Apelin enhances IL-1β expression in human synovial fibroblasts by inhibiting miR-144-3p through the PI3K and ERK pathways

Much data suggests intersecting activities between the adipokine apelin (APLN) and the pathologic processes of obesity and osteoarthritis (OA), with APLN modulating cartilage, synovium, bone, and various immune cell activities. The synovium plays an important role in the pathogenesis of OA. We investigated the crosstalk between APLN, a major OA-related adipokine, and interleukin 1 beta (IL-1β), a major proinflammatory cytokine, in human OA synovial fibroblasts (OASFs). We showed that APLN stimulated the synthesis of IL-1β in a concentration- and time-dependent manner, which was mitigated by blockade of the PI3K and ERK pathway. We also showed that APLN inhibited the expression of miRNA-144-3p, which blocks IL-1β transcription; this suppression activity was reversed via blockade of the PI3K and ERK pathway. Moreover, pathologic changes in OA cartilage were rescued when APLN was silenced by shAPLN transfection both in vitro and in vivo. Our evidence is the first to show that APLN stimulates the expression of IL-1β by activating the PI3K and ERK pathway and suppressing downstream expression of miRNA-144-3p in OASFs. We also demonstrate that knockdown of APLN expression by shAPLN transfection ameliorated changes in OA cartilage severity. These results shed light on OA pathogenesis and suggest a novel treatment pathway.


INTRODUCTION
Osteoarthritis (OA) is a multifactorial disease that manifests with synovial inflammation, cartilage destruction, joint swelling and pain [1,2]. Among the various risk factors, obesity has been linked to the risk of developing OA [3,4], with the Framingham Heart Study demonstrating a 1.5-to 2-fold higher risk of AGING developing knee OA among people who are obese compared with slimmer individuals [5]. In a US population-based study investigating communitydwelling older individuals (aged ≥70 years), a 5 kg/m 2 increase in BMI increased the likelihood of developing knee OA by 32% [6]. However, despite the observations suggesting a correlation between obesity and OA, the detailed mechanisms underlying this correlation are far from clear. What is known is that adipokines, multifunctional molecules secreted by adipose tissue, act as an intersecting link between obesity and OA by modulating the activities of cartilage, synovium, bone, and various immune cells [3,7].
The synovium plays an important role in the pathogenesis of OA. The synthesis of chondrolytic enzymes and proinflammatory mediators by the inflamed synovium leads to cartilage destruction, which enhances synovial inflammation, forming a vicious cycle [8,9]. OA synovial cells sustain arthritic pathology by excreting chondrolytic enzymes and inflammatory mediators [8,10,11]. Synoviumtargeted therapy could theoretically slow OA progression and lessen the severity of OA symptoms [12,13].
Apelin (APLN) is one adipokine that is established as a pivotal player in OA pathogenesis [14]. Early in vitro investigations suggested that APLN stimulates chondrocyte proliferation and significantly increases transcript levels of the catabolic factors matrix metalloproteinase (MMP)-1, -3 and -9, as well as the expression of the proinflammatory cytokine interleukin 1 beta (IL-1β) [14]. IL-1β is a major chondrolytic enzyme that induces the degradation of proteoglycan from cartilage and suppresses new proteoglycan synthesis [15][16][17].
Non-coding, single-stranded micro-ribonucleic acids (miRNAs) mediate the expression of target genes at the post-transcriptional level [18,19]. 3'-untranslated region (3'-UTR) miRNAs base-pair with the seed sequence of target mRNA molecules and effectively suppress target gene expression [1,20]. While both APLN and IL-1β are known to be involved in the pathogenesis of OA, no details exist as to any interaction between these molecules in OA synovial cells. In view of the importance of synovial cells in OA pathogenesis, we explored the crosstalk between APLN and IL-1β in human osteoarthritis synovial fibroblasts (OASFs). Myriads of miRNAs are involved in OA pathogenesis [1,8]. We hypothesized that APLN upregulates IL-1β expression by modulating miRNA expression in OASFs.

APLN expression is positively correlated with IL-1β expression in OA patients
To decipher crosstalk between APLN and IL-1β in the OA cohort, we used IHC staining to examine normal and OA synovial tissue samples. Levels of APLN and IL-1β expression were significantly higher in OA tissue than in normal tissue according to IHC staining ( Figure 1A-1C, respectively). A positive correlation was observed between APLN and IL-1β in IHC stain ( Figure 1D).

APLN stimulates IL-1β expression in human OASFs
Both APLN and IL-1β are known to act as proinflammatory mediators in arthritic disease [3]. However, no detailed information exists regarding any crosstalk between them in the pathogenesis of OA nor on how such an interaction may influence the synovium-induced inflammation in OASFs. APLN (0-10 ng/mL) dose-dependently stimulated IL-1β transcription and translation (Figure 2A and 2B, respectively) and the excretion of IL-1β protein by OASFs ( Figure 2C). Treatment of OASFs with APLN (10 ng/mL) for 24 hours stimulated IL-1β gene transcription and translation, as well as IL-1β protein excretion, in a time-dependent manner, as determined by RT-qPCR Western blot and ELISA assays, respectively ( Figure 2D-2F). However, stimulation of OASFs with APLN did not significantly increase TNFα expression (Supplementary Figure 1). These findings indicate that APLN enhances the downstream expression of IL-1β in human OASFs, via concentration-and time-dependent manners.
The protein kinase intracellular signaling molecule, extracellular signal-regulated kinase (ERK), regulates cellular proliferation and differentiation. ERK is AGING phosphorylated by various stimuli, including growth factors, cytokines and carcinogens, and is implicated in Wogonin-mediated anti-inflammatory and protective effects in human OA chondrocytes [23]. The Ras-ERK and PI3K-mechanistic target of rapamycin (PI3K-mTOR) signaling pathways have long been considered to act as linear signaling conduits that are activated by different stimuli, but recently it has become clear that ERK and PI3K regulate each other by intersecting and co-regulating downstream functions [24]. In this study, treatment of OASFs with ERK inhibitors (PD98059, U0126) and the transfection of OASFs with ERK siRNA prior to APLN administration significantly mitigated APLN-enhanced IL-1β synthesis ( Figure 4A-4D). In Western blot analysis, APLN time-dependently stimulated ERK phosphorylation ( Figure 4E), which was mitigated by PI3K inhibitors (LY294002, Wortmannin) ( Figure 4F). These findings suggest that APLN enhances IL-1β expression through the PI3K and ERK signaling pathways and that PI3K transmits information upstream of ERK in this cascade.

APLN enhances IL-1β expression by inhibiting miR-144-3p synthesis
Various miRNAs demonstrate differential expression patterns between OA and normal cartilage and are involved in the pathogenesis of OA [1,8]. However, the miRNA networks involved in OA pathogenesis are far from clear. We used open-source software (TargetScan, miRMap, RNAhybrid, and miRWalk) to identify miRNAs that could possibly interfere with IL-1β transcription ( Figure 5A; Supplementary Table 1). Among the 15 candidate miRNAs that could possibly bind to the 3'UTR region of IL-1β mRNA, levels of miR-144-3p expression were significantly decreased by the greatest extent after APLN administration. To confirm these findings, we compared levels of miR-144-3p expression in OASFs treated with APLN 1-10 ng/mL. APLN concentration-dependently inhibited miR-144-3p expression ( Figure 5B). To further determine whether APLN stimulates IL-1β expression by inhibiting miR-144-3p synthesis, we transfected OASFs with miR-144-3p mimic and observed AGING reductions in APLN-enhanced IL-1β mRNA and protein secretion ( Figure 5C and 5D).

Knockdown of APLN mitigates histologic features of OA
To validate the in vivo role of APLN, we investigated the effects of shRNA-mediated APLN knockdown in mitigating disease severity in the ACLT OA model. Compared with control samples, ACLT samples transfected with control shRNA exhibited significantly lower cartilage thickness in Safranin-O and H&E staining (Figure 6A), and a significantly higher proportion of IL-1β-and APLN-positive synovial cells in IHC analysis ( Figure 6B-6D). ACLT-induced histologic changes were reversed by downregulating APLN expression.

DISCUSSION
APLN has a crucial role in the pathogenesis of arthritic diseases, including rheumatoid arthritis and OA [3]. Early in vitro investigations indicated that APLN significantly increases MMP-1, -3 and -9 transcript levels, and IL-1β protein synthesis in chondrocytes [14]. In that same study, rats administered intra-articular injections of APLN (1 nM) experienced significant increases in gene expression levels of MMP-1, -3 and -9, A disintegrin and metalloproteinase with thrombospondin motifs 4 (ADAMTS-4) and -5, and IL-1β, while the gene expression of collagen II was significantly decreased by APLN; moreover, protein syntheses of MMP-3, -9, and IL-1β were significantly increased and synthesis of collage II decreased by APLN, compared with control samples. APLN also lowered cartilage proteoglycan production in articular cartilage [14]. The impact of APLN on OASFs has not been previously clarified. Here we reported that APLN levels were higher in OA tissue than in normal tissue. We did not examine APLN expression in other articular tissues such as cartilage and muscle, blood vessels, or inflammatory cells. The autocrine-and paracrine-driven aspects of APLN-induced effects in the articular microenvironment deserve further study. In this study, we found that APLN stimulates IL-1β expression in OASFs by stimulating PI3K and ERK phosphorylation and suppressing the downstream expression of miR-144-3p. These results add to the existing literature on OA pathogenesis.
Many miRNAs are involved in OA pathogenesis [25,26]. We used open-source software (TargetScan, miRMap, RNAhybrid, and miRWalk) to predict which miRNAs potentially interfere with APLN transcription. Among all candidate miRNAs, miR-144-3p was suppressed to the greatest extent by APLN. The analyses also showed that transfection of OASFs with miR-144-3p mimic mitigates APLN-stimulated IL-1β expression. These findings underscore the importance of miR-144-3p in the process of APLN-stimulated IL-1β expression.
The APJ receptor is a major receptor of APLN. The APLN/APJ system is a critical regulator of various physiological functions, such as glycometabolism, liver disease and macrophage activation [27][28][29]. This study did not examine APJ expression in synovial tissues. Further investigation needs to determine whether similar interactions are involved in APLN-induced promotion of IL-1β expression in OA synovial fibroblasts. The roles of PI3K and ERK in OA pathogenesis have been explored in previous work. The PI3K/Akt pathway is involved in both the degradation of extracellular matrix and chondrocyte death [30]. Inhibition of the PI3K/AKT/mTOR signaling pathway promotes autophagy of articular chondrocytes and attenuates inflammatory responses in rats with OA [31]. PI3K plays a pivotal role in allicin-suppressed IL-1β expression in chondrocytes [32]. In a dog model of surgically-induced OA, activation of ERK1/2, JNK and p38 was higher in OA tissue compared with normal tissue [33]. ERK is involved in the pathologic interaction between OA subchondral osteoblasts and articular chondrocytes as well as the hypertrophic differentiation of articular chondrocytes [34]. In this study, we show that APLN stimulates IL-1β expression via PI3K and ERK phosphorylation. Our data emphasize the importance of PI3K and ERK in OA pathogenesis. Transcriptional and AGING post-transcriptional modulations affect miRNA activation and inhibition [35]. In this study, treatment of OASFs with PI3K and ERK inhibitors antagonized APLNinhibited miR-144-3p expression, indicating that APLN suppresses the expression of miR-144-3p via PI3K/ERK signaling. Whether PI3K/ERK modulates miR-144-3p expression via post-transcriptional regulation warrants further examination.
In summary, our study shows that APLN treatment of OASFs triggers PI3K and ERK phosphorylation and contributes to a decline in miR-144-3p expression (Figure 7). These results improve our understanding about the role of OASFs in the pathogenesis of obesity and OA and may lead to the design of more effective therapy for OA patients.

MATERIALS AND METHODS
Antibodies against IL-1β, PI3K, and ERK were all bought from Santa Cruz (Santa Cruz, CA, USA). Antibodies against p-PI3K and p-ERK were purchased from Cell Signaling (Cell Signaling, UK). ON-TARGETplus siRNAs for IL-1β, PI3K, ERK and the control siRNA were purchased from Dharmacon (Lafayette, CO, USA). PI3K inhibitors (LY294002, Wortmannin) and ERK inhibitors (PD98059, U0126) were supplied by Calbiochem (San Diego, CA, USA). Cell culture supplements were purchased from Invitrogen (Carlsbad, CA, USA). A Dual-Luciferase ® Reporter Assay System was bought from Promega (Madison, WI, USA). The qPCR primers and probes, as well as the Taqman ® onestep PCR Master Mix, were supplied by Applied Biosystems (Foster City, CA, USA). All other chemicals not mentioned above were supplied by Sigma-Aldrich (St. Louis, MO, USA).

Cell culture
Synovial tissue from the suprapatellar pouch of the knee was obtained from 20 patients diagnosed with Ahlbäck stage IV OA. Synovial fibroblasts were cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS), 50 units/mL penicillin and 50 μg/mL streptomycin, as previously described [1,36].

Clinical samples
Synovial tissue samples were obtained from patients with OA undergoing knee replacement surgery and those undergoing arthroscopy after trauma/joint derangement (these were used as healthy controls) at China Medical University Hospital. The study protocol was approved by the Institutional Review Board (IRB) of China Medical University Hospital and all methods were performed in accordance with the IRB's guidelines and regulations. Informed written consent was obtained from all patients.

Real-time quantitative PCR analysis of mRNA and miRNA
Total RNA was extracted from human synovial fibroblasts by TRIzol; reverse transcription used 1 μg of total RNA transcribed into cDNA by oligo (dT) primers. Real-time quantitative PCR (RT-qPCR) used the Taqman ® One-Step RT-PCR Master Mix. All RT-qPCR assays were performed using the StepOnePlus sequence detection system (Applied Biosystems) [37,38].

Western blot analysis
Cell lysate was separated by SDS-PAGE electrophoresis then transferred to polyvinylidene difluoride (PVDF) membranes, following the method described in our previous work [39]. After blocking the blots with 4% bovine serum albumin, the blots were treated with primary antibody and then secondary antibody. Enhanced chemiluminescent imaging of the blots was visualized with the UVP Biospectrum system (UVP, Upland, CA, USA) [40][41][42].

Plasmid construction and luciferase assays
Wild-type and mutant IL-1β 3'-UTRs were generated on the miR-144-3p target recognition sites (seed sequences). The wild-type 3'-UTRs of IL-1β were cloned into the pmirGLO-luciferase reporter vector using Nhe1 and Xho1 restriction sites. The primer sequences used were defined as the IL-1β forward primer (CGGCTAGCAGAAACCACGGCCACATTT) and the reverse primer (GGCTCGAGTTCAGTG AAGTTTATTTCAGAACCA). All constructs were sequenced to verify that they contained the 3'-UTR inserts. The mutant 3'UTR region of IL-1β mRNA (mut-IL-1β 3'-UTR) was purchased from Invitrogen. Luciferase activity was assayed using the method described in our previous publications [1,8,43].

Lentiviral production
Recombinant lentiviruses were produced by transient cotransfection of 293T cells with short hairpin (sh)RNA-expressing plasmid (TRCN0000004877) with the packaging plasmid pCMVDR8.91 and the VSV-G envelope glycoprotein expression plasmid pMD.G. All were obtained from the National RNAi Core Facility at the Academia Sinica in Taiwan. After 48 hours, lentiviral particles carrying shAPLN (Lenti-shAPLN) were isolated from the supernatant of 293T cells. A plaque assay using serial dilutions was performed in AGING OASFs and determined the viral titer of Lenti-shAPLN to be ~7.1 x 10 6 plaque-forming units (PFU)/mL.

Experimental OA model
Sprague-Dawley (SD) rats (8 weeks of age, weighing 300-350 g) were purchased from the National Laboratory Animal Center in Taiwan and maintained under conditions complying with the Guidelines of the Animal Care Committee of China Medical University, Taichung, Taiwan, as described in our previous work [44]. We followed the protocol established by Wang et al. for our ACLT model to induce OA in rats [45]. In brief, the right knee was prepared in a surgically sterile fashion. Using a medial parapatellar mini-arthrotomy approach, the ACL fibers were transected with a scalpel and the entire medial meniscus was excised. After surgery, the joint surface was washed with sterile saline solution and the capsule and skin were sutured. The rats received prophylactic antibiotic with ampicillin 50 mg/kg body weight for 5 days post-surgery. After surgery (day 0), the rats were divided into 3 groups (n=8 per group), including a control group, an ACLT group, and an shAPLN-transfected ACLT group. For 6 weeks, the shAPLN-transfected ACLT group were given weekly intra-articular injections of ~7.1 x 10 6 PFU of APLN-shRNA. All rats were allowed to move freely in plastic cages until their necropsies at 6 weeks post-surgery.

Histological analysis
Immunohistochemistry (IHC) staining was performed on serial sections of the mice knee joints. After fixing the knee joints in 1% formaldehyde, the specimens were decalcified in 10% EDTA and dehydrated in ethanol/xylene, following previous work [46]. All sections were stained with primary anti-IL-1β (1:200) (Santa Cruz Biotechnology). Biotin conjugated goat anti-rabbit immunoglobulin G (IgG) was used as the secondary antibody and 3,3ʹ-diaminobenzidine tetrahydrochloride as the substrate for color development. Some specimens were also stained with Safranin-O/Fast-green or hematoxylin and eosin (H&E) [44,46].
The cartilage degeneration score evaluates overall cartilage pathology and includes the important parameters of collagen matrix fibrillation/loss and chondrocyte death/loss, with chondrocyte loss being the primary determinant of the score (Grade 0 = no changes; Grade 1 = minimal degeneration, with 5-10% of the total projected cartilage area affected by matrix or chondrocyte loss; Grade 2 = greater than mild degeneration, with 11-25% of the area affected; Grade 3 = greater than moderate degeneration, with 26-50% of the area affected; Grade 4 = marked degeneration, with 51-75% of the area affected; Grade 5 = more severe degeneration, with over 75% of the area affected).
Synovial membrane inflammation was scored as follows: Grade 0 = no changes (1-2 layers of synovial lining cells); Grade 1 = an increased number of lining cell layers (≥3-4 layers) or slight proliferation of subsynovial tissue; Grade 2 = an increased number of lining cell layers (≤3-4 layers) and/or proliferation of subsynovial tissue; Grade 3 = an increased number of lining cell layers (>4 layers) and/or proliferation of subsynovial tissue and infiltration of few inflammatory cells; Grade 4 = an increased number of lining cell layers (>4 layers) and/or proliferation of subsynovial tissue with infiltration of a large number of inflammatory cells.

Statistics
All values are given as the mean ± standard error of the mean (S.E.M.). The Student's t-test assessed betweengroup differences. A p value of <0.05 was considered to be statistically significant.

CONFLICTS OF INTEREST
None of the authors has any financial or personal relationships with other people or organizations that could inappropriately influence this work.