Materials and Methods

Sample Collection.

Surplus tissue (5–10 mg) available from patients undergoing percutaneous liver biopsy was snap frozen in liquid nitrogen and stored at −70°C until processing. Informed written consent was obtained from each patient, and the study protocol was approved by the St. Vincent's Hospital Human Ethics Committee.

Extraction of RNA.

RNA was extracted from the tissue by homogenization with a Polytron homogenizer in Trizol (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. The extracted RNA was dissolved in 200 μl of water treated with 1% diethyl pyrocarbonate, and contaminating genomic DNA was digested with 1.5 units of DNase I (Promega, Madison, WI). The solution was extracted with Tris-equilibrated phenol, and the RNA precipitated with 2.5 volumes of ethanol. Finally, the RNA was dissolved in 100 μl of water, and aliquots removed for spectroscopic quantitation and agarose gel electrophoresis to assess RNA integrity. Approximately 5 μg of RNA was reverse transcribed in a reaction containing 10 U avian myeloblastosis virus reverse transcriptase, 1 mM dNTPs, 40 U RNasin (Promega), and 0.5 μg of random 10-mer oligonucleotides (GenSet Pacific, Lismore, Australia) for 10 min at room temperature and then for 60 min at 42°C. The resultant cDNA was diluted 10- to 20-fold and stored at −70°C until use.

PCR Primers.

Sequences of the PCR primers used for the detection of UGT isoforms were selected from UGT sequences aligned using DNAMAN software (Lynnon corporation, Quebec, Canada). Primers were made for the detection of UGT isoforms 1A1, 1A3, 1A4, 1A6, 1A9, 2B4, 2B7, 2B10, 2B11, 2B15, and 2B17. In addition, for the detection of all family 1 isoforms, primers were made to regions spanning exons 3 and 4 and, for the detection of all known family 2 isoforms, to regions spanning exons 1 to 3. Primers were also made for the detection of CYP1A2, IL-1β, and TGF-β. Sequences of all primers used in the PCR are shown in Table1.

Real-Time PCR.

PCR was performed on 5 μl of cDNA using 10 pmol of forward and reverse primers (Table 1; Sigma Genosys, Castle Hill, Australia) on a GenAmp model 5700 cycler using SYBR Green detection in a total volume of 20 μl with 10 μl of SYBR Green PCR Master-Mix (Applied Biosystems, Foster City, CA). Cycling conditions were 1 cycle for 2 min at 60°C, 1 cycle at 95°C for 10 min, followed by 40 cycles of a 15-s denaturation at 95°C, and a 1-min annealing/extension at 60°C. The threshold value of detection (C_t) is the cycle at which PCR product rises above an arbitrary number and was set at the exponential phase of amplification. Values were determined from standard curves generated for each gene from serial dilution of genomic or cDNA and normalized to the value of 18S ribosomal RNA measured for each sample. PCR products were of the correct size, as determined by agarose gel electrophoresis. Statistical significance was determined with StatView software (SAS Institute, Inc., Cary, NC) using Fisher's protected least significant difference. Experimental values are the means ± S.E.

Grading of Samples.

Samples were graded (by J.L.S. and P.V.D.) for inflammation or fibrosis according to the Metavir scoring system (Bedossa, 1994), and the mean mRNA levels was determined for each group. To determine the effect of inflammation, samples were graded into three groups according to the inflammation component of their Metavir score (Table2). Groups consisted of samples with low (score 0), medium (score 1), and high inflammation scores (score 2 or 3). To determine the effect of fibrosis, samples were separated into four groups consisting of samples with no fibrosis (score 0), mild fibrosis (score 1), medium fibrosis (score 2), and high fibrosis or established cirrhosis (score 3 or 4; Table 2).

Because insufficient cirrhotic (score 4) or inflammation (score 3) samples were available, samples with a fibrosis score of 3 or 4 or an inflammation score of 2 or 3 were grouped together for statistical purposes. The mean value for samples with a fibrosis or inflammation score of 0 was set at 100%, and means in other groups are expressed relative to this number. Graphs show the mean mRNA levels ± S.E. for each group.

Results

Interindividual Variation.

Most UGT isoforms, with the exception of UGT2B17, were readily detectable. C_t values were between 20 and 30 (Fig. 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Distribution of C_t values.

Distribution C_t values at which measurement of the PCR product was made.

Interindividual variability was estimated by analyzing results from samples that had minor or no inflammatory activity or fibrosis (inflammation and fibrosis scores of 1 or 0). We found a large interindividual variation in enzyme mRNA levels, ranging from 5- to 15-fold for most isoforms. Of the UGT enzymes, isoform 2B17 had the greatest interindividual variability (Fig. 1; Table3).

Effect of Fibrosis.

We found no difference in UGT mRNA levels between the low-score (fibrosis 0) and high-score (fibrosis 3–4) groups using primers designed to amplify all (Fig. 2A) or individual isoforms (Fig. 2B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Effect of fibrosis on mRNA levels.

Samples were separated into four groups according to their fibrosis score. Values are arbitrary units determined by reference to a standard curve generated for each isoform and normalized to the level of 18S ribosomal RNA determined for each sample. The fibrosis samples with a score of 0 are set to 100%. Data show means ± S.E. ∗, indicates p < 0.05. A, levels of UGT, CYP1A2, TGF-β, and IL-1β mRNA. UGT family 1 and family 2 genes were measured using primers designed to amplify all known isoforms. B, levels of individual UGT isoforms.

Levels of mRNA for CYP1A2, an enzyme representative of the cytochrome P450 family, tended to be elevated in fibrosis groups with a score of 1 and 2, and these were significantly higher than levels in the fibrosis groups with a high score (3 to 4). However, levels of the mRNA for TGF-β, a cytokine known to promote matrix deposition (Border and Noble, 1994), tended to be higher in the high-score fibrosis group (Fig. 2A). Levels of the mRNA for the cytokine IL-1β, frequently associated with reduced cytochrome P450 expression (Abdel-Razzak et al., 1995; Calleja et al., 1998; Kudo and Kawano, 1999) showed no apparent trend.

Effect of Inflammation.

UGT family 1A and family 2B mRNA levels as measured by primers designed to amplify all isoforms within each family tended to be lower in high-inflammation score samples (Fig.3A). Measurement of individual isoforms using specific primers showed a consistent downward trend in the levels of most UGT isoforms in high-inflammation score samples, reaching statistical significance for isoforms 1A4, 2B4, and 2B7 (Fig. 3B), which were reduced to 43.0 ± 12.3, 39.6 ± 19.5, and 38.4 ± 20.3% of control levels, respectively.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Effect of inflammation on mRNA levels.

Samples were separated into three groups according to their inflammation score. Values are arbitrary units determined by reference to a standard curve generated for each isoform and normalized to the level of 18S ribosomal RNA determined for each sample. The inflammation samples with a score of 0 are set to 100%. Data show means ± S.E. ∗, indicates p < 0.05. A, levels of UGT, CYP1A2, TGF-β, and IL-1β mRNA. UGT family 1 and family 2 genes were measured using primers designed to amplify all known isoforms. B, levels of individual UGT isoforms.

High-score inflammation samples had lower CYP1A2 and higher IL-1β mRNA levels than samples with a score of 1 (Fig. 3A). TGF-β mRNA levels in the group with score from 2 to 3 were similar to those with a score of 0.

Discussion

Although it is commonly accepted that liver disease results in a decreased activity of the cytochrome P450 system, the effect on glucuronidation is controversial (Hoyumpa and Schenker, 1991). Glucuronidation has commonly been determined from either pharmacokinetic studies in humans or experimental models of liver disease in animals. UGT isoforms have overlapping specificities, and a metabolite may be the product of several isoforms. Furthermore, extra hepatic metabolism may be significant (McGurk et al., 1998; Nowell et al., 1999), hepatic blood flow may be impaired, or the binding of drug to plasma proteins may be affected (Davies and Skjodt, 2000). The direct measurement of UGT mRNA in hepatocytes overcomes limitations and assumptions of the pharmacokinetic approach. A number of studies in animals and hepatic cell lines (Vargas et al., 1998; Kardon et al., 2000; Walle et al., 2000; Luquita et al., 2001) and in humans (Ritter et al., 1999) indicate that hepatic UGT protein and/or activity, for at least some isoforms, in general follow the mRNA levels.

In these studies, we used real-time PCR to examine the effect of inflammation and fibrosis on the levels of individual isoforms of the UGT family in patients with liver disease. We found that a reduction in the expression of UGT mRNA in our samples was associated with the degree of inflammation. There was no correlation with the level of fibrosis. In addition our data indicates that the large interindividual variability in drug metabolism, which is often observed across patients, may at least partly be a result of differential mRNA levels of these enzymes. This may be a result of genetic factors or even induction by dietary components.

In a number of studies in human liver (Debinsky et al., 1995) or animal models of cirrhosis (Debinsky et al., 1996), results have indicated that levels of the glucuronidating enzymes as measured by immuno or hybridization histochemistry are maintained or elevated when compared with control tissue. Our results support findings that glucuronidation is unaffected in the fibrotic liver and that hepatic UGT mRNA levels are not reduced.

The mechanism by which the inflammatory process may repress these enzymes is complex and poorly understood (Morgan et al., 1998). The reduction in UGT mRNA in our high-score inflammation samples may be the result of substances released by resident or infiltrating inflammatory cells, as evidenced by the increase in interleukin-1β mRNA in the high-score inflammation group. A number of studies have found that cytokines can affect the expression of UGT (Monshouwer et al., 1996;Levesque et al., 1998; Strasser et al., 1998) or cytochrome P450 enzymes (Ferrari et al., 1993; Abdel-Razzak et al., 1995; Monshouwer et al., 1996; Sindhu et al., 1996; Guillen et al., 1998; Levesque et al., 1998; Kudo and Kawano, 1999; Monshouwer and Witkamp, 2000). There is evidence for the involvement of nuclear factor-κB in the repression of the rat CYP2C11 gene in response to bacterial lipopolysaccharide or interleukin-1β, but interleukin-6 induced repression occurs by an alternative mechanism (Iber et al., 2000). In addition, because our data are standardized to 18S ribosomal RNA, some reduction in hepatic UGT mRNA is expected as a result of an increase in nonhepatic 18S RNA by infiltrating inflammatory cells.

Because of the consistent downward trend observed across most isoforms, the relatively small number of samples with severe liver disease, and the large variation, more comprehensive studies will need to be done to establish the precise relationship between inflammatory liver disease and the expression of individual UGT isoforms. However, the reduction in mRNA levels reached statistical significance for three of the isoforms that we measured, indicating that inflammatory liver disease may have differential effects on the various UGT isoforms.

Our data indicate that there is the potential for decreased drug glucuronidation in the liver of patients with an active inflammatory condition and some hepatic fibrosis, but glucuronidation may recover once the inflammation subsides. This may help explain the uncertainty surrounding drug glucuronidation in liver disease using pharmacokinetic approaches in which experimental results may be from pools of patients with mixed inflammatory/cirrhotic conditions.