To investigate whether IgG-degrading enzyme of Streptococcus pyogenes (IdeS), a bacterial cysteine endopeptidase that cleaves human IgG in the hinge region, can be used for blocking the development of arthritis.
To investigate whether IgG-degrading enzyme of Streptococcus pyogenes (IdeS), a bacterial cysteine endopeptidase that cleaves human IgG in the hinge region, can be used for blocking the development of arthritis.
Recombinant IdeS was purified and tested for specificity against mouse IgG. IdeS was injected intravenously into mice with collagen antibody–induced arthritis (CAIA), collagen-induced arthritis (CIA), or relapsing CIA, and its effects on arthritis development and severity were assessed.
IdeS efficiently cleaved mouse IgG2a/c and IgG3 in vitro. Even at low dosage (10 μg), IdeS specifically cleaved IgG2a in vivo without any apparent side effects. IdeS treatment efficiently blocked CAIA induced by IgG2a antibodies. No effect was observed when arthritis was induced with IgG2b anti–type II collagen antibodies; since IdeS does not cleave IgG2b, this indicated that IgG cleavage was the mechanism of action. IdeS treatment reduced the severity of arthritis if administered within 24 hours after the onset of clinical arthritis, but did not block ongoing severe arthritis. IdeS treatment also significantly prevented an antibody-induced relapse in mice that had chronic arthritis, and delayed the onset and reduced the severity of arthritis in classic CIA.
IdeS has therapeutic potential in IgG antibody–mediated autoimmune arthritis, representing a new and unique means of blocking pathogenic antibodies.
Rheumatoid arthritis (RA) is a degenerative inflammatory autoimmune disease involving articular cartilage, in which both the cellular and the humoral immune mechanisms act in concert to mediate disease progression. B cell regulation is disturbed in RA, as evidenced by elevated levels of various autoantibodies, such as rheumatoid factor, autoantibodies to cyclic citrullinated peptide, type II collagen (CII), glucose-6-phosphate isomerase, and many other self antigens (1). The observation that some of these are arthritogenic in animal models emphasizes the importance of antibodies and B cells in arthritis disease induction. Arthritogenicity of CII-specific monoclonal antibodies (mAb) in vivo has now been well documented (for review, see ref. 2). Collagen antibody–induced arthritis (CAIA) is Fcγ receptor dependent (3) and complement dependent (4), but not B cell or T cell dependent (5); neutrophils and macrophages are the major mediators of this inflammation (6), and secretion of tumor necrosis factor α and interleukin-1β is pathogenic (7). CAIA provides an opportunity to study the inflammatory phase without involving the priming phase of the immune response.
The efficacy of anti-CD20 therapy in RA (8, 9) confirms that B cells play a pathogenic role and presents new treatment possibilities. The presence of antibodies to CD20 leads to depletion of B cells but not of long-lived plasma cells, and it takes time for the antibody levels to decrease. To develop therapeutic agents directly targeting the pathogenic antibodies, we sought to mimic the action of infectious agents in trying to avoid the immune system. Diversion of the adaptive immune system by cleaving immunoglobulins is a common mechanism utilized by many pathogenic bacterial species (for review, see ref. 10). The gram-positive bacterium Streptococcus pyogenes (group A Streptococcus) is a common human pathogen, often causing relatively mild clinical conditions such as pharyngitis or impetigo (11). However, invasive strains can penetrate into deeper tissues and cause severe or life-threatening infections such as necrotizing fasciitis, sepsis, and streptococcal toxic shock syndrome. In addition, postinfectious sequela such as rheumatic fever and poststreptococcal glomerulonephritis sometimes follow acute streptococcal infections.
The IgG-degrading enzyme of S pyogenes (IdeS) (also designated Mac-1) (12) is a cysteine endopeptidase secreted by group A streptococcal strains during infection. It cleaves the heavy chains of IgG with a unique specificity, thus generating 2 monomeric Fab′ fragments and 1 Fc fragment (13–15). By removing the Fc part from the antigen recognizing Fab, immune responses such as complement deposition and Fc-mediated phagocytosis are inhibited. This proteolytic degradation promotes inhibition of opsonophagocytosis and interferes with the killing of group A Streptococcus (13, 16). In this study, we demonstrated that IdeS completely blocked antibody-induced arthritis in mice in vivo, delayed the onset and reduced the severity of collagen-induced arthritis (CIA), and significantly inhibited antibody-induced relapses in a model of chronic arthritis.
All mice used in this study were bred in the animal facilities of the Section for Medical Inflammation Research, Lund University, but original founders for the B10.RIII and B10.Q mice were from Professor Jan Klein (Tubingen, Germany), and original founders for the DBA/1J and BALB/c mice were from The Jackson Laboratory (Bar Harbor, ME). All animal experiments were approved by the local animal welfare authorities.
Recombinant IdeS (1 μg) was incubated with 5 μg of purified mAb (M287 [IgG2a], CIIC1 [IgG2a], M2139 [IgG2b], M284 [IgG1], or B10 [IgG3]) in 10 μl phosphate buffered saline (PBS) for 3 hours at 37°C. Samples were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions for 1 hour at 120V on 4–20% Precise Protein Gradient Gels (Pierce, Rockford, IL), and stained with Coomassie blue (0.25% brilliant blue R). IgG1, IgG2a, and IgG2b are mouse mAb binding to CII, and were affinity purified from culture supernatants; IgG3 mAb was purchased from Southern Biotechnology (Birmingham, AL).
CII-specific B cell hybridomas were generated and characterized (17–19). The clones were cultured in CL-1000 flasks (Integra Biosciences, Walliselen, Switzerland) using Dulbecco's Glutamax-I containing ultra-low bovine IgG (Gibco BRL, Invitrogen, Stockholm, Sweden). Antibodies from clones M2139 (IgG2b) and M287 (IgG2a) bind J1 epitopes of CII (551–564; GERGAAGIAGPK), CIIC1 (IgG2a) binds C1I epitopes of CII (359–363; ARGLT), CIIC2 (IgG2b) binds D3 epitopes of CII (687–698; RGAQGPPGATGF), and UL1 (IgG2b) binds U1 epitopes of CII (494–504; GLVGPRGERGF). Antibodies were purified using γ-Bind Plus affinity gel matrix.
The expression of recombinant glutathione S-transferase (GST)–IdeS in Escherichia coli has been described previously (13). Briefly, the fusion protein was purified on Glutathione Sepharose 4B Fast Flow (Amersham Biosciences, Uppsala, Sweden) according to standard protocols, dialyzed against PBS, and sterilized with a 0.22 Millex-GP filter unit (Millipore, Bedford, MA). The GST tag was removed with factor Xa, Glutathione Sepharose 4B Fast Flow, and X-Arrest Sepharose (Amersham Biosciences). The protein concentration was determined using Advanced Protein Assay Reagent (Cytoskeleton, Denver, CO), and the samples were checked for impurities by SDS-PAGE and Limulus endotoxin assay (Cape Cod, East Falmouth, MA).
Arthritis was induced using a CII-specific mAb cocktail as described previously (6, 20). Mice were injected intravenously (IV) with 0.4–0.5-ml volumes of antibodies. IdeS in PBS was injected IV. Lipopolysaccharide (LPS) from E coli O55:B5 (25 μg/mouse) was injected intraperitoneally when indicated, to enhance arthritis. Mice were examined daily and scored for arthritis using a scoring system based on the number of inflamed joints in each paw, with inflammation being defined by swelling and redness (21). Paws from each group of mice (4–5 per group) were dissected on day 15, fixed, decalcified, dehydrated, and embedded in paraffin. Sections (6 μm) were stained with hematoxylin and eosin using a standard protocol.
To induce CIA, DBA/1 mice were injected at the base of the tail with 100 μg of rat CII emulsified 1:1 in Freund's complete adjuvant (CFA; Difco, Detroit, MI) in a 100-μl volume. Arthritis was monitored at least twice per week. IdeS was injected IV on the indicated days. Mice were bled before IdeS injection and 3 days after the final IdeS injection. CII-specific IgG and IgG2a antibody levels in sera were determined by quantitative enzyme-linked immunosorbent assay (22). For induction of chronic arthritis, (BALB/c × B10.Q)F1 mice were injected at the base of the tail with 100 μg of rat CII emulsified 1:1 in CFA on day 0 and boosted on day 35 with 50 μg of rat CII in Freund's incomplete adjuvant (Difco). Arthritis was scored for a minimum of 210 days. Mice that developed chronic arthritis (defined as severe arthritis for a minimum of 120 days) were selected for injection of an anti-CII antibody cocktail (M287 plus CIIC1) 9 months after the initial CII immunization. At the time of induction of arthritis relapse with antibodies, none of the mice exhibited active arthritis. It is possible to induce relapses in mice with chronic arthritis by injecting anti-CII antibodies alone, and LPS was not injected in these mice.
The Mann-Whitney U test and chi-square test were used for statistical comparisons of the severity and incidence of arthritis; data were analyzed using StatView, version 5.0.1. P values less than 0.05 were considered significant.
SDS-PAGE with specific purified mAb showed that recombinant IdeS specifically cleaved the IgG2a/c and IgG3 subclasses of mouse IgG, but not other isotypes, in vitro (Figure 1A). To ascertain the therapeutic potential of IdeS in antibody-mediated autoimmune diseases, we established an antibody-induced arthritis mouse model using a cocktail of CII-specific IgG2a mAb binding to 2 dominant B cell epitopes of CII (J1 and C1I). IdeS cleaved the arthritogenic mAb cocktail when injected 3 hours before the antibody injection, and none of the mice receiving IdeS treatment developed arthritis (Figure 1B). Similarly, there was no clinical or histologic evidence of arthritis development if IdeS was given 3 hours after the mAb injection (data not shown). In contrast, 83% of the mice in the non–IdeS-treated group developed arthritis (P < 0.01 versus IdeS-treated mice) with a mean ± SEM maximum score of 17 ± 5 (P = 0.01 versus IdeS-treated mice). Massive infiltration of immune cells, pannus formation, fibrin deposition, and distinct bone and cartilage erosions were observed in the antibody-injected, non–IdeS-treated group (Figure 1C), whereas IdeS-treated animals showed completely normal joint architecture with no cellular infiltration (Figure 1D).
IdeS may have unknown protein interactions and effects in vivo, apart from its IgG2a/c cleavage capacity. To determine whether its mechanism of action is dependent on the specific cleavage of IgG2a and not on some other unknown effect in vivo, we used arthritogenic anti-CII IgG2b antibodies (which were not cleaved by IdeS) to induce CAIA, and treated the mice with IdeS. Three different IgG2b antibodies (M2139, UL1, and CIIC2) were injected into B10.RIII mice (Figures 2A and B). IdeS had no effect in this CAIA model, excluding other mechanisms that could explain its antiarthritic effect.
To explore the role of the IdeS application site in inhibition of arthritis, we administered either systemic (IV) or intraarticular injections of IdeS after mAb transfer in naive mice. The mice were administered an anti-CII IgG2a cocktail plus LPS and subsequently injected with IdeS at various time points (3, 12, 24, 48, or 72 hours). In mice not injected with IdeS, clinical arthritis was seen as early as 24 hours after anti-CII injection. Injection of antibodies before the onset of arthritis (at 3 hours or 12 hours) almost completely blocked arthritis development, and treatment started after the onset of arthritis (at 24 hours and 48 hours) stopped the further progression of arthritis. In mice treated at 72 hours (i.e., 1–2 days after the onset of clinical arthritis), there was no significant effect on the disease course (Figure 3A). Since antibodies are bound to cartilage within 3 hours after injection (23), we concluded that IdeS could block the further escalation of the inflammatory process even if the antibodies had already reached the joint, but that it could not affect the established inflammatory process beyond a later point.
The above results suggested that IdeS might have a very efficient direct effect on antibodies in the joints. To investigate this, we injected IdeS locally into the talocrural joint of 1 hind paw (Figure 3B). In mice that received this injection 3 hours after the injection of anti-CII antibodies, the treatment completely blocked the arthritis, and surprisingly, arthritis was blocked in all joints, as was observed with IV injection of IdeS at the same dose. Injection of half the dose at 3 hours and the other half at 24 hours still blocked arthritis in the injected joints but allowed the development of arthritis in other joints, suggesting that local injection of IdeS could be more efficient in preventing arthritis development, and a sufficient concentration of IdeS is needed to block spreading of arthritis to other joints.
Next we performed experiments to determine whether IdeS had a dominant effect on arthritis protection, i.e., whether IdeS treatment could block arthritis induced by a mixture of IgG antibodies that included arthritogenic IgG2b antibodies, which are resistant to IdeS cleavage. In these experiments we used an IgG2b antibody specific for the J1 epitope (M2139) in combination with the IgG2a antibody CIIC1 to induce arthritis. Treatment with IdeS at 100 μg, or even in doses as low as 10 μg, led to suppression of arthritis induced with both M2139 and CIIC1 (P < 0.05, cumulative incidence of arthritis in 100 μg IdeS–treated versus non–IdeS-treated mice) (Figure 4), demonstrating that IdeS has the capacity to suppress arthritis even in the presence of arthritogenic antibodies that cannot be cleaved.
The above findings suggested that it might be possible to treat classic CIA with IdeS. In CIA, CII-reactive antibodies of all IgG subclasses are arthritogenic, and the antibodies in serum are mainly dominated by IgG1, with high levels also of IgG2a/c and IgG2b. Injection of IdeS on days 22, 26, and 30, starting just before the onset of arthritis in DBA/1 mice immunized with rat CII, led to a delay in arthritis onset, with reduced severity of the disease during the initial stages (Figure 5A). However, the effect was not sustained, and IdeS did not affect the final severity of established disease on day 35, in spite of repeated IdeS treatment (data not shown). Although overall concentrations of anti-CII antibody levels were found to be reduced, similar to concentrations of anti-CII IgG2a antibodies (Figure 5B), we still detected substantial levels of anti-IgG antibodies after IdeS treatment. IdeS cleaved circulating and cartilage-bound IgG2a antibodies in animals with CIA, but did not affect new antibody synthesis. Furthermore, IdeS activity persisted only for 3 hours in vivo (data not shown). Hence, there was a reduction, but not a complete abolition, of anti-CII IgG2a. The transient effect of IdeS on CIA suggests that cleavage of only IgG2a antibodies significantly delays, but is not enough to completely block, the development of CIA. This indicates that other IgG isotypes or other mechanisms independent of antibodies have a role in the arthritis pathogenesis.
To investigate whether IdeS could be efficacious in the treatment of chronic established, but relapsing, arthritis, we used a newly developed arthritis model in the (BALB/c × B10.Q)F1 mouse strain. Immunization of (BALB/c × B10.Q)F1 mice with rat CII in CFA leads to the development of arthritis, which later exhibits a chronic relapsing pattern with no healing (Nandakumar KS, et al: unpublished observations). The relapses in this model are unpredictable but can be synchronized with injection of anti-CII antibodies. To test whether IdeS treatment could block such an antibody-induced relapse during the chronic stage of arthritis, we injected anti-CII IgG2a antibodies into (BALB/c × B10.Q)F1 mice with chronic arthritis that had been present for at least 120 days. IdeS treatment of these mice significantly suppressed the relapses (P < 0.005 for cumulative incidence, P < 0.001 for mean maximum score) (Figure 6), demonstrating that IdeS may also have an effect on chronic relapsing arthritis.
We have identified a new therapeutic concept for the treatment of antibody-mediated arthritis. Injection of the streptococcal enzyme IdeS cleaves IgG2a antibodies in vivo and protects mice against collagen antibody–induced arthritis. IdeS also delays the development of collagen-induced arthritis, reduces the severity of arthritis during the initial stage of disease, and protects against antibody-induced relapses in chronic arthritis.
IdeS modulates immune responses by cleaving the heavy chains of IgG in the hinge region. We have shown that mouse IgG2a is rapidly degraded into Fab and Fc fragments in the blood after IdeS administration. Although circulating Fab and Fc fragments are cleared from the blood circulation within hours (24), Fab fragments from cleaved autoantibodies have a retained affinity against self structures, and thus can block epitopes that otherwise would have been recognized by intact autoantibodies. In contrast to other IgG-cleaving proteinases (e.g., streptococcal pyrogenic exotoxin B, papain, and pepsin), IdeS has an extreme proteolytic specificity, and apart from IgG, no other substrates have been identified. This extreme specificity not only contributes to the enzyme's lack of toxicity, but might also enhance its therapeutic effects; IgG autoantibodies usually are pathogenic, whereas IgM appears to suppress the development of autoimmune diseases (25). However, similar to the classic streptococcal cysteine proteinase streptococcal pyrogenic exotoxin B, IdeS contains an RGD motif (12, 13), which is involved in the interaction of IdeS with vitronectin (αVβ3) and platelet receptors (αIIβ3) (26) and likely confers other, as-yet-unknown, properties to the enzyme.
To determine the specificity of IdeS activity in vivo in cleaving IgG2a antibodies and also to ascertain whether other properties of IdeS could account for its arthritis-inhibitory activity, we established 2 different CAIA models using IgG2a mAb (binding to CII epitopes J1 and C1I) and IgG2b mAb (binding to CII epitopes J1, D3, and U1) and administered IdeS treatment in both of these models. IdeS specifically cleaved the IgG2a mAb cocktail but not the IgG2b mAb cocktail, demonstrating that specific cleavage of mouse IgG2a mAb in vivo by IdeS led to inhibition of arthritis and ruling out other properties of IdeS in modulating arthritis. However, we did not test the cleaving capacity of IdeS on IgG3 mAb in vivo.
Antibodies might play an important role in inducing relapses in arthritis. In the CIA model in the mouse, the antibody response to the predominant B cell epitopes of CII (C1, J1, and U1) correlated well with arthritis, and this association was maintained throughout the chronic relapsing course of the disease (19). After observing that IdeS completely inhibited arthritis in a model of acute arthritis in naive mice, we used an mAb cocktail to induce relapses in mice with chronic arthritis, in order to investigate the effectiveness of IdeS in a model that mimics RA. It is important to note that antibodies to the major epitopes of CII are positively associated with RA (27). These dominant CII epitopes are also shared between mice, rats, and humans. It is also important to note that IdeS will cleave any IgG2a antibody in the mouse, irrespective of the specificity. Further extrapolating our findings to humans, it is of considerable interest that IdeS cleaves all human IgG subclasses (13). Thus, any IgG antibody involved in attacking joints or any other self tissue will be cleaved by IdeS.
IdeS is active in the host only for ∼3 hours. Thus, cleavage of the antibodies is only transient, and new antibodies will be produced. Moreover, results from the CIA experiments suggest that although circulating and cartilage-bound antibodies are cleaved by IdeS, antibody synthesis is not affected at all. Hence, IdeS treatment does not result in a permanent lack of antibodies to fight infections. Plasmapheresis and immunoadsorption are normally used to remove pathogenic antibodies in many autoimmune diseases, and IdeS presents an alternative treatment strategy in such conditions.
In the mouse, IgG cleavage by IdeS is extremely rapid, efficient, and specific, and the IdeS treatment was well tolerated in the mice in the present study; none of the treated mice exhibited any easily observable signs of distress, such as fever, diarrhea, weight loss, or passive behavior, and there was no mortality. Furthermore, local injection of a 10-fold higher concentration of IdeS than the effective concentration needed to block arthritis in the talocrural joint (i.e., 100 μg rather than 10 μg) proved to be effective not only in blocking arthritis in the injected joint, but also in preventing the spread of arthritis to other joints. Thus, this novel concept of treatment based on cleaving of IgG antibodies using IdeS may be of clinical importance, since it would provide a new way to rapidly block the action of pathogenic autoantibodies.
IgG autoantibodies binding to human antigens contribute to pathogenesis that can lead to various clinical conditions. Patients who have antibodies that are agonistic (Graves' disease), neutralizing (acquired factor VIII deficiency), complement and macrophage activating (myasthenia gravis, Goodpasture's syndrome, autoimmune bullous skin diseases), depleting (autoimmune hemolytic anemia, idiopathic thrombocytopenic purpura), or immune complex associated (RA, systemic lupus erythematosus) could benefit from treatments targeting elimination of pathogenic IgG. The demonstration that IdeS inhibits arthritis by cleaving IgG2a antibodies in vivo, and the cleavage of all subclasses of human IgG (13) but not IgM, IgA, IgD, or IgE, present opportunities for further exploring the therapeutic potential of bacterial enzymes, such as IdeS, that are used by bacteria for immune evasion.
Dr. Nandakumar 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. Nandakumar, Björck, Holmdahl.
Acquisition of data. Nandakumar, Johansson.
Analysis and interpretation of data. Nandakumar, Johansson.
Manuscript preparation. Nandakumar, Johansson, Björck, Holmdahl.
Statistical analysis. Nandakumar.
Animal experiments. Nandakumar.
We thank Emma Mondoc for technical assistance with the histologic studies and Carlos Palestro for care of the study animals.