HDAC10 upregulation contributes to interleukin 1β‐mediated inflammatory activation of synovium‐derived mesenchymal stem cells in temporomandibular joint

Abstract

Histone deacetylases (HDACs) are important in chronic inflammation, and inflammatory responses affect synovium‐derived mesenchymal stem cell (SMSC) function in temporomandibular joint repair. However, the effect of HDACs on SMSC inflammatory activation remains unclear. In this study, temporomandibular joint fibroblast‐like synoviocytes obtained from osteoarthritis patients met the minimal mesenchymal stem cell criteria. Interleukin 1β (IL‐1β) upregulated IL‐6 and IL‐8 expression in SMSCs through nuclear factor‐κB (NF‐κB) pathway activation. IL‐6 and IL‐8 upregulation were blocked by broad‐acting HDAC inhibitors SAHA and LBH589. MC1568 alleviated IL‐1β activation of SMSCs, whereas CI994 and FK228 produced a minimal or opposite effect in vitro. We also found HDAC10 was highly associated with localized IL‐1β expression in vivo and in vitro. HDAC10 knockdown alleviated IL‐1β‐mediated SMSC activation and blocked NF‐κB pathway activation. Conversely, HDAC10 overexpression promoted IL‐6 and IL‐8 expression and IL‐1β‐mediated NF‐κB pathway activation. In conclusion, HDAC10 upregulation contributed to IL‐1β‐mediated inflammatory activation of SMSCs, indicating that HDAC10 may be a novel therapeutic target.

The inflammation activation of synovium‐derived mesenchymal stem cells (SMSCs) mediated by interleukin 1β (IL‐1β) could be blocked by the histone deacetylase (HDAC) inhibitors, vorinostat, and panobinostat, by inhibiting nuclear factor‐κB (NF‐κB) pathway activation. HDAC10 upregulation contributed to inflammatory activation by IL‐1β through the NF‐κB pathway in SMSCs, indicating that it might be a new therapeutic target of temporomandibular joint disorders.

Abbreviations

HAT

histone acetyltransferase

HDAC

histone deacetylase

miR

microRNA

MMP‐13

matrix metallopeptidase‐13

MSC

mesenchymal stem cell

NF

nuclear factor

OA

osteoarthritis

siRNA

short interfering RNA

SMSC

synovium‐derived mesenchymal stem cell

TMD

temporomandibular joint disorders

TMJ

temporomandibular joint

TNF

tumor necrosis factor

1. INTRODUCTION

Temporomandibular joint disorders (TMD) are characterized by synovial inflammation, articular destruction, and jaw movement abnormalities, and are some of the most prevalent diseases of the stomatognathic system. The ability of self‐repair and regeneration of articular cartilage is limited because of its avascular nature, which allows early cartilage injuries to rapidly develop to severe osteoarthropathy. Recent studies on cartilage repair using synovium‐derived mesenchymal stem cells (SMSCs) have yielded encouraging results. SMSCs play an important role in joint homeostasis and natural cartilage repair (McIntyre, Jones, Han, & Vangsness, 2017; Richter, Zech, & Andreas, 2017). Furthermore, mesenchymal stem cells (MSCs) derived from synovial joint tissues show superior chondrogenic potential compared with those derived from other tissues (Jones & Pei, 2012; Yoshimura et al., 2007). These findings suggest that SMSCs may be a promising source of cells for cell‐based therapy in TMD.

Although no specific single cause has been identified to date for TMD, inflammatory cytokines are recognized as important factors contributing to the pathogenesis of TMD. The inflammatory milieu may also determine the biological fate of MSCs (Najar et al., 2016). The immunomodulatory effects and differentiation potential of MSCs can be influenced by cytokines, indicating that they have a wide variety of functions (Heldens et al., 2012; Ousema et al., 2012; L. Wang et al., 2017). Interleukin 1β (IL‐1β) is one of the most important proinflammatory cytokines involved in TMD. It has been reported that the expression of IL‐1β is substantially promoted in the synovial fluid of patients with TMD (Kellesarian et al., 2016). In our previous study, we found that IL‐1β impedes the chondrogenic differentiation of synovial fluid MSCs by promoting IL‐6 expression in the temporomandibular joint (TMJ; Liu et al., 2017). IL‐6 and IL‐8, which are also important factors contributing to the pathogenesis of TMD, can be highly upregulated by IL‐1β in MSCs in the synovial fluid (Kellesarian et al., 2016). Therefore, inflammatory activation of MSCs in the synovial fluid by IL‐1β is a major obstacle for cell‐based therapies. Our previous study revealed that SMSCs are an important source of MSCs in the synovial fluid (Sun et al., 2014). Thus, we hypothesized that SMSCs may also induce a similar inflammatory activation in the synovial fluid after IL‐1β stimulation. However, the underlying regulatory mechanism for inflammatory activation of SMSCs induced by IL‐1β is not fully understood.

Epigenetics is an important regulator of gene expression that does not involve any changes in the DNA sequence and is associated with the pathogenesis of many different human diseases. The acetylation and deacetylation of histone are one of the most important mechanisms of epigenetic regulation. Histone acetyltransferases (HATs) modify the N‐terminal domains of histones to transform chromatin into a hyperacetylated state, which reduces the condensation of chromatin and allows transcription factors to bind to the transcriptional regulatory region of the target gene. On the contrary, deacetylation is catalyzed by histone deacetylases (HDACs) and increases the affinity between the DNA and histones, leading to chromatin condensation and repression of gene transcription (Wolffe, 1996). According to the DNA sequence similarity and activities of the HDAC proteins, HDACs are classified into four groups: Class I (HDAC1, HDAC2, HDAC3, and HDAC8); II (HDAC4, HDAC5, HDAC6, HDAC7, HDAC9, and HDAC10), which includes class IIa comprised of HDAC4, HDAC5, HDAC7, and HDAC9 and class IIb comprised of HDAC6 and HDAC10; III (SIRT1–SIRT7); and IV (HDAC11; X. J. Yang & Seto, 2008). The change in HDAC/HAT activity has been identified as a critical factor contributing to chronic inflammatory diseases. Moreover, studies have demonstrated that using HDAC inhibitors (HDACi) can regulate inflammatory responses. The studies verified that the HDACi trichostatin A and ITF2357 suppressed IL‐1β‐induced IL‐6 production in fibroblast‐like synoviocytes in rheumatoid arthritis (Grabiec, Korchynskyi, Tak, & Reedquist, 2012). Therefore, HDACi have been suggested as new treatment options for inflammatory diseases, including inflammatory joint diseases (Grabiec, Tak, & Reedquist, 2008; Shuttleworth, Bailey, & Townsend, 2010). It has proven that a specific HDACi can prevent ECM degradation in human osteoarthritis (OA) chondrocytes (Carpio & Westendorf, 2016; Im & Choi, 2013). Although the broad‐acting HDACi, which mainly target classes I and II HDAC enzymes, were found to be beneficial in treating many diseases and produce lesser side effects, research continues an effort to target specific HDAC subtypes to improve their efficacy and reduce the risk of side effects (Cantley & Haynes, 2013; Vojinovic et al., 2011). Thus, it is important to identify the precise HDAC isoforms playing significant roles in the initiation and development of inflammatory diseases.

In the current study, we aimed to determine the cytokine profile in SMSCs stimulated by IL‐1β. We hypothesized that HDACs would exert a significant effect on regulating the IL‐1β‐induced secretion of IL‐6 and IL‐8 in SMSCs in the TMJ. We also attempted to identify the specific HDAC subtype that primarily contributed to these processes.

2. MATERIALS AND METHODS

2.1. Collection of synovial membrane samples

From January 2017 to March 2018, 13 patients with TMJ OA undergoing open TMJ surgery each donated a synovium sample. The patients had no systemic diseases, were aged from 15 to 64 years, and included 11 females and two males. The study was approved by the Institutional Ethics Board of the Hospital of Stomatology, Sun Yat‐sen University (Guangzhou, China; ERC‐2017‐8). All patients provided written informed consent before surgery.

2.2. SMSCs culture

The synovial specimens from patients suffering from TMJ OA were obtained during the surgery. The specimens were washed three times in phosphate‐buffered saline (PBS; Gibco, Carlsbad, CA, USA), cut into small pieces, and placed in a petri dish. Next, 1 ml (4 mg/ml) type I collagenase (Sigma‐Aldrich, St. Louis, MO) was added, and the mixture was incubated for 2.5 hr at 37°C. After incubation, the mixture was centrifuged at 350g for 5 min. The precipitate was suspended in complete culture medium composed of 10% fetal bovine serum (FBS; Gibco, Australia), ×1 GultaMax (Gibco, Carlsbad, CA), and α‐minimum essential medium (α‐MEM; Gibco, Carlsbad, CA) and the suspended cells were seeded into Petri dishes and cultured at 37°C and 5% CO₂. When the cells were grown to 80% confluency, they were passaged at a ratio of 1:3. During the passage, the culture medium was removed and then the cells in the Petri dishes were washed with PBS. TrypLE (Gibco, Carlsbad, CA) was added to the dishes and incubated for 2 min to harvest the adherent cells. A triple volume of PBS was added to terminate the digestion and the mixture was then transferred to a 15‐ml centrifuge tube. After centrifugation, the cells were resuspended in culture medium and seeded into new Petri dishes. In the following experiments, the cells derived from pools of three patients were mixed to produce primary cultures. A total of four primary cultures produced from 12 of the 13 original samples were used in this study. Cells from passage 3–5 were used in the experiments.

2.3. Preparation of reagents

IL‐1β (PeproTech, Rocky Hill, NJ) was dissolved in double distilled water containing 5% trehalose to a final concentration of 10 μg/ml and stored at −20°C until use. Upon use, the stock solution of IL‐1β was diluted to 1 or 10 ng/ml with a culture medium. Inhibitors BAY117082, SAHA, and LBH589 (Santa Cruz Biotechnology, Dallas, TX), MC1568 (Absin Bioscience Inc., Shanghai, China), CI994 (Selleck, Houston, TX), and FK228 (Selleck) were dissolved in dimethyl sulfoxide (DMSO; MP Biomedicals, Santa Ana, CA) and diluted in culture medium to the appropriate concentrations.

2.4. Identification of surface antigens

For each sample to be analyzed, 2 × 10⁵ cells were harvested and mixed with antibodies specific for different surface markers including CD90, CD105, CD44, CD45, CD34, CD11b, CD19, and HLA‐DR (each purchased from BD Biosciences, San Jose, CA; 1:20), CD73 (purchased from Miltenyi Biotec, Bergisch Gladbach, Germany; 1:11), and isotype control antibody (1:20; BD Biosciences). The tubes were incubated in the dark for 30 min at 4°C, then centrifuged and resuspended in 300 µl PBS. Antibody labeling was detected using an FC500 Flow Cytometer (Beckman Coulter, Atlanta, GA) and analyzed using CXP Software (Beckman Coulter).

2.5. Osteogenic induction

SMSCs (5,000 cells/well) were seeded into six‐well plates and grown to 80% confluency. Osteogenic differentiation medium was added to the plates and incubated for 4 weeks. The osteogenic differentiation medium was composed of DMEM with high glucose (Gibco, Carlsbad, CA) containing 10% FBS, 10 mM sodium β‐glycerophosphate (Santa Cruz Biotechnology), 100 nM dexamethasone (MP Biomedicals) and 50 µg/L ascorbic acid‐2‐phosphate (Wako, Osaka, Japan). The medium was changed every 3 days. After 4 weeks of osteogenic induction, the cells in the plates were fixed in 4% paraformaldehyde (Biosharp, Hefei, China) for 10 min and then stained with alizarin red solution (Cyagen, Santa Clara, CA) for 15 min. The expression levels of RUNX2 and OCN messenger RNA (mRNA) were quantitated by reverse‐transcription quantitative polymerase chain reaction (RT‐qPCR), as detailed below.

2.6. Adipogenic induction

The adipogenic differentiation medium consisted of DMEM with high glucose containing 10% FBS, 0.5 mM isobutylmethylxanthine (Sigma‐Aldrich), 10 mg/ml insulin (Telenbiotech, Guangzhou, China), 1 mM dexamethasone (MP Biomedicals), and 200 mM indomethacin (Sigma‐Aldrich). When cell monolayers were grown to 80% confluency, the adipogenic induction medium was added to the plates and incubated for 4 weeks. The medium was changed every 2–3 days. Cells were fixed in 4% paraformaldehyde for 10 min and stained with Sudan Black (Cyagen) for 5 min to assess adipogenesis induction. The cells were observed using an inverted phase contrast microscope (Axiovert 40; Carl Zeiss AG, Oberkochen, Germany). The expression levels of LPL and PPARG2 were determined by RT‐qPCR.

2.7. Chondrogenic induction

SMSCs (3 × 10⁵) were harvested with TrypLE and placed in each 15‐ml centrifuge tube. After centrifugation at 450g for 10 min, the digestion mixture was replaced with 500 μl chondrogenetic induction medium (Stem Chondro Diff Kit; Invitrogen, Waltham, MA). The tubes were then incubated at 37°C with the caps loosened. The medium was replaced with fresh medium every 3–4 days. After 4 weeks of chondrogenic induction, the cartilage pellets were fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin for histological sectioning. The sections were rehydrated using ethanol gradient, and stained with Safranin O (Cyagen) or using hematoxylin and eosin (H&E) Kit (Nanjing Jiancheng Tech Inc., Nanjing, China) following the manufacturers' instructions. The sections for immunohistochemical staining were rehydrated using an ethanol gradient and then sequentially incubated with 0.3% Triton and 3% H₂O₂ solutions for 10 and 8 min, respectively. Antigen retrieval was performed in 10 mM sodium citrate buffer using a microwave oven. The cells were immunostained by incubation overnight with a specific antibody for COL2 (1:100; Cell Signaling Technology, Danvers, MA). A 3,3′‐diaminobenzidine Kit (CW Biotech, Beijing, China) was used as a chromogen for visualization of antibody labeling. The stained sections were rinsed in distilled water, mounted onto microscope slides, and evaluated by light microscopy (Axioskop40; Carl Zeiss AG). Cartilage pellets (four to six) were combined and used for total RNA extraction. The expression of the COL2 and SOX9 at gene level was detected by RT‐qPCR. The induced group samples were compared with the control group samples, which were cultured with complete culture medium instead of chondrogenetic induction medium.

2.8. Immunofluorescence cell assay

SMSCs (1 × 10⁴) were seeded on a confocal dish and incubated overnight at 37°C and 5% CO₂. The cells were then fixed in 4% paraformaldehyde for 15 min and subsequently, washed three times in PBS and treated with 0.3% Triton X‐100 (MP Biomedicals) for 15 min. The cells were then incubated in 5% nonfat milk for 1 hr at room temperature. Nuclear factor‐κB (NF‐κB) P65 rabbit monoclonal antibody (1:100; Cell Signaling Technology) was added to the treated dishes and incubated overnight at 4°C. As a negative control, PBS instead of antibody was added to one dish. The DyLight 594 AffiniPure Goat Anti‐Rabbit IgG (1:100; EarthOx, Millbrae, CA) was added as a secondary antibody and incubated in the dark for 1 hr. The cells were then washed three times with PBS and 4,6‐diamidino‐2‐phenylindole (Beyotime, Shanghai, China) was used for 5 min to stain the nuclei. The cells were observed by laser confocal microscopy (Carl Zeiss AG).

2.9. Cell proliferation assay

SMSCs (5,000 cells/well) were seeded into 96‐well plates and incubated overnight. SAHA and LBH589 were diluted to concentrations of 0.5, 1, 3, 5, 10, 20, 50, or 100 μM in 0.2% DMSO. MC1568 was diluted to concentrations of 0.01, 0.1, 0.5, 1, 3, 5, 10, 20, or 50 μM. CI994 was diluted to concentrations of 1, 2, 5, 10, 20, 50, or 100 μM. FK228 was diluted to concentrations of 1, 2, 5, 10, 20, 50, or 100 nM. The SMSCs were treated with SAHA, LBH589, CI994, or FK228 at the different concentrations for 48 hr. The supernatant was then removed and the cells in each well were washed with DMEM. Next, 100 μl of 10% Cell Counting Kit (CCK)‐8 (Dojindo, Kumamoto, Japan) was added to each well and incubated for 2 hr. Five wells without cells were used as background controls. The optical density (OD) of the experimental and control wells were measured at an absorbance of 450 nm. The background OD was subtracted from each sample before calculations were made. The cell proliferation rate (%) was calculated as the average OD of the experimental wells divided by the average OD of control wells × 100.

2.10. Transfection of short interfering RNA (siRNA)

A total of 30,000 SMSCs/ml were plated into six‐well or 24‐well plates and grown to 70–80% confluency. The medium was replaced before transfection. The siRNAs were dissolved in RNase‐free water (Invitrogen) to a concentration of 20 mM siRNA and stored at −20°C until use. Next, 5 μl of 20 mM siHDAC10 (Invitrogen) or Stealth siRNA Negative Control (NC, 12935400; Invitrogen) was mixed with 120 μl Opti‐MEN medium (Invitrogen). Similarly, 5 μl of Lipofectamine RNAiMAX (Invitrogen) was added to 120 μl Opti‐MEM. The two media containing the siRNA and the lipofectamine were mixed and incubated for 10 min. After the incubation, 250 μl aliquots were added to each well of the six‐well plates. In the 24‐well plates, 2.5 μl siRNA and Lipofectamine RNAiMAX were separately added to 22.5 μl Opti‐MEM, mixed and incubated for 5 min. Thereafter, 50 μl of the mixtures were added to each well (an equal volume of culture medium was removed before from each well before adding the mixture). The final concentration of siRNA was 50 nM. The following sequences were used for each siRNA: siHDAC10‐1, 5´‐GGUGGUUUCCUGAGCUGCAUCUUGG‐3´; siHDAC10‐2, 5´‐CGGAGUCAGUGUGCAUGACAGUACA‐3´. Cells were transfected for 48 hr before further experiments were performed.

2.11. Construction and transduction of a lentiviral vector

Lentiviruses expressing the green fluorescent protein (GFP) empty vector and GFP vector overexpressing HDAC10 (LV‐HDAC10; NM_032019), were constructed by GeneChem Co., Ltd. (Shanghai, China). SMSCs (100,000 cells/well) were seeded in a six‐well plate and incubated overnight. The virus solution was diluted to 1 × 10⁸ TU/ml with enhanced transfection solution and 20 µl was added to each well. Polybrene (5 mg/ml) was added to the culture medium for 12 hr to induce transduction. The medium was replaced with complete culture medium and cultured for 48 hr and then the transformed cells were selected using puromycin (1 μg/ml). The selected cells were sorted by flow cytometry to obtain a GFP‐positive rate of at least 95%.

2.12. RNA extraction and RT‐qPCR

Total RNA was extracted using TRIzol (Invitrogen). The synovial membrane tissue from the 13 TMJ OA patients or from the rats or chondrogenic differentiation pellets were snap frozen and pulverized in liquid nitrogen before being added to the TRIzol reagent. Reverse transcription was performed to generate complementary DNA (cDNA) using 1 μg of total RNA as a template with the SuperScript™ Reverse‐Transcription System (Takara Biotechnology, Dalian, China), according to the manufacturer's instructions. Relative mRNA expression levels were detected using an SYBR Green qPCR Kit (Roche, Basel, Switzerland). The relative expression of the target genes was calculated according to the $2^{-\mathrm{\Delta }\mathrm{\Delta }{C}_{\mathrm{t}}}$ method by normalizing to glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) as an endogenous control. The primer sequences for the target gene are listed in Table 1. The ΔΔC _t of the synovial membrane tissue from the 13 TMJ OA patients was calculated using the following formula: $C_{{\mathrm{t}}_{\mathrm{t}\mathrm{}\text{arg}\mathrm{e}\mathrm{t}\mathrm{}\mathrm{}\mathrm{g}\mathrm{e}\mathrm{n}\mathrm{e}}}-C_{{\mathrm{t}}_{\mathrm{G}\mathrm{A}\mathrm{D}\mathrm{P}\mathrm{H}}}\mathrm{.}$

Table 1.

Primer sequences for polymerase chain reaction

Genes	Forward	Reverse
human GAPDH	GACAGTCAGCCGCATCTTCT	TTAAAAGCAGCCCTGGTGAC
human RUNX2	TCAACGATCTGAGATTTGTGGG	GGGGAGGATTTGTGAAGACGG
human OCN	CCACCGAGACACCATGAGAG	TCAGCCAACTCGTCACAGTC
human PPARG2	GCAAACCCCTATTCCATGCTG	CACGGAGCTGATCCCAAAGT
human LPL	CAAGAGTGAGTGAACAAC	AATTATGCTGAAGGACAAC
human COL2	GGCAATAGCAGGTTCACGTACA	CGATAACAGTCTTGCCCCACTT
human SOX9	ACACACAGCTCACTCGACCTTG	AGGGAATTCTGGTTGGTCCTCT
human MMP‐13	GACTGGTAATGGCATCAAGGGA	CACCGGCAAAAGCCACTTTA
human IL‐1β	CCAGGGACAGGATATGGAGCA	TTCAACACGCAGGACAGGTACAG
human IL‐6	ACTCACCTCTTCAGAACGAATTG	CCATCTTTGGAAGGTTCAGGTTG
human IL‐8	ACTGAGAGTGATTGAGAGTGGAC	AACCCTCTGCACCCAGTTTT
human HDAC10	ACTGCACTTGGGAAGCTCCTGTA	GCCTCTCCGAACAGCCACAT
rat GAPDH	AGACAGCCGCATCTTCTTGT	CTCGTGGTTCACACCCATCA
rat IL‐1β	TTGAGTCTGCACAGTTCCCC	GTCCTGGGGAAGGCATTAGG
rat IL‐6	CCGGAGAGGAGACTTCACAG	CAGAATTGCCATTGCAACAAC
rat IL‐8	TGGCCAGAGAAAGAAGTGCC	TGTCTTCAATCCATCCCAGAGC
rat MMP‐1	GCCATTACTCACAACAATCCTCG	AACACAATATCACCTTCCTCCTCAA
rat MMP‐3	TGGACCAGGGACCAATGGA	GGCCAAGTTCATGAGCAGCA
rat MMP‐13	AGACCTCCAGTTTGCAGAGCGCGT	ACCCACATCAGGAA CCCCGCAT
rat TIMP‐1	CTGGCATCCTCTTGTTGCTATCATT	CTTATAACCAGGTCCGAGTTGCAGA
rat TIMP‐2	GGGACACGCTTAGCATCACC	CATCCAGAGGCACTCATCCG
rat ADAMST‐4	CAGCCATACCCAGAGCGTCAC	GCATCCGAAACCCTGTCAACT
rat ADAMST‐5	GTTTCGAGGTGCGGGGTTATT	TCTGCCTGCAAGGGAAATGTG
rat NF‐κB	CAAGCACTGTGAGGACGGCATA	GCACCAGAAGTCCAGGGTTATAGC
rat HDAC10	GTG CCC TGG AGT CTA TC	CCA AGG CAA CAG CTA TG

2.13. Protein collection and western blot analysis

Protein in the cytoplasm and nucleus were extracted following the instructions of PARIS Kit Protein and RNA Isolation System (Invitrogen). For total protein, RIPA buffer (Beyotime) was used to lyse the cells. The concentration of each protein sample was measured using a BCA Protein Assay Kit (Beyotime) following the instructions included with the kit. Next, 40 μg protein was mixed with loading buffer and then denatured by heating for 10 min at 99°C. Protein samples were electrophoretically separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (Beyotime). The protein was then transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Boston, MA) by applying a 200‐mA steady current. The PVDF membranes were blocked with 5% nonfat milk (BD Biosciences) for 1 hr at room temperature and then incubated with phospho‐NF‐κB P65 (p‐P65), NF‐κB P65, phospho‐IKB‐α (p‐IKB‐α), IKB‐α, β‐actin, or GADPH rabbit monoclonal antibodies (1:1,000; Cell Signaling Technology), or HDAC10 rabbit monoclonal antibody (1:1,000; Abcam, Cambridge, UK) overnight at 4°C. The membranes were then incubated with IgG‐HRP (1:5,000; Cell Signaling Technology) as a secondary antibody for 1 hr. Specific protein bands were visualized using an ECL Kit (Millipore).

2.14. Cytometric bead array (CBA) assay

SMSCs (30,000 cells/well) were seeded into each well of 24‐well plates and treated according to the experiment grouping. The cell culture supernatant was collected and centrifuged at 3,000g for 10 min. The protein levels of IL‐6 and IL‐8 in the culture medium were determined using human IL‐6 and IL‐8 CBA Kits (BD Biosciences). Samples were prepared following the manufacturer's instruction. Data were measured using a CytoFLEX Flow Cytometer (Beckman Coulter).

2.15. Dual‐Glo luciferase assay

SMSCs (10,000 cells/well) were seeded into each well of 24‐well plates and grown to 80% confluency. The culture medium was replaced with 400 μl Opti‐MEM medium (Gibco, Carlsbad, CA). A total of 500 ng pGL4.15 [luc2P/Hygro] (Promega, Madison, WI) or pGL4.32 [luc2P/NF‐κB‐RE/Hygro] vector (Promega) and 50 ng pGL4.74 [hRluc/TK] vector (Promega) were dissolved in Opti‐MEM medium and 25 μl medium was prepared for each well. A total of 1.1 μl Lipofectamine 2000 (Gibco, Carlsbad, CA) was dissolved in Opti‐MEM medium and 25 μl medium was prepared for each well. The two media containing vector and Lipofectamine 2000 were mixed and incubated for 10 min. A total of 50 μl mixture was added to each well and incubated for 5 hr. The medium was replaced with complete culture medium. The cells were then treated for 1 hr with 0 or 10 ng/mL IL‐1β, or with 10 ng/ml IL‐1β preincubated with SAHA, or LBH568.

For this assay, a Dual‐Glo Luciferase Assay System (Promega) was used. All the reagents were heated in a water bath at room temperature for several hours before use. The contents of one bottle of Dual‐Glo luciferase buffer were transferred to one bottle of Dual‐Glo luciferase substrate and 1.5 ml dispensed into each EP tube and stored at −70°C. The amount of Dual‐Glo Stop and Glo reagent needed to perform the desired experiments was calculated. The reaction mixtures were mixed by inversion until the substrate was thoroughly dissolved. A total of 250 μl culture medium was removed and 250 μl Dual‐Glo luciferase reagent was added to each well and incubated for 10 min at room temperature to prepare the lysate. A total of 100 μl lysate was added to each well of the 96‐well black plate and firefly luminescence was measured using a microplate reader (Promega). A total of 50 μl Dual‐Glo stop reagent was added and incubated for 10 min following the measurement of Renilla luminescence.

2.16. Establishment of OA model of TMJ in rat

Twenty male 8‐week‐old Sprague–Dawley rats (weight 300–400 g) were housed in a temperature‐controlled room with a 12 hr light–12 hr dark cycle. The study was approved by the Animal Ethical and Welfare Committee of Sun Yat‐sen University. The rats were randomly divided into two groups: 10 for the normal control group (NC) and 10 for the OA group. A total of 200 μl 10 mg/ml collagenase (Sigma‐Aldrich) was injected bilaterally into the TMJ articular upper cavity of the OA group of rats. No treatment was given to the NC group. Four weeks after injection, the rats were sacrificed under deep anesthesia by an overdose injection of 10% chloral hydrate (Shanghai Hushi, Shanghai, China). TMJs, including the synovium, were dissected and collected. Five OA synovium samples and five normal synovium samples were collected, mixed, and stored in TRIzol for RNA extraction. Images of the TMJ samples were acquired using a stereoscopic microscope. The specimens were then fixed in 4% paraformaldehyde and samples from each group were scanned by micro‐computed tomography (CT, µ CT50; Scanco, Brüttisellen, Switzerland). The expression of IL‐1β, IL‐6, IL‐8, HDAC10, MMP‐1, MMP‐3, MMP‐13, TIMP1, TIMP2, ADAMTS4, and ADAMTS5 at the gene level were determined by RT‐qPCR.

2.17. Statistical analysis

For each assay, three technical and biological replicates were performed. Data were expressed as the mean ± SD. Statistical significance between two groups was determined using Student's t test. Analysis of variance was used to determine the significance between more than two groups. The correlation between the expression of IL‐1β and HDAC1–10 was analyzed using Spearman's rank test. p < 0.05 was considered statistically significant.

3. RESULTS

3.1. Fibroblast‐like synoviocytes in the TMJ met criteria minimum for MSCs

After incubation for 1 week, migration of fibroblast‐like cells from the synovial membrane tissue was observed. After passaging, the cells adopted a homogeneous fibroblast‐like morphology (Figure 1a), could be induced to differentiate into osteogenic, adipogenic, and chondrogenic lineages in vitro. After 4 weeks of culturing in osteogenic induction medium, calcium deposits were observed, which were confirmed by Alizarin red staining (Figure 1b). The expression of RUNX2 (Figure 1g), OCN (Figure 1h) in the osteogenically induced cells was 1,160 ± 93.47 and 12.30 ± 0.32‐fold higher (p < 0.0001), respectively, compared with that in the control group. After culturing in adipogenic induction medium for 28 days, the SMSCs developed into Sudan Black B‐positive, lipid‐laden fat cells (Figure 1c). The expression of PPARG2 (Figure 1j), LPL (Figure 1i) in the adipogenically induced cells was 35.32 ± 2.74 and 9,240 ± 460.50‐fold higher (p < 0.0001), respectively, compared with that in the control group. After 28 days of culturing in chondrogenic induction medium, the cells formed cartilage nodules (Supporting Information Figure S1). Histological sections of the cartilage nodules stained positively for COL2 (Figure 1d), with Safranin O (Figure 1e), a typical chondrocyte morphology was observed after H&E staining (Figure 1f). The expression of COL2 (Figure 1k), SOX9 (Figure 1l) in the cartilage nodules was 136,890 ± 8,635 and 2.736 ± 0.22‐fold higher (p < 0.0001), respectively, compared with that in the control group. Flow cytometry showed that over 95% of the fibroblast‐like cells from the synovial membrane tissue expressed CD90, CD44, CD105, and CD73. In contrast, <2% of the cells expressed CD45, CD34, CD11b, CD19, and HLA‐DR (Figure 1m).

Figure 1.

Characteristics of fibroblast‐like synoviocytes in the TMJ. (a) Typical morphology of SMSCs. (b) Alizarin red staining of SMSCs after 4‐week postinduction of osteogenic differentiation. (c) Sudan Black staining of SMSCs 4‐week postinduction adipogenic differentiation. (d) Immunohistochemical staining of COL2 in cartilage nodule histological sections after chondrogenic induction. (e, f) Safranin O (e) and H&E (f) staining of cartilage nodule histological sections. (g, h) Expression of RUNX2 (g) and OCN (h) was upregulated after osteogenic induction. (i, j) Expression of LPL (i) and PPARG (j) was upregulated after adipogenic induction. (k, l) Expression of COL2 (k) and SOX9 (l) was upregulated after chondrogenic induction. (m) Flow cytometry assay of SMSCs surface markers. Red: target marker; black: isotype control group. *p < 0.05. (a–f) Scale bars = 100 µm. GAPDH: glyceraldehyde 3‐phosphate dehydrogenase; H&E: hematoxylin and eosin; SMSC: synovium‐derived mesenchymal stem cell; TMJ: temporomandibular joint [Color figure can be viewed at wileyonlinelibrary.com]

3.2. IL‐6 and IL‐8 expression in SMSCs was upregulated by IL‐1β through activating the NF‐κB pathway

Inflammatory cytokines, such as IL‐1β, IL‐6, IL‐8, tumor necrosis factor α (TNF‐α), IL‐10, and IL‐12p, are central causative factors of TMD. We found that IL‐6 and IL‐8 expression increased significantly in SMSCs after stimulation with IL‐1β, whereas the expression of TNF‐α, IL‐10, and IL‐12p did not markedly change (Supporting Information Figure S2). The expression of IL‐6 and IL‐8 peaked at 6 hr after stimulation with IL‐1β and then gradually decreased over time (Figure 2a). The concentrations of IL‐6 and IL‐8 in the culture media gradually increased over time following stimulation of the cells with IL‐1β (Figure 2b). Western blot analysis revealed that p‐P65 expression increased substantially in response to IL‐1β compared with that in the control group (Figure 2c,d). Compared with that in the control group, the levels of the NF‐κB P65 subunit markedly increased in the nuclei of SMSCs after stimulation with IL‐1β, which was confirmed by immunofluorescent staining (Figure 2e). Treatment with IκB kinase inhibitor BAY117082 considerably reduced IL‐6 and IL‐8 expression levels compared with that in the IL‐1β stimulation group at both the gene and protein levels (Figure 2f–h).

Figure 2.

Cytokine expression and the regulatory mechanism underlying IL‐1β‐mediated inflammation in SMSCs. (a) Gene expression levels of IL‐6 and IL‐8 were measured by RT‐qPCR after SMSCs were stimulated by 10 ng/ml IL‐1β for 2, 6, 12, and 24 hr and 1 ng/ml IL‐1β for 2, 12, and 24 hr. *p < 0.05 compared with the control group; ^# p < 0.05 for the 10 ng/ml stimulation groups compared with the 1 ng/ml stimulation group at the same time point. (b) Activity of IL‐6 and IL‐8 protein in the culture supernatant determined by CBA assays after SMSCs were stimulated by 10 ng/ml IL‐1β for 2, 6, 12, and 24 hr. *p < 0.05 compared with the control group; ^# p < 0.05 compared with the different time‐point groups. (c) Western blot showing p‐P65 expression in SMSCs after IL‐1β stimulation. Each group is shown in triplicate. (d) The bands in Figure 2c were quantified and the results from the IL‐1β group was compared with the control group. *p < 0.05 compared with the control group. (e) Immunofluorescence assay showing the cellular localization of P65 in SMSCs 6‐hr poststimulation with 10 ng/ml IL‐1β. Scale bars = 100 µm. The gene expression of IL‐6 and IL‐8 (f, g) and protein expression (h) were measured after SMSCs were pretreated for 1 hr with 3 µM BAY117082, 10 µM SAHA, or 3 µM LBH589 and then stimulated with 10 ng/ml IL‐1β for 6 hr. *p < 0.05 compared with the control group. ^# p < 0.05 compared with the IL‐1β group. CBA: cytometric bead array; GAPDH: glyceraldehyde 3‐phosphate dehydrogenase; IL: interleukin; RT‐qPCR: reverse‐transcription quantitative polymerase chain reaction; SMSC: synovium‐derived mesenchymal stem cell [Color figure can be viewed at wileyonlinelibrary.com]

3.3. HDAC inhibitors reduced IL‐1β‐mediated IL‐6 and IL‐8 expression and NF‐κB pathway activation

Treatment with the type I and type II HDAC inhibitors SAHA or LBH589, respectively, considerably reduced IL‐6 and IL‐8 expression at both the gene and protein level compared with that in the IL‐1β stimulated group (Figure 2f–h). The concentrations of the drugs were confirmed to have low cytotoxicity based on results of the cell proliferation assays (Supporting Information Figure S3). Immunofluorescence staining showed that the levels of the NF‐κB P65 subunit were markedly reduced in the nuclei of SMSCs after cotreatment with SAHA or LBH589 compared with that in the IL‐1β‐treated group (Figure 3a). IL‐1β activated the NF‐κB pathway, whereas SAHA and LBH589 attenuated the activation mediated by IL‐1β, which was confirmed by Dual‐Glo luciferase assays (Figure 3b). Western blot analysis showed that the IL‐1β‐mediated increases in p‐P65 and P65 could be reversed by LBH589 treatment (Figure 3c,d). The expression of P65 in the cytoplasm and nucleus increased after treatment with IL‐1β compared with that in the control group, whereas the expression of p‐P65 increased primarily in the nucleus following treatment with IL‐1β. Interestingly, cotreatment of cells with LBH589 and IL‐1β caused increased expression of p‐P65 and P65 in the cytoplasm and reduced expression in the nucleus (Figure 3e, f).

Figure 3.

The effect of SAHA and LBH589 on IL‐1β‐mediated IL‐6 and IL‐8 expression in SMSCs. (a) Immunofluorescence assay showing the cellular localization of P65 after SMSCs were pretreated for 1 hr with 10 µM SAHA or 3 µM LBH589 and then stimulated with 10 ng/ml IL‐1β for 6 hr. (b) Dual‐Glo luciferase assay detected the activation of NF‐κB. (c) Western blot showed p‐P65 and P65 expression in SMSCs after IL‐1β stimulation with or without pretreatment of LBH589. (d) Quantitation of the blot shown in (c). (e) Expression of p‐P65 and P65 in the cytoplasm and nucleus was detected by western blot analysis. (f) Quantitation of the blot shown in (e). *p < 0.05 compared with the control group. ^# p < 0.05 compared with the IL‐1β group. DAPI: 4,6‐diamidino‐2‐phenylindole; IL: interleukin; NF‐κB: nuclear factor‐κB; SMSC: synovium‐derived mesenchymal stem cell [Color figure can be viewed at wileyonlinelibrary.com]

3.4. Type II HDAC inhibitor reduced IL‐1β‐mediated IL‐6 and IL‐8 expression and NF‐κB pathway activation

When type I HDAC inhibitors CI994 and FK228 were added to the culture medium, gene expression of IL‐6 and IL‐8 increased compared with that in the group stimulated with IL‐1β but not treated with an inhibitor (Figure 4a,c). However, when type II HDAC inhibitor MC1568 was added to the culture medium, the expression of IL‐6 and IL‐8 were both reduced compared with those in the IL‐1β only treated group (Figure 4e). The concentrations of these factors were confirmed to have low cytotoxicity based on the results of the cell proliferation assays (Figure 4b,d,f). Western blot analysis showed that the expression of p‐P65 and P65 at the protein level were reduced after cotreatment of 5 µM MC1568 and IL‐1β compared with that in the control group (Figure 4g). Compared with the IL‐β only group, cotreatment with IL‐β and MC1568 increased the expression of p‐P65 and P65 in the cytoplasm but reduced their expression in the nucleus (Figure 4h). Quantitation of the results shown in Figure 4g,h were presented in Supporting Information Figure S4.

Figure 4.

Effect of type I and type II HDAC inhibitors on IL‐1β‐mediated IL‐6 and IL‐8 expression in SMSCs. Effect of CI994 (b), FK228 (d) and MC1568 (f) on cells proliferation was determined by CCK‐8 assay. *Significant difference compared with the control group (p < 0.05). (a) SMSCs were preincubated with 1, 5, 10, or 20 µM CI994 for 1 hr, and then treated with 10 ng/ml IL‐1β for 6 hr. (c) SMSCs were preincubated with 1, 5, 10, or 50 nM FK228 for 1 hr and then treated with 10 ng/ml IL‐1β for 6 hr. (e) SMSCs were preincubated with 1, 3, and 5 µM MC1568 for 1 hr and then treated with 10 ng/ml IL‐1β for 6 hr. RT‐qPCR was used to quantitate the expression of IL‐6 and IL‐8. (g) Western blot showed p‐P65 and P65 expression in SMSCs after pretreatment with 5 µM MC1568 for 1 hr followed by treatment with 10 ng/ml IL‐1β for 6 hr. (h) Expression of p‐P65 and P65 in the cytoplasm and nucleus detected by western blot analysis. *Significant difference compared with the control group (p < 0.05). ^#Statistical difference versus the IL‐1β group (p < 0.05). GAPDH: glyceraldehyde 3‐phosphate dehydrogenase; HDAC: histone deacetylase; IL: interleukin; RT‐qPCR: reverse‐transcription quantitative polymerase chain reaction; SMSC: synovium‐derived mesenchymal stem cell

3.5. HDAC10 positively correlated with IL‐1β both in vivo and in vitro

The expression of IL‐1β and type I and type II HDACs in the synovium samples obtained from 13 TMJ OA patients was evaluated by RT‐qPCR. We then analyzed the correlation between IL‐1β and HDAC1–10. According to the results from Spearman's rank correlation test, HDAC10 was highly positively correlated with IL‐1β in the synovium samples of TMD patients, while only a weak or moderate correlation was noted between IL‐1β and HDAC1–9 (Figure 5a). To further verify the results in vivo, we established the TMJ OA rat model to identify the changes in HDAC10 expression in the synovial membrane of the OA model compared with that in the normal control group and determine its relationship with IL‐1β. After injecting collagenase into the TMJ cavity of the rats, destruction of the articular disc was observed in the OA group using a stereoscopic microscope (Figure 5b). Results of micro‐CT analysis also revealed condyle destruction in the OA group (Figure 5c). The expression the MMP‐1, MMP‐3, MMP‐13, TIMP1, TIMP2, and ADAMTS4 in the OA group were all higher than that expressed in the control group, which suggested inflammation and destruction of the synovium. The above results confirmed that we had successfully established the rat model of TMJ OA.

Figure 5.

HDAC10 expression in vivo and in vitro in SMSCs (a) Correlation analysis for the expression of inflammatory cytokines IL‐1β and HDAC1–10 in the synovial membrane tissue from 13 TMD patients. (b) Evaluation of the condyle articular surface in the normal group and OA group using a stereo microscope. (c) Micro‐CT evaluation of the condyle bone destruction in the normal rat group and OA rat group. (d) The gene expression of IL‐1β, IL‐6, IL‐8, HDAC10, MMP‐1, MMP‐3, MMP‐13, TIMP‐1, TIMP‐2, ADAMST‐4, ADAMST‐5, and NF‐κB in synovial membrane tissue of normal rats and OA rats determined by RT‐qPCR. (e) Gene expression of HDAC10 in vitro measured using RT‐qPCR after SMSCs were treated with 10 ng/ml IL‐1β for 2, 6, 12, or 24 hr and 1 ng/ml IL‐1β for 2, 12, or 24 hr. (f) The protein expression of HDAC in vitro was measured by western blot analysis after SMSCs were treated with 10 ng/ml IL‐1β for 2, 6, 12, or 24 hr. (g) Quantitation of the blots shown in (f). *Significant difference compared with the control group (p < 0.05). CT: computed tomography; GAPDH: glyceraldehyde 3‐phosphate dehydrogenase; HDAC: histone deacetylase; IL: interleukin; MMP: matrix metallopeptidase; NF‐κB: nuclear factor‐κB; OA: osteoarthritis; RT‐qPCR: reverse‐transcription quantitative polymerase chain reaction; SMSC: synovium‐derived mesenchymal stem cell; TMD: temporomandibular joint disorder [Color figure can be viewed at wileyonlinelibrary.com]

The expression levels of inflammatory factors IL‐1β, IL‐6, IL‐8, and HDAC10 in the OA group were higher than those in the normal group (Figure 5d). After the human SMSCs were treated with different concentrations of IL‐1β, the expression of HDAC10 was markedly upregulated at different time points (Figure 5e). According to our western blot analysis results, the protein levels of HDAC10 showed a similar trend (Figure 5f,g).

3.6. Knockdown of HDAC10 inhibited IL‐1β‐induced IL‐6 and IL‐8 expression in SMSCs

We then utilized two valid siRNA sequences for HDAC10 (siHDAC10‐1 and siHDAC10‐2) to knockdown HDAC10 in SMSCs. The knockdown was confirmed by measuring the gene expression (mRNA) and protein levels of HDAC10 after transfection (Figure 6a–c). The siRNAs against HDAC10 successfully inhibited IL‐1β‐induced gene expression and protein secretion of IL‐6 and IL‐8 (Figure 6d–h). As shown in Figure 6j, we successfully constructed SMSCs that overexpressed HDAC10. The overexpression of HDAC10 promoted the IL‐1β‐induced expression of IL‐6 and IL‐8 (Figure 6k,l).

Figure 6.

Effect of siHDAC10 on IL‐1β‐mediated IL‐6 and IL‐8 expression. (a, b) The efficiency of HDAC10 knockdown using two siHDAC10 sequences confirmed by RT‐qPCR (a) and western blot analysis (b). (c) Quantitation of the blot shown (b). ^# p < 0.05 compared with the NC group. SMSCs were transfected with siHDAC10‐1 or siHDAC10‐2 for 48 hr and then incubated with 10 ng/ml IL‐1β for 6 hr. The expression of IL‐6 (d and f) and IL‐8 (e and g) was measured by RT‐qPCR. (h) Protein level of IL‐6 and IL‐8 determined by CBA assay. *p < 0.05 compared with the control group. ^# p < 0.05 compared with the NC + IL‐1β group. (j) Efficiency of overexpression of HDAC10 from lentiviral transfection analyzed by RT‐qPCR. (k, l) Gene expression of IL‐6 and IL‐8 determined by RT‐qPCR. *p < 0.05 compared with the GFP group. ^# p < 0.05 compared with the GFP + IL‐1β group. CBA: cytometric bead array; GAPDH: glyceraldehyde 3‐phosphate dehydrogenase; GFP: green fluorescent protein; HDAC: histone deacetylase; IL: interleukin; NC: negative control; RT‐qPCR: reverse‐transcription quantitative polymerase chain reaction; SMSC: synovium‐derived mesenchymal stem cell

3.7. Knockdown of HDAC10 inhibited IL‐1β‐induced activation of the NF‐κB pathway

Compared with the negative control group transfected with empty vector and stimulated with IL‐1β, IL‐1β‐induced levels of the NF‐κB P65 subunit in the nuclei of SMSCs was lower after transfection of siHDAC10 (siHDAC10‐1 sequences were used to knockdown the expression of HDAC10). This reduction was demonstrated by immunofluorescent staining (Figure 7a). The IL‐1β‐induced expression levels of p‐P65, P65, and p‐IKB‐α were all significantly reduced in SMSCs transfected with siHDAC10‐1 compared with that in the negative control group (Figure 7b,c). The transfection of siHDAC10‐1 combined with IL‐1β stimulation increased the expression of p‐P65 and P65 in the cytoplasm and reduced their expression in the nucleus (Figure 7d,e). In contrast, overexpression of HDAC10 increased the IL‐1β‐stimulated expression of P65 and p‐P65 compared with that in the GFP vector group (Figure 7f). The transfection of LV‐HDAC10 to cause the overexpression of HDAC10 increased the IL‐1β‐induced expression of p‐P65 and P65 in both the cytoplasm and nucleus (Figure 7g). The results of the statistical analysis of the data shown in Figure 7f,g are presented in Supporting Information Figure S5.

Figure 7.

Effect of siHDAC10 on IL‐1β‐mediated NF‐κB pathway activation. Transfection of siHDAC10‐1 sequences was used to knockdown the expression of HDAC10. (a) Immunofluorescence assay showing the cellular localization of P65 in SMSCs after transfection with NC or siHDAC10 for 48 hr and then incubated with 10 ng/ml IL‐1β for 6 hr. (a, b) Scale bars = 50 µm. Western blot of p‐IKB‐α, IKB‐α, p‐P65, P65, and actin expression in SMSCs transfected with HDAC10 siRNA and then stimulated with IL‐1β for 6 hr. (c) Quantitation of the blot shown in (b). (d) Expression of p‐P65 and P65 in the cytoplasm and nucleus detected by western blot analysis. (e) Quantitation of the blot shown in (d). (f) Expression of p‐P65 and P65 detected by western blot analysis after transfection with GFP or LV‐HDAC10 vectors and then incubated with 10 ng/ml IL‐1β for 6 hr. (g) Expression of p‐P65 and P65 in the cytoplasm and nucleus detected by western blot analysis. GFP: green fluorescent protein; HDAC: histone deacetylase; IL: interleukin; NC: negative control; p‐IKB‐α: phospho‐IKB‐α; SiRNA: short interfering RNA; SMSC: synovium‐derived mesenchymal stem cell [Color figure can be viewed at wileyonlinelibrary.com]

4. DISCUSSION

The synovium plays an essential regulatory role in maintaining the physiological balance of the TMJ. One of the most important regulatory mechanisms is that of the synovium tissue serving as the major source for MSCs in the synovial fluid. In the current study, we found that fibroblast‐like synoviocytes derived from the TMJ synovium tissues met the minimal criteria needed for consideration as MSCs without immune isolation. In addition, our previous study found that the intima of synovium in TMD patients can shed into the synovial fluid in the form of single dissociated cells or as cell masses and these intima‐derived cells exhibit characteristics similar to MSCs that exist in the synovial fluid. The role of MSCs in the synovial fluid is still not fully understood. The belief that they could migrate to the site of damaged condylar cartilage to repair it through their chondrogenic differentiation potential is one of the most widely accepted opinions (Sun et al., 2014).

The immune regulating abilities of MSCs in the synovial fluid also play an important role. Inflammatory cytokines in the TMJ, such as IL‐1β, IL‐6, IL‐8, TNF‐α, IL‐10, and IL‐12p, are strongly associated with the development of TMD; their expression is positively correlated with the severity of joint disease (Kellesarian et al., 2016). Of these inflammatory cytokines, IL‐1β is one of the most prevalent and first cytokines produced in response to excessive loading of the TMJ (Kim et al., 2012). IL‐6 can be produced by lymphocytes, fibroblasts, endothelial cells, and synovial fluid‐derived MSCs and plays a central role in chronic inflammation (de Alcantara et al., 2017). IL‐8, another IL‐1β‐targeting cytokine, is one of the most powerful chemokines for neutrophils and T lymphocytes and promotes the homing and activation of mononuclear cell leukocytes into the synovium (Larsen, Anderson, Appella, Oppenheim, & Matsushima, 1989). IL‐8 can also promote some pathogenic situations, such as the release of oxidative products, chondrocyte apoptosis, production of MMP‐13 by articular chondrocytes, and the loss of proteoglycans resulting in cartilage destruction (Borzi, Mazzetti, Marcu, & Facchini, 2004; Matsukawa et al., 1995). In our previous study, we found that IL‐6 and IL‐8 in response to IL‐1β are primarily produced by the synovial fluid‐derived MSCs (Nishimura, Segami, Kaneyama, Suzuki, & Miyamaru, 2002). Based on the theory that the synovium tissue is the major source for synovial fluid‐derived MSCs, we speculated that SMSCs exhibited a similar immune response with synovial fluid‐derived MSCs when stimulated by IL‐1β. In the current study, we verified that the production of IL‐6 and IL‐8 in SMSCs was greatly increased after IL‐1β stimulation. Our results were consistent with previous studies. It has been reported that synovium‐derived MSCs express higher levels of IL‐6, IL‐8, and MIP2‐α transcripts in the knee joint (Djouad et al., 2005). In the TMJ, upregulation of IL‐6 may inhibit the chondrogenic potential of MSCs in the synovial fluid and may contribute to the upregulation of MMP‐13, which may then influence osteoclastic bone resorption and cartilage repair. All these results were consistent with the idea that SMSCs inflamed by IL‐1β had a negative effect on restoring the physiological balance in the TMJ. The regulatory mechanism underlying the IL‐1β‐induced activation in SMSCs is currently not fully understood. In our previous study, we found that IL‐1β substantially increases the secretion of IL‐6 and IL‐8 in the synovial fluid‐derived MSCs by activating the NF‐κB pathway. In the current study, we demonstrated that the upregulation of IL‐6 and IL‐8 in SMSCs inflamed by IL‐1β also occurred through the activation of the NF‐κB pathway.

Changes in HDAC expression contribute to the initiation and development of joint diseases. Nuclear HDAC activity and HDAC1 expression are significantly higher in rheumatoid arthritis tissues than in OA synovial tissues and are associated with TNF‐α (Kawabata et al., 2010). Ziesché et al. (2013) found that HDAC3 plays a positive role in the transcription of the majority of IL‐1β‐induced human or murine genes involved in the deacetylation of NF‐κB P65. X. Yang et al. (2016) found that microRNA (miR)‐365 is upregulated by cyclic loading, that IL‐1β stimulation in the articular chondrocytes stimulates the transcription factor NF‐κB, and that HDAC4 is a direct target of miR‐365 (X. Yang et al., 2016). In rheumatoid arthritis, synovial class I HDAC expression is associated with an increase in the local expression of TNF, whereas HDAC5 expression is inversely correlated with disease severity. The upregulation of IL‐1β or TNF selectively suppresses HDAC5 expression in synovioblasts obtained from rheumatoid arthritis patients (Angiolilli et al., 2016). In OA patients, HDAC7 induces a major increase in cartilage production (Higashiyama et al., 2010). Therefore, the expression patterns of HDACs seem to vary depending on the specific disease and tissue.

In this study, we found that HDAC inhibitors SAHA and LBH589 reduced the IL‐1β‐mediated expression of IL‐6 and IL‐8 in SMSCs. These HDAC inhibitors, which primarily target class I and class II HDACs, are FDA‐approved anticancer drugs (Duvic et al., 2007; Garnock‐Jones, 2015). Although vorinostat (SAHA, MK0683) is the most well‐known and well‐studied HDAC pan‐inhibitor, its inhibition mechanism upon binding to HDAC is still controversial (Zhou, Wu, & Luo, 2015). In chondrocytes obtained from OA patients, SAHA inhibits the release of nitric oxide, MMP‐1, and MMP‐13 induced by IL‐1β (Makki & Haqqi, 2016). Another study clarified that SAHA and LBH589 enhance the negative regulation of IL‐1β signaling by increasing the expression of microRNA‐146a, which subsequently inhibits IRAK1 and TRAF6 expression, impairing NF‐κB activity, and suppressing the expression of NF‐κB target genes, such as IL‐6, IL‐8, and TNF‐α (J. H. Wang, Shih, Wu, Wang, & Yang, 2013). In the current study, we also found that SAHA and LBH589 impaired NF‐κB activity and suppressed the expression of NF‐κB target genes IL‐6 and IL‐8. The detailed regulatory mechanism in our study was not fully clear, but the results of preliminary experiments suggested that SAHA and LBH589 increased the negative regulation of the NF‐κB pathway by increasing the expression of NKILA (data not shown). These results will be verified and expanded in future studies.

We also further confirmed that in SMSCs class I HDAC inhibitors FK228 and CI994 increased the expression of IL‐1β‐mediated IL‐6 and IL‐8, whereas the class II HDAC inhibitor MC1568 reduced the expression of IL‐6 and IL‐8 and NF‐κB pathway activation. MC1568 is one of the known selective inhibitors of class IIa HDACs, which was reported in 2005 (Mai et al., 2005). MC1568 suppresses overexpression of the IL‐8 gene by inhibiting hyperacetylation of histones 3 and 4 and the expression of c‐Jun in melanoma cells (Venza et al., 2013). HDACi not only inhibits HDAC enzymatic activity, but also affect global acetylation levels by controlling HDAC expression (Dokmanovic et al., 2007). Our results suggested that some upregulated class II HDAC members may contribute to the biological activity of SMSCs.

In our study, the expression of HDAC10 most positively correlated with the expression of IL‐1β, which was consistent with the results from the in vitro cell stimulation experiments and rat OA model. Knockdown of HDAC10 reversed the inflammatory activation of SMSCs mediated by IL‐1β. Conversely, upregulation of HDAC10 promoted the inflammatory activation. These results suggested that HDAC10 contributed to the inflammatory activation of SMSCs mediated by IL‐1β. Therefore, a specific inhibitor against these proteins may be beneficial to stem cell‐based therapies to treat TMD. Furthermore, we found that HDAC10 siRNA reduced IL‐1β‐mediated activation of the NF‐κB pathway, indicating that HDAC10 could regulate IL‐1β‐induced inflammatory activation of SMSCs through the NF‐κB pathway. HDAC10 was first identified and classified as a class II HDAC in 2002 and is widely expressed in human tissues and cultured cells. HDAC10 possesses a unique leucine‐rich domain, which suggests that HDAC10 may have distinct biological functions. Unlike other class II members, HDAC10 is enriched in the cytoplasm, suggesting that HDAC10 may function outside the nucleus (Tong, Liu, Bertos, & Yang, 2002). HDAC10 has a potent repressive ability (Guardiola & Yao, 2002; Kao, Lee, Komarov, Han, & Evans, 2002), whereas other class II HDACs, including HDAC4, HDAC5, HDAC7, and HDAC9, are known to function as transcriptional co‐repressors (Bertos, Wang, & Yang, 2001; Fischle, Kiermer, Dequiedt, & Verdin, 2001). It has been reported that HDAC10 is associated with DNA mismatch repair gene and is a predictor of good prognosis in colon carcinoma. In contrast, it also promotes cell proliferation via AKT phosphorylation in lung cancer (Tao, Yan, Lu, & Chen, 2017). Therefore, we suspect that the biological function of HDAC10 is more complex and may be dependent on the type of tissue in which it is being expressed. The regulatory mechanism underlying how HDAC10 inhibits the activation of the NF‐κB pathway in SMSCs still requires further study.

5. CONCLUSIONS

In this study, we found that IL‐1β markedly upregulated the expression of IL‐6 and IL‐8 in SMSCs by activating the NF‐κB pathway. Furthermore, HDAC inhibitors LBH589, SAHA, and MC1568 were able to alleviate the inflammatory response and the IL‐1β‐mediated NF‐κB pathway activation. We confirmed that the upregulation of HDAC10 contributed to this biological process, indicating that HDAC10 may be an important factor in inducing the pathogenesis of TMD and may be useful as a novel therapeutic target.

ACKNOWLEDGMENT

This study was supported by grants from Guangdong Medical Research Foundation (A2018385), Natural Science Foundation of Guangdong Province (2018A030310329), and National Natural Science Foundation of China (81800996).

AUTHOR CONTRIBUTIONS

W. L. and J. S. carried out the molecular genetic studies and drafted the manuscript. W. Liu and W. Li participated in the statistical analyses of the data. J. J. and F. O. participated in the cell culture and identification. K. S. and Y. Z. collected the synovial membrane tissue specimens from the patients. Z. Z. and Y. S. conceived the study and participated in its design. All authors read and approved the final manuscript.

AVAILABILITY OF DATA AND MATERIAL

All relevant data are within the paper and its supporting information files.

CONFLICTS OF INTEREST

The authors declare that there are no conflicts of interest.

Supporting information

Supporting information