ROR Inverse Agonist Suppresses Insulitis and Prevents Hyperglycemia in a Mouse Model of Type 1 Diabetes

Abstract

Hyperglycemia associated with type 1 diabetes is a consequence of immune-mediated destruction of insulin producing pancreatic β-cells. Although it is apparent that both CD8⁺ T cells and T_H1 cells are key contributors to type 1 diabetes, the function of T_H17 cells in disease onset and progression remains unclear. The nuclear receptors retinoic acid receptor-related orphan receptors-α and γt (RORα and RORγt) play critical roles in the development of T_H17 cells and ROR-specific synthetic ligands have proven efficacy in several mouse models of autoimmunity. To investigate the roles and therapeutic potential for targeting the RORs in type 1 diabetes, we administered SR1001, a selective RORα/γ inverse agonist, to nonobese diabetic mice. SR1001 significantly reduced diabetes incidence and insulitis in the treated mice. Furthermore, SR1001 reduced proinflammatory cytokine expression, particularly T_H17-mediated cytokines, reduced autoantibody production, and increased the frequency of CD4⁺Foxp3⁺ T regulatory cells. These data suggest that T_H17 cells may have a pathological role in the development of type 1 diabetes, and use of ROR-specific synthetic ligands targeting this cell type may prove utility as a novel treatment for type 1 diabetes.

Type 1 diabetes is a chronic autoimmune disease precipitating in genetically susceptible individuals in collaboration with unknown environmental factors (1). The body's immune system selectively destroys the insulin-producing pancreatic-β cells, resulting in insulin deficiency and hyperglycemia. Type 1 diabetes is treated with insulin replacement therapy and is required for the remainder of the patient's life. Treatment options for type 1 diabetes are limited, focusing mainly on controlling blood glucose with insulin therapy, which has little effect on the autoimmune process. Therefore, identifying factors that can modulate the autoimmune destruction may provide new approaches for the treatment of type 1 diabetes.

T cells play a significant role in the development of type 1 diabetes with cytotoxic CD8⁺ T cells and CD4⁺ T_H1 cells considered key mediators of pathogenesis in both rodent models and human patients (2). However, the discovery that T_H17 cells are pathological mediators of several autoimmune diseases has led many to investigate their role in type 1 diabetes. Evidence for the pathogenicity of T_H17 cells in type 1 diabetes originates from studies in which nonobese diabetic (NOD) mice were treated with neutralizing IL-17 antibodies or IL-25, both of which antagonized T_H17 differentiation in vivo and prevented the development of disease (3). Moreover, studies of type 1 diabetes patient samples showed elevated levels of IL-17-producing CD4⁺ T cells in the peripheral blood and pancreatic lymph nodes as well as increased populations of peripheral blood monocytes that could promote T_H17 cell differentiation (4–7). In contrast, several studies have demonstrated that induction of T_H17 cells and/or IL-17 expression is protective in mouse models of type 1 diabetes (8–10). Adding to this complicated issue is the recent evidence delineating the inherent plasticity of T_H17 cells. These studies have demonstrated that T_H17 cells can convert into interferon (IFN)-γ-producing T_H1-like cells, considered the most pathogenic (11, 12). Thus, the role for T_H17 cells in the pathogenesis of type 1 diabetes remains controversial.

Nuclear receptors (NRs) are ligand-regulated transcription factors, and numerous therapeutics used clinically have been developed targeting several members of the NR superfamily. The retinoic acid receptor-related orphan receptors (RORs)-α and -γt [RORα (NR1F1) and RORγ (NR1F3)] are members of the NR superfamily with critical roles in several metabolic processes, including glucose and lipid metabolism, and the development and function of T_H17 cells (13). A significant body of work has focused on the roles of the RORs in immune function, and elegant genetic studies have established that the combined deletion of both RORα and RORγ completely abolishes T_H17 cell development, suggesting a synergism between the two transcription factors in the generation of this cell type (14). T_H17 cells preferentially secrete IL-17A, IL-17F, IL-21, and IL-22, all of which are important during tissue inflammation and play a role in antimicrobial immunity at epithelial/mucosal barriers (15). Interestingly, polymorphic variants of the common γ-chain cytokine IL-21 and its receptor have been associated with susceptibility to type 1 diabetes (16). Several studies have established that deletion of IL-21 or the IL-21 receptor protects mice from developing type 1 diabetes, suggesting that inhibition of IL-21 expression or signaling may be of benefit for type 1 diabetes treatment (17, 18). These data suggest that the inhibition of cytokines secreted by T_H17 cells, such as IL-21, may be an effective therapeutic option.

We have identified several high-affinity synthetic ligands specific for the RORs and demonstrated their roles in the regulation of immunity. Two RORγ-specific ligands, SR1555 and SR2211, inhibited T_H17 cell development and function (19, 20). Interestingly, SR1555 also enhanced the number of in vitro-derived Foxp3⁺ T regulatory cells (19). Finally, we also identified a high-affinity synthetic ligand, SR1001, a dual-RORα/γ modulator, that inhibits T_H17 cell differentiation and function (21). SR1001 is effective at delaying the onset and severity of symptoms in a mouse model of multiple sclerosis (MS), experimental autoimmune encephalomyelitis (21). Although several groups have assessed the immune-regulating effects of anti-IL-17 treatment, silencing of IL-17, or adjuvant treatment to indirectly assess the roles of T_H17 cells in the pathogenesis of type 1 diabetes, no study has investigated the direct consequences of modulating the transcription factors that control T_H17 development and function, the RORs (3, 9). Therefore, the unclear role of T_H17 cells in type 1 diabetes and the genetic evidence suggesting that both RORα and RORγt are required for full T_H17 cell development prompted us to investigate whether suppression of ROR activity affected T_H17 cell function and pathogenesis of type 1 diabetes and determine whether selective targeting of the RORs is a viable therapeutic option.

Materials and Methods

Mice and treatments

Female NOD/LtJ mice were obtained from Jackson Labs and housed in specific pathogen-free conditions in accordance with protocols approved by the Institutional Animal Care and Use Committee of the Scripps Research Institute (Jupiter, Florida). NOD mice were injected ip twice a day [2.5 mg/mL, equaling 25 mg/kg (1 μL/g body weight of mouse)] with vehicle or SR1001, which was dissolved in a 10% dimethylsulfoxide and 10% Tween 80 solution. Mice were either 5 or 10 weeks of age at the initiation of all experiments. Diabetes monitoring was performed once per week in nonfasted mice, at approximately 4:00 pm, with a One Touch Ultra glucometer (LifeScan), with animals considered diabetic after two consecutive blood glucose readings of greater than 18 mmol/L or one reading of 25 mmol/L.

Histology and immunofluorescence

Pancreata were removed and fixed in 4% paraformaldehyde overnight at 4ºC and transferred to 30% sucrose buffer. The tissue was subsequently frozen in O.C.T. (optimum cutting temperature) (Tissue-Tek), and 10-μM-thick sections were cut 150 μm apart to prevent double counting of the same islets and placed onto Fisherbrand Superfrost Plus slides using a Leica CM1950 cryostat (Leica Biosystems). Hematoxylin and eosin staining was performed using air-dried and rehydrated slides. After dehydration by xylene washes, all slides were coverslipped using DPX (Fisher Scientific) as a mounting medium. All stained sections were viewed and photographed using an Olympus microscope.

Islet counts and insulitis scoring

Islet counts were performed using methods outlined by Wang et al (22). Islet counts included the number of visible islets per slide counted (five mice per group, eight slides per mouse) using sections spanning the pancreas. For insulitis scoring, six sections per pancreas that had been stained with hematoxylin and eosin were analyzed by light microscopy. Slides were assigned a random number corresponding to an animal Identification to perform a blind analysis. Insulitis scoring was performed according to the following criteria: severe insulitis, 50% or greater of the islet area is infiltrated; mild insulitis, less than 50% of the islet area is infiltrated; periinsulitis, infiltration is restricted to the periphery of the islet; no insulitis, absence of cell infiltration. For assessment of insulin, slides were air dried and rehydrated in PBS supplemented with 0.05% Tween 20.

To prevent nonspecific binding of primary antibodies, sections were blocked in a solution of 5% normal serum, 1% BSA, and 0.05% Tween 20 for 1 hour at room temperature. Slides were rinsed in PBS (three washes of 10 min each) and incubated in primary antibody in PBS, 1% BSA, and 0.05% Tween 20 overnight at 4ºC. Guinea pig antiinsulin (Abcam) was used at a 1:1000 dilution. After an incubation in primary antibody, slides were rinsed in PBS (three washes of 10 min each) and incubated in with goat anti-guinea pig Alexafluor 647 (Invitrogen) secondary antibody (diluted 1:200). Slides were rinsed in PBS three times and coverslipped using Vectashield with 4′,6′-diamino-2-phenylindole (Vector Labs) as a mounting medium.

GAD65 (Glutamic acid decarboxylase, isoform 65) and insulin autoantibody assays

At the termination of the experiments, blood was collected via heart puncture in heparin-coated microtubes and centrifuged for 10 minutes at 4°C, 2000 rpm, to collect plasma from both vehicle- and drug-treated animals in the 1-month studies. GAD65 (IgG) and insulin (IgG) autoantibody levels were determined using detection kits from Alpha Diagnostic International according to the manufacturer's protocols. Briefly, plasma samples, run in duplicate, were diluted 1:100 (GAD65) or 1:50 (insulin) in working sample diluent and incubated for 1 hour at room temperature (RT) in GAD65 or insulin-coated 96-well plates. Plates were washed and incubated for 30 minutes at RT with antimouse IgG-horseradish peroxidase, followed by a substrate incubation (15 min, RT). Reactions were stopped (stop solution) and analyzed at a 450-nm wavelength using a Biotek Synergy 2 plate reader. Concentrations were calculated based on a calibrator curve.

Cytokine assays

Plasma was isolated from animals following the procedures outlined above. Plasma cytokines were run in duplicate and measured using a multiplexed kit (Mouse Th17 magnetic bead panel; Millipore) and a Luminex 200 xPONENT 3.1 instrument (Luminex Corp) following the manufacturer's protocol.

Quantitative PCR

Splenocyte total RNA was extracted using RNeasy Plus micro kits (QIAGEN) and reverse transcribed using iScript cDNA biosynthesis kits (Bio-Rad Laboratories) after being stimulated for 5 hours with plate bound anti-CD3 (2 μg/mL; eBioscience). Total RNA from the pancreas of NOD mice was isolated using the TRIzol reagent (Invitrogen) followed by deoxyribonuclease 1 treatment (Invitrogen) and reverse transcribed using an iScript cDNA biosynthesis kit. Quantitative RT-PCR was performed with a 7900HT Fast real-time PCR system (Applied Biosystems) using SYBR Green (Roche) as previously described. Primer sequences can be found in Table 1 (21).

Table 1.

Primer Sequences

Gene	Forward (5-′–3-′)	Reverse (5-′–3-′)
Ins1	CCTCTGGGAGCCCAAACCCA	TCCACTTCACGGCGGGACTT
Ins2	CATGTCCCGCCGTGAAGTGG	AATGCCACGCTTCTGCTGGG
Il17a	CTCCAGAAGGCCCTCAGACTAC	AGCTTTCCCTCCGCATTGACACAG
Il17f	GAGGATAACACTGTGAGAGTTGAC	GAGTTCATGGTGCTGTCTTCC
Il21	CAGATCGCCTCCTGATTAGACT	CTCACAGTGCCCCTTACATC
il22	TCATCGGGGAGAAACTGTTC	CATGTAGGGCTGGAACCTGT
il23r	TGTGGATCCTGTCCTTACAGAG	CCTACAGGACAGTCTCTTGTCTCA
Ifng	TGGCTGTTTCTGGCTGTTACT	GCTCTGCAGGATTTTCATGTC
Prf1	CCCTGCACACATTACTGGAA	GCACTCACACTGGCATGAAT
Foxp3	CATAGCTCCCAGCTTCTCCTT	GAGCCAGAAGAGTTTCTCAAGC
Fas	ATCGCCTATGGTTGTTGACC	TGAGGCATTCATTGGTATGG
Tnfa	TCAGCCGATTTGCTATCTCAT	TGGAAGACTCCTCCCAGGTAT
IL12p40	GGAAGCACGGCAGCAGAATA	AACTTGAGGGAGAAGTAGGAATGG
Il6	CATGTTCTCTGGGAAATCGTG	TCCAGTTTGGTAGCATCCATC

T-cell differentiation in vitro

The conditions for the different T_H cell subsets were as follows: 20 μg/mL anti-IL-4, 20 ng/mL IL-12 (R&D Systems), and 10 ng/mL IL-2 (R&D Systems) for T_H1 conditions; 20 μg/mL anti-IFNγ, 20 μg/mL anti-IL-4, 1 ng/mL TGFβ, and 10 ng/mL IL-6 (R&D Systems) for T_H17 conditions. All T cells were stimulated with 2 μg/mL anti-CD3 (eBiosciences) and 10 μg/mL anti-CD28 (eBiosciences). For naïve T-cell differentiation, CD4⁺ T cells were isolated from 6- to 8-week-old NOD mice using Dynal beads (Invitrogen) according to the manufacturer's instructions. CD4⁺CD25⁻CD62L^hiCD44^lo cells were sorted with a fluorescence-activated cell sorter (FACS) on a BD FACSAriaII. Four days after activation, all cells were restimulated with 5 ng/mL phorbol-12-myristate-13-acetate (Sigma) and 500 ng/mL ionomycin (Sigma) for 2 hours followed by the addition of GolgiStop (BD Bioscience) for an additional 2 hours before intracellular staining. Cells were cultured in RPMI 1640 medium (Invitrogen) with 10% fetal bovine serum and antibiotics.

Flow cytometry

Cell suspensions were prepared from various tissue and were stained at 4°C in PBS containing 2% BSA and 0.5% EDTA after FcγRII/III blocking. All surface and intracellular staining was performed with antibodies from BD Pharmingen or eBioscience [fluorescein isothiocyanate anti-CD4 (GK1.5), Alexafluor 647 anti-Foxp3 (FJK-16s); phycoerythrin-conjugated antimouse IL-17A (eBio17B7), Alexafluor 647 antimouse IFNγ (XMG1.2), and fluorescein isothiocyanate anti-CD8 (clone 53–6.7). For intracellular staining, cells were stimulated with 5ng/mL phorbol-12-myristate-13-acetate (Sigma) and 500 ng/mL ionomycin (Sigma) for 2 hours followed by the addition of GolgiStop (BD Bioscience) for an additional 2 hours prior to fixation and permeabilization using the Foxp3 staining buffer set (eBioscience). Flow cytometric analysis was performed on a BD LSRII (BD Biosciences) instrument and analyzed using FlowJo software (TreeStar).

Statistical analysis

Data are expressed as mean ± SEM (n = ≥ 5). A Kaplan-Meier analysis was used for survival curves. Unpaired, two-tailed Student's t test were used to analyze data between two groups. All data was analyzed using GraphPad Prism 6.0 software. Significance was established as follows: *, P < .05; **, P < .01; ***, P < .001.

Results

SR1001 treatment prevents the development of autoimmune diabetes

Most mechanistic data and potential intervention therapies for type 1 diabetes have been derived from the NOD mouse, a commonly used mouse model of spontaneous disease onset with many similarities to human type 1 diabetes (23). To determine whether inhibition of RORα and RORγ activity, and consequently inhibition of T_H17 cells, would affect the natural development of diabetes, we administered SR1001 (25 mg/kg⁻¹, twice a day), a RORα/RORγ-specific synthetic ligand, to 5- and 10-week-old female NOD mice and followed up the incidence of diabetes, referred to as the long-term studies (21). Dosage was based on SR1001's selectivity, pharmacokinetic profile, and dose response determined in a mouse model of MS (21). The ages of the mice were chosen based on previous histological analysis demonstrating that immune infiltrates begin to surround pancreatic islets at approximately 5 weeks of age (periinsulitis, initiation phase), whereas by approximately 10 weeks of age, immune cells have begun their invasion of the islets (insulitis, effector phase) (23). After 16 weeks of treatment, when the mice were 22–24 weeks of age, none of the SR1001-treated animals had developed diabetes compared with greater than 70% of the vehicle control group (Figure 1, A and D). No changes in behavior or body weight were observed over the course of the experiment, indicating the drugs were not toxic (Supplemental Figure 1). Although one mouse in the vehicle-treated group that was 5 weeks of age at the start of the experiment developed diabetes very early (∼9 wk of age), the remainder of the vehicle-treated mice developed diabetes at rates and ages typical of female mice acquired from Jackson Labs, suggesting diabetes was progressing at the normal rate (23).

Figure 1.

SR1001 inhibits the onset of diabetes. Graph representing diabetes incidence (A), representative hematoxylin and eosin-stained islets (B), and total islet count (C) from female NOD mice treated with SR1001 or vehicle control starting at 5 weeks of age. Data from panels A–C were from mice that were 22 weeks of age at the time of analysis. Graph representing diabetes incidence (D), representative hematoxylin and eosin-stained islets (E), and total islet count (F) from female NOD mice treated with SR1001 or vehicle control, starting at 10 weeks of age. Data from panels D–F were from mice that were 24 weeks of age at the time of analysis. Bold lines in graphs represent the SR1001-treated group, and dotted lines represent the vehicle-treated group. Images are presented at ×20 magnification. Number of islets counted from hematoxylin and eosin-stained sections are represented in panels C and F. Data are mean values ± SEM (n = 10/group for A, B, D, and E; n = 5/group for C and F). *, P < .05; **, P < .01.

To assess the extent of immune infiltration in the pancreatic islets, we performed a histological analysis on pancreas sections from vehicle- and SR1001-treated animals. Not surprisingly, there were very few visible pancreatic islets remaining in the diabetic vehicle control mice from both groups of animals (Figure 1, B and E). In contrast, the islets of the SR1001-treated animals had reduced immune infiltrates present (Figure 1, B and E). When counted, we observed that there were significantly more visible islets remaining in the SR1001-treated groups relative to vehicle control (Figure 1, C and F).

SR1001 affects insulin expression in NOD mice

Because SR1001 affected the development of autoimmune diabetes, we next wanted to determine whether SR100 treatment helped preserve insulin expression. We performed immunostaining on pancreas sections from vehicle- and SR1001-treated animals from Figure 1 and observed very few to no insulin-positive cells remaining in the vehicle control groups (Figure 2, A and C). Conversely, there was a significant amount of insulin-positive cells remaining in the SR1001-treated animals despite observing immune infiltration. An RT-PCR analysis of pancreatic insulin gene expression corroborated our immunostaining results, demonstrating that the SR1001-treated animals had a significant amount of insulin mRNA expression remaining at the termination of the experiment relative to the vehicle controls (Figure 2, B and D).

Figure 2.

SR1001 preserves insulin expression. A, Representative immunofluorescence staining of islets from SR1001- or vehicle-treated mice (starting at 5 wk of age) showing insulin content (red) and reduced immune infiltration. B, mRNA expression of insulin (Ins) genes in the pancreas of the same mice depicted in panel A. Data from panels A and B were from mice that were 22 weeks of age at the time of analysis. C, Representative immunofluorescence staining of islets from SR1001- or vehicle-treated mice (starting at 10 wk of age) showing insulin content (red) and reduced immune infiltration. D, mRNA expression of insulin (Ins) genes in the pancreas of the same mice depicted in panel C. Data from panels C and D were from mice that were 24 weeks of age at the time of analysis. Images are presented at ×20 magnification. Data are mean ± SEM values (n = 5/group). *, P < .05. DAPI, 4′,6′-diamino-2-phenylindole.

Reduction in pancreatic immune infiltration with SR1001 treatment

To better understand the roles for the RORs in the early events leading to diabetes onset, we administered vehicle or SR1001 to 5- and 10-week-old female NOD mice for 30 days (25 mg/kg⁻¹, twice a day), referred to throughout as short term. Weights and blood glucose levels of all mice remained relatively stable, suggesting the drug and treatment regimen was not toxic. Only one mouse in the 10-week-old, vehicle-treated group developed diabetes (Supplemental Figure 2, A and B). Histological analysis of the pancreatic islets indicated that treatment with SR1001 significantly reduced the degree of mononuclear cell infiltration in the short-term group of mice that were 10 weeks old at the start of the study (Figure 3A). A similar analysis was performed on the pancreas from the short-term, 5-week-old treated mice, and there appeared to be very little difference between the SR1001- and vehicle-treated groups (Supplemental Figure 3, A and B). Immunostaining of pancreas sections from the short-term, 10-week-old animals showed that SR1001 treatment led to more insulin-positive cells than vehicle alone (Figure 3B). RT-PCR analysis of pancreatic insulin gene expression supported our immunostaining results and suggested that inhibition of ROR activity slows pancreatic immune infiltration and helps maintain insulin expression in NOD mice (Figure 3C and Supplemental Figure 3C).

Figure 3.

Short-term SR1001 treatment delays immune infiltration and preserves insulin expression. A, Representative hematoxylin and eosin-stained and scored islets from SR1001- or vehicle-treated female NOD mice. Ten-week-old animals were treated for 30 days. An analysis was performed on mice that were 14 weeks of age at the termination of the experiment. B, Representative immunofluorescence staining of islets from SR1001- or vehicle-treated mice (depicted in panel A) showing insulin content (red) and reduced immune infiltration. C, mRNA expression of insulin (Ins) genes in the pancreas of the same mice depicted in panel A. Images are presented at ×20 magnification. Data are mean ± SEM values (n = 5/group). *, P < .05.

Inhibition of ROR prevents autoantibody formation

Due to the decreased immune infiltration observed in the islets from the SR1001-treated animals, we wanted to determine the immune-modulating effects of SR1001 in these mice. Anti-islet autoantibodies, produced by autoreactive B cells, are important markers of immune cell activation against β-cells and are detected before the clinical onset of type 1 diabetes. Although the precise role of B cells in type 1 diabetes onset is still debated, autoantibodies reflect a prelude to autoimmunity and insulin is considered a primary autoantigen (1). Whereas autoantibody expression does not always correlate with diabetes onset, the more autoantibodies present, the greater the probability of developing type 1 diabetes (1). Therefore, to determine whether SR1001 was affecting the development of autoantibodies from autoreactive B cells in NOD mice, plasma samples from both the short-term treated 5- and 10-week-old mice were analyzed for anti-GAD65 and anti-insulin autoantibodies. Samples analyzed from the 5-week-old treated mice showed no difference in anti-GAD65 or anti-insulin autoantibody levels between groups, suggesting that a pronounced immune response had yet to occur (Figure 4). However, there was a significant difference in autoantibody levels between the SR1001- and vehicle-treated samples from the 10-week-old treated mice (Figure 4). These data suggest that inhibition of RORα and RORγ can inhibit autoantibody formation during the initiation period of diabetes.

Figure 4.

SR1001 inhibits autoantibody formation. An ELISA analysis of insulin autoantibody formation (IgG) (A) or GAD65 autoantibody formation (IgG) (B) from serum in 30-day SR1001- or vehicle-treated female NOD mice. Mice were either 5 weeks or 10 weeks of age at the start of the experiment and 9 or 14 weeks of age at the termination of the experiment, respectively. Data are mean ± SEM values (n = 10/group). *, P < .05; **, P < .01.

SR1001 inhibits T_H17 cells in NOD mice

Inhibition of T_H17 cells in other mouse models of autoimmunity has resulted in a measurable reduction in autoantibody formation (24, 25). In light of this fact, we wanted to determine the effects of SR1001 on T_H17 cells in NOD mice. We isolated splenocytes from the short-term treated 5- and 10-week-old animals and analyzed the mRNA expression of T_H17-secreted cytokines, specifically Il17a, Il17f, Il21, Il22, as the IL-23 receptor (Il23r). Although there was some inhibition of T_H17-mediated cytokines in the mice that were 5 weeks old at the start of the experiment, there was significantly greater inhibition in the 10-week-old treated group (Figure 5A). Intracellular cytokine staining assessing IL-17A expression supported our mRNA data (Figure 5B and Supplemental Figure 4). Interestingly, Il21 expression was suppressed in both drug-treated sample groups, suggesting that the RORs may be central to facilitating IL-21 expression in NOD mice (16–18). Additionally, the expression of IFNγ (Ifng) and perforin (Prf-1), a cytolytic protein found in the granules of cytotoxic T lymphocytes, including CD8⁺ T cells, was suppressed in both 5-week and 10-week treated mice (Figure 5A). SR1001 also suppressed the expression of Fas (Fas) and TNFα (Tnfa) and led to increased expression of Foxp3, the master transcription factor of T-regulatory cells (26). Finally, the expression of IL-6 (Il6), a cytokine critical for the development of T_H17 cells, and the IL-12p40 subunit (Il12p40), which contributes to both the IL-12 and IL-23 heterodimers, were also suppressed in SR1001-treated animals, suggesting that SR1001 was having a significant effect on various aspects of disease development outside of T_H17 cells (Figure 5A) (27, 28).

Figure 5.

Inhibition of proinflammatory cytokine expression in the spleen of NOD mice. A, Splenic mRNA expression of T_H17 cytokines, IFNγ, T-cell cytotoxic responses, and other proinflammatory cytokines from animals treated with SR1001 or vehicle for 30 days. Mice were either 5 weeks or 10 weeks of age at the start of the experiment and 9 or 14 weeks of age at the termination of the experiment, respectively (n = 10/group). FACS analysis of IL-17A-expressing CD4⁺ T cells (B), and IFNγ-expressing CD4⁺ (C) and CD8⁺ (D) T cells in 30-day treated animals is shown. Mice were 10 weeks old at the start of the experiment and 14 weeks old at the termination of the experiment. Data are mean ± SEM values (n = 5/group). *, P < .05; **, P < .01.

IFNγ is secreted by T_H1 and CD8⁺ T cells and is thought to be a key cytokine in diabetes onset. To determine whether SR1001 was affecting T_H1- or IFNγ-expressing CD8⁺ T cells, we performed intracellular cytokine staining on splenocytes from 30-day treated 5-week-old and 10-week-old NOD mice. We observed less IFNγ expression from CD4⁺ and CD8⁺ T cells in both the 5-week and 10-week SR1001-treated groups despite no difference in the number of CD4⁺ or CD8⁺ T cells between the vehicle or SR1001-treated groups (Figure 5, C and D, and data not shown). Analysis of IL-17A⁺IFNγ⁺ cells (double positive), converted T_H17 cells into T_H1-like cells, revealed very few double positives in our experimental groups, and of the few present, there was no difference between the drug-treated and vehicle groups (data not shown). Furthermore, we have published that SR1001 does not inhibit IFNγ expression from T_H1 cells or their development (21). However, these assays were performed in C57BL/6J mice, so we wanted to determine whether SR1001 inhibited the development of T_H1 cells from NOD mice. Using naïve CD4⁺ T cells isolated from NOD mice, we differentiated T_H1 and T_H17 cells in the presence of SR1001 or vehicle control. Similar to our previous report, SR1001 did not inhibit IFNγ expression in T_H1 cells, whereas SR1001 did inhibit the development of T_H17 cells, suggesting that in vivo inhibition of IFNγ was not occurring due to direct effects on T_H1 cells (Figure 6).

Figure 6.

SR1001 does not inhibit the in vitro development of T_H1 cells from NOD mice. Naïve CD4⁺ T cells were isolated from spleens and lymph nodes of 6- to 8-week-old NOD mice and differentiated into either T_H1 (top panels) or T_H17 (bottom panels) cells in the presence of SR1001 (5 mM) or vehicle (DMSO). Graphs to the left of the FACS plots represent the average of the data. Data are mean values ± SEM (n = 10/group). *, P < .05. DMSO, dimethylsulfoxide.

Inhibition of RORα and RORγ affects proinflammatory cytokine expression while increasing the expression of Foxp3

Based on our results, we next wanted to determine whether SR1001 treatment was also influencing the expression of T_H17-mediated and proinflammatory cytokines in the periphery and pancreas. We first measured plasma levels of the effector cytokines of T_H17, T_H1, and CD8⁺ T cells. Consistent with our mRNA results, we found that although there was some inhibition of T_H17-mediated cytokines and IFNγ in the mice that were 5 weeks old at the start of the experiment, there was significantly greater inhibition in the 10-week treated group evidenced by a decreased expression of IL-17F, IL-21, IL-22, and IFNγ (Figure 7). Although the plasma levels of cytokines were relatively low at these time points, it is apparent that they do rise over time, similar to what was observed with autoantibodies in Figure 4. Levels of IL-17A were extremely low at these time points, which is consistent with our FACS data.

Figure 7.

SR1001 suppresses the expression of proinflammatory cytokines in the periphery. Plasma samples from 30-day SR1001- or vehicle-treated female NOD mice. Mice were either 5 or 10 weeks of age at the start of the experiment and 9 or 14 weeks of age at the termination of the experiment, respectively. Levels of T_H17-mediated cytokines and IFNγ were analyzed using Luminex assays. Data are mean values ± SEM (n = 8–10/group). *, P < .05; **, P < .01.

Next, mRNA analysis of pancreas from 30-day treated, 5-week-old and 10-week-old NOD mice showed similar results to what was observed in the spleen and plasma, including a decreased expression of Il17a, Il17f, Il21, and Il22 (Figure 8A). Consistently, we also observed a decreased expression of Ifng and Prf-1 (Figure 8A). Consistent with what was observed in the spleen, the mRNA expression of Foxp3 was increased in the pancreas of SR1001-treated animals relative to vehicle controls (Figure 8A), suggesting that SR1001 might be effecting the T-regulatory cell population. To further explore the increased Foxp3 expression, FACS analysis of splenocytes from 5-week-old, short-term treated NOD mice showed a small but significant increase in the percentage of CD4⁺Foxp3⁺ cells in the SR1001-treated mice relative to vehicle control. There appeared to be little effect on the CD4⁺Foxp3⁺ cells in short-term, 10-week-old mice (Figure 8B and Supplemental Figure 5). However, FACS analysis of splenocytes from NOD mice treated with SR1001 for 4 months showed an even greater effect on the CD4⁺Foxp3⁺ population in both groups relative to vehicle control (Figure 8C and Supplemental Figure 5). These data suggest that the inhibition of ROR activity suppresses proinflammatory T-cell populations and cytokine expression, including T_H1 and T_H17 cells in vivo.

Figure 8.

SR1001 inhibits proinflammatory cytokine expression in the pancreas of NOD mice and increases the frequency of splenic CD4⁺ Foxp3⁺ T cells. mRNA expression of T_H17 cytokines, IFNγ, and Foxp3 in the pancreas of 30-day treated SR1001 or vehicle-treated mice (A) is shown. Mice were either 5 or 10 weeks of age at the start of the experiment and 9 or 14 weeks of age at the termination of the experiment, respectively (n = 10/group). FACS analysis was performed of splenic CD4⁺ T cells expressing Foxp3 in 30-day (B) and long-term (C)-treated SR1001- and vehicle-treated NOD mice. Mice were either 9 or 24 weeks of age at the termination of the experiment, respectively. Data are mean ± SEM values (n = 5–8/group). *, P < .05.

Discussion

Due to the known roles for the RORs in T_H17 cell development and autoimmunity, we initiated these studies to determine whether targeting the RORs may be a viable therapeutic option for the treatment of type 1 diabetes. Here we show that use of a synthetic RORα/γ-specific ligand, SR1001, facilitates a cascade of events that confers beneficial effects regulating immune-mediated pathology in NOD mice. SR1001 suppressed the immune response, including T_H17 cells in vitro and in vivo, T_H1 cells in vivo, autoantibody production, maintained insulin levels, and increased Foxp3 expression.

T_H17 cells have been linked to a number of organ-specific autoimmune diseases, including MS, and since we described the first RORα/γ-selective synthetic ligand that had utility in a mouse model of MS, SR1001, we believed that this approach was a feasible means to specifically address the roles of T_H17 cells in type 1 diabetes pathogenesis. We found that SR1001 did inhibit T_H17 cells and T_H17-mediated cytokine expression in NOD mice, supporting a pathological role for T_H17 cells in type 1 diabetes pathogenesis. Our data are in agreement with several studies that demonstrated that T_H17 cells and its effector cytokine, IL-17A, is needed for initiation of diabetes, suggesting that T_H17 cells are pathogenic (3, 22). Martin-Orozco et al (12) demonstrated that T_H17 cells are found surrounding the pancreatic islet, resulting in macrophage recruitment to the islets, promoting further immune infiltration. Furthermore, anti-IL-17 treatment reduced islet inflammation and prevented diabetes development when it was administered to NOD mice during the effector stage of type 1 diabetes onset (3, 12). Our data are consistent with this because we observed a decreased immune infiltration with SR1001 treatment.

T_H17 cells also secrete other cytokines, including IL-21, and several studies have established that deletion of IL-21, IL-21 receptor, or neutralization of IL-21 protects mice from diabetes onset (18, 29–31). We observed suppressed IL-21 expression with SR1001 treatment, which in combination with decreased IL-17 expression may account for inhibition of disease onset. These data suggest that targeting the RORs, which in turn can target multiple cytokines, is more protective than merely targeting one cytokine. Support for this hypothesis comes from studies using recombinant IL-25, a naturally occurring cytokine that has been shown to potently inhibit T_H17 cells and protect from experimental autoimmune encephalomyelitis (32). When IL-25 was administered to NOD mice, it was more effective at inhibiting disease onset than anti-IL-17 alone (3). Despite this evidence for the roles of T_H17 cells in type 1 diabetes pathogenesis, there have been studies suggesting that T_H17 cells are protective, and therefore, inhibiting T_H17 exacerbates diabetes. However, these studies addressed T_H17 cells indirectly via adjuvant treatment or knockdown of IL-17 in vivo (8, 9). Because T_H17 cells secrete many cytokines and these cytokines may collaborate to induce disease onset, it is possible that targeting one cytokine may not be sufficient; rather targeting the cell type as a whole may be more efficacious.

Further characterization of the immune repertoire in drug-treated NOD mice revealed that SR1001 treatment suppressed IFNγ expression from both CD4⁺ and CD8⁺ T cells. Although decreased IFNγ is an unexpected result, it potentially proves beneficial for diabetes outcome. How this may be occurring is still unknown because SR1001 is specific for the RORs. However, T_H17 cells have been demonstrated to convert to T_H1 cells in an inflammatory environment, including NOD mice (11, 12). Converted T_H17 cells are thought to be the pathogenic cells driving diabetes onset in NOD mice as well as other autoimmune diseases (11). Although we did not observe any changes in the double-positive cell population, we may not have assessed the correct time points. Therefore, it is conceivable that SR1001 may be inhibiting the conversion of T_H17 cells into T_H1 cells. It stands to reason that if there are fewer T_H17 cells, there are fewer cells available for conversion. We have not seen any substantial inhibition of conversion in vitro using SR1001, but in vitro conditions rarely mimic in vivo inflammatory conditions, and factors still unknown may be involved in this process (data not shown).

Another possibility for decreased IFNγ expression from T_H1 cells includes ROR-dependent, indirect effects on this population. Considering that both RORα and RORγ are expressed in various tissues throughout the body, it is possible that SR1001, through the suppression of ROR function in other cell types, is indirectly affecting the development of T_H17 and T_H1 cells. For example, both RORα and RORγt are expressed in macrophages, and inhibition of macrophage function could result in decreased cytokine expression or decreased antigen presentation (33–35). We observed decreased mRNA expression of the IL-12p40 subunit, which contributes to both the IL-12 and IL-23 heterodimers (27). RORα-deficient mice present with decreased IL-12 expression in a mouse model of ischemia-induced angiogenesis, and IL-12 is critical for the development of T_H1 cells (36). Furthermore, we observed decreased mRNA expression of IL-6, a cytokine that is critical for the development of T_H17 cells, and among other cell types, is also secreted by macrophages (28, 37, 38). Finally, the decreased IFNγ expression from CD8⁺ T cells may be attributed to decreased T_H17 and T_H1 responses because diabetogenic CD8⁺ T cells may require CD4⁺ T cell help to reach their maximum diabetogenic potential (39). Therefore, the inhibition of the RORs in vivo may have both direct and indirect effects, affecting the development of type 1 diabetes. Further studies assessing the expression and function of the RORs in various immune cell populations contributing to diabetes onset will be needed to dissect these potential mechanisms.

Our data also demonstrated that inhibition of RORα/γ activity affected B-cell reactivity to several autoantigens. ELISAs demonstrated that SR1001 treatment inhibited the formation of autoantibodies to insulin and GAD65. This inhibitory effect is consistent with data from a mouse model of inducible autoimmune myasthenia gravis, which demonstrated that IL-17 knockout animals are resistant to the development of autoantibody formation resulting from the loss of T_H17 help (40). In line with this, another study observed significantly reduced GAD65 autoantibodies in NOD mice after anti-IL-17 treatment (3). Finally, we can not rule out a direct effect of SR1001 on the formation of autoantibodies in NOD mice because RORα has been demonstrated to play a critical role in the generation and maintenance of memory IgA⁺ B cells (41). Thus, these data also allude to potential indirect effects of ROR inhibition on disease processes in this model. However, further investigation into how SR1001 affects autoantibody formation in NOD mice is needed to elucidate this process.

Our data also suggest that SR1001 influenced Foxp3⁺ mRNA expression and protein expression in CD4⁺ T cells. We have observed a similar phenomenon in vitro with SR1555, a RORγ inverse agonist, suggesting that inhibition of ROR activity could positively affect the balance between proinflammatory and tolerogenic T cells (19). Increased Foxp3⁺ T-regulatory cell numbers could equate to greater suppression and another potential mechanisms of how SR1001 protects from diabetes. However, inhibition of ROR activity alone may not be sufficient for full T regulatory cell development because the cytokine milieu also plays a significant role in the development of T_H17 over T-regulatory cells (42). Further studies exploring the generation and effector function of T-regulatory cells over T_H17 cells with ROR inhibitors is warranted to better define the mechanisms driving the T_H17/T-regulatory cell balance and ascertain whether ROR-specific, small-molecule inhibitors may be a viable therapeutic option to exploit and alter this balance.

Whereas autoimmunity is the predominant effector mechanism in type 1 diabetes, underlying metabolic disorders preside, specifically dysregulated glucose metabolism. Gluconeogenesis is inappropriately active, resulting in a large amount of circulating glucose contributing to both fasting and postprandial hyperglycemia (43, 44). Prior to their identification as required factors for full T_H17 cell development, the RORs were extensively studied due to their roles in various physiological processes, including the regulation of the circadian rhythm, lipid, and glucose metabolism (45). Strict control of blood glucose levels is considered essential to delay or prevent the development of diabetic complications (46). Thus, in addition to the immune mediating effects observed with SR1001 treatment, modulation of ROR activity in the liver may have beneficial effect on the regulation of blood glucose levels, suggesting that ROR regulation of hepatic glucose output may be an attractive therapeutic target for the regulation of glucose in type 1 diabetes patients. RORα and RORγ are also expressed in the pancreas and islets, respectively, and regulate the expression of insulin and glucagon in a circadian manner, suggesting a role for the circadian rhythm and the RORs in type 1 diabetes. Work elucidating the roles for the RORs in these processes needs further exploration.

Although there have been considerable advances made in the understanding of type 1 diabetes pathogenesis, there are still many unknown aspects surrounding disease onset and progression. Current therapies for type 1 diabetes are limited, with new efforts focusing on either the metabolic or immune aspects of the disease. Thus, there remains a lack of therapeutic options for type 1 diabetes patients. Our studies reveal the global effects of inhibiting ROR activity on the pathogenesis of type 1 diabetes. However, further work to elucidate the individual roles of each ROR as well as the tissue specific effects of ROR inhibition is warranted. These studies would enable us to gain a broader perspective of the roles of the RORs in type 1 diabetes and enable the design of more focused therapeutics to treat type 1 diabetes. In conclusion, we have demonstrated that small-molecule regulation of ROR activity positively affects type 1 diabetes disease outcome, suggesting that small-molecule inhibitors to the ROR may prove beneficial for type 1 diabetes treatment because they target many arms of type 1 diabetes disease pathogenesis.