ADP-ribosylating bacterial enzymes for the targeted control of mucosal tolerance and immunity.

The questions of whether mucosal tolerance and IgA immunity are mutually exclusive or can coexist and whether they represent priming of the local immune system through the same or different activation pathways are addressed. Two strategies were attempted: the first using cholera toxin (CT) or the enzymatically inactive receptor-binding B subunit of CT (CTB), and the second using CTA1-DD or an enzymatically inactive mutant thereof, CTA1R7K-DD. The CTA1-DD adjuvant is a fusion protein composed of the ADP-ribosylating part of CT, CTA1, and DD, which is derived from Staphylococcus areus protein A and targets the molecule to B cells. Here, we provide compelling evidence that delivery of antigen in the absence of ADP ribosylation can promote tolerance, whereas ADP-ribosyltransferase activity induces IgA immunity and prevents tolerance. By linking antigen to the ADP-ribosylating enzymes we could show that CT, although potentially binding to all nucleated cells, in fact, bound preferentially to dendritic cells (DCs) in vivo. On the other hand, DD-bound antigen was distinctly targeted to B cells and probably also to follicular dendritic cells (FDCs) in vivo. Interestingly, the CT and CTA1-DD adjuvants gave equally enhancing effects on mucosal and systemic responses, but appeared to target different APCs in vivo. CT- or CTB-conjugated antigen accumulated in mucosal and systemic DCs. Whereas only CT promoted an active IgA response, CTB induced tolerance to the conjugated antigen. Following intravenous injection of CT-conjugated antigen, DCs in the marginal zone (MZ) of the spleen were selectively targeted. Interestingly, CTB delivered antigen to the same MZ DCs, but failed to induce maturation and upregulation of costimulatory molecules in these cells. Thus, ADP-ribosylation was necessary for a strong enhancing effect of immune responses following CT/CTB-dependent delivery of antigen to the MZ DCs. Moreover, using CTA1-DD, antigen was targeted to the B cell follicle and FDC in the spleen after intravenous injection. Only active CTA1-DD, but not the inactive mutant CTA1R7K-DD, provided enhancing effects on immune responses. By contrast, antigen delivered by the CTA1R7K-DD stimulated specific tolerance in adoptively transferred T cell receptor transgenic CD4(+) T cells. Whether targeting of B cells suffices for tolerance induction or requires participation of DCs remains to be investigated. With CT we found that enzyme-dependent modulation of DCs affects migration, maturation, and differentiation of DCs, which resulted in CD4(+) T cell help for IgA B cell development. On the contrary, antigen presentation in the absence of ADP-ribosylating enzyme, as seen with CTB or CTA1R7K-DD, appears to expand specific T cells to a similar extent as enzymatically active CT or CTA1-DD, but fails to recruit help for germinal center (GC) formation and the necessary expansion of activated B cells. Also, the CD41 T cells that are primed in a suboptimal, tolerogenic, fashion do not migrate to the B cell follicle to provide T cell help. Thus, ADP-ribosylating enzymes may be used to selectively control the induction of an active IgA response or promote the development of tolerance. In particular, on the targeted APC, modulation of the expression of costimulatory molecules, CD80, CD86, CD83, and B7RP-1, plays an important role in the effect of the ADP-ribosylating CTA1-based adjuvants on the development of tolerance or active IgA immunity. For example, the expression of CD86 in vivo was a prominent feature of the enzymatically active CT or CTA1-DD adjuvants. By contrast, CD80 expression appeared not to be important in CTA1-augmented APCs for an adjuvant function.