Osteoarthritis and Cartilage (2001) 9, 416–422© 2001 OsteoArthritis Research Society International 1063–4584/01/050416+07 $35.00/0doi:10.1053/joca.2000.0407, available online at http://www.idealibrary.com on

Basic calcium phosphate crystals up-regulate metalloproteinases butdown-regulate tissue inhibitor of metalloproteinase-1 and -2 in humanfibroblastsG. Bai*†, D. S. Howell*, G. A. Howard*†, B. A. Roos*‡ and H. S. Cheung*Research Service and Geriatric Research, Education, and Clinical Center, *VA Medical Center, andDepartments of †Medicine, Biochemistry and Molecular Biology, and ‡Neurology, University of Miami Schoolof Medicine, Miami, Florida 33125, U.S.A.

Summary

Objective: To examine the effect of basic calcium phosphate (BCP) crystals on expression of tissue inhibitors of metalloproteinases (TIMP)-1and -2 in human fibroblasts.

Method: Using a semi-quantitative reverse transcription-polymerase chain reaction method and phosphocitrate (PC), a specific inhibitor ofthe biological effects of BCP crystals, we examined the effects of BCP on the steady state transcript levels of metalloproteinase (MMP)-1,-3, -9 and -13 and TIMP-1 and -2 in human fibroblasts. DNA primers against elongation factor were used as internal controls. RNAs isolatedfrom human fibroblasts treated with BCP crystals (50 �g/ml) in the presence or absence of PC (10−3 M) were used as templates, and RNAfrom untreated control cultures and cultures treated with Interleukin-1-beta (IL-1�) were used as negative and positive controls, respectively.

Results: We observed increases in MMP-1, -3, -9 and -13 transcripts by BCP crystals. BCP crystal down-regulated TIMP-1 and -2 overuntreated controls. Western blot analysis confirmed that BCP crystals down-regulate the synthesis of TIMP-1 and -2. While IL-1�up-regulated MMP-1, -3, -9 and -13, it had no significant effect on expression of either TIMP. In all cases, PC specifically reversed thedifferential regulation of MMPs and TIMPs by BCP crystals but had no effect on IL-1� induction of MMP expression.

Conclusion: The ability of BCP to induce the synthesis of degradative MMPs while down-regulating the synthesis of the naturally occurringcounterpart TIMPs may explain the changes consistent with a role of BCP crystal in the pathogenesis of degenerative changes inosteoarthritis. The ability of PC to reverse both degradative effects of BCP crystal suggests that PC can be a potential therapeutic agent forBCP crystal deposition diseases. © 2001 OsteoArthritis Research Society International

Key words: MMP, TIMP, BCP Crystal, PC, RT-PCR.

Abbreviations: HF, human fibroblasts; MMP, metalloproteinase; TIMP, tissue inhibitor of metalloproteinase; BCP, basic calcium phosphatecrystals; PC, phosphocitrate; RT-PCR, reverse transcription-polymerase chain reaction.

Received 1 June 2000; revision requested 8 September 2000;revision received 13 November; accepted 27 November 2000.

Address correspondence to: Dr Herman S. Cheung, ArthritisDivision (D-26), Dept. of Medicine, University of Miami Schoolof Medicine, 1201 N.W. 16th Street, Miami, Florida 33135,U.S.A. Tel: (305) 243-5735; Fax: (305) 243-5655; E-mail:hcheung@med.miami.edu

Introduction

Basic calcium phosphate (BCP) crystals (hydroxyapatite,octacalcium phosphate, tricalcium phosphate) are commonin osteoarthritic knee effusions and are often associatedwith low-grade synovial proliferation and inflammation1.The presence of BCP crystals correlates well with radio-graphic evidence of cartilaginous degeneration and isassociated with larger joint effusions when compared withjoint fluid from osteoarthritic knees without BCP crystals1,2.BCP crystals are uniformly present in a distinctive destruc-tive arthropathy predominantly affecting the shoulders andknees called Milwaukee Shoulder Syndrome and charac-terized by synovial proliferation and widespread loss ofintrasynovial collagenous structures3.

Calcium-containing crystals are known to have a numberof growth factor-like effects. For example, BCP crystals

416

and epidermal growth factor exert similar biologic effectson cultured cells, such as stimulation of prostaglandin E2(PGE2) production via the phospholipase A2/cyclo-oxygenase pathway, activation of phospholipase C andmitogen-activated protein knase, induction of proto-oncogenes, and induction of metalloproteinase (MMP) 1, 3,9 and 13 (see reviews, 4–6).

The erosion of connective tissue (cartilage, tendon andbone) associated with arthritis is postulated to be mediatedlargely by MMP such as MMP-1 (collagenase-1), MMP-3(stromelysin-1) produced by synovial fibroblasts andchondrocytes7. MMPs belong to a major family of zinc-dependent endopeptidases known for their ability to cleaveone or several extracellular matrix proteins. They sharecommon structural and functional elements and are prod-ucts of several different genes. MMPs play important bio-logical roles in embryo development, morphogenesis andtissue remodeling8. They also play complex roles in patho-logical processes such as arthritis, tissue ulceration andtumor progression9.

The active forms of all of the MMPs are inhibited by afamily of specific inhibitors, the TIMPs. Inhibition representsa major level of control of MMP activity and is thus atherapeutic target. A detailed knowledge of the mechanism

Osteoarthritis and Cartilage Vol. 9, No. 5 417

Materials and methods

CELL CULTURE

Our studies employ the use of human foreskin fibroblasts(HF) as a model for the synovial fibroblasts that interactwith crystals found associated with articular cartilage invivo. These HF and synoviocytes exhibit similar biologicresponses to crystals and growth factors in vitro25. Primarycultures of HF were established from explants and trans-ferred as reported26. Cells were grown and maintained inDulbecco’s modified Eagle’s medium (DMEM) supple-mented with 10% fetal bovine serum (FBS) containingpenicillin and streptomycin. All experiments were per-formed on confluent cell monolayers of third or fourth-passage cells that had been rendered quiescent byremoving the medium, washing with DMEM containing

0.2% FBS (heat-inactivated) and subsequently incubatingthe cells in this medium for 48 h. The cells were thentreated with or without IL-1� (10 ng/ml), BCP crystals(50 mg/ml) and PC (10−3 M) for 24 h. The cell growthconditions and treatments followed previous protocols22,26,where the time course and dose-response of the chemicaltreatments were established.

CRYSTAL SYNTHESIS

BCP crystals were synthesized by a modification ofpublished methods21,27. Minerals prepared by thesemethods had a calcium/phosphate molar ratio of 1.59 andcontained partially carbonate-substituted hydroxyapatite(HA) with admixed octacalcium phosphate (OCP) as shownby Fourier transform infrared spectroscopy. The crystalswere crushed and sieved to yield 10–20 �m aggregates,which were sterilized and rendered pyrogen free by heatingat 200°C for 90 min. Heating to 200°C did not alter thecharacter of the crystal or the relationship of HA and OCP,as confirmed by X-ray diffraction and Fourier transforminfrared spectroscopy. Crystals were weighed and sus-pended by sonication in DMEM. The amount of BCPcrystals used was based on our previous studies of dose–response relationships between BCP crystals and themitogenic response28. In those studies, maximal mito-genic responses were achieved with BCP levels of50–100 �g/ml. These levels are approximately equivalentto concentrations of BCP crystals found in vivo.

RNA ISOLATION

RNA was isolated from human fibroblasts using RNAzolB (Tel-Test B, Inc., Friendswood, TX, U.S.A.) following themanufacturer’s instructions29. The isolated RNA was rou-tinely treated with RNase-free, DNase (Promega, Madison,WI, U.S.A.) at 37°C for 1–2 h then phenol/chloroform (24:1amyl ethanol) was extracted and ethanol precipitated. TheRNA was dissolved in RNase-free TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.0) and stored in aliquots at −80°C.

DNA PRIMERS

Published DNA primers against human sequences ofMMP-1, -930, -3 and -1331, TIMP-1 and -230 were synthe-sized by the University of Miami DNA core facility. Syntheticprimers for rat elongation factor 1� (ELF)32 were used forinternal control of RT-PCR reactions. The DNA primerswere diluted to a concentration of 5 nmol/ml and10–15 pmol of each primer pair were used for PCR ampli-fication of specific genes. The sizes of all tested DNAproducts ranged from 125 to 422 bp. The synthetic DNAprimers used for these studies are listed below.GeneForward primerReverse primerMMP-1GGAAAACACATGGTGTGAGTCCGTTAGAAGAGTTATCCCTTGCCMMP-3ATGAAGAGTCTTCCAATCCTGTCCTTTCTCCTAACAAACTMMP-9TGGGCTACGTGACCTATGACATGCCCAGCCCACCTCCACTCCTC

controlling TIMP gene expression is therefore important.TIMP1 and TIMP2 are capable of inhibiting the activities ofall known MMPs and, as such, play a key role in maintain-ing the balance between extracellular matrix depositionand degradation in different physiological processes.TIMP1 and TIMP2 can inhibit tumor growth, invasion, andmetastasis in experimental models10. Both inhibitors exhibitgrowth factor-like activity and can inhibit angiogenesis10.TIMP-1 and TIMP-2 have many structural and functionalsimilarities but appear to have distinct physiological rolesbecause they are differently expressed and regulated innormal and transformed cells11–14.

It is believed that coordinated control over the expres-sion of both MMP and TIMP is necessary for connectivetissue homeostasis. There is now good evidence that lossof this coordinated regulation, resulting in an imbalance ofdegradative proteinases over protective TIMP activities,underlies the pathological tissue destruction in diseaseslike rheumatoid arthritis and osteoarthritis (OA)15,16.Differential regulation of some MMPs and TIMPs has beenreported in several cell types; for example, in culturedepithelial cells from human bronchial explants17 and rabbitaortic smooth muscle cells18. The regulation of TIMPs byBCP crystals is unknown, however.

Phosphocitrate (PC) is a naturally occurring compoundfound in crab hepatopancreas19 and in mammalian mito-chondria20. Our previous work demonstrated that PCspecifically reversed the biological effect of BCP crystals.For example, PC inhibited BCP crystal induced mito-genesis, proto-oncogene and MMP mRNA expression andMMP protein expression in human foreskin fibroblasts(HF)21,22. Additionally, PC inhibits BCP crystal-inducedphosphorylation of the mitogen-activated protein kinasesignal transduction p42/44 MAPK in HF in a dose-dependent fashion23. Moreover, PC prevents the diseaseprogression in murine progressive ankylosis, an animalmodel of crystal deposition disease24.

In the present study, we compared the steady-statelevels of MMP and TIMP mRNAs by semi-quantitativeRT-PCR in primary cultured HF following treatments withIL-1� and BCP crystals. The fibroblasts were treated withIL-1� or BCP crystals, in the presence or absence of PC.We found that BCP crystals up-regulate MMP-1, -3, -9 and–13, but down-regulate TIMP-1 and -2 transcript levelsresulting in an increase in the MMP/TIMP transcript ratio.PC specifically reversed the effect of crystals on MMPs andTIMPs transcripts but had no effect on IL-1� inducedchanges in cultured HF.

418 G. Bai et al.: Basic calcium phosphate crystals

MMP-13TGCTGCATTCTCCTTCAGGAATGCATCCAGGGGTCCTGGCTIMP-1ACCACCTTATACCAGCGTTAAAACAGGGAAACACTGTGCATIMP-2CCACCCAGAAGAAGAGCCTGAATGGACCCATGGGATGACTGTTTELFAGGTGATTATCCTGAACCATCCAGTGGAGGGTAGTCAGAGAAGC

QUANTIFICATION OF POLYMERASE CHAIN REACTION (PCR)

METHOD

RNA samples isolated from human fibroblasts undernormal growth conditions (DMEM+10% FBS) were usedfor quantification. The RNA was diluted twice to give aserial dilution of 0.25, 0.5 and 1 �g template per reaction.Yeast tRNA was used to supplement RNA concentrationequal to 1 �g in every test tube. Aliquots of PCR reactionswere removed after every two cycles following an initial 20PCR cycles to give 20, 22, 24, 26 and 28 cycle amplifica-tions. Equal amounts of PCR reactions were loaded onto a1.5% agarose gel with ethidium bromide staining. DNAsamples were size fractionated by gel electrophoresis, a

picture of the gel was taken on Polaroid film 667 (SIGMA,St Louis, MO, U.S.A.) under UV light and the DNA bandswere scanned using Corel Photo-Paint 7. The intensities ofDNA bands were compared by a computer programImageCalc33.

MMP-1

ELF

Fig. 1. Up-regulation of MMP-1 in human fibroblasts by BCPcrystals and IL-1�. Photograph of gel of MMP-1 with ELF amplifiedseparately (24 cycles). Lanes are labeled as follows: 1: BCPtreatment (50 �g/ml); 2: BCP+PC (10−3 M); 3: IL-1� treat-ment (10 ng/ml); 4: IL-�+PC (10−3 M); and 5: untreated control.Computer-scanned DNA densities of MMP-1 in comparison withuntreated control set as 100% (lane 5). The data were normalizedwith ELF as the internal control. Bars represent the mean andstandard deviation of three repeated experiments. * indicates

significantly different from control (P<0.02).

MMP-9

ELF

Fig. 2. Up-regulation of MMP-9 by BCP crystals and IL-1�.Photograph depicts the agarose gel of PCR products of MMP-9(36 cycles) with ELF. Computer-scanned DNA densities of MMP-9compared to untreated control set as 100% (lane 5). Lanes arelabeled as follows: 1: BCP treatment (50 �g/ml); 2: BCP+PC(10−3 M); 3: IL-1� treatment (10 ng/ml); 4: IL-�+PC (10−3 M); and5: untreated control. Bars represent mean and standard deviationof three repeated experiments, and * indicates significantly

different from control (P<0.04).

SEMI-QUANTITATIVE REVERSE TRANSCRIPTASE (RT)-PCR

RNA was reverse transcribed into single strand cDNAusing an RNA PCR kit (Perkin-Elmer, Foster City, CA,U.S.A.) with synthetic oligo dT primer and MLV reversetranscriptase at 42°C for 30 min. After reverse transcription,the enzyme was inactivated by heating to 95°C for 5 min,then the reaction mix was distributed into testing groups(equal to 0.5 or 1 �g RNA per reaction in 50 �l volume)which included different MMP and TIMP primers. DNAprimers for elongation factor were also used for internalcontrol of PCR reactions in separate tubes. Different PCRcycles were chosen at linear range amplification of eachindividual primer pair (see Results for details), and three orfour repeated experiments were performed for each assay.Samples without RNA or reverse transcriptase wereroutinely included in the RT-PCR reactions as negativecontrols.

Osteoarthritis and Cartilage Vol. 9, No. 5 419

WESTERN BLOT ANALYSIS

Cell culture medium, conditioned for 48 h and 72 h, wasconcentrated ten-fold (2 ml into 0.2 ml) with Centriconconcentrators (Amicon Inc., Beverly, MA, U.S.A.) and elec-trophoresed on 7.5% SDS-polyacrylamide gel, followed bytransfer to a polyvinylidene flouride (PVDF) filter paper.After elimination of non-specific binding and several wash-ings in between, the blots were incubated consecutivelywith rabbit polyclonal antibody against TIMP 1 and 2(Cehmicon International Inc,Temecula, CA, U.S.A.) andhorseradish peroxidase-linked goat anti-rabbit IgG. Theprotein bands were subsequently detected with the Super-Signal Chemiluminescent Substrate (Pierce, Rockford, IL,U.S.A.) and Kodak X-OMAT film.

STATISTICAL ANALYSIS

Sigma Plot (Jandel Scientific, San Rafael, CA, U.S.A.)was used for chart preparation of computer scanned DNAdensities of PCR amplified MMPs and TIMPs with anaverage of three or four repeated experiments per datapoint shown. Error bars representing standard deviationsare included. A Student t-test was used for statisticalanalysis (P<0.05 is considered significant).

MMP-13

ELF

Fig. 3. Up-regulation of MMP-13 by BCP crystals and IL-1�.Photograph depicts agarose gel of PCR products of MMP-13 (36cycles) and ELF. Lanes are labeled as follows: 1: BCP treatment(50 �g/ml); 2: BCP+PC (10−3 M); 3: IL-1� treatment (10 ng/ml); 4:IL-�+PC (10−3 M); and 5: untreated control. Computer-scannedDNA densities of MMP-13 compared to untreated control set as100% (lane 5). Bars represent mean and standard deviation ofthree repeated experiments, and * indicates significantly different

from control (P<0.04).

TIMP-1

ELF

Fig. 4. Down-regulation of TIMP-1 by BCP crystals. Panel A.Photograph of an agarose gel of PCR products of TIMP-1 with ELF(24 PCR cycles). Lanes are labeled as follows: 1: BCP treatment(50 �g/ml); 2: BCP+PC (10−3 M); 3: IL-1� treatment (10 ng/ml); 4:IL-�+PC (10−3 M); and 5: untreated control. Computer-scannedDNA densities of TIMP-1 normalized to internal controls; untreatedcontrol (lane 5) was set as 100%. Bars represent mean andstandard deviation of three repeated experiments, and * indicates

significantly different from control (P<0.03).

Results

Optimal PCR amplification conditions for every primerpair used for these studies was determined (see individual

figures). For separate amplification of MMPs and TIMPswith ELF, all initial reverse transcription reactions weredone in the same master reaction tubes for each RNAsample in order to eliminate variations caused by reversetranscription.

Up-regulation of MMP-1 by IL-1� and BCP crystals wasobtained with individually amplified ELF as internal con-trols. Up-regulation of MMP-1 served as a positive controlfor RT-PCR determination methods. The human fibroblastswere serum deprived for 48 h, and then the cells weretreated with either IL-1� (10 ng/ml) or BCP crystals (50 �g/ml) for 24 h with or without PC (10−3 M) before isolatingRNA (see Materials and methods). The RNA template(0.5 �g) was used from each sample, 24 PCR cycles forindividual amplification with ELF (Fig. 1). PC blockedup-regulation of MMP-1 by BCP crystals (P<0.01) but notby IL-1�. Similar data were observed with BCP induction ofMMP-3 (data not shown).

Although their basal expression levels were relativelylow (approximately 10% of the level of MMP-1, data notshown), both BCP crystals and IL-1� stimulated the expres-sions of MMP-9 (P<0.04, Fig. 2) and MMP-13 (P<0.04,Fig. 3). PC blocked the BCP effects on MMP-9 and -13specifically (P<0.01) but had no effect on IL-1� regulation.

TIMP-1 was down-regulated by BCP crystals to 73%(P<0.03) of untreated controls. There was no significantchange in TIMP-1 expression with IL-1� treatment. PCblocked the effect of BCP crystals on TIMP-1 (Fig. 4).

420 G. Bai et al.: Basic calcium phosphate crystals

TIMP-2, another important inhibitor of MMPs, was down-regulated by BCP crystals to 66% (P<0.01) of untreatedcontrols. PC again abolished the effect of crystal on TIMP-2(Fig. 5). IL-1� did not cause any significant change inTIMP-2 expression. Western blot analysis (Fig. 6) con-firmed that BCP down-regulated both TIMP-1 and -2expression and this inhibitory effect was reversed by PC.

TIMP-2

ELF

Fig. 5. Down-regulation of TIMP-2 by BCP crystals. Photograph ofan agarose gel of PCR products of TIMP-2 with ELF (24 cycles).The DNA products are indicated. Lanes are labeled as follows: 1:BCP treatment (50 �g/ml); 2: BCP+PC (10−3 M); 3: IL-1� treat-ment (10 ng/ml); 4: IL-�+PC (10−3 M); and 5: untreated control.Computer-scanned DNA densities of TIMP-2 normalized to internalcontrols; untreated control (lane 5) was set as 100%. Bars repre-sent mean and standard deviation of three repeated experiments.

* indicates significantly different from control (P<0.01).

M 1 2 3 1 2 3

TIMP-1

TIMP-2

WESTERN BLOTTING

48hr 72hr

31

21

Fig. 6. BCP down-regulate both TIMP-1 & 2. Lanes 1, 2, and 3 are samples from 48 h while Lane 4, 5, and 6 are samples from 72 hoursconcentrated spent media. Lanes are labeled as 1: BCP treatment; 2: BCP plus PC; 3: Control untreated samples. 4: BCP; 5: BCP+PC and

6: untreated control.

Discussion

Besides confirming the earlier work on BCP crystalup-regulation of the synthesis of MMP-1, -3, and -9 inHF22,34, we demonstrate for the first time that BCP crystalsstimulate the expression of MMP-13 but down-regulateTIMP-1 and -2. Our data indicate that there is a differentialregulation of MMPs and TIMPs by BCP crystals in favor ofraising MMP levels.

Since PCR is a very sensitive method, precautions weretaken to include several controls in these studies. Theseincluded quantification of PCR methods using multiple RNAdilutions and PCR cycles (with amplification of MMP-1),ELF as an internal controls for normalizing the data, andreproduction of known positive results with IL-1� treat-ments. The up-regulation of MMP-1 and -3 by BCP crystalsserved as positive controls for the RT-PCR determinations.

MMP-13, an enzyme originally isolated and cloned inhuman carcinoma35, is expressed at a high level bychondrocytes in human osteoarthritic cartilage and has amuch higher catalytic rate (5–10-fold) than MMP-1 on typeII collagen31,35,36. Although the MMP-13 level induced byboth BCP crystals and IL-1� was low, the biological conse-quence may be highly significant.

TIMP-1 and TIMP-2, which are capable of inhibitingMMP family members10, are both down-regulated by BCPcrystals over the untreated controls. IL-1� treatment had nosignificant effect on either TIMP-1 or -2 expression. Thedown-regulation of TIMP-1 and -2 further shifts the balanceof MMP/TIMP activities in addition to the up-regulation ofMMP-1, -3, -9 and -13 by BCP crystals. The down-regulation of TIMP-2 is particularly interesting since TIMP-2was reported constitutively expressed in many systems,suggesting that basal TIMP-2 levels may be needed forcartilage integrity37.

Burger et al. has demonstrated that direct contact withstimulated T lymphocyte triggered the synthesis of PGE2and MMP-1, but not TIMP-1 in synoviocytes and dermalfibroblasts38. Divergent regulation of MMP especiallygelatinase B (MMP-9) and TIMP had also been reported indifferent cell systems, such as in human bronchial epithelialcells and rabbit aortic smooth muscle cells. In humanbronchial epithelial cells, MMP-9 mRNA level wasincreased two-fold in response to lipopolysaccharide andIL-1� while TIMP-1 mRNA remained unchanged in

Osteoarthritis and Cartilage Vol. 9, No. 5 421

response to both cytokines but clearly decreased to 70% inthe presence of TNF-�. These measurements were doneon both protein and RNA levels (RNA was also determinedby RT-PCR) and the results were in agreement with eachother17. In the case of rabbit aortic smooth muscle cells,expression of MMP-9 was increased by phorbol myristateacetate, FBS, thrombin and IL-1�, while TIMP-1 andTIMP-2 were highly expressed and their synthesis was notaffected by growth factors or cytokines18. Both casesimplied that pathological processes were related to animbalance of MMP/TIMP activities.

It has been suggested that the loss of coordinatedregulation of MMPs/TIMPs, resulting in an imbalance ofdegradative proteinases over protective TIMP activities,underlies the pathological tissue destruction in diseaseslike RA and OA15,16. In their analysis of the joint fluids ofpatients with acute ‘pyrophosphate arthropathy’ (calciumpyrophosphate dihydrate in joint fluids and OA),Lohmander et al.39 reported that these crystal-containingfluids had increased levels of collagenase (MMP-1) andstromelysin (MMP-3), and the highest ratio of MMPs: TIMPas compared to fluids from patients with other arthriticdiseases, suggesting that a high ratio of MMP/TIMP maypromote cartilage matrix degradation.

In summary, we demonstrated that BCP crystals stimu-late the expression of MMPs but down-regulate the syn-thesis of TIMPs. PC blocks both of these effects. The abilityof BCP to induce the synthesis of degradative MMPs whiledown-regulating the synthesis of the naturally occurringcounterpart TIMPs may explain the changes consistentwith a role of BCP crystal in the pathogenesis of degenera-tive changes in OA. The ability of PC to reverse bothdegradative effects of BCP crystal suggests that PC can bea potential therapeutic agent for BCP crystal depositiondiseases.

Acknowledgments

The authors gratefully acknowledge L.Wenger, A. Aguedaand M. Cruz for their outstanding technical assistance. Thiswork was supported in part by NIH AR38421-10, a VAResearch Career Scientist Award and a VA Merit Review(HSC) and NIH AR08662-33 (DSH).

References

1. Halverson PB, McCarty DJ. Basic calcium phosphate(apatite, octacalcium phosphate, tricalcium phos-phate) crystal deposition disease; calcinosis. In:Koopman WJ, Ed. Arthritis and Allied Conditions.Baltimore, MD: Williams and Wilkins 1997:212–46.

2. Carroll GJ, Stuart RA, Armstrong JA, Breidahl PD,Laing BA. Hydroxyapatite crystals are a frequentfinding in osteoarthritic synovial fluid, but are notrelated to increased concentrations of keratan sulfateor interleukin 1b. J Rheumatol 1991;18:861–6.

3. McCarty DJ, Halverson PB, Carrera GF, Brewer BJ,Kozin F. ‘Milwaukee Shoulder’: Association of micro-spheroids containing HA crystals, active collagenaseand neutral protease with rotator cuff defects. I.Clinical aspects. Arthritis Rheum 1981;24:464–73.

4. Ryan LM, Cheung HS. The role of crystals in osteo-arthritis. In: Brandt KD, Ed. Osteoarthritis RheumaticDisease of America, Vol. 25. Philadelphia: W.B.Saunders Co 1999:257–67.

5. Cheung HS. The role of crystals in articular tissuedegeneration. Current Rheumatology Reports 1999;1:128–3.

6. Cheung HS. Calcium crystal effects on the cells of thejoint: Implications for pathogenesis of the disease.Current Opinion in Rheumatology, 2000;12(3):223–7.

7. Krane SM. Mechanisms of tissue destruction in rheu-matoid arthritis. In: McCarty DJ, Ed. Arthritis andAllied Conditions. 11th ed. Philadelphia: Lea andFebiger 1989:698–714.

8. Nagase N. Activation mechanisms of matrix metallo-proteinases. Biol Chem 1997;378:151–60.

9. Powell WC, Matrisian LM. Complex roles of matrixmetalloproteinases in tumor progression. Curr TopMicrobiol Immunol 1996;213:1–21.

10. Gomez DE, Alonso DF, Yoshiji H, Thorgeirsson UP.Tissue inhibitor of metalloproteinases: structure,regulation and biological functions. Eur J Cell Biol1997;74:111–22.

11. Hammani K, Blakis A, Morsette D, Bowcock AM,Schmuttes C, Henriet P, et al. Structure andcharacterization of human tissue inhibitor ofmetalloproteinase-2 gene. J Biol Chem1996;271:25498–505.

12. Leco KI, Hayden LJ, Sharma RR, Rocheleau H,Grenberg AH, Edward DR. Differential regulationof TIMP-1 and TIMP-2 mRNA expression in normaland HA-Ras-transformed murine fibroblasts. Gene1992;117:209–17.

13. DeClerck YA, Darville MI, Eeckhout Y, Rousseau GG.Characterization of the promotor of the gene encod-ing human tissue inhibitor of metalloproteinase-2(TIMP-2). Gene 1994;139:185–91.

14. Stetler-Stevenson WG, Brown PD, Onisto M, Levy AT,Liotta LA. Tissue inhibitor of metalloproteinase-2(TIMP-2) mRNA expression in tumor cell lines andhuman tumor tissues. J Biol Chem 1990;265:13933–38.

15. Dean DD, Martel-Pelletier J, Pelletier JP, Howell DS,Woessner JF Jr. Evidence for metalloproteinase andmetalloproteinase inhibitor imbalance in humanosteoarthritic cartilage. J Clin Invest 1989;84:678–85.

16. Martel-Pelletier J, McCollum R, Fujimoto N, Obata K,Cloutier JM, Pelletier J. Excess of MMP over TIMPmay contribute to cartilage degradation in OA andRA. Lab Invest 1994;70:807–15.

17. Yao PM, Maitre B, Delacourt C, Buhler JM, Harf A,Lafuma C. Divergent regulation of 92-kDa gelatinaseand TIMP-1 by HBECs in response to IL-1 and TNF.J American Physiological Society 1997;1040-0605:L866–74.

18. Fabunmi RP, Baker AK, Murray EJ, Booth RF, NewbyAC. Divergent regulation by growth factors andcytokines of 95 kDa and 72 kDa gelatinases andtissue inhibitors or metalloproteinases-1, -2, and -3 inrabbit aortic smooth muscle cells. Biochem J1996;315:335–42.

19. Howard JE. Studies on urinary stone formation: a sagaof clinical investigation. Johns Hopkins Med J1976;139:239–52.

20. Tew WP, Malis CD, Howard JE, Lehninger AJ.Phosphocitrate inhibits mitochondrial and cytosolicaccumulation of calcium in kidney cells in vivo. ProcNatl Acad Sci USA 1981;78:5528–32.

21. Cheung HS, Sallis JD, Mitchell PG, Struve JA.Inhibition of basic calcium phosphate crystal-induced

422 G. Bai et al.: Basic calcium phosphate crystals

mitogenesis by phosphocitrate. Biochem BiophysRes Comm 1990;171:20–5.

22. Cheung HS, Sallis JD, Struve JA. Specific inhibition ofbasic calcium phosphate and calcium pyrophosphatecrystal-induction of metalloproteinase synthesis byphosphocitrate. Biochim Biophys Acta 1996;1315:105–11.

23. Nair D, Misra RP, Sallis JD, Cheung HS. Phospho-citrate inhibits a BCP and CPPD crystal-inducedmitogen-activated protein kinase cascade signaltransduction pathway. J Biol Chem 1997;272:18920–5.

24. Krug HE, Mahowald ML, Halverson PB, Cheung HS.Phosphocitrate prevents disease progression inmurine progressive ankylosis. Arthritis Rheum1993;36:1603–11.

25. Cheung HS, Story MT, McCarty DJ. Mitogenic effectsof HA and CPPD crystals on cultured mammaliancells. Arthritis Rheum 1984;27:668–74.

26. McCarthy GM, Mitchell PG, Struve JA, CheungHS. Basic calcium phosphate crystals cause coordi-nate induction and secretion of collagenase andstromelysin. J Cell Physiol 1992;153:140–6.

27. Bett JAS, Christner LG, Hall WK. Studies of thehydrogen held by solid. X11. Hydroxyapatitecatalysts. J Am Chem Soc 1967;89:5535–41.

28. Cheung HS, Story MT, McCarty DJ. Mitogenic effectsof hydroxyapatite and calcium pyrophosphatedihydrate crystals on cultured mammalian cells.Arthritis Rheum 1984;27:668–74.

29. Chomczynski AP, Sacchi N. Single-step method ofRNA isolation by acid guanidinium thoicyanate-pheno-chloroform extraction. Anal Biochem 1987;162:156–9.

30. Tamarina NA, McMillan WD, Shively VP, Pearce WH.Expression of matrix metalloproteinases and theirinhibitors in aneurysms and normal aorta. Surgery1997;122:264–72.

31. Reboul P, Pelletier JP, Tardif G, Cloutier JM, PelletierJM. The new collagenase, collagenase-3, isexpressed and synthesized by human chondrocytesbut not by synoviocytes. J Clin Invest 1996;97:2011–9.

32. Grumbles RM, Shao L, Jeffrey JJ, Howell DS. Regu-lation of rat interstitial collagenase gene expressionin growth cartilage and chondrocytes by vitamin D3,interleukin-1, and okadaic acid. J Cell Biochem1996;63:395–409.

33. van de Lest CHA, Veerkamp JH, van KuppeveltTH. ImageCalc: A microsoft WindowsY applicationfor quantitative image analysis and comparison.BioTechniques 1995;18:1051–5.

34. McCarthy GM, Macius AM, Christopherson PA, RyanLM, Pourmotabbed T. Basic calcium phosphatecrystals induce synthesis and secretion of 92 kDagelatinase (gelatinase B/matrix metalloprotease 9) inhuman fibroblasts. Ann Rheum Dis 1998;57:56–60.

35. Freije JMP, Diez-Itza I, Balbin M, Sanchez LM, BlascoR, Tolivia T, et al. Molecular cloning and expressionof collagenase-3, a novel human matrix metallo-proteinase produced by breast carcinomas. J BiolChem 1994;269:16766–73.

36. Mitchell PG, Magna HA, Reeves LM, Lopresti-MorrowLL, Yocum SA, Rosner PJ, et al. Cloning, expression,and type II collagenolytic activity of matrixmetalloproteinase-13 from human osteoarthriticcartilage. J Clin Invest 1996;97:761–8.

37. Zafarullah M, Su S, Martel-Pelletier J, DiBattista JA,Costello BG, Stetler-Stevenson WG, et al. Tissueinhibitor of metalloproteinase-2 (TIMP-2) mRNA isconstitutively expressed in bovine, human normal,and osteoarthritic articular chondrocytes. J CellBiochem 1996;60:211–7.

38. Burger D, Rezzonco R, Li JM, Modoux C, Pierce RA,Welgus HG, et al. Imbalance between interstitialcollagenase and tissue inhibitor of metalloproteinase1 in synoviocytes and fibroblasts upon direct contactwith stimulated T lymphocytes. Arthritis Rheum1998;41:1748–59.

39. Lohmander LS, Hoerrner LA, Lark MW. Metallo-proteinases, tissue inhibitor, and proteoglycan frag-ments in knee synovial fluid in human osteoarthritis.Arthritis Rheum 1993;36:181–9.