Time Dependent Gene Expression Changes in the Liver of Mice Treated with Benzene.

Benzene is used as a general purpose solvent. Benzene metabolism starts from phenol and ends with p-benzoquinone and o-benzoquinone. Liver injury inducted by benzene still remains a toxicologic problem. Tumor related genes and immune responsive genes have been studied in patients suffering from benzene exposure. However, gene expression profiles and pathways related to its hepatotoxicity are not known. This study reports the results obtained in the liver of BALB/C mice (SLC, Inc., Japan) administered 0.05 ml/100 g body weight of 2% benzene for six days. Serum, ALT, AST and ALP were determined using automated analyzer (Fuji., Japan). Histopathological observations were made to support gene expression data. c-DNA microarray analyses were performed using Affymetrix Gene-chip system. After six days of benzene exposure, twenty five genes were down regulated whereas nineteen genes were up-regulated. These gene expression changes were found to be related to pathways of biotransformation, detoxification, apoptosis, oxidative stress and cell cycle. It has been shown for the first time that genes corresponding to circadian rhythms are affected by benzene. Results suggest that gene expression profile might serve as potential biomarkers of hepatotoxicity during benzene exposure.


Introduction
Benzene is widely used as a general purpose industrial solvent. However, it is now used principally in the synthesis of other chemicals. Its toxic effects on the blood and bone marrow include leucopenia, pancytopenia, aplastic anemia and leukemia (Snyder, 2002). Several studies on benzene have focused on its metabolic pathways to determine its hematotoxicity and leukemogenicity (Henderson, 1996;Snyder and Hedli, 1996). Metabolism of benzene to reactive metabolites by hepatic enzymes mainly cytochrome P 450 2E1 (CYP2E1) is a prerequisite to its cyto and genotoxicity (Gut et al. 1996;Valentine et al. 1996). Primary benzene metabolites include phenol, hydroquinone, catechol and trans trans muconic acid (Ross, 2000). They interact synergistically and further exacerbate benzene toxicity (Chen and Eastmond, 1995). This mechanism of multimetabolite genotoxicity is another unique aspect of benzene that distinguishes it from other chemicals in terms of mechanism of its toxicity. Subsequent secondary activation of its metabolites by myeloperoxidase (MPO) present in bone marrow results in the production of genotoxic quinines and reactive oxygen species. They induce hemopoetic cellular damage and dysfunction of bone marrow stromal cells (Farris et al. 1997). Despite intensive studies over several decades, mechanisms underlying benzene induced toxicity and leukemogenicity are not yet fully understood. They are complicated by various pathways including those of metabolism (Snyder and Hedli, 1996), growth factor regulation (Niculescu, 1995), oxidative stress (Laskin et al. 1996), DNA damage (Lee and Garner, 1991), cell cycle regulation (Yoon et al. 2001b) and programmed cell death (Ross et al. 1996). To elucidate molecular mechanisms of its toxicity, cDNA microarray analyses have been performed by several workers. Yoon et al. (2003) performed cDNA microarray analysis to study the mechanisms of benzene induced hematotoxicity in bone marrow of mouse. Effects of oral exposure to benzene on gene expression in rat liver were reported by Heijne et al. (2005). Gene expression changes were attributed to pathways of biotransformation viz. glutathione synthesis, fatty acid and cholesterol metabolism. Kim et al. (2005) suggested that hydroquinone at least in part, may enhance allergic immune responses of benzene by inhibiting the production of IL-12 in macrophages. Microarray analysis of peripheral blood of benzene exposed workers was performed by Forrest et al. (2005). They showed that the altered expression of CXCL16, ZNF331, JUN and PF4 as potential biomarkers of benzene exposure. Tumor related ) and immune responsive genes  have also been recorded in patients suffering from benzene poisoning.
Since global gene expression studies are now being increasingly used in toxicology (Waters, et al. 2003), we hypothesized that microarrays could identify changes in gene expression in the liver that could be used as new biomarkers of exposure, identify early effects and provide information on mechanism of benzene toxicity. In this study we investigated the changes in gene expression induced by 24 hrs and six days of exposure to benzene in the liver of mice to probe further molecular mechanisms of benzene toxicity. The overall objective of the study is to identify potential molecular markers of early effects of benzene in liver.

Animals and treatments
Approximately 9-week old, male BALB/c mice (SLC, Inc., Japan) were kept in a 12 h light/dark cycle under controlled temperature and humidity for 2 weeks in the animal room prior to the experiment. The mice were offered standard food pellets and water ad libitum. Their body weight averaged from 24 to 27 g (mean ± S.D. -25.75 ± 0.97).
Pure benzene (99%) was diluted by corn oil so as to make a 2% (v/v) solution of benzene in corn oil. 0.05 ml of diluted benzene/100 g body weight was intraperitoneally injected to mice. 5 benzene treated mice were sacrifi ced after 24 h of benzene treatment. While 5 were sacrifi ced after 6 days of benzene treatment. Parallel controls were also sacrifi ced simultaneously.

Biochemical analysis
Blood was collected from the inferior vena cava. Serum was separated by centrifugation. Serum activities of alanine aminotransferase (ALT), aspartate aminotransferase (AST) and alkaline phosphatase (ALP) were determined using automated Clinical Chemistry Analyzer (Fuji Dri-Chem 3500s, Fujifi lm Co., Japan) at each time point. Data was statistically calculated and signifi cance difference between treated and control groups was tested using one way analysis of variance (ANOVA) followed by the Dunnett test. P values less than 0.05 were considered to be signifi cant.

Histopathology
Small pieces of liver collected from the mid liver lobe of all treated and control mice were fi xed in 10% neutral buffered formalin, dehydrated in alcohol, cleared in xylene and embedded in paraffin. Four microns thick paraffin sections thus prepared using rotary microtome (2165, Leica, Germany) were stained with hematoxylin and eosin and examined under light microscope (E400, Nikon, Japan).

Isolation of RNA
Liver samples were homogenized with 2 ml Trizol ® (Invitrogen, U.S.A.) and mixed with 0.4 ml chloroform. Following centrifugation, aqueous phase was collected and mixed with isopropanol for precipitation. RNA precipitate was washed by ethanol. Isolated total RNA was purifi ed using RNeasy mini kit (Qiagen, Germany) according to the manufacturer's protocol. The total RNA was quantifi ed using NanoDrop ® ND-1000 (NanoDrop, U.S.A.) and RNA integrity was determined by Agilent Bioanalyzer 2100 (Agient Technologies, U.S.A.).

Microarray analysis
Affymetrix GeneChip ® Mouse 430A 2.0 containing over 22,600 probe sets was used for microarray experiment. Double-stranded cDNA was synthesized from 5 ug of template total RNA using One-Cycle cDNA Synthesis kit (Affymetrix, U.S.A.), and biotin labeled cRNA was synthesized using IVT Labeling kit (Affymetrix, U.S.A.). The labeled cRNA was purified with GeneChip ® Sample Cleanup Module (Affymetrix, U.S.A.). The quality and quantity of the cRNA was checked by conducting gel electrophoresis and NanoDrop ® ND-1000 (NanoDrop, U.S.A.), respectively. Fifteen micrograms of the purifi ed cRNA of each sample was fragmented and hybridized to arrays for 16 h at 45 °C. All arrays were washed and stained automatically by using a fluidics Station 450 (Affymetrix, U.S.A.) and scanned by GeneChip ® scanner 3000 (Affymetrix, Inc, U.S.A.). All procedures were performed according to the manufacturer's protocols (Affymetrix, U.S.A.). Image processing was performed using Affymetrix GeneChip ® Operating System (GCOS).

Gene expression analysis
The preprocessing procedure of resultant cell intensity files (CEL) and following statistical analysis were performed using GenePlex software version 2.3 (ISTECH Inc., Korea). Microarray data was globally normalized and used for further analysis. Total 22,600 probe sets representing over 14,000 mouse genes were applied to gene expression analysis as well as hierarchical clustering. The differentially expressed genes were selected at a minimum 2-fold change after 1 day and 6 days of benzene treatment compared with the control group. Data were statistically analyzed by Student's t-test and P-values Ͻ0.01 were considered as statistically signifi cant. The selected genes were fi ltered by extension GCOS fi ltering (reference M1). In addition, the differentially expressed genes at 6 days treatment group compared with 1day treatment group were also selected considering the same criteria (Ն2-fold, P Ͻ 0.01). All selected genes were analyzed by two-dimensional hierarchical clustering based on Pearson correlation and Complete Linkage (Eisen, 1998). Discrepancies among control, 1 day, and 6 days group were visualized by Principal Component Analysis (PCA). The classifi cation of pathway for interesting genes was performed using KEGG pathway database. The selected genes were annotated based on NetAffx, linked at http://www.affymetrix.com.

Blood biochemistry
Aspartate aminotransferase (AST) and alanine aminotransferase (ALT) values increased significantly after 24 hours of benzene treatment. However, on treatment for 6 days, a signifi cant decline in their activity was recorded. Activity of serum alkaline phosphatase also increased after 24 hours but declined after six days of benzene treatment (Fig. 1).

Liver histopathology
Hepatocytes near the centrilobular regions of the lobule were compared in all the three groups. No significant histopathological changes were observed in the liver of mice treated with benzene for 6 days. However, hypertrophic lesions were recorded at a few places (Fig. 4). However, microbalooning in the hepatocytes was observed in the liver of mice after 24 hr of benzene treatment. Large necrotic spaces were wanting (Fig. 3). No pathological lesions were observed in the liver of control mice (Fig. 2).

Gene expression profi le
Microarray analysis was carried out to determine differences in hepatic gene expression between benzene and vehicle (corn oil) treated mice. RNA integrity was intact (data not shown). Individual liver samples from treated animals were analysed. Array based observations were made in triplicate. Gene expression was analyzed only in animals treated with benzene for 24 hours and 6 days.
A total of 136 reliable genes were fi ltered from 22600 probe sets by standard deviation of control strength after per spot and per chip normalization. Genes were sorted according to Student's "t" test (P values from 0 to Ͻ0.05). Forty four genes were selected based on the fold difference in at least three of nine conditions using ANOVA. A summary of these analyses is presented in table 1. Table 2 lists genes that changed signifi cantly (p Ͻ 0.05→0.01) in mice liver treated with benzene for six consecutive days. Results exhibiting time dependent changes in the expression of genes corresponding to control mice are explained in Figure 5. Hierarchical clustering of the genes and duration of treatment with benzene also revealed time dependent differences in gene expression (Fig. 5). Interestingly, proteins involved in biosynthesis of steroids (Hmger), cell cycle (Mcm5 and Mee), fatty and biosynthesis (Fasn), glutathione metabolism (Gstm3), glycerolipid metabolism (Lipg), glycolysis (Pklr), gluconeogenesis, MAPK signaling pathways (Hspa1a) and purine metabolism (Ecgf1) were found to be induced by benzene.

Analysis of pathway level
After t test and fold differences fi ltering, 44 genes were found to be up or down regulated by benzene treatment. Analysis of differentially regulated genes showed the involvement of calcium signaling pathway, cell adhesion molecules pathway, circadian rhythms, glutathione-S-transferases and heat shock proteins in benzene toxicity.

Discussion
It has been accepted now that omic technologies have signifi cant potential in generating novel biomarkers of exposure, susceptibility and response to different xenobiotics. Different xenobiotics produce HC Figure 2. Normal hepatocytes without any atrophy or hypertrophy were observed in the liver of control mice. ×400. Figure 3. T.S. of the liver of a mice treated with benzene for 24 hrs shows microbalooning of the hepatocytes (HC). A few binucleated cells are also observed. ×400. Figure 4. T.S. of the liver of mice treated with benzene for 6 days shows no microbalooning of the hepatocytes and no paranchymal degradation. However, mild hypertrophy was observed at certain places. Enlarged nuclei are also observed. ×400. Table 1. Gene expression profi le in the liver of mice affected by 6 days of benzene treatment.   Table 2. Comparison of gene expression profi le in the liver of mice between two time points (One day treatment and six days treatment). specifi c gene expression patterns in target organs. Minami et al. (2005) in their study on acetaminophen (APAP), bromobenzene, carbontetrachloride, dimethylnitrosamine and thioacetamide identifi ed potential biomarkers of their hepatotoxicity. However, gene expression changes induced by benzene in the liver are yet to be established. Smith et al. (2005) showed that toxicogenomics have signifi cant potential in generating biomarkers of benzene toxicity. They identifi ed more than 100 differentially expressed genes, those related to apoptosis and immune function were most signifi cantly affected in humans occupationally exposed to benzene. Yoon et al. (2003) investigated benzene induced hematotoxicity and leukemogenicity using mouse bone marrow and making cDNA microarray analysis in C57BL/6, wild type (WT), and p53knock out (KO) mice. They suggested that dysfunction of p53 gene induced fatal problems fi nally resulting in hemopoietic malignancies. Cluster 5

HC
Cluster 6 Figure 5. Hierarchical clustering map of gene expression profi les of all selected genes at the statistical criteria of Ն2-fold change and p Ͻ 0.01. Identifi cation of genes that can differentiate between 6 days and 1 day groups after benzene treatment (The gene expression pattern of control group shows similar to 1 day group after benzene treatment). Each row represents a gene on the microarray and each column represents an individual hepatic mRNA sample of mice. The red and green indicate increased and decreased expression relative to the control, respectively. The colored branch on each column indicates each sample group: control (Green), 1 day (blue) and 6 days (red) group after benzene treatment.
Benzene induced gene expression changes in rat liver after 28 days of oral exposure to benzene showed disturbances in pathways of biotransformation, glutathione synthesis, fatty acid and cholesterol metabolism (Heijne et al. 2005).  in their study on peripheral white blood cell gene expression profi le of benzene poisoned patients identifi ed 38 genes to be associated with benzene toxicity. Chen et al. (2005) made cDNA microarray analysis of leukocytes in patients of benzene poisoning and found 25 genes to be differentially expressed. Amongst these 16 genes were upregulated and 9 genes were down regulated. Some cDNA replication and some repair genes associated with benzene poisoning also showed differential expression. Time and dose dependent study on gene expression made by us shows that after six days of benzene treatment, 44 genes were affected. 25 genes were down regulated and 19 genes were upregulated. This information was further analysed through KEGG pathway database. Important pathways affected after 24 hrs of benzene treatment included β-cell receptor signaling pathways (Ifi tm1), cell cycle; signaling pathway; focal adhesion; Jak-STAT signaling pathway, circadian rthym (Cry1), cytokine-cytokine receptor interaction (Ccr5), glutathione metabolism (Gstm3), hematopoietic cell lineage (Cd24a) and TGF-beta signaling pathway (Id3 and Id1).
We for the fi rst time report that circadian rhythm pathway (Cry1) is also affected by benzene treatment. Rana et al. (2007) observed that circadian rhythms effect lipid peroxidation induced by benzene. Susceptibility of humans and rats to benzene during morning and evening was conducted by Leung and Harrison (1998) and Starek et al. (1990). Down regulation of Per3 and Arnt7 observed after six days of benzene treatment clearly shows the infl uence of circadian rhythms on benzene toxicity. Circadian rhythms are further regulated by protein synthesis and we show that these genes too were infl uenced by benzene. Present report confi rms the infl uence of circadian rhythms on benzene hepatotoxicity. Rana and Verma (2005) reviewed the biochemical toxicity of benzene. However, this new information is expected to help in understanding the molecular mechanisms of benzene toxicity.
Biochemical and histopathological observations show that six days treatment with benzene encouraged an adaptive mechanism in the liver. Gene expression studies especially those of CYP2610, CYP4A10 that were found to be down-regulated fi nd support from the biochemical and histopathological results. Upregulation of Gstm3 and Hspa1a further suggest a protective mechanism.
The pathway analysis shows the involvement of calcium signaling, transferases and heat shock proteins in benzene toxicity. Present observations offer additional information on the molecular mechanisms of benzene toxicity. The cDNA microarray system showed alterations in the number of benzene affected genes including physiologic and toxicologic gene repertoires. This communication provides valuable targets for future investigations on the benzene induced toxicity and leukernogenicity.