Genomics and Proteomics Analysis of Acetaminophen Toxicity in Mouse Liver

Stefan U. Ruepp, Robert P. Tonge, Joanne Shaw, Nicola Wallis and François Pognan,1

AstraZeneca, Alderley Park, Macclesfield SK10 4TG, Cheshire, United Kingdom

Received June 7, 2001; accepted September 18, 2001


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Overdose of acetaminophen (APAP) causes severe centrilobular hepatic necrosis in humans and experimental animals. Here, to explore its mechanism, we administered APAP at subtoxic (150 mg/kg ip) and toxic (500 mg/kg ip) doses to overnight fasted mice. Animals were sacrificed at different time points from 15 min to 4 h postinjection. We assessed liver toxicity by plasma ALT activity and by electron microscopy. Using nylon filter arrays and RTQPCR, we performed genomics analysis in liver. We ran proteomics on liver mitochondrial subfractions using the newly developed quantitative fluorescent 2D-DIGE© method (Amersham Pharmacia Biotech UK Limited). As soon as 15 min postinjection, centrilobular hepatocyte mitochondria were already slightly enlarged and GSH total content dropped by a third at top dose. GM-CSF mRNA, which is a granulocyte specific gene likely coming from resident Kupffer cells, was also induced to its maximum of 3-fold at both doses. Chaperone proteins Hsp10 and Hsp60 were readily decreased by half in mitochondria at both doses, most likely by leaking into cytoplasm. Although APAP is known as an apoptotic trigger, no apoptosis was observed at any time point. Most of the protein changes in mitochondria were present at 15 min postinjection, thus preceding most of the gene regulations. The decrease of ATP synthase subunits and ß-oxidation pathway proteins indicated a loss of energy production. As the morphology of mitochondria was also affected very early at top dose, we concluded that APAP toxicity was a direct action of its known reactive metabolite NAPQI, rather than a consequence of gene regulation. However, the latter will either worsen the toxicity or lead toward cell recovery depending on the cellular damage level.

Key Words: acetaminophen; APAP; genomics; in vivo; Kupffer cells; mitochondria; mouse; oxidative stress; paracetamol; proteomics.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Acetaminophen or paracetamol, as it is known in Europe (N-acetyl-para-aminophenol; APAP) is a widely used analgesic and antipyretic drug throughout the world. At pharmacological doses, APAP is mainly metabolized by sulfation and glucuronidation (Vermeulen et al., 1992Go). A small proportion is metabolized through cytochromes CYP2E1 and to a lesser extent CYP1A2 and CYP3A4, which produce a reactive metabolite, N-acetyl-p-benzoquinone imine, NAPQI (Corcoran et al., 1980Go; Dahlin et al., 1984Go). Following therapeutic doses, NAPQI is efficiently detoxified by conjugation with glutathione, GSH (Miners et al., 1984Go). However, overdose causes severe centrilobular hepatic necrosis (Davidson and Eastham, 1966Go) and tubular necrosis in kidney (Kleinman et al., 1980Go) in humans and experimental animals (Boyd and Bereczky, 1966Go; Mudge et al., 1978Go). In this condition, a large amount of APAP is metabolized through the P450s, leading to GSH depletion by NAPQI conjugation followed by covalent binding of NAPQI to proteins (Mitchell et al., 1973Go). However, transgenic mice overexpressing glutathione synthetase, leading to a hepatic GSH build-up, displayed a more severe hepatotoxicity and nephrotoxicity caused by APAP (Rzucidlo et al., 2000Go). Also, APAP is toxic during mouse embryo development without depleting GSH (Laub et al., 2000Go). Thus, even though GSH has an obvious role in protecting cells against acetaminophen toxicity, its action is ambivalent and not the only mechanism by which cells are managing this toxicity. Subsequent to GSH depletion, covalent binding to proteins and inhibition of their functions has been postulated to be the main mechanism of toxicity (Hinson and Roberts 1992Go; Jollow et al., 1973Go). However, the pharmacologically inactive and nontoxic APAP counterpart, N-acetyl-meta-aminophenol (AMAP) also produces reactive metabolites that bind covalently to proteins to a higher or similar extent than APAP, but with different subcellular distribution particularly with regard to mitochondrial fraction (Rashed et al., 1990Go). Therefore, it has been postulated that APAP should inactivate critical proteins, left unchanged by AMAP. A number APAP protein targets have been identified, including GST{pi} and glutathione peroxidase, which increase APAP toxicity by preventing a rebuilding of GSH pool and thus allow further NAPQI binding to more proteins (Qiu et al., 1998Go). Again, other transgenic mice overexpressing intracellular glutathione peroxidase (EC 1.11.1.9.) also showed an increased sensitivity to APAP toxicity. This has been explained by a slow recovery of GSH in these animals (Mirochnitchenko et al., 1999Go). Other experiments are in apparent contradiction with the "protein-binding" hypothesis. Treatment with antioxidants like prometazine (McLean and Nutiall, 1978Go) or {alpha}-tocopherol and diphenylphenylenediamine (Albano et al., 1983Go) have shown to protect liver from APAP toxicity without reducing the amount of protein binding. The protective action of antioxidants points out the possible central role of reactive oxygen species (ROS) in APAP toxicity. Indeed, depletion of GSH is depriving cells of their first defense line against oxidative stress arising from their own metabolism as well as from APAP biotransformation. Mitochondria have been shown many times to be involved in ROS production as well as being victims of excessive oxidative stress. Treatment of primary rat hepatocytes in vitro with pro-oxidants like hydrazine or hydrogen peroxide induces lipid peroxidation and swollen megamitochondria (Karbowski et al., 1997Go). It appears that megamitochondria are always accompanied by lipid peroxidation, whereas lipid peroxidation can appear alone (Karbowski et al., 1997Go). Mice treated with APAP not only produce lipid peroxides in liver (Albano et al., 1983Go), but also megamitochondria as early as 1 h postdosage at 600 mg/kg po (Placke et al., 1987Go) and 1.5 h postdosage at 500 mg/kg po (Walker et al., 1980Go). Therefore, mitochondria may play a central role in APAP toxicity, as they have been shown to be involved in cell death, either by apoptosis or necrosis (Bernardi et al., 1999Go). In addition, mitochondrial GSH (mGSH) is the only defense mechanism against ROS in mitochondria (Fernandez-Checa et al. 1998Go), which is largely depleted by NAPQI production (Vendemiale et al., 1996Go). The level of mGSH controls the amount of ROS in cells, which in turn is an activator for transcription factors like AP-1 (Pinkus et al., 1996Go) and NF-{kappa}B (Fernandez-Checa et al., 1998Go; Pinkus et al., 1996Go). As a matter of fact, NF-{kappa}B and NF-IL6 binding capacities are almost completely abolished at 4 h postdosage in mice treated with 500 mg/kg ip (Blazka et al., 1995Go), whereas AP-1 binding activity was increased (Blazka et al., 1996Go). These transcription factors have pleiotropic actions on a wide range of different genes, which in turn may have a many different functions.

Among the 23 identified protein targets for NAPQI covalent binding, 6 are mitochondrial and these include the ATP synthetase {alpha}-subunit, a protein vital for energy production, and thus for cell survival (Qiu et al., 1998Go). Also, changes in calcium homeostasis may be pivotal in APAP toxicity, with a dramatic increase 2 h after APAP overdosing mice (Corcoran et al., 1987Go). An increase in intracellular and intramitochondrial calcium concentration initiates numerous cascades of events, leading to either necrosis or apoptosis (Williams, 1998Go). Calcium homeostasis is an energy (ATP) dependent mechanism and lack of energy, leads to Ca2+ influx from extracellular environment into cells (Williams, 1998Go). It has been shown that NAPQI induces a very rapid loss of ATP in treated hepatocytes (Andersson et al., 1990Go), thus linking mitochondria, ATP production, covalent modification of ATP synthetase {alpha}-subunit, Ca2+ homeostasis, and APAP toxicity together.

There are many controversial findings regarding acetaminophen toxicity, but most of them are alleviated by studies using CYP2E1 and CYP1A2 double knock-out mice that are extremely resistant to toxic doses of APAP. These mice do not show a significant amount of NAPQI or a drop of GSH, confirming the pivotal role of these P450s and NAPQI in liver toxicity and GSH in its protective role (Zaher et al., 1998Go). This is consistent with the observed colocalization by immunohistology of acetaminophen metabolite adducts in target organs with necrotic tissues (Hart et al., 1995Go; Roberts et al., 1991Go). These findings are anchoring the production of NAPQI as the very initial step of APAP toxicity. However, the understanding of the molecular basis remains largely unknown, and the proteins involved in covalent binding (Qiu et al., 1998Go) do not allow a full understanding of acetaminophen toxicity. Previously, only proteins covalently modified by NAPQI have largely been studied but it is possible that other reactive APAP and oxidative stress-derived metabolites interact with critical proteins leading to the observed toxicity. Thus, we have investigated by genomics (mRNA profiling by nylon microarray membranes) and proteomics (on mitochondrial preparations using 2D-DIGE technology) the effect of nontoxic and toxic doses of acetaminophen in mouse liver. Acetaminophen clinical and biochemical side effects are known well enough to be used as a reference compound to assess strengths and weaknesses of genomics and proteomics technologies as toxicological tools. Linked to protein regulations and mitochondrial toxicity, the complex gene regulations involved in APAP toxicity, as suggested by transcription factor modulations, could be integrated into a complex full picture of direct and delayed toxicity, as well as attempts of recovery.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Animal treatment.
All animal procedures were performed under United Kingdom Home Office License in accordance with the Animals Act (1986). Five adult Alderley Park (CD-1) male mice per group were fasted overnight before treatment. All animals were dosed by ip and sacrificed at different times by CO2 inhalation. Group 1 received sterile 0.9% (w/v) sodium chloride (control vehicle) and was sacrificed 60 min after treatment. Groups 2 and 3 were dosed with 150 mg/kg of APAP (nontoxic dose; Sigma A5000) and sacrificed at 15 and 120 min, respectively. Groups 4, 5, 6, 7, and 8 were dosed with 500 mg/kg of APAP (toxic dose) and sacrificed at 15, 30, 60, 120, and 240 min, respectively. Blood samples were withdrawn by cardiac puncture for estimation of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) activities, and for determination of APAP concentration. The livers were immediately removed and divided into several parts for RNA isolation, protein extraction, and histopathological examination.

Liver enzymes and APAP concentration in plasma.
ALT and AST were quantified using colorimetric kits (Sigma 505 and 505-P respectively) and APAP plasma concentrations were determined using a colorimetric kit based on nitrous acid reaction with APAP (Sigma 430-A).

Hepatic nonprotein thiol (HNPT) concentrations.
HPNT levels were taken as an estimate of hepatic glutathione concentrations and were carried out on liver homogenates using the Ellman's reagent method as described previously (Tonge et al., 1998Go).

Histology.
A small sample of the 4 major lobes (left and right lateral, left and right medial lobes) was taken and preserved in neutral buffered formalin. One histological section was prepared from each sample, stained with H and E, and examined by light microscopy. A small sample of the left lateral lobe was taken and immersed in 2.5% glutaraldehyde fixative and examined by electron microscopy.

RNA preparation.
Mouse liver samples were homogenized in 5 ml of lysis buffer (4M guanidine thiocyanate, 25 mM sodium citrate pH 7.0, 0.5 % sarcosyl, 0.72% ß-mercaptoethanol [Sigma, St. Louis, MO]). Total RNA was purified according to Chomczynski and Mackey (1995). Poly-A+ RNA was isolated with the PolyATtract mRNA Isolation System (Promega, Madison, WI). Good quality of mRNA was confirmed by Northern blotting using G3PDH as a reference gene.

Microarrays.
Mouse ToxBlots were produced by AstraZeneca. Mouse ToxBlots were spotted with PCR products of 450 mouse genes (Pennie et al., 2000Go). Each PCR product was spotted in quadruplicate. GDA filters were purchased from Genome Systems, Inc. (St. Louis, MO). GDA filters were spotted with bacterial clones that were lysed on the filter and subsequently UV crosslinked; 18,378 clones were spotted in duplicate; 1037 of these clones represent known genes, the remaining ones are ESTs.

Labeling and hybridization.
Identical amounts of Poly(A)+ RNAs from the 5 animals within each group were pooled and 1 µg of pooled Poly(A)+ RNA was reverse transcribed at 42°C for 2 h in the presence of 30 µCi {gamma}-33P dCTP (Amersham Pharmacia Biotech, Amersham, UK), 10 pmol dATP, 10 pmol dGTP, 10 pmol dTTP, 100 fmol dCTP, 100 pmol oligo dT18VN, and 200 units of SuperScript II RT (Life Technologies, Paisley, UK). Unincorporated nucleotides were removed using a ProbeQuant G-50 Micro column (Amersham Pharmacia Biotech, Amersham, UK). Microarrays were prehybridized at 65°C for at least 2 h in 8 ml (ToxBlots) or 12 ml (GDA filters) of hybridization solution (Soares et al., 1994Go). After prehybridization, solutions were replaced with the same volume of hybridization solution containing 107 (ToxBlots) or 1.5 x 107 cpm (GDA filters) radiolabeled probes and incubated at 65°C for 16 h in a roller oven. Microarrays were rinsed and washed twice for 30 min in 2x SSC/1% SDS at 68°C followed by 2 washes for 30 min in 0.6x SSC/1% SDS at 68°C.

Image analysis.
Filters were exposed for 2 days to a low energy phosphor screen (Molecular Dynamics, Sunnyvale, CA) and subsequently scanned (with a Fuji Phosphorimager). Images were analyzed using Array Vision software (Array Vision version 4, Molecular Research Inc., St. Catharines, Ont., Canada). Spot intensities were normalized to the median intensity of the control blot. Only genes exhibiting a good coefficient of variation in their duplicates (GDA) or quadruplicates (ToxBlot), more than a 2-fold variation and a consistent response over time have been selected.

RTQ-PCR.
One µg of pooled Poly(A)+ RNAs was reverse transcribed at 42°C for 15 min using a reverse transcription system (Promega, Madison, WI). Absence of DNA contamination in the mRNA preparation was confirmed by performing the same reverse transcription reactions, but without reverse transcriptase followed by RTQ-PCR analysis. Resulting first-strand cDNA was diluted 100-fold and 1 µl was used as template for RTQ-PCR analysis. RTQ-PCR analysis was performed using a LightCycler (Roche Molecular Biochemicals, Mannheim, Germany). Clones identified on the mouse ToxBlots were sequenced and a selection of sequence-verified genes was analyzed by RTQ-PCR. Clones identified on GDA filters were analyzed by RTQ-PCR only. Most IMAGE clones used by Genome systems represented only short, partial sequences of genes rendering it difficult to select appropriate primer pairs. Therefore most primer pairs were selected using the corresponding full-length gene sequences found in GenBank. Sequence-specificity of selected primers was confirmed by sequence similarity comparison with public and proprietary databases using BLAST 2 (NCBI). Amplified PCR products were analyzed by melting peak and melting curve analysis using the supplied Light Cycler Software Version 3 (Roche Molecular Biochemicals, Mannheim, Germany). A relative standard curve of control group cDNA was used in each PCR reaction at following dilutions: 1:2.5, 1:5, 1:10, 1:40, 1:160, 1:640, and 1:2560. Based on the expected fold up- or downregulation (indicated by microarray analysis) the 5 most appropriate dilutions were used for standard curves of each gene. In every set of PCR reactions was a negative control reaction included substituting cDNA with H2O. Based on microarray results GAPDH was used for normalization of input cDNA amounts. Table 1Go lists the primer pairs (all HPSF-grade from MWG-Biotech, Milton Keynes, UK) that were used for PCR analysis.


View this table:
[in this window]
[in a new window]
 
TABLE 1 Primer Pairs Used for RTQ-PCR Analysis
 
Mitochondria preparation.
Liver samples from each individual animal were initially processed separately. Liver samples (approximately 0.5 g) were homogenized into 10 ml ice-cold homogenization buffer (10 mM Tris–HCl [pH 7.4] containing 0.25 M sucrose, 1 mM EDTA, 0.5 mg/ml bovine serum albumin, 2.5 mg/l each of aprotinin, antipain, and leupeptin, and 1mM PMSF) with a Potter-type homogenizer (10 strokes). Whole homogenate were centrifuged at 250 x g for 10 min (4°C) to remove unbroken cells and connective tissue. Two ml of each supernatant were taken, snap frozen in liquid N2 and stored at –70°C. Some of this whole cell homogenate was used for HNPT estimations. The remaining supernatants from within a dose group were pooled and mitochondria were prepared essentially as per Rice and Lindsay (1997). The final mitochondrial pellet was resuspended in 1 ml isolation buffer (10 mM Tris–HCl [pH 7.4], 0.3 M sucrose, 0.5 mM EDTA) and assayed for protein concentration using the BCA method (Pierce, manufacturer's protocols). Samples were stored at –70°C until use.

Protein labeling with cyanine dyes.
CyTM2, Cy3, Cy5 (1mM solutions in DMF), and Pharmalytes (3–10) were from APBiotech (Amersham Pharmacia Biotech, Amersham, UK). This was essentially as per Tonge et al. (2001). Mitochondria preparations were solubilized in lysis buffer (4 M urea, 2 M thiourea, 2% CHAPS, 2% SB3-10, 0.5% Triton X-100; final protein concentration was 5 mg/ml) by vortexing at room temperature for 1–2 h. The pH of sample was adjusted to pH 8.5 by adding Tris–HCl to final concentration of 50 mM. One mM stock cyanine dye (in DMF) was diluted 1:5 with fresh dry DMF. This 200 µM dye solution was added to final ratio of 50 µg protein:200 pmol dye. Sample was vortexed and left for 30 min on ice in dark. The reaction was quenched by adding 1 µl 10 mM lysine (10 nmol) per 200 pmol dye used. Sample was vortexed and left for 10 min on ice in dark. An equal volume of 2D-sample buffer was added so that the final concentration was 4 M urea, 2 M thiourea, 2% CHAPS, 2% SB3-10, 0.5% Triton X-100, 0.8% (v/v) Pharmalytes, and 4 mg/ml DTT. Sample was vortexed for 15 min at room temperature. Sample was frozen at –70°C for later use or used immediately to rehydrate IPG strips.

Fluorescence 2-dimensional differential gel electrophoresis (2D-DIGE).
This was carried out essentially as Tonge et al. (2001). Mitochondrial samples were labeled with either Cy2, Cy3, or Cy5 and analyzed by 2D-DIGE. Three samples were run per 2D gel (10 gels in total): Gels 1 and 2, Control versus 150mg/kg 15 min versus 150mg/kg 30 min; Gels 3 and 4, Control versus 500mg/kg 15 min versus 500mg/kg 30 min; Gels 5 and 6, Control versus 500mg/kg 30 min versus 500mg/kg 60 min; Gels 7 and 8, Control versus 500mg/kg 120 min versus 500mg/kg 240 min; Gels 9 and 10, Control versus 150mg/kg 15 min versus 150mg/kg 15 min. Each gel contained the same control sample to act as an internal standard and in each case, the first sample was labeled with Cy2, the second with Cy3, and the third with Cy5. Samples were labeled and 75 µg of each was loaded onto the gel (total protein load per gel = 225 µg). Isoelectric focusing was carried out using a Multiphor II (APBiotech) for 120 kVh. IPG strips were pH 3–10, 18 cm (APBiotech). The IPG strips were then loaded and run on a 12.5% acrylamide Laemmli SDS–PAGE gel (modified ESA Investigator; 12.8% T, 0.86% C poured between Pyrex low fluorescence glass plates, essentially as Ausubel et al., 1993Go) until the bromophenol blue dye front had just run off the base of the gel (20°C, 300V, 300W, 6mA/gel 2 h, 10mA/gel 24 h). After SDS–PAGE, Cyanine dye-labeled protein gels were scanned directly using a 2920–2D Master Imager (APBiotech) and saved as 16-bit TIFF files. Preparative 2D gels of unlabeled protein were also run (1 gel per treatment group, 400 µg protein) and were visualized using colloidal Coomassie (Neuhoff et al., 1988Go) and digitized at 150 dpi resolution, using a flat bed scanner. These images were converted to 8-bit TIFF files using BioImage 2-D Analyzer (V6.1) Software (Genomics Solutions). Spots identified as different from the fluorescent analysis were excised from these preparative gels for MS analysis.

Image and data analysis.
Of the 10 fluorescent gels above, the gel from each duplicate with the best technical quality by eye was chosen and used for image analysis. Gels of best technical quality were those with least streaking and other artifactual features. Images were analyzed using ImageMaster (V3 Beta 1 and V3.0, APBiotech) as described earlier (Tonge et al., 2001Go) and normalized local background subtracted spot volumes determined. Previous investigations had defined threshold ratios outside which the ratio of 2 test spots must fall to be considered different with this experimental system (Tonge et al., 2001Go). These threshold ratios were based on the normal variability found when comparing identical samples and are dependent on spot volume and the particular dye combination in question. Thus, within each gel, spot ratios were generated and compared with those ratios required to assign a quantitative change.

Tryptic digest and MALDI-MS.
This was essentially as Tonge et al. (2001). Spots of interest were excised and placed into 96-well plates. The gel pieces were washed, digested with trypsin, and peptides were extracted. Concentrated peptide extracts were spotted onto a MALDI-TOF target plate along with {alpha}-cyano-4-hydroxycinnamic acid matrix using a Symbiot robot (Perseptive Biosystems). {alpha}-cyano-4-hydroxycinnamic acid (Aldrich, Dorset, UK) was recrystallized and matrix solution was made up as a 10 mg/ml solution in 0.1% TFA/60% acetonitrile. The plate was run automatically on the Voyager DE-STR (Perseptive Biosystems) using close offset external calibration. If the digest was weak and no identity was found, the samples were cleaned using C18 ZipTips (Milipore) for further MALDI-TOF analysis. Data files were searched in a user-generated Mammalia database subset of the SwissProt and TrEMBL database using Zips (an in-house developed search engine). Protein identifications were based on criteria previously described (Tonge et al., 2001Go). Protein spots that failed to give convincing database hits, or spots that gave spectra with unidentified masses were analyzed further by MS/MS using a Q-Tof mass spectrometer (Micromass, UK) equipped with a nanospray microcapillary inlet. A small volume (1–2 µl) of the concentrated and desalted sample from C18 ZipTip was loaded into the nanospray microcapillary needle for analysis. From an MS scan over the mass range 400–1600 Da, doubly charged peptide ions (excluding known keratin masses and trypsin autodigest products) were identified and selected for subsequent fragmentation by MS/MS analysis. The resultant MSMS spectra were searched against a mammalian subset of the Swissprot and TrEMBL databases using the Mascot program (http://www.matrixscience.com) and were also partially manually interpreted to generate sequence tags. These tags were at least 6 consecutive amino acids in length, usually greater, and were also searched against a mammalian subset of the Swissprot and TrEMBL databases. Longer tags with no hits were searched for homology allowing single mismatches. When a peptide selected for MS/MS fragmentation matched for sequence and theoretical tryptic mass, then the other peptides predicted from the theoretical tryptic digest and present were examined for confirmation.

Western blotting.
This was essentially as per Tonge et al. (1998). Mini SDS–PAGE gels of mitochondrial samples (10 µg for HSP60 and 50 µg per lane for the remaining proteins of interest) were run (200V, 4°C) using Mini Protean II electrophoresis tanks (BioRad). Stacking gels were 4% acrylamide and resolving gels were either 8% (GRP75), 10% (HSP60 and APAP-adducts), or 15% (HSP10) acrylamide. Gels were electroblotted to PVDF (Immobilon-P, 0.45µm, Millipore. 100 V, 2 h, 4°C). Primary antibodies were 1:250 affinity purified anti(APAP) (Tonge et al., 1998Go), 1:10000 anti(HSP60), 1:5000 anti(HSP10), or 1:5000 anti(GRP75) and all were incubated with the blots for 3 h at room temperature. Anti-HSP60 antibody was polyclonal rabbit anti-cyanobacteria HSP60 (product number SPA-804), anti-HSP10 antibody was polyclonal rabbit anti-Cpn10 peptide (product number SPA-110), and anti-GRP75 antibody was monoclonal mouse anti-human GRP75 (product number SPA-825); all were from StressGen (BioQuote Ltd., York, UK). Secondary antibodies were 1:1000 goat anti-rabbit IgG-alkaline phosphatase conjugate and were incubated for 2 h at room temperature. Bound antibodies were visualized colourimetrically with NBT/BCIP. Secondary antibody was from Sigma, UK and BCIP/NBT (5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium) color development substrate was from Promega, UK. Blots were digitized and quantified using a flat bed scanner (300 dpi), a Macintosh computer, and the public domain NIH Image program (available at http://rsb.info.nih.gov/nih-image/).


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Clinical Chemistry
Mice were chosen because they have been shown to be more sensitive to acetaminophen than rats (Davis et al., 1974Go). APAP plasma maxima were reached as quickly as 15 min after injection at both dosages (Table 2Go). The ratio was about 3-fold between doses and also between plasma concentrations at the first time point, but APAP was almost cleared from blood at 120 min postinjection at the low dose, while it was still significantly high at the top dose at 240 min. Hepatic GSH was reduced at 120 min postinjection at low dose, while it was already decreased at 15 min at the high dose compared to control (Table 3Go). Plasma ALT and AST (Control ALT: 50.3 U/ml ± 12.4, AST: 114 U/ml ± 29.4) showed a statistically significant increase only at 120 (ALT: 129.3 U/ml ± 35.4, AST: 178.6 U/ml ± 25.5, p < 0.01) and 240 min (ALT: 229.3 U/ml ± 37.8, AST: 244.6 U/ml ± 7.5, p < 0.001) at the top dose only.


View this table:
[in this window]
[in a new window]
 
TABLE 2 Plasma APAP Concentrations in µM
 

View this table:
[in this window]
[in a new window]
 
TABLE 3 Hepatic GSH Concentrations in µM per Gram of Tissue
 
Histology
Histopathological examinations showed hepatocyte vacuolation at the light microscope level, 60 min after administration of 500 mg/kg APAP. The vacuolated hepatocytes were swollen and this led to sinusoidal narrowing and congestion. The most severely congested sinusoids were those in the mid-zone as a result of red cells not being able to pass through the narrowed centrilobular sinusoids. As the toxic injury became more severe with time, ballooning degeneration and hepatocyte necrosis was seen in the centrilobular zone with a mild degree of associated hemorrhage. The changes in the animals dosed at 150 mg/kg were less severe although there were minor changes after 120 min. At electron microscopy level (Fig. 1Go), minor changes could be detected as early as 15 min after dosing with 500 mg/kg (Fig. 1CGo). The changes started in the centrilobular zone and increased in severity and in distribution with time. It was only after 60 min that the degree of mitochondrial dilatation was sufficient to be visible as vacuolation by light microscopy. These mitochondria were likely not only swollen, but also fused together as their number was visibly lower and their shape suggested such a process (Fig. 1EGo). The changes in the mitochondria at 150 mg/kg were much less severe (Fig. 1BGo). There were no mitochondria sufficiently enlarged to be seen as vacuoles in the samples in light microscopy.



View larger version (164K):
[in this window]
[in a new window]
 
FIG. 1. Electron microscopy of the liver. (A) Control liver at 60 min posttreatment. (B) Treated liver with 150 mg/kg at 120 min posttreatment. Mitochondria were swollen to a similar extent as 30-min high dose. (C) Treated liver with 500 mg/kg at 15 min posttreatment. Mitochondria were already slightly enlarged compared to control. (D) Treated liver with 500 mg/kg at 30 min posttreatment. (E) Treated liver with 500 mg/kg at 120 min posttreatment. Mitochondria swollen and fused together. N, nucleus; M, mitochondria.

 
Genomics
Gene induction or repression as measured by their mRNA level, either by real time quantitative PCR (RTQ-PCR) using the Light Cycler or by micro-array were considered as real above 2-fold. Gene profiles obtained with AstraZeneca mouse ToxBlot I array and RTQ-PCR correlated generally very well (Table 4Go). The RTQ-PCR seemed to fail to detect a downregulation at low dose, at 120 min postinjection, as all other time points at both doses were downregulated. On contrary, many results obtained with GDA filters were not in agreement with the RTQ-PCR (Table 5Go, italic numbers), being either neutral or upregulated in PCR and downregulated on arrays (e.g., Interleukin 6 receptor {alpha} and Calpain-like cysteine protease) or vice versa (e.g., BAX and STAT3). The poor correlation of GDA results with RTQ-PCR is most likely caused by the fact that Genome Systems, Inc. failed to sequence the clones obtained from the IMAGE collection prior to arraying. The IMAGE database has an estimated 20% error rate (http://www.incyte.com/reagents/catalog.jsp?page=expression/comparison/clone). Therefore, all other upregulated and downregulated genes and ESTs as measured by GDA arrays are only trustworthy if validated. This validation is not straightforward and needs more work in quantitative PCR for conclusive interpretations. Information can be gained by knowing the apogee and nadir of gene regulation, however the profile of the time course is also very important. Figure 2Go showed that both c-fos or EGR-1, known as immediate early responsive genes in case of stress (Liu et al., 1998Go), seemed to plateau at 120 min postinjection at 500 mg/kg, whereas their mRNA production at 150 mg/kg seemed still on a growing slope at the same time. However due to the design of the experiment, it is not known if the low dose would reach the same plateau later on or have a different profile. The curve shapes for c-fos were identical with RTQ-PCR and micro-array (Fig. 2AGo), which was the case for all other genes checked by both methods with the mouse ToxBlot I (data not shown), indicative of reliable kinetics. Mad is a protein, which is known to be a downregulator of c-myc activity by sequestering max protein, the second half of the active heterodimer myc/max (Dang, 1999Go). Therefore, we have monitored mRNA level of these 3 genes by RTQ-PCR (Fig. 3Go). C-myc was more quickly induced at low dose than high dose, whereas mad showed a fairly identical slope up to 120 min. Max was slightly induced to a maximum 2.9-fold at high dose at 240 min, but the curves remained rather flat. It is therefore possible that the hetero-duplex max/mad has been preferably formed to the complex myc/max. EGR-1 has been shown to be involved in pro-apoptotic actions, either by activating directly apoptotic pathways through p53 or tumor necrosis factor-{alpha}, or indirectly by inhibiting growth arrest pathways through c-myc activation (Liu et al., 1998Go). As a c-myc mRNA increase (3-fold increase at 120 min top dose) was observed together with EGR-1 mRNA (8-fold at 30 min top dose), indicating a possible apoptotic pathway trigger, p53 and TNF-{alpha} mRNA levels have been checked by RTQ-PCR. The anti-oncogene and pro-apoptotic p53 mRNA level remained flat at all time points at both dose levels (not shown), whereas TNF-{alpha} mRNA showed a 5.8-fold increase at 120 min for 150 mg/kg and a plateau at 3.5-fold from 60 min onward at high dose (Fig. 4AGo). TNF-{alpha} has been shown as being synthesized by Kupffer cells, which are resident macrophages in liver (Hoebe et al., 2001Go). GM-CSF is known to be secreted by circulating macrophages and monocytes and to trigger EGR-1 synthesis that in turn activates TNF-{alpha} during apoptosis (Liu et al., 1998Go). Therefore, we checked GM-CSF mRNA level that was actually increased about 2 to 3-fold as soon as 15 min at both doses (Fig. 4BGo).


View this table:
[in this window]
[in a new window]
 
TABLE 4 AstraZeneca Mouse Toxblot and RTQ-PCR Analysis of Hepatic mRNA, as Fold Induction or Repression in Treated Liver, Compared to Controls
 

View this table:
[in this window]
[in a new window]
 
TABLE 5 GDA Mouse Array and RTQ-PCR Analysis of Hepatic mRNA, as Fold Induction or Repression in Treated Liver, Compared to Controls
 


View larger version (21K):
[in this window]
[in a new window]
 
FIG. 2. C-fos and EGR-1 induction monitored by RTQ-PCR and microarray. (A) C-fos induction assessed by micro-array and quantitative RT-PCR. (B) EGR-1 monitored by quantitative RT-PCR.

 


View larger version (14K):
[in this window]
[in a new window]
 
FIG. 3. Time course of c-myc, max, and med mRNA expression. (A) c-myc; (B) max; (C) mad. (Monitored by quantitative RT-PCR.)

 


View larger version (19K):
[in this window]
[in a new window]
 
FIG. 4. Time course of TNF-{alpha} and GM-CSF expression. (A) TNF-{alpha}; (B) GM-CSF. (Monitored by quantitative RT-PCR.)

 
In summary, for genes that have a role in apoptosis and were induced in this study, their order of appearance at top dose, independently of their maximum level of induction, was: EGR-1 = GM-CSF (15 min) > c-fos (30 min) > TNF-{alpha} = mad (60 min) > c-myc (120 min).

APAP-Protein Adducts
Mitochondrial samples from livers of animals treated with APAP were analyzed by immunoblotting for the presence of APAP modified proteins (Fig. 5Go). Following 500 mg/kg APAP, a range of different protein-adducts were produced in a time-dependent fashion beginning at 60 min postdose. No adducts were seen at 15 and 30 min. At 150 mg/kg, adducts were observed at the 120 min time point only and the spectrum of adducts was different to that seen at this time point with 500 mg/kg APAP.



View larger version (91K):
[in this window]
[in a new window]
 
FIG. 5. Western blot demonstrating presence of APAP-protein adducts produced in liver mitochondria following exposure of mice to 150 or 500 mg/kg APAP. Minigel run as described in Materials and Methods. Protein adducts appeared at 120 min posttreatment low dose, whereas they were seen from 60 min onward for the high dose. However, protein adducts may have appeared earlier in the low dose group. The protein-adduct pattern was different with more bands at high dose, showing a stronger effect.

 
Proteomics
Protein spots found to be differentially regulated in at least 1 cyanine dye combination from 1 2D-DIGE gel can be seen in Figure 6Go. Compared to control animals, 70 protein abundance changes were highlighted, both up (30), down (39), and variable (1) changes. Of these 70 spots, only 45 could be found on preparative gels stained with colloidal coomassie and of these 45, 20 gave positive identifications by mass spectrometry. These are listed in Table 6Go. The magnitude of ratio changes ranged from 0.3 downregulation (ATP synthase {alpha} chain and HSP10) to 4.2 upregulation (mitochondrial stress 70 protein P1). A number of proteins demonstrated changes at certain time zones in the experiment (e.g., aldehyde dehydrogenase was downregulated at later time points in both low and high APAP doses while mitochondrial stress 70 protein P1 was upregulated at these times. HSP10 and HSP60 appeared to be downregulated in all time points at both APAP doses.



View larger version (135K):
[in this window]
[in a new window]
 
FIG. 6. Quantitative fluorescent 2D-DIGE example. Spots being differentially regulated in at least 1 cyanide dye combination from 1 gel are numbered and referred as to Master Spot Numbers (MSN) in the text and Table 5Go.

 

View this table:
[in this window]
[in a new window]
 
TABLE 6 Protein Abundance Changes in Mouse Mitochondria following APAP Exposure (150 or 500 mg/kg) for 15–240 Min
 
Western Blotting
In an attempt to further confirm the nature of the protein abundance changes observed using fluorescence 2D-DIGE, we examined the change in abundance in HSP60, GRP75, and HSP10 with APAP-treatment by immunoblotting. These proteins were chosen as antibodies to them, were commercially available, and they appeared to show marked dose and time dependent effects by 2D-DIGE. Although by eye there were no obvious protein abundance changes, densitometric analysis of the immunoblots revealed abundance changes with APAP-treatment (Fig. 7Go). The difference trends seen with HSP60 and HSP10 were similar with 2D-DIGE and immunoblotting but GRP75 did not appear to change in abundance appreciably by immunoblotting while it was observed to be upregulated with APAP-treatment in a time and dose-dependent fashion by 2D-DIGE.



View larger version (22K):
[in this window]
[in a new window]
 
FIG. 7. Western blot demonstrating levels of HSP60, GRP75, and HSP10 in liver mitochondria following exposure of mice to 150 or 500 mg/kg APAP for 15–240 min. Densitometric analysis on the immunoblots (A) enabled determination of band peak areas that could then be expressed as a percentage

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
This study has demonstrated the dose and time-dependent production of gene and mitochondrial protein changes following APAP exposure of mice in vivo. Along with these molecular responses, the time course of toxicity was monitored by more conventional and accepted clinical chemistry and histological parameters, enabling us to place the molecular changes in the context of the wider biological picture.

The first necropsy time point was 15 min after ip injection of APAP, at which time the maximum plasma APAP concentration was also measured. Thus, the APAP concentration peak was likely to be earlier, although it is hard to determine the exact time. Hepatic GSH level was reduced by a third only 15 min postinjection at high dose, an effect on mitochondria was already detectable by electron microscopy and some gene induction at the mRNA level was readily detectable. Traditional liver toxicity marker enzymes showed a plasma increase only after 2 h postinjection high dose, thus being a comparatively late event reflecting an already established toxicity, whereas the former were more the onset of the hepatotoxicity.

The mitochondrial toxicity was progressive with time and dose, starting with swollen mitochondria at 15 min and finally leading to swollen and fused mitochondria from 60 min onward, to so-called megamitochondria (Karbowski et al., 1999Go). Therefore, it appeared that GSH depletion was concomitant to mitochondrial damage, which is known to lead to an increase of reactive oxygen species (ROS; Fernandez-Checa et al., 1998Go). This increase of ROS contributes to a more profound GSH depletion, while NAPQI can conjugate with more GSH and protein and induce more damage. It has also been shown that a decrease of cytoplasmic GSH induces a parallel and quick mitochondrial GSH (mGSH) depletion (Fernandez-Checa et al., 1998Go). The only supposed mitochondrial defense line against ROS is mGSH (Fernandez-Checa et al., 1998Go). A drop in mGSH leads to rapid protein damage, cascading to ATP fall, elevation of matrix Ca2+ concentration, inversion of the inter-membrane potential ({Delta}{Psi}m) and subsequent release of the pro-apoptotic factors, cyt-c, Hsp10, and Hsp60 (Bernardi et al., 1999Go). Such a GSH depletion leads to ROS mediated gene activation (like c-fos, c-myc family, EGR-1, TNF-{alpha}, etc.) seen during this study. Many of these genes are involved in cell regeneration or apoptotic pathways.

Induction of apoptosis by cell death effectors like FAS ligands decreases c-fos transcriptional activity in Jurkat T cells (Bertolotto et al., 2000Go). As this proto-oncogene is well known for its proliferative and transforming activity when expressed in cells, induction of its mRNA in APAP-treated liver may reveal an attempt of cell rescue at both low and high doses. As c-fos induction was less significant at 150 mg/kg, cells may grade its induction upon the gravity of injury. Therefore, c-fos, which has already been postulated as a simple response to injury in case of acute stress-induced in heart, stomach, and liver (Fernandez et al., 2000Go), may be seen as a biomarker of cell suffering.

The role of c-myc was ambivalent as it was not paralleled by a max increase, which would have implied a cell division activation. On the opposite, mad, which is c-myc negative "controller," was going up with c-myc. Therefore, myc gene transcription activation is likely to be affected in one way or another by other gene regulation. Unfortunately, c-jun was not present on both arrays and it has not been possible to monitor its behaviors. However, mice treated with APAP lead to c-fos and c-jun induction after 1 h and these gene inductions correlated with increased AP-1 binding activity (Kitteringham et al., 2000Go). C-myc and c-fos inductions are usually a sign of toxicity and therefore should not be underestimated.

GM-CSF is produced by a variety of cells including circulating macrophages and Kupffer cells resident in liver. GM-CSF mRNA induction plateaued as soon as 15 min postinjection. Therefore, this increase cannot be attributable to the accumulation of blood and circulating macrophages, which started at 60 min postinjection at the top dose. If Kupffer cells alone are responsible for this observation, this represents a dramatic boost of GM-CSF transcription, as these cells represent only 2% of total liver (Blouin, 1977Go). However important this activation was, it appeared to be a physiological answer, as the mRNA increases were identical at both doses. Therefore, this is unlikely to be the starting event of APAP toxicity observed at high dose only, but combined with other gene regulations, GM-CSF may orientate the cell reactions either to death or survival. GM-CSF is known to trigger EGR-1 (Kharbanda et al., 1991Go), which in turn activates TNF-{alpha} (Ahmed et al., 1997Go). It was remarkable to notice that GM-CSF induction occurred first, followed successively by EGR-1 (30 min) and TNF-{alpha} (60 min) at both doses. EGR-1 induction was about 25-fold at 500 mg/kg APAP and 15-fold at 150 mg/kg APAP at 120 min postinjection. EGR-1 is a transcription factor that plays a role in macrophage/monocyte differentiation, but it is also capable of triggering TNF-{alpha} production (Ahmed et al., 1997Go; Nguyen et al., 1993Go). This cytokine is produced by macrophages only, and though modest, the average 3-fold increase of TNF-{alpha} mRNA at top dose, was likely to originate from Kupffer cells only. Therefore, as for GM-CSF, this induction may represent a huge transcription boost. It appeared that the TNF-{alpha} increase was higher at low dose than at high dose. TNF-{alpha} has been shown to be involved in necrosis (Simpson et al., 2000Go), but also in triggering apoptosis (Nagata, 1997Go) through release of ceramides that in turn act on mitochondria (Garcia-Ruiz et al., 1997Go), the central effector of programmed cell death. It is also a feedback activator of EGR-1 (Mechtcheriakova et al., 2001Go), which initiates apoptosis through induction of p53 activation (Nair et al., 1997Go). However, p53 was not induced or repressed at any time point at both doses. Moreover, there was no sign of apoptosis on any light or electron microscope pictures. TNF-{alpha} alongside with IL-6 are also known to induce NF-{kappa}B and STAT3 transcription factors that cascade to liver regeneration, in which EGR-1 also plays a role (Kirillova et al., 1999Go). The discrepancy between the GDA array and the RTQ-PCR about STAT-3 induction did not allow us to state a clear hypothesis on it. However, the trend was indicating transcriptional activation. It seemed that early gene inductions were aimed at both survival and death. However, for any reason, the balance was poised toward necrosis at 500 mg/kg. The decisive weight in the balance could be a direct damage of mitochondria as they were clearly rapidly affected at top dose treatment. Damage to mitochondria can lead either to necrosis or to apoptosis (Bernardi et al., 1999Go). Also, TNF-{alpha} is targeting mitochondria as an apoptotic effector. Here, mitochondria were swollen and fused together at high dose as soon as 60 min postdosing. However, this is unlikely the result of TNF-{alpha} action as its mRNA level seemed to reach its maximum at 60 min postinjection. Thus, TNF-{alpha} may have an enhancer role at later time points, but was not the initiator of mitochondrial swelling. Furthermore, TNF-{alpha} was unlikely to have any direct role in the observed necrosis for the same timing reasons and also it has been shown as not being responsible for APAP-induced necrosis in mouse liver (Simpson et al., 2000Go).

GM-CSF, EGR-1, and TNF-{alpha} are all known to trigger programmed cell death but in this study, no apoptotic events have been observed. On the contrary, necrosis was clearly apparent. The megamitochondria are supposedly ATP depleted (Karbowski et al., 1999Go), which could mean a lack of energy for the hepatocyte to recover from the NAPQI injuries. Also, apoptosis is ATP-dependent, thus, if mitochondria were not able to provide enough energy to activate the caspase cascades, cells would likely die from necrosis instead. The inability of cells to quickly replace GSH depleted by NAPQI, due to a lack of ATP, would further enhance toxic protein binding and letting ROS to have a pleiotropic action on any molecule available, such as lipids, proteins, and nucleic acids. The situation as summarized in Figure 8Go, was that necrosis would overwhelm apoptosis, which could not go further because of a lack of energy. The vicious circle initiated by the drop of GSH is likely to be further enhanced by the inability of mitochondria to work normally. According to the damage level, the pleiotropic gene regulation observed in this study will either aggravate the toxicity or lead the liver to recovery. Therefore the gene transcriptions were likely not to have an effect on acute APAP toxicity, but certainly have a role in delayed toxicity as well as in case of repeated dosing. This could explain apparent discrepancies described in the introduction of this paper. It is likely that observation of liver treated with 150 mg/kg at later time points would have revealed apoptotic events, triggered by the GM-CSF/EGR-1/TNF-{alpha} cascade. This will be studied in subsequent experiments.



View larger version (33K):
[in this window]
[in a new window]
 
FIG. 8. Toxic cascade of events in APAP-overdosed mouse liver. The scheme summarizes published data and the present results. The sequence of events seemed to start with NAPQI direct toxicity, such as protein binding, cytoplasmic and mitochondrial GSH depletion. This leads to a broad range of gene up- and downregulation. A number of these genes are involved in apoptotic pathways, but necrosis shunts programmed cell death because of mitochondrial failure and lack of energy.

 
APAP is known to both inhibit and induce the synthesis of a number of proteins in the liver during the course of toxicity (Bruno et al., 1992Go; Myers et al., 1995Go; Salminen et al., 1998Go) and thus, in an attempt to better understand the nature of the toxicity, we have used newly developed quantitative fluorescent 2D-DIGE methods (Tonge et al., 2001Go) to examine the changes in the mitochondrial proteome following APAP exposure. Overall, differences were relatively modest in magnitude (maximum 3.3-fold downregulation [0.3 ratio] and 4.2-fold upregulation) and thus, sensitive quantitative techniques are essential for detection of changes.

Aldehyde dehydrogenase (ADH) is involved in the oxidation of aldehydes in many different metabolic pathways, from amino-acids metabolism to fatty acid metabolism and has been reported to be covalently modified by APAP (Landin et al., 1996Go). The extent of covalent modification correlated with loss in ADH activity and progressed from 1–4 h following treatment. The most significant loss of ADH protein was at late time points (> 2 h) and thus, may be a consequence of protein degradation caused by covalent modification. The loss of this activity may impair the fatty acid ß-oxidation pathway, which is one of the two major sources of energy of the cells, together with glycolysis. This was going along with a similar depletion pattern of thiolase, the third enzyme in the fatty acid ß-oxidation spiral and enoyl-CoA isomerase, an auxiliary enzyme required to link unsaturated fatty acids to the ß-oxidation pathway (Filppula et al., 1998Go). This would logically create a lack of acetyl-CoA that is the main final product of the fatty acid ß-oxidation pathway. This lack would also impair the normal functioning of the citric acid cycle, as it cannot work without acetyl-CoA. The observed increase of acetyl-CoA acetyl transferase at both doses and mainly at early time points, may reflect an attempt by mitochondria to compensate the failure of acetyl-CoA production coming from the ß-oxidation of fatty acids. A major part of this failure was possibly the downregulation of ATP synthase {alpha}-chain. Of the 4 isoforms detected on the 2D-gels, only 1 seemed not to be affected. At top dose treatment ATP synthase was more affected than at the low dose. ATP synthase {alpha}-chain subunit has been described as covalently bound by NAPQI (Qiu et al., 1998Go) and its disappearance was not unexpected. Noteworthy, the ß-chain of the ATP synthase complex was not affected. This picture was a strong indication that energy supply in hepatocytes suffering from acute APAP intoxication was a crucial point that could likely be the second event after GSH depletion that led to necrosis, instead of apoptosis or cell recovery.

HSP10 and HSP60 are chaperone proteins and are found in the mitochondrial intermembrane space in a complex with procaspase-3 (Samali et al., 1999Go). When apoptosis is induced, HSP10 and 60 dissociate from procaspase-3 and are released from the mitochondria. Once in the cytoplasm, they act to accelerate caspase activation in the process of apoptosis. Some evidence has recently been shown however, that APAP does not activate caspase-3 and does not cause apoptosis (Lawson et al., 1999Go). The observed decrease in our study was likely due to a release of HSP10 and HSP60 from mitochondria into the cytoplasm. Thus, although HSP10 and HSP60 were released from mitochondria and some pro-apoptotic gene transcriptions were induced, no apoptotic events were noted. This may be due to deactivation of caspases by APAP-metabolites (Lawson et al., 1999Go) or possibly by ATP depletion as ATP is required for procaspase-3 activation and apoptosis.

This study aimed to dissect the roles of gene and protein regulation in acute APAP toxicity in fasted mouse liver, compared to a subtoxic dosing, for assessment against well established pathological and biochemical techniques The overall picture of mRNA transcription was similar for 150 mg/kg and 500 mg/kg. The differences were only in the extent of induction or repression, with a more pronounced effect at the top dose. However, it is suspected that the same signal or group of signals may have a different effect in one cell, depending on the strength with which it is delivered (Pawson and Saxton, 1999Go). Although less pronounced, the same pattern was observed for mitochondrial proteins. One can argue that 150 mg/kg was not a nontoxic dose, as a statistically significant GSH decrease and some mitochondrial dilatations were observed at 120 min postinjection. However, the extent of histopathological changes was dramatically more pronounced at top dose and could not be compared to the lower dose. Therefore, the difference between recovery and extended liver damage resides in a fairly limited difference in gene and protein regulation span, unless some critical and rare genes and/or proteins were not observed during this study. The fact that only a part of the murine transcriptome was represented on the arrays in this study and that it is well documented that 2D-gels are not sufficiently sensitive to detect very low abundance proteins (Gygi et al., 2000Go) does not allow us to exclude the possibility that we missed some important changes. The fact that a great number of genes changed also at the low dose and that those genes belonged to a variety of functional categories (oncogenes, pro-apoptotic factors, transcription factors, stress genes, etc.) demonstrates the effect of a subtoxic dose on the transcriptome. These gene regulations are very likely to have a role in either recovery by sacrificing heavily damaged cells by apoptosis or stimulating cell proliferation, or in long-term toxicity in case of repeated dosing. However, it appeared that the high acute dose (500 mg/kg) resulted in severe toxicity that could not be explained by the observed transcriptional changes. This can be concluded from the timing of the events, where the first pathological effects in liver were concomitant to gene upregulation and protein up- and downregulation. Since toxicity was already evident after 15 min, cells likely lacked the time to alter transcription of specific genes, create mature proteins, and allow them to produce an effect. Observed transcriptional changes themselves are probably more relevant in terms of amplifying the primary toxic insult, but also reflect attempts of regeneration processes. Therefore, protein damage was more likely to play a critical role in acute APAP overdosing. However, APAP-protein adducts were only seen from 60 min onward. Thus, APAP acute toxicity is possibly affecting proteins before and independently of binding to them. One of these changes that may be the most biologically significant was the inability of the cells to produce a normal energy level, due to the depletion of fatty acid oxidation pathway and ATP synthase {alpha} subunits. Three recent reports have studied the molecular aspects of APAP toxicity. At the proteome level at 8 h posttreatment with 100 and 300 mg/kg APAP (Fountoulakis et al., 2000Go) and at 4 h posttreatment with 500 mg/kg APAP (Tonge et al., 2001Go), both in fasted mice. At the genome level, the studies were at 6 h posttreatment with 300 mg/kg APAP in fasted mice (Reilly et al., 2001Go). The protein regulations observed corroborated our own findings, as for example HSP 60 and GRP 75, but with a more limited breadth. This can be due to the lower doses used as well as the relatively late time point used after injection. Indeed, it appeared that protein changes occurred extremely quickly, as soon as 15 min postinjection. Therefore, the relevant window of observation was before the release of ALT and AST in the blood. As for the genomics study (Reilly et al., 2001Go), most of the regulations were in accordance to ours. However, again due to the relative late unique time point observed and unique dose used, genes participating in the early onset of APAP toxicity could be missed, as for example GM-CSF.

In the study reported here, the kinetic design and the 2 different doses used, allowed us to reject a direct implication of gene regulation in acute APAP toxicity, but also pointed out the crucial role of mitochondria and energy production. Unfortunately, the gene comparison of the 2 different doses did not allow us to find one or more genes that would have been specific to APAP toxicity and could have been subsequently used as a toxicological biomarker. In summary, transcript profiling and proteomic investigations have detected some APAP-treatment related cellular responses that have been previously reported, but also succeeded in identifying new and interesting changes. Dose and time courses were essential for interpretation of the often-surprising responses detected. Transcript profiling and proteomics did not usually detect expression level changes of one mRNA and the corresponding protein, but genes, proteins, and pathways identified by transcript profiling and by proteomics, told a similar story.


    ACKNOWLEDGMENTS
 
The authors would like to thank Matthew Davison and Terry Orton for helpful discussions and suggestions and Janice Young, Rachel Rowlinson, and Steve Rayner for mass spectrometric identifications.


    NOTES
 
S. Ruepp and R. Tonge have participated equally in this work.

Portions of this data were presented at the 9th annual North American ISSX meeting, October 1999, Nashville, TN.

1 To whom correspondence should be addressed at AstraZeneca, Molecular and Investigative Toxicology, 1800 Concord Pike, PO Box 15437, Wilmington, DE 19850. Fax: (302) 886-2341. E-mail: francois.pognan{at}astrazeneca.com. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Ahmed, M. M., Sells, S. F., Venkatasubbarao, K., Fruitwala, S. M., Muthukkumar, S., Harp, C., Mohiuddin, M., and Rangnekar, V.M. (1997). Ionizing radiation-inducible apoptosis in the absence of p53 linked to transcription factor EGR-1. J. Biol. Chem. 26, 33056–33061.

Albano, E., Poli, G., Chiarpotto, E., Biasi, F., and Dianzani, M. U. (1983). Paracetamol-stimulated lipid peroxidation in isolated rat and mouse hepatocytes. Chem. Biol. Interact. 47, 249–263.[ISI][Medline]

Andersson, B. S., Rundgren, M., Nelson, S. D., and Harder, S. (1990). N-acetyl-p-benzoquinone imine-induced changes in the energy metabolism in hepatocytes. Chem. Biol. Interact. 75, 201–211.[ISI][Medline]

Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K., Eds. (1993). Current Protocols in Molecular Biology, Vol. 2. John Wiley, Canada.

Bernardi, P., Scorrano, L., Colonna, R., Petronilli, V., and Di Lisa, F. (1999). Mitochondria and cell death. Mechanistic aspects and methodological issues. Eur. J. Biochem. 264, 687–701.[Abstract/Free Full Text]

Bertolotto, C., Ricci, J. E., Luciano, F., Mari, B., Chambard, J. C., and Auberger, P. (2000). Cleavage of the serum response factor during death receptor-induced apoptosis results in an inhibition of the c-FOS promoter transcriptional activity. J. Biol. Chem. 275, 12941–12947.[Abstract/Free Full Text]

Blazka, M. E., Bruccoleri, A., Simeonova, P. P., Germolec, D. R., Pennypacker K. R., and Luster, M. I. (1996). Acetaminophen-induced hepatotoxicity is associated with early changes in AP-1 DNA binding activity. Res. Commun. Mol. Pathol. Pharmacol. 92, 259–273.[ISI][Medline]

Blazka, M. E., Germolec, D. R., Simeonova, P., Bruccoleri, A., Pennypacker, K. R., and Luster, M. I. (1995). Acetaminophen-induced hepatotoxicity is associated with early changes in NF-{kappa}B and NF-IL6 DNA binding activity. J. Inflamm. 47, 138–150.[Medline]

Blouin, A. (1977). Morphometry of liver sinusoidal cells. In Kupffer Cells and other Liver Sinusoidal Cells (E. Wisse and D. L. Knook, Eds.), pp. 61–77. Elsevier/North-Holland Biomedical Press, Amsterdam.

Boyd, E. M., and Bereczky, G. M. (1966). Liver necrosis from paracetamol. Br. J. Pharmacol. 26, 606–614.[Medline]

Bruno M. K., Cohen, S. D., and Khairallah, E. A. (1992). Selective alterations in the patterns of newly synthesized proteins by acetaminophen and its dimethylated analogues in primary cultures of mouse hepatocytes. Toxicol. Appl. Pharmacol. 112, 282–290.[ISI][Medline]

Chomczynski, P., and Mackey, K. (1995). Short technical reports. Modification of the TRI reagent procedure for isolation of RNA from polysaccharide- and proteoglycan-rich sources. Biotechniques 19, 942–945.[ISI][Medline]

Corcoran, G. B., Mitchell, J. R., Vaishnav, Y. N., and Horning, E. C. (1980). Evidence that acetaminophen and N-hydroxyacetaminophen form a common arylating intermediate, N-acetyl-p-benzoquinoneimine. Mol. Pharmacol. 18, 536–542.[Abstract]

Corcoran, G. B., Wong, B. K., and Neese, B. L. (1987). Early sustained rise in total liver calcium during acetaminophen hepatotoxicity in mice. Res. Commun. Chem. Pathol. Pharmacol. 58, 291–305.[ISI][Medline]

Dahlin, D. C., Miwa, G. T., Lu, A. Y., and Nelson, S. D. (1984). N-acetyl-p-benzoquinone imine: A cytochrome P-450-mediated oxidation product of acetaminophen. Proc. Natl. Acad. Sci. U.S.A. 81, 1327–1331.[Abstract]

Dang, C. V. (1999). C-Myc target genes involved in cell growth, apoptosis and metabolism. Mol. Cell. Biol. 19, 1–11.[Free Full Text]

Davidson, D. G., and Eastham, W. N. (1966). Acute liver necrosis following overdose of paracetamol. Br. Med. J. 5512, 497–499.[Medline]

Davis, D. C., Potter, W. Z., Jollow, D. J., and Mitchell, J. R. (1974). Species differences in hepatic glutathione depletion, covalent binding and hepatic necrosis after acetaminophen. Life Sci. 14, 2099–2109.[ISI][Medline]

Fernandez, G., Mena, M. P., Arnau, A., Sanchez, O., Soley, M., and Ramirez, I. (2000). Immobilization stress induces c-fos accumulation in liver. Cell Stress Chaperones 5, 306–312.[ISI][Medline]

Fernandez-Checa, J. C., Garcia-Ruiz, C., Colell, A., Morales, A., Mari, M., Miranda, M., and Ardite, E. (1998). Oxidative stress: Role of mitochondria and protection by glutathione. Biofactors 8, 7–11.[ISI][Medline]

Filppula, S. A., Yagi, A. I., Kilpelainen, S. H., Novikov, D., FitzPatrick, D. R., Vihinen, M., Valle, D., and Hiltunen, J. K. (1998). {Delta}3,5-{Delta}2,4-dienoyl-CoA isomerase from rat liver. Molecular characterization. J. Biol. Chem. 273, 349–355.[Abstract/Free Full Text]

Fountoulakis, M., Berndt, P., Boelsterli, U. A., Crameri, F., Winter, M., Albertini, S., and Suter, L. (2000). Two-dimensional database of mouse liver proteins: Changes in hepatic protein levels following treatment with acetaminophen or its nontoxic regioisomer 3-acetamidophenol. Electrophoresis 21, 2148–2161.[ISI][Medline]

Garcia-Ruiz, C., Colell, A., Mari, M., Morales, A. and Fernandez-Checa, J. C. (1997). Direct effect of ceramide on the mitochondrial electron transport chain leads to generation of reactive oxygen species. Role of mitochondrial glutathione. J. Biol. Chem. 272, 11369–11377.[Abstract/Free Full Text]

Gygi, S. P., Corthals, G. L., Zhang, Y., Rochon, Y. and Aebersold, R. (2000). Evaluation of two-dimensional gel electrophoresis-based proteome analysis technology. Proc. Natl. Acad. Sci. U.S.A. 97, 9390–9395.[Abstract/Free Full Text]

Hart, S. G., Cartun, R. W., Wyand, D. S., Khairallah, E. A., and Cohen, S. D. (1995). Immunohistochemical localization of acetaminophen in target tissues of the CD-1 mouse: Correspondence of covalent binding with toxicity. Fundam. Appl. Toxicol. 24, 260–274.[ISI][Medline]

Hinson, J. A., and Roberts, D. W. (1992). Role of covalent and non-covalent interactions in cell toxicity: Effects on proteins. Annu. Rev. Pharmacol. Toxicol. 32, 471–510.[ISI][Medline]

Hoebe, K. H., Witkamp, R. F., Fink-Gremmels, J., Van Miert, A. S. and Monshouwer, M. (2001). Direct cell-to-cell contact between Kupffer cells and hepatocytes augments endotoxin-induced hepatic injury. Am. J. Physiol. Gastrointest. Liver Physiol. 280, G720–728.[Abstract/Free Full Text]

Jollow, D. J., Mitchell, J. R., Potter, W. Z., Davis, D. C., Gillette, J. R., and Brodie, B. B. (1973). Acetaminophen-induced hepatic necrosis. II. Role of covalent binding in vivo. J Pharmacol Exp. Ther. 187, 195–202.[ISI][Medline]

Karbowski, M., Kurono, C., Nishizawa, Y., Horie, Y., Soji, T., and Wakabayashi, T. (1997). Induction of megamitochondria by some chemicals inducing oxidative stress in primary cultured rat hepatocytes. Biochim. Biophys. Acta. 1349, 242–250.[ISI][Medline]

Karbowski, M., Kurono, C., Wozniak, M., Ostrowski, M., Teranishi, M., Nishizawa, Y., Usukura, J., Soji, T., and Wakabayashi, T. (1999). Free radical-induced megamitochondria formation and apoptosis. Free Radic. Biol. Med. 26, 396–409.[ISI][Medline]

Kharbanda, S., Nakamura, T., Stone, R., Hass, R., Bernstein, S., Datta, R., Sukhatme, V. P., and Kufe, D. (1991). Expression of the early growth response 1 and 2 zinc finger genes during induction of monocytic differentiation. J. Clin. Invest. 88, 571–577.[ISI][Medline]

Kirillova, I., Chaisson, M., and Fausto, N. (1999). Tumor necrosis factor induces DNA replication in hepatic cells through nuclear factor {kappa}B activation. Cell Growth Differ. 10, 819–828.[Abstract/Free Full Text]

Kitteringham, N. R., Powell, H., Clement, Y. N., Dodd, C. C., Tettey, J. N., Pirmohamed, M., Smith, D. A., McLellan, L. I., and Park, K. B. (2000). Hepatocellular response to chemical stress in CD-1 mice: Induction of early genes and gamma-glutamylcysteine synthetase. Hepatology 32, 321–333.[ISI][Medline]

Kleinman, J. G., Breitenfield, R. V., and Roth, D. A. (1980). Acute renal failure associated with acetaminophen ingestion: Report of a case and review of the literature. Clin. Nephrol. 14, 201–205.[ISI][Medline]

Landin, J. S., Cohen, S. D., and Khairallah, E. A. (1996). Identification of a 54-kDa mitochondrial acetaminophen-binding protein as aldehyde dehydrogenase. Toxicol. Appl. Pharmacol. 141, 299–307.[ISI][Medline]

Laub, D. N., Elmagbari, N. O., Elmagbari, N. M., Hausburg, M. A., and Gardiner, C. S. (2000). Effects of acetaminophen on preimplantation embryo glutathione concentration and development in vivo and in vitro. Toxicol. Sci. 56, 150–155.[Abstract/Free Full Text]

Lawson, J. A., Fisher, M. A., Simmons, C. A., Farhood, A., and Jaeschke, H. (1999). Inhibition of Fas receptor (CD95)-induced hepatic caspase activation and apoptosis by acetaminophen in mice. Toxicol. Appl. Pharmacol. 156(3), 179–186.[ISI][Medline]

Liu, C., Rangnekar, V. M., Adamson, E., and Mercola, D. (1998). Suppression of growth and transformation and induction of apoptosis by EGR-1. Cancer Gene Ther. 5, 3–28.[ISI][Medline]

McLean, A. E., and Nutiall, L. (1978). An in vitro model of liver injury using paracetamol treatment of liver slices and prevention of injury by some antioxidants. Biochem. Pharmacol. 27, 425–430.[ISI][Medline]

Mechtcheriakova, D., Schabbauer, G., Lucerna, M., Clauss, M., De Martin, R., Binder, B. R. and Hofer, E. (2001). Specificity, diversity, and convergence in VEGF and TNF-{alpha} signaling events leading to tissue factor upregulation via EGR-1 in endothelial cells. FASEB J. 15, 230–242.[Abstract/Free Full Text]

Miners, J. O., Drew, R., and Birkett, D. J. (1984). Mechanism of action of paracetamol protective agents in mice in vivo. Biochem. Pharmacol. 33, 2995–3000.[ISI][Medline]

Mirochnitchenko, O., Weisbrot-Lefkowitz, M., Reuhl, K., Chen, L., Yang, C., and Inouye, M. (1999). Acetaminophen toxicity. Opposite effects of two forms of glutathione peroxidase. J. Biol. Chem. 274, 10349–10355.[Abstract/Free Full Text]

Mitchell, J. R., Jollow, D. J., Potter, W. Z., Davis, D. C., Gillette, J. R., and Brodie, B. B. (1973). Acetaminophen-induced hepatic necrosis. I. Role of drug metabolism. J. Pharmacol. Exp. Ther. 187, 185–194.[ISI][Medline]

Mudge, G. H., Gemborys, M. W., and Duggin, G. G. (1978). Covalent binding of metabolites of acetaminophen to kidney protein and depletion of renal glutathione. J. Pharmaco. Exp. Ther. 206, 218–226.[Abstract]

Myers, T. G., Dietz, E. C., Anderson, N. L., Khairallah, E. A., Cohen, S. D., and Nelson, S. D. (1995). A comparative study of mouse liver proteins arylated by reactive metabolites of acetaminophen and its nonhepatotoxic regioisomer, 3`-hydroxyacetanilide. Chem. Res. Toxicol. 8, 403–413.[ISI][Medline]

Nagata, S. (1997). Apoptosis by death factor. Cell 88, 355–365.[ISI][Medline]

Nair, P., Muthukkumar, S., Sells, S. F., Han, S. S., Sukhatme, V. P., and Rangnekar, V. M. (1997). Early growth response-1-dependent apoptosis is mediated by p53. J. Biol. Chem. 272, 20131–20138.[Abstract/Free Full Text]

Neuhoff, V., Arold, N., Taube, D., and Ehrhardt, W. (1988). Improved staining of proteins in polyacrylamide gels including isoelectric focusing gels with clear background at nanogram sensitivity using Coomassie Brilliant Blue G-250 and R-250. Electrophoresis 9, 255–262.[ISI][Medline]

Nguyen, H. Q., Hoffman-Liebermann, B., and Liebermann, D. A. (1993). The zinc finger transcription factor Egr-1 is essential for and restricts differentiation along the macrophage lineage. Cell 72, 197–209.[ISI][Medline]

Pawson, T., and Saxton, T.M. (1999). Signaling networks—Do all roads lead to the same genes? Cell 97, 675–678.[ISI][Medline]

Pennie, W. D., Tugwood, J. D., Oliver, G. J. A., and Kimber, I. (2000). The principles and practice of toxicogenomics: Applications and opportunities. Toxicol. Sci. 54, 277–283.[Free Full Text]

Pinkus, R., Weiner, L. M., and Daniel, V. (1996). Role of oxidants and antioxidants in the induction of AP-1, NF-{alpha}B, and glutathione S-transferase gene expression. J. Biol. Chem. 271, 13422–13429.[Abstract/Free Full Text]

Placke, M. E., Ginsberg, G. L., Wyand, D. S., and Cohen, S. D. (1987). Ultrastructural changes during acute acetaminophen-induced hepatotoxicity in the mouse: A time and dose study. Toxicol. Pathol. 15, 431–438.[ISI][Medline]

Qiu, Y., Benet, L. Z., and Burlingame, A. L. (1998). Identification of the hepatic protein targets of reactive metabolites of acetaminophen in vivo in mice using two-dimensional gel electrophoresis and mass spectrometry. J. Biol. Chem. 273, 17940–17953.[Abstract/Free Full Text]

Rashed, M. S., Myers, T. G., and Nelson, S. D. (1990). Hepatic protein arylation, glutathione depletion, and metabolite profiles of acetaminophen and a non-hepatotoxic regioisomer, 3`-hydroxyacetanilide, in the mouse. Drug Metab. Dispos. 18, 765–770.[Abstract]

Reilly, T. P., Bourdi, M., Brady, J. N., Pise-Masison, C. A., Radonovich, M. F., George, J. W., and Pohl, L. R. (2001). Expression profiling of acetaminophen liver toxicity in mice using microarray technology. Biochem. Biophys. Res. Commun. 282, 321–328.[ISI][Medline]

Rice, J. E. and Lindsay, J. G. (1997). Subcellular fractionation of mitochondria. In Subcellular Fractionation, a Practical Approach (J. M. Graham and D. Rickwood, Eds.), pp. 107–142. IRL Press, Oxford.

Roberts, D. W., Bucci, T. J., Benson, R. W., Warbritton, A. R., McRae, T. A., Pumford, N. R., and Hinson, J. A. (1991). Immunohistochemical localization and quantification of the 3-(cystein-S-yl)-acetaminophen protein adduct in acetaminophen hepatotoxicity. Am. J. Pathol. 138, 359–371.[Abstract]

Rzucidlo, S. J., Bounous, D. I., Jones, D. P., and Brackett, B. G. (2000). Acute acetaminophen toxicity in transgenic mice with elevated hepatic glutathione. Vet. Hum. Toxicol. 42, 146–150.[ISI][Medline]

Salminen, W. F., Jr., Voellmy, R., and Roberts, S. M. (1998) Effect of N-acetylcysteine on heat shock protein induction by acetaminophen in mouse liver. J. Pharmacol. Exp. Ther. 286, 519–524.[Abstract/Free Full Text]

Samali, A., Cai, J., Zhivotovsky, B., Jones, D. P., and Orrenius, S. (1999). Presence of a pre-apoptotic complex of pro-caspase-3, Hsp60 and Hsp10 in the mitochondrial fraction of jurkat cells. EMBO J. 18, 2040–2048.[Abstract/Free Full Text]

Simpson, K. J., Lukacs, N. W., McGregor, A. H., Harrison, D. J., Strieter, R. M., and Kunkel, S. L. (2000). Inhibition of tumour necrosis factor-{alpha} does not prevent experimental paracetamol-induced hepatic necrosis. J. Pathol. 190, 489–494.[ISI][Medline]

Soares, M. B., Bonaldo, M. D. F, Jelene, P., Su, L., Lawton, L., and Efstratiadis, A. (1994). Construction and characterization of a normalized cDNA library. Proc. Natl. Acad. Sci. U.S.A. 91, 9228–9232.[Abstract/Free Full Text]

Tonge, R. P., Kelly, E. J., Bruschi, S. A., Kalhorn, T., Eaton, D. L., Nebert, D. W., and Nelson, S. D. (1998). Role of CYP1A2 in the hepatotoxicity of acetaminophen: Investigations using Cyp1a2 null mice. Toxicol. Appl. Pharmacol. 153, 102–108.[ISI][Medline]

Tonge, R, Shaw, J., Middleton, B., Rowlinson, R., Rayner, S., Young, J., Pognan, F., Hawkins, E., Currie, I., and Davison, M. (2001). Validation and development of fluorescence two-dimensional differential gel electrophoresis proteomics technology. Proteomics 1, 377–396.[ISI][Medline]

Vendemiale, G., Grattagliano, I., Altomare, E., Turturro, N., and Guerrieri, F. (1996). Effect of acetaminophen administration on hepatic glutathione compartmentation and mitochondrial energy metabolism in the rat. Biochem. Pharmacol. 52, 1147–1154.[ISI][Medline]

Vermeulen, N. P., Bessems, J. G., and Van de Straat, R. (1992). Molecular aspects of paracetamol-induced hepatotoxicity and its mechanism-based prevention. Drug Metab. Rev. 24, 367–407.[ISI][Medline]

Walker, R. M., Racz, W. J., and McElligott, T. F. (1980). Acetaminophen-induced hepatotoxicity in mice. Lab. Invest. 42, 181–189.[ISI][Medline]

Williams, R. J. (1998). Calcium: Outside/inside homeostasis and signalling. Biochim. Biophys. Acta 1448, 153–165.[ISI]

Zaher, H., Buters, J. T., Ward, J. M., Bruno, M. K., Lucas, A. M., Stern, S. T., Cohen, S. D., and Gonzalez, F. J. (1998). Protection against acetaminophen toxicity in CYP1A2 and CYP2E1 double-null mice. Toxicol. Appl. Pharmacol. 152, 193–199.[ISI][Medline]