PR11‐364P22.2/ATF3 protein interaction mediates IL‐1β‐induced catabolic effects in cartilage tissue and chondrocytes

Abstract Osteoarthritis (OA) is a degenerative joint disease which lacks effective medical treatment due to ill‐defined molecular mechanisms underlying the pathology. Inflammation is a key factor that induces and aggravates OA. Therefore, the current study aims to explore roles of the dysregulated long non‐coding RNAs in the pro‐inflammatory cytokine IL‐1β‐mediated catabolic effects in cartilage tissue and chondrocytes. We identified RP11‐364P22.2 as dysregulated in OA patient‐derived cartilage tissues and highly responsive to IL‐1β stimulus. RNA pull‐down coupled with mass spectrometry demonstrated that RP11‐364P22.2 physically binds to activating transcription factor 3 (ATF3) and thus increases the protein stability and facilitates its nuclear translocation. Loss‐ and gain‐of‐function assays indicated that the interaction between RP11‐364P22.2 and ATF3 is indispensable for the detrimental effects of IL‐1β including growth inhibition, apoptosis induction as well as degradation of the key chondrocyte structural proteins of type II collage and Aggrecan and synthesis of the extracellular matrix‐degrading enzyme MMP13 in chondrocytes. In vivo, depletion of the RP11‐364P22.2 effector ATF3 drastically prevented OA development in the rats with surgical destabilization of the medial meniscus (DMM). These results highlight the important roles of lncRNAs in the pathogenesis of OA and indicate the RP11‐364P22.2/ATF3 regulatory axis as a potential therapeutic target of inflammation‐induced OA.

arthritis, synovial fluid white blood cell count in OA barely exceeds 1000-2000/mL. 6 At the molecular level, the inflammation in OA joints is characterized by the presence of massive amounts of proinflammatory mediators including cytokines such as TNF, IL-1β and IL-6 and chemokines which are involved in the innate immune response to joint damage. 7 Long non-coding RNA (lncRNA) is a superfamily of transcribed RNA molecules which are more than 200 nucleotides (nt) in length and do not encode proteins. 8 LncRNA is pathogenic to many human diseases including cancer, Alzheimer's disease and diabetes. 9 Increasing evidence shows that lncRNAs are also key regulators of cartilage formation, bone and cartilage homeostasis, and thus affect the progress of OA. 10 However, the roles of lncRNAs in OA development, especially the inflammation process, remain largely unknown.
In the current study, we demonstrated that RP11-364P22.2 is not only over-expressed in OA cartilage tissues and chondrocytes, but also highly responsive to the stimulus of the pro-inflammatory cytokine IL-1β. Binding of RP11-364P22.2 to activating transcription factor 3 (ATF3) stabilizes the protein and facilitates its nuclear translocation, which is indispensable for OA development in the rats underwent surgical destabilization of the medial meniscus (DMM) in vivo. In vitro mechanism studies displayed that the interaction between RP11-364P22.2 and ATF3 is essential for IL-1β-induced growth inhibitory and apoptotic effects as well as degradation of

| Patient samples
Osteoarthritis articular cartilage tissues were collected from OA patients (n = 10) undergoing the entire knee replacement surgery, while the amputations from non-OA and rheumatoid arthritis patients (n = 10) were used as controls. All the patients were informed of their participation in the current study and gave their informed consent. This study was approved by the Ethics Committee of the Ethics Committee of Xiangya Hospital (Changsha, China) and performed in accordance with the Declaration of Helsinki.

| Cell line and reagents
Human chondrosarcoma cell line SW1353 was purchased from Chinese Academy of Sciences Cell Bank (Shanghai, China) and cultured in Dulbecco's Modified Eagle Medium (DMEM)/F12 containing 10% foetal bovine serum (FBS) in humidified atmosphere 5% CO2 in air at 37°C. IL-1β was obtained from Roche Diagnostics (Germany), and cycloheximide (CHX) was purchased from Sigma-Aldrich.

| Gene expression
For lncRNAs and mRNA expression studies, total RNA was extracted using TRIzol reagent (Invitrogen) followed by reversed transcription using RevertAid TM H Minus First Strand cDNA Synthesis Kit (Fermentas). Quantitative real-time PCR was conducted using SYBR Green PCR Master Mix (ABI 4309155) in ABI7900 PCR machine. The sequences of primers are shown in Table 1.

| Primary human chondrocytes culture
Preparation of primary human articular chondrocytes was performed as described by Manning and Bonner. 11 Briefly, the disinfected cartilage was first cropped into small fragments. Chondrocytes were then released by sequential digestion with 0.25% trypsin (Gibco) and 1 mg/ mL type II collagenase (Sigma) at 37°C and cultured with DMEM/F12 containing 10% foetal bovine serum (FBS) in a 37°C water-jacketed CO 2 incubator. The primary chondrocytes exceeding five passages were abandoned due to an irrepressible dedifferentiation caused by repetitive subculture.

| Immunofluorescence
The monolayer chondrocytes on the glass cover slide were fixed with 4% paraformaldehyde and blocked with 5% normal goat serum. After an overnight incubation with primary antibodies (Table 2) at 4°C, cells were washed three times with PBS and incubated with Alexa Fluor-conjugated secondary antibody for 1 hour at room temperature. DAPI was used for nuclear counterstain.

| Cell proliferation
The time-dependent chondrocytes proliferation was assessed using MTT Cell Proliferation and Cytotoxicity Detection Kit (KeyGEN Biotech, #KGA312). Briefly, cells were seeded at a density of 1 × 10 4 cells per well in 96-well plates and treated with IL-1β for the indicated time. By the end of the treatment, an equal volume of MTT reagent was added into each well, and the plate was incubated at 37°C for 3 hours. Carefully remove the supernatant without disturbing the attached cells, add 150 µL MTT solvent into each well followed by shaking on an orbital shaker for at least 15 minutes in the dark and read the absorbance at OD = 590 nm.

| Cell apoptosis
Apoptosis and cell death in chondrocytes were examined through flow cytometry using Annexin V-PE/7AAD apoptosis kit (Abnova, #KA3809) based on the instruction of the manual. Briefly, the treated chondrocytes were washed twice with cold staining buffer and then resuspended in Annexin V Binding Buffer. Add sequentially PE Annexin V and 7AAD viability staining solution, and incubate for 15 minutes at room temperature in the dark. Add Annexin V Binding Buffer up to 500 µL of total volume in each tube and analyse by flow cytometry.

| Western blot
Total protein was extracted using RIPA buffer supplemented with protease inhibitor cocktail (Sigma #P8340) from each experimental group of chondrocytes. 30 µg of whole cell extracts was fractionated by SDS-PAGE and transferred to a nitrocellulose membrane. After blocking with 5% non-fat milk for 1 hour at room temperature, the TA B L E 1 Sequence information Bio-control oligo 5' TCCTTCCTCTCTTTCTCTCCCTTGTGA 3' membrane was incubated with the primary antibodies (see Table 2)

| Fluorescence in situ hybridization (FISH)
The monolayer chondrocytes grown on glass slides were fixed by 4%

| RNA pull-down assay
The full-length sense and antisense RP11-364P22.2 were amplified using primers listed in Table 1  RNA probes and proceeded with the standard pull-down procedure. The bound proteins in the pull-down product were analysed by Western blotting using ATF3 antibody ( Table 2).

| RNA immunoprecipitation (RIP)
RNA immunoprecipitation was performed using Magna RIP TM RNA-

| RNA interference
Sequences of small interfering RNA (siRNA) oligonucleotides targeting RP11-364P22.2 and the negative control siRNA are listed in

| Bioinformatics
Data for LncRNA expression Venn diagram were retrieved from three independent human OA patient-derived data sets. Among lage from human subjects with and without osteoarthritis was conducted by microarray. 15 For comparing ATF3 expression between normal and OA cartilage tissues, the expression data were retrieved, respectively, from the data sets of GSE11 7999 and GSE51588 in the study by Chou et al (2013), in which microarray was applied to assess gene expression profiling in human OA subchondral bone. 16 In silico prediction of RP11-364P22.2/ATF3 interaction was conducted using catRAPID (http://s.tarta glial ab.com/page/catra pid_group).

| Safranin-O staining
Knee joints were harvested and fixed using zinc-buffered forma-

| Statistical analysis
Data are expressed as mean ± standard deviation of three independent experiments. Two-tailed Student's t test was performed for determining a difference between two groups, while one-way ANOVA analysis followed by Tukey's honestly significant difference post hoc test was conducted for comparing two-group difference among the multiple groups using GraphPad Prism 8 software. For MTT assay results, only treatment or genotype effect was considered by comparing each experimental group mean with the control group mean within each time point using two-way ANOVA analysis followed by Dunnett's multiple comparisons test. *P < 0.05, **P < 0.01, ***P < 0.001, n.s., non-significant.

| RP11364P22.2 is highly expressed in OA cartilage tissues and responsive to IL-1β
Osteoarthritis is accompanied with a simultaneous inflammation from early stage, and IL-1β is one of the key inflammatory cytokines involved in the pathogenesis of the disease. 18 To explore those essential lncRNAs involved in IL-1β-induced OA, we retrieved and interrogated lncRNAs that are not only upregulated in OA cartilage tissues but also highly responsive to IL-1β exposure from three independent previous studies. [12][13][14] Interestingly, RP11-364P22.2 turned out as the unique signature in the intersection of all three databases ( Figure 1A and B). We confirmed by qPCR the expression of RP11-364P22. In addition to the morphology, the chondrocyte phenotype of the primary cultures was confirmed by immunofluorescence of type II collagen and Aggrecan which are two major components of the extracellular matrix (ECM) of articular cartilage 19 ( Figure 1D). IL-1β exhibited a dose-dependent growth inhibitory effect on both the primary OA chondrocytes and chondrosarcoma cell line SW1353 ( Figure 1E). As IL-1β at 10 ng/mL concentration exerted a robust inhibition compa- rable to the higher one tested, it was adopted for the subsequent in vitro experiments. Upon treatment with IL-1β, we detected that RP11-

364P22.2 was mostly upregulated in both cell types in comparison
with the untreated cells ( Figure 1F). RP11-399K21.10 was also significantly induced in the primary chondrocytes but not SW1353 cells.

| IL-1β inhibits chondrocytes proliferation and induces apoptosis in vitro
Growing evidence suggests that chondrocyte cell death occurs and contributes to OA development. 20 Figure 3E). The nuclear factor kap-paB (NF-κB) is a general and key inflammatory mediator in many cell types. 22 The activation of NF-κB-dependent signalling pathway in the IL-1β-treated chondrocytes ( Figure 3E) indicated that IL-1β might intensify the inflammatory reaction during OA development.
Interestingly, such role of IL-1β relied to a large extent on the expression of RP11-364P22.2 as knock-down of the lncRNA attenuated while its forced expression promoted the activation of NF-κBdependent signalling pathway ( Figure 3E).

| ATF3 is a target of RP11-364P22.2
The abundant cytosolic RP11-364P22.2 in the IL-1β-treated chondrocytes reminds the protein modulator-related function. 23 Next, we performed RNA pull-down coupled with mass spectrometry to identify the RP11-364P22.2-interacting proteins. Mass spectrum analysis of the unique protein bands appeared in the sense but not antisense RP11-364P22.2-incubated cell lysates by silver staining revealed ATF3 as a potential target ( Figure 4A). Consistently, Western blot assay detected ATF3 in sense but not antisense RP11-364P22.2 pull-down cell lysates as well ( Figure 4B). In turn, RIP assay using ATF3 antibody combined with qPCR analysis showed that ATF3 specifically precipitated RP11-364P22.2 in chondrocytes cell lysates, and the precipitation was increased accordingly with an ectopic expression of RP11-364P22.2 ( Figure 4C).
We further performed an in silico catRAPID analysis 24 to define the domains in RP11-364P22.2 that physically interact with ATF3. As the output interaction profile ( Figure 4D) and interaction matrix ( Figure 4E)

| RP11-364P22.2 is required for protein stability and IL-1β-induced nuclear translocation of ATF3
Previous studies had indicated the involvement of ATF3 in pathogenesis of early OA development. 25,26 Indeed, ATF3 expression was significantly increased in the OA cartilage tissues than non-OA controls ( Figure 5A). We continued to exam the effects of RP11-364P22.2 in ATF3 expression. The qPCR results showed that while IL-1β greatly induced ATF3 in the chondrocytes, depletion of RP11-364P22.2 did not affect its expression at the transcription level ( Figure 5B). Instead, induction of ATF3 protein by IL-1β in the chondrocytes was largely abolished upon knocking down of RP11-364P22.2 ( Figure 5C). In chondrocytes, most majority of ATF3 was preferentially expressed in the nucleus with a small portion being detected in the cytoplasm ( Figure 5D and E). IL-1β treatment promoted a nuclear translocation of the cytosol ATF3 protein ( Figure 5D and E), which was completely prohibited in the absence of RP11-364P22.2 ( Figure 5D and E). This was due largely to the fact that loss of RP11-364P22.2 decreased ATF3 protein stability as indicated by the accelerated protein degradation in the presence of cycloheximide, a protein synthesis inhibitor in eukaryotes that inhibits translation elongation, 27 while the opposite was true as well ( Figure 5F).

| RP11-364P22.2-stabilized ATF3 is required for the catabolic effects of IL-1β in chondrocytes in vitro
Next, we assessed the role of interaction between RP11-364P22.  Figure 6C).
Importantly, the IL-1β-activated NF-κB signalling pathway was also reversed to a normal level upon breaking up the interaction between RP11-364P22.2 and ATF3 ( Figure 6C).

| ATF3 is indispensable for the surgical DMMinduced OA development in vivo
Our data from the in vitro chondrocytes-based studies dem- sham group rats ( Figure 7B). Western blot results also showed the increased synthesis of MMP-13 and decreased expression of type II collage and Aggrecan in the cartilage tissues of these rats ( Figure 7C). Strikingly, lentiviral shRNA-mediated stable knockdown of ATF3 in the cartilage tissue of knee joints retained in a great degree the Safranin-O staining ( Figure 7A) and synthesis of the chondrocyte structural proteins ( Figure 7C) as indicative of an alleviated OA development after the surgical DMM. In accordance, cartilage surface in the ATF3-depleted rats after the surgery was displayed with the regularly arranged cells and clear tide line but little fibrillation ( Figure 7A). The OARSI score was significantly lower than that of in DMM group as well ( Figure 7B). Taken all together, our study demonstrated that the highly expressed RP11-364P22.2 is essential for the IL-1β-induced catabolic effects in cartilage tissue and chondrocytes by stabilizing ATF3.

| D ISCUSS I ON
Inflammation is an important concomitant throughout initiation and development of OA. Aetiologies such as joint injury, obesity and ageing that cause OA also contribute to inflammatory processes. [28][29][30][31] Emerging evidence has implicated the involvement of the pre- Consistent with previous report that inflammatory cytokines induce ATF3 expression through NF-κB pathway, 25 we detected both the activated NF-κB signalling pathway and upregulated ATF3 in the IL-1β-treated chondrocytes. Interestingly, RP11-364P22.2 turned out to be essential for IL-1β-induced activation of NF-κB pathway.
We speculate that it might because of the post-translational regulation of the lncRNA on ATF3. The amount of ATF3 protein dramatically decreased in the chondrocytes lack of RP11-364P22.2, which in turn attenuates the phosphorylation of IκB and p65. 25

CO N FLI C T S O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets generated/analysed in the present study are available upon reasonable request from the corresponding author.