Type I interferondependent CD86high marginal zone precursor B cells are potent T cell costimulators in mice.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 63, No. 4, April 2011, pp 1054–1064 DOI 10.1002/art.30231 © 2011, American College of Rheumatology Type I Interferon–Dependent CD86high Marginal Zone Precursor B Cells Are Potent T Cell Costimulators in Mice John H. Wang,1 Qi Wu,1 PingAr Yang,1 Hao Li,1 Jun Li,1 John D. Mountz,2 and Hui-Chen Hsu1 Objective. To investigate the role of CD86high marginal zone (MZ) precursor B cells in type I interferon (IFN)–induced T cell–dependent responses in autoimmune BXD2 mice. Methods. Confocal microscopic imaging was used to determine the location of plasmacytoid dendritic cells (DCs), MZ precursor B cells, and CD4 T cells in the spleens of BXD2 and BXD2-Ifnarⴚ/ⴚ mice. Immunohistochemical staining was used to determine IgGbright cells in the spleens of BXD2 and BXD2-Ifnarⴚ/ⴚ mice. Enzyme-linked immunosorbent assay was used to determine serum levels of IFN␣ and autoantibodies, and 4-hydroxy-3-nitrophenylacetyl hapten (NP)–chicken ␥-globulin (CGG) (NP-CGG)– or NP-Ficoll–induced anti-NP2 antibody titers. Real-time quantitative polymerase chain reaction was used to determine the levels of type I IFN transcripts. T cell proliferation was measured using 3H-thymidine. The expression of CD86 and CD80 was determined by fluorescence-activated cell sorting analysis. Results. The deletion of type I IFN receptor abrogated the development of IgGbright cells and sup- pressed a T cell–dependent antibody response. Type I IFN signaling was associated with the expression of CD86, but not CD80, on follicular, MZ, and MZ precursor B cells. However, MZ precursor B cells demonstrated the highest expression of CD86 and the highest capacity for T cell costimulation with intact type I IFN receptor. This effect was blocked by an antibody that neutralizes CD86. In IFN receptor–intact BXD2 mouse spleens, MZ precursor B cells clustered at the T cell–B cell border. CD86 deletion suppressed germinal center formation, autoantibody production, and development of autoimmune diseases in BXD2 mice. Conclusion. Type I IFN can promote autoimmune responses in BXD2 mice through up-regulation of CD86high expression on MZ precursor B cells and trafficking of MZ precursor B cells to the T cell–B cell border to provide costimulation of CD4 T cells. Overexpression of type I interferon (IFN)– inducible genes, known as the “type I IFN” signature, has been observed in the peripheral blood of patients with systemic lupus erythematosus (SLE) (1,2). Type I IFN is produced primarily by CD11clow-expressing dendritic cells (DCs) that express the phenotypic markers B220, Gr-1, and a more specific surface marker, plasmacytoid DC antigen 1 (PDCA-1) (3,4). These DCs are known as plasmacytoid DCs (3–6). A T cell–dependent antibody response requires antigen presentation by class II major histocompatibility complex and costimulation via CD80 or CD86 expressed on antigen-presenting cells (7). Studies of human peripheral blood revealed increased expression of CD80 and CD86 on B cells from patients with SLE compared with healthy individuals (8,9). The severity of lupus disease is positively correlated with the levels of CD80 and CD86 expression (9). However, only CD86 expression was significantly increased in lupus patients with renal disease, the hallmark of SLE, while differences Supported by grants to Dr. Mountz from the American College of Rheumatology Research and Education Foundation (Within Our Reach research grant), the Alliance for Lupus Research (Target Identification in Lupus program grant), the Department of Veterans Affairs (Merit Review grant 1I01BX000600-01), DaiichiSankyo Co., Ltd, and the NIH (grants 1-AI-071110-01A1, ARRA-3R01-AI-71110-02S1, and P30-AR-48311) and by grants to Dr. Hsu from the Lupus Research Institute (Novel Research project grant) and the Arthritis Foundation (Arthritis Investigator Award). 1 John H. Wang, PhD, Qi Wu, BS, PingAr Yang, BS, Hao Li, MS, Jun Li, MD, PhD, Hui-Chen Hsu, PhD: University of Alabama at Birmingham; 2John D. Mountz, MD, PhD: University of Alabama at Birmingham and Birmingham VA Medical Center, Birmingham, Alabama. Address correspondence to Hui-Chen Hsu, PhD, Department of Medicine, Division of Clinical Immunology & Rheumatology, University of Alabama at Birmingham, 1825 University Boulevard, SHELB 311, Birmingham, AL 35294. E-mail: email@example.com. Submitted for publication June 17, 2010; accepted in revised form December 28, 2010. 1054 TYPE I IFN IN T CELL–DEPENDENT AUTOANTIBODY RESPONSES in CD80 expression levels were statistically insignificant (10). Other studies have corroborated the importance of CD86 (but not CD80) by demonstrating that only the expression of CD86 on B cells is elevated in patients with inactive SLE, and that the level of CD86 expression is further elevated in conjunction with active disease (11,12). Our group previously demonstrated that BXD2 mice spontaneously produce pathogenic autoantibodies that can induce and exacerbate glomerulonephritis and erosive arthritis (13). Blocking of the interaction of B7 and CD28 in young BXD2 mice, using AdCTLA-4Ig, dramatically suppressed the expression of activationinduced cytidine deaminase, which is the essential enzyme to promote B cell somatic hypermutation and class switch recombination (14). This treatment also prevented the development of both nephritis and arthritis in BXD2 mice (14). Although expression of CD86 was shown to be increased on the B cells of BXD2 mice (14), it has not been specifically determined whether increased expression of CD86 is associated with the pathogenesis of autoimmunity in BXD2 mice. Also, the stage(s) of germinal center (GC) development at which CD86high B cells encounter CD28⫹ CD4 T cells and the mechanisms that are involved in driving the encounter of these cells remain unclear. Recently, we identified a subpopulation of B cells that have surface expression of CD1d high IgM high CD21highCD23high in BXD2 mice; this B cell population was significantly increased in the spleens of BXD2 mice at the expense of reduced marginal zone (MZ) B cell counts (15). This population of CD19⫹ splenocytes is commonly known as MZ precursor B cells (16). The immunopathogenesis of MZ precursor B cells in BXD2 mice was demonstrated by their high-affinity binding to an exogenous antigen, trinitrophenyl (TNP)– Ficoll (15). Importantly, our previous study also showed that high levels of type I IFN produced by plasmacytoid DCs in the marginal sinus play an important role in up-regulating CD69 and facilitating TNP-positive MZ precursor B cell migration to the light zone border of GCs (15). In the current study, we examined the role of type I IFN in regulating the surface expression of the costimulatory molecules CD80 and CD86 on follicular, MZ, and MZ precursor B cells. We also determined whether type I IFN signaling is required for MZ precursor localization at the critical T cell–B cell border before a spontaneous GC response is initiated. The results showed that type I IFN–induced up-regulation of CD86 on MZ precursor B cells and the direction of MZ 1055 precursor localization to the T cell–B cell border are important in promoting an IgG antibody response and autoimmune disease. MATERIALS AND METHODS Mice. Female homozygous C57BL/6J (B6) mice, BXD2 recombinant inbred mice, and B6-Cd86ⴚ/ⴚ mice were obtained from The Jackson Laboratory; B6-Ifnar⫺/⫺ mice were obtained from Dr. Jocelyn Demengeot (Instituto Gulbenkian de Ciência, Oeiras, Portugal). BXD2-Ifnar⫺/⫺ and BXD2Cd86⫺/⫺ mice were generated by backcrossing B6-Ifnar⫺/⫺ and B6-Cd86⫺/⫺ mice with BXD2 mice for 7 generations. All mice were housed in the University of Alabama at Birmingham (UAB) Mouse Facility under specific pathogen–free conditions, and all procedures were approved by the UAB Institutional Animal Care and Use Committee. Unless specified otherwise, all mice were killed at 8–12 weeks of age. In vivo treatments and enzyme-linked immunosorbent assays (ELISAs). To induce the production of IFN␣, 5 g of CpG-A oligonucleotide was dissolved with 30 l of DOTAP Liposomal Transfection Reagent (Hoffman-La Roche) in 120 l of phosphate buffered saline (PBS). The CpG-A–DOTAP mixture was intravenously injected into mice. An IFN␣ ELISA kit (PBL Biomedical) was used to assay serum levels of IFN␣. Immunizations in mice were carried out by intraperitoneal injection of 50 g of 4-hydroxy-3-nitrophenylacetyl hapten (NP)–chicken ␥-globulin (CGG) (NP-CGG; Biosearch Technologies) adsorbed onto 1.3 mg of alum (Sigma-Aldrich) in a total volume of 100 l of PBS or 50 g of NP-Ficoll in PBS. Antibodies to NP2–bovine serum albumin (BSA) were determined by ELISA, using horseradish peroxidase (HRP)– linked anti-mouse IgM (Southern Biotechnology), anti-mouse IgG2b (Southern Biotechnology), and anti-mouse IgG2c (Southern Biotechnology) antibodies, as previously described (17). Urinary albumin was analyzed using the competitive Albuwell M ELISA kit (Exocell) according to the manufacturer’s protocol (13). Flow cytometric analysis. For the analysis of costimulatory molecules, phycoerythrin (PE)–conjugated anti-CD80 or PE-conjugated anti-CD86 antibodies were used. For the analysis of B cell subpopulations, Alexa 700–conjugated antiCD19, peridinin chlorophyll A protein–Cy5.5–conjugated antiCD93/AA4, PE-Cy7–conjugated anti-IgM (Southern Biotechnology), fluorescein isothiocyanate–conjugated anti-CD21, and allophycocyanin (APC)–conjugated anti-CD23 antibodies were used. All antibodies were obtained from BioLegend, except where indicated otherwise. GC B cells were analyzed using PE-conjugated anti-Fas (BioLegend) and biotinconjugated peanut agglutinin (PNA) (Vector) followed by APC-conjugated streptavidin (BioLegend). All cells were fixed in 1% paraformaldehyde/fluorescence-activated cell sorting (FACS) solution before analysis by flow cytometry using a BD LSR II flow cytometer (BD Biosciences). The analysis was performed using FlowJo software (Tree Star). Forward-angle light scattering was used to exclude dead and aggregated cells. For analyses of follicular, MZ, and MZ precursor B cells, CD19⫹ splenocytes were first gated. For follicular, MZ, and MZ precursor B cells, AA4 ⫺ IgM low IgD high CD21 low CD23 high , AA4 ⫺ IgM high CD21 high CD23 low , and 1056 WANG ET AL Figure 1. Activated plasmacytoid dendritic cells (PDCs) and type I interferon (IFN) in BXD2 mice. A, Representative images of spleen sections obtained from BXD2 and C57BL/6J (B6) mice, stained with anti–plasmacytoid DC antigen 1 (anti–PDCA-1; green) and MAdCAM-1 (red) are shown. B, A representative spleen section from a BXD2 mouse, with peanut agglutinin (PNA)–binding positive germinal center (GC; blue) inside follicles demarcated with MAdCAM-1 (red) and surrounded by layers of plasmacytoid DCs (green) is shown. C, CpG-A–DOTAP was injected intravenously into BXD2 and B6 mice as described in Materials and Methods. Serum IFN␣ levels were determined by enzyme-linked immunosorbent assay (n ⫽ 3–6). ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001 versus B6 mice. D, RNA was extracted from peripheral blood cells. Transcripts of Ifna1, Ifna4, Ifna11, and Ifnb isoforms from BXD2 and B6 mice in different age groups were quantitated relative to copy counts of Actin, as described in Materials and Methods (n ⫽ 6). For BXD2 mice, ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001 versus mice ages 3–6 months. For B6 mice, ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001 versus age group–matched BXD2 mice. Values in C and D are the mean ⫾ SEM. Original magnification ⫻ 40 in A; ⫻ 20 in B. AA4⫺IgMhighCD21highCD23high splenocytes, respectively, were gated using the method described by Allman and Pillai (18) (additional information is available from the corresponding author). For determination of the dose response to IFN␣, follicular, MZ, and MZ precursor B cells sorted by FACS were cultured with increasing doses of IFN␣ (0, 200, 400, and 800 units/ml; PBL Biomedical) for 12 hours prior to flow cytometric analysis for CD80⫹ or CD86⫹ cells. Histologic analysis and confocal microscopic imaging. Confocal microscopic imaging of frozen sections was used for the detection of CD1d, CD23, CD4, IgM, and PNA binding (15). For analysis of the plasmacytoid DC and marginal sinus locations, frozen spleen sections were blocked with 10% normal rat serum and stained with anti–PDCA-1 (rat IgG2b; Miltenyi Biotec), followed by Alexa 555–conjugated anti-rat IgG (Invitrogen) and Alexa 488–labeled anti–MAdCAM-1 (BioLegend). All tissue sections were mounted in Fluormount-G (Southern Biotechnology) and viewed with a Leica DM IRBE inverted Nomarski/epifluorescence microscope outfitted with Leica TCS NT laser confocal optics. For analysis of IgM or IgG deposition in the spleens and IgG deposition in the kidneys, immunohistochemical analysis was performed on 10% formalin–fixed spleen tissue sections (4 M). Endogenous peroxidase quenching was performed by the addition of 3% H2O2 followed by 0.25% pepsin antigen retrieval. Tissues were subsequently blocked with 1.5% BSA, followed by incubation with HRP-linked antimouse IgM and/or anti-mouse IgG (both from Southern Biotechnology) and treated with 3,3⬘,5,5⬘-tetramethylbenzidine (Sigma). Tissue sections were imaged using an Olympus BX41 microscope. Quantitation of IgG-positive glomeruli was carried out using the ImageJ program (NIH Image, National Institutes of Health; online at: http:// rsbweb.nih.gov/ij/). For each image, the background intensity was subtracted. Costimulatory proliferation assay. B cells from the mouse spleens were enriched by positive selection using antiCD19 magnetic microbeads (Miltenyi Biotech). Subsequently, CD19⫹ B cells were sorted by flow cytometry into follicular, MZ, and MZ precursor B cells according to the combinations of surface markers as determined by FACS. For IFN␣ stimulation, sorted B cells were incubated with IFN␣ (400 units/ml; PBL Biomedical) for 12 hours and washed with culture medium twice prior to irradiation. For costimulation, irradiated (3,000 rads) follicular, MZ, or MZ precursor B cells (5 ⫻ 105) were cocultured with 5 ⫻ 105 magnetically activated cell– sorted CD4 T cells derived from the spleens of BXD2 mice at a 1:1 ratio. Cells were cultured for 48 hours in triplicate round-bottomed wells in 96-well plates (Costar) in the presence of anti-CD3 (2.5 g/ml; clone 145-2C11) plus anti-CD86 (10 g/ml; clone GL-1), anti-CD80 (10 g/ml; clone 16-10A1), or rat IgG2a isotype control (10 g/ml; clone RTK2758). TYPE I IFN IN T CELL–DEPENDENT AUTOANTIBODY RESPONSES The proliferative response was measured by a standard 3Hthymidine incorporation assay (17) in which 1 Ci of 3Hthymidine (PerkinElmer) was added to each well during the last 12 hours of culture. RNA quantitation. The expression of type I IFN isoforms Ifna1, Ifna4, Ifna11, and Ifnb in peripheral blood was determined using a real-time quantitative polymerase chain reaction (qPCR) method, as previously described (15,17). The real-time qPCR mixtures contained SYBR Green PCR Master Mix (Bio-Rad) with the following primers: for Ifna1, forward AGTGAGCTGACCCAGCAGAT, reverse GGTGGAGGTCATTGCAGAAT; for Ifna4, forward TCTGCAATGACCTCCATCAG, reverse TATGTCCTCACAGCCAGCAG; for Ifna11, forward CCCAGCAGATCTTGAACCTC, reverse GGTGGAGGTCATTGCAGAAT; for Ifnb, forward CTCCACCACAGCCCTCTC, reverse CATCTTCTCCGTCATCTCCATAG; for Actin, forward CGTTGACATCCGTA, reverse GGAAGGTGGACAGTGAGG. Statistical analysis. All results are shown as the mean ⫾ SEM. A 2-tailed t-test was used when 2 groups were compared, and analysis of variance was used when ⬎2 groups were compared. P values less than 0.05 were considered significant. RESULTS Increased type I IFN/plasmacytoid DCs in lupusprone BXD2 mice. Clustering of plasmacytoid DCs was increased in the MZ and outside the follicle, the boundary of which is demarcated by the marginal sinus (19), in BXD2 mice compared with B6 mice (Figure 1A). We also observed developing GCs in BXD2 mice that were located inside the follicle, with clusters of these plasmacytoid DCs at the MZ (Figure 1B). Intravenous treatment with CpG-A, a known Toll-like receptor 9 (TLR-9) ligand, in BXD2 mice dramatically increased serum IFN␣ levels in the first 2 hours; these levels peaked at 4 hours and dramatically declined at 12 hours (Figure 1C). Although the changes in IFN␣ were comparable in similarly treated B6 mice, the levels of IFN␣ in B6 mice were significantly lower than those in BXD2 mice (Figure 1C). Peripheral blood cells obtained from untreated BXD2 and B6 mice at different ages were assayed for various type I IFN isoform transcripts (Figure 1D). In BXD2 mice aged ⬍2 months, 9–12 months, and ⬎1 year, transcript levels of Ifna1, Ifna4, and Ifna11 were significantly lower than those in mice aged 3–6 months (Figure 1D). With the exception of Ifnb, all type I IFN transcripts were significantly lower in B6 mice compared with age-matched BXD2 mice in the groups aged ⬍2 months and 3–6 months (Figure 1D). No significant differences were detected between older mice (age ⬎9 months). 1057 Type I IFN–promoted IgG class switch. Histologic analysis of spleen tissues at 20⫻ magnification and 40⫻ magnification for the identification of IgMbright and IgGbright large B cells was carried out to verify that spleens from BXD2, BXD2-Ifnarⴚ/ⴚ, and B6 mice all contained IgMbright large B cells (Figures 2A and B), but significant numbers of IgGbright large B cells were generated only by BXD2 mice. In contrast, IgGbright cells were virtually undetectable in BXD2-Ifnarⴚ/ⴚ mice, similar to B6 mice (Figures 2A and B). These results suggest that type I IFN acts mainly through a GCdependent mechanism to promote autoimmunity in BXD2 mice. Promotion of a T cell–dependent antibody response by type I IFN. To determine whether type I IFN promotes mainly a T cell–dependent or a T cell– independent antibody response, B6, BXD2, and BXD2Ifnarⴚ/ⴚ mice were immunized with either a known T cell–dependent antigen (NP-CGG) or a known T cell– independent antigen (NP-Ficoll) and intravenously treated with either placebo or CpG-A 17 hours later (Figure 3). Immunization with NP-CGG induced dramatically higher levels of high-affinity IgG2b and IgG2c anti-NP2 antibody isotypes in BXD2 mice compared with normal B6 mice (Figures 3B and C), while IgM titers remained low (Figure 3A). CpG-A treatment further significantly increased high-affinity IgG isotype anti-NP2 titers in BXD2 mice compared with untreated NP-CGG–immunized controls (Figures 3B and C). Deletion of type I IFN receptor in BXD2 mice suppressed IgG2b and IgG2c titers (Figures 3B and C), regardless of treatment with CpG-A, indicating that intact type I IFN receptor is required to complete a T cell–dependent antibody response. NP-Ficoll did not induce dramatically increased IgG titers (Figures 3B and C) but did generate substantial IgM titers (Figure 3A). In the presence of intact IFN receptor or CpG-A stimulation, type I IFN signaling did not significantly class switch IgM to IgG in response to NP-Ficoll challenge (Figure 3A). These results suggest that type I IFN promotes a T cell–dependent antibody response, and that type I IFN may play a significant role at the pre-GC stage, prior to antibody maturation and plasma cell differentiation. Type I IFN promotes MZ precursor B cell– induced T cell costimulation and positioning. We previously showed that the follicle-oriented migration response of MZ precursor B cells into the light zone border of well-formed GCs is highly associated with type I IFN (15). In the current study, we investigated the location of MZ precursor B cells at the pre-GC stage. Histologic analysis of spleen sections from BXD2 mice 1058 WANG ET AL Figure 2. Type I interferon receptor deletion abrogates IgG class switch in 4-month-old BXD2 mice. Spleen sections from BXD2, BXD2-Ifnarⴚ/ⴚ, and C57BL/6J (B6) mice were stained with anti-IgM and anti-IgG antibodies as described in Materials and Methods. A, Representative images of spleen sections showing the presence of follicles and IgMbright or IgGbright B cells. B, High-magnification view of spleen sections, showing the presence of a follicle (FO) whose boundary is denoted by broken lines. demonstrated that MZ precursor B cells formed clusters adjacent to CD4 T cells at the T cell–B cell boundary (Figure 4A, top). The significant presence of clustering MZ precursor B cells adjacent to CD4 T cells in BXD2 mice contrasted dramatically with the poor MZ precursor B cell clustering that was observed in BXD2-Ifnarⴚ/ⴚ mice at the T cell–B cell boundary (Figure 4A, bottom). Instead, in BXD2-Ifnarⴚ/ⴚ mice, the pool of B cells Figure 3. Type I interferon receptor deletion abrogates the T cell–dependent antibody response. C57BL/6J (B6), BXD2, and BXD2-Ifnarⴚ/ⴚ mice were immunized with either 4-hydroxy-3-nitrophenylacetyl hapten (NP)–chicken ␥-globulin (CGG) (NP-CGG) or NP-Ficoll. Seventeen hours later, the mice were treated with placebo or CpG-A, as described in Materials and Methods. Sera were collected on day 17 after immunization. Levels of IgM (A), IgG2b (B), and IgG2c (C) were determined by enzyme-linked immunosorbent assay. Bars show the mean ⫾ SEM results from 6 mice per group. For untreated (control) BXD2 mice, ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01 versus CpG-A–treated mice. For B6 or BXD2-Ifnarⴚ/ⴚ mice, ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001 versus BXD2 mice. OD450 ⫽ optical density at 450 nm. TYPE I IFN IN T CELL–DEPENDENT AUTOANTIBODY RESPONSES 1059 Figure 4. CD4 T cell costimulatory capability of each of the B cell subpopulations. A, Left, Histologic analysis of follicular (CD23⫹CD1d⫺; red), marginal zone (CD23⫺CD1d⫹; green), and marginal zone precursor (MZ-P) (CD23⫹CD1d⫹; yellow) B cells (B), and CD4 T cells (T) (blue) located on spleen tissue sections from representative BXD2 and BXD2-Ifnar⫺/⫺ mice. Original magnification ⫻ 40. Right, Higher-magnification views of the corresponding boxed areas (left). The broken white line indicates the junction at the T cell–B cell border. B, Proliferative response as measured by standard 3H-thymidine incorporation assay of irradiated MZ precursor, MZ, and follicular (FO) B cells from BXD2 and BXD2-Ifnar⫺/⫺ mice cocultured with CD4 T cells and an agonistic anti-CD3 antibody for 48 hours. Bars show the mean ⫾ SEM results from 3 independent experiments (n ⫽ 2 mice per experiment). C and D, Representative results of flow cytometric analyses indicating CD86 (C) and CD80 (D) expression on splenic MZ precursor, MZ, and follicular B cells from BXD2 and BXD2-Ifnar⫺/⫺ mice. The values shown are the mean ⫾ SEM of CD86⫹ or CD80⫹ B cell subpopulations in BXD2 (numerator) and BXD2-Ifnar⫺/⫺ (denominator) mice. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001, BXD2-Ifnar⫺/⫺ mice versus BXD2 mice, or as indicated. clustered at the T cell–B cell border was biased toward a predominance of follicular B cells (Figure 4A, bottom). To test the CD4 T cell costimulatory capability of each of the B cell subpopulations, we sorted and cocultured irradiated B cell subpopulations from BXD2 and BXD2-Ifnarⴚ/ⴚ mice with anti-CD3–stimulated CD4 T cells. Consistent with the literature (20), MZ B cells were better costimulatory agents than were follicular B cells (Figure 4B). MZ precursor B cells from BXD2 mice, however, provided the greatest costimulation of CD4 T cells, compared with the costimulation provided by MZ or follicular B cells (Figure 4B). MZ precursor B cells from BXD2-Ifnarⴚ/ⴚ mice demonstrated significantly attenuated T cell costimulation compared with MZ precursor B cells from BXD2 mice (Figure 4B). MZ B cells also displayed attenuated costimulatory capability when type I IFN receptor was deleted (Figure 4B). Consistent with these results, the percentage of CD86high MZ precursor B cells from BXD2 mice was increased significantly compared with the percentage of MZ and follicular B cells (Figure 4C). BXD2-Ifnarⴚ/ⴚ mice exhibited an ⬃50% decrease in the expression of CD86high by all B cell subpopulations (Figure 4C). Interestingly, MZ precursor B cells from BXD2-Ifnarⴚ/ⴚ mice maintained higher basal CD86 surface expression than did their MZ and follicular B cell counterparts (Figure 4C). Both MZ and MZ precursor B cells maintained higher CD80 expression than did follicular B cells (Figure 4D). However, CD80 expression in all 3 subpopulations of B cells was not significantly different between BXD2 mice and BXD2-Ifnarⴚ/ⴚ mice (Figure 4D). Type I IFN stimulates the costimulatory function of MZ precursor B cells via the induction of CD86. To test whether IFN␣ has a direct effect on the surface expression of CD86, splenic MZ precursor, MZ, and follicular B cells from BXD2 mice and BXD2-Ifnar⫺/⫺ mice were sorted and either were left unstimulated or were stimulated with IFN␣. Surface expression of CD86 on all 3 B cell subpopulations was determined 12 hours after stimulation with IFN␣ (Figure 5A). The highest expression of CD86 in response to IFN␣ was observed on MZ precursor B cells compared with either follicular or MZ B cells. Deletion of type I IFN receptor resulted in significant blunting of CD86 expression, especially on MZ precursor B cells (Figure 5A). In contrast, CD80 surface expression on follicular, MZ, or MZ precursor B cells was not dependent on IFN␣ signaling (Figure 5A). Stimulation of follicular, MZ, and MZ precursor B cells with increasing doses of IFN␣ (0, 200, 400, and 1060 WANG ET AL Figure 5. Suppression of interferon-␣ (IFN␣)–induced costimulatory function of MZ precursor B cells by CD86 neutralization. A, Representative histograms showing the cell surface expression of CD86 and CD80 in the indicated population of B cells. MZ precursor, MZ, and follicular B cells from BXD2 and BXD2-Ifnarⴚ/ⴚ mice were sorted by fluorescence-activated cell sorting and stimulated with medium or IFN␣ (400 units/ml), as described in Materials and Methods. B, IFN␣-induced dosedependent induction of the expression of CD86 and CD80 on sorted MZ precursor, MZ, and follicular B cells. C, Proliferative response as measured by standard 3 H-thymidine incorporation assay of irradiated MZ precursor, MZ, and follicular B cells from BXD2 and BXD2-Ifnarⴚ/ⴚ mice cocultured with CD4 T cells and an agonistic anti-CD3 antibody for 48 hours. Sorted B cells were left unstimulated or stimulated with IFN␣, as described in Materials and Methods. Cells were cultured in the presence of rat IgG2a isotype control (10 g/ml), anti-CD86 (10 g/ml), and anti-CD80 (10 g/ml). Results are the mean ⫾ SEM values from 3 independent experiments (n ⫽ 2 mice per experiment). ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001. WT ⫽ wild-type; KO ⫽ knockout (see Figure 4 for other definitions). 800 units/ml) demonstrated a dose-related up-regulation of CD86 surface expression on follicular, MZ, and MZ precursor B cells but most significantly on MZ precursor B cells (Figure 5B, left), whereas no up-regulation of CD80 surface expression was observed in all 3 B cell subpopulations (Figure 5B, right). To test whether type I IFN promotes the CD4 T cell costimulatory capability of each of the B cell subpopulations via increased expression of CD86, sorted B cells were stimulated with IFN␣ and then cocultured with anti-CD3–stimulated CD4 T cells, in the presence of rat IgG2a isotype control, anti-CD86, or anti-CD80. IFN␣-prestimulated MZ precursor B cells exhibited the most dramatic effect in promoting an anti-CD3–induced T cell proliferative response (Figure 5C); a lesser response was observed in MZ and follicular B cells (Figure 5C). Anti-CD86 antibody effectively blocked T cell costimulation by MZ precursor B cells, either with or without IFN␣ prestimulation (Figure 5C). In contrast, anti-CD80 did not significantly block T cell costimulation provided by MZ precursor B cells (Figure 5C). Role of CD86 in autoimmune disease in BXD2 mice. To demonstrate that CD86 is a key pathogenic molecule in the induction of autoimmune disease, the phenotype of BXD2-Cd86ⴚ/ⴚ mice was determined. CD86 deletion significantly reduced spleen weight and the cell count (Figure 6A). The expression of IgG autoantibodies to histone, BiP, and DNA was significantly reduced in BXD2-Cd86ⴚ/ⴚ mice compared with that in BXD2 mice (Figure 6B). Histologic analysis confirmed that the generation of IgGbright cells in the spleen was dramatically reduced in BXD2-Cd86ⴚ/ⴚ mice (Figure 6C). Flow cytometric and histologic analyses showed a significantly reduced percentage of PNApositive GC B cells in BXD2-Cd86ⴚ/ⴚ mice compared with BXD2 mice (Figure 6D). Consistent with these findings, mesangial proliferation was diminished dramatically, the deposition of IgG-containing immune TYPE I IFN IN T CELL–DEPENDENT AUTOANTIBODY RESPONSES 1061 Figure 6. Critical role of CD86 for autoantibody and germinal center (GC) formation in BXD2 mice. Spleens from 10-month-old BXD2 mice and BXD2-Cd86ⴚ/ⴚ mice were harvested. A, Spleen weights and total cell counts, as determined with a hemocytometer, in BXD2 mice and BXD2Cd86ⴚ/ⴚ mice. ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001 versus BXD2-Cd86⫺/⫺ mice. B, Serum titers of autoantibodies to histone, BiP, and DNA in B6, BXD2, and BXD2-Cd86⫺/⫺ mice. ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001 versus BXD2 mice. C, Histologic analysis of representative anti-IgG antibody– stained spleen tissue sections obtained from BXD2 mice and BXD2-Cd86ⴚ/ⴚ mice. D, Left, Flow cytometric analysis of splenic GC B cells (gated on CD19⫹ B cells first) obtained from BXD2 mice and BXD2-Cd86ⴚ/ⴚ mice. Boxed areas show the percentages of Fas-positive, peanut agglutinin (PNA)–binding positive cells. Right, Histologic analysis of the corresponding spleen sections. ⴱⴱⴱ ⫽ P ⬍ 0.001 versus BXD2 mice. OD450 ⫽ optical density at 450 nm. Values in A, B, and D are the mean ⫾ SEM results from 6 mice per group. complexes in the glomeruli of the kidney was diminished, and the urinary albumin level was significantly reduced in BXD2-Cd86ⴚ/ⴚ mice at 10 months of age compared with that in age-matched BXD2 mice (additional information is available from the corresponding author). In BXD2-Cd86ⴚ/ⴚ mice, the development of erosive arthritis decreased dramatically, as demonstrated by significantly lower levels of inflammatory infiltration, synovial hyperplasia, and marginal erosion, compared with the arthritis that developed in wild-type BXD2 mice (additional information is available from the corresponding author). DISCUSSION Type I IFN is known to be required for autoantibody production (21). Deletion of type I IFN receptor eliminated the development of autoantibodies in various murine models of lupus (22–24). Although the results of studies conducted in nonautoimmune hosts have sug- gested pathways by which type I IFNs drive antibody responses in lupus (21), the precise mechanism by which type I IFNs generate autoantibodies is not clear. The dominant paradigm is that type I IFNs generate autoantibodies in lupus by directly promoting the maturation of autoreactive B cells into antibody-secreting plasmablasts (4,25). Although this paradigm may hold true, our studies suggest additional or alternative mechanisms by which type I IFNs generate plasmablasts and/or plasma cells by promoting T cell–dependent autoantibody responses. We observed increased numbers of plasmacytoid DCs located in clustering layers of cells in the MZ outside the follicle, which is demarcated by MAdCAM-1. Consistent with a noninflammatory phenotype, plasmacytoid DCs in B6 mice exhibited significantly less clustering in the MZ and a more diffuse distribution inside the follicle. However, in BXD2 mice, clustering plasmacytoid DCs were observed in the MZ, 1062 consistent with an inflammatory environment in which immune complexes carrying TLR-binding deoxynucleoproteins may trigger plasmacytoid DC–induced type I IFNs and plasmacytoid DC clustering in the MZ (26). Because plasmacytoid DCs are also systemically circulated between the peripheral blood and the secondary lymphoid organs (27), significant amounts of type IFN transcripts were also observed in the peripheral blood of BXD2 mice. Treatment with CpG-A further confirmed that BXD2 mice have the potential to generate high serum levels of IFN␣. Previously, we demonstrated that the spleens of BXD2 mice have increased counts of type I IFN– producing plasmacytoid DCs compared with normal B6 mice (15). The messenger RNA (mRNA) levels of most Ifna mouse species increase until 6 months of age, at which time the levels peak and then subsequently decline. Interestingly, the increase and peak in mRNA levels were similar to the age-related progression of certain autoantibodies (13), consistent with our findings that type I IFNs can modulate CD4 T cell costimulation and induce antibody responses. Declining type I IFN message levels in older BXD2 mice (⬎9 months) may be attributable to plasmacytoid DCs senescing/maturing in response to antigen exposure (28) and in response to an increasing antigen–autoantibody load (13). We observed that intact type I IFN signaling is required for the formation of IgGbright large plasmablasts or plasma B cells. Deletion of type I IFN receptor in BXD2 mice eliminated the presence of IgGbright B cells while maintaining the presence of IgMbright B cells; this finding would not be expected if type I IFNs were responsible only for the maturation of antibodyproducing B cells into plasmablasts. Immunization with a T cell–dependent antigen but not a T cell–independent antigen showed that type I IFN can produce a highaffinity IgG antibody response. Also, the T cell– dependent response was significantly dependent on the induction of IFN␣ by CpG-A. Deletion of type I IFN receptor, in spite of CpG-A treatment, suppressed IgG titers to baseline levels. Our findings suggest that type I IFN can promote the generation of plasmablasts or plasma B cells in addition to their effects to induce post-GC B cell maturation. Previous studies showed that costimulation of CD4 T cells that is blocked, for instance, by CTLA-4Ig (which competitively binds to either CD80 or CD86) anergizes CD4 T cells (29). The finding that B cells from patients with SLE display high levels of costimulatory receptors, especially CD86, corroborates the assertion that the T cell–dependent humoral response plays a WANG ET AL prominent role in generating autoantibodies (8–12). BXD2 mice display an autoantibody response that is significantly dependent on costimulation signaling brought about by CD80/86–CD28 interactions, because AdCTLA-4Ig treatment in BXD2 mice almost completely blocked the expression of the gene encoding activation-induced cytidine deaminase and suppressed the production of autoantibodies (14). As demonstrated in the current study, CD86 expression is especially important in the autoimmune phenotype of BXD2 mice. Deletion of CD86 significantly inhibited GC formation, suppressed the presence of IgGbright class-switched cells, attenuated autoantibody titers, and obviated the natural course of glomerulonephritis and erosive arthritis in BXD2 mice. Secondary lymphoid organs, such as the spleen, comprise a heterogeneous assortment of B cells (30–32). Most of these consist of mature follicular B cells and a smaller population of MZ B cells, with a series of “transitional” B cells or “intermediates,” including MZ precursor B cells (32). We previously showed that this follicle-oriented migration of MZ precursor B cells can be regulated via type I IFN–induced CD69-dependent down-regulation of sphingosine-1-phosphate receptor 1, and that MZ precursor B cells can transport antigens to the light zone end of existing GCs (15). In the present study, we further demonstrated that MZ precursor B cells exhibit close physical contact with CD4 T cells at the T cell–B cell border during the pre-GC stage. The deletion of type I IFN receptor redistributes MZ precursor B cells away from this T cell–B cell border and toward the MZ. This result, together with our previous findings (15), suggests that type I IFN plays an important role in promoting the follicle-oriented migration of MZ precursor B cells both before and after GC formation has occurred in BXD2 mice. The present study further showed that an important function of type I IFN–promoted MZ precursor B cells at the T cell–B cell border is the provision of costimulation to CD4 T cells. MZ B cells have been reported to express higher levels of CD86 and higher costimulation of CD4 T cells compared with follicular B cells (20). We observed that MZ precursor B cells have higher CD86 expression compared with either follicular or MZ B cells, and that MZ precursor B cells provide even stronger costimulation to CD4 T cells than do MZ B cells. Significantly, CD86 expression on B cells is under the regulation of type I IFN, while expression of CD80 is not. Disruption of type I IFN signaling lowered CD86 expression on follicular, MZ, and MZ precursor B cells. Although MZ precursor B cells retained higher TYPE I IFN IN T CELL–DEPENDENT AUTOANTIBODY RESPONSES levels of CD86 in spite of disrupted type I IFN signaling compared with MZ or follicular B cells, diminished CD86 expression via the deletion of type I IFN receptor significantly decreased MZ precursor B cell–induced T cell costimulation to levels comparable with those induced by follicular or MZ B cells. These results suggest that there is a critical CD86 expression threshold below which T cell costimulation becomes ineffective, and that the signal provided by type I IFN is important to elevate the expression of CD86 to levels above this threshold. Anti-CD28 costimulation normalized the costimulatory effects of MZ precursor B cells, with or without IFN␣ prestimulation, on CD4 T cell proliferation (data not shown), suggesting that IFN␣ acts mainly through the CD86 pathway to enhance the costimulatory function of MZ precursor B cells. Interestingly, CD80 expression is not under the regulation of type I IFN and appears not to offer as strong a costimulatory molecule compared with CD86, at least in BXD2 mice. Our observation that these highly potent CD86high MZ precursor B cell costimulators cluster at the splenic T cell–B cell border in BXD2 mice suggests a geographic advantage of providing costimulation in the very initial stages of GC formation. In summary, the results of this study suggest that type I IFN can offer a new avenue by which T cell– dependent antibody responses are generated. First, type I IFN is critical in bringing MZ precursor B cells to the T cell–B cell border into the follicular interior at the pre-GC stage. Second, type I IFN significantly increases the levels of CD86 on MZ precursor B cells, providing potent costimulation to activated CD4 T cells. This dual-level function of type I IFN prior to the formation of GCs offers new insights into how type I IFNs can induce T cell–dependent autoreactive antibody responses in lupus-prone BXD2 mice and ultimately, perhaps, in human SLE. ACKNOWLEDGMENTS We thank Ms Enid Keyser at the UAB Arthritis and Musculoskeletal Disease Center Analytic and Preparative Cytometry Facility and Mr. Marion L. Spell at the UAB Center for AIDS FACS Core for operating the FACS instrument. Confocal imaging was carried out at the UAB Arthritis and Musculoskeletal Disease Center High Resolution Imaging Facility. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Hsu had full access to all of the 1063 data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Wang, Mountz, Hsu. Acquisition of data. Wang, Wu, Yang, H. Li, J. Li, Mountz, Hsu. Analysis and interpretation of data. Wang, Wu, Yang, H. Li, J. Li, Mountz, Hsu. REFERENCES 1. Bengtsson AA, Sturfelt G, Truedsson L, Blomberg J, Alm G, Vallin H, et al. Activation of type I interferon system in systemic lupus erythematosus correlates with disease activity but not with antiretroviral antibodies. Lupus 2000;9:664–71. 2. Bennett L, Palucka AK, Arce E, Cantrell V, Borvak J, Banchereau J, et al. Interferon and granulopoiesis signatures in systemic lupus erythematosus blood. J Exp Med 2003;197:711–23. 3. Nakano H, Yanagita M, Gunn MD. CD11c⫹B220⫹Gr-1⫹ cells in mouse lymph nodes and spleen display characteristics of plasmacytoid dendritic cells. J Exp Med 2001;194:1171–8. 4. Colonna M, Trinchieri G, Liu YJ. Plasmacytoid dendritic cells in immunity. Nat Immunol 2004;5:1219–26. 5. Asselin-Paturel C, Brizard G, Pin JJ, Briere F, Trinchieri G. Mouse strain differences in plasmacytoid dendritic cell frequency and function revealed by a novel monoclonal antibody. J Immunol 2003;171:6466–77. 6. Siegal FP, Kadowaki N, Shodell M, Fitzgerald-Bocarsly PA, Shah K, Ho S, et al. The nature of the principal type 1 interferonproducing cells in human blood. Science 1999;284:1835–7. 7. Wang S, Chen L. T lymphocyte co-signaling pathways of the B7-CD28 family. Cell Mol Immunol 2004;1:37–42. 8. Wong CK, Lit LC, Tam LS, Li EK, Lam CW. Aberrant production of soluble costimulatory molecules CTLA-4, CD28, CD80 and CD86 in patients with systemic lupus erythematosus. Rheumatology (Oxford) 2005;44:989–94. 9. Dolff S, Wilde B, Patschan S, Durig J, Specker C, Philipp T, et al. Peripheral circulating activated b-cell populations are associated with nephritis and disease activity in patients with systemic lupus erythematosus. Scand J Immunol 2007;66:584–90. 10. Liu W, Li J. Expression of CD80/CD86 and CTLA-4 mRNA in peripheral blood mononuclear cells of the patients with systemic lupus erythematosus. J Huazhong Univ Sci Technolog Med Sci 2004;24:247–9. 11. Bijl M, Horst G, Limburg PC, Kallenberg CG. Expression of costimulatory molecules on peripheral blood lymphocytes of patients with systemic lupus erythematosus. Ann Rheum Dis 2001; 60:523–6. 12. Folzenlogen D, Hofer MF, Leung DY, Freed JH, Newell MK. Analysis of CD80 and CD86 expression on peripheral blood B lymphocytes reveals increased expression of CD86 in lupus patients. Clin Immunol Immunopathol 1997;83:199–204. 13. Hsu HC, Zhou T, Kim H, Barnes S, Yang P, Wu Q, et al. Production of a novel class of polyreactive pathogenic autoantibodies in BXD2 mice causes glomerulonephritis and arthritis. Arthritis Rheum 2006;54:343–55. 14. Hsu HC, Wu Y, Yang P, Wu Q, Job G, Chen J, et al. Overexpression of activation-induced cytidine deaminase in B cells is associated with production of highly pathogenic autoantibodies. J Immunol 2007;178:5357–65. 15. Wang JH, Li J, Wu Q, Yang P, Pawar RD, Xie S, et al. Marginal zone precursor B cells as cellular agents for type I IFN-promoted antigen transport in autoimmunity. J Immunol 2010;184:442–51. 16. Srivastava B, Quinn WJ III, Hazard K, Erikson J, Allman D. Characterization of marginal zone B cell precursors. J Exp Med 2005;202:1225–34. 17. Hsu HC, Yang P, Wang J, Wu Q, Myers R, Chen J, et al. Interleukin 17-producing T helper cells and interleukin 17 orches- 1064 18. 19. 20. 21. 22. 23. 24. trate autoreactive germinal center development in autoimmune BXD2 mice. Nat Immunol 2008;9:166–75. Allman D, Pillai S. Peripheral B cell subsets. Curr Opin Immunol 2008;20:149–57. Kraal G, Schornagel K, Streeter PR, Holzmann B, Butcher EC. Expression of the mucosal vascular addressin, MAdCAM-1, on sinus-lining cells in the spleen. Am J Pathol 1995;147:763–71. Attanavanich K, Kearney JF. Marginal zone, but not follicular B cells, are potent activators of naive CD4 T cells. J Immunol 2004;172:803–11. Ronnblom L, Alm GV, Eloranta ML. Type I interferon and lupus. Curr Opin Rheumatol 2009;21:471–7. Agrawal H, Jacob N, Carreras E, Bajana S, Putterman C, Turner S, et al. Deficiency of type I IFN receptor in lupus-prone New Zealand mixed 2328 mice decreases dendritic cell numbers and activation and protects from disease. J Immunol 2009;183:6021–9. Nacionales DC, Kelly-Scumpia KM, Lee PY, Weinstein JS, Lyons R, Sobel E, et al. Deficiency of the type I interferon receptor protects mice from experimental lupus. Arthritis Rheum 2007;56: 3770–83. Santiago-Raber ML, Baccala R, Haraldsson KM, Choubey D, Stewart TA, Kono DH, et al. Type-I interferon receptor deficiency reduces lupus-like disease in NZB mice. J Exp Med 2003;197:777–88. WANG ET AL 25. Jego G, Palucka AK, Blanck JP, Chalouni C, Pascual V, Banchereau J. Plasmacytoid dendritic cells induce plasma cell differentiation through type I interferon and interleukin 6. Immunity 2003;19:225–34. 26. Asselin-Paturel C, Brizard G, Chemin K, Boonstra A, O’Garra A, Vicari A, et al. Type I interferon dependence of plasmacytoid dendritic cell activation and migration. J Exp Med 2005;201:1157–67. 27. Yoneyama H, Matsuno K, Zhang Y, Nishiwaki T, Kitabatake M, Ueha S, et al. Evidence for recruitment of plasmacytoid dendritic cell precursors to inflamed lymph nodes through high endothelial venules. Int Immunol 2004;16:915–28. 28. McKenna K, Beignon AS, Bhardwaj N. Plasmacytoid dendritic cells: linking innate and adaptive immunity. J Virol 2005;79:17–27. 29. Mirshahidi S, Huang CT, Sadegh-Nasseri S. Anergy in peripheral memory Cd4⫹ T cells induced by low avidity engagement of T cell receptor. J Exp Med 2001;194:719–32. 30. Lopes-Carvalho T, Kearney JF. Development and selection of marginal zone B cells. Immunol Rev 2004;197:192–205. 31. Martin F, Kearney JF. Positive selection from newly formed to marginal zone B cells depends on the rate of clonal production, CD19, and btk. Immunity 2000;12:39–49. 32. Pillai S, Cariappa A. The follicular versus marginal zone B lymphocyte cell fate decision. Nat Rev Immunol 2009;9:767–77.