Rheumatoid Arthritis: Anticytokine Therapies on the Horizon

The agents currently used to treat aggressive rheumatoid arthritis, the so-called disease-modifying antirheumatic drugs, were generally not developed for that purpose. At best, they reduce the destructive component of the disease modestly and inconsistently.

Although the cause of rheumatoid arthritis remains unknown, recent advances in molecular technology have facilitated the identification of distinct cell subsets, cell surface markers, and cell products that contribute to the inflammatory and destructive components of the disease. This enhanced understanding affords new opportunities to direct specific therapies at relevant disease molecules and cells. In particular, increasing evidence implicates interleukin-1 and tumor necrosis factor-α (TNF-α) as major contributors to the inflammatory and destructive manifestations of rheumatoid arthritis and, therefore, as promising targets for improved therapy.

In this connection, several natural endogenous inhibitors of interleukin-1 and TNF-α have been identified, including interleukin-1 receptor antagonist, soluble interleukin-1 receptors, and soluble TNF-α receptors. These natural inhibitors are present in normal persons. Although increased levels of these inhibitors are found in serum specimens and at sites of inflammation in patients with rheumatoid arthritis, interleukin-1 and TNF-α are present in excess relative to the amounts of their respective inhibitors, a balance that favors the proinflammatory actions of these cytokines. An obvious potential approach to treating rheumatoid arthritis is thus to simply neutralize the cytokines that are suspected of producing the damage.

Studies in animal models of arthritis have provided support for this approach. Interleukin-1 and TNF inhibitors have been shown to alleviate established collagen-induced arthritis [2, 3]. Mice that are transgenic for the human TNF-α gene develop spontaneous chronic inflammatory arthritis, and treating these mice with anti-TNF monoclonal antibody or recombinant soluble TNF receptor fusion proteins abrogates the arthritis [4, 5].

Although interleukin-1 and TNF-α have many overlapping biological properties, each may have distinct roles in the pathogenesis of rheumatoid arthritis. For example, depending on the stage of collagen-induced arthritis, inhibitors of TNF and interleukin-1 differ markedly in their ability to suppress the disease [6]. In some animal models of arthritis, blocking TNF-α suppressed inflammation, whereas blocking interleukin-1 more effectively prevented cartilage destruction [7]. These studies have obvious theoretical implications for the treatment of human disease; they suggest that therapy with inhibitors of both TNF-α and interleukin-1 would be more effective than therapy with either type of inhibitor alone.

Biological agents aimed at inhibiting the pro-inflammatory activities of these cytokines have thus far included cytokine receptor antagonists, anticytokine monoclonal antibodies, fusion molecules consisting of soluble cytokine receptors combined with human Fc constructs or polyethylene glycol (PEG), and counterregulatory cytokines (for example, interleukin-10, −4,or −11)that oppose the actions of the target cytokine. Inhibitors of the processing and synthesis of interleukin-1 or TNF-α are also in development.

Interleukin-1 receptor antagonist, a naturally occurring specific inhibitor, blocks the binding of interleukin-1 to its cell surface receptors but does not possess agonist activity [8]. In an initial pilot trial, patients with rheumatoid arthritis were given recombinant human interleukin-1 receptor antagonist in various dosages for 3 weeks [9]. A beneficial effect on clinical and biochemical responses (C-reactive protein levels) was noted in this short study. A subsequent large, double-blind, placebo-controlled trial lasting 6 months [10] produced statistically significant improvements according to American College of Rheumatology clinical response criteria. Additional placebo-controlled trials are in progress to evaluate the long-term efficacy and safety of interleukin-1 receptor antagonist.

Clinical trials in patients with rheumatoid arthritis using a chimeric anti-TNF monoclonal antibody, cA2, have provided the first direct evidence that TNF inhibitors might be useful therapeutic agents. First in an open-label trial [11] and later in a double-blind, placebo-controlled trial [12], swollen joint counts and serum C-reactive protein levels significantly improved. In the placebo-controlled trial, 58% of patients who received a single dose of cA2 (10 mg/kg of body weight) had at least a 50% improvement according to the Paulus clinical response criteria. Retreatment with cA2 has been effective, but human antichimeric antibody responses to cA2 that may limit the duration of retreatment have been noted [13]. Preliminary studies indicate that these responses may be diminished by concomitant treatment with low-dose methotrexate. The initial results of treatment of rheumatoid arthritis with humanized anti-TNF monoclonal antibodies have been encouraging (these antibodies have also been shown to be effective in treating refractory Crohn disease) [14, 15]. The mechanism of action of cA2 is still uncertain, but the observed anti-inflammatory effects may be partly attributable to downregulation of cytokine-induced vascular adhesion molecules and reduction of inflammatory cell movement into the joints.

The biological activities of TNF-α and TNF-β are mediated through two distinct membrane-bound receptors-type I, or p60 [p55], and type II, or p80 [p75]-that are expressed by numerous cell types [16]. Soluble TNF receptors have been isolated; they arise from the enzyme-mediated cleavage of the extracellular portion of the membrane-bound type I and II molecules. In an attempt to increase the half-life, affinity, and bioavailability of soluble TNF receptors, monomeric soluble TNF receptors have been fused to the Fc portion of an IgG₁ immunoglobulin, forming a recombinant soluble TNF receptor fusion protein. A different soluble TNF receptor preparation combines soluble TNF receptor p55 with PEG molecules and is called PEG-rsTNF-RI.

In a dose-escalation safety study in patients with refractory rheumatoid arthritis, the administration of soluble TNF receptor fusion protein (p75) produced no clinically significant side effects and decreased C-reactive protein levels and numbers of swollen joints [17]. These initial encouraging clinical results were reproduced in a multicenter, placebo-controlled trial [18]. Seventy-five percent of patients receiving the highest dose of soluble TNF receptor achieved at least a 20% clinical response according to American College of Rheumatology criteria. Moreover, 58% of patients receiving this dose had at least 50% improvement in disease activity measures. No human antibodies to the study agent were detected.

A p55 soluble TNF receptor fusion protein has also been evaluated in patients with rheumatoid arthritis; a dose response has been noted but no definite dose range has been established. Antibodies to this protein have been detected but are of uncertain clinical significance. Further clinical trials evaluating the long-term safety and efficacy of recombinant TNF receptor fusion protein are in progress.

Other approaches to inhibiting the proinflammatory activities of TNF and interleukin-1 are currently under investigation and include treatment with recombinant interleukin-10, −4,or −11.Interleukin-4 and −10 both inhibit the release and function of interleukin-1, TNF-α, and other proinflammatory cytokines (interleukin-6 and −8);inhibit the production of matrix metalloproteinases; and increase the secretion of natural inhibitors of cytokines, such as interleukin-1 receptor antagonist and soluble TNF receptor [19, 20]. Thus, interleukin-4 and −10 exert anti-inflammatory effects and probably serve as endogenous “buffers” or “natural” modulators of the activities of several proinflammatory cytokines. Interleukin-11 is an anti-inflammatory cytokine that can reduce production of TNF-α; interleukin-1, −12,and −6;and nitric oxide [21]. Clinical trials evaluating recombinant interleukin-4, −10,and −11 in rheumatoid arthritis are now in progress.

The encouraging clinical results of short-term trials of TNF-α and interleukin-1 inhibitors, including soluble TNF receptor fusion proteins, anti-TNF monoclonal antibodies, and rhuIL-1 receptor antagonist, clearly warrant further studies to determine not only whether these agents can modify the destructive component of the disease but also whether they can be administered safely for long periods. These questions are being addressed in trials in progress.

Given the many pathways that seem to be involved in the pathogenesis of rheumatoid arthritis, a combination of two or more biological agents may be necessary to suppress joint destruction, as suggested above. Furthermore, administering these agents together with traditional disease-modifying antirheumatic drugs might enhance their usefulness.

Enthusiasm about these agents that is based on the encouraging clinical results obtained to date must be tempered by concern about toxicities arising from the long-term neutralization of cytokines. To date, adverse events noted in a small number of patients include lymphoma, infections, and the development of antinuclear antibodies and antibodies to the biological agent. With the exception of the latter, it is not yet possible to determine whether these events are directly related to cytokine inhibition or represent complications of the underlying disease and its conventional treatment. Long-term follow-up of patients enrolled in trials evaluating these agents is necessary to determine the true frequency and causal mechanisms of apparent side effects.

The benefits of therapies directed against TNF-α and interleukin-1 support important roles for these cytokines in mediating the clinical manifestations of rheumatoid arthritis. The optimization of this approach, through increasing experience and extended, careful surveillance of patients, should assure anti-cytokine agents a place in the therapy for rheumatoid arthritis in the near future. And, finally, these agents may be useful in the treatment of other autoimmune and inflammatory disorders. What we learn from studying their therapeutic effects in one class of diseases may advance both our understanding and our clinical management of other classes.

• Copyright ©2004 by the American College of Physicians