Plasmacytoid dendritic cells and type 1 interferon promote peripheral expansion of forkhead box protein 3⁺ regulatory T cells specific for the ubiquitous RNA‐binding nuclear antigen La/Sjögren's syndrome (SS)‐B

Summary

RNA‐binding nuclear antigens are a major class of self‐antigen to which immune tolerance is lost in rheumatic diseases. Serological tolerance to one such antigen, La/Sjögren's syndrome (SS)‐B (La), is controlled by CD4⁺ T cells. This study investigated peripheral tolerance to human La (hLa) by tracking the fate of hLa‐specific CD4⁺ T cells expressing the transgenic (Tg) 3B5.8 T cell receptor (TCR) after adoptive transfer into lymphocyte‐replete recipient mice expressing hLa as a neo‐self‐antigen. After initial antigen‐specific cell division, hLa‐specific donor CD4⁺ T cells expressed forkhead box protein 3 (FoxP3). Donor cells retrieved from hLa Tg recipients displayed impaired proliferation and secreted interleukin (IL)−10 in vitro in response to antigenic stimulation. Transfer of highly purified FoxP3‐negative donor cells demonstrated that accumulation of hLa‐specific regulatory T cells (T_reg) was due primarily to expansion of small numbers of donor T_reg. Depletion of recipient plasmacytoid dendritic cells (pDC), but not B cells, severely hampered the accumulation of FoxP3⁺ donor T_reg in hLa Tg recipients. Recipient pDC expressed tolerogenic markers and higher levels of co‐stimulatory and co‐inhibitory molecules than B cells. Adoptive transfer of hLa peptide‐loaded pDC into mice lacking expression of hLa recapitulated the accumulation of hLa‐specific T_reg. Blockade of the type 1 interferon (IFN) receptor in hLa Tg recipients of hLa‐specific T cells impaired FoxP3⁺ donor T cell accumulation. Therefore, peripheral expansion of T_reg specific for an RNA‐binding nuclear antigen is mediated by antigen‐presenting pDC in a type 1 IFN‐dependent manner. These results reveal a regulatory function of pDC in controlling autoreactivity to RNA‐binding nuclear antigens.

Introduction

Ubiquitous nuclear and cytoplasmic self‐antigens that constitutively bind RNA are targeted selectively by the autoimmune response in several rheumatic diseases. Autoantibodies directed to one such antigen, La/Sjögren's syndrome (SS)‐B, occur at high prevalence in SS 1 and mothers of infants with congenital heart block or neonatal lupus syndrome 2, and at lower prevalence in systemic lupus erythematosus (SLE) 3. La‐targeting antibodies co‐occur typically with antibodies specific for Ro/SS‐A 3, associate with a higher frequency of extraglandular manifestations of SS including vasculitis, leukopenia, hypergammaglobulinaemia, restrictive pulmonary disease and internal organ involvement 4, 5, 6 and have been reported to associate selectively with vasculitis in SLE 7. Deciphering mechanisms of normal immunological tolerance to RNA‐binding nuclear antigens is essential for understanding the development of autoimmunity to such antigens and for designing effective therapies for mitigating the disorders in which they occur.

Anti‐La autoantibody production is T helper cell‐dependent, as suggested by the association with specific class II human leucocyte antigen (HLA) alleles 8, class‐switched isotypes 9, somatic hypermutation 9 and high serum concentration 10. Using mice expressing transgenic (Tg) human La (hLa) with ubiquitous, nuclear distribution at physiological levels, we have previously reported detectable immunological tolerance in hLa‐specific CD4⁺ T cells, but not in B lymphocytes 11, 12. Further studies revealed that human La‐specific T cells in double Tg mice for an I‐E^k‐restricted T cell receptor (TCR) specific for hLa 67–76 peptide and the hLa neo‐self‐antigen undergo drastic but incomplete thymic clonal deletion 13. A portion of residual, hLa‐specific thymocytes in these mice developed into natural thymic regulatory T cells (T_reg), and the mice displayed high proportions of forkhead box P3 (FoxP3)‐expressing, hLa‐specific CD4⁺ T cells in peripheral lymphoid organs that functioned to maintain immunological self‐tolerance and prevent autoimmune lung pathology. However, it was possible that the lymphopenic environment caused by extensive thymic clonal deletion in this TCR/neo‐self‐antigen double Tg model was responsible for the increased proportion of T_reg in the periphery of these mice.

CD4⁺FoxP3⁺ T_reg are important for controlling autoimmunity and can develop in response to self‐antigen in the thymus 14 or in the periphery from mature, naive T cells that have completed T cell development 15, 16. These cells suppress the responses of conventional T cells by several mechanisms, including immunosuppressive cytokines such as interleukin (IL)−10 or transforming growth factor (TGF)‐β and various cell contact‐dependent mechanisms 17. Both types of cells contribute to the circulating T_reg pool, where they are regulated by homeostatic factors including IL‐2 18, 19, TCR stimulation 20, 21 and co‐stimulation 22. TCR and co‐stimulatory signals promoting maintenance or conversion of T_reg in the periphery have been shown to be delivered by dendritic cells (DC) 23, 24 and, in some instances, B lymphocytes 25, 26, 27, but the mechanisms involved are incompletely understood. DC subsets that have been implicated in peripheral T_reg induction in settings of transplantation, low dose peptide‐induced tolerance and the steady state include CD103⁺ gut DC 28, migratory dermal DC 29 and plasmacytoid pDC (pDC) 30, 31, 32, 33. pDC may drive T_reg development and expansion through low levels of antigen presentation and have been shown to promote T_reg through production of indoleamine‐2,3‐dioxygenase (IDO), co‐stimulatory molecules, secretion of granzyme B and production of retinoic acid 34, 35. Existing studies have not addressed mechanisms of induction of T_reg specific for ubiquitous, RNA‐binding nuclear antigens often targeted in systemic rheumatic diseases.

In the present study, we assessed the fate of hLa‐specific T cells introduced into the periphery of lymphocyte‐replete mice expressing the hLa neo‐self‐antigen in order to understand peripheral T_reg induction and homeostasis in T cells specific for a representative RNA‐binding human nuclear autoantigen. We show that pDC induce expansion of hLa‐specific FoxP3⁺ T_reg in a type I interferon (IFN)‐dependent manner. These studies uncover a regulatory mechanism that should be considered in terms of prophylactic and therapeutic approaches for systemic rheumatic diseases that employ type I IFN blockade.

Materials and methods

Mice

C57BL/6 (B6) mice congenic for H‐2^k (B6.AK‐H2^k/FlaEgJ; Jackson Laboratory, Bar Harbor, ME, USA) were crossed to B6 mice congenic for Thy1.1 (B6.PL‐Thy1^a/CyJ; Jackson Laboratory) to generate B6.H‐2^k/k.Thy1.1^+/+ mice. Line 3 hLa Tg mice described previously 11 were back‐crossed to B6 mice at least 12 generations and crossed to B6.H‐2^k/k.Thy1.1^+/+ congenic mice to generate heterozygous hLa transgenic (hLa Tg) or non‐Tg H‐2^k/k Thy1.1^+/+ and H‐2^k/b Thy1.1^+/+ recipient mice. 3B5.8⁺ hLa‐specific TCR Tg mice described previously 13 were crossed to Tcra ^–/– (B6.129S2‐Tcra^tm1Mom/J; Jackson Laboratory) and B6.H‐2^k congenic mice to generate B6.3B5.8^+/–.Tcra^–/–.H‐2^k/k and B6.3B5.8^+/–Tcra^‐/–.H‐2^k/b donor mice that are naturally Thy1.2⁺. In addition, these donor mice were crossed with C57BL/6‐FoxP3^tm1Flv/J mice (Jackson Laboratory), which express monomeric red fluorescent protein (mRFP)‐tagged FoxP3, to generate B6.3B5.8^+/–.Tcra^–/–.H‐2^k/k.FoxP3‐mRFP⁺ donor mice that are naturally Thy1.2⁺.

Animals were maintained under specific pathogen‐free barrier conditions in the OMRF Laboratory Animal Resource Center until experiments were carried out at 5–12 weeks of age. All studies were approved by the Oklahoma Medical Research Foundation Institutional Animal Care and Use committee.

Cell preparation

Splenocyte suspensions were obtained by passing spleens through 40 μm nylon filters, treating with Tris ammonium chloride solution (TAC; 0·14 M NH₄Cl in 17 mM Tris, pH 7·2) to lyse red blood cells and washing in Dulbecco's modified Eagle's medium (Sigma‐Aldrich, Inc., St Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS), 1× non‐essential amino acids (Life Technologies, Grand Island, NY, USA), 2 mM L‐glutamine, 10 μg/ml penicillin/streptomycin, 50 μM β‐mercaptoethanol and 2 mM sodium pyruvate. Cells from lymph nodes were obtained similarly but without TAC treatment. Cells were quantified using trypan blue exclusion.

Bulk CD4⁺ T cells and sorted T_reg were obtained from B6.3B5.8^+/–.Tcra^–/–.H‐2^k/k.FoxP3‐mRFP donor mice by pre‐purification using CD4 (L3T4) MicroBeads (Miltenyi Biotec, San Diego, CA, USA) positive selection according to the manufacturer's protocol, followed by sorting of mRFP^‐ (FoxP3^–) cells on an Influx cell sorter (BD Biosciences, San Diego, CA, USA).

To obtain adequate numbers of conventional DC (cDC) and pDC, mice were implanted with 5 × 10⁶ Fms‐like kinase 3 ligand (Flt3L)‐secreting B16 melanoma cells by subcutaneous injection to induce DC expansion 36. Within 2 weeks of implantation, spleens were harvested, diced and digested at 37°C for 45 min in 10 ml RPMI‐1640 supplemented with 10% FCS, 5 mM ethylenediamine tetraacetic acid (EDTA), 15 mM HEPES, 1 mg/ml collagenase D (Sigma‐Aldrich) and 0·1mg/mL DNase I (Roche Life Sciences, Indianapolis, IN, USA) followed by TAC treatment. CD11c⁺ cells were enriched using CD11c MicroBeads (Miltenyi Biotec), according to the manufacturer's instructions. cDC and pDC were sorted further using antibodies directed to CD11c, CD19, CD317/PDCA1 and CD45R/B220 on a high‐speed MoFlo XDP1 cytometer (Beckman Coulter, Brea, CA, USA) at > 98% purity.

Donor T cells retrieved from recipient spleens for in‐vitro experiments were selected positively using anti‐Thy1.2 MicroBeads (Miltenyi Biotec) and further purified by MoFlo, sorting for Thy1.2⁺ Thy1.1^– CD4⁺ Vβ10⁺ donor cells. The T cell‐depleted fraction was irradiated (2200 rads) and used as antigen‐presenting cells (APC).

Adoptive transfer experiments

For T cell transfers, 4–6 × 10⁷ unlabelled or 5‐(and 6)‐carboxyfluorescein diacetate succinimidyl ester (CFSE; Life Technologies, Grand Island, NY, USA)‐labelled total splenocytes, 6 × 10⁶ CFSE‐labelled CD4⁺ T cells or 2·5 × 10⁶ CD4⁺ mRFP^– (FoxP3^–) cells harvested from B6.3B5.8^+/–.Tcra^–/–.H‐2^k/k or B6.3B5.8^+/–.Tcra^–/–.H‐2^k/b donor mice were resuspended in 0·2 ml phosphate‐buffered saline (PBS) and injected retro‐orbitally into hLa Tg or non‐Tg H‐2^k/k Thy1.1^+/+ or H‐2^k/b Thy1.1^+/+ recipient mice, respectively. Seven days later (unless noted otherwise), spleens and lymph nodes from recipient mice were harvested and analysed by flow cytometry to assess cell proliferation, T_reg accumulation and B cell and DC phenotype.

Recipient pDC depletion was carried out by treating recipient mice with intraperitoneal (i.p.) injections of 400 μg anti‐pDC antibody (120G8.04; Imgenex; Novus Biologicals, Littleton, CO, USA) 37 or immunoglobulin (Ig)G isotype control (BioXCell, West Lebanon, NH, USA) every 48 h beginning 2 days prior to cell transfer. B cell depletion was carried out by i.p. injection of 250 μg anti‐CD20 antibody (MB20‐11; generously provided by Dr Thomas Tedder) or IgG isotype control antibody (BioXCell) every 48 h beginning 1 day prior to cell transfer. To block the type I IFN receptor, recipient mice were treated i.p. with anti‐IFN‐alpha/beta receptor alpha chain (IFNAR1) 38 [monoclonal antibody mouse IFNAR1 (MAR1)−5A3; Leinco Technologies, St Louis, MO, USA] or IgG1 isotype control antibody (BioXCell). Antibody dosage began daily at 500 μg for the first 3 days (beginning 1 day prior to cell transfer) followed by 250 μg every 48 h thereafter until completion of the experiment.

Sorted splenic pDC were pulsed for 2 h with 10 μM hLa 61–84 peptide or hen egg‐white lysozyme (HEL) 46–61 peptide in Dulbecco's modified Eagle's medium supplemented with 10% FBS, 1× non‐essential amino acids, 2 mM L‐glutamine, 10 μg/ml penicillin/streptomycin, 50 μM β‐mercaptoethanol and 2 mM sodium pyruvate at 1 × 10⁶ cells/ml, washed twice and transferred subsequently (2 × 10⁵ cells/recipient) into non‐Tg H‐2^k/k Thy1.1⁺ recipient mice. The next day, 3B5.8⁺ donor splenocytes were transferred into the same recipient mice. Euthanasia and cell retrieval were conducted 7 days post‐3B5.8 T cell transfer.

Flow cytometry

Single‐cell suspensions were treated with Fc block (anti‐CD16/32 antibodies) for 5 min on ice, then stained with combinations of the following fluorophore‐conjugated monoclonal antibody clones: CD4 (RM4‐5), Vβ10 (B21.5), CD90.1(OX‐7), CD90.2 (53‐2.1), CD45R/B220 (RA3‐6B2), CD11c (N418), CD19 (1D3), CD25 (PC61.5), CD86(GL1), CD80(16‐10A1), MHC II(14‐4‐4S), CD200(OX‐90), CD9 (KMC8), CCR9 (242503), pDC/IPC (120G8.04), FoxP3 (FJK‐16s), PDCA‐1(JF05‐1C2.4.1), B7‐DC (TY25) and B7‐H1 (MIH5). FoxP3 staining was performed using a commercial kit (FoxP3 Staining Kit; eBiosciences, San Diego, CA, USA), according to the manufacturer's instructions. Data were collected on an LSRII cytometer (BD Biosciences, San Diego, CA, USA) and analysed with FACSDiva (BD Biosciences) or FlowJo (Treestar, Ashland, OR, USA) software.

In‐vitro assays

Antibodies directed to the hLa antigen were measured by enzyme‐linked immunosorbent assay (ELISA) as described previously 11 from sera of recipient mice up to 20 days after adoptive transfer of 3B5.8 cells.

In‐vitro proliferation assays were performed as described previously with slight modifications 39. Briefly, B6.3B5.8^–.H‐2^k/k.Tcra^–/– donor cells were transferred into hLa Tg and non‐Tg mice as described above. Seven days post‐transfer, donor CD4⁺ T cells were isolated from recipients by positive selection with anti‐Thy1.2 MicroBeads followed by cell sorting as described above. T cell depleted splenocytes were irradiated (2200 rads) and used as APC. Co‐cultures containing a 1 : 4 T : APC ratio in the presence of 3·7 μM hLa 61–84 peptide were conducted for 72 h; 1 μCurie/well of [³H]‐thymidine was added for the last 18 h. Cultures were harvested onto glass fibre filters and counted by liquid scintillation. Supernatant from cell cultures was obtained at 48 h in order to measure IL‐10 (Invitrogen, now Life Technologies) and TGF‐β (eBiosciences, San Diego, CA, USA) by ELISA according to the manufacturers' protocols.

To determine DC antigen presentation capacity, cDC and pDC were purified from Flt3L‐induced Tg or non‐Tg mouse spleens as detailed above. Varying numbers of DC were cultured with 1 × 10⁵ 3C5.5 hLa 61‐84‐specific T cell hybridoma cells 40 or 3A9 HEL 46–61‐specific 3A9 hybridoma cells 41 in V‐bottomed 96‐well plates for 24 h. Culture supernatants were collected to measure IL‐2 production by ELISA (OptEIA ELISA Set; BD Biosciences).

Statistical analysis

Pairwise comparisons were evaluated using Student's t‐test. P‐values less than 0·05 were considered significant.

Results

The ubiquitous nuclear hLa neo‐self‐antigen induces expansion of antigen‐specific T_reg in the periphery of lymphocyte replete mice

To assess the post‐thymic fate of T cells specific for a representative RNA‐binding nuclear antigen, CFSE‐labelled 3B5.8 TCR⁺ hLa‐specific T cells from Thy 1.2⁺ B6.3B5.8^+/‐H‐2^k/bTcra^‐/‐ donor mice (referred to hereafter as 3B5.8 TCR Tg) were transferred adoptively into groups of H‐2^k/b Thy1.1^+/+ recipient mice expressing the hLa neo‐self antigen (hereafter referred to as hLa Tg). Recipients were Tg for the complete human gene for La, including its natural promotor, and expressed nuclear‐localized hLa ubiquitously at levels similar to the endogenous mouse La protein 11. Donor cells were recovered from secondary lymphoid organs of recipient mice 7 days post‐transfer and examined for proliferation status. As assessed by the fraction of cells diluting CFSE, donor CD4⁺ T cells divided extensively in hLa Tg (median 84·9% CFSE^low fraction, range 82·1–86.7%, n = 4) but not non‐Tg (median 7·1%, range 7·1–12·6%, n = 3) recipients by the 7‐day time point (Fig. 1a). Similar results were observed in H‐2^k/k recipients (Supporting information, Fig. S1).

Figure 1.

Anti‐human La (hLa)‐specific 3B5.8 T cells differentiate into forkhead box protein 3 (FoxP3)‐expressing cells in human La transgenic recipients. Representative (of four mice/group) flow cytometry plots depicting 5‐(and 6)‐carboxyfluorescein diacetate succinimidyl ester (CFSE) dilution (a) and regulatory T cell (T_reg) gating (b). Numbers in dot‐plots refer to quadrant percentages. (c) Graphical representation of T_reg as a frequency of the total donor cell population (left) or absolute cell numbers (right) obtained from individual recipient mice (four mice/group). Statistics calculated by Student's t‐test. *P < 0·05; **P < 0·01.

The hLa Tg recipient mice produced low, but detectable, levels of anti‐La antibodies 14 and 20 days post‐transfer (Supporting information, Fig. S2); total body pathological examination 20 days post‐transfer revealed no evidence of lymphoid cell infiltrates in any tissues or organs, and all blood cell counts were normal (data not shown). As these observations suggested potential control of autoreactivity by tolerogenic mechanisms, the potential development of peripheral T_reg in the model was evaluated. Donor splenocytes were transferred retro‐orbitally into groups of hLa Tg and non‐Tg recipient mice as discussed above. Seven days post‐transfer, splenocytes and lymph node cells were recovered from recipients and analysed by flow cytometry for the expression of CD25 and the transcription factor, FoxP3. Increased frequencies and absolute numbers of donor cells with CD25⁺FoxP3⁺ and CD25^–FoxP3⁺ T_reg phenotypes were recovered from hLa Tg recipient spleens (Fig. 1b,c) and lymph nodes (not shown) compared to those of non‐Tg recipient mice. Similar results were observed in H‐2^k/k and H‐2^k/b recipient mice; thus, H‐2^k/k and H‐2^k/b transfer models were used interchangeably in subsequent experiments. In summary, 3B5.8 donor T cells recognize and respond to the hLa antigen as identified by cellular proliferation and accumulation of cells with a T_reg phenotype.

To determine if the accumulation of T_reg in this model arises from de ‐ novo differentiation or marked expansion of small numbers of pre‐existing T_reg, donor mice expressing a red fluorescent protein (RFP)‐tagged FoxP3 locus 42 were employed. In this case total purified CD4⁺ T cells, which included a small percentage of contaminating RFP⁺ T_reg cells, or RFP^‐CD4⁺ (FoxP3^–) non‐T_reg cells purified (at ∼99% purity) from FoxP3‐RFP donor mice (Fig. 2a) were transferred into recipient mice as described above. Spleen cells recovered from recipients 7 days post‐transfer were analysed by flow cytometry for the frequency of donor RFP⁺ (FoxP3⁺) T cells. Both the bulk CD4⁺ and highly purified RFP^–CD4⁺ donor T cell populations were capable of generating RFP⁺ T_reg following transfer into hLa Tg recipient mice, while RFP⁺ cells were absent following transfer into non‐Tg recipients (Fig. 2b). Furthermore, more extensive amplification of T_reg cells occurred in hLa Tg recipients that received bulk CD4⁺ T cells compared to those that received sorted CD4 ⁺ RFP^– non‐T_reg cells (Fig. 2b). Therefore, the increased T_reg in hLa Tg recipients is due primarily to the expansion of pre‐existing T_reg, although de ‐ novo differentiation occurs to a lesser extent.

Figure 2.

hLa antigen‐specific T_reg accumulation is due primarily to the expansion of pre‐existing natural regulatory T cells (T_reg). (a) Post‐sort dot‐plots showing forkhead box protein 3 (FoxP3)‐monomeric red fluorescent protein (mRFP) expression from total CD4⁺ T cell sorts (left panel) and CD4⁺ RFP^– cell sorts (right panel); (b) 7 days after adoptive transfer, donor T cells were analysed in individual recipient mice (four mice/group for hLa transgenic (Tg) recipients and two mice for non‐Tg recipients) by flow cytometry for the frequency of FoxP3‐mRFP‐expressing T_reg. T_reg were observed in hLa‐expressing spleens receiving either bulk CD4⁺ cells or CD4⁺ RFP^– cells from 3B5·8⁺ donor mice. Statistics calculated by Student's t‐test. *P < 0·05.

To examine further the characteristics of hLa‐specific donor T cells after retrieval from hLa Tg mice, proliferation and production of immunosuppressive cytokines in response to in‐vitro antigenic challenge were assessed. Donor cells retrieved from recipient mice 7 days post‐transfer and challenged with hLa 61‐84 peptide in vitro exhibited defective proliferation and secreted IL‐10 (Fig. 3). TGF‐β1 was undetectable under similar conditions (data not shown).

Figure 3.

3B5.8 donor T cells retrieved from human La transgenic (Tg) recipients secrete interleukin (IL)‐10 and are hypo‐proliferative in vitro. Following adoptive transfer, donor T cells were purified from pooled (nine mice/group) recipient splenocytes and co‐cultured with irradiated antigen‐presenting cells (APC) in the presence of 3·7 μM of hLa 61–84 peptide for 72 h for measurement of cell proliferation by incorporation of [³H]‐thymidine (a) and 48 h for evaluation of IL‐10 in supernatants by enzyme‐linked immunosorbent assay (ELISA) (b). All values are expressed as mean ± standard deviation (s.d.) of triplicate wells and are representative of similar results from two independent experiments.

Plasmacytoid dendritic cells but not B lymphocytes promote expansion of hLa‐specific T_reg

B lymphocytes and pDC have been implicated selectively in pathogenic responses to RNA‐binding nuclear antigens such as La 43 and have also been shown to aid in the generation of peripheral T_reg. To determine if these cell types drive peripheral expansion of T_reg in the present model, recipient mice were depleted selectively of these cells throughout the duration of the experiment. Treatment of hLa Tg recipient mice with B cell‐ or pDC‐depleting antibodies resulted in elimination of B lymphocytes (Supporting information, Fig. S3) or pDC (Supporting information, Fig. S4), respectively. B lymphocytes were not required for the increased appearance of donor‐derived T_reg (Fig. 4a,b). In contrast, depletion of pDC impaired the development of hLa‐specific T_reg in hLa Tg recipient mice significantly (Fig. 4c,d). Thus, pDC but not B lymphocytes promote peripheral expansion of T_reg specific for a ubiquitous, RNA‐binding nuclear antigen.

Figure 4.

Plasmacytoid dendritic cell (pDC) but not B cell depletion reduces regulatory T cell (T_reg) accumulation. (a,c) Representative (of three to four mice/group) histograms showing forkhead box protein 3 (FoxP3) expression among donor T cells transferred into B cell‐depleted (MB20; (a) or pDC‐depleted [120G8; (c)] hLa transgenic (Tg) (Tg; four mice/group) or non‐Tg (three mice/group) recipient mice. (b,d) Frequencies of FoxP3⁺ donor T cells 7 days post‐transfer. Graphs show mean ± standard error of the mean (s.e.m.) of individual mice. Statistics were calculated by Student's t‐test. *P < 0·05.

To determine whether previously proposed tolerogenic (CD199/CCR9 32, CD200 44 and CD9 45), co‐inhibitory (CD274/B7‐H1/PDL‐1 and CD273/B7‐DC/PD‐L2) or co‐stimulatory (CD80 and CD86) markers were expressed selectively on pDC following adoptive transfer of hLa‐specific donor cells, recipient pDC (CD11c⁺CD45R/B220⁺CD317/120G8⁺CD317/PDCA1⁺) and B lymphocytes (CD45R/B220 ⁺CD11c^–), were characterized phenotypically 7 days post‐transfer. Compared to B cells, pDC exhibited increased expression of MHC class II, CD199/CCR9, CD200, CD9, CD80 and CD86, as well CD274/B7‐H1 and CD273/B7‐DC (Fig. 5). Compared to cDC (CD11c⁺CD45R/B220^–CD317/120G8^–CD317/PDCA1^–), pDC expressed lower levels of MHC class II and CD9 and increased expression of CD199/CCR9, CD200, CD274/B7‐H1and CD273/B7‐DC (Supporting information, Fig. S5).

Figure 5.

Cell surface phenotype of plasmacytoid dendritic cells and B cells following T cell adoptive transfer. Following adoptive transfer, recipient splenocyte B cell and pDC populations were analysed for expression of surface molecules. (a) Gating strategies distinguishing B cells (blue box) from pDC (red boxes). (b) Representative (of four mice/group) histograms showing expression of major histocompatibility complex (MHC) class II (MHCII), CCR9, CD200 and CD9 (top) and co‐stimulatory molecules CD80, CD86, B7‐H1 and B7‐DC (bottom). Numbers indicate mean fluorescence intensities (MFI) ± standard error of the mean (s.e.m.) of the respective cell surface proteins from individual mice (four mice/group), with text colours corresponding to flow cytometry gates depicted in (a). Statistics were calculated using Student's t‐test. *P < 0·05; **P < 0·01; ***P < 0·001; ****P < 0·0001.

hLa Tg pDC constitutively present the hLa 61‐84 epitope and autonomously induce 3B5.8 T_reg in vivo

We have shown previously that hLa Tg B lymphocytes constitutively present hLa T cell epitopes 11. To determine whether pDC and cDC constitutively present the hLa 61‐84 epitope in vivo in hLa Tg recipient mice, pDC and cDC from both hLa Tg and non‐Tg mice were sorted and co‐cultured with an I‐E^k‐restricted hLa 61–84‐specific T cell hybridoma in the absence of any exogenous hLa antigen. Following 24‐h co‐culture, IL‐2 production was measured from cell culture supernatants. Compared to those from microcultures containing non‐Tg DC, hLa‐specific T cells from microcultures containing hLa Tg cDC and, to a lesser extent, hLa Tg pDC, secreted increasing IL‐2 with increasing DC numbers (Fig. 6a), indicating constitutive presentation of the hLa 61–84 epitope by cDC and pDC in hLa Tg mice. In contrast, cDC and pDC isolated from hLa Tg mice failed to elicit responses from the 3A9 T cell hybridoma that is specific for an I‐A^k‐restricted epitope of the irrelevant HEL antigen.

Figure 6.

Plasmacytoid dendritic cells constitutively present hLa antigen to T cells. (a,b) conventional DC (cDC) and pDC sorted from hLa‐transgenic (Tg) and non‐Tg mice were co‐cultured with the hLa‐specific 3C5.5 T cell hybridoma or the irrelevant hen egg‐white lysozyme (HEL)‐specific 3A9 hybridoma for 24 h as indicated. Interleukin (IL‐2) in the supernatant was measured by enzyme‐linked immunosorbent assay (ELISA). Values are the mean ± standard deviation (s.d.) of triplicate wells. Similar results were observed in three independent experiments. (c) Peptide‐pulsed non‐Tg pDC and 3B5.8 splenocytes were transferred adoptively into non‐Tg B6.H‐2^k/k.Thy1.1 ^+/+ recipient mice on consecutive days. Seven days post‐transfer, spleen cells were recovered from recipient mice for analysis of T_reg development. Representative (of four mice/group) histograms show forkhead box protein 3 (FoxP3) expression within donor T lymphocytes from mice receiving hLa peptide‐loaded pDC (left) or HEL peptide‐loaded pDC (right). (d) Frequencies of FoxP3‐expressing T_reg cells from individual recipient mice (four mice/group) after adoptive transfer. Graphs show mean ± standard error of the mean (s.e.m.). Statistics were calculated by Student's t‐test. **P < 0·01.

To determine whether hLa peptide‐loaded pDC alone from non‐Tg mice could induce 3B5.8 T_reg expansion in the absence of endogenously expressed hLa antigen in vivo, highly purified pDC were loaded with either hLa 61–84 peptide or irrelevant control I‐A^k‐restricted HEL 46–61 peptide and transferred subsequently into non‐Tg recipient mice. Peptide loading was confirmed by the capacity of peptide‐loaded DC to specifically stimulate the relevant hybridomas (Supporting information, Fig. S6). The following day, cells from 3B5.8⁺ donor mice were also transferred into recipient mice. Seven days after donor T cell transfer, spleens were harvested and donor 3B5.8 T_reg were assessed by flow cytometry as in earlier experiments. pDC loaded with hLa peptide, but not pDC loaded with control peptide, induced accumulation of 3B5.8 TCR‐expressing T_reg (Fig. 6c,d). Thus, pDC not only present the hLa antigen constitutively in hLa Tg mice, but can also promote the generation of antigen‐specific T_reg when provided as the sole hLa 61–84 peptide‐presenting cell in vivo.

Type 1 IFN enhances hLa‐specific T_reg expansion

pDC are specialized type I IFN‐producing cells 43. In the present model, pDC are also important mediators in promoting self‐antigen‐specific T_reg. To determine whether type I IFN contributes to the enhanced generation of hLa‐specific T_reg in the periphery of hLa Tg mice, hLa Tg recipients were treated with a blocking monoclonal antibody specific for the IFN‐α/β (IFN‐α/β) receptor, IFNAR1, throughout the T cell adoptive transfer experiment. Numbers of hLa‐specific T_reg were reduced significantly among anti‐IFNAR1 antibody‐treated hLa Tg recipient mice compared to those that had been treated with an isotype control antibody (Fig. 7). Therefore, type I IFN promotes accumulation of donor hLa‐specific FoxP3⁺ regulatory T cells in the periphery of hLa Tg recipient mice.

Figure 7.

Type I interferon (IFN) is important for induction of a forkhead box protein 3 (FoxP3⁺) T_reg phenotype in hLa‐specific 3B5.8 donor cells. Recipient mice were treated with the IFN alpha receptor 1 (IFNAR1) blocking monoclonal antibody mouse IFNAR1 (MAR1)‐5A3 or a matched isotype control antibody before and during adoptive transfer. Seven days post‐transfer, donor T cells were examined by flow cytometry to assess T_reg development. (a) Representative (of four mice/group) histograms of FoxP3 expression among donor T cells obtained from recipient mice treated with MAR1‐5A3 antibody (left) or isotype control antibody (right). (b) Frequencies (left) and absolute numbers (right) of T_reg among donor T cells (four mice/group). Graphs show mean ± standard error of the mean (s.e.m.) of individual mice. Statistics were calculated by Student's t‐test. **P < 0·01.

Discussion

RNA‐binding antigens are key targets of the autoimmune response in several rheumatic diseases, a phenomenon that is thought to arise from the capacity of RNA‐binding nuclear antigen‐containing immune complexes to stimulate RNA sensors in accessory cells such as B lymphocytes and pDC 43. Mechanisms of normal immunological tolerance to this class of self‐antigen have not been well studied. Moreover, although T_reg are recognized as critical for the control of autoimmune disease, mechanistic details of their regulation are still incomplete. We observed recently that thymic clonal deletion and development of thymic T_reg are important mechanisms of tolerance to the hLa neo‐self‐antigen that act to prevent high titres of anti‐hLa autoantibodies, cellular autoimmunity and pulmonary pathology 13. Herein we investigated the regulation of hLa antigen‐specific T_reg in the periphery of mice using an adoptive transfer system.

The data presented show that a representative, ubiquitously expressed RNA‐binding nuclear antigen can elicit antigen‐specific T_reg in the periphery of mice. Antigen exposure without concurrent inflammation may promote conversion of FoxP3^– T cells into FoxP3⁺ T_reg 46. Although there was evidence for de‐novo T_reg differentiation in the hLa‐specific donor T cell/hLa Tg recipient adoptive transfer model, the major effect was expansion of pre‐existing T_reg. In the present study, donor T cells proliferated in recipient mice and induced low, but detectable, levels of autoantibodies directed to the hLa neo‐self‐antigen. Using adoptive transfer of hLa 61–84 peptide‐primed polyclonal T cells, we established previously that non‐tolerant donor cells can, at least transiently, elicit autoantibodies in hLa Tg recipient mice, while CD4⁺ T cells from tolerant hLa Tg donor mice cannot 11. The La autoantibodies observed in the present study were of lower titre at all time‐points compared to the prior study, despite transfer of much higher numbers of hLa‐specific donor T cells. We speculate that the low autoantibody levels and lack of autoimmune pathology following the transfer of large numbers of hLa‐specific T cells from TCR Tg donor mice were limited by the observed expansion of T_reg. This thesis is supported by our recent study showing that impairment of thymic T_reg differentiation and T_reg function in a hLa/TCR double Tg mouse model led to high titre hLa autoantibody responses and subsequent autoimmune lung pathology 13. Understanding the mechanisms of RNA‐binding nuclear antigen‐specific T_reg accrual may permit the design of rheumatic disease‐targeted therapies that can preserve this beneficial regulatory response selectively while simultaneously curtailing pathogenic cell subsets.

The pDC and B lymphocyte subsets are major pathogenic cell types in lupus and related disorders such as SS and are the targets of existing or developing therapies 47, 48. The present studies show clearly that pDC but not B cells are required for efficient peripheral T_reg expansion, as their elimination in recipient mice hampered hLa‐specific T_reg accumulation severely in the adoptive transfer model. The mechanism of pDC expansion involved antigen presentation, as hLa peptide‐loaded pDC alone could elicit hLa‐specific T_reg expansion in vivo in the absence of any other hLa antigen source. Moreover, pDC from hLa Tg recipient mice presented endogenous hLa peptide constitutively, but at a lower level than conventional DC as measured by the readout of hLa 61–84 peptide‐specific hybridoma stimulation and by cell surface MHC class II levels. This is consistent with high rates of MHC class II turnover among activated pDC that keeps presentation of specific antigens low 49 and with induction of T_reg by low peptide concentrations 31.

Cell surface phenotyping of recipient T_reg‐inducing pDC revealed expression of tolerogenic pDC markers reported by others, as well as enhanced expression of both co‐stimulatory and co‐inhibitory markers, all of which could contribute mechanistically to T_reg expansion. In particular, pDC in our model expressed high levels of CD199/CCR9, reported to be a marker for tolerogenic, immature DC that control graft‐versus‐host disease by enhancing T_reg function 32. Tolerogenic pDC in our model also expressed CD200, which can induce T_reg through an IDO‐dependent mechanism 44. Interestingly, only activated pDC can secrete IDO 44. Low levels of CD9 were reported previously to mark tolerogenic immature pDC that produced negligible levels of IFN‐α, while CD9⁺ pDC were immunostimulatory and secreted IFN‐α in response to Toll‐like receptor ligation 45. In the present study, the tolerogenic pDC expressed CD9 levels that were intermediate between B cells and cDC, consistent with a phenotype permissive for IFN‐α secretion. Although the role of immature DC in inducing T_reg is established, mature DC expressing high levels of co‐stimulatory molecules can also expand T_reg preferentially 50, 51. Moreover, B7‐H1 has been implicated in T_reg expansion in the settings of infection 52, 53 and allergy 54. A similar role for CD273/B7‐DC can be hypothesized, as these two ligands share the same receptor. Taken together, the phenotype of tolerogenic pDC in hLa Tg recipient mice is consistent with recently activated immature pDC. Further studies will be required to explore systematically the relevance of individual surface markers to the expansion of T_reg specific for ubiquitous RNA‐binding nuclear antigens.

Although elevated type 1 IFN levels are pathogenic in SLE and related disorders, we observed that the mechanism of expansion of T_reg specific for the representative RNA‐binding nuclear antigen La was promoted by type 1 IFN. Thus, blockade of the type 1 IFN receptor IFNAR1 hampered T_reg accumulation of hLa‐specific T cells severely after transfer into hLa Tg mice in this tolerogenic model. The present studies have not established the cellular source of T_reg‐promoting type 1 IFN in the hLa adoptive transfer model; however, it is reasonable to speculate that the source may be pDC. CD200 can modulate type 1 IFN secretion by myeloid cells 55. Thus, elevated CD200 expression and intermediate levels of CD9 45 may mark pDC with the capacity to produce tolerogenic, non‐pathogenic levels of type 1 IFN that promote T_reg accumulation. While the present paper was in preparation, Metidji and colleagues used IFNAR1‐deficient mice to show that type 1 IFN promotes development of thymic T_reg and survival of T_reg in the periphery of mice in a competitive or stress environment 56.

We conclude that pDC and type 1 IFN are important mechanisms controlling accumulation of RNA‐binding nuclear antigen‐specific T_reg in the periphery. These results inform prophylactic and therapeutic approaches for preventing and treating rheumatic autoimmune diseases wherein B cell depleting therapies may spare T_reg, while IFN‐α blockade may inhibit not only pathogenic responses but also beneficial T_reg expansion.

Disclosure

The authors have no competing interests to disclose.

Author contributions

A. D. F. and Z. P. designed the study. Z. P. and C. L. performed the experiments. Z. P., C. H. and A. D. F. analysed the data. C. H., Z. P. and A. D. F. wrote the paper.

Supporting information

Additional Supporting information may be found in the online version of this article at the publisher's web‐site:

Fig. S1. hLa‐specific T cells from 3B5.8 T cell receptor (TCR) transgenic (Tg) mice divide upon adoptive transfer into hLa Tg recipient mice. 5‐(and 6)‐Carboxyfluorescein diacetate succinimidyl ester (CFSE)‐labelled 3B5.8 splenocytes from B6.3B5.8^+/–.H‐2^k/k donors were transferred retro‐orbitally into B6.H‐2^k/k hLa Tg (four mice) or non‐Tg (four mice) recipient mice. Recipients in this experiment were not congenic for Thy1.1. Splenocytes recovered from recipient mice 24 h, 48 h and 5 days (d) post‐transfer were analysed by flow cytometry. Histograms show CFSE dilution of CD4⁺Vβ10⁺ cells (both unlabelled recipient cells and CFSE‐labelled donor cells are visible). Percentages of undivided donor cells are shown.

Fig. S2. Anti‐La immunoglobulin (Ig)G antibody titres from recipient mice at 0, 14 and 20 days following adoptive transfer of hLa‐specific T cells. IgG antibody titres to (a) recombinant 6xhis‐hLa or (b) irrelevant recombinant 6xhis‐protective antigen (PA) from Bacillus anthracis. Each symbol represents the response of an individual mouse. Filled symbols represent hLa transgenic (Tg) recipients (three mice), open symbols represent non‐Tg recipients (four mice).

Fig. S3. Treatment with anti‐CD20 antibody (MB20) ablates CD19⁺ cells. hLa transgenic (Tg) (four mice/group) or non‐Tg (three) recipient mice were treated with B cell‐depleting monoclonal antibody (MB20‐11) or isotype control immunoglobulin (Ig)G prior to and during the duration of adoptive transfer. (a) Representative flow cytometry profiles of B cells from MB20‐treated (left) or isotype control‐treated (middle) hLa Tg recipient mice and untreated non‐Tg recipient mice (right). (b) Frequencies of CD19⁺ B cells from recipient mice. (c) Absolute numbers of CD19⁺ B cells from recipient mice. Graphs show mean ± standard error of the mean (s.e.m) of individual mice. Statistics were calculated by Student's t‐test. ****P < 0·0001.

Fig. S4. Treatment with 120G8 antibody depletes plasmacytoid dendritic cells (pDC). hLa transgenic (Tg) (four mice/group) or non‐Tg (three) recipient mice were treated with pDC‐depleting monoclonal antibody 120G8 or isotype control immunoglobulin (Ig)G before and during adoptive transfer of donor cells. (a) Representative gating and flow profiles of pDC from 120G8 antibody treated (left) or isotype control treated (middle) hLa Tg recipients and untreated non‐Tg recipients (right). Plots were gated on CD11c⁺ cells. (b) Percentages of pDC from recipient mice. (c) Absolute numbers of pDC from recipient mice. Graphs show mean ± standard error of the mean (s.e.m) of individual mice. Statistics were calculated by Student's t‐test. ****P < 0·0001.

Fig. S5. Expression of cell surface molecules on recipient plasmacytoid dendritic cell (pDC) and conventional DC (cDC) populations following T cell adoptive transfer. Splenocytes were recovered from recipient mice 7 days post‐transfer and analysed subsequently for cell surface markers. (a) Gating strategies distinguishing pDC (top; red box) from cDC (bottom; blue box). (b) Representative histograms showing expression of major histocompatibility complex (MHC) class II (MHCII), CCR9, CD200 and CD9 (top) and co‐stimulatory molecules CD80, CD86, B7‐H1 and B7‐DC (bottom). Numbers indicate mean fluorescence intensity (MFI) of the respective cell surface proteins, where colours are representative of the flow cytometry gates depicted in (a). Cells were pre‐gated on MHC class II⁺ cells for the evaluation of MHC class II MFI values. Black line indicates isotype control from CD11c⁺B220^– cells. Statistics were calculated using Student's t‐test. *P < 0·05; **P < 0·01; ***P <0·001; ****P < 0·0001.

Fig. S6. Verification of peptide loading onto major histocompatibility complex (MHC) class II of plasmacytoid dendritic cells (pDC). Sorted pDC from Fms‐like kinase 3 ligand (Flt3L)‐induced wild‐type B6 mice congenic for H‐2^k (isolated from splenocytes pooled from two to three mice/group) were loaded with 10 μM of hLa 61–84 peptide (a) or hen egg‐white lysozyme (HEL) 46–61 peptide (b) in vitro for 2 h, washed twice, then tested (8 × 10³ peptide‐loaded DC per well) for the capacity to stimulate the 3C5.5 hLa 61–84‐specific T cell hybridoma or the 3A9 HEL 46–61‐specific T cell hybridoma (1 × 10⁵ hybridoma cells/well). In comparison, direct stimulation of 3C5.5 (c) and 3A9 (d) with peptides in the presence of irradiated splenic antigen‐presenting cells (APC) (3 × 105 cells/well) demonstrates the greater inherent responsiveness of the 3C5.5 hybridoma compared to the 3A9 hybridoma for specific antigenic peptide. All results are the mean ± standard deviation (s.d.) of triplicate wells.