ToxSci Advance Access originally published online on November 12, 2008
Toxicological Sciences 2009 107(2):544-552; doi:10.1093/toxsci/kfn237
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Parallelogram Approach Using Rat-Human In Vitro and Rat In Vivo Toxicogenomics Predicts Acetaminophen-induced Hepatotoxicity in Humans
,1
* Department of Health Risk Analysis and Toxicology, Maastricht University, Maastricht, The Netherlands
Business Unit Biosciences, TNO Quality of Life, Zeist, The Netherlands
Department of Surgery, University Hospital Maastricht and Nutrition and Toxicology Research Institute (NUTRIM), Maastricht University, Maastricht, The Netherlands
NIEHS Microarray Group, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709
1 To whom correspondence should be addressed at Department of Health Risk Analysis and Toxicology, University of Maastricht, PO Box 616, 6200 MD Maastricht, The Netherlands. Fax: +31 43 388 4146. E-mail: anne.kienhuis{at}rivm.nl.
Received September 6, 2008; accepted November 6, 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
The frequent use of rodent hepatic in vitro systems in pharmacological and toxicological investigations challenges extrapolation of in vitro results to the situation in vivo and interspecies extrapolation from rodents to humans. The toxicogenomics approach may aid in evaluating relevance of these model systems for human risk assessment by direct comparison of toxicant-induced gene expression profiles and infers mechanisms between several systems. In the present study, acetaminophen (APAP) was used as a model compound to compare gene expression responses between rat and human using in vitro cellular models, hepatocytes, and between rat in vitro and in vivo. Comparison at the level of modulated biochemical pathways and biological processes rather than at that of individual genes appears preferable as it increases the overlap between various systems. Pathway analysis by T-profiler revealed similar biochemical pathways and biological processes repressed in rat and human hepatocytes in vitro, as well as in rat liver in vitro and in vivo. Repressed pathways comprised energy-consuming biochemical pathways, mitochondrial function, and oxidoreductase activity. The present study is the first that used a toxicogenomics-based parallelogram approach, extrapolating in vitro to in vivo and interspecies, to reveal relevant mechanisms indicative of APAP-induced liver toxicity in humans in vivo.
Key Words: hepatocyte-based in vitro models; liver injury; acetaminophen; interspecies extrapolation; gene expression profiling; T-profiler.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Toxicological and pharmacological assessments of hepatotoxic risks of chemical compounds and drugs in humans rely on extrapolation from animal experiments. Increasingly, studies in well-established hepatic in vitro assays, such as primary hepatocytes, liver slices, and hepatic cell lines, precede in vivo animal experiments in an attempt to identify potential hepatotoxicity in early stages of toxicity testing and to decrease attrition rates of drugs during lead optimization (Dambach et al., 2005
). The general use of rodent hepatic in vitro systems challenges not only the extrapolation of in vitro results to the situation in vivo (Guillouzo, 1998) but also interspecies extrapolation from rodents to humans (Kern et al., 1997; LeCluyse, 2001; Maurel, 1996).
One of the most extensively studied hepatotoxicants is acetaminophen or paracetamol (N-acetyl-para-aminophenol; APAP). APAP is converted by cytochrome P450 (CYP) enzymes (CYP2E1, CYP3A, and to a lesser extent CYP1A2; Wolf et al., 2007) to the toxic metabolite N-acetyl-p-benzoquinone imine (NAPQI). NAPQI is rapidly metabolized by conjugation to intracellular glutathione (GSH) forming a nontoxic APAP-GSH conjugate, which is excreted in the urine (Bessems and Vermeulen, 2001; Jaeschke and Bajt, 2006; Wolf et al., 2007). APAP overdose or excessive CYP activation may lead to depletion of GSH stores and, consequently, the inability to detoxify reactive metabolites. Despite the numerous experiments performed to identify the toxic mode of action of APAP, the exact mechanism by which this compound induces liver injury is not completely clear. In the recent past, numerous studies showed the suitability of toxicogenomics and proteomics approaches for unraveling mechanisms of liver cell injury in animal models exposed to APAP. Common findings in these studies are gene expression patterns reflecting cellular energy loss and mitochondrial damage in rats and mice after APAP intoxication (Beyer et al., 2007; Heinloth et al., 2004; Kikkawa et al., 2006; Ruepp et al., 2002).
Besides contributing to the identification of the mode of action of compounds, systems biology approaches like toxicogenomics would be particularly suitable for comparative toxicology, e.g., on effects induced by different drugs or compounds, comparison of in vitro systems, of species, or of the in vitro and in vivo situation, as toxicogenomics allows comparison of thousands of gene expression modifications. Toxicogenomics studies have previously been performed for the purpose of identifying (in)consistencies in responses between APAP and other compounds inducing similar end points of hepatotoxicity (de Longueville et al., 2003), between subtoxic and overtly toxic doses (Heinloth et al., 2004), and between primary human hepatocytes and HepG2 cells (Harris et al., 2004). Recently, a multicenter study of APAP hepatotoxicity was performed to uncover robust genomic signatures of APAP-induced toxicity in mice (Beyer et al., 2007). Furthermore, in a recent study, harmful levels of APAP exposure were successfully predicted in humans by interspecies comparison of gene expression data measured in blood cells of APAP-exposed rats and humans (Bushel et al., 2007). None of these studies seized the challenge to fully compare hepatotoxic effects caused by APAP both in vitro and in vivo as well as in rodent and human, in order to predict mechanisms of human liver injury. The parallelogram approach, originally introduced by Sobels in the late seventies (Sobels, 1977, 1984, 1989), can be applied to extrapolate these toxicogenomics results from in vitro to in vivo and interspecies in order to estimate toxicity which cannot be assessed directly.
The aims of the present study therefore were (1) to compare APAP-induced gene expression profiles from rat liver cells in vitro with responses to APAP in the liver in vivo, (2) to perform interspecies extrapolation based upon rat-human in vitro models, and (3) to combine these findings in order to predict mechanistic changes that may occur in vivo in man, using a toxicogenomics-based parallelogram approach. For this purpose, APAP-induced gene expression profiles in sandwich-cultured primary human and rat hepatocytes were compared to a previously published rat in vivo study (Gene Expression Omnibus [GEO] database [Heinloth et al., 2007], see "Materials and Methods" section).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Chemicals.
Ca2+- and Mg2+-free Hanks’ buffered salt solution (HBSS), Dubecco's minimal essential medium (DMEM), fetal calf serum (FCS), penicillin-streptomycin, PBS, and TRIzol were obtained from Invitrogen, Breda, The Netherlands. Bovine serum albumin (BSA), ascorbic acid, collagenase type IV, insulin, dexamethasone (DEX), Percoll, phenobarbital (PB), β-naphthoflavone (β-NF), and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were purchased form Sigma-Aldrich, Zwijndrecht, The Netherlands. BD ITS+ Premix and rat-tail collagen I were obtained from BD Biosciences, Alphen aan den Rijn, The Netherlands. The RNeasy MinElute kit and RNeasy mini spin columns were obtained from Qiagen, Westburg B.V. (Leusden, The Netherlands). Cyanine 5-cytosine triphosphate (CTP) and cyanine 3-CTP were purchased from PerkinElmer (Boston, MA). Agilent's low RNA input fluorescent linear amplification kit and the hybridization solution were obtained from Agilent Technologies (Palo Alto, CA). APAP, CAS-no 103-90-2, purity minimum 99% according to manufacturer, was obtained from Sigma-Aldrich. All other chemicals were of analytical grade.
Culture of human hepatocytes.
Human hepatocytes were obtained from resection specimens of patients undergoing partial hepatectomy for colorectal metastases in an otherwise normal liver at the University Hospital Maastricht, The Netherlands. Surgery was performed as described before (Dejong and Garden, 2003), and in none of the cases hepatic inflow occlusion was applied during liver transaction. All patients provided written informed consent, and the study was approved by the Medical Ethics Committee of the University Hospital Maastricht. Isolation of human hepatocytes from resection specimens was performed according to the method described by LeCluyse et al. (2005). This method has been adopted by an interlaboratory consortium sponsored by the European Centre for the Validation of Alternative Methods for the isolation and cultivation of primary human hepatocytes for testing the potential of new drugs to induce liver enzyme expression (LeCluyse et al., 2005; Richert et al., 2004). Essentially, directly after removal of the liver tissue, an encapsulated wedge, weighing approximately 50–100 g, was cut off. Blood vessels on the cut surface were immediately flushed with ice-cold wash buffer consisting of Ca2+- and Mg2+-free HBSS supplemented with 0.5mM EDTA, 0.5% (wt/vol) BSA, and 50 µg/ml ascorbic acid. Liver tissue was transported to the laboratory in ice-cold buffer. Transportation time did not exceed 15 min. After a two-step collagenase perfusion, first with wash buffer, then with digestion medium consisting DMEM supplemented with 0.05% (wt/vol) collagenase type IV and 0.5% BSA, hepatocytes were dispersed from the digested liver and collected in ice-cold attachment medium (5% FCS and penicillin-streptomycin [100 U/ml and 100 µg/ml, respectively], 0.1 U/ml insulin, and 1µM DEX in DMEM). After passing of the cell suspension through a 100-µm nylon mesh, the suspension was washed by low-speed centrifugation three times at 75 x g for 5 min. Pellets were resuspended in ice-cold suspension medium after each centrifugation step. Viability was assessed by trypan blue exclusion. Hepatocyte preparations with viability greater than 75% were included for further studies. Cell suspension with viability below 85% was purified using a Percoll gradient, as previously described (LeCluyse et al., 2005). Cells were cultured on collagen gel–precoated 12-well plates at a density of 6.5 x 105 cells per well. Human hepatocyte sandwich cultures were essentially prepared according to the method of Beken et al. (2004). After attachment for 4 h in attachment medium, dead cells were removed by washing and the upper collagen layer was applied. Thereafter, cells were kept in DMEM containing 0.1µM DEX, 6.25 µg/ml insulin, 6.25 µg/ml transferin, and 6.25 ng/ml selenium (BD ITS+ Premix) (Hamilton et al., 2001). Cultures were incubated at 37°C in a humidified incubator gassed with 5% CO2 in air. Medium was changed on a daily basis during a period of 72 h.
Culture of rat hepatocytes.
Male Wistar rats (Crl: (WI) WU BR), 9–12 weeks of age, 180–250 g, were obtained from Charles River GmbH, Sulzfeld, Germany. During the acclimatization period and until sacrifice, animals were housed individually in macrolon cages with wire tops and sawdust bedding at 22°C and 50–60% humidity. The light cycle was 12 h light/12 h dark. Feed and tap water were available ad libitum. Hepatocytes were isolated according to a two-step collagenase perfusion technique as described by Seglen (1976) with minor modifications. Hepatocyte preparations with viability greater than 85% as determined by trypan blue exclusion were used and cultured on collagen gel–precoated 6-well plates at a density of 1.3 x 106 cells per well. Sandwich cultures of rat hepatocytes were prepared similar to sandwich cultures of human hepatocytes as described above with culture conditions as described previously (Kienhuis et al., 2006). An inducer mix used to increase the metabolic competence of conventional rat sandwich cultures consisted of 1mM PB, 10µM DEX, and 5µM β-NF (Kienhuis et al., 2007). PB was added as a concentrated stock solution in PBS. DEX and β-NF were added as concentrated stock solutions in dimethylsulfoxide (DMSO). The final concentration of DMSO was equalized in all culture media and did not exceed 0.2% (vol/vol). Cultures were incubated at 37°C in a humidified incubator gassed with 5% CO2 in air. Medium was changed on a daily basis during a period of 72 h.
Rat in vivo study.
In vivo data from male F344/N rats, 8–12 weeks old, were retrieved from the GEO database on the NCBI Web site (NCBI), series GSE5860
[NCBI GEO]
, submitted by Heinloth and Paules, 19 September 2006 (Heinloth et al., 2007). The data represent log10 expression ratios of messenger RNA (mRNA) from livers from rats single dosed by oral gavage with 1.5 g/kg body weight APAP in 0.5% ethyl cellulose hybridized against liver mRNA from rats oral gavaged with vehicle only. Three rats were used per APAP-treated and vehicle group. Animals were not fasted prior to dosing with APAP. At 24 h following dosing, animals were exposed to carbon dioxide (CO2/O2) mixture from a regulated source and livers were rapidly removed. A section from the mid-sagittal section of the left lobe was taken for histopathological evaluation, and the remainder of the left lobe was rapidly cubed (approximately 0.5 x 0.5 x 0.5 cm samples) and frozen in liquid nitrogen. Experiments were performed according to the guidelines established in the National Institute of Health Guide for the Care and Use of Laboratory Animals, and an approved animal study protocol was on file prior to the initiation of the study. For more information, refer to Heinloth et al. (2007).
Hepatocyte treatment and cytotoxicity analysis.
After 72 h of culture, five independent human hepatocyte cultures and three independent rat hepatocyte cultures were exposed to two concentrations of APAP (5 and 10mM) for 24 h. APAP was dissolved in culture medium. Control cultures were maintained in medium only. Cytotoxicity was determined by the MTT reduction method on hepatocytes from subjects 4, 5, and 6 (Table 1) and on all rat hepatocyte cultures either conventionally cultured (standard culture) or cultured in medium to increase the metabolic competence (modified culture). The in vivo dose and in vitro concentrations selected for the present study to enable comparison were based on results from previous studies comparing APAP profiles (de Longueville et al., 2003; Harris et al., 2004; Heinloth et al., 2004, 2007).
|
Total RNA extraction.
Following removal of culture medium of sandwich-cultured human and rat hepatocytes, Trizol was added onto the upper collagen layer and cells were collected. RNA was purified using the RNeasy MinElute kit including an additional DNA digestion step. RNA concentrations were determined spectrophotometrically by UV absorbance at 260 nm. RNA quality was determined using the Bioanalyzer (Agilent Technologies). All samples contained intact total RNA with a 28S/18S ribosomal RNA ratio >1.5 and an RIN number >8. RNA extractions of hepatocyte cultures exposed to APAP of five human individuals and three rats were used for microarray analysis. For the rat in vivo study, total liver RNA was later isolated using Qiagen RNeasy Maxi Kits and equal amounts of control animal RNA from each time were pooled and compared with liver RNA samples from individual APAP-treated rats at each dose for that time point.
Labeling.
RNA samples from control hepatocyte cultures and cultures exposed to APAP were labeled with cyanine 5-CTP. Cyanine 5-CTP–labeled samples from one individual (rat or human) were hybridized against cyanine 3-CTP–labeled RNA samples from control hepatocyte cultures of the same individual. As a result, control samples labeled with cyanine 5-CTP hybridized against cyanine 3-CTP samples can be considered as a self-self hybridization. Labeling was performed using Agilent's low RNA input fluorescent linear amplification kit following manufacturer's instruction. Briefly, double-stranded complementary DNA was synthesized using moloney murine leukemia virus-reverse transcriptase with T7 promoter primer, starting with 1 µg of total RNA. Cyanine-labeled complementary RNA (cRNA) targets were transcribed using T7 RNA polymerase. The amplified cRNA was purified using RNeasy mini spin columns. Synthesized cRNA products were quantified spectrophotometrically.
Hybridization.
For microarray hybridization, cyanine 5–labeled samples and cyanine 3–labeled samples were combined. cRNAs were fragmented at 60oC for 30 min with fragmentation solution followed by hybridization on Agilent 22 K format 60-mer oligo microarrays (G4130A for rat and G4110B for human from Agilent Technologies) for 17 h at 60oC with Agilent hybridization solution. Arrays were washed according to manufacturer's instruction. Microarrays were scanned using a Packard Scanarray Express confocal laser scanner (PerkinElmer). Resulting tagged image file format images were loaded into Imagene 5.0 (Biodiscovery Inc., El Segundo, CA) to further process and collect the gene expression data. For the rat in vivo study, hybridization was performed in a dye-swap design on Agilent Rat Oligo Microarrays G4130A.
Data analysis.
Data were transferred to GeneSight 4.1 (Biodiscovery Inc.). Flagged spots, consisting of poor quality spots and negative and positive control spots, were excluded. For each spot, median local background intensity was subtracted from the median spot intensity, and spots from low expression genes (with a net intensity of <40 in both channels) were excluded from further analysis. These background-corrected median intensities were log-transformed by base 2. Data were normalized using the Lowess algorithm (Yang et al., 2002). Resulting gene expression ratios were loaded into Excel (Microsoft Corporation, Redmond, WA). For further analysis, 60% of values per gene had to be available. For rat and human in vitro studies, a subset of significantly modulated genes was selected by performing a Student's t-test (p < 0.05) on genes with mean fold changes above or below a threshold of 1.5 compared to control cultures (Guo et al., 2006). Similar human and rat genes, orthologs, were identified based on gene symbol or based on orthologs as provided by Resourcerer database of The Computational Biology and Functional Genomics Laboratory at the Dana-Farber Cancer Institute and Harvard School of Public Health (Resourcerer). For the rat in vivo study, we averaged the log10 expression data of dye-swap array pairs and calculated fold changes. Genes with fold changes above or below a threshold of 1.5 were considered modulated. Modulated genes were considered significant when average log10 expression values minus two times their SD were greater than zero. The rat and human in vitro and the rat in vivo data sets are publically available in the GEO database on the NCBI. Accession numbers for the rat in vitro, human in vitro, and rat in vivo data are GSE13465
[NCBI GEO]
..., GSE13430
[NCBI GEO]
, and GSE5860
[NCBI GEO]
, respectively.
T-profiler (Boorsma et al., 2005) was used to identify transcriptional regulation of biochemical pathways and biological processes in the complete data set of genes without preselecting only significantly modulated genes. T-profiler uses the t-test to score the difference between the mean expression level of predefined groups of genes and that of all other genes within the complete data set of genes (Boorsma et al., 2005). To determine significance, a Bonferroni corrected p value, E value, is generated. Pathways and processes were significant when E values were below 0.05. Pathways and processes provided by T-profiler included GO terms (www.geneontology.org), KEGG pathways (http://www.genome.jp/kegg/pathway.html), and gene sets including GenMAPP pathways (http://www.genmapp.org), gene sets from BioCarta (http://www.biocarta.com), T-profiler–curated (manually curated) gene sets, and gene sets retrieved from literature. A T-profiler web-based tool that allows for analysis of rat, mouse, and human transcriptional data is currently under development.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Cytotoxicity was determined using the MTT reduction assay. After exposure to 5mM APAP, no cytotoxicity was observed in human and rat hepatocytes, the latter maintained in either standard or modified culture medium. Exposure to 10mM APAP resulted in on average 10% loss in viability in human hepatocytes and in on average 20% loss in viability in rat hepatocytes cultured in modified culture medium. At 10mM, no cytotoxicity was observed in rat hepatocytes cultured in standard culture medium, without enzyme inducers. The in vivo dose of 1.5 g/kg body weight from which gene expression data were selected was overtly hepatotoxic as indicated with histopathology and clinical chemistry (Heinloth et al., 2004, 2007
). Treatment of human hepatocytes with 5 and 10mM APAP resulted in significant modulation of 1624 genes (991 and 1401 modulated genes after treatment with 5 or 10mM, respectively; all 767 overlapping genes were regulated in the same directions). APAP treatment resulted in significant modulation of 368 genes in rat hepatocytes cultured in standard medium (208 and 207 genes modulated after treatment with 5 or 10mM, respectively; 47 genes in the overlap were, with two exceptions, regulated in similar directions), whereas 1289 genes were significantly changed in rat hepatocytes cultured in modified medium (327 and 1132 modulated genes after treatment with 5 or 10mM, respectively; all 172 genes in the overlap were in the same directions). In vivo treatment of rats with 1.5 g/kg body weight APAP resulted in significant modulation of 1349 genes.
Overlap of significantly modulated genes between rats in vivo and rat hepatocytes in vitro is shown in Figure 1. Overlap is highest between rat hepatocytes cultured in standard and modified medium. When comparing with in vivo, more commonly modulated genes were found in rat hepatocytes cultured in modified medium as compared to those cultured in standard medium. However, only 18 of 43 genes common between the modified system and in vivo are regulated in the same directions. Furthermore, only two of the five significantly modulated genes in the standard system and in vivo are regulated in a similar direction. This indicates that rat hepatocytes cultured in modified medium better represent effects induced by APAP in rats in vivo in contrast to the cells cultured in standard medium. The two genes significantly modulated by all three systems are downregulated in vitro and upregulated in vivo.
|
In order to compare the gene expression results in rat and human hepatocytes, only rat-human orthologs were used, based on gene symbol and the Resourcerer database. The Venn diagram in Figure 2 presents the number of significantly modulated genes between human hepatocytes and rat hepatocytes cultured in modified medium (comparison with those cultured in standard medium are also done but not shown as less overlap was observed) and between rat hepatocytes cultured in modified medium and rats in vivo. These results show that overlap is highest between human hepatocytes and rat hepatocytes in modified culture medium and that most genes in the overlap are regulated in the same direction in both species.
|
Application of T-profiler, a pathway analysis method for which no preselection of subsets of modulated genes is needed (Boorsma et al., 2005
), to the data sets resulted in identification of large numbers of biochemical/cellular pathways and biological processes based on gene sets that were differentially expressed, thereby enabling thorough functional comparison between systems. All differentially expressed biochemical/cellular pathways identified by T-profiler in rats in vivo and rat and human hepatocytes in vitro can be viewed online at http://ntc.voeding.tno.nl/tbase/acetaminophen/. Tables 2–4 show pathways and processes enriched in at least two systems: human hepatocytes, rat hepatocytes (either standard or modified), or rats in vivo. Eighteen predominant biochemical pathways and biological processes were identified in rat hepatocytes cultured in standard medium, of which eight were found in at least one other system. Six of these eight pathways were regulated in opposite directions compared to regulation in any other system. The correspondence of APAP-induced biochemical pathways and biological processes was much better between the other systems. Therefore, the rat hepatocytes cultured in standard medium is excluded in the following presentation. The Venn diagram in Figure 3 shows the number of biochemical pathways and biological processes as modified by APAP and identified by T-profiler per system and the overlap thereof between human hepatocytes, rat hepatocytes cultured in modified medium, and rats in vivo. Comparisons were made regardless of the concentrations/doses of APAP. Overlap between human hepatocytes and rat hepatocytes cultured in modified medium is highest. The six pathways and processes in the center of the Venn diagram included metabolism in general, lipid and fatty acid metabolism, electron transport, processes occurring in peroxisomes, and degradation of branched-chain amino acids (valine, leucine, and isoleucine degradation).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
The present study further contributes to the information on acute toxicity of APAP as it presumably occurs in the human in vivo situation by comparison of in vitro and in vivo toxicogenomics data as well as interspecies toxicogenomics results. To this purpose, sandwich-cultured human and rat hepatocytes were treated with APAP and the resulting gene expression profiles were compared with gene expression profiles in livers of rats following exposure to APAP in vivo using published data (Heinloth et al., 2007
).
Our findings showed that more genes were similarly modulated in rat hepatocytes cultured in modified medium and rats in vivo as compared to rat hepatocytes cultured in standard medium and rats in vivo. These results are concordant with previous findings which already showed that the modified medium increased the metabolic competence of the rat hepatocyte system (Kienhuis et al., 2007), resulting in a more vivo-like response to compounds compared to the response in rat hepatocytes cultured in standard medium (Kienhuis et al., 2006). Specifically, gene expression as well as activity of the enzymes responsible for conversion of APAP to its toxic metabolites, CYP2E1, CYP3A, and, to a lesser extent, CYP1A2 (Wolf et al., 2007), demonstrated to be better preserved in rat hepatocytes cultured in modified medium compared to rats in vivo (Kienhuis et al., 2007). In the present study, correspondence between the number of genes significantly expressed and in the direction of regulation in rat hepatocytes in vitro and rats in vivo might have been higher when similar rat strains would have been compared. However, the use of different rat strains, Wistar for in vitro and F344 for in vivo, presumably increases robustness. Furthermore, both strains have been used frequently for APAP hepatotoxicity studies and have not been reported to differ in CYP enzyme activity responsible for APAP conversion.
In vitro-in vivo as well as interspecies extrapolation of APAP toxicogenomics data is complicated based upon the outcome of listings of differentially expressed single genes, even though these genes may share similar biology. Pathway analysis instead allows translation of these genes to a common and therefore easy to compare vocabulary (Ashburner et al., 2000; Currie et al., 2005). In particular, a major limitation of interspecies comparison between rats and humans at the modulated gene level is the poor annotation of the rat genome. Due to this, for only 30% of the rat genes, ortholog identification was available on the human Agilent microarrays. Therefore, in the present study, GO terms, pathway, and gene set analysis, all ways to compile individual gene information into biological categories, were used to compare biology between in vitro and in vivo (rat hepatocytes in vitro and rats in vivo) and between species (human and rat hepatocytes). Our results may be improved if better annotation and ortholog identification are available.
The preference of pathway analysis above single gene analysis is demonstrated in the present study, as in contrast to results based on single gene comparison (Fig. 2), pathway analysis indicates that APAP effects in rat hepatocytes cultured in modified medium are quite similar to effects in rats in vivo, at least concerning those pathways differentially expressed (Fig. 3). Moreover, T-profiler pathway analysis includes not only genes that were categorized as differentially expressed based on statistics but also uses all genes within the complete data set to statistically evaluate the distribution of gene expression of predefined groups of genes (GO terms, KEGG pathways, and gene sets retrieved from GenMAPP, BioCarta, etc.), in comparison to the distribution of expression of all genes on the array. The advantage of such a continuous pathway analysis approach is that a group of genes can be scored as significantly up- or downregulated even if none of its individual member genes are significantly modulated (Boorsma et al., 2005).
Following the toxicogenomics-based parallelogram approach, we identified six biological pathways and processes that are relevant to humans in vivo, as they were differentially expressed upon APAP treatment in rat and human hepatocytes in vitro as well as in rat in vivo. These six pathways include lipid metabolism, among which fatty acid metabolism occurs in peroxisomes, degradation of the branched-chain amino acids valine, leucine, and isoleucine, and electron transport in mitochondria and were repressed after APAP intoxication. Repression of these processes is strongly associated with loss of hepatocellular energy production. Furthermore, mitochondrial function was clearly impaired in rat and human hepatocytes in vitro, and oxidoreductase activity and related pathways including CYP activity were downregulated in rat and human hepatocytes in vitro and rat in vitro-in vivo. The major energy-consuming pathways, lipid metabolism and electron transport, were found to be robust genomic signatures of APAP-induced toxicity, repressed in mice treated with acute toxic doses of APAP in a multicenter study (Beyer et al., 2007). Moreover, the loss of mitochondrial function and concomitant generation of oxidative stress, as found in the present study, have previously been observed in rodent studies upon exposure to toxic doses of APAP (Beyer et al., 2007; Burcham and Harman, 1991; Heinloth et al., 2004; Ruepp et al., 2002) and were proposed to be mechanisms of APAP-induced hepatotoxicity following toxic doses in humans as well (Bessems and Vermeulen, 2001; Jaeschke and Bajt, 2006; James et al., 2003). Our observations that, unlike other studies, include effects on human hepatocytes in vitro thus support the hypothesis that APAP-induced effects on energy-consuming biochemical pathways, mitochondria, and oxidoreductase activity are likely to drive liver injury in humans and therefore contribute to further unraveling the mechanism of APAP hepatotoxicity.
The toxicogenomics-based parallelogram approach used in the present study, however, holds some limitations for human risk assessment, especially concerning the use of a hepatocyte-based in vitro model for interspecies extrapolation. The use of such a model limits identification of APAP-induced responses which involve other cell types. For example, there is a growing body of evidence suggesting the role of inflammation in APAP-induced hepatotoxicity (Larson, 2007). This effect has recently been identified by gene expression profiling in blood cells of APAP-intoxicated rats and humans (Bushel et al., 2007). Development and propagation of hepatocyte injury are in this case mediated by infiltration of mononuclear cells, macrophages, Kupffer cells, and neutrophils, cell types absent in hepatocyte-based in vitro models.
Moreover, attention must be paid to the difference in pharmacokinetics of APAP between in vitro and in vivo, e.g., bioavailability of APAP is more continuous in vitro compared to in vivo. As a consequence, the so-called GSH threshold at which GSH is depleted which leaves NAPQI to induce hepatic injury might occur at different time points in vitro and in vivo (Slikker et al., 2004). Due to decrease of bioavailability of APAP in time in vivo and continuous exposure in vitro, it might have been of interest to compare the 24-h in vitro data with, e.g., 6-h in vivo data made available by Heinloth et al. (2007). Therefore, we additionally analyzed T-profiler pathways in rats in vivo treated with 1.5 g/kg body weight at the 6-h time point (data not shown, but available at http://ntc.voeding.tno.nl/tbase/acetaminophen/). However, no KEGG pathways and gene sets and only four GO terms were significant at this time point. None of these GO terms were similarly significantly expressed in other systems.
Furthermore, the present study provides no quantitative assessment of human hepatotoxic risks, as doses of APAP in rats in vivo are one order of magnitude beyond the generally accepted hepatotoxic dose for an average adult within 24 h, which is 150 mg/kg. However, factors that influence CYP activity (e.g., genetic variability, food constituents, alcohol) might considerably increase APAP hepatotoxicity and decrease the hepatotoxic dose, even to a level below the maximum therapeutic dose of 4 g per 24 h (Larson et al., 2005). Therefore, liver injury observed in the present study might still be representative for certain situations in humans.
In conclusion, the present study is the first that used a toxicogenomics-based parallelogram approach using in vitro to in vivo as well as interspecies extrapolations to reveal relevant mechanisms of APAP-induced liver toxicity in humans in vivo. Gene expression profiling combined with T-profiler pathway analysis uncovered impairment of energy-consuming biochemical pathways and biological processes, mitochondrial function, and oxidoreductase activity as most relevant for hepatotoxicity in humans.
|
FUNDING |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
The Netherlands Organization for Health Research and Development, program Alternatives to Animal Experiments (3170.0049); the Dutch Ministry of Economic Affairs.
|
ACKNOWLEDGMENTS |
|---|
The authors would wish to thank Dr R. Deligt, Dr C. Krul, and M. Schut for providing primary rat hepatocytes.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONFUNDINGREFERENCES |
|---|
Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al. Gene ontology: Tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. (2000) 25:25–29.[CrossRef][Web of Science][Medline]
Beken S, Vanhaecke T, De Smet K, Pauwels M, Vercruysse A, Rogiers V. Collagen-gel cultures of rat hepatocytes: Collagen-gel sandwich and immobilization cultures. In: Cytochrome P450 Protocols—Phillips IR, Shephard EA, eds. (2004) Humana Press Inc. Totowa, NJ.
Bessems JG, Vermeulen NP. Paracetamol (acetaminophen)-induced toxicity: Molecular and biochemical mechanisms, analogues and protective approaches. Crit. Rev. Toxicol. (2001) 31:55–138.[CrossRef][Web of Science][Medline]
Beyer RP, Fry RC, Lasarev MR, McConnachie LA, Meira LB, Palmer VS, Powell CL, Ross PK, Bammler TK, Bradford BU, et al. Multicenter study of acetaminophen hepatotoxicity reveals the importance of biological endpoints in genomic analyses. Toxicol. Sci. (2007) 99:326–337.
Boorsma A, Foat BC, Vis D, Klis F, Bussemaker HJ. T-profiler: Scoring the activity of predefined groups of genes using gene expression data. Nucleic Acids Res. (2005) 33:W592–W595.
Burcham PC, Harman AW. Acetaminophen toxicity results in site-specific mitochondrial damage in isolated mouse hepatocytes. J. Biol. Chem. (1991) 266:5049–5054.
Bushel PR, Heinloth AN, Li J, Huang L, Chou JW, Boorman GA, Malarkey DE, Houle CD, Ward SM, Wilson RE, et al. Blood gene expression signatures predict exposure levels. Proc. Natl. Acad. Sci. U.S.A. (2007) 104:18211–18216.
Currie RA, Orphanides G, Moggs JG. Mapping molecular responses to xenoestrogens through gene ontology and pathway analysis of toxicogenomic data. Reprod. Toxicol. (2005) 20:433–440.[CrossRef][Web of Science][Medline]
Dambach DM, Andrews BA, Moulin F. New technologies and screening strategies for hepatotoxicity: Use of in vitro models. Toxicol. Pathol. (2005) 33:17–26.
Dejong CHC, Garden OJ. Neoplasms in the liver. In: Advanced Surgical Practice—Majid AA, Kingsnorth A, eds. (2003) London: Greenwich Medical Media. 146–156.
de Longueville F, Atienzar FA, Marcq L, Dufrane S, Evrard S, Wouters L, Leroux F, Bertholet V, Gerin B, Whomsley R, et al. Use of a low-density microarray for studying gene expression patterns induced by hepatotoxicants on primary cultures of rat hepatocytes. Toxicol. Sci. (2003) 75:378–392.
Guillouzo A. Liver cell models in in vitro toxicology. Environ. Health Perspect. (1998) 106(Suppl. 2):511–532.[CrossRef][Web of Science][Medline]
Guo L, Lobenhofer EK, Wang C, Shippy R, Harris SC, Zhang L, Mei N, Chen T, Herman D, Goodsaid FM, et al. Rat toxicogenomic study reveals analytical consistency across microarray platforms. Nat. Biotechnol. (2006) 24:1162–1169.[CrossRef][Web of Science][Medline]
Hamilton GA, Jolley SL, Gilbert D, Coon DJ, Barros S, LeCluyse EL. Regulation of cell morphology and cytochrome P450 expression in human hepatocytes by extracellular matrix and cell-cell interactions. Cell Tissue Res. (2001) 306:85–99.[CrossRef][Web of Science][Medline]
Harris AJ, Dial SL, Casciano DA. Comparison of basal gene expression profiles and effects of hepatocarcinogens on gene expression in cultured primary human hepatocytes and HepG2 cells. Mutat. Res. (2004) 549:79–99.[Web of Science][Medline]
Heinloth AN, Boorman GA, Foley JF, Flagler ND, Paules RS. Gene expression analysis offers unique advantages to histopathology in liver biopsy evaluations. Toxicol. Pathol. (2007) 35:276–283.
Heinloth AN, Irwin RD, Boorman GA, Nettesheim P, Fannin RD, Sieber SO, Snell ML, Tucker CJ, Li L, Travlos GS, et al. Gene expression profiling of rat livers reveals indicators of potential adverse effects. Toxicol. Sci. (2004) 80:193–202.
Jaeschke H, Bajt ML. Intracellular signaling mechanisms of acetaminophen-induced liver cell death. Toxicol. Sci. (2006) 89:31–41.
James LP, Mayeux PR, Hinson JA. Acetaminophen-induced hepatotoxicity. Drug Metab. Dispos. (2003) 31:1499–1506.
Kern A, Bader A, Pichlmayr R, Sewing KF. Drug metabolism in hepatocyte sandwich cultures of rats and humans. Biochem. Pharmacol. (1997) 54:761–772.[CrossRef][Web of Science][Medline]
Kienhuis AS, Wortelboer HM, Hoflack JC, Moonen EJ, Kleinjans JC, van Ommen B, van Delft JH, Stierum RH. Comparison of coumarin-induced toxicity between sandwich-cultured primary rat hepatocytes and rats in vivo: A toxicogenomics approach. Drug Metab. Dispos. (2006) 34:2083–2090.
Kienhuis AS, Wortelboer HM, Maas WJ, van Herwijnen M, Kleinjans JC, van Delft JH, Stierum RH. A sandwich-cultured rat hepatocyte system with increased metabolic competence evaluated by gene expression profiling. Toxicol. In Vitro (2007) 21:892–901.[CrossRef][Web of Science][Medline]
Kikkawa R, Fujikawa M, Yamamoto T, Hamada Y, Yamada H, Horii I. In vivo hepatotoxicity study of rats in comparison with in vitro hepatotoxicity screening system. J. Toxicol. Sci. (2006) 31:23–34.[CrossRef][Medline]
Larson AM. Acetaminophen hepatotoxicity. Clin. Liver Dis (2007) 11:525–548. vi.[CrossRef][Medline]
Larson AM, Polson J, Fontana RJ, Davern TJ, Lalani E, Hynan LS, Reisch JS, Schiodt FV, Ostapowicz G, Shakil AO, et al. Acetaminophen-induced acute liver failure: Results of a United States multicenter, prospective study. Hepatology (2005) 42:1364–1372.[CrossRef][Web of Science][Medline]
LeCluyse EL. Human hepatocyte culture systems for the in vitro evaluation of cytochrome P450 expression and regulation. Eur. J. Pharm. Sci. (2001) 13:343–368.[CrossRef][Web of Science][Medline]
LeCluyse EL, Alexandre E, Hamilton GA, Viollon-Abadie C, Coon DJ, Jolley S, Richert L. Isolation and culture of primary human hepatocytes. Methods Mol. Biol. (2005) 290:207–229.[Medline]
Maurel P. The use of adult human hepatocytes in primary culture and other in vitro systems to investigate drug metabolism in man. Adv. Drug Delivery Rev. (1996) 22:105–132.[CrossRef][Web of Science]
Richert L, Alexandre E, Lloyd T, Orr S, Viollon-Abadie C, Patel R, Kingston S, Berry D, Dennison A, Heyd B, et al. Tissue collection, transport and isolation procedures required to optimize human hepatocyte isolation from waste liver surgical resections. A multilaboratory study. Liver Int. (2004) 24:371–378.[CrossRef][Web of Science][Medline]
Ruepp SU, Tonge RP, Shaw J, Wallis N, Pognan F. Genomics and proteomics analysis of acetaminophen toxicity in mouse liver. Toxicol. Sci. (2002) 65:135–150.
Seglen PO. Preparation of isolated rat liver cells. Methods Cell Biol. (1976) 13:29–83.[Medline]
Slikker W Jr, Andersen ME, Bogdanffy MS, Bus JS, Cohen SD, Conolly RB, David RM, Doerrer NG, Dorman DC, Gaylor DW, et al. Dose-dependent transitions in mechanisms of toxicity: Case studies. Toxicol. Appl. Pharmacol. (2004) 201:226–294.[CrossRef][Web of Science][Medline]
Sobels FH. Some problems associated with the testing for environmental mutagens and a perspective for studies in "comparative mutagenesis". Mutat. Res. (1977) 46:245–260.[Web of Science][Medline]
Sobels FH. Problems and perspectives in genetic toxicology. In: Mutations in Man—Obe G, ed. (1984) Berlin: Springer. 1–19.
Sobels FH. Models and assumptions underlying genetic risk assessment. Mutat. Res. (1989) 212:77–89.[Web of Science][Medline]
Wolf KK, Wood SG, Allard JL, Hunt JA, Gorman N, Walton-Strong BW, Szakacs JG, Duan SX, Hao Q, Court MH, et al. Role of CYP3A and CYP2E1 in alcohol-mediated increases in acetaminophen hepatotoxicity: Comparison of wild-type and Cyp2e1(-/-) mice. Drug Metab. Dispos. (2007) 35:1223–1231.
Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J, Speed TP. Normalization for cDNA microarray data: A robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res. (2002) 30:e15.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||