Many citrullinated proteins have been identified in the inflamed joints of RA patients and a number of these have been shown react in vitro
with ACPA. A prominent candidate to be the main target for ACPA is fibrin(ogen), which is known to be present in large amounts in inflamed joints and which is known to be present in a citrullinated form in the synovial fluid of RA patients [3
]. ACPA can indeed bind to citrullinated fibrinogen, in particular the alpha chain. Recently, it was shown that in inflamed joints also significant amounts of peptides can be found that are derived from fibrinogen, due to endogenous proteolytic activity [15
]. Here, we showed that citrullinated versions of these peptides can be found in the synovial fluid of inflamed joints. While some of these citrullinated peptides were equally abundant in the RA patients and the control group with patients suffering from other inflammatory joint diseases, others appeared to be more abundant in RA patients (Table ). Consistent with that observation, PAD activity in joints is known to be elevated in such other inflammatory diseases as well. The higher level of some of the peptides in RA patients suggests that a different immune response toward these peptides might occur in RA patients, leading to, for example, prolonged retention of these peptides, which, in turn, could contribute to increased immune system activation. Although there was significant diversity in the levels of citrullinated peptides in the control group, with especially patient S2 (ankylosing spondylitis) showing very high levels, no obvious relation with their diagnosed disease was apparent (Additional file 1
, Table S1). This diversity among the patients in the control group confirms that the citrullinated peptides arose from inflammatory events that can also occur in diseases other than RA, which is not unexpected as many rheumatic diseases show overlap in the symptoms observed. An ideal control material would be synovial fluid from healthy individuals, but for obvious reasons this material is less readily available, not only due to ethical issues, but also because the amount of synovial fluid in noninflamed joints is much smaller.
The citrullinated peptides we found to be most abundantly present in RA patients covered mainly two distinct regions in the alpha chain of fibrinogen, amino acids 20-35 and 414-433. These regions have been shown to be susceptible to citrullination in the past and it was shown that the arginine residues in these sequences (R35, R425 and R426) can be citrullinated by both the PAD2 as well as the PAD4 enzymes [22
], both known to be present in inflammatory cells in the synovium [23
]. Together, these PAD enzymes are known to be able to citrullinate two-thirds of all arginine residues in fibrinogen [22
], but little is known about which citrullinated residues contribute most to eliciting the anti-citrullinated protein immune response. The fact that specific citrullinated fibrinogen-derived peptides reside in the synovial fluid makes them prime candidates for having a role in the immune response against citrullinated epitopes.
The fibrinogen region between amino acids 414-433 is known to be autoantigenic in RA patients [6
], but for the citrullinated fibrinopeptide A (amino acids 20-35) no data on autoantigenicity is available, most likely because it is not part of the mature fibrinogen protein that has been used in such studies most frequently.
Interestingly, the fact that fibrinopeptide A, after cleavage from the fibrinogen precursor, contains a C-terminal arginine residue makes it an excellent substrate for PAD enzymes, which are able to efficiently convert such residues to citrulline (data not shown).
Both sequence elements were represented by a variety of peptides, differing in either their N- or C-terminal sequence (Figures and S1), most likely due to in vivo
proteolytic activities. While it is difficult to predict the enzyme responsible for their generation, proteases abundantly present in inflammatory cells, such as matrix metalloproteinases, elastases or caspases are likely involved. The two cleavage sites generating the full amino acids 20-35 fragment, however, are obvious, as this corresponds exactly to the borders of fibrinopeptide A, which is released from the A-alpha chain by cleavage at the N-terminal side of amino acid 20 by the signal peptidase and at the C-terminal side of amino acid 35 by thrombin [7
] The latter cleavage is one of the first steps of the clotting cascade and therefore the presence of (derivatives of) fibrinopeptide A is not very surprising, although its presence in a citrullinated fashion has not been reported before. Because thrombin cleavage at that position is dependent on the presence of an arginine residue, the citrullination of the C-terminal arginine of fibrinopeptide A has most likely occurred after thrombin cleavage. Citrullination of R35 in the uncleaved fibrinogen protein would inhibit thrombin cleavage and might, therefore, have implications on the efficiency of clotting in inflamed joints [25
]. The cleavage generating the N-terminus of the amino acids 414-433 fragment is most likely mediated by cathepsin D, which can cleave fibrinogen at that position in the process of clearing fibrin cloths [26
Fibrinogen is certainly not the only citrullinated antigen present in synovial fluid as, for example, citrullinated fibronectin and vimentin have also been found [21
]. However, the presence of high levels of endogenous citrullinated peptides derived from the protein is quite unique to fibrinogen, and could potentially provide a direct link between with the generation of an immune response targeting citrullinated epitopes. From a diagnostic point of view, such peptides might be very interesting to discriminate RA from closely related inflammatory diseases that also affect the joints. This could be achieved either by employing targeted mass spectrometric techniques or by the development of alternative, for example, antibody-based methods that would allow the specific detection of these citrullinated peptides in synovial fluid samples.