Blimp-1 has been identified as a key regulator of plasma cell differentiation in B cells and effector/memory function in T cells. We demonstrate that Blimp-1 in dendritic cells (DCs) is required to maintain immune tolerance in female but not male mice. Female mice lacking Blimp-1 expression in DCs (DCBlimp-1ko) or haploid for Blimp-1 expression exhibit normal DC development but an altered DC function and develop lupus-like autoantibodies. Although DCs have been implicated in the pathogenesis of lupus, a defect in DC function has not previously been shown to initiate the disease process. Blimp-1ko DCs display increased production of IL-6 and preferentially induce differentiation of follicular T helper cells (TFH cells) in vitro. In vivo, the expansion of TFH cells is associated with an enhanced germinal center (GC) response and the development of autoreactivity. These studies demonstrate a critical role for Blimp-1 in the tolerogenic function of DCs and show that a diminished expression of Blimp-1 in DCs can result in aberrant activation of the adaptive immune system with the development of a lupus-like serology in a gender-specific manner. This study is of particular interest because a polymorphism of Blimp-1 associates with SLE.
Tolerance to self-antigens is a key feature of the immune system. Systemic autoimmune diseases, such as systemic lupus erythematosus (SLE), result from dysregulation of B and T cell activation and altered function of macrophages and DCs (Shlomchik, 2009), leading to pathogenic autoantibodies, which are IgG isotype switched, high affinity to self-antigens, and somatically hypermutated. (Rothfield and Stollar, 1967; Diamond and Scharff, 1984). Although B cells are the proximal cells in the phenotypic manifestations of SLE, interactions with other immune cell types are dynamically involved (MacLennan, 1994; Shlomchik et al., 2001; Craft, 2011).
Recently, a polymorphism of Blimp-1 has been identified as a risk factor in SLE by genome-wide association studies, suggesting a critical function of Blimp-1 in SLE (Gateva et al., 2009; Han et al., 2009). Blimp-1 negatively regulates expression of IFN-β in both humans and mice (Keller and Maniatis, 1991; Turner et al., 1994). In B cells, Blimp-1 is a key regulator of plasma cell development (Shapiro-Shelef et al., 2003). In T cells, Blimp-1 regulates the differentiation of TH1 and the function of regulatory T cells (Martins et al., 2006). Blimp-1 was suggested to be a survival factor in monocytes (Chang et al., 2000); however, a more recent study using a Tie2–CRE system has suggested that Blimp-1 may regulate the differentiation and activation of DCs (Chan et al., 2009). In that study, IL-6 and MCP-1 were shown to be direct targets of Blimp-1 and the deletion of Blimp-1 resulted in increased expression of proinflammatory cytokines. The physiological importance of Blimp-1 specifically in DCs, however, could not be addressed in this mouse model in which Blimp-1 was deleted in all hematopoietic cell lineages.
DCs are important in lupus pathogenesis, although a primary defect in DC function has not been reported. Because DCs were discovered by Steinman and Cohn (1973), they have been recognized as the key immune-regulating cells. DCs can mediate both immune tolerance and immune activation (Cools et al., 2007). DCs can acquire tolerogenic phenotype after phagocytosis of apoptotic cells (Qiu et al., 2009). They can also generate regulatory T (Treg) cells or cause immune suppression by secretion of cytokines (Yamazaki et al., 2008). Although differentiation of DCs into immunogenic or tolerogenic DCs has not been fully characterized, it is generally accepted that DC maturation status rather than DC lineage alone determines the functionality of DCs (Cools et al., 2007).
In this study, DCBlimp-1ko mice were generated by mating Blimp-1flox mice to CD11c-CRE+ mice to identify the in vivo consequences of Blimp-1 deficiency to DC function. Female, but not male, DCBlimp-1ko mice developed autoantibodies with extensive mutations, suggesting their maturation in a germinal center (GC) response. Consistent with this observation, female mice display increased GC formation in the basal state and after immunization, accompanied by an increased frequency of TFH cells. Finally, DCs from female DCBlimp-1ko mice produced increased IL-6 and preferentially induced differentiation of TFH cells. All aspects of the phenotype were abolished in DCBlimp-1ko mice haploid for IL-6. Together, these observations suggest that a defect restricted to DCs can alter T cell differentiation resulting in the production of high titers of lupus-like autoantibodies in a gender-specific fashion.
RESULTS AND DISCUSSION
Gender-dependent development of autoantibodies in DCBlimp-1ko mice
Blimp-1 expression was measured by Western blotting (Fig. 1 A). CD11chi DCs were purified from spleens of age-matched DCBlimp-1ko (Blimp-1flox/flox; CD11c-CRE+) and control (either Blimp-1+/+; CD11c-CRE+ or Blimp-1flox/flox; CD11c-CRE−) mice. Splenic DC expression of Blimp-1 in DCBlimp-1ko mice was barely detectable compared with control DCs in both female and male. DCs from Blimp-1flox/+; CRE+ mice displayed an intermediate level of expression. Deletion of Blimp-1 is DC specific because the Blimp-1 level was unaffected in B cells, NK cells, macrophages, and α-CD3/CD28–activated T cells (Fig. S1 A).
Consistent with a previous study (Chan et al., 2009), development of DCs was normal in both conventional DCs (cDCs; CD11chiSiglec-H−) and plasmacytoid DCs (pDCs; CD11cloSiglec-H+) in the spleen or in BM-derived DCs (BM-DCs; Fig. S1 B). However, female DCBlimp-1ko mice developed autoantibodies as early as 4–5 mo of age. Sera from age-matched female DCBlimp-1ko mice and control mice were assayed for anti-nuclear antibody (ANA), anti-double-stranded (ds) DNA, and anti-ENA5 by ELISA. All the ANA-positive immunoglobulin was IgG; IgM ANA was negligible (Fig. 1 B and not depicted). Moreover, sera from DCBlimp-1ko mice displayed IgG reactivity to both dsDNA and ENA5 (Fig. 1 E). Mice haploid for Blimp-1 also displayed IgG reactivity to dsDNA antibodies. An isotype-specific ELISA demonstrated that IgG2b is the major isotype of anti-dsDNA antibodies (Fig. 1 C). Proteinuria was developed, and kidney deposition of IgG and mesangial cell proliferation and inflammatory infiltrates were observed in 10-mo-old DCBlimp-1ko mice (Fig. 1 D). In contrast to the results from female mice, serum from male DCBlimp-1ko mice displayed no anti-dsDNA or ENA5 reactivity (Fig. 1 E). Similarly, total serum immunoglobulin levels and number of splenocytes are significantly increased in female but not in male DCBlimp-1ko mice (Fig. S2 A and Table S1). The gender disparity in DCBlimp-1ko mice is interesting because most lupus mouse models demonstrate a lupus-like phenotype in both genders. Overall, these data suggest that expression of Blimp-1 in DCs plays a critical role in tolerance against self-reactivity. Moreover, haplosufficient expression is not sufficient for immune tolerance, and, most surprisingly, downstream effects of Blimp-1 deficiency are gender specific.
Increased expression of IL-6 in Blimp-1ko DCs
To understand the alterations in Blimp-1ko DCs responsible for the generation of autoantibodies, we analyzed the characteristics of the DCs. Because the expression of IL-6 is regulated by Blimp-1 (Chan et al., 2009) and enhanced expression of IL-6 is associated with SLE, and possibly related to DC activation (Colonna et al., 2006; Jeon et al., 2010), we examined IL-6 production. We observed an increased production of IL-6 by splenic Blimp-1ko DCs compared with control DCs (Fig. 2 A) and by BM-DCs after LPS stimulation (Fig. 2 B). In male mice, however, there was no significant difference in the level of IL-6 produced by either splenic DCs or BM-DCs from control or DCBlimp-1ko mice (Fig. 2, A and B). In fact, production of IL-6 was higher in female than in male control DCs, implicating a sex difference in cytokine production even in wild type DCs. This observation is consistent with data in patients showing an association of an estrogen-sensitive polymorphism of the IL-6 promoter with susceptibility to type I diabetes in women (Kristiansen et al., 2003). The level of expression of several genes also increased in Blimp-1ko DCs as measured by quantitative (q) PCR (Table S2). Bcl-6, a molecule negatively regulated by Blimp-1, was up-regulated in Blimp-1ko DCs. Expression of XBP, which has been demonstrated to be a survival factor for DCs (Iwakoshi et al., 2007), was equivalent in control and Blimp-1ko DCs, supporting our observation that activation not development is affected by Blimp-1 deficiency.
IL-6 overexpression was of particular interest because IL-6 has been shown to affect several B cell functions, including GC formation and antibody secretion by plasma cells (Kopf et al., 1998; Cassese et al., 2003), and to participate in the differentiation of TFH cells (Nurieva et al., 2009). To test the importance of increased IL-6 in autoantibody production in DCBlimp-1ko mice, we generated DCBlimp-1ko mice haploid for IL-6 (IL-6+/− DCBlimp-1ko). DCs from IL-6+/− DCBlimp-1ko mice express the same level of IL-6 as DCs from control mice after LPS stimulation (Fig. 2 C). Immunization of control and IL-6+/− DCBlimp-1ko mice with NP-CGG showed that IL-6+/− DCBlimp-1ko mice mount an antibody response that is indistinguishable from that of control mice (Fig. S3 A), demonstrating that B cells from IL-6+/− DCBlimp-1ko mice are not defective in antibody production or affinity maturation. Although IL-6+/− DCBlimp-1ko mice had a normal antibody response to immunization, they did not develop autoantibodies (Fig. 2 D). These data suggest that the increased expression of IL-6 by DCs contributes to the generation of autoantibodies in female DCBlimp-1ko mice.
Enhanced GC formation in DCBlimp-1ko mice
Because we observed only IgG, and not IgM, autoantibodies, we asked whether the autoantibodies were derived from GC-experienced B cells. We generated hybridomas of splenocytes from 4-mo-old DCBlimp-1ko mice. There were 27 ANA-positive clones from 304 IgG-secreting clones (∼10%). From the 27 clones, 13 and 16 clones were successfully sequenced for heavy and light chain, respectively. Sequence analysis revealed a high incidence of mutation in most clones (Fig. 3 A). Interestingly, 3 out of 13 clones contained an arginine residue acquired by point mutation in the complementary determining region 3 in the heavy chain, which is often seen in high-affinity anti-DNA antibodies. These data suggest that autoantibodies are produced by GC-matured plasma cells.
Flow cytometry and immunohistochemistry (IHC) demonstrated an enhanced GC response in young mice (8–12 wk old); many more spontaneous GC B cells, as well as GCs in spleens, were present in DCBlimp-1ko mice (Fig. 3 B). In contrast to the response in IL-6+/− DCBlimp-1ko mice, there was an enhanced immune response in DCBlimp-1ko mice after NP-CGG immunization with an increased high-affinity anti-NP IgG response (Fig. S3 B). There was an increased number of GC B cells (Fig. S3 C) and an increased number of total GCs (Fig. S3 D) as well as antigen-specific GCs (7/12 λ+ GCs in DCBlimp-1ko mice and 1/3 λ+ GCs in control mice).
Blimp-1ko DCs induce expansion of TFH cells in vivo and in vitro
TFH cells in GCs provide direct help to antigen-specific B cells (Garside et al., 1998). The importance of tight regulation of expression of the costimulatory molecule inducible co-stimulator (ICOS) on CD4+ TFH cells has been demonstrated in studies of lupus-prone sanroque mice (Vinuesa et al., 2005); moreover, ICOS blockade inhibits lupus in NZB/W F1 mice (Hu et al., 2009).
To address whether there were more activated CD4+ T cells in DCBlimp-1ko mice, we measured ICOS expression on T cells. There was increased expression of ICOS in CD4+ T cells from DCBlimp-1ko mice compared with control mice (Fig. S4 A). In addition, the percentage and the number of TFH cells were also increased in DCBlimp-1ko mice (Fig. 4 A and Fig. S4 B).
Several studies have suggested that specific subsets of DCs preferentially induce different helper T cells (Maldonado-López et al., 1999). We reasoned that the increased number of TFH was dependent on DC function in DCBlimp-1ko mice; therefore, we assayed to assess in vitro differentiation of TFH cell with purified DCs. To avoid the possibility of indirect effects on T cells from DCBlimp-1ko mice, we used naive T cells from control mice. Naive T cells were cultured with DCs from either control or DCBlimp-1ko mice in the presence of differentiation factors. Fig. 4 B shows that the generation of TFH cells was highly enhanced in cultures with Blimp-1ko DCs compared with control DCs. We confirmed the phenotype of TFH cells by demonstrating expression of Bcl-6, which is a principle transcription factor for TFH differentiation and is expressed exclusively in TFH. To test whether the increased TFH was a result of the increased level of IL-6 from DCs, DCs from IL-6+/− Blimp-1ko mice were assayed in culture. There was no increased differentiation of TFH in the presence of DCs from IL-6+/− Blimp-1ko mice, suggesting a direct involvement of IL-6 from DCs in TFH differentiation (Fig. 4 B). The importance of IL-6 was further demonstrated in vitro as α–IL-6 neutralizing antibody inhibited the differentiation of T cells co-cultured with control DCs and Blimp-1ko DCs into TFH (Fig. S5). These observations suggest that the increased generation of TFH cells results from increased IL-6 production by DCs in DCBlimp-1ko mice. Interestingly, there was no significant increase in TH17 cells, another subset for which IL-6 is critical (unpublished data), suggesting that additional requirements for the generation of TH17 must exist and are not provided by Blimp-1ko DCs.
IL-6–dependent generation of GC B cells and TFH cells
Because increased TFH cells and GC response are important mechanisms for autoantibody production in DCBlimp-1ko mice and the phenotype was impaired in IL-6+/− DCBlimp-1ko mice, we compared GC and TFH cells in IL-6+/− DCBlimp-1ko mice. IL-6+/− DCBlimp-1flox/+ mice showed a reduced number of TFH cells as well as a reduced number of GC cells in the spleen (Fig. 5). These data suggest that the increased expression of IL-6 in Blimp-1ko DCs is a major molecular mechanism responsible for the expansion of TFH cells and enhanced GC formation, leading to the generation of autoantibodies in DCBlimp-1ko mice.
In summary, we propose a new mechanism for the development of a lupus-like phenotype mediated by Blimp-1, which is required to maintain tolerogenic function in DCs in a gender-dependent manner. The loss of function of Blimp-1 in female DCs results in increased secretion of the critical proinflammatory cytokine IL-6, and in increased differentiation of TFH cells and increased GC responses. It is of considerable interest that a polymorphism of the Blimp-1 gene has now been implicated in both rheumatoid arthritis and SLE and that the phenotype of the DCBlimp-1ko mouse is analogous to human SLE with a female bias, enhanced TFH cells, and increased IL-6 levels (Swaak et al., 1989).
MATERIALS AND METHODS
Blimp-1flox mice were provided by K. Calame (Columbia University, NY, NY) and backcrossed with C57BL/6 for eight generations. CD11c-CRE mice were generated in the Reizis laboratory. DCBlimp-1ko mice and control mice were bred in the animal facility of The Feinstein Institution for Medical Research (FIMR) in specific pathogen-free conditions. IL-6ko mice (The Jackson Laboratory) were bred with DCBlimp-1ko mice in the animal facility of FIMR.
Purification of splenic DCs and in vitro generation of BM-DCs.
CD11c+ splenic DCs were enriched with an EasySep kit (STEMCELL technologies) according to the manufacturer’s protocol. Then, CD11chi Siglec-H− DCs were further purified by cell sorter (FACSAria; BD). Cell purity was routinely >95%.
To generate BM-DCs, BM cells were harvested from the femur with PBS. T cells and B cells were depleted by incubation with antibodies from hybridoma cell lines (American Type Culture Collection; TIB-120, TIB-211, TIB 207, and TIB-146) with rabbit complement (Pel-Freeze Biologicals). The remaining cells were cultured in RPMI 1640 with 10% FCS and 200 ng/ml Flt3L (PeproTech) for 8 d. The nonadherent cells were collected. To measure cytokines in the supernatants, 106 DCs/ml were cultured overnight in medium with or without 1 µg/ml LPS (Sigma-Aldrich).
Anti-dsDNA and anti-ENA5 (Sm, RNP, SS-A, SS-B, and Scl-70) antibodies were measured by a QUANTA Lite ELISA kit (INOVA Diagnostics). Assays were performed as described in the protocol provided by manufacturer. In brief, serum samples were diluted in sample diluent at 1:101 and incubated in antigen-precoated plates for 30 min at room temperature (rt). Horse radish peroxidase (HRP)–conjugated isotype-specific anti-mouse IgG (1:2,000 SouthernBiotech) was added for 30 min at rt. Plates were washed after incubation, and TMB substrate was added for development. Absorbance (OD) was read for each well at 450 nm. L-6 ELISAs were performed with specific cytokine kits according to the manufacturer’s protocol (BD).
Histology of spleen and kidney.
Spleens from 6–10-wk-old DCBlimp-1ko and control mice were fixed with 4% PFA and transferred to a 30% sucrose solution. The fixed spleens were snap frozen in Tissue-Tek O.C.T. compound (Sakura) and sliced to 7 µm. On the day of staining, sections were fixed with ice-cold acetone and blocked with blocking buffer. After blocking, samples were incubated with fluorochrome-conjugated antibodies diluted in dilution buffer for 1 h at room temperature. After incubation with antibodies, slides were washed with PBS three times.
Kidneys were harvested from 8-mo-old mice and fixed with formaldehyde and 70% ethanol. Fixed tissues were paraffin embedded and sliced to 5-µm thickness. Tissue was stained with standard hematoxylin and eosin. Images were visualized using a fluorescence microscope (AxioCam II; Carl Zeiss) and analyzed by OpenLab software (PerkinElmer).
Total RNA was extracted from purified DCs with RNeasy kit (Invitrogen) according to the manufacturer’s instructions and subjected to reverse transcription with iScript cDNA synthesis kit (Bio-Rad Laboratories). cDNA was analyzed by qPCR using LightCycler 480 probes master with various primers (Applied Biosystems). Relative induction of each gene of interest was calculated by ΔΔCt.
TFH cell in vitro differentiation.
105 purified DCs (CD11chi) from control and DCBlimp-1ko mice and 5 × 105 naive T cells (CD4+, CD62Lhi, and CD44lo) were co-cultured with anti-CD3ε antibody precoated (145-2C11: 5 µg/ml) 96-well plates. To induce TFH cell differentiation, TFH medium (10 µg/ml anti-IL-4 [11B11], 10 µg/ml anti–IFN-γ [XMG1.2], 10 µg/ml anti–TGF-β [1D11], 30 ng/ml IL-6 [PeproTech], and 50 ng/ml IL-21 [PeproTech]) was added. PBS alone and anti-CD3ε alone without TFH medium were used as a negative control. T cells were cultured for 4 d and TFH cells were analyzed by flow cytometry as described in Fig. S4 B. In some experiments, anti–IL-6 neutralizing antibody (eBioscience) was added during the culture. CD4+ T cells were purified by anti-CD4 microbeads (Miltenyi Biotec) for Bcl-6 expression.
Information on serum immunoglobulin ELISA, Western blotting, ANA, antibodies, immunization with NP(16)-CGG, proteinuria and IgG deposition in kidney, hybridoma generation, and Ig sequencing are available in the supplemental Materials and methods.
Unpaired two-tailed Student t tests were used for statistical analysis with Prism software (GraphPad Software). P < 0.05 was considered to be significantly different.
Online supplemental material.
Fig. S1 shows Blimp-1 expression in hematopoietic lineages and additional phenotypes of DCBlimp-1ko mice. Fig. S2 shows gender-dependent serology and secretion of IL-6 from B cells. Fig. S3 shows enhanced antibody and GC response in DCBlimp-1ko mice. Fig. S4 shows analysis of CD4+ T cells in the spleen of control and DCBlimp-1ko mice. Fig. S5 shows blocking of TFH by anti–IL-6 antibodies in vitro. Additional information is provided in the supplemental Materials and methods.
We thank K. Calame (Columbia University) for Blimp-1 floxed mice, and P. Gregersen, J. Cohen-Solal, and P.M. Osorio for valuable discussions and critical reading of the manuscript. The animal experimental procedure was approved by Ping Wang, a chairperson of the Feinstein Institute for Medical Research Institutional Animal Care and Use Committee.
This study was supported by National Institutes of Health grant AR047984 and an SLE Foundation Fellowship to S.J. Kim.
The authors have no financial conflicts of interest to declare.
systemic lupus erythematosus