Article Text

Download PDFPDF

Concise report
The concentration of anticitrullinated protein antibodies in serum and synovial fluid in relation to total immunoglobulin concentrations
  1. Annemiek Willemze1,
  2. Jing Shi1,
  3. Marlies Mulder1,
  4. Gerrie Stoeken-Rijsbergen1,
  5. Jan W Drijfhout2,
  6. Tom W J Huizinga1,
  7. Leendert A Trouw1,
  8. Rene E M Toes1
  1. 1Department of Rheumatology, Leiden University Medical Center (LUMC), Leiden, The Netherlands
  2. 2Department of Immunohematology and Blood Transfusion, Leiden University Medical Center, Leiden, The Netherlands
  1. Correspondence to Annemiek Willemze, Department of Rheumatology, Leiden University Medical Center, P.O. Box 9600, Leiden 2300 RC, The Netherlands; a.willemze{at}


Background Anticitrullinated protein antibodies (ACPA) are one of the best predictors for the development of rheumatoid arthritis. Nonetheless, relatively little information is present on the absolute concentration of ACPA in relation to total immunoglobulin (Ig) concentrations. Such information would be of relevance to compare ACPA levels to other antibody levels. Here, we estimated the relative abundance of ACPA Ig in serum and synovial fluid using a quantitative approach.

Methods ACPA were purified using HiTrap Streptavidin columns coupled with biotinylated cyclic citrullinated peptide (CCP2). Total Ig and anti-CCP2 isotype reactivities were measured by ELISA.

Results ACPA were successfully isolated as substantial antibody amounts were eluted from sera of ACPA-positive patients and neglectable antibody amounts were eluted from sera of ACPA-negative patients. Up to 1 in 80 IgG-molecules were estimated to be ACPA. Strikingly, IgM-ACPA was most abundant in synovial fluid (with the highest enrichment in the range of one IgM-ACPA for every eight IgM-antibodies).

Conclusions ACPA-IgG levels are estimated to be within the range of peak levels of protective antibody responses against recall antigens. IgM-ACPA is abundantly present in synovial fluid, suggesting the presence of a continuous ongoing autoimmune response in the synovial compartment.

  • Rheumatoid Arthritis
  • Autoantibodies
  • Synovial fluid

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Anticitrullinated protein antibodies (ACPA) have been shown to initiate and enhance disease in murine models of arthritis1 ,2 and to activate Fc receptor (FcR)-positive cells3 ,4 and the complement system,5 arguing that they could play a role in disease pathogenesis. ACPA are highly specific for rheumatoid arthritis (RA) and can be detected years before clinical manifestations.6 ,7 Activation of naïve B cells upon the first antigen encounter results in proliferation and differentiation in IgM-secreting cells. During their differentiation upon further contact with T cells, some B cells undergo isotype switching and affinity maturation. The ACPA-response uses all isotypes.4 ,8–10 Intriguingly, an expanded ACPA-isotype profile associates with more severe radiographic damage, indicating that the extent of the ACPA-response impacts on disease progression.11

It has been shown that anticyclic citrullinated peptide (anti-CCP2) antibody levels are higher in synovial fluid than serum.12–14 However, limited information on absolute ACPA levels in either synovial fluid or serum is present as the levels are generally expressed as arbitrary units. Nonetheless, information on the absolute concentration of ACPA is of interest as it allows the comparison of the ACPA response to other antibody responses in quantitative terms. To quantitate ACPA levels it is required to isolate ACPA. Here we present data on the estimation of ACPA quantities in serum and synovial fluid.

Materials and methods

Patient population

Patients analysed were derived from the Leiden Early Arthritis clinic cohort.15 All patients fulfilled the American College of Rheumatology 1987 revised criteria for the classification of RA.16 Serum samples of 10 ACPA-positive and two ACPA-negative patients with RA and two healthy controls were obtained for ACPA purification. Knee synovial fluid was obtained at the time of therapeutic arthrocentesis. Samples were obtained from five ACPA-positive (including two paired synovial fluid-serum samples selected for high IgM-ACPA levels) and one ACPA-negative patient with RA and one patient with osteoarthritis, attending the outpatient clinic of the rheumatology department in Leiden. The protocols were approved by the relevant local ethics committee and all participants provided informed consent.

Affinity-purification of ACPA

Synovial fluid samples were treated with hyaluronidase type IV from bovine testes (Sigma Aldrich;100 μg/ml) and protease-inhibitor (Sigma Aldrich;1:50). ACPA from synovial fluid and serum were purified using HiTrap streptavidin HP1ml columns (GE-Healthcare) coupled with biotinylated CCP2 peptides as described previously.17 The CCP2 peptide was obtained from Dr Drijfhout, Department of IHB, LUMC, The Netherlands. Antibodies were eluted with 0.1 M Glycine-hydrogen chloride (HCL) pH2.5 and neutralised with 2 M Tris. A control column coated with CCP2 arginine was attached before the CCP2 citrulline column. This was necessary to control for non-specific adherence of antibodies to the bead material. We observed that, indeed, a considerable amount of IgG (mean:10 ug/ml) adhered to the control column (data not shown). After running the sample the columns were disconnected and eluted separately to guarantee the citrulline specificity of the purified antibodies.

Anti-CCP2 assays

Anti-CCP2-IgG was measured by ELISA (Immunoscan-RA Mark 2; Eurodiagnostica, Arnhem, The Netherlands). Samples with values >25 units/ml were considered positive according to the manufacturer's instructions. Anti-CCP-positive individuals were considered ACPA-positive.

Quantitative IgA, IgM, IgG ELISA

Quantitative Ig analyses were performed using the Human IgA-ELISA, IgG-ELISA and IgM-ELISA Quantitation Set (Bethyl Laboratories, USA) according to the manufacturer's instructions.


Affinity-purification of ACPA

ACPA were successfully isolated using HiTrap Streptavidin columns loaded with CCP2 peptides as substantial amounts of antibodies were eluted from sera of ACPA-positive patients with RA and neglectable amounts from sera of ACPA-negative patients or healthy controls (range:0–2.1 μg/ml) (table 1). Two affinity-columns, one coated with biotinylated CCP2 arginine and one coated with biotinylated CCP2 citrulline, were used to purify ACPA (figure 1A). The CCP2 arginine coated column was essential in the purification procedure, as considerable amounts of, most likely, non-specific antibodies were eluted from this control column (mean:10 μg IgG/ml). The flow-through was collected and antibodies were eluted with glycine-HCL and directly neutralised. The flow-through was completely devoid of anti-CCP2 reactivity, except for four high-titre samples (figure 1B), probably due to overloading of the column. Recovery of ACPA-activity was not complete. In some samples only 20% of ACPA activity could be detected. The recovery rate of ACPA of the different immunoglobulins (Ig) after purification was comparable. Together, these data indicate that ACPA can be purified from ACPA-positive sera and indicate that numbers of absolute ACPA levels are possibly an underestimation as not all ACPA activity originally present in the samples has been recovered.

Table 1

Relative abundance of ACPA- Ig in serum and synovial fluid in relation to total Ig levels

Figure 1

Affinity-purification of anticitrullinated protein antibodies (ACPA). ACPA was purified using two affinity columns. A control column coated with CCP2-arginine was attached to the column with CCP2-cittrulline to guarantee the citrulline specificity of the purified antibodies (A). In the elution step the two columns were disconnected and the antibodies were eluted off separately. Start material (serum or synovial fluid) (1), flow-through (2), elution CCP2-arginine (3) and elution CCP2-citrulline (4) were measured for ACPA presence and total immunoglobulin (Ig) content. ACPA were effectively purified using HiTrap Streptavidin columns (B). Hardly any ACPA remained in the flow-through (FL). However, not all ACPA activity could be recovered as in some patients only 20% of the ACPA activity present in the start material was present in the elution, indicating an underestimation of the amount of quantified ACPA. The elution of the arginine column contained barely any ACPA. The percentage in the graph represents the amount of ACPA present in relation to the start material. The different fractions were measured with an anti-CCP2 ELISA and the amount was measured in AU/ml.

Quantification of ACPA immunoglobulins present in serum and synovial fluid

After affinity-purification, ACPA-IgG levels of ACPA-positive patients were quantified by measuring the amount of IgG present. In sera of ACPA-positive patients, we measured up to 60 μg/ml IgG-ACPA (mean:28 μg/ml). As hardly any IgG could be isolated and measured from sera of ACPA-negative patients with RA (mean:1.5 μg IgG/ml), we conclude that considerable amounts of ACPA are present in ACPA-positive sera (table 1). These levels were estimated to be on average almost 1 in 183 IgG-antibodies. Within patients with high levels even up to 1 in 80 antibodies were ACPA (figure 2). IgA-ACPA is less abundant, only four patients displayed a level higher than 1 μg/ml in serum. The abundance of IgA-ACPA appeared, therefore, limited with a maximum of 0.36% of all IgA present in serum. IgM-ACPA was abundantly present in some subjects, in two patients around 10% of the serum IgM appeared to be ACPA.

Figure 2

Anticitrullinated protein antibodies (ACPA) immunoglobulins (Ig) in serum and synovial fluid of patients with rheumatoid arthritis (RA). ACPA-IgG in relation to total IgG (AU/mg total IgG) is not significantly higher in synovial fluid as compared with serum, only a trend was found (Wilcoxon signed-rank test p=0.09) (A). ACPA-IgA (AU/mg total IgA) and ACPA-IgM (AU/mg total IgM) are significantly enriched in synovial fluid as compared with serum (Wilcoxon signed-rank test: IgA, p=0.002, IgM, p=0.0003) (B–C). Percentage ACPA-IgG, IgA and IgM of total IgG, IgA and IgM in serum are depicted for 10 ACPA-positive patients with RA, two ACPA-negative patients with RA and two healthy controls (D). Percentage ACPA IgG, IgA and IgM of total IgG, IgA and IgM in synovial fluid are depicted for five ACPA-positive patients with RA, one ACPA-negative patient with RA and one patient with osteoarthritis (E).

Next we compared ACPA levels in paired serum and synovial fluid samples. A trend towards higher levels of ACPA-IgG in synovial fluid as compared with serum was observed, however this difference does not reach statistical significance (p=0.07). In contrast, a significant increase in IgM-ACPA was noted in synovial fluid as compared with serum (figure 2A). To quantify these ACPA levels we next purified ACPA from synovial fluid to investigate the local presence of the ACPA isotypes in the joint. Surprisingly, IgM-ACPA was most abundantly present in synovial fluid, where up to 13% of IgM-antibodies were estimated to be ACPA (figure 2). These two patients displayed high IgM ACPA levels in synovial fluid (∼5000 AU/ml). Up to 101 μg/ml IgM-ACPA was present in synovial fluid (range:2.8–101 μg/ml).

IgG ACPA was found in synovial fluid up to 103 μg/ml (table 1). Based on the total amount of IgG present, this indicates that in synovial fluid up to 2.3% of total IgG is ACPA.

In line with the small amount of IgA-ACPA in serum, little IgA-ACPA was also present in synovial fluid. We included synovial fluid of one ACPA-negative patient with RA and one patient with osteoarthritis as controls. In these samples, no IgA-ACPA and IgG-ACPA were detected. The purified sample of the ACPA-negative patient contained 0.5 μg/ml IgM.


In this study, we quantified the abundance of different ACPA-Ig in serum and synovial fluid of patients with RA. IgG-ACPA is present in relatively high concentrations in serum and synovial fluid as up to 1 out of 80 IgG antibodies can be ACPA. These findings extend a previous observation describing the estimation of IgG-ACPA amounts using three RA sera.18 Furthermore, we now report that IgM-ACPA can be abundantly present in synovial fluid (up to 13% of total IgM). Next to IgM-ACPA, also IgG-ACPA was found in relatively high concentrations in synovial fluid (up to 2.3% of total IgG). The estimations presented are conceivably an underestimation of the true quantity as not all ACPA activity could be recovered after purification possibly as a result from the isolation procedure. Nonetheless, we feel that the estimations presented provide a good reflection of total Ig-ACPA levels as also evidenced by the good correlation between ACPA-IgG levels in serum and purified ACPA-IgG concentrations (Spearman's rank correlation coefficient: 0.931. p<0.01).

The amounts of IgG-ACPA (mean:28 μg/ml) present in the sera are remarkably in line with the peak levels of IgG directed against tetanus following repetitive vaccinations (20–28 μg/ml).19 Protective antibody titres after vaccination have been described as titres above 1 μg/ml against, for example, Haemophilus influenza type b and group B streptococci.19–22 Surprisingly, the ACPA concentrations found exceed these protective antibody titres and are in the same range as the amount of antibody present shortly after vaccination. The relatively high ACPA antibody levels might be related to the continuous presence of citrullinated antigens in the joint, which could activate ACPA-producing B cells.

Furthermore, the results indicate the abundant presence of IgM-ACPA in synovial fluid, as up to one in eight IgM antibodies can be ACPA in patients with high ACPA levels. IgM responses against T cell dependent antigens are, in general, not continuously present. In the setting of vaccination, for example, levels of antigen-specific IgM decrease within weeks after immunization against rabies.23 Therefore the high concentrations of IgM-ACPA in sustained disease is intriguing and suggest the continuous local production and conceivably persistence of autoreactive B cell clones at the site of the inflamed joint. Previously, we reported that some IgG-ACPA-positive patients harbour IgM-ACPA 7 years after the initial presence of IgG-ACPA.10 Since IgM antibodies have a half-life of days and long-lived plasma cells producing IgM against protein-antigen have not been described in humans, the continuous presence of IgM against T cell-dependent antigens, indicate the continuous triggering of newly generated B cells. This suggests also that novel IgM-producing B cells are continuously recruited into the ACPA response, indicating that the ACPA response is continuously reactivated during the course of arthritis.10

Nonetheless we feel that the data presented on IgM-ACPA levels should be taken with caution as we can formally not exclude that the measured IgM-ACPA levels are influenced by IgM-rheumatoid factor (RF) bound to ACPA. However, previously reported data by our group showed that the depletion of IgM-RF did not result in reduction of IgM-ACPA levels. Furthermore mixing sera of RF-positive with IgM-ACPA-negative patients did not change the reactivity.10 Likewise, not all IgM-ACPA-positive patients included in this manuscript were IgM-RF-positive, excluding a contribution of IgM-RF, at least in these patients.

In conclusion, high concentrations of ACPA are present in serum and synovial fluid of ACPA-positive patients with RA, exceeding protective antibody levels against recall antigens. Furthermore, the abundance of IgM-ACPA in synovial fluid indicates ongoing recruitment of new B cells into the ACPA response, reflecting a continuous (re)activation of the RA-specific ACPA response during the course of arthritis.


Grant support: This study was supported by grants from the Centre for Medical Systems Biology (CMSB) and Netherlands proteomics platform within the framework of the Netherlands Genomics Initiative (NGI); by the European Union (Sixth Framework Programme integrated project Autocure, Seventh Framework Programme integrated project Masterswitch) and Innovative Medicines Initiative (IMI) funded project BeTheCure. AW's work is supported by the Dutch Organisation for Scientific Research (AGIKO grant). REMT's work is supported by a NWO-ZonMW VICI grant from the Dutch Organisation for Scientific Research. LAT's work is supported by an NWO-ZonMW VIDI grant and a fellowship from Janssen Biologics BV.



  • Handling editor Tore K Kvien

  • AW and JS contributed equally.

  • Contributors All authors have contributed to the preparation, analyses and finalizing of this manuscript.

  • Competing interests None.

  • Ethics approval Local ethics committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.