Combination cytokine therapy: The next generation of rheumatoid arthritis therapy?
The approach to the treatment of patients with rheumatoid arthritis (RA) has undergone dramatic evolution over the past few years. Impetus for this sweeping transformation has come from several areas, including a growing appreciation of the severity of this pernicious condition and a greater understanding of the immunopathogenesis of RA. Perhaps the most important factor underlying this renewed therapeutic vigor for RA has been the introduction of biologic agents targeting tumor necrosis factor (TNF). Results from rigorous clinical trials have demonstrated the substantial efficacy of TNF inhibitors (1–3). Although few comparative studies have been performed to date, it has been suggested that the extent of clinical benefit demonstrated with these agents may exceed that of traditional antirheumatic agents, with a number of patients achieving remission or near-remission. Moreover, not only have these agents improved the usual clinical parameters such as tender and swollen joint counts, there is also data suggesting that treatment with TNF inhibitors can substantially improve quality of life and may inhibit the progression of radiographic structural damage. Although there are still considerations that need to be addressed fully to optimize their use (4), experience in the clinic with the TNF inhibitors etanercept and infliximab suggests that their effectiveness in actual practice may approximate the promising results noted in trials.
The dramatic results seen with TNF inhibitors have effected a paradigm shift in the treatment of RA. In previous years, the “therapeutic pyramid” had been the standard. This approach to treating RA patients called for slow and cautious introduction of antirheumatic therapies to minimize potential toxicity (5). Information gleaned over the ensuing decades highlighted the aggressive, rapidly destructive nature of RA and the poor outcomes associated with persistent disease activity. Appreciation of the severity of RA altered the approach to treating patients, with the focus shifting toward earlier and more aggressive use of disease-modifying antirheumatic drugs (DMARDs) (5). Despite the successes of some treatment regimens, there were a sizable proportion of RA patients who, having run the gauntlet of traditional treatments, still had persistent disease activity. For these patients, there was a sense of futility at the lack of tenable treatment options. The introduction of TNF inhibitors, whose efficacy was initially established in RA patients with severe and refractory disease, added a powerful new tool to the therapeutic armamentarium for RA. Accompanying the appreciation of their therapeutic potential has been a mounting dissatisfaction with residual disease activity. Just as the appreciation of serious outcomes of RA resulted in the abandonment of the old therapeutic pyramid, the success of the TNF inhibitors has raised therapeutic expectations. Borrowing the parlance of protest movements, the contemporary goal of treating RA might be said to be “to eliminate inflammation, by whatever means necessary.” The focus for the next generation of RA therapy has become using therapeutic interventions in such a way that disease activity can be eradicated to the greatest extent possible for the largest number of patients.
The treatment of RA might be modified by a number of exciting avenues to achieve such a sublime goal. The use of combinations of traditional and newer DMARDs has been an area of renewed interest and investigation (6). Innovative treatment paradigms using TNF inhibitors and several DMARDs are being tried in the clinic. Another potential approach, and one with which there have been impressive results in animal studies, involves the use of combinations of cytokine-targeted therapies.
The optimal targeting of cytokines, either alone or in various combinations, begins with consideration of their activities. Upon binding specific cell-surface receptors, cytokines can exert diverse biologic activities (7, 8). Regarding function, cytokines may exhibit pleiotropy (i.e., one cytokine can mediate diverse functions), redundancy (i.e., several cytokines may mediate the same activity), or antagonism (i.e., the effects of one cytokine may be inhibited by another cytokine or by soluble forms of the cytokine receptor). A simplified classification of the dominant functions of some cytokines is shown in Table 1. Although cytokines can be considered individually, in vivo they function in complex networks and cascades. Thus, the overall outcome of a response reflects the balance between proinflammatory factors (e.g., inflammatory cytokines) and antiinflammatory factors (e.g., soluble forms of cytokine receptors, cytokines with antiinflammatory function) present in the local milieu. Given the myriad associations of various cytokines, it is probably naive to consider their interactions as if they occur in linear fashion. In vivo, their chaotic, cascading interactions allow for significant amplification of small perturbations. Focused therapeutic interventions might achieve outcomes seemingly far greater than might have been anticipated.
Table 1. Functions of cytokines relevant to their roles in rheumatoid arthritis*
|Regulation of the inflammatory response|
| Proinflammatory: TNFα, IL-1, IL-6, IL-7, chemokines (IL-8, MIP-1α/β, RANTES)|
| Antiinflammatory: IL-10, TGFβ, IL-4; IL-1Ra, sTNFR I/II, IL-1R II|
|Modulation of the immune response|
| Modify Th1/Th2 bias|
| Th1: IL-12, IL-18, IFNγ|
| Th2: IL-4, IL-13, IL-10, chemokines|
| Mediate activation/apoptosis|
| IL-2, IL-15, IFNγ, TNFα, IL-3, IL-5, IL-7|
| Bone and cartilage destruction|
| IL-1, TNFα, OPG/RANKL(TRANCE)|
| Angiogenesis/growth factors|
| TNFα, VEGF, TGFβ|
In RA, there is substantial evidence that cytokines subserve a crucial role in disease propagation and expression (7, 8). Indeed, a reductionist view of the pathophysiology of RA is that there is cytokine imbalance, with the excess activity of proinflammatory factors, such as TNF and interleukin 1 (IL-1), being incompletely offset by naturally occuring antiinflammatory factors, such as soluble TNF receptors, IL-1 receptor antagonist (IL-1Ra), and IL-10. This has made cytokine therapy an attractive intervention at present and has also provided a strong rationale for combination cytokine therapy in the future.
Combination Cytokine Therapy
Potential benefits. Potential benefits that might be achieved using combinations of therapies that target distinct cytokines include improved efficacy, lessened toxicity, and greater immunomodulation (Table 2).
Additive efficacy would be a simple increase in the extent of response or number of responders compared to single-agent therapy. Rationale supporting the potential benefit of such an approach comes from the long-appreciated observation that cytokines function in networks. Indeed, before TNF inhibitors had been shown effective, some skeptics predicted that such a focused approach would not work given the diverse, redundant network of proinflammatory cytokines present within the rheumatoid synovium. Now that the utility of TNF inhibitors has been proven conclusively, and the goals and paradigms for treating RA have changed, the question turns to why all patients do not respond; why all responders do not achieve remission; and what can be done to optimize efficacy. Interestingly, significant heterogeneity regarding cytokine mRNA expression has been observed in synovial biopsy specimens from patients with RA (9). As a correlate of this observation, it might be hypothesized that for some RA patients, TNF plays such a key role that its inhibition is sufficient to completely suppress disease activity. However, for other patients, or at other stages of disease, TNF might assume a role of somewhat lesser importance. For such patients, who may achieve modest but incomplete clinical benefit from TNF inhibition, the addition of another cytokine-targeted therapy might provide additional effect. Interestingly, it has been shown in animal models that IL-1 and TNF are synergistic in causing arthritis; this provides some support for targeting both cytokines in combination (10). If the amount of benefit achieved by the combination of 2 agents exceeds the sum of improvements from each used singly, the combination could be said to be synergistic. Such synergy, which would be a highly desirable outcome of combination therapy, could be achieved either by targeting cytokines with overlapping distinct functions.
Table 2. Potential benefits of combination therapy in rheumatoid arthritis
|Efficacy at lower doses|
| Lesser toxicity|
| Greater cost effectiveness|
| Less “resistance”|
| Facilitate induction of tolerance|
Synergy could also refer to the induction of other effects achieved by combinations of cytokines that might be distinct from those seen with monotherapy. Of note, TNF and IL-1 share a number of activities relevant to such systemic inflammatory diseases as RA (Table 3). Each of these activities has been demonstrated in experimental models, but the relative hierarchy or dose dependency for these effects in vivo is not clearly defined. For example, it is not clear what concentration of cytokine is required in a particular milieu to achieve all or some of these effects. Is it possible that specific concentrations of IL-1 and TNF allow responses to certain inflammatory stimuli, but not to others? Looked at another way, could inhibition of TNF and IL-1 to a specific extent decrease synovial inflammation and attenuate bony destruction, yet spare the response of the innate immune system to challenge foreign antigens? If this could be achieved, it would be a tremendous advantage for combination cytokine therapy.
Table 3. Activities of TNFα and IL-1*
|Effects on the vasculature|
| Upregulate adhesion receptors (ICAM-1, VCAM-1, E-selectin) via activation of NF-κB|
| Stimulate angiogenesis|
| Alter endothelium toward procoagulant activities [TNF]|
|Effects on cells|
| Activate lymphocytes: increases antibody production [IL-1], modifies CD44 adhesion [TNF]|
| Dendritic cells: maturation and migration into secondary lymphoid organs [TNF]|
| Activate neutrophils and platelets|
| Induce proliferation of fibroblasts/synoviocytes|
|Effects on mediators|
| Induce synthesis of proinflammatory cytokines, e.g., IL-6, GM-CSF, IL-1 [TNF]|
| Induce synthesis of proinflammatory chemokines (RANTES, IL-8, MIP-1α, MCP-1)|
| Induce other inflammatory mediators: PGE2, PAF, nitric oxide, reactive oxygen species|
| Induce synthesis of metalloproteinases that mediate bone and cartilage damage|
| Mediate pain and fever; cachexia [TNF]|
| Mobilize calcium from bone|
| Modulate apoptosis [TNF]|
| Upregulate antiinflammatory factors (sTNFR, IL-10, IL-1Ra)|
If enhanced efficacy could be achieved with combination therapy, that might engender other benefits as well. For example, the doses of agents required to be used might be decreased, which could then reduce the risk of adverse effects. A potentially relevant example of this has been seen in an animal model. In collagen-induced arthritis, high doses of a TNF inhibitor or cyclosporine can both effectively reduce arthritis, whereas lower doses of either agent are considerably less effective. However, the combination of both agents when used at lower, suboptimal doses achieves an additive effect (11). This could potentially allow for safer use of such agents in the clinic. Another benefit of combination therapy could be improved cost efficacy. Although biologic agents can be quite costly, estimation of their cost efficacy requires consideration of all relevant costs, both direct and indirect (12). Because efficacy is the major determinant of indirect costs, and indirect costs are the major contributor to total costs of therapy, enhanced efficacy of combination therapies might give them an incremental cost effectiveness compared to monotherapy.
Another potential advantage that may derive from combination cytokine therapy is the possibility for greater degrees of immunomodulation. Because different components of the inflammatory and immune responses function in cascades, combination therapy targeting distinct parts may exert a greater overall effect. As noted above, a clinical benefit of this could be synergy in terms of efficacy. In addition, it is possible that a combination could make the individual biologic agents less immunogenic and thereby better tolerated than if they were used alone. Concurrent targeting of different cytokines might also minimize the development of “resistance,” either pharmacokinetic or pharmacodynamic, to monotherapy. By analogy to therapeutic approaches used in oncology and infectious disease medicine, combinations of agents with distinct effects lessen the chances for emergence of resistance to therapy.
Perhaps the most desirable facet of combination therapy and its potential for greater immunomodulation would be the induction of immunologic tolerance. The propagation and sustenance of rheumatoid inflammation reflect an active, ongoing, immunologically driven process. Data from numerous models indicate that therapy targeting components of the immune response can induce specific immunologic tolerance to the relevant etiopathogenic antigens, reestablishment of immunologic homeostasis, and true remission. In this regard, the well recognized failure of a number of T cell-directed therapies (e.g., anti-CD4, anti-CD5, IL-2 inhibitors) in patients with RA has been profoundly disappointing, because these strategies are often effective in animal models of immune-driven inflammatory diseases. It is possible that part of the reason for the failure of T cell-directed therapies in RA is that the abundance of inflammatory factors in the local synovial milieu of patients with established disease precludes such efficacy. Of note, a potential role for inflammatory cytokines in generating specific autoimmune reactions has been postulated (13). Also, in animal models, the induction of tolerance with immune-modulating therapies is often more successful when the disease and its attendant inflammation are less well established. Therefore, combination cytokine therapy, by decreasing the total burden of inflammation, could conceivably facilitate the achievement of this ultimate long-term goal of treating RA, namely, cure by induction of specific immune tolerance.
Possible regimens. A variety of strategies for combining cytokine-directed therapies are possible (Table 4). Several combination regimens have been tested in animal models of arthritis with some intriguing results.
Table 4. Possible combinations of cytokine-targeted therapies*
|Antiinflammatory + antiinflammatory|
| TNF inhibitor + IL-1 inhibitor|
| TNF or IL-1 inhibitor + IL-10|
| IL-10 + IL-4|
| TNF or IL-1 inhibitor + chemokine inhibitor|
|Antiinflammatory + immunomodulator|
| TNF or IL-1 inhibitor + IL-2 inhibitor|
| TNF or IL-1 inhibitor + IL-12 or IL-18 inhibitor or IL-4|
|Other combination (cytokine targeted + other biologic agent)|
| TNF or IL-1 inhibitor + nondepleting anti-CD4 antibody|
| TNF or IL-1 inhibitor + inhibitor of co-stimulatory molecules (e.g., CD40/CD40L[CD154]; CD28/CD80; CD28/CD86)|
| TNF or IL-1 inhibitor + inhibitor of adhesion molecules (e.g., LFA-1/ICAM-1; VLA-4/VCAM-1)|
| TNF or IL-1 inhibitor + angiogenesis inhibitor|
Because TNF and IL-1 appear to be the most important proinflammatory cytokines, and because many of the key activities of these cytokines overlap (Table 3), combinations of agents targeting them would seem to be an attractive approach toward maximizing efficacy in arthritis. In animals, it has been suggested that TNF plays a more vital role in synovial inflammation, whereas IL-1 is more important for bone destruction (14); this would also be a rationale for combination therapy targeting these cytokines.
The combination of IL-1Ra and the soluble form of type I TNF receptor has been tested in rats with adjuvant arthritis (15). Cachexia, inflammation, joint damage, and bone density were assessed at various time points in this study. With both agents used as monotherapy, there was a dose-dependent benefit seen in all measured outcomes. For example, the greatest inhibition of inflammation seen with TNF blockade was 77%; that seen with IL-1 blockade was 33%. Regarding joint structure, blocking IL-1 attenuated <20% of damage and blocking TNF inhibited <40%. However, combination therapy with inhibitors of both cytokines commonly achieved synergistic benefit. Moreover, at the highest doses used, inflammation was completely inhibited, and joint damage was inhibited by more than 80%. The same combination of agents has been tested in other animal models, including collagen-induced arthritis in rats (16). Again, both agents were effective when used alone, and the combination achieved additive or synergistic efficacy in inhibition of inflammation and joint damage. Combination therapy at the higher doses tested almost completely inhibited disease. In patients with RA, the IL-1 inhibitor IL-1Ra has been studied and shown to be effective in alleviating signs and symptoms of disease, as well as slowing the progression of radiographic damage (17). The remarkable results seen with the combination of inhibitors of TNF and IL-1 in animal studies have stimulated preliminary studies testing this type of approach in RA patients.
Another strategy for combination therapy would be to use an inhibitor of TNF or IL-1 in conjunction with an antiinflammatory cytokine, such as IL-10. IL-10 has immune-modulating properties, and there is reason to hypothesize that it could be a promising therapy for RA. Although results in animal models have been promising, results from using IL-10 as monotherapy in RA have been mixed (18). In a study of collagen-induced arthritis, animals were treated with adenovirus encoding a viral form of IL-10, soluble TNF receptor-immunoglobulin construct, or both (19). Although neither therapy alone had a substantial effect, the combination synergistically suppressed arthritis. Of relevance to potential treatment of human disease, the combination was effective even when given after the onset of disease in the animal model.
In other animal studies, IL-10 was combined with another cytokine that would be predicted to have a salubrious effect in RA: IL-4. Both cytokines have been effective in controlling inflammation in some, but not all animal studies. However, the combination of IL-10 and IL-4 was shown to be synergistic (reviewed in reference 20). Interestingly, neither cytokine alone achieved dramatic results in that model; IL-4–treated animals had a joint score very similar to placebo-treated animals, whereas IL-10 treatment achieved about 20% inhibition. Combination therapy with IL-4 and IL-10 inhibited disease by more than 50% and also seemed to prevent cartilage damage. Each of these cytokines alone have been assessed in preliminary trials in RA patients with results that were modest, at best. Although the combination of IL-4 and IL-10 has not been tried in RA, it would seem to be a viable approach based on the animal studies.
Combination cytokine therapy that controls inflammation and also modifies the immune response could have significant potential for disease modification. In RA, the subset of T cells considered to be most vital in orchestrating immune-driven rheumatoid inflammation are memory CD4+ T cells with a Th1 pattern of cytokine secretion. Bias towards Th1 or Th2 phenotype is regulated to a large extent by a balance of cytokines; cytokines with a key role driving cells toward a Th1 phenotype include IL-12, IL-18, and interferon-γ. Therefore, inhibition of IL-12 might be able to steer the immune system in such a way as to dampen rheumatoid inflammation. Combination therapy with antibodies directed against IL-12 and TNF, alone and in combination, have been tested in murine collagen-induced arthritis (21). Anti–IL-12 treatment improved the clinical score of affected animals by up to 40%, whereas anti-TNF decreased it by approximately 50%. However, the combination demonstrated synergy to the point of nearly preventing any clinical evidence of disease. This approach to combination therapy, an antiinflammatory agent along with an immune modulator, may hold the most exciting promise for the future. Other approaches that may achieve the same effect include combinations of antiinflammatory biologic agents plus inhibitors of CD4, inhibitors of costimulatory molecules, and inhibitors of adhesion molecules (Table 4). Some of these approaches have shown promising effects in animal studies (11, 22)
Tempering the enthusiasm surrounding combination cytokine therapy and other potent immunomodulating regimens is the consideration that such therapy might be associated with untoward effects. Because immunosurveillance is critical to defense against nonself and altered-self, powerful immunomodulating treatments have the theoretical risk of increased proclivity to infection and malignancy. Clinicians have rightfully called for assiduous followup of patients treated with TNF inhbitors to evaluate for infectious complications (23). The addition of other agents targeting distinct components of the immune response should require further vigilance.
From the discussions of the cytokine functions, inhibiting more than one cytokine would seem to be rational and the approach straightforward. However, because they act in circuits and cascades, their interactions can be nonlinear and chaotic. On occasion, unanticipated and even paradoxical results have been seen in experimental systems (24, 25). Likewise in intact organisms, there is the possibility that a combination that seemingly makes sense might exert unanticipated effects.
There are a number of exciting areas for additional research in combination therapy. For example, this type of therapeutic approach would presumably be useful not only in RA, but in other inflammatory conditions, including Crohn's disease, psoriatic arthritis and psoriasis, ankylosing spondylitis, ulcerative colitis, and congestive heart failure, among others. Interestingly, experience to date with TNF inhibitors does not allow us to prospectively identify patients who might be expected to have exceptional efficacy with minimal toxicity. Additional studies may shed light on this question, which would be of great relevance to combination therapy as well. Allelic polymorphisms in genes encoding TNF, IL-1, and other cytokines have been identified. Preliminary results suggest that such genetic variation may be relevant to disease outcome in RA (26); this could provide critical insights into optimal therapeutic strategies.
In addition to developing other macromolecules, such as monoclonal antibodies and soluble receptors, there has been a tremendous effort to develop novel, small-molecule inhibitors of TNF and IL-1 (27). Such agents, including protease inhibitors, peptidomimetics, and enzyme inhibitors, could be as effective yet less expensive than current agents, and perhaps even be orally available. Finally, when considering combination cytokine therapy with biologics, it should be recalled that traditional agents (e.g., corticosteroids, cyclosporine, methotrexate) can exert effects on cytokine synthesis and function. This may have to be considered when using biologic agents concurrently with other drugs.
Summary and Conclusions
The approach to treatment of RA has undergone a paradigm shift. The considerable efficacy seen with TNF inhibitors has invigorated clinicians, stimulating them to redouble their efforts to control rheumatoid inflammation. A number of exciting avenues, such as combination cytokine therapy, may allow us to achieve the goal of eradicating disease activity. If the promising results from early animal studies of combination cytokine therapy can be reproduced in trials in RA, we may be able to achieve true disease modification and even remission for many of our patients with RA. Further developments in this area are eagerly awaited.