Flt3L-Derived Antigen-Presenting Cell Transfer in Neonatal NOD Mice Reduces the Incidence of Type 1 Diabetes

Abstract

Type 1 diabetes (T1D) is an autoimmune disease characterized by progressive stages culminating in T-cell–mediated destruction of the β-cells at the islets of Langerhans. The immune mechanisms that initiate T1D are not fully resolved but likely involve an interaction between proinflammatory antigen-presenting cells (APCs) and autoreactive T cells that initiate immune infiltration and activation. Previous studies have tested the use of tolerogenic APCs in adult female NOD mice to delay or prevent T1D with only slight to intermediate success. Moreover, immune infiltration begins as early as age 4 weeks; therefore, targeting autoreactive T cells with tolerogenic APCs in adult mice may not impact later stages of diabetes. Thus, we hypothesize that the transfer of tolerogenic APCs at the neonatal stage prior to priming and immune infiltration will result in effective protection from autoimmunity. Our studies demonstrate that immature APCs travel to the pancreatic draining lymph nodes, alter the cytokine milieu in young mice, divert autoreactive CD4⁺ T cells to anergy, and drastically decrease proliferation and function of cytotoxic lymphocytes in adult prediabetic mice, leading to a significant reduction in the incidence of T1D.

Article Highlights

• Neonatal transfer of immature dendritic cell–enriched Flt3L splenocytes significantly reduces the incidence of type 1 diabetes in female NOD mice.

• Early time points are associated with accumulation of anergic T cells.

• In adult mice, there is a reduction in CD4 T helper 1 cells and reduced proliferation and perforin of CD8 T cells.

• Our work demonstrates how targeting the neonatal window of tolerance alters autoimmunity outcome.

Graphical Abstract

Introduction

Type 1 diabetes (T1D) is an autoimmune disease orchestrated by the T-cell–mediated destruction of the insulin-producing β-cells in the islets of Langerhans (1). The NOD mouse model is a well-established system that shares many characteristics of T1D, including polygenetic predisposition, spontaneous initiation of disease, and immune cell involvement (2). T1D is known as a stage-dependent disease requiring interactions among innate and adaptive immune cells to drive overt diabetes (3). Studies in NOD mice have indicated that antigen-presenting cells (APCs) play a crucial role in the progression to T1D. In the absence of both dendritic cells (DCs) and macrophages, NOD mice are protected from diabetes (4–6). In the early stages of T1D, inflammatory APCs establish a pathogenic microenvironment at the pancreatic draining lymph nodes (pLNs), but as immune infiltrates accumulate in the pancreas, the pLN niche loses importance as removing the pLNs in adult mice does not prevent onset of T1D (6–9). Together, this suggests a time-specific role for pLNs in disease progression.

Prior studies using either in vivo Flt3-ligand (Flt3L) treatment or bone marrow–derived DC transfers into adult female NOD mice showed variable, but low levels of protection to T1D (10–14). NOD mice and human patients with T1D have abnormalities in the myeloid DC compartment, with a lower frequency observed (15,16). However, soluble Flt3L treatment can increase both myeloid-derived DCs and plasmacytoid DCs, thus compensating for defective APCs (11,12). The addition of cytokines interleukin-6 (IL-6) and IL-4 to the Flt3L-DC derivation process slightly improved tolerance; however, such studies failed to acknowledge early priming of autoreactive T cells and immune infiltration to the pancreas (8,9,13,14,17). Interestingly, DCs isolated from pLNs of adult (age 8–20 weeks) female NOD mice conveyed 73% protection from diabetes onset in 4-week-old female mice, underscoring the role of site-specific early tolerance (18). The idea of a neonatal window of tolerance during which the immune system can be altered to prevent autoimmunity is not a new concept to the field of T1D (19–21). An example of a neonate time-dependent effect is the pleiotropic nature of tumor necrosis factor α (TNF-α). When TNF-α levels increase systemically in neonatal mice, disease is accelerated, while TNF-α treatment in adult NOD mice confers protection (21–23). γ-Interferon (IFN-γ) is another example of a pleiotropic cytokine in T1D, exhibiting both pathogenic and regulatory functions. IFN-γ is known to drive pathogenic T helper 1 (Th1) T-cell differentiation and regulate CD8⁺ cytotoxic lymphocyte (CTL) function (24,25). Additionally, β-cell apoptosis induced by autoimmune CD4⁺ T cells is dependent on IFN-γ (25). However, DCs stimulated with IFN-γ can prevent T1D, and protective immunization with Complete Freund’s Adjuvant is dependent on IFN-γ (26–28), suggesting that the proinflammatory milieu could lead to tolerance, depending on timing, location, and cellular targets.

The progressive development of T1D suggests that earlier interventions targeting antigen presentation in the draining lymph nodes may enhance protection by disrupting T-cell priming. To test this hypothesis, we developed a model of adoptive transfers of in vivo immature DC-enriched NOD.Scid splenocytes (Flt3L-APCs) to neonatal (day 0–17) NOD mice. After APC transfers into neonatal mice, we observed a significant reduction in progression to T1D, an event likely driven by modifying the pLN landscape at a critical time point in autoreactive T-cell priming. Consequently, we saw a drastic long-term change to autoreactive CD4⁺ and CD8⁺ T-cell function in adult mice, supporting the idea of strengthening tolerance specifically during the neonatal window to prevent autoimmunity.

Research Design and Methods

Mice

NOD/ShiLtJ (NOD), NOD.Scid, NOD.Scid.CD45.2, NOD.BDC2.5, NOD.CD45.2, NOD.Foxp3.GFP, and NOD.Thy1.1 mice were a gift from Dr. Dario Vignali (University of Pittsburgh School of Medicine) and maintained at the University of Utah animal facility under specific pathogen-free conditions. Experimental protocols were approved by the Institutional Animal Care and Use Committee of the University of Utah, Salt Lake City, Utah. For preweaning experiments, both male and female NOD mice were used; for prediabetic experiments, only female NOD mice were analyzed.

B16-Flt3L Culture and Flt3L-Derived APC Cell Injections

The B16-Flt3L melanoma cell line was used to generate APCs in vivo using NOD.Scid mice as previously described (23). Spleens were isolated 14–19 days post–Flt3L-tumor injection, treated with red blood cell lysis buffer for up to 3 min, processed, and filtered to obtain a single-cell suspension in sterile (1×) PBS. Neonatal 0–2-day-old pups were given 5 × 10⁶ APCs in a 100-μL i.p. injection every 3 days for a total of six injections.

Diabetes Monitoring

Urine glucose level was determined by Diastix (Ascencia Diabetes Care). NOD and NOD.CD45.2 female mice were used for diabetes incidence. Hyperglycemia was assessed by blood glucose level using a FreeStyle Lite glucometer (Abbott). Mice with a blood glucose level of >400 mg/dL or two consecutive readings of >300 mg/dL were considered diabetic.

Organ Preparation and Flow Cytometry Analysis

Pancreata were perfused with 3 mL of 30,000 units/mL collagenase IV solution (Worthington/Fisher Scientific) via the pancreatic duct. Pancreata were incubated in 3 mL of 30,000 units/mL collagenase IV for 30–45 min at 37ºC. Individual islets were picked under a microdissection scope and dissociated using enzyme-free cell dissociation buffer (Gibco) for 15 min at 37ºC. Ghost BV510 (Tonbo Biosciences) and FcBlock (BioLegend) were used to exclude dead cells and block nonspecific staining, respectively. Antibodies and dyes used in the study are listed in Supplementary Table 1. IGRP H2-Kd/KYNKANVFL tetramer was obtained from the National Institutes of Health Tetramer Core Facility. Intracellular staining was performed overnight at 4ºC using fixation and permeabilization buffer (Foxp3 staining kit; eBioscience).

LEGENDplex Cytokine Analysis and In Vitro Stimulations

Cytokine analysis was performed according to the manufacturer’s instructions using undiluted supernatant from stimulation cultures. Duplicates were run for each standard and sample, and the average reading from each duplicate was plotted. For CD3⁻ stimulations, the LEGENDplex Mouse Inflammation Panel (13-plex) was used. For CD3⁺ stimulations, the LEGENDplex MU Th1/Th2 Panel (8-plex) V03 was used. Single-cell suspensions obtained from pLNs were stained 1:100 with CD3 biotin (Clone 17A2, catalog no. 100243; BioLegend) and enriched or depleted with antibiotin beads (Miltenyi). CD3⁻ cells were plated at a density of 1 × 10⁶/mL and stimulated with IFN-γ (1 µg/mL), lipopolysaccharide (LPS) (1 µg/mL), and CpG (1 µg/mL) for 24 h. CD3⁺ cells were plated at a density of 1 × 10⁶/mL and stimulated with phorbol myristic acid (50 ng/mL; Sigma-Aldrich) and ionomycin (100 ng/mL; Sigma-Aldrich) for 24 h.

Transgenic T-Cell Adoptive Transfers

Neonatal 10–12-day-old Thy1.1 NOD mice treated with Flt3L-APCs received one i.p. injection of 750,000 NOD.BDC2.5 Thy1.2 transgenic-naive CD62L⁺CD4⁺CD25⁻ T cells sorted from spleen and lymph nodes (90% purity) using an MA900 Cell Sorter (Sony Biotechnology). Five weeks after BDC2.5 T-cell transfer, NOD.Thy1.1 mice were euthanized, and organs were processed as described above for flow cytometry analysis.

In Vitro Cell Stimulation and Intracellular Cytokine Analysis

pLNs and islets were processed to a single-cell suspension as described above. Dissociated islets and pLNs were stimulated for 6 h with 10 ng/mL phorbol myristic acid (and 1 μmol/L ionomycin in the presence of brefeldin A (BFA) (5 mg/mL) and monensin (2 mmol/L) (38). pLN samples were cultured at 1 × 10⁶ cells; islets were cultured at one mouse per well in 24-well plates. Cells were washed and stained for cell membrane markers followed by overnight staining for intracellular markers and cytokines.

BrdU In Vivo Administration and Analysis

Prediabetic 16-week-old NOD mice were injected with 200 μL of 10 mg/mL BrdU i.p. and analyzed 6 h after injection. Organ suspension and surface staining of single-cell organ suspensions was performed as described above, followed by live/dead staining and fixation with fixation and permeabilization buffer (Foxp3 staining kit) for 30 min at 4ºC. Cells were washed with permeabilization buffer and stained overnight for intracellular markers at 4ºC. Next, cells were washed and resuspended in BD Cytofix/Cytoperm buffer, incubated at room temperature for 10 min, and washed with BD permeabilization wash buffer. Cells were incubated in PBS with DNase I at 37ºC for 30 min, followed by BD permeabilization buffer wash and staining with anti-BrdU conjugated to phycoerythrin at room temperature for 30 min.

In Vivo Cytokine Analysis

Analysis of in vivo cytokine production was performed as previously described (29). BFA powder was dissolved in DMSO (25 μg/mL) and stored at −20ºC. BFA was thawed and diluted to a working concentration of 0.5 mg/mL, mixed with 0.5 mg/mL monensin in 500 μL of sterile PBS, and injected via tail vein. Six hours postinjection, mice were euthanized; pancreatic islets and lymph nodes were analyzed by flow cytometry as described above.

Statistical Analysis

GraphPad Prism 10 was used for all statistical analyses of the experiments. Data are presented as mean ± SD unless otherwise specified in the figure legends. P values were calculated using the Mann-Whitney U test for comparisons between two groups or Kruskal-Wallis and Dunn test for comparisons of more than two groups. Diabetes incidence was compared using the log-rank (Mantel-Cox) test.

Data and Resource Availability

The data sets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Results

Adoptive Transfer of Flt3L-Derived Immature APCs to Neonatal NOD Mice Delays Incidence of T1D

To determine whether transfer of Flt3L-derived APCs to neonates establishes tolerance and confers protection to T1D, we used an in vivo system to expand APCs for sufficient numbers in adoptive transfers studies (Fig. 1A) (30). Approximately 14–19 days following subcutaneous B16-Flt3L tumor implant, NOD.Scid mice developed splenomegaly, and the frequency of CD11c⁺ cells increased to ∼60% compared with ∼20% in untreated NOD.Scid mice or NOD mice (Fig. 1A and B). The total number of splenic CD11c⁺ cells increased by 2 logs in NOD.Scid + Flt3L mice compared with nontreated NOD.Scid and NOD mice (Fig. 1C). In the CD11c^−/lo fraction, we observed a twofold increase in the frequency of CD11c⁻CD11b⁺ myeloid cells (80%), with a decrease in the frequency of CD11c^lo B220⁺ plasmacytoid DCs (pDCs), natural killer cells (Nkp46⁺), and neutrophils (CD11b Gr-1⁺) (Fig. 1D). We also assessed the maturation state of CD11c⁺ cells, given that DC maturation was previously connected to T1D development (18,31,32). We found no changes in the expression of the costimulatory molecule CD40 but a significant decrease in expression of MHC class II (MHC-II) and another costimulatory molecule, CD80, in Flt3L-APCs (CD11c⁺) compared with untreated NOD (Fig. 1E). We classified these CD11c⁺ cells as being immature based on low levels of MHC-II, low expression of costimulatory molecule CD80, but comparable levels of CD40 and CCR7, a chemokine known to drive APCs to T-cell zones (Fig. 1E) (31). However, Flt3L-APCs were capable of maturation based on their ability to upregulate MHC-II and costimulatory molecules upon stimulation with LPS and CpG (Supplementary Fig. 1). Having established the phenotype of Flt3L-APCs, we next treated female neonatal NOD mice to test whether Flt3L-APCs could delay the onset of T1D. In support of our hypothesis, we observed a reduction in disease incidence with 70% protection in neonatal NOD mice treated with Flt3L-APCs (Fig. 1F). This constituted a 20–40% increase in protection compared with previous studies of DC transfers or in vivo Flt3L treatment of adult mice (10–14,17,18).

Figure 1.

Adoptive transfer of immature APCs early in life significantly delays the onset of diabetes in NOD mice. A: Experimental schematic of in vivo–generated APCs in the spleens of NOD.Scid mice immunized with 2 × 10⁶ Flt3L-producing melanoma tumor cells. Flt3L-derived APCs were injected into neonatal NOD mice every 3 days from days 0–2 for a total of six injections. B and C: Representative flow plots and frequency and numbers of CD11c⁺ cells in spleens of NOD, NOD.Scid, and NOD.Scid + Flt3L-APC mice and cell counts of CD11c⁺ cells (n = 7 per group; Tukey one-way ANOVA). D: Frequency of subsets in CD11c⁻ in spleens of NOD.Scid and NOD.Scid + Flt3L mice (n = 3–4 per group). E: Expression of maturation markers MHC-II, CD40, CD80, and CCR7 of CD11c⁺ splenocytes in NOD, NOD.Scid, and NOD.Scid + Flt3L mice (n = 3–4 per group; one-way ANOVA). F: Diabetes incidence for NOD mice treated with Flt3L-derived APCs or PBS. Gray bar indicates treatment period with APC adoptive transfers from age 0 to 15–17 days. Log-rank (Mantel-Cox) test. Data are mean ± SD. **P ≤ 0.01, ****P ≤ 0.0001. pDC, plasmacytoid DC.

Donor APCs Travel to the pLNs and Alter T-Cell Priming

To determine whether donor APCs travel to major lymphoid organs, we used congenic NOD.Scid.CD45.2 mice as APC donors to neonatal NOD.CD45.1 recipients. Twenty-four hours after the last Flt3L-APC injection we observed trafficking of donor CD45.2⁺CD11c and CD45.2⁺CD11c⁻ cells to the thymus, spleen, pLNs, and islets of recipients (Fig. 2A and B). While the highest frequency of donor CD45.2⁺CD11c⁺ cells was found in the thymus (30–60%) (Fig. 2B), we did not observe major changes to the thymic cellular composition, Foxp3⁺ regulatory T-cell development, or pancreas-specific T cells (Supplementary Fig. 2). The activation state of DCs has been linked to progression to diabetes (10,17,18,31,32); thus, we compared the activation state of donor CD11c⁺ cells to endogenous CD45.2⁺ DCs 1 week after the last injection. We found that donor DCs expressed comparable levels of MHC-II, CD40, and CD80 to endogenous CD45.1⁺CD11c⁺ cells and PBS controls (Fig. 2C), indicating that Flt3L-derived DCs responded to the local environment and increased their capacity as APCs, while endogenous DCs were not affected by transferred cells. However, we also observed a significant reduction in MHC-II, CD40, and CD80 expression from the donor CD11c⁻ cells compared with CD11c⁺ cells, indicating an immature and possibly tolerogenic APC population present in the pLNs (Fig. 2C, purple bar). Looking more closely at the CD11c⁻CD11b⁺ donor population, the CD11b cells expressed significantly less MHC-II in the pLNs and accumulated at higher rates in the islets compared with the pLNs (Fig. 2D). Prior work demonstrated the necessity of pLNs to prime autoreactive T-cells (6–9,32). To resolve whether Flt3L-APC transfers alter the pLN cytokine microenvironment, CD3-depleted cells were obtained from pLNs of PBS- and APC-treated mice 1 week after the last Flt3L-APC injection and stimulated with LPS, CpG, and IFN-γ, and supernatants were analyzed using Luminex. The results indicated that the CD3⁻ fraction of the pLNs secreted higher levels of TNF-α compared with PBS control mice (Fig. 2E). TNF-α has been described as a pleiotropic cytokine, exerting both tolerogenic and inflammatory functions in T1D (20–22). The increase in TNF-α was not due to pLN-derived T cells (Supplementary Fig. 3A). To distinguish whether the CD11c⁺ and/or CD11c⁻ fractions of APCs could suppress proliferation in an antigen-specific manner, we sorted diabetogenic Thy1.2 BDC2.5 naive (CD4⁺CD62L⁺) T cells and cocultured with NOD CD11c⁺ DCs, Flt3L-derived CD11c⁺ DCs, CD11c⁻ APCs, or bulk APCs with or without 0.1 µmol/L 2.5 hybrid insulin peptide for 96 h (Supplementary Fig. 4). Although we did not observe any differences in in vitro proliferation, there was a significant reduction in downstream T-cell receptor (TCR) signaling and activation markers (CD44, CD69, Nur77, and Nor1) in BDC2.5 T cells when cultured with Flt3L-derived APCs, with the biggest reduction cultured with Flt3L-derived CD11c⁻ APCs. Therefore, donor APCs that travel to the pLNs early in life may alter T-cell priming possibly by exposure to a combination low MHC and costimulator molecules and TNF-α.

Figure 2.

Immature DCs migrate to the pLNs and alter T-cell priming. NOD mice were analyzed 24 h after the last splenocyte injection at days 16–19. A: Number of donor CD45.2⁺ cells within recipient thymus, spleen, pLNs, and islets of DC-treated mice and the ratio of donor/host cells in the pLNs and islets (two-way ANOVA). B: Frequency of CD11c⁺ and CD11c⁻ cells in the CD45.2⁺ donor population within recipient thymus, spleen, pLNs, and islets of APC-treated mice. C: Geometric mean fluorescence intensity (gMFI) of maturation markers MHC-II, CD40, and CD80 in endogenous CD11c⁺ populations or donor CD11c⁺ and CD11c⁻ populations within pLNs (ordinary one-way ANOVA). D: Percentage CD11b⁺ in pLNs and islets (two-way ANOVA) and gMFI of MHC comparing host and donor in pLNs (Mann-Whitney U test). E: Twenty-four-hour supernatant cytokine profile of CD3⁻ cells from pLNs of PBS- and APC-treated mice stimulated with LPS (1 µg/mL), CpG (1 µg/mL), and IFN-γ (1 µg/mL). Measured using LEGENDplex multiplex immunoassay. Data combined from six independent experiments (n = 3–4 mice pooled; Mann-Whitney U test). Data are mean ± SD. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Flt3L-Induced APCs Increase CD4⁺ T-Cell Anergy and Boost CD4⁺CD25⁺ T-Cell Population

To further determine the impact of APC transfers on the T-cell phenotype, we examined pLNs of 3-week-old preweaning mice. APC recipients had a twofold increase in the total number of CD3⁺ T cells with an increase in both CD4⁺ and CD8⁺ T-cell compartments (Supplementary Fig. 5A). The increased cellularity was localized to the pLNs, as we did not detect changes to the nondraining lymph nodes (ndLNs) or the spleen (Supplementary Fig. 5B). The increase in the levels of IL-2 (Supplementary Fig. 3B) at the pLNs led us to investigate the generation of CD4⁺CD25⁺ T cells, documented to compete with effector T cells for IL-2, facilitate conversion of anergy, and produce IL-10 (33–36). First, using FR4 and CD73 as markers of T-cell anergy (33,34), we observed an increase in anergic cells within activated CD4⁺CD44^hi T cells in both frequency and number in APC-treated mice compared with PBS (Fig. 3A and Supplementary Fig. 5C). In accordance with hallmarks of anergy (37,38), CD4⁺Foxp3⁻ T cells demonstrated a reduction in active proliferation (Fig. 3B, top) and a twofold increase in Egr2⁺PD1⁺ cells (Fig. 3B, bottom). To determine whether such changes persist into adulthood, we analyzed 16-week-old prediabetic NOD mice. The cellular expansion and anergy seen in the pLNs of young mice were transient, and cell numbers and anergy in pLNs were normalized by age 16 weeks (Supplementary Fig. 5D). Although upregulation of CD25 is usually associated with T-cell activation, CD4⁺CD25⁺ T cells were previously reported to be protective in autoimmunity and generated in response to DC-derived TNF-α (22,35,36,39,40). Interestingly, we observed an increase in frequency and number of CD4⁺CD25⁺Foxp3⁻ T cells in preweaned APC-treated mice, as well as in the 16-week-old prediabetic mice (Fig. 3C and D). This may indicate that some autoreactive T cells are driven to anergy when exposed to high numbers of immature APCs secreting TNF-α (Fig. 2C) early in activation while other T cells are driven to a CD4⁺CD25⁺Foxp3⁻ phenotype (36). To test whether the tolerogenic effects are extended to β-cell antigen–specific T cells in APC recipient NOD mice, we sorted diabetogenic Thy1.2 BDC2.5 naive (CD4⁺CD62L⁺) T cells and transferred 0.75 × 10⁶ cells into 10–12-day-old Thy1.1 NOD mice (Fig. 3E). Five-weeks after the mice received Flt3L-APCs or PBS, we analyzed the islets and pLNs. We detected similar frequencies of donor BDC2.5 Thy1.2⁺CD4⁺ T cells in both PBS- and APC-treated mice in the islets and pLNs (Supplementary Fig. 5E), indicating that trafficking and expansion of autoreactive CD4⁺ T cells was not impacted by APCs. However, the donor BDC2.5 Thy1.2⁺CD4⁺ T cells infiltrating the islets adopted an anergic phenotype where ∼20% of BDC2.5 T cells at the islets became anergic compared with ∼5% in the PBS group (Fig. 3F). Additionally, when we analyzed donor activated (CD11a⁺) T cells, we saw an increase in CD4⁺CD25⁺Foxp3⁻ frequencies at the islets and pLNs of APC mice (Fig. 3G). Our data demonstrate that by targeting the neonatal window with APC transfers, we can alter antigen-specific autoreactive CD4⁺ T-cell fate into adulthood.

Figure 3.

Adoptive transfer of immature APCs transiently expands the pLNs, rendering CD4⁺ T-cell anergy and a CD4⁺CD25⁺Foxp3⁻ population. Lymphoid organs were isolated from 3-week-old or 16-week-old NOD mice and analyzed by flow cytometry. A: Representative flow plots and frequencies of FR4/CD73⁺ cells gated on lymphocytes/singlets/live/CD3⁺CD4⁺Foxp3⁻CD44^hi from three independent experiments (n = 7 PBS-treated mice and n = 8 APC-treated mice; Mann-Whitney U test). B: As gated in A, representative flow plot and frequencies of Egr2⁺PD1⁺Foxp3⁻ T cells (n = 4 PBS-treated and n = 7 APC-treated mice; unpaired t test). C: Representative flow plots and frequencies of CD44⁺CD25⁺Foxp3⁻ T cells from three independent experiments (n = 8 PBS-treated and n = 6 APC-treated mice; Mann-Whitney U test). D: Sixteen-week-old NOD mouse anergy and CD4⁺CD25⁺Foxp3⁻ frequency data combined from three independent experiments (n = 11 PBS-treated and n = 10 APC-treated mice; Mann-Whitney U test. E: Experimental schematic of Thy1.2 BDC2.5-naive (CD4⁺CD62L⁺CD25⁻) T-cell transfer to 10–12-day-old Thy1.1 neonatal NOD mice treated with PBS or APCs. F: Representative flow plots and frequencies of anergic FR4⁺CD73⁺CD4⁺Foxp3⁻ T cells within donor BDC2.5 cells gated on lymphocytes/singlets/live donor Thy1.2 cells (Mann-Whitney U test). G: Representative flow plots and frequencies of CD4⁺CD25⁺Foxp3⁻ T cells gated on donor cells from two independent experiments (n = 5–7 per group; Mann-Whitney U test). Data are mean ± SD. *P ≤ 0.05, **P ≤ 0.005. tg, transgenic.

Adoptive Transfers of Immature Flt3L-Derived APCs to Neonatal NOD Mice Decreases CD4⁺ Th1 Effector Function in Prediabetic NOD Mice

The increase in CD4⁺CD25⁺Foxp3⁻ T cells sustained through disease development (Fig. 3C) in both islets and pLNs led us to investigate whether these T cells can produce IL-10. To test our hypothesis, we stimulated T cells isolated from the pLNs of PBS- and APC-treated mice at weaning in vitro to measure production of IL-10 via intracellular staining. Indeed, we observed a threefold increase in frequency of IL-10⁺CD25⁺Foxp3⁻ T cells in the APC-treated group compared with the PBS controls (Fig. 4A). However, we did not detect a difference in IL-10⁺CD25⁺Foxp3⁻ T cells in prediabetic mice (Supplementary Fig. 6), indicating that other mechanisms may be involved in long-term tolerance.

Figure 4.

APC neonatal transfers to NOD mice increase the frequency of IL-10⁺ T cells in young mice and lead to a sustained decrease in Th1 CD4⁺ T cells. Lymphoid organs from preweaning and prediabetic NOD mice were isolated and analyzed via flow cytometry. A: Representative flow plots of IL-10⁺ T cells from preweaning mice stimulated in vitro. Analysis was gated on lymphocytes/singlets/live/CD4⁺TCR-b⁺/Foxp3⁻/CD25⁺CD44⁺ T cells from three independent experiments (n = 11–14 per group; Mann-Whitney U test). B: Representative flow plots and frequencies of CXCR3⁺ T-bet⁺ cells gated on lymphocytes/singlets/live/CD4⁺TCR-b⁺. C: Representative flow plots of CD4⁺TCR-b⁺ cells stimulated in vitro and gated as described above. Shown are frequencies and geometric mean fluorescence intensity (gMFI) normalized to unstimulated samples of neonatal PBS- and APC-treated mice (n = 6–7 per group; Mann-Whitney U test). D: Representative flow plots, frequencies, and numbers of CXCR3⁺ T-bet⁺ cells gated as above in the islets of prediabetic (16-week-old) PBS- and APC-treated mice (Mann-Whitney U test). E: Representative flow plots and frequencies of T-bet⁺ IFN-γ⁺ T cells from islets and pLNs of prediabetic (16-week-old) PBS- and APC-treated mice. Cells were stimulated in vitro and gated as described above from two independent experiments (n = 8–10 per group; Mann-Whitney U test). Data are mean ± SD. *P ≤ 0.05, **P ≤ 0.005, ***P ≤ 0.001. SSC-A, side scatter area.

To further understand how tolerance is maintained long term, we investigated whether alterations in TNF-α production observed in APCs (Fig. 2C) may inhibit lineage commitment to disease-driving Th1 T cells (23,39,40). We therefore analyzed markers of Th1 lineage, T-bet and CXCR3, in preweaning and prediabetic mice (41–43). CXCR3⁺T-bet⁺ T cells decreased in the pLNs of 3-week-old APC recipients (20% vs. 30%), with no change in the ndLNs (Fig. 4B). To test Th1 functionality, we isolated T cells from the pLNs of PBS- and APC-treated preweaning mice and stimulated cells in vitro to measure inflammatory cytokines by intracellular staining. We observed a significant increase in IFN-γ⁺T-bet⁺ Th1 cells in 3-week-old APC-treated mice compared with controls (Fig. 4C). It is possible that the Th1 phenotype that we observed in young mice is transient; therefore, we also analyzed the frequency and function of Th1 T cells in the islets and lymph nodes of prediabetic NOD mice. Consistent with reduced autoimmunity in APC-treated mice, we observed a 50% decrease in the frequency of CXCR3⁺T-bet⁺ T cells in APC recipients from ∼15% to ∼30% Th1 cells in PBS-treated mice (Fig. 4D). IFN-γ⁺ cytokine production by T-bet⁺ cells was also diminished in prediabetic APC recipients, with a 50% decrease in the frequency of IFN-γ⁺ Th1 cells within the islets (Fig. 4E). Consequently, Flt3L-APC transfers alter the pLN cytokine milieu early in disease, increasing the frequency of both IL-10– and IFN-γ–producing T cells in young mice but compromising Th1 T-cell function at later stages of the disease.

Neonatal NOD Mice Treated With Immature APCs Show a Decrease in Proliferation and Function of Diabetogenic CD8⁺ T Cells

Islet infiltration and proliferation of autoreactive CD8⁺ T cells are thought to be indispensable for T1D development, an event dependent on perforin secretion by autoreactive CTLs driving β-cell destruction (6,44–46). For autoreactive CD8⁺ T cells to become activated, interactions with both APCs and CD4⁺ T cells are necessary (23,40). Our data thus far indicate that we have altered both the APC niche and the frequency and function of Th1 cells (Figs. 2 and 4). Therefore, we next evaluated the impact on CD8⁺ T cells in 16-week-old prediabetic mice. Similar to CD4⁺ T cells, we observed no change in total CD8⁺ T-cell infiltration at 16 weeks (Supplementary Fig. 7), suggesting that T-cell trafficking was not impaired. Nonetheless, we observed a decrease in CD8⁺ T-cell proliferation (Ki67⁺ and BrdU labeling) in the islets of APC-treated NOD mice compared with islet-infiltrating T cells from controls (Fig. 5A). To evaluate whether autoreactive CD8⁺ T-cell clones are affected in adulthood by neonatal APC transfers, we analyzed frequencies of tetramer⁺ CD8⁺ T cells specific for glucose-6-phosphatase catalytic-subunit–related protein (Igrp) peptide targeted by a highly pathogenic T-cell clone (45). We observed a 50% decrease in the frequency of CD8⁺Igrp tetramer⁺ cells in the islets of prediabetic mice (Fig. 5B). Hence, early treatment with immature APCs impacts the proliferation of infiltrating CD8⁺ T cells (Fig. 5A), including pathogenic Igrp-specific CD8⁺ T cells (Fig. 5B). To determine whether infiltrating autoreactive CD8⁺ T cells are functional and consequently drive β-cell destruction, we used in vivo treatment with BFA and monensin (29) to measure perforin production. In accordance with the observed protection (Fig. 1E), we saw a 50% decrease in the frequency of perforin⁺ CD8⁺ T cells at the islets of prediabetic APC-treated mice, where ∼20% of CD8⁺ islet infiltrates were perforin⁺ in controls compared with ≤10% perforin⁺ CD8⁺ T cells in the APC recipient group (Fig. 5C). Collectively, our results propose a mechanism of protection by which neonatal APC transfers decrease the function of CTLs later in life to preserve β-cells.

Figure 5.

Decrease in activation, frequency, and cytolytic capacity of autoreactive CD8⁺ T cells in APC-treated mice. Islets and pLNs were isolated from prediabetic PBS- or APC-treated 16-week-old NOD mice. A: Representative flow plots and frequencies of BrdU and Ki67⁺CD8⁺ T cells gated on lymphocytes/singlets/live/CD8⁺TCR-b⁺. Data are combined from three independent experiments (n = 8–10 per group; Mann-Whitney U test). B: Representative flow plots and frequencies of Igrp tetramer⁺CD8⁺ T cells gated as described above from two independent experiments (n = 8–9 per group; Mann-Whitney U test). C: Representative flow plots and frequencies of ex vivo perforin⁺CD8⁺ T cells gated as described above from two independent experiments (n = 4–8 per group; Mann-Whitney U test). Data are mean ± SD. *P ≤ 0.05, **P ≤ 0.005. SSC-A, side scatter area.

Discussion

Soluble Flt3L treatment and DC transfers to prevent progression to diabetes has been demonstrated with low to intermediate success in adult NOD mice (10–14). This may be due to early immune cell infiltrates primed in pLNs contributing to pathogenesis during the first weeks of life (8,9). Therefore, we hypothesized that APC/DC-driven protection is localized to pLNs and islets, consistent with the important role for pLNs in early disease (7–9,17). Supporting our hypothesis, we observed an increase in anergic CD4⁺ T cells and a shift to a CD25⁺Foxp3⁻ phenotype specifically in pLNs and islets. Prior studies categorized CD25⁺ cells as tolerogenic (33,35,36,39,40) and suggested that anergic CD4⁺ T cells serve as a reservoir for IL-10–secreting T cells (36). Indeed, we found an accumulation of both anergic and IL-10⁺Foxp3⁻ T cells in 3-week-old APC-treated mice but not in older prediabetic mice. Thus, we hypothesize that autoreactive CD4⁺ T cells become anergic in the presence of suboptimal APC engagement, leading to long-term T-cell dysfunction with a decrease in CTL proliferation and Th1 differentiation (Figs. 4 and 5). Our results support the importance of the neonatal time window not only for development of central tolerance but also for peripheral tolerance.

Early events in self-reactive T-cell priming and T-cell lineage commitment are likely influenced by a combination of multiple signals, including the quality of the APCs and the cytokine milieu. The specific role of TNF-α in early stages of immune cell priming is unclear; however, TNF-α is clearly a pleotropic cytokine with multiple functions. Systemic TNF-α administration to neonatal NOD mice accelerates T1D, but treatment past 4 weeks of age delays disease onset (21). The pleiotropic nature of TNF-α is speculated to depend on which receptor overtakes intracellular signaling, TNF receptor 1 (TNFR1) or TNFR2 (24). Induction of TNFR1 preferentially activates canonical nuclear factor-κB while binding to TNFR2 activates the noncanonical pathway known to induce cell survival and proliferation (24). It is yet to be determined whether there is a preferential role for a TNFR1 or TNFR2 intracellular response leading to long-term tolerance in our studies; however, notable functional and transcriptional differences between adults and neonate follicular helper T cells and regulatory T cells have been reported (47–49).

In this study, we used transfers of DC-enriched splenocytes at the neonate stage to prevent diabetes onset. The immature APCs (CD11c⁺ and CD11⁻) engage and tolerize naive T cells, resulting in a 70% decrease in disease incidence. The increase in IL-10–producing T cells and the increase in FR4/CD73 anergic T cells early in life leads to an overall decrease in functional Th1 T cells and CTLs at the islets of Langerhans later in life, demonstrating the advantage of targeting the neonate window and the pLN to alter the onset of T1D.

This article contains supplementary material online at https://doi.org/10.2337/figshare.30347107.

Funding Statement

This study was supported by The Robert and Janice McNair Foundation (to M.L.B. and M.B.), National Institutes of Health grants R01DK114456, R01AI136963, and R01AI173406 (to M.L.B.); R01AI175494 (to M.B.); F31AI161946 and T32AI138945 (to S.O.); and T32AI138945 (to H.M.R.).