Gene expression profiling in rat liver treated with various hepatotoxic-compounds inducing coagulopathy.

A large-scale transcriptome database of rat liver (TG-GATEs) has been established by the Toxicogenomics Project in Japan. In the present study, we focused on 8 hepatotoxic compounds within TG-GATEs, i.e., clofibrate, omeprazole, ethionine, thioacetamide, benzbromarone, propylthiouracil, Wy-14,643 and amiodarone, which induced coagulation abnormalities. Aspirin was selected as a reference compound that directly causes coagulation abnormality, but not through liver toxicity. In blood chemical examinations, for all the coagulopathic compounds there was little elevation of aspartate aminotransferase (AST) and/or alanine aminotransferase (ALT), suggesting no severe cell death by treatment with the compounds. We extracted 344 probe sets from the data for these 8 typical drugs, which induced this phenotype at any time from 3 to 28 days of repeated administration. Principal component analysis using these probe sets clearly separated dose- and time-dependent clusters of the treated groups from their controls, except aspirin and propylthiouracil, both of which were considered to cause coagulopathy not due to their hepatotoxicity but due to their direct effects on the blood coagulation system. Reviewing the extracted genes, changes in lipid metabolism were found to be dominant. Genes related to blood coagulation were generally down-regulated by these drugs except that vitamin K epoxide reductase complex subunit 1 (Vkorc1) like 1, a paralogous gene of Vkorc1, was up-regulated. As expected, expression changes of these genes were least prominent in aspirin or propylthiouracil-treated liver. We concluded that these probe sets could be a good starting point in developing mechanism-based biomarkers for diagnosis or prognosis of hepatotoxicity-related coagulation abnormalities in the early stage of drug development.

A toxicogenomics approach for early assessment of potential non-genotoxic hepatocarcinogenicity of chemicals in rats.

For assessing carcinogenicity in animals, it is difficult and costly, an alternative strategy has been desired. We explored the possibility of applying a toxicogenomics approach by using comprehensive gene expression data in rat liver treated with various compounds. As prototypic non-genotoxic hepatocarcinogens, thioacetamide (TAA) and methapyrilene (MP) were selected and 349 commonly changed genes were extracted by statistical analysis. Taking both compounds as positive with six compounds, acetaminophen, aspirin, phenylbutazone, rifampicin, alpha-naphthylisothiocyanate, and amiodarone as negative, prediction analysis of microarray (PAM) was performed. By training and 10-fold cross validation, a classifier containing 112 probe sets that gave an overall success rate of 95% was obtained. The validity of the present discriminator was checked for 30 chemicals. The PAM score showed characteristic time-dependent increases by treatment with several non-genotoxic hepatocarcinogens, including TAA, MP, coumarin, ethionine and WY-14643, while almost all of the non-carcinogenic samples were correctly predicted. Measurement of hepatic glutathione content suggested that MP and TAA cause glutathione depletion followed by a protective increase, but the protective response is exhausted during repeated administration. Therefore, the presently obtained PAM classifier could predict potential non-genotoxic hepatocarcinogenesis within 24 h after single dose and the inevitable pseudo-positives could be eliminated by checking data of repeated administrations up to 28 days. Tests for carcinogenicity using rats takes at least 2 years, while the present work suggests the possibility of lowering the time to 28 days with high precision, at least for a category of non-genotoxic hepatocarcinogens causing oxidative stress.

Gene expression profiling in rat liver treated with compounds inducing phospholipidosis.

We have constructed a large-scale transcriptome database of rat liver treated with various drugs. In an effort to identify a biomarker for diagnosis of hepatic phospholipidosis, we extracted 78 probe sets of rat hepatic genes from data of 5 drugs, amiodarone, amitriptyline, clomipramine, imipramine, and ketoconazole, which actually induced this phenotype. Principal component analysis (PCA) using these probes clearly separated dose- and time-dependent clusters of treated groups from their controls. Moreover, 6 drugs (chloramphenicol, chlorpromazine, gentamicin, perhexiline, promethazine, and tamoxifen), which were reported to cause phospholipidosis but judged as negative by histopathological examination, were designated as positive by PCA using these probe sets. Eight drugs (carbon tetrachloride, coumarin, tetracycline, metformin, hydroxyzine, diltiazem, 2-bromoethylamine, and ethionamide), which showed phospholipidosis-like vacuolar formation in the histopathology, could be distinguished from the typical drugs causing phospholipidosis. Moreover, the possible induction of phospholipidosis was predictable by the expression of these genes 24 h after single administration in some of the drugs. We conclude that these identified 78 probe sets could be useful for diagnosis of phospholipidosis, and that toxicogenomics would be a promising approach for prediction of this type of toxicity.

Gene expression profiling of methapyrilene-induced hepatotoxicity in rat.

The present study was conducted as a model case of the toxicogenomics approach for analyzing toxicological mechanisms and toxicity assessments in the early stage of drug development by comparing with classical toxicology data. Methapyrilene (MP) 100 mg/kg produced obvious histopathological changes in liver of rats by single or repeated dose up to 28 days with significant elevation of ALT and AST. In the middle dose groups (30 mg/kg MP), no apparent changes were noted in blood biochemical data by single dosing or repeated dosing up to one week, and no obvious histopathological changes were observed except a slight hypertrophy in the hepatocytes. Comprehensive gene expression changes were analyzed using Affymetrix GeneChip and differentially expressed probe sets were statistically extracted. These contained many genes related to "glutathione metabolism", "apoptosis", "MAPK signaling pathway" and "regulation of cell cycle", which were all thought to be involved in the development of presently observed phenotypes. In the high dose groups, TGP1 scores (developed in our system in order to overview the responsiveness of drugs to multiple marker gene lists) for these categories were markedly increased from the early time point after single dose and kept their high expression throughout the repeated dose period. In the middle dose groups, the increment of the scores were noted not only at the time points when apparent pathological changes emerged, but also at the earlier stage of repeated dosing and even after single dosing. We conclude that toxicogenomics would enable a more sensitive assessment at the earlier time point than classical toxicology evaluation.

Gene expression profiling of rat liver treated with serum triglyceride-decreasing compounds.

We have constructed a large-scale transcriptome database of rat liver treated with various drugs. In an effort to identify a biomarker for interpretation of plasma triglyceride (TG) decrease, we extracted 218 probe sets of rat hepatic genes from data of 15 drugs that decreased the plasma TG level but differentially affected food consumption. Pathway and gene ontology analysis revealed that the genes belong to amino acid metabolism, lipid metabolism and xenobiotics metabolism. Principal component analysis (PCA) showed that 12 out of 15 compounds were separated in the direction of PC1, and these 12 were separated in the direction of PC2, according to their hepatic gene expression profiles. It was found that genes with either large or small eigenvector values in principal component PC 2 were those reported to be regulated by peroxisome proliferator-activated receptor (PPAR)alpha or constitutive androstane receptor (CAR), respectively. In fact, WY-14,643, clofibrate, gemfibrozil and benzbromarone, reported to be PPARalpha activators, distributed to the former, whereas propylthiouracil, omeprazole, phenobarbital, thioacetamide, methapyrilene, sulfasalazine and coumarin did to the latter. We conclude that these identified 218 probe sets could be a useful source of biomarkers for classification of plasma TG decrease, based on the mechanisms involving PPARalpha and CAR.

Identification of glutathione depletion-responsive genes using phorone-treated rat liver.

To identify candidate biomarker gene sets to evaluate the potential risk of chemical-induced glutathione depletion in livers, we conducted microarray analysis on rat livers administered with phorone (40, 120 and 400 mg/kg), a prototypical glutathione depletor. Hepatic glutathione content was measured and glutathione depletion-responsive gene probe sets (GSH probe sets) were identified using Affymetrix Rat Genome 230 2.0 GeneChip by the following procedure. First, probe sets, whose signal values were inversely correlated with hepatic glutathione content throughout the experimental period, were statistically identified. Next, probe sets, whose average signal values were greater than 1.5-fold compared to those of controls 3 hr after phorone treatment, were selected. Finally, probe sets without unique Entrez Gene ID were removed, ending up with 161 probe sets in total. The usefulness of the identified GSH probe sets was verified by a toxicogenomics database. It was shown that signal profiles of the GSH probe sets in rats treated with bromobenzene were strongly altered compared with other chemicals. Focusing on bromobenzene, time-course profiles of hepatic glutathione content and gene expression revealed that the change in gene expression profile was marked after the bromobenzene treatment, whereas hepatic glutathione content had recovered after initial acute depletion, suggesting that the gene expression profile did not reflect the hepatic glutathione content itself, but rather reflects a perturbation of glutathione homeostasis. The identified GSH probe sets would be useful for detecting glutathione-depleting risk of chemicals from microarray data.

Utilization of a one-dimensional score for surveying chemical-induced changes in expression levels of multiple biomarker gene sets using a large-scale toxicogenomics database.

A large-scale toxicogenomcis database has now been constructed in the Toxicogenomics Project in Japan (TGP). To facilitate the analytical procedures for such large-scale microarray data, we developed a simple one-dimensional score, named TGP1 which expresses the trend of the changes in expression of biomarker genes as a whole. To evaluate the usefulness of the TGP1 score, microarray data of rat liver and rat hepatocytes deposited in the TGP database were scored for three biomarker gene sets, i.e., carcinogenesis-related, PPARalpha-regulated and glutathione depletion-related gene sets. The TGP1 scoring system gave reasonable results, i.e., the scores for carcinogenesis-related genes were high in omeprazole-, chlorpromazine-, hexachlorobenzene-, sulfasalazine- and Wy-14,643-treated rat livers, that for PPARalpha-regulated genes were high in clofibrate-, Wy-14,643-, gemfibrozil-, benzbromarone- and aspirin-treated rat livers as well as rat hepatocytes, and for glutathione deficiency-related genes were high in omeprazole-, bromobenzene-, acetaminophen- and coumarin-treated rat liver. We concluded that the TGP1 score is useful for surveying the expression changes in multiple biomarker gene sets for a large-scale toxicogenomics database, which would reduce the time of doing conventional multivariate statistical analysis. In addition, the TGP1 score can be applied to screening of compatible biomarker gene sets between rat liver and rat hepatocytes, like PPARalpha-regulated gene sets, which will allow us to develop an appropriate in vitro system for drug safety assessment in vivo.

Comparison of gene expression profiles among papilla, medulla and cortex in rat kidney.

The aim of this study was to compare gene expression profiles in the different kidney regions as the basis for toxicogenomics. Rat kidney was separated into papilla, medulla and cortex, and total RNA was isolated from these and from the whole slice. Gene expression profiling was performed using Affymetrix Rat Genome 230 2.0 Array. When global normalization was applied, the expression of beta-actin or GAPDH varied among the regions. It was considered that such a comparison could not be made, especially between papilla and other portions, since the production of total mRNA in the former was relatively low. In fact, ANOVA was performed on the gene expression values with global normalization in papilla, medulla, cortex, and whole slice, and the numbers of genes appeared to be the highest in papilla. It was also observed that many genes showed their maximum or minimum in the whole slice, which was theoretically impossible. To overcome the problems associated with global normalization, the "percellome" normalization (a way to obtain the values directly related to the copies of mRNA per cell) was employed to compare the regions. In applying this procedure, probe sets with regional difference in expression were efficiently extracted by ANOVA. When they were sorted by the fold difference to other regions, the higher rank was occupied by genes characteristic of the functions of kidney, i.e., channels, transporters and metabolic enzymes. Some of them were consistent with the literature and were related to pathophysiological phenomena. Comprehensive comparison of data of gene expression in the renal anatomical area will greatly enhance studies of the physiological function and mechanism of toxicity in kidney.

Gene expression profile in liver of differing ages of rats after single oral administration of acetaminophen.

In order to verify the influence of the rat age on hepatotoxicity, male Sprague-Dawley rats of 6 (young) and 12 (adult) weeks of age were orally administered acetaminophen (APAP), isoniazid (INH), or carbon tetrachloride (CCl4). Liver samples were obtained in a time-course manner, and changes in gene expression examined by an Affymetrix GeneChip. APAP caused more prominent hepatic injury with respect to pathology and blood biochemistry in adults than in young rats, whereas no obvious age-related differences were observed in INH- or CCl4-treated rats. Comparing gene expression in control rats, CYP3A13 was higher and GSTY2c was lower in adults, suggesting that production of the active metabolite of APAP is higher and its detoxification is lower in adults. The total amount of glutathione and total SH in rat liver was found to be higher in adult rats whereas the extent of its reduction by APAP was larger in adults. A detailed analysis of genes showing age-related differences revealed that some of them were different not in their extent but in their time course, i.e., the stress responses occurred earlier in the young than in the adult, resulting in a difference at 24 hr after dosing. These results suggest that the age-related difference in toxicity would be attributed to a higher expression of CYP3A13, producing the active metabolite of APAP as well as the lower expression of the detoxification enzyme, GSTY2c, in adult rats. Furthermore, these differences affect the time course of APAP toxicity. The present study clearly depicts the advantage of the multi-time, multi-dose protocol employed in our project for analyzing the mechanism of toxicity by gene expression profiling.

Profiling of gene expression in rat liver and rat primary cultured hepatocytes treated with peroxisome proliferators.

The Toxicogenomics project has been constructing a large-scale database of about 150 compounds exposed to rat (single dose, 3, 6, 9, 24 hrs and repeated dose for 3, 7, 14 28 days with 3 dose levels) and rat hepatocytes (2, 8, 24 hr with 3 concentrations) and data of transcriptome in liver using GeneChip, and the related toxicological measures are being accumulated. In the present study, the data of three ligands of peroxisome proliferator activated receptor alpha (PPARalpha), i.e., clofibrate, WY-14643 and gemfibrozil in our database were analyzed. Many of the beta-oxidation-related genes were commonly induced in vivo and in vitro, whereas expression changes in genes related to cell proliferation, apoptosis, were detected in vivo (single and repeated dose) but not in vitro. Changes in those related to the immune response, coagulation and the stress response were also detectable exclusively in vivo. Using the genes mobilized in two or three PPARalpha agonists, hierarchical clustering was performed on 32 compounds stored in our database. In the profiling of an in vivo single dose, benzbromarone and aspirin were located in the same cluster of the three PPARalpha agonists. The clustering of in vitro data revealed that benzbromarone, three NSAIDs (aspirin, indomethacin and diclofenac sodium) and valproic acid belonged to the same cluster of PPARalpha agonists, supporting the reports that benzbromarone,valproic acid and some NSAIDs were reported to be PPARalpha agonists. Using the genes commonly up-regulated both in vivo and in vitro, principal component analysis was performed in 32 compounds, and principal component 1 was found to be the convenient parameter to extract PPARalpha agonist-like compounds from the database.

Evaluation of methods for duration of preservation of RNA quality in rat liver used for transcriptome analysis.

In The Toxicogenomics Project, about 150 chemicals are administered to rats, and gene expression in the liver analyzed by Affymetrix GeneChip and stored in the database. As the quality of RNA greatly influences the accuracy of gene expression data, conditions of the storage of the sample are very important. Recently, an RNA stabilization solution, RNAlater, has become commercially available. In this study, the new storage method was compared with the traditional storage method (stored in freezer or liquid nitrogen) under various conditions by looking at the degradation of RNA assessed by its total yield, OD260/280 ratio, 28S/18S ratio, and quantity of beta-actin. It was confirmed that RNAlater preserved the liver tissue sample by maintaining the quality of RNA for one year (in liquid N(2) or -80 degrees C), for 3 days (4 degrees C), or for 2 hr (room temperature) without degradation of RNA. Quality of RNA samples dissolved in buffer RLT and stored at -20 degrees C tended to decrease, but samples stored at -80 degrees C were almost equivalent to those stored in liquid nitrogen. In conclusion, we recommend the following procedure for preservation of liver tissue for extraction of RNA: 1) tissues removed should be put into chilled RNAlater as soon as possible; 2) samples in RNAlater must be stored overnight or longer at 4 degrees C and can be left for as long as 2 weeks without freezing; 3) samples in RNAlater can be stored for at least one year under less than -20 degrees C and 4) samples dissolved in buffer RLT can be preserved at least for one year under -80 degrees C.

Effect of the difference in vehicles on gene expression in the rat liver--analysis of the control data in the Toxicogenomics Project Database.

The Toxicogenomics Project is a 5-year collaborative project by the Japanese government and pharmaceutical companies in 2002. Its aim is to construct a large-scale toxicology database of 150 compounds orally administered to rats. The test consists of a single administration test (3, 6, 9 and 24 h) and a repeated administration test (3, 7, 14 and 28 days), and the conventional toxicology data together with the gene expression data in liver as analyzed by using Affymetrix GeneChip are being accumulated. In the project, either methylcellulose or corn oil is employed as vehicle. We examined whether the vehicle itself affects the analysis of gene expression and found that corn oil alone affected the food consumption and biochemical parameters mainly related to lipid metabolism, and this accompanied typical changes in the gene expression. Most of the genes modulated by corn oil were related to cholesterol or fatty acid metabolism (e.g., CYP7A1, CYP8B1, 3-hydroxy-3-methylglutaryl-Coenzyme A reductase, squalene epoxidase, angiopoietin-like protein 4, fatty acid synthase, fatty acid binding proteins), suggesting that the response was physiologic to the oil intake. Many of the lipid-related genes showed circadian rhythm within a day, but the expression pattern of general clock genes (e.g., period 2, arylhydrocarbon nuclear receptor translocator-like, D site albumin promoter binding protein) were unaffected by corn oil, suggesting that the effects are specific for lipid metabolism. These results would be useful for usage of the database especially when drugs with different vehicle control are compared.

Stereoselective Syntheses of Cis- and Trans-Isomers of alpha-Hydroxy-alpha,beta-dibenzyl-gamma-butyrolactone Lignans: New Syntheses of (+/-)-Trachelogenin and (+/-)-Guayadequiol.

Cis- and trans-isomers of alpha-hydroxy-alpha,beta-dibenzyl-gamma-butyrolactone lignans 1a,d-g and 2a,c,d were stereoselectively synthesized in good yields based on the electrophilic addition to the metal enolate of alpha-benzyl-gamma-butyrolactone derivatives 1l-o and 3 as a key step. This method was applied to the syntheses of (+/-)-trachelogenin and (+/-)-guayadequiol, representative examples of the trans- and cis-isomers of alpha-hydroxy-alpha,beta-dibenzyl-gamma-butyrolactone lignan series.