Several different cell populations and mechanisms of activity have been shown to be operative in IVIg-mediated suppression of a wide variety of autoimmune diseases [2-5, 33-36]. Among these different pathways, IVIg glycoforms rich in terminal sialic acid residues have been suggested to be critical for IVIg activity in several but not all mouse model systems [6, 7, 11-13, 37]. A variety of studies suggest that myeloid cells are modulated via sialic acid rich IgG glycovariants by interaction with cell surface receptors such as SignR1 in mice and DC-Sign in humans, providing a plausible explanation for the reduced effector cell-dependent tissue inflammation [6, 8, 11, 12]. Whereas it is known since a long time that B-cell development and responses may be modulated via the presence of anti-idiotypic antibodies, which are able to recognize the B-cell receptor in the IVIg preparation, it was noted only recently that FcγRIIB expression on B cells becomes upregulated by IVIg infusion in mice and humans [18, 19, 38, 39]. As B cells, besides their obvious role in being the source of autoantibody production, can also have strong direct antiinflammatory and immunomodulatory functions, we set out to test if they play a direct role in IVIg-mediated suppression of ITP and RA. Indeed, a previous study showed that B cells may have the capacity to recognize sialic acid rich IgG via CD22 . Moreover, IVIg was suggested to induce apoptosis in primary B cells and in different B cells lines most dramatically 24 h after coincubation with IVIg in vitro . In vivo, however, we did not observe a reduction in peripheral B-cell counts, despite the capacity of IVIg to bind to B cells and other cell types in the blood consistent with other studies showing that the proliferation of human B cells is not influenced by IVIg in vitro. More importantly, this binding was independent of the presence of terminal sialic acid residues and did not require FcγRs or CD22 suggesting that other as yet unidentified molecules on mouse peripheral blood cells can recognize IVIgs. Apart from the induction of apoptosis, IVIgs might be able to stimulate B cells to dampen an ongoing innate immune response via release of ILx02010;10 as shown to be critical for dampening EAE [20, 21]. However, IVIg infusion did not result in a dramatic systemic cytokine release, which is consistent with previous studies showing that IVIg activity was preserved in mouse strains deficient in a variety of cytokines or functional cytokine receptors in different models of mouse ITP [35, 40]. Finally, neither B cells nor CD22 were involved in IVIg-dependent suppression of autoantibody-induced ITP and RA, in which terminal sialic acid residues were shown to be absolutely required for the antiinflammatory activity. It should be noted, that the previous study showing an involvement of CD22 in binding to sialic acid rich IgG glycoforms used intact IVIg preparations to enrich for IgG fractions rich in terminal sialic acid residues . As about 20% of serum IgG can have an additional and more accessible sugar moiety attached to the IgG variable fragment it seems likely that predominantly F(ab)2 sialylated IgG glycoforms were enriched, which may bind to CD22 but do not have an antiinflammatory activity as described by others [10, 11, 41-43]. This may explain at least in part the lack of a role of CD22 in the immunosuppressive function of the immediate antiinflammatory activity of IVIgs in vivo.
Nonetheless, human and mouse B cells have clearly been shown to respond to IVIg therapy by upregulation of FcγRIIB expression for example. As FcγRIIB is a critical checkpoint of humoral tolerance in mice and humans this may indicate that IVIgs might rather have a delayed effect on B cell and antibody responses by raising the threshold for B-cell activation. Moreover, isolated cross-linking of FcγRIIB by immune complexes on plasma cells has been suggested to induce apoptosis and open new niches for de novo generated plasma cells [44-46]. Moreover, IVIg was shown to modulate TLR9-dependent activation of B cells and interfere with B-cell dependent antigen presentation [23, 47]. Therefore, future studies focusing on IVIg-mediated modulation of B-cell development and antibody responses will be necessary to understand if IVIg, and especially the sialic acid rich fraction, affects humoral immune responses in vivo. Supporting this notion, previous reports suggested that the anti-idiotype specific fraction in the IVIg preparation can affect B-cell development and function .
In general, care should be taken when trying to transfer data from mouse model systems directly to humans. Although some IVIg-mediated effects, such as the upregulation of FcγRIIB on myeloid cells and B cells can be observed in mice and humans, other players involved in the antiinflammatory activity are quite different. This may be exemplified by the cellular expression pattern of the molecules critical for recognition of sialic acid rich IgG in mice and humans. Whereas SIGNR1 is present on macrophage subpopulations in mouse lymphoid tissue, the human orthologue DC-SIGN shows a more DC-restricted expression pattern [6, 12, 48]. Along the same lines, the genetic background of different mouse strains critically affects the penetrance of autoimmunity, which may explain why FcγRIIB deletion reduced the antiinflammatory activity of IVIg in some but not all strains of mice [8, 10, 16, 17, 49-51]. Ultimately, integrating our knowledge from both, mouse in vivo and human in vitro studies, may allow us to find common denominators of IVIg-mediated suppression of autoimmune pathology.