Skip Navigation


Rheumatology Advance Access originally published online on March 27, 2006
Rheumatology 2006 45(9):1101-1109; doi:10.1093/rheumatology/kel060
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary data
Right arrow All Versions of this Article:
45/9/1101    most recent
kel060v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (15)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Catterall, J. B.
Right arrow Articles by Cawston, T. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Catterall, J. B.
Right arrow Articles by Cawston, T. E.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2006. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Development of a novel 2D proteomics approach for the identification of proteins secreted by primary chondrocytes after stimulation by IL-1 and oncostatin M

J. B. Catterall, A. D. Rowan, S. Sarsfield1, J. Saklatvala1, R. Wait1 and T. E. Cawston

Musculoskeletal Research Group, School of Clinical Medical Sciences, The Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH and 1Kennedy Institute, Imperial College, London, UK.

Correspondence to: T. E. Cawston, Musculoskeletal Research Group, School of Clinical Medical Sciences, The Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH, UK. E-mail: T.E.Cawston{at}ncl.ac.uk


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 References
 
Objectives. To develop a proteomics approach to study changes in the secreted protein levels of primary human chondrocytes after stimulation by the pro-inflammatory cytokines interleukin-1 and oncostatin M.

Methods. Using both the primary human articular and bovine nasal chondrocyte-conditioned mediums, methods were investigated to enable the separation of proteins by two-dimensional (2D) gel electrophoresis. Differentially regulated proteins were identified using tandem electrospray mass spectrometery.

Results. We discovered that proteoglycans and glycosylaminoglycans (GAGs) secreted by chondrocytes significantly interfered with 2D gel focusing. Several different methods for GAG removal were attempted including enzymic digestion, cetyl pyridinium chloride precipitation and anion exchange in high salt. The anion exchange proved to be the most effective. Even from these initial gels, we were able to identify eight proteins produced by human chondrocytes: matrix metalloproteinase (MMP)-1, MMP-3, YKL40, cyclophilin A, ß2-microglobulin, transthyretin, S100A11, peroxidine 1 and cofilin. MMP-1, MMP-3, YKL40 and cyclophilin A were all identified as processed, smaller peptide fragments.

Conclusions. We were able to develop a novel sample preparation protocol to allow the reproducible sample preparation of secreted proteins from human chondrocytes. From the initial data, we were able to show that at least some of the proteins produced were cleaved to smaller fragments as a result of proteolysis. Therefore, this technique provides valuable information about protein processing which gene-based arrays do not.

KEY WORDS: Chondrocyte, IL-1, MMP, OSM, Proteomics.


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 References
 
The cartilage tissue is composed predominately of an extracellular matrix (ECM), where proteoglycan provides the resistance to compression whilst collagen gives the cartilage its tensile strength. Embedded within this ECM are chondrocytes that regulate the turnover and remodelling of the cartilage matrix [1]. In normal adult cartilage, a steady state exists in which the turnover of the matrix molecules is in equilibrium. Irreversible degradation of the cartilage matrix is a major feature in both osteoarthritis (OA) and rheumatoid arthritis (RA), leading to a loss of joint function. Cartilage model systems respond to the pro-inflammatory cytokine interleukin-1 (IL-1) with a rapid release of proteoglycan from the tissue [2–5], although this can be resynthesized quickly [4] if the stimulus is removed. In contrast, collagen is much less readily broken down and hence represents the irreversible step in cartilage degradation [5]. Thus, collagen degradation is believed to be a key control point for cartilage degradation.

The matrix metalloproteinases (MMPs) are a family of zinc-dependent endopeptidases that collectively cleave most, if not all, of the components of the ECM. They are involved in the normal turnover of connective tissue matrix that takes place during growth and development. They are also implicated in the pathological destruction of tissues in a variety of diseases including tumour invasion, metastasis [6], and cartilage and bone destruction in the arthritides [7, 8]. The MMPs can be divided either into groups that differ in structure or in substrate specificity such as collagenases, stromelysins, gelatinases and membrane-type MMPs (MT-MMPs). There are three main human collagenases: MMP-1 (interstitial collagenase), MMP-8 (neutrophil collagenase) and MMP-13 (collagenase-3 [9–11]). These enzymes specifically cleave fibrillar collagen into 3/4 and 1/4 fragments. In addition, gelatinase A (MMP-2) and MT1-MMP (MMP-14) have also been shown to cleave type I collagen [12, 13]. Degradation of the collagenous network is excessive in human arthritis [14], and breakdown of type II collagen, the main type of collagen in articular cartilage, correlates with MMP activity in cartilage explant cultures [15]. Raised levels of collagenases have been localized within diseased cartilage from rheumatic diseases [16] and in rheumatoid synovial fluid [17]. Collagenases are also present at the cartilage–pannus junction and in synovial tissue from RA and OA joints [18–20]. Collagenases are therefore key enzymes involved in collagen turnover [21, 22].

The regulation of MMP activity is tightly controlled at three levels: synthesis, pro-enzyme activation and inhibition of activated enzymes. All MMPs are synthesized as zymogens that must undergo cleavage to produce the active enzyme. MMPs can be activated in vitro by proteinases and by non-proteolytic compounds such as thiol agents, organomercurials, reactive oxygen species and by heat treatment [23]. Once activated, MMPs are subject to further control by a group of inhibitors known as tissue inhibitors of metalloproteinases (TIMPs) [24]. The production of MMPs and inhibitors is mediated by a variety of cytokines and growth factors including IL-1, oncostatin M (OSM) and tumour necrosis factor {alpha} (TNF-{alpha}) [25, 26].

Many different pro-inflammatory cytokines have been demonstrated to be up-regulated in RA, including TNF-{alpha}, IL-1, OSM and IL-17, while anti-inflammatory cytokines such as IL-4 are virtually absent. In model systems such as the bovine nasal cartilage assay, these pro-inflammatory cytokines have been shown to stimulate the destruction of cartilage in explant cultures while the addition of anti-inflammatory cytokines can, in some cases, inhibit this destruction. Cytokines, however, are never found in isolation within an RA joint, and the effects of one cytokine can modulate the effect of others. This is demonstrated by IL-1 and OSM: alone IL-1 is capable of causing the degradation of cartilage in the bovine nasal system; however, if OSM is added there is a synergistic increase in the amount of destruction observed. This was a novel finding as OSM alone increases the expression of TIMP-1, the natural inhibitor of MMPs that are involved in cartilage destruction during rheumatoid disease. This cytokine combination was also shown to synergistically regulate the expression of the collagenase MMP-1 in chondrocytes and synovial fibroblasts, and has been strongly linked to cartilage destruction in RA [27]. A later work [28] demonstrated that mRNA for a number of MMPs was rapidly up-regulated in response to this cytokine combination. The destruction of cartilage, however, is unlikely to be caused by the up-regulation of a single enzyme. It is known that the removal of proteoglycan, which always precedes the removal of collagen, is due to a family of proteinases belonging to a disintegrin and metalloprotinease with thrombospondin motifs (ADAMTS) family [29]. These proteins are associated with the surface of the cell and are responsible for initiating the dismantling of the ECM. The localization of proteins to specific areas of the cell surface is likely to involve a large number of different proteinases, inhibitors and assessory proteins. The up-regulation and/or down-regulation of such proteins leads to a shift in the balance between active proteinases and inhibitors at the interface of the cell and the matrix to promote cartilage destruction. It is obvious that the discovery of the precise sequence of events that leads to these imbalances will give new insight into the mechanism of action and could identify new therapeutic targets.

It is clear that to try and understand the interplay of the different proteins produced by chondrocytes after stimulation with disease-causing cytokines, a method needed to be developed to study all the proteins produced. In this article, we have developed methodologies to look at the proteins produced by primary chondrocytes in monolayer culture after stimulation with combinations of the pro-inflammatory cytokines IL-1 and OSM. This ability to identify changes in protein profiles after pro-inflammatory cytokine stimulation should enable a better understanding of some of the underlying mechanisms in cartilage breakdown [30].


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 References
 
This study was approved by the Newcastle and North Tyneside Local Research Ethics Committee.

Reagents
Chemicals were obtained from the following suppliers: IL-1{alpha} was a generous gift from Glaxo Smith Kline (Stevenage, UK), OSM [31] from Prof. J Heath (University of Birmingham, Birmingham, UK). All other reagents were commercially available analytical grade or better, obtained from Fisher Scientific UK (Loughborough, UK), BDH Chemicals (Poole, UK), Sigma-Aldrich (Poole, UK) and Amersham Bioscience (Little Chalfont, Buckinghamshire, UK).

Tissue culture
Primary human articular chondrocytes (HACs) were extracted from human cartilage obtained after joint replacement surgery. Four samples were obtained from both sexes with established OA and ages which ranged from 71–85 years. The bovine nasal chondrocytes where obtained from the septum cartilage of freshly slaughtered cows. Briefly, cartilage was removed from the joints, and sequentially digested with hyaluronidase [1 mg/ml in phosphate buffered saline (PBS), type IV bovine testicular], trypsin (0.25% w/v in PBS) and finally, bacterial collagenase (1 mg/ml). Cells were resuspended and seeded into T75 flasks in Dulbecco's modification of Eagle's medium (DMEM) supplemented with 10% bovine fetal calf serum, 2 mM glutamine, 200 IU/ml of penicillin, 200 µg/ml of streptomycin, 40 units/ml nystatin and 50 µg/ml gentomycin.

Cells were grown to pre-confluence before serum starving for 24 h in medium minus bovine fetal calf serum. The medium was replaced and cells stimulated for 24 h with IL-1 (1 ng/ml) and OSM (10 ng/ml) combinations. For the final 4 h, the medium was again replaced but with DMEM without cysteine and methionine. This medium was further supplemented with 5.3 MBq of 35S-labelled cysteine and methionine (ProMix, Amersham Bioscience) to label the newly synthesized proteins and cytokines. At harvest, all proteolytic activity within the samples was inhibited using proteinase inhibitors (complete inhibitor tablets, Roche), and any floating cells or debris was removed by centirfugation.

Cetyl pyridinium chloride (CPC) precipitation of GAG
Medium samples were collected after the 24 h stimulation. The levels of glycosylaminoglycan (GAG) were measured using dimethyl-methylene blue (DMB) binding [32]. The GAGs present were precipitated at RT for 30 min using 3 mg of CPC for every milligram of GAG. Precipitated GAGs were pelleted by centrifugation at 4°C, 35 000 g for 10 min. The medium was collected and the pellet washed briefly with 100 nM NaCl, both the wash and the medium were combined. The proteins contained within the medium were precipitated overnight at 4°C using an 80% saturated solution of ammonium sulphate and pelleted by centrifugation at 4°C, 35 000 g for 30 min. Protein pellets were resuspended in ddH2O and dialysed extensively against 100 mM ammonium bicarbonate at 4°C before freeze drying. Comparison samples were treated identically but with the omission of the CPC step.

Sample preparation using enzymes
Conditioned medium was treated with either chondrotinase ABC (1 µg/ml, 0.001 units/ml) or hylauronidase (1 µg/ml, 0.36 units/ml, type IV bovine testicular) at 37°C for up to 24 h. To prevent further proteolysis of the samples, proteinase inhibitors were added (complete inhibitor tablets, Roche). Proteins were collected by 80% saturated solution of ammonium sulphate precipitation followed by extensive dialysis against 100 mM ammonium bicarbonate and concentration by freeze drying.

Sample preparation using Q sepharose
Conditioned medium was rapidly dialysed into 50 mM sodium cacodylate buffer, pH 7.5, supplemented with 0.05% (w/v) sodium azide and 5 mM CaCl2 at 4°C with multiple changes. Dialysed medium was adjusted to 1 M NaCl before passing through a Q sepharose resin batch column (1 ml resin per 10 ml of medium). Proteins in the flow-through were precipitated using 10% trichloro-acetic acid (TCA) and resuspended in 100 mM ammonium bicarbonate. Samples were desalted by dialysis against ammonium bicarbonate before concentration by freeze drying. The GAGs, as determined by DMB assay, were retained on the column.

Two-dimensional (2D) gel electrophoresis
The first-dimension isoelectric focusing (IEF) was performed using the Amersham Bioscience recommended conditions. Briefly, freeze-dried samples were resuspended in 450 µl of rehydration buffer (8M urea, 2% w/v CHAPS, 2.8 mg/ml DTT, 2% v/v IPG buffers). Protein concentration was estimated using the Bradford assay (Sigma-Aldrich) and volumes corrected so that equal amounts of protein were added per gel. Samples were rehydrated directly into 24 cm immobiline pH gradient strips (Amersham Bioscience) overnight at room temperature. The proteins were focused for 28 h on a Multiphor system (Amersham Bioscience) using the following conditions: 150 V for 1 h, 300 V for 3 h, 3500 V for 24 h at a constant 20°C.

For the second-dimension SDS–PAGE, focused immobiline strips were first equilibrated into SDS–PAGE sample buffer by incubating in 20 ml of equilibration buffer [50 mM Tris–HCl pH 8.8, 6 M urea, 30% (v/v) glycerol, 2% (w/v) SDS, trace bromophenol blues] with 10 mg/ml DTT for 10 min. Strips were further incubated for another 10 min in 20 ml of equilibration buffer containing 25 mg/ml iodoacetamide before sealing onto the top of a 12.5% SDS–PAGE gel using agarose sealing solution [1% (w/v) low melting point agarose, 25 mM Tris–HCl, 192 mM glycine, 0.1% (w/v) SDS, trace bromophenol blue]. Up to 12 gels at a time were run overnight at 10°C and 2 W per gel on an Ettan Dalt II 12 gel system (Amersham Bioscience).

Gels were stained with a mass spectrometry compatible silver stain [33]. Briefly, gels were fixed for at least 30 min in fixer [50% (v/v) methanol, 10% (v/v) acetic acid] before washing for 10 min in 50% (v/v) methanol. Gels were rehydrated with a further wash for 10 min in ddH2O before a 2 min sensitizing step in 0.02% (w/v) sodium thiosulphate. Gels were given two further 1 min washes in water before incubation at 4°C in the dark in 0.1% (w/v) silver nitrate solution for at least 40 min. Excess silver nitrate was removed with two brief washes with ddH2O before the gels were developed [0.04% formalin in 2% (w/v) sodium carbonate] for 2–5 min. The developing reaction was stopped using 10% (w/v) acetic acid, and the gels were stored in 1% (w/v) acetic acid until dried. Before drying, gels were incubated in a drying solution [30% (v/v) ethanol, 5% (v/v) glycerol] for 20 min to prevent cracking and aid subsequent rehydration.

Gel analysis
Silver-stained images were captured using a flat bed scanner (Hewlett Packard) and the 35S-labelled images were captured using phosphor-screens on a Storm 860 phosphorimager (Amersham Biosciences). Differential protein expression was compared between gels using the Progenesis software (Nonlinear Dynamics, Newcastle upon Tyne). Protein spots of interest were manually excised ready for identification by mass spectrometry.

Mass spectrometry
In-gel digestions with trypsin were performed as described [34] and samples were analysed by tandem electrospray mass spectromtery (ESI MS/MS). Spectra were recorded using a Q-Tof spectrometer (Micromass, Manchester, UK) interfaced to a Micromass CapLC capillary chromatograph. Samples were dissolved in 0.1% formic acid, and aliquots were injected onto a 300 µm x 15 mm Pepmap C18 column (LC Packings, Amsterdam, NL) and eluted with an acetonitrile/0.1% formic acid gradient. Capillary voltage was 3500 V and data-dependent MS/MS acquisitions were performed on precursors with charge states of 2, 3 or 4 over a survey mass range 540–1000. The collision gas was argon, and the collision voltage was 18–45 V depending on the precursor charge state and mass. Proteins were identified by correlation of uninterpreted spectra to entries in SwissProt and TREMBL using ProteinLynx Global Server (v 1.0, Micromass).


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 References
 
Separation of stimulated primary HAC medium
Primary HACs were stimulated with IL-1, OSM and IL-1 + OSM combinations, and the secreted proteins precipitated and dialysed before separation using 2D gel electrophoresis (Fig. 1). A high percentage gel (15%) which gives resolution down to 10 kDa was used to ensure that all protein species present were represented and that smaller molecular weight proteins were not lost. Only the lower molecular weight proteins, below 36 kDa, separated well into distinct protein spots (Fig. 1). Above 36 kDa, the proteins smeared, making it difficult to identify distinct protein spots or perform reliable analysis on changes in protein expression in this region. Proteins that were present in chondrocytes after all treatments were removed from the lower regions of the gels to test the reliability of the technique.


Figure 1
View larger version (123K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 1. Initial 2D gels from primary HAC stimulated with pro-inflammatory cytokines. Conditioned medium was prepared by stimulating primary HAC with cytokine combinations for 24 h at 37°C. Proteins were prepared for 2D by ammonium sulphate precipitation, dialysis and freeze drying before resuspension in 2D sample buffer. Proteins were first separated according to their isoelectric point on 24 cm 3–10 NL strips using 300 µg of protein per strip for 28 h at 20°C. The second dimension was performed on 15% SDS–PAGE gels run O/N at 2W per gel at 10°C before visualization using silver stain.

 
Identified proteins
Although the higher molecular weight regions of the gels (shown in Fig. 1) were poorly resolved, we identified eight protein spots in the lower molecular weight region of the gel (Table 1). We were able to identify more proteins on these initial gels but the eight listed in Table 1 were found on gels from two separate patients (data not shown). Four of the proteins were cytoplasmic; two of which S100A11 (calgazzarin) and cofilin were present constitutively while both cyclophilin A (peptidyl-proyl cis–trans isomerase A) and peroxirdoxin 1 appeared to be up-regulated by cytokine stimulation. In the case of cyclophilin A, we could only identify it on the IL-1 and OSM gels but could not exclude its up-regulation by IL-1 and OSM combinations as the resolution in this region was poor. To exclude cytotoxicity of the cytokines as an explanation for the secretion of intracellular proteins, we performed FACs looking for apoptosis using annexin V and necrosis using propidium iodide incorporation. We were able to show very low levels of cell death and this level was unaffected by any of the cytokine treatments (data not shown).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Proteins identified from initial screen of chondrocytes medium

 
The four remaining proteins, YKL40 (chitinase 3 like 1), ß2-microglobulin, MMP-1 and MMP-3 were all secreted proteins. Of these secreted proteins, only MMP-3 showed clear cytokine up-regulation as it was present in all the cytokine-stimulated gels but absent in the control gels. MMP-1 was only found on the IL-1 gel but could not be ruled out from any of the other gels as it was found in the poorly focusing regions. The remaining secreted proteins, YKL40 and ß2-microglobulin, were found on all gels suggesting constitutive expression.

We determined the molecular weights of the proteins and compared these with the expected molecular weights from the SWISS-PRO database. Only four of the proteins: ß2-microglobulin, S100A11, cofilin and peroxirdoxin 1 were found to migrate with the correct molecular weight as estimated by protein markers. The remaining five proteins all ran with an apparent molecular weight that was less than expected, and so were presumably breakdown products. IL-1 and OSM medium is known to contain active proteolytic activity [27].

Development of a protocol to separate higher molecular weight proteins
Although we were able to identify a significant number of lower molecular weight proteins of interest from the initial gels, it was evident that many of the higher molecular weight proteins present within the sample were not identifiable. Furthermore, when 35S incorporation was examined (Fig. 2), the majority of the incorporation was in the higher molecular weight regions of the gel. A new sample preparation protocol was required before a more detailed analysis of the protein expression profile for stimulated human chondrocytes could be performed.


Figure 2
View larger version (129K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 2. Identification of newly synthesized proteins by 35S incorporation. Conditioned medium was prepared by stimulating primary HAC with IL-1/OSM cytokine combinations for 24 h at 37°C. For the final 4 h of culture, the cells were pulsed with 35S-labelled Met/Cyt to label up newly synthesized proteins. Proteins were prepared for 2D by ammonium sulphate precipitation, dialysis and freeze drying before resuspension in 2D sample buffer. Proteins were first separated according to their isoelectric point on 24 cm 3–10 NL strips using 300 µg of protein per strip for 28 h at 20°C. The second dimension was performed on 12.5% SDS–PAGE gels run O/N at 2W per gel at 10°C. The gel was visualized using: (A) silver stain and (B) autoradiography.

 
Removal of proteoglycans
Chondrocytes secrete both hyaluronic acid (HA) and large aggregating proteoglycans such as aggrecan. Proteoglycans can interfere in both the first and second dimensions of 2D runs leading to poor reproducibility. Levels of sulphated GAGs (not HA) were measured using the DMB assay and found to vary between 3 and 30 µg/ml in different samples, which would be equivalent to 30–300 µg per IEF of sulphated GAG. We evaluated three different methods of GAG removal, namely, precipitation, enzymic digestion and anion-exchange chromatography, for their effects upon 2D gel resolution.

Enzymic digestion
The effect of treating the samples with either hyaluronidase or chondrotinase ABC was investigated. Chondrotinase ABC degrades all forms of chondroitin sulphate but only slowly degrades HA, while hyaluronidase degrades HA and chondroitin sulphate A/C but not B. Both enzymes digest the GAG chains to disaccharide units that are lost in the subsequent dialysis step. Treatments were performed in the presence of proteinase inhibitors to prevent any degradation either from contaminating proteinases or from endogenous proteinases within the samples. The chondroitinase treatment did not improve resolution (data not shown) while hyaluronidase treatment (Fig. 3B) was found to improve the definition of the protein spots but horizontal streaking was still present due to poor IEF. Increasing the incubation time from 4 h to overnight had no noticeable improvements in either treatment (data not shown).


Figure 3
View larger version (95K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 3. Comparison of hyaluronidase treatment with CPC precipitation for the removal of GAG before 2D separation. Pooled conditioned medium was prepared by stimulating primary HAC with IL-1/OSM cytokine combinations for 24 h at 37°C. Proteins were prepared for 2D by treating as follows: (A) left untreated, (B) treated with hyaluronidase for 2 h at 37°C, (C) treated with CPC, (D) treated with hyaluronidase for 2 h at 37°C before CPC precipitation. Proteins were collected by ammonium sulphate precipitation, dialysis and freeze drying before resuspension in 2D sample buffer. 100 µg of protein per strip were first separated according to their isoelectric point on 24 cm 3–10 NL strips for 28 h at 20°C. The second dimension was performed on 12.5% SDS-PAGE gels run O/N at 2W per gel at 10°C before visualization with silver stain.

 
Precipitation by CPC
Precipitation of GAGs using the detergent CPC was investigated as a potential alternative to enzymic treatment. Precipitation of proteoglycans could lead to the loss of proteins that interact with the GAG side-chains, many of which are of potential significance to arthritis biology such as the ADAMTS enzymes and cytokines.

Treatment of the samples with CPC led to a reduction in the smearing of proteins across the gel (Fig. 3C). Precipitation with CPC led to the loss of up to 40% of the total starting protein levels as determined by Bradford protein assay, although all the gels in Fig. 3 had an equal protein load. The focusing of samples which contained no detectable GAG was unaffected by CPC treatment and a single precipitation step was sufficient to reduce GAG levels to below detection level. Increasing the amount of CPC added or repeated precipitation appeared to produce no further improvements (data not shown).

Combined enzymic digestion and CPC precipitation
Neither precipitation nor enzymic treatment alone led to resolution in the higher molecular weight regions of the gels. We investigated whether combining both treatments gave a significant improvement. When a 4 h treatment with hyaluronidase followed by CPC precipitation was performed, a significant improvement in resolution was observed (Fig. 3D). Treating with CPC first and then with hyaluronidase had no noticeable effect, presumably the CPC interfered with the hyaluronidase (data not shown). These results illustrate the importance of removing charged GAGs from the samples before attempts are made to focus protein samples produced by chondrocytes.

Anionic exchange chromatography
As an alternative to the combined method outlined previously, we exploited the highly negatively charged nature of GAG side-chains using anion exchange. The binding characteristics of two common anionic exchange resins, Q sepharose and DEAE sepharose, were determined using bovine nasal chondrocyte-conditioned medium (Fig. 4). The medium was dialysed into cacodylate buffer pH 7.5 and batch eluted with increasing concentrations of NaCl, 0–3 M. With the weaker DEAE resin, the protein eluted from the column at 1 M salt while the sulphated GAG began to elute at 2 M salt (Fig. 4A). For the stronger Q sepharose resin (Fig. 4B), the proteins had finished eluting in 1–2 M salt while even at 3 M, the sulphated GAG remained bound to the column matrix and this column was selected for further work. Conditioned BNC medium was dialysed into cacodylate buffer in 1 M salt and passed through a Q sepharose batch column. The flow was collected, prepared and separated by 2D (Fig. 4C and 4D). This method led to improved focusing in the higher molecular weight regions of the gel. When this protocol was repeated on human cells, similar improvements in protein separation were observed (data not shown).


Figure 4
View larger version (70K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 4. Anion-exchange pre-treatment of chondrocyte-conditioned medium. Pooled IL-1/OSM-conditioned bovine nasal chondrocyte medium was dialysed into sodium cacodylate buffer, pH 7.5. Dialysed medium was passed through either: (A) DEAE column or (B) Q sepharose anion-exchange resins. Columns were washed in a step form using increasing salt concentrations 0–3 M in cacodylate buffer and the eluent collected. Protein content was determined using the Bradford protein assay and sulphated GAG content using the DMB assay. For (C) and (D), conditioned medium was dialysed into cacodylate buffer pH 7.5 before making up to 1 M salt. GAG was either retained (C) or removed (D) by passing through a Q sepharose column. Proteins were precipitated using 10% TCA before dialysis into 100 mM ammonium bicarbonate before concentration by freeze drying. First-dimension IEF was performed using 500 µg of protein per 24 cm 3–10 NL strips for 28 h at 20°C. The second-dimension separation was performed on 12.5% SDS–PAGE gels run O/N at 2W per gel at 10°C. The gels were visualized using silver stain.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
References
 
RA is a complex inflammatory disease where many different cytokines are found to be elevated within the joint. In our previous work, we have identified two cytokines, IL-1 and OSM [25], which are both found at elevated levels within the synovial fluid of RA and which act in a synergistic manner in vitro and in an animal model leading to enhanced cartilage breakdown [25, 27]. Chondrocyte cells are the only cell type found within cartilage, and under normal pathology are responsible for the maintenance of the cartilage structure. However, during inflammatory disease states such as RA, pro-inflammatory cytokines act upon chondrocyte cells to produce enzymes that lead to the eventual destruction of the cartilage and a resulting loss in joint function. While many enzymes including the MMPs and the ADAMTS family have been implicated in this destruction, a better understanding of the disease process could be achieved if a more complete understanding of the changes in protein expression could be determined. In this study, we have developed a novel 2D approach which has the advantage of looking directly at the proteins produced rather than a gene-array-based study that only looks at mRNA expression.

Even though our initial simple sample preparation protocol gave very poor protein separation in the upper regions of the gel, we were still able to identify some proteins of interest (Table 1). Two of the proteins identified, MMP-1 and MMP-3, are both proteinases that have already been strongly linked to arthritic destruction [17, 18]. They are known to be produced by chondrocytes after stimulation by IL-1/OSM [28], and our previous work has confirmed these findings at both mRNA and protein levels [27, 28, 35]. MMP-1, or collagense 1, is an important candidate enzyme for collagen breakdown during arthritic disease, while MMP-3 has also been closely linked to arthritic disease and may play a role in activating other proteinases [35]. We have already demonstrated regulation of both MMP-1 and MMP-3 in HAC by inflammatory cytokines and so were expecting these enzymes to be present [28]. MMP-3 is known to be secreted into the medium in both the pro and active forms. The data suggest that these enzymes can be further cleaved when secreted. Another protein identified that may be of some importance during disease state was YKL40, which is a member of the glycosyl hydrolase family of enzymes but has no apparent enzyme activity. It is an extracellular protein that is not synthesized by healthy chondrocytes in vivo, but is a major secreted protein of cultured chondrocytes, and there is some evidence to suggest that it is present in both RA cartilage [36] and in increased levels in synovial fluids [37]. A role as a growth factor has been postulated for YKL40 as it can stimulate chondrocyte cell division and activate several different signalling pathways [38] and may even be able to down-regulate some of the catabolic effects of the inflammatory cytokines IL-1 and TNF-{alpha} [38]. We identified this protein on all gels looked at so far not only in the native form but also as many different breakdown products suggesting that this protein is actively turned over during culture. Both YKL40 and MMP-3 showed multiple lower molecular weight protein spots suggesting that these were breakdown products of the full-length proteins after proteolytic cleavage. The fact that these lower molecular weight proteins are likely to be degradation products could explain the lack of consistency observed in this region of the gels between samples prepared from different patients. These degradation products were present even though the samples were treated with proteinase inhibitors at collection prior to separation, indicating significant proteolysis occurring either at the cell surface or within the medium after secretion of the protein from the cell.

Treatment of chondrocytes with pro-inflammatory cytokines has previously been shown to lead to an increased production of reactive oxygen species [39]. In this preliminary screen, we were able to identify two enzymes, perxiredoxin 1 and superoxide dismutase that are involved in protecting cells from free-radical damage. Peroxirdoxin 1 was found up-regulated by cytokine treatment and has not been previously described in chondrocytes although another member of this family, peroxiredoxin 5, has been shown to be up-regulated by pro-inflammatory cytokines in OA cartilage [40]. We found constitutive expression of superoxide dismutase although this can be up-regulated by pro-inflammatory cytokines in chondrocytes [39]. We were also able to identify cyclophilin A which has a wide-spread expression pattern and is involved in aiding the correct folding of proteins by catalysing the cis–trans proline imidic peptide bond formation. It is found mainly as an intracellular protein, but can be secreted [41] and has been found in increased levels in RA synovial fluid [42]. Apart from its enzymic activity, it is also a chemotactic agent for both eosinophils and neutrophils [43], and its levels in an RA joint appear to correlate with both increased numbers of neutrophils and increased joint swelling [42].

We were also able to further identify three proteins, ß2-microglobulin, S100A11 and cofilin. ß2-Microglobulin was found expressed constitutively on one set of gels and is associated with the heavy chain of class I major histocompatability complex at the cell surface [44]. Among many of the functions associated with ß2-microglobulin, it can act as an osteoblast growth factor and can also stimulate the production of MMP-3 [45] and MMP-1 [46] in synovial fibroblasts. The two intracellular proteins, S100A11 involved in calcium signalling and the cytoskeletal component cofilin, were found on all gels. As there was no significant amount of the intracellular enzyme lactate dehydrogenase found after cytotoxicity testing, the intracellular proteins identified are unlikely to be present due to cell death.

Using a very simple sample preparation protocol we were able to identify proteins of interest but it was clear that we needed to develop a way of cleaning the sample so that we could reproducibly resolve proteins in all regions of the gels. In vivo and during culture, chondrocytes produce large amounts of GAGs as both monomeric proteoglycan and HA [47], and the presence of differing amounts of highly charged GAG side-chains are very likely to interfere with IEF. As in a recent publication [48], we began to investigate methods for removal of GAGs in the hope that this would improve the focusing. Either precipitation with CPC or enzymic digestion with hyaluronidase was unable to resolve the focusing problems sufficiently although a combination of both treatments allowed the upper regions of the gels to be resolved. Treatment of the samples with hyaluronidase gave more discrete protein spots while CPC treatment appeared to improve the IEF across the full pH range. CPC is a long chained aliphatic quaternary amine that binds to the highly charged side groups of GAG chains. Precipitation by CPC works best on large, highly sulphated GAGs and the efficiency is dependent upon the size of the GAG chains and the charge density with sulphate groups binding CPC more strongly than carboxyl groups [49]. CPC precipitation may well precipitate some HA, which contains no sulphate but only carboxyl side groups under the salt conditions used, but will instead be much more efficient at precipitating the large proteoglycan molecules such as aggrecan that contain many highly sulphated chondroitin sulphate side-chains. This suggests that CPC precipitation would remove the large proteoglycan molecules and possibly the larger molecules of HA but would leave the smaller HA molecules in solution to interfere with IEF. Enzymic digestion by hyaluronidase on the other hand would tend to be more effective at removing the smaller contaminating HA chains and be less effective at removing the proteoglycans.

While combining enzymic digestion and precipitation succeeded in solving some of our earlier focusing problems, it raised the problem of protein loss (about 25% by Bradford protein assay) during GAG precititation. The ADAMTS enzymes, which include the aggrecanases ADAMTS 4 and 5, contain multiple thrombospondin type 1 repeats which are known to interact with proteoglycans [29] and so are likely to be lost during the proteoglycan precipitation. Some of the cytokines and chemokines that have been implicated in arthritic diseases have also been shown to interact with proteoglycans including IL-8 [50] and IL-10 [51] and may also be removed by the precipitation step. This leaves the possibility that some proteins of considerable interest to cartilage biology could be removed using this method. The addition of hyaluronidase adds more proteins to an already complicated mix which could cause interference. To overcome some of the problems associated with enzymic digestion and precipitation, we investigated using either DEAE or Q sepharose anion-exchange resins to remove the charged GAG side-chains. The Q sepharose was found to have ideal binding characteristics as we could load the column in 1 M high salt conditions and collect the proteins of interest in the flow while the GAG remained bound to the column. Loading the column in high salt conditions also had the added bonus of disrupting protein–protein or protein–GAG interactions, and so we are less likely to lose the proteins which under normal conditions would be bound to GAG. We will, of course, still lose any protein secreted that contains a GAG side-chain.

We have been able to identify some low molecular weight proteins of potential interest. While this is still early data and requires further work, we were still able to identify nine proteins produced by chondrocytes (Table 1), which show a broad range of functions. In order to identify high molecular weight proteins, we have developed a method that allows us to use IEF and separate all the proteins produced by chondrocytes. By using an ion-exchange step to pre-clean samples to remove contaminating GAGs, we are able to successfully focus on proteins secreted from human chondrocytes. We are now using this protocol in conjunction with real-time PCR studies to investigate changes in the protein expression of primary human chondrocytes.

Formula


   Acknowledgments
 
This work was generously supported by the Arthritis Research Campaign (ARC), UK and the Dunhill Medical Trust.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 

  1. Muir H. (1995) The chondrocyte, architect of cartilage. Biomechanics, structure, function and molecular biology of cartilage matrix macromolecules. Bioessays 17:1039–48.[CrossRef][Web of Science][Medline]
  2. Saklatvala J, Pilsworth LM, Sarsfield SJ, Gavrilovic J, Heath JK. (1984) Pig catabolin is a form of interleukin 1. Cartilage and bone resorb, fibroblasts make prostaglandin and collagenase, and thymocyte proliferation is augmented in response to one protein. Biochem J 224:461–6.[Web of Science][Medline]
  3. Dingle JT, Page Thomas DP, King B, Bard DR. (1987) In vivo studies of articular tissue damage mediated by catabolin/interleukin 1. Ann Rheum Dis 46:527–33.[Abstract/Free Full Text]
  4. Page Thomas DP, King B, Stephens T, Dingle JT. (1991) In vivo studies of cartilage regeneration after damage induced by catabolin/interleukin-1. Ann Rheum Dis 50:75–80.[Abstract/Free Full Text]
  5. Jubb RW and Fell HB. (1980) The breakdown of collagen by chondrocytes. J Pathol 130:159–67.[CrossRef][Web of Science][Medline]
  6. Sternlicht MD and Werb Z. (2001) How matrix metalloproteinases regulate cell behavior. Annu Rev Cell Dev Biol 17:463–516.[CrossRef][Web of Science][Medline]
  7. Cawston T, Plumpton T, Curry V, Ellis A, Powell L. (1994) Role of TIMP and MMP inhibition in preventing connective tissue breakdown. Ann N Y Acad Sci 732:75–83.[Web of Science][Medline]
  8. Nagase H and Woessner JF. (1993) Role of endogenous proteinases in the degradation of cartilage matrix. In Woessner JF and Howell DS (Eds.). Joint Cartilage Degradation (Marcel Dekker, New York).
  9. Stricklin GP, Bauer EA, Jeffrey JJ, Eisen AZ. (1977) Human skin collagenase: isolation of precursor and active forms from both fibroblast and organ cultures. Biochemistry 16:1607–15.[CrossRef][Medline]
  10. Hasty KA, Jeffrey JJ, Hibbs MS, Welgus HG. (1987) The collagen substrate specificity of human neutrophil collagenase. J Biol Chem 262:10048–52.[Abstract/Free Full Text]
  11. Freije JM, Diez-Itza I, Balbin M, et al. (1994) Molecular cloning and expression of collagenase-3, a novel human matrix metalloproteinase produced by breast carcinomas. J Biol Chem 269:16766–73.[Abstract/Free Full Text]
  12. Aimes RT and Quigley JP. (1995) Matrix metalloproteinase-2 is an interstitial collagenase. Inhibitor-free enzyme catalyzes the cleavage of collagen fibrils and soluble native type I collagen generating the specific 3/4- and 1/4-length fragments. J Biol Chem 270:5872–6.[Abstract/Free Full Text]
  13. Ohuchi E, Imai K, Fujii Y, Sato H, Seiki M, Okada Y. (1997) Membrane type 1 matrix metalloproteinase digests interstitial collagens and other extracellular matrix macromolecules. J Biol Chem 272:2446–51.[Abstract/Free Full Text]
  14. Hollander AP, Pidoux I, Reiner A, Rorabeck C, Bourne R, Poole AR. (1995) Damage to type II collagen in aging and osteoarthritis starts at the articular surface, originates around chondrocytes, and extends into the cartilage with progressive degeneration. J Clin Invest 96:2859–69.[Web of Science][Medline]
  15. Kozaci LD, Buttle DJ, Hollander AP. (1997) Degradation of type II collagen, but not proteoglycan, correlates with matrix metalloproteinase activity in cartilage explant cultures. Arthritis Rheum 40:164–74.[Web of Science][Medline]
  16. Wolfe GC, MacNaul KL, Buechel FF, et al. (1993) Differential in vivo expression of collagenase messenger RNA in synovium and cartilage. Quantitative comparison with stromelysin messenger RNA levels in human rheumatoid arthritis and osteoarthritis patients and in two animal models of acute inflammatory arthritis. Arthritis Rheum 36:1540–7.[Web of Science][Medline]
  17. Clark IM, Powell LK, Ramsey S, Hazleman BL, Cawston TE. (1993) The measurement of collagenase, tissue inhibitor of metalloproteinases (TIMP), and collagenase-TIMP complex in synovial fluids from patients with osteoarthritis and rheumatoid arthritis. Arthritis Rheum 36:372–9.[Web of Science][Medline]
  18. Tetlow LC and Woolley DE. (1998) Comparative immunolocalization studies of collagenase 1 and collagenase 3 production in the rheumatoid lesion, and by human chondrocytes and synoviocytes in vitro. Br J Rheumatol 37:64–70.[Abstract/Free Full Text]
  19. Konttinen YT, Ainola M, Valleala H, et al. (1999) Analysis of 16 different matrix metalloproteinases (MMP-1 to MMP-20) in the synovial membrane: different profiles in trauma and rheumatoid arthritis. Ann Rheum Dis 58:691–7.[Abstract/Free Full Text]
  20. Westhoff CS, Freudiger D, Petrow P, et al. (1999) Characterization of collagenase 3 (matrix metalloproteinase 13) messenger RNA expression in the synovial membrane and synovial fibroblasts of patients with rheumatoid arthritis. Arthritis Rheum 42:1517–27.[CrossRef][Web of Science][Medline]
  21. Shingleton WD, Hodges DJ, Brick P, Cawston TE. (1996) Collagenase: a key enzyme in collagen turnover. Biochem Cell Biol 74:759–75.[Web of Science][Medline]
  22. Billinghurst RC, Dahlberg L, Ionescu M, et al. (1997) Enhanced cleavage of type II collagen by collagenases in osteoarthritic articular cartilage. J Clin Invest 99:1534–45.[Web of Science][Medline]
  23. Nagase H. (1997) Activation mechanisms of matrix metalloproteinases. Biol Chem Hoppe Seyer 378:267–83.
  24. Brew K, Dinakarpandian D, Nagase H. (2000) Tissue inhibitors of metalloproteinases: evolution, structure and function. Biochim Biophys Acta 1477:267–83.[CrossRef][Medline]
  25. Cawston TE, Curry VA, Summers CA, et al. (1998) The role of oncostatin M in animal and human connective tissue collagen turnover and its localization within the rheumatoid joint. Arthritis Rheum 41:1760–71.[CrossRef][Web of Science][Medline]
  26. MacNaul KL, Chartrain N, Lark M, Tocci MJ, Hutchinson NI. (1992) Differential effects of IL-1 and TNF alpha on the expression of stromelysin, collagenase and their natural inhibitor, TIMP, in rheumatoid human synovial fibroblasts. Matrix Suppl 1:198–9.[Medline]
  27. Cawston TE, Milner JM, Catterall JB, Rowan AD. (2003) Cytokine synergy, collagenases and cartilage collagen breakdown. Biochem Soc Symp 125–33.
  28. Koshy PJ, Lundy CJ, Rowan AD, et al. (2002) The modulation of matrix metalloproteinase and ADAM gene expression in human chondrocytes by interleukin-1 and oncostatin M: a time-course study using real-time quantitative reverse transcription-polymerase chain reaction. Arthritis Rheum 46:961–7.[CrossRef][Web of Science][Medline]
  29. Tang BL. (2001) ADAMTS: a novel family of extracellular matrix proteases. Int J Biochem Cell Biol 33:33–44.[CrossRef][Web of Science][Medline]
  30. Patterson SD. (2004) Proteomics: beginning to realize its promise? Arthritis Rheum 50:3741–4.[CrossRef][Web of Science][Medline]
  31. Staunton D, Hudson KR, Heath JK. (1998) The interactions of the cytokine-binding homology region and immunoglobulin-like domains of gp130 with oncostatin M: implications for receptor complex formation. Protein Eng 11:1093–102.[Abstract/Free Full Text]
  32. Farndale RW, Buttle DJ, Barrett AJ. (1986) Improved quantitation and discrimination of sulphated glycosaminoglycans by use of dimethylmethylene blue. Biochim Biophys Acta 883:173–7.[Medline]
  33. Yan JX, Wait R, Berkelman T, et al. (2000) A modified silver staining protocol for visualization of proteins compatible with matrix-assisted laser desorption/ionization and electrospray ionization-mass spectrometry. Electrophoresis 21:3666–72.[CrossRef][Web of Science][Medline]
  34. Leung KY, Wait R, Welson SY, et al. (2001) A reference map of human lung MRC-5 fibroblast proteins using immobilized pH gradient-isoelectric focusing-based two-dimensional electrophoresis. Proteomics 1:787–94.[CrossRef][Web of Science][Medline]
  35. Milner JM, Elliott SF, Cawston TE. (2001) Activation of procollagenases is a key control point in cartilage collagen degradation: interaction of serine and metalloproteinase pathways. Arthritis Rheum 44:2084–96.[CrossRef][Web of Science][Medline]
  36. Hakala BE, White C, Recklies AD. (1993) Human cartilage gp-39, a major secretory product of articular chondrocytes and synovial cells, is a mammalian member of a chitinase protein family. J Biol Chem 268:25803–10.[Abstract/Free Full Text]
  37. Matsumoto T and Tsurumoto T. (2001) Serum YKL-40 levels in rheumatoid arthritis: correlations between clinical and laborarory parameters. Clin Exp Rheumatol 19:655–60.[Web of Science][Medline]
  38. Ling H and Recklies AD. (2004) The chitinase 3-like protein human cartilage glycoprotein 39 inhibits cellular responses to the inflammatory cytokines interleukin-1 and tumour necrosis factor-alpha. Biochem J 380:651–9.[CrossRef][Web of Science][Medline]
  39. Lo YY and Cruz TF. (1995) Involvement of reactive oxygen species in cytokine and growth factor induction of c-fos expression in chondrocytes. J Biol Chem 270:11727–30.[Abstract/Free Full Text]
  40. Wang MX, Wei A, Yuan J, Trickett A, Knoops B, Murrell GA. (2002) Expression and regulation of peroxiredoxin 5 in human osteoarthritis. FEBS Lett 531:359–62.[CrossRef][Web of Science][Medline]
  41. Price ER, Zydowsky LD, Jin MJ, Baker CH, McKeon FD, Walsh CT. (1991) Human cyclophilin B: a second cyclophilin gene encodes a peptidyl-prolyl isomerase with a signal sequence. Proc Natl Acad Sci USA 88:1903–7.[Abstract/Free Full Text]
  42. Billich A, Winkler G, Aschauer H, Rot A, Peichl P. (1997) Presence of cyclophilin A in synovial fluids of patients with rheumatoid arthritis. J Exp Med 185:975–80.[Abstract/Free Full Text]
  43. Xu Q, Leiva MC, Fischkoff SA, Handschumacher RE, Lyttle CR. (1992) Leukocyte chemotactic activity of cyclophilin. J Biol Chem 267:11968–71.[Abstract/Free Full Text]
  44. Bjorkman PJ, Saper MA, Samraoui B, Bennett WS, Strominger JL, Wiley DC. (1987) Structure of the human class I histocompatibility antigen, HLA-A2. Nature 329:506–12.[CrossRef][Medline]
  45. Migita K, Eguchi K, Tominaga M, Origuchi T, Kawabe Y, Nagataki S. (1997) Beta 2-microglobulin induces stromelysin production by human synovial fibroblasts. Biochem Biophys Res Commun 239:621–5.[CrossRef][Web of Science][Medline]
  46. Moe SM, Singh GK, Bailey AM. (2000) ß2-Microglobulin induces MMP-1 but not TIMP-1 expression in human synovial fibroblasts. Kidney Int 57:2023–34.[CrossRef][Web of Science][Medline]
  47. Bassleer C, Gysen P, Bassleer R, Franchimont P. (1987) Proteoglycans synthesized by human chondrocytes cultivated in clusters. Am J Med 83:25–8.[Web of Science][Medline]
  48. Hermansson M, Sawaji Y, Bolton M, et al. (2004) Proteomic analysis of articular cartilage shows increased type II collagen synthesis in osteoarthritis and expression of inhibin betaA (activin A), a regulatory molecule for chondrocytes. J Biol Chem 279:43514–21.[Abstract/Free Full Text]
  49. Roden L, Baker J, Cifonelli A, Matthew MB. (1972) Academic Press Vol. 28:73–8.
  50. Goger B, Halden Y, Rek A, et al. (2002) Different affinities of glycosaminoglycan oligosaccharides for monomeric and dimeric interleukin-8: a model for chemokine regulation at inflammatory sites. Biochemistry 41:1640–6.[CrossRef][Medline]
  51. Salek-Ardakani S, Arrand JR, Shaw D, Mackett M. (2000) Heparin and heparan sulfate bind interleukin-10 and modulate its activity. Blood 96:1879–88.[Abstract/Free Full Text]
Submitted 12 July 2005; revised version accepted 31 January 2006.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary data
Right arrow All Versions of this Article:
45/9/1101    most recent
kel060v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (15)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Catterall, J. B.
Right arrow Articles by Cawston, T. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Catterall, J. B.
Right arrow Articles by Cawston, T. E.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?