CARTILAGE NUCLEOSIDE TRIPHOSPHATE PYROPHOSPHOHYDROLASE

11. Role in Extracellular Pyrophosphate Generation and Nucleotide Metabolism

LAWRENCE M. RYAN, ROBERT L. WORTMANN, BARBARA KARAS, and DANIEL J

Extracellular generation of inorganic pyrophos- phate (PPi) in cartilage organ culture is markedly augmented by ATP. ATP, not an ATP metabolite (ADP, AMP, adenosine), is necessary for this augmentation. Excess PPi production is effectively blocked by known inhibitors of nucleoside triphosphate (NTP) pyro- phosphohydrolase (EDTA, EGTA, dithiothreitol). Ex- cess 32P-PPi is generated directly from g2P-ATP by cartilage, as substrate and product have similar specific activities. These findings strongly favor ecto-NTP pyro- phosphohydrolase as the source of extracellular PPi generation in the presence of NTP. Additionally, active nucleotide and nucleoside catabolism is demonstrated in these cartilage organ cultures.

Calcium pyrophosphate dihydrate (CPPD) crys- tal deposition disease is defined by the presence of CPPD crystals in articular fluid or tissues. These crystals often cause an inflammatory arthritis (pseudo-

From the Section of Rheumatology, Department of Medi- cine, The Medical College of Wisconsin, Milwaukee, Wisconsin, and Wood Veterans Administration Medical Center, Wood, Wis- consin.

Supported by USPHS grants AM-18074-1 1 and AM-31395; a Merit Review grant from the Veterans Administration; and a grant from the Wisconsin Chapter of The Arthritis Foundation.

Lawrence M. Ryan, MD, FACP: Associate Professor of Medicine, Section of Rheumatology, and recipient of New Investi- gator Award, NIADDK; Robert L. Wortmann, MD, FACP: Assist- ant Professor of Medicine, Section of Rheumatology ; Barbara Karas, MA: Research Associate, Section of Rheumatology; Daniel J. McCarty, Jr., MD, FACP: Will and Cava Ross Professor and Chairman, Department of Medicine, Medical College of Wisconsin.

Address reprint requests to Lawrence M. Ryan, MD, Section of Rheumatology, Medical College of Wisconsin, 8700 W. Wisconsin Avenue, Milwaukee, WI 53226.

Submitted for publication July 9, 1984; accepted in revised form October 16, 1984.

McCARTY. JR.

gout) and even more frequently, are associated with a chronic degenerative arthritis in a distribution of af- fected joints different from that of primary osteoarthri- tis. The inorganic pyrophosphate (PPi) which partici- pates in crystal formation appears to originate from articular cartilage itself, although the mechanism of its generation is unclear (1,2).

Tenenbaum et a1 ( 3 ) have reported increased activity of nucleoside triphosphate pyrophosphohy- drolase (NTPPPHase) in Triton X-100 extracts of degenerated articular cartilage from patients with CPPD deposition in comparison with extracts from osteoarthritic or normal cartilage. This enzyme gener- ates PPi from the nucleoside triphosphates in the reaction:

nucleoside triphosphate (NTP) e nucleoside monophosphate (NMP) + PPi

Tenenbaum et a1 also noted increased 5’nucleotidase (5’NT) and decreased inorganic pyrophosphatase (PPiase) in CPPD cartilage.

Abnormalities of these 3 enzymes theoretically could lead to increased PPi accumulation in tissue by the following mechanisms: 1) increased NTP pyro- phosphohydrolase activity could generate increased amounts of PPi; 2) removal of the NMP coproduct by S’NT could favor NTP degradation; and 3 ) low PPiase activity could decrease PPi removal. We have charac- terized NTPPPHase as an ecto-enzyme in canine cartilage organ culture (4). Therefore, in the presence of suitable substrate, the nature and concentration of which are still obscure, PPi could be generated by this enzyme extracellularly in cartilage.

The PPi which appears in the presence of NTP may result from 2 processes other than ecto-

Arthritis and Rheumatism, Vol. 28, No. 4 (April 1985)

414 RYAN ET AL

NTPPPHase. First, since spontaneous “background” extracellular PPi elaboration occurs in the absence of exogenous NTP (1,2), it is possible that part or all of the PPi liberated in the presence of NTP results from augmented endogenous secretion by chondrocytes stimulated by the NTP or by an NTP metabolite. Second, extracellular ATP has been demonstrated to increase the permeability of some cell types (5,6) and cause degranulation of others (7), thus providing an- other route by which extracellular PPi could appear following exposure of cells to NTP.

We sought to determine the role of NTPPPHase in NTP-stimulated extracellular PPi generation. All excess PPi generated by cartilage in the presence of exogenous ATP was derived directly from ATP via NTPPPHase since 1) ATP metabolites did not stimu- late significant PPi elaboration; 2) elaboration of ex- cess PPi in the presence of ATP is substantially depressed by the addition of known NTPPPHase in- hibitors; and 3) 32P-PPi elaborated by chondrocytes in the presence of g2P-ATP has a specific activity simi- lar to that of the initial ”P-ATP. Additionally, extra- cellular nucleotide catabolism occurred in cartilage, as has been demonstrated in other tissues.

MATERIALS AND METHODS Reagents. Na4P207. 10 H20, ATP, ADP, AMP,

adenosine, dithiothreitol, theophylline, adenosine diphos- phonate (AMPCP), dipyridamole, cysteine. and tetramisole were purchased from Sigma Chemical Co. (St. Louis, MO). Erythro-9-[3-(2-hydroxy-nonyl)] adenine (EHNA) was pur- chased from Burroughs Wellcome (Research Triangle Park, NC); g2P-ATP was obtained from New England Nuclear (Boston, MA). All reagents were prepared in doubly deion- ized, glass-distilled, charcoal-filtered water.

Assays for PPi and ATP metabolites. PPi was deter- mined by a modification of the method of Cheung and Suhadolnik (8). Briefly, I4C-UDPG was quantitatively con- verted enzymatically to 14C-glucose- 1-PO4 and UTP in the presence of PPi. Differential charcoal adsorption of 14C before and after reaction in the assay was directly related to PPi concentration. Appropriate standard PPi determinations at concentrations from 0.5-20 pM were made with each assay. Standard stock PPi solutions (16 mm) were assayed for inorganic phosphate content by the method of Fiske and SubbaRow (9) before and after hydrolysis by yeast inorganic pyrophosphatase or by acid hydrolysis as described previ- ously (10). This is necessary since stock Na4P207 . 10 H2O powder readily gains or loses waters of hydration (lo), which makes calculations of molarity by weight inaccurate.

Concentrations of ATP and its metabolites were determined by high performance liquid chromatography (HPLC) (Varian 5300, Sunnyvale, CA). Samples were pre- pared for analysis using 6% trichloroacetic acid and 0.5M tri- N-octyl amine in freon as described by Chen et al (11).

Samples were analyzed immediately or stored at -20°C. ATP, ADP, AMP, and IMP were chromatographed with a Micropak AX-10 anion exchange column and a gradient of 0.01M KH2P04, pH 2.85 (buffer A) and 0.75M KH2P04, pH 4.4 (buffer B) at a flow rate of 2.0 ml/minute. A linear gradient was developed over 37.5 minutes at 5% B to 100% B followed by elution with 100% B for 12.5 minutes (12). Nucleosides and bases were quantified with a MCH-10 reverse phase column isocratically using 0.05M KH2P04, pH 6.0 (buffer C) for 11 minutes and then buffer C in 25% methanol (volume/volume) for 15 minutes at a flow rate of 2.0 ml/minute (13).

Cartilage organ cultures. Hyaline articular cartilage was obtained from the femoral condyles and fibrocartilage from the menisci of freshly killed adult mongrel dogs and transported in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum (FCS). Cartilage was minced with a scalpel using sterile techniques into =5-mg fragments and washed 3 times with DMEM over 1 hour prior to study in order to remove debris and FCS, which contains NTP pyrophosphohydrolase activity. Con- trol cartilage specimens were incubated in a sevenfold (volume/wet weight) volume of DMEM containing 50 mM Tris, pH 7.2 at 37°C in a 10% C02 atmosphere for 30-240 minutes, as previously described (2). Experimental plates contained cartilage from the same animal incubated with identical media to which I mM ATP, ATP metabolites, and/ or inhibitors of NTP pyrophosphohydrolase had been added in various combinations.

At the conclusion of incubations, media was immedi- ately removed and centrifuged. A portion of the supernatant was frozen at -20°C for PPi analysis. The remainder was extracted for HPLC. PPi production was expressed as nmoles/hour/mg wet weight of cartilage.

Effect of ATP metabolites on PPi production. To determine whether a metabolite of ATP is indirectly stimu- lating PPi extrusion from cartilage, canine cartilage was incubated in the presence of 1.0 mM adenosine, AMP, or ADP in single experiments. With each metabolite a compan- ion plate containing cartilage from the same animal was incubated in 1 mM ATP. Results were expressed as:

PPiM (nmoles/hour/lOO mg cartilage) % = x 100

PPiATp (nmoles/hour/lOO mg cartilage)

Inhibitors of nucleoside production. Nucleosides are produced during incubation of cartilage with ATP (see Results). To determine its effect upon PPi elaboration, nucleoside production was inhibited by 0.5 mM AMPCP (inhibits 5’NT), by 0.5 mM AMPCP plus 1 mM AMP (further inhibits 5‘NT), by 10 pM dipyridamole (inhibits transport of adenosine into cells where it is further catabolized), or by EHNA (inhibits intracellular adenosine deaminase). Por- tions of cartilage from the same dog were incubated in DMEM containing ATP or ATP plus inhibitor in single experiments. Results were expressed as:

PPiI (nmoles/hour/lOO mg cartilage) PPiATp (nmoles/hour/lOO mg cartilage)

% PPi generation =

where PPiI = PPi produced on the plate with inhibitor plus 1

x 100

CARTILAGE NTP PYROPHOSPHOHYDROLASE 415

mM ATP, and PPiATp = PPi produced on the plate with ATP alone.

Inhibitors of NTP pyrophosphohydrolase. The follow- ing were studied to assess their effect upon PPi generation by ATP: (a) known inhibitors of NTPPPHase (14) (10 mM EGTA, 10 mM dithiothreitol, 5 mM EDTA and 1 mM cysteine); (b) an inhibitor of alkaline phosphatase which does not affect NTPPPHase (14) ( 2 mM tetramisole); and (c) an inhibitor of Na', K+ ATPase (15) (0.1 mM ouabain). PPi production ip the presence of ATP and ATP plus inhibitor was measured on 2 aliquots of cartilage from the same dog. Results were expressed as:

% PPi production =

PPi produced per hour with inhibitor + 1 mM ATP PPi produced per hour with 1 mM ATP

x 100

Specific activity of 32P-PPi produced in the presence of "P-ATP. If the PPi generated by cartilage in the presence of ATP is derived entirely from ATP via NTP yrophosphohy-

from d2P-ATP substrate should be identical to that of the starting 32P-ATP. Production of endogenous unlabeled PPi would dilute 32P-PPi produced via NTP pyrophosphohydro- lase, thereby lowering the specific activity of the total 32P- PPi produced. Organ cultures of articular cartilage were incubated with 1 mM f2P-ATP in DMEM. The specific activity of the starting P-ATP was determined as follows: ATP was measured by the method of Bucher (16) as modi- fied by Adams (17). 32P-ATP counts were determined in 100 pl of media after precipitation of contaminating 32P inorganic phosphate (Pi) (18) by measuring Cerenkov radiation in 0.5 N HCl in a Packard Tri-Carb scintillation counter with a 50- 1,000 window at 50% gain. The specific activity of the media PPi at the conclusion of the organ culture incubations was determined as follows: PPi concentration was measured as previously described. 32P-PPi counts were determined by differential counting of 32P-Pi precipitated by the method of Sugino and Miyoshi (18) before and after incubation with

drolase, then the specific activity of the P ' P-PPi produced

Table 1. Appearance of PPi and ATP catabolites in media of cartilage specimens incubated with ATP*

Hours of incubation

Substrate 0

ATP ADP AMP Adenosine Inosine H ypoxanthine PPi

882 f 257 53 f 20 12 2 9 1 1 f 7

<2 <2

1.8 f 1.5

2 4

586 t 394 435 2 303 85 f 17 102 2 44 19 * 3 20 t 9 83 f 25 120 f 39 74 f 49 181 f 137 54 t 23 136 2 74

182 2 75 275 2 71

* Values represent media concentration (pw, mean * SD (n = 5 ) . PPi = inorganic pyrophosphate. For comparison, plates containing cartilage but no ATP routinely increased media PPi concentration by 2 p M after 4 hours. ATP incubated in media alone exhibited no spontaneous breakdown to PPi during a 4-hour incubation.

Table 2. and fibyocartilage*

NTP pyrophosphohydrolase activity of articular hyaline

Dog 1 Dog 2 Dog 3

Inner V, meniscus 105 68 192 Outer % meniscus 105 115 24 1 Femoral condyle cartilage 84 111 -

* Inorganic pyrophosphate produced/hour/lOO mg wet weight of cartilage incubated in DMEM plus ATP, in nmoles.

Type I11 yeast inorganic PPiase (Sigma) in a mixture contain- ing 50 p1 media, 50 pl of 100p,lml PPiase, and 100 pl of 13 mM MgC12, 15 mM Tris C1,5 mM EDTA, pH 7.2, for 1 hour at room temperature.

RESULTS When hyaline articular cartilage was incubated

in the presence of 1 mM ATP, PPi was liberated linearly over the next 4-hour period, or until nearly 300 pM PPi had accumulated in the media (Table 1). Both articular hyaline and fibrocartilages generated PPi (Table 2). Concurrently, nucleoside and nucleotide coproducts appeared in the media (Table 1). Mean measurements for 5 consecutive incubations of carti- lage specimens from different dogs are summarized in Table 1. PPi generation from ATP varied widely and raised media PPi concentrations between 75 and 250 pM after 2 hours of incubation (28-90 nmoles PPi generated/hour/lW mg wet weight cartilage). In each incubate, ATP degradation exceeded PPi accumula- tion, implying avenues of degradation other than the pyrophosphohydrolase reaction. The generation of significant levels of ADP reinforces this inference. The difference between ATP degradation and PPi produc- tion cannot be explained by hydrolysis of a portion of the newly-formed PPi since 32P-PPi added to the media was not substantially hydrolyzed, precipitated, or ad- sorbed during the incubation period (data not shown).

Addition of ATP metabolites did not cause the same increase in PPi elaboration into the cartilage culture media (Table 3) as did ATP. However, minor increases in PPi generation were observed with CAMP, adenosine, and especially ADP. Addition of com- pounds known to interfere with nucleoside and nucleo- tide metabolism did not prevent the enhanced PPI liberation seen in the presence of ATP (Table 4). The inhibitors effectively slowed primary nucleotide catab- olism at the concentrations noted: AMPCP increased AMP accumulation 146% by inhibiting 5'NT; dipyrida- mole increased extracellular adenosine accumulation 117% by preventing adenosine ingress into cells; and

416 RYAN ET AL

Table 3. generation by canine hyaline articular cartilage

Effect of ATP metabolites upon pyrophosphate (PPi) Table 5. Inhibition of cartilage NTP pyrophosphohydrolase

% PPi production* Inhibitor

No. Dithiothreitol (10 mM) 13 Additive % PPi generated*

ADP 4.8 2 Ouabain (0.1 mM) 63 AMP 0.3 4 EDTA (10 mM) 1 Adenosine 0.8 1 EGTA (5 mM) 5 CAMP 1.8 1 Tetramisole (2 mM) 81 None 0.3 2 Cysteine ( 1 mM) 80

PPiM * % = - ; PPiM = PPi produced by plate containing 1 mM

PPiATP metabolite; PPiATp = PPi produced by companion plate containing 1 mM ATP.

EHNA decreased inosine accumulation 94% by inhib- iting adenosine deaminase.

We studied the effect produced by addition of NTPPPH inhibitors, upon ATP-induced PPi genera- tion. EDTA (10 mM) and EGTA (5 mM) inhibited PPi production by 99 and 95%, respectively (Table 5). Incubation of cartilage with these chelators for one- half hour prior to adding ATP maximized inhibition. Dithiothreitol inhibited 90% of NTPPPHase activity in cartilage extracts (14) and inhibited PPi production by 87% in our study. Other substances tested, including ouabain, tetramisole, and cysteine, were weak PPi production inhibitors.

If PPi is generated entirely by cleavage of the a- p pyrophosphate bond of ATP, then the specific activity of g2P-ATP added to cartilage media and 32P- PPi generated from it should be identical. Results of studies of 5 femoral condylar cartilage specimens from different dogs are summarized in Table 6. Ninety- seven -+ 20% (mean & SD) of the PPi produced was radiolabeled, indicating a direct origin from ATP via NTP pyrophosphohydrolase.

DISCUSSION Howell et al first reported spontaneous extra-

cellular release of PPi by human osteoarthritic carti-

Table 4. presence of inhibitors of nucleotide and nucleoside catabolism

Pyrophosphate (PPi) generation from cartilage in the

Additive % PPi generated*

1 mMATP I00 125 p M AMPCP + I mM ATP 1 I9 500 p M AMPCP + I mM ATP 97 500 p M AMPCP + 1 mM AMP + 1 mM ATP 85 5 I.M EHNA + 1 mM ATP 92 10 pM dipyridamole + 1 mM ATP 1 I5

PPi

PPiATP *%=I. , PPi, = PPi produced by plate with inhibitor + 1 m M

ATP; PPiATp = PPi produced on plate with I mM ATP alone. AMPCP = adenosine diphosphonate.

* % pyrophosphate (PPi) production =

PPi production in presence of inhibitor + 1 mM ATP

PPi production in presence of 1 mM ATP x 100.

lage in culture (l), and we have previously demonstrat- ed that normal articular hyaline and fibrocartilage from several species spontaneously elaborates PPi (2). When any one of a number of nucleoside triphosphates is added to the incubation media, PPi generation is markedly augmented (4). This augmentation presum- ably results from the action of ecto-NTP pyrophos- phohydrolase, which cleaves nucleoside triphosphates into nucleoside monophosphates and PPi (3,4,14).

We have further studied the interaction of ATP and cartilage to determine whether PPi is derived entirely or partially via NTPPPHase or whether PPi production is an independent process resulting from chondrocyte stimulation by ATP or an ATP metabo- lite.

During in vitro incubation of cartilage with exogenous ATP, PPi was generated linearly, which confirmed our prior results (4). ATP is concomitantly catabolized to nucleotides, nucleosides, and purine bases. Since ATP degradation always exceeded PPi accumulation, and since ADP appeared in the media, it is clear that extracellular ATP is also catabolized by an ATPase, nonspecific phosphatase, or a protein kinase. Similar amounts of PPi generation were not observed when nucleoside or nucleotide catabolites of ATP were substituted for ATP. Moreover, inhibition of various steps in ATP degradation with AMPCP, dipy-

Table 6. Specific activity of 32P-PPi generated in the presence of ~'=P-ATP

PPi Hours Specific activity PPi

Don ( W M ) incubated Specific activity ATP* - -

1 40 2 0.66 2 50 2 1.16 3 64 2 1.16 4 137 2 I .04 5 40 1 0.83

* 32P-PPi counts per nmole pyrophosphate (PPi) divided by 32P counts per nmole ATP. See Table I for definitions.

CARTILAGE NTP PYROPHOSPHOHYDROLASE 417

ridamole, or EHNA did not interfere with PPi produc- tion. These two latter lines of evidence indicate that ATP and not an ATP derivative is essential for aug- mented PPi production.

Muniz et a1 (14) reported that divalent cation chelators and dithiothreitol are excellent inhibitors of NTPPPHase in human cartilage extracts. EDTA, EGTA, and dithiothreitol all blocked excess PPi gener- ation in our organ cultures, further supporting the role of NTPPPHase in PPi generation. However, the ef- fects of these inhibitors are highly nonspecific in living cells and results must be interpreted cautiously. Tetra- misole and ouabain, which inhibit other phosphatases, only moderately affected PPi generation from ATP. Cysteine was not an effective inhibitor in canine cartilage organ culture, in contrast to findings in detergent extracts of human cartilage (14). These variances may be a result of species difference or represent the enhanced inhibition by cysteine in deter- gent extracts in comparison with native preparations, as described for cartilage extracellular fluid alkaline phosphatase (19).

Studies of the specific activity of 32P-PPi gener- ated in the presence of y 32P-ATP showed that excess extracellular PPi is not derived from unlabeled intra- cellular or extracellular pool. The similar specific activities of 32P-PPi product and the 32P-ATP substrate indicate that the majority of PPi is formed from the y, and probably p, phosphate of ATP.

The function of this enzyme is unclear. It has been identified in multiple nonarticular tissues (19,20), as well as in cartilage (3,4,14,21) and matrix vesicles (22-24). Flodgaard and Torp-Pedersen hypothesized that it represents a membrane-associated, ATP-depen- dent, calcium pump (25). Its ecto-position makes this function less likely. It may represent a means of salvaging purines which escape from damaged cells, as suggested by Tran-Thi et a1 (26). Our data support this possibility by clearly demonstrating that in cartilage, nucleoside triphosphates can be converted sequential- ly to nucleoside monophosphates, then to nucleosides which gain access to cells, and finally to purine bases which may be re-utilized (Table 1).

The significance of NTP pyrophosphohydrolase in the pathogenesis of CPPD crystal deposition disease is uncertain, since a naturally occurring extracellular substrate has not been identified. It seems likely that exogenous ATP leads directly to PPi production via NTP pyrophosphohydrolase. A concerted effort to identify possible natural substrates seems justified. Perhaps chondrocytes may become “leaky” under certain physiologic or pathologic conditions, much as

endothelial cells or platelets release ATP in response to trauma or enzymatic insult (27). Leaked substrate could be converted to PPi which might accumulate in areas relatively deficient in pyrophosphatase, as dis- cussed by Howell et a1 (24).

Tenenbaum et a1 (3) reported and Muniz et a1 (14) confirmed an excess of 5’NT in extracts of carti- lage from patients with CPPD deposition. They hy- pothesized that 5’NT might accelerate AMP removal, thereby favoring further PPi production (3). Our stud- ies indicate that in the presence of supraphysiologic concentrations of AMP and a specific inhibitor of 5’NT, PPi generation from ATP is not appreciably altered (Table 4). Assuming that human and canine cartilage behave similarly, this finding may diminish the potential import of 5‘NT in influencing PPi produc- tion. However, 5’NT may be an important modulator at lower concentrations of NTP.

NTP pyrophosphohydrolase activity in normal canine cartilage is adequate to hydrolyze large amounts of substrate at physiologic pH. The excess NTP pyrophosphohydrolase noted in extracts of carti- lage from patients with CPPD deposition disease may, therefore, be less important than postulated. Perhaps an excess of substrate is more crucial in the generation of extracellular PPi in CPPD deposition disease.

ACKNOWLEDGMENTS We wish to acknowledge the helpful suggestions of

Dr. Herman Cheung, the donation of canine cartilage by Dr. Garrett Gross, and the manuscript preparation by Dorothy Shrode.

REFERENCES 1. Howell DS, Muniz 0, Pita JC, Enis JE: Extrusion of

pyrophosphate into extracellular media by osteoarthritic cartilage incubates. J Clin Invest 56: 1473-1480, 1975

2. Ryan LM, Cheung HS, McCarty DJ: Release of pyro- phosphate by normal mammalian articular hyaline and fibrocartilage in organ culture. Arthritis Rheum 24: 1522- 1527, 1981

3. Tenenbaum J , Muniz 0, Schumacher HR, Good AE, Howell DS: Comparison of pyrophosphohydrolase ac- tivities from articular cartilage in calcium pyrophosphate deposition disease and primary osteoarthritis. Arthritis Rheum 24:492-500, 1981

4. Ryan LM, Wortmann RL, Karas B, McCarty DJ Jr: Cartilage nucleoside triphosphate (NTP) pyrophospho- hydrolase. I. Identification as an ecto-enzyme. Arthritis Rheum 27:404-409, 1984

5. Stewart C, Gosic G, Hempling H: Effect of exogenous ATP on the volume of TA? ascites cells. J Cell Physiol 73:125-132, 1969

418 RYAN ET AL

6. Makan N: Induction of permeability change and restora- tion of membrane permeability barrier in transformed cell cultures. Exp Cell Res 114:417-427, 1978

7. Dahlquist R, Diamant B, Kruger P: Increased perme- ability of the rat mast cell membrane to sodium and potassium caused by extracellular ATP and its relation- ship to histamine release. Int Arch Allergy Appl Im- munol 46:655-675, 1974

8. Cheung CP, Suhadolnik RJ: Analysis of inorganic pyro- phosphate at the picomole level. Anal Biochem 83:61- 63, 1977

9. Fiske CH, SubbaRow Y: The colorimetric determina- tion of phosphorus. J Biol Chem 66:375-400, 1925

10. Ryan LM, Kozin F, McCarty DJ: Quantification of human plasma inorganic pyrophosphate. I. Normal val- ues in osteoarthritis and calcium pyrophosphate dihy- drate crystal deposition disease. Arthritis Rheum

1 1 . Chen SC, Brown PR, Rosie DM: Extraction procedures for use prior to HPLC nucleotide analysis using micro- particle chemically bonded packings. J Chromatogr Sci

12. Edelson EH, Lawless JG, Wehr CT, Abbott SR: Ion- exchange separation of nucleic acid constituents by high-performance liquid chromatography. J Chromatogr

13. Hartwick RA, Brown PR: Evaluation of tnicroparticle chemically bonded reversed-phase packings in high- pressure liquid chromatography analysis of nucleosides and their bases. J Chromatogr 126:679-691, 1976

14. Muniz 0, Pelletier J-P, Martel-Pelletier J , Morales S, Howell DS: NTP pyrophosphohydrolase in human chondrocalcinotic and osteoarthritic cartilage. I. Some biochemical characteristics. Arthritis Rheum 27: 186- 192, 1984

15. Manery JF, Dryden EE: Ecto-enzymes concerned with nucleotide metabolism, Physiological and Regulatory Functions of Adenosine and Adenine Nucleotides. Edit- ed by HP Baer, GI Drummond. New York, Raven Press, 1981, pp 323-339

16. Bucher T: Uber ein phosphatubertragendes garungsfer- meut. Biochim Biophys Acta 1:292, 1947

22~886-891, 1979

15:218-221, 1977

174~409-419, 1979

17. Adams H: Adenosine 5'-triphosphate determination with phosphoglycerate kinase, Methods of Enzymatic Analysis. Edited by HU Bergmeyer. New York, Aca- demic Press, 1963, pp 539-543

18. Sugino Y, Miyoshi Y : The specific precipitation of orthophosphate and some biochemical applications. J Biol Chem 239:2360-2364, 1964

19. Howell DS, Blanco L, Pita JC, Muniz 0: Further characterization of a nucleational agent in hypertrophic cell extracellular cartilage fluid. Metab Bone Dis Relat Res 1:155-159, 1978

20. Torp-Pedersen C, Flodgaard H, Saermark T: Studies on a Ca*+-dependent nucleoside triphosphate pyrophos- phohydrolase in rat liver plasma membrahes. Biochim Biophys Acta 571:94-104, 1979

21. Cartier P, Picard J: La mineralization du cartilage ossi- fiable. 111. La mecanisme de la reaction ATP asique du cartilage. Bull Soc Chem Biol 37: 1159-1 176, 1955

22. Hsu H: Purification and partial characterization of ATP pyrophosphoh ydrolase from fetal bovine epiph yseal car- tilage. J Biol Chem 258:3463-3468, 1983

23. Siege1 S, Hummel C, Carty P: The role of nucleoside triphosphate pyrophosphohydrolase in in vitro nucleo- side triphosphate-dependent matrix vesicle calcification. J Biol Chem 258:8601-8607, 1983

24. Howell DS, Martel-Pelletier J , Pelletier J-P, Morales S, Muniz 0: NTP pyrophosphohydrolase in human chon- drocalcinotic and osteoarthritic cartilage. 11. Further studies on histologic and subcellular distribution. Arthri- tis Rheum 27: 193-199, 1984

25. Flodgaard H, Torp-Pedersen C: A calcium-dependent adenosine triphosphate pyrophosphohydrolase in plas- ma membrane from rat liver. Biochem J 171:817-820, 1978

26. Tran-Thi T, Phillips J, Schulze-Specking A, Rosenack J, Decker K: Properties and biosynthetic connection of the nucleotide pyrophosphatases of rat liver plasma mem- brane and endoplasmic reticulum. Hoppe-Selyers Z Physiol Chem 362:305-316, 1981

27. Pearson JD, Gordon JL: Vascular endothelial and smooth muscle cells in culture selectively release ade- nine nucleotides. Nature 281:384-386, 1979