BAFF and the plasticity of peripheral B cell tolerance

BAFF and its receptors play a crucial role in peripheral B cell selection and survival, by dictating the set point for mature primary B cell numbers and adjusting thresholds for specificity-based selection during transitional differentiation. The notion that selective stringency can be varied on the basis of homeostatic demands reveals a previously unappreciated degree of plasticity in B cell tolerance, and suggests a paradigm that unites peripheral negative and positive selection with the maintenance of mature B cell numbers. Moreover, it implies a developmentally regulated coupling of BCR and BAFF receptors at the transitional stages and beyond.

Introduction

Within the confines of clonal selection, acquired tolerance is accommodated by purging autoreactive specificities at key checkpoints during lymphocyte differentiation. Assessments of B cell dynamics are consistent with this idea, inasmuch as only ∼5% of newly formed B cells survive to enter mature primary pools. Although BCR signal strength was long assumed the sole factor determining cell death versus survival at these checkpoints; recent evidence reveals a more complex mechanism that integrates tolerogenic elimination, positive selection, and the maintenance of mature B cell numbers. This unifying conceptual advance derives from observations showing that in the penultimate stages of B cell maturation, BCR signaling mediates not only tolerogenic elimination but also survival on the basis of relative fitness to compete for BAFF (BLyS). Herein, we summarize seminal observations leading to these ideas, emphasizing recent literature linking BAFF with B cell selection and homeostasis.

Several observations provided impetus for revisiting the accepted notions of B cell tolerance mechanisms. Key among these was the discovery of a peripheral developmental intermediate, coupled with evidence that BCR-mediated negative and positive selection occur at this stage.

The peripheral differentiative intermediate for primary B cells, now termed the transitional (TR) B cell, was originally characterized by high levels of Heat Stable Antigen (CD24) and has subsequently been refined with additional markers. Several findings emerged from these studies [1, 2]. First, these cells resembled neonatal B cells in terms of responses to BCR cross-linking and other signaling characteristics. Second, they displayed rapid turnover, indicating that residency in the TR compartment is short—on the order of days. Last, comparisons of TR versus mature B cell production rates indicated that two thirds of TR cells formed die before maturation, suggesting this stage may undergo negative or positive selection.

Historically, deletional versus anergic B cell tolerance mechanisms were considered disparate, yielding rapid cell death in the bone marrow versus a refractory but persistent clone in the periphery. These notions, originally proposed by Nossal and Pike [3, 4], were apparently supported by in vivo studies in the HEL/anti-HEL transgenic model, where high avidity BCR interactions led to the elimination of HEL-reactive cells in the bone marrow, while low avidity interactions indeed yielded self-reactive cells in the periphery [5]. However, following description of the TR compartment, Fulcher and Basten demonstrated that these ‘anergic’ B cells achieved entrance to the TR pools, but nonetheless died rapidly before entering mature pools [6]. This relationship holds in humans, as demonstrated by studies showing the loss of autoreactive and polyreactive clonotypes at the TR stages [7^••], and apparent failure of this checkpoint in SLE patients [8].

Concomitant with these emerging notions of peripheral maturation and tolerance, mounting evidence indicated a positive role for BCR signaling in the development and maintenance of peripheral B cells. The notion that BCR signals are required for peripheral B cell differentiation or survival was first suggested by the studies of Gu et al. [9], which revealed clear differences in V_H distribution among immature versus mature bone marrow B cells. Because these findings preceded definition of TR pools, the developmental stage where such bias was imposed remained unclear. Subsequent studies, using analogous assessments of light chain V gene usage, showed that this bias was introduced between the TR and mature peripheral stages [10]. Together, these findings implied that BCR signal strength could specify any of three alternative fates at the TR stage: differentiative failure, continued survival or tolerogenic elimination. This tripartite distribution of BCR-mediated effects was directly demonstrated by Wang and Clarke, using a transgenic BCR model coupled with differing doses of anti-clonotypic antibody [11]. Finally, subthreshold BCR signals are not only essential for continued differentiation at the TR stage, but are also required for survival in the mature peripheral pools. This was definitively shown by studies where induced ablation of either the BCR itself or BCR signaling complex molecules led to the rapid death of most mature B cells [12, 13].

Considered together, these findings carried the overarching implication that a developmental switch transpires during the TR stages, whereby BCR signaling no longer only mediates elimination, but is also required for continued survival.

An equally significant contribution to evolving thought about selection in the peripheral B cell compartments has been the notion of competitive survival. Most studies demonstrating such competition relied on mixed bone marrow chimeras. The earliest such studies showed that developing B cells from xid mice could persist in the periphery but failed to do so when competing with wild-type B lineage progenitors [14]. More recent studies from the Freitas laboratory indicated that intraclonal competition, in terms of BCR specificity, determined relative competitive advantage among peripheral B cells [15, 16]. Thus, comparisons of turnover rates in mixtures of BCR Tg and non-Tg cells revealed that a mature B cell's lifespan is not absolute, but instead depends on the clonotypic identities of co-existing competitors. Cyster et al. showed similar specificity-based competition as cells navigate TR stages [17]. In these studies, B cells bearing an autoreactive BCR transgene persisted in the absence of competition, but were excluded from splenic follicles and died rapidly when competing with B cells of normal receptor diversity.

These findings substantially impacted perceptions of how BCR signals mediate peripheral B cell survival in several ways. First, they showed that B cell survival capacity, and hence lifespan, is neither fixed nor absolute but variable and relative. Second, they implied that the survival-promoting effects of BCR signals are not cell intrinsic, but are instead coupled to the acquisition of limiting, cell-extrinsic resources.

The keystone for assembling a conceptual paradigm that could accommodate peripheral selection, survival and competition was provided by the discoveries of BAFF and BAFFr. Initial descriptions of BAFF revealed profound positive effects on mature B cells both in vivo and in vitro [18, 19]. The subsequent discovery of BAFFr, a B-lineage restricted receptor whose sole natural ligand is BAFF, showed that BAFF–BAFFr interactions were crucial to maintaining normal peripheral B cell numbers [20, 21].

Striking parallels with the emerging concepts of B cell maturation, survival, and competition in the periphery accumulated rapidly. First, in the absence of BAFF or BAFFr, differentiation failed at the early TR stages, suggesting this is the developmental point where reliance on BAFF is imposed [20, 21, 22]. Consistent with this, BAFFr expression begins at the TR stages and increases steadily as cells mature and join the follicular or marginal zone pools [23]. Second, BAFF signaling via BAFFr proved the key resource for which mature peripheral B cells compete, since B cells from BAFFr mutants failed to thrive and were diluted rapidly when competing with BAFFr-sufficient B cells [24]. Last, BAFFr signaling competence dictated mature B cell lifespan, since intermediate turnover rates were observed among B cells heterozygous for a non-functional BAFFr mutation [24]. Enhanced survival has proven the major downstream consequence of BAFF–BAFFr signaling, as demonstrated by in vivo labeling showing that the lifespan of TR, follicular (FO) and marginal zone (MZ) cells is reduced in BAFFr mutants [22] and is extended with exogenous BAFF administration [23]. These effects involve upregulation of Bcl-2 family members, prevention of PKC-δ nuclear accumulation, and activation of the non-classical Nuclear Factor of kappa B (NF-κB) pathway [23, 25, 26•].

These findings fundamentally altered notions of peripheral B cell survival mechanisms because they revealed a second, BCR independent, signaling requirement. Perhaps more importantly, they carried a profound implication for BCR-mediated selection in the periphery: If the likelihood of successful TR differentiation can be modified by BAFF signaling, then the threshold for negative selection at this stage must be variable. This implied plasticity in the stringency of TR selection was directly interrogated in two elegant studies using the HEL transgenic model. The results clearly showed that increased BAFF levels rescued the maturation of autoreactive cells normally lost at the TR checkpoint [27, 28••]. Mirroring previous results, the likelihood of such rescue depended on the competing cells, such that when mixed with enough competitors of normal BCR diversity, autoreactive clonotypes failed to persist regardless of excess BAFF [28^••]. Importantly, high avidity interactions that yielded deletion at bone marrow immature stage were not influenced by BAFF overexpression, again pinpointing the TR stage as the point where BCR-mediated selection and BAFFr signaling become integrated.

Overall, these observations suggest a model in which the set point for B cell numbers is reached when free BAFF becomes limiting, and the BCR signal strength suitable for successful TR differentiation varies on the basis of available BAFF [29]. Accordingly, knowledge of the molecular mechanisms integrating BCR and BAFFr signals is fundamental to understanding peripheral B cell tolerance. Clearly, BAFF thwarts death mechanisms driven by strong BCR signals. For example, BCR-driven Bim dephosphorylation and consequent caspase-mediated cell death [30] is blocked by BAFF. While these findings disclose distal events through which BAFF opposes BCR-driven death at the TR stages, the reason why BCR signals are also necessary for BAFF-mediated survival remains unclear. Several recent studies suggest that cross-talk between members of the NF-κB family may provide this link. Indeed, disruption of either the classical or non-classical NF-κB pathways blocks peripheral B cell development [31, 32]. Because the BCR primarily if not exclusively, utilizes the classical NF-κB pathway while BAFFr uses the non-classical NF-κB pathway, the activation of parallel or interacting downstream systems may be required to yield survival. Recent studies by Sasaki et al. demonstrated that constitutive classical NF-κB pathway activation allowed B cell development in the absence of BAFFr [33^•], but BAFF treatment still led to increased survival over that provided by forced classical NF-κB activation alone, indicating a non-redundant function for BAFFr. A separate study by Enzler et al. showed that all major components of the non-classical NF-κB pathway are crucial for BAFF-mediated survival [26^•]. Further, while knockout mice for the components of each pathway differ slightly, all are essentially similar to BAFF^−/− or BAFFr^−/− mice [20, 34, 35, 36]. Thus, it seems likely that the BCR and BAFFr are coupled through these two signaling systems, either via simultaneous activation targeting separate promoters of survival, or through direct cross-talk. Regardless of exact mechanism, determining the details of this relationship should eventually afford the opportunity to deliberately manipulate selection and survival within peripheral B cell subsets.