Lew.1AV1 and Dark Agouti (DA) rats, originating from the Zentralinstitut für Versuchstierzucht, Hannover, Germany, were bred, kept and used under pathogen-free conditions at the animal department of Karolinska University Hospital, Stockholm, Sweden. Animals were fed with standard rodent food and water ad libitum. Experiments were performed with female Lew.1AV1 or male DA rats, aged 9 to 14 weeks, with approval from the Stockholm North Ethical Committee.

CII was prepared from Swarm's rat chondrosarcoma by pepsin digestion and purification as previously described 3738. CII was stored freeze-dried at -20°C until dissolved in 0.01 M acetic acid and dialysed against the citrullination buffer, 0.1 M Tris-HCl (pH 7.6) containing 10 mM CaCl₂ and 5 mM dithiothreitol. RSA (Sigma-Aldrich, Steinheim, Germany) was dissolved directly in citrullination buffer. Proteins were incubated with rabbit skeletal muscle PAD (Sigma), at a concentration of 2 U/mg protein, for 2 hours at 37°C. Citrullination was terminated by adding 20 mM EDTA and subsequently dialysed successively against 2 mM EDTA (pH7.6) and Milli-Q water containing 10 mM Tris-HCl (pH7.6) at 4°C. Control proteins were treated similarly, apart from the addition of PAD. Protein solutions were then freeze-dried and stored at -20°C. For experimental use, freeze-dried proteins were dissolved in appropriate buffer, namely RSA in PBS and CII in 0.01 M acetic acid.

To verify the purity and amount of proteins, aliquots of samples were run in SDS-PAGE gels followed by Coomassie Brilliant Blue staining. Total protein quantification was also performed with a modified version of the Bradford method, using the Coomassie Plus Protein Assay Reagent Kit (Pierce Biotechnology, Rockford, IL, USA) in accordance with the manufacturer's instructions. To verify successful citrullination, immunoblotting was performed with RA3 39, a human recombinant antibody directed against citrulline (Fig. 1). In brief, 5 μl of citrullinated CII (Cit-CII), CII, Cit-RSA and RSA (concentration 2 mg/ml) were applied to a nitrocellulose membrane (BioTrace NT, product no. 66485; Pall Life Sciences, Ann Arbor, MI, USA). The membrane was incubated for 1 hour in blocking buffer (PBS containing 0.05% Tween 20 with 5% (w/v) non-fat dried milk) then incubated for a further 1 hour with primary antibody RA3, diluted 1:5 in blocking buffer, at room temperature (20–25°C). After incubation with a horseradish peroxidase (HRP)-conjugated donkey anti-human antibody (diluted 1:1,000 in blocking buffer; Amersham Pharmacia Biotech, Little Chalfont, Bucks., UK) for 1 hour at room temperature, the protein–antibody complexes were detected by enhanced chemiluminescence with the ECL system (Amersham Pharmacia Biotech, Uppsala, Sweden).

To increase the likelihood of detecting citrullinated proteins and the enzyme PAD4 in the joints of arthritic animals, we chose to perform the immunohistochemical stainings in the DA rat, because this strain has proven to develop the most severe CIA.

In contrast, the Lew.1AV1 rat strain develops a milder disease and was therefore selected when investigating the additive arthritogenic effects of Cit-CII. Lew.1AV1 rats were also used when studying the effects of citrullination on tolerance to RSA.

Anaesthetised rats were immunised in the base of the tail: male DA rats with 150 μg of CII dissolved in 100 μl of 0.01 M acetic acid, emulsified with an equal volume of Freund's incomplete adjuvant (FIA; Difco, Detroit, MI, USA) (intradermally); female Lew.1AV1 rats with 120 μg of CII or Cit-CII dissolved in 100 μl of 0.01 M acetic acid plus 100 μl of FIA (intradermally) or with 180 μg of RSA or Cit-RSA dissolved in 150 μl of PBS plus 150 μl FIA (subcutaneously). The animals immunised with CII and Cit-CII were monitored for signs of arthritis from day 12 after immunisation, in accordance with a previously described procedure 40. Each paw was divided into three groups of joints, the interphalangeal joints of the digits, the metacarpophalangeal and wrist joints in the forepaws and the metatarsophalangeal and ankle joints in the hind paws. In brief, 1 point signifies swelling of one group of joints, 2 points signifies two groups of swollen joints, 3 points signifies three groups of swollen joints and 4 points signifies swelling of the entire paw. The maximum score was thus 16 for each rat.

Individual serum samples from Lew.1AV1 rats immunised with RSA and Cit-RSA were obtained at different time points, namely 12, 24 and 35 days after immunisation (by tail bleeding) as well as at 61 days after immunisation (by heart puncture). ELISA plates (96-well Maxisorp; Nunc, Roskilde, Denmark) were coated with 10 μg/ml RSA or Cit-RSA (diluted in PBS), overnight at 4°C. Plates were washed with PBS-Tween (0.05%) and sera (serially diluted in PBS containing 0.05% Tween) were added in duplicate. After 2 hours of incubation at room temperature, plates were washed as above and incubated with alkaline phosphatase-conjugated goat anti-rat IgG (diluted 1:5,000 in PBS containing 0.05% Tween; Jackson Immunoresearch Lab, West Grove, PA, USA) for a further 2 hours at room temperature. Finally, a phosphatase substrate (Sigma, St Louis, MO, USA) was added and absorbance was determined at 405 nm with an Emax precision microplate reader (Molecular Devices).

To investigate the risk of contamination by PAD in our Cit-RSA sample, giving a false positive result in our Cit-RSA ELISA, we performed an anti-PAD ELISA. Plates were coated with the same amount of PAD expected to contaminate the Cit-RSA sample in the Cit-RSA ELISA, namely 0.0432 μg/ml. The ELISA was performed with the same procedure as described for Cit-RSA and RSA (see above).

In addition, to quantify antibodies specific for anti-citrullinated protein, an in-house CCP ELISA was used. Control reactions with corresponding cyclic arginine-containing peptides were also included. In brief, Streptawell plates were coated with 1 μg of biotinylated cfc1-cyc citrullinated peptide or cf0-cyc control peptide (diluted in PBS, 0.1% BSA) overnight at 4°C. Rat serum samples (diluted 1:50 in PBS containing 0.05% Tween and 1% BSA) were added in duplicate and incubated for 1 hour in a 37°C humid chamber. Plates were washed before the addition of HRP-conjugated rabbit anti-rat IgG (Sigma P216; diluted 1:1,000 in PBS containing 0.05% Tween and 1% BSA), followed by incubation for 1 hour in a humid chamber at 37°C. Bound antibodies were revealed by the addition of 3,3',5,5'-tetramethylbenzidine (Sigma). The enzymic reaction was stopped with 0.5 M H₂SO₄ after 30 min, and absorbance was determined at 450 nm.

Inguinal lymph nodes from Lew.1AV1 rats immunised with RSA and Cit-RSA were removed 10 days after immunisation; single-cell suspensions were prepared and resuspended in Dulbecco's modified Eagle's medium supplemented with glutamine, penicillin, streptomycin, HEPES and 10% fetal calf serum (Life Technologies, Paisley, Renfrewshire, UK). Cells were cultured in vitro for 72 hours at 10⁶ cells per well in triplicate (96-well flat-bottomed culture plates; Nunc) in the presence of RSA (10 μg/ml), Cit-RSA (10 μg/ml), PBS or concanavalin A (2 μg/ml; Sigma-Aldrich). [³H]Thymidine (PerkinElmer Life Sciences Inc., Boston, MA, USA), was added (1 μCi per well) for the final 16 hours of culture. Cells were harvested with a Tomtec cell harvester and [³H]thymidine incorporation was measured as counts per minute (c.p.m.) in a Wallac Trillux 1450 microbeta counter. Stimulation was calculated and expressed as stimulation index (C.p.m. after stimulation/C.p.m. of background).

Recombinant anti-citrullinated protein single-chain variable fragment (scFv) antibody RA3, selected from RA-patient-derived phage display libraries and control scFv human anti-U1-70K have been described elsewhere 39. Rabbit antibodies directed against chemically modified citrulline (anti-MC antibodies), developed by Dr Tatsuo Senshu and colleagues 541, were purchased from Upstate Biochemicals (Lake Placid, NY, USA). Polyclonal antibodies recognising PAD4 (SN823) were produced by immunising rabbits with PAD4 isotype-specific peptides (amino acids 210 to 225 and 517 to 531) and by affinity purification with a CNBr-activated Sepharose 4B column, as described previously 10. Preimmune rabbit IgG was used as control for PAD4.

Zamboni-fixed 42 cryopreserved sections of synovial tissue from CII-immunised DA rats were stained for expression of citrullinated proteins and PAD4. Endogenous peroxidase activity was blocked by treatment for 30 min in darkness at room temperature with 1% hydrogen peroxide and 2% sodium nitride dissolved in PBS. After washing with PBS, 2% normal goat serum was added to block non-specific binding sites. Sections were then incubated with 200 μl of primary antibody (RA3 (undiluted) or anti-PAD4 (diluted 1:24)) overnight in a humid chamber at 4°C. An isotype-matched recombinant human anti-U1-70K antibody and a preimmune rabbit IgG were used as respective controls. Slides were then washed in PBS before incubation with 200 μl of appropriate biotin-labelled secondary antibody (donkey anti-human IgG (Jackson Immuno Research) diluted 1:1,000 or goat anti-rabbit IgG (Vector, Burlingame, CA, USA) diluted 1:800) for 1 hour at room temperature. After further washing, 200 μl of avidin-biotin-HRP (Vectastain; Vector) prepared in accordance with the manufacturer's directions was applied for 60 min at room temperature. A final wash was followed by addition of the substrate diaminobenzidine (Peroxidase Substrate Kit; Vector) for 5 min. The colour reaction was stopped by washes in tap water, after which sections were counterstained with Mayer's haematoxylin. Finally the slides were dried and mounted with buffered glycerol, and evaluation was performed by microscopy with a Polyvar II microscope (Reichert-Jung, Vienna, Austria) connected to a charge-coupled device colour camera (DXC-750P; Sony Corp., Tokyo, Japan).

In addition, anti-MC antibodies were used to confirm the citrulline staining obtained with RA3. Before incubation with primary antibody, sections were treated for 3 hours at 37°C in a chemical modification solution consisting of one part solution A (0.025% (w/v) FeCl₃, 4.6 M H₂SO₄, 3.0 M H₃PO₄) plus one part solution B (0.5% diacetyl monoxime, 0.25% antipyrine, 0.5 M acetic acid) (anti-Citrulline (Modified) Detection Kit; Upstate Biochemicals). Control sections were incubated in a solution containing one part solution A and one part Milli-Q water. After extensive washing in PBS, slides were incubated for 40 min at room temperature with 1% H₂O₂, washed in PBS and incubated for 30 min at room temperature with 5% normal goat serum in PBS plus 1% BSA. Primary antibody (anti-MC), diluted 1:1,000 in PBS plus 1% BSA, was added and slides were incubated overnight at room temperature. Slides were then washed in PBS, followed by incubation for 30 min at room temperature with secondary antibody (biotinylated goat anti-rabbit IgG; Vector) diluted 1:800 in PBS plus 1% BSA. The subsequent steps were performed as described above.

All data were evaluated with the Mann–Whitney U-test for independent groups, except for arthritis incidence, which was evaluated by Kaplan–Meier survival analysis.

To examine the presence of citrullinated proteins and the enzyme PAD4 in the joints at different stages of experimental arthritis, histological analyses of rat ankle joints were performed by immunohistochemistry. Citrullinated proteins could be detected in the joints of arthritic animals with the RA3 antibody (Fig. 2 and Table 1). The first appearance of protein citrullination was noted after disease onset, at day 21 after immunisation, and increased staining was observed as the disease progressed into a more severe, chronic state (namely 28 and 38 days after immunisation). Unimmunised animals and time points before clinical signs of arthritis (namely 0, 3, 6 and 10 days after immunisation) as well as the time of disease onset (namely 15 days after immunisation) were negative for citrulline. Some infiltrating cells as well as the cartilage surface stained positively for citrulline and the major sources of citrullinated proteins in the arthritic joint were extracellular deposits, presumably fibrin deposits. The occurrence of citrullinated proteins in the joints is specific, because other investigated organs such as lung, ear, spinal cord, spleen, lymph nodes and salivary glands were all negative (data not shown).

For comparison with previous studies in humans 43 and mice 13 we also used antibodies targeting chemically modified citrulline. This method circumvents the risk of epitope blocking by residues flanking citrulline, because the citrulline side-chain becomes so bulky through the chemical modification that antibody recognition cannot be influenced by other amino acids 541. Here we confirmed the results obtained with RA3. A similar staining pattern was observed, with positive cells, cartilage and extracellular deposits, whereas control stainings of non-modified sections were negative (data not shown).

PAD4, which has been reported to be present in mouse and human arthritic synovia 1013, could also be detected in synovial tissue of this rat arthritis model from 21 days after immunisation (Fig. 3 and Table 1). The number of positive cells localised to the synovial infiltrate increased by days 28 and 38 after immunisation. PAD4 was not evident in healthy synovial tissue, nor in sections from 3, 6 and 10 days after immunisation; although an apparent inflammation was observed, PAD4 could not be detected 15 days after immunisation.

Lew.1AV1 rats were immunised with the non-immunogenic autoantigen RSA or the modified counterpart Cit-RSA and the differences in induced immune responses were analysed in vitro. The kinetics of the antigen-specific IgG response was investigated and the results revealed not only a response towards the modified protein, but also cross-reactivity to the unmodified form of RSA (Fig. 4). Antibodies were detected from 12 days after immunisation, had increased by day 24 after immunisation and persisted for a further 40 days. Cross-reactivity was demonstrated at all time points. No B cell response was noted in animals immunised with unmodified RSA.

Animals immunised with Cit-RSA and RSA were also tested for anti-PAD IgG responses. The anti-PAD ELISA was negative both for animals immunised with Cit-RSA and for animals immunised with RSA (data not shown), indicating that the amount of PAD contaminating the Cit-RSA sample did not result in any false positive result in the Cit-RSA ELISA, described above.

In addition, by performing both an anti-CCP ELISA and a control ELISA containing cyclic arginine-containing peptides, we could confirm that the animals immunised with Cit-RSA produced antibodies against citrullinated epitopes and also antibodies recognising arginine epitopes (data not shown).

The T cell response was evaluated by [³H]thymidine incorporation 10 days after immunisation, as depicted in Fig. 5. Here, a proliferative response was demonstrated in animals immunised with unmodified RSA, both when stimulated with the same antigen and when stimulated with the citrullinated antigen. Cells from animals immunised with Cit-RSA showed a stronger proliferation than the RSA-immunised rats in response to the modified protein. The same tendency was observed when using RSA as a stimulus, although it was not statistically significant.

Unmodified or citrullinated rat CII was, together with adjuvant, injected into Lew.1AV1 rats; arthritis development was monitored by blinded macroscopic evaluation. The citrullinated form of CII induced arthritis with significantly higher incidence and earlier onset than did the same amount of unmodified CII (Fig. 6). The incidence was 35% higher in the Cit-RSA group during the early phase of the disease and 15% higher during the end stage. Disease onset in this group occurred 13 days after immunisation, compared with 16 days after immunisation in the control group. A trend towards higher mean arthritis score was also observed in the affected animals in the Cit-CII group compared with the CII group, although this was not statistically significant.