br Experimental Procedures br Author Contributions br Acknow
Introduction Germinal centers (GCs) form in secondary lymphoid organs of vertebrates in response to challenge with T-cell-dependent antigens. After extrafollicular and follicular interactions with cognate antigen and T helper cells, responding B cell clones can enter the GC response, where they undergo iterative cycles of proliferation, somatic hypermutation (SHM), and selection such that clones acquiring increased affinity for the eliciting antigen preferentially accumulate (Gatto and Brink, 2010, Victora and Nussenzweig, 2012). This process is also associated with the differentiation of memory pde inhibitors (MBCs) and plasma cells (PCs) from GC B cell precursors. MBCs and PCs are responsible for effective long-term immunity against infectious pathogens and underpin the efficacy of almost all current vaccines (Plotkin et al., 2008). Whereas long-lived PCs exit the GC response and primarily contribute to immune protection by homing to the bone marrow and secreting high-affinity antibodies, MBCs emerge from the GC as recirculating cells that are primed to elicit rapid antibody responses upon secondary antigen challenge. Multiple subsets of both mouse and human MBCs have been defined according to their immunoglobulin (Ig) isotype, surface phenotype, longevity, survival requirements, SHM characteristics, and functional response to antigen re-challenge (Good-Jacobson and Shlomchik, 2010, Kurosaki et al., 2015, Tangye and Tarlinton, 2009, Tarlinton, 2006). One point of contention has centered on the precise origins of MBCs. BCL6-deficient mice produce isotype-switched, unmutated MBCs in response to a T-cell-dependent antigen (Toyama et al., 2002), indicating a GC-independent pathway for MBC generation. Furthermore, early, isotype-switched MBCs have been shown to arise outside of the GC response and to almost invariably carry no somatic mutations (Chan et al., 2009, Kaji et al., 2012, Taylor et al., 2012). Nevertheless, it is broadly accepted that GCs give rise to long-lived MBCs (Shlomchik and Weisel, 2012, Victora and Nussenzweig, 2012) because of the predominance of Ig-variable-region gene SHM events among MBCs (Gray, 1993, Tangye and Tarlinton, 2009) and the greatly reduced numbers of somatically mutated MBCs and impaired recall responses in mice and humans who are unable to form robust GCs (Good-Jacobson and Shlomchik, 2010, Shlomchik and Weisel, 2012, Tangye et al., 2012). The transcription factor BLIMP-1 is required during the initiation of PC differentiation and has been used extensively for investigating GC-derived PC precursors (plasmablasts) and their further differentiation into PCs (Kallies et al., 2004). Taking advantage of this marker, we previously showed that PC differentiation in the GC is not stochastic but rather is induced only among high-affinity GC B cells, even if low-affinity cells still predominate (Phan et al., 2006). A similar study of MBC differentiation has been hampered by the lack of a “master” transcription factor or other marker that identifies B cells destined to exit the GC as MBCs. Here, we demonstrate that expression of the chemokine receptor CCR6 specifically identifies the precursors of MBCs within the GC. CCR6+ MBC precursors were found to be highly enriched in the light zone (LZ) of both mouse and human GCs, were unique within the GC in undergoing cell-cycle exit, and possessed a gene expression signature consistent with an MBC transition. SHM and antigen-binding analyses revealed that MBCs emerge predominantly from low-affinity CCR6+ B cells in the GC LZ but also develop from rarer, high-affinity CCR6+ LZ progenitors. The differentiation of MBCs and PCs from GC B cells is therefore driven by fundamentally different stimuli that establish the distinct roles of these two GC-derived cell types in mediating long-term immunity against reinfection.