Article Text

Download PDFPDF

Greater glycosaminoglycan content in human patellar tendon biopsies is associated with more pain and a lower VISA score
  1. Mohamed Attia1,2,
  2. Alexander Scott3,
  3. Gilles Carpentier1,
  4. Øystein Lian4,
  5. Toin Van Kuppevelt5,
  6. Camille Gossard6,
  7. Dulce Papy-Garcia1,
  8. Marie-Claude Tassoni2,6,
  9. Isabelle Martelly1
  1. 1Laboratoire CRRET, Université Paris-Est Créteil, Créteil, France
  2. 2Département de mécanobiologie, Cogitobio, Cachan, France
  3. 3Department of Physical Therapy, Centre for Hip Health and Mobility, University of British Columbia, Vancouver, British Columbia, Canada
  4. 4Kristiansund Hospital, Kristiansund, Norway
  5. 5The Department of Biochemistry, Radboud University Nijmegen Medical Centre, Nijmegen Centre for Molecular Life Sciences, Nijmegen, The Netherlands
  6. 6Osteobio, Ecole Supérieure d'Ostéopathie et de Biomécanique Appliquée, Cachan, France
  1. Correspondence to Professor Isabelle Martelly, Laboratoire CRRET CNRS EAC 7149, Université Paris-Est Créteil, 61 Av du Général de Gaulle 94010 Créteil Cedex France; martelly{at}


Background People with patellar tendinopathy experience chronic pain and activity limitation, but a pertinent biochemical marker correlated with these clinical features has not been identified. The Victoria Institute of Sport Assessment (VISA) questionnaire is a condition-specific patient-rated outcome measure. Since the quantity of glycosaminoglycans (GAGs) increases with advancing tendon pathology, we hypothesised that there would be a correlation between the quantity of GAGs in the patellar tendon and the VISA score.

Methods Tissue biopsies from athletes with chronic patellar tendinopathy (confirmed by clinical examination and MRI) were recruited (n=7), as well as controls with no history of knee pain (n=4). The quantity of sulphated GAGs in the human patellar tendons was determined with a dimethyl methylene blue (DMMB) assay; this method was first validated with rat tendon tissue. The extent and distribution of GAG species and proteoglycans (decorin, versican and aggrecan) in the human tendon biopsies were examined using immunohistochemistry.

Results Greater sulphated GAG content of the patellar tendon was correlated with the greater tendon dysfunction (R2=0.798). The quantity of aggrecan in the tendon, a chondroitin sulphate-rich proteoglycan, also increased with advancing tendon pathology.

Conclusions Increased GAGs in the pathological human patellar tendon are related to a worse clinical status. These findings indicate that the VISA score reflects the extent of tendon tissue pathology.

  • Orthopaedics
  • Biomechanics
  • Tendons
  • Knee injuries
  • Shoulder injuries

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Chronic patellar tendinopathy remains a significant challenge, with no treatment capable of consistently reversing the profound structural changes that develop in response to excessive and repetitive mechanical loading.1 ,2 The pathology underlying patellar tendinopathy is known as tendinosis, a degenerative condition characterised by an increase in the levels of glycosaminoglycans (GAGs),3 which represent the polysaccharidic moiety of proteoglycans (PGs). This increase begins during the development of tendinosis in the asymptomatic stage, and becomes more prominent and extensive in patients with longstanding symptoms.1 ,4 ,5 The presence of excessive GAGs in tendon, and also the increased presence of PGs such as aggrecan (which is more characteristic of cartilaginous tissue) would be expected to impair the tensile load-bearing function of the tissue. This impairment could contribute to patients’ symptoms and activity limitation. However, to our knowledge the potential relationship between GAGs and clinical status has not yet been examined.

In this report, we hypothesised that the quantity of GAGs in the patellar tendon (Pt) would negatively correlate with the Victorian Institute Sport Assessment (VISA) score, a widely used outcome measure which assesses the functional status and pain levels of patients with patellar tendinopathy.6 The methodology for quantification of GAGs was first validated using rat supraspinatus tendons.


Rat supraspinatus tendon sampling and processing

Control and tendinopathic rat tendons were obtained in order to validate the histological technique used for GAGs assessment. Supraspinatus tendons were removed from a cohort of Sprague-Dawley adult rats (n=12) subjected to 4 weeks of downhill training, as originally described in ref. 7. Training was performed with the approval of the French and Canadian local Ethics Review Board. Three other animals that did not run were used as controls. Tendons of these animals were snap frozen in liquid N2 and stored until use. In addition, histological studies on rat tissues were performed using paraffin blocks from a previous cohort of animals (or their age-matched controls) subjected to the same overuse protocol. Five running and three control rats were studied. We used these pre-existing tissue blocks to reduce the required number of animals in accordance with the principles of the Canadian Council on Animal Care.

Human Pt sampling and processing

Human Pt was obtained from patients with patellar tendinopathy with structural tendon changes confirmed by clinical examination and MRI. Each participant filled out the VISA. The VISA score is a questionnaire-based index of disability and pain specifically designed and validated for use in patients with patellar tendinopathy.6 The control group consisted of patients undergoing intramedullary nailing for tibial fractures with no current or previous knee pain. In total, seven patients (six males and one female) and four control participants (three males and one female) were included in this study (table 1). These patients are a convenience sample from a previously characterised cohort.3 ,8–10 The subset of patients in the current study was selected based on the availability of clinical data and of sufficient tissue for analysis. All patients gave their written informed consent, and the research was performed with the approval of the local Committee for Research Ethics in Norway.

Table 1

Information on patient biopsies and their corresponding Victorian Institute Sport Assessment (VISA) score

Pt biopsies were taken from the proximal osteotendinous junction as described in ref. 3. Tendon biopsies were fixed in Zamboni's solution and washed in 0.1 M phosphate-buffered NaCl, pH 7.2, with phosphate buffered saline (PBS) 15% sucrose (w/v) and 0.1% natrium azide. After fixation, the biopsies were stored in PBS at 4°C, dehydrated, embedded in paraffin and serially sectioned at 7 μm thickness.

GAG extraction and quantification assay

The extraction of GAGs from rat supraspinatus tendon was performed as in ref. 11. The 1,9-dimethyl methylene blue (DMMB) assay was used to quantify total sulphated GAGs. This assay is based on the binding of sulphated GAGs to the cationic DMMB. After formation of the GAG-dye complex, it was dissociated and the optical density of DMMB was measured. Calibration curves, constructed with known amounts of chondroitin sulphate (CS)-A standards, ranging from 0.5 to 4.0 µg, were included in every assay. All data points were acquired in triplicate.

Histological quantification of total sulphated GAG

The DMMB staining technique was derived from the biochemical technique described above. In brief, paraffin sections of 7 μm thickness from human Pts and rat supraspinatus tendons were deparaffinised and incubated overnight with DMMB solution in the dark at room temperature, as previously described.12 The labelling was validated using an enzymatic treatment (Chondroitinase ABC and heparanase, Sigma, Saint-Quentin-Fallavier, France). Digital images were acquired with a CoolSNAP camera (Princeton Instruments; Acton, Massachusetts, USA). GAGs quantification was performed using colour segmentation with ImageJ software13 as described in ref. 12. For each tendon sample, nine microscopic fields were captured on six tissue sections (54 fields per sample in total).


Deparaffinised Pt sections were blocked with PBS-bovine serum albumin 3% (w/v) and incubated overnight with phage display single chain anti-heparan sulphate (HS; clone AO4B08), anti-dermatan sulphate (DS; clone LKN1) and anti-CS; clone IO3H10 tagged with the vesicular stomatitis virus (VSV) protein (antibodies produced by T Van Kuppevelt's laboratory). Bound antibodies were detected with mouse antibody P5D4 raised against the VSV tag. The identity of labelled GAGs was confirmed by elimination of the signal after specific digestion with chondroitinase ABC, chondroitinase B (Sigma, Saint-Quentin-Fallavier, France) or heparinase (Iduron Ltd, Manchester, UK). PGs were detected using anti-decorin (1/100, from Larry Fisher PhD, NIH, Bethesda), anti-versican (1/100, Santa-Cruz Biotechnology, Heidelberg, Germany) and anti-aggrecan (1/100, Santa-Cruz Biotechnology) antibodies. Bound antibodies were detected with the appropriate Alexa-conjugated IgG (Fluoprobes Interchim, Montluçon, France). Nuclei were labelled using 4′,6-diamidino-2-phenylindole (DAPI). Fluorescence images were obtained using a CCD monochrome camera (CFW-1310M; Scion; Frederick, MD) fitted to a BH-2 epifluorescence microscope (Olympus; Rungis, France).

Statistical analysis

Data are expressed as mean±SD and were statistically compared by one-way analysis of variance followed by Bonferroni's multiple comparison tests using GraphPad software (San Diego, California, USA). Statistical differences were determined at p<0.05 as minimum level of significance; p<0.001 (***) and p<0.01 (**).


Validation of GAG quantification—rat tendons

In this study, we first used rat tendon samples to validate the assessment of GAGs using a computerised histological method, with reference to the standard evaluation of GAGs using a quantitative biochemical technique. The rat overuse training has been previously shown to induce tendinosis-like changes in supraspinatus tendons from asymptomatic animals, especially a progressive increase in GAGs over time.12 As expected, the histochemical staining with the DMMB showed that GAG staining was increased in pathological supraspinatus rat tendons compared to control (figure 1a). The GAG quantification by image analysis (figure 1b) or by biochemical method (figure 1c) both demonstrated a similar magnitude of change in tendinopathic tendons compared to controls (up to 2.5-fold increase in GAG levels).

Figure 1

Sulphated glycosaminoglycan (GAG) expression in healthy and pathological rat supraspinatus tendons (SSts) (a) GAG staining of control and overused tendons at 4 weeks with dimethyl methylene blue (DMMB) assay. (b) GAG amount assessed by image analysis of DMMB labelling in control (C) and overused tendons (R). (c) Sulphated GAG amount assessed biochemically in control (C) and overused (R) rat supraspinatus tendons. Values are mean±SD of nine microscopic fields from six different sections of tendons from three different rats; p<0.001 (***) and p<0.01 (**). Scale bar 50 μm.

Correlation of sulphated GAGs with VISA score

All tendinopathic human samples exhibited a stronger GAG labelling than healthy tendons (figure 2A) and the increase in GAG quantity ranged from 3 to 4 times the level found in non-pathological Pts (figure 2B). There was a strong negative correlation (R2=0.798) between the amount of sulphated GAGs in human Pt and the VISA score (figure 2C).

Figure 2

Sulphated glycosaminoglycan (GAG) expression and quantification in human patellar tendons (Pts) according to the severity of tendinopathy determined by the Victorian Institute Sport Assessment (VISA) score. (A) Representative images of human Pts from patients with different VISA scores where sulphated GAGs have been labelled by dimethyl methylene blue (DMMB) assay. Scale bar 100 μm. (B) GAGs amount assessed by image analysis in human Pts of the different participants. (C) GAGs amount in relation to VISA score. For each sample, nine microscopic fields of six different sections were analysed. Values are mean±SD.

Distribution of PGs and GAG species in human Pt biopsies

We next examined the extent of PGs typically associated with GAGs in human Pts. Decorin, versican and aggrecan were all clearly increased in patellar tendinopathy samples compared to controls (figure 3). In particular, the expression of aggrecan in Pts increased progressively with worsening VISA score (figure 4). Finally, the molecular types of sulphated GAGs in control and tendinopathic tendons were identified by immunohistochemistry (figure 5). Whereas the amount of HS did not seem to vary between groups, the amounts of DS and CS were clearly increased in pathological Pts. DS and CS were prominent in the immediate cellular environment of tenocytes in healthy Pts, but were more extensively present throughout pathological tendons.

Figure 3

Immunohistological staining of decorin, versican and aggrecan in human healthy and pathological patellar tendons. Nuclei were stained with 4',6-diamidino-2-phenylindole (DAPI). Scale bar 25 μm.

Figure 4

Aggrecan expression in human patellar tendons according to the severity of the tendinopathy determined by the Victorian Institute Sport Assessment (VISA) score. Aggrecan labelling was mostly pericellular in healthy tendon, but in pathological tendons it frequently appeared as a disorganised network between tenocytes arrays. Nuclei were stained with DAPI. Scale bar 50 μm.

Figure 5

Glycosaminoglycan (GAG) species expression (DS=dermatan sulphate, CS=chondroitin sulphate and HS=heparan sulphate) in human healthy and pathological patellar tendons using phage-display antibodies. Nuclei were stained with DAPI. The specificity of the antibodies was verified by enzymatic or chemical treatments as in ref. 12. Scale bar 25 μm.


An increased level of GAGs in pathological tendons has previously been described in humans4 ,5 and animals.12 ,14 In this study, we detected a linear relationship between the quantity of GAGs in the human Pt biopsies and the clinical status (VISA score). In particular, the PG aggrecan was highly increased in biopsies of pathological tendons and was accompanied by an increase in CS. The parallel increase of aggrecan and CS is expected, given that this PG carries high amount of CS species as its GAG moiety. The enrichment of aggrecan and CS may reflect chondrocyte metaplasia, a pathological process known to occur in patellar tendinopathy.3 ,15 Other PGs such as versican and decorin were also increased in altered Pts; however, these components are also characteristic of healthy tendons and cannot be taken as specific indicators of pathology per se.

The VISA score quantifies pain with tendon loading, and activity limitation due to tendon pain. The mechanisms of pain in tendinopathy remain under investigation, and several diffusible molecules present in chronically painful tendon have been shown to promote the sensitisation of nociceptive afferents.16–19 It is tempting to speculate that, despite their high molecular weight, aggrecan with its bound CS chains (more than 100 chains/molecule)20 may diffuse and reach nociceptive targets, and the progressive disruption of tendon structure could promote this diffusion. Moreover, since significant degradation of PGs has been shown in patellar tendinopathy notably through proteases,21 including aggrecanase activities,2 ,5 aggrecan derived peptides with their CS chains could also be involved in nociceptive activation or sensitisation due to their high negative charge; this charge would tend to attract cations, thus modifying the environment of nociceptors.1 Alternatively, increased hydration due to accumulation of GAGs and PG breakdown products in pathological Pt may raise the internal pressure, as in cartilage.22 This hypothesis is in line with recent findings suggesting that CS is involved in mechanical hyperalgesia,23 which is a common finding in patients with tendinopathy.24

We speculate that an elevated water content in the pathological Pt is responsible for the typical increase of MRI signal seen in this condition.1 ,2,5 This increased water content could result from the local GAG enrichment, as observed in the current study. In a recent study, an increase in intratendinous MR signal was associated with Achilles tendon pain and functional impairment, independently of tendon volume.2,6 Thus, the correlation between the amount of GAGs and the extent of tendon pain/disability suggested here is in keeping with the previously demonstrated relation between increased MR signal and tendon pain.2,6

There are some limitations in this study. First, the measurement of GAG amounts based on histological staining could be considered semiquantitative; however, the method was validated by obtaining parallel GAG measurements using a more quantitative biochemical technique performed on animal tendons. The number of samples in this study is rather small and it would be interesting in future to broaden the study by including data from patients with less severe pathology (eg, VISA scores higher than 53). The analysis was performed on the proximal portion of the Pt. Many patients do also develop pathological changes in the middle and distal regions of the Pt, and it is possible that the changes observed in this study do not reflect the status of the entire tendon.

Nevertheless, this study demonstrates that GAGs, along with aggrecan content, increase linearly with pain/disability assessed by VISA score in human patellar tendinopathy. GAGs, in particular CS, could be taken as potential markers of tendon pathology and pain. If GAGs could be detected either in the circulation, or within the tissue using GAG-enhanced imaging methods such as gadolinium-enhanced MRI, perhaps patients with early asymptomatic tendinosis could be identified and measures could be taken to prevent the development of patellar tendinopathy.

What are the new findings?

  • This study demonstrates that glycosaminoglycans (GAGs), particularly those associated with aggrecan, increase linearly with advancing severity of patellar tendinopathy.

  • The results underscore the occurrence of excessive GAGs, and their relation to clinical status, in the pathological patellar tendon.

How might it impact on clinical practice in the near future?

  • The findings provide biochemical evidence that the Victorian Institute Sport Assessment (VISA) score represents a pertinent indicator of patellar tendon pathology.

  • Additionally, non-invasive methods for glycosaminoglycans detection such as gadolinium-enhanced MRI may be useful in quantifying patellar tendon pathology.


The authors would like to thank Eric Huet, Suzanne Menashi, Arlette Duchesnay, Stéphanie Thierart, Sarah Etienne and Jean Delbé for their valuable discussions and help.



  • M-CT and IM contributed equally.

  • Funding Cogitobio, Association Nationale de Recherche et Technologie (ANRT), Michael Smith Foundation for Health Research.

  • Competing interests MA received financial support from Cogitobio and ANRT. AS received support from the Michael Smith Foundation for Health Research.

  • Patient consent Obtained.

  • Ethics approval The local Committees for Research Ethics in Norway and Canada.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles