Hepatic Gene Expression Changes in Mice Associated with Prolonged Sublethal Microcystin Exposure

Abstract

Microcystin-LR (MCLR) is an acute hepatotoxicant and suspected carcinogen. Previous chronic studies have individually described hepatic morphologic changes, or alterations in the cytoskeleton, cell signaling or redox pathways. The objective of this study was to characterize chronic effects of MCLR in wild-type mice utilizing gene array analysis, morphology, and plasma chemistries. MCLR was given daily for up to 28 days. RNA from the 28-day study was hybridized onto mouse genechip arrays. RNA from 4 hours, 24 hours, 4 days, 1 day, and 28 days for selected genes was processed for quantitative-PCR. Increases in plasma hepatic enzyme activities and decreases in total protein, albumin and glucose concentrations were identified in MCLR-treated groups at 14 and 28 days. Histologically, marked hepatokaryomegaly was identified in the 14-day MCLR group with the addition of giant cells at 28 days. Major gene transcript changes were identified in the actin organization, cell cycle, apoptotic, cellular redox, cell signaling, albumin metabolism, and glucose homeostasis pathways, and the organic anion transport polypeptide system. Using toxicogenomics, we have identified key molecular pathways involved in chronic sublethal MCLR exposure in wild-type mice, genes participating in those critical pathways and related them to cellular and morphologic alterations seen in this and other studies.

Introduction

Microcystin-LR (MCLR) is cyclic heptapeptide produced by the cyanobacteria (blue-green algae), Microcystis aeruginosa (Dawson, 1998). Under appropriate conditions of temperature, nutrients and light, massive growth of algal blooms can occur (Bischoff, 2001). These blooms have been associated with acute, often lethal toxicity in various species of domestic animals, wildlife and humans (Azevedo et al., 2002; Frazier et al., 1998; Puschner et al., 1998) with death due to severe hepatic necrosis (Frazier et al., 1998; Puschner et al., 1998). Experimentally induced toxicosis in swine and sheep has identified massive centrilobular hepatic necrosis and hemorrhage as the major lesions (Jackson et al., 1984; Beasley et al., 2000). In mice and rats, acute MCLR-induced liver toxicity is characterized by severe centrilobular to midzonal hepatocyte apoptosis/necrosis (Hooser et al., 1989). As observed clinically, the liver is the primary target organ of MCLR due to a hepatocyte-specific organic anion transporting polypeptide (OATP) membrane transport system that carries MCLR into hepatocytes (Fischer et al., 2005). Human illnesses which have been associated with microcystin exposure include: gastroenteritis, allergic reactions and liver disease (Bell and Codd, 1994). In one case, microcystin contamination of the water supply used for kidney dialysis (Azevedo et al., 2002) resulted in 100 out of 131 patients developing acute liver failure with 52 deaths. Microcystin also has been associated with tumor promoting activity (Falconer and Buckley, 1989; Nishiwaki-Matsushima et al., 1992; Ito et al., 1997; Hu et al., 2002; Zhou et al., 2002). Epidemiologic studies in China have suggested that microcystins in contaminated water may play a role in the higher incidences of primary human hepatocellular carcinomas (Ueno et al., 1996; Yu et al., 2001). In animals, intraperitoneal injections of microcystin have induced tumor promotion in the livers of rats initiated with diethylnitrosamine (Nishiwaki-Matsushima et al., 1992; Hu et al., 2002). In an additional study, mice receiving sublethal doses of microcystin (20 μg/kg) for 28 weeks developed neoplastic liver nodules (Ito et al., 1997).

A number of studies have also characterized the pathologic effects associated with chronic microcystin exposure. These studies have described tumor promotion, hepatocellular degeneration, single cell necrosis, fibrosis, elevated liver enzymes, and neutrophil infiltration (Elleman et al., 1978; Falconer et al., 1988; Nishiwaki-Matsushima et al., 1992; Carbis et al., 1994). Alterations associated with chronic sublethal exposure to MCLR have also been characterized. Balb/c mice administered sublethal intraperitoneal (45 μg/kg/d) doses of MCLR for 4 or 7 days exhibited marked hepatocytomegaly and karyomegaly with numerous multinucleated cells (Guzman and Solter, 2002). In rats, prolonged sublethal (16, 32, and 48 μg/kg/day) exposure to MCLR for 28 days was found to have multiple dose-dependent hepatotoxic effects including increased hepatic apoptosis, steatosis and centrilobular fibrosis (Solter et al., 1998). In addition, oxidative stress has been shown to play a role in microcystin-induced cytotoxicity through induction of increased lipid peroxidation in the livers of mice and rats treated with microcystin (Guzman and Solter, 1999; Gehringer et al., 2003, 2004; Moreno et al., 2005).

Although the cellular mechanisms of microcystin-induced hepatotoxicity have not been fully elucidated, after hepatocyte uptake the majority of the toxin has been found to be present in the cytosol of hepatocytes (Hooser et al., 1991b) where it is a potent and rapid inhibitor of the serine/threonine protein phosphatases-1 and –2A (PP). This widespread inhibition leads to hyperphosphorylation of a diverse array of cellular proteins (Honkanen et al., 1990; MacKintosh et al., 1990; Yoshizawa et al., 1990), culminating in the hepatotoxic effects associated with microcystin. These effects include cytoskeletal alterations and morphologic changes including extensive cellular apoptosis and necrosis at high doses, or tumor promotion at low, chronic doses (Eriksson et al., 1990; Yoshizawa et al., 1990; Yoshida et al., 1997; Solter et al., 1998; Hooser, 2000). The mechanisms of tumor promotion have not been fully elucidated but most likely involve protein phosphatase inhibition leading to disruption of the dynamic equilibrium of protein phosphorylation/dephosphorylation and the hyperphosphorylation of many cellular proteins (Guy et al., 1992; Fujiki and Suganuma, 1999) and activation of the mitogen- activated protein kinase (MAPK) pathway (Davis et al., 1996), which may lead to an increase in cellular proliferation (Toivola and Eriksson, 1999).

The pathogenesis and molecular pathways involved in chronic, sublethal MCLR exposure are not completely understood. The objective of this study was to relate the chronic effects of MCLR in wild-type mice utilizing gene array analysis, morphology, and plasma chemistries. Using toxicogenomics, we have identified key molecular pathways involved in chronic sublethal MCLR exposure in wild-type mice, genes participating in those critical pathways and related them to cellular and morphologic alterations seen in this and other studies.

Materials and Methods

Animals

Twenty-four B6.129-Trp53 ^+,+ N5 mice (wild-type litter mates to B6.129-Trp53 ^{tm1Brd ,+} N5 p53 heterozygous knockout mice) were obtained from Taconic, Inc. (Germantown, PA) and allowed to acclimate at the Purdue Laboratory Animal Care facility for 2 weeks or longer before use. All experiments were approved by the Purdue Animal Care and Use Committee. Animals had free access to a commercial laboratory chow and water, and were maintained on a 12-hour light:dark cycle. Three or 4 mice were assigned to the various dosing groups (24 hours, 28-day control groups and 4 hours, 24 hours, 4 days, 14 days, and 28-days treatment groups) immediately prior to dosing and given free access to food and water for the duration of the study. Microcystin-LR (MCLR) (approximately 95% pure) from Microcystis aeruginosa was purchased from Calbiochem, Inc. (San Diego, CA). The MCLR was dissolved in sterile 0.9% NaCl and diluted prior to intraperitoneal administration. Forty micrograms of MCLR per kilogram of body weight or vehicle control were given daily by intraperitoneal injection. At 4 hours, 24 hours, 4 days, 14 days, and 28 days after the first MCLR dose, 3 or 4 mice from each group were anesthetized with carbon dioxide and exsanguinated. Control mice were similarly anesthetized and exsanguinated at 24 hours and 28 days postdosing. Heparinized blood was collected for plasma chemistries from the caudal vena cava. The liver was removed, patted dry, and weighed. A section of liver was fixed in 10% neutral-buffered formalin for light microscopy. An additional piece of the liver from the left lateral liver lobe was placed in RNAlater (Ambion, Inc., Austin, TX) for microarray analysis. The remainder was frozen in liquid nitrogen and stored at −80°C. Sections of lungs, kidneys, spleen, small intestine, pancreas, adrenal gland, and brain were also fixed by immersion in 10% neutral-buffered formalin. Tissues were routinely processed, embedded in paraffin, cut at 4 to 6 μm and stained with hematoxylin and eosin (H&E) for light microscopic examination.

Plasma Chemistries

Whole blood for clinical chemistry determinations was collected from the caudal vena cava using heparinized syringes, and samples were placed on ice. The plasma was collected following centrifugation and stored at −20°C until further examination. Concentrations of total bilirubin, glucose, cholesterol, triglycerides, total protein, albumin, and the activities of alkaline phosphatase (ALP), gamma glutamyltransferase (GGT), alanine aminotransferase (ALT), aspartate aminotransferase (AST), and creatine kinase in plasma were determined using standard methods on a Hitachi 917 Automatic Analyzer (Roche Diagnostics Corporation, Indianapolis, IN).

RNA Isolation

At necropsy, liver sections of approximately 500 mg from the left lateral liver lobe were excised and stored in RNAlater (Ambion, Inc., Austin, TX) as per manufacturer’s recommendations. Total RNA was isolated from approximately 100 mg of liver tissue by homogenization using guanidium isothiocyanatephenol (RNA-Stat 60,^™ Tel-Test, Inc., Friendswood). RNA was further purified using RNeasy^™ columns (Qiagen, Valencia) according to package instructions. RNA quality was assessed spectrophotometrically using a Gene Quant II instrument (Amersham Biosciences Corp, Piscataway).

cDNA Synthesis

First-strand cDNA was synthesized from 5 μg of each RNA sample using Superscript II RT (Invitrogen Corp, Carlsbad). Second-strand cDNA was then synthesized using DNA polymerase I (Invitrogen Corp, Carlsbad). The double-stranded cDNA was purified using Genechip Sample Cleanup Module (Affymetrix Inc., Santa Clara). In vitro transcription reaction was with MEGAscript T7 RNA polymerase (Ambion, Inc., Austin), and the cDNA was linearly amplified and labeled with biotinylated ribonucleotides (Affymetrix Inc., Santa Clara). Labeled RNA was further purified using RNeasy^™ columns (Qiagen, Valencia) and fragmented. Fifteen micrograms of labeled and fragmented cRNA was then hybridized onto Mouse Genome 430A 2.0 GeneChip arrays (Affymetrix Inc., Santa Clara) for 16 hours at 45°C. Labeling, hybridization, staining and washing were performed according to manufacturer’s protocols (Affymetrix Expression Analysis Technical Manual). The arrays were scanned with an Affymetrix Gene Chip Scanner, and images were quantified using MAS 5.1 software (MicroArray Suite, Affymetrix). Data analysis and mining were performed using Affymetrix Microarray suite (MAS4.0) and data mining tool (DMT 2.0) software.

Quantitative PCR

The gene transcripts for RT-PCR analysis were selected from those biological pathways shown to be altered by chronic MCLR exposure. Representative gene transcripts from actin organization, cell cycle, apoptotic, cellular redox and cell signaling pathways were examined by RT-PCR to confirm our microarray data and to also further examine the mechanisms of chronic MCLR toxicity. Oligonucleotide primers used for PCR confirmation studies were identified using the PrimerBank website 〈http://pga.mgh.harvard.edu/primerbank/index.html〉K. Oligonucleotide sequences are shown in Table 1. Total RNA was isolated as described above, and treated with DNase I for 30 minutes at 37°C. DNase I was inactivated (DNAfree,^™ Ambion Inc., Austin) and the RNA was used to generate first-strand cDNA (High Capacity cDNA Archive Kit, Applied Biosystems, Foster City) for use in quantitative (SybrGreen^™) PCR analysis. Quantitative real-time PCR was performed using the ABS 7700 sequence detector. Each sample was assayed in duplicate, and relative quantitation was performed using the C_t method (ABS user bulletin number 2) using beta-actin as the normalization gene. The primer/probe sets were determined to have the same amplification efficiencies as the beta-actin control sets (data not shown). Primer concentrations and cycle number were optimized to ensure that reactions were analyzed in the linear phase of amplification.

Statistics

Liver percent body weight and plasma chemistries were statistically analyzed using the Students’ t-test with a statistical significance at p ≤ 0.05. For the microarray data, an ANOVA analysis was used to compare the treatment groups of interest with p ≤ 0.05.

Results

Gross Pathologic Findings

A statistically significant loss in body weight was observed throughout the time course of the study in microcystin-treated animals (Table 2). This was most prominent in the mice receiving 14 or 28 daily doses of microcystin. The livers of the mice from the 14- and 28-day MCLR treatment groups were pale and moderately enlarged, while those from all other groups appeared normal. Statistically significant increases in liver to body weight ratios were seen in the 14- and 28-day treated groups (Table 2).

Plasma Chemistry Data

There were statistically significant increases in plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities in the 14- and 28-day treated groups (Table 2). The plasma ALT activity began to increase in the 4-day treated group, but no changes were identified in ALT or AST activities in the 4-hour or 24-hour treated groups (data not shown). Statistically significant decreases in plasma total protein, plasma albumin and plasma glucose were identified in mice treated with MCLR for 14 and 28 days (Table 2). An increase in plasma total bilirubin was identified in the 28-day treated group. No changes were identified in plasma gamma-glutamyltransferase throughout the study.

Histopathology

Liver histology from the control mice revealed normal hepatic cord pattern, lobules and hepatocytes (Figure 1A). Histopathologic changes in the livers of MCLR-treated mice were not seen until after 4 days of treatment. The 4-day treated group had subtle centrilobular changes that were most consistent with hepatocellular hypertrophy (Figure 1B). These changes were more prominent in the 14-day treatment group and were much more pronounced in the 28-day treatment group. In the 14-day treated group, the hepatocellular hypertrophy was primarily centrilobular to midzonal in location. The enlarged hepatocytes had moderate karyomegaly and had a loss of the cytosolic vacuolation that was seen in livers from healthy nonfasted mice. The areas of hypertrophy were sharply demarcated from the surrounding normal periportal areas (Figure 1C). Livers from the mice treated for 28 days were characterized by marked hepatocytomegaly and karyomegaly that involved the entire hepatic lobule but appeared to be worse in the centrilobular areas (Figure 1D). Many multinucleated hepatocytes containing as many as 7 giant nuclei per cell were also identified (Figure 1E). Inflammatory foci consisting primarily of neutrophils and lymphocytes were identified throughout the hepatic parenchyma. These foci appeared to be associated with necrotic hepatocytes.

Microarray Analysis

The biological processes most affected by chronic microcystin exposure for 28 days included cell organization, cell death and the response to stimulus pathways. Selected genes associated with organic ion and bile acid transporters (primarily downregulated), albumin metabolism (down-regulated), glucose homeostasis (both upregulated and downregulated), cytoskeletal organization and actin binding (primarily upregulated), GTPase activity (upregulated), apoptosis (upregulation of both pro- and anti-apoptotic related genes), redox and phase II metabolism (upregulated), cellular signaling pathways (upregulated), and calcium binding and transport (upregulated) along with fold changes and p values are listed in Table 3. There was an approximately 3-fold downregulation in Oatp1a4 (Slco1a4) and Oatp1b2 (Slco1b2) transporters in the liver following 28 days of MCLR treatment when compared to the saline treated animals. Interestingly, an approximately 133-fold decrease in the expression of Oatp1a1 (Slco1a1) transporter in the livers of the microcystin-treated animals was identified. In addition, an approximately 10-fold decrease in expression was also identified in the sodium-dependent bile acid transporter, Ntcp (Slc10a1) in the microcystin-treated group.

Real Time-PCR

RT-PCR was used to confirm the gene expression changes of 18 selected genes in the 28 day treated group that were identified with the gene chip (Table 1). RT-PCR was also used to examine the gene expression changes of 5 selected genes involved in apoptotic, cellular redox and cell signaling pathways that occurred over the 4 hours to 28-day MCLR treatment period. An increase in gene expression as early as 4 hours after treatment was identified for jun and cystatin B, and also in the redox genes gstm3, gss, and nqo1 (Table 4). These continued to increase reaching their maximums at 14 days or 28 days.

Discussion

Microcystin exposure through the ingestion of contaminated water has become a concern worldwide (Dietrich and Hoeger, 2005) because it is speculated that chronic microcystin exposure through this route may contribute to the higher incidence of liver cancer in some areas (Yu et al., 2001). However, the exact mechanism of chronic microcystin toxicity is not completely understood. In previous studies, multiple pathways affected by microcystin have been identified and include the apoptotic pathway, cellular redox pathway, actin filament/cytoskeletal reorganization pathways, and multiple other signaling pathways (Gehringer, 2004). This study is the first to report the use of global expression analysis to identify and characterize increases and decreases in gene transcripts associated with these pathways and with physiological alterations related to MCLR exposure. In addition, we describe here the MCLR-related changes in genes related to MCLR uptake, OATP transporters.

The liver is the primary target organ of MCLR due to the presence of a hepatocyte-specific organic anion-transporting polypeptide (Oatp) membrane transport system (Robinson et al., 1989; Runnegar et al., 1991; Fischer et al., 2005). In the current study, downregulation of several OATP transporters was identified following the administration of MCLR (Table 3). The downregulation of Oatp1a1 is marked and suggests a potential role of Oatp1a1 in the uptake of MCLR. Alternatively, the difference in expression between Oatp1a1 and the other 2 Oatp transporters (Oatp1a4 and Oatp1b2) may be associated with the upregulation of these 2 Oatps in response to MCLR exposure. Additionally, in the current study, an increase in total plasma bilirubin (cholestasis) was identified in the 28-day MCLR treated group (data not shown). In cholestatic liver disease, the expression of Oatp transporters has been shown to be decreased (Jung and Kullak-Ublick, 2003). Previous studies in vivo, in isolated, perfused liver and in hepatocytes have shown that competitive inhibition of the Oatps membrane transport system with a variety of substrates protects against hepatotoxicity by MCLR (Robinson et al., 1989; Runnegar et al., 1991, 1995). Recently, it was shown that Oatp1a1, 1a4, 1b2, and 2b1 are expressed at high levels in the liver of mice (Cheng et al., 2005) and it was demonstrated that rat Oatp1b2, human OATP1B1, human OATP1A2, and human OATP1B3 mediate microcystin-LR uptake in a Xenopus laevis oocyte expression system (Fischer et al., 2005). Therefore, it is probable that the Oatps system is responsible for microcystin transport into hepatocytes. The downregulation of these transporters may be a compensatory mechanism by these animals to limit the hepatic uptake of MCLR and in addition, be associated with cholestasis. Further in vivo studies are needed to determine the dynamics of Oatps in MCLR heptotoxicity.

The clinical pathologic changes associated with acute and chronic, sublethal microcystin exposure in vivo have been documented (Hermansky et al., 1990; Solter et al., 1998). In this study, the increases in alanine aminotransferase (ALT) and aspartate aminotransferase (AST) in the 14- and 28-day treatment groups were consistent with hepatocellular injury. In addition, it is of interest to note that in the 14- and 28-day treatment groups decreases in total protein, albumin and glucose were identified. These serum chemistry changes are supportive of a decrease in functional mass (hepatic insufficiency) associated with chronic, sublethal microcystin exposure. To examine this, the gene expression changes associated with the decreases in plasma albumin and glucose were evaluated. Albumin is a negative acute phase protein. Several genes that are associated with albumin were downregulated in the 28-day microcystin-treated group including albumin (alb1), prealbumin or transthyretin (ttr), and alpha-albumin (afm). These data suggest that the downregulation of these genes may be responsible for the decrease in plasma albumin in the 28-day microcystin-treated group. The mechanisms of hypoglycemia associated with hepatic insufficiency are not completely understood. It is known that G6Pase (glucose-6-phosphatase) has an important role in glucose homeostasis and its activity is regulated predominantly at the level of gene expression. In addition, the activation of the MEK/ERK pathways can suppress G6Pase gene expression (Grempler et al., 2004). In the current study, hypoglycemia may have resulted partially from the greater than 20-fold downregulation of glucose-6-phosphatase (G6pc) in the MCLR-treated animals (Table 3). The downregulation of G6pc and the upregulation of both hexokinase 2 (hk2) and glucokinase (gck) are important in the production of glucose-6-phosphate (G6P). G6P is the first substrate involved in the pentose phosphate pathway and is used to generate NADPH, which is important in reductive biosynthesis reactions. To further support these findings, we also found that 6-phosphogluconate dehydrogenase (pgd), an important enzyme involved in the pentose phosphate pathway, was upregulated. In addition, several NADPH-dependent enzymes that are likely involved in maintaining the redox state of the liver following MCLR exposure were upregulated (Table 3). Together, these changes would indicate that the decrease in plasma glucose is associated with an increase in the expression of genes in the pentose phosphate pathway that are necessary for the generation of NADPH, which is required for cell viability.

Few studies have examined the microscopic pathologic changes associated with chronic, sublethal microcystin exposure. In one of those, Balb/c mice receiving 4 or 7 daily doses of microcystin-LR had marked hepatocytomegaly and karyomegaly in the centrilobular regions of the liver (Guzman and Solter, 2002). In the present study, similar changes were identified in the livers of microcystin-treated mice. In the 4- and 14-day treatment groups, the pathologic changes were primarily centrilobular in location, however by 28 days, the lesions involved the entire hepatic lobule. Hepatic hypertrophy identified in the 14- and 28-day treatment groups was associated with an increase in liver weight (Table 2) possibly due to the increase in cell size. In the present study hepatic karyomegaly and multinucleation were identified in animals treated for 14 or 28 days. Hepatic cytomegaly, karyomegaly, and multinucleation may all be the result of mitotic arrest associated with protein phosphatase inhibition. A similar inhibitor of protein phosphatases 1 and 2A, okadaic acid, has been shown to induce cell cycle arrest in vitro (You and Bird, 1995). Although the precise mechanism of okadaic acid-induced cell cycle arrest is largely unknown, it has been shown that the treatment of transformed HeLa cells with okadaic acid resulted in an increase in the levels of genes involved in cell cycle regulation. Those genes included cdk2, cyclin A, cyclin B and Rb (You and Bird, 1995). In our present study, a statistically significant increase in the cyclin-dependent protein kinase inhibitor 1A (p21, Waf1, Cip1) expression was identified in the MCLR-treated group. Statistically significant increases in cyclin D1 and cyclin G were also identified in the MCLR-treated group (Table 3).

In addition to alterations in genes associated with the cell cycle, mitotic arrest may be related to hepatocyte cytoskeletal damage associated with microcystin exposure. In this current chronic study, it was found that sublethal exposure to microcystin for up to 28 days results in gene expression changes associated with cytoskeletal components. Cytokeratin 8 (krt2-8) and cytokeratin 18 (krt1-18) were upregulated in the liver following 28 daily injections with microcystin. This upregulation could be a response to the reorganization and loss of function of cytokeratin associated with microcystin exposure. In addition, several genes involved in actin regulation were upregulated (Table 3). The majority of these genes play a role in actin binding and/or actin polymerization. Alterations in the hepatocyte cytoskeleton attributed to disorganization of cytokeratin intermediate filaments (Falconer and Yeung, 1992; Toivola et al., 1997) and actin microfilaments have been demonstrated with acute MCLR toxicity (Eriksson et al., 1989; Ghosh et al., 1995; Hooser et al., 1991a; Wickstrom et al., 1995; Batista et al., 2003). Inhibition of protein phosphatases 1 and 2A results in hyperphosphorylation of the intermediate filaments, cytokeratins 8 and 18, contributing to these cytoskeletal alterations (Falconer and Yeung, 1992; Ohta et al., 1992; Toivola et al., 1997). Acute microcystin exposure in vitro and in vivo has resulted in the collapse and aggregation of both cytokeratin intermediate filaments and actin microfilaments around the nucleus (Hooser et al., 1991a; Falconer and Yeung, 1992; Ghosh et al., 1995; Wickstrom et al., 1995; Khan et al., 1996). Changes in actin polymerization and actin filament organization have previously been associated with the activation of Rho GTPases (Jaffe and Hall, 2005). The studies reported here also identified the upregulation of numerous genes with GTPase activities. These genes included hepatic Rho GTPases; CDC42, Rhob, Rhoc, and Rhoq. In addition, it is also known that Rho and CDC42 can activate the signal transduction pathways, JNK and p38 MAPK (Jaffe and Hall, 2005). Interestingly, the expression of Rhoc is significantly increased in some forms of hepatocellular carcinoma as it is in this study (Wang et al., 2004). This is the first study that identifies the upregulation of Rho GTPases following chronic, sublethal microcystin exposure.

Microcystin can induce apoptosis both in vitro (Hooser et al., 1991a; Khan et al., 1995; McDermott et al., 1998; Fladmark et al., 1999; Ding et al., 2000a) and in vivo (Solter et al.,1998; Yoshida et al., 1998; Hooser, 2000; Gupta et al., 2003). Although the mechanisms of microcystin-induced apoptosis are largely unknown, cysteine proteases such as calpains and caspases, may play a role as they do in other models of apoptosis (Fladmark et al., 1999; Ding et al., 2002). However, the exact role of caspases in microcystin-induced apoptosis is unresolved. In this study, we identified upregulation of both proapoptotic and antiapoptotic genes. No major upregulation in caspase gene expression was identified in the microcystin-treated animals. However, there was a small increase in the expression of the calpain subunits, capn2 and capns1 indicating a potential role of calpains in chronic, sub-lethal microcystin toxicity. It is possible that the lack of increase in caspase gene expression reflects the normal intracellular situation in which caspases are preformed in cells ready to respond to intracellular or extracellular signaling. It has been shown that ZVAD-fmk, a general caspase inhibitor, can inhibit microcystin-induced apoptosis, thus suggesting caspase involvement (Fladmark et al., 1999; Ding et al., 2002). However, ZVAD-fmk has also been shown to also inhibit other proteases such as calpain. In primary rat hepatocytes, it was found that MCLR treatment resulted in an increase in calpain activity and that two calpain inhibitors significantly blocked MCLR-induced calpain activation and subsequent cell death (Ding et al., 2002). Also in CaCo2 and MCF-7 cells, a dose-dependent increase in calpain activity 24 hours after MCLR exposure was identified (Botha et al., 2004). Interestingly, several cysteine protease inhibitor genes were upregulated in this study following chronic microcystin exposure including cystatin B and the calpain inhibitor, cast. This upregulation further suggests a role of cysteine proteases in chronic microcystin toxicity. To characterize the timing of microcystin-induced apoptosis in the chronic toxicity of microcystin, we examined the gene expression pattern of the cysteine protease inhibitor, cystatin B, over the time course of the study. Using RT-PCR, we were able to show that the increased expression of cystatin B began as early as 4 hours post-treatment with microcystin and the increased expression continued throughout the study (Table 4). These changes would indicate a role of cysteine proteases in the pathogenesis of chronic, sublethal microcystin exposure. In addition, we were also able to identify the upregulation of the bcl proapoptotic genes, bax and bim (bcl211). The products of these genes promote the release of cytochrome c and are involved in the intrinsic, mitochondrial apoptotic pathway (Harada and Grant, 2003). Therefore, in our studies we noted upregulation of both pro- and anti-apoptotic genes suggesting increased activity in response to decreased control of cell proliferation.

Many studies have demonstrated the role that oxidative stress plays in MCLR-induced hepatotoxicity (Ding and Nam Ong, 2003). Both acute and chronic microcystin toxicity studies have found increases in the formation of reactive oxygen species and in lipid peroxidation following microcystin exposure (Guzman and Solter, 1999; Moreno et al., 2005). However, global gene transcript changes in the glutathione pathway associated with chronic, sublethal microcystin exposure have not been previously reported. Following 28 days of sublethal microcystin exposure, this study identified large increases in the expression of a multitude of genes involved in glutathione metabolism (Table 3). An increase in glutathione S-tranferase mu 3 (gstm3) expression greater than 20-fold was measured in MCLR-treated animals. This increase suggests that it is involved in the Phase II detoxification of MCLR with glutathione. Other glutathione S-tranferases were also found to be upregulated following MCLR exposure, but not to the extent of gstm 3 (Table 3). Both glutathione peroxidases, gpx2 and gpx3, were upregulated greater than 10-fold suggesting a role of glutathione peroxidase in the intracellular protection from MCLR. In addition, there was also an increase in the expression of gamma-glutamylcysteine synthetase (gclc) and glutathione synthetase (gss), which are both key enzymes in glutathione synthesis. An increase in the expression of other oxidative stress-related genes including NAD(P)H dehydrogenase quinone 1 (nqo1), ATX1 antioxidant protein 1 homolog 1 (atox1), and thioredoxin reductase 1 (txnrd1) were also found in the MCLR-treated group (Table 3). In addition, we examined changes in the expression patterns of several redox genes over the time- course of the study. Using RT-PCR we found a similar pattern of gene expression for gstm3, gss and nqo1. An upregulation of these genes was identified as early as 4 hours postdosing and their expression was increased throughout the duration of the study (Table 4). Earlier studies have shown that microcystin detoxification involves conjugation to glutathione (Kondo et al., 1996) via the action of glutathione S-transferases in the liver of microcystin-exposed animals (Pflugmacher et al., 1998; Takenaka, 2001) and that changes in glutathione, glutathione peroxidase, and glutathione reductase activities associated with microcystin exposure occur both in vivo and in vitro (Ding et al., 2000b; Bouaicha and Maatouk, 2004; Gehringer et al., 2004; Moreno et al., 2005). Mice given a single 75% LD50 dose intraperitoneally showed a statistically significant increase in lipid peroxidation 16 hours post-exposure. A significant increase in glutathione peroxidase activity, a slight increase in mean soluble glutathione S-transferase and an overall increase in the mean level of oxidized glutathione (GSSG) were also identified in the livers of all toxin-treated mice. However, no significant changes were measured in total glutathione levels in the livers of the treated animals. The increase in both glutathione peroxidase (GPX2 and GPX4) and glutathione S-transferase (GST m2 and GST m5) activities corresponded to an increase in their transcription levels. In addition, there was also an increase in the transcription of gamma-glutamyl transferase, the rate-limiting enzyme in glutathione synthesis (Gehringer et al., 2004). Together, the data in our study and numerous previous studies strongly support the role of glutathione in both the Phase II metabolism of MCLR and in the defense against reactive oxygen species associated with MCLR exposure.

Previous work has shown that activation of the mitogen-activated protein kinase (MAPK) pathway occurs following treatment with protein phosphatase inhibitors (Davis et al., 1996). It has been suggested that this activation could lead to an increase in protooncogene transcription, resulting in an increase in cellular proliferation and tumor-promotion (Toivola and Eriksson, 1999). In this study, we identified an increase in expression of several genes involved in MAPK and janus kinase signaling including janus kinase 1 (jak1), map3k1, mapk3, and mapk9. We also identified an increase in the expression of the cellular oncogenes jun, jund1, and myc. The increased expression of jun was shown by RT-PCR to begin as early as 4 hours posttreatment with microcystin and continued to increase throughout the study (Table 4). It has been documented that PP2A is a major negative regulator of several steps in the MAPK signaling pathways (Hunter, 1995). The inhibition of PP2A by microcystin could then result in the activation of these pathways resulting in the transcription of genes involved in cellular proliferation including jun, fos, and myc. The upregulation of these signaling pathways and cellular oncogenes in this study would support a potential role of these signaling pathways in the tumor promoting activities of microcystin. Determination of the expression of these genes at later time points during hepatocarcinogenesis will be important.

This study is the first to measure changes in global hepatic gene expression associated with chronic, sublethal microcystin exposure in mice and associate them with functional and histological changes in the liver. Previous studies have reported increases of hepatic oxidant damage due to microcystin administration and related this to tumor promotion. Our results support these findings and significantly add to them by quantifying increases in numerous additional genes related to oxidant injury and phase II metabolism. In addition, alterations in the expression of genes related to glucose homeostasis and albumin metabolism are manifested as reduced liver function that appear clinically as decreases in total serum albumin and glucose. Since previous studies have shown that microcystin is carried into hepatocytes via hepatocyte-specific organic anion-transporting polypeptides, it is significant that the genes regulating these transport peptides are downregulated with chronic microcystin administration. Microcystin has been shown epidemiologically to be associated with liver cancer in humans, and has been shown to be a promoter of hepatocellular carcinoma in animals. An extremely significant finding of the current study is that chronic microcystin exposure results in histological changes of hepatocyte hypertrophy and karyomegaly with the appearance of numerous multinucleated hepatocytes. It is likely that with continued microcystin dosing, these lesions would have progressed to preneoplastic foci followed by hepatocellular carcinoma as has been described previously in longer studies. In our current study, the suggestive preneoplastic hepatocellular changes and apparent mitotic arrest (multinucleated hepatocytes) were accompanied by alterations in the expression of genes related to intracellular signaling, GTPase activity, apoptosis, and cytoskeletal organization all of which can be related to neoplastic transformation. Therefore, these gene changes may well represent some of the earliest alterations in the transformation of normal hepatocytes to neoplastic ones and be early indications of the formation of hepatocellular carcinoma. Clearly, further studies are indicated to better define the gene changes related to the pathogenesis of liver cancer.