Resolve, revise, and relax: The 3 Rs of B cell repertoire adjustment

Abstract

Competition for limited, cell extrinsic survival factors is a general feature of peripheral selection checkpoints involved in B lymphocyte maturation and activation. Perhaps the best-characterized example involves BLyS (B lymphocyte stimulator), which modulates the size and composition of mature naïve B cell pools, but evidence for analogous competitive checkpoints is emerging for both germinal center B cells and plasma cells. Here we discuss how deliberate alteration of BLyS levels might be used to manipulate B cell repertoire selection in order to restore self-tolerance in autoimmunity, remodel the repertoire to accommodate neo-self antigens introduced through transplantation and gene therapy, or expand repertoire diversity to reveal novel, therapeutically useful specificities.

1. Introduction

The adaptive immune system generates clonally distributed B- and T-cell specificities through random genetic processes. While these afford great receptor diversity, they also impose a need to screen and eliminate cells with self-reactive or ineffective antigen receptors. In general, this is accomplished through negative and positive selection based on antigen receptor signal strength. Among B-lymphocytes, these processes occur during the development of preimmune pools, as well as during the generation of antibody secreting effectors and memory cells. In toto, these selective events shape the composition of naïve and antigen-experienced specificity repertoires.

Recent insights into the molecular processes underlying B cell selection offer the potential to intentionally adjust repertoire selection. Indeed, some settings have long been considered potential targets for such manipulation, such as resolving failed negative selection in humoral autoimmunity. However, the ability to intentionally manipulate specificity-based selection during B cell development and activation raises further groundbreaking possibilities, such as revising specificity repertoires to accommodate neo-self antigens introduced through transplantation or gene therapies, or intentionally relaxing selection to promote the availability of otherwise rare specificities desired for vaccine efficacy. Herein, we consider each of these broad categories of repertoire manipulation in the context of such translational examples.

2.1 Overview of B cell selective checkpoints

Selection based upon BCR specificity occurs at several well-established checkpoints among both developing and activated B cells. These checkpoints vary in terms of the differentiation stages where they act, as well as in the underlying mechanisms that determine cell survival versus elimination. Although some selection occurs at earlier developmental stages based on pre-BCR interactions, the first checkpoint where selection involves a complete BCR occurs during the immature (IMM) stage in the bone marrow (BM). The mechanisms active at this so-called central tolerance checkpoint have been reviewed extensively elsewhere [1–5], and are cell intrinsic, apparently based on BCR signal strength alone. Thus, IMM B cells receiving strong BCR signals are induced to either undergo secondary Ig gene rearrangements or die. This results in the loss of nearly 90% of all IMM B cells prior to their migration to the periphery [6–8].

The remaining selective checkpoints are imposed in peripheral tissues, and act to shape both the pre-immune and antigen-experienced B cell repertoires. The first of these occurs as IMM B cells that have survived central deletion exit the BM and enter the transitional (TR) pools, where they undergo further specificity-based selection [7–10]. In contrast to the exclusively BCR-dependent processes acting among IMM B cells, selection at the TR stages is governed by interplay between BCR signal strength and availability of the cytokine B lymphocyte stimulator (BLyS, also termed BAFF). Survival through this checkpoint allows TR B cells to complete differentiation and organization into the mature, pre-immune follicular (FO) or marginal zone (MZ) compartments.

Although mature B cells in these quiescent, preimmune pools undergo some degree of protracted BCR-mediated selection, the next selective checkpoints critical to effective immunity and tolerance transpire during antigen-driven activation. The best-established post-activation checkpoint is within the germinal center (GC), where B cells undergo somatic hypermutation (SHM) of Ig genes. GC selection involves positive selection for GC B cells with improved antigen affinity, as well as negative selection against autoreactive specificities arising from SHM [11–13]. Selection mechanisms in the GC remain poorly understood, but current models for GC positive selection invoke competition between GC B cells for antigen and T follicular helper interactions that provide co-stimulation and key survival cytokines [14,15]. Two long-lived subsets are derived from GC reactions, memory B cells and long-lived plasma cells (LLPCs). While there is evidence for competitive survival within these populations, the nature and selectivity involved are not yet well characterized.

Because all known peripheral checkpoints involve competition for B cell extrinsic signals, their stringency should - at least in theory - be manipulable by either restricting or enhancing the availability of these cell-extrinsic factors. The discovery that BLyS serves as the limiting factor mediating TR B cell selection affords the potential to adjust the selective stringency imposed on the emerging naïve repertoire, thus opening previously unappreciated possibilities for translational applications in autoimmunity, transplantation, gene and cell therapies, and vaccine development. Moreover, this may serve as a conceptual template for intervention strategies in antigen-experienced pools as the selective mechanisms active in these pools are revealed.

2.2 The BLyS family of cytokines and receptors

Over the last decade, members of the BLyS cytokine and receptor family have emerged as key players in the selection and homeostasis of nearly all mature B lineage pools (reviewed in [16–18]). This tumor necrosis factor (TNF) sub-family consists of three receptors and two ligands. The receptors are B cell maturation antigen (BCMA), TACI (transmembrane activator and calcium-modulator and cyclophilin ligand interactor), and BR3 (BLyS Receptor 3, also termed BAFF-R). The two cytokines in this family are BLyS (also termed BAFF), and APRIL (a proliferation inducing ligand). BLyS can bind to any of the three receptors, with decreasing affinity in the following order: BR3 > TACI > BCMA. In contrast, APRIL binds only BCMA and TACI.

BR3 expression is first observed on IMM B cells in bone marrow, then increases through the TR stages and is highest on FO and MZ B cells ([19] and reviewed in [20]). Once a B cell exits the BM, its persistence in the preimmune pools depends on its ability to compete for BLyS binding and signaling via BR3 [17,21]. Following antigen encounter, the predominance of the BLyS/BR3 axis in the life of a B cell may begin to shift toward different BLyS family members [22,23]. For example, the short-lived plasma cells that develop within a few days of TLR ligation or early in a T-dependent (TD) response upregulate TACI and may require signaling through this receptor for terminal differentiation [24–26], and signaling through TACI or BR3 can induce class switch recombination [20,27–29]. To date, BCMA expression on B cells has been reported only for GC, memory, and LLPC subsets [20,29]; however, BLyS family receptor expression patterns on highly heterogeneous B cell groups, such as those found in the GC or in the memory compartment, are not yet well-characterized. As discussed elsewhere in this review, APRIL signaling through BCMA likely plays a major role in long-term plasma cell retention and survival in the BM [30–32].

2.3 BLyS integrates selection and homeostasis of the preimmune repertoire

Overwhelming evidence indicates that BLyS controls the size and composition of mature preimmune B cell pools because it is the survival factor for which TR, FO, and MZ B cells compete (reviewed in [33]). Thus, eliminating BLyS through either genetic manipulation or exogenous neutralizing agents yields profound decreases in these pools, although developing BM B cell subsets are spared [34,35]. Conversely, BLyS overexpression or exogenous administration yields dose-dependent increases in TR, FO and MZ B cell numbers [36,37]. These effects are mediated primarily through BR3, since deficiencies or mutations in this receptor yield similarly reduced B cell numbers in mice and humans [38,39]. Moreover, TACI-deficient mice have normal or elevated primary B cell numbers, ruling out a central role for this receptor [40].

Importantly, the efficiency with which B cells garner and process BLyS-mediated survival signals is tied to BCR signaling. This suggests that BCR and BR3 signaling are integrated through downstream signaling crosstalk, although the molecular mechanisms involved in this signaling are complex and subject to debate [19,21,41–44]. Regardless of the exact basis, this crosstalk allows BLyS to control the stringency of TR B cell selection and thus determine the proportion of cells that complete TR development and join the FO or MZ subsets (reviewed in [45]). Thus, under conditions when BLyS availability is elevated, a higher proportion of B cells survive through TR development, and clonotypes that would normally be deleted instead enter the mature FO or MZ subsets [37,46]. The likelihood of successful TR development is increased when more BLyS is present – indeed, increased BLyS signaling through BR3 results in the “rescue” of autoreactive B cells normally lost at the TR checkpoint [46–48]. Thus, the stringency of this peripheral B cell tolerance checkpoint shows “plasticity” [43]. Taken together, these observations suggest that BLyS levels might be purposefully increased or decreased, in order to either relax or tighten the stringency of TR selection, thereby leading to an expanded or restricted mature B cell repertoire.

2.4 Selection and homeostasis in antigen-experienced B cell subsets

In contrast to naïve B cells, less is known about homeostatic controls and selection checkpoints for antigen-experienced B cell subsets. Nonetheless, there is increasing evidence for a continued dual-control nature following antigen encounter, involving both BCR character and competition for extrinsic factors. It is not yet known if any single factor functions to integrate selection and homeostasis in GC, memory B, or LLPC subsets, as is the case for BLyS and preimmune subsets; however, there are tantalizing clues that suggest similarly elegant -- and potentially manipulable – control mechanisms.

Long-lived subsets are generally thought to be the “outcome” of GC reactions, wherein antibody genes undergo SHM as well as class switch recombination. Both positive and negative selection of B cells occur during a GC reaction, such that cells with improved antigen affinity are selected for, and those with decreased or self-antigen affinity are eliminated. GC “evolution” culminates with relatively few B cell clones that possess very high affinity for antigen [15]. Selection mechanisms appear to involve competition between GC B cells for limiting amounts of antigen as well as for factors such as CD40L and cytokines produced by T cells (reviewed in [49,50]).

Roles played by BLyS family members in this process are under investigation. Both BLyS and BR3 knockouts, as well as a BR3 signaling mutant mouse, can initiate GCs upon antigen challenge; however, GC kinetics in these animals are altered, GC structure is generally smaller, and the antibody response is reduced (reviewed in [51]). Interestingly, SHM is not affected in BR3 knockouts or BLyS-deficient mice [52,53]. In addition, TD immune responses in knockout mice lacking TACI, BCMA, or APRIL are largely normal, albeit with subtle changes in antibody responses or isotype representation. Work with ex vivo human B cells indicates that BLyS enhances differentiation of Ig-secreting cells in TD responses, but attenuates differentiation of such cells in TI responses; moreover, access to BLyS may be limiting in GCs [54].

There is increasing evidence in both mice and humans for checkpoints during the GC reaction, as well as during differentiation and maintenance of memory B cells and plasma cells [12,13,55–59]. Dysregulated GC formation and responses, characterized by positive selection of pathogenic specificities, ultimately result in disease in the sanroque mouse model of lupus [60]. In this model, competition among B cells for T cell help may be relaxed as the result of an expansion of T follicular helper cells (TFH). Additional mouse models in which T cell help is aberrant also suggest a competition-based GC checkpoint, where B cells compete for extrinsic factors [61,62].

In healthy humans, cells expressing self-reactive specificities that result from SHM are selected against, and therefore present at very low frequency in the IgM+ memory B cell subset; moreover, mutational analyses -- including reversion to germline -- indicate that this selection occurs before SHM, suggesting an early GC checkpoint [63]. Another study of self-reactive IgG+ memory B cells in healthy people indicates selection after SHM [64]. Work using a tetrameric ds-DNA “mimeotope” to identify and track ds-DNA-specific B cells in human peripheral blood indicates the existence of an antigen naïve-to-experienced checkpoint that may be defective in SLE patients [65]. Moreover, the exclusion of an autoreactive clonotype from early germinal centers is defective in SLE but not RA patients, indicating a disease-specific tolerance defect [66].

Long-lived subsets include memory B cells and long-lived plasma cells. LLPC are originally generated through the GC reactions of primary immune responses, but also when memory B cells expressing high-affinity antibody undergo a second antigen encounter. LLPC are so named because they may survive for the lifetime of the animal [67]. This subset occupies a homeostatic niche that is independent of the much larger / more numerous naïve compartment [35,68]. Despite apparent redundancy in LLPC survival cytokines, recent studies indicate a major role for APRIL in LLPC persistence in the bone marrow [69–71]. Chemokine and homing interactions also contribute to LLPC maintenance, both in BM and extramedullary sites [72–75]. Thus, mechanisms affording a degree of plasticity in the size and composition of the LLPC subset seem likely. The current model is based on competition for survival factor(s) between recently-generated and previously-generated plasma cells, as well as selective processes, such that memory for pathogens (antigens) encountered in an animal's distant past is retained / maintained to some degree [72].

Recent work has shown unexpected phenotypic heterogeneity among memory B cells in both mice and humans, with diversity in surface markers as well as BCR/antibody isotype, and including cells that have not undergone somatic hypermutation [76–79]. This suggests different origins, functions, and/or or homeostatic niches for various memory B subsets. Little is known about maintenance of the memory B compartment; however, it is clear that memory B cells are long-lived, occupy several anatomical locations, and can differentiate into antibody-secreting plasma cells upon secondary antigen encounter – all suggesting homeostatic mechanisms that are likely integrated for maintenance of memory B and LLPC compartments. Neutralization studies indicate that neither BLyS nor APRIL is required for memory B cell survival or function [68]; however, unswitched memory cells are sensitive to BLyS depletion, while IgG-bearing memory cells are not [35]. Moreover, the proapoptotic BH3-only Bcl-2 family member Puma has recently been shown to regulate memory B cell survival in both mice and humans [80]. These studies raise the intriguing possibility that distinct memory B subsets have distinct survival requisites -- potentially affording targets for repertoire adjustment.

3. Resolve, Revise, and Relax: The 3 Rs of repertoire adjustment

A key feature of all known peripheral checkpoints is that, in addition to depending on BCR signaling and function, competition for extrinsic differentiation and survival signals plays an equal role. The plasticity of selection implied by competition affords an opportunity to purposefully manipulate B cell selection, in either direction, to suit a desired outcome [81]. As detailed above, the best understood point of competitive repertoire selection is at the TR checkpoint, where BLyS is the limiting survival factor. Accordingly, this checkpoint may afford a proving ground for repertoire adjustment through the manipulation of selection.

Three general categories of intentional repertoire adjustment can be envisioned and, in fact, are already at varying levels of realization. The notion of reducing available BLyS levels has recently been implemented clinically in SLE (reviewed in [82]), and would be expected to increase the stringency of TR selection. Such an approach might be viewed as “resolving” otherwise aberrant or failed negative selection, with obvious applicability to humoral autoimmunity; however, it is also more broadly applicable to resolving perturbed selection during reconstitution following B ablative therapies in general. The second potential manipulation might be to remodel the repertoire around neo-self antigens. This notion of “revising” the repertoire might be implemented when long term tolerance to recently introduced molecular components would be advantageous, such as during tissue transplantation or gene therapy. Finally, and clearly the least developed concept in terms of implementation, temporarily increasing BLyS levels could broaden the available repertoire by ”relaxing” the stringency of TR selection. This may afford an approach for intentionally expanding the B cell repertoire in order to discover or elicit otherwise rare or counterselected specificities.

3.1. Resolving selection

By far the most advanced approaches embracing the idea of modifying B cell repertoire selection involve humoral autoimmunity – spurred in part by the success of B cell depletion therapy (BCDT). Indeed, therapies targeting B cells for ablation in humoral autoimmune syndromes such as RA, SLE, and Sjøgren's syndrome have been increasingly adopted over the last decade. For example, Rituximab (anti-CD20), originally developed and approved for the treatment of non-Hodgkin's lymphoma and used to treat other B cell cancers, was subsequently approved for treatment of RA [83,84], and has been used in “off-label” settings for other autoimmune disorders [85–88]. While initially conceived as a straightforward elimination of neoplastic or autoreactive clonotypes, a challenge to the overall success of B cell ablative therapies is to not only obtain a “clean slate,” free of offending neoplastic or autoreactive cells, but to subsequently rebuild a mature B cell pool that has undergone appropriate selection. Indeed, we and others have suggested the temporary but deliberate reduction of BLyS levels as an approach for increasing the stringency of TR selection, thereby “rectifying” inappropriate selection at the TR checkpoint (reviewed in [89]).

Several lines of evidence have suggested that this might be a feasible and efficacious approach. Elevated BLyS levels that correlate with autoantibody production and other aspects of disease activity are observed in SLE and RA (as well as mouse models of autoimmune disease). Moreover, there is evidence that peripheral tolerance is defective in these diseases. For example, frequencies of self-reactive Ab-producing cells are equal between the TR and FO stages in SLE and RA patients, whereas in healthy humans, this frequency decreases by half between the TR and mature stages [90,91]. Furthermore, repertoire analyses indicate that the TR checkpoint is defective in both of these diseases, leading to increased representation of autoreactive specificities in the mature naïve pool [91–94]. Experimental manipulation of BLyS levels in mice shows that elevated BLyS causes increased TR B cell throughput as well as relaxed TR selection, indicated by significantly increased representation of autoreactive or otherwise rare specificities in mature naïve subsets [46,95]. Conversely, long-term BLyS neutralization in NOD mice not only restores negative selection at the TR checkpoint, but also results in reduced insulin autoantibody levels, arrested islet cell destruction, and delayed disease onset [96].

Resolving selection at the TR checkpoint might thus be expected to yield a primary B cell compartment with significantly reduced representation of potentially or frankly pathogenic clonotypes. Clinical trial results with Rituximab and belimumab have been informative in this regard. Rituximab (anti-CD20) treatment ablates mature naïve B cells, and leads to a predominance of TR B cells during a prolonged autoreconstitution period [83,97,98]. However, BLyS levels transiently increase 2- to 3-fold for up to 4 months following rituximab treatment, raising the possibility that selection at the TR checkpoint may be temporarily relaxed [99,100]. Treatment with anti-BLyS might be expected to circumvent this possibility. Belimumab (Benlysta), a human anti-BLyS monoclonal antibody that was recently approved by the FDA for treatment of SLE, ablates mature naïve B cell subsets [82]. Although the mechanism for the therapeutic efficacy of belimumab remains unknown [101], it is tempting to speculate that it acts by reducing serum BLyS levels, thereby ablating primary B cells and simultaneously increasing the stringency of TR selection, thus reducing the likelihood that autoreactive specificities transit to mature naïve subsets during reconstitution [81,102].

Neither rituximab nor belimumab appears to have significant or lasting effects on autoantibodies. Memory B cells may remain depleted for a much longer time than primary B cells following rituximab treatment, and this correlates with clinical efficacy in some cases; nevertheless, autoantibody titers generally show only modest reduction (49, 62–67). Belimumab treatment largely spares some memory B subsets as well as plasma cells, and also has limited effects on autoantibodies (81). Sequence analysis of a heavy-chain immunoglobulin family (VH4) in patients with active RA showed increased mutation frequency and ratio of amino acid replacement:silent mutations during early autoreconstitution following rituximab [97]. The most highly mutated cells were derived from antigen-experienced, class-switched memory and plasmablast compartments, suggesting that antigen-experienced B cells (re)generated following rituximab treatment may represent a distinct subset resulting from affinity maturation and/or GC selection [97]. Thus, although both rituximab and belimumab may have transient effects on the size and composition of long-lived subsets, pathogenic specificities existing prior to treatment or generated after treatment may be less tractable. It would be of particular interest to determine which of the various memory B subsets is most affected by each agent, since in mice, IgM-bearing memory cells were more sensitive to anti-BLyS treatment compared to IgG-bearing memory [35].

The notion of resolving selection may extend to B ablative approaches in general, rather than only in the context of autoimmunity. By definition, B ablative therapies yield varying levels of B lymphopenia. This paucity of BLyS “consumers” generally results in elevated BLyS levels - as happens in some SLE patients following B cell depletion. Accordingly, as individuals treated with B ablative regimes undergo self-reconstitution, these high BLyS levels may lead to a lack of appropriate selective stringency at the TR checkpoint, increasing the risk of admitting autoreactive or polyreactive clonotypes into the primary pools, or a lack of appropriate germinal center selection upon antigen challenges faced during the reconstitution process [81]. Using titered doses of anti-BLyS to maintain selective stringency during autoreconstitution of B cells following B cell depletion therapy done for any purpose (e.g., autoimmune therapy, cancer therapy) may prove to be an important tactic in the strategy of repertoire adjustment.

3.2. Revising selection to tolerate neo-self antigens

A growing number of therapeutic interventions introduce structures that act as antigens, either transiently or chronically. Perhaps foremost among these are alloantigens introduced by tissue and organ transplantation. The presence or induction of alloantibodies is increasingly recognized as a major contributing factor to acute and chronic rejection, and clearly play direct roles in sensitization [103,104]. Similarly, gene and cell therapeutics involve long-term introduction of structural gene products, as well as the presence of delivery vectors [105,106] that may elicit alloantibody formation.

While less well developed than the concept of rectifying the B cell repertoire in autoimmunity, the notion of repertoire “revision” around neo-self antigens to afford long-term tolerance has been proposed in recent years [81,107,108]. Lending support to this idea were observations and results implicating B cells as key players in allograft rejection (reviewed in [109–111]). Several strategies have been tested for achieving B cell tolerance in adult transplant patients, and most of these involve B cell depletion following transplantation (reviewed in [108,112,113]). However, despite long-term graft survival and function, many patients ultimately generate donor-specific antibodies or antibody-mediated transplant rejection, indicating a failure of B cell tolerance. Analogous to BCDT in cancer and autoimmunity, while these regimes provide a setting where new antigens might be introduced with impunity, successful subsequent reconstitution will require repertoire selection under conditions sufficiently stringent to prevent the maturation of clones reactive to the neo-self antigens introduced. If this can be accomplished without excess BLyS -- for example using titered doses of a BLyS neutralizing agent -- then transitional selection stringency can be rigorous as B cells reconstitute, and neo-self specificities should be deleted [108].

Hence the proposed approach: deplete B cells immediately before or at transplantation, then use titered doses of anti-BLyS to maintain the stringency of TR selection during B cell reconstitution. There is evidence from an alloantigen-specific Ig transgenic mouse model that “permanent” B cell tolerance can be achieved when the primary repertoire is rebuilt in the presence of a tissue graft, with apparently complete clonal deletion of alloreactive specificities from the primary repertoire [114]. However, these results were obtained in a globally immune cell depleted context (scid mice). Results currently in press for mice illustrate the potential effectiveness of B cell depletion at transplant followed by periodic, gradually decreasing doses of anti-BLyS [115]. In this work, fully allogeneic transplanted pancreatic islets were maintained for more than 6 months after cessation of anti-BLyS treatment; moreover, subsequent new transplants of the same MHC haplotype were accepted and maintained. Thus, revising the B cell repertoire in the presence of allograft transplants shows promise as an approach for achieving long-term B cell tolerance.

3.3. Relaxing selection to capture rare specificities

A final and comparatively unexplored notion is that under some circumstances, temporarily relaxing the stringency of selection might prove informative or beneficial. While a seemingly dangerous assault on the basic notion of horror autotoxicus [116,117], the recruitment of polyreactive and self-reactive specificities may prove useful in some scenarios, such as autologous tumor vaccination. Indeed, healthy people may have a range of autoantibodies and self-reactive B cells in both immature and mature pools [90,91,118,119]. Moreover, a weak immune response to tumor-associated antigens – which may be recognized as self-antigens -- has presented one roadblock in cancer vaccine development [120], and it has been suggested that current data support a beneficial effect of variation in both T and B cell repertoires (via epitope spreading) in cancer vaccines [121].

Similarly, deliberately relaxing the stringency of selection might reveal key specificities for vaccine development that are normally lost during selective processes [81]. A relevant contemporary example of this possibility is the difficulty in generating vaccines capable of eliciting broadly neutralizing antibodies (bNAb) to human immunodeficiency virus (HIV). bNAb are extremely rare and may appear only after years of chronic infection [122,123]. Moreover, attempts to elicit bNAb through conventional approaches have been largely unsuccessful to date [124–126]. However, recent observations suggest that antibodies with the desired characteristics may be those that are more likely to be lost during selective processes. For example, extensive SHM appears to be a hallmark of naturally occurring bNAb, suggesting a GC origin [127,128] as well as the potential for elimination by negative selection. Further, B cell clonotypes that generate polyspecific bNAb, including those with known autoreactivity, may be strongly counterselected [129,130]. In transgenic mouse models, such clonotypes are subjected to central and peripheral selection mechanisms, and immature cells can be “rescued” by culture with IL-7 and BLyS [130,131]. In contrast, there is evidence that polyspecific antibodies found in humans are positively selected if they include high-affinity pathogen specificity [132]. During the response to HIV, naïve and memory B cells can produce polyreactive antibodies that recognize two sites on the HIV surface; moreover, mutations that increase affinity also increase both breadth and neutralizing potency [133,134].

Thus, broadly neutralizing clonotypes may be rare because they are under-represented in the mature naïve B cell pool; because they mutate in the “wrong direction” (toward self-reactivity or polyreactivity with decreased affinity for HIV) and are therefore eliminated during GC reactions; because of attrition from long-lived subsets; or for a combination of all of these reasons. Under conditions involving naturally or artificially elevated BLyS levels, TR selection stringency should be relaxed, leading to a generally broadened naive B cell repertoire. This in turn may result in increased representation of potentially broadly neutralizing clones in FO and MZ subsets. Manipulation of selection checkpoints in the GC and beyond, with BLyS or APRIL, or other limiting survival factors, should also afford the possibility of enhancing recruitment and retention of broadly neutralizing B cell clonotypes. Recent evidence indeed suggests enhanced neutralizing antibody titers using APRIL containing viral vaccine constructs, as well as in mice treated with BLyS prior to vaccination [135,136].

Whether regimes designed to relax selection temporarily could be implemented safely remains unclear. Nonetheless, this approach might be used immediately in experimental animals to discover and characterize rare, therapeutically useful B cell clonotypes, as well as to reveal the basis for their relative paucity under normal conditions.

Abbreviations

Ab

antibody

BCMA

B cell maturation antigen

BCDT

B cell depletion therapy

BCR

B cell receptor

BLyS

B lymphocyte stimulator

BM

bone marrow

bNAb

broadly neutralizing antibody

BR3

BLyS receptor 3

FO

follicular

GC

germinal center

HIV

human immunodeficiency virus

Ig

immunoglobulin

IMM

immature LLPC long-lived plasma cell

MZ

marginal zone

MHC

major histocompatibility complex

RA

rheumatoid arthritis

SHM

somatic hypermutation

SLE

systemic lupus erythematosus

TACI

transmembrane activator and calcium-modulator and cyclophilin ligand interactor

TD

T-dependent

TFH

T follicular helper cell

TI

T-independent

TLR

Toll-like receptor

TNF

tumor necrosis factor

TR

transitional