Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Associations of allelic variants of the multidrug resistance gene (ABCB1 or MDR1) and Inflammatory Bowel Disease and their effects on disease behavior: A case-control and meta-analysis study

Associations of allelic variants of the multidrug resistance gene (ABCB1 or MDR1) and... Background Allelic variants of the ATP-binding cassette, subfamily B member 1 (ABCB1), also known as the multidrug resistance gene (MDR1) that encodes the membrane-bound efflux transporter P-glycoprotein 170 (PGP-170), have been associated with inflammatory bowel disease but with conflicting results. Methods The present study examined the association of ABCB1 C3435T and G2677T/A in a large British case-control cohort of 828 Crohn's disease, 580 ulcerative colitis (UC) cases, and 285 healthy controls. The effect of these variants was further examined with respect to phenotypic and epidemiological characteristics. A meta-analysis was carried out of our results and those from 8 previously published association studies of the C3435T variant in inflammatory bowel disease. Results The 2677T allele was significantly increased in British UC cases compared with controls (45.2% vs. 39.6%; P = 0.034). In particular, the TT genotype was significantly associated with severe UC (odds ratio [OR] 1.90; 95% CI 1.01–3.55) and the use of steroids in UC (OR 1.77; 95% CI 1.08–2.88). No significant association was seen with C3435T and UC, Crohn's disease, or any clinical subgroup. A meta-analysis of 9 association studies of C3435T showed a significant association of the 3435T allele with UC (OR 1.12; 95% CI 1.02–1.23; P = 0.013) but not with CD. Conclusions These results indicate that ABCB1 sequence variants are associated with a small increase in the risk of developing UC and may influence disease behavior. Crohn's disease, ulcerative colitis, association, genetics, meta-analysis Crohn's disease (CD) and ulcerative colitis (UC) are the 2 common clinical subtypes of idiopathic inflammatory bowel disease (IBD) characterized by chronic inflammation of the gastrointestinal tract. The prevalence of IBD in Western countries is ≈63 to 445/100,000,1 and together with its clinical course and high morbidity, its prevalence represents a significant burden to health care. The underlying pathogenesis remains unclear but may involve a combination of persistent bacterial infection, a defective mucosal barrier, or an imbalance in the regulation of the immune system.2 Epidemiological and family studies suggest that genetic factors play a significant role in determining susceptibility to IBD, with a sibling relative risk of 15 to 35 for CD3,4 and 6 to 9 for UC.3,5 Recently, a member of the adenosine triphosphate-binding cassette superfamily ABCB1 (previously known as MDR1), which encodes the membrane transport protein P-glycoprotein (PGP)-170, has been proposed as a candidate gene for IBD on the basis of its location within a region of suggestive linkage detected on chromosome 7q.6 A role for PGP-170 in IBD was suggested by a mouse knockout model for ABCB1 (mdr1a−/−) in which developmentally normal mice spontaneously develop a colitis resembling UC in humans.7 PGP-170 is expressed in several human tissues, including peripheral blood lymphocytes, and is localized to the apical membrane of epithelial cells in the jejunum and colon. In addition to its role in the secretion of drugs, PGP-170 is thought to play a protective role in prevention of cell apoptosis,8 as well as cell-mediated immunity, by promoting migration of antigen-presenting cells from the periphery to lymph nodes to initiate a T lymphocyte response.9 The ABCB1 gene spans 200 kb comprising 28 exons with 2 transcription start sites and has been shown to be highly polymorphic.10,–12 Recent attention has focused on 2 single nucleotide polymorphisms (SNPs), C3435T in exon 26 and a triallelic SNP G2677T/A (Ala893Ser/Thr), which have been shown to correlate with the activity and expression of PGP-170.10,11 However, reports of an association between these 2 SNPs and IBD have been conflicting. The initial study by Schwab et al13 identified a significant association of the 3435T allele with UC but not CD, a finding that has been replicated by some14,15 but not others.16,17 An association also has been reported between the G allele of G2677T/A and IBD in case-control analyses17; this study also found significant overtransmission of the G allele to CD offspring and nonsignificant overtransmission of the same allele to UC offspring in a family-based test of association. Others, however, have found an association with the 2677T allele in UC, with a trend of an association with refractory CD.14 Significant linkage disequilibrium has been observed between these 2 SNPs in different white populations.14,15 More recently, Ho et al15 undertook a genotype-phenotype analysis showing a significant association of C3435T with extensive UC and a higher frequency of the G2677 allele with severe IBD and the need for surgery, although Schwab et al13 previously failed to detect an association of ABCB1 genotypes with specific phenotypes. In view of these conflicting findings and the potential importance of this gene to the understanding of the pathogenesis of IBD, we have investigated the contribution of ABCB1 C3435T and G2677T/A to IBD in a large cohort of British CD and UC patients. We have examined the relationship between genotypic and phenotypic expression, as well as drug exposure and epidemiological factors, in a set of clinically well-characterized patients. Evidence for epistasis between the CD-associated gene CARD15 and ABCB1 in CD patients also was investigated. We have interpreted our association study results in the context of previous studies by performing a comprehensive meta-analysis of 9 independent populations genotyped for C3435T. Materials and Methods Patient Collection A cohort of unrelated British patients with IBD were recruited after ethical review and obtaining informed consent from Guy's and St Thomas' Hospitals and St Mark's Hospital (United Kingdom). The diagnoses of CD and UC were made by established criteria of clinical, radiological, and endoscopic analyses and from histology reports.18,19 Healthy controls were recruited from the greater London region to match the recruitment area and ethnicity of IBD cases. The ethnicity of both cases and controls was ≈94% white and 6% nonwhite; there was no evidence of population stratification in these cohorts in previous studies of > 100 SNP and microsatellite markers. Additional data on site of disease, disease behavior, and therapy were available for a subset of 385 UC and 597 CD patients. Genotyping Genotyping of C3435T (rs1045642) was performed by TaqMan (Applied Biosystems, Warrington, UK) available from ABI assay on demand. G2677T/A was genotyped by Pyrosequencing to determine the allele frequencies of all 3 variants. Genotyping of the CARD15 mutations R702W and L1007fs was also performed by Pyrosequencing, and G908R was performed by TaqMan (ABI assay by design). All of the SNP assays were verified with a subset of at least 48 samples that had been genotyped by DNA sequencing and by checking for Hardy-Weinberg equilibrium in control populations. Sequence and reaction settings are available on request. Phenotypic Assessment UC phenotype was classified by disease extent at diagnosis and latest follow-up (Table 1). We defined extensive disease as disease extending past the splenic flexure, left-sided colitis as disease extending up to the splenic flexure, and proctitis as disease confined to the rectum. When there were discrepancies between the extent of disease as determined by macroscopic and microscopic evidence, histological results were used.15 Severe colitis was determined by the need for surgery either for disease refractory to treatment or for acute severe disease (excluding surgery for dysplasia), as well as for any patient with a severe attack of UC requiring admission to the hospital. Table 1. Demographics and Clinical Characteristics of 385 Non-Jewish Patients With UC     View Large Table 1. Demographics and Clinical Characteristics of 385 Non-Jewish Patients With UC     View Large The clinical characteristics of the CD patients are shown in Table 2. Location was considered the maximum extent of disease involvement at any time before the first surgical resection. Ileal disease was determined as being limited to the terminal ileum with or without spillover into the cecum; colonic disease was any lesion located between the ascending colon and rectum; and ileocolonic disease was defined as any disease of the terminal ileum with or without spillover into the cecum and any location between the ascending colon and rectum. In addition, a further location subset of the presence of perianal disease, defined by the presence of perianal fistulas, ulcers, fissures, and abscesses but not skin tags, was recorded. Disease behavior was classified by current or past behavior subtypes. Penetrating disease was determined by the presence of any abnormal connection between 2 epithelial surfaces or the formation of intraabdominal abscesses or perforations. This was subdivided into intraabdominal fistulas and perianal fistulas, including perianal and ischiorectal abscesses. Any postoperative perforating complications were excluded. Stenotic disease was recorded on the basis of radiological, endoscopic, or histopathological findings, together with the presence of obstructive symptoms and/or evidence of prestenotic dilatation on barium examination. Inflammatory disease was diagnosed in the absence of stenotic or perforating behavior. Surgical history was also recorded in detail and divided into the number of laparotomies with bowel resection and the number of examinations under anesthesia for drainage of perianal sepsis or treatment of fistulas. The presence or absence of granulomas from histopathology specimens also was noted when available. Table 2. Demographics and Clinical Characteristics of 597 Non-Jewish Patients With CD     View Large Table 2. Demographics and Clinical Characteristics of 597 Non-Jewish Patients With CD     View Large For both UC and CD, the requirement for treatment with systemic steroids was assessed and an estimate of the cumulative amount of steroid needed to treat active disease was made by a detailed examination of patients' notes. The need for immunosuppressant therapy also was recorded, as was the need for treatment with biological agents such as infliximab. Other phenotypic details, including age at onset of symptoms, family history, extraintestinal symptoms, and years of follow-up, were noted. Data on smoking habit (current smoker, ex-smoker, or never smoked) were collected at the time of the latest follow-up; cases were stratified into “ever smoked” or “never smoked” for statistical analysis. Statistical Methods Genotype and allele frequencies were compared between cases and controls through the use of χ2 tests. Two-locus haplotype frequencies were estimated by an expectation-maximization algorithm implemented through the use of COCAPHASE, a module of UNPHASED.20 Significance values are reported uncorrected for multiple testing. Linkage disequilibrium between markers was assessed with multiallelic D′21,22 calculated with the GOLD program.23 Meta-analysis Published association studies of ABCB1/MDR1 in IBD were identified through PubMed by a search of the keywords “MDR1,” “ABCB1,” “association,” “inflammatory bowel disease,” “Crohn's disease,” and “ulcerative colitis” in various combinations. A total of 7 publications involving an association study with ABCB1 C3435T were identified; only 3 had genotyped the G2677T/A polymorphism. Two of these performed genotyping by TaqMan, which did not identify the rare 2677A allele.14,15 Thus, individuals heterozygous for this allele appeared homozygous for the opposite allele, and AA homozygous genotypes would not be detected. Consequently, this technique would bias estimates of disease risk in these studies. Therefore, and in view of the low number of studies, the meta-analysis was restricted to the C3435T polymorphism. When full genotype counts were not presented or not provided separately for CD and UC, these data were requested from individual study authors. One study24 was excluded from the meta-analysis because no C3435T CC homozygotes were observed, resulting in significant deviation from Hardy-Weinberg equilibrium. All of the remaining studies were included, regardless of ascertainment method and/or distribution of clinical subphenotypes. When individual studies included > 1 population, each population was treated independently in the meta-analysis. Characteristics of each study are shown in Table 3. After these results were pooled with our own data, a total of 1743 UC cases, 2311 CD cases, and 2931 controls were included in the meta-analysis. Table 3. Published Studies Included in Meta-analysis of MDR1 C3435T     View Large Table 3. Published Studies Included in Meta-analysis of MDR1 C3435T     View Large Genotype and allele-specific odds ratios (ORs) and 95% CIs were calculated from genotype data for each study. ORs were combined across studies using both fixed- and random-effects models. A fixed-effects model assumes that all studies are estimating a single true underlying effect size (i.e., no heterogeneity between studies). A random-effects model assumes that the true effect size in each study comes from a random distribution of effect sizes with fixed mean and variance (i.e., heterogeneity between studies). In view of the low number of studies included in the meta-analysis but large within-study sample sizes, the inverse-weighted method25 was selected for a fixed-effects model26; the random-effects model used was that of DerSimonian and Laird.27 Heterogeneity was assessed by use of the χ2 Q statistic.28 Results Effect of ABCB1 C3435T and G2677T/A Polymorphisms on Disease Susceptibility A total of 828 CD cases, 580 UC cases, and 285 British healthy controls were genotyped for C3435T and G2677T/A. Control and IBD case genotypes for both C3435T and G2677T/A were in Hardy-Weinberg equilibrium. Our combined case collection contained a total of 85 individuals of Jewish ancestry (39 UC, 46 CD). In view of the significant differences in the frequency of ABCB1 polymorphisms between populations29 and particularly Jewish and non-Jewish that have been previously reported,17,30 allele frequencies at the 2 loci were compared between Jewish and non-Jewish cases. We observed significant differences in allele frequencies in British Jewish cases compared with non-Jewish cases (e.g., 3435T: 39.7% versus 55.5%, respectively; P = 0.01; Tables 4 and 5); therefore, the data were analyzed both including and excluding Jewish cases. Table 4. Genotypes and Allele Frequencies of MDR1 G2677T/A in Cases and Controls     View Large Table 4. Genotypes and Allele Frequencies of MDR1 G2677T/A in Cases and Controls     View Large Table 5. Genotypes and Allele Frequencies of MDR1 C3435T in Cases and Controls     View Large Table 5. Genotypes and Allele Frequencies of MDR1 C3435T in Cases and Controls     View Large The frequencies of the ABCB1 G2677T/A alleles were similar in CD cases (G2677, 57.8%; 2677T, 40.8%; 2677A, 1.4%) and controls (G2677, 57.9%; 2677T, 39.6%; 2677A, 2.5%; Table 4). For C3435T, there were no significant differences in the frequency of the 3435T allele between CD cases (51.9%) and controls (53.9%; Table 5). Allele frequencies and ORs for both SNPs were similar when Jewish cases were included. CD patients genotyped for C3435T and G2677T/A were stratified by carriage of the 3 common associated CARD15 mutations to look for evidence of epistasis between the 2 loci. We did not detect any significant difference in the frequency of C3435T in CD cases who carried 1 or 2 CARD15 mutations compared with both controls and CD cases with no CARD15 mutations (Table 5), in agreement with previous studies.13,16 Similarly, we found no evidence of an interaction between G2677T/A and CARD15 (Table 4). Marginally significant differences in allele frequencies of G2677T/A were observed in UC cases compared with controls (G2677: cases, 52.7%; controls, 57.9%; P = 0.050; 2677T: cases, 45.2%; controls, 39.6%; P = 0.034;Table 4). The frequency of the rare allele 2677A was 2.1% in UC cases and 2.5% in controls. For SNP C3435T, we did not detect any significant difference in the frequency of the 3435T allele between UC cases (55.5%) and controls (53.9%). The frequency of the TT genotype, previously reported as exhibiting the strongest association with UC,13,15 also was not significantly different between cases (31.8%) and controls (29.6%). Allele frequencies and ORs for both SNPs were similar when Jewish patients were included (Tables 4 and 5). Because C3435 and G2677T/A are in linkage disequilibrium (multiallelic D′ = 0.85), we performed 2-locus haplotype association tests with UC and CD. We found no significant C3435T/G2677T/A haplotype association with either CD or UC. The T-T haplotype was 42.2% in UC cases compared with 38.6% in controls, but this difference was not significant. The frequency of the T-G haplotype, previously reported to be associated with UC,15 was not significantly different in UC cases (12.0%) and controls (12.9%). Genotype-Phenotype Analysis Detailed phenotypic information was available on 597 CD and 385 UC British non-Jewish cases. Tables 1 and 2 summarize the clinical characteristics of CD and UC patients, respectively. The proportion of males was higher in UC (50.1%) than CD (43.2%). The median age at onset of symptoms for CD and UC patients was 21 and 29 years, respectively, with a median duration of follow-up of 14 years (UC) and 16 years (CD). For UC, 71.9% of our patients had extensive colitis at their latest follow-up. The frequencies of the C3435T and G2677T/A variants in the different phenotypic subgroups were compared with their frequencies in controls to investigate whether the association may be restricted to specific patterns of disease behavior. In CD cases, we did not find any association of C3435T with disease behavior, disease location, treatment, or surgical history (all P values > 0.05; data not shown). The frequency of the 2677 TT genotype was increased in CD cases without ileal disease (OR 1.86; 95% CI 1.09–3.16; P = 0.022), but given the lack of association of 2677T with the overall phenotype of CD and multiple testing of phenotypic subgroups, this finding requires independent replication. In UC cases, the highest risk associated with G2677T/A was observed for the TT genotype in cases with severe disease (OR 1.90; 95% CI 1.01–3.55; P = 0.045) and in patients who had received steroids (OR 1.77; 95% CI 1.08–2.88; P = 0.022; Table 6). In addition, we detected a significant association between the TT genotype and extensive disease (OR 1.67; 95% CI 1.01–2.75; P = 0.045), although a similar nonsignificant risk was observed in 85 cases with only left-sided disease. No association was observed with the presence of extraintestinal manifestations. We did not detect any association of either the T allele or TT genotype of C3435T with disease extent, severity, treatment, or surgical history for UC. In particular, we did not detect an association between the 3435T allele or the TT genotype and extensive colitis. The highest risk of UC associated with the T allele occurred in patients who had received immunosuppressants (OR 1.24; 95% CI 0.98–1.56) and steroid treatment (OR 1.25; 95% CI 0.90–1.73), although neither of these reached statistical significance. Table 6. C3435T and G2677T/A Disease Risks in Clinical Subgroups of 385 UC Patients     View Large Table 6. C3435T and G2677T/A Disease Risks in Clinical Subgroups of 385 UC Patients     View Large Finally, we did not detect an association between C3435T or G2677T/A and age of onset of disease symptoms in either UC or CD. C3435T Meta-analysis The frequency of the 3435T allele, which previously has been reported to be associated with UC,13,–15 was not significantly higher in our cases compared with controls. Therefore, we sought to assess our results in the context of all of the published studies of this SNP by meta-analysis. Control genotype frequencies were in Hardy-Weinberg equilibrium in all studies included in the meta-analysis, and the frequency of the 3435T allele in our control population was within the range of those reported for other non-Jewish populations. Study-specific disease ORs and 95% CIs associated with frequency of the 3435T allele are shown in Figure 1. No significant between-study heterogeneity was observed for UC (Q = 9.42; P = 0.40) or CD (Q = 13.0; P = 0.16). Under a fixed-effects model, meta-analysis of all 9 studies showed a significant association (P = 0.013) of the 3435T allele with UC (OR 1.12; 95% CI 1.02–1.23; Figure 1A). Similar results were observed under a random-effects model (OR 1.12; 95% CI 1.02–1.23). Of the 9 studies, 7 produced a study-specific OR > 1 for the 3435T allele, although only the initial study13 and 1 other study15 were significant. The 95% CIs for the ORs estimated by individual studies were consistent with the pooled OR. Several studies have suggested that a higher disease risk is associated with the TT genotype.13,15,31 The pooled OR for the TT genotype in UC was similar to that for the T allele (OR 1.13; 95% CI 0.98–1.31). Figure 1. View largeDownload slide Meta-analysis for MDR1 C3435T association studies. ORs and 95% CIs are displayed for individual studies. Pooled ORs are shown for fixed- and random-effects models. Figure 1. View largeDownload slide Meta-analysis for MDR1 C3435T association studies. ORs and 95% CIs are displayed for individual studies. Pooled ORs are shown for fixed- and random-effects models. Disease risk estimates reported by initial published associations tend to be inflated.32 Therefore, the influence of the first reported association of ABCB1 and UC13 was assessed by excluding this study from the meta-analysis. Results from a meta-analysis that did not include this initial study were similar to those obtained from analysis of all studies. For example, under a fixed-effects model, the pooled OR for UC was 1.10 (95% CI 1.00–1.21), suggesting that inclusion of the first published association study in the meta-analysis had a minimal influence on the pooled OR. In view of the lower frequency of the T allele (35%) in the Jewish population,17 we also considered the effect of excluding these data from the meta-analysis. The disease risk obtained after excluding the Jewish population was close to that obtained from all studies (fixed-effects model: OR 1.11; 95% CI 1.02–1.22). For CD, the pooled ORs were not significant for either the fixed-effects (OR 0.98; 95% CI 0.91–1.07) or the random-effects (OR 0.99; 95% CI 0.89–1.10) model (Figure 1B). Only 5 of the 9 studies produced a study-specific OR > 1 for the 3435T allele. Discussion Meta-analysis of 9 independent populations has confirmed association between C3435T and UC, with an OR of 1.12 (95% CI 1.02–1.23), and all 9 individual studies identified study-specific risk estimates that were consistent with this small effect. The fact that no association was detected with CD in a study of > 2300 CD cases indicates that this variant of ABCB1 is unlikely to confer an increased risk of this phenotype of IBD. In our own British cohort, we did not find a significant association of C3435T with UC or CD. We did identify significant differences in the frequency of the 3435T allele between Jewish and non-Jewish cases, which is in agreement with the differences between Ashkenazi Jewish and non-Jewish subjects observed by Brant et al17 and Ostrovsky et al30 at this locus, as well as among different world populations.29 However, including Jewish subjects in the meta-analysis did not affect the pooled estimate of risk. The tight CIs obtained in the meta-analysis for both UC and CD indicate that the risk estimates we have derived are likely to be robust. The lack of replication of the association with UC in some studies is consistent with the small size of the effect; our study has only 19% power to detect the association, yet it is the most highly powered study reported to date. This underlines the value of a meta-analysis in detecting effects of this size. To investigate the possibility that the ABCB1 associations were restricted to particular phenotypic features of IBD, we classified a large subset of our cohort into rigorously phenotyped clinical subgroups. We did not find an association of C3435T with any clinical group of CD or UC. We did not replicate the finding by Ho et al15 of an association of extensive UC with the T allele and TT genotype of C3435T, despite the fact that 71.9% (n = 264) of our large cohort of phenotyped UC patients had extensive colitis. The high proportion of UC cases with extensive disease may reflect not only the tertiary care institutions from which cases were recruited but also the use of histological extent of disease. Discrepancies of up to 39% between colonoscopy and histology as regards establishing the distribution of inflammation in the colon have been reported,33 with a suggestion that histological changes are a more sensitive marker of inflammation than endoscopic changes.34 We did find a significant association of G2677T/A with UC. Brant et al17 also found an association of G2677T/A with IBD, although it was with the G allele rather than the T allele observed in both our cohort and as part of the associated haplotype found by Potocnik et al.14 Others have failed to observe an association between G2677T/A and IBD.15 The reasons for the observed differences are unclear; there may be phenotypic heterogeneity between the cohorts (phenotype subgroups are not reported by Brant et al),17 and some studies may lack the statistical power to detect the moderate effect size. We also detected a significant association of the TT genotype of G2677T/A with both severe and extensive UC. A similar estimated risk was observed in patients with left-sided UC, although it was not significant, perhaps because of the smaller sample size (n = 85) in this group. This finding suggests that the association with the TT genotype is not specific to extensive colitis. It is important to acknowledge that the genotype-phenotype correlations reported here do not incorporate statistical correction for multiple testing because the application of a stringent correction to testing of multiple disease classes may result in loss of a signal from a genuine association. Replication of the associations in independent data sets is required to validate these findings. The 2677T allele of ABCB1 (Ser893) has been associated with enhanced activity of PGP-170.11 Our finding of an association of the TT genotype of G2677T/A with severe UC is thus consistent with the observation by Farrell et al35 that high PGP-170 expression is associated with UC patients with severe glucocorticoid-resistant disease. The association of the TT genotype with requirement for steroid therapy is also consistent with their observation that high peripheral blood lymphocyte PGP-170-expressing patients are more likely to require steroids than patients with low PGP-170 expression, although they did not investigate the role of genetic variants of the ABCB1 gene. However, our results appear to be at variance with the association of colitis with ABCB1 deficiency in the mdr1a−/− mouse7 and the hypothesis that low levels of PGP-170 expression increase susceptibility to IBD.15 Whether this reflects an etiological difference between human and animal models of disease remains unclear. Studies investigating expression of PGP-170 variants are conflicting, with Kimchi-Sarfaty et al36 failing to demonstrate any difference in cell surface distribution and function of PGP-170 in G2677T/A variants compared with wild-type. Recently, it has been reported that the 3435T allele of ABCB1 is associated with a moderate reduction in mRNA stability.37 It is also possible that an indirect association with disease susceptibility exists with both C3435T and G2677T/A that is caused by linkage disequilibrium with the true causal variant, which may affect the expression of PGP-170.38 If so, the true degree of disease risk attributable to ABCB1 may be higher than that associated with either of these polymorphisms. Sequencing of the coding regions of the ABCB1 gene in IBD patients revealed 2 intronic SNPs in linkage disequilibrium with the risk haplotype, but their possible effect on PGP-170 expression is not known.14 Conclusions The meta-analysis we have carried out supports the results of several individual studies that DNA sequence variation in ABCB1 influences susceptibility to UC but suggests that the size of the effect is small. The C3435T variant, however, does not appear to confer susceptibility to CD, but the other variants need to be tested in a sample of comparable size to resolve the question of the involvement of ABCB1 in CD. These results may have important implications for the organization of studies that investigate the association of candidate genes that may have real but moderate effects on disease risk in IBD. In addition, our finding of an association between the 2677T allele and severe UC suggests that ABCB1 may influence disease behavior and would be worthy of follow-up in other cohorts. An analysis of ABCB1 expression and function in primary cells of cases and controls with defined ABCB1 genotypes would help to clarify the role of sequence variation in this gene in disease susceptibility and behavior in IBD.38 Acknowledgments We thank Dr Uros Potocnik (University of Ljubljana, Ljubljana, Slovenia), Dr Judy Cho (University of Chicago, Chicago, Ill), and Dr Peter Croucher (University of Kiel, Kiel, Germany) forproviding additional information for the meta-analysis. All of theother data were extracted directly from publications. References 1. Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease: incidence, prevalence, and environmental influences. Gastroenterology.  2004; 126: 263– 271. CrossRef Search ADS PubMed  2. Sartor RB. Clinical applications of advances in the genetics of IBD. Rev Gastro Dis.  2003; 3: S9– S17. 3. Kuster W, Pascoe L, Purrmann J, et al. The genetics of Crohn disease: complex segregation analysis of a family study with 265 patients with Crohn disease and 5,387 relatives. Am J Med Genet.  1989; 32: 105– 108. Google Scholar CrossRef Search ADS PubMed  4. Meucci G, Vecchi M, Torgano G, et al. Familial aggregation of inflammatory bowel disease in northern Italy: a multicenter study: the Gruppo di Studio per le Malattie Infiammatorie Intestinali (IBD Study Group). Gastroenterology.  1992; 103: 514– 519. Google Scholar CrossRef Search ADS PubMed  5. Orholm M, Munkholm P, Langholz E, et al. Familial occurrence of inflammatory bowel disease. N Engl J Med.  1991; 324: 84– 88. Google Scholar CrossRef Search ADS PubMed  6. van Heel DA, Fisher SA, Kirby A, et al. Inflammatory bowel disease susceptibility loci defined by genome scan meta-analysis of 1952 affected relative pairs. Hum Mol Genet.  2004; 13: 763– 770. Google Scholar CrossRef Search ADS PubMed  7. Panwala CM, Jones JC, Viney JL. A novel model of inflammatory bowel disease: mice deficient for the multiple drug resistance gene,mdr1a, spontaneously develop colitis. J Immunol.  1998; 161: 5733– 5744. Google Scholar PubMed  8. Smyth MJ, Krasovski E, Sutton VR, et al. The drug efflux protein, P-glycoprotein, additionally protects drug-resistant tumor cells from multiple forms of caspase-dependent apoptosis. Proc Natl Acad Sci U S A.  1998; 95: 7024– 7029. Google Scholar CrossRef Search ADS PubMed  9. Randolph GJ, Beaulieu S, Pope M, et al. A physiologic function for p-glycoprotein (MDR-1) during the migration of dendritic cells from skin via afferent lymphatic vessels. Proc Natl Acad Sci U S A.  1998; 95: 6924– 6929. Google Scholar CrossRef Search ADS PubMed  10. Hoffmeyer S, Burk O, von Richter O, et al. Functional polymorphisms of the human multidrug-resistance gene: multiple sequence variations and correlation of one allele with P-glycoprotein expression and activity in vivo. Proc Natl Acad Sci U S A.  2000; 97: 3473– 3478. Google Scholar CrossRef Search ADS PubMed  11. Kim RB, Leake BF, Choo EF, et al. Identification of functionally variant MDR1 alleles among European Americans and African Americans. Clin Pharmacol Ther.  2001; 70: 189– 199. Google Scholar CrossRef Search ADS PubMed  12. Cascorbi I, Gerloff T, Johne A, et al. Frequency of single nucleotide polymorphisms in the P-glycoprotein drug transporter MDR1 gene in white subjects. Clin Pharmacol Ther.  2001; 69: 169– 174. Google Scholar CrossRef Search ADS PubMed  13. Schwab M, Schaeffeler E, Marx C, et al. Association between the C3435T MDR1 gene polymorphism and susceptibility for ulcerative colitis. Gastroenterology.  2003; 124: 26– 33. Google Scholar CrossRef Search ADS PubMed  14. Potocnik U, Ferkolj I, Glavac D, et al. Polymorphisms in multidrug resistance 1 (MDR1) gene are associated with refractory Crohn disease and ulcerative colitis. Genes Immun.  2004; 5: 530– 539. Google Scholar CrossRef Search ADS PubMed  15. Ho GT, Nimmo ER, Tenesa A, et al. Allelic variations of the multidrug resistance gene determine susceptibility and disease behavior in ulcerative colitis. Gastroenterology.  2005; 128: 288– 296. Google Scholar CrossRef Search ADS PubMed  16. Croucher PJ, Mascheretti S, Foelsch UR, et al. Lack of association between the C3435T MDR1 gene polymorphism and inflammatory bowel disease in two independent Northern European populations. Gastroenterology.  2003; 125: 1919– 1920. Google Scholar CrossRef Search ADS PubMed  17. Brant SR, Panhuysen CI, Nicolae D, et al. MDR1 Ala893 polymorphism is associated with inflammatory bowel disease. Am J Hum Genet.  2003; 73: 1282– 1292. Google Scholar CrossRef Search ADS PubMed  18. Lennard-Jones JE. Classification of inflammatory bowel disease. Scand J Gastroenterol.  1989; 170: 2– 6. Google Scholar CrossRef Search ADS   19. Podolsky DK. Inflammatory bowel disease. N Engl J Med.  1991; 325: 928– 937. CrossRef Search ADS PubMed  20. Dudbridge F. Pedigree disequilibrium tests for multilocus haplotypes. Genet Epidemiol.  2003; 25: 111– 115. 21. Lewontin RC. The interaction of selection and linkage, I: general considerations; heterotic models. Genetics.  1964; 49: 49– 67. PubMed  22. Hedrick P. Gametic disequilibrium measures: proceed with caution. Genetics.  1987; 117: 331– 341. Google Scholar PubMed  23. Abecasis GR, Cookson WOC. GOLD: graphical overview of linkage disequilibrium. Bioinformatics.  2000; 16: 182– 183. Google Scholar CrossRef Search ADS PubMed  24. Gazouli M, Zacharatos P, Gorgoulis V, et al. The C3435T MDR1 gene polymorphism is not associated with susceptibility for ulcerative colitis in Greek population. Gastroenterology.  2004; 126: 367– 369. Google Scholar CrossRef Search ADS PubMed  25. Sutton AJ, Brams KR, Jones DR, et al Methods for Meta-Analysis in Medical Research . Chichester, UK: Wiley; 2000. 26. Fleiss JL. Statistical Methods for Rates and Proportions  ( 2nd ed.). New York: Wiley; 1981. 27. DerSimonian R, Laird N. Meta analysis in clinical trials. Controlled Clin Trials.  1986; 7: 177– 188. Google Scholar CrossRef Search ADS   28. Cochran WG. The combination of estimates from different experiments. Biometrics.  1954; 10: 101– 129. CrossRef Search ADS   29. Tang K, Wong LP, Lee EJ, et al. Genomic evidence for recent positive selection at the human MDR1 gene locus. Hum Mol Genet.  2004; 13: 783– 797. Google Scholar CrossRef Search ADS PubMed  30. Ostrovsky O, Nagler A, Korostishevsky M, et al. Genotype and allele frequencies of C3435T polymorphism of the MDR1 gene in various Jewish populations of Israel. Ther Drug Monit.  2004; 26: 679– 684. Google Scholar CrossRef Search ADS PubMed  31. Glas J, Torok HP, Schiemann U, et al. MDR1 gene polymorphism in ulcerative colitis. Gastroenterology.  2004; 126: 367. Google Scholar CrossRef Search ADS PubMed  32. Ioannidis JPA, Ntzani EE, Trikalinos TA, et al. Replication validity of genetic association studies. Nat Genet.  2001; 29: 306– 309. Google Scholar CrossRef Search ADS PubMed  33. Niv Y, Bat L, Ron E, et al. Change in the extent of colonic involvement in ulcerative colitis: a colonoscopic study. Am J Gastroenterol.  1987; 82: 1046– 1051. Google Scholar PubMed  34. Moum B, Ekbom A, Vatn MH, et al. Change in the extent of colonoscopic and histological involvement in ulcerative colitis over time. Am J Gastroenterol.  1999; 94: 1564– 1569. Google Scholar CrossRef Search ADS PubMed  35. Farrell RJ, Murphy A, Long A, et al. High multidrug resistance (P-glycoprotein 170) expression in inflammatory bowel disease patients who fail medical therapy. Gastroenterology.  2000; 118: 279– 288. Google Scholar CrossRef Search ADS PubMed  36. Kimchi-Sarfaty C, Gribar JJ, Gottesman MM. Functional characterization of coding polymorphisms in the human MDR1 gene using a vaccinia virus expression system. Mol Pharmacol.  2002; 62: 1– 6. Google Scholar CrossRef Search ADS PubMed  37. Wang D, Johnson AD, Papp AC, et al. Multidrug resistance polypeptide 1 (MDR1, ABCB1) variant 3435C>T affects mRNA stability. Pharmacogenet Genomics.  2005; 15: 693– 703. Google Scholar CrossRef Search ADS PubMed  38. Ho GT, Gaya DR, Satsangi J. Multidrug resistance ( MDR1 ) gene in inflammatory bowel disease: key player? Inflamm Bowel Dis.  2005; 11: 1013– 1019. Copyright © 2006 by Lippincott Williams & Wilkins. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Inflammatory Bowel Diseases Oxford University Press

Associations of allelic variants of the multidrug resistance gene (ABCB1 or MDR1) and Inflammatory Bowel Disease and their effects on disease behavior: A case-control and meta-analysis study

Download PDF
10 pages

Loading next page...
 
/lp/oxford-university-press/associations-of-allelic-variants-of-the-multidrug-resistance-gene-Sur6WOYcZE

References (44)

Copyright
Copyright © 2006 by Lippincott Williams & Wilkins.
ISSN
1078-0998
eISSN
1536-4844
DOI
10.1097/01.MIB.0000209791.98866.ba
pmid
16633048
Publisher site
See Article on Publisher Site

Abstract

Background Allelic variants of the ATP-binding cassette, subfamily B member 1 (ABCB1), also known as the multidrug resistance gene (MDR1) that encodes the membrane-bound efflux transporter P-glycoprotein 170 (PGP-170), have been associated with inflammatory bowel disease but with conflicting results. Methods The present study examined the association of ABCB1 C3435T and G2677T/A in a large British case-control cohort of 828 Crohn's disease, 580 ulcerative colitis (UC) cases, and 285 healthy controls. The effect of these variants was further examined with respect to phenotypic and epidemiological characteristics. A meta-analysis was carried out of our results and those from 8 previously published association studies of the C3435T variant in inflammatory bowel disease. Results The 2677T allele was significantly increased in British UC cases compared with controls (45.2% vs. 39.6%; P = 0.034). In particular, the TT genotype was significantly associated with severe UC (odds ratio [OR] 1.90; 95% CI 1.01–3.55) and the use of steroids in UC (OR 1.77; 95% CI 1.08–2.88). No significant association was seen with C3435T and UC, Crohn's disease, or any clinical subgroup. A meta-analysis of 9 association studies of C3435T showed a significant association of the 3435T allele with UC (OR 1.12; 95% CI 1.02–1.23; P = 0.013) but not with CD. Conclusions These results indicate that ABCB1 sequence variants are associated with a small increase in the risk of developing UC and may influence disease behavior. Crohn's disease, ulcerative colitis, association, genetics, meta-analysis Crohn's disease (CD) and ulcerative colitis (UC) are the 2 common clinical subtypes of idiopathic inflammatory bowel disease (IBD) characterized by chronic inflammation of the gastrointestinal tract. The prevalence of IBD in Western countries is ≈63 to 445/100,000,1 and together with its clinical course and high morbidity, its prevalence represents a significant burden to health care. The underlying pathogenesis remains unclear but may involve a combination of persistent bacterial infection, a defective mucosal barrier, or an imbalance in the regulation of the immune system.2 Epidemiological and family studies suggest that genetic factors play a significant role in determining susceptibility to IBD, with a sibling relative risk of 15 to 35 for CD3,4 and 6 to 9 for UC.3,5 Recently, a member of the adenosine triphosphate-binding cassette superfamily ABCB1 (previously known as MDR1), which encodes the membrane transport protein P-glycoprotein (PGP)-170, has been proposed as a candidate gene for IBD on the basis of its location within a region of suggestive linkage detected on chromosome 7q.6 A role for PGP-170 in IBD was suggested by a mouse knockout model for ABCB1 (mdr1a−/−) in which developmentally normal mice spontaneously develop a colitis resembling UC in humans.7 PGP-170 is expressed in several human tissues, including peripheral blood lymphocytes, and is localized to the apical membrane of epithelial cells in the jejunum and colon. In addition to its role in the secretion of drugs, PGP-170 is thought to play a protective role in prevention of cell apoptosis,8 as well as cell-mediated immunity, by promoting migration of antigen-presenting cells from the periphery to lymph nodes to initiate a T lymphocyte response.9 The ABCB1 gene spans 200 kb comprising 28 exons with 2 transcription start sites and has been shown to be highly polymorphic.10,–12 Recent attention has focused on 2 single nucleotide polymorphisms (SNPs), C3435T in exon 26 and a triallelic SNP G2677T/A (Ala893Ser/Thr), which have been shown to correlate with the activity and expression of PGP-170.10,11 However, reports of an association between these 2 SNPs and IBD have been conflicting. The initial study by Schwab et al13 identified a significant association of the 3435T allele with UC but not CD, a finding that has been replicated by some14,15 but not others.16,17 An association also has been reported between the G allele of G2677T/A and IBD in case-control analyses17; this study also found significant overtransmission of the G allele to CD offspring and nonsignificant overtransmission of the same allele to UC offspring in a family-based test of association. Others, however, have found an association with the 2677T allele in UC, with a trend of an association with refractory CD.14 Significant linkage disequilibrium has been observed between these 2 SNPs in different white populations.14,15 More recently, Ho et al15 undertook a genotype-phenotype analysis showing a significant association of C3435T with extensive UC and a higher frequency of the G2677 allele with severe IBD and the need for surgery, although Schwab et al13 previously failed to detect an association of ABCB1 genotypes with specific phenotypes. In view of these conflicting findings and the potential importance of this gene to the understanding of the pathogenesis of IBD, we have investigated the contribution of ABCB1 C3435T and G2677T/A to IBD in a large cohort of British CD and UC patients. We have examined the relationship between genotypic and phenotypic expression, as well as drug exposure and epidemiological factors, in a set of clinically well-characterized patients. Evidence for epistasis between the CD-associated gene CARD15 and ABCB1 in CD patients also was investigated. We have interpreted our association study results in the context of previous studies by performing a comprehensive meta-analysis of 9 independent populations genotyped for C3435T. Materials and Methods Patient Collection A cohort of unrelated British patients with IBD were recruited after ethical review and obtaining informed consent from Guy's and St Thomas' Hospitals and St Mark's Hospital (United Kingdom). The diagnoses of CD and UC were made by established criteria of clinical, radiological, and endoscopic analyses and from histology reports.18,19 Healthy controls were recruited from the greater London region to match the recruitment area and ethnicity of IBD cases. The ethnicity of both cases and controls was ≈94% white and 6% nonwhite; there was no evidence of population stratification in these cohorts in previous studies of > 100 SNP and microsatellite markers. Additional data on site of disease, disease behavior, and therapy were available for a subset of 385 UC and 597 CD patients. Genotyping Genotyping of C3435T (rs1045642) was performed by TaqMan (Applied Biosystems, Warrington, UK) available from ABI assay on demand. G2677T/A was genotyped by Pyrosequencing to determine the allele frequencies of all 3 variants. Genotyping of the CARD15 mutations R702W and L1007fs was also performed by Pyrosequencing, and G908R was performed by TaqMan (ABI assay by design). All of the SNP assays were verified with a subset of at least 48 samples that had been genotyped by DNA sequencing and by checking for Hardy-Weinberg equilibrium in control populations. Sequence and reaction settings are available on request. Phenotypic Assessment UC phenotype was classified by disease extent at diagnosis and latest follow-up (Table 1). We defined extensive disease as disease extending past the splenic flexure, left-sided colitis as disease extending up to the splenic flexure, and proctitis as disease confined to the rectum. When there were discrepancies between the extent of disease as determined by macroscopic and microscopic evidence, histological results were used.15 Severe colitis was determined by the need for surgery either for disease refractory to treatment or for acute severe disease (excluding surgery for dysplasia), as well as for any patient with a severe attack of UC requiring admission to the hospital. Table 1. Demographics and Clinical Characteristics of 385 Non-Jewish Patients With UC     View Large Table 1. Demographics and Clinical Characteristics of 385 Non-Jewish Patients With UC     View Large The clinical characteristics of the CD patients are shown in Table 2. Location was considered the maximum extent of disease involvement at any time before the first surgical resection. Ileal disease was determined as being limited to the terminal ileum with or without spillover into the cecum; colonic disease was any lesion located between the ascending colon and rectum; and ileocolonic disease was defined as any disease of the terminal ileum with or without spillover into the cecum and any location between the ascending colon and rectum. In addition, a further location subset of the presence of perianal disease, defined by the presence of perianal fistulas, ulcers, fissures, and abscesses but not skin tags, was recorded. Disease behavior was classified by current or past behavior subtypes. Penetrating disease was determined by the presence of any abnormal connection between 2 epithelial surfaces or the formation of intraabdominal abscesses or perforations. This was subdivided into intraabdominal fistulas and perianal fistulas, including perianal and ischiorectal abscesses. Any postoperative perforating complications were excluded. Stenotic disease was recorded on the basis of radiological, endoscopic, or histopathological findings, together with the presence of obstructive symptoms and/or evidence of prestenotic dilatation on barium examination. Inflammatory disease was diagnosed in the absence of stenotic or perforating behavior. Surgical history was also recorded in detail and divided into the number of laparotomies with bowel resection and the number of examinations under anesthesia for drainage of perianal sepsis or treatment of fistulas. The presence or absence of granulomas from histopathology specimens also was noted when available. Table 2. Demographics and Clinical Characteristics of 597 Non-Jewish Patients With CD     View Large Table 2. Demographics and Clinical Characteristics of 597 Non-Jewish Patients With CD     View Large For both UC and CD, the requirement for treatment with systemic steroids was assessed and an estimate of the cumulative amount of steroid needed to treat active disease was made by a detailed examination of patients' notes. The need for immunosuppressant therapy also was recorded, as was the need for treatment with biological agents such as infliximab. Other phenotypic details, including age at onset of symptoms, family history, extraintestinal symptoms, and years of follow-up, were noted. Data on smoking habit (current smoker, ex-smoker, or never smoked) were collected at the time of the latest follow-up; cases were stratified into “ever smoked” or “never smoked” for statistical analysis. Statistical Methods Genotype and allele frequencies were compared between cases and controls through the use of χ2 tests. Two-locus haplotype frequencies were estimated by an expectation-maximization algorithm implemented through the use of COCAPHASE, a module of UNPHASED.20 Significance values are reported uncorrected for multiple testing. Linkage disequilibrium between markers was assessed with multiallelic D′21,22 calculated with the GOLD program.23 Meta-analysis Published association studies of ABCB1/MDR1 in IBD were identified through PubMed by a search of the keywords “MDR1,” “ABCB1,” “association,” “inflammatory bowel disease,” “Crohn's disease,” and “ulcerative colitis” in various combinations. A total of 7 publications involving an association study with ABCB1 C3435T were identified; only 3 had genotyped the G2677T/A polymorphism. Two of these performed genotyping by TaqMan, which did not identify the rare 2677A allele.14,15 Thus, individuals heterozygous for this allele appeared homozygous for the opposite allele, and AA homozygous genotypes would not be detected. Consequently, this technique would bias estimates of disease risk in these studies. Therefore, and in view of the low number of studies, the meta-analysis was restricted to the C3435T polymorphism. When full genotype counts were not presented or not provided separately for CD and UC, these data were requested from individual study authors. One study24 was excluded from the meta-analysis because no C3435T CC homozygotes were observed, resulting in significant deviation from Hardy-Weinberg equilibrium. All of the remaining studies were included, regardless of ascertainment method and/or distribution of clinical subphenotypes. When individual studies included > 1 population, each population was treated independently in the meta-analysis. Characteristics of each study are shown in Table 3. After these results were pooled with our own data, a total of 1743 UC cases, 2311 CD cases, and 2931 controls were included in the meta-analysis. Table 3. Published Studies Included in Meta-analysis of MDR1 C3435T     View Large Table 3. Published Studies Included in Meta-analysis of MDR1 C3435T     View Large Genotype and allele-specific odds ratios (ORs) and 95% CIs were calculated from genotype data for each study. ORs were combined across studies using both fixed- and random-effects models. A fixed-effects model assumes that all studies are estimating a single true underlying effect size (i.e., no heterogeneity between studies). A random-effects model assumes that the true effect size in each study comes from a random distribution of effect sizes with fixed mean and variance (i.e., heterogeneity between studies). In view of the low number of studies included in the meta-analysis but large within-study sample sizes, the inverse-weighted method25 was selected for a fixed-effects model26; the random-effects model used was that of DerSimonian and Laird.27 Heterogeneity was assessed by use of the χ2 Q statistic.28 Results Effect of ABCB1 C3435T and G2677T/A Polymorphisms on Disease Susceptibility A total of 828 CD cases, 580 UC cases, and 285 British healthy controls were genotyped for C3435T and G2677T/A. Control and IBD case genotypes for both C3435T and G2677T/A were in Hardy-Weinberg equilibrium. Our combined case collection contained a total of 85 individuals of Jewish ancestry (39 UC, 46 CD). In view of the significant differences in the frequency of ABCB1 polymorphisms between populations29 and particularly Jewish and non-Jewish that have been previously reported,17,30 allele frequencies at the 2 loci were compared between Jewish and non-Jewish cases. We observed significant differences in allele frequencies in British Jewish cases compared with non-Jewish cases (e.g., 3435T: 39.7% versus 55.5%, respectively; P = 0.01; Tables 4 and 5); therefore, the data were analyzed both including and excluding Jewish cases. Table 4. Genotypes and Allele Frequencies of MDR1 G2677T/A in Cases and Controls     View Large Table 4. Genotypes and Allele Frequencies of MDR1 G2677T/A in Cases and Controls     View Large Table 5. Genotypes and Allele Frequencies of MDR1 C3435T in Cases and Controls     View Large Table 5. Genotypes and Allele Frequencies of MDR1 C3435T in Cases and Controls     View Large The frequencies of the ABCB1 G2677T/A alleles were similar in CD cases (G2677, 57.8%; 2677T, 40.8%; 2677A, 1.4%) and controls (G2677, 57.9%; 2677T, 39.6%; 2677A, 2.5%; Table 4). For C3435T, there were no significant differences in the frequency of the 3435T allele between CD cases (51.9%) and controls (53.9%; Table 5). Allele frequencies and ORs for both SNPs were similar when Jewish cases were included. CD patients genotyped for C3435T and G2677T/A were stratified by carriage of the 3 common associated CARD15 mutations to look for evidence of epistasis between the 2 loci. We did not detect any significant difference in the frequency of C3435T in CD cases who carried 1 or 2 CARD15 mutations compared with both controls and CD cases with no CARD15 mutations (Table 5), in agreement with previous studies.13,16 Similarly, we found no evidence of an interaction between G2677T/A and CARD15 (Table 4). Marginally significant differences in allele frequencies of G2677T/A were observed in UC cases compared with controls (G2677: cases, 52.7%; controls, 57.9%; P = 0.050; 2677T: cases, 45.2%; controls, 39.6%; P = 0.034;Table 4). The frequency of the rare allele 2677A was 2.1% in UC cases and 2.5% in controls. For SNP C3435T, we did not detect any significant difference in the frequency of the 3435T allele between UC cases (55.5%) and controls (53.9%). The frequency of the TT genotype, previously reported as exhibiting the strongest association with UC,13,15 also was not significantly different between cases (31.8%) and controls (29.6%). Allele frequencies and ORs for both SNPs were similar when Jewish patients were included (Tables 4 and 5). Because C3435 and G2677T/A are in linkage disequilibrium (multiallelic D′ = 0.85), we performed 2-locus haplotype association tests with UC and CD. We found no significant C3435T/G2677T/A haplotype association with either CD or UC. The T-T haplotype was 42.2% in UC cases compared with 38.6% in controls, but this difference was not significant. The frequency of the T-G haplotype, previously reported to be associated with UC,15 was not significantly different in UC cases (12.0%) and controls (12.9%). Genotype-Phenotype Analysis Detailed phenotypic information was available on 597 CD and 385 UC British non-Jewish cases. Tables 1 and 2 summarize the clinical characteristics of CD and UC patients, respectively. The proportion of males was higher in UC (50.1%) than CD (43.2%). The median age at onset of symptoms for CD and UC patients was 21 and 29 years, respectively, with a median duration of follow-up of 14 years (UC) and 16 years (CD). For UC, 71.9% of our patients had extensive colitis at their latest follow-up. The frequencies of the C3435T and G2677T/A variants in the different phenotypic subgroups were compared with their frequencies in controls to investigate whether the association may be restricted to specific patterns of disease behavior. In CD cases, we did not find any association of C3435T with disease behavior, disease location, treatment, or surgical history (all P values > 0.05; data not shown). The frequency of the 2677 TT genotype was increased in CD cases without ileal disease (OR 1.86; 95% CI 1.09–3.16; P = 0.022), but given the lack of association of 2677T with the overall phenotype of CD and multiple testing of phenotypic subgroups, this finding requires independent replication. In UC cases, the highest risk associated with G2677T/A was observed for the TT genotype in cases with severe disease (OR 1.90; 95% CI 1.01–3.55; P = 0.045) and in patients who had received steroids (OR 1.77; 95% CI 1.08–2.88; P = 0.022; Table 6). In addition, we detected a significant association between the TT genotype and extensive disease (OR 1.67; 95% CI 1.01–2.75; P = 0.045), although a similar nonsignificant risk was observed in 85 cases with only left-sided disease. No association was observed with the presence of extraintestinal manifestations. We did not detect any association of either the T allele or TT genotype of C3435T with disease extent, severity, treatment, or surgical history for UC. In particular, we did not detect an association between the 3435T allele or the TT genotype and extensive colitis. The highest risk of UC associated with the T allele occurred in patients who had received immunosuppressants (OR 1.24; 95% CI 0.98–1.56) and steroid treatment (OR 1.25; 95% CI 0.90–1.73), although neither of these reached statistical significance. Table 6. C3435T and G2677T/A Disease Risks in Clinical Subgroups of 385 UC Patients     View Large Table 6. C3435T and G2677T/A Disease Risks in Clinical Subgroups of 385 UC Patients     View Large Finally, we did not detect an association between C3435T or G2677T/A and age of onset of disease symptoms in either UC or CD. C3435T Meta-analysis The frequency of the 3435T allele, which previously has been reported to be associated with UC,13,–15 was not significantly higher in our cases compared with controls. Therefore, we sought to assess our results in the context of all of the published studies of this SNP by meta-analysis. Control genotype frequencies were in Hardy-Weinberg equilibrium in all studies included in the meta-analysis, and the frequency of the 3435T allele in our control population was within the range of those reported for other non-Jewish populations. Study-specific disease ORs and 95% CIs associated with frequency of the 3435T allele are shown in Figure 1. No significant between-study heterogeneity was observed for UC (Q = 9.42; P = 0.40) or CD (Q = 13.0; P = 0.16). Under a fixed-effects model, meta-analysis of all 9 studies showed a significant association (P = 0.013) of the 3435T allele with UC (OR 1.12; 95% CI 1.02–1.23; Figure 1A). Similar results were observed under a random-effects model (OR 1.12; 95% CI 1.02–1.23). Of the 9 studies, 7 produced a study-specific OR > 1 for the 3435T allele, although only the initial study13 and 1 other study15 were significant. The 95% CIs for the ORs estimated by individual studies were consistent with the pooled OR. Several studies have suggested that a higher disease risk is associated with the TT genotype.13,15,31 The pooled OR for the TT genotype in UC was similar to that for the T allele (OR 1.13; 95% CI 0.98–1.31). Figure 1. View largeDownload slide Meta-analysis for MDR1 C3435T association studies. ORs and 95% CIs are displayed for individual studies. Pooled ORs are shown for fixed- and random-effects models. Figure 1. View largeDownload slide Meta-analysis for MDR1 C3435T association studies. ORs and 95% CIs are displayed for individual studies. Pooled ORs are shown for fixed- and random-effects models. Disease risk estimates reported by initial published associations tend to be inflated.32 Therefore, the influence of the first reported association of ABCB1 and UC13 was assessed by excluding this study from the meta-analysis. Results from a meta-analysis that did not include this initial study were similar to those obtained from analysis of all studies. For example, under a fixed-effects model, the pooled OR for UC was 1.10 (95% CI 1.00–1.21), suggesting that inclusion of the first published association study in the meta-analysis had a minimal influence on the pooled OR. In view of the lower frequency of the T allele (35%) in the Jewish population,17 we also considered the effect of excluding these data from the meta-analysis. The disease risk obtained after excluding the Jewish population was close to that obtained from all studies (fixed-effects model: OR 1.11; 95% CI 1.02–1.22). For CD, the pooled ORs were not significant for either the fixed-effects (OR 0.98; 95% CI 0.91–1.07) or the random-effects (OR 0.99; 95% CI 0.89–1.10) model (Figure 1B). Only 5 of the 9 studies produced a study-specific OR > 1 for the 3435T allele. Discussion Meta-analysis of 9 independent populations has confirmed association between C3435T and UC, with an OR of 1.12 (95% CI 1.02–1.23), and all 9 individual studies identified study-specific risk estimates that were consistent with this small effect. The fact that no association was detected with CD in a study of > 2300 CD cases indicates that this variant of ABCB1 is unlikely to confer an increased risk of this phenotype of IBD. In our own British cohort, we did not find a significant association of C3435T with UC or CD. We did identify significant differences in the frequency of the 3435T allele between Jewish and non-Jewish cases, which is in agreement with the differences between Ashkenazi Jewish and non-Jewish subjects observed by Brant et al17 and Ostrovsky et al30 at this locus, as well as among different world populations.29 However, including Jewish subjects in the meta-analysis did not affect the pooled estimate of risk. The tight CIs obtained in the meta-analysis for both UC and CD indicate that the risk estimates we have derived are likely to be robust. The lack of replication of the association with UC in some studies is consistent with the small size of the effect; our study has only 19% power to detect the association, yet it is the most highly powered study reported to date. This underlines the value of a meta-analysis in detecting effects of this size. To investigate the possibility that the ABCB1 associations were restricted to particular phenotypic features of IBD, we classified a large subset of our cohort into rigorously phenotyped clinical subgroups. We did not find an association of C3435T with any clinical group of CD or UC. We did not replicate the finding by Ho et al15 of an association of extensive UC with the T allele and TT genotype of C3435T, despite the fact that 71.9% (n = 264) of our large cohort of phenotyped UC patients had extensive colitis. The high proportion of UC cases with extensive disease may reflect not only the tertiary care institutions from which cases were recruited but also the use of histological extent of disease. Discrepancies of up to 39% between colonoscopy and histology as regards establishing the distribution of inflammation in the colon have been reported,33 with a suggestion that histological changes are a more sensitive marker of inflammation than endoscopic changes.34 We did find a significant association of G2677T/A with UC. Brant et al17 also found an association of G2677T/A with IBD, although it was with the G allele rather than the T allele observed in both our cohort and as part of the associated haplotype found by Potocnik et al.14 Others have failed to observe an association between G2677T/A and IBD.15 The reasons for the observed differences are unclear; there may be phenotypic heterogeneity between the cohorts (phenotype subgroups are not reported by Brant et al),17 and some studies may lack the statistical power to detect the moderate effect size. We also detected a significant association of the TT genotype of G2677T/A with both severe and extensive UC. A similar estimated risk was observed in patients with left-sided UC, although it was not significant, perhaps because of the smaller sample size (n = 85) in this group. This finding suggests that the association with the TT genotype is not specific to extensive colitis. It is important to acknowledge that the genotype-phenotype correlations reported here do not incorporate statistical correction for multiple testing because the application of a stringent correction to testing of multiple disease classes may result in loss of a signal from a genuine association. Replication of the associations in independent data sets is required to validate these findings. The 2677T allele of ABCB1 (Ser893) has been associated with enhanced activity of PGP-170.11 Our finding of an association of the TT genotype of G2677T/A with severe UC is thus consistent with the observation by Farrell et al35 that high PGP-170 expression is associated with UC patients with severe glucocorticoid-resistant disease. The association of the TT genotype with requirement for steroid therapy is also consistent with their observation that high peripheral blood lymphocyte PGP-170-expressing patients are more likely to require steroids than patients with low PGP-170 expression, although they did not investigate the role of genetic variants of the ABCB1 gene. However, our results appear to be at variance with the association of colitis with ABCB1 deficiency in the mdr1a−/− mouse7 and the hypothesis that low levels of PGP-170 expression increase susceptibility to IBD.15 Whether this reflects an etiological difference between human and animal models of disease remains unclear. Studies investigating expression of PGP-170 variants are conflicting, with Kimchi-Sarfaty et al36 failing to demonstrate any difference in cell surface distribution and function of PGP-170 in G2677T/A variants compared with wild-type. Recently, it has been reported that the 3435T allele of ABCB1 is associated with a moderate reduction in mRNA stability.37 It is also possible that an indirect association with disease susceptibility exists with both C3435T and G2677T/A that is caused by linkage disequilibrium with the true causal variant, which may affect the expression of PGP-170.38 If so, the true degree of disease risk attributable to ABCB1 may be higher than that associated with either of these polymorphisms. Sequencing of the coding regions of the ABCB1 gene in IBD patients revealed 2 intronic SNPs in linkage disequilibrium with the risk haplotype, but their possible effect on PGP-170 expression is not known.14 Conclusions The meta-analysis we have carried out supports the results of several individual studies that DNA sequence variation in ABCB1 influences susceptibility to UC but suggests that the size of the effect is small. The C3435T variant, however, does not appear to confer susceptibility to CD, but the other variants need to be tested in a sample of comparable size to resolve the question of the involvement of ABCB1 in CD. These results may have important implications for the organization of studies that investigate the association of candidate genes that may have real but moderate effects on disease risk in IBD. In addition, our finding of an association between the 2677T allele and severe UC suggests that ABCB1 may influence disease behavior and would be worthy of follow-up in other cohorts. An analysis of ABCB1 expression and function in primary cells of cases and controls with defined ABCB1 genotypes would help to clarify the role of sequence variation in this gene in disease susceptibility and behavior in IBD.38 Acknowledgments We thank Dr Uros Potocnik (University of Ljubljana, Ljubljana, Slovenia), Dr Judy Cho (University of Chicago, Chicago, Ill), and Dr Peter Croucher (University of Kiel, Kiel, Germany) forproviding additional information for the meta-analysis. All of theother data were extracted directly from publications. References 1. Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease: incidence, prevalence, and environmental influences. Gastroenterology.  2004; 126: 263– 271. CrossRef Search ADS PubMed  2. Sartor RB. Clinical applications of advances in the genetics of IBD. Rev Gastro Dis.  2003; 3: S9– S17. 3. Kuster W, Pascoe L, Purrmann J, et al. The genetics of Crohn disease: complex segregation analysis of a family study with 265 patients with Crohn disease and 5,387 relatives. Am J Med Genet.  1989; 32: 105– 108. Google Scholar CrossRef Search ADS PubMed  4. Meucci G, Vecchi M, Torgano G, et al. Familial aggregation of inflammatory bowel disease in northern Italy: a multicenter study: the Gruppo di Studio per le Malattie Infiammatorie Intestinali (IBD Study Group). Gastroenterology.  1992; 103: 514– 519. Google Scholar CrossRef Search ADS PubMed  5. Orholm M, Munkholm P, Langholz E, et al. Familial occurrence of inflammatory bowel disease. N Engl J Med.  1991; 324: 84– 88. Google Scholar CrossRef Search ADS PubMed  6. van Heel DA, Fisher SA, Kirby A, et al. Inflammatory bowel disease susceptibility loci defined by genome scan meta-analysis of 1952 affected relative pairs. Hum Mol Genet.  2004; 13: 763– 770. Google Scholar CrossRef Search ADS PubMed  7. Panwala CM, Jones JC, Viney JL. A novel model of inflammatory bowel disease: mice deficient for the multiple drug resistance gene,mdr1a, spontaneously develop colitis. J Immunol.  1998; 161: 5733– 5744. Google Scholar PubMed  8. Smyth MJ, Krasovski E, Sutton VR, et al. The drug efflux protein, P-glycoprotein, additionally protects drug-resistant tumor cells from multiple forms of caspase-dependent apoptosis. Proc Natl Acad Sci U S A.  1998; 95: 7024– 7029. Google Scholar CrossRef Search ADS PubMed  9. Randolph GJ, Beaulieu S, Pope M, et al. A physiologic function for p-glycoprotein (MDR-1) during the migration of dendritic cells from skin via afferent lymphatic vessels. Proc Natl Acad Sci U S A.  1998; 95: 6924– 6929. Google Scholar CrossRef Search ADS PubMed  10. Hoffmeyer S, Burk O, von Richter O, et al. Functional polymorphisms of the human multidrug-resistance gene: multiple sequence variations and correlation of one allele with P-glycoprotein expression and activity in vivo. Proc Natl Acad Sci U S A.  2000; 97: 3473– 3478. Google Scholar CrossRef Search ADS PubMed  11. Kim RB, Leake BF, Choo EF, et al. Identification of functionally variant MDR1 alleles among European Americans and African Americans. Clin Pharmacol Ther.  2001; 70: 189– 199. Google Scholar CrossRef Search ADS PubMed  12. Cascorbi I, Gerloff T, Johne A, et al. Frequency of single nucleotide polymorphisms in the P-glycoprotein drug transporter MDR1 gene in white subjects. Clin Pharmacol Ther.  2001; 69: 169– 174. Google Scholar CrossRef Search ADS PubMed  13. Schwab M, Schaeffeler E, Marx C, et al. Association between the C3435T MDR1 gene polymorphism and susceptibility for ulcerative colitis. Gastroenterology.  2003; 124: 26– 33. Google Scholar CrossRef Search ADS PubMed  14. Potocnik U, Ferkolj I, Glavac D, et al. Polymorphisms in multidrug resistance 1 (MDR1) gene are associated with refractory Crohn disease and ulcerative colitis. Genes Immun.  2004; 5: 530– 539. Google Scholar CrossRef Search ADS PubMed  15. Ho GT, Nimmo ER, Tenesa A, et al. Allelic variations of the multidrug resistance gene determine susceptibility and disease behavior in ulcerative colitis. Gastroenterology.  2005; 128: 288– 296. Google Scholar CrossRef Search ADS PubMed  16. Croucher PJ, Mascheretti S, Foelsch UR, et al. Lack of association between the C3435T MDR1 gene polymorphism and inflammatory bowel disease in two independent Northern European populations. Gastroenterology.  2003; 125: 1919– 1920. Google Scholar CrossRef Search ADS PubMed  17. Brant SR, Panhuysen CI, Nicolae D, et al. MDR1 Ala893 polymorphism is associated with inflammatory bowel disease. Am J Hum Genet.  2003; 73: 1282– 1292. Google Scholar CrossRef Search ADS PubMed  18. Lennard-Jones JE. Classification of inflammatory bowel disease. Scand J Gastroenterol.  1989; 170: 2– 6. Google Scholar CrossRef Search ADS   19. Podolsky DK. Inflammatory bowel disease. N Engl J Med.  1991; 325: 928– 937. CrossRef Search ADS PubMed  20. Dudbridge F. Pedigree disequilibrium tests for multilocus haplotypes. Genet Epidemiol.  2003; 25: 111– 115. 21. Lewontin RC. The interaction of selection and linkage, I: general considerations; heterotic models. Genetics.  1964; 49: 49– 67. PubMed  22. Hedrick P. Gametic disequilibrium measures: proceed with caution. Genetics.  1987; 117: 331– 341. Google Scholar PubMed  23. Abecasis GR, Cookson WOC. GOLD: graphical overview of linkage disequilibrium. Bioinformatics.  2000; 16: 182– 183. Google Scholar CrossRef Search ADS PubMed  24. Gazouli M, Zacharatos P, Gorgoulis V, et al. The C3435T MDR1 gene polymorphism is not associated with susceptibility for ulcerative colitis in Greek population. Gastroenterology.  2004; 126: 367– 369. Google Scholar CrossRef Search ADS PubMed  25. Sutton AJ, Brams KR, Jones DR, et al Methods for Meta-Analysis in Medical Research . Chichester, UK: Wiley; 2000. 26. Fleiss JL. Statistical Methods for Rates and Proportions  ( 2nd ed.). New York: Wiley; 1981. 27. DerSimonian R, Laird N. Meta analysis in clinical trials. Controlled Clin Trials.  1986; 7: 177– 188. Google Scholar CrossRef Search ADS   28. Cochran WG. The combination of estimates from different experiments. Biometrics.  1954; 10: 101– 129. CrossRef Search ADS   29. Tang K, Wong LP, Lee EJ, et al. Genomic evidence for recent positive selection at the human MDR1 gene locus. Hum Mol Genet.  2004; 13: 783– 797. Google Scholar CrossRef Search ADS PubMed  30. Ostrovsky O, Nagler A, Korostishevsky M, et al. Genotype and allele frequencies of C3435T polymorphism of the MDR1 gene in various Jewish populations of Israel. Ther Drug Monit.  2004; 26: 679– 684. Google Scholar CrossRef Search ADS PubMed  31. Glas J, Torok HP, Schiemann U, et al. MDR1 gene polymorphism in ulcerative colitis. Gastroenterology.  2004; 126: 367. Google Scholar CrossRef Search ADS PubMed  32. Ioannidis JPA, Ntzani EE, Trikalinos TA, et al. Replication validity of genetic association studies. Nat Genet.  2001; 29: 306– 309. Google Scholar CrossRef Search ADS PubMed  33. Niv Y, Bat L, Ron E, et al. Change in the extent of colonic involvement in ulcerative colitis: a colonoscopic study. Am J Gastroenterol.  1987; 82: 1046– 1051. Google Scholar PubMed  34. Moum B, Ekbom A, Vatn MH, et al. Change in the extent of colonoscopic and histological involvement in ulcerative colitis over time. Am J Gastroenterol.  1999; 94: 1564– 1569. Google Scholar CrossRef Search ADS PubMed  35. Farrell RJ, Murphy A, Long A, et al. High multidrug resistance (P-glycoprotein 170) expression in inflammatory bowel disease patients who fail medical therapy. Gastroenterology.  2000; 118: 279– 288. Google Scholar CrossRef Search ADS PubMed  36. Kimchi-Sarfaty C, Gribar JJ, Gottesman MM. Functional characterization of coding polymorphisms in the human MDR1 gene using a vaccinia virus expression system. Mol Pharmacol.  2002; 62: 1– 6. Google Scholar CrossRef Search ADS PubMed  37. Wang D, Johnson AD, Papp AC, et al. Multidrug resistance polypeptide 1 (MDR1, ABCB1) variant 3435C>T affects mRNA stability. Pharmacogenet Genomics.  2005; 15: 693– 703. Google Scholar CrossRef Search ADS PubMed  38. Ho GT, Gaya DR, Satsangi J. Multidrug resistance ( MDR1 ) gene in inflammatory bowel disease: key player? Inflamm Bowel Dis.  2005; 11: 1013– 1019. Copyright © 2006 by Lippincott Williams & Wilkins.

Journal

Inflammatory Bowel Diseases – Oxford University Press

Published: Apr 1, 2006

There are no references for this article.