Abnormal expression of long noncoding RNA FGD5-AS1 affects the development of periodontitis through regulating miR-142-3p/SOCS6/NF-κB pathway

Abstract Periodontitis refers to the inflammation of gums and the surrounding structures and caused by a bacterial infection. The infection occurs owing to poor oral hygiene and could destroy the bone and the gum over time if left untreated. The present study identified the involvement of a key long noncoding RNA (lncRNA), i.e. FGD5-AS1, in the pathogenesis of periodontitis by assessing its expression in the gingival tissues of patients diagnosed with chronic periodontitis (CP). Overexpression of FGD5-AS1 in primary human periodontal ligament cells (PDLCs) significantly reduced the lipopolysaccharide (LPS)-induced periodontitis, whereas its suppression aggravated this injury. Moreover, the miR-142-3p was markedly expressed in the gingival samples of patients diagnosed with CP and LPS-induced PDLCs. We found that the FGD5-AS1-mediated reduction in the inflammation was mediated through downside regulation of miR-142-3p, as evident from the upregulation of SOCS6, a target gene of miR-142-3p. Furthermore, the association between FGD5-AS1 and NF-κB pathway was detected. FGD5-AS1 was found to protect against LPS-stimulated PDLC injury through restraining the NF-κB signals. Based on these findings, we conclude that up-regulation of lncRNA FGD5-AS1 could protect against periodontitis via regulating the miR-142-3p/SOCS6/NF-κB signals. Therefore, the FGD5-AS1/miR-142-3p/SOCS6 axis may act as an important indicator in explaining the pathogenesis of periodontitis.


Introduction
Periodontitis is a common dental disease that is present worldwide. It is characterized by inflammatory responses to elevated levels of subgingival pathogens, causing damage to the soft tissues and alveolar bones supporting the teeth [1,2]. It is reported that mild to moderate periodontitis affects up to 50% of adults [3]. However, severe periodontitis may raise the risks of systemic diseases, such as diabetes, rheumatoid arthritis, atherosclerosis and adverse pregnancy outcomes, thereby detrimentally affecting the systemic health of patients [4][5][6]. Therefore, it is imperative to enhance the apprehensiveness of biological mechanisms underlying the pathogenesis of periodontitis to find better therapies.
Recently, endogenous long noncoding RNAs (lncRNAs) with more than 200 nucleotides have possessed extensive attention owing to their important roles in a large-scale biological processes [7][8][9]. Moreover, an increasing number of studies have revealed several lncRNAs to be abnormally expressed in chronic periodontitis (CP), thus playing a crucial role in periodontitis pathogenesis [10,11]. However, the exact roles of differentially expressed key lncRNAs in periodontitis pathogenesis have not been completely understood. Highthroughput RNA sequencing and microarray data on periodontitis have revealed the abnormal expression of lncRNA FGD5-AS1 to be associated with CP [11]. However, these data require verification through physiological experiments including the role of lncRNA FGD5-AS1 and its impact of action.
In the current study, we determined the levels of FGD5-AS1 lncRNA in the gingival tissues of patients diagnosed with CP. Furthermore, to understand the possible mechanism of FGD5-AS1 in periodontitis, periodontal ligament cells (PDLCs) were isolated and treated with lipopolysaccharide (LPS) to induce periodontitis-caused inflammatory damage. The effects of abnormal expression of FGD5-AS1 on LPS-disposed PDLC injury were then evaluated by detecting cell viability, apoptosis, and production of inflammatory cytokines. Several studies have pointed out that lncRNAs serving as competitive endogenous RNAs (ceRNAs) and sponging miRNAs in various diseases [12][13][14]. For example, FGD5-AS1 lncRNA, as a ceRNA, was proved to target miR-142-3p [11]. Based on these data, we investigated whether FGD5-AS1 regulates LPS-induced PDLC injury via sponging miR-142-3p. Moreover, a regulatory pattern between miR-142-3p and suppressor of cytokine signalling 6 (SOCS6) was detected. Furthermore, NF-jB signals have been reported to be the central bridge involved in immune and inflammatory responses [15,16]. We, therefore, studied the association between FGD5-AS1 and NF-jB pathway. We believe that our discoveries would supply a theoretical basis for illuminating the pathogenesis of periodontitis.

Gingival sample collection
Totally, 26 individuals suffered from CP and 20 healthy controls from April 2016 to December 2018 were chosen as subjects in this study. Gingival tissues from cases who were suffered from CP were collected surgically, followed by an original ineffectual nonsurgical scaling and root planning in accordance with the established professional and required oral hygiene of patients. Gingival tissues were also obtained while crown-lengthening routines from healthy individuals with following inclusion criteria: clinical attachment loss < 4 mm, probing depth (PD) < 4 mm, and no alveolar bone destruction at radiographic level. All collected gingival tissue samples were frozen in liquid nitrogen and then stored immediately at -80 C for further experiments. Exclusion criteria included patients with smoking, rheumatoid arthritis, the appearance of any systemic disease, for instance, allergy type 1 and 2 diabetes, and uptake of drugs containing corticosteroids, and antibiotics within 1 month before the surgery. This study obtained the approval of the local ethics committee, and written consent was acquired from all enrolled individuals.

Cell culture and treatment
As previously described, we isolated PDLCs from the explanted healthy periodontal ligament in the middle third of the periodontal membrane root of impacted third molars [17]. PDLCs were then cultured in DMEM (Dulbecco's Modified Eagle Medium; Sangon Biotech; Shanghai, China) including 10% fetal bovine serum, 100,000 mg/mL streptomycin sulfate, 10,000 IU/mL penicillin G, 1% L-glutamine, and 25 mg/mL amphotericin B. Cells were cultivated at 37 C with 5% CO 2 . After 80% confluency was achieved, PDLCs were treated with Porphyromonas gingivalis LPS (100 ng/mL; Sangon Biotech) for different time points.

Real-time qPCR
We used RNAiso Plus (Takara Bio; Tokyo, Japan) to extract the total RNA from both tissues and cells. The purified RNA was used for producing the complementary DNA (cDNA) by reverse-transcribed method using the PrimeScript RT Master Mix (for gene amplification, Perfect Real Time) or Mir-X miRNA First-Strand Synthesis Kit (for miRNA amplification, Takara). RT-qPCR assay was carried out by the SYBR V R Premix Ex Taq TM II kit (Takara) for detecting the expression of target genes. The products were run on an ABI 7500 RT-PCR system (Bio Red). U6 and b-actin were used as references for miRNAs and RNAs, respectively. 2 -DDCt method was chosen for analyzing the relative expression standards of target genes.

MTT assay
We seeded the PDLCs into 96-well plates (1 Â 10 4 cells/well). After various dispose, the obtained PDLCs in each well were mixed with 500 lg/mL of [3-(4, 5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide solution (MTT; Sigma-Aldrich) for further experiment. After 3 h of incubation, dimethyl sulfoxide solution (200 lL, Sigma-Aldrich) was mixed with cells in each well and mixed with the cells to melt the reaction product formazan. Finally, microplate reader (BioRad Model 550; Hercules, CA, USA) was chosen for detecting the absorbance of each well at 570 nm.

Apoptosis assay
PDLCs were harvested and washed in phosphate-buffered saline (PBS) for further use. The washed PDLCs were fixed in 70% ethanol. FITC Annexin V (50 lg/mL RNase A) (Sigma-Aldrich) was chosen for the subsequent staining. After 1 h incubation at room temperature darkly, FACScan system (BioRad) was used for assessment of the percentage of apoptotic cells. FlowJo software was used for analyzing the data.

ELISA assay
Following the manufacturer's protocol, the cytokine production in the supernatants of treated and untreated PDLCs was analyzed using the human IL-6, IL-8, IL-1b and TNF-a ELISA (Duoset kits; R&D Systems, USA). The optical density at 450 nm in each group was generated with 570 nm compensation.

Target prediction and luciferase reporter analysis
Based on the data obtained from TargetScan, SOCS6 was predicted and chosen as the latent target of miR-142-3p. To verify the prediction, we amplified the two kinds of types of (wild-type, WT; mutant, MUT) 3'-UTR regions in the SOCS6 sequences including the binding sequence of the putative miR-142-3p, and then inserted the amplified sequence into the pMIR GLO vector (Promega; Madison, WI, USA) at XhoI and XbaI restriction sites to produce the luciferase reporter plasmids (SOCS6-WT or SOCS6-MUT). PDLCs were then cotransfected with the following vectors, luciferase reporter plasmids (SOCS6-WT or SOCS6-MUT, 1 lg) and miR-142-3p vector (mimics or mimic NC, 50 pmoL) using the Lipofectamine 2000 transfection reagent. At 48 h of posttransfection, dual-luciferase activity assay system (Sangon Biotech) was chosen for the assessment of the luciferase activity of each luciferase reporter plasmid in each group. The Renilla luciferase activity was chosen as the internal control.

Statistical analysis
Three independent experiments were conducted for each assay. All values are calculated and expressed as mean values ± standard deviations (SDs). Significance was analyzed using one-way analysis of variance (ANOVA) or a two-tailed Student's t-test. p < .05 was chosen as of significance.

FGD5-AS1 is poorly expressed in gingival samples of patients diagnosed with CP
Our results indicated that the expression of FGD5-AS1 was dramatically low in the gingival samples of patients diagnosed with CP in comparison with that in the gingival samples obtained from healthy controls (p < .05, Figure 1(A)). These findings indicated that FGD5-AS1 participates in the development of CP.

FGD5-AS1 level is decreased in LPS-stimulated PDLCs
As mentioned in the methods section, PDLCs were treated with LPS for different time points to generate the periodontitis cell model. Relative to the control, LPS dispose significantly reduced the cell viability (Figure 1(B)); induced cell apoptosis (Figure 1(C)); increased the levels of apoptotic proteins including Bax/Bcl-2, cleaved/pro-caspase-3 and cleaved/pro-caspase-9; but accelerated the levels of inflammatory cytokines of TNF-a, IL-6, IL-1b and IL-8 ( Figure  1(D-H)) in a time-dependent way (p < .05). These results implied that LPS induced inflammatory damage in PDLCs. Furthermore, the expression of FGD5-AS1 was also found to decrease significantly in PDLCs following LPS dispose in a time-dependent way (p < 0.05, Figure 1(I)).

SOCS6 is targeted by miR-142-3p
A downside expression of SOCS6 was discovered in periodontal disease [18]. Also, the targeted evidence about miR-142-3p vs. SOCS6 was reported in CD4þ T cells of multiple sclerosis [19] and in nasopharyngeal carcinoma [20]. We discovered that SOCS6 was a possible target of miR-142-3p (http://www.targetscan.org/cgi-bin/targetscan/vert_71/) using TargetScan (version 7.1). Their complementary pairing sequences are shown in Figure 4(A). Moreover, only the luciferase activity of SOCS6-WT was markedly decreased after co-transfection of miR-142-3p mimic relative to mimic NC (p < .05, Figure 4(B)), suggesting SOCS6 is targeted by miR-142-3p. In addition, the data revealed that SOCS6 levels in the miR-142-3p mimic group were markedly reduced relative to that in the mimic NC group, whereas its expression in the miR-142-3p inhibitor group was markedly higher compared with the inhibitor NC group (p < .01, Figure 4(C,D)), confirming SOCS6 is targeted by miR-142-3p.

FGD5-AS1 protects against LPS-induces PDLCs injury through inhibiting NF-jB signals
We further explored the regulatory pattern between FGD5-AS1 and NF-jB signals in LPS-stimulated PDLCs ( Figure 5(J)). Our results revealed that LPS treatment markedly augmented the level of p/t-p65 but cut down the level of p/t-IjBa, indicating the LPS-mediated activation of the NF-jB signals. Overexpression of FGD5-AS1 significantly decreased p/t-p65 and increased p/t-IjBa expression in LPS-stimulated PDLCs, indicating that overexpression of FGD5-AS1 could reverse the LPS-induced activation of NF-jB signal. Moreover, simultaneous overexpression of FGD5-AS1 and miR-142-3p could reverse the influences of overexpression of FGD5-AS1 alone on activating the NF-jB signals; these effects were further changeover after co-transfection of pc-SOCS6 ( Figure 5(K)).

Discussion
Numerous studies have highlighted the regulatory effects of lncRNAs in several human diseases [21][22][23]. However, crucial lncRNAs that are known to participate in the development of diseases comprise only the tip of the iceberg. In this current research, our team explored the expression and effects of FGD5-AS1 in the pathogenesis of periodontitis. Our research revealed that FGD5-AS1 was down-regulated in patients with CP and LPS-disposed PDLCs. Moreover, overexpression of FGD5-AS1 alleviated the LPS-disposed PDLC injury, whereas its suppression aggravated this injury. Furthermore, FGD5-AS1 negatively regulated miR-142-3p, whereas si-FGD5-AS1 alleviated the LPS-induced cell injury via negatively regulating miR-142-3p. In addition, SOCS6 was targeted directly by miR-142-3p, and it was revealed that miR-142-3p alleviated the LPS-induced injury in PDLCs by targeting SOCS6. Another significant discovered result of the present study was that FGD5-AS1 protected against LPS-disposed PDLCs injury through depressing the NF-jB signals. Nevertheless, these data merit further discussion.
There are growing studies that suggest lncRNAs to play a pivotal role in disease pathogenesis by acting as ceRNAs to modulate the gene expression through sponging miRNAs [24,25]. Previous evidence pointed out that miR-142-3p is targeted by FGD5-AS1 as a part of the ceRNA network [11]. Moreover, miR-142-3p is considered as a mediator of inflammation and functions in regulating proinflammatory signalling pathways [26,27]. Previous evidence has confirmed that miR-142-3p gets dysregulated in several inflammatory diseases, such as keratitis [28], psoriasis [29], ulcerative colitis [30], osteoarthritis [31] and dermatitis [32]. Notably, Perri et al. demonstrated that miR-142-3p was markedly added in periodontitis [33], suggesting its potential role in periodontitis development. In the current study, we found that FGD5-AS1 negatively regulated miR-142-3p and its suppression enhanced LPS-disposed cell injury via up-regulation of miR-142-3p. Given the key role of miR-142-3p in inflammation, we believe that FGD5-AS1 could hold back the development of periodontitis via negatively regulating the miR-142-3p.
In addition, SOCS6 has been discovered as targeted by miR-142-3p. SOCS6 is a member of SOCS family proteins that are known as regulators of inflammatory responses induced by bacterial infections [34]. Also, SOCS6 is identified as a cytokine-inducible negative regulator of cytokine signalling [35] and is involved in negative regulation of receptor signalling and inducing apoptosis by targeting mitochondrial proteins [36]. Evidence also points out that SOCS6 is implicated in diverse human cancers [37,38]. In the present research, SOCS6 level was discovered as very low in gingival samples of patients with CP and LPS-disposed PDLCs. Moreover, the influences of inhibition of miR-142-3p on LPS-resulted cell injury in LPS-resulted PDLCs were remarkably reversed after simultaneous inhibition of miR-142-3p and knocking down of SOCS6. Although the role of SOCS6 in periodontitis has not been deciphered, based on the association between SOCS6 and cytokine signalling, we speculate that the si-SOCS6 could accelerate the development of periodontitis and miR-142-3p may prevent periodontitis by targeting SOCS6. Furthermore, the NF-jB signal has been proved to be the central mediator that regulates inflammation, periodontitis, and atherogenesis [39]. Nakajima et al. demonstrated that anandamide could block LPS-triggered activation of NF-jB signalling and consequently regulated hyperinflammatory reactions in periodontitis [40]. Similarly, Dong et al. revealed that osteopontin promoted bone destruction in periapical periodontitis via activating the NF-jB signal [40]. In the present study, overexpression of FGD5-AS1 could changeover the LPS-resulted activation of NF-jB signal. Moreover, simultaneous overexpression of FGD5-AS1 and miR-142-3p could reverse the influences of overexpression of FGD5-AS1 alone on activating NF-jB signal that was further changeover after co-transfection with pc-SOCS6. We thus hypothesize that FGD5-AS1 could prevent the development of periodontitis through regulating the miR-142-3p/SOCS6 axis so as to realize the suppressed effects on activating NF-jB signals.
In conclusion, the results of the current study reported a down-regulation of FGD5-AS1 in patients with CP. Up-regulation of RNA FGD5-AS1 could protect against periodontitis via regulating the miR-142-3p/SOCS6/NF-jB pathway. Thus, the FGD5-AS1/miR-142-3p/SOCS6 axis may provide a promising tactics for the treatment of periodontitis.

Disclosure statement
No potential conflict of interest was reported by the authors.