B. Vander Cruyssen & H. Peeters contributed equally to this work.
CARD15 polymorphisms are associated with anti-Saccharomyces cerevisiae antibodies in caucasian Crohn's disease patients
Article first published online: 30 MAR 2005
Clinical & Experimental Immunology
Volume 140, Issue 2, pages 354–359, May 2005
How to Cite
Cruyssen, B. V., Peeters, H., Hoffman, I. E. A., Laukens, D., Coucke, P., Marichal, D., Cuvelier, C., Remaut, E., Veys, E. M., Mielants, H., De Vos, M. and De Keyser, F. (2005), CARD15 polymorphisms are associated with anti-Saccharomyces cerevisiae antibodies in caucasian Crohn's disease patients. Clinical & Experimental Immunology, 140: 354–359. doi: 10.1111/j.1365-2249.2005.02759.x
- Issue published online: 30 MAR 2005
- Article first published online: 30 MAR 2005
- Accepted for publication 6 January 2005
- Crohn's disease;
Carriage of CARD15 gene polymorphisms and the serological marker anti-Saccharomyces cerevisiae antibodies (ASCA) are two markers for Crohn's disease (CD). Similar phenotypes have been associated with both markers. In the present study we analysed whether both markers were associated with each other and, if so, whether this association could be explained by a direct link or by an indirect association with those phenotypes. Therefore, we included 156 consecutive Caucasian CD patients and assessed the prevalence of the three common single nucleotide polymorphisms in the CARD15 gene. Serum samples were analysed for IgA and IgG ASCA by ELISA. CD patients with CARD15 polymorphisms were more frequently ASCA positive (OR 2·7 (1.4–5.2); P = 0·002) and had higher titres for ASCA IgA (P = 0·005) and ASCA IgG (P < 0·001) compared to patients carrying the wild type polymorphisms. Multivariate analysis demonstrated that this association was independent from ileal disease, penetrating disease and stricturing disease, the need for resective bowel surgery, familial cases, smoking habits and early age at onset. Homozygotes or compound heterozygotes for CARD15 polymorphisms had significantly more frequent ASCA positivity compared to single heterozygotes (OR 9·1 (1.1–74.2), Pc (corrected P-value) = 0·030). These data indicate that there is a significant association between the carriage of CARD15 polymorphisms and ASCA, independent of the described phenotypes. Moreover, ASCA positivity is more frequent in CD patients carrying 2 CARD15 polymorphisms compared to single heterozygotes.
Although the exact pathogenesis of Crohn's disease (CD) remains unclear, it is well accepted that an impaired microbial immune response, triggered by environmental and genetic factors, is important [1–4]. Recently, 3 polymorphisms in the CARD15 gene (two missense mutations (R702W and G908R) and one frame shift mutation (1007fs)), were independently associated with CD [5,6]. It has been estimated that heterozygotes have a 3-fold risk to develop CD and homozygotes or compound heterozygotes a 40-fold risk to develop the disease [5,6]. Genotype-phenotype studies showed different possible associations of these CARD15 polymorphisms with ileal and stricturing disease [2,7–10], familial cases  and early onset of disease .
Anti-Saccharomyces cerevisiae antibodies (ASCA) are directed against the cell wall mannan of Saccharomyces cerevisiae, commonly known as baker's or brewer's yeast . ASCA are considered as a serological marker for CD. However, their pathophysiological role is not yet clear . Sensitivity and specificity of ASCA for CD range from 39% to 65% and 80% to 97·5%[15–18], respectively. Combinations of ASCA with other serological markers as pANCA, I2 and OmpC antibodies are under investigation to obtain a better serological diagnostic tool for IBD . In CD patients, ASCA are linked with earlier onset of disease , ileal involvement, penetrating and stricturing disease and need for resective bowel surgery [19–22]. One study also pointed at a possible negative association with smoking behaviour . Previous family and twin studies already suggested a genetic influence on ASCA formation. Indeed, family and twin studies revealed that unaffected twins and unaffected relatives have higher ASCA titres compared to healthy controls [20,24–27]. The aim of the present study was to analyse whether CARD15 and ASCA are related and if so, whether this association can be explained by a direct link or an association by phenotypes.
Materials and methods
Study population and assessment of clinical characteristics
The study population consisted of 156 consecutive, Caucasian CD patients (57 male, 99 female). The diagnosis of CD was based on clinical, endoscopical, histological and/or radiological findings. All patients were seen by a gastroenterologist. Family and personal medical history, onset of disease, localizations of inflammation, amount and types of surgical interventions, history or current presence of fistulae, past and present intake of medication and smoking habits were recorded. Familial Crohn's disease was considered when at least one first, second or third degree relative also had a proven Crohn's disease. The specific anatomical localization was based on documented areas of macroscopical and/or microscopical inflammation and was subdivided into 3 categories: ileal, ileocolonic or colonic disease. Penetrating disease was defined by the history or actual presence of fistulae, abscesses or perforations. Fistulae secondary to surgery were excluded. Stricturing disease was considered in those patients without fistulae, who had radiological or surgical evidence for stenosis.
Additionally, we tested ASCA in 188 controls including 24 osteoarthritis patients, 56 rheumatoid arthritis patients and 108 blood donors of whom 87 could also be typed for CARD15 variants.
Detection of ASCA by ELISA
The Medizym ASCA kits (Medipan Diagnostica, Germany) were used. Tests were performed according to instructions from the manufacturer, including the use of cut-offs that were determined at 20 U/ml for both IgA and IgG ASCA. Briefly, serum was diluted 1 : 50 and applied to the microtitre plates (100 µl/well), coated with cell wall mannan from a mixture of different Saccharomyces strains. The plates were incubated for one hour at 37°C. To remove unbound serum components, plates were washed five times. Consequently 100 µl of conjugate, specific for either IgG of IgA, coupled with horseradish peroxidase was added, followed by an incubation period of 30 min at 37°C. The plates were washed again five times, after which substrate was added (3,3′,5,5′-tetramethylbenzidine in citrate buffer containing hydrogen peroxide). Plates were incubated in the dark at room temperature for 10 min. The reaction was stopped using a stop solution containing sulphuric acid, turning the colour of the solution from blue to yellow. Plates were read at a wavelength of 450 nm.
ASCA IgG and IgA levels were determined using a standard curve, for which the manufacturer supplied calibrators. Study personnel were blinded for diagnosis during these assays. Each sample was tested in duplo and 2 positive control samples were run on each plate. The mean values for the IgG-samples were 60 U/ml and 29 U/ml with a coefficient of variation (CV%) of 2% and 4·5%, respectively. The mean values for the ASCA IgA samples were 18 U/ml and 28 U/ml with a CV% of 6·9% and 5·6%, respectively. The mean CV% between duplo's of all samples was 3·15% for ASCA IgA and 4·23% for ASCA IgG. Unless otherwise specified, ASCA were considered positive when either ASCA IgA or IgG was positive.
Genotyping of R702W, G908R and 1007fs and sequencing
Genomic DNA was extracted from whole blood using Qiagen blood and cell culture DNA kit (Qiagen, Germany). All patients were genotyped for R702W, G908R and 1007fs using a RFLP-PCR technique, followed by separation on 2·5% agarose gel. The missense mutation R702W (GenBank accession number G67950) abolishes the restriction site for MspI (5′-CAGCCCTGATGACATTTCTCTT-3′ and 5′-AGC CGCTCCTCCTGCATCTCGTA-3′), resulting in an intact 130-bp band for mutant alleles compared to two bands of 54- and 76-bp for wild type alleles. The missense mutation G908R (GenBank accession number G67951) creates a restriction site for HinP1l. The frame shift mutation 1007fs (GenBank accession number G67955) creates a restriction site for NlaIV (5′-CTGAGCCTTTGTTGATGAGC-3′ and 5′-TCTTCCAACCACATCCCCATT-3′). The presence of a mutant allele results in two bands of 219 and 41 base pairs, while the wild type allele produces a single 260-bp product.
Statistical analysis was performed using SPSS software (SPSS inc., Chicago, Illinois, USA). Groups were compared with Mann–Whitney U-test because normality was not achieved. Dichotomous data were analysed using Pearson's χ2 test or with Fisher's exact test when the expected count was less than five in at least one cell. Odds ratios were calculated with their corresponding 95% confidence interval (CI). When indicated, a corrected P-value (Pc) was calculated using Bonferroni's correction. We also calculated conditional odds ratios and their corresponding 95% CI by binary logistic regression.
The study was approved by the local ethics committee. All patients gave written informed consent.
Thirty-eight patients had ileal disease, 79 patients had ileocolonic disease and 39 patients had only colonic involvement. Seventy-three patients needed resective small bowel surgery, 92 patients had penetrating disease and 33 patients had stricturing disease. Thirty patients had at least one affected relative. Forty patients were ex-smokers and 44 patients smoked at the time of evaluation. Mean age was 38 years (range 18–80 years) and mean age at diagnosis was 27 years (range 9–66 years). In 48 patients, onset of disease was before the age of 20.
Prevalence of CARD15 mutations
Seventy-seven (49·3%) of 156 CD patients carried CARD15 polymorphisms. Forty-three CD patients carried at least one R702W polymorphism, 14 patients carried at least one G908R polymorphism and 27 patients carried at least one 1007fs polymorphism. Fourteen patients carried two polymorphisms of which 7 patients were homozygous and 7 patients compound heterozygous (Table 1). In a local control population, 19/87 (21·8%) controls carried at least one CARD15 variant.
|CD patients (n = 156) n (%)|
|R702W heterozygotes||33 (21·3%)|
|G908R heterozygotes||9 (5·8%)|
|1007fs heterozygotes||21 (13·5%)|
|R702W homozygotes||4 (2·6%)|
|G908R homozygotes||3 (1·9%)|
|R702W + G908R compound heterozygotes||1 (0·6%)|
|G908R + 1007fs compound heterozygotes||1 (0·6%)|
|R702W + 1007fs compound heterozygotes||5 (3·2%)|
Prevalence of ASCA positivity
Eighty-two (52·6%) CD patients were ASCA positive, 71 (45·5%) patients had ASCA IgA and 57 (36·5%) patients ASCA IgG. In a control population, 5/188 (2·6%) tested positive for ASCA (3/24 OA patients, 0/56 RA patients and 2/108 blood donors), confriming the high specificity of the test. Only 1/87 controls who were typed for CARD15 polymorphisms had ASCA. This control carried the 1007fs CARD15 polymorphism.
Univariate analysis of association between CARD15 genotype and ASCA positivity
Fifty of 77 (64·9%) carriers of CARD15 polymorphisms had positive ASCA IgA or IgG, in contrast to 32 (40·5%) of 79 wild type patients (OR 2·7; 95% CI (1.4–5.2), P = 0·002).
Forty-four (57·1%) of 77 carriers of CARD15 polymorphisms were positive for ASCA IgA versus 27 of (34·2%) 79 wild type patients (OR 2·87; 95% CI (1.49–5.50), P = 0·004) and 37 (48·1%) of 77 carriers tested positive for ASCA IgG versus 20 (25·3%) of 79 wild type patients (OR 2·7; 95% CI (1.4–5.4), P = 0·003) (Table 2).
|ASCA positivity (IgA or IgG)||ASCA IgA positive||ASCA IgG positive|
|Carriage of CARD15 polymorphisms|
|No||32 (40·5%)||P = 0·002||27 (34·2%)||P = 0·004||20 (25·3%)||P = 0·003|
|Yes||50 (64·9%)||44 (57·1%)||37 (48·1%)|
|Total||82 (52·6%)||71 (45·5%)||57 (36·5%)|
Carriage of 1 versus 2 CARD15 polymorphisms and ASCA positivity
Carriers of 2 CARD15 polymorphisms were more frequently ASCA positive compared to patients who carried only 1 polymorphism: 13 (93%) of 14 patients who carried 2 CARD15 variants were ASCA positive compared to 37 (58·7%) of 63 carriers of 1 polymorphism (OR 9·135, 95% CI (1.1–74.2), P = 0·015, Pc = 0·030). When looking at ASCA IgA and ASCA IgG separately, we could again find higher prevalences of ASCA IgA or ASCA IgG in those patients who carried 2 polymorphisms, but it did not reach significance for ASCA IgG (IgA: OR 5·81, 95% CI (1·0–28·1), P = 0·017, Pc = 0·034; IgG: OR 3·33, 95% CI (0·9–77·8) P = 0·053, Pc = 0·106).
Association between carriage of CARD15 polymorphisms and ASCA titres
When considering the ASCA titres as continuous data, they were significantly higher in patients who carried a common CD associated CARD15 polymorphism compared to those who carried only wild type polymorphisms (ASCA IgA: median 23·5 U/ml versus 14·4 U/ml, P = 0·005; ASCA IgG: median 19·4 U/ml versus 7·2 U/ml, P < 0·001).
Carriage of 1 versus 2 CARD15 polymorphism and ASCA titres
Carriers of 2 CARD15 variants had higher titres compared to patients who carried only 1 variant for ASCA IgA (Median 46·3 U/ml versus 20·5 U/ml, P = 0·033, Pc = 0·066) and ASCA IgG (Median 31·6 U/ml versus 17·5 U/ml, P = 0·187, Pc = 0·37), but this was not statistically significant.
Association between carriage of different CARD15 polymorphisms and ASCA positivity
The data on ASCA positivity in relation to the different CARD15 polymorphisms are shown in Table 3. These results provide no arguments for a higher association between one specific polymorphism and ASCA positivity.
|Carriage of at least 1 R702W polymorphism||15||28||P = 0·575|
|Carriage of at least 1 G908R polymorphism||3||11|
|Carriage of at least 11007fs polymorphism||10||17|
Multivariate analysis by logistic regression
Multivariate analysis was performed with a logistic regression model, using ASCA as dependent variable and familial CD, onset before the age of 20 years, ileal, penetrating and stricturing disease, need of resective bowel surgery, present smoking and carriage of CARD15 polymorphisms as covariates (Table 4). In this model, carriage of CARD15 polymorphisms and ileal disease were significant and independent predictors for ASCA positivity.
|Carriage of CARD15 polymorphisms||0·028||2·26||1·09–4·67|
|Age of onset before 20||0·608||1·23||0·56–2·71|
|Need of resective surgery||0·143||1·91||0·82–4·56|
This study describes the association between a genetic predictor of CD (carriage of a CARD15 polymorphism) and a serological marker for CD (ASCA).
So far, data on the association between CARD15 and ASCA are very disparate. Data from Abreu et al.  suggested an association between the 1007fs CARD15 polymorphism and ASCA positivity, but the authors eventually rejected that hypothesis since this association could not be confirmed in a second, smaller test population (in which however, the prevalence of ASCA positivity was significantly different from the prevalence in the first test population) . In a Scottish population, where the prevalence of CARD15 polymorphisms is much lower, no significant association between CARD15 polymorphisms and ASCA could be found, although there was a trend towards an association between the carriage of 2 copies and ASCA positivity . Another recent study in a Dutch population, described an association between ASCA positivity or atypical pANCA positivity and carriage of the 1007fs or G908R polymorphism .
In our present study, we show that there is an association between the carriage of at least one of the three common CARD15 polymorphisms (R702W, G908R and 1007fs) and ASCA. We also found that patients, carrying 2 CARD15 polymorphisms were more frequently ASCA positive compared to patients carrying 1 polymorphism. Multivariate analysis, showed that the presence of CARD15 polymorphisms is a significant predictor of ASCA positivity, conditionally independent from the related phenotypes. This indicates that this association between the carriage of CARD15 polymorphisms and ASCA can not be explained only by an indirect link with the common associated, clinical phenotypes. In the same analysis, ileal disease was also a significant, independent predictor of ASCA. This might suggest that other (environmental?) factors might also contribute to the formation of ASCA. The fact that data on a possible association between carriage of CARD15 polymorphisms and ASCA are so disparate might be explained by differences in environmental factors and different genetic backgrounds. Associations between genetic markers and serologic markers have previously been described for several diseases. In insulin dependent diabetes, disease associated autoantibodies are associated with susceptibility HLA class II alleles . Coeliac associated antibodies can identify healthy first-degree relatives who express coeliac associated HLA haplotypes . In rheumatoid arthritis, rheumatoid factor is associated with the HLA shared epitope and the − 2849 IL-10 promoter polymorphism [32,33]. Carriage of different HLA class II alleles is associated with a specific antibody response to nuclear antigens, in particular anti-SSA/Ro52, both in primary Sjögren's syndrome and lupus [34,35] and may be more strongly associated with the antibody subsets than with the disease status itself . Other nuclear antibodies (anti-SSB/La), associated with HLA class II genes, have also recently been linked with carriage of polymorphisms in the genes for transforming growth factor β and tumour necrosis factor α in patients with primary Sjögren's syndrome . Different hypotheses can be generated to explain how the carriage of different HLA types or different gene polymorphisms on the promoters of cytokines, might influence antibody responses [33,35]. How the different polymorphisms on the CARD15 gene are involved in the antibody response against mannan from Saccharomyces cerevisiae is still unclear. The CARD15 gene encodes for the Nod2/CARD15 protein, a member of the Apaf-1/Ced-4 family of apoptosis regulators and an intracellular protein expressed in monocytes, macrophages, epithelial cells, granulocytes and dendritic cells. Nod2 activates the NF-κB pathway after stimulation by bacterial products. Initial reports suggested lipopolysaccharide as a possible ligand for Nod2, but recently, 2 independent groups highlighted muramyl dipeptide as the major activator of the Nod2-receptor [6,38–41]. The activation of the NF-κB–pathway is followed by an enhanced production and secretion of proinflammatory cytokines .
It has been suggested that the described polymorphisms in the CARD15 gene give rise to an impaired NF-κB activation by a deficient recognition of microbial antigens. This might result in an impaired killing of intracellular microbes, which leads to inflammation [38,40]. Whether ASCA is just a side-effect of this inflammation or whether ASCA have themselves a pathological role by cross-reactivity with self-antigens in human gut or other tissues is still unknown .
In conclusion, our data suggest that there is an association between the presence of CARD15 gene polymorphisms and the serological marker ASCA in a West-European CD population. This could be interpreted in a double way. CARD15, as an immune response gene, may modulate in some way humoral immunity predisposing to the generation of ASCA. Or, carriage of CARD15 polymorphisms and ASCA may be both associated with a particular, as yet undefined, phenotypical subset of CD. However, we proved in the present study that the association between carriage of CARD15 polymorphisms and ASCA could not be explained by an indirect link with the common CARD15 and ASCA related phenotypes.
We thank Dr O. Thas, PhD, Department of Applied Mathemathics of the Ghent University, for his help with the statistical analysis and Ms Virgie Baert for excellent technical contribution. This study was supported by a concerted action grant GOA 2001/12051501 of the Ghent University, Belgium. Ilse Hoffman is supported by a research grant of the ‘Bijzonder Onderzoeksfonds’, Ghent University.
- 21Diagnostic Value of anti-Saccharomyces cerevisiae and anti neutrophil cytoplasmic auto antibodies in inflammatory bowel disease. Am J Gastroenterol 2001; 96: 730–4., , , , , .Direct Link:
- 36The genetics of Human lupus. In: WallaceDJ, Hannahs HahnB, eds. Dubois’ Lupus Eythematosus. Philadelphia: Lippincott Williams & Wilkins, 2002: 101–2..