No Evidence for a Role of FcγRIIB in Suppression of In vivo Antibody Responses to Erythrocytes by Passively Administered IgG

In a paper in this issue of Scandinavian Journal of Immunology, Sinclair addresses the role of Fc-signalling in the modulation of immune responses by passive antibody [1]. This is a broad field, encompassing aspects of immune responses from autoimmune disease models to cell signalling in vitro. A substantial part of the paper deals in a critical way with recent results from our laboratory concerning the ability of immunoglobulin (Ig)G to feedback-regulate in vivo antibody responses in mice [2,3] and we will limit our reply to this area. In a field as complicated as antibody feedback regulation, it is important to define exactly what is discussed. Antibodies of many classes have the capacity to modulate antibody responses, either positively or negatively [4]. IgG has a dual role, suppressing the response to particulate antigens such as erythrocytes [5–9], and enhancing the response to soluble antigens such as keyhole limpet hemocyanine (KLH), bovine serum albumin (BSA), and ovalbumin (OVA) [3,9–11]. The same monoclonal IgG anti-trinitriphenyl (anti-TNP) can suppress or enhance, depending on whether TNP is conjugated to an erythrocyte or a soluble antigen [9,11]. Considering the strikingly different structure of these two types of immune complexes this is not surprising. While an IgG molecule is much smaller than an erythrocyte (10 nm versus 7 µm in diameter), IgG is on average twice as large as BSA. When discussing IgG-mediated immunoregulation, it is therefore vital to keep apart studies using particulate versus soluble antigens. Major differences between in vitro and in vivo systems for studying antibody responses exist (e.g., lack of lympoid organ structure in vitro) and unless specifically stated, only in vivo systems will be considered here. Why IgG has the capacity to completely inhibit a response to SRBC is a question that has been debated extensively over the years. Hypotheses put forward are: (i) epitope masking by IgG, preventing recognition of the antigen by B cells, (ii) effective elimination of IgG-coated antigen via Fc-dependent phagocytosis, and (iii) central inhibition of B-cell activation via cocrosslinking of the B-cell receptor and the inhibitory FcγRIIB. Sinclair argues that the mechanism behind the IgG-mediated suppression of erythrocyte-specific responses was resolved 30 years ago and that his so called Tripartite Inactivation Model ((iii) see above) explains the phenomenon [12]. Some textbooks and reviews also state for a fact that the mechanism behind the IgG-mediated suppression has been proven to be FcγRIIB-mediated negative signalling to B cells. We do not share this opinion. Although this receptor undoubtedly plays an inhibitory role in many experimental systems, there are many observations calling for scepticism about its role in inhibition of erythrocyte responses by passively administered IgG. Firstly, data regarding the Fc-dependence of suppression are conflicting. The efficiency of F(ab′)2 fragments varies [13–16] and, as discussed previously [1], the technical difficulties working with F(ab′)2 fragments will always leave this type of experiments open to criticism. A finding used to support Fc-dependence has been the ability of IgG, specific for one epitope on the erythrocyte, to suppress the response also to other epitopes, not recognized by IgG [6,7,9,11]. On the other hand, observations of epitope-specific suppression exist [17,18]. The conclusion that the nonepitope-specific suppression implicates Fc-dependence is probably only correct in situations where IgG binds to the erythrocyte at a low density. IgG, as well as F(ab′)2 fragments, binding to one epitope on an erythrocyte will most likely sterically hinder B cells from accessing neighbouring epitopes. When the density of IgG on the erythrocyte surface is high, steric hindrance may lead to nonepitope specific suppression independently of the Fc-portion. Another reason for questioning the role of FcγRIIB is that both monoclonal IgG3 [7,8] and IgM [7] can suppress although neither isotype binds to the receptor. More recently, we have shown that monoclonal IgE is suppressive [2] also in the absence of FcγRIIB and FcγRIII [19], receptors also reported to bind IgE. A few years ago, mice lacking FcγRIIB (FcγRIIB-/-), as a result of gene targeting, became available. We set out to test whether IgG could suppress SRBC-specific responses in these animals. Wildtype and FcγRIIB-/- mice were immunized with 0.1, 0.4, 2, or 50 µg of IgG (polyclonal IgG anti-SRBC, monoclonal TNP-specific IgG2a, IgG2b or IgG1) together with SRBC or SRBC-TNP. IgG was administered either 1 h before antigen or in preformed complexes with the antigen. The SRBC-specific primary response in wildtype and FcγRIIB-/- mice was equally well suppressed in a series of 14 experiments [2]. In a follow-up we studied the ability of IgG to suppress secondary IgM- and IgG-responses and the ability of IgG administered 2 days after SRBC to suppress. No difference between wildtype and FcγRIIB-/- mice were observed [19]. Thus, in all in vivo experiments performed so far, IgG inhibits antibody responses to SRBC to the same extent in the absence of FcγRIIB as in its presence. There are various ways of interpreting these results. We think that the most logical interpretation is that FcγRIIB does not play a major role in the suppression of SRBC-responses by passively administered IgG. Sinclair argues against this view along two lines. Firstly he postulates that if we had used other experimental conditions, we would have found evidence of FcγRIIB-involvement. This type of argument will by necessity always be impossible to refute. However, as detailed above, under all experimental conditions we tested, no influence of FcγRIIB-/- was found. Secondly, Sinclair suggests a number of alternative explanations for our inability to demonstrate a role for FcγRIIB, e.g. unspecific suppression by contaminating transforming growth factor (TGF)-β or interleukin (IL)-4-mediated prevention of FcγRIIB-signalling. We have no reason to believe that (all) our antibody preparations contain TGF-β. Besides, that would presumably inhibit the response also to horse erythrocytes (HRBC) which were always coinjected with IgG and SRBC as a specificity control in our experiments. No suppression of the HRBC-specific response was seen [2]. Moreover, IgG-mediated suppression of SRBC-specific antibody responses is unperturbed in transgenic mice, overexpressing the IL-4 (Heyman, unpublished observation), contrary to Sinclair's latter speculation. There is to date no evidence for involvement of FcγRIIB in suppression of in vivo antibody responses to SRBC by passively administered IgG. Its role has never been directly demonstrated, but merely inferred from observations suggesting Fc-dependence. Even if these observations were thought to be indisputable (which they are not), they would only indicate that any FcγR, or the complement system, was involved and not specifically point to FcγRIIB. Our demonstration of efficient suppression in FcγRIIB-/- mice, however, proves that suppression can take place in the absence of this receptor. Notably, it does not exclude that FcγRIIB could play a crucial role during other experimental conditions. It is also possible that it acts in a redundant fashion simultaneously with the FcγRIIB-independent mechanisms observed [2,19]. These possibilities are however hypothetical. The inhibitory role of FcγRIIB has been convincingly demonstrated in many other experimental systems and we do not question these findings. There is ample evidence of its capacity to inhibit cell activation via receptors containing immunoreceptor tyrosine-based activation motifs (ITAM) in vitro[21]. In vivo, FcγRIIB-/- mice have a 5-fold higher antibody response to SRBC and TNP–KLH and an augmented anaphylactic response [20]. These animals spontaneously develop lupus [22] and are more susceptible to experimentally induced autoimmune diseases than wild-type mice [23–25]. We have shown that FcγRIIB inhibits the ability of IgG1, IgG2a and IgG2b to enhance the antibody response to soluble proteins [3]. In wild-type mice, the response to immunization with e.g. IgG2a anti-TNP/BSA–TNP was 10-fold higher than the response to BSA–TNP alone. In the FcγRIIB-/- mice, the response to the same immune complexes was 268-fold higher than the response to BSA–TNP (i.e. 27-fold higher than the response to IgG2a anti-TNP/BSA–TNP in wild-type mice). In other words, the presence of FcγRIIB in wild-type mice did not prevent IgG from upregulating the antibody response, but set the upper level for the magnitude of the response. We believe that this observation gives us an important clue to the way the receptor acts as an in vivo regulator of antibody responses. Most likely, FcγRIIB primarily acts to prevent an overshoot of antibody-production, and not to completely prevent a response. The IgG binds to other FcγRs than FcγRIIB and also activates the complement system which leads to an array of possible mechanisms for suppression of antibody responses. As mentioned before, determining the Fc-dependence using F(ab′)2 fragments or epitope-specificity of suppression is inherently difficult. In order to avoid these problems, we investigated the ability of the IgG to suppress in mice lacking FcγRI + FcγRIII (owing to the deletion of the common FcRγ subunit, FcRγ-/-), FcγRIIB + FcγRI + FcγRIII (double knockouts: FcRγ-/- x FcγRIIB-/-) or the neonatal FcRn (owing to a lack of the associated β2m chain, β2m-/-). In these experiments mice were immunized with IgG either 1 h before SRBC or with IgG in preformed complexes with SRBC. No difference between suppression of primary IgM-responses in wild-type and knockout mice was found [2]. Thus, the suppression took place in the absence of all known FcγRs. In the same study we showed that F(ab′)2 fragments of a TNP-specific monoclonal IgG2a were efficient suppressors of primary responses to SRBC–TNP [2]. Based on these findings, and the realization that most in vivo data would be compatible with the idea, we have suggested that the major explanation for IgG-mediated suppression of SRBC-responses in vivo is epitope-masking. This means that passively administered antibodies bind to the antigen and hides it for specific B cells, thereby preventing them from responding. Our arguments have been extensively presented in previous work [2,4,26] and are briefly summarized in Table 1. There is convincing evidence for an inhibitory role of FcγRIIB in various in vivo and in vitro models, but in the special case of in vivo feedback suppression of murine erythrocyte-specific responses by passively administered IgG, such evidence is lacking. Studies aimed at specifically investigating the role of FcγRIIB in feedback suppression have failed to demonstrate an effect [2,19]. This does not exclude that the receptor operates in a redundant fashion, simultaneously with e.g. epitope-masking, or that the receptor during other experimental conditions may prove to be the sole mechanism for feedback suppression. This, however, remains to be proven. In contrast, even excluding all data based on the use of F(ab′)2 fragments or epitope-specificity, there is clear evidence of suppression in the absence of FcγRIIB [2, Karlsson et al., unpublished observations] or other FcγRs [2]. The demonstration of suppression by monoclonal IgG3 [7,8] and IgM [7], neither of which bind FcγRIIB, also argues against a role of this receptor. We find that the epitope-masking model best explains the majority of available observations. As most scientific hypotheses, it may have to be modulated/abandoned as new experimental data become available. Work in the authors' laboratory was supported by Ellen, Walter and Lennart Hesselman's Foundation, Hans von Kantzow's Foundation, King Gustaf V:s 80 years Foundation, the Swedish Medical Research Council, the Swedish Foundation for Health Care Sciences and Allergy Research and by a grant from the 'Network for Inflammation research' funded by the Swedish Foundation for Strategic Research.