CTGF facilitates cell‐cell communication in chondrocytes via PI3K/Akt signalling pathway

Abstract Purposes Gap junction intercellular communication (GJIC) is essential for articular cartilage to respond appropriately to physical or biological stimuli and maintain homeostasis. Connective tissue growth factor (CTGF), identified as an endochondral ossification genetic factor, plays a vital role in cell proliferation, migration and adhesion. However, how CTGF regulates GJIC in chondrocytes is still unknown. This study aims to explore the effects of CTGF on GJIC in chondrocytes and its potential biomechanism. Materials and methods qPCR was performed to determine the expression of gene profile in the CCN family in chondrocytes. After CTGF treatment, CCK‐8 assay and scratch assay were performed to explore cell proliferation and migration. A scrape loading/dye transfer assay was adopted to visualize GJIC in living chondrocytes. Western blot analysis was done to detect the expression of Cx43 and PI3K/Akt signalling. Immunofluorescence staining was used to show protein distribution. siRNA targeting CTGF was used to detect the influence on cell‐cell communication. Results The CTGF (CCN2) was shown to be the highest expressed member of the CCN family in chondrocytes. CTGF facilitated functional gap junction intercellular communication in chondrocytes through up‐regulation of Cx43 expressions. CTGF activated PI3K/Akt signalling to promote Akt phosphorylation and translocation. Suppressing CTGF also reduced the expression of Cx43. The inhibition of PI3K/Akt signalling decreased the expressions of Cx43 and thus impaired gap junction intercellular communication enhanced by CTGF. Conclusions For the first time, we provide evidence to show CTGF facilitates cell communication in chondrocytes via PI3K/Akt signalling pathway.

biological stimuli and maintain tissue homeostasis with activation of signalling pathways such as proliferation, differentiation and death. 2 Gap junction proteins have significant roles in such cellular communication, including connexins (Cxs) and pannexins (pxs). [2][3][4] The connexons are formed by the oligomerization of six connexins, function as hemichannels, regulating the release of small signalling molecules such as calcium, other ions, ATP and prostaglandins between cells and extracellular environment. 5,6 In addition, two connexons from adjacent cells dock with each other to form gap junctions, thereby providing a channel for direct exchange of small signalling molecules (<1 kDa) between communicating cells. Cx43 is the most widely expressed connexins in these tissues, which are thought to play an important role in organogenesis and homeostasis. Cx43 was located at the edge of the cell and expressed in each layer of articular cartilage, while mainly expressed in the superficial zone of the growth plate. [7][8][9] Connexins dysfunction may cause joint diseases such as rheumatoid arthritis (RA), osteoarthritis (OA). 7,10 F I G U R E 1 CTGF promotes cell-to-cell comunication in chondrocytes. A, Quantitative real-time PCR showing the highest gene expression of CCN2 (CTGF) in CCN family. Relative gene profiles of CCN family were calculated by normalizing to internal GAPDH. Data were derived from the three independent experiments (n = 3). B, CCK-8 assay showing increased cell proliferation of chondrocytes induced by CTGF. Data were derived from the three independent experiments (n = 3). *Significant difference was in comparison to the control group. Cont = control. C, Scratch wound closure assay was performed and captured by phase-contrast microscopy. Data were based on three independent experiments (n = 3). D, Quantification was done to access the wound closure in Figure 1C. *Significant difference was in comparison to the control group. Cont = control. E, The scrape loading/dye transfer (SL/DT) assay showing the increased cell-to-cell transmission of Lucifer yellow through gap junction in chondrocytes induced by CTGF (50 ng/mL) within 7 and 10 min by CLSM. The results were derived from three independent experiments (n = 3). F, Quantification was done to access the gap junction formation in Figure 1E. *Significant difference was in comparison to the control group. G, Gap junction formations in living chondrocytes were shown to be increased in the CTGF group (50 ng/mL) by CLSM after Lucifer yellow loading at 1 mg/mL. The dye transfer images were collected at 10 min after Lucifer yellow loading. The results were based on three independent experiments (n = 3). The white arrows indicated the lightened gap junction channels. H, Western blot showing that small interfering RNA targetting CTGF at 50 nmol/L at 48 h inhibited the CTGF expression successfully at protein level. Scramble means scrambled control group. Si-CTGF means the small interfering RNA targetting CTGF group. The data were based on three independent experiments (n = 3). I, Quantification was performed to show the protein changes of in Figure 1H. The statistical analyses were based on three independent experiments (n = 3). *P < .05. J, Gap junction formations in living chondrocytes were reduced in the siRNA group at 50 nmol/L siRNA at 48 h by CLSM after Lucifer yellow loading at 1 mg/mL. The dye transfer images were collected at 10 min after Lucifer yellow loading. The results were derived from three independent experiments (n = 3). The white arrows indicated the lightened gap junction channels. K, Quantification was done to access the gap junction formation in Figure 1J. **P < .01. L, Representative changes of intercellular connection in chondrocytes in response to CTGF (50 ng/mL) via SEM at 12 h treatment. The white arrows represent cell-cell connecting junctions among chondrocytes. White boxed areas in the first column were zoomed in and displayed in the second one. The images were derived from three independent experiments (n = 3). M, Representative changes of intercellular connection in chondrocytes with Si-CTGF (siRNA) via SEM at 48 h. White boxed areas in the first column were zoomed in and displayed in the second one. The white arrows represent cell-cell connecting junctions among chondrocytes. The images were derived from three independent experiments (n = 3) Connective tissue growth factor (CTGF), known as CCN2, is a 38kDa extracellular matrix protein, a member of CCN (cellular communication network factor) protein family. These proteins are made up of five domains, including secretory signal peptide (SP), insulin-like growth factor-binding protein (IGFBP) module, von Willebrand factor type C repeat (VWC), thrombospondin type 1 repeat (TSP1), and C-terminal (CT) module. 11 CTGF, identified as an endochondral ossification genetic factor, plays vital roles in regulating various cellular functions, including proliferation, migration, adhesion, survival, differentiation and synthesis of ECM proteins in skeletal development. [12][13][14] CTGF participates in a variety of biological processes and plays multiple roles, such as angiogenesis, chondrogenesis, osteogenesis, wound healing, fibrosis and tumour formation. [13][14][15] In addition, applying CCN2 together with gelatin hydrogel to cartilage defects can repair articular cartilage. 16 Transfection with siRNA targeting CTGF ribonucleic acid in cardiomyocytes reduced Cx43 expression. 17 However, the relationship between CTGF and cell-cell communication is not yet known and needs be further investigated.
In our study, we used recombinant CTGF to explore the relationship between CTGF and the gap junction in chondrocytes.
We examined the effects of CTGF on intercellular communication, Cx43 expression, cell proliferation and migration. Furthermore, we demonstrated that CTGF induced Cx43 formation through the phosphatidylinositol 3-kinase/Akt (PI3K/Akt) signalling pathway.

| Preparation of tissues and cells
The animal materials used in this study meet the requirements of ethical principles, and all protocols have been reviewed and approved by the institutional review board of our hospital (IRB, institutional review board of West China Stomatological Hospital, number: wchsirb-D-2017-029). Articular cartilage was isolated from newborn C57BL mice by using the enzymatic digestion method. The articular cartilages were cut into pieces and trypsinized with 0.25% protease solution dissolved in Dulbecco's modified Eagle's medium (high-glucose DMEM, 0.1 mmol/l non-essential amino acids, 4 mmol/l L-glutamine, 1% penicillinstreptomycin solution, HyClone, Logan, UT, USA) for 30 minutes at 37°C, and then was digested with 0.2% type II collagenase (sigma) at 37°C for 12 hours. After centrifugation with 8 minutes at 1000 rpm, the chondrocytes were seeded onto 25-cm 2 cell culture flasks in DMEM supplemented with 10% foetal bovine serum (FBS) and 1% penicillin/streptomycin and incubated with 5% CO 2 at 37°C. The culture media were changed every 2 days.

| Quantitative real-time PCR (qPCR)
Total RNA was extracted from primary articular chondrocytes with the  Table 1.  Figure 2D after CTGF treatment (50 ng/mL). The statistical analyses were based on three independent experiments (n = 3). *P < .05. F, Representative IF staining by CLSM showing the enhanced expression and distribution of Cx43 in chondrocytes induced by CTGF (50 ng/mL). Cytoskeleton was stained with FITC-phalloidin (green) and nuclei were stained with Dapi (blue). White arrows showed Cx43 accumulation was along gap junction. The results were derived from three independent experiments (n = 3). G, Fluorescence optical density (OD) assay was performed to show the distribution of Cx43 cross the cell body in Figure 2F. H, Western blot showing that siRNA of CTGF in chondrocytes reduced Cx43 expression within 48 h. The data were based on three independent experiments (n = 3). I, Quantification was performed to show the protein changes of Cx43 in Figure 2H. The statistical analyses were based on three independent experiments (n = 3). *P < .05. J, Representative IF staining by CLSM showing the decreased expression of Cx43 in siRNA of CTGF-treated chondrocytes. Cytoskeleton was stained with FITC-phalloidin (green) and nuclei were stained with dapi (blue). White arrows showed Cx43 accumulation was along gap junction. The results were derived from three independent experiments (n = 3). K, qPCR showed reduced CTGF gene expression after siRNA of CTGF transfection in chondrocytes. The results were derived from three independent experiments (n = 3). **P < .01. L, qPCR showed reduced Cx43 gene expression after siRNA of CTGF transfection in chondrocytes. The results were derived from three independent experiments (n = 3). *P < .05

| CTGF Small interfering RNA (siRNA) transfection
Cells were seeded onto 35 mm cell culture dishes for 12 hours and transiently transfected with specific siRNA oligonucleotides (hanbio, Shanghai, China) using Lipofectamine RNAiMAX

| Western blot analysis
The expression levels of Cx43, CTGF, essential proteins (AKT, p-AKT) involved in the PI3k/AKT signalling were analysed by Western blot. Briefly, chondrocytes were seeded at a density of 2 × 10 5 cells/well and cultured with DMEM containing 10% FBS.
After seeding, cells were starved using 2% FBS DMEM for 12 hours.
After starvation, the culture medium was changed to 1% fresh The immunocomplexes were visualized with using a Immobilon ® Western (P90719, Millipore). The protein expression levels were analysed with Image J software (NIH, Bethesda, MD) and normalized against β-actin.

| Scanning electron microscopy (SEM)
Chondrocytes were seeded onto 35 mm cell culture dishes with 70% confluence. After incubating with CTGF or LY294002 for 12 hours, the culture medium was discarded and replaced by 2.5% glutaraldehyde for fixing for 2 hours. Besides, chondrocytes were seeded onto 35 mm cell culture dishes for transfection with 50 nmol/L siRNA that targeting CTGF gene. After 48 hours, the culture medium was also discarded and replaced by 2.5% glutaraldehyde for fixing for 2 hours. Then these cells were subsequently dehydrated with graded ethanol at concentration from 30%, 50%, 70%, 80%, 90%, to 100% for 15 minutes at each level. Subsequently, the samples were coated with a gold layer and then ready to be visualized for cell junctions by SEM.

| Scratch assay
After articular chondrocytes were grown to a full confluent monolayer in a 6-well plate, we applied a thin scratch in the middle  We then monitored LY dye transfer by confocal laser scanning microscopy (CLSM, Olympus, FV3000, Japan) to evaluate intercellular communication.

| Immunofluorescence and confocal laser scanning microscopy (CLSM)
Chondrocytes were seeded and cultured in observation dishes speci- were applied to stain the nuclei and cytoskeleton. The cells then were observed with CLSM. All experiments were repeated at least three times.

| Statistical analysis
Experiments were performed in triplicate. Quantitative results are presented as the mean ± standard deviation (SD) and plotted with (GraphPad Prism Inc, San Diego, CA). Data were assessed through one-way ANOVA followed by Tukey's protected least-significant difference post hoc test for multiple comparisons. P value < .05 was considered to be statistically significant.

| CTGF promotes intercellular communication and gap junction formation in chondrocytes
To begin with, we investigated the gene expressions of CCN family in primary chondrocytes. The CCN family comprises six members with a similar mosaic primary structure. They are Cyr61/CCN1, CTGF/CCN2, Nov/CCN3, WISP1/CCN4, WISP2/CCN5 and WISP3/ CCN6. 13 Our qPCR results indicated that mRNA of CTGF (CCN2) ranked No.1 in primary chondrocytes ( Figure 1A).
Subsequently, CCK-8 assay confirmed that the cell proliferation was also enhanced with the stimulation of CTGF in a dosedependent manner at 24 hours ( Figure 1B). Then the scratch wound closure assay was applied to access the migration rates of chondrocytes at 12 and 24 hours after CTGF treatment with different concentrations. The analysis of the captured images indicated that cell migration was significantly accelerated in the CTGF group compared with the control group ( Figure 1C&D). Intercellular communication was closely correlated with cell proliferation and migration due to the requirement of full cell confluence. Thus, we then explored the influence of CTGF on the gap junction formations in chondrocytes.
With the scrape loading/dye transfer assay, the CTGF-treated group showed higher ability to transfer Lucifer yellow (LY) dye between chondrocytes when compared with the control group ( Figure 1E&F).
Chondrocytes extended many cytoplasmic cilia to neighbouring cells to establish intercellular communication. The entire cell including nuclei was illuminated by Lucifer yellow, forming multiple transmission chains to transfer yellow fluorescent molecules to a farther distant cell. 18,19 Our results confirmed that CTGF promoted this gap junction formation in chondrocytes ( Figure 1G

| CTGF promotes gap junction formation through the increase of Cx43 protein in chondrocytes
Cx43, one of gap junction proteins, is widely expressed in 80%-100% of chondrocytes in each zone. 20 Our previous report had confirmed that the mRNA level of Cx43 was the highest among the connexin family in the primary chondrocytes. 19 In order to explore the effect of CTGF on Cx43, we performed western blot and immunofluorescence. Western blot assay confirmed that the increase of Cx43 protein was in line with the increased concentrations of CTGF ( Figure 2A). In addition, quantification confirmed that the expression of Cx43 after CTGF treatment (50 ng/mL) was much higher than that of the control group at both 12 hours ( Figure 2B) and 36 hours To further explore the distribution of Cx43 in CTGF-induced chondrocytes, immunohistochemical staining was performed and it was found that Cx43 is mainly located in the cytoplasm, cell membrane and intercellular connection ( Figure 2F). After treatment with 50 ng/ mL CTGF, we found that Cx43 was accumulated increasingly in the cytoplasm, cell membrane and gap junction plaques between cells which were characterized by continuous dotted lines ( Figure 2Fboxed area, yellow). Further fluorescent qualification also confirmed that the expression of Cx43 in the CTGF-treated group was significantly enhanced ( Figure 2G). Furthermore, suppressing CTGF expression by siRNA reduced the expression of Cx43 protein when compared to the scrambled group ( Figure 2H&I). CLSM also showed siRNA in chondrocytes reduced the Cx43 expression in cell membranes and cell-cell connection ( Figure 2J). At mRNA levels, inhibiting the levels of CTGF also reduced the expression of Cx43 mRNA ( Figure 2K&L).

| CTGF triggers PI3K/AKT signalling to mediate the expression of connexin43
The PI3K/Akt signalling pathway was an important regulator of cellular proliferation and differentiation in chondrocytes. 21,22 We found that CTGF could activate the protein level of p-Akt in a dosedependent manner ( Figure 3A&B). At 50 ng/mL, CTGF increased the expressions of p-Akt in a time-dependent manner ( Figure 3C&D).
We then used LY294002, a specific PI3K inhibitor for PI3K/Akt signalling and found that LY294002 reduced the expression of p-Akt at 10 and 15 μmol/L. Furthermore, we found that LY294002 could reduce the expression of Cx43 at 10 and 15 μmol/L ( Figure 3E&F).
The immunofluorescence assay further confirmed LY294002 could specifically reduce the expression of Cx43 compared with the control group ( Figure 3G).
Further experiment by western blot confirmed that intense expression of p-Akt protein induced by CTGF could be attenuated by the addition of LY294002 ( Figure 4A&B). Accordingly, the increased Cx43 expression induced by CTGF can also be reduced by pretreatment with LY294002 at 2 hours ( Figure 4C&D). Moreover, immunofluorescence was further performed to identify the expression of p-Akt in chondrocytes induced by CTGF. The p-Akt protein was located predominantly in the cytoplasm of no-stimulated cells.
Rather, p-Akt accumulation was observed at the nuclear region at 2 hours with the stimulation with 50 ng/mL CTGF ( Figure 4E). The enhanced p-Akt expression induced by CTGF could be neutralized by addition of the PI3k inhibitor LY294002. (Figure 4E). Most importantly, immunofluorescence also showed CTGF alone increased the expression of Cx43, but this increase was significantly impaired at the presence of the inhibitor LY294002 ( Figure 4F). Detailed research by western blot also revealed that inhibition of CTGF F I G U R E 6 Schematic diagram illustrates the biomechanism of CTGF-regulated Cx43 in chondrocytes. In brief, CTGF induced the PI3K/ Akt signalling pathway to promote the expression of Cx43 and facilitate the gap junction formation expression reduced both the Cx43 and p-Akt expression at protein levels ( Figure 4G-J). Together, these results demonstrate the significant role of PI3K/AKT signalling in mediating CTGF-induced Cx43 expression.

| PI3K/Akt signalling shows its great importance in CTGF-induced cell-cell communication in chondrocytes
To further investigate whether the PI3K/Akt signalling could mediate the CTGF-induced intercellular communication, we used scrape loading/dye transfer assay in living chondrocytes ( Figure 5A Figure 5A); this phenomenon also could be reverted by LY294002. SEM images also showed the cell junctions induced by CTGF could be reduced by LY294002 ( Figure 5C). Our study provides evidence for the first time that CTGF stim- Other studies also confirmed that Cx43 was transferred from the endoplasmic reticulum to the Golgi, and then oligomerized in the trans-Golgi network, and finally, transported to the plasma membrane along an intact cytoskeleton. 33,34 This phenomenon might indicate that Cx43 proteins were in a constant state of flux, providing a dynamic membrane domain in response to different stimuli, which is in consistent with previous study. 34,35 Align with previous study, 21,36 CTGF also effectively promoted chondrocytes proliferation and migration, which may owe to dual function of Cx43. That is, connexin proteins form gap junctions for the direct exchange of tissue homeostasis determinants between nearby cells and also act as an intercellular signalling reservoir by reversibly interacting with homeostasis determinants such as β-catenin to modulate the production of cell growth and cell death. 37 Consequently, we supposed that Cx43 not only form the gap junction but also directly interact with cell growth and cell death regulators under the stimulation of CTGF in chondrocytes. Further researches are required to verify this hypothesis.

| D ISCUSS I ON
The serine/threonine kinase protein kinase B (PKB)/Akt is a central player of various physiological functions such as metabolism, proliferation, survival, growth, angiogenesis, migration and invasion.
Akt/PKB can be activated in a PI3K-dependent manner with the stimulation of numerous growth factors, cytokines and hormones such as IGF-1 and PTH. [38][39][40] PI3K/AKT signalling is closely associated with Cx43. Inhibiting PI3K/ Akt signal transduction decreases Cx43 expression at mRNA and protein levels in MC3T3E1 cells. 39 Previous studies have also revealed CTGF promoted cell proliferation, differentiation, extracellular matrix protein production, cell migration and actin cytoskeleton reorganization through PI3k/Akt pathway. 21,41,42 To explore the potential mechanism of CTGF-induced Cx43 expression, we examined whether PI3K/Akt signalling pathway was activated. In accordance with previous study, 41,43 Akt was phosphorylated and transported into the nucleus upon stimulation of CTGF in chondrocytes, where it regulated the activity of transcription, indicating the activation of the PI3K/Akt pathway ( Figure 4E).

Si-CTGF in chondrocytes inhibited the expression of p-Akt. Akt is
initially synthesized on the endoplasmic reticulum, but translocated to the plasma membrane to be activated, and then repositions in various regions, including the cytosol and the cell nucleus, 44 suggesting manifestation of differentially regulated phosphatase activity in these two compartments. From the results of confocal microscopy, p-Akt was primarily concentrated in nucleus under the stimulation of CTGF. 45 Accordingly, we speculate CTGF could play biological roles by inducing p-Akt translocation in nucleus rather cytosol. Akt predominantly localizes in the cytoplasm, but it also translocates to the nucleus with treatment of growth factors. 46 Nuclear PI3K/Akt signalling is also involved in cell survival, cell cycle progression control, cell differentiation, mRNA processing and exportation, DNA repair and tumorigenesis. 47 But how Akt regulates cx43 expression remains unclear. There may be several potential mechanisms as below. Firstly, the transcription factor CREB (cAMP response element-binding protein) is phosphorylated and activated by Akt on serine 133. 48 Once phosphorylated on serine 133, CREB can induce transcription of connexin 43 gene. 48 So we guess CTGF may induce Cx43 expression through PI3K/Akt signalling pathway to activate the CREB phosphorylation. Secondly, the half-life of Cx43 mRNA was shortened in both LY294002-treated cells and DN (dominant-negative)-Akt-expressing-MC3T3E1 cells, indicating Akt is associated the stability of mRNA. 39 The AdenyMte/uridylate-rich elements (ARE) are located in the 3' untranslated region (3'UTR), interacting with ARE-binding proteins (ARE-BPs), then regulating the mRNA stability of short-lived mRNAs.
Active Akt1 can directly phosphorylate ARE-BPs, which could impair the decay of ARE containing transcripts. 49,50 Mouse and rat Cx43 3'UTRs contain four AREs and human Cx43 3'UTR contains five AREs, so Cx43 mRNAs may escape from the mRNA decay machinery by the phosphorylation of ARE-BPs through PI3K/Akt activity.
Consequently, CTGF may enhance PI3K/Akt activity to phosphorylate ARE-BPs, then helping increase the stability of Cx43 mRNAs. At last, Akt could also phosphorylate Connexin43 protein on Ser373 to enlarge gap junctions by eliminating interaction between Cx43 and zona occludens-1 (ZO-1). 51 More detailed researches need to be confirmed the above mechanism assumption. Additionally, our results showed that inhibition of PI3K/Akt pathway by LY294002 blocked the effects of CTGF on p-Akt and Cx43-gap junction activation, indicating the PI3K/Akt pathway mediated the stimulatory effect of CTGF on intercellular communication.
In conclusion, we demonstrated for the first time that CTGF promoted the Cx43 expression, gap junction formation as well as proliferation in chondrocytes with the involvement of PI3K/Akt signalling pathway ( Figure 6). These findings suggested that CTGF actively participated in gap junction intercellular communication in chondrocytes.

ACK N OWLED G EM ENT
This work was supported by the funding of NSFC grants (81600840, 81771047) to Jing Xie.

CO N FLI C T O F I NTE R E S T
The authors report no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
No publicly available data or shared data are cited. All original data supporting the conclusion of the current study are available from the corresponding author on request.