These authors contributed equally to the work.
The +1059G/C polymorphism in the C-reactive protein (CRP) gene is associated with involvement of the terminal ileum and decreased serum CRP levels in patients with Crohn's disease
Version of Record online: 12 SEP 2006
Alimentary Pharmacology & Therapeutics
Volume 24, Issue 7, pages 1105–1115, October 2006
How to Cite
THALMAIER, D., DAMBACHER, J., SEIDERER, J., KONRAD, A., SCHACHINGER, V., PFENNIG, S., OTTE, J.-M., CRISPIN, A., GÖKE, B., OCHSENKÜHN, T., LOHSE, P. and BRAND, S. (2006), The +1059G/C polymorphism in the C-reactive protein (CRP) gene is associated with involvement of the terminal ileum and decreased serum CRP levels in patients with Crohn's disease. Alimentary Pharmacology & Therapeutics, 24: 1105–1115. doi: 10.1111/j.1365-2036.2006.03093.x
- Issue online: 12 SEP 2006
- Version of Record online: 12 SEP 2006
- Publication data Submitted 19 June 2006 First decision 9 July 2006 Resubmitted 13 July 2006 Accepted 14 July 2006
Background Serum C-reactive protein (CRP) levels influence the response to anti-tumour necrosis factor (TNF) therapies.
Aim To analyse the influence of the +1059G/C CRP polymorphism on CRP serum levels and disease susceptibility in patients with Crohn's disease (CD).
Methods Using restriction fragment length polymorphism (RFLP) analysis, genomic DNA from 241 CD patients and 199 unrelated controls was analysed for the +1059G/C substitution in the CRP gene and the common caspase-activation recruitment domain 15 (CARD15) variants.
Results Homozygous C/C carriers were detected only among CD patients (P = 0.066). Patients with ileal involvement (L1 and L3 phenotype) were found in only 58.4% of patients with the wildtype G/G genotype but in 88.2% of the heterozygous G/C carriers (OR 5.26; 95% CI 1.19–23.92) and four of the five C/C homozygous carriers (80%; OR 4.55; 95% CI 1.64–16.67; P = 0.008 for hetero- and homozygous carriers vs. wildtype) which was independent of the presence of CARD15 variants. Increased CD activity was associated with increased CRP serum levels (P < 0.005). For Crohn's disease activity index (CDAI) < 150, C/C homozygosity for the +1059 G/C polymorphism was associated with significantly lower CRP serum levels (P < 0.01).
Conclusions The C allele of the CRP +1059G/C polymorphism is associated with decreased serum CRP levels and increased likelihood of disease involvement of the terminal ileum in CD patients.
Human C-reactive protein (CRP) is the prototype of acute phase proteins released during the early host response to inflammatory injury.1 In addition to the release of proinflammatory cytokines and complement activation products, the liver secretes large amounts of CRP into the circulation during acute inflammation such as active Crohn's disease (CD). Therefore, blood CRP levels are frequently monitored to determine the degree of inflammation in these patients. CRP, which is a member of the pentraxin family, is a pattern recognition molecule and is composed of five identical subunits.2 In the presence of Ca2+, each subunit interacts with phosphocholine.3 The B (binding) face has five ligand-binding sites that can bind to a variety of biologic ligands including bacteria forming CRP-ligand complexes. On the opposite side of each subunit is an effector molecule-binding site that mediates the protein's interaction with C1q,4 Fcγ RI,5 and FcγRII.6
The gene coding for CRP has been mapped to 1q23. As the basal values of CRP appear to be significantly heritable,7–10 it is very likely that polymorphisms in the CRP gene and in genes controlling CRP expression influence CRP levels. There are five published functional polymorphisms which have an influence on CRP levels: a dinucleotide repeat in the only intron,11 a CRP (GC) +1059 single nucleotide polymorphism (SNP) in exon 2,12 a CRP (CT) +1444 polymorphism13 and a CRP 4AG substitution,14 both in the 3′ untranslated region (UTR), as well as a −717GA exchange in the promoter region.13 A recent study demonstrated that, in healthy male blood donors, the carriers of the CRP +1059 C-allele have significantly lower CRP values than G/G homozygotes,15 while no significant difference was found in women.15 However, in men +1059 G/G homozygotes had only low CRP concentrations if they also were T allele carriers of the IL-1B +3954 polymorphism.15 The genotype CRP +1059 GG/IL-1B +3954 CC was associated with an almost threefold increased risk of a higher baseline CRP value.15 Therefore the authors concluded that both IL-1B +3954 (CT) and CRP +1059 (GC) polymorphisms influence baseline CRP values and act independently of each other in male subjects.15 The 1444C/T substitution located in the 3′ UTR was also associated with variations in CRP serum concentrations: CRP concentrations were higher in 1444T/T homozygous healthy volunteers, both at baseline and after exercise, than in 1444C carriers.13 In patients undergoing coronary artery bypass surgery, mean CRP concentrations after surgery were also higher in 1444T/T homozygotes.16 Recently, in a cohort of patients suffering from systemic lupus erythematosus (SLE) and their family members, it was confirmed that the CRP haplotype containing the 1444T allele was associated with high basal CRP concentrations.14 In the same study, a new polymorphism named CRP 4, also located in the 3′ UTR, was identified and linked with SLE; the rare CRP 4A allele was associated with low basal CRP concentrations, having a gene dose effect.14 In contrast, a −717G/A substitution in the CRP gene was not associated with variations in CRP concentrations.13
Recently, we demonstrated that polymorphisms in certain genes involved in the acute phase response during inflammation are associated with particular CD subtypes.17–19 Therefore, we analysed if polymorphisms in the gene encoding CRP, the prototypic acute phase response protein, are also associated with CD and certain CD phenotypes. We focused on the CRP (GC) +1059 SNP in exon 2, which is so far the most extensively characterized CRP SNP and has been shown to decrease CRP serum levels15 while in contrast no association for three other CRP polymorphisms (−717G/A, 1444C/T, CRP 4A/G) with CD and CRP levels after infliximab treatment could be demonstrated in CD patients.20 Moreover, a recent study demonstrated that patients with CD and a persistently low CRP despite active disease were characterized by an almost exclusive ileal disease distribution.21 Given the previously described CRP lowering effect of the +1059G/C CRP polymorphism,15 we therefore hypothesized that this polymorphism may be a risk factor for ileal disease location.
In this study, 241 patients with CD and 199 healthy, unrelated controls were included after giving written informed consent. All study participants were Caucasians. The diagnosis of CD was assessed according to established clinical guidelines and endoscopic, radiological, and histopathologic criteria. Cases with an indeterminate colitis were excluded from the study. Severity of disease in CD patients was defined using the Vienna classification regarding age at diagnosis (A), behaviour (B) and localization of the disease (L)22 at the time of patient enrollment. Since disease behaviour as defined by the Vienna classification22 changes during the course of the disease,23, 24 we also performed multivariate analysis to account for disease duration in these patients (see Statistics). By retrospective analysis of clinical charts and the evaluation of detailed questionnaires, demographic and routine clinical data (including behaviour and localization of CD, disease-related complications, and the use of immunosuppressive therapy e.g. azathioprine, 6-mercaptopurine, 6-thioguanin, methotrexate, or infliximab) were collected by two independent investigators blind to CRP genotype. The study was approved by the Ethics Committee of the Medical Faculty of the University of Munich.
DNA extraction and genotyping of the CRP polymorphism
Genotyping of the +1059G/C CRP polymorphism was performed as previously described.25 Briefly, a single 50-μL PCR was performed, using approximately 200 ng of genomic DNA and 15 pmoles each of the following primers: CRP-1059F 5′-GATCTGTGTGATCTGAGAAACCTCT-3′ and CRP-1059R 5′-GAGGTACCAGAGACAGAGACGTG-3′. Other conditions were as follows: 2.0 mm MgCl2, 175 μm of each dNTP, 1.5 U Taq polymerase and 1x buffer (Roche Diagnostics, Mannheim, Germany). Amplification reactions were performed using 35 cycles of 95 °C, 69 °C, and 72 °C for 30 s each, preceded by a single cycle of 95 °C for 3 min and followed by a single cycle of 72 °C for 10 min. In order to genotype the +1059G/C polymorphism, 5 μL of the resulting 744 bp PCR product were then digested overnight with 3 U of Mae III (Roche) at 55 °C. Digestion of the less common 1059C allele produced two smaller fragments with sizes of 434 and 310 bp. Digestion of the more common 1059G allele resulted in three fragments 310, 233 and 201 bp in length. Restriction digests were analysed by electrophoresis of the digestion mixture in a 2% agarose gel stained with ethidium bromide.
Genotyping of the CARD15/NOD2 variants
Caspase-activation recruitment domain 15 (CARD15) exons 4, 8, and 11 were amplified and sequenced as previously described.17 Sequences were analysed on an ABI PRISM 377 DNA Sequencer (Applied Biosystems, Foster City, CA, USA) using the Sequence Analysis program version 3.4.5 (Applied Biosystems).
Serum CRP measurements
Serum for CRP measurements was collected together with whole blood for genotyping from all patients. At the day of the CRP measurement, the Crohn's disease activity index (CDAI) as defined by Best et al.26 was determined. Patients, who did not fill out a provided patient diary containing information on CDAI calculation or were unable to give the required information for CDAI calculation, were excluded from the CRP serum level analysis. In 133 CD patients, for whom the CDAI was available, serum CRP concentrations were measured using a latex based high-sensitivity immunoassay (N Latex CRP mono, Dade Behring GmBH, Marburg, Germany, Behring Laser Nephelometer). The CD patients had different disease activities ranging from quiescent disease/remission to very active disease. As CRP serum levels are variable depending on the disease activity, we included the CDAI in the statistical analysis. Based on the CDAI and classifications used in previous studies,27 patients were grouped into three categories: CDAI < 150 (disease in remission), CDAI between 150 and 300 (moderate disease), and CDAI > 300 (active disease).
Fisher's exact test or χ2-test, where appropriate, were used for comparison between categorical variables. The Mann–Whitney U-test was used to compare continuous variables. Power calculations were performed using PS: Power and Sample Size Calculation Software (http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/PowerSampleSize). All tests were two-tailed. P-values <0.05 were considered as significant. To analyse the influence of the CRP polymorphism on various phenotypic characteristics, we first performed univariate analyses. Because of the exploratory character of this study, no correction for multiple testing was performed. Next, to adjust for potential confounding effects of CARD15 variants, age, disease duration and gender, we performed logistic regression analyses. Predictors were selected using forward selection based on Wald tests. The significance level for entry into the model was 0.05. Logistic regression analysis was performed using SAS 9.1 software for Linux (SAS Institute, Cary, NC, USA).
Patient characteristics and CRP genotype frequencies in CD patients and healthy controls
Table 1 shows the demographic characteristics of the 241 CD patients and the 199 controls. The Vienna classification22 was used to classify all CD patients according to age at diagnosis (A), disease location (L) and disease behaviour (B). The mean age at diagnosis (27.7 ± 11.7 years; Table 1) indicates an overall early disease onset with 83.2% of the patients diagnosed before the age of 40 years (Vienna A1; Table 2). As shown in Table 2, the majority of the patients investigated had a severe disease phenotype with 54.8% belonging to the Vienna group B3 (penetrating disease) and 25.7% belonging to the B2 phenotype (stricturing disease). The respective genotype frequencies regarding the CRP +1059G/C polymorphism in the 241 CD patients and 199 controls are shown in Table 3. The overall frequency of the three CRP genotypes in CD patients did not differ significantly from that observed in healthy controls and was not significantly different from the predicted Hardy–Weinberg distribution although there were no C/C homozygotes found in the control group compared to five C/C homozygous carriers in the CD group (P = 0.066). In the CD group, 90.9% were wildtype (G/G), 7.1% heterozygous (G/C) and 2.1% homozygous (C/C). In the control population, no homozygous C/C carriers were detected and 7.0% were G/C heterozygotes, while 93.0% had two wildtype alleles.
|CD (n = 241)||Controls (n = 199)|
|Male||117 (48.5%)||100 (50.3%)|
|Female||124 (52.5%)||99 (49.7%)|
|Mean ± s.d.||39.7 ± 12.3||46.4 ± 15.2|
|Age at diagnosis (year)|
|Mean ± s.d.||27.7 ± 11.7|
|Disease duration (year)|
|Mean ± s.d.||11.6 ± 8.2|
|Family history of IBD (%)|
|Positive||36 (14.9%)||0 (0%)|
|Negative||183 (75.9%)||199 (100%)|
|Data not available||22 (9.1%)||0 (0%)|
|(1) G/G wildtype (n = 219)||(2) G/C heterozygous (n = 17)||(3) C/C homozygous (n = 5)||(1) vs. (2) P-value||(1) vs. (3) P-value||(2) vs. (3) P-value||(1) vs.(2+3) P-value|
|n (%)||107 (48.9%)||9 (52.9%)||1 (20.0%)||0.805||0.371||0.323||0.825|
|Body mass index|
|Mean ± s.d.||23.0 ± 4.0||24.9 ± 3.3||21.0 ± 2.7||0.043||0.2194||0.032||0.231|
|Age at diagnosis|
|Mean ± s.d.||27.7 ± 11.9||28.1 ± 10.6||27.8 ± 8.8||0.884||0.974||0.957||0.883|
|Mean ± s.d.||11.7 ± 8.2||10.5 ± 8.5||8.2 ± 9.0||0.587||0.435||0.624||0.374|
|Age at diagnosis|
|<40 years (A1)||185 (84.5%)||13 (76.5%)||5 (100.0%)||0.489||1.000||0.535||0.759|
|>40 years (A2)||34 (16.4%)||4 (23.5%)||0|
|Terminal ileum (L1)||16 (7.3%)||4 (23.5%)||2 (40.0%)||0.044||0.053||0.585||0.008|
|Colon (L2)||46 (21.0%)||0||1 (20.0%)||0.050||1.000||0.227||0.087|
|Ileocolon (L3)||112 (51.1%)||11 (64.7%)||2 (40.0%)||0.322||0.679||0.609||0.510|
|Upper GI (L4)||45 (20.6%)||2 (11.8%)||0||0.536||0.586||1.000||0.265|
|Any ileal CD involvement*||163 (74.4%)||17 (100.0%)||4 (80.0%)||0.015||0.651||1.000||0.032|
|Non-stricturing/ Non-penetrating (B1)||40 (18.3%)||6 (35.3%)||1 (20%)||0.109||1.000||1.000||0.155|
|Stricturing (B2)||60 (27.4%)||1 (5.9%)||1 (20.0%)||0.080||1.000||0.411||0.074|
|Penetrating (B3)||119 (54.3%)||10 (58.8%)||3 (60.0%)||0.804||1.000||0.411||0.074|
|Fistulas||120 (54.8%)||10 (58.8%)||3 (60.0%)||0.805||1.000||1.000||0.823|
|Stenosis/strictures||120 (54.8%)||10 (58.8%)||3 (60.0%)||0.176||0.641||1.000||0.154|
|Abscesses||82 (37.4%)||4 (23.5%)||2 (40.0%)||0.304||1.000||0.585||0.487|
|Surgery (related to CD)‡||129 (58.9%)||8 (47.1%)||3 (60.0%)||0.445||0.647||0.586||0.498|
|Extraintestinal manifestations§||154 (70.3%)||10 (58.8%)||4 (80.0%)||0.412||1.000||0.613||0.627|
|Pos. family history||35 (16.0%)||1 (5.9%)||0||0.482||1.000||1.000||0.214|
|Use of immunosupp. Agents¶||180 (82.2%)||15 (88.2%)||3 (60%)||0.744||0.227||0.210||1.000|
|CRP Genotype + 1059 SNP||CD (n = 241)||Controls (n = 199)||P-value|
|G/G||219 (90.9%)||185 (93.0%)||0.486|
|G/C||17 (7.1%)||14 (7.0%)||1.000|
|C/C||5 (2.1%)||0 (0%)||0.066|
Genotype-phenotype correlations: the +1059G/C CRP polymorphism is associated with ileal involvement
The disease characteristics of patients with the wildtype CRP genotype and of those heterozygous and homozygous for the +1059G/C CRP polymorphism are displayed in Table 2. There were no significant differences in the use of immunosuppressive agents, extra-intestinal involvement and other complications of CD such as abscess formation and fistulas between carriers of a C allele compared to those with the G/G wildtype genotype.
Patients with L1 and L3 phenotype were found in only 128 of the 219 patients with the wildtype G/G genotype (58.4%) but in 15 of 17 (88.2%) of the heterozygous G/C carriers (OR 5.26; 95% CI 1.19–23.92) and four of the five C/C homozygous carriers (80%; OR 4.55; 95% CI 1.64–16.67; P = 0.008 for hetero- and homozygous carriers vs. wildtype carriers). Moreover, 35 of the 45 G/G wildtype patients and both G/C heterozygous carriers with L4 phenotype (upper GI involvement) had additional disease involvement of the terminal ileum. Therefore, in the combined patient population with any ileal involvement (L1+L3 and ileal subgroup of L4), disease involvement of the terminal ileum was more often in G/C heterozygotes than in G/G wildtype patients (P = 0.015; Table 2). Similar findings were obtained in the combined group of G/C heterozygotes and C/C homozygotes compared to wildtype patients (P = 0.032; Table 2). Consistent with this finding was the observation, that none of the G/C heterozygotes had disease involvement restricted to the colon only (L2 phenotype according to the Vienna classification), while the L2 phenotype was found in 21.0% of patients with two G alleles (P = 0.05, Table 2).
Associations between serum CRP levels, disease activity and CRP +1059G/C genotype
Next, we analysed the influence of the +1059 G/C polymorphism on serum CRP levels, which were adjusted for disease activity using the Crohn's disease activity index (CDAI).26 Of 241 CD patients, CRP serum levels and corresponding CDAI values were available for 133 patients which were included in this analysis. Based on their CDAI levels, these patients were divided into three groups: CDAI < 150 (quiescent disease, remission), CDAI between 150 and 300 (moderate disease), and CDAI > 300 (active disease). With increasing disease activity, mean serum CRP levels increased significantly (Figure 1a; P < 0.005 vs. CDAI < 150) and serum CRP levels correlated with disease activity in all patients (r = 0.54). This correlation was higher in C/C homozygotes (r = 0.97) than in G/C heterozygotes (r = 0.49) and G/G wildtype patients (r = 0.51). However, as demonstrated in Figure 1b, in patients with CDAI values <150, the 1059 C allele was associated with lower mean CRP levels than observed in patients with the G/G wildtype genotype. The difference in the CRP serum levels between G/G wildtype carriers and carriers with C/C homozygous genotype was statistically significant (P < 0.01). For the two other categories with CDAI values >150, no sufficient number of patients was available for further analysis.
A recent study demonstrated that, in men, the carriers of the CRP +1059C allele have significantly lower CRP values than G/G homozygotes,15 while no significant difference was found in women.15 Therefore, we analysed if the observed differences in the serum CRP levels between the different genotypes were gender-dependent. The mean CRP serum concentrations did not differ between men (1.76 mg/dL) and women (1.72 mg/dL) although there was a higher proportion of female patients with CDAI values between 150 and 300 (43.2%) than male patients with such CDAI scores (25.4%, P = 0.044; Table 4). However, the CRP serum levels were significantly lower in male patients carrying one or two C alleles compared to wildtype G/G carriers (P = 0.013; Figure 1c). In contrast, there was no statistically significant difference in female patients (data not shown).
|CDAI||Male (n = 59)||Female (n = 74)||P-value|
|<150||36 (61.0%)||37 (50.0%)||0.224|
|150–300||15 (25.4%)||32 (43.2%)||0.044|
|>300||8 (13.6%)||5 (6.8%)||0.244|
Analysis of CARD15 variants
The CD involvement of the terminal ileum has also been described for several CARD15 variants.28–32 We recently demonstrated an increased frequency of intestinal stenoses especially in the terminal ileum in homozygous carriers of the CARD15 variant L1007fsinsC (3020insC).28 Therefore, we determined in all CD patients, whether the three common CD-associated CARD15 variants R702W, G908R, L1007fsinsC were present or not. The combined prevalence for these three variants in the CD cohort was 42.7%. The R702W mutation was the only CARD15 mutation which was represented significantly more often among G/C heterozygotes compared to wildtype carriers (35.3% vs. 13.7%, P = 0.029). However, in univariate analysis, there was no significant association of this mutation with ileal disease location. Similar genotype frequencies were observed for the 1007fs CARD15 variant, for which we demonstrated ileal involvement,28 in wildtype (17.8%) and heterozygous carriers (23.5%; P = 0.522), thereby confirming that the observed association between the +1059 G/C polymorphism and ileal disease location is independent of the CARD15 1007fs variant. However, the overall prevalence of CARD15 variants was significantly higher in +1059G/C CRP heterozygous carriers (76.5%) compared with G/G wildtype carriers (39.7%, P = 0.0043). In addition, in univariate analysis, there was a significant association of the presence of any CARD15 variant with ileal disease location (P = 0.021), although this association was much weaker than the association of the CRP +1059C allele with ileal involvement (P = 0.008). We therefore performed multivariate analysis to examine if the association of the +1059G/C CRP polymorphism with ileal involvement is independent of CARD15 variants. Multivariate analysis, which included the CRP and CARD15 genotype status, age, disease duration and gender, confirmed the association between the 1059C allele and the disease location in the terminal ileum (L1) (G/C heterozygotes: P = 0.04; OR 3.72; 95% CI 1.06–13.06; C/C homozygotes: P = 0.04; OR 7.16; 95% CI 1.06–48.52), thus demonstrating that the association between carriers of one or two C alleles and involvement of the terminal ileum is independent of the CARD15 genotype.
Although the etiopathogenesis of CD is still not completely understood, it is widely accepted that this disease is the result of an exaggerated immune response in a genetically susceptible host to a – so far undefined – luminal antigen, probably derived from the microbial flora. Variants in the CARD15 gene encoding nucleotide binding and oligomerization domain (NOD)2 have been identified as a major genetic risk factor for developing CD.33–35 However, these variants are found in up to 20% of the normal population and therefore account only for a small part of the overall risk of developing CD. We have previously shown that genes of several proteins involved in the acute phase response of inflammation such as toll-like receptor (TLR) 4,17CX3CR1,18 and macrophage migration inhibitory factor (MIF)19 are associated with certain CD phenotypes. In this study, we analysed if the (GC) +1059 polymorphism in the CRP gene encoding the prototypic acute phase response protein is also associated with CD or certain CD phenotypes. This polymorphism is so far the most extensively studied CRP SNP and has been shown to affect CRP serum levels in healthy male blood donors.15 However, in contrast to other CRP polymorphisms,20 its role in CD has not been analysed yet.
Our results demonstrate that the CRP +1059G/C polymorphism is not overrepresented in CD patients compared to the normal population, although homozygous C/C carriers were detected only among CD patients (P = 0.066). However, while only 74.4% of the patients with the wildtype G/G genotype had an involvement of the terminal ileum, the terminal ileum was affected in all of the heterozygous G/C (100%) and nearly all C/C homozygous carriers (80%). Similar genotype frequencies, in contrast, were observed for the CARD15 variant 1007fs, for which we recently demonstrated a strong association with ileal disease involvement,28 in wildtype (17.8%) and heterozygous carriers (23.5%) suggesting that the observed association between the +1059 G/C polymorphism and ileal disease location is independent of this particular CARD15 variant. In contrast, the prevalence of all CARD15 variants was significantly higher among G/C heterozygotes than in wildtype carriers. However, multivariate analysis confirmed that CD restricted to the terminal ileum only (L1) is significantly dependent on the CRP genotype, while the combined group with disease involvement of the terminal ileum and the ileocolon is significantly influenced by the presence of at least one CARD15 variant. Consistent with this finding was the observation, that none of the G/C heterozygotes had disease involvement restricted to the colon only (L2 phenotype according to the Vienna classification), while the L2 phenotype was found in 21.0% of patients with two G alleles (P = 0.05).
The predominant ileal disease location in carriers of the +1059 G/C polymorphism points to an association of this polymorphism with an impaired recognition of intestinal bacteria. Microbial density rises from approximately 108 bacteria/g in the distal ileum to 1011–1012 organisms/g in the colon36 which is the highest increase in luminal bacterial load in the entire gastrointestinal tract. The evidence for a pathophysiological role of certain luminal bacteria strains in the pathogenesis of CD is supported by a number of animal models.37–41 CRP itself binds with high avidity to bacteria.2 Similar to the association of the +1059 G/C polymorphism with ileal disease location, we recently demonstrated that the 1007fs variant in the CARD15 gene encoding NOD2, which recognizes bacterial muramyl dipeptide, is a strong predictor for ileal disease.28 In addition, we demonstrated that the T280M polymorphism in the CX3CR1 gene, which encodes the receptor for the chemokine fractalkine, is also associated with ileal disease.18 This is consistent with the essential role of CX3CR1-expressing dendritic cells in the luminal sampling of bacteria in the terminal ileum which we demonstrated in a previous study.42 These observations suggest that genetic variants, which are associated with deficits in the recognition of the intestinal bacteria or bacterial products, are associated with a predominant ileal disease location, triggered by the 103- to 104-fold increase of microbial density from the terminal ileum to the colon.36 The important role of CRP in bacterial recognition has been known for a long time. CRP was initially defined as a protein in serum that binds to the C polysaccharide of Streptococcus pneumoniae,1, 43 although exact details on the mechanisms of CRP in bacterial recognition and the host response have been published only recently.44–47 For example, a very recent study demonstrates that CRP increases the cytokine responses after bacterial infection through interactions with Fcγ receptors.45 However, functional studies are needed to analyse if the +1059 G/C polymorphism leads to an impaired recognition of bacteria by CRP which would explain a similar CD phenotype as observed for the 1007fs CARD15 variant28 and the T280M CX3CR1 genotype.18
The influence of the CRP +1059G/C polymorphism on CD has not been studied so far. Here, we demonstrate that the presence of a C allele, particularly in male patients and CD patients with a CDAI < 150, is associated with lower serum CRP levels. The reason for this gender dependent effect is unknown, although a likely explanation is the modulation of serum CRP levels by oestradiol in women.48 Moreover, there was a higher proportion of female patients with a CDAI between 150 and 300 (43.2%) compared with male patients (25.4%, P = 0.044) which may have influenced the results in female patients. Interestingly, a similar gender dependent effect was observed in another study analysing healthy blood donors.15 However, the +1059G/C CRP polymorphism does not result in an amino acid exchange49 and the detailed mechanisms by which this polymorphism regulates CRP serum levels have not been identified yet.
One major problem for analysing the data of this study are potential type 2 errors related to the very low genotype frequency of C/C homozygous carriers in the normal population. The largest population data on the +1059 G/C polymorphism (rs 1800947) are collected in the SNP data base of the National Center of Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/SNP) and contains genotype data from 1016 individuals. In this database, the genotype frequencies for the +1059 G/C polymorphism were 0.1%, 8.4% and 91.5% for homo-, heterozygous and wildtype carriers, respectively. For example, assuming a similar genotype frequency in our control population, a sample size of more than 25 000 patients would be needed to provide a power of 80% to detect a twofold increase in the genotype frequency of C/C homozygotes in CD patients with a two-tailed alpha = 0.05 (type 1 error probability of 5%). Even with the higher C/C genotype frequency among CD patients detected in our study (2.1%), a total sample size of 1042 individuals would be required to have 80% power for detecting a significant difference between CD patients and controls, assuming that the C/C genotype frequency among controls is 0.1%. Therefore, the influence of C/C homozygosity on serum CRP levels and CD phenotype can be only analysed in very large trials which is beyond the scope of a single clinical center. However, comparing the ‘historic’ controls from the NCBI SNP data base (n = 1016 with one C/C homozygous carrier) with the CD population of our study (n = 241 with 5 C/C homozygotes), there was a significant difference in the C/C genotype frequency between the two populations (P = 0.0013).
In another recent study, no association was observed between three other CRP polymorphisms (−717G/A, 1444C/T, CRP 4A/G) and CRP serum concentrations in active CD before infliximab treatment or clinical and biological response to infliximab.20 These observations are in apparent contrast to results obtained in coronary heart disease or in SLE, where the CRP 1444T allele and the CRP 4A allele were associated with higher and lower CRP concentrations, respectively.13, 14 However, in a recent study of a cohort of patients with coronary heart disease and a population of controls, this association could not be confirmed.50 In this context, it is of interest that a response to infliximab was associated with elevated CRP serum concentrations before treatment in CD patients.51 Similar observations were made in a recent study analysing the efficacy of certolizumab pegol [a polyethylene-glycolated Fab fragment of anti-tumour necrosis factor (TNF), also named CDP870]52 and the anti-TNF agent CDP571.53 Therefore, the knowledge of CRP genotypes influencing CRP levels would be of great value for selecting patients for anti-TNF therapy. Although 88 patients received infliximab in our study, only six of these patients were hetero- or homozygous for the +1059G/C polymorphism. Because of the small sample size, the influence of the +1059G/C polymorphism on the response to infliximab could not been determined in this study.
One of the main reasons for the discrepancies regarding the effect of certain CRP polymorphisms on CRP serum levels in different studies13, 14, 20 may be that those CRP variants lead to slight variations in low CRP levels in healthy individuals or diseases which do not result in a significant elevation of CRP.50 Furthermore, an association with higher levels of CRP in diseases such as CD, which is associated with a pronounced elevation of this protein, is difficult to analyse, because the heterogeneity in the degree of inflammatory reaction is a major confounding factor. It has been hypothesized that the paradoxical absence of a CRP increase in some patients with clinically active CD may be because of a genetic predisposition to produce less CRP.20 Indeed, family studies have suggested that 30–40% of the variation in plasma CRP concentration is genetically determined.7 CRP elevation in inflammatory bowel disease (IBD) patients is associated with clinical disease activity, endoscopic inflammation, severely active histologic inflammation (in CD patients), and several other biomarkers of inflammation.54 To adjust for disease activity, we compared the mean serum CRP levels in three different patient groups according to their disease activity based on the CDAI. Our results show that with increasing disease activity, as defined by CDAI, serum CRP levels increased. Furthermore, the data demonstrate that hetero- and homozygous carriers of the CRP +1059G/C polymorphism have lower mean serum CRP levels than wildtype carriers. Interestingly, a recent study demonstrated that patients with CD and a persistently low CRP in the face of active disease were characterized by an almost exclusive ileal disease distribution.21 This study included data sets for 223 CD patients, with a mean disease duration of 12 years.21 Of these, 22 patients fulfilled the criteria for low CRP. The low CRP group showed significant differences for disease site (P < 0.01) and for BMI (P = 0.006) compared with the raised CRP group.21 Specifically, CD patients with low CRP had a predominance of pure ileal disease (95% vs. 53%).21 This observed association may be due to different +1059 G/C genotypes given the results of our study, in which carriers of the +1059 C allele were also characterized by a predominant ileal disease location and lower mean serum CRP values than patients with wildtype G/G genotype.
In summary, heterozygosity for the CRP +1059G/C polymorphism is associated with disease involvement of the terminal ileum in CD independent of CARD15 variant alleles. Homozygous C/C carriers were found only in the CD group and were characterized by lower CRP serum levels which were adjusted for CD activity. Given the overall low frequency of the homozygous C/C genotype, larger trials are necessary to evaluate if homozygosity for the C allele is associated with CD and if the presence of this polymorphism defines the efficacy of anti-TNF therapies.
We are indebted to Mrs P. Lohse, H. Rübsamen, G. Simon and Z. Sisic for their expert technical assistance.
This study was supported by grants of the Deutsche Forschungsgemeinschaft/DFG (BR 1912/3–1 and BR 1912/5–1), the Else-Kröner-Fresenius-Stiftung (P50/05//EKMS 05/62) and by a grant of the University of Munich (FöFoLe Nr. 264). This work contains parts of the unpublished degree theses of D. Thalmaier and J. Dambacher at the Ludwig-Maximilians-University Munich.
- 3Studies on the binding specificity of human C-reactive protein for phosphorylcholine. Fed Proc 1978; 37: 1495., , .
- 15Epistatic effect of C-reactive protein (CRP) single nucleotide polymorphism (SNP) +1059 and interleukin-1B SNP +3954 on CRP concentration in healthy male blood donors. Int J Immunogenet 2005; 32: 229–32., , .
- 18Increased expression of the chemokine fractalkine in Crohn's disease and association of the fractalkine receptor T280M polymorphism with a fibrostenosing disease phenotype. Am J Gastroenterol 2006; 101: 99–106., , , et al.Direct Link:
- 19Macrophage migration inhibitory factor (MIF) -173 G/C polymorphism influences upper gastrointestinal tract involvement in Crohn's disease (abstract). Gastroenterology 2006; 130 (Suppl. 2): A359., , , et al.
- 28Homozygosity for the CARD15 frameshift mutation 1007fs predicts early onset of Crohn's disease with ileal stenosis, entero-enteral fistulas, and frequent need for surgical intervention with high risk of re-stenosis. Scand J Gastroenterol 2006; in press., , , et al.
- 31Homozygosity for the 3020insC mutation of the CARD15/NOD2 gene is a strong predictor for a very severe subtype of Crohn's disease characterized by long-segment bulky fibrostenoses of the terminal ileum (abstract). Gut 2004; 53(Suppl. VI): A142., , , et al.