Toll-like receptors and antibody responses (Reply)

Abstract

Using a range of different experimental approaches (and not just analyses of MyD88-deficient mice), we showed that, contrary to the prevailing view, a T-helper response is not sufficient for the generation of optimal T-dependent antibody responses, and that Toll-like-receptor (TLR) signalling in B cells is also required¹. Similar results have since been obtained in human B cells². Nemazee et al.³ now question our conclusions¹.

Main

We showed that the requirement for TLR signalling in antibody responses depends on the antibody isotype¹: the IgM and IgG1 isotypes are largely, but not completely, TLR-dependent; IgG2 isotypes are entirely TLR-dependent; and IgE and IgA responses are TLR-independent. But in any event, our main discovery was that TLR signalling has to happen in B cells in order for the optimum T-cell-dependent antibody response to be generated. It was not, as claimed by Nemazee et al.³, simply that TLR-dependent B-cell responses depend on TLR signalling.

Figure 1 of Nemazee et al.³ shows responses obtained under suboptimal conditions, without essential comparisons and controls¹: for example, their antibody responses would seem negligible if compared with the response induced in the presence of TLR ligands. The results do not therefore present a meaningful counterargument to our conclusions.

Nemazee et al. claim that TLR signalling is not essential for responses to “typical vaccination”, which very much depends on what is meant by ‘typical’. Most successful vaccines, such as attenuated pathogens, certainly do trigger TLR signalling, which is one of the reasons why they are successful. The subunit vaccines, on the other hand, generally do not work; the few that do, do so because they either activate TLRs or some TLR-equivalent component of the innate immune system.