Nonalcoholic fatty liver disease (NAFLD) comprises a spectrum of pathologic lesions that ranges from hepatic steatosis to a form of fatty liver hepatitis known as nonalcoholic steatohepatitis (NASH). Histologic features independently associated with the diagnosis of NASH in human biopsies include hepatic steatosis, hepatocyte ballooning, lobular inflammation, Mallory's hyaline, and perisinusoidal fibrosis. Moreover, NASH has the potential to advance to cirrhosis requiring liver transplant (Powell et al., 1990; Abdelmalek et al., 1995; Charlton et al., 2001; Caldwell and Hespenheide, 2002). A recent overview of clinical data estimated that NAFLD and NASH affect 17 to 33% and 5.7 to 17% of American adults, respectively. Although the etiology of NASH is not completely understood, most investigators agree that a critical baseline of steatosis requires a second “hit” of oxidative stress that promotes the subsequent inflammation and fibrosis of NASH. NAFLD is primarily associated with features of the metabolic syndrome, including obesity, diabetes, dyslipidemia, and insulin resistance (McCullough, 2006). Importantly, these NAFLD-associated features invariably require pharmacologic intervention with one or more therapeutic agents. It therefore follows that a significant portion of the NAFLD demographic group is already using prescription medication.

In the present study, NAFLD was recapitulated in two of its prominent forms, simple fatty liver (SFL) and NASH, respectively. A high-fat diet was used for the SFL model, while a methionine- and choline-deficient (MCD) diet was used for the NASH model. The histologic changes that occur in the MCD model have been thoroughly characterized and determined to be remarkably similar to those seen in human NASH (George et al., 2003). Histologic features of NASH necessarily include hepatic steatosis, mild, lobular inflammatory infiltration, and hepatocellular ballooning degeneration (Brunt and Tiniakos, 2005).

Because the vast majority of drug metabolism occurs in the liver, it is particularly important to understand the transport processes involved in the elimination of drug metabolites formed in this organ, especially a diseased one. In the liver, there are two basic types of transporters. These include the uptake transporters, e.g., organic anion-transporting polypeptides and organic cation transporters, which facilitate the extraction of drugs from portal blood into hepatocytes, and the efflux transporters, e.g., multidrug resistance-associated proteins (Mrps), bile salt excretory protein (Bsep), and breast cancer resistance protein (Bcrp), which extrude their substrates out of hepatocytes into bile or blood. Although not all drugs require uptake transport to cross the sinusoidal membrane into hepatocytes, virtually all of their metabolites require efflux transporters to be excreted into bile and blood for biliary and urinary elimination, respectively.

Acetaminophen (APAP) provides a fine example of the integral role played by efflux transporters in drug metabolite excretion from liver. After passive diffusion into hepatocytes, APAP is mainly conjugated with sulfate (SULF) or glucuronic acid (GLUC) to form APAP-SULF and APAP-GLUC, respectively (Moldéus, 1978; Grafström et al., 1979). A small portion of a dose of APAP will undergo cytochrome P450 metabolism to the reactive metabolite N-acetyl-p-benzoquinoneimine, which is subsequently conjugated with glutathione (GSH) to form APAP-GSH. In rats, excretion of APAP-GLUC and APAP-GSH metabolites from the liver occurs predominantly via the biliary route, whereas the APAP-SULF metabolite predominantly undergoes sinusoidal efflux and subsequent elimination in the urine (Gregus et al., 1988). Importantly, all APAP metabolites require efflux transport to be excreted from the liver, and each can be detected in both bile and urine (Gregus et al., 1988; Xiong et al., 2002; Chen et al., 2003; Manautou et al., 2005; Zamek-Gliszczynski et al., 2006a).

In vivo disposition studies and in vitro functional transport experiments indicate that Mrp2, Mrp3, Mrp4, and Bcrp each have the ability to transport a variety of unconjugated and conjugated drugs, including APAP metabolites (Büchler et al., 1996; Keppler et al., 1997; Xiong et al., 2000; Nakanishi et al., 2003; Zamek-Gliszczynski et al., 2005, 2006b). It is important to emphasize the distinctive location of these four transporters in hepatocytes, as well as their respective APAP metabolite substrates. Mrp2 and Bcrp are localized to the canalicular (apical) membrane of hepatocytes from which they excrete their substrates into the bile canaliculi. Accordingly, in a healthy liver, biliary excretion of the SULF, GLUC, and GSH conjugates of APAP is predominantly mediated by Mrp2, whereas Bcrp appears also to contribute to excretion of APAP-SULF conjugates (Büchler et al., 1996; Keppler and König, 1997, 2000; Borst et al., 2000; Keppler et al., 2000; Chen et al., 2003; Zamek-Gliszczynski et al., 2005). Mrp3 and Mrp4 are expressed at the sinusoidal (basolateral) membrane of hepatocytes and cholangiocytes from which they expel their substrates into the blood (König et al., 1999; Donner and Keppler, 2001; Soroka et al., 2001). Sinusoidal excretion of the APAP-GLUC metabolite from hepatocytes is predominantly mediated by Mrp3, whereas Mrp4 appears to mediate excretion of APAP-SULF metabolites (Chen et al., 2001; Manautou et al., 2005; Zamek-Gliszczynski et al., 2006a). Recent studies indicate that Mrp3 and Mrp4 have an equal role in the efflux of APAP-SULF (Zamek-Gliszczynski et al., 2006b).

During other liver disease states, such as septic cholestasis, obstructive cholestasis, and alcoholic and viral hepatitis, human and rodent studies have demonstrated alterations in cytochrome P450-mediated drug metabolism, transport, and pharmacokinetics (Forrest et al., 1977; Farrell et al., 1978, 1979; Narang et al., 1981, 1982, 1985; Figg et al., 1995; Westphal and Brogard, 1997; Kubitz et al., 1999; Nadai et al., 2001; Congiu et al., 2002). However, there is little information on the effect of NAFLD on these same parameters. Therefore, the purpose of the present study was to determine the effect of NAFLD on efflux transporter expression in liver and disposition of APAP metabolites.

Materials and Methods

Materials. APAP and urethane were purchased from Sigma-Aldrich (St. Louis, MO). The analytical column and guard cartridges for high performance liquid chromatography (HPLC) analysis were purchased from Bodman (Aston, PA). HPLC-grade water, HPLC-grade methanol, Tris base, SDS, sodium chloride, and sucrose were purchased from Fisher Scientific Co. (Pittsburgh, PA), and all other chemicals used were of reagent grade or better.

Treatment of Animals. Male Sprague-Dawley rats weighing 200 to 250 g were purchased from Harlan (Bar Harbor, ME). Animals were housed in a temperature-, light-, and humidity-controlled environment in hanging cages with hardwood chips. Rats (n = 5-9/group) were randomly placed into one of three diet treatment groups: control group to simulate healthy liver (Teklad 7001 4% Mouse/Rat Diet; Harlan Teklad, Madison, WI), HF group to simulate SFL [18% butter (w/w) diet], and MCD group to simulate NASH (methionine- and choline-deficient diet) (Dyets Incorporated; Bethlehem, PA) ad libitum for 8 weeks. According to its manufacturer, the Teklad 7001 4% Mouse/Rat Diet is a “complete and balanced, fixed formula diet formulated to provide consistent nutrition to research rodents.” The Institutional Animal Care and Use Committee of the University of Arizona approved the animal studies described below.

Histology. Sections of liver from control, SFL, and NASH rats were embedded in Tissue-Tek OCT compound (Electron Microscopy Sciences; Hatfield, PA), rapidly frozen in ethanol cooled with dry ice, and stored at -80°C until use. With use of a Microm HM 550 cryostat (Richard Allen Scientific, Kalamazoo, MI), 5-μm slices from each diet group were generated and mounted onto Superfrost Plus slides (Fisher Scientific Co., Houston, TX). The mounted tissues were stained with hematoxylin and eosin (H&E) or Masson's trichrome stain. Histologic analyses were conducted under light microscopy (40× objective) using coded slides to avoid observer bias. Parameters of the histologic analyses included degree of steatosis, inflammatory infiltration, and fibrosis.

APAP Excretion Experiments. Under urethane-induced anesthesia (1.1 g/kg, 5 ml/kg, i.p.), the left femoral artery and right femoral vein were cannulated with PE 50 polyethylene tubing. The bile duct was subsequently cannulated with PE 10 polyethylene tubing distal to the bile duct bifurcation. Before injection of APAP, the femoral artery and vein cannula were flushed with approximately 0.5 ml of saline-5% mannitol solution containing 50 units of heparin. APAP, dissolved in saline containing 5% mannitol, was injected into the femoral vein cannula at a dose of 1 mmol/kg (10 ml/kg) using a 3-ml syringe with a 23-gauge needle. After APAP injection, the femoral vein cannula was flushed with a 0.5-ml injection of 0.5% saline-5% mannitol solution. At 0, 2, 10, 20, 40, 60, and 90 min after APAP administration, ∼250-μl blood samples were collected from the femoral artery in heparinized tubes. Bile was collected at 15, 30, 45, 60, 75, and 90 min after APAP administration. To maintain urine flow, 1 ml/kg saline-5% mannitol solution was injected through the femoral vein cannula every 15 min after APAP administration. Urine was collected from the bladder at 90 min after APAP administration by ligation of the bladder and removal with a 22-gauge needle and 3-ml syringe. Livers and kidneys were also harvested at the end of each APAP disposition experiment and snap-frozen in liquid nitrogen. All samples were stored at -80°C until analysis.

Detection of APAP and Its Metabolites in Bile, Plasma, and Urine Samples. APAP and its metabolites in bile, plasma, and urine samples were quantified by high-performance liquid chromatography analysis based on previously described methods (Howie et al., 1977; Chen et al., 2000; Slitt et al., 2003). APAP and its metabolites were resolved using a Zorbmax SB-C₁₈ reverse-phase 4.6-mm × 25-cm column with a Phenomenex Security Guard Column Guard and eluted using a mobile phase composed of 12.5% HPLC-grade methanol, 1% acetic acid, and 86.5% water, run isocratically at a flow rate of 1.2 ml/min. The elution of metabolites was monitored at a wavelength of 254 nm. Retention times of APAP and its metabolites were determined by comparison with those of authentic standards. Because this HPLC method does not separate the cysteinylglycine and cysteine conjugates of APAP, they were quantified together as APAP-CG/CYS.

Samples were analyzed using a Beckman System Gold HPLC system (Beckman Coulter, Fullerton, CA) equipped with a 128-nm solvent module and a 166-nm detector. Quantitation was based on integrated peak areas. The concentrations of APAP and its metabolites were calculated using an APAP standard curve since the molar extinction coefficients of APAP and its conjugated metabolites are approximately the same (Howie et al., 1977). To precipitate proteins in bile, plasma, and urine, samples were diluted 1:2, 1:2, and 1:3, respectively, with ice-cold methanol and centrifuged at 4,000g for 30 min at 4°C. The resulting supernatants were collected and diluted in mobile phase 1:3 (bile and urine) or 1:2 (plasma) before HPLC analysis. Liver samples were homogenized (1:9 w/v) in ice-cold HPLC-grade methanol. Homogenates were centrifuged at 4,000g for 20 min at 4°C, and the supernatants were filtered through a 0.45-μm filter. Samples were further diluted 1:3 with mobile phase before HPLC analysis.

RNA Extraction. Total RNA from liver and kidney tissue was extracted using RNAzol B reagent (Tel-Test Inc., Friendswood, TX) according to the manufacturer's protocol. The quality of RNA samples was judged by the integrity and relative ratio of 28S and 18S rRNA bands after agarose gel electrophoresis.

Messenger RNA Expression Analysis. Rat Mrp1-7, Bsep, and Bcrp mRNA levels were measured using the branched DNA signal amplification assay. To measure the levels of Mrp1-6, Bsep, and Bcrp mRNA, respectively, oligonucleotide probe sets containing multiple mRNA transcript-specific capture, label, and blocker probes were designed using ProbeDesigner software (v1.0; Bayer Corp., Diagnostics Div., Tarrytown, NY). The probe sets used in this study have been described previously, including Mrp1-3 (Cherrington et al., 2002), Mrp4-6, Bsep, Mrp7, and Bcrp (Tanaka et al., 2005). Probes were designed with an annealing temperature of approximately 63°C, which enabled the hybridization conditions to be held constant (i.e., 53°C) during each hybridization step and for each probe set. Every probe developed in ProbeDesigner was submitted to the National Center for Biotechnological Information for nucleotide comparison by the basic logarithmic alignment search tool (BLASTn), to ensure minimal cross-reactivity with other known rat sequences and expressed sequence tags. Oligonucleotides with a high degree of similarity (≥80%) to other rat gene transcripts were excluded from the design.

Total RNA (1 μg/μl; 10 μl/well) was added to each well of a 96-well plate containing capture hybridization buffer and 50 μl of each diluted probe set. For each gene, total RNA was allowed to hybridize to the probe set overnight at 53°C. Subsequent hybridization and quantitation steps were performed according to the manufacturer's protocol (QuantiGene Reagent System; Panomics, Fremont, CA).

Western Blot Analysis of Mrp2, Mrp3, Mrp4, and Bcrp Protein Expression. Liver crude membrane preparations were made as described previously (Slitt et al., 2003). Liver and kidney protein concentrations were determined using the BCA Protein Assay Kit (Pierce Biotechnology, Rockford, IL). Liver (40 μg/lane) membrane proteins were electrophoretically resolved on Tris-glycine SDS-polyacrylamide gels. Proteins were transblotted onto nitrocellulose membranes (Bio-Rad, Hercules, CA). Membranes were blocked for 1 h in 5% nonfat dry milk in Tris-buffered saline-Tween buffer before incubation with antibodies. Primary antibody clones, membrane incubation times and dilutions in 5% nonfat dry milk were as follows. For Mrp2, conditions included a 1:2000 dilution of mouse anti-human M₂III-6 monoclonal antibodies (ID Labs Inc., London, ON, Canada). For Mrp3, conditions included a 1:5000 dilution of affinity-purified antibodies against rat Mrp3 protein (Slitt et al., 2003). For Mrp4, conditions included a 1:500 dilution of goat anti-human polyclonal antibodies (Abcam Inc., Cambridge, MA). For Bcrp, conditions included a 1:500 dilution of mouse anti-human bxp-21 monoclonal antibodies (Kamiya Biomedical, Seattle, WA). Blots were subjected to four 15-min washes in Tris-buffered saline-Tween buffer before incubation for 1 h at room temperature with secondary antibody. The secondary antibody dilutions included 1:2000 (Mrp2), 1:5000 (Mrp3 and Mrp4), and 1:3000 (Bcrp). Protein-antibody complexes were detected using ECL Western blotting reagents (GE Healthcare, Little Chalfont, Buckinghamshire, UK) and Blue Lite Autoradiography Film (ISC BioExpress, Kaysville, UT).

Statistics. For excretion studies, statistical differences between diet groups at each time point were determined using a one-way analysis of variance followed by a Duncan's multiple range post hoc test. For mRNA and protein expression analyses, statistical differences between diet groups were also determined using a one-way analysis of variance followed by Duncan's multiple range post hoc test.

Results

Liver Histology.Figure 1 demonstrates the effect of the respective high-fat and MCD diets on liver histology after 8 weeks. Hepatic steatosis was apparent at 2 weeks in rats fed the MCD diet but not in rats fed the control standard diet. Notably, inflammation and fibrosis were plainly absent in the MCD group at this early time point (data not shown). However, after 8 weeks of consuming the diet, histologic analyses of H&E (A-C)- and trichrome-stained sections (D-F) were indicative of two distinct NAFLD pathologies that corresponded satisfactorily with simple hepatic steatosis and early-stage NASH in human patients (Brunt and Tiniakos, 2005). Histology of H&E-stained liver sections from the control group was unremarkable, with minimal periportal inflammatory infiltrates (A) and minimal fibrosis (D). Sections from the high-fat diet group were characterized by predominantly periportal, microvesicular steatosis (B) and minimal fibrosis (E). Sections from the MCD diet group revealed diffuse, predominantly macrovesicular steatosis (C) with mild biliary oval cell hyperplasia, mild periportal inflammatory infiltrates, and fibrosis (F).

Efflux Transporter mRNA Levels in the Liver After 8 Weeks of Feeding on the Respective High-Fat and MCD Diets. After 2 weeks of feeding on the control and MCD diets, there were no discernible differences in mRNA levels of Mrp2, Mrp3 Mrp4, or Bcrp (data not shown). This finding suggests that acute dietary deficiency of methionine and choline does not alter transcriptional regulation of these four major efflux transporters. Figure 2 illustrates the effect of 8 weeks of feeding on the respective high-fat and MCD diets on efflux transporter mRNA levels in liver. In high-fat rats, Mrp5 mRNA levels were increased 350%, whereas Mrp6 and Bsep were each decreased by 50%. Mrp1-4 and Bcrp mRNA levels were unchanged in livers of high-fat rats. In MCD rats, Mrp3, Mrp4, and Bcrp mRNA expression levels were increased 2990, 405, and 59%, respectively. Mrp1 and Mrp2 mRNA levels were not altered by the high-fat or MCD diet.

Efflux Transporter Protein Levels in Liver After 8 Weeks of Feeding on the Respective High-Fat and MCD diets.Figure 3 shows the effect of the respective high-fat and MCD diets on Mrp2, Mrp3, Mrp4, and Bcrp protein levels in liver. In MCD rats, Mrp2, Mrp3, Mrp4, and Bcrp protein expression levels were increased 185, 253, 246, and 71%, respectively. Notably, no changes in protein levels of the major APAP metabolite transporters were observed in the livers of high-fat rats.

Effect of High-Fat and MCD Diets on Bile Flow.Figure 4 shows the cumulative amount of bile collected during the 90 min after APAP administration for control, high-fat, and MCD rats. The cumulative bile volume in high-fat rats was significantly greater than the control by 19% at 60 min and 26% at 90 min. MCD rat bile flow did not differ from the control. Importantly, the data from these models of NAFLD do not indicate any alteration of bile secretory function.

Effect of High-Fat and MCD Diets on Biliary Excretion of APAP and APAP Metabolites.Figure 5, top left, shows the concentrations of APAP and its major metabolites in bile over 90 min. The concentration of APAP in the bile of MCD rats was significantly elevated compared with the control (20% higher) at 15 min. It is important to note that parent APAP that does not undergo metabolism within a hepatocyte can passively diffuse across the canalicular or basolateral membrane. Thus, any intergroup differences in the biliary or urinary concentration of APAP would not be reflective of alterations in function by efflux transporters.

The biliary concentration of each major metabolite of APAP was recurrently observed to be significantly decreased in MCD rats relative to control and high-fat rats. In contrast with APAP parent, each of these metabolites, i.e., APAP-GLUC, APAP-GSH, and APAP-SULF, does require active transport to be excreted from hepatocytes.

Figure 5, top right, illustrates the concentration of APAP-GLUC in bile of control, high-fat, and MCD groups over the 90-min collection period. APAP-GLUC was significantly lower in MCD rats than in control rats throughout the 90 min, with the maximal difference (73% lower) being observed by 45 min. The biliary APAP-GLUC levels in MCD rats became significantly lower than those in the high-fat group by 30 min. Of the three major metabolites measured in the bile, APAP-GLUC was the most abundant metabolite detected throughout the time course in each group.

Figure 5, bottom left, illustrates the concentration of APAP-SULF in bile of control, high-fat, and MCD rats over the 90-min collection period. APAP-SULF was significantly lower in MCD rats than in control rats throughout the 90 min, with the maximal difference (67% lower) being observed by 30 min. In addition, the biliary APAP-SULF levels in MCD rats also became significantly lower than those in the high-fat group by 15 min. The biliary concentration of APAP-SULF in high-fat rats was lower than that for the control rats (by 23%) only at 15 min.

Figure 5, bottom right, illustrates the concentration of APAP-GSH in bile of control, high-fat, and MCD groups over the 90-min collection period. In the MCD group, APAP-GSH levels in bile were significantly lower than those in the control group throughout the 90 min. The maximal difference from the control group (93% lower) was observed by 30 min.

Effect of High-Fat and MCD Diets on APAP and APAP-GLUC Metabolite Concentrations in Liver.Figure 6 shows the concentrations of APAP and its APAP-GLUC metabolite in the livers of control, high-fat, and MCD rats, respectively, at the completion of the study. The APAP parent concentration in the livers of high-fat rats was 85% lower than that of the control rats. Notably, the concentration of APAP-GLUC in the livers of MCD rats was 79% lower than that of the control rats. The APAP-GSH, APAP-SULF, and APAP-CG/CYS metabolites were not detectable in livers in any group.

Effect of High-Fat and MCD Diets on Plasma Concentration of APAP and APAP Metabolites.Figure 7 shows the concentrations of APAP, APAP-GLUC, and APAP-SULF in plasma over 90 min. APAP-GSH was not detectable in the plasma of rats in these studies. For APAP (Fig. 7, top), there were no plasma level differences between the control, high-fat, and MCD groups throughout the time course. The APAP-GLUC metabolite (Fig. 7, center) was detected in the plasma of the control and high-fat groups by 10 min after APAP administration. In contrast, in the MCD group, the APAP-GLUC metabolite was present in the plasma within 2 min of APAP administration. It is interesting that by 10 min, the concentration of the APAP-GLUC metabolite in the MCD animals was 4.8-fold higher than in controls. Plasma levels of APAP-GLUC in MCD rats remained significantly elevated above those of the control group throughout the remainder of the time course by an average of 3.3-fold. APAP-SULF was present in the plasma of control, high-fat, and MCD rats within 10 min of administration of the APAP (Fig. 7, bottom). Within 40 through 90 min, the plasma concentration of APAP-SULF in MCD rats was significantly lower than the control group by an average of 50%.

Effect of High-Fat and MCD Diets on Urinary Excretion of APAP and APAP Metabolites.Figure 8 shows the concentration of APAP and its metabolites in the urine of control, high-fat, and MCD rats at 90 min after APAP administration. There were no group differences for APAP, APAP-SULF, APAP-CG/CYS, and APAP-NAC, nor were there differences in the net (i.e., total) concentration of metabolites between control, high-fat, and MCD rats. However, the concentration of the APAP-GLUC metabolite was 80% higher in the urine of MCD rats compared with control rats. Importantly, this same metabolite was elevated in the plasma of MCD rats throughout the 90-min time course. APAP-GSH was not detected in the urine of any group.

Discussion

The current study presents novel data that demonstrate the effects of experimental NAFLD on transporter expression and drug disposition. The clinical relevance of these topics is underscored by the fact that the NAFLD patient population, a demographic that uses a great quantity and variety of prescription medications, is expected to continue growing at an alarming rate (Kopelman, 2000; Brunt and Tiniakos, 2005). To expand our comprehension and/or potentially develop pharmacologic treatment for this disorder, extensive studies of accurate animal models will serve as an invaluable tool. It is thus pertinent to note that the current models of NAFLD (i.e., high-fat and MCD diets) provided accurate histologic representations of SFL and NASH, respectively.

It is important to note that there were no alterations in bile flow between the control and modified diet groups. The cumulative bile flow was significantly elevated in high-fat rats but only at the later time points. Each of these groups is a contrast to the obese Zucker rat (a model of insulin resistance and obesity) in which bile secretory function in fatty liver is impaired, nonetheless in the absence of cholestatic injury (Pizarro et al., 2004; Geier et al., 2005).

This is the first study to examine liver efflux transporter expression in high-fat and MCD dietary models of simple hepatic steatosis and NASH in rats. Both models resulted in multiple alterations in liver expression of efflux transporters. It is interesting to note that when the mRNA levels of a given efflux transporter were altered with the high-fat diet, the MCD model generally generated the same effect (i.e., increased Mrp5 in liver and decreased Mrp6 in liver). In contrast, only MCD rats were characterized by up-regulation of mRNA and protein levels of the hepatic efflux transporters Mrp2, Mrp3, Mrp4, and Bcrp.

The obvious alterations in transporter expression and drug disposition in the current study encourage further investigation into mechanisms of transcriptional regulation during NAFLD that would be likely to provide greater insight into necessary differences at the molecular level between simple fatty liver and steatohepatitis in rats. It is useful to note that, at the histologic level, the major principal difference between the high-fat and MCD rats was the presence of inflammatory infiltrates and fibrosis in MCD rat livers. Thus, one important factor to consider with regard to efflux transporter regulation during NAFLD is the role of proinflammatory cytokines. Elevated levels of the proinflammatory cytokines tumor necrosis factor (TNF)-α and interleukin (IL)-6 have been detected in the plasma of patients with biopsy-proven NASH but not in patients with simple hepatic steatosis (Kugelmas et al., 2003; Bahcecioglu et al., 2005; Abiru et al., 2006). Consistent with the current model, the early stages of MCD-induced NASH are characterized by up-regulation of hepatic IL-6 mRNA and elevated plasma levels of TNF-α (Chawla et al., 1998; Stärkel et al., 2003). Although there is little information on the role of IL-6 in the regulation of hepatic transporters in rats, the role of TNF-α in this case has been well studied by using the endotoxin-induced model of sepsis (Whiting et al., 1995; Green et al., 1996; Geier et al., 2003; Cherrington et al., 2004). Specifically, down-regulation of Mrp2 mRNA during endotoxemia has been demonstrated to be dependent on TNF-α (Geier et al., 2003). In contrast, in obese Zucker rats, which develop elevated plasma levels of TNF-α, abrogation of TNF-α activity with etanercept has no effect on Mrp2 protein levels (Liu et al., 2002; Pizarro et al., 2004). Although there are limited data on the hepatic and serum levels of proinflammatory cytokines in rat models of NAFLD, it seems reasonable to propose that cytokines may be important in transporter regulation in the MCD model of NASH.

Importantly, the biochemical mechanisms by which the high-fat and MCD diets induce their respective pathologies in the liver are markedly distinct. Thus, direct comparison of the effects, e.g., gene expression, between the high-fat and MCD groups was deemed inappropriate for statistical analysis in these studies. In the high-fat dietary model, induction of hepatic steatosis is most likely a consequence of increased net flux of lipids through the liver. In contrast, the MCD model achieves its histologic recapitulation of NASH via a series of events consistent with lipid peroxidation as a pathogenic mechanism for initiation and perpetuation of liver injury, inflammatory recruitment, stellate cell activation, and fibrogenesis (George et al., 2003). Although the initial steatosis of the MCD model is not associated with the peripheral insulin resistance that is frequently observed in human NASH, the consequent damage, including hepatocellular injury, inflammatory recruitment, and zone 3 fibrosis does resemble that observed in human NASH (George et al., 2003).

It is pertinent to note that the difference between the “control” and MCD diets used in the current study was not exclusive to the presence or absence of methionine and choline. An arguably more appropriate dietary control would be achieved by merely resupplementing the MCD diet with methionine and choline. Nonetheless, both diets applied in the current study are, in fact, fully defined in terms of their amino acid content, and it is thus known that the control diet (Teklad 7001 4% Mouse/Rat Diet) is methionine- and choline-sufficient, resulting in a histology reflective of a healthy liver (Fig. 1). At the same time, there is no question that the methionine and choline deficiency is responsible for the pathology observed (Fig. 1F). However, it is not entirely clear whether the alterations in transporter expression (and thus APAP-GLUC metabolite excretion) are the consequence of a chronic dietary deficiency in methionine and choline, or, conversely, a regulatory response to the progressive damage and/or steatosis that occur as a result of the deficiency. According to the mRNA and histologic analysis of control and MCD rat livers after 2 weeks of the respective diets, acute biochemical deficiency of methionine and choline is clearly not responsible for inducing alterations in transporter expression.

The latter part of the current study examined the disposition of acetaminophen metabolites during NAFLD. In high-fat rats, there were no significant differences in the concentration of APAP metabolites in the plasma or urine and relatively few differences in the biliary excretion of metabolites. It is therefore apparent that simple hepatic steatosis has no discernible effect on the disposition of APAP metabolites. In contrast, there were obvious alterations in the disposition of the APAP-GLUC, APAP-GSH, and APAP-SULF metabolites in MCD rats.

One of the most important observations in the current study was the altered disposition of APAP-GLUC metabolites in the rats with MCD-induced NASH. The marked reduction in the biliary concentration of APAP-GLUC was accompanied by a concurrent increase in its plasma and urine concentrations. These were in the range of those observed after pretreatment of the same strain of rats with various Mrp3-inducing compounds, including phenobarbital (Brouwer and Jones, 1990; Xiong et al., 2002; Slitt et al., 2003). In hepatocytes of normal rat liver, Mrp3 protein is expressed at much lower levels than Mrp2. Thus, although Mrp3 has a much higher affinity for APAP-GLUC, this metabolite is normally excreted predominantly into bile in rats (Brouwer and Jones, 1990; Gregus et al., 1990). However, in the livers of rats with MCD-induced NASH, the expression levels of Mrp2 and Mrp3 appear to have become relatively equal, probably resulting in competition for the APAP-GLUC substrate. It therefore follows that the apparent shift from biliary to urinary excretion is most likely due to 1) the increased Mrp3 protein levels in liver and 2) the higher affinity of Mrp3 for the APAP-GLUC metabolite.

The concentration of APAP-SULF in bile was significantly lower in MCD rats compared with the control rats throughout the entire time course. This result was unanticipated, given the substantial up-regulation of both Mrp2 and Bcrp, which are responsible for the canalicular excretion of APAP-SULF. However, the plasma concentration of APAP-SULF was also consistently lower in MCD rats, and although the decrease of APAP-SULF in MCD rat urine did not reach statistical significance, it too appears to be slightly lower than the control. Therefore, it would seem that the differences in disposition may have resulted from a decrease in sulfotransferase expression, sulfotransferase activity, and/or the sulfotransferase cofactor 3′-phosphoadenosine-5′-phosphosulfate.

The concentration of APAP-GSH in bile was also decreased in MCD rats throughout the time course. This result was also unforeseen, given the pronounced induction of Mrp2 protein levels. In contrast, the concentrations of APAP-GSH derivatives APAP-CG/CYS and APAP-NAC in urine were not altered in MCD rats. These data indicate that the sinusoidal efflux transporters responsible for the hepatovascular excretion of APAP-GSH and its derivatives were able to maintain this function in MCD rat livers. However, it is also clear that fewer APAP-GSH metabolites and derivatives were formed in MCD rat livers. This effect may be partly due to the reduction ex vivo in cytochrome P450-mediated bioactivation of APAP to the reactive N-acetyl-p-benzoquinoneimine (data not shown), which undergoes a spontaneous electrophilic reaction with GSH to form APAP-GSH. Hepatic GSH levels were not altered in either model of NAFLD (data not shown).

The metabolic capacity of the kidneys to form derivatives of the APAP-GSH conjugate is also an important consideration in the case of APAP disposition. After efflux of APAP-GSH conjugates from the liver into the sinusoidal blood, they are rapidly degraded into APAP-CG, APAP-CYS, and APAP-NAC by the kidneys (Moldéus, 1978). Thus, only the degradation products of APAP-GSH, including APAP-CG, APAP-CYS, and APAP-NAC, are normally detectable in the urine (Gregus et al., 1988). Mrp2, which is expressed in the brush border membrane of rat proximal tubule cells, is presumably responsible for the tubular secretion of these APAP-GSH derivatives (Schaub et al., 1997). There were no alterations in the urinary excretion of the APAP-GSH derivatives. Bcrp and Mrp4 are also present at high levels in rat kidneys and appear to be important in the secretion of sulfate conjugates from proximal tubule cells (Jonker et al., 2002; Mizuno et al., 2004; Tanaka et al., 2005). In contrast, Mrp3 is expressed at the basolateral membrane of renal proximal tubule cells and may participate in efflux of APAP-GLUC into blood (Kuroda et al., 2004). Although Mrp3, Mrp4, and Bcrp are each expressed to a considerable extent in the kidney, their contribution to renal excretion of APAP-GLUC (by Mrp2) and APAP-SULF (by Mrp4 and Bcrp) is probably very limited. This is because, in rats and humans, the vast majority of APAP biotransformation, including sulfation, glucuronidation, and glutathione conjugation, occurs in the liver, where sulfotransferase, UDP-glucuronosyl transferase, and glutathione S-transferase activities make their greatest contribution to drug metabolism (Levy and Yamada, 1971; Slattery and Levy, 1979; Galinsky and Levy, 1981; Vendemiale et al., 1996; Lohr et al., 1998).

In summary, we report that diet-induced models of simple hepatic steatosis and NASH produce multiple changes in efflux transporter expression in rats. Importantly, the MCD dietary model of NASH resulted in decreased biliary excretion of APAP-GLUC, APAP-SULF, and APAP-GSH. Furthermore, the MCD model of NASH altered the vectorial excretion of APAP-GLUC, shifting excretion from bile to blood instead of blood to bile. This shift in vectorial excretion of APAP-GLUC correlated with induction of the basolateral efflux transporters Mrp3 and Mrp4. Thus, the current study provides proof of principle that changes in drug transporter expression during liver disease, i.e., NASH, can result in profound alterations in the drug disposition process. More recent studies also indicate that the therapeutic efficacy of drugs is likewise integrally linked to the function of drug transporters, which can distribute pharmacologically active concentrations of the drug to its target site, e.g., ezetimibe glucuronide and morphine 6-glucuronide (Zelcer et al., 2005; Oswald et al., 2006a,b). Therefore, future studies should place greater emphasis on determining the role of drug transporters in the pharmacokinetics and pharmacodynamics of drug therapy administered during NAFLD.