Introduction
Psoriatic arthritis (PsA) is a chronic inflammatory autoimmune disease that is characterized by inflammatory arthritis and skin psoriasis. It is commonly classified as a subtype of the spondyloarthropathy (SpA) concept, together with ankylosing spondylitis (AS), reactive arthritis, arthritis associated with inflammatory bowel disease, and undifferentiated spondyloarthropathy (USpA) [
1]. Indeed, PsA shares a number of phenotypic characteristics with the other SpA subtypes, such as asymmetrical synovitis, enthesitis and familial aggregation. Typical SpA features such as sacroiliitis, presence of HLA-B27, uveitis and chronic inflammatory gut lesions are also found in PsA, albeit at a lower frequency than in other SpA subtypes [
2‐
4]. Interestingly, the peripheral joint involvement is mostly asymmetrical and oligoarticular, as in SpA, but can also mimic rheumatoid arthritis (RA) with eventually destructive involvement of multiple small hand and feet joints [
5]. Finally, skin and nail involvement, total ankylosis of destructed peripheral joints, predominant distal interphalyngeal joint involvement and arthritis mutilans sets PsA apart from both SpA and RA. Therefore, there is at present no clear consensus regarding whether PsA belongs to the SpA concept, whether it is a separate disease entity, or whether it is a heterogeneous disease group with oligoarticular and axial forms belonging to the SpA concept on the one hand and destructive, polyarticular forms resembling RA on the other [
6,
7]. This question will probably only be answered once the pathogenesis of the disease is fully elucidated, leading to an aetiological rather than phenotypical disease classification.
In previous histopathological synovial studies we considered PsA as part of the SpA concept and compared it as such with other inflammatory joint diseases. Although some early data are available on PsA synovial histology based on small studies of surgical or blind needle specimens, only few studies systematically evaluated PsA synovium obtained from actively inflamed joints [
8,
9]. Moreover, most of those studies used RA as a control and did not compare PsA and nonpsoriatic SpA, or oligoarticular and polyarticular PsA. In this context it must be highlighted that the few reports dealing with PsA synovium did not specify whether these patients had oligoarticular or polyarticular disease. The present study is the first to address these issues systematically and quantitatively in a large set of actively inflamed synovium samples by (1) conducting a detailed analysis of PsA synovitis in comparison with AS–USpA on the one hand and RA on the other, using a large panel of histopathological features; (2) comparing oligoarticular versus polyarticular PsA with respect to synovial features that discriminate between RA and SpA; and (3) analyzing the expression of S100A12, which was reported to be an interesting marker in PsA [
10].
Discussion
Because the synovial membrane is one of the primary tissue targets in PsA, detailed histopathological analysis of PsA synovitis might be a useful approach to gain new insights into this disease and to analyze biological similarities with or differences from other chronic inflammatory joint disorders such as RA and SpA. This is of particular importance because PsA is thought to resemble either RA or SpA, depending on the clinical pattern of peripheral joint involvement, with respectively polyarticular symmetrical disease and oligoarticular asymmetrical involvement. In this context, a detailed analysis of the synovial histopathology in PsA not only may provide new insights in the disease process but also may help to provide a biological rationale for the classification of PsA. However, until now such analyses are lacking.
Synovial histopathology in PsA is generally characterized by neovascularization, and inflammatory infiltration with predominantly mononuclear cells (T lymphocytes, B lymphocytes and plasma cells, and macrophages), although PMCs can also be detected [
28]. Mild to moderate synovial lining hyperplasia is observed in a considerable percentage of cases. Whereas these characteristics are found as well in other types of inflammatory arthritis, the availability of synovial biopsies from needle arthroscopy and new histopathological tools (monoclonal antibodies, RNA probes) have permitted a more detailed comparison of PsA and RA synovitis. One of the most prominent differences was the higher degree and typical morphology of synovial vascularity, which appeared to be related to angiogenic factors such as vascular endothelial growth factor, matrix metalloproteinase (MMP)-9 and angiopoietins, and was accompanied by a higher expression of E-selectin [
8,
29,
30]. Lining layer hyperplasia appeared to be more pronounced in RA, both by altered apoptosis of lining cells as by increased presence of CD68+ macrophage-like synoviocytes, although macrophage-derived pro-inflammatory cytokines such as tumour necrosis factor-α were also found abundantly in PsA synovium [
8,
31‐
33]. Apart from a difference in CD68+ macrophages and the somewhat unexpected presence of PMCs in PsA synovitis [
8,
28], no major differences in infiltrating cell populations were described between PsA and RA. Although these differences between PsA and RA are of interest, it should be highlighted that these reports did not clearly distinguish between different clinical subtypes of PsA and did not compare PsA with SpA.
Considering that the oligoarticular form of PsA might belong to the SpA concept, both the present study and previous reports of our group indicate similar differences between SpA and RA as those described between PsA and RA: higher lining layer thickness in RA, but more pronounced vascularity, presence of polymorphonuclear cells and presence of the CD163+ macrophage subset in SpA.[
14,
16,
19]. However, we could not confirm earlier reported differences in the presence of CD68+ macrophages between the disease groups [
8]. These findings of the present study are in agreement with our previous reports in SpA [
16,
19], in which we indicated that the increased synovial infiltration with CD163+ macrophages – a subset that may play an important role in innate immune inflammation – is not paralleled by an increase in global macrophage number, as detected using the pan-macrophage marker CD68. This is also in agreement with a recent independent study [
34] comparing PsA with RA, in which no difference was found in the number of CD68+ macrophages between the diseases.
In order to confirm whether similarities between SpA and PsA exist, as suggested by these independent studies, the present study provides a direct comparison of oligoarticular PsA with RA and with nonpsoriatic SpA. Our findings clearly demonstrate the resemblance between oligoarticular PsA and SpA synovium, whereas oligoarticular PsA synovium is significantly different from RA in terms of lining layer hyperplasia and PMC infiltration. Having previously indicated the importance of local disease activity but not disease duration in the assessment of synovial histopathology, the present study furthermore provides evidence that the obtained results are not biased by differences in treatment between the groups. Confirming the observations of other recent studies [
19,
35,
36], these data certainly do not suggest that DMARD and/or corticosteroid treatment have no influence on synovial histopathology [
23‐
27] but they indicate that global synovial histopathology is not altered in the case of persistent synovitis despite DMARD and/or systemic corticosteroid treatment.
Bearing in mind that subclassification of PsA is based on the pattern of peripheral joint involvement [
6,
7], it is surprising that, until now, no synovial histology data were available detailing oligoarticular PsA, which is thought to be related to SpA, and polyarticular PsA, which can mimic RA. Using histological parameters that discriminate between SpA and RA, such as lining layer thickness, degree of vascularity, presence of PMCs and CD163 expression, we were unable to demonstrate any significant differences between oligoarticular and polyarticular PsA. In addition, earlier reported markers distinguishing PsA from RA, such as CD68 and E-selectin, appeared not to be differentially expressed between oligoarticular and polyarticular PsA [
8,
37]. Considering the discrepancies with regard to published data on CD68 and E-selectin expression between PsA and RA, one should consider potential biases in study populations or in treatment schedules. Moreover, intracellular citrullinated proteins and Mab-12A staining, which are highly specific for RA and were found in approximately half of the RA synovia [
17,
18], were not detected in the PsA cohorts. These data provide the first clear evidence that polyarticular PsA does not exhibit RA-specific synovial features and that the synovial histopathology of PsA, either oligoarticular or polyarticular, is more closely related to SpA than to RA.
Apart from global histological features, a number of studies have analyzed more specific molecular systems in PsA synovium, including S100 proteins. The calcium-binding granulocyte protein S100A12 was recently described in PsA synovium [
10]: although the number of analyzed samples was too small to allow statistical comparison, a specific expression in PsA as compared with RA was suggested. In the present study we confirmed both the presence and the staining pattern of S100A12 in PsA synovium, but neither synovial tissue analysis nor synovial fluid measurements identified differences between PsA, SpA and RA. Two other S100 proteins (S100A8 and S100A9, respectively named myeloid-related protein [MRP]8 and MRP14), which are found in both infiltrating PMCs and monocytes, were recently described in different types of inflammatory arthritis [
36‐
40]. Similar to the findings of the present study, the expressions of these MRPs in synovium and synovial fluid were different between SpA and RA but not between PsA and nonpsoriatic SpA [
36]. Furthermore, we recently provided evidence that the effect of tumour necrosis factor-α blockade with infliximab on MRPs, as well as on global histological features, was similar in PsA and in nonpsoriatic SpA, further supporting the concept that peripheral joint disease is closely related in both diseases [
41,
42].
Nevertheless, the data provided by the present study and the resulting hypothesis require some critical comments. First, it should be noted that all patients included in the present study had an active knee synovitis, which could bias the study population by excluding polyarticular PsA with predominant or exclusive involvement of hand and feet joints. In addition, DMARD therapy has been shown to represent a confounding factor in terms of the pattern and number of swollen joints in established PsA; patients with polyarticular joint involvement may therefore be under-represented in our patient cohort [
7]. Moreover, because occasional evolution to polyarticular disease suggests that PsA is a more systemic disease than is nonpsoriatic SpA, it should be further investigated whether there are specific histological alterations in clinically uninvolved joints in PsA versus AS and USpA. Second, more sensitive scoring (digital image analysis versus semiquantitative scoring) or detection methods might have revealed small differences that could have been overlooked in the present study. However, such small differences should be considered in the light of their likelihood of being reproduced on the one hand and their biological relevance on the other [
43].
A third criticism of the present study is that it addressed a wide range of histopathological markers that are known to discriminate either PsA or SpA from RA, but the resulting concept requires confirmation by studies of cellular and molecular players, which are crucially involved in the pathogenesis of peripheral synovitis. Therefore, novel mediators identified
in vitro or in animal models of PsA should be further evaluated. Considering the specific radiological features of peripheral joint involvement in PsA, an interesting issue – which was not addressed in the present study – is the synovial mechanism involved in cartilage and/or bone destruction. However, we recently investigated a set of metalloproteinases and their inhibitors in a similar study [
35] that indicated that there was no difference in their synovial expression between SpA and RA. Surprisingly, the minor differences between PsA and nonpsoriatic SpA indicated less pronounced expression of MMP-1 and MMP-2 in PsA.
Another recent study [
44] demonstrated that synovial osteoclasts and the receptor activator of nuclear factor-κB (RANK)/RANK ligand system, which are known to play a crucial role in bone degradation in RA, were also clearly present in PsA synovium. However, a systematic comparison of synovial osteoclasts and RANK/RANK ligand in PsA, SpA and RA remains to be conducted. Taken together, these three issues indicate that the absence of histopathological differences between PsA and non-psoriatic SpA in the present study does not indicate that these conditions are similar. Further studies are needed to address these issues in greater detail.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
EK collected the samples, participated in the immunohistochemistry, performed statistical analysis and the interpretation of the study, and prepared the manuscript. DB designed the study, participated in its coordination, participated in the interpretation of the results, and drafted the manuscript. LDR participated in the collection of the samples, in the immunohistochemistry and in the interpretation of the results. BV participated in the collection of the samples and in the interpretation of the results. DF provided the polyclonal affinity-purified rabbit antisera against human S100A12 and participated in the interpretation of the results. JR provided the polyclonal affinity-purified rabbit antisera against human S100A12 and participated in the interpretation of the results. JC participated in the collection of the samples and in the interpretation of the results. AB provided the antibody detecting the MHC class II-HC gp39 peptide complexes and participated in the interpretation of the results. EMV supervised the collection of the samples as well as the design of the study. FDK participated in the design of the study and in its coordination, and participated in the interpretation of the results. All authors read and approved the final manuscript.