Is TNF Inhibitor Therapy Effective at Preventing Bony Ankylosis in Ankylosing Spondylitis?

TNF-α blockade in patients with ankylosing spondylitis (AS) has been shown to improve clinical symptoms of disease and physical function.  However, a benefit in reducing the incidence or severity of axial ankylosis has not been demonstrated.  In fact, qualified observational data has suggested that radiographic progression of ankylosis and inflammation may be unrelated in AS.  Here, Lories et al (Arthritis Rheum 2007; 56(2): 489) examine the relationship between TNF-α inhibition and joint ankylosis in a mouse model of AS.

Methods and Results

Experiment 1: In order to test the effect of anti-TNF therapy on inflammatory arthritis in mice, inflammatory arthritis was induced by articular injection of bovine serum albumin (BSA) and IL-1.  In this group, a single injection of etanercept was administered four days after the induction of arthritis.  Compared to untreated mice, mice treated with etanercept had significantly reduced TNF-α expression, joint inflammation, pannus formation, and bone and cartilage destruction.

Experiment 2: In order to test the effect of anti-TNF therapy on spontaneous arthritis, male mice prone to developing an AS-like disease were treated with twice weekly etanercept starting at 12 weeks of age or placebo and followed for the development of spontaneous arthritis.  In this setting, there were no qualitative or quantitative differences in the incidence or severity of arthritis.  Histologically, chondrogenesis, bony proliferation, and ankylosis were not statistically different in etanercept treated vs. placebo treated mice, despite the expression of TNF-α in regions of enthesial inflammation and damage.  Etanercept antibodies were common, but did not affect the ability of etanercept to inhibit TNF-α.  Bone morphogenic protein (BMP) and Smad signaling, steps involved in early chondrogenesis, were unaffected by etanercept therapy.

Experiment 3: In an in vitro model of enthesial chondrogenesis, TBF-β and BMP induced chondrogenesis was inhibited by the presence of TNF-α.  Osteogenesis was not affected by TNF-α.


Although effective in reducing inflammation and improving signs and symptoms of inflammatory synovitis, etanercept was ineffective at reducing enthesitis or subsequent ankylosis.  Maximal inhibition of these processes may actually require the presence of TNF-α.

Editorial Comment

These are very compelling findings which suggest that the processes of acute inflammation and ongoing structural damage may be independent in AS.  It is discouraging to imagine that therapies effective in treating the signs and symptoms of AS and related axial arthropathies may have no long term effect on structural progression.  Indeed, these results suggest that TNF-α may actually be protective against some forms of longterm damage, and that blocking TNF-α could result in more, rather than less, structural progression.  However, to date, there is no evidence to confirm these preliminary findings in humans.  Longer term follow-up of radiographic outcomes from clinical trials data is needed to more carefully address these concerns.

Jon Giles, MD

About Jon Giles, MD