TLR2- and Dectin 1–Associated Innate Immune Response Modulates T-Cell Response to Pancreatic β-Cell Antigen and Prevents Type 1 Diabetes

Innate immunity, initiated primarily by environmental factors such as microbes, plays a key role in initiating or preventing the T-cell response to pancreatic β-cell-Ag in type 1 diabetes (T1D). Although it has been suggested that the proinflammatory response mediated by pathogen recognition receptors (PRRs) facilitates β-cell-Ag presentation by activated antigen-presenting cells (APCs) (1), environmental factors such as bacterial and viral infections are also known to have a protective effect in T1D (2–5). Innate immune response is mediated by an array of PRRs such as Toll-like receptors (TLRs) and C-type lectin receptors (CLRs) that primarily recognize microbial products. In recent years, studies, including ours, have shown that innate immune responses induced through TLR2 and Dectin 1 using zymosan, a fungal cell wall component, are regulatory in nature and involve, in addition to proinflammatory factors, the expression of IL-2, IL-10, TGF-β1, and retinaldehyde dehydrogenase 1A2 (Raldh1A2) by one or other type of APCs (6–13). Importantly, the innate immune response induced by zymosan has the ability to prevent/delay disease in T1D and experimental autoimmune encephalomyelitis (EAE) models, even upon disease onset (6–11).

In this report, we show that zymosan-induced innate immune response facilitates regulatory T-cell (Treg) induction and/or expansion and Th1 to Th17 skewing of the T-cell response to pancreatic β-cell-Ag. Importantly, treatment with zymosan along with β-cell-Ag resulted in a significant delay in hyperglycemia in NOD mice even when the treatment was initiated at an early hyperglycemic stage as compared with treatment with zymosan alone. These observations show that zymosan has therapeutic values as a tolerogenic adjuvant and can be used for promoting β-cell-Ag–specific tolerance and to reverse early-stage hyperglycemia in T1D.

Wild-type (WT) NOD/LtJ, NOD-BDC2.5-TCR transgenic (TCR-Tg), NOD-Scid, C57BL/6, and OT-II-TCR-Tg mice were purchased from The Jackson Laboratory (Bar Harbor, ME). OT-II-TCR-Tg-Foxp3-GFP-knockin, NOD-Foxp3-GFP, and NOD-BDC2.5-Foxp3-GFP mice were generated at our animal facilities. To detect hyperglycemia in NOD mice, glucose levels in blood collected from the tail vein of WT NOD-Ltj or NOD-Scid mice were monitored using the Ascensia Microfill blood glucose test strips (Bayer, Mishawaka, IN). All animal studies were approved by the animal care and use committee of University of Illinois at Chicago (UIC) and the Medical University of South Carolina (MUSC).

Immunodominant β-cell-Ag peptides, viz. 1) insulin B (9–23), 2) GAD65 (206–220), 3) GAD65 (524–543), 4) IA-2β (755–777), 5) IGRP (123–145), 6) BDC2.5 TCR reactive peptide (YVRPLWVRME; referred to as BDC peptide), and 7) OVA (323–339) peptides, were described in our earlier studies (14–16). Peptides 1–5 were pooled at an equal molar ratio and used as β-cell-Ag for in vitro and in vivo experiments.

Purified zymosan A of Saccharomyces cerevisiae was prepared as described previously (6,7). Bacterial lipopolysaccharide (LPS; Escherichia coli origin, ion-exchange purified), curdlan, phorbol myristic acid (PMA), ionomycin, brefeldin A, and monensin were purchased from Sigma-Aldrich, BD Biosciences, eBioscience, Invivogen, and Invitrogen. Normal rat serum, various fluorochrome-conjugated reagents and antibodies (Abs), and isotype control Abs (Invitrogen, BD Biosciences, eBioscience, R&D Systems, and Biolegend Laboratories) were used for FACS. Magnetic bead-based total and CD4⁺ T-cell and CD11c⁺ dendritic cell (DC) isolation kits (Miltenyi Biotec and Invitrogen) were used for enriching or depleting T cells and DCs. Paired Abs and standards for ELISA were purchased from R&D Systems, BD Biosciences, Invitrogen, and eBioscience.

Twelve-week-old euglycemic (glucose levels <110 mg/dL; prediabetic age) and 10–20-week-old early hyperglycemic (glucose levels between 140 and 250 mg/dL; early hyperglycemic stage) WT female NOD/Ltj mice were treated with zymosan and/or β-cell-Ag. Although insulitis in NOD mice is very heterogeneous at any given age, 12-week-old euglycemic mice represent the prediabetic stage because the hyperglycemia begins to appear at this stage. Mice were injected with zymosan (i.v. 25 μg/mouse/day on days 1, 3, 5, 16, 18, and 20) in PBS. Some groups of mice were injected intravenously with β-cell-Ag (0.5 μg/mouse/day on days 5 and 20) in PBS. Mice with glucose levels >250 mg/dL for two consecutive weeks were considered diabetic.

Four-week-old NOD-BDC2.5-Foxp3-GFP mice were injected intravenously with bacterial LPS (5 μg/mouse/day) or zymosan (25 μg/mouse/day) for three consecutive days and BDC2.5 peptide (5 μg/mouse) on day 3. Spleen cells from these mice (2 × 10⁶/mouse) were injected into 4-week-old WT NOD mice and examined for blood glucose levels every other day. In a separate experiment, 4-week-old WT NOD mice were treated with zymosan and/or β-cell-Ag as described above and injected with 2 × 10⁶ ex vivo–activated BDC2.5-TCR-Tg T cells. The recipients were monitored for hyperglycemia every other day. In one set of experiments, 12-week-old euglycemic WT NOD mice were treated with zymosan and/or β-cell-Ag as described above, and CFSE-labeled BDC2.5-TCR-Tg T cells were injected intravenously (2 × 10⁶/mouse). After 96 h, spleen and pancreatic lymph node (PnLN) cells from these mice were stained for Vβ4 and CD4 and examined for proliferating T cells by FACS.

Splenic CD11c⁺ cells and total T cells were enriched using magnetic separation reagents. Bone marrow–derived DCs (BMDCs) were generated as described previously (6). Splenic CD11c⁺ DCs and BMDCs were incubated with zymosan (25 μg/mL) or LPS (2 μg/mL) for different time points. Pilot experiments using varying amounts of these agents were performed to determine the optimum culture conditions based on TNF-α production. cDNA prepared from DCs was subjected to qualitative and real-time quantitative PCR using SYBR Green PCR Master Mix (ABI Prism). DCs from 36-h cultures were examined for the levels of surface activation markers after staining with fluorochrome-labeled specific Abs. Spent media were tested for cytokine levels by ELISA.

Purified total T cells (1 × 10⁵ cells/well) were incubated with zymosan-exposed or unexposed DCs (2 × 10⁴ DCs/well) in the presence of anti-CD3 Ab (2 μg/mL). In some assays, DCs were incubated with β-cell-Ag (5 μg/mL) and zymosan (25 μg/mL) or LPS (2 μg/mL) agents overnight, washed, and cultured with CD4⁺ T cells. In some assays, spleen and PnLN cells (2 × 10⁵ cells/well) from treated and control mice were stimulated with anti-CD3 Ab (2 µg/mL) or β-cell-Ag (5 μg/mL) for 48 h. Spent media from these cultures were tested for cytokines.

Freshly isolated and ex vivo–cultured cells were washed using PBS supplemented with 2% FBS and 10 mmol/L EDTA (pH 7.4) and blocked with anti-CD16/CD32 Fc block Ab or 5% rat serum on ice for 15 min. For surface staining, cells were incubated with fluorescein isothiocyanate–, phycoerythrin–, and phycoerythrin-cyanine 5– or phycoerythrin-Texas Red–labeled appropriate Abs, in different combinations, on ice for 45 min and washed three times before analysis. Cells were also stained using isotype-matched control Abs for determining the background.

BMDCs were cultured in the presence of zymosan (25 μg/mL) or LPS (2 μg/mL) with β-cell-Ag (5 μg/mL) or Ag alone for 24 h, washed, and injected into 8-week-old prediabetic female NOD mice (i.v. 5 × 10⁶ cells/mouse). In some experiments, total splenocytes (5 × 10⁶ cells/mouse) or purified T cells (2 × 10⁶ cells/mouse) from the spleens of control and treated mice (15 days posttreatment) were transferred into euglycemic female NOD mice and monitored as described above. In some experiments, freshly isolated T cells from BDC2.5-TCR-Tg mice were labeled with CFSE and injected into zymosan and/or β-cell-Ag–treated WT NOD mice. The recipients were killed 4 days post–T-cell transfer to determine the proliferation of donor T cells. In addition, BDC2.5-TCR-Tg T cells were activated using control and zymosan-exposed and BDC2.5 peptide–pulsed DCs, and T cells purified from these cultures were injected into 6-week-old WT mice (intravenously).

Pancreata were fixed in 10% formaldehyde and 5-µm paraffin sections were made and stained with hematoxylin-eosin (H-E). Stained sections were analyzed using a grading system in which 0 = no evidence of infiltration, 1 = peri-islet infiltration (<5%), 2 = 5–25% islet infiltration, 3 = 25–50% islet infiltration, and 4 = greater than 50% islet infiltration, as described in our earlier studies (6,14–17). About 150 islets were examined for every group. In some experiments, pancreatic sections were stained using anti-insulin and antiglucagon Abs followed by Alexa Fluor 488– and 568–linked secondary Abs and DAPI and scored for insulitis based on DAPI-positive cells in islet areas and insulin expression. Insulitis was scored as described for H-E–stained sections, and insulin-positive and -negative islets were counted. Areas that appeared to have completely lost islets were considered as grade 5 and included in this grading approach. In some experiments, H-E–stained kidney and liver sections were examined for inflammation and tissue damage. Further, serum samples collected from zymosan-treated and control mice were tested for C-reactive protein (CRP) levels by ELISA (R&D Systems).

Mean, SD, and statistical significance (P value) were calculated using Microsoft Excel, GraphPad online, and/or other online statistical applications. Two-tailed Student t test was used unless specified for values from in vitro and ex vivo assays. Log-rank analysis was performed to compare T1D incidence (hyperglycemia) of the test group with that of the control group. Fisher exact test was used for comparing the total number of infiltrated islets in test versus control groups. A P value of ≤0.05 was considered significant. In most experiments, unless specified, individual treated groups were compared with the nontreated group for calculating P values.

In this study, we compared the innate immune response induced by zymosan through TLR2 and Dectin 1 with that induced through TLR4 by bacterial LPS. BMDCs (generated using GM-CSF and IL-4, represent primarily myeloid DCs) and peripheral DCs were activated using LPS or zymosan and examined for the amount of cytokines secreted. As observed in Fig. 1A, whereas zymosan-exposed DCs produced large amounts of the immune regulatory cytokines IL-2, IL-10, and TGF-β1 along with proinflammatory cytokines (TNF-α, IL-6, IL-1β, and IL-12), LPS induced only proinflammatory cytokines in these DCs. We have also examined the ability of zymosan-exposed or unexposed DCs from NOD mice to express Raldh1A2 in comparison with DCs that are treated with LPS. As observed in Fig. 1B, exposure to zymosan, but not LPS, upregulated Raldh1A2 expression in both splenic and BMDCs, albeit at a much lower level (∼10-fold induction from basal level) than previously reported (∼200-fold induction from basal level) (10). However, expression of this enzyme, in combination with the observed immune regulatory cytokines, by DCs suggests that zymosan-exposed APCs may promote a Treg response upon Ag presentation.

Next, zymosan DCs were examined for their ability to modulate T-cell response using purified T cells from OT-II-TCR-Tg mice. As observed in Fig. 1C, DCs that were exposed to LPS and zymosan induced similar levels of proliferation in Ag-specific T cells. However, the zymosan-exposed DCs induced significantly higher frequencies of Foxp3⁺ cells in culture as compared with unexposed or LPS-exposed DCs. Examination of the cytokine profiles revealed that T cells activated using DCs that were exposed to zymosan produced significantly higher amounts of IFN-γ, IL-17, and IL-10 and reduced amounts of IL-4 as compared with untreated control DC-activated T cells (Fig. 1D). However, T cells that were activated using the TLR4 ligand produced high amounts of IFN-γ and IL-17 and lower amounts of TGF-β1, compared with control DC-activated T cells. These observations show that TLR2- and Dectin 1–engaged DCs and TLR4-engaged DCs modulate T-cell function differently.

Both BDC2.5-TCR-Tg cells and T cells from WT NOD mice demonstrate diabetogenic properties upon activation (14,17). Therefore, to further understand the effect of TLR2- and Dectin 1–mediated innate immune response on T cells in a T1D model, BDC2.5-TCR-Tg and WT NOD mouse T cells were activated in the presence of zymosan-exposed splenic DCs. The zymosan-exposed DCs induced significantly higher frequencies of Foxp3⁺ cells in the cultures of WT NOD mouse T cells as compared with unexposed or LPS-exposed DCs (Fig. 2A). On the other hand, when LPS-exposed DCs were used, considerably low frequencies of T cells in the cultures expressed Foxp3 as compared with controls (Fig. 2A). T cells from WT NOD mice produced higher IL-10 and TGF-β1 when activated in the presence of zymosan-exposed, but not LPS-exposed, DCs. (Fig. 2B). Similar to WT T cells, zymosan-exposed, but not LPS-exposed, DCs (both splenic and BMDCs) induced an increase in the frequency of Foxp3⁺ cells in BDC2.5-TCR-Tg T-cell cultures (Fig. 2C). BDC2.5-TCR-Tg T-cell cultures also showed significantly higher amounts of IL-10, TGF-β1, and IL-17 along with IFN-γ when zymosan-exposed DCs were used for Ag presentation (Fig. 2D). These observations show that innate immune response induced by zymosan, but not LPS, has the ability to modulate the properties of diabetogenic T cells and induce T cells with a regulatory phenotype.

To assess the immune regulatory nature of Ag presentation by zymosan-exposed DCs in vivo as compared with TLR4-engaged DCs, 8-week-old NOD mice were injected with zymosan- or LPS-exposed, β-cell-Ag–pulsed BMDCs and monitored for hyperglycemia. NOD mice that received zymosan-exposed, but not LPS-exposed, β-cell-Ag–pulsed DCs showed a significant delay in hyperglycemia as compared with β-cell-Ag–loaded control DC recipient mice (Fig. 3A). One set of euglycemic mice from a similar experiment was killed 30 days posttreatment and pancreatic tissues were examined for insulitis. As observed in Fig. 3B, mice that received zymosan-exposed and β-cell-Ag–pulsed DCs had significantly higher numbers of islets with less severe insulitis or no immune cell infiltration compared with untreated or control DC recipient mice. Spleen cells from one set of mice that were killed 15 days posttreatment were examined for the frequency of Foxp3⁺ T cells. As shown in Fig. 3C, the number of Foxp3-expressing CD4⁺ Tregs is significantly higher in mice that were injected with β-cell-Ag–pulsed, zymosan-exposed DCs as compared with β-cell-Ag–loaded (none) DC recipients or mice that received DCs that were not pulsed with β-cell-Ag (control). Although LPS-exposed, β-cell-Ag–loaded DC recipients showed a modest increase in Foxp3⁺ T cells compared with controls, they developed a more or less similar level of insulitis as the control DC recipients. These results suggest that β-cell-Ag presentation by the zymosan-exposed DCs has the ability to promote induction/expansion of Tregs in vivo.

To assess whether zymosan and LPS induce proinflammatory adjuvant effects in vivo and promote the generation of diabetogenic T cells, young BDC2.5-Foxp3-GFP mice were injected with bacterial LPS or zymosan along with the BDC2.5 peptide. Zymosan-injected NOD-BDC2.5-Foxp3-GFP mice showed a profound increase in the CD4⁺GFP⁺ T-cell numbers, compared with the control group of mice (Fig. 3D). On the other hand, LPS recipients showed lower CD4⁺GFP⁺ T-cell frequency compared with control mice. Importantly, T cells from only LPS-recipient, but not zymosan-recipient, mice could induce hyperglycemia in young NOD mice (Fig. 3E). T cells from mice that received BDC2.5 peptide alone did not induce hyperglycemia, perhaps due to tolerogenic Ag presentation by steady-state APCs in the absence of adjuvant treatment in these peptide-recipient mice. Whether intravenous injection with zymosan can lead to a strong inflammatory response was also assessed by determining the serum levels of CRP in zymosan-treated mice. As observed in Supplementary Fig. 1, a marginal increase in the serum CRP levels was detected 48 h after zymosan treatment. On the other hand, intravenous injection with LPS induced a fivefold increase in the serum CRP levels as compared with control. These results indicate that zymosan does not have a strong proinflammatory adjuvant property, as opposed to that of LPS.

Next, we examined whether zymosan treatment–induced modulation of immune response to β-cell-Ag affects T-cell proliferation in vivo. Twelve-week-old (prediabetic age) NOD mice were treated with zymosan and/or β-cell-Ag, injected with CFSE-labeled BDC2.5-TCR-Tg T cells alone or along with BDC2.5 peptide, and examined for CFSE dilution in donor cells by FACS. As anticipated, PnLN, but not the spleen, of control mice showed profound proliferation of BDC2.5 T cells in the absence of BDC2.5 peptide injection (Fig. 4A). Importantly, PnLN of all groups of mice that were treated with zymosan and/or β-cell-Ag also showed a significant number of proliferating BDC2.5 T cells. PnLN from all BDC2.5 peptide–injected mice showed a more or less similar extent of BDC2.5 T-cell proliferation (Fig. 4A). In addition, upon peptide injection, spleen cells from zymosan-treated and untreated mice showed a similar extent of BDC2.5 T-cell proliferation. These results, in association with the observations of Figs. 2 and 3, suggest that although the zymosan-induced innate immune response may not have a significant suppressive effect on T-cell proliferation, it may influence the functionality of these T cells.

Our previous report (14) has shown that young WT NOD mice injected with in vitro–activated TCR-Tg T cells can serve as an effective adoptive transfer model for T1D to assess if the Ag-specific T cells are pathogenic. In this model, if the TCR-Tg T cells are pathogenic, then 100% of recipient mice develop overt hyperglycemia within 6–8 days. If the T cells are not pathogenic in nature, then the recipients fail to develop T1D for months. Therefore, to further assess the regulatory nature of TLR2- and Dectin 1–dependent innate immune response, young WT NOD mice were treated with zymosan and/or β-cell-Ag, followed by injection of in vitro–activated diabetogenic BDC2.5-TCR-Tg T cells. As observed in Fig. 4B, diabetogenic BDC2.5-TCR-Tg induced overt hyperglycemia in control and β-cell-Ag–treated mice within 8 days. However, these T cells failed to induce hyperglycemia in zymosan- and zymosan plus β-cell-Ag–treated groups of mice. These results suggest that zymosan-induced innate immune response alters the function of diabetogenic T cells in vivo.

Various effects of TLR2- and Dectin 1–dependent innate immune response on secondary lymphoid organs as well as nonlymphoid organs were examined to further understand the potential side effects of zymosan treatment. Twelve-week-old (prediabetic age) mice were treated with zymosan, and various tissues were examined for cell death. Cells from the pancreas, PnLN, and spleen were examined for dead/dying (both apoptotic and necrotic) cells by Annexin V and 7-AAD staining. While spleen cells from zymosan- and zymosan plus β-cell-Ag–recipient mice did not show a difference in cell death, PnLN of these treated mice had a higher number of Annexin V– and/or 7-AAD–positive cells compared with control mice (Supplementary Fig. 2A). Importantly, CD4⁺ T cells from pancreas and PnLN of only zymosan plus β-cell-Ag–recipient mice, but not zymosan alone or the control group of mice, showed considerable cell death (Supplementary Fig. 2B). These results suggest that zymosan treatment causes immune cell death at sites of high inflammation, and in addition to the regulatory cytokine–mediated effect, apoptotic elimination of activated immune cells, including T cells, could be a feature of zymosan-induced immune modulation.

To assess whether the zymosan treatment affects DC numbers and their functionality, leading to T-cell death, pancreatic, PnLN, and spleen cells were examined for the frequencies of DCs. As observed in Supplementary Fig. 3A, treatment with zymosan at the dose selected for this study did not induce a considerable increase in the number of CD11c⁺ DCs in the pancreas, PnLN, or spleen. Further, the expression levels of apoptotic ligands FAS-L and TRAIL were comparable in the pancreas, PnLN, or spleen DCs of control and zymosan-treated groups of mice. Further analysis revealed that myeloid (CD8a⁻), lymphoid (CD8a⁻), and plasmacytoid (PDCA1) DC frequencies were not noticeably different in the spleen and PnLN of control and zymosan-treated mice (Supplementary Fig. 4). Although the Ag-presenting function of DCs from zymosan-treated and control mice may be different in vivo, these observations indicate that the frequency of DCs is not affected by zymosan treatment and they may not have a direct role in causing T-cell death in the pancreatic microenvironment of zymosan-treated mice.

We have also examined whether zymosan treatment induces side effects such as inflammation in nonlymphoid organs like kidney and liver. As observed in Supplementary Fig. 5, intravenous injection with low-dose zymosan does not cause noticeable inflammation in the nonlymphoid organs such as kidney and liver. To assess whether zymosan and β-cell-Ag treatment induces general immune suppression, the ability of T cells from treated NOD mice to proliferate against alloantigen in a mixed lymphocyte reaction assay was tested. As observed in Supplementary Fig. 6, T cells from zymosan- and zymosan plus β-cell-Ag–treated mice showed a more pronounced proliferative response than T cells from untreated or β-cell-Ag–treated NOD mice. While this enhanced proliferative ability of T cells from zymosan-treated mice, compared with control T cells, could be an effect of the cytokines IL-2 and TNF-α produced by zymosan-exposed APCs, this result suggests that zymosan treatment does not produce general immune suppression.

As shown in our previous report (6), the innate immune response induced in NOD mice by injecting a higher dose of zymosan (100 μg/injection, i.p.) could achieve long-term protection from T1D. Results from experiments using zymosan-exposed DCs presented in Fig. 3 suggest that a relatively short-term treatment, with smaller doses of zymosan, may be sufficient to achieve a long-lasting therapeutic effect in conjunction with β-cell-Ag. Twelve-week-old prediabetic age NOD mice were coadministrated with zymosan and β-cell-Ag and monitored for hyperglycemia. As observed in Fig. 5A, mice that were treated with zymosan and β-cell-Ag remained diabetes free for a significantly longer duration than the mice that received zymosan alone. In addition, as observed in our previous study (13), mice that received zymosan also showed significant protection from T1D as compared with untreated mice.

The pancreatic islets of prediabetic mice that received zymosan and β-cell-Ag showed significantly less severe immune cell infiltration and insulitis compared with peptide-treated and untreated control mice within 30 days posttreatment (Fig. 5B). Whereas the mice that received low-dose zymosan alone also showed significantly less severe insulitis compared with untreated mice, the difference in insulitis observed between pancreatic islets of mice that received β-cell-Ag alone and untreated controls was not statistically significant, at 30 days posttreatment. These results suggest that exposure of APCs to zymosan in vivo leads to the skewing of T-cell response from pathogenic to protective type.

Encouraged by the results of Fig. 5, we examined whether modulating the immune response against β-cell-Ag, with the help of zymosan-induced innate immune response, can control autoimmunity in early hyperglycemic mice and restore euglycemia. Early hyperglycemic (glucose 140–250 mg/dL) mice were treated with zymosan alone or along with β-cell-Ag and examined for blood glucose levels every 3rd day. One hundred percent of noninjected control mice were diabetic within 6 weeks posttreatment. However, the majority of the mice that received zymosan along with β-cell-Ag showed normal glucose levels and remained diabetes free for >20 weeks posttreatment (Fig. 6A and Supplementary Fig. 7). Although less profound, mice that received zymosan alone also showed significant protection from diabetes compared with control mice.

Examination of the pancreatic tissue from one set of these mice, 30 days posttreatment, revealed that zymosan- and β-cell-Ag–treated mice have a significantly higher number of infiltration-free islets or islets with less severe insulitis compared with other groups after 30 days posttreatment (Supplementary Fig. 8A). Another set of mice from a similar experiment were killed on day 30, and the pancreatic tissue sections were stained for insulin and glucagon to assess islet function. Although differences in the overall insulitis severities between zymosan and zymosan plus β-cell-Ag groups were not statistically significant in Supplementary Fig. 6, the latter group showed a profound increase in the number of insulin-positive islets compared with zymosan recipients (Fig. 6C). Overall, the percentage of insulin-positive islets was significantly low in control and β-cell-Ag–treated mice compared with zymosan and zymosan plus β-cell-Ag recipients. These results suggest that the zymosan-induced innate immune response can significantly affect the infiltration of lymphocytes into the pancreatic islets, resulting in protection of insulin-producing islets from autoimmune destruction.

Spleen and PnLN cells obtained from mice that were killed 30 days posttreatment were examined for their ability to respond to an ex vivo challenge with β-cell-Ag. Both CD4⁺ and CD8⁺ T cells from mice that received β-cell-Ag or zymosan plus β-cell-Ag showed a comparable extent of CFSE dilution upon ex vivo challenge, which was significantly higher than that of their counterparts from untreated control mice (Supplementary Fig. 8B). However, the cytokine profiles of spleen and PnLN cells from zymosan- and β-cell-Ag–recipient groups were significantly different when challenged with β-cell-Ag ex vivo (Fig. 7A). Whereas all β-cell-Ag recipient groups produced comparable levels of IFN-γ, which is significantly higher than mice that have not received β-cell-Ag, zymosan plus β-cell-Ag recipients produced significantly higher amounts of IL-10 and IL-17 as compared with mice that received β-cell-Ag alone. Further, as observed in Fig. 7B, whereas the CD4⁺IFN-γ⁺ T-cell frequencies were comparable in all groups upon nonspecific activation using PMA and ionomycin, similar to the observations of Fig. 7A, all zymosan-recipient groups had significantly higher frequencies of IL-10– and IL-4–expressing cells. Importantly, CD8⁺ T cells not only showed a similar trend but also the difference in frequencies of IL-10– and IL-4–producing cells was more pronounced among these T cells in zymosan plus β-cell-Ag–treated mice, compared with untreated controls. Moreover, zymosan β-cell-Ag–recipient mice showed a higher number of IL-17–producing CD4⁺ and CD8⁺ T cells as compared with control groups. On the other hand, higher numbers of CD8⁺IFN-γ⁺ T cells were detected in β-cell-Ag–treated, but not zymosan plus β-cell-Ag–treated, mice. These observations suggest that zymosan-exposed APCs skew the T-cell response against β-cell-Ag from IFN-γ production to IL-17, IL-10, and/or IL-4 production (from pathogenic Th1 and TC1 to less pathogenic or T1D-protective Th17 and TC17, and Th2 and TC2) in vivo.

To realize whether zymosan and/or β-cell-Ag treatment can induce and /or expand T cells with regulatory properties, spleen and PnLN cells from zymosan- and/or β-cell-Ag–treated mice were examined for Foxp3⁺ and LAP⁺ T cells by FACS. As observed in Fig. 8A, significantly higher numbers of Foxp3⁺CD4⁺ T cells were found in the PnLN of mice treated with zymosan and β-cell-Ag as compared with control mice. However, the splenic Foxp3⁺CD4⁺ T-cell frequencies in zymosan- and β-cell-Ag–treated mice were not different. Significantly higher numbers of splenic CD4⁺ T cells from mice that were treated with zymosan alone or along with β-cell-Ag were positive for surface LAP, and most of these LAP-positive cells were found to be CD25⁻ (not shown). Splenic Tregs from the mice that were killed 15 days posttreatment were adoptively transferred into 8- and 10-week-old female NOD mice to determine their potential to prevent hyperglycemia. CD4⁺CD25⁺ and LAP⁺ T cells isolated from the spleen of treated and untreated mice did not show significant differences in their ability to suppress autoimmunity, when transferred into young WT female mice (not shown), suggesting that the per cell suppressive abilities of these purified populations from treated mice are not superior to their control counterparts. On the other hand, as observed in Fig. 8B, prediabetic (8-week-old) mice that received total T cells from the spleen of zymosan plus β-cell-Ag–treated mice showed higher protection from T1D compared with mice that received T cells from control groups of mice. Ten-week-old mice that received T cells from zymosan alone or along with β-cell-Ag also developed hyperglycemia relatively slowly compared with mice that received T cells from control mice. Moreover, WT NOD mice that received total spleen cells from treated and untreated mice (Supplementary Fig. 9) and NOD-Scid mice that received purified T cells also showed similar disease incidence trend as that of WT NOD mice (not shown). These observations, in association with Fig. 7, not only suggest that T cells play a key role in zymosan plus β-cell-Ag treatment–induced protection from T1D but also indicate that the overall increase in multiple populations of T cells with regulatory phenotype, and not the superior ability of specific subsets, contributes to this disease protection.

Previously, we and others have reported that inducing innate immune response through PRRs like TLR2 and Dectin 1 using the fungal cell wall agent zymosan could suppress autoimmunity and prevent T1D in NOD mice (6,8). Here, we show that coadministration of β-cell-Ag and zymosan leads to prolonged protection of NOD mice from diabetes, and this effect appears to be dependent on a combination of zymosan-induced regulatory innate immune response involving IL-10, TGF-β1, IL-2, and Raldh1A2 and modulation of T-cell response by zymosan-exposed APCs.

It has been shown that engagement of TLR2 by its ligands induces both pro- and anti-inflammatory cytokine responses by APCs and IL-10 in T cells (18–22). However, our observations show that the TLR2 ligands Pam3Cys and Pam2Cys do not induce IL-2 and TGF-β1 in APCs independently of Dectin 1 engagement (Supplementary Fig. 10). The synergistic signaling through TLR2 and Dectin 1 (23–26), although it does not induce a significant increase in the expression of coinhibitory ligands such as PD-L1 and PD-L2, triggers the production of high levels of immune regulatory factors IL-10, TGF-β1, IL-2, and Raldh1A2 (6,7,10). Considering the involvement of these factors in immune regulation (10,27–30), we hypothesized that Ag presentation by zymosan-exposed APCs that produce these factors can skew the T-cell response from proinflammatory to anti-inflammatory/regulatory type. Our observations that Ag presentation by zymosan-exposed DCs can promote IL-10 and Foxp3 expression in T cells support this notion.

Since T1D is the result of gradual expansion of autoreactive T cells and an increase in inflammation, exposure of APCs to the immunodominant self β-cell-Ag–specific peptides in early hyperglycemic NOD mice can result in the induction and/or expansion of pathogenic T cells, leading to early hyperglycemia (14,15). Our observation that treatment with β-cell-Ag and zymosan induces protection from T1D in prediabetic NOD mice, more significantly, as compared with treatment with zymosan or β-cell-Ag alone, suggests that the zymosan-exposed APCs modulate T-cell response against β-cell-Ag in vivo, and this modulated response is protective in nature. This notion is substantiated by the observation that zymosan-exposed, β-cell-Ag–loaded DCs could induce better Treg response and prevent T1D compared with β-cell-Ag–loaded, nontreated DCs in NOD mice. Importantly, although the treatment of NOD mice with low-dose zymosan does not cause an increase in the frequencies of DCs or their subsets in the spleen and pancreatic microenvironment, the Ag-presenting ability of these cells may profoundly be altered by exposure to zymosan in vivo.

The protective nature of T cells from zymosan- and β-cell-Ag–treated mice was evident from higher frequencies of T cells with the regulatory phenotype, in PnLNs in particular, and their ability to produce IL-10 and IL-4 upon challenge/exposure to β-cell-Ag. Interestingly, T cells from mice that were treated with zymosan and β-cell-Ag also produce relatively higher amounts of IL-17. These observations indicate that zymosan-induced innate immune response contributes to the skewing of the T-cell response against β-cell-Ag from IFN-γ production (pathogenic Th1 and TC1) to IL-17, IL-10, and/or IL4 production (less pathogenic or T1D protective Th17 and TC17, and Th2 and TC2) in vivo. Although T1D is generally considered to be a Th1-mediated autoimmune disease, skewing immune response to Th17 and Th2 type results in protection of NOD mice from the disease (31,32). Recent studies, including ours, have shown that a shift in the immune response in NOD mice from Th1 to Th17 under various conditions can lead to prevention of the disease or delayed hyperglycemia (31,32). On the other hand, the potential contribution of a shift in the T-cell response toward IL-4 production when peptide vaccine, oral tolerance, and adjuvant therapy approaches are used has been well recognized for many years (33–37). Interestingly, we observed that the percentages of IL-10– and IL-4–positive T cells were higher in the spleen of both zymosan alone and zymosan plus β-cell-Ag groups, indicating that these cells may be the result of zymosan-induced innate immune response. On the other hand, an increase in the percentage of IL-17–producing cells was observed primarily in zymosan plus β-cell-Ag–treated mice, indicating a T-cell response specific to this Ag under conditions of zymosan-induced innate immune response. In addition to the difference in cytokine profiles of T cells from zymosan- and β-cell-Ag–treated and control mice, these treated mice showed higher Foxp3⁺ T-cell frequencies in the PnLN, but not spleen, compared with zymosan-treated mice. This suggests selective trafficking of Foxp3⁺ cells to the pancreatic microenvironment.

Treatment with proinflammatory agents such as bacterial LPS (TLR4 ligand) and poly I:C (TLR3 ligand) can delay hyperglycemia in NOD mice when the treatment is initiated at early stages (38–44). Although the potential mechanism is not known, a previous study has shown that the prolonged treatment of NOD mice at different stages of disease progression with LPS from Salmonella enterica results in an increase in the Treg numbers and protection from the T1D (39). Our observations show that treatment with E. coli LPS-exposed DCs can also result in a modest increase in Foxp3⁺ cells in vivo, but they fail to produce protection of the NOD mice from T1D. A role for TNF-α produced by APCs upon exposure to these TLR ligands and other microbial agents and a better immune regulation achieved through immune potentiation and activation of Tregs, especially at early stages of insulitis, have been suggested (45). In fact, previous reports have shown that TNF-α has contradictory effects at different stages of disease development in NOD mice (46,47). On the other hand, we and others have reported the unique ability of zymosan to induce the expression of IL-10, TGF-β1, IL-2, PD-L1, and/or Raldh1A2 in APCs (6–10,48) and promote protection from autoimmune diseases, even at prediabetic age and early hyperglycemic stage. Increased cell death in the PnLN, but not spleen, of zymosan-treated mice suggests an additional mechanism to eliminate inflammatory immune cells and suppress autoimmunity. Lack of significant difference in the expression levels of FAS-L and TRAIL on DCs from zymosan-treated mice suggests that cell death in the pancreatic microenvironment could be DC independent. On the other hand, our current observations that only zymosan, but not LPS, could induce immune regulatory factors such as IL-2, IL-10, and TGF-β1 in DCs, suppress the diabetogenic T cell function, and promote protection from T1D further demonstrate the regulatory nature of the zymosan-associated innate immune response.

In spite of the disease-promoting effect of zymosan in a genetically susceptible mouse model of arthritis (49), the ability of zymosan to ameliorate and/or prevent disease progression has been demonstrated in T1D and EAE models (6–11). Whereas low-dose zymosan protected mice from EAE, treatment with a high dose caused initiation of EAE through Th17 cells (11,50). Th17 cells, unlike in the T1D model where they are protective/less pathogenic (31,32), play a pathogenic role in EAE, arthritis, and many other autoimmune conditions (51–53). However, in association with previous reports, including ours (31,32), we demonstrate that skewing the T-cell response to IL-17 production instead of IFN-γ, in combination with other factors such as Tregs, IL-10, and IL-4, results in suppressed autoimmunity and delayed hyperglycemia in NOD mice.

In conclusion, our observations show that the TLR2- and Dectin 1–dependent innate immune response induced by zymosan is significantly different from that induced by a TLR4 ligand E. coli LPS. Whereas cooperative signaling though TLR2 and Dectin 1 induced by zymosan (23–26) involves both pro- and anti-inflammatory responses, signaling through TLR4 induces primarily proinflammatory responses in DCs. Although the role of zymosan-induced cell death at the site of inflammation in reduced insulitis and protection of NOD mice from T1D requires further investigation, our current study shows that the synergistic innate immune response mediated by TLR2 and Dectin 1 has the unique ability to not only produce regulatory innate immunity but also modulate T-cell response to β-cell-Ag and protect NOD mice from T1D. Importantly, our observations show that an overall increase in T cells with regulatory properties and a skewing in the immune response (cells that express Foxp3, LAP, IL-10, IL-17, and IL-4), rather than the superior suppressive ability of an individual population, is responsible for the protection of NOD mice from T1D. We conclude that zymosan has the potential to be used as a tolerogenic/immune regulatory adjuvant for promoting β-cell-Ag–specific immune modulation and reverse hyperglycemia, at least at early stages.

Acknowledgments. The authors thank Biorepository & Tissue Analysis Shared Resource, Hollings Cancer Center, MUSC, for the histology service.

Funding. This work was supported by internal funds from MUSC and UIC, National Institutes of Health grant R01-AI-073858, American Diabetes Association grant ADA-1-13-IN-57, and JDRF regular grant JDRF-32-2008-343 to C.V.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.K.-M. designed experiments and researched and analyzed data. M.H.S. and N.P. researched data. R.G. researched and analyzed data and reviewed and edited the manuscript. B.M.J. researched data and reviewed and edited the manuscript. C.V. designed experiments; analyzed data; and wrote, reviewed, and edited the manuscript. C.V. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.