induced hepatic injury is enhanced by polychlorinated biphenyls.
Environ Health Perspect 104:634-640.

Buchweitz JP, Ganey PE, Bursian SJ and Roth RA (2002) Underlying endotoxemia

augments toxic responses to Chlorpromazine: is there a relationship to drug idiosyncrasy?
J Pharmacol Exp Ther 300:460-467.

203

 

Buerke M, Weyrich AS and Lefer AM (1994) Isolated cardiac myocytes are sensitized by
hypoxia-reoxygenation to neutrophil-released mediators. Am J Physiol 266:H128-H136.

Calkins CM, Bensard DD, Shames BD, Pulido EJ, Abraham E, Fernandez N, Meng X,
Dinarello CA and McIntyre RC, Jr. (2002) IL-l regulates in vivo C-X-C chemokine
induction and neutrophil sequestration following endotoxerrria. J Endotoxin Res 8:59-67.

Camus G, Nys M, Poortrnans JR, Venneman I, Monfils T, by—Dupont G, JuchmeS-Ferir
A, Deby C, Lamy M and Duchateau J (1998) Endotoxaemia, production of tumour
necrosis factor alpha and polymorphonuclear neutrophil activation following strenuous
exercise in humans. Eur J A ppI Physiol Occup Physiol 79:62-68.

Camus G, Poortrnans J, Nys M, by—Dupont G, Duchateau J, Deby C and Lamy M (1997)
Mild endotoxaenria and the inflammatory response induced by a marathon race. Clin Sci
(Lond) 92:415-422.

Chamulitrat W, Blazka ME, Jordan SJ, Luster MI and Mason RP (1995) Tumor necrosis
factor-alpha and nitric oxide production in endotoxin-primed rats administered carbon
tetrachloride. Lifi: Sci 57:2273-2280.

Chen D, Giannopoulos K, Shiels PG, Webster Z, McVey JH, Kemball-Cook G,
Tuddenham E, Moore M, Lechler R and Dorling A (2004) Inhibition of intravascular
thrombosis in murine endotoxemia by targeted expression of hirudin and tissue factor
pathway inhibitor analogs to activated endothelium. Blood 104:1344-1349.

Choi JW, Park SC, Kang GI-I, Liu J0 and Youn HD (2004) Nur77 activated by hypoxia-
inducible factor-lalpha overproduces proopiomelanocortin in von Hippel-Lindau-
mutated renal cell carcinoma. Cancer Res 64:35-39.

Chosay JG, Essani NA, Dunn CJ and Jaeschke H (1997) Neutrophil margination and
extravasation in sinusoids and venules of liver during endotoxin-induced injury. Am J
Physiol 272:Gl 195-GI 200.

Christen U, Burgin M and Gut J (1991) Halothane metabolism: Kupffer cells carry and
partially process trifluoroacetylated protein adducts. Biochem Biophys Res Commun
175:256-262.

Cohen A and Fabre L (1983) Tolerance to repeated intravenous doses of ranitidine HCl
and cimetidine HCl in normal volunteers. Cur. Ther. Res. 34: 475-482.

Colin-Jones DG, Ireland A, Gear P, Golding PL, Ramage JK, Williams JG, Leicester RJ,
Smith CL, Ross G, Bamforth J and . (1984) Reducing overnight secretion of acid to heal
duodenal ulcers. Comparison of standard divided dose of ranitidine with a single dose
administered at night. Am JMed 77:116-122.

Colman RW, MarderVJ, Salzrnan EW, and Hirsh J (1994) Overview of hemostasis, in
Hemostasis and Thrombosis (Colman RW, Hirsh J, Marder J, Salzrnaned, eds) p 3,
Lippincott, Philadelphia.

204

Cook JA, Marconi EA and Di Luzio NR (1974) Lead, cadmium, endotoxin interaction:
effect on mortality and hepatic function. Toxicol Appl Pharmacol 28:292-302.

Copple BL, Banes A, Ganey PE and Roth RA (2002a) Endothelial cell injury and fibrin
deposition in rat liver after monocrotaline exposure. T oxicol Sci 65:309-318.

Copple BL, Ganey PE and Roth RA (2004a) Hypoxia in rat liver after monocrotaline
exposure. T oxicol Sci 78 (S-l), 491 .

Copple BL, Moulin F, Hanumegowda UM, Ganey PE and Roth RA (2003) Thrombin and
protease-activated receptor-1 agonists promote lipopolysaccharide-induced hepatocellular
injury in perfused livers. J Pharmacol Exp Ther 305:417-425.

Copple BL, Rondelli CM, Maddox J F , Hoglen NC, Ganey PE and Roth RA (2004b)
Modes of cell death in rat liver after monocrotaline exposure. T oxicol Sci 77:172-182.

Copple BL, Woolley B, Banes A, Ganey PE and Roth RA (2002b) Anticoagulants
prevent monocrotaline-induced hepatic parenchymal cell injury but not endothelial cell
injury in the rat. T oxicol Appl Pharmacol 180:186-196.

Cottam DR, Mattar SG, Barinas-Mitchell E, Eid G, Kuller L, Kelley DE and Schauer PR
(2004) The chronic inflammatory hypothesis for the morbidity associated with morbid
obesity: implications and effects of weight loss. Obes Surg 14:589-600.

Coutellier A, Pelletier S, Ekert P, Hauw JJ, Soubrie C and Herson S (1993) Inflammatory
myopathy during ranitidine therapy. J Rheumatol 20: 1453-1454.

Dalgleish AG and O'Byme KJ (2002) Chronic immune activation and inflammation in
the pathogenesis of AIDS and cancer. Adv Cancer Res 84:231-276.

dc GH and N011 T (1984) The crucial role of hypoxia in halothane-induced lipid
peroxidation. Biochem Biophys Res Commun 119:139-143.

Deaciuc IV, McDonough KH, Bagby GJ and Spitzer JJ (1993) Alcohol consumption in
rats potentiates the deleterious effect of gram-negative sepsis on hepatic hyaluronan
uptake. Alcohol Clin Exp Res 17:1002-1008.

Derevianko A, D'Amico R and Simms H (1996) Polymorphonuclear leucocyte (PMN)-
derived inflammatory cytokines-regulation by oxygen tension and extracellular matrix.
Clin Exp Immunol 106:560-567.

Desaint B, Alperine M, Legendre C, Florent C and Levy VG (1985) [Hepatitis caused by
ranitidine]. Gastroenterol Clin Biol 9:746-747.

Devuyst O, Lefebvre C, Geubel A and Coche E (1993) Acute cholestatic hepatitis with
rash and hypereosinophilia associated with ranitidine treatment. Acta Clin Belg 48: 109-
114.

205

Dieckhaus CM, Thompson CD, Roller SG and Macdonald TL (2002) Mechanisms of
idiosyncratic drug reactions: the case of felbamate. Chem Biol Interact 142:99-117.

Duijvestijn AM, van GH, Klatter F, Majoor GD, van BE and van B, V (1992) Antibodies
defining rat endothelial cells: RECA-l, a pan-endothelial cell-specific monoclonal
antibody. Lab Invest 66:459-466.

El-Bassiouni EA, bo-Ollo MM, Helrny MH, Ismail S and Ramadan MI (1998) Changes
in the defense against fiee radicals in the liver and plasma of the dog during hypoxia
and/or halothane anaesthesia. Toxicology 128:25-34.

Engelmann B, Luther T and Muller I (2003) Intravascular tissue factor pathway—a model
for rapid initiation of coagulation within the blood vessel. Thromb Haemost 89:3-8.

Essani NA, Bajt ML, Farhood A, Vonderfecht SL and Jaeschke H (1997) Transcriptional
activation of vascular cell adhesion molecule-1 gene in vivo and its role in the
pathophysiology of neutrophil-induced liver injury in murine endotoxin shock. J
Immunol 158:5941-5948.

Essani NA, Fisher MA, Farhood A, Manning AM, Smith CW and Jaeschke H (1995)
Cytokine-induced upregulation of hepatic intercellular adhesion molecule-1 messenger
RNA expression and its role in the pathophysiology of murine endotoxin Shock and acute
liver failure. Hepatology 21 :1632-1639.

Essayan DM (1999) Cyclic nucleotide phosphodiesterase (PDE) inhibitors and
immunomodulation. Biochem Pharmacol 57:965-973.

Faich GA and Moseley RH (2001) Troglitazone (Rezulin) and hepatic injury.
Pharmacoepidemiol Drug Saf 10:53 7-547.

F assoulaki A, Eger EI, Johnson BH, Ferrell LD, Smuckler EA, Harper MH, Eger RR and
Cahalan MK (1984) Brief periods of hypoxia can produce hepatic injury in rats. Anesth
Analg 63:885-887.

F rancavilla A, Panella C, Polimeno L, Di Leo A, Makowka L, Barone M, Amoruso A,
Ingrosso M and Starzl TE (1989) Effect of cimetidine, ranitidine, famotidine and
omeprazole on hepatocyte proliferation in vitro. J Hepatol 8:32-41.

F ujii K, Morio M, Kikuchi H, Ishihara S, Okida M and Ficor F (1984) In vivo spin-trap
study on anaerobic dehalogenation of halothane. Life Sci 35:463-468.

Fujiwara K, Ogata I, Ohta Y, Hirata K, Oka Y, Yamada S, Sato Y, Masaki N and Oka H
(1988) Intravascular coagulation in acute liver failure in rats and its treatment with
antithrombin III. Gut 29:1103-1108.

Funk C, Ponelle C, Scheuermann G and Pantze M (2001) Cholestatic potential of
troglitazone as a possible factor contributing to troglitazone-induced hepatotoxicity: in

206

vivo and in vitro interaction at the canalicular bile salt export pump (Bsep) in the rat. Mol
Pharmacol 59:627-635.

Furst SM and Gandolfi AJ (1997) Interaction of lymphocytes with Kupffer cells from
halothane-exposed guinea pigs. Int Arch Allergy Immunol 1 14:46-53.

Furst SM, Luedke D and Gandolfi AJ (1997) Kupffer cells from halothane-exposed
guinea pigs carry trifluoroacetylated protein adducts. Toxicology 120:119-132.

Ganey PE, Luyendyk JP, Maddox JF, and Roth RA. (2004). Adverse hepatic drug
reactions: inflammatory episodes as consequence and contributor. Chem Biol Interact.
150:35-51.

Ganey PE, Bailie MB, VanCise S, Colligan ME, Madhukar BV, Robinson JP and Roth
RA (1994) Activated neutrophils from rat injured isolated hepatocytes. Lab Invest 70:53-
60.

Ganey PE, Barton YW, Kinser S, Sneed RA, Barton CC and Roth RA (2001)
Involvement of cyclooxygenase-2 in the potentiation of allyl alcohol-induced liver injury
by bacterial lipopolysaccharide. T oxicol App] Pharmacol 174:1 13-121.

Ganey PE and Roth RA (2001) Concurrent inflammation as a determinant of
susceptibility to toxicity from xenobiotic agents. Toxicology 169:195-208.

Gao JJ, Diesl V, Wittrnann T, Morrison DC, Ryan JL, Vogel SN and Follettie MT (2002)
Regulation of gene expression in mouse macrophages stimulated with bacterial CpG-
DNA and lipopolysaccharide. J Leukoc Biol 72: 1234-1245.

Garcia RL, Wallander MA and Stricker BH (1997) The risk of acute liver injury
associated with cimetidine and other acid-suppressing anti-ulcer drugs. Br J Clin
Pharmacol 43:183-188.

Gardiner KR, Halliday MI, Barclay GR, Milne L, Brown D, Stephens S, Maxwell RJ and
Rowlands BJ (1995) Significance of systemic endotoxaerrria in inflammatory bowel
disease. Gut 36:897-901.

Gereda JE, Klinnert MD, Price MR, Leung DY and Liu AH (2001) Metropolitan home
living conditions associated with indoor endotoxin levels. J Allergy Clin Immunol
107:790-796.

Gill RQ and Sterling RK (2001) Acute liver failure. J Clin Gastroenterol 33:191-198.

Gomez-Lechon MJ, Ponsoda X, O'Connor E, Donato T, Castell JV and Jover R (2003a)
Diclofenac induces apoptosis in hepatocytes by alteration of mitochondrial function and
generation of ROS. Biochem Pharmacol 66:2155-2167.

Gomez-Lechon MJ, Ponsoda X, O'Connor E, Donato T, Jover R and Castell JV (2003b)
Diclofenac induces apoptosis in hepatocytes. Toxicol In Vitro 17:675-680.

207

Graham DY, Opekun AR, Smith JL and Schwartz JT (1985) Ranitidine and
hepatotoxicity. Ann Intern Med 102:4] 6.

Gray MO, Karliner J S and Mochly-Rosen D (1997) A selective epsilon-protein kinase C
antagonist inhibits protection of cardiac myocytes from hypoxia-induced cell death. J
Biol Chem 272:30945-30951.

Gruber F, Hufnagl P, Hofer-Warbinek R, Schmid J A, Breuss JM, Huber-Beckmann R,
Lucema M, Papac N, Harant H, Lindley I, de MR and Binder BR (2003) Direct binding
of Nur77/NAK-1 to the plasminogen activator inhibitor 1 (PAI-l) promoter regulates
TNF alpha -induced PAI-l expression. Blood 101:3042-3048.

Gujral J S, Hinson JA, Farhood A and Jaeschke H (2004) NADPH oxidase-derived
oxidant stress is critical for neutrophil cytotoxicity during endotoxemia. Am J Physiol
Gastrointest Liver Physiol 287 :G243-6252.

Gunawan B and Kaplowitz N (2004) Clinical perspectives on xenobiotic-induced
hepatotoxicity. Drug Metab Rev 36:301-312.

Gut J, Christen U and Huwyler J (1993) Mechanisms of halothane toxicity: novel
insights. Pharmacol Ther 58:133-155.

Halparin LS (1984) Adverse effects of ranitidine therapy. Can Med Assoc J 130:668, 672.

Hamaguchi E, Takamura T, Shimizu A and Nagai Y (2003) Tumor necrosis factor-alpha
and troglitazone regulate plasminogen activator inhibitor type 1 production through
extracellular signal-regulated kinase- and nuclear factor-kappaB-dependent pathways in
cultured human umbilical vein endothelial cells. J Pharmacol Exp Ther 307 :987-994.

Hannah S, Mecklenburgh K, Rahman I, Bellingan GJ, Greening A, Haslett C and
Chilvers ER (1995) Hypoxia prolongs neutrophil survival in vitro. FEBS Lett 372:233-
237.

Hansen J, Cherwitz DL and Allen J] (1994) The role of tumor necrosis factor-alpha in
acute endotoxin-induced hepatotoxicity in ethanol-fed rats. Hepatology 20:46] -474.

Hashimoto F, Davis RL and Egli D (1994) Hepatitis following treatments with
famotidine and then cimetidine. Ann Pharmacother 28:37-39.

Hastings KL, Thomas C, Brown AP and Gandolfi AJ (1995) Trifluoroacetylation
potentiates the humoral immune response to halothane in the guinea pig.
Immunopharmaco] Immunotoxico] 17:201-213.

Hayashi T and Faustrnan D (2000) Essential role of human leukocyte antigen-encoded

proteasome subunits in NF-kappaB activation and prevention of tumor necrosis factor-
alpha-induced apoptosis. J Biol Chem 275:5238-5247.

208

Heinrich J, Gehring U, Douwes J, Koch A, Fahlbusch B, Bischof W and Wichmann HE
(2001) Pets and vermin are associated with high endotoxin levels in house dust. Clin Exp
Allergy 31:1839-1845.

Helfgott SM, Sandberg-Cook J, Zakirn D and Nestler J (1990) Diclofenac-associated
hepatotoxicity. JAMA 264:2660-2662.

Hewett J A, Jean PA, Kunkel SL and Roth RA (1993) Relationship between tumor
necrosis factor-alpha and neutrophils in endotoxin-induced liver injury. Am J Physiol
265:G1011-61015.

Hewett JA and Roth RA (1993) Hepatic and extrahepatic pathobiology of bacterial
lipopolysaccharides. Pharmacol Rev 45:382-41 1.

Hewett J A and Roth RA (1995) The coagulation system, but not circulating fibrinogen,
contributes to liver injury in rats exposed to lipopolysaccharide fi'om gram-negative
bacteria. J Pharmacol Exp Ther 272:53-62.

Hewett J A, Schultze AE, VanCise S and Roth RA (1992) Neutrophil depletion protects
against liver injury from bacterial endotoxin. Lab Invest 66:347-361.

Hiesse C, Cantarovich M, Santelli C, Francais P, Charpentier B, Fries D and Buffet C
(1985) Ranitidine hepatotoxicity in renal transplant patient. Lancet 1:1280.

Hirschowitz BI, Berenson MM, Berkowitz JM, Bright-Asare P, DeLuca VA, Jr.,
Eshelman FN, Font RG, Griffin JW, Jr., Kozarek RA, McCray RS et a]. (1986) A
multicenter study of ranitidine treatment of duodenal ulcers in the United States. J Clin
Gastroenterol 8:3 59—366.

Ho J S, Buchweitz JP, Roth RA and Ganey PE (1996) Identification of factors from rat
neutrophils responsible for cytotoxicity to isolated hepatocytes. J Leukoc Biol 59:716-
724.

Honche] R, Marsano L, Cohen D, Shedlofsky S and McClain CJ (1991) Lead enhances
lipopolysaccharide and tumor necrosis factor liver injury. J Lab Clin Med 117:202-208.

Homyak SC, Gehlsen KR and Haaparanta T (2003) Histamine dihydrochloride protects
against early alcohol-induced liver injury in a rat model. Inflammation 27:317-327.

Horton MR, Olrnan MA, Bao C, White KB, Choi AM, Chin BY, Noble PW and
Lowenstein CJ (2000) Regulation of plasminogen activator inhibitor-1 and urokinase by
hyaluronan fragments in mouse macrophages. Am J Physiol Lung Cell Mo] Physio]
279:L707-L71 5.

Ilyin GP, Rissel M, Malledant Y, Tanguy M and Guillouzo A (1994) Human hepatocytes
express trifluoroacetylated neoantigens afier in vitro exposure to halothane. Biochem
Pharmacol 48:561-567.

209

Ishii K, Ito Y, Katagiri H, Matsumoto Y, Kakita A and Majima M (2002) Neutrophil
elastase inhibitor attenuates lipopolysaccharide-induced hepatic microvascular
dysfunction in mice. Shack 18:163-168.

Jacob AI, Goldberg PK, Bloom N, Degenshein GA and Kozinn PJ (1977) Endotoxin and
bacteria in portal blood. Gastroenteralagy 72:1268-1270.

Jaeschke H, Farhood A, Fisher MA and Smith CW (1996) Sequestration of neutrophils in
the hepatic vasculature during endotoxemia is independent of beta 2 integrins and
intercellular adhesion molecule-1. Shock 6:351-356.

Jaeschke H, Farhood A and Smith CW (1991) Neutrophil-induced liver cell injury in
endotoxin shock is a CD1 1b/CD18-dependent mechanism. Am J Physiol 261:G1051-
G1056.

Jaeschke H and Smith CW (1997) Mechanisms of neutrophil-induced parenchymal cell
injury. J Leukac Biol 61:647-653.

Jeukendrup AE, Vet-Joop K, Sturk A, Stegen JH, Senden J, Saris WH and Wagenmakers
AJ (2000) Relationship between gastro-intestinal complaints and endotoxaemia, cytokine
release and the acute-phase reaction during and after a long-distance triathlon in highly
trained men. Clin Sci (Lond) 98:47-55.

Jimenez-Saenz M, rguelleS-Arias F, Herrerias-Gutierrez JM and Duran-Quintana JA
(2000) Acute cholestatic hepatitis in a child treated with famotidine. Am J Gastraentera]
95:3665-3666.

J in SL and Conti M (2002) Induction of the cyclic nucleotide phosphodiesterase PDE4B
is essential for LPS-activated TNF-alpha responses. Proc Nat] Acad Sci U S A 99:7628-
7633.

Jones BE, Lo CR, Liu H, Srinivasan A, Streetz K, Valentino KL and Czaja MJ (2000)
Hepatocytes sensitized to tumor necrosis factor-alpha cytotoxicity undergo apoptosis
through caspase-dependent and caspase-independent pathways. J Biol Chem 275:705-
712.

Jones DB, Hanson RG, Mihaly GW, Johnson RD, Louis WJ, Yeomans ND and
Smallwood RA (1982) Double-blind controlled trial of ranitidine versus cimetidine in the
treatment of duodenal ulceration. Aust N Z J Med 12:547-549.

Ju C and Uetrecht JP (2002) Mechanism of idiosyncratic drug reactions: reactive
metabolite formation, protein binding and the regulation of the immune system. Curr
Drug Metab 3:367-3 77.

Kaplowitz N (2004) Drug-induced liver injury. Clin Infect Dis 38 (S-2):S44-8.

Kaplowitz N (2004b) Acetaminophen hepatoxicity: what do we know, what don't we
know, and what do we do next? Hepatology 40:23-26.

210

Karachalios GN (1985) Ranitidine and hepatitis. Ann Intern Med 103:634-635.

Kenna JG, Martin JL, Satoh H and Pohl LR (1990) Factors affecting the expression of
trifluoroacetylated liver microsomal protein neoantigens in rats treated with halothane.
Drug Metab Dispos 18:788-793.

Kenna JG, Neuberger J and Williams R (1987) Specific antibodies to halothane-induced
liver antigens in halothane-associated hepatitis. Br J Anaesth 59:1286-1290.

Kerb R, Hoffmeyer S and Brinkmann U (2001) ABC drug transporters: hereditary
polymorphisms and pharmacological impact in MDR] , MRP1 and MRP2.
Pharmacogenomics 2:51 -64.

Khan S and O'Brien PJ (1997) Rapid and specific efflux of glutathione before hepatocyte
injury induced by hypoxia. Biochem Biophys Res Commun 238:320-322.

Khan S and O'Brien PJ (1995) Molecular mechanisms of tirapazamine (SR 4233, Win
59075)-induced hepatocyte toxicity under low oxygen concentrations. Br J Cancer
71 :7 80-785.

Kharasch ED, Hankins D, Mautz D and Thummel KE (1996) Identification of the
enzyme responsible for oxidative halothane metabolism: implications for prevention of
halothane hepatitis. Lancet 347 :1367-1371 .

Kietzrnann T, J ungerrnann K and Gorlach A (2003) Regulation of the hypoxia-dependent
plasminogen activator inhibitor 1 expression by MAP kinases. Thromb Haemost 89:666-
673.

Kietzrnann T, Roth U and Jungermann K (1999) Induction of the plasminogen activator
inhibitor-1 gene expression by mild hypoxia via a hypoxia response element binding the
hypoxia-inducible factor-1 in rat hepatocytes. Blood 94:4177-4185.

Kinser S, Copple BL, Roth RA and Ganey PE (2002) Enhancement of allyl alcohol
hepatotoxicity by endotoxin requires extrahepatic factors. T axica] Sci 69:470-481.

Kinser S, Sneed R, Roth R and Ganey P (2004) Neutrophils contribute to endotoxin
enhancement of allyl alcohol hepatotoxicity. J T oxicol Environ Health A 67 :91 1-928.

Kitteringham NR, Kenna JG and Park BK (1995) Detection of autoantibodies directed
against human hepatic endoplasmic reticulum in sera from patients with halothane-
associated hepatitis. Br J Clin Pharmacol 40:379-386.

Knolle PA and Gerken G (2000) Local control of the immune response in the liver.
Immunol Rev 174:21-34.

Kobayashi H, Horikoshi K, Yamataka A, Yamataka T, Okazaki T, Lane GJ and Miyano

T (1999) Hyaluronic acid: a Specific prognostic indicator of hepatic damage in biliary
atresia. J Pediatr Surg 34:1791-1794.

2]]

Kohlroser J, Mathai J, Reichheld J, Banner BF and Bonkovsky HL (2000) Hepatotoxicity
due to troglitazone: report of two cases and review of adverse events reported to the
United States Food and Drug Administration. Am J Gastroenterol 95:272-276.

Kretz-Romme] A and Boelsterli UA (1995) Cytotoxic activity of T cells and non-T cells
fiom diclofenac-immunized mice against cultured syngeneic hepatocytes exposed to
diclofenac. Hepatalagy 22:213-222.

Kretz-Rommel A and Boelsterli UA (1993) Diclofenac covalent protein binding is
dependent on acyl glucuronide formation and is inversely related to P450-mediated acute
cell injury in cultured rat hepatocytes. Toxicol Appl Pharmacol 120: 1 55-161 .

Kretz-Rommel A and Boelsterli UA (1994) Selective protein adducts to membrane
proteins in cultured rat hepatocytes exposed to diclofenac: radiochemical and
immunochemical analysis. Ma] Pharmacol 45:23 7-244.

Labib R, Turkal] R and bdel-Rahman MS (2002) Endotoxin potentiates the hepatotoxicity
of cocaine in male mice. J Toxicol Environ Health A 65:977-993.

Lahat N, Rabat MA, Ballan M, Weiss-Cerem L, Engelmayer M and Bitterman H (2003)
Hypoxia reduces CD80 expression on monocytes but enhances their LPS-stimulated
TNF-alpha secretion. J Leukoc Biol 74:197-205.

Lane DA and Grant PJ (2000) Role of hemostatic gene polymorphisms in venous and
arterial thrombotic disease. Blood 95:1517-1 532.

Lane SR, Nicholls PJ and Sewell RD (2004) The measurement and health impact of
endotoxin contamination in organic dusts from multiple sources: focus on the cotton
industry. Inha] Toxicol 16:21 7-229.

Lasser KE, Allen PD, Woolhandler SJ, Himmelstein DU, Wolfe SM and Bar DH (2002)
Timing of new black box warnings and withdrawals for prescription medications. JAM
287:2215-2220.

Lauritsen K, Havelund T and RaSk-Madsen J (1984) Ranitidine and hepatotoxicity.
Lancet 2:147].

Lazarou J, Pomeranz BH and Corey PN (1998) Incidence of adverse drug reactions in
hospitalized patients: a meta-analysis of prospective studies. JAM 279: 1200-1205.

Lee Fl, Reed PI, Crowe JP, McIsaac RL and Wood JR (1986) Acute treatment of
duodenal ulcer: a multicentre study to compare ranitidine 150 mg twice daily with
ranitidine 300 mg once at night. Gut 27:1091-1095.

Lemasters JJ, J i S and Thurman RG (1981) Centrilobular injury following hypoxia in
isolated, perfused rat liver. Science 213:661-663.

212

Lepper PM, Held TK, Schneider EM, Bolke E, Gerlach H and Trautrnann M (2002)
Clinical implications of antibiotic-induced endotoxin release in septic Shock. Intensive
Care Med 28:824-833.

Lewis J.H. (2003) Nonsteroidal anti-inflammatory drugs: pathology and clinical
presentation of hepatoxicity, in Drug-Induced Liver Disease (Kaplowitz N and DeLeve
LD ed) pp 377-404, Marcel Dekker, New York.

Li AP (2002) A review of the common properties of drugs with idiosyncratic
hepatotoxicity and the "multiple determinant hypothesis" for the manifestation of
idiosyncratic drug toxicity. Chem Biol Interact 142:7-23.

Li C, Liu J, Waalkes MP and Jaeschke H (2003) Gene array analysis of the hepatic
response to endotoxin in glutathione peroxidase-deficient mice. T oxica] Lett 144:397-
406.

Li X, Klintrnan D, Liu Q, Sato T, Jeppsson B and Thorlacius H (2004) Critical role of
CXC chemokines in endotoxemic liver injury in mice. J Leukoc Biol 75:443-452.

Liberopoulos EN, Nonni AB, Tsianos EV and Elisaf MS (2002) Possible ranitidine-
induced cholestatic jaundice. Ann Pharmacother 36:172.

Lin JH (1991) Pharmacokinetic and pharmacodynamic properties of histamine H2-
receptor antagonists. Relationship between intrinsic potency and effective plasma
concentrations. Clin Pharmacokinet 20:218-236.

Lind L (2003) Circulating markers of inflammation and atherosclerosis. Atherosclerosis
169:203-214.

Lind RC and Gandolfi AJ (1997) Late dimethyl sulfoxide administration provides a
protective action against chemically induced injury in both the liver and the kidney.
T oxico] App] Pharmacol 142:201-207.

Lind RC, Gandolfi AJ and Hall PM (1992) Glutathione depletion enhances subanesthetic
halothane hepatotoxicity in guinea pigs. Anesthesiology 77:721-727.

Lind RC, Gandolfi AJ and Hall PM (1994) A model for fatal halothane hepatitis in the
guinea pig. Anesthesiology 81:478-487.

Lind RC, Gandolfi AJ, Sipes IG and Brown BRJ (1984) The involvement of endotoxin in
halothane-associated liver injury. Anesthesiology 61:544-550.

Liu H, Lo CR and Czaja MJ (2002) NF -kappaB inhibition sensitizes hepatocytes to TNF-
induced apoptosis through a sustained activation of JNK and c-Jun. Hepatology 35:772-
778.

Luparini RL, Rotundo A, Mattace R and Marigliano V (2000) [Possibly ranitidine-
induced autoimmune hepatitis]. Ann Ital Med Int 15:214-217.

213

Luyendyk JP, Copple BL, Barton CC, Ganey PE and Roth RA (2003a) Augmentation of
aflatoxin B1 hepatotoxicity by endotoxin: involvement of endothelium and the
coagulation system. T axica] Sci 72:171-1 81 .

Luyendyk JP, Maddox JF, Cosma GN, Ganey PE, Cockerell GL and Roth RA (2003b)
Ranitidine treatment during a modest inflammatory response precipitates idiosyncrasy-
like liver injury in rats. J Pharmacol Exp Ther 307 :9-1 6.

Luyendyk JP, Mattes WB, Burgoon LD, Zacharewski TR, Maddox JF, Cosma GN,
Ganey PE and Roth RA (2004c) Gene expression analysis points to hemostasis in livers
of rats cotreated with lipopolysaccharide and ranitidine. Toxicol Sci 80:203-213.

Luyendyk JP, Shores KC, Ganey PE and Roth RA (2002) Bacterial lipopolysaccharide
exposure alters aflatoxin B(1) hepatotoxicity: benchmark dose analysis for markers of
liver injury. T axica] Sci 68:220-225.

Luyendyk JP, Maddox JF, Ganey PE, and Roth RA. (2004a). Coagulation system
activation and fibrin deposition are critical factors in idiosyncrasy-like liver injury from
lipopolysaccharide and ranitidine coexposure. FASEB J in press. Abstract

Luyendyk JP, Maddox JF, Green, CD Ganey PE, and Roth RA. (2004b) Role of hepatic
fibrin in idiosyncrasy-like liver injury from lipopolysaccharide-ranitidine coexposure in
rats. Hepatology in press.

Mackman N (1996) Regulation of tissue factor gene expression in human monocytic and
endothelial cells. Haemostasis 26 Suppl 1:17-19.

Maddox JF, Domzalski AC, Roth RA and Ganey PE (2004) 15-deoxy prostaglandin J2
enhances ally] alcohol-induced toxicity in rat hepatocytes. T axical Sci 77:290-298.

Maher JJ, Scott MK, Saito JM and Burton MC (1997) Adenovirus-mediated expression
of cytokine-induced neutrophil chemoattractant in rat liver induces a neutrophilic
hepatitis. Hepatology 25:624-630.

Marotto ME, Thurman RG and Lemasters JJ (1988) Early midzonal cell death during
low-flow hypoxia in the isolated, perfused rat liver: protection by allopurinol. Hepatology
8:585-590.

Marra F, DeFranco R, Grappone C, Milani S, Pinzani M, Pellegrini G, Laffi G and
Gentilirri P (1998) Expression of the thrombin receptor in human liver: up-regulation
during acute and chronic injury. Hepatology 27 :462-471 .

Marshall LJ, Ramdin LS, Brooks T, DPhil PC and Shute JK (2003) Plasminogen
activator inhibitor-1 supports IL-8-mediated neutrophil transendothelial migration by
inhibition of the constitutive shedding of endothelial IL-8/heparan sulfate/syndecan-l
complexes. J Immunol 171:2057-2065.

214

Martin JL, Kenna JG, Martin BM, Thomassen D, Reed GF and Pohl LR (1993a)
Halothane hepatitis patients have serum antibodies that react with protein disulfide
isomerase. Hepatology 18:858-863.

Martin JL, Reed GF and Pohl LR (1993b) Association of anti-58 kDa endoplasmic
reticulum antibodies with halothane hepatitis. Biochem Pharmacol 46: 1247-1250.

Martino G, Adorini L, Rieckmann P, Hillert J, Kalhnann B, Comi G and Filippi M (2002)
Inflammation in multiple sclerosis: the good, the bad, and the complex. Lancet Neural
1:499-509.

Masubuchi Y, Nakayama S and Horie T (2002) Role of mitochondrial permeability
transition in diclofenac-induced hepatocyte injury in rats. Hepatology 35:544-551.

Matsumaru K, Ji C and Kaplowitz N (2003) Mechanisms for sensitization to TNF-
induced apoptosis by acute glutathione depletion in murine hepatocytes. Hepatology
37: 1425-1434.

Mattes WB (2004) Annotation and cross-indexing of array elements on multiple
platforms. Environ Health Perspect 112:506-510.

Mavier P, Preaux AM, Guigui B, Lescs MC, Zafiani ES and Dhumeaux D (1988) In vitro
toxicity of polymorphonuclear neutrophils to rat hepatocytes: evidence for a proteinase-
mediated mechanism. Hepatology 8:254-258. -

McGirr LG, Khan S, Lauriault V and O'Brien PJ (1990) Molecular mechanisms for
bromotrichloromethane cytotoxicity in isolated rat hepatocytes. Xenobiotica 20:933-943.

McLain GE, Sipes IG and Brown BR, Jr. (1979) An animal model of halothane
hepatotoxicity: roles of enzyme induction and hypoxia. Anesthesiology 51:321-326.

Meyer 0, Hoffmann T, Aslan T, Ahrens N, Kiesewetter H and Salama A (2003)
Diclofenac-induced antibodies against RBCS and platelets: two case reports and a concise
review. T ransfitsion 43:345-349.

Mills JG, Koch KM, Webster C, Sirgo MA, Fitzgerald K and Wood JR (1997) The safety
of ranitidine in over a decade of use. Aliment Pharmacol Ther 11:129-137.

Minoda Y and Kharasch ED (2001) Halothane-dependent lipid peroxidation in human
liver microsomes is catalyzed by cytochrome P4502A6 (CYP2A6). Anesthesiology
95:509-5 14.

Mizoguchi Y, Shin T, Mizuno M, Seki S, Kobayashi K and Morisawa S (1990) The

enhancement of liver injury by lipopolysaccharide in an experimental drug-induced
allergic hepatitis model. Gastroentero] Jpn 25 :458-462.

215

Montes R, Declerck PJ, Calvo A, Montes M, Hermida J, Munoz MC and Rocha E (2000)
Prevention of renal fibrin deposition in endotoxin-induced DIC through inhibition of
PAI-l . Thromb Haemost 84:65-70.

Moulin F, Copple BL, Ganey PE and Roth RA (2001) Hepatic and extrahepatic factors
critical for liver injury during lipopolysaccharide exposure. Am J Physiol Gastrointest
Liver Physiol 281:G1423-G1431.

Moulin F, Pearson JM, Schultze AE, Scott MA, Schwartz KA, Davis JM, Ganey PE and
Roth RA (1996) Thrombin is a distal mediator of lipopolysaccharide-induced liver injury
in the rat. J Surg Res 65: 149-158.

Murphy EJ, Davem TJ, Shakil AO, Shick L, Masharani U, Chow H, F reise C, Lee WM
and Bass NM (2000) Troglitazone-induced fulminant hepatic failure. Acute Liver Failure
Study Group. Dig Dis Sci 2000 45:549-553.

Nagai H, Matsumaru K, Feng G and Kaplowitz N (2002) Reduced glutathione depletion
causes necrosis and sensitization to tlnnor necrosis factor-alpha-induced apoptosis in
cultured mouse hepatocytes. Hepatology 36:55-64.

Nakamoto Y and Kaneko S (2003) Mechanisms of viral hepatitis induced liver injury.
Curr Mol Med 3:537-544.

Neville LF, Matlriak G and Bagasra O (1997) The irnmunobiology of interferon-gamma
inducible protein 10 kD (IP-10): a novel, pleiotropic member of the C-X-C chemokine
superfanrily. Cytokine Growth F actor Rev 8:207-219.

Njoku DB, Greenberg RS, Bourdi M, Borkowf CB, Dake EM, Martin J L and Pohl LR
(2002) Autoantibodies associated with volatile anesthetic hepatitis found in the sera of a
large cohort of pediatric anesthesiologists. Anesth Analg 94:243-9, table.

Nolan JP (1989) Intestinal endotoxins aS mediators of hepatic injury--an idea whose time
has come again. Hepatology 10:887-891.

Offit KO (1984) Ranitidine: Correspondence. N. Engl. J. Med. 310:1603-1604.

Okajima K, Harada N and Uchiba M (2002) Ranitidine reduces ischemia/reperfirsion—
induced liver injury in rats by inhibiting neutrophil activation. J Pharmacol Exp Ther
301:]157-1165.

Oldham HG, Norman SJ and Chenery RJ (1985) Primary cultures of adult rat
hepatocytes--a model for the toxicity of histamine 112-receptor antagonists. Toxicology
36:21 5-229.

Ozawa K, Kuwabara K, Tamatani M, Takatsuji K, Tsukamoto Y, Kaneda S, Yanagi H,
Stern DM, Eguclri Y, Tsujimoto Y, Ogawa S and Tohyama M (1999) 150-kDa oxygen-
regulated protein (ORPl 50) suppresses hypoxia-induced apoptotic cell death. J Biol

Chem 274:6397-6404.

216

Park JH, Spiegehnan DL, Burge HA, Gold DR, Chew GL and Milton DK (2000)
Longitudinal study of dust and airborne endotoxin in the home. Environ Health Perspect
108:1023-1028.

Patterson RM, Stachlewitz R and Germolec D (2003) Induction of apoptosis by 2,3,7,8-
tetrachlorodibenzo-p—dioxin following endotoxin exposure. Toxicol Appl Pharmacol
190: 120-134.

Pawlinski R, Pedersen B, Kehrle B, Aird WC, Frank RD, Guha M and Mackman N
(2003) Regulation of tissue factor and inflammatory mediators by Egr-l in a mouse
endotoxenria model. Blood 101:3940-3947.

Pawlinski R, Pedersen B, Schabbauer G, Tencati M, Holscher T, Boisvert W, ndrade-
Gordon P, Frank RD and Mackman N (2004) Role of tissue factor and protease-activated
receptors in a mouse model of endotoxemia. Blood 103: 1342-1347.

Pearson JM, Brown AP, Schultze AE, Ganey PE and Roth RA (1996a) Gadolinium
chloride treatment attenuates hepatic platelet accumulation afier lipopolysaccharide
administration. Shack 5:408-415.

Pearson JM, Schultze AB, Jean PA and Roth RA (1995) Platelet participation in liver
injury fi'om gram-negative bacteria] lipopolysaccharide in the rat. Shock 4: 1 78-186.

Pearson JM, Schultze AE, Schwartz KA, Scott MA, Davis JM and Roth RA (1996b) The
thrombin inhibitor, hirudin, attenuates lipopolysaccharide-induced liver injury in the rat.
J Pharmacol Exp Ther 278:378-383.

Pellock JM (1999) Felbamate. Epilepsia 40 Supp15:SS7-S62.

Pfeilschifier J, Koditz R, Pfohl M and Schatz H (2002) Changes in proinflammatory
cytokine activity after menopause. Endocr Rev 23:90-1 l9.

Pirmohamed M, Breckenridge AM, Kitteringham NR and Park BK (1998) Adverse drug
reactions. BMI316:1295-1298.

Pirmohamed M, Madden S and Park BK (1996) Idiosyncratic drug reactions. Metabolic
bioactivation as a pathogenic mechanism. Clin Pharmacokinet 31:215-230.

Pirmohamed M, Naisbitt DJ, Gordon F and Park BK (2002) The danger hypothesis--
potential role in idiosyncratic drug reactions. Toxicology 181-182:55-63.

Platz EA and De Marzo AM (2004) Epidemiology of inflammation and prostate cancer. J
Urol 171:S36-S40.

Price CJ, Warburton EA and Menon DK (2003) Human cellular inflammation in the
pathology of acute cerebral ischaenria. J Neural Neurosurg Psychiatry 74: 1476-1484.

Proctor JD (1984) Hepatitis associated with ranitidine. JAMA 251:1554.

217

Pumford NR, Martin BM, Thomassen D, Burris JA, Kenna JG, Martin JL and Pohl LR
(1993a) Serum antibodies from halothane hepatitis patients react with the rat endoplasmic
reticulum protein ERp72. Chem Res T axicol 6:609-615.

Pumford NR, Myers TG, Davila J C, I-Iighet RI and Pohl LR (1993b) Irnmunochemical
detection of liver protein adducts of the nonsteroidal antiinflammatory drug diclofenac.
Chem Res Toxicol 6: 147-150.

Purcell P, Henry D and Melville G (1991) Diclofenac hepatitis. Gut 32: 1 381-1385.

Ramrakhiani S, Brunt EM and Bacon BR (1998) Possible cholestatic injury from
ranitidine with a review of the literature. Am J Gastraentera] 93:822-826.

Ribeiro JM, Lucas M, Baptista A and Victorino RM (2000) Fatal hepatitis associated
with ranitidine. Am J Gastroenterol 95:559-560.

Rice SA, Maze M, Smith CM, Kosek JC and Mazze RI (1987) Halothane hepatotoxicity
in Fischer 344 rats pretreated with isoniazid. T axica] Appl Pharmacol 87 :41 1-419.

Roth RA, Harkema JR, Pestka JP and Ganey PE (1997) Is exposure to bacterial
endotoxin a determinant of susceptibility to intoxication from xenobiotic agents? T oxicol
App] Pharmacol 147 :300-3 1 l.

Roth RA, Luyendyk JP, Maddox JF and Ganey PE (2003) Inflammation and drug
idiosyncrasy—is there a connection? J Pharmacol Exp Ther 307:1-8.

Rumbeiha WK, Fitzgerald SD, Braselton WE, Roth RA and Kaneene JB (2000)
Potentiation of mercury-induced nephrotoxicity by endotoxin in the Sprague-Dawley rat.
Toxicology 149:75-87.

Rush GF, Ripple M and Chenery R (1985) Mechanism of oxmetidine (SK&F 92994)
cytotoxicity in isolated rat hepatocytes. J Pharmacol Exp Ther 233:741-746.

Sachs UJ, Santoso S, Roder L, Smart E, Bein G and Kroll H (2004) Diclofenac-induced
antibodies against red blood cells are heterogeneous and recognize different epitopes.
T ransfitsian 44:1226-1230.

Sallustio BC and Holbrook FL (2001) In vivo perturbation of rat hepatocyte canalicular
membrane firnction by diclofenac. Drug Metab Dispos 29:1535-1538.

Sato N, F uj ii K, Yuge O and Moria M (1990) The association of halothane-induced lipid
peroxidation with the anaerobic metabolism of halothane: an in vitro study in guinea pig
liver microsomes. Hiroshima J Med Sci 39: 1-6.

Satoh H, Fukuda Y, Anderson DK, Ferrans VJ, Gillette JR and Pohl LR (1985)
Immunological studies on the mechanism of halothane-induced hepatotoxicity:
immunohistochemical evidence of trifluoroacetylated hepatocytes. J Pharmacol Exp Ther
233:857-862.

218

Savov JD, Brass DM, Berman KG, McElvania E and Schwartz DA (2003) F ibrinolysis in
LPS-induced chronic airway disease. Am J Physiol Lung Cell Mol Physiol 285:L940-
L948.

Sawdey MS and Loskutoff DJ (1991) Regulation of murine type 1 plasminogen activator
inhibitor gene expression in viva. Tissue Specificity and induction by lipopolysaccharide,
tumor necrosis factor-alpha, and transforming growth factor-beta. J Clin Invest 88: 1346-
1 3 53.

Scarpignato C, Tramacere R and Zappia L (1987) Antisecretory and antiulcer effect of
the H2-receptor antagonist famotidine in the rat: comparison with ranitidine. Br J
Pharmacol 92:153-159.

Schnitt SJ, Stilhnan IE, Owings DV, Kishimoto C, Dvorak HF and Abelmann WH
(1993) Myocardial fibrin deposition in experimental viral myocarditis that progresses to
dilated cardiomyopathy. Circ Res 72:914-920.

Schultze AE and Roth RA (1998) Chronic pulmonary hypertension--the monocrotaline
mode] and involvement of the hemostatic system. J Toxicol Environ Health B Crit Rev
1 :271-346.

Schwabe RF, Uchinami H, Qian T, Bennett BL, Lemasters JJ and Brenner DA (2004)
Differential requirement for c-Jun NH2-termina] kinase in TNFalpha- and Pas-mediated
apoptosis in hepatocytes. FASEB J 18:720-722.

Seguin B and Uetrecht J. (2003). The danger hypothesis applied to idiosyncratic drug
reactions. Curr Opin Allergy Clin Immunol 3:23 5-242.

Seitz S, Kretz-Rommel A, Oude Elferink RP and Boelsterli UA (1998) Selective protein
adduct formation of diclofenac glucuronide is critically dependent on the rat canalicular
conjugate export pump (Mrp2). Chem Res Toxicol 11:513-519.

Semenza GL (1999) Regulation of mammalian 02 homeostasis by hypoxia-inducible
factor 1. Annu Rev Cell Dev Biol 15:551-578.

Sgro C, Clinard F, Ouazir K, Chanay H, Allard C, Guilleminet C, Lenoir C, Lemoine A
and Hillon P (2002) Incidence of drug-induced hepatic injuries: a French population-
based study. Hepatology 36:451-455.

Shayiq RM, Roberts DW, Rothstein K, Snawder JE, Benson W, Ma X and Black M
(1999) Repeat exposure to incremental doses of acetaminophen provides protection
against acetaminophen-induced lethality in mice: an explanation for high acetaminophen
dosage in humans without hepatic injury. Hepatology 29:451-463.

Shen ES, Garry VF and Anders MW (1982) Effect of hypoxia on carbon tetrachloride
hepatotoxicity. Biochem Pharmacol 31 :3787-3 793.

219

Shi L, Kishore R, McMullen MR and Nagy LE (2002) Lipopolysaccharide stimulation of
ERK1/2 increases TNF-alpha production via Egr-l. Am J Physiol Cell Physiol
282:C1205-C1211.

Shibayama Y (1987) Enhanced hepatotoxicity of endotoxin by hypoxia. Pathol Res Pract
182:390-3 95.

Shishehbor MH and Hazen SL (2004) Inflammatory and oxidative markers in
atherosclerosis: relationship to outcome. Curr Atheroscler Rep 62243-250.

Silva JM, Rao DN and O'Brien PJ (1992) Modulation of trenimon-induced cytotoxicity
by DT-diaphorase in isolated rat hepatocytes under aerobic versus hypoxic conditions.
Cancer Res 52:3015-3021.

Smith AC, Roberts SM, Berman LM, Harbison RD and James RC (1988) Effects of
piperonyl butoxide on halothane hepatotoxicity and metabolism in the hyperthyroid rat.
Toxicology 50:95-105.

Sneed RA, Buchweitz JP, Jean PA and Ganey PE (2000) Pentoxifylline attenuates
bacterial lipopolysaccharide-induced enhancement of allyl alcohol hepatotoxicity.
Toxicol Sci 56:203-210.

Sneed RA, Grimes SD, Schultze AE, Brown AP and Ganey PE (1997) Bacterial
endotoxin enhances the hepatotoxicity of allyl alcohol. T axicoI App] Pharmacol 144:77-
87.

Souza Lima MA (1984) Hepatitis associated with ranitidine. Ann Intern Med 101:207-
208.

Spracklin DK, Emery ME, Thummel KB and Kharasch ED (2003) Concordance between
trifluoroacetic acid and hepatic protein trifluoroacetylation after disulfiram inhibition of
halothane metabolism in rats. Acta Anaesthesia] Scand 47 :765-770.

Spracklin DK, Hankins DC, Fisher JM, Thummel KB and Kharasch ED (1997)
Cytochrome P450 2E1 is the principal catalyst of human oxidative halothane metabolism
in vitro. J Pharmacol Exp Ther 281:400-41 1.

Spracklin DK, Thummel KB and Kharasch ED (1996) Human reductive halothane
metabolism in vitro is catalyzed by cytochrome P450 2A6 and 3A4. Drug Metab Dispos
24:976-983.

Stadrricki A (1984) Ranitidine and haemostasis in man. Hepatagastroenteralogy 31 :230-
232.

Stier A (1964) TRIFLUOROACETIC ACID AS METABOLITE OF HALOTHANE.
Biochem Pharmacol 13:1544.

220

Stroka DM, Burkhardt T, Desbaillets I, Wenger RH, Neil DA, Bauer C, Gassmann M and
Candinas D (200]) HIF-l is expressed in norrnoxic tissue and displays an organ-Specific
regulation under systemic hypoxia. FASEB J 15:2445-2453.

Su GL, Klein RD, Aminlari A, Zhang HY, Steinstraesser L, Alarcon WH, Remick DG
and Wang SC (2000) Kupffer cell activation by lipopolysaccharide in rats: role for
lipopolysaccharide binding protein and toll-like receptor 4. Hepatology 31:932-936.

Suzuki H and Sugiyama Y (2002) Single nucleotide polymorphisms in multidrug
resistance associated protein 2 (MRP2/ABCC2): its impact on drug disposition. Adv Drug
Deliv Rev 54:13] 1-1331.

Tai JH and Pestka JJ (1988) Synergistic interaction between the trichothecene T-2 toxin
and Salmonella typhirnurium lipopolysaccharide in C3H/HeN and C3H/HeJ mice.
T axico] Lett 44: 191-200.

Tamura DY, Moore BB, Partrick DA, Johnson JL, Offner PJ and Silliman CC (2002)
Acute hypoxemia in humans enhances the neutrophil inflammatory response. Shock
17:269-273.

Tang W (2003) The metabolism of diclofenac-enzymology and toxicology perspectives.
Curr Drug Metab 4:319—329.

Thiel G and Cibelli G (2002) Regulation of life and death by the zinc finger transcription
factor Egr-l . J Cell Physiol 193:287-292.

Tolman KG and Chandramouli J (2003) Hepatotoxicity of the thiazolidinediones. Clin
Liver Dis 7:369 -79

Tracy RP (2003) Emerging relationships of inflammation, cardiovascular disease and
chronic diseases of aging. Int J Obes Relat Metab Disard 27 Suppl 3:S29-S34.

Tredger JM, Thuluvath P, Williams R and Murray-Lyon IM (1995) Metabolic basis for
high paracetamol dosage without hepatic injury: a case study. Hum Exp Toxicol 14:8-12.

Tuder RM and Voelkel NF (1998) Pulmonary hypertension and inflammation. J Lab Clin
Med 132:16-24.

Uetrecht J (2001) Prediction of a new drug's potential to cause idiosyncratic reactions.
Curr Opin Drug Discov Devel 4:55-59.

Uetrecht J (1991) Metabolism of drugs by activated leukocytes: implications for drug-
induced lupus and other drug hypersensitivity reactions. Adv Exp Med Biol 283: 121-132.

Ulrich RG, Bacon JA, Brass EP, Cramer CT, Petrella DK and Sun EL (2001) Metabolic,

idiosyncratic toxicity of drugs: overview of the hepatic toxicity induced by the anxiolytic,
panadiplon. Chem Biol Interact 134:251-270.

22]

Valois M, Cooper MA and Shear NH (2003) Clinical pharmacology consultations:
consultation requests may be misleading - an organized approach to drug-induced
hepatitis. Can J Clin Pharmacol 10:59-62.

van Bommel EF and Meyboom RH (1992) [Liver damage caused by ranitidine]. Ned
Tijdschr Geneeskd 136:435-437.

Van Dyke TE and Serhan CN (2003) Resolution of inflammation: a new paradigm for the
pathogenesis of periodontal diseases. J Dent Res 82:82-90.

Van Pelt FN and Kenna JG (1994) Formation of trifluoroacetylated protein antigens in
cultm'ed rat hepatocytes exposed to halothane in vitro. Biochem Pharmacol 48:461-471.

Vannier E, Miller LC and Dinarello CA (1991) Histamine suppresses gene expression
and synthesis of tumor necrosis factor alpha via histamine H2 receptors. J Exp Med
174:281-284.

Vergani D, Mieli-Vergani G, Alberti A, Neuberger J, Eddleston AL, Davis M and
Williams R (1980) Antibodies to the surface of halothane-altered rabbit hepatocytes in
patients with severe halothane-associated hepatitis. N Eng] J Med 303:66-7 l .

Vial T, Goubier C, Bergeret A, Cabrera F, Evreux JC and Descotes J (199]) Side effects
of ranitidine. Drug Saf 6:94-1 17.

Visser M, Bouter LM, McQuillan GM, Wener MH and Harris TB (1999) Elevated C-
reactive protein levels in overweight and obese adults. JAMA 282:2]31-2135.

Visser M, Bouter LM, McQuillan GM, Wener MH and Harris TB (200]) Low-grade
systemic inflammation in overweight children. Pediatrics 107:E13.

Vollmar B, Ruttinger D, Wanner GA, Leiderer R and Menger MD (1996) Modulation of
kupffer cell activity by gadolinium chloride in endotoxemic rats. Shack 6:434—441 .

Wagner JG, Van Dyken SJ, Hotchkiss JA and Harkema JR (2001) Endotoxin
enhancement of ozone-induced mucous cell metaplasia is neutrophil-dependent in rat
nasal epithelium. Toxicol Sci 60:338-347.

Walker AM (1997) Quantitative studies of the risk of serious hepatic injury in persons
using nonsteroidal antiinflammatory drugs. Arthritis Rheum 40:201-208.

Whitcomb DC (2004) Inflammation and Cancer V. Chronic pancreatitis and pancreatic
cancer. Am J Physiol Gastrointest Liver Physiol 287 :G3 1 5-G3 l 9.

Willerson JT and Ridker PM (2004) Inflammation as a cardiovascular risk factor.
Circulation 109:112-10.

Wood CL, Gandolfi AJ and Van Dyke RA (1976) Lipid binding of a halothane
metabolite. Relationship to lipid peroxidation in vitro. Drug Metab Dispos 4:305-313.

222

Yamasawa H, Ishii Y and Kitamura S (1999) Cytokine-induced neutrophil
chemoattractant in a rat model of lipopolysaccharide-induced acute lung injury.
Inflammation 23:263-274.

Yamazoe K, Inaba T, Bonkobara M, Matsuki N, Ono K and Kudo T (1998) Changes of
hepatic tissue phospholipid peroxidation, malondialdehydes, and antioxidative enzyme
activities in dogs with halothane inhalation. J Vet Med Sci 60: 15-21 .

Yan SF, Lu J, Zou YS, Sch-Won J, Cohen DM, Buttrick PM, Cooper DR, Steinberg SF,
Mackman N, Pinsky DJ and Stern DM (1999) Hypoxia-associated induction of early
growth response-l gene expression. J Biol Chem 274: 15030-15040.

Yang SQ, Lin HZ, Lane MD, Clemens M and Diehl AM (1997) Obesity increases
sensitivity to endotoxin liver injury: implications for the pathogenesis of steatohepatitis.
Proc Nat] Acad Sci U S A 94:2557-2562.

Yasar U, Eliasson E, Forslund-Bergengren C, Tybring G, Gadd M, Sjoqvist F and Dahl
ML (2001) The role of CYP2C9 genotype in the metabolism of diclofenac in vivo and in
vitro. Eur J Clin Pharmacol 57:729-735.

Yee SB, Ganey PE and Roth RA (2003a) The role of Kupffer cells and TNF-alpha in
monocrotaline and bacteria] lipopolysaccharide-induced liver injury. T axica] Sci 71: 124-
132.

Yee SB, Hanumegowda UM, Copple BL, Shibuya M, Ganey PE and Roth RA (2003b)
Endothelial cell injmy and coagulation system activation during synergistic
hepatotoxicity from monocrotaline and bacterial lipopolysaccharide coexposure. Toxicol
Sci 74:203-214.

Yee SB, Hanumegowda UM, Hotchkiss JA, Ganey PE and Roth RA (2003c) Role of
neutrophils in the synergistic liver injury from monocrotaline and bacterial
lipopolysaccharide exposure. Toxicol Sci 72:43-56.

Yee SB, Harkema JR, Ganey PE and Roth RA (2003d) The coagulation system
contributes to synergistic liver injury from exposure to monocrotaline and bacterial
lipopolysaccharide. T axico] Sci 74:457-469.

Yee SB, Kinser S, Hill DA, Barton CC, Hotchkiss JA, Harkema JR, Ganey PE and Roth
RA (2000) Synergistic hepatotoxicity from coexposure to bacterial endotoxin and the
pyrrolizidine alkaloid monocrotaline. T oxico] Appl Pharmacol 166:173-185.

Yokoyama M, Yokoyama A, Mori S, Takahashi HK, Yoshino T, Watanabe T, Watanabe
T, Ohtsu H and Nishibori M (2004) Inducible histamine protects mice from P. acnes-
primed and LPS-induced hepatitis through H2-receptor stimulation. Gastroenterolagy
1272892-902.

223

Younes M, Heger B, Wilhehn KP and Siegers CP (1988) Enhanced in vivo-lipid
peroxidation associated with halothane hepatotoxicity in rats. Pharmacol T axico] 63:52-
56.

Zhang P, Xie M, Zagorski J and Spitzer JA (1995) Attenuation of hepatic neutrophil
sequestration by anti-CINC antibody in endotoxic rats. Shock 4:262-268.

Zhao B, Stavchansky SA, Bowden RA and Bowman PD (2003) Effect of interleukin-
lbeta and tumor necrosis factor-alpha on gene expression in human endothelial cells. Am
J Physiol Cell Physiol 284:C1577-ClS83.

Zhong H, Willard M and Sirnons J (2004) NS398 reduces hypoxia-inducible factor
(I-IIF)-lalpha and HIF-l activity: Multiple-level effects involving cyclooxygenase-2
dependent and independent mechanisms. Int J Cancer 1 12:585.

Zhou HR, Harkema JR, Hotchkiss JA, Yan D, Roth RA and Pestka JJ (2000)
Lipopolysaccharide and the trichothecene vorrritoxin (deoxynivalenol) synergistically
induce apoptosis in murine lymphoid organs. Toxicol Sci 53:253-263.

Zimmerman HJ (1978) Drug-induced liver disease. Drugs 16:25-45.
Zimmerman HJ (1993) Hepatotoxicity. Dis Man 39:675-787.

Zimmerman HJ, Jacob L, Bassan H, Gillespie J, Lukacs L and Abernathy CO (1986)
Effects of H2-blocking agents on hepatocytes in vitro: correlation with potential for
causing hepatic disease in patients. Proc Soc Exp Biol Med 182:511-514.

224