No association between the functional CARD4 insertion/deletion polymorphism and inflammatory bowel diseases in the German population

Inflammatory bowel diseases (IBD, OMIM 601458) are represented by two main subtypes, Crohn's disease (CD, OMIM 266600) and ulcerative colitis (UC, OMIM 191390). The first and most widely replicated susceptibility gene for CD is CARD15 (caspase recruitment domain family, member 15) that encodes NOD2 (nucleotide binding oligomerisation domain protein 2), a protein involved in the pathogen associated molecular pattern recognition system (for review see Schreiber and colleagues¹). Identification of functional CARD15 variants revealed the important role of impaired barrier integrity and host defence in the pathogenesis of CD and other inflammatory diseases.¹ Recently, an association between IBD and variants in the CARD4 gene, which encodes NOD1, a structural homologue of NOD2, has been demonstrated in the British population.² In that study, McGovern et al identified the common (deletion) allele of the marker ND₁+32656 as a risk factor for IBD. Conversely, the minor ND₁+32656 allele, in haplotype combination with its adjacent marker rs2907748, exhibited a protective effect.² ND₁+32656 may alter the splicing of CARD4 by affecting the binding of an unknown nuclear factor.³ The more frequent deletion variant could result in a NOD1 protein with reduced numbers of the leucine rich repeats that are essential for pathogen recognition. The odds ratio (OR) associated with this allele was calculated to be 2.0.² Association of CARD4 polymorphisms has also been observed with asthma and atopic eczema,³,⁴ suggesting that CARD4 may be a common barrier disease susceptibility gene.

In the present study, we evaluated whether the variants previously reported in the British population are also associated with IBD in Germany. In addition to ND₁+32656, we included the adjacent single nucleotide polymorphisms (SNPs) ND₁+27606 (rs2075822) and ND₁+45343 (rs2907748) in our analyses. ND₁+233 and ND₁+21984, which were only weakly associated with IBD in the study of McGovern et al, were not examined here. SNP genotyping was undertaken on a case control (1015 IBD patients (676 CD, 344 UC), 886 unrelated controls) and an independent family based panel (775 mother‐father affected child trios for IBD (328 CD, 447 UC)) using TaqMan Assays‐by‐Design (Applied Biosystems, Foster City, California, USA). Given our sample sizes, our study provided >96% power to detect an OR of 2.0, assuming a risk allele frequency of 79% and 1% type I error. However, no significant association was detected with either the case control or independent family based CD/UC samples (tables 1, 2). McGovern et al also reported a strong association with specific disease subgroups.² We therefore stratified our sample for patients with early onset IBD (n = 449 <25 years) and patients with fistulae and stenoses (n = 491). Analyses of both subgroups showed no significant associations (data not shown).

Table 1 Summary of single marker association statistics for CARD4 (caspase recruitment domain family, member 4).

Phenotype	Variant*	MAFcontrols†	MAFcases†	CCA‡	CCG§	TDT¶
				p Value	p Value	p Value
CD	rs2075822	0.81	0.80	0.569	0.308	1.000
	ND1+32656	0.79	0.79	0.847	0.977	0.091
	rs2907748	0.79	0.77	0.420	0.578	0.306
UC	rs2075822	0.81	0.78	0.152	0.357	0.480
	ND1+32656	0.79	0.76	0.124	0.295	0.951
	rs2907748	0.79	0.75	0.073	0.195	0.956
IBD	rs2075822	0.81	0.79	0.263	0.327	0.580
	ND1+32656	0.79	0.78	0.510	0.804	0.248
	rs2907748	0.79	0.77	0.132	0.279	0.547

*All three assays had a call rate >95% and showed no significant deviations from the Hardy‐Weinberg equilibrium in the control sample.

†MAF, major allele frequency in cases and controls.

‡CCA, case control analysis for alleles.

§CCG, case control analysis for genotypes.

¶TDT, transmission disequilibrium test.

Table 2 Two marker haplotype frequencies, transmission, and association statistics for CARD4 (caspase recruitment domain family, member 4) in inflammatory bowel diseases.

Variation	Haplotype*	fcontrols†	fcases†	p Value‡	fT§	fNT§	p Value¶	D′**
rs2075822 + ND1+32656	1 – 1	0.771	0.762	0.846	0.455	0.486	0.094	0.91
	1 – 2	0.042	0.040		0.140	0.101
	2 – 1	0.013	0.015		0.026	0.057
	2 – 2	0.174	0.184		0.380	0.357
ND1+32656 + rs2907748	1 – 1	0.784	0.775	0.727	0.473	0.519	0.719	0.98
	1 – 2	0.002	0.002		0.011	0.011
	2 – 1††	0.006	0.004		0.014	0.008
	2 – 2	0.208	0.220		0.503	0.462

*Allele 1 is defined as the major allele.

†Frequencies of haplotypes in cases and controls estimated by the expectation maximisation algorithm using COCAPHASE.[6]

‡Global significance value obtained after 10 000 permutations with COCAPHASE.

§Frequencies of transmitted (f_T) and non‐transmitted (f_NT) haplotypes observed using TDTPHASE.[6]

¶Global significance value obtained after 10 000 permutations with TDTPHASE.

**D′ value as a measure of linkage disequilibrium in the control sample.

††Protective haplotype previously identified by McGovern and colleagues.[2]

In conclusion, there is no evidence for an association of the IBD phenotype with the previously reported CARD4 susceptibility variants in the German population. The discrepant findings in the German and British samples may reflect true genetic heterogeneity, as has been demonstrated for CARD15 associations.⁵ This heterogeneity is partly evident in the different linkage disequilibrium (LD) patterns of the three polymorphisms analysed. While LD was incomplete in the British,² there was strong LD in the German patients (table 2). ND₁+32656 could therefore be a population specific marker for an as yet unidentified causative variant in the vicinity. Replication in other British and European samples will be necessary to further examine the role of CARD4 variants in IBD susceptibility.

Acknowledgements

This work was supported by grants from the German National Genome Research Network (NGFN) and the German Federal Ministry of Education and Research (BMBF).