RNA expression in the early characterization of hepatotoxicants in Wistar rats by high-density DNA microarrays.

High-density microarrays are useful tools to study gene expression for the purpose of characterizing functional tissue changes in response to the action of drugs and chemicals. To test whether high-density expression data can identify mechanisms of toxicity and to identify an unknown sample through its RNA expression pattern, groups of male Wistar rats were administered 6 hepatotoxicants. The compounds chosen for this study were microcystin-LR (MLR), phenobarbital (PB), lipopolysaccharide (LPS), carbon tetrachloride (CT), thioacetamide (THA), and cyproterone acetate (CPA). These hepatotoxicants are known to induce adverse liver effects through different mechanisms. Liver mRNA was isolated and used to generate biotinylated cRNA for hybridization to a custom 1,600-rat gene DNA microarray. Treatment correlation matrices analyzed hybridization data from a hepatotoxicant-blinded sample, with gene expression coefficients (GEC) evaluated by means of hierarchical cluster analysis and visual representation as dendrograms. The experimental liver toxicity from the different treatments was confirmed by means of concurrent histopathology, liver enzymes, and bilirubin assays. This toxico genomic analysis identified multiple genes and groups of genes that were affected by the hepatotoxicants on study, indicating that high-density microarray expression data are useful to identify groups of genes involved in toxicity. In addition, the mRNA expression profile of an unidentified sample can be accurately identified when compared with the expression profiles resident in the data set. This study supports the use of gene expression-profiling technology to determine or to predict toxic liver effects.

High-throughput toxicology: practical considerations.

The current pace of drug development has necessitated that the pharmaceutical industry consider alternative approaches for rapid toxicity screening of novel compounds. Genomics technology surfaced the concept of comprehensive toxicity evaluation from a single sample. Metabonomics is a new technology employing (1)H-NMR to rapidly evaluate peripheral samples for indices of toxicity enabling higher-throughput evaluation of in vivo toxicity. Although both approaches show significant promise, much remains to be done before either can become a routine part of a toxicologist's portfolio. Distinguishing efficacy and adaptive responses from toxicity is a problem common to both technologies. In addition, the novelty of the approaches has precluded standardization to this point. Clearly the traditional role of a toxicologist is changing, new demands for speed call for new ideas. However, novelty does not supplant quality, and the answers we provide to these very basic questions will determine how effectively these new technologies will become an accepted part of the drug discovery process.

In vitro photogenotoxic activity of clinafloxacin: a paradigm predicting photocarcinogenicity.

Fluoroquinolone antiinfective drugs exhibit phototoxic, photogenotoxic, and photocarcinogenic activities in experimental systems which may be interrelated. Clinafloxacin (CLX), a new fluoroquinolone, is a potent antiinfective agent being developed for use in life-threatening infections. While this drug has previously been demonstrated to be phototoxic, this report evaluated the photogenotoxic and photocarcinogenic potential of CLX. When Skh-1 mice were administered CLX in the presence of ultraviolet light (UVA) at the maximum tolerated dose expected for a photocarcinogenicity bioassay, induction of DNA strand breakage was noted in keratinocytes isolated from these animals. When compared with other well-studied fluoroquinolones in vitro, CLX and Lomefloxacin (LMX) were equally effective in producing chromosome damage and DNA strand breakage in Chinese hamster ovary (CHO) cells exposed to UVA. Treatment of CHO cells with CLX in the presence of UVA also resulted in hydroxyl radical formation. However, coincubation of CHO cells with CLX and various antioxidants markedly reduced hydroxyl radical formation, but inhibited photogenotoxicity only to a limited extent. Thus, while reactive oxygen species contribute to the photogenotoxic activity of CLX, other factors may be involved. Since CLX exhibits both phototoxic and photogenotoxic activity, we predict that CLX would be photocarcinogenic in vivo. The present study suggests that under conditions of human exposure, the potential risk for CLX-induced photocarcinogenicity is small.

The mechanism of thioacetamide-induced apoptosis in the L37 albumin-SV40 T-antigen transgenic rat hepatocyte-derived cell line occurs without DNA fragmentation.

The hepatotoxicant thioacetamide (TH) has classically been used as a model to study hepatic necrosis; however, recent studies have shown that TH can also induce apoptosis. In this report we demonstrate that 2.68+/-0.54% of the albumin-SV40 T-antigen transgenic rat hepatocytes undergo TH-induced apoptosis, a level comparable to other in vivo models of liver apoptosis. In addition, TH could induce apoptosis and necrosis in the L37 albumin-SV40 T-antigen transgenic rat liver-derived cell line. Examination of dying L37 cells treated with 100 mM TH by electron microscopy revealed distinct morphological characteristics that could be attributed to apoptosis. Quantitation of apoptosis by FACS analysis 24 h after treatment with 100 mM TH revealed that 81.3+/-1.6% of the cells were undergoing apoptosis. In contrast, when L37 cells were treated with 250 mM TH, cells exhibited characteristics consistent with necrotic cell death. DNA fragmentation ladders were produced by growth factor withdrawal-induced apoptosis; however, in 100 mM TH-induced apoptosis, DNA fragmentation ladders were not observed. Analysis of endonuclease activity in L37 cells revealed that the enzymes were not inactivated in the presence of 100 mM TH. The data presented in this report indicate that the L37 cell line could be used to study the mechanism of TH-induced apoptosis that was not mediated through a mechanism requiring DNA fragmentation.

Cell lines with heterogeneous phenotypes result from a single isolation of albumin-sv40 T-antigen transgenic rat hepatocytes.

Several immortalized cell lines were established from the livers of two transgenic rats expressing the simian virus protein large T antigen under the control of the albumin promoter. Hepatocytes from transgenic rats were isolated by a two-step perfusion procedure from a normal-appearing liver and from a liver that contained a single neoplasm. Cells were also isolated from the dissociated liver neoplasm. After 5 weeks in culture, cell colonies were isolated and subcloned as individual cell lines. Electron microscopy revealed that all cell lines had a high nuclear-to-cytoplasmic ratio compared with normal hepatocytes. The cytoplasm contained numerous organelles, including smooth and rough endoplasmic reticulum, Golgi, mitochondria, peroxisomes, and annulate lamellae. However, the lines exhibited a variety of different cell morphologies. All cell lines, including those derived from neoplastic cells, exhibited a similar doubling time of 26 hours and the ability to grow in soft agar. Northern blot analysis revealed that the cell lines differentially expressed hepatocyte markers. Large T antigen was expressed in the cultured cell lines at much higher levels than was observed in transgenic hepatocytes in vivo. This suggests that the viral protein is required to maintain cell viability in culture. In addition, the tumor suppressor proteins, p53 and Rbp105, were detected in all cultured cell lines. In contrast, these same proteins were not detected by Western blot in transgenic hepatocytes in vivo. All cell lines expressed the oncogene c-myc, yet growth factor-dependent and independent growth were observed. The data presented show for the first time the establishment and characterization of a number of cell lines derived from hepatocytes isolated from an alb-SV40 large T-antigen transgenic rat. These cell lines exhibited varied morphological and biochemical hepatocellular characteristics in vitro, suggesting that the expression of the transgene in hepatocytes leads to considerable phenotypic diversity analogous to that seen in hepatocellular carcinomas induced by chemical carcinogenesis.

The translational inhibitor cycloheximide represses growth factor depletion-induced apoptosis in an alb-SV40T transgenic rat liver cell line.

A transgenic rat line carrying the alb-SV40A transgene has been described by this laboratory. Several cell lines have been established from the livers of two of these rats. One of these cell lines, L37, exhibits a large nuclear/cytoplasmic ratio and a well-differentiated cytoplasm containing numerous organelles. When L37 cells are placed into culture medium lacking necessary growth factors, cellular proliferation continues for 48 hours after medium change. Subsequent to the initial 48 hours, cells begin to shrink and lose contact with adjacent cells, eventually sloughing off the culture plate surface, with most cell deaths occurring between 48 and 96 hours after medium change. Microscopic examination of sloughing cells indicates they possess highly convoluted and blebbed plasma membranes, a morphological characteristic of apoptosis. Ultrastructural studies demonstrate the ubiquitous presence of apoptotic bodies. When DNA isolated from growth factor-depleted cells is resolved on agarose gels, DNA fragmentation ladders are observed at times of maximum apoptotic change. Quantitative analysis of L37 cells between 48 and 96 hours after the removal of the culture medium shows that 59% +/- 2% of the cells undergo apoptosis. When cycloheximide, puromycin, or actinomycin D is added to the L37 cultures, only cycloheximide is able to repress apoptosis, indicating that the mechanism of apoptosis in the L37 liver-derived cell line requires a cycloheximide-sensitive translational event. The extremely high rate of apoptosis, together with the maintenance of hepatocellular characteristics, indicates the usefulness of this cell line as a model in which to study the mechanisms of hepatocellular apoptosis.

Acetaminophen-arylated proteins are detected in hepatic subcellular fractions and numerous extra-hepatic tissues in CD-1 and C57B1/6J mice.

To identify acetaminophen (APAP)-bound proteins in addition to the major 44 and 58 kDa APAP-binding proteins (Bartolone et al., 1992, Toxicol. Appl. Pharmacol. 113. 19-9; Pumford et al., 1992, Biochem. Biophys. Res. Commun. 182, 1348-1355; Bulera et al., 1995, Toxicol, Appl. Pharmacol. 134, 313-320), we investigated subcellular localization of liver proteins and tissue distribution of proteins arylated by a hepatotoxic dose of APAP in CD-1 and C57B1/6J mice. Western blot analysis with affinity-purified, anti-APAP antibodies allowed the detection of covalently bound proteins in liver mitochondria, nuclei, membrane, cytosol, and microsomes. Enzyme market assays revealed that subcellular fractions were 90-98% pure. The lack of contamination from other isolated subcellular fractions indicates that covalently bound proteins were specific to the particular subcellular fraction. APAP-arylated proteins with molecular weights similar to those detected in the liver were found in cytosolic fractions from kidney, lung, pancreas, heart, skeletal muscle, and stomach. The presence of arylated proteins in extra-hepatic organs suggests that other organs may be susceptible to APAP toxicity and may contain critical protein targets that are important in APAP toxicity. In contrast, covalently bound proteins were not detected in cytosols isolated from spleen, small intestine, brain, and testis. The characterization of the APAP-arylated proteins identified in this study will aid in elucidating the mechanism of APAP-induced toxicity.

Identification of the mouse liver 44-kDa acetaminophen-binding protein as a subunit of glutamine synthetase.

Identification of proteins that covalently bind acetaminophen (APAP) is essential for a clearer understanding of the hepatotoxicity that results after an APAP overdose. Birge et al. (1991) have reported that in our mouse model system a membrane-associated 44-kDa acetaminophen-binding protein is the earliest target detected immunochemically following administration of an hepatoxic dose of APAP. To identify this 44-kDa protein, liver microsomes from mice administered 600 mg APAP/kg were extracted with Triton X-114 and the resulting aqueous fraction was adsorbed to DEAE-cellulose. Further purification of the DEAE eluate by reverse-phase HPLC and by two-dimensional (2D) gel electrophoresis indicated that four proteins of approximately 44 kDa with pIs ranging from 6.7-7.0 were targeted by APAP. The most highly arylated of these 44-kDa isoforms (pI 7.0) was excised from 2D gels, digested with trypsin, and the resulting peptides were separated by reverse-phase HPLC. The two best resolved peptides were sequenced and 14 of the 15 amino acids detected were found to be identical to subunits of mouse liver glutamine synthetase (EC 6.3.1.2). Purification of glutamine synthetase from APAP-treated mice confirmed that the enzyme is indeed targeted by APAP in vivo. Since the intracellular activity of glutamine synthetase is significantly decreased after the administration of APAP to hepatocytes in culture, it is likely that the arylation of this enzyme may be involved in the ensuing hepatotoxicity.

Purification, antibody production, and partial amino acid sequence of the 58-kDa acetaminophen-binding liver proteins.

Immunochemical analysis of electrophoretically resolved liver proteins from mice administered hepatotoxic doses of acetaminophen has identified two proteins of 44 and 58 kDa as major targets for acetaminophen arylation. In the present study the 58-kDa acetaminophen-binding protein (58-ABP) was purified from mouse liver cytosol by gel permeation chromatography, preparative isoelectric focusing, and polyacrylamide gel electrophoresis. The acetaminophen adducts were visualized on immunoblots using affinity-purified anti-acetaminophen antibodies after each step of the purification. Gel permeation chromatography, under nondenaturing conditions, indicated that the protein is a monomer. Two-dimensional gel electrophoresis demonstrated that the 58-ABP consists of a cluster of four immunochemically reactive isoforms with isoelectric points ranging from 6.2 to 6.6. V-8 protease digestion of the isoforms suggested that they contained similar peptide fragments. The purified 58-ABP was utilized to produce polyclonal antibodies and to determine the amino acid composition and partial sequence of the protein. These antibodies revealed a protein cluster of similar molecular weight and isoelectric points in the cytosol of a human liver specimen. Amino acid analysis of the purified protein indicated that it contains eight cysteine residues (about 1.4% by weight). This low cysteine content raises the possibility that at hepatotoxic doses acetaminophen may also bind to non-thiol sites on the protein. The amino acid sequence of two cyanogen bromide/tryptic peptide fragments revealed that the major immunochemically detectable acetaminophen target in the cytosol is homologous to a selenium-binding protein which has been recently sequenced.

The arylation of microsomal membrane proteins by acetaminophen is associated with the release of a 44 kDa acetaminophen-binding mouse liver protein complex into the cytosol.

When analyzed by Western blotting with affinity purified antibodies against acetaminophen, proteins of molecular weight 44 and 58 kDa appear to be the major macromolecular targets in livers of mice administered hepatotoxic concentrations of acetaminophen. In this study, we have examined the characteristics and biochemical properties of the 44 kDa acetaminophen-binding protein in mouse liver. Data are presented which indicate that the 44-kDa protein is the earliest detectable protein targeted by acetaminophen; 30 min after acetaminophen administration in vivo, the binding to the 44 kDa protein is primarily localized in the microsomal fraction. After 1 hr, the 44 kDa acetaminophen-binding protein can be detected in both the microsomes and the cytosol. Extractions of microsomes with Triton X-114 or 1 M NaCl suggests that the acetaminophen-bound 44-kDa protein behaves as a peripheral membrane protein associated with the endoplasmic reticulum by ionic interactions. The cytosolic and microsomal 44-kDa proteins possess similar biochemical properties; both exist natively as components of a protein complex of greater than 200 kDa and both consist of two major isovariants with isoelectric points of 7.0 and 7.1 on two-dimensional gels. When N-acetyl-p-benzoquinone imine, the reactive metabolite of acetaminophen, is incubated with cytosolic or microsomal fractions from control liver, targeting of a 44-kDa protein is only observed in the microsomes. However, when acetaminophen is activated in an NADPH-regenerating microsomal system in vitro, some of the microsomal 44-kDa protein complex can be solubilized and released into the cytosol. Thus, acetaminophen administration can alter the subcellular distribution of at least one protein target in the cell.