Genetic and cellular dissection of the activation of AM14 rheumatoid factor B cells in a mouse model of lupus

Multiple genetic and cellular factors contribute to the activation of autoreactive B cells in the NZM2410 mouse model of lupus.

Abstract

The RF-specific AM14 tg BCR has been used as a model to dissect the mechanisms of B cell tolerance to ICs containing nucleic acids. We have shown previously that AM14 RF B cells break tolerance in the TC mouse model of lupus through the dual engagement of the AM14 BCR and TLR9. In this study, we showed that neither the expression of Sle1 or Sle2 susceptibility loci alone was sufficient to activate AM14 RF B cells, suggesting that the production of antichromatin IgG2a^a autoAg mediated by Sle1 and an intrinsically higher B cell activation mediated by Sle2 were required. We also showed that the B6 genetic background enhanced the selection of AM14 RF B cells to the MZB cell compartment regardless of the expression of the Sle loci and therefore, of their activation into AFCs. Furthermore, some AM14 RF B cells were selected into the B-1a compartment, where they did not differentiate into AFCs. Therefore, it is unlikely that the selection of AM14 RF B cells to the MZB or B-1a cell compartments in TC.AM14^a mice is responsible for their breach of tolerance. Finally, we showed that the presence of expression of Sle1 in non-tg cells, most likely T cells, is necessary for the activation of AM14 RF B cells into AFCs. Overall, these results suggest a threshold model of activation of AM14 RF B cells on the B6 background with additive genetic and cellular contribution of multiple sources.

Introduction

A large body of work has demonstrated that establishment and maintenance of B cell tolerance against self-antigens are multistep processes in humans and mice [1]. Impairments at any of the tolerance checkpoints result in the production of autoAb, which may lead to pathogenic systemic autoimmunity. Central tolerance is largely a B cell intrinsic process depending on the binding specificity of the BCR, the strength of the signal that it induces, and its ability to undergo receptor editing. Peripheral tolerance is a more complex and diverse process, in which different B cell subsets display different thresholds of functional autoreactivity. In addition, extrinsic factors, such the presence of CD4⁺ T cell help or the level of BAFF availability, fully participate in the maintenance of peripheral B cell tolerance.

BCR tg mouse models have shown that multiple breaches in B cell tolerance can result in the production of autoAb directed against nuclear antigens [2]. One of the best models of B cell tolerance against nuclear antigens has been offered by the AM14 RF tg BCR [3]. In this model, the AM14 HC combined to endogenous Vκ8 results in the expression of a BCR specific for IgG2a^a autoAg. AM14 HC Vk8 Id⁺ B cells become activated and differentiate into AFCs upon the BCR binding to ICs composed of IgG2a^a bound to chromatin that become internalized to trigger TLR9 signaling [4]. AM14 RF B cells developing in a nonautoimmune background are clonally ignorant [5] but become activated in the lupus-prone MRL/lpr mice expressing the IgG2a^a self-antigen to differentiate into extrafollicular PBs that secrete Id⁺ RF [6, 7]. The main contribution of the MRL/lpr autoimmune genetic background in this model is the production of antichromatin IgG2a^a that is necessary and sufficient to activate AM14 RF B cells in a TLR9-dependent manner [8]. Accordingly, AM14 RF B cells are activated in BALB/c or MRL/+ nonautoimmune mice by immunization with antichromatin IgG2a^a [8], supporting the hypothesis that in these strains, the breach of tolerance of AM14 RF B cells is controlled by factors extrinsic to the tg B cells. In the presence of antichromatin IgG2a^a, BALB/c AM14 RF B cells do not require T cell help for activation, although CD40L and IL-21 signals significantly enhanced the magnitude of the AM14 RF response [9]. However, B cell intrinsic factors can influence the AM14 RF B cell activation. Deficiency in actin related gene 1, a gene that regulates CD40 signaling, results in spontaneous BALB/c AM14 RF B cell activation through a GC rather than extrafollicular route [10].

We have recently characterized the fate of AM14 RF B cells in another mouse model of lupus, the TC strain, which expresses 3 NZM2410 lupus susceptibility loci on a B6 background [11]. We showed that in the TC but not B6 mice expressing the IgG2a^a autoAg, AM14 RF B cells differentiate into AFCs through the production of short-lived extrafollicular PBs [12]. This indicated that MRL/lpr and TC lupus-prone backgrounds induce the spontaneous differentiation of AM14 B cells into AFCs through the same extrafollicular route. However, contrary to MRL/lpr or BALB/c mice, immunization of TC.AM14 mice with antichromatin IgG2a^a activated AM14 RF B cells but was not sufficient to induce the production of Id⁺ RF. Moreover, the immunization of B6.AM14 mice with antichromatin IgG2a^a had no effect on AM14 RF B cells. This indicated that the mechanisms of activation of AM14 RF B cells are different between the B6/TC and BALB/c /MRL genetic backgrounds.

This study was conducted to dissect the genetic and cellular factors contributing to AM14 RF B cells in TC.AM14^a mice. We compared the individual contribution of the Sle1 and Sle2 loci with the process. Sle1 is a locus that is functionally expressed in B and T cells [13] and that is strongly associated with the production of antichromatin IgG [14]. If the production of antichromatin IgG is sufficient to activate AM14 RF B cells, then the phenotype of AM14 RF B cells should be similar between Sle1.AM14^a and TC.AM14^a mice. Sle2, a locus functionally expressed in B cells [15], is associated with polyclonal B cell activation [16], expansion of the B-1a cell compartment [15], and selection to the MZB compartment [17, 18]. If high intrinsic B cell activation contributes to the breach of tolerance of AM14 RF B cells in TC.AM14^a mice, then AM14 RF B cells should also be activated in Sle2.AM14^a mice. We also examined the recruitment of AM14 RF B cells to the MZB and B-1a cell compartments, both of which are enriched for autoreactive clones and expanded in TC mice [19, 20]. We used a mixed BM chimera approach to determine whether TC or Sle1 non-tg cells contributed to the activation of AM14 RF B cells. We showed that neither the expression of Sle1 nor Sle2 alone was sufficient to activate AM14 RF B cells, suggesting that the production of antichromatin IgG2a^a and an intrinsically higher B cell activation were required. We also showed that the B6 background enhanced the selection of AM14 RF B cells to the MZB compartment regardless of the expression of the Sle loci and therefore, of their activation into AFCs. Furthermore, some AM14 RF B cells were selected into the B-1a compartment, where they did not differentiate into AFCs. Therefore, it is unlikely that the selection of AM14 RF B cells to the MZB or B-1a cell compartments in TC.AM14^a mice is responsible for their breach of tolerance. Finally, we showed that the presence of expression of Sle1 in non-tg cells, most likely T cells, is necessary for the activation of AM14 RF B cells into AFCs. Overall, these results suggest a threshold model of activation of AM14 RF B cells on the B6 background with additive genetic and cellular contribution of multiple sources.

MATERIALS AND METHODS

Mice

The TC congenic strain has been described previously [11]. B6 and TC mice expressing the AM14 HC tg, with or without the IgH^a allotype (B6.AM14^a, B6.AM14, TC.AM14^a, and TC.AM14, respectively), have been described already [12]. To produce the Sle1.AM14, Sle1.AM14^a, Sle2.AM14, and Sle2.AM14^a strains, the B6.Sle1 or B6.Sle2 strain [21] was crossed to B6.AM14 or B6.AM14^a. B6.p18^−/−.AM14^a mice were produced by crossing B6.p18^−/− [22] and B6.AM14^a mice. The IgH^a allotype was detected by IgG2a^a and IgG2a^b ELISA in serum from 2-month-old mice by use of B6 and B6^a mice as negative and positive controls, respectively. The AM14 HC tg was detected by BCR, as described previously [3]. Microsatellite markers were used as described previously [21] to genotype for the Sle1 or Sle2 locus. Only female mice were used in this study. All mice were bred and maintained at the University of Florida in specific pathogen-free conditions. Animal protocols were approved by the Institutional Animal Care and Use Committee of the University of Florida.

Flow cytometry

RBC-depleted, single-cell suspensions of splenocytes were blocked with 10% rabbit serum and anti-CD16/32 (2.4G2) antibody and then stained with predetermined amounts of the following fluorophore-conjugated or biotinylated antibodies: B220 (RA3-6B2), Bcl6 (K112-91), CD4 (RM4-5), CD19 (1D3), CD44 (IM7), CD62L (MEL-14), CD69 ([¹H].2F3), CD90.1 (OX-7), CD90.2 (53-2.1), CXCR5 (2G8), Fas (Jo2), Foxp3 (FJK-16S), Ly-77 (GL7), IgM (II/41), IgM^a (DS-1), IgM^b (AF6-78), and PD-1 (RMP1-30). For the detection of AM14 Id⁺ B cells, membrane staining with 4-44-biotin was followed by intracellular staining with 4-44-Alexa 647, by use of the Cytofix/Cytoperm kit (BD Biosciences, San Jose, CA, USA), as described previously [3]. For detection of biotinylated antibodies, streptavidin conjugated to PE, PerCP-Cy5.5, or allophycocyanin-Cy7 was used. Stained cells were analyzed on a CyAn (Beckman Coulter, Brea, CA, USA) or Fortessa (BD Biosciences) flow cytometer, and at least 100,000 cells were acquired/sample. Dead cells were excluded on the basis of forward- and side-scatter characteristics.

ELISPOT

Id-specific AFCs were quantified by ELISPOT, as described previously [12]. In brief, multiscreen filter plates (EMD Millipore, Billerica, MA, USA) were coated overnight with 5 μg/ml goat anti-mouse IgM µ chain antibody (EMD Millipore). Serially diluted splenocytes or BM cells were cultured overnight in duplicate in RPMI 1640 (Cellgro, Mediatech, Manassas, VA, USA), supplemented with 5% FBS at 37°C. Bound antibodies were detected with 4-44-biotin and streptavidin conjugated to HRP (Vector Laboratories, Burlingame, CA, USA). Plates were developed with 3-amino-9-ethylcarbazole (Sigma-Aldrich, St. Louis, MO, USA), and AFCs were counted by use of a Bioreader 4000 Pro-X (Bio-Sys, Miami, FL, USA).

ELISA

Id-specific (4-44⁺) IgM was detected as described previously [3]. In brief, Immulon 4 plates (Thermo Fisher Scientific, Waltham, MA, USA) were coated with 5 μg/ml anti-mouse IgM µ chain antibody (EMD Millipore). 4-44-Biotin was used as a secondary antibody, followed by streptavidin conjugated to AP (SouthernBiotech, Birmingham, AL, USA). Serum samples were diluted 1:100. For in vitro stimulation, 2 × 10⁶ splenocytes from 7-month-old mice were cultured for 5 days with 1 μg/ml CpG-B (InvivoGen, San Diego, CA, USA), and culture supernatant was assayed for Id⁺ RF secretion at 1:10 dilution. Total IgM was measured in culture supernatants diluted 1:100 in Immulon 2 plates coated with anti-mouse IgM antibody (EMD Millipore) and with anti-IgM-AP as a secondary antibody (SouthernBiotech). IgG2a^a-specific IgM RF was quantified in Immulon 2 plates coated with 25 μg/ml TNP-BSA (Biosearch Technologies, Petaluma, CA, USA), followed by anti-TNP IgG2a^a (BD Biosciences). Sera were diluted 1:100, and the secondary antibody was anti-IgM-AP (SouthernBiotech). To quantify antichromatin antibodies, Immulon 2 plates were coated with 50 μg/ml dsDNA (Sigma-Aldrich), followed by 10 μg/ml total histone (Roche Diagnostics, Indianapolis, IN, USA). Serum samples were diluted 1:100, and specific isotypes were detected with the following secondary antibodies: anti-IgG2b-AP (SouthernBiotech), rabbit anti-IgG2a^a or IgG2a^b (Nordic Immunologic Labs, Susteren, The Netherlands), followed by goat anti-rabbit IgG-AP (Sigma-Aldrich). For both antichromatin and the IgG2a^a-specific IgM RF ELISA, pooled TC.IgH^a sera were used to standardize results into relative units.

Hybridoma immunizations

Immunization of IgH^b allotype mice with the antichromatin IgG2a^a hybridoma clone PL2-3 has been described previously [12]. Two-month-old mice were injected intraperitoneally with 250 μl pristane (Sigma-Aldrich) on days 0 and 7 and then with 10⁷ hybridoma cells on day 10 and euthanized on day 17 when the presence of B cells expressing intracellular 4-44 Id was evaluated by flow cytometry; serum hybridoma antibody secretion was measured by ELISA.

BM chimeras

Mixed BM chimeras were conducted as described previously [23]. Lethally irradiated B6 hosts received a combination of 2 million T cell-depleted BM cells at a ratio of 1 TC.AM14^a cell to 8 non-tg cells. B6^a cells express CD90.1 (Thy1^a), but this allotypic marker was lost during the breeding of TC^a mice, which express only IgH^a. Chimeric mice were euthanized 3 months after transplant to analyze B and T cell phenotypes by flow cytometry and the presence of Id⁺ AFCs by ELISPOT.

Statistical analysis

Data analysis was performed with Prism 6.0 software (GraphPad Software, La Jolla, CA, USA). Unless indicated, graphs show means with sem or medians with interquartile ranges, and statistical significance between strains was determined by 2-tailed by t tests or Mann-Whitney tests, depending on whether the data were distributed normally. Multiple test corrections were applied when several strains were compared. Significance levels are indicated in the figures.

RESULTS

The Sle1 and Sle2 loci affect the peripheral expansion of AM14 HC tg B cells

We have shown previously that the number of lymphocytes was reduced in B6 and TC AM14 HC tg mice, but the TC background maintained the relative expansion of the number of splenocytes compared with B6 [12]. Lymphoid expansion in Sle1 and Sle2 AM14 HC tg mice was decreased significantly compared with TC.AM14 mice, with or without IgH^a expression, as measured by spleen weight or splenocytes numbers (Fig. 1A and B). The number of splenocytes in Sle1.AM14^a/Sle1.AM14 and Sle2.AM14^a/Sle2.AM14 mice was similar to that of B6.AM14^a/B6.AM14 mice (Fig. 1B). Thus, as with the non-tg mice [11], the single Sle1 and Sle2 loci have limited effect on lymphoid expansion. The percentage and number of total B cells were also lower in Sle1.AM14^a/Sle1.AM14 mice compared with TC.AM14^a/TC.AM14 mice (Fig. 1C and D). However, the percentage and number of total B cells in Sle2.AM14^a mice were similar to TC.AM14^a and significantly higher than in B6.AM14^a mice, which is likely a result of the association of the Sle2 locus with B cell hyperactivity [16]. The percentage and number of CD4⁺ T cells were similar among the Sle1, Sle2, and B6 AM14^a/AM14 mice (data not shown). The fate of IgM^a AM14 HC tg B cells can be compared with that of endogenous non-tg B cells only in IgH^b allotype mice. We have shown previously that the TC genetic background expanded the percentage and number of AM14 HC tg B cells compared with the endogenous B cells [12]. Sle1, but not Sle2, resulted in a similar expansion of the percentage of AM14 HC tg B cells (Fig. 1E). The absolute number of AM14 HC tg B cells in Sle1.AM14 mice was, however, intermediate between TC.AM14 and B6.AM14 (Fig. 1F), as a result of their overall lower number of B cells (Fig. 1C and D). These results show that Sle1 is responsible for the expansion of the AM14 tg B cells in the TC model and suggest that its negative effect on endogenous B cells is compensated by Sle2 expression in TC mice. Finally, we compared the surface expression of the Id⁺ BCR between the IgH^a and IgH^b allotypes and showed that it was decreased significantly by the presence of the autoAg to a similar extent in the 4 strains (Fig. 1G). To the extent that the IgM expression level can be used as a marker of anergy [24], this result suggests that exposure to IgG2a^a anergizes Id⁺ RF B cells on a B6 genetic background. Furthermore, it shows that the expression of the Sle loci does not impair this process.

Figure 1. Characterization of B cells in Sle1 and Sle2 AM14 mice compared with TC.AM14 and B6.AM14 mice.

For each strain, values were compared between mice expressing the IgH^a or IgH^b allotype (such as TC.AM14^a and TC.AM14). Spleen weight (A), total number of splenocytes (B), percentage (C), and total number (D) of B220⁺ splenocytes are shown. An age-matched non-tg B6 value is given as reference. (A and B) Statistical significance is indicated for multiple-comparison tests with TC.AM14^a or TC.AM14 for IgH^a or IgH^b strains, respectively. (C and D) Statistical significance corresponds to comparisons indicated with brackets. Percentage (E) and absolute number (F) of B cells expressing the AM14 HC tg, measured as IgM^a+ in IgH^b AM14 tg mice. (G) IgM^a expression measured as mean fluorescence intensity (MFI) on CD19⁺ IgMa⁺ Id⁺-gated splenocytes. (G, right) Representative overlays between the IgH^a or IgH^b allotypes for each of the 4 strains. Each symbol corresponds to values obtained from individual 7- to 9-month-old mice. Results show means and sem values. Significance levels correspond to Holm-Sidak’s multiple comparisons tests (A–F) and 1-tailed t tests between the IgH^a or IgH^b allotypes for each strain (G). *P < 0.05; **P < 0.01; ***P < 0.001.

The Sle1 and Sle2 loci induce a modest production of RF in the presence of IgG2a^a

The combination of the 3 Sle loci and autoAg results in the production of serum Id⁺ RF and RF AFCs in TC.AM14^a mice [12]. The amount of Id⁺ or anti-IgG2a^a IgM RF found in the serum of Sle1 and Sle2 AM14 mice with IgH^a or IgH^b allotype was significantly lower than in TC.AM14^a mice (data not shown). However, the level of serum RF is poorly correlated with the number of Id⁺ RF AFCs (R² = 0.38), most likely as a result of circulating ICs that decrease the sensitivity of RF ELISA measurements. Either Sle1.AM14^a or Sle2.AM14^a spleens showed Id⁺ AFC numbers intermediate between that of TC.AM14^a and B6.AM14^a (Fig. 2A). As we have shown previously for TC.AM14^a mice [12], the number of Id⁺ AFCs was significantly higher in the spleen than in the BM of Sle1.AM14^a or Sle2.AM14^a mice (data not shown). This excludes the possibility that the lower number of Id⁺ AFCs found in the spleen of these mice is a result of their migration to the BM. However, the numbers of Id⁺ AFCs were significantly higher in Sle1.AM14^a or Sle2.AM14^a splenocytes than in Sle1.AM14 or Sle2.AM14, respectively, indicating that the strains expressing Sle1 or Sle2 in the presence of the autoAg produce more Id⁺ AFCs than in the absence of the autoAg.

Figure 2. Individual expression of Sle1 or Sle2 supports a limited production of Id⁺ RF.

Splenic Id⁺ AFCs (A), serum antichromatin IgG2a^a (B), and serum antichromatin IgG (C) in Sle1 and Sle2 AM14 mice compared with TC.AM14 and B6.AM14 mice expressing the IgH^a or IgH^b allotype. Results show medians and interquartile ranges. Statistical significance is indicated for comparisons with the TC.AM14^a values. In addition, significant pair comparisons are indicated with brackets (A). (D) Correlation between serum antichromatin IgG2a^a and the numbers of splenic Id⁺ AFCs. The lines correspond to the linear regression for each of the 3 strains (TC.AM14^a, Sle1.AM14^a, and Sle2.AM14^a). The correlation was statistically significant (Pearson’s correlation) only for Sle1.AM14^a mice. Each symbol corresponds to values obtained from individual 7- to 9-month-old mice. *P < 0.05; **P < 0.01; ***P < 0.001.

Sle1 is characterized by the production of large amounts of antichromatin IgG [14]. Therefore, we expected high levels of antichromatin IgG2a^a that would correspond to high numbers of Id⁺ AFCs in Sle1.AM14^a mice. Surprisingly, very little antichromatin IgG2a^a was found in Sle1.AM14^a sera (Fig. 2B). Sle1.AM14^a mice also produced significantly lower levels of total antichromatin IgG compared with TC.AM14^a (Fig. 2C) and B6.Sle1 non-tg mice (data not shown). As expected, the amounts of antichromatin IgG2a^a and antichromatin IgG were highly correlated in all strains (P < 0.001). Sle1.AM14^a sera also contained significantly lower levels of anti-dsDNA IgG (P < 0.001), as well as total IgG2a^a (P < 0.01) compared with TC.AM14^a. This suggests that Sle1 expands AM14 HC tg B cells relative to the endogenous B cells in Sle1.AM14^a mice, as we have shown in Sle1.AM14 mice (Fig. 1E). As the tg B cells produce only IgM, it limits the production of all IgGs, including antichromatin IgG2a^a. We observed a highly significant correlation (R² = 0.88) between the production of antichromatin IgG2a^a and the number of Id⁺ AFCs in Sle1.AM14^a mice (Fig. 2D), further suggesting that autoAg production is the limiting factor in the production of Id⁺ RF in these mice.

Sle2 expression is associated with the production of polyreactive IgM and to a lesser level, IgG antibodies, which include low-affinity autoAb [16]. As for Sle1, the expression of Sle2 resulted in a higher number of Id⁺ AFCs in IgH^a than IgH^b allotype mice (Fig. 2B). Antichromatin IgG2a^a and IgG were found in the sera of Sle2.AM14^a mice at levels intermediate between TC.AM14^a and B6.AM14^a (Fig. 2B and C). However, there was no correlation (R² = 0.07) between antichromatin IgG2a^a and the number of Id⁺ AFCs in the spleen of Sle2.AM14^a mice (Fig. 2D). This suggests that AM14 Id⁺ B cells expressing Sle2 are activated by a different mechanism than the Id⁺ B cells expressing Sle1. As more Id⁺ AFCs were found in Sle2.AM14^a than in Sle2.AM14 mice (Fig. 2A), this suggests that the expression of IgG2a^a plays a role, although not in the context of antichromatin ICs.

The Sle1 and Sle2 loci promote the spontaneous activation of AM14 B cells in the presence of IgG2a^a

The production of Id⁺ RF is preceded by the activation of Id⁺ B cells from a stage in which only mId⁺ is expressed to iId expression in PBs (mId⁺ iId⁺ PBs) and then PCs (mId⁻ iId⁺ PCs) [12]. The percentage of mId⁺ iId⁻ B cells among splenocytes was similar between strains, except for a small but significant decrease in Sle1.AM14 compared with Sle1.AM14^a mice, which corresponds to a similar trend between TC.AM14^a and TC.AM14 Id⁺ B cells (Fig. 3A). Thus, the individual lupus susceptibility loci do not have a major effect on the expansion of quiescent Id⁺ B cells.

Figure 3. Individual expression of Sle1 or Sle2 supports a limited a expression of iId⁺ IgM in the presence of autoAg.

Splenocytes from TC, Sle1, Sle2, and B6 AM14 mice, expressing the IgH^a or IgH^b allotype, were identified as B cells expressing only membrane mId (A), PBs expressing mId and iId (B), and PCs expressing only iId (C). (D) Total iId⁺ (PB + PC) B cells. Each symbol corresponds to values obtained from individual 7- to 9-month-old mice. Results show medians and interquartile ranges. Significant pair comparisons are indicated with brackets. In addition, statistical significance was reached between TC.AM14^a and B6.AM14^a values with Dunn's multiple comparison tests (B and D; shown by a single asterisk). *P < 0.05; **P < 0.01; ***P < 0.001.

mId⁺ iId⁺ PBs are the major producers of Id⁺ RF in the MRL/lpr and TC strains [12, 25]. TC.AM14^a spleens contain a higher percentage of PBs than B6.AM14^a spleens [12], but either Sle1.AM14^a or Sle2.AM14^a spleens contained intermediate percentages of mId⁺ iId⁺ PBs (Fig. 3B). However, as for TC.AM14^a, the percentage of mId⁺ iId⁺ PBs was significantly higher in Sle1.AM14^a and Sle2.AM14^a than in Sle1.AM14 and Sle2.AM14, respectively (Fig. 3B). Similar results were obtained when mId⁺ iId⁺ PBs were compared as percentages of B cells (data not shown). Meanwhile, all 3 Sle loci were necessary for B cells to terminally differentiate into mId⁻ iId⁺ PCs, as only TC.AM14^a mice had significantly higher levels of Id⁺ PCs compared with all of the other strains (Fig. 3C). Finally, the total percentage of splenocytes that expressed iId was significantly higher in TC.AM14^a and Sle2.AM14^a spleens than in their IgH^b counterparts, whereas the same trend for Sle1.AM14^a did not reach statistical significance (Fig. 3D). These flow cytometry results are in agreement with the results obtained by ELISPOTS. Therefore, the Sle1 and Sle2 loci are each sufficient in the presence of IgG2a^a for Id⁺ B cells to activate and differentiate into PBs, although not in a majority of mice, as it is the case for the combination of the 3 Sle loci in TC.AM14^a mice.

Dual BCR and TLR ligation does not activate Sle1.AM14 or Sle2.AM14 B cell

We showed previously that as in the MRL background, RF B cells in the TC genetic background require dual BCR and TLR9 ligation for activation [12]. To determine if the Sle1 or Sle2 locus was sufficient to induce RF B cell activation in the same conditions, we immunized young Sle1.AM14 and Sle2.AM14 mice with the IgG2a^a antichromatin-secreting PL2-3 hybridoma by use of immunization of TC.AM14 and B6.AM14 mice as positive and negative controls, respectively. The sera of most recipient mice showed high levels of antichromatin IgG2a^a, which indicated that the hybridomas secreted their antibody (Fig. 4A). However, unlike in TC.AM14 mice, this PL2-3-produced antichromatin IgG2a^a was not sufficient to induce Id⁺ B cell activation and differentiation into PBs in Sle1.AM14 and Sle2.AM14 mice (Fig. 4B). Therefore, the expression of Sle1 or Sle2 alone is not sufficient to allow AM14 Id⁺ B cells to be activated by dual BCR and TLR ligation and differentiate into PBs. Although antichromatin IgG2a^a might be the limiting factor for the activation of AM14 Id⁺ B cells in Sle1.AM14^a mice (Fig. 2), the results from the hybridoma immunization also suggest that a chronic stimulation by the autoAg is necessary in these mice.

Figure 4. Dual BCR and TLR ligation is not sufficient to activate Sle1 or Sle2 AM14 B cells.

Two-month-old mice pretreated with pristane were immunized with IgG2a^a antichromatin secreting hybridomas (PL2-3) and analyzed 1 week later. Serum antichromatin IgG2a^a secreted by the hybridomas (A) and percentage of splenic mId⁺ iId⁺ PBs (B) in the recipient mice. Each symbol corresponds to values obtained from individual recipient mice. Results show medians and interquartile ranges. Significance levels correspond to Dunn's multiple comparison tests with TC.AM14 values. *P < 0.05; ***P < 0.001.

Id⁺ B cells are selected in the MZB compartment

We have shown that mId⁺ B cells are selected to the MZB compartment more frequently in TC.AM14^a than in TC.AM14 or B6.AM14^a mice, which suggested that the MZ route could be a part of Id⁺ B cell activation in TC.AM14^a mice [12]. We have explored in more details this issue of peripheral selection and expansion of AM14 HC tg B cells and Id⁺ B cells, first by comparing the MZ distribution of IgM^a AM14 HC tg B cells with endogenous IgM^b B cells in IgH^b mice. With the use of an IgM⁺ CD1d^hi-gating scheme for MZB cells (Fig. 5A), we found that within the same mouse, IgM^a AM14 HC tg B cells are found at a higher frequency in the MZB compartment than endogenous IgM^b B cells in all 4 AM14 strains (Fig. 5B). The same results were found for the absolute numbers of IgM^a relative to IgM^b MZBs in TC.AM14 and B6.AM14 mice but not in Sle1.AM14 and Sle2.AM14 mice (Fig. 5C), which was linked to their total and IgM^a B cell numbers, respectively (Fig. 1). In addition, the absolute number of IgM^a MZBs was significantly higher in TC.AM14 mice than in all of the 3 other strains (Fig. 5C). Conversely, the number of IgM^a FOBs was lower than the number of IgM^b FOBs in all 4 strains (Fig. 5D). Similar results were obtained when B cells were gated, first as MZB or FOB cells, and the expression level of IgM^a was significantly higher on MZB than on FOB cells in all strains (data not shown). These results suggest that the AM14 HC tg B cells are preferentially recruited or expanded in the MZB compartment on the B6 genetic background and that this process is enhanced by the combination of TC susceptibility loci, which not only expands the total number of B cells [11], but specifically increases the number of MZB [19].

Figure 5. AM14 HC tg B cells and mId⁺ B cells are preferentially selected to the MZB compartment.

(A) Gating scheme for CD1d^hi IgM⁺ MZB and CD1d^lo IgM⁺ FOB subsets in the IgM^a AM14 tg and IgM^b endogenous B cell populations (left), as well as in the IgM^a 4-44^- Id⁻ and IgM^a 4-44⁺ Id⁺ populations (right). Percentages (B) and absolute numbers (C) of IgM^a AM14 tg and IgM^b endogenous B cells in the MZB subset in TC.AM14, Sle1.AM14, Sle2.AM14, and B6.AM14 mice. (D) Absolute numbers of IgM^a AM14 tg and IgM^b endogenous B cells in the FOB subset in the same mice as in B and C. Percentages of mId⁺ (+) and Id⁻ (−) IgM⁺ cells in the MZB gate in the IgH^a (E) and in the IgH^b AM14 tg strains (F). Percentages (G) and absolute numbers (H) of iId⁺ FOB (FO) and MZB (MZ) in TC.AM14^a and B6.AM14^a mice. Linked symbols represent values within the same mouse, and each pair represents a different mouse. Comparisons between IgM^a and IgM^b cells (B–D), Id⁺ and Id⁻ cells (E and F), and FOB and MZB (G and H) within a strain were performed with Wilcoxon ranked tests. The number of IgM^a MZB cells (C; P < 0.001) and the number of IgM^a or IgM^b FOB cells (D; P < 0.05) were significantly higher in TC.AM14 mice than in any of the 3 other strains (Dunn's multiple comparisons tests). (H) Comparisons between strains are also indicated (Mann-Whitney tests) *P < 0.05; **P < 0.01; ***P < 0.001.

We next compared the distribution of mId⁺ RF B cells (the subset of AM14 HC tg B cells expressing Vκ8) and Id⁻ B cells (which contain AM14 HC tg and endogenous B cells) with the MZB compartment. In all 4 IgH^a strains, the Id⁺ B cells were found more frequently in the MZB compartment than Id⁻ B cells (Fig. 5E). This was also observed to a similar extent in the IgH^b tg strains (Fig. 5F). This suggested that B cells with the Id specificity were preferentially selected to the MZ subset on the B6 background regardless of the expression of the Sle susceptibility loci. Similar results were obtained when B cells were gated, first as MZB or FOB cells, and the expression level of the Id⁺ BCR was significantly higher on MZB than on FOB cells in all strains (data not shown). mId⁺ B cells differentiated toward the expression of iId with similar frequency in the MZB and FOB compartments (Fig. 5G). However, the number of iId⁺ FOB cells was significantly higher than the number of iId⁺ MZB cells (Fig. 5H), as a result of the overall difference in cell numbers between these compartments. Furthermore, the number of iId⁺ FOB and iId⁺ MZB was significantly higher in TC.AM14^a than in B6.AM14^a mice (Fig. 5H), indicating that the TC genetic background promotes PB differentiation similarly in MZB and FOB. Overall, these results show that in TC.AM14^a mice, mId⁺ B cells are preferentially recruited to the MZB compartment, where they can differentiate into Id⁺ RF AFCs. The bulk of Id⁺ PBs and hence, Id⁺ RF is, however, produced by cells of follicular origin.

Id⁺ B cells are selected to the B-1a cell compartment, where they do not get activated

The number of B-1a cells is greatly expanded in the PerC of TC mice, largely as a result of the low expression of cyclin-dependent kinase 4 inhibitor (p18) in the Sle2 locus [22]. Accordingly, the TC, Sle2, and p18^−/− genetic backgrounds result in similar phenotypes in B-1a cells [26], which are autoreactive and have been proposed to contribute, at least indirectly, to autoimmune pathology [26]. Therefore, we examined whether the Id⁺ B cells were selected to the PerC B-1a compartment in the TC.AM14^a, Sle2.AM14^a, and p18^−/−.AM14^a strains compared with B6.AM14^a and whether Id⁺ B-1a cells differentiated into AFCs. The total PerC cell counts were significantly higher in TC.AM14^a and p18^−/−.AM14^a mice compared with B6.AM14^a mice, with a similar trend for Sle2.AM14^a mice (Fig. 6A). The number of PerC B-1a cells (Fig. 6B) as well as the B-1a/B-2 ratio (Fig. 6C) were, however, similar among the 4 tg strains. The proportion of B-1a relative to B-2 cells was significantly lower in TC.AM14^a, Sle2.AM14^a, and p18^−/−.AM14^a PerCs than in their non-tg counterparts (Fig. 6C). This indicates that the expression of the AM14 HC tg largely blunts the expansion of B-1a cells in TC mice. Similar numbers of mId⁺ B-1a cells were found in the PerC of the 4 tg strains (Fig. 6D), and similar results were obtained with the percentage of mId⁺ B-1a cells relative to either as total PerC cells or PerC B cells (data not shown). Similar PerC mId⁺ B-1a/B-2 cell ratios were also observed among the 4 tg strains (Fig. 6E). However, the PerC mId⁺ B-1a/B-2 cell ratio was significantly higher than the total PerC B-1a/B-2 ratio (Fig. 6F), suggesting that on the B6 genetic background and regardless of the Sle susceptibility locus expression, mId⁺ B cells are selected to the B-1a cell compartment relatively more frequently than total B cells. We did not detect the expression of iId in PerC B-1a cells in any of the 4 tg strains (data not shown). This indicated that the relatively enhanced selection of mId⁺ to the B-1a cell compartment is not productive and that these cells do not contribute to the production of Id⁺ RF autoAb.

Figure 6. mId⁺ B cells are selected to the B1-a cell compartment.

Absolute number of total PerC (A) and B-1a (B) cells in TC.AM14^a, Sle2.AM14^a, and p18^−/−.AM14^a and B6.AM14^a mice. (C) PerC B-1a/B-2 cell ratio in AM14^a strains (empty symbols) compared with the corresponding non-tg strains (filled symbols). Absolute number of PerC mId⁺ B-1a cells (D) and PerC B-1a/B-2 mId⁺ cell ratio (E) in TC.AM14^a, Sle2.AM14^a, and p18^−/−.AM14^a and B6.AM14^a mice. (F) Paired PerC B-1a/B-2 cell ratios for total B cells (B; empty symbols) and mId⁺ B cells (mId⁺; filled symbols). Linked symbols represent values within the same mouse, and each pair represents a different mouse. A–E) Results show medians and interquartile ranges, and statistical significance is indicated for Dunn's multiple comparison tests with B6.AM14^a values. (C) Comparisons between the tg and non-tg strains were performed with Mann-Whitney tests. (F) Wilcoxon ranked tests were performed between total B cell and mId⁺ B cell values for each strain. *P < 0.05; **P < 0.01; ***P < 0.001.

Nonintrinsic factors contribute to AM14 Id⁺ B cell activation in TC mice

To address whether the activation of Id⁺ B cells in TC.AM14^a mice was a result of intrinsic factors, we used several strategies. Adoptive transfers of purified TC.AM14^a or B6.AM14^a B cells in B6.Rag1^−/− or sublethally irradiated B6 mice, as used successfully in BALB/c mice [5], did not yield sufficient numbers of Id⁺ B cells for informative results. We tested next whether TC^a-derived non-tg cells could induce B6.AM14^a Id⁺ B cells to differentiate into AFCs in mixed BM chimera. In these gain-of-function conditions, however, TC^a-derived cells always completely outgrew B6.AM14^a B cells, even at a 1:9 ratio. Finally, we devised a loss-of-function, mixed BM chimera to test whether the presence of B6^a non-tg cells would decrease the activation of TC.AM14^a Id⁺ B cells compared with chimeras in which all tg and non-tg cells are from the TC^a origin (Fig. 7A). Three months after transfer, the number of splenocytes was significantly reduced in the chimeras containing B6^a cells (Fig. 7B), and a similar trend was observed for the number of iId⁺ B cells (Fig. 7C), although there was no difference in the number of mId⁺ B cells (data not shown). The proportion of Id⁺ AFCs was also significantly reduced in the chimeras containing B6^a cells (Fig. 7D). It should be noted that the proportion of Id⁺ AFCs in the B6^a cell containing chimeras was similar to that of untouched B6.AM14^a mice, whereas chimeras containing TC^a cells were similar to untouched TC.AM14^a mice (compare Fig. 7D with Fig. 2A). The B6^a cell containing chimeras also showed a decreased CD4⁺ T cell activation, as measured by CD69 (Fig. 7E) and CD44 expression (Fig. 7F). This was associated with a significant reduction in the percentage of total GC B cells (Fig. 7G). Interestingly, this was not the case for Id⁺ GCs, which were found at a similar low frequency in the 2 types of chimeras. This shows that as in intact TC.AM14^a mice [12], the GC route was not followed preferentially by activated AM14 B cells in the BM chimeras. Overall, these results showed that the activation of TC.AM14^a B cells requires the presence of other cells expressing the same genetic background. Our experimental design did not allow testing of which cell type was contributing to the Id⁺ B cell activation, but our data showed that it correlated with a strong CD4⁺ T cell activation.

Figure 7. Extrinsic factors contribute to the activation of TC.AM14^a Id⁺ B cells.

(A) Experimental strategy for the BMT 1 BM chimeras with TC.AM14^a BM cells mixed with TC^a or B6^a BM cells. Total splenocytes numbers (B), total number of iId⁺ splenocytes (C), splenic Id⁺ AFCs (D), and percentage of CD69⁺ CD4⁺ T cells (E) in recipient mice that have received TC.AM14^a and TC^a or B6^a BM cells. CD44 mean fluorescence intensity on CD4⁺-gated splenocytes (F), percentage of Fas⁺ GL7⁺ GC B220⁺-gated B cells (G), and percentage of Id⁺ IgM^a cells within these GC B cells (H) in recipient mice that have received TC.AM14^a and TC^a or B6^a BM cells. (E–H) Representative FACS plots are shown below with the corresponding gates. The x-axis of each graph shows the strain of origin on the non-tg BM cells combined with TC.AM14^a cells. Each symbol represents an individual chimeric mouse. Graphs show means ± sem (B and E–G) or median and interquartile range (C, D, and H) and the corresponding significance values of t- or Mann-Whitney tests. *P < 0.05; **P < 0.01; ***P < 0.001.

Extrinsic Sle1 expression activates TC^a AM14 Id⁺ B cells

We have shown that each of the major Sle1 subloci, Sle1a, Sle1b, and Sle1c, confers a strong autoreactive CD4⁺ T cell phenotype [23, 27, 28]. Therefore, we postulated that Sle1 expressing non-tg cells should be sufficient to activate TC.AM14^a Id⁺ B cells. We used the same loss-of function, mixed BM strategy to test the ability of Sle1 non-tg cells to activate TC^a AM14 Id⁺ B cells compared with B6 or TC non-tg cells (Fig. 8A). Sle1 chimeras presented low numbers of splenocytes (Fig. 8B), in which the size of the B cell compartment was considerably reduced (Fig. 8C). Interestingly, the majority of the B cells in the Sle1 chimeras were of tg origin (Fig. 8D), which was consistent with Sle1 favoring BCR tg B cell expansion. Accordingly, the percentage of mId⁺ B cells was much higher than in any of the 2 other chimeras (Fig. 8E), and there was a trend for an increase in the percentage of splenocytes expressing iId (Fig. 8F). Most importantly, Sle1 non-tg cells supported the activation of TC^a AM14 Id⁺ B cells to AFC differentiation to a level similar to that of TC non-tg cells (Fig. 8G). It should be noted that although the non-tg B cells did not express the IgH^a allotype in these chimeras, IgH^a expression in the tg cells was sufficient to induce the production of Id⁺ AFCs from TC^a AM14 Id⁺ B cells in the presence of TC or Sle1 non-tg cells (Fig. 8G). This correlated with an increased percentage of CD69⁺ CD4⁺ T cells (Fig. 8H), the expansion of CD44⁺ CD62L⁻ effector memory over CD44⁻ CD62L⁺ naïve CD4⁺ T cells (Fig. 8I), as well as the expansion of the CD4⁺ CXCR5^high PD-1⁺ Bcl6⁺ Foxp3⁻ T_FH cells (Fig. 8J). Surprisingly, the percentage of GC B cells was not different between strains (Fig. 8K). A very small percentage of Id⁺ cells was GC B cells in the TC chimeras (Fig. 8L), confirming earlier results (Fig. 7H). This was not the case for the Sle1 chimeras, in which the percentage of Id⁺ GC B cells was equivalent to that of the B6 chimeras, both of which were significantly higher than TC ⁺ GC B cells. This could be a result of the large percentage of Id⁺ B cells in the Sle1 chimeras, as we have shown that unmanipulated TC.AM14^a mice correlates with a relatively high percentage of Id⁺ B cells found in GCs [12]. Overall, these results demonstrate that Sle1 expression in non-AM14 tg cells was sufficient to activate Id⁺ B cells into AFC differentiation, most likely through T cell help.

Figure 8. Sle1 expression in non-tg cells is sufficient to activate TC.AM14^a Id⁺ B cells.

(A) Experimental strategy for the BMT 2 BM chimeras with TC.AM14^a BM cells mixed with B6, Sle1, or TC BM cells. Total splenocyte numbers (B), percentage of B cells (C), IgM^a (tg) B cells (D), mId⁺ B cells (E),iId⁺ splenocytes (F), and splenic Id⁺ AFCs (G) in recipient mice that have received TC.AM14^a and TC, Sle1, or B6 BM cells. Percentage of CD69⁺ CD4⁺ T cells (H), ratio of CD44⁺ CD62L⁻ effector memory (T_EM) over CD44⁻ CD62L⁺ naïve (T_N) CD4⁺ T cells (I), and percentage of CD4⁺ CXCR5^high PD-1⁺ Bcl6⁺ Foxp3⁻ T_FH cells in the chimeric mice (J). Percentage of Fas⁺ GL7⁺ GC B220⁺-gated B cells (K) and percentage of Id⁺ IgM^a cells within these GC B cells (L). The x-axis of each graph shows the strain of origin on the non-tg BM cells combined with TC.AM14^a cells. Each symbol represents an individual chimeric mouse. Graphs show means + sem and the corresponding significance values of t-tests. *P < 0.05; **P < 0.01; ***P < 0.001.

DISCUSSION

The combination of the 3 NZM2410-derived Sle susceptibility loci is required to replicate the NZM2410 fully penetrant lupus phenotype on the B6 genetic background [11]. Each of these 3 loci plays a role in B cell tolerance, directly or indirectly [17, 18, 29, 30]. The AM14 model provided an ideal system to investigate the role of each locus in maintaining tolerance, as we have shown that AM14 RF B cells are spontaneously activated in mice expressing Sle1, Sle2, and Sle3 on a B6 background in the presence of the autoAg [12]. We have focused in this study on Sle1 and Sle2, both of which are intrinsically expressed in B cells [13, 15]. Moreover, genes directly affecting B cell tolerance have been identified for both loci [22, 30, 31]. Here, we show that Sle1 and Sle2 each promotes a modest level of AM14 RF B cell activation in the presence of autoAg. However, the frequency of Sle1.AM14^a or Sle2.AM14^a mice producing splenic RF AFCs or PBs was not significantly different from B6.AM14^a mice. Furthermore, unlike the combination of the 3 loci, neither Sle1 nor Sle2 was sufficient to promote the terminal differentiation of activated Id⁺ B cells into PCs. AM14 RF is largely produced by short-lived splenic PBs in the MRL/Lpr [25] and TC [12] mice. Moreover, the migration of long-lived PCs to the BM is defective in the TC model [32, 33]. Here, we showed that AM14 RF was also largely produced in the spleen of Sle1.AM14^a or Sle2.AM14^a mice, suggesting that it is a characteristic of the antibody specificity and/or affinity rather than the genetic background of the B cells.

AM14 HC tg and Id⁺ RF B cells showed a different fate among TC.AM14^a, Sle1.AM14^a, and Sle2.AM14^a mice. Sle1 expression decreased the percentage and number of total B cells but favored the expansion of AM14 HC tg B cells over endogenous B cells, and this was independent of the expression of the autoAg. Consistent with this result, it has been shown that Sle1 locus favored the expansion of DNA-specific 56R HC tg B cells relative to the endogenous B cells [34]. In the AM14 model, the expansion of the tg B cells resulted in a relatively low production of antichromatin IgG2a^a by endogenous B cells, thereby limiting autoAg availability for Id⁺ RF B cells. The correlation between Id⁺ AFCs and antichromatin IgG2a^a in Sle1.AM14^a mice strongly suggests that autoAg availability as a consequence of endogenous B cell exclusion is the limiting factor that prevents a major break of tolerance in Sle1.AM14^a mice. Sle2.AM14^a mice presented an expanded total B cell population, which can be partly attributed to the B cell hyperactivity associated with this locus [16]. The presence of autoAg increased the number of total B cells in Sle2.AM14^a mice but did not correlate with an elevated number of Id⁺ B cells. The level of antichromatin IgG2a^a did not correlate with the number of RF AFCs in Sle2.AM14^a mice, although the number of RF AFCs was significantly higher in Sle2.AM14^a than in Sle2.AM14 mice. This suggests an indirect role of IgG2a^a expression in the activation of Sle2.AM14^a RF B cells, which has yet to be determined. We have shown previously that Sle2 favored the expansion of endogenous B cells, as opposed to 56R tg B cells [18]. The expansion of AM14 tg B cells was significantly lower in Sle2.AM14^a than in TC.AM14^a mice. This suggests that Sle2 counterbalances the effect of Sle1 and allows a more robust expansion of endogenous B cells in TC.AM14^a mice, in which they secrete a sustained amount of antichromatin IgG2a^a to activate RF AM14⁺ B cells.

BCR expression was down-regulated in RF Id⁺ B cells in the presence of autoAg, not only in B6.AM14^a but also in TC.AM14^a, Sle1.AM14^a, and Sle2.AM14^a mice. On the BALB/c genetic background, RF Id⁺ B cells are not anergic, as they do not down-regulate IgM and respond to antigen-specific T cell help [35]. The low number of RF Id⁺ B cells on the B6 genetic background prevented us from testing their response to antigen-specific T cell help (data not shown). The down-regulation of surface Id⁺ expression suggests, however, that the autoreactive RF Id⁺ B cells are anergized in these mice, which is a major difference that can be attributed to the B6 genetic background. This suggests that the breach of tolerance in TC.AM14^a relies on extrinsic factors, such as T cell help.

Previous studies have shown that T cells, although not essential, play an important role in the modulation of the RF response in the MRL/lpr model [36] and in a PL2-3-induced model of activation of BALB/c.AM14 B cells [9]. The help was provided to the autoreactive B cells by extrafollicular antigen-specific CD4⁺ T through CD40L and IL-21. TC as well as B6.Sle1 mice have an increased number of follicular and extrafollicular T cells (this work and ref. [37]). More specifically, the Sle1 subloci are responsible for the production of activated nucleosome-specific T cells [38], expanded Th1 cells [23], and impaired T regulatory cells [27], as well as expanded T_FH [unpublished results]. Here, we show with BM chimeras that TC and Sle1 non-tg cells are required, at least in this experimental setting, for the activation of TC.AM14^a RF B cells. Therefore, it is most likely that follicular or extrafollicular T cells represent the major extrinsic factor contributing to the activation of the TC.AM14^a Id⁺ B cells. We have also shown that TC DCs produce an abnormal cytokine profile, including high levels of IL-6, IFN-α, and IFN-γ, which contributes to B cell activation [39, 40], and may play a role in TC.AM14^a Id⁺ B cell activation. High levels of IL-6 are produced by Sle1 DCs [41], and Sle3 expression results in the production of proinflammatory DCs that activate T cells [42] and could also activate B cells, directly or indirectly through T cells. Overall, our results indicate that intrinsic B cell activation is not sufficient for spontaneous differentiation of autoreactive Id⁺ RF AFCs in TC mice and suggest that Sle1-expressing T cells provide a necessary help.

In addition to being insufficient to promote a robust, spontaneous activation of AM14 RF B cells in the presence of autoAg, Sle1 or Sle2 expression did not sustain AM14 RF B cells activation by the dual BCR and TLR9 ligation triggered by the PL2-3 hybridoma. We have shown that unlike in BALB/c.AM14 mice, RF B cell activation could not be induced in B6.AM14 mice by PL2-3 immunization [12]. Thus, in the B6 genetic background, a stronger immune activation is necessary for AM14 RF B cells to respond to BCR and TLR9 coligation, which is indirectly supported by their requirement for extrinsic factors. This level of immune activation is provided by the combination of the 3 Sle loci but by neither Sle1 nor Sle2 alone. In conclusion, the Sle1 and Sle2 loci play intermediate roles in the breakdown of tolerance of RF B cells. Each locus can spontaneously activate RF B cells in the presence of the antigen and promote their differentiation toward antibody-secreting PBs, although at a low frequency. There is good evidence for a contribution of Sle1-expressing T cells to the process, but evidence for an intrinsic contribution of Sle1 to AM14 RF B cells is impeded by the issue of endogenous versus tg B cell selection associated with this locus. Nonetheless, these results demonstrated that the loss of tolerance by AM14 RF B cells in TC.AM14^a mice is multigenic and multifactorial, involving at least the coexpression of Sle1 and Sle2.

In addition to genetic contribution, the TC.AM14^a model allowed us to examine the effect of the AM14 RF B cell selection to the various mature B cell subsets. In the MRL/lpr or BALB/c strains, AM14 RF B cells accumulate in B cell follicles but not in the MZ [6, 10, 43]. Furthermore, BALB/c mice expressing site-directed HC and LC AM14 RF B cells lack MZB and B-1 cells [44]. We have shown a relative selection of AM14 RF B cells to the MZB subset in TC.AM14^a mice [12]. We examined this issue in more details, including in the single congenic AM14 mice. We found a preferential recruitment or expansion of MZB cells for AM14 HC B cells, which form a polyclonal population, and for Id⁺ B cells, regardless of the expression of the cognate antigen. These results suggest that the MZB selection may not be linked to autoreactivity but rather, associated with the expression of the AM14 HC itself, as it is the case for the antiphosphorylcholine M167 clone [45]. The selection of AM14 HC tg B cells to the MZB cells is, however, only observed in the B6 genetic background, establishing the existence of interactions between BCR specificity and genetic factors to determine the AM14 tg B cell fate. In the BALB/c background, Bruton's tyrosine kinase deficiency increased the selection of AM14 B cells to the MZ subset [44], predicting that a weaker BCR signaling of the AM14 B cells on a B6 (including TC) than on a BALB/c background. Nonetheless, the preferential selection of AM14 RF B cells to the MZ subset is unlikely to contribute to the breach of tolerance in TC.AM14^a mice for 2 reasons: first, it occurs to the same extent in all of the B6-based strains that we examined, whether or not AM14 RF B cells breached tolerance, and second, there were still a larger number of Id⁺ PBs from follicular origin than from MZ origin. We also found that a small number of AM14 RF B cells were selected to the B1-a cell subset in TC.AM14^a mice, but no evidence for RF production from these cells was found, indicating that this naturally autoreactive B cell compartment that is expanded in the TC model does not participate in the breach of tolerance by AM14 RF B cells.

The results reported in this study, combined with our initial characterization of the TC.AM14^a mice, suggest a multistep pathway for the activation and breakdown in tolerance of RF AM14 B cells (Fig. 9). In the absence of the TC Sle susceptibility loci and the IgG2a^a autoAg, B6 AM14 B cells remain inactivated regardless of the stimulation. The B6 genetic background is less permissive than BALB/c or MRL for the activation of RF AM14 B cells, as shown by the apparent anergy of autoreactive RF Id⁺ B cells, not only in B6.AM14^a but also in TC.AM14^a, Sle1.AM14^a, and Sle2.AM14^a mice. Therefore, strong extrinsic factors are required to breach tolerance. Expression of the autoAg and the presence of the Sle1 or Sle2 locus promote the activation and antibody secretion of AM14 B cells only in a subset of mice. Inflammatory cytokines secreted by DCs may also be required to maximize the activation of RF B cells in the presence of the Sle1 or Sle2 locus. When the 3 Sle loci are expressed in TC.AM14 mice, exogenous administration of BCR and TLR7/TLR9 ligands is sufficient to induce activation of RF AM14 B cells. Lastly, the highest level of activation is attained in aged TC.AM14^a mice, in which a continuous secretion of antichromatin IgG2a^a results in the production of a large amount of ICs that chronically stimulate BCR and TLR9. Therefore, this study has shown that multiple factors are responsible for the breach of tolerance of B cells in lupus and that the nature of these factors depends on the genetic background.

Figure 9. Proposed multistep mechanism for the activation of AM14 RF B cells in the TC model.

In the absence of lupus susceptibility loci and the autoAg, the RF B cells remain in an inactivated state. The presence of at least one Sle locus and the autoAg provides ligation for the BCR and the endosomally located TLR7/TLR9. Expression of a single Sle locus results in an intermediate phenotype, where only a subset of mice display activated and antibody-secreting RF B cells. This is achieved most likely through Sle1-induced CD4⁺ T cell help and Sle2-induced B cell-intrinsic activation. Cytokines produced by DCs expressing Sle1, Sle3, or a combination of both may also contribute to the activation of AM14 RF B cells. Exogenous administration of BCR and TLR7/TLR9 ligands is sufficient to promote the activation and BCR intracellular localization of the Id⁺ receptor in TC.AM14 mice but does not elicit antibody secretion. Maximal AM14 B cell activation and antibody secretion are observed in aged TC.AM14^a mice as a result of chronic activation of the B cells by ICs.

AUTHORSHIP

A.S. and Y.Y.Z. designed and performed experiments and analyzed data. S.-C.C. performed experiments and analyzed data. L.Z. performed experiments. L.M. designed experiments, analyzed data, and prepared the manuscript.

Glossary

AFC

antibody-forming cell

AP

alkaline phosphatase

autoAb

autoantibodies

autoAg

autoantigen

B6

C57BL/6

B6.AM14

B6.AM14.IgH^b

B6.AM14^a

B6.AM14.IgH^a

B6^a

C57BL/6J-Cg-Igh^aThy1aGpila/J

BM

bone marrow

BMT 1

experiment with donor BM cells from TC.AM14^a plus either B6.IgH^a.Thy1^a or TC.IgH^a mice

BMT 2

experiment with donor BM cells from TC.AM14^a plus either B6, B6.Sle1, or TC mice

CD40/62L

cluster of differentiation 40/62 ligand

DC

dendritic cell

FOB

follicular B cell

Foxp3

forkhead box p3

GC

germinal center

HC

heavy chain

IC

immune complex

Id

idiotype

iId

intracellular idiotype

mId

membrane idiotype

MRL

Murphy Roths Large

MZ

marginal zone

MZB

marginal zone B cell

PB

plasmablast

PC

plasma cell

PD-1

programmed death 1

PerC

peritoneal cavity

RF

rheumatoid factor

Sle1.AM14

B6.Sle1.AM14

Sle1.AM14^a

B6.Sle1.AM14.IgH^a

Sle2.AM14

B6.Sle2.AM14

Sle2.AM14^a

B6.Sle2.AM14.IgH^a

Sle

systemic lupus erythematosus

TC

B6.NZM-Sle1^NZM2410/Aeg Sle2^NZM2410/Aeg Sle3^NZM2410/Aeg/LmoJ triple congenic strain

TC^a

TC.IgH^a

TC.AM14

TC.AM14.IgH^b

TC.AM14^a

TC.AM14.IgH^a

T_FH

T follicular helper

tg

transgenic

DISCLOSURES

The authors declare no conflict of interest. The paper’s contents are solely the responsibility of the authors and do not necessarily represent the official views of NIAID, NIAMS, or NIH.