Anti-TNF Therapy for Rheumatoid Arthritis

by Joan Bathon, M.D. (Founder and Director of Johns Hopkins Arthritis Center and Johns Hopkins Arthritis Center Website, 1999-2010; Current affiliation, Columbia University)

Website Director’s Note: Dr. Bathon led one of the seminal studies of etanercept demonstrating its efficacy in rheuamtoid arthritis called the ERA study. Her contributions to the field of RA extend from clinical trials to better understanding cardiovascular disease in RA. Although some information concerning available TNFs in this particular article may be out of date, we have kept this “classic” article as part of our archive to provide important historical perspective of these agents which now are a standard part of our treatment armamentarium.


The development of inhibitors of tumor necrosis factor (TNF) evolved from a targeted bench-to-bedside approach in which lessons learned from basic pathophysiological research were tested in patients with debilitating chronic inflammatory diseases, in particular rheumatoid arthritis (RA). Insofar as all prior treatments for RA evolved primarily from serendipitous observations, the TNF inhibitors represent the first “rationally based” treatment, as well as the first FDA-approved recombinant proteins (“biologics”) for the treatment of RA.

While a T cell mediated, antigen-specific process is undoubtedly critical to the initiation of RA, sustained inflammation is at least equally dependent on cytokine production by synovial macrophages and fibroblasts which may act on each other in an autocrine or paracrine manner. Tumor necrosis factor-a(TNF-a) and interleukin-1 (IL-1) are the major macrophage-derived cytokines present in the rheumatoid joint and both induce the synthesis and secretion from synovial fibroblasts of matrix-degrading proteases, prostanoids, interleukin-6 (IL-6), interleukin-8 (IL-8) and granulocyte-macrophage colony stimulating factor (GM-CSF). Consequently, attention has focused on inhibition of TNF-aas a way to treat RA.

TNF and TNF Receptors

Tumor necrosis factor-awas originally named for its ability to trigger necrosis of transplanted tumor cells in mice. The purification and cloning of a molecule called “cachectin”, which causes wasting in chronic diseases, was subsequently found to be identical to TNF-a. TNF is produced primarily by macrophages and, to a lesser extent, by lymphocytes. It is one of 17 known members of a family of polypeptides that bind to a corresponding family of receptors. The polypeptide ligands are characterized by a common core sequence predicted to contain 10 b-sheet forming sequences, and include TNF-a, lymphotoxin-aand -b, Fas ligand, CD40 ligand and others (Table I below). TNF-ais initially synthesized and expressed as a transmembrane molecule, the extracellular portion of which is subsequently cleaved by TNF-aconverting enzyme (TACE) to release the soluble 17 kDa molecule. Soluble TNF-acirculates as a homotrimer and engages its cognate receptors on cell surfaces.

Table I Selected members of the TNF ligand/receptor superfamily*.
TNF-R1 and -RII
*This is not a complete list. For complete list of TNF ligands and receptors, and updated nomenclature, see reference 4.

In contrast to the relatively restricted synthesis of TNF-aby macrophages and T cells, TNF receptors (TNF-R) are expressed by nearly every mammalian cell. This ubiquitous expression, in conjunction with cell-specific effector molecules that are triggered by the TNF-R, may explain the variety of effects of TNF which include apoptosis, the synthesis of protein and lipid inflammatory molecules, and transcription factors. Unlike other ligands of the TNF-R family that bind to a single receptor, TNF and lymphotoxin-a are capable of binding to each of the two TNF-R designated as TNF-RI (or p55) and TNF-RII (or p75). Interaction of TNF with its receptor triggers a conformational change and dimerization or clustering of receptors which, in turns, triggers the cellular response. TNF-R, like their ligand, can be cleaved from the cell surface by TACE but soluble TNF-R are believed to be present only in small amounts relative to membrane-bound TNF-R.

Development of TNF Inhibitors

The two strategies for inhibiting TNF that have been most extensively studied to date consist of monoclonal anti-TNF antibodies and soluble TNF receptors (sTNF-R) (Table II). Both constructs will theoretically bind to circulating TNF-a, thus limiting its ability to engage cell membrane-bound TNF receptors and activate inflammatory pathways. Soluble TNF-R, but not anti-TNF antibodies, would also be expected to bind lymphotoxin. [An alternate strategy is to develop a TACE inhibitor in order to limit the amount of circulating soluble TNF.]

The best studied of the monoclonal anti-TNF antibodies is infliximab (Remicade®), originally referred to as cA2. Infliximab is a chimeric human/mouse monoclonal anti-TNFaantibody composed of the constant regions of human (Hu) IgG1κ, coupled to the Fv region of a high-affinity neutralizing murine anti-HuTNFaantibody. The antibody exhibits high affinity (Ka 1010/mol) for recombinant and natural huTNFa, and neutralizes TNF-mediated cytotoxicity and other functions in vitro. Because of the potential for an immune reaction to the mouse protein components of a chimeric antibody, an alternate strategy has been to develop a fully human anti-TNF monoclonal antibody. One such antibody, known as D2E7, also known as adalumimab, was generated by phage display technology. A high affinity murine anti-TNF monoclonal antibody was used as a template for guided selection, which involves complete replacement of the murine heavy and light chains with human counterparts and subsequent optimization of the antigen-binding affinity. D2E7 (HumiraTM) received FDA approval in in December, 2002.

In the second approach to TNF inhibition, soluble TNF-R have been engineered as fusion proteins in which the extracellular ligand-binding portion of the huTNF-RI or huTNF-RII is coupled to a human immunoglobulin-like molecule. Although TNF-RI is thought to mediate most of the biological effects of TNF in vivo, engineered sTNF-RI and sTNF-RII constructs both appear to be effective in vivo inhibitors of TNF. Etanercept (sTNF-RII:Fc; Enbrel®) is the best studied of the sTNF-R and is approved for the treatment of rheumatoid arthritis in adults and in children. It is a dimeric construct in which two sTNF-RII (p75) are linked to the Fc portion of human IgG1. The dimeric receptor has a significantly higher affinity for TNF-a than the monomeric receptor (50-1000-fold higher), and the linkage to the Fc structure significantly prolongs the half-life of the construct in vivo. Although it also has an unnatural linkage site, anti-etanercept antibodies have been infrequent. Another mechanism for prolonging the half-life of monomeric receptors is via conjugation with polyethylene glycol. One such construct, PEG-sTNF-RI (p55), has shown efficacy in several animal models of arthritis and is now in early clinical trials (see below).

Table II TNF Inhibitors Currently Approved or in Development
Mouse-human chimeric anti-huTNF mAb
FDA approved
Fully human anti-huTNF mAb
FDA approved
p75sTNF-RII-Fc (dimeric)
FDA approved
PEG-p55sTNF-RI (monomeric)
Development terminated
p55sTNF-RI-IgG1 (dimeric)
Development terminated

Animal Studies

Several lines of evidence exist in animal models that support the importance of TNF-ain the pathogenesis of human RA. Although no animal model of inflammatory arthritis is thought to completely mimic human RA, studies in animals have provided important information on inflammatory mediators and their potential as therapeutic targets in human disease. Most compelling are the following findings:

  • Elevated levels of TNF-a in the joints of mice with collagen-induced arthritis (CIA)
  • Amelioration or prevention of CIA with anti-TNF blocking antibodies
  • Spontaneous development of inflammatory arthritis in transgenic mice overexpressing TNF-a

Studies in Mice with CIA.

Collagen-induced arthritis (CIA) in the mouse is induced by immunization of susceptible mice strains with native type II collagen. Macroscopically evident arthritis occurs between day 28-35 after immunization and persists for several months until the joints ankylose. CIA shares several histopathologic features with RA including mononuclear cell infiltration and synovial cell hyperplasia resulting in pannus formation with bone and cartilage destruction. In both RA and CIA, disease susceptibility is restricted by MHC class II alleles and autoreactive T cells are prominent in the joint with restriction in VbT cell receptor usage. Because of these similarities, CIA is a widely used experimental model for RA.

Similar to RA, several studies with CIA mice have demonstrated elevated TNF levels in the arthritic joints. Recently, to assess the level of cytokine expression during the course of CIA, CIA mice were sacrificed on a weekly basis starting at day 21, before the onset of clinical arthritis. Cytokine mRNA levels in joint tissue were measured by highly quantitative RNA protection assays. Levels of TNF-a were elevated early in the CIA disease course and persisted at high levels through the later stages. Elevated TNF-amRNA levels were also found in macroscopically and microscopically uninvolved joints. Weekly treatment of CIA mice with a neutralizing hamster monoclonal antibody to TNF-a prior to the onset of arthritis, ameliorated the severity of the disease both histologically and clinically although the incidence of arthritis did not change. Antibody treatment starting soon after the onset of arthritis had a similar but less pronounced effect on decreasing the severity of the arthritis. Anti-TNF therapy was ineffective if given 7 days after onset of arthritis. Of interest, use of a polyclonal antibody against IL-1aand IL-1bwas effective in CIA in both early and late disease. The effectiveness of anti-TNF therapy only in early disease in murine CIA is in marked contrast to the efficacy of TNF inhibitors in humans with both early and longstanding RA (see Clinical Trials below).

Similar results were also obtained using a sTNF-R1- IgG1 fusion protein construct. Administration of the sTNF-R1-IgG1 starting prior to the onset of arthritis decreased severity of arthritis but differed from the monoclonal antibody studies in decreasing the incidence of arthritis as well. Mice deficient in TNF-R1 by gene targeting were resistant to development of CIA confirming the importance of TNF-R1, possibly through mediating TNF induced adhesion molecule expression and mononuclear cell infiltration into the joint space.

Transgenic mice overexpressing human TNF-a.

Transgenic mice expressing a modified human TNF-atransgene spontaneously develop a chronic polyarthritis providing further evidence for the direct involvement of TNF in the pathogenesis of human RA. Mice carrying a human TNF transgene with a modified 3’region from a human globin gene show deregulated human TNF expression resulting in low level expression of TNF in the joints and a variety of other organs. In contrast, mice carrying a wild type human TNF transgene showed appropriately regulated TNF expression. Mice with deregulated TNF expression developed a chronic symmetric polyarthritis with histologic features similar to human RA. This process did not require a specific genetic background in the target mice.

Despite the differences with human RA, these animal models strongly support TNF as an important target for therapeutic intervention.

Human Clinical Trials

Because the safety of the TNF inhibitors in humans was unknown, early trials in RA targeted patients with severe, longstanding disease that had failed to respond adequately to conventional treatments such as methotrexate, gold salts, immunosuppressives and others. More recently, as the safety of these agents unfolded, patients with juvenile RA and adults with early RA have been targeted. An evolution in the selection of study outcomes has also occurred in that earlier trials focused on clinical parameters as endpoints, while more recent trials have focused on structural (radiographic) endpoints.

Some of the clinical data that led to the FDA approval of two anti-TNF therapies for the treatment of RA will be presented. The agents to be discussed are:

  • infliximab – mouse-human chimeric anti-human TNF antibody
  • etanercept – soluble p75 TNF receptor coupled to Fc portion of IgG

Relevant outcomes that are assessed in clinical trials of potential therapies for RA include individual clinical parameters, composite scores that integrate multiple clinical parameters, and radiographic scores.

Examples of individual clinical parameters include:

  • tender joint count
  • swollen joint count
  • erythrocyte sedimentation rate (ESR) or C-reactive protein (CRP)
  • visual analog scales for pain, function, global assessment

An example of a composite score is the ACR20 score.

However, the “gold standard” for evaluating the efficacy of a treatment in rheumatoid arthritis is its ability to slow or halt radiographic progression of the disease – that is, the treatment must slow or halt the development of new or enlarging erosions and slow the development of new or progressive joint space narrowing. Several scoring systems have been developed for quantifying these radiologic manifestations of RA.

Both infliximab and etanercept have been studied extensively in human subjects, and both are now FDA approved for the treatment of RA. Initially, because of the experimental nature of these treatments, only patients with long-standing, severe RA were evaluated and the identified outcome was clinical (but not radiologic) improvement. Although many of these patients had failed multiple conventional treatments for RA (such as methotrexate, gold, etc.), clinical responses to the TNF inhibitors were gratifyingly robust and rapid. More recently, patients with early disease have been targeted for study and their responses have been similarly robust. Even more compelling are radiographic data in both early and late disease patients which demonstrate the ability of anti-TNF agents to slow or halt radiographic damage in the majority of patients.

These data provide “proof of concept” in humans that TNF is indeed an important pathogenic mediator of joint damage in RA. Perhaps most intriguing about these studies is that targeting (inhibiting) a single cytokine can profoundly alter the natural history of this disease.

Potential Side Effects and Precautions

TNF plays an important role in host defenses, particularly in the killing of intracellular microorganisms such as Listeria and mycobacteria, and in inducing apoptosis of some tumor cells. Consequently, there has been some concern that long-term inhibition of TNF could lead to an increased incidence of infection and of malignancy. In addition, as these agents are genetically engineered proteins that will be given repeatedly over long periods for the treatment of chronic diseases, issues of immunogenicity and injection reactions require scrutiny.

Injection Site Reactions

With both etanercept and infliximab, injection reactions represent the most frequent and consistent side effect, although rarely limiting administration of the drugs. Reactions occur early after initiation of treatment, are generally mild and self limited, decrease and then resolve completely with repeated dosing.


Multiple studies in humans and animals demonstrate the importance of TNF-aas a defense against infection with intracellular organisms such as Listeria and mycobactaeria, raising concerns about the potential for increased infections with chronic TNF inhibition. TNF is increased in the systemic circulation after administration of endotoxin or bacteria, and TNF together with IL-1 are responsible for the physiologic alterations seen in septic shock. Mice deficient in TNF-a> by gene targeting lack primary B cell follicles and demonstrate impaired humoral immune responses to both T dependent and T independent antigens. Mice deficient in TNF-a, TNFR1 (p55), or TNFR2 (p75) are highly susceptible to infection by Listeria monocytogenes. In a human clinical trial, treatment of septic shock with etanercept resulted in increased mortality in patients with gram positive organisms.

Despite these concerns, controlled clinical trials with etanercept alone or in combination with methotrexate did not show an increase in either frequency, type or severity of infections. In one study with infliximab, however, investigators reported one death each due to tuberculosis and coccidiomycosis. Since infliximab and etanercept have been FDA approved and available commercially, a higher than expected number of caes of TB and fungal infections have been reported. These appear to be more frequent with infliximab, perhaps due to the differences in stoichiometry and the slower “off-rate” of infliximab compared to etanercept. Careful screening of patients for latent and active TB prior to intiation of anti-TNF therapy is now recommended. Careful monitoring for infection in patients treated with TNF inhibitors is also indicated.


The immune system has an important role in surveillance for malignancy, and the role of TNF , in particular, in triggering apoptosis of some tumor cell types has already been noted. Thus, an increased risk of malignancy is of theoretical concern with chronic long-term TNF inhibition. Unfortunately, short-term clinical trials cannot adequately address this question. At 3-year follow-ups of patients treated with TNF inhibitors in clinical trials, however, no apparent increase in the rate or type of malignancies has surfaced yet. However, definitive answers to the risk of malignancy await long-term treatment data in a wider population. Registries have been established to collect these data.


Infliximab is a chimeric monoclonal antibody containing 25% mouse sequence at the binding site for TNF. Of concern is the potential of the mouse sequence to elicit an anti-infliximab or human anti-chimeric antibody response that would limit the therapeutic efficacy. Such antibodies have indeed been found but they can be suppressed by the use of concomitant methotrexate. The effect of these antibodies on therapeutic efficacy remains unclear.

Although etanercept is composed entirely of human sequence, neoepitopes might be generated at the joining regions of the TNF receptor and the immunoglobulin Fc region which could elicit an anti-etanercept antibody response. This does not appear to be relevant. In the two published trials, non-blocking anti-etanercept antibodies were found in only 2 patients and did not have a notable effect on efficacy.

Of unclear etiology and clinical signficance is the development of low titers of anti-double stranded DNA (anti-ds-DNA) antibodies in patients treated with infliximab and etanercept. Anti-ds-DNA antibodies are considered to be specific for systemic lupus erythematosus. However, in generall, patients treated with infliximab or etanercept who developed these antibodies do not exhibit lupus like illnesses.


In vitro studies suggested that TNF is a critical and proximal mediator of the inflammatory pathway in the rheumatoid joint. Proof-of-concept for this hypothesis has now been provided by animal studies and clinical trials. Not only does TNF inhibition dramatically reduce markers of inflammation but it also slows or halts structural damage, and these effects appear to be as potent in early disease as they are in late disease. In human terms, these efficacies should translate to less functional disability and higher quality of life.

The robust responses to treatment with TNF inhibitors in rheumatoid arthritis and inflammatory bowel disease are likely to be the tip of the iceberg. Any chronic (noninfectious) inflammatory disease that is primarily macrophage-driven could be a potential target for anti-TNF therapy. For example, pilot trials are now underway to evaluate the efficacy of TNF inhibitors in Wegeners granulomatosis, psoriatic arthritis, congestive heart failure and others illnesses.

The potential contribution of interleukin-1, independent of TNF-a, in chronic inflammatory states remains to be clarified but it is likely that a combined approach to inhibit both monokines will be even more potent than either solitary approach. Finally, the rebound in disease activity that occurs after cessation of anti-TNF therapy is a sobering reminder that the inflammatory cascade has been interrupted by neutralizing TNF, but that the underlying cause(s) of the disease itself, has not been addressed.


  1. Moreland LW, Heck Jr. LW, and Koopman WJ. Biologic agents for treating rheumatoid arthritis. Arthritis Rheum40:397409, 1997.
  2. Le J, and Vilcek, J. 1Tumor necrosis factor and interleukin-1: Cytokines with multiple overlapping biological activities. Lab Invest56:234248, 1987.
  3. Wallach Varfolomeev EE, Malinin NL, Goltsev YV, Kovalenko AV, and Boldin MP. Tumor necrosis factor receptor and Fas signaling mechanisms. Annu Rev Immunol17:331367, 1999.
  4. Bazzoni F, and Beutler, B. The tumor necrosis factor ligand and receptor families. New Engl J Med334:17171725, 1996.
  5. Elliott MJ, Maini RN, Feldmann M, Long-Fox A, Charles P, Katsikis P, Brennan FM, Walker J, Bijl H, Ghrayeb J, and Woody JN. Treatment of rheumatoid arthritis with chimeric monoclonal antibodies to tumor necrosis factor a. Arthritis Rheum12: 16811690, 1993.
  6. Mori L, Iselin S, De Libero G, and Lesslauer W. Attenuation of collagen-induced arthritis in 55-kDa TNF receptor type 1 (TNFR1)-IgG1 treated and TNFR1 deficient mice. J Immunol157:3178-3182, 1996.
  7. Keffer J, Probert, L., Cazlaris, H., Georgopoulos, S., Kasalaris, E., Kioussis, D., and Kollias, G. Transgenic mice expressing human tumour necrosis factor: a predictive genetic model of arthritis. EMBO10:4025-4031, 1991.
  8. Maini RN, Breedveld FC, Kalden J.R., et al. Therapeutic efficacy of multiple intravenous infusions of anti-tumor necrosis factor a monoclonal antibody combined with low-dose weekly methotrexate in rheumatoid arthritis. Arthritis Rheum41:1552-1563, 1998.
  9. Maini R, St. Clair EW, Breedveld F, Furst D, et al. Infliximab (chimeric anti-tumour necrosis factor a monoclonal antibody) versus placebo in rheumatoid arthritis patients receiving concomitant methotrexate: a randomized phase III trial. Lancet354:1932-1939, 1999.
  10. Lipsky, P, van der Heijde DM, St. Clair EW, Furst DE, et al. Infliximab and methotrexate in the treatment of rheumatoid arthritis. New Engl J Med 343:1594-1602, 2000.
  11. Moreland, C., Smith, M.F., Eidlen, D., Arend, W.P. Interleukin 1 receptor antagonist (IL1RA) iL.W., Baumgartner, S. W., Schiff, M.H., Tindall, E.A., Fleischmann, R.M., Weaver, A.L., Ettlinger, R.E., Cohen, S., Koopman, W.J., Mohler, K., Widmer, M.B., and Blosch, C.M. 1997. Treatment of rheumatoid arthritis with a recombinant human tumor necrosis factor receptor (p75)-Fc fusion protein. New Engl J Med337:141-147, 1997.
  12. Bathon J, Martin RW, Fleischmann, et al. A comparison of etanercept and methotrexate in patients with early rheumatoid arthritis. New Engl J Med343:1586-1593, 2000.
  13. Rothe J, Lesslauer W, Lotscher H, Lang Y, Koebel P, Kontgen F, and Althage A. Mice lacking the tumor necrosis factor receptor 1 are resistant to TNF-mediated toxicity but highly susceptible to infection by Listeria monocytogenes. Nature364:798-802, 1993.
  14. Fisher CJ, Agosti JM, Opal SM, Lowry SF, Balk RA, Sadoff JC, Abraham E, Schein RM, and Benjamin E. Treatment of septic shock with the tumor necrosis factor receptor:Fc fusion protein. N Engl J Med334:1697-1702, 1996.

Updated: November 1, 2011

Arthritis Center

About Arthritis Center

Founded in 1998, the Arthritis Center at Johns Hopkins is dedicated to providing quality education to patients and healthcare providers alike.