soluble tumour necrosis factor receptors in synovial fluids from temporomandibular joints with...

Soluble tumour necrosis factor receptors insynovial fluids from temporomandibular jointswith painful anterior disc displacement withoutreduction and osteoarthritis$

Junji Uehara, Takuo Kuboki*, Takuo Fujisawa, Shunji Kojima,Kenji Maekawa, Hirofumi Yatani

Department of Oral and Maxillofacial Rehabilitation, Okayama University Graduate School ofMedicine and Dentistry, 2-5-1 Shikata-cho, Okayama 700-8525, JapanAccepted 22 July 2003

Introduction

Osteoarthritis (OA) is a widespread multifactorialdisease that is characterised by a gradual loss ofarticular cartilage, sclerotic and cystic changes inlocal bone and osteophyte formation. It is also wellrecognised that anterior disc displacement withoutreduction (ADDwoR) can cause osteoarthriticchanges (ADDwoR/OA) in the temporomandibular

Archives of Oral Biology (2004) 49, 133—142

KEYWORDS

Temporomandibular

joint (TMJ);

Osteoarthritis (OA);

TNFa; Soluble-form

receptor; Synovial fluid

Summary The objective of this study was to detect soluble-form tumour necrosisfactor receptors (sTNFRs) in temporomandibular joint (TMJ) synovial fluid aspirates,and to compare the sTNFR concentrations between painful anterior disc displacementwithout reduction and osteoarthritis (ADDwoR/OA) and asymptomatic TMJs.

Synovial fluid was sampled from the superior TMJ cavity of 11 painful ADDwoR/OAcases (mean age: 36.9 years) and 10 asymptomatic females (mean age: 24.7 years) bydiluted aspiration. The concentrations of sTNFR-I and -II in the synovial fluid weremeasured using human sTNFR-I and -II enzyme-linked immunosorbent assays. The totalprotein concentrations in synovial fluids were measured using a bicinchoninic acidprotein assay kit. All data were normalised to the total protein concentration of eachsample.

Two-way factorial analysis of variance and post hoc multiple comparison revealedthat: (1) mean normalised sTNFR-I and -II concentrations were higher in TMJ synovialaspirates from ADDwoR/OA patients than from healthy controls; (2) in the ADDwoR/OApatients and the healthy controls, the sTNFR-I concentration in TMJ synovial aspirateswas higher than the sTNFR-II concentration; and (3) high TMJ synovial aspirate sTNFR-IIseemed to be associated with less TMJ pain and a less restricted range of mouthopening in the ADDwoR/OA patients.

The concentrations of sTNFRs in TMJ synovial fluid are higher in the presence ofpainful ADDwoR/OA, which could modulate intracapsular inflammation.� 2003 Elsevier Ltd. All rights reserved.

Abbreviations: ADDwoR, anterior disc displacement withoutreduction; ANOVA, analysis of variance; ELISA, enzyme-linkedimmunosorbent assay; OA, osteoarthritis; TNFa, tumour necrosisfactor alpha; TNFR, tumour necrosis factor receptor

$A preliminary report was presented at the General Sessionand Exhibition of the Japanese Association of Dental Research,Matsudo, Chiba, Japan, 2000.

*Corresponding author. Tel.: þ81-86-235-6681;fax: þ81-86-235-6684.E-mail address: [email protected] (T. Kuboki).

0003–9969/$ — see front matter � 2003 Elsevier Ltd. All rights reserved.doi:10.1016/j.archoralbio.2003.07.001

joint (TMJ).1,2 However, the exact mechanism ofthe degenerative changes in the ADDwoR/OA TMJremains unclear. Evidence suggests that pro-inflam-matory cytokines play important roles in mediatingcartilage degradation in OA and rheumatoid arthritis(RA).

Tumour necrosis factor alpha (TNFa) is a potentpro-inflammatory cytokine produced by several celltypes.3 It suppresses the synthesis of cartilagematrix such as type II collagen and proteoglycansby cultured chondrocytes4 and enhances the synth-esis of matrix degrading enzymes including matrixmetalloproteinases, prostaglandins and plasmino-gen activators by synovial fibroblasts and articularchondrocytes in vitro.5—7 Indeed, TNFa has beendetected in synovial fluids in OA joints,8,9 includingTMJs.10—12

TNFa exerts these biological activities by bindingto two specific high-affinity cell surface receptors,TNF receptor (TNFR)-I and -II.13 The extracellulardomains of these receptors, which are cleaved byproteolytic activity14 or shed,15 exist in solubleforms.16 Recent findings suggest that these solublereceptors (sTNFR-I and -II) inhibit the biologicalactivity of TNFa by binding to it antagonisti-cally.17—20 Based on these findings, TNFa antago-nists, infliximab and etanercept, have already beenintroduced in the treatment of RA and Crohn’sdisease. Potential new uses for these drugs in amultitude of other inflammatory conditions areleading to new clinical trials.21

Since sTNFRs in synovial fluids from ADDwoR/OATMJs have not been documented, we examinedthe presence of sTNFRs in the synovial fluid frompainful TMJ ADDwoR/OA patients and tried tocorrelate their signs and symptoms to the sTNFRlevels. The null hypothesis tested in this study wasthat the mean sTNFR concentration of the synovialfluid does not differ between painful ADDwoR/OAand asymptomatic TMJs, and there is no relation-ship between the sTNFR concentration and theseverity of signs and symptoms of the ADDwoR/OA patients.

Materials and methods

Participants

Eleven female TMJ ADDwoR/OA patients with amean age of 36:9 � 16:9 years (range: 17—65 years)were selected from a consecutive series of patientswho attended our clinic in Okayama UniversityDental Hospital from April 1998 to March 2000.Inclusion criteria for patient selection were: (1)they had complained of chronic TMJ pain for more

than 3 months, even though they had receivedsome palliative treatments [soft-food diet instruc-tion, non-steroidal anti-inflammatory drug (NSAID)administration, gentle mouth-opening exercises,stabilisation appliance]; (2) there was tomographicand magnetic resonance imaging (MRI) evidence ofobvious degenerative changes in the hard tissueand ADDwoR of the affected TMJ (pain-persistentside of the TMJ); (3) they had moderate or severeTMJ pain aggravated by jaw movement; and (4) theyreceived pumping manipulation or arthrocentesisprocedure in the clinic to manage their chronicTMJ pain. They did not complain of any other jointpain. If they were diagnosed as having polyarthritis(e.g. RA, American Rheumatism Association, 1987),they were excluded from the sample.

Ten female volunteers with a mean age of 24:7�1:7 years (range: 22—27 years) were recruited fromamong the students and staff of Okayama UniversityDental School to serve as controls. Inclusion criteriafor the controls were: (1) good physical health;(2) no history of TMJ pain and dysfunction, norany other joint pain; (3) maximum range of mouthopening (RMO, interincisal distance) of over 40 mm;and (4) no joint noise, no condylar translatoryrestriction and no tenderness during palpation ofthe lateral aspect of the TMJs during maximalmouth open-and-close cycles. This study protocolwas approved by a human protective committee atour institution (approval number #21), and informedconsent was obtained from each participant prior tothe commencement of this study.

Synovial fluid samples

Synovial fluid samples were collected from thesuperior joint cavity using diluted aspiration asdescribed by Kanyama et al.22 The skin surfacewas disinfected, then a local anesthetic (1% xylo-caine, 1.0 ml) was injected hypodermally into thelateral aspect of the target TMJ. Synovial fluidswere aspirated from the affected side of theADDwoR/OA patients, and from the right TMJ ofcontrols. Physiological saline (1.0 ml) was injectedinto the upper joint cavity using a 23-gauge hypo-dermic needle. The solution was aspirated after itwas repeatedly injected and withdrawn more than10 times. The mean recovery of the diluted synovialfluid was 0.85 ml (range: 0.50—1.0 ml). The aspi-rates were centrifuged at 3000 � g for 15 min toremove cells and particulate matter. The totalprotein concentration of the aspirated synovial fluidwas measured using a bicinchoninic acid (BCA) pro-tein assay kit (Protein Assay Reagent Kit, PIERCE, IL,USA). Aliquots of the supernatant were frozen andstored at �80 8C. In all of the patients, it was the

134 J. Uehara et al.

first experience of arthrocentesis or pumping pro-cedure.

Measurement of sTNFR concentrations insynovial fluid aspirates

The concentrations of sTNFR-I and -II in the synovialfluids were measured using human sTNFR-I and -IIsandwich enzyme-linked immunosorbent assay(ELISA) kits (Quantikine, R&D Systems, MN, USA).Briefly, 50 ml of assay diluent and 200 ml of synovialfluid sample were pipetted into 96-well polystyrenemicroplates coated with a murine monoclonal anti-body against sTNFR-I (or sTNFR-II) and incubated for2 h at room temperature. The plate was washedthree times with 400 ml of wash buffer, then eachsoluble receptor conjugate (polyclonal antibodyagainst each soluble receptor conjugated to horse-radish peroxidase) was added to the wells andincubated for 1 h (sTNFR-II) or 2 h (sTNFR-I) at roomtemperature. The plate was washed three timeswith wash buffer, and 200 ml of substrate was addedto the wells and incubated for 20 min at roomtemperature. The incubation for sTNFR-II pro-ceeded in the dark. The reaction was stopped with50 ml of stop solution (2N sulphuric acid), thenabsorbance at 450 nm was measured using a micro-plate reader. All reagents were from the ELISA kit.The minimum detectable doses of sTNFR-I and -IIwere 3.0 and 1.0 pg/ml, respectively.

Joint imaging

Images of bilateral TMJs were obtained from eachpatient using multiple-layered X-ray tomographyand MRI. Tomography using an Optiplanimat (Sie-mens Co. Ltd., Germany) with spiral 458 projectionswas performed in the sagittal plane at 2 mm thick-ness and 2 mm intervals to detect hard tissue defor-mities in the joint. Changes in the joints visualisedon multiple-layered tomograms were classified intothree levels:22 OA1 represents changes (erosion,sclerosis, etc.) with no obvious joint outline defor-mity; OA2 represents clear joint outline deformities(flattening, beaking, etc.) without shortening of thecondyle; and OA3 represents clear condylar short-ening with all types of changes. Joint deformitieswere estimated by an expert who was blind to theclinical findings and ELISA data. Joints were alsoscanned by MRI using a 1.5 T Magnetom (SiemensCo. Ltd., Germany) with proton density and T2-weighted imaging sequences. The scanning planewas set perpendicular to the long axis of the coronalcondylar outline at a thickness of 3 mm. The discposition was diagnosed, and joint pathology was subcategorised according to Orsini et al.23

Clinical examinations

The participants were examined clinically using astructured protocol. Inter- and intra-examiner relia-bility levels of each examination item in the protocolwere confirmed as acceptable prior to the study.24

Clinical examination items consisted of painintensity and RMO (mm) at synovial fluid sampling.Pain intensity was measured with a 100 mm visualanalogue scale (VAS), based on the subject’s ownperception of jaw pain at rest and while chewing atthe synovial fluid sampling visit. Anchors to bothscales were labelled as ‘no pain’ and ‘extremelysevere pain’. VAS values were measured by a milli-metre ruler and expressed in mm/100 mm. The RMOwas measured at maximal voluntary painless open-ing (painless), maximal voluntary opening (max),and maximal opening while the examiner appliedpressure between the maxillary and mandibularteeth using the fingers (stretch), just before thesynovial fluid sampling.

Statistical analysis

The mean and standard deviation of total proteinand sTNFR-I and -II concentrations in the ADDwoR/OA patients and controls were calculated and com-pared statistically using a two-way factorial analysisof variance (ANOVA) test with two predictors:sTNFR-I versus sTNFR-II and ADDwoR/OA versuscontrol group. These statistical analyses were pro-ceeded simultaneously using non-parametric meth-ods: the Kruskal—Wallis one-way ANOVA followedby the Mann—Whitney U-test. However, the statis-tical trend was the same as in the ANOVA test.

In addition, we analysed correlations between ageand sTNFR concentrations (Spearman’s rank correla-tion coefficient), between symptom severity (TMJpain, RMO) and sTNFR concentrations (Pearson’scorrelation coefficient), and between hard tissuechange (OA levels) in the TMJ and sTNFR concentra-tions (Spearman’s rank correlation coefficient) usinga statistical package (Abacus Concepts, Stat View4.5J, Abacus Concepts Inc., Berkeley, USA). Thecut-off level for significance was set at a ¼ 0:05.

Results

Demographic and symptom severitycharacteristics of ADDwoR/OA patients andcontrols

Table 1 shows a comparison of the demographicdata and severity of the signs and symptomsbetween ADDwoR/OA and controls. The mean age

Soluble from TNF receptors and osteoarthritic TMJ 135

Table 1 Demographic information and signs and symptoms of the ADDwoR/OA patients and asymptomatic controls.

Subjectno.

Age(years)

Pain level (VAS) RMO (mm) Imaging data (right joint) Imaging data (left joint) NSAIDs restdurationa

At rest Chewing Painless Max Stretch MRI findings Effusion X-ray findings MRI findings Effusion X-ray findings

Control1 23 0 0 49 49 50 b b b b b b —2 22 0 0 50 50 51 b b b b b b —3 23 0 0 44 44 45 b b b b b b —4 24 0 0 48 48 49 b b b b b b —5 25 0 0 44 44 45 b b b b b b —6 25 0 0 44 44 45 b b b b b b —7 25 0 0 42 42 43 b b b b b b —8 26 0 0 50 50 51 b b b b b b —9 27 0 0 43 43 44 b b b b b b —10 27 0 0 43 43 44 b b b b b b —

Mean 24.7 0 0 45.7 45.7 46.7S.D. 1.7 0 0 3.2 3.2 3.2

ADDwoR/OA1 26 12 43 25 27 34 ADDwoRc � OA1 ADDwoR � OA1 2 weeks2 31 10 32 24 27 29 ADDwoRc þ OA1 ADDwR � Normal >1 month3 38 70 78 27 31 32 Normal � Normal ADDwoRc þ OA2 04 60 34 35 28 29 30 Normal � OA1 ADDwoRc þ OA1 >1 month5 25 0 10 35 41 42 ADDwoRc � OA1 Normal � Normal >1 month6 24 0 67 24 27 29 ADDwoRc þ OA1 Normal � OA1 1 week7 58 5 5 38 42 46 ADDwoRc þ OA3 ADDwR � Normal >1 month8 22 0 36 29 32 34 Normal � Normal ADDwoRc þ OA1 4 days9 65 20 33 40 43 44 Normal � Normal ADDwoRc � OA1 >1 month10 17 0 27 29 35 36 ADDwoRc þ OA1 Normal � OA1 011 40 18 45 25 27 34 ADDwoRc � OA2 ADDwoR � OA2 2 weeks

Mean 36.9 15.4 37.4 29.5 32.8 35.5S.D. 16.9 21.1 21.4 5.7 6.4 6.0P-value 0.035 0.033 <0.001 <0.001 <0.001 <0.001

VAS: visual analogue scale of pain (1—100); RMO: maximum range of mouth opening; ADDwR: anterior disc displacement with reduction; ADDwoR: anterior discdisplacement without reduction; OA1: no severe joint outline deformity; OA2: joint with clear outline deformity without shortening; OA3: joint with clear shorteningof the condyle; NSAID: non-steroidal anti-inflammatory drug. Effusion was diagnosed by T2-weighted magnetic resonance imaging. P-values were calculated by one-tailed Student’s t-test.a Since all of the ADDwoR/OA patients were prescribed NSAIDs as a palliative care, NSAIDs administration status was expressed as ‘NSAIDs rest duration’ beforesynovial fluid sampling.b Data not taken.c Examined side.

136J.

Uehara

et

al.

of the ADDwoR/OA patients was higher than that ofthe asymptomatic controls. The maximal RMO wassignificantly lower in the ADDwoR/OA patients, andmean joint pain severity (at rest and during chew-ing) was significantly higher in the ADDwoR/OApatients than in the controls.

Regarding the symptoms of the contralateral sideTMJ, no symptoms were experienced in 10 of the 11ADDwoR/OA patients. One patient (#11) had slightTMJ pain in the contralateral side during maximaljaw opening, but the main symptomatic side wasthe joint fluid sampling side.

Since all of the ADDwoR/OA patients had beenprescribed an NSAID (diclofenac sodium, Voltaren,Ciba-Geigy, Tokyo, Japan), 25 mg, three times perday, it is possible that the medication affects thesTNFR concentration. As shown in Table 1, NSAIDadministration status was investigated as NSAID restduration before synovial fluid sampling. There wasno clear relationship between the sTNFR levels andthe NSAID administration status before synovialfluid sampling.

Total protein concentration in the synovialfluid samples

The mean total protein concentration in theaspirated synovial fluid of the ADDwoR/OA TMJswas higher than that in the control TMJs. However,the difference was not statistically significant(P ¼ 0:062, Table 2).

Normalised sTNFR-I and -II concentrationsin synovial fluid samples

The mean and standard deviation of sTNFR-I and -IIconcentrations in synovial fluid from the ADDwoR/OA and control TMJs are shown in Table 2. Normal-isation against the total protein concentration ineach synovial fluid sample did not affect the generalstatistical trend.

Two-way factorial ANOVA performed on the nor-malised sTNFR-I and -II data revealed significantmain effects for the two predictors (ADDwoR/OAversus control: P < 0:001; sTNFR-I versus sTNFR-II

Table 2 The results of biological assays of the synovial fluid aspirates from the ADDwoR/OA and control TMJs.

Subjectno.

Total protein (TP)(mg/ml)

sTNFR-I(pg/ml)

sTNFR-II(pg/ml)

sTNFR-I/TP(pg/mg)

sTNFR-II/TP(pg/mg)

Control1 1580.8 94.2 39.1 59.6 24.72 1409.9 96.8 46.1 68.7 32.73 1224.9 85.4 43.2 69.7 35.34 1976.5 105.7 48.4 53.5 24.55 941.0 57.2 11.6 60.8 12.36 675.0 32.5 12.1 48.1 17.97 614.0 23.8 8.4 38.8 13.78 1590.0 96.2 12.7 60.5 8.09 950.0 41.8 18.8 44.0 19.810 993.0 38.7 14.0 39.0 14.1

Mean 1195.5 67.2 25.4 54.3 20.3S.D. 439.2 31.5 16.5 11.4 8.9

ADDwoR/OA1 1272.5 145.6 72.5 114.4 57.02 1964.0 211.2 57.0 107.5 29.03 4408.0 594.9 165.9 135.0 37.64 3130.0 172.3 141.8 55.0 45.35 2971.3 297.1 198.4 100.0 66.86 1868.0 207.2 73.8 110.9 39.57 997.0 212.8 90.4 213.4 90.78 1020.0 111.0 69.6 108.8 68.29 1284.0 116.7 76.2 90.9 59.310 1070.7 108.4 41.2 101.2 38.511 1362.0 166.9 42.0 122.5 30.8

Mean 1940.7 213.1 93.5 114.5 51.2S.D. 1108.1 138.5 52.0 38.5 19.0P-value 0.062 0.004 0.001 <0.001 <0.001

P-values were calculated by one-tailed Student’s t-test.


difference: P < 0:001). The interactions betweenthem were statistically significant (P ¼ 0:047).

One-way factorial ANOVA followed by Fisher’sPLSD analysis revealed significant mean concen-tration differences between ADDwoR/OA sTNFR-Iand ADDwoR/OA sTNFR-II (P < 0:001), betweenADDwoR/OA sTNFR-I and control sTNFR-I(P < 0:001), between ADDwoR/OA sTNFR-II andcontrol sTNFR-II (P ¼ 0:004), and between controlsTNFR-I and control sTNFR-II (P ¼ 0:002).

Correlation between age and normalisedsTNFR-I and -II concentrations

Since age was not closely matched between theADDwoR/OA and controls, we analysed the relation-ship between the normalised sTNFR concentrationsand age. The normalised sTNFR concentrations werenot significantly related to age (sTNFR-I versusage: r ¼ 0:142, n ¼ 21, P ¼ 0:525; sTNFR-II versusage: r ¼ 0:093, n ¼ 21, P ¼ 0:679). The normalised

Figure 1 Correlation between TMJ symptom severity and normalised sTNFR-I concentration in ADDwoR/OA synovialfluid aspirates. Pearson’s correlation coefficients were calculated and scattergrams were constructed. (A) TMJ pain atrest vs. normalised sTNFR-I. (B) TMJ pain during chewing vs. normalised sTNFR-I. (C) RMO-painless vs. normalisedsTNFR-I. (D) RMO-max vs. normalised sTNFR-I. (E) RMO-stretch vs. normalised sTNFR-I. No examined TMJ symptomseverity indices were significantly related to sTNFR-I concentrations in ADDwoR/OA synovial fluid aspirates.


sTNFR-I concentration was significantly related tothe concentration of sTNFR-II (r ¼ 0:771, n ¼ 21,P < 0:001).

Correlation between symptom severity andnormalised sTNFR-I and -II concentrations inADDwoR/OA synovial fluid aspirates

We analysed the relationship between symptomseverity (TMJ pain and RMO) and normalised

sTNFR-I and -II concentrations in the ADDwoR/OApatients using a correlation analysis (Figs. 1 and 2).The normalised sTNFR-I concentration was not sig-nificantly related to any symptom severity indices(for example, TMJ pain during chewing versus nor-malised sTNFR-I: r ¼ �0:193, P ¼ 0:580; RMO-stretch versus normalised sTNFR-I: r ¼ 0:458,P ¼ 0:161). However, the normalised sTNFR-II con-centration was significantly and negatively relatedto some of the symptom severity indices (TMJ pain

Figure 2 Correlation between TMJ symptom severity and normalised sTNFR-II concentration in ADDwoR/OA synovialfluid aspirate. Pearson’s correlation coefficients were calculated and scattergrams were constructed. (A) TMJ pain atrest vs. normalised sTNFR-II. (B) TMJ pain during chewing vs. normalised sTNFR-II. (C) RMO-painless vs. normalisedsTNFR-II. (D) RMO-max vs. normalised sTNFR-II. (E) RMO-stretch vs. normalised sTNFR-II. Some examined TMJ symptomseverity indices, e.g. TMJ pain during chewing, RMO-painless, RMO-max, RMO-stretch, were significantly related tosTNFR-II concentrations in ADDwoR/OA synovial fluid aspirates.


during chewing versus normalised sTNFR-II:r ¼ �0:619, P ¼ 0:041; RMO-stretch versus normal-ised sTNFR-II; r ¼ 0:768, P ¼ 0:004). In other words,as the sTNFR-II concentration decreased, theADDwoR/OA signs and symptoms became moresevere (e.g. high pain level, low RMO).

Correlation between OA level detected bytomography and normalised sTNFR-I and -IIconcentrations in ADDwoR/OA synovialfluid aspirates

We analysed the relationship between OA leveldetected by tomography and normalised sTNFR-Iand -II concentrations in ADDwoR/OA patients.The normalised sTNFR-I concentration was signifi-cantly related to the OA level (r ¼ 0:783, n ¼ 11,P ¼ 0:013), but the normalised sTNFR-II concentra-tion was not (r ¼ �0:029, n ¼ 11, P ¼ 0:927).

Discussion

Since TNFa has been identified in some TMJ synovialfluid,10,25—29 this implied that sTNFRs should bedetected in the synovial fluid of OA TMJs. Wedetected sTNFRs in the TMJ synovial fluid fromOA patients and from asymptomatic controls. Toour knowledge, we are the first to detect sTNFRs inTMJ synovial fluid.

Interest in the roles of sTNFRs in a variety ofdiseases has increased recently.30—32 For example,Girardin et al. noted that levels of sTNFR-I and -IIinitially increased as levels of TNFa increased inchildren with severe meningococcal septicaemia.However, at TNFa levels above 500 pg/ml, the con-centration of soluble receptors did not increase,further and this was associated with a high mortalityrate.30 Thus, the investigators concluded that animbalance between TNFa and its naturally occurringinhibitors is partly responsible for the increasedmorbidity and mortality. Van Zee et al. observedthat sTNFR concentrations increased with inflamma-tion in experimental bacteraemia, and that suffi-cient levels circulate in critically ill patients to blockTNFa cytotoxicity in vitro and in vivo.31 This suggeststhat in experimental bacteraemia, sTNFRs protectagainst the TNF-mediated effects associated withseptic shock. Rooney et al. also demonstrated thatthe increased joint destruction associated with thepolyarthritis subtype of juvenile chronic arthritis(JCA) compared with the systemic OA-like subtypesis related to lower sTNFRs/TNFa ratios.19

These findings indicate that the sTNFR concen-trations increase as TNFa is upregulated at inflam-matory sites. However, the ratio of sTNFR versus

the TNFa concentration varies in accordance withthe nature of disease. For example, active polyar-thritic conditions (RA, JCA) are associated with arelatively low sTNFR/TNFa ratio, presumably due tohigh TNFa expression related to the auto-immunereaction occurring in local arthritic tissues. In con-trast to the systemic polyarthritis conditions, theTNFa level in the OA joint is known to be relativelylow, while the mean sTNFR concentrations in the OATMJs were also upregulated.19 This finding supportsthe notion that the increased level of sTNFR is aprotective reaction that blocks or inhibits theexacerbating side effects induced by secreted TNFa.Since sTNFRs can be upregulated by pro-inflamma-tory cytokines such as IL-1b and TNFa, this type offeedback inhibition mechanism seems to be effec-tive in preventing aggressive tissue destruction inTMJ OA. This would explain why TMJ OA is not aprogressive/degenerative disease by nature.

In this study, the mean synovial fluid sTNFR-Iconcentration in the OA joints was significantlyhigher than that of sTNFR-II. In contrast, the oppo-site is true of polyarthritic conditions. This discre-pancy can be explained by the findings in humanchondrocyte culture, indicating that normal humanchondrocytes express mainly TNFR-I, the levels ofwhich remain unchanged, while the expression ofTNFR-II is significantly upregulated by IL-1b, TNFaand bFGF.33 In other words, concentrations ofTNFR-I do not change significantly with variationsin the disease condition, whereas those of TNFR-IIincrease as intra-articular inflammation worsens.Since the pro-inflammatory cytokine level in thesynovial aspirates is lower in OA than in the poly-arthritis conditions, we speculate that the sTNFR-IIlevel would also be lower in OA. This would accountfor the lower sTNFR-II/I ratio found in the presentstudy. Since Prieur et al. have also shown thatsTNFR-II but not TNFa correlates with fever spikesin children with systemic polyarthritic JCA,34 thesTNFR-II/I ratio could be used as a good diseasemarker that would help diagnose the involvement ofauto-immuno-reaction-related components in anunclear pathological condition.35

Interestingly, the synovial fluid sTNFR-II (but notsTNFR-I) concentration was negatively correlated tothe symptom severity of ADDwoR/OA in the presentstudy. In other words, as the sTNFR-II concentrationdecreased, the ADDwoR/OA symptoms becamemore severe. At a glance, this finding does not seemto be consistent with the above discussion that thesTNFR-II correlates positively with the severity ofjoint inflammation. However, when OA is not asso-ciated with a component related to an auto-immuno-reaction, the incremental drive for sTNFR-II seemsvery low. Therefore, the constitutive expression


level of the natural inhibitors (soluble receptors)would be a determinant of symptom modificationability in OA joints. Secreted TNFa could bindrapidly to the soluble receptors and consume themin an unknown mechanism. Regardless, the anti-inflammatory feedback mechanism of the solublereceptor may sometimes be insufficient due to thelow constitutive sTNFR-II level in the OA joints.

On the other hand, normalised sTNFR-I (but notsTNFR-II) concentration was positively related tothe OA bone change level detected by X-ray tomo-graphy. Since the sTNFR-I concentration was notrelated to ongoing joint pain and RMO, increasedsTNFR-I in some patients with severe joint defor-mity can be related to the past history of intra-articular inflammation and degenerative process.Unfortunately, there were only three patients withmoderate and severe joint deformity in thisADDwoR/OA sample. To get a valid and reliableconclusion on this point, we should increase thesample size in the future study.

In summary, this study shows that: (1) meannormalised sTNFR-I and -II concentrations werehigher in TMJ synovial aspirates from theADDwoR/OA patients than from the healthy con-trols; (2) in the ADDwoR/OA patients and thehealthy controls, the sTNFR-I concentration inTMJ synovial aspirates was higher than the sTNFR-II concentration; and (3) high TMJ synovial aspiratesTNFR-II seemed to be associated with less TMJ painand less restricted RMO in the ADDwoR/OA patients.

Therefore, novel therapies may include agentsthat specifically mimic the anti-inflammatory feed-back mechanism of sTNFR-II in vivo. Enbrel (etaner-cept), a potent inhibitor (sTNFR) of TNFa andlymphotoxin, has recently received approval forthe treatment of refractory RA. This product hashad a revolutionary effect on the treatment of RA,and may also play a role in the treatment of chronicrefractory TMD.36,37

Acknowledgements

This research was supported in part by a Grant-In-Aid from the Ministry of Education, Culture, Sports,Science and Technology of Japan (#12557169,#14571843 and #15592050).

References

1. Ali AM, Sharawy MM. Histopathological changes in rabbitcraniomandibular joint associated with experimentallyinduced anterior disk displacement (ADD). J Oral PatholMed 1994;23:364—74.

2. Stegenga B. Osteoarthritis of the temporomandibular jointorgan and its relationship to disc displacement. J OrofacPain 2001;15:193—205.

3. Aggarwal BB, Natarajan K. Tumour necrosis factors: devel-opments during the last decade. Eur Cytokine Netw 1996;7:93—124.

4. Saklatvala J. Tumour necrosis factor alpha stimulatesresorption and inhibits synthesis of proteoglycan in carti-lage. Nature 1986;322:547—9.

5. Campbell IK, Piccoli DS, Roberts MJ, Muirden KD, HamiltonJA. Effects of tumor necrosis factor alpha and beta onresorption of human articular cartilage and production ofplasminogen activator by human articular chondrocytes.Arthritis Rheum 1990;33:542—52.

6. Mitchell PG, Cheung HS. Tumour necrosis factor alpha andepidermal growth factor regulation of collagenase andstromelysin in adult porcine articular chondrocytes. J CellPhysiol 1991;149:132—40.

7. Arend WP, Dayer JM. Cytokines and cytokine inhibitors orantagonists in rheumatoid arthritis. Arthritis Rheum1990;33:305—15.

8. Westacott CI, Whicher JT, Barnes IC, Thompson D, Swan AJ,Dieppe PA. Synovial fluid concentration of five differentcytokines in rheumatic diseases. Ann Rheum Dis 1990;49:676—81.

9. Di Giovine FS, Nuki G, Duff GW. Tumour necrosis factor insynovial exudates. Ann Rheum Dis 1988;47:768—72.

10. Takahashi T, Kondoh T, Fukuda M, Yamazaki Y, Toyosaki T,Suzuki R. Proinflammatory cytokines detectable in synovialfluids from patients with temporomandibular disorders.Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1998;85:135—41.

11. Nordahl S, Alstergren P, Kopp S. Tumour necrosis factor-alpha in synovial fluid and plasma from patients withchronic connective tissue disease and its relation totemporomandibular joint pain. J Oral Maxillofac Surg2000;58:525—30.

12. Fu K, Ma X, Zhang Z, Chen W. Tumour necrosis factor insynovial fluid of patients with temporomandibular disor-ders. J Oral Maxillofac Surg 1995;53:424—6.

13. Bazzoni F, Beutler B. The tumor necrosis factor ligand andreceptor families. N Engl J Med 1996;334:1717—25.

14. Porteu F, Brockhaus M, Wallach D, Engelmann H, Nathan CF.Human neutrophil elastase releases a ligand-binding frag-ment from the 75-kDa tumor necrosis factor (TNF) receptor:comparison with the proteolytic activity responsible forshedding of TNF receptors from stimulated neutrophils.J Biol Chem 1991;266:18846—53.

15. Ohe H, Takashiba S, Naruishi K, et al. Tumour necrosisfactor-alpha (TNF-alpha)-induced and interleukin-1 beta(IL-1 beta)-induced shedding of TNF receptors from gingivalfibroblasts. J Interferon Cytokine Res 2000;20:1077—82.

16. Arend WP, Dayer JM. Inhibition of the production andeffects of interleukin-1 and tumor necrosis factor alpha inrheumatoid arthritis. Arthritis Rheum 1995;38:151—60.

17. Camussi G, Lupia E. The future role of anti-tumour necrosisfactor (TNF) products in the treatment of rheumatoidarthritis. Drugs 1998;55:613—20.

18. Alaaeddine N, DiBattista JA, Pelletier JP, et al. Osteoar-thritic synovial fibroblasts possess an increased level oftumor necrosis factor-receptor 55 (TNF-R55) that mediatesbiological activation by TNF-alpha. J Rheumatol 1997;24:1985—94.

19. Rooney M, Varsani H, Martin K, Lombard PR, Dayer JM, WooP. Tumour necrosis factor alpha and its soluble receptors injuvenile chronic arthritis. Rheumatology (Oxford) 2000;39:432—8.


20. Pisetsky DS. Tumour necrosis factor blockers in rheumatoidarthritis. N Engl J Med 2000;342:810—1.

21. Andreas MR. New indications for treatment of chronicinflammation by TNF-blockade. Am J Med Sci 2003;325:75—92.

22. Kanyama M, Kuboki T, Kojima S, et al. Matrix metallopro-teinases and tissue inhibitors of metalloproteinases insynovial fluids of patients with temporomandibular jointosteoarthritis. J Orofac Pain 2000;14:20—30.

23. Orsini MG, Kuboki T, Terada S, Matsuka Y, Yatani H,Yamashita A. Clinical predictability of temporomandibularjoint disc displacement. J Dent Res 1999;78:650—60.

24. Matsuka Y, Itoh S, Minakuchi H, Kuboki T, Yamashita A.Validity of questionnaire for epidemiological studies onsymptoms of temporomandibular disorders. J Jpn Soc TMJ1997;9:80—91.

25. Shafer DM, Assael L, White LB, Rossomando EF. Tumournecrosis factor-alpha as a biochemical marker of pain andoutcome in temporomandibular joints with internal de-rangements. J Oral Maxillofac Surg 1994;52:786—91.

26. Emshoff R, Puffer P, Rudisch A, Gassner R. Temporoman-dibular joint pain: relationship to internal derangementtype, osteoarthrosis, and synovial fluid mediator level oftumor necrosis factor-alpha. Oral Surg Oral Med Oral PatholOral Radiol Endod 2000;90:442—9.

27. Sandler NA, Buckley MJ, Cillo JE, Braun TW. Correlation ofinflammatory cytokines with arthroscopic findings in pa-tients with temporomandibular join internal derangements.J Oral Maxillofac Surg 1998;56:534—43.

28. Segami N, Miyamaru M, Nishimura M, Suzuki T, Kaneyama K,Murakami K. Does joint effusion on T2 magnetic resonanceimages reflect synovitis? Part 2. Comparison of concentra-tion levels of proinflammatory cytokines and total protein insynovial fluid of the temporomandibular joint with internalderangements and osteoarthrosis. Oral Surg Oral Med OralPathol Oral Radiol Endod 2002;94:515—21.

29. Kaneyama K, Segami N, Nishimura M, Suzuki T, Sato J.Importance of proinflammatory cytokines in synovial fluidfrom 121 joints with temporomandibular disorders. Br JOral Maxillofac Surg 2002;40:418—23.

30. Girardin E, Roux-Lombard P, Grau GE, et al. Imbalancebetween tumor necrosis factor-alpha and soluble TNFreceptor concentrations in severe meningococcaemia.Immunology 1992;76:20—3.

31. Van Zee KJ, Kohno T, Fischer E, Rock CS, Moldawer LL,Lowry SF. Tumour necrosis factor soluble receptors circu-late during experimental and clinical inflammation and canprotect against excessive tumor necrosis factor alpha invitro and in vivo. Proc Natl Acad Sci USA 1992;89:4845—9.

32. Feldmann M. What is the mechanism of action of anti-tumour necrosis factor-alpha antibody in rheumatoidarthritis? Int Arch Allergy Immunol 1996;111:362—5.

33. Alsalameh S, Mattka B, Al-Ward R. Preferential expressionof tumor necrosis factor receptor 55 (TNF-R55) on humanarticular chondrocytes: selective transcriptional upregula-tion of TNF-R75 by proinflammatory cytokines interleukin1beta, tumor necrosis factor-alpha, and basis fibroblastgrowth factor. J Rheumatol 1999;26:645—53.

34. Prieur ARM, Oux-Lombard P, Dayer JM. Dynamics of feverand the cytokine network in systemic juvenile arthritis. RevRhum Engl Ed 1996;63:163—70.

35. Kuboki T, Hattori T, Mizushima T, et al. Detection ofspecific antibodies against human cultured chondrosarcoma(HCS-2/8) and osteosarcoma (Saos-2) cells in the serum ofpatients with osteoarthritis of the temporomandibularjoint. Arch Oral Biol 1999;44:403—14.

36. Garrison L, McDonnell ND. Etanercept: therapeutic use inpatients with rheumatoid arthritis. Ann Rheum Dis 1999;58:165—9.

37. Kacena MA, Merrel GA, Konda SR, Wilson KM, Xi Y, HorowitzMC. Inflammation and bony changes at the temporoman-dibular joint. Cells Tissues Organs 2001;169:257—64.


soluble tumour necrosis factor receptors in synovial fluids from temporomandibular joints with...

Documents