review the influence of ageing and exercise on tendon...

Comparative Biochemistry and Physiology Part A 133(2002) 1039–1050

1095-6433/02/$ - see front matter� 2002 Elsevier Science Inc. All rights reserved.PII: S1095-6433Ž02.00148-4

Review

The influence of ageing and exercise on tendon growth anddegeneration—hypotheses for the initiation and prevention of strain-

induced tendinopathies�

R.K.W. Smith *, H.L. Birch , S. Goodman , D. Heinegard , A.E. Goodshipa, b a d b,c˚

Department of Veterinary Clinical Sciences, The Royal Veterinary College, Hawkshead Lane, North Mymms, Hatfield,a

Herts. AL9 7TA, UKDepartment of Veterinary Basic Sciences, The Royal Veterinary College, London, UKb

Institute of Orthopaedic and Musculoskeletal Sciences, University College London, London, UKc

Section for Connective Tissue Biology, Department of Cell and Molecular Biology, University of Lund, Lund, Swedend

Received 16 January 2002; received in revised form 13 May 2002; accepted 15 May 2002

Abstract

Strain-induced tendinopathy is a common injury in both human and equine athletes, with increasing incidenceassociated with greater involvement in sport and an increasingly aged population. This paper reviews our studies on theabundant non-collagenous protein, cartilage oligomeric matrix protein(COMP), in equine tendons. Its variation betweentendon type and site, age and exercise has provided an insight into how age and exercise influence tendon growth andmaturation. Tendons can be broadly divided into two types, reflecting their different matrix composition and function:the energy-storing tendons used for weight-bearing and locomotion, which suffer a high incidence of strain-inducedtendinopathy, and positional tendons involved in limb placement or manipulative skills. It would appear that whileenergy-storing tendon can respond to the mechanical forces applied to it during growth, there is no evidence that it cando so after skeletal maturity. Instead, cumulative fatigue damage causes degeneration at the molecular level, potentiallyweakening it and increasing the risk of clinical injury. Appropriate exercise regimes early in life may help to improvethe quality of growing tendon, thereby reducing the incidence of injury during ageing or subsequent athletic career.� 2002 Elsevier Science Inc. All rights reserved.

Keywords: Tendon; Ageing; Exercise; Physiology; Degeneration; Tendinopathy; Cartilage oligomeric matrix protein(COMP); Matrix

1. Introduction

Strain-induced tendinopathy is an extremelycommon injury in human and animal athletes.Studies in man have shown a high incidence ofinjury in the Achilles tendon of professional ath-

� This paper was presented at ‘Tendon – Bridging the Gap’,a symposium at the 2002 Society of Integrative and Compar-ative Biology. Participation was funded by SICB, The ShrinersHospitals for Children, and the National Science Foundation(IBN-0127260).

*Corresponding author. Tel:q44-1707-666297; fax:q44-1707-660671.

E-mail address: [email protected](R.K.W. Smith).

letes, while the incidence in the sedentary popu-lation is low (Gibbon et al., 1999). In horses,recent epidemiological studies have revealed anincidence in National Hunt horses(those horsesthat race over jumps) in training in the UK of43% (Pickersgill, 2000) and 46% of injuries onracecourses being due to tendon strain injuries(Williams et al., 2001). These specific strain-induced injuries are associated with specific ten-dons, the Achilles and patellar tendonyligament inhumans, and the superficial digital flexor tendon(SDFT) of the forelimb in the horse(Fig. 1).These particular energy-storing tendons serve not

1040 R.K.W. Smith et al. / Comparative Biochemistry and Physiology Part A 133 (2002) 1039–1050

Fig. 1. The normal anatomy of the equine distal limb. Note the relative positions of the superficial digital flexor tendon, which supportsthe metacarpophalangeal joint under weight bearing, and the common digital extensor tendon, which is a positional tendon and doesnot receive any weight-bearing load.

only to transmit muscle-generated force to bone tomove joints, but more importantly, they act likesprings to store elastic energy for energy-efficientlocomotion(Alexander, 1988). Interestingly, othertendons that do not serve this physiological role,but instead act to position segments of the limbsprimarily as a function of muscle contraction, suchas the digital extensor tendon in the horse(Fig. 1)and the anterior tibialis and hand and arm tendonsin humans, have a low incidence of strain-inducedtendinopathy. Epidemiological data in both humansand man have shown a strong association betweenthe incidence of strain-induced tendinopathy andboth age and exercise(Gibbon et al., 1999; Pick-ersgill, 2000). With an increasing interest in exer-cise for cardiac and bone health in an increasinglyageing human general population, there has beenan increase in the incidence of Achilles tendino-pathy in both younger athletes and in the elderlyin recent years(Moller et al., 1996; Houshian etal., 1998; Maffulli et al., 1999).

This difference in tendon-specific incidence ofclinical injury implies a difference in the responseof these two groups of tendons to physiologicaland biomechanical influences associated with age

and exercise. Most musculoskeletal tissues accom-modate changes in mechanical demand by under-going functional adaptation involving alterationsin mass and structural architecture. In the case oftendons, there have been relatively few studies toinvestigate the relationship between physiologicaldemand and associated changes in molecular com-position. Studies in miniature swine demonstrateda gross change in size in digital extensor, but notflexor tendons(Woo et al., 1982). This has beensupported by studies in thoroughbred horses, whichshowed exercise-induced hypertrophy of the digitalextensors, but not the energy-storing flexor tendons(Birch et al., 1999).

This difference between the two tendon types isreflected by differences in mechanical properties.The energy-storing tendons in both horses(Batsonet al., 2000) and humans(Birch et al., 2001) showsignificantly lower elastic modulus than the posi-tional tendons, which would support optimisationfor their specific physiological roles.

Tendon comprises connective tissue characteri-sed by cell populations within an extracellularmatrix. The functional role of both the completetendon and site-specific regions within a tendon is

1041R.K.W. Smith et al. / Comparative Biochemistry and Physiology Part A 133 (2002) 1039–1050

Fig. 2. Cartoon of the COMP molecule(courtesy of Dr K.Rosenberg).

reflected in the molecular composition of thematrix. Thus, tendons exhibit anisotropy in matrixcomposition between regions loaded primarily intension(‘tensile regions’, although also subjectedto compression when stretched) and regions sub-jected to additional external compression when thetendon changes direction around a bony promi-nence (‘compressed regions’). The site-specificfunctional regions are reflected in terms of matrixcomposition, with the tensile regions containing alarger abundance of the species of small proteogly-can (the leucine-rich repeat proteins, such asdecorin and fibromodulin), and the compressedregions having a more fibrocartilagenous pheno-type, with the large proteoglycans of aggrecan andversican predominating(Vogel and Heinegard,˚1985; Vogel and Koob, 1989; Evanko and Vogel,1990; Vogel et al., 1994). This site-specific matrixcomposition has been shown to be linked tofunction by tendon transfer experiments in vivo(Gillard et al., 1979) and explant loading experi-ments in vitro(Koob and Vogel, 1987; Koob etal., 1992; Evanko and Vogel, 1993), suggestingsome evidence for functional adaptation to theseextremes of biomechanical environment.

Another non-collagenous protein recently foundto be abundant in tendon is cartilage oligomericmatrix protein (COMP), a five-armed protein,bound via disulfide bonds at their N-termini andwith globular C-terminal domains(Hedbom et al.,1992; Efimov et al., 1994) (Fig. 2). It is a largeprotein that can be visualised as a ‘bouquet’ with

rotary shadowing electron microscopy(Morgelin¨et al., 1992). Its precise function is unknown,although it has been shown to bind, via its globularC terminal domains, to fibrillar collagen molecules(Type I, II), as well as to more complex fibril-binding collagen IX, with a zinc-dependent mech-anism(Rosenberg et al., 1998. Thur et al., 2001).This protein has proved an interesting componentin tendon because of its restricted distributionwithin the body to those soft tissues for which theprimary role is to resist load(tendon, ligament,cartilage, meniscus and intervertebral disc), whichsuggests it has a function in this role.

As further evidence of its importance in tendon,mutations in COMP have been shown to causepseudoachondroplasia in man(Briggs et al., 1995;Hecht et al., 1995; Holden et al., 2001; Thur etal., 2001), which is characterised by short stature,early-onset osteoarthritis, and lax tendons andligaments. Thus, the phenotype corresponds to thedistribution of COMP, but this mutation would notappear to provide conclusive proof of the impor-tance of COMP for tissue structural integrity, as itresults in retention of the protein in the endoplasm-ic reticulum, thereby potentially disrupting therelease of other matrix molecules.

Despite the evidence to suggest the adaptabilityof a site-specific matrix, the high incidence ofinjury, influenced by ageing, and a failure to showchanges in size of the flexor tendons with exercisesuggest dissociation between functional demandand matrix adaptation in energy-storing tendons.The horse is an ideal model with which to studythe effects of age and exercise on these tendons,as it is a species that has phases of growth,maturation and ageing, while also being athletic.The superficial digital flexor tendon in the horseis subjected to very high loads—10–20 kN(1–2t of weight; Goodship et al., 1994) on a structurethat is approximately 100 mm in cross-sectional2

area and 45 cm in length. In doing so, it stretchesup to 16% in vivo during high-speed locomotion(Stephens et al., 1989), which is close to its failurestrain (superficial digital flexor tendons rupture atbetween 12 and 20% in vitro; Goodship et al.,1994).

Since this particular tendon contributes to ener-getic efficiency in an animal evolved and furtherselectively bred for high-speed locomotion, it func-tions close to failure with a low safety margin.This may explain the extremely high incidence ofpartial rupture of this specific tendon. Once ten-


dons are injured, the slow process of repair withfibrous scar tissue occurs. This scar tissue isessentially the same as that forming in skin woundsand does not have the same matrix composition asthe normal tendon. While, given time, this repairtissue can become mechanically strong(Crevier-Denoix et al., 1997), there is an increase in elasticmodulus, and hence the tissue is compromised infunctional efficiency. The high stiffness of scartissue also puts increased strain on adjacent, rela-tively less injured and remote areas of the sametendon, resulting in a high incidence of recurrentinjury—a common feature in both equine andhuman athletes. Prognosis for these injuries islargely dependent on the severity of the initialinjury, with little contribution from the therapeuticregime subsequently applied. Current treatmentmodalities are primarily aimed at ‘educating’ thereparative tissue to be more functional, but none,at present, are able toregenerate normal tendontissue. It is the authors’ belief that the greatestcontribution to restoration of function is by astructured rehabilitation programme with appropri-ate monitoring. This imposition of a controlledfunctional stimulus has the potential to modulatethe repair process in terms of both the structuralmorphology of the repair tissue and the molecularcomposition of the matrix.

Previous post mortem studies on equine tendonsuggested that the SDFT had a preceding stage ofsub-clinical damage that predisposed the horse toclinical injury (Webbon, 1977). This has beensupported by the identification of ‘sub-clinical’central discoloured regions of flexor tendons thatappeared to be normal on clinical examinationprior to euthanasia(Goodship et al., 1994). Sub-sequent analysis of similar tendons using moreadvanced modern ultrasound imaging can nowdetect these lesions. The injured tendon isenlarged, and matrix analyses of the red-stainedcentral core lesions support the interpretation thatscar tissue is present(high collagen III and gly-cosaminoglycans(Birch et al., 1998)). Notwith-standing this finding, Achilles micro-tears havebeen identified in asymptomatic professional ath-letes(Gibbon et al., 1999), and histological eval-uation of rotator cuff tendons(Riley et al., 1994)and electron microscopy examination of tendons(Kannus and Jozsa, 1991) have both identifieddegenerative pathology, which, as in the equineSDFT, may precede overt clinical injury. Further-more, many strain-induced tendinopathies in both

man(Gibbon et al., 1994) and the horse(Webbon,1977) are bilateral, which is inconsistent with asudden onset injury with no prior pathology.

In support of the hypothesis that clinical tendi-nopathy is a consequence of preceding matrixdegeneration are the epidemiological studies dem-onstrating a strong association with age(Gibbonet al., 1999; Pickersgill, 2000). Only smallamounts of degeneration dramatically increase therisk of progression to clinical tendinopathy,because the tendon is operating close to its toler-ance limits. Therefore, preventing such cumulativedegeneration potentially has a large effect in reduc-ing the incidence of tendon injury in a populationof athletes.

Furthermore, we have hypothesised that age andexercise are synergistic in the adult horse in initi-ating tendon degeneration(Smith et al., 1999).Understanding the processes that control matrixcomposition in relation to mechanical pathophy-siological influences will give us strategies withwhich we can potentially prevent tendon injuries.

This paper reviews our data on the effects ofboth age and exercise on tendon matrix composi-tion, from which we propose rational strategies forthe prevention of strain-induced tendinopathy.

2. General materials and methods

2.1. Exercise studies

2.1.1. Mature horsesTwo exercise studies were performed, each with

12 thoroughbred fillies. These have been describedin detail elsewhere(Patterson-Kane et al., 1997;Smith et al., 1999). In brief, for the first experi-ment (long-term study), five 21-month-old horseswere exercised at high speed on a treadmill threetimes per week for 18 months before euthanasia(at 3 years 3 months of age). A control group ofsix horses was given only walking exercise duringthe same period.

In the second experiment(short-term study),similar treadmill exercise was given for 4.5 monthsto six horses from 19 months of age. Anothercontrol group of six horses of the same age wasgiven only walking exercise.

2.1.2. Immature horsesTwo individual clinical cases were analysed, as

these had had non weight-bearing lameness on onelimb for 6 weeks. The first case was a 5-week-old


foal that suffered a radial fracture, which wasrepaired with internal fixation. However, althoughthe foal was ambulatory, it failed to bear weighton the affected limb for 6 weeks. At this point(at11 weeks of age) it was euthanased because ofthe lack of clinical progress. The digital flexor andextensor tendons(and metacarpophalangeal jointcartilage) from both limbs were analysed forCOMP.

In the second case, an 18-month-old horse suf-fered an infected elbow, which rendered it non-weight-bearing on one limb for a similar period(6 weeks). Euthanasia was performed at the endof this period because of the failure in eliminationof sepsis and the associated hopeless prognosis.The digital flexor tendons were immediately recov-ered post mortem for COMP analysis.

In an exercise experiment on foals, 43 warm-blood foals were divided into three groups from 1week of age. The exercise protocol has beendescribed in detail elsewhere(van Weeren andBarneveld, 1999). In brief, Group 1 receivedrestricted exercise(box rest), Group 2 was kept ina box, but given enforced, high-speed exercise forshort periods for 5 days of the week(training),and Group 3 was allowed free exercise(pasture).Half of the foals were euthanased at 5 months ofage and the remainder were grouped together andgiven low-level free exercise until they were 11months old, when the remainder of the foals wereeuthanased. The digital flexor tendons were imme-diately recovered post mortem and analysed forCOMP.

2.2. Tendon sample recovery from cadavers

The forelimb superficial digital flexor tendonand, in some cases, the common digital extensortendon were immediately recovered post mortemfrom horses of known age euthanased for reasonsother than tendon disease. Samples from the SDFTcomprised of tensile(metacarpal) and compressed(metacarpophalangeal) regions. Only the metacar-pal region of the common digital extensor tendonwas used, since this tendon does not have anyportion under compressive strain.

The tendons were either stored frozen for com-positional analysis or divided fresh into small 2-mm cubes for tenocyte recovery by collagenasedigestion. Tenocytes recovered in this way werecultured to confluence and passaged for analysisof COMP synthesis in response to TGF-b and

biaxial strain in the Flexercell apparatus(Buckleyet al., 1988; Brown et al., 2000) with or withoutthe presence of TGF-b.

2.3. Cartilage oligomeric matrix protein (COMP)analysis

A section of tendon(approximately 500 mg wetweight) was frozen in liquid nitrogen and groundin a tissue disintegrator(Retsch). In some cases,dry weights were first obtained by freeze-dryingbefore grinding to give a measurement of hydra-tion. For tendon ground without freeze-drying, 20vyw of extraction buffer(4 M guanidine hydro-chloride with protease inhibitors) was added to thefrozen powder and extraction carried out at 48Cfor 24 h with continuous agitation. For the freeze-dried samples, 100 vyw of extraction buffer wasused.

After 24 h, the supernatant was recovered bycentrifugation. A 10-ml aliquot of this supernatantwas twice precipitated in 95% ethanoly50 mMsodium acetate. After the second precipitation, theprecipitate was dried and dissolved in samplebuffer. Appropriate dilutions of the samples weremade, determined from parallel inhibition curves,and the COMP in the sample was quantified intriplicate using a homologous inhibition ELISA aspreviously described(Smith et al., 1997), usingpurified equine COMP as plate coating, and stan-dards and anti-equine COMP polyclonal antiserumas primary antibody. This assay has previouslybeen shown to detect COMP at concentrationsbetween 0.005 and 0.05mgyml. The COMP con-centration in the extract was then used to calculatethe COMP content of the tissue in mgyg wetweight.

3. Results

3.1. COMP variation with age (Smith et al., 1997)

Matrix analysis of these cadaveric tendons dem-onstrated considerable variation in COMP levels,both between tendon type and site, and with age.COMP levels were universally low in all tendonsat birth (-1 mgyg wet weight), but increased inan exponential fashion as a function of age onlywithin the tensile region of the digital flexortendons, to peak at approximately 2 years(;10mgyg wet weight). Thereafter, COMP declined tolower levels(;2 mgyg wet weight) after approx-


Fig. 3. COMP levels in tensile(metacarpal) and compressed(metacarpophalangeal) regions of the superficial digital flexortendon (SDFT) in the horse. The tensile region shows analmost exponential increase up to 2 years of age, and thereafterfalls precipitously. In contrast, the compressed region shows amore gradual rise to lower levels that are maintained. The com-mon digital extensor tendon(CDET) over the same age rangeis largely unchanged at the same level as in the neonate flexortendons (long dashed line; E. Batson, personalcommunication).

Fig. 4. COMP levels in the tensile region of the superficialdigital flexor tendon of the horse following intensive treadmillexercise(E) or just walking(controls, NE). The light-grey barsrefer to the first experiment, when 18 months of exercise wasgiven and the horses were over 3 years at the time of analysis;the dark-grey bars refer to the second experiment, when ashorter period of exercise was given and the horses were only2 years old when analysed. Compare these ages with Fig. 3.There was a significantdecrease in COMP levels in compar-ison to controls, but only with the longer period of exercisewhen the animals were older and levels of COMP weredeclining.

imately 5 years of age(Fig. 3). Within the com-pressed region of the equine superficial digitalflexor tendon, levels peaked at a similar age, butat approximately half that observed in the tensileregion, and this level appeared to be maintainedduring maturity. However, the levels in the digitalextensor tendons remained essentially constant ata low level (-1 mgyg wet weight; E. Batson,personal communication). This pattern of COMPexpression between tendon types appeared to besimilar in human positional and energy-storingtendons, as there were low levels in anterior tibialisand hand and arm tendons, and higher levels inhuman Achilles tendon(Heinegard, Riley and˚Smith, unpublished data).

3.2. Correlation of COMP levels with mechanicalproperties (Smith et al., 2000)

When COMP levels were compared to mechan-ical properties, such as ultimate tensile stress andstiffness, there was a statistically positive correla-tion between the two(Rs0.72, Ps0.008; andRs0.61, Ps0.003, respectively) but only whenanimals of the same age at skeletal maturity werecompared.

3.3. Exercise studies

3.3.1. Mature horses (Smith et al., 1999)No significant alterations between exercised and

control groups were observed for COMP levels inthe short-term study when the horses were euthan-ased after 4 months of exercise at 2 years of age.In contrast, the horses given 18 months of exercise(3 years old at euthanasia) had a lower level ofCOMP, as would be predicted from the cadavericdata, but also showed a significant decrease inCOMP levels with exercise in comparison to thecontrols(Fig. 4).

3.3.2. Immature horsesIn the foal with non-weight-bearing on one limb

for 6 weeks at 5 weeks of age, there was consid-erably higher levels(four-fold) of COMP in theweight-bearing tendons(and articular cartilage) incomparison to the non-loaded limb. In contrast,the digital extensor tendons, which normally havelow levels of COMP, showed no difference inCOMP levels. However, in a similar situation in anear-adult horse, when COMP has already accu-mulated in the tendon, there was little effect ofreduced loading(Fig. 5).

In the exercise study on foals, COMP levelswere significantly lower in the foals given forced


Fig. 5. COMP levels in the tendons(and metacarpophalangeal joint cartilage) of two clinical cases that suffered non-weight-bearinglameness for 6 weeks. One case was 11 weeks old at euthanasia and the other closer to skeletal maturity at 18 months of age. Thereis a failure to accumulate COMP in the digital flexor tendons and cartilage in the non-weight-bearing limb in the 11-week-old animal.The effect is less in the 18-month-old animal, which has already accumulated considerable amounts within its matrix. Note that thelevels in the digital extensor tendon of the 11-week-old animal are unaltered, as not only is the mechanical environment unaltered forthis tendon, but this tendon is also characterised by not accumulating COMP during growth. Key: T, tensile(metacarpal) region; C,compressed(metacarpophalangeal) region; MCP, metacarpophalangeal.

exercise in comparison to those at pasture at 5months of age(Fig. 6) (Cherdchutham et al.,1999). Following a further 6 months of lower-intensity exercise, although there was a similarpattern to the COMP levels between groups, therewere no longer any significant differences.

3.4. Mechanism of COMP accumulation: the influ-ence of growth factors

Tenocytes showed a dose-dependent increase inCOMP production with TGF-b isoforms(1, 2 and3) (Fig. 7). This effect was greater than that ofmechanical load, which had a synergistic effect onCOMP synthesis.

Tenocytes recovered from digital flexor tendonshowed a greater COMP synthetic response thancells from extensor tendons(Goodman et al.,2000). Immature flexor tenocytes appeared to syn-thesise greater amounts of COMP than matureflexor tenocytes or extensor tenocytes from anyage, although the latter group were still able tosynthesise appreciable amounts of COMP(Good-man et al., 2000).

4. Discussion

Together with a recent finding demonstratingthat the COMP promoter is most active in tendonand ligament(Deere et al., 2001), the finding thatCOMP is most abundant in the tensile region ofenergy-storing tendons in young animals indicatesthat the presence of this protein is unrelated to anycartilage phenotype.

The positive correlation of COMP levels withultimate tensile strength and stiffness in equinedigital flexor tendons at skeletal maturity suggeststhat COMP is in some way linked to the formationof a strong tendon matrix. Mechanical analysis ofcadaveric tendons has shown a large(up to two-fold) variation in mechanical, structural and mate-rial properties within a given population, with nosignificant difference between tendons from leftand right limbs within individual animals(Birchet al., 2000). This suggests that some individualshave developed weaker tendons than others, andthat it is possible that these individuals are mostsusceptible to tendinopathy.


Fig. 6. COMP levels in the tensile(metacarpal) region of the superficial digital flexor tendon in growing foals at 5(5m) and 11 months(11m) of age. The foals were given three different exercise regimes: box rest(maintained in a stall, with only minimal exercise);training (maintained in a stall, but given short periods of intense, enforced exercise); and pasture(maintained out at pasture, wherenatural regular self-exercise occurs). After 5 months, the foals were grouped together and given lower levels of exercise to determineif any deleterious effects of the enforced training would resolve. There was a significant decrease in COMP levels between the pastureand trained groups at 5 months. No significant differences were observed at 11 months.

Fig. 7. Dose response for digital flexor and extensor tenocytes(ns3; from neonatal tensile tendon) of COMP synthesis with additionof TGF beta 1. Note the increased COMP synthesis by tenocytes derived from the digital flexor tendon in comparison to those fromthe digital extensor tendon.

The maximum accumulation of COMP duringtendon development and subsequent loss aftergrowth is finished would suggest that its mainfunction is in the formation of tensile tendonmatrix. As the C-terminal domain of COMP bindsto a collagen molecule at four points equallydistributed along the collagen molecule(Rosen-berg et al., 1998), the five-armed structure of

COMP acts to align collagen molecules and assistin early events in collagen fibrillogenesis(Rosen-berg, Morgelin and Heinegard, unpublished data).¨ ˚This has led to the speculation that COMP facili-tates the early stages of collagen fibril formationin tissue. Adequate levels of COMP may thereforebe necessary for the formation of a structurallycompetent collagen matrix. Thus COMP appears


to be most highly expressed in those tissues syn-thesising matrix designed to withstand the largesttensile(or compressive) forces.

The data reviewed and presented in this paperare consistent with the energy-storing tendon hav-ing two phases in its life—a growth phase and anageing phase. At birth, the tendon appears to be‘blank’—i.e. homogeneous throughout its lengthand unadapted to load-bearing, in a similar fashionto that suggested for cartilage(Little and Ghosh,1997; Brama et al., 2000). Site-specific variationsrelated to external compressive forces have alsobeen shown to develop after birth(Vogel andHeinegard, 1985; Vogel and Evanko, 1987). Dur-˚ing growth, the tendon accumulates tendon matrix(as illustrated by COMP), which is modulated bythe biomechanical environment. Appropriate exer-cise therefore appears to be necessary to stimulateoptimal tendon development, although the tendonmay also be potentially more easily injured duringthis phase if the loadingyexercise is excessive. Insupport of this, enforced exercise in foals resultedin lower COMP levels at 5 months and reducedDNA and polysulfated glycosaminoglycans after11 months compared to the restricted or freeexercise groups(Cherdchutham et al., 1999). Othermatrix analyses in these growing tendons haveprovided further evidence that a combination ofimmobilisation and enforced exercise appears moredamaging to developing tendon than more regular,limited or free exercise(Cherdchutham et al.,1999).

During growth, growth factors potentially actsynergistically with load to stimulate tendon matrixsynthesis. The distribution of TGF-b isoforms inequine tendon shows an age-dependent distribu-tion, with these isoforms being distributed through-out the tendon fascicles and endotenon duringgrowth. However, after skeletal maturity, theimmunohistochemical staining pattern changes,with TGF-b becoming solely located within theendotenon septa separating the tendon fascicles(Cauvin et al., 2000; Cauvin, 2001).

However, while energy-storing tendon canpotentially adapt to loading(exercise) duringgrowth, it has little ability to do so after skeletalmaturity. After this period, this type of tendonappears to age, with the loss andyor disruption ofmatrix—a change which is accelerated by exercise.The persistence of COMP levels in the compressedregions of the digital flexor tendons is consistentwith matrix synthesis still being active in these

regions during life. This observation in equinetendon correlates very well with the gene expres-sion data for a variety of matrix proteins deter-mined for different regions of bovine digital flexortendons at different ages(Perez-Castro and Vogel,1999). In this study using in situ hybridisation,gene expression was evident in all areas of thetendon during growth and in the adult compressedtendon, but absent from adult tensile tendon.

After skeletal maturity, the energy-storing flexortendon appears to cease matrix synthesis and ispoorly, if at all, responsive to loadingyexercise.Exercise given during this phase serves only toaccelerate the ageing effect of accumulating micro-trauma, resulting in tendon matrix degeneration.

This hypothesis is supported by other analysesof these tendons. Collagen fibril analysis hasshown an increased predominance of small-diam-eter fibrils in the central core region of the long-term exercised group(Patterson-Kane et al., 1997).While this could represent the formation of newcollagen, there was no significant increase intendon size or collagen content, and the collagen‘age’ determined by autofluorescence of non-enzy-matic glycation products was not significantlydifferent between the two groups. Both thesefindings suggest that exercise causes disruption ofthe collagen and non-collagenous matrix, ratherthan adaptive hypertrophy.

This may appear contra-intuitive, as other skel-etal tissues, such as muscle and bone, have beenshown to dramatically respond to load. However,it is now known that, at least, the bone responseis also very age-dependent, with the highest activ-ity during puberty and the lowest activity in theaged. As the SDFT of the horse has to essentiallyfunction as a spring to store energy on loading,tissue stiffness is critical for optimising this func-tion. Too stiff or too elastic a ‘spring’ would makeit less efficient for the given weight of animal. Itis therefore possible that, once an optimised tendonis formed at skeletal maturity, there is a much-reduced need to synthesise further matrix, and sothe synthetic machinery is ‘turned off’.

The control of expression of genes for the matrixprotein composition in the tendon during growth,exercise and ageing is not clear. The low level ofmatrix synthesis could be a consequence of cellularsenescence, loss of mechanotransduction, orgrowth factor stimulus. Allied studies in our lab-oratories have shown that immunohistochemicalstaining of the prototypic anabolic growth factor,


Fig. 8. Stylised hypothetical response of tendons to the earlyintroduction of exercise. The two dotted lines refer to two sep-arate individuals: one that develops poor-quality tendon duringgrowth (long dashed line) and one that develops good-qualitytendon(short dashed line). Both accumulate damage withinthe tendon associated with post-skeletal exercise and ageing,which inevitably and progressively weaken the tendon. How-ever, in the former, this weakening predisposes the individualto clinical tendonitis during its racing career, while the latteris able to survive its racing career without suffering such injury.We hypothesise that by providing ‘appropriate’ conditioningexercise during growth, the quality of the growing tendon canbe improved, resulting in an individual more resistant to sub-sequent injury.

TGF-b, is lost from tendon fascicular tissue afterskeletal maturity. Furthermore, there is no detect-able TGF-b gene expression after birth(Cauvin etal., 2000; Cauvin, 2001). These findings are mostconsistent with a mechanism more related to theabsence of necessary growth factors than theabsence of cellular synthetic ability, although areduction in cell activity and the loss of a co-ordinated mechanotransduction response(such asthe loss of gap junctions between cells) could alsocontribute (McNeilly et al., 1996). Should themajor influence on the synthetic ability of thetendon be the lack of appropriate growth factors,this presents a possible therapeutic target for theprevention of tendon degeneration.

The mature tendon normally endures the naturallife-style of the horse. It ages and accumulatesmicro-damage, but only to levels that would notendanger the functional competence of the tissue.However, the artificial superimposition ofincreased exercise on this ageing mechanism(through training and racing horses, or increasedathleticism and longevity in man) potentiallyaccelerates this degenerative process to weakenthe tissue sufficiently to allow the initiation ofclinical tendinopathy when loads overcome thefunctional capacity of the tendon.

There are several potential candidates for amechanism of matrix degeneration. These includedirect matrix damage from physical forces, hyper-thermia of tendon tissue caused by energy lost asheat through hysteresis(Wilson and Goodship,1993) and relative hypoxia(Stromberg and Tufves-¨son, 1969; Kannus and Jozsa, 1991; Jarvinen etal., 1997), all of which could either act directlyon the matrix or on tenocytes to cause a releaseof destructive proteolytic enzymes.

From the results of these studies, the followinghypothesised strategies for the prevention of strain-induced tendinopathies can be proposed.

1. Maximise the quality of tendon prior to skeletalmaturity with:– Early introduction of controlled exercise reg-

imens with appropriate monitoring(Fig. 8).The exact level quantity and intensity ofexercise necessary is currently unknown, buta ‘window’ of opportunity in both time andamountyintensity probably exists. If tendon,at this stage, responds in fashion similar tobone, then relatively short periods of highand diverse strain rates would be most appro-

priate. It is interesting to speculate that thenormal gambolling activity of young foals atpasture, which includes a large amount ofprancing jumps, may be perfectly designedfor this function.

– Improved genetic determinants.2. Reduce the degeneration after skeletal maturity:

– Avoid training directed at tendon adaptationin the adult, as it will serve only to acceleratedegeneration, and hence long periods of highpeak loads should be avoided.

– Prevent the cumulative microdamage in adulttendon by identifying and preventing theprocesses involved(e.g. the release of prote-olytic enzymes).

– Reactivate resident cell populations to repairmicrodamage, or at least re-activate matrixsynthesis.

3. Reduce the risk factors for provoking clinicaltendinopathy(e.g. high peak loads caused byincoordination caused by fatigue, hard groundsurfaces on which the horse is running, etc.).

We are currently evaluating the first of thesestrategies in vivo. We have been able to demon-


strate a significant increase in the rate of increasein SDFT cross-sectional area with a short periodof additional exercise from 3 to 15 months of age(Kasashima et al., in press). Matrix analysis willbe needed to confirm that this change is trueadaptation rather than injury or abnormal matrixcomposition.

Acknowledgments

We acknowledge the funding of the HorseraceBetting Levy Board, the Home of Rest for Horses,The Worshipful Company of Saddlers and theWellcome Trust, who have made these studiespossible.

References

Alexander, R.M., 1988. Elastic Mechanisms in Animal Move-ment. Cambridge University Press, Cambridge.

Batson, E.L., Paramour, R.J., Smith, T.J., Birch, H.L., Good-ship, A.E., 2000. A comparison of the elastic modulus ofthe equine superficial digital flexor tendon with the commondigital extensor tendon. Proceedings of the XVIIth FECTSMeeting. p. L14.

Birch, H.L., Bailey, A.J., Goodship, A.E., 1998. Macroscopic‘degeneration’ of equine superficial digital flexor tendon isaccompanied by a change in extracellular matrix composi-tion. Equine Vet. J. 30, 534–539.

Birch, H.L., McLaughlin, L., Smith, R.K.W., Goodship, A.E.,1999. Treadmill exercise-induced tendon hypertrophy:assessment of tendons with different mechanical functions.Equine Vet. J. 30, 222–226.

Birch, H.L., Smith, T.J., Lawes, T.L., Goodship, A.E., 2000.Mechanical properties of equine flexor tendons show sym-metry between right and left forelimbs within individualsbut a wide variation in strength and stiffness betweenindividuals. Proceedings of the XVIIth FECTS Meeting. p.L13.

Birch, H.L., Smith, T., Goodship, A.E., 2001. Properties ofhuman lower extremity tendons are structure specific. J.Bone Joint Surg.wBr.x 83B, 342.

Brama, P.A., Tekoppele, J.M., Bank, R.A., Barneveld, A.,van Weeren, P.R., 2000. Functional adaptation of equinearticular cartilage: the formation of regional biochemicalcharacteristics up to age one year. Equine Vet. J. 32,217–221.

Briggs, M.D., Hoffman, S.M.G., King, L.M., et al., 1995.Pseudoachondroplasia and multiple epiphyseal dysplasia dueto mutations in the cartilage oligomeric matrix protein gene.Nat. Genet. 10, 330–336.

Brown, T.D., Bottlang, M., Pedersen, D.R., Banes, A.J., 2000.Development and experimental validation of a fluidystruc-ture-interaction finite element model of a vacuum-drivencell culture mechanostimulus system. Comput. MethodsBiomech. Biomed. Eng. 3, 65–78.

Buckley, M.J., Banes, A.J., Levin, L.G., et al., 1988. Osteo-blasts increase their rate of division and align in response

to cyclic, mechanical tension in vitro. Bone Miner. 4,225–236.

Cauvin, E.R.J., 2001. An Investigation into the Roles ofTransforming Growth Factor-beta(TGFb) in the Develop-ment, Adaptation and Repair of Equine Tendons, PhDThesis. University of London.

Cauvin, E.R., Goodman, S., Bird, J., et al., 2000. Expressionof TGF beta isoforms during tendon development andmaturation. Proceedings of the BSMB Meeting ‘MolecularCell Biology of the Synovial Joint’. p. 12.

Cherdchutham, W., Becker, C., Smith, R.K.W., Barneveld, A.,van Weeren, P.R., 1999. Age-related changes and effect ofexercise on the molecular composition of immature equinesuperficial digital flexor tendons. Equine Vet. J. 31, 86–94.

Crevier-Denoix, N., Collobert, C., Pourcelot, P., et al., 1997.Mechanical properties of pathological equine superficialdigital flexor tendons. Equine Vet. J. 23, 23–26.

Deere, M., Rhoades Hall, C., Gunning, K.B., LeFebvre, V.,Ridall, A.L., Hecht, J.T., 2001. Analysis of the promoterregion of human cartilage oligomeric matrix protein(COMP). Matrix Biol. 19, 783–792.

Efimov, V.P., Lustig, A., Engel, J., 1994. The thrombospondin-like chains of cartilage oligomeric matrix protein are assem-bled by a five-strandeda-helical bundle between residues20 and 83. FEBS Lett. 341, 54–58.

Evanko, S.P., Vogel, K.G., 1990. Ultrastructure and proteogly-can composition in the developing fibrocartilagenous regionof bovine tendon. Matrix 10, 420–436.

Evanko, S.P., Vogel, K.G., 1993. Proteoglycan synthesis infetal tendon is differentially regulated by cyclic compressionin vitro. Arch. Biochem. Biophys. 307, 153–164.

Gibbon, W.W., Cooper, J.R., Radcliffe, G.S., 1999. Sonograph-ic incidence of tendon microtears in athletes with chronicAchilles tendinosis. Br. J. Sports Med. 33, 129–130.

Gibbon, W., Winter, F., Johnson, S., 1994. Ultrasound appear-ances of the asymptomatic Achilles tendon in patients withunilateral Achilles tendon symptoms. Proceedings of the26th Annual Meeting of the British Medical UltrasoundSociety. p. 2.

Gillard, G.C., Reilly, H.C., Bell-Booth, P.G., Flint, M.H., 1979.The influence of mechanical forces on the glycosaminogly-can content of the rabbit flexor digitorum profundus tendon.Connect. Tissue Res. 7, 37–46.

Goodman, S., May, S.A., Smith, R.K.W., 2000. Cyclical strainenhances the TGFb-induced production of cartilage oligo-meric matrix protein by equine tenocytes. Proceedings ofthe BSMB Meeting ‘Molecular Cell Biology of the SynovialJoint’. p. 28.

Goodship, A.E., Birch, H.L., Wilson, A.M., 1994. The patho-biology and repair of tendon and ligament injury. Vet.Clinics N. Am.: Equine Pract. 10, 323–349.

Hecht, J.T., Nelson, L.D., Crowder, E., et al., 1995. Mutationsin exon 17B of cartilage oligomeric matrix protein(COMP)cause pseudoachondroplasia. Nat. Genet. 10, 325–329.

Hedbom, E., Antonsson, P., Hjerpes, A., et al., 1992. Cartilagematrix proteins. An acidic oligomeric protein(COMP)detected only in cartilage. J. Biol. Chem. 267, 6132–6136.

Holden, P., Meadows, R.S., Chapman, K.L., Grant, M.E.,Kadler, K.E., Briggs, M.D., 2001. Cartilage oligomericmatrix protein interacts with type IX collagen, and disrup-tions to these interactions identify a pathogenetic mechanismin a bone dysplasia family. J. Biol. Chem. 276, 6046–6055.


Houshian, S., Tscherning, T., Riegels-Nielsen, P., 1998. Theepidemiology of Achilles tendon rupture in a Danish county.Injury 29, 651–654.

Jarvinen, M., Jozsa, L., Kannus, P., Jarvinen, T.L., Kvist, M.,Leadbetter, W., 1997. Histopathological findings in chronictendon disorders. Scand. J. Med. Sci. Sports 7, 86–95.

Kasashima, Y., Smith, R.K.W., Birch, H.L., Takahashi, T.,Kusano, K., Goodship, A.E. Exercise-induced tendon hyper-trophy: cross-sectional area changes during growth is influ-enced by exercise. Equine Vet. J., in press.

Kannus, P., Jozsa, L., 1991. Histopathological changes preced-ing spontaneous rupture of a tendon. A controlled study of891 patients. J. Bone Joint Surg. Am. 73, 1507–1525.

Koob, T.J., Clark, P.E., Hernandez, D.J., Thurmond, F.A.,Vogel, K.G., 1992. Compression loading in vitro regulatesproteoglycan synthesis by tendon fibrocartilage. Arch.Biochem. Biophys. 298, 303–312.

Koob, T.J., Vogel, K.G., 1987. Proteoglycan synthesis in organcultures from regions of bovine tendon subjected to differentmechanical forces. Biochem. J. 246, 589–598.

Little, C.B., Ghosh, P., 1997. Variation in proteoglycan metab-olism by articular chondrocytes in different joint regions isdetermined by post-natal mechanical loading. OsteoarthritisCartilage 5, 49–62.

McNeilly, C.M., Banes, A.J., Benjamin, M., Ralphs, J.R.,1996. Tendon cells in vivo form a three-dimensional networkof cell processes linked by gap junctions. J Anat. 189,593–600.

Maffulli, N., Waterston, S.W., Squair, J., Reaper, J., Douglas,A.S., 1999. Changing incidence of Achilles tendon rupturein Scotland: a 15-year study. Clin. J. Sport Med. 9, 157–160.

Moller, A., Astron, M., Westlin, N., 1996. Increasing incidenceof Achilles tendon rupture. Acta Orthop. Scand. 67,479–481.

Morgelin, M., Heinegard, D., Engel, J., Paulsson, M., 1992.¨ ˚Electron microscopy of native cartilage oligomeric matrixprotein purified from the Swarm rat chondrosarcoma revealsa five-armed structure. J. Biol. Chem. 267, 6137–6141.

Patterson-Kane, J.C., Wilson, A.M., Firth, E.C., Parry, D.A.D.,Goodship, A.E., 1997. Comparison of collagen fibril popu-lations in the superficial digital flexor tendons of exercisedand non-exercised thoroughbreds. Equine Vet. J. 29,121–125.

Perez-Castro, A.V., Vogel, K.G., 1999. In situ expression ofcollagen and proteoglycan genes during development offibrocartilage in bovine deep flexor tendon. J. Orthop. Res.17, 139–148.

Pickersgill, C., 2000. MVM Thesis. University of Glasgow.Riley, G.P., Harrall, R.L., Constant, C.R., Chard, M.D., Caws-

ton, T.E., Hazleman, B.L., 1994. Glycosaminoglycans ofhuman rotator cuff tendons: changes with ages and inchronic rotator cuff tendinitis. Ann. Rheum. Dis. 53,367–376.

Rosenberg, K., Olsson, H., Morgelin, M., Heinegard, D., 1998.¨ ˚Cartilage oligomeric matrix protein shows high affinity zinc-dependent interaction with triple helical collagen. J. Biol.Chem. 273, 20397–20403.

Smith, R.K., Birch, H., Patterson-Kane, J., et al., 1999. Shouldequine athletes commence training during skeletal develop-ment?: changes in tendon matrix associated with develop-ment, ageing, function and exercise. Equine Vet. J. 31,201–209.

Smith, R.K.W., Gerard, M., Dowling, B., Dart, A.J., Heinegard,˚D., 2000. Cartilage oligomeric matrix protein(COMP)levels in tendon are correlated with mechanical propertiesat skeletal maturity. Proceedings of the 1st Symposium ofthe International Society for Matrix Biology. p. V-9.

Smith, R.K.W., Zunino, L., Webbon, P.M., Heinegard, D.,˚1997. The distribution of cartilage oligomeric matrix protein(COMP) in tendon and its variation with tendon site, age,and load. Matrix Biol. 16, 255–271.

Stephens, P.R., Nunamaker, D.M., Butterweck, D.M., 1989.Application of a Hall-effect transducer for the measurementof tendon strain in horses. Am. J. Vet. Res. 50, 1089–1095.

Stromberg, B., Tufvesson, G., 1969. Lesions of the superficial¨flexor tendons in racehorses. A microangiographic andhistopathologic study. Clin. Orthop. 62, 113–123.

Thur, J., Rosenberg, K., Nitsche, D.P., et al., 2001. Mutationsin cartilage oligomeric matrix protein causing pseudoach-ondroplasia and multiple epiphyseal dysplasia affect bindingof calcium and collagen I, II, and IX. J. Biol. Chem. 276,6083–6092.

van Weeren, P.R., Barneveld, A., 1999. Study design toevaluate the influence of exercise on the development ofthe musculoskeletal system of foals up to age 11 months.Equine Vet. J. 31, 4–8.

Vogel, K.G., Evanko, S.P., 1987. Proteoglycans of fetal bovinetendon. J. Biol. Chem. 262, 13607–13613.

Vogel, K.G., Heinegard, D., 1985. Characterization of proteo-˚glycans from adult bovine tendon. J. Biol. Chem. 260,9298–9306.

Vogel, K.G., Koob, T.J., 1989. Structural specialisation intendons under compression. Int. Rev. Cytol. 115, 267–293.

Vogel, K.G., Sandy, J.D., Pogany, G., Robbins, J.R., 1994.Aggrecan in bovine tendon. Matrix Biol. 14, 171–179.

Webbon, P.M., 1977. Post mortem study of equine digitalflexor tendons. Equine Vet. J. 9, 61–67.

Williams, R.B., Harkins, L.S., Hammond, C.J., Wood, J.L.N.,2001. Racehorse injuries, clinical problems and fatalitiesrecorded on British racecourses from flat racing and NationalHunt racing during 1996, 1997 and 1998. Equine Vet. J. 33,478–486.

Wilson, A.M., Goodship, A.E., 1993. Exercise-induced hyper-thermia as a possible mechanism for tendon degeneration.J. Biomech. 27, 899–905.

Woo, S.L., Gomez, M.A., Woo, Y.K., Akeson, W.H., 1982.Mechanical properties of tendons and ligaments. II. Therelationships of immobilization and exercise on tissueremodeling. Biorheology 19, 397–408.

review the influence of ageing and exercise on tendon...

Documents