Materials and Methods

Human Liver Specimens.

Samples of frozen and formalin-fixed, paraffin-embedded adult explant livers [normal, n = 20; steatosis, n = 11; NASH (fatty liver), n = 10; NASH (no longer fatty), n = 11] were obtained from the Liver Tissue Cell Distribution System at the University of Minnesota, Virginia Commonwealth University, and the University of Pennsylvania. Histological slides were diagnosed using criteria from a scoring system for human NAFLD (Kleiner et al., 2005). Steatotic liver was diagnosed when >10% of hepatocytes showed fat deposition. NASH with fatty liver diagnosis was defined as having marked inflammation, fibrosis, and >5% of hepatocytes with fat deposition. NASH no longer fatty liver was diagnosed by marked inflammation, fibrosis, and <5% of hepatocytes with fat deposition. Diagnosis was first established by a Liver Tissue Cell Distribution System medical pathologist and confirmed by histological examination at the University of Arizona in a blinded fashion. Information on donors, including age and gender, can be seen in Supplemental Data Table 1.

Total RNA Isolation and Reverse Transcription.

Approximately 50 mg of each human liver sample was homogenized in 3 ml of Nucleic Acid Purification Lysis Solution (Applied Biosystems, Foster City, CA). Total RNA was isolated from each sample using the Applied Biosystems 6100 Nucleic Acid Prepstation. For reverse transcription, approximately 200 ng of total RNA from each sample was converted to cDNA following the manufacturer protocol for the Applied Biosystems High Capacity cDNA Archive Kit.

Quantitative Reverse Transcription-Polymerase Chain Reaction (TaqMan) Analysis.

CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4 cDNA from human liver samples was analyzed using gene-specific TaqMan primer/probe sets (Applied Biosystems). Reactions with the specific primer/probes for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were analyzed as an endogenous control for P450 expression. Amplifications were performed on an ABI 7900HT real-time polymerase chain reaction system (Applied Biosystems) in relative quantification mode for 40 amplification cycles using standard conditions for TaqMan-based assays. Threshold cycle (C_T) determinations were performed by the ABI 7900HT system software for both P450 and GAPDH gene. Relative-fold mRNA content was determined for each sample relative to the endogenous control gene expression (GAPDH) using the relationship: Relative-fold mRNA Content = 2^−ΔΔCt.

Human Liver Microsome Isolation.

Human liver samples (∼300 mg) were homogenized in 3 ml of buffer A (50.00 mM Tris-HCl, 1.00 mM EDTA, and 154.00 mM KCl, pH 7.4) with a Dounce homogenizer. Homogenate was centrifuged at 10,000g for 30 min at 4°C, and the supernatant was collected. After centrifugation at 100,000g for 60 min at 4°C, the supernatant was discarded, and the pellet was resuspended in 600 μl of buffer B (100.0 mM sodium pyrophosphate and 0.1 mM EDTA, pH 7.4). Samples in buffer B were centrifuged at 100,000g for 60 min at 4°C. The supernatant was resuspended in 300 μl of buffer C (10.0 mM KPO₄, 1.0 mM EDTA, and 20.0% glycerol, pH 7.4) and stored at −80°C until analysis.

Western Blot Analysis of Microsomal P450 Levels.

Microsomal protein concentrations were determined using a Bio-Rad Protein Assay Reagent Kit (Bio-Rad Laboratories, Hercules, CA) as described by the manufacturer. Microsomal protein levels of P450s and GAPDH (loading control) were determined using a mouse monoclonal antibody specific for human CYP1A2 (Abcam Inc., Cambridge, MA), polyclonal rabbit anti-human CYP2A6, CYP2B6, CYP2C8, CYP2D6, CYP3A4, and CYP2E1 (XenoTech LLC, Lenexa, KS), CYP2C9 and CYP2C19 (Fitzgerald Industries International Inc., Concord, MA), and a monoclonal rabbit anti-GAPDH antibody (Cell Signaling Technology, Inc., Danvers, MA). Microsomes (10 μg/well) or 10 μg of respective recombinant human P450 protein (BD Gentest from BD Biosciences, San Jose, CA) was separated by SDS-polyacrylamide gel electrophoresis as previously reported (Augustine et al., 2008). Quantification of relative protein expression was determined using image processing and analysis with Image J in JAVA (National Institutes of Health, Bethesda, MD) and normalized to respective GAPDH protein expression.

P450 Enzymatic Activity Determination.

Microsomal activities for human CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4/5 were determined using specific marker substrates according to established procedures listed in Supplemental Data Table 2. Human liver microsomes (∼0.01–0.04 mg/ml) incubated at 37°C in potassium phosphate buffer (100 mM, pH 7.4), NADPH (1 mM), and substrate were added to each incubation in methanol or acetonitrile so that the final solvent concentration was 0.1%. Reactions were started by addition of NADPH and stopped after indicated time points by addition of organic solvent. The amount of product formed was quantified using validated liquid chromatography/tandem mass spectrometry methodologies. In each analytical run, at least six calibration standards and 12 quality control samples (at three different concentration levels) were used to ensure the quality of the analytical run.

Statistics.

P450 and other expression levels are continuous outcomes with skewed distribution. This asymmetry suggests that median rather than mean values should be compared. Therefore, graphs in this manuscript show box-whisker plots rather than mean values with error bars. The disease groups examined in the current study can be ordered by their severity: normal < steatosis < NASH with fatty liver < NASH no longer fatty. We initially performed two-sample comparisons between each disease state and normal using the Wilcoxon rank-sum test. This is the appropriate alternative to the two-sample t test when skewed continuous outcomes exist in samples of modest size. For comparing all the groups together, we performed a nonparametric test for trend (Cuzick, 1985) rather than using analysis of variance, which assumes unordered categories. Thus, instead of simply looking for differences between categories, we used the more appropriate and logical approach of testing whether expression levels increased or decreased consistently over the ordered disease states. Under Results, NAFLD progression refers to increasing severity of disease state. A significance level of 0.05 was used.

Immunohistochemical Staining of Paraffin-Embedded Liver Samples.

Formalin-fixed sections of paraffin-embedded livers were deparaffinized in xylenes and rehydrated through a graded alcohol series. Antigen retrieval was performed by incubating slides in citrate buffer (10 mM) for 10 min in a Kenmore 1200-W microwave set (Sears Roebuck and Co.) on defrost, and endogenous peroxidase activity was blocked with 3% (v/v) H₂O₂ for 10 min at room temperature. Deparaffinized sections were incubated overnight with either a rabbit polyclonal interleukin 1β (IL-1β) antibody (H-153; Santa Cruz Biotechnology, Inc., Santa Cruz, CA), mouse monoclonal tumor necrosis factor α (TNFα), or mouse monoclonal hypoxia-induced factor 1α (HIF-1α) (Abcam Inc.) diluted 1:50 in phosphate-buffered saline. Protein-antibody complexes were visualized using the Vectastain Elite ABC kit and developed with 3,3′-diaminobenzidine as per manufacturer's protocol (Vector Laboratories, Burlingame, CA). Negative control staining of human liver sections was performed by incubating without primary antibody.

Results

Histopathology of Human Livers with Progressive Stages of NAFLD.

Hematoxylin and eosin staining of donor livers was used to assess the severity of NAFLD. Representative images of hematoxylin and eosin-stained livers from normal, steatotic, NASH with fatty liver, and NASH no longer fatty can be seen in Fig. 1.

Hepatic P450 mRNA Expression during NAFLD Progression.

There were decreasing trends of CYP1A2 and CYP2C19 mRNA expression associated with progressive stages of NAFLD (p values of 0.225 and 0.193, respectively); however, these trends were not statistically significant (Fig. 2). CYP2E1 mRNA expression showed a statistically significant decreasing trend (p = 0.001) with NAFLD progression. Conversely, CYP2C9 mRNA expression tended to increase with NAFLD progression but did not reach statistical significance (p = 0.220). Likewise, CYP2A6 and CYP2B6 mRNA expression significantly increased with NAFLD progression, with p values of 0.002 and 0.003, respectively. NAFLD progression had little effect on CYP2C8, CYP2D6, or CYP3A4 mRNA expression levels.

Microsomal P450 Protein Expression in Progressive Stages of NAFLD.

Representative Western blots of microsomal CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9/19, CYP2D6, CYP2E1, CYP3A4, and GAPDH are shown in Fig. 3. In addition, relative protein expression of P450s for all the donor samples was determined by densitometry and normalized to GAPDH expression and is seen in Fig. 4. Similar to mRNA expression, CYP2A6 protein expression was significantly increased with NAFLD progression (p = 0.019). CYP2C8, CYP2D6, and CYP3A4 protein expression tended to decrease with progression of NAFLD (p values of 0.191, 0.068, and 0.112, respectively); however, this trend was not statistically significant. CYP1A2 (p = 0.0001), CYP2C19 (p = 0.01), and CYP2E1 (p = 0.01) protein levels significantly decreased with NAFLD progression.

Microsomal P450 Enzyme Activity during NAFLD Progression.

Phenacetin O-dealkylation by CYP1A2 significantly decreased (p = 0.001) as the severity of NAFLD increased (Fig. 5). Similar to CYP1A2, there was a decreased rate of CYP2C19-mediated mephenytoin 4′-hydroxylation with NAFLD progression (p = 0.05). CYP2D6 and CYP3A4 activity toward dextromethorphan and testosterone, respectively, also displayed a decreasing trend with NAFLD progression (p values of 0.062 and 0.18), but these trends did not reach statistical significance. In contrast, CYP2A6 hydroxylation of coumarin was significantly increased with NAFLD progression (p = 0.02). Finally, the enzymatic activity of CYP2C9 was determined using two specific substrates of this enzyme. CYP2C9 enzyme activity, determined by diclofenac 4′-hydroxylase and hydroxytolbutamide metabolite formation, was significantly increased with NAFLD progression, with p values of 0.0001 and 0.004, respectively.

Results of Two-Group Comparisons between Each Disease State and Normal.

The rank-sum tests did not reveal statistically significant differences between each disease state considered separately versus normal. Given the modest sample size and the high degree of variability observed in the outcomes, this was not unexpected. However, separate consideration of each disease state discards important information available from the inherent ordering of the disease states. Statistical analysis for trends across ordered categories has the greater power to detect systematic differences than two-sample tests. Thus, the balance of the analyses focused on the use of a nonparametric trend test (Cuzick, 1985) to detect such systematic changes in outcome as a function of NAFLD progression.

Immunohistochemical Staining of HIF-1α during NAFLD Progression.

To determine whether NAFLD induces hypoxia, immunohistochemical staining of donor livers from normal and progressive stages of NAFLD was used to identify expression of known markers, specifically, HIF-1α. Although staining was not observed in normal livers (Fig. 6), and only moderate staining was observed in steatotic livers, there was pronounced HIF-1α expression in the cytosol of NASH fatty liver samples and both cytosolic and nuclear staining in NASH no longer fatty liver samples, suggesting that hypoxia occurs in the later stages of NAFLD.

Immunohistochemical Staining of Proinflammatory Cytokines in Progressive Stages of NAFLD.

Little to no cytokine staining was observed in normal or steatotic liver tissue (Fig. 6). However, there was marked increased expression of TNFα and IL-1β in both stages of NASH, strongly suggesting the presence of inflammation in these stages of NAFLD.

Discussion

P450s have been shown to be particularly susceptible to alterations in expression and activity (Frye et al., 2006). Decreased P450 enzymatic activity can potentially lead to reduced metabolism of therapeutics, ultimately leading to increased bioavailability and possible toxicity. Conversely, increased activity of hepatic P450s present the potential to increase the metabolism of known substrates, thereby decreasing their pharmacotherapeutic effect or increasing the generation of reactive metabolites and oxidative stress. The aim of the current study was to determine whether expression and function of the major drug-metabolizing P450s are altered in human livers diagnosed with progressive stages of NAFLD. To our knowledge, this is the first report of P450 enzyme expression and activity in progressive stages of human NAFLD.

Previous studies have reported up to a 50% decrease in hepatic CYP1A2 protein levels in cirrhotic liver patients when compared with normal liver (George et al., 1995b; Congiu et al., 2002). Guengerich and Turvy (1991) noted similar findings in CYP1A2 immunohistochemical staining of livers with sclerosing cholangitis and cirrhosis. CYP1A2 metabolic activity has also been shown to be decreased in primary biliary cirrhosis, alcoholic steatohepatitis, and cirrhotic patients as seen by reduced clearance of known substrates antipyrine, theophylline, and caffeine (Bechtel et al., 2000; Lelouët et al., 2001; Villeneuve and Pichette, 2004). Although we report only a slight downward trend in mRNA expression of CYP1A2, the protein (p = 0.0001) and enzyme activity levels (p = 0.001) were significantly decreased with NAFLD progression. CYP1A2 has been reported to be significantly decreased in the presence of proinflammatory cytokines TNFα and IL-1β (Zhou et al., 2008) and may explain decreased expression and function in the current study.

CYP2A6 plays a role in the metabolism of several clinically relevant drugs, including halothane, disulfiram, and valproic acid (Raunio et al., 2001). In the current study, we show that mRNA (p = 0.002), protein (0.019), and enzyme activity (p = 0.020) of CYP2A6 increased with progressive stages of NAFLD. Significantly elevated levels of CYP2A6 enzymatic activity have been reported in patients with hepatitis, primary biliary cirrhosis, and alcoholic cirrhosis (Kirby et al., 1996; Bechtel et al., 2000; Lelouët et al., 2001). In addition, induction of CYP2A5, the mouse ortholog of human CYP2A6, has been shown to be induced during oxidative injury to the endoplasmic reticulum, as well as during altered redox status (Gilmore et al., 2003; Nichols and Kirby, 2008). It is well documented that oxidative stress occurs in NAFLD patients, as shown by NASH patients who show significantly increased systemic levels of lipid peroxidation products (Chalasani et al., 2004; Videla et al., 2004). Therefore, it is possible that oxidative stress induced during NAFLD plays a role in the increased CYP2A6 expression and activity reported in this study.

CYP2C9 is generally accepted as the second-most abundant P450 in the human liver and is responsible for the metabolism of a number of clinically relevant drugs substrates, including S-warfarin, losartan, rosiglitazone, fluoxetine, and tamoxifen (Danielson, 2002). Hepatic CYP2C9 mRNA expression showed an increasing trend with progressive stages of NAFLD (p = 0.220); however, there was little overall change in protein expression between samples. Nevertheless, CYP2C9 metabolism of diclofenac was significantly increased with the severity of NAFLD (p = 0.0001). To verify these results, CYP2C9 enzymatic activity was also determined using a second high affinity substrate, tolbutamide. Similar to diclofenac, hydroxytolbutamide formation by CYP2C9 significantly increased with progressive states of NAFLD (p = 0.004).

Several studies have shown that CYP2C9 activity is increased during hypoxia, potentiating metabolism of arachidonic acid into 11,12-epoxyeicosatrienoic acid (Michaelis et al., 2005). 11,12-Epoxyeicosatrienoic acid in turn attenuates vascular smooth muscle cell hyperpolarization and resultant vasoconstriction during acute and chronic hypoxic conditions (Earley et al., 2003). Although no data are available with regard to hypoxia in cases of NAFLD, an experimental model of ethanol-induced steatohepatitis in rats showed a significant increase in HIF-1α expression in hepatocytes (Li et al., 2006a). In an attempt to explain the observed increase in CYP2C9 activity, we investigated the possibility of hypoxia occurring during NAFLD progression. Figure 6 shows increased cytosolic expression of HIF-1α in NASH with fatty liver samples, whereas both cytosolic expression and nuclear accumulation of HIF-1α was observed in NASH no longer fatty samples. The increased expression and nuclear localization of HIF-1α suggest that hypoxia occurs in the later stages of NAFLD and provides a plausible mechanism for the elevated CYP2C9 activity reported in the current study.

CYP2C19 has been identified as one of the more sensitive P450s to the presence of liver diseases such as hepatocellular carcinoma, hepatitis C, and chronic hepatitis and cirrhosis (Ohnishi et al., 2005; Frye et al., 2006). In addition, it has also been shown to be affected earlier than the other important drug-metabolizing P450s (Villeneuve and Pichette, 2004). Results in the current study support these observations as the protein expression (p = 0.010) and enzymatic activity (p = 0.05) showed statistically significant decreases with progressive states of NAFLD.

Hepatic CYP2C8 mRNA and enzyme activity appeared unaffected by different stages of the disease. It is interesting to note that although CYP2C8 expression and activity remained relatively constant with NAFLD progression, two other relevant members of this subfamily were alternately affected by NAFLD, with increased CYP2C9 activity and decreased CYP2C19 activity. The alternative regulation of CYP2C enzymes observed in this study have not been reported in other liver diseases and warrant further investigation to determine whether these changes are specific to NAFLD.

In addition to CYP1A2 and CYP2C19, CYP3A4 regulation and expression have been reported to be particularly sensitive to other liver diseases; however, its activity can selectively change with different disease states (Villeneuve and Pichette, 2004). A study by Yang et al. (2003) reported that CYP3A4 activity was significantly altered in patients with cirrhosis yet remained unchanged by obstructive jaundice. In the current study CYP3A4 mRNA expression was not different between disease groups, but CYP3A4 protein expression and activity showed decreasing trends with respect to the severity of NAFLD (p = 0.112 and p = 0.180, respectively). Although those trends did not reach statistical significance, they do seem to indicate that CYP3A4 expression and activity may be decreased with the progression of NAFLD.

CYP2E1 mRNA and protein expression significantly decreased with NAFLD progression, whereas CYP2E1 metabolism of chlorzoxazone was unaltered by NAFLD progression. The finding that CYP2E1 mRNA and protein decrease during NAFLD conflicts with previously published data, which noted increases in livers from patients with NAFLD (Weltman et al., 1998; Chtioui et al., 2007).

In the current study we report that a number of the major hepatic drug-metabolizing P450s are differentially regulated in progressive stages of NAFLD. The expression and activity of CYP1A2, CYP2C19, CYP2D6, and CYP3A4 tended to decrease with increasing severity of NAFLD. However, CYP2A6 and CYP2C9 enzyme activity was significantly increased with progressive stages of NAFLD. With the incidence of NAFLD increasing at an alarming rate, the effect of this disease on major drug-metabolizing enzymes is of critical importance. The current study offers a comprehensive analysis of the major hepatic P450 expression and activity in three progressive stages of NAFLD and may provide a valuable framework for physicians when determining the pharmacotherapeutic options and dosing regimens to patients with this disease.