Gene Expression Analysis Points to Hemostasis in Livers of Rats Cotreated with Lipopolysaccharide and Ranitidine

James P. Luyendyk*, William B. Mattes{ddagger},2, Lyle D. Burgoon*, Timothy R. Zacharewski{dagger}, Jane F. Maddox*, Gregory N. Cosma{ddagger},3, Patricia E. Ganey* and Robert A. Roth*,1

Departments of * Pharmacology and Toxicology and {dagger} Biochemistry and Molecular Biology, National Food Safety and Toxicology Center, Center for Integrative Toxicology, Michigan State University, East Lansing, Michigan 48824; and {ddagger} Investigative Toxicology, Pharmacia Corporation, Kalamazoo, Michigan

Received February 26, 2004; accepted April 2, 2004


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Supplemental data
 REFERENCES
 
Studies in rats have demonstrated that modest underlying inflammation can precipitate idiosyncratic-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 be distinguished among vehicle-, LPS-, RAN- and LPS/RAN-treated rats before the onset of significant liver injury in the 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. Hierarchal clustering of gene expression data from Affymetrix U34A rat genome array 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 products involved in inflammation, hypoxia, and cell death. Some were enhanced synergistically by LPS/RAN cotreatment. Real-time PCR confirmed robust changes in expression of B-cell translocation gene 2, early growth response-1, and plasminogen-activator inhibitor-1 (PAI-1) in cotreated rats. The increase in PAI-1 mRNA was reflected in an increase in serum PAI-1 protein concentration in LPS/RAN-treated rats. Consistent with the antifibrinolytic activity of PAI-1, significant fibrin deposition occurred only in livers of LPS/RAN-treated rats. The results suggest the possibility that expression of PAI-1 promotes fibrin deposition in liver sinusoids of LPS/RAN-treated rats and are consistent with the development of local ischemia and consequent tissue hypoxia.

Key Words: inflammation; ranitidine; liver; gene array; plasminogen activator inhibitor-1; hypoxia.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Supplemental data
 REFERENCES
 
Idiosyncratic drug reactions are adverse responses of unknown etiology that occur in a small fraction of people taking a drug, with the liver being a frequent target of toxicity. One drug associated with idiosyncratic hepatotoxicity is the histamine 2-receptor antagonist, ranitidine (RAN), a drug used for the treatment of duodenal ulcers, gastric hypersecretory diseases, and gastroesophageal reflux disease. Mechanisms underlying human idiosyncratic liver injury from RAN remain unclear. Recently it has been suggested that idiosyncratic reactions to some drugs, including RAN, may result from episodic inflammation (Roth et al., 2003Go).

Bacterial lipopolysaccharide (endotoxin, LPS) is a potent inflammagen, and exposure to small, noninjurious amounts of LPS renders rats susceptible to liver injury from normally nontoxic doses of hepatotoxic xenobiotics, including some drugs (Roth et al., 2003Go). For example, rats cotreated with a small, normally nonhepatotoxic dose of LPS and either a small nonhepatotoxic dose of RAN or the phenothiazine antipsychotic drug chlorpromazine develop hepatotoxicity similar to idiosyncratic reactions observed in people treated with either drug (Buchweitz et al., 2002Go; Luyendyk et al., 2003Go). Halothane, another agent associated with human idiosyncratic hepatotoxicity, causes liver injury in hypoxic rats when they are coexposed to LPS (Lind et al., 1984Go). Thus, these studies support the hypothesis that concurrent inflammation might precipitate some idiosyncratic drug reactions. However, mechanisms of toxicity in these animal models are not understood.

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, 2002Go). An important component of LPS activation of inflammatory cells such as macrophages is transcriptional activation of numerous genes (Gao et al., 2002Go). Many of these gene products such as tumor necrosis factor {alpha} (TNF-{alpha}) can further activate transcription of other cytokines, adhesion molecules, and neutrophil (PMN) chemokines in other cell types such as endothelial cells (Zhao et al., 2003Go). Increased TNF-{alpha} mRNA can be detected in livers of rats treated with LPS, and serum TNF-{alpha} concentration is markedly increased after LPS exposure (Barton et al., 2001Go; Hewett et al., 1993Go). Interestingly, TNF-{alpha} is important for liver injury from large doses of LPS in a mechanism dependent on PMN activation (Hewett et al., 1993Go). Thus, enhanced expression of certain genes after LPS exposure is important for liver injury, and understanding these changes could further 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 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., 2001Go). In this model, COX-2 inhibition affords partial protection from liver injury, suggesting that augmented COX-2 gene expression is important for AA/LPS-induced liver injury (Ganey et al., 2001Go). In rats cotreated with nonhepatotoxic doses of aflatoxin B1 (AFB1) and LPS, TNF-{alpha} mRNA is increased in liver to a level similar to rats treated with LPS alone. However, the serum concentration of TNF-{alpha} is significantly greater in AFB1/LPS-treated rats, and this cytokine is causally involved in the potentiation of hepatocellular injury (Barton et al., 2001Go). 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 (Li et al., 2003Go).

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 plasminogen activator inhibitor-1 (PAI-1), and hepatic fibrin deposition was evaluated to determine if the enhanced PAI-1 expression was associated with tissue fibrin deposition as a functional consequence.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Supplemental data
 REFERENCES
 
Materials. Unless otherwise noted, all chemicals were purchased from Sigma Chemical Co. (St. Louis, MO). Lipopolysaccharide derived from E. coli serotype O55:B5 with an activity of 6.6 x 106 endotoxin units (EU)/mg was used for these studies. This activity was determined using a colorometric, kinetic Limulus Amebocyte Lysate (LAL) assay (Kit #50–650 U) purchased from Cambrex (East Rutherford, NJ).

Animals. Male Sprague-Dawley rats (Crl:CD (SD)IGS BR; Charles River, Portage, MI) weighing 250–350 g were used for this study. 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.

Experimental protocol. Rats fasted for 24 h were given 44.4 x 106 EU/kg LPS or its saline vehicle, iv. Two h later, 30 mg/kg RAN or sterile phosphate-buffered saline (PBS) vehicle was administered iv. RAN solution was administered at 2 ml/kg body weight at a rate of approximately 0.15 ml/min. Accordingly, rats were treated with either saline and PBS (Veh group), LPS and PBS (LPS group), saline and RAN (RAN group), or LPS and RAN (LPS/RAN group). Three h 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 isolation. Groups treated with either vehicle, LPS, or RAN contained 3 rats, whereas 4 rats comprised the LPS/RAN group.

Hepatotoxicity assessment. Hepatic parenchymal cell injury was estimated as an increase in serum alanine aminotransferase (ALT) activity. ALT activity was determined spectrophotometrically using Infinity-ALT reagent from Sigma Chemical Co. (St. Louis, MO). Sinusoidal endothelial cell function was estimated using a commercially available, enzyme-linked immunosorbent assay (ELISA) for hyaluronic acid (Corgenix Medical Corporation, Westminster, CO). Formalin-fixed liver samples (3–4 samples/rat) were embedded in paraffin, sectioned at 5 µm, stained with hematoxylin and eosin (H&E), and examined by light microscopy.

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).

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 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 an 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 is the number of data points on the array and XX is the number of arrays. Normalized signal data were imported into the JMP (Release 5.0.1.2, SAS Institute, Inc.) software for principal component analysis.

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. Probe sets 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 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 and Hochberg, 1995Go). The FDR provided an approach capable of decreasing, to a user-determined 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 vehicle-treated group were identified for each treatment (i.e., LPS, RAN, or LPS/RAN) by an FDR criterion with {alpha} = 0.05. Each probeset that changed relative to its expression in vehicle-treated rats was then assigned to a set defined by the treatment or treatments that produced a change in its activity (i.e., LPS/RAN [LR], L, R, L{cap}R, LR{cap}L, LR{cap}R, LR{cap}L{cap}R, where "{cap}" indicates intersection of sets). The resulting sets were visualized using a Venn diagram (Fig. 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, LR{cap}L, LR{cap}R and LR{cap}L{cap}R 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 LR{cap}L and LR{cap}L{cap}R or in the LR{cap}R and LR{cap}L{cap}R 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 LR{cap}L sets.



View larger version (25K):
[in this window]
[in a new window]
 
FIG. 3. Venn diagram depiction of probeset activity relative to Veh/Veh-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). Three h after RAN administration, RNA was isolated from liver and gene expression evaluated by Affymetrix U34A Rat Genome Array. The number of Affymetrix probesets increased ({uparrow}) or decreased ({downarrow}) 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 ({cap}): LR, probesets changed only after treatment with LPS/RAN (Supplemental Table 1); L, probesets changed only after treatment with LPS (Supplemental Table 2); R, probesets changed only after treatment with RAN (Supplemental Table 3); L{cap}R, probesets changed after treatment both with LPS alone and with RAN alone (Supplemental Table 4); LR{cap}L, probesets changed after treatment both with LPS/RAN and with LPS alone (Supplemental Table 5); LR{cap}R, probesets changed after treatment both with LPS/RAN and with RAN alone (Supplemental Table 6); LR{cap}L{cap}R, probesets changed after treatment with LPS/RAN, LPS alone, and RAN alone (Supplemental Table 7).

 
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, 2004Go). 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.

Reverse transcription and real-time PCR. RNA quantification was performed on the same samples from the gene array experiment using a Spectramax Microplate Spectrophotometer (Molecular Devices, Sunnyvale, CA). Random priming was performed in a final volume of 12.5 µl using 500 ng of total RNA, 7.5 mM Random Hexamers (Amersham Biosciences, Piscataway, NJ), and 1 mM dNTPs (Invitrogen, Carlsbad, California). RNA was denatured at 65°C for 5 min and chilled on ice. Reverse Transcription (RT) master mix was prepared in a final volume of 12.5 µl with a final concentration of 20 U/ml of Superscript II RNAse H-Reverse Transcriptase, 4 U/ml RNaseOut, 2 mM dithiothreitol, and 1X 1st Strand RT buffer (Invitrogen, Carlsbad, California). Denatured RNA and RT master mix were combined in total volume of 25 µl. The reverse transcriptase reaction was performed at room temperature for 10 min, at 42°C for 60 min, and then at 70°C for 15 min in an 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 mRNA levels of the following genes by real-time PCR analyses. Early growth response-1 (egr-1): 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-1): upper primer- 5' AAC CCA GGC CGA CTT CA 3'; lower primer- 5' CAT GCG GGC TGA GAC TAG AAT 3'. Ribosomal protein L19 (Rpl19): upper primer- 5' CTC GAT GCC GGA AGA ACA C 3'; lower primer- 5' CGA GCG TTG 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 µl per reaction, with 1 ng of cDNA and 1X SYBR Green PCR Master Mix (PE Applied Biosystems, Foster City, CA). Primer mixture was prepared in 5 µl per reaction with a final concentration of 0.3 µM 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 from all livers of vehicle-treated rats.

Immunohistochemistry. Liver samples evaluated for fibrin immunohistochemistry were from the 3-h post-treatment time point in a previous study (Luyendyk et al., 2003Go). A 1 cm3 block of liver cut from the left medial lobe was frozen for 8 min in liquid nitrogen-chilled isopentane, then stored at – 80°C until processing. Eight-mm-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 except for cross-linked fibrin; therefore, only cross-linked fibrin remains in sections of liver (Schnitt et al., 1993Go). Sections were blocked with PBS containing 10% horse serum (i.e., blocking solution; Vector Laboratories) for 30 min, 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 h with donkey anti-goat 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 a fluorescent microscope. No staining was observed in controls for which the primary or secondary antibody was eliminated from the staining protocol. Liver sections from all treatment groups that were compared morphometrically were stained at the same time.

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 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)Go. Ten random fields analyzed for each liver section were averaged and counted as a replicate, i.e., each replicate represents a different rat.

Evaluation of serum PAI-1 concentration. Serum total PAI-1 concentration (i.e., inactive, active, and bound to plasminogen activator) was evaluated using a commercially available ELISA purchased from American Diagnostica, Inc. (Greenwich, CT.).

Statistical analysis. 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.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Supplemental data
 REFERENCES
 
Development of Hepatic Parenchymal Cell Injury
Given alone, the doses of LPS and RAN are not hepatotoxic up to 24 h after administration (Luyendyk et al., 2003Go). Confirming earlier results (Luyendyk et al., 2003Go), no significant change in ALT was observed in rats given RAN or LPS (Fig. 1), and LPS/RAN cotreatment did not cause a statistically significant increase in ALT by 3 h (Fig. 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).



View larger version (12K):
[in this window]
[in a new window]
 
FIG. 1. Evaluation of hepatic parenchymal cell injury after LPS/RAN cotreatment: Rats were treated with 44.4 x 106 EU/kg (endotoxin unit per kilogram) LPS or its Veh (iv), then two h later with 30 mg/kg RAN or its Veh (iv). Hepatic parenchymal cell injury was estimated 3 h after RAN administration by increases in serum ALT (alanine aminotransferase) activity (n = 3 for rats given Veh/Veh, LPS/ Veh, or Veh/RAN (n = 4 for rats given LPS/RAN). Data are expressed as mean ± SEM (standard error of the mean). No treatment was found to be significantly different from Veh-treated rats.

 
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 2. Four clusters resolved from this analysis, segregating animals by their specific treatment (Veh, LPS, RAN, or LPS/RAN). Additionally, animals treated with LPS, RAN or LPS/RAN clustered separately from Veh-treated animals.



View larger version (86K):
[in this window]
[in a new window]
 
FIG. 2. Hierarchical clustering of hepatic gene expression after LPS/RAN-cotreatment: Rats were treated with 44.4 x 106 EU/kg LPS or its vehicle (iv), then two h later with 30 mg/kg RAN or its vehicle (iv). Three h 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.

 
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 vehicle 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., LR{cap}L) contains probesets changed after LPS/RAN treatment and after LPS treatment. The resulting sets are summarized as a Venn diagram in Figure 3. 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, signal intensities relative to vehicle 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 from this treatment (Luyendyk et al., 2003Go). Several probesets were also changed only after LPS treatment (163 increased; 46 decreased) or only after 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., 2003Go); however, the potential exists for interaction of one of these gene products with another, resulting in liver injury. Probesets changing after LPS/RAN treatment and after either agent given alone (i.e., LR{cap}L, LR{cap}R) are 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, LR{cap}L, LR{cap}R, and LR{cap}L{cap}R sets (Table 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 LR{cap}L and LR{cap}L{cap}R sets) or in LPS/RAN- and RAN-treated rats (from LR{cap}R and LR{cap}L{cap}R 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 LR{cap}L 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.


View this table:
[in this window]
[in a new window]
 
TABLE 1 Expression Pattern Groups and Gene Product Classification

 
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 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 were hypoxia-inducible (Table 1). For example, hypoxia-inducible genes, including early growth response-1 (egr-1), glucose transporter-1 (GLUT-1), insulin-like growth factor binding protein-1 (igfbp-1), and plasminogen activator inhibitor-1 (PAI-1), had greater expression after LPS/RAN treatment as compared to expression after treatment with LPS or RAN alone (Table 2). Furthermore, genes involved in inflammation such as the chemokine (C-X-C motif) ligand 10 (CxCl10) and the cell cycle regulator B-cell translocation gene-2 (btg-2) segregated into this group (Table 2).


View this table:
[in this window]
[in a new window]
 
TABLE 2 Genes With A Greater Signal in Livers from LPS/RAN-treated Rats Compared to Other Treatments

 
Table 1 shows that groups of genes expressed to a similar degree after 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 1). Specific genes comprising each of these groups are available online in Supplemental Tables 8–11.

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 Genomics. Real-time PCR was performed for three of the genes with increased expression in LPS/RAN-treated rats: PAI-1, egr-1, and btg-2. PAI-1 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, Fig. 4A). This pattern of expression was consistent with signal intensities for the corresponding Affymetrix probeset (circles, Fig. 4A) identified as active by the FDR filter. For egr-1 (Fig. 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 (Fig. 4B). 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 (Fig. 4C). Although a single probeset for btg-2 (M60921_at) is shown for comparison (circles, Fig. 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).



View larger version (21K):
[in this window]
[in a new window]
 
FIG. 4. Real-time PCR confirmation of gene-array data: 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). Three h after RAN administration, RNA was isolated from whole liver, and SYBR Green real-time PCR was performed for (A) plasminogen activator inhibitor-1 (PAI-1), (B) early growth response-1 (egr-1), and (C) B-cell translocation gene-2 (btg-2). 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 ({Delta}{Delta}DCt) method (bars). Normalized signal intensities for corresponding Affymetrix probesets are graphed for comparison (circles). Although a single probeset for btg-2 (M60921_at) is shown for comparison, all 3 probesets for this gene determined to be active by the FDR filter had similar expression patterns (M60921_at, M60921_g_at, rc_AA944156_s_at).Data are expressed as mean ± SEM; n = 3–4; *significantly different from Veh/Veh-treated rats (p < 0.05); #significantly different from all other treatments (p < 0.05).

 
Evaluation of Serum PAI-1
To determine if the change in hepatic gene expression of PAI-1 resulted in altered serum concentration of PAI-1 protein, total PAI-1 protein was measured. Serum PAI-1 was significantly increased after either LPS or RAN treatment by 450- and 70-fold, respectively. Serum PAI-1 in LPS/RAN-treated rats was significantly greater than that of rats treated with either agent given alone (Fig. 5).



View larger version (13K):
[in this window]
[in a new window]
 
FIG. 5. Serum plasminogen activator inhibitor-1 (PAI-1) concentration after LPS/RAN treatment: 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). Serum concentration of PAI-1 in its active, latent, and complex forms was evaluated three h later, using an ELISA; n = 3–4 rats per group. Data are expressed as mean ± SEM; *significantly different from Veh/Veh-treated rats (p < 0.05); #significantly different from all other treatments (p< 0.05).

 
LPS/RAN Treatment and Sinusoidal Endothelial Cells (SECs)
In addition to expression in hepatic parenchymal cells, PAI-1 is expressed by endothelial cells exposed to factors such as LPS and inflammatory cytokines (Colman et al., 1994Go). 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 HA in the blood is cleared by SECs in the liver (Kobayashi et al., 1999Go). Accordingly, increased plasma HA concentration suggests altered SEC function, and this has been used as a biomarker after toxic insult (Copple et al., 2002Go; Deaciuc et al., 1993Go). 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 (Fig. 6). PAI-1 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).



View larger version (12K):
[in this window]
[in a new window]
 
FIG. 6. Effect of LPS/RAN cotreatment on serum hyaluronic acid (HA) concentration: Rats were treated with 44.4 x 106 EU/kg LPS or its vehicle (iv), then two h later with 30 mg/kg RAN or its vehicle (iv). Altered sinusoidal endothelial cell function was estimated 3 h after RAN administration by increases in serum HA activity;. n = 3–4 rats per group. Data are expressed as mean ± SEM; *significantly different from vehicle/vehicle-treated rats (p < 0.05); #significantly different from all other treatments (p < 0.05).

 
Hepatic Fibrin Deposition
Enhanced PAI-1 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 7 shows representative images of hepatic fibrin staining in livers. Fibrin staining in the intima of the larger vessels of vehicle-treated rats (Fig. 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 vehicle-treated rats. Similarly, no sinusoidal staining was observed in rats given RAN alone (Fig. 7C). Slight fibrin staining was observed in livers of rats treated with LPS alone (Fig. 7B). In LPS/RAN-treated rats (Fig. 7D), a much more pronounced, panlobular fibrin staining occurred in sinusoids. Morphometry revealed a statistically significant increase in fibrin staining only in livers from animals cotreated with LPS/RAN (Fig. 7E).



View larger version (28K):
[in this window]
[in a new window]
 
FIG. 7. Effect of LPS/RAN cotreatment on hepatic fibrin deposition: 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 immunohistochemistry as described in Materials and Methods. Representative images of fibrin staining in livers from (A) Veh- and (C) RAN-treated rats, showing minimal stain (black). (B) Representative image from rat treated with LPS showing slight panlobular fibrin staining. (D) Representative image from 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, 100X fields per tissue was determined morphometrically, as described in Materials and Methods. Data are expressed as mean ± SEM; n = 3; #significantly different from all other treatments (p < 0.05).

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Supplemental data
 REFERENCES
 
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 h 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., 2003Go). At this time, hierarchical cluster analysis of hepatic gene expression changes was able to segregate rats by treatment group (Fig. 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 vehicle-treated rats (Fig. 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 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.

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/RAN {approx} LPS in Table 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-{alpha} and interleukin-1ß 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-1 (LOX-1), CD14 and CD38, the transcription factors CCAAT/enhancer binding protein (C/EBP)-{delta} and –ß, and products involved in signal transduction such as Janus kinase 2 (Jak2) and phosphodiesterase 4B (PDE4B) fit this pattern of expression.

PDE4B is an important regulator of inflammatory responses, including expression of cytokines and activation of inflammatory cells such as PMNs (Jin and Conti, 2003Go; Essayan, 1999Go). In this regard, it is of interest that hepatic lesions that develop in LPS/RAN-treated rats are laden with PMNs (Luyendyk et al., 2003Go). Regulation of inflammation by PDE4B 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 PMN-derived factors (Luyendyk et al., 2003Go). It therefore seems possible that RAN sensitizes hepatocytes to become injured from otherwise noninjurious 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 LPS/RAN < LPS in Table 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 (150 kDa) (ORP150), have been implicated in protection against ischemic stress (Gray et al., 1997Go; Ozawa et al., 1999Go). This suggests the possibility that livers from 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-{alpha}-induced apoptosis (Hayashi and Faustman, 2000Go). Overall, these and other genes expressed to a greater degree after LPS treatment (heat shock 70 kDa protein 1A, heme oxygenase-2, dnaJ homolog subfamily ß member 9) compared to LPS/RAN cotreatment may confer some degree of cytoprotection, perhaps by decreasing sensitivity to hypoxia, inflammatory mediators, or oxidative stress.

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 and >RAN in Table 1, Supplemental Table 8, Table 2). Several of the hypoxia-inducible genes in this group can also participate in inflammatory responses. For example, CxCl10 (interferon-inducible cytokine IP-10) is induced under hypoxic conditions and modulates recruitment and retention of inflammatory cells (Neville et al., 1997Go). The hypoxia-inducible transcription factor egr-1 is involved in cell death signaling (Thiel and Cibelli, 2002Go), but it can also influence inflammatory responses by altering cytokine expression (Yan et al., 1999Go; Shi et al., 2002Go). 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-1. PAI-1 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., 2002Go; Kietzmann et al., 1999Go; Hamaguchi et al., 2003Go). Although the cell source and mechanism of enhanced PAI-1 expression in livers of LPS/RAN-cotreated rats is not known, one possibility is that RAN may indirectly augment expression by increasing levels of cytokines known to induce PAI-1, such as TNF-{alpha} 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 PAI-1. Perturbation of SECs by hypoxia (Kietzman et al., 1999Go) or altered signal transduction may be responsible for augmented PAI-1 expression in endothelial cells in LPS/RAN-treated rats. P38 mitogen-activated protein kinase is involved in induction of PAI-1 during hypoxia but does not appear to be important in induction of PAI-1 by TNF-{alpha} (Hamaguchi et al., 2003Go; Kietzmann et al., 2003Go). Another possibility is that cotreatment with RAN influences PAI-1 expression at the transcriptional level. Gruber et al. (2003)Go demonstrated that the PAI-1 promoter contains a response element for the orphan receptor Nur77 (NGFI-B, TR3). Furthermore, Nur77 overexpression in human umbilical vein endothelial cells activates a luciferase reporter gene controlled by a PAI-1 promoter, and Nur77 is necessary for induction of PAI-1 expression by TNF-{alpha} (Gruber et al., 2003Go). Since Nur77 is also a hypoxia-inducible gene (Choi et al., 2004Go), its expression might be expected to be enhanced in LPS/RAN-treated rats, since numerous other hypoxia-inducible genes, including PAI-1, were identified in this group (Table 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 vehicle-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 PAI-1 in LPS/RAN-treated rats. However, additional experiments are necessary to prove such a connection.

The greater expression of the PAI-1 gene in livers of LPS/RAN-treated rats (Fig. 4A) was reflected in enhanced concentration of PAI-1 protein in serum (Fig. 5), suggesting a functional consequence to its induction. PAI-1 has many physiological roles, including inhibition of fibrinolysis and modulation of inflammatory cell migration (Binder et al., 2002Go; Marshall et al., 2003Go). 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 PAI-1 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 injury. This hypothesis is consistent with the observation that numerous hypoxia-inducible genes were expressed in livers of LPS/RAN-cotreated rats (Table 3). Inasmuch as the PAI-1 gene is hypoxia-inducible, the presence of hypoxia might further enhance its expression, triggering a cascade of hemostatic dysregulation. Additionally, hypoxic upregulation of egr-1 could facilitate coagulation system activation by upregulating tissue factor on liver-cell membranes (Pawlinski et al., 2003Go). Overall, the data suggest that enhanced PAI-1 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 from liver injury by fibrinolytic or anticoagulant drugs in LPS/RAN-treated rats support this hypothesis (Luyendyk et al., 2004Go).

In summary, rats were treated with either a nonhepatotoxic dose of LPS or its vehicle and with either RAN or its vehicle. Of the four treatments, only LPS/RAN treatment results in liver injury (Luyendyk et al., 2003Go). 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 PAI-1 and egr-1, were expressed to a greater degree in livers after LPS/RAN treatment compared to either agent given alone. The enhanced PAI-1 gene expression was reflected in increased PAI-1 protein in serum. Significant fibrin deposits in liver sinusoids were observed only in LPS/RAN-cotreated rats, consistent with the antifibrinolytic activity of PAI-1. 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.


    Supplemental data
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Supplemental data
 REFERENCES
 
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 for download as Tables 8–11 in Microsoft 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.oupjournals.org.


    ACKNOWLEDGMENTS
 
This work was supported by grants from Pharmacia Corporation and the National Institutes of Health (DK061315). The authors thank Christopher Green, Rohan Pradhan, and Jennifer Fostel for technical assistance. Lyle Burgoon and James Luyendyk were partially supported by training-grant number 5 T32 ES07255 from the National Institute of Environmental Health Sciences (NIEHS), NIH, and James Luyendyk partly by the Barnett Rosenberg Fellowship from Michigan State University.


    NOTES
 
2 Present address: Gene Logic, Inc., Gaithersburg, MD 20878. Back

3 Present address: Bristol Myers Squibb Co., New Brunswick, NJ. Back

1 To whom correspondence should be addressed at the Department of Pharmacology and Toxicology, National Food Safety and Toxicology Center, Michigan State University, East Lansing, MI 48824. Fax: (517) 432-2310. E-mail: rothr{at}msu.edu.


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Supplemental data
 REFERENCES
 
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.[CrossRef][ISI][Medline]

Benjamini,Y., and Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. J. R. Statist. Soc. 57, 289–300.[ISI]

Beutler, B. (2002). TLR4 as the mammalian endotoxin sensor. Curr. Top. Microbiol. Immunol. 270, 109–120.[ISI][Medline]

Binder, B. R., Christ, G., Gruber, F., Grubic, N., Hufnagl, P., Krebs, M., Mihaly, J., and Prager, G. W. (2002). Plasminogen activator inhibitor 1: Physiological and pathophysiological roles. News Physiol. Sci. 17, 56–61.[Abstract/Free Full Text]

Buchweitz, J. P., Ganey, P. E., Bursian, S. J., and Roth, R. A. (2002). Underlying endotoxemia augments toxic responses to chlorpromazine: Is there a relationship to drug idiosyncrasy? J. Pharmacol. Exp. Ther. 300, 460–467.[Abstract/Free Full Text]

Choi, J. W., Park, S. C., Kang, G. H., Liu, J. O., and Youn, H. D. (2004). Nur77 activated by hypoxia-inducible factor-1{alpha} overproduces pro-opiomelanocortin in von Hippel-Lindau-mutated renal cell carcinoma. Cancer Res. 64, 35–9.[Abstract/Free Full Text]

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.[Abstract/Free Full Text]

Deaciuc, I. V., McDonough, K. H., Bagby, G. J., and Spitzer, J. J. (1993). Alcohol consumption in rats potentiates the deleterious effect of gram-negative sepsis on hepatic hyaluronan uptake. Alcohol Clin. Exp. Res. 17, 1002–1008.[ISI][Medline]

Essayan, D. M. (1999). Cyclic nucleotide phosphodiesterase (PDE) inhibitors and immunomodulation. Biochem. Pharmacol. 57, 965–973.[CrossRef][ISI][Medline]

Ganey, P. E., Barton, Y. W., Kinser, S., Sneed, R. A., Barton, C. C., and Roth, R. A. (2001). Involvement of cyclooxygenase-2 in the potentiation of allyl alcohol-induced liver injury by bacterial lipopolysaccharide. Toxicol. Appl. Pharmacol. 174, 113–121.[CrossRef][ISI][Medline]

Gao, J. J., Diesl, V., Wittmann, T., Morrison, D. C., Ryan, J. L., Vogel, S. N., and Follettie, M. T. (2002). Regulation of gene expression in mouse macrophages stimulated with bacterial CpG-DNA and lipopolysaccharide. J. Leukoc. Biol. 72, 1234–1245.[Abstract/Free Full Text]

Gray, M. O., 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.[Abstract/Free Full Text]

Gruber, F., Hufnagl, P., Hofer-Warbinek, R., Schmid, J. A., Breuss, J. M., Huber-Beckmann, R., Lucerna, M., Papac, N., Harant, H., Lindley, I., de Martin, R., and Binder, B. R. (2003). Direct binding of Nur77/NAK-1 to the plasminogen activator inhibitor 1 (PAI-1) promoter regulates TNF-{alpha}-induced PAI-1 expression. Blood 101, 3042–3048.[Abstract/Free Full Text]

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-{kappa}B-dependent pathways in cultured human Umbilical vein endothelial cells. J. Pharmacol. Exp. Ther. 307, 987–994.[Abstract/Free Full Text]

Hayashi, T. and Faustman, 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.[Abstract/Free Full Text]

Hewett, J. A., Jean, P. A., Kunkel, S. L., and Roth, R. A. (1993). Relationship between tumor necrosis factor-alpha and neutrophils in endotoxin-induced liver injury. Am. J. Physiol. 265, G1011–1015.[ISI][Medline]

Jin, S. L., and Conti, M. (2003). Induction of the cyclic nucleotide phosphodiesterase PDE4B is essential for LPS-activated TNF-alpha responses. Proc. Natl. Acad. Sci. 99, 7628–7633.[CrossRef][ISI]

Kietzmann, T., Jungermann, K., and Gorlach, A. (2003). Regulation of the hypoxia-dependent plasminogen activator inhibitor 1 expression by MAP kinases. Thromb. Haemost. 89, 666–673.[ISI][Medline]

Kietzmann, 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.[Abstract/Free Full Text]

Kobayashi, H., Horikoshi, K., Yamataka, A., Yamataka, T., Okazaki, T., Lane, G. J., and Miyano, T. (1999). Hyaluronic acid: A specific prognostic indicator of hepatic damage in biliary atresia. J. Pediatr. Surg. 34, 1791–1794.[ISI][Medline]

Li, C., Liu, J., Waalkes, M. P., and Jaeschke, H. (2003). Gene array analysis of the hepatic response to endotoxin in glutathione peroxidase-deficient mice. Toxicol. Lett. 144, 397–406.[CrossRef][ISI][Medline]

Lind, R. C., Gandolfi, A. J., Sipes, I. G., and Brown, B. R. J. (1984). The involvement of endotoxin in halothane-associated liver injury. Anesthesiology 61, 544–550.[ISI][Medline]

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[Abstract/Free Full Text]

Luyendyk, J. P., Maddox, J. F., Ganey, P. E., and Roth, R. A. (2004). Coagulation system activation and fibrin deposition are critical factors in idiosyncrasy-like liver injury from lipopolysaccharide and ranitidine coexposure (Abstract). FASEB J. (in press).

Marshall, L. J., Ramdin, L. S., Brooks, T., Charlton, P., and Shute, J. K. (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-1 complexes. J. Immunol. 171, 2057–2065.[Abstract/Free Full Text]

Mattes, W. B. (2004). Annotation and cross-indexing of array elements on multiple platforms. Environ. Health Perspect. 112, 506–510.[ISI][Medline]

Neville, L. F., Mathiak, G., and Bagasra, O. (1997). The immunobiology of interferon-{gamma}-inducible protein 10 kD (IP-10): a novel, pleiotropic member of the C-X-C chemokine superfamily. Cytokine Growth Factor Rev. 8, 207–219.[CrossRef][Medline]

Ozawa, K., Kuwabara, K., Tamatani, M., Takatsuji, K., Tsukamoto, Y., Kaneda, S., Yanagi, H., Stern, D. M., Eguchi, Y., Tsujimoto, Y., Ogawa, S., and Tohyama, M. (1999). 150-kDa oxygen-regulated protein (ORP150) suppresses hypoxia-induced apoptotic cell death. J. Biol. Chem. 274, 6397–6404.[Abstract/Free Full Text]

Pawlinski, R., Pedersen, B., Kehrle, B., Aird, W. C., Frank, R. D., Guha, M., and Mackman, N. (2003). Regulation of tissue factor and inflammatory mediators by Egr-1 in a mouse endotoxemia model. Blood 101, 3940–3947.[Abstract/Free Full Text]

Roth, R. A., Luyendyk, J. P., Maddox, J. F., and Ganey, P. E. (2003). Inflammation and drug idiosyncrasy—is there a connection? J. Pharmacol. Exp. Ther. 307, 1–8.[Abstract/Free Full Text]

Schnitt, S. J., Stillman, I. E., Owings, D. V., Kishimoto, C., Dvorak, H. F., and Abelmann, W. H. (1993). Myocardial fibrin deposition in experimental viral myocarditis that progresses to dilated cardiomyopathy. Circ. Res. 72, 914–920.[Abstract]

Shi, L., Kishore, R., McMullen, M. R., and Nagy, L. E. (2002). Lipopolysaccharide stimulation of ERK1/2 increases TNF-alpha production via Egr-1. Am. J. Physiol. Cell Physiol. 282, C1205–1211[Abstract/Free Full Text]

Thiel, G., and Cibelli, G. (2002). Regulation of life and death by the zinc finger transcription factor Egr-1. J. Cell Physiol. 193, 287–292.[CrossRef][ISI][Medline]

Yan, S. F., Lu, J., Zou, Y. S., Soh-Won, J., Cohen, D. M., Buttrick, P. M., Cooper, D. R., Steinberg, S. F., Mackman, N., Pinsky, D. J., and Stern, D. M. (1999). Hypoxia-associated induction of early growth response-1 gene expression. J. Biol. Chem. 274, 15030–15040.[Abstract/Free Full Text]

Zhao, B., Stavchansky, S. A., Bowden, R. A., and Bowman, P. D. (2003). Effect of interleukin-1ß and tumor necrosis factor-{alpha} on gene expression in human endothelial cells. Am. J. Physiol. Cell Physiol. 284, C1577–C1583[Abstract/Free Full Text]