Effect of a putative B cell superantigen on complement.

Staphylococcal protein A (SPA), a cell wall component of S. aureus, binds not only to the Fc fragment of IgG, but also via an alternative site to the Fab fragments of immunoglobulins (Igs) independent of heavy chain isotype. ‘ J In man, this binding is restricted to Igs with vH3 heavy chains and occurs at a site(s) outside the conventional antigen-binding r e g i ~ n . ~ . ~ An increasing body of evidence has indicated that SPA shares many analogous properties with T cell superantigens, and it has thus been characterized as a B cell ~uperant igen .~ ,~ Unlike a T cell superantigen, however, a B cell superantigen has the ability to react with potentially large amounts of soluble B cell antigen receptors in the serum. Encountering such a large reservoir of reactive Igs could have deleterious effects on the host, particularly if this interaction leads to activation of the complement cascade with resultant tissue inflammation. The addition of SPA to serum has been shown to cause activation of the complement cascade. This activation has been attributed to the binding of the classical site on SPA to the Fc region of IgG.5 It is also possible, however, that complement activation is caused by the interaction of SPA with Fab regions of vH3+ Igs. The generation of such complement-activating “immune complexes” by SPA may represent a novel and important biologic activity of a B cell superantigen. Therefore, we sought to determine in vitro if the interaction of the alternative site on SPA with the Fab region of Ig molecules causes complement activation. Using a total hemolytic complement assay, we demonstrated that SPA, abrogated of its IgG Fc-binding activity by hyperiodination (mod-SPA), causes complement consumption when incubated with human serum (TABLE 1). To further test this hypothesis, we determined whether the interaction of SPA with polyclonal IgM or a panel of monoclonal IgM proteins (representative of human VH gene families) led to Clq binding in an ELISA. These proteins were first analyzed in an SPAbinding ELISA to determine which bind to SPA. Our results (data not shown) demonstrated that polyclonal IgM bound SPA as well as four out of six VH3+ IgM proteins. No proteins from other VH gene families had binding activity. Because only a subset of vH3+ Ig molecules is known to bind SPA, it was anticipated that some of our VH3+ IgM monoclonals would also not bind SPA.3 To determine if the interaction of the IgM proteins with SPA led to binding of Clq, biotinylated SPA was incubated with either human polyclonal IgM or the monoclonal IgM