The Journal of Immunology
B Cell Tolerance Checkpoints That Restrict Pathways of Antigen-Driven Differentiation1 Jacqueline William,* Chad Euler,† Nicole Primarolo,† and Mark J. Shlomchik2*,† Autoreactive B cells can be regulated by deletion, receptor editing, or anergy. Rheumatoid factor (RF)-expressing B lymphocytes in normal mice are not controlled by these mechanisms, but they do not secrete autoantibody and were presumed to ignore self-Ag. Surprisingly, we now find that these B cells are not quiescent, but instead are constitutively and specifically activated by self-Ag. In BALB/c mice, RF B cells form germinal centers (GCs) but few Ab-forming cells (AFCs). In contrast, autoimmune mice that express the autoantigen readily generate RF AFCs. Most interestingly, autoantigen-specific RF GCs in BALB/c mice appear defective. B cells in such GCs neither expand nor are selected as efficiently as equivalent cells in autoimmune mice. Thus, our data establish two novel checkpoints of autoreactive B cell regulation that are engaged only after initial autoreactive B cell activation: one that allows GCs but prevents AFC formation and one that impairs selection in the GC. Both of these checkpoints fail in autoimmunity. The Journal of Immunology, 2006, 176: 2142–2151.
H
igh autoantibody titers are a hallmark of systemic autoimmune diseases such as rheumatoid arthritis (RA)3 and systemic lupus erythematosus (SLE) (1, 2). This must represent a breakdown of tolerance mechanisms that normally regulate autoreactive B cells. Central tolerance prevents the development of high-affinity self-reactive B cells. Mechanisms of central tolerance, such as deletion, anergy, and receptor editing, were demonstrated using mice with Ig transgenes (Tg) specific for artificial autoantigens, like hen egg lysozyme or MHC class I proteins (3– 6). There are also peripheral tolerance mechanisms for B cells (7), but these are much less well defined. To initiate systemic autoimmunity, either central tolerance mechanisms must be overcome or mechanisms that regulate peripheral B cells must break down. Our laboratory and others have used Ig-Tg models to study how B cells specific for disease-related self-Ags are regulated (8 –15). In SLE, only certain Ags efficiently elicit autoreactive B cells: chromatin, ribonucleoproteins, and IgG (2, 16). These autoantigens must have special properties that render them preferred targets when immune regulation is abnormal. These properties, which include the ability to provide an endogenous ligand for certain Toll-like receptors (17, 18), are currently being elucidated. It is thus important to study B cells specific for relevant autoantigens. We have focused on the regulation of rheumatoid factor (RF)expressing B cells that bind to the constant region of Ig (19). RFs are prevalent in RA patients but are also produced by patients with a number of other systemic autoimmune diseases such as Sjogren’s *Section of Immunobiology and †Department of Laboratory Medicine, Yale University School of Medicine, New Haven, CT 06520 Received for publication July 7, 2005. Accepted for publication November 28, 2005. 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. 1 This work was supported by National Institute of Health Grants P01 AI36529 and R01 AR44077. 2 Address correspondence and reprint requests to Dr. Mark J. Shlomchik, Department of Laboratory Medicine, Yale University School of Medicine, 333 Cedar Street, Box 208035, New Haven, CT 06520-8035. E-mail address:
[email protected] 3 Abbreviations used in this paper: RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; Tg, transgene; RF, rheumatoid factor; AFC, Ab-forming cell; GC, germinal center; PNA, peanut agglutinin; TNP, trinitrophenol; KLH, keyhole limpet hemocyanin; MLN, mesenteric lymph node; IC, immune complex; LN, lymph node.
Copyright © 2006 by The American Association of Immunologists, Inc.
syndrome, mixed cryoglobulinemia, and SLE (2, 20, 21), and by Fas-deficient lpr mice (22). To study RF B cells, we have used AM14 H chain Tg mice. When this H chain is paired with a Tg-encoded or endogenous V8 L chain, the result is an Ab that binds the Fc portion of IgG2aa, but not IgG2ab (23). The AM14 Tg mouse is an ideal model because it combines specificity for a disease-related autoantigen and the ability to study the regulation of the B cells in the presence or absence of autoantigen (in IgHa or IgHb congenic mice). This Tg system uniquely allows for measurement and control of the presence or absence of a disease-related autoantigen. This ability is useful for distinguishing specific from nonspecific B cell activation. We have found that B cells in autoantigen-positive (IgHa) BALB/c mice Tg for both the AM14 H and L chains were neither anergized nor deleted during development (23, 24). However, in Fas-deficient H and HL Tg mice, RF B cells differentiated into numerous Ab-forming cells (AFCs). The formation of AFCs was autoantigen-driven, because IgHb congenic mice that lacked the autoantigen failed to produce RF AFCs (25). We have recently further investigated the fates of RF B cells in autoimmune-prone mice, revealing a dominant short-lived plasmablast response (26, 27). However, the differences between RF B cell fates as revealed by our recent studies of MRL/lpr mice and our earlier work on BALB/c mice (23, 24) indicated the presence of additional mechanisms that regulate naive but fully developed RF B cells. These mechanisms are presumably intact in BALB/c mice but are breached in MRL/lpr mice. In this study, we have used the AM14 model to identify these mechanisms by comparing the regulation of RF B cells in MRL/lpr to the BALB/c and MRL ⫹/⫹ strains. We found that normal BALB/c mice do not have RF AFCs; at this level, the mice remain apparently self-tolerant. However, we were surprised to find that despite the lack of RF AFCs, RF B cells are not quiescent in BALB/c mice. Rather, they actively participate in germinal centers (GCs). Importantly, this activation of RF B cells is autoantigenspecific, because it is absent in the autoantigen-negative IgHb strains. Furthermore, RF B cells undergo somatic hypermutation in GCs of BALB/c mice, which could potentially lead to the generation of high-affinity autoantibodies. Nonetheless, analysis of somatic hypermutation patterns of V regions from RF B cells in 0022-1767/06/$02.00
The Journal of Immunology BALB/c mice compared with those from autoimmune mice revealed a failure to efficiently select and expand RF B cell clones. Overall, our data show that Ag-specific autoimmunity can proceed much further than originally thought even in normal animals, but that diversion from AFC production and regulation in the GC can prevent the pathologic consequences.
Materials and Methods Mice AM14 H Tg MRL ⫹/⫹ mice were made by backcrossing Tg BALB/c mice (23) at least 10 generations to MRL ⫹/⫹ mice (The Jackson Laboratory). Congenic Tg BALB/c IgHb mice (CB.17) were made by crossing AM14Tg BALB/c to CB.17 mice (Taconic). Tg MRL ⫹/⫹ IgHb mice were made by similar crosses between the AM14 Tg MRL ⫹/⫹ mice and MRL ⫹/⫹ IgHb mice. MRL ⫹/⫹ IgHb mice were made by a series of crosses of MRL/lpr IgHb (a gift from Robert Eisenberg, University of Pennsylvania, Philadelphia, PA (28)) with MRL ⫹/⫹ mice and screening for homozygous Faswt mice. Mice were housed under specific pathogen-free conditions.
Histology Cells were stained as described previously (29). The following Abs were used: 4-44 FITC (anti-Id), anti-CD21/35-biotin (CR1; BD Pharmingen), peanut agglutinin (PNA)-biotin (Vector Laboratories), 30H12-biotin (antiCD90.2), and 30H12-FITC. Unless purchased as indicated, Abs were made and conjugated as described previously (23). For immunohistochemistry, anti-FITC-alkaline phosphatase (Molecular Probes) and streptavidin-HRP (Southern Biotechnology Associates) were developed with Fast Blue BB or 3-amino-9-ethyl-carbazole (Sigma-Aldrich), as described previously (24).
ELISPOTs ELISPOTs were performed as described previously (24).
Sequencing 4-44⫹ clusters were microdissected from stained sections, V8/J4 rearranged DNA amplified with PFU Turbo (Stratagene), cloned, and sequenced as described previously (25). Sequences were aligned using Lasergene DNA analysis software (DNASTAR), and all mutation data was compiled into a custom database program written in Filemaker Pro for further statistical analysis.
Cell isolation and FACS analysis FACS analysis was performed as described previously (24). The following Abs were used: 4-44-biotin (anti-Id), 4-44-Alexa 488, anti-phagocytic glycoprotein-1-FITC (anti-CD44), anti-CD80-PE (BD Pharmingen), antiCD22.2-FITC (BD Pharmingen), and anti-CD22.2-PE (BD Pharmingen). Streptavidin-allophycocyanin (Serologicals) was used to detect biotinylated reagents. Propidium iodide was used to detect dead cells.
Immunizations H Tg mice were immunized i.p. with 50 g of the 2,4,6-trinitrophenol (TNP)-specific IgG2a Ab, Hy1.2, complexed with 12.5 g of TNP-keyhole limpet hemocyanin (KLH) in alum. Alum-precipitated Hy1.2-TNP-KLH complexes (24) were made by incubating purified Hy1.2 with TNP-KLH at the appropriate ratio in PBS for 1 h at 37°C. Spleens were harvested on days 5, 8, and 12 of immunization.
Results
Autoantigen drives the formation of RF⫹ GCs, but very few AFCs, in the spleens of BALB/c and MRL ⫹/⫹ mice In AM14 H Tg mice, a few percentage of B cells that naturally express certain V8 L chains are specific for IgG2aa but not IgG2ab (23, 25–27). This population is detected by the mAb 4-44, which is specific for the AM14 H chain when combined with one of two homologous V8 genes, each of which confers RF specificity. H Tg mice are ideal to study how an autoreactive B cell is regulated because of the ease of detecting clonal expansion and activation of RF B cells. To determine how autoantigen determines the fate of RF B cells in normal (BALB/c), autoimmune-prone (MRL ⫹/⫹), or autoim-
2143 mune-prone/Fas-deficient (MRL/lpr) mice, spleens and mesenteric lymph nodes (MLNs) of IgHa and IgHb Tg mice of ages ranging from 10 to 44 wk were examined by histology and FACS. Unexpectedly, we found that RF⫹ (i.e., 4-44⫹) GCs were common in the follicles of most normal Tg BALB/c spleens (Fig. 1G, M). GCs were also found in MRL ⫹/⫹ spleens (Fig. 1, H, J, N, and P). In contrast, CB.17 (IgHb) and IgHb MRL ⫹/⫹ congenic strains that lacked the autoantigen had very few 4-44⫹ cells and no detectable 4-44⫹ GCs (Fig. 1, E, F, K, L, Q, and R). The absence of RF⫹ GCs in the IgHb mice demonstrates that the RF⫹ GCs in BALB/c and MRL ⫹/⫹ mice were spontaneously induced by self-Ag. These data are quantitated in Fig. 2A and corroborated by FACS data presented below. Although both normal BALB/c and autoimmune-prone MRL ⫹/⫹ strains had RF⫹ GCs, they had few dark staining 4-44⫹ cells outside of follicles (Fig. 1, A–C, G–I, and S–U). In contrast, MRL/ lpr mice had very few RF⫹ GCs but rather had large numbers of RF AFCs at the T zone-red pulp border (Figs. 1, C, I, and O, and 2A), as reported previously (25). These darkly staining extrafollicular cells very likely represent AFCs, as documented in detail in Ref. 27. In agreement with the histology, the median numbers of AFCs/million splenocytes in the Tg BALB/c, MRL ⫹/⫹ were 25and 16-fold lower than in MRL/lpr mice, respectively (Fig. 2B). Because MRL ⫹/⫹ mice develop autoimmunity much later in life than MRL/lpr mice (30), we also examined spleens from Tg MRL ⫹/⫹ IgHa mice 40 wk of age and older. These spleens had RF⫹ GCs (Fig. 1, J and P), like those of younger mice. However, in contrast to the younger mice, the spleens of older mice had clusters of darkly staining 4-44⫹ cells at the T zone-red pulp border, similarly to younger Tg MRL/lpr mice (Fig. 1D). Thus, with sufficient age, RF B cells were driven by autoantigen to AFC production, even in Fas-intact but autoimmune-prone mice. Failure to regulate RF B cell AFC differentiation was specific to the MRL strains, because even older BALB/c mice never developed appreciable numbers of such cells (data not shown). Autoantigen drives the accumulation of RF B cells in BALB/c and MRL ⫹/⫹ mice In addition to differentiating into either GC cells or AFCs, as just described, RF⫹ (4-44⫹) cells accumulated in the spleens of IgHa mice. Mice were examined at 18 – 46 wk of age (Fas-sufficient strains) or 11–22 wk of age (Fas-deficient strains). Among total live lymphocytes, 4-44⫹ cells comprised 4.5% in Tg BALB/c spleens vs 1.6% in CB.17 ( p ⬍ 0.0024) and 4.4% in MRL ⫹/⫹ vs 1.9% in MRL ⫹/⫹ IgHb ( p ⬍ 0.0001). However, presence of the autoantigen did not cause a significant increase in 4-44⫹ cells in the MRL/lpr vs MRL/lpr.IgHb (1.5 vs 1.2%; p ⫽ 0.68; Fig. 3, A–F and M), despite the increased numbers of AFCs (Fig. 2B). Because 4-44⫹ AFCs in Tg MRL/lpr IgHa mice express normal levels of surface Ig (27), they would have been included in the FACS analysis. Their failure to accumulate is most likely because the AFCs are plasmablasts, and their rapid death rate prevents substantial numerical accumulation (27). CD22 is down-regulated on AFCs, but expressed at normal levels on mature and activated B cells (27). Consistent with the large numbers of RF AFCs, a significant proportion of the 4-44⫹ cells in IgHa MRL/lpr mice was CD22low (36%). However, only a very small percentage of the 4-44⫹ cells in BALB/c and MRL ⫹/⫹ mice was CD22low (2.4 and 3.7%, respectively; Fig. 3N), again in concert with the paucity of AFCs in these strains. The frequencies of CD22low cells in BALB/c and MRL ⫹/⫹ mice were not significantly different from their IgHb congenics. Thus, the increase in 4-44⫹ cells in BALB/c and MRL ⫹/⫹ mice is due to relative
2144
REGULATING AUTOREACTIVE B CELLS AFTER ACTIVATION
FIGURE 1. Histologic sites of autoantigen-driven RF B cell activation and differentiation. A–F, 4-44 spleen sections from IgHa and IgHb Tg BALB/c, MRL ⫹/⫹, and MRL/lpr mice were stained with 4-44 (blue) and anti-Thy 1.2 (red). 4-44⫹ cells are distributed in a follicular and marginal zone pattern in IgHa BALB/c and MRL ⫹/⫹ spleens (A and B). 4-44⫹ cells in IgHa MRL/lpr spleens have very dark (blue) cytoplasmic staining and accumulate outside the B cell follicles in the T zone-red pulp borders (C). Adjacent sections (G–L) were stained with 4-44 (blue) and PNA (red), to identify GCs (arrows). M–R, Magnifications of the boxed regions shown in G–L. Note 4-44⫹ cells in the GCs of BALB/c and MRL ⫹/⫹ Tg mice (G, H, M, and N). However, Tg MRL/lpr spleens rarely have 4-44⫹ cells in their GCs, as shown by the 4-44-negative GC in I and O. Spleens from aged (⬎40-wk-old) IgHa Tg MRL ⫹/⫹ mice (J and P) partially resemble Tg MRL/lpr spleens (D), but are notable for the additional presence of 4-44⫹ GCs (arrows). No accumulation of 4-44⫹ cells was noted in IgHb mice, which lack the autoantigen (E, F, K, L, Q, and R). S–X, Adjacent spleen sections of IgHa Tg BALB/c (S and V), MRL ⫹/⫹ (T and W), and MRL/lpr (U and X) mice were stained with anti-Thy1.2 (red)/4-44 (blue) (S–U) and CR1 (red)/Thy1.2 (blue) (V–X). These panels demonstrate the follicular (BALB/c and MRL ⫹/⫹) and extrafollicular (MRL/lpr) locations of 4-44⫹ cells. For Thy1.2/PNA/4-44 stains to document GCs, data are representative of 18 BALB/c, 17 MRL ⫹/⫹, and 72 MRL/lpr mice. CR1 staining was performed on at least four mice per group.
expansion of the CD22high subpopulation in the IgHa vs IgHb strains. To determine the location of the expanded population of RF B cells in the Tg BALB/c and MRL ⫹/⫹mice, we stained sections with anti-CD21, which darkly stains follicular dendritic cells (Fig. 1, S–X). In Tg IgHa BALB/c and MRL ⫹/⫹ mice, 4-44⫹ cells are in follicles, colocalizing with CD21, whereas 4-44⫹ B cells in the Tg MRL/lpr mice generally do not colocalize with CD21 and are extrafollicular. Thus, the presence of self-Ag not only stimulates a GC response in BALB/c and MRL ⫹/⫹ mice, it also leads to accumulation of CD22high, RF B cells that reside in follicles and are neither in GCs nor are AFCs. It is important to note in comparing FACS and histologic data that the latter has a much lower dynamic range and that rare scattered cells are difficult to appreciate if weakly stained. In contrast, clusters of activated 4-44⫹ cells, most likely with increased intracellular and surface Ig, are more easily appreciated on histologic analysis. In the presence of autoantigen, 4-44⫹ CD22high cells have increased expression of costimulatory molecules To determine whether this expansion of CD22high RF B cells in the Fas-sufficient strains is in part due to activation of RF B cells, we
analyzed expression of B cell activation markers by FACS. The percentage of CD22high RF B cells that express CD44 and CD80 was substantially higher in the presence of autoantigen (IgHa strains; Fig. 3, G, H, J–K, P, and Q). Interestingly, although 4-44⫹/ CD22high compartment of cells did not expand as a percentage of total lymphocytes in MRL/lpr mice compared with MRL/lpr.IgHb congenics, they were nonetheless activated, as indicated by the increased fraction of them that expressed CD44 and CD80 (Fig. 3, I, L, P, and Q). Because GC cells are CD22high and express activation markers like CD80, it was not surprising that we found increased numbers of 4-44⫹/CD22high cells that also had higher levels of CD44 and/or CD80 in IgHa Fas-sufficient mice compared with their IgHb counterparts (Fig. 3, P and Q; see legend for p values). However, we were surprised to find that in these same strains, quiescent 4-44⫹/ CD22high cells were also increased in frequency in IgHa vs IgHb mice. This was indicated by the percentages of 4-44⫹/ CD22highCD80low cells (Fig. 3O), which are substantially higher in IgHa than IgHb strains (BALB/c; p ⫽ 0.06; MRL ⫹; p ⫽ 0.0067). Analysis of 4-44⫹/CD22highCD44low cells led to similar conclusions (data not shown). Thus, Fas-sufficient mice, including the normal strain BALB/c, not only activate RF B cells in the
The Journal of Immunology
2145
FIGURE 2. Numbers of RF⫹ AFCs and GCs in Tg BALB/c, MRL ⫹/⫹, and MRL/lpr mice. A, The number of 4-44⫹ GCs per plane of section (bars are 1 SEM). The number of 4-44⫹ GCs in MRL/lpr spleens compared with BALB/c and MRL ⫹/⫹ spleens is significantly lower (p ⬍ 0.0001). B, 4-44⫹ AFCs were detected by ELISPOT. Each diamond is an individual mouse and bars represent means. Tg IgHa MRL/lpr spleens have much greater numbers of 4-44⫹ AFCs compared with Tg IgHa BALB/c and MRL ⫹/⫹ spleens (p ⬍ 0.0001).
presence of autoantigen, but they also accumulate RF B cells with a resting phenotype. Such accumulation could represent resting RF memory B cells that had been activated previously and/or naive B cells that are positively selected in the presence of autoantigen but without frank activation (31–33).
that other accessory factors in the mice (e.g., costimulatory signals) differ in the induced vs spontaneous response. Finally, because GCs but not AFCs were induced in both IgHa and IgHb congenics, we conclude that the presence of the endogenous autoantigen does not suppress the RF response.
Immunization with IgG2a-containing immune complexes (ICs) induces a strong GC response but a very weak AFC response in all strains
Activation of RF B cells in the lymph node (LN)
Data presented above show that Fas-sufficient mice activate B cells and make GCs, but generate few AFCs; whereas, Fas-deficient mice generate few GCs in spleen but large numbers of AFCs. This could be due to inherent differences between B cells in these strains. Alternatively, it is possible that the nature of the autoantigen stimulus is different in each strain and that this determines the quality of the response. To investigate these possibilities, we immunized Tg BALB/c and MRL/lpr mice (and their IgHb allotype congenics) with IgG2a-containing ICs comprised of IgG2aa anti-TNP and TNPKLH in alum. We used 5- to 6-wk-old mice; Tg MRL/lpr mice at this age do not have spontaneous 4-44⫹ AFCs, and their splenic architecture is normal (26). All strains formed RF⫹ GCs, histologically similar to those that formed spontaneously in Tg BALB/c mice (Fig. 4, E–H). However, very few 4-44⫹ AFCs were formed, even in the Fas-deficient or IgHb mice, as indicated by a lack of darkly staining 4-44⫹ cells outside of follicles (compare with Fig. 1, I and J) and as confirmed by ELISPOT analysis (data not shown). Responses in all strains were also similar on days 5 (data not shown) and 12 (Fig. 4, I and J, and data not shown). These results indicate that ICs of IgG2a and foreign protein are sufficient to stimulate RF B cells, but they elicit mainly GCs rather than AFCs, regardless of strain background. Because Tg MRL/lpr mice form GCs in response to immunization, we conclude that the absence of spontaneous RF⫹ GCs in MRL/lpr spleens (Fig. 1) was due neither to an inherent inability of the RF B cells to form GCs nor to an environment that could not support GCs. Rather, these results make it more likely that either the endogenous form of the IgG2a autoantigen is different from injected ICs, leading to a different quality of B cell stimulus and thus differentiation path, or
Having defined autoantigen-specific RF B cell activation in the spleen, we next investigated LNs. Like the spleen, MLNs of Tg BALB/c and MRL ⫹/⫹ mice frequently had 4-44⫹ GCs (Fig. 5). As for the spleen, such GCs were absent in IgHb congenics (data not shown), indicating that those in IgHa mice are autoantigen driven. Despite the absence of RF⫹ GCs in their spleens, a fraction of MRL/lpr mice did have RF⫹ GCs in MLNs (Fig. 5, C and G) and Peyer’s patches (data not shown). BALB/c and MRL/lpr mice had similar average numbers of 4-44⫹ GCs in their MLNs, whereas MRL ⫹/⫹ mice had three times as many (Fig. 5D). In addition to GCs, most MRL/lpr and some MRL ⫹/⫹ MLNs also had numerous darkly staining 4-44⫹ cells occupying large portions of the medullary region (Fig. 5, B and C); the dark staining and characteristic location indicate that these cells represent AFCs. A qualitative assessment of these indicated that MRL/lpr MLNs had approximately three times as many of these cells as the MRL ⫹/⫹ MLNs (Fig. 5H). These dark-staining cells in the medulla were absent in BALB/c MLNs (Fig. 5, A and H). Thus, at the level of GC formation, regulation of autoreactive RF B cells is lost in all three strains in LN, but only MRL/lpr and MRL ⫹ mice had AFCs. Consequently, in the LN, like in the spleen, a tolerance mechanism that prevents AFC formation operates in BALB/c mice but is only partially effective in MRL ⫹ mice and is ineffective in MRL/lpr mice. Different patterns of somatic mutations in BALB/c and MRL/lpr GCs Considering that BALB/c mice are not autoimmune-prone, we did not expect to find autoantigen-driven RF⫹ GCs in both spleen and LNs. It would seem risky to allow selection of higher affinity mutant (autoreactive) B cells, as could occur in the GC. Therefore, it was important to determine whether such selection was regulated
2146
REGULATING AUTOREACTIVE B CELLS AFTER ACTIVATION
FIGURE 3. Phenotype of activated 4-44⫹ cells in Tg BALB/c, MRL ⫹/⫹, and MRL/lpr mice. A–F, Splenocytes from Tg IgHa BALB/c (A), MRL ⫹/⫹ (B), MRL/lpr (C), and congenic IgHb Tg mice (D–F) were analyzed by FACS for CD22 (x-axis) and 4-44 (y-axis). G–L, Histograms of gated 4-44⫹, CD22high splenocytes showing expression of CD44 (G–I) and CD80 (J–L) for IgHa (bold line) and IgHb (dotted line) mice. M–Q, Each diamond is an individual mouse, and bars represent means. M, Percentage of splenocytes that are 4-44⫹. N, Percentage of gated 4-44⫹ splenocytes that are CD22low. O, Percentage of live splenocytes that are resting 4-44⫹ cells, as indicated by a 4-44⫹, CD22high and CD80low phenotype. P and Q, Percentage of gated 4-44⫹, CD22high splenocytes that are CD44high and CD80high, respectively. IgHa strains have significantly higher percentages than their IgHb counterparts (CD44: BALB/c, p ⬍ 0.01; MRL ⫹/⫹ and MRL/lpr p ⫽ 0.0005; CD80: BALB/c, p ⫽ 0.023; MRL ⫹/⫹, p ⫽ 0.0011; MRL/lpr, p ⬍ 0.005).
in any way within BALB/c GCs. Because GCs were present in the MLN of all three strains, there was an opportunity to compare across strains the degree of somatic hypermutation and the patterns of these mutations that reflect selection.
4-44⫹ GCs from MLNs of Tg BALB/c, MRL ⫹/⫹, and MRL/ lpr mice were microdissected, and the endogenous V8 L chains were sequenced (25). PCR products using V8-specific primers (and containing V8 –19 or V8 –28 genes; Ref. 25) were readily
The Journal of Immunology
2147
FIGURE 4. Immunization of Tg BALB/c and MRL/lpr mice with IgG2a-containing ICs results in the formation of RF⫹ GCs but few AFCs. Tg IgHa and Tg IgHb BALB/c and MRL/lpr mice, as indicated on the individual panels, were immunized with either IgG2a-containing complexes (I.C.) in alum (E–J) or with alum alone (A–D). All spleen sections were stained with 4-44 (blue) and PNA (red). Images are from spleens harvested at day 8 (A–H) or day 12 (I and J). Alum-only immunized mice form no purple 4-44⫹ GCs, although there are GCs of unknown specificity (red). Note the lack of histologically identifiable 4-44⫹ AFCs, which would stain more darkly than the GCs and would be at the border of the red pulp or in the red pulp. Each panel is representative of at least three mice per group, all with similar appearance.
recovered from nearly all microdissections, confirming the histologic identification of 4-44⫹ GCs. Sequence analysis revealed significant differences among the phenotypically similar GCs in the three strains. There were 2.6, 4.2, and 5.1 mutations per sequence in the BALB/c, MRL ⫹/⫹, and MRL/lpr GCs, respectively ( p ⬍ 0.001 for BALB/c vs MRL/lpr; p ⫽ 0.03 for BALB vs MRL ⫹⫹; and p ⫽ not significant for MRL ⫹/⫹ vs MRL/lpr). The BALB/c GCs were also more polyclonal: for every 100 cells within the BALB/c, MRL ⫹/⫹, and MRL/lpr GCs, there were 12, 9, and 5 different clones, respectively (Fig. 6, A–C, and analysis not
shown). The smaller number of mutations and greater number of clones in BALB/c GCs both indicate less clonal expansion compared with the MRL strains. The differences in mutation frequency and clonal composition also suggest that selection of RF B cells was different in these GCs. Differences in selection are also reflected in the shapes of the genealogic trees that are derived from the patterns of shared and unique mutations (34). From each microdissection, we used maximum parsimony to reconstruct genealogic trees that reflect clonal evolution. To quantitate the extent of branching in these trees, each
FIGURE 5. MLNs in all three strains contain 4-44⫹ GCs, but AFCs are found only in MRL strains. A–C and E–G, MLNs of IgHa Tg BALB/c, MRL ⫹/⫹, and MRL/lpr mice stained with PNA (red) and 4-44 (blue). A–C, Magnification, ⫻100. E–G, Enlargement of boxed areas in A–C, demonstrating the presence of numerous 4-44⫹ cells in GCs in all three strains. Note the large number of darkly staining 4-44⫹ cells in the medullary region of MLNs of Tg IgHa MRL ⫹/⫹ and MRL/lpr (B and C) but not BALB/c (A) mice. D, The number of 4-44⫹ GCs per LN section was tallied in each mouse, and averages were plotted (bars are 1 SEM). Data are compiled from representative single sections from 11 BALB/c, 14 MRL ⫹/⫹, and 44 MRL/lpr mice with p values for MRL ⫹/⫹ vs the other two strains ⬍0.002, and p ⫽ not significant between BALB/c and MRL/lpr. H, The number of AFCs in each LN section was estimated using an AFC score that represents the approximate percentage of each MLN spanned by cells with dark cytoplasmic staining (see Materials and Methods). Histograms represent averages of 11 BALB/c, 14 MRL ⫹/⫹, and 54 MRL/lpr LN examined. BALB/c vs MRL ⫹/⫹ and vs MRL/lpr are both highly significant (p ⬍ 0.0001), whereas MRL ⫹/⫹ and MRL/lpr are not significantly different (p ⫽ 0.07).
2148
REGULATING AUTOREACTIVE B CELLS AFTER ACTIVATION
FIGURE 6. Decreased clonal expansion and diversification in Tg IgHa BALB/c MLN GCs. Sample genealogical trees generated from sequences derived from microdissections of individual MLN GCs of Tg IgHa BALB/c (A), MRL ⫹/⫹ (B), and MRL/lpr (C) mice. Individual clones were assigned to a position in a genealogical tree on the basis of shared and unique mutations. All members of a tree are from a single GC, although some GCs yielded more than one tree (e.g., A). Bases of trees show the six nucleotides at the VJ junction of the parental germline rearrangement. By convention, sequences that shared junctions but no other mutations are depicted in separate trees, because there are many clones with the same canonical VJ junction. The serial numbers of sequenced clones from each pick are written inside the circles, with the pick name followed by the clone number. Numbers in parenthesis indicate the number of identical clones retrieved from each pick. For example, in the BALB/c tree (A), the sequence 20b4-47 was identical with four other sequences obtained from the 20b4 pick. Mutations and their codon position are written on the sides of branches, with affected regions (either CDR or framework region (FR)) in parenthesis. Amino acid changes are written below the codons. Silent mutations are in italics. Sequences written in red are nonviable sequences that have resulted from an insertion, deletion, or a stop codon mutation. Dark circles represent inferred precursors and intermediates in clonal evolution, which were not recovered. D–F, Pie charts representing the distribution of mutations among the branches of genealogical trees inferred from V region sequences from GCs of each strain. Mutations were classified according to their occurrence in the genealogical trees as follows: T(Un) and T(Br) are both mutations shared by all sequences of a tree. Unbranched (Un) mutations are those in trunks without branches above them, and branched (Br) mutations are those in trunks that do have branches extending from them. Branch mutations were labeled B(1-n), depending on the level of branching on the tree. Examples of each of these categories are written in blue on the trees (A–C). Data are derived from microdissections of 10 BALB/c, 9 MRL ⫹/⫹, and six MRL/lpr GCs from which were obtained 62, 49, and 36 unique sequences.
mutation was given a “branch” or “trunk” designation, depending on its position in the genealogical tree (see Fig. 6, A–C, and legend); then the distributions of all mutations from each strain were plotted as pie charts (Fig. 6, D–F). Genealogic trees without any selection should have many branches, and nearly all mutations should be branch mutations, a result of the mutation rate of ⬃1 mutation/cell/division (35). Selection of mutant cells (either positive or negative) will distort the “natural” shape of the trees, and the type of selection will affect the resultant shape (36). It is clear that trees from BALB/c LN are shaped differently from the MRL ⫹/⫹ and MRL/lpr-derived trees. BALB/c trees had shapes resembling blades of grass. Most of the mutations were “trunk unbranched” mutations, with very few branch mutations. This indicates that most of the BALB/c mutants
did not survive, either because of negative selection against autoreactive cells or more likely failure to positively select. Lack of selection that would lead to oligoclonal expansion is consistent with the larger number of clones and few mutations/sequence found in BALB/c GCs. In contrast, trees from Tg MRL ⫹/⫹ and MRL/lpr GCs were bushier, with a much higher proportion of branch mutations, as well as trunks with branches above them (Fig. 6, B, C, E, and F). This indicates that, at the least, there is positive selection for certain mutants and may also indicate less stringent negative selection. This view is also consistent with the low replacement: silent ratios in CDRs of BALB/c GCs (data not shown). These differences in mutation number, clonal composition, and tree shape all indicate that BALB/c GCs, in contrast to MRL ⫹/⫹ or MRL/lpr GCs, do not efficiently generate expanded clones of mutant RF B cells.
The Journal of Immunology
Discussion B cells that recognize self can escape the dominant mechanisms of central tolerance (23, 37–39). It has been widely assumed that these escaped cells are quiescent in the peripheral lymphoid tissues of normal animals (40). It is felt that peripheral autoreactive B cells are regulated by clonal anergy or are clonally ignorant and do not sense the presence of self-Ag; or if activated, they are quickly eliminated (41). In this study, we demonstrate that this is not always the case: we show that RF B cells are activated and participate in GC reactions in the periphery of BALB/c mice. Moreover, this activation specifically requires the presence of self-Ag. This demonstration that disease-related autoreactive B cells can be activated by autoantigen to the point of GC development in normal individuals is important because it expands and alters our view of how autoimmunity is normally kept in check. All of the previously known regulatory checkpoints for self-reactive B cells operate to prevent initial activation or to eliminate cells quickly after they are activated (41– 46). In this study, we have found two new ways by which autoreactive B cells are prevented from causing harm to the host even after they have been activated. First, the generation of RF AFCs was suppressed in BALB/c mice, whereas MRL/lpr mice and older MRL ⫹/⫹ mice generated large numbers of AFCs (Figs. 1–3). We propose that the suppression of AFC formation serves to prevent the harmful effects of pathogenic autoantibodies. Blocking AFC generation, however, would not fully prevent the adverse effects of activating autoreactive B cells. Such B cells could also promote autoimmunity by acting as APCs for pathogenic self-reactive T cells (47, 48). Thus, the generation in GCs of high-affinity RF B cells is likely to be detrimental. The second means by which RF responses were restricted in BALB/c mice addresses this potential problem: there was less clonal expansion and weaker selection of mutants in GCs of normal BALB/c mice than in MRL/lpr or MRL ⫹/⫹ GCs (Figs. 5 and 6). Such regulation of clonal expansion and selection of mutants would naturally limit the generation of high-affinity RF B cells. We propose that controlling selection and clonal expansion of autoreactive cells in the normal GC is an additional form of self-tolerance. How are AFC formation and selection in GCs regulated in BALB/c mice compared with MRL mice? Such regulation could be B cell intrinsic, or relate to exogenous factors like T cell help or regulation or even the form of autoantigen present in the different strains. Whether the regulation is B cell intrinsic or simply reflects the requirement for certain forms of Ag that are more prevalent in autoimmune-prone mice, such mechanisms would constitute tolerance checkpoints in that they result in the restriction of autoreactive B cell immune responses in terms of differentiation and affinity maturation. With regard to B cell intrinsic mechanisms, MRL ⫹ or MRL/lpr B cells themselves may be more likely to undergo AFC differentiation and/or to survive in the GC. The differences between Fasdeficient and sufficient strains demonstrate that Fas plays a role. This could be intrinsic to the B cell, because GC B cells express Fas (49 –51). There is genetic evidence in MRL/lpr of a B cellintrinsic defect that leads to autoimmunity (52–54). Fas, which transmits a death signal (55), could therefore serve to regulate GC responses as has been suggested for GC responses to model hapten Ags (56). Fas is also expressed on AFCs (49, 50), but there is little data on whether Fas normally influences AFC development. Because older MRL ⫹/⫹, but not BALB/c, mice do make RF AFCs, Fas cannot be the only explanation for RF AFC development in the autoimmune-prone strains. Some of the polymorphic genes associated with the MRL autoimmune phenotype (57, 58)
2149 must play a role in promoting AFC formation in the Fas-sufficient MRL mice. Although chimera experiments suggest a role for intrinsic B cell defects in MRL ⫹ mice (53, 59), whether any of the already-identified MRL genes are expressed in B cells has not been reported. Nonetheless, in New Zealand Black/New Zealand Whitebased models of systemic autoimmunity, there are clear examples of genes with polymorphisms associated with autoimmunity that are expressed in B cells (60 – 62). B cell-extrinsic factors could also determine whether autoreactive B cells differentiate into AFCs and are selected in GCs. These include the form and concentration of the autoantigen as well as stimulatory and regulatory roles of T cells. Certain forms of the RF autoantigen may be required to provide the signals needed to induce AFCs. IgG2a complexed with chromatin or CpG-containing DNA (17, 63) is a potent mitogen for RF B cells in vitro. This ligand provides a combination of BCR and TLR-mediated signals. These signals also promote the initiation of AFC differentiation (Ann Marshak-Rothstein, personal communication). In contrast, IgG2a complexed with proteins is a poor inducer of RF B cell proliferation in vitro (17, 63). It is possible that these in vitro findings are paralleled in vivo in AM14 Tg mice. Supporting this notion, immunization with IgG2a anti-protein ICs did not induce AFCs in any of the strains but instead elicited a GC response (Fig. 4). If form of autoantigen controls the nature of the response, we would expect BALB/c mice to have relatively small amounts of DNA-containing ICs and possibly more of classical antiprotein ICs. In this view, MRL/lpr mice should have high levels of DNAcontaining ICs, as has been shown to be the case (64). A number of genes that predispose to systemic autoimmunity might control the levels of DNA-containing ICs and thus the form of the RF autoantigen. For example, natural and artificial mutations that affect clearance of apoptotic cells and ICs, such as deficiency in early complement components or the tyrosine kinase Mer, are associated with systemic autoimmunity (65, 66). We think that this explanation could well account for the differences between normal and autoimmune-prone mice and could be a way that genetic predisposition could be linked to regulation of B cell activation. The influence of T cells could also be important in controlling the fates of activated RF B cells. This is easiest to understand for GCs, the complete development of which is known to be T celldependent. T cells have been shown to be hyperreactive in MRL mice and in patients with SLE (67, 68); hyperreactive T cells in MRL mice could provide signals that override normal controls that operate in BALB/c mice. Conversely, there is some evidence that CD4⫹CD25⫹ T cells can regulate the extent of the GC reaction in normal mice (69). Such cells can also regulate organ-specific autoimmunity in a variety of model systems (70). This has been less well-investigated in systemic autoimmunity, with evidence for a role in some but not other model systems (71). Lupus patients are reportedly deficient in CD4⫹CD25⫹ cell numbers (72). However, we are not aware of a direct test of the function of these cells in regulating the activation of autoreactive B cells in MRL mice, although in a transfer system they can prevent the activation of anti-DNA B cells by T cells that recognize Ag constitutively presented by these B cells (73). Although BALB/c spleens and LNs lack RF⫹ AFCs, the presence of RF⫹ GCs in BALB/c secondary lymphoid tissue poses a constant risk for functional autoimmunity. Left unchecked, proliferating and mutating low-affinity self-reactive GC B cell could differentiate into high-affinity memory clones or long-lived AFCs (74 –77). Thus, the normal control of GC dynamics in BALB/c LNs could be an important fail-safe mechanism of tolerance. If GC
2150
REGULATING AUTOREACTIVE B CELLS AFTER ACTIVATION
tolerance mechanisms were to be breached, the type of autoimmunity associated with SLE would result: the generation of highaffinity, isotype-switched autoantibodies. There is indeed evidence that selection in GCs of humans with SLE is impaired (78). It is interesting to speculate that failure of GC tolerance mechanisms is what eventually leads to autoimmunity in MRL ⫹/⫹ mice, the GCs of which do not appear as stringently regulated as those in BALB/c. Prior work has suggested that there could be self-tolerance in the GC (79 – 81). These studies did not directly demonstrate this with autoreactive B cells in that two of them used a synthetic hapten toleragen (79, 80). Hande et al. (81) came closest to this point in showing that Tg-mediated overexpression of bcl-2 led to accumulation of autoreactive B cells during the immune response to a nominal Ag. Our results make a direct link by showing regulation of autoreactive B cells responding to self-Ag and how this is relaxed in lupus-prone mice. Taking all of these results into account, we predict that GCs that cannot properly regulate autoreactive B cells, such as GCs that form ectopically or that are comprised of lymphocytes with intrinsic signaling defects, would be associated with pathogenic autoimmune responses. Indeed, ectopic (and poorly organized) GCs are associated with RA, Sjogren’s syndrome, and possibly other autoimmune diseases (82– 84). The MRL/lpr spleen may represent an extreme in which mutation and selection are entirely dissociated from normal GC structure (25). The existence of checkpoints that might regulate autoreactive B cells after they have already been activated highlights the complexity of the problem of preventing autoimmunity while promoting immunity. Although such checkpoints do exist, it is unclear how self and nonself could be discriminated at this stage. Regulation of activated lymphocytes could thus have the side effect of inhibiting responses to pathogens as well as responses to self. However, regulating autoreactive B cells even after their initial activation may have been an evolutionary adaptation to further control autoimmunity while at the same time maintaining immunity to pathogens-two inherently conflicting requirements of a functional immune system (40, 41).
Acknowledgments We thank Martin Weigert, Ann Marshak-Rothstein, and Ann Haberman for useful comments on the manuscript and discussion. We thank the technical staff of Yale Animal Resources for expert animal care that enabled the success of these studies.
Disclosures The authors have no financial conflict of interest.
References 1. Lipsky, P. E. 2001. Systemic lupus erythematosus: an autoimmune disease of B cell hyperactivity. Nat. Immunol. 2: 764 –766. 2. Tan, E. M. 1989. Antinuclear antibodies: diagnostic markers for autoimmune diseases and probes for cell biology. Adv. Immunol. 44: 93–151. 3. Tiegs, S. L., D. M. Russell, and D. Nemazee. 1993. Receptor editing in selfreactive bone marrow B cells. J. Exp. Med. 177: 1009 –1020. 4. Nemazee, D. A., and K. Burki. 1989. Clonal deletion of B lymphocytes in a transgenic mouse bearing anti-MHC class I antibody genes. Nature 337: 562– 566. 5. Goodnow, C. C., J. Crosbie, S. Adelstein, T. B. Lavoie, S. J. Smith-Gill, R. A. Brink, H. Pritchard-Briscoe, J. S. Wotherspoon, R. H. Loblay, K. Raphael, et al. 1988. Altered immunoglobulin expression and functional silencing of selfreactive B lymphocytes in transgenic mice. Nature 334: 676 – 682. 6. Hartley, S. B., J. Crosbie, R. Brink, A. B. Kantor, A. Basten, and C. C. Goodnow. 1991. Elimination from peripheral lymphoid tissues of self-reactive B lymphocytes recognizing membrane-bound antigens. Nature 353: 765–769. 7. Russell, D. M., Z. Dembic, G. Morahan, J. F. A. P. Miller, K. Burki, and D. Nemazee. 1991. Peripheral deletion of self-reactive B cells. Nature 354: 308 – 311. 8. Erikson, J., M. Z. Radic, S. A. Camper, R. R. Hardy, and M. G. Weigert. 1991. Expression of anti-DNA immunoglobulin transgenes in non-autoimmune mice. Nature 349: 331–334.
9. Chen, C., Z. Nagy, M. Z. Radic, R. R. Hardy, D. Huszar, S. A. Camper, and M. Weigert. 1995. The site and stage of anti-DNA B-cell deletion. Nature 373: 252–255. 10. Li, Y., H. Li, and M. Weigert. 2002. Autoreactive B cells in the marginal zone that express dual receptors. J. Exp. Med. 195: 181–188. 11. Santulli-Marotto, S., M. W. Retter, R. Gee, M. J. Mamula, and S. H. Clarke. 1998. Autoreactive B cell regulation: peripheral induction of developmental arrest by lupus-associated autoantigens. Immunity 8: 209 –219. 12. Li, H., Y. Jiang, H. Cao, M. Radic, E. L. Prak, and M. Weigert. 2003. Regulation of anti-phosphatidylserine antibodies. Immunity 18: 185–192. 13. Friedmann, D., N. Yachimovich, G. Mostoslavsky, Y. Pewzner-Jung, A. Ben-Yehuda, K. Rajewsky, and D. Eilat. 1999. Production of high affinity autoantibodies in autoimmune New Zealand Black/New Zealand white F1 mice targeted with an anti-DNA heavy chain. J. Immunol. 162: 4406 – 4416. 14. Mandik-Nayak, L., A. Bui, H. Noorchashm, A. Eaton, and J. Erikson. 1997. Regulation of anti-double-stranded DNA B cells in nonautoimmune mice: localization to the T-B interface of the splenic follicle. J. Exp. Med. 186: 1257–1267. 15. Mandik-Nayak, L., S. Seo, A. Eaton-Bassiri, D. Allman, R. R. Hardy, and J. Erikson. 2000. Functional consequences of the developmental arrest and follicular exclusion of anti-double-stranded DNA B cells. J. Immunol. 164: 1161–1168. 16. Shan, H., M. J. Shlomchik, A. Marshak-Rothstein, D. S. Pisetsky, S. Litwin, and M. G. Weigert. 1994. The mechanism of autoantibody production in an autoimmune MRL/lpr mouse. J. Immunol. 153: 5104 –5120. 17. Leadbetter, E. A., I. R. Rifkin, A. M. Hohlbaum, B. C. Beaudette, M. J. Shlomchik, and A. Marshak-Rothstein. 2002. Chromatin-IgG complexes activate B cells by dual engagement of IgM and Toll-like receptors. Nature 416: 603– 607. 18. Christensen, S. R., M. Kashgarian, L. Alexopoulou, R. A. Flavell, S. Akira, and M. J. Shlomchik. 2005. Toll-like receptor 9 controls anti-DNA autoantibody production in murine lupus. J. Exp. Med. 202: 321–331. 19. Haberman, A. M., J. William, C. Euler, and M. J. Shlomchik. 2003. Rheumatoid factors in health and disease: structure, function, induction and regulation. Curr. Dir. Autoimmun. 6: 169 –195. 20. Howard, T. W., M. J. Iannini, J. J. Burge, and J. T. Davis. 1991. Rheumatoid factor, cryoglobulinemia, anti-DNA, and renal disease in patients with systemic lupus erythematosus. J. Rheumatol. 18: 826 – 830. 21. Markusse, H., H. Otten, T. Vroom, T. Smeets, N. Fokens, and F. Breedveld. 1993. Rheumatoid factor isotypes in serum and salivary fluid of patients with primary Sjogren’s syndrome. Clin. Immunol. Immunopathol. 66: 26 –32. 22. Izui, S., M. Abdelmoula, Y. Gyotoku, G. Lange, and P. H. Lambert. 1984. IgG rheumatoid factors and cryoglobulins in mice bearing the mutant gene lpr (lymphroliferation). Rheumatol. Int. 4(Suppl): 45– 48. 23. Shlomchik, M. J., D. Zharhary, T. Saunders, S. A. Camper, and M. G. Weigert. 1993. A rheumatoid factor transgenic mouse model of autoantibody regulation. Int. Immunol. 5: 1329 –1341. 24. Hannum, L. G., D. Ni, A. M. Haberman, M. G. Weigert, and M. J. Shlomchik. 1996. A disease-related rheumatoid factor autoantibody is not tolerized in a normal mouse: implications for the origins of autoantibodies in autoimmune disease. J. Exp. Med. 184: 1269 –1278. 25. William, J., C. Euler, S. Christensen, and M. J. Shlomchik. 2002. Evolution of autoantibody responses via somatic hypermutation outside of germinal centers. Science 297: 2066 –2070. 26. William, J., C. Euler, E. Leadbetter, A. Marshak-Rothstein, and M. J. Shlomchik. 2005. Visualizing the onset and evolution of an autoantibody response in systemic autoimmunity. J. Immunol. 174: 6872– 6878. 27. William, J., C. Euler, and M. J. Shlomchik. 2005. Short-lived plasmablasts dominate the early spontaneous rheumatoid factor response: differentiation pathways, hypermutating cell types, and affinity maturation outside the germinal center. J. Immunol. 174: 6879 – 6887. 28. Halpern, M. D., S. Y. Craven, P. L. Cohen, and R. A. Eisenberg. 1993. Regulation of anti-Sm autoantibodies by the immunoglobulin heavy chain locus. J. Immunol. 151: 7268 –7272. 29. Hannum, L. G., A. M. Haberman, S. M. Anderson, and M. J. Shlomchik. 2000. Germinal center initiation, variable gene region hypermutation, and mutant B cell selection without detectable immune complexes on follicular dendritic cells. J. Exp. Med. 192: 931–942. 30. Hewicker, M., and G. Trautwein. 1987. Sequential study of vasculitis in MRL mice. Lab. Anim. 21: 335–341. 31. Levine, M. H., A. M. Haberman, D. B. Sant’Angelo, L. G. Hannum, M. P. Cancro, C. A. Janeway, Jr., and M. J. Shlomchik. 2000. A B-cell receptorspecific selection step governs immature to mature B cell differentiation. Proc. Natl. Acad. Sci. USA 97: 2743–2748. 32. Hayakawa, K., M. Asano, S. A. Shinton, M. Gui, D. Allman, C. L. Stewart, J. Silver, and R. R. Hardy. 1999. Positive selection of natural autoreactive B cells. Science 285: 113–116. 33. Wang, H., and S. H. Clarke. 2003. Evidence for a ligand-mediated positive selection signal in differentiation to a mature B cell. J. Immunol. 171: 6381– 6388. 34. McKean, D., K. Huppi, M. Bell, L. Straudt, W. Gerhard, and M. Weigert. 1984. Generation of antibody diversity in the immune response of BALB/c mice to influenza virus hemagglutinin. Proc. Natl. Acad. Sci. USA 81: 3180 –3184. 35. Kleinstein, S. H., Y. Louzoun, and M. J. Shlomchik. 2003. Estimating hypermutation rates from clonal tree data. J. Immunol. 171: 4639 – 4649. 36. Mehr, R., H. Edelman, D. Sehgal, and R. Mage. 2004. Analysis of mutational lineage trees from sites of primary and secondary Ig gene diversification in rabbits and chickens. J. Immunol. 172: 4790 – 4796.
The Journal of Immunology 37. Souroujon, M., S. M. White, J. Andreschwartz, M. L. Gefter, and R. S. Schwartz. 1988. Preferential autoantibody reactivity of the preimmune B cell repertoire in normal mice. J. Immunol. 140: 4173– 4179. 38. Nemazee, D., and V. Sato. 1983. Induction of rheumatoid factors in the mouse: regulated production of autoantibody in secondary humoral response. J. Exp. Med. 158: 529 –545. 39. Welch, M. J., S. Fong, J. Vaughan, and D. Carson. 1983. Increased frequency of rheumatoid factor precursor B lymphocytes after immunization of normal adults with tetanus toxoid. Clin. Exp. Immunol. 51: 200 –304. 40. Nossal, G. J. V. 1994. Negative selection of lymphocytes. Cell 76: 229 –239. 41. Goodnow, C. C. 1996. Balancing immunity and tolerance: deleting and tuning lymphocyte repertoires. Proc. Natl. Acad. Sci. USA 93: 2264 –2271. 42. Satoh, M., and W. H. Reeves. 1994. Induction of lupus-associated autoantibodies in BALB/c mice by intraperitoneal injection of pristane. J. Exp. Med. 180: 2341–2346. 43. Bolland, S., and J. V. Ravetch. 2000. Spontaneous autoimmune disease in Fc␥RIIB-deficient mice results from strain-specific epistasis. Immunity 13: 277–285. 44. Morel, L., B. P. Croker, K. R. Blenman, C. Mohan, G. Huang, G. Gilkeson, and E. K. Wakeland. 2000. Genetic reconstitution of systemic lupus erythematosus immunopathology with polycongenic murine strains. Proc. Natl. Acad. Sci. USA 97: 6670 – 6675. 45. Lang, J., M. Jackson, L. Teyton, A. Brunmark, K. Kane, and D. Nemazee. 1996. B cells are exquisitely sensitive to central tolerance and receptor editing induced by ultralow affinity, membrane-bound antigen. J. Exp. Med. 184: 1685–1697. 46. Nemazee, D., and M. Weigert. 2000. Revising B cell receptors. J. Exp. Med. 191: 1813–1817. 47. Chan, O. T. M., M. P. Madaio, and M. J. Shlomchik. 1999. The central and multiple roles of B cells in lupus pathogenesis. Immunol. Rev. 169: 107–121. 48. Shinde, S., R. Gee, S. Santulli-Marotto, L. K. Bockenstedt, S. H. Clarke, and M. J. Mamula. 1999. T cell autoimmunity in Ig transgenic mice. J. Immunol. 162: 7519 –7524. 49. Nishiuchi, R., T. Yoshino, Y. Matsuo, I. Sakuma, L. Cao, Y. Seino, K. Takahashi, and T. Akagi. 1996. The Fas antigen is detected on immature B cells and the representative cell lines show Fas-mediated apoptosis. Br. J. Haematol. 92: 302–307. 50. Mandik, L., K. A. Nguyen, and J. Erikson. 1995. Fas receptor expression on B-lineage cells. Eur. J. Immunol. 25: 3148 –3154. 51. Smith, K. G., G. J. Nossal, and D. M. Tarlinton. 1995. FAS is highly expressed in the germinal center but is not required for regulation of the B-cell response to antigen. Proc. Natl. Acad. Sci. USA 92: 11628 –11632. 52. Sobel, E. S., T. Katagiri, K. Katagiri, S. C. Morris, P. L. Cohen, and R. A. Eisenberg. 1991. An intrinsic B cell defect is required for the production of autoantibodies in the lpr model of murine systemic autoimmunity. J. Exp. Med. 173: 1441–1449. 53. Sobel, E. S., V. N. Kakkanaiah, J. Schiffenbauer, E. A. Reap, P. L. Cohen, and R. A. Eisenberg. 1998. Novel immunoregulatory B cell pathways revealed by lpr-⫹ mixed chimeras. J. Immunol. 160: 1497–1503. 54. Nemazee, D., C. Guiet, K. Buerki, and A. Marshak-Rothstein. 1991. B lymphocytes from the autoimmune-prone mouse strain MRL/lpr manifest an intrinsic defect in tetraparental MRL/lpr DBA/2 chimeras. J. Immunol. 147: 2536 –2539. 55. Itoh, N., S. Yonehara, A. Ishii, M. Yonehara, S.-I. Mizushima, M. Sameshima, A. Hase, Y. Seto, and S. Nagata. 1991. The polypeptide encoded by the cDNA for human cell surface antigen fas can mediate apoptosis. Cell 66: 233–243. 56. Takahashi, Y., H. Ohta, and T. Takemori. 2001. Fas is required for clonal selection in germinal centers and the subsequent establishment of the memory B cell repertoire. Immunity 14: 181–192. 57. Watson, M. L., J. K. Rao, G. S. Gilkeson, P. Ruiz, E. M. Eicher, D. S. Pisetsky, A. Matsuzawa, J. M. Rochelle, and M. F. Seldin. 1992. Genetic analysis of MRL-lpr mice: relationship of the Fas apoptosis gene to disease manifestations and renal disease-modifying loci. J. Exp. Med. 176: 1645–1645. 58. Vidal, S., D. H. Kono, and A. N. Theofilopoulos. 1998. Loci predisposing to autoimmunity in MRL-Faslpr and C57BL/6-Faslpr mice. J. Clin. Invest. 101: 696 –702. 59. Kakkanaiah, V. N., E. S. Sobel, G. C. MacDonald, R. L. Cheek, P. L. Cohen, and R. A. Eisenberg. 1997. B cell genotype determines the fine specificity of autoantibody in lpr mice. J. Immunol. 159: 1027–1035. 60. Rozzo, S. J., J. D. Allard, D. Choubey, T. J. Vyse, S. Izui, G. Peltz, and B. L. Kotzin. 2001. Evidence for an interferon-inducible gene, Ifi202, in the susceptibility to systemic lupus. Immunity 15: 435– 443. 61. Mohan, C., L. Morel, P. Yang, and E. K. Wakeland. 1997. Genetic dissection of systemic lupus erythematosus pathogenesis: Sle2 on murine chromosome 4 leads to B cell hyperactivity. J. Immunol. 159: 454 – 465.
2151 62. Jongstra-Bilen, J., B. Vukusic, K. Boras, and J. E. Wither. 1997. Resting B cells from autoimmune lupus-prone New Zealand Black and (New Zealand Black ⫻ New Zealand White) F1 mice are hyperresponsive to T cell-derived stimuli. J. Immunol. 159: 5810 –5820. 63. Viglianti, G. A., C. M. Lau, T. M. Hanley, B. A. Miko, M. J. Shlomchik, and A. Marshak-Rothstein. 2003. Activation of autoreactive B cells by CpG dsDNA. Immunity 19: 837– 847. 64. Bjorkman, L., C. F. Reich, and D. S. Pisetsky. 2003. The use of fluorometric assays to assess the immune response to DNA in murine systemic lupus erythematosus. Scand. J. Immunol. 57: 525–533. 65. Manderson, A. P., M. Botto, and M. J. Walport. 2004. The role of complement in the development of systemic lupus erythematosus. Ann. Rev. Immunol. 22: 431– 456. 66. Scott, R. S., E. J. McMahon, S. M. Pop, E. A. Reap, R. Caricchio, P. L. Cohen, H. S. Earp, and G. K. Matsushima. 2001. Phagocytosis and clearance of apoptotic cells is mediated by MER. Nature 411: 207–211. 67. Vratsanos, G., S. Jung, Y. Park, and J. Craft. 2001. CD4⫹ T cells from lupusprone mice are hyperresponsive to T cell receptor engagement with low and high affinity peptide antigens: a model to explain spontaneous T cell activation in lupus. J. Exp. Med. 193: 329 –338. 68. Liossis, S.-N. C., X. Z. Ding, G. J. Dennis, and G. C. Tsokos. 1998. Altered pattern of TCR/CD3-mediated protein-tyrosyl phosphorylation in T cells from patients with systemic lupus erythematosus. J. Clin. Invest. 101: 1448 –1457. 69. Bystry, R. S., V. Aluvihare, K. A. Welch, M. Kallikourdis, and A. G. Betz. 2001. B cells and professional APCs recruit regulatory T cells via CCL4. Nat. Immunol. 2: 1126 –1132. 70. Maloy, K. J., and F. Powrie. 2001. Regulatory T cells in the control of immune pathology. Nat. Immunol. 2: 816 – 822. 71. Bagavant, H., C. Thompson, K. Ohno, Y. Setiady, and K. S. Tung. 2002. Differential effect of neonatal thymectomy on systemic and organ-specific autoimmune disease. Int. Immunol. 14: 1397–1406. 72. Liu, M. F., C. R. Wang, L. L. Fung, and C. R. Wu. 2004. Decreased CD4⫹CD25⫹ T cells in peripheral blood of patients with systemic lupus erythematosus. Scand. J. Immunol. 59: 198 –202. 73. Seo, S. J., M. L. Fields, J. L. Buckler, A. J. Reed, L. Mandik-Nayak, S. A. Nish, R. J. Noelle, L. A. Turka, F. D. Finkelman, A. J. Caton, and J. Erikson. 2002. The impact of T helper and T regulatory cells on the regulation of anti- doublestranded DNA B cells. Immunity 16: 535–546. 74. Takahashi, Y., P. R. Dutta, D. M. Cerasoli, and G. Kelsoe. 1998. In situ studies of the primary immune response to (4-hydroxy-3- nitrophenyl)acetyl. V. Affinity maturation develops in two stages of clonal selection. J. Exp. Med. 187: 885– 895. 75. Shlomchik, M. J., A. H. Aucoin, D. S. Pisetsky, and M. G. Weigert. 1987. Structure and function of anti-DNA antibodies derived from a single autoimmune mouse. Proc. Natl. Acad. Sci. USA 84: 9150 –9154. 76. Diamond, B., and M. D. Scharff. 1984. Somatic mutation of the T15 heavy chain gives rise to an antibody with autoantibody specificity. Proc. Natl. Acad. Sci. USA 81: 5841–5844. 77. Hoyer, B. F., K. Moser, A. E. Hauser, A. Peddinghaus, C. Voigt, D. Eilat, A. Radbruch, F. Hiepe, and R. A. Manz. 2004. Short-lived plasmablasts and long-lived plasma cells contribute to chronic humoral autoimmunity in NZB/W mice. J. Exp. Med. 199: 1577–1584. 78. Pugh-Bernard, A. E., G. J. Silverman, A. J. Cappione, M. E. Villano, D. H. Ryan, R. A. Insel, and I. Sanz. 2001. Regulation of inherently autoreactive VH4 –34 B cells in the maintenance of human B cell tolerance. J. Clin. Invest. 108: 1061–1070. 79. Pulendran, B., G. Kannourakis, S. Nouri, K. G. Smith, and G. J. Nossal. 1995. Soluble antigen can cause enhanced apoptosis of germinal-centre B cells. Nature 375: 331–334. 80. Shokat, K. M., and C. C. Goodnow. 1995. Antigen-induced B-cell death and elimination during germinal-centre immune responses. Nature 375: 334 –338. 81. Hande, S., E. Notidis, and T. Manser. 1998. Bcl-2 obstructs negative selection of autoreactive, hypermutated antibody V regions during memory B cell development. Immunity 8: 189 –198. 82. Hansen, A., P. E. Lipsky, and T. Dorner. 2003. New concepts in the pathogenesis of Sjogren syndrome: many questions, fewer answers. Curr. Opin. Rheumatol. 15: 563–570. 83. Berek, C., and A. E. Schroder. 1997. A germinal center-like reaction in the nonlymphoid tissue of the synovial membrane. Ann. NY Acad. Sci. 815: 211–217. 84. Randen, I., O. J. Mellbye, O. Forre, and J. B. Natvig. 1995. The identification of germinal centres and follicular dendritic cell networks in rheumatoid synovial tissue. Scand. J. Immunol. 41: 481– 486.
