Genetic polymorphisms influencing xenobiotic metabolism and transport in patients with primary biliary cirrhosis

Authors


  • Conflict of interest: Nothing to report.

Abstract

Epidemiological data suggest that environmental factors may trigger autoimmunity in genetically susceptible individuals. In primary biliary cirrhosis (PBC), it has been postulated that halogenated xenobiotics can modify self-molecules, facilitating the breakdown of tolerance to mitochondrial antigens. The transport and metabolism of xenobiotics is highly dependent on key genetic polymorphisms that alter enzymatic phenotype. We analyzed genomic DNA from 169 patients with PBC and 225 geographically and sex-matched healthy subjects for polymorphisms of genes coding for cytochromes P450 (CYPs) 2D6 (CYP2D6*4, CYP2D6*3, CYP2D6*5, and CYP2D6*6) and 2E1 (c1/c2), multidrug resistance 1 (MDR1 C3435T) P-glycoprotein, and pregnane X receptor (PXR C-25385T, C8055T, and A7635G). We compared the genotype frequencies in patients and controls and also correlated polymorphisms with PBC severity. The distributions of the studied genotypes did not significantly differ between patients and controls. However, when clinical characteristics of patients with PBC were compared according to genotype, the CYP2E1 c2 allele was associated with signs of more severe disease. In conclusion, genetic polymorphisms of CYP 2D6 and 2E1, PXR, and MDR1 do not appear to play a role in the onset of PBC. (HEPATOLOGY 2005;41:55–63.)

The etiology of primary biliary cirrhosis (PBC) remains elusive, but recent data suggest that the breaking of tolerance to the highly conserved 2-oxoacid dehydrogenase complex lipoylated domains of the mitochondrial autoantigen could result from molecular mimicry initiated by an immune response directed toward xenobiotic structural analogues of lipoic acid.1 For example, sera from patients with PBC present antimitochondrial antibodies (AMAs) recognizing a number of synthetic structures that mimic a xenobiotic-modified lipoyl hapten conjugated to a peptide from the E2 subunit of the pyruvate dehydrogenase complex.2 In addition, rabbits immunized with a xenobiotic (6-bromohexanoate) bovine serum albumin conjugate produce immunoglobulin G autoantibodies that react not only with xenobiotic but also self-reactive AMAs.3 Although many genetic factors conferring susceptibility to PBC have been suggested in population and family studies,4, 5 no definitive genetic association with the onset of the disease or its outcome has yet been found. The liver is the primary organ involved in the metabolism and disposition of foreign chemicals. In such an environment, chemicals and/or their reactive metabolites may modify cellular proteins to form neoantigens.6 The recent findings of a possible role of molecular mimicry prompted us to determine whether polymorphisms in the genes involved in xenobiotic metabolism could contribute to the pathogenesis of PBC by genotyping a large population of patients with PBC and controls searching for a range of single nucleotide polymorphisms (SNPs). In particular, we decided to concentrate our efforts on mechanisms controlling the absorption of xenobiotics from the intestinal lumen, as well as their secretion into the bile (multidrug resistance 1 [MDR1]). Moreover, we were interested in analyzing the enzymes directly responsible for the metabolism of exogenous compounds (cytochromes P450 [CYPs]) or influencing the activity of the latter enzymes through regulation of their transcription (pregnane X receptor [PXR]).

Within the group of CYPs, CYP2D6 (debrisoquine/sparteine hydroxylase) is involved in the metabolism of approximately 20% of drugs.7 Several coding genetic polymorphisms have been identified that are associated with significant reduction of drug metabolism rates in vivo and in vitro; these are known as “poor metabolizers” (PMs).8 In particular, the PM phenotype is found in as many as 10% of Caucasian subjects9 and is most commonly due to the presence of null alleles for single base pair mutations (CYP2D6*3, CYP2D6*4, and CYP2D6*6, among others) or deletion of the whole gene (allele CYP2D6*5).7 As such, four alleles—CYP2D6*3, CYP2D6*4, CYP2D6*5, and CYP2D6*6—account for 93% to 97% of the PM phenotypes in Caucasians.10

CYP2E1 metabolizes several compounds, including ethanol, estrogenic metabolites, and halothane, and its activity is altered by nicotine.11 These characteristics make this enzyme particularly interesting in PBC because of a striking female predominance.12 Furthermore, PBC shares several characteristics (including the presence of serum AMAs) with halothane-induced hepatitis,13 and is diagnosed more commonly among smokers.14 A specific genetic polymorphism of CYP2E1 (RsaI restriction polymorphisms alleles c1/c2) has been widely investigated15 and is associated with reduced activity per se16 or altered phenotype following interaction with specific compounds (e.g., ethanol or isoniazid), despite similar baseline activity.17, 18

The MDR1 gene encodes for the P-glycoprotein (PGP), a molecule that controls the cellular trafficking of substrates such as bilirubin and cancer drugs. In particular, PGP plays a role in excreting toxic xenobiotics and metabolites into the intestinal lumen as well as urine and bile. Although multiple mutations have been identified in MDR1, the exon 26 C3435T SNP is of special interest because of its association with a lower PGP expression in the intestine.19 It has also been suggested that the C3435T SNP might be associated with susceptibility to ulcerative colitis,20 although the latter observation was not confirmed in another study.21

PXR is a nuclear receptor for steroid hormones and select xenobiotics whose activity regulates the expression of CYP3A4 and MDR1 in the liver and intestine. Importantly, there is an association of specific PXR SNPs with the CYP3A4 phenotype.22 We report the prevalence of CYP2D6*4, CYP2E1 c1/c2, MDR1 C3435T, and PXR (C-25385T, C8055T, A7635G) polymorphisms in patients with PBC and controls and have identified a correlation between CYP2E1 c1/c2 genotype and disease severity.

Abbreviations

PBC, primary biliary cirrhosis; CYP, cytochrome P450; MDR, multidrug resistance; PXR, pregnane X receptor; AMA, antimitochondrial antibody; SNP, single nucleotide polymorphism; PM, poor metabolizer; PGP, P-glycoprotein; PCR, polymerase chain reaction.

Patients and Methods

Patients.

A total of 169 Italian patients with PBC who attended the Liver Unit at San Paolo Hospital (Milan, Italy) were enrolled in the study (Table 1). The diagnosis of PBC was based on internationally accepted criteria,23, 24 and the AMA status of each patient was verified via indirect immunofluorescence. Of these 169 patients, 30 (18%) were AMA negative and 139 (82%) were AMA positive. All patients were negative for hepatitis B surface antigen and antibodies to hepatitis C virus and denied alcohol abuse during the previous 12 months. The disease duration was calculated as the time between the date of the earliest recorded evidence of liver disease and the date of blood sampling. The latter evidence was determined through an extensive search of all laboratory data for alterations in cholestasis indicators (alkaline phosphatase > 1.5 normal values with or without altered γ-glutamyltransferase). All patients had undergone liver biopsy during the 12 months before blood sampling. Patients who did not have fibrosis according to liver histology (i.e., Stages I-II according to Ludwig et al.25) were considered to have early-stage PBC. Patients with liver fibrosis or cirrhosis (i.e., Stages III-IV according to Ludwig et al.25), patients who had a history of a major complication from cirrhosis (e.g., ascites or gastrointestinal bleeding caused by portal hypertension), and patients who had undergone orthotopic liver transplantation for PBC were considered to have advanced-stage PBC. Advanced disease was found in 92 individuals (54.4%). Based on age, serum bilirubin and albumin, prothrombin time, and the presence of ascites, the Mayo score, the only validated prognostic index in PBC,26 was calculated at the time of blood sampling or orthotopic liver transplantation. Two hundred twenty-five healthy subjects (blood donors) geographically and sex-matched with patients with PBC were used as a control population. Namely, 4 matched controls were obtained for every 3 patients. The study protocol followed the ethical guidelines of the most recent Declaration of Helsinki (Edinburgh, 2000); all patients provided written informed consent.

Table 1. Characteristics of Patients With PBC at Time of Enrollment
 All Patients (n = 169)PBC
Early Disease (n = 77)Advanced Disease (n = 92)P Value
  1. NOTE. Continuous variables are expressed as the mean ± SD.

  2. Abbreviations: NS, not significant; INR, international normalized ratio; n.v., normal value.

Female sex (n)153 (91%)68 (88%)85 (92%)NS
Age at enrollment (yr)62 ± 1258 ± 1266 ± 11< .001
Disease duration (mo)124 ± 7290 ± 51152 ± 75< .001
AMA-positive (n)139 (82%)61 (79%)78 (85%)NS
Total bilirubin (mg/dL) (n.v. < 1.0)1.6 ± 3.20.7 ± 0.352.4 ± 4.2< .001
Albumin (g/dL) (n.v. >3.5)4.2 ± 0.94.9 ± 0.73.7 ± 0.7< .001
Prothrombin time (INR) (n.v. < 1.2)1.04 ± .0170.99 ± 0.081.08 ± 0.21.001
With ascites (n)22 (13%)022 (24%)< .001
Mayo score5.7 ± 1.44.7 ± .076.5 ± 1.4< .001

Genotyping.

Whole blood samples were obtained from each patient and control and stored at −20°C before DNA extraction. DNA extraction was performed using a commercially available kit (Instagene Matrix; Bio-Rad Laboratories, Segrate, Italy). SNP genotypes were determined with polymerase chain reaction (PCR)/restriction fragment length polymorphism or the TaqMan SNP detection system (Applied Biosystems, Foster City, CA).

CYP2D6.

CYP2D6*3 (2549A deletion) and CYP2D6*6 (1707T deletion) variants were analyzed using TaqMan-based methods. Amplification was performed using TaqMan universal master mix (Applied Biosystems) and 40X primers (forward 5′-CCTGACCCAGCTGGATGAG-3′, reverse 5′GCCAGGAAGGCCTAGT-3′ for CYP2D6*3 and forward 5′-GGGCCTGGGCAAGAAGTC-3′, reverse 5′-CGAAGGCGGCACAAAGG-3′ for CYP2D6*6) and allele-specific probes (VIC-ACTGAGCACAGGATGA-8NFQ, FAM-TAACTGAGCACGGATGA-8NFQ for CYP2D6*3 and VIC-CACCCACTGCTCCAG-8NFQ, FAM-TCACCCCTGCTCCAG-8NFQ for CYP2D6*6). PCR amplification included one cycle at 95°C for 10 minutes followed by 40 cycles of 92°C for 15 seconds and 65°C for 1 minute. Allelic discrimination was performed on the post-PCR product by the 7900HT Sequence Detection System (Applied Biosystems). The presence of homozygosity for the CYP2D6*5 allele (deletion of the whole gene) was determined as described previously.27

The CYP2D6 G1934A SNP (CYP2D6*4) was investigated as described by Brown et al.8 Briefly, using primer pairs 5′-GGTGTTCCTCGCGCGCTATG-3′ and 5′-CTCGGTCTCTCGCTCCGCAC-3′, DNA amplifications were performed with the PCR System 9700 (Applied Biosystems). PCR amplification consisted of an initial denaturation for 5 minutes at 94°C followed by 35 cycles of denaturation at 94°C for 1 minute, annealing at 60°C for 1 minute, and extension at 72°C for 1 minute. The terminal elongation was performed at 72°C for 7 minutes. PCR products were digested with restriction enzyme BstNI (New England BioLab, Beverly, MA) overnight at 60°C, resolved in 3% agarose in Tris-acetate buffer, and visualized using ultraviolet ethidium bromide staining. Based on the pattern of the detected bands, restriction fragment patterns were scored as CYP2D6*4 homozygous (77 and 344 bp), non-CYP2D6*4 homozygous (77, 161, and 183 bp), or heterozygous (Fig. 1).

Figure 1.

Restriction fragment length polymorphism analysis of (A) CYP2D6 G1934A, (B) CYP2E1 c1/c2, and (C) MDR1 C3435T SNPs. PCR products of CYP2D6, CYP2E1, and MDR1 were digested with restriction enzymes BstNI, RsaI, and Sau3AI, respectively, and resolved via agarose gel electrophoresis. Restriction fragment patterns were scored as: (A) non-CYP2D6*4 homozygotes (77, 161, and 183 bp) in lane 1, heterozygous in lane 2, or CYP2D6*4 homozygous (77 and 344 bp) in lane 3; (B) CYP2E1 wild-type homozygous c1/c1 (360 and 50 bp) in lane 1, heterozygous c1/c2 in lane 2, or mutant-type homozygous c2/c2 (410 bp) in lane 3; and (C) MDR1 homozygous TT (197 bp) in lane 1, heterozygous CT in lane 2, or homozygous CC (158 and 39 bp) in lane 3. CYP, cytochrome P450; MDR, multidrug resistance.

CYP2E1 c1/c2.

CYP2E1 c1/c2 SNP was investigated as described by Choi et al.28 Briefly, using primer pairs 5′-CCAGTCGAGTCTACATTGTCA-3′ and 5′-TTCATTCTGTCTTCTAACTGG-3′, DNA amplification was performed with an initial denaturation of 4 minutes at 94°C followed by 34 cycles at 94°C for 60 seconds, 60°C for 60 seconds, and 72°C for 60 seconds. The terminal elongation was performed at 72°C for 4 minutes. PCR products were digested with restriction enzyme RsaI (Invitrogen, Carlsbad, CA) at 37°C for 3 hours, resolved on 3% agarose gels in Tris-acetate buffer, and visualized with ethidium bromide staining. Based on the size of detected bands, samples were identified as either wild-type homozygous c1/c1 (360 and 50 bp), mutant-type homozygous c2/c2 (410 bp), or heterozygous c1/c2 (see Fig. 1).

MDR1 C3435T.

The MDR1 C3435T SNP was determined as described by Cascorbi et al.29 Briefly, using primer pairs 5′-TGTTTTCAGCTGCTTGATGG-3′ and 5′-AAGGCATGTATGTTGGCCTC-3′, DNA amplification was performed with an initial denaturation of 2 minutes at 94°C followed by 35 cycles of 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds. The terminal elongation was performed at 72°C for 7 minutes. PCR products were digested with restriction enzyme Sau3AI (New England BioLab) overnight at 37°C, resolved on 3.8% agarose gels in Tris-acetate buffer, and visualized with ethidium bromide staining. Based on the size of the detected bands, samples were identified as C/C homozygous (158 and 39 bp), T/T homozygous (197 bp), or heterozygous (Fig. 1).

PXR.

Three PXR SNPs—C-25385 T, A7635G, and C8055T—were analyzed using a TaqMan-based method. Using the TaqMan universal master mix and 40× assay mix including primers and probes 5′-ACCACGATTGAGCAAACAGGTA-3′, 5′-ACCTGAAGACAACTGTGGTCATT-3′, VIC-TCCCAGGTTCTCTTTT-8NFQ, and FAM-TCCCAGGTTTTCTTTT-8NFQ (Assays-by-Design; Applied Biosystems), the PXR C-25385T SNP was determined. PCR amplification consisted of one cycle at 95°C for 10 minutes followed by 40 cycles of 92°C for 15 seconds and 65°C for 1 minute. Allelic discrimination was performed on the post-PCR product using the 7900HT Sequence Detection System (Applied Biosystems). Potentially, there were four clusters of points that corresponded to the three genotypes (CC, CT, TT) or to no amplification. Similarly, PXR A7635G SNP was analyzed using primers and probes 5′-CACAGTCATCCTCAGGGAAAGG-3′, 5′-CAGCCATCCCATAATCCAGAAGT-3′, VIC-CTCTTCCTCTCACCCCCA-8NFQ, and FAM-CTTCCTCTCGCCCCCA-8NFQ. PXR C8055T SNP was analyzed using primers and probes 5′-GGGATGATTAGATCTTGGTCAGCTT-3′, 5′-CTGGAAGCCACCTGTGGAT-3′, VIC-CCCCTCCATCCTGTTAC-8NFQ′, and FAM-CCCCTCCATTCTGTTAC-8NFQ′.

Statistical Analysis.

Fisher exact tests were used for the analysis of categorical variables. In the case of continuous variables, the Mann-Whitney U test was used to compare two groups, and the Kruskal-Wallis nonparametric one-way ANOVA was used to compare more than two groups. In the presence of statistically significant differences in categorical variables, odds ratios and 95% confidence intervals were calculated, and P values were corrected for age at enrollment and duration of disease. Statistical comparisons were made using Stata Statistical Software (Stata Corp., College Station, TX) or SAS (SAS Institute Inc., Cary, NC). All of the analyses were two-sided, and P values of less than .05 after correction were considered statistically significant.

Results

CYP2D6.

The CYP2D6*4 allele frequencies in patients with PBC and controls are shown in Table 2. The obtained genotype frequencies corresponded to the Hardy-Weinberg equilibrium in our healthy Caucasian sample. Among 169 patients with PBC, 3 (2%) were found to be homozygous for the mutant allele (*4/*4), 118 (70%) were homozygous for the wild-type allele (N/N), and 48 (28%) were heterozygous (*4/N). No significant differences were observed between patients and controls in the frequencies of CYP2D6 genotypes. When the clinical characteristics of patients with PBC were analyzed according to the presence of the CYP2D6*4 allele (Table 3), no significant differences in genotype distributions were observed for age at enrollment, disease duration, or AMA status. Patients carrying the *4 allele presented signs of a slightly more advanced disease as indicated by higher prevalence of ascites (22% vs. 9%, P = .031 after correction for age and disease duration); higher but nonsignificant total bilirubin levels (2.2 ± 4.3 vs.1.4 ± 2.7; P value not significant) were also observed. Moreover, the presence of the *4 allele was found to be associated with ascites (odds ratio, 2.76; 95% CI, 1.1-6.92). Data obtained from a preliminary study of CYP2D6*3 and CYP2D6*6 showed low allele frequencies in both patients (n = 90; 5/180 for CYP2D6*3 and 1/180 for CYP2D6*6) and controls (n = 90; 0/180 for CYP2D6*3 and 6/180 for CYP2D6*6). A similar preliminary study showed that homozygous CYP2D6*5 was not found in 80 patients with PBC and 80 controls. Based on these findings, further genotyping for these variants was not performed.

Table 2. Distribution of CYP2D6, CYP2E1, and MDR1 Genotypes and Alleles in Patients With PBC and Controls
 ControlsPBC
TotalEarly DiseaseAdvanced Disease
  • *

    P = 0.043 after correction for age and disease duration; odds ratio 11.39 (95% CI 1.09–119.47) for c1/c2 genotype determining advanced disease.

CYP2D6    
 N/N164/225 (73%)118/169 (70%)57/77 (74%)61/92 (66%)
 *4/N57/225 (25%)48/169 (28%)19/77 (25%)29/92 (32%)
 *4/*44/225 (2%)3/169 (2%)1/77 (1%)2/92 (2%)
 *4 allele frequency0.1440.1590.1360.179
CYP2E1    
 C1/C1212/223 (95%)159/169 (94%)76/77 (99%)*83/92 (90%)*
 C1/C29/223 (4%)10/169 (5%)1/77 (1%)9/92 (10%)
 C2/C22/223 (1%)000
 C2 allele frequency0.0290.0330.0060.049
MDR1    
 3435 C/C60/225 (27%)47/169 (28%)23/77 (30%)24/92 (26%)
 3435 C/T110/225 (49%)88/169 (52%)39/77 (51%)49/92 (53%)
 3435 T/T55/225 (24%)34/169 (20%)15/77 (19%)19/92 (21%)
 T allele frequency0.4880.4610.4480.472
Table 3. Clinical Features of Patients With PBC and the Presence of the CYP2D6*4 Allele
 N/N (n = 118)*4/N +*4/*4 (n = 51)P Value (Corrected)
  • NOTE. Continuous variables are expressed as the mean ± SD.

  • Only P values below .2 for the comparison between patient groups are reported before and after correction for age and disease duration.

  • *

    Odds ratio 2.76 (95% CI, 1.1–6.92) for *4 allele determining the presence of ascites.

  • Abbreviations: NS, not significant; INR, international normalized ratio; n.v., normal value.

Female sex (n)106 (90%)47 (92%)NS
Age at enrollment (yr)62 ± 1362 ± 11NS
Disease duration (mo)122 ± 71127 ± 76NS
AMA-positive (n)99 (84%)40 (78%)NS
Total bilirubin (mg/dL) (n.v. < 1.0)1.4 ± 2.72.2 ± 4.3.194
Albumin (g/dL) (n.v. >3.5)4.3 ± 0.94.2 ± 0.9NS
Prothrombin time (INR) (n.v. < 1.2)1.04 ± 0.181.03 ± 0.14NS
Ascites (n)11 (9%)11 (22%).031 (.031)*
Mayo score5.6 ± 1.45.9 ± 1.5NS

CYP2E1 c1/c2.

The CYP2E1 genotype and allele frequencies among patients with PBC and controls are shown in Table 2. The obtained genotype frequencies corresponded to the Hardy-Weinberg equilibrium in our healthy Caucasian sample. 159 (94%) of 169 patients with PBC were found to be homozygous c1/c1 and 10 (6%) heterozygous c1/c2. No significant differences were observed between patients and controls in the frequencies of CYP2E1 genotypes. When patients with PBC were analyzed according to disease stage, the frequency of the c1/c2 genotype was found to be significantly higher in advanced-stage PBC compared with early-stage PBC (P = .043 after correction for age and disease duration). Moreover, the presence of the c2 allele was found to be associated with advanced disease (odds ratio, 11.39; 95% CI, 1.09-119.47). When the clinical characteristics of patients with PBC were analyzed according to the presence of the c2 allele (Table 4), patients with this allele presented signs of more advanced disease as indicated by higher total bilirubin levels (3.6 ± 5.3 vs. 1.5 ± 3.0; P = .054), lower serum albumin levels (3.6 ± 0.9 vs. 4.3 ± 0.8; P = .036 after correction for age and disease duration), and higher Mayo score values (6.9 ± 1.9 vs. 5.6 ± 1.4; P = .015 after correction for age and disease duration).

Table 4. Clinical Features of Patients With PBC and the Presence of the CYP2E1 c2 Allele
 c1/c1 (n = 159)c1/c2 (n = 10)P Value (Corrected)
  1. NOTE. Continuous variables are expressed as the mean ± SD. Only P values below .2 for the comparison between patient groups are reported before and after correction for age and disease duration.

  2. Abbreviations: NS, not significant; INR, international normalized ratio; n.v., normal value.

Female sex (n)143 (90%)10 (100%)NS
Age at enrollment (yr)62 ± 1264 ± 13NS
Disease duration (mo)122 ± 69155 ± 112NS
AMA-positive (n)131 (82%)8 (80%)NS
Total bilirubin (mg/dL) (n.v. < 1.0)1.5 ± 3.03.6 ± 5.3.054
Albumin (g/dL) (n.v. > 3.5)4.3 ± 0.93.6 ± 0.9.011 (.036)
Prothrombin time (INR) (n.v. < 1.2)1.03 ± 0.161.13 ± 0.20.109
Ascites (n)19 (12%)3 (30%).125
Mayo score5.6 ± 1.46.9 ± 1.9.011 (.015)

MDR1 C3435T.

The MDR1 C3435T genotypes and allele frequencies among patients with PBC and controls are shown in Table 2. The obtained genotype frequencies corresponded to the Hardy-Weinberg equilibrium in our healthy Caucasian sample. Forty-seven (28%) of 169 patients with PBC were found to be homozygous C/C, 34 (20%) were homozygous T/T, and 88 (52%) were heterozygous C/T. No significant differences were observed between patients and controls in the frequencies of MDR1 genotypes. No significant differences in genotype distributions were observed between patients with early and advanced disease (T allele frequency 0.442 in early disease vs. 0.473 in advanced disease) nor in the clinical characteristics across the three genotypes (data not shown).

PXR.

The PXR1 C-25385T genotypes and allele frequencies among patients with PBC and controls are shown in Table 5. The obtained genotype frequencies corresponded to the Hardy-Weinberg equilibrium in our healthy Caucasian sample. Seventy-two (43%) of 167 patients with PBC were found to be homozygous C/C, 24 (14%) were homozygous T/T, and 71 (43%) were heterozygous C/T. No significant differences in genotype distribution were observed between patients and controls nor between patients with early and advanced disease. Similarly, no significant differences were found in the clinical features of patients with different genotypes (data not shown).

Table 5. Distribution of PXR Genotypes and Alleles in Patients With PBC and Controls
 ControlsPBC
TotalEarly DiseaseAdvanced Disease
−24385 C/C76/225 (34%)72/167 (43%)36/77 (47%)36/90 (40%)
−24385 C/T119/225 (53%)71/167 (43%)31/77 (40%)40/90 (44%)
−24385 T/T30/225 (13%)24/167 (14%)10/77 (13%)14/90 (16%)
−24385 T allele frequency0.3970.3560.3310.378
7635 A/A32/102 (31%)28/100 (28%)12/42 (29%)16/58 (28%)
7635 A/G49/102 (49%)53/100 (53%)25/42 (59%)28/58 (48%)
7635 G/G20/102 (20%)19/100 (19%)5/42 (12%)14/58 (24%)
7635 G allele frequency0.4360.4550.4170.483
8055 C/C66/102 (65%)63/99 (63%)29/43 (68%)34/56 (61%)
8055 C/T29/102 (28%)34/99 (34%)13/43 (30%)21/56 (37%)
8055 T/T7/102 (7%)2/99 (2%)1/43 (2%)1/56 (2%)
8055 T allele frequency0.2110.1920.1740.205

The PXR A7635G genotypes and allele frequencies among patients with PBC and controls are shown in Table 5. Twenty-eight (28%) of 100 patients with PBC were found to be homozygous A/A, 19 (19%) were homozygous G/G, and 53 (53%) were heterozygous A/G. This genotype distribution was similar to what was observed among controls. Patients with PBC presented similar clinical characteristics across the three genotypes (data not shown).

The PXR C8055T genotypes and allele frequencies among patients with PBC and controls are shown in Table 5. No significant differences were observed between patients and controls in the frequencies of C8055T genotypes nor between patients with early and advanced disease. When the clinical characteristics of patients with PBC were analyzed according to their PXR genotypes, similar features were encountered in all groups (data not shown).

Discussion

We investigated several key SNPs of CYP2D6, CYP2E1, MDR1, and PXR in patients with PBC and geographically and sex-matched controls to determine if particular alleles contribute to a link between xenobiotics and PBC. We did not identify an association between such alleles and PBC. In addition, we correlated the genotypes of the investigated genes with histological stages and other clinical and biochemical features of patients in a cross-sectional fashion and identified a significant association of the frequency of CYP2E1 c2 allele with PBC severity. Our series, one of the largest series of PBC cases ever genotyped for candidate genes, included an unusually high frequency of AMA-negative patients, but in all cases the diagnosis was verified as described herein. Moreover, data from AMA-negative patients were similar to the AMA-positive group. Indeed, our comparisons (between patients and controls as well as among affected individuals) should therefore be regarded as statistically powerful. PBC has a wide spectrum of disease progression—some patients remain asymptomatic for decades after diagnosis, while others present a rapidly progressing disease leading to orthotopic liver transplantation or death. Although a number of genetic factors have been proposed to explain such differences, results obtained thus far have proven to be weak or limited to specific geographical areas.4, 5, 30, 31

Four CYP2D6 alleles (CYP2D6*3, CYP2D6*4, CYP2D6*5, and CYP2D6*6) are known to account for 93% to 97% of PM cases,7 with CYP2D6*4 alone accounting for approximately 75% of these10, 32; this is the most widely studied allele in association studies. Brown et al.8 demonstrated a link between the CYP2D6*4 allele and ankylosing spondylitis and postulated that the poor metabolism of xenobiotics by a defective CYP2D6 polymorphism might explain this association. Others have suggested that a PM state might increase susceptibility to Parkinson's disease because of impaired detoxification of neurotoxins.33 It is interesting to note that patients with CYP2D6*4 had slight signs of more severe clinical features as indicated by the prevalence of ascites, though differences in other variables (e.g., total bilirubin) did not reach statistical significance. The low or null allelic frequencies observed for CYP2D6*3, CYP2D6*5, and CYP2D6*6 in a preliminary study performed on both patients and controls did not provide sufficient statistical power and did not warrant further investigation. It is important to note that the observed allelic frequencies are similar to previous reports for all the CYP2D6 polymorphisms studied herein.34

CYP2E1 is involved in the metabolism of drugs, chemicals (including ethanol), and carcinogens, and its c1/c2 polymorphism has been shown to influence enzyme activity per se16 or when induced by specific agents.17, 18 Interestingly, Tsutsumi et al.35 demonstrated that the c2 allele is associated with the development of alcoholic liver disease, likely through the reduced observed activity.16 Conversely, subjects with the c1/c1 genotype had higher CYP2E1 activity induced by isoniazid, suggesting an association between such a genotype and susceptibility to antituberculosis drug-induced hepatitis.18 We can assume that the c1/c2 SNP is involved in determining CYP2E1 activity possibly induced by specific agents, thus influencing the clinical manifestations of several diseases.18, 35–37 Many CYP2E1 substrates have been identified, including halothane, isoniazid, acetone, acetonitrile, estrogen metabolites, and ethanol. Halothane is of great interest in the association of xenobiotic metabolism and susceptibility to PBC, because the sera of patients with halothane-induced hepatitis have autoantibodies to the pyruvate dehydrogenase complex similar to PBC sera.13 Moreover, previous data have also indicated that both oxidative and reductive metabolism of halothane can lead to active xenobiotics in vivo.2 On the other hand, the influence of low doses of nicotine on the expression of liver CYP2E1 in animal models38 is also interesting considering that smoking is a risk factor for PBC.14

The frequency of the c1/c2 genotype in patients with advanced stage PBC was significantly higher than that of early-stage PBC. Accordingly, among other variables, the Mayo score values in patients carrying the c2 allele were significantly higher than those observed in patients with the c1/c1 genotype. A possible confounding effect of age (one of the factors in the calculation of such prognostic index) or disease duration was considered in this comparison, and we also note that these variables were not significantly different in patients with different genotypes. Our results indicate that genetically determined alterations of the metabolism of xenobiotics possibly mediated by other compounds might play a role in determining disease severity. Interestingly, in one of our previous studies, AMA from patients with PBC often reacted with a higher titer against the xenobiotically modified peptide than with the native lipoyl domain; that is, altered lipoic acid actually increased antibody binding.2 Considering such data and the results described herein, we hypothesize that higher CYP2E1 activity induced by the presence of the c2 allele may make patients with PBC produce more active xenobiotics, including drugs, resulting in an enhanced T-cell reactivity to organic modified autoepitopes. Although our findings on the CYP2E1 c2 allele were limited to a small subgroup of patients, this polymorphism in PBC should be further assessed as a prognostic marker. Our data therefore support the previously suggested hypothesis, derived from the weak association between PBC and human leukocyte antigen haplotypes,39 that susceptibility and progression are most likely caused by a combination of several factors, including more than one genetic determinant (e.g., abnormalities in sex chromosomes40) together with specific genomic variations.

The MDR1 gene is responsible for the production of PGP, which is highly expressed in intestinal epithelial cells, and its genetic polymorphisms play a major role in determining local defense against bacteria and xenobiotics.41 Individuals carrying the homozygous MDR1 3435TT genotype had on average a two-fold lower intestinal level of PGP expression compared with the CC genotype, resulting in decreased transport of PGP substrates into the gut lumen and higher absorption from the gastrointestinal tract.20 We have postulated that xenobiotic modification of lipoic acid occurs on microbial proteins.42 In particular, we reported that sera from patients with PBC react in a highly directed and specific fashion against proteins from the ubiquitous xenobiotic-metabolizing bacterium Novosphingobium aromaticivorans.43 In the data presented herein, homozygous TT genotype was found in 20% of patients with PBC, and no significant differences were observed between patients and controls in the frequencies of MDR1 genotypes and alleles (T allele frequency 0.461 in patients with PBC vs. 0.488 in controls). In the future, studies designed to investigate MDR1 genetic polymorphisms linked to alterations in PGP expression or SNPs of other genes related to gastrointestinal protection against bacteria and xenobiotics should be addressed. Based on a similar hypothesis, we note that Pauli-Magnus et al.44 recently investigated the genetic polymorphisms of two different adenosine triphosphate–dependent binding cassettes (ABC B11 and B6, or MDR3) and reported a lack of correlation between polymorphisms of such genes and susceptibility to PBC.

PXR is a nuclear hormone receptor that acts as a xenobiotic sensor to transcriptionally regulate many important genes such as CYP3A4 and MDR1. Although there is the extent of CYP3A4 phenotypic variation, few allelic polymorphisms with altered activity have been reported.45 Functional genetic variations in PXR, on the other hand, influence the expression of CYP3A4.22 Among PXR SNPs, C-25385T was associated with CYP3A4 inducibility phenotype in the liver, while A7635G and C8055T were involved in intestinal CYP3A induction. In particular, 3A4 is the most abundant CYP detectable in the liver (18.4% of the total CYP activity) and that an increased CYP3A4 induction in the liver is associated with the PXR -25385CC genotype.46 We postulate that higher liver CYP3A4 activity in patients with PBC directly caused by the PXR -25385 CC genotype may alter the metabolism of xenobiotics, thus leading to the induction of autoimmunity through the enhanced production of xenobiotic structural analogues of lipoic acid. Our data, however, showed that the frequency of the latter genotype among patients with PBC was not significantly different compared with matched controls (43% vs. 33%; P = .06). Similarly, no differences were observed in the prevalence of A7635G and C8055T genotypes. Because PBC presents a striking female predominance,23 it is interesting to note that progesterone is one of the substrates of CYP3A4.47

In summary, we concentrated on genes for which sound evidence has suggested characteristics constituting a possible link with PBC. Second, we note that our design allowed to investigate the crucial steps of xenobiotic transport and metabolism. Third, we chose variations within such genes that were demonstrated to be “coding” for phenotypic differences, although in the case of CYP2E1 such differences can be mediated by other compounds. Our findings herein, along with our previous data, point toward a “multi-hit” pathogenesis of PBC, with different genetic factors leading to onset and severity of disease.

Ancillary