Annals ofthe Rheumatic Diseases 1995; 54: 970-975

Determination of interstitial collagenase (MMP-1)in patients with rheumatoid arthritis

Shingo Maeda, Takashi Sawai, Miwa Uzuki, Yuichi Takahashi, Hiroshi Omoto,Masayuki Seki, Minoru Sakurai

AbstractObjectives-To investigate whether inter-stitial collagenase (MMP-1) concentrationin synovial fluid can be useful as a markerfor disease activity in rheumatoid arthritis(RA), to determine the main route bywhich collagenase degrades the matrix ofarticular cartlage, and to investigate ifan imbalance between metalloproteinases(MMPs) and tissue inhibitor of metallo-proteinases (TIMP) is responsible for theactivity ofMMPs in RA.Methods-Collagenase concentrations weremeasured in synovial fluid and pairedserum samples using a specific sandwichenzyme linked immunosorbent assay.Coliagenase activities were also assayed insynovial fluid samples. Synovial tissuesobtained from the same patient wereexamined by inmunohistochemical stain-ing and the numbers of cells expressingcollagenase were counted.Results-Collagenase concentrations insynovial fluid did not correlate with Creactive protein and collagenase levels inserum, but did correlate positively withthe degree of synovial inflammation, andincreased with increasing numbers ofcells identified as expressing collagenasein synovial tissue. Collagenase activitiesdid not correlate with TIMP-1 concen-trations, but did correlate strongly withthe ratios of collagenase concentration toTIMP-1 (r=0-73).Conclusion-The collagenase concen-tration in synovial fluid cannot be used asa marker for systemic disease activity, butcan be used as a marker for the degreeof synovial inflammation in the jointfrom which the sample is aspirated. Inadvanced RA, most of the collagenase isprobably produced in synovial lining cellsand released into synovial fluid, where itdegrades the matrix of articular cartilage.An imbalance between MMP and TIMPmay be of importance in the degradationof extraceilular matrix of articular carti-lage in RA.

(Ann Rheum Dis 1995; 54: 970-975)

Rheumatoid arthritis (RA) is characterisedby chronic inflammation, which leads todegradation of the extracellular matrix anddistortion of the architecture and function of

the joints. The matrix metalloproteinases(MMPs) are a family of enzymes that candegrade all components of the extracellularmatrix.' Interstitial collagenase (E.C.3.4.24.7;MMP-1) is a member of the MMP gene familythat cleaves the collagen triple helix to yieldcharacteristic 1/4-3/4 products.2

Explants from human rheumatoid synovialtissue have been found to release large amountsof collagenase,3 and increased collagenaseactivity has been reported in rheumatoidsynovial fluid.4 Collagenase was localised byimmunostaining on the extracellular matrixcomponents at the cartilage-pannus junction inRA,5 and immunohistochemical staining hassubsequently shown it to be localised intra-cellularly in hyperplastic synovial lining cells ofthe rheumatoid synovium.6 However, it hasalso been reported that chondrocytes produceproteinases capable of degrading osteoarthriticand normal human cartilage,7 8 and that thematrix of cartilage was severely degraded whenliving cartilage was placed in contact withsynovium.9

Collagenases in cell and tissue cultures arecommonly synthesised and secreted as inactive,latent forms.'0 Procollagenase can be activatedin culture medium by trypsin,"1 12 chaotrophicagents,'13 14 and mercurial compounds,'5 andin vivo by stromelysin (MMP-3).'6 Van Wartet al have proposed the 'cysteine switch'activation mechanism for members of thematrix metalloproteinase gene family,'7 but theprocess by which activation is achieved in vivoremains unclear. Cawston et al purified acollagenase inhibitor synthesised by newbornrabbit bone in tissue culture'8 that bound topurified collagenase with a 1:1 stoichiometry.The inhibitor, named TIMP (tissue inhibitorof metalloproteinase), interacted with theactive form of collagenase to form a tightenzyme-inhibitor complex that differed fromlatent collagenase.'9 Dean et al proposedthat an imbalance of MMPs and TIMP con-tributed to the pathogenesis of osteoarthritis(OA).20There have been no reports of studies in

which synovial fluid, serum, and synoviumwere obtained from the same RA patient andMMP contents of the samples measured. Inthis study we assayed collagenase concen-trations in synovial fluid and serum to inves-tigate whether those in synovial fluid couldbe useful as a marker for disease activity inRA. We also compared synovial fluid con-centrations of collagenase with the numbers

Department ofOrthopaedic Surgery,Tohoku University,School ofMedicine,Sendai 980-77, JapanS MaedaM SekiM SakuraiDepartment of Clinicaland LaboratoryMedicineY TakahashiDepartment ofPathology,Tohoku UniversityHospital,School ofMedicine,Sendai 980-77, JapanM UzukiT SawaiKanebo Institute forCancer Research,Osaka 534, JapanH OmotoCorrespondence to:Dr Shingo Maeda,Department of OrthopaedicSurgery,Tohoku University,School of Medicine,Sendai 980-77, Japan.Accepted for publication28 July 1995

970

on Novem

ber 11, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.54.12.970 on 1 D

ecember 1995. D

ownloaded from

Determination of interstitial collagenase (MMP-1) in patients with rheumatoid arthritis

of cells expressing collagenase in synovialtissue (demonstrated by immunohistochemicalstaining) in order to determine the main path-way taken by collagenase to degrade the matrixof articular cartilage in RA. In addition, wemeasured collagenase activities and concen-trations of collagenase and TIMP to investigateif an imbalance between MMPs and TIMP isresponsible for the activity ofMMPs in RA.

Patients and methodsMONOCLONAL ANTIBODIES AGAINST

INTERSTITIAL COLLAGENASE (MMP-1)Mouse monoclonal antibodies (clone K5E 1,K2F7, K4H 1 1) against interstitial collagenase(MMP-1) purified from serum free culturemedium of P12-4 fibrosarcoma were preparedby means of a modification of a methoddescribed previously.2' Purified procollagenasehad an Mr value of 52 000. Analysis of theeffect of monoclonal antibodies on collagenaseby immunoblotting showed that the antibodiesreacted specifically with collagenase that hadidentical Mr values.

PATIENT SELECTION AND PREPARATION OF

SAMPLESSynovial tissue, synovial fluid, and serumwere obtained from patients with RA, all ofwhom met the American College of Rheu-matology (formerly the American RheumatismAssociation) criteria for the diagnosis of RA.22Synovial tissue was obtained at the time of totalknee joint replacement surgery, and synovialfluid was aspirated from the same joint. Thesynovial tissue was fixed in 4% para-formaldehyde dissolved in phosphate bufferedsaline solution (PBS) for two hours, and thenwrapped in paraffin. Synovial fluid and serumsamples were centrifuged to remove cells anddebris, and stored frozen at -70°C untilrequired for assay.

IMMUNOHISTOCHEMICAL STUDIES AND

QUANTIFICATION OF POSmIVE CELLSFor the immunohistochemical studies, sectionsof synovial tissue 2 ,um thick were mountedon slides. Endogenous peroxidase activity wasblocked with 1% hydrogen peroxide inmethanol for 30 minutes, followed by rinsingwith PBS for 15 minutes. Normal horse serumwas applied for 15 minutes as a blocking agent.This was followed by rinsing with PBS andapplication of the primary antibodies. The nextday, the sections were rinsed with PBS andbiotinylated secondary antibodies applied for30 minutes. Streptavidin peroxidase was appliedfor 30 minutes. The sections were then exposedto diaminobenzidine and dilute hydrogenperoxide for seven minutes. The reaction wasstopped with tap water, and the sectionsdehydrated and mounted.The numbers of cells shown to express

collagenase were counted on at least 20different microscopic fields and the numbersaveraged and graded on the following semi-quantitative scale: grade 0 = absent; 1 + = 1-5

cells per high power field (hpf) (x400);2+ = 6-15 cells/hpf; 3+ = >15 cells/hpf.

DOUBLE ANTIBODY SANDWICH ENZYME LINKEDIMMUNOSORBENT ASSAY (ELISA) FORCOLLIAGENASECollagenase concentrations were assayed usinga specific sandwich ELISA. Ninety six wellmicrotitre plates were coated overnight at 4°Cwith 10 ,ug/ml of K5E1 in carbonate bufferedsaline solution. The plates were rinsed in0 05% Tween 20-PBS, and incubated for onehour at room temperature (15°C) with BlockAceR (Dainippon Seiyaku Co Ltd) in PBS toblock non-specific protein binding in the wells.(Block AceR is a blocking reagent that contains1% milk protein and organic acid buffer.) Theblocker was rinsed from the plates and standardrecombinant procollagenase or samples addedfor two hours at room temperature. Synovialfluid samples were diluted 1:1000 for assaywith Block AceR; serum samples were diluted1:10. The plates were then rinsed three timesand incubated with K2F7 for two hours atroom temperature. The antibodies were rinsedfrom the plates, which were then incubatedwith streptavidin peroxidase for 30 minutesat room temperature before rinsed fourtimes, followed by 10 minutes incubation witho-phenylenediamine. The reaction was stoppedby addition of normal sulphuric acid, andthe absorbance at 490 nm was measuredspectrophotometrically.The assays detected both procollagenase

and activated collagenase binding TIMPwith almost the same sensitivity. Free activecollagenase was assayed to a slightly differentdegree of sensitivity, but a dose-response curvefor free active collagenase was obtained. TIMPcould not be assayed adequately.

ONE STEP SANDWICH ENZYME IMMUNOASSAY(EIA) FOR TIMP-1TIMP-1 concentrations in synovial fluid weremeasured by means of sandwich EIA Kitspurchased from Fuji Chemical Industries Ltd.The sandwich EIA quantitatively measures notonly free TIMP but also TIMP complexedwith collagenase.2>25

INFIAMMATION SCORESThe degree of synovial inflammation wasanalysed on haematoxylin and eosin stainedsections using a modification of a methoddescribed previously.26 Inflammation scoreswere determined as the sum of three com-ponents: synovial lining thickness (0 = 1-2 cells;1+ = 3-4 cells; 2+ = 5-6 cells; 3+ = >6 cells),lymphoid follicles (0 = 0-1 per low powerfield (lpf) (X40); 1+ = 2-3/lpf; 2+ = 4-6/lpf;3+ = >6/lpf), and interfollicular features(0 = complete fibrosis; 1+ = mononuclear cellinfiltration <33% of lpf; 2+ = 33-67%;3+ = >67%). They were then classified intofour groups according to the semiquantitativescores: 0 = class 0; 1-3 = 1+; 4-6 = class 2+;7-9 = class 3+.

971

on Novem

ber 11, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.54.12.970 on 1 D

ecember 1995. D

ownloaded from

Maeda, Sawai, Uzuki, et al

COLLAGENASE ACTIVITY ASSAYS

Collagenase activities were assayed using themethod reported previously.27 In this study,synovial fluids were assayed for collagenaseactivities after activation with L-1-tosylamide-2-phenylethyl chloromethyl ketone (TPCK)treated trypsin. Fluorescein isothiocyanatelabelled collagen was used as substrate. Oneunit of collagenase activity was defined as theamount of enzyme degrading 1 ,ug of collagenper minute under the conditions used.

STATISTICAL ANALYSISData were analysed by the non-parametricMann-Whitney U test, Student's paired t test,and Pearson correlation coefficients. P valuesless than 0-05 were considered significant.

ResultsPATIENTS AND SAMPLE PREPARATION

Synovial fluid and serum samples wereobtained from 41 patients (37 women andfour men; mean age 60 years (range 32 to 89years)). Synovial tissues were taken from 35 ofthese patients. TIMP-1 concentrations wereassayed in samples from 35 patients, andcollagenase activities measured in samplesfrom 23.

RELATIONSHIP BETWEEN COLLAGENASECONCENTRATIONS IN PAIRED SYNOVIAL FLUID

AND SERUM SAMPLESWe measured collagenase concentrations inpaired synovial fluid and serum samples from41 patients. Collagenase concentrations insynovial fluids (mean 2361 (SD 2628) ng/ml)were significantly greater than those in sera(25-6 (21 9) ng/ml) (p < 0-0001). No obviousrelationship existed between the levels ofcollagenase in synovial fluid and serum(r=0-207, p=0-187), and we found norelationship between the concentrations insynovial fluid and those of C reactive proteinin serum, or between collagenase concen-tration in synovial fluid and the ARA (1949)Stage & Class criteria for the progression ofrheumatoid arthritis and the functional

: i Ir

c

00

(n-

0) cmCa-

0U-CL)

12 000

10 000

8000

6000

4000

2000

0

**

*-I a11

8k8+ a

A

0

AIA

0 1+ 2+ 3+Inflammation scores (class)

Figure 1 Relationship between synovialfluidconcentrations ofcollagenase and inflammation scores.*p < 0 05; **p < 0 005.

capacity of patients with the disease (data notshown).28

RELATIONSHIP BETWEEN SYNOVIAL FLUID

CONCENTRATIONS OF COLLAGENASE ANDINFLAMMATION SCORESFigure 1 illustrates the relationship betweenthe concentration of collagenase in synovialfluid and the inflammation score classes. Sixpatients were in class 3+, 11 were in class 2+,12 were in class 1 +, and six were in class 0. Theconcentrations of collagenase in synovial fluidfrom patients with class 3+ inflammationranged from 1991 to 11712 ng/ml (mean (SD)4954 (3723) ng/ml (n = 6)); the values for class2+ patients ranged from 478 to 11158 ng/ml(3790 (3163) ng/ml (n = 11)); those of class 1+patients were from 320 to 6398 ng/ml (2400(2182) ng/ml (n = 12)); and those of class 0patients ranged from 97 to 1820 ng/ml (953(707) ng/ml (n = 6)). A positive correlation wasobserved between the degree of synovialinflammation and the collagenase concen-tration in synovial fluid.

RELATIONSHIP BETWEEN GRADE FOR POSITIVE

CELL COUNT AND COLLAGENASECONCENTRATIONS IN SYNOVIAL FLUID

Collagenase was immunolocalised in hyper-plastic synovial lining cells in rheumatoidsynovium. Figure 2 shows a haematoxylinand eosin stained section of synovial tissue

Figure 2 Haematoxylin and eosin stained section ofrheumatoid synovial tissue (left), and immunolocalisation ofcollagenase in the same rheumatoid synovial tissue, showing positive cells in the hyperplastic synovial lining cells (right).Bar represents 100 ,um.

972

T6 .-

il )'- .....54.

on Novem

ber 11, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.54.12.970 on 1 D

ecember 1995. D

ownloaded from

Determination of interstitial coUagenase (MMP-1) in patients with rheumatoid arthritis

12 000 -*_C 10 0000

(D 8000CE

6000

C,~~~~~ ,o 4000 o3 !{ 1u_ 20000 1

0 1+ 2+ 3+Number of positive cells (grade)

Figure 3 Relationship between the density ofpositive ceUsdemonstrated by immunohistochemical staining andcoUagenase concentrations in synovialfluid (SF).*p < 0-01; **p < 0-001.

and a serial section from the same synoviumanalysed for expression of collagenase byimmunohistochemical staining. Very littlestaining was found in chondrocytes and fibro-blasts at the cartilage-pannus junction.With our system of grading by the number

of cells immunohistochemically positive forcollagenase expression, nine patients weregrade 3+, nine were grade 2+, five weregrade 1 +, and 11 were grade 0. Figure 3 showsthe collagenase concentrations in synovial fluidtaken from patients belonging to each grade;concentrations in patients from grade 3+ranged from 1189 to 11712 ng/ml (mean (SD)4430 (3197) ng/ml (n = 9)); in grade 2+patients they were from 924 to 11158 ng/ml(4954 (3157) ng/ml (n = 9)); those in grade 1+patients ranged from 1609 to 3451 ng/ml(2346 (754) ng/ml (n = 5)); and concen-trations in grade 0 patients ranged from 96 to2040 ng/ml (842 (712) ng/ml (n= 11)).Collagenase levels in synovial fluid tended toincrease with increasing numbers of positivecells, and a positive correlation existed betweenthe degree of collagenase expression in synovialtissue and the collagenase concentrations insynovial fluid.

RELATIONSHIP BETWEEN CONCENTRATIONS OF

COLLAGENASE AND TIMP-1 IN SYNOVIAL FLUID

In 35 synovial fluid samples measured, nopositive correlation was found between theconcentrations of collagenase and TIMP-1(r = 0-155, p = 0-3744).

220020001800

1600140012001000

800600400200'

0 0

C°ba

0a

a

00

r= 0.235p=0.2814

0 10 20 30 40 50 60 70 80Collagenase activity (U/ml)

Figure 4 Relationship between collagenase activities andTIMP-1 concentrations in synovialfluid.

positive correlation between collagenase activityand these ratios (r= 0 73, p = 0 0001).

DiscussionThis study has shown: (i) collagenase concen-trations in synovial fluid did not correlate withC reactive protein or collagenase concen-trations in serum; (ii) collagenase levels insynovial fluid correlated positively with thedegree of synovial inflammation; (iii) collagenaselevels in synovial fluid increased withincreasing numbers of cells positive for theexpression of collagenase as demonstrated byimmunohistochemical staining; (iv) no positivecorrelation between collagenase and TIMP-1concentrations; (v) collagenase activity did notcorrelate with TIMP-1 concentrations, but (vi)did correlate strongly with the ratios of theconcentrations of collagenase and TIMP-1.

In this study we examined collagenaseconcentrations in paired synovial fluid andserum samples taken from RA patients. Noobvious relationship existed between the levelsof collagenase in synovial fluid and serum, andthe concentrations in synovial fluid correlatedwith neither C reactive protein concentrationsin the serum nor ARA Stage or Class of diseaseprogression. These data imply that synovialfluid collagenase concentration cannot be usedas a marker for systemic disease activity in RA.

RELATIONSHIP BETWEEN COLLAGENASEACTIVITIES AND CONCENTRATIONS OF TIMP-1

IN SYNOVIAL FLUIDWe found no obvious relationship between thecollagenase activities and the concentrations ofTIMP- 1 in 23 synovial fluid samples measured(r=0-235,p=0-2814) (fig4).

RELATIONSHIP BETWEEN COLLAGENASEACTIVITIES AND THE RATIOS OF COLLAGENASECONCENTRATION TO TIMP-1 IN SYNOVIAL FLUID

Figure 5 illustrates the relationship betweencollagenase activities and the ratios of theconcentrations of collagenase and TIMP-1 in21 synovial fluid samples. There was a strong

6

CU

axn

CoCU

cJ0)Co

._00

Co

5

4

. .

.31

2

1 * .Ua

0 10 20 30 40 50 60 70 80Collagenase activity (U/ml)

Figure 5 Relationship between collagenase activities andthe ratios of the concentration ofcollagenase to that ofTIMP-1 in synovialfluid.

973

on Novem

ber 11, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.54.12.970 on 1 D

ecember 1995. D

ownloaded from

Maeda, Sawai, Uzuki, et al

We also compared synovial fluid concen-trations with the degree of synovial inflam-mation. Using in situ hybridisation, Firesteinet al reported a positive correlation between thedegree of synovial inflammation and expressionof collagenase mRNA in synovial tissue.29 Inour study, concentrations of collagenase insynovial fluid correlated with the degree ofsynovial inflammation. These data comp-lement previous reports that cytokines derivedfrom inflammatory cells, such as interleukin-1and tumour necrosis factor, can modulate theamount of collagenase synthesised.3"34 Whilesynovial fluid concentration of collagenasecannot be used as a marker for systemic diseaseactivity, it can be used as a marker for thedegree of synovial inflammation in the jointfrom which the sample is aspirated.

Collagenase has been immunolocalised inhyperplastic lining cells in rheumatoid syn-ovium using an immunohistochemical method,6and in situ hybridisation studies have alsodemonstrated production of collagenase mRNAin rheumatoid synovial lining cells.29 35 36We applied monospecific antibodies againstcollagenase to synovium, cartilage, and thecartilage-pannus junction in the RA joint andwere able to immunolocalise collagenase insynovial lining cells; very little staining wasfound in chondrocytes and fibroblasts atthe cartilage-pannus junction. Woolley et alreported immunolocalisation of collagenase inthe extracellular matrix at the cartilage-pannusjunction, but the cells producing collagenasecould not be identified.5 We graded synovialtissue according to the number of cells shownby immunohistochemical staining to expresscollagenase, and compared the grades withthe concentrations of collagenase found inthe synovial fluid. The concentrations ofcollagenase in the synovial fluid tended toincrease with increasing numbers of positivelystaining cells.Three apparently distinct processes are

involved in cartilage degradation: hyper-trophic synovial cells produce MMPs andrelease them into synovial fluid, where theenzymes degrade the articular cartilage matrix;fibroblasts at the cartilage-pannus junctionsecrete MMPs and directly degrade thecartilage matrix; and chondrocytes produceand secrete MMPs and then digest their ownmatrix. Fell et al reported marked degradationof both proteoglycans and collagen of thematrix when living cartilage was placed incontact with synovium,9 and Sapolsky et alhave shown that chondrocytes of normal adultcartilage produce proteinases that are capableof degrading proteoglycans.7 Ehrlich et aldemonstrated a collagenolytic enzyme inosteoarthritic and normal human cartilage,8and Shinmei et al have suggested that autocrinemodulation of chondrocyte metabolism maybe the cause of cartilage breakdown in OA.37

In the present study, collagenase wasimmunolocalised in synovial lining cells, butvery little staining was found in the chondro-cytes of articular cartilage and in the fibroblastsat the cartilage-pannus junction. A relationshipwas demonstrated between the degree of

expression of collagenase in synovial tissue andthe concentration of collagenase in synovialfluid. These data suggest that, as far ascollagenase is concerned, synovial cells aremore important in the breakdown of thecartilage matrix than are chondrocytes. It hasalso been suggested that collagenase partici-pates in the destruction of the entire cartilagesurface, not just that of a localised area suchas the cartilage-pannus junction. Hasty et alstudied rats with collagen induced arthritis andproposed that chondrocytes may play a directpart in the earliest stages of degradation oftheirown matrix.38 In the present study, weobtained all samples at the time of total kneejoint replacement surgery and thus cannotcomment on the importance of chondrocytesin the initial stage of RA. Our findings dosuggest that, in advanced RA, most of thecollagenase is produced in synovial lining cellsand released into synovial fluid, where itdegrades the articular cartilage matrix.Both active and latent collagenase have been

shown to be present in rheumatoid synovialfluid,39 40 and an inhibitor of collagenasedemonstrated in newborn rabbit bone, TIMP,has been shown to bind to activatedcollagenase.18 19 It has been proposed that animbalance between MMPs such as collagenaseand TIMP contributes to the pathogenesis ofOA.20 We measured collagenase activities andTIMP- 1 concentrations in synovial fluid, usingan assay that included activation with trypsin.It is known that collagenase inhibitors such asTIMP-1, TIMP-2, and O2-macroglobulin maybe inactivated by trypsin treatment, and thusthe results of our assay may be overestimates.However, little active collagenase has beendetected in synovial fluid,4" and the assay is notcapable of measuring collagenase activity with-out the addition of an activating agent such astrypsin. We found no positive correlationbetween collagenase and TIMP- 1 concen-trations. This result is consistent with thatreported by Clark et al.42

It has been reported that procollagenase issecreted as both a minor glycosylated form anda major polypeptide.43 Both the 57 kDa and the52 kDa procollagenase could be activated bytrypsin, generating their respective 47 kDa and42 kDa active enzyme forms.43 In the presentstudy, the extent to which procollagenase con-tributes to the total (latent + active) concen-tration of collagenase in synovial fluid couldnot be determined; even when the concen-tration of total collagenase was assayed, anaccurate molarity could not be calculated. It haspreviously been reported that no appreciableamount of apparently active collagenase waspresent in synovial fluid, the majority of thecollagenase being in the latent form.4' If this isso, the ratio of concentrations of collagenaseand TIMP will correspond approximately withthe ratio of the molarity between them. Wefound a strong positive correlation betweencollagenase activity and the collagenase:TIMPratio (r= 073, p = 00001). These data supportthe concept that an imbalance between MMPand TIMP is responsible for the degradation ofextracellular matrix which occurs in RA.

974 on N

ovember 11, 2020 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.54.12.970 on 1 Decem

ber 1995. Dow

nloaded from

Determination of interstitial collagenase (MMP-1) in patients with rheumatoid arthritis

We thank Dr A Rikimaru, Narugo National Hospital of Miyagi,and Dr K Sato, Tohoku Rosai Hospital of Sendai, for providingsamples obtained from patients. We also thank Mr K Shoji forhis skilful help with photographs.

1 Matrisian L. Metalloproteinases and their inhibitors inmatrix remodelling. Trends Genet 1990; 6: 121-5.

2 Evanson J M, Jeffrey J J, Krane S M. Human collagenase:identification and characterization of an enzyme fromrheumatoid synovium in culture. Science 1967; 158:499-502.

3 Evanson J M, Jeffrey J J, Krane S M. Studies on collagenasefrom rheumatoid synovium in tissue culture. Jf Clin Invest1968; 47: 2639-51.

4 Harris E D Jr, DiBona D R, Krane S M. Collagenase inhuman synovial fluid. CGin Invest 1969; 48: 2104-13.

5 Woolley D E, Crossley M J, Evanson J M. Collagenase atsites of cartilage erosion in the rheumatoid joint. ArthritisRheum 1977; 20: 1231-9.

6 Okada Y, Gonoji Y, Nakanishi I, Nagase H, Hayakawa T.Immunohistochemical demonstration of collagenase andtissue inhibitor of metalloproteinases (TIMP) in synoviallining cells of rheumatoid synovium. Virchows Arch B CellPathol 1990; 59: 305-12.

7 Sapolsky A T, Keiser H, Howell D S, Woessner J F Jr.Metalloproteinasesss of human articular cartilage thatdigest cartilage proteoglycan at neutral and acid pH.J Clin Invest 1976; 58: 1030-41.

8 Ehrlich M G, Mankin H J, Jones H, Wright R, Crispen C.Vigliani G. Collagenase and collagenase inhibitors inosteoarthritis and normal human cartilage. J Clin Invest1977; 59: 226-33.

9 Fell H B, Jubb R W. The effect of synovial tissue on thebreakdown of articular cartilage in organ culture. ArthritisRheum 1977; 20: 1359-71.

10 Nagase H, Jackson R C, Brinckerhoff C E, Vater C A,Harris E D Jr. A precursor form of latent collagenaseproduced in a cell-free system with mRNA from rabbitsynovial cells. J Biol Chem 1981; 256: 11951-4.

11 Vaes G. The release of collagenase as an inactive proenzymeby bone explants in culture. Biochem J 1972; 126:275-89.

12 Stricklin G P, Bauer E A, Jeffrey J J, Eisen A Z. Human skincollagenase: isolation of precursor and active forms fromboth fibroblasts and organ cultures. Biochemistry 1977;16: 1607-15.

13 Shinkai H, Nagai Y. A latent collagenase from embryonichuman skin explants. J Biochem 1977; 81: 1261-8.

14 Sakamoto S, Sakamoto M, Matsumoto A, Goldhaber P,Glimcher M J. Latent collagenase from culturemedium of embryonic chick bones. FEBS Lett 1978; 88:53-8.

15 Vater C A, Mainardi C L, Harris E D. Activation in vitroof rheumatoid synovial collagenase from cell cultures.JClin Invest 1978; 62: 987-92.

16 Murphy G, Cockett M I, Stephens P E, Smith B J,Docherty A J P. Stromelysin is an activator of pro-collagenase. BiochemJ 1987; 248: 265-8.

17 Van Wart H E, Birkedal-Hansen H. The cystein switch: aprinciple of regulation of metalloproteinase activity withpotential applicability to the entire matrix metallo-proteinase gene family. Proc NadAcad Sci USA 1990; 87:5578-82.

18 Cawston T E, Galloway W A, Mercer E, Murphy G,Reynolds J J. Purification of rabbit bone inhibitor ofcollagenase. BiochemJz 1981; 195: 159-65.

19 Cawston T E, Murphy G, Mercer E, Galloway W A,Hazleman B L, Reynolds J J. The interaction of purifiedrabbit bone collagenase with purified rabbit bone metallo-proteinase inhibitor. BiochemJ 1983; 211: 313-8.

20 Dean D D, Martel-Pelletier J, Pelletier J P, Howell D S,Woessner J F. Evidence for metalloproteinase andmetalloproteinase inhibitor imbalance in human osteo-arthritic cartilage. J7 Clin Invest 1989; 84: 678-85.

21 Birkedal-Hansen B, Birkedal-Hansen H, Robert E T,Ajit S B, Birkedal-Hansen H. Monoclonal antibodies tohuman fibroblast collagenase. Inhibitor of enzymaticactivity, affinity purification of the enzyme, and evidencefor clustering of epitopes in the NH2-terminal end of theactivated enzyme. Biochemnistry 1988; 27: 67541-8.

22 Arnett F C, Edworthy S M, Bloch D A, et al. The AmericanRheumatism Association 1987 revised criteria for theclassification of rheumatoid arthritis. Arthritis Rheum1988; 31: 315-24.

23 Kodama S, Yamashita K, Kishi J, Iwata K, Hayakawa T. Asandwich enzyme immunoassay for collagenase inhibitorusing monoclonal antibodies. Matrix 1989; 9: 1-6.

24 Kodama S, Iwata K, Iwata H, Yamashita K, Hayakawa T.Rapid one-step sandwich enzyme immunoassay for tissueinhibitor of metalloproteinases: an application forrheumatoid arthritis serum and plasma. Jf ImmunolMethods 1990; 127: 103-8.

25 Hayakawa T, Yamashita K, Kodama S, Iwata H, Iwata K.Tissue inhibitor of metalloproteinases and collagenaseactivity in synovial fluid of human rheumatoid arthritis.Biomedical Res 1991; 12: 169-73.

26 Rooney M, Whelan A, Feighery C, Breshinhan B. Changesin lymphocyte infiltration of membrane and the clinicalcourse of rheumatoid arthritis. Arthritis Rheum 1989; 32:361-9.

27 Terato K, Hashida R, Miyamoto K, et al. Histological,immunological and biochemical studies on type IIcollagen-induced arthritis in rats. Biomedical Res 1982; 3:495-505.

28 Steinbrocker 0, Traeger C H, Batterman R C. Therapeuticcriteria in rheumatoid arthritis. JAMA 1949; 140:659-72.

29 Firestein G S, Paine M M, Littman B H. Gene expression(collagenase, tissue inhibitor of metalloproteinases,complement, and HLA-DR) in rheumatoid arthritis andosteoarthritis synovium: quantitative analysis and effect ofintraarticular corticosteroids. Arthritis Rheum 1991; 34:1094-105.

30 Murphy G, Reynolds J J, Werb Z. Biosynthesis of tissueinhibitor of metalloproteinases by human fibroblasts inculture: stimulation by 12-0-tetradecanoylphorbol13-acetate and interleukin 1 in parallel with collagenase.J Biol Chem 1985; 260: 3079-83.

31 Mizel S B, Dayer J M, Krane S M, Merenhagen S E.Stimulation of rheumatoid synovial cell collagenaseand prostaglandin production by partially purifiedlymphocyte-activating factor (interleukin 1). Proc NailAcad Sci USA 1981; 78: 2474-7.

32 Dayer J M, Beutler B, Cerami A. Cachectin/tumor necrosisfactor stimulates collagenase and prostaglandin E2production by human synovial cells and dermal fibro-blasts. JExp Med 1985; 162: 2163-8.

33 Postlethwaite A E, Raghow R, Stricklin G P, Poppletoro H,Seyer J M, Kang A H. Modulation of fibroblast functionsby interleukin 1: increased steady-state accumulation oftype 1 procollagen messenger RNAs and stimulation ofother functions but not chemotaxis by humanrecombinant interleukin 1 ax and ,3. J Cell Biol 1988; 106:311-8.

34 McCachren S S, Greer P K, Niedel J E. Regulation ofhuman synovial fibroblast collagenase messenger RNA byinterleukin-1. Arthritis Rheum 1989; 32: 1539-45.

35 McCachren S S, Haynes B F, Niedel J E. Localization ofcollagenase mRNA in rheumatoid arthritis synovium byin situ hybridization histochemistry. Jf Clin Immunol 1990;10: 19-27.

36 Gravallese E M, Darling J M, Ladd A L, Katz J N,Glimcher L H. In situ hybridization studies of stromelysinand collagenase messenger RNA expression in rheu-matoid synovium. Arthritis Rheum 1991; 34: 1076-84.

37 Shinmei M, Masuda K, Kikuchi T, Shimomura Y,Okada Y. Production of cytokines by chondrocytesand its role in proteoglycan degradation. J Rheumatol(suppl 27) 1991; 18: 89-91.

38 Hasty K A, Reife R A, Kang A H, Stuart J M. The role ofstromelysin in the cartilage destruction that accompaniesinflammatory arthritis. Arthritis Rheum 1990; 33: 388-97.

39 Abe S, Nagai Y. Evidence for the presence of a latent formof collagenase in human rheumatoid synovial fluid.JBiochem 1972; 71: 919-22.

40 Wize J. A latent collagenase from rheumatoid synovial fluid:purification and partial characterization. Biochim BiophysActa 1980; 615: 199-207.

41 Abe S, Shinmei M, Nagai Y. Synovial collagenase and jointdiseases: the significance of latent collagenase with specialreference to rheumatoid arthritis. J Biochem 1973; 73:1007-11.

42 Clark I M, Powell L K, Ramsey S, Hazleman B L,Cawston T E. The measurement of collagenase, tissueinhibitor of metaloproteinases (TIMP), and collagenase-TIMP complex. in synovial fluids from patients withosteoarthritis and rheumatoid arthritis. Arthritis Rheum1993; 36: 372-9.

43 Goldberg G I, Wilhelm S M, Kronberger A, Bauer E A,Grant G A, Eisen A Z. Human fibroblast collagenase._J Biol Chem 1986; 261: 6600-5.

975 on N

ovember 11, 2020 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.54.12.970 on 1 Decem

ber 1995. Dow

nloaded from