SEARCH

SEARCH BY CITATION

Keywords:

  • Genetics;
  • Human leukocyte antigens;
  • Crohn's disease;
  • Ulcerative colitis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

The aim of this study was to identify major histocompatibility complex alleles associated with the development and clinical features of inflammatory bowel disease (IBD). Genotyping at the human leukocyte antigen (HLA) DRB1 and DQB1 loci was performed on individuals from 118 Caucasian IBD sibling pair families and on 216 healthy controls. Both population- and family-based association tests were used to analyze data obtained on the entire study population and on clinical subgroups stratified by diagnosis, ethnicity, and disease distribution. HLA DRB1*0103 was significantly associated with IBD (OR = 6.0, p = 0.0001) in a case–control analysis of non-Jewish IBD-affected individuals. This association was apparent among both Crohn's disease (OR = 5.23, p = 0.0007) and ulcerative colitis (OR = 7.9, p = 0.0001) patients and was confirmed in the non-Jewish IBD population by results of family-based association analysis using the transmission disequilibrium test. HLA DQB1*0501 was also associated with IBD (OR = 1.64, p = 0.02) in the non-Jewish population, but statistically significant association of this allele with disease was not detected for Crohn's disease and ulcerative colitis separately. No significant associations were identified among the Jewish patients. In the non-Jewish IBD families, IBD was as strongly associated with the DRB1*0103 DQB1*0501 haplotype as with the DRB1*0103 allele alone. The carrier frequency of the DRB1*0103 allele was found to be 10-fold higher in Crohn's disease patients with pure colonic involvement than in healthy controls (38.5% vs. 3.2%; p = 0.0002). These data demonstrate the association of the HLA DRB1*0103 allele with both Crohn's disease and ulcerative colitis and with large intestine–restricted disease in non-Jewish IBD patients and therefore identify HLA DRB1*0103 as a potentially important contributor to disease susceptibility and to expression of colonic involvement in IBD.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Data from epidemiological studies have indicated that genetic factors play an important role in the pathogenesis of idiopathic inflammatory bowel disease (IBD) (1). To delineate the gene variants contributing to IBD susceptibility, a number of groups have undertaken microsatellite marker–based genome-wide scans of IBD-affected sibling pair families, the goal being to localize and then positional clone the genes of interest (2–8). Another approach to IBD gene identification has involved the search for disease association with specific alleles of given candidate genes. Both of these methods have recently been successfully used to identify NOD2/CARD15 as a susceptibility gene for Crohn's disease (CD) (9,10).

Among the candidate genes studied to date in relation to IBD, the major histocompatibility complex (MHC) genes encoding human leukocyte antigen (HLA) proteins are of particular relevance because of their major role in regulating the immune response (11) and their well-established association with such immunologic conditions as rheumatoid arthritis (12), uveitis (13), diabetes (14), and celiac disease (15). HLA specificities have also been shown in many studies to correlate with ulcerative colitis (UC), CD, and the expression of extraintestinal manifestations in IBD patients (16–30). In addition, linkage data from our group and others have revealed IBD to be linked to the MHC gene–containing region on chromosome 6p (the IBD3 locus) (8,31,32). While these data are consistent with a key role for HLA proteins in IBD susceptibility, the extent to which specific HLA alleles associate with IBD remains unclear, with the results differing considerably from one study to another. Thus, for example, a significant association between UC and HLA DR2 or DRB1*1501 has been detected in some studies but has not been replicated in several others (17,21,29). The reasons for such disparities are not completely clear but are likely to reflect such factors as variability in typing techniques (serological vs. DNA-based typing), inadequate sample sizes, disparate statistical methodology (e.g., failure to correct results for multiple tests), poorly matched control populations, and type I error. Genetic and clinical heterogeneity are also likely to represent major complicating factors as results from several groups have shown that phenotypic variations, such as perinuclear antineutrophil cytoplasmic antibody (pANCA) positivity, disease severity, and extraintestinal manifestations, may define clinical subgroups with distinct profiles of HLA susceptibility alleles (18,27,30).

To address these issues and elucidate the individual HLA alleles that predispose to IBD, we have used a cohort of Canadian Caucasian IBD sibling pair families to evaluate association of HLA class 2 genes with CD and UC. HLA genotyping of this IBD population has allowed us to use both case–control and family-based association analyses to explore the extent to which specific HLA DRB1 and DQB1 alleles contribute to IBD susceptibility and course.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Subjects

The IBD families studied were recruited from the Mount Sinai Hospital IBD Center in Toronto and represent a cohort of Caucasian IBD sibling pair families ascertained for the purposes of a microsatellite marker–based genome scan (8). In 97 of these 118 families, both parents of the affected probands were available for study. Affection status as well as severity, extent, and sites of disease were established by review of the clinical, endoscopic, radiological, and histological reports using standard criteria. As the patients were taken exclusively from the MSH IBD Center, all charts were retrieved for review and update of the clinical data. Jewish ethnicity was defined as an individual with at least three Jewish grandparents. A group of 216 healthy, unrelated, Caucasian, non-Jewish individuals from the Toronto population served as a matched control group. HLA DRB1 and DQB1 allele frequencies previously published for 132 Ashkenazi Jewish individuals were also used as a control data set (33).

Molecular HLA Genotyping

Genomic DNA was extracted from venous blood samples by standard methods. HLA genotyping at the DRB1 and DQB1 loci was carried out by PCR amplification of DNA using locus- and group-specific primers, and the PCR amplicons were then identified by sequence-specific oligonucleotide probes (34). The presence of DNA sequences representing specific HLA alleles was further resolved using either sequence-specific primers and/or sequence-based typing on an automated sequencer (35).

Statistical Methods

Population-Based Association

Population associations were evaluated first by comparing HLA specificity and allele frequencies in IBD patient groups and ethnically matched control groups using two-tailed Fisher exact tests. Several HLA specificities and alleles were initially selected for a priori hypothesis testing based on previously reported associations with IBD (DR1, DR3, DR4, DR7, DQ1, DQ4 for CD and DR1, DR2, DR4 for UC); for specificities/alleles not classified as a priori, p values were corrected for multiple testing using a Bonferroni adjustment. Specificities/alleles that showed significant frequency differences at the 5% significance level were further analyzed using a logistic regression model with patient/control status considered a covariate and presence/absence of a specific allele as a binary response. The Haldane correction was applied for specificity/allele frequencies of zero (36). A generalized estimating equations (GEE) method was applied to adjust the logistic regression results for familial correlation among siblings using the SAS GENMOD procedure. Observations from the same family were treated as a cluster, assuming an independence working correlation and standard errors were calculated using a robust empirical method (37–40).

The logistic regression GEE method used compares the log odds of carrying allele DRB1*0103 in cases from sibling pair families to the comparable log odds in independent controls. The log odds for the cases is unaffected by having siblings because the odds is a ratio of counts (provided all families have the same probability of being included in the sample), but the usual variance estimate may be too small. The variance for the difference of log odds, used in constructing the test statistic and the confidence interval, consists of two variance terms, one for the cases and one for the controls. The GEE adjustment for familial correlation among sibling pairs involves inflation of the case variance term by a factor (1 + r) where r depends on the correlation between genotypes for a pair of affected siblings. If the sibling genotypes are completely uncorrelated (which is unlikely), then the variance is unchanged and the effective number of cases is twice the number of sibling pairs, assuming one sibling pair per family. However, if sibling pair genotypes were completely correlated, then r would be 1 and the variance would be doubled so that the effective number of cases would be the number of sibling pair families and the result would be equivalent to using only one sibling from each family. In practice, the genotype correlation will be between 0 and 1, and the effective number of cases will be reduced. For example, for the combined UC and CD non-Jewish families, the value of the variance inflation factor (1 + r) is 1.88, and the information in 169 cases from 79 families with at least one sibling pair is approximately equivalent to 88 independent cases. Therefore, if genotyping costs have already been incurred, then some additional information can be obtained from utilizing additional siblings.

For analysis of HLA DRB1*0103 association with disease severity and localization, Fisher exact tests and GEE logistic regression were used to compare carrier frequencies of genotypes involving HLA DRB1*0103 among groups.

Family-Based Association

Confirmatory family-based association analyses were also performed for alleles DRB1*0103 and DQB1*0501 and the haplotype DRB1*0103 DQB1*0501 in families in which both parents were available by using the transmission disequilibrium test (TDT). These alleles were analyzed as biallelic markers with all other alleles combined. The TDTEX option of the software package SAGE (S.A.G.E., Release 3.1; Case Western Reserve University, Cleveland, OH) was used to analyze siblings as either independent individuals or as a sibling pair (SIBS option of TDTEX). The exact distribution of the test statistic was approximated via a Markov chain Monte Carlo algorithm (41). Tests of association that included all siblings were verified using the software program FBAT, which applies an empirical adjustment for the correlation induced by linkage (42). Two locus HLA DRB1 DQB1 haplotypes transmitted and nontransmitted from parents to affected offspring were also tabulated using the haplotypes constructed in GENEHUNTER 2.0 (43), and exact p values to test the transmission probabilities of the DRB1*0103 DQB1*0501 haplotype were calculated from the binomial distribution.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

The population studied here included 118 Caucasian IBD sibling pair families ascertained for the purposes of IBD susceptibility gene discovery. As shown in Table 1, these families contained 238 affected siblings, of whom 186 were diagnosed with CD and 52 were diagnosed with UC. Fifty-nine patients from 29 CD families were of Ashkenazi Jewish descent and were analyzed independently of the non-Jewish families/individuals. There were insufficient numbers of Jewish individuals with UC to be included in this study.

Table Table 1. Study population and distribution by diagnostic subtype and ethnicity
  No families with 
  1. CD, Crohn's disease; NJ, not of Jewish descent; J, of Ashkenazi Jewish descent; UC, ulcerative colitis.

Family typeNo. families1 affected sibling2 affected siblings≥ 3 affected siblingsNo. affected siblings
CD NJ614507127
CD J29126259
UC NJ28620252
Total118119611238

All of the patients included in the study were initially typed for HLA DRB1 specificities. The results of this analysis revealed HLA DR1 to be positively associated with IBD as a whole (19.1% vs. 11.1%; p = 0.01) and CD (17.5% vs. 11.1%; p = 0.04) and UC (23.1% vs. 11.1%; p = 0.01) in non-Jewish individuals. By contrast, as consistent with previously reported data, HLA DR4 was found to be negatively associated with the non-Jewish UC population (6.7% vs. 19.9%; p = 0.003). No significant differences in the distribution of HLA DR specificities or alleles were apparent in a comparison of Jewish IBD patients with matched controls.

To further investigate the association between IBD and the HLA DRB1 locus, the frequencies of HLA DRB1 alleles were compared between patients and ethnically matched controls (Table 2). The results of this analysis revealed HLA DRB1*0103 to be significantly associated with IBD in non-Jewish individuals (9.0% vs. 1.6%; OR = 6.0, p = 0.0001). Importantly, this allele was also positively associated with both CD (7.9% vs. 1.6%; OR = 5.23, p = 0.0007) and UC (11.5% vs. 1.6%; OR = 7.9, p = 0.0001). All HLA DRB1*0103-positive individuals were heterozygous at this locus. No significant association was detected between IBD and any of the other HLA DRB1 alleles (including subtypes of HLA DR4).

Table Table 2. Frequency distribution of HLA DRB alleles among cases and controls
 CasesControls
 IBD NJUC NJCD NJCD JHealthy NJHealthy J
DRB allelen%n%n%n%n%n%
  • a,b, a

    aa priori for CD.

  • a,b, b

    ba priori for UC.

  • c

    cp = 0.0001.

  • d

    dp = 0.0007.

  • e

    ep = 0.0001.

  • HLA. human leukocyte antigen; IBD. inflammatory bowel disease; NJ. not of Jewish descent; UC. ulcerative colitis; CD, Crohn's disease; J. of Ashkenazi Jewish descent.

0101329.01211.5207.910.9337.652.0
010241.100.041.697.781.9259.9
0103a,b329.0c1211.5d207.9e00.071.631.2
0301a3710.487.72911.543.44710.9176.7
030200.000.000.000.010.200.0
0401277.632.9249.532.6429.741.6
040210.300.010.4119.440.9228.7
040310.311.000.032.630.752.0
0404133.721.9114.400.0225.141.6
0405a72.000.072.810.961.410.4
040730.800.031.200.061.400.0
040810.311.000.000.020.552.0
0410a00.000.000.000.010.200.0
07014412.4109.63413.5119.46715.53212.6
1301215.9109.6II4.454.3225.1104.0
1302a205.654.8156.097.71.12.5124.7
130330.811.020.843.481.900.0
1501a,b3610.1II10.6259.943.44911.372.8
1502b41.143.800.054.351.283.2
150300.000.000.000.010.2o0.0
1601b10.300.010.400.071.631.2
160230.811.020.800.010.220.8

HLA DQB1 specificities and alleles were also studied in the affected individuals. The results of these analyses revealed a significant positive association between HLA DQ1 and UC in the non-Jewish population (56.7% vs. 38.4%; p = 0.04). When individual alleles were evaluated in this population, the HLA DQB1*0501 allele was found to be significantly associated with IBD (19.7% vs. 13.0%; OR = 1.64, p = 0.02) but not with CD and UC individually (Table 3). Similarly, no other individual DQB alleles were found to be significantly associated with disease. Again, among the Jewish individuals studied, no significant differences in the frequencies of specific HLA DQ types or alleles were identified between affected individuals and controls. The results shown in Table 4 provide a summary of the HLA specificities and alleles showing positive or negative association with IBD in this study. These results are conservatively calculated utilizing corrections for multiple tests, and analysis of related individuals (siblings) and are in accord with other published data suggesting that HLA DR1 and the DRB1*0103 allele are associated with IBD (21,28,29).

Table Table 3. Frequency distribution of HLA DQB alleles among cases and controls
 CasesControls
 IBD NJUC NJCD NJCD JHealthy NJHealthy J
DRB allelen%n%n%n%n%n%
  • a

    a a priori for CD.

  • b

    bp = 0.02.

  • HLA. human leukocyte antigen: IBD. inflammatory bowel disease: NJ, not of Jewish descent; UC. ulcerative colitis; CD. Crohn's disease: J. of Ashkenazi Jewish descent.

0401a00.000.000.000.010.200.0
0402a123.421.9104.054.2133.031.3
0501a7019.7b2524.04517.91411.95613.03615.1
050220.611.010.400.0112.572.9
050330.832.900.032.5112.572.9
060141.143.800.054.261.493.8
0602a3710.41110.62610.332.55011.6114.6
0603a205.6109.6104.054.2133.0166.7
0604164.543.8124.865.181.920.8
Table Table 4. Statistics for selected HLA specificities/alleles with significant difference between cases and controls
Specificity/allelePopulationCase frequency (%)Control frequency (%)Adjusted p valueBAFOdds ratio95% confidence interval
  1. BAF, Bonferroni adjustment factor; CD. Crohn's disease; NJ, not of Jewish descent; UC, ulcerative colitis; IBD, inflammatory bowel disease.

DR1CD NJ17.511.10.0411.691.02, 2.82
 UC NJ23.111.10.0112.401.20, 4.79
 IBD NJ19.111.10.0111.891.19, 2.99
DR4UC NJ6.719.90.00310.290.13, 0.66
DRB 1*0103CD NJ7.91.60.000715.232.02, 13.63
 UC NJ11.51.60.000117.922.76, 22.71
 IBD NJ9.01.60.000116.002.49, 14.44
DQ1UC NJ56.738.40.0442.101.19, 3.72
DQB 1*0501CD NJ17.913.00.1211.460.90, 2.35
 UC NJ24.013.00.56202.121.08, 4.16
 IBD NJ19.713.00.0211.641.06, 2.53

Because of the availability of sibling pair families for this study and because family-based association analysis diminishes confounding effects of population stratification, the potential role of the HLA DRB1 and DQB1 alleles identified by the population-based analysis was also evaluated using the TDT. When applied to multiple affected siblings from the same families, the TDT incorporates the assumption that allele transmissions to siblings are independent events. This assumption is valid in the absence of linkage but does not necessarily apply if linkage is present. The TDT analyses in this report, therefore, were conducted by considering each affected sibling independently or by treating the sibling pair as the unit of analysis and utilizing only heterozygous parents that transmit the same allele to both affected offspring (41). As shown in Table 5, analysis of the non-Jewish IBD families revealed significant excess transmission to affected offspring for the HLA DRB1*0103 allele as well as for the HLA DQB1*0501 allele. Interestingly, analysis of two-locus haplotypes constructed by GENEHUNTER 2.0 revealed that DRB1*0103 occurred nearly exclusively with DQB1*0501, and TDT analysis revealed the DRB1*0103 DQB1*0501 haplotype to be as strongly associated with IBD as the DRB1*0103 allele alone (Table 5).

Table Table 5. Results of family-based association with exact p values for the transmission disequilibrium test statistic calculated with TDTEX in SAGE
  All siblingsaSibling pailb
Allele/haplotypePopulationTNTPTNTp
  • TDTEX is an option in the SAGE software package (Case Western Reserve University. Cleveland, OH, U.S.A.)

  • a

    aAll available siblings are analyzed.

  • b

    bUnit of analysis is the sibling pair.

    T. number of alleles transmitted; NT, the number of alleles not transmitted; CD, Crohn's disease; NJ, not of Jewish descent; UC, ulcerative colitis; IBD. inflammatory bowel disease; J, of Ashkenazi Jewish descent.

DRB1*0103CD NJ1640.01600.03
 UC NJ620.29200.50
 IBD NJ2260.004800.01
 CD J0000
DQB 1*0501CD NJ34180.041020.04
 UC NJ1680.15510.22
 IBD NJ50260.011530.01
 CD J1170.48310.62
DRB1*0103/DQB 1*0501CD NJ1330.025(l0.06
 UC NJ620.29200.50
 IBD NJ1950.007700.02
 CD J0000

Because of the strong associations detected here between the HLA DRB1*0103 allele and both CD and UC, the hypothesis that this allele associates with severity, extent, and/or localization of disease was also investigated by analysis of 169 CD patients from 90 families in whom the relevant clinical data was available. The analysis in both non-Jewish and Jewish individuals with CD revealed the HLA DRB1*0103 allele to be present in 5 of 14 individuals (35.7%) with disease restricted to the colon (“pure” colonic CD) as compared with 15 of 155 (9.7%) CD patients with only small bowel involvement or with small bowel and colonic involvement (OR = 5.2, p = 0.02;Table 6). When the non-Jewish individuals with CD were evaluated separately, 5 of 13 (38.5%) with pure colonic disease carried HLA DRB1*0103 versus only 3.2% of health controls (OR = 18.7, p = 0.0002). As individuals with UC by definition have a pure colonic disease phenotype, 52 UC patients for whom information on HLA DRB1 alleles was available were analyzed by grouping them together with CD patients with pure colonic disease, creating a large bowel–only-affected IBD phenotypic subgroup. In this analysis, the HLA DRB1*0103 allele was found to occur in 26.2% (17/65) of non-Jewish IBD-affected individuals with colonic disease only (5 with CD and 12 with UC) as compared with 3.2% of healthy controls (OR = 10.6, p = 0.0001). By contrast, the presence of HLA DRB1*0103 did not distinguish between disease severity (defined by those individuals with ≥3 intestinal sites affected or number of past surgical procedures), extent (as measured by the number of intestinal sites affected), or sites of involvement other than the large intestine.

Table Table 6. Frequencies of individuals carrying HLA DRB1*0103 by site of disease
 No. and frequency of DRB*0103Case vs. control GEE regression analysis
GroupCase phenotypeNo. (%)Control groupNo. (%)POR95% CI
  • a

    aAll CD individuals for whom phenotypic data was available.

  • b

    bCD confined to large bowel only (i.e. pure colonic CD with no small bowel involvement; for IBD NJ analysis all individuals with UC were included in this group).

  • c

    cIndividuals with pure SB CD and those with SB and colonic Crohn's disease included.

  • HLA. human leukocyte antigen: GEE, generalizing estimation equations; OR. odds ratio; CI, confidence interval; CD, Crohn's disease; NJ, not of Jewish descent; IBD, inflammatory bowel disease; LB. large bowel; SB. small bowel.

All CDa (n = 169)Only LBb (n = 14)5(35.7)SB ± LB CDc (n = 155)15 (9.7)0.025.21.3, 20.1
CD NJ (n = 114)Only LBb (n = 13)5 (38.5)SB ±LBCDc (n = 101)15(14.9)0.073.60.9, 14.5
 Only LBb(n = 13)5 (38.5)Healthy (n = 216)7 (3.2)0.000218.74.0, 86.8
IBD NJ (n = 167)Only LBb (n = 65)17 (26.2)SB + LBCDc (n = 101)15 (14.9)0.12.00.8, 5.3
 Only LBb (n = 65)17(26.2)Healthy (n = 216)7 (3.2)0.000110.63.7, 30.0

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

In the study presented here, HLA DRB1*0103 was found to be associated with both CD and UC in a non-Jewish Caucasian IBD patient population. This is consistent with data identifying an association between the HLA DRB1*0103 allele and UC (21,27,28) and with results suggesting this allele is also associated with CD (29). Moreover, Ahmad et al. (44) have recently reported important associations of HLA alleles with colonic and perianal CD, in contrast to ileal CD, where NOD2/CARD15 alleles play a predominant role. The current observations extend these data by providing evidence showing HLA DRB1*0103 to be significantly associated with pure colonic disease in non-Jewish CD patients and with the presence of a large bowel disease–only phenotype in the non-Jewish IBD population as a whole. In addition, the TDT analysis provides further evidence of HLA DRB1*0103 association with IBD in non-Jewish families and also reveals this association to involve a two-locus haplotype comprised of the HLA DRB1*0103 and DQB1*0501 alleles, a result which also confirms the previous report from Trachtenberg et al. (29). No HLA–disease associations were detected in the Jewish IBD population studied here, a result which may reflect insufficient patient numbers. Taken together, these data, which were generated using both population- and family-based association methods and by applying stringent corrections for multiple tests and for familial correlation, provide compelling evidence of a role for HLA DRB1*0103 in predisposing to IBD and, in particular, to the phenotype of colonic disease.

Although the relevance of HLA specificities to IBD has been primarily investigated via association studies, a role for these genes in IBD has also been implied by linkage data emanating from genome scans of IBD sibling pair families. Thus, for example, the results of a genome-wide scan from our center have revealed possible linkage between IBD and a region on chromosome 6p (IBD3) in proximity to the MHC class 2 region (peak LOD score = 2.3) (8). IBD linkage to this chromosomal region has also been detected in several other independent populations (31,32) but, importantly, has not been found in all such analyses (4,7,45). A preliminary evaluation to determine the degree to which the association data described here account for the linkage evidence found in our previous genome-wide scan was carried out by stratification of families by the presence or absence of HLA DRB1*0103. This analysis indicated that the excess allele sharing at the IBD3 locus seen in our patient population could not be completely explained by the association with DRB1*0103 as excess sharing was still present in DRB1*0103-negative families (data not shown). Thus, while the current population- and family-based data demonstrating IBD to be associated with a specific HLA DR allele strongly support a role for MHC genes in IBD susceptibility, it appears likely that reported variations in the detection of HLA association and linkage with IBD reflect, at least in part, real disparities in the extent to which HLA genes contribute to IBD in different populations and may also be reflective of an additional gene in linkage disequilibrium with HLA DRB1 in this region.

At present, the mechanisms whereby MHC class 2 gene products may predispose to IBD are not defined. However, the association of the HLA DRB1*0103 allele in particular with IBD is of interest in that this allele encodes an HLA DR molecule with a unique sequence within the pocket of the peptide-binding groove. This structural distinction from other HLA DRB1 alleles may imbue HLA DRB1*0103 molecules with the capacity to bind and present distinct and potentially disease-associated peptides and may therefore account for the association of this allele with IBD (29). It is also possible, however, that the association of IBD with HLA alleles does not reflect these alleles per se but rather linkage disequilibrium between particular HLA alleles and a non–HLA-related IBD-causative gene within or near the MHC class 2 region. This area contains a number of immunoregulatory genes that represent credible candidates for IBD susceptibility genes. These include, for example, genes encoding peptide transporters and other proteins involved in antigen processing, complement components, tumor necrosis factor α, and the inhibitor of κB-like (IKBL) gene, which encodes a protein structurally resembling the cytokine transcription regulator, IκBα. A polymorphism within the IKBL gene has recently been shown to be associated with extensive and more severe UC, and as this allele appears to be in linkage disequilibrium with HLA DRB1*0103, it has been suggested that the HLA DRB1*0103 allele may confer susceptibility to UC, while the IKBL polymorphism may predispose to more severe disease (46). It therefore remains to be determined whether and how variants of these or other genes within and around the MHC region interact with HLA alleles such as DRB1*0103 to influence susceptibility to IBD.

In the current study, carrier frequency of the HLA DRB1*0103 allele was found to be increased by 10-fold in CD patients with pure colonic disease as compared with the frequency in healthy controls. These observations indicate that this allele is associated not only with CD, but also with the propensity to colonic involvement in CD patients. The association of this allele with more extensive disease in UC has been previously reported but could not be tested here as a result of the relatively small numbers of UC patients studied. Similarly, insufficient sample size or lack of clinical information may have interfered with discovery of the relationship between HLA DRB1*0103 and colonic involvement in CD in previous studies. The relatively few patients in this study with colonic CD is likely reflective of the fact that the group studied here is a familial IBD population in which the prevalence of colonic CD appears to be lower than in sporadic cases of CD (47).

The results reported here also extend the hypothesis that there is a subgroup of CD patients with colonic disease which may be defined by the presence of a unique serum marker profile (48) and now the genetic marker HLA DRB1*0103 (in contrast to the association of NOD2/CARD15 with ileal CD (44)). Taken together, these data as well as clinical data suggesting that colonic CD may have different epidemiological (47) features and treatment responses (49) support the concept that colonic CD may be a unique form of IBD with divergent etiologic contributors. These data also confirm the importance of clinical phenotyping and the value of subgroup analyses as aids to further delineate the genetic factors contributing to IBD.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References