Article Text

Download PDFPDF

The early phase of psoriatic arthritis
  1. Dennis McGonagle1,
  2. Zoe Ash1,2,
  3. Laura Dickie2,
  4. Michael McDermott1,2,
  5. Sibel Zehra Aydin1,2
  1. 1Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, University of Leeds, Leeds, UK
  2. 2NIHR Leeds Musculoskeletal Biomedical Research Unit, Leeds Teaching Hospitals NHS Trust, Leeds, UK
  1. Correspondence to Dr Dennis McGonagle, Leeds Institute of Molecular Medicine, Section of Musculoskeletal Disease, University of Leeds, Chapel Allerton Hospital, Leeds LS7 4SA, UK; D.G.McGonagle{at}


Evaluation of the preclinical phases of the classic autoimmune diseases including rheumatoid arthritis has been facilitated by the availability of autoantibody and genetic markers that point firmly towards the early dysregulation of the adaptive immune responses. The association of psoriatic disease with the human leucocyte antigen—Cw0602 (HLA-Cw0602) gene has likewise led to the perception that autoimmunity has a pivotal role in early psoriatic arthritis (PsA). However, this HLA-Cw0602 genetic association does not appear to hold for PsA or associated nail, scalp and intergluteal skin involvement. Of note, these three sites of psoriasis are predictive of PsA evolution. For initiation of both skin and nail disease there is a link with Koebnerisation, or site-specific trauma. Nail disease is most common in the dominant hand thumbnail, pointing towards local tissue factors as disease initiators Likewise, for PsA, there is also good evidence for a history of previous joint trauma and histological studies showing microdamage in normal entheses which are typical locations where PsA frequently occurs. Furthermore, subclinical enthesopathy including osteitis is common in subjects with psoriasis but without arthritis. Collectively, these findings indicate that the classic model of adaptive immune dysregulation does not generally hold for the early stages of PsA. The way in which knowledge pertaining to tissue-specific factors in PsA, combined with the emerging data relating to monogenic disorders and animal models, points towards perturbation in the healing response and dysregulation of innate immune responses in early PsA is discussed. The way in which this model explains the clinical disconnect between skin and joint disease and the emerging human data that support it are demonstrated.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The purpose of this article is to examine the events that take place in the preclinical phase or the ‘early events’ that ultimately culminate in joint inflammation in psoriatic arthritis (PsA). To put this into a relevant immunopathological framework, we first describe the course of early autoantibody-associated immune diseases as a comparative platform for understanding early PsA.

Classical autoantibody-associated autoimmune disease: early events

Study of the early stages of the classic autoimmune diseases is facilitated by the availability of biomarkers, including anticitrullinated protein antibodies in rheumatoid arthritis, anti-mitochondrial antibodies in primary biliary cirrhosis and an array of autoantibodies in lupus and other disorders.1,,4 The use of various techniques, including the assessment of family members at higher risk of disease, combined with the predictive value of some major histocompatibility complex class II associations and the ability to detect subclinical disease has permitted robust models that envisage the failure of immunological tolerance in the primary and secondary lymphoid organs with inflammatory processes supervening in a particular target organ at a later date.5 6 As a rule, disease-associated autoantibodies lead to a progressive and inexorable attack on an organ until it is completely functionless—for example, the β cells of the pancreas in diabetes and the thyroid in autoimmune thyroid disease.7 In the case of anticitrullinated protein antibody-positive rheumatoid arthritis, recent genetic studies have confirmed multiple associations with proteins that are largely related to T- and B-cell function, including protein tyrosine phosphatase, non-receptor type 22 (PTPN22) and cytotoxic T-lymphocyte antigen 4 (CTLA4),8,,10 which vindicates the classic paradigm of a disease process that is genetically dependent on the failure of tolerance of the key effector cells of the adaptive immune response.

Admittedly, it is not presently clear why some antibodies against ubiquitously expressed proteins lead to organ-specific autoimmunity such as in dermatomyositis, but tantalising clues are emerging such as dysregulated disposal of relevant autoantigens in apoptotic blebs in target tissues.11 Despite this, several problems remain unresolved, including the apparently stochastic onset of autoimmune diseases several months or many years after the emergence of autoantibodies and questions relating to whether specific autoantibodies are primary or secondary to the underlying inflammatory processes. Nevertheless, autoantibodies represent an important biomarker for the classic autoimmune disease and give strong pointers to the early disease-associated events.

Psoriatic disease: early events

Psoriasis, PsA and psoriatic nail disease, now collectively labelled as psoriatic disease,12 have been viewed as archetypal autoimmune disorders.13 A reliable and predictive biomarker such as a disease-associated autoantibody, as set out above, has not been defined. At the population level, working out the early events in psoriatic disease is complicated by the erratic temporal associations between skin and joint disease. Skin disease predates arthritis in about 60% of cases, arthritis predates skin involvement in 20% and onset is contemporaneous in the remainder.14 15 Furthermore, the patterns of skin and joint disease are highly variable in psoriatic disease compared with classic autoimmunity such as that in the thyroid, pancreas or synovium where a target tissue is predictably attacked and eventually destroyed. This extreme heterogeneity of PsA itself, with new bone formation in some cases and mutilating arthritis in others,15 makes it even more difficult to conceptualise a unifying immunological basis.

In this perspective we tackle these problems by looking at psoriatic disease from the joint target tissue perspective rather than from the adaptive immune response. This is based on the immunological disease continuum model whereby many diseases formerly considered as autoimmune are in fact related to ‘tissue-specific’ dysregulation that culminates in secondary inflammatory responses.16 This classification scheme is underpinned by an understanding of the genetic basis for inflammatory disease in man.

Understanding the genetics of the pathogenesis of psoriatic disease

The seronegative spondyloarthritides (SpA) belong to a related group of conditions as set out in figure 1. The spectrum of disease associated with PsA includes chronic multifocal recurrent osteomyelitis (CRMO) and the SAPHO syndrome (synovitis, acne, pustulosis, hyperostosis and osteitis).17,,20 An autosomal recessive variant of CMRO termed Majeed syndrome has recently been characterised and is related to mutations in a gene called lipin.21 Of paramount importance is that this protein is expressed in cells of the innate immune system and, possibly, stromal cells where it may be involved in cell motility.22

Figure 1

Getting a genetic handle on psoriatic arthritis. A schematic figure demonstrating how the spondyloarthritis group of diseases, including psoriatic arthritis, which are collectively associated with osteitis, synovitis and colitis have clinical overlaps with monogenic disorders including CRMO and the PAPA syndrome. These rare disorders represent autoinflammatory disorders that are driven by innate immunopathology. The boxes in red depict some of the genes involved in SpA-related conditions and point towards AS, ankylosing spondylitis; CRMO, chronic multifocal recurrent osteomyelitis; IBD, inflammatory bowel disease; PAPA, pyogenic arthritis, pyoderma gangrenosum and acne; PsA, psoriatic arthritis; ReA, reactive arthritis, SAPHO, synovitis, acne, pustulosis, hyperostosis, and osteitis. CARD15, caspase recruitment domain gene; LCE, late cornified envelope; PSTPIP1, proline-serine-threonine phosphatase interacting protein 1.

A related autoinflammatory condition is the PAPA syndrome (pyogenic arthritis, pyoderma gangrenosum and acne), which shares overlapping features with PsA, including monoarthropathy and pustular skin disease.23 Collectively these clinical manifestations strongly overlap with inflammatory bowel disease and PsA phenotypes.24 In the case of the PAPA syndrome the causal mutations are in a protein called proline-serine-threonine phosphatase interacting protein 1 (PSTPIP1), which is expressed in cell of monocytic lineage.25

It is also striking to note that in a related animal model with mutations in PSTPIP2, bone destruction and nail disease are common26 and that these occur in recombinase-activating gene knockouts,27 indicating that the phenotype is totally dependent on cells of the innate immune system and the tissue to where the inflammatory reactions localises. Several other models firmly support the role of tissue-specific factors at the enthesis as key early players in PsA. In a tumour necrosis factor transgenic model of SpA that is characterised by sacroiliitis, arthritis and colitis, disease onset in the joints is exclusively dependent on the stromal cells of the enthesis and adjacent tendons rather than myeloid cells or lymphoid cells.28 The DBA/1 mouse model gets an oligoarthritis, dactylitis and nail disease and disease can be prevented by noggin, which antagonises bone morphogenetic proteins, illustrating how experimental PsA phenotypes can be modulated by molecules involved, not primarily in adaptive immune responses, but in the regulation of bone and joint biology.29,,32

These findings from human monogenic disorders that are supported by surrogate animal models are also backed up by the genetics of some polygenic diseases that fall within the SpA umbrella (figure 1). For example, the nucleotide-binding oligomerisation domain 2 and other mutations in Crohn's disease, which is part of the SpA disease spectrum, position the SpA category of disease as being further linked to an innate driven immunopathology (figure 1).33 Because the relevant cells that express these mutated proteins in both monogenic and polygenic SpA-related disease have a widespread tissue distribution, yet disease is confined to restricted territories, it was realised that tissue-specific factors are likely to be key events in the early stages of psoriatic disease in man.16 Viewing psoriatic disease from this perspective helps to explain the widely divergent phenotype between skin and joint disease, which are quite dissociated in time, severity and extent. That being the case, one might expect different tissue-specific factors to be involved in initiation of both skin and joint disease. In that respect, tissue-specific dysregulation of the late cornified envelope proteins has recently been associated with psoriasis but not PsA thus representing a tissue-specific factor for skin but not joint disease.34 35 Of course, in keeping with the overlap between skin and joint disease there may be some overlapping genes important for expression at both sites. Both the interleukin 23 receptor and interleukin 12 have been linked to psoriasis but show weaker associations with joint disease, which needs further investigation.36 It is fair to say that most genes involved in this dual site of skin and joint disease propensity await definition.

The claim that PsA is not primarily autoimmune in origin seems heretical at first because of the longstanding association with carriage of the HLA-Cw0602 gene in type 1 psoriasis.37 The HLA-Cw0602 association with skin disease has been confirmed38 and the link between HLA-Cw0602 and the extent of trunk and skin disease is well established.39 40 This compelling genetic evidence for adaptive immune response underlining the immunopathogenesis of type 1 psoriasis is supported by data showing the efficacy of T-cell directed treatments for skin disease.41 Thus far, the failure to conclusively identify specific autoantigens has hampered any effort to use current techniques such as peptide tetramer technology to track and study specific antigen-associated cell populations,42 thus making it difficult to work out the precise molecular pathways underpinning early skin disease.43 However, as discussed below, the HLA-Cw0602 association is linked only to certain cutaneous manifestations such as the extent of trunk and limb psoriasis and not specifically arthritis or nail disease.39 40 We will briefly discuss the role of the target tissues in early psoriatic disease as a prerequisite for an integrated genetic model for disease.

Tissue-specific factors predicting emergent arthritis in patients with psoriasis

The site-specific tissue dysregulation model of PsA leading to activation of innate immunity provides a new method of looking at the early phases of PsA. The question then becomes ‘what tissue-specific factors in the skin may be relevant for predicting subsequent arthritis?’ Data are scarce but a very elegant study by Wilson et al in nearly 1600 patients followed up for at least 10 years showed that lesions in the scalp and intergluteal regions had a hazard ratio of almost four and two for the subsequent evolution of arthritis.44 Why lesions at these anatomical sites are specifically associated with the development of PsA at this time remains enigmatic.

The original studies by Wright et al and others showed the association between nail disease and subsequent arthritis.45 46 Indeed the work by Wilson et al recently showed that nail disease had an OR of three for subsequent arthritis evolution.44 Presently the conceptualisation between the link between nail disease and arthritis is much clearer. The nail is intimately linked with the network of entheses fibres extending from the extensor tendon and collateral ligament enthesis that anchor the nail directly to the skeleton.47 Therefore from a tissue-specific factor it seems that nail dystrophy and PsA are anatomically linked to enthesopathy.48,,50

The role of trauma as a specific factor in the early phases of PsA

Early studies by Moll and Wright reported an association between trauma and the subsequent evolution of PsA in the same digit, with digital trauma to interphalangeal joints being associated with subsequent arthritis mutilans at that site.51,,53 This was termed as a ‘deep Koebner response’ and coined on the basis of the well-recognised Koebner response in psoriasis.54 A recent study by Rich et al found that nail dystrophy was most common in the dominant hand thumbnail, which points towards local tissue factors, likely to be injury related, as key drivers of nail disease.55 Indeed, several epidemiological surveys have since demonstrated the association between PsA and trauma at the population level.56 57 Historically, the connection between skin and joint has been viewed in relationship to synovial inflammation but we showed over a decade ago that early PsA and related SpA are associated with clinically unrecognised enthesitis and osteitis in the same joint.58 A similar pattern of enthesitis and related osteitis pathology has been demonstrated at virtually every site afflicted by PsA, including disease of the spine, SAPHO syndrome and dactylitis.59

What is interesting to note is that we and others have shown that mechanical problems and degenerative arthritis in the peripheral and axial skeleton have a similar MRI pattern of disease to PsA.60 61 Furthermore, following strenuous exercise, elite soldiers get patterns of MRI bone oedema reminiscent of those in PsA.62 Therefore, mechanically induced pathologies have a similar MRI appearances to PsA.

Based on these MRI observations, we have asked the question whether normal entheses and immediately adjacent bones, which are subject to high mechanical stressing, are sites of microscopic damage and inflammation. At both these sites we have found prominent evidence of micro damage and microscopic inflammatory changes, presumably related to a physiological healing process.63 64 We have recently used ultrasound to visualise the enthesis fibrocartilage in both patients with SpA and in young normal subjects and have confirmed evidence for microdamage of this site in health.65

Subclinical skeletal abnormalities in psoriasis

Although the Koebner response is well recognised in psoriasis,66 67 it is now becoming increasingly clear that asymptomatic entheseal or bone abnormalities have been frequently reported in patients with psoriasis without clinical arthropathy (table 1 and figure 2).68 A recent study in this regard found subclinical enthesopathy in over 40% of psoriasis cases69 and similar findings have been demonstrated with MRI,70 observations that support the deep Koebnerisation response. Not surprisingly the monogenic disorders of innate immunity and the PAPA syndrome, in particular, have been associated with trauma as a precipitating factor,23 24 71 which collectively suggests that microdamage is a pivotal factor in any array of inflammatory arthropathies related to the SpA spectrum.

Figure 2

In patients with psoriatic arthritis there is evidence of enthesopathy at clinically asymptomatic sites (A, B). Likewise, similar changes can be seen in patients with psoriasis and without evident arthritis (C, D), suggesting that the psoriatic disease spectrum is associated with extensive subclinical enthesopathy. (A) Longitudinal scan of the Achilles tendon (AT) insertion. Entheseal thickening (white line), hypoechogenicity of the tendon and the enthesis (*) and cortical irregularities of the calcaneus (cal). (B) Longitudinal scan of the patellar tendon (PT). Entheseal thickening (white line), hypoechogenicity of the tendon and the enthesis (*) and cortical irregularities of the patella (pat) with power Doppler signals. (C) Longitudinal scan of AT insertion. Thickening of the tendon (white line) with hypoechogenicity (*), loss of fibrillar echotexture and bone irregularities. (D) Longitudinal scan of the quadriceps tendon (QT) insertion. Power Doppler signals around the enthesophytes (+).

Table 1

Musculoskeletal imaging in patients with psoriasis patients shows a high prevalence of subclinical enthesopathy and related osteitis

Other factors in early psoriatic disease

Of course, the psoriasis and thus PsA spectrum are likely to be highly heterogeneous. Some types of disease are clearly infection related. For example, post-streptococcal psoriasis, which can evolve into chronic psoriasis in about 40% of cases, often has a short incubation phase following the inciting infection and it is associated with carriage of the HLA-Cw0602 antigen.72 This can also be associated with an inflammatory arthritis, which in our experience often behaves like a reactive arthritis and it is unclear whether there is any strong pathophysiological correlation between the two other than having a common inciting event. There may also be an association between subclinical gut inflammation and PsA in a subset of patients which needs to be considered.73

Divergent genetics between skin, joint and nail disease

It is fascinating to note that the immunopathogenesis of psoriatic skin disease that predicts arthritis (nail and scalp) is fundamentally different from classic type 1 psoriasis. For example, the HLA-Cw0602 association which is strong for type 1 psoriasis has not been proved in the case of PsA.74 Two studies have shown no association between HLA-Cw0602 and nail disease.40 75 In fact there is a negative association between nail disease and carriage of the HLA-Cw0602 antigen.39 Likewise, a negative association has been found between scalp disease and carriage of the HLA-Cw0602 antigen.39 These observations are intriguing because the genetics strongly support the recent clinical observation where scalp and nail disease are predictive of PsA as set out above.44


There seems to be a common biomechanical theme running between the skin and joints and involvement at other sites.76 There is a strong overlap between skin, nail and joint disease at the population level with respect to subclinical trauma. And it is relatively simple to conceptualise the wide phenotypic variation in psoriasis and PsA based on regional factors taking place in a particular joint—for example, a patient with a large knee effusion in PsA and a solitary plaque at another part of the body.

However, many patients with PsA have a polyarticular disease and even in those with oligoarticular disease clinically, ultrasound studies have demonstrated subclinical synovitis at clinically normal sites.77 It is highly unlikely that this simply reflects trauma and nothing else and the underlying immunopathogenic basis for this remains to be defined.76 Several possibilities merit consideration including other non-HLA-Cw0602 cytotoxic CD8 T-cell-driven responses analogous to the classic T-cell model for type 1 skin psoriasis in subsets of PsA.78

It is also clear that the tissue-specific factors which play a part in the regulation of repair, including inflammation, angiogenesis and molecular cascades involved in both soft tissue and bone healing, could all be important for joint and skin disease.79 In this regard, the innate lymphocyte populations that have roles in immunosurveillance and tissue repair, including natural killer (NK) cells, γδ T cells and NK T cells, can be comfortably accommodated within a model where restoration or maintenance of tissue integrity at skin barrier sites or entheses is required (figure 3). Of note, it has been suggested, that NK cells are involved in PsA disease pathogenesis based on putative associations with genes involved in NK-cell function, including major histocompatibility complex class I chain-related A genes.80 81 Recently, the results of five genome-wide association studies have appeared in the literature. These showed an association between PsA and psoriasis with a gene termed TRAF3IP2, which has multifaceted roles including interleukin-17 signalling.82 83 The link between type 1 psoriasis and adaptive immunity perturbation via the HLA-Cw6 pathway was further strengthened by an association with ERAP-1, which is involved in peptide loading via the major histocompatibility complex.84

Figure 3

Proposed scheme to explain the pathogenesis of early psoriatic disease. (Left) The classic autoimmune diseases such as autoantibody-positive rheumatoid arthritis and systemic lupus erythematosus arise mostly as a consequence of dysregulated adaptive immunity in the primary and secondary lymphoid organs. (Middle) In comparison, in type I psoriasis, tissue-specific insults or dysregulation in tissue homoeostasis are the key initiator for the entire spectrum of early psoriatic disease. At the population level there is a link between HLA-Cw0602 and type I psoriasis, which is generally more extensive than type II psoriasis. After an initial insult dendritic cells traffic to regional nodes, with subsequent activation of CD8 effector T cells that lead to an exaggerated inflammatory reaction in the skin. To what extent such a classic antigen–conventional T-cell model, if any, operates in psoriatic arthritis, at the population level awaits definition. (Right) In psoriatic arthritis and in some types of psoriasis it appears that disease localisation is determined by site-specific injury or dysregulated tissue homoeostasis, with a resultant inflammatory response as a consequence of innate immune cells, including macrophages, osteoclasts, neutrophils and innate immune lymphocytes including natural killer cells, γδ T cells and others. How and why this leads to chronic inflammation in some cases and the role of conventional T cells awaits definition. MHC, major histocompatibility complex.

In conclusion, as an alternative model to classic autoimmunity, we have reviewed the evidence for ‘deep Koebnerisation’ responses in PsA-prone sites and how this offers a fundamentally different immunopathogenesis concept. The emerging genetic and microanatomical features of psoriatic disease point towards a ‘differential immunopathology’ between skin and joint disease, with pure skin disease being an intermediate between autoimmunity and innate immunity, and joint disease itself being more linked to innate immune pathway perturbation.49 In support of this scenario, we have shown how animal models of destructive arthritis and nail disease are critically dependent on innate immunity and we looked at the role of trauma at insertions as propagators of PsA. We have focused less on the inflammatory cascades that are pivotal to final disease expression but have emphasised the factors that lead to site-specific inflammation in PsA in the first instance. This model for early PsA has implications for a better understanding of the PsA disease process.


View Abstract


  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.