Original Article
Copyright ©2010 Baishideng. All rights reserved.
World J Gastroenterol. Jan 14, 2010; 16(2): 176-183
Published online Jan 14, 2010. doi: 10.3748/WJG.v16.i2.176
Interaction of the major inflammatory bowel disease susceptibility alleles in Crohn’s disease patients
Veronika Csöngei, Luca Járomi, Enikő Sáfrány, Csilla Sipeky, Lili Magyari, Bernadett Faragó, Judit Bene, Noémi Polgár, Lilla Lakner, Patrícia Sarlós, Márta Varga, Béla Melegh
Veronika Csöngei, Luca Járomi, Enikő Sáfrány, Csilla Sipeky, Lili Magyari, Bernadett Faragó, Judit Bene, Noémi Polgár, Béla Melegh, Department of Medical Genetics, University of Pécs, Pécs 7624, Hungary
Lilla Lakner, Department of Medicine and Gastroenterology, Markusovszky Hospital, Szombathely 9700, Hungary
Patrícia Sarlós, 3rd Department of Internal Medicine, University of Pécs, Pécs 7624, Hungary
Márta Varga, 3rd Department of Medicine and Gastroenterology, Réthy Pál Hospital, Békéscsaba 5600, Hungary
Author contributions: Csöngei V and Járomi L performed the genotyping and analyzed the data; Sáfrány E, Sipeky C, Magyari L, Faragó B and Bene J designed and performed the genotyping and were also involved in editing the manuscript; Polgár N overviewed the genotyping and revised the paper; Lakner L, Sarlós P and Varga M contributed human subjects and DNA samples; Melegh B designed and overviewed the study; Csöngei V wrote the manuscript.
Supported by Grant of Hungarian Scientific Research Foundation, No. OTKA T 73430
Correspondence to: Béla Melegh, MD, PhD, DSc, Professor, Chairman, Department of Medical Genetics, University of Pécs, Pécs 7624, Hungary. bela.melegh@aok.pte.hu
Telephone: +36-72-536427 Fax: +36-72-536032
Received: July 24, 2009
Revised: September 22, 2009
Accepted: September 29, 2009
Published online: January 14, 2010

Abstract

AIM: To investigate the interaction of interleukin-23 receptor (IL23R) (rs1004819 and rs2201841), autophagy-related 16-like 1 (ATG16L1) (rs2241880), caspase recruitment domain-containing protein 15 (CARD15) genes, and IBD5 locus in Crohn’s disease (CD) patients.

METHODS: A total of 315 unrelated subjects with CD and 314 healthy controls were genotyped. Interactions and specific genotype combinations of a total of eight variants were tested. The variants of IBD5 locus (IGR2198a_1 rs11739135 and IGR2096a_1 rs12521868), CARD15 (R702W rs2066845 and L1007fs rs2066847), ATG16L1 (rs2241880) and IL23R (rs1004819, rs2201841) genes were genotyped by PCR-RFLP, the G908R (rs2066844) in CARD15 was determined by direct sequencing.

RESULTS: The association of ATG16L1 T300A with CD was confirmed [P = 0.004, odds ratio (OR) = 1.69, 95% CI: 1.19-2.41], and both IL23R variants were found to represent significant risk for the disease (P = 0.008, OR = 2.05, 95% CI: 1.20-3.50 for rs1004819 AA; P < 0.001, OR = 2.97, 95% CI: 1.65-5.33 for rs2201841 CC). Logistic regression analysis of pairwise interaction of the inflammatory bowel disease (IBD) loci indicated that IL23R, ATG16L1, CARD15 and IBD5 (IGR2198a_1) contribute independently to disease risk. We also analysed the specific combinations by pair of individual ATG16L1, IL23R rs1004819, rs2201841, IGR2198a_1, IGR2096a_1 and CARD15 genotypes for disease risk influence. In almost all cases, the combined risk of susceptibility pairs was higher in patients carrying two different risk-associated gene variants together than individuals with just one polymorphism. The highest OR was found for IL23R rs2201841 homozygous genotype with combination of positive CARD15 status (P < 0.001, OR = 9.15, 95% CI: 2.05-40.74).

CONCLUSION: The present study suggests a cumulative effect of individual IBD susceptibility loci.

Key Words: Gene interaction, Interleukin-23 receptor, Autophagy-related 16-like 1, IBD5, Caspase recruitment domain-containing protein 15, Crohn’s disease, Inflammatory bowel disease


Citation: Csöngei V, Járomi L, Sáfrány E, Sipeky C, Magyari L, Faragó B, Bene J, Polgár N, Lakner L, Sarlós P, Varga M, Melegh B. Interaction of the major inflammatory bowel disease susceptibility alleles in Crohn’s disease patients. World J Gastroenterol 2010; 16(2): 176-183
INTRODUCTION

Two main clinical presentations of inflammatory bowel disease (IBD) are Crohn’s disease (CD) and ulcerative colitis (UC)[1]. IBD is now widely believed to originate from an uncontrolled mucosal immunity of the gastrointestinal tract[2]. Twin and family studies have reported that besides environmental factors genetic susceptibility is essential in IBD development[3]. Up to now, many novel candidate genes have been found to confer increased risk for the disease, some loci seem to be specific to CD or UC, others have been reported to confer susceptibility to IBD overall; at the moment the most replicated loci are interleukin-23 receptor (IL23R) and autophagy-related 16-like 1 (ATG16L1).

Duerr et al[4] identified IL23R gene as an IBD-associated gene in a genome-wide association study. Subsequent genome-wide association studies provided replication and confirmed the role of IL23R in CD[5-7].

In a European genome-wide association study the coding variant T300A (rs2241880) within the ATG16L1 gene was reported to be highly associated with CD, and to carry the whole disease risk exerted by this locus[8]. The association of ATG16L1 gene and CD was replicated in numerous studies[9-12]. IL23R and ATG16L1 T300A were also proved to be risk variants in the Hungarian CD population[13].

The rising number of CD candidate genes gives us the possibility to evaluate gene-gene interactions among susceptibility genes. Playing a role in biomolecular mechanisms, these interactions, or epistases are ubiquitous features of the genetic architecture of common human diseases[14]; their existence has been proved by several studies[15-17]. Since gene-gene interactions cannot only enhance but also weaken the individual gene effects, which can explain the lack of replication of single-locus results[14,18], complex gene-gene interactions may be considered more important than independent effects of single susceptibility genes.

Therefore our aim was to join the major susceptibility genes into a gene-gene interaction analysis in the Hungarian CD population: two IL23R gene risk variants, namely the intronic rs2201841 and rs1004819, the ATG16L1 gene variant T300A, the three well-known SNPs (R702W, L1007fs and G908R) of caspase recruitment domain-containing protein 15 (CARD15) gene and two markers located in IBD5 (IGR2198a_1 and IGR2096a_1) were tested for statistical interaction.

MATERIALS AND METHODS
Patients

We examined 315 unrelated patients with CD (151 males, 164 females, mean age 38.65 ± 0.79 years). The CD group included mixed Caucasian patients who had typical symptoms and diagnosis. A group of 314 clinically healthy subjects (170 males, 144 females, mean age 40.8 ± 0.80 years) with no IBD or other autoimmune disease were collected for the study. The origin of DNA samples was the central Biobank governed by the University of Pecs, as part of the National Biobank Network of Hungary (www.biobank.hu ), which belongs also to the pan-European Biobanking and Biomolecular Resources Research Infrastructure preparatory phase project (http://bbmri.eu/bbmri/ ). The governance, maintenance and management principles of the Biobank had been approved by the national Scientific Research Ethics Committee, Budapest (ETT TUKEB).

During the entire investigation period the guidelines and regulations of the 1975 Helsinki Declaration and the currently operative national laws were followed; the patients gave their informed consent for use of their collected, anonymized DNA samples for research purposes.

Genotyping

Genomic DNA was extracted from peripheral blood leukocytes with a routine salting out method. For genotyping the variants of IBD5 locus (IGR2198a_1 rs11739135 and IGR2096a_1 rs12521868), CARD15 (R702W rs2066845 and L1007fs rs2066847), ATG16L1 (rs2241880) and IL23R (rs1004819, rs2201841) genes PCR-RFLP methods were applied, the primers designed and used are given in Table 1. The PCR amplifications were performed on MJ Research PTC-200 thermal cyclers (Bio-Rad, Hercules, CA, USA) using the following conditions: initial denaturation at 96°C for 2 min followed by 35 cycles of denaturation at 95°C for 30 s, annealing for 45 s at 54°C (rs1004819), 55°C (rs2241880 and rs2201841), 58°C (rs11739135, rs12521868), 60°C (rs2066847), and 62°C (rs2066845), extension at 72°C for 45 s and final extension at 72°C for 5 min. Each polymerase chain reaction contained 200 μmol/L of each dNTP, 1 U of Taq polymerase, 5 μL of reaction buffer (100 mmol/L Tris HCl, pH = 9.0; containing 500 mmol/L KCl, 15 mmol/L MgCl2), 0.2 μmol/L of each primer and 1 μL DNA to be amplified in a final volume of 50 μL. The amplicons were digested by allele-specific restriction endonucleases Hin1II (rs11739135), TrulI (rs12521868), HinP1I (rs2066845), BspLI (rs2066847), LweI (rs2241880), TaaI (rs1004819) and HpyF3I (rs2201841). The amplicon contained an obligate cleavage site of the restriction enzyme for the suitable visual control of the efficacy of the digestion. The restriction fragments were separated by electrophoresis on 3% agarose gels containing ethidium bromide and visualized by UV transillumination. The genotyping of G908R (rs2066844) in CARD15 was carried out by direct sequencing by BigDye Terminator labelling with ABI 3100 automatic sequencer (Foster City, CA, USA).

Table 1 Primer sequences for the analysed variants.
GeneSNPPrimers (5’-3’)
IL23Rrs1004819F: GCATTCTAGGACCGTTTTGG
R: ATCTGGTGGAAATATGTGAAACCTA
IL23Rrs2201841F: GGCAAAAGGGAATTGAGAGG
R: GGCCTATGATTATGCTTTTTCCTG1
ATG16L1rs2241880F: CTCTGTCACCATATCAAGCGTGG
R: TCTAGAAGGACAGGCTATCAACAGATG
CARD15rs2066844F: GAGCCGCACAACCTTCAGATC
R: ACTTGAGGTGCCCAACATTCAG
CARD15rs2066845F: GAGGCCACTCTGGGATTGAG
R: TAGACTCTGAAGCTTACCTGCGC1
CARD15rs2066847F: TGGCTAACTCCTGCAGTCTC
R: GATCCTCAAAATTCTGCCATTC
IGR2198a_1rs11739135F: AGACACTGGGACATCATCTGTCTG
R: GGGCAATTCTATGAGGACATTTAGA
IGR2096a_1rs12521868F: CAAGATTTCTGCCATAGCCTCCT
R: GGAGGGTGGTGTAGCCAGAGTAG
Statistical analysis

Statistical analysis was carried out using the SPSS 15.0. package for Windows (SPSS Inc., Chicago, IL, USA). The allele frequencies were compared with Pearson’s χ2 test. Haploview 4.1 was used to test linkage disequilibrium. The r2 values for IGR2198a_1 and IGR2096a_1; and for IL23R rs1004189 and rs2201841 were below 0.8 (r2 = 0.63 and r2 = 0.62, respectively). Binary logistic regression analysis was applied to observe the individual contributions of IBD5, CARD15, ATG16L1 and IL23R, and to test for pairwise statistical interaction. An association was considered significant if a P value of < 0.05 was attained. CARD15 status was classified as - (wild type) or + (at least one mutation in any of the three CARD15 SNPs). The individual CARD15 genotypes and CARD15 status were stratified by IBD5 genotypes; IL23R and ATG16L1 genotypes were stratified by CARD15 genotypes, CARD15 status and IBD5 genotypes; and ATG16L1 genotype was stratified by IL23R genotypes. The odds ratios (ORs) and confidence intervals for these specific combinations of IBD5, CARD15, ATG16L1 and IL23R were derived from χ2 in 2x2 contingency tables.

RESULTS

All analysed allele frequencies and genotype distributions were in Hardy-Weinberg equilibrium both in patients and controls; results are shown in Table 2.

Table 2 Case-control genotypes and allele frequencies of variants in IL23R, ATG16L1, CARD15 and IBD5 n (%).
CD (n = 315)Controls (n = 314)OR (95% CI)1P
IL23R (rs1004189)
GG119 (37.8)151 (48.1)
GA152 (48.3)140 (44.6)
GA+AA196 (62.2)163 (51.9)1.50 (1.09-2.08)0.013a
AA44 (14.0)23 (7.3)2.05 (1.20-3.50)0.008a
RAF0.3810.2960.001a
IL23R (rs2201841)
TT131 (41.6)152 (48.4)
TC139 (44.1)145 (46.2)
TC+CC184 (58.4)162 (25.8)1.28 (0.93-1.76)0.14
CC45 (14.3)17 (5.4)2.97 (1.65-5.33)< 0.001a
RAF0.3630.2850.003a
ATG16L1 T300A (rs2241880)
AA56 (17.8)72 (22.9)
AG151 (47.9)163 (51.9)
AG+GG259 (82.2)242 (77.1)1.45 (0.98-2.17)0.06
GG108 (34.3)79 (25.2)1.69 (1.19-2.41)0.004a
RAF0.5830.5110.011a
CARD15 R702W (rs2066844)
CC275 (87.3)294 (93.6)
CT38 (12.1)18 (5.7)
CT+TT40 (12.7)20 (6.4)2.13 (1.19-3.80)0.011a
TT2 (0.6)2 (0.6)0.62 (0.06-6.98)0.70
RAF0.0670.0350.011a
CARD15 G908R (rs2066845)
GG299 (94.9)305 (97.1)
GC16 (5.1)9 (2.9)
GC+CC16 (5.1)9 (2.9)1.65 (0.71-3.86)0.24
CC0 (0.0)0 (0.0)NCNC
RAF0.0250.0140.17
CARD15 L1007fs (rs2066847)
- -264 (83.8)291 (92.7)
- insC42 (13.3)23 (7.3)
- insC+insC insC51 (16.2)23 (7.3)2.57 (1.51-4.36)< 0.001a
insC insC9 (2.9)0 (0.0)NCNC
RAF0.0950.037< 0.001a
CARD15 status2
-220 (69.8)262 (83.4)
+95 (30.2)52 (16.6)2.17 (1.46-3.21)< 0.001a
IGR2198a_1 (rs11739135)
GG91 (28.9)120 (38.2)
GC160 (50.8)144 (45.9)
GC+CC251 (79.7)194 (61.8)1.54 (1.10-2.15)0.013a
CC64 (20.3)50 (15.9)1.32 (0.87-1.99)0.19
RAF0.4570.3890.014a
IGR2096a_1 (rs12521868)
GG94 (29.8)120 (38.2)
GT149 (47.3)142 (45.2)
GT+TT296 (94.0)262 (83.4)1.44 (1.03-2.02)0.034a
TT72 (22.9)52 (16.6)1.45 (0.97-2.17)0.07
RAF0.4650.3920.009a

The risk allele frequencies of IL23R rs1004819 and rs2201841, ATG16L1 rs2241880, IGR2198a_1 (rs11739135), IGR2096a_1 (rs12521868), CARD15 R702W (rs2066845) and CARD15 L1007fs (rs2066847) were significantly higher in patients compared to controls (Table 2). After adjusting for age and gender, logistic regression analysis showed that IL23R rs1004819 A and rs2201841 C, IGR2198a_1 C, IGR2096a_1 T, CARD15 702W, CARD15 1007fsinsC variants confer risk for CD; OR values ranged from 1.44 to 2.57 (P < 0.02).

For IL23R rs1004819 and rs2201841, and for ATG16L1 a highly significant increase in risk was observed in homozygous individuals (P = 0.008, OR = 2.05, 95% CI: 1.20-3.50 for rs1004189 AA; P < 0.001, OR = 2.97, 95% CI: 1.65-5.33 for rs2201841 CC; P = 0.004, OR = 1.69, 95% CI: 1.19-2.41 for ATG16L1 GG genotype).

The CARD15 G908R (rs2066845) didnot show any association with CD.

Next, we analysed the possible statistical interactions by pairs of ATG16L1, IL23R variants, CARD15 status and IBD5 (IGR2198a_1). No evidence of interactions between these seven markers was found. None of the P values was significant; the lowest P value was 0.25 (Table 3).

Table 3 Pairwise analysis of interactions of CARD15, IBD5 (IGR2198a_1), ATG16L1 and IL23R to risk of CD.
ModelOR (95% CI)P
IL23R rs1004189 * ATG16L11.51 (0.73-3.11)0.26
IL23R rs2201841 * ATG16L11.01 (0.29-3.57)0.99
IL23R rs1004189 * CARD15 status11.20 (0.54-2.67)0.65
IL23R rs2201841 * CARD15 status1.57 (0.30-8.31)0.60
IL23R rs1004189 * IGR2198a_11.30 (0.71-2.36)0.40
IL23R rs2201841 * IGR2198a_10.46 (0.12-1.75)0.25
ATG16L1 * CARD15 status1.49 (0.59-3.80)0.40
ATG16L1 * IGR2198a_10.95 (0.45-2.00)0.89
CARD15 status * IGR2198a_10.75 (0.33-1.71)0.49

Finally, we observed the specific combinations of single markers by pairs, the combined ORs are shown in Tables 4, 5 and 6. The individual IL23R and ATG16L1 genotypes were stratified by CARD15 genotypes and CARD15 status; results are shown in Table 4. ATG16L1 and both IL23R variants showed significant association with CD on the background of a wild type CARD15 genotype. The rs2201841 CC genotype in itself was found to be a stronger CD risk factor than CARD15. The ORs were higher in individuals carrying one of the CARD15 variants with either the ATG16L1 GG genotype or IL23R rs1004189 variant (P < 0.001, OR = 3.33, 95% CI: 1.96-5.67 for rs1004189 A carriers; and P < 0.001, OR = 3.82, 95% CI: 1.86-7.86 for ATG16L1 GG genotype on the background of CARD15 + status). The IL23R rs2201841 CC genotype and + CARD15 status together showed by far the highest OR (P < 0.001, OR = 9.15, 95% CI: 2.05-40.74).

Table 4 Genotype-specific CD odds ratios1 (with 95% CI) for combinations of variants in IL23R, ATG16L1 and CARD15.
CARD15 R702W
CARD15 G908R
CARD15 L1007fs
CARD15 status
ATG16L1
CCCT+TTGGGC+CC- -- insC+insC insC-+AA+AGGG
IL23R rs1004189
GG12.35 (1.03-5.34)a11.55 (0.46-5.21)12.85 (1.28-6.35)a12.100 (1.17-3.76)a11.16 (0.68-1.99)
GA+AA1.56 (1.12-2.19)a3.18 (1.45-6.97)a1.51 (1.09-2.09)a3.23 (0.99-10.57)1.57 (1.12-2.20)a3.40 (1.69-6.82)a1.50 (1.04-2.16)a3.33 (1.96-5.67)a1.34 (0.92-1.95)2.51 (1.55-4.08)a
IL23R rs2201841
TT+TC12.25 (1.26-4.03)a11.49 (0.62-3.59)12.43 (1.42-4.18)a12.12 (1.42-3.16)a11.48 (1.03-2.14)a
CC3.04 (1.67-5.57)a4.75 (0.53-42.81)2.69 (1.49-4.86)aNC2.89 (1.57-5.32)a8.52 (1.04-69.74)a2.70 (1.42-5.15)a9.15 (2.05-40.74)a2.67 (1.31-5.44)a4.68 (1.72-12.78)a
ATG16L1
AA+AG12.20 (1.14-4.26)a11.44 (0.56-3.72)12.37 (1.29-4.34)a12.12 (1.35-3.31)a--
GG1.57 (1.09-2.25)a3.18 (1.11-9.08)a1.51 (1.06-2.14)a6.91 (0.83-57.92)1.54 (1.07-2.22)a4.27 (1.54-11.79)a1.54 (1.04-2.27)a3.82 (1.86-7.86)a--
Table 5 Genotype-specific CD odds ratios1 (with 95% CI) for combinations of variants in IBD5, IL23R and ATG16L1.
IL23R rs1004189
IL23R rs2201841
ATG16L1
GGGA+AATT+TCCCAA+AGGG
IGR2198a_1
GG11.43 (0.80-2.53)14.83 (1.52-15.37)a11.71 (0.94-3.09)
GC+CC1.45 (0.84-2.48)2.44 (1.43-4.15)a1.58 (1.11-2.24)a3.66 (1.81-7.41)a1.60 (1.07-2.38)a2.38 (1.46-3.87)a
IGR2096a_1
GG11.55 (0.88-2.73)14.03 (1.39-11.62)a11.63 (0.91-2.92)
GT+TT1.49 (0.87-2.56)2.41 (1.42-4.09)a1.50 (1.06-2.13)a3.71 (1.80-7.67)a1.50 (1.01-2.24)a2.32 (1.42-3.79)a
Table 6 Genotype-specific CD odds ratios1 (with 95% CI) for combinations of variants in IBD5 and CARD15.
CARD15 R702W
CARD15 G908R
CARD15 L1007fs
CARD15 status
CCCT+TTGGGC+CC- -- insC+insC insC-+
IGR2198a_1
GG13.17 (1.15-8.69)a11.79 (0.39-8.22)13.04 (1.30-7.14)a12.65 (1.36-5.17)a
GC+CC1.60 (1.13-2.27)a2.819 (1.39-5.72)a1.52 (1.08-2.13)a2.69 (0.97-7.45)1.61 (1.13-2.31)a3.58 (1.80-7.16)a1.63 (1.11-2.39)a3.19 (1.90-5.37)a
IGR2096a_1
GG13.07 (1.20-7.86)a12.19 (0.51-9.41)12.43 (1.06-5.59)a12.46 (1.29-4.69)a
GT+TT1.55 (1.09-2.20)a2.75 (1.32-5.70)a1.47 (1.05-2.06)a2.41 (0.86-6.77)1.48 (1.03-2.11)a3.74 (1.85-7.54)a1.54 (1.05-2.26)a3.18 (1.87-5.39)a

The IL23R rs2201841 C variant in homozygous form increased the susceptibility for the disease on the ATG16L1 nonhomozygous (i.e. wild type and heterozygous subjects together) background significantly; we could not detect a similar effect for rs1004189 A carriers. ATG16L1 was associated with CD in the absence of IL23R rs2201841 CC genotype, but not in rs1004189 noncarriers. We found that bearing the ATG16L1 GG genotype together with one of the IL23R susceptibility variants enhanced the risk for CD (P < 0.001, OR = 2.51, 95% CI: 1.55-4.08 for rs1004189; P = 0.001, OR = 4.68, 95% CI: 1.72-12.78 for rs2201841 homozygous genotype).

The ORs calculated for specific combinations of IBD5 with IL23R and ATG16L1 genotypes are shown in Table 5. We detected significantly increased risk in patients carrying the IL23R rs2201841 CC genotype on a wild type IBD5 background. The IGRs showed significant association with the disease on an IL23R rs2201841 nonhomozygous or ATG16L1 nonhomozygous background. ATG16L1 didnot significantly influence the susceptibility of CD except in the presence of IGR variants (P < 0.001, OR = 2.38, 95% CI: 1.46-3.87 for IGR2198a_1 C; and P = 0.001, OR = 2.32, 95% CI: 1.42-3.79 for IGR2096a_1 T background). Moreover, for the combinations of IGRs and IL23R rs1004189 we could detect significantly elevated high ORs only in carriers of IGR2198a_1 C and rs1004189 A, or in patients with IGR2096a_1 T and rs1004189 A variants (P = 0.001, OR = 2.44, 95% CI: 1.43-4.15 for IGR2198a_1 C; and P = 0.001, OR = 2.41, 95% CI: 1.42-4.09 for IGR2096a_1 T). The IGRs and IL23R rs2201841 CC genotypes together resulted in higher risk than the IGRs in themselves, but this OR value was lower than the OR calculated for the rs2201841 CC genotype alone (P < 0.001, OR = 3.66, 95% CI: 1.81-7.41 for IGR2198a_1 C; and P < 0.001, OR = 3.71, 95% CI: 1.80-7.67 for IGR2096a_1 T in patients carrying IL23R rs2201841 homozygous genotype).

The ORs for individual CARD15 genotypes stratified by IBD5 IGR2198a_1 and IGR2096a_1 genotypes are summarized in Table 6. Both the IBD5 markers significantly increased the risk of CD in the absence of CARD15 mutations; the CARD15 variants were confirmed to be stronger risk factors for CD than IGRs. The IBD5 variants showed higher significant risk together with + CARD15 status (P < 0.001, OR = 3.19, 95% CI: 1.90-5.37 for IGR2198a_1; and P < 0.001, OR = 3.18, 95% CI: 1.87-5.39 for IGR2096a_1 on the background of CARD15 + status).

DISCUSSION

Since the identification of NOD2/CARD15 as the first susceptibility gene for CD in 2001[19,20], several additional loci have been implicated in CD and confirmed by replication, among others the IBD5, IL23R and ATG16L1[4-6,8,9,21,22] loci.

Recently the idea was raised that exploring gene-gene interactions might lead to a better understanding of disease cause and might help the prediction of disease risk. So far numerous studies have assessed the risk for the development of CD by combining information from the known genetic risk variants associated with the disease. Though Hampe et al[8] found a modest but significant association between ATG16L1 and CARD15 in their pioneer study, no interaction was demonstrated between the two loci in the majority of subsequent Caucasian studies[10,23,24]. Prescott et al[12] also demonstrated that ATG16L1 increased the susceptibility for CD irrespective of CARD15 status; at the same time they detected increased ATG16L1 G allele frequency in CARD15 carriers compared with noncarriers, which may indicate a weak interaction between these candidate CD susceptibility genes. Moreover, an additive effect was reported between ATG16L1 and CARD15[25]. The independence of ATG16L1 and IBD5 variants was also confirmed in IBD[12,23].

Besides the individual risk of IL23R variants several studies examined their epistatic interaction with other IBD genes like CARD15[24-26]. Mostly the well-replicated IL23R R381Q protecting variant was implied in gene-gene interaction analyses and was reported to act independently of CARD15. In one study an additive effect was found between this IL23R variant and CARD15 gene[25]. The epistasis of the intronic rs1004189 risk variant with CARD15 and IBD5 was examined in a German CD population, but no gene-gene interaction was found[26]. No evidence of epistatic interaction between CARD15 and IBD5 was demonstrated in an Italian IBD study either[27]. Besides combining the most associated CD susceptibility variants by pair, multilocus analyses with three or more loci were also performed. Latiano et al[24] didnot found any significant interaction between ATG16L1, IL23R, CARD15 and IBD5 by triplets. Prescott et al[12] established a combined additive risk for all high risk genotypes in ATG16L1, CARD15 and IBD5, which was 20 fold that of the baseline risk for individuals carrying none of the risk alleles. Weersma et al[28] found an association between the increase in the number of risk alleles (ATG16L1, IL23R, CARD15, IBD5 and DLG5) and an increased risk for the development of CD.

Here we selected two risk-conferring variants of IL23R and ATG16L1 T300A mutation, and studied them together with the well-replicated CARD15 and IBD5 loci in Hungarian CD patients. Besides confirming our previous results with respect to IL23R rs2201841[29], we found a significant positive association between IL23R rs1004819 variant and CD observed in the Hungarian population for the first time. For ATG16L1 rs2241880 the estimated risk derived from the 315 CD patients showed a significant 1.7-fold increase in risk for the homozygous genotype. Our data are in line with most previous reports in Hungarian and other Caucasian populations[12,13]. Our results also verify that the previously identified R702W and L1007fs alterations in CARD15 gene act as CD susceptibility factors in the Hungarian population[30,31]. Since neither SLC22A4 nor SLC22A5 variants seem to confer risk for CD in the Hungarian population[32], we tested two other disease-associated IBD5 markers, IGR2198a_1, (rs11739135), IGR2096a_1 (rs12521868) for gene-gene interactions[22,33,34].

First we observed the statistical pairwise interactions between IL23R (rs1004189, rs2201841), ATG16L1 (rs2241880), CARD15 status and IGR2198a_1 (rs11739135). No evidence of epistatic interaction was found by logistic regression suggesting that all examined loci contribute independently to disease risk.

Next, we analysed the specific combinations of individual genotypes by pairs with respect to CD risk. We detected a significant association with CD for IGR2198a_1 and IGR2096a_1, respectively, on the background of wild type (-) CARD15 status, indicating that these two IBD5 markers and CARD15 are independent determinants of disease risk. Also high, significant ORs were found for ATG16L1 and IL23R variants (rs1004189, rs2201841), respectively, both in the presence and absence of CARD15 mutations, suggesting that these genes also act independently on CD risk. We detected a significant association in patients bearing IL23R rs2201841 CC genotype and wild type IBD5 background together, and vice versa. In carriers of IGR variants a significantly high risk was detected on IL23R rs2201841 nonhomozygous background; accordingly they may play an independent role in CD susceptibility. Similar to the results of previous studies[12,25], ATG16L1 increased the disease risk both in the presence and absence of CARD15, moreover in our study it was independent from IL23R rs2201841 genotype, but not from rs1004189 and IBD5 status. For combinations of IL23R rs1004189 and IBD5 markers, significant association was seen only in individuals carrying together the rs1004189 mutation and one of the two IGR variants.

In almost all specific pairwise combinations, the highest OR was found in patients with two different risk-associated gene variants, this cumulative OR was by far the highest in individuals with IL23R rs2201841 CC genotype and + CARD15 status.

However, we cannot detect significant statistical interaction between the analysed IL23R, ATG16L1, CARD15 and IBD5 risk alleles. The results of the present study suggest that these susceptibility factors may have a possible cumulative effect in the Hungarian CD population. By combining information from the known common risk polymorphisms significant predictive value from genetic markers might be gained; accordingly further large, well-powered studies should be performed to clarify the exact nature of these possible correlations.

COMMENTS
Background

Up to now, strong evidence has been provided that genetic factors play a significant role in determining the susceptibility of individuals to inflammatory bowel disease (IBD), especially for Crohn’s disease (CD). After the identification of the disease-associated NOD2 [caspase recruitment domain-containing protein 15 (CARD15)] gene, huge genome-wide linkage-analyses and meta-analyses have reported several CD susceptibility regions like IBD5 locus, DLG5, interleukin-23 receptor (IL23R), and autophagy-related 16-like 1 (ATG16L1) gene.

Research frontiers

Besides establishing the risk of carrying single variants, numerous studies have performed gene-gene interaction analyses for the major well-replicated susceptibility genes (CARD15, ATG16L1, IL23R genes and IBD5 locus). Mostly the independence of these main loci has been reported; nevertheless some studies have found an increased disease risk for carrying two or more certain risk variants together compared to non-carriers or individuals with only one susceptibility variant. In the present study two SNPs of IL23R, one of the ATG16L1, three of the CARD15 genes and two of IBD5 locus were genotyped and involved in interaction analysis in the Hungarian CD population.

Innovations and breakthroughs

The present study confirms the reported association between IL23R rs2201841 and ATG16L1 rs2241880 variants and CD susceptibility. The authors examined the IL23R rs1004189 in the Hungarian CD population for the first time, and found it significantly more frequent in patients compared to healthy controls. The analysis of statistical pairwise interactions between IL23R, ATG16L1, CARD15 status and IBD5 confirmed the independence of these susceptibility genes, while the specific combinations by pair showed the highest odds ratio in patients with two different risk-associated gene variants, suggesting that they may have a cumulative effect in this Hungarian CD population.

Applications

The exploration of epistatic interactions between the major susceptibility genes and the specification of high risk genotype combinations could support the better understanding of the development of CD and could facilitate the diagnosis of high-risk patients.

Peer review

The authors investigated the interactions of the major IBD susceptibility alleles in a Hungarian CD cohort. This is of high clinical significance as this methodology could help the diagnosis of high-risk patients.

Footnotes

Peer reviewers: Ferenc Sipos, MD, PhD, Cell Analysis Laboratory, 2nd Department of Internal Medicine, Semmelweis University, Szentkirályi u. 46., Budapest 1088, Hungary; Dr. Morten Frisch, MD, PhD, DSc, Department of Epidemiology Research, Statens Serum Institut, Building 206, Room 317, 5 Artillerivej, DK-2300 Copenhagen S, Denmark

S- Editor Tian L L- Editor O’Neill M E- Editor Zheng XM

References
1.  Podolsky DK. Inflammatory bowel disease. N Engl J Med. 2002;347:417-429.  [PubMed]  [DOI]
2.  Van Limbergen J, Russell RK, Nimmo ER, Ho GT, Arnott ID, Wilson DC, Satsangi J. Genetics of the innate immune response in inflammatory bowel disease. Inflamm Bowel Dis. 2007;13:338-355.  [PubMed]  [DOI]
3.  Weersma RK, van Dullemen HM, van der Steege G, Nolte IM, Kleibeuker JH, Dijkstra G. Review article: Inflammatory bowel disease and genetics. Aliment Pharmacol Ther. 2007;26 Suppl 2:57-65.  [PubMed]  [DOI]
4.  Duerr RH, Taylor KD, Brant SR, Rioux JD, Silverberg MS, Daly MJ, Steinhart AH, Abraham C, Regueiro M, Griffiths A. A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science. 2006;314:1461-1463.  [PubMed]  [DOI]
5.  Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007;447:661-678.  [PubMed]  [DOI]
6.  Libioulle C, Louis E, Hansoul S, Sandor C, Farnir F, Franchimont D, Vermeire S, Dewit O, de Vos M, Dixon A. Novel Crohn disease locus identified by genome-wide association maps to a gene desert on 5p13.1 and modulates expression of PTGER4. PLoS Genet. 2007;3:e58.  [PubMed]  [DOI]
7.  Raelson JV, Little RD, Ruether A, Fournier H, Paquin B, Van Eerdewegh P, Bradley WE, Croteau P, Nguyen-Huu Q, Segal J. Genome-wide association study for Crohn's disease in the Quebec Founder Population identifies multiple validated disease loci. Proc Natl Acad Sci USA. 2007;104:14747-14752.  [PubMed]  [DOI]
8.  Hampe J, Franke A, Rosenstiel P, Till A, Teuber M, Huse K, Albrecht M, Mayr G, De La Vega FM, Briggs J. A genome-wide association scan of nonsynonymous SNPs identifies a susceptibility variant for Crohn disease in ATG16L1. Nat Genet. 2007;39:207-211.  [PubMed]  [DOI]
9.  Rioux JD, Xavier RJ, Taylor KD, Silverberg MS, Goyette P, Huett A, Green T, Kuballa P, Barmada MM, Datta LW. Genome-wide association study identifies new susceptibility loci for Crohn disease and implicates autophagy in disease pathogenesis. Nat Genet. 2007;39:596-604.  [PubMed]  [DOI]
10.  Cummings JR, Cooney R, Pathan S, Anderson CA, Barrett JC, Beckly J, Geremia A, Hancock L, Guo C, Ahmad T. Confirmation of the role of ATG16L1 as a Crohn's disease susceptibility gene. Inflamm Bowel Dis. 2007;13:941-946.  [PubMed]  [DOI]
11.  Weersma RK, Zhernakova A, Nolte IM, Lefebvre C, Rioux JD, Mulder F, van Dullemen HM, Kleibeuker JH, Wijmenga C, Dijkstra G. ATG16L1 and IL23R are associated with inflammatory bowel diseases but not with celiac disease in the Netherlands. Am J Gastroenterol. 2008;103:621-627.  [PubMed]  [DOI]
12.  Prescott NJ, Fisher SA, Franke A, Hampe J, Onnie CM, Soars D, Bagnall R, Mirza MM, Sanderson J, Forbes A. A nonsynonymous SNP in ATG16L1 predisposes to ileal Crohn's disease and is independent of CARD15 and IBD5. Gastroenterology. 2007;132:1665-1671.  [PubMed]  [DOI]
13.  Lakatos PL, Szamosi T, Szilvasi A, Molnar E, Lakatos L, Kovacs A, Molnar T, Altorjay I, Papp M, Tulassay Z. ATG16L1 and IL23 receptor (IL23R) genes are associated with disease susceptibility in Hungarian CD patients. Dig Liver Dis. 2008;40:867-873.  [PubMed]  [DOI]
14.  Moore JH. The ubiquitous nature of epistasis in determining susceptibility to common human diseases. Hum Hered. 2003;56:73-82.  [PubMed]  [DOI]
15.  Moore JH. A global view of epistasis. Nat Genet. 2005;37:13-14.  [PubMed]  [DOI]
16.  Malmberg RL, Held S, Waits A, Mauricio R. Epistasis for fitness-related quantitative traits in Arabidopsis thaliana grown in the field and in the greenhouse. Genetics. 2005;171:2013-2027.  [PubMed]  [DOI]
17.  Segrè D, Deluna A, Church GM, Kishony R. Modular epistasis in yeast metabolism. Nat Genet. 2005;37:77-83.  [PubMed]  [DOI]
18.  Culverhouse R, Suarez BK, Lin J, Reich T. A perspective on epistasis: limits of models displaying no main effect. Am J Hum Genet. 2002;70:461-471.  [PubMed]  [DOI]
19.  Hugot JP, Chamaillard M, Zouali H, Lesage S, Cézard JP, Belaiche J, Almer S, Tysk C, O'Morain CA, Gassull M. Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease. Nature. 2001;411:599-603.  [PubMed]  [DOI]
20.  Ogura Y, Bonen DK, Inohara N, Nicolae DL, Chen FF, Ramos R, Britton H, Moran T, Karaliuskas R, Duerr RH. A frameshift mutation in NOD2 associated with susceptibility to Crohn's disease. Nature. 2001;411:603-606.  [PubMed]  [DOI]
21.  Peltekova VD, Wintle RF, Rubin LA, Amos CI, Huang Q, Gu X, Newman B, Van Oene M, Cescon D, Greenberg G. Functional variants of OCTN cation transporter genes are associated with Crohn disease. Nat Genet. 2004;36:471-475.  [PubMed]  [DOI]
22.  Silverberg MS, Duerr RH, Brant SR, Bromfield G, Datta LW, Jani N, Kane SV, Rotter JI, Philip Schumm L, Hillary Steinhart A. Refined genomic localization and ethnic differences observed for the IBD5 association with Crohn's disease. Eur J Hum Genet. 2007;15:328-335.  [PubMed]  [DOI]
23.  Glas J, Konrad A, Schmechel S, Dambacher J, Seiderer J, Schroff F, Wetzke M, Roeske D, Török HP, Tonenchi L. The ATG16L1 gene variants rs2241879 and rs2241880 (T300A) are strongly associated with susceptibility to Crohn's disease in the German population. Am J Gastroenterol. 2008;103:682-691.  [PubMed]  [DOI]
24.  Latiano A, Palmieri O, Valvano MR, D'Incà R, Cucchiara S, Riegler G, Staiano AM, Ardizzone S, Accomando S, de Angelis GL. Replication of interleukin 23 receptor and autophagy-related 16-like 1 association in adult- and pediatric-onset inflammatory bowel disease in Italy. World J Gastroenterol. 2008;14:4643-4651.  [PubMed]  [DOI]
25.  Roberts RL, Gearry RB, Hollis-Moffatt JE, Miller AL, Reid J, Abkevich V, Timms KM, Gutin A, Lanchbury JS, Merriman TR. IL23R R381Q and ATG16L1 T300A are strongly associated with Crohn's disease in a study of New Zealand Caucasians with inflammatory bowel disease. Am J Gastroenterol. 2007;102:2754-2761.  [PubMed]  [DOI]
26.  Glas J, Seiderer J, Wetzke M, Konrad A, Török HP, Schmechel S, Tonenchi L, Grassl C, Dambacher J, Pfennig S. rs1004819 is the main disease-associated IL23R variant in German Crohn's disease patients: combined analysis of IL23R, CARD15, and OCTN1/2 variants. PLoS One. 2007;2:e819.  [PubMed]  [DOI]
27.  Latiano A, Palmieri O, Valvano RM, D'Incà R, Vecchi M, Ferraris A, Sturniolo GC, Spina L, Lombardi G, Dallapiccola B. Contribution of IBD5 locus to clinical features of IBD patients. Am J Gastroenterol. 2006;101:318-325.  [PubMed]  [DOI]
28.  Weersma RK, Stokkers PC, van Bodegraven AA, van Hogezand RA, Verspaget HW, de Jong DJ, van der Woude CJ, Oldenburg B, Linskens RK, Festen EA. Molecular prediction of disease risk and severity in a large Dutch Crohn's disease cohort. Gut. 2009;58:388-395.  [PubMed]  [DOI]
29.  Faragó B, Magyari L, Sáfrány E, Csöngei V, Járomi L, Horvatovich K, Sipeky C, Maász A, Radics J, Gyetvai A. Functional variants of interleukin-23 receptor gene confer risk for rheumatoid arthritis but not for systemic sclerosis. Ann Rheum Dis. 2008;67:248-250.  [PubMed]  [DOI]
30.  Bene J, Magyari L, Talián G, Komlósi K, Gasztonyi B, Tari B, Várkonyi A, Mózsik G, Melegh B. Prevalence of SLC22A4, SLC22A5 and CARD15 gene mutations in Hungarian pediatric patients with Crohn's disease. World J Gastroenterol. 2006;12:5550-5553.  [PubMed]  [DOI]
31.  Nagy Z, Karádi O, Rumi G, Rumi G Jr, Pár A, Mózsik G, Czirják L, Süto G. Crohn's disease is associated with polymorphism of CARD15/NOD2 gene in a Hungarian population. Ann N Y Acad Sci. 2005;1051:45-51.  [PubMed]  [DOI]
32.  Bene J, Komlósi K, Magyari L, Talián G, Horváth K, Gasztonyi B, Miheller P, Figler M, Mózsik G, Tulassay Z. Plasma carnitine ester profiles in Crohn's disease patients characterized for SLC22A4 C1672T and SLC22A5 G-207C genotypes. Br J Nutr. 2007;98:345-350.  [PubMed]  [DOI]
33.  Lakner L, Csöngei V, Sarlós P, Járomi L, Sáfrány E, Varga M, Orosz P, Magyari L, Bene J, Miheller P. IGR2096a_1 T and IGR2198a_1 C alleles on IBD5 locus of chromosome 5q31 region confer risk for Crohn's disease in Hungarian patients. Int J Colorectal Dis. 2009;24:503-507.  [PubMed]  [DOI]
34.  Noble CL, Nimmo ER, Drummond H, Ho GT, Tenesa A, Smith L, Anderson N, Arnott ID, Satsangi J. The contribution of OCTN1/2 variants within the IBD5 locus to disease susceptibility and severity in Crohn's disease. Gastroenterology. 2005;129:1854-1864.  [PubMed]  [DOI]