OsteoArthritis and Cartilage (2005) 13, 225e234 2004 OsteoArthritis Research Society International. Published by Elsevier Ltd.

. All rights reserved. doi:10.1016/j.joca.2004.11.005

International Cartilage Repair Society

Characterizing osteochondrosis in the dog: potential roles for matrix metalloproteinases and mechanical load in pathogenesis and disease progression1
K. Kuroki D.V.M., Ph.D., J. L. Cook D.V.M., Ph.D.*, A. M. Stoker M.S., S. E. Turnquist D.V.M., Ph.D., J. M. Kreeger D.V.M., Ph.D. and J. L. Tomlinson D.V.M., M.V.Sc.
Comparative Orthopaedic Laboratory, University of Missouri, USA Summary
Objective: To address possible roles of matrix metalloproteinases (MMPs) and mechanical stress in the pathogenesis of osteochondrosis (OC). Methods: Naturally-occurring canine OC lesions (n Z 50) were immunohistochemically analyzed for MMP-1, -3, and -13, and normal canine articular cartilage explants (n Z 6) cultured under 0-, 2-, or 4-MPa compressive loads (0.1 Hz, 20 min every 8 h up to 12 days) were compared to OC samples (n Z 4) biochemically and molecularly. Results: MMP-1 and -3 immunoreactivities were readily detected in both OC samples and control tissues obtained from age-matched dogs (n Z 11) whereas MMP-13 was only detectable in OC samples. MMP-13 gene expression as determined by real-time reverse transcription polymerase chain reaction was elevated in OC samples and cartilage explants cultured without mechanical stimuli (0 MPa groups) compared to normal cartilage (day 0 controls). Glycosaminoglycan content (per weight) in cartilage explants cultured under no load was signicantly (P ! 0.05) lower on day 12 than in the day 0 controls. Gene expression levels of aggrecan and type II collagen in OC samples were lower than those in the day 0 controls. High levels of aggrecan and collagen II expression were seen in the 2 MPa groups. Conclusions: These ndings imply that impaired biochemical characteristics in OC-affected cartilage may be attributable to decreased extracellular matrix production that may stem from disruption of normal weight bearing forces. 2004 OsteoArthritis Research Society International. Published by Elsevier Ltd. All rights reserved. Key words: Osteoarthritis, Matrix metalloproteinase, Compressive load, Dog.

Introduction
Osteochondrosis (OC) is a developmental orthopedic disease commonly diagnosed across species1. Although many factors including genetics, nutrition, hormonal disturbances, trauma and ischemia are thought to play roles in OC development, the disease mechanisms are not fully understood. Since the pathophysiology of the disease has high similarities between humans and animals1, studies of OC using the dog as a model may deepen our insights into the disease and contribute to elucidating its pathogenesis in both humans and animals. Tomlinson et al.2 reported that glycosaminoglycan (GAG) content and immunoreactivity of collagen types I, II, and X were different between cartilage affected by OC and normal cartilage in dogs. Alterations in the biochemical characteristics of cartilage extracellular matrix in OC lesions have also been reported in other species3e5. Findings of these studies suggest that disruption of the integrity of cartilage
1 Supported by Hohn-Johnson Research Award, Veterinary Orthopaedic Society. *Address correspondence and reprint requests to: James L. Cook, D.V.M., Ph.D., Comparative Orthopaedic Laboratory, University of Missouri, 379 East Campus Drive, Columbia, MO 65211, USA. Tel: 1-573-882-7821; Fax: 1-573-884-5444; E-mail: cookjl@missouri.edu Received 8 November 2003; revision accepted 10 November 2004.

extracellular matrix may play a role in the disease development. The question remains whether abnormalities in the extracellular matrix of cartilage affected by OC are the result of decreased production of integral matrix molecules, increased loss of these molecules, or both. A recent study from our laboratory6 showed that cellular viability and matrix composition in three-dimensional chondrocyte cultures from OC lesions were inferior to those from normal cartilage. The study indicated that at least a portion of matrix alterations in cartilage affected by OC can be attributed to disruption of matrix turnover by chondrocytes6. However, whether these changes represent a cause or an effect of the disease is still unknown. In addition to impaired production of matrix macromolecules by chondrocytes, abnormal remodeling and degradation of extracellular matrix may also contribute to alterations in the biochemical characteristics of cartilage affected by OC. Matrix metalloproteinases (MMP) are known to play important roles in articular cartilage extracellular matrix turnover. In diarthrodial joints, MMP-1 (interstitial collagenases), MMP-3 (stromelysin-1), and MMP-13 (collagenase-3) have been given great attention because of their known ability to degrade cartilage-specic proteoglycans and collagens. However, little information as to the expression of these MMPs in naturally-occurring OC lesions is present in the literature. Inappropriately high MMP levels would implicate increased protease activity as a cause of abnormal extracellular matrix in cartilage affected by OC, whereas 225

226 inappropriately low MMP levels would suggest extracellular matrix abnormalities result from retarded matrix turnover. Douglas and Rang7 have reported that traumatic mechanical stress on joints plays an important role in the pathogenesis of OC. Since biomechanical forces have been incriminated as a possible etiologic factor in OC, further investigation involving precisely controlled mechanical forces on articular cartilage may help to delineate the roles of mechanical force in disease development. Recently, a computer-driven dynamic pressure transmitter system has been used to investigate the effects of mechanical forces on cultured cells and tissues making this type of investigation possible8e11. Therefore, the objectives of the present study were to determine immunoreactivity of MMP-1, -3, and -13 in naturally-occurring OC of dogs and to delineate the effects of compressive forces on cellular and extracellular matrix characteristics of canine cartilage explants. The hypotheses of the present study were that altered expression of MMP-1, -3, and -13 would be detected in naturally-occurring OC lesions of dogs and that repetitive high compressive loads on normal canine articular cartilage explants in vitro would alter characteristics of chondrocytes and extracellular matrix similar to those of naturally-occurring canine OC.

K. Kuroki et al.: MMP and mechanical stress in OC International). Load was applied at a frequency of 0.1 Hz at 5.6 lb or 11.2 lb for 20 min, every 8 h through the BioFlex plate piston (Flexcell International). Substantial loads applied on cartilage explants were calculated based on the formula; pressure Z force/area. With a 4 mm diameter explant in each well, the pressures applied in the present study: 5.6 lb and 11.2 lb correspond to loads of 2 MPa and 4 MPa applied onto the cartilage, respectively. Cartilage explants cultured in the identical environment with no compressive load applied served as 0 MPa groups. The explants were incubated at 37(C with 5% CO2 and 95% humidity. The explants were collected on days 6 and 12 and the media were collected and refreshed every 3 days for evaluation of the effects of compressive loading.
MMP IMMUNOHISTOCHEMISTRY

Materials and methods

SAMPLE COLLECTION

Immunohistochemistry was performed on ex vivo cartilage samples (naturally-occurring OC and normal cartilages) as well as cartilage explants undergoing mechanical loading using mouse monoclonal antibodies against human MMP-1, -3, and -13 (Oncogene Research Products, Cambridge, MA) as previously described12. The immunostained sections were subjectively evaluated for the presence and location of staining and scored by two investigators who were blinded to sample group on the basis of the presence (1) or the absence (0) of positive staining.
CELL VIABILITY

For the MMP immunohistochemical study, detached articular cartilage aps (n Z 50) from OC lesions of the humeral head were collected from 39 client-owned dogs (range 4e30 months of age, mean 10.4 months) either arthroscopically or after arthrotomy. Grossly normal, fullthickness articular cartilage samples (n Z 11) were obtained from the caudocentral humeral head of 11 young adult dogs (range 7e18 months of age, mean 15.4 months) immediately after euthanasia performed for reasons unrelated to this study. Cartilage samples were placed in 10% buffered formalin for routine histologic processing and 5micron sections were obtained. For the in vitro mechanical loading study, full-thickness articular cartilage slices were aseptically obtained from the caudocentral portion of the humeral head of six young adult canine cadavers (range 12e24 months of age) via arthrotomy performed immediately after euthanasia. The dogs were euthanatized for reasons unrelated to this study and were apparently healthy with grossly normal humeral head cartilage. Using a 4-mm dermal punch, uniform-sized full-thickness cartilage explants were obtained from each dog. Cartilage samples collected immediately after euthanasia served as day 0 controls for analyses of histology, biochemistry, and gene expression. Remaining cartilage explants from each dog were placed in 6-well BioFlex plates (Flexcell International, Hillsborough, NC) with one explant per well. Cartilage explants from each dog were prepared in triplicate for histology, biochemistry, and gene expression analyses and were bathed with 3 ml of Dulbeccos modied eagle medium supplemented with 10% fetal bovine serum, antibiotics (100 IU/ml penicillin, 100 mg/ml streptomycin, 2.5 mg/ml amphotericin B), and ascorbate (50 mg/ml). A FX-4000C Flexercell Compression Plus Unit (Flexcell International) was used to apply cyclic compressive loads on cartilage explants. Cyclic compression was driven by air pressure, which was monitored with an in-line manometer and controlled by solenoid valves using FlexSoft software (Flexcell

Chondrocyte viability in cartilage explants was determined with use of ethidium homodimer-1 (13 ml/ml) and calcein AM (0.4 ml/ml) uorescent stains (Live/Dead Viability/Cytotoxicity Kit, Molecular Probes, Eugene, OR) and a confocal laser microscopy (BioRad Radiance 2000 confocal system coupled to an Olympus IX70 inverted microscope) equipped with KryptoneArgon mixed lasers and a red diode laser. Approximately 1.5 mm thickness cartilage cross-sections were incubated with the stains in phosphate buffered saline for 30 min at room temperature. The number of calcein AM stained live cells (green uorescence) and ethidium homodimer-1 stained dead cells (red uorescence) of all explants were objectively evaluated using image analysis software (Image-Pro Plus, MediaCybernetics, Carlsbad, CA).
EXTRACELLULAR MATRIX EVALUATION

To evaluate proteoglycan distribution within the tissue matrix, samples were processed using routine histological procedures and 5 mm sections were stained with toluidine blue (T-blue). In addition, GAG in cartilage explants and in liquid media were quantitated by DMMB assay13. Briey, collected cartilage explant samples were lyophilized (FreeZone, Labconco, Kansas City, MO), weighted (dry weight), and digested in 1 ml of 0.5 mg/ml papain (Sigma, St. Louis, MO) in 20 mM sodium phosphate buffer (pH 6.8) containing 1 mM EDTA at 60(C overnight. A 5 ml aliquot of the digest solution was assayed for total GAG content (GAG/dry weight) by addition of 245 ml of DMMB solution and spectrophotometric determination of absorbance at 525 nm with known concentrations of chondroitin sulfate as standards. Results were standardized to correct for differences in sample weight. Total GAG content for samples was reported in micrograms per milligram. GAG content in liquid medium samples (5 ml) were assayed by

Osteoarthritis and Cartilage Vol. 13, No. 3 addition of 245 ml of DMMB solution and spectrophotometric determination of absorbance at 525 nm. GAG loss from the explant was determined by calculating GAG content in the liquid media as a percentage of total GAG in the explant for each sample. For collagen matrix evaluation, total collagen contents of the cartilage explants were determined by measuring hydroxyproline (HP) using a colorimetric procedure14 as previously described15. Presence and distribution of type I, II, and X collagens in cartilage explants were subjectively determined with use of an avidinebiotin immunohistochemical technique as previously described2.
REAL-TIME REVERSE TRANSCRIPTION POLYMERASE CHAIN REACTION (REAL-TIME RT-PCR)

227 each set produced a single product. The PCR products were then sequenced and compared to known sequences using BLAST17 to ensure the specicity of the primer sets. The melt curve data obtained during this testing were then utilized to determine the specicity of subsequent PCR tests in this study. The PCR prole for all tests consisted of an initial incubation at 94(C for 15 min, followed by 55 cycles of 5 s at 94(C, 10 s at 57(C and 20 s at 72(C. After the PCR prole a melt curve analysis was done to ensure specic amplication for each sample. SYBR green uorescence was monitored during the extension step of the PCR prole, and take-off values and amplication efciencies were determined using the Rotor-Gene software. Relative expression levels for the target genes studied were determined using Q-Gene and expressed as a ratio to the level of GAPDH expression18. Real-time PCR analysis was performed at least twice for each cartilage sample. Means (%) of S.E.M. of normalized expression of samples between assays were smaller than 20%, indicating the reproducibility of real-time PCR was high. The no-RT controls were negative for each primer set utilized, indicating genomic DNA contamination was undetectable.
STATISTICAL ANALYSIS

Total RNA was extracted from naturally-occurring OC lesions of the humeral head collected from four clientowned dogs (n Z 4, range 6e7 months of age, mean 6.4 months) and the cartilage explants from the in vitro mechanical loading study and analyzed by RT-PCR using primers (Table I) corresponding to cDNA sequences for structural matrix macromolecules (type I and II collagens, aggrecan, and decorin), proteinases (MMP-3 and MMP-13), proteinase inhibitors (TIMP-1, TIMP-2, and TIMP-3), and the house keeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Snap frozen cartilage samples were powdered, transferred to a 0.5 ml screw cap tube lled with 1.0 mm diameter Zirconia Beads (BioSpec Products, Bartlesville, OK) and Trizol reagent (Invitrogen, Carlsbad, CA), and homogenized using a mini-bead beater (BioSpec Products) at 5000 rpm for 2 min. RNA was extracted from the homogenates using the TRIspin method as described16. RNA was converted to cDNA using random hexamer primers and the StrataScript RT enzyme (StrataGene, La Jolla, CA). RealTime PCR was performed with the Rotor-Gene RG-3000 (Corbett Research, Sydney, Australia) using the Quantitect SYBR green PCR kit (Qiagen, Valencia, CA) following the manufacturers guidelines. Primer sets were designed for each gene, and tested using this system to ensure that

All statistical analyses were performed using a computer software program (SigmaStat, Jandel Scientic, San Rafael, CA). In the study of MMP immunohistochemistry for OC lesions, the presence or absence of staining of each MMP was compared using a ManneWhitney rank sum test. Case correlation coefcients among the different parameters were determined by Spearman rank order correlation. For the in vitro mechanical loading study, data from each collection day in each group were combined and means G standards errors (S.E.M.) were determined. A one-way ANOVA was performed to determine differences among treatment groups with respect to each assay at each collection time. When signicant differences among groups were obtained, an all pair-wise multiple comparison (Tukey test) was performed to determine which groups were different from each other. Differences compared to day

Table I Oligonucleotide primers used for RT-PCR (fw: forward primer; rv: reverse primer) Target template Col 2 Co1 1 Aggrecan Decorin MMP-3 MMP-13 TIMP-1 TIMP-2 TIMP-3 GAPDH PCR primers fw: 5#-TGGAGCAGCAAGAGCAAGGA-3# rv: 5#-ATCAGGTCAGG TCAGCCATTC-3# fw: 5#-ACCTGCCGCGACCTCAAGATG-3# rv: 5#-CAACGTCCA AGGTGCCACA-3# fw: 5#-ACT GTTTCCCAGGAACTTGC-3# rv: 5#-TGG GTCTCCTCTCCTGAGT-3# fw: 5#-ATCTCCGCTGTTGACAATGG-3# rv: 5#-TGGCAGAGCGCACGTAGAC-3# fw: 5#-ATGGCATCCAGTCCCTGTAT-3# rv: 5#-AAAGAACAGGAACTCTCCCC-3# fw: 5#-GCCTTCCTCTTCTT GAGCTG-3# rv: 5#-TTGGACCACTTGAGAGTTCG-3# fw: 5#-GCCCAGAG AGACTCACCAGA-3# rv: 5#-CTCAC AGCCAGCAGCATAGG-3# fw: 5#-GACCAGACAAGGACATAG AG-3# rv: 5#-TCAGAGCTGGACCAGT CTAA-3# fw: 5#-AACTCCGACATCGTGATCCG-3# rv: 5#-GTAGTAGCAGGACTTG ATCT-3# fw: 5#-CCACGGCAAATTCCAC GGCACAG-3# rv: 5#-GGGGTCCCTCCGATGCCTG CTTC-3# Amplication product size (bp) 563 551 607 290 161 320 472 488 347 652

Primers correspond to cDNA sequences deposited in GeneBank. Col 2 Z collagen II, Col 1 Z collagen I.

228 0 controls with respect to each assay at different collection times were analyzed in a similar manner. Signicance was established at P ! 0.05.

K. Kuroki et al.: MMP and mechanical stress in OC were positive for MMP-3. When present, distribution of MMP-1 [Fig. 1(D)] and MMP-3 [Fig. 1(E)] in OC samples was similar to those in controls. MMP-13 was detected in 21 OC-affected cartilage samples [Fig. 1(F)]. There was a statistically signicant difference between groups with signicantly (P ! 0.001) more cartilage affected by OC staining positive for MMP-13 than controls.

Results
MMP IMMUNOHISTOCHEMISTRY FOR NATURALLY-OCCURRING OC

In normal cartilage controls (n Z 11), MMP-1 and -3 were detected in eight samples. When detected, MMP-1 was identied in chondrocytes in the middle and deep zones [Fig. 1(A)], whereas MMP-3 was diffusively distributed in interterritorial regions of each zone of the matrix, except for zone of calcication [Fig. 1(B)]. A weak, positive correlation (P Z 0.03, r Z 0.28) was found between the presence of MMP-1 and -3 in controls. No MMP-13 staining was present in control samples [Fig. 1(C)]. In OC samples (n Z 50), 22 specimens stained positive for MMP-1 and 20 specimens

CELL VIABILITY

The effects of in vitro compressive loading of cartilage explants were evaluated with confocal laser microscopic analysis for cell viability. The majority of chondrocytes in cartilage explants in the 0 MPa and 2 MPa groups [Fig. 2(A,B)] emitted green uorescence at each sample collection time indicating the maintenance of cell viability throughout the study period. In contrast, many chondrocytes in explants from the 4 MPa group emitted red

Fig. 1. Representative photomicrographs of MMP-1 (A, D), MMP-3 (B, E), and MMP-13 (C, F) immunostaining in controls (A, B, C) and in OCaffected cartilages (D, E, F). MMP-1 was identied in the cytoplasm of chondrocytes in the middle and deep zones (A, D), whereas MMP-3 was diffusively distributed in interterritorial regions of each zone of the matrix, except for the calcication zone (B, E). Note the intensity of immunostaining in OC-affected cartilages (D, E) was apparently stronger than those in controls (A, B). No MMP-13 staining was detectable in control samples (C), while it was detectable in OC-affected cartilages (F). Avidinebiotineperoxidase stain; original magnication Z 20!.

Osteoarthritis and Cartilage Vol. 13, No. 3

229

Fig. 2. Photomicrographs of confocal laser microscopy of tissue cross-sections from articular cartilage explants cultured under 0 MPa (A), 2 MPa (B), and 4 MPa (C) for 12 days. On each image, the top is the cartilage surface. The green dots are live chondrocytes uorescent signals emitted by calcein AM, and the red dots are dead chondrocytes uorescent signals emitted by ethidium homodimer. Note increased number of red dots in the supercial zone in the 4 MPa (C). (E) and (F) correspond to green dots and red dots in (C), respectively. Original magnication Z 10!.
100

DNA/Tissue Weight (g/mg)

uorescence indicating reduced cell viability [Fig. 2(C)]. Analysis of chondrocytes emitting green uorescence or red uorescence in each sample [Fig. 2(D,E)] showed that cell viability in the 4 MPa group was signicantly (P ! 0.05) lower than in the 2 MPa group on day 6 and the 0 MPa group on day 12 [Fig. 3(A)]. To verify the ndings of cell viability evaluated by confocal laser microscopy, diphenylamine assay19 was conducted to quantify total DNA content in cartilage explants, as described previously15. Similar to the results shown by confocal laser microscopy, DNA content in the 4 MPa group was decreased during the study period and the reduction on day 6 and day 12 was signicant (P ! 0.05) compared to day 0 controls [Fig. 3(B)]. These results demonstrated that cell viability could be maintained well under regimens of 0 MPa and 2 MPa compressive loads at 0.1 Hz for 20 min, every 8 h, while the 4 MPa regimen led to signicant cell death in this model.

(A)
* *

Live Cell (mean-%)

80 60 40 20 0

Control 0 MPa 2 MPa 4 MPa

Day 0

Day 6

Day 12
Control 0 MPa 2 MPa 4 MPa

5 4

(B)
*

EXTRACELLULAR MATRIX EVALUATION

3 2 1 0

The effects of different compressive loads on cartilage explant extracellular matrix macromolecules were evaluated. Compared to the day 0 controls, GAG content was signicantly (P Z 0.036) decreased in the 0 MPa group on day 12 (Fig. 4). In addition, T-blue histochemical analysis showed the presence of decreased stained areas in the supercial to middle layers in the 0 MPa group (Fig. 5). These ndings corresponded well with each other and indicated that the lack of mechanical stimuli on cartilage leads to depletion of proteoglycans. Moreover, aggrecan gene expression was apparently higher in the day 0 controls compared to OC samples and the gene expression was better preserved in the 2 MPa group (Table II). Gene expression of decorin between OC samples and the controls was unchanged (Table II). In contrast to GAG content, collagen content in the 0 MPa group as determined by the HP assay was minimally affected during the study period (Fig. 6). A signicant

Day 0

Day 6

Day 12

Fig. 3. (A) Mean (GS.E.M.) percent of live chondrocytes in articular cartilage explants as determined by confocal laser microscopy with image analysis software. The number of live cells was signicantly lower in the 4 MPa group than in the 2 MPa group on day 6 and than in the 0 MPa group on day 12. Reduction of live cells in the 4 MPa group on day 6 was also signicant compared to the day 0 controls. (B) Mean (GS.E.M.) DNA content (DNA/dry weight) in articular cartilage explants as determined by diphenylamine assay. DNA content in the 4 MPa group was signicantly decreased on day 6 and day 12 compared to the day 0 controls (*P ! 0.05).

230
Control 0 MPa 2 MPa 4 MPa

K. Kuroki et al.: MMP and mechanical stress in OC higher in the day 0 controls compared to OC samples and collagen II gene expression was best preserved in the 2 MPa groups among the three treatment regimens (Table II). Type X collagen immunohistochemistry showed collagen X to be concentrated in the territorial and interterritorial regions of chondrocytes in supercial to middle layers [Fig. 7(E)] and immunostaining was well preserved in those samples cultured under compressive loads [Fig. 7(FeH)]. No cartilage explants were positive for type I collagen (data not shown) and collagen I gene expression was signicantly (P ! 0.05) higher in OC samples than the day 0 controls. None of the treatment regimens induced collagen I gene expression (Table II).
MMP EVALUATION

300

(A)

GAG/Tissue Weight (g/mg)

250 200 150 100 50 0

Day 0
12 10

Day 6

Day 12
0 MPa 2 MPa 4 MPa

(B)

%-GAG loss

8 6 4 2 0

Day 6

Day 12

Fig. 4. (A) Mean (GS.E.M.) GAG content in articular cartilage explants (GAG/dry weight) as determined by dimethylmethylene assay. GAG in the 0 MPa group was signicantly decreased on day 12 compared to the day 0 control (*P Z 0.036). (B) GAG loss from the cartilage explant calculated as a mean (GS.E.M.) percentage of total GAG. %-GAG loss was not varied among groups throughout the study period.

In order to investigate whether MMPs play a role in the alterations noted in the matrix molecules of loaded cartilage explants, immunohistochemistry for MMP-1, -3, and -13 and real-time RT-PCR for MMP-3 and -13 were performed. Similar to those data seen in the MMP immunohistochemistry on normal cartilage samples, some cartilage samples in day 0 controls were positive for MMP-1 (5/6) and MMP-3 (2/6) and none of day 0 controls stained positive for MMP13 (0/6). When MMP staining was detected, staining patterns were similar to those seen in the rst study (Fig. 1) and the presence and distribution of each MMP were not signicantly varied throughout the culture period (data not shown). Interestingly gene expression of MMP-13 was apparently higher in OC samples compared to the day 0 controls and its expression was higher in the 0 MPa groups than the day 0 controls with a signicant (P ! 0.05) difference on day 6 (Table II). In contrast, MMP-3 gene expression in OC samples and the day 0 controls was not signicantly different. The signicantly (P ! 0.05) lower levels of TIMP-2 and TIMP-3 gene expression in OC samples compared to the day 0 controls implies imbalance between MMP and TIMP in cartilage affected by OC (Table II).

(P ! 0.05) decrease in collagen content for the 4 MPa group was found on day 6 compared to the 0 MPa group and the day 0 controls. Immunohistochemical analyses for type I, II, and X collagens were conducted to further elucidate the effects of compressive loads on collagen extracellular matrix. Type II collagen expression was diffusively present throughout the tissue with more intense immunostaining in surface layers in the day 0 controls [Fig. 7(A)]. Although all cartilage samples were positive for type II collagen [Fig. 7(BeD)], immunoreactivity was subjectively less intense in all treated groups compared to the day 0 controls. In addition, collagen II gene expression was

Discussion
Only a limited number of studies have looked into potential roles for MMPs in the pathophysiology of OC. Brama et al.20,21 reported that MMPs in OC-affected equine synovial uid were not signicantly different from those in synovial uid obtained from normal horses and proposed that MMP-mediated extracellular matrix destruction is not of major importance in OC. Similarly, a study of naturallyoccurring canine OC showed that concentrations of MMP-3 in synovial uid obtained from OC-affected joints were not

Fig. 5. Representative photomicrographs of cross-sections of articular cartilage explants harvested on day 0 (A) and cultured under 0 MPa (B), 2 MPa (C), and 4 MPa (D) for 12 days. Sections were stained with toluidine blue. Note the denite loss of the stain from the supercial to the middle zones in the 0 MPa (C). Original magnication Z 20!.

Osteoarthritis and Cartilage Vol. 13, No. 3

(G2.3) (G0.2) (G8.5)* (G0.1) (G5.1) (G0.1) (G4.1) (G42.7) (G0.2)

231
Control 0 MPa 2 MPa 4 MPa

4 MPa

20

HP/Tissue Weight (g/mg)

7.1 2.0 57.8 0.3 10.9 0.2 6.9 71.1 0.5

16 12 8 4 0

Table II Mean (GS.E.M.) of relative gene expression levels for the genes studies were expressed as a ratio to the level GAPDH as a reference gene

Day 12

2 MPa

10.1 3.5 146.5 0.7 13.6 0.2 1.1 8.9 0.1

(G4.6) (G1.6) (G59.3) (G0.5) (G7.1) (G0.1) (G0.6) (G4.7)* (G0.0)*

Day 0
(G1.4) (G2.3) (G34.1)* (G0.1) (G1.4) (G1.2) (G1.1) (G1.4)* (G0.0)*

Day 6

Day 12

4.2 5.4 109.8 0.2 3.4 3.7 3.1 5.3 0.1

Fig. 6. Mean (GS.E.M.) hydroxyproline (HP) content (HP/dry weight) in articular cartilage explants. HP content was signicantly decreased in the 4 MPa group on day 6 compared to the 0 MPa group and the day 0 control (*P ! 0.05).

0 MPa

different from those in normal synovial uid22. However, MMPs are produced by various cell types within multiple joint tissues, and therefore, MMPs in the synovial uid may not reect all potential roles of MMPs in the pathogenesis of OC. Clegg and Carter23 reported gelatinase (MMP-2 and -9) production by OC-affected cartilages obtained from horses. Similarly, a recent investigation of equine OC using in situ gelatin zymography demonstrated increased gelatinase activity in OC-affected cartilages24. However, these studies could not provide denitive evidence as to whether the gelatinase production was a causative or resultant event in OC. To the authors knowledge, there are no data in the literature that conclude that increases in MMPs are directly responsible for loss of cartilage matrix integrity in OC. The results of MMP analyses in the present study demonstrated that MMP-13 immunoreactivity was only detectable in OC samples and MMP-13 gene expression was higher in OC than normal cartilage tissues. Although the expression of MMPs as determined by immunohistochemistry and real-time RT-PCR may not accurately reect enzyme activity, the ndings in the present study are not contradictory to a recent study that showed increased collagenase-generated COL2-3/4short in equine OC cartilage25. Although MMP-13 expression identied in OCaffected cartilage samples may represent a cause or an effect of OC, elevated gene expression of MMP-13 in those explants cultured without dynamic compressive loads in the present study implies that an increase in MMP-13 expression in OC samples might be related to a disruption of normal weight bearing forces in OC-affected joints. A decline in TIMP-2 and TIMP-3 synthesis in OC, as suggested by gene expression data from the present study, may magnify the imbalance between MMP and TIMP, leading to further matrix degradation. Caution is necessary with respect to interpretation of gene expression data in this study based on the number of samples studied and the inherent variability in gene expression seen among different animals. In humans, normal physiological stresses on hip joint cartilage have been reported to be as high as 18 MPa26. Although pressures in canine shoulder joints under physiological stresses are not known, it can be speculated that 2 MPa falls within the range of physiological pressures, while 4 MPa may approach supraphysiological pressures based on the weight of these dogs (between 25 and 30 kg),

4 MPa 2 MPa 0 MPa Day 6

3.5 0.9 21.0 0.2 3.5 10.7 2.4 4.2 0.1 Aggrecan Decorin Col 2 Col 1 MMP-3 MMP-13 TIMP-1 TIMP-2 TIMP-3 4.5 6.3 275.0 7.0 0.6 4.5 6.4 2.0 0.1 (G0.6) (G2.5) (G36.3) (G3.8)* (G0.1) (G1.1) (G1.0) (G0.5)* (G0.0)* 14.4 8.7 404.9 0.5 0.9 0.5 1.8 85.1 0.8 (G4.7) (G3.0) (G155.0) (G0.3) (G0.5) (G0.2) (G0.6) (G27.8) (G0.3)

(G2.0)* (G0.4) (G5.7)* (G0.1) (G0.6) (G6.8)* (G1.0) (G1.7)* (G0.0)*

*

1.7 2.8 123.0 0.2 5.7 4.0 1.0 11.0 0.1 Signicant (P ! 0.05) difference compared to the day 0 controls.

OC

Day 0 (controls)

(G0.8)* (G1.4) (G42.6)* (G0.1) (G0.8)* (G1.6) (G0.3) (G5.5)* (G0.0)*

2.3 2.1 18.2 0.1 10.0 0.5 1.7 10.9 0.4

(G1.3)* (G0.8) (G1.8)* (G0.0) (G6.1) (G0.2) (G0.6) (G6.9)* (G0.2)*

232

K. Kuroki et al.: MMP and mechanical stress in OC

Fig. 7. Representative photomicrographs of cross-sections of articular cartilage explants harvested on day 0 (A, E) and cultured under 0 MPa (B, F), 2 MPa (C, G), and 4 MPa (D, H) for 12 days. Sections were stained with antibodies against type II collagen (AeD) or type X collagen (EeH). Type II collagen (AeD) was detected in the territorial and interterritorial regions of cartilage matrix and its immunoreactivity was subjectively less intense in all treatment groups (C, D, E), especially samples from the 4 MPa group compared to the day 0 controls (A). Type X collagen (EeH) was detectable in the territorial and interterritorial regions in the supercial to middle zones. Note type X collagen immunoreactivity was better preserved in explants cultured under compressive forces (G, H). Avidinebiotineperoxidase stain; original magnication Z 20!.

distribution of weight bearing to the forelimbs in dogs, and data from the human literature26. The present in vitro study clearly demonstrated that compressive loading of 4 MPa on canine articular cartilage explants with a frequency of 0.1 Hz for 20 min every 8 h causes increased cell death, especially in more supercial zones. Although cartilage necrosis is a common pathological feature associated with OC in dogs2, the area of necrosis is usually a focal area within the intermediate and deep zones of developing articular or epiphyseal cartilage. Therefore, the mechanism causing cell death associated with high compressive loading is likely different from that seen in naturallyoccurring OC. Destruction of proteoglycan matrix as determined by GAG loss to the media was not signicantly different among groups throughout the study period and this nding was consistent with a recent study in which no evidence of increased proteoglycan degradation was observed in equine OC24. In addition, the apparent decrease in aggrecan gene expression levels in OC samples compared to controls was in accordance with previous studies2e4 reporting decreased proteoglycan content in OC-affected cartilage. Preservation of GAG content despite increased cell death in explants in the 4 MPa group and decreased GAG content despite well preserved cell viability in the 0 MPa group may be explained by differential effects of compressive load on chondrocytes based on phenotype and location as previously reported8,9. In fact, GAG levels in the 4 MPa group normalized by DNA content were the highest among the three treatment groups followed by the 2 MPa group (data not shown). These data suggest that physical destruction of cells and matrix in the supercial zone of loaded explants in this in vitro system with concurrent stimulus to the middle and deep zones appropriate for proteoglycan synthesis may occur. It then follows that disruption of proteoglycan matrix integrity in OC-affected cartilage reported in previous studies2e4 may

be attributable to a lack of normal weight bearing stimuli secondary to OC. Further evidence for incriminating a lack of normal weight bearing stimuli as a cause for diminutive proteoglycan content in OC-affected cartilage is present in the proteoglycan depletion in articular cartilage seen with prolonged unloading of a joint27,28. Decreased collagen content in cartilage explants under 4 MPa dynamic compressive loads and maintenance of collagen content in the 0 MPa group throughout the study period despite increased MMP-13 gene expression levels indicates that this alteration more likely stems from mechanical destruction rather than enzymatic degradation. The differential effects on collagen content and proteoglycan content in this study may be explained by the structure of articular cartilage in which dense collagen bers are distributed in the supercial layer, which has a relatively low amount of proteoglycan. It can be speculated that the compressive loading associated with direct contact of cartilage and the bioreactor in this system had a negative impact on the collagen-dense supercial zone, resulting in prominent reduction of collagen content. The result of collagen II gene expression is suggestive of the importance of mechanical stimuli for transcriptional regulation of type II collagen expression, and further, both lack of load and excess load may cause alterations in the biochemical characteristics of cartilage extracellular matrix. Increased type I collagen gene expression in OC samples veried our previous study2 in which 28 of 42 OC-affected canine cartilage samples were positive for collagen type I while normal cartilage had no detectable type I collagen as determined by immunohistochemistry. Although the present study indicated that compressive loads alone do not affect this shift in collagen synthesis, future research is needed to determine whether type I collagen expression is a cause or an effect of the disease. Type X collagen immunoreactivity was well preserved for cartilage explants cultured under compressive loads in the

Osteoarthritis and Cartilage Vol. 13, No. 3 present study. Although type X collagen has been considered as a specic product of hypertrophic chondrocytes at sites of endochondral ossication, it has also been identied in the surface of normal articular cartilage in various species including adult dogs29,30. The observation of type X collagen in supercial to middle zones and the persistence of its expression in cartilage explants cultured under compressive loads in this study may support previous studies that indicated that type X collagen could contribute to the structural integrity of articular cartilage31 and that expression of this collagen might be regulated by mechanical forces32,33. A previous study2 showed that OC-affected cartilage in dogs had signicantly less type X collagen than did normal cartilage. Further investigation including gene expression analysis is necessary to elucidate whether the decrease in type X collagen is relevant to the disturbance of endochondral ossication or due to the disruption of normal weight bearing secondary to OC. The unknown duration of the disease at the time of surgical intervention is always a major issue in the investigation of OC using naturally-occurring samples. The median age for the onset of lameness associated with humeral head OC in canines has been reported to range between 5 and 6 months34. The initiation of the disease process would occur even earlier than notable lameness. The time lag from the onset of disease, to the diagnosis, and subsequent surgical intervention may conceal critical pathologic alterations in OC cartilage samples. OC is a very common cause of secondary OA in dogs. In addition, disruption of the normal weight bearing forces on OCaffected articular cartilage is likely to cause tissue atrophy. Future investigation using OC cartilage samples should attempt to adapt a rigorous scientic method to select samples which represent the early stages of the disease processes. Although in vitro compressive loading did not alter the gene expression proles in cartilage explants similar to those in OC samples, our ndings suggested that the extracellular matrix composition and integrity altered by OC and loads appears to be related to viability and synthetic capabilities of chondrocytes, perturbations of cell phenotype, and alterations and imbalances in MMP and TIMP synthesis. Further dening the in vivo processes of disease and renement of this model may allow for attainment of a representative in vitro model for the study of OC. Development of an OC model would then allow for directed research into the mechanisms of disease and provide avenues for optimal prevention and treatment across species.

233 cartilage affected by osteochondritis dissecans in the humeral head of dogs. Am J Vet Res 2001;62: 876e81. Lillich JD, Bertone AL, Malemud CJ, Weisbrode SE, Ruggles AJ, Stevenson S. Biochemical, histochemical, and immunohistochemical characterization of distal tibial osteochondrosis in horses. Am J Vet Res 1997;58:89e98. Nakano T, Thompson JR, Aherne FX. Cartilage proteoglycans from normal and osteochondrotic porcine joints. Can J Comp Med 1985;49:219e26. Wardale RJ, Duance VC. Characterization of articular and growth plate cartilage collagens in porcine osteochondrosis. J Cell Sci 1994;107:47e59. Kuroki K, Cook JL, Tomlinson JL, Kreeger JM. In vitro characterization of chondrocytes isolated from naturally occurring osteochondrosis lesions of the humeral head of dogs. Am J Vet Res 2002;63:186e93. Douglas G, Rang M. The role of trauma in the pathogenesis of the osteochondroses. Clin Orthop 1981;158:28e32. Sah RL, Kim YJ, Doong JY, Grodzinsky AJ, Plaas AH, Sandy JD. Biosynthetic response of cartilage explants to dynamic compression. J Orthop Res 1989;7: 619e36. Buschmann MD, Kim YJ, Wong M, Frank E, Hunziker EB, Grodzinsky AJ. Stimulation of aggrecan synthesis in cartilage explants by cyclic loading is localized to regions of high interstitial uid ow. Arch Biochem Biophys 1999;366:1e7. Chowdhury TT, Bader DL, Lee DA. Dynamic compression inhibits the synthesis of nitric oxide and PGE(2) by IL-1beta-stimulated chondrocytes cultured in agarose constructs. Biochem Biophys Res Commun 2001;285:1168e74. Graff RD, Lazarowski ER, Banes AJ, Lee GM. ATP release by mechanically loaded porcine chondrons in pellet culture. Arthritis Rheum 2000;43:1571e9. Kuroki K, Kreeger JM, Cook JL, Tomlinson JL, Johnson GC, Pace LW, et al. Immunohistochemical analysis of matrix metalloproteinase-1, -3, and -13 in naturally occurring cartilaginous tumors of dogs. Am J Vet Res 2002;63:1285e91. Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation and discrimination of sulphated glycosaminoglycans by use of dimethylmethylene blue. Biochim Biophys Acta 1986;883:173e7. Reddy GK, Enwemeka CS. A simplied method for the analysis of hydroxyproline in biological tissues. Clin Biochem 1996;29:225e9. Kuroki K, Cook JL, Kreeger JM, Tomlinson JL. The effects of TIMP-1 and -2 on canine chondrocytes cultured in three-dimensional agarose culture system. Osteoarthritis Cartilage 2003;11:625e35. Reno C, Marchuk L, Sciore P, Frank CB, Hart DA. Rapid isolation of total RNA from small samples of hypocellular, dense connective tissues. Biotechniques 1997;22:1082e6. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol 1990;215: 403e10. Muller PY, Janovjak H, Miserez AR, Dobbie Z. Processing of gene expression data generated by quantitative real-time RT-PCR. Biotechniques 2002; 32:1372e8. Natarajan N, Shambaugh GE 3rd, Elseth KM, Haines GK, Radosevich JA. Adaptation of the diphenylamine

3.

4. 5. 6.

7. 8.

9.

10.

11. 12.

13.

14. 15.

Acknowledgments
The authors would like to thank Dr Gary Gibson (The Henry Ford Hospital, Detroit, MI) for providing a type X collagen antibody. 16.

17.

References
1. Olsson SE, Reiland S. The nature of osteochondrosis in animals. Summary and conclusions with comparative aspects on osteochondritis dissecans in man. Acta Radiol Suppl 1978;358:299e306. 2. Tomlinson JL, Cook JL, Kuroki K, Kreeger JM, Anderson MA. Biochemical characterization of 18.

19.

234 (DPA) assay to a 96-well plate tissue culture format and comparison with the MTT assay. Biotechniques 1994;17:166e71. Brama PA, TeKoppele JM, Beekman B, van El B, Barneveld A, van Weeren PR. Inuence of development and joint pathology on stromelysin enzyme activity in equine synovial uid. Ann Rheum Dis 2000;59:155e7. Brama PA, TeKoppele JM, Beekman B, van Weeren PR, Barneveld A. Matrix metalloproteinase activity in equine synovial uid: inuence of age, osteoarthritis, and osteochondrosis. Ann Rheum Dis 1998;57: 697e9. Hegemann N, Kohn B, Brunnberg L, Schmidt MF. Biomarkers of joint tissue metabolism in canine osteoarthritic and arthritic joint disorders. Osteoarthritis Cartilage 2002;10:714e21. Clegg PD, Carter SD. Matrix metalloproteinase-2 and -9 are activated in joint diseases. Equine Vet J 1999; 31:324e30. Al-Hizab F, Clegg PD, Thompson CC, Carter SD. Microscopic localization of active gelatinases in equine osteochondritis dissecans (OCD) cartilage. Osteoarthritis Cartilage 2002;10:653e61. Laverty S, Okouneff S, Ionescu M, Reiner A, Pidoux I, Webber C, et al. Excessive degradation of type II collagen in articular cartilage in equine osteochondrosis. J Orthop Res 2002;20:1282e9. Hodge WA, Fijan RS, Carlson KL, Burgess RG, Harris WH, Mann RW. Contact pressures in the human hip joint measured in vivo. Proc Natl Acad Sci U S A 1986;83:2879e83.

K. Kuroki et al.: MMP and mechanical stress in OC 27. Haapala J, Arokoski JP, Hyttinen MM, Lammi M, Tammi M, Kovanen V, et al. Remobilization does not fully restore immobilization induced articular cartilage atrophy. Clin Orthop 1999;362:218e29. 28. Kiviranta I, Tammi M, Jurvelin J, Arokoski J, Saamanen AM, Helminen HJ. Articular cartilage thickness and glycosaminoglycan distribution in the young canine knee joint after remobilization of the immobilized limb. J Orthop Res 1994;12:161e7. 29. Rucklidge GJ, Milne G, Robins SP. Collagen type X: a component of the surface of normal human, pig, and rat articular cartilage. Biochem Biophys Res Commun 1996;224:297e302. 30. Lammi PE, Lammi MJ, Hyttinen MM, Panula H, Kiviranta I, Helminen HJ. Site-specic immunostaining for type X collagen in noncalcied articular cartilage of canine stie knee joint. Bone 2002;31:690e6. 31. Eerola I, Salminen H, Lammi P, Lammi M, von der Mark K, Vuorio E, et al. Type X collagen, a natural component of mouse articular cartilage: association with growth, aging, and osteoarthritis. Arthritis Rheum 1998;41:1287e95. 32. Gibson G, Lin DL, Francki K, Caterson B, Foster B. Type X collagen is colocalized with a proteoglycan epitope to form distinct morphological structures in bovine growth cartilage. Bone 1996;19:307e15. 33. Wong M, Siegrist M, Gaschen V, Park Y, Graber W, Studer D. Collagen brillogenesis by chondrocytes in alginate. Tissue Eng 2002;8:979e87. 34. Johnston SA. Osteochondritis dissecans of the humeral head. Vet Clin North Am Small Anim Pract 1998;28: 33e49.

20.

21.

22.

23. 24.

25.

26.