Lupus is an Ab-mediated autoimmune disease. One of the potential contributors to the development of systemic lupus erythematosus is a defect in naturally occurring CD4+CD25+ regulatory T cells. Thus, the generation of inducible regulatory T cells that can control autoantibody responses is a potential avenue for the treatment of systemic lupus erythematosus. We have found that nasal administration of anti-CD3 mAb attenuated lupus development as well as arrested ongoing lupus in two strains of lupus-prone mice. Nasal anti-CD3 induced a CD4+CD25−latency-associated peptide (LAP)+ regulatory T cell that secreted high levels of IL-10 and suppressed disease in vivo via IL-10- and TFG-β-dependent mechanisms. Disease suppression also occurred following adoptive transfer of CD4+CD25−LAP+ regulatory T cells from nasal anti-CD3-treated animals to lupus-prone mice. Animals treated with nasal anti-CD3 had less glomerulonephritis and diminished levels of autoantibodies as measured by both ELISA and autoantigen microarrays. Nasal anti-CD3 affected the function of CD4+ICOS+CXCR5+ follicular helper T cells that are required for autoantibody production. CD4+ICOS+CXCR5+ follicular helper T cells express high levels of IL-17 and IL-21 and these cytokines were down-regulated by nasal anti-CD3. Our results demonstrate that nasal anti-CD3 induces CD4+CD25−LAP+ regulatory T cells that suppress lupus in mice and that it is associated with down-regulation of T cell help for autoantibody production.
Systemic lupus erythematosus (SLE)3 is an autoimmune disease characterized by autoantibody production and glomerulonephritis. T cell help is required for the production of high-affinity IgG autoantibodies, which are closely linked to tissue damage in lupus (1, 2, 3, 4). A distinct subset of T cells, follicular helper T cells, selects mutated high-affinity B cells within germinal centers (5). Follicular helper T cells are emerging as a cellular subset with a functional program different from that of extrafollicular Th1 or Th2 T cells: they express high levels of ICOS (6, 7) and have distinct patterns of gene expression of cytokines (predominantly IL-21) and chemokine receptor CXCR5 (8, 9). The expression of CXCR5 by these T cells allows them to localize to B cell follicles where they provide help to B cells. Mice lacking CXCR5 display major aberrations in splenic follicular architecture and reduced numbers of lymph nodes and Peyer’s patches (10, 11). Increased autoantibodies could result from a lack of control of cognate interaction between helper T and B cells due to defective T cell regulation. Recently it has been demonstrated that the proinflammatory cytokine IL-17 orchestrates the spontaneous formation of autoreactive germinal centers by arresting the migration of B cells (12). This arrest provides an optimal microenvironment for the generation of autoantibodies that contribute to the development of glomerulonephritis in BXD2 mice (12).
One of the mechanisms of defective tolerance in lupus may be a defect in the number and/or function of naturally occurring CD4+CD25+ regulatory T cells. Defects in regulatory T cells have been reported in SLE patients (13, 14, 15, 16) and lupus-prone mice (17, 18). We have studied the generation of inducible regulatory T cells by oral anti-CD3 (19, 20). These regulatory T cells express surface latency-associated peptide (LAP) and suppress experimental autoimmune encephalomyelitis (EAE) (19) and diabetes (20) in mice by TGF-β-dependent mechanisms. LAP identifies a class of regulatory T cells that function in a TGF-β-dependent fashion (21, 22, 23, 24, 25). LAP is the amino-terminal domain of the TGF-β precursor peptide and remains noncovalently associated with the TGF-β peptide after cleavage, forming the latent TGF-β complex. CD4+LAP+ T cells appear to be distinct from naturally occurring CD4+CD25+ regulatory T cells, although it has been reported that CD4+CD25+ T cells may express TGF-β on their surface and mediate their suppressive function by presenting TGF-β to a receptor on target cells via cell-to-cell contact (22, 24, 25, 26).
Herein we investigate whether nasal anti-CD3-induced LAP+ regulatory T cells could control the cognate interaction between helper T and B cells and suppress disease in lupus-prone mice. We chose the nasal route, as there have been reports that oral tolerance may be defective in some lupus mouse strains (27, 28). Anti-CD3 mAbs given i.v. have been used to treat both animal models of autoimmunity (29, 30, 31, 32, 33) and transplantation (34, 35, 36), but have never been tested in lupus models or other Ab-mediated diseases. Intravenously administered anti-CD4 has been tested for the treatment of murine lupus (37, 38) (reviewed in Ref. 39). Other studies have used a strategy to expand the naturally occurring CD4+CD25+ regulatory T cells in vivo for the treatment of lupus using the histone peptide H471 (40) and a peptide based on anti-DNA Ig sequences (41, 42). We show herein that nasal anti-CD3 induces an IL-10-secreting CD4+CD25−LAP+ regulatory T cell that suppresses lupus in mice and is associated with a down-regulation of IL-17 and IL-21 expression by CD4+ICOS+CXCR5+ follicular helper T cells.
Materials and Methods
(NZB × SWR)F1 (SNF1) mice were bred and maintained at our facility at the Harvard Institutes of Medicine. Parental female NZB and male SWR mice and lupus-prone female (NZB × NZW)F1 (BWF1) mice were purchased from The Jackson Laboratory. Only female mice were used in our experiments. All mice were housed in a specific pathogen-free environment according to the animal protocol guidelines of the Committee on Animals of Harvard Medical School, which also approved the experiments.
Ags and Abs
The histone peptide H471 based on the amino acid sequence of histone protein H4 at positions 71–93 was synthesized by F-moc chemistry (Biopolymer Laboratory, Harvard Medical School). The peptide was purified by HPLC using a gradient of water and acetonitrile and was analyzed by mass spectrometry for purity (97%). Abs specific to CD3 (145-2c11) and CD28 (37.51) were used to stimulate T cells in vitro. In some experiments neutralizing anti-mouse IL-10 (JES5-2A5) or TGF-β (1D11) or relevant isotype control Ab (BioXCell) was added to cell culture. Fluorescent (FITC or PE)-conjugated anti-mouse Abs used in flow cytometry were CD4-specific (H129.19), CD25-specific (PC61), CXCR5-specific (2G8), ICOS-specific (7E17G9), IL-17-specific (TC11), 7-aminoactinomycin D, and streptavidin-allophycocyanin (all from BD Biosciences). For Fcγ receptor blocking we used CD16/CD32-specific Ab. Affinity-purified biotinylated goat LAP-specific polyclonal Ab was from R&D Systems.
Nasal administration of Ab and immunization
Mice were nasally dosed for 5 consecutive days with hamster IgG CD3-specific F(ab′)2-specific Ab (clone 145-2C11, BioXCell) or hamster IgG control F(ab′)2 Ab (BioXCell) dissolved in PBS or PBS alone. In disease studies we nasally treated mice with multiple courses (five consecutive doses as one course) of Ab at 1-wk intervals. To accelerate lupus development in SNF1 mice, each mouse received an intradermal injection of H471 peptide (100 μg) emulsified in CFA (Sigma-Aldrich) and an i.p. injection of H471 (100 μg) in IFA 10 days later.
To test the in vivo regulatory function of LAP+ T cells, we immunized SNF1 mice at 4–6 wk of age twice with H471 peptide to accelerate disease. Immediately after the second immunization we transferred PBS or freshly isolated whole CD4+ or CD4+ T cells depleted of LAP+ cells in PBS from SNF1 donors nasally treated with isotype control (IC) or anti-CD3 F(ab′)2 Ab. Each recipient received 2 × 106 T cells i.p. We then followed the recipients for 90 days. To test the in vivo role of antiinflammatory cytokines, mice received five i.p. injections of 50 μg of IC or neutralizing anti-IL-10 or anti-TGF-β Ab on alternative days starting from the day before T cell transfer. We then followed the recipients for 35 days. To estimate mouse kidney function we measured the level of proteinuria in urine using Albustix (Bayer) weekly.
T cell proliferation and in vitro Ab production
Cervical (superficial and deep, CLNs) or peripheral (axial, brachial, and inguinal, PLNs) lymph node cells or splenocytes were cultured in triplicate at 1.5 × 106/ml in the presence of various amounts of Ag or Abs or alone in 96-well round-bottom microtiter plates (Corning) for 96 h at 37°C with 5% CO2 in a humid incubator. CD4+ T cells were negatively selected using a cocktail of Abs against other cell types (R&D Systems). The purity of selected cells was checked by flow cytometry. In all experiments, selection efficiency was >95%. For cell sorting, CD4+ T cells or whole lymphocytes were incubated with biotinylated goat LAP-specific polyclonal Ab at 1 μg per million cells before being stained with fluorescent anti-mouse CD4, CD25 and Sav-allophycocyanin (all at 0.5 μg per million cells). CD4+CD25−LAP− or CD4+CD25−LAP+ T cells were sorted using a FACSVantage SE (BD Biosciences). The purity of each population was >98% by flow cytometric analysis. Tissue culture medium was RPMI 1640 with 4.5 g/L glucose and l-glutamine (BioWhittaker) supplemented with 2% penicillin and streptomycin (BioWhittaker) and 1% FCS. Cultures were pulsed with 0.25 μCi tritiated thymidine ([3H]d-Thd; PerkinElmer) for the last 6 h. [3H]d-Thd incorporation was measured using a liquid scintillation beta counter (Wallac, PerkinElmer). Cell proliferation was expressed in Δcpm. To measure T cell help and Ab production by B cells in vitro, sorted CD4+ICOS+CXCR5+ follicular helper T cells were cocultured with positively selected CD19+ B cells or B cells were cultured alone for 5 days. Cells were centrifuged at 1200 rpm at 4°C for 10 min before supernatant was collected, filtered through a cell strainer (0.22 μm), and centrifuged at 10,000 × g for 20 min. Culture supernatant was either used fresh or stored at −20°C in the presence of protease inhibitor until used in ELISA for detection of anti-dsDNA autoantibodies.
Histology and immunofluorescent staining
Mouse kidneys were fixed in 10% formalin (Fisher Scientific). Before periodic acid-Schiff (PAS) staining, kidneys were embedded in paraffin (Tissue-Tek) and 5-μm kidney sections were cut on a cryostat. For immunofluorescent staining, frozen kidney sections were air dried at −80°C for 30 min, fixed in 100% alcohol for 1 min, and bleached with 0.1% sodium borohydride (Sigma-Aldrich) in PBS for 10 min at room temperature. Sections were then washed twice with PBS and nonspecific binding sites were blocked with 10% normal rat serum in PBS for 1 h at room temperature. Following two further washes with PBS, kidney sections were stained with FITC- or PE-conjugated anti-mouse IgG Ab at 1/10,000 dilution (Molecular Probes) or FITC-conjugated anti-mouse complement C3 Ab at 1/200 dilution (Fisher Scientific) for 1 h at room temperature in dark. Unbound Abs were washed away with PBS and sections were embedded in Vectashield containing DAPI (4′,6-diamidino-2-phenylindole) (Vector Laboratories). Slides were examined using an Axioskop 2 Plus fluorescence illumination system (Zeiss) in a blinded fashion.
The level of cytokines produced in vitro by cell cultures was determined using BD OptEIA ELISA set and reagent set B (BD Biosciences). Samples were tested in triplicate using the manufacturer’s recommended assay procedure. Cell culture supernatant was harvested at different time points (24 h for IL-2; 72 h for IFN-γ and IL-4; 90 h for IL-10 and TGF-β) for the detection of cytokines. For quantitative PCR, RNA was extracted from FACS-sorted cells using RNAeasy columns (Qiagen) after 5 h of stimulation with PMA (50 ng/ml) and ionomycin (1000 ng/ml). cDNA was transcribed as recommended (Applied Biosystems) and used as a template for quantitative PCR. Primer/probe mixtures for mouse IL-21 was obtained from Applied Biosystems. The TaqMan analysis was performed on the AB 7500 Fast system (Applied Biosystems). Gene expression was normalized to the expression of β-actin.
Cells were washed (12,000 rpm, 5 min at 4°C) with PBS containing 2% BSA (PBS/BSA, BioWhittaker). Fcγ receptors were blocked by incubation with anti-CD16/CD32 Ab for 30 min at 4°C. Cells were washed twice before being stained with FITC-, PE-, or allophycocyanin-conjugated anti-mouse cell-surface molecule Abs (1 μg/106 cells/test) or relevant IC Ab for 30 min at 4°C in the dark. After staining, cells were washed again with PBS/BSA before flow cytometry (FACScan, BD Biosciences). For intracellular IL-17 staining, cells (10 × 107 cells/ml) in culture medium containing 1 μl/ml GolgiStop (BD Biosciences) were stimulated with PMA (50 ng/ml) and ionomycin (1000 ng/ml) for 4 h at 37°C with 5% CO2 in humid incubator. After incubation, cells were fixed and permeabilized before being stained with PE-conjugated anti-mouse IL-17 Ab. All FACS data were analyzed using FlowJo software (Tree Star).
ELISA for serum autoantibodies
Autoantibodies were measured as described previously (43). Briefly, dsDNA was used at 20 μg/ml or ssDNA was used at 10 μg/ml or histone peptide H471 was used at 5 μg/ml. For the detection of total IgG or IgG1 or IgG2a Abs, 50 μl/well of HRP-conjugated rat anti-mouse Ab (BD Biosciences) at 0.001 μg/ml was added and incubated at 37°C for 1 h.
Lupus autoantigen microarray
A total of 39 lupus-associated autoantigens were purchased from Sigma-Aldrich. Glomeruli extracts were prepared according to a previously established protocol (44). Glomeruli extracts were used at 10 μg/ml. Lupus-associated autoantigens were spotted onto epoxy slides (TeleChem) in replicates of six as described previously (45). Nonspecific binding sites on microarrays were blocked with 1% BSA in PBS at 37°C for 1 h and mouse serum diluted 1/200 in blocking buffer was added and incubated at 37°C for 2 h. Following incubation, microarrays were washed before Cy3-conjugated goat anti-mouse IgG or Cy5-conjugated goat anti-mouse IgM detecting Abs (Jackson ImmunoResearch Laboratories) diluted 1/500 in blocking buffer was added and incubated at 37°C for 45 min. After incubation the arrays were scanned using a ScanArray 4000X scanner (GSI Luminomics) and the IgM and IgG results were recorded separately. Raw data were normalized and analyzed using the GeneSpring software (Silicon Genetics). Raw data have been deposited at the Gene Expression Omnibus (GEO) website (www.ncbi.nlm.nih.gov/projects/geo) under the entry name “Antibody reactivity in anti-CD3 treated SLE-mice” with GEO accession (no. GPL7200). Serum reactivity was defined as the mean fluorescent intensity of the replicates. Results were analyzed with the Wilcoxon rank-sum test, which is a nonparametric test robust to outliers, and the Benjamini-Hochberg false discovery method (46). To perform the hierarchical clustering of the samples we used a pairwise average linkage algorithm based on Spearman’s rank correlation as a distance measurement (46). The classification resulting from the hierarchical clustering was significant in a leave-one-out cross-validation (46).
Statistical differences in cell proliferation and circulating IgG levels were derived from two-way ANOVA and Student’s t test, respectively. The Wilcoxon rank-sum test was used for all of the pairwise group comparisons. A closed testing procedure was used to control for multiple comparisons, and the appropriately adjusted p values were reported. A p-value of <0.05 was considered significant.
Nasal administration of anti-CD3 before systemic challenge with H471 peptide suppresses lupus development in SNF1 mice
Before disease studies, we performed dosing experiments and measured the immunologic effects of nasal administration of anti-CD3 F(ab′)2 Ab on T cell responses. We chose to use the F(ab′)2 fragment of anti-CD3 Ab to eliminate mitogenic effects of anti-CD3 related to the interaction of the Fc portion with FcRs present on phagocytes and NK cells. SNF1 mice were nasally dosed with IC or anti-CD3 Ab (0.5 or 5 μg/day) for 5 consecutive days. We chose a regimen of 5 consecutive days based on previous protocols for both oral (19) and i.v. administered anti-CD3 (31, 33). Pooled CLN cells were isolated 72 h after the last nasal dose and stimulated with soluble anti-CD3 and anti-CD28 for 96 h. We found significant suppression of T cell proliferation following nasal 0.5 μg anti-CD3 (Δcpm 10,000 ± 230 vs 56,000 ± 450 for IC-treated controls, two-way ANOVA test, p = 0.0005). No effect was observed at the 5 μg dose, which is in concordance with our studies of oral anti-CD3 in the EAE model where lower doses were more effective than higher doses (19). Given these results, we initiated studies in lupus disease models using 0.5 μg nasal anti-CD3.
Systemic immunization of SNF1 mice with peptide H471 expressing a dominant pathogenic T cell epitope in histone protein H4 rapidly accelerates the development of lupus in these mice by promoting a cognate interaction between nucleosome-reactive T and B cells (43, 47). Young (4–6 wk old) female SNF1 mice received nasally three 5-day courses of 0.5 μg nasal IC or anti-CD3 Ab every other week during a 6-wk period before immunization with H471 peptide. As shown in Fig. 1,A, nasal anti-CD3-treated mice had significantly lower incidence of proteinuria compared with IC-treated controls. PAS staining of kidney sections revealed hypercellularity and thickening of glomerular basement membrane in mice treated with IC, whereas nasal anti-CD3-treated mice did not show hypercellularity or inflammation (Fig. 1,B). When we quantified glomerulonephritis, we found a significantly lower incidence of severe glomerulonephritis (>grade 3) in these mice (p = 0.004, Fig. 1,C). Furthermore, nasal anti-CD3 suppressed the production of IgG autoantibodies to both dsDNA and H471 peptide (Fig. 1, D and E, respectively), splenocyte proliferation to H471 stimulation (Fig. 1,F), and CD4+ T cell proliferation to plate-bound anti-CD3 stimulation (Fig. 1 G). Thus, nasal anti-CD3 given before the onset of lupus resulted in attenuation of disease severity and progression.
Nasal anti-CD3 induces CD4+CD25−LAP+ regulatory T cells
We have previously shown that anti-CD3 mAb given orally induced a CD4+CD25−LAP+ regulatory T cell that suppresses EAE (19) and diabetes (20). We investigated LAP expression following nasal anti-CD3. We observed an up-regulation of CD4+CD25−LAP+ T cells with no changes in CD4+CD25+ or CD4+CD25+LAP+ T cells (Fig. 2,A) following nasal anti-CD3. To investigate the regulatory function of CD4+CD25−LAP+ T cells induced following nasal administration of anti-CD3 Ab, SNF1 mice were nasally dosed with a 5-day course of 0.5 μg IC or anti-CD3 Ab before being immunized with H471 peptide. Ten days following immunization, CD4+CD25−LAP+ T cells were sorted from PLNs and cocultured with splenocytes as responders in the presence of H471. As shown in Fig. 2,B, CD4+CD25−LAP+ T cells from nasal anti-CD3-treated mice suppressed the proliferation of responder cells to H471 stimulation in vitro, whereas no suppression was observed in IC control-treated animals. Furthermore, CD4+CD25−LAP+ T cells were sorted from CLNs after a 5-day course of 0.5 μg IC or anti-CD3 and cocultured with CD4+CD25−LAP− responder T cells from naive mice in the presence of soluble anti-CD3 and anti-CD28. As shown in Fig. 2 C, CD4+CD25−LAP+ T cells from nasal anti-CD3-treated animals suppressed the proliferation of responder T cells in coculture. Because less than 1% of cells were CD4+CD25−LAP+ T cells in IC-treated animals, additional mice were used to obtain these cells for the suppression assay.
We investigated the cytokine profile of CD4+CD25−LAP+ regulatory T cells following nasal anti-CD3. As shown in Fig. 2,D, CD4+CD25−LAP+ T cells from anti-CD3-treated animals produced significantly higher levels of IL-10 and lower levels of IL-2 and IFN-γ compared with control animals. No differences were observed in IL-4 and TGF-β production. We next asked if in vitro suppression by CD4+CD25−LAP+ regulatory T cells is mediated by soluble IL-10 and whether in vitro suppression was dependent on cell contact. CD4+CD25−LAP+ T cells from nasally treated mice were cocultured with naive CD4+CD25−LAP− responder T cells in the presence of neutralizing anti-IL-10 or anti-TGF-β Ab. We also tested cells separated by a permeable membrane in a transwell system. Fig. 2 E shows partial reversal of suppression by neutralizing IL-10, and that suppression did not require physical contact between the regulator and the responder cell. We also determined the expression of Foxp3 in CD4+CD25−LAP+ T cells from SNF1 mice nasally treated with a 5-day course of 0.5 μg IC or anti-CD3. Intracellular staining revealed that Foxp3 expression was only marginally up-regulated in CD4+CD25−LAP+ T cells following nasal anti-CD3 (from <1% in IC treated controls to 6% in anti-CD3 treated animals). Thus, nasal anti-CD3 induces a CD4+CD25−LAP+ regulatory T cell that suppresses the proliferation of effector T cells in an IL-10-dependent but cell contact-independent manner.
In vivo suppression of lupus by LAP+ regulatory T cells
To test the in vivo suppressive function of LAP+ regulatory T cells in the accelerated lupus SNF1 model, we transferred whole CD4+ T cells or CD4+ T cells depleted of LAP+ cells from nasally treated SNF1 donors immediately after the second immunization with H471 peptide and measured proteinuria weekly for 90 days. Fig. 3,A shows a marked reduction in the incidence of proteinuria in recipients of CD4+ T cells from nasal anti-CD3-treated donors. Depletion of LAP+ T cells effectively abrogated the protective effect on kidney function in recipients. Histological analysis of mouse kidneys revealed no signs of severe glomerulonephritis in recipients of CD4+ T cells of nasal anti-CD3-treated donors, whereas 50% of recipients of CD4+ T cells depleted of LAP+ cells developed severe glomerulonephritis (Fig. 3,B). As shown in Fig. 3,B, all recipients of CD4+ T cells of IC-treated donors developed severe glomerulonephritis. Almost no IgG anti-dsDNA autoantibody was detected in recipients of CD4+ T cells of nasal anti-CD3-treated donors compared with control recipients (Fig. 3 C), and there was also a significant (p = 0.05) suppression of IgG anti-ssDNA autoantibodies (IC/CD4+ 49 mean U, anti-CD3/CD4+ 21 mean U vs anti-CD3/CD4+LAP− 45 mean U, p = 0.05).
In vivo suppression of lupus by LAP+ regulatory T cells is IL-10 and TGF-β dependent
To test the role of IL-10 and TGF-β in in vivo suppression of disease, we performed adoptive transfer experiments in the accelerated lupus SNF1 model. We transferred whole CD4+ T cells or CD4+ T cells depleted of LAP+ cells from nasally treated SNF1 donors immediately after the second immunization with H471 peptide and treated some recipients with neutralizing anti-IL-10 or anti-TGF-β Ab or a control Ab. Mice were followed for 35 days and proteinuria was measured weekly. Fig. 4,A shows a marked reduction in the incidence of proteinuria in recipients of CD4+ T cells from nasal anti-CD3-treated donors. Depletion of LAP+ T cells effectively abrogated the positive effect on kidney function. Furthermore, injection of anti-IL-10 Ab abrogated the protective effect of adoptively transferred regulatory T cells. An effect was also observed with anti-TGF-β, but its effect was delayed. This suggests that nasally induced LAP+ regulatory T cells exert suppression in vivo initially via IL-10 and that TGF-β is involved later. Additionally, as shown in Fig. 4,B, the decreased glomerular damage and IgG and complement C3 deposition in kidneys of animals that received regulatory T cells were reversed by neutralizing IL-10 and TGF-β in vivo. Recipients of regulatory T cells did not have glomerulonephritis, and neutralization of IL-10 or TGF-β reversed protection from kidney damage (Fig. 4,C). In vivo neutralization of IL-10 or TGF-β also reversed the suppressive effect of LAP+ regulatory T cells on autoantibody production (Fig. 4,D, IgG anti-dsDNA and Fig. 4 E, IgG anti-H471 and IgG anti-ssDNA PBS 58 mean U, LAP-depleted CD4+ 49 mean U, CD4+ + IC 9 mean U, CD4+ + anti-IL-10 50 mean U, CD4+ + anti-TGF-β 46 mean U, CD4+ 7 mean U; CD4+ + IC vs CD4+ + α-IL-10, p = 0.001; CD4+ + IC vs CD4+ + anti-TGF-β, p = 0.001). These results demonstrate that LAP+ regulatory T cells mediate disease suppression in vivo via IL-10 and TGF-β.
Nasal anti-CD3 suppresses the function of, and IL-17 and IL-21 expression by, CD4+ICOS+CXCR5+ follicular helper T cells
To investigate the effect of nasal anti-CD3 on follicular helper T cells, SNF1 mice at 7 mo of age with established spontaneous lupus (three consecutive weekly proteinuria readings of >300 mg/dl) were nasally treated with six 5-day courses of IC or anti-CD3 Ab, after which splenic CD4+ICOS+CXCR5+ T cells were isolated and cocultured with naive CD19+ B cells for 5 days. We found that CD4+ICOS+CXCR5+ T cells from anti-CD3 nasally treated mice induced less IgG, IgG1, or IgG2a anti-dsDNA Ab production compared with IC-treated controls (Fig. 5,A). Furthermore, nasal anti-CD3 resulted in significant down-regulation of IL-21 expression by CD4+ICOS+CXCR5+ T cells (Fig. 5,B, p = 0.004). As Th17 cells express high levels of IL-21 (48), we investigated the effect of nasal anti-CD3 on the expression of IL-17 by CD4+ICOS+CXCR5+ T cells. As shown in Fig. 5 C, IL-17 is predominantly expressed by CD4+ICOS+CXCR5+ follicular helper T cells and only minimally in CD4+ICOS−CXCR5− T cells. We found a significant decrease in the percentage of IL-17 expressing CD4+ICOS+CXCR5+ T cells in mice nasally treated with anti-CD3 compared with IC-treated controls (p = 0.03). Of note, we followed mice for 90 days after nasal anti-CD3 treatment and observed that mice treated with anti-CD3 (n = 5) had a lower incidence of proteinuria compared with control mice (n = 5). Nasal anti-CD3-treated mice survived beyond 90 days after treatment, whereas no control mice survived beyond 32 days after treatment. These results suggest that suppression of lupus by nasal anti-CD3 is associated with decreased follicular helper T cell function, thus leading to reduced levels of autoantibodies.
Nasal anti-CD3 suppresses established spontaneous lupus in BWF1 mice
BWF1 mice spontaneously exhibit clinical and immunologic features that closely resemble those of SLE in humans, including anti-dsDNA Abs and severe immune complex-mediated glomerulonephritis (49). In female BWF1 mice, autoantibodies become overt at an early stage of life, and proteinuria occurs soon thereafter, with most animals dying by the age of 10 mo (49). To investigate the effect of nasal anti-CD3 on spontaneous disease, 6-mo-old female BWF1 mice that had developed proteinuria (>300 mg/dl) were nasally treated with six 5-day courses of 0.5 μg IC or anti-CD3 Ab every other week. On day 63 after the last treatment, 100% of the nasal anti-CD3-treated mice were alive compared with no survival in the IC-treated control group (Fig. 6,A). In fact, 100% of mice nasally treated with anti-CD3 were still alive 90 days after treatment. We sacrificed some mice on day 50 and examined kidneys, autoantibodies, and measured T cell proliferation. As shown in Fig. 6,B (top panel), there was extensive crescent formation and even collapsed glomerular loops due to extensive interstitial inflammation in control mice, whereas only minor inflammatory infiltrates and thickening of capillaries were observed in nasal anti-CD3-treated mice. Immunofluorescent staining revealed global and diffuse IgG immune complex deposition in kidneys of IC-treated mice compared with focal and segmental IgG immune complex deposition in kidneys of nasal anti-CD3-treated mice (Fig. 6,B, middle panel). There was also a marked reduction in complement C3 deposition in kidneys of nasal anti-CD3-treated mice compared with IC-treated controls (Fig. 6,B, bottom panel). Seventy-five percent of control mice showed end-stage glomerulonephritis compared with 13% of nasal anti-CD3-treated mice where 74% showed only mild glomerulonephritis (Fig. 6,C). This was accompanied by a significant reduction in IgG anti-dsDNA (Fig. 6 D) and ssDNA autoantibody production (IC 65 mean U vs anti-CD3 35 mean U, p = 0.008). Suppression of disease pathology was also associated with a CD4+ T cell hyporesponsiveness to plate-bound anti-CD3 stimulation in vitro compared with untreated control mice with established disease (6-mo-old control) or IC-treated control mice (6 mo control 143,000 Δcpm, IC 97,000 Δcpm vs anti-CD3 51,000 Δcpm, p = 0.05). These results indicate that nasal administration of anti-CD3 mAb led to control of ongoing disease processes and prolonged mouse survival. As with oral anti-CD3 (19), nasal anti-CD3 did not modulate the TCR/CD3 complex on T cells or deplete T cells (data not shown). Furthermore, we did not observe adverse effects (such as cytokine release syndrome) in mice even after 30 nasal administrations, and there was no anti-globulin response against anti-CD3.
Reduction of multiple autoantibody reactivities in BWF1 mice by nasal anti-CD3 as measured by Ag microarrays
Ag microarrays represent a comprehensive method for measuring serum reactivity against large numbers of autoantigens. In lupus, Ag microarrays have been used to characterize autoantibodies associated with the disease both in humans (50, 51) and mice (51). We investigated the effect of nasal anti-CD3 on autoantibody responses in BWF1 mice with spontaneous lupus using a panel of 40 Ags (Table 1). Nasal anti-CD3 led to a significant reduction in IgG responses against 32 of 40 autoantigens tested (Fig. 7). The heat map in Fig. 7 shows that samples from anti-CD3-treated mice were clustered together and separate from samples taken from IC-treated (p = 0.0003) and 9-mo-old untreated mice (p = 0.0014). The IgG responses to autoantigens showing a significant reduction after nasal anti-CD3 are shown in Fig. 8 according to the intensity of reactivity of the 9-mo untreated animals. The level of IgG responses to these autoantigens in nasal anti-CD3-treated mice was comparable to 1-mo-old naive mice (Fig. 8). Of note, we did not observe a significant difference in the IgM autoantibody reactivity. This finding is consistent with the effect of nasal anti-CD3 on T cell-dependent Ab production. Thus, nasal anti-CD3 ameliorates spontaneous lupus by suppressing a broad array of autoantibody production.
|Ag .||Source .|
|ECM (extracellular matrix)||Sigma-Aldrich|
|GBM (glomerular basement membrane)||Sigma-Aldrich|
|Histone I (F1 H1)||ImmunoVision|
|Histone III and IV||ImmunoVision|
|Histones IV and IIa||ImmunoVision|
|Ag .||Source .|
|ECM (extracellular matrix)||Sigma-Aldrich|
|GBM (glomerular basement membrane)||Sigma-Aldrich|
|Histone I (F1 H1)||ImmunoVision|
|Histone III and IV||ImmunoVision|
|Histones IV and IIa||ImmunoVision|
We show herein that nasal anti-CD3 induces an IL-10-secreting CD4+CD25−LAP+ regulatory T cell that suppresses lupus in mice and is associated with down-regulation of the function of IL-17+CD4+ICOS+CXCR5+ follicular helper T cells. The induction of LAP+ regulatory T cells following nasal administration of anti-CD3 Ab suppressed lupus both in an accelerated model in SNF1 mice and a spontaneous model in BWF1 mice.
Parental anti-CD3 has been studied in animal models of autoimmunity, with most studies having been conducted in the NOD model of autoimmune diabetes (Refs. 29 , 31 and reviewed in Ref. 52). In autoimmune diabetes, parental CD3-specific Ab induced tolerance that involved transferable T cell tolerance by CD4+CD25+ T cells that functioned in a TGF-β-dependent fashion (31). LAP+ T cells were not examined. These regulatory T cells did not appear immediately after i.v. treatment, suggesting that i.v. CD3-specific Ab may have deleted pathogenic Th1 cells, which then allowed the development of regulatory T cells. In EAE, i.v. CD3-specific Ab reversed established EAE but was not effective before induction of EAE, and therapeutic effects were postulated to be secondary to the induction of T cell tolerance, apoptosis, or alterations in cell trafficking, with no evidence of active regulation (33). There have been no studies of anti-CD3 in lupus models and no studies of anti-CD3 in an Ab-mediated autoimmune condition. In a study using i.v. anti-CD4-depleting Ab in a human autoantibody-induced lupus model, positive effects were observed but only when Ab was given before disease establishment (53).
We found that nasal anti-CD3 given in SNF1 mice with accelerated disease or at the height of spontaneous disease in BWF1 mice lessened damage to kidneys by reducing immune complex and complement C3 deposition. This reduction was consistent with a significant down-regulation of anti-DNA autoantibody production following nasal anti-CD3 Ab treatment. We hypothesize that nasal anti-CD3 in lupus-prone mice resulted in the induction of LAP+ regulatory T cells that suppressed the differentiation of follicular helper T cells (reviewed in Ref. 8), leading to the disruption of autoaggressive Th and B cell cognate interactions. We found that nasal anti-CD3-treated mice had reduced IL-21 expression by CD4+ICOS+CXCR5+ helper T cells that is needed to initiate plasma cell generation (54, 55) and Ab production (56, 57). This resulted in markedly reduced production of autoantibodies by B cells. The CD4+ICOS+CXCR5+ helper T cells express high levels of proinflammatory cytokine IL-17. This result correlates well with the finding by Hsu et al. (12), which shows that autoimmune BXD2 mice express heightened levels of IL-17 and that IL-17 is required for the spontaneous development of germinal centers and production of pathogenic autoantibodies in these mice. Injection of antagonistic Ab to IL-17 in mice disrupted T and B cell interaction and the formation of germinal centers and autoantibodies (12). Herein, we showed that nasal anti-CD3 treatment in SNF1 mice with spontaneous lupus led to a down-regulation of IL-17 expression in CD4+ICOS+CXCR5+ helper T cells. We are currently investigating the effect of nasal anti-CD3 in the development of germinal centers in lupus-prone mice. The differentiation and maintenance of IL-17-producing “Th17” cells that are highly proinflammatory and induce severe autoimmunity are mediated by a combination of IL-6 and TGF-β and IL-23, respectively (58). We are now examining the role of IL-6, TGF-β, and IL-23 in the differentiation and maintenance of IL-17+CD4+ICOS+CXCR5+ follicular helper T cells in lupus animals. It is conceivable that excessive Th cell responses due to insufficient control by a defective peripheral CD4+CD25+ regulatory T cell pool contribute to the development of SLE. Indeed, the Sanroque mouse, which has a mutation that acts within mature T cells to cause formation of excessive numbers of follicular helper T cells and excessive production of IL-21, develops high titers of autoantibodies and SLE-like pathology (59). Thus, induction of LAP+ regulatory T cells by nasal anti-CD3 appears to correct an immune imbalance associated with murine models of SLE.
We have previously shown that orally administered anti-CD3 suppressed EAE (19) and diabetes (20) in a dose-dependent fashion, with lower doses being more effective. Suppression was mediated by CD4+CD25−LAP+ regulatory T cells that suppressed in vitro and in vivo in a TGF-β-dependent manner. Oral anti-CD3 did not enhance the frequency or function of the naturally occurring Foxp3+CD4+CD25+ regulatory T cells, and thus we focused our investigation on the CD4+CD25−LAP+ regulatory T cells following nasal anti-CD3 in lupus-prone mice. Herein, we show that anti-CD3 is effective when given nasally and exhibited a dose-response effect, with lower doses being more effective. A similar dose effect occurs with nasal histone peptide Ags, where suppression of T cell responses was only seen at lower peptide dosages (43). We believe that this dose effect of nasal anti-CD3 relates to delivering a “weak signal” to the T cell via the TCR/CD3 complex that preferentially induces regulatory T cells (60, 61). As with oral anti-CD3, nasal anti-CD3 did not modulate the TCR/CD3 complex on T cells or deplete T cells. Notably, a log less Ab was administered nasally as compared with the oral route and, as would be predicted, we observed immunologic effects of nasal anti-CD3 in the CLNs. This further demonstrates the direct activation of T cells by mucosal administration of anti-CD3 Ab.
The regulatory T cell induced by nasal anti-CD3 was different from that induced by oral anti-CD3, although there were similarities. In both instances there was up-regulation of LAP expression on CD4+ T cells following mucosal anti-CD3. The increase in LAP expression was exclusively found on CD4+ T cells that do not express CD25. Additionally, like oral anti-CD3, we did not detect a significant increase in Foxp3 expression both by flow cytometry and messager RNA detection. Unlike oral anti-CD3, however, the primary cytokine responsible for the immunologic effects of the nasally induced CD4+CD25−LAP+ regulatory T cell was IL-10. The in vitro suppressive effects of nasally induced CD4+CD25−LAP+ regulatory T cells were partially reversed by anti-IL-10; no in vitro effect of anti-TGF-β was observed. Physical separation of the regulator and responder cell by a permeable membrane did not affect suppression. The in vitro effect was Ag nonspecific, as CD4+CD25−LAP+ T cells isolated from SNF1 mice nasally treated with anti-CD3 suppressed proliferation of naive CD4+CD25−LAP− T cells as well as H471 peptide-experienced T cells. It is known that the cytokine profile of regulatory T cells induced by the oral route is different than that induced by the nasal route, with TGF-β preferentially induced after oral Ag and IL-10 after nasal Ag (reviewed in Ref. 62). This may be related to different properties of dendritic cells at different mucosal sites (63) (reviewed in Ref. 64). It appears from our work that the same holds true for nasally or orally administered anti-CD3. Interestingly, in vivo, the suppressive function of nasal anti-CD3-induced CD4+CD25−LAP+ regulatory T cells appears to depend on both IL-10 and TGF-β. Anti-IL-10 acted immediately to reverse suppression by adoptively transferred regulatory T cells, whereas anti-TGF-β appeared to act in a delayed fashion. This suggests that complex in vivo processes are associated with suppression by these cells. Given the cytokine profile of the cells induced by nasal anti-CD3 and the cytokines used to suppress in the lupus model, we may be inducing Tr1-like regulatory T cells with nasal anti-CD3 (65). We previously demonstrated that nasal administration of H471 peptide in lupus-prone mice led to the induction of T cell anergy (43), and it was associated with Ag presentation by B cells that lack CD80 and CD86 expression (66). We showed that the H471-anergized T cells were immunologically unresponsive to further stimulation, but these cells did not possess suppressive function over naive T cell response (43). In contrast to nasal anti-CD3, nasal administration of H471 peptide did not induce regulatory T cells. Thus, it appears that the type of agent being administered nasally dictates the immunologic outcome; that is, peptide Ag leads to T cell anergy and anti-CD3 leads to the generation of LAP+ regulatory T cells. Interestingly, the in vivo effect of adoptively transferred CD4+CD25−LAP+ regulatory T cells lasted more than 90 days and suppressed established, spontaneous lupus. It is possible that CD4+CD25−LAP+ regulatory T cells induced additional regulatory T cells in vivo, for example, infectious tolerance (67), which then functioned to suppress disease at a later time point. Our finding of a delayed effect of in vivo neutralization of TGF-β on disease suppression is consistent with this view.
Ag microarray experiments demonstrated that nasal anti-CD3 treatment of mice with spontaneous lupus reduced the pathogenic IgG autoantibody response to a level comparable to pre-disease state. Interestingly, nasal anti-CD3 significantly affected IgG but not IgM responses. This indicates that nasal anti-CD3 affected the Ab isotype switching process that is dependent on CD4 help. Our finding of a down-regulation of the function of CD4+ICOS+CXCR5+ follicular helper T cells suggests that LAP+ regulatory T cells may control CD4+ T cell help for autoantibody production by B cells in vivo.
We performed our experiments with a F(ab′)2 Ab to eliminate any potential side effects related to the Fc portion of the molecule that might occur after multiple administrations of the Ab nasally. We observed no mitogenic effect of nasal hamster CD3-specific F(ab′)2 Ab in mice and no evidence of cytokine release syndrome (wasted appearance, ruffled fur), even after 30 nasal administrations. Furthermore, we did not observe an anti-globulin response against anti-CD3 in mice nasally treated with anti-CD3 (data not shown). Nasal administration of CD3-specific Ab is applicable for chronic therapy and would not be expected to have side effects including cytokine release syndromes and anti-globulin responses.
In summary, we have shown that nasal anti-CD3 Ab induces a CD4+CD25−LAP+ regulatory T cell that suppressed Th cell function, leading to down-regulation of autoantibody production and glomerulonephritis in murine models of lupus.
We thank Drs. V. K. Kuchroo and B. H. Waksman for critical assessment of the manuscript, Anneli Jäger for help in IL-21 detection, Dr. Alexandre Basso for help in Th cell assay, and Brian C. Healy for performing statistical analyses.
The authors have no financial conflicts of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grants AI435801 and NS38037 (to H.L.W.).
Abbreviations used in this paper: SLE, systemic lupus erythematosus; LAP, latency-associated peptide; EAE, experimental autoimmune encephalomyelitis; SNF1, (NZB × SWR)F1; BWF1, (NZB × NZW)F1; IC, isotype control; CLN, cervical lymph node; PLN, peripheral lymph node; PAS, periodic acid-Schiff.