a complete expression profile of matrix-degrading metalloproteinases in dupuytren’s disease

DiCfefi

A Complete Expression Profile ofMatrix-Degrading Metalloproteinases in

Dupuytren’s DiseasePhillip Johnston, MB BChir, Adrian J. Chojnowski, MB BChir,

Rose K. Davidson, BSc, Graham P. Riley, PhD, Simon T. Donell, MD,Ian M. Clark, PhD

From the Institute of Orthopaedics, Norfolk and Norwich University Hospital, Norwich, UK; Schoolof Biological Sciences, University of East Anglia, Norwich, UK; and the Rheumatology Research Unit,Addenbrooke’s Hospital, Cambridge, UK.

Purpose: Dupuytren’s disease (DD) is a common fibrotic condition of the palmar fascia,leading to deposition of collagen-rich cords and finger contractions. The metzincin super-family contains key enzymes in the turnover of collagen and other extracellular matrixmacromolecules. A number of broad-spectrum matrix metalloproteinase inhibitors, used incancer clinical trials, caused side effects of DD-like contractures. We tested the hypothesisthat changes in the expression of specific metalloproteinases underlie or contribute to thefibrosis and contracture seen in DD.Methods: We collected tissue from patients with DD and used normal palmar fascia as acontrol. We profiled the expression of the entire matrix metalloproteinase (MMP), tissueinhibitor of metalloproteinases (TIMP), and a disintegrin and metalloproteinase domain withthrombospondin motif (ADAMTS) gene families in these tissues using real-time reverse-transcription polymerase chain reaction.Results: A number of metalloproteinases and inhibitors are regulated in DD. The expressionof 3 key collagenases, MMP1, MMP13, and MMP14 is increased significantly in the DDnodule, as is the expression of the collagen biosynthetic enzyme ADAMTS14. The expressionof MMP7, an enzyme with broad substrate specificity, is increased in the DD nodule andremains equally expressed in the DD cord. TIMP1 expression is increased significantly in DDnodule compared with normal palmar fascia.Conclusions: This study measured the expression of all MMP, ADAMTS, and TIMP genes inDD. Contraction and fibrosis may result from: (1) increased collagen biosynthesis mediatedby increased ADAMTS-14; (2) an increased level of TIMP-1 blocking MMP-1– and MMP-13–mediated collagenolysis; and (3) contraction enabled by MMP-14–mediated pericellularcollagenolysis (and potentially MMP-7), which may escape inhibition by TIMP-1. Thecomplete expression profile will provide a knowledge-based approach to novel therapeuticstargeting these genes. (J Hand Surg 2007;32A:343–351. Copyright © 2007 by the AmericanSociety for Surgery of the Hand.)Key words: ADAMTS, Dupuytren’s disease, gene expression, MMP, TIMP.

btpstaa

upuytren’s disease (DD) is a common con-dition with 4% to 6% of Caucasian popula-tions having evidence of disease,1 and affect-

ng up to 20% of men older than 65 years of age.2

linically it is characterized by fibrosis of the palmarascia; progression and potentially onset of the dis-ase involves the formation and proliferation of myo-
broblasts. Three distinct histologic phases have a
een described,3 with a proliferative phase leading tohe development of a nodular lesion, an involutionalhase in which cells align themselves to lines oftress, and a residual phase leaving a scar-like cordissue. Thus, the nodules are thought to represent anctive phase of disease with myofibroblast prolifer-tion, whereas the cords represent late-stage disease
nd the absence of myofibroblasts.
The Journal of Hand Surgery 343

ioglMrMcOtafa(amdANotidifTi(ftbcmlpNbfip

iitdaadsrpa

dsdtthgamiamaimmt

TpctPTttmPfusassTttettMf

Midtmlc

344 The Journal of Hand Surgery / Vol. 32A No. 3 March 2007

The matrix metalloproteinases (MMPs) are a fam-ly of 23 enzymes in human beings that include thenly mammalian enzymes able to degrade the colla-en triple helix in a specific manner under physio-ogic conditions.4 The classic collagenases are

MP-1, -8, and -13 of the human enzymes. Moreecently, MMP-2 (gelatinase A) and MMP-14 (MT1-

MP) also have been shown to make this specificleavage, although with less catalytic efficiency.5–7

ther MMPs also have been implicated in collagenurnover (eg, MMP-3, by virtue of its ability toctivate the pro-enzyme form of MMP-1).8 A relatedamily of metalloproteinases, a disintegrin and met-lloproteinase domain with thrombospondin motifADAMTSs), have 19 members in human beings andlso are implicated in extracellular matrix (ECM)etabolism. These include enzymes capable of

egrading the proteoglycan, aggrecan (at leastDAMTS-1, -4, -5, -8, -9, and -15), and 3 procollagen-propeptidases (ADAMTS-2, -3, and -14). Manyther members of this family are of unknown func-ion.9 A family of 4 specific inhibitors, the tissuenhibitors of metalloproteinases (TIMPs), have beenescribed.10 Although the ability of the 4 TIMPs tonhibit MMPs is largely promiscuous, a number ofunctional differences have been noted; for example,IMP-2, -3, and -4, but not TIMP-1, are effective

nhibitors of the membrane-type metalloproteinaseMT-MMP) subclass. Specificity among the TIMPsor inhibition of the ADAMTS family of metallopro-einases also has been described, with TIMP-3eing a potent inhibitor of ADAMTS-4 (aggre-anase-1) and ADAMTS-5 (aggrecanase-2).11 Inany fibrotic diseases such as those affecting the

iver, lung, and skin, MMPs (and related metallo-roteinases) and TIMPs play an important role.ormal ECM turnover can be viewed as a balanceetween proteinase and inhibitor activities, withbrosis coming from an imbalance away fromroteolysis.4

In the 1980s and 1990s, several small moleculenhibitors of MMPs were involved in clinical trialsn a variety of cancers.12 The major side effect ofhese drugs was a so-called musculoskeletal syn-rome, often referred to as musculoskeletal painccompanied by tendonitis.13,14 This was dose-nd time-dependent and reversible on treatmentiscontinuation, but did not respond well to non-teroidal anti-inflammatory drugs or low-dose ste-oid treatment. The clinical presentation, when re-orted in detail, is described as frozen shoulder or
condition resembling DD.15 Both of these con- c
itions involve similar fibrotic mechanisms (of thehoulder joint capsule in the case of frozen shoul-er), the laying down of a collagen-rich ECM, andhe involvement of myofibroblast-mediated con-raction.16,17 Although DD and frozen shoulderave different natural histories (the former a pro-ressive disease, the latter usually self-limitingnd resolving in time), they may well share com-on pathways leading to contracture.16 The MMP

nhibitors that cause the musculoskeletal syndromere pan-MMP inhibitors, showing an approxi-ately nanomolar (or lower) inhibition constant

gainst many of the MMPs tested. Moreover, theres good evidence that they also may inhibit relatedetalloproteinases (eg, ADAMTSs). Indeed, theusculoskeletal syndrome usually is ascribed to

he inhibition of nontarget metalloproteinases.The measurement of a small subset of MMPs and

IMPs (MMP-1, -2, -9, TIMP-1 and -2) in the sera ofatients with DD, compared with patients who hadarpal tunnel release, lends some support to the tenethat expression of these enzymes is altered in DD.18

atients with DD had significantly higher serumIMP-1 levels and the investigators18 suggested that

here is a systemic change in the MMP-to-TIMP ratiohat may lead to increased collagen deposition. Aicroarray analysis of gene expression in DD andeyronie’s disease, compared with normal palmarascia and tunica albuginea, respectively, showedpregulation of MMP2 and MMP9 in diseased tis-ue.19 A reverse-transcription polymerase chain re-ction (PCR)–based comparison of Dupuytren’s tis-ue with frozen shoulder or control shoulder tissueshowed expression of MMP1, MMP2, MMP9, andIMP1 in all tissues, with MMP3 absent from DD

issue and MMP14 increased compared with shoulderissue.20 An earlier report also showed an increasedxpression of MMP-2 and MMP-9 in Dupuytren’sissue in response to mechanical load.21 None ofhese studies were able to look at all members of the

MP, TIMP, and related metalloproteinase geneamilies.

From these data we proposed the hypothesis thatMPs, ADAMTSs, and TIMPs may play a key role

n the onset or progression of DD and related disor-ers. Hence, we used a quantitative reverse-transcrip-ion PCR methodology to profile the expression of allembers of the MMP, ADAMTS, and TIMP fami-

ies in nodule and cord tissue from DD patients andompared this with normal palmar fascia taken at
arpal tunnel release.

MTAwlitwoNordfs

RT(uG1ATrrmaetWssatpW

SCtIotu

QPOpAf

daceinApe(npw

ftwcsMBnC5c6

wtpeptb(gg(dafampCu

iwpp

Johnston et al / Metalloproteinases in Dupuytren’s Disease 345

aterials and Methodsissue Samplesll surgery was performed at the Norfolk and Nor-ich University Hospital under approval from the

ocal research ethics committee and all patients gavenformed consent. Tissue from patients with DD wasaken at fasciectomy (n � 20; age range, 42–83 y; 3omen, 20 men). Samples were divided into regionsf nodule and cord according to gross morphology.ormal palmar fascia was taken from patients with-ut DD who had carpal tunnel release (n � 20; ageange, 25–84 y; 17 women, 3 men). Tissue wasissected into approximately 5-mm pieces and snaprozen in liquid nitrogen within 15 to 30 minutes ofurgery.

NA Extractionissue was weighed and homogenized in reagent

TRIzol; Invitrogen Paisley, UK) (1 mL/0.1 g tissue)sing an homogenizer (UltraTurrax, IKA, Staufen,ermany). Particulates were pelleted at 10,000 g for0 minutes at 4°C and the supernatant was recovered.

total of 600 �L chloroform was added per 1 mLRIzol, vortexed for 15 seconds, and incubated at

oom temperature for 10 minutes. The TRIzol/chlo-oform solution was centrifuged at 10,000 g for 15inutes at 4°C and the aqueous layer was placed intofresh tube. A total of 0.5 � volume of 100%

thanol was added and mixed. Samples were appliedo spin columns (RNeasy Mini Kit; Qiagen, Crawley,

est Sussex, UK), centrifuged at full speed for 15econds, and the flow through was discarded as de-cribed by Price et al.22 Columns then were washednd eluted according to the manufacturer’s instruc-ions. RNA samples were quantified using a spectro-hotometer (NanoDrop; NanoDrop Technologies,ilmington, DE) and stored at �80°C.

ynthesis of Complementary DNAomplementary DNA was synthesized from 1 �g of

otal RNA using reverse transcriptase (Superscript II,nvitrogen) and random hexamers in a total volumef 20 �L according to the manufacturer’s instruc-ions. Complementary DNA was stored at �20°Cntil used in downstream PCR.

uantitative Real-Timeolymerase Chain Reactionligonucleotide primers and fluorescent-labeledrobes were designed (Primer Express 1.0 software;pplied Biosystems, Warrington, UK). Sequences

or MMP and TIMP primers and probes were as p

escribed by Nuttall et al23 and ADAMTS primersnd probes were as described by Porter et al.24 Toontrol against amplification of genomic DNA, prim-rs were placed within different exons close to anntron/exon boundary, with the probe spanning 2eighboring exons where possible. Basic Locallignment Search Tool (BLAST) searches for all therimer and probe sequences also were conducted tonsure gene specificity. The 18S ribosomal RNArRNA) gene was used as an endogenous control toormalize for differences in the amount of total RNAresent in each sample; 18S rRNA primers and probeere purchased from Applied Biosystems.The relative quantification of genes was per-

ormed using the ABI Prism 7700 sequence detec-ion system (Applied Biosystems) in accordanceith the manufacturer’s protocol. Polymerase

hain reactions contained 5 ng of reverse-tran-cribed RNA (1 ng for 18S analyses), 50% 2Xaster Mix (Tagman 2X Master Mix, Appliediosystems), 100 nmol/L of each primer, and 200mol/L of probe in a total volume of 25 �L.onditions for the PCR reaction were 2 minutes at0°C, 10 minutes at 95°C, then 40 cycles eachonsisting of 15 seconds at 95°C and 1 minute at0°C.The threshold cycle (CT), the cycle number at

hich the signal is detectable above the baseline, wasransformed in 2 ways. To gain an approximate com-arison across all genes measured, the amplificationfficiency was assumed to be identical across allrimer sets and normalized to 18S expression using aransformation proportional to normalized copy num-er (2��CT), where �CT is CT (target gene) - CT

18S). When comparing the expression of a singleene across sample groups, standard curves for eachene were generated using the complementary DNAcDNA) from one sample and making 2-fold serialilutions across an appropriate range. Relative inputmounts of template cDNA then were calculatedrom CT using the standard curves; data are presenteds relative levels of messenger RNA (mRNA) nor-alized to 18S rRNA. As a final quality control for the

urified RNA samples, only those cDNAs within �1.5

T of the median value for 18S for all samples weresed in the downstream study.To ascertain that the amplification product was

ndeed that of the desired target gene, productsere subcloned and sequenced. All primer androbe sets have been shown to amplify specificroducts from appropriate human tissue sam-
les.23,24

SAgcttWSttscfbd

cc

REsltsTfc

oa

Felbd

Felbd


tatistical Analysisnalyses were performed on 19 samples from eachroup because 1 sample in each group failed qualityontrol as detailed previously. Differences betweenhe 3 groups were defined using a Kruskal-Wallisest or pair-wise comparison using a 2-sided Mann-

hitney U test (in GraphPad Prism 4; GraphPadoftware Inc, San Diego, CA); these nonparametric

ests make no prior assumption on the data distribu-ion. To account for differences in age and gender,tandardized expression data using the standardurve method were subjected to a logarithmic trans-ormation to approximate normalization and testedy analysis of covariance (ANCOVA; SPSS for Win-ows; SPSS Inc, Chicago, IL). Further analysis of

igure 1. Comparative expression of MMP genes in normxpression level of each gene was determined as describeevel of 18S rRNA gene expression using �CT (CT [target goxes, nodule. The box-and-whisker plot shows the medietected.

igure 2. Comparative expression of ADAMTS genes in noxpression level of each gene was determined as describeevel of 18S rRNA gene expression using �CT (CT [target goxes, nodule. The box-and-whisker plot shows the medi
etected.
orrelation was performed using the Spearman rankorrelation (SPSS for Windows).

esultsxamining the overall pattern of gene expressionhown in Figures 1 to 3, a striking feature is that theevel of TIMP gene expression is higher than that ofhe proteinase genes. Broadly, the level of steady-tate mRNA across the 3 gene families assayed isIMP � ADAMTS � MMP. Table 1 shows the dif-erences in gene expression between the DD nodule andord tissue compared with normal palmar fascia.

Although DD obviously involves the laying downf excess collagenous matrix, this matrix is presum-bly also remodeled and shortened during disease

lmar fascia versus cord and nodule tissue from DD. Thehe Materials and Methods section and normalized to theCT [18S]). Open boxes, normal; filled boxes, cord; dottedd each section represents a quartile of the data. N.D., not

almar fascia versus cord and nodule tissue from DD. Thehe Materials and Methods section and normalized to theCT [18S]). Open boxes, normal; filled boxes, cord; dottedd each section represents a quartile of the data. N.D., not

al pad in t

ene] -an, an

rmal pd in t

ene] -an, an

pcsAicsbange

tlipiaaac

Dtca

g

AApApDrli

tai

roMaaosfhcAgsiFi(tmTfFw

DImarouiacsi

Fpetlgdm


rogression either as a consequence of tension orontributing to increased tension. Thus, the expres-ion of 3 procollagen N-propeptidases, ADAMTS2,DAMTS3, and particularly ADAMTS14, is increased

n the DD nodule, consistent with an increase inollagen synthesis. Conversely, the expression ofeveral enzymes capable of mediating collagenreakdown, MMP1, MMP2, MMP13, and MMP14,lso is increased in the DD nodule compared withormal palmar fascia, consistent with increased colla-en turnover. The increase of all of the earlier-describednzymes is to some extent attenuated in the DD cord.

Comparing the DD nodule with the cord, it is clearhat many genes that are induced in the nodule are ateast less induced in the cord, with many not signif-cantly different in the cord compared with normalalmar fascia. The only gene whose expression isncreased in the DD nodule and remains increased tosimilar level in the DD cord is MMP7; MMP-7 hasbroad substrate specificity,25 and this may representcontinued matrix remodeling process in the DD

ord (see later).The most significantly downregulated gene in the

D nodule is MMP3. This appears to be a recurringheme in diseases involving matrix destruction be-ause similar findings have been reported in osteo-rthritic cartilage26 and degenerative tendinopathy.27

Six ADAMTS enzymes have been shown to be ag-

igure 3. Comparative expression of TIMP genes in normalalmar fascia versus cord and nodule tissue from DD. Thexpression level of each gene was determined as described inhe Materials and Methods section and normalized to theevel of 18S rRNA gene expression using �CT (CT [targetene] - CT [18S]). Open boxes, normal; filled boxes, cord;otted boxes, nodule. The box-and-whisker plot shows theedian, and each section represents a quartile of the data.

recanases, at least in vitro, ADAMTS-1, ADAMTS-4, t

DAMTS-5, ADAMTS-8, ADAMTS-9, andDAMTS-15. ADAMTS15 is expressed in normalalmar fascia, but not regulated in DD tissue.DAMTS4, ADAMTS5, and ADAMTS9 all are ex-ressed in normal palmar fascia and increased in theD nodule, but not the cord. ADAMTS1 is expressed

obustly in normal palmar fascia and expression isower in the DD nodule. ADAMTS8 is undetectablen most tissue samples.

Of the TIMPs, TIMP1 expression is increased inhe DD nodule, but TIMP2, TIMP3, and TIMP4 allre expressed at lower levels in the DD nodule thann normal palmar fascia.

Analysis of covariance showed that there is aelationship between the patient’s age and the levelf gene expression for MMP3, MMP16, MMP17,MP25, TIMP1, TIMP2, ADAMTS1, ADAMTS5,

nd ADAMTS18. Analysis of covariance also showedweak relationship between gender and expression

f MMP16, TIMP2, ADAMTS3, and ADAMTS5. Thetatistically significant differences in gene expressionor these genes between groups (detailed previously),owever, are paralleled in this analysis in allases apart from ADAMTS18. The expression ofDAMTS18 shows no statistical difference betweenroups under ANCOVA analysis but shows weakignificance when analyzed by either Kruskal-Wallis orn pair-wise analysis using the Mann-Whitney U test.or the expression of MMP3, TIMP1, and TIMP2, there

s a strong correlation with age in normal palmar fasciathe former 2 decreasing with age, the latter increasing)hat is lost in DD tissues. This shows that regulatory

echanisms for these genes are aberrant in disease.able 2 shows a Spearman rank correlation analysis

or the 9 genes showing an effect of age in ANCOVA.igure 4 shows the correlation of MMP3 expressionith age in the normal palmar fascia.

iscussionn screens of gene expression the need to account forultiple testing in the statistical analyses is problem-

tic. The Bonferroni correction is often conservative,emoving type I error (false positive) at the expensef type II error (false negative). This would limit thetility of gene expression studies in which the valid-ty of any multiple testing procedure has yet to bescertained.28,29 In Table 1, we have not applied anyorrection for multiple testing, although at the 5%ignificance level (p � .05) a false-positive rate of 1n 20 would be expected.

Of prime importance to the phenotypic outcome of
hese changes in gene expression is the final balance

opmo

cma


f proteinase versus inhibitor. Although this is im-ossible to ascertain from a study at the steady-stateRNA level such as this, it is clear that TIMP1 is the

Table 1. Comparison of Gene Expression BetweenFrom DD

GeneNormal versus nodule(�, higher in nodule)

MMP1 � .0009MMP2 � �.0001MMP3 - �.0001MMP7 � �.0001MMP8 NSMMP9 NSMMP10 NSMMP11 � �.0001MMP12 NSMMP13 � �.0001MMP14 � �.0001MMP15 � .0144MMP16 � �.0001MMP17 � �.0001MMP19 � �.0001MMP20 NDMMP21 � .0071MMP23 NSMMP24 NSMMP25 NSMMP26 NDMMP27 - .0003MMP28 NSADAMTS1 - .011ADAMTS2 � .0012ADAMTS3 � .0034ADAMTS4 � .0024ADAMTS5 � .0066ADAMTS6 - .0043ADAMTS7 NSADAMTS8 - .0056ADAMTS9 � .0037ADAMTS10 NSADAMTS12 � �.0001ADAMTS13 NSADAMTS14 � �.0001ADAMTS15 NSADAMTS16 � .0223ADAMTS17 NSADAMTS18 � .036ADAMTS19 - .0056ADAMTS20 - .0034TIMP1 � .0008TIMP2 - .0115TIMP3 - .004TIMP4 - .004

Statistical analysis was performed pair-wise using the Mann-WhitnNS, not significant; ND, not detected.

nly TIMP whose expression is increased signifi- n

antly in the DD nodule compared with normal pal-ar fascia. Indeed, TIMP2, TIMP3, and TIMP4 all

re reduced in the DD nodule. Thus, in the DD

al Palmar Fascia, Cord, and Nodule Tissue

� p value

Normal versus cord(�, higher in cord)

Nodule versus cord(�, higher in cord)

NS NS.0043 - .0043.0002 � .0091

�.0001 NSNS � .044NS NSNS NS

�.0001 - .004NS NS.0003 - .0012NS – .0061NS - .0098.036 – .004NS - �.0001NS - .0028ND NDNS NSNS NSNS NSNS NSND NDNS � .002NS - .04NS � .0043NS - .012.018 NSNS - .0005NS - .017NS � .04NS NSNS � .0043NS NSNS NS.0028 - .002NS NSNS - .0002NS � .036NS - .0024NS NS.0144 NSNS � .021NS � .0031NS - .028NS � .0005NS NSNS � .03

est.

Norm

�-�

�

�

�

�

�

�

ey U t

odule, the situation may be similar to that described

idipiMMtTcezSeCtTt

ebdtbtnIitwwdt

optchofdmaDgepe

toptts

eTedwt

RD

FmmM1gsr


n liver fibrosis, in which the potential for matrixegradation is present, even in advanced disease, buts prevented by TIMP expression.30 Of potential im-ortance here is the fact that TIMP-1 is a poornhibitor of MMP-14, MMP-15, MMP-16, and

MP-24.10 Because expression of MMP14 andMP16 is significantly increased in the DD nodule,

hese proteinases might escape inhibition fromIMP-1. This also may provide a mechanism forollagen remodeling, particularly in the pericellularnvironment (by MMP-14, a membrane-bound en-yme) during contraction in the presence of TIMP-1.imilarly, MMP-7 (whose expression is increasedqually in both the DD nodule and the cord), lacks a-terminal hemopexin-like domain that is common

o most MMPs, decreasing its affinity for at leastIMP-1,31 and potentially resisting inhibition in pa-

hology.Several studies have compared the contractile prop-

rties of Dupuytren’s cord- or nodule-derived fibro-lasts with control cultures.32–35 Results variedepending on the model of contraction used, buthe consensus appears to be that Dupuytren’s fi-roblasts can generate significantly increased con-ractile force compared with control cells, withodule-derived cells showing greatest contraction.n such models, broad-spectrum synthetic MMPnhibitors have been shown to inhibit contrac-ion,36,37 and this is paralleled by the inhibition ofound contraction in a mouse model of skinound healing.38 This latter study also reported aecrease in �-smooth muscle actin expression inhe presence of the MMP inhibitor. This blockade

Table 2. Correlation of Gene Expression WithAge in Normal Palmar Fascia, Cord, or NoduleTissue From DD

Gene

p value (r value)

Cord Nodule Normal

MMP3 .318 .273 �.001 (�.774)MMP16 .204 .307 .063 (�.435)MMP17 .322 .024 (�.516) .616MMP25 .323 .358 .115TIMP1 .632 .557 .012 (�.565)TIMP2 .280 .416 .013 (.558)ADAMTS1 .051 (.454) .263 .473ADAMTS5 .627 .011 (�.567) .743ADAMTS18 .022 (�.523) .468 .348

Genes shown are those identified by ANCOVA. Statistical anal-ysis was performed using the Spearman rank correlation. TheSpearman r value is shown where correlation is significant (pos-itive value indicates an increase in expression with age, negative
value indicates a decrease in expression with age).
f contraction is difficult to dovetail with the re-orted side effects of some of these MMP inhibi-ors in clinical trials in which Duputyren’s-likeontracture and frozen shoulder were observed13–15;owever, this may reflect the fact that culture modelsf collagen lattice contraction only represent a singleacet of a more complex disease process. Again,rawing on the analogy with liver fibrosis, a treat-ent that downregulates TIMPs but increases the

ctivity of MMPs may be an appropriate therapy forD.30 One such possibility is relaxin, an insulin-likerowth factor hormone reported to decrease collagenxpression, increase MMP, and decrease TIMP ex-ression in a variety of in vitro and in vivo mod-ls.39–42

It also may be possible to target collagen biosyn-hesis via blockade of ADAMTS-14. Functionalverlap between the 3 ADAMTS procollagen-N-ropeptidases (ADAMTS-2, -3, and -14) provideshe potential to inhibit ADAMTS-14 without disrup-ion of global collagen synthesis,43 but this may beufficient to slow or halt collagen deposition in DD.

The current study examines the expression of thentire MMP, ADAMTS, and TIMP families in DD.hese enzymes and inhibitors have critical roles inxtracellular matrix homeostasis, which is clearlyysregulated in DD. These data will allow futureork to focus on the function of specific metallopro-

einases in the disease process.

eceived for publication October 25, 2006; accepted in revised formecember 15, 2006.

20 30 40 50 60 70 80 900.0

0.5

1.0

1.5

2.0

2.5

Age (years)

Sta

nd

ard

ized

CT

igure 4. Correlation of MMP3 expression with age in nor-al palmar fascia. The expression level of MMP3 was deter-ined using the standard curve method as described in theaterials and Methods section and normalized to the level of

8S rRNA gene expression using an adjusted �CT (CT [targetene] - CT [18S], allowing for efficiency of primer-probeets). The solid line shows the best fit as determined by linearegression, with dotted lines defining 95% confidence limits.

The authors received a research grant from Action Arthritis.

ee

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3


Corresponding author: Ian M. Clark, PhD, School of Biological Sci-nces, University of East Anglia, Norwich, NR4 7TJ, United Kingdom;-mail: [email protected].

Copyright © 2007 by the American Society for Surgery of the Hand0363-5023/07/32A03-0007$32.00/0doi:10.1016/j.jhsa.2006.12.010

eferences1. Thurston AJ. Dupuytren’s disease. J Bone Joint Surg 2003;

85B:469–477.2. Mikkelsen OA. Epidemiology of a Norwegian population.

In: McFarlane RM, McGrouther DA, Flint MH, eds. Du-puytren’s disease: biology and treatment. New York:Churchill Livingstone, 1990:191–200.

3. Luck JV. Dupuytren’s contracture; a new concept of thepathogenesis correlated with surgical management. J BoneJoint Surg 1959;41A:635–664.

4. Brinckerhoff CE, Matrisian LM. Matrix metalloproteinases:a tail of a frog that became a prince. Nat Rev Mol Cell Biol2002;3:207–214.

5. Knauper V, Lopez-Otin C, Smith B, Knight G, Murphy G.Biochemical characterization of human collagenase-3. J BiolChem 1996;271:1544–1550.

6. Ohuchi E, Imai K, Fujii Y, Sato H, Seiki M, Okada Y.Membrane type 1 matrix metalloproteinase digests intersti-tial collagens and other extracellular matrix macromole-cules. J Biol Chem 1997;272:2446–2451.

7. d’Ortho MP, Will H, Atkinson S, Butler G, Messent A,Gavrilovic J, et al. Membrane-type matrix metalloprotein-ases 1 and 2 exhibit broad-spectrum proteolytic capacitiescomparable to many matrix metalloproteinases. Eur J Bio-chem 1997;250:751–757.

8. Murphy G, Cockett MI, Stephens PE, Smith BJ, DochertyAJ. Stromelysin is an activator of procollagenase. A studywith natural and recombinant enzymes. Biochem J 1987;248:265–268.

9. Porter S, Clark IM, Kevorkian L, Edwards DR. The AD-AMTS metalloproteinases. Biochem J 2005;386:15–27.

0. Baker AH, Edwards DR, Murphy G. Metalloproteinaseinhibitors: biological actions and therapeutic opportunities.J Cell Sci 2002;115:3719–3727.

1. Kashiwagi M, Tortorella M, Nagase H, Brew K. TIMP-3 isa potent inhibitor of aggrecanase 1 (ADAM-TS4) and ag-grecanase 2 (ADAM-TS5). J Biol Chem 2001;276:12501–12504.

2. Brown PD. Ongoing trials with matrix metalloproteinaseinhibitors. Expert Opin Investig Drugs 2000;9:2167–2177.

3. Bird J, Montana JG, Wills RE, Baxter AD, Owen DA.Selective matrix metalloproteinase (MMP) inhibitors havingreduced side-effects. Chemical Abstracts 1999;129:225751.

4. Drummond AH, Beckett P, Brown PD, Bone EA, DavidsonAH, Galloway WA, et al. Preclinical and clinical studies ofMMP inhibitors in cancer. Ann N Y Acad Sci 1999;878:228–235.

5. Hutchinson JW, Tierney GM, Parsons SL, Davis TR. Du-puytren’s disease and frozen shoulder induced by treatmentwith a matrix metalloproteinase inhibitor. J Bone Joint Surg1998;80B:907–908.

6. Smith SP, Devaraj VS, Bunker TD. The association betweenfrozen shoulder and Dupuytren’s disease. J Shoulder Elbow
Surg 2001;10:149–151. 3
7. Schaer H. Die atiologie der periarthritis humeroscapularis.Ergebn Chir Orthop 1936;29:11.

8. Ulrich D, Hrynyschyn K, Pallua N. Matrix metalloprotein-ases and tissue inhibitors of metalloproteinases in sera andtissue of patients with Dupuytren’s disease. Plast ReconstrSurg 2003;112:1279–1286.

9. Qian A, Meals RA, Rajfer J, Gonzalez-Cadavid NF. Com-parison of gene expression profiles between Peyronie’s dis-ease and Dupuytren’s contracture. Urology 2004;64:399–404.

0. Bunker TD, Reilly J, Baird KS, Hamblen DL. Expression ofgrowth factors, cytokines and matrix metalloproteinases infrozen shoulder. J Bone Joint Surg 2000;82B:768–773.

1. Tarlton JF, Meagher P, Brown RA, McGrouther DA, BaileyAJ, Afoke A. Mechanical stress in vitro induces increasedexpression of MMPs 2 and 9 in excised Dupuytren’s diseasetissue. J Hand Surg 1998;23B:297–302.

2. Price JS, Waters JG, Darrah C, Pennington C, Edwards DR,Donell ST, Clark IM. The role of chondrocyte senescence inosteoarthritis. Aging Cell 2002;1:57–65.

3. Nuttall RK, Pennington CJ, Taplin J, Wheal A, Yong VW,Forsyth PA, Edwards DR. Elevated membrane-type matrixmetalloproteinases in gliomas revealed by profiling pro-teases and inhibitors in human cancer cells. Mol Cancer Res2003;1:333–345.

4. Porter S, Scott SD, Sassoon EM, Williams MR, Jones JL,Girling AC, et al. Dysregulated expression of adamalysin-thrombospondin genes in human breast carcinoma. ClinCancer Res 2004;10:2429–2440.

5. Clark IM, Parker AE. Metalloproteinases: their role in ar-thritis and potential as therapeutic targets. Expert Opin TherTargets 2003;7:19–34.

6. Kevorkian L, Young DA, Darrah C, Donell ST, ShepstoneL, Porter S, et al. Expression profiling of metalloproteinasesand their inhibitors in cartilage. Arthritis Rheum 2004;50:131–141.

7. Jones GC, Corps AN, Pennington CJ, Clark IM, EdwardsDR, Bradley MM, et al. Expression profiling of metallopro-teinases and tissue inhibitors of metalloproteinases in normaland degenerate human Achilles tendon. Arthritis Rheum2006;54:832–842.

8. Jung SH, Bang H, Young S. Sample size calculation formultiple testing in microarray data analysis. Biostatistics2005;6:157–169.

9. Bretz F, Landgrebe J, Brunner E. Multiplicity issues in microar-ray experiments. Methods Inf Med 2005;44:431–437.

0. Iredale JP. Cirrhosis: new research provides a basis forrational and targeted treatments. BMJ 2003;327:143–147.

1. Baragi VM, Fliszar CJ, Conroy MC, Ye QZ, Shipley JM,Welgus HG. Contribution of the C-terminal domain of met-alloproteinases to binding by tissue inhibitor of metallopro-teinases. C-terminal truncated stromelysin and matrilysinexhibit equally compromised binding affinities as comparedto full-length stromelysin. J Biol Chem 1994;269:12692–12697.

2. Bisson MA, Mudera V, McGrouther DA, Grobbelaar AO.The contractile properties and responses to tensional loadingof Dupuytren’s disease—derived fibroblasts are altered: acause of the contracture? Plast Reconstr Surg 2004;113:611–624.

3. Moyer KE, Banducci DR, Graham WP III, Ehrlich HP.

3

3

3

3

3

3

4

4

4

4


Dupuytren’s disease: physiologic changes in nodule andcord fibroblasts through aging in vitro. Plast Reconstr Surg2002;110:187–196.

4. Tarpila E, Ghassemifar MR, Wingren S, Agren M, FranzenL. Contraction of collagen lattices by cells from Dupuytren’snodules. J Hand Surg 1996;21B:801–805.

5. Rayan GM, Tomasek JJ. Generation of contractile force bycultured Dupuytren’s disease and normal palmar fibroblasts.Tissue Cell 1994;26:747–756.

6. Daniels JT, Cambrey AD, Occleston NL, Garrett Q, Tar-nuzzer RW, Schultz GS, Khaw PT. Matrix metallopro-teinase inhibition modulates fibroblast-mediated matrixcontraction and collagen production in vitro. Invest Oph-thalmol Vis Sci 2003;44:1104 –1110.

7. Scott KA, Wood EJ, Karran EH. A matrix metalloproteinaseinhibitor which prevents fibroblast-mediated collagen latticecontraction. FEBS Lett 1998;441:137–140.

8. Mirastschijski U, Haaksma CJ, Tomasek JJ, Agren MS.Matrix metalloproteinase inhibitor GM 6001 attenuateskeratinocyte migration, contraction and myofibroblastformation in skin wounds. Exp Cell Res 2004;299:465–
475.
9. Unemori EN, Beck LS, Lee WP, Xu Y, Siegel M, Keller G,et al. Human relaxin decreases collagen accumulation invivo in two rodent models of fibrosis. J Invest Dermatol1993;101:280–285.

0. Unemori EN, Amento EP. Relaxin modulates synthesis andsecretion of procollagenase and collagen by human dermalfibroblasts. J Biol Chem 1990;265:10681–10685.

1. Unemori EN, Pickford LB, Salles AL, Piercy CE, GroveBH, Erikson ME, Amento EP. Relaxin induces an extracel-lular matrix-degrading phenotype in human lung fibroblastsin vitro and inhibits lung fibrosis in a murine model in vivo.J Clin Invest 1996;98:2739–2745.

2. Williams EJ, Benyon RC, Trim N, Hadwin R, Grove BH,Arthur MJ, et al. Relaxin inhibits effective collagen deposi-tion by cultured hepatic stellate cells and decreases rat liverfibrosis in vivo. Gut 2001;49:577–583.

3. Le Goff C, Somerville RP, Kesteloot F, Powell K, Birk DE,Colige AC, Apte SS. Regulation of procollagen amino-propeptide processing during mouse embryogenesis by spe-cialization of homologous ADAMTS proteases: insights oncollagen biosynthesis and dermatosparaxis. Development
2006;133:1587–1596.

a complete expression profile of matrix-degrading metalloproteinases in dupuytren’s disease

Documents