1. Top of page
  2. Abstract
  6. Acknowledgements


Autoantibodies to citrullinated proteins (ACPAs) are specific for rheumatoid arthritis (RA) and probably are involved in its pathophysiology. Citrullyl residues, posttranslationally generated by peptidyl arginine deiminase (PAD), are indispensable components of ACPA-targeted epitopes. The aim of this study was to identify which PAD isotypes are expressed in the synovial tissue (ST) of patients with RA and are involved in the citrullination of fibrin, the major synovial target of ACPAs.


Expression of all PAD isotypes, including the recently described PAD type 6 (PAD-6), was explored by reverse transcription–polymerase chain reaction and immunoblotting, first in blood-derived mononuclear leukocytes from healthy donors, then in ST samples from 16 patients with RA and 11 control patients (4 with other arthritides and 7 with osteoarthritis [OA]). In ST samples from patients with RA, PADs were localized by immunohistochemistry.


In lymphocytic and monocytic cells and, similarly, in ST samples from patients with RA, the PAD-2, PAD-4, and PAD-6 genes were found to be transcribed, but only PAD-2 and PAD-4 enzymes were detected. PAD-2 was also expressed in ST from control patients, including those with OA, while PAD-4 was preferentially expressed in ST from patients with other arthritides. In RA, the expression levels of PAD-2 and PAD-4 were correlated with the intensity of inflammation (cell infiltration, hypervascularization, and synovial lining hyperplasia), and both enzymes were demonstrable within or in the vicinity of citrullinated fibrin deposits.


PAD-2 and PAD-4 are the only PAD isotypes expressed in the ST of patients with RA and those with other arthritides. Inflammatory cells are a major source, but PAD-4 also comes from hyperplastic synoviocytes. Both isotypes are probably involved in the citrullination of fibrin.

Chronic inflammation of synovial joints with frequent extraarticular manifestations is an essential characteristic of rheumatoid arthritis (RA), the most common human autoimmune disease. The presence of autoantibodies to “citrullinated” (deiminated) proteins (anti–citrullinated protein antibodies [ACPAs]) in the serum of patients also constitutes a major feature of RA. These autoantibodies very probably play a significant role in the pathophysiology of RA. Indeed, ACPAs are most likely the most disease-specific of the RA-associated autoantibodies. They are detectable in the serum years before the onset of arthritis symptoms, and a significant positive correlation exists between the serum titer and clinical, biologic, and radiologic data related to RA activity and/or severity (for review, see ref.1). In addition, ACPAs are produced by plasma cells of rheumatoid synovial tissue (ST) and therein not only concentrate (2) but also probably interact with citrullinated proteins, among which citrullinated fibrin constitutes their major target (3).

Even if not sufficient, citrullyl residues are essential to the formation of ACPA epitopes on the target antigens (4–6). These epitopes result from the posttranslational transformation of the positively charged guanidino group of arginyl residues into the uncharged ureido group of citrullyl residues. A Ca2+-dependent enzyme activity, designated peptidyl arginine deiminase (PAD; also called protein-L-arginine iminohydrolase [EC]), catalyzes such conversion. Five PAD isotypes (PAD-1, PAD-2, PAD-3, PAD-4, and PAD-6), encoded by 5 paralogous genes (PADI1, PADI2, PADI3, PADI4, and PADI6, respectively) clustered on chromosome 1p35–36, have been described in humans (7). Their cellular expression pattern has not yet been extensively explored, particularly at the protein level. PAD-1 has been detected in the epidermis, in hair follicles, in arrector pili muscles, and in sweat glands (8–10). PAD-2 has been widely detected, notably in brain astrocytes (11, 12), sweat glands (9, 13), arrector pili muscles (9), macrophages (14), and epidermis (8, 15). To date, PAD-3 expression has been reported only in the upper layers of epidermis and in hair follicles (8–10, 16). PAD-4 expression has so far been described only in hematopoietic cells (14, 17–19). PAD-4 differs from other PAD isotypes by its capacity to undergo nuclear translocation (20). No data are available concerning expression of PAD-6 in human tissue, but PADI6 messenger RNA (mRNA) is present mainly in ovary, testis, and peripheral blood leukocytes (7).

Protein “citrullination” (deimination) leads to alterations in intramolecular and intermolecular interactions of the protein targets (21). It is implicated in terminal differentiation of epidermis (22) and in brain development (23). Nuclear PAD-4 could regulate gene expression via chromatin remodeling (24, 25). Diseases in which citrullination has been shown or suggested to play a role include not only RA but also multiple sclerosis, as well as psoriasis, Alzheimer's disease, primary open-angle glaucoma, and obstructive nephropathy (26).

One or several PADs are necessarily responsible for generation of the ACPA-targeted epitopes in rheumatoid ST, but therein the presence of the 5 PAD isotypes, including the most recently described PAD-6, has not yet been systematically explored. In the present study, we determined which of the 5 PADs are expressed in rheumatoid ST, at both the mRNA level and the protein level. After assessing the expression of PADs in human blood-derived mononuclear leukocytes, PADs were investigated in ST samples obtained from a large series of patients with RA, compared with patients with other arthritides or osteoarthritis (OA). Moreover, correlations were sought between the levels of PAD expression and the intensity of ST inflammation, detected histologically, both parameters being evaluated by semiquantitative methods. Finally, to identify the PAD(s) most likely to be involved in the generation of citrullinated fibrin, immunohistochemical analyses of ST samples from patients were RA were performed to analyze whether this major ACPA target is colocated with the ST-expressed PADs.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Patients and serum samples.

ST samples were obtained consecutively from 27 patients undergoing surgery of the hand, wrist, elbow, or knee at the Department of Orthopedic and Traumatic Surgery of the Purpan Hospital of Toulouse. Rheumatologists established the clinical diagnoses according to international disease classification criteria (27–29). RA was diagnosed in 16 patients (patients RA1 to RA16; 12 women and 4 men, age range 36–87 years [median age 56.5 years]). The 11 control patients (7 women and 4 men, age range 21–80 years [median age 56 years]) included 4 patients with non-RA arthritides (2 with ankylosing spondylitis [patients AS1 and AS2], 1 with undifferentiated spondylarthritis [patient uSpA1], and 1 with psoriatic arthritis [patient PsA1]) and 7 patients with OA (patients OA1 to OA7). The presence or absence of ACPAs in contemporaneous serum samples was assessed by enzyme-linked immunosorbent assay (ELISA) on in vitro–citrullinated human fibrinogen (30). All control patients were ACPA-negative, and 9 of the 16 patients with RA were ACPA-positive.

ST specimens.

Each ST sample was wrapped in a gauze compress soaked in normal saline and kept at 4°C until dissected into identical sets of fragments representing all of its different macroscopic aspects. Sets destined for protein or RNA extraction were snap-frozen in liquid nitrogen. Sets destined for histologic analyses were fixed overnight in Bouin's solution and successively dehydrated for 24 hours in 7.5%, 15%, and 30% sucrose solutions. All sets were stored at −80°C until further processed.

Preparation of different blood-derived cell populations.

Different subpopulations of blood-derived cells were enriched from buffy coats from healthy adult donors (Établissement Français du Sang Pyrénées-Méditerranée, Toulouse, France). Buffy coats were diluted to one-half in phosphate buffered saline (PBS) containing 0.6% sodium citrate and 0.1% bovine serum albumin, and peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation over Ficoll 1,077 gm/ml (Biocoll; Biochrom, Berlin, Germany). A portion of the PBMC fraction was further fractionated using magnetic beads coated with anti-CD14 antibodies (CD14 MicroBeads; Miltenyi Biotec, Paris, France) according to the manufacturer's instructions. The CD14-negative and the CD14-positive mononuclear cells corresponded to the lymphocyte-enriched fraction and to purified monocytes, respectively. A portion of the monocytes was allowed to differentiate into macrophages, by 7-day culture in perfluoroalkoxy polymer culture inserts (VWR International, Fontenay-sous-Bois, France) containing a medium designed for macrophages (Macrophage-SFM; Invitrogen, Cergy-Pontoise, France) supplemented with 10% fetal calf serum (BioWest, Nuaillé, France) and 100 ng/ml macrophage colony-stimulating factor (M-CSF; PeproTech, Levallois-Perret, France). Macrophages were harvested either directly or after 24-hour activation with 0.5 μg/ml lipopolysaccharide (LPS) from Escherichia coli (Sigma-Aldrich, St. Quentin Fallavier, France). All cell samples were washed 3 times in PBS and stored as pellets at −80°C until further processed.

Analysis of PADI mRNA expression.

The RNeasy Mini Kit (Qiagen, Courtaboeuf, France) was used for total RNA preparation from ST or blood-derived cells, according to the manufacturer's instructions. For standard reverse transcription–polymerase chain reaction (RT-PCR) analysis, complementary DNAs (cDNA) were synthesized using 1 μg of RNA and an RT kit with oligo(dT) primers and random hexamers (ImProm-II Reverse Transcription System; Promega, Charbonnières, France) according to the manufacturer's instructions. PCRs were performed in 15 μl (total reaction volume), using 0.75 unit of Taq DNA polymerase (MP Biomedicals, Illkirch, France), 1.5 mM MgCl2, 0.2 mM of each dNTP, 5 pmoles of each primer, and 1 μl of cDNA (equivalent to 50 ng of RNA). The PCR conditions were as follows: 94°C for 3 minutes, followed by 37 cycles at 94°C for 30 seconds, at either 58°C (PADI1, PADI2, PADI3, and PADI4) or 62°C (PADI6) for 30 seconds, and at 72°C for 20 seconds. PCR amplification of the cDNA of G3PDH or HPRT1 was performed as control. Positive control reactions also included amplification of human epidermis cDNA (31), using primers specific for PADI1 and PADI3, of all-trans-retinoic acid–treated HL-60 cell cDNA (8) using primers specific for PADI2 and PADI4, and of human testis cDNA (BD Biosciences, Erembodegem, Belgium) using primers specific for PADI6.

For real-time RT-PCR analysis, total RNA from blood-derived cells was treated with amplification-grade DNase I (Invitrogen) before cDNA were synthesized, as described above. Amplification assays were performed and analyzed with the 7300 Real-Time PCR System and corresponding software (Applied Biosystems, Foster City, CA). PCRs were carried out using Power SYBR Green PCR Master Mix (Applied Biosystems) with 250 nM of each primer and cDNA templates corresponding to 20 ng of RNA. Positive control reactions included amplification of human skin cDNA using primers specific for PADI1 and PADI3. Fluorescence was quantified as threshold cycle (Ct) values. Samples were analyzed in triplicate, and Ct values exhibiting differences less than 0.3 cycles were averaged. The relative amount of each PADI gene was normalized to HPRT1 by using the difference in Ct (ΔCt) method. Negative control wells (corresponding to PCR without template) emitted no significant fluorescence after 40 cycles.

PADI amplification primers were designed after checking for the absence of similarity to any of the other PADI sequences, using BLAST (Basic Local Alignment Research Tool) analysis. Positions in the cDNA sequence, according to the indicated GenBank accession numbers, were as follows (when 2 primer pairs are mentioned, the first corresponds to standard PCR and the second to real-time PCR): for PADI1 (AB033768), nucleotides (nt) 1644–1663 and 1858–1877 and nt 1768–1788 and 1977–1998; for PADI2 (AB030176), nt 1807–1826 and 1984–2003 and nt 1619–1642 and 1759–1783; for PADI3 (AB026831), nt 1754–1773 and 1943–1962 and nt 1601–1625 and 1797–1821; for PADI4 (AB017919), nt 1780–1799 and 1926–1945; and for PADI6 (AY422079), nt 1683–1707 and 1813–1834.

Protein extraction.

ST samples were homogenized with an Ultra-Turrax homogenizer (T25 basic; IKA Labortechnik, Staufen, Germany) in an ice-cold lysis solution consisting of a 60-mM Tris HCl buffer (pH 7.4) containing 150 mM NaCl, 40 mM EDTA, 0.02% sodium azide, 1 mM phenylmethylsulfonyl fluoride, 10 mM N-ethylmaleimide, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 1.2 μM aprotinin, 25 μM leupeptin, 18 μM pepstatin A, 15 μM E-64, and 40 μM bestatin (all from Sigma-Aldrich). Pellets of blood-derived cells were placed in the lysis solution supplemented with 0.5% (volume/volume) Nonidet P40 (Sigma-Aldrich) and sonicated on ice at 8–10W (3 times for 15 seconds). ST and cell lysates were then centrifuged for 15 minutes at 15,000g and 4°C, and the supernatants (low-salt extracts) were collected and stored at −80°C until analyzed.

Anti-PAD antibodies.

Previously described and validated isotype-specific rabbit antipeptide antibodies directed against human PAD types 1–4 (purified antipeptide antibodies) were used (8, 31). An antiserum to human PAD-6 was elicited in rabbits by injection of 3 synthetic peptides conjugated to keyhole limpet hemocyanin via a natural or added N-terminal cysteine residue. These peptides were chosen in the most isotype-specific regions of the predicted amino acid sequence of human PAD-6 observed after multialignment of all human, mouse, and rat PAD sequences: peptide A6, (C)IYRNGQVEMSSDKQA (aa 130–144), peptide B6, (C)VEESQDPSIPETVLY (aa 280–294), and peptide C6, CLEKLTNIPSDQQPKRS (aa 591–607). Antipeptide antibody titers were determined by ELISA (CovalAb, Lyon, France). The antiserum was then affinity-purified on an equimolar mixture of the reactive peptides (peptides A6 and B6), coupled to an agarose-activated affinity column (SulfoLink Kit), essentially as described by the manufacturer (Perbio Science, Brebières, France). The sensitivities of detection of all the anti-PAD antibodies were roughly equalized by immunoblot titration against equimolar amounts of each PAD isotype, corresponding to PAD-2 purified from rabbit skeletal muscle (Sigma-Aldrich), recombinant human PAD-1, PAD-3, PAD-4, produced as previously described (10, 16, 17), and recombinant PAD-6 (see below).

Production of recombinant human PAD-6.

The entire coding sequence of human PADI6 previously cloned into the pCRII-TOPO plasmid (Invitrogen) (7) was subcloned into the same vector in carboxy-terminal fusion with a hexahistidine tag. The plasmid insert was verified by sequencing. Expression of recombinant PAD-6 was induced by IPTG (Sigma-Aldrich) treatment of the E coli TOP10 F′ transformants (Invitrogen). After cell lysis with 0.1 mg/ml lysozyme (Sigma-Aldrich) in phosphate buffer, PAD-6 was purified from the bacterial extract by chromatography on a HiTrap Chelating/Ni2+ column (GE Healthcare, Orsay, France), as suggested by the manufacturer. The PAD-6–containing fractions were pooled, precipitated in ice-cold absolute ethanol, and then redissolved in sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer.

Immunoblot detection of PAD.

Equal amounts (∼50 μg) of total proteins from ST or cell extracts (as adjusted using the Bradford total protein assay and successive visual examinations of Coomassie blue–stained gels) were separated by SDS-PAGE on 7.5% polyacrylamide gels and then electrotransferred onto reinforced nitrocellulose membranes (Hybond-C Extra; GE Healthcare). After blocking in 40 mM Tris HCl buffer (pH 8.0) containing 150 mM NaCl, 0.05% Tween 20, and 2.5% powdered skimmed milk (blotting buffer), the membranes were probed overnight at 4°C with the antibodies to PAD-1, PAD-2, PAD-3, PAD-4, and PAD-6 diluted in blotting buffer at 2.7 μg/ml, 0.9 μg/ml, 2.5 μg/ml, 1.1 μg/ml, and 5 μg/ml, respectively. Bound rabbit antibodies were detected with a peroxidase-conjugated goat antiserum to rabbit IgG (H+L; Southern Biotechnology, Birmingham, AL) diluted to 1:10,000 in blotting buffer. Peroxidase activity was visualized using enhanced chemiluminescence Western blotting reagents (GE Healthcare). Negative controls consisted of probing with the secondary antibody only. Rabbit skeletal muscle PAD-2 and recombinant human PAD-1, PAD-3, PAD-4, and PAD-6 were used as positive controls. In addition, protein extracts of human epidermis (8), HL-60 cells (17), and human adult ovary (BioChain Institute, Hayward, California) were used as sources of tissular PAD types 1–3, PAD-4, and PAD-6, respectively. The intensity of immunodetection of the various PADs in the ST samples was separately evaluated by summing the scores given by 2 readers, on a scale ranging from 0 to 3 (0.25 steps).

Histology and immunohistochemistry.

Bouin's–fixed and dehydrated ST samples were embedded in OCT compound (Tissue-Tek II; Miles, Elkhart, IN), and serial 4-μm cryosections were generated. To evaluate ST inflammation, histologic parameters were scored by a pathologist blinded to the diagnosis and clinical data, as previously described (32), after staining with hematoxylin and eosin (H&E). Briefly, the mean lining thickness was evaluated by counting the number of synovial cell layers in 6 randomly selected regions and semiquantitatively scoring the mean (0 = 1–2, 1 = 3–4, 2 = 5–6, 3 = ≥7 cell layers). Similarly, infiltration of the sublining layer with inflammatory cells was evaluated by scoring the global number of lymphocytes, plasma cells, monocyte/macrophages, and polymorphonuclear cells, and vascularity was evaluated by scoring the number of blood vessels. Both infiltration and vascularity were semiquantitatively scored on a scale ranging from 0 (lowest level) to 4 (highest level).

In immunohistochemical analyses, PAD-2 and PAD-4 were detected by using their corresponding antibody probes at 3 μg/ml, fibrin was probed with a rabbit antiserum to the β-chain of human fibrin(ogen) (1:4,000 dilution; Cambio), CD68 was stained using the PG-M1 mouse monoclonal antibody (1:200 dilution; DakoCytomation, Trappes, France), and citrullinated proteins were probed with rabbit IgG to modified citrullyl residues (a generous gift from T. Senshu, Yokohama City University, Yokohama, Japan) used at 1.6 μg/ml after in situ modification of citrullyl residues, as previously described (3). In simple immunostainings, bound rabbit antibodies were detected using a peroxidase-labeled polymer conjugated to goat anti-rabbit immunoglobulin (EnVision+ Dual Link System; DakoCytomation) followed by incubation with 3-amino-9-ethylcarbazole and hematoxylin counterstaining. In double immunostainings, bound anti-PAD antibodies were first detected using anti-rabbit immunoglobulin conjugated to a peroxidase micropolymer (ImmPRESS Kit; CliniSciences, Montrouge, France) followed by incubation with diaminobenzidine. Next, sections were probed with PG-M1, which was visualized using biotinylated goat anti-mouse immunoglobulin (DakoCytomation) followed by a mixture of avidin and biotinylated alkaline phosphatase (both from a Vectastain ABC-AP Kit; Vector, Burlingame, CA) and incubation with blue chromogen (Vector Blue; Vector). In all staining conditions, a normal rabbit antiserum was used as negative control, and only specific staining was taken into account.

Statistical analysis.

Data analyses were performed using Statistica for Windows (StatSoft, Tulsa, OK). Wilcoxon's matched pairs test was used to compare PAD-2 and PAD-4 levels in the different patient groups. Differences in the median scores obtained for PAD expression and for the evaluation of ST inflammation in the different patient groups were tested with the Mann-Whitney U test. Correlations were sought by computing Spearman's rank correlation coefficients. P values less than or equal to 0.05 were considered significant.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Expression of PAD in mononuclear leukocytes.

We set out to explore the expression of PAD-6 in blood-derived mononuclear leukocytes in comparison with expression of the other PAD isotypes. PBMCs were isolated from the blood of healthy donors and separated into a purified monocyte fraction and a lymphocyte-enriched fraction. Macrophages were obtained from monocytes, and activated macrophages were generated by stimulation with LPS. The results presented in Figure 1 are representative of those obtained for mononuclear leukocytes from 3 healthy donors. Expression of each PAD was analyzed at the mRNA level using isotype-specific primers, first by standard RT-PCR (Figure 1A). PADI1 and PADI3 mRNA could not be detected in any of the cell types and culture conditions tested. In contrast, PADI4 and PADI6 mRNA were present in the lymphocyte-enriched and monocyte fractions as well as in monocyte-derived macrophages, where their levels did not vary significantly upon LPS stimulation. PADI2 mRNA was observed clearly in the monocyte fraction and in resting macrophages and weakly in LPS-stimulated macrophages and in the lymphocyte-enriched fraction.

thumbnail image

Figure 1. Standard reverse transcription–polymerase chain reaction (RT-PCR) and immunoblot analysis of peptidyl arginine deiminase (PAD) expression in mononuclear leukocytes. A, Expression of the different PAD isotypes in peripheral blood mononuclear cells (PBMCs), a CD14-negative lymphocyte-enriched fraction (lymphocyte), CD14-positive purified monocytes (monocyte), monocyte-derived macrophages (macrophage), and lipopolysaccharide (LPS)–stimulated macrophages (macrophage-LPS) was assessed at the mRNA level by RT-PCR. Amplification of HPRT1 mRNA was used to control the evenness of cDNA input. Positive and negative control reactions included cDNA preparations of tissues or cells expressing the tested PAD isotype (results not shown; see Patients and Methods) or PCR without template (− control), respectively. Values on the right show the size of the PCR products. B–D, Expression of PAD was also assessed at the protein level. B, The specificity of the anti–PAD-6 antibodies was checked by immunoblotting equal amounts of recombinant (r) forms of human PAD-1, PAD-3, PAD-4, and PAD-6, and PAD-2 purified (p) from rabbit skeletal muscle. C, Equalization of the sensitivities of the anti–PAD types 1–4 and anti–PAD-6 antibodies was checked on the immunoblots of 0.8–13.5 fmoles of rPAD-1, pPAD-2, rPAD-3, rPAD-4, and rPAD-6, respectively. D, Proteins from the low-salt extracts of the different mononuclear leukocyte fractions were immunodetected with the anti-PAD antibodies (PAD types 1–4 and PAD-6), used at these adjusted sensitivities in parallel to relevant control tissue or cell extracts (control extract; see Patients and Methods). The positions of molecular mass standards (B) and apparent molecular mass of the different PAD isotypes (C and D) are shown on the right.

Download figure to PowerPoint

To further explore variations in the PADI mRNA levels between the different mononuclear cell populations, real-time RT-PCR analysis was performed using leukocytes from a fourth healthy donor (Table 1). The results confirmed the absence of PADI1 and PADI3 and the presence of PADI2, PADI4, and PADI6 transcripts. The results also confirmed variations in the relative expression of PADI2 between the different cell populations, with monocytes exhibiting 10-fold higher amounts of transcripts than LPS-stimulated macrophages or lymphocytes. In addition, elevated Ct values (Ct >32) strongly suggested very low expression levels of PADI4 in resting or LPS-stimulated macrophages and of PADI6 in all cell fractions.

Table 1. Real-time RT-PCR analysis of PADI expression in blood-derived cells*
GeneCell type
  • *

    Values are the theshold cycle (Ct) (relative expression). In the different cell fractions, the relative expression of each PADI gene was calculated in reference to the monocyte fraction, using the formula 2math image, where ΔCt represents the difference between the Ct value of the PADI gene and the Ct value of the reference gene (HPRT1), and ΔΔCt represents the difference between the ΔCt value obtained with the indicated cell fraction and the ΔCt value obtained with the monocyte fraction. RT-PCR = reverse transcription–polymerase chain reaction; PBMCs = peripheral blood mononuclear cells; LPS = lipopolysaccharide; ND = not detectable (no amplification occurred, because no fluorescence was detected after 40 cycles); NA = not applicable (the relative expression was not calculated, because only very low amplification occurred, preventing accurate determination of the Ct value).

PADI227 (0.4)28 (0.1)24 (1.0)26 (0.6)29 (0.1)
PADI427 (0.4)28 (0.1)24 (1.0)>32 (NA)>32 (NA)
PADI6>32 (NA)>32 (NA)>32 (NA)>32 (NA)>32 (NA)

Expression of PAD was then analyzed at the protein level, using previously produced antibodies specific for PAD-1, PAD-2, PAD-3, and PAD-4. In addition, antibodies specific for PAD-6 were generated, and their specificity was verified by immunoblotting on recombinant human PAD-6. The apparent molecular mass of the band recognized by the anti–PAD-6 antibodies (∼78 kd) corresponds to that expected for recombinant PAD-6 (Figure 1B). Moreover, these antibodies did not cross-react with rabbit muscle PAD-2 or with human recombinant PAD-1, PAD-3, and PAD-4 (Figure 1B). Each set of anti-PAD antibodies was used at a concentration allowing detection of at least ∼1 fmole of its PAD target (Figure 1C).

In the different blood-derived cell fractions, PAD expression was explored by immunoblotting equal amounts of low-salt–extracted proteins separated by SDS-PAGE (Figure 1D). The obtained results were very consistent with the PCR results described above. Indeed, PAD-1 and PAD-3 were not detected in any of the cell fractions, PAD-2 was found in all cell fractions, and monocytes exhibited the highest expression levels. PAD-4 was detected in the lymphocyte-enriched and purified monocyte fractions but not in cells where corresponding mRNA was rare, i.e., in resting or LPS-stimulated macrophages. Similarly, PAD-6 was not detected in any of the cell fractions, even though, in simultaneously probed extracts of human adult ovary, the anti–PAD-6 antibodies stained an ∼78-kd band (control extract, Figure 1D).

Expression of PAD in ST.

The presence of the different PAD isotypes was then assessed in ST samples from 16 patients with RA. These were compared with ST from control patients, corresponding to 4 patients with non-RA arthritides and 7 patients with OA. PAD expression was first examined at the protein level by immunoblot analysis of equal amounts of proteins from low-salt extracts of the ST samples, using the anti-PAD antibodies at their adjusted concentrations (Figure 2A). PAD-1, PAD-3, and PAD-6 were detected in neither the ST of patients with RA nor that of control patients. In contrast, PAD-2 was present in all 16 patients with RA, all 4 control patients with other arthritides, and in 6 of the 7 patients with OA. PAD-4 was also expressed both in patients with RA and in control patients. However, it was detected in 13 of 16 patients with RA, all 4 patients with other arthritides, but in only 2 of the 7 patients with OA. The presence of mRNA transcripts corresponding to the 5 PAD isotypes was also explored by RT-PCR in a subset of patients corresponding to 3 patients with RA (patients RA3, RA7, and RA15) and 2 control patients (patients AS1 and OA4) (Figure 2B). This analysis showed the absence of expression of PADI1 and PADI3, and expression of PADI2 and PADI4 in the ST of all patients with RA and all control patients. PADI6 transcripts were detected in 4 of the 5 ST samples (in all samples except that from patient RA15).

thumbnail image

Figure 2. Immunoblot and RT-PCR analyses of PAD expression in synovial tissue (ST) samples from patients with rheumatoid arthritis (RA) or other arthritides. A, For immunoblot analysis, proteins from low-salt extracts of the ST samples were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and electrotransferred onto membranes. Total proteins were stained with ponceau red and then probed with the different indicated anti-PAD antibodies at their adjusted concentrations. Positive controls included rPAD-1, rPAD-3, rPAD-4, rPAD-6, or pPAD-2, detected by the relevant corresponding anti-PAD antibodies. Values on the right represent the apparent molecular mass of each PAD. B, The presence of mRNA transcripts of the 5 PADI genes was evaluated by RT-PCR in ST samples from 3 patients with RA and 2 control patients. Amplification of G3PDH mRNA was used to control the evenness of cDNA input. Positive and negative control experiments were performed as indicated in Figure 1A. Values on the right show the size of the PCR products. AS = ankylosing spondylitis; uSpA = undifferentiated spondylarthritis; PsA = psoriatic arthritis; OA = osteoarthritis (see Figure 1 for other definitions).

Download figure to PowerPoint

Search for correlations between PAD expression and ST inflammation.

The levels of PAD-2 and PAD-4 were semiquantitatively estimated by scoring the labeling intensity of the corresponding bands on immunoblots (Table 2). No statistically significant differences were observed between the levels of PAD-2 and those of PAD-4, in either patients with RA or patients with other arthritides, although a trend toward higher levels of PAD-4 was observed. In contrast, in patients with OA, the scores obtained for PAD-2 were significantly higher than those for PAD-4 (P = 0.028). Accordingly, whereas the median level of PAD-2 expression did not significantly differ between the various groups of patients, that of PAD-4 was significantly higher in patients with arthritides compared with patients with OA (P = 0.0055).

Table 2. PAD-2 and PAD-4 expression and synovial tissue (ST) inflammation*
Patient codePAD-2PAD-4VascularityInfiltrationMean lining thickness
  • *

    The levels of peptidyl arginine deiminase (PAD) expression were evaluated by semiquantitative scoring of the labeling intensity of the PAD bands on immunoblots of ST extracts. ST inflammation was evaluated by semiquantitative scoring of the vascularity, infiltration of the sublining layer by inflammatory cells, and mean synovial lining thickness. See Patients and Methods for details. RA = rheumatoid arthritis; AS = ankylosing spondylitis; uSpA = undifferentiated spondylarthritis; PsA = psoriatic arthritis; OA = osteoarthritis.


The intensity of inflammatory lesions in all ST samples was evaluated histologically on H&E-stained sections, by scoring the mean lining thickness, vascularity, and infiltration by cells of hematologic origin (Table 2). This analysis indicated that not all of the synovial membranes from patients with arthritides were actively inflamed, and that 1 ST sample from a patient with OA (OA1) was obtained during a phase of high inflammation. Globally, however, as expected, ST samples obtained from the group of patients with arthritides tended to have higher inflammation scores than those from patients with OA, although the differences did not reach statistical significance.

In Figure 2A, the lanes corresponding to ST samples from the 3 groups of patients (RA, non-RA arthritides, and OA) are classified according to decreasing density of inflammatory infiltration. Interestingly, the labeling intensities of the bands corresponding to PAD-2 and PAD-4 tended to decrease in parallel, showing a correlation between the expression levels of PAD-2 or PAD-4 and the density of ST cell infiltration. Furthermore, expression of PAD-2 and PAD-4 was significantly correlated not only with the density of infiltration by inflammatory cells but also with the vascularity of the deep synovium and the synovial lining thickness, when considering either all patients, only those with arthritides (including RA), or only patients with RA (Table 3). Interestingly, however, in patients with arthritides, PAD-2 was more significantly correlated with the density of ST infiltration than with the synovial lining thickness (P = 0.00023 versus P = 0.045), while the opposite was observed for PAD-4 (P = 0.02 versus P = 0.0053). Remarkably, in the group of patients with OA, the level of PAD-2 expression was correlated only with the density of ST infiltration (Table 3). Finally, in the group of patients with RA, no relationships were observed between the expression levels of PAD-2 or PAD-4 and the serum titers of ACPAs.

Table 3. Correlations between PAD-2 and PAD-4 expression and ST inflammation*
Patient groupMean lining thicknessVascularityInfiltration
  • *

    Values are Spearman's rank correlation coefficients (P values). P values less than or equal to 0.05 were considered significant. PAD-2 = peptidyl arginine deiminase type 2; ST = synovial tissue; NS = not significant.

All patients   
 PAD-20.43 (0.025)0.53 (0.0048)0.67 (0.00015)
 PAD-40.47 (0.014)0.47 (0.013)0.55 (0.0029)
Rheumatoid arthritis   
 PAD-20.59 (0.015)0.68 (0.0038)0.77 (0.00043)
 PAD-40.61 (0.012)0.70 (0.0023)0.54 (0.032)
All arthritides   
 PAD-20.45 (0.045)0.66 (0.0017)0.73 (0.00023)
 PAD-40.60 (0.0053)0.59 (0.0063)0.51 (0.02)
 PAD-2NSNS0.80 (0.032)

Immunolocalization of PAD-2 and PAD-4 in the ST of patients with RA.

Finally, we sought to specify the location of PAD-2 and PAD-4 in the ST of patients with RA (Figure 3). Simple immunoperoxidase stainings were performed on serial sections of ST obtained from a subset of 10 patients with RA. Figure 3A illustrates the results obtained in ST samples from 4 patients with RA, in areas where fibrin and citrullinated proteins were detected. Depending on the sample examined, fibrin was evidenced as compact, well-defined deposits (patient RA2) or as diffuse deposits in the deep synovium (patient RA3). In other cases, fibrin deposits were observed in the synovial lining layer and appeared as either well defined (patient RA6) or diffuse (patient RA9).

thumbnail image

Figure 3. Immunolocalization of peptidyl arginine deiminase type 2 (PAD-2) and PAD-4 in the synovial tissue of patients with rheumatoid arthritis (RA). A and B, In simple immunoperoxidase stainings, fibrin, citrullinated proteins, PAD-2, and PAD-4 were localized on adjacent sections. PAD-2 and PAD-4 are shown in areas containing citrullinated fibrin (A) and in fibrin-free areas (B). C, In double immunostainings, CD68 (blue) and PAD-2 or PAD-4 (brown) were simultaneously localized. Large boxed areas show higher-magnification views of small boxed areas. Bars in A and B = 50 μm; bar in C = 10 μm.

Download figure to PowerPoint

The antibody to citrullinated proteins revealed extensive citrullination of the detected fibrin (patients RA2, RA3, and RA9) or of intracellular proteins in cells located within the fibrin deposits (patient RA6). PAD-2 was observed extracellularly and in the cytoplasm of cells within the citrullinated fibrin areas (see patient RA2 for an example of both locations). It was also found in the cytoplasm of palisadic histiocytes encircling amorphous deposits rich in citrullinated proteins (patient RA3). PAD-4 was observed extracellularly in some fibrin-rich areas (patient RA6) and in the cytoplasm and/or nucleus of cells within (patient RA2) or in the close vicinity of (patients RA6 and 9) fibrin deposits. More generally, the presence of PAD-2 and PAD-4 in the ST was observed in 6 and 10 patients with RA, respectively. The observation of zones simultaneously containing PAD-2– and PAD-4–positive cells was quite uncommon, suggesting the rarity of cells simultaneously expressing both enzymes. PAD-2 staining was cytoplasmic and generally detected in mononuclear cells scattered in the sublining or deep synovium and rarely in synovial lining cells, while PAD-4 staining involved the cytoplasm and/or nucleus of more numerous and generally grouped mononuclear cells located in all tissue areas, and synovial lining cells (Figure 3B). The morphology of both PAD-2– and PAD-4–positive cells frequently evoked that of macrophages, and, indeed, double stainings with the antibodies to PAD-2 or PAD-4 and an anti-CD68 antibody showed that most PAD-expressing cells were CD68 positive and therefore of myelomonocytic origin (Figure 3C).


  1. Top of page
  2. Abstract
  6. Acknowledgements

In the first part of our study, the fact that none of the mononuclear leukocyte populations we examined expressed PADI1 and PADI3 fits with the reported absence of the corresponding mRNA transcripts in peripheral blood leukocytes (8, 31). It also was consistent with a recent study by Vossenaar et al, in which the authors reported the absence of mRNA for PADI1 and PADI3 in PB-derived T cells (CD3+ PBMCs), B cells (CD19+ PBMCs), monocytes (CD14+ PBMCs), and natural killer cells (CD56+ PBMCs) from 1 blood donor (14). Both studies are also in agreement concerning the presence of PADI2 mRNA in lymphocytes, monocytes, and monocyte-derived macrophages. However, at the protein level, we found PAD-2 in monocytes and macrophages, while it was detected in macrophages but not in monocytes by the other authors (14). This discrepancy may originate from differences in the methods used to obtain blood monocytes, because we purified monocytes using positive selection of CD14-positive PBMCs, while Vossenaar et al isolated monocytes from PBMCs by plastic adherence. Another discrepancy concerns PAD-4, which we could not detect in macrophages, while Vossenaar et al made an opposite observation. Again, this could originate from methodologic differences: to obtain monocyte-derived macrophages, we used a 7-day culture of the CD14-positive PBMCs in the presence of M-CSF and prevented adherence, while in the other study, macrophages were obtained by a 7-day culture of adherent PBMCs.

Concerning PAD-6, our analysis of its protein expression using specially developed and validated antibodies is original. The apparent molecular mass of the band that these antibodies detected in a protein extract of human adult ovary corresponds to that predicted for human PAD-6. Therefore, this protein is probably present at least in ovaries, similar to its mouse ortholog, ePAD (33, 34). In addition, we confirm previous reports mentioning detection of transcripts of PADI6 in peripheral blood leukocytes (7, 35) and, furthermore, specify their presence in several types of mononuclear leukocytes. However, real-time RT-PCR analysis showed that the expression level of PADI6 is very low, and this probably accounts for the fact that the corresponding protein could not be detected. Therefore, and on the whole, the results concerning expression of PAD in different mononuclear leukocyte populations indicate that only PAD-2 and PAD-4 reach detectable expression at the protein level in cells of lymphocyte and monocyte lineages.

The second part of our study constitutes the first systematic exploration of synovial expression of the 5 PAD isotypes in a large series of patients with RA, non-RA arthritides, or OA. We clearly demonstrate the absence of PAD-1 and PAD-3 in all disease types. Moreover, we show that PAD-2 is consistently expressed in the course of both arthritides and OA, while expression of PAD-4 is mainly associated with arthritides. Finally, no PAD-6 protein was found in any of the disease groups.

Concerning PAD-2 and PAD-4, our results can be compared with those of other less systematic studies analyzing their presence in human rheumatoid ST. The research group led by K. Yamamoto mentioned the presence of PAD-4 in the sublining region of ST from patients with RA (36, 37). However, in a much more detailed confocal microscopy analysis using the same anti–PAD-4 antibody, the presence of PAD-4 was detected in all areas of ST samples from patients with RA, in cells that were identified as T lymphocytes, B lymphocytes, macrophages, granulocytes, fibroblasts, and endothelial cells (19). The same group of investigators also reported immunohistologic detection of PAD-2 in the lining, sublining, and deep regions of 1 ST sample from a patient with RA (36).

In ST samples obtained from a series of patients including 19 patients with RA and 19 control patients with inflammatory or noninflammatory rheumatism, De Rycke et al detected PAD-2 in 59% of patients with RA but also in 17% of control patients (38). This is consistent with our observations, even though our immunoblot analysis with optimized antibody concentrations appears to be more sensitive (0.5–1 fmole), because it allowed us to detect the presence of PAD-2 in all RA and almost all control ST samples. Additionally, PAD-2 and PAD-4 have been explored in the ST of mice with collagen-induced arthritis (CIA) or streptococcal cell wall–induced arthritis, with identical results in both models (39). Transcripts for Padi2 were detected in naive and arthritic mice, while transcripts for Padi4 were detected only in arthritic mice. Both corresponding PAD-2 and PAD-4 proteins were absent in the ST of naive mice and, in the inflamed ST of arthritic mice, only PAD-4 was detected and reported to be expressed by neutrophils (39). PAD-4 has also been observed in the inflamed ST of dark agouti rats with CIA, but, in this case, mononuclear cells were found to be responsible for its expression (40).

An important point of our study is that it clearly shows that PAD expression in the ST is not specific for RA, which is consistent with our finding that citrullination of fibrin in the ST occurs during a variety of synovitides (41), and that citrullinated proteins are present in several other inflamed tissues, such as the muscle, the colon, or the lung of patients with polymyositis, inflammatory bowel disease, or interstitial pneumonia, respectively (42, 43). Similarly, in patients with RA, we could not demonstrate a relationship between the serum ACPA titer and the level of PAD-2 or PAD-4 in the ST, which is consistent with the absence of correlation between the serum ACPA titer and the amount of citrullinated fibrin in the ST (41). This emphasizes that, in the ST, PAD expression and citrullination of target proteins are not specifically associated with RA, in contrast to our findings with ACPAs.

We actually show that PAD expression in the ST is an inflammation-dependent phenomenon. A close correlation is observed in arthritides and OA between the level of PAD-2 and the degree of infiltration by inflammatory cells, suggesting basal expression of this isotype by these cells in these 2 disease types. Because PAD-4 is also correlated with the degree of infiltration, inflammatory cells also constitute a source for this isotype. However, PAD-4 is confined to the arthritides. Differences in the composition of the infiltrate, but also in the state of differentiation and/or activation of infiltrating cells, probably influenced by differences in their cytokine environment, may account for PAD-4 expression in arthritides but not OA. Moreover, PAD-4 is frequently observed in the lining layer, and its expression level is particularly correlated with the synovial lining thickness. This suggests that the hyperplastic cells of the synovial lining constitute another important source of PAD-4 in these diseases. It would be interesting to evaluate whether the proportion of PAD-4–positive cells exhibiting nuclear staining is related to the concentration of tumor necrosis factor α in the ST, because it was recently shown that this cytokine induces nuclear translocation of this PAD isotype in murine and human oligodendroglial cell lines (44).

Studies of PAD polymorphisms have shown that a haplotype of PADI4 was associated with RA in Japanese (37, 45) and Korean populations (46). These results were confirmed in a Caucasian population from North America (47) but were not reproduced in various European Caucasian populations (47–51). In the initial study by Suzuki et al (37), in vitro results indicated that the presence of this haplotype could lead to more stable mRNA, suggesting that PAD-4 expression was increased in the ST of patients with RA, leading to enhanced levels of citrullinated proteins. However, so far the question of whether the levels of PAD-4 expression are related to the PADI4 haplotype has not been explored. Given the association between PAD-4 expression and ST inflammation, it would be interesting to evaluate the influence of PADI4 haplotypes in groups of patients stratified according to the degree of ST inflammation.

Owing to the use of anti-PAD antibodies with equalized sensitivities of detection, we could observe that, in the ST of patients with RA, the levels of PAD-4 detected by immunoblotting tended to be higher than those of PAD-2. This is corroborated by the observation of higher numbers of PAD-4–positive cells than PAD-2–positive cells by immunohistochemical analysis. However, our results strongly indicate that both PAD-2 and PAD-4 are involved in the process of fibrin citrullination, because even if their simultaneous detection in the same area was quite rare, both were observed directly or in the close vicinity of fibrin deposits in the ST. This is compatible with the observation that, in vitro, recombinant forms of both human PAD-2 and human PAD-4 are able to citrullinate human fibrinogen (36). Moreover, the presence of citrullinated fibrin in samples of inflamed ST from patients with OA (41), in which PAD-4 is rarely seen (ref.19, and the present study), is necessarily attributable to a PAD-2 activity.

In conclusion, the present study has provided the first extensive exploration of all existing PAD enzymes in the ST of patients with RA. Even if it remains to be precisely determined which factors induce their production and how/why they are released and activated to be able to target extracellular proteins such as fibrin, most probably both PAD-2 and PAD-4 play a role in the generation of synovial ACPA targets. In this respect, inhibition of the activity or expression of PAD-2 and/or PAD-4 may constitute a valuable therapeutic strategy for RA.


  1. Top of page
  2. Abstract
  6. Acknowledgements

We thank Prof. B. Fournié, Dr. L. Zabraniecki, and Dr. O. Lemaire (Rheumatology Department, Hôpital Purpan, Toulouse), and Prof. A. Cantagrel and Dr. A. Constantin (Rheumatology Department, Hôpital Rangueil, Toulouse) for providing patient data and sera. We also thank Prof. M. Mansat, Dr. M. Rongières, Prof. P. Mansat, and Prof. P. Bonnevialle (Department of Orthopaedic and Traumatic Surgery, Hôpital Purpan) for providing samples of synovial membrane, Dr. A. Gomez-Brouchet (Pathology Department, Hôpital Rangueil) for providing access to her diagnostic pathologic analysis of the synovial membranes, Prof. J.-P. Chavoin (Plastic Surgery Department, Hôpital Rangueil) for providing samples of human skin, and Ms F. Capilla (Experimental Pathology Platform of Institut Fédératif de Recherche 30) for her assistance in histologic analyses. The skillful technical assistance of Ms R. Llobera, M.-T. Ribouchon, M.-P. Henry, and M.-F. Isaïa is gratefully acknowledged.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Dr. Serre had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Sebbag, Serre.

Acquisition of data. Foulquier, Sebbag, Clavel, Chapuy-Regaud, Al Badine.

Analysis and interpretation of data. Foulquier, Sebbag, Chapuy-Regaud, Al Badine, Guerrin.

Manuscript preparation. Foulquier, Sebbag, Chapuy-Regaud, Méchin, Simon, Guerrin, Serre.

Statistical analysis. Vincent.

Provision of indispensable reagents. Clavel, Méchin, Nachat, Yamada, Takahara.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  • 1
    Vincent C, Nogueira L, Clavel C, Sebbag M, Serre G. Autoantibodies to citrullinated proteins: ACPA [review]. Autoimmunity 2005; 38: 1724.
  • 2
    Masson-Bessiere C, Sebbag M, Durieux JJ, Nogueira L, Vincent C, Girbal-Neuhauser E, et al. In the rheumatoid pannus, anti-filaggrin autoantibodies are produced by local plasma cells and constitute a higher proportion of IgG than in synovial fluid and serum. Clin Exp Immunol 2000; 119: 54452.
  • 3
    Masson-Bessiere C, Sebbag M, Girbal-Neuhauser E, Nogueira L, Vincent C, Senshu T, et al. The major synovial targets of the rheumatoid arthritis-specific antifilaggrin autoantibodies are deiminated forms of the α- and β-chains of fibrin. J Immunol 2001; 166: 417784.
  • 4
    Schellekens GA, de Jong BA, van den Hoogen FH, van de Putte LB, van Venrooij WJ. Citrulline is an essential constituent of antigenic determinants recognized by rheumatoid arthritis-specific autoantibodies. J Clin Invest 1998; 101: 27381.
  • 5
    Girbal-Neuhauser E, Durieux JJ, Arnaud M, Dalbon P, Sebbag M, Vincent C, et al. The epitopes targeted by the rheumatoid arthritis-associated antifilaggrin autoantibodies are posttranslationally generated on various sites of (pro)filaggrin by deimination of arginine residues. J Immunol 1999; 162: 58594.
  • 6
    Sebbag M, Moinard N, Auger I, Clavel C, Arnaud J, Nogueira L, et al. Epitopes of human fibrin recognized by the rheumatoid arthritis-specific autoantibodies to citrullinated proteins. Eur J Immunol 2006; 36: 225063.
  • 7
    Chavanas S, Mechin MC, Takahara H, Kawada A, Nachat R, Serre G, et al. Comparative analysis of the mouse and human peptidylarginine deiminase gene clusters reveals highly conserved non-coding segments and a new human gene, PADI6. Gene 2004; 330: 1927.
  • 8
    Nachat R, Mechin MC, Takahara H, Chavanas S, Charveron M, Serre G, et al. Peptidylarginine deiminase isoforms 1-3 are expressed in the epidermis and involved in the deimination of K1 and filaggrin. J Invest Dermatol 2005; 124: 38493.
  • 9
    Nachat R, Mechin MC, Charveron M, Serre G, Constans J, Simon M. Peptidylarginine deiminase isoforms are differentially expressed in the anagen hair follicles and other human skin appendages. J Invest Dermatol 2005; 125: 3441.
  • 10
    Mechin MC, Enji M, Nachat R, Chavanas S, Charveron M, Ishida-Yamamoto A, et al. The peptidylarginine deiminases expressed in human epidermis differ in their substrate specificities and subcellular locations. Cell Mol Life Sci 2005; 62: 198495.
  • 11
    Sambandam T, Belousova M, Accaviti-Loper MA, Blanquicett C, Guercello V, Raijmakers R, et al. Increased peptidylarginine deiminase type II in hypoxic astrocytes. Biochem Biophys Res Commun 2004; 325: 13249.
  • 12
    Ishigami A, Ohsawa T, Hiratsuka M, Taguchi H, Kobayashi S, Saito Y, et al. Abnormal accumulation of citrullinated proteins catalyzed by peptidylarginine deiminase in hippocampal extracts from patients with Alzheimer's disease. J Neurosci Res 2005; 80: 1208.
  • 13
    Urano Y, Watanabe K, Sakaki A, Arase S, Watanabe Y, Shigemi F, et al. Immunohistochemical demonstration of peptidylarginine deiminase in human sweat glands. Am J Dermatopathol 1990; 12: 24955.
  • 14
    Vossenaar ER, Radstake TR, van der Heijden A, van Mansum MA, Dieteren C, de Rooij DJ, et al. Expression and activity of citrullinating peptidylarginine deiminase enzymes in monocytes and macrophages. Ann Rheum Dis 2004; 63: 37381.
  • 15
    Ishigami A, Ohsawa T, Asaga H, Akiyama K, Kuramoto M, Maruyama N. Human peptidylarginine deiminase type II: molecular cloning, gene organization, and expression in human skin. Arch Biochem Biophys 2002; 407: 2531.
  • 16
    Kanno T, Kawada A, Yamanouchi J, Yosida-Noro C, Yoshiki A, Shiraiwa M, et al. Human peptidylarginine deiminase type III: molecular cloning and nucleotide sequence of the cDNA, properties of the recombinant enzyme, and immunohistochemical localization in human skin. J Invest Dermatol 2000; 115: 81323.
  • 17
    Nakashima K, Hagiwara T, Ishigami A, Nagata S, Asaga H, Kuramoto M, et al. Molecular characterization of peptidylarginine deiminase in HL-60 cells induced by retinoic acid and 1α,25-dihydroxyvitamin D3. J Biol Chem 1999; 274: 2778692.
  • 18
    Asaga H, Nakashima K, Senshu T, Ishigami A, Yamada M. Immunocytochemical localization of peptidylarginine deiminase in human eosinophils and neutrophils. J Leukoc Biol 2001; 70: 4651.
  • 19
    Chang X, Yamada R, Suzuki A, Sawada T, Yoshino S, Tokuhiro S, et al. Localization of peptidylarginine deiminase 4 (PADI4) and citrullinated protein in synovial tissue of rheumatoid arthritis. Rheumatology (Oxford) 2005; 44: 4050.
  • 20
    Nakashima K, Hagiwara T, Yamada M. Nuclear localization of peptidylarginine deiminase V and histone deimination in granulocytes. J Biol Chem 2002; 277: 495628.
  • 21
    Tarcsa E, Marekov LN, Mei G, Melino G, Lee SC, Steinert PM. Protein unfolding by peptidylarginine deiminase: substrate specificity and structural relationships of the natural substrates trichohyalin and filaggrin. J Biol Chem 1996; 271: 3070916.
  • 22
    Senshu T, Kan S, Ogawa H, Manabe M, Asaga H. Preferential deimination of keratin K1 and filaggrin during the terminal differentiation of human epidermis. Biochem Biophys Res Commun 1996; 225: 7129.
  • 23
    Moscarello MA, Wood DD, Ackerley C, Boulias C. Myelin in multiple sclerosis is developmentally immature. J Clin Invest 1994; 94: 14654.
  • 24
    Wang Y, Wysocka J, Sayegh J, Lee YH, Perlin JR, Leonelli L, et al. Human PAD4 regulates histone arginine methylation levels via demethylimination. Science 2004; 306: 27983.
  • 25
    Cuthbert GL, Daujat S, Snowden AW, Erdjument-Bromage H, Hagiwara T, Yamada M, et al. Histone deimination antagonizes arginine methylation. Cell 2004; 118: 54553.
  • 26
    Mechin MC, Sebbag M, Arnaud J, Nachat R, Foulquier C, Adoue V, et al. Update on peptidylarginine deiminases and deimination in skin physiology and severe human diseases [review]. Int J Cosmet Sci 2007; 29: 14768.
  • 27
    Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, et al. Development of criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee. Arthritis Rheum 1986; 29: 103949.
  • 28
    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 1988; 31: 31524.
  • 29
    Dougados M, van der Linden S, Juhlin R, Huitfeldt B, Amor B, Calin A, et al, for the European Spondylarthropathy Study Group. The European Spondylarthropathy Study Group preliminary criteria for the classification of spondylarthropathy. Arthritis Rheum 1991; 34: 121827.
  • 30
    Chapuy-Regaud S, Nogueira L, Clavel C, Sebbag M, Vincent C, Serre G. IgG subclass distribution of the rheumatoid arthritis-specific autoantibodies to citrullinated fibrin. Clin Exp Immunol 2005; 139: 54250.
  • 31
    Guerrin M, Ishigami A, Mechin MC, Nachat R, Valmary S, Sebbag M, et al. cDNA cloning, gene organization and expression analysis of human peptidylarginine deiminase type I. Biochem J 2003; 370(Pt 1): 16774.
  • 32
    Baeten D, Demetter P, Cuvelier C, Van den Bosch F, Kruithof E, Van Damme N, et al. Comparative study of the synovial histology in rheumatoid arthritis, spondyloarthropathy, and osteoarthritis: influence of disease duration and activity. Ann Rheum Dis 2000; 59: 94553.
  • 33
    Wright PW, Bolling LC, Calvert ME, Sarmento OF, Berkeley EV, Shea MC, et al. ePAD, an oocyte and early embryo-abundant peptidylarginine deiminase-like protein that localizes to egg cytoplasmic sheets. Dev Biol 2003; 256: 7388.
  • 34
    Liu M, Oh A, Calarco P, Yamada M, Coonrod SA, Talbot P. Peptidylarginine deiminase (PAD) is a mouse cortical granule protein that plays a role in preimplantation embryonic development. Reprod Biol Endocrinol 2005; 3: 42.
  • 35
    Zhang J, Dai J, Zhao E, Lin Y, Zeng L, Chen J, et al. cDNA cloning, gene organization and expression analysis of human peptidylarginine deiminase type VI. Acta Biochim Pol 2004; 51: 10518.
  • 36
    Nakayama-Hamada M, Suzuki A, Kubota K, Takazawa T, Ohsaka M, Kawaida R, et al. Comparison of enzymatic properties between hPADI2 and hPADI4. Biochem Biophys Res Commun 2005; 327: 192200.
  • 37
    Suzuki A, Yamada R, Chang X, Tokuhiro S, Sawada T, Suzuki M, et al. Functional haplotypes of PADI4, encoding citrullinating enzyme peptidylarginine deiminase 4, are associated with rheumatoid arthritis. Nat Genet 2003; 34: 395402.
  • 38
    De Rycke L, Nicholas AP, Cantaert T, Kruithof E, Echols JD, Vandekerckhove B, et al. Synovial intracellular citrullinated proteins colocalizing with peptidyl arginine deiminase as pathophysiologically relevant antigenic determinants of rheumatoid arthritis–specific humoral autoimmunity. Arthritis Rheum 2005; 52: 232330.
  • 39
    Vossenaar ER, Nijenhuis S, Helsen MM, van der Heijden A, Senshu T, van den Berg WB, et al. Citrullination of synovial proteins in murine models of rheumatoid arthritis. Arthritis Rheum 2003; 48: 2489500.
  • 40
    Lundberg K, Nijenhuis S, Vossenaar ER, Palmblad K, van Venrooij WJ, Klareskog L, et al. Citrullinated proteins have increased immunogenicity and arthritogenicity and their presence in arthritic joints correlates with disease severity. Arthritis Res Ther 2005; 7: R45867.
  • 41
    Chapuy-Regaud S, Sebbag M, Baeten D, Clavel C, Foulquier C, De Keyser F, et al. Fibrin deimination in synovial tissue is not specific for rheumatoid arthritis but commonly occurs during synovitides. J Immunol 2005; 174: 505764.
  • 42
    Makrygiannakis D, af Klint E, Lundberg IE, Lofberg R, Ulfgren AK, Klareskog L, et al. Citrullination is an inflammation-dependent process. Ann Rheum Dis 2006; 65: 121922.
  • 43
    Bongartz T, Cantaert T, Atkins SR, Harle P, Myers JL, Turesson C, et al. Citrullination in extra-articular manifestations of rheumatoid arthritis. Rheumatology (Oxford) 2007; 46: 705.
  • 44
    Mastronardi FG, Wood DD, Mei J, Raijmakers R, Tseveleki V, Dosch HM, et al. Increased citrullination of histone H3 in multiple sclerosis brain and animal models of demyelination: a role for tumor necrosis factor-induced peptidylarginine deiminase 4 translocation. J Neurosci 2006; 26: 1138796.
  • 45
    Ikari K, Kuwahara M, Nakamura T, Momohara S, Hara M, Yamanaka H, et al. Association between PADI4 and rheumatoid arthritis: a replication study. Arthritis Rheum 2005; 52: 30547.
  • 46
    Kang CP, Lee HS, Ju H, Cho H, Kang C, Bae SC. A functional haplotype of the PADI4 gene associated with increased rheumatoid arthritis susceptibility in Koreans. Arthritis Rheum 2006; 54: 906.
  • 47
    Plenge RM, Padyukov L, Remmers EF, Purcell S, Lee AT, Karlson EW, et al. Replication of putative candidate-gene associations with rheumatoid arthritis in >4,000 samples from North America and Sweden: association of susceptibility with PTPN22, CTLA4, and PADI4. Am J Hum Genet 2005; 77: 104460.
  • 48
    Barton A, Bowes J, Eyre S, Spreckley K, Hinks A, John S, et al. A functional haplotype of the PADI4 gene associated with rheumatoid arthritis in a Japanese population is not associated in a United Kingdom population. Arthritis Rheum 2004; 50: 111721.
  • 49
    Caponi L, Petit-Teixeira E, Sebbag M, Bongiorni F, Moscato S, Pratesi F, et al. A family based study shows no association between rheumatoid arthritis and the PADI4 gene in a white French population. Ann Rheum Dis 2005; 64: 58793.
  • 50
    Harney SM, Meisel C, Sims AM, Woon PY, Wordsworth BP, Brown MA. Genetic and genomic studies of PADI4 in rheumatoid arthritis. Rheumatology (Oxford) 2005; 44: 86972.
  • 51
    Martinez A, Valdivia A, Pascual-Salcedo D, Lamas JR, Fernandez-Arquero M, Balsa A, et al. PADI4 polymorphisms are not associated with rheumatoid arthritis in the Spanish population. Rheumatology (Oxford) 2005; 44: 12636.