Relationship between Hepatic Gene Expression Profiles and Hepatotoxicity in Five Typical Hepatotoxicant-Administered Rats

Keiichi Minami*, Toshiro Saito{dagger}, Masatoshi Narahara{dagger}, Hiroyuki Tomita{dagger}, Hirokazu Kato{dagger}, Hisashi Sugiyama{dagger}, Miki Katoh*, Miki Nakajima* and Tsuyoshi Yokoi*,1

* Drug Metabolism and Toxicology, Division of Pharmaceutical Sciences, Kanazawa University, Kanazawa, Japan, and {dagger} Life science group, Hitachi Ltd., Saitama, Japan

1 To whom correspondence should be addressed at Drug Metabolism and Toxicology, Division of Pharmaceutical Sciences, Kanazawa University, Kakuma-machi, Kanazawa 920-1192, Japan. Fax: +81-76–234–4407. E-mail: tyokoi{at}kenroku.kanazawa-u.ac.jp.

Received May 27, 2005; accepted June 2, 2005


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY DATA
 REFERENCES
 
In the field of gene expression analysis, DNA microarray technology has a major impact on many different areas including toxicogenomics, such as in predicting the adverse effects of new drug candidates and improving the process of risk assessment and safety evaluation. In this study, we investigated whether there is relationship between the hepatotoxic phenotypes and gene expression profiles of hepatotoxic chemicals measured by DNA microarray analyses. Sprague-Dawley rats (6 weeks old) were administered five hepatotoxicants: acetaminophen (APAP), bromobenzene, carbon tetrachloride, dimethylnitrosamine, and thioacetamide. Serum biochemical markers for liver toxicity were measured to estimate the maximal toxic time of each chemical. Hepatic mRNA was isolated, and the gene expression profiles were analyzed by DNA microarray containing 1,097 drug response genes, such as cytochrome P450s, other phase I and phase II enzymes, nuclear receptors, signal transducers, and transporters. All the chemicals tested generated specific gene expression patterns. APAP was sorted to a different cluster from the other four chemicals. From the gene expression profiles and maximal toxic time estimated by serum biochemical markers, we identified 10 up-regulated genes and 10 down-regulated genes as potential markers of hepatotoxicity. By Quality-Threshold (QT) clustering analysis, we identified major up- and down-regulated expression patterns in each group. Interestingly, the average gene expression patterns from the QT clustering were correlated with the mean value profiles from the biochemical markers. Furthermore, this correlation was observed at any extent of hepatotoxicity. In this study, we identified 17 potential toxicity markers, and those expression profiles could estimate the maximal toxic time independently of the hepatotoxicity levels. This expression profile analysis could be one of the useful tools for evaluating a potential hepatotoxicant in the drug development process.

Key Words: gene expression profiles; hepatotoxicity; DNA microarray.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY DATA
 REFERENCES
 
The aim of toxicological studies is to detect adverse effects of a chemical on an organism based on observed toxicity markers (i.e., serum biochemical markers and chemical-specific gene expression) or phenotypic outcomes. In the past several years, novel systems, especially microarray technology, have been developed, allowing the simultaneous measurement of gene expression at the RNA level. Microarray technology can be used to elucidate the mechanisms of chemical-induced toxicity and have the possibility of being used as a tool for predicting the adverse effects of new drug candidates and improving the process of risk assessment and safety evaluation.

The liver is one of the first organs to be exposed to peroral-administered chemicals via the portal vein. Chemical concentrations in the liver are often much higher than the peak plasma concentration. The liver is also the major site for xenobiotic metabolism, and various chemicals can lead to the formation of active metabolites with toxic effects. The high concentration exposure and metabolic activity make the liver one of the primary targets for various types of chemical-induced toxicity.

The five typical hepatotoxicants chosen in this study were acetaminophen (APAP, p-acetamidophenol), bromobenzene (BB), carbon tetrachloride (CT), dimethylnitrosamine (DMN), and thioacetamide (TA). APAP is known as a mild analgesic drug, but it is a potent hepatotoxicant at high doses and in persons with enhanced susceptibility. APAP is largely (apparently more than 80%) converted to conjugates of glucuronate and sulfate. A minor amount, less than 5%, is metabolized to an active metabolite, mainly N-acetyl-p-benzoquinone imine (NAPQI) by cytochrome P450 (CYP) 2E1, which binds promptly to glutathione (GSH). Other metabolites (5 to 15%) appear to have no toxicity (Zimmerman, 1999Go). When an active metabolite exceeds the GSH content, excess metabolite binds to tissue molecules and manifests toxicity such as necrosis. BB, a traditional hepatotoxicant, is subjected to cytochrome P450-mediated epoxidation, and a major metabolite is 3,4-epoxide. Detoxication of the active metabolite is by GSH conjugation. At high BB doses, due to the conjugation to the epoxides, liver GSH shortage and secondary reactions such as lipid peroxidation, intracellular calcium alteration, and mitochondrial dysfunction finally lead to cell death (Heijne et al., 2003Go, Zimmerman, 1999Go). CT is a potent hepatotoxicant, and a single dose leads promptly to severe necrosis and steatosis. CT liver necrosis is caused by trichloromethyl free radicals from a CYP2E1-mediated pathway. The covalent binding of trichloromethyl to cell protein is considered the initial step of sequential events leading to membrane lipid peroxidation and, finally, to cell necrosis (Jeong, 1999Go; Zimmerman, 1999Go). DMN is the most potent toxicant of all dialkylnitrosamines and leads to hemorrhagic necrosis and steatosis. The DMN toxicity process is as follows: first, demethylation of CYP2E1 to monomethylnitrosamine; second, spontaneous change of diasomethane; finally, methylation of cell components (Zimmerman, 1999Go). Thioacetamide (TA) is a potent hepatotoxicant that requires metabolic activation by mixed-function oxidases. Generally, CYP2B, CYP2E1, and FMOs metabolize TA to its toxic metabolites (Hunter et al., 1977Go; Wang et al., 2000Go), and these intermediate metabolites might bind to cellular proteins by the formation of acetylimidolysine derivatives (Dyroff and Neal, 1981Go). TA is apparently converted to thioacetamide-S-oxide and is presumably converted to an active toxic metabolite that binds covalently to tissue molecules, provoking necrosis (Zimmerman, 1999Go).

The aims of this study were to find available toxicity marker genes, to investigate the correlation between biochemical markers and gene expression profiles, and to create a new evaluation method using DNA microarray. In this study, we attempted to apply our microarray of drug-response gene expressions for the evaluation of chemical-induced hepatotoxicity in rats.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY DATA
 REFERENCES
 
Animals and chemicals.
Male Sprague-Dawley rats (5-week old, 130–150 g) were obtained from SLC Japan (Hamamatsu, Japan). Animals were housed in a controlled environment (temperature 25 ± 1°C, humidity 50 ± 10%, and 12-h light/12-h dark cycle) in the institutional animal facility with access to food and water ad libitum. Animals were acclimatized for a week before use in experiments. Animal maintenance and treatment were conducted in accordance with the National Institutes of Health Guide for Animal Welfare of Japan, as approved by Institutional Animal Care and Use Committee of Kanazawa University, Japan. APAP, BB, CT, DMN, and TA were obtained Wako Pure Chemical Industries (Osaka, Japan). ISOGEN, RNA extraction reagent was from Nippon Gene (Tokyo, Japan). N-hydroxysuccinimide (NHS)-ester Cy3 or Cy5 was from GE Healthcare Amersham Biosciences (Buckinghamshire, UK). ReverTra Ace (Moloney Murine Leukemia Virus Reverse Transcriptase RnaseH Minus) was from Toyobo (Tokyo, Japan). Random hexamer and SYBR® Premix Ex TaqTM (Perfect Real Time) were from Takara (Osaka, Japan). All primers were commercially synthesized at Hokkaido System Sciences (Sapporo, Japan). Other chemicals were of the highest grade commercially available.

Administration of chemicals and assessment of liver injury.
Eighty-eight rats were assigned to 22 groups (four rats/group). The dosing solutions were prepared as follows with each vehicle. APAP: 500 mg/kg in corn oil; BB: 2.5 mmol/kg in corn oil; CT: 1 ml/kg in corn oil; DMN: 20 mg/kg in saline; TA: 400 mg/kg in saline; control: vehicle for saline or corn oil group. The chemicals were intraperitoneally injected in a single bolus at a volume of 2 ml/kg. At the indicated time (6, 12, 24, 48 h after administration), the rats were sacrificed, and the liver and serum samples were collected. Four typical biochemical markers for hepatotoxicity (aspartate aminotransferase, AST; alanine aminotransferase, ALT; lactate dehydrogenase, LDH; alkaline phosphatase, ALP) were measured by SRL, Inc. (Tokyo, Japan).

RNA isolation.
Total hepatic RNA was isolated using ISOGEN. Approximately 100 mg of whole liver were lysed with 1.0 ml of the lysis solution. Chloroform (200 µl) was added and vortexed vigorously for 15 s. The mixture was centrifuged at 15,000 x g for 15 min at 4°C. The aqueous phase was transferred carefully to a new tube, and the RNA was precipitated with 0.5 ml of isopropyl alcohol for 10 min at room temperature. The mixture was centrifuged at 15,000 x g for 10 min. After washing with 75% ethanol, the pellet was dissolved in diethylpyrocarbonate-treated water. Equal amounts of total mRNA from each hepatotoxicant-administered sample were pooled and used for the microarray analysis and real-time reverse transcriptase (RT)-PCR.

In vitro amplification and DNA microarray.
cDNA targets were prepared from pooled total RNA by in vitro transcription reaction as described previously (Luo et al., 1999Go). Amplified RNA (6 µg) was reverse transcribed by random hexamer and aminoallyl-dUTP. The synthesized cDNA was labeled with NHS-ester Cy3 or Cy5 (Hughes et al., 2001Go). The labeled cDNA was applied to the cDNA microarray (Rat Drug Response Chip containing 1,097 genes, Hitachi, Tokyo, Japan). In order to confirm the microarray data in the APAP group, a Rat cDNA Microarray kit G4105A containing 14,815 genes (Agilent Technologies, Palo Alto, CA) was used. Hybridization was performed at 62°C for 14 h. After washing, the microarray was scanned on a ScanArray 5000 (Packard BioChip Technologies, Billerica, MA), and the image was analyzed using QuantArray software (Packard BioChip Technologies). The signal intensity of each spot was calibrated by subtraction of the intensity of the control.

Real-time RT-PCR.
Rat Cathepsin L (CtsL), Diazepam binding inhibitor (Dbi), Heme oxygenase 1 (Hmox1), Sulfotransferase 1a2 (Sult1a2), T-cell death associated gene (Tdag), and GAPDH were quantified by real-time RT-PCR. Primer sequences using in this study were as follows: CtsL, 5'-TCT ACT ATG AAC CCA ACT G-3' and 5'-GAT TCA AGT ACC ATG GTC T-3'; Dbi, 5'-CCA ACT GAT GAA GAG ATG CTG T-3' and 5'-CCC TAA CAT ATC AGA GCC ATG T-3'; Hmox1, 5'-ATA GAG CGA AAC AAG CAG A-3' and 5'-TAG AGC TGT TTG AAC TTG G-3'; Sult1a2, 5'-TCA TTG AGT GGA CTT TGC CTT-3' and 5'-CAC TTT TCC AGC TTT GAA CTG-3'; Tdag, 5'-CCA AGC AGG TAC AAC ATC AG-3' and 5'-TTC TGC CTC GTA GAC TTG AC-3'. For RT process, total RNA (4 µg) and 150 ng random hexamer were mixed and incubated at 70°C for 10 min. RNA solution was added to a reaction mixture containing 100 units of ReverTra Ace, reaction buffer, and 0.5 mM dNTPs in a final volume of 40 µl. The reaction mixture was incubated at 30°C for 10 min, 42°C for 1 h, and heated at 98°C for 10 min to inactivate the enzyme. Real-time PCR was performed using the Smart Cycler® (Cepheid, Sunnyvale, CA) with Smart Cycler® software (Ver. 1.2b). PCR mixture contained 1 µl of template cDNA, SYBR® Premix Ex TaqTM solution and 10 pmol of sense and antisense primers. The PCR condition for GAPDH and Sult1a2 was as follows: after an initial denaturation at 95°C for 30 s, the amplification was performed by denaturation at 94°C for 4 s, annealing and extension at 64°C for 20 s for 45 cycles. The PCR condition for other genes was as follows: after an initial denaturation at 95°C for 20 s, the amplification was performed by denaturation at 95°C for 5 s, annealing at 55°C for 10 s, and extension at 72°C for 15 s for 45 cycles. Amplified products were monitored directly by measuring the increase of the dye intensity of the SYBR Green I (Molecular Probes, Eugene, OR) that binds to double-strand DNA amplified by PCR.

Data management.
Fold-change determination, experiment normalization, and clustering analysis were performed with GeneSpring software (Agilent Technologies). Gene expression values for each chip were normalized to the intensity-dependent (LOWESS) normalization built in GeneSpring. In the Quality-Threshold (QT) clustering analyses, a standard correlation method was used in this study. Fold-change filters included the requirement that the genes be present in at least 200% of the administered samples for up-regulated genes, and 50% of controls for down-regulated genes.

Statistics.
The t-test was used to detect significant differences between the means of two groups relative to the observed variance within groups.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY DATA
 REFERENCES
 
Assessment of Liver Toxicity
The serum biochemical markers in the five chemical-administered groups were measured at 6, 12, 24, and 48 h after administration (Table 1). The AST activities of the APAP-, BB-, CT-, DMN-, and TA-administered groups were significantly high at 12, 24, 6, 48, and 24 h after administration, respectively. The changes of the ALT activities showed almost the same pattern as the AST activities in all groups. The LDH activity increased significantly in the BB-, CT-, and TA-administered groups at 24, 6, and 24 h, respectively. In the DMN-administered group, the LDH activity significantly decreased at 24 and 48 h in a time-dependent manner. In the APAP-administered group, no significant change was observed. The ALP activity decreased significantly in the APAP-, BB-, and CT-administered groups at 6 and 12, 12, and 12 and 24 h, respectively. In the TA-administered group, the ALP activity significantly increased at 12 h. In the DMN-administered group, no significant change of the ALP activity was observed. Taking these results into consideration, the maximal toxic times of APAP, BB, CT, DMN, and TA were estimated as 12, 24, 6, 48, and 24 h, respectively.


View this table:
[in this window]
[in a new window]
 
TABLE 1 Changes of Biochemical Markers in Five Chemical-Administered Rats

 
Genes Significantly Changed by at Least Four Chemicals at the Toxic Time Points
Analysis of the gene expression profile showed that 52 (25 up-regulated and 27 down-regulated) of 1,097 genes were changed above 2-fold by at least four chemicals at the maximal toxic times described above (Table 2). Eleven genes were changed above 2-fold in all the chemical administered groups (3 up-regulated and 8 down-regulated, Table 2), and 40 genes in four of the five chemical administered groups (21 up-regulated and 19 down-regulated, Table 2).


View this table:
[in this window]
[in a new window]
 
TABLE 2 Genes Changed of Their Expression in At Least Four Chemical-Administered Rats

 
The genes in Table 2 are presented by two-way hierarchical clustering (Fig. 1). In the up-regulated genes at the maximal toxic times (Fig. 1A), CT-, TA-, and BB-administered groups were sorted in a similar cluster. However, all APAP-administered groups were sorted in a different cluster from other groups. The DMN-administered groups from 12 to 48 h were sorted in a similar cluster. Among the down-regulated genes at the maximal toxic times (Fig. 1B), those of the CT- and TA-administered groups, BB- and DMN-administered groups, and APAP-administered groups were sorted in a similar cluster, respectively. In both cluster analyses, 6 and 12 h of the CT-administered groups and 24 h of the TA-administered groups were sorted in a similar cluster.



View larger version (59K):
[in this window]
[in a new window]
 
FIG. 1. Two-way hierarchical cluster images of genes whose expressions were altered more than 2-fold at the maximal toxic times by at least four chemicals. The results of hierarchical cluster analyses are shown with a dendrogram for (A) 25 up-regulated genes and (B) 27 down-regulated genes listed in Table 2. Gene expression data are expressed as fold of control values, and the range of change represented by colors at the bottom of the cluster images. The expression pattern of each gene is displayed here as a horizontal strip.

 
Gene Expression Profiles and Maximal Toxic Time
The peak profiles of 10 up- or down-regulated genes in Table 2 were each overlapped at the maximal toxic times (Fig. 2). The expression profiles in each of the chemical-administered groups showed similar patterns at all the time points investigated in the present study.



View larger version (28K):
[in this window]
[in a new window]
 
FIG. 2. Time-dependent changes of the expressions of 20 representative genes in Table 2. The peaks of the expression in up- or down-regulated genes were overlapped with the maximal toxic times examined in the present study. Red arrows indicate the maximal toxic time in each chemical-administrated group estimated from the change of biochemical markers. Official gene name is described in Table 2.

 
To confirm the gene expression profiles of DNA microarray as shown in Figure 2, real-time RT-PCR was performed (Fig. 3). The expression profiles of all five genes were almost the same as those of DNA microarray. The extent of these gene expressions was higher than that of DNA microarray.



View larger version (25K):
[in this window]
[in a new window]
 
FIG. 3. Real-time RT-PCR analysis of the expression of the representative genes. Total RNA samples from four rats were pooled and used for real-time RT-PCR analysis. This figure contains the representative three up-regulated and two down-regulated genes as shown in Figure 2. Official gene name is described in Table 2.

 
QT Clustering for Minimal Correlation
In order to estimate the majority of the gene expression profiles, QT clustering analysis was performed. In this process, we used the GeneSpring QT clustering algorism. The analysis setting for the minimal cluster size was 100 genes, and the minimal correlation coefficient was 0.5 for the Rat Drug Response Chip containing 1,097 genes. Probes with a certain expression level higher than 0.01 in all the administered samples (759 of 1,097 genes) were used. In all groups, the up-regulated and down-regulated types of clusters were identified and expressed as average values (Fig. 4A). In the up-regulated type cluster, the APAP-, BB-, CT-, DMN-, and TA-administered groups contained 109, 105, 143, 139, and 153 genes, respectively (upper part of Fig. 4A). The down-regulated type clusters of the APAP-, BB-, CT, DMN-, and TA-administered groups contained 122, 159, 180, 144, and 163 genes, respectively (lower part of Fig. 4A). As a result, all clusters except those of DMN reflected the maximal toxic time.



View larger version (16K):
[in this window]
[in a new window]
 
FIG. 4. QT clustering analysis of hepatic gene expression in rats administered the five chemicals. (A) The clustering setting for the correlation coefficient was 0.5 and contained more than 100 genes. The method for QT clustering analysis was described in Materials and Methods. Each figure shows an expression profile using the average of the clustered genes. Arrows indicate the maximal toxic time in each chemical-administered group estimated from the change of biochemical markers. In the upper group, the APAP, BB, CT, DMN, and TA clusters contained 109, 105, 143, 139, and 153 genes, respectively. In the lower group, the APAP, BB, CT, DMN, and TA clusters contained 122, 159, 180, 144, and 163 genes, respectively. (B) QT clustering analysis in APAP-administered rats performed using Agilent Rat cDNA microarray kit G4105A. The clustering setting for the correlation coefficient was 0.68 and contained more than 1,000 genes. The upper cluster contained 1,058 genes, and the lower cluster contained 1,106 genes.

 
In order to confirm the data with the Rat Drug Response Chip, analysis with Agilent Rat cDNA microarray kit G4105A was performed only in the APAP-administered group. QT clustering was also performed, and the analysis setting for the minimal cluster size was 1,000 genes, and the minimal correlation coefficient was 0.68. Probes with a certain expression level higher than 0.01 in all administered samples (14,474 of 14,815 genes) were used. Up-regulated and down-regulated type clusters contained 1,058 and 1,106 genes, respectively, and were expressed as mean values (Fig. 4B). As expected, both clusters of APAP showed almost the same profiles as those of the Rat Drug Response Chip.

Genes Appeared in All Five of the Chemical-Administered Groups
For further analysis, the gene expression profiles identified in the gene clusters of all chemicals in Figure 4A were displayed. Three genes were contained in all the up-regulated type clusters and 17 genes in the down-regulated type clusters of all the chemical-administered groups as shown in Figure 4. The expression profiles of 17 down-regulated genes are shown in Figure 4A. The blue lines indicate the expression profiles of the genes that appeared not only in Figure 2 but also in Figure 4. The expression profiles of the genes that appeared in all chemical groups but only in Figure 4 are indicated by black lines. The three common genes in the up-regulated type cluster were not listed in Figure 2 (data not shown). In the down-regulated type cluster, eight genes of all the regulated groups were also listed in Figure 2. The 3 up-regulated and 17 down-regulated genes were analyzed in the following QT clustering.

QT Clustering for the Genes That Appeared in at Least Four of Five Chemical-Administered Groups
For further analysis, the genes that appeared in at least four of the five chemical-administered groups were analyzed. Thirty-seven genes were identified in the up-regulated type cluster and 60 genes in the down-regulated type cluster, as shown in Figure 4. In the up-regulated type cluster, 7 of 37 genes were overlapped with the genes of the up-regulated groups shown in Figure 2. In the down-regulated type cluster, all genes of the down-regulated groups in Figure 2 were overlapped. Further QT clustering was performed using the 37 up-regulated or 60 down-regulated genes identified in Figure 4A. The analysis setting for the minimal cluster size was 10 genes and the minimal correlation coefficient was 0.65. Twenty-two of 37 genes were identified as the up-regulated type cluster, and 44 of 60 genes as the down-regulated type cluster. The profiles of the 22 and 44 identified genes and the average profile are shown in Figures 5B and 5C, respectively. The expression profiles of the genes that also appeared in Figure 2 are indicated by blue lines. The expression profiles of the clustered gene that did not appear in Figure 2 are indicated by black lines. Seven of 10 genes in the up-regulated group in Figure 2 and all 10 genes in the down-regulated group in Figure 2 were overlapped as a result of the clustering (Figs. 5B and 5C). As shown in Figures 5A, 5B, and 5C, the average up-regulated or down-regulated peaks were correlated with the maximal toxic times.



View larger version (38K):
[in this window]
[in a new window]
 
FIG. 5. Gene expression profiles of each group of chemical-administered rats identified by QT clustering analysis. The blue lines show the expression profiles of the genes that appeared in Figure 2. The red lines indicate the average of the expression. The red arrows indicate the maximal toxic time in each chemical-administered group estimated from the change of biochemical markers. (A) The expression profiles of genes down-regulated in the five chemical-administered groups. The expression profiles of genes that did not appear in Figure 2 are indicated by black lines. (B) The expression profiles of genes down-regulated in four of five chemical-administered groups. (C) The expression profiles of genes up-regulated in four of five chemical-administered groups. The expression profiles of the clustered genes that did not appear in Figure 2 are indicated by black lines. The analysis setting for the correlation coefficient was 0.65 and contained more than 10 genes.

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY DATA
 REFERENCES
 
Gene expression changes have been used routinely to obtain specific mechanistic information concerning the type of action of a toxicant. Toxicogenomics is an approach that applies microarray technology to toxicological evaluation paradigms. Microarrays are now available that contain huge numbers of genes, even those whose functions are not clear, and the effects of a chemical on the gene expression cannot be assessed with a single microarray experiment.

In this study, we evaluated five typical hepatotoxic chemicals these were thought to cause zone-3 necrosis (Zimmerman, 1999Go). The dose levels of APAP (Price and Jollow, 1982Go; Sato and Izumi, 1989Go), BB (Chakrabarti and Brodeur, 1984Go), CT (Theocharis et al., 2001Go), DMN (Asakura et al., 1998Go), and TA (Wang et al., 2000Go; Zaragoza et al., 2000Go) were selected based on their association with detectable hepatotoxicity as previously reported. These data confirmed that the hepatotoxicity models of all the chemical-administered groups were successfully conducted, and the toxic time points of APAP, BB, CT, DMN, and TA were estimated as 12, 24, 6, 48, and 24 h, respectively. The maximal toxic times of BB (Chakrabarti and Brodeur, 1984Go), DMN (Asakura et al., 1998Go), and TA (Wang et al., 2000Go; Zaragoza et al., 2000Go, respectively) in rats were the same as previously reported. In the CT-administered rats, AST and ALT elevated significantly at 6 and 48 h in the present study, but the CT toxicity assessed by serum AST and ALT increased at 6 to 24 h (AST) or 6 to 36 h (ALT) in a time-dependent manner, respectively (Zimmerman, 1999Go). In APAP-administered rats, AST and ALT elevated significantly at 6 and 12 h, but the serum AST and ALT were previously reported to be evaluated at 24 h by APAP administration (Hong et al., 1992Go; Wang et al., 1999Go). In the present study, the biochemical markers reflected the major gene expression profiles (Figs. 2, 3, 4, and 5).

We performed hierarchical clustering using gene groups whose expression levels were distinctively changed at the toxic time points (Fig. 1). At the maximal toxic time, CT and TA were sorted in a relatively close cluster. At the maximal toxic time, BB and DMN were sorted in a similar cluster. However, all APAP-administered groups were sorted in a different cluster. We performed many other types of hierarchical clustering by using other gene groups such as enzymes, signal transductions, and so on, and the results were almost the same as shown in Figure 1 (data not shown). Although studies concerning many hepatotoxicants including APAP, BB, CT, DMN, and TA administered to Sprague-Dawley rats have been reported (Kier et al., 2004Go, McMillian et al., 2004aGo,bGo), there has been no attempt of such a hierarchical clustering analysis using five chemicals. Thus, that the gene expression profiles of APAP administration were different from other those of the four chemicals constitutes new information.

In handling microarray data, it is necessary to consider what kinds of effects are relevant to the purpose of the experiments. We identified 20 representative genes whose up- or down-regulation peaks overlapped with the maximal toxic time (Fig. 2). In each of the chemical-administered rats, almost all genes identified in the present study showed similar expression profiles. The expression profile was confirmed in five representative genes, resulting in the overlapped profile with that of DNA microarray. Moreover, 17 of 20 genes were also identified by QT clustering analysis (Fig. 5). Data from QT clustering are independent of the hepatotoxicity estimated by serum biochemical markers. The present results showed the potential usefulness of 17 identified genes as toxicity markers. Most of the identified genes were not described previously as having a relationship with hepatotoxic chemicals. For example, TA administration up-regulated rat aldorase A mRNA (Bulera et al., 2001Go). CtsL was up-regulated at the hepatic mRNA level after 24 h in BB- and TA-administered rats (Heijne et al., 2003Go; Bulera et al., 2001Go, respectively). Hmox1 was reported to be up-regulated by four chemicals, APAP (Chiu et al., 2002Go), BB (Heijne et al., 2003Go), CT (Montosi et al., 1998Go), and TA (Bulera et al., 2001Go; Matsuura et al., 1988Go), in rats in vivo. In the present study, the expression of CYP2E1, which possibly catalyzes the induced toxicity of the five chemicals (Jeong, 1999Go; Lauriault et al., 1992Go; Wang et al., 2000Go; Zimmerman, 1999Go) was only slightly changed (data not shown), suggesting that the induction of CYP2E1 would not be involved.

QT clustering analysis is usually performed to determine the specific gene expression patterns. The resulting clusters gave us a good indication of the types of gene expression patterns that existed in the data (Heyer et al., 1999Go). The gene expression patterns obtained from QT clustering analysis were major patterns expressed with each type of chemical administration (Fig. 4A). However, the extent of toxicity estimated by serum biochemical markers was different for each chemical. The average of each gene expression profile from QT clustering was overlapped with the changes of the serum biochemical markers. Furthermore, we performed QT clustering using Agilent Rat cDNA microarray kit G4105A in APAP-administered rat samples, the results of which showed almost the same expression profile (Fig. 4B). From this cDNA microarray, we confirmed the reproducibility of the new QT clustering data.

In conclusion, we identified 17 potential toxicity markers. It was not clarified whether all of the genes were related to the development of toxicity or whether these genes were related to each other. In the present study, we found that the expression profile analysis of the chemical administration could be used to estimate the maximal toxic time independently of the toxicity grade. This expression profile analysis could also be a tool for identifying potential hepatotoxicants. This would be a new approach for determining hepatotoxicity by microarray analysis. We demonstrated that these two approaches, serum biochemical markers and two different QT clustering analyses, yielded the same results. For further studies, detailed time courses, multidose levels, and evaluation of other hepatotoxicants will be performed.


    SUPPLEMENTARY DATA
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY DATA
 REFERENCES
 
Supplementary data are available online at www.toxsci.oxfordjournals.org.


    ACKNOWLEDGMENTS
 
This work was supported in part by a grant from the Ministry of Education, Science, Sports, and Culture of Japan, and by Research on Advanced Medical Technology, Health and Labor Science Research Grants from the Ministry of Health, Labor and Welfare of Japan. We thank Mr. Brent Bell for reviewing the manuscript.


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUPPLEMENTARY DATA
 REFERENCES
 
Asakura, S., Daimon, H., Sawada, S., and Sagami, F. (1998). A short-term assessment of tumor-promotion activity in the livers of rats treated with two genotoxic methylating agents: Dimethylnitrosamine and methylnitrosourea. Toxicol. Lett. 98, 155–167.[CrossRef][ISI][Medline]

Bulera, S. J., Eddy, S. M., Ferguson, E., Jatkoe, T. A., Reindel, J. F., Bleavins, M. R., and De La Iglesia, F. A. (2001). RNA expression in the early characterization of hepatotoxicants in Wistar rats by high-density DNA microarrays. Hepatology 33, 1239–1258.[CrossRef][ISI][Medline]

Chakrabarti, S., and Brodeur, J. (1984). Dose-dependent metabolic excretion of bromobenzene and its possible relationship to hepatotoxicity in rats. J. Toxicol. Environ. Health 14, 379–391.[ISI][Medline]

Chiu, H., Brittingham, J. A., and Laskin, D. L. (2002). Differential induction of heme oxygenase-1 in macrophages and hepatocytes during acetaminophen-induced hepatotoxicity in the rat: Effects of hemin and biliverdin. Toxicol. Appl. Pharmacol. 181, 106–115.[CrossRef][ISI][Medline]

Dyroff, M. C., and Neal, R. A. (1981). Identification of the major protein adduct formed in rat liver after thioacetamide administration. Cancer Res. 41, 3430–3435.[Abstract]

Heijne, W. H., Stierum, R. H., Slijper, M., van Bladeren, P. J., and van Ommen, B. (2003). Toxicogenomics of bromobenzene hepatotoxicity: A combined transcriptomics and proteomics approach. Biochem. Pharmacol. 65, 857–875.[CrossRef][ISI][Medline]

Heyer, L. J., Kruglyak, S., and Yooseph, S. (1999). Exploring expression data: Identification and analysis of coexpressed genes. Genome Res. 11, 1106–1115.[CrossRef]

Hong, R. W., Rounds, J. D., Helton, W. S., Robinson, M. K., and Wilmore, D. W. (1992). Glutamine preserves liver glutathione after lethal hepatic injury. Ann. Surg. 215, 114–119.[ISI][Medline]

Hughes, T. R., Mao, M., Jones, A. R., Burchard, J., Marton, M. J., Shannon, K. W., Lefkowitz, S. M., Ziman, M., Schelter, J. M., Meyer, M. R., et al. (2001). Expression profiling using microarrays fabricated by an ink-jet oligonucleotide synthesizer. Nat. Biotech. 19, 342–347.[CrossRef][ISI][Medline]

Hunter, A. L., Holscher, M. A., and Neal, R. A. (1977). Thioacetamide-induced hepatic necrosis. I. Involvement of the mixed-function oxidase enzyme system. J. Pharmacol. Exp. Ther. 200, 439–448.[Abstract]

Jeong, H. G. (1999). Inhibition of cytochrome P450 2E1 expression by oleanolic acid: Hepatoprotective effects against carbon tetrachloride-induced hepatic injury. Toxicol. Lett. 105, 215–222.[CrossRef][ISI][Medline]

Kier, L. D., Neft, R., Tang, L., Suizu, R., Cook, T., Onsurez, K., Tiegler, K., Sakai, Y., Ortiz, M., Nolan, T., et al. (2004). Applications of microarrays with toxicologically relevant genes (tox genes) for the evaluation of chemical toxicants in Sprague Dawley rats in vivo and human hepatocytes in vitro. Mutat. Res. 549, 101–113.[ISI][Medline]

Lauriault, V. V., Khan, S., and O'Brien, P. J. (1992). Hepatocyte cytotoxicity induced by various hepatotoxins mediated by cytochrome P-450IIE1: Protection with diethyldithiocarbamate administration. Chem. Biol. Interact. 81, 271–289.[CrossRef][ISI][Medline]

Luo, L., Salunga, R. C., Guo, H., Bittner, A., Joy, K. C., Galindo, J. E., Xiao, H., Rogers, K. E., Wan, J. S., Jackson, M. R., et al. (1999). Gene expression profiles of laser-captured adjacent neuronal subtypes. Nat. Med. 5, 117–122.[CrossRef][ISI][Medline]

Matsuura, Y., Fukuda, T., Yoshida, T., and Kuroiwa, Y. (1988). Inhibitory effect of zinc-protoporphyrin on the induction of heme oxygenase and the associated decrease in cytochrome P-450 content in rats. Toxicology 50, 169–180.[CrossRef][ISI][Medline]

McMillian, M., Nie, A. Y., Parker, J. B., Leone, A., Bryant, S., Herlich, J., Yieh, L., Bittner, A., Liu, X., Wan, J., et al. (2004a). Inverse gene expression patterns for macrophage activating hepatotoxicants and peroxisome proliferators in rat liver. Biochem. Pharmacol. 67, 2141–2165.[CrossRef][ISI][Medline]

McMillian, M., Nie, A. Y., Parker, J. B., Leone, A., Bryant, S., Kemmerer, M., Herlich, J., Liu, Y., Yieh, L., Bittner, A., et al. (2004b). A gene expression signature for oxidant stress/reactive metabolites in rat liver. Biochem Pharmacol. 68, 2249–2261.[CrossRef][ISI][Medline]

Montosi, G., Garuti, C., Iannone, A., and Pietrangelo, A. (1998). Spatial and temporal dynamics of hepatic stellate cell activation during oxidant-stress-induced fibrogenesis. Am. J. Pathol. 152, 1319–1326.[Abstract]

Price, V. F., and Jollow, D. J. (1982). Increased resistance of diabetic rats to acetaminophen-induced hepatotoxicity. J. Pharmacol. Exp. Ther. 220, 504–513.[Abstract]

Sato, C., and Izumi, N. (1989). Mechanism of increased hepatotoxicity of acetaminophen by the simultaneous administration of caffeine in the rat. J. Pharmacol. Exp. Ther. 248, 1243–1247.[Abstract]

Theocharis, S. E., Margeli, A. P., Skaltsas, S. D., Spiliopoulou, C. A., and Koutselinis, A. S. (2001). Induction of metallothionein in the liver of carbon tetrachloride intoxicated rats: An immunohistochemical study. Toxicology 161, 129–138.[CrossRef][ISI][Medline]

Wang, T., Shankar, K., Ronis, M. J., and Mehendale, H. M. (2000). Potentiation of thioacetamide liver injury in diabetic rats is due to induced CYP2E1. J. Pharmacol. Exp. Ther. 294, 473–479.[Abstract/Free Full Text]

Wang, P. Y., Kaneko, T., Wang, Y., and Sato, A. (1999). Acarbose alone or in combination with ethanol potentiates the hepatotoxicity of carbon tetrachloride and acetaminophen in rats. Hepatology 29, 161–165.[CrossRef][ISI][Medline]

Zaragoza, A., Andres, D., Sarrion, D., and Cascales, M. (2000). Potentiation of thioacetamide hepatotoxicity by phenobarbital pretreatment in rats. Inducibility of FAD monooxygenase system and age effect. Chem. Biol. Interact. 124, 87–101.[CrossRef][ISI][Medline]

Zimmerman, H. J. (1999). Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver, 2nd ed., pp. 229–274. Lippncott Williams & Wilkins, Philadelphia.





This Article
Abstract
Full Text (PDF)
Supplementary Data
All Versions of this Article:
87/1/296    most recent
kfi235v1
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (1)
Disclaimer
Request Permissions
Google Scholar
Articles by Minami, K.
Articles by Yokoi, T.
PubMed
PubMed Citation
Articles by Minami, K.
Articles by Yokoi, T.