JNK Regulatory Molecule G5PR Induces IgG Autoantibody–Producing Plasmablasts from Peritoneal B1a Cells

Autoantibodies are often associated with autoimmune pathogenesis, as indicated by correlations with various symptomatic manifestations, laboratory findings, deposition of immune complexes in pathological specimens, and recapitulation of diseases by serum transfer (1–4). Peritoneal B cells are thought to be the source of autoantibody-producing long-lived plasma cells in various autoimmune-prone mice such as New Zealand Black (NZB) and BXSB mice (5, 6), and mice genetically engineered for various molecules involved in signal transduction pathways of B cell survival and apoptosis (7–14). Expression of CD5 and CD11b classifies B1a (CD5^high CD11b^low), B1b (CD5⁻ CD11b^high), and B2 (CD5⁻ CD11b⁻) cells from the peritoneal cavity and spleen (15, 16). B1 cells are thought to originate from different B-lineage precursors of conventional B (B2) cells, and show a unique character associated with their localization, self-renewal potential, BCR signaling, and survival potential in the peritoneal cavity (17–20). B1a cells express a restricted repertoire of BCR specificity and are often reactive to self-antigens, as well as T cell–independent Ags including pathogens (21, 22). B1a cell–derived Abs are produced as natural Abs involved in innate immunity, but also include autoantibodies that are usually the IgM class with reactivity to self-antigens During aging of autoimmune-prone mice, the number of B1a cells increases abnormally, which generates plasmablasts (PBs) that produce autoantibodies against ssDNA and dsDNA (5, 23).

B2 cells are selected as naive B cells of the non–self-reactive primary repertoire in the bone marrow, which are distributed to peripheral lymphoid organs. Ag stimulation with the help of Th cells induces rapid expansion of naive B cells in the germinal centers (GCs) of lymphoid follicles, where Ag-reactive B cells undergo somatic hypermutation (SHM) of IgV region genes and class switch recombination (CSR) at the S regions by AID (24–29). Thus, B2 cells gain IgV region SHM and CSR in the GCs of lymphoid follicles to become high-affinity Ag-reactive and IgG Ab-producing plasma cells (30).

B1 and B2 cells might be originally derived from different lineages or alternatively generated by the consequence of different activation signals provided by some restricted Ag properties that induce B cell activation in a T cell–independent manner (31–33). In Ig transgenic mice, ligand-dependent signaling, presumably through the BCR, affects the generation and maintenance of B1 cells in the peritoneal cavity. Thus, the molecules involved in differentiation of GC B cells may be also associated with differentiation of B1a cells that appear in the peritoneal cavity. G5PR is a kind of regulatory subunit for protein phosphatases (34), which is characterized as a molecule expressed in GC B cells (14). G5PR is associated with the catalytic subunit of protein phosphatase 2A and is involved in suppression of JNK activation, leading to Bim dephosphorylation (35, 36). B cell–conditional targeting of g5pr in cd19-cre transgenic mice results in the impairment of GC B cell survival (35). Interestingly, transgenic G5PR overexpression (G5PR^Tg) under the control of the Lck proximal promoter and Eμ enhancer regions leads to abnormal survival of B1a cells in the peritoneal cavity and development of autoimmunity after aging in female mice (14). These findings suggest that G5PR is involved in both the survival of GC B cells and the differentiation of peritoneal B1a cells into autoantibody-producing plasma cells.

Autoimmunity is associated with the generation of high-affinity autoantibodies that are strongly reactive to self-antigens and class switching, typically with IgG class anti-ssDNA and anti-dsDNA Abs produced by long-lived plasma cells (37–39). Such high-affinity IgG-switched autoantibodies might be produced from cells that are generated by a distinct differentiation process in a yet undetermined GC-like structure localized at the adjacent region of the peritoneal cavity (40). Alternatively, autoantibodies may be produced by cells resulting from the entry and differentiation of peritoneal B1a cells in the GCs of lymphoid follicles. Autoimmune-prone mice such as NZB, NZB × New Zealand White F₁ (BWF1), and BXSB show spontaneous GC formation in the spleen without Ag immunization (41). Stimulation of B1a cells by LPS and IL-5 induces production of autoantibodies in vitro (42). B1a cells might be capable of differentiating into IgG autoantibody–producing PBs in the spleen, as suggested by introduction of a purified and enriched B1a cell population obtained from the peritoneal cavity of autoimmune-prone mice (43). The production of pathogenic IgG autoantibodies is dependent on CD40L that is also required for GC formation in NZB mice (44). Thus, IgM B1a cells in the peritoneal cavity are the likely precursors of PBs secreting high-affinity IgG autoantibodies against self-antigens associated with autoimmunity (19). However, the mechanisms underlying IgV-region SHM and S-region CSR remain to clarified in B1a cells, and it is unclear where B1a cells undergo such maturation processes, as is the case for conventional Ag-reactive B cells at GCs during immune responses.

In this study, we investigated whether IgM B1a cells can directly mature into GC-like B cells and differentiate into IgG autoantibody–producing plasma cells, and we clarify the stimuli and molecule(s) responsible for these differentiation processes. To this end, we used an in vitro culture system that supports long-term growth of GC B cells (iGB) on fibroblasts expressing CD40L and BAFF (45), and observed the cells in the GC area after adoptive transfer in vivo. As a result, the B1a cells that were highly enriched from the peritoneal cavity of BWF1 and G5PR^Tg mice differentiated into GC-like B cells. The GC-like B cells derived from B1a cells underwent differentiation into IgG Ab-producing PBs at a high frequency compared with conventional B2 cells of the spleen through regulation of JNK activity. This approach may resolve the origin and molecular mechanism of autoantibody-producing PBs in autoimmunity.

C57BL/6 (B6) and BWF1 mice were purchased from Kyudo (Fukuoka, Japan) and SLC (Hamamatsu, Japan), respectively. G5PR^Tg mice were maintained under specific pathogen-free conditions at the Center for Animal Resources and Development, Kumamoto University. All procedures were carried out according to Center for Animal Resources and Development regulations for animal care.

Cells from the spleen and peritoneal cavity were depleted of RBCs by ammonium chloride buffer (Sigma-Aldrich, St. Louis, MO). The cells were stained with various combinations of mAbs after blocking the FcR with an anti-CD16/32 mAb (eBioscience, San Diego, CA). The mAbs used in this study were Alexa Fluor 488–anti-CD11b, allophycocyanin–anti-IgG₁ (Biolegend, San Diego, CA), FITC–anti-GL7, PE–anti-CD5, PE–anti-CD138, PE–anti-Fas (BD Biosciences, San Jose, CA), FITC–anti-IgG₁, allophycocyanin–anti-B220, and allophycocyanin–anti-IgM (eBioscience). After washing, the cells were stained with 7-AAD (BD Biosciences) to exclude the dead cells from analysis. For intracellular staining, the cells were fixed and permeabilized with CytoFix/CytoPerm buffer (BD Biosciences) and then stained with an Alexa Fluor 647–anti–p-JNK mAb (Cell Signaling Technology, Danvers, MA), Alexa Fluor 647–anti-BCL6 mAb, or an isotype control (BD Biosciences). Stained cells were analyzed by a FACSAria, FACSCalibur, or FACSVerse (BD Biosciences) and FlowJo software (Tree Star, Ashland, OR). B1a (B220⁺ CD5^high CD11b^low), B1b (B220⁺ CD5⁻ CD11b^high), and B2 (B220⁺ CD5⁻ CD11b⁻) cells from the peritoneal cavity and spleen B2 (B220⁺ CD5⁻) cells were isolated by the FACSAria (>98% purity). For GC B cell isolation, B6 mice (female, 8–12 wk old) were immunized with nitrophenyl-chicken γ-globulin (Biosearch Technologies, Novato, CA) in alum. At 10 d after immunization, B220⁺ Fas⁺ GL7⁺ cells were isolated by the FACSAria.

Sorted peritoneal B1a, B1b, B2, and spleen B2 cells (5 × 10⁵ cells) were cultured on irradiated feeder cells with 1 ng/ml IL-4 (PeproTech, Princeton, NJ). Feeder cells expressing CD40L and BAFF (40LB) were prepared as described previously (45). After 4 d of culture, the cells were transferred onto new feeder cells with 10 ng/ml IL-21 (PeproTech) and cultured for 3 d in the presence or absence of 10 μg/ml JNK inhibitor SP600125 (Wako, Osaka, Japan). To detect anti-dsDNA Ab production, we recovered the cells at day 4 of iGB culture; then 1 × 10⁵ cells were cultured with fresh medium without cytokines for 7 d.

Total RNA from sorted cells was purified using an RNeasy Micro kit (Qiagen, Hilden, Germany). cDNA was prepared with Superscript III (Invitrogen, Carlsbad, CA). Quantitative PCR was performed with TaqMan gene expression assays (aicda, Mm01184115; bcl6, Mm00477633; gapdh, Mm99999915; g5pr, Mm01257828; prdm1, MM00476128; xbp1, Mm00457360; Applied Biosystems, Foster, CA) or Thunderbird SYBR qPCR mix (Toyobo, Osaka, Japan) with primer sets for each gene on an ABI7500 FAST or ABIViiA7 (Applied Biosystems) using default cycling conditions. Gene expression was analyzed by the relative standard curve method and normalized to gapdh expression. The following primer sets were used: gapdh, sense 5′-GGA GAA ACC TGC CAA GTA TGA-3′ and antisense, 5′-CCC TGT TGC TGT AGC CGT ATT-3′; irf-4, sense 5′-CTA CCC CAT GAC AGC ACC TT-3′ and antisense, 5′-CCA AAC GTC ACA GGA CAT TG-3′.

Anti-dsDNA Abs in culture supernatants were detected by ELISA as described previously (14). In brief, culture supernatants were incubated for 1 h at room temperature in 96-well plates coated with 500 ng dsDNA (Sigma-Aldrich). After washing, the captured Abs were detected by alkaline phosphatase–conjugated anti-mouse IgG (γ-chain specific) or anti-mouse IgM (μ-chain specific; Sigma-Aldrich) in combination with p-nitrophenyl phosphate substrate (Sigma-Aldrich). The Ag–Ab reaction was measured by the absorbance at 405 nm using a MULTISKAN FC plate reader (Thermo Fisher Scientific, Waltham, MA). Total amounts of secreted IgM and IgG in culture supernatants were measured by mouse IgM and IgG ELISA quantitation sets (Bethyl Laboratories, Montgomery, TX), respectively.

B1a cells from BWF1 mice (female, 8–12 wk) were cultured for 3 d on irradiated 40LB cells with IL-4. IgG₁⁻ CD138⁻ cells were purified by the FACSAria and then labeled with 5 μM CFSE (Invitrogen). CFSE-labeled cells (5 × 10⁶ cells) were transferred into aged female BWF1 mice (over 8 mo) via the tail vain. The spleens of the aged mice were then analyzed at 2 d after transfer.

Surgically excised spleens were embedded in OCT compound (Sakura Finetech, Tokyo, Japan) and frozen immediately in liquid nitrogen. The frozen blocks were cut into 6-μm-thick sections using a Cryotome (Thermo Shandon, Cheshire, U.K.). The sections were fixed with acetone, blocked with BlockAce (DS Pharma Biomedical, Osaka, Japan), and then stained with the following reagents: biotin-peanut agglutinin (PNA; Vector Laboratories, Burlingame, CA), allophycocyanin–anti-IgG₁, allophycocyanin–anti-IgD (Biolegend), purified anti-CD138 (BD Biosciences), Alexa Fluor 488–anti-FITC, Alexa Fluor 594–streptavidin, Alexa Fluor 647–anti-rat IgG, and DAPI (Invitrogen). Signals were observed using a FV1200 confocal microscope with FluoView software (Olympus, Tokyo, Japan).

Data are presented as the mean ± SD. Statistical analysis was performed by the Student t test. A p value <0.05 was considered to be statistically significant.

Similar to spleen B cells, we examined whether peritoneal B1a cells can differentiate into GC-like B cells by iGB culture with IL-4, CD40L, and BAFF (45). B220⁺ B-lineage cells from the peritoneal cavity were sorted into CD5^high CD11b^low (B1a), CD5⁻ CD11b^high (B1b), and CD5⁻ CD11b⁻ (B2) cells. These cells were then subjected to iGB culture (Fig. 1A). Peritoneal B1a, B1b, and B2 cells, as well as spleen B2 cells, differentiated into GC-like B cells with surface marker expression of Fas⁺ GL7⁺ in vitro (Fig. 1B), which was similar to spleen GC B cells obtained from Ag-immunized mice (Supplemental Fig. 1). Similar to spleen GC B cells, iGB cells obtained from peritoneal B-lineage populations and spleen B2 cells showed upregulation of BCL-6 (Fig. 1C). These results indicate that peritoneal B1a cells possess the potential to differentiate into GC-like B cells when stimulated with IL-4, CD40L, and BAFF in vitro.

iGB culture of spleen IgM⁺ B cells induces maturation and maintains the long-term proliferative ability of GC-like B cells in vitro. However, some spleen iGB cells further differentiate toward class-switched IgG₁⁺ B cells and/or PBs (45). Thus, we examined later GC-stage differentiation of peritoneal B1a-derived GC-like B cells. After 4 d of culture, iGB cells derived from B1a cells showed a marked decrease of IgG₁ class switching (14.8 ± 3.9%) compared with that of iGB cells derived from spleen B2 cells (38.0 ± 2.7%; Fig. 2A). This tendency was also observed in peritoneal B1b cells (22.8 ± 3.2%), but not in peritoneal B2 cells (34.6 ± 2.4%). In contrast, iGB culture augmented the differentiation of peritoneal B1a cells into CD138⁺ PBs (15.9 ± 6.8%) compared with that of spleen B2 cells (4.8 ± 1.6%; Fig. 2B). These results suggest that B1a cells differ from spleen B2 cells in terms of their potential for differentiation toward PBs. GC-phenotype (GL7⁺CD138⁻) cells from 2-d iGB culture of B1a cells were enriched, checked by postsort analysis, and then recultured on fresh feeder cells to address how PBs differentiate from B1a cells. Up to 38.9% of these cells differentiated into CD138⁺ PB cells, suggesting that B1a cells also differentiate into Ab-secreting PBs through the GC-like stage (Fig. 2C).

To understand how peritoneal B1a cells differ in the GC-like B cell maturation process compared with the earlier B cell populations, we measured transcripts of genes associated with maturation and differentiation of B cells in GCs (Fig. 3). Peritoneal B1a cells similarly showed augmented expression of aicda and g5pr as the early-stage transcripts of GC B cells after iGB culture. g5pr expression is selectively upregulated in GC B cells of the light zone area (14). B1a cells showed increased expression of g5pr. As a GC maturation-associated molecule, B1a-derived iGB cells did not show an increase in bcl6 expression (Fig. 3), whereas a marked increase was observed in BCL6 protein expression (Fig. 1C). BCL6 expression is presumably regulated by posttranscriptional mechanisms as reported previously (46, 47). Interestingly, B1a cells showed a unique transcription profile with higher levels of prdm1 and irf4 expression compared with those in spleen B2 cells. B1a-derived iGB cells showed higher levels of prdm1, irf4, and xbp1 expression, implying that B1a cell differentiation is biased toward the PB stage.

To determine whether B1a cell maturation leads to autoantibody production in autoimmune-prone mice, we examined the differentiation of peritoneal B1a cells of BWF1 mice in comparison with spleen B2 cells. Peritoneal B1a cells of female BWF1 mice showed a similar decrease in class switching to IgG₁⁺ cells (13.7 ± 5.3%) compared with that of spleen B2 cells (Fig. 4A). However, the peritoneal B1a cells showed increased differentiation into PBs (47.9 ± 2.1%) during iGB culture (Fig. 4B). The B1a-derived PBs generated during iGB culture secreted high levels of total IgM (Fig. 4C) and anti-dsDNA Abs of both IgM and IgG classes in vitro (Fig. 4D).

Next, we investigated whether the GC-like B cells generated from peritoneal B1a cells migrate into the spleen of secondary lymphoid organs in vivo. GC-like B cells were obtained from peritoneal B1a cells of BWF1 mice by iGB culture for 3 d, labeled with CFSE, and then adoptively transferred into aged BWF1 mice (Fig. 5A). At 2 d after adoptive transfer, spleen sections were subjected to CFSE, IgG₁, PNA, CD138, IgD, and DAPI staining. In three-color analysis, CFSE-labeled cells were found to localize in the GC area and were positive for both IgG₁ and PNA staining (Fig. 5B, upper left, boxed). The cells were then analyzed by DAPI instead of PNA staining (Fig. 5B, upper right). CFSE⁺ cells showed IgG₁ staining in their periphery with clear nuclear DNA staining (merge). When stained for CD138 and IgD, CFSE-labeled cells were merged with CD138 staining, indicating that PBs are localized in the extrafollicular area (Fig. 5B, middle left, boxed) and then merged with DAPI staining (Fig. 5B, middle right). In addition, B1a-derived, CFSE-labeled cells appeared at the marginal zone, merged with IgD (Fig. 5B, lower boxed) and DAPI staining. Flow cytometric analysis showed that CFSE-labeled cells in the spleen also expressed either IgG₁ or CD138 (Fig. 5C), which confirmed recruitment of B1a-derived cells to the spleen. These results demonstrate that peritoneal B1a cells maintained in iGB culture indeed appear at the GC region of the spleen in autoimmune-prone BWF1 mice.

BCR-mediated signaling is regulated by the JNK pathway, leading to activation of proapoptotic protein Bim in the B cells of autoimmune-prone mice (14). Therefore, we compared the JNK activation status of peritoneal B1a cells in BWF1 and B6 mice. Flow cytometric analysis revealed a decrease of intracellular p-JNK in B1a cells from BWF1 mice (Fig. 6A). To confirm that this effect is mediated through JNK, we added the specific JNK inhibitor SP600125 to the culture. B1a cell differentiation into PBs increased from 30.7 ± 12.3 to 46.7 ± 4.3% in the presence of SP600125, but showed decreased differentiation to IgG₁ cells from 41.6 ± 8.0 to 30.2 ± 13.4% (Fig. 6B). This tendency was similar in spleen B2 cells (from 3.3 ± 2.2 to 8.5 ± 2.5%), but more marked in B1a cells. Collectively, JNK signaling is critical for suppression of B1a cell differentiation into PBs.

CSR is initiated by expression of AID, which is regulated by JNK signaling (48). Female G5PR^Tg mice produce autoantibodies against dsDNA of IgM and IgG classes after aging (14). Thus, we investigated the differentiation of B1a cells of G5PR^Tg mice in terms of class switching from IgM to IgG₁ during iGB culture. B1a cells from G5PR^Tg mice showed a marked decrease of differentiation to IgG₁⁺ B cells compared with that of B1a cells from wild-type (wt) B6 mice (40.0 ± 10.2 versus 51.1 ± 4.3%; p < 0.05; Fig. 7A). Spleen B2 cells from G5PR^Tg and wt mice displayed similar frequencies of IgG₁⁺ cells (Fig. 7A). B1a cells from G5PR^Tg mice showed an increase of differentiation to PBs (from 19.1 ± 5.3 to 28.9 ± 8.9%; p < 0.05; Fig. 7B). The effect of G5PR overexpression was marked in B1a cells, but not in spleen B2 cells, as similar frequencies of class switching were observed (Fig. 7B).

We next investigated the expression of transcription factors and molecules involved in early differentiation of B1a cells into PBs in G5PR^Tg mice (Fig. 7C). B1a cells of G5PR^Tg mice showed a marked increase of prdm1 expression compared with that in B1a cells of wt mice. This bias appeared to decline along with the differentiation of GC-like stage cells after iGB culture. The expression of aicda involved in Ig class switching was slightly lower in B1a cells than that in spleen B2 cells (Fig. 7C). The approximate 30% reduction of aicda expression was the likely cause of the decreased frequency of IgG₁ B cells (Fig. 7A).

Considering that B1a cells from G5PR^Tg mice tended to undergo differentiation toward PBs, we addressed whether these PBs include autoantibody producers. The culture supernatants of G5PR^Tg B1a cells showed a greater increase of the total IgM level and anti-dsDNA Abs of both IgM and IgG classes compared with those in supernatants of wt B1a cells (Fig. 7D, 7E). iGB culture of spleen B2 cells did not reveal any differences of autoantibody levels in G5PR^Tg and wt mice. The increase of G5PR expression augmented the differentiation of B1a cells into autoantibody-producing PBs. These results demonstrate that dysregulation of JNK signaling has a critical effect on the generation and differentiation of autoantibody-producing PBs from peritoneal B1a cells in an autoimmune state.

Similar to spleen B2 cells, in vitro culture with stimulation by IL-4, CD40L, and BAFF induced differentiation of peritoneal B1a cells into mature GC-like B cells expressing Fas and GL7. These results indicate that peritoneal B1a cells respond to these stimuli and undergo differentiation similar to GC B cells induced by T cell–dependent Ags in vivo. However, compared with spleen B2 cells, B1a cells displayed unique properties in GC differentiation processes after the GC-like stage in iGB culture. These properties included decreased class switching toward IgG⁺ GC B cells and, in contrast, increased differentiation into Ab-producing PB stage cells. The GC-like B cells from the B1a cell culture also showed lower induction of aicda expression, but higher levels of prdm1 and irf4 expression. This finding indicates that B1a-derived, GC-like B cells have a unique expression profile of transcription factors and/or molecules associated with later maturation of GC B cells. Such an expression profile might be caused by the difference of microenvironments in the spleen and peritoneal cavity that is influenced by the visceral organs or microbial flora of the gut. Alternatively, the altered genomic state of B1a cells might be critical for the difference in B1a and spleen B2 cells. Further study of purified B1a-derived GC-like B cells obtained by iGB culture would help to elucidate the chromosomal state of B1a cells.

Autoimmune-prone female BWF1 mice produce autoantibodies of IgM and IgG classes against ssDNA and dsDNA, and aged BWF1 mice display formation of GC-like microscopic architecture with PNA staining in the spleen without exogenous Ag immunization (41). The spontaneous development of GC-like structures might be caused by entry of peritoneal B1 cells into the spleen. To investigate direct communication of peritoneal B1a cells and the spleen GC region, we labeled iGB cells generated from peritoneal B1a cells with CFSE and adoptively transferred them into autoimmune BWF1 mice (Fig. 5). B1a-derived iGB cells showed localization similar to that of B2-derived iGB cells as reported previously (45). Because iGB-cultured cells are unselected for Ag specificity, they presumably migrate into the extrafollicular region. Thus, more CFSE⁺ cells were detected in the extrafollicular region as shown in Fig. 5B (lower image). Some of the B1a-derived iGB cells indeed localized to the GC area and appeared as IgG⁺ cells and PBs in vivo. Thus, the aberrant GC formation might be associated with the generation of autoantibodies in the autoimmune state.

For spleen B2 cells, the rescue and survival of B cells appeared to be strictly dependent on BCR signaling. B cell–conditional targeting of G5PR that upregulates JNK signaling leading to Bim activation causes a severe loss of spleen B cells in mice (35). In contrast, G5PR overexpression in G5PR^Tg mice preferentially affects the survival of B1a cells (14). One possible explanation for this cell-type–specific effect might be continual stimulation of B1a cells by autoantigens in the peritoneal cavity. BCR signaling of autoreactive B1a cells might allow the increase and facilitated differentiation of B1a cells in G5PR^Tg mice. Alternatively, B1a cells possess a unique program of GC maturation compared with that of spleen B2 cells. Furthermore, the majority of B1a cells probably do not respond to autoantigens in this culture system. Without autoantigens in this culture system, B1a cells of G5PR^Tg mice showed a marked difference in differentiation compared with that of spleen B2 cells. JNK signaling is activated in peritoneal B1a cells through a BCR-independent pathway in iGB culture. In support of this effect of JNK signaling, the JNK inhibitor SP600125 induced similar prompt maturation of B1a cells into PBs (Fig. 6B).

B cells undergo differentiation into GC B cells or PBs by cognate interactions with Ags and costimulatory signals in the GC microenvironment through regulation of transcription factors. IRF4 might play a pivotal role in the mutual regulation of transcription factors in the differentiation processes toward GC B cells and terminally to PBs. GC B cell differentiation accompanies the expression of bcl6, pax5, bach2, and obf1, but PB differentiation requires the expression of blimp1/prdm1 and xbp1 under the regulation of IRF4 (49). IRF4 has been proposed to form a complex with PU.1, Spi-B, or BATF containing AP-1 heterodimers on the Ets-IRF composite motif or AP-1–IRF composite motif. The kinetics and expression level of irf4 affect global interactions at various transcription-competent elements for transcription initiation and regulation of various genes as reported by Ochiai et al. (50). Our results might agree with this model, which is presumably associated with G5PR-dependent regulation of JNK signaling strength. In G5PR^Tg mice, B1a cells reduce the extent of JNK activation, leading to activation and transcription of c-jun (14). The altered c-Jun interaction with AP-1 could be a target of G5PR in regulation of B1a cell maturation.

B1a-derived GC-like B cells undergo CSR to IgG₁ in vitro, implying that the class switching of B1a cells can occur in the extraperitoneal region in the presence of costimulatory signals by CD40L and BAFF. In culture, aicda expression was rather low in both B1a and B2 cells compared with that in GC B cells obtained from the spleens of Ag-immunized mice (Fig. 3). GC-like B cells showed less aicda expression compared with that in GC B cells generated in vivo. In addition to the peritoneal cavity, it would be important to determine whether the maturation of BCRs can also occur in B1a-derived GC-like B cells against autoantigens in the extraperitoneal region. Various autoimmune diseases display the formation of GC-like architecture at inappropriate or aberrant regions such as the conjunctiva, the synovial membrane of the joint, the extrafollicular region of the spleen or lymphoid tissues, and many kinds of epithelial membranes (51, 52). In the autoimmune-prone mouse model, B cell maturation can occur at the extra GC region (53), which may implicate the existence of aberrant stimulatory signals that induce expression of aicda and the associated molecules under the autoimmune-prone status.

The B1a cells of autoimmune-prone G5PR^Tg mice showed prompt differentiation into PBs in iGB culture. Overexpression of the phosphatase component G5PR facilitates the differentiation of GC-like B cells into plasma cells. In contrast, JNK activation in naive B cell is associated with differentiation into GC B cells (54, 55). The B cells of G5PR^Tg mice show anti-IgM–induced JNK activation during the initial period (5 min) similar to wt B cells. However, the phosphorylation does not continue over 30 min because of the phosphatase activity (14). This result suggests that G5PR^Tg B cells differentiate into GC B cells, but then tend to the prompt differentiation into plasma cells under both in vitro and in vivo conditions. Overexpression of G5PR increases the number of peritoneal B1a cells and the generation of anti-dsDNA Abs, leading to autoimmunity with immune complex deposition in the kidneys of aged female mice (14). The B1a cells from G5PR^Tg mice generated higher numbers of PBs compared with those from wt mice, indicating that downregulation of JNK activation by G5PR results in early entry into the PB differentiation pathway. This notion is in accordance with previous reports that show the regulation of plasma cell differentiation of B-lineage cells by increased expression of blimp1/prdm1 and xbp1, and suppressed expression of bcl6 via B cell signals through the JNK pathway (56). G5PR may be a critical regulator of GC B cell differentiation through the signals provided to B1a cells.

In conclusion, we used an in vitro culture system of GC B cells for proliferation and differentiation of B1a lineage cells. Using B1a cells isolated from the peritoneal cavity, we demonstrate that B1a cells can differentiate into IgG⁺ autoantibody-producing PBs in vitro and appear in the GCs of peripheral lymphoid organs. The regulation of JNK signaling in B1a cells might be a critical target to prevent autoantibody production in autoimmune diseases.

We thank Mika Hirota for technical and secretarial assistance.

The authors have no financial conflicts of interest.