Role of miR-146a in the Regulation of Inflammation in an In Vitro Model of Graves' Orbitopathy

Abstract

Purpose: To investigate the role of microRNA 146a (miR-146a) in the regulation of inflammation in an in vitro model of Graves' orbitopathy (GO).

Methods: The level of miR-146a expression in orbital adipose tissue was compared between GO and non-GO by quantitative real-time PCR (qPCR). The effects of interleukin 1β (IL-1β) on miR-146a expression were analyzed in orbital fibroblasts by qPCR. To investigate the molecular mechanism underlying IL-1β–induced miR-146a expression, the effects of inhibitors of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), mitogen-activated protein kinase/extracellular signal–regulated kinases (MEK)-1/2, c-Jun N-terminal kinases (JNK)-1/2, p38 MAP kinase, and phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K) were analyzed. The effects of miR-146a mimics and inhibitors on IL-1β–induced IL-6 release were examined by ELISA and Western blotting.

Results: The level of miR-146a expression was significantly higher in GO orbital adipose tissue than in non-GO (P = 0.032). Interleukin 1β induced a time- and concentration-dependent increase in miR-146a expression. Interleukin 1β (10 ng/mL, 16 hours) induced an approximately 17.5-fold increase in miR-146 expression. The increase in miR-146a expression by IL-1β was significantly inhibited by NF-κB, JNK-1/2, and PI3K inhibitors (1.94  ±  0.25, 5.28  ±  0.34 and 9.73  ±  2.32-fold, respectively, P < 0.05 compared with IL-1β–induced miR-146 expression, independent t-test). Interleukin 1β–induced IL-6 protein production was further decreased by miR-146a mimics, but not by inhibitors of miR-146a.

Conclusions: MicroRNA 146a was upregulated by inflammatory stress in orbital fibroblasts. Our results indicated that miR-146a had a positive effect on the anti-inflammatory process. MicroRNA 146a may play a role in the regulation of inflammation in orbital fibroblasts, and may participate in the pathogenesis of GO.

Graves' orbitopathy (GO) is an inflammatory autoimmune disorder of the orbit. Previous studies have indicated that the thyroid stimulating hormone (TSH) receptor, which is expressed on orbital fibroblasts, is the autoimmune target of GO.^1–4 Binding of autoantibodies to TSH receptors, expressed on orbital fibroblasts, activates the T cell–dependent inflammatory process. Thus, GO is believed to be related to T cell–mediated autoimmunity to an antigen present in orbital fibroblasts. Activated CD4⁺ T cells secrete IL-1, IFN-γ, and TNF-α, inducing the expression of TSH receptor and CD40 on the surface of orbital fibroblasts, which promote the secretion of IL-6, −8, fibronectin, type 1 collagen, and glycosaminoglycans.^5–7 Interaction with CD4⁺ T cells enhances orbital fibroblast activation, proliferation, differentiation, and lipid accumulation. 

MicroRNAs (miRNAs) are endogenous, single-stranded, noncoding RNAs, 18 to 24 nucleotides in length that can play important regulatory roles by targeting mRNAs for cleavage or translational repression.^8–10 Thus, they negatively regulate gene expression at the posttranscriptional level. Several recent studies have shown that inflammatory autoimmune diseases, such as rheumatoid arthritis (RA),¹¹ systemic lupus erythematosus (SLE),^12,13 ulcerative colitis,¹⁴ and psoriasis¹⁵ have been reported to be associated with altered miRNA expression. Thus, based on reports that inflammatory autoimmune conditions are related to altered miRNA expression, we postulated that specific miRNAs may also be associated with GO. 

Located in the LOC285628 gene on human chromosome 5, miRNA 146a (miR-146) is a relatively well known miRNA in inflammatory autoimmune diseases.^10,16 Studies have indicated that miR-146a plays an important role in the pathogenesis of several autoimmune disorders, such as RA,^11,17 SLE,¹³ osteoarthritis (OA),¹⁸ and Sjögren syndrome.¹⁹ MicroRNA 146a seems to act through inhibition of the nuclear factor kappa-light-chain-enhancer of activated B cell (NF-κB) pathway by downregulation of its target genes, such as TNF receptor–associated factor 6 (TRAF6) and IL-1 receptor–associated kinase 1 (IRAK1).²⁰ This leads to termination or mitigation of an inflammatory response. Based on this background, we focused on miR-146a in an inflammatory cellular model of orbital fibroblast inflammation induced by IL-1β. 

In this study, we first compare the expression levels of miR-146a in orbital adipose tissue between GO and non-GO. Then, we determine the role of miR-146a in the regulation of inflammation in an in vitro model of GO. 

Methods

Subjects and Cell Culture Protocol

Orbital adipose/connective tissue explants were obtained from 19 GO patients and from 17 age- and sex-matched control subjects with no history of GO. All GO patients underwent orbital decompression for proptosis correction and control subjects underwent cosmetic upper and lower blepharoplasty. Informed written consent was obtained from all subjects and this study was approved by the Institutional Review Board of Severance Hospital, Yonsei University College of Medicine. All GO patients were euthyroid status at the time of surgery and had not been treated with steroids or radiation therapy for at least 3 months. 

Orbital fibroblast cell cultures were performed according to the methods described previously.^21–23 Briefly, for primary cell cultures, tissue explants were minced and placed in plastic culture dishes containing Dulbecco's modified Eagle's medium (DMEM):F12 (1:1) (Lonza, Basel, Switzerland), 20% fetal bovine serum (FBS; Life Technologies, Carlsbad, CA, USA), and penicillin–streptomycin (Life Technologies). After orbital fibroblasts had grown out from the explants, monolayers were passaged serially by gentle treatment with trypsin/EDTA, and cells were incubated in DMEM with 10% FBS and antibiotics. Cell cultures were grown in a humidified 5% CO₂ incubator at 37°C. Cells were stored in liquid N₂ until needed and used between the third and seventh passage. 

Microarray Analysis

For the microarray analysis, the total RNA in each orbital adipose/connective tissue sample obtained from eight GO patients and six non-GO normal controls was extracted using a commercial reagent (TRI Reagent; Molecular Research Center, Cincinnati, OH, USA), according to the manufacturer's instructions. The quality and quantity of total RNA were assessed using a bioanalyzer (Agilent Bioanalyzer 2100; Agilent Technologies, Santa Clara, CA, USA). Starting with 250 ng of total RNA, the labeling process began by adding a poly(A) tail to each RNA strand using poly(A) polymerase, followed by ligation of biotin-labeled 3DNA dendrimer. Biotinylated RNA strands were hybridized at 48°C for 18 hours on a miRNA array (Affymetrix GeneChip miRNA 4.0 Array; Affymetrix, Santa Clara, CA, USA). The miRNA array (Affymetrix), with 2578 human mature miRNA, was washed and stained in an array cartridge system (Affymetrix Fluidics Station 450; Affymetrix). Amplified fluorescent signals were scanned using a commercial scanner (Affymetrix GeneChip Scanner 3000 7G; Affymetrix). 

The arrays were analyzed using a scanner with associated software (Agilent). The miRNA expression levels were calculated with a commercial console (Expression Console 1.4; Affymetrix). Relative signal intensities for each miRNA were generated using the robust multi-array average algorithm. The data were processed based on the quantile normalization method using commercial software (GeneSpring GX 13.1; Agilent Technologies). This normalization method aims to achieve a consistent distribution of intensities for each of a set of arrays. The normalized and log-transformed intensity values were then analyzed using commercial software (Agilent Technologies). The fold changes in the miRNA expression between the GO and non-GO samples were calculated from the signal values. MicroRNA expression was considered significantly different if the fold change exceeded 1.5. Target prediction was performed using a cutoff at the 95% percentile using the TargetScan6.2 database in the public domain (http://www.targetscan.org/). 

Measurement of miR-146a Expression by Quantitative Real-Time PCR

For quantitative real-time PCR (qPCR), the total RNA (1 μg) of each orbital adipose/connective tissue sample obtained from another five GO patients and five non-GO normal controls was isolated using a miRNA isolation kit (mirVana; Ambion, Austin, TX, USA), and reverse-transcribed into complementary DNA using a microRNA reverse transcription kit (TaqMan; Applied Biosystems, Carlsbad, CA, USA). The resulting cDNA was amplified using a thermocycler (ABI StepOnePlus Real Time PCR; Applied Biosystems) with a universal PCR master mix (TaqMan No AmpErase UNG; Applied Biosystems) and the recommended PCR conditions for quantitative assessment of gene transcript levels in the tissue samples. All PCRs were performed in triplicate. The catalog number of the primers used was 000468 for miR-146a. RNU6B expression was used for normalization, and the results were expressed as relative fold changes of threshold cycle (Ct) value relative to the control group using the 2^–ΔΔCt method.²⁴ 

Cell Stimulation

Orbital fibroblasts were plated onto 6-well plates for assessment of cytokine release and RNA extraction. Cells were stimulated in triplicate in DMEM:F12 with the indicated IL-1β (R&D Systems, Minneapolis, MN, USA) concentration (0, 1, 5, 10, and 20 ng/mL) or with 10 ng/mL of IL-1β for the indicated time (0, 3, 6, 16, and 24 hours). 

To assess the molecular mechanism of IL-1β–induced miR-146a expression, the effects of inhibitors of MEK-1/2, JNK-1/2, p38 MAP kinase, and PI3-K were investigated using GO orbital fibroblasts. In the present study, we used MEK-1/2 (PD098059; Sigma-Aldrich Corp., St. Louis, MO, USA), JNK-1/2 (SP600125; Sigma-Aldrich Corp.), p38 MAP kinase (SB203580; Sigma-Aldrich Corp.), and PI3-K (LY294002; Sigma-Aldrich Corp.). To examine the effects on IL-1β–induced miR-146a expression, these inhibitors were added at a concentration of 20 μM 60 minutes prior to addition of IL-1β (10 ng/mL). 

To assess the possible involvement of proinflammatory transcription factors, including the NF-κB pathway, the effects of preincubation with SC-514 (Calbiochem, La Jolla, CA, USA), a selective inhibitor of kappa light polypeptide gene enhancer in B-cells (IκB) kinase-2 inhibitor, and dexamethasone were examined.²⁵ Following 1-hour pretreatment with SC-514 (100 μM) and dexamethasone (0.1 μM), orbital fibroblasts were stimulated with IL-1β (10 ng/mL) and the expression of miR-146a was determined at 16 hours. 

Transfection With miR-146a Mimics and Inhibitors

Orbital fibroblasts were transfected with miR-146a mimics, inhibitors, and each control according to the respective manufacturer's protocol. Three separate experiments were performed using cells from three different individuals. The microRNA 146a mimics were obtained from Ambion/Applied Biosystems and miR-146a inhibitors were obtained from Exiqon (Vedbaek, Denmark). We used a commercial reagent (Lipofectamine 2000; Life Technologies) as a negative control. Orbital fibroblasts were plated onto six-well plates for assessment of cytokine release. Cells were transfected with the indicated miR146a mimics and inhibitors at concentrations of 0, 10, 30, or 100 nM using commercial reagents (Lipofectamine RNAiMAX; Life Technologies). 

Measurement of IL-6 Secreted by ELISA

The effects of miR-146a mimics and inhibitors on IL-1β–induced IL-6 release expression were analyzed using a human cytokine ELISA kit (R&D Systems) according to the manufacturer's protocol for three GO and three non-GO orbital fibroblasts from different individuals. 

Transfected cells were plated into six-well plates and left to adhere overnight. The cells were then starved for 6 hours prior to stimulation with 10 ng/mL IL-1β. Supernatants were removed at 24 hours and IL-6 levels were determined by ELISA. 

Western Blotting

The effects of miR-146a mimics on IL-1β–induced IL-6, cyclooxygenase (COX)-2, and intercellular adhesion molecule (ICAM)-1 release in GO orbital fibroblasts were analyzed by Western blotting. Transfected cells were washed with ice-cold PBS, and whole-cell lysates were obtained by incubation on ice for 30 minutes in cell lysis buffer (20 mM HEPES, pH 7.2, 10% (vol/vol) glycerol, 10 mM Na₃VO₄, 50 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 0.1 mM dithiothreitol, 1 μg/mL leupeptin, 1 μg/mL pepstatin, and 1% (vol/vol) Triton X-100). Reagents were purchased from Sigma-Aldrich Corp. Lysates were centrifuged at 12,000g for 10 minutes and the cell homogenate fractions were stored at −70°C until use. 

Protein concentrations were determined by the Bradford assay.^21,22 Equal amounts of protein (50 μg) were boiled in sample buffer and resolved by 10% (wt/vol) SDS-PAGE. Proteins were transferred onto polyvinylidene difluoride membranes (Immobilon; Millipore, Billerica, MA, USA). The samples were probed overnight with primary antibodies (IL-6, COX-2, and ICAM-1) in tris-buffered saline containing Tween 20 (TBST), and washed three times with TBST. Immunoreactive bands were detected with horseradish peroxidase-conjugated secondary antibody and developed using an enhanced chemiluminescence kit (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and exposed to x-ray film (Amersham Pharmacia Biotech). The immunoreactive bands were quantified by densitometry and normalized relative to the β-actin level in the same sample. 

Data Analysis

All experiments were performed at least three times independently, and using at least three cell cultures harvested from different individuals. The results are presented as means  ±  SD. Differences between groups were assessed by independent and paired t-tests. In all analyses, P < 0.05 was taken to indicate statistical significance. 

Results

miR-146a Expression in GO

To determine which miRNAs are involved in the pathogenesis of GO, we performed microarray analysis using GO (n = 8) and non-GO (n = 6) orbital connective tissue. Microarray analyses showed that 38 miRNAs were upregulated and 7 miRNAs were downregulated in orbital connective tissue from GO patients compared to healthy controls (Fig. 1). The heat map represents the results using a color intensity scale with the highest and lowest expression levels corresponding to bright red and bright blue, respectively (Fig. 1A). 

Figure 1

 

MicroRNA expression profiling in GO. The heat map represents the expression value of each specimen (eight GO and six non-GO) using color intensity. Gene names of microRNA, showing a more than 1.5-fold change, are listed to the right. In total, 38 miRNAs were upregulated and 7 microRNAs were downregulated in orbital connective tissue from GO patients, compared with healthy normal controls. Six of the eight GO subjects had higher miR-146a expression level than the mean control value. miR, microRNA; NL, normal.

Among the miRNAs overexpressed in GO, we noted that miR-146a, which is a relatively well-known miRNA in inflammatory autoimmune diseases, was increased by 3.15-fold compared with normal controls. Six of the eight GO subjects had miR-146a expression higher than the mean control value (Fig. 1). To confirm the results of microarray analysis, we performed qPCR of miR-146a expression in RNA samples obtained from another five GO patients and five non-GO control subjects. The expression of miR-146a was significantly higher in orbital adipose tissues from GO than from non-GO subjects (P = 0.032, independent t-test; Fig. 2). 

Figure 2

 

Expression of miR-146a in GO. Expression of miR-146a was significantly higher in orbital adipose tissues from GO than from non-GO subjects. (*P < 0.05). The results are expressed as the means  ±  standard deviation of five individual samples and the graphs are representative of three independent experiments.

Effects of IL-1β on miR-146a Expression

Orbital fibroblasts from three GO and three non-GO patients were used to measure miR-146a expression following exposure to IL-1β. In an experiment using GO orbital fibroblasts, IL-1β (10 ng/mL) caused a 5-fold increase in miR-146a expression level at 6 hours (4.96  ±  0.73-fold, P = 0.019 compared with control, paired t test), which continued to rise to 17.05  ±  1.38-fold at 16 hours and 18.63  ±  1.41-fold at 24 hours (Fig. 3A). Interleukin 1β (16 hours) induced a concentration-dependent increase in miR-146a expression. Interleukin 1β caused an increase in miR-146a expression level by about 12-fold at 1 ng/mL (12.38  ±  1.18-fold, P < 0.001 compared with the control, paired t-test), which continued to rise to 14.35  ±  2.10-fold at 5 ng/mL, 17.54 ± 2.70-fold at 10 ng/mL, and 18.75  ±  2.42-fold at 20 ng/mL (Fig. 3B). Interleukin 1β induced a time- and concentration-dependent increase in miR-146a expression. 

Figure 3

 

Effects of IL-1β on miR-146a expression in GO and non-GO orbital fibroblasts. Interleukin 1β induced a time- and concentration dependent increase in miR-146a expression in orbital fibroblasts from both GO and non-GO specimens (*P < 0.05, versus time and concentration-matched controls). We found that non-GO cells showed a similar response as GO orbital fibroblasts, but higher levels of miR-146a expression were observed in non-GO cells under some conditions (16 hours and 24 hours: 10 ng/mL; 16 hours: 1, 5, 10, and 20 ng/mL), compared with GO orbital fibroblasts (**P < 0.05, comparison between GO and healthy non-GO with IL-1β treatment). The results are expressed as the means  ±  standard deviation of three individual samples and the graphs are representative of three independent experiments.

The same experiments were performed using orbital fibroblasts from non-GO patients. The results showed a similar response to GO orbital fibroblasts, but higher levels of miR-146a expression were observed in non-GO cells under some conditions (16 and 24 hours: 10 ng/mL; 16 hours: 1, 5, 10, and 20 ng/mL), compared with GO orbital fibroblasts (independent t-test, P < 0.05; Fig. 3). 

Effects of Inhibitors of MEK-1/2, JNK-1/2, p38 MAP kinase, PI3-K and NF-κB on IL-1β-Induced miR-146a Expression

To investigate the molecular mechanism underlying IL-1β–induced miR-146a expression, the effects of inhibitors of MEK-1/2, JNK-1/2, p38 MAP kinase, and PI3-K were analyzed using GO orbital fibroblasts. Interleukin 1β (10 ng/mL, 16 hours) induced an increase of about 17.5-fold in miR-146 expression. This increase in miR-146a expression induced by IL-1β was significantly inhibited by JNK-1/2 (5.28  ±  0.34-fold, P = 0.030 compared with IL-1β–induced miR-146 expression, independent t test) and PI3K inhibitors (9.73  ±  2.32-fold, P = 0.039), but not by those of MEK-1/2 and p38 MAP kinase (Fig. 4). 

Figure 4

 

Effects of inhibitors of MEK-1/2, JNK-1/2, p38 MAP kinase, PI3-K, NF-κB and dexamethasone on IL-1β–induced miR-146a expression in GO orbital fibroblasts. The increase in miR-146a expression induced by IL-1β (17.54  ±  3.84-fold) was inhibited by JNK-1/2 (5.28  ±  0.34-fold, P = 0.030) and PI3K inhibitors (9.73  ±  2.32-fold, P = 0.039), but not by MEK-1/2 or p38 MAP kinase inhibitors (*P < 0.05, versus IL-1β–induced miR-146 expression). The level of miR-146a expression was significantly decreased by pretreatment with SC-514 (1.94  ±  0.25-fold) and dexamethasone (4.95  ±  0.48-fold) compared with the IL-1β–induced miR-146 expression (17.54  ±  3.84-fold, P = 0.019 and 0.029, respectively). The results are expressed as the means  ±  standard deviation of three individual samples and the graphs are representative of three independent experiments. Dexa, dexamethasone; LY, PI3-K inhibitor; PD, MEK-1/2 inhibitor; SB, p38 MAP kinase inhibitor; SC, NF-κB inhibitor; SP, JNK-1/2 inhibitor.

To assess the involvement of the NF-κB pathway, samples were pretreated with SC-514 and the miR-146a expression was examined. The SC-514 pretreatment significantly decreased the level of miR-146a expression by (1.94  ±  0.25-fold) compared with the IL-1β–induced miR-146 expression (17.54  ±  3.84-fold, P = 0.019). 

Glucocorticoids have powerful anti-inflammatory actions. Therefore, we examined whether dexamethasone affects the IL-1β–induced miR-146a expression. Dexamethasone pretreatment significantly decreased the level of miR-146a expression (4.95  ±  0.48-fold) compared with the IL-1β–induced miR-146 expression (17.54  ±  3.84-fold, P = 0.029). 

Effects of miR-146a Mimics and Inhibitors on IL-1β–Induced IL-6 Protein Production

We transfected three GO and three non-GO orbital fibroblasts with 0, 10, 30, and 100 nM miR-146a mimics and inhibitors. In an experiment using GO orbital fibroblasts, IL-1β induced IL-6 protein production (526.59 ± 19.55 pg/mL) compared with baseline (86.82  ±  3.01 pg/mL). Interleukin 1β–induced IL-6 protein production was decreased by miR-146a mimics (480.59  ±  17.80 pg/mL at 30 nM and 418.15  ±  15.83 pg/mL at 100 nM, P = 0.017 and 0.001, respectively, compared with control mimics). Mimics of miR-146a induced a concentration-dependent decrease in IL-1β–induced IL-6 protein production, whereas miR-146a inhibitors did not induce any changes in IL-6 protein production (Fig. 5). 

Figure 5

 

Effects of miR-146a mimics and inhibitors on IL-1β–induced IL-6 protein production expression in GO and non-GO orbital fibroblasts. Interleukin 1β induced less IL-6 production by non-GO orbital fibroblasts (481.05 ± 20.01 pg/mL) than GO orbital fibroblasts (526.59 ± 19.55 pg/mL, **P = 0.048). Interleukin 1β–induced IL-6 protein production was decreased by miR-146a mimics (30 and 100 nM, *P < 0.05 compared with control mimics), whereas miR-146a inhibitors did not induce any changes in IL-6 protein production in an experiment using GO orbital fibroblasts. Interleukin 1β–induced IL-6 protein production was decreased by miR-146a mimics (10, 30 and 100 nM, *P < 0.05 compared with control mimics), whereas miR-146a inhibitors did not induce any changes in IL-6 protein production in an experiment using non GO orbital fibroblasts. The results are expressed as the means  ±  standard deviation of three individual samples and the graphs are representative of three independent experiments.

The same experiments were performed using non-GO orbital fibroblasts. The results showed a similar response to GO orbital fibroblasts. Mimics of miR-146a induced a decrease in the IL-1β–induced IL-6 protein production at concentrations of 10, 30, and 100 nM, whereas miR-146a inhibitors did not induce any change in the IL-6 protein production. 

Of note, IL-1β induced less IL-6 production in non-GO orbital fibroblasts (481.05 ± 20.01 pg/mL) than in GO orbital fibroblasts (526.59 ± 19.55 pg/mL, P = 0.048). In an experiment using non-GO orbital fibroblasts, IL-1β–induced IL-6 protein production was significantly decreased by 10 nM miR-146a mimics (413.39 ± 15.34 pg/mL, P = 0.011), whereas 10 nM miR-146a mimics did not induce any changes in IL-6 protein production in GO orbital fibroblasts (Fig. 5). 

Effects of miR-146a Mimics on IL-1β–Induced IL-6, COX-2, and ICAM-1 Protein Production

We further investigated the effects of miR-146a mimics on the expression of inflammatory proteins, such as IL-6, COX-2, and ICAM-1, by Western blotting using orbital fibroblasts from GO. Stimulation with IL-1β increased the levels of IL-6, COX-2, and ICAM-1 expression. The increases in IL-6 and ICAM-1 were inhibited by miR-146a mimics (30 nM), but no significant effect was observed on COX-2 expression (Fig. 6). 

Figure 6

 

Effects of miR-146a mimics on IL-1β–induced IL-6, COX-2, and ICAM-1 protein production in GO orbital fibroblasts. The increases in IL-6 and ICAM-1 were inhibited by miR-146a mimics (30 nM) (*P < 0.05, versus IL-1β–induced IL-6 and ICAM-1 production). However, COX-2 did not show significant results. The results are expressed as the means  ±  standard deviation of three individual samples and the graphs are representative of three independent experiments.

Discussion

This study investigated the role of miR-146a in regulation of inflammation in an in vitro model of GO. We compared the expression level of miR-146a between GO and non-GO orbital adipose tissue. The expression of miR-146a was significantly higher in orbital adipose tissues from GO than from non-GO subjects (P = 0.032). GO is an inflammatory autoimmune disease, and thus miR-146a seems to be upregulated by inflammatory stress, as in GO. Similarly, there have been several reports that expression of miR-146a was increased in diseased tissue.^11,15,17,18 Nakasa et al.¹¹ reported that miR-146a was highly expressed in RA synovial tissue, and Yamasaki et al.¹⁸ reported strong miR-146a expression in cartilage in low-grade OA. Sonkoly et al.¹⁵ reported that miR-146a was significantly overexpressed in psoriatic lesional skin compared with healthy skin. 

Several recent pharmacologic studies investigated the mechanism of IL-1β–induced miR-146a expression using different cell lines.^26,27 Larner-Svensson et al.²⁷ reported that IL-1β induced a time-dependent, 100-fold induction in miR-146a expression in primary human airway smooth muscle cells. Perry et al.²⁶ also reported that IL-1β induced a time- and concentration-dependent increase in miR-146a expression in human lung alveolar epithelial cells. In this study, we found that IL-1β induced time- and concentration-dependent increases in miR-146a expression in both GO and non-GO orbital fibroblasts. Interestingly, the expression of miR-146a was upregulated to a lesser extent in GO orbital fibroblasts than in non-GO orbital fibroblasts after stimulation with IL-1β (P < 0.05). This result is in line with studies that showed that GO orbital fibroblasts produced significantly lower levels of IL-1 receptor antagonist (IL-1RA) compared with non-GO orbital fibroblasts.²⁸ Because IL-1RA acts as a competitive inhibitor of IL-1α and IL-1β and reduces the inflammatory process in GO orbital tissues,^28–30 we believe that these findings support the hypothesis that GO orbital fibroblasts have a reduced capacity to control the inflammatory response. 

Inhibitors of MEK-1/2, JNK-1/2, p38 MAP kinase, and PI3-K were used to investigate which signal cascade controls the increase in miR-146a expression by IL-1β. The results indicated that the increase in miR-146a expression induced by IL-1β was inhibited by JNK-1/2 and PI3K inhibitors, suggesting that activation of JNK and PI3K pathways was required for IL-1β–induced miR-146a expression in orbital fibroblasts. Larner-Svensson et al.²⁷ reported similar experimental results indicating that IL-1β–induced miR-146a was regulated by MEK-1/2 and JNK-1/2 in human airway smooth muscle cells. In their report, the extent of the decrease in miR-146a expression was greater for JNK-1/2 than MEK-1/2. In the present study, the JNK-1/2 pathway, but not MEK1/2, was associated with IL-1β–induced upregulation of miR146a. As the function and mechanism of action of miR-146a are dependent on the cell type,²⁷ we assumed that this discrepancy was due to the different cell lines used in these studies. The observation that the PI3K pathway was required for IL-1β–induced miR-146a expression in orbital fibroblasts can be explained. Recently, a cAMP-independent cascade was shown to increase PI3K activity, a signaling pathway that plays a central role in the pathogenesis of GO.³¹ Kumar et al.³¹ reported that a stimulatory TSH receptor antibody enhanced adipogenesis via PI3K activation in GO, and suggested that inhibition of PI3K signaling may represent a potential novel therapeutic approach in GO. 

Dexamethasone, a corticosteroid, attenuates the actions of multiple proinflammatory transcription factors, including NF-κB. We found that IL-1β–induced miR146a expression was inhibited after treatment with dexamethasone. Moreover, we found that the level of miR-146a expression was significantly decreased by SC-514, which is a selective NF-κB inhibitor. According to a previous report,²⁰ miR-146a acts through inhibition of the NF-κB pathway by downregulating its target genes, such as TRAF6 and IRAK1. This leads to the termination or mitigation of the inflammatory response. 

In this study, we found that IL-1β–induced IL-6 protein production was further decreased by miR-146a mimics, confirming the positive effect of miR-146a on the anti-inflammatory process. We further investigated whether miR-146a mimics could affect the levels of other inflammatory cytokines induced by IL-1β and found that increases in IL-6 and ICAM-1 expression by IL-1β were inhibited by miR-146a mimics (30 nM). The increase in COX-2 expression by IL-1β was decreased by miR-146a mimics, although the effect was not statistically significantly. Recently, Wei et al.³² reported significantly lower circulating serum levels of miR-146a in GO patients compared with controls, and suggested that weakened miR-146a activity caused an increase in inflammation. The authors further showed that the serum levels of miR-146a were significantly correlated with clinical activity score, which indicates the disease inflammatory activity of GO. 

Attenuation of miR-146a activity using an miR-146a inhibitor had no significant effect on IL-1β–induced IL-6 release in this study. According to a previous review,³³ there are multiple and complex pathways for the release of IL-6 by cells, and cells use different pathways, organelles, carriers, and molecules to control the release of cytokines. As the molecular mechanism of IL-1β–induced IL-6 release in orbital fibroblasts has not yet been fully elucidated, further physiological studies are needed to determine why miR-146a inhibitors did not induce any changes in IL-1β–induced IL-6 release. Although several questions could not be answered by the results of this study, we showed for the first time that miR-146a may play a role in the regulation of inflammation in orbital fibroblasts from GO and that it participates in the pathogenesis of GO. 

In conclusion, miR-146a seems to contribute to GO pathogenesis by modulating inflammatory protein expression and cellular functions in orbital fibroblasts. MicroRNA 146a was upregulated by inflammatory stress, such as IL-1β, and our results indicated a positive effect of miR-146a on the anti-inflammatory process. Further studies are required to examine the potential of miR-146a as a target for therapeutic strategies and as a biomarker. 

Acknowledgments

Supported by a grant from the National Research Foundation of Korea (2014R1A1002754) and in part by the Soonchunhyang University Research Fund. 

Disclosure: S.Y. Jang, None; M.K. Chae, None; J.H. Lee, None; E.J. Lee, None; J.S. Yoon, None