Influence of a nucleotide oligomerization domain 1 (NOD1) polymorphism and NOD2 mutant alleles on Crohn's disease phenotype

Abstract

AIM: To examine genetic variation of nucleotide oligomerization domain 1 (NOD1) and NOD2, their respective influences on Crohn's disease phenotype and gene-gene interactions.

METHODS: (ND₁+32656*1) NOD1 polymorphism and SNP8, SNP12 and SNP13 of NOD2 were analyzed in 97 patients and 50 controls. NOD2 variants were determined by reaction restriction fragment length polymorphism analysis. NOD1 genotyping and NOD2 variant confirmation were performed by specific amplification and sequencing.

RESULTS: The distribution of NOD1 polymorphism in patients was different from controls (P = 0.045) and not altered by existence of NOD2 mutations. In this cohort, 30.92% patients and 6% controls carried at least one NOD2 variant (P < 0.001) with R702W being the most frequent variant. Presence of at least one NOD2 mutation was inversely associated with colon involvement (9.09% with colon vs 36.4% with ileal or ileocolonic involvement, P = 0.04) and indicative of risk of penetrating disease (52.63% with penetrating vs 25.64% with non-penetrating or stricturing behavior, P = 0.02). L1007finsC and double NOD2 mutation conferred the highest risk for severity of disease (26.3% with penetrating disease vs 3.8% with non-penetrating or stricturing behavior presented L1007finsC, P = 0.01 and 21.0% with penetrating disease vs 2.5% with non-penentrating or stricturing behavior carried double NOD2 mutation, P = 0.007). Exclusion of patients with NOD2 mutations from phenotype/NOD1-genotype analysis revealed higher prevalence of *1*1 genotype in groups of younger age at onset and colonic location.

CONCLUSION: This study suggests population differences in the inheritance of risk NOD1 polymorphism and NOD2 mutations. Although no interaction between NOD1-NOD2 was noticed, a relationship between disease location and Nod-like receptor molecules was established.

Key Words: Crohn's disease; Nucleotide oligomerization domain 1; Nucleotide oligomerization domain 2

INTRODUCTION

Crohn's disease (CD) is a chronic inflammatory disorder of the gastrointestinal tract. Although the etiopathogenesis of this disease remains poorly understood, both genetic and environmental factors have been suggested to predispose to CD. Various disease phenotypes, including age at diagnosis, sex, family history, location of disease, response to medical therapies and behavior of the disease may be genetically determined.

Experimental and observational data suggest that intestinal inflammation arises from abnormal immune reactivity to bacterial flora in the intestine of individuals who are genetically predisposed[1]. The analysis of the molecules that participate in the response of commensal organisms revealed that gastric and intestinal cells are largely deficient in TLR signaling and must rely on alternative systems, such as Nod-like receptors (NLRs) for the detection of pathogens. The mammalian NLR family is composed of more than 20 members that share a modular domain organization of a C-terminal leucine-rich repeat (LRR) domain, a central nucleotide-binding site domain and a N-terminal protein-protein-interaction domain composed of a CARD (caspase activation and recruitment domain), pyrin domain or Bir domain[2].

The first NLRs reported to have a direct function as intracellular pattern recognition molecules were Nucleotide oligomerization domain 1 (NOD1) (CARD4) and NOD2 (CARD15); both proteins detect distinct substructures from bacterial peptidoglycan. NOD1 detects a unique tripeptide motif found in Gram-negative bacterial peptidoglycan and also in specific Gram-positive bacteria such as Listeria and Bacillus spp[3]. NOD2 detects muramyl dipeptide, the largest molecular motif common to Gram-negative and Gram-positive bacteria[4]. It is expressed in intestinal epithelial cells, with high expression in Paneth cells in the small intestine, intestinal myofibroblasts, granulocytes, endothelial, and monocyte-derived cells[5,6].

Identification of NOD2 as the first susceptibility gene for CD was a breakthrough in understanding inflammatory bowel disease (IBD) pathogenesis. NOD2 gene is located at the CD susceptibility locus (IBD1) on chromosome 16q12[7,8] and it has more than 60 sequence variants. Although, disease-associated NOD2 mutations linked to Blau syndrome and early onset of sarcoidosis have been found in the region encoding the nucleotide-binding site domain[9,10], the three common genetic mutations linked to CD are mapped within or adjacent to the LRR region of NOD2 (leading to protein changes at R702W, G908R, L1007finsC)[7,8]. These mutations are associated with an altered NF-κB activation and the linkage is particularly strong with ileal and ileocolonic CD[11,12]. NOD2 variants are associated with early surgery due to stenosis, postsurgical recurrence, familial CD[13] and stricturing and penetrating forms of CD[14].

CD association with NOD2 has been widely replicated. However, investigations into the inheritance of the three risk alleles in NOD2 associated with susceptibility to CD have demonstrated a remarkable heterogeneity across ethnicities and populations with regional variation across Europe[15,16].

The discovery of NOD2-related innate immune defects in certain CD cases has led to speculation about defects in other pattern recognition receptors and downstream signaling molecules. The gene encoding NOD1 (CARD4) is located within the chromosome 7p14 IBD locus, a region that contains an IBD susceptibility locus in British families[17]. An association between a complex insertion/deletion polymorphism (ND₁+32656*1) in NOD1 and susceptibility to IBD has been described. Particularly, this polymorphism has been associated to age at diagnosis and to the presence of IBD extraintestinal manifestations[18]. This NOD1 polymorphism has also been associated to increased susceptibility to asthma[19,20]. In both diseases, the mutation has been found to be an insertion/deletion polymorphism in an intron of NOD1. Convincing replication of these findings is pending, since no evidence of association between ND₁+32656*1 and IBD was found in two recent well-powered data sets[21,22].

The present study examines the genetic variation in NOD1 and NOD2 and their respective influences on the CD phenotype (age at diagnosis, disease location and behavior) in a cohort of well-characterized CD patients. Since NOD1 and NOD2 share structure and functions, a potential interaction between NOD1 and NOD2 variants in CD phenotype was analyzed. After stratifying patients by their NOD2 genotype, the distribution of NOD1 polymorphism was determined and the contribution of each genotype was studied in regard to the disease phenotype.

MATERIALS AND METHODS

Patients

Ninety-seven CD patients attending the IBD outpatient clinic of Hospital Sant Pau (Barcelona, Spain) were prospectively included in the study. Fifty healthy controls matched for age, sex and geography were also evaluated. CD diagnoses were based on clinical, radiologic, endoscopic and pathologic bases. Patients with CD were classified according to Montreal classification for age at onset, disease location and behavior[23]. All patients and healthy controls gave informed consent and the study was approved by the local ethics committee.

Genotyping

Analysis of NOD2 variants was performed as previously described, using genomic DNA extracted from blood samples by Qiagen kit (Qiagen, Heiden, Germany). A panel of 3 single nucleotide polymorphisms (SNP8, 12 and 13) was detected by a polymerase chain reaction (PCR)-restriction fragment length polymorphism analysis (PCR-RFLP)[7]. Each NOD2 variant was initially amplified by PCR using specific primers (Table 1). The PCR products were subsequently analyzed by restriction enzyme cleavage and gel electrophoresis. For assay of the SNP8, the PCR product (185 bp) was digested with MspI, resulting in the following fragments: 20, 35, 54 and 76 bp in R702 homozygous; 20, 35 and 130 bp in 702W homozygous and 20, 35, 54, 76, and 130 bp in heterozygous. For assay of the SNP12, the PCR product (163 bp) was digested with HhaI, resulting in the following fragments: 163 bp in G908 homozygous; 27 and 136 bp in 908R homozygous and 27, 136 and 163 in heterozygous. In order to detect the SNP13, the PCR product (151 bp) was digested with ApaI, resulting in the following fragments: 151 bp for Leu1007 homozygous; 20 and 131 bp in 1007Pro homozygous and 20, 131 and 151 bp in heterozygous.

Table 1 Primers for NOD2 and NOD1 genotyping.

	Forward	Reverse	Size (bp)
NOD2
R702W	5´-AGATCACAGCAGCCTTCCTG-3´	5´-CACGCTCTTGGCCTCACC-3´	185
G908R	5´-CTCTTTTGGCCTTTTCAGATTCTG-3´	5´-CAGCTCCTCCCTCTTCACCT-3´	163
L1007finsC NOD1	5´-GGCAGAAGCCCTCCTGCAGGGCC-3´	5´-CCTCAAAATTCTGCCATTCC-3´	151
ND1+32656	5´-TGACTGTGTGTGACTCTCTCTGC-3´	5´-TGGTGAAAGCTCTCCACTATCTC-3´	250

Genotyping of NOD1 (ND₁+32656) polymorphism and confirmation of the three NOD2 mutations were performed by specific amplification with the primers described in Table 1 and the subsequent sequencing of the amplified products. Sequencing reaction was performed using ABI PRISM BigDye terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) and analyzed by Genescan analysis on an ABI Prism 3100 Genetic Analyser according to the manufacturer's protocol (Applied Biosystem).

Statistical analysis

Genotype and allele frequencies of the patients and controls were compared by the χ² test or Fisher exact test in 2 × 2 contingency tables with at least 1 expected value < 5. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to estimate relative risks. A two-tailed P value ≤ 0.05 was considered significant. Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS) version 14.0 for Windows (SPSS Inc., Chicago, III).

RESULTS

Frequencies of three NOD2 mutant alleles and one NOD1 polymorphism in CD patients and healthy controls

NOD2 gene mutations (R702W, G908R and L1007finsC) were determined in 97 CD patients and 50 healthy controls. Frequencies are summarized in Table 2. The distribution of genotypes at each mutation was significantly different in CD patients versus controls. R702W was the most frequent variant in CD and controls (21.65% and 4%, respectively, P = 0.004), and the only homozygous mutant patient in this cohort was found for this SNP8. Carriage of R702W was associated to the highest risk for CD in our cohort of patients (OR = 7.04; 95% CI: 1.58-31.30). Genotype frequency of the L1007finsC variant was lower than R702W, but showed a tendency to be higher in CD patients than in controls (8.25% vs 2%, P = 0.167; OD 4.40, 95% CI: 0.53-36.25). The NOD2 variant with the lowest frequency in CD patients and in controls was G908R (5.15% vs 0%,P = 0.166). No homozygous NOD2 mutant was found for L1007finsC or G908R. In this CD cohort, 30.92% of patients carried at least one variant of NOD2 compared with 6% of healthy controls (P < 0.001) (Table 3),conferring a high risk for CD (OR 7.01; 95% CI: 2.02-24.30). Six CD patients but no controls carried two NOD2 variant alleles.

Table 2 Genotype at NOD2 polymorphisms in CD cases and healthy controls.

Mutation	Group	Genotype count, n (%)			OR (95% CI)1	P2
		WT/WT	Heterozygous	Homozygous
R702W	CD	75 (77.32)	21 (21.65)	1 (1.03)	7.04 (1.58-31.30)	0.004
	Controls	48 (96.00)	2 (4.00)	0
G908R	CD	92 (94.85)	5 (5.15)	0		0.166
	Controls	50 (100)	0	0
L1007finsC	CD	89 (91.75)	8 (8.25)	0	4.40 (0.53-36.25)	0.167
	Controls	49 (98.00)	1 (2.00)	0

¹ORs and probability values for disease status associated with carriage of at least 1 mutant allele (heterozygous, compound heterozygous and homozygous were grouped together). ²P-values were calculated with the Fisher's Exact test when comparing controls and CD patients.

Table 3 Distribution of NOD2 mutations in CD patients and controls.

NOD2 genotype	CD Patients,n (%)	Controls,n (%)	OR (95% CI)1	P2
At least one variant3	30 (30.92)	3 (6)	7.01 (2.02-24.30)	< 0.001
Heterozygous	24 (24.74)	3 (6)	5.61 (1.59-19.0)	0.003
Compound heterozygous4	5 (5.15)	0
Homozygous	1 (1.03)	0

¹ORs and probability values for disease status associated with NOD2 genotype. ²P-values are calculated with the Fisher's Exact test when comparing controls and CD patients.

³At least one variant was considered if any subject had at least one copy of the variant allele.

⁴Compound heterozygous was defined as the presence of two different variants.

NOD1 complex insertion/deletion polymorphism (ND1+32656) was examined in the same cohort of patients and controls (Table 4). Fifty-two percent of controls were *1*1, 34% were *1*2 and 14% were *2*2, whereas 59.79% of CD patients were *1*1, 37.11% were *1*2 and only 3.09% were *2*2. The distribution of NOD1 genotype according to the ND₁+32656 polymorphism in CD patients and controls was statistically different (P = 0.045). Frequency of CD patients carrying *1 allele was 96.8% whereas in controls it was 86%, conferring a significant risk to develop the disease (OR 5.10; 95% CI: 1.25-20.68, P = 0.032).

Table 4 Genotype frequencies of ND₁+32656 in CD patients and controls.

Group	NOD2	n1	NOD1 genotype, n (%)
			*1*1	*1*2	*2*2
CD		97	58 (59.79)2	36 (37.11)	3 (3.09)
	WT/WT	67	40 (59.70)	25 (37.31)	2 (2.98)
	Mutant3	30	18 (60.00)	11 (36.66)	1 (3.33)
Controls		50	26 (52.00)	17 (34.00)	7 (14.00)
	WT/WT	47	23 (48.93)	17 (36.17)	7 (14.89)
	Mutant	3	3 (100)	0	0

CD: Crohn's disease; WT: Wildtype. ¹Number of CD patients or controls;

²Results are expressed as number of CD patients (% of CD patients);

³NOD2 mutant was considered any subject that inherited at least one copy of the variant allele.

Distribution of NOD1 genotype according to WT or mutant NOD2 was analyzed in CD patients to assess potential interactions between NOD1 and NOD2 (Table 4). Among those patients carrying at least one NOD2 mutant allele, 60% of the patients were *1 *1, 36.66% were *1 *2 and 3.33% were *2 *2. Similarly, 59.70% of NOD2 WT/WT were *1 *1, 37.31% were *1 *2 and 2.98% were *2 *2. The presence of NOD2 mutant alleles had therefore no influence on the NOD1 polymorphism distribution(P = 0.99), suggesting no gene-gene interactions.

Clinical characteristics of CD patients according to the NOD2 genotype

CD patients were classified according to Montreal classification, with minor modifications as indicated in Table 5. The association of NOD2 mutations to each CD phenotype was analyzed using each mutant genotype. The presence of at least one risk allele or the joint analysis of compound heterozygous and homozygous for NOD2 mutations were considered as a single independent variable.

Table 5 Clinical characteristics of CD patients according to NOD2 genotype.

Clinical features	n1	NOD2 genotype, n (%)					P2OR (95% CI)
		WT/WT	R702W/WT	G908R/WT	L1007fsinsC/WT	Heter.compound & homozygous
Age at diagnosis
< 40 yr (A1 + A2)	80	54 (55.67)3	13 (13.40)	3 (3.09)	4 (4.12)	6 (6.18)	0.57 1.56 (0.46-5.27)
> 40 yr (A3)	17	13 (13.40)	3 (3.09)	0	1 (1.03)	0
Location
Ileal (L1)	34	21 (21.65)	8 (8.25)	1 (1.03)	2 (2.06)	2 (2.06)	0.26 1.67 (0.69-4.06)
Colonic (L2)	23	20 (20.62)	0	1 (1.03)	1 (1.03)	1 (1.03)	0.04 0.26 (0.07-0.96)
Ileocolonic (L3)	40	26 (26.80)	8 (8.25)	1 (1.03)	2 (2.06)	3 (3.09)	0.5 1.38 (0.57-3.29)
Behavior
Non-stricturing, non-penetrating (B1)	37	24 (24.74)	8 (8.25)	3 (3.09)	2 (2.06)	0	0.5 1.37 (0.56-3.29)
Stricturing (B2)	41	34 (35.05)	5 (5.15)	0	0	2 (2.06)	0.01 0.29 (0.11-0.78)
Penetrating (B3)	19	9 (9.28)	3 (3.09)	0	3 (3.09)	4 (4.12)	0.02 3.22 (1.14-9.06)

CD: Crohn's disease; WT: Wildtype. ¹Number of CD patients in each subgroup; ²P-values, odds ratios and confidence intervals refer to the comparison of presence versus absence of the at least one mutant NOD2 allele;

³Results are expressed as number of CD patients (% of CD patients).

A high proportion of patients in this cohort were diagnosed under the age of 40 (A1 + A2, n = 2 + 78), whereas 17 patients were diagnosed over 40 years (A3). All 6 patients with 2 mutant NOD2 alleles were diagnosed before 40 years of age.

To study the association between genotype and disease location, one L1 + L4 patient was included in the L1 group (n = 34, 35.05%), and another L2 + L4 patient was included in the L2 group (n = 23, 23.71%). NOD2 WT/WT patients were similarly distributed in L1, L2 and L3 groups: in 21.65% patients disease location was terminal ileum (L1), in 20.62%, it was colonic (L2), and in 26.80%, it was ileocolonic (L3). However, location of disease in NOD2 mutant patients was not identical to NOD2 WT/WT patients (P = 0.08). Despite the fact that R702W polymorphism was the most frequent in this cohort, only one patient with colon location carried this mutation (the patient had a double NOD2 mutation). The presence of at least one mutant NOD2 gene was inversely associated with exclusively colonic involvement (L2) (P = 0.04, OR 0.26; 95% CI: 0.07-0.96). When analyzing compound heterozygous and homozygous NOD2 mutations, 2.06% of patients were L1, 1.03% of patients were L2 and 3.09% of patients were L3, indicating a comparable distribution.

Similarly to location, behavior groups in CD patients were simplified as follows: B1 group included 4 patients B1p (n = 37, 38%), B2 included 4 patients B2p (n = 41, 42%) and B3 included 4 patients B3p (n = 19, 19.5%). Distribution of NOD2 mutations was different depending on disease behavior (P = 0.003). The presence of at least one mutant NOD2 allele was indicative of risk of penetrating disease (B3) (P = 0.02, OR 3.22, 95% CI: 1.14-9.06) with the allele L1007finsC being indicative of the highest risk (P = 0.007, OR 8.92; 95% CI: 1.91-41.68). The frequency of double NOD2 mutants was significantly higher in the B3 group than in the B2 and B1 groups (66% of the double NOD2 mutants were B3, 33.3% were B2 and non-double NOD2 mutants were B1). Presence of two NOD2 mutant alleles was therefore indicative of risk for severity of disease (P = 0.01, OR 10.13; 95% CI: 1.70-60.40). The exam of the behavior through the course of disease showed an expected changing pattern[24]. There was a progressive reduction in the proportion of patients in the group B1 (51.6% patients with 5 years of disease, 34.9% patients with 6-9 years of disease and 21.7% patients with 10-15 years of disease). Inversely, there was a progressive increase in the proportion of patients in the groups of more complicated forms, B2 (34.5% patients with 5 years of disease, 44.2 patients with 6-9 years of disease and 52.2% patients with 10-15 years of disease) and B3 (13.8% patients with 5 years of disease, 20.9% patients with 6-9 years of disease and 26.1% patients with 10-15 years of disease). Selecting the group of patients with 6-9 years of disease (n = 43), the presence of at least one NOD2 mutation was indicative of risk of penetrating disease (the B3 group) (P = 0.046, OR = 5.55, 95% CI: 1.14-27.01) and 66.7% of the double NOD2 mutants were included in the B3 group. Twelve patients with perianal disease (B1p, B2p, and B3p) were analyzed separately. Four of these patients presented one NOD2 mutation, one was a compound heterozygous and the rest were NOD2 WT/WT.

Clinical characteristics of CD patients according to the NOD1 genotype

Phenotype of CD patients was analyzed according to the ND₁+32656 polymorphism of NOD1 gene. Distribution of NOD1 genotype according to older age at diagnosis (A3 P = 0.64), location (L1 P = 0.28, L2 P = 0.56 and L3 P = 0.26) and behavior (B1 P = 0.55, B2 P = 0.99 and B3 P = 0.99) of the disease were not different from healthy controls. Similarly to a previous report, ND1+32656 genotype distribution in the group of early-onset CD (A1 + A2) was different from that observed in healthy controls. Only 2.5% of CD patients in the early-onset group had *2*2 genotype compared to 14% of healthy controls (P = 0.04).

Since NOD2 mutations have a strong association with some CD clinical characteristics, and in particular with the ileal location, 30 CD patients that presented at least one NOD2 mutation were excluded from the phenotype/genotype study to prevent any influence of NOD2 (Table 6). When comparing the distribution of NOD1 polymorphism in each phenotype group with healthy controls, a higher prevalence of *1*1 was observed in the group A1 + A2 (P = 0.04). Interestingly, there was a clear tendency of the colonic group (L2) to have a higher frequency of *1*1 and lower frequency of *1*2 and *2*2 than controls, but the distribution of the ND₁+32656 polymorphism was not statistically different because the low number of cases decreased the power of the test. Distribution of this NOD1 polymorphism in the other clinical subgroups of CD patients was comparable to healthy controls. Seven of the 12 patients with perianal disease (B1p + B2p + B3p) were NOD2 WT/WT. In this subgroup of patients, NOD1 polymorphism analysis showed that three of them were *1*1 and four were *1*2.

Table 6 Clinical characteristics of NOD2 WT/WT CD patients according to NOD1 genotype.

Clinical features	n1	NOD2 WT/WT
		NOD1:	*1*1	*1*2	*2*2
Age at diagnosis
< 40 yr (A1 + A2)	54		62.962	35.18	1.85
> 40 yr (A3)	13		46.15	46.15	7.69
Location
Ileal (L1)	21		57.14	38.09	4.76
Colonic (L2)	20		70	25	5
Ileocolonic (L3)	26		53.84	46.15	0
Behavior
Non-stricturing,	24		62.5	33.33	4.16
non-penetrating (B1)
Stricturing (B2)	34		61.76	35.29	2.94
Penetrating (B3)	9		44.44	55.55	0
Controls	47		48.93	36.17	14.89

WT: Wildtype. ¹Number of CD patients or controls in each subgroup;

²Values are expressed as the percent of patients in each clinical subgroup.

DISCUSSION

The frequency of NOD2 mutant alleles associated to CD in our cohort of patients was within the European range, but deviated somewhat from populations of nearby geographic regions[25,26]. The frequency of R702W was one of the highest described in Caucasian populations, whereas the frequency of L1007finsC was lower than in other studies[27]. This observation is consistent with marked racial and regional differences described in the inheritance of the three risk NOD2 alleles[16].

As expected, carriage of NOD2 mutations conferred a high risk for developing CD, but this was neither necessary nor sufficient for CD development. The three NOD2 mutations were not equally involved in CD susceptibility. The presence of R702W showed the strongest risk for CD in our cohort of patients. However, this mutation was not associated to CD in Galician, Finnish or Scottish populations[26]. The mutation with the strongest CD association in several familial and non-familial studies was L1007finsC[14], but this was not so in our cohort. Although the frequency of L1007finsC was noticeably elevated in CD patients, the absence of controls with this genotype precluded a statistical comparison.

We found an association between the polymorphism located at the intron IX-exon IX boundary of NOD1 and susceptibility to CD in our cohort of patients. These results confirm a previous report associating this NOD1 polymorphism with early IBD-onset and extraintestinal manifestations[18]. Although one recent study did not show a significant association with IBD[21], this NOD1 non-coding polymorphism showed a strong association with asthma and the presence of elevated IgE levels in three independent panels of subjects[20]. Other NOD1 polymorphisms in the coding sequence have been previously examined and showed no influence in CD susceptibility[28]. Mutations with phenotypic effects should be predominantly found at the coding sequence but complex disease susceptibility is often mediated through regulatory polymorphisms. In this case, ND₁+32656 may affect the binding of an unknown nuclear factor[20]. The involvement of NOD1 gene is not surprising, since NOD1, similarly to NOD2, is involved in the recognition of intracellular bacterial pathogen-associated molecular patterns[29] and the two molecules share structure and functional similarities. Certain polymorphisms and mutations in these molecules may, therefore, result in abnormalities during bacterial recognition with direct implications for CD pathogenesis. Given the importance of these results, further confirmatory studies are warranted in more and larger IBD populations. In order to maximize the opportunities to compare clinical subgroups, location was kept simple and genotyping was specifically blinded to clinical status. Mutations of the NOD2 gene were rare among our patients with disease limited to the colon (L2). This is in accordance with recent studies showing that NOD2 mutations (particularly L1007finsC) are strongly related to an increased risk of developing ileal CD. In our cohort of patients we only found this association after combining ileal and ileocolonic patients. This could be the consequence of the low rates of limited ileal CD in our cohort of patients compared to other studies (ranging from 40% to 50% in CD patients)[25,26]. Since location remains relatively stable during the course of the disease[24], the low rates of ileal CD seen in our patients could be attributable to the impact of interobserver disagreement[30], variation of disease location among different backgrounds[31] and even differences in diagnostic techniques. The present study suggests a relationship between disease location and different Nod-like receptor molecules, with relevant clinical implications. Distinctive subcellular location, trafficking, and expression of each Nod-like receptors could be confining the association of NOD1 and NOD2 with location of the disease at different parts of the gastrointestinal tract. In healthy humans, NOD2 is expressed in Paneth cells within the crypts of the small intestine but not in colonic epithelium[6]. On the other hand, colon intestinal epithelial cells constitutively express NOD1[32]. NOD1 or NOD2 prevalence in colon or ileum could also be due to the predominance of different intracellular organisms or enteroinvasive bacteria for which they are receptors[33]. Further studies are needed to better clarify this subject.

A higher genetic load of NOD2 mutations increased the susceptibility to CD and determined an aggressive course of the disease. Although CD behavior is a dynamic process progressing towards complicated forms in 80% of patients[24], the presence of NOD2 variants could predict a stricturing and penetrating disease[14]. In addition, NOD2 variants have been associated with early surgery due to stenosis and with CD recurrence after surgery[13]. No association was established between the NOD1 polymorphism and disease behavior.

When comparing these results with other published genotype/phenotype associations, potential confounding factors should be taken into account to understand the differences. Agreement in Montreal classification, modification of the phenotype during follow-up, as well as the mixture of populations in some studies could be masking the particularities of each population. Our study adds two novel approaches to previous studies. First, two functionally related genes were analyzed for the first time in the same population, and second, the association phenotype/NOD1 genotype was established after ruling out the strong influence of NOD2. Although this work emphasized the importance of NOD1 and NOD2 on CD disease phenotype, the complexity of IBD genetics should not be ignored. Individual combinations of genetic risk factors from other molecules such as OCTN, DLG5, TUCAN, MDR1, TNF and TLRs[34-40] would depict a specific clinical picture for each CD patient.

ACKNOWLEDGMENTS

We thank Carolyn Newey for editorial assistance and Ignasi J Gich for advice in the statistical analysis.

COMMENTS

Background

Crohn's disease (CD) is a chronic inflammatory disorder of the gastrointestinal tract. Genetic and environmental factors have been suggested to predispose to CD. CD association with NOD2 mutations has been widely replicated but with remarkable heterogeneity across populations. Similarly to NOD2, NOD1 has a direct function as intracellular pattern recognition molecules but detecting different substructures from bacterial peptidoglycan. The present study examines the genetic variation in NOD1 and NOD2 and their respective influences on the CD phenotype in a cohort of well-characterized CD patients.

Research frontiers

Individual combinations of genetic risk factors from NOD2, NOD1 and other molecules, such as OCTN, TNF and TLRs, would depict a specific clinical picture for each CD patient.

Innovations and breakthroughs

This study adds two novel approaches to previous studies. First, NOD2 and NOD1 were analyzed for the first time in the same population and second, the association phenotype/NOD1 genotype was established after ruling out the strong influence of NOD2. The present results suggest a relationship between disease location and different Nod-like receptor molecules.

Applications

This is an association study that compares the allele or genotype frequencies of two genes between affected and unaffected individuals of Crohn's disease. Exploring new gene variants associated with inflammatory bowel disease would make possible the identification of proteins located in certain pathophysiological pathways.

Terminology

NODs are cytosolic proteins that contain a nucleotide-binding oligomerization domain (NOD). As sensors of bacterial components, NOD1 and NOD2 are triggered by host recognition of specific motifs in bacterial peptidoglycan and, upon activation, induce the production of proinflammatory mediators.

Peer review

This is a very well written paper. Authors examined genetic variation of NOD1 and NOD2, their respective influences on Crohn's disease phenotype and gene-gene interactions. This study suggests population differences in the inheritance of risk NOD1 polymorphism and NOD2 mutations.