A cornerstone of the project was the construction of an integrated experimental-computational pipeline capable of reading, for each individual B cell, its full transcriptome together with the paired immunoglobulin heavy- and light-chain sequences. The pipeline has been released under an open-source license. All raw and processed datasets have been deposited in ArrayExpress, and their reuse by external groups has already generated more than thirty data-driven publications, attesting to the resource’s broad utility.
Using this platform, we first characterized the fate of B cells specific for wild-type influenza hemagglutinin (HA) across spleen, lymph nodes and lungs. We revealed that memory cells destined for the lung are generated in lymphoid organs and later acquire a tissue-resident phenotype upon arrival in the lung. This work provided the first high-resolution map of lung-resident B-cell memory and has become a frequently cited reference in the field. The discovery prompted two follow-up investigations. In the first we found that Streptococcus pneumoniae, after influenza, disrupted the recruitment signal, thereby preventing memory B cells from establishing residency and compromising protection against a subsequent influenza challenge. In the second we demonstrated the presence of active germinal centers in the nasal mucosa of mice and humans and documented their clonal relationship to systemic responses.
Parallel experiments addressed the role of T-cell help and of pre-existing humoral immunity. We showed that CD4 T cells present before infection accelerate extrafollicular plasmablast production without increasing germinal-centre entry . A more ambitious series of reinfection, monoclonal-antibody transfer and mixed-cell-transfer models clarified how antibody valency and topology dictate whether memory clones re-enter germinal centres or whether naïve B cells dominate secondary responses. We found that antibody feedback does not hamper vaccine responses.
The COVID-19 amendment yielded two complementary outputs. Longitudinal sampling of hospitalised patients, combined with the single-cell pipeline from aim 1, uncovered B-cell clones that originated from seasonal coronavirus memory, rapidly differentiated into plasmablasts and gradually broadened their neutralisation breadth. Furthermore, we also identified a transcriptomic “affinity signature” that identifies high-affinity SARS-CoV-2-specific B cells for therapeutic antibody isolation.
Overall, the unifying model posit that viral infections elicit persistent germinal centers in lymphoid organs that, in turn, fuel the generation of variable memory that takes residency in lungs and upper respiratory tract and it is equipped to defend us against new emerging viral variants.
Over the funding period we produced twelve peer-reviewed articles, two invited reviews and three preprints; delivered approximately fifteen invited lectures at invited seminars or major conferences (Keystone, EMBO, WIRM, SSI, among others); and made all code, data and protocols freely accessible. Early-career researchers received training and moved on to new positions.
