Anti-MBL autoantibodies in patients with rheumatoid arthritis: prevalence and clinical significance

Journal of Autoimmunity 27 (2006) 125e133 www.elsevier.com/locate/issn/08968411

Anti-MBL autoantibodies in patients with rheumatoid arthritis: prevalence and clinical significance Bhawna Gupta a, Sunil Kumar Raghav a, Charu Agrawal a, Ved Prakash Chaturvedi b, Rakha Hari Das a, Hasi Rani Das a,* a

Institute of Genomics and Integrative Biology, Delhi University Campus, Mall Road, Delhi 110 007, India b Department of Rheumatology, Army Hospital, N. Delhi 110 010, India Received 23 March 2006; revised 30 June 2006; accepted 1 July 2006

Abstract Occurrence of autoantibodies in patients’ sera is the characteristic feature of autoimmune disorders. We assessed the presence of antimannose binding lectin (MBL) autoantibodies in the sera of 107 rheumatoid arthritis (RA) patients and 121 control subjects by enzyme immunoassay. Elevated levels of anti-MBL autoantibodies in the sera of RA patients (P < 0.0001) was detected for the first time. The ratios of anti-MBL positive in RA patients and controls were respectively 60.7% and 1.65%. Experiments were then designed to understand the functional relevance of these autoantibodies. An inverse correlation of anti-MBL autoantibodies with serum MBL levels (P ¼ 0.001) and MBL complex activity (P ¼ 0.02) was observed without genetic association between MBL polymorphisms and anti-MBL autoantibody secretion. A significant increase (P ¼ 0.038) in the level of anti-MBL autoantibodies was observed in 23 synovial fluid samples in comparison to the serum samples. Moreover, the anti-MBL autoantibodies were found to be more often present in the sera of RA patients (60.75% sensitivity, 98.35% specificity and 0.913 area under the ROC curve) in comparison to the IgM and IgG isotypes of rheumatoid factors (RF). Anti-MBL autoantibodies were still positive in 25.23% RA patients when both the RF isotypes were negative. Also, in RA patients, at all stages of disease activity and joint deformity, anti-MBL autoantibodies were more often present than both the RF isotypes. Therefore, the significant presence of anti-MBL autoantibodies enunciates that anti-MBL autoantibodies might have a diagnostic value; however, more studies are needed to confirm the role of anti-MBL autoantibodies in the diagnosis of rheumatoid arthritis. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Autoantibodies; Mannose binding lectin; Rheumatoid arthritis; Rheumatoid factor

1. Introduction Rheumatoid arthritis (RA) is one of the most common human systemic autoimmune diseases affecting approximately 1% of the world’s population. Although the precise etiology of RA is unknown, genetic and environmental factors seem to be involved in its pathogenesis [1]. We observed an association of tumor necrosis factor (TNF)-a microsatellites and the single nucleotide polymorphisms of mannose binding

* Corresponding author. Tel.: þ91 11 2766 2581; fax: þ91 11 2766 7471. E-mail addresses: [email protected], [email protected] (H.R. Das). 0896-8411/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.jaut.2006.07.002

lectin (MBL2) with susceptibility and progression of RA in Indian population [2,3]. The frequency of the B variant (codon 54) of the MBL2 gene was observed to be lower in RA patients in comparison to the controls (P ¼ 6.35  106). However, no statistically significant difference of serum MBL was observed between the RA patients and the controls [3]. RA is known to be associated with the presence of a number of autoantibodies [4,5]. Rheumatoid factor (RF) is a well known autoantibody directed against the Fc part of human IgG0 (agalactosyl IgG with an exposed N-acetyl glucosamine), which is more prevalent in RA [6]. This agalactosylated IgG with the exposed GlcNAc is also shown to be an easy target to the serum acute phase protein, mannose-binding lectin (MBL) [7].

126

B. Gupta et al. / Journal of Autoimmunity 27 (2006) 125e133

Serum MBL is enhanced by inflammatory stimuli and binds carbohydrates on the surfaces of pathogenic microorganisms and particulate materials via its C-terminal domain in a calcium dependent manner and activates lectin complement pathway in an antibody-independent manner [7,8]. MBL consists of trimeric subunits with collagenous domain that assemble to high order structures (hexamers) resembling the complement component C1q. On binding to the specific carbohydrate, its associated serine proteases (MBL associated serine proteases or MASPs) get activated, leading to activation of lectin complement pathway [6,7]. Consequently MBL has a significant role in eliciting the inflammatory response and thus has been well associated with the pathogenesis of various infectious and autoimmune diseases [9e11]. Particularly in rheumatoid arthritis (RA), an exposed GlcNAc on IgG0 being an easy target for binding with MBL leads to generation of inflammatory response. The presence of MBL and IgG0 in the synovial fluid of these patients supports such association [12]. MBL deficiency has been found in the general population and mostly correlated with the mutations in MBL gene. Its gene maps to chromosome 10 and mutations in the MBL gene and promoter polymorphisms determine the protein levels [3]. The variant alleles and altered serum MBL level are frequently associated with RA [13e15]. Recently in systemic lupus erythematosus (SLE), a related autoimmune disorder, autoantibodies against C1q and MBL have been reported and interestingly, the anti-MBL autoantibody level in this disorder has been shown to decrease the functional activity of MBL [16,17]. These reports prompted us to investigate the presence of MBL autoantibodies in the sera and synovial fluid of RA patients. The significant presence of the anti-MBL autoantibodies in RA directed us to assess their functional importance. The clinical significance of these autoantibodies was further measured against the conventionally known isotypes of rheumatoid factor (IgM RF and IgG RF) in RA.

obtained from each patient and normal individual. The human ethics committee approved the project. Medical records of the patients stating the disease duration, duration of morning stiffness, acute phase response as erythrocytes sedimentation rate (ESR) and C-reactive protein (CRP), presence of extra-articular manifestations and presence of bone deformities were collected (Table 1). Disease activity of the patients with RA was assessed at their first visit, using the DAS-28 disease score calculator, according to number of tender joints involved, swollen joints, ESR (mm/1st h), visual analogue score (VAS) for general health as subjectively estimated by the patients and clinician’s assessment of physical function [19]. The RA patients (n ¼ 107) were later categorized into severe (n ¼ 75) and less severe (n ¼ 32) groups on the basis of DAS-28 disease score calculator as well as other medical records stating the disease duration, duration of morning stiffness, presence of extra-articular manifestations, the presence of bone deformities, and the titer of CRP (as in Table 1 of this paper). The examiner was blinded to the anti-MBL results at the time of diagnosis. Blood samples from controls and patients were obtained at first clinical presentation and sera separated were stored at 20  C until assayed for anti-MBL autoantibody. Synovial fluid (SF) samples were obtained from actively inflamed joints of 23 patients with RA. Disease duration in these RA patients was 5  3.2 years, as measured from the first clinical signs of arthritis, irrespective of which joint was initially affected. Clinical inflammation was defined as both joint swelling and pain at the time of physical examination. The laboratory assessment of these RA patients included disease duration, duration of morning stiffness, ESR, CRP, presence of extra-articular manifestations, presence of bone deformities and rheumatoid factor (RF). The samples were ejected out by the rheumatologists under aseptic conditions and were immediately stored at -70  C until assayed. No prior freezee thawing was allowed.

2. Patients and methods

Table 1 Characteristics of 107 patients with rheumatoid arthritis (RA)

2.1. Patients and controls

Clinical and paraclinical variables

RA patients Mean (SD)

Severe patients Mean (SD)

Less severe patients Mean (SD)

Females (%) Age (years) Disease duration (years) DAS Tender joint count Swollen joint count ESR (mm/1st h) CRP (mg/L) Morning stiffness (min) RA patients with nodules (%) Anti-MBL positive (%) IgM RF positive (%) IgG RF positive (%)

60 45 (6) 3 (2.8) 5.1 (1.61) 10 (6) 8 (3) 25.8 (14.2) 12 (7) 80.6 (69.5) 10 60.74 51.4 28.97

50 43.5 (3.6) 3.4 (1.5) 5.1 14.6 (3.8) 10.3 (2.1) 30.2 (10.5) 15.5 (4.2) 98.7 (73.8) 100 65.34 56 36

50 41 (4.6) 2 (1.5) <5.1 6.2 (2.4) 5.6 (4.2) 26.4 (7.6) 12 (7.3) 28.7 (23.2) 0 50 40.62 12.5

Patients visiting the Rheumatology outpatient center of the Department of Rheumatology, Army Hospital, Research and Referral, N. Delhi, India were considered for the present study. All patients fulfilled the American College of Rheumatology classification criteria for the disease [18] that finally served as the gold standard for the diagnosis of rheumatoid arthritis. Blood samples from 107 RA patients were collected. To analyze the sensitivity and specificity of the tests, our control cohort of 121 individuals included healthy controls and 13 individuals with degenerative and other inflammatory joint diseases including psoriatic arthritis (n ¼ 3), osteoarthritis (n ¼ 7) and spondylarthropathy (n ¼ 3). All controls and the RA patients included in this study were age, sex and ethnicity matched. Informed consent was

DAS, disease activity score; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; RF, rheumatoid factor.

B. Gupta et al. / Journal of Autoimmunity 27 (2006) 125e133

2.2. Detection of anti-MBL autoantibodies in the sera and synovial fluid samples The detection of IgG binding to MBL was done by the modified method of Seelen et al. [16]. Nunc plates were coated for 2 h at 37  C with 100 ml/well of serum purified MBL (0.5 mg/ ml, US Biologicals M2250) in a 0.1 M carbonate-buffer. Purity of MBL was checked by SDSePAGE (12%). Anti MBL antibody (US Biologicals M2250-03) was used as a positive control and bovine serum albumin (BSA, Sigma), a nonsense antigen was included as a negative control in each assay performed. The wells incubated with Tris-buffered saline containing 0.05% Tween-20 (TBST) instead of serum were used as blanks. All samples were analyzed in triplicate. After blocking the unoccupied blocking sites by 1% BSA, 100 ml/well of serum samples diluted to 1:100 in TBST containing 0.3% BSA and 10 mM EDTA were added to each well and incubated at 37  C for 2 h. EDTA was included to inhibit the Ca2þ-dependent binding of MBL to carbohydrates present on the Fc portion of IgG. One hundred microliters/ well peroxidase conjugated goat antihuman IgG F(ab 0 )2 (Sigma, A2290) diluted 1:5000 in TBSTeEDTA, was added to each well and incubated at 37  C for 1 h. After washing, 100 ml/well of TMB substrate (BD Biosciences, Cat. No. 555214) was added. The plates were incubated for 10 min at room temperature. The reaction was stopped using 2 N phosphoric acid and optical densities (OD) at 450 nm were measured with the ELISA plate reader (BioRad). The concentration of IgG reactive with MBL is expressed in arbitrary units/ml of serum (AU/ml) considering the autoantibody level of a patient with higher OD value as 1000 AU/ml. The standard curve was generated with each assay performed using the serial dilutions of the said sample. The synovial fluid samples were diluted 1:100 with the TBST containing 0.3% BSA and 10 mM EDTA and assayed for the binding of IgG to MBL as above. A few of the sera samples tested previously were also included in each assay performed to verify the accuracy of the detection process. For the assessment of intra-assay variation, one sample was tested in 12 wells on one occasion while the inter-assays reproducibility was ascertained using three samples containing low, intermediate, and high anti-MBL antibody concentrations, six times on different days. The coefficient of variation (CV), standard deviation (SD), and mean values were calculated (CV ¼ SD/mean  100%). 2.3. Inhibition assay for anti-MBL autoantibodies by purified MBL Anti-MBL positive sera (1:100) were preincubated with different concentrations of purified MBL and binding of IgG was detected for each sample as described above. 2.4. Measurement of serum MBL Serum concentration of MBL in RA patients and control individuals was measured by ELISA as described earlier [3].

127

2.5. MBL complex activity MBL complex activity was measured using the method described by Peterson et al. with a few modifications [20]. In brief, mannan coated plates were incubated with serum, diluted 1:100 in MBL binding buffer at 4  C for 16 h. C4bdepositing capacity was assessed by incubation with 1 mg/ml C4 (Sigma, C8195) diluted in TBST-Ca2þ at 37  C and deposited C4b was detected by the addition of rabbit anti-human C4 (Sigma, Cat. No. C 3402) followed by the incubation with HRP conjugated anti-rabbit IgG (Sigma) for 1.5 h. After addition of TMB substrate, OD at 450 nm was measured. 2.6. MBL2 genotyping As reported earlier [3], MBL2 gene in our cases/controls was typed for the three polymorphisms in exon 1 by homogenous mass extend (hME) assay and one-directional sequencing. The genotypes were compared for association with anti-MBL autoantibody secretion. 2.7. Measurement of the serum IgG and IgM rheumatoid factors IgM and IgG isotypes of rheumatoid factors in the sera of RA patients and the controls was measured by the modified method of Teitsson et al. [21]. Briefly, microtiter plates were coated with 100 ml/well of normal rabbit IgG (10 mg/ml, Fluka 56830) in carbonate buffer at 37  C for 2 h. The sera dilutions used were 1/10, 1/50, 1/100 and 1/1000 for both IgM and IgG RF but the final dilutions were kept 1/100 for their low reactivity at the higher dilutions. The bound IgM and IgG RF were detected with 1:10 000 dilution of peroxidase conjugate anti-human IgM F(ab 0 )2 (Sigma) with TBST and 1:5000 diluted peroxidase conjugate anti-human IgG F(ab 0 )2 (Sigma) at 37  C for 1 h respectively. The TMB substrate was added and the reaction was stopped by 2 N phosphoric acid. The OD values were determined at 450 nm. BSA, a non-sense antigen was included as a negative control in each assay performed and the wells incubated with TBST instead of serum were used as the blanks. The concentration of IgM and IgG RF is expressed in arbitrary units/ml of serum (AU/ml) considering the RF level of a patient with higher OD value as 1000 AU/ml. The standard curve was generated with each assay performed using the serial dilutions of the said sample. 2.8. Statistical analysis Means, standard deviations (SD), confidence intervals, Student t-test and Chi-square analysis were used wherever appropriate. McNemar test and Chi-square analysis were used to examine the significance of differences in distributions of categorical variables across different groups. ManneWhitney Utest was used to examine continuous variables across different groups. Correlations between the tests used were determined by Spearman’s rank correlation test. Two sided P values less than or equal to 0.05 were considered significant. For the assay

128

B. Gupta et al. / Journal of Autoimmunity 27 (2006) 125e133

for anti-MBL autoantibody and both the RF isotypes, we calculated the following indices: sensitivity and specificity under Bayesian model, positive and negative predictive values (PPV and NPV respectively), positive and negative likelihood ratio (PLR and NLR respectively) to examine the ratio of the probabilities of the test result in RA patients and the controls. In addition, a ROC (receiver operating characteristic) curve analysis was carried out to compare test characteristics independently of predefined cutoff points across different tests and calculated the area under the curve (AUC). All statistical analyses were done with SPSS 13.0 and other web-based statistical software. Moreover, in preparing this report, the guidelines proposed by Bruns et al. were followed [22]. 3. Results 3.1. Detection of autoantibodies to MBL in the sera samples The level of anti-MBL autoantibody in RA patients was found to be significantly higher (P  0.0001 by Manne Whitney U-test) than the controls (Fig. 1A). The high OD value of a patient was designated as 1000 AU/ml and the titer of anti-MBL autoantibodies in the RA patients and the controls was thus computed. The mean  standard deviation (SD) was 427.487  129.39 AU/ml in the controls and

747.722  192.42 AU/ml in the RA patients. A cutoff level of 2 SD above the average of the anti-MBL autoantibodies in the control population was calculated as 686.28 AU/ml (as indicated by the dotted line in Fig. 1A). The number of subjects having a titer of more than 2 SD above the average of the controls was 65 out of 107 RA patients (60.7%) as compared to 2 out of 121 (1.65%) controls (Fig. 1A). This difference was also found to be statistically significant (P ¼ 0.00001). A titration curve was drawn using serial dilutions of the serum sample (Fig. 1B). It was also observed that titer of anti-MBL autoantibody in the 13 non-RA controls (Psoriatic arthritis: mean anti-MBL  SD ¼ 476.81  26.688 AU/ml; Osteoarthritis: mean anti-MBL  SD ¼ 488.196  113.573 AU/ml; Spondylarthropathy: mean anti-MBL  SD ¼ 450.238  67.192 AU/ ml) was below the cutoff level (2 SD above average of controls). The intra-assay CV was 4.8e10%, and the inter-assay CV was 7e18%.

3.2. Inhibition assay Anti-MBL positive sera (diluted 1:100) preincubated with purified MBL showed an inhibition of binding of autoantibodies to solid phase MBL in a dose dependent manner (Fig. 1C).

Fig. 1. (A) Autoantibodies to mannose binding lectin (MBL) in serum samples of 121 controls and 107 RA patients. P value by ManneWhitney U-test. AU, arbitrary units. The dotted line indicates 2 SD above average in the control subjects. (B) Titration curve for anti-MBL autoantibodies using serial dilutions of the standard serum. (C) Inhibition assay for anti-MBL autoantibodies.

B. Gupta et al. / Journal of Autoimmunity 27 (2006) 125e133

129

3.3. Detection of the anti-MBL autoantibodies in the synovial fluid samples The titer of anti-MBL autoantibodies in the synovial fluid samples was calculated (Mean MBL  SD ¼ 866.60  236.61) considering 1000 AU/ml of the serum sample described above. The mean value of the anti-MBL autoantibodies in the synovial fluid samples was statistically higher than the mean value of the anti-MBL autoantibodies in the sera samples (P ¼ 0.038) of the RA patients.

3.4. Correlation of anti-MBL autoantibodies and serum MBL level A statistically strong inverse correlation was observed (Fig. 2) when anti-MBL autoantibodies were compared with the serum MBL levels of the RA patients (P ¼ 0.001, r ¼ 0.331). However, a statistically significant difference in the concentration of serum MBL in RA patients (mean MBL  SD ¼ 1.273  0.796 mg/ml) and controls (mean MBL  SD ¼ 1.287  0.807 mg/ml) was not found.

3.5. Association of MBL complex activity with the presence of anti-MBL autoantibodies Any correlation of the MBL complex activity of the RA patients with the anti-MBL autoantibodies was calculated using Spearman’s correlation coefficient. An inverse correlation (Fig. 3) between the two was identified (P ¼ 0.02, r ¼ 0.221).

3.6. Association of MBL2 genotypes with the presence of anti-MBL autoantibodies The genotypes of RA patients and the controls were analyzed for an association with the titer of anti-MBL autoantibody. We did not observe any statistically significant association among the cases/controls (Table 4).

Fig. 2. Association between titers of anti-MBL autoantibodies and concentration of MBL in RA patients. P value by Spearman’s rank correlation test. Trend line depicts a negative linear regression. AU, arbitrary units.

Fig. 3. Correlation between titers of anti-MBL autoantibodies and MBL complex activity in RA patients. Trend line depicts a negative linear regression. AU, arbitrary units.

3.7. Comparison between the presence of anti-MBL autoantibodies, IgM and IgG rheumatoid factors in rheumatoid arthritis IgM RF was evaluated, in both the RA patients (Mean  SD ¼ 1359.23  682.62 AU/ml) and the controls (Mean  SD ¼ 645.02  294.89 AU/ml), and the difference was found to be statistically significant (P < 0.001 by ManneWhitney U-test). The level of IgG RF in the RA patients (Mean  SD ¼ 1055.06  292.75 AU/ml) was significantly increased (P < 0.001 by ManneWhitney U-test) as compared to the controls (Mean  SD ¼ 840.85  146.07 AU/ml). IgM and IgG RF isotypes were measured by ELISA and a cutoff level of 2 SD above average of the controls was calculated. The number of cases having the titer of IgM RF more than 2 SD above average of the controls (1234.814 AU/ml) was 55 out of 107 RA patients and 5 out of 121 controls. However, the number of cases with the amount of IgG RF more than 2 SD above average of the controls (1133 AU/ml) was 31 out of 107 RA patients and 4 out of 121 control individuals. For both the RF isotypes this difference was statistically significant (P  0.001). We compared the occurrence and clinical significance of all the three antibodies in the sera of RA patients. For comparisons between the anti-MBL autoantibodies and the RF isotypes, the cutoff level of 2 SD above average of the control population was maintained. With respect to the cutoff level, the sensitivity, specificity, predictive values and likelihood ratios for each assay were calculated (Table 2). The assay for anti-MBL autoantibodies was found to give highest sensitivity

130

B. Gupta et al. / Journal of Autoimmunity 27 (2006) 125e133

Table 2 Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR) and negative likelihood ratio (NLR) of serological markers in 107 RA patients Serological marker

Sensitivity (%)

Specificity (%)

PPV (%)

NPV (%)

PLR

NLR

Anti-MBL autoantibody IgM RF IgG RF Anti-MBL þ IgM RF Anti-MBL þ IgG RF Anti-MBL þ IgM RF þ IgG RF

60.75 51.4 28.97 28.94 22.43 15.89

98.35 95.87 96.69 100 98.35 100

97.01 91.67 88.57 100 92.31 100

73.91 69.05 60.62 61.42 58.91 57.35

36.75 12.44 8.76 f 13.57 f

0.399 0.507 0.734 0.710 0.789 0.841

(60.75%) and specificity (98.35%) in comparison to both of the RF isotypes. Examination of the likelihood ratios for various test results confirmed the difference between the tests. We undertook a ROC (receiver operating characteristic) analysis (Fig. 4) and calculated the area under the curve (AUC) with standard error (SE) and 95% confidence interval (CI). The ROC analysis displays the pairs of sensitivity and specificity for different cutoff points of anti-MBL autoantibodies, IgM RF and IgG RF concentrations. The AUC was best for anti-MBL autoantibodies (AUC ¼ 0.913, SE ¼ 0.019, 95% CI ¼ 0.876e0.949) while a decrease was observed for IgM RF (AUC ¼ 0.822, SE ¼ 0.029, 95% CI ¼ 0.766e0.879) and IgG RF (AUC ¼ 0.761, SE ¼ 0.032, 95% CI ¼ 0.698e0.824). We also analyzed the benefits of combined use of all the three antibody assays but found a decrease in the sensitivity in all the cases. However, the combined assays were more specific for the detection of rheumatoid arthritis (Table 2). Among the RA patients investigated, only 15.89% (17 out of 107) were positive for all the three antibodies, however, 43 (40.18%) patients with clinically defined RA showed an absence of RF isotypes (IgM RF and IgG RF). Among these 43 RF negative RA patients 62.79% (27) were positive for antiMBL autoantibodies (Table 3).

If only the IgM RF was used as a single RF test to detect rheumatoid arthritis (most laboratories and Rheumatology departments in hospitals only measure IgM RF and not the other RF isotypes), as many as 52 (49.59%) patients with RA remained undetected. However, in these 52 IgM RF negative RA patients, anti-MBL autoantibodies were positive in 65.38% (34) patients. The data therefore demonstrates that the anti-MBL autoantibodies are also frequent in RF-negative patients. 3.8. Association of anti-MBL autoantibodies, IgM and IgG RF with clinical and paraclinical parameters of rheumatoid arthritis Considering the cutoff level of 2 SD above average of the control subjects, the number of RA patients positive for antiMBL autoantibodies at all stages of disease activity (calculated according to DAS score), was more when compared to both the RF isotypes (in patients with DAS < 5.1, 19/38 are anti-MBL positive compared to 14/38 for RF IgM, in patients with DAS > 5.1, 44/69 are anti-MBL positive vs. 41/69 for RF) as shown in Table 3. However, the difference was not statistically significant. The anti-MBL autoantibody reactivity was also observed to be increased with the disease progression of severity (calculated according to DAS and other medical records) but without statistical significance. Eleven of the 27 (40.74%) RA patients categorized as less severe were tested positive for anti-MBL autoantibodies but tested negative for both RF isotypes. 4. Discussion

Fig. 4. Receiver operating characteristic (ROC) curve analysis of anti-MBL autoantibodies (Anti-MBL), IgM rheumatoid factor (RFM) and IgG rheumatoid factor (RFG).

In recent years, a growing body of evidence has shown the importance of innate immunity involving lectin pathway of complement activation wherein MBL plays a crucial role [7e9]. In RA and other related disorders like SLE, low levels of MBL have been reportedly associated [11,13,14]. In the present study, for the first time we have detected significant presence of anti-MBL autoantibodies in the sera of RA patients as compared to the controls. The specific detection of anti-MBL autoantibodies was confirmed by the addition of EDTA in the enzyme immunoassay inhibiting the Ca2þ-dependent carbohydrate binding of MBL; use of anti-human IgG F(ab 0 )2 fragment confirmed that the binding was only on Fab region of the anti-MBL autoantibodies. Moreover, preincubation of sera with different concentrations of

B. Gupta et al. / Journal of Autoimmunity 27 (2006) 125e133

131

Table 3 Data distribution depending on Anti-MBL, IgG RF and IgM RF positivity in 107 RA patients Clinical and paraclinical variables

Anti-MBL Positive

Negative

All RA patients (n ¼ 107) Less severe RA (n ¼ 32) Severe RA (n ¼ 75) DAS: <3.49 (n ¼ 16) DAS: 3.49e5.1 (n ¼ 22) DAS: 5.1e6.71 (n ¼ 50) DAS: > 6.71 (n ¼ 19) Tender joints (10) Swollen joints (8) IgM RF negative (n ¼ 52) IgG RF negative (n ¼ 76) IgM and IgG RF negative (n ¼ 43)

65 16 49 8 11 32 12 44 34 34 41 27

42 16 26 8 11 18 7 26 20 18 35 16

P value

IgM RF

P value

Positive

Negative

0.0078 0.197 0.0001 0.27 0.23 0.003 0.071 0.0013 0.0041

55 13 42 6 8 32 9 36 31

52 19 33 10 14 18 10 34 23

IgG RF

0.1 0.065 0.044 0.106 e 0.003 0.241 0.126 0.047

Positive

Negative

31 4 27 2 3 18 8 23 20

76 28 48 14 19 32 11 47 34

AU, arbitrary units; RF, rheumatoid factor. P value for anti-MBL and IgM RF (that were more frequently present) by Fisher exact test.

MBL inhibited the binding of anti-MBL autoantibody to the MBL coated plates in a dose-dependent manner (Fig. 1C) similar to that observed by Takahashi et al., [17]. These findings strongly suggest that IgG binding to MBL represents a true antigen-antibody reaction. The anti-MBL autoantibodies might not be exclusive to rheumatoid arthritis as their presence was also observed in lupus patients [16,17]. We also observed the presence of IgM type of anti-MBL autoantibodies (data not shown) in the patients’ sera as reported by Mok et al. in SLE [23]. However, in the present study we report the presence and clinical significance of IgG type anti-MBL autoantibodies due to their higher levels. Different authors have provided genetic mutations either in exonic regions affecting the functional activity of MBL or in the promoter regions as a possible explanation of MBL deficiency [13,15], but there is considerable controversy about it [24e26]. We therefore examined any association between levels of anti-MBL autoantibodies and the MBL genotypes. We did not find any significant association between the structural allelic variants (exon 1) of MBL gene and the level of anti-MBL autoantibody (Table 4), indicating that genetic polymorphisms possibly have no role in determination of autoantibody generation/level in our population. As an acute phase protein, expression of MBL is expected to rise in any inflammatory condition including RA [27]. However, in our previous study [3] we reported a similar level of serum MBL in both the cases and the controls. In other population studies, the serum MBL level is reported

to be low in RA as compared to the controls [11,14]. Recently it was reported that the decreased MBL levels may result from gene polymorphisms, consumption during disease, or binding of autoantibodies [28]. A similar or low serum MBL level may therefore be hypothesized as an affect of consumption/clearance due to autoantibody production. In our present study, the significant rise in autoantibody against MBL in RA and an inverse correlation between anti-MBL autoantibodies and the serum MBL level further supports such clearance. However, further works are required to establish such a hypothesis. We continued to examine the effect of anti-MBL autoantibodies on MBL complex activity. In the present study, the role of classical pathway in the activation of C4 was negated by using MBL binding buffer as described by Petersen et al. [20]. A significant negative correlation was observed between the levels of anti-MBL autoantibodies and the MBL complex activity of circulating MBL. These findings again confirm those of Seelen et al. [16]. The MBL complex activity represents the ability of the MBL-MASP complexes to activate C4. The association between anti-MBL autoantibodies and decreased MBL function suggests that these autoantibodies interfere with the MBL-MASP complex in either (i) inhibiting MBL binding to mannan, (ii) interfering with the MASP2 binding with MBL, or (iii) may themselves bind to MBLMASP complex and prevent MASP2 activation. The significant presence of anti-MBL autoantibodies in the sera of RA patients further enunciated their possible

Table 4 Association of MBL genotypes of exon 1 and titers of anti-MBL autoantibodies in the sera of controls and RA patients MBL genotype

AA AB AC AD

No. of Controls (n ¼ 121) 57 43 8 13

Anti-MBL autoantibodies (AU/ml) Mean

SD

424.29 424.16 420.77 450.02

137.73 119.68 128.13 144.45

P value

0.996NS 0.942NS 0.559NS

No. of RA patients (n ¼ 107) 71 13 8 15

Anti-MBL autoantibodies (AU/ml) Mean

SD

750.88 765.77 741.96 720.16

185.39 227.89 216.13 198.52

P value

0.827NS 0.914NS 0.583NS

AA is the wild genotype, AB is the heterozygous codon 54 mutation, AC is the heterozygous codon 57 mutation and AD is the heterozygous codon 52 mutation. AU, arbitrary units. P value by Student’s t-test. NS, not significant.

132

B. Gupta et al. / Journal of Autoimmunity 27 (2006) 125e133

presence in the synovial fluid samples of the RA patients. A higher level of the anti-MBL autoantibodies in the synovial fluid in comparison to the serum of the RA patients can be explained either as an over-production of these antibodies or their preferential accumulation in the synovial fluid. The data suggests a possible role of these antibodies in the synovial fluid yet the exact mechanism of which is still to be explored. After a careful analysis of the results we verified the clinical significance of anti-MBL autoantibodies in comparison to both the isotypes of rheumatoid factors. The presence of high levels of anti-MBL autoantibodies in the samples with negative IgM/IgG/both IgM and IgG rheumatoid factors further confirms that there are no false positives due to binding of rheumatoid factors on the solid support. The highest sensitivity (60.75%), specificity (98.35%) and AUC (0.93) for the anti-MBL autoantibodies suggest the clinical significance of these antibodies. When both the RF isotypes (IgM and IgG RF) were negative, antiMBL autoantibodies were still positive in 25.23% of RA patients. Categorizing the RA patients according to disease severity, disease activity and joint deformity, we found that the antiMBL autoantibodies were more often present in the RA patients in comparison to both the RF isotypes. Although the exact mechanism in the persistence of the autoantibodies against MBL is still not clear but genetic effect can be easily ruled out. Therefore a possibility of some posttranslational modifications of the MBL protein arises that may be responsible for the production of these antibodies. However, owing to an increased exposure of immune system to MBL, a subsequent consumption of elevated serum MBL by the increased autoantibodies might be a cause of their generation. In conclusion, our data suggest an important role of antiMBL autoantibodies in rheumatoid arthritis. Further studies such as the correlation between anti-MBL autoantibody and radiological destruction, comparison of anti-MBL prevalence with Anti-CCP antibodies, and determination of the exact status of MBL autoantibody in the pathogenesis of rheumatoid arthritis are necessary.

[3]

[4]

[5]

[6]

[7]

[8] [9]

[10] [11] [12]

[13]

[14]

[15]

[16]

[17]

Acknowledgements The authors wish to thank Mr. A.N. Shirkey for his help in obtaining the serum samples from the RA patients. The study is supported by the Council of Scientific and Industrial Research, India.

[18]

[19]

References [1] Tokuhiro S, Yamada R, Chang X, Suzuki A, Kochi Y, Sawada T, et al. An intronic SNP in a RUNX1 binding site of SLC22A4, encoding an organic cation transporter, is associated with rheumatoid arthritis. Nat Genet 2003;35:341e8. [2] Agrawal C, Raghav SK, Gupta B, Das RH, Chaturvedi VP, Goswami K, et al. Tumor necrosis factor-alpha microsatellite polymorphism

[20]

[21]

association with rheumatoid arthritis in Indian patients. Arch Med Res 2005;36:555e9. Gupta B, Agrawal C, Raghav SK, Das SK, Das RH, Chaturvedi VP, et al. Association of mannose binding lectin gene (MBL2) polymorphisms with rheumatoid arthritis in an Indian cohort of case-control samples. J Hum Genet 2005;50:583e91. Nell VP, Machold KP, Stamm TA, Eberl G, Heinzl H, Uffmann M, et al. Autoantibody profiling as early diagnostic and prognostic tool for rheumatoid arthritis. Ann Rheum Dis 2005;64:1731e6. Vencovsky J, Machacek S, Sedova L, Kafkova J, Gatterova J, Pesakova V, et al. Autoantibodies can be prognostic markers of an erosive disease in early rheumatoid arthritis. Ann Rheum Dis 2003;62: 427e30. Bond A, Alavi A, Axford JS, Youinou P, Hay FC. The relationship between exposed galactose and N-acetylglucosamine residues on IgG in rheumatoid arthritis (RA), juvenile chronic arthritis (JCA) and Sjogren’s syndrome (SS). Clin Exp Immunol 1996;105:99e103. Ohta M, Okada M, Yamashina I, Kawasaki T. The mechanism of carbohydrate-mediated complement activation by the serum mannan-binding protein. J Biol Chem 1990;265:1980e4. Presanis JS, Kojima M, Sim RB. Biochemistry and genetics of mannanbinding lectin (MBL2). Biochem Soci Trans 2003;31:748e52. Eisen DP, Minchinton RM. Impact of mannose-binding lectin on susceptibility to infectious diseases. Clin Infect Dis 2003;37: 1496e505. Kilpatrick DC. Mannan-binding lectin: clinical significance and applications. Biochim Biophys Acta 2002;1572:401e13. Turner MW, Hamvas RM. Mannose-binding lectin: structure, function, genetics and disease association. Rev Immunugenet 2000;2:305e22. Malhotra R, Wormald MR, Rudd PM, Fischer PB, Dwek RA, Sim RB. Glycosylation changes of IgG associated with rheumatoid arthritis can activate complement via the mannose-binding protein. Nat Med 1995;1: 237e43. Lipscombe RJ, Sumiya M, Hill AV, Lau YL, Levinsky RJ, Summerfield JA, et al. High frequencies in African and non-African populations of independent mutations in the mannose binding protein gene. Hum Mol Genet 1992;1:709e15. Tsutsumi A, Sasaki K, Wakamiya N, Ichikawa K, Atsumi T, Ohtani K, et al. Mannose-binding lectin gene: polymorphisms in Japanese patients with systemic lupus erythematosus, rheumatoid arthritis and Sjogren’s syndrome. Genes Immun 2001;2:99e104. Madsen HO, Garred P, Thiel S, Kurtzhals JA, Lamm LU, Ryder LP, et al. Interplay between promoter and structural gene variants control basal serum level of mannan-binding protein. J Immunol 1995;155: 3013e20. Seelen MA, Trouw LA, van der Hoorn JW, Fallaux-van den Houten FC, Huizinga TW, Daha MR, et al. Autoantibodies against mannose-binding lectin in systemic lupus erythematosus. Clin Exp Immunol 2003;134: 335e43. Takahashi R, Tsutsumi A, Ohtani K, Goto D, Matsumoto I, Ito S, et al. Anti-mannose binding lectin antibodies in sera of Japanese patients with systemic lupus erythematosus. Clin Exp Immunol 2004;136: 585e90. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 2004;31: 315e24. Prevoo ML, van ’t Hof MA, Kuper HH, van Leeuwen MA, van de Putte LB, van Riel PL. Modified disease activity scores that include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of patients with rheumatoid arthritis. Arthritis Rheum 1995;38:44e8. Petersen SV, Thiel S, Jensen L, Steffensen R, Jensenius JCJ. An assay for the mannan-binding lectin pathway of complement activation. Immunol Methods 2001;257:107e16. Teitsson I, Withrington RH, Seifert MH, Valdimarsson H. Prospective study of early rheumatoid arthritis. I. Prognostic value of IgA rheumatoid factor. Ann Rheum Dis 1984;43:673e8.

B. Gupta et al. / Journal of Autoimmunity 27 (2006) 125e133 [22] Bruns DE, Huth EJ, Magid E, Young DS. Toward a checklist for reporting of studies of diagnostic accuracy of medical tests. Clin Chem 2000;46:893e5. [23] Mok MY, Jack DL, Lau CS, Fong DY, Turner MW, Isenberg DA, et al. Antibodies to mannose binding lectin in patients with systemic lupus erythematosus. Lupus 2004;13:522e8. [24] Stanworth SJ, Donn RP, Hassall A, Dawes P, Ollier W, Snowden N. Absence of an association between mannose-binding lectin polymorphism and rheumatoid arthritis. Br J Rheumatol 1998;37:186e8. [25] Horiuchi T, Tsukamoto H, Morita C, Sawabe T, Harashima S, Nakashima H, et al. Mannose binding lectin (MBL) gene mutation is not a risk factor for systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) in Japanese. Genes Immun 2000;1:464e6.

133

[26] Larsen F, Madsen HO, Sim RB, Koch C, Garred P. Disease-associated mutations in human mannose-binding lectin compromise oligomerization and activity of the final protein. J Biol Chem 2004;279:21302e11. [27] Thiel S, Holmskov U, Hviid L, Laursen SB, Jensenius JC. The concentration of the C-type lectin, mannan-binding protein, in human plasma increases during an acute phase response. Clin Exp Immunol 1992;90: 31e5. [28] Kravitz MS, Pitashny M, Shoenfeld Y. Protective molecules C-Reactive Protein (CRP), Serum Amyloid P (SAP), Pentraxin3 (PTX3), MannoseBinding Lectin (MBL), and Apolipoprotein a1 (Apo a1), and their autoantibodies: prevalence and clinical significance in autoimmunity. J Clin Immunol 2005;25:582e91.