The Influence of Effector T Cells and Fas Ligand on Lupus-Associated B Cells 1
http://www.100md.com
免疫学杂志 2005年第13期
Abstract
Circulating autoantibodies against dsDNA and chromatin are a characteristic of systemic lupus erythematosus in humans and many mouse models of this disease. B cells expressing these autoantibodies are normally regulated in nonautoimmune-prone mice but are induced to secrete Abs following T cell help. Likewise, anti-chromatin autoantibody production is T cell-dependent in Fas/Fas ligand (FasL)-deficient (lpr/lpr or gld/gld) mice. In this study, we demonstrate that Th2 cells promote anti-chromatin B cell survival and autoantibody production in vivo. FasL influences the ability of Th2 cells to help B cells, as Th2-gld/gld cells support higher titers of anti-chromatin Abs than their FasL-sufficient counterparts and promote anti-chromatin B cell participation in germinal centers. Th1 cells induce anti-chromatin B cell germinal centers regardless of FasL status; however, their ability to stimulate anti-chromatin Ab production positively correlates with their level of IFN- production. This distinction is lost if FasL-deficient T cells are used: Th1-gld/gld cells promote significant titers of anti-chromatin Abs regardless of IFN- production levels. Thus, FasL from effector T cells plays an important role in determining the fate of anti-chromatin B cells.
Introduction
Antibodies that bind dsDNA and chromatin are a serologic hallmark of systemic lupus erythematosus and are found in several murine models of systemic lupus erythematosus (1). Using the VH3H9 Ig transgenic (Tg) 3 model, a population of anti-chromatin B cells (VH3H9/V1) has been characterized in both healthy and autoimmune mice (2, 3, 4, 5, 6, 7). In healthy mice, these anti-chromatin B cells persist in the periphery but their Abs are undetected (2, 5). They have a decreased life span, are predominantly developmentally arrested and activated, and localize to the edges of the B cell follicles near the T cell areas in the spleen (5, 8). We hypothesize that this phenotype is a consequence of chronic exposure to Ag, in the absence of T cell help (9, 10, 11, 12, 13).
To dissect the responses of anti-chromatin B cells to CD4+ Th cells, we established an in vivo model of cognate interaction between these two cell types. Mice engineered to express the neoself Ag hemagglutinin (HA) on MHC class II-bearing cells (including B cells) were mated to VH3H9 Ig Tg mice (14, 15). Anti-HA CD4+ T cells from TCR Tg mice and HA-expressing anti-chromatin B cells were then transferred together into a third-party recipient mouse and their fates tracked (16, 17). Using this strategy, we have demonstrated that anti-chromatin B cells from healthy mice respond to CD4+ Th cells in vivo by producing autoantibodies (16). These findings contrast with those obtained from the hen-egg lysozyme model of B cell tolerance (18, 19), where autoreactive anti-hen-egg lysozyme B cells were shown to be resistant to CD4+ T cell help (20, 21, 22, 23). Importantly, Fas/ Fas ligand (FasL) interactions mediated this resistance (20, 21, 24).
Both Th1 and Th2 cells can help nonautoreactive B cells to produce Abs (25, 26, 27), although Th2 cells appear more efficient in this regard (28, 29, 30, 31). Th1 cells reportedly express higher levels of FasL than Th2 cells (32, 33, 34, 35), which could modulate T-B interactions by either limiting the availability of Th cells (32, 33, 35) and/or by direct killing of the autoreactive B cells (20, 21). Furthermore, Th1 cells may be more susceptible to Fas-mediated death even in cases where they express comparable levels of FasL (32, 33, 35).
Once activated, B cells can follow a number of differentiation pathways–toward short-lived Ab-forming cells (AFCs), memory B cells, or long-lived AFCs. Although much has been learned about the transcription factors that govern the fate decisions of B cells (36), less is known about the precise cues that determine these decisions. Given the critical role that CD4+ T cells play in autoantibody production, and the seemingly contradictory data on autoreactive B cell responses to T cell help (20, 21, 22, 23), we have investigated the impact of Th1 and Th2 cells, with and without FasL, on anti-chromatin B cells. We report here the differential capabilities of T effector cells to influence anti-chromatin B cells in vivo, and suggest that FasL may play a role in controlling the differentiation pathways and magnitude of the autoreactive B cell response.
Materials and Methods
Mice
TS1 BALB/c (FasL-sufficient or -deficient) and VH3H9 Tg/HACII/Ig–/– mice were bred in specific-pathogen-free conditions at The Wistar Institute under the approval and supervision of the Institutional Animal Care and Use Committee (IACUC), and genotyped as described (2, 6, 16, 17, 37). Similarly, site-directed VH3H9 Tg BALB/c mice in which the VH3H9 Tg was targeted to the JH locus (Refs. 38, 39, 40 ; generously provided by Dr. M. Weigert (University of Chicago, Chicago, IL) were bred to produce site-directed VH3H9 Tg/HACII/Ig–/– mice. CB17 mice were purchased from Charles River Laboratories. Only young (6–12 wk old) TS1 BALB/c-gld/gld mice were used. All other mice were used at 6–16 wk, and both genders were used.
Th1 and Th2 cell cultures
TS1 BALB/c or TS1 BALB/c-gld/gld lymph nodes were depleted (>90%) of CD8+ cells using anti-CD8 Dynalbeads (Dynaltech). A total of 0.5 x 106 CD8-depleted lymphocytes were then cultured for Th1/Th2 deviation as previously described (41). rIL-12 was a generous gift from Dr. G. Trinchieri (National Institutes of Health, Bethesda, MD) or was purchased from R&D Biosystems or PeproTech. Cells received fresh media containing IL-2 at days 3 and 5, and were rested in the absence of IL-2 at day 7. At day 9, cells were harvested and an aliquot cultured with PMA and ionomycin (Sigma-Aldrich) in the presence of brefeldin A (Cytofix/Cytoperm kit; BD Pharmingen) for another 4–6 h. Cells were stained for CD4 expression, fixed, permeabilized, and stained for intracellular cytokines, using anti-IFN--FITC, anti-IL-4-PE, and/or anti-IL-10-PE (BD Pharmingen) (42). The cells that were not restimulated with PMA/ionomycin were purified by centrifugation with Lympholyte M (Cedarlane Laboratories) before injection.
T cell/influenza virus injections
A total of 5–10 x 106 Th1- or Th2-deviated CD4+ cells from in vitro cultures were suspended in sterile PBS with 1000 hemagglutinating units (43) of purified PR8 influenza virus (15) and injected i.v.
Anti-chromatin B cell injections
Splenocytes from VH3H9 Tg/HACII/Ig–/– or site-directed VH3H9 Tg/HACII/Ig–/– mice were depleted of RBC and an aliquot was stained by flow cytometry to determine the frequency of anti-chromatin B cells (B220+ Ig1+). CB17 recipient mice (allotype Igb) were injected with spleen preparations containing 4–10 x 106 anti-chromatin B cells (allotype Iga). Control mice received B cells without previous injection of Th cells or virus.
Blockade of CD40-CD154 interactions
Mice were i.p. injected with 250 μg of purified anti-CD154 mAb (MR1) (a kind gift of Dr. R. Noelle, Dartmouth Medical School, Lebanon, NH) on the same day as B cell transfer, and 3 days afterward.
Chromatin ELISAs
Chromatin (a generous gift of Dr. M. Monestier, Temple University, Philadelphia, PA) was diluted to 2 μg/ml and plated overnight as described (44). ELISAs were done as described (45) with the following modifications: the block used was 1% BSA/PBS/azide, developing Abs were anti-Ig1, anti-IgMa, anti-IgG1, or anti-IgG2aa (all biotinylated, BD Pharmingen), followed by avidin-alkaline phosphatase (AP; Southern Biotechnology Associates). Plates were developed for 14–18 h with ImmunoPure PNPP (Pierce) as the substrate. OD values were recorded and background values were subtracted out (background was defined as the OD value generated by a hybridoma supernatant of irrelevant specificity, typically 0.07). All developing Abs were allotype-marked (Iga) except for IgG1, (due to poor sensitivity of the anti-IgG1a Ab in ELISAs). Although the IgG1 reagent was not allotype-marked, sera from uninjected CB17 mice showed no significant staining above background for IgG1 anti-chromatin Abs. Points derived from the linear range of the ELISAs were used for generating graphs. For studies examining whether long-lived plasma cells are generated when T cell help is provided, mice were serially bled and their sera tested by ELISA. When anti-chromatin OD values returned to baseline levels for a mouse, they were no longer bled or tested.
Flow cytometry
A total of 0.5–1 x 106 splenocytes were surface stained (46) using the following Abs: anti-B220-FITC (RA3-6B2), anti-CD4-PE (GK1.5), anti-IgMa-PE or -bio (DS-1), anti-Ig1-bio (R11-153), IgDa-bio (217–170), IgG1a-bio (10.9), IgG2aa-bio (8.3) (BD Pharmingen), and 6.5-biotin (grown as supernatant and biotinylated). Biotinylated peanut agglutinin (PNA; Vector Laboratories) was also used to mark germinal center (GC) B cells.
Determination of cell recovery
The frequency of IgMa+Ig1+ B cells or CD4+6.5+ T cells in the spleen was determined by flow cytometry and multiplied by the total number of live splenocytes to determine the absolute number of cells. The percent recovery of transferred B or T cells was determined by dividing the absolute number of cells recovered by the number of cells injected.
Immunohistochemistry
Spleens were frozen, sectioned, and stained (5) using the following Abs: anti-CD4-bio (GK1.5), anti-CD22-FITC or -bio (Cy34.1), anti-IgMa-FITC (DS-1), anti-IgG1a-bio (10.9), anti-IgG2aa-bio (8.3) (BD Pharmingen), and/or PNA-bio (Vector Laboratories). Secondary reagents were anti-FITC-AP, anti-FITC-HRP, streptavidin-AP, or streptavidin-HRP (Southern Biotechnology Associates). Developed slides were read by multiple (at least four) investigators without prior knowledge of the experimental condition.
Statistical analyses
Statistical significance was determined via the unpaired, two sample Student’s t test provided by Microsoft Excel software unless otherwise noted. Significance was ascribed when p < 0.05.
Results
The VH3H9 H chain paired with the V1 L chain generates an Ab that binds dsDNA and chromatin (47, 48). Therefore, the VH3H9 Tg can be used to directly track autoreactive B cells in vivo by staining for endogenous V1. To test the impact of cognate T cell help on anti-chromatin B cells in vivo, HA-bearing anti-chromatin B cells were injected into third-party recipient CB17 mice that had received distinct subsets of anti-HA CD4+ T cells (Fig. 1; Ref. 16).
Th1-deviated cells derived from wild-type or gld/gld mice produced little or no IL-4 or IL-10, and varied widely according to the amount of IFN- they made (Fig. 2). One group of Th1 cells had 40–70% of the cells producing IFN- without detectable IL-4 or IL-10, and such cells are termed IFN-high Th1 cells. In contrast, the second group, which also did not make IL-4 or IL-10, yielded fewer IFN--producing cells (2–35%), and less IFN- per cell, as visualized by lower intracellular IFN- fluorescent intensity (Fig. 2). This second group is termed IFN-low Th1 cells. This variation in IFN- production proved to be significant in terms of the ability of the T cells to help anti-chromatin B cells (see below) and appeared dependent upon the source of rIL-12 in cultures. To directly examine the effects of limiting IL-12 on deviated Th cells, cultures were set up in the absence of exogenous rIL-12, but in the presence of anti-IL-4 Ab to prevent acquisition of a Th2-like phenotype (cells resulting from such cultures are termed IFN-low cells). A low frequency of CD4+ T cells (2–8%) from these cultures produced IFN-, with low levels of IFN- per cell, and <1% produced IL-4 or IL-10. Both the IFN-high and IFN-low intracellular cytokine levels are within the range of those reported by others for Th1-deviated cultures (41, 42, 49, 50).
Anti-chromatin Ab production is induced by Th2 cells and IFN-high Th1 cells in a CD40-dependent manner
Th2 cells promoted anti-chromatin autoantibodies from the transferred B cells (Fig. 3). Autoantibody production from mice given Th1 cells was variable and correlated (p < 0.01) with the level of IFN- produced by the Th1 cells before their injection. IFN-high Th1 cells induced levels of anti-chromatin autoantibodies similar to those seen in Th2 recipients, whereas levels in recipients of IFN-low and IFN-low Th1 cells were not above the baseline (Fig. 3). The observation that IFN- levels affect the ability of Th1 cells to help B cells may resolve a controversy regarding the ability of Th1 cells to serve as helpers (25, 26, 27, 28, 29, 30, 31), as laboratories using distinct in vitro deviation protocols may generate differing frequencies of IFN--producing Th1 cells.
CD40-CD154 interactions play a vital role in Th cell activity, including autoimmune settings (51, 52, 53, 54). Likewise, the blocking anti-CD154 mAb (MR1) abrogated anti-chromatin Ab production stimulated by either Th2 or IFN-high Th1 cells (Fig. 3).
FasL influences the quality of T cell help for anti-chromatin B cells
Injection of Th2-deviated gld/gld cells enhanced the production of anti-chromatin Abs in recipient mice compared with the titers observed with Th2 cells derived from wild-type mice (Fig. 3). There was no difference in anti-chromatin Ab titers between IFN-high Th1 cells and IFN-high Th1-gld/gld cells (Fig. 3, p = 0.34), whereas IFN-low Th1-gld/gld T cells promoted higher titers of anti-chromatin Abs than their IFN-low FasL-sufficient counterparts (Fig. 3).
Anti-chromatin B cells undergo isotype switching in response to Th2- or Th1-type help
The site-directed VH3H9 Tg was used to monitor anti-chromatin isotype switching. CB17 mice that received Th2 cells or IFN-high Th1 cells and site-directed Tg anti-chromatin B cells had detectable titers of Ig1 anti-chromatin Abs (Fig. 4A). These Abs included not only IgMa Abs (Fig. 4B) but also isotype-switched Abs (Fig. 4, C and D). IFN-high Th1 cell help resulted in higher titers of IgG2aa autoantibodies than mice given either Th2 cells or B cells alone. Th2 recipients produced high levels of IgG1 anti-chromatin Abs and low, but also significant (relative to recipients of B cells alone), titers of IgG2aa anti-chromatin Abs. Like what was observed using the non-site-directed VH3H9 Tg donors, IFN-low Th1 cells did not induce anti-chromatin Abs.
Th2-gld/gld cells induced higher titers of Ig1 anti-chromatin Abs from site-directed VH3H9 Tg B cells than their FasL-sufficient counterparts (Fig. 4). This increase included higher titers of the IgMa and IgG2aa isotypes, but similar levels of the IgG1 isotype. The increased level of IFN- production by gld/gld-derived Th2 cells (Fig. 2) may account for the higher IgG2a isotype anti-chromatin Abs (25). In contrast, the absence of functional FasL on IFN-high Th1 cells did not have an effect on the levels of Ig1 anti-chromatin Abs produced or their isotype distribution (Fig. 4, p 0.05 for Ig1, IgMa, or IgG2aa ELISAs).
Anti-chromatin B cell recovery correlates with autoantibody production
In the absence of exogenous Th cells, very few anti-chromatin B cells remained 7 days after transfer (Fig. 5), consistent with their rapid in vivo turnover rate (8). Both Th2 and IFN-high Th1 cells supported anti-chromatin B cell recovery and this was dependent upon CD40-CD154 interactions (Fig. 5). FasL-deficient Th2 and IFN-high Th1 cells promoted similar recoveries compared with their wild-type counterparts. In contrast, IFN-low Th1 cells derived from gld/gld mice promoted anti-chromatin B cell survival well above that from FasL-sufficient IFN-low Th1 cells (Fig. 5).
When FasL-deficient Th2 cells were administered, tight clusters of anti-chromatin AFCs were observed, similar to their FasL-sufficient counterparts (Fig. 6A). Anti-chromatin B cells from Th2-gld/gld recipients, however, were also present in GCs with a higher frequency compared with mice given FasL-sufficient Th2 cells (Fig. 6A, arrows, and Fig. 7). Recipients of IFN-high Th1-gld/gld T cells had anti-chromatin GCs as did their wild-type counterparts, but their EFF were even more diffuse. Most strikingly, IFN-low Th1-gld/gld cells induced not only anti-chromatin GCs but also AFCs (data summarized in Table I).
Discussion
The abilities of Th1 and Th2 cells, with and without FasL, to elicit autoantibody production, promote autoreactive B cell survival, and trigger participation in the GC reaction in vivo are summarized in Table I. These studies were done using anti-chromatin B cells from both VH3H9 Tg and site-directed VH3H9 Tg mice because of differences documented in signaling between IgM/IgD and IgG receptors (56, 57) as well as disparities seen between randomly integrated transgenes and those that have been targeted to the JH locus (58). The only distinction we detected between the site-directed and the non-site-directed transgenes was that the former generated isotype-switched Abs in response to T cell help.
Although polarized CD4+ T cell populations are typically described by the production of Th1 vs Th2 cytokines, it has been documented that within these populations a wide range of cytokines can be produced (29, 59). In this study, this variability was most pronounced for the Th1 polarized cells and proved to be significant in terms of B cell fate. Although Th2 cells consistently induced high titers of anti-chromatin Abs in a CD40-dependent manner, the ability of Th1-deviated cells to help anti-chromatin B cells correlated with T cell IFN- production. IFN-high Th1 cells induced anti-chromatin AFCs and GCs, but IFN-low Th1 cells promoted only GCs, and a much lower degree of anti-chromatin B cell recovery. Unlike the tightly clustered anti-chromatin AFCs observed in Th2 recipients, the AFCs found under IFN-high Th1 conditions were more diffuse and localized at the border of the T cell area, as has been described for other autoimmune AFCs in lpr/lpr or gld/gld mice (4, 60, 61, 62). This phenotype is even more pronounced in recipients of Th1-gld/gld cells, which is consistent with the possibility that the T cells driving the autoimmune response in Fas/FasL-mutant mice are of the Th1 type (61, 62, 63, 64, 65). The significance of the distinct localization sites for AFCs has not been determined, but in one model, these extrafollicular cells were somatically mutated (60).
FasL plays an important role in influencing the outcome of T cell help for anti-chromatin B cells. Relative to their FasL+ counterparts, Th2-gld/gld cells induced higher titers of IgM and IgG2a isotype anti-chromatin Abs, and more frequent anti-chromatin B cell GCs. Strikingly, the inability of IFN-low Th1 cells to induce anti-chromatin Ab production appears dependent on FasL, as only FasL-deficient IFN-low Th1 cells supported significant anti-chromatin B cell survival and high titers of anti-chromatin Abs.
FasL may alter anti-chromatin B cell fate indirectly by limiting T cell help (32, 33, 35) or by direct B cell killing (20). T cell loss due to FasL expression, however, was not observed in our study. IL-4, which has been shown to protect B cells from Fas-mediated death (66), may be responsible for the similar recovery of anti-chromatin B cells in recipients of Th2 FasL-sufficient and -deficient cells (see Fig. 5). The finding that IFN-high, but not IFN-low, Th1 cell help results in anti-chromatin Ab production and B cell survival, prompts us to consider that, like IL-4 (66), IFN- may impart resistance to Fas-mediated death. Consistent with this hypothesis, one report showed a synergistic effect on B cells if IFN- was combined with stimulatory CpG oligonucleotides (67), but the direct effect of IFN- on B cells remains controversial (68, 69, 70, 71, 72, 73). An alternate hypothesis is that the IFN-low Th1 cells have diverged to a distinct differentiation pathway, such that they provide unique helper functions compared with the IFN-high Th1 cells. Because we have previously documented that undeviated Th cells can promote autoantibody production (16), the failure of the IFN-low Th1 cells to induce autoantibodies is not likely to be a consequence of them being less differentiated.
It has been postulated that Fas-mediated killing is involved in B cell negative selection within GCs (74, 75, 76, 77). Our experiments provide no evidence that B cell death via FasL borne on Th cells curtails anti-chromatin GCs: Th1 cells (both IFN-high and IFN-low) induce anti-chromatin GCs with or without FasL, and although Th2-gld/gld but not FasL-sufficient Th2 cells induce anti-chromatin GCs, a difference in B cell survival is not observed. Rather, we found a correlation between IFN- production by Th cells and B cell GC differentiation. In all cases where some degree of IFN- is produced (including Th2-gld/gld conditions), anti-chromatin B cells are found in GCs. Notably, older Fas/FasL-deficient mice have a predominance of IFN-+ T cells (78), and IFN- is critical for autoantibody production and autoimmune disease in MRL-lpr/lpr mice (79, 80). Studies to examine the direct effect of IFN- on anti-chromatin B cell proliferation, activation, and sensitivity to Fas-mediated death are under way.
Although we have documented that anti-chromatin B cells can form GCs, there is no evidence that long-lived AFCs are generated by the addition of T cell help (Table II). Other reports have described anti-DNA B cells in GCs in the absence of detectable serum Ab production (81, 82), suggesting a fail-safe mechanism that guards against memory formation and long-term autoantibody production. The observation documented here that only short-lived AFCs are generated in healthy mice by provision of T cell help suggests a means of curtailing the liability of autoreactive B cells. This conclusion is tempered, however, as another study, again using a transfer model but this time with nonautoreactive B cells, also failed to induce long-lived AFCs (26). This raises the possibility that their generation may have particular requirements that are not met under the transfer conditions used. Studies are underway to determine whether autoantibodies that arise naturally in autoimmune settings are derived from long-lived or short-lived AFCs. In the New Zealand Black/White and New Zealand M2410 models, the presence of long-lived AFCs has been documented, and interestingly, they have been shown to reside at unique sites (83, 84, 85). Clearly, more studies are warranted to better understand what governs B cell fate decisions in healthy vs autoimmune settings, the outcome of which may have important implications for B cell depletion strategies currently used to treat human autoimmune diseases (86, 87).
Acknowledgments
We thank A. Acosta and J. Faust of The Wistar Institute Flow Cytometry Facility; S. Alexander and A. Pagán for excellent technical assistance; and R. Noelle, M. Weigert, G. Trinchieri, and M. Monestier for reagents.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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 Funding has been provided by the National Institutes of Health (AI32137, AR47913, and 2T32AI007518) and the Commonwealth Universal Research Enhancement Program, PA Department of Health.
2 Address correspondence and reprint requests to Dr. Jan Erikson, The Wistar Institute, Room 276, 3601 Spruce Street, Philadelphia, PA 19104. E-mail address: jan@wistar.upenn.edu
3 Abbreviations used in this paper: Tg, transgenic; HA, hemaglutinin; FasL, Fas ligand; AFC, Ab-forming cell; AP, alkaline phosphatase; PNA, peanut agglutinin; GC, germinal center; EFF, extrafollicular foci.
Received for publication January 19, 2005. Accepted for publication April 12, 2005.
References
Tan, E. M.. 1989. Antinuclear antibodies: diagnostic markers for autoimmune diseases and probes for cell biology. Adv. Immunol. 44: 93-151
Erikson, J., M. Z. Radic, S. A. Camper, R. R. Hardy, C. Carmack, M. Weigert. 1991. Expression of anti-DNA immunoglobulin transgenes in non-autoimmune mice. Nature 349: 331-334.
Roark, J. H., A. Bui, K.-A. Nguyen, L. Mandik, J. Erikson. 1997. Persistence of functionally compromised anti-dsDNA B cells in the periphery of non-autoimmune mice. Int. Immunol. 9: 1615-1626.
Mandik-Nayak, L., S.-j. Seo, C. Sokol, K. M. Potts, A. Bui, J. Erikson. 1999. MRL-lpr/lpr mice exhibit a defect in maintaining developmental arrest and follicular exclusion of anti-double-stranded DNA B cells. J. Exp. Med. 189: 1799-1814.
Mandik-Nayak, L., A. Bui, H. Noorchashm, A. Eaton, 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.[
Fields, M. L., C. L. Sokol, A. Eaton-Bassiri, S.-j. Seo, M. P. Madaio, J. Erikson. 2001. Fas/Fas ligand deficiency results in altered localization of anti-double-stranded DNA B cells and dendritic cells. J. Immunol. 167: 2370-2378.
Fields, M. L., J. Erikson. 2003. The regulation of lupus-associated autoantibodies: immunoglobulin transgenic models. Curr. Opin. Immunol. 15: 709-717.(Michele L. Fields, Simone)
Circulating autoantibodies against dsDNA and chromatin are a characteristic of systemic lupus erythematosus in humans and many mouse models of this disease. B cells expressing these autoantibodies are normally regulated in nonautoimmune-prone mice but are induced to secrete Abs following T cell help. Likewise, anti-chromatin autoantibody production is T cell-dependent in Fas/Fas ligand (FasL)-deficient (lpr/lpr or gld/gld) mice. In this study, we demonstrate that Th2 cells promote anti-chromatin B cell survival and autoantibody production in vivo. FasL influences the ability of Th2 cells to help B cells, as Th2-gld/gld cells support higher titers of anti-chromatin Abs than their FasL-sufficient counterparts and promote anti-chromatin B cell participation in germinal centers. Th1 cells induce anti-chromatin B cell germinal centers regardless of FasL status; however, their ability to stimulate anti-chromatin Ab production positively correlates with their level of IFN- production. This distinction is lost if FasL-deficient T cells are used: Th1-gld/gld cells promote significant titers of anti-chromatin Abs regardless of IFN- production levels. Thus, FasL from effector T cells plays an important role in determining the fate of anti-chromatin B cells.
Introduction
Antibodies that bind dsDNA and chromatin are a serologic hallmark of systemic lupus erythematosus and are found in several murine models of systemic lupus erythematosus (1). Using the VH3H9 Ig transgenic (Tg) 3 model, a population of anti-chromatin B cells (VH3H9/V1) has been characterized in both healthy and autoimmune mice (2, 3, 4, 5, 6, 7). In healthy mice, these anti-chromatin B cells persist in the periphery but their Abs are undetected (2, 5). They have a decreased life span, are predominantly developmentally arrested and activated, and localize to the edges of the B cell follicles near the T cell areas in the spleen (5, 8). We hypothesize that this phenotype is a consequence of chronic exposure to Ag, in the absence of T cell help (9, 10, 11, 12, 13).
To dissect the responses of anti-chromatin B cells to CD4+ Th cells, we established an in vivo model of cognate interaction between these two cell types. Mice engineered to express the neoself Ag hemagglutinin (HA) on MHC class II-bearing cells (including B cells) were mated to VH3H9 Ig Tg mice (14, 15). Anti-HA CD4+ T cells from TCR Tg mice and HA-expressing anti-chromatin B cells were then transferred together into a third-party recipient mouse and their fates tracked (16, 17). Using this strategy, we have demonstrated that anti-chromatin B cells from healthy mice respond to CD4+ Th cells in vivo by producing autoantibodies (16). These findings contrast with those obtained from the hen-egg lysozyme model of B cell tolerance (18, 19), where autoreactive anti-hen-egg lysozyme B cells were shown to be resistant to CD4+ T cell help (20, 21, 22, 23). Importantly, Fas/ Fas ligand (FasL) interactions mediated this resistance (20, 21, 24).
Both Th1 and Th2 cells can help nonautoreactive B cells to produce Abs (25, 26, 27), although Th2 cells appear more efficient in this regard (28, 29, 30, 31). Th1 cells reportedly express higher levels of FasL than Th2 cells (32, 33, 34, 35), which could modulate T-B interactions by either limiting the availability of Th cells (32, 33, 35) and/or by direct killing of the autoreactive B cells (20, 21). Furthermore, Th1 cells may be more susceptible to Fas-mediated death even in cases where they express comparable levels of FasL (32, 33, 35).
Once activated, B cells can follow a number of differentiation pathways–toward short-lived Ab-forming cells (AFCs), memory B cells, or long-lived AFCs. Although much has been learned about the transcription factors that govern the fate decisions of B cells (36), less is known about the precise cues that determine these decisions. Given the critical role that CD4+ T cells play in autoantibody production, and the seemingly contradictory data on autoreactive B cell responses to T cell help (20, 21, 22, 23), we have investigated the impact of Th1 and Th2 cells, with and without FasL, on anti-chromatin B cells. We report here the differential capabilities of T effector cells to influence anti-chromatin B cells in vivo, and suggest that FasL may play a role in controlling the differentiation pathways and magnitude of the autoreactive B cell response.
Materials and Methods
Mice
TS1 BALB/c (FasL-sufficient or -deficient) and VH3H9 Tg/HACII/Ig–/– mice were bred in specific-pathogen-free conditions at The Wistar Institute under the approval and supervision of the Institutional Animal Care and Use Committee (IACUC), and genotyped as described (2, 6, 16, 17, 37). Similarly, site-directed VH3H9 Tg BALB/c mice in which the VH3H9 Tg was targeted to the JH locus (Refs. 38, 39, 40 ; generously provided by Dr. M. Weigert (University of Chicago, Chicago, IL) were bred to produce site-directed VH3H9 Tg/HACII/Ig–/– mice. CB17 mice were purchased from Charles River Laboratories. Only young (6–12 wk old) TS1 BALB/c-gld/gld mice were used. All other mice were used at 6–16 wk, and both genders were used.
Th1 and Th2 cell cultures
TS1 BALB/c or TS1 BALB/c-gld/gld lymph nodes were depleted (>90%) of CD8+ cells using anti-CD8 Dynalbeads (Dynaltech). A total of 0.5 x 106 CD8-depleted lymphocytes were then cultured for Th1/Th2 deviation as previously described (41). rIL-12 was a generous gift from Dr. G. Trinchieri (National Institutes of Health, Bethesda, MD) or was purchased from R&D Biosystems or PeproTech. Cells received fresh media containing IL-2 at days 3 and 5, and were rested in the absence of IL-2 at day 7. At day 9, cells were harvested and an aliquot cultured with PMA and ionomycin (Sigma-Aldrich) in the presence of brefeldin A (Cytofix/Cytoperm kit; BD Pharmingen) for another 4–6 h. Cells were stained for CD4 expression, fixed, permeabilized, and stained for intracellular cytokines, using anti-IFN--FITC, anti-IL-4-PE, and/or anti-IL-10-PE (BD Pharmingen) (42). The cells that were not restimulated with PMA/ionomycin were purified by centrifugation with Lympholyte M (Cedarlane Laboratories) before injection.
T cell/influenza virus injections
A total of 5–10 x 106 Th1- or Th2-deviated CD4+ cells from in vitro cultures were suspended in sterile PBS with 1000 hemagglutinating units (43) of purified PR8 influenza virus (15) and injected i.v.
Anti-chromatin B cell injections
Splenocytes from VH3H9 Tg/HACII/Ig–/– or site-directed VH3H9 Tg/HACII/Ig–/– mice were depleted of RBC and an aliquot was stained by flow cytometry to determine the frequency of anti-chromatin B cells (B220+ Ig1+). CB17 recipient mice (allotype Igb) were injected with spleen preparations containing 4–10 x 106 anti-chromatin B cells (allotype Iga). Control mice received B cells without previous injection of Th cells or virus.
Blockade of CD40-CD154 interactions
Mice were i.p. injected with 250 μg of purified anti-CD154 mAb (MR1) (a kind gift of Dr. R. Noelle, Dartmouth Medical School, Lebanon, NH) on the same day as B cell transfer, and 3 days afterward.
Chromatin ELISAs
Chromatin (a generous gift of Dr. M. Monestier, Temple University, Philadelphia, PA) was diluted to 2 μg/ml and plated overnight as described (44). ELISAs were done as described (45) with the following modifications: the block used was 1% BSA/PBS/azide, developing Abs were anti-Ig1, anti-IgMa, anti-IgG1, or anti-IgG2aa (all biotinylated, BD Pharmingen), followed by avidin-alkaline phosphatase (AP; Southern Biotechnology Associates). Plates were developed for 14–18 h with ImmunoPure PNPP (Pierce) as the substrate. OD values were recorded and background values were subtracted out (background was defined as the OD value generated by a hybridoma supernatant of irrelevant specificity, typically 0.07). All developing Abs were allotype-marked (Iga) except for IgG1, (due to poor sensitivity of the anti-IgG1a Ab in ELISAs). Although the IgG1 reagent was not allotype-marked, sera from uninjected CB17 mice showed no significant staining above background for IgG1 anti-chromatin Abs. Points derived from the linear range of the ELISAs were used for generating graphs. For studies examining whether long-lived plasma cells are generated when T cell help is provided, mice were serially bled and their sera tested by ELISA. When anti-chromatin OD values returned to baseline levels for a mouse, they were no longer bled or tested.
Flow cytometry
A total of 0.5–1 x 106 splenocytes were surface stained (46) using the following Abs: anti-B220-FITC (RA3-6B2), anti-CD4-PE (GK1.5), anti-IgMa-PE or -bio (DS-1), anti-Ig1-bio (R11-153), IgDa-bio (217–170), IgG1a-bio (10.9), IgG2aa-bio (8.3) (BD Pharmingen), and 6.5-biotin (grown as supernatant and biotinylated). Biotinylated peanut agglutinin (PNA; Vector Laboratories) was also used to mark germinal center (GC) B cells.
Determination of cell recovery
The frequency of IgMa+Ig1+ B cells or CD4+6.5+ T cells in the spleen was determined by flow cytometry and multiplied by the total number of live splenocytes to determine the absolute number of cells. The percent recovery of transferred B or T cells was determined by dividing the absolute number of cells recovered by the number of cells injected.
Immunohistochemistry
Spleens were frozen, sectioned, and stained (5) using the following Abs: anti-CD4-bio (GK1.5), anti-CD22-FITC or -bio (Cy34.1), anti-IgMa-FITC (DS-1), anti-IgG1a-bio (10.9), anti-IgG2aa-bio (8.3) (BD Pharmingen), and/or PNA-bio (Vector Laboratories). Secondary reagents were anti-FITC-AP, anti-FITC-HRP, streptavidin-AP, or streptavidin-HRP (Southern Biotechnology Associates). Developed slides were read by multiple (at least four) investigators without prior knowledge of the experimental condition.
Statistical analyses
Statistical significance was determined via the unpaired, two sample Student’s t test provided by Microsoft Excel software unless otherwise noted. Significance was ascribed when p < 0.05.
Results
The VH3H9 H chain paired with the V1 L chain generates an Ab that binds dsDNA and chromatin (47, 48). Therefore, the VH3H9 Tg can be used to directly track autoreactive B cells in vivo by staining for endogenous V1. To test the impact of cognate T cell help on anti-chromatin B cells in vivo, HA-bearing anti-chromatin B cells were injected into third-party recipient CB17 mice that had received distinct subsets of anti-HA CD4+ T cells (Fig. 1; Ref. 16).
Th1-deviated cells derived from wild-type or gld/gld mice produced little or no IL-4 or IL-10, and varied widely according to the amount of IFN- they made (Fig. 2). One group of Th1 cells had 40–70% of the cells producing IFN- without detectable IL-4 or IL-10, and such cells are termed IFN-high Th1 cells. In contrast, the second group, which also did not make IL-4 or IL-10, yielded fewer IFN--producing cells (2–35%), and less IFN- per cell, as visualized by lower intracellular IFN- fluorescent intensity (Fig. 2). This second group is termed IFN-low Th1 cells. This variation in IFN- production proved to be significant in terms of the ability of the T cells to help anti-chromatin B cells (see below) and appeared dependent upon the source of rIL-12 in cultures. To directly examine the effects of limiting IL-12 on deviated Th cells, cultures were set up in the absence of exogenous rIL-12, but in the presence of anti-IL-4 Ab to prevent acquisition of a Th2-like phenotype (cells resulting from such cultures are termed IFN-low cells). A low frequency of CD4+ T cells (2–8%) from these cultures produced IFN-, with low levels of IFN- per cell, and <1% produced IL-4 or IL-10. Both the IFN-high and IFN-low intracellular cytokine levels are within the range of those reported by others for Th1-deviated cultures (41, 42, 49, 50).
Anti-chromatin Ab production is induced by Th2 cells and IFN-high Th1 cells in a CD40-dependent manner
Th2 cells promoted anti-chromatin autoantibodies from the transferred B cells (Fig. 3). Autoantibody production from mice given Th1 cells was variable and correlated (p < 0.01) with the level of IFN- produced by the Th1 cells before their injection. IFN-high Th1 cells induced levels of anti-chromatin autoantibodies similar to those seen in Th2 recipients, whereas levels in recipients of IFN-low and IFN-low Th1 cells were not above the baseline (Fig. 3). The observation that IFN- levels affect the ability of Th1 cells to help B cells may resolve a controversy regarding the ability of Th1 cells to serve as helpers (25, 26, 27, 28, 29, 30, 31), as laboratories using distinct in vitro deviation protocols may generate differing frequencies of IFN--producing Th1 cells.
CD40-CD154 interactions play a vital role in Th cell activity, including autoimmune settings (51, 52, 53, 54). Likewise, the blocking anti-CD154 mAb (MR1) abrogated anti-chromatin Ab production stimulated by either Th2 or IFN-high Th1 cells (Fig. 3).
FasL influences the quality of T cell help for anti-chromatin B cells
Injection of Th2-deviated gld/gld cells enhanced the production of anti-chromatin Abs in recipient mice compared with the titers observed with Th2 cells derived from wild-type mice (Fig. 3). There was no difference in anti-chromatin Ab titers between IFN-high Th1 cells and IFN-high Th1-gld/gld cells (Fig. 3, p = 0.34), whereas IFN-low Th1-gld/gld T cells promoted higher titers of anti-chromatin Abs than their IFN-low FasL-sufficient counterparts (Fig. 3).
Anti-chromatin B cells undergo isotype switching in response to Th2- or Th1-type help
The site-directed VH3H9 Tg was used to monitor anti-chromatin isotype switching. CB17 mice that received Th2 cells or IFN-high Th1 cells and site-directed Tg anti-chromatin B cells had detectable titers of Ig1 anti-chromatin Abs (Fig. 4A). These Abs included not only IgMa Abs (Fig. 4B) but also isotype-switched Abs (Fig. 4, C and D). IFN-high Th1 cell help resulted in higher titers of IgG2aa autoantibodies than mice given either Th2 cells or B cells alone. Th2 recipients produced high levels of IgG1 anti-chromatin Abs and low, but also significant (relative to recipients of B cells alone), titers of IgG2aa anti-chromatin Abs. Like what was observed using the non-site-directed VH3H9 Tg donors, IFN-low Th1 cells did not induce anti-chromatin Abs.
Th2-gld/gld cells induced higher titers of Ig1 anti-chromatin Abs from site-directed VH3H9 Tg B cells than their FasL-sufficient counterparts (Fig. 4). This increase included higher titers of the IgMa and IgG2aa isotypes, but similar levels of the IgG1 isotype. The increased level of IFN- production by gld/gld-derived Th2 cells (Fig. 2) may account for the higher IgG2a isotype anti-chromatin Abs (25). In contrast, the absence of functional FasL on IFN-high Th1 cells did not have an effect on the levels of Ig1 anti-chromatin Abs produced or their isotype distribution (Fig. 4, p 0.05 for Ig1, IgMa, or IgG2aa ELISAs).
Anti-chromatin B cell recovery correlates with autoantibody production
In the absence of exogenous Th cells, very few anti-chromatin B cells remained 7 days after transfer (Fig. 5), consistent with their rapid in vivo turnover rate (8). Both Th2 and IFN-high Th1 cells supported anti-chromatin B cell recovery and this was dependent upon CD40-CD154 interactions (Fig. 5). FasL-deficient Th2 and IFN-high Th1 cells promoted similar recoveries compared with their wild-type counterparts. In contrast, IFN-low Th1 cells derived from gld/gld mice promoted anti-chromatin B cell survival well above that from FasL-sufficient IFN-low Th1 cells (Fig. 5).
When FasL-deficient Th2 cells were administered, tight clusters of anti-chromatin AFCs were observed, similar to their FasL-sufficient counterparts (Fig. 6A). Anti-chromatin B cells from Th2-gld/gld recipients, however, were also present in GCs with a higher frequency compared with mice given FasL-sufficient Th2 cells (Fig. 6A, arrows, and Fig. 7). Recipients of IFN-high Th1-gld/gld T cells had anti-chromatin GCs as did their wild-type counterparts, but their EFF were even more diffuse. Most strikingly, IFN-low Th1-gld/gld cells induced not only anti-chromatin GCs but also AFCs (data summarized in Table I).
Discussion
The abilities of Th1 and Th2 cells, with and without FasL, to elicit autoantibody production, promote autoreactive B cell survival, and trigger participation in the GC reaction in vivo are summarized in Table I. These studies were done using anti-chromatin B cells from both VH3H9 Tg and site-directed VH3H9 Tg mice because of differences documented in signaling between IgM/IgD and IgG receptors (56, 57) as well as disparities seen between randomly integrated transgenes and those that have been targeted to the JH locus (58). The only distinction we detected between the site-directed and the non-site-directed transgenes was that the former generated isotype-switched Abs in response to T cell help.
Although polarized CD4+ T cell populations are typically described by the production of Th1 vs Th2 cytokines, it has been documented that within these populations a wide range of cytokines can be produced (29, 59). In this study, this variability was most pronounced for the Th1 polarized cells and proved to be significant in terms of B cell fate. Although Th2 cells consistently induced high titers of anti-chromatin Abs in a CD40-dependent manner, the ability of Th1-deviated cells to help anti-chromatin B cells correlated with T cell IFN- production. IFN-high Th1 cells induced anti-chromatin AFCs and GCs, but IFN-low Th1 cells promoted only GCs, and a much lower degree of anti-chromatin B cell recovery. Unlike the tightly clustered anti-chromatin AFCs observed in Th2 recipients, the AFCs found under IFN-high Th1 conditions were more diffuse and localized at the border of the T cell area, as has been described for other autoimmune AFCs in lpr/lpr or gld/gld mice (4, 60, 61, 62). This phenotype is even more pronounced in recipients of Th1-gld/gld cells, which is consistent with the possibility that the T cells driving the autoimmune response in Fas/FasL-mutant mice are of the Th1 type (61, 62, 63, 64, 65). The significance of the distinct localization sites for AFCs has not been determined, but in one model, these extrafollicular cells were somatically mutated (60).
FasL plays an important role in influencing the outcome of T cell help for anti-chromatin B cells. Relative to their FasL+ counterparts, Th2-gld/gld cells induced higher titers of IgM and IgG2a isotype anti-chromatin Abs, and more frequent anti-chromatin B cell GCs. Strikingly, the inability of IFN-low Th1 cells to induce anti-chromatin Ab production appears dependent on FasL, as only FasL-deficient IFN-low Th1 cells supported significant anti-chromatin B cell survival and high titers of anti-chromatin Abs.
FasL may alter anti-chromatin B cell fate indirectly by limiting T cell help (32, 33, 35) or by direct B cell killing (20). T cell loss due to FasL expression, however, was not observed in our study. IL-4, which has been shown to protect B cells from Fas-mediated death (66), may be responsible for the similar recovery of anti-chromatin B cells in recipients of Th2 FasL-sufficient and -deficient cells (see Fig. 5). The finding that IFN-high, but not IFN-low, Th1 cell help results in anti-chromatin Ab production and B cell survival, prompts us to consider that, like IL-4 (66), IFN- may impart resistance to Fas-mediated death. Consistent with this hypothesis, one report showed a synergistic effect on B cells if IFN- was combined with stimulatory CpG oligonucleotides (67), but the direct effect of IFN- on B cells remains controversial (68, 69, 70, 71, 72, 73). An alternate hypothesis is that the IFN-low Th1 cells have diverged to a distinct differentiation pathway, such that they provide unique helper functions compared with the IFN-high Th1 cells. Because we have previously documented that undeviated Th cells can promote autoantibody production (16), the failure of the IFN-low Th1 cells to induce autoantibodies is not likely to be a consequence of them being less differentiated.
It has been postulated that Fas-mediated killing is involved in B cell negative selection within GCs (74, 75, 76, 77). Our experiments provide no evidence that B cell death via FasL borne on Th cells curtails anti-chromatin GCs: Th1 cells (both IFN-high and IFN-low) induce anti-chromatin GCs with or without FasL, and although Th2-gld/gld but not FasL-sufficient Th2 cells induce anti-chromatin GCs, a difference in B cell survival is not observed. Rather, we found a correlation between IFN- production by Th cells and B cell GC differentiation. In all cases where some degree of IFN- is produced (including Th2-gld/gld conditions), anti-chromatin B cells are found in GCs. Notably, older Fas/FasL-deficient mice have a predominance of IFN-+ T cells (78), and IFN- is critical for autoantibody production and autoimmune disease in MRL-lpr/lpr mice (79, 80). Studies to examine the direct effect of IFN- on anti-chromatin B cell proliferation, activation, and sensitivity to Fas-mediated death are under way.
Although we have documented that anti-chromatin B cells can form GCs, there is no evidence that long-lived AFCs are generated by the addition of T cell help (Table II). Other reports have described anti-DNA B cells in GCs in the absence of detectable serum Ab production (81, 82), suggesting a fail-safe mechanism that guards against memory formation and long-term autoantibody production. The observation documented here that only short-lived AFCs are generated in healthy mice by provision of T cell help suggests a means of curtailing the liability of autoreactive B cells. This conclusion is tempered, however, as another study, again using a transfer model but this time with nonautoreactive B cells, also failed to induce long-lived AFCs (26). This raises the possibility that their generation may have particular requirements that are not met under the transfer conditions used. Studies are underway to determine whether autoantibodies that arise naturally in autoimmune settings are derived from long-lived or short-lived AFCs. In the New Zealand Black/White and New Zealand M2410 models, the presence of long-lived AFCs has been documented, and interestingly, they have been shown to reside at unique sites (83, 84, 85). Clearly, more studies are warranted to better understand what governs B cell fate decisions in healthy vs autoimmune settings, the outcome of which may have important implications for B cell depletion strategies currently used to treat human autoimmune diseases (86, 87).
Acknowledgments
We thank A. Acosta and J. Faust of The Wistar Institute Flow Cytometry Facility; S. Alexander and A. Pagán for excellent technical assistance; and R. Noelle, M. Weigert, G. Trinchieri, and M. Monestier for reagents.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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 Funding has been provided by the National Institutes of Health (AI32137, AR47913, and 2T32AI007518) and the Commonwealth Universal Research Enhancement Program, PA Department of Health.
2 Address correspondence and reprint requests to Dr. Jan Erikson, The Wistar Institute, Room 276, 3601 Spruce Street, Philadelphia, PA 19104. E-mail address: jan@wistar.upenn.edu
3 Abbreviations used in this paper: Tg, transgenic; HA, hemaglutinin; FasL, Fas ligand; AFC, Ab-forming cell; AP, alkaline phosphatase; PNA, peanut agglutinin; GC, germinal center; EFF, extrafollicular foci.
Received for publication January 19, 2005. Accepted for publication April 12, 2005.
References
Tan, E. M.. 1989. Antinuclear antibodies: diagnostic markers for autoimmune diseases and probes for cell biology. Adv. Immunol. 44: 93-151
Erikson, J., M. Z. Radic, S. A. Camper, R. R. Hardy, C. Carmack, M. Weigert. 1991. Expression of anti-DNA immunoglobulin transgenes in non-autoimmune mice. Nature 349: 331-334.
Roark, J. H., A. Bui, K.-A. Nguyen, L. Mandik, J. Erikson. 1997. Persistence of functionally compromised anti-dsDNA B cells in the periphery of non-autoimmune mice. Int. Immunol. 9: 1615-1626.
Mandik-Nayak, L., S.-j. Seo, C. Sokol, K. M. Potts, A. Bui, J. Erikson. 1999. MRL-lpr/lpr mice exhibit a defect in maintaining developmental arrest and follicular exclusion of anti-double-stranded DNA B cells. J. Exp. Med. 189: 1799-1814.
Mandik-Nayak, L., A. Bui, H. Noorchashm, A. Eaton, 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.[
Fields, M. L., C. L. Sokol, A. Eaton-Bassiri, S.-j. Seo, M. P. Madaio, J. Erikson. 2001. Fas/Fas ligand deficiency results in altered localization of anti-double-stranded DNA B cells and dendritic cells. J. Immunol. 167: 2370-2378.
Fields, M. L., J. Erikson. 2003. The regulation of lupus-associated autoantibodies: immunoglobulin transgenic models. Curr. Opin. Immunol. 15: 709-717.(Michele L. Fields, Simone)