Activation of NF-κB in Synovium versus Cartilage from Patients with Advanced Knee Osteoarthritis: A Potential Contributor to Inflammatory Aspects of Disease Progression This information is current as of September 26, 2021. Aisha S. Ahmed, Per Gedin, Anders Hugo, Georgy Bakalkin, Alkass Kanar, David A. Hart, Henrik Druid, Camilla Svensson and Eva Kosek J Immunol 2018; 201:1918-1927; Prepublished online 22 August 2018; Downloaded from doi: 10.4049/jimmunol.1800486 http://www.jimmunol.org/content/201/7/1918 http://www.jimmunol.org/ Supplementary http://www.jimmunol.org/content/suppl/2018/08/21/jimmunol.180048 Material 6.DCSupplemental References This article cites 51 articles, 8 of which you can access for free at: http://www.jimmunol.org/content/201/7/1918.full#ref-list-1

Why The JI? Submit online. by guest on September 26, 2021

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2018 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Activation of NF-kB in Synovium versus Cartilage from Patients with Advanced Knee Osteoarthritis: A Potential Contributor to Inflammatory Aspects of Disease Progression

Aisha S. Ahmed,* Per Gedin,† Anders Hugo,† Georgy Bakalkin,‡ Alkass Kanar,x,{ David A. Hart,‖ Henrik Druid,x,{ Camilla Svensson,# and Eva Kosek*,**

The aim was to assess the activation and association of the NF-kB system across synovial membrane (SM) and articular cartilage (AC) in patients with knee osteoarthritis (OA) and ascertain its potential effects on catabolic mediator expression in advanced OA. SM and AC were obtained from 40 OA patients undergoing total knee arthroplasty and from 19 postmortem control subjects. NF-kB subunit RelA in nuclear and cytosolic fractions and NF-kB1–DNA binding in nuclear extracts was assessed by ELISA,

whereas NFKB1, RELA, IL-8, IL-6, and MMP3 expression were analyzed by reverse transcriptase–quantitative PCR in Downloaded from tissues. We observed higher SM nuclear RelA levels and upregulated NF-kB1–DNA binding in OA patients compared with postmortem controls. However, in AC, lower nuclear RelA levels were observed compared with cytosolic extracts in patients. Nuclear RelA levels correlated positively with NF-kB1–DNA binding in SM and AC in patients. SM RELA and MMP3 mRNA levels were upregulated, whereas IL-8 and IL-6 as well as AC RELA were downregulated in patients compared with controls. In SM, nuclear RelA levels correlated positively with MMP3 in patients. A negative correlation was observed between SM nuclear RelA levels and AC NF-kB1–DNA binding, and SM nuclear NF-kB1-DNA binding correlated negatively http://www.jimmunol.org/ with AC MMP3 and NFKB1 mRNA levels in patients. These findings highlight NF-kB–triggered cross-talk and feedback mech- anisms between SM and AC in OA. Further, our findings strongly support a role for an activated NF-kB system in the transcrip- tional mechanism of inflammatory processes, especially in SM of patients with advanced OA. The Journal of Immunology, 2018, 201: 1918–1927.

steoarthritis (OA) is the most prevalent joint disease and a comprises five : RelA/p65, RelB, c-Rel, NF-kB1/p50, and leading cause of chronic disability worldwide. Pain is the NF-kB2/p52 (6). Under normal conditions, NF-kB is retained in main symptom, followed by progressive loss of joint the cytosol in an inactive state bound to the inhibitory protein, IkB O by guest on September 26, 2021 structural integrity. Although structural changes are reported for (7, 8). Proinflammatory cytokines, excessive mechanical stress, all knee joint compartments (i.e., cartilage, subchondral bone, and and/or matrix degradation enzymes can initiate a cascade of re- synovial membrane [SM]), studies highlighting the associations actions leading to the phosphorylation and subsequent proteasome among joint tissues are scarce. system–mediated degradation of IkB by the ubiquitin proteasome The role of synovial inflammation in OA pathogenesis is system. Proteasomal degradation of IkB leads to the release of emerging, with recent reports demonstrating an association with active NF-kB (RelA/NF-kB1 dimer), which gets translocated to knee OA pain and cartilage erosion (1, 2). Synovial inflammation is the nucleus to induce gene transcription (7). Activation of NF-kB a major source of proinflammatory cytokines, metalloproteinases has been reported in vascular endothelium and type A synovial (MMPs), and aggrecanases, which contribute to cartilage matrix lining cells, as well as adjacent to the cartilage–pannus junction in degradation in OA (3, 4). Among multiple pathways and media- patients with rheumatoid arthritis (RA) (9, 10). NF-kB1 and RelA tors influencing the development and persistence of synovial in- expression is upregulated in synovial tissues in OA patients with a flammation, NF-kB is reported to play a prominent role (5). more prominent upregulation in early OA compared with later NF-kB is a heterodimer protein consisting of two subunits; phases of the disease (11). Furthermore, upregulation of NF-kB1 RelA/p65 and NF-kB1/p50. The mammalian NF-kB subfamily expression in cartilage tissues has been reported in patients with

*Department of Clinical Neuroscience, Karolinska Institutet, 171 77 Stockholm, LundbladandfamilyandfromEliLillyand Co., United States. There was no Sweden; †Ortho Center Stockholm, Lo¨wenstro¨mska Hospital, 194 89 Upplands role for any funding agency regarding the study design, collection and analysis, Va¨sby, Sweden; ‡Department of Pharmaceutical Biosciences, Uppsala University, interpretation of data, manuscript writing, or decision to submit the manuscript for 751 05 Uppsala, Sweden; xDepartment of Oncology-Pathology, Karolinska Institutet, publication. 171 77 Stockholm, Sweden; {Swedish National Board of Forensic Medicine, 171 65 ‖ Address correspondence and reprint requests to Dr. Aisha S. Ahmed, Department of Solna, Sweden; McCaig Institute for Bone and Joint Health, University of Calgary, Clinical Neuroscience, Karolinska Institutet, Nobel Va¨g 9, 171 77 Stockholm, Sweden. Calgary, Alberta T2N 1N4, Canada; #Department of Physiology and Pharmacology, E-mail address: [email protected] Karolinska Institutet, 171 77 Stockholm, Sweden; and **Stockholm Spine Center, Lo¨wenstro¨mska Hospital, 194 89 Upplands Va¨sby, Sweden The online version of this article contains supplemental material. ORCIDs: 0000-0002-8483-5309 (A.S.A.); 0000-0002-5701-8338 (P.G.); 0000-0002- Abbreviations used in this article: AC, articular cartilage; BMI, body mass index; 8074-9833 (G.B.); 0000-0002-9198-023X (H.D.). FLS, fibroblast-like synoviocyte; MMP, metalloproteinase; OA, osteoarthritis; PM, postmortem; RA, rheumatoid arthritis; RQI, RNA quality index; RT-qPCR, reverse Received for publication April 3, 2018. Accepted for publication July 22, 2018. transcriptase–quantitative PCR; SM, synovial membrane. This work was supported by funding from the Swedish Research Council (K2013- 52X-22199-01-3 to E.K. and 542-2013-8373 to C.S.); the Swedish Rheumatism Copyright Ó 2018 by The American Association of Immunologists, Inc. 0022-1767/18/$35.00 Association (to A.S.A.); the Ulla and Gustav af Uggla Foundation, Sweden (to A.S.A.); the King Gustav V Foundation, Sweden (to A.S.A.); donations from Leif www.jimmunol.org/cgi/doi/10.4049/jimmunol.1800486 The Journal of Immunology 1919 advanced knee joint OA (12). It is not known if there is a concurrent (Hs00765730_m1), RELA (Hs00153294_m1), IL-8 (Hs00174103_m1), activation of NF-kB across different joint compartments in OA. IL-6 (Hs00985639_m1), and MMP3 (Hs00968305_m1) were used to de- Altered expression of numerous involved in inflammation tect targets. Relative abundance was calculated using a standard curve generated by adding a dilution series of cDNA from primary fibroblast-like or extracellular matrix degradation are reported in OA cartilage synoviocytes (FLS). Data were normalized to the mean of HPRT (13). Activated NF-kB boosts the expression of inflammation- (Hs02800695_m1) and GAPDH (Hs00266705-g1) gene expression and related cytokines and chemokines, such as IL-6, IL-8, adhesion presented as relative expression units. Because of technical or handling molecules, and matrix degrading enzymes such as matrix MMPs complications, detectable SM RNA could be extracted from 39/40 OA patients and 19/19 PM controls, and for AC, it was from 24/40 patients and (14, 15). IL-8 and IL-6 were shown to activate chondrocytes and 7/19 PM controls (Table I). No differences were noted for RQI among fibroblasts to secrete collagenases and proteases contributing to patients and PM controls for SM and AC tissues, with values measured for cartilage destruction in primary cell cultures (16, 17). It has been SM as 8.04 6 0.5 for OA and 7.83 6 0.5 for PM controls and for AC as reported that OA patients have elevated MMP3, IL-8, and IL-6 7.30 6 1.54 for OA and 7.41 6 1.42 for PM controls, respectively. levels in the synovial fluid (18, 19), and damaged chondrocytes Nuclear and cytoplasmic protein extraction release cytokines including IL-8 and IL-6 (17). In addition, OA patients have increased serum IL-6 levels, and serum IL-6 levels Frozen SM and AC tissues collected from OA patients and PM controls were powdered using a Mikro-Dismembrator in liquid nitrogen and mixed with have been highlighted as a predictor of radiological OA (20). hypotonic buffer consisting of 20 mM Tris-HCl, 10 mM NaCl, 3 mM MgCl2 The aim of the current study was to investigate whether NF-kBis containing 10% phosphatase, and 10% protease inhibitor mixture (Thermo activated in both cartilage and synovial tissues and if it is associated Fisher Scientific). After addition of 10% liquid NP-40, homogenates were 3 with elevated levels of cytokines and proteinases in both compart- centrifuged at 3000 g for 10 min at 4˚C. Supernatant fractions corre- sponding to the cytoplasmic extracts were collected, aliquoted, and stored ments. For this purpose, we have quantified and compared RelA at 280˚C. The pellets were resuspended in extraction buffer containing Downloaded from levels, NF-kB1–DNA binding, and RELA and NFKB1 as well as 10 mM Tris, 100 mM NaCl, 1 mM EDTA, 1% Triton X-100, and 0.1% IL-8, IL-6,andMMP3 gene expression in SM and articular cartilage SDS deoxycholate supplemented with 2% protease inhibitor mixture and (AC) collected from patients with advanced knee OA and from 1 mM PMSF, incubated at 4˚C for 30 min with vortex every 10 min, 3 postmortem (PM) controls. Considering that OA affects the entire followed by centrifugation at 14,000 g for 30 min at 4˚C. Supernatants corresponding to nuclear extracts were collected, aliquoted, and stored at joint involving all joint compartments, and cells in each compart- 280˚C. Because of technical and handling complications, SM cytoplasmic ment are capable of initiating the cascade of reactions leading to the and nuclear extracts could be made from 30/40 OA patients and 18/19 PM http://www.jimmunol.org/ degeneration of cartilage independently as well as in association controls, and for AC, it was from 35/40 patients and 14/19 PM controls with adjacent tissues, we hypothesized that NF-kB(RelA/NF-kB1) (Table I). Mean protein concentration in cytoplasmic fractions from SM was 3.16 6 1.30 mg/ml for the patient samples and 3.09 6 1.78 mg/ml for activation and the regulation of inflammatory degenerative marker the PM control samples, whereas for the nuclear fractions it was 3.73 6 transcription would correlate positively across compartments. 2.01 mg/ml for the patients and 2.91 6 1.67 mg/ml for the PM controls, with no significant group differences. Mean protein concentration in AC cytoplasmic fractions was 1.35 6 1.32 mg/ml for the patient samples and Materials and Methods 1.07 6 0.52 mg/ml for the PM control samples, whereas for the nuclear Study subjects fractions it was 3.15 6 4.25 mg/ml for the patients and 3.33 6 4.7 mg/ml for the PM controls, with no significant differences between these groups.

Patients. Forty patients with knee joint OA undergoing total knee arthro- by guest on September 26, 2021 plasty were included in the study. All patients were diagnosed according to TNF-stimulated human FLS were included as positive controls for the the American College of Rheumatology criteria for OA (21), had OA pain cytosolic and nuclear protein extraction procedures. Protein content was as their dominant complaint, and were recruited from the waiting list for measured using a BCA Assay Kit (Thermo Fisher Scientific). total knee replacement at the Ortho Center, Stockholm, Sweden. The ex- Detection of RelA by ELISA clusion criteria were as follows: chronic pain due to causes other than knee OA, such as fibromyalgia, degenerative disc disease, disc herniation, in- RelA expression was measured in the cytoplasmic and nuclear extracts from flammatory rheumatic disease, or neurologic disease or previous knee SM and AC collected from OA patients and PM controls using a com- surgery at the knee planned for total knee arthroplasty. mercially available ELISA kit (Invitrogen Life Technologies). Cytosolic All participants were informed about the study procedure, and written and nuclear extracts from TNF-stimulated FLS were included as assay consent was obtained. Knee joint AC and SM was collected during surgery controls. Assays were performed according to the protocol set by the and immediately frozen at 280˚C for future analysis. manufacturer. Briefly, the assay employed Abs specific for RelA coated on PM controls. Nineteen PM subjects with no history of knee or hip OA or 96-well plates. All reagents, samples, and standards were prepared as inflammatory rheumatic diseases were included as controls. AC and SM instructed. Subsequently, standards or samples were added to each well and from the medial side were collected during an autopsy procedure and incubated for 2 h at room temperature followed by the addition of 100 mlof immediately frozen at 280˚C for future analysis. AC was macroscopically biotin Ab. Subsequently, streptavidin solution was added, followed by examined during the autopsy procedure for any characteristic signs of OA. incubation for 45 min at room temperature. After addition of the stop Patient and PM control demographic data are reported in Table I. This solution, absorbance was measured at 450 nm. study was approved by the Regional Ethical Review Board in Stockholm, k Sweden (reference number 2011/2036-31-1, 2012/2006-32) and followed NF- B1–DNA binding activity the guidelines of the Declaration of Helsinki. NF-kB1–DNA binding was analyzed using a commercially available ELISA kit (Rockland) in SM and AC nuclear extracts from OA and control RNA isolation and reverse transcriptase–quantitative PCR synovial tissues, following the protocol set by the manufacturer. The k Frozen SM and AC samples collected from OA patients and PM controls ELISA kit employed Abs specific for NF- B1 coated on 96-well plates. were homogenized using a Mikro-Dismembrator (B. Braun Biotech In- All reagents, samples, and standards were prepared as instructed. ternational) and dissolved in 2–3 vol of TRIzol reagent (Invitrogen Life TNF-stimulated FLS nuclear extract was included as a positive control in Technologies). RNA was then extracted and purified using RNeasy the assays. The absorbance was read at 450 nm. MiniKit (Qiagen) following the manufacturer’s protocol. Quantification of Statistics RNA was determined using a Nanodrop ND-1000 Spectrophotometer (Isogen Life Science), and RNA quality was measured as the RNA quality The data were analyzed using the IBM SPSS Statistics 22 software index (RQI) using the Experion electrophoresis system (Bio-Rad). First- program (International Business Machines). Student t test was applied strand cDNA were synthesized from 1 mg of total RNA using a first-strand to detect differences among groups, and between-group differences cDNA Synthesis Kit (Roche). were measured by paired t test. The significance of differences among Reverse transcriptase–quantitative PCR (RT-qPCR) was performed us- groups was analyzed by ANOVA followed by Fisher protected least ing TaqMan Gene Expression Assays (Applied Biosystems, Carlsbad, significant difference test. A univariate analysis of covariance was CA) by using the GeneAmp 7500 Fast Sequence Detection system applied to observe age- and gender-related group differences. The (Applied Biosystems). Predeveloped specific primers for NFKB1 significance of correlations was determined by Spearman correlation 1920 NF-kB ACTIVATION IN KNEE JOINT OSTEOARTHRITIS coefficient, and effects of age, gender, ongoing pain, and PM interval group, no differences were observed between SM and AC for were determined by partial correlation analysis. A p value #0.05 was RELA, whereas NFKB1 mRNA levels were higher (p = 0.002) in considered significant. the SM compared with AC (Supplemental Fig. 1A, 1B). Results NF-kB (NFKB1, RELA) expression and activity Patients’ clinical and demographic parameters Synovial membrane. ELISA of cytosolic and nuclear fractions from The studies included 40 patients with knee OA (average age synovial tissues showed higher nuclear RelA levels in OA patients 64.5 6 5.7 y) and 19 PM controls (average age 43 6 12.8 y). All (p = 0.007) compared with PM controls. Within the PM control patients reported pain during the previous 24 mo. Body mass in- group, RelA levels were higher in the cytosolic fraction compared dex (BMI) for patients was recorded as 24.25 6 3.3, and for PM with nuclear extracts (p = 0.037). In the OA patient group, no controls, it was recorded as 28.37 6 5.1. Mean visual analogue significant differences in RelA level between the cellular fractions score for ongoing pain was recorded as 41.28 6 29.4. Eighteen were observed, although levels in the nuclear extracts were higher patients received nonsteroidal anti-inflammatory drugs for pain (Fig. 2A). No age, sex, or BMI-related associations were detected relief; however, 2 wk prior to the surgical procedure, these had for RelA expression in cytosolic and nuclear extracts between OA been stopped (Table I). All OA patients’ and PM controls’ de- patients and PM controls. mographics and those from which RNA and protein could be Analysis of nuclear extracts prepared from OA patients and PM extracted are presented as Table I. controls revealed that NF-kB1–DNA binding activity was signifi- cantly (p = 0.027) higher in OA patients compared with the PM NF-kB (NFKB1, RELA) gene expression controls (Fig. 2B), demonstrating NF-kB system activation in a Downloaded from Synovial membrane. Extracted RNA was of good quality with no diseased state. No gender or BMI-related changes were detected for significant differences in RQI between PM controls and OA pa- NF-kB1–DNA binding activity between OA patients and PM con- tients. Quantitative RT-PCR analysis showed significantly upreg- trols. However, age was detected as a dominant factor for group ulated (p = 0.044) synovial RELA mRNA levels in OA patients differences as indicated by univariate analyses of covariance. compared with the PM controls (Fig. 1A). No significant differ- Articular cartilage. No significant differences were observed between

ences in NFKB1 mRNA levels between OA patients and the PM patients and PM controls for nuclear RelA levels. Within the patient http://www.jimmunol.org/ controls were observed (Fig. 1B). group, RelA levels were higher in the cytosolic fraction compared with Articular cartilage. Extracted RNA was of good quality with no nuclear extracts (p = 0.032) (Fig. 2C), indicating an inactive NF-kB significance differences in RQI between PM controls and OA patients. state in AC extracted from late-stage OA patients. No age, sex, or A significant downregulation (p =0.003)ofRELA mRNA levels in BMI-related changes were detected for RelA expression in cytosolic OA patients compared with the PM controls was observed (Fig. 1C). and nuclear extracts between OA patients and PM controls. However, no significant differences in NFKB mRNA levels between Analysis of nuclear extracts prepared from OA patients and PM OA patients and the PM controls were detected (Fig. 1D). control revealed no statistically significant differences between Gene expression among SM and AC. As measured by paired NF-kB1–DNA binding activity in OA patients compared with the analysis, gene expression levels were significantly higher in the SM PM controls (Fig. 2D). by guest on September 26, 2021 compared with the AC for RELA (p = 0.0001) and NFKB1 Expression among SM and AC. Nuclear RelA levels were higher in (p = 0.0001) in OA patients (Fig. 1E, 1F). In the PM control the SM compared with AC nuclear RelA levels (p = 0.0001) in OA

Table I. Descriptive data for study subjects

OA Patients PM Controls Significance Total study subjects n 40 19 Gender (F/M) 17/23 6/13 Average age (y 6 SD) 64.5 (6 5.7) 43.0 (6 12.8) p = 0.0001 BMI (kg/m2) 24.25 (6 3.3) 28.37 (6 5.1) p = 0.0001 PM interval (h) — 49.3 (6 15) Ongoing pain VAS (1–100) 41.28 (6 29.4) — Anti-inflammatory medication 18/40 — Subjects included in biochemical analysis RT-qPCR (SM) n 39/40 19/19 Gender (F/M) 17/22 6/13 Age, y (6 SD) 64.7 (6 5.8) 43.0 (6 12.8) p = 0.0001 ELISA (SM) n 30/40 18/19 Gender (F/M) 12/18 6/12 Age, y (6 SD) 64.25 (6 6.4) 43.4 (6 13.0) p = 0.0001 RT-qPCR (AC) n 24/40 7/19 Gender (F/M) 10/14 3/4 Age, y (6 SD) 65.92 (6 5.7) 43.53 (6 13.5) p = 0.0001 ELISA (AC) n 35/40 14/19 Gender (F/M) 15/20 4/10 Age, y (6 SD) 64.38 (6 5.7) 41.7 (6 12.8) p = 0.0001 Descriptive and biochemical data for all OA patients and PM controls and those from which biochemical data could be collected. Data presented as mean 6 SD. —, not applicable; F, female; M, male; VAS, visual analogue scale. The Journal of Immunology 1921

FIGURE 1. Relative expression of RELA and NFKB1 mRNA in SM and AC. (A) RELA and (B) NFKB1 gene expression in the SM retrieved from OA patients and PM controls. Values reported are mean 6 SEM. A total of n = 36 OA patients and 18 PM control samples. (C) RELA and (D) NFKB1 gene expression in the AC retrieved from OA patients and PM controls. Values reported are mean 6 SEM. A total of n =24

OA patients and 7 PM control samples. *p # 0.05, Downloaded from **p # 0.01 compared with PM control calculated by Students t test. (E) RELA mRNA levels in SM and AC extracted from OA patients. Data presented as scatterplot with n = 20. Samples were compared using paired Student t test. ##p # 0.01, comparison between SM and AC. (F) NFKB1 mRNA levels in SM and AC extracted from OA patients. Data presented as scat- http://www.jimmunol.org/ terplot with n = 21. Samples were compared using paired Student t test. ##p # 0.01, comparison between SM and AC. The FLS used as an internal control in (A) were also used for comparative analysis regarding (B)–(D). by guest on September 26, 2021

patients (Fig. 2E). Similarly, higher NF-kB1–DNA binding ac- Association between SM and AC. We further assessed elements of tivity (p = 0.046) was observed in the SM samples compared with the NF-kB system for associations among joint tissues and ob- AC samples in the patient group (Fig. 2F), indicating an upregu- served negative correlation between synovial nuclear RelA levels lated activity in the SM. and cartilage NF-kB1–DNA binding (r = 20.547; p = 0.019) in OA patients (Fig. 3C). Association among RelA and NF-kB1–DNA binding within and between tissue compartments IL-8, IL-6, and MMP3 gene expression levels Synovial membrane. A positive correlation (r = 0.717; p = 0.001) Synovial membrane. Quantitative RT-PCR analysis showed signifi- was observed between nuclear RelA levels and nuclear NF-kB1– cantly downregulated (p = 0.002) synovial IL-8 and IL-6 (p = 0.0001) DNA binding in OA patients (Fig. 3A). No detectable effects of mRNA levels in OA patients compared with the PM controls. age, gender, BMI, or ongoing pain on an association between However, increased MMP3 (p = 0.0001) mRNA levels in OA patients nuclear RelA and NF-kB1–DNA binding activity were observed were noted compared with values for the controls. (Fig. 4A–C). in OA patients, as assessed by partial correlation analysis. Articular cartilage. No significant differences in IL-8 and MMP3 Articular cartilage. A positive correlation (r = 0.370; p = 0.037) mRNA levels between OA patients and the PM controls were de- was observed between nuclear RelA levels and nuclear NF-kB1– tected. As IL-6 mRNA levels were only detectable in two PM sub- DNA binding in tissue samples from OA patients (Fig. 3B). jects and four OA patients, no analysis was made for IL-6 because of Similarly, the cytosolic and nuclear RelA levels correlated posi- the limited number of subjects positive for the signal (Fig. 4D–F). tively (r = 0.451; p = 0.008) in the OA patients. No detectable Gene expression in SM and AC. As assessed by the paired analysis, effects of age, gender, BMI, or ongoing pain on an association the gene expression levels were significantly higher in the SMs among nuclear RelA and NF-kB1–DNA binding activity or compared with AC for IL-8 (p =0.004)andMMP3 (p =0.037) between cytosolic and nuclear RelA levels were observed for the for the OA patients (Fig. 4G, 4H). In the PM control group, no OA patients, as assessed by partial correlation analysis. differences were noted between SM and AC for IL-8 and 1922 NF-kB ACTIVATION IN KNEE JOINT OSTEOARTHRITIS

FIGURE 2. NF-kB expression and activation state in SM and AC. (A) RelA levels in the nuclear and cyto- solic fractions of SM collected from patients with OA and PM controls. Values reported are mean 6 SEM. A total of n = 15 OA patients samples and 6 PM controls for cytosolic fractions and nuclear fractions from 20 OA patients samples and 10 PM controls. *p # 0.05, comparison between cytosolic and nuclear fractions in PM controls. ##p # 0.01, comparison between nuclear fractions compared with PM controls by one-way ANOVA followed by Fisher protected least significant difference test. (B) Nuclear NF-kB1–DNA binding activity in SM extracted from OA patients and PM controls. Values reported are mean 6 SEM. A total of n = 30 for OA patients and 18 for PM controls. Sam- ples were compared using Student t test. *p # 0.05, comparison between SM and AC. (C) RelA levels in the nuclear and cytosolic fractions of AC collected from patients with OA and PM controls. Values reported are mean 6 SEM. A total of n = 36 OA pa- Downloaded from tients samples and 14 PM controls for cytosolic frac- tions and nuclear fractions from 33 OA patient samples and 12 PM control samples. *p # 0.05, comparison between cytosolic and nuclear fractions in OA patients by one-way ANOVA followed by Fisher protected least significant difference test. (D) Nuclear NF-kB1–DNA

binding activity in AC extracted from OA patients and http://www.jimmunol.org/ PM controls. Values reported are mean 6 SEM. A total of n = 36 for OA patients and 15 for PM controls. (E) Nuclear RelA levels in SM and AC extracted from OA patients. Data presented as scatterplot with n = 20. Samples were compared using paired Student t test. ##p # 0.01, comparison between SM and AC. (F) Nuclear NF-kB1–DNA binding activity in SM and AC extracted from OA patients. Data presented as scat- terplot with n = 26. Samples were compared using paired Student t test. #p # 0.05, comparison between by guest on September 26, 2021 SM and AC.

MMP3 expression levels. No analysis was performed regarding NF-kB1–DNA binding activity and IL-8, IL-6, or MMP3 gene IL-6 levels because of the limited number of subjects exhibiting expression levels for either the OA patients or the PM controls. detectable results. A positive correlation was detected between RELA and MMP3 gene expression (r = 0.623; p = 0.0001) and between RELA and k Associations among NF- B and IL-8, and IL-6 and IL-8 mRNA (r = 0.499; p = 0.002) levels in synovial tissues MMP3 gene expression levels within and between the collected from OA patients (Fig. 5B, 5C). No statistically signif- tissue compartments icant correlations were observed between RELA and either NFKB1 Synovial membrane. A positive correlation was observed between (Supplemental Table I) or IL-6 gene expression or between nuclear RelA levels and MMP3 gene expression (r = 0.572; p = NFKB1 and either IL-8, IL-6,orMMP3 gene expression levels in 0.011) (Fig. 5A). No correlations were observed between nuclear tissues from the OA patients. No effects of age, gender, BMI, or RelA levels and IL-8 or IL-6 gene expression or between nuclear ongoing pain were observed on outcomes of any association with

FIGURE 3. Correlations among and between tissues for NF-kB subunit expression and activity in OA patients. (A) Correlation among nuclear RelA levels and NF-kB1–DNA binding activity in SM from OA patients (r = 0.717; p = 0.001. n = 19; Pearson correlation coefficient). (B) Correlation among nuclear RelA levels and NF-kB1–DNA binding activity in AC from OA patients (r = 0.451; p = 0.008. n = 33; Pearson correlation coefficient) and (C) correlation among synovial nuclear RelA levels and cartilage NF-kB1–DNA binding in OA patients (r = 20.547; p = 0.019. n = 18; Pearson correlation coefficient). The Journal of Immunology 1923 Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 4. Relative expression of IL-8, IL-6, and MMP3 mRNA in SM and AC. (A) IL-8 (n = 38 OA patients and 19 PM controls), (B) IL-6 (n =35OA patients and 19 PM controls), and (C) MMP3 gene expression (n = 38 OA patients and 19 PM controls) in SM extracted from OA patients and PM controls. Values reported are mean 6 SEM. **p # 0.01, comparison between OA patients and PM controls calculated by Student t test. (D) IL-8 (n = 24 OA patients and 7 PM controls). (E) IL-6 (n = 4 OA patients and 2 PM controls) and (F) MMP3 gene expression (n = 19 OA patients and 6 PM controls) in AC samples extracted from OA patients and PM controls. Values reported are mean 6 SEM. The FLS used as an internal control in (A) were also used for comparative analysis regarding (B)–(F). (G) Relative IL-8 gene expression in SM and AC extracted from OA patients. Data presented as scatterplot with median value; n = 22. Samples were compared using paired Student t test. ##p # 0.01, comparison between SM and AC. (H) MMP3 gene expression in SM and AC extracted from OA patients. Values reported as scatterplot with median with n = 17. Samples were compared using paired Student t test. #p # 0.05, comparison between SM and AC. the OA patients except for a BMI association with RELA and IL-8 (r = 20.549; p = 0.028) in OA patients (Fig. 5G, 5H). However, gene expression. the synovial IL-6 mRNA levels correlated positively with cartilage Articular cartilage. A positive correlation was observed between nuclear NF-kB1–DNA binding (r = 0.356; p = 0.046) (Fig. 5I). NFKB1 and RELA gene expression (r = 0.615; p = 0.001) and Synovial IL-8 and RELA mRNA levels correlated positively with between NFKB1 and MMP3 gene expression (r = 0.498; p = 0.03) cytosolic RelA levels in the cartilage (r = 0.358; p = 0.035 and r = in AC tissues collected from OA patients (Fig. 5D, 5E). Moreover, 0.522; p = 0.002, respectively). Moreover, synovial IL-8 mRNA MMP3 and IL-8 mRNA levels were found to be positively cor- expression was found to be correlated positively with cartilage IL-8 related (r = 0.567; p = 0.011) in the patient group (Fig. 5F). No mRNA levels (r = 0.458; p = 0.024), and synovial RELA mRNA statistically significant correlations were observed between nu- were correlated positively with cartilage IL-8 mRNA (r = 0.358; clear RELA and either NFKB1, IL-8,orMMP3 gene expression p = 0.037) levels in patients, indicating across-tissue association. levels or between RelA and IL-8 or MMP3 gene expression levels for the OA patients. No effects of age, gender, BMI, or ongoing Discussion pain were evident for the associations observed for the OA To our knowledge, this is the first study reporting NF-kB ex- patients. No analysis was performed for IL-6 gene expression pression and activity and its association with inflammatory and because of the limited number of subjects with the sufficient degenerative markers across SM and AC from late-stage OA pa- amplification. tients. The main findings of the study are that tissue-specific Associations between SM and AC. Synovial nuclear NF-kB1–DNA NF-kB unit activation, RELA mRNA and nuclear RelA protein binding correlated negatively with cartilage NFKB1 mRNA levels, and NF-kB1-DNA binding activity was higher in OA sy- levels (r = 20.576; p = 0.008) and with MMP3 mRNA levels novial tissues compared with controls, whereas in AC, RELA 1924 NF-kB ACTIVATION IN KNEE JOINT OSTEOARTHRITIS Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021 FIGURE 5. Relationship among NF-kB and IL-8, IL-6, and MMP3 gene expression in OA patients. All associations were made by Pearson correlation coefficient. Correlations between (A) nuclear RELA levels and MMP3 gene expression (r = 0.572; p = 0.011. n = 19), (B) RELA and MMP3 gene expression (r = 0.623; p = 0.0001. n = 35), and (C) RELA and IL-8 gene expression (r = 0.499; p = 0.002. n = 36) in SM. (D) Correlations between NFKB1 and RELA gene expression (r = 0.615; p = 0.001. n = 24) and (E) among NFKB1 and MMP3 gene expression (r = 0.498; p = 0.03. n = 19) and (F) IL-8 and MMP3 gene expression (r = 0.567; p = 0.011. n = 19) in AC. (G) Correlations between synovial nuclear NF-kB1–DNA binding and cartilage NFKB1 gene expression (r = 20.549; p = 0.028. n = 16). (H) Correlations among synovial nuclear NF-kB1–DNA binding and cartilage MMP3 gene expression (r = 20.576; p = 0.008. n = 20) and (I) correlation among synovial IL-6 gene expression and cartilage NF-kB1–DNA binding (r = 0.356; p = 0.046. n = 32). mRNA and nuclear RelA protein levels were lower in OA patients. IKKa and IKKb complex, is reported to be associated with in- Within tissues, synovial and cartilage nuclear levels of RelA flammation and results in phosphorylation and subsequent deg- correlated positively with NF-kB1–DNA binding in OA samples. radation of the IkB molecule by the ubiquitin proteasome system However, contrary to our hypothesis of a positive correlation for (22). The ubiquitin proteasome system also processes mature NF-kB pathway elements, we observed negative associations NF-kB1/p50 from its precursor p105 (22). The noncanonical across tissues; synovial nuclear RelA levels correlated negatively pathway, in contrast, probably depends on IKKa activation and with NF-kB1-DNA binding in the cartilage. Moreover, synovial nuclear localization of p52/RelB heterodimer and is reported to be NF-kB1–DNA binding activity exhibited a negative association involved in adaptive immunity (23, 24). Our present results are in with cartilage MMP3 and NFKB1 gene expression. In the synovial line with previous findings of the activation of the canonical tissues, an upregulated MMP3 and IL-8 gene expression was ob- NF-kB pathway during synovial inflammation, as we observed served and MMP3 mRNA levels correlated positively with RelA upregulated nuclear RelA levels in the SM that correlated posi- nuclear levels in OA patients. In summary, these results suggest tively with upregulated NF-kB1–DNA binding activity in OA that chronic, advanced knee OA is associated with what can be patients, indicating translocation of active NF-kB (RelA/NF-kB1 considered a pathological activation of NF-kB RelA/NF-kB1 in dimer) to the nucleus, an essential step in the activation and the synovium, which can in turn regulate the inflammatory, cata- subsequent induction of transcription of inflammatory and de- bolic processes contributing to cartilage degeneration. generative genes. Increased synovial NF-kB1 expression has been A balanced NF-kB response is critical for physiological reported at sites adjacent to the cartilage–pannus junction in pa- processes and to maintain tissue integrity and homeostasis. Under tients with RA (10), and RelA DNA binding complexes are pathological states, this balance is shifted toward pathological reported in nuclear extracts from TNF-a stimulated human expression of algogenic and inflammatory mediators, leading to synovial explants cultures (25). Upregulated NF-kB1 and RELA catabolic processes, cell death, or apoptosis. For NF-kB signaling, expression are associated with inflammation and increased COX-2 the canonical pathway, mediated by the activation of IkB kinases, expression in knee joint synovial tissues retrieved from patients The Journal of Immunology 1925 with early-phase OA (11). In animal models of arthritis, an up- chondrocyte cell death (via apoptosis and/or necrosis) leading to regulated NF-kB expression and NF-kB–DNA binding activity in loss of cartilage in chronic OA (34). Indeed, complete deletion of inflamed SM has been reported by us and others (26, 27). Our RelA has been shown to result in chondrocyte apoptosis, whereas present findings strengthen previous observations of NF-kB acti- the heterozygous RelA knockout led to suppression of OA as a vation in synovial tissues during inflammatory conditions but also result of decreased catabolic gene expression in AC in tissue- provide proof of continued NF-kB–mediated synovial inflamma- specific knockout mice models (35). The NF-kB system has tion in late-stage OA. been shown to regulate expression of both pro- and antiapoptotic Chronic synovial inflammation in late-stage OA may play a factors (36). The presence of both anabolic and catabolic genes crucial role in the activation or deactivation of the NF-kB system and the role of NF-kB pathway has been identified in OA cartilage in other connective tissues of the knee, such as AC. In several by large gene expression profiling studies (37–39). Taken together, studies, it has been suggested that NF-kB–mediated signaling an active NF-kB system regulates cell survival and proliferation, results in inflammatory responses by chondrocytes and subsequent whereas a restrained NF-kB system might reflect chondrocyte progressive extracellular matrix damage and cartilage destruction apoptosis or necrosis in late-stage OA. (28, 29). Upregulated NF-kB1 expression, both at the mRNA and NF-kB signaling pathways mediate critical events in the in- protein level, has been reported in cartilage tissues retrieved from flammatory response by promoting transcription of genes encod- patients with late-stage knee joint OA (12). However, we did not ing for cytokines such as TNF-a and IL-1b, which in turn induce find any differences in NFKB1 expression in residual cartilage production of IL-6, IL-8, and COX-2 and stimulate production of retrieved from OA patients compared with PM controls, and RELA MMPs (MMP-1, 3, 9, 13) by synoviocytes, mononuclear cells, or gene expression was lower in the OA cartilage. Furthermore, RelA chondrocytes (17, 40). In addition to regulation of TNF-a and Downloaded from protein levels were higher in the cytosolic compared with the IL-1b expression, NF-kB is also activated by these cytokines (41), nuclear fractions. Thus, we were unable to detect activation of the leading to the amplification and probable chronic activation of NF-kB system in AC from patients with advanced OA. Mediators NF-kB in synoviocytes or mononuclear cells and thus promoting produced and released from synovial cells as a consequence of synovial inflammation. The presence of a wide range of proin- NF-kB signaling can have an impact on NF-kB activation in the flammatory genes has been reported in extracted SM obtained

AC. Several in vivo and in vitro studies have provided evidence of from late-stage OA patients by genome-based microarray analysis http://www.jimmunol.org/ cross-talk and feedback mechanisms between subchondral bone techniques (42, 43), indicating the presence of synovial inflam- and cartilage in OA joints (30–32). For example, mediators de- mation in chronic OA. IL-6 and IL-8 levels are elevated in sy- rived from bone cultures altered cartilage matrix metabolism (31), novial fluid from OA patients (19), and these cytokines have been and factors secreted by chondrocytes induced osteoclastogenesis associated with inflammation and initiation of a cascade of reac- and subsequent subchondral bone loss in rats (30). On the con- tions that lead to the activation of proteases and subsequent car- trary, chondrocytes in hypertrophic cartilage stimulated osteoblast tilage degradation (44). Unexpectedly, we found that IL-6 and differentiation and subsequent bone matrix deposition (32). Fur- IL-8 gene expression were downregulated in synovial tissues de- thermore, the Wnt/b-catenin signaling pathway has been shown to spite increased NF-kB activation in patients with chronic OA in modulate bone-to-cartilage biochemical processes leading to the comparison with PM controls. Although an unexpected finding of by guest on September 26, 2021 onset of OA in mice (30), highlighting across-tissue communi- the current studies, decreased IL-6 and IL-8 expression may cation at the transcription level. Taken together, there are represent different stages of synovial inflammation with the mounting data indicating communication between bone and car- presence or absence of inflammatory and cartilage degradation tilage in the joint that can result in enhancement of either catabolic phases during OA progression. Indeed, clinical observations of or anabolic processes. At present, few studies have focused on good analgesic effects of nonsteroidal anti-inflammatory drugs in communication between synovium and cartilage. We found a some but not all OA patients are coupled to the degree of synovial negative association between synovial nuclear RelA levels and inflammation during OA progression. Further, it has been reported cartilage NF-kB–DNA binding as well as between synovial that NF-kB binding sites exist not only in the promotor regions of NF-kB–DNA binding and cartilage NFKB1 and MMP3 gene ex- the genes but also at the noncoding regions (45), and thus in- pression, findings that could indicate that synovial processes lead creased NF-kB–DNA binding in SM might not be reflected as to chondrocyte impairment and result in AC loss in humans. increased IL-6 or IL-8 gene expression in OA. However, further studies are warranted to delineate the underlying For the current studies, it is important to note that the PM control mechanisms of synovial/cartilage communication leading to car- group is not ideal. Although we did not observe any compromise on tilage degradation. the quality of the RNA isolated from the PM group, as the RQI of We did not observe higher nuclear RelA levels or NF-kB–DNA PM tissues isolates was not statistically different from those of the binding activity in cartilage tissues than were observed in the OA patients, the age and BMI differences between OA and PM synovial tissues, potentially indicating an inactive proteasome controls or the limited number of samples in PM group could have activity leading to modified NF-kB expression in cartilage. The influenced our results. Another, more intriguing possibility is that downregulated ubiquitin proteasome system activity in chon- our data reflect epigenetic changes, which have been reported as drocytes can lead to reduced maturation of p105 and less degra- one consequence of long-term activation of NF-kB (46). In support dation of the inhibitor IkB protein, and as a result, RelA/NF-kB1 of this notion, altered DNA methylation and histone acetylation would be retained in the cytoplasm (24), as we recorded higher have been observed in synovial fibroblasts collected from patients cytosolic RelA levels in the cartilage. Interestingly, an inactive with RA (47), and increased DNA methylation and decreased proteasome system has been identified in dying cells in Alzheimer histone acetylation in the IL-6 gene promoter in synovial fibro- disease that effects the translocation of RelA/NF-kB1 into the blasts collected from OA patients was reported recently (48). nucleus (33). However, the expressional imbalance and activity Thus, it is possible that long-term activation of NF-kB in the SM of proteasome system within distinct cells types of the SM and led to epigenetic changes that lowered the production of cytokines AC, with respect to their role in OA, needs to be further ex- such as IL-6 and IL-8. plored. Another possible cause behind an inactive NF-kB system MMP3 is one of the several proteases involved in the patho- is that it could be due to decreased chondrocyte proliferation or genesis of OA through its ability to degrade proteoglycans (15). 1926 NF-kB ACTIVATION IN KNEE JOINT OSTEOARTHRITIS

Elevated MMP3 levels, along with other proteases, have been 13. Ramos, Y. F., W. den Hollander, J. V. Bove´e, N. Bomer, R. van der Breggen, N. Lakenberg, J. C. Keurentjes, J. J. Goeman, P. E. Slagboom, R. G. Nelissen, reported in serum and synovial fluid from patients with both RA et al. 2014. Genes involved in the osteoarthritis process identified through ge- and OA (49, 50). NF-kB regulates transcription of MMP1 and nome wide expression analysis in articular cartilage; the RAAK study. PLoS One MMP3 in human OA synoviocytes (40). In the current studies, we 9: e103056. 14. Pulai, J. I., H. Chen, H. J. Im, S. Kumar, C. Hanning, P. S. Hegde, and found that, although not changed in cartilage, MMP3 expression R. F. Loeser. 2005. NF-kappa B mediates the stimulation of cytokine and che- was increased in the SM in late-stage OA. Moreover, an associ- mokine expression by human articular chondrocytes in response to fibronectin ation between synovial nuclear RELA and MMP3 was observed, fragments. J. Immunol. 174: 5781–5788. 15. Delhalle, S., R. Blasius, M. Dicato, and M. Diederich. 2004. A beginner’s guide indicating strong potential for NF-kB–mediated regulation of to NF-kappaB signaling pathways. Ann. N. Y. Acad. Sci. 1030: 1–13. inflammation-related changes in OA. In our earlier work, we 16. Goldring, S. R., and M. B. Goldring. 2004. The role of cytokines in cartilage k matrix degeneration in osteoarthritis. Clin. Orthop. Relat. Res. 427(Suppl.): S27– found that inhibition of NF- B led to downregulated MMP3 ex- S36. pression in the cartilage and reduced cartilage destruction and 17. Yang, Y., S. G. Gao, F. J. Zhang, W. Luo, J. X. Xue, and G. H. Lei. 2014. Effects joint swelling in an animal model of OA (51). The upregulated of osteopontin on the expression of IL-6 and IL-8 inflammatory factors in human knee osteoarthritis chondrocytes. Eur. Rev. Med. Pharmacol. Sci. 18: 3580–3586. MMP-3, along with several other proinflammatory and catabolic 18. Kaneko, S., T. Satoh, J. Chiba, C. Ju, K. Inoue, and J. Kagawa. 2000. Interleukin- genes, has been reported from microarray analysis of SM and AC 6 and interleukin-8 levels in serum and synovial fluid of patients with osteoar- collected from OA patients (38, 42). Thus, present and previous thritis. Cytokines Cell. Mol. Ther. 6: 71–79. 19. Mahmoud, R. K., A. K. El-Ansary, H. H. El-Eishi, H. M. Kamal, and N. H. El- findings suggest that synovial MMP3 activity is a possible path- Saeed. 2005. Matrix metalloproteinases MMP-3 and MMP-1 levels in sera and ophysiological mechanism underlying OA disease progression in synovial fluids in patients with rheumatoid arthritis and osteoarthritis. Ital. J. Biochem. 54: 248–257. humans. 20. Livshits, G., G. Zhai, D. J. Hart, B. S. Kato, H. Wang, F. M. Williams, and Taken together, our present findings provide comprehensive ev- T. D. Spector. 2009. Interleukin-6 is a significant predictor of radiographic knee Downloaded from idence of NF-kB activation in synovial tissues of patients with osteoarthritis: the Chingford study. Arthritis Rheum. 60: 2037–2045. k 21. Altman, R., E. Asch, D. Bloch, G. Bole, D. Borenstein, K. Brandt, W. Christy, chronic knee joint OA and raise the possibility of NF- B–mediated T. D. Cooke, R. Greenwald, M. Hochberg, et al. 1986. Development of criteria cross-talk between joint tissues. It also provides a foundation for the for the classification and reporting of osteoarthritis. Classification of osteoar- future development of effective pharmacotherapies targeting NF-kB thritis of the knee. Arthritis Rheum. 29: 1039–1049. 22. Magnani, M., R. Crinelli, M. Bianchi, and A. Antonelli. 2000. The ubiquitin- for inhibiting the synovial contributions to this disorder. dependent proteolytic system and other potential targets for the modulation of

nuclear factor-kB (NF-kB). Curr. Drug Targets 1: 387–399. http://www.jimmunol.org/ 23. Oeckinghaus, A., and S. Ghosh. 2009. The NF-kappaB family of transcription Acknowledgments factors and its regulation. Cold Spring Harb. Perspect. Biol. 1:a000034. We thank orthopedic surgeon Ingemar Gladh at the Ortho Center Stockholm 24. Olivotto, E., M. Otero, K. B. Marcu, and M. B. Goldring. 2015. Pathophysiology for patient recruitment and taking tissue samples during surgery. We thank of osteoarthritis: canonical NF-kB/IKKb-dependent and kinase-independent a Carola Ska¨rvinge for excellent assistance with logistics and database man- effects of IKK in cartilage degradation and chondrocyte differentiation. RMD Open 1(Suppl. 1): e000061. agement and the Stockholm Spine Center and Ortho Center Stockholm for 25. Lehmann, T., C. Murphy, D. G. Zahra, and M. L. Handel. 2002. Reduction of providing space for sample collection, preparation, and storage. induced nuclear factor-kappaB nuclear translocation and DNA binding by dexamethasone in human osteoarthritic synovial tissue ex- plants. J. Rheumatol. 29: 787–795. Disclosures 26. Han, Z., D. L. Boyle, A. M. Manning, and G. S. Firestein. 1998. AP-1 and NF-

The authors have no financial conflicts of interest. kappaB regulation in rheumatoid arthritis and murine collagen-induced arthritis. by guest on September 26, 2021 Autoimmunity 28: 197–208. 27. Ahmed, A. S., J. Li, M. Ahmed, L. Hua, T. Yakovleva, M. H. Ossipov, G. Bakalkin, and A. Stark. 2010. Attenuation of pain and inflammation in References adjuvant-induced arthritis by the proteasome inhibitor MG132. Arthritis Rheum. 1. Neogi, T., A. Guermazi, F. Roemer, M. C. Nevitt, J. Scholz, L. Arendt-Nielsen, 62: 2160–2169. C. Woolf, J. Niu, L. A. Bradley, E. Quinn, and L. F. Law. 2016. Association of 28. Roman-Blas, J. A., and S. A. Jimenez. 2006. NF-kappaB as a potential thera- joint inflammation with pain sensitization in knee osteoarthritis: the multicenter peutic target in osteoarthritis and rheumatoid arthritis. Osteoarthritis Cartilage osteoarthritis study. Arthritis Rheumatol. 68: 654–661. 14: 839–848. 2. Ayral, X., E. H. Pickering, T. G. Woodworth, N. Mackillop, and M. Dougados. 29. Chen, L. X., L. Lin, H. J. Wang, X. L. Wei, X. Fu, J. Y. Zhang, and C. L. Yu. 2005. Synovitis: a potential predictive factor of structural progression of medial 2008. Suppression of early experimental osteoarthritis by in vivo delivery of the tibiofemoral knee osteoarthritis – results of a 1 year longitudinal arthroscopic adenoviral vector-mediated NF-kappaBp65-specific siRNA. Osteoarthritis study in 422 patients. Osteoarthritis Cartilage 13: 361–367. Cartilage 16: 174–184. 3. Kapoor, M., J. Martel-Pelletier, D. Lajeunesse, J. P. Pelletier, and H. Fahmi. 30. Jiao, K., L. N. Niu, M. Q. Wang, J. Dai, S. B. Yu, X. D. Liu, and J. Wang. 2011. 2011. Role of proinflammatory cytokines in the pathophysiology of osteoar- Subchondral bone loss following orthodontically induced cartilage degradation thritis. Nat. Rev. Rheumatol. 7: 33–42. in the mandibular condyles of rats. Bone 48: 362–371. 4. Pelletier, J. P., J. Martel-Pelletier, and S. B. Abramson. 2001. Osteoarthritis, an 31. Westacott, C. I., G. R. Webb, M. G. Warnock, J. V. Sims, and C. J. Elson. 1997. inflammatory disease: potential implication for the selection of new therapeutic Alteration of cartilage metabolism by cells from osteoarthritic bone. Arthritis targets. Arthritis Rheum. 44: 1237–1247. Rheum. 40: 1282–1291. 5. Brown, K. D., E. Claudio, and U. Siebenlist. 2008. The roles of the classical and 32. Nurminskaya, M., C. Magee, L. Faverman, and T. F. Linsenmayer. 2003. alternative nuclear factor-kappaB pathways: potential implications for autoim- Chondrocyte-derived transglutaminase promotes maturation of preosteoblasts in munity and rheumatoid arthritis. Arthritis Res. Ther. 10: 212. periosteal bone. Dev. Biol. 263: 139–152. 6. Oeckinghaus, A., and S. Ghosh. 2009. The NF-kappaB family of transcription 33. Ross, J. M., L. Olson, and G. Coppotelli. 2015. Mitochondrial and ubiquitin factors and its regulation. Cold Spring Harb. Perspect. Biol. 1: a000034. proteasome system dysfunction in ageing and disease: two sides of the same 7. Baldwin, A. S., Jr. 1996. The NF-kappa B and I kappa B proteins: new dis- coin? Int. J. Mol. Sci. 16: 19458–19476. coveries and insights. Annu. Rev. Immunol. 14: 649–683. 34. Ku¨hn, K., D. D. D’Lima, S. Hashimoto, and M. Lotz. 2004. Cell death in car- 8. Naumann, M., and C. Scheidereit. 1994. Activation of NF-kappa B in vivo is tilage. Osteoarthritis Cartilage 12: 1–16. regulated by multiple phosphorylations. EMBO J. 13: 4597–4607. 35. Kobayashi, H., S. H. Chang, D. Mori, S. Itoh, M. Hirata, Y. Hosaka, 9. Marok, R., P. G. Winyard, A. Coumbe, M. L. Kus, K. Gaffney, S. Blades, Y. Taniguchi, K. Okada, Y. Mori, F. Yano, et al. 2016. Biphasic regulation of P. I. Mapp, C. J. Morris, D. R. Blake, C. Kaltschmidt, and P. A. Baeuerle. 1996. chondrocytes by Rela through induction of anti-apoptotic and catabolic target Activation of the nuclear factor-kappaB in human inflamed genes. Nat. Commun. 7: 13336. synovial tissue. Arthritis Rheum. 39: 583–591. 36. Matteucci, C., A. Minutolo, F. Marino-Merlo, S. Grelli, C. Frezza, A. Mastino, 10. Benito, M. J., E. Murphy, E. P. Murphy, W. B. van den Berg, O. FitzGerald, and and B. Macchi. 2015. Characterization of the enhanced apoptotic response to B. Bresnihan. 2004. Increased synovial tissue NF-kappa B1 expression at sites azidothymidine by pharmacological inhibition of NF-kB. Life Sci. 127: 90–97. adjacent to the cartilage-pannus junction in rheumatoid arthritis. Arthritis 37. Aigner, T., K. Fundel, J. Saas, P. M. Gebhard, J. Haag, T. Weiss, A. Zien, Rheum. 50: 1781–1787. F. Obermayr, R. Zimmer, and E. Bartnik. 2006. Large-scale gene expression 11. Benito, M. J., D. J. Veale, O. FitzGerald, W. B. van den Berg, and B. Bresnihan. profiling reveals major pathogenetic pathways of cartilage degeneration in os- 2005. Synovial tissue inflammation in early and late osteoarthritis. Ann. Rheum. teoarthritis. Arthritis Rheum. 54: 3533–3544. Dis. 64: 1263–1267. 38. Xu, Y., M. J. Barter, D. C. Swan, K. S. Rankin, A. D. Rowan, M. Santibanez- 12. Gu, R., N. Liu, S. Luo, W. Huang, Z. Zha, and J. Yang. 2016. MicroRNA-9 Koref, J. Loughlin, and D. A. Young. 2012. Identification of the pathogenic regulates the development of knee osteoarthritis through the NF-kappaB1 pathways in osteoarthritic hip cartilage: commonality and discord between hip pathway in chondrocytes. Medicine (Baltimore) 95: e4315. and knee OA. Osteoarthritis Cartilage 20: 1029–1038. The Journal of Immunology 1927

39. Geyer, M., S. Gra¨ssel, R. H. Straub, G. Schett, R. Dinser, J. Grifka, S. Gay, Distribution of NF-kappaB-binding sites across human 22. Proc. E. Neumann, and U. Mu¨ller-Ladner. 2009. Differential transcriptome analysis of Natl. Acad. Sci. USA 100: 12247–12252. intraarticular lesional vs intact cartilage reveals new candidate genes in osteo- 46. Imagawa, K., M. C. de Andre´s, K. Hashimoto, D. Pitt, E. Itoi, M. B. Goldring, arthritis pathophysiology. Osteoarthritis Cartilage 17: 328–335. H. I. Roach, and R. O. Oreffo. 2011. The epigenetic effect of glucosamine and a 40. Wang, Z., Y. Qiu, J. Lu, and N. Wu. 2013. Connective tissue growth factor nuclear factor-kappa B (NF-kB) inhibitor on primary human chondrocytes–im- promotes interleukin-1b-mediated synovial inflammation in knee osteoarthritis. plications for osteoarthritis. Biochem. Biophys. Res. Commun. 405: 362–367. Mol. Med. Rep. 8: 877–882. 47. Wada, T. T., Y. Araki, K. Sato, Y. Aizaki, K. Yokota, Y. T. Kim, H. Oda, 41. Fan, Z., B. Bau, H. Yang, and T. Aigner. 2004. IL-1beta induction of IL-6 and R. Kurokawa, and T. Mimura. 2014. Aberrant histone acetylation contributes to LIF in normal articular human chondrocytes involves the ERK, p38 and elevated interleukin-6 production in rheumatoid arthritis synovial fibroblasts. NFkappaB signaling pathways. Cytokine 28: 17–24. Biochem. Biophys. Res. Commun. 444: 682–686. 42. van Lent, P. L., A. B. Blom, R. F. Schelbergen, A. Slo¨etjes, F. P. Lafeber, 48. Yang, F., S. Zhou, C. Wang, Y. Huang, H. Li, Y. Wang, Z. Zhu, J. Tang, and W. F. Lems, H. Cats, T. Vogl, J. Roth, and W. B. van den Berg. 2012. Active M. Yan. 2017. Epigenetic modifications of interleukin-6 in synovial fibroblasts involvement of alarmins S100A8 and S100A9 in the regulation of synovial ac- from osteoarthritis patients. Sci. Rep. 7: 43592. tivation and joint destruction during mouse and human osteoarthritis. Arthritis 49. Yoshihara, Y., H. Nakamura, K. Obata, H. Yamada, T. Hayakawa, K. Fujikawa, Rheum. 64: 1466–1476. and Y. Okada. 2000. Matrix metalloproteinases and tissue inhibitors of metal- 43. Lambert, C., J. E. Dubuc, E. Montell, J. Verge´s, C. Munaut, A. Noe¨l, and loproteinases in synovial fluids from patients with rheumatoid arthritis or oste- Y. Henrotin. 2014. Gene expression pattern of cells from inflamed and normal oarthritis. Ann. Rheum. Dis. 59: 455–461. areas of osteoarthritis synovial membrane. Arthritis Rheumatol. 66: 960–968. 50. Green, M. J., A. K. Gough, J. Devlin, J. Smith, P. Astin, D. Taylor, and P. Emery. 44. Pola, E., P. Papaleo, R. Pola, E. Gaetani, F. C. Tamburelli, L. Aulisa, and 2003. Serum MMP-3 and MMP-1 and progression of joint damage in early C. A. Logroscino. 2005. Interleukin-6 gene polymorphism and risk of osteoar- rheumatoid arthritis. Rheumatology (Oxford) 42: 83–88. thritis of the hip: a case-control study. Osteoarthritis Cartilage 13: 1025–1028. 51. Ahmed, A. S., J. Li, H. Erlandsson-Harris, A. Stark, G. Bakalkin, and 45. Martone, R., G. Euskirchen, P. Bertone, S. Hartman, T. E. Royce, M. Ahmed. 2012. Suppression of pain and joint destruction by inhibition of the N. M. Luscombe, J. L. Rinn, F. K. Nelson, P. Miller, M. Gerstein, et al. 2003. proteasome system in experimental osteoarthritis. Pain 153: 18–26. Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021