SOX9 is a regulator of ADAMTSs-induced cartilage degeneration at the early stage of human osteoarthritis

Osteoarthritis and Cartilage 23 (2015) 2259e2268

SOX9 is a regulator of ADAMTSs-induced cartilage degeneration at the early stage of human osteoarthritis Q. Zhang y ¶ a, Q. Ji y a, X. Wang x, L. Kang k, Y. Fu y, Y. Yin y, Z. Li y, Y. Liu y, X. Xu z **, Y. Wang y * y Department of Orthopaedics, The General Hospital of Chinese People's Liberation Army, Beijing 100853, China z Beijing Institute of Biotechnology, Beijing 100850, China x Department of Neurology, The General Hospital of Chinese People's Liberation Army, Beijing 100853, China k Department of Nuclear Medicine, Peking University First Hospital, Beijing 100034, China ¶ Department of Orthopaedic Surgery, Royal Liverpool University Hospital, Prescot Street, Liverpool, UK

a r t i c l e i n f o

s u m m a r y

Article history: Received 26 January 2015 Accepted 29 June 2015

Objective: To identify whether cartilage master regulator SRY-related protein 9 (SOX9) mediates A disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS) dysregulation during osteoarthritis (OA) cartilage degeneration. Method: Twenty-two randomly selected OA patients were evaluated using Outerbridge Classification via arthroscopy. Haematoxylineeosin (HE), Safranin O and Masson staining were performed for the histopathological assessment. The expression of ADAMTSs, collagen 2A1 (COL2A1), aggrecan (ACAN), cartilage oligomeric matrix protein (COMP) and SOX9 were examined using real-time quantitative Polymerase Chain Reaction (PCR) (RT-qPCR) and western blotting analysis. Immunohistochemistry (IHC) analysis was performed to investigate the production of ADAMTSs in cartilage tissues. The association between SOX9 production and ADAMTSs, COL2A1, ACAN, and COMP expression was established by full-depth cartilage biopsies. Results: ADAMTSs expression levels were repressed at stage 1, while a significant increase was observed at the progressive stage of OA. SOX9 was upregulated at stage 1 and suppressed at a later stage of cartilage development, particularly in cartilage with severe damage. In addition, SOX9 repressed the expression of ADAMTSs and promoted COL2A1, ACAN and COMP expression in human chondrocytes. SOX9 was recruited to the promoters of ADAMTS-4 and ADAMTS-7. SOX9 expression was negatively correlated with ADAMTSs production and was positively associated with COL2A1, ACAN and COMP expression. Inhibition of ADAMTSs markedly increased the production of COL2A1, ACAN and COMP in chondrocytes isolated from the early stage of OA. Conclusions: These findings indicated that SOX9 upregulation might mediate ADAMTSs suppression at the early stage of human OA. In addition, SOX9 could be used as a potential therapeutic agent for human OA at an early stage. © 2015 The Authors. Published by Elsevier Ltd and Osteoarthritis Research Society International. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Osteoarthritis Cartilage Chondrocyte ADAMTS SOX9

Introduction * Address correspondence and reprint requests to: Y. Wang, Department of Orthopedics, General Hospital of Chinese People's Liberation Army, Beijing 100853, China. Tel: 86-010-66939439. ** Address correspondence and reprint requests to: X. Xu, Beijing Institute of Biotechnology, Beijing 100850, China. Tel: 86-010-66931830. E-mail addresses: [email protected] (X. Xu), [email protected] (Y. Wang). a These authors contributed equally to this work. The General Hospital of Chinese People's Liberation Army and Beijing Institute of Biotechnology contributed equally to this work.

Osteoarthritis (OA) is a chronic and complex multifactorial joint disease that is characterized by the progressive degradation of articular cartilage and joint inflammation1e3. The extracellular matrix (ECM), a network of proteins and macromolecules, provides both strength and nutrients for chondrocytes. The destruction of ECM components, such as the proteoglycan aggrecan (ACAN), cartilage oligomeric matrix protein (COMP) and type II collagen, usually induce OA4,5. These processes involve the coordinated

http://dx.doi.org/10.1016/j.joca.2015.06.014 1063-4584/© 2015 The Authors. Published by Elsevier Ltd and Osteoarthritis Research Society International. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

2260

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

regulation of inducers and inhibitors during OA progression6. However, the underlying pathophysiological process of the early stage of OA has not been completely elucidated. A disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS) family and matrix metalloproteinase (MMP) play an important role in OA pathogenesis7e9 by degrading matrix and nonmatrix substrates and by shifting the balance from the anabolic toward the catabolic state with gradual progressive cartilage loss. Limited studies have shown that ADAMTSs can degrade the ECM and are highly expressed in OA cartilage. ADAMTS-4 and ADAMTS-5 have been reported to be crucial for degrading the matrix10. ADAMTS-7 forms a positive feedback loop with tumour necrosis factor (TNF)-a in the pathogenesis of OA11. ADAMTS-12 contributes to the progressive loss of cartilage and joint dysfunction12. However, the molecular mechanisms underlying the activity of ADAMTSs at the early stage of OA are poorly characterized. SRY-related protein 9 (SOX9) is an essential transcription factor regulating the expression of many ECM genes, such as ACAN and collagen type II13e15. Haploinsufficiency of SOX9 can induce campomelic dysplasia, which is a severe syndrome caused by inadequate cartilage formation during development, underlying its importance to the chondrocyte phenotype16,17. Nevertheless, little is known regarding SOX9 expression at the early stage of OA as well as with regard to the relationship between SOX9 and ADAMTSs during cartilage degeneration in human OA pathogenesis. In this study, we found that SOX9 upregulation mediated the suppression of ADAMTSs at the early stage. Furthermore, knockdown of SOX9 significantly enhanced ADAMTS-4, ADAMTS-5,

ADAMTS-7 and ADAMTS-12 production and downregulated collagen 2A1 (COL2A1), ACAN and COMP expression in human chondrocytes, indicating that SOX9 may play an important role in cartilage protection. Taken together, these findings comprehensively elucidate their properties at different stages of OA and open new possibilities for new therapeutic strategies against OA. Methods Reagents Antibodies against ADAMTS-5 (A6727) were obtained from SigmaeAldrich (St. Louis, MO, USA). Antibodies against ADAMTS-4 (sc25582), ADAMTS-7 ADAMTS-12 (sc-25583), COL2A1 (sc-28887), ACAN (sc-25674), b-actin (sc-47778) and SOX9 (sc-20095) were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Antibodies against COMP (ab128893) were obtained from Abcam (Cambridge, MA, USA). Antibodies against ADAMTS-7 (HPA034581) were obtained from Atlas antibodies AB (St. Louis, MO, USA). Specimen collection and patients Human articular cartilage specimens were obtained from knee joints at the time of total knee replacement surgery from patients [n ¼ 22, mean ± 95% confidence intervals (CIs), age: 55.7 years (52.8e56.9), 54.5% men] graded using an arthroscopic diagnosis. Relatively normal cartilage evaluated as stage 0 was regarded as the control group. Tissues used as controls were already significantly

B

A

Anterial Stage 4

Stage 3

X

Stage 1 Stage 0

Medial

a Stage 0

Stage 1

Stage 2

Posterial

Y Stage 3

Stage 4

Masson

Safranin-O/ Fast green

H&E

C

b

Lateral

Stage 2

Fig. 1. Extraction and stage of the OA tissue. The grades of articular cartilage were classified using arthroscopy observations and were marked by pinning points (A, arrows). Marked points, located in the medial tibial plateau (M) or lateral tibial plateau (L), were recorded from anterior to posterior (X1 2 3) and from medial to lateral (Y1 2 3). For example, (a) the cartilage of this position was classified as Grade II, and this position was recorded as (M:X3Y2, II), which was the most posterior and the second medial point among the three markers in the medial tibial plateau (A). Cartilage was obtained from the tibial plateau by drilling holes of the same diameter (4.5 mm) with the reference point marked using arthroscopic observation (B). H&E staining (upper panel), Safranin O/Fast green staining (middle panel), and Masson staining (lower panel) of articular cartilage from control samples and OA samples (C). Scale bar, 50 mm. The sections shown are representative of OA tissue from one patient.

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

different from what was expected in tissues obtained from a healthy control. All of tissue samples were obtained from joints affected by OA and the areas of each individual patient's articular cartilage were selected and graded as stage 0e4 according to the Outerbridge Classification scheme. Cartilage specimens that included all of the cartilage layers and subchondral bone were separately harvested by drilling holes (4.5 mm) [Fig. 1 (B)], the locations of which were determined under arthroscopy [Fig. 1(A)]. The present study was reviewed and approved by the institutional ethics review board (No. 20090611-3) of the General Hospital of People's Liberation Army (Beijing, China). Written informed consent was obtained from each patient. Cell culture Human chondrocytes were isolated from autopsy donors that were graded macroscopically according to a modified Outerbridge scale. Only normal (Grade 0: normal and intact cartilage) or severe OA (Grade III: maximal fibrillation) cartilage was used for the studies. Chondrocytes were incubated in Dulbecco's modified Eagle's (DMEM) high glucose medium containing 10% foetal calf serum (FCS), 100 IU/mL penicillin and 100 mg/ml streptomycin at 37 C in an atmosphere of 5% CO2. First-passage chondrocytes at 85% confluence were used for all experiments. Transfection To construct the pXJ40-myc-tagged human Sox9, Sox9 cDNA was amplified using two primer sets (Forward 50 -CGGGATCCATGAATCTCCTGGACCCCTTCA-30 and Reverse 50 -CCGCTCGAGTCAAGGTCGAGTGAGCTGTGTG-30 ) that contained BamHI and XhoI restriction-enzyme sites. The PCR products were digested with BamHI and XhoI and cloned into the pXJ40-myc vector. Human chondrocytes were seeded at a density of 4  104 in 12- or 24-well plates. On the following day, cells were transfected with Sox9 using Lipofectamine 2000 (Invitrogen, CA, USA), according to the manufacturer's instructions: 1 day before transfection, the cells were plated in growth medium without antibiotics such that the cells were 80% confluent at the time of transfection. The optimal ratio of DNA/Lipofectamine 2000 in our system was 1:1.5. The solution was mixed gently and incubated for 20 min at room temperature, and the DNA/Lipofectamine 2000 complexes were then added to each well. After 4e6 h, the culture medium was changed. Cell lysates were collected for different types of experiments after 48 h of transfection. To monitor the transfection efficiency, the pEGFP-C1 plasmid (Clontech Laboratories, CA, USA) was transfected each time in a parallel well. Total RNA extraction, reverse transcription and real-time quantitative PCR (RT-qPCR) Total RNA from tissues or cultured chondrocyte samples containing miRNA was extracted and reverse-transcribed to cDNA using the RNeasy Mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer's protocol. Total RNA (2 mg) was reverse-transcribed in a total volume of 25 mL containing RT buffer (Takara), 200 U of Moloney murine leukemia virus Reverse Transcriptase (M-MLV RT) (Takara), 20 U of RNasin (Takara), and 0.2 mg of a random primer (Takara). The reaction was performed at 42 C for 60 min and 95 C for 10 min cDNA was stored at 20 C until further use. RTqPCR reactions were performed using the SYBR Premix Ex Taq Master Mix (2) (Takara, Japan) as previously described. The relative quantification value of the target, normalized to the control, was calculated using the comparative Ct method. The primers used for RT-qPCR analysis are shown in Table I.

2261

Western blotting analysis Proteins were extracted from tissues or chondrocytes using RIPA buffer (50 mM TriseHCl (pH 7.4), 150 mM NaCl, 20 mM EDTA, 1% Triton X-100, 1% sodium deoxycholate, 1% SDS and protease inhibitors, Wheaton Science) on ice for 30 min and analysed using SDSPAGE electrophoresis. Membranes were subsequently incubated with primary antibodies against b-actin (Santa Cruz, 1:500 dilution), ADAMTS-4 (Santa Cruz, 1:100 dilution), ADAMTS-5 (Sigma, 1:100 dilution), ADAMTS-7 (Atlas antibodies, 1:200 dilution), ADAMTS-12 (Santa Cruz, 1:100 dilution), SOX9 (Santa Cruz, 1:500 dilution), COL2A1 (Santa Cruz, 1:50 dilution) and ACAN (Santa Cruz, 1:50 dilution) overnight at 4 C, followed by incubation with a horseradish peroxidase conjugate. The immunocomplexes were visualized by chemiluminescence using an ECL kit (Amersham Biosciences). Small interfering RNA (siRNA) pLenti-H1 vector-containing siRNA constructs for ADAMTS-4, ADAMTS-5, ADAMTS-7, ADAMTS-12 and SOX9 were generated using recombinant PCR. The empty vector was treated as the control. The sequences of siRNAs were as follows: ADAMTS-4: 50 AAGCATCCGCAATCCTGTCAG-30 ; ADAMTS-5: 50 -AAGATAAGCGCTTAATGTCTT -30 ; ADAMTS-7: 50 -ACCTAA AGATCACGCACCA-30 ; ADAMTS-12: 50 -ACACATCACACACACCCAA-30 ; SOX9: 50 -CAGCGAACGCACATCAAGA-30 . To produce the lentivirus, HEK293T cells were co-transfected with recombinant lentiviral vectors (pLenti-H1/ADAMTS-4, pLenti-H1/ADAMTS-5, pLenti-H1/ ADAMTS-7, pLenti-H1/ADAMTS-12 and pLenti-H1/SOX9) and pPACK Packaging Plasmid Mix (System Biosciences) using the Megatran reagent (Origene) as previously described28. To infect human chondrocytes, lentiviruses were collected after 48 h of transfection and added to the medium of target cells with 8 mg/ ml polybrene (SigmaeAldrich). Chondrocytes were harvested for measurement after 72 h of transfection. Chromatin immunoprecipitation (ChIP) A ChIP assay was performed using a commercially available kit (Millipore, MA, USA) according to the manufacturer's instructions29. Briefly, chondrocytes isolated from autopsy donors were cross-linked in 3.7% formaldehyde, pelleted, and resuspended in lysis buffer. Cells were sonicated, followed by centrifugation to Table I Primers used for RT-qPCR Gene (Genebank accession number)

Sequence

ADAMTS-4 (NM_005099.4) ADAMTS-5 (NM_007038.3) ADAMTS-7 (NM_014272.3) ADAMTS-12 (NM_030955.2) Col2A1 (NM_001844.4) ACAN (NM_001135.3) COMP (NM_000095.2) SOX9 (NM_000346.3) b-actin (NM_001101.3)

50 -GGCCACCGGAGCATCTACTTGGCC-30 (S) 50 -GGCTGCAGTGGCCCCGCTGTAGC-30 (AS) 50 -GAACTATAGCGGTTGGAGCCACAG-30 (S) 50 -GCTGCCATGACTAGTGACAGAG-30 (AS) 50 -GACAGGGCTGCCCGAGGAAGACAG-30 (S) 50 -CAGCGTCTCGCAGAACCCGAAGG-30 (AS) 50 -GAAGACCAAGACCAATGTCTATG-30 (S) 50 -GTCTGGCAGAAACTGGCTGACAG-30 (AS) 50 -GCCTCGCGGTGAGCCATGATC-30 (S) 50 -CTCCATCTCTGCCACGGGGT-30 (AS) 50 -TAGAGGATGTGAGTGGTCTT-30 (S) 50 -TCCACTAAGGTACTGTCCAC-30 (AS) 50 -AACACGGTCACGGATGACGACTATG-30 (S) 50 -CACAGAGCGTTCCGCAGCTGTTC-30 (AS) 50 -GAGCTGAGCAGCGACGTCATCT-30 (S) 50 -GGCGGCGCCTGCTGCTTGGACA-30 (AS) 50 -ATCACCATTGGCAATGAGCG-30 (S) 50 -TTGAAGGTAGTTTCGTGGAT-30 (AS)

S sense, AS antisence.

2262

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

remove insoluble material. The supernatants were collected, and Pellet Protein G magnetic beads were added to the sample and incubated for 1 h at 4 C with antibodies against SOX9. Chromatin was collected, purified, and de-crosslinked at 62 C for 2 h and then at 95 C for 10 min. The precipitated DNA fragments were quantified using RT-qPCR analysis. ChIP analysis was performed using the “Fold Enrichment Method”, which was also called “relative to the no-antibody control”. Using this method, the ChIP signals were represented as the fold increase in signal of SOX9 relative to the IgG signal. The primers used for the ChIP assay are displayed in Table II.

of the RT-qPCR data was performed using one-way ANOVA with Tukey's post hoc test. Analysis of the statistical correlation between the expression of SOX9, ADAMTS-4, ADAMTS-5, ADAMTS-7, ADAMTS-12, COL2A1, ACAN and COMP was performed with tissues that were freshly isolated from 22 OA patients using Spearman rank analysis with GraphPad PRISM 6 (GraphPad, San Diego, CA, USA). Results Identification of different stages of OA using histopathological staining

Histopathological assessment Cartilage samples were fixed in 4% buffered paraformaldehyde for 48 h and then decalcified with buffered Ethylene Diamine Tetraacetic Acid (EDTA) (20% EDTA, pH 7.4). The tissues were then embedded in paraffin, sectioned and stained with haematoxylineeosin (HE), Safranin O/Fast green and Masson's trichrome. The histopathological changes were scored using the modified Mankin grading system in a blind manner18.

We primarily investigated the different stages of cartilage tissues. Human articular cartilage specimens were obtained from knee joints at the time of total knee replacement surgery from patients who were graded using an arthroscopic diagnosis. Knee cartilage was obtained by drilling holes [Fig. 1(B)], the locations of which were determined using arthroscopy [Fig. 1(A)]. The stages of OA were diagnosed using H&E, Safranin O/Fast green, and Masson staining [Fig. 1(C)].

Immunohistochemistry (IHC)

Expression of ADAMTSs at the early stage of OA

Articular cartilage sections were pretreated for 10 min with trypsin (0.05%) before treatment with 3% (vol/vol) H2O2 for 15 min. The sections were then blocked with 10% goat serum for 1 h at room temperature. After washing with Phosphate Buffered Saline (PBS), antibodies against ADAMTS-4 (sc-25582, Santa Cruz, 1:25 dilution), ADAMTS-5 (A6727, Sigma, 1:50 dilution), ADAMTS-7 (HPA034581, Atlas antibodies, 1:50 dilution) and ADAMTS-12 (sc25583, Santa Cruz, 1:50 dilution) were applied to the sections and incubated for 1 h at room temperature. The sections were then washed with PBS and incubated for 15 min with a biotinylated secondary antibody using a Histostain Plus kit (Invitrogen, Carlsbad, CA, USA). The sections were washed and incubated for 2 min with 3, 30 -diaminobenzidine (DAB) substrate. Using light microscopy, all of the IHC staining was assessed by two pathologists who were blind to the origin of the specimen. The German semiquantitative scoring system is widely accepted with consideration of the staining intensity and extent of the area used30, where 0 indicates no staining; 1 indicates weak staining; 2 indicates moderate staining; and 3 indicates strong staining. In addition, the percentage of staining was given a score of 0 (<5%), 1 (5e25%), 2 (25e50%), 3 (51e75%), or 4 (>75%). These two scores were multiplied, resulting in the final score. For ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12, we defined a score of 0 as negative and a score of 1e12 as positive.

To identify the different expression of the ADAMTSs in the control sample (stage 0) and early stage OA (stage 1) cartilage, we first evaluated the mRNA levels of ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 using RT-qPCR. We found that the mRNA levels of ADAMTS-4 and ADAMTS-5 [Fig. 2 (A and B)] as well as ADAMTS-7 and ADAMTS-12 [Fig. 2(C and D)] could be detected in both the control and OA human articular cartilage. Interestingly, the transcriptional levels of the ADAMTSs were significantly repressed in stage 1 compared with the control sample group [Fig. 2(AeD)]. To further assess the potential involvement of ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 in the early OA process, we performed IHC and western blotting analysis to analyse whether this trend also occurred at the protein level. Western blotting analysis confirmed that the expression of ADAMTS-4 and ADAMTS-5 [Fig. 3 (A and B)] as well as ADAMTS-7 and ADAMTS-12 [Fig. 3(C and D)] was remarkably decreased at the early stage of OA (stage 0 vs stage 1). IHC staining demonstrated that ADAMTS-4-, ADAMTS-5-, ADAMTS-7- and ADAMTS-12-positive chondrocytes were observed mainly in the superficial layers of the control sample and OA human articular cartilage and that ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 were mainly present in the cytoplasm [Fig. 3(E)]. Similarly, the expression of ADAMTSs was also suppressed at stage 1 (P < 0.01) [Fig. 4 ], which was consistent with the mRNA expression levels of ADAMTSs at stage 1.

Statistical analysis Expression of ADAMTSs at the later stage of OA The data were expressed as the mean ± 95% CIs. Statistical analysis was performed using KruskaleWallis test for multiple comparison, and P < 0.05 was indicated statistical significance. The ManneWhitney U test or Wilcoxon signed-rank test was used to account for multiple testing. All experiments were performed using samples obtained from at least three different donors with duplicate or triplicate replication as indicated in the figure legends. Evaluation

Next, we examined the mRNA expression of ADAMTSs at the later stage of OA. RT-qPCR analysis suggested that the mRNA levels of ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 at the later stage of OA cartilage (stage 2 to stage 4) were upregulated in contrast to stage 1 [Fig. 2(AeD)]. In addition, the expression of ADAMTSs was gradually enhanced from stage 2. Consistent with the mRNA levels,

Table II Primers sequences for ChIP Name ADAMTS-4 ADAMTS-4 ADAMTS-7 ADAMTS-7

promoter upstream promoter upstream

Forward (50 / 30 )

Reverse (50 / 30 )

CTCCTTCCTGGGGATTTCCTG ACCCTTTGTCCGACTCCTCTA AATGAAAGGGCCCAGGAGATT AGCAAGACCAGCTGGTGTAAG

GTAGGGGCAATTTCTGCTCA ACTTGCCTGACCCTGGTTCT TGTGCCCCCTGCCCTCTGAG TGCTTGACTCCTGGCCCCTA

5 4 3

Stage Stage Stage Stage Stage

0 1 2 3 4

**

B

** *

2 1

*

0

Relative mRNA expression

A

Relative mRNA expression

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

3.5 3 2.5 2

Stage Stage Stage Stage Stage

0 1 2 3 4

1 0.5

8 6

Stage Stage Stage Stage Stage

4 2 0

0 1 2 3 4

* *

*

0

ADAMTS-5

** ** **

* ADAMTS-7

D

Relative mRNA expression

Relative mRNA expression

10

**

1.5

ADAMTS-4

C

2263

7 6 5 4

Stage Stage Stage Stage Stage

0 1 2 3 4

3

** ** *

2 1

**

0

ADAMTS-12

Fig. 2. Analysis of the mRNA levels of ADAMTSs at different stages of OA. Total RNA from the cartilage of 22 patients were extracted and transcribed into cDNA. The mRNA expression of ADAMTS-4 (A), ADAMTS-5 (B), ADAMTS-7 (C) and ADAMTS-12 (D) in cartilages obtained, from stage 0 to stage 4. The mRNA levels were analysed using RT-qPCR and normalized using b-actin. Bars indicate the mean ± 95% CI of triplicate experiments from each sample obtained from 22 patients. *P < 0.05; **P < 0.01.

the results obtained from western blotting analysis and IHC staining also showed similar trends [Fig. 3]. This sustainable increase in the mRNA and protein levels of the ADAMTSs was significantly observed in OA cartilage compared with control cartilage. Expression of SOX9 and ECM components at different stages of OA To examine the effects of the ADAMTSs on the ECM components, we identified the expression of SOX9 and the proteoglycans COMP and COL2A1 at different stages of OA using RT-qPCR [Fig. 4(A)] and western blotting analysis [Fig. 4(B)]. Expectedly, the production of COL2A1, ACAN, COMP and SOX9 demonstrated a transient upregulation at stage 1 compared with control cartilage and subsequently decreased gradually at the later stage of OA, indicating an inverse trend compared to the ADAMTSs. Associations between SOX9, ADAMTSs, COL2A1, ACAN and COMP at the early stage of OA The potential pathophysiological of SOX9 in early stage (stage 1) OA cartilage was investigated by examining the relationship between SOX9 and the anabolic factors ACAN and COMP as well as between SOX9 and ADAMTS-4, ADAMTS-5, ADAMTS-7, and ADAMTS-12. An association analysis indicated a remarkable positive correlation between SOX9 mRNA expression levels and those of COL2A1 [Fig. 4(C)], ACAN [Fig. 4(D)] and COMP [Fig. 4(E)]. In contrast, the expression of SOX9 was inversely correlated with ADAMTS-4 [Fig. 4(F)], ADAMTS-5 [Fig. 4(G)], ADAMTS-7 [Fig. 4(H)], and ADAMTS-12 [Fig. 4(I)]. Effects of SOX9 on ADAMTSs, COL2A1, ACAN and COMP in human chondrocytes To identify whether the associations previously found in freshly isolated tissues obtained from patients were correlated, we next studied whether SOX9 could modulate catabolic activity in chondrocytes isolated from stage 1 cartilage samples. RT-qPCR and

western blotting analysis of supernatant cultures suggested that SOX9 overexpression repressed ADAMT-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 expression, while knockdown of SOX9 increased ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 activity (P < 0.05). In addition, the expression of ADAMTSs was inhibited after SOX9 rescue in SOX9 knockdown chondrocytes [Fig. 5 (A)]. In addition, SOX9 promoted the expression of cartilage matrix genes, such as COL2A1, ACAN and COMP, and the inhibition of SOX9 significantly suppressed the expression of COL2A1, ACAN and COMP [Fig. 5(B)], suggesting that SOX9 induced the composition of human cartilage ECM components and antagonized the progression of OA. Taken together, these findings indicated a crucial role of SOX9 in cartilage protection that is potentially correlated with cartilage homeostasis.

Inhibition of ADAMTSs increased the expression of cartilage matrix genes at an early stage of OA Among the ECM components, COL2A1, ACAN and COMP are regarded as important biomarkers of cartilage, which determines the levels of OA degeneration. To further study whether the upregulation of cartilage genes at the early stage of OA was caused by ADAMTSs, human chondrocytes isolated from human knee cartilage at stage 1 were independently transfected with siRNAs for ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12. Alternations in the mRNA and protein levels of COL2A1, ACAN and COMP were evaluated using RT-qPCR and western blotting analysis. Our data suggested that the production of COL2A1, ACAN and COMP were upregulated when ADAMTSs were knocked down [Fig. 5(DeG)], indicating that the knockdown of ADAMTSs stimulated cartilage matrix gene expression at the early stage of OA. To further confirm the direct regulation of ADAMTS by SOX9, we performed a ChIP assay. These results indicated that SOX9 could be recruited to the promoters of ADAMTS-4 and ADAMTS-7, but not to a region that was approximately 2-kb upstream of the ADAMTS-4 and ADAMTS-7 promoters [Fig. 5(HeI)]. Taken together, these data strongly suggested that SOX9 inhibits the transcription of ADAMTSs via regulated recruitment of SOX9 to the ADAMTS promoter.

2264

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

Fig. 3. Analysis of the protein levels of ADAMTSs at different stages of OA. Expression levels of ADAMTS-4 (A), ADAMTS-5 (B), ADAMTS-7 (C), and ADAMTS-12 (D) in human cartilage of different stages using RT-qPCR and western blotting analyses. The different stages were identified using the Outerbridge Classification. Box-and-whisker plots below the immunoblotting images showed the average expression of ADAMTS. The expression of ADAMTS at different stages of OA was plotted and compared (ManneWhitney U test). In boxand-whisker plots, the horizontal bars indicate the medians, the boxes indicate the twenty fifth to seventy fifth percentiles, and the whiskers indicate the tenth and ninetieth percentiles. *P < 0.05 vs stage 0; *P < 0.01 vs stage 0. Immunohistochemical analysis of ADAMTSs at different stages of OA. ADAMTS-4 (aee) ADAMTS-5 (fej), ADAMTS-7 (keo) and ADAMTS-12 (pet) protein levels at different stages of OA were detected using IHC on sections of paraffin-embedded pallets. Images were obtained using a 200 objective (E). Scale bar, 50 mm. Sections shown are representative of OA tissue from one patient. Western blots are representative of experiments from six different patients with similar results.

Discussion ADAMTSs consist of a family of 19 secreted enzymes that have been implicated in numerous physiological processes, such as development, invasion, angiogenesis and coagulation19,20.

Abnormal expression of ADAMTS family members has been identified in the pathogenesis of various types of diseases, including OA. ADAMTS-7 and ADAMTS-12 are important in the degradation of COMP, and collagen type II is important in the progression of arthritis12,21; the levels of ADAMTS-7 and ADAMTS-12 are

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

2265

Fig. 4. Correlation analysis between SOX9 and COL2A1, ACAN, and COMP in human OA cartilage samples. The expression of COL2A1, ACAN, COMP and SOX9 in human cartilages at different stages (A, B). mRNA expression was identified using RT-qPCR. The protein profiles were examined using western blotting analyses. Data obtained from stage 0 was treated as the control group. Data were shown as the mean ± 95% CIs. *P < 0.05; **P < 0.01. (CeH) Correlations between the expression levels of SOX9 and COL2A1(C), ACAN (D), COMP (E), ADAMTS-4 (F), ADAMTS-5 (G), ADAMTS-7 (H) and ADAMTS-12 (I) (P < 0.01) in human cartilage at stage 1. COL2A1, ACAN, COMP and SOX9 relative mRNA expression levels at stage 1 were examined using RT-qPCR. The scores of ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 at stage 1 of OA were identified using immunohistochemical analysis.

significantly elevated in the cartilage and synovium of patients with arthritis. ADAMTS-4 and ADAMTS-5 are also considered to be key ECM degradation aggrecanases underlying OA22,23. Although several reports have studied the mechanisms of ECM degradation, inflammation and other factors involved in the OA process, our understanding of OA degeneration and the precise mechanism of ADAMTS-mediated degeneration at the early stage of OA have not been fully elucidated. Thus, we investigated the expression of ADAMTSs at the early stage of OA using arthroscopic classification and found that they were downregulated at the early stage, but

demonstrated a sustainable increase in the later stage of OA, which simultaneously demonstrates their importance during the course of human cartilage ECM degradation. ADAMTSs are known to be involved in the degradation of ECM components, such as ACAN, COMP and COL2A1, as well as other factors. In our study, we found that knockdown of these enzymes not only affected the protein levels of ECM components but also affected their mRNA levels, which indicated that ADAMTSs plays a dual role in controlling the levels of ECM components. With the exception of their classical enzyme activities to degrade substrates,

2266

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

Fig. 5. SOX9 repressed ADAMTSs expression levels in human chondrocytes. Human chondrocytes were transfected with control siRNA, SOX9 siRNA, empty vector or the SOX9 plasmid construct (A, B). SOX9 was also rescued in SOX9 knockdown chondrocytes after 48 h of transfection. The mRNA expression levels of ADAMTS-4, ADAMTS-5, ADAMTS-7, and ADAMTS-12 (A) and COL2A1, ACAN, and COMP (B) in human chondrocytes after 72 h of transfection using RT-qPCR. The protein profiles of ADAMTS-4, ADAMTS-5, ADAMTS-7, and ADAMTS-12 (A) and COL2A1, ACAN, and COMP (B) in human chondrocytes were identified using western blotting analyses. Data are shown as the mean ± 95% CIs. *P < 0.05; **P < 0.01. (C) Bright field (top), fluorescence (middle) and merged (bottom) microscopy of human chondrocytes 96 h following transfection with vector labelled with Green Fluorescent Protein (GFP). (DeG) Human chondrocytes were transfected with siRNAs for ADAMTS-4, ADAMTS-5, ADAMTS-7 or ADAMTS-12. Control siRNA was treated as the control. The mRNA and protein levels of COL2A1, ACAN and COMP obtained from chondrocytes after 72 h of transfection were measured using quantitative RT-PCR and western blotting analysis. RT-PCR and western blots are representative of experiments from three different patients with similar results (n ¼ 3). (HeI) ChIP assay for SOX9 occupancy on the ADAMTS-4 promoter (H), ADAMTS-7 promoter (I) or upstream of the promoter in human OA chondrocytes (passage 1). Data are shown as the mean ± 95% CIs. *P < 0.05; **P < 0.01.

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

they also affect the transcription of these substrates by changing the expression of transcriptional factors (TFs) or by changing the recruitment of TFs to the promoters of the substrates. However, which transcription factors participate in this process requires further investigated. Importantly, SOX9 is essential for the ability of chondrocytes to produce cartilage matrix13. Fukui N et al. revealed that the mRNA expression of SOX9 tended to be reduced in the degenerated areas, particularly in the upper region of the degenerated cartilage in OA cartilage. Within OA cartilage, the expression of cartilage matrix genes was significantly correlated with SOX9 expression24. Considering the importance of SOX9 in the development and maintenance of the chondrocyte phenotype, its reduction in OA may contribute to cartilage pathology. To investigate the potential mechanism underlying the decreased expression of ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 in stage 1 of OA and to assess the potential role of SOX9 in OA pathogenesis, we determined the expression of SOX9 at different stages of OA and found that SOX9 was upregulated in the early stage of OA, but was downregulated in the later stage of OA. Inhibition of SOX9 markedly suppressed COL2A1, COMP, and ACAN production, but increased the expression of ADAMTS-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12, indicating the important function of SOX9 in cartilage equilibrium between matrix deposition and degradation, suggesting that stage 1 may be a generation period of OA [Fig. 6 ]. However, cartilage has a limited repair and regeneration capacity, and thus, cartilage damage often results in dysfunction and OA pathogenesis, suggesting that stage 2 to stage 4 is the lost generation period. The transcription factor SOX9 plays an essential role in regulation of chondrocyte development. It is known that both SOX9 directly upregulates the gene expression of ECM components, including COL2A1, COMP and ACAN, while ADAMTSs downregulate them. However, the relationship between SOX9 and ADAMTSs is largely unknown. A research group examined the expression pattern and levels of Sox9 in ADAMTS-5 null mice. They observed an increase in Sox9 expression in ADAMTS-5/ valve mesenchyme, although the mechanism is undefined25. In our previous study, we found that ADAMTS-7 and ADAMTS-12 could downregulate the expression of SOX9 in endplate cells26. Taken together, these findings indicated that ADAMTS could regulate SOX9. However, there are some clues as to whether SOX9 regulates ADAMTSs

2267

or vice versa. Meelis Kadaja et al. used RNA sequencing (RNA-seq) to identify SOX9-dependent transcriptional changes and ChIP and deep sequencing (ChIP-seq) to identify SOX9-bound genes in hair follicles (HF-SCs) and found a large cohort of ECM remodelling factors, including ADAMTS-10, ADAMTS-14, and ADAMTS-1, which were upregulated in SOX9 KO cells. This finding indicated that SOX9 could directly downregulate ADAMTS27. In our study, we used more direct methods by performing both SOX9 overexpression and knockdown experiments to determine both the mRNA and protein expression levels of ADAMT-4, ADAMTS-5, ADAMTS-7 and ADAMTS-12 and found that overexpression of SOX9 suppressed the activity of the ADAMTSs, while knockdown of SOX9 increased the activity of the ADAMTSs. Expression of ADAMTS was inhibited after SOX9 rescue in SOX9 knockdown chondrocytes. On the basis of the RNA and CHIP-Seq analysis, we hypothesized that SOX9 directly binds to the promoters of ADAMTSs and affects their transcription. Thus, we performed a ChIP assay and confirmed that SOX9 was recruited to the promoters of ADAMTS-4 and ADAMTS-7. These data collectively indicated that SOX9 could negatively and directly regulate the transcription of ADAMTSs. Our finding that ADAMTSs may represent a novel target of SOX9 is consistent with previous findings. We observed that the transfection efficiency of myc-SOX9 was approximately 40%. As a result, the mRNA levels of the entire population might be expected to be 40% lower. However, in the case of ADAMTS-7, these levels were reduced to 20e30% of the controls. This finding might be due to experimental variability, an indirect effect of SOX9 involving communication between transfected and non-transfected cells, or a preferential transfection of a subpopulation of cells that are higher producers of ADAMTS-7. In conclusion, SOX9 upregulation mediates the repression of ADAMTSs at the early stage of OA. In addition, SOX9 expression was negatively correlated with the production of ADAMTSs, suggesting that SOX9 could function as a protective regulator in human articular cartilage. However, further investigation accounting for the dysregulation of SOX9 and ADAMTSs at the early stage of OA requires further studies. Thus, the identification of the mechanisms underlying this process may highlight the development of a therapeutic target for OA, which could provide an alternative treatment strategy for this disease that is distinct from surgical intervention.

Fig. 6. A schema depicting the hypothesized regulatory mechanism of SOX9 in the expression of ADAMTSs at stage 1 of human OA. Inflammatory cytokines, such as TNF-a and interleukin-1 beta (IL-1b), repress SOX9 expression and activate the expression of ADAMTSs. In addition, inflammatory cytokines inhibit the recruitment of the repressor SOX9 to the promoters of ADAMTSs, thereby enhancing the expression of ADAMTSs at the early stage of human OA, and resulting in an increase in COL2A1, ACAN and COMP expression. Thus, SOX9 can be regarded to be a repressor of ADAMTSs and a protector of the early stage of OA, either by directly regulating ECM markers or by suppressing ADAMTS-mediated ECM marker downregulation.

2268

Q. Zhang et al. / Osteoarthritis and Cartilage 23 (2015) 2259e2268

Author contributions QZ, QJ, and YW designed the study; QZ, QJ, XX, XW, and LK acquired the data; QZ and QJ analysed and interpreted the data; ZL and YL provided technical and material support; and QZ, XX and YW obtained funding and participated in the critical revision of the manuscript and provided important intellectual content and study supervision. QZ and QJ provided statistical analysis. All of the authors read and approved the final manuscript. Conflicts of interest The authors declare that they have no competing interests. Acknowledgements The authors thank Drs Jiying Chen, Zhi Gang Wang and Xu Cai for collecting the data and Dr Min Wei for valuable comments and for providing some samples. This work was supported by the National Natural Science Foundation of China (81330053, 81101387, 81371976, 81472589, 31100604, and 81372161), Beijing Natural Science Foundation (7152135) and Beijing Nova Program (Z141102001814055). References 1. Glyn-Jones S, Palmer AJ, Agricola R, Price AJ, Vincent TL, Weinans H, et al. Osteoarthritis. Lancet 2015. pii: S01406736(14)60802-3. [Epub ahead of print]. 2. Lee AS, Ellman MB, Yan D, Kroin JS, Cole BJ, van Wijnen AJ, et al. A current review of molecular mechanisms regarding osteoarthritis and pain. Gene 2013;527(2):440e7. 3. Berenbaum F. Osteoarthritis as an inflammatory disease (osteoarthritis is not osteoarthrosis!). Osteoarthritis Cartilage 2013;21(1):16e21. 4. Maldonado M, Nam J. The role of changes in extracellular matrix of cartilage in the presence of inflammation on the pathology of osteoarthritis. Biomed Res Int 2013;2013:284873. 5. Jordan JM. Cartilage oligomeric matrix protein as a marker of osteoarthritis. J Rheumatol Suppl 2004;70:45e9. 6. Tsezou A. Osteoarthritis year in review 2014: genetics and genomics. Osteoarthritis Cartilage 2014;22(12):2017e24. 7. Arai M, Anderson D, Kurdi Y, Annis-Freeman B, Shields K, Collins-Racie LA, et al. Effect of adenovirus-mediated overexpression of bovine ADAMTS-4 and human ADAMTS-5 in primary bovine articular chondrocyte pellet culture system. Osteoarthritis Cartilage 2004;12(8):599e613. 8. Akagi R, Sasho T, Saito M, Endo J, Yamaguchi S, Muramatsu Y, et al. Effective knock down of matrix metalloproteinase-13 by an intra-articular injection of small interfering RNA (siRNA) in a murine surgically-induced osteoarthritis model. J Orthop Res 2014;32(9):1175e80. 9. Jackson MT, Moradi B, Smith MM, Jackson CJ, Little CB. Activation of matrix metalloproteinases 2, 9, and 13 by activated protein C in human osteoarthritic cartilage chondrocytes. Arthritis Rheumatol 2014;66(6):1525e36. 10. Majumdar MK, Askew R, Schelling S, Stedman N, Blanchet T, Hopkins B, et al. Double-knockout of ADAMTS-4 and ADAMTS5 in mice results in physiologically normal animals and prevents the progression of osteoarthritis. Arthritis Rheum 2007;56(11):3670e4. 11. Lai Y, Bai X, Zhao Y, Tian Q, Liu B, Lin EA, et al. ADAMTS-7 forms a positive feedback loop with TNF-alpha in the pathogenesis of osteoarthritis. Ann Rheum Dis 2014;73(8):1575e84.

12. Luan Y, Kong L, Howell DR, Ilalov K, Fajardo M, Bai XH, et al. Inhibition of ADAMTS-7 and ADAMTS-12 degradation of cartilage oligomeric matrix protein by alpha-2-macroglobulin. Osteoarthritis Cartilage 2008;16(11):1413e20. 13. Lefebvre V, de Crombrugghe B. Toward understanding SOX9 function in chondrocyte differentiation. Matrix Biol 1998;16(9):529e40. 14. Hardingham TE, Oldershaw RA, Tew SR. Cartilage, SOX9 and Notch signals in chondrogenesis. J Anat 2006;209(4): 469e80. 15. de Crombrugghe B, Lefebvre V, Behringer RR, Bi W, Murakami S, Huang W. Transcriptional mechanisms of chondrocyte differentiation. Matrix Biol 2000;19(5):389e94. 16. Pritchett J, Athwal V, Roberts N, Hanley NA, Hanley KP. Understanding the role of SOX9 in acquired diseases: lessons from development. Trends Mol Med 2011;17(3):166e74. 17. Akiyama H, Lefebvre V. Unraveling the transcriptional regulatory machinery in chondrogenesis. J Bone Miner Metab 2011;29(4):390e5. 18. Pearson RG, Kurien T, Shu KS, Scammell BE. Histopathology grading systems for characterisation of human knee osteoarthritisereproducibility, variability, reliability, correlation, and validity. Osteoarthritis Cartilage 2011;19(3):324e31. 19. Fosang AJ, Rogerson FM. Identifying the human aggrecanase. Osteoarthritis Cartilage 2010;18(9):1109e16. 20. Vo NV, Hartman RA, Yurube T, Jacobs LJ, Sowa GA, Kang JD. Expression and regulation of metalloproteinases and their inhibitors in intervertebral disc aging and degeneration. Spine J 2013;13(3):331e41. 21. Guo F, Lai Y, Tian Q, Lin EA, Kong L, Liu C. Granulin-epithelin precursor binds directly to ADAMTS-7 and ADAMTS-12 and inhibits their degradation of cartilage oligomeric matrix protein. Arthritis Rheum 2010;62(7):2023e36. 22. Verma P, Dalal K. ADAMTS-4 and ADAMTS-5: key enzymes in osteoarthritis. J Cell Biochem 2011;112(12):3507e14. 23. Nagase H, Kashiwagi M. Aggrecanases and cartilage matrix degradation. Arthritis Res Ther 2003;5(2):94e103. 24. Fukui N, Ikeda Y, Ohnuki T, Tanaka N, Hikita A, Mitomi H, et al. Regional differences in chondrocyte metabolism in osteoarthritis: a detailed analysis by laser capture microdissection. Arthritis Rheum 2008;58(1):154e63. 25. Dupuis LE, McCulloch DR, McGarity JD, Bahan A, Wessels A, Weber D, et al. Altered versican cleavage in ADAMTS5 deficient mice; a novel etiology of myxomatous valve disease. Dev Biol 2011;357(1):152e64. 26. Zhang Q, Huang M, Wang X, Xu X, Ni M, Wang Y. Negative effects of ADAMTS-7 and ADAMTS-12 on endplate cartilage differentiation. J Orthop Res 2012;30(8):1238e43. 27. Kadaja M, Keyes BE, Lin M, Pasolli HA, Genander M, Polak L, et al. SOX9: a stem cell transcriptional regulator of secreted niche signaling factors. Genes Dev 2014;28(4):328e41. 28. Xu X, Jiang C, Wang S, Tai Y, Wang T, Kang L, et al. HPIP is upregulated in liver cancer and promotes hepatoma cell proliferation via activation of G2/M transition. IUBMB Life 2013;65(10):873e82. 29. Xu X, Fan Z, Kang L, Han J, Jiang C, Zheng X, et al. Hepatitis B virus X protein represses miRNA-148a to enhance tumorigenesis. J Clin Invest 2013;123(2):630e45. 30. Feng Y, Xu X, Zhang Y, Ding J, Wang Y, Zhang X, et al. HPIP is upregulated in colorectal cancer and regulates colorectal cancer cell proliferation, apoptosis and invasion. Sci Rep 2015;5:9429.