Germinal Center B Cells Replace Their Antigen Receptors in Dark Zones and Fail Light Zone Entry when Immunoglobulin Gene Mutations are Damaging

摘要

class="head no_bottom_margin" id="sec1title">IntroductionThe affinity and breadth of antibodies improves over the course of immune responses as a consequence of antibody affinity maturation in germinal centers (GCs) (, ). GCs are specialized transient structures that form within B cell follicles in the days following an infection or immunization and can persist for periods of a few of weeks to many months depending upon the nature of the challenge. GCs are anatomically divided into two distinct zones known as light zones (LZs) and dark zones (DZs), with the former forming proximal to the site of antigen entry. LZs are clearly distinguishable by the presence of specialized stromal cells known as follicular dendritic cells (FDCs), which express very high amounts of Fc and complement receptors that sequester antigen-containing immune complexes. DZs are identified by the relative absence of FDCs and by the higher density of proliferating GC B cells.Antibody affinity enhancements occur through iterative rounds of immunoglobulin variable region (IgV) gene somatic hypermutation (SHM) and selection involving GC B cells transitioning back and forth between DZs and LZs. Movement of GC B cells between the two zones is associated with changes in behavior and phenotype (, , ). Expression of Aicda (, , ), the gene encoding activation induced cytidine deaminase (AID), is higher in DZ GC B cells. AID catalyzes the deamination of cytosines to uracils, which in turn recruits various error prone repair pathways that cause nucleotide substitutions, additions, or deletions. Mutations in IgV genes are introduced at frequencies of approximately 10⁻³ nucleotides/division, approximately 10⁶-fold above background mutation rates (). Following antibody diversification in DZs, somatically mutated GC B cells reduce surface expression of the chemokine receptor CXCR4 and migrate to LZs, where they test their newly minted receptors through “selection,” a competitive process that involves the acquisition of antigen from FDCs and the presentation of processed peptides to T follicular helper (Tfh) cells (, , , , ). Rare cells in which somatic mutations cause antibody affinity improvements are thought to be preferentially selected as a result of them internalizing and processing more antigen to limiting numbers of Tfh cells, thereby driving antibody affinity maturation.It has become increasingly evident that humoral immune responses are not always best served by rapidly expanding the highest affinity B cell clones at the expense of all others (, href="#bib31" rid="bib31" class=" bibr popnode">Mesin et al., 2016). For example, the unmutated common ancestors of HIV broadly neutralizing antibodies (bnAbs) often are of very low affinities and only acquire neutralizing potential after accumulating a slew of somatic mutations (href="#bib19" rid="bib19" class=" bibr popnode">Kelsoe and Haynes, 2017, href="#bib54" rid="bib54" class=" bibr popnode">West et al., 2014). While HIV bnAbs may represent extreme examples when it comes to mutation loads, simpler antibodies such as those arising during responses to influenza A infections may also mature in a stepwise process (href="#bib25" rid="bib25" class=" bibr popnode">Lingwood et al., 2012). Moreover, secondary memory B cell responses to viral variants and re-assortments benefit from antibody diversification in the GC during the primary challenge (href="#bib34" rid="bib34" class=" bibr popnode">Pappas et al., 2014, href="#bib38" rid="bib38" class=" bibr popnode">Purtha et al., 2011). As such, the preferential selection and expansion of high affinity B cells should not come at the expense of retaining breadth. Consistent with such a notion, two recent studies tracked clonal participation in GCs formed following immunization with complex protein antigens and reported that, while affinity enhancements with time were evident, GCs were remarkably permissive to the retention of low to moderate affinity cells (href="#bib22" rid="bib22" class=" bibr popnode">Kuraoka et al., 2016, href="#bib47" rid="bib47" class=" bibr popnode">Tas et al., 2016). Therefore, GC B cell selection might be as much about screening new pools of somatically mutated cells for their ability to still bind antigen as it is about expanding the very best clones (href="#bib6" rid="bib6" class=" bibr popnode">Bannard and Cyster, 2017). This is necessary because the random nature of SHM means that it is far more likely to negatively impact antigen-binding or be harmful to antibody structure than it is to increase affinity.The dynamics of GC B cell responses have been intensively studied in recent years with direct measurements of cell movement between zones and states revealing that cells undergo on average two cell divisions between each selection event (href="#bib15" rid="bib15" class=" bibr popnode">Gitlin et al., 2015, href="#bib14" rid="bib14" class=" bibr popnode">Gitlin et al., 2014), inclusive of the one initiated in the LZ, and that approximately half of all DZ cells transition back to the LZ every 4 hr period (href="#bib31" rid="bib31" class=" bibr popnode">Mesin et al., 2016, href="#bib50" rid="bib50" class=" bibr popnode">Victora et al., 2010). Therefore, the period between a cell somatically mutating its immunoglobulin genes and it entering the selection process is probably quite short (href="#bib8" rid="bib8" class=" bibr popnode">Bannard et al., 2016). This raises the question of whether GC B cells actively replace their BCRs during the intervening period so that they are selected while only expressing the “new” mutated variants. When membrane immunoglobulin half-lives have been measured in follicular B cells, they were found to be long, in the range of 20–80 hr (href="#bib5" rid="bib5" class=" bibr popnode">Andersson et al., 1974). While it is easy to rationalize that GC B cells acquiring affinity enhancing mutations can be preferentially selected without need for complete receptor replacement, it is less clear how detrimental or affinity lowering mutations might be screened out, as is required for diversification, without BCR turnover.Here, we determined the stage at which pre-SHM antigen receptor complexes are degraded in GC B cells by correlating the acquisition of harmful mutations with the presence or absence of membrane BCRs. Within the DZ, although some cells displayed high surface BCR levels despite carrying damaging mutations in their immunoglobulin genes, two thirds of the cells with that mutation type had low to negligible surface BCR expression, indicating that complete receptor turnover had already occurred. GC B cells carrying damaging mutations did not accumulate in LZ populations as might be predicted if newly minted receptors are first tested at that phase. Instead, apoptosis of cells lacking surface BCR complexes was triggered while cells were still in the DZ state. We propose that the turnover of BCR complexes and the testing of antibody functional integrity following SHM in the DZ prevents the accumulation of cells carrying damaging mutations and facilitates accurate selection for antigen-binding ability when GC B cells reach the LZ.