Abstract
Regulatory T cells hold promise as targets for therapeutic intervention in autoimmunity, but approaches capable of expanding antigen-specific regulatory T cells in vivo are currently not available. Here we show that systemic delivery of nanoparticles coated with autoimmune-disease-relevant peptides bound to major histocompatibility complex class II (pMHCII) molecules triggers the generation and expansion of antigen-specific regulatory CD4+ T cell type 1 (TR1)-like cells in different mouse models, including mice humanized with lymphocytes from patients, leading to resolution of established autoimmune phenomena. Ten pMHCII-based nanomedicines show similar biological effects, regardless of genetic background, prevalence of the cognate T-cell population or MHC restriction. These nanomedicines promote the differentiation of disease-primed autoreactive T cells into TR1-like cells, which in turn suppress autoantigen-loaded antigen-presenting cells and drive the differentiation of cognate B cells into disease-suppressing regulatory B cells, without compromising systemic immunity. pMHCII-based nanomedicines thus represent a new class of drugs, potentially useful for treating a broad spectrum of autoimmune conditions in a disease-specific manner.
Main
Autoimmune diseases such as type 1 diabetes (T1D), multiple sclerosis and rheumatoid arthritis result from chronic autoimmune responses involving T cells and B cells recognizing numerous antigenic epitopes on incompletely defined lists of autoantigens1,2,3. Eliminating or suppressing all polyclonal autoreactive T-cell specificities (known and unknown) in each individual autoimmune disorder without compromising systemic immunity is not currently possible.
Adoptive transfer of polyclonal FOXP3+CD4+CD25+ regulatory T (Treg) cells expanded ex vivo has been proposed as an alternative therapeutic approach4. The potential for bystander immunosuppression, the lack of effective strategies for expanding antigen-specific Treg cells in vitro, and the lineage instability of FOXP3+ Treg cells, have hindered the clinical translation of this approach5,6,7. TR1 FOXP3−CD4+CD25− T cells, which produce the cytokines IL-10 and IL-21, and express the surface markers CD49b and LAG-3 and the transcription factor c-Maf 8, constitute another regulatory T-cell subset recently exploited for the treatment of human inflammatory diseases9,10,11. However, as with FOXP3+ Treg cells, there are no pharmacological approaches that can expand autoantigen- or disease-specific TR1-like cells in vivo.
We recently discovered that systemic delivery of nanoparticles (NPs) coated with T1D-relevant pMHC class I complexes could blunt the progression of T1D by expanding subsets of CD8+ T cells with regulatory potential but conventional memory-like phenotype12. As this outcome could be replicated with different pMHC class I complexes, we reasoned that pMHC–NP therapy may utilize a naturally occurring negative feedback regulatory loop, whereby chronic autoantigenic exposure (and exposure to pMHC–NPs) would trigger the differentiation of autoreactive T cells into regulatory T-cell progeny. By this reasoning, we predicted that NPs coated with disease-relevant pMHCII complexes might be able to expand disease-specific regulatory CD4+ T cells in vivo.
Expansion of disease-specific TR1 cells
We treated non-obese diabetic (NOD) and NOD Foxp3-eGFP mice (expressing enhanced green fluorescent protein (eGFP) under the control of the mouse Foxp3 promoter) with uncoated nanoparticles or nanoparticles coated with a pMHC, 2.5mi/IAg7 (ref. 13), recognized by the diabetogenic BDC2.5-specific T-cell receptor (TCR), or with 2.5mi/IAg7 monomers. Nanoparticles coated in 2.5mi/IAg7 induced expansion of cognate CD4+ T cells in blood and spleens of all mice (Fig. 1a, b). These cells had a memory-like (CD44hiCD62Llow) FOXP3− TR1-like phenotype, expressing ICOS, latent-associated TGF-β and the TR1 markers CD49b and LAG-3 (Fig. 1c, d). A similar outcome was observed in mice treated with 2.5mi/IAg7–NPs upon depletion of CD4+CD25+ T cells (Extended Data Fig. 1a). Unlike their tetramer− counterparts, these cells proliferated and secreted IL-10 and to a lesser extent IFNγ, but not IL-2, IL-4 or IL-17, in response to dendritic cells (DCs) pulsed with the 2.5mi peptide (Extended Data Fig. 1b). Real-time reverse-transcription (RT)–PCR analyses confirmed the TR1-like phenotype of these cells (Supplementary Table 1).
a, b, Tetramer-staining profiles (a) and percentages of tetramer+CD4+ T cells (b). Data correspond to pre-diabetic NOD females treated for 5 weeks (blood: n = 5, 8 and 6; spleen: n = 5, 18 and 6, respectively). Tet, tetramer. c, Tetramer-staining of splenic CD4+ T cells from treated or untreated NOD Foxp3-eGFP mice. d, The tetramer+CD4+ T cells of 2.5mi/IAg7–NP-treated mice display a TR1-like phenotype. e, Incidence of diabetes in T-cell-reconstituted NOD scid hosts transfused with CD4+ T cells from different donors ± 2.5mi/IAg7–NPs (n = 11, 5, 7 and 6 from top). f, Percentages of tetramer+CD4+ T cells in 2.5mi/IAg7–NP-treated or untreated NOD scid hosts (n = 4–5 per group). g, Incidence of disease reversal in diabetic mice treated with pMHC–NPs (n = 9, 7, 7, 7 from top left to right), or IGRP4–22 peptide16 and IGRP4–22 peptide–NP (n = 9). h, Percentage of tetramer+CD4+ T cells in diabetic mice at onset, in response to 2.5mi/IAg7–NP therapy and age-matched non-diabetic controls (n = 8, 6, 2 and 7 from left). i, Insulitis scores (n = 6, 4, 3 and 6 from left). Bottom, representative images. j–m, C57BL/6 mice were immunized with pMOG35–55. j, EAE scores of mice treated from day 14 (n = 4 each). k, EAE scores of mice treated from day 21 (n = 10, 7 and 3 from top). l, Percentage of tetramer+CD4+ T cells in spleen and blood of mice from j and k (n = 13, 14 and 5 from top). m, Representative flow profiles of CD4+ T cells from mice in j and k. n, Representative microglial IBA1 stainings and relative rank scores in the cerebellum of mice from k (n = 4–5). P values were calculated via Mann–Whitney U-test, log-rank (Mantel–Cox) test or two-way ANOVA. Error bars, s.e.m.
To determine if pMHCII–NPs could directly trigger TR1 marker and IL-10 expression on cognate CD4+ T cells, we cultured naive and anti-CD3 plus anti-CD28 monoclonal antibody (mAb)-preactivated 2.5mi/IAg7-tetramer+CD4+ T cells from BDC2.5-TCR-transgenic NOD Foxp3-eGFP or NOD Il10GFP mice (carrying an eGFP insertion in the Il10 locus)14 in the presence of 2.5mi/IAg7–NPs, 2.5mi peptide or 2.5mi/IAg7 monomer. Naive T cells expressed neither CD49b nor LAG-3, even after incubation with 2.5mi/IAg7–NPs, 2.5mi/IAg7 monomer or 2.5mi peptide (Extended Data Fig. 1c, d). However, preactivated T cells upregulated both markers as well as eGFP (IL-10) only in response to 2.5mi/IAg7–NPs (Extended Data Fig. 1d, e). In agreement with this, expression of IL-10 in NOD Il10GFP mice treated with 2.5mi/IAg7–NPs was largely restricted to the CD49b+LAG-3+CD4+ subset (Extended Data Fig. 1f).
In vitro, the tetramer+CD4+ T cells of pMHC–NP-treated mice suppressed the proliferation of non-cognate (islet-specific glucose-6-phosphatase catalytic subunit-related protein (IGRP)- or LCMV Gp33-specific) CD8+ T cells in response to peptide-pulsed DCs, in an IL-10- and TGF-β-dependent manner (Extended Data Fig. 1g). In vivo, splenic CD4+ T cells from donors treated with pMHC–NPs suppressed diabetes development in T-cell-reconstituted NOD scid (also known as NOD Prkdcscid) hosts (Fig. 1e), an effect that was potentiated by treating hosts with pMHC–NPs (Figs 1e, f).
We next investigated whether 2.5mi/IAg7–NPs or NPs coated with IGRP4–22/IAg7 or IGRP128–145/IAg7, targeted by sub-dominant pools of autoreactive CD4+ T cells15, could restore normoglycaemia in diabetic NOD mice. Unlike mice treated with nanoparticles coated with hen egg-white lysozyme (HEL)14–22/IAg7, 90–100% of the mice that received nanoparticles coated with 2.5mi/IAg7, IGRP4–22/IAg7 or IGRP128–145/IAg7 reverted to stable normoglycaemia (Fig. 1g, Extended Data Fig. 1h) and displayed systemic expansion of cognate TR1-like T cells (Fig. 1h, Extended Data Fig. 2a–g). Treatment with peptide16 or peptide-coated nanoparticles but without MHC could not reproduce any of these effects (Fig. 1g, Extended Data Figs 1h and 2h). Treatment withdrawal resulted in loss of the normoglycaemic state in 25–60% of mice (Extended Data Fig. 1i), in association with the loss of the tetramer+CD4+ T-cell pools (Fig. 1h, Extended Data Fig. 2a). The animals that maintained normoglycaemia had normal postprandial serum insulin levels, fasting glucose tolerance (Extended Data Fig. 1j–m) and reduced insulitis (Fig. 1i). In addition, their pancreatic lymph nodes (PLNs) could not support the proliferation of carboxyfluorescein succinimidyl ester (CFSE)-labelled IGRP206–214/Kd-specific CD8+ T cells in vivo (Extended Data Fig. 1n).
We next tested the ability of nanoparticles coated with myelin oligodendrocyte glycoprotein (pMOG)38–49/IAb to blunt the progression of pMOG35–55-induced experimental autoimmune encephalomyelitis (EAE, a model of multiple sclerosis) in C57BL/6 mice. pMOG38–49/IAb–NP therapy dampended disease progression when given on day 14 after immunization and restored motor function in paralytic mice when given on day 21 (Fig. 1j, k). These effects were mirrored by weight gain, and were associated with systemic expansion of cognate TR1-like T cells, reductions in activated macrophage/microglia in the cerebellum, fewer inflammatory foci and areas of demyelination in the white matter of the cerebellum and decreased demyelination of the spinal cord (Fig. 1l–n, Extended Data Figs 3a–f). Similar therapeutic effects were seen in HLA-DR4-IE-transgenic C57BL/6 IAbnull mice (MHCII knockout mice expressing a transgenic hybrid MHCII molecule composed of the peptide-binding domain of human HLA-DR4 and the membrane-proximal domain of mouse IE (DR4-IE)) immunized with human (h) proteolipid protein (hPLP)175–192 or hMOG97–108 peptides and treated with hPLP175–192/DR4-IE or hMOG97–108/DR4-IE–NPs upon developing limb paralysis (Extended Data Figs 4a–d).
Disease and organ specificity
Studies in another autoimmune disease model, collagen-induced arthritis (CIA), showed that nanoparticles displaying mouse collagen (mCII)259–273/DR4-IE could reduce joint inflammation in arthritic HLA-DR4-IE-transgenic C57BL/10.M mice in association with systemic expansions of cognate TR1-like T cells (Figs 2a–e, Extended Data Fig. 4e). In contrast, nanoparticles coated with hMOG97–108/DR4-IE complexes had no effect (Fig. 2a–c)
a–f, C57BL/10.M HLA-DR4-IE-transgenic mice immunized with bovine collagen. a, Left, changes in joint swelling (top) and clinical scores (bottom) in response to uncoated NPs, pMHC–NPs, peptide s.c.16 or peptide-coated MPs i.v17. Treatment was initiated when joint swelling reached 130% of baseline (data normalized to the initiation of treatment (100% value)) (n = 4, 4, 4 and 8 from top). Right, percentage increase in joint swelling relative to pre-immunization baseline (100% value). b, Representative haematoxylin and eosin (first row) and O-safranin/fast-green/haematoxylin (second and third rows) knee joint staining images. Third row shows enlarged images of lacunae on the bone and meniscal articular surfaces. 1, panus formation; 2, cellular infiltration of the meniscus; 3, bone erosion; 4, proteoglycan depletion; 5, loss of chondrocyte/lacunnae. c, Average pathology scores (n = 3–4 per group). d, Percentage of tetramer+CD4+ T cells. e, Representative flow cytometry profiles for TR1 markers in mCII259–273/DR4–NP-treated. f–h, C57BL/6 IAbnull HLA-DR4-IE-transgenic mice immunized with hPLP. f, Changes in EAE scores ((n = 5, 4, 13 (4–9 per group), 5, 19 (4–5 per group, see also Extended Data Fig. 4h), 4 and 5 from top). g, Percentage of tetramer+CD4+ T cells in the spleen of mice from f (n = 4, 5, 4, 6, 15, 3 and 3 from left). h, Representative flow cytometry profiles for TR1 markers. Data were compared using Mann–Whitney U-test or two-way ANOVA. Error bars, s.e.m.
To investigate further the disease-specificity of pMHC–NP therapy, we induced EAE in C57BL/6 IAbnull HLA-DR4-IE-transgenic mice by immunization with hPLP175–192 and treated diseased mice with hPLP175–192/DR4-IE–NPs (positive control), uncoated nanoparticles (negative control), EAE-relevant hMOG97–108/DR4-IE–NPs (which display a peptide from a CNS autoantigen other than that used to induce disease), or CIA-relevant mCII259–273/DR4-IE–NPs. Whereas hMOG97–108/DR4-IE–NPs blunted EAE as efficiently as the positive control, mCII259–273/DR4-IE–NPs had no therapeutic activity (Fig. 2f, Extended Data Fig. 4f, g). Here, too, therapeutic activity was associated with systemic expansions of cognate TR1-like T cells (Fig. 2g, h). Administration of mCII259–273 peptide16 or of mCII259–273-peptide-coated microparticles (MPs)17 to arthritic C57BL/10.M HLA-DR4-IE-transgenic mice (Fig. 2a–d, Extended Data Fig. 4e), or of hMOG97–108 peptide, hMOG97–108/DR4-IE monomers or hMOG97–108-coated nanoparticles or microparticles to C57BL/6 IAbnull HLA-DR4-IE-transgenic mice failed to both expand cognate TR1-like cells and blunt disease (Fig. 2f, g, Extended Data Fig. 4f–i). Thus, the biological and therapeutic effects of pMHCII–NPs are disease-specific and dissociated from the pathogenic role of epitopes (disease-triggering versus downstream autoantigenic targets), suggesting that these compounds act on pre-activated autoreactive T cells and can generate TR1-like cell expansions from rare T-cell precursor pools.
Soluble mediators
Blockade of IL-10, TGF-β and IL-21R (but not IFNγ) abrogated the anti-diabetogenic properties of 2.5mi/IAg7–NPs or IGRP4–22/IAg7–NPs (Fig. 3a, Extended Data Fig. 5a). With the exception of IL-21R blockade (known to inhibit CD8+ T-cell activation), these interventions also abrogated the suppression of autoantigen crosspresentation by the pMHC–NP-expanded TR1-like T cells in the PLNs (Extended Data Figs 5b, c, 6). Studies using diabetic NOD Ifng−/− and NOD Il10−/− mice revealed that development of the TR1 precursors and/or TR1-like cells that expand in response to therapy requires IFNγ in addition to IL-10 (Extended Data Figs 5d–f, 6). mAbs against IL-10, TGF-β and IL-21R also abrogated the anti-encephalitogenic activity of hPLP175–192/DR4-IE–NPs in C57BL/6 IAbnull HLA-DR4-IE-transgenic mice (Extended Data Fig. 5g–j). pMOG35–55-immunized C57BL/6 Il27r−/− mice responded to pMOG38−49/IAb−NPs like their wild-type counterparts (Fig. 1j–n, Extended Data Fig. 5k–n). Thus, IFNγ and IL-10, but not IL-2718, are necessary for pMHC–NP-induced TR1-like cell development; and autoreactive TR1-like T cells use IL-10, TGF-β and IL-21 (but not IFNγ) to suppress disease.
a, Blood glucose levels in diabetic NOD mice treated with 2.5mi/IAg7–NPs and blocking antibodies (n = 8, 4, 6, 6, 5 and 4 from top to right). b, Expression of IL-10 (eGFP) and upregulation of CD5 and CD1d by eGFP− 2.5mi-pulsed splenic B cells from NOD Il10GFP donors in 2.5mi/IAg7–NP-treated NOD hosts. c, Averaged results from b (n = 4, 3, 3 and 7 from left). d, Incidence of diabetes in T-cell-reconstituted NOD scid hosts left alone or transfused with PLN CD19+ cells (n = 7, 13 and 7 from top). e, Incidence of diabetes in T-cell-reconstituted NOD scid hosts transfused with CD19+ and/or CD4+ cells (n = 7, 6, 3, 7, 8, 11 and 13 from top). f, Cytokine and chemokine profiles of PLN and MLN CD11b+ cells from 2.5mi/IAg7–NP-treated NOD mice in response to LPS (n = 3–4 each). g, Percentage of tetramer+CD4+ T cells in the spleens (left), and viral titres in the ovaries (right) of treated compared with untreated NOD mice 4 and 14 days after vaccinia virus infection (n = 3 per group). h, Percentages of tetramer+CD4+ T cells in the spleens (left) and serum anti-dinitrophenyl (DNP) antibody titres (right) in treated and untreated NOD mice immunized with keyhole limpet haemocyanin (KLH)–DNP (n = 3–5 per group). Data were compared using Mann–Whitney U-test, log-rank test or two-way ANOVA. Error bars, s.e.m.
Downstream effectors and network formation
The PLNs (but not the mesenteric lymph nodes (MLNs) or spleens) of pMHC–NP-treated NOD mice harboured increased percentages of B cells compared with the PLNs of mice treated with pMHCII–NPs not relevant for T1D (Extended Data Fig. 5o). Studies of mice treated with a range of pMHC–NP doses revealed that the sizes of the PLN (but not splenic) B-cell and TR1-like cell pools were correlated (Extended Data Fig. 5p). Unlike their splenic or MLN counterparts, the PLN B cells of these mice could not effectively present peptide to cognate CD4+ T cells ex vivo (Extended Data Fig. 5q). In addition, these cells produced IL-10 in response to lipopolysaccharide (LPS) (Extended Data Fig. 5r), suggesting that pMHC–NP-induced TR1-like cells might trigger the formation and expansion of regulatory B (Breg) cells in the PLNs. In fact, 2.5mi-pulsed B cells, but not DCs, underwent expansion in 2.5mi/IAg7–NP-treated hosts within a week of transfer (Extended Data Fig. 5s, t).
To probe this further, we transfused NOD Il10GFP splenic B cells that were either pulsed with 2.5mi or a negative control peptide (GPI282–292), into 2.5mi/IAg7–NP-treated NOD or NOD Il10−/− hosts. Seven days later, the hosts were analysed for the presence of IL-10-producing (eGFP+) CD5+CD1dhigh B cells. NOD (but not NOD Il10−/−) mice treated with 2.5mi/IAg7–NPs efficiently induced formation of Breg cells specifically from 2.5mi-pulsed B cells, and IL-21R but not IL-10 or TGF-β blockade suppressed this effect (Fig. 3b, c, Extended Data Fig. 5u).
In vitro, the PLN B cells of 2.5mi/IAg7–NP-treated mice had a moderate suppressive effect on the proliferative activity of BDC2.5 CD4+ T cells in response to peptide-pulsed DCs (Extended Data Fig. 5v). In vivo, these B cells suppressed diabetes development in T-cell-reconstituted NOD scid hosts as compared to PLN B cells from control mice (Fig. 3d). Co-transfer of PLN B cells and bulk or 2.5mi/IAg7-tetramer+ splenic CD4+ T cells from 2.5mi/IAg7–NP-treated mice resulted in >95% suppression, as compared to PLN B cells with or without tetramer–CD4+ T cells from 2.5mi/IAg7–NP-treated mice, to CD4+ T cells with or without MLN B cells from 2.5mi/IAg7–NP-treated mice (~40%), or to CD4+ T cells from untreated or control–NP-treated mice (0%) (Fig. 3e), supporting synergistic effects. In agreement with this, treatment of newly diabetic NOD mice with a B-cell-depleting anti-CD20 mAb abrogated the anti-diabetogenic activity of 2.5mi/IAg7–NPs (Fig. 3a, Extended Data Fig. 5x).
Comparison of the cytokine and chemokine profile of CD11b+ cells derived from the PLN or MLN of pMHC–NP-treated NOD mice further revealed that CD11b+ cells from the PLN produced lower levels of the pro-inflammatory mediators IL-3, IL-17, IL-6, IFNγ, CXCL9 and CXCL10 in response to LPS than their MLN counterparts did (Fig. 3f). Importantly, the effects of pMHC–NP therapy on antigen-presenting cells (APC)s from draining lymph nodes were not associated with impaired systemic immunity because pMHC–NP-treated NOD mice cleared an acute viral infection and mounted antibodies against an exogenous antigen as efficiently as untreated mice (Figs 3g, h).
Antigen-experienced T cells as targets
The memory-like phenotype and the upregulation of T-bet mRNA in the expanded TR1-like cells, coupled with the inability of pMHC–NPs to expand cognate TR1-like cells in non-diseased mice or NOD Ifng−/− mice suggested that pMHC–NPs expand pre-existing TR1 cells that arise from antigen-experienced precursors; and/or trigger the differentiation of antigen-experienced TH1 cells into TR1-like cells. Indeed, whereas diabetic NOD G6pc2−/− mice (which lack IGRP) responded to 2.5mi/IAg7–NPs like wild-type NOD mice, they did not respond to IGRP4–22/IAg7–NPs (Figs 4a, b). In vitro, 2.5mi/IAg7–NPs triggered the expression of CD49b and LAG-3 and the upregulation of c-maf, Il21, Il10 and Lag3 mRNA exclusively in anti-CD3 plus anti-CD28 mAb-activated, but not naive BDC2.5 CD4+ T cells (Fig. 4c, Extended Data Fig. 1d).
a, Percentage of tetramer+CD4+ T cells in hyperglycaemic NOD G6pc2−/− compared with NOD mice treated with IGRP4–22/IAg7– (n = 4 and 7) or 2.5mi/IAg7–NPs (n = 6 and 9). b, Blood glucose levels in hyperglycaemic NOD G6pc2–/– mice in response to pMHC–NP therapy (n = 4–6 per group). c, Upregulation of TR1 transcripts by anti-CD3/anti-CD28 mAb-activated eGFP–CD4+ T cells from BDC2.5 NOD Foxp3-eGFP mice in response to different in vitro stimuli (n = 4 mice each). d, Changes in TR1-relevant transcripts in naive or memory BDC2.5 CD4+ T cells in response to 2.5mi/IAg7-NPs in vivo (n = 6, 6, 5 and 4 from left). e, LAG-3 and CD49b profiles (blue; compared with isotype control in red) of Thy1b+ cells from d. f, Proliferation of CFSE-labelled memory BDC2.5 CD4+ T cells in NOD.Thy1a hosts in response to 2.5mi/IAg7–NPs. g, Incidence of diabetes in T-cell-reconstituted NOD scid hosts transfused with naive or memory BDC2.5 CD4+ T cells and treated with bi-weekly doses of 2.5mi/IAg7–NPs (n = 4 and 3) or left untreated (n = 4 and 6). P values were calculated via Mann–Whitney U-test or log-rank (Mantel–Cox) tests. Error bars, s.e.m.
To investigate this further, we transfused naive (CD44lowCD62Lhi) or memory-like (CD44hiCD62Llow) BDC2.5 CD4+ T cells into hosts of the congenic NOD.Thy1a strain and measured changes in their expression of LAG-3 and CD49b protein and c-maf, Il21, Il10, Ifng, Lag3 and Cd49b mRNA, both upon 2.5mi/IAg7–NP therapy and in the absence of therapy. Notably, the memory T cells from pMHC–NP-untreated hosts expressed about one hundred-fold higher levels of c-maf and Il21 and, to a lesser extent, Lag3 and Cd49b, but not Il10 mRNA than their naive counterparts (Fig. 4d). This is in accordance with the observed demethylation of Il21 and the c-Maf/IL-10- and IL-21-expression competency of effector/memory CD4+ T cells18,19,20,21,22, and suggests that the memory T-cell pool is enriched for uncommitted TR1 precursors, expressing a TR1-poised transcriptional program. Remarkably, whereas pMHC–NP therapy only upregulated Lag3 mRNA and, to a lesser extent, LAG-3 protein in naive BDC2.5 CD4+ T cells, it promoted the upregulation of Il10 mRNA and LAG-3 and CD49b protein, and the proliferation of memory BDC2.5 CD4+ T cells (Fig. 4d–f). Similar results were observed using memory eGFP– (FOXP3–) BDC2.5 CD4+ cells from BDC2.5 NOD Foxp3-eGFP mice (Extended Data Fig. 5y).
These effects on antigen-experienced T cells were accompanied by acquisition of anti-diabetogenic properties: whereas pMHC–NP therapy afforded 100% diabetes protection to T-cell-reconstituted NOD scid hosts bearing memory BDC2.5 T cells, therapy was inconsequential in hosts receiving naive BDC2.5 T cells (Fig. 4g). Therefore, pMHC–NP therapy promotes the differentiation (and expansion) of c-Maf-expressing antigen-experienced CD4+ T cells into TR1 progeny.
Translational potential
We determined the ability of human T1D-relevant pMHCII–NPs to expand cognate TR1-like T cells in NOD scid Il2rg−/− (NSG) hosts reconstituted with peripheral blood mononuclear cells (PBMCs) from T1D patients (Supplementary Table 2). Initial assay development focused on NSG hosts reconstituted with PBMCs from five DRB1*0401+ recent-onset T1D patients and treated with nanoparticles coated in either human glutamic acid decarboxylase-65 (GAD65)555–567(F557I)/DR4 or preproinsulin (PPI)76–90(K88S)/DR4 (Fig. 5a, b, Supplementary Table 2). We then repeated these experiments using NSG hosts reconstituted with PBMCs from 7 DRB1*0301+ T1D patients and a third T1D-relevant pMHC–NP type (hIGRP13–25/DR3–NPs) given at a higher dose. We saw expansion of tetramer+CD49b+LAG-3+CD4+ T cells in the spleen and/or PLNs (endogenous mouse (m)IGRP13–25 is highly homologous to hIGRP13–25) from all seven pMHC–NP-treated mice and none of the untreated controls (Fig. 5c, d, Supplementary Table 2, Extended Data Fig. 7a). The average percentage and numbers of tetramer+CD4+ T cells in IGRP13–25/DR3–NP-treated mice were significantly greater than in untreated littermates (Fig. 5d) and expressed Il10 mRNA (Fig. 5e). These responses could not be induced with peptide or peptide-coated nanoparticles or microparticles (Fig. 5d, Extended Data Fig. 7b and Supplementary Table 2).
a, Expansion of cognate CD4+ T cells by GAD555–567(557I)/DR4–NPs (top) or PPI76–90(88S)/DR4–NPs (bottom) in NSG mice engrafted with PBMCs from DR4+ T1D patients. b, CD49b and LAG-3 marker expression on the sample at the bottom of a. c, Expansion of cognate TR1-like CD4+ T cells in NSG mice engrafted with PBMCs from DR3+ T1D patients in response to IGRP13–25/DR3–NP-therapy. d, Percentages (left) and numbers (right) of tetramer+CD4+ T cells in mice engrafted with T1D PBMCs in response to treatment (n for spleen and PLN per treatment = 9/6, 7/6 and 14/1 from left legend). e, Expression of Il10 mRNA in IGRP13–25/DR3 tetramer+CD4+ T cells from mice treated with IGRP13–25/DR3–NPs (n = 3 each). f, The PLNs of responder mice contained increased numbers of lymphocytes compared to the other groups (n = 6, 3, 4, 3 from top legend). g, h, Correlation between the absolute numbers of IGRP13–25/DR3 tetramer+ cells in the PLNs (g) or spleen (h) and the percentage or absolute number of PLN or splenic B cells in IGRP13–25/DR3–NP-treated mice (n = 6 and 7). i, Secretion of IL-10 by LPS-stimulated CD19+ cells (ex vivo, for 24 h) isolated from the PLNs or spleens of hPBMC-engrafted NSG mice treated with IGRP13–25/DR3–NPs (n = 3 each). P values were calculated by Mann–Whitney U-test or Pearson correlation test. Error bars, s.e.m.
The PLNs of the pMHC–NP-treated mice that harboured increased percentages of tetramer+CD4+ T cells had increased cellularity (Fig. 5f). Furthermore, there were correlations between the number of PLN tetramer+CD4+ T cells and the percentage and absolute number of PLN human B cells, and the PLN B cells, unlike their splenic counterparts, produced IL-10 in response to LPS (Fig. 5g–i), suggesting Breg formation and/or recruitment. No such responses were seen in patient hPBMC-reconstituted NSG mice treated with peptide or peptide–NPs/MPs (Fig. 5f).
Discussion
We have shown that systemic therapy with nanoparticles coated with autoimmune-disease-relevant pMHC class II complexes triggers the expansion of cognate TR1-like CD4+ T cells, restores normoglycaemia in spontaneously diabetic mice and motor function in paralyzed EAE mice, and resolves joint swelling and destruction in arthritic mice, without compromising systemic immunity. We demonstrate that this outcome is dissociated from genetic background and type of autoimmune disease and can be replicated with ten different human or murine autoimmune-disease-relevant pMHC–NP types. The cell surface phenotype, cytokine secretion pattern, transcriptional profile and function of the TR1-like cell pools expanded by pMHCII-based nanomedicines are consistent with those described for murine TR1-like CD4+ T cells and remarkably similar to TR1 cells derived from healthy individuals and autoimmune disease patients8. We demonstrate key roles for prior autoantigenic experience and IFNγ- and IL-10-expression competence in the developmental biology of autoreactive TR1 cells. We show that pMHCII–NPs promote IL-10 transcription and the upregulation of TR1 markers in TR1-poised, antigen-experienced CD4+ T cells in an APC- and IL-27-independent manner, followed by systemic expansion. The need for IFNγ, the expression of the TH1 transcription factor T-bet, the c-Maf/IL-10- and IL-21-expression competency of effector and memory CD4+ T cells18,19,20, and the ability of pMHCII–NPs to turn T cells primed by active immunization into TR1 suppressors suggest that these TR1 precursors are effector/memory TH1 cells. We define the mechanisms of action and uncover a cascade of cellular interactions downstream of the pMHC–NP-expanded TR1-like cells, including Breg cell formation, that coordinately lead to the resolution of inflammation in an antigen-dependent but antigen-non-specific manner (Extended Data Fig. 8).
Collectively, our data support the contention that any single pMHC involved in a given autoimmune disease could be used to blunt complex autoimmune responses via this approach. Consistent with this prediction, the 20 pMHCI/II-based nanomedicines tested to date have similar efficacy, regardless of antigen prevalence, dominance or role in the disease process. Neither pMHC monomers nor peptides or peptide-coated nanoparticles/microparticles trigger cognate TR1 cell formation/expansion from the polyclonal T-cell repertoires or reverse T1D, CIA or EAE in the chronic models tested here. pMHC-based nanomedicines thus represent a new class of therapeutics in autoimmunity, capable of resolving cellularly and antigenically complex autoimmune responses in a disease- and organ-specific manner without compromising systemic immunity.
Methods
Mice
NOD/Ltj, NOD scid, BDC2.5-NOD, NOD Il10−/−, C57BL/6, C57BL/6 Il27r−/−, C57BL/10.M, NOD Foxp3-egfp and NOD scid Il2rg−/−(NSG) mice were purchased from the Jackson Lab. NOD Ifng−/− and LCMV Gp33-specific TCR-transgenic NOD mice were from D. Serreze (Jackson Lab). HLA-DR4-IE-transgenic C57BL/6 IAbnull mice were purchased from Taconic Farms. NOD Il10GFP (tiger) mice were obtained by backcrossing the Il10GFP allele from C57BL/6 Il10GFP mice (Jackson Lab) onto the NOD/Ltj background for 10 generations. 8.3-NOD and NOD G6pc2–/– mice have been described elsewhere23,24. These studies were approved by the corresponding institutional animal care committees. No statistical methods were used to predetermine sample size.
Antibodies, tetramer staining and flow cytometry
FITC, PE, PerCP or biotin-conjugated mAbs against mouse CD4 (RM4-5), CD8α (53-6.7), B220 (RA36B2), CD62L (MEL-1), CD69 (H1.2F3), CD44 (IM7), and CD49b (DX5) and streptavidin–PerCP were purchased from BD Pharmingen. The antibody against murine LAG-3 (C9B7W) was from eBioscience. Anti-latent-associated-TGF-β antibody (TW7-16B4) was from BioLegend. PE-conjugated pMHC class II tetramers were prepared using biotinylated pMHC monomers. Peripheral blood mononuclear cells, splenocytes, lymph node and bone marrow CD4+ T cells were incubated with avidin for 15 min at room temperature and stained with tetramer (5 μg ml−1) in FACS buffer (0.05% sodium azide and 1% FBS in PBS) for 30–120 min at 4 °C or 37 °C, depending on the tetramer, washed, and incubated with FITC-conjugated anti-CD4 (5 μg ml−1) and PerCP-conjugated anti-B220 (2 μg ml−1; as a ‘dump’ channel) for 30 min at 4 °C. Cells were washed, fixed in 1% paraformaldehyde (PFA) in PBS and analysed with FACScan, FACSaria, or BD LSRII flow cytometers. For other phenotypic analyses, single-cell suspensions were stained with pMHC tetramers and antibodies diluted 1:100 in FACS buffer (all used at 4 °C except anti-LAG-3, which was used at 37 °C), washed, fixed in 1% PFA, and analysed by FACS. All phenotypic staining were performed in the presence of an anti-CD16/CD32 mAb (2.4G2; BD Pharmingen) to block Fc receptors. Analysis was done using FlowJo software.
NSG-engrafted human T cells were analysed using the following mAbs: FITC-conjugated anti-CD4 (OKT4, BioLegend), APC-conjugated anti-CD19 (HIB19, BD Pharmingen), PerCP-conjugated polyclonal goat anti-LAG-3 IgG (R&D Systems), biotin-conjugated anti-CD49b (AK7, Pierce Antibodies, Thermo Scientific), and EF450-conjugated streptavidin (eBioscience). Briefly, splenocytes and pancreatic lymph node cells were incubated with avidin (0.25 mg ml−1 in FACS buffer) for 30 min at room temperature, washed and stained with tetramer (5 μg ml−1) for 1 h at 37 °C, washed, and incubated with FITC-conjugated anti-CD4 (2/100), APC-conjugated anti-CD19 (5/100; used as a ‘dump’ channel), PerCP-conjugated anti-LAG-3 (8/100) and biotin-conjugated anti-CD49b (4/100) at 4 °C for 45 min. After washing, the cells were incubated with EF450-conjugated streptavidin for 30 min at 4 °C, washed, fixed in 1% PFA in PBS and cells within the hCD4+/hCD19− gate analysed with a FACSCanto II (BD Bioscience).
Peptides and pMHCs
Unless specified otherwise, recombinant pMHC class II monomers were purified from culture supernatants of induced Drosophila SC2 cells transfected with constructs encoding I-Aβ and I-Aα chains carrying c-Jun or c-Fos leucine zippers, respectively, and a BirA and 6×His tags. In these constructs, the peptide-coding sequence was tethered to the amino-terminal end of the I-Aβ chain via a flexible Gly-Ser linker as described13. GAD65555(557I)–567/DR4, PPI76–90(88S)/DR4 and IGRP13–25/DR3 monomers were produced by loading the corresponding peptides onto DR4 and DR3 complexes purified from supernatants of induced SC2 cells, as described25. Other constructs (those encoding 2.5mi/IAg7, pMOG35–55/IAb, hMOG97–108/DR4-IE, hPLP175–192/DR4-IE and mCII259–273/DR4-IE) were purified from supernatants of Chinese Hamster Ovary (CHO) cells transduced with lentiviruses encoding a monocistronic message in which the peptide–MHCβ and MHCα chains of the complex were separated by the ribosome skipping P2A sequence26. These monomers were engineered to encode a BirA site, a 6×His tag and a free Cys at the carboxyterminal end of the construct. The self-assembled pMHC class II complexes were purified by nickel chromatography and used for coating onto nanoparticles or processed for biotinylation and tetramer formation as described above. The epitopes encoded in the different monomeric constructs used here include: 2.5mi; AHHPIWARMDA)13; IGRP128–145 (TAALSYTISRMEESSVTL) and IGRP4–22 (LHRSGVLIIHHLQEDYRTY)15; HEL14–22 (RHGLDNYRG); GAD65555(557I)–567 (NFIRMVISNPAAT)27; PPI76–90(88S) (SLQPLALEGSLQSRG)28; IGRP13–25 (QHLQKDYRAYYTF)25; pMOG38–49 (GWYRSPFSRVVH); hMOG97–108 (TCFFRDHSYQEE); hPLP175–192 (YIYFNTWTTCQSIAFPSK); and mCII259–273 (GIAGFKGDQGPKGET). IGRP4–22, IGRP128–145 and GPI282–292 (LSIALHVGFDH) or 2.5mi, pMOG35–55 (MEVGWYRSPFSRVVHLYRNGK), pMOG38–49, hMOG97–108 and hPLP175–192 peptides were purchased from Sigma Genosys, Mimotopes or Genscript.
Nanoparticles, pMHC–NP, peptide–NP and peptide–MP synthesis and purification
We coated pMHCs onto crosslinked dextran-coated or pegylated iron oxide NPs (CLIO– or PFM–NPs, respectively). Briefly, CLIO–NPs were treated with ammonia to produce amino groups (NH2). Avidin was oxidized with sodium periodate and added to the amino–NPs. Further incubation with sodium cyanoborohydride was used to generate a stable covalent bond. Finally, biotinylated monomers were added to the nanoparticles at a molar ratio of 4 mol biotin/mol avidin29. PFM–NPs were produced by thermal decomposition of Fe(acac)3 in the presence of 2 kDa methoxypolyethylene glycol maleimide (S. Singha et al., unpublished data). The NPs were purified using magnetic (MACS) columns (Miltenyi Biotec) or an IMag cell separation system (BD BioSciences). To conjugate pMHC or free peptide to PFM–NPs, we incubated pMHCs or peptide carrying a free carboxyterminal Cys with nanoparticles in 40 mM phosphate buffer, pH 6.0, containing 2 mM EDTA, 150 mM NaCl overnight at room temperature. The pMHC-conjugated nanoparticles were separated from free pMHC or peptide using magnetic columns, sterilized by filtration through 0.2 μm filters and stored in water or PBS at 4 °C. Quality control was performed using transmission electron microscopy, dynamic light scattering, and native and denaturing gel electrophoresis. pMHC or peptide content was measured using different approaches, including Bradford assay (Thermo Scientific), denaturing SDS–PAGE, amino acid analysis (HPLC-based quantification of 17 different amino acids in hydrolyzed pMHC–NP preparations) or dot-ELISA (Singha et al., unpublished data).
Peptide-coated microparticles were made using carboxylated 500 nm diameter polystyrene beads from Polysciences (Warrington, PA) as previously described17. The peptides were conjugated to polystyrene beads via carbodiimide chemistry following the manufacturer’s instructions. Briefly, we incubated 250 μl PSB (containing ~9 × 1011 beads) with 250 μg peptide in 0.1 M MES buffer, pH 5.0 at room temperature with gentle rolling in the presence of 1 mg EDC for 2 h. The peptide-conjugated polystyrene beads were washed with PBS to remove unconjugated peptides and analysed with native and denaturing PAGE against serial dilutions of unconjugated peptide and microparticle controls.
pMHC–NP and peptide or peptide–NP therapy in NOD mice
Experiments in pre-diabetic NOD mice involved treating (i.v.) cohorts of 10-week-old female mice with 7.5 μg of pMHC–NPs, or equivalent amounts of soluble pMHC monomers or uncoated nanoparticles twice weekly for 5 consecutive weeks. Experiments in diabetic mice involved following cohorts of 10-week-old female NOD/Ltj, NOD G6pc2−/−, NOD Il10−/− or NOD Ifng−/−mice for diabetes development by measuring blood glucose levels with Accucheck Strips (Roche) twice a week. Mice displaying two consecutive measurements >11 mM were considered diabetic and treated twice weekly with 7.5 μg pMHC–NPs, nanoparticles delivering a molecular equivalent of peptide or free peptide (8 μg per dose)16, until stably normoglycaemic (defined as 8 consecutive measurements <11 mM) or until hyperglycaemia was considered irreversible (3 measurements >25 mM). In Figs 1g and 4b and Extended Data Figure 1h), mice were randomized into treatment with 2.5mi/IAg7–NPs or HEL14–22/IAg7–NPs (Fig. 1g) or with 2.5mi/IAg7–NPs or IGRP4–22/IAg7–NPs (Fig. 4b). In Fig. 1g, IGRP4–22/IAg7 and IGRP128–145/IAg7 were tested in separate cohorts of mice. Mice treated with peptide or peptide–NPs (Fig. 1g) were randomized into either treatment within the same experiment. In vivo cytokine neutralization experiments involved administering mAbs against CD20 (5D2, a gift from A. Chan, Genentech; three doses of 250 μg i.v. on days 0–2 relative to the onset of hyperglycaemia) or 500 μg of HRPN (rIgG1), IFNγ (R4-6A2), IL-10 (JES5-2A5), TGF-β (1D11) or IL-21R (4A9) (BioXcell) i.p. twice a week for 2 weeks, followed by 200 μg per dose for 3 additional weeks. Mice were randomized into cytokine-blocking mAb-treatment (IFNγ, IL-10, TGFβ) or HRPN rat-IgG1 groups. Anti-CD20 and anti-IL21R mAbs were tested in separate cohorts of diabetic mice (Fig. 3a). Animals were assessed daily for glycosuria (corresponding to >16 mM blood glucose) and given human insulin isophane (1 IU per day) s.c. if positive. Upon treatment withdrawal, NOD mice were monitored for recurrence of hyperglycaemia until 50 weeks of age.
Peptide, pMHC, pMHC–NP, peptide–NP or peptide–MP therapy in EAE
Six- to eight-week-old female C57BL/6, C57BL/6 Il27r−/− or HLA-DR4-IE-transgenic C57BL/6 IAbnull mice were immunized with 150 μg of pMOG35–55 or hMOG97–108 or hPLP175–192, respectively in CFA s.c. at the base of the tail, under isofluorane anaesthesia. The mice received 300 ng of Pertussis toxin i.v. on days 0 and 3. Mice were weighed and scored daily starting on day 10 after immunization. The score system used was been reported elsewhere30 and plotted over a 5-point scale. When most of the mice showed signs of advanced disease (day 14) or reached maximum disease scores (day 21), mice were divided into different treatment groups, synchronized for weight and disease score averages, and treated twice a week with pMHC-coated and uncoated nanoparticles, an identical amount of pMHC monomer, peptide-coated nanoparticles (at an equivalent dose of peptide), free peptide (8 μg per dose i.v. or s.c.)16, peptide-conjugated microparticles (15 μg of peptide per dose)17 or unconjugated microparticles for 5 weeks. Mice were randomized into treatment with pMHC–NPs (one or two different types, depending on the experiment, as described in Figs 1j–n and 2f–h and Extended Data Figures 3 and 4), uncoated nanoparticles or no treatment. Peptide, peptide–MPs, peptide–NPs, pMHC monomers and uncoated microparticles were tested together; mice were randomized into each treatment group as mice reached the indicated disease score (Fig. 2f and Extended Data Figures 4f–i). An additional control cohort was treated with a single dose of peptide-conjugated microparticles (Extended Data Fig. 4h). Anti-cytokine and cytokine receptor mAb blocking studies (Extended Data Fig. 5g) involved randomization of mice into each treatment group.
Peptide, pMHC-NP or peptide-MP therapy in CIA
Bovine collagen II (bCII) dissolved in 0.05M acetic acid at 2 mg ml−1 was emulsified in CFA (v/v) containing 4 mg ml−1 of killed Mycobacterium tuberculosis (H37Ra). Eight- to twelve- week-old HLA-DR4-IE-transgenic C57BL/10.M mice were immunized intradermally at the base of the tail with 100 μg of bCII in CFA and boosted with 100 μg of bCII in IFA on days 14 and 28. The size of all four paws was measured using a caliper before immunization (day 0) and daily upon disease onset. Disease progression was measured as percentage increase in joint swelling relative to day 0. When this value reached 130%, mice were divided into different treatment groups and treated with pMHC-NPs, Cys-coated (pMHC unconjugated) NPs (25 μg of pMHC for pMHC-NPs, or an equivalent amount of iron for Cys-conjugated NPs), free peptide (8 μg per dose s.c.)16 or peptide-conjugated MPs (15 μg of peptide per dose)17 i.v. twice a week for 5 weeks. Mice were randomized into treatment with either pMHC–NP or uncoated nanoparticles, or into peptide or peptide–MP, respectively (Fig. 2a and associated Extended Data Data Figures). Mice were also assessed for clinical signs of disease up to a maximum clinical score of 12 as reported elsewhere31.
Peptide, pMHC–NP, peptide–NP and peptide–MP therapy in human PBMC-reconstituted NSG hosts
PBMCs from new or recently diagnosed HLA-DRB1*0401+ or -DRB1*0301+ T1D patients (recruited with informed consent, approved by the Institutional Review Board at Hospital Clinic) were depleted of CD8+ T cells using anti-CD8 mAb-coated magnetic beads (Miltenyi Biotech) and injected i.v. (2 × 107) into 8–10-week-old NSG hosts. Mice were treated with pMHC–NPs at the indicated doses, peptide-coated-NPs (at an equivalent dose of peptide), peptide alone (8 μg per dose s.c.)16 or peptide-conjugated microparticles (15 μg of peptide per dose)17 starting on day 5 after PBMC transfusion, twice a week for 5 consecutive weeks, or left untreated. Individual patient samples were processed separately and injected into two (for pMHC–NP and peptide–NP experiments) or three separate mice (for peptide and peptide–MP experiments); one or two of the two-to-three hosts used in each of these experiments were treated and the other was left untreated (Supplementary Table 2). Therapy-induced expansion of cognate CD4+ T cells was measured in PLNs and/or spleen as described above. The HLA genotype, gender, age, months from diagnosis and type of pMHC–NP tested for each patient are summarized in Supplementary Table 2.
Intraperitoneal glucose tolerance tests
Animals were fasted overnight and challenged with 2 mg kg−1 of d-glucose i.p. Blood glucose was monitored from the tail vein with a glucometer at different time points before and after glucose challenge. Serum insulin content was measured using the Mouse Ultrasensitive Insulin ELISA (ALPCO).
Evaluation of systemic cellular and humoral immunity
For the evaluation of cellular responses, pMHC–NP-treated and untreated female mice were injected with 2 × 106 plaque-forming units (pfu) of recombinant Vaccinia Virus (rVV) i.v. Cohorts of mice were killed on day 4 and 14 after infection and processed for pMHC tetramer staining and rVV titre measurements. Briefly, the ovaries were weighed, homogenized using a pestle in 300 ul of RPMI-1640 containing 10% FBS, freezed-thawed 3 times followed by 3 rounds of sonication (20 seconds each). Serial dilutions of the lysates were added to confluent BSC-1 cell cultures in 6-well plates, incubated at 37 °C for 2 h, washed twice with PBS and cultured in DMEM10. On day 2, the supernatants were discarded and the cell layers were stained with crystal violet to reveal plaques.
To evaluate humoral immunity, pMHC–NP-treated and untreated mice were immunized i.p. with 100 μg of DNP–KLH (Alpha Diagnostic International) in CFA. An identical boost was performed 3 weeks later. Mice were killed 10 days later. Anti-DNP antibody titres were measured by diluting serum samples in PBS containing 0.05% Tween 20. Anti-DNP antibodies were semi-quantified using an anti-DNP Ig ELISA Kit (Alphadiagnostic International) following the manufacturer’s instructions.
Proliferation and cytokine secretion assays
CD4+ T cells from pMHC–NP-treated mice were enriched from peripheral lymphoid organs using a BD Imag enrichment kit, stained with pMHC tetramers as described above and sorted by flow cytometry. For assays using memory and naive BDC2.5 CD4+ T cells, cells were enriched using Stem Cell Technologies enrichment kit, stained with antibodies and sorted. FACS-sorted cells (2–3 × 104) were co-cultured with bone marrow-derived DCs (2 × 104) pulsed with 2 μg ml−1 of peptide. Supernatants were collected 48 h later for measurement of cytokines via Luminex and the cells were pulsed with 1 microcurie (μCi) of (3H)-thymidine and collected after 24 h to measure thymidine incorporation in triplicates.
To ascertain whether pMHC–NP therapy promoted the generation of IL-10-secreting B-cells in the PLNs of PBMC-engrafted NSG hosts, we stained the PLN and splenic cell suspensions of individual mice with anti-hCD4–FITC, anti-hCD19–APC and tetramer–PE as described above, and sorted B-cells by flow cytometry (FACSAria-BD Biosciences). The B cells sorted from each organ were stimulated with LPS (1 μg ml−1, Sigma) for 24 h in RPMI-1640 supplemented with 10% human AB serum. The IL-10 content in the supernatants was measured in duplicates via Meso Scale technology using a V-PLEX Custom Human Cytokine kit for hIL-10 (Meso Scale Discovery). Data were normalized to the splenic B-cell values and reported as fold-change.
Isolation and in vitro stimulation of CD11b+ cells from the PLNs and MLNs
CD11b+ cells from LNs were obtained by digestion in collagenase D (1.25 μg ml−1) and DNase I (0.1 μg ml−1) for 15 min at 37 °C followed by purification with CD11b (BD Imag) mAb-coated magnetic beads. Cells were stimulated for 3 days with LPS (2 μg ml−1) and the supernatants analysed for cytokine content with a Luminex multiplex cytokine assay.
In vitro suppression assays
FACS-sorted 2.5mi/IAg7 tetramer positive or negative cells (2 × 104) were co-cultured with bone marrow-derived DCs (2 × 104) pulsed with 2 μg ml−1 ‘suppressor’ (2.5mi or GPI282–292) and ‘responder’ (gp33 or NRP–V7) peptides. Responder cells were CD8+ T cells (2 × 104) purified from from 8.3-NOD or LCMV-Gp33-specific TCR-transgenic NOD mice using BD-Imag beads. These cells were labelled with CFSE (5 μM) and added to the DC cultures in duplicates or triplicates. Dilution of CFSE in the responder cells was measured 48 h later by FACS. In other experiments, the wells were supplemented within 24 h of co-culture with HRPN rIgG, anti-IFNγ, anti-IL10 or anti-TGF-β (all 10 μg ml−1) or the IDO inhibitor, 1-methyl tryptophan (1-MT; 400 μM).
In vivo suppression of crosspresentation
For crosspresentation assays in non-transgenic mice, we transfused CFSE-labelled 8.3-CD8+ reporter cells (5–10 × 106) into untreated or pMHC–NP-treated mice and measured CFSE dilution in the hosts’ lymphoid organs within 7 days after transfer.
Adoptive transfer of suppression
Splenic CD4+ or CD8+ T cells (107) from untreated mice or mice treated with 10 doses of 2.5mi/IAg7–NPs or uncoated nanoparticles were transfused into 5–10 week-old NOD scid females. The hosts were transfused 24 h later with 2 × 107 CD4+ or CD8+ T-cell splenocyte mixtures purified from female NOD donors. The hosts were monitored for development of diabetes for at least 90 days after transfer (Fig. 1e). In another experiment, the hosts were treated twice a week with 2.5mi/IAg7–NPs (Fig. 1e). In other experiments (Fig. 3e), CD4+ or CD8+ T-cell-reconstituted 5–6-week-old NOD scid females were transfused with 5 × 105 CD19+ cells purified from the PLNs of mice treated with 10 doses of uncoated or 2.5mi/IAg7-coated NPs during the preceding 5 week (Fig. 3d). B-cells were purified using the EasySep Mouse CD19-positive selection Kit II (StemCell Technologies). Other cohorts, studied separately (Fig. 3e), received PLN or MLN CD19+ cells (5 × 105) plus total splenic CD4+ T cells (107) or 2.5mi/IAg7 tetramer+ (2 × 105) or tetramer− CD4+ T cells (107) from 2.5mi/IAg7–NP-treated donors. The hosts were randomized into each transfusion group and monitored for development of diabetes together. Figure 3e includes data from the corresponding cohors studied in Figs 1e and 3d. Isolation of 2.5mi/IAg7 tetramer+ and tetramer− cells from total splenic CD4+ T cells of 2.5mi/IAg7–NP-treated mice was performed using anti-PE mAb-coated microbeads and MACS LD columns (Miltenyi Biotec).
B-cell proliferation and Breg induction in vivo and Breg suppression in vitro
To isolate splenic DCs, spleens were digested in collagenase D and DNase for 15 min at 37 °C and DCs purified using anti-CD11c mAb-coated magnetic beads (MACS). The cells were pulsed with 10 μg ml−1 of 2.5mi or GPI282–292 peptide for 2 h at 37 °C and labelled with CFSE (0.5 μm) or PKH26 (2 μM), respectively. Labelled cells (5–10 × 106; mixed at 1:1 ratio) were administered i.v. into pMHC–NP-treated or untreated NOD mice. Three days later, we compared the ratios of CFSE+ versus PKH26+ cells in the spleens of the different hosts by FACS. Similar experiments were done using peptide-pulsed splenic B cells isolated from female donor mice using anti-B220 mAb-coated magnetic beads (MACS).
For in vivo Breg induction assays, B cells from NOD Il10GFP (tiger) mice were enriched using a CD19 enrichment kit (Stem Cell Technologies) and pulsed with 2.5mi or GPI282–292 peptides (10 μg ml−1) for 2 h at 37 °C. The peptide-pulsed B cells were washed twice with PBS, labelled with PKH26 and transfused (1 × 106) into pMHC–NP-treated or untreated mice. The hosts were killed 7 days later and their spleens labelled with anti-B220-APC and biotinylated anti-CD1d or anti-CD5 mAbs and Streptavidin-PerCP. PKH26+ cells were analysed for presence of eGFP+CD1dhigh or CD5+ cells by flow cytometry.
To determine the role of TR1-derived cytokines in Breg formation (Extended Data Fig. 5u), we repeated the experiments described above but using 3 × 106 B cells and hosts treated with 250 or 500 μg (given i.p. daily from day −3 to day 6 relative to B-cell transfer) of anti-HRPN (rIgG1), anti-IL-10 (JES5-2A5), anti-TGFβ (1D11) or anti-IL-21R (4A9) mAbs (BioXcell). Hosts were randomized into each antibody-treatment group and studied together.
To measure the ability of the TR1-induced Breg cells to suppress the antigen-induced activation of T cells in vitro, we isolated CD19+ B cells from the PLNs of age-matched untreated NOD mice or NOD mice treated with 10 doses of 2.5mi/IAg7–NPs and cultured these cells with LPS (10 μg ml−1) overnight. We then cultured these cells (2 × 104) with 2.5mi-peptide-pulsed (0.1 μg ml−1) bone marrow-derived DCs (2 × 104) and CFSE-labelled BDC2.5 CD4+ cells (4 × 104). Dilution of CFSE in CD4+ cells was measured 3 days later.
CD25+CD4+ Treg depletion
NOD mice were treated with 500 μg of anti-CD25 (PC61.5.3, BioXcell) i.p. 3 times weekly from 8 weeks of age, followed by 10-injections of pMHC–NPs given twice weekly starting at 10 weeks of age. Average CD4+CD25+ and FOXP3–eGFP+CD4+ T-cell depletion was 90% and 70%, respectively.
Histology
Tissues were fixed in 10% formalin and embedded in paraffin. H&E-stained pancreata were scored for insulitis as reported23. Briefly, insulitis was scored as: 0, none; 1, peri-insulitis; 2, infiltration covering <25% of the islet; 3, covering 25–50% of the islet; and 4, covering >50% of the islet.
Spinal cord and brain tissues were fixed in 10% buffered formalin for a minimum of 24 h, embedded in paraffin and sectioned at 6 μm. Slides from paraffin-embedded tissues were deparaffinized and subjected to antigen retrieval by steaming the slides in 10 mM sodium citrate buffer (pH 6.0) for 20 min and cooling at room temperature for 20 min. For immunohistochemistry, slides were fixed with 10% formalin and treated with 3% H2O2 in methanol at −20 °C. Sections were permeabilized with 0.25% Triton-X 100 and blocked with a skim milk blocking solution. Rabbit anti-IBA1 (Wako, 1:500) or rat anti-MBP (Abcam) were incubated at 4 °C overnight followed by respective biotinylated secondary antibodies (1:500), avidin-biotin complex, and 3,3′-diaminobenzidine. Sections were counterstained with haematoxylin and eosin, dehydrated with graded ethanol and mounted with Acrytol. For histological myelin staining, slides were fixed with 10% formalin or deparaffinized, dehydrated with graded ethanol, and incubated with 0.2% luxol fast blue in 95% ethanol at 65 °C. Slides were developed in 0.05% lithium carbonate, counterstained with haematoxylin and eosin, and mounted with Acrytol. Images of cerebellum were taken on an Olympus bright-field microscope. Inflammatory foci (dense nuclear clusters or perivascular cuffs with corresponding demyelination) were counted and their size measured using ImageJ software. For quantification of relative IBA1 intensity, blinded observers ranked images from highest to lowest intensity.
Knee joints from bCII-immunized mice were fixed in 4% buffered formalin overnight, and decalcified with 14% EDTA over 3 weeks. Decalcified paws were embedded in paraffin, sectioned at 8 μm and stained with haematoxylin and eosin to score infiltration and pannus formation on a scale of 5, where 5 corresponds to erosive arthritis, with severe infiltration and pannus covering 60% of the joint space. Proteoglycan depletion at the articular surface of the tibia and femur was assessed by the loss of safranin-O stain intensity. For this, sections were deparaffinized, hydrated and stained with haematoxylin before staining with 0.05% aqueous fast green for 5 min. Slides were fixed with 1% acetic acid and stained with 0.1% aqueous safranin-O for 2 min, dehydrated with graded ethanol, cleared with xylene and mounted with DPX. Scoring was done on a scale of 0 to 3 corresponding to: 0, 0% depletion, 1, low (<25%), 2, moderate (25–50%), and 3, severe (>50%). Destruction of articular cartilage included an assessment of the presence of dead chondrocytes (empty lacunae) and was scored on a scale of 3 (0, no empty lacunae; 3, complete loss of chondrocytes on articular cartilage/severe cartilage erosion).
Isolation of CNS-infiltrating lymphocytes
Mice were anesthetized with Ketamine-Xylazine and perfused with PBS through the heart left ventricle. The brain and spinal cord were isolated manually, cut into small fragments and digested with a solution of collagenase D (1.25 μg ml−1) and DNase I (1% w/v) in HBSS for 30 min at 37 °C. The digested CNS was passed through a 70 μm cell strainer. Cells were resuspended in DMEM (supplemented with 2% FBS and 10 mM HEPES) and 100% Percoll (to a final Percoll concentration of 30%). The solution was layered onto 65% Percoll and centrifuged at 380g for 30 min at room temperature. The mononuclear cell layer lying at the interphase was washed with RPMI before further analyses.
Quantitative RT–PCR
RNA was extracted from 2.5mi/IAg7 tetramer+ or tetramer− CD4+ T cells sorted from 2.5mi/IAg7–NP-treated NOD mice and stimulated in vitro with anti-CD3/anti-CD28 mAb-coated dynabeads.
Each tetramer+ sample corresponded to cells pooled from 2–3 mice. RNA was reverse transcribed and cDNA plated in Mouse Immunology 384 StellArray qPCR plates (Bar Harbour BioTechnology) with 2X SYBR Green Master Mix (Applied Biosystems). The plate was run in a 7900HT Applied Biosystems real-time PCR instrument, and the raw data was analysed using the Global Pattern Recognition (GPR) analysis tool (http://www.gene-quantification.com/qpcr-array.html). mRNA isolated from additional samples was subjected to RT–qPCR using primers specific for IL-21 (Forward: 5′-TCATCATTGACCTCGTGGCCC-3′; Reverse: 5′-ATCGTACTTCTCCACTTGCAATCC-3′), IL-10 (Forward: 5′-CTTGCACTACCAAAGCCACA-3′; Reverse: 5′-GTTATTGTCTTCCCGGCTGT-3′), c-Maf (Forward: 5′-AGCAGTTGGTGACCATGTCG-3′; Reverse: 5′-TGGAGATCTCCTGCTTGAGG-3′), IFN-γ (Forward: 5′-TGAACGCTACACACTGCATCTTGG-3′; Reverse: 5′-CGACTCCTTTTCCGCTTCCTGAG-3′), LAG-3 (Forward: 5′-TCCCAAATCCTTCGGGTTAC-3′; Reverse: 5′-GAGCTAGACTCTGCGGCGTA-3′), CD49b (Forward: 5′-CCGGGTGCTACAAAAGTCAT-3′; Reverse: 5′-GTCGGCCACATTGAAAAAGT-3′), Aryl Hydrocarbon Receptor (Forward: 5′-CGTCCCTGCATCCCACTACTT-3′; Reverse: 5′-GGACATGGCCCCAGCATAG-3′) and ICOS (Forward: 5′-TGACCCACCTCCTTTTCAAG-3′; Reverse: 5′-TTAGGGTCATGCACACTGGA-3′).
pMHC–NP-induced upregulation of TR1 transcripts in in vitro-activated CD4+ T cells was performed by culturing mouse naive eGFP−BDC2.5-CD4+ T cells from BDC2.5 NOD Foxp3-eGFP mice (CD62LhiFOXP3−eGFP−; 1.5 × 106 ml−1) with anti-CD3/anti-CD28 mAb-coated microparticles (1 bead per cell) for three days in the absence of APCs, followed by a one day culture of re-purified (microparticle-free) CD4+ T cells in rhIL-2 (30 IU ml−1), and a 6-day culture with 2.5mi peptide (10 μg ml−1), 2.5mi/IAg7 monomers (25 μg pMHC per ml), 2.5mi/IAg7–NPs (25 μg pMHC per ml and 50 μg ml−1 iron), or unconjugated nanoparticles (50 μg iron per ml). Relative gene expression was calculated using unstimulated cultures as controls.
pMHC–NP-induced upregulation of TR1 transcripts in naive compared to memory BDC2.5 CD4+ T cells in vivo was done by transfusing naive (CD44medCD62Lhi) or memory (CD44hiCD62Llow) eGFP−CD4+ T cells from BDC2.5-TCR-transgenic NOD or NOD Foxp3-eGFP mice (Thy1b+) (1–1.5 × 106 cells per host) into NOD.Thy1a hosts and by treating the hosts with four doses of 2.5mi/IAg7–NPs over two weeks or leaving them untreated. Two and a half weeks later, Thy1b+CD4+ T cells were sorted from the hosts and challenged with anti-CD3 and anti-CD28-coated magnetic Dynabeads for 3 days before mRNA extraction and RT–qPCR using primers specific for c-Maf, IL-21, IL-10, IFNγ, LAG-3 and CD49b.
To compare levels of IL-10 mRNA in the tetramer+ compared with tetramer− CD4+ T cells of pMHC–NP-treated PBMC-engrafted NSG hosts, we stained splenocytes with anti-hCD4-FITC, anti-hCD19-APC and tetramer–PE as described above, and sorted tetramer+ and tetramer− cells from individual hosts by FACS (FACSAria-BD Biosciences). Sorted cells were cultured for 72 h in RPMI-1640 containing 10% human AB serum, in the presence of Dynabeads Human T-Activator CD3/CD28 (LifeTechnologies) using a 1:1 cell to bead ratio. Total RNA from cell pellets was reverse-transcribed using a dual reverse transcriptase/lysis solution containing 5 mM DTT, 2 U ml−1 RNAase, 500 mM dNTPs, 10 U ml−1 of Superscript reverse transcriptase (Invitrogen, LifeTechnologies), 100 mg ml−1 bovine serum albumin, 1% Triton X-100, 25 ng ml−1 Oligo dT (Invitrogen), 0.5 nM spermidine, and 1X First Strant buffer (Invitrogen) in 20 μl for 60 min at 50 °C and 15 min at 70 °C. We then mixed 1 μl of the cDNA reaction volume with 12.5 μl of Power SyBRGreen PCR master mix solution (Applied Biosystem) and amplified with a real-time PCR machine (7900HT, Applied Biosystems) using the following primers: β-actin (Forward: 5′-CTGGAACGGTGAAGGTGACA-3′; Reverse: 5′-AAGGGACTTCCTGTAACAATGCA-3′), IL-10 (Forward: 5′-AAGACCCAGACATCAAGGCG-3′; Reverse: 5′-AATCGATGACAGCGCCGTAG-3′).
Statistical analyses
The sample size values described in the figure legends correspond to the number of individual mice tested (not replicates) and data shown correspond to pooled data from different experiments. Data were compared by Student’s t-test, Mann–Whitney U-test, chi-square, log-rank (Mantel–Cox), Pearson correlation or two-way ANOVA tests. Statistical significance was assumed at P < 0.05.
References
Santamaria, P. The long and winding road to understanding and conquering type 1 diabetes. Immunity 32, 437–445 (2010)
Babbe, H. et al. Clonal expansions of CD8+ T cells dominate the T cell infiltrate in active multiple sclerosis lesions as shown by micromanipulation and single cell polymerase chain reaction. J. Exp. Med. 192, 393–404 (2000)
Firestein, G. S. Evolving concepts of rheumatoid arthritis. Nature 423, 356–361 (2003)
Sakaguchi, S. et al. Foxp3+CD25+CD4+ natural regulatory T cells in dominant self-tolerance and autoimmune disease. Immunol. Rev. 212, 8–27 (2006)
Zhou, X. et al. Instability of the transcription factor Foxp3 leads to the generation of pathogenic memory T cells in vivo. Nature Immunol. 10, 1000–1007 (2009)
Komatsu, N. et al. Pathogenic conversion of Foxp3+ T cells into TH17 cells in autoimmune arthritis. Nature Med. 20, 62–68 (2014)
Bailey-Bucktrout, S. L. et al. Self-antigen-driven activation induces instability of regulatory T cells during an inflammatory autoimmune response. Immunity 39, 949–962 (2013)
Gagliani, N. et al. Coexpression of CD49b and LAG-3 identifies human and mouse T regulatory type 1 cells. Nature Med. 19, 739–746 (2013)
McLarnon, A. IBD: regulatory T-cell therapy is a safe and well-tolerated potential approach for treating refractory Crohn's disease. Nature Rev. Gastroenterol. Hepatol. 9, 559 (2012)
Desreumaux, P. et al. Safety and Efficacy of Antigen-Specific Regulatory T-Cell Therapy for Patients With Refractory Crohn's Disease. Gastroenterology 143, 1207–1217 (2012)
Roncarolo, M. G., Gregori, S., Lucarelli, B., Ciceri, F. & Bacchetta, R. Clinical tolerance in allogeneic hematopoietic stem cell transplantation. Immunol. Rev. 241, 145–163 (2011)
Tsai, S. et al. Reversal of autoimmunity by boosting memory-like autoregulatory T cells. Immunity 32, 568–580 (2010)
Stratmann, T. et al. Susceptible MHC alleles, not background genes, select an autoimmune T cell reactivity. J. Clin. Invest. 112, 902–914 (2003)
Kamanaka, M. et al. Expression of interleukin-10 in intestinal lymphocytes detected by an interleukin-10 reporter knockin tiger mouse. Immunity 25, 941–952 (2006)
Mukherjee, R., Wagar, D., Stephens, T., Le-Chan, E. & Singh, B. Identification of CD4+ T cell-specific epitopes of islet-specific glucose-6-phosphatase catalytic subunit-related protein: a novel β cell autoantigen in type 1 diabetes. J. Immunol. 174, 5306–5315 (2005)
Burton, B. R. et al. Sequential transcriptional changes dictate safe and effective antigen-specific immunotherapy. Nature Commun. 5, 4741–4747 (2014)
Getts, D. R. et al. Microparticles bearing encephalitogenic peptides induce T-cell tolerance and ameliorate experimental autoimmune encephalomyelitis. Nature Biotechnol. 30, 1217–1224 (2012)
Pot, C. et al. Cutting edge: IL-27 induces the transcription factor c-Maf, cytokine IL-21, and the costimulatory receptor ICOS that coordinately act together to promote differentiation of IL-10-producing Tr1 cells. J. Immunol. 183, 797–801 (2009)
Spensieri, F. et al. Human circulating influenza-CD4+ICOS1+IL-21+ T cells expand after vaccination, exert helper function, and predict antibody responses. Proc. Natl Acad. Sci. USA 110, 14330–14335 (2013)
Hale, J. S. et al. Distinct memory CD4+ T cells with commitment to T follicular helper- and T helper 1-cell lineages are generated after acute viral infection. Immunity 38, 805–817 (2013)
Sato, K. et al. Marked induction of c-Maf protein during Th17 cell differentiation and its implication in memory Th cell development. J. Biol. Chem. 286, 14963–14971 (2011)
Saraiva, M. et al. Interleukin-10 production by Th1 cells requires interleukin-12-induced STAT4 transcription factor and ERK MAP kinase activation by high antigen dose. Immunity 31, 209–219 (2009)
Verdaguer, J. et al. Spontaneous autoimmune diabetes in monoclonal T cell nonobese diabetic mice. J. Exp. Med. 186, 1663–1676 (1997)
Wang, J. et al. In situ recognition of autoantigen as an essential gatekeeper in autoimmune CD8+ T cell inflammation. Proc. Natl Acad. Sci. USA 107, 9317–9322 (2010)
Yang, J. et al. Islet-specific glucose-6-phosphatase catalytic subunit-related protein-reactive CD4+ T cells in human subjects. J. Immunol. 176, 2781–2789 (2006)
Holst, J. et al. Generation of T-cell receptor retrogenic mice. Nature Protocols 1, 406–417 (2006)
Reijonen, H. et al. Detection of GAD65-specific T cells by major histocompatibility complex class II tetramers in type 1 diabetic patients and at-risk subjects. Diabetes 51, 1375–1382 (2002)
Yang, J. et al. CD4+ T cells from type 1 diabetic and healthy subjects exhibit different thresholds of activation to a naturally processed proinsulin epitope. J. Autoimmun. 31, 30–41 (2008)
Moore, A., Grimm, J., Han, B. & Santamaria, P. Tracking the recruitment of diabetogenic CD8+ T cells to the pancreas in real time. Diabetes 53, 1459–1466 (2004)
Giuliani, F. et al. Additive effect of the combination of glatiramer acetate and minocycline in a model of MS. J. Neuroimmunol. 158, 213–221 (2005)
Leavenworth, J. W., Tang, X., Kim, H. J., Wang, X. & Cantor, H. Amelioration of arthritis through mobilization of peptide-specific CD8+ regulatory T cells. J. Clin. Invest. 123, 1382–1389 (2013)
Acknowledgements
We thank S. Thiessen, J. DeLongchamp, J. Erickson, J. Luces, R. Barasi and K. Umeshappa for technical contributions; L. Kennedy, L. Robertson and Y. Liu for flow cytometry; F. Jirik for help with histological analyses of arthritic mice; J. Elliott and K. Suzuki for Meso Scale measurements; M. Fritzler for Luminex; and P. Colarusso for assistance with microscopy. This work was funded by the Canadian Institutes of Health Research (CIHR), the Diabetes Research Foundation, the Juvenile Diabetes Research Foundation (JDRF), the Canadian Diabetes Association (CDA), the Multiple Sclerosis Society of Canada (MSSC), the Brawn Family Foundation, National Research Council of Canada–Industrial Research Assistance Program (NRC-IRAP), Instituto de Investigaciones Sanitarias Carlos III (ISCIII) Integrated Projects of Excellence and FEDER, the Ministerio de Economia y Competitividad of Spain (MINECO), the European Association for the study of diabetes (EASD), the Sardà Farriol Research Programme, and the European Community’s Seventh Framework Programme. X.C.C. was supported by studentships from the AXA Research Fund and the endMS network. P.A. was supported by the endMS network. J.B. was suported by a Rio Hortega fellowship and by a grant from the Spanish Society for Diabetes. S.T. was supported by a studentship from the Alberta Heritage Foundation of Medical Research (AHFMR). J.W. was funded by a fellowship from the CDA. P.Se. is an investigator of the Ramon y Cajal reintegration program and is supported by a JDRF Career Development Award. P.Sa. is a Scientist of the Alberta Innovates-Health Solutions and a scholar of ISCIII. The JMDRC is supported by the Diabetes Association (Foothills) and the CDA.
Author information
Authors and Affiliations
Contributions
X.C.-C. executed most of the experiments in Figs 1, 3a–d, h–j, 4a–f, Supplementary Table 1 and Extended Data Figs 1b–n, 2, 3, 5a–f, n–t 6, with contributions from J.Y., P.A., S.T. and J.W., and contributed to writing the manuscript with P.Sa. J.Y. executed the experiments in Figs 3e–g, 4g–i and Extended Data Figs 1a, c–f, 2h and 5u, v. P.A. executed all of the experiments described in Figs. 2, 3b, and Extended Data Figs 4 and 5g–m. J.B. recruited T1D patients and healthy controls and performed the experiments leading to Fig. 5, Extended Data Fig. 7 and Supplementary Table 2 under the supervision of P.Se. S.S., Y.Y. and A.M. produced nanoparticles and pMHC–NP conjugates for the study. C.F. produced 2.5mi/IAg7 class II monomers for mechanistic experiments. S.A. and M.K. contributed to the execution of the EAE experiments, and analysed histological sections for histopathological features of CNS inflammation and demyelination under the supervision of V.W.Y. E.J. provided human T1D-relevant pMHC class II monomers and tetramers. N.G., C.I. and T.S. produced the pMHC class II monomers used for the studies on the reversal of T1D. P.Sa. designed the study, supervised and coordinated its execution and wrote the manuscript.
Corresponding author
Ethics declarations
Competing interests
P.Sa. is the scientific founder of Parvus Therapeutics Inc. and has a financial interest in the company.
Extended data figures and tables
Extended Data Figure 1 Sustained expansion of cognate TR1-like CD4+ T cells by pMHCII–NP therapy restores normal glucose homeostasis in diabetic NOD mice by suppressing antigen presentation and the activation of non-cognate autoreactive T cells in the PLNs and the progression of insulitis.
a, Top left, expansion of cognate CD4+ T cells by 2.5mi/IAg7–NPs in anti-CD25 mAb-treated NOD Foxp3-eGFP mice. Data correspond to 8-week-old mice treated three times a week with 500 μg of a depleting anti-CD25 mAb i.p. or control anti-HPRN mAbs, followed by 10 doses of 2.5mi/IAg7–NPs starting at 10 weeks of age (two doses per week; n = 4 mice each). Bottom, the tetramer+CD4+ T cells from anti-CD25 mAb-treated mice express TR1 markers. Right, percentage of circulating FOXP3+eGFP+CD4+ (top) and CD25+CD4+ cells (bottom). b, Tetramer+CD4+ T cells sorted from 2.5mi/IAg7–NP-treated mice proliferate and produce IL-10 and, to a lesser extent IFNγ in response to stimulation with 2.5mi peptide-pulsed DCs (n = 3 mice). c, Representative cell surface CD49b and LAG-3 profiles on tetramer+CD4+ T cells from BDC2.5 NOD Foxp3-eGFP mice compared with tetramer−CD4+ T cells from transgenic or wild-type NOD mice (n = 4). d, Upregulation of CD49b and LAG-3 on anti-CD3/anti-CD28 mAb-activated BDC2.5 CD4+ T cells from BDC2.5 NOD Foxp3-eGFP mice in response to 2.5mi/IAg7–NP (25 μg pMHC per ml) versus 2.5mi peptide (10 μg ml−1) or 2.5mi/IAg7 monomers (25 μg pMHC per ml). e, Upregulation of eGFP (IL-10) in anti-CD3/anti-CD28 mAb-activated BDC2.5 CD4+ T cells from BDC2.5 NOD Il10GFP mice in response to 2.5mi/IAg7–NP as a function of CD49b and LAG-3 expression. f, Expression of eGFP (IL-10) in the CD4+ T cells of 2.5mi/IAg7–NP-treated NOD Il10GFP mice (2 doses per week for 5 weeks) as a function of CD49b and LAG-3 expression (left, representative profiles; right, eGFP MFI values) (n = 8). g, Proliferation of CFSE-labelled 8.3-TCR-transgenic CD8+ T cells (IGRP206–214/NRP–V7-specific) in response to 2.5mi/NRP–V7–peptide-pulsed or unpulsed DCs in the presence of tetramer− or tetramer+ CD4+ T cells from 2.5mi/IAg7-NP-treated mice and in the presence or absence of cytokine-blocking mAbs, rat IgG (negative control) or 1-methyl-l-tryptophan (1-MT; an IDO inhibitor). Data correspond to average of proliferated cells in 3–7 experiments per condition. h, Changes in blood glucose levels of spontaneously hyperglycaemic (>11 mM) female NOD mice treated with 2.5mi/IAg7–NP, IGRP4–22/IAg7–NP, IGRP128–145/IAg7–NP or HEL14-22/IAg7–NP (n = 6–9 per group), IGRP4–22 peptide or IGRP4–22 peptide–NPs (n = 9, 4–5 each). Mice received two doses per week until irreversibly hyperglycaemic or normoglycaemic for 4 consecutive weeks, at which point treatment was withdrawn. i, Incidence and timing of disease relapse in hyperglycaemic female NOD mice rendered stably normoglycaemic by treatment with 2.5mi/IAg7–NP, IGRP4-22/IAg7–NP or IGRP128–145/IAg7–NPs upon treatment withdrawal (after 4 consecutive weeks of normoglycaemia). Data correspond to responder mice in Fig. 1g. j, Post-prandial serum insulin levels in pMHC–NP-treated mice that reverted to normoglycaemia until 50 weeks of age (n = 6) versus newly diabetic (n = 12) and non-diabetic age-matched untreated controls (n = 10). k, Intra-peritoneal glucose tolerance tests (IPGTT) of the mice in h. l, Areas under the curve (AUC) in the IPGTTs shown in k. m, IPGTT serum insulin levels corresponding to the mice in k. n, Proliferation of CFSE-labelled IGRP206–214-reactive 8.3-CD8+ T cells in the PLNs compared with MLNs of 2.5mi/IAg7–NP-treated mice that reverted to normoglycaemia until 50 weeks of age, non-diabetic age-matched untreated controls and newly diabetic mice. Left panels show representative FACS profiles. Right panel compares percentages of proliferated cells in the PLNs after subtraction of the background proliferation values in non-draining MLNs (n = 6–8 mice per group). P values were calculated by Mann–Whitney U-test, log-rank (Mantel–Cox) test or two-way ANOVA. Data are averages ± s.e.m.
Extended Data Figure 2 Nanoparticles coated with different T1D-relevant pMHCII complexes expand cognate TR1-like CD4+ T cells in vivo to similar extent, regardless of epitope dominance or role of the target T-cell specificity in the disease process.
a, Percentage of tetramer+CD4+ T cells in the PLN, MLN and bone marrow (BM) of 2.5mi/IAg7–NP-treated mice that reverted to normoglycaemia until 50 weeks of age (n = 5–6 mice per lymphoid organ) or relapsed (n = 1–2) compared with newly diabetic (n = 5–6) and non-diabetic age-matched untreated controls (n = 4–6). b, Percentage of tetramer+CD4+ T cells in the splenic CD4+ T cells of 2.5mi/IAg7–NP-treated mice that reverted to normoglycaemia until 50 weeks of age or of age-matched non-diabetic untreated mice, stained with two T1D-relevant but non-cognate pMHCII tetramers (n = 3–4 per group). c, Percentage of tetramer+CD4+ T cells in blood, spleen, PLN, MLN and bone marrow of IGRP4–22/IAg7–NP-treated mice that reverted to normoglycaemia until 50 weeks of age (n = 5–6 mice per lymphoid organ) compared with newly diabetic (n = 5–8) and non-diabetic age-matched untreated controls (n = 4–6). d, Percentage of tetramer+CD4+ T cells in blood, spleen, PLN, MLN and bone marrow of IGRP128–145/IAg7–NP-treated mice that reverted to normoglycaemia until 50 weeks of age (n = 5–7 mice per lymphoid organ) compared with newly diabetic (n = 4–7) and non-diabetic age-matched untreated controls (n = 5–7). e, Representative IGRP4–22/IAg7, IGRP128–145/IAg7 and GPI282–292/IAg7 tetramer staining profiles for splenic CD4+ T cells from IGRP4–22/IAg7–NP- and IGRP128–145/IAg7–NP-treated compared with untreated NOD mice. f, Percentages of blood CD4+ T cells of IGRP4–22/IAg7–NP- or IGRP128–145/IAg7–NP-cured, HEL14–22/IAg7–NP-treated and age-matched non-diabetic untreated mice stained with non-cognate pMHCII tetramers (n = 3–7 per group). g, The tetramer+CD4+ T cells of mice treated with IGRP128–145/IAg7–NP (top) and IGRP4-22/IAg7–NP (bottom) proliferate and produce IL-10 specifically in response to stimulation with IGRP4–22 or IGRP128–145-peptide-pulsed DCs, respectively (n = 3 mice each). cpm, counts per minute. h, Percentages of IGRP4–22/IAg7 tetramer+CD4+ T cells in blood, spleen, PLN, MLN and bone marrow of NOD mice at the onset of hyperglycaemia or upon treatment with IGRP4–22/IAg7–NPs, or IGRP4–22 peptide or IGRP4–22 peptide-coated nanoparticles (n = 5–9 mice per organ). P values were calculated by Mann–Whitney U-test. Data are averages ± s.e.m.
Extended Data Figure 3 EAE-relevant pMHCII–NPs expand cognate IL-10-secreting TR1-like CD4+ T cells and ameliorate established clinical and pathological signs of EAE.
a, b, Changes in the average weights of C57BL/6 mice immunized with pMOG35–55 and treated with pMOG38–49/IAb–NPs or uncoated nanoparticles starting on days 14 (a) or 21 (b) after immunization. c, Percentage of pMOG38–49/IAb tetramer+CD4+ T cells in peripheral lymph nodes, bone marrow and central nervous system (CNS) of mice from a and b. d, The tetramer+CD4+ T cells of pMOG38–49/IAb–NP-treated mice proliferate and produce IL-10 and, to a lesser extent, IFNγ in response to stimulation with pMOG38–49 peptide-pulsed DCs. e, Left and middle, representative luxol fast blue (LFB)/H&E cerebellum staining images from untreated and treated mice from b showing presence of inflammatory foci and areas of demyelination (red arrows). Right, average number of inflammatory foci per section. Data corresponds to 4 untreated and 5 treated mice. f, Representative LFB/H&E-stained spinal cord sections from mice in b. Data were compared with Mann–Whitney U-test. Data are averages ± s.e.m.
Extended Data Figure 4 EAE- or CIA-relevant pMHCII–NPs expand cognate TR1-like CD4+ T cells and ameliorate clinical and pathological signs of EAE or CIA in HLA-DR4-IE-transgenic C57BL/6 IAbnull or C57BL/10.M mice.
a, Changes in the average EAE scores of HLA-DR4-IE-transgenic C57BL/6 IAbnull mice immunized with hPLP175–192 or hMOG97–108 and treated with hPLP175–192 /DR4-IE or hMOG97–108/DR4-IE–NPs or uncoated nanoparticles starting on the day when mice reached a score of 1.5 (to synchronize the groups for disease activity) (n = 3–4 per group). b, Percentage of tetramer+CD4+ T cells in spleen, blood, cervical and inguinal LNs and CNS of mice from a. Data correspond to 4 pMHC–NP-treated and 6 control-NP-treated mice. c, Changes in the average weights of HLA-DR4–IE-transgenic C57BL/6 IAbnull mice from a, immunized with hPLP175–192 or hMOG97–108 and treated with hPLP175–192/DR4-IE–NPs, hMOG97–108/DR4-IE–NPs or uncoated nanoparticles when the mice reached a score of 1.5. d, LFB/H&E staining of the cerebellum of HLA-DR4-IE-transgenic C57BL/6 IAbnull mice from a showing reductions in inflammation and demyelination in mice treated with hPLP175–192 /DR4-IE or hMOG97–108/DR4-IE–NPs compared with controls. e, Percentage of tetramer+CD4+ T cells in lymph nodes and bone marrow of the mice in Fig. 2a (C57BL/10.M HLA-DR4-IE mice immunized with bovine collagen) at the end of follow-up (10 doses, 5 weeks). f, Changes in the average weights of HLA-DR4-IE-transgenic C57BL/6 IAbnull mice immunized with hPLP175–192 from Fig. 2f. g, Representative LFB/H&E staining of the cerebellum of HLA-DR4-IE-transgenic C57BL/6 IAbnull mice immunized with hPLP175–192 and treated with hPLP175–192/DR4-IE–NPs, hMOG97–108/DR4-IE–NPs, hMOG97–108 peptide i.v. or s.c. (8 μg per dose), hMOG97–108/DR4-IE monomer (25 μg per dose), hMOG97–108 peptide–NPs (using the molar equivalent of peptide delivered via pMHC–NPs; 0.68 μg per dose), or hMOG97–108 peptide–MPs (15 μg peptide per dose) compared with mice left untreated or treated with uncoated NPs or MPs (at the same NP/MP number). h, Changes in the average EAE scores and body weights of HLA-DR4-IE-transgenic C57BL/6 IAbnull mice immunized with hPLP175–192 in response to treatment with hMOG97–108 peptide i.v. or s.c. (8 μg per dose16), hMOG97–108/DR4-IE monomer (25 μg per dose), hMOG97–108 peptide–NPs (0.68 μg peptide per dose), hMOG97–108 peptide–MPs (15 μg peptide per dose17), or a single dose of hMOG97–108 peptide–MPs (15 μg peptide17) compared with mice left untreated or treated with uncoated NPs or MPs (at the same NP/MP number) (n = 4–5 per group). The cohort of mice treated with one dose had to be terminated after 2.5 weeks, owing to rapid progression of disease. i, Percentages of tetramer+CD4+ T cells in spleen, blood, cervical and inguinal LNs and bone marrow of mice from h (n = 3–9 per group). Data were compared with Mann–Whitney U-test or two-way ANOVA. Data are averages ± s.e.m.
Extended Data Figure 5 Disease reversal by pMHC–NPs is driven by the TR1 cytokines IL-21, IL-10 and TGF-β and involves several downstream cellular targets.
a, Changes in blood glucose levels in diabetic NOD mice (>11 mM) treated with IGRP4–22/IAg7–NPs and blocking anti-IL-10, anti-IFNγ or anti-TGF-β mAbs or anti-HRPN rat-IgG (n = 4–6 per group). b, c, Percentages of tetramer+CD4+ T cells in the spleens (b), and proliferation of CFSE-labelled 8.3-CD8+ T cells in the PLNs verus MLN of the mice from Fig. 3a at the end of follow up (c). d, Changes in blood glucose in hyperglycaemic NOD, NOD Il10−/− and NOD Ifng−/− mice (n = 3−6 per group) in response to 2.5mi/IAg7–NPs. e, f, Percentages of tetramer+CD4+ T cells in the spleens (e), and proliferation of CFSE-labelled 8.3-CD8+ T cells in the PLNs versus MLN of the mice from d at the end of follow up (f). g, EAE scores of mice treated with pMHC–NPs and rat-IgG or blocking mAbs (n = 4 per group). h, LFB/H&E staining of the cerebellum of HLA-DR4-IE-transgenic C57BL/6 IAbnull mice from g, highlighting differences in inflammation and demyelination in mice treated with hPLP175–192/DR4-IE–NPs and rat-IgG versus blocking anti-IL-10, anti-TGF-β or anti-IL-21R mAbs. i, Changes in the average body weights of HLA-DR4-IE-transgenic C57BL/6 IAbnull mice from g. j, Percentage of tetramer+CD4+ T cells in spleen, blood and inguinal LNs of mice from g (n = 4 per group). k, l, Changes in the average EAE scores (k) and body weights (l) of C57BL/6 Il27r−/− mice immunized with pMOG35–55 and treated with pMOG38–49/IAb–NPs or uncoated nanoparticles starting on the day when mice reached a score of 1.5 (to synchronize the groups for disease activity) (n = 7 and 4, respectively). m, Representative IBA1 and LFB/H&E stainings of the cerebellum and the corresponding relative rank scores of mice from k (n = 3 and 4, respectively). n, Percentage of tetramer+CD4+ T cells in spleen, blood, inguinal LNs and bone marrow of mice from k (left), and representative CD49b and LAG-3 staining profiles of tetramer+ versus tetramer− cells (right). o, Percentage of B220+ cells in the PLNs or MLNs of 2.5mi/IAg7–NP- or HEL14-22/IAg7–NP-treated mice (n = 4 per group). p, Correlation between the percentages of PLN and splenic B220+ cells and 2.5mi/IAg7 tetramer+CD4+ T cells in additional cohorts of mice treated with 2.5mi/IAg7–NPs, over a range of total pMHC dose (0.75–25 μg of total pMHC) (n = 24–28). q, Left, in vitro proliferation of CFSE-labelled BDC2.5 CD4+ T cells against 2.5mi or GPI282–292 peptide-pulsed B cells purified from the PLNs or MLNs of untreated NOD mice or mice treated with 2.5mi/IAg7–NPs (n = 5–6 per group). Right, representative CFSE dilution profiles. Briefly, profiles show the extent of CFSE dilution in CFSE-labelled BDC2.5 CD4+ T cells cultured in the presence of 2.5mi or GPI282–292 peptide-pulsed B cells purified from the PLNs or MLNs of untreated or 2.5mi/IAg7–NP-treated NOD mice. r, PLN-derived B cells (105) from 2.5mi/IAg7–NP-treated mice secrete IL-10 ex vivo in response to LPS (1 μg ml−1). Data correspond to 6 pMHC-treated and 5 untreated NOD mice. s, t, Changes in the percentages of 2.5mi (PKH26-labelled) compared with GPI282–292 peptide-pulsed (CFSE-labelled) B cells (s) or DCs (t) 7 days after transfer (at 1:1 ratio) into untreated or 2.5mi/IAg7–NP-treated NOD mice. Histograms show averaged ratios for each cell type and condition (n = 3–4 mice per cell type and condition). u, Percentages of CD5+CD1dhieGFP+B220+ cells in mice treated as in Fig. 3b plus blocking Abs (n = 4 each). v, LPS-stimulated PLN B cells from NOD mice treated with 10 doses of 2.5mi/IAg7–NPs suppress the proliferation of CFSE-labelled BDC2.5 CD4+ T cells by 2.5mi peptide-pulsed DCs in vitro, as compared to LPS-stimulated PLN B cells from untreated controls. x, Percentage of CD19+CD3− cells in blood before and after 3 doses of 250 μg of anti-CD20 mAb (n = 4). y, 2.5mi/IAg7–NP-induced upregulation of IL-21 and IL-10 mRNA in memory eGFP− BDC2.5 CD4+ T cells from BDC2.5-TCR-transgenic NOD Foxp3-eGFP donors in NOD Thy1a hosts (n = 5). P values were calculated by Pearson correlation, Mann–Whitney U-test or two-way ANOVA. Data are averages ± s.e.m.
Extended Data Figure 6 Effects of cytokine blockade or genetic deficiency on the cytokine profile of cognate CD4+ T cells expanded by 2.5mi/IAg7–NPs.
n = 3 mice each. Data are averages ± s.e.m.
Extended Data Figure 7 Human T1D-relevant pMHCII–NPs, but not free peptide or peptide-coated nanoparticles or microparticles, expand cognate TR1-like CD4+ T cells in human PBMC-engrafted NSG hosts.
a, FACS profiles (cognate versus control tetramer staining in hCD4+ T cells) of samples from mice identified as responders in Supplementary Table 2. Numerical data on tetramer+ T cells are presented on Supplementary Table 2. b, Representative FACS profiles (cognate versus control tetramer staining in splenic hCD4+ T cells) of human healthy control PBMC-engrafted NSG hosts treated with IGRP13-25/DR3–NPs (left), or human T1D PBMC-engrafted NSG hosts treated with IGRP13–25 peptide, IGRP13–25 peptide-coated nanoparticles, IGRP13–25 peptide-coated microparticles, or left untreated (right). See Fig. 5 legend for details.
Extended Data Figure 8 Schematic of the proposed mode of operation of pMHCII-based nanomedicines.
pMHCII-coated NPs (pMHC–NP, lacking costimulatory molecules) promote the differentiation of disease-primed (antigen-experienced) IFNγ-producing CD4+ TH1-cells into memory TR1-like CD4+ T cells followed by systemic expansion. This differentiation process (but not the subsequent expansion) requires both IFNγ and IL-10, whereas IL-27 is dispensable. The pMHC–NP-expanded (mono-specific) autoreactive TR1-like CD4+ T cells then suppress other autoreactive T-cell responses by secreting IL-21, IL-10 and TGF-β, which act on local APCs (B cells, CD11c+ and CD11b+ cells) that have captured the cognate autoantigen and thus present cognate pMHCII complexes to the expanded TR1-like cells. This interaction inhibits the proinflammatory function of the targeted APCs and blocks their ability to present other pMHC class I and class II complexes to non-pMHC–NP-cognate autoreactive T-cell specificities (note that the local APCs uptake both cognate and non-cognate autoantigens shed into the milieu simultaneously). Suppression of antigen-presentation requires IL-10 and TGF-β but not IFNγ or IL-21. Furthermore, cognate interactions between the pMHC–NP-expanded TR1 CD4+ T cells and autoreactive B cells specific for the cognate autoantigen (able to display the cognate pMHCII complex on the surface) promotes their differentiation into Breg cells in an IL-21-dependent manner, which contribute to promote local immunosuppression, likely by secreting IL-10. Suppression of antigen presentation selectively targets APCs displaying the cognate pMHC, but as local APCs that capture the cognate autoantigen also capture other autoantigens simultaneously, the autoregulatory CD4+ T cells expanded by pMHC–NPs blunt the presentation of other autoantigenic pMHC complexes to a broad range of autoreactive T cells. This suppression is disease-specific and self-limiting.
Supplementary information
Supplementary Tables 1 and 2
This file comprises: Table 1 - Transcriptional profile of pMHC-NP-expanded CD4+ T-cells. (a) qRT-PCR for a panel of 384 immunological markers in 2.5mi/IAg7 tetramer+ versus tetramer– CD4+ T-cells sorted from NOD mice treated with 10 doses of 2.5mi/IAg7-NPs from 10-15 weeks of age (n=3 and 4 samples, respectively). The cells were stimulated in vitro with anti-CD3/anti-CD28 mAb-coated Dynabeads before RNA collection. Panel summarizes the most significant differences. (b) qRT-PCR for 8 TR1-relevant markers, including markers that were not represented in the primer set used in a. Data correspond to four additional 2.5mi/IAg7 tetramer+ and seven tetramer– CD4+ T-cell samples; Table 2 - Human T1D donors and outcome of pMHC-NP, peptide, peptide-NP and peptide-MP therapy in PBMC-engrafted NSG hosts. See text for details. (PDF 729 kb)
Rights and permissions
About this article
Cite this article
Clemente-Casares, X., Blanco, J., Ambalavanan, P. et al. Expanding antigen-specific regulatory networks to treat autoimmunity. Nature 530, 434–440 (2016). https://doi.org/10.1038/nature16962
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/nature16962
This article is cited by
-
The interaction between nanoparticles and immune system: application in the treatment of inflammatory diseases
Journal of Nanobiotechnology (2022)
-
Stem cells differentiation into insulin-producing cells (IPCs): recent advances and current challenges
Stem Cell Research & Therapy (2022)
-
Regulation of activated T cell survival in rheumatic autoimmune diseases
Nature Reviews Rheumatology (2022)
-
Membranous nephropathy: new pathogenic mechanisms and their clinical implications
Nature Reviews Nephrology (2022)
-
Re-programming mouse liver-resident invariant natural killer T cells for suppressing hepatic and diabetogenic autoimmunity
Nature Communications (2022)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.




