Induction of broadly neutralizing antibodies against rapidly evolving viruses is a “holy grail” of current vaccine research. A study in this issue reveals that it’s not just the properties of the immunogen that matter, but also the anatomical site of the B cell response.

Influenza infection causes substantial morbidity and mortality each year despite the wide use of an annually updated vaccine. A major limitation of the current vaccination approach is the weak ability to induce antibodies that recognize variant forms of the hemagglutinin (HA) antigen, which arise continually by antigenic drift (point mutations in an existing strain) and more rarely by antigenic shift (major changes in the HA gene by viral mixing). In this issue, Adachi et al. examine the requirements for the generation of cross-reactive antibodies that recognize different influenza strains. By comparing the properties of germinal center (GC) and memory B cells (Bmem) isolated from lung, draining LNs, and spleen of intranasally infected mice, they discovered that cross-reactive B cells are preferentially generated in the lung. Using conditional Bcl6 ablation, they established that ongoing GC responses are needed for generation of these late-arising cross-reactive memory cells. Using a novel procedure of intratracheal exposure to low amounts of 5-ethynyl-20-deoxyuridine (EdU) to locally label dividing cells, they show that cross-reactive Bmem are preferentially generated and maintained in the lung.

Although GCs are most readily induced in secondary lymphoid organs, they can also form ectopically at sites of inflammation. Remarkably, this study shows that ectopic GC formation doesn’t just provide additional antibody-generating capacity, it provides a qualitatively unique environment tailored for the induction of cross-reactive antibodies. As the authors suggest, the enhanced cross-reactive Bmem induction in the lung might reflect the presence of greater amounts of viral antigen, drifted variants that arose before the virus was cleared, or forms of HA that are more accessible to the BCR. It will be exciting to see whether these factors operate together to make the local response the most efficacious.

Interestingly, the proliferation rate of HA-binding GC B cells in the lung was greater than in LNs and spleen. The basis for this effect was not examined but might reflect properties associated with the inflammatory milieu or perhaps even the higher oxygen levels in the lung. Whatever the explanation, the faster proliferation rate may contribute to more rapid accrual of mutations and possibly greater diversification of the Bmem pool that is generated. While the mechanistic studies move forward, this work is a reminder that anatomy and not just molecular components matter in the efforts to develop vaccine strategies that induce broadly neutralizing antibody responses.

, et al
J. Exp. Med.