In this study we show that BCR affinity and subset identity make unique contributions to anergy. Analysis of anti-Smith (Sm) B cells of different affinities indicates that increasing affinity improves anergy’s effectiveness while paradoxically increasing the likelihood of marginal zone (MZ) and B-1 B cell differentiation rather than just follicular (FO) B cell differentiation. Subset identity in turn determines the affinity threshold and mechanism of anergy. Subset-specific affinity thresholds for anergy induction allow discordant regulation of low-affinity anti-Sm FO and MZ B cells and could account for the higher frequency of autoreactive MZ B cells than that of FO B cells in normal mice. The mechanism of anergy changes during differentiation and differs between subsets. This is strikingly illustrated by the observation that blockade of BCR-mediated activation of FO and MZ B cells occurs at different levels in the signaling cascade. Thus, attributes unique to B cells of each subset integrate with signals from the BCR to determine the effectiveness, affinity threshold, and mechanism of anergy.
Immune tolerance is paramount to the prevention of autoimmune diseases, and the evolution of multiple checkpoints to prevent anti-self B cell activation reflects this. These checkpoints span the full range of B cell differentiation and activation, from the initial expression of a BCR to the pre-plasma cell (PC)3 stage following Ag activation (1, 2, 3, 4, 5, 6, 7). Anergy, defined as hyporesponsiveness to BCR and TLR activating signals, is a significant phenomenon in mice and man (8, 9). Although progress has been made (10, 11, 12, 13, 14), we currently lack a detailed understanding of the underlying mechanisms of anergy. The observation that some autoreactive B cells are long lived (15), thereby providing a significant window of opportunity for anergy to fail, underscores the importance of understanding these mechanisms.
Multiple factors influence the timing and mechanism of anergy, including the nature and availability of self-Ag, the affinity of the BCR for Ag, and when during differentiation B cells encounter Ag. B cell subset identity may also contribute to anergy induction. Some autoreactive B cells arrest at a transitional (Tr) stage, termed An1 (8), but others go on to become mature follicular (FO), marginal zone (MZ), and B-1 B cells (15, 16, 17, 18, 19, 20). The MZ and B-1 repertoires are distinctly more anti-self than the FO B cell repertoire (21), suggesting that anti-self FO and MZ B cells are regulated differently. An understanding of how these factors integrate to regulate anti-self B cells may reveal new targets to block anti-self B cell activation in disease.
B cells specific for the self-Ag Smith (Sm), a ribonucleoprotein commonly targeted in systemic lupus erythematosus, are ideally suited to addressing how BCR affinity and subset identity integrate to regulate autoreactive B cells. Our previous analysis of anti-Sm transgenic (Tg) mice (2-12H) indicates that anti-Sm B cells can differentiate to the FO, MZ, and B-1 subsets (15, 16, 17). Significant to this study, anti-Sm FO B cells are anergic (15), whereas anti-Sm MZ and B-1 B cells appear to be functional (17, 22). In this report, we compare the regulation of anti-Sm B cells of 2-12H mice and mice of two L chain-restricted anti-Sm Tg lines, 2-12H/Vκ8 and 2-12H/Vκ4. 2-12H and 2-12H/Vκ8 mice generate low-affinity anti-Sm B cells (15, 23). We add to this analysis the high-affinity anti-Sm B cells of 2-12H/Vκ4 mice, and present evidence that BCR affinity and subset identity control different key features of anergy.
Materials and Methods
Splenic B cells were enriched to 92–98% purity by negative selection using IMag Streptavidin Particles Plus (BD Biosciences) following the manufacturer’s protocol. For cell sorting experiments, splenic cells were stained for CD21, CD23, B220, CD138 (BD Biosciences), and propidium iodide and sorted on a MoFlo high-speed sorter (DakoCytomation) to a purity of >95%.
In vitro stimulation
B cells were labeled with CFSE (Molecular Probes-Invitrogen) and cultured in 24-well plates (1 × 106/ml) with the indicated LPS or CpG concentrations for 72 h. For Ab secretion, cells were cultured with 10 μg/ml LPS or 1 μg/ml CpG for 4 days, and secreted IgM was measured by ELISA. For BCR cross-linking, cells were stimulated for 72 h with 10 μg/ml F(ab′)2 goat anti-mouse IgM (Jackson ImmunoResearch Laboratories) and stained with the LIVE/DEAD dead cell stain kit (Molecular Probes-Invitrogen).
Single-cell suspensions of splenocytes were prepared and stained as previously described (6, 15, 22) and analyzed at the University of North Carolina Flow Cytometry Facility using a FACSCalibur (BD Biosciences) or a CyAn (Dako) flow cytometer. Analysis was performed with Summit software (Dako). For intracellular staining of transcription factors, cells were fixed with 2% paraformaldehyde and permeabilized with ice-cold MeOH for 30 min and stained with mouse anti-Pax-5 (Ms IgG1; BD Biosciences), rabbit anti-X box-binding protein (XBP-1), or goat anti-IFN-response factor 4 (IRF-4) (Santa Cruz Biotechnology). Goat anti-rabbit Alexa Fluor 647, goat anti-mouse IgG1Alexa Fluor 488 (Molecular Probes-Invitrogen), and donkey anti-goat Cy5 (Jackson ImmunoResearch) were used as secondary Abs. For cell signaling analysis, cells were rested at 37°C for 30 min and then stimulated with 10 μg/ml F(ab′)2 of a goat anti-mouse IgM. At the indicated times, cells were fixed and permeabilized before staining for the indicated cell surface markers and for either phospho-ERK(p44/42) (Cell Signaling Technology), phospho-Syk, or phospho-Tyr (BD Biosciences) for 1 h.
Quantification of anti-Sm and IgM in mouse serum and supernatants was done by ELISA as previously described (23).
Affinity controls subset differentiation
To investigate the role of affinity in anti-Sm B cell regulation, we generated 2-12H/Vκ4 mice that produce high-affinity anti-Sm B cells for comparison with 2-12H and 2-12H/Vκ8 B cells. 2-12H/Vκ4 mice generated Sm-staining B cells of the splenic Tr, FO, and MZ subsets and the peritoneal B-1 subset (Fig. 1,A and data not shown). This resembles 2-12H mice but differs from 2-12H/Vκ8 mice, which generate few or no MZ or B-1 B cells (Fig. 1,A and Ref. 15). Mice of all three strains had similar numbers of splenic B cells, which was about half the number found in non-Tg mice due to a decrease in FO B cell numbers (Fig. 1,B). This reduction was not caused by a rapid turnover rate (Fig. 1,C) or PC differentiation, because BrdU incorporation was similar to that of non-Tg mice and anti-Sm Ab levels were not elevated (Fig. 1 D and Refs. 15 and 22).
2-12H and 2-12H/Vκ4 B cells of the FO and MZ subsets (distinguished by CD21 and CD23 staining) stained with biotinylated Sm with differing intensities (Fig. 1,E). Because IgM expression levels influence the intensity of Sm staining, we determined the ratio of the median fluorescence intensity (MFI) for Sm binding to the MFI for IgM to assess the relative affinities of anti-Sm B cells in these mice. Differences in epitope specificity between B cells could also affect this ratio if the epitopes differed in availability. However, because mice of all three strains use the same H chain, their B cells likely recognize the same Sm epitope, although we have not formally demonstrated this. Based on this ratio, 2-12H/Vκ4 FO and MZ B cells were of high affinity, and 2-12H and 2-12H/Vκ8 B cells were of low affinity (Fig. 1 F). The lower Sm:IgM ratio for MZ B cells relative to FO B cells in 2-12H and 2-12H/Vκ4 mice, although not statistically significant, may be due to the low expression of endogenous H chains by MZ B cells (data not shown). These data, together with the subset analysis, indicate that the high-affinity anti-Sm B cells of 2-12H/Vκ4 mice have greater differentiative potential than the low-affinity anti-Sm B cells of 2-12H/Vκ8 mice.
Affinity controls effectiveness of anergy; subset identity controls its induction
Hyporesponsiveness to TLR and BCR activation is a hallmark of anergic B cells (1), and we previously demonstrated that low-affinity 2-12H/Vκ8 B cells are hyporesponsive to LPS (13, 15). To determine TLR responsiveness of 2-12H/Vκ4 B cells, we stimulated sorted Tr, FO, and MZ B cells from non-Tg, 2-12H, and 2-12H/Vκ4 mice (Fig. 2,A) with LPS and CpG and measured secreted Ab levels. 2-12H Tr B cells were not different in Ab secretion from non-Tg Tr B cells, whereas FO B cells secreted substantially less and MZ B cells only somewhat less Ab than their non-Tg counterparts (Fig. 2,B). In contrast, 2-12H/Vκ4 B cells of all subsets secreted significantly less Ab than did their non-Tg and 2-12H counterparts (Fig. 2,B). LPS and CpG hyporesponsiveness was not due to an overall diminished responsiveness to these TLR agonists, because both induced robust proliferation of 2-12H and 2-12H/Vκ4 B cells (Fig. 2 C). To our knowledge, this is the first demonstration of hyporesponsive MZ B cells. Thus, although the low- and high-affinity anti-Sm B cells of 2-12H and 2-12H/Vκ4 mice, respectively, have similar differentiative potential, they differ significantly in function. The discordant regulation of FO and MZ B cells in 2-12H and the concordant regulation of FO and MZ B cells in 2-12H/Vκ4 mice suggest that anergy induction in FO and MZ B cell is independent and that MZ B cells have a higher affinity threshold for anergy than FO B cells.
Subset identity determines mechanism
To determine the contribution of subset identity to anergy, we examined survival and PC differentiation following LPS stimulation. PC differentiation in response to LPS was readily detected by the increased number of IgMhighB220lowCD138+ B cells that also had high levels of the PC-specific transcription factors XBP-1 and IRF-4 (Fig. 3,A) (25, 26). Because these changes occurred in concert, we used the frequency of IgMhighXBP-1high cells as a measure of PC differentiation. As shown in Fig. 3,B, high-affinity 2-12H/Vκ4 Tr B cells were three times less likely to survive LPS stimulation than non-Tg Tr B cells (∼15 vs 45%), and consistently fewer of the surviving cells underwent PC differentiation (∼25% vs 37%). Low-affinity 2-12H Tr B cells exhibited only modest differences from non-Tg B cells (Fig. 3 B). Thus, high-affinity anti-Sm Tr B cells are less responsive to LPS than low-affinity Tr B cells due principally to decreased survival.
The mechanism of LPS hyporesponsiveness changed with Tr to FO B cell differentiation. 2-12H/Vκ4 B cell survival improved ∼2-fold (∼15% to 30%), but the ability of surviving B cells to undergo PC differentiation decreased ∼5-fold (from ∼25 to 5%) (Fig. 3 B). 2-12H B cells exhibit a similar shift to a blockade of PC differentiation. Because 2-12H/Vκ4 Tr and FO B cells express the same BCR, we conclude that differences inherent to B cells of each subset are responsible for this shift in mechanism.
MZ B cells were the most responsive B cells in both 2-12H and 2-12H/Vκ4 mice. 2-12H/Vκ4 MZ B cells were more likely to survive LPS activation and to undergo PC differentiation than either 2-12H/Vκ4 Tr or FO B cells, although they were nonetheless hyporesponsive relative to non-Tg MZ B cells (Fig. 3,B). In contrast, LPS responsiveness by 2-12H MZ B cells was indistinguishable from that of non-Tg MZ B cells (Fig. 3,B), consistent with their Ab secretion ability (Fig. 2 B). These data corroborate the findings with Ab secretion, indicating regulation of high-affinity, but not low-affinity, MZ B cells.
To examine the responsiveness of anti-Sm B cells to BCR-mediated activation signals, we initially determined their proliferative response to BCR cross-linking. Non-Tg B cells cultured for 72 h with an F(ab′)2 anti-IgM proliferated extensively as opposed to 2-12H and 2-12H/Vκ4 B cells, which died without proliferating, similar to cultures of nonstimulated B cells (Fig. 4 A). Because only FO B cells proliferate in response to BCR cross-linking (Tr and MZ B cells normally die in response to BCR cross-linking) (27, 28, 29), these data indicate that anti-Sm FO B cells are defective in response to BCR cross-linking.
To determine the ability of B cells from individual subsets to respond to BCR signaling, we measured general protein Tyr phosphorylation as well as Syk and ERK phosphorylation by flow cytometry (30). To control for experimental variation, we mixed non-Tg B cells with either 2-12H or 2-12H/Vκ4 B cells at a 1:1 ratio for in vitro stimulation and distinguished 2-12H or 2-12H/Vκ4 B cells from non-Tg B cells during analysis with IgMa staining. CD21, CD23, and B220 expression distinguished FO and MZ B cells (Fig. 4,B). BCR cross-linking induced a similar increase in phospho-Tyr and phospho-Syk levels in 2-12H, 2-12H/Vκ4, and non-Tg FO B cells (Fig. 4,C). ERK phosphorylation was similar in non-Tg and 2-12H/Vκ4 B cells, but consistently greater in 2-12H FO B cells (Fig. 4 C). Thus, high- and low-affinity anti-Sm FO B cells respond as well or better to BCR cross-linking than non-Tg FO B cells, indicating that the inability to survive in response to BCR cross-linking is not due to a general failure of BCR signal transduction.
In contrast, high- and low-affinity anti-Sm MZ B cells differed in BCR signaling ability. 2-12H/Vκ4 MZ B cells consistently exhibited low phospho-Tyr, phospho-Syk, and phospho-ERK levels after BCR cross-linking in comparison to 2-12H and non-Tg MZ B cells (Fig. 4,C), whereas 2-12H MZ B cells exhibited high levels of protein phosphorylation, equal in magnitude and duration to those of non-Tg MZ B cells. Thus, high-affinity anti-Sm MZ B cells exhibit a functional uncoupling of the BCR from its signalsome, whereas low-affinity anti-Sm MZ B cells appear to be functional and undergo normal BCR signaling, paralleling the findings with TLR responsiveness (Figs. 2 and 3). Taken together, the analysis of BCR-mediated activation of 2-12H/Vκ4 B cells indicates that the mechanism of anergy in FO and MZ B cells is inherently different.
We demonstrate in this study that BCR affinity and subset identity make unique contributions to autoreactive B cell regulation. Increasing BCR affinity for Sm induces more effective blockade of TLR- and BCR-mediated activation while simultaneously promoting selection into multiple mature B cell subsets. In turn, subset identity determines the mechanism of anergy and the affinity threshold for anergy induction.
BCR affinity and anergy
Affinity for Sm determines anergy’s effectiveness. With increasing affinity, anti-Sm B cells of all subsets secrete less Ab, are less likely to survive, and are less able to undergo PC differentiation in response to LPS or CpG (Figs. 2 and 3 and Ref. 15). 2-12H and 2-12H/Vκ4 B cells exhibit the same shift upon Tr to FO B cell differentiation from a primary reliance on cell death to a blockade of PC differentiation (Fig. 3,B), suggesting that increasing affinity does not induce a fundamental change in mechanism. Thus, the primary effect of BCR affinity to anti-Sm B cell anergy appears to be in determining its effectiveness. A similar relationship between affinity and the effectiveness of receptor editing has been demonstrated with anti-hen egg lysozyme (HEL) B cells (31). One mechanism that could account for these data is that chronic BCR engagement with self-Ag in vivo has led to blockade of PC differentiation through activation of the Ras/MEK/ERK pathway, as demonstrated by Goodnow and colleagues for anti-HEL B cells (10, 11). Although we have not added Sm to the LPS and CpG activation cultures (Figs. 2 and 3), apoptotic cells expose Sm and, thus, Ag is likely to be present in anti-Sm B cell cultures and to increase over the 3- to 4-day culture period. However, inhibition of ERK phosphorylation does not restore LPS-induced Ab secretion by 2-12H or 2-12H/Vκ4 B cells (R.D. and S.H.C., unpublished observation), arguing against the involvement of the Ras/MEK/ERK pathway in anti-Sm B cells. Thus, investigation into the mechanism for LPS and CpG hyporesponsiveness by anti-Sm B cells may identify additional biochemical pathways that block T-independent activation of autoreactive B cells by bacterial products or endogenous DNA.
BCR affinity indirectly affects the mechanism of anergy through its influence on subset differentiation. High-affinity anti-Sm B cells differentiate into FO and MZ B cells, whereas low-affinity anti-Sm B cells only differentiate into FO B cells (Fig. 1 and Ref 15). This agrees with models in which a stronger BCR signal is required for MZ and B-1 B cell differentiation than for FO B cell differentiation (32). Affinity may not be the only determining factor in MZ B cell differentiation, but assuming that 2-12H/Vκ8 and 2-12H/Vκ4 B cells recognize the same Sm epitope, affinity is the most likely factor responsible for the differences observed here. Because BCR signaling can induce cell death of MZ B cells, it is likely that the signal delivered by Sm binding exceeds the threshold level required for positive selection into the MZ B cell subset, but not that required for deletion. Alternatively, because Sm is associated with TLR ligands, such as RNA and DNA (33, 34), the ability to induce a TLR signal may block any BCR-mediated death signal or provide a survival signal. It is notable that high-affinity 2-12H/Vκ4 B cells are not selected into the MZ and B-1 B cell subsets to the exclusion of the FO B cell subset. This resembles B cells specific for glucose-6-phosphate isomerase (35) but differs from B cells specific for phosphatidyl choline, Thy-1, and certain B cell clonotypes expressing VH81X (24, 32, 36, 37). The affinity for Sm may be borderline for MZ B cell selection, or Sm availability may be limited, either of which could restrict the number of B cells that receive a selecting signal. Regardless, the increased likelihood of high-affinity anti-Sm B cell differentiation into the MZ and B-1 subsets may be relevant to autoimmunity, because B cells of both subsets have a preactivated phenotype that allows rapid PC differentiation (38, 39), and anti-Sm MZ and B-1 B cells are activated in autoimmune mice (22).
The discordant regulation of low-affinity FO and MZ B cells in 2-12H mice indicates that anergy induction in FO and MZ B cells occurs independently and that FO and MZ B cells have different affinity thresholds for anergy induction. Thus, the affinity threshold for anergy in FO B cells is lower than that for MZ B cells. The 2-12H/Vκ4 affinity likely exceeds even the higher affinity threshold for MZ B cell anergy, accounting for the regulation of both FO and MZ B cells. Different affinity thresholds for FO and MZ B cell anergy would explain how the MZ B cell repertoire in normal mice becomes substantially more self-reactive than the FO B cell repertoire (21). This mechanism could account for why allelically included anti-self B cells generated by receptor editing are specifically selected into the MZ B cell subset (40).
High-affinity anti-Sm B cell differentiation into FO and MZ B cells is dependent on the absence of competition with nonself B cells. 2-12H/Vκ4 B cells do not mature past the Tr B cell stage in the presence of a high frequency of nonself B cells, and already mature 2-12H/Vκ4 B cells undergo rapid cell death when transferred to non-Tg littermates (R. Busick, R. Diz, and S. H. Clarke, manuscript in preparation). Competitive elimination regulates anti-HEL B cells, because anti-self B cells have a greater dependence on the B cell survival factor BAFF (41). 2-12H/Vκ4 B cells are also poorly responsive to BAFF, suggesting that a similar mechanism controls high-affinity anti-Sm B cells. We previously demonstrated that low-affinity 2-12H/Vκ8 B cells are not regulated by competitive elimination, indicating that BCR affinity for Sm controls this process (15). Thus, competitive elimination is an additional layer of regulation for high-affinity, but not low-affinity, anti-Sm B cells, providing a greater degree of protection from the potentially more pathogenic anti-Sm B cell population. The regulation of high-affinity anti-Sm FO and MZ B cells described in this article would maintain tolerance if competitive elimination is overcome, as may occur in systemic lupus erythematosus, where BAFF concentrations are high (42, 43).
Subset identity and anergy
Our analysis of 2-12H/Vκ4 B cells indicates that the B cell subset determines the mechanism of anergy. First, the mechanism changes during differentiation. The primary mechanism of TLR hyporesponsiveness shifts from failure to deliver a survival signal at the Tr B cell stage to failure to undergo PC differentiation at the FO B cell stage (Fig. 3,B). Differentiation to a MZ B cell results in an improvement in survival and PC differentiation, even though it remains hyporesponsive to LPS (Figs. 2 and 3) and is clearly deficient in the ability to be activated by a BCR signal (Fig. 4,C). Second, BCR signal disruption in anti-Sm FO and MZ B cells occurs at different levels in the cascade. 2-12H/Vκ4 FO B cells undergo Tyr, Syk, and ERK phosphorylation in response to BCR cross-linking as well or better than non-Tg FO B cells (Fig. 4,C), indicating that, similar to anti-insulin B cells (44), the failure to survive BCR cross-linking (Fig. 4 A) is not due to a gross failure of BCR signaling. In contrast, 2-12H/Vκ4 MZ B cells exhibit a functional uncoupling of the BCR from its signalsome. Because Syk phosphorylation is blocked, the defect in BCR signaling must be early in the signaling cascade, possibly through physical dissociation of the BCR from its Igα/Igβ signaling component (45). BCR affinity is not responsible for inducing these intersubset differences, because B cells of both subsets express the same BCR, but MZ B cells may receive a stronger BCR signal than FO B cells because of higher IgM levels or greater access to Ag by virtue of their anatomical location. Thus, FO and MZ B cells are intrinsically different in the mechanism of anergy.
The differences in the affinity threshold for anergy and the mechanism of BCR and TLR nonresponsiveness between FO and MZ B cells are in line with other differences in BCR signaling consequences between B cells of these two subsets. In addition to differences in signaling requirements for differentiation (21), MZ B cells differentiate into plasma cells much faster than FO B cells in response to Ag engagement with the BCR (46). They also become competent Ag presenters to T cells much more rapidly than FO B cells (47). In vitro, BCR cross-linking induces MZ B cell death, but the same signal induces FO B cell proliferation. This difference in response to BCR signaling is evident in vivo, as B cell superantigen containing bacteria induce an acute loss of MZ B cells due to cell death with no apparent effect on FO B cell numbers (48). Thus, the differences between MZ and FO B cells in the mechanism of anergy and affinity threshold setting are reflective of a fundamental difference in the consequences of BCR signaling between the B cells of these subsets.
In summary, we present evidence that increasing affinity for Sm improves anergy’s effectiveness while simultaneously driving anti-Sm B cell differentiation into the MZ and B-1 B cell subsets. Subset identity in turn determines the mechanism and the affinity threshold. Understanding the underlying mechanisms for these subset-specific differences in anergy may prove useful for developing therapeutic strategies that target specific subsets.
We gratefully acknowledge the Flow Cytometry Facility at the University of North Carolina (Chapel Hill, NC) for their assistance with this work, and thank Dr. Barbara Vilen for critical review of the manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grants AI29576 and AI43587 and a grant from the Arthritis Foundation.
Abbreviations used in this paper: PC, plasma cell; FO, follicular; HEL, hen egg lysozyme; IRF-4, interferon response factor 4; MFI, mean fluorescence intensity; MZ, marginal zone; Sm, Smith; Tg, transgenic; Tr, transitional; XBP-1, X box-binding protein 1.