Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

Journal of Crohn's and Colitis (2013) 7, 134–141

Low mannose-binding lectin (MBL) is associated withpaediatric inflammatory bowel diseases and ilealinvolvement in patients with Crohn diseaseMarta Kovacs a, Maria Papp b, Peter Laszlo Lakatos c, Silvia Jacobsen d,Eva Nemes e, Marianne Polgar f, Eniko Solyom g, Piroska Bodi h,Agnes Horvath i, Kriszta Molnar j, Doloresz Szabo j, Aron Cseh j,Katalin Eszter Muller j, Antal Dezsofi j, Andras Arato j, Gabor Veres j,⁎

a Department of Paediatrics, Petz Aladár County and Teaching Hospital, Győr, Hungaryb 2nd Department of Medicine, University of Debrecen, Debrecen, Hungaryc 1st Department of Medicine, Semmelweis University, Budapest, Hungaryd Euroimmun Medizinische Labordiagnostika AG, Luebeck, Germanye Department of Paediatrics, University of Debrecen, Debrecen, Hungaryf Madarász Children's Hospital, Budapest, Hungaryg Paediatric Health Centre, Borsod-A-Z County and University Teaching Hospital, Miskolc, Hungaryh Department of Paediatrics, Pándy Kálmán Hospital, Gyula, Hungaryi Department of Paediatrics, Cholnoky Ferenc County Hospital, Veszprém, Hungaryj 1st Department of Paediatrics, Semmelweis University, Budapest, Hungary

Received 30 October 2011; received in revised form 11 March 2012; accepted 12 March 2012

⁎ Corresponding author at: 1st DeparE-mail address: vergab@gyer1.sote

1873-9946/$ - see front matter © 2012doi:10.1016/j.crohns.2012.03.008

KEYWORDSMannose-binding lectin(MBL);MBL deficiency;Inflammatory boweldisease;Crohn disease;Ulcerative colitis;Paediatric

Abstract

Background: Mannose-binding lectin (MBL) is a pattern-recognition molecule of the innateimmune system and may be involved in the pathogenesis of inflammatory bowel disease(IBD). Our aim was to assess the prevalence of MBL deficiency in a cohort of patients withpaediatric-onset IBD and study whether it is associated with the clinical manifestations, serumantibody formation, or genetic factors.Methods: This prospective study included 159 paediatric patients (mean age: 14.0 years) with IBD[107 patients with Crohn disease (CD) and 52 patients with ulcerative colitis (UC)]. Furthermore, 95controls were investigated. Serum sampleswere determined for MBL by enzyme-linked immunosor-

bent assay (ELISA) and for serologic markers [autoantibodies against Saccharomyces cerevisiae(ASCA) and perinuclear components of neutrophils (pANCA)] by indirect immunofluorescent

tment of Paediatrics, Semmelweis University, 53 Bókay Street, 1083 Budapest, Hungary..hu (G. Veres).

European Crohn's and Colitis Organisation. Published by Elsevier B.V. All rights reserved.

135Low MBL is associated with paediatric IBD and ileal involvement in CD patients

assay. NOD2/CARD15 variants were tested by polymerase chain reaction/restriction fragmentlength polymorphism.Results: The MBL serum concentration was significantly lower in IBD patients(both with CD andUC) compared to controls (IBD, p=0.007, CD, p=0.04, UC p=0.004). Prevalence of low MBL level(b500 ng/mL) was significantly higher in both CD and UC groups compared to controls (p=0.002and p=0.006). Furthermore, low MBL level was associated with isolated ileal involvement(p=0.01) and MBL deficiency (b100 ng/mL) with male gender (p=0.004) in patients with CD.We failed to confirm any correlation between MBL deficiency and serum autoantibodies orNOD2/CARD15 variants.Conclusions: Our results suggest that low MBL associated with paediatric-onset IBD and ileal CDmay be considered an additional marker of the IBD pathogenesis.© 2012 European Crohn's and Colitis Organisation. Published by Elsevier B.V. All rights reserved.

1. Introduction

The inflammatory bowel diseases (IBD), including Crohn dis-ease (CD) and ulcerative colitis (UC), are chronic relapsingand remitting disorders of the gastrointestinal tract. Thepathogenesis of IBD is complex and multifactorial. Currentevidence suggests that IBD results from aberrant immuneresponse and loss of tolerance to the normal intestinal flora,leading to chronic inflammation of the gut in a geneticallysusceptible host.1,2 This idea is supported by the occurrenceof antibodies directed to microbial antigens and by the identi-fication of NOD2/CARD15 as a gene conferring susceptibility toCD.3,4 Both CD and UC are characterized by intestinal inflam-mation caused by a disturbance in the balance between cyto-kines and increased complement activation.5

Mannose-binding lectin (MBL) is an important componentof innate immunity. MBL is a pattern-recognition moleculethat activates the lectin pathway of the complement systemirrespective of an antibody. MBL can directly opsonize mi-croorganisms and enhance their uptake by phagocyticcells.6 MBL is primarily synthesized in the liver and predom-inantly circulates as a serum protein. In addition, Seyfarth etal. showed intestinal MBL2 gene expression using a commer-cial cDNA library.7 In contrast, a more recent study could notfind meaningful MBL2 gene expression in intestinal tissues.8

Furthermore, MBL binds to apoptotic and necrotic cells andfacilitates uptake by macrophages.9,10 Recently, a newrole for MBL as a Toll-like receptor (TLR), co-receptor indirecting intracellular signalling, has been identified.11

Three different point mutations of exon1 of the MBL2gene (in codons 52, 54 and 57) have the greatest influenceon serum levels and functional activity of MBL, but promoterregion polymorphism is also important.12,13 The inter-individual levels of MBL vary from approximately 5 ng/mLto more than 10000 ng/mL, but the level of MBL in eachindividual is genetically determined, quite stable throughoutthe life. At birth, the level is about 2/3 of the adult level,which is reached in a month, and there is a minor declineat elderly age.11,14

Low MBL level was observed in up to 40% and the preva-lence of MBL deficiency is 8–10% in the normal popula-tion.15,16 MBL deficiency has been associated withincreased susceptibility and severity of infections, especiallyin children and in immune-compromised patients.17–19

Furthermore, the MBL pathway may be involved in the

development of various autoimmune diseases (e.g., IBD andcoeliac disease).20,21

The role of MBL in IBD is still controversial. Rector et al.reported a significant decrease in the frequency of the MBLvariants in sporadic adult UC patients.22 In contrast, Sivaramet al. observed a significantly higher frequency of mutationsat codon 54 in UC patients than in healthy controls, whichwas accompanied by low MBL concentrations.23 In otherstudies, the frequency of MBL polymorphism, MBL deficien-cy, and the median level of MBL were not significantly differ-ent from controls — neither in adult CD nor in UC.24–26 Onlyone Polish survey was conducted in paediatric IBD, in whichthe frequency of MBL2 gene variants responsible for MBLdeficiency was significantly higher in CD patients than incontrols or in children with UC. However, only a smallnumber of patients were involved in this study (CD:30,UC:26).27 Up until now, the prevalence of MBL deficiencyhas not been investigated in paediatric IBD.

The aim of our study was to determine the serum MBLlevels and the prevalence of MBL deficiency in a cohort ofpatients with paediatric-onset IBD and to evaluate possibleassociations with the clinical presentation, response totreatment, and extraintestinal manifestations in IBD. Ad-ditionally, we studied the relationship between MBL defi-ciency and serum antibody formation, and polymorphism ofNOD2/CARD15.

2. Patients and methods

2.1. Patients

One hundred and seven consecutive patients with paediatric-onset CD [male/female (m/f) ratio: 64/43, mean age:14.1 years (range: 5.3–20 years)], 52 patients with paediatric-onset UC [m/f ratio: 22/30, mean age: 14.0 years (range:6–19.7 years)], and 95 age-and sex-matched controls wereincluded in this study.

Diagnosis of IBD was established according to clinical, ra-diographic, endoscopic, and histological criteria.28,29 Dis-ease activity was evaluated according to the PaediatricCrohn's Disease Activity Index (PCDAI) in children with CDand according to the Paediatric Ulcerative Colitis Index(PUCAI) for the patients with UC.30,31 Activity index N30 isdefined as moderate-severe disease, the index between 11

136 M. Kovacs et al.

and 30 indicates mild disease, and the index ≤10 refers toinactive disease.

Age, age at onset, presence of extraintestinal manifesta-tion (EIM); arthritis, ocular manifestations, skin lesions, andhepatic manifestations, frequency of flare-ups (frequentflare up: N1/year), therapeutic effectiveness (e.g.: needfor steroid and/or immunosuppressive therapy, steroid resis-tance, as defined in the European Crohn's and Colitis Organi-sation Consensus (ECCO) Report,32) or short-term responseto infliximab therapy, need for surgery (resection), the pres-ence of familial IBD were collected by the clinical investiga-tor reviewing the medical charts and completing aquestionnaire. In CD, an additional parameter, perianal in-volvement, was also investigated. Localization and pheno-type of disease were based on the Montreal classificationcriteria. Only patients with a confirmed diagnosis for morethan 1 year were enrolled. In UC, disease extent was definedbased on the maximum extent during the follow-up.

Prospectively, blood samples were obtained for measure-ment of mannan-binding lectin (MBL), for serologic markers[anti-Saccharomyces cerevisiae antibodies (ASCA), perinuc-lear anti-neutrophil cytoplasmic antibodies (pANCA)], com-plete blood count and C-reactive protein (CRP). Whenblood samples were taken, patients underwent a clinical as-sessment, including calculation of clinical disease activityscores. Sera for MBL and serological marker determinationwere coded to maintain blinding. After serum separation,blood samples were stored at −80 °C until further analysis.

The study protocol was approved by the Ethical andScience Committee of the Semmelweis University. Each par-ent of the subject, who had been informed about the natureof the study before, signed the informed consent form.

2.2. MBL assay

Serum concentration of MBL was detected by the use of thedouble-antibody sandwich enzyme-linked immunosorbentassay (ELISA) system adopted from Minchinton et al.33

using monoclonal mouse antihuman MBL antibody (BioPortoDiagnostics A/S, Gentofte, Denmark). A vial of standardMBL solution (BioPorto Diagnostics A/S) was supposed tohave an MBL content of 1000 AU and we accepted that it cor-responded to 3200 ng/mL oligomerized MBL, as declared bythe manufacturer. The detection limit was 4.86 ng/mL.According to the recommendations of the manufacturer,low MBL concentration was defined as a serum levelb500 ng/mL and MBL deficiency as b100 ng/mL. The MBLassay was performed at the Clinical Research Center of theDebrecen University in a blinded manner, without havingknown the patients' diagnosis or other clinical informationbefore.

2.3. Antibody assays for ASCA and pANCA antibodies

Presence of ASCA and pANCA antibodies was determined in acommercially available indirect immunofluorescent (IIF)assay (EUROIMMUN AG, Luebeck, Germany) according tothe manufacturers' instructions. IIF coated cover glasseswith several biological substrates were cut into millimetre-sized fragments (BIOCHIPs, EUROIMMUN AG) and used sideby side in the same reaction field, giving the opportunity

to investigate one serum on several tissues or preparedcells: ASCA: Saccharomyces cerevisiae fungal smear, ANCA:ethanol (EOH)-fixed and formalin (HCHO)-fixed human gran-ulocytes. Sera were incubated at a 1:10 dilution (ASCAIgG:1:1000, ASCA IgA: 1:100) in phosphate-buffered saline(PBS)/Tween for 30 min. After being washed with the buffer,slides were incubated with fluorescein-labelled goat antihu-man IgG or IgA antibodies for another 30 min. After anotherwashing process, evaluation and classification of the pat-terns were performed with the help of a fluorescence micro-scope (EUROIMMUN LED, EUROStar Bluelight, EUROIMMUNAG, Luebeck, Germany) under ultraviolet light (UV).

2.4. Detection of NOD2/CARD15 mutations

Genomic DNA was isolated from whole blood using the QIAampDNA Blood Mini Kit (QIAGEN, Germany). The three NOD2/CARD15 variants, Arg702Trp, Gly908Arg and Leu1007fs, weretyped using polymerase chain reaction/restriction fragmentlength polymorphism as previously described.34 NOD2/CARD15variants were detected by denaturing high-performance liquidchromatography (dHPLC, Wave DNA Fragment Analysis System,Transgenomic, UK). Sequence variation, observed in the dHPLCprofile, was sequenced on both strands to confirm thealteration. Sequencing reactions were performed with theABI BigDye Terminator Cycle Sequencing Kit v1.1 (AppliedBiosystems, Foster City, CA) and samples were sequenced onan ABI Prism 310 Genetic Analyzer (Applied Biosystems).Genotyping was carried out at the National Haematology andImmunology Institute, Budapest, Hungary.

2.5. Statistical methods

Statistical analysis was carried out using Graph Pad Prism 5(GraphPad, San Diego CA, USA). T-test with separate vari-ance estimates and Fisher's exact test were used to evaluatedifferences between IBD and control patients as well aswithin subgroups of IBD patients. Results are expressed asodds ratio (OR) with 95% confidence intervals (95% CI).Spearman's rank order correlation was calculated to testthe association between MBL levels, CRP, and clinical activ-ity indexes. A pb0.05 was considered as significant.

3. Results

3.1. MBL levels and MBL deficiency in patients withIBD and their association with other serologicalmarkers

Low MBL levels (defined as b500 ng/mL) were found in 34 of107 (31.8%) patients with CD, in 18 of 52 patients (34.6%)with UC, and in 13 of 95 (13.7%) controls. MBL deficiency(defined as b100 ng/mL) was found in 10 (9.4%) patientswith CD, 6 (11.5%) patients with UC and 9 controls (9.5%)(Table 1). We noted significantly lower median MBL levelsin both CD and UC patients compared to controls (CD,p=0.04, UC, p=0.004, IBD, p=0.007). Furthermore, preva-lence of low MBL level (b500 ng/mL) was significantly higherin both CD and UC groups than in controls (CD, p=0.002, UC,p=0.006, IBD, p=0.001). Nevertheless, prevalence of MBL

Table 1 Mannose-binding lectin (MBL) levels in patients with inflammatory bowel disease (IBD) and controls.

Median MBL levelsng/mL (IQR)

MBL deficiency(b100 ng/mL), (n,%)

Low level of MBL(b500 ng/mL), (n,%)

CD (n=107) 1591 (416.9–2958) 10 (9.4%) 34 (31.8%)UC (n=52) 1133 (383.2–2447) 6 (11.5%) 18 (34.6%)Controls (n=95) 2304 (931–3186) 9 (9.5%) 13 (13.7%)

IQR = inter-quartile range.

137Low MBL is associated with paediatric IBD and ileal involvement in CD patients

deficiency (b100 ng/mL) was similar in IBD patients and incontrols. No significant difference was found in the preva-lence of MBL deficiency and low MBL level as well as in me-dian MBL levels between UC and CD patients. Moreover, wedid not find any significant association between low MBLlevel or MBL deficiency and ASCA/pANCA in CD or in UC(Figs. 1, 2).

3.2. Association between MBL levels, clinicalphenotype, CRP, and actual disease activity inpatients with IBD

The clinical phenotype of patients with CD and UC is shownin Tables 2 and 3 according to different MBL levels. LowMBL level (b500 ng/mL) was associated with isolated ileal in-volvement (p=0.01) in patients with CD (Fig. 3). In addition,MBL deficiency (b100 ng/mL) was related to male gender(p=0.004) in CD. Nevertheless, the MBL level or deficiencywas not associated with medical therapy, need for surgery,or extraintestinal manifestations — neither in CD nor in UC.Additionally, MBL level was not associated with CRP andactual PCDAI in CD or with CRP and PUCAI in UC. Moreover,CRP values correlated positively with PCDAI in CD (n=86,R: 0.617, pb0.0001). However, CRP values were not signifi-cantly associated with PUCAI in UC.

MB

L (

ng

/mL

)

ASCA posit

ive C

D pat

ients

ASCA neg

ative

CD p

atien

ts0

2000

4000

6000

8000

10000

Figure 1 Correlation between mannose-binding lectin (MBL)levels and ASCA antibodies, based on 89 serum samples in patientswith Crohn disease [ASCA positive CD patients n=66, ASCA nega-tive CD patients n=23, Mann–Whitney test, p=0.19].

3.3. Association between MBL levels andNOD2/CARD15 genotype in CD

The prevalence of NOD2/CARD15 genotypes was available in 44patients. MBL deficiency was not associated with NOD2 vari-ants. The prevalence of NOD2/CARD15mutationswere not sta-tistically different between patients with high (N500 ng/mL,20.6%) and low (b500 ng/mL, 33.3%) MBL levels.

4. Discussion

This is the largest study to assess MBL levels and MBL defi-ciency as well as relevant associations with phenotype inpaediatric patients with IBD. The MBL serum concentrationwas significantly lower in IBD (both in CD and UC) comparedto controls. Prevalence of low MBL level (b500 ng/mL) wassignificantly higher in both CD and UC groups compared tocontrols. Furthermore, low MBL level was associated withisolated ileal involvement and MBL deficiency (b100 ng/mL)with male gender in patients with CD.

The role of MBL in IBD is still controversial. A strong corre-lation exists between MBL2 genotypes and serum MBL levels,however the current knowledge of the genetics of MBL2 is in-adequate to predict serum MBL concentration.33,35

MB

L (

ng

/mL

)

pANCA posit

ive U

C pat

ients

pANCA neg

ative

UC p

atien

ts0

2000

4000

6000

8000

10000

Figure 2 Correlation between mannose-binding lectin (MBL)levels and pANCA antibodies, based on 47 serum samples inpatients with ulcerative colitis [pANCA positive UC patientsn=34, pANCA negative UC patients n=13, Mann–Whitney test,p=0.41].

Table 2 Clinical phenotype, serological status, and NOD2/CARD15 genotype in patients with Crohn's disease (CD, n=107)according to different mannose binding lectin (MBL) categories.

MBL deficiency(b100 ng/mL)n=10

Low level of MBL(b500 ng/mL)n=34

MBL competence(N500 ng/mL)n=73

Totaln=107

Male/female 9/1 (90/10%) 24/10 (70/30%) 40/33 (54/46%) 64/43 (60/40%)Age (y) 13.5 (range:8.7–19.2) 13.8 (range:5.8–20) 14.2 (range:5.3–19.6) 14.1 (range:5.3–20)Age at presentation (y) 12.1 (range:5.5–16.2) 12.4 (range:5.5-17.6) 12.8 (range:2–18) 12.7 (range:2–18)Duration (y) 1.4 (range:0–3.2) 1.3 (range:0–7) 1.4 (range:0–9) 1.4 (range:0–9)LocationIleum (L1) 3/10 (30%) 11/34 (32.4%) 8/73 (11%) 19/107 (16.8%)Colon (L2) 4/10 (40%) 10/34 (29.4%) 19/73 (26%) 29/107 (24.8%)Ileocolon (L3) 3/10 (30%) 13/34 (38.2%) 46/73 (63%) 59/107 (58.4%)Upper GI (+L4) 2/10 (20%) 5/34 (14.7%) 17/73 (23.3%) 22/107 (20.5%)BehaviourNon-stricturing,non-penetrating (B1)

6/10 (60%) 22/34 (64.7%) 41/73 (56.1%) 63/107 (58.9%)

Stricturing (B2) 1/10 (10%) 5/34 (14.7%) 16/73 (22%) 21/107 (20%)Penetrating (B3) 3/10 (30%) 7/34 (20.6%) 22/73 (30.1%) 29/107 (28.7%)Perianal disease 2/10 (20%) 5/34 (14.7%) 17/73 (23.3%) 22/107 (21%)Frequent relapse 3/10 (30%) 14/34 (41.2%) 27/73 (37%) 41/107 (38.3%)ASCA (IgA or IgG) positive 5/9 (55.5%) 18/27 (66.6%) 48/62 (77.4%) 66/89 (74.1%)NOD2/CARD15 carrier 2/3 (66.6%) 5/15 (33.3%) 6/29 (20.6%) 11/44 (25%)

138 M. Kovacs et al.

Rector et al. reported a highly significant decrease in thefrequency of the MBL variants (mutations in codons 52, 54and 57) in sporadic adult UC patients (36/124, 29%) com-pared to CD (93/216, 43.1%) and controls (126/308, 40.9%),suggesting a protective effect of functional MBL mutationsin UC.22 In contrast, in UC patients, Sivaram et al. observeda significantly higher frequency of mutations at codon 54than in healthy controls, accompanied by low MBL concen-trations.23 In other studies, the frequency of MBL deficiencyand the median level of MBL in adult CD patients25,26,36 andin UC patients24,26,36 were comparable. Only one Polishsurvey has been conducted in paediatric IBD with a smallnumber of patients, in which the frequency of MBL2 genevariants responsible for MBL deficiency was significantlyhigher in CD patients than in controls or in children withUC.27 In accordance with the results of this report, the

Table 3 Clinical phenotype and serological status of patients wibinding lectin (MBL) categories.

MBL deficiency(b100 ng/mL)n=6

Low level(b500 ng/mn=18

Male/female 3/3 (50/50%) 9/9 (50/50Age (y) 13.6 (range:10–17.5) 14.4 (rangAge at presentation (y) 10.3 (range:2–15.7) 11.6 (rangDuration (y) 3.4 (range:0–8) 2.7 (rangeLocationProctitis (E1) – –Left sided (E2) 3/6 (50%) 4/18 (22.2Extensive (E3) 3/6 (50%) 14/18 (77.Frequent relapse 2/6 (33.3%) 8/18 (44.4pANCA (IgA or IgG) positive 5/6 (83.3%) 14/16 (87.

median MBL concentration in our present study was signifi-cantly lower in CD than in controls.

Furthermore, our results were similar to the above-mentioned adult data concerning the prevalence of MBL defi-ciency in IBD patients and in controls.25,26 However, in contrastto these studies conducted in adults, we found that the preva-lence of low MBL levels was more common and median MBLlevels were significantly lower in paediatric patients with IBDthan in the control paediatric group. We cannot explain satis-factorily the reason for the differences between paediatricand adult MBL levels in IBD. Different genetic and/or immuno-logic factors could be responsible for this finding.37 Neverthe-less, further simultaneous studies conducted in paediatric andadult IBD patients are necessary to answer this question.

In the present study, we found no association of MBL de-ficiency with ASCA and pANCA antibodies — neither in CD

th ulcerative colitis (n=52) according to different mannose

of MBLL)

MBL competence(N500 ng/mL)n=34

Totaln=52

%) 13/21 (38/62%) 22/30 (42/58%)e:8.5–19) 13.7 (range:6–19.7) 14.0 (range:6–19.7)e:1.2–17.8) 11.6 (range:4.6–18) 11.6 (range:1.2–18):0–13.4) 2.1 (range:0–7) 2.3 (range:0–13.4)

5/34 (14.7%) 5/52 (9.6%)%) 11/34 (32.3%) 15/52 (28.8%)8%) 18/34 (52.9%) 32/52 (61.6%)%) 11/34 (32.3%) 19/52 (36.5%)5%) 20/31 (64.5%) 34/47 (72.3%)

Location of Crohn disease

Nu

mb

er o

f p

atie

nts

L1 non L10

20

40

60

80

Low MBL level(<500 ng/mL)

MBL competence(>500 ng/mL)

Figure 3 Association between lowmannose binding (MBL) levels(b500 ng/mL) and isolated terminal ileal (L1) involvement inpatients with Crohn disease (n=107, low MBL level, L1: isolatedterminal ileum involvement, n=11, non L1: colonic and ileocolonicinvolvement, n=23, MBL competence, L1: n=8, non L1: n=65)[Fisher's exact test, OR: 3.9, CI: 1.3–10.8, p=0.01].

139Low MBL is associated with paediatric IBD and ileal involvement in CD patients

nor in UC. Previous studies proposed an association betweengenetic predisposition (NOD2/CARD15) for CD and the pres-ence of ASCA antibodies.3,38 As a component of the innateimmune system, MBL may have a role in this process. MBLcan bind to Saccharomyces cerevisiae. MBL deficiency wasalso proposed to contribute to ASCA production after impair-ment of the normal processing of mannan-expressing micro-bial antigens, such as those which had been found on the cellsurface of many common microorganisms.39 The enhancedmannan exposure could stimulate specific immune responsesand the production of ASCA.39 However, the associationbetween genetically determined MBL deficiency and ASCApositivity was controversial in previous studies (positivecorrelation, ref. 39–41 no correlation, ref. 26,42).

In the present study, we found that low MBL levels wereassociated with ileal involvement in paediatric patientswith CD. In a study published by Seibold et al., all but 2 pa-tients with a low MBL phenotype and genotype had small in-testinal involvement.39 In addition, in previous studies, MBLdeficiency was associated with ASCA positivity,39–41 andwith complicated disease phenotype (B2 and B3).39 In con-trast, no correlation was found between MBL deficiencyand IBD phenotype25,26 in other studies conducted in adultpatients with IBD.

Up until now, association between MBL levels and clinicalphenotype has not been evaluated in paediatric IBD. Our ob-servation regarding an association between low MBL levelsand ileal involvement in CD could be explained by the failureof extrahepatic transcription of MBL2 gene in the small intes-tine.7,43 However, in amore recent study, no clinically signif-icant MBL2 gene expression was found in intestinal tissues.8

Although, Müller et al. showed that MBL deficiency resultedin enhanced ASCA production and excessive experimental co-litis in response to mannose-expressing mild gut pathogens inMBL deficient mice.8 It is conceivable that MBL2 polymor-phisms, resulting in lower MBL levels, may be associatedwith an increased number of bacteria in the small intestine.It is possible that MBL is responsible for maintaining the rela-tive sterility of the small bowel as a protective factor by en-hancing phagocytosis44 or by inhibiting the attachment ofmicroorganisms on intestinal cell surfaces.45 Furthermore,MBL deficiency impairs the normal removal and clearanceof apoptotic cells that may subsequently reveal previously

hidden self-antigen, causing loss of self-tolerance andspreading of autoimmunity.9,10,21,46

In concordance with the results of the study by Pappet al.,26 MBL level was not associated with CRP level and ac-tual disease activity index in CD or in UC in the presentstudy. CRP concentration may increase rapidly up to 1000-fold, compared to modest (2–3 fold) and slow (1–2 weeksafter the inducing event) increase of MBL levels during inflam-mation.47,48 The elevated CRP levels in patients with IBD indi-cate disease activity rather than bacterial complicationscaused by the enhanced production of pro-inflammatory cyto-kines in the mesenteric fat.26,49

In accordance with the results of Tilakaratne et al., CRPcorrelated positively with PCDAI in CD in our present study.Children with active disease had significantly higher CRPvalues compared to children with inactive disease.50 Fur-thermore, in concordance with our findings, Henriksenet al. described that patients with CD had a stronger CRPresponse than those with UC.49

In addition, the prevalence of a low MBL level was notrelated to the presence of major NOD2/CARD15 mutations26

in our study, which is in agreement with the results of thestudy of Papp et al.

In conclusion, MBL levels were lower in paediatric pa-tients with IBD than in controls. Nevertheless, MBL levels inpatients with CD and UC were similar. Furthermore, lowMBL level was associated with ileal involvement in paediatricpatients with CD. We found no association between MBL anddifferent serological markers or genetic polymorphism ofNOD2/CARD15.

Conflict of interest

There was no conflict of interest in this study.

Acknowledgements

We thank Petar Mamula (Chlidren's Hospital Philadelphia, USA) forthe language editing. Gabor Veres is holder of the János BolyaiResearch grant; this article was supported by the János BolyaiResearch Scholarship of the Hungarian Academy of Sciences.

References

1. Xavier RJ, Podolsky DK. Unravelling the pathogenesis of inflam-matory bowel diseases. Nature 2007;448:427–34.

2. Vind I, Riis L, Jespersgaard C, Jess T, Knudsen E, Pedersen N,et al. Genetic and environmental factors as predictors of dis-ease severity and extent at time of diagnosis in an inception co-hort of inflammatory bowel disease, Copenhagen County andCity 2003–2005. J Crohns Colitis 2008;2:162–9.

3. Papp M, Altorjay I, Dotan N, Palatka K, Foldi I, Tumpek J, et al.New serological markers for inflammatory bowel disease are as-sociated with earlier age at onset, complicated disease behav-iour, risk for surgery, and NOD2/CARD 15 genotype in aHungarian IBD cohort. Am J Gastroenterol 2008;103:665–81.

4. Rogler G. The effects of NOD2/CARD15 mutations on the func-tion of the intestinal barrier. J Crohns Colitis 2007;2:53–60.

5. Halstensen TS, Mollnes TE, Fausa O, Brandtzaeg P. Deposits ofterminal complement complex (TCC) in muscularis mucosaeand submucosal vessels in ulcerative colitis and Crohn's diseaseof the colon. Gut 1989;30:361–6.

140 M. Kovacs et al.

6. Brouwer N, Dolman KM, van Houdt M, Sta M, Roos D, KuijpersTW. Mannose-binding lectin (MBL) facilitates opsonophagocyto-sis of yeasts but not of bacteria despite MBL binding. J Immunol2008;180:4124–32.

7. Seyfarth J, Garred P, Madsen HO. Extra-hepatic transcription ofthe human mannose-binding lectin gene (mbl2) and theMBL-associated serine protease 1–3 genes. Mol Immunol2006;43:962–71.

8. Müller S, Schaffer T, Flogerzi B, Seibold-Schmid B, Schnider J,Takahashi K, et al. Mannan-binding lectin deficiency results inunusual antibody production and excessive experimental colitisin response to mannose-expressing mild gut pathogens. Gut2010;59:1493–500.

9. Nauta AJ, Raaschou-Jensen N, Roos A, Daha MR, Madsen HO,Borrias-Essers MC, et al. Mannose-binding lectin engagementwith late apoptotic and necrotic cells. Eur J Immunol2003;33:2853–63.

10. Ogden CA, de Cathelineau A, Hoffmann PR, Bratton D,Ghebrehiwet B, Fadok VA, et al. C1q and mannose binding lec-tin engagement of cell surface calreticulin an CD19 initiatesmacropinocytosis and uptake of apoptotic cells. J Exp Med2001;194:781–95.

11. Ip WK, Takahashi K, Moore KJ, Stuart LM, Ezekowitz RA.Mannose-binding lectin enhances Toll-like receptors 2 and 6 sig-naling from phagosome. J Exp Med 2008;205:169–81.

12. Garred P, Larsen F, Madsen HO, Koch C. Mannose-binding lectindeficiency-revisited. Mol Immunol 2003;40:73–84.

13. Garred P, Larsen F, Seyfarth J, Fujita R, Madsen HO. Mannose--binding lectin and its genetic variants. Genes Immun 2006;7:85–94.

14. Ytting H, Christensen IJ, Thiel S, Jensenius JC, Svendsen MN,Nielsen L, et al. Biological variation in circulating levels ofmannan-binding lectin (MBL) and MBL-associated serineprotease-2 and the influence of age, gender and physical exer-cise. Scand J Immunol 2007;66:458–64.

15. Wortley DL, Bardy PG, Mullighan CG. Mannose-binding lectin:biology and clinical implications. Intern Med J 2005;35:548–55.

16. Sorensen CM, Hansen TK, Steffensen R, Jensenius JC, Thiel S.Hormonal regulation of mannan-binding lectin synthesis in he-patocytes. Clin Exp Immunol 2006;145:173–82.

17. Koch A, Melbye M, Sorensen P, Homøe P, Madsen HO, Mølbak K,et al. Acute respiratory tract infections and mannose-bindinglectin insufficiency during early childhood. JAMA 2001;285:1316–21.

18. Peterslund NA, Koch C, Jensenius JC, Thiel S. Associationbetween deficiency of mannose-binding lectin and severe infec-tions after chemotherapy. Lancet 2001;358:637–8.

19. Kronborg G, Garred P. Mannose-binding lectin genotype as a riskfactor for invasive pneumococcal infection. Lancet 2002;360:1176.

20. Bowmann LH, Roep BO, Roos A. Mannose-binding lectin: clinicalimplications for infection, transplantation, and autoimmunity.Hum Immunol 2006;67:247–56.

21. Boniotto M, Braida L, Baldas V, Not T, Ventura A, Vatta S, et al.Evidence of a correlation between mannose binding lectin andceliac disease: a model for other autoimmune diseases. J MolMed 2005;83:308–15.

22. Rector A, Lemey P, Laffut W, Keyaerts E, Struyf F, Wollants E,et al. Mannan-binding lectin (MBL) gene polymorphisms in ulcer-ative colitis and Crohn's disease. Genes Immun 2001;2:323–8.

23. Sivaram G, Tiwari SK, Bardia A, Manoj G, Santhosh B, Saikant R,et al. Association of genetic variants of mannan-binding (MBL)lectin-2 gene, MBL levels and function in ulcerative colitis andCrohn's disease. Innate Immun Nov 18 2010 [Epub ahead of print].

24. Wang FY, Arisawa T, Tahara T, Nagasaka M, Fujita H, Hirata I,et al. The role of mannan-binding lectin (MBL) gene polymor-phisms in ulcerative colitis. J Clin Biochem Nutr 2008;42:54–8.

25. Nielsen RG, Vind I, Munholm P. Genetic polymorphisms of mannanbinding lectin (MBL), serum levels of MBL the MBL associated

serine protease and H-ficolin in patients with Crohn's disease.Gut 2007;56:446–54.

26. Papp M, Lakatos PL, Harsfalvi J, Farkas G, Palatka K, Udvardy M,et al. Mannose-binding lectin level and deficiency is not associ-ated with inflammatory bowel diseases, disease phenotype, se-rology profile, and NOD2/CARD15 genotype in a large Hungariancohort. Hum Immunol 2010;71:407–13.

27. Bak-Romaniszyn L, Szala A, Sokolowska A, Mierzwa G,Czerwionka-Szaflarska M, Swierzko AS, et al. Mannan-bindinglectin deficiency in pediatric patients with inflammatorybowel disease. Scand J Gastroenterol 2011;46:1275–8.

28. Van Assche GV, Dignass AI, Panes J, Beaugerie L, Karagiannis J,Allez M, et al. The second European evidence-based Consensuson the diagnosis and management of Crohn's disease: definitionsand diagnosis. J Crohns Colitis 2010;4:7–27.

29. Stange EF, Travis SPL, Vermeire S, Reinisch W, Geboes K,Barakauskiene A, et al. European evidence-based Consensuson the diagnosis and management of ulcerative colitis: defini-tions and diagnosis. J Crohns Colitis 2008;2:1–23.

30. Kundhal PS, Critch JN, Zachos M, Otley AR, Stephens D, GriffithsAM. Paediatric Crohn disease activity index: response toshort-term change. J Pediatr Gastroenterol Nutr 2003;36:83–9.

31. Turner D, Otley AR, Mack D, Hyams J, de Bruijne J, Uusoue K,et al. Development, validation, and evaluation of a paediatric ul-cerative colitis activity index: a prospective multicenter study.Gastroenterology 2007;133:423–32.

32. Stange EF, Travis SP, Vermeire S, Beglinger C, Kupcinkas L,Geboes K, et al. European evidence based consensus on the di-agnosis and management of Crohn's disease: definition and diag-nosis. Gut 2006;55:i1–i15.

33. Minchinton RM, Dean MM, Clark TR, Heatley S, Mullighan CG.Analysis of the relationship between mannose-binding lectin(MBL) genotype, MBL levels and function in an Australian blooddonor population. Scand J Immunol 2002;56:630–41.

34. Heliö T, Halme L, Lappalainen M, Fodstad H, Paavola-Sakki P,Turunen U, et al. CARD15/NOD2 gene variants are associatedwith family occurring and complicated forms of Crohn's disease.Gut 2003;52:558–62.

35. Swierzko AS, Szala A, Cedzynski M, Domzalska-Popadiuk I,Borkowska-Klos M, Jopek A, et al. Mannan-binding lectin geno-types and genotype-phenotype relationship in a large cohortof Polish neonates. Hum Immunol 2009;70:68–72.

36. Hoffmann C, Hoffmann P, Lun A, Büning C, Hiepe H, Scherer HU,et al. Is there a role for mannan-binding lectin in the diagnosis ofinflammatory bowel disease? Immunogenetics 2010;62:231–5.

37. Levine A. Pediatric inflammatory bowel disease: is it different?Dig Dis 2009;27:212–4.

38. Dassapoulos T, Frangakis C, Cruz-Correa M, Talor MV, Burek CL,Datta L, et al. Antibodies to Saccharomyces cerevisiae in Crohn'sdisease: higher titer are associated with a greater frequency ofmutant NOD2/CARD15 alleles and with a higher probability ofcomplicated disease. Inflamm Bowel Dis 2007;13:143–51.

39. Seibold F, Konrad A, Flogerzi B, Seibold-Schmid B, Arni S,Jüliger S, et al. Genetic variants of the mannan-binding lectinare associated with immune reactivity to mannans in Crohn'sdisease. Gastroenterology 2004;127:1076–84.

40. Seibold F, Boldt AB, Seibold-Schmid B, Schoepfer AM, FlogerziB, Müller S, et al. Association of deficiency formannan-binding lectin with anti-mannan antibodies in Crohn'sdisease: a family study. Inflamm Bowel Dis 2007;13:1077–82.

41. Schoepfer AM, Flogerzi B, Seibold-Schmid B, Schaffer T, Kun JF,Pittet V, et al. Low mannan-binding lectin serum levels are as-sociated with complicated Crohn's disease and reactivity to oli-gomannan (ASCA). Am J Gastroenterol 2009;104:2508–16.

42. Joossens S, Pierik M, Rector A, Vermeire S, Ranst MV, RutgeertsP, et al. Mannan binding lectin (MBL) gene polymorphisms arenot associated with anti-Saccharomyces cerevisiae (ASCA) in pa-tients with Crohn's disease. Gut 2006;55:746.

141Low MBL is associated with paediatric IBD and ileal involvement in CD patients

43. Uemura K, Saka M, Nakagawa T, Kawasaki N, Thiel S, JenseniusJC, et al. L-MBP is expressed in epithelial cells of mouse smallintestine. J Immunol 2002;169:6945–50.

44. Kuhlman M, Joiner K, Ezekowitz RA. 1989. The humanmannose-binding protein functions as an opsonin. J Exp Med1989;169:1733.

45. Zuo DM, Zhang LY, Lu X, Liu Y, Chen ZL. Protective role ofmouse MBL-C on intestinal mucosa during Shigella flexneri inva-sion. Int Immunol 2009;21:1125–34.

46. Nakamura N, Nonaka M, Ma BY, Matsumoto S, Kawasaki N, AsanoS, et al. Characterization of the interaction between serummannan-binding protein and nucleic acid ligands. J Leukoc Biol2009;86:737–48.

47. Sidoroff M, Karikoski R, Raivo T, Savilahti E, Kolho KL.High-sensitivity C-reactive protein in paediatric inflammatorybowel disease. World J Gastroenterol 2010;16:2901–6.

48. Thiel S, Frederiksen PD, Jensenius JC. Clinical manifestations ofmannan-binding lectin deficiency. Mol Immunol 2006;43:86–96.

49. Henriksen M, Jahnsen J, Lygren I, Stray N, Sauar J, Vatn MH,et al. C-reactive protein: a predictive factor and marker of in-flammation in inflammatory bowel disease. Results from a pro-spective population-based study. Gut 2008;57:1518–23.

50. Tilakaratne S, Lemberg DA, Leach S, Day AS. C-reactive proteinand disease activity in children with Crohn's disease. Dig Dis2010;55:131–6.