Jetal.BMCMusculoskeletalDisorders2012,13:53 http://www.biomedcentral.com/1471-2474/13/53 RESEARCH ARTICLE Open Access Local biochemical and morphological differences in human Achilles tendinopathy: a case control study Pingel J1*, Fredberg U2, Qvortrup K3, Larsen JO4, Schjerling P1, Heinemeier K1, Kjaer M1 and Langberg H5 Abstract Background: The incidence of Achilles tendinopathy is high and underlying etiology as well as biochemical and morphological pathology associated with the disease is largely unknown. The aim of the present study was to describe biochemical and morphological differences in chronic Achilles tendinopathy. The expressions of growth factors, inflammatory mediators and tendon morphology were determined in both chronically diseased and healthy tendon parts. Methods: Thirty Achilles tendinopathy patients were randomized to an expression-study (n = 16) or a structural- study (n = 14). Biopsies from two areas in the Achilles tendon were taken and structural parameters: fibril density, fibril size, volume fraction of cells and the nucleus/cytoplasm ratio of cells were determined. Further gene expressions of various genes were analyzed. Results: Significantlysmallercollagenfibrilsandahighervolumefractionofcellswereobservedinthe tendinopathicregionofthetendon.Markersforcollagenanditssynthesiscollagen1,collagen3,fibronectin, tenascin-c,transforminggrowthfactor-bfibromodulin,andmarkersofcollagenbreakdownmatrixmetalloproteinase- 2,matrixmetalloproteinase-9andmetallopeptidaseinhibitor-2weresignificantlyincreasedinthetendinopathic region.Noalteredexpressionsofmarkersforfibrillogenesis,inflammationorwoundhealingwereobserved. Conclusion: The present study indicates that an increased expression of factors stimulating the turnover of connective tissue is present in the diseased part of tendinopathic tendons, associated with an increased number of cells in the injured area as well as an increased number of smaller and thinner fibrils in the diseased tendon region. As no fibrillogenesis, inflammation or wound healing could be detected, the present data supports the notion that tendinopathy is an ongoing degenerative process. Trial registration: Current Controlled Trials ISRCTN20896880 Keywords: Collagen, Gene expression, Patients, Growth factors, Tissue turnover Background adaptation to loading can ultimately leadto increases in Tendons connect muscle to bone and enable transmis- CSAandcollagencontent inchronicallyloaded tendons sionofforcesfromcontractingmuscletobone,resulting [5].Despitethisphysiologicalabilitytoadapt,tendinopa- in joint movement. They possess the ability to adapt to tic tendons represents a large and constantly growing changes in loading [1] and studies have shown that col- clinical problem affecting both recreational and profes- lagensynthesisisincreasedasaresultofbothacuteexer- sional athletes as well as people involved in repetitive cise [2,3] and prolonged physical training [4]. The labour [6,7]. Years of research have unfortunately not provided much insight into the pathogenesis of chronic tendinopathy [8]. Indeed, the etiology of tendinopathy *Correspondence:[email protected] 1InstituteofSportsMedicine,DepartmentofOrthopedicSurgeryM. hasbeenrelated to repeated micro strainbelow the fail- BispebjergHospitalandCenterforHealthyAging,FacultyofHealthSciences, ure threshold as an initiating stimulus for degenerative UniversityofCopenhagen,Copenhagen,Denmark processes[9,10].Other authors,however,haveproposed Fulllistofauthorinformationisavailableattheendofthearticle ©2012Jetal;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommons AttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,andreproductionin anymedium,providedtheoriginalworkisproperlycited. Jetal.BMCMusculoskeletalDisorders2012,13:53 Page2of13 http://www.biomedcentral.com/1471-2474/13/53 that mechano-biological under-stimulation results in a factor (bFGF) are decreased causing a reduced healing degenerative cascade, through the production of a pat- capacityintheinjuredareaoftheAchillestendons. ternofcatabolicgeneexpressionthatleadsultimatelyto extracellular matrix degeneration [11]. Tendinopathy is Methods characterized by activity-related pain, focal tendon ten- Design derness, and decreased local movement in the affected Thirty patients with chronic Achilles tendon pain were area[12,13].Thegeneralopinionisthatnoinflammatory includedinthisstudyapprovedbythelocalEthicalCom- cellsarepresentinthetendinopathictissue[14]andthat mitteeoftheCapitalRegionCopenhagen(H-1-2009-114) tendinopathyistheresultofa degenerativeprocesswith and in compliance with the Helsinki Declaration. Addi- collagen disorganization, collagen fibre separation, tionally the study was registered at Current Controlled increasedcellularity,neovascularizationandfocal necro- Trials (ISRCTN20896880). Due to limitations in the sis[15]. amount of tissue gained from the tendon biopsies Previousstudieshaveshownanalteredconcentrationof patients were randomly assigned to either a Structural certain matrix metalloproteinases MMPs, AdAMt’s and study (n = 14) or a Biochemical study (n = 16) by the TIMP’sinnormalanddegeneratehumanAchillestendon envelope method. Allsubjectswererecreationalathletes [16].Additionallyseveralcytokines[9,10]canbefoundin or workers with a long-term history of chronic Achilles tendonsandfibroblastsaftercyclicmechanicalstretching tendonpain(>1/2year)(Table4)andconventionalcon- in healthy tendon tissue. However, the published data servative treatments (eccentric rehabilitation, NSAIDs arises from the comparison of tendinopathic tissue with andcorticosteroidinjections)hadbeentriedinallindivi- either control tissue from different anatomical tendons duals with no effect. Intake of NSAID or corticosteroid [17]orwithtissuesfromidenticalanatomicaltendonsbut injectionwasnotallowed6monthspriortoinclusionin from different subjects [18]. Since considerable micro- the present study. All subjects were recruited from the scopic structure differences have been demonstrated in Rheumatology Department, Silkeborg Hospital, Den- anatomicallydifferenttendons[19],thislimitstheconclu- mark, and the biopsies from the Achilles tendons were sions that may be drawn from these studies. Taking the takenaspartofastandardprocedureinordertoexamine aforementionedlimitationsintoaccount,currentdatacon- for deposits ofcholesterol, uric acid, and amyloid in the cerninglocalbiochemicaldifferenceswithintendinopathic injuredAchillestendons. tendons,seemtoindicatethatanalteredexpressionofcol- lagen [20], proteoglycans [21] and matrix metalloprotei- Biopsy procedure nases[16,22]existsintendinopathic tendons.Inaddition The subjects were locally anesthetized, in the peritendi- thelevelofcytokines[23]VEGFandfibronectin[24]has nous space from both the medial and lateral side of the been shown to be significantly different in the tendino- tendon with injections of 2 × 10 ml 1% Lidocain, using pathicarea.Howeveranalysesoflocalbiochemicaldiffer- ultrasound guidance. Biopsies were taken with a semi- encestogetherwithmorphologicaldifferencesarelacking. automatic biopsy needle (14 GA, 9 cm; Angiotech) also The aim of the present study was to elucidate if any usingultrasound(US)guidance. Aninitialtendonbiopsy local structural differences are present in tendinopathic was taken inthe maximally sickarea evaluatedusing US areas of human Achilles tendons compared to healthy (definedastheareawithmaximalincreasedtendonthick- areas in the same tendon. Furthermore, we wanted to ness,neovascularisation,hypoeccogenicity).Thisareawas investigate which proteoglycans, growth factors and usually3-5cmabovetheattachment of the Achillesten- cytokines that were involved in the local structural dif- don to the calcanaeus bone. A second biopsy was taken ferences observed. fromthesametendon4cmproximaltothefirstbiopsyin Wehypothesizethatseveralmarkerssuchascollagen3 a region of the tendon tissue that was deemed normal wouldbelocallyupregulatedindicatingformationofscar usingUS. tissue withinthe tendon[25] and higher concentrations BiopsysamplesintendedforanalysisusingTransmission ofMMP-2 andMMP-9indicating an enhanced degrada- ElectronMicroscopywereimmersedin2%glutaraldehyde tion of collagen structures in the tendinopathic area (t- in 0.05 M sodium buffer (pH 7.2), and the samples for area)whencomparedwiththehealthyarea(h-area)ofthe gene expression were snap-frozen and stored at -80°C sametendon.Furthermoreitishypothesizedthatcertain untilanalysis. proteoglycans would have altered expression in the two tendon regions, e.g. an increased expression of decorin Transmission electron microscopy of tendon biopsies which might cause the collagen turnovertobe increased Fourteen tendon biopsy pairs were cut into small pieces also in chronic tendinopathic tendons. Additionally we and were immersion-fixed in 2% glutaraldehyde for hypothesize that growth factors like fibroblast growth 24 hours. Following three rinses in 0.15 M sodium Jetal.BMCMusculoskeletalDisorders2012,13:53 Page3of13 http://www.biomedcentral.com/1471-2474/13/53 phosphate buffer(pH 7.2) the specimens werepost-fixed ofthetendontissue.Theestimatedparameterswere:the in 1% OsO in 0.12 M sodium cacodylic buffer for volume fraction of cells within the tendon, the volume 4 2hours.Thespecimensweredehydratedinagradedseries fraction of the nucleus within the cell, and the volume ofethanol(70%,96%and100%),transferredtopropylene fraction of cytoplasm within the cell. A single experi- oxide and embedded in Epon (VWR Bie&Berntsen) in encedinvestigatorperformedallstereologicalanalysesin three steps according to standard procedures. For each a blinded fashion. The investigator was blinded for all biopsy one ultra thinsectionwas cutapproximately per- subject characteristics, and whether the sample was pendiculartothelengthaxisof thetendon witha Reich- obtainedfromthetendinopathicorthehealthyregionof ert-JungultracutEmicrotome.Thesectionwascollected thetendon. on a one-hole copper grid with a Formvar supporting membrane and stained with uranyl acetate and lead RNA extraction and real time-PCR analysis citrate. The sections were examined using a Phillips CM Total RNAisolation: Total RNA wasextracted fromfro- 100 transmission electron microscope operated at an zen tendon samples from 16 subjects (sample weight: acceleratingvoltageof80kV.Digitalimageswereobtained mean23.2±6.4mg)byusing1mlofTRIReagent(Mole- withaMegaViewIIcameraandananalysissoftwarepack- cularResearchCentre,Cincinnati,OH)5steelbeads(2.3 age. From each ultra thin section the intercellular tissue mm) and 4 silica beads (1.0 mm Silicon Carbide Beads was examined by taking a simple, random sample of ten (454grams)BioSpecProductsInc.).Glycogenwasadded digitizedTEMimagesoftheintercellulartissue.Thecellu- (120 μg per ml of TriReagent) to the tendon samples to lar component of the tendon was examined in eleven improveRNAprecipitation. biopsypairsbytaking6times6imagesinthreerandomly ExtractedRNAwasprecipitatedfromtheaqueousphase positioned regions of the section. The 6 times 6 images with isopropanol and was washed with ethanol [75%], were spliced into one image using multiple image align- driedandsuspendedin10μlofnuclease-free water.The ment(MIA)tools, so foreachexamined biopsyatotalof RNA concentration was determined using a RiboGreen threeMIAimageswereobtained. RNAQuantitationkit200-2000Assays,MolecularProbes USA.RNAqualitywasdeterminedonthebasisofaRNA Stereology 6000nanoChipassaykit,AgilentTechnologies,Germany. TheStereologicalanalysesoftheimageswerecarriedout TheRNAsampleswerestoredfrozenat-20°Cuntilsubse- on a computer monitor onto which the digitized EM quentuseinreal-timeRT-PCRprocedures. image was merged with a graphic representation of the cDNA synthesis: 100 ng RNA was reverse transcribed stereological test systems for just 12 of the 14 biopsy foreachtendonsampleinatotalvolumeof20μlbyusing pairs (2 biopsies was unfortunately not useable for theQiagenOmniscriptRTKitat37°Cfor1hourfollowed stereologyanalyses)(C.A.S.T.-gridsoftware,TheInterna- by70°C for15minutes.Theresulting cDNA wasdiluted tional Stereology Center at Olympus). The intercellular twentytimesindilutionbuffer(10mMTrisEDTAbuffer: tissuewasanalyzedatafinalmagnificationof210.000in Sigma Germany) + Salmon Testes DNA (1 ng/μl; Sigma the ten ordinary TEMimages. Thevolume fraction (Vv) Germany),andsampleswerestoredat-20°Cuntilusedin ofcollagenfibrilsperintercellulartissuevolumewasesti- thePCRreactionsforspecificmRNAanalysis. matedwiththe pointcounting technique asthe number PolymeraseChainReaction:TheReal-timePCR-method ofpointshittingcollagenfibresdividedbythenumberof using Glyceraldehyde 3-phosphate dehydrogenase pointshittingtheintercellular tissue (includingcollagen (GAPDH)and60SacidicribosomalproteinP0(RPLP0)as fibrils) using a point grid of 36 points. The number of reference genestostudyspecific mRNA’sofinterestwas collagenfibrilspercrosssectionalareaofintercellulartis- applied.TheprimerswerepurchasedfromMWGBiotech. sue (NA) was counted in 16 uniformly positioned, ForeachtargetcDNAthePCRreactionswerecarriedout unbiased counting frames, each with an area of 0.0426 underidenticalconditionsbyusing5μldilutedcDNAina mm2 (42.6μm2), and the individual diameters (d) ofthe total volume of 25 μlQuantiTect SYBR Green PCR Mix sampledcollagenfibrilsweremeasuredasthelargestdia- (Qiagen)and100nMofeachprimer(Table2).Theampli- meterperpendiculartothelongestaxis(i.e.thelengthof fication was monitored in real-time using a MX3005P theminoraxisoftheellipse) usingthe“measure-length” real-time PCR machine (Stratagene, CA). The threshold featureoftheCAST-gridsystem.Theunbiasedcounting cycle (C) values were related to a standard curve made t frameensuresthatallprofiles,regardlessofshape,sizeor withclonedPCRproductstodeterminetherelativediffer- orientation, have an equal probability of being sampled ence between the unknown samples, accounting for the within area probe. The MIA images were analyzed at a PCR efficiency. The specificity of the PCR reaction was finalmagnificationof115,000.The pointcountingtech- confirmed by melting curve analysis after amplification. nique,usingapointgridwithapproximately1000points, Thereal-timePCRconditionswereasfollows:todenatu- estimatedthevolumefractionsofthecellularcomponent ratetheDNAstrandsthereactionmixwasheated above Jetal.BMCMusculoskeletalDisorders2012,13:53 Page4of13 http://www.biomedcentral.com/1471-2474/13/53 the melting temperature of DNA (95°C) for 10 minutes, ofthefibrils,asignificantlyhighernumberoffibrilswitha followedby50cycleseachof15secondsat95°C,followed diameterinthelowerrange(10-40nm)wasfoundinthe bytheannealingstepwhereoptimalprimerhybridization tendinopathicareacomparedtothehealthyarea(diameter conditionswereobtainedbyloweringthetemperatureto 10-20nm:Tendinopathicarea:32±7fibrils/μm2;Healthy 58°C for 30 seconds, and the extension step, where the area 13 ± 4 fibrils/μm2; diameter: 20-30 nm: Tendino- reaction mix was heated to 63°C for 90 seconds. Two pathic area: 68 ± 10 fibrils/μm2; Healthy area: 26 ± housekeeping genes GAPDH and RPLP0 were used as 4 fibrils/μm2; diameter 30-40 nm: Tendinopathic area: reference genes. The RPLP0gene hadbeenchosenasan 74± 10 fibrils/μm2; Healthy area: 42 ± 5 fibrils/μm2; see internal control, assuming RPLP0 to be constitutively Figure1).Inadditionasignificantlyhighervolumefraction expressed. To validate this assumption GAPDH mRNA ofcellswasobservedinthetendinopathicareaoftheten- wasmeasuredasanotherunrelated“constitutive”andnor- don (Figure 2). The volume fraction of the cytoplasm malized with RPLP0, showing no difference between withinthecellwasfoundto beidenticalinthetwoareas the healthy and the tendinopathic region of the tendon (Figure 2) implying an increased number of cells in the (Figure3). tendinopathicarea. Statistical analysis Gene expression analysis ThePCRdatawerelogtransformedandaPairedStudents The gene expression of 24 genes was analyzed in the t-test was performed to compare the results from the biopsies of 16 tendinopathy patients (Table 2). As men- healthyareaofthetendonwiththetendinopathicareaof tioned in the methods section, GAPDH mRNA served thetendon,withexceptionoftheresultsfromIL-6,IL-1b, as a normalization factor for the genes of interest, and ki67andHGF-1.Thesegenetargetscouldnotbedetected RPLP0 mRNA was used to validate the stability of the in all samples. In these cases Chi2 tests were performed. expression of GAPDH mRNA. Specific gene targets AllPCRdataarepresentedasthegeomean±backtrans- were selected covering the areas of: Extracellular matrix formed SEM. The collagen fibril data were divided into (ECM) components, degradation components, Growth area and diameter fractions, and a paired Students t-test factors, inflammatory markers and fibrillogenesis. was performed to compare each fraction between the Extracellularmatrixcomponents:Theexpressionofsev- healthy and the tendinopathic area of the tendon. Like- eral structural components (collagen 1 and collagen 3) wise,thevolumefractionofcellsandthevolumefraction togetherwithmRNAoffibronectin,tenascin-candfibro- of the nucleus within the cell were compared using a modulinwasfoundtobesignificantlyupregulatedinthe PairedStudentt-testcomparingthetwoareasoftheten- t-area(Figure3).Versicanwasfoundtobeunchangedand don.AP-value<0.05wasconsideredtobesignificantand decorintendedtobedecreasedinthet-areacomparedto alldatadespiteofthesubjectcharacteristicsareshownas thatoftheh-areaoftheAchillestendon(Figure3). Mean±SEM. Degradationfactors:ExpressionofMMP-2,MMP-9and TIMP-2 was significantly increased in the t-areas com- Results paredtothat oftheh-areaswithnodifferenceinexpres- Structural composition of the tendon sionofTIMP-1(Figure3). The density, volume fraction and mean area of the col- Growthfactors: A significant up-regulation ofTGF-b1 lagenfibrilsweremeasuredinbiopsiesfrom14oftheten- expressionwasobservedinthet-areacomparedtothatof dinopathy patients (Table 1). The density of collagen the h-area, whereas bFGF and cmet expression in the fibrils wasfound to be significantly higher andthe mean sameareawassignificantlydecreased.Nodifferenceswere areaofthecollagenfibrilswassignificantlysmallerinten- observedintheexpressionofCTGF,VEGF-A1andIGF-1 dinopathic tendon region compared to that of healthy (Figure4).TheexpressionofHGF-1couldnotbedetected controlregion,additionallyatrendtowardssignificantdif- in all samples but no significant differences were found ference was found in volume fraction (Table 1). When betweenthetworegionsofthetendon(Table3). analysingtheindividualbinsin the diameterdistribution Inflammatory markers: Data on IL-6, IL-1b and ki67 could not be detected in all samples and no differences wereobservedwhenthetworegionswerecompared(chi Table 1Tendonfibril characteristics squared test). Ki67 showed a significant decrease in SickTendon HealthyTendon expression in the tendinopathic areas. No expression Tissue Tissue Mean SD Mean SD P-value could be detected of COX-2 and TNF-a in any of the samples(resultsnotshown).COX-1,IL-1RandCCL2was Density 155,73 48,80 111,09 46,72 0,04 detectedinallsamples,buttherewasnosignificantdiffer- VolumeFraction 0,56 0,08 0,61 0,08 0,06 encebetweenthet-areaandtheh-areaofthetendon(Fig- MeanArea 2963,54 1693,29 5239,15 2298,83 0,02 ure 4). Fibrillogenesis: There were no differences in Jetal.BMCMusculoskeletalDisorders2012,13:53 Page5of13 http://www.biomedcentral.com/1471-2474/13/53 Figure1Fibrildiameterdistribution.Fibrildiameterofthetendinopathicandthehealthyareaofthesametendondividedintofractions. ErrorbarsrepresentSEM(P<0.05). expression of scleraxis, tenomodulin and Lysyloxidase significantly larger fibril area and a lower collagen fibril (LOX)betweenthetwoareas(Figure5) density was observed in patellar tendinopathy [27]. This discrepancy might partly be explained by the fact of the Discussion patella tendon functions more like a ligament (ligament The major finding of the present study is that tendino- patella) with the primary role of ensuring a fixed dis- pathy causes focal biochemical and morphological differ- tance between the patella apex and the insertion of the ences in the human Achilles tendon. The data obtained tendon on to the tibia bone, in contrast to the Achilles using TEM indicate that the structural composition of tendon which functions more like a spring, providing a the t-area has a significantly increase in the number of means of shock absorption via the connective tissue, smaller collagen fibrils compared to the h-area of the during periods of muscle contraction and loading [28]. same tendon. This supports our hypothesis that loca- SincethefunctionofthepatellarandtheAchillestendon lized structural differences are present in tendinopathic differ,itislikelythatthestructureofthetendonse.g.the Achilles tendons, potentially as a result of an increased cross-linkingandlength offibrilsetc. of thetwotendons turnover of the tissue in an attempt to heal potentially reflects this difference. A further explanation might be injured tissue. This fits with previous findings where a thatthehealthytissuewastakenfromthesametendonin site-specific loss of larger collagen fibrils and an increase thepresent study, whileKongsgaardet al.[27] used con- of fibrils with a small diameter was observed in Achilles trol tissue taken from another tendon of healthy control tendons after tendon rupture [26]. However the present subjects. The present design has as all study designs findings differ somewhat from the results observed in advantages and limitations. The advantage is that this another study from our laboratory [27], in which a design enables to investigate local differences in the Jetal.BMCMusculoskeletalDisorders2012,13:53 Page6of13 http://www.biomedcentral.com/1471-2474/13/53 Figure2Volumefractionofcellsandcytoplasmwithinthecells.ErrorbarsrepresentSEM.(P<0.05). tendon,thelimitationisthatnocontroltissueisavailable tenomodulin was analysed. Both gene targets have pre- fromtendonsthatneverhadanysymptoms.Changesthat viouslybeenassociatedwithtendonformationanddevel- occurinthewholetendoncanthereforebeoverlooked.A opment[30,31].Theabsentdifferenceoftheexpressionof previousstudyhasshownthathistologicalchangesinthe scleraxisandtenomodulinsuggeststhatnofibrillogenesis tendon were not only present at the site of rupture but tookplaceinthet-areaatthisverylatestageinthedisease alsointhemacroscopicalnormalpartofthetendon,indi- (Figure5).FurthermoreasimilarexpressionofLysyloxi- catingthatlocalalterationsofthetissuemightnotneces- dase (LOX) in both regions of the tendon indicates that sarilybelocal[29].Whethertendinopathyshowsthesame thetissuedoesnotcompensateforthelocalizedstructural pattern is unclear and needs further investigation. How- changesbyinitiatingcross-linkstomaintainthemechani- everwebelievethatthepresentlocaldifferencesbetween cal properties of the tendon. Previously findings showed thet-areaandtheh-areaarestrongenoughtojustify the that training increases the expression of LOX in healthy conclusions of the present study. To investigate if the tendon tissue in rats [32]. The present data suggest that increasednumberofsmallcollagenfibrilscouldbedueto thisadaptationdoesnottakeplaceint-areaofthetendon. genesis of collagen fibrils or degradation of previously Several abnormalities of the tendon structure have been much larger fibrils, the gene expression of scleraxis and investigatedwithhistopathologicalanalysisincludingfibre Jetal.BMCMusculoskeletalDisorders2012,13:53 Page7of13 http://www.biomedcentral.com/1471-2474/13/53 Table 2PCRprimers mRNA SensePrimer AntisensePrimer Collagen1 GGCAACAGCCGCTTCACCTAC GCGGGAGGACTTGGTGGTTTT Collagen3 CACGGAAACACTGGTGGACAGATT ATGCCAGCTGCACATCAAGGAC Fibronectin TTTGCTCCTGCACATGCTTT TAGTGCCTTCGGGACTGGGTTC TenascinC CAACCATCACTGCCAAGTTCACAA GGGGGTCGCCAGGTAAGGAG Fibromodulin CAGTCAACACCAACCTGGAGAACC TGCAGAAGCTGCTGATGGAGAA Versican AGTCAGTGGAAGGCACGGCAATCT CCGTTAAGGCACGGGTTCATTT Decorin GGTGGGCTGGCAGAGCATAAGT TGTCCAGGTGGGCAGAAGTCA MMP-2 CCGCCTTTAACTGGAGCAAAAACA TTGGGGAAGCCAGGATCCATTT MMP-9 AGCGAGGTGGACCGGATGTT AGAAGCGGTCCTGGCAGAAATAG TIMP-1 CGGGGCTTCACCAAGACCTACA TGGTCCGTCCACAAGCAATGA TIMP-2 CTCGCTGGACGTTGGAGGAAAG GTGTCCCAGGGCACGATGAAGT CTGF TGCGAAGCTGACCTGGAAGAGA GCCGTCGGTACATACTCCACAGAA bFGF TGACGGGGTCCGGGAGAAGA ATAGCCAGGTAACGGTTAGCACACAC HGF TGAAATATGTGCTGGGGCTGAAA ACAAACAAGTGGGCCACCATAATCC cMet AACCCGAATACTGCCCAGACCC TGATATCCGGGACACCAGTTCAG VEGFA-1 ATGACGAGGGCCTGGAGTGTGT CTCCTATGTGCTGGCCTTGGTG IGF-1 GACATGCCCAAGACCCAGAAGGA CGGTGGCATGTCACTCTTCACTC TGFb-1 GAGGTCACCCGCGTGCTAATG CACGGGTTCAGGTACCGCTTCT Cox-1 GGTTTGGCATGAAACCCTACACCT CCTCCAACTCTGCTGCCATCT IL-1R GGAAGGGATGACTACGTTGGGGA CCAGCCAGCTGAAGCCTGATGTT IL-1b TCCAGGGACAGGATATGGAGCA AGGCCCAAGGCCACAGGTATTT KI67 CGGAAGAGCTGAACAGCAACGA GCGTCTGGAGCGCAGGGATA CCL GCCCTTCTGTGCCTGCTGCT GCAGGTGACTGGGGCATTGATT IL-6 GAGGCACTGGCAGAAAACAACC CCTCAAACTCCAAAAGACCAGTGATG TNF-a TTCCCCAGGGACCTCTCTCTAATC GAGGGTTTGCTACAACATGGGCTAC RPLP0 GGAAACTCTGCATTCTCGCTTCCT CCAGGACTCGTTTGTACCCGTTG GAPDH CCTCCTGCACCACCAACTGCTT GAGGGGCCATCCACAGTCTTCT structure,fibrearrangement,nuclearroundingandcellu- thetendon.Ithaspreviouslybeenshownthatnormalten- larity [15]. In the present study a significant increased don tissue express matrix metalloproteinases and that a volumefractionofcellswasobservedinthetendinopathic homeostaticturnoverisnecessaryfortendonregeneration area of the tendon using TEM. This is in line with pre- and maintenance [16]. A increased collagen turnover is viousanimalstudiesofSoslowskyandcolleagues[33-35], usuallyassociatedwithadaptationtoexercise[2]orheal- where rats ran with a velocity of 17 m/minute, 5 days/ ingofthetendon[38].Itisstillpuzzlingwhytheincreased week,1hour/day,eitheruphillordownhillforaperiodof collagen turnover inthe tendons of chronic patients like between2-16weeks.Insuchexperiments,adecreasedcol- thepresenthasnotresultedinadecreaseinsymptomsor lagen fibre organization and increased numbers of cell a healing of the tissue (symptoms range: 0.5-10 years nuclei were observed [36,37]. The present TEM analysis Table4).Howeverthesedataareconfirmedinotherstu- didunfortunatelynotallowfordistinguishingbetweenthe diesalsoshowingincreasedcollagenturnoverintendino- celltypesthatwerecounted,andthusitwasnotpossible pathic tendons [22,24] and in tendon ruptures [39]. toexcludethatothercelltypesthanjustfibroblastsmight Alterations in proteoglycans have previously been asso- have migrated into the t-area of the tendon. The signifi- ciated with tendon pathology [40,41]. Proteoglycans and cantly higher mRNA expression of both collagen I and glycoproteinsareessentialforthemaintenanceofhomeos- collagenIIIinthet-areashowsahighercollagensynthesis tasisoftheECMofthetendonandachievethisbyregulat- of the tendon. At the same time indicates the higher ingthecollagenfibrilassembly[42].Theupregulationof expression of MMP-2 and MMP-9 in the t-area an tenascin-C,andfibronectinisconsistentwithearlierfind- increased collagen matrix degradation. Together these ings [24,43]. The observed unchanged levels of versican findingsdisplayahighercollagenturnoverinthet-areaof contrasts earlier findings, where a significant down Jetal.BMCMusculoskeletalDisorders2012,13:53 Page8of13 http://www.biomedcentral.com/1471-2474/13/53 Figure 3 Collagen turnover. Gene expressions of collagens, non-collagenous matrix components, matrix metalloproteinases and metallopeptidaseinhibitors,shownasarelativeratiobetweenthetendinopathicandthehealthyregionofthetendon.Thehealthyregion equals1.ErrorbarsrepresentSEM.(P<0.05). regulationofversicanwasobservedinbothtendinopathic patientshadaverylonghistoryoftendonpain(range:0.5- andruptured tendontissue [40]. This discrepancy might 10 years). Thus the current biochemical situation in the lie in the medical history of the patients. The present tissueofthesepatientsmayhavechangedovertime.The Jetal.BMCMusculoskeletalDisorders2012,13:53 Page9of13 http://www.biomedcentral.com/1471-2474/13/53 Figure4 Tendon healing.GeneexpressionsofGrowthfactorsanddifferentinflammatory markers,shown asa relativeratio betweenthe tendinopathicandthehealthyregionofthetendon.Thehealthyregionequals1.ErrorbarsrepresentSEM.(P<0.05). observedtendencytoadecreaseofdecorinexpressionin depressed decorin is part of the healing response of the thet-areaofthetendonwascontrarytoourhypothesis.It tendons and thus beneficial for the patients is unclear. hasbeenpreviouslyshownthatadown-regulationofdec- Additionally the increased expressionof fibromodulinin orinusinganti-sensedecorininjectionsimprovedligament thet-areamaypartlyexplaintheobservationofmanythin healing in rabbits [44]. Whether the present finding of a collagen fibrils since fibromodulin participates in the Jetal.BMCMusculoskeletalDisorders2012,13:53 Page10of13 http://www.biomedcentral.com/1471-2474/13/53 Table 3Inflammatorymarkers after one hour of exercise in the form of running, in Healthy Tendinopathy Chitest healthyyoungmen[57].However,theissueastowhether local injections of IL-6 in tendinopathy patients may be (detected/notdetected) (detected/notdetected) beneficial to the healing process of a damaged tendon is IL-6 9/7 10/6 0.7 stillunknown.Thepainthattendinopathypatientsexperi- IL-1b 4/12 6/10 0.4 ence has previously among other been related to Sub- ki67 10/6 8/8 0.5 stance-P,aneuro-peptidewithvariousbiologicalfunctions HGF-1 10/6 7/9 0.3 includingpaintransmission[58,59].Sincenodifferencein total 16 16 Substance P expression were observed between the two areas, the present data might indicate that other factors matrixassemblyleadingtoadelayedfibrilformationand thansubstancePcanberesponsibleforthepainintendi- formationofthinnerfibrils[45,46].Treatmentwithinjec- nopathic tendons. It is however also possible that the tions of growth factors for tendon injuries has received expression of Substance-P is increased in the whole ten- much attentioninrecentyears. Growthfactorsare poly- donandthereforeoverlookedduetothepreviouslymen- peptidemoleculesthataredecisiveregulationofcellmeta- tionedlimitationsofthepresentdesign.Althoughtherole bolism and cell proliferation and are associated with of inflammation in tendinopathy is one that is often dis- tendonhealing[47-51].Studiesusinglocaladministration cussed, it has long been known that tendinopathy is pri- ofbFGF[52,53],HGF[54,55]andIGF[56]haveallshown marily a degenerative condition, in which inflammatory beneficialeffectsinthehealingprocessoftendoninjuries, cells in or around the lesion are absent. All markers of but not all injurieswere tendinopathiesthough. Exogen- inflammationthatweremeasuredshowednoupregulated ous injections of bFGF in human patellar tendons have expression in the t-area of the tendon (Table 3). This been shown to increase wound healing both in vitro and underlinesthatlong-termtendinopathyisnotprimarilyan invivoinpatellartendonmodelsaftersurgery[52,53],and inflammatoryprocess,butratheranongoingdegenerative likewise, collagen type III and cell proliferation was process.Althoughinflammationisabsentintendinopathy increasedafterexogenousbFGFinjectionsinpatellarten- atthislatestage,itdoesnotruleoutthataninflammation dons after surgery in vivo [53]. Recently, our group insult was present at the initiation of the tendinopathic showedthatthecytokineIL-6couldactasagrowthfactor process [60,61]. In fact, various inflammatory mediators intendontissue[57].Moreover,localinjectionsofrhIL-6 like TNF-alpha, IL-6, IL-15, IL-18 have been shown to havebeenshowntoincreasecollagensynthesisinhumans play a role especially in wound healing after injury [62] Figure5Fibrillogenesis.Geneexpressionsofmarkersforfibrillogenesis,shownasarelativeratiobetweenthetendinopathicandthehealthy regionofthetendon.Thehealthyregionequals1.ErrorbarsrepresentSEM(P<0.05).