Expression and function of CCN2-derived circRNAs in chondrocytes

Cellular communication network factor 2 (CCN2) molecules promote endochondral ossification and articular cartilage regeneration, and circular RNAs (circRNAs), which arise from various genes and regulate gene expression by adsorbing miRNAs, are known to be synthesized from CCN2 in human vascular endothelial cells and other types of cells. However, in chondrocytes, not only the function but also the presence of CCN2-derived circRNA remains completely unknown. In the present study, we investigated the expression and function of CCN2-derived circRNAs in chondrocytes. Amplicons smaller than those from known CCN2-derived circRNAs were observed using RT-PCR analysis that could specifically amplify CCN2-derived circRNAs in human chondrocytic HCS-2/8 cells. The nucleotide sequences of the PCR products indicated novel circRNAs in the HCS-2/8 cells that were different from known CCN2-derived circRNAs. Moreover, the expression of several Ccn2-derived circRNAs in murine chondroblastic ATDC5 cells was confirmed and observed to change alongside chondrocytic differentiation. Next, one of these circRNAs was knocked down in HCS-2/8 cells to investigate the function of the human CCN2-derived circRNA. As a result, CCN2-derived circRNA knockdown significantly reduced the expression of aggrecan mRNA and proteoglycan synthesis. Our data suggest that CCN2-derived circRNAs are expressed in chondrocytes and play a role in chondrogenic differentiation. Graphical abstract Production and role of CCN2-derived RNAs in chondrocytes


Introduction
Cellular communication network factor 2 (CCN2) (Perbal et al. 2018) is a secreted protein expressed in various mesenchymal cells such as chondrocytes, fibroblasts, osteoblasts, and vascular endothelial cells (Chaqour 2020;Kubota and Takigawa 2011;Twigg 2018).In chondrocytes, CCN2 promotes cell proliferation and differentiation (Nakanishi et al. 2000;Nishida et al. 2017Nishida et al. , 2008)).CCN2 belongs to the CCN family, which consists of six members from CCN1 to CCN6.CCNs are composed of four modules starting from the N-terminal insulin-like growth factor binding protein-like module (IGFBP), followed by the von Willebrand factor type C module (VWC), the thrombospondin-1 type 1 repeat module (TSP1), and the C-terminal cystine knot module (CT) (Kubota and Takigawa 2011;Leask 2020;Perbal 2018).The mRNAs of CCNs are characterized by the retention of a long 3′-untranslated region (UTR).The half-lives and translation efficacies of these CCN mRNAs are controlled by the ciselements in the 3′-UTR, typically miRNA targets (Kubota and Takigawa 2015).Structurally, CCN5 is the only exception, in which the CT module is not encoded, and a short 3′-UTR is included (Lau 2016;Perbal 2013).In the case of CCN2, multiple miRNA targets are indicated in the 3′-UTR.Since CCN2 is highly expressed in a variety of tissues during development, the production and significant role of the circular RNAs (circRNAs) derived from CCN2 in these tissues are suspected.
Circular RNAs were first reported in 1976 as viroids, the smallest plant pathogens (Sanger et al. 1976).Subsequently, various circRNAs were discovered in eukaryotic cells in several organisms including mammals (Capel et al. 1993;Ford and Ares 1994;Grabowski et al. 1981).Circular RNAs are generated via back-splicing during the maturation of linear pre-mRNA and forms rings containing one or more exons without poly-A tails (Chen and Yang 2015).Moreover, it is known that multiple circRNAs can be generated from a single pre-mRNA, such as RAN binding protein 17 (RANBP17) pre-mRNA, depending on the combination of exons and introns (Chen and Yang 2015;Szabo and Salzman 2016).The functions of circRNAs are diverse (Bi et al. 2018;Du et al. 2017Du et al. , 2016)).For example, circSMARCA5, which is the SWI/SNF-related, matrix-associated, actin-dependent regulator of chromatin subfamily A member 5-derived cir-cRNA, performs RNA-DNA hybridization and transcription regulation (Conn et al. 2017;Xu et al. 2020).In contrast, ciRS-7 (hsa_circ0002484) functions as a miRNA sponge, binding to miRNAs and inhibiting their function (Memczak et al. 2013;Thomson and Dinger 2016;Zhong et al. 2019;Zou et al. 2019).
Among the CCNs, no circRNAs produced from CCN3-6 pre-mRNAs have been reported.On the other hand, CCN1 and 2 have been shown to yield a few circRNAs.In particular, CCN2-derived circRNA is known to be expressed in human vascular endothelial cells (http:// www.circb ase.org) (Fig. 1).In chondrocytes, however, the biological functions or presence of such CCN2-derived circRNA remains unclear.In this study, we investigated the expression and biological function of CCN2-derived circRNAs in human and murine chondrocytic cells.Herein, we show that CCN2derived circRNAs were expressed in chondrocytes, but their structures were different from those previously reported in other cell types (http:// www.circb ase.org).Furthermore, our results suggest that CCN2-derived circRNA plays a role in chondrocyte differentiation.

Cell culture
Human chondrocytic cell line HCS-2/8 cells that stably retain mature chondrocytic phenotype and murine chondrocyte progenitor cell line ATDC5 cells were cultured in Dulbecco′s Modified Eagle's Medium (DMEM) and alphamodified Eagle's minimum essential medium (αMEM) containing 10% fetal bovine serum (FBS), respectively.All cells were cultured at 37 °C in humidified air with 5% CO 2 .To induce ATDC5 cell differentiation, the culture medium was replaced with fresh αMEM containing 5% FBS and ITS (insulin-transferrin-selenium; Corning, Corning, NY, USA) every 3 days after the cells reached confluence.

RNA extraction
HCS-2/8 cell RNA was harvested 3 days after cell seeding.ATDC5 cell RNA was harvested every 3 days from 0 to 21 days after differentiation induction.RNA extraction from HCS-2/8 and ATDC5 cells was performed with the RNeasy® Mini Kit (Qiagen, Hilden, Germany) according to the protocol provided by the manufacturer.

Reverse transcription (RT) polymerase chain reaction (PCR)
Five hundred nanograms of RNA was reverse-transcribed with PrimeScript cDNA synthesis kits (Takara, Japan).Thereafter, PCR was performed according to the protocol of Quick Taq HS DyeMix (Toyobo, Osaka, Japan).The number of PCR cycles was within the range of 30-40 cycles in each experiment.For circRNA detection, primer sets were designed to specifically amplify the segments containing the back-splice junctions of human and murine CCN2-derived circRNAs (Table 1).

Agarose gel electrophoresis
PCR products were separated via electrophoresis on 1.5% or 2% agarose gels with Tris-acetate-EDTA buffer containing ethidium bromide and visualized with a UV transilluminator (Kurabo, Osaka, Japan).The intensities of the visualized bands were quantified using the image processing software ImageJ (https:// imagej.nih.gov/ ij/).

DNA sequencing
The bands containing the target DNA fragments were excised from the agarose gels.The DNA was purified according to the protocol of the QlAquick Gel Extraction Kit (Qiagen).The purified DNA was ligated into pGEM®-T Easy Vector (Promega, Madison, WI, USA) using the attached ligase mixture.Thereafter, plasmid DNA amplified with Escherichia coli was extracted using the GenEluteTM Plasmid Miniprep Kit (Merck).The nucleotide sequences of the plasmid DNA from the Sp6 and T7 primer binding sites were determined using a commercial service (FASMAC, Atsugi, Japan).The obtained nucleotide sequences were analyzed with the online tools and databases provided by the National Center for Biotechnology Information (NCBI: https:// www.ncbi.nlm.nih.gov).

RNAi
A 27-mer siRNA duplex containing 2 deoxynucleotides (5′-AAU AUU GUG UGU GUG ACG AGC CC dAdA-3′and 5′-UUG GGC UCG UCA CAC ACA CAA UAU UUA-3′) was designed for the specific knockdown of the CCN2-derived circRNA (A), targeting the back-splice junction.This siRNA duplex and a negative control siRNA were obtained from Nippon Gene (Tokyo, Japan).The siRNA duplex was transfected into HCS-2/8 cells (400,000 cells per well in 6-well plates) using siPORT™ NeoFX™ Transfection Agent (Thermo Fisher Scientific, Waltham, MA, USA) and incubated at 37 °C in humidified air with 5% CO 2 .After 24 h, the medium was replaced with a fresh one, and the cells were incubated for another 24 h.Thereafter, the RNAs were collected for subsequent analyses.

Quantitative real-time PCR (qPCR)
The real-time quantitative monitoring of PCR amplification was performed using the SYBR Green Realtime PCR Master Mix (Toyobo) according to the manufacturer's protocol.The fluorescent signals were detected and analyzed using the StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific).The primers used for the qPCR analysis are described in Table 2.

Glycosaminoglycans (GAG) assay
After 72 h of HCS-2/8 cell culture, the culture supernatants were collected.Cellular proteins were extracted by 2%Triton X-100/PBS.The lysate was treated at 65 °C for 90 min with 1 mg/ml actinase E in 0.2 M Tris-HCl (pH 7.5) and 5 mM CaCl 2 .Twenty microliters of the supernatant or cell lysate were added to 200 μl of 1,9 dimethylmethlyene blue (DMMB) reagent (38.46 μM DMMB, 40.5 mM glycine, 27.38 mM NaCl, 0.6% acetic acid).Thereafter, the absorbance was read using a plate reader at a wavelength of 525 nm immediately.GAG amount was normalized to that of the total protein.The total protein amount was measured with the BCA protein assay kit (Thermo Fisher Scientific) according to the manufacturer's protocol.
Firefly luciferase activity was normalized to Renilla luciferase activity.

RNA analysis in silico
The secondary structure prediction of human CCN2 pre-mRNA was performed online with UNAfold (http:// www.unafo ld.org).Putative miRNA binding sites were searched

Statistical analysis
The results are presented as mean ± standard deviation.The statistical significance of the differences between the mean values was determined with Dunnett's test or Student's t-test.Differences between the mean values were considered significant at a P-value of < 0.05.
The expected size of the PCR product using this primer to amplify this circRNA was 615 bp.The results demonstrate that the primer set yielded a distinct PCR product from the RNAs even after the treatment with high concentrations of RNase R (Fig. 2b).However, contrary to our expectation, the size of the PCR product was approximately 200 bp.In addition, RT-PCR was also performed using another antisense (hCCN2IGFBP-BA) primer and sense (hCCN2UTR-BS) primer (UTR-IGFBP set), which were designed in the IGFBP region and the 3′-UTR of a known CCN2-derived circRNA, respectively (circBase ID, hsa_circ_0077864; circBanc ID: hsa_circ_CTGF_001) (Fig. 2a).The expected size of the PCR product using this primer set to detect this particular circRNA was 770 bp.However, contrary to our expectation, two different amplicon sizes of approximately 250 bp and 400 bp were detected in the RNAs treated with high or low concentrations of RNase R, respectively (Fig. 2b).To confirm that RNase R was working effectively under these conditions, RT-PCR for CCN3 mRNA was performed.The results show that bands of approximately 200 bp in length disappeared in the presence of 5 U and 10 U of RNase R (Fig. 2c).These results suggest that the PCR product in Fig. 2b is from RNase R-resistant CCN2-derived circRNA.Furthermore, the nucleotide sequence of the PCR product in Fig. 2b was determined by DNA sequencing, and the accurate size of the PCR product obtained with the UTR-TSP primer set (A in Fig. 2d) was 180 bp with a backsplice junction within the flanking "GTG TGT G" sequence (700th-802nd bases of VWC-and TSP1-encoding region and 1784th-1853rd bases of 3′-UTR in human CCN2 mRNA: accession: NM_001901.4)(Fig. 2d).On the other hand, the sizes of the PCR products B-1 and B-2 (Fig. 2d) obtained with the UTR-IGFBP primer set were 226 bp and 402 bp, respectively.The nucleotide sequences of the PCR products B-1 and B-2 from CCN2-derived circRNAs exhibited backsplice junctions within the flanking "GCCT" sequences.

Structural basis of CCN2 pre-mRNA that yields CCN2-derived circRNA (A)
To gain more insight into the mechanism that enables the back-splicing that forms the CCN2-derived circRNA (A), possible secondary structures of the CCN2 mRNA precursor were predicted employing the mFold server provided on the UNAfold website (http:// www.unafo ld.org).Because of a current limitation of the system, the prediction of the entire structure of the full-length CCN2 pre-mRNA was not possible.However, the secondary structure prediction of a 2400-base RNA segment including the back-splicing sites revealed a stable secondary structure that could gather the back-splicing donor (BSD) and acceptor (BSA) nearby (Fig. 3).Interestingly, the BSD is located in an unstable portion of a stem-loop, in which G-U base paring is prevented due to conformational hindrance, facilitating the initiation of back-splicing events.

Changes in Ccn2-derived circRNA expression levels with murine ATDC5 cell differentiation
To explore the function of Ccn2-derived circRNAs in chondrocyte differentiation, we employed a cell culture system with a murine chondrocyte progenitor cell line ATDC5 as a model system of chondrocyte differentiation.The chondrocytic differentiation of ATDC5 cells was observed by measuring the changes in the expression levels of chondrogenic differentiation marker genes via RT-qPCR (Fig. 4a).The results demonstrate that the expression levels of the type II collagen α1 chain gene (Col2a1) and aggrecan core protein gene (Acan), which are typical marker genes of mature chondrocytes, increased until day 18 and then decreased until day 21.The expression levels of type X collagen α1 chain (Col10a1), which is a terminal chondrocyte differentiation marker, increased along with chondrocyte differentiation and remained high until the end.Under these conditions, we investigated the expression of Ccn2-derived circRNAs.Unlike in human cells, no Ccn2-derived circRNAs have been reported in murine cells.Therefore, the antisense (mCcn2TSP-BA) and sense (mCcn2UTR-BS) primers used herein were designed in the TSP1-encoding region and 3′-UTR, respectively, so that they could detect possible orthologs of the known human CCN2-derived circRNA, hsa_circ_0077863 (Fig. 4b).
Agarose gel electrophoresis after RT-PCR revealed multiple bands, including when the RNAs treated with 10 U of RNase R were used for amplification (Fig. 4c), which suggested the involvement of multiple circRNA species.Furthermore, quantitative analysis of the photographic density of each band indicated that the 400 bp signal significantly increased on day 9 and day 12, whereas the 500 and 600 bp signals peaked on day 12 (Fig. 4d), that is to say, the expression of the circRNAs producing these signals increased during the early steps of maturation in the chondrocytes.These results suggest that these circRNAs may play certain roles in chondrocyte differentiation.

Structures of murine Ccn2-derived circRNAs
The nucleotide sequences of the murine PCR products were determined using Sanger DNA sequencing (Fig. 5).
The results indicate that multiple novel circRNAs were present in murine chondrocytes.In contrast to the products detected in human cells, these PCR products were characterized by variations in the nucleotide sequences in comparison with that deposited in GenBank.The PCR products of 625 bp (Fig. 5a), 483 bp (Fig. 5b), and 516 bp (Fig. 5c) in length shared a common AACA insertion between the 2132nd and 2133rd bases in murine Ccn2 mRNA (accession: NM_010217.2),suggesting a genetic variation.Moreover, these PCR products displayed ATA TAT (1930, ATA TAT AT (1930-1937th bases), and ATA TAT ATAT (1930-1939th bases) deletions in the same location, respectively, indicating instability in the dinucleotide repeats on the RNA.Regarding the back-splice junctions, the murine circRNAs showed more diversity.The back-splice junctions of several PCR products (Fig. 5a, b, d, and e) were located within the indicated 3-6 nucleotides.For example, in the case of Fig. 5f, back-splicing occurred within the trinucleotide that originated either from the 3′-end of the 3′-UTR or the 5′-end of the TSP1-encoding exon.In contrast with these four PCR products, the boundaries of the back-spliced exons were distinct in the other two amplicons (Fig. 5c and f).

CCN2-derived circRNA knockdown in HCS-2/8 cells
To investigate the function of a CCN2-derived circRNA in chondrocyte differentiation in more detail, human CCN2derived circRNA (A) was knocked down via RNA interference in HCS-2/8 cells (Fig. 6a).For this purpose, we designed and used a 27-mer siRNA duplex targeting the back-splice junction.This siRNA with two deoxynucleotides exerts high knockdown efficacy after being processed by the DICER complex.The RT-PCR results showed that CCN2-derived circRNA (A) was efficiently knocked down in HCS-2/8 cells (Fig. 6a).Importantly, no significant change was observed in the CCN2 mRNA expression (Fig. 6b) and CCN2 protein production (Fig. 6c) levels upon the circRNA knockdown in HCS-2/8 cells.These results indicated that the circRNA-specific knockdown was successful.

Effect of CCN2-derived circRNA knockdown on chondrocytic phenotype in HCS-2/8 cells
Under the abovementioned conditions, the expression levels of the chondrocyte differentiation marker genes ACAN, SOX9, COL2A, and COL10A were evaluated using RT-qPCR (Fig. 7a).The results indicate that the expression level of ACAN mRNA significantly decreased in CCN2derived circRNA-knocked-down HCS-2/8 cells compared with the control cells (Fig. 7a).However, there was no change in the expression levels of SOX9, COL2A1, and COL10A1 (Fig. 7a).These results suggest that CCN2derived circRNA (A) has a function in the regulation of ACAN mRNA expression specifically.Of note, this cir-cRNA contains a miR-181-5p target in its 3′-UTR segment (Ling et al. 2020), and our previous study clearly showed that exogenous miR-181a-5p drastically decreased the ACAN mRNA in HCS-2/8 cells, strongly suggesting their direct interaction (Sumiyoshi et al. 2013).Therefore, to investigate the suppression mechanism of ACAN mRNA expression by knockdown of CCN2-derived circRNA (A), a reporter gene assay was performed using a firefly luciferase reporter that contains a target sequence of miR-181-5p (Sumiyoshi et al. 2013).The result of the reporter gene assay showed that luciferase activity was significantly decreased in CCN2-derived circRNA-knocked-down HCS-2/8 cells (Fig. 7b).These results suggest that CCN2derived circRNA (A) enhances ACAN mRNA expression by the function as a miRNA sponge for miR-181-5p.At
Computed Gibbs free energy of this structure was -707.33 kcal/mol at 37 °C.The primary structures of BSD and BSA areas are made visible in the enlarged views of the boxed areas.Watson-Crick base pairing "G-C" and "A-U" are indicated with red and blue bars, respectively, whereas wobble "G-U" base pairing is shown in green.Note that these areas gathered in the stable secondary structure both contain unstable bulged structures, and a few hydrogen bonds between G and U around BSD may not have formed due to steric hindrance a protein level, no obvious difference was appreciable in the production of chondrocyte differentiation marker proteins by immunoblotting compared to the control (Fig. 7c).However, in the GAG assay, the amount of GAG in both culture supernatant and cell lysate was significantly reduced in CCN2-derived circRNA-knocked-down HCS-2/8 cells compared to the control cells (Fig. 7d).
Considering that aggrecan is the major proteoglycan in cartilage and possesses tremendous amount of glycosaminoglycan chains, this result suggests that knockdown of the CCN2-derived circRNA (A) decreases aggrecan production.
Fig. 4 Changes in Ccn2-derived circRNA expression along with murine ATDC5 cell differentiation.a RNAs were collected from ATDC5 cells every 3 days during chondrocytic differentiation, and the expression levels of chondrocyte differentiation marker genes were measured with RT-qPCR.Acan and Col2a1 expression levels increased from day 1 until day 18.Col10a1 expression level increased gradually and was retained at higher levels until the end.All of the values were standardized against the expression level on day 0. RNAs from 3 independent cultures were analyzed.Asterisks represent significant differences at P < 0.05.b The primers were designed to specifically detect the back-spliced segments of the 3′-UTR and TSP1 exon in murine Ccn2-derived circRNAs.c Agarose gel electrophoresis after PCR using the primers in "b" showed RNase R-resistant, multiple bands.Four bands subjected to subsequent analysis are indicated with arrowheads d The photographic densities of bands at approximately 200, 400, 500, and 600 bp in "c (arrowheads)" were quantified using Image J. The bands of approximately 200 bp showed no significant changes.The intensities of the bands approximately 400 bp in length significantly increased 9 and 12 days after the induction of differentiation.The bands of approximately 500 and 600 bp displayed significantly higher intensities 12 days after the induction of differentiation.All of the values were standardized against the expression level on day 0. The values are presented as averages ± SD. *P < 0.05

Discussion
CCN2-derived circRNAs are known to be expressed in several types of human cells including vascular endothelial cells, as revealed in previous studies (http:// www.circb ase.org).However, in chondrocytes, the presence of CCN2derived circRNAs has not been reported.In this study, we revealed the expression of CCN2-derived circRNAs in human chondrocytic HCS-2/8 cells and murine chondroblastic ATDC5 cells.To date, no Ccn2-derived circRNAs have been detected in any murine cell.Therefore, this is the first study in which Ccn2-derived circRNAs have been discovered.Moreover, one of these CCN2-derived circRNAs was revealed to play a role in chondrocyte differentiation.
In our study, the identified portions of circRNAs contained only two exon fragments around the back-splice junction.However, considering them alongside the structure of the known circRNA (circBase ID, hsa_circ_0077863: circBank ID, hsa_circ_CTGF003), the novel CCN2-derived circRNA (A) is estimated to contain a 3′-UTR fragment,  The production level of CCN2 protein was examined by immunoblotting.Beta-actin was used as a loading control entire TSP1-and CT-module-encoding exons, and a fragment of the VWC-module-encoding exon (Fig. 8a).Similarly, the results suggest that the other two novel human CCN2-derived circRNAs (B-1) and (B-2) retain entire regions encoding the VWC, TSP1, and CT modules between the fragments of the IGFBP-encoding exon and the 3′-UTR, as observed in the other known circRNA (circBase ID, hsa_ circ_0077864; circBank ID: hsa_circ_CTGF001 (Fig. 8a).
The biosynthetic pathway of these CCN2-derived cir-cRNAs remains unclear.However, circRNAs are generally known to be produced from linear pre-mRNA, forming a circular structure via the fusion of a downstream 3′ back-splicing donor (BSD) with an upstream 5′ backsplicing acceptor (BSA) (Chen and Yang 2015).Although the mechanism of BSD-BSA fusion is poorly understood, several models, such as those of intron-pairing-driven cyclization (Zang et al. 2020) and cyclization carried out by RNA-binding proteins (RBPs) (Chen et al. 2021a, b), are currently being considered.In the intron-pairingdriven cyclization model, when a pair of inverse repetitive elements such as Alu elements, which are abundant in humans, or non-repetitive complementary sequences are present in adjacent introns, the downstream donor and upstream acceptor sites are drawn to each other, and cyclization occurs (Chen and Yang 2015;Ivanov et al. 2015;Zhang et al. 2016).The other "cyclization by RBPs" model proposes that RBPs bind to introns near the splice site and subsequently congregate the BSD and BSA (Conn et al. 2015).In our study, we discovered a functional CCN2-derived circRNA (A) that was stably expressed in human HCS-2/8 cells.According to the analysis using RepeatMasker (https:// www.repea tmask er.org), the human CCN2 gene was found to contain no pairs of short interspersed nuclear elements (SINEs), such as Alu elements, which would facilitate back-splicing (Fig. 8b).Similarly, even fewer repetitive elements were identified in the murine CCN2 gene.However, the secondary structure formation analysis with mFold suggested that BSD and BSA could be gathered nearby in a stable structure formed in CCN2 pre-mRNA.Therefore, we suspect that such a complex secondary structure enables the stable formation of CCN2-derived circRNA (A), probably with the assistance of particular RBPs stabilizing this structure.The other human CCN2-derived circRNA (B) and all of the murine Ccn2-derived circRNAs were found to be differentially expressed, suggesting that multiple circRNAs can be produced from the same pre-mRNA that is folded into multiple structures depending on the cellular conditions.As such, the production of Ccn2-derived circRNAs in ATDC 5 cells was regulated along with chondrocytic differentiation.RBPs are supposed to play significant roles in the selective stabilization of RNA structures therein.
Circular RNAs have various functions, one of which is acting as miRNA sponges (Hansen et al. 2013;Misir et al. 2022).Circular RNAs that function as miRNA sponges have miRNA binding sites in their nucleotide sequences.Therefore, such circRNAs can regulate target mRNA expression via the inhibition of miRNA activity by adsorbing miRNAs (Ebert et al. 2007).For example, ciRS-7, an oncogenic circRNA, plays an important role in the development and progression of various tumors by functioning as a miR-7 sponge (Chen et al. 2021a, b).In this study, human CCN2derived circRNA (A) knockdown significantly decreased the expression level of ACAN mRNA, which is a chondrocyte differentiation marker gene that encodes aggrecan core protein (Fig. 7c).That is to say, CCN2-derived circRNA was found to promote the expression of the ACAN mRNA that supported the chondrocytic phenotype.Importantly, the 3′-UTR of CCN2-derived circRNA contains a binding site for miR-181-5p that could inhibit ACAN mRNA expression (Ling et al. 2020;Fig. 8c).Indeed, our present data indicated that binding of miR-181-5p to a target in the reporter mRNA was competed by this circRNA (Fig. 7).Therefore, CCN2derived circRNA may inhibit miR-181-5p as a miRNA sponge and subsequently promote ACAN mRNA expression (Fig. 8d) by preventing its degradation.In contrast to ACAN, human CCN2-derived circRNA (A) did not alter the mRNA levels of SOX9 and COL2A1 (Fig. 7a).However, all of the three mRNAs were shown to be upregulated by a miR-181a-5p inhibitor in a previous study (Ye et al. 2022).In this study with chondrogenic mesenchymal stem cells, the effect of the miRNA inhibitor was the strongest on ACAN mRNA, which is consistent with our results.The sensitivity to a miRNA varies among target RNAs, and thus effect of the circRNA on miR-181-5p was detectable only through ACAN mRNA with the highest sensitivity (Fig. 7a).This idea also accounts for why CCN2 mRNA level was not affected by knocking down of the CCN2-derived circRNA.It should be also noted that although ACAN is a target of miR-181-5p in human chondrocytes, no miR-181-5p target is found in murine Acan mRNA by Targetscan.Considering together that circRNAs from human CCN2 (Fig. 2) and those from murine Ccn2 (Fig. 4) are structurally different, circRNAs are produced and regulating chondrocyte differentiation in a species-specific fashion.Further investigation is currently ongoing to clarify this issue.
In conclusion, this study revealed the expression of novel CCN2-derived circRNAs in human and murine chondrocytic cells.Moreover, the results indicate that these circRNAs play important roles in chondrocyte differentiation.Among chondrocyte markers, ACAN was the only one that was under the regulation of a particular cir-cRNA.This finding raises the intriguing possibility that distinct circRNAs may regulate different chondrocyte differentiation markers.Therefore, the biological function of other CCN2-derived circRNAs should be further investigated in future research.
otherwise in a credit line to the material.If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Fig. 2
Fig. 2 Detection of CCN2-derived circRNAs.a Positions of the primers designed in CCN2.The primer hCCN2TSP-BA antisense primer was designed on the TSP1 exon for the detection of a known CCN2-derived circRNA (hsa_circ_0077863), whereas hCCN2UTR-BS (sense primer) was designed on the UTR that is common between two CCN2-derived circRNAs (hsa_circ_0077863 and hsa_ circ_0077864).The other hCCN2IGFBP-BA antisense primer was designed on the IGFBP exon to detect hsa_circ_0077864.b Detection of novel circRNAs in HCS-2/8 cells.RNAs of human chondrocytic HCS-2/8 cells were treated with 0 U (control), 1 U, 5 U, or 10 U of RNase R, followed by reverse transcription (RT) with random hexamers, PCR, and electrophoresis.The experiments were repeated twice, as indicated by Rep. 1 and Rep. 2. RT-PCR with hCCN2UTR-BS and hCCN2TSP-BA (UTR-TSP); a distinct band approximately 200 bp in length was stably observed regardless of the RNase R concentration.In contrast, with the primer set composed of hCCN2UTR-BS and

Fig.
Fig. CCN2-derived circRNA (A) knockdown in HCS-2/8 cells.a Knockdown of CCN2-derived circRNA (A) was confirmed with RT-PCR via 3 independent sets of experiments.Specific signals of CCN2-derived circRNA (A) are indicated with a rectangle.b The expression level of CCN2 mRNA was measured using RT-qPCR.cThe production level of CCN2 protein was examined by immunoblotting.Beta-actin was used as a loading control

Fig. 7 Fig. 8 a
Fig.7Effect of CCN2-derived circRNA (A) knockdown on chondrocytic phenotype of HCS-2/8 cells.a The expression level of ACAN mRNA significantly decreased in the CCN2-derived circRNA (A) knocked-down cells compared with the control.No significant differences were observed in the expression levels of SOX9, COL2A1, and COL10A1.The values are presented as the averages ± SD of 9 independent cultures.*P < 0.05.b Reporter gene assay was performed by using pGL3-181TS in control and CCN2-derived circRNA (A)-knocked-down HCS-2/8 cells.Luciferase activity was significantly reduced in CCN2-derived circRNA (A)knocked-down HCS-2/8 cells.The values are presented as the averages ± SD of 3 independent cultures.*P < 0.05.c The production levels of aggrecan, SOX9, type II collagen proteins were examined by immunoblotting.Beta-actin was used as a loading control.d GAG assay was performed to measure the GAG levels in the culture supernatant and cell lysate obtained from CCN2-derived circRNA (A) knockeddown cells and the control cells.The values are presented as the averages ± SD of 6 independent cultures.*P < 0.05

Table 1
Primers for amplifying CCN2-and Ccn2-drived circRNAs BS, sense primer for detecting back-spliced product; BA, anti-sense primer for detecting back-spliced product

Table 2
Primers for real-time PCR S, sense primer; AS, anti-sense primer