Detection of autoantibodies to citrullinated BiP in rheumatoid arthritis patients and pro-inflammatory role of citrullinated BiP in collagen-induced arthritis

Serum samples were obtained from 100 RA patients, 60 systemic lupus erythematosus (SLE) patients, and 30 healthy volunteers. All of the RA patients fulfilled the 1987 and 2010 American College of Rheumatology criteria for RA [23, 24]. SLE diagnosis was made according to the diagnostic criteria of the American College of Rheumatology [25, 26]. Written informed consent was obtained before blood samples were taken. This study was approved by the ethics committee of Tokyo University Hospital.

DBA/1J female mice (age: 6-8 weeks) were obtained from SLC Japan (Shizuoka, Japan). All animal experiments were conducted in accordance with institutional and national guidelines.

A full-length fragment of mouse BiP cDNA (GenBank: AJ002387) was subcloned, and the plasmid vector, which contained a mouse BiP cDNA sequence with a His⁶-tag at its N-terminal, was prepared using a Champion pET Directional TOPO Expressing Kit (K100-01; Invitrogen, Carlsbad, CA, USA). The vector was transduced into BL21 Star One Shot Chemically Competent Escherichia coli (Invitrogen) by using a heat shock procedure. The transformed BL21 E. coli were incubated in LB medium with ampicillin (100 μg/mL). When the absorbance (O.D. 650 nm) rose to 0.5, 1 mM of isopropylthio-beta-galactoside (Invitrogen) was added to the medium. After a 6 h incubation, the pellet was dissolved using the Fast Break Cell Lysis Reagent (Promega, Madison, WI, USA) and the lysate was purified using an Ni-NTA Fast Start Kit (QIAGEN, Valencia, CA, USA) according to the manufacturer's procedure. Mouse fibrinogen was purchased from Sigma-Aldrich (St. Louis, MO, USA). For endotoxin removal, purified proteins were passed through a Detoxi-Gel AffinityPak prepared column (Pierce, Rockford, IL, USA) according to the manufacturer's protocol.

For protein citrullination, 2 U of rabbit PADI2 (Sigma-Aldrich) were added to 1 mg of the protein in buffered medium (0.1 M Tris-HCl, 10 mM CaCl₂, 5 mM DTT) at 37°C for 2 h. EDTA (20 mM) was added to stop the reaction [15]. The samples were then passed through the Ni-NTA column to remove PADI2.

The peptides (Tables 1 and 2) were synthesized using NeoMPS (San Diego, CA, USA). Native BiP-derived peptides were also prepared that covered the entire human BiP sequence (GenBank: CAA61201) (Table 1). BiP includes 27 arginine residues that could be replaced by citrulline; therefore, we designed 27 synthesized peptides that each contained 1 citrulline residue (Table 2). In some experiments, we also prepared original peptides without citrullination. Peptide purity was >80%.

Murine BiP and citBiP were emulsified with an equal volume of complete Freund's adjuvant (CFA) (Chondrex, Redmond, WA, USA), and DBA/1J mice were immunized with 100 μg of murine BiP or citBiP intradermally at the tail base. Spleen samples for culture were obtained at 14 days post-immunization and serum samples for antibody titer measurements were obtained at 28 days post-immunization.

The serum levels of the anti-CCP antibodies were measured using MESACUP FCCP200 (Axis-Shield Diagnostics Ltd., Dundee, UK). Anti-CCP antibodies in the human serum samples were measured according to the manufacturer's protocol. Goat anti-mouse IgG-alkaline phosphate (Chemicon, Temecula, CA, USA) was used at a 3000-fold dilution as the detection antibody for measuring mouse serum anti-CCP antibodies [15]. In the competitive assay, the indicated concentrations of citrullinated peptides were added to human anti-CCP antibody-positive standard serum samples (a mixture of serum from 18 ACPA-positive RA patients), the samples were incubated for 2 h at room temperature, and the anti-CCP antibody levels were then measured.

To detect antibodies to other proteins, 5 μg/mL of each antigen was dissolved in 0.1 M NaHCO₃ and 100 μL of each were plated in 96-well plates (Immuron4 Multiplate; Dynatech Laboratories, Chantilly, VA, USA) at 4°C for 16 h. After the plates were blocked using PBS containing 3% skimmed milk and 0.5% Tween-20, ×100 diluted samples were applied. Goat anti-human IgG-horseradish peroxidase (HRP) (Vector, Burlingame, CA, USA) or goat anti-mouse IgG-HRP antibodies (Zymed, South San Francisco, CA, USA) were used at a 5000-fold dilution as detection antibodies and TMB solution (KPL, Gaithersburg, MD, USA) was used to develop the colors. An automatic microplate reader (Bio-Rad 550; Bio-Rad, Hercules, CA, USA) was used to measure optical density.

After 1 μg BiP or citBiP was electrophoresed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis, the samples were blotted to Hybond ECL (GE Healthcare). The membranes were blocked with PBS containing 10% skimmed milk at room temperature for 1 h and ×100 diluted serum samples were then applied at 4°C for 16 h. A goat anti-human IgG-HRP antibody was used at a 5000-fold dilution as the detection antibody. The membranes were treated with ECL reagent (GE Healthcare, Buckinghamshire, UK) and the film images were developed. A polyclonal rabbit anti-BiP antibody (Thermo Fisher Scientific, Rockford, IL, USA) was used at a 100-fold dilution as a positive control. A Senshu antibody (rabbit polyclonal anti-modified citrulline antibody; Upstate Biochemicals, Lake Placid, NY, USA) was used to detect the citrullinated antigens, and blotting was performed in accordance with the manufacturer's protocol [27]. Band density was quantified using NIH ImageJ software.

Rat esophageal sections (SCIMEDX Corporation, Denville, NJ, USA) were used, and citrullinated filaggrin was detected according to the manufacturer's protocol. In short, mouse serum samples were treated at room temperature for 1 h and Alexa Fluor 488-conjugated goat anti-mouse IgG antibody (Invitrogen) was used as the detection antibody. The samples were observed under a fluorescent microscope [15].

CIA was induced as described previously [28]. In short, bovine type II collagen (Chondrex) was emulsified with an equal volume of CFA or incomplete Freund's adjuvant (IFA) (Chondrex). Fourteen days after immunization with BiP or citBiP, DBA/1J mice were immunized with 50 μg bovine type II collagen intradermally at the tail base on day 14 with CFA and on day 35 with IFA. The arthritis score was determined by the degree of erythema, swelling, or ankylosis observed on each paw, as described elsewhere [29]. The mice were sacrificed at 50 days after the first CIA immunization. Tissue samples were embedded in paraffin wax after 10% formaldehyde fixation and decalcification. The sections were stained with hematoxylin and eosin. Synovial tissues were graded by mononuclear cell infiltration, pannus formation, and cartilage erosion, as described previously [30].

We first measured the serum levels of anti-BiP and anti-citBiP antibodies in RA patients, SLE patients, and healthy controls (Figure 1A). Among these patients, 95 RA patients and no SLE patients were positive for anti-CCP antibodies. We chose SLE patients as disease controls because ~30% of SLE patients are positive for anti-BiP antibodies, while serum anti-CCP antibodies are detected in <5% of SLE patients [31]. In RA patients, the levels of anti-BiP and anti-citBiP antibodies were significantly higher than those observed in SLE patients and healthy controls (p < 0.001). Immunoblotting analysis demonstrated higher serum concentrations of anti-citBiP antibodies than anti-BiP antibodies in the serum of each anti-CCP antibody-positive RA patient (Figure 1B). Seventy-two percent of RA patients showed higher anti-citBiP antibody levels than anti-BiP antibody levels, whereas there was no significant difference in the serum titers of anti-BiP and anti-citBiP antibodies in the non-RA controls (Figure 1C). A previous report suggested that serum anti-BiP antibodies were detected in SLE patients [31]; in our study, the serum levels of anti-BiP and anti-citBiP antibodies in SLE patients were mildly increased in comparison with those in healthy controls (p < 0.05) (Figure 1A). However, in contrast to RA, there was no significant difference in the serum levels of anti-BiP and anti-citBiP antibodies in SLE patients (Figure 1C). In RA patients, the serum levels of anti-citBiP antibodies were strongly correlated with those of anti-CCP antibodies (R² = 0.59), whereas the levels of anti-BiP antibodies were less correlated with those of anti-CCP antibodies (R² = 0.38) (Figure 1D). Therefore, the serum levels of anti-citBiP antibodies are higher than those of anti-BiP antibodies and have a clear association with anti-CCP antibodies in RA patients.

Because BiP is an immunoglobulin chaperon protein, there is the possibility that immunoglobulins bind to BiP, independent of epitope recognition by the antigen-binding site. To confirm the existence of specific epitope recognition, epitope mapping was performed by using native human BiP-derived peptides. The mapping revealed that BiP1 (BiP_21-40), BiP16 (BiP_246-265), BiP39 (BiP_591-610), BiP40 (BiP_606-625), and BiP42 (BiP_636-655) were the epitopes recognized by the anti-BiP antibodies (Figure 1E). Most of these major epitopes were located in the C and N terminals. Therefore, anti-human BiP antibodies truly exist in the serum of RA patients.

To identify the citrullinated epitopes of the anti-citBiP antibodies in RA, we prepared citrullinated peptides in which the arginine residues of the native BiP sequences were replaced with citrulline residues. Serum antibody levels to each of the citBiP-derived peptides were measured in RA patients with high anti-citBiP antibody levels (n = 18) and healthy controls (n = 5). The serum antibody levels were increased against several citrullinated peptides (Figure 2A). Four of the major citBiP-derived epitopes were selected for further investigation, and the serum levels of the anti-native peptide antibodies and the anti-citrullinated peptide antibodies were compared (Figure 2B). The serum antibody levels to No. 12 (BiP_279-298 R289citrulline), 15 (BiP_295-314 R305citrulline), and 23 (BiP_500-519 R510citrulline) citrullinated peptides were significantly increased compared with those of the native peptides in RA patients. Thus, these citrullinated residues were thought to be important residues for autoantibody recognition in RA. Thus, the positions of the major citrullinated epitopes (BiP_279-298, BiP_295-314, and BiP_500-519) were different from the major epitopes of the native BiP protein (BiP_21-40, BiP_246-265, BiP_591-610, BiP_604-625, and BiP_636-655). In particular, the No. 23 citrullinated peptide inhibited serum antibody binding to CCP, whereas the No. 22 peptide did not (Figure 2C). No. 12 and 15 citrullinated peptides did not bind to CCP either (data not shown). Therefore, we identified several citrullinated anti-citBiP antibody epitopes, one of which showed a clear cross-reaction with CCP.

We then examined whether active immunization with citBiP could induce ACPAs in a mouse model. DBA/1J mice were immunized with BiP or citBiP and serum samples were obtained after 28 days. Anti-CCP antibodies were detected in the serum samples of citBiP-immunized mice (Figure 3A). Although BiP immunization also induced the development of anti-CCP antibodies, the anti-CCP antibody levels were relatively low. Because Vossenaar et al. pointed out that some mouse models of inflammatory arthritis developed false-positive anti-CCP antibodies [32], we verified the presence of ACPAs using different methods. The serum levels of anti-citrullinated fibrinogen antibodies were only detected in the sera of citBiP-immunized mice (Figure 3B). We stained rat esophageal horny layer containing citrullinated filaggrin with sera from citBiP-immunized mice (Figure 3C). Therefore, we confirmed that citBiP immunization not only induced anti-citBiP antibodies but also induced other ACPAs.

Finally, we tested the pro-inflammatory effects of citBiP in the CIA mouse model. Similar to the findings in RA, CIA mouse sera contained anti-BiP and anti-citBiP antibodies (Figure 4A). The serum levels of anti-citBiP antibodies were higher than those of anti-BiP antibodies. Since the development of ACPAs precedes the onset of RA, some immune responses to citrullinated antigens are hypothesized to occur before RA becomes clinically obvious. We assessed this hypothesis by pre-immunizing mice with citBiP before inducing CIA, and determined whether the immune responses to citBiP were associated with the enhancement of inflammatory arthritis. Fourteen days after BiP or citBiP immunization, the mice were immunized with bovine type II collagen to induce arthritis. Joint inflammation developed sooner and was significantly exacerbated in the citBiP-pre-immunized mouse group as compared to the control group (Figure 4B and Additional file 1). BiP-pre-immunized mice did not develop severe arthritis. There was a significant increase in the histological scores of the citBiP-pre-immunized mouse group (Figure 4C). The serum levels of anti-CCP antibodies were also significantly elevated in the citBiP-pre-immunized mice (Figure 4D). Therefore, citBiP immunization exacerbated inflammatory arthritis in mice, whereas BiP immunization had no significant effect on CIA.