The tumor necrosis factor (TNF) family has anti-tumor and immune regulating properties. There are two distinct members of the TNF family, TNF-α and β. TNF-α is a regulator of apoptosis and proinflammatory cytokines with pleiotropic actions [1]. Thus, dysregulation of TNF-α is related to the pathophysiology of various inflammatory diseases such as rheumatoid arthritis, inflammatory bowel disease, ankylosing spondylitis, psoriatic arthritis, and psoriasis. The efficacy and safety of TNF-α blockers have been proven by several randomized controlled trials in psoriasis patients. However, because of the increased use of TNF-α blockers, side effects from these drugs have been reported. The adverse effects include opportunistic infections, reactivation of tuberculosis, various cutaneous side effects, lupus like reactions, demyelinating disease, heart failure and lymphoma [2, 3, 4].

Although the administration of a TNF-α blocker is indicated in the treatment of psoriasis, some case reports have shown that TNF-α blocker might induce psoriasis flare-ups and changes in morphology [5, 6, 7, 8].

A recent randomized controlled trial of IL-12/23 monoclonal antibody ustekinumab showed a favorable response for the treatment of psoriasis and psoriatic arthritis [9, 10]. Tocilizumab, an interleukin (IL)-6 receptor blockade, is used to treat disease modifying anti-rheumatic drugs (DMARDs) refractory rheumatoid arthritis (RA), and a favorable response has been observed without using concomitant methotrexate [11]. Here, we report on a refractory monoarthritis patient who previously had no skin lesions, but experienced psoriatic skin lesions after administration of a TNF-α blocker. The psoriasis of this patient responded to treatment with IL-12/23 blockade; however, the arthritis responded well to treatment with IL-6 receptor blockade.

A 53-year-old oriental woman visited a rheumatology clinic for right knee swelling and pain. She had a history of hypertension. She had never smoked. Her occupation is a housewife. She did not have a family history of spondyloarthritis. The patient experienced right knee swelling and pain after mountain climbing in the spring of 2009. Her symptoms were wax and wane. After several years of observation, the patient finally visited Seoul National University Hospital in February 2011. There was no inflammatory back pain during that period. For a diagnosis, a knee magnetic resonance imaging (MRI) was done which revealed diffuse synovial hypertrophy (Figure 1A). A diagnostic synovectomy was done for the knee arthritis, and histopathologically, there was chronic inflammation, lymphoplasmacytic infiltration, lymphoid follicles and synovial hyperplasia suggesting RA (Figure 2). At that time, rheumatoid factor, anti-cyclic citrullinated protein antibody and HLA-B27 were all negative. In a simple X-ray, there was no sacroiliitis. We tentatively diagnosed the patient with seronegative rheumatoid arthritis when taking into consideration the age, sex, hypertrophic synovial membrane and pathology, even though her case did not match the classification criteria. We used low dose glucocorticoids and methotrexate, leflunomide, or sulfasalazine. However, those medications lead to a loss of hair, abdominal pain, nausea and generalized edema; thus, we decided to stop all medications. Adalimumab injection was started in July 2012 and continued biweekly for 8 weeks, after which, the joint swelling in her right knee was reduced. However, after the 8th injection of adalimumab, a pustular skin rash appeared on her face, scalp, and trunk. We stopped the adalimumab injection because the newly appeared skin lesions were considered to be drug-related (Figure 3A and 3B). Thereafter, there was no more exacerbation of the skin rash after discontinuing the medication, but it still persisted. Because of the previous improvement in knee swelling after the use of a TNF-α blocker, we tried etanercept as a therapy in January 2013. After 4 injections of etanercept, there were psoriatic skin rashes over her whole body (Figure 3C and 3D). Thus, we stopped the medication again. A skin biopsy on her back revealed hyperkeratosis and irregular acanthosis in the epidermis suggesting active psoriasis. A focal decrease in the granular layer and inflammatory cells, especially lymphocyte infiltration at the dermis, was observed (Figure 4). Her skin lesions eventually developed to a palmar and shin pustular type indicating active psoriasis (Figure 3E and 3F).

Figure 1 (A) Right knee magnetic resonance imaging (MRI) reveals diffuse synovial hypertrophy (arrows) at diagnosis (December 2010). This finding is compatible to chronic arthritis. (B) Follow-up right knee MRI shows still synovial hypertrophy (arrows) in the synovial cavity but scattered and a semimembranous bursa (February 2013). Click for larger imageDownload as PowerPoint slide

Figure 2 Right knee synovectomy specimen shows chronic inflammation. In (A) lymphoplasmacytic infiltration, lymphoid follicles and synovial hypertrophy are seen (H&E, ×100), (B) is a magnification view of the lymphoid follicles (H&E, ×200). (C, D) show synovial hypertrophy (H&E; C: ×200, D: ×400). Click for larger imageDownload as PowerPoint slide

Figure 3 (A, B) After 8 separate injections at different time points, a pustular skin rash appeared in the face, scalp, and trunk. The rash was considered to be a drug related reaction. We stopped the Adalimumab injection because of the rash. At that time, we thought it was just a simple skin reaction. We started etanercept injection again 1 month later. (C, D) After 4 separate injections at different time points, there was a psoriatic rash over her whole body. We performed a skin biopsy on her back. That finding revealed psoriasis. (E, F) A little bit later in February 2013, there was a palmar and shin pustular psoriatic rash. This is a typical finding in psoriasis. (G, H) After two ustekinumab injections, psoriatic skin lesions nearly disappeared. Click for larger imageDownload as PowerPoint slide

Figure 4 Skin biopsy can confirm the diagnosis of psoriasis. There were hyperkeratosis and irregular acanthosis at the epidermis. The decreased granular layer was focal. Inflammatory cells, lymphocytes, and infiltration were observed at the dermis. Numerous neutrophils were in the stratum corneum. In the upper dermis, lympho-dominant inflammatory cell infiltration is seen (H&E, ×200). Click for larger imageDownload as PowerPoint slide

Four weeks later, starting from her last injection of etanercept, we started ustekinumab injections in February 2013 (45 mg, subcutaneously) to control her psoriasis and active synovitis in the right knee. Two weeks later, her knee swelling and pain had been slightly reduced, and then, 2 more injections of ustekinumab were given at 4 and 12 weeks after the 1st injection. Although the psoriatic skin lesions had much improved after the 3rd injection of ustekinumab (Figure 3G and 3H), the right knee synovitis did not respond to the ustekinumab treatment, and the synovial swelling became aggravated after transient improvement.

We examined some cytokine levels from her synovial fluid. Before the ustekinumab injection, the levels of IL-6 and IL-17 in the synovial fluid were elevated compared to the levels 1.5 years prior (Figure 5). After the first ustekinumab injection, the synovial cytokine levels were markedly decreased. However in the following period, joint symptoms became aggravated again, and the levels of IL-6 and IL-17 were elevated simultaneously. After 10 weeks from the third injection of ustekinumab, we started tocilizumab (400 mg, intravenously, monthly), a humanized monoclonal antibody against the IL-6 receptor, to treat the refractory synovitis. Joint swelling of her right knee and tenderness decreased after the 2nd injection of tocilizumab without any exacerbation of skin psoriasis. Tocilizumab was well tolerated and has not shown any skin or arthritis flare-ups.

Figure 5 This graph shows the interleukin (IL)-6 and IL-17 concentration in synovial fluid. Before the ustekinumab injection, the cytokine levels were elevated to the peak level. Interestingly, after the ustekinumab injection, the level of cytokines decreased dramatically. However, in the following period, joint symptoms become aggravated, and the cytokine levels in the synovial fluid increased. After 3 injections of ustekinumab, we started tocilizumab on Aug 16, 2013. We examined the IL-1β, IL-12 and tumor necrosis factor-α levels. Those data are very low out of the standard value range. The analysis was performed with Bio-Plex 200&Luminex; Bio-Rad, Hercules, CA, USA. Click for larger imageDownload as PowerPoint slide

TNF-α blockers are widely used in dermatologic, rheumatic, and autoimmune disorders. The safety of these drugs has been shown by many reports [2, 3, 4]. TNF-α blockers have been used to treat psoriasis. However, paradoxically, psoriasis newly develops de novo after therapy. Several case reports have revealed this phenomenon [5, 7]. This has been shown to be because of the interaction between TNF-α and interferon-α (IFN-α) [7]. This concept explains that plasmacytoid dendritic cells are the main source of IFN-α with increased levels in early psoriatic skin lesions. IFN-α induces the expression of CXCR3 in T cells, and then, T cells migrate to the skin. IFN-α accelerates the secretion of proinflammatory cytokines from myeloid dendritic cells. Thus, IFN-α injection can induce the development or exacerbation of psoriasis [12, 13]. TNF-α blocks IFN-α production through 2 ways. One way is to block the maturation of plasmacytoid dendritic cells. The other way is to block the release of IFN-α [8]. Following these theories, the development and exacerbation of psoriasis can be explained as follows: TNF-α blockers affect the function of TNF-α which regulates IFN-α production.

We tentatively diagnosed that the patient had seronegative rheumatoid arthritis from the synovial biopsy and knee MRI. Other disease criteria, including psoriatic arthritis, did not match her symptoms and signs. We started administering a TNF-α blocker because of the patient's intolerance to oral DMARDs (methotrexate, methylprednisolone, leflunomide, and sulfasalazine). After several injections of this drug, psoriatic skin lesions appeared. Because the skin biopsy pathology was compatible to psoriasis, we tried ustekinumab, a monoclonal antibody that binds to the shared p40 subunit of the IL-12/23 monoclonal antibody. IL-12/23 is critical in the pathophysiology of psoriasis, and ustekinumab is a novel drug in the treatment of psoriasis and psoriatic arthritis [9]. After the injection of ustekinumab, the skin change looked promising. After 12 weeks later, the skin lesions all disappeared. At the time of the 1st ustekinumab injection, a follow-up MRI was done. There was still scattered synovial hypertrophy in the right knee joint cavity. Because the beneficial role of synovectomy for the knee has been clearly shown for inflammatory arthritis, we also considered therapeutic synovectomy. However, at that time, the patient felt better in her joint than before the diagnostic synovectomy, and she was reluctant to undergo surgery. Thus, we just performed medical treatment. We did an analysis of the cytokines in the joint fluid including IL-6 and 17 (Bio-Plex 200&Luminex; Bio-Rad, Hercules, CA, USA) to further establish a management plan. Before the ustekinumab injection, the cytokine levels were elevated to 60,655.64 (IL-6)/78.51 (IL-17) pg/dL. After the first injection, the synovial cytokine levels were greatly decreased (Figure 5). IL-17 production has been observed in RA synovial tissue, and IL-6 is also known to be abundant in both the synovial fluid and sera of RA patients [11, 14, 15]. The patient showed transient improvement in synovitis after IL-12/23 receptor blockade, and arthritis flare-up accompanied the increased level of IL-6 (20,878.86 pg/dL) and IL-17 (68.70 pg/dL). IL-6 affects the function of inflammatory cells including neutrophils, T cells, B cells, monocytes, and osteoclasts which induce joint destruction and active arthritis. Tocilzumab, a monoclonal antibody against the IL-6 receptor could be a good treatment choice in this case because arthritis flare could be associated with increased IL-6 production in the inflamed synovium [11]. Her initial diagnosis was equivocal because of unmatched classification criteria; however, through a retrospective review, we ascertained that the exceptional response to tocilizumab was a key factor in her disease. We also checked other inflammatory cytokines such as TNF-α, IL-1, and IFN-γ but did not see any significant change according to flare-ups (data not shown).

TNF-α blockers can promote psoriasis development and flare-ups, and IL-12/23 blockade or IL-6 blockade could be a good alternative treatment for inflammatory arthritis with psoriasis induced after TNF-α blocker treatment. Evaluation of inflammatory cytokines within the involved joint synovial fluid can be used to select the appropriate biological treatment in refractory monoarthritis.