A feature of the immune system of critical importance is its ability to mount a much stronger antibody response the second time a pathogen or antigen is encountered (1). This property underlies the need for "booster" doses to ensure the effectiveness of vaccination. The booster effect derives in part from the generation, by the primary response, of expanded clones of memory B cells (MBCs). Upon re-exposure to antigen, MBCs rapidly proliferate and differentiate into plasma cells, generating high titers of serum antibody over a short period of time. Because most MBCs that respond to boosting have also undergone affinity maturation in germinal centers (GCs) during the primary response (2), MBC-derived antibody has both higher affinity and higher cross-variant breadth than primary antibody (3).
A common assumption in the vaccinology field has been that, in addition to forming plasma cells, MBCs will also generate secondary GCs upon boosting with high efficiency, allowing them to re-evolve their immunoglobulins to adapt to variant strains of a pathogen (4-6). This assumption forms the basis of multiple attempts to guide B cell clones towards broad reactivity to influenza or HIV by iteratively recruiting them to germinal centers by sequential immunization.
Recent work using genetic tracing of memory B cell clones in mice questions this assumption (7). Using fate-mapping of primary-GC-derived B cells, results showed that mouse MBCs are remarkably inefficient at forming secondary GCs, which instead consist predominantly (\>90%) of cells derived from naïve precursors engaged only by the boost but not by the prime (2). Partly in contrast to the mouse findings, a recent study using fine-needle aspirates (FNA) to sample vaccine-draining lymph node GCs in healthy humans immunized with a quadrivalent influenza vaccine showed that MBC participation in recall GCs can vary depending on the individual sampled, ranging from approximately 5%, a proportion similar to that found in mice, to up to \~65% in one individual (8). This wide range was not recapitulated in our mouse models. A key unknown in this study was the degree of prior exposure of each individual to influenza antigens, via either infection or vaccination. Long histories of exposure can generate influenza-specific MBC compartments in blood that represent \>1% of all B cells (9). On the other hand, simpler exposure regimens, such as HPV, tetanus, and HBV vaccination, generate much lower frequencies of antigen-specific MBCs, in the high tens to low hundreds per million B cells, comparable to frequencies found after single immunization or infection in mice (10-14).
There is a hypothesize that the wide range of MBC re-entry into recall GCs observed after human influenza vaccination reflects differences in the history of exposure of individuals to influenza antigens, rather than more fundamental differences in MBC biology between humans and mice. There is expectation the findings of this study will be of critical value to the general understanding of vaccination, as well as to those seeking to generate influenza and HIV broadly neutralizing antibodies (bNAbs) by vaccination. These studies will also be critical to determine the extent to which mouse models reliably predict human responses at the clonal level and can therefore be used to test sequential immunization regimens.
