Disease exacerbation by etanercept in a mouse model of alphaviral arthritis and myositis




Mosquito-borne alphaviruses such as chikungunya virus, o'nyong-nyong virus, and Ross River virus (RRV) cause sporadic, sometimes large, outbreaks of rheumatic disease worldwide. This study was designed to test the effect of treating RRV-induced arthritis using the anti–tumor necrosis factor (anti-TNF) drug etanercept in a mouse model of rheumatic disease.


Mice were infected with RRV and treated with etanercept. Weight gain was measured, tissue viral titers were determined, and histologic changes in muscle and joint tissues were assessed.


RRV-infected mice treated with etanercept showed decreased weight gain, higher viral titers in muscle, joints, and blood, and more tissue damage and inflammatory cell recruitment than RRV-infected mice without treatment.


Anti-TNF therapy is unlikely to be useful in treating alphaviral arthritides. During alphaviral epidemics, careful monitoring of patients being treated with anti-TNF agents may be warranted.

A wide range of viruses are known to cause arthritis/arthralgia, including human immunodeficiency virus (HIV), hepatitis C virus (HCV), human parvovirus B19, Epstein-Barr virus, rubella virus, dengue virus, influenza virus, and several members of the genus Alphavirus (1). This genus includes the mosquito-borne arthrogenic arboviruses Mayaro, o'nyong-nyong, Barmah Forest, chikungunya, and Ross River virus (RRV). Chikungunya virus has recently reemerged to produce the largest epidemic ever recorded for this virus, with ∼260,000 cases reported in Reunion Island and >3 million cases reported in India (1). Numerous imported cases have been reported in Europe and the US (2) and, more recently, extensive outbreaks have occurred in Asia and the Indian Ocean region (3). RRV affects up to 8,000 individuals each year in Australia, with the disease primarily manifesting as polyarthritis/arthralgia, with myalgia, fever, and rash also often present. Although RRV-induced arthritis is generally self-limiting, the disease can be severe and can persist for months. Nonsteroidal antiinflammatory drugs are the current primary therapy, but these often provide only partial relief (1).

We have recently developed a mouse model of RRV-induced arthritis and myositis to better understand the pathogenesis of alphavirus-induced disease and to test new treatment options (4, 5). In this model, RRV-infected mice develop hind limb dysfunction, limb weakness, and muscle wasting. The disease is associated with prolific macrophage infiltration in joint and muscle tissues, and high levels of proinflammatory cytokines, including tumor necrosis factor α (TNFα), monocyte chemotactic protein 1 (MCP-1), and interferon-γ. The disease appears to have a significant immunopathologic component, as macrophage depletion, or treatment with inhibitors of MCP-1 or NF-κB, results in improved outcome (5, 6).

TNFα antagonists, which include monoclonal antibodies (infliximab and adalimumab) and a soluble TNFα receptor/Fc receptor fusion protein (etanercept), have revolutionized the treatment of rheumatoid arthritis (7) and certain other inflammatory diseases, including inflammatory bowel disease, psoriatic arthritis, juvenile chronic arthritis, psoriasis, and ankylosing spondylitis (8). TNFα inhibitors have also been used to treat autoimmune arthritides in patients concurrently infected with hepatitis B virus (HBV), HCV, or HIV, without apparent viral flares (9–12). TNFα is also induced during many viral arthritides, including RRV-induced rheumatic disease (1), perhaps suggesting that TNFα inhibitors might be used in the treatment of viral arthritides. However, TNFα may often be important for the development of antiviral immunity (13) and TNFα inhibitors have been associated with an increased risk of herpes zoster (14). We thus sought to determine whether anti-TNFα treatment of RRV-induced arthritis in a mouse model would provide therapeutic benefit or result in an exacerbation of disease.



RRV was generated from the T48 (RRV-T48) infectious plasmid clone as described previously (4, 5). Virus titers were determined by plaque assay as described previously (4, 5).

Mice and treatment.

Twenty-one–day-old C57BL/6J mice were obtained from Animal Resources Facility. These animals, weighing 6–7 gm each, were inoculated subcutaneously in the axillary area with 103 plaque-forming units of RRV (T48 strain) in 100 μl phosphate buffered saline (PBS) and monitored for changes in weight up to 10 days after infection. Some of the mice were inoculated with etanercept (400 μg/gm body weight) (Wyeth Pharmaceuticals) intraperitoneally on days 1, 3, 5, and 7 after infection. Control mice were injected with 100 μl PBS only. Muscle, ankle joint, and serum samples were collected on days 3, 5, and 10 after mice were infected, and processed for measurement of viral titers. For histologic analysis, muscle and ankle joint tissues were collected on day 10 after infection and fixed in formalin, embedded in paraffin, and processed for hematoxylin and eosin staining at the histology facilities of the John Curtin School of Medical Research (Australian National University, Canberra, Australia). In a separate experiment, mice were monitored twice per day up to day 15 after being infected. The experiments were approved by the Animal Ethics Committee of the University of Canberra.

Statistical analysis.

Data on weight differences between groups were analyzed by one-way analysis of variance, followed by Bonferroni adjustment for multiple comparisons. Data on difference in virus titers between experimental groups were assessed by unpaired 2-tailed t-test. Statistical analyses were performed using GraphPad Prism software version 4.0b. P values less than 0.05 were considered significant.


This mouse model of RRV produces a highly reproducible, acute, self-limiting arthritis that closely reflects the pathology of human RRV disease (15). RRV-infected mice also show a consistent decrease in weight gain (Figure 1A). Mice infected with RRV and receiving etanercept treatment on days 1, 3, 5, and 7 after infection exhibited a decrease in weight gain compared with untreated RRV-infected mice, which was apparent by day 3 after infection, and was statistically significant by days 4, 7, and 10 (Figure 1A). Uninfected mice that received etanercept showed no significant difference from animals receiving no treatment (Figure 1A).

Figure 1.

Ross River virus (RRV)–induced disease in C57BL/6J mice treated with etanercept. Mice were infected subcutaneously with 103 plaque-forming units of RRV. Mock-inoculated mice were injected with phosphate buffered saline alone. Mice received peritoneal injections of etanercept on days 1, 3, 5, and 7 after infection. A, Amount of weight gain after RRV inoculation. Values are the mean ± SD (n = 5 mice per group). ∗ = P < 0.05 by one-way analysis of variance followed by Bonferroni adjustment for multiple comparisons. Results are representative of 3 independent experiments. B, RRV titers in muscle (quadriceps), joint (ankle), and serum of RRV-infected and etanercept-treated RRV-infected mice at 3, 5, and 10 days after infection. The amount of virus present was determined by plaque assay on Vero cells. Values are the mean ± SD. Data shown are representative of 2 separate experiments (n = 5). ∗ = P < 0.05 by unpaired 2-tailed t-test. C, Survival rates in RRV-infected and etanercept-treated RRV-infected mice (n = 10 per group).

Since TNFα can contribute to antiviral activity (13), we tested whether etanercept treatment affected RRV replication. At 5 and 10 days after the mice were infected, the virus titers in the muscle and serum were significantly higher in etanercept-treated RRV-infected mice compared with untreated RRV-infected mice (Figure 1B), and titers in the ankle were significantly higher at 10 days. TNFα thus appears to be important for antiviral activity against RRV.

In a separate experiment, mice were monitored beyond 10 days after infection. By day 14, all etanercept-treated RRV-infected mice had died (Figure 1C).

We histologically examined the effect of etanercept treatment on muscle and joint tissue in RRV-infected mice. More pronounced disruption of striated muscle fibers and cell infiltrate was observed in the quadriceps muscle of etanercept-treated RRV-infected mice (Figure 2c) 10 days after infection, when compared to the untreated RRV-infected mice (Figure 2b). The ankle joint in etanercept-treated RRV-infected mice (Figures 2f and i) showed higher cell infiltrate in the synovial tissue when compared to the untreated RRV-infected mice (Figures 2e and h). The increased weight loss, muscle tissue disruption, and increased cell infiltrate in etanercept-treated RRV-infected mice are thus likely due to increased virus replication.

Figure 2.

Effect of etanercept on Ross River virus (RRV)–infected muscle and joint tissue. Mice were infected subcutaneously with 103 plaque-forming units of RRV. Mock-inoculated mice were injected with phosphate buffered saline alone. Mice received peritoneal injections of etanercept at 1, 3, 5, and 7 days after infection. a–c, At 10 days after infection, mice were perfused with 4% paraformaldehyde and 5-μm–thick paraffin-embedded sections generated from quadriceps muscle were stained with hematoxylin and eosin (H&E). Muscle specimens from uninfected mice (a), untreated RRV-infected mice (b), and etanercept-treated RRV-infected mice (c) are shown. d–i, Following decalcification, 5-μm–thick paraffin-embedded sections generated from ankle joints were stained with H&E. Synovial tissue specimens from uninfected mice (d and g), untreated RRV-infected mice (e and h), and etanercept-treated RRV-infected mice (f and i) are shown. g–i are higher magnification views of d–f. Arrows indicate cell infiltration. Each image is representative of 4 mice per group.


Treatment of RRV-infected mice with the TNFα inhibitor etanercept resulted in dramatic disease exacerbation, increased virus titers, and death in 100% of the mice. These results suggest that TNFα has critical antiviral activity against RRV infection and that anti-TNFα therapy is thus probably ill advised for treatment of alphaviral arthritides. This is the first study testing TNF inhibitors for the treatment of a viral arthritis, and suggests that this class of drugs may need to be used with caution for this group of diseases.

Infection with Mycobacterium tuberculosis, including reactivation of latent infection, has been a consistently observed side effect of anti-TNFα treatment, and it is now recommended that all patients be screened carefully for the presence of latent tuberculosis before starting anti-TNFα therapy (8). In contrast, little is known about the risks of viral infection in patients receiving anti-TNFα therapy. Anti-TNFα therapy can safely be given to at least some patients with HIV (11), HBV, or HCV (12), although it may be associated with an increased risk of herpes zoster (14). Our results suggest that patients receiving anti-TNFα therapy may be vulnerable to RRV infection. This scenario may be diagnostically challenging, and we suggest that clinicians in alphaviral endemic regions need to be particularly aware of this possibility. One potential strategy is that patients in alphaviral endemic regions could be screened for preexisting antiviral antibodies prior to the initiation of anti-TNFα therapy. The presence of such antibodies could indicate the presence of effective antiviral immunity against RRV, and might define a patient population in which anti-TNFα therapy could be safely given. Careful monitoring of patients receiving anti-TNFα therapy would thus appear to be warranted during alphaviral epidemics, in areas where arthrogenic alphaviruses are endemic, and for patients at high risk of mosquito exposure. Unfortunately, there is currently no licensed alphavirus vaccine available for humans.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Mahalingam 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 conception and design. Zaid, Rulli, Mahalingam.

Acquisition of data. Zaid, Rulli.

Analysis and interpretation of data. Zaid, Rulli, Rolph, Suhrbier, Mahalingam.