Induction of Innate Immune Response through TLR2 and Dectin 1 Prevents Type 1 Diabetes¹

Autoimmune type 1 diabetes (T1D)³ is induced by environmental factors affecting the genetic susceptibility of individuals, ultimately leading to T cell-mediated destruction of insulin-secreting β cells in the pancreatic islets (1). Innate immunity plays a key role in initiating an effector T cell response as well as in maintaining tolerance to pancreatic β cell Ags. The correlation observed between the decline in infectious diseases due to improved hygiene, and increase in the incidence of T1D in developed countries suggests a role for innate immunity in preventing autoimmunity against pancreatic β cells (2, 3). Innate immune response is mediated through a number of pattern recognition receptors such as TLRs and c-type lectins that are stimulated primarily by microbial agents (4). On interaction with appropriate ligands, these receptors on APCs and other innate immune cells activate signaling pathways that lead to the production of a wide array of proinflammatory cytokines and/or up-regulation of costimulatory ligands such as CD80 and CD86 (5). Due to the ability of signaling through innate immune receptors to trigger the expression and/or up-regulation of these molecules on APCs, they have been considered important in linking innate and adaptive immune responses (6, 7).

Although the role of innate immune receptors such as TLRs in initiating T1D is not clearly understood, inflammatory effects of TLR3-, TLR2-, and TLR9-mediated signals have been implicated in virus-triggered autoimmune diabetes, sensitizing pancreatic β cells for apoptosis, and/or activating APCs for efficient self-Ag presentation (8, 9, 10). Polyinosinic-polycytidylic acid, a TLR3 ligand, and CpG, a TLR9 ligand, and viral infections in the presence of self-peptides could trigger autoimmune responses leading to T1D in mouse and rat models (8, 10, 11, 12, 13). Recently, it has been suggested that TLR2 ligation on APCs by necrotic/late apoptotic pancreatic β cells contributes to the initiation of autoimmune diabetes through the activation of Ag-captured APCs (9).

TLR2 recognizes lipoproteins/lipopeptides (LP) and peptidoglycan of microorganisms (14, 15). This receptor is unique in its ability to form heteromers with TLR1 or TLR6 to mediate intracellular signaling, and bacterial LPs signal through TLR2 in a TLR1- or TLR6-dependent manner (16, 17, 18). However, TLR2 is also known to signal in a TLR1- and TLR6-independent manner depending on the chain length of the O-acylated fatty acids as well as the assembling of the polypeptide tail (19, 20). TLR1- and TLR6-dependent and -independent signals lead to identical patterns of gene activation, indicating that heteromerization of TLR2 with TLR1 or TLR6 is evolutionarily developed to expand the ligand spectrum to enable the innate immune system to recognize different structures of LPs present in various pathogens (21).

A significant number of studies have shown that TLR2 ligation by its ligands can trigger a balanced innate immune response that includes both pro- and anti-inflammatory components (22, 23, 24, 25). Further, TLR2 agonists have shown the potential not only to induce significant amounts of suppressor cytokines but also to enhance CD4⁺CD25⁺ regulatory T cell (Treg) function (26). In contrast, recent studies have shown that TLR2 regulate the expansion and function of natural Tregs (27, 28). These reports have shown that TLR2 signaling enhanced proliferation of Tregs in vitro and in vivo resulted in a transient suppression of foxhead box p3 transcription factor (Foxp3) expression and loss of their suppressor function. However, these expanded Tregs regained suppressor function and showed the ability to prevent inflammatory condition such as spontaneous colitis (27).

Dectin-1, a c-type lectin that plays a crucial role in the recognition of fungi such as Candida and Aspergillus through β-glucans, is now known to collaborate with TLR2 in inducing a balanced pro- and anti-inflammatory innate immune response by APCs (29, 30). This indicates that dectin 1 signaling might also contribute significantly to regulatory T cell expansion and function and that approaches to engage both TLR2 and dectin 1 concurrently would have therapeutic values in preventing and treating autoimmune disorders such as T1D. Recent studies using zymosan, a fungal cell wall component, have demonstrated its ability to interact with both TLR2 and dectin 1 simultaneously and to induce large amounts of suppressor cytokines along with TNF-α (31, 32). Therefore, we explored the potential use of this agent in therapeutically modulating autoimmunity in T1D by triggering a regulatory innate immune response in diabetes-prone female NOD mice at pre- and early diabetic stages. Our results demonstrate that zymosan treatment protects NOD mice from T1D even at an early hyperglycemic stage. The disease protection was associated with a profound increase in IL-10, TGF-β1, IL-2, and TNF-α responses by APCs; enhanced natural Treg function; and changes in the effector T cell phenotype and cytokine profile.

Wild-type female NOD/LtJ and NOD-Scid mice were purchased from The Jackson Laboratory. All animal studies were approved by the animal care and use committee of the University of Illinois at Chicago. Nonfasting glucose levels in the tail vein blood samples of wild-type and NOD-Scid mice were monitored with Ascensia Contour blood glucose test strips and glucose meter (Bayer USA).

Purified Saccharomyces cerevisiae zymosan was purchased from Sigma- Aldrich and was boiled for 30 min, washed extensively, and suspended in PBS as described earlier (31). Purified anti-mouse TGF-β1 (clone A75-2), anti-CD16/CD32 (FC block) Abs; FITC-conjugated anti-mouse CD11c, CD4, and CD25 Abs and streptavidin; PE-labeled anti-mouse, F4/80, CD80, CD86, CD40, I-A^k (which cross-reacts with I-A^g7), TLR2, dectin 1, CD4, CD25, glucocorticoid-induced TNFR, programmed death-1, CTLA-4, CD8, L-selectin (CD62L), CD44, CD127, CCR7, Foxp3, and streptavidin; biotin-labeled anti-mouse/human TGF-β1 (clone A75-3) and anti-LAP (latent TGF-β1) affinity-purified Ab; PE-Cy5-labeled anti-mouse CD62L Ab and streptavidin; PE-Texas Red-labeled anti-mouse CD4 Ab; and different fluorochrome-labeled isotype control Abs (Caltag Laboratories, BD Biosciences, eBioscience, R&D Systems, Serotec, and Biolegend Laboratories) were used for FACS. Magnetic bead-conjugated total T cell, CD4⁺CD25⁺ T cell, and CD4⁺ T cell, CD11c⁺ dendritic cell (DC) isolation kits (Miltenyi Biotec) were used for enriching or depleting T cell subpopulations and DCs. PE-labeled anti-F4/80 or CD11c Ab and anti-PE Ab-linked magnetic beads were used for enriching F4/80⁺ macrophages and CD11c⁺ DCs. Multiplex reagents and paired Abs and standards for ELISA to detect TGF-β1, TNF-α, IL-10, IL-2, IL-4, IL-17, IFN-γ, IL-12, IL-1β, and IL-6 were purchased from R&D Systems, Millipore, and eBiosciences.

Short-term treatment of NOD/Ltj mice of different age groups were conducted by injecting with zymosan (i.p.; 25, 100, or 200 μg/mouse) in PBS on days 1, 3, and 5. A zymosan dose of 100 μg/mouse was found to be the adequate and optimum dose based on our preliminary observation that lower doses failed to produce a significant delay in hyperglycemia when the treatment was initiated at 8 or 12 wk of age. Further, early hyperglycemic mice failed to show a significant delay in the disease when treated intermittently even with higher doses. Therefore, two different treatment approaches were undertaken: 1) intermittent long-term treatment of 8- and 12-wk-old prediabetic mice. These mice were first injected with zymosan (100 μg/mouse) on days 1, 3, and 5 (three injections per wk); again injected three times/week on wk 3, 5, 7, 9, and 11 posttreatment initiation; and monitored for hyperglycemia for at least 30 wk; 2) continuous treatment of early hyperglycemic mice. Early hyperglycemic mice (glucose levels, 140–200 mg/ml) were injected with zymosan (100 μg/mouse/day) for 30 consecutive days and monitored for hyperglycemia for up to 27 wk after treatment initiation. Control mice were injected with PBS. Glucose levels of test and control mice were tested weekly. Mice with glucose levels >250 mg/dl for 2 consecutive wk were considered diabetic.

BM cells were cultured in complete RPMI 1640 containing 10% heat-inactivated FBS in the presence of 20 ng/ml GM-CSF for 2 days and for an additional 3 days in fresh medium containing 20 ng/ml GM-CSF and 5 ng/ml IL-4 (Invitrogen). Cells from 6-day cultures were used for this study.

Peripheral CD11c⁺ DCs and F4/80⁺ macrophages were enriched from spleen and pancreatic lymph nodes (PnLN) and pancreatic cells using PE-labeled anti-CD11c and F4/80 Abs and anti-PE Ab magnetic beads. Pancreatic tissues and islets were digested with collagenase (Roche) followed by trypsin-EDTA to generate single-cell suspension. CD11c⁺ and F4/80⁺ cell populations were enriched to >93% using magnetic separation reagents.

Total T cells were enriched to >98% using a T cell negative selection kit (Miltenyi Biotec). CD4⁺, CD4⁺CD25⁻, and CD4⁺CD25⁺ T cell subpopulations were enriched to >95% using CD25⁺ T cell isolation kit from Miltenyi Biotec according to the manufacturer’s directions. Isolated cells were washed, stained with FITC-, PE-, and allophycocyanin-labeled appropriate Abs, and tested for purity by FACS before use. Enriched CD4⁺CD25⁺ cells were also analyzed for Foxp3 expression by FACS.

Freshly isolated and ex vivo cultured cells were washed using PBS supplemented with 2% FBS and 10 mM 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 FITC-, PE-, and PECy5 or PE-TR-labeled appropriate Abs in different combinations on ice for 30 min and washed three times before analysis. For intracellular staining, surface-stained cells were fixed and permeablized using reagents from eBiosciences, incubated with fluorochrome labeled appropriate Abs, and washed three times before analysis. Stained cells were acquired using a FACSCalibur or LSR (BD Biosciences) flow cytometer, and the data were analyzed using WinMDI or Weasel application. Cells were also stained using isotype-matched control Abs and considered background controls. Specific regions were marked, and the gates and quadrants were set while analyzing the data based on these isotype control background staining. At least 10,000 cells were analyzed in each experiment.

Cell-free supernatants collected from DC, macrophage, and T cell cultures were tested for cytokine levels by ELISA or Luminex multiplex assays. Assays were conducted for detecting IL-2, IL-4, IL-10, IFN-γ, TGF-β1, IL-17, TNF-α, IL-10, IL-12, IL-1β, and IL-6. ELISA was conducted using paired Abs and standards (eBioscience and R&D Systems). Multiplex assay was conducted as per the manufacturer’s (Millipore) directions, and the data were acquired and analyzed using a Luminex-100 instrument and application from Bio-Rad. The amount of a particular cytokine was determined using an appropriate cytokine-specific standard curve.

Total T cells were enriched from spleen cells of control and zymosan-treated mice using a magnetic bead-based pan T cell isolation (negative selection) kit (Miltenyi Biotec). Enriched cells, which contained both CD4⁺ and CD8⁺ populations, were transferred to 6-wk-old NOD-Scid mice (i.v.; 2 × 10⁶ cells/mouse); glucose levels were tested every third day for up to 75 days.

All in vitro assays with mouse primary cells were done in RPMI medium supplemented with 10% FBS, l-glutamine (2 mM), HEPES (15 mM), sodium pyruvate (1 mM), 2-ME (5 × 10⁻⁵ M), penicillin (100 U/ml), streptomycin (0.l mg/ml), and Fungizone (l μg/ml). BMDCs and CD11c⁺ and F4/80⁺ cell fractions enriched from spleen were cultured with or without zymosan (10 or 50 μg/ml), for different time points, and the supernatants obtained from these cultures were used for cytokine analysis. Cells were examined for surface marker expression by FACS.

In some assays, spleen and PnLN cells (2 × 10⁵ cells/well) from zymosan-treated and control mice were stimulated with anti-CD3 Ab (2 μg/ml) for 48 h. Enriched CD4⁺CD25⁺ and CD4⁺CD25⁻ T cells from treated and untreated mice were stimulated with anti-CD3 and CD28 Abs (2 μg/ml each) with or without zymosan (50 μg/ml). Spent media from stimulated and nonstimulated wells were tested for cytokines. Cells harvested from these cultures were tested for the levels of expression of surface markers by FACS.

For coculture assay, CD4⁺CD25⁻ T cells from untreated control mice were labeled with CFSE, cultured with CD4⁺CD25⁺ T cells from zymosan-treated or control mice in the presence of anti-CD3 and APCs. Effector and Tregs were cocultured at varying ratios, but the total number of cells was maintained at 5 × 10⁵/well to avoid discrepancy in the proliferative response due to the difference in the cell density. On day 5, to test for CFSE dilution, cells from these cultures were collected, stained with PE-labeled to CD4, and analyzed by FACS.

Pancreata were fixed in 10% formaldehyde, and 5-μm paraffin sections were stained with H&E. Stained sections were analyzed in a blinded manner, using a grading system, in which 0 = no evidence of infiltration, 1 = peri-islet infiltration, 2 = <25% infiltration, 3 = 25–50% infiltration, 4 = 50–100% infiltration of each islet, and 5 = complete loss or only remnants of islets seen as described earlier (33). For each group, ∼100 islets were examined.

Mean, SD, and statistical significance (p value) were calculated using Microsoft Excel, SPSS, or GraphPad statistical application. Because the same number of test (zymosan treated or exposed) and control values (data points) were compared, a paired two-tailed t test was used unless specified. Log-rank analysis was performed to compare TID incidence (hyperglycemia) of the test group with that of the control group. Fisher’s exact test was used for comparing the total number of infiltrated islets in the test groups vs the control group. A p value of ≤0.05 was considered significant.

It has been shown that zymosan can interact with both TLR2 and dectin 1 (23). To examine the effect of engagement of TLR2 and dectin 1 on APCs from NOD mice, BMDCs, splenic DCs, and peritoneal and splenic macrophages were incubated with zymosan at different time points, and examined for surface activation markers and secreted cytokines. BMDCs and peritoneal macrophages, but not splenic DCs and macrophages, showed a significant difference in the levels of expression of costimulatory molecules on the surface upon exposure to zymosan compared with untreated control APCs (Fig. 1 A). Whereas BMDCs showed considerable up-regulation of CD80, CD86, and MHC class II (MHC II), peritoneal macrophages showed up-regulation of CD86 and CD40, but a marked down-regulation of CD80 upon in vitro exposure to zymosan. Although all macrophage and DC preparations showed considerable levels of TLR2 and dectin 1 on the surface, exposure to zymosan had no significant effect on splenic DCs and macrophages in terms of the expression levels of activation markers, CD80, CD86, CD40, and MHC II. These results suggest that zymosan can influence different populations of APCs differently and, upon treatment, might modulate T cell function in vivo.

Culture supernatants from control and zymosan-exposed APCs were examined for secreted pro- and anti-inflammatory cytokines. BM and splenic DCs and peritoneal and splenic macrophages produced profound amounts of IL-10, TGF-β1, TNF-α, and IL-2 compared with control DCs (Fig. 1 B). Although zymosan induced a significantly greater amount of IL-6 in BMDCs, the IL-6 response by peritoneal macrophages upon zymosan exposure was significantly lower than for the unexposed control. Splenic DCs and macrophages did not produce significant amounts of IL-6. Similarly, exposure to zymosan did not trigger secretion of detectable amounts of IL-12(p70), and IL-1β by DCs and macrophages (not shown). These observations show that zymosan can trigger both anti- and proinflammatory innate immune responses by APCs.

To examine the in vivo effect of TLR2 and dectin 1 engagement on APCs, zymosan-treated mice were tested for their DC and macrophage phenotypes and cytokine profiles. Fig. 2,A demonstrates that macrophages and DCs from spleen, pancreas, and PnLNs of zymosan-treated mice are not significantly different from cells of control mice in terms of the levels of expression of Ag presentation-related activation markers, TLR2, and dectin 1 on the surface. This observation was similar to that observed in vitro in Fig. 1 A. However, peritoneal macrophages from zymosan-treated mice expressed relatively lower levels of CD86, but high MHC II compared with controls, which were different from that observed upon in vitro exposure of these cells to zymosan. This difference could be due to prolonged exposure of macrophages to zymosan and various cytokines in vivo.

As shown in Fig. 2,B, CD11c⁺ DCs and F4/80⁺ macrophages from spleen, pancreas, and PnLNs produced significant amounts of IL-10, TGF-β1, IL-2, and TNF-α as observed for BM-derived, splenic, and peritoneal APCs in vitro. Similar to in vitro effects, whereas zymosan-exposed splenic APCs did not produce significant amounts of IL-6, peritoneal APCs and PnLN DCs produced significantly lower levels of IL-6 than did control cells (Fig. 2 B, bottom). In addition, these APCs did not produce significant levels of IL-1β and IL-12 upon exposure to zymosan (not shown). These results demonstrate that zymosan induces production of anti- and proinflammatory cytokines that are known to influence natural Treg function and adoptive immune response.

Considering the effect of TLR2 and dectin 1 ligation on APCs in terms of suppressor cytokine production, we examined whether treatment with zymosan can influence the T1D outcome in NOD mice. Younger (4-wk-old) mice, but not prediabetic mice, treated for a short term (on days 1, 3, and 5) with zymosan produced a significant delay in hyperglycemia. These 4-wk-old mice showed delay of hyperglycemia up to 7 wk upon short-term treatment with lower (25 μg/mouse) or higher (100 μg/mouse) doses of zymosan as compared with controls (not shown). However, 8- and 12-wk-old mice failed to show significant delay in hyperglycemia when doses lower than 100 μg/mouse were used for short-term or intermittent treatment. Therefore, prediabetic NOD mice 8 and 12 wk of age were treated intermittently with 100 μg zymosan/mouse for a prolonged period (three times on alternate weeks for 6 wk) and monitored weekly for hyperglycemia at least until 40 wk of age (30 wk after treatment initiation). As observed in Fig. 3,A, treatment of 8- and 12-wk-old mice with zymosan prevented hyperglycemia in a majority of mice for at least 30 wk after treatment initiation and for ∼20 wk after the termination of treatment. Pancreatic tissues of mice from parallel experiments were examined for insulitis 1 wk after the termination of treatment. Zymosan-treated mice showed significantly suppressed immune cell infiltration and destruction of pancreatic islets (Fig. 3, B and C). Significant numbers of islets were free of infiltration (∼50% islets), or with minimum infiltration (∼40% islet with grade 1 and 2 insulitis), in zymosan-treated mice compared with control mice (>70% islets with grade 3 and 4 insulitis).

To examine whether prolonged treatment with zymosan can suppress autoimmunity and prevent hyperglycemia in mice that show early signs of hyperglycemia, mice with glucose levels of 140–200 mg/dl were treated for 30 consecutive days with zymosan. Fig. 4 shows that whereas 100% of control mice turned hyperglycemic within 9 wk of treatment initiation, >70% of zymosan-treated early hyperglycemic mice remained nondiabetic for at least 20 wk after the treatment was terminated. Further, examination of the pancreatic tissue in nondiabetic mice 20 wk after the termination of treatment (ages between 36 and 45 wk) demonstrated that zymosan-treated mice have a significant number of infiltration-free islets even when compared with control mice that were at early hyperglycemic stages (glucose, 140–200 mg/dl; ages between 12 and 20 wk). These results indicate that lymphocyte infiltration into pancreatic islets is abrogated by innate immune response induced through TLR2 and dectin 1 by zymosan. Considering the fact that a majority of the remaining islets in the control mice had severe insulitis, it appears that anti-inflammatory cytokines induced through TLR2 and dectin 1 engagement have a role in suppressing immune cell infiltration into islets and reducing the insulitis severity.

Because zymosan treatment suppressed autoimmunity and prevented hyperglycemia in female NOD mice and the immune cell infiltration into pancreatic islets was significantly affected, we assumed that T cell function in zymosan-treated mice might also have been altered along with the APC function as observed in Figs. 1 and 2. Earlier studies have shown that splenocytes or purified splenic T cells of diabetic and prediabetic NOD mice can induce hyperglycemia in immunodeficient NOD-Scid mice upon adoptive transfer (34, 35). Therefore, diabetogenicity of purified total T cells that include both CD4⁺ and CD8⁺ populations from zymosan-treated mice was compared with that of T cells from untreated mice in immunodeficient NOD-Scid mice. Fig. 5 demonstrates that whereas T cells from control mice induced hyperglycemia in 100% of NOD-Scid mice within 55 days, T cells from zymosan-treated mice failed to induce diabetes in 80% of the mice for at least 75 days. These results indicated that T cell function is altered as a result of innate immune response induced by zymosan.

To examine the difference between T cells from control and zymosan-treated mice, PnLN and spleen cells were stimulated using anti-CD3 Ab, and spent medium was tested for various cytokines. As observed in Fig. 6, both PnLN and spleen cells secreted large amounts of IL-10 and TGF-β1 compared with cells from control mice upon activation. In contrast, IFN-γ response by T cells from zymosan-treated mice was significantly lower than those from control cells. Further, T cells from both control and zymosan-treated mice produced comparable levels of IL-4 and IL-17 (data not shown). These results indicated that T cell function in zymosan-treated mice is also modulated, perhaps, by IL-10, TGF-β1, and IL-2 secreted by APCs upon signaling through TLR2 and dectin 1.

CD62L⁺ T cells are considered less diabetogenic and have regulatory properties in NOD mice (36, 37). Therefore, we examined the frequency of CD62L⁺ T cells in zymosan-treated mice compared with untreated controls. A higher frequency of CD62L^high CD4⁺ T cells was detected in both spleen and PnLNs in zymosan-treated mice than in controls (Fig. 7,A). Whereas the majority of T cells in PnLNs of prediabetic NOD mice appeared to be CD62L^low, >50% of these cells in zymosan-treated mice were found to be CD62L^high. Further, the CD62L^high and CD62L^low T cell distribution appeared to be similar in both PnLNs and spleens of zymosan-treated mice. Similar CD62L expression pattern was observed on CD8⁺ T cells from the spleen and PnLN of zymosan-treated mice compared with control mice (not shown). T cells from zymosan-treated and control mice were also examined for other memory markers such as CD44, CD127, and CCR7. As observed in Fig. 7,B, significantly higher numbers of spleen and PnLN cells from zymosan-treated mice appeared to have a naive T cell phenotype (CD44^lowCD62L^high) than did these cells from control mice. Although the majority of CD62L^low cells from PnLNs of control mice were CD44^low, a significant number of these cells in zymosan-treated mice were CD44^high. Further analysis showed that significantly lower proportions of PnLN and spleen T cells from zymosan-treated mice were CD127⁺ and CCR7⁺ compared with T cells from control mice. These observations indicate that the T cell phenotype is altered and/or T cells with memory phenotype are replaced by T cells with naive and/or regulatory phenotype in zymosan-treated mice. This result, in conjunction with the observation from Fig. 6, suggests that altered T cell function and distribution may have contributed to reduced immune cell infiltration into pancreatic islets and delay in hyperglycemia.

Ineffective natural Treg (CD4⁺CD25⁺ T cells) function has been implicated in the progression of T1D in NOD mice. Although the Treg frequency appears to be significantly not different at different stages of the disease and in different age groups, their ability to exert a suppressive effect on effector T cells is diminished at later stages in NOD mice and in T1D patients (38, 39). Therefore, we examined whether naturally existing Tregs are affected by zymosan treatment. Although there was no selective increase in the frequency of CD62L^highCD4⁺CD25⁺ population in zymosan-treated mice, a higher proportion of CD4⁺CD25⁺ T cell population in these mice was CD62L^high compared with control mice (Fig. 8,A). In other words, increase in the CD62L^high Treg frequency in zymosan-treated mice was comparable with that of total CD4⁺ T cells (right panels of Figs. 7,A and 8A). As observed in Fig. 8 B, Foxp3⁺ Treg frequencies in spleen and PnLNs of zymosan-treated mice were not significantly different from those of untreated control mice. These results indicate that CD62L^high subpopulations may have a role in the reduced diabetogenicity of splenic T cells from zymosan-treated mice.

The ability of CD4⁺CD25⁺ T cells from zymosan-treated mice in suppressing the effector T cell (CD4⁺CD25⁻) response was tested in comparison with cells from control mice in a coculture assay. CD4⁺CD25⁺ T cells enriched from zymosan-treated and control mice were tested for Treg frequency by examining for Foxp3 expression before use in T cell suppression assays. As observed in Fig. 9,A, ∼94% of Tregs enriched from both zymosan-treated and control mice were Foxp3⁺. Although CD4⁺CD25⁺ T cells from both control and zymosan-treated NOD mice demonstrated a significant ability to suppress CD3 ligation-induced proliferation of effector T cells, the suppressive ability per cell appeared to be significantly higher in CD4⁺CD25⁺ cells from zymosan-treated mice compared with controls (Fig. 9 B). This shows that natural Treg function is enhanced by innate immune response through TLR2 and dectin 1, and it could partially explain the disease protective effect of zymosan therapy.

Because CD4⁺CD25⁺ T cells from zymosan-treated mice demonstrated enhanced ability to suppress effector T cell proliferation, levels of surface molecules and secreted cytokines of these cells were compared with that of control CD4⁺CD25⁺ T cells. The frequency of CD4⁺CD25⁺ T cells with surface-bound active and latent TGF-β1 (LAP), expression levels of glucocorticoid-induced TNFR, programmed death-1, and CTLA-4 on these cells were comparable with that of cells from control mice (not shown). However, activation using anti-CD3 and CD28 Abs revealed that CD4⁺CD25⁺ T cells from zymosan-treated mice can produce greater amounts of TGF-β1 and IL-10 than control CD4⁺CD25⁺ T cells do (Fig. 10). Alternatively, CD4⁺CD25⁻ T cells from zymosan-treated mice produced significantly higher amounts of IL-10, but not TGF-β1, compared with corresponding cells from control mice. These results show that IL-10 and TGF-β1 could be major mediators of enhanced suppressor function of CD4⁺CD25⁺ T cells from zymosan-treated mice.

With the exception of a recent study that suggested a role for TLR2 engagement by dying pancreatic β cells in initiating T1D (9), the effect of TLR2 and dectin 1 signaling in T1D is largely unknown. Therefore, to understand the effect of innate immune response induced through TLR2 and dectin 1 in modulating T1D, prediabetic and early hyperglycemic NOD mice were treated with zymosan. We found that induction of innate immune response using zymosan in these mice leads to their protection from T1D. This disease-protective effect appears to be contributed by copious amounts of anti-inflammatory cytokines produced by APCs and enhanced function of the regulatory wing of the immune system.

Consistent with earlier reports (31, 32), exposure of DCs and macrophages to zymosan resulted in the secretion of significant amounts of IL-10, TNF-α, and IL-2. Zymosan also triggered the production of large amounts of TGF-β1 by DCs and macrophages in vivo and in vitro, suggesting that treatment with this agent will produce suppressive effects on autoimmunity in T1D. Unlike signaling through inflammatory TLRs such as TLR3, TLR4, and TLR9 that induces profound amounts of IL-6, IL-1β, and/or IL-12 (40, 41), zymosan did not trigger these inflammatory cytokines by peripheral APCs such as macrophages and DCs, indicating that signaling through TLR2 and dectin 1 produces a regulatory innate immune response. TNF-α has contradictory effects in NOD mice depending on the insulitis stage (42). It is destructive at early stages but protective at later stages of insulitis. TNF-α also indirectly promotes natural Treg function through aiding effective APC function (43, 44). This cytokine, in the presence of large amounts of IL-10, TGF-β1, and IL-2, may produce a protective effect at all stages of T1D. IL-10 and TGF-β1 have proven their protective effects in T1D and other inflammatory diseases by promoting Treg function and by suppressing pathogenic T cell response (45, 46, 47). IL-2, on the other hand, is essential for maintaining peripheral tolerance through promoting the survival and propagation of natural Tregs (48, 49). Therefore, we assumed that induction of regulatory innate immune response through TLR2 and dectin 1 using zymosan will suppress the ongoing autoimmunity in NOD mice and protect them from T1D.

A recent report indicated that TLR2 signaling by necrotic pancreatic β cells contributes to the activation of APCs and initiation of the β cell-specific autoimmune response in NOD mice (9). However, in contrast to inflammatory responses induced by ligands for other TLRs, stimulation of TLR2 induced large amounts of proinflammatory cytokine, TNF-α as well as anti-inflammatory cytokines (30, 31, 32). Recent studies have also shown that TLR2 engagement enhances natural Treg function (26) and helps to expand Treg numbers (27, 28). These reports, and our observation that zymosan can trigger significant amounts of TGF-β1, IL-10, and IL-2 secretion by APCs from NOD mice prompted us to examine the potential of zymosan, which is known to bind to TLR2 and the synergistic receptor, dectin 1, in modulating T1D. TLR2 and dectin 1 receptors are also known to collaborate in activating NF-κB and NFAT, leading to the production of IL-2 and IL-10 (32, 50) and indicating that profound disease modulation could be achieved using zymosan.

Treatment of prediabetic and early hyperglycemic NOD mice with zymosan clearly demonstrated that the innate immune response induced using zymosan can result in profound protection of these mice from diabetes. Studies have shown that insulitis can be detected in NOD mice as early as 3–4 wk of age, and therapeutic approaches have been widely effective at this stage. Short-term treatment of NOD mice at 4 wk of age with zymosan was sufficient to delay the disease for ∼7 wk. However, because an effective treatment at, or close to, hyperglycemic stage is one of the major challenges in treating T1D, the efficacy of prolonged treatment in preventing the disease was tested. Our observations show that disease in older age groups (8 and 12 wk) and in early hyperglycemic mice (glucose, 140–200 mg/dl) could be prevented for a significantly long time by intermittent and/or prolonged treatment using zymosan, suggesting that innate immune response through receptors such as TLR2 and dectin 1 can have a protective effect on pre- and early hyperglycemic stages of the disease.

T1D can be induced in immunodeficient NOD-Scid mice by adoptively transferring T cells from diabetic and prediabetic NOD mice (34, 35, 51). We assumed that an immunotherapy that suppresses the disease outcome could negatively affect the pathogenicity of peripheral T cells, and we therefore tested the diabetogenic ability of T cells from zymosan-treated mice in comparison with control T cells using NOD-Scid mice. Our results show that T cells from zymosan-treated mice failed to induce hyperglycemia in the majority of NOD-Scid mice, indicating that innate immune response induced through TLR2 and dectin 1 has a profound suppressive effect on the T cell function.

Lowered ability of T cells from zymosan-treated mice to induce diabetes is associated with a reduction in the frequency of CD62L^low T cells in lymphoid organs including PnLNs. In NOD mice, diabetogenicity of CD62L^low cells is known to be higher than that of CD62L^high T cells (36, 37). In addition, CD62L^highCD4⁺ T cells have been found to have regulatory properties, and the CD4⁺CD25⁺CD62L^high population has been considered a more efficient Treg population than CD4⁺CD25⁺CD62L^low cells (36, 52, 53). A significantly higher number of PnLNs T cells are CD62L^low than those of spleen and other peripheral lymphoid organs at any given stage and are indicative of the inflammatory nature of pancreatic microenvironment in NOD mice (33). However, unlike controls, zymosan-treated mice had similar CD62L^high and CD62L^low T cells ratios in both spleen and PnLNs.

Memory T cells are highly heterogeneous in terms of expression of surface markers including chemokine and cytokine receptors. Mouse memory T cell populations with varying levels of expression of CD62L, CD44, CD127, CD122, and CCR7 have been reported (54, 55, 56, 57). Reduced frequency of CD62L^low cells in zymosan-treated mice compared with controls indicated that memory T cells may be affected by zymosan treatment. Our observation that relatively higher proportions of T cells from the PnLNs and spleens of zymosan-treated mice have naive T cell phenotype (CD62L^highCD44^lowCCR7⁻) compared with controls supported this notion. In addition, although a significant portion of CD62L^low T cells from the PnLNs of control mice expressed CD127, CCR7, and low levels of CD44, the majority of cells from zymosan-treated mice expressed CD127 and high levels of CD44, indicating that innate immune response induced upon TLR2 and dectin 1 engagement has a significant effect on T cell phenotype. Although further studies are needed to understand the functional significance of these phenotypic alterations, our observations suggest that innate immune response induced through TLR2 and dectin 1 affect the pathogenicity of T cells, and zymosan treatment leads to repopulation of nonpathogenic T cells in the pancreatic microenvironment. Our observations that PnLN and spleen T cells from zymosan-treated mice produce higher amounts of IL-10 and TGF-β1 but significantly lower levels of IFN-γ upon activation further indicate that the T cell function is, in fact, altered in zymosan-treated mice.

TGF-β1 and IL-2 play a critical role in the maintenance of effective natural Treg function (45, 48, 49). In contrast, IL-10 and TGF-β1 can negatively affect the production of proinflammatory cytokines (such as IL-1β and IL-6 by APCs) that have an abrogative effect on Foxp3⁺ natural Tregs (58, 59, 60). Although zymosan treatment did not result in a significant increase in the natural Foxp3⁺ Treg numbers in vivo, Tregs from treated mice secreted profoundly higher amounts of TGF-β1 and IL-10 upon CD3 ligation and demonstrated an enhanced ability to suppress effector T cells compared with Tregs from untreated mice. These observations stress not only on a positive role of innate immune response through receptors such as TLR2 and dectin 1 on natural Treg function, but also on the importance of innate immunity through these receptors in maintaining peripheral tolerance.

It is now known that mutually dependent steady state functions of APCs and Tregs are important to maintain peripheral T cell tolerance (61). Two major defects in the immune system that have widely been reported in T1D are Ag-presenting function and Treg properties (62, 63). Our observation that DCs and macrophages from zymosan-treated mice produce significant amounts of IL2, IL-10, and TGF-β1 indicate that these factors could be responsible, at least in part, for the alteration of T cell phenotype and function.

TLRs and other innate immune receptors are widely expressed on APCs, and cytokines produced by them upon ligation of innate immune receptors bridge innate and adaptive arms of the immune system (7). However, recent reports indicated that TLRs are also expressed on various other immune cell populations including T and B cells, and they might have a direct role in modulating adaptive immune responses upon Ag-specific activation (64, 65). Therefore, it is likely that T cell function in zymosan-treated mice is affected, in part, also by direct ligation of TLR2 and/or dectin 1.

A number of studies have demonstrated the ability of microbial agents in preventing T1D in NOD mice (66, 67, 68, 69). Injection of mycobacteria or CFA could prevent T1D for a long duration in NOD mice (67, 68, 70). Although the mechanism is not known, a role of TNF-α produced upon microbial triggering of innate immune response in the disease prevention has been suggested (70). Recent studies have demonstrated that dectin 1 functions together with TLR2 to mediate macrophage activation by mycobacteria (71, 72). Our observations suggest that these synergistically acting receptors, TLR2 and dectin 1, have the unique ability to induce cytokines such as IL-10, TGF-β1, and IL-2 by APCs upon ligation by zymosan and that they prevent T1D in NOD mice. These observations also indicate that protection from T1D achieved by others (67, 68, 70) through treating with mycobacteria or CFA could involve innate immune response induced through receptors such as TLR2 and dectin 1.

Based on the observation that zymosan can trigger the secretion of large amounts of suppressor cytokines by APCs, it can be assumed that adaptive self-Ag-specific T cell response could be skewed toward protective regulatory type. Further, considering the fact that innate immune receptors are expressed not only on APCs but also on T cells (64, 65), direct effects of ligation of these receptors in the alteration of T cell phenotype and function cannot be ruled out. Therefore, additional studies are needed to understand the effect of ligation of these receptors on APCs and T cells in influencing adaptive self-Ag-specific T cell response. Studies are also needed to understand the effect of innate immune response triggered through receptors such as TLR2 and dectin 1 on target tissue and insulitis. Nevertheless, our study clearly demonstrates the protective role of innate immune response through TLR2 and dectin 1 by zymosan in T1D, and agonistic compounds for these receptors could prove effective in preventing/treating T1D in humans. These observations also suggest that environmental factor induced prevention of T1D in NOD mice may be dependent on innate immune response induced through receptors such as TLR2 and dectin 1.

The authors have no financial conflict of interest.