Rheumatoid arthritis (RA) is a chronic inflammatory disorder of unknown etiology characterized by synovitis and articular destruction. Recently, novel anti–tumor necrosis factor α (anti-TNFα) agents have been reported to be effective for treatment of RA. This is strongly suggestive of the involvement of TNFα in the pathogenesis of RA (1).
TNFα is a potent proinflammatory cytokine exerting pleiotropic effects on various cell types. TNFα is generated as a precursor form called transmembrane TNFα, which is expressed as a 26-kd cell surface type II polypeptide on activated macrophages and lymphocytes as well as on other cell types (2–4). After being processed by TACE, a soluble form of TNFα (17 kd) is released and mediates its biologic activities through type I and type II TNF receptors (TNFRI and TNFRII) (5–8). Transmembrane TNFα is also known to exert its unique biologic functions, such as cytotoxic activity and polyclonal B cell activation, in a cell-to-cell contact manner (2, 9, 10).
Anti-TNF agents can be structurally classified into 2 types; one type is generated as an antibody against human TNFα and the other type is engineered from human TNF receptors. The former type includes infliximab and adalimumab, and the latter type includes etanercept (11). Infliximab is a chimeric mouse/human anti-TNFα monoclonal antibody (mAb) composed of a murine variable region and a human IgG1 constant region. Adalimumab is a fully humanized anti-TNFα mAb generated by recombinant DNA techniques, and its structure is indistinguishable from the normal human IgG1. Etanercept is composed of the extracellular portion of the 2 human type II TNF receptors linked to the Fc portion (CH2 and CH3 domains). All of these 3 agents are able to bind to a soluble form of TNFα and exert potent clinical effects on RA (12–14). In contrast, etanercept did not show such an effect on Crohn's disease (15) and Wegener's granulomatosis (16), while infliximab is effective for these granulomatous disorders (17, 18). Recently, the efficacy of adalimumab in induction therapy for Crohn's disease has been reported (19, 20). This difference in clinical outcome among anti-TNF agents cannot be explained solely by neutralization of the soluble TNFα, which is the action commonly shared by these 3 agents. The effects of anti-TNF agents on transmembrane TNFα–bearing cells (TNF-producing cells) should also be taken into account.
To our knowledge, only infliximab has been studied for activities of complement-dependent cytotoxicity (CDC) and antibody-dependent cell-mediated cytotoxicity (ADCC) against a transmembrane TNFα–bearing mouse cell line (21). Definitive data on CDC or ADCC have not yet been demonstrated for etanercept and adalimumab. The precise estimation of the effects of anti-TNF agents on transmembrane TNFα has not been performed, because TNF-producing cells, such as activated monocytes and lymphocytes, carry soluble TNFα, transmembrane TNFα, and TNF receptors as well as soluble TNFα bound to TNF receptors. Thus, it is too complex to estimate the specific effects of anti-TNF agents on transmembrane TNFα.
Our group has developed a simple system of transmembrane TNFα–expressing Jurkat T cells that stably express an uncleavable form of transmembrane TNFα on their cell surfaces and that are negative for TNF receptors (22, 23). This model enabled us to analyze actual biologic effects of anti-TNF agents on transmembrane TNFα without interference from other factors such as soluble TNFα and TNF receptors. By using this system, we have recently shown that infliximab, not etanercept, induced apoptosis and cell cycle arrest through outside-to-inside (reverse) signals by transmembrane TNFα (24). In the present study, we extended our research to perform a CDC and ADCC head-to-head comparison of all 3 clinically available anti-TNF agents. In addition, apoptosis and cell cycle arrest induced by outside-to-inside signals through transmembrane TNFα was demonstrated as a novel mechanism for the destruction of transmembrane TNFα–bearing cells by adalimumab.
- Top of page
- MATERIALS AND METHODS
- AUTHOR CONTRIBUTIONS
In the present study, we directly compared biologic effects of 3 available anti-TNF agents (infliximab, etanercept, and adalimumab) on transmembrane TNFα by using Jurkat T cells stably expressing transmembrane TNFα on their surfaces. We have demonstrated for the first time that these anti-TNF agents exert different biologic effects when examined in the evaluation systems such as CDC, ADCC, and outside-to-inside signals through transmembrane TNFα.
Infliximab and adalimumab, both of which were generated as anti-TNF mAb, showed potent CDC activities, but etanercept, an agent derived from a soluble form of TNFRII, did not. As shown in Figure 2B, there is a marked difference in CDC between infliximab/adalimumab and etanercept. For etanercept, 100 μg/ml is required, which is 2–3 logs more than for infliximab and adalimumab, and the effect is still markedly less than that of infliximab and adalimumab at 1 μg/ml. Since complement activation of the classical pathway is initiated by binding of C1 to the CH2 domains of IgG, it is conceivable that the anti-TNF antibodies possess CDC activities. Activation of the classical pathway of the complement system leads to the eventual formation of membrane attack complex and the resultant cell lysis. Etanercept also contains the Fc portion of IgG1; however, it is lacking the CH1 domain and hinge region (31). Considering that C3 exclusively binds to IgG1 within a narrow region of 23 residues in the CH1 domain (32), it is likely that etanercept initiates complement activation by its Fc portion, but does not provide an appropriate platform for C3, which culminates in the impaired activation of subsequent complement components. Moreover, the lack of hinge region in the Fc portion of etanercept resulted in structural rigidity compared with natural antibody, which would eventually be the conformational hindrance to the proper access of complement.
As for ADCC, all of the 3 anti-TNF agents exerted almost equal activities. Because all of them contain the CH2 domain of the Fc portion of IgG1, a domain recognized by Fc receptors, it is likely that ADCC activities are equivalent in these anti-TNF agents. A number of anti-CD20 mAb have been generated for clinical use, and these have been classified into 2 functional subgroups, type I and type II (33). Type I mAb induce translocation of CD20 into lipid rafts, while type II mAb do not. Both types of anti-CD20 mAb exert almost equivalent ADCC activities. In contrast, CDC is strongly induced only by type I mAb, but not by type II mAb (34). This close relationship between the ability of mAb to translocate CD20 into lipid rafts and the ability to induce complement activation prompted us to investigate whether anti-TNF agents need similar molecular events for their CDC. The lipid raft inhibitor MCD did not abolish CDC by anti-TNF agents (data not shown). It is thus suggested that translocation of transmembrane TNFα into lipid rafts is not required for complement activation by anti-TNF agents.
Apoptosis and cell cycle arrest were induced by adalimumab as a direct effect mediated by transmembrane TNFα, a novel biologic activity identified in addition to its CDC and ADCC activities. Adalimumab was recently shown to induce apoptosis in activated human monocyte cell line THP-1 cells (35). Since THP-1 cells express soluble and transmembrane forms of TNFα, as well as TNF receptors, it is difficult to determine what mechanism is involved in this apoptosis. There are several possible explanations, as follows: 1) direct outside-to-inside (reverse) signals through transmembrane TNFα, 2) neutralization of the effect of soluble and membrane forms of TNF on TNFRI or TNFRII, and 3) stimulation of receptor-bound soluble TNFα. Our experimental system is simple because only the uncleavable form of transmembrane TNFα is expressed on human Jurkat T cells, which do not express the TNF receptors TNFRI and TNFRII (24). By using this system, we have previously reported that infliximab induced apoptosis and cell cycle arrest by an outside-to-inside signal through transmembrane TNFα (24). Here we demonstrated that adalimumab also induced apoptosis and cell cycle arrest as in the case of infliximab. It is of note that outside-to-inside signals are transmitted not only in transmembrane TNFα–expressing Jurkat T cells, but also in transmembrane TNFα–bearing CD4+ T cells from human peripheral blood (22).
Transmembrane TNFα has recently been shown to contribute to the host defense against acute Mycobacterium tuberculosis infection (36, 37). Moreover, transmembrane TNFα, in the absence of soluble TNFα, induces colitis in a mouse model (38). Infliximab seems to be more potent than etanercept in the elimination of transmembrane TNFα–bearing T cells. Thus, infliximab may lead to more efficient inhibition of granuloma formation by these cells in tuberculosis or in Crohn's disease, as compared with etanercept. Specimens from patients treated with infliximab lack granuloma formation (39). In the case of tuberculosis, in which granuloma formation is a protective reaction for host defense, infliximab might be more harmful compared with etanercept. In fact, postmarketing surveillance in the US has identified more cases of tuberculosis in patients treated with infliximab than in those treated with etanercept (40). This difference in the reported number of cases of tuberculosis might be explained by many other factors; however, our present data concerning the different effects of these 2 agents on transmembrane TNFα might at least partly explain the cause.
Considering that our assay system utilizes a human T cell line with accumulated transmembrane TNFα expression, our study is not without its limitations, especially in the case of monocytes, another potent source of TNFα in addition to T lymphocytes. The physiologic role of transmembrane TNFα in the action of anti-TNF agents should be studied further using various cell types, both in vitro and in vivo.
In conclusion, we demonstrated that infliximab and adalimumab have the abilities of CDC, ADCC, and outside-to-inside signals (apoptosis/cell cycle arrest) through transmembrane TNFα in our assay system. Although these findings are in vitro data, it is likely that these agents could induce efficient elimination of TNF-producing cells. However, etanercept had only ADCC activity and lacked activities of CDC and outside-to-inside signals through transmembrane TNFα, suggesting that etanercept may not be as effective as infliximab and adalimumab in the destruction of TNF-producing cells. Since all of the 3 anti-TNF agents neutralize soluble TNFα and are equally effective for treatment of RA, it is likely that soluble TNFα, rather than transmembrane TNFα, is primarily important in the pathogenesis of RA. In contrast, transmembrane TNFα may play a more critical role in the pathogenesis of granulomatous diseases such as Crohn's disease and Wegener's granulomatosis, because etanercept, which exerts only limited effects on transmembrane TNFα, is not clinically effective for these granulomatous diseases.
- Top of page
- MATERIALS AND METHODS
- AUTHOR CONTRIBUTIONS
Dr. Horiuchi 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 design. Mitoma, Horiuchi, Hatta.
Acquisition of data. Mitoma, Tamimoto, Kimoto, To.
Analysis and interpretation of data. Mitoma, Horiuchi, Tsukamoto, Uchino, Harashima.
Manuscript preparation. Mitoma, Horiuchi, Harada.
Statistical analysis. Mitoma.