TOXICOLOGICAL SCIENCES 100(1), 267–280 (2007)

doi:10.1093/toxsci/kfm209

Advance Access publication August 13, 2007

The Role of Tumor Necrosis Factor Alpha in Lipopolysaccharide/Ranitidine-Induced Inflammatory Liver Injury

Francis F. Tukov,*,† James P. Luyendyk,*,† Patricia E. Ganey,*,† and Robert A. Roth*,†,1

*Center for Integrative Toxicology and †Department of Pharmacology and Toxicology, Michigan State University, East Lansing, Michigan 48824

Received March 20, 2007; accepted July 26, 2007

Exposure to a nontoxic dose of bacterial lipopolysaccharide

(LPS) increases the hepatotoxicity of the histamine-2 (H2) re-

ceptor antagonist, ranitidine (RAN). Because some of the

pathophysiologic effects associated with LPS are mediated

through the expression and release of inflammatory mediators

such as tumor necrosis factor alpha (TNF), this study was

designed to gain insights into the role of TNF in LPS/RAN

hepatotoxicity. To determine whether RAN affects LPS-induced

TNF release at a time near the onset of liver injury, male

Sprague-Dawley rats were treated with 2.5 3 106 endotoxin units

(EU)/kg LPS or its saline vehicle (iv) and 2 h later with either 30

mg/kg RAN or sterile phosphate-buffered saline vehicle (iv). LPS

administration caused an increase in circulating TNF concentra-

tion. RAN cotreatment enhanced the LPS-induced TNF increase

before the onset of hepatocellular injury, an effect that was not

produced by famotidine, a H2-receptor antagonist without

idiosyncrasy liability. Similar effects were observed for serum

interleukin (IL)-1beta, IL-6, and IL-10. To determine if TNF plays

a causal role in LPS/RAN-induced hepatotoxicity, rats were given

either pentoxifylline (PTX; 100 mg/kg, iv) to inhibit the synthesis

of TNF or etanercept (Etan; 8 mg/kg, sc) to impede the ability of

TNF to reach cellular receptors, and then they were treated with

LPS and RAN. Hepatocellular injury, the release of inflammatory

mediators, hepatic neutrophil (PMN) accumulation, and bio-

markers of coagulation and fibrinolysis were assessed. Pretreat-

ment with either PTX or Etan resulted in the attenuation of liver

injury and diminished circulating concentrations of TNF, IL-1b,IL-6, macrophage inflammatory protein-2, and coagulation/

fibrinolysis biomarkers in LPS/RAN-cotreated animals. Neither

PTX nor Etan pretreatments altered hepatic PMN accumulation.

These results suggest that TNF contributes to LPS/RAN-induced

liver injury by enhancing inflammatory cytokine production and

hemostasis.

Key Words: tumor necrosis factor alpha; inflammation; liver

injury; lipopolysaccharide; ranitidine; adverse drug reactions;

coagulation; hemostasis; hepatotoxicity.

Previous studies in rats showed that modest inflammation

triggered by bacterial lipopolysaccharide (LPS) decreases the

threshold for xenobiotic hepatotoxicity (reviewed in Roth

et al., 1997). One of the xenobiotics for which this is true is the

histamine-2 (H2) receptor antagonist, ranitidine (RAN; Luyendyk

et al., 2003). RAN has been associated with idiosyncratic

adverse drug reactions (IADRs) in people, with the liver as

a frequent target. Idiosyncratic hepatotoxicity occurs in less

than 0.1% of people taking RAN (Vial et al., 1991), and most

RAN-induced liver reactions are mild and reversible (Ribeiro

et al., 2000). In contrast, another H2-receptor antagonist,

famotidine (FAM), is not associated with idiosyncratic hepa-

totoxicity in humans: very few reports of FAM-associated

hepatotoxicity have been published (Ament et al., 1994;

Hashimoto et al., 1994; Jimenez-Saenz et al., 2000), and the

contribution of FAM to liver injury in these cases has been

questioned (Luyendyk et al., 2003).

Although it did not produce liver injury in healthy

experimental animals, RAN (30 mg/kg) was rendered hepato-

toxic in rats undergoing a mild inflammatory response induced

by LPS (Luyendyk et al., 2003). By contrast, a dose of FAM

that was equal in pharmacological efficacy (6 mg/kg) did not

interact with LPS to produce liver injury. LPS/RAN-cotreated

animals developed midzonal hepatocellular necrosis and a

liver-related clinical chemistry pattern resembling human cases

of RAN idiosyncrasy. Hepatocellular oncotic necrosis in LPS/

RAN-treated rats was preceded by hepatic neutrophil (PMN)

accumulation, and depletion of circulating PMNs attenuated

LPS/RAN-induced liver injury, suggesting a critical role for

these cells in this response (Luyendyk et al., 2005). In other

studies, activation of the coagulation system was shown to be

crucial to the pathogenesis of LPS/RAN hepatotoxicity

(Luyendyk et al., 2004b).

These results support the hypothesis that inflammation and

coagulation initiated by exposure to small amounts of LPS are

involved in the mechanism of liver injury from LPS/RAN

cotreatment. Evidence from both in vivo and in vitro studies

suggests that tumor necrosis factor alpha (TNF) is a critical and

proximal mediator of the inflammatory and hemostatic path-

ways stimulated by LPS (Aderka et al., 1992; Beutler and

Cerami, 1988; van der Poll et al., 1997). Furthermore, at a

1 To whom correspondence should be addressed at Department of

Pharmacology and Toxicology, National Food Safety and Toxicology Center,

221 Food Safety and Toxicology Building, Michigan State University, East

Lansing, MI 48824. Fax: (517) 432-2310. E-mail: rothr@msu.edu.

� The Author 2007. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved.For Permissions, please email: journals.permissions@oxfordjournals.org

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

relatively larger dose of LPS, liver injury is caused by complex

mechanisms involving the interaction of numerous soluble

mediators such as TNF and inflammatory cells, including

Kupffer cells (KCs), which are thought to be the major source

of TNF in liver (Ganey and Roth, 2001; Hewett et al., 1992,

1993). These findings raise the possibility that TNF contributes

to liver damage caused by cotreatment with LPS and RAN.

The purpose of this study was to characterize the expression

of inflammatory mediators during LPS/RAN interaction in vivoand in vitro and to determine the role of TNF in the ability of

LPS to render a nontoxic dose of RAN hepatotoxic. Toward

these ends, the effects of RAN and LPS on production of TNF

and other inflammatory factors in rats and in KC/hepatocyte

(KC/HPC) cocultures were determined and compared with

those elicited by FAM. Finally, by neutralizing TNF, the

hypothesis that TNF is a crucial mediator in LPS/RAN-induced

hepatotoxicity was tested.

MATERIALS AND METHODS

Materials. LPS derived from Escherichia coli serotype O55:B5 (Sigma

Chemical Co., St Louis, MO) with an activity of 13 3 106 endotoxin units

(EU)/mg was used for these studies. This activity was determined using

a limulus amebocyte lysate colorimetric end-point assay (kit 50-6480)

purchased from Cambrex Corp. (East Rutherford, NJ). Pentoxifylline (PTX),

and fetal bovine serum (FBS) were obtained from Sigma Chemical Co.

Etanercept (Etan; Enbrel; Immunex Corporation, Thousand Oaks, CA) was

obtained from the Michigan State University Clinical Center Pharmacy.

Antibiotic/antimycotic solution, gentamicin and RPMI 1640 (supplemented

with 2mM L-glutamine) were purchased from Gibco BRL (Rockville, MD).

The WEHI-13VAR mouse fibrosarcoma cell line was obtained from ATCC

(Manassas, VA). Unless otherwise stated, all other materials were purchased

from Sigma Chemical Co.

Animals. Male Sprague-Dawley rats (Crl:CD [SD]IGS BR; Charles River,

Portage, MI) weighing 250–350 g 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 before use. All procedures on animals followed the

guidelines for humane treatment set by the American Association of Laboratory

Animal Sciences and the University Laboratory Animal Research Unit at

Michigan State University.

Assessment of the effects of RAN on LPS-induced TNF release

in vivo. Rats fasted for 24 h were given 2.5 3 106 EU/kg LPS or its saline

vehicle, iv via a tail vein. Two hours later, 30 mg RAN/kg or sterile phosphate-

buffered saline (PBS) vehicle was administered through a tail vein. At 0, 2, 3, 6,

or 12 h after RAN administration, rats were anesthetized with sodium

pentobarbital (50 mg/kg, ip) and killed by exsanguination. For the collection of

plasma, blood was collected into vacutainer tubes (Becton Dickenson, Franklin

Lakes, NJ, USA) containing sodium citrate (final concentration, 3.8%).

Remaining blood was allowed to clot at room temperature, and serum was

collected and stored at � 80�C until use. TNF concentration was determined by

enzyme-linked immunosorbent assay (ELISA; see below). In a separate study

to examine inflammatory mediator release before the onset of injury, rats were

treated as above except that a group given 6 mg/kg FAM instead of RAN was

included, and the animals were killed 1, 2, or 3 h later. For these studies, RAN

and FAM were administered at doses that have equal pharmacological efficacy

in humans (Lin, 1991; Scarpignato et al., 1987) to simulate the ratio of doses

taken by people. As a result, the dose of RAN was fivefold greater than the dose

of FAM. We have shown previously that cotreatment of rats with LPS and

a dose of FAM equimolar or equally efficacious to the dose of RAN used in

these studies does not result in liver injury (Luyendyk et al., 2003).

Determination of PTX efficacy. PTX, a xanthine oxidase inhibitor, has

been used as a tool for exploring the role of TNF in xenobiotic-induced

hepatotoxicity (Sneed et al., 2000; Yee et al., 2003). PTX inhibits transcription

of TNF mRNA through a mechanism that involves phosphodiesterase

inhibition and the elevation of intracellular cyclic adenosine monophosphate

(Doherty et al., 1991; Schandene et al., 1992). Rats fasted for 24 h received

PTX (100 mg/kg) or its saline vehicle intravenously 1 h before treatment with

LPS (2.5 3 106 EU/kg) or its saline vehicle, iv. This treatment protocol for

PTX has been shown to prevent the LPS-induced rise in plasma TNF

concentration (Hewett et al., 1993). Two hours after LPS or vehicle exposure,

rats were anesthetized with sodium pentobarbital and killed by exsanguination.

Blood was allowed to clot at room temperature, and serum was collected and

stored at � 80�C until use. TNF concentration was determined using ELISA.

Determination of Etan efficacy. Rats fasted for 24 h received 8 mg Etan/kg

or its sterile water vehicle sc 1 h before LPS (2.5 3 106 EU/kg) or its saline

vehicle. Two hours later, they were anesthetized and bled as described above.

This Etan treatment has been shown to inactivate TNF activity (Geier et al.,

2003). A cytolytic cell assay (WEHI assay) was used to measure serum TNF

activity as described previously (Eskandari et al., 1990; Espevik and Nissen-

Meyer, 1986; Hewett et al., 1993; Sneed et al., 2000) with slight modifications.

Briefly, dilutions of serum samples from rats previously exposed to either veh/

LPS or Etan/LPS were assayed in triplicate for cytotoxic activity in WEHI-

13VAR mouse fibrosarcoma cells seeded at a density of 5 3 104 cells/well in

RPMI 1640 medium supplemented with 10% FBS, 1% antibiotic/antimycotic

solution, and 50 lg/ml gentamicin. The extent of cell death was measured with

3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide using a Bio-Tek

plate reader. Absorbance was read at 570 nm. A recombinant rat TNF (BD

Biosciences, San Diego, CA) was used as standard to calculate TNF activity in

the samples.

Assessment of LPS/RAN-induced hepatotoxicity in rats after PTX or Etan

pretreatment. Rats were fasted for 24 h before experiments. One hour before

administration of LPS (2.5 3 106 EU/kg) or its saline vehicle, rats received

either PTX (100 mg/kg) or its saline vehicle intravenously or Etan (8 mg/kg) or

its water vehicle sc. Two hours after LPS treatment, they were treated with

either 30 mg RAN/kg or its vehicle, and 6 h later they were anesthetized, bled,

and killed as described above. Blood was allowed to clot at room temperature,

and serum was collected by centrifugation and assayed for alanine

aminotransferase (ALT) activity using a diagnostic kit (ALT [GPT]; Ref:

TR7 111-125, Infinity; Thermo Electron, Melbourne, Australia).

Histopathology. Livers were fixed by immersion in 10% neutral-buffered

formalin for at least 3 days before being processed for histologic analysis.

Formalin-fixed liver samples from the left lateral liver lobe (3 samples/rat) were

embedded in paraffin, sectioned at 5 lm, stained for hematoxylin and eosin,

and examined by light microscopy for lesion size as described previously by

Luyendyk et al., (2003). All tissues were examined without knowledge of

treatment. The nature of lesions did not differ qualitatively among groups, and

each was assigned a score of 0 (no injury) to 6 based on increasing size of the

lesions examined at 1003 magnification.

Concentrations of cytokines, plasminogen activator inhibitor-1, and

thrombin-antithrombin dimers. Serum TNF concentration was measured

using a commercial kit (Rat TNF ELISA Kit II, BD OptEIA; BD Biosciences)

with recombinant rat TNF as standard. Measurements were performed in

duplicate. The concentrations of interleukin (IL)-1b, IL-6, IL-10, interferon-

gamma, and macrophage inflammatory protein-2 (MIP-2) in serum were

determined using enzyme immunoassay (EIA) kits obtained from BioSource

International, Inc. (Camarillo, CA). The assays were performed according to

the manufacturer’s instructions, and all samples were assayed in duplicate. The

concentration of plasminogen activator inhibitor-1 (PAI-1) in serum was

evaluated using an ELISA purchased from American Diagnostica, Inc.

268 TUKOV ET AL.

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

(Greenwich, CT). This ELISA measures total PAI-1 including active, inactive,

and tissue plasminogen activator–complexed forms. Plasma thrombin-anti-

thrombin (TAT) concentration was determined using an EIA kit (Enzygnost

TAT micro; #OWMG15) from Dade Behring Inc. (Deerfield, IL). Serum

concentrations of prostaglandin E2 (PGE2) were determined using a commer-

cially available EIA kit (PGE2 Express EIA Kit; Cayman Chemical, Ann Arbor,

MI). EIA was performed according to the manufacturer’s instructions, and all

samples were assayed in duplicate.

Detection of hepatic fibrin. Fibrin immunohistochemistry and quantifica-

tion were performed as described previously (Copple et al., 2002; Luyendyk

et al., 2004a). Eight micrometer thick sections of frozen liver were fixed in 10%

buffered formalin containing 2% acetic acid for 30 min at room temperature.

This fixation protocol solubilizes all fibrinogen and fibrin species except for

cross-linked fibrin. Sections were blocked with PBS containing 10% horse

serum (i.e., blocking solution; Vector Laboratories, Burlingame, CA) for

30 min, and this was followed by incubation overnight at 4�C with goat antirat

fibrinogen diluted (1:1000; ICN Pharmaceuticals, Aurora, OH) in blocking

solution. Next, sections were incubated for 3 h with donkey antigoat secondary

antibody conjugated to Alexa 594 (1:1000; Molecular Probes, Eugene, OR) in

blocking solution for 3 h. Sections were washed three times, 5 min each, with

PBS and visualized using an Olympus AX-80T fluorescence microscope

(Olympus, Lake Success, NY). Ten randomly chosen digital images (1003

magnification) were captured using a SPOT II camera and SPOT advanced

software (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. (2002). Results from 10 random fields analyzed per

liver section were averaged and counted as a replicate, that is, each replicate

represents a different rat.

Evaluation of hepatic PMN accumulation. PMN immunohistochemistry

was performed on formalin-fixed liver sections as described previously

(Luyendyk et al., 2004a; Yee et al., 2003). Briefly, the paraffin-embedded

liver tissue was cut into 6 lm thick slices (three serial liver sections per slide).

Paraffin was removed from the liver tissues with xylene before staining. PMNs

within liver sections were stained with a rabbit anti-PMN Ig isolated from

serum of rabbits immunized with rat PMNs as described by Hewett et al.

(1992). After incubation with the primary antibody, tissue sections were

incubated with biotinylated goat antirabbit IgG, avidin-conjugated alkaline

phosphatase, and Vector Red substrate to stain PMNs. The number of PMNs in

10–20 randomly selected, high-power fields (4003) was counted for each liver.

The average of these numbers was calculated, and hepatic PMN accumulation

is presented as the mean for each treatment group.

Isolation and culture of rat HPCs and KCs. HPCs and KCs were isolated

from the same rat liver as described by Tukov et al. (2006). After the isolation

of HPCs, cell viability was assessed by trypan blue exclusion. If cells were

greater than 80% viable, they were kept on ice for later use, at which time the

HPCs were plated in 24-well culture plates at a density of 2.5 3 105 cells/well.

Similarly, the viability of the isolated KCs was determined by trypan blue

exclusion and was usually > 90%. The cell concentration was adjusted to 1 3

106 viable cells/ml, and the KCs (1 3 106 cells/well) were plated in 24-well

plastic culture plates (Costar, Cambridge, MA) for 20 min at 37�C in

a humidified incubator (95% air/5% CO2). After this time, nonadherent cells

were removed by replacing the culture medium with fresh complete medium

(RPMI 1640 supplemented with 10% FBS, glutamine, 0.1% gentamicin, and

1% antibiotic/antimycotic solution). Isolated KCs were identified after staining

using ED1 antibody (mouse antirat CD68, Serotec; 1:500 in blocking serum),

and a secondary antibody coupled to fluorescein isothiocyanate (Alexa Fluor

594 goat anti-mouse IgM [microchain]; Molecular Probes) as described

previously (Tukov et al., 2006). These techniques revealed a purity of � 90%

KCs after differential plating. In experiments using cocultures, HPCs (2.5 3

105/well) were added after differential plating of KCs, and the plates were

returned to the incubator.

Exposure of KC/HPC cocultures and KC monocultures to test agents. After

the cells were plated as described above, culture medium in KC/HPC cocultures or

KC monocultures was replaced with fresh, complete RPMI medium and allowed to

incubate overnight at 37�C. The medium was then removed and replaced with fresh

medium prior to treatment. Cell cultures were treated with LPS, RAN, FAM,

vehicle (culture medium), or a combination of LPS/RAN or LPS/FAM and

returned to the incubator for an additional 6 h, after which medium from each well

was collected and centrifuged at 9000 3 g for 5 min. A portion of the medium was

used for the determination of activity of ALT, and the remainder was stored at �80�C for future use.

Assessment of HPC toxicity in vitro. ALT activity in culture supernatants

was assayed as a measure of HPC injury using a diagnostic kit (ALT (GPT);

Ref: TR7 111-125, Infinity; Thermo Electron). Enzyme leakage into the

medium was expressed as a percentage of total intracellular ALT content,

which was determined in vehicle-treated control cells lysed with Triton X.

Experiments were considered valid only if the ALT activity in the medium of

vehicle-treated cells was less than 25% of total activity. TNF and PGE2 in

culture medium were determined as described above for serum.

Statistical analysis. Results are presented as means ± SEM. For these

studies, if the variances were homogenous, data were analyzed by Student’s

t-test or one-way or two-way ANOVA as appropriate. For data sets with

nonhomogenous variances, Kruskal-Wallis nonparametric ANOVA was used

or data were log-transformed prior to analysis using Student’s t-test or

ANOVA. Data points identified as statistical outliers using Grubb’s test were

not included in statistical analyses. For ANOVA, group means were compared

using Student-Newman-Keul post hoc test. Histopathology scores were

compared using Mann-Whitney rank sum test. For studies in vitro, two-way

repeated measures ANOVA was applied after appropriate data transformation,

and group comparisons were made using the Student Newman Keuhl post hoc

test. The criterion for significance was p < 0.05 for all studies.

FIG. 1. Time-dependent effect of LPS/RAN cotreatment on serum TNF-aconcentration. Rats fasted for 24 h were given 2.5 3 106 EU/kg LPS or its saline

vehicle iv. Two hours after LPS, either 30 mg RAN/kg or its vehicle was

administered iv. At 2, 3, 6, or 12 h after RAN administration, rats were killed, and

serum TNF-a concentration was determined using ELISA. Data are expressed as

mean ± SEM. n ¼ 4–13 rats/group. a, Significantly different from LPS/Veh

group at the same time; b, significantly different from Veh/Veh group at the same

time; c, significantly different from Veh/RAN group at the same time.

ROLE OF TNF-a IN LPS/RAN-INDUCED LIVER INJURY 269

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

Hours after drug1 2 3

Hours after drug1 2 3

Hours after drug1 2 3

Hours after drug1 2 3

Hours after drug1 2 3

Hours after drug1 2 3

Hours after drug1 2 3

Hours after drug1 2 3

ALT

(U/L

)

0

100

200

300

400

500

a,b,c

a,b,c

a,b,c

a,b,ca,b,c

a,b,ca,b,c

a,b,c

a,b,c

a,b,c

a,b,c

a,b,c

a,b,ca,b,c

a,b,c

b,c b,c

a,b

a,ba,b

a,b

a,ba,b

a,b,c

a,cTN

F (p

g/m

l)

0

1000

2000

3000Veh/VehVeh/RANVeh/FAMLPS/VehLPS/RANLPS/FAM

a

aa

aa a

BAIL

-1 b

eta

(pg/

ml)

0

100

200

300

400

500

600

700

aa

a a

a

C

IL-6

(pg/

ml)

0

20000

40000

60000

80000

100000D

Veh/VehVeh/RANVeh/FAMLPS/VehLPS/RANLPS/FAM

Veh/VehVeh/RANVeh/FAMLPS/VehLPS/RANLPS/FAM

Veh/VehVeh/RANVeh/FAMLPS/VehLPS/RANLPS/FAM

Veh/VehVeh/RANVeh/FAMLPS/VehLPS/RANLPS/FAM

Veh/VehVeh/RANVeh/FAMVeh/LPSLPS/RANLPS/FAM

Veh/VehVeh/RANVeh/FAMLPS/VehLPS/RANLPS/FAM

Veh/VehVeh/RANVeh/FAMLPS/VehLPS/RANLPS/FAM

IFN

-gam

ma

(pg/

ml)

020

2000

4000

6000

8000

10000

12000

a

a,b

a

a

a

a a

G

PGE 2

(ng/

ml)

0

50

100

150

200

250

b

a a

H

F

PAI-1

(ng/

ml)

050

500

1000

1500

2000

2500

3000

aa

a a

a

a

IL-1

0 (p

g/m

l)

0

200

400

600

800

1000

1200

a aa a

a a

E

FIG. 2. Early liver injury and inflammatory mediator production after LPS/RAN cotreatment. Rats fasted for 24 h were given 2.5 3 106 EU/kg LPS or its

saline vehicle iv. Two hours after LPS, 30 mg RAN/kg, a pharmacologically equally efficacious dose of FAM (6 mg/kg) or its vehicle was administered, iv, and

rats were killed 1, 2, or 3 h later. Liver injury was assessed from the activity of ALT in serum (A). The concentrations of TNF-a (B), IL-1 beta (C), IL-6 (D), IL-10 (E),

PAI-1 (F), interferon-gamma (IFN) (G), and PGE2 (H) were determined as described in ‘‘Materials and Methods’’ section. Data are expressed as mean ± SEM.

n ¼ 4–9 rats/group. a, Significantly different from respective group in the absence of LPS; b, significantly different from respective group in the absence of drug;

c, significantly different from respective group in the presence of FAM.

270 TUKOV ET AL.

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

RESULTS

Time-Course of Serum TNF Concentration in LPS/RAN-Cotreated Rats

At the time of RAN administration, the concentrations of

TNF in serum of rats treated 2 h earlier with LPS or saline were

32 ± 1 and 0.14 ± 0.05 ng/ml, respectively. This large LPS-

induced increase in TNF is consistent with previous reports

(Hewett et al., 1993). Neither vehicle nor RAN alone caused

a significant increase in serum TNF concentration over the 12-h

examination period (Fig. 1). Serum TNF concentration was

greater at all times in LPS-treated rats compared with vehicle-

treated rats. RAN cotreatment maintained plasma TNF

concentration at significantly greater values compared with

rats given LPS alone (Fig. 1). Serum TNF concentrations were

approximately twofold greater in LPS/RAN-cotreated animals

compared with LPS/veh-treated animals at 2 and 3 h after RAN

administration.

Early Inflammatory Mediator Release and Comparisonto FAM

Results from the experiment above suggested that RAN

enhanced the LPS-induced elevation of serum TNF around the

time that hepatocellular injury began to occur (i.e., 2–3 h

[Luyendyk et al., 2003, 2004b]). An additional experiment was

designed to (1) explore the effect of RAN on LPS-induced

TNF release before the onset of hepatocellular injury,

(2) compare the TNF-enhancing effects of RAN to those of

FAM, which does not interact with LPS to produce liver injury

in rats and does not have IADR liability in humans (Luyendyk

et al., 2006), and (3) elucidate the effects of RAN on other

inflammatory mediators. Rats were treated with a nontoxic

dose of LPS or its vehicle followed 2 h later by a non-

hepatotoxic dose of RAN, its vehicle, or a dose of FAM that is

pharmacologically equally efficacious to that of RAN (Luyendyk

et al., 2003). Times were chosen to encompass those before

(1 h) and during (2–3 h) the onset of hepatocellular injury. Of

the six treatments, only LPS/RAN was hepatotoxic, confirm-

ing previous results (Luyendyk et al., 2003, 2006). In this

group, serum ALT activity was unchanged 1 h after RAN

treatment, slightly elevated at 2 h, and more markedly elevated

by 3 h (Fig. 2A). Serum TNF concentration was not affected

by either RAN or FAM given alone (Fig. 2B). LPS exposure

elevated TNF concentration relative to control at all three of

these early times; RAN cotreatment enhanced this LPS effect

at all times, whereas FAM cotreatment did not influence it

(Fig. 2B). Similar effects of LPS-drug treatment were ob-

served for serum IL-1b, interferon-gamma, IL-6, IL-10, and

PAI-1 (Figs. 2C–G).

PGE2 is a product of cyclooxygenase-2 (COX-2), which is

expressed by cells during inflammatory responses. A previous

study indicated that COX-2 mRNA was upregulated in livers

of LPS/RAN-cotreated rats (Luyendyk et al., 2006). Further-

more, it has been reported that TNF increases production of

PGE2 and that PGE2 diminishes LPS-stimulated production of

TNF (Kreydiyyeh et al., 2007; Scales et al., 1989). Serum

PGE2 concentration was elevated slightly by either RAN or

FAM as well as by LPS (Fig. 2H). In LPS-treated rats,

cotreatment with either RAN or FAM enhanced serum PGE2

concentration compared to LPS treatment alone at all three

times.

Effect of PTX and Etan on LPS/RAN-Induced Liver Injury

In a preliminary efficacy study, the concentration of

circulating TNF was determined 2 h after administration of

LPS to rats pretreated with PTX or Etan. The LPS-induced

increase in TNF was markedly attenuated by pretreatment with

PTX (Fig. 3A). Similarly, Etan pretreatment significantly

reduced the increase in circulating TNF activity (Fig. 3B).

FIG. 3. Efficacy of PTX (A) and Etan (B) as inhibitors of TNF-a. Rats

fasted for 24 h received either PTX (100 mg/kg, iv) or its vehicle or Etan (8 mg/

kg, sc) or its vehicle 1 h before treatment with LPS (2.5 3 106 EU/kg, iv). Two

hours after LPS, rats were killed, and serum was collected. TNF-aconcentration was determined as described in ‘‘Materials and Methods’’

section. Data are expressed as mean ± SEM; n ¼ 4 rats/group. a, Significantly

different from the Veh/LPS group.

ROLE OF TNF-a IN LPS/RAN-INDUCED LIVER INJURY 271

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

LPS/RAN treatment increased serum ALT activity at 6 h

(Figs. 4A and 4B). Serum ALT activity was significantly less

in animals given either PTX (Fig. 4A) or Etan (Fig. 4B) prior to

LPS/RAN treatment. In preliminary studies, ALT activity was

not altered in Etan/veh/veh-treated animals compared with veh/

veh/veh-treated rats (data not shown). Similarly, we demon-

strated previously that the dose of PTX used for these studies

was not hepatotoxic in rats (Hewett et al., 1993). These

findings were supported by histologic examination of liver

sections. Liver from veh/veh/veh-treated animals appeared

normal (Fig. 5A). Sections from LPS/RAN-cotreated rats taken

6 h after administration of RAN had acute, multifocal,

midzonal hepatocellular necrosis (Fig. 5B). These lesions have

been described in detail previously (Luyendyk et al., 2003).

In livers from animals that received PTX (Fig. 5C) or Etan

(Fig. 5D) prior to LPS/RAN cotreatment, the size and

frequency of the necrotic foci were significantly reduced

compared with the LPS/RAN-treated rats.

Effect of PTX or Etan on LPS/RAN-Induced Increasesin Serum IL-1b and IL-6 Concentrations

Cotreatment of rats with LPS/RAN significantly increased

circulating concentrations of IL-1b and IL-6 compared with

vehicle-treated rats (Fig. 6). Pretreatment with PTX reduced the

LPS/RAN-induced increase in IL-1b and IL-6 (Figs. 6A and

6B). Etan reduced the LPS/RAN-induced increase in circulat-

ing IL-1b (Fig. 6C) but did not significantly affect IL-6 con-

centration, although there was a trend toward reduced

concentration (Fig. 6D).

Effect of PTX or Etan on LPS/RAN-Induced Changesin the Hemostatic System

Increased TAT concentration in plasma is a biomarker of

coagulation system activation. PAI-1 is the major endogenous

inhibitor of fibrinolysis, thus an increase in its concentration in

plasma suggests impaired fibrinolysis. Previous studies dem-

onstrated that both TAT and PAI-1 concentrations were

enhanced early after LPS/RAN treatment (Fig. 2F; Luyendyk

et al., 2004a,b). Confirming these results, LPS/RAN treatment

increased plasma TAT concentration and total PAI-1 concen-

tration in serum (Fig. 7). These increases were significantly

attenuated by either PTX or Etan pretreatment (Fig. 7).

FIG. 4. Effect of PTX (A) or Etan (B) on the LPS/RAN-induced elevation

in serum ALT activity. Rats fasted for 24 h received either PTX (100 mg/kg, iv)

or its vehicle or Etan (8 mg/kg, sc) or its vehicle 1 h before treatment with LPS

(2.5 3 106 EU/kg, iv) or its vehicle. Two hours after LPS, either RAN (30 mg/kg,

iv) or its vehicle was administered. Hepatocellular injury was assessed 6 h after

RAN as increased serum ALT activity. Data are expressed as mean ± SEM; n ¼4–6 rats/group; a, significantly different from the veh/veh/veh group; b,

significantly different from the veh/LPS/RAN group.

FIG. 5. Representative photomicrographs of livers after LPS/RAN

cotreatment. Rats were treated as described in the legend to Fig. 4. At 6 h

after RAN administration, rats were killed and livers were removed and fixed in

10% neutral-buffered formalin, stained with H&E, and examined by light

microscopy. (A) veh/veh/veh, (B) veh/LPS/RAN, (C) PTX/LPS/RAN, and

(D) Etan/LPS/RAN-treated rats. The photomicrographs were taken at 1003

magnification. Acute, multifocal, midzonal hepatocellular necrosis (arrows)

developed in LPS/RAN-cotreated rats. Reduction in the size and frequency of

necrotic lesions was observed in PTX- and Etan-pretreated rats.

272 TUKOV ET AL.

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

Liver sections from veh/veh/veh-treated animals showed

very little fibrin deposition, with some staining of the larger ves-

sels due to fibrin deposition during animal sacrifice (Luyendyk

et al., 2003) (Fig. 8A). Fibrin deposition in liver tissues from

veh/LPS/RAN-treated rats (Fig. 8B) was significantly in-

creased compared with control. Pretreatment with either PTX

(Fig. 8C) or Etan (Fig. 8D) attenuated the increase in fibrin

deposition caused by veh/LPS/RAN treatment. Morphometric

analysis confirmed this result (Figs. 8E and 8F).

Effect of PTX or Etan on LPS/RAN-Induced Hepatic PMNAccumulation and Serum MIP-2 Concentration

Previous results suggested that PMNs play a causal role in

LPS/RAN-induced hepatocellular injury (Luyendyk et al.,2005). To test the possibility that TNF mediates liver injury

after LPS/RAN exposure by promoting PMN accumulation in

the liver, the effects of PTX or Etan pretreatment on the LPS/

RAN-induced increase in liver PMNs were evaluated.

Quantitative analysis of liver sections immunostained for

PMNs revealed that PMN accumulation occurred within 6 h

in livers from animals treated with LPS/RAN. PMN numbers

in the livers of LPS/RAN-treated animals were unaffected by

either PTX or Etan (Figs. 9A and 9C). PMN distribution was

panlobular in livers of LPS/RAN-cotreated rats, although foci

of PMNs were noted in lesioned areas.

MIP-2 is a chemokine that can cause transmigration of

PMNs from vessels and PMN activation. The serum concentra-

tion of MIP-2 was significantly increased in rats cotreated with

LPS and RAN. This increase was attenuated by pretreatment

with either PTX or Etan (Figs. 9B and 9D).

Mediator Production In Vitro

KCs are major producers of the inflammatory mediators

altered by LPS/RAN cotreatment in vivo. Accordingly, release

of mediators was evaluated in vitro in KC/HPC cocultures at

noncytotoxic drug and LPS concentrations. Concentrations of

RAN were chosen to approximate the maximum concentration

likely to occur in plasma in the study in vivo above (i.e., approx-

imately 100lM) and one larger, but noncytotoxic concentration

(i.e., 500lM). Neither LPS, RAN, FAM nor combinations of

LPS and the drugs caused release of ALT in KC/HPC

cocultures within 6 h of incubation (Figs. 10A and 10B). LPS

increased TNF release (Figs. 10C and 10D). RAN enhanced

slightly the LPS-induced TNF release, but only at the larger

RAN concentration, whereas both FAM concentrations led to

a modest enhancement. Unlike the results in KC/HPC

cocultures, RAN was without effect on LPS-stimulated

TNF release when it was added to KCs cultured by themselves

(Fig. 11A), suggesting that the presence of HPCs modified

the ability of RAN to alter this response. At the larger

FIG. 6. Effect of TNF-a inhibition on proinflammatory cytokine release after LPS/RAN cotreatment. Rats were treated as described in the legend to Fig. 4. At

6 h after RAN administration, rats were killed, and serum IL-1b (A and C) and IL-6 (B and D) concentrations were determined by ELISA. Data are expressed as

mean ± SEM; n ¼ 5–6 rats/group; a, Significantly different from the veh/veh/veh group; b, significantly different from the veh/LPS/RAN group.

ROLE OF TNF-a IN LPS/RAN-INDUCED LIVER INJURY 273

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

concentration, FAM by itself elicited TNF release in KC

cultures (Fig. 11B).

MIP-2, the release of which was selectively enhanced by

RAN in LPS-cotreated rats (Luyendyk et al., 2006), was un-

affected by RAN or FAM cotreatment in KC/HPC cocultures

(Figs. 12A and 12B). Interestingly, PGE2 release in vitro was

largely unaffected by LPS but enhanced markedly by RAN

(Fig. 12C). In contrast, FAM caused only a slight enhancement

(Fig. 12D).

DISCUSSION

TNF is an important proinflammatory cytokine involved in

normal physiological immune and inflammatory processes.

A causative role for TNF in promoting liver injury has been

identified for inflammation interaction with a number of

xenobiotic agents (Barton et al., 2001; Endo et al., 1999;

Yee et al., 2002) as well as for large, hepatotoxic doses of LPS

(Hewett et al., 1993). The present work explored the role of

TNF in an animal model of RAN-induced idiosyncratic liver

injury. LPS exposure induces a large and rapid increase in

plasma TNF concentration (Hewett et al., 1993). In the current

study, despite increased plasma TNF concentration, rats given

a small dose of LPS did not develop liver injury by 3 h. This

indicates that an increase in TNF of this magnitude and

duration is not sufficient to cause hepatocellular damage.

Despite this insufficiency, this large increase in TNF con-

centration could be critical for the genesis of hepatocellular

injury when combined with other, independent effects of RAN

cotreatment. Indeed, RAN-cotreated rats sustained greater

plasma TNF levels compared with rats given LPS alone and

developed hepatotoxicity. In contrast, FAM at an equally

efficacious dose neither enhanced TNF production nor

synergized with LPS to produce liver injury. Thus, the increase

in LPS-stimulated TNF production distinguished a drug that

causes human IADRs from one that does not. This also raises

the possibility that the effect of RAN on LPS-induced TNF

production could contribute to hepatotoxicity in LPS/RAN-

treated rats.

To determine the role of TNF in LPS/RAN-induced

hepatotoxicity, we used two approaches. PTX is a methylxanthine

that inhibits the synthesis of TNF (Barton et al., 2001; Dezube

et al., 1993; Yee et al., 2003). However, PTX also has several

other pharmacological effects (Banfi et al., 2004). Accordingly,

we also used Etan, a dimeric fusion protein that contains a soluble

TNF receptor capable of selectively neutralizing TNF. PTX and

Etan significantly reduced serum TNF concentration and activity,

respectively, and were both effective in reducing hepatocellular

FIG. 7. Effect of PTX or Etan on plasma TAT and serum PAI-1 concentrations after LPS/RAN cotreatment. Rats were treated as described in the legend to

Fig. 4. At 6 h after RAN administration, they were killed, and plasma TAT (A and C) and serum PAI-1 (B and D) concentrations were determined by ELISA. Data

are expressed as mean ± SEM; n ¼ 5–6 rats/group; a, significantly different from the veh/veh/veh group; b, significantly different from the veh/LPS/RAN group.

274 TUKOV ET AL.

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

injury in LPS/RAN-cotreated rats. These results indicate that

TNF is involved in LPS/RAN-induced liver injury.

Although this enhancement of TNF concentration by RAN

was small compared with the pronounced increase caused by

LPS exposure, the prolongation of TNF levels above

a particular threshold could have pathogenic importance since

it occurred prior to injury in LPS/RAN-treated rats. The

importance of the RAN-induced potentiation of LPS-induced

TNF release is difficult to test. We have shown previously that

LPS given alone at the dose used in these studies does not

cause liver injury up to 24 h (Luyendyk et al., 2003).

Accordingly, LPS-induced TNF is insufficient to cause liver

damage. TNF induces the expression of several cytokines

including IL-1b, IL-6, IL-8, MIP-2, as well as other gene

products capable of damaging the liver directly or in concert

with TNF (Aggarwal and Natarajan, 1996; Locksley et al.,2001). For example, IL-1b and IL-6 display partially over-

lapping activities with TNF (Eigler et al., 1997). TNF and

FIG. 8. Effects of PTX or Etan on hepatic fibrin deposition after LPS/RAN cotreatment. Rats were treated as described in the legend to Fig. 4. Livers were

removed 6 h after RAN administration and processed for fibrin immunohistochemistry as described in ‘‘Materials and Methods’’ section. Fibrin appears as white on

a gray background. Representative liver section from (A) a veh/veh/veh-treated rat showing minimal fibrin staining, (B) a veh/LPS/RAN-treated rat with sinusoidal

fibrin deposits in both periportal and centrilobular regions and (C) a PTX/LPS/RAN-treated animal, and (D) an Etan/LPS/RAN-treated animals showing reductions in

fibrin staining compared with those from veh/LPS/RAN-treated rats. Fibrin deposition was quantified as described in ‘‘Materials and Methods’’ section. (E and F).

Data are expressed as mean ± SEM; n¼ 5–6 rats/group. a, significantly different from the veh/veh/veh group; b, significantly different from the veh/LPS/RAN group.

ROLE OF TNF-a IN LPS/RAN-INDUCED LIVER INJURY 275

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

IL-1b activate PMNs (Klebanoff et al., 1986), induce their own

production, stimulate production of other inflammatory medi-

ators, and antagonize anti-inflammatory cytokines such as

IL-10 (Ji et al., 2004). In the present study, LPS induction of

IL-1b and IL-6 was enhanced by RAN cotreatment but not by

cotreatment with FAM. The LPS/RAN-induced increase in

these cytokines was inhibited by TNF neutralization, suggest-

ing that the RAN-dependent increase in TNF is required for

enhanced expression of inflammatory genes in LPS/RAN-

treated rats (Figs. 2C and 2D and 6B and 6C).

In addition to enhancing cytokine expression, TNF–

dependent gene expression might contribute to LPS/RAN-

induced liver injury by enhancing coagulation. RAN cotreatment

enhanced LPS-induced coagulation prior to liver injury, and

anticoagulants reduced liver damage in LPS/RAN-treated rats

(Luyendyk et al., 2004a). TNF and other cytokines, including

IL-1b and IL-6, induce tissue factor expression on endothelial

cells and monocytes/macrophages in vitro (Neumann et al.,1997). Tissue factor is generally accepted as the pivotal

initiator of coagulation system activation in endotoxemia and

bacteremia (de Jonge et al., 2000; Levi et al., 1994; Pawlinski

et al., 2004; Taylor et al., 1991). In addition, in the presence of

PMNs, TNF causes sinusoidal endothelial cell (SEC) damage

in vitro (Smedly et al., 1986; Takei et al., 1995), which can

activate the coagulation system (Colman et al., 1994). Indeed,

PTX and Etan inhibited coagulation activation in LPS/RAN-

treated rats, as marked by reduced plasma TAT concentration

(Fig. 7). This result indicates that TNF contributes to LPS/

RAN-induced liver injury at least in part by enhancing

activation of the coagulation system.

Activation of the coagulation system in LPS/RAN-treated rats

results in the formation of fibrin clots in liver sinusoids, and

prevention of fibrin deposition was associated with reduced

hepatocellular injury (Luyendyk et al., 2004a). Neutralization of

TNF reduced LPS/RAN-induced liver fibrin deposition (Fig. 8).

This reduction could be a consequence of reduced coagulation

and/or increased fibrinolysis. TNF has been shown to play a role

in the induction of PAI-1 in the liver and plasma of LPS-treated

mice (Fearns and Loskutoff, 1997). Like TNF, the concentration

of PAI-1 in the plasma was enhanced by RAN cotreatment

(Fig. 2F, Luyendyk et al., 2006), and TNF neutralization

inhibited the induction of PAI-1 in LPS/RAN-treated rats

(Figs. 7B and 7D). The cellular source and mechanism of

enhanced PAI-1 expression in livers of LPS/RAN-cotreated rats

is not known. TNF and IL-1 can stimulate the expression and

release of PAI-1 by endothelial cells (Schleef et al., 1988).

Indeed, RAN cotreatment enhanced the LPS-induced alteration

of SEC function (Luyendyk et al., 2004a and b). Taken together,

these results suggest that TNF enhancement of PAI-1 production

promotes fibrin deposition and injury in this animal model.

FIG. 9. Effect of PTX or Etan on serum MIP-2 concentration and the accumulation of PMNs in liver after LPS/RAN cotreatment. Rats were treated as

described in the legend to Fig. 4. Serum MIP-2 concentration (B and D) was evaluated 6 h after RAN treatment. Liver PMN accumulation (A and C) was evaluated

in 10–20 randomly selected 4003 fields 6 h after RAN administration as described under ‘‘Materials and Methods’’ section. Data are expressed as mean ± SEM;

n ¼ 5–6 rats/group. a, Significantly different from the veh/veh/veh group. b, Significantly different from the veh/LPS/RAN group.

276 TUKOV ET AL.

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

Coagulation-dependent liver damage in LPS/RAN-cotreated

rats could activate inflammatory cells such as PMNs. Depletion

of circulating PMNs attenuated LPS/RAN-induced liver injury,

suggesting their importance in the pathogenesis (Luyendyk

et al., 2005). In models of PMN-dependent liver injury, PMNs

not only accumulate in liver sinusoids but must also trans-

migrate through sinusoidal endothelium and become activated

in close proximity to HPCs to cause injury to these cells

(Jaeschke, 2006). Interestingly, anticoagulation did not inhibit

PMN accumulation in LPS/RAN-treated rats, but it did reduce

plasma concentration of the PMN chemokine, MIP-2

(Luyendyk et al., 2006), which can activate PMNs and elicit

the release of toxic products (Biedermann et al., 2000; Wang

and Thorlacius, 2005). TNF can prompt the accumulation of

PMNs in tissues by activating endothelial cells (Bradham et al.,1998; Vassalli, 1992) and can also prime PMNs for activation

FIG. 10. Release of ALT and TNF-a from HPC/KC cocultures exposed to LPS and RAN or FAM. HPCs and KCs were isolated and cocultured as described

in ‘‘Materials and Method’’ section. LPS (10 ng/ml) and drug at the indicated concentrations were added at the same time, and the medium was collected after 6 h.

ALT activity and TNF-a concentration were measured in the cell-free supernatant. ALT activity (A and B) in the medium is expressed as a percent of the total ALT

activity. TNF-a concentration (C and D) was determined by ELISA. n ¼ 4–5 isolations from different rats. a Significantly different from vehicle-treated group at

the same concentration of drug; b, significantly different from corresponding vehicle-treated group in the absence of drug.

FIG. 11. Release of TNF-a from KCs exposed to LPS and RAN or FAM. KCs were isolated and cultured as described in ‘‘Materials and Method’’ section.

LPS (10 ng/ml) and drug at the indicated concentrations were added at the same time, and medium was collected after 6 h. TNF-a concentration was measured in

the cell-free supernatant. n ¼ 3–4 isolations from different rats. a, Significantly different from corresponding vehicle-treated group at the same concentration of

drug; b significantly different from vehicle-treated group in the absence of drug.

ROLE OF TNF-a IN LPS/RAN-INDUCED LIVER INJURY 277

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

(Kushimoto et al., 1996; Nagaki et al., 1991; Vassalli, 1992).

Analogous to the protection afforded by anticoagulants in this

model, inhibition of TNF did not impact hepatic PMN

accumulation but did reduce serum MIP-2 concentration

(Fig. 9). The observations that TNF inhibition reduced the

concentration of MIP-2 in plasma and attenuated coagulation

system activation and that inhibition of coagulation reduced

MIP-2 concentration (Luyendyk et al., 2006) are consistent

with the hypothesis that TNF activation of coagulation induces

the expression of MIP-2, which could activate PMNs

accumulated in the liver.

Recently, we reported that inflammatory mediator expres-

sion elicited by xenobiotic/LPS coexposure in rats could be

recapitulated using a KC/HPC coculture system (Tukov et al.,2006). In the present study, RAN enhanced TNF release in the

presence of LPS in KC/HPC cocultures but only at a large

RAN concentration that is unlikely to be achieved in the LPS/

RAN animal model (Fig. 10). FAM had a similar effect. Thus,

this in vitro cytokine system failed to distinguish a drug that

causes IADRs in humans from one that does not. The MIP-2

response in vitro also differed from the response in rats

(Luyendyk et al., 2006, and Fig. 12). The disparity between

these in vitro and in vivo responses suggests that the RAN-

induced enhancement of TNF and MIP-2 production seen

in vivo likely occurs by indirect mechanisms, perhaps involving

other cell types.

LPS/RAN treatment of rats selectively increased COX-2

mRNA expression, whereas LPS/FAM treatment did not

(Luyendyk et al., 2006). Interestingly, treatment with RAN

produced a similar increase in PGE2 compared with treatment

FIG. 12. Release of MIP-2 and PGE2 from HPC/KC cocultures exposed to LPS and RAN or FAM. HPCs and KCs were isolated and cocultured as described

in ‘‘Materials and Method’’ section. LPS (10 ng/ml) and drug at the indicated concentrations were added at the same time, medium was collected after 6 h, and

MIP-2 (A and B) and PGE2 (C and D) were measured in the cell-free supernatant. n ¼ 3–5 isolations from different rats. a, Significantly different from vehicle-

treated group at the same concentration of drug; b, significantly different from corresponding vehicle-treated group in the absence of drug.

FIG. 13. Working hypothesis for the pathogenesis of LPS/RAN-induced hepatotoxicity. Previous studies (see ‘‘Discussion’’ and Luyendyk et al. 2004a and b;

2005) have implicated the hemostatic system and neutrophils (PMNs) as critical players in the hepatotoxic interaction between LPS and RAN. The present results

suggest that TNF-a is important as a proximal mediator in the chain of events leading to hepatocellular injury.

278 TUKOV ET AL.

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

with FAM, irrespective of LPS cotreatment (Fig. 2H). In KC/

HPC cocultures, RAN prompted the release of PGE2, whereas

FAM had little effect (Figs. 12C and 12D). These drug effects

were not influenced by LPS treatment of the cells. Thus,

increased PGE2 release did distinguish the two drugs in vitro,

whereas it failed to do so in vivo.

In summary, RAN cotreatment potentiated the LPS-medi-

ated production of TNF at a time before the onset of liver injury

in LPS/RAN-cotreated rats. Inhibition of TNF biosynthesis or

signaling significantly attenuated LPS/RAN-induced hepato-

cellular injury. Interference with TNF also reduced the serum

concentrations of other proinflammatory cytokines and chemo-

kines in LPS/RAN-treated rats. In addition, TNF inhibition

decreased coagulation system activation, PAI-1 production,

and fibrin deposition. Together, these results suggest that TNF

contributes to LPS/RAN-induced liver injury by activating the

hemostatic system, inducing inflammatory cytokines, and

possibly by contributing to PMN activation (Fig. 13).

FUNDING

Grant DK061315 from the National Institutes of Health;

Training Grant 2 T32 ES07255 from the National Institute of

Environmental Health Sciences to F.F.T. and J.P.L.; the

Society of Toxicology’s Colgate Palmolive Postdoctoral

Fellowship in In Vitro Toxicology to F.F.T.; and Society of

Toxicology’s Novartis Predoctoral Fellowship to J.P.L.

ACKNOWLEDGMENTS

The authors are grateful to Theresa Eagle and Sandra Newport for technical

assistance.

REFERENCES

Aderka, D., Engelmann, H., Wysenbeek, A. J., and Levo, Y. (1992). The

possible role of tumor necrosis factor (TNF) and its natural inhibitors, the

soluble-TNF receptors, in autoimmune diseases. Isr. J. Med. Sci. 28,

126–130.

Aggarwal, B. B., and Natarajan, K. (1996). Tumor necrosis factors: Develop-

ments during the last decade. Eur. Cytokine Netw. 7, 93–124.

Ament, P. W., Roth, J. D., and Fox, C. J. (1994). Famotidine-induced mixed

hepatocellular jaundice. Ann. Pharmacother. 28, 40–42.

Banfi, C., Sironi, L., De Simoni, G., Gelosa, P., Barcella, S., Perego, C.,

Gianazza, E., Guerrini, U., Tremoli, E., and Mussoni, L. (2004). Pentoxifyl-

line prevents spontaneous brain ischemia in stroke-prone rats. J. Pharmacol.

Exp. Ther. 310, 890–895.

Barton, C. C., Barton, E. X., Ganey, P. E., Kunkel, S. L., and Roth, R. A.

(2001). Bacterial lipopolysaccharide enhances aflatoxin B1 hepatotoxicity in

rats by a mechanism that depends on tumor necrosis factor alpha.

Hepatology 33, 66–73.

Beutler, B., and Cerami, A. (1988). Tumor necrosis, cachexia, shock, and

inflammation: A common mediator. Annu. Rev. Biochem. 57, 505–518.

Biedermann, T., Kneilling, M., Mailhammer, R., Maier, K., Sander, C. A.,

Kollias, G., Kunkel, S. L., Hultner, L., and Rocken, M. (2000). Mast cells

control neutrophil recruitment during T cell-mediated delayed-type hyper-

sensitivity reactions through tumor necrosis factor and macrophage inflam-

matory protein 2. J. Exp. Med. 192, 1441–1452.

Bradham, C. A., Plumpe, J., Manns, M. P., Brenner, D. A., and Trautwein, C.

(1998). Mechanisms of hepatic toxicity. I. TNF-induced liver injury. Am.

J. Physiol. 275, G387–G392.

Colman, R. W., Marder, V. J., Salzman, E. W., and Hirsh, J. (1994). Overview

of hemostasis. In Hemostasis and Thrombosis (R. W. Colman, J. Hirsh,

V. J. Marder, and E. W. Salzman, Eds.), p. 3. Lippincott, Philadelphia.

Copple, B. L., Banes, A., Ganey, P. E., and Roth, R. A. (2002). Endothelial cell

injury and fibrin deposition in rat liver after monocrotaline exposure.

Toxicol. Sci. 65, 309–318.

de Jonge, E., Dekkers, P. E. P., Creasey, A. A., Hack, C. E., Paulson, S. K.,

Karim, A., Kesecioglu, J., Levi, M., van Deventer, S. J. H., and van der

Poll, T. (2000). Tissue factor pathway inhibitor dose-dependently inhibits

coagulation activation without influencing the fibrinolytic and cytokine

response during human endotoxemia. Blood 95, 1124–1129.

Dezube, B. J., Sherman, M. L., Fridovich-Keil, J. L., Allen-Ryan, J., and

Pardee, A. B. (1993). Down-regulation of tumor necrosis factor expression

by pentoxifylline in cancer patients: A pilot study. Cancer Immunol.

Immunother. 36, 57–60.

Doherty, G. M., Jensen, J. C., Alexander, H. R., Buresh, C. M., and

Norton, J. A. (1991). Pentoxifylline suppression of tumor necrosis factor

gene transcription. Surgery 110, 192–198.

Eigler, A., Sinha, B., Hartmann, G., and Endres, S. (1997). Taming TNF:

Strategies to restrain this proinflammatory cytokine. Immunol. Today 18,

487–492.

Endo, Y., Shibazaki, M., Yamaguchi, K., Kai, K., Sugawara, S., Takada, H.,

Kikuchi, H., and Kumagai, K. (1999). Enhancement by galactosamine of

lipopolysaccharide (LPS)-induced tumour necrosis factor production and

lethality: Its suppression by LPS pretreatment. Br. J. Pharmacol. 128, 5–12.

Eskandari, M. K., Nguyen, D. T., Kunkel, S. L., and Remick, D. G. (1990).

WEHI 164 subclone 13 assay for TNF: Sensitivity, specificity, and

reliability. Immunol. Investig. 19, 69–79.

Espevik, T., and Nissen-Meyer, J. (1986). A highly sensitive cell line, WEHI

164 clone 13, for measuring cytotoxic factor/tumor necrosis factor from

human monocytes. J. Immunol. Methods 95, 99–105.

Fearns, C., and Loskutoff, D. J. (1997). Induction of plasminogen activator

inhibitor type 1 gene expression in murine liver by lipopolysaccharide.

Cellular localization and role of endogenous tumor necrosis factor-a. Am.

J. Pathol. 150, 579–590.

Ganey, P. E., and Roth, R. A. (2001). Concurrent inflammation as a determinant

of susceptibility to toxicity from xenobiotic agents. Toxicology 169, 195–208.

Geier, A., Dietrich, C. G., Voigt, S., Kim, S. K., Gerloff, T., Kullak-

Ublick, G. A., Lorenzen, J., Matern, S., and Gartung, C. (2003). Effects of

proinflammatory cytokines on rat organic anion transporters during toxic

liver injury and cholestasis. Hepatology 38, 345–354.

Hashimoto, F., Davis, R. L., and Egli, D. (1994). Hepatitis following treatments

with famotidine and then cimetidine. Ann. Pharmacother. 28, 37–39.

Hewett, J. A., Jean, P. A., Kunkel, S. L., and Roth, R. A. (1993). Relationship

between tumor necrosis factor- and neutrophils in endotoxin-induced liver

injury. Am. J. Physiol. 265, G1011–G1015.

Hewett, J. A., Schultze, A. E., Van Cise, S., and Roth, R. A. (1992). Neutrophil

depletion protects against liver injury from bacterial endotoxin. Lab.

Investig. 66, 347–361.

Jaeschke, H. (2006). Mechanisms of liver injury. II. Mechanisms of neutrophil-

induced liver cell injury during hepatic ischemia-reperfusion and other acute

inflammatory conditions. Am. J. Physiol. Gastrointest. Liver Physiol. 290,

G1083–G1088.

ROLE OF TNF-a IN LPS/RAN-INDUCED LIVER INJURY 279

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

Ji, Q., Zhang, L., Jia, H., and Xu, J. (2004). Pentoxifylline inhibits endotoxin-

induced NF-kappa B activation and associated production of proinflamma-

tory cytokines. Ann. Clin. Lab. Sci. 34, 427–436.

Jimenez-Saenz, M., Arguelles-Arias, F., Herrerias-Gutierrez, J. M., and Duran-

Quintana, J. A. (2000). Acute cholestatic hepatitis in a child treated with

famotidine. Am. J. Gastroenterol. 95, 3665–3666.

Klebanoff, S. J., Vadas, M. A., Harlan, J. M., Sparks, L. H., Gamble, J. R.,

Agosti, J. M., and Waltersdorph, A. M. (1986). Stimulation of neutrophils by

tumor necrosis factor. J. Immunol. 136, 4220–4225.

Kreydiyyeh, S. I., Riman, S., Serhan, M., and Kassardjian, A. (2007). TNF-

alpha modulates hepatic Naþ-Kþ ATPase activity via PGE2 and EP2

receptors. Prostaglandins Other Lipid Mediat. 83, 295–303.

Kushimoto, S., Okajima, K., Uchiba, M., Murakami, K., Okabe, H., and

Takatsuki, K. (1996). Role of granulocyte elastase in ischemia/reperfusion

injury of rat liver. Crit. Care Med. 24, 1908–1912.

Levi, M., ten Cate, M., Bauer, K. A., van der Poll, T., Edgington, T. S.,

Buller, H. R., van Deventer, S. J., Hack, C. E., ten Cate, J. W., and

Rosenberg, R. D. (1994). Inhibition of endotoxin-induced activation of

coagulation and fibrinolysis by pentoxifylline or by a monoclonal anti-tissue

factor antibody in chimpanzees. J. Clin. Investig. 93, 114–120.

Lin, J. H. (1991). Pharmacokinetic and pharmacodynamic properties of

histamine H2-receptor antagonists: Relationship between intrinsic potency

and effective plasma concentrations. Clin. Pharmacokinet. 20, 218–236.

Locksley, R. M., Killeen, N., and Lenardo, M. J. (2001). The TNF and TNF

receptor superfamilies: Integrating mammalian biology. Cell 104, 487–501.

Luyendyk, J. P., Lehman-McKeeman, L. D., Nelson, D. M., Bhaskaran, V. M.,

Reilly, T. P., Car, B. D., Cantor, G. H., Maddox, J. F., Ganey, P. E., and

Roth, R. A. (2006). Unique gene expression and hepatocellular injury in the

lipopolysaccharide-ranitidine drug idiosyncrasy rat model: Comparison with

famotidine. Toxicol. Sci. 90, 569–585.

Luyendyk, J. P., Maddox, J. F., Cosma, G. N., Ganey, P. E., Cockerell, G. L.,

and Roth, R. A. (2003). Ranitidine treatment during a modest inflammatory

response precipitates idiosyncrasy-like liver injury in rats. J. Pharmacol.

Exp. Ther. 307, 9–16.

Luyendyk, J. P., Maddox, J. F., Green, C. D., Ganey, P. E., and Roth, R. A.

(2004a). Role of hepatic fibrin in idiosyncrasy-like liver injury from

lipopolysaccharide-ranitidine coexposure in rats. Hepatology 40, 1342–1351.

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

Maddox, J. F., Cosma, G. N., Ganey, P. E., and Roth, R. A. (2004b). Gene

expression analysis points to hemostasis in livers of rats cotreated with

lipopolysaccharide and ranitidine. Toxicol. Sci. 80, 203–213.

Luyendyk, J. P., Shaw, P. J., Green, C. D., Maddox, J. F., Ganey, P. E., and

Roth, R. A. (2005). Coagulation-mediated hypoxia and neutrophil-dependent

hepatic injury in rats given lipopolysaccharide and ranitidine. J. Pharmacol.

Exp. Ther. 314, 1023–1031.

Nagaki, M., Muto, Y., Ohnishi, H., and Moriwaki, H. (1991). Significance of

tumor necrosis factor (TNF) and interleukin-1 (IL-1) in the pathogenesis of

fulminant hepatitis: Possible involvement of serine protease in TNF-

mediated liver injury. Gastroenterol. Jpn. 26, 448–455.

Neumann, J. F., Ott, I., Marx, N., Luther, T., Kenngott, S., and Gawaz, M.

(1997). Effect of human recombinant interleukin-6 and interleukin-8 on

monocyte procoagulant activity. Arterioscler. Thromb. Vasc. Biol. 17,

3399–3405.

Pawlinski, R., Pedersen, B., Schabbauer, G., Tencati, M., Holscher, T.,

Boisvert, W., Andrade-Gordon, P., Frank, R. D., and Mackman, N. (2004).

Role of tissue factor and protease-activated receptors in a mouse model of

endotoxemia. Blood 103, 1342–1347.

Ribeiro, J. M., Lucas, M., Baptista, A., and Victorino, R.-R. M. (2000). Fatal

hepatitis associated with ranitidine. Am. J. Gastroenterol. 95, 559–560.

Roth, R. A., Harkema, J. R., Pestka, J. P., and Ganey, P. E. (1997). Is exposure

to bacterial endotoxin a determinant of susceptibility to intoxication from

xenobiotic agents? Toxicol. Appl. Pharmacol. 147, 300–311.

Scales, W. E., Chensue, S. W., Otterness, I., and Kunkel, S. L. (1989).

Regulation of monokine gene expression: Prostaglandin E2 suppresses tumor

necrosis factor but not interleukin-1 alpha or beta-mRNA and cell-associated

bioactivity. J. Leukoc. Biol. 45, 416–421.

Scarpignato, C., Tramacere, R., and Zappia, L. (1987). Antisecretory and

antiulcer effect of H2-receptor antagonist famotidine in the rat: Comparison

with ranitidine. Br. J. Pharmacol. 92, 153–159.

Schandene, L., Vandenbussche, P., Crusiaux, A., Alegre, M. L.,

Abramowicz, D., Dupont, E., Content, J., and Goldman, M. (1992).

Differential effects of pentoxifylline on the production of tumour necrosis

factor-alpha (TNF-alpha) and interleukin-6 (IL-6) by monocytes and T cells.

Immunology 76, 30–34.

Schleef, R. R., Bevilaqua, M. P., Sawdey, M., Gimbrone, M. A., and

Loskutoff, D. J. (1988). Cytokine activation of vascular endothelium. Effects

on tissue-type plasminogen activator and type 1 plasminogen activator

inhibitor. J. Biol. Chem. 263, 5797–5803.

Smedly, L. A., Tonnesen, M. G., Sandhaus, R. A., Haslett, C., Guthrie, L. A.,

Johnston, R. B., Jr, Henson, P. M., and Worthen, G. S. (1986). Neutrophil-

mediated injury to endothelial cells. Enhancement by endotoxin and essential

role of neutrophil elastase. J. Clin. Investig. 77, 1233–1243.

Sneed, R. A., Buchweitz, J. P., Jean, P. A., and Ganey, P. E. (2000).

Pentoxifylline attenuates bacterial lipopolysaccharide-induced enhancement

of allyl alcohol hepatotoxicity. Toxicol. Sci. 56, 203–210.

Takei, Y., Kawano, S., Nishimura, Y., Goto, M., Nagai, H., Chen, S. S.,

Omae, A., Fusamoto, H., Kamada, T., Ikeda, K., et al. (1995). Apoptosis:

A new mechanism of endothelial and Kupffer cell killing. J. Gastroenterol.

Hepatol. 10(Suppl. 1), S65–S67.

Taylor, F. B., Jr, Chang, A., Ruf, W., et al. (1991). Lethal E. coli septic shock is

prevented by blocking tissue factor with monoclonal antibody. Circ. Shock

33, 127–134.

Tukov, F. F., Maddox, J. F., Amacher, D. E., Bobrowski, W. F., Roth, R. A.,

and Ganey, P. E. (2006). Modeling inflammation-drug interactions in vitro:

A rat Kupffer cell-hepatocyte coculture system. Toxicol. In Vitro 20,

1488–1499.

van der Poll, T., Coyle, S. M., Levi, M., Jansen, P. M., Dentener, M.,

Barbosa, K., Buurman, W. A., Hack, C. E., ten Cate, J. W., Agosti, J. M.,

et al. (1997). Effect of a recombinant dimeric tumor necrosis factor receptor

on inflammatory responses to intravenous endotoxin in normal humans.

Blood 89, 3727–3734.

Vassalli, P. (1992). The pathophysiology of tumor necrosis factors. Annu. Rev.

Immunol. 10, 411–452.

Vial, T., Goubier, C., Bergeret, A., Cabrera, F., Evreux, J.-C., and Descotes, J.

(1991). Side effects of ranitidine. Drug Saf. 6, 94–117.

Wang, Y., and Thorlacius, H. (2005). Mast cell-derived tumour necrosis factor-

alpha mediates macrophage inflammatory protein-2-induced recruitment of

neutrophils in mice. Br. J. Pharmacol. 145, 1062–1068.

Yee, S. B., Ganey, P. E., and Roth, R. A. (2002). Temporal relationships in

the augmentation of monocrotaline hepatoxicity by bacterial endotoxin.

J. Toxicol. Environ. Health A 65, 961–976.

Yee, S. B., Ganey, P. E., and Roth, R. A. (2003). The role of Kupffer cells and

TNF-alpha in monocrotaline and bacterial lipopolysaccharide-induced liver

injury. Toxicol. Sci. 71, 124–132.

280 TUKOV ET AL.

by guest on January 21, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from