Introduction Crystal-related arthropathies

Ann Rheum Dis (1983), 42, Supplement p 1

Introduction

Crystal-related arthropathiesPAUL DIEPPE, MICHAEL DOHERTY, DIANA MACFARLANE

From the University Department ofMedicine, Bristol Royal Infirmary, Bristol BS2 8HW

Historical introduction

Antoni Van Leuwenhoek was the firstperson to identify crystals in relation toa diseased joint. Using his earlymicroscope he saw and drew crystalsderived from the gouty tophus of afriend in 1679.' Sir Alfred BaringGarrod first proposed that crystalsmight cause inflammation; the thirdand fourth of his 10 propositions ongout state: 'The deposit is crystallineand interstitial ... (and) may belooked upon as the cause and not theeffect of the gouty inflammation'.' Hisand his pupil Freudweiler then carriedout the first experiments oncrystal-induced inflammation at theturn of the century.3 Radiologicalevidence of crystalline deposits in andaround the joints was noted soon afterthe introduction of radiology todiagnostic medicine,4 and reports ofinflammation in relation to thesedeposits appeared early in the 20thcentury.' 6

In spite of these early observationslinking crystals to joint disease, themodern history of this subject issurprisingly short; many of thehistorical observations noted wereonly rediscovered when Hollander,McCarty, and others started to explorethe subject in depth in the early 1960s.Examination of synovial fluids bypolarised light microscopy firstestablished that monosodium uratemonohydrate crystals were a feature ofgout,7 and then led to the discovery ofcalcium pyrophosphate dihydratecrystals in 'pseudogout'.' 9 Thepseudogout syndrome was then linkedwith the earlier discovery of familialchondrocalcinosis by Zitnan andSitaj."' A rapid expansion ofexperimental work followed thesediscoveries. Crystals were shown to becapable of reproducing the acute

synovitis of gout and pseudogout" andthe phlogistic properties of the crystalswere demonstrated in vitro.'2 Asimple, elegant model ofcrystal-related arthropathies emerged(Fig. la).

1(a)SOLUTE EXCESS

FORMATION OF CRYSTALS(urate of pyrophosphate)

ACUTE SYNOVITIS(gout, pseudogout)

Fig. 1(a) The simple concept ofcrystal-related arthropathies developedin the 1 960s.

During the past decadetechnological advances have helpedproduce a further expansion ofknowledge, if not of understanding.Sophisticated analytical techniqueshave identified a far wider range ofcrystal species in joint tissue than waspreviously contemplated.'3- i6 Inaddition, the potential complexity ofthe interaction between crystalsurfaces and biological systems hasbeen explored, 1718 and therelationships between joint diseaseand crystal deposition have beenre-examined critically .19 20

Consequently, present concepts of thepathways involved in a crystal-relatedarthropathy are less clear cut thanthose of two decades ago (Fig. lb).l(b) SOLUTE LOCAL TISSUE

EXCESS \ CONDITIONS

Activators/Inhibitors

'IDISSOLUTION f- --CRYSTAL --- ASYMPTOMATIC

FORMATION DEPOSIT

/ 1s \InSflamation ?Otber eochanisas?

TISSUE DAMAGE

Fig. 1(b) Some ofthe pathways nowthought to be involved in crystal-relatedjoint disease.

Problems

The fact that crystals are often presentin diseased joints is not in dispute, butthe relevance of that finding is beingquestioned. Four fundamentalproblems have arisen.

(I) THE IDENTIFICATION OFCRYSTALSTable 1 lists the increasing number ofcrystal species and morphologies that

Table 1 Some ofthe crystalsidentified in human articular andperiarticular tissue

Monosodium urate monohydrateUrate spherulitesUltrasmall urate crystalsMonoclinic calcium pyrophosphate

dihydrateTriclinic calcium pyrophosphate dihydrateUltrasmall pyrophosphate crystalsHydroxyapatiteCarbonate apatiteOctacalcium phosphateApatite spherulitesDicalcium phosphate dihydrateCalcium triphosphateCalcium carbonateCalcium oxalateCholesterolLiquid lipid crystalsMixtures of crystals

have been identified in humanarticular tissue. Faced with such a listand appreciating the wide range ofdifferent techniques used for tissuesampling, particle extraction, andcrystal identification, the sensitivityand selectivity of the methods usedmust be established. Errors andartefacts are easily introduced by themethods, and it is easier to find aparticle than to say either what it is orwhen it appeared and where it camefrom.

copyright. on July 25, 2022 by guest. P

rotected byhttp://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.42.Suppl_1.1-a on 1 January 1983. D

ownloaded from

http://ard.bmj.com/

Suppl p 2 Annals of the Rheumatic Diseases

It is almost impossible to exclude thepresence of crystals in any individualpatient.

(II) THE ORIGIN OF CRYSTALSThe body contains a number ofcomplex systems designed to activatecrystal formation where it isbeneficial-for example, bones andteeth-and inhibit it where it isundesirable ('high risk' areas such asexcretory organs contain crystalpoisons-for example, salivary andurinary inhibitory proteins).Generalised or localised metabolicabnormalities may cause a soluteexcess that can overcome inhibitionand lead to extensive crystallisation.Examples include patients undergoinghaemodialysis with high phosphateconcentrations who deposit calciumphosphates in a variety of sites(generalised solute excess) andpatients with hyperuricosuria whodevelop renal stones (localised soluteexcess). In other crystal depositiondiseases a local tissue abnormalityseems to be more important thansolute excess as, for example, in thedystrophic calcification developing intuberculous lesions.

In joint diseases local or generalisedsolute excess may occur-for example,localised excess of inorganicpyrophosphate in pyrophosphatearthropathy and hyperuricaemia ingout. It is often difficult, however, toestablish true solute concentrationsbecause of the complexity of thebiological system. Joint damage mayalso lead to removal of naturalinhibitors of crystallisation orintroduce local promoters of crystalnucleation. Small local changes intemperature, pH, and theconcentration of a variety of ions couldalso be caused by joint damage andcould promote or inhibit crystalformation.There could therefore be complex

interactions between general and localmetabolic factors and localised jointdamnage, leading to the origin of crystaldeposits.

(III) CRYSTAL-INDUCED TISSUE

DAMAGECrystal deposition is associated withtwo types of joint disease: acute selflimiting att8cks of synovitis, andchronic destructive joint changes.

Crystal-induced inflammation isprobably one of the most widely

investigated and best understoodpathogenic mechanisms in rheumaticdiseases. It is, however, far from beingfully understood. The relevance ofcrystal surfaces, crystal-proteininteractions, crystal-cell interactions,and the release of a variety ofmediators and modulators ofinflammation is still being explored.The possible mechanisms whereby

crystals might contribute to chronicdestructive joint diseases have hardlybeen investigated. The effects ofcrystal-induced surface wear and ofcrystal deposition on the complianceand other mechanical properties ofcartilage and soft tissues remainunexplored. Several recentdevelopments, however, indicate thatcalcium-containing crystals may have awider range of biological and cellulareffects than was previouslyappreciated. They may, for example,cause release of destructive enzymessuch as collagenase from synovial cellsin culture, and this may be onemechanism involved in jointdestruction. Rapid developments inthis field may be expected.

(IV) RELATIONSHIP BETWEEN

CRYSTALS AND JOINT DISEASE

Most investigators have assumed thatcrystal deposition is pathological and acause of disease. There are now manyreasons to doubt this basic assumption.

Figure 2 summarises some of the

(1)X -> JOINT DISEASE

CRYSTALS

(3)JOINT DISEASE

i CRYSTALS

(2)JOINT DISEASE

X - CRYSTALS

possible relationships between crystaldeposition and joint disease. Crystaldeposits may be found inasymptomatic, otherwise normal,joints. Chondrocalcinosis is often achance radiological finding, and uratecrystals have recently been found freein joint fluids in the absence of anyevidence of inflammation. Theidentification problems mentionedand the paucity of data on normal jointtissue, add to the difficulty inestablishing the true relationshipbetween crystals and joint disease.The experimental data make a

compelling case for acute synovitisbeing 'caused' by crystals in somepatients. It remains to be seen why thisdoes not always happen and whatlimits attacks. The relationshipbetween chronic joint damage andcrystals may be revealed by carefulprospective studies of patients, andmay depend on interactions betweentissue damage and crystallisation.

Conclusions

Crystal deposition is a feature of manyacute and chronic disorders of joints.The mechanisms involved have beenexplored only in the past 20 years.Improved technology has led to theidentification ofmany new crystals andmany new problems. The proceedingsof this symposium indicate our ideas,and some of the main subjects forresearch in 1982. We expect that thiswork will soon be outdated by furtheradvances.

We would like to thank the Arthritis and(4)X > JOINT DISEASE Rheumatism Council for Research for

financial support.

Y CRYSTALS

(5) JOINT DISEASE

CRYSTALS

Fig. 2 Some of the possiblerelationships between crystaldeposition and joint disease. (1) Jointdisease causes crystallisation. (2)Crystals cause joint disease. (3) Crystaldeposits are a by-product ofthe processcausing arthritis. (4) Crystal depositionandjoint disease are independent, and a

chance relationship exists. (5) Jointdamage and crystal deposition interacttogether.

References

1 Van Leuwenhoek A. See McCarty D J.Pathogenesis and treatment of crystal-produced inflammation. In: McCarty D J,ed. Arthritis and allied conditions. 9thed. New York: Lea & Febiger, 1979:1245.

2 Garrod A B. The nature and treatment ofgout and rheumatic gout. London:Walton and Maberley, 1859.

3 Freudweiler M. Experimentelleuntersuchungen uber die Enistehungder Gichtknoten. Dtsch Arch Klin Med1901 69: 155 (English translationArthritis Rheum 1965, 8: 270).

4 Painter C F. Subdeltoid bursitis. BostonMed Surg J 1907; 156: 345-9.

5 Schmitt J. Bursitis calcarea am



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Crystal-related arthropathies Suppl p 3

epicondylus externus humeri. EinBetrag zur Pathogenese derEpicondylitis. Archiv Fur Orthopadicheund Unfallchirurgie 1921; 19: 215-21.

6 Sanstrom C. Peritendinitis calcarea:common disease in middle life; itsdiagnosis, pathology and treatment.AJR 1938; 40: 1-21.

7 McCarty D J, Hollander J L.Identification of urate crystals in goutysynovial fluid.Ann Intern Med 1961; 54:452-60.

8 McCarty D J, Kohn N N, Faires J S. Thesignificance of phosphate crystals inthe synovial fluid of arthritis patients:the pseudogout syndrome. Ann InternMed1965; 56: 711-37.

9 Kohn N N, Hughes R E, McCarty D J,Faires J S. The significance of calciumphosphate crystals in the synovial fluidof arthritic patients: the 'pseudogoutsyndrome'. II Identification of crystals.Ann Intern Med 1962; 56: 738-45.

10 Zitnan D, Sitaj S. Chondrocalcinosisarticularis I. Clinical and radiological

study. Ann Rheum Dis 1963; 22:142-52.

11 Phelps P, McCarty D J. Crystal-inducedinflammation in canine joints. II.Importance of polymorphonuclearleucocytes. J Exp Med 1966; 124:115-26.

12 Weissmann G, Rita G A. Molecularbasis of gouty inflammation: interactionof monosodium urate crystals withlysosomes and liposomes. Nature, NewBiology 1972; 240: 167-72.

13 Dieppe P A, Crocker P, Huskisson E C,Willoughby D A. Apatite depositiondisease. A new arthropathy. Lancet1976; i: 266-8.

14 Gaucher A, Faure G, Netter P, Pourel J.Single crystal identification of calciumhydrogen phosphate dihydrate in thedestructive arthropathies ofchondrocalcinosis. European Journal ofRheumatology and Inflammation 1978;1: 120-4.

15 Fam A G, Pritzker K P H, Cheng P-T,Little A H. Cholesterol crystals in

osteoarthritic joint effusions. JRheumatol 1981; 8: 273-80.

16 Hoffman G S, Schumacher H R, Paul H,Cherian V, Reed R, Ramsay A G,Franck W A. Calcium oxalatemicrocrystalline-associated arthritis inend-stage renal disease.Ann Intern Med1982; 97: 36-42.

17 Mandel N S, Mandel G S. Structures ofcrystals that provoke inflammation. InWeisman G, ed. Advances ininflammation research. Vol 4. NewYork: Raven Press, 1982: 73-94.

1 8 Hasselbacher P. Crystal-proteininteractions in crystal-induced arthritis.In: Weisman G, ed. Advances ininflammation research. New York:Raven Press, 1982: 25-44.

19 Dieppe P A. Osteoarthritis and crystaldeposition. In: Panayi G, ed. Scientificbasis of rheumatology. Edinburgh:Churchill Livingstone, 1982: 224-40.

20 Dieppe P A, Calvert P. Crystals andjoint disease. London: Chapman andHall, 1983.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Crystals and inflammationPHILIP PLATT AND W. CARSON DICK

From the Department ofRheumatology, University ofNewcastle upon Tyne

The inflammatory response is one ofthe most fundamental defencemechanisms in human biology and isconsidered under the headings 'acute'and 'chronic'. Acute inflammation isassociated with small molecular weightmediators such as the biogenic amines,products of the lipoxygenase andcycloxygenase pathways and smallpeptides such as kinins. The cellscommonly considered under thisheading are polymorphonuclearleucocytes, eosinophils, and perhapsbasophils and mast cells. It iscustomary to think of chronicinflammation in terms of largermolecular weight molecules such asacidic and neutral proteolyticenzymes, complement componentproteins, proteins of the coagulationand fibrinolytic systems, lymphokines,interleukins, and immunoglobulinsand to visualise the cellularinvolvement in terms of lymphocytes,cells of the mononuclear phagocytelineage, endothelial cells, andfibroblasts.That such a convenient but artificial

compartmentalisation is incorrect isabundantly clear to anyhistopathologist accustomed toviewing damaged human tissue wherethese processes may all be seen in thesame histological section. When oneappreciates the additional confusion ofthe dynamism of events in humandisease, the rate of flow or traffic ofincoming cells through an inflammedsite, then a simplistic view ofinflammation becomes untenable. It iswith this insecure backcloth in mindthat we have to view the questionscurrently being asked about the role ofcrystals in the promotion ofinflammation.What are the cellular and

extracellular events and biochemicalprocesses which occur when a crystalinteracts with a biological membrane?To begin to think about this it isnecessary to consider our presentconcept of the cell membrane at which

point so many important events occur.The lipid bilayer is a three dimensionalnetwork which is continuouslyrecycling. Surface membranefragments are in equilibrium withmembrane segments within the cell sothat in many circumstances these maybe freely interchangeable. Thus thenuclear membrane is well adapted tointerchange freely with intracellularstructural membrane and with cellsurface membrane. It is important toemphasise the fluidity of themembrane; an insulin or antigenreceptor may be present and expressedat one time of day yet not at another.Hydrophilic protein segments float inconstrained fashion within the lipidbilayer. The glycosylated end of themolecule is orientated towards theinside of the cell, while the superficialpolysaccharide segments project intothe extracellular milieu. All of themembrane functions which are takenfor granted and too often studied inisolation occur in an integrated fashiondependent on events occurring close toand also at a considerable distancefrom the reaction site. Thus oxygenuptake and metabolism are linked todisulphide exchange interactions andto monovalent and divalent cationtransport, all of these being required asfuel to provide energy and to maintainthe integrity of the intact cell.Hormone recognition and in someinstances internalisation and theprocess of antigen recognition and'restriction' by the majorhistocompatability complex are alsodependent on the integrity of themembrane and on the availability ofenergy and substrates. Phagocytosis,endocytosis, 'regurgitation duringfeeding', migration, blasttransformation, and secretion whetherit be of hormone, structural protein,leukotriene, immunoglobulin,interleukin or lymphokine are allintegrally dependent on thesefunctions.We have recorded some of the many

interactions between crystals andbiological membranes.

Mechanisms of crystal-induceddamage

CELLS

Polymorphonuclear leucocytes,endothelial lining cellsOne of the earliest contenders asputative effector cells in theproduction of crystal-induced tissuedestruction was the poly-morphonuclear leucocyte. It waswithin or around these cells thatcrystals were first discovered.' Thestage was set, therefore, for a unitaryhypothesis that would explaininflammation and tissue destruction indiseases such as gout and calciumpyrophosphate. This was proposed byHollander and McCarty and wasfocused on the polymorphonuclearleucocyte as the central cell. It alsostated, if only implicitly, that theresponsible crystals were formed in theextracellular space. The model was anextension of the 'rupture from within'hypothesis proposed in the context ofpulmonary disease produced bysilicate crystals.2 After ingestion of anon-metabolisable crystal it issuggested that the crystal surface bindsto and disrupts the membrane of thesecondary phagolysosome-producingrelease of lytic enzymes within the celland cell death. Thereafter both theunmetabolised crystal and these samelytic enzymes are released into theextracellular space where both arethen in a position to produce furthertissue destruction.

Early work by Phelps and McCarty3showed pronounced suppression ofinflammation in a canine urate crystalmodel when neutropaenia wasinduced by the administration ofvinblastine. The inflammatoryresponse was restored by transfusionof normal blood. Unfortunately,vinblastine exerts effects on many



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Crystals and inflammation Suppl p 5

systems in addition to thepolymorphonuclear leucocyte. Thiscriticism is also applicable to workwhich overinterprets results withcobra venom depleted animals. In anattempt to overcome this problemChang and Gralla4 produced similarresults using heterologousantipolymorphonuclear serum, againfocusing attention on the requirementfor polymorphonuclear leucocytes.A considerable amount of excellent

work was focused on the lateenzymatic events that could beresponsible for matrix destruction andcell death and it was documented thatthis train of events could be set inmotion during phagocytosis ofpreformed crystals. Discharge ofenzymes possessing maximal activityeither at acidic or at neutral pH wasshown to occur during phagocytosisand endocytosis even in the intact celland was described as 'regurgitationduring feeding'. It was also shown thatthese cellular enzymes possessed all ofthe various activities required to digesteach of the components of normal orabnormal connective tissue collagen,glycoprotein and proteoglycan matrix.It seems likely that these enzymesexert their maximum effect in thepericellular area or intracellularlyrather than in plasma, synovial fluid,or the extracellular space.A considerable volume of work

published subsequently is in accordwith this model. Hoffstein andWeissman' exploited the peculiarcharacteristics of the large lysosomesof the smooth dogfish shark in a mostelegant series of studies. The timecourse of cell death after exposure toprefonned crystals was monotored. Celldeath began within 10 to 15 minutes offusion of lysosomes withcrystal-containing phagosomes,primary crystal-free phagosomesremaining unaffected. The earlyrelease of the cytoplasmic enzymelactic dehydrogenase in addition tolysosomal enzymes is an indication ofthe peculiarly destructive effects ofsome crystals, particularlymonosodium urate, on polymorphs.This response was inhibited bycytochalasin B, which impairsphagocytosis, and by agents whichincrease cyclic adenosinemonophosphate activity such astheophylline. These results wereconsistent with the Hollander and

McCarty 'suicide-sac' hypothesis. Aconsiderable number of publishedreports now derive from sequentialstudies of the interaction of preformedurate, calcium pyrophosphate, andother crystals with polymorphonuclearleucocytes in vitro to establish thesequence of events which followscontact and subsequent ingestion ofcrystals by these cells. Differencesbetwen crystals tend to be quantitativerather than qualitative and maydepend on differences in size, surfacecharge, and interaction with variousadsorbed proteins. Again most reportssupport the validity of the suicide-sacmodel at least in vitro.Although much of the work on the

interaction of polymorphs and crystalshas centred on the release of lysosomalenzymes the first reaction,chronologically, of the polymorph tocrystals is the production of oxygenradicals. These toxic oxygen-derivedproducts include superoxide,peroxides, and hydroxyl radicals andtheir release is triggered within 30seconds of contact betweenpolymorph membrane and crystals.6 Aseries of methods of measuring theseproducts have been used includingcytochrome reduction,7 nitrobluetetrazolium reduction8 andluminol-dependent chemi-luminescence6 and have demonstratedmodifying effects of protein coating ofcrystals and complement activation.7The exact role of these potential toxicproducts is at present unclear,although if nothing else they do offer avery early marker of polymorphactivation.Much of the core of these ideas may

stand the test of time and subsequentexperiment but at the moment certainobservations militate against thesimple assumption that thepolymorphonuclear leucocyte isindeed the sole cell responsible.Schumacher et al.9 have documentedclearly in thorough sequential studiesthat the first cell to interact with amonosodium urate crystal is thesynoviocyte and that only thereafterwere polymorphonuclear leucocytesrecruited.

Glatt et al. 0 in parallel studies butusing the rat pleural cavity,demonstrated only 50% reduction ininflammatory response whenneutropaenia was induced withcyclophosphamide and concluded that

the pleural lining cells made animportant contribution. The samecriticism that may be levelled againstthe selectivity and specificity of cobravenom or vinblastine is applicable tocyclophosphamide.

Finally, crystal-induced tissuedestruction and inflammation hasbeen recorded in man at a time whenthere could have been only a very fewpolymorphonuclear leucocytespresent." "These observations do notcontradict most of the previous in vitrowork but they do focus attention onthe importance of cells other thanpolymorphonuclear leucocytes in theproduction of inflammation and tissuedestruction in vivo.

PlateletsPlatelets are capable of phagocytosingmonosodium urate crystals'3 andrespond in two phases, an active phasewith release of adenosine diphosphateand triphosphate and serotoninfollowed by a lytic phase with releaseof lactic dehydrogenase. How muchthis interaction of crystals with plateletmembranes contributes to the totalinflammatory response has yet to bequantified but again the interaction ismodulated by proteins that areadsorbed on to the surface of thecrystal (vide infra). An earlynonspecific interaction may occurthrough direct contact and hydrogenbonding. Thereafter more specificinteraction occurs between exposed Fcgroups projecting from the crystalsurface (to which they are bound bytheir Fab binding site) and plateletmembrane Fc receptors in a manneranalogous to the binding of crystals tonucleated cell membranes. If thesesuggestions are true then there must beconsiderable differences of bindingenergy at these sites and this has yet tobe demonstrated.

NON-CELLULAR MEDIATORS OFINFLAMMATION

ImmunoglobulinsThe first group to call attention to thebinding of protein to the surface ofcrystalswasScheeletal. in 1954.'4Theinteraction of crystal surfaces withspecific proteins such as complementcomponents and clotting factors is welldocumented. However, many other



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


proteins bind to crystal surfaces. IgGbinds in the greatest quantitiesfollowed by lysozyme, serum albumin,and ovalbumin in that order."5 Thefunctions that are affected by thisinteraction include electrophoreticmobility, isoelectric point and severaltitratable groups, and functionalchanges occur in Hageman factor,complement, and IgG. One of themost interesting points that has beenemphasised by Hasselbacher"6 is thatconformational changes occur in thetertiary structure of the adsorbedprotein which may have profoundeffects upon its function.The most intense interest in

crystal-protein interaction is focusedon immunoglobulin binding, whichseems to occur by the Fab binding site,leaving the Fc fragment free to interactwith any cell or structure whichpossesses an Fc receptor. It is likelythat these subsequent interactions atthe Fc site account for a considerableproportion of the biological effects ofcrystal-protein interaction. Cationicspecies of IgG are boundpreferentially to anionic crystals suchas calcium pyrophosphate (CPP),monosodium urate, calciumhydroxyapatite, and amphiboleasbestos, whereas the only positivelycharged crystal, chryositile asbestos,binds anionic IgG. Thus electrostaticbonds are important even if otherspecies of bonding such as Van derWaals, hydrophobic or hydrogenbonds may also participate in differentcircumstances.'6 These surface chargeinteractions may dictate the subclass ofimmunoglobulin bound and even theorientation of the molecule.

It should be emphasised thatadsorption of proteins of variousspecies may operate to protect againstas well as to enhance membranolysis'7and it is probably too early to speculatefrom the meagre data available on theresultant effect in vivo.

ComplementThere has been considerable interestin the role of complement in theproduction of crystal-induceddiseases. Decreased complement andthe presence of complement splitproducts has been noted in acutegout.'" In vitro activation of bothclassic and alternative pathway, withand without the presence ofimmunoglobulins, has been

documented by several authors.Complement activation was at firstthought to be produced only whenendotoxin was adsorbed to the crystalsurface but this hypothesis does notexplain all of the experimental data.Subsequent work showed conflictingresults due, at least in part, todifferences in crystal numbers andsurface characteristics on the one handand other experimental conditions onthe other. Naff and Byers" describedthe activation of complement bymonosodium urate crystals. Theynoted a small decrease in Cl activitybut pronounced decreases in C2, C3,C4, and C5 and suggested amechanism other than the classicpathway of activation. Considerablyless activation of complementoccurred in agammaglobulinaemicsera. Hasselbacher reported activationof C3 by monosodium urate, calciumpyrophosphate, and hydroxyapatite,and concluded on the basis of calciumdependency and abolition of theactivation with EDTA and EGTA thatthe classic pathway was the majorpathway of activation." Ginsberg et al.reported decreased Cl, C4, and C3haemolytic activity and againconcluded that the classic pathway wasthe significant mechanism.2' They alsosuggested that the activation occurredindependently of IgG, by direct bindingof the crystal to Clq. Doherty andDieppe reported activation of C3 in amanner suggestive of alternatepathway activation, and againsuggested that the mechanism wasindependent of immunoglobulin."These results may be interpreted as

showing that crystals alone appearedcapable of activating both thealternate and classic pathways, where-as in the presence of adsorbedimmunoglobulin activation of theclassical pathway by binding throughtheir exposed Fc fragmentspredominates.

Hageman factor and kininsHageman factor has been suggested asa possible mediator of crystal-inducedinflammation through its role in thegeneration of kallikrein and kinin. Itspresence in synovial fluid andactivation by monosodium uratecrystals has been documented.'3Activation of Hageman factor (80 000Daltons) releases two peptidefragments (52 000 and 28 000

Daltons respectively) and these havebeen detected after the addition ofmonosodium urate crystals toabnormal synovial fluid or to dilutenormal plasma. Kininogen andpre-kallikrein are necessary co-factorsHowever, an acute gouty arthritis hasbeen found in animals lackingHageman factor.24 These observationssuggest that activation of Hagemanfactor is not a major mediator incrystal-induced arthritis.

There is abundant evidence ofactivation of the kinin system in vitrobut the data is less complete in vivo. Itis possible that the kinin systemoperates in synergy with otherinflammatory mediators such ascyclo-oxygenase products to producethe exquisite pain which is socharacteristic of podagra.2"

Acute phase proteinsThe interaction of acute phaseproteins and crystals in thepathogenesis of inflammation needsfurther study, in particular the role ofacute phase proteins in mechanisms oftermination of the inflammatoryresponse. Information is becomingavailable on the structure, function,and biosynthesis of acute phaseproteins in particular C-reactiveprotein26"28 and alpha-i-acidglycoprotein and the controlmechanisms involved.29Thisinformation should now be applied tocrystal-induced inflammation.

Liposomes, lysosomes, cell membranefragments, etc.Crystals of monosodium urate or silicarupture natural and artificial cellmembranes in media that do notconstrain hydrogen bonding. Lysis isenhanced by the presence in thesemembranes of androgens such astestosterone and reduced by theincorporation of oestrogens such as17 /3 oestradiol.'0 It has beensuggested that this interaction maycontribute to the explanation of thedifference in predeliction of the sexesto the development of gouty arthritis.

In summary, the mechanisms ofcrystal-induced inflammation arecomplex and almost certainlymultifactorial. It is likely thatindividual crystal types invokedifferent mechanisms to variabledegrees. Furthermore: 'There is noinvariant host response to a given



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Crystals and inflammation Suppl p 7

crystal' (Hasselbacher 1982).Knowledge of mechanisms ofcrystal-induced inflammation isapplicable not only to crystal-induceddiseases in rheumatology but also toindustrial lung diseases2 and possiblyatherosclerosis (P N Platt and AMalcolm, p 000).We believe that the first priority is to

accumulate reliable reproducible datadesigned to dissect the detail of themultiplicity of mechanisms involved inthe interaction of different crystalsurfaces with different varieties of cellmembranes in vitro. Armed withconfident and secure data we may thenerect testable hypotheses that may beexamined in vivo. All of the importantmethods are available in thelaboratory and it might be better tofocus attention for a while on thisaspect of the subject rather than tocontinue to erect more and moreelaborate models on the presentfragmentary and insecure data.

References

1 McCarty D J, Hollander J L.Identification of urate crystals in goutysynovial fluid.Ann Intern Med 1961; 54:452-60.

2 Alison A C, Harington J S, Birbeck M.An examination of the cytotoxic effectsof silica on macrophages. J Exp Med1966; 124: 141-61.

3 Phelps P, McCarty D J. Crystal inducedinflammation in canine joints. IIImportance of polymorphonuclearleucocytes. J Exp Med 1966; 124:115-25.

4 Chang Y H, Garalla E J. Suppression ofurate crystal induced canine jointinflammation by heterologousanti-polymorphonuclear leucocyteserum. Arthritis Rheum 1968; 11:145-50.

5 Hoffstein S, Weissman G. Mechanismsof lysosomal enzyme release fromleucocytes. IV interaction ofmonosodium urate with dogfish andhuman leucocytes. Arthritis Rheum1975; 18: 153-65.

6 Platt P N, Jeffery S, Wilson L,Griffiths I D. The interaction of crystals

and polymorphonuclear monitored byluminol dependent chemiluminescence.Ann Rheum Dis 1981; 40: 615.

7 Abramson S, Hoffstein S T, Weiss-mann G. Superoxide generation byneutrophils exposed to monosodiumurate: effect of protein adsorption andcomplement activation. Arthritis Rheum1982; 25: 174-81.

8 Simchowitz L, Atkinson J P, Spilberg I.Stimulation of the respiratory burst inhuman neutrophils by crystalphagocytosis. Arthritis Rheum 1982; 25:181-8.

9 Schumacher H R, Phelps P, Agudelo C A.Urate crystal induced inflammation indog joints: sequence of synovialchanges. J Rheumatol 1974; 1: 102-13.

10 Glatt M, Dieppe P A, Willoughby P A.Crystal induced inflammation, enzymerelease and the effects of drugs in the ratpleural space. J Rheumatol 1979; 6:251-8.

11 Ortel R W, Newcombe D S. Acutegouty arthritis and response tocolchicine in the virtual absence ofsynovial fluid leucocytes. N Engl J Med1974; 290: 1363-4.

12 Matthay M, Lindamood M, Steiger-wald J C. Acute pseudogout in theabsence of synovial fluid leucocytes. JRheumatol 1977; 4: 303-6.

13 Schumacher H R, Phelps P. Sequentialchanges in human polymorphonuclearleucocytes after urate crystalphagocytosis. An electron microscopystudy. Arthritis Rheum 1971; 14:513-26.

14 Scheel L D, Smith B, Van Riper J,Fleischer E. Toxicity of silica; IIcharacteristics of protein films adsorbedby quartz. Archives of IndustrialHygiene 1954; 9: 29-36.

15 Kozin F, McCarty D J. Protein bindingto monosodium urate monohydrate,calcium pyrophosphate dihydrate andsilicon dioxide crystals. I Physicalcharacteristics. J Lab Clin Med 1977;89: 1314-25.

16 Hasselbacher P. Crystal-proteininteractions in crystal-induced arthritis.In: Weissmann G, ed. Advances ininflammation research. Vol. 4. NewYork: Raven Press, 1982.

17 Malawista S E, Van Blaricon G,Crettela S B, Swartz M L. The phlogisticpotential of urate in solution; studies ofthe phlogistic process in human

leucocytes. Arthritis Rheum 1979; 22:728-36.

18 Pekin T J, Zvaifler N J. Haemolyticcomplement in synovial fluid. J ClinInvest 1964; 43: 1372-82.

19 Naff G B, Byers P H. Complement as amediator of inflammation in acute goutyarthritis. Studies on the reactionbetween human serum complement andurate crystals.J Lab Clin Med 1973; 81:747-60.

20 Hasselbacher P. C3 activation bymonosodium urate monohydrate andother material. Arthritis Rheum 1979;22: 571-8.

21 Ginsberg M H. The mechanisms ofcomplement activation by monosodiumurate. Arthritis Rheum 1979; 22: 612.

22 Doherty M, Dieppe P A, Whicher J.Activation of the alternate pathwaycomplement by monosodium uratemonohydrate crystals and otherinflammatory particles. Ann RheumDis 1983; 42: 285-91.

23 Kellermyer R W, Breckenridge R T.The inflammatory process in acutegouty arthritis. I Activation of HagemanFactor by sodium urate crystals. J LabClin Med 1965; 67: 455-60.

24 Spilberg L. Urate crystal arthritis inanimals lacking Hageman Factor.Arthritis Rheum 1974; 17: 142-8.

25 Kellermyer R W, Naff G B. Chemicalmediators of inflammation in urategouty arthritis. Arthritis Rheum 1975;18: 765-70.

26 Volanakis J E, Narkates J. Interactionsof C-reactive protein with artificialphosphatidylcholine bilayers andcomplement. J Immunol 1981; 125:1820-5.

27 Oliveira E B, Gotschlich E C, Liu T-Y.Primary structure of C-reactive protein.J Biol Chem 1979; 254: 489-502.

28 Mortensen R F, Osmand A P, Lind T F,Gewurz H. Interaction of C-reactiveprotein with lymphocytes andmonocytes complement dependentadherence and phagocytosis.J Immunol1976; 117: 774-81.

29 Kushner J, Feldmann G C. Control ofthe acute phase response. J Exp Med1978; 148: 466-521.

30 Weissmann G, Rita G A. Molecularbasis of gouty inflammation; interactionof monosodium urate crystals withlysosomes and liposomes. Nature, NewBiology 1972; 240: 167-72.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Human purine metabolism: some recent advances andrelationships with immunodeficiencyGEORGE NUKI

From the Rheumatic Diseases Unit, Department ofMedicine (WGH), University ofEdinburgh

Studies of human purine metabolismbegan some 200 years ago withScheele's identification of uric acid as aconstituent of a renal calculus' andWollaston's demonstration of urate ina tophus from his own ear.2 Just asthese were among the earliestexperiments in the new science ofbiochemistry, so Garrod's (1854)3thread test must rank as the first assayin what we now call the discipline ofclinical chemistry. Seegmiller'sdiscovery, 15 years ago, that the severepurine overproduction and goutobserved in boys with Lesch-Nyhansyndrome resulted from a specificdefect of the purine salvage enzymehypoxanthine-guanine phospho-ribosyl transferase (HGPRT)4 led towhat has been described as the 'purinerevolution' and was soon followed bythe identification of other inbornerrors of metabolism associated withprimary purine overproduction andgout. Oxypurine and deoxypurinebases, nucleosides, and nucleotides do,however, have much wider physio-logical importance as intermediates inenergy metabolism, components ofcoenzymes, vasoactive and neuro-transmitter substances, mediators ofhormonal action, and as the buildingblocks for transmitters of geneticinformation (DNA and RNA). Mostrecently interest has been focused onthe role of purines, and especiallydeoxypurines, in the regulation ofimmune responses.

Gout and hyperuricaemia

In most individuals with gout orasymptomatic hyperuricaemia the risein serum urate concentration is theresult of a decrease in renal urateclearance. In 15% or less there isevidence of excessive purine synthesisand hyperuricosuria. A 24-hoururinary urate excretion greater than

0-6 g (3-6 mmol) while on a 10-9 MJ(2600 kilocalorie), 70 g protein,purine-free diet is generally taken tobe abnormal,5 but recent data onhealthy Australians on low purinediets showed mean urinary urates of3-9 mmol (0 65 g)/24 hrs for men and2-9 mmol (0.5 g)/24 hrs for women.6A variety of mutations of three

enzymes are associated with excessivepurine synthesis in patients with gout,but these account for only a smallproportion of those identified asprimary overproducers of urate.

Purine nucleotide synthesis inmammalian cells is regulated by a

balanced interaction of severalbiochemical pathways (Fig. 1). Purinebiosynthesis de novo consists of aseries of 10 enzymatic steps in whichglutamine, glycine, carbon dioxide,aspartate, and 1-carbon-formylderivatives of folate are added to theribose moiety of phosphoribosylpyrophosphate (PRPP) to forminosinic acid (IMP). Adenosinemonophosphate (AMP) andguanosine monophosphate (GMP) areeach formed from IMP by twoenzymatic steps before phosphory-lation to their respective diphosphatesand triphosphates. Alternatively, the

Ribose - 5 - P + ATP

5- Phosphoribosyl -1- Pyrophosphate PRPP + GlutamineDeoxynu~cetc Nuclic A * ,Nucleic Dexynucli~c

Acids Acids ^ Acids Acids

dGTPGT* 5-Phosphoibosyl-1-Amine *dGTP GTP AT,- A O P:dAAP\\ \\\ F?1 ormate 1 /77NHCO3

' AspartateFormate

Guanylic Acd A-rIosindcAcid lo Adeyi Acid Adene d AMP

Guanosie HGPRT Io ne eR Adenosone N

9f % 1 > 2,8-dihydroxyademne

Guanine Hypoxanthine deoxyinosre -i' dexyaadenoswe

Xanthine

Uric cid

Fig. 1 Pathways ofpurine metabolism in man.ADA = Adenosine deaminase.APRT = Adenine phosphoribosyl transferase.HGPRT = Hypoxanthine-guanine phosphoribosyl transferase.NP = Nucleoside phosphorylase.5'NT = 5'nucleotidase.PAT = Phosphoribosyl pyrophosphate amidotransferase.PPRPS = Phosphoribosyl pyrophosphate synthetase.XO = Xanthine oxidase.Reproduced from: Nuki G, Advanced Medicine 15. London: Pitman Medical,1979: 140 with permission of publishers.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Human purine metabolism: some recent advances Suppl p 9

nucleotide monophosphates may besynthesised by phosphorylation ofpurine bases by phosphoribosyl trans-ferase enzymes reacting with PRPP,and AMP may be formed byphosphorylation of adenosine byadenosine kinase. In vitro studies ofamido phosphoribosyl transferase(PAT), the first enzyme in the de novopathway, show it to have allostericproperties that one often associateswith a rate-limiting enzyme, and haveprovided a molecular model for themechanism whereby the rate of purinenucleotide synthesis may be regulated.The protein exists in two forms; aninactive dimer and a catalyticallyactive monomer. PRPP converts thelarger form of the enzyme to the activemonomer, while purine nucleotidemonophosphates have the reverseeffect.7

HGPRT deficiency

In Lesch-Nyhan syndrome and thesevere familial form of X-linked gout,which is also associated with deficiencyof HGPRT, it seems logical toconclude that the considerablyaccelerated rate of de novo purinesynthesis results from bothaccumulation of PRPP and diminishedfeedback inhibition of PAT by purinenucleotides. However, whileintracellular concentrations of PRPPare increased in HGPRT-deficientcells, steady state purine nucleotideconcentrations do not appear to bereduced,8 and concentrations ofpyrimidine nucleotides are actuallyincreased.9 A good deal of evidencesuggests that PRPP, which is present incells at limiting concentrations, is themajor regulator of rates of de novopurine synthesis whenever purineoverproduction occurs in vivo. On theother hand, the importance of purinebase 'salvage' is emphasised byexperiments which show that rates ofpurine synthesis in normal humanlymphoblasts may be increased tothose observed in HGPRT-deficientmutants if hypoxanthine is carefullyexcluded from the culture medium."0Moreover, Hershfield hasdemonstrated co-ordinated inhibitionof IMP conversion to adeninenucleotides with enhancement of totalguanine nucleotide synthesis after theaddition of adenosine to humanlymphoblasts in culture, emphasising

the potential regulatory importance ofthe distal inosinic acid branch point."There is increasing evidence of

genetic heterogeneity at the HGPRTlocus."2 13 Although the severity ofclinical manifestations is usuallyproportional to the severity of theenzyme defect, this is not alwaysapparently the case. In rare instancesmutations resulting in abnormalenzyme kinetics may be associated onthe one hand with the classic clinicalLesch-Nyhan phenotype andapparently normal red cell HGPRTactivity assayed at saturating substrateconcentrations, or on the other handwith X-linked gout withoutneurological features and apparentlyabsent enzyme activity in standardassays. Immunochemical studies haveshown that most mutations at theHGPRT locus are structural genemutations resulting in catalyticallydefective enzyme protein that does notcross react with antibody raisedagainst highly purified normal humanHGPRT (CRM negative mutants).'4Recent studies have provided

evidence for a single amino acidsubstitution in one such structuralvariant (HGPRTmunich)," while othergroups have now succeeded in cloningthe HGPRT gene.'

APRT deficiency

Adenine phosphoribosyl transferasedeficiency is not associated with goutor purine overproduction. Althoughheterozygous, partial APRTdeficiency was originally described ingouty patients it is clear from familystudies that the enzyme defect andhyperuricaemia are inherited asindependent traits.'8 The apparentassociation arose spuriously becauseAPRT enzyme assays wereundertaken only in patients with gout,although heterozygous APRTdeficiency may occur in as many as 1 %of the population.'9 More recently, anumber of patients have beenidentified in whom severe,homozygous APRT deficiency wasassociated with symptomatic urinarylithiasis and calculi composed of 2,8-dihydroxyadenine.20 21 X-raydiffraction, infrared and ultravioletabsorption spectrometry, and ionexchange chromatography of HPLCextracts of these stones have beenrequired to differentiate the adenine

metabolites from uric acid." Thesepatients may be treated successfullywith a low purine diet andallopurinol," which inhibits theformation of 2,8-dihydroxyadenine.

Superactive PRPP synthetase

Structural mutations resulting insuperactive phosphoribosyl pyro-phosphate synthetase are now wellestablished as a cause of X-linked goutand purine overproduction. Initiallyincreased enzyme activity was associ-ated with decreased sensitivity tofeedback inhibition by purine nucleo-tides in one family" and a primaryincrease in catalytic activity perenzyme molecule of the mutant pro-tein in another.'4 More recently thegenetic heterogeneity and diversity ofabnormalities in kinetic mechanismsleading to superactivity of this enzymehas been extended to include mutantswith increased V max, various abnor-malities of purine nucleotide feedbackinhibitor responsiveness, increasedaffinity for the substrate ribose-5-phosphate, and a combination ofcatalytic and regulatory defects."

Most affected males have presentedwith gouty arthritis and/or uric acidurolithiasis in early adult life, but twofamilies have been described wherehomozygous boys developed symp-toms in childhood and sharedmetabolic abnormalities and nervedeafness with their heterozygousmothers.'6 27

Type I glycogen storage disease

Glucose-6-phosphatase deficiency isassociated with both overproductionand underexcretion of uric acid, so thatgout and hyperuricaemia becomeappreciable clinical problems inaffected individuals fortunate enoughto survive to adult life. Persistent lacticacidaemia is the major cause ofimpairment of uric acid excretion andthe simultaneous increase in de novopurine synthesis observed in thesepatients is thought to result fromexcessive production of PRPP after'shunting' of metabolites through thepentose phosphate pathway.'8 It hasbeen difficult to test this hypothesisdirect, as glucose-6-phosphataseactivity is limited to the liver, kidney,and intestinal mucosa. Recently it hasbeen suggested that increased purine



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


production in this disorder may besecondary to accelerated purinedegradation after recurrentglycogenolysis and depletion ofATP29; rather like the mechanismaccounting for the purineoverproduction and hyperuricaemiawhich follow ingestion or infusion offructose.30 Van den Berghe and hiscolleagues believe that the hepaticcatabolism of adenine nucleotides andthe formation of uric acid areregulated by the activity of AMPdeaminase.3'Type I glycogen storage disease is

usually characterised by completeabsence of hepatic glucose-6-phosphatase. We have recentlyreported the case of a child with a par-tial enzyme deficiency where there wasa striking absence of hypoglycaemia.32This suggests that type I glycogen stor-age disease with partial deficiency ofhepatic glucose-6-phosphatase shouldperhaps be considered in patients withgout or hyperuricaemia associatedwith hypertriglyceridaemia and lacticacidaemia, even in the absence ofhypoglycaemia.

Other enzyme defects and gout

We are still unable to define themetabolic abnormality responsible forpurine overproduction in most of thefew patients with gout who are primaryhyperexcretors of uric acid. Thoughsome of these may well be attributableto subtle kinetic variations of knownenzyme defects not detectable instandard in vitro assays, it seemsprobable that other primary purineenzyme defects remain to bediscovered. For example, in vivo33 andin vitro34 studies with drugs that inhibitIMP dehydrogenase suggest thatdeficiency of this enzyme might well beassociated with primary purineoverproduction.

PURINE ENZYMEABNORMALITIES ANDIMMUNODEFICIENCYA great deal of interest has beenrecently aroused by the association oftwo inborn errors of purinemetabolism with immune deficiencysyndromes.

Severe deficiency of adenosinedeaminase (ADA) was first reportedin red cell haemolysates from twochildren with severe combinedimmunodeficiency in 1972.3 Three

years later Eloise Giblett also reporteda gross deficiency of red cell purinenucleoside phosphorylase (PNP) inchildren with recurrent infections anda predominant T-lymphocytedeficiency.36The causal relationship between

these inherited enzyme abnormalitiesand the immunodeficiency syndromesis now established and has highlightedthe need for an intact purine catabolicpathway to maintain normal cellularand humoral immunity. In patientswith deficiencies of ADA and PNPplasma concentrations of deoxy-adenosine37 and deoxyguanosine3"are raised. These deoxynucleosidesare selectively phosphorylated andtrapped by T-lymphocytes with highactivities of deoxycytidine kinase andlow activities of intracellular deoxy-nucleotidase. As a result deoxy-adenosine39 and dexoyguanosine4accumulate in the cells, leading toinhibition of ribonucleotide reductaseand DNA synthesis. An alternativemechanism for the toxicity of deoxy-adenosine in ADA deficiency hasbeen proposed by Hershfield,4" whoshowed that it causes irreversible 'sui-cide' inactivation of the enzymeS-adenosyl homocysteine hydrolase.Consistent with this hypothesisS-adenosyl homocysteine hydrolaseactivity is found to be considerablyreduced in the red cells of childrenwithADA deficiency,42 but an alterna-tive explanation needs to be found toexplain the T-cell deficiency in PNPdeficiency, as deoxyguanosine doesnot inhibit the S-adenosyl homocys-teine hydrolase enzyme.To explain the difference in the

immunodeficiency syndromesassociated with ADA and PNP defi-ciency, evidence has recently beenproduced to show that intracellularATP activities are depleted withaccumulation of deoxyadenosine butnot with deoxyguanosine, so killing thenon-dividing helper T-lymphocytes.43Humoral immunity persists in patientswith PNP deficiency because T-hel-per function is relatively proliferationindependent.Decreased activity of the ectopurine

5' nucleotidase (5'NT) enzyme hasbeen associated with X-linked"4 andacquired adult onset (common vari-able) hypergammaglobulinaemia.45Deficiency of this enzyme does not,however, result from a primary gene

mutation and there is no evidence tosuggest that it is causally related toB-cell deficiency. 5'NT deficiency inX-linked agammaglobulinaemialargely reflects the deficit in circulatingB-cells which have high 5'NT activity.B-cell numbers are, however, almostnormal in common variable hypo-gammaglobulinaemia and in this situa-tion it seems likely that reduction in5'NT is associated with a population ofrelatively immature B-lymphocytes.5'NT has been previously shown to bea marker of differentiation in both T46and B47 lymphocytes.

References

1 Scheele K W. Examen chemicum calculiurinarii. Opuscula 1976; 73: 2.

2 Wollaston W H. On gouty and urinaryconcretions. Philos Trans R Soc Lond1797; 87: 386.

3 Garrod A B. On the blood and effusedfluids in gout, rheumatism and Bright'sDisease. Transactions of theMedico-Chirugical Society ofEdinburgh1854; 37: 49.

4 Seegmiller J E, Rosenbloom F M,Kelley W N. Enzyme defect associatedwith a sex-linked neurological disorderand excessive purine synthesis. Science1967; 155: 1682-4.

5 Seegmiller J E, Grayzel A I, Laster L,Liddle L. Uric acid production in gout.JClin Invest 1961; 40: 1304-14.

6 Stafford W, Emmerson B T. Effect ofpurine restriction on serum and urinaryurate (Abstract). J Clin Chem ClinBiochem 1982; 20: 422-3.

7 Holmes E W, Wyngaarden J B,Kelley W N. Human glutamine, phos-phoribosyl pyrophosphate amido trans-ferase: two molecular forms inter-convertible by purine ribonucleotidesand phosphoribosyl pyrophosphate. JBiol Chem 1973; 248: 6035-40.

8 Rosenbloom F M, Henderson J F,Caldwell I C, Kelley W N, Seeg-miller J E. Biochemical bases of accele-rated purine biosynthesis de novo inhuman fibroblasts lacking hypo-xanthine-guanine phosphoribosyl trans-ferase. J Biol Chem 1968; 243:1166-73.

9 Nuki G, Astrin K, Brenton D,Cruikshank M, Lever L, Seegmiller J E.Purine and pyrimidine nucleotideconcentrations in cells with decreasedhypoxanthine-guanine phosphoribosyltransferase (HGPRT) activity.Adv ExpMed Biol 1977; 76A: 326-9.

10 Hershfield M S, Seegmiller J E.Regulation of de novo purine synthesisin human lymphoblasts: similar rates ofde novo synthesis during growth bynormal cells and mutants deficient inhypoxanthine-guanine phosphoribosyl



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Human purine metabolism: some recent advances Suppl p 11

transferase activity. J Biol Chem 1977;252: 6002-10.

11 Hershfield M S, Seegmiller J E.Regulation of de novo purinebiosynthesis in human lymphoblasts:co-ordinate control of proximal ratedetermining steps and the inosinic acidbranch point. J Biol Chem 1976; 251:7348-54.

12 Emmerson B T, Thompson L. Thespectrum of hypoxanthine-guaninephosphoribosyl transferase deficiency.Q J Med 1973; 42: 423-40.

13 Nyhan W L. Genetic heterogeneity atthe locus for hypoxanthine-guaninephosphoribosyl transferase. In: Purineand pyrimidine metabolism, CibaFoundation Symposium No 48.Amsterdam. Excerpta Medica: 65-81.

14 Upchurch K S, Leyva A, Arnold W J,Holmes E W, Kelley W N. Hypo-xanthine phosphoribosyl transferasedeficiency: association of reducedcatalytic activity with reduced levels ofimmunologically detectable enzyme

protein. Proc Natl Acad Sci USA 1975;72: 4142-6.

15 Wilson J M, Tait G E, Grobner W, Zol-lner N, Kelley W N. Evidence for an

amino acid substitution (Ser Arg) in thehypoxanthine-guanine phosphoribosyltransferase variant HGPRTmunnCh (Abs-tract). J Clin Chem Clin Biochem 1982;

16 Jolly D J, Esty A, Friedmann T. Thecloned human hypoxanthine-guaninephosphoribosyl transferase gene(Abstract). J Clin Chem Clin Biochem1982; 20: 380.

17 Chinault A C, Konecki D S, Brennand J,Patel P, Melton D W, Caskey C T.Cloning and characterisations ofmammalian hypoxanthine-guaninephosphoribosyl transferase genes(Abstract). J Clin Chem Clin Biochem1982; 20: 442.

18 Emmerson B T, Gordon R B,Thompson L. Adenine phosphoribosyltransferase deficiency in a female withgout. Adv Exp Med Biol 1974; 41:327-31.

19 Fox I H, Lacroix S, Planet G, Moore N.Partial deficiency of adeninephosphoribosyl transferase in man.

Medicine 1977; 56: 515-26.20 Cartier M B, Hanset M. Une nouvelle

maladie metabolique: le deficit complexen adenine phosphoribosyl transferaseavec lithiase de 2, 8-dihydroxyadenine.Comptes Rendus Hebdomadaires desSeances de lAcademie des Sciences(Paris) 1974; 270: 883-6.

21 Simmonds H A, Van Acker K J,Cameron J S, Sneddon W. Theidentification of 2, 8-dihydroxyadenine,a new component of urinary stones.Biochem J 1976; 157: 485-7.

22 Simmonds H A, Van Acker K J,Cameron J S, McBurney A. Purineexcretion in complete adenine

phosphoribosyl transferase deficiency:effect of diet and allopurinol therapy.Adv Exp Med Biol 1977; 76B: 304-11.

23 Sperling Q, Persky-Brosh S, Boer P, DeVries A. Human erythrocyte phospho-ribosyl pyrophosphate synthetase muta-tionally altered in relatory properties.Biochem Med 1973; 7: 389-95.

24 Becker M A, Kostel P J, Meyer L J,Seegmiller J E. Human phosphoribosylpyrophosphate synthetase: increasedenzyme activity in a family with gout andexcessive purine synthesis. Proc NatlAcad Sci USA 1973; 70: 2749.

25 Becker M A. Phosphoribosyl pyro-phosphate synthetase superactivity:detection, characterization of underly-ing defects and treatment (Abstract). JClin Chem Clin Biochem 1982; 20: 347.

26 Becker M A, Raivio K 0, Bakay B,Adams W B. Variant human phos-phoribosyl pyrophosphate synthetasealtered in regulatory and catalytic func-tions. J Clin Invest 1980; 65: 109-20.

27 Simmonds H A, Webster D R, Wilson J,Lingham S. An X-linked syndromecharacterised by hyperuricaemia, deaf-ness and neurodevelopmental abnor-malities. Lancet 1982; fi: 68-70.

28 Howell R R. The interrelationship ofcolycogen storage disease and gout.Arthritis Rheum 1965; 8: 780-5.

29 Greene H C, Wilson F A, Hefferan P, etal. ATP depletion, a possible role in thepathogenesis of hyperuricaemia inglycogen storage disease type I. J ClinInvest 1978; 62: 321-8.

30 Raivio K 0, Becker M A, Meyer L J,Greene M L, Nuki G, Seegmiller J E.Effects of fructose infusion on urateproduction in man. Metabolism 1975;24: 861-9.

31 Van den Berghe G, Bronfman M,Vanneste R, Hers H G. The mechanismof adenosine triphosphate depletion inthe liver after a load of fructose.Biochem J 1977; 162: 601-9.

32 Nuki G, Parker J. Clinical andenzymological studies in a child withtype I glycogen storage diseaseassociated with partial deficiency ofhepatic glucose-6-phosphatase. AdvExp Med Biol 1979; 122A: 189-202.

33 Seegmiller J E, Grayzel A I, Liddle L,Wyngaarden J B. The effect of2-ethyl-amino-1,3,4-thiadizole on theincorporation of glycine into urinarypurines and uric acid in man.Metabolism 1963; 12: 507-15.

34 Willis R C, Carson D A, Seegmiller J E.Adenosine kinase initiates the majorroute of ribavirin activation in a culturedhuman cell line. Proc NatlAcad Sci USA1978; 75: 3042-4.

35 Giblett E R, Anderson J E, Cohen F,Pollara B, Menwissen H J. Adenosinedeaminase deficiency in two patientswith severely impaired cellularimmunity. Lancet 1972; ii: 1067-9.

36 Giblett E R, Ammann A J, Warn D W,

Sandiman R, Diamond L K. Nucleosidephosphorylase deficiency in a child withseverely defective T-cell immunity andnormal B-cell immunity. Lancet 1975;i: 1010-3.

37 Simmonds H A, Panayi G S, Corigal V.A role for purine metabolism in theimmune response, adenosine deaminaseactivity and deoxyadenosinecatabolism. Lancet 1978; i: 60-3.

38 Cohen A, Doyle D, Martin D W Jr,Ammann A J. Abnormal purinemetabolism and purine over-productionin a patient deficient in purinenucleoside phosphorylase. N EngI JMed 1976; 295: 1449-54.

39 Donofino J, Coleman M S, Hutton J J,Duoud A, Lomplein B, Difminski J.Over production of adeninedeoxynucleosides and deoxynucleotidesin adenosine deaminase deficiency withsevere combined immunodeficiencydisease. J Clin Invest 1978; 62: 884.

40 Cohen A, Goudas L J, Ammann A J,Staal G E J, Martin D W Jr.Deoxyguanosine tri-phosphate as apossible toxic metabolite in purinenucleoside phosphorylase deficiency. JClin Invest 1978; 61: 1405.

41 Hershfield M S. Suicide inactivation ofhuman lymphoblast S-adenosylhomocysteine hydrolase by2'-deoxyadenosine and adeninearabinoside: a basis for direct toxiceffects of analogs of adenosine. J BiolChem 1979; 254: 22.

42 Hershfield M S, Knedich N M,Ownby D R, Ownby H, Buckley R J. Invivo inactivation of erythrocyteS-adenosyl homocysteine hydrolase by2'-deoxyadenosine in adenosinedeaminase deficient patients. J ClinInvest 1979; 63: 807.

43 Carson D A, Wasson D B, Lakow E,Kamatami M. Biochemical basis forlymphocyte dysfunction in adenosinedeaminase and purine nucleosidephosphorylase deficiencies.J Clin ChemClin Biochem 1982; 20: 355.

44 Edwards M L, Magilavy D B, Cas-sidy J T, Fox I H. Lymphocyte ecto-5'-nucleotidase deficiency in agammag-lobulinaemia. Science 1978; 201:628-30.

45 Johnson S M, North M E, Asher-son G L, Allsop J, Watts R W E,Webster A D B. Lymphocyte purine5'-nucleotidase deficiency in primaryhypogammaglobulins. Lancet 1977; i:168-70.

46 Edwards M L, Gelfund E W, Burk L,Dosch H M, Fox I H. Distribution of 5'nucloeotidase in human lymphoidtissues. Proc Natl Acad Sci USA 1979;76: 3474-6.

47 Boss G R, Thompson L F, Spregel-heng H L, et al. Lymphocyte ecto-5'-nucleotidase activity as a marker ofB-cell maturation. Trans Assoc AmPhysicians 1979; 92: 309-15.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Studies of urate crystallisation in relation to gout

R. W. FIDDIS, N. VLACHOS, P.D. CALVERT

From the School of Chemistry and Molecular Sciences, University ofSussex, Brighton

Introduction Solubility

Gout is associated with the appearanceof crystals of monosodium uratemonohydrate (hereafter called sodiumurate) in the synovial fluid, causing aninflammatory reaction. There is a goodcorrelation between the incidence ofgout and raised serum uric acidconcentrations. In particular theoccurrence of gout increases rapidlywith concentration above thesaturation solubility of sodium urate inphysiological saline, about 0 4 mmolIl(7 mg/100 ml). Apparently we canview the development of gout asstemming simply from the process ofprecipitation from a supersaturatedsolution. We discuss what may belearnt about the development of thedisease by laboratory studies of uratecrystallisation and what this may tell usof the cause and treatment of gout.An analysis of the precipitation

process may be divided into a numberof stages. Firstly, we must establishthat the solution is supersaturated. Inundersaturated solutions any crystalswill dissolve, usually much faster thanthey grew, so we must be sure thatsupersaturation occurs at theprecipitation site and is constantlymaintained. Nucleation, the initiationof new crystals, often requires quitelarge supersaturations to occurspontaneously. The nucleation ratemay be enhanced by the presence ofsmall quantities of foreign particlessuch as dust so it is often quite difficultto determine why nucleation is easyunder some circumstances anddifficult under others. Havingnucleated, crystals grow at a rate whichdepends on the supersaturation. Atlow supersaturations growth rates maybe so slow that the crystals apparentlydo not grow at all. By combining aknowledge of physiological urateconcentrations with in vitroobservations on nucleation and growthrates we can estimate the time scale forcrystals to appear in vivo.

There have been severalmeasurements of the solubility ofsodium urate in water.1`3 This is astrong function of temperature but ismore or less independent of pH overthe range 6-5-6-9. Expressed as asolubility product [Na+l [HU-],, thesolubility is somewhat dependent onionic strength, increasing by about50% in physiological saline over thelow ionic strength value.' Figure 1shows our measurements for solubilitycompared to those of other workers.Measurements made by dissolution ofcrystals in a microscope hot stage tendto underestimate solubility as the stageoverheats while the crystals slowlydissolve.

60

E 40

20

/x*co

a1IA .0

0 ;

* a

* ox

0 3 06 0-9Concentrata n (rr,wo1/1)

Fig. 1 Solubility ofsodium urate.(-x-) This work, (-) Allen et al.,(l) Wilcox et al., (a) Wilcox et al. byhot stage, (0) Lam Erwin andNancollas.

One might expect solubility in vivoto be modified by binding of the urateto macromolecules or tissuecomponents. Only the free urate andfree sodium contribute to thesaturation. Although there are reportsof binding of urate to serum albuminand other proteins3 4 these effectsseem to be quite small. Neither weakbinding to an abundant protein norstrong binding to a protein with a low

concentration will have much effect.Also in vivo other anions will competeeffectively for the available bindingsites.These in vitro measurements predict

a saturation concentration of urate inserum of 0 5 8 mmol/l (9 7 mg/100 ml)at 37°C which is a bit higher than thevalue of 0 4 mmol/l (7 mg/100 ml)usually taken as the criticalconcentration for the onset of gout.The concentrations achieved by theincubation of urate crystals withhuman serum are in the range of0-3-0-5 mmol/l (6-8 mg/100 ml).5This lower value compared to salinesolutions probably reflects the effect ofthe non-electrolytes dissolved inserum.

Nucleation

Nucleation is easy to observe. Onesimply allows solutions to cool to adesired temperature and looks with amicroscope for precipitates aftersuitable times have elapsed. Theproblem is that nucleation may readilybe induced by small numbers ofsubmicroscopic foreign particleswhichprovide surfaces for heterogenousnucleation. The nucleating efficiencyof various surfaces may becharacterised by the degree ofsupersaturation at which they inducenucleation. Thus, according to thepicture in Fig. 2, we can expectnucleating agents to vary from

Sarationsolublity

Stable

Sbbk tb - -.Se.aabt/ _~~~~~- Supersolublity

/ #*kde ,-" linilit ysok,m,,"P1

Conoerration

Fig. 2 Schematic ofprecipitationregions.

T



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Studies of urate crystallisation in relation to gout Suppl p 13

inefficient ones, which induceprecipitation close to the'spontaneous' (supersolubility limit)line to efficient ones, which cause

precipitation at low supersaturations.Wilcox and co-workers have madeextensive studies of the effect ofvarious agents on nucleating uratesolutions.2 6 Calcium, decreasing pH,and mechanical shock have beenshown to enhance nucleation at highsupersaturations. Synovial fluids frompatients with gout were effective innucleating physiological saline uratesolutions at 1-8-2-1 mmolVl (30-35mg/100 ml), whereas no nucleationwas observed below 5 mmolIl (85mg/100 ml) in urate alone. Dialysisand uricase treatments of the synovialfluid did not remove this effect,implying the presence in patients withgout of a component that cannot beremoved by dialysis or dissolved inuricase and that is capable of inducingurate crystallisation. However, theeffect may be due to other crystalsformed during the handling of thesynovial fluid. No one has yet observednucleation in hyperuricaemic synovialfluids or at equivalent super-saturations in vitro.

Crystal growth

Figure 3 shows measurements ofcrystal growth rate as a function ofsupersaturation. In our laboratory we

69

i I

so I

0 /*° I

/

/s't/ x

/. x

2 4 6

Suprsctraion

Fig. 3 Growth rates ofsodium urateneedles against supersaturation (([Na ]

[HU-i"2'([Na+] [HU-]Sp)"2 - 1).(-I-) This work, 37'C, Na HU; (x)this work, 37'C, 0-14 M Na+; (O) thiswork, 37'C, 0 14M Na+ corrected; (U)Lam Erwin and Nancollas, 37'C, NaHU; (0) Allen et al., 50'C; 0-14 MNa+; (S) Allen et al., 50'C, 0 14 MNa+ corrected.

have followed microscopically thegrowth of individual crystals at 37°Cboth in aqueous solution and in saline.This technique does not allow us toextend our measurements down intothe range of hyperuricaemic fluids andextrapolation is necessary. Allen et al.used a similar technique to makemeasurements at 50'C and relativelyhigh concentrations." Lam Erwin andNancollas have recently reportedmeasurements of growth rate made byfollowing the solute depletion fromseeded solutions."° They express theirresults in terms of the reaction kineticsbut rve have extracted approximatevalues for the linear growth rates,which are also shown in Fig. 3.The important finding from our

results is that the growth rate is verystrongly dependent on super-

saturation. Expressed as a power

law, the linear growth rate varies as thesupersaturation, defined as (([Na+][HU I-)"2/([Na+] [HU-i,sp)"2 - 1), tothe 4-5 power. Such a strong depen-dence does not agree with the squarelaw, which is normally observed andfits the screw dislocation model of crystalgrowth but may be fitted to theexponential law of the surfacenucleation model."' 12 Ourmeasurements of urate growth insaline solution rather than equimolarsodium urate suggest that the same

supersaturation dependence isobserved if the data are normalised bydivision by the urate concentration(Fig. 3). Using this information we haveextrapolated our growth rate results bythe exponential law in order to

estimate growth rates in thephysiological range, these estimatesare given in Table 1.Lam Erwin and Nancollas initiated

their experiments in a narrow range ofconcentrations. They have, however,fitted the rate of solute depletion usinga square law dependence of growthrate on supersaturation. Thisprocedure also has its problems in thatone cannot easily take account ofcrystal multiplication and the chemicalanalysis of the solution must be very

precise. Table 1 also gives growth ratesbased on extrapolation according tothe square law.

It is important to know which ofthese laws applies to lowsupersaturations as this will make agreat difference to the time scale overwhich urate deposits can be assumedto develop in gout. The conclusions ofourselves and Lam Erwin andNancollas are not necessarilyincompatible in that cases are knownfor melt crystallisation whereexponential law growth is found unlessthe crystals are deliberately damagedwhen square law growth occurs.13 Toour knowledge, however, this effecthas never been observed in growthfrom solution. We certainly believe thegrowth is very slow at physiologicalsupersaturations as we have neverbeen able to observe growth of seedcrystals in hyperuricaemic serum at37°C.

DISSOLUTION

Lam Erwin and Nancollas alsomeasured urate dissolution rates. This

Table 1 Growth rates extrapolated to low supersaturation in saline

Temperature Urate Growth rate(OC) concentration

(mmol/l) pum/min punlyr

Exponential law extrapolation37 04 0 0

0-6 1- 12x 1-11 59x 10-60-8 Sixt07 0-27

32 0 4 17X 1013 88x 107-0-6 21x10- 0110-8 1-5x10- 7-7

27 0 4 8 8x10' 0-050-6 9 8x 10- 5-20-8 9-9x 10- 52

Square law extrapolation37 04 0 0

0-6 3 x O-' 150-8 2 x 10-4 100

Conversion: SI to traditional units-Urate: 1 mmolIl 17 mg/ 100 ml.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Suppl p 14 Annals ofthe Rheumatic Diseases

process is very fast compared togrowth and is limited by the rate atwhich the urate can diffuse away fromthe crystal surface.

MORPHOLOGYBy observing the morphology of uratecrystals grown in vivo one ought to beabile to deduce something about theirgrowth conditions. Rinaudo andBoistelle have recently studied themorphology of needle-like urates indotail. 4 The most significant facts arethat they are weak, brittle needles andhave a strong tendency to formspherical aggregates due to epitaxialnrcleation when crystallisation occursat large supersaturations. Similarspherical: aggregates are found,particularly in tophaceous gout."5Although there have been nosystematic studies it seems that thesupersaturations needed to producespherical aggregates on quiescentcrystallisation in vitro are much higherthan the supersaturations present invivo.We have also found that continuous

stirring or short bursts of ultrasonicirradiation may dramatically increasethe crystallisation rate of sodium uratecompared to that in unstirredsolutions. The crystallisation times, ofseveral days, are too long for this to besimply a mixing effect. We believe thatthis is the result of the fracture of thefine needles leading to a rapid increasein the number of effective growingcrystals and so to a great increase in thecrystallisation rate as measured bysolute depletion. In chemicalengineering studies this process isknown as secondary crystallisation. Inthe relatively immobile circumstancesof crystals embedded in cartilage or-synovium this could lead to sphericalcrystal aggregates forming.

POISONSCrystallisation inhibitors can act in anumber of ways to produce the sameend result of slowing the overallcrystallisation. They may bind thesolute and so reduce the availableconcentration; this will have an effectonly if the pool of solute is notreplenished by equilibrium with anoutside soifce such 'as'he blood andrequires quite high concentrations ofthe binder. Cartilage proteoglycan hasbeen implicated in'urate nucleation bysudden release -of bound urate'6 but

this binding now seems to haan artefact of the preptechnique."7 18 In veryconcentrations, poisons mEprevent nucleation by bin(surfaces that would otherwiseheterogeneous nucleationgenerally difficult to provepoison is binding to the undesmall surface area of an unucleating agent. Instead of bithe nucleating agent the poisbind to the surface of thecrystal and slow or prevyaddition of further molNancollas and Gardner havesuch an effect of pyrophospoxalate crystal growth.'9 A nudyes, including Bismarck brcmethylene blue, poison urate gthis way.8 As shown in Fig. 4studied this effect with neulserum albumin and a nunpolymers. Addition of 33neutral red to a 70 mmolI sodiisolution increased the crystatime by a factor of 50. Serumat about 10 g/l, as in synoviincreased the crystallisation tifactor of about 4, which is noeffect (Fig. 4). Additionsynovial fluid similarly incre;crystallisation time four fold.of the fluid showed that this w-a high molecular weight conTWe also found that heparin awas an effective crystalinhibitor. Lam Erwin and Nreport no effect from hepphosphonates at 10 ppm,"may be due to the very higharea of seed crystals usedstudies; this might simpladsorbed all the poisoIcrystallisations were unseedec

Log poison comention (A.n /I

Fig. 4 Effect ofpoison on

crystallisation halftimes, 0 07NaHU, 37°C. (-0-) Neutra(-X-) Serum albumin.

ve beenarative

Discussion

small The study of urate crystallisation hasay also brought up many interesting questionsding to and highlighted gaps in our knowledgeinduce of crystallisation from solution inIt is general. It should, however, be

that a possible to discuss crystallisation intectably gout without first having to solve allinknown the problems of crystallisation ininding to general.son may Solubility of sodium urate has beengrowing measured by many groups withent the generally good agreement. There is alecules. small but explicable discrepancystudied between measurements made in salinehate on solutions and in serum.imber of Nucleation is not observed inDwn and laboratory experiments atrowth in concentrations of less than 1- 8 mmoJlIwe have (30 mg/100 ml). This suggests thattral red, nucleation in vivo is probably a verynber of slow process. The work of Tak et al.mmol/l suggests that other particles in jointum urate fluid or cartilage may enhance theillisation nucleation rate.7 Urate crystals arealbumin very brittle so that once they appearial fluid, they can multiply rapidly throughime by a mechanical fracture.it a large The role of cartilage as the favouredof 4% nucleation site is still unclear. Crystalsased the forming in joints are subjected toDialysis mechanical fracture which will lead toas due to rapid development of a deposit. Theyiponent. are protected from early phagocytosis0t.01% and removal. The high charge densityIlisation of cartilage may also favour the[ancollas deposition of ionic crystals by reducing)arin or their effective surface energy.but this Crystal growth rates at low

tsurface supersaturations are still uncertain butin their all the evidence suggests that the timesly have taken for the growth of the crystalsn. Our usually seen in polarised lightd. microscopy are months or years rather

than hours. These rates are verysensitive to the precise value of the

or ° solubility, the instantaneouso concentrations, and the temperature./ The progress of nucleation and/x crystallisation will thus both be very

much enhanced by short termfluctuations to high uric acidconcentrations or low temperatures,

X_X- the extreme values are more importantthan the averages. There are few jointtemperature measurements available;Hollander et al. report normal kneetemperatures as being 33°C and ankle

mol/l temperatures as 29"C2° though this willl red, obviously vary with time. In the

absence of accessible data on



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Studies of urate crystallisation in relation to gout Suppl p 15

peripheral joint temperatures,measurements were made using athermocouple of the temperaturebetween two (strapped together) toesover a period of three hours. Thetemperature varied from 26 to 34°Cdepending on the extent of exertioncompared to a mouth temperature of37 5°C. Table 1 shows the effect ofsuch temperatures on the estimatedcrystallisation rates.The current therapeutic approach to

gout is to counter the inflammationwith colchicine and to permanentlyreduce the serum urate concentrationswith allopurinol or uricosurics.Growth poisons would be a possibletreatment but to maintain such aconstant concentration of methyleneblue, for instance, seems unpromising.If a specific nucleating species isidentified treatment aimed at itsremoval could be contemplated. Theslow growth rates and relatively rapiddissolution rate does suggest thatperiodic short term lowering of serumurates might be as effective as aconstant treatment. If theprecipitation cycle does take yearswhile the dissolution can be achievedin days then a few days ofnormouricaemia or hypouricaemia ayear would suffice to eliminate smalldeposits. This would be particularlyeffective if it was started ashyperuricaemia commenced ratherthan waiting for gout to appear.Accurate timing of such treatmentwould depend on a somewhat

improved understanding of therelationship between urateconcentration and precipitation invivo.

We would like to thank the MRC for thesupport of RWF during this work.

References

1 Wilcox W R, Khalaf A, Weinberger A,Kippen I, Klinenberg J R. Solubility ofuric acid and monosodium urate.Medical Biology & Engineering 1972;10: 522-31.

2 KhalafA A, WilcoxW R. Solubility andnucleation of monosodium urate inrelation to gouty arthritis. Journal ofCrystal Growth 1973; 20: 227-32.

3 Kippen I, Klinenberg J R, Weinberg A,Wilcox W R. Factors affecting uratesolubilityin vitro. Ann Rheum Dis 1974;33: 313-7.

4 Campion D S, Olsen R, Bluestone R,Klinenberg J R. Binding of urate byserum proteins. Arthritis Rheum 1975;18: 747-9.

5 Seegmiller J E. The acute attack ofgouty arthritis. Arthritis Rheum 1965; 8:714-25.

6 Wilcox W R, KhalafA A. Nucleation ofmonosodium urate crystals.Ann RheumDis 1975; 34: 332-9.

7 Tak H K, Cooper S M, Wilcox W R.Studies on the nucleation ofmonosodium urate at 37°C. ArthritisRheum 1980; 23: 574-80.

8 Allen D J, Milosovich G, Mattocks A M.Inhibition of monosodium urate needlecrystal growth. Arthritis Rheum 1965; 8:1123-33.

9 Allen D J, Milosovich G, MattocksA M.

Crystal growth of sodium acid urate. JPharm Sci 1965; 54: 383.

10 Lam Erwin C-Y, Nancollas G H. Thecrystallisation and dissolution of sodiumurate. Journal of Crystal Growth 1981;53: 215-23.

11 Fiddis R W D. Brighton: University ofSussex, 1982. Crystal deposition in rela-tion to physiological systems. Thesis.

12 Uhlmann D R. Crystal growth fromsolution: interface structure andinterface kinetics. In: Finlayson B,Hench L L, Smith L H, eds. Urolithiasis,physical aspects. Washington: NationalAcademy of Sciences, 1971: 169-82.

13 Ovsienko D E, Alfintsev G A. Crystalgrowth from the melt. In: Frey-hardt H C, ed. Crystals: growth andproperties. Berlin: Springer-Verlag,1980: 119-70.

14 Rinaudo C, Boistelle R. Theoretical andexperimental growth morphologies ofsodium urate crystals. Journal ofCrystalGrowth 1982; 57: 432-42.

15 Fiechtner J J, Simkin P A. Monosodiumurate monohydrate as spherulites. AdvExp Med Biol 1980; 122: 141-3.

16 Katz W A. Deposition of urate crystalsin gout. Arthritis Rheum 1975; 18,suppl: 751-6.

17 Perricone E, Brandt K D. Enhancementof urate solubility by connective tissue I.Arthritis Rheum 1978; 21: 453-60.

18 Perricone E, Brandt K D. Enhancementof urate solubility by connective tissueII. Ann Rheum Dis 1979; 38: 467-70.

19 Nancollas G H. The growth of crystals insolution. Advances in Colloid andInterface Science 1979; 10: 215-52.

20 Hollander J L, Stoner E K, Brown E M,De Moor P. Joint temperaturemeasurement in the evaluation ofanti-arthritic agents. J Clin Invest 1951;30: 701-6.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


GoutJ. T. SCOTT

From the Kennedy Institute ofRheumatology and Charing Cross Hospital, London W6

The pathogenesis of gout is now wellrecognised as depending on sustainedhyperuricaemia, leading in some cases(but only some) to urate crystaldeposition and then again to theinflammatory changes of acute goutyarthritis.

Hyperuricaemia

Epidemiological studies have shownserum urate concentrations to bedistributed on a normal curve, skewedtowards upper values, with those ofmen lying above those of women. In arecent study in the United Kingdommean concentrations were 330 AmoVl(5 5 mg/ 100 ml) t SD 60 ,umoVl (1- 0mg/100 ml) for men and 234,umol/1(3- 9 mg/ 100 ml) ± SD 48 ,umoVl (0 8mg/100 ml) for women.' Thus it ispossible to define hyperuricaemia instatistical terms (2 standard deviationsabove the mean), but this presentsconceptual difficulties in biologicalterms, especially in view of the skeweddistribution. Hyperuricaemia may alsobe defined in physicochemical terms,2depending on the solubility of urate inplasma at 37'C of about 420 ,umollI(7 0 mg/100 ml), but here again thereare uncertainties about such influencesas protein binding of urate, andcertainly prolonged supersaturationmay exist without crystal precipitation.For practical purposes, the upperlimits of normal may be taken as 420,umoVl (7`0 mg/100 ml) for men and360 ,umol/l (6-0 mg/100 ml) forwomen.

Individual factors contributing tohyperuricaemia consist of (a) thoseassociated with increased formation ofuric acid (Table 1), (b) thoseassociated with diminished excretionof uric acid (Table 2), and (c) those inwhich the exact nature of theassociation remains uncertain (Table3). These factors have beenextensively described3 and I will notdiscuss them further. Most patients

Table 1 Some factors associated withincreased formation of uric acid

Proliferative haemopoietic disorders:-for example, polycythaemia veraSpecific enzyme abnormalities:

Increased activity PRPP synthetaseDecreased activity HGPRT(Lesch-Nyhan)

Decreased activity glucose-6-phosphatase

Increased activity glutathionereductase

Drugs:2-ethylamino-1,3,4-thiadiazoleMethylene blueFructose

Diet

Table 2 Some factors associated withdiminished excretion of uric acid

Alterations in renal function and fluidvolumeDrugs:

PyrazinamideDiuretics

Lactic acidaemia:Toxaemia of pregnancyExerciseType 1 glycogen storage

diseasesAlcoholBeryllium disease

Starvation and ketosisEssential hypertensionLead poisoningHypercalcaemiaMyxoedema

Table 3 Other factors related touric acid concentrations

RaceSexBody weightHyperlipidaemiaIntelligence and 'drive'MongolismPsoriasisMyocardial infarctionSurgery

with primary gout have a genetictendency to hyperuricaemia(metabolic and renal) to whichenvironmental factors such as diet,alcohol, and drugs may in some casesexert a considerable additionalinfluence.The management of hyperuricaemia

and gout is usually a relatively easy andrewarding therapeutic exercise,4 theacute attack responding to colchicineor a wide range of non-steroidalanti-inflammatory drugs, andhyperuricaemia being controlled, ifnecessary, by uricosuric drugs or thexanthine oxidase inhibitor allopurinol.Currie reported that in generalpractice in the United Kingdom 45%of patients were receiving continuoustreatment with allopurinol comparedwith only 7% receiving uricosuricdrugs.5 The remaining 48% weretaking no regular drug treatment. Inthis context the anti-inflammatoryagent azapropazone is of someinterest: it is effective in the treatmentof acute gout and is also a uricosuricagent, though less powerful thanprobenecid' (C. S. Higgins and J. T.Scott, unpublished observations).Tienilic acid emerged as another drugof great potential value7; this druguniquely combines the properties of adiuretic and uricosuric agent, but it hasbeen withdrawn from study in theUnited States and Britain after reportsof hepatotoxicity.The question as to whether

allopurinol, by controlling urateconcentrations, has any effect inpreventing deterioration in renalfunction, is fully discussed in theimportant and detailed studies ofGibson and his colleagues' whoassessed renal function over a two yearperiod in two groups of patients, onereceiving colchicine and allopurinol,the other receiving colchicine only. Nosequential change was observed inthose receiving allopurinol, whereaspatients receiving colchicine alone



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Gout Suppl p 17

showed a statistically significant fall inglomerular filtration rate and urineconcentrating ability. Inspection of thefigures, however, shows that at theonset of the study these values werehigher in the colchicine group than inthe allopurinol group, concentrationsat one and two years closelyapproximating in both groups, so thequestion may still perhaps be regardedas open.The problem of asymptomatic

hyperuricaemia is also one ofconsiderable current interest.Asymptomatic hyperuricaemia hasbeen proposed as a risk factor in (a)gouty arthritis, (b) renal damage, and(c) ischaemic heart disease. There islittle doubt that hyperuricaemia mustpredispose to the development of goutbut the figures from the Framinghamstudy,9 which indicated a 90% risk inmen with serum urate concentrationsof over 540 ,umoVl (9 0 mg/100 ml)referred to very small numbers ofsubjects: subsequent observationshave indicated that the risk is not sogreat-for example, only three of 66such subjects developing gout afterfour years.'" In any case, gout can bedealt with if and when it occurs. Theproblem of renal damage is more

difficult, since hyperuricaemia maysometimes lead to the deposition ofinterstitial urate and renal uric acidcalculi. Here again, however, the riskmay not be as great as is sometimesfeared. The matter is an extremelydifficult one to investigate, but it isvery fully disussed by Fessel," whosestudies indicated that azotaemiaattributable to hyperuricaemia isgenerally mild and probably of noclinical importance, at least untilserum urate concentrations are veryhigh, and that the risk of urolithiasis issufficiently low to justify awaiting theoccurrence of a stone before loweringthe serum urate concentration. Thereare conflicting findings relating to uricacid as a risk factor in coronary heartdisease, but detailed studies" 13

indicate that hyperuricaemia is not anindependent risk factor after allowingfor variables of body weight, bloodpressure, and diuretic usage; there iscertainly no evidence that reducing theserum uric acid per se decreases therisk of coronary heart disease.

Urate crystal deposition

Table 4 lists some suggested

Table 4 Mechanism ofcrystalformation

1 Supersaturation of serum or synovialfluid with monosodium urate

2 Protein binding of urate3 Turnover of proteoglycans4 Temperature5 Trauma and exercise6 Altered hydrogen ion concentration7 Resorption of extracellular fluid8 Aging and avascularity

mechanisms relating to urate crystalformation. Some of these will perhapsbe discussed later in the symposium,but the degree and duration ofhyperuricaemia must surely be themost significant influences. This is wellillustrated by our experience of a

patient who developed the classicalneurological features of theLesch-Nyhan syndrome shortly afterbirth; diagnosis was confirmed at theage of 9 years by the demonstration ofvirtually complete absence oferythrocyte HGPRT.'4 Serum uricacid concentration at that time andsubsequently was about 720 AmolI(12-0 mg/100 ml) and urinary uric acidwas 10-8 mmol (1800 mg)/24 hours,but for certain reasons the parentsdecided that he should not receiveallopurinol. For 23 years his grosshyperuricaemia remained untreated

Fig. 1 Explosive onset ofacute goutwith tophaceous deposits in a youngman with the Lesch-Nyhan syndromeafter 23 years ofsevere asymptomatichyperuricaemia.

without the development of any jointsymptoms-or, incidentally, anydeterioration in renal function orevidence of urolithiasis. Then,suddenly, there was the explosiveonset of polyarticular gout withdeposition of numerous subcutaneoustophi (Fig. 1), events which the parentsfinally accepted as sufficientindications for treatment withallopurinol. Why had multifocal uratedeposition and inflammation takenplace so rapidly after such a longasymptomatic period?

Acute gouty arthritis

Table 5 lists some of the events thathave been proposed for thedevelopment of acute gouty arthritis.The fact that urate crystals can liewithin joints without producing acuteinflammation was first brought ratherstrikingly to my attention by the caseof a man with frequently recurrentacute polyarticular gouty arthritis, inwhom arthroscopy showed abundantdeposits of urate in synovialmembrane and articular cartilage,confirmed on biopsy. Control of serumurate concentration by allopurinolproduced rapid and complete freedomfrom symptoms, but appearance onarthroscopy and biopsy six monthslater was unchanged." Why the uratecrystals had ceased to causeinflammation is a mystery, but thestudies of Malawista et al, who foundthat the inflammatory response of cellsingesting silica crystals was greater in amedium containing a highconcentration of urate than in acontrol system,'6 may have a bearingon the problem. Certainly crystalshave been aspirated from the asymp-tomatic metatarsophalangeal joints ofpatients with gout," " observationswith which my colleagues and I agree.

In the evolution of gout the

Table 5 Events in acute gout

1 Crystallisation (or release from tophi)2 IgG coating3 IgG coated crystals meet Fc receptors of

neutrophils4 Phagocytosis5 Release of lysozomal enzymes and cell

lysis6 Activation complement, Hageman

factor, kinins and oxygen-derived freeradicals



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


occurrence of acute attacks normallyprecedes the deposition of clinicallyevident tophi, usually by a matter ofseveral years. The development oftophi in the absence of acute gout isrem-arkably rare. With my colleaguesDrs Hollingworth and Burry, I saw fivepatients with such developments,"9and it was of some interest to us thatfour of these five patients had varyingdegrees of renal functionalimpairment, while the fifth hadcoexistent rheumatoid arthritis.The rarity of acute gouty arthritis in

association with chronic renal failurehas been commented on from time totime; it has been thought that perhapsduration of renal disease may havebeen insufficient to allow crystaldeposition and acute gout.20 Ourpatients with tophi but without acutegout, however, allow speculation as toa further possibility-namely, thathyperuricaemia and crystal depositionoccur in patients with renalimpairment, but that the inflammatoryresponse is inhibited. Buchanan et alshowed that skin reactivity tointradermal or subcutaneous injectionof urate crystals was depressed inpatients with uraemia.2' Similarly, therarity of association of gout andrheumatoid arthritis has beenattributed to difference in age and sexdistribution and to diagnosticdifficulty, but again, it is possible thatorystal deposition is taking place inrheumatoid patients with hyper-uricaemia but that the inflammatoryresponse is inhibited by factors such asalteration in protein coating,22 syno-vial hypocomplementaemia,22 ordiminished phagocytic function ofpolymorphs in the synovial fluid ofpatients with rheumatoid arthritis.24

References1 Sturge R A, Scott J T, Kennedy A C,

Hart D P, Buchanan W W. Serum uricacid in England and Scotland. AnnRheum Dis 1977; 36: 420-7.

2 Wyngaarden J B, Kelley W N. Gout andhyperuricaemia. New York: Grune andStratton, 1976: 21.

3 Kelley W N, Weiner I M. Uric acid.Handbook of experimental pharma-cology, Vol 51. Berlin, Heidelberg, NewYork: Springer-Verlag, 1978.

4 Scott J T. Long term management ofgout and hyperuricaemia. Br Med J1980; 281: 1164-6,

5 Currie W J C. The gout patient ingeneral practice. J Rheum Rehab 1978;17: 205-18.

6 Dieppe P A, Doherty M, Whicher J T,Walters G. The treatment of gout withazapropazone: clinical and experimen-tal studies. European Journal ofRheumatology and Inflammation 1980;4: 392-400.

7 Reardon J A, Scott J T. Controlled in-patient study of tienilic acid in treatmentof gout and hypertension. Ann RheumDis 1980; 39: 367-72.

8 Gibson T, Rodgers V, Potter C, Sim-monds H A. Allopurinol and its effecton renal function in gout: a controlledstudy.Ann Rheum Dis 1982; 41: 59-65.

9 Hall A P, Barry P E, Dawber T R,McNamara P M. Epidemiology of goutand hyperuricaemia. Am J Med 1967;42: 27-3.

10 Fessel W J. Hyperuricaemia in healthand disease. Semin Arthritis Rheum1972; 1: 275-99.

11 Fessel W J. Renal outcomes of gout andhyperuricaemia. Am J Med 1979; 67:74-82.

12 Klein R, Klein B E, Cornoni J, Mar-eady J, Cassel J C, Tyroler H A. Serumuric acid. Its relation to coronary heartdisease risk factors and cardiovasculardisease. Arch Intern Med 1973; 132:401.

13 Persky V W, Dyer A R, ldris-Soven E,et al. Uric acid: a risk factor forcoronary heart disease? Circulation1979; 59: 969-76.

14 Bunn D N, Moss I K, Nicholls A,Scott J T, Snaith M L, Watson M R.

Clinical and biochemical observationson three cases of hypoxanthine-guaninephosphoribosyltransferase deficiency.Ann Rheum Dis 1975; 34: 249-55.

15 Scott J T. New knowledge of thepathogenesis of gout. J Clin Pathol[Suppl] 1978; 12: 204-13.

16 Malawista S E, Van Blaricom G, Cre-tella S B, Schwartz M L. The phlogisticpotential of urate in solution: studies ofthe phagocytic process in human leuko-cytes. Arthritis Rheum 1979; 22:728-36.

17 Agudelo C A, Weinberger A,Schumacher R, Tumer R, Molina J.Definitive diagnosis of gout by identifi-cation of urate crystals in asymptomaticmetatarsophalangeal joints. ArthritisRheum 1979; 22: 559-60.

18 Roualt T, Caldwell D S, Holmes E W.Aspiration of the asymptomaticmetatarsophalangeal joint in goutpatients and hyperuricemic controls.Arthritis Rheum 1982; 25: 209-12.

19 Holiingworth P, Scott J T, Burry H C.Non-articular gout: hyperuricemia andtophus formation without gouty arth-ritis. Arthritis Rheum 1983; 26:98-101.

20 Sorensen C B. The pathogenesis ofgout.Arch Intern Med 1962; 109: 55-66.

21 Buchanan W W, Klinenberg J R, Seeg-miller J E. The inflammatory responseto injected microcrystalline mono-sodium urate in normal, hyperuricemic,gouty and uremic subjects. ArthritisRheum 1965; 8: 361-7.

22 Abramson S, Hoffstein S T, Weiss-man G. Superoxide anion generation byhuman neutrophils exposed to mono-sodium urate.Arthritis Rheum 1982; 25:174-80.

23 Hasselbacher P. Immunoelectro-phoretic assay for synovial fluid C3 withcorrection for synovial fluid globulin.Arthritis Rheum 1979; 22: 243-9.

24 Turner R A, Schumacher H R,Myers A R. Phagocytic function ofpolymorphonuclear leukocytes inrheumatic diseases. J Clin Invest 1973;52: 1632-5.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Pathology of articular deposition of calcium salts andtheir relationship to osteoarthrosisDRAGOSLAV R. MITROVIC

From the Department of Cartilage Biology, U-18 ofINSERM, Lariboisiere Hospital, Pari&

Since the reports of Zitnan and Sitaj in1963,' articular chondrocalcinosis,also known as calcium pyrophosphatedihydrate (CPPD) crystal depositiondisease,' is now a well recognisedclinical entity. It is characterised byintra-articular deposition of CPPDcrystals, which encrust hyaline andfibrocartilages, synovium andsometimes intra-articular ligaments.The common finding of crystals in

synovial fluid provides a simple clinicaltest for assessing the microcrystallinenature of an acute pseudogoutattack,3 4the latter being a characteris-tic, though not most common, clinicalfeature associated with chondro-calcinosis.2Numerous radiological and clinical

reports have established three majorcircumstances that may influence theappearance of intra-articularcalcification or a pseudogout attack.

Occasionally, it is a familialdisorder' -1" with an autosomaldominant mode of transmissionestablished in a few kindreds."9 Moreoften, it is a sporadic disease of adultsusually found in association withhyperparathyroidism,'- 'S haemo-chromatosis,'6" ' Wilson's disease,'ochronosis" " and other hormonaland metabolic disorders.23 24 Ofinterest, though of minor importance,are those cases associated with hypo-phosphatasia3' and hypo-magnesaemia." 2 Trauma has beengiven as the explanation for meniscalcalcifications and more recentlychondrocalcinosis of the knee has beenlinked to meniscectomy.28An unusually high incidence of

chondrocalcinosis is found in theelderly,29 34 in whom calcificationsoften remain clinically silent,35 being

Correspondence to Doctor D Mitrovic,U-18 INSERM, 6, rue Guy Patin, 75010Paris.

discovered accidentally or after apseudogout attack, which maycomplicate infections, andcardiovascular diseases,36 majorsurgical interventions-particularlyparathyroidectomy37-or simplyprolonged confinement in bed.The frequent association of

chondrocalcinosis and osteoarthrosishas aroused speculation about thepathophysiology and causalrelationship of these twoconditions.29-34 383S Evidence suggeststhat CPPD crystal deposition mayinduce severe arthropathies thatresemble osteoarthrosis, particularlyin a few familial forms ofchondrocalcinosis' 7 where CPPDcrystal deposits affect youngerindividuals and thus precede thedevelopment of osteoarthrotic likearthropathies by many years.Osteoarthrosis of the wrist joint isunusual except in chondrocalcinosis'and is considered by some as thedistinct feature of disease. Moreimportantly, certain arthropathies ofthe large joints follow much morerapid and severe courses whenassociated with CPPD crystaldeposits.43 43 Finally, the frequentfinding of CPPD crystals in the tissuesor synovial fluid of osteoarthroticjoints suggest more than a causallink between these two conditions.

In earlier papers'" 44 we describedthe histological and histochemicalchanges that have occurred in thecartilage, menisci, and synovium of theknee joints of two patients affected bysporadic chondrocalcinosis, one ofthem being associated withhaemochromatosis.'9 In more recentstudies34 I we have conducted postmortem examination of more than 250knees searching for articular cartilage,degenerative changes and CPPDcrystals deposits. We report here ourmain conclusions and try to summarise

the known data on the pathology ofCPPD crystals deposition.

Mineral deposits in articularchondrocalcinosis

Intra-articular mineral deposits,unlike all others found under variousphysiological or pathologicalconditions,47 are composed ofmonoclinic and triclinic CPPDcrystals.45"49 Other forms of calciumsalts, such as hydrogen phosphatedihydrate,46 hydroxyapatite,46 andoccasionally brushite,47 or evensodium phosphate were found, but it isstill unclear whether such crystals arecontaminants from calcified cartilageand bone or are formed during theprocess of abnormal intra-articularmineralisation. Recent studies havepointed out the possibility of changesin the crystal organisation andstructure during in vitro handling ofbiological materials (S Wilhelm andothers, paper presented at 15thInternational Congress ofRheumatology, Paris, 1981).Examined under polarised light, the

deposits show weak to strong positivebirefringence which disappear afterdecalcification. Microincineration ofthe non-decalcified sections followedby dark field examination underreflected light is useful for identifyingmineral deposits. Using this technique,we have been able to demonstrate theabsence of iron in CPPD tissuedeposits in one case ofhaemochromatosis.1'With routine haematoxylin-eosin

procedure, deposits appear even ondecalcified sections as characteristicintense blue spots. They show onlyweak metachromasia with toluidineblue44 and faintly stain with alcianblue.' 44 Histochemical procedures forcalcium salts give irregular results andusually reveal only small deposits or



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


those located at the periphery of largeones.

Seen under the scanning electronmicroscope, CPPD crystals of varioussizes appear as triclinic parallelopiped,rhomboid or rod-like structures.45Under transmission electron

microscope, two different types ofcrystals are reported.4 48-50 Most oftenthe crystals are identified as CPPDmonoclinic and triclinic rod-likeforms of various sizes.2 4 4 4 Thesecrystals are resistant to a variableextent under electron beam, some

remain homogenous and opaque whileothers acquire a foamy likeappearance (Fig. 1).4 40

" The secondtype of crystals seen on electronmicroscopy is a structure composed ofnumerous small electron-denseneedles that resemble hydroxyapatite(Fig. 2).5 In one of our cases thesewere seen, as reported elsewhere,50 inassociation with typical electronopaque and foamy CPPD crystals.

In articular chondrocalcinosisdeposits are usually located withinfibrocartilages and hyaline cartilagesof knees, wrists, symphysis pubis, andintervertebral discs, although virtuallyany joint may be affected. Synovialtissue and intra-articular ligaments are

affected less often.5" In a few instancesCPPD crystal deposits have beenreported in extra-articular structures

Fig. 1 Cartilagenous CPPD crystaldeposit observed on unstained sectionwith transmission electron microscope.The rod-like crystals exhibit bubbledappearance under electron beam. Asurrounding material is homogeneousand weakly stained. H and E x 48000(original magnification). Bar: 0 5 ,um.

Fig. 2 Structure seen under theelectron microscope in articularcartilage that exhibited CPPD andamyloid deposits. The 'crystal' iscomposed of tightly packed smallgranules or needles which may resemblehydroxyapatite. Surrounding materialis presumably amyloid. Unstainedsection x 24000 (originalmagnification). Bar: 1 p,m.

such as ligaments,52-54tendons47 5 anddura.56 57 In most of these cases, theextra-articular deposits were

associated with multiple articularchondrocalcinosis.5255

Meniscal and articular cartilagedeposits

CPPD crystal deposits in the kneejoint have been studied thoroughly. Inan earlier paper, McCarty et al.described three radiological aspects ofmeniscal calcifications, presumablycorresponding to different types ofcrystals.46 Indeed, using similarmethods, we have observed in all buttwo of 31 joints affected bymeniscocalcinosis the aspect shown inFig. 3 which, according to McCarty etal. should contain CPPD crystals. Inone case the calcification had an

appearance of ectopic ossification andin the other, it was seen as diffuse

Fig. 3 Roentgenograms using highcontrastfilms ofmenisci ofknee joint of73 year old woman. Mineral depositsappear as irregular granules andlamellae infiltrating 2/3 of the internalportion ofmenisci (ref ).

infiltration of severely damagedmedial meniscus. Unfortunately, ourpreliminary crystallographic analysis ofthese deposits is too incomplete tocome to a definite conclusion aboutthe nature of these crystals.Under a light microscope, meniscal

and chondral calcifications usually areseen as numerous sharply delimitedmultifocal deposits of various sizescutting off the matrix beneath thearticular surface (Fig. 4). In familialcases and in younger individuals,48 5"60the articular surface was reported to beintact. Most cases of sporadic andsenile chondrocalcinosis, however,have damaged surfaces."8 44 50 5 Thelocation of the deposits varies fromone case to another and from one sitein the joint to another. The depositsare sometimes seen as relatively largeplaques that may be easily detached,leaving behind an eroded but smoothsurface. On histological sections ofarticular cartilage they may lie undernormal and fibrillated surfacesdepending on the area chosen forsampling. Usually, fibrillated andvertically fissured cartilage is seen inassocation with rather large depositsand in elderly persons. In menisci ofthe knee joint, CPPD crystal depositsare usually located in thefibrocartilagenous internal portionbeneath the surface of both facets butsometimes they are located moredeeply in the tissue. The affectedmenisci usually look normal except forthe presence of CPPD crystal deposits,though in the case of severe

Fig. 4 CPPD crystal deposits (inblack) are seen in femoral condylararticular cartilage of54 year old manaffected by haemochromatosis.Deposits lying beneath a fibrillatedsurface which shows a few verticalclefts. Tissue below deposits hasnormal appearance. Von Kossa'sstaining x 50 (original magnification).Bar: 250 pAm.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Pathology of articular deposition ofcalcium salts Suppl p 21

osteoarthrosis they may be frayed anddeformed. In a few cases of severelydamaged articular cartilage andmenisci, we have seen CPPD crystaldeposits as diffuse infiltrations locatedexclusively in the superficial fibrillatedportion of the tissue. These deposits,unlike the multifocal ones describedabove, exhibit intense metachromaticstaining. Seen under the transmissionelectron microscope, the CPPDcrystals may lie in a matrix of normalappearance,"S-2 58 59 61-63 or be closelypacked in some areas (Fig. 5),radiating from the centre of theclusters.48 S Certain authors haveobserved CPPD crystals surroundedby an electron dense granularmaterial50 and by cell debris.58 59 Thismaterial, seen in two of the menisci,was intensely stained by rutheniumred.-"CPPD crystals were occasionally

seen close to cell lacunae"8 S but neverwithin chondrocytes. Recently, weobserved a CPPD crystal within avacuole of a chondrocyte with anecrotic appearance (Fig. 6). This was,however, in necropsy material and wasan isolated finding. In one reportCPPD crystals were seen in thecalcified layer of articular cartilagelying within a hydroxyapatiteimpregnated matrix,63 but this was notconfirmed elsewhere. In ourexperience, the calcified basal layermay be affected by invasion of largecrystal deposits from soft uncalcifiedcartilage.Changes in articular cartilage and

menisci have been replrted by severalauthors.44 " Usually, a loss ofmetachromasia of the superficial layerinvaded by CPPD crystal deposits isseen in association with cartilagefibrillation.44 The chondrocytes mayproliferate and form clones that areusually found along the cartilage clefts.Slightly stained zones of matrixdegeneration not related to CPPDcrystal deposits have also beenreported.44 48 62 In one of our sporadiccases, these stained differently from asurrounding matrix, and exhibited noperiodic acid Schiff colouration. Seenunder electron microsopy, abnormalareas of uncalcified matrix containedfragmented collagen fibres and matrixof a higher electron density.59 62 Thechondrocytes were studiedhistochemically in one case of sporadicchondrocalcinosis associated with

Fig. 5 Electron micrograph ofunstained section showing CPPD crystal depositin articular cartilage. Electron opaque rod-like crystals ofvarious sizes are runningin all directions inside an area surrounded by a rim ofcondensed matrix and almostdevoid ofother visible material. However, in areas oflower crystal density anhomogenous hyaline matrix, presumably collagen, is present in between thecrystals. x 20400 (original magnification). Bar: I p,m.

haemochromatosis.'8 It was found thatthe cells located between CPPDcrystal deposits were larger, had ahigher content of acid proteins andsulphydryl groups, and weresurrounded by a rim of periodic acidSchiff + material. As these cells wereloaded with iron-containing material,it was not possible to ascribe thesechanges to the presence of CPPDcrystals. Necrotic chondrocytes havefrequently been found close to CPPDcrystal deposits.4"50 58 59 62

Synovial deposits

Synovial deposits are not ascommon or extensive as meniscal andchondral calcifications inchondrocalcinosis. When found (threeout of 22 joints in our experience) they

are usually seen in the fibrous tissue asround or oval, sharply delimitedcalcifications, and as diffuse depositsinfiltrating superficial synovioblasticand subintimal layers. Most oftenthese calcifications do not induce anycell changes. In one case they werefound in the superficial subintimallayer and were surrounded bynumerous histiocytic andmultinucleated giant cells, whichresembled small tophi. Fibrin anddiscrete perivascular cell infiltrateswere present. The synovium of 21knee joints affected by senilemeniscochondrocalcinosis comparedhistologically with that of joints ofnormal subjects matched for age,revealed no significant difference inthe intensity of synovialinflammation.64



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


.1trg. ,::o t t ' '%.,. ...yo..............A f . 4w

Fig. 6 Electron micrograph showing typical CPPD crystal (arrow) in vacuole ofa rather active chondrocyte. Anothercrystal (arrow) is present in matrix next to thechondrocyte's lacunae. Acetate uranyl-lead citrate staining. x 11200 (originalmagnification). Bar: 2 p.m.

In a few cases with intra-articularCPPD deposits, clusters of crystalswere seen under polarised light in thesuperficial synovioblastic layer.47 50

Ultrastructural studies performed onthe synovium in the CPPD jointsusually revealed the presence ofcrystals in the interstitial space44 and inthe vacuoles of macrophage-likesynovioblasts,49- 50 65- 67 but onlyoccasionally in the endoplasmicreticulum of secretory cells,50suggesting their possible localproduction.

Synovial fluid deposits

These crystals are often found onsimple microscopical examination of adrop of unstained synovial fluid. Thishas become a useful clinical test forestablishing CPPD crystal depositiondisease.3 4 CPPD and, on a fewoccasions, hydroxyapatite6"71 crystalsare most often found in the vacuoles ofpolymorphonuclear and mononuclearmacrophage-like cells. They disappearon addition of a few drops of acetic

acid or a solution of edetic acid, so thatthey may easily be differentiated frommonosodium urate crystals.3 I

The origin of crystals in the synovialfluid is not known. Most investigatorsbelieve that these crystals derive fromarticular cartilage and menisci byshedding of the deposits. They may,however, be formed in synovial fluid bythe crystallisation of soluble CPPsalts.72When calcium hydroxyapatite

crystals are identified, they may, atleast in part, derive from the calcifiedlayer of articular cartilage andsubchondral bone. In severedestructive arthropathies, theresorption of calcifiedosteocartilagenous fragments by thesynovial cells (Fig. 7) is a commonfinding.73

Ultrastructural studies have shownthat CPPD crystals occur in thevacuoles of phagocytic cells togetherwith other engulfed material such asfibrin, immunoglobulins, and celldebris.4 4 5 7 The membranepreservation of phagocytic vacuoles

Fig. 7 Synovium taken from anosteoarthrotic knee joint. Severalosteocartilaginous fragments (arrow)are seen per section. No calcificationswere recorded in menisci and articularcartilage. H and E x 300 (originalmagnification). Bar: 50 ,um.

that contain crystals was found to besatisfactory, but in vitro studies havesuggested that CPPD crystals mayinduce a membranolytic effect thoughto a much lesser extent thanmonosodium urate crystals.75CPPD crystals may absorb proteins

such as IgG and induce phagocytosisthrough surface Fc receptor sites, orthey may activate Hageman factor andcomplement.76 The addition of CPPDcrystals to the culture medium inducesan appreciable increase in theproduction and release of collagenase,neutral proteases, and prostaglandinsby human rheumatoid and normalrabbit synovioblasts, though the effectis much more pronounced in thepresence of calcium hydroxyapatitecrystals.77 The phlogistic action ofCPPD crystals is treated in more detailelsewhere in this issue.

Relationship between CPPD crystaldeposits and osteoarthrosis

Clinical and radiological evidencesuggests a relationship between CPPDcrystal deposits and osteoarthrosis.The possible undermining effect ofCPPD crystals deposits on theanatomical integrity of articularcartilage and other joint tissues doesnot rule out the existence of morespecific interactive processes that mayfavour either CPPD crystal formationor cartilage degeneration.

Chondrocalcinosis appears to be anindependeq,t clinical entity in thefamilial forms of disease where youngindividuals are affected. In such casesthe occurrence of CPPD crystaldeposits in anatomically normal tissueis followed by the development of



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Pathology ofarticular deposition ofcalcium salts Suppl p 23

degenerative-like arthropathies whichappear in the later stages of disease,probably as a result of both theirundermining effect andcrystal-induced inflammation."The problem is more difficult to

analyse in sporadic and senilechondrocalcinosis. Arthropathies are

common in certain metabolic dis-orders such as haemochromatosis, 19

ochronosis,2" 22 and Wilson's disease,'which are also known to have highincidence of intra-articular CPPDcrystal deposits. Whether sucharthropathies develop in the absenceof CPPD crystal deposits is not wellestablished. In senile forms ofchondrocalcinosis, the frequentassociation with osteoarthrosis may beaccidental and due to the increasedincidence of both disorders in elderlyindividuals.29-34We examined more than 250 knee

joints of unselected cadavers obtainedfrom a pluridisciplinary hospital inParis.34 3 64 The joints were examinedwith the naked eye before and afterstaining their articular surfaces withIndia ink. Articular and meniscalcalcifications and joint lesions were

recorded on specially designeddrawings. Then the menisci andarticular cartilage fragments fromfemoral condyles were dissected andsubjected to radiological andmicroscopical examination. Multiplesynovial fragments were taken from100 joints of 50 cases.

CPPD crystal deposits wereidentified by soft x-ray radiography ofthe menisci and cartilage fragments as

described previously,34 s and on thehistological sections using routinestaining procedures and polarisedlight. The incidence of CPPD crystaldeposits, their distribution within thejoint and in the tissues, and theirrelationship to the age and sex of thesubjects and to osteoarthrosis, was

thus established. Osteoarthrosis was

diagnosed in cases of extensive andsevere cartilage fibrillation and/orulceration associated with peripheralosteophytosis. CPPD crystal meniscaland chondral deposits were found in18-5% of all subjects: 21-5% ofwomen and 15-8% of men. In about40% of these cases meniscocalcinosiswas not associated withchondrocalcinosis. In all cases

chondral calcification affected themenisci.

In two cases not included in our

series, an abundant chalk-like powderdeposited on all tissue surfaces wasobserved in the joint but no

calcification was detected on the x-rayfilms of menisci and cartilagefragments. This material may havebeen present on the surface and notwithin the tissue and might thereforehave been lost while processing thetissue. Unfortunately, we were notable to perform any identificationstudies on this material.The mean age of the subjects

affected by CPPD crystal deposits was78 years for men and 82 years forwomen. The incidence of depositsincreased with age: none was seen inthose under 60; the incidence thenrose to 11-7% in those aged 60-69,21-2% in those aged 70-79, 26-9% inthose aged 80-89 and 50% in thoseaged over 90 years. In all those over

80, the incidence was 32%. The mean

age of the subjects affected bymeniscal deposition without chondralcalcifications was slightly, butinsignificantly, lower. Lateral menisciwere more often affected.

Calcifications were seen more

frequently on condyles than in thefemoropatellar joint, which was

affected in about 20% of all cases.

Calcifications were usually found infibrillated areas of articular cartilage:apparently normal surroundingcartilage was also affected but to a

much lesser extent. In addition, 20%of the joints with meniscal andmeniscochondral calcifications were

normal.We attempted to relate the presence

of intra-articular CPPD crystaldeposits to the degree of knee jointdamage.3 Out of 78 knees of thesubjects below 75 years of age (this ageappeared to be critical in our serieswith respect to CPPD crystaldeposition), meniscal or menisco-chondral calcifications or both were

found in three, femorotibial arthrosisin seven, and cartilage fibrilation (notcharacterised as osteoarthrosis) in 35joints. Two out of three joints affectedby CPPD deposits were

osteoarthrotic, and one showedfibrillation of femoral condyles. Inthose subjects over 75 years of age

(110 knees), CPPD crystal depositswere identified in 35, femorotibialosteoarthritis in 34, and fibrillation ofcondylar articular cartilage in 39

joints. Cartilage fibrillation of thetibial plateau was found in almost alljoints but was not considered in thisstudy. In this group 19 out of 35 jointsaffected by CPPD crystal deposits hadfemorotibial osteoarthrosis (54 2%);four had fibrillated cartilage surfaces,and 12 (34 2%) were normal. In thoseover 75 years of age who did not haveintra-articular CPPD deposits (75joints), femorotibial osteoarthrosiswas found in 15 (20%) and fibrillationin 39 joints. The difference inincidence of osteoarthrosis betweenthose joints with CPPD crystaldeposits (54-2%) and those without(from cases aged over 75) is highlysignificant (p<0-001, x2 test). Theincidence of osteoarthrosis was foundto be even higher (80%) in those jointswhere cartilage deposits were seen.Not only were osteoarthrotic lesions

more frequent in the joints affected byCPPD crystal deposits but these jointsappeared to be much more severelydamaged.

Discussion

No definite conclusion could bedrawn from the results obtained in theyounger subjects, considering thesmall amount of positive findings.Nevertheless, a high incidence ofcartilage fibrillation occurred in theabsence of visible CPPD crystaldeposits, suggesting that the CPPDcrystal deposits are not the cause offibrillation at least in these cases.Our findings are in agreement with

those of previous reports41-43 andclearly show that osteoarthroticlesions are more common and moresevere in the joints affected by CPPDcrystal deposits than in those that arenot.

Cartilage fibrillation, which isthought to be the initial lesion inosteoarthrosis, may occur, at least inits first stages of development, in theabsence of CPPD crystal deposits.However, CPPD deposits in

articular cartilage seem to affect thefibrillated areas more frequently andmore heavily, and in a few instancesare found exclusively in the superficialfibrillated layer of the tissue. Thiswould suggest that local factors areimportant in determining the tissuelocalisation of CPPD crystal deposits,and that fibrillated cartilage somehowpredisposes to disease. The finding of



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


higher concentrations of inorganicpyrophosphate in the synovial fluid ofosteoarthrotic patients78 and in vitroproduction by incubates ofosteoarthrotic articular cartilage,'would further support this view.

However, as shown in this andprevious studies, CPPD crystaldeposits may also occur inanatomically normal joints.47-50 58-6163 64 This is all the more significant ascartilage fibrillation is a particularlycommon phenomenon in patientsover 75. However, it is unclearwhether the affected tissue is normalor not. Zones of abnormal cartilagematrix coexist with calcified areas andmay thus represent the initial lesion inchondrocalcinosis.44 58 62 Recent invitro studies of CPPD crystaldeposition from solution in artificialand cartilage matrices861 "3 arepromising for study into the nature ofthose local factors which may triggercrystal nucleation and growth.

In conclusion, most currentevidence suggests that cartilagefibrillation and CPPD crystaldeposition are independent processes.Mineral deposition may, however,cause or accelerate cartilage damageby secondarily altering chondrocytemetabolism or simply by decreasingtissue strength and elasticity. On theother hand, particular disorders intissue metabolism and structure, notexpressed in all cases of osteoarthroticlesion, may constitute a basicabnormality which would account forCPPD crystal deposition. Morebiochemical and metabolic studies onselected cases are needed to elucidatethese and other unexplained points.

We thank Professor A Ryckewaert for hishelpful discussions; Doctors: A Stankovic, JMorin, 0 I Borda, M Quintero, and M Uzanfor their collaboration in recent years; andMrs F Aprile for typing the manuscript.

This work was supported by the grants ATPNo 6978101 and ATP No 7679108 ofINSERM.

References

1 Zitnan D, Sitaj S. Chondrocalcinosisarticularis. Section I. Clinical radiologystudy. Ann Rheum Dis 1963; 22,142-52.

2 McCarty D J. Calcium pyrophosphatedihydrate crystal deposition disease1975 Arthritis Rheum, 1976; 19, suppl:275-85.

3 McCarty D J, Kohn N N, Faires J S. Thesignificance of calcium phosphatecrystals in the synovial fluid of arthritispatients: the 'pseudogout syndrome'. I.Clinical aspects. Ann Intern Med; 1962;56: 711-37.

4 Solnica J, Mitrovic D. Kahn M F.L'intfret de la mise en evidence desformations cristalines dans le liquidesynovial: la differenciationmorphologique et ultrastructurale descristaux. Sem Hop Paris 1967; 43:2573-80.

5 Louyot P, Peterschmitt J, Berthelme P.Chondrocalcinose articulaire diffusefamiliale. Rev Rhum Mal Osteoartic1964; 31: 659-63.

6 Van der Korst J K, Geerards J,Driessens F C M. A hereditary type ofidiopathic articular chondrocalcinosis.Survey of a pedigree. Am J Med 1974;56: 307-14.

7 Reginato A J, Hollander J L, MartinezV, et al. Familial chondrocalcinosis inthe Chiloe Islands, Chile. Ann RheumDis 1975; 34: 260-8.

8 Bjelle A, Edvinson U, Hagstram A.Pyrophosphate arthropathy in twoSwedishfamilies.ArthritisRheum 1982;25: 66-74.

9 Gaucher A, Faure G, Netter P, Pourel J,et al. Hereditary diffuse articularchondrocalcinosis. Dominantmanifestation without close linkage withthe HLA system in a large pedigree.Scand J Rheumatol 1977; 6: 217-21.

10 Rodriguez-Valverde V, Tinture T,Zuniga M, Pena J, Gonzales A. Familialchondrocalcinosis. Arthritis Rheum1980; 23: 471-8.

1 1 Moskowitz R W, Katz D.Chondrocalcinosis andchondrocalcinosynovitis (Pseudogoutsyndrome). Analysis of twenty-fourcases. Am J Med 1967; 43: 322-34.

12 Bywaters E G L, Dixon A St, Scott J T.Joint lesions of hyperparathyroidism.Ann Rheum Dis 1963; 22: 171-87.

13 Dodds W J, Steinbach H L. Primaryhyperparathyroidism and articularcartilage calcification. AJR 1968; 104:884-93.

14 Serre H, Simon L, Thevenet A, BarjonM C. Hyperparathyroidie etchondrocalcinose articulaire. JournalBelge de Medecine Physique et deRhumatologie (Bruxelles) 1966; 21:223-42.

15 Ryckewaert A, Solnica J, Lanham C,Seze S de. Les manifestationsarticulaires de l'hyperparathyroidie.Presse Medicale 1966; 74: 2599-603.

16 Dorfmann H, Solnica J, Dimenza Cl,Seze S de. Les arthropathies deshemochromatoses. Sem Hop Paris1969; 8: 516-523.

17 Dymock J W, Hamilton E B, Law J W,Williams R. Arthropathy ofhaemochromatosis: clinical and

radiological analysis of 63 patients withiron overload. Ann Rheum Dis 1970;29: 469-76.

18 Mitrovic D, Mazabraud A, Jaffres R, etal. Etude histologique et histochemiquedes lesions articulaires de lachondrocalcinose survenant au coursd'une hemochromatose. Archivesd'Anatomie Pathologique (Paris) 1966;14: 264-70.

19 Seze S de, Hubault A, Kahn M F,Welfling J, Jaffres R, Mitrovic D,Solinica J. Les arthropathies deshemochromatoses. Sem Hop Paris1966; 42: 2472-82.

20 Boudin G, Pepin B, Hubault A. Lesarthropathies de la maladie de Wilson.Rev Rhum Mal Osteoartic 1974; 10:594-98.

21 Reginato A J, Schumacher H R,Martinez V A. Ochronotic arthropathywith calcium pyrophosphate crystaldeposition. A light and electronmicroscopic study. Arthritis Rheum1973; 16: 795-814.

22 Rynes R I, Sosman J L, Holdsworth LE.Pseudogout in ochronosis. ArthritisRheum 1975; 18: 21-5.

23 McCarty D J, Silcox D C, Coe F.Diseases associated with calciumpyrophosphate dihydrate crystaldeposition. A control study. Am J Med1974; 56: 704-1 4.

24 Sany J, Rosenberg F, Bataille F R, SerreH. Chondrocalcinoses secondaires etassociees. Rev Rhum Mal Osteoartic1977; 44: 565-77.

25 O'Duffy J K. Hypophosphatasiaassociated with pyrophosphatedihydrate deposits in cartilage. ArthritisRheum 1970; 13: 381-8.

26 Rapado A, Castillo J M, Abad J A.Chondrocalcinosis e hypomagnesemia:un nuevo syndrome. Rev Esp RhumEnferm Osteoartic 1976; 3: 283-91.

27 Milazzo S C, Ahem M J, Cleland L G,Henderson D R F. Calciumpyrophosphate deposition disease andfamilial hypomagnesemia. J Rheumatol1981; 8: 767-71.

28 Doherty M, Watt I, Dieppe P A.Localised chondrocalcinosis inpost-meniscectomy knees. Lancet 1982;i: 1207-10.

29 Ellman M H, Levin D.Chondrocalcinosis in elderly persons.Arthritis Rheum 1975; 18: 43-7.

30 Delauche M C, Stehle E, Cassou B,Verret J M, Kahn M F. Frequence de lachondrocalcinose radiologique apres 80ans. Une 6tude prospective. Rev RhumMal Osteoartic 1977; 44: 555-7.

31 Memin Y, Monvilie Cl, Ryckewaert A.La chondrocalcinose articulaire apres80 ans. Rev Rhum Mal Osteoartic 1978;45: 77-82.

32 Leonard A, Solnica J, Cauvin M, et al.La chondrocalcinose: etude de safrequence radiologique et de sesrapports avec l'arthrose. Etude du taux



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Pathology of articular deposition ofcalcium salts Suppl p 25

de parathormone. Rev Rhum MalOsteoartic 1977; 44: 559-64.

33 Megard M, Vignon E, Arlot M, et al.Etude de la chondrocalcinose articulairedans une population de 200 vieillards.Ses liaisons avec l'antigene HLA. Lesrelations avec l'arthrose. Lyon Medicine1981; 245: 365-70.

34 Mitrovic D, Stankovic A, Morin J, et al.Frequence anatomique de lamenisco-chondrocalcinose du genou.Rev Rhum Mal Osteoartic 1982; 49:495-99.

35 Emeriau J P, Borde C, Chapoulart H, etal. Chondrocalcinose articulaireasymptomatique chez le sujet age.Bordeaux Medicine 1977; 10: 825-30.

36 O'Duffy J D. Clinical studies of acutepseudogout attacks. Arthritis Rheum1976; 19: 349-52.

37 Glass J S, Graham R. Chondrocalcinosisafter parathyroidectomy. Ann RheumDis 1976; 35: 521-29.

38 Dieppe P A, Doyle D V, Huskisson E C,Willoughby D A, Crocker P R. Mixedcrystal deposition disease andosteoarthritis. Br MedJ 1978;i: 150.

39 Stankovic A, Mitrovic D, Morin J,Memin Y, Seze S de, Ryckewaert A.Relation entre la chondrocalcinosearticulaire et l'arthrose du genou chez lesujet age. Rheumatology 1982; 34:35-9.

40 Utsinger P D, Resnick D, Zvaifler N J.Wrist arthropathy in calciumpyrophosphate dihydrate depositiondisease. Arthritis Rheum 1975; 18:485-92.

41 Menkes C J, Simon F, Churaki M,Ecoffet M, Amor B, Delbarre F. Lesarthropathies destructives de lachondrocalcinose. Rev Rhum MalOsteoartic 1973; 40: 115-23.

42 Richards A J, Hamilton E B D.Destructive arthropathy inchondrocalcinosis articularis. AnnRheum Dis 1974; 33: 196-203.

43 Villiaumey J, Galle P, Amouroux J,Larget-Piet B, Rotterdam M, DimenzaC. Arthropathies lytiques etchondrocalcinose articulaire. Sem HopParis 1974; 50: 3175-3190.

44 Seze S de, Fressinaud L, Beeson J,Mazabraud A, Mitrovic D. Etudeanatomo-clinique d'un cas dechondrocalcinose articulaire diffuse.Sem Hop Paris 1963; 39: 1515-25.

45 Gaucher A, Faure G, Netter P,Malaman B, Steinmetz J. Identificationof microcrystals in synovial fluids bycombined scanning electron microscopyand x-ray diffraction: application totriclinic calcium pyrophosphatedihydrate. Biomedicine 1977; 27:242-4.

46 McCarty D J, Hogan J M, Gatter R A,Grossman M. Studies on pathologicalcalcifications in human cartilage. I.Prevalence and type of crystal depositsin the menisci in two hundred and

fifteen cadavera. J Bone Joint Surg[Am] 1966; 48: 309-25.

47 Lagier R. L'approcheanatomo-pathologique du concept dechondrocalcinose articulaire.Rheumatology 1981; 23: 421-37.

48 Bjelle A 0. Morphological study ofarticular cartilage in pyrophosphatearthropathy. Chondrocalcinosisarticularis or calcium pyrophosphatedihydrate crystal deposition disease.Ann Rheum Dis 1972; 31: 449-56.

49 Reginato A J, Schumacher H R,Martinez V A. The articular cartilage inarticular chondrocalcinosis. ArthritisRheum 1974; 17: 977-92.

50 Schumacher H R. Ultrastructuralfindings in chondrocalcinosis andpseudogout. Arthritis Rheum 1976; 19:413-25.

51 Bywaters E G L. Calciumpyrophosphate deposits in synovialmembrane. Ann Rheum Dis 1972; 31:219-1.

52 Bywaters E G L, Hamilton E B D,Williams R. The spine in idiopathichaemochromatosis. Ann Rheum Dis1971; 30: 453-65.

53 Dehais J, Senegas J, Bauduceau B,Bulier R, David-Chausse J. Stenose ducanal rachidien lombaire au cours d'unechondrocalcinose articulaire diffuse.Rev Rhum Mal Osteoartic 1977; 44:585-8.

54 Ellman M H, Vazquez T, Fergusson L,Mandel N. Calcium pyrophosphatedeposition in ligamentum flavum.Arthritis Rheum 1978; 21: 611-3.

55 Gerster J C, Lagier R, Boivin G.Achilles tendinitis associated withchondrocalcinosis. J Rheumatol 1980;7: 82-8.

56 Baud C A. Etude biophysique descalcifications intra-craniennes. J BelgeRadiol 1973; 56: 265-9.

57 Grahame R, Sutor D J, Mitchener M B.Crystal deposition inhyperparathyroidism. Ann Rheum Dis1971; 30: 597-604.

58 Bjelle A 0. Articular cartilage inhereditary pyrophosphate arthropathy.Morphological and biochemical studies.Rheumatology 1981; 33: 439-51.

59 Bjelle A 0, Sundstrom K G B. Anultrastructural study of the articularcartilage in calcium pyrophosphatedihydrate (CPPD) crystal depositiondisease (chondrocalcinosis articularis).Calc Tiss Res 1975; 19: 63-71.

60 Vignon E, Vignon G. Lachondrocalcinose articulaire diffuse.Lyon Medicine 1977; 238: 283-92.

61 Ducastelle Ch, Hemet J, Deshayes P.Introduction anatomopathologique al'etude de la chondrocalcinose. RevRhum Mal Osteoartic 1978; 45: 107-10.

62 Bjelle A 0. Cartilage matrix inhereditary pyrophosphate arthropathy.J Rheumatol 1981; 8: 959-64.

63 Botvin G, Lagier R, Baud D. Aspects

ultrastructuraux de la chondrocalcinosearticulaire. R6sultats preliminaires.Rheumatology 1981; 33: 127-30.

64 Stankovic A. Struktume histohemijskeibiohemiske promene u Zglobniumtkivima Nig, Yugoslavia: Faculty ofMedicine, 1981. Thesis (inSerbo-Croat).

65 Schumacher H R. The synovitis ofpseudogout: electron microscopicobservation. Arthritis Rheum 1968; 11:426-35.

66 Moskowitz R W, Harris B K, SchwartzA, et al. Chronic synovitis as amanifestation of calcium crystaldeposition disease. Arthritis Rheum1971; 14: 109-116.

67 Kariya M, Terayama K, Taguchi Y, et al.A study of crystal induced synovitis.Thirty three cases of calcification of themenisci of the knee joint. Journal oftheJapanese Orthopaedic Association1970; 44: 1099-113.

68 Dieppe P A, Huskisson E C, Crocker P,Willoughby D A. Apatite depositiondisease: a new arthropathy. Lancet1976; i: 266-8.

69 Halverson P B, McCarty D J.Identification of hydroxyapatite crystalsin synovial fluid. Arthritis Rheum 1979;22: 389-95.

70 Schumacher H R, Miller J L, LudovicoC, Jessar R A. Erosive arthritisassociated with crystal deposition.Arthritis Rheum 1981; 24: 31-7.

71 Fam A G, Pritzker K R H. Stein J L, etal. Apatite associated arthropathy: aclinical study of 14 cases and of twopatients with calcific bursitis. JRheumatol 1979; 6: 461-71.

72 Doherty M, Dieppe P A. Acutepseudogout: 'crystal sheding' or acutecrystallization? Arthritis Rheum 1981;24: 954-7.

73 Guiraudon C, Bontoux D, Sapporta L.Renseignements fournis par l'examinhistologique de la synoviale dans lesrhumatismes inflammatoireschroniques. Gazette Medicale de France1967; 74: 3461-70.

74 Taguchi Y, Kariya M, Terayama K. Anelectron microscopic study on thesynovial cells and the leucocytes in thesynovial fluid from pseudogoutsyndrome. Journal of the JapaneseOrthopaedic Association 1968; 42:1119-29.

75 Mandel N S. The structural basis ofcrystal-induced membranolysis.Arthritis Rheum 1976; 19: 439-45.

76 Kozin F, McCarty D J. Proteinadsorption to monosodium urate,calcium pyrophosphate dihydrate andsilica crystals. Relationship to thepathogenesis of crystal inducedinflammation. Arthritis Rheum 1976;19: 433-8.

77 Cheng H S, Halverson P B, McCartyD J. Release of collagenase, neutralprotease and prostaglandins from



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


cultured mammalian synovial cells byhydroxyapatite and calciumpyrophosphate dihydrate crystals.Arthritis Rheum 1981; 24: 1338-44.

78 Silcox D C, McCarty D J. Elevatedinorganic pyrophosphateconcentrations in synovial fluid inosteoarthritis and pseudogout. J LabClin Med 1974; 83: 518-31.

79 Russell R G G. Metabolism of inorganicpyrophosphate (PPi). Arthritis Rheum1976; 19, suppl: 465-78.

80 Howel D S, Muntz O, Pita J C, Enis J E.Extrusion of pyrophosphate into extracell media by osteoarthrotic cartilageincubates. J Clin Invest 1975; 56:1473-80.

81 Pritzker P H K, Cheng P T, Adams M E,Nyberg S C. Calcium pyrophosphatedihydrate crystal formation in modelhydrogels. J Rheumatol 1981; 5:469-73.

82 Pritzker P H K, Cheng P T, Omar S A,Nyberg S C. Calcium pyrophosphate

crystal formation in model hydrogels. II.Hyaline articular cartilage as gel. JRheumatol 1981; 8: 45 1-5.

83 Cheng PT, Pritzker P H K, Adams M E,Nyberg S C, Omar A S. Calciumpyrophosphate crystal formation inaqueous solutions. J Rheumatol 1980;7: 609-16.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Pathogenesis of chondrocalcinosis and pseudogout.Metabolism of inorganicpyrophosphate and production of calciumpyrophosphate dihydrate crystalsA. CASWELL, D. F. GUILLAND-CUMMING, P. R. HEARN, M. K. B. McGUIRE,R. G. G. RUSSELLFrom Department ofHuman Metabolism and Clinical Biochemistry, University of Sheffield Medical School, Beech Hill

Road, Sheffield S10 2RX.

Introduction General considerations

Calcium pyrophosphate dihydrate(CPPD) is one of several types ofcrystal that may be deposited in thebody. The term chondrocalcinosis isused to describe the radiographicappearance of calcified deposits,notably in the articular cartilage andmenisci of the knees but also in otherjoints.' Not all calcified deposits withinjoints are due to calciumpyrophosphate; however, occasionaldeposition of apatite,2 calciumdihydrogen phosphate,' and oxalatealso occurs.McCarty was the first to identify

CPPD crystals in synovial fluids takenfrom patients thought to have gouty

arthritis,"5 He found the crystals to bethe calcium pyrophosphate dihydratesalt in its triclinic form. Sheddingcrystals into the synovial space

produces acute or chronic attacks ofpseudogout.6

Certain considerations apply to allforms of crystal deposition, inparticular the factors determiningwhether or not crystallisation occurs.

This largely concerns the activityproducts of the ions involved, althoughthe presence of nucleating agents or

inhibitory agents, or both, must beconsidered. Solubilisation processesmay also be important.The purpose of this presentation is

twofold: (a) to review currentknowledge of the metabolism ofinorganic pyrophrosphate and (b) todescribe some of the mechanisms thatmay be involved in the production ofCPDD crystals, both in vitro and invivo.

Table 1 summarises the factors to beconsidered in relation to thedeposition of CPPD crystals and theassociation of chondrocalcinosis withendocrine and metabolic diseases. Thereasons for these clinical associationsare not always clear but they indicatethat a variety of metabolic disturbancesmay lead to deposition ofpyrophosphate crystals.CPPD deposition is also more

common in the presence of other jointdisease, including osteoarthropathy,5neuropathic (Charcot) joint disease,7and gout.5 In such cases the destructivechanges that occur may inducealterations of pyrophosphatemetabolism within the joint leading toincreased crystal formation.The increased incidence of CPPD

deposition with aging may also berelated to the changes that occur incartilage with age.

A familial form of chondrocalcinosiswas described by Zitnan and Sit'aj in1958," although at that time it wasnot known that CPPD was the calciumsalt responsible. Subsequently otherfamilial forms have beendescribed.'0"'- The existence of theseinherited forms allow a comparisonwith classic gout where specificenzyme defects have occasionally beenidentified--for example, in theLesch-Nyhan syndrome.'3 Theunderlying metabolic defect in thefamilial forms of chondrocalcinosis isstill not determined but may entailenzyme abnormalities.

Hypophosphatasia is an inheriteddisorder associated with

pseudogout." '" Here there is anenzyme defect, a deficiency of alkalinephosphatase with resultant rise ofinorganic pyrophosphate (PP,) in bodyfluids. Hypophosphatasia is the bestexample of how an abnormality inpyrophosphate metabolism maycontribute to the production ofchondrocalcinosis, but raises theintriguing question of why CPPDcrystals appear in the typical sites incartilage, even though PP,concentrations are raisedsystemically.

It is apparent that chondrocalcinosismust be considered a multifactoralproblem in which several metabolicand physiochemical factors probablyinteract to produce CPPD crystals.Before considering the mechanisms ofcrystallisation involved in productionof CPPD crystals, we will first reviewcurrent knowledge of pyrophosphatemetabolism.

Intracellular metabolism of inorganicpyrophosphate

GENERAL CONDITIONS

Inorganic pyrophosphate (PP,) isproduced at one or more steps in awide variety of biochemical pathwaysthat lead to the synthesis of most of themajor cell constituents. Hencegeneration of PPi occurs during thebiosynthesis of proteins, lipids,phospholipids, nucleotides, andnucleic acids, urea, steroids, structuralpolysaccharides, and glycogen.Breakdown of pyrophosphate isbrought about by a hydrolysis reactioncatalysed by inorganic



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Table 1 Conditions associated with deposition of crystals ofcakium pyrophosphate dihy-drate (CPPD) in joints (chondrocalcinosis). Note the large number ofpossible underlyingmechanisms.

Possible mechanisms involved

Inherited forms:Described from Possible abnormality in inorganic pyrophosphateCzechoslovakia, Chile, Netherlands, (PP,) metabolism-for example, over-Sweden, France, USA production of PP, or decreased degradation of

PP1 (possible changes in pyrophosphatases).Reduced inhibitors of crystallisation.

General associations:Aging

Other joint disease:OsteoarthritisNeuropathic joint disease

(Charcot joints)Destructive arthropathyOchronosisRheumatoid arthritisUrate gout

Metabolic disorders:Hyperparathyroidism

HypothyroidismHypophosphatasiaHaemochromatosis

Long-term steroid therapyPossible associations:

HypertensionRenal insufficiency

AcromegalyPaget's diseaseDiabetes mellitusWilson's disease

pyrophosphatases, in which 2 mol oforthophosphate (Pi) are produced permol of PP1 cleaved. Such enzymesinclude glucose- 6-phosphatase andalkaline phosphatase, as well as more

specific inorganic pyrophosphatases.The metabolic importance of PP1

has yet to be defined. The assumptionthat intracellular concentrations of PP1are very low led Kornberg and others'61 to point out that the removal of PP,provides a means of drivingpyrophosphorylase reactions in thedirection of synthesis, essentiallyrendering them irreversible. However,observations of detectable amounts ofPPi in rat liver, for example,'8 '9question this assumption, and,moreover, question whether the

Disturbed PP1 metabolism.Physical damage to chondrocytes.Release of nucleating agents or decreased

inhibitory activity (perhaps via action of releasedproteases to destroy inhibitors). Increasedcartilage permeation by Ca"+ and PP1.

Epitaxy on apatite crystals.pH may be lowered during inflammation whichpromotes crystal transformations to moreinsoluble forms.

Epitaxy on urate crystals

Raised extracellular Ca, and/or raised PP1 due toincreased adenylate cyclase activity.

Metabolic changes in cartilageRaised PP1 due to alkaline phosphatase deficiencyFe as nucleating agent or

pyrophosphatase inhibitorSecondary hyperparathyroidism

High PP, in chronic renal failure; acidosis promotescrystal transformation to insoluble forms

Metabolic alterations in cartilageAge related

Copper as crystal nucleating agent or

pyrophosphatase inhibitor

hydrolysis reaction for PP1 is atequilibrium in vivo. An alternativepossibility may be that inorganicpyrophosphatase is a nonequilibriumenzyme and that its activity is thelimiting factor in the removal of PP1.Additional determinants of theintracellular steady stateconcentration of PP1 may be theintracellular concentration of Pi, bothas a result of the effect of Pi on theequilibrium reaction, and because it isa competitive inhibitor ofpyrophosphatases. If thepyrophosphatase reaction is not inequilibrium, the concentration of PPiwill be determined by the balancebetween the rate of formation andbreakdown of PPi. In this case, the

concentration of intracellular PPiwould be under metabolic control andcould respond to changes in the ratesof either its synthesis or degradation.Furthermore, if the intracellularconcentration of PP1 does change inresponse to different metabolicconditions, it becomes possible for thision itself to be involved in theregulation of metabolism.

IS THE PPi HYDROLYSIS

REACTION IN EQUILIBRIUM?Attempts to determine theequilibrium constant and free energychange for hydrolysis of PP, understimulated physiological conditionshave produced variable results.20 22However, the reported tissueconcentrations of PPi appear to exceedthose calculated from values found forthe equilibrium constant,'8 19 23supporting the view that the hydrolysisreaction is not in equilibrium. Usingfreeze-clamped rat tissues we haverecently demonstrated that the totalPP, content of skeletal muscle (50nmol/g) is substantially higher thanthat found in liver, kidney, heart, andlung tissues (20 nmol/g) (unpublishedobservations). Similarly, with isolatedhuman cells, it has been observed thatfibroblasts and synovial cells containless PP1 than chondrocytes and bonecells. Furthermore, the higher PP1content of chondrocytes relative tofibroblasts appears to be correlatedwith differences in the rate ofproteoglycan synthesis in these twocell types.24.26

Several studies have shown thatboth the total PP1 content and thecalculated cytoplasmic free PP1content of freeze-clamped rat livermay change quite considerably aftervarious short term or long termmetabolic manipulations-forexample, after administration ofacetate or butyrate or after 48 hours'starvation."9 21 23 Such studies haveled to the suggestion that hepaticglucose uptake and phosphorylationare regulated predominantly bychanges in the concentration of freePPi in cytoplasm,2' thereby indicatinga role for PP1 in metabolic regulation.An increase in the intracellularconcentration of PP1 in response to thestimulation of a biosynthetic pathwayproducing PP1 has also been observedin isolated human articularchondrocytes after enhancement of



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Pathogenesis ofchondrocalcinosis and pseudogout Suppl p 29

glycosaminoglycan synthesis bytreatment with xylosides.2"

Raised intracellular concentrationsof PPi have also been reported inisolated fibroblasts and lymphoblastsderived from patients of a Frenchfamily with an hereditary form ofchondrocalcinosis.12 27 Thisobservation raises the question ofwhether some of the hereditary formsof chondrocalcinosis are associatedwith overproduction of PP, in cells,due to an abnormality in one of thebiosynthetic pathways generating PP,.Reports of abnormal cartilage matrixproduction in patients from twoSwedish families with a hereditaryform of this disease might support thisview.'2 28Hence evidence appears to indicate

that the intracellular concentration ofPP, does vary with such factors as celltype, metabolic conditions, and thepresence of disease states, supportingthe hypothesis that the hydrolysisreaction is not in equilibrium and thatthe concentration of this ion is undermetabolic control.

CONTROL OF THEINTRACELLULARCONCENTRATION OF PPiFactors involved in the control of theintracellular PP, concentration haveyet to be defined. As discussed above,PPi production will be controlled bythose factors which regulatebiosynthetic pathways containing PPigenerating steps and it will benecessary to delineate the relativecontribution of the different pathwaysto the overall rate of PP, production inthe cell. PP, breakdown will becontrolled by those factors whichregulate inorganic pyrophosphateactivity.

TISSUE DISTRIBUTION ANDPROPERTIES OF INORGANICPYROPHOSPHATASEA recent study of inorganicpyrophosphatase activity in the ratshowed that the total activity variesquite considerably between tissues andthat the total hepatic activity changesduring development.2' These resultsimply that the total tissue activity ofthis enzyme is regulated, but it shouldbe noted that the activities measuredin this study under optimal conditionsin vitro do not necessarily reflect thereal activities in vivo.

Inorganic pyrophosphatase appearsto be principally located in thecytosol.a 32 Smaller amounts of theactivity occur in mitochondria.33Inorganic pyrophosphatase activityhas also been demonstrated inendoplasmic reticulum in some tissues,but this may be due to the presence of aglucose-6-phosphatase,4 35 which alsopossesses pyrophosphatase activity.

Studies of pyrophosphatase in thecytosol from various tissues (red bloodcells, polymorphonuclear leukocytes,cartilage, dental pulp, etc.) havesuggested that this enzyme is specificfor PPi31 32 ...39 has a pH optimum of7-8, requires Mg++ for activity, isstrongly inhibited by other divalentmetal cations even in the presence ofMg++-for example, Ca++, Fe", andCu'' and is also inhibited by fluorideions. Several known inhibitors ofalkaline phosphatase-for example,Pi, imidazole, and CN- have littleinhibitory effect on this enzyme andthis, together with its specificity,suggest that this inorganicpyrophosphatase activity is a functionof a discrete enzyme and is not simplya function of alkaline phosphatase.Reports of the Km for PP1 of acytosolic inorganic pyrophosphataseare in the range 11-40 mmol, whichmay imply that the enzyme is notsaturated with substrate underphysiological conditions.Mitochondrial inorganic

pyrophosphatase activity occurs in twoforms, one of which is membranebound. Both forms resemble thecytosolic activity in that they arespecific for PPi, require Mg++ foractivity, have pH optimum of 7-8 andare inhibited by divalent cations-forexample, Ca++-and by fluoride.33 4 4

The effects of Mg++ and Ca++ may beof particular relevance and we are atpresent examining the influence of theavailability of these cations on theintracellular concentration of PP, inhuman articular chondrocytes andbone cells.

Extracellular metabolism of inorganicpyrophosphate

GENERAL CONSIDERATIONSAlthough PPi occurs extracellularly inbody fluids-for example, serum,plasma, urine, saliva, and synovialfluid, and large quantities also occuradsorbed to bone mineral, there is

little information about the movementof PPi across cell membranes or aboutwhich organs make appreciablecontributions to the PP, content ofextracellular fluid. There is evidencethat this PP, is of endogenous origin;thus it is not derived direct from thediet, as dietary PP1 andpolyphosphates appear to becompletely hydrolysed to Pi within theintestinal lumen, probably by theaction of alkaline phosphatase, whichis present in the brush bordermembranes of enterocytes.42

Studies of PP, turnover have beenrestricted to the examination ofextracellular PP,. Studies using32P-labelled PPi indicate that plasmaPP, turns over extremely rapidly indogs and in man.4" In dogs thehydrolysis to Pi accounts for at least25% of the loss of PP1 from the plasmacompartment whereas urinaryexcretion accounts for only 10%.43Hence hydrolysis to Pi would appear tobe a major mechanism for the removalof PP, from the plasma compartment.32P-labelled PP1 added to whole bloodin vitro, however, is only relativelyslowly hydrolysed, implying that themajor hydrolytic enzymes are notcirculating but are located on or withincells.43

ORIGIN OF EXTRACELLULAR PPiIN ARTICULAR CARTILAGEIf CPPD crystals are first formedoutside cells rather than inside, thenthe origin of extracellular PP, andmechanisms controlling the localconcentrations of this ion becomeimportant. (Fig. 1)The PP, present in the extracellular

space of articular cartilage could eitherarise from the intracellular compart-ment or it could be synthesisedextracellularly or released from sub-chondral bone. Release of PP, fromthe intracellular compartment couldoccur by a variety of mechanisms--forexample, by a specific membrane car-rier for PP,, in conjunction with thesecretion of matrix components or fol-lowing cell damage or death.

THE RELEASE OF PP1 FROMARTICULAR CARTILAGEGeneration of PPi in vitro by cartilagefragments was first reported byHowell's group,45 who stated that PPiwas released from growth platecartilage and from articular cartilage



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


INTRACELLULARSPACE

Posextr

Nucleosidetriphosphate

(NTfP) -~~~/BLoaynthetic reactionseg proteoglycan synthesis7

NUCLEOSIDE TRIPHOSPHATEP YROPHOSPHORYIASES

derived from young rabbits and frompatients with osteoarthritis but notfrom rabbit ear cartilage or fromarticular cartilage derived frommature rabbits or 'normal' humans.However, in a recent study, McCarty'sgroup,4" using a highly sensitive assayand correcting for PP1 hydrolysisduring the incubation by.the use of32P-labelled PPi, have observed release

EXTRACELLULARSPACE

cAMP

2P.

release ofproteoglycans

plus PP iin vesicles

leakage ?

cell damage ?

NUCLEOSIDE TRIPHOSPHATEPYROPHOSPHOHYDROLASE

(ectoenzyme)

+ PPI

of PPi from fragments of articularcartilage and fibrocartilage derivedfrom both young and adult rabbits andfrom fragments of 'normal' humanarticular cartilage. Extrusion of PP,may therefore be a feature common toall types and.age of cartilage.

It has been reported that, if washedmonolayers of human articularchondrocytes are incubated in

phosphate-buffered saline or mediumwithout serum, no detectable releaseof PP, from these cells occurs.Furthermore, when the concentrationof PPi in the medium is raised toaround 100 ,.tmol/l rapid hydrolysis ofPP; occurs.24 47 However, using a moresensitive assay, we have recently beenable to demonstrate the release ofsmall amounts of PP, from washedmonolayers of human articularchondrocytes incubated in mediumwithout serum. Additionally, we haveobserved that under our conditions, inwhich the extracellular PP,concentration is generally less than 1,tmol/l PP1 hydrolysis determined with32P-labelled PPi in the medium is veryslow (A. M. Caswell and others, p. 99).This difference between our resultsand those of others47 may imply thepresence of an extracellular inorganicpyrophosphatase activity with arelatively high Km for PP,.These various results suggest that

chondrocytes possess the ability torelease limited amounts of PP,. In thiscontext, it may be relevant that it hasnot been possible to show the passageof PPi across the membrane of the redcell.46 It is also of interest that, in thestudy of McCarty's group describedabove, release of PPi from rabbitcartilage fragments, but not fromhuman cartilage fragments, waspositively correlated with release ofuronic acid.46 However, this does notnecessarily establish that PP, anduronic acid are released from the celltogether since, as suggested earlier,changes in the rates of synthesis ofmatrix components could result inparallel changes in intracellular PP,.Such changes could in turn influencethe rate of release of PPi from the cellby any putative carrier mechanismsspecific for PPi. Release of PP, into theextracellular space of cartilage afterdamage or death of chondrocytes mayoccur and could explain why.patientswho have had meniscectomies have ahigher incidence of chondrocalcinosisin the operated knee than in theunoperated knee.49 The extent towhich this release mechanism occurs inundamaged cartilage or in other'normal' tissues is not known.

PRODUCTION OF PPiEXTRACELLULARLYAnother possible origin ofextracellular PP, is that it is generated

* PPi

PYROPHOSPRATASE(ectoenzyme)eg alkalinephosphatase

2PI

2Pi

Fig. 1 Pyrophosphate metabolism: a schematic representation ofpossiblesources of both intracellular and extracellular PPi. PPi introduced in theintracellular compartment may be either co-secreted with products ofintracellularbiosynthetic reactions-for example, proteoglycans-or may leak to theextracellular compartment when cells are damaged, as happens in degenerativejoint disease. Extracellular PP, may arise from the activity of the ectoenzyme,nucleoside triphosphate pyrophosphohydrolase, acting presumably on nucleosidetriphosphates 'leaked' to the extracellular compartment.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


outside the cell by membrane boundenzymes. One such enzyme is adenylatecyclase which catalyses the reaction:

ATP---3' ,5'-cyclic AMP + PP,.

This enzyme is located in cellmembranes and is activated inresponse to hormone and otheragents-for example, certain drugs.50The cyclic AMP formed is releasedinto the intracellular compartment butthe fate of PP, formed is unknown andit is not known whether any is releasedinto the extracellular fluid.Another enzyme of considerable

interest is nucleoside triphosphatepyrophosphohydrolase, whichcatalyses the reaction:NTP--NMP+ PP,, where NTP de-

notes nucleoside triphosphate andNMP nucleoside monophosphate.

In rat liver the enzyme is located inthe plasma membrane, is calciumdependent and hydrolyses bothpyrimidine and purine nucleosidetriphosphates with Kms in the Molarrange.5"-53 The enzyme has beenproposed to function either in calciumtransport52 or together with5'-nucleotidase in the salvage ofnucleoside triphosphates that leakfrom the cell.53 In rats this enzyme hasbeen shown to exhibit a wide tissuedistribution and the activity appears tovary quite considerably with tissuetype, the highest activities beingobserved in liver, small intestine, andkidney and much lower activities beingobserved in brain, thymus, andblood.52The presence of a nucleoside

triphosphate pyrophosphohydrolaseactivity in human articular cartilagehas been inferred from a recent studyof Howell's group, who demonstratedthe generation of PP, after adding 1mmol/l ATP to cartilagehomogenates.54 However, the use ofhomogenates precludes any statementabout the location of this enzymeactivity or the number of possibleactivities involved.We have been able to demonstrate

the generation of substantial amountsof PPi extracellularly after adding ATP(concentration range 6-25-400 Amol/1)to washed human articularchondrocyte monolayers incubated inmedium without serum (A. M. Caswelland others, p. 00). The effect can alsobe observed with other nucleosidetriphosphates-for example,

guanosine triphosphate, cytidinetriphosphate, and uridinetriphosphate. These preliminaryobservations suggest that presence of ahighly active nucleoside triphosphatepyrophosphohydrolase on the outsideof the plasma membrane of humanarticular chondrocytes with anapparently low Km for nucleosidetriphosphates. The function of thisactivity in chondrocytes is unclear butcould be similar to that suggestedearlier for nucleoside triphosphatepyrophosphohydrolase activities inother tissues-for example, a role incalcium transport.

It is therefore interesting that thetotal concentration of nucleotides inthe extracellular fluid of chickenepiphyseal cartilage has been reportedto range from 200 ,u mol/l in theproliferating zone to 740 molIl in thehypertrophic zone.55 Thus there is apotential mechanism for theextracellular generation of PPi, at leastin growth cartilage, since appreciableamounts of PP, could be generated ifnucleoside triphosphates comprisedeven a small fraction of the observedtotal nucleotide concentration.

It is also interesting to note that inthe recent study by Howell's groupdescribed above,54 nucleosidetriphosphate pyrophosphohydrolaseactivity was greater in articularcartilage homogenates derived frompatients with chondrocalcinosis than inthose derived from patients withosteoarthritis. One sample derivedfrom a 'normal' patient containedvirtually no activity. This apparentincrease in the activity in a situation inwhich increased amounts of PP1 occurextracellularly supports the view thatthis enzyme is important in thegeneration of PP, extracellularly inarticular cartilage. Furthermore, theremay be a role for this activity in thepathogenesis of chondrocalcinosisafter cartilage damage, since theleakage of metabolites from disruptedcells could result in the release ofnucleoside triphosphates, which couldthen serve as substrates for thegeneration of PP,.

RELEASE OF PPi FROMSUBCHONDRAL BONE

Local release of PP, from subchondralbone has been suggested as a furtherpossible source of extracellular PP, inarticular cartilage. Some polarity

would be necessary so that PP, couldreach subchondral bone through theblood supply on the metaphyseal sideand then subsequently be releaseddown a concentration gradienttowards the articular surface. Thiscould account for the observation thatcalcium pyrophosphate dihydratedeposits in the mid zone of articularcartilage, since this would be the pointat which PP1 from the subchondralbone and calcium from the joint spacewould come together.56

ALKALINE PHOSPHATASE ANDREMOVAL OF PPi FROM THEEXTRACELLULAR SPACEAs in the case of intracellular PP,metabolism, the major mechanism forremoving PPi from the extracellularcompartment is enzymatic hydrolysis.The addition of inorganicpyrophosphatase, to extracellularfluids--for example, plasma-results ina pronounced reduction in the PP1concentration and therefore thereaction:

PP 2Picannot be in equilibrium and thebreakdown of extracellular PP, mustbe limited by the activity of hydrolyticenzymes.The physiological importance of

alkaline phosphatase activity in theextracellular hydrolysis of PP; hasbeen demonstrated mainly by studiesof patients with hypophosphatasia inwhom a deficiency in alkalinephosphatase activity occurs."4 Thereis a fourfold increase in the plasma PP1concentration with a correspondingreduction in the rate constant ofhydrolysis in the removal of PP, fromthe extracellular compartment.5 4 5Such observations establish thatalkaline phosphatase does function asa pyrophosphatase in vivo and that isan ectoenzyme capable of hydrolysingsubstrates present in the extracellularcompartment. It can be calculated that-alkaline phosphatase may beresponsible for the removal of as muchas 80% of the PPi delivered to theextracellular compartment in normalindividuals.

PROPERTIES OF THE PPjHYDROLYTIC ACTIVITY OFALKALINE PHOSPHATA-SEThe properties of the hydrolyticactivity of alkaline phosphatasetowards PP1 has been studied in a



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


variety of tissues-for example, liver,kidney, and small intestine,56bone6' 62 and epiphyseal cartilage andmatrix vesicles.63-' This enzyme exhibitsremarkably similar properties in awide variety of tissues. The pHoptimum of the PP, hydrolytic activitydecreases with decreasing PP1concentration, so that at physiologicalconcentrations of PPi approaches pH7- 0. Magnesium ions affect the activityin a complex manner and this effectappears to be influenced by theavailability of substrate with anoptimal activity occurring at aMg+/ PPi concentration ratio of 1: 1.This led to the suggestion thalt the truesubstrate for the enzyme is MgP2O72-and that this cannot form if PP1 ispresent in excess over Mg++, whereas ifMg++ is present in excess, MgP2O7forms, which is inhibitory. In thepresence of Mg++, Ca++ is only slightlyinhibitory but the activity is subject toinhibition by substrate and product(Pi). The Km of alkaline phosphatasefor PPi inhibition has been reported torange from 40-85 molIl but some ofthese differences may be accounted forby the use of different concentrationratios of Mg'+ to PP1.

In a recent study, Howell's group54have extracted and examined theproperties of alkaline phosphatasederived from adult human articularcartilage. Virtually no activity could beobtained from 'normal' samples butappreciable amounts of the activitywere obtained from patients withchondrocalcinosis and even higheractivities were obtained from patientswith osteoarthritis.

In articular cartilage derived frompatients with osteoarthritis orchondrocalcinosis, there appear to betwo forms of alkaline phosphatase,both of which possess hydrolytic-activity towards PP, but whereas oneform of the enzyme resembles theactivity found in other tissues, theother appears to be activated ratherthan inhibited when Mg` is in excessover PPi. It is not known, however,whether both forms of the enzyme arefound in normal adult human articularcartilage.Taken together, these studies of the

hydrolytic activity of alkalinephosphatase towards PP, demonstratethat this activity can vary in responseto a variety of physiologicalagents-for example, Mg++, Ca,- and

Pi, and further studies are needed todetermine whether such agentsregulate the breakdown ofextracellular PP, in vivo.

IS A DEFECT IN CLEARANCE OF

PPi FROM ARTICULARCARTILAGE INVOLVEDIN THE PATHOGENESIS OFCHON DROCALCINOSIS?There is little direct evidence for adefect in PP, clearance from articularcartilage in patients with idiopathicchondrocalcinosis. There have beenreports that inorganic pyrophosphataseactivity due to both alkalinephosphatase and glucose-6-phosphatase is reduced in jointfluids from patients withchondrocalcinosis.66... In otherstudies, however, no change wasobserved either in the hydrolyticactivity of alkaline phosphatasetowards PP, or in an inorganicpyrophosphatase activity, with an acidpH optimum, in joint fluids from thesepatients.70 Moreover, as noted earlier,Howell's group observed an increaserather than decrease in the PPihydrolytic activity of alkalinephosphatase in articular cartilagederived from patients withchondrocalcinosis.54

However, some of the diseaseassociations observed suggest thatthere is a defect in PP, clearancefrom articular cartilage at least in somecases of chondrocalcinosis-forexample, in hypophosphatasia.7'Similarly, the associations betweenchondrocalcinosis and hyperpara-thyroidism,7274 haemochromatosis7'and hypomagnesaemia73 couldreflect the influence of Ca++, Fe++,and Mg++ respectively on the pyro-phosphatase activity of alkaline phos-phatase.

CLINICAL CONDITIONS IN WHICHDISORDERS OF EXTRACELLULARPP1 METABOLISM OCCURMeasurements of PP, in serum,plasma, or urine in a variety of clinicalconditions have suggested thatdisorders of extracellular PP,metabolism do occur in some diseasestates, hypophosphatasia being thebest example. However, plasma andserum PPi concentrations are alsoraised in about one third of patientswith chronic renal failure, and thevalues return to normal after

haemodialysis or renal transplant.76 'An increase in the PP, content of bonehas also been noted in some patientswith chronic renal failure and it hasbeen suggested that a relationshipexists between the bone content of PP,and the extent of soft tissuecalcification in these patients.79The plasma concentration of PP, has

been reported to be raised in somecases of acromegaly and this increaseappears to be correlated with anincrease in the plasma Piconcentration.' "' Plasma PP, is alsoraised in some patients withosteomalacia due to vitamin Ddeficiency, but apparently not in someof the inherited forms of vitaminD-resistant renal tubular rickets, norin osteomalacia associated with totalparenteral nutrition,'5 (M. K. B.McGuire and others, paper presentedat 14th Annual Meeting of AmericanSociety of Nephrology, 1981).A defect in urinary PP, excretion

may contribute to formation of renalstones in some cases. In several studiesit has been reported that urinary PP1excretion is reduced in men who formstones but not women, with this effectbeing most marked in the 30-40 agegroup."-83A defect in extracellular PPi

metabolism is unlikely in osteogenesisimperfecta as we have been unable toconfirm the observation of Solomon'sgroup that the serum PPiconcentration is raised in thisdisorder."5

In this context it is important to notethat measurements of PP, in serumgive values two to four-fold higherthan in plasma. This is due to therelease during blood clotting of PP,stored in the dense granules ofplatelets.

There is little evidence for anychange in the plasma PP,concentration in either rheumatoidarthritis or osteoarthritis.'7 8 86-86

In summary, there appears to be nosystemic disorder in PP1 metabolismassociated with most cases ofchondrocalcinosis.

Physicochemical studies of calciumpyrophosphate crystal formation

Little is known about thephysicochemical conditions necessaryfor the formation of calciumpyrophosphate dihydrate (CPPD)



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


crystals in articular cartilage and othersites in vivo. The problem can beconsidered from several points ofview:(a) What factors are necessary forinitiating crystal deposition? (b) Whatdetermines the interconversions ofdifferent crystal forms of CPPD? (c)What influences the formation of thecrystals, their solubility and removalfrom the joint?

INITIATION OF CRYSTALDEPOSITIONThe formation of CPPD crystals inpathological conditions may bepromoted in various ways. Firstly, theconcentrations of calcium or PPi maybe raised; the increasedconcentrations of PP, could result fromeither enhanced production ordecreased removal from the joint, as,for example, in hypophosphatasia." "712By analogy, increased plasma calciumconcentrations occur inhyperparathyroidism, althoughwhether or not the association withdeposition of CPPD crystals ismediated via raised calciumconcentrations is at present unknown.Crystal formation might also befavoured by the presence of nucleatingagents or by the removal of any neutralinhibitors of crystal formation,although currently very little is knownabout these possible mechanisms. Inorder to study some of these questionswe have devised a simple method todefine the conditions necessary forcrystal deposition in vitro.89 90The formation of crystals was

studied in simple synthetic solutionswhich mimic extracellular and synovialfluids. At physiological concentrationsof calcium (1- 5 mmol/) at pH 7- 4 andphysiological ionic strength, crystals ofcalcium pyrophosphate were found toform within three days at 370C, whenthe PPi concentration was 40 ,umol/l orhigher. In the presence of Mg++ atphysiological concentrations (0.5mmolIl), crystals formed only whenthe PP, concentration reached 175mol/l. This contrasts with theconcentrations of PP, found in normalsynovial fluids (mean 3 molIl; range 1to 4 mol) and in pseudogout fluid(mean 20 ,umolVl; range 5 to 60,umol/l). In the presence ofphysiological concentrations (1mmol/l) of inorganic phosphate(orthophosphate) the PPi

concentration required for crystalinitiation was lowered to 75 molIl. Thisreflects conditions which are theclosest to physiological that have so farbeen explored. It appears, therefore,that the formation of CPPD crystals insynovial fluid would not be favoured invivo. To explain where and whycrystals form in the body one thereforeneeds to invoke additionalmechanisms such as increased localconcentrations of either calcium orPP,, or the presence of nucleatingagents.The concept of nucleating

mechanisms leading to crystalformation is an attractive one whichrequires further study. The associationof pyrophosphate arthropathy withhaemochromatosis7" was investigatedand the possibility that iron salts mightact as nucleating agents for crystalformation was tested in the crystalgrowth system in vitro. The presenceof low concentrations of ferric salts(Fe++ at 25 umol/l) promoted crystalgrowth with the result that the amountof PPi needed for crystal formationwas reduced to one quarter of thatrequired in the absence of iron.89 It wasthought that this effect may have beendue to the colloidal nature of Fe-+ saltsin solution at neutral pH. Urate atphysiological concentrations also has asmall promoting effect on theprecipitation of calciumpyrophosphate'9 but it is doubtfulwhether this effect is potent enough toexplain the association found betweenurate gout and pyrophosphatearthropathy.

Crystal formation in vitro increasesrapidly as the pH rises through therange of 7 2 to 7-4 suggesting that invivo a small change in pH could inducecrystal formation without such largechanges in the concentration of Ca andPP, being required. These results areinteresting in relation to studies ofHowell and Pita's group,9'" who haverecorded high pH in extracellular fluidaspirates from epiphyseal cartilage bymicropuncture techniques. The pH offluid within articular rather thanepiphyseal cartilage is, however, notknown. In our studies of crystalformation in vitro a high pH promotedcrystal formation.The possible role of nucleation by

epitaxy has also been explored andhydroxyapatite crystals were tested fortheir effect in our crystal growth

experiments. At 40 4g/ml the addedcrystals raised the amount ofpyrophosphate required to initiateCPPD crystal growth, presumablyexplained by their ability to absorbpyrophosphate, thereby making itunavailable. At lower concentrations(4,g/ml) preliminary results suggestthat hydroxyapatite crystals arewithout effect, suggesting a balancebetween nucleating effects and surfaceadsorption of pyrophosphate. Furtherwork is required to clarify this point.

CRYSTAL TRANSFORMATIONSThe type of CPPD crystals foundnaturally in vivo are predominantly inthe triclinic form, although McCarty etal94 and Bywaters et al95 have reportedthe occurrence of the monoclinic form.Crystals produced in the three dayincubations under the experimentalconditions of Hearn and Russell' 90were shown by x-ray diffraction to beorthorhombic in the absence of Mg-+and amorphous in the presence ofphysiological concentration of Mg+.Longer incubations of one month ormore with added magnesium appear toallow the slow formation of crystals ofthe monoclinic type, followed by aslow transition to the triclinic form.The presence of 1 mmolI phosphateconsiderably increased the rate of thistransition, but even then failed toconvert a significant proportion ofcrystals to the triclinic form.The use of nucleating agents has so

far failed to influence the type ofcrystal formed from solution.Hydroxyapatite crystals promoteCPPD crystal formation but the typesformed are orthorhombic andmonoclinic. Similarly, addition ofpreformed CPPi crystals of these typesleads only to further growth of thosecrystals and new growth of similarcrystals, with no increase in transitionto the triclinic forms.These studies are based on

microscopical identification and awaitconfirmation by x-ray diffractiontechniques. Work from the soilchemists of the Tennessee RiverValley Authority,96 who investigatedcalcium pyrophosphates in relation totheir use as fertilisers, providesvaluable information about thepotential transformations that canoccur. A study of their results indicatesthat the monoclinic and triclinicvarieties of CPPD crystals appear to



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


represent the stable end products of anumber of potential crystaltransformations. Regardless of thephysical nature of the first crystals todeposit, all crystals may thereforeconvert ultimately to monoclinic andtriclinic varieties in vivo.These studies shed light on the

mystery ofwhy the naturally occurringcrystals are usually of the triclinic andoccasionally of the monoclinic variety.

In an interesting study, Pritzker etal97 showed that the monoclinic andtriclinic crystals could be made to formwithin silica and gelatin gels in vitro.Their conditions were close to, but didnot match, the natural physiologicalstate. Thus the concentrations ofcalcium and PPi they started with wereconsiderably in excess of normal andthe pH was below 6 0.Our own current work also suggests

that these crystal conversions actuallyoccur under simulated physiologicalconditions, but often take a long time,which means that the existence ofthese specific crystal forms in vivo maysimply be an indication of the length oftime available for crystal conversionsto take place.

CRYSTAL FORMATION,SOLUBILITY AND REMOVAL FROMTHE JOINTIn cartilage CPPD crystal depositsappear initially in rims around cells93and the large deposits are associatedwith empty cartilage lacunaeindicating chondrocyte death.9" In thesynovium, in contrast, CPPD crystalsare found only in the phagocytic cellsof the synovial membrane.99Experimental studies using labelledcrystals support the concept that these

.c..ystal deposits are sequestered fromt,he synovial fluid rather than beingformed in the synovium itself.

It seems probable, therefore, that!CPPD crystals are formed initially inarticular cartilage, and appear in thesynovial fluid via a process ofsheddingrather than growth in situ of newCPPD crystals.Bone is another tissue with a

possible role in formation of CPPD.Bone lies close to the joint and the PP,concentrations might be expected tobe high due to the ability of PP, toadsorb to hydroxyapatite surfaces. It isalso possible that nucleating agents arelocated at the sites where crystalgrowth starts. The preferential

location of CPPD deposits in articularcartilage and other joint structuressuggests that there may be someabnormality of PPi metabolism,metabolite diffusion, or of nucleationat these sites. PPi concentrations areraised in the synovial fluid but not inthe plasma of patients withchondrocalcinosis and osteoarthritis,with chronically symptomaticjoints.70"6 100 101 It is uncertain whetherthis reflects local abnormalities in PP,metabolism or whether it is a result ofdissolution of CPPD within the jointtissues.As crystals are not regularly found

in synovial fluid in osteoarthropathy,but only in pseudogout, despite thereported similarity in synovial fluidconcentrations of PP, levels, it wouldappear that either the PP, is derivedfrom different sources in the twoconditions, or that specificmechanisms (such as nucleation ofcrystal growth) exist in the joints ofpatients with chondrocalcinosis whichlead to CPPD crystal formation.The high negative fixed charge

density of articular cartilage102 mayfavour retention of cations.Maroudas103 has suggested thatcartilage may show some selectivity forthe retention of calcium, which wouldlead to a rise in total calciumconcentrations within the cartilage.The free ionic calcium concentrationwould, however, be expected toremain similar to that in synovial fluid.However, in pathological conditions,bound calcium might be releasedduring degradation of proteoglycansand thereby initiate the formation ofCPPD crystals.

Finally, it is possible that the raisedPPi concentrations arise fromdissolution of CPPD crystals.However, Camerlain et al104 found thatcrystals incubated in vitro with jointfluid showed very little exchange with321PPPi. This contrasts with the rapidturnover of PP, in vivo and suggeststhat either the solubility product forcalcium pyrophosphate had alreadybeen attained or that one or morecomponents of the turnover system invivo were lacking. The problems ofsolubility are complex, however, andsynthetic crystals behave differentlyfrom natural ones-possibly becauseof proteins which coat the crystals andimpair dissolution. The phagocytosisof CPPD crystals is probably

important in provoking the acuteinflammatory reaction within thejoint. CPPD, as with other crystaltypes which can provokeinflammation, has been shown toinduce cell lysis. This may be the basisfor the findings of chondrocytelacunae in the vicinity of CPPDcrystals in articular cartilage.99 Thismay be due to the intracellulardissolution of crystals releasing largeamounts of calcium and therebypoisoning the cell.The question of why only certain

patients with CPPD crystal depositiondisease experience attacks of crystalsynovitis remains to be determined.Presumably certain conditions have tobe met before crystals will be 'shed'from the cartilage to provoke anattack. There appears to be anassociation of such attacks withmajor surgical interevention,especially where there is a significantpostoperative decrease in serumcalcium concentrations (para-thyroidectomy, major surgery inthe abdomen or thorax, etc).101 Anyabrupt fall in serum calcium will lowerthe [Ca] x [PPi] ion product in theextracellular fluid bathing the CPPDcrystals and may result in a looseningof these deposits within the lacunaedue to crystal dissolution. The suddendecrease in size of the crystals mayenhance their release from sites wherethey were hitherto tightly packed.Other mechanisms for release includeminor surface fractures induced bytrauma or wear.

We are grateful to the Arthritis andRheumatism Council for their support ofthis work. A. Caswell acknowledges thereceipt of a research fellowship from theMedical Research Council.

References

1 McCarty D J. Clinical assessment ofarthritis. In: Hollander J L, ed.Arthritis and allied conditions. 9th ed.Philadelphia: Lea & Febiger, 1979:131-47.

2 Dieppe P A, Huskisson E C, CrockerP, Willoughby D A. Apatite depositiondisease. A new arthropathy. Lancet1976; i: 266-8.

3 Gaucher A, Faur6 G, Netter P, PourelJ, Duheille J. Identification descristaux observes dans lesarthropathies destructrices de lachondrocalcinose. Rev Rhum MalOsteoartic 1977; 44: 407-14.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


4 Hoffman G S, Schumacher H R, PaulH, et al. Calcium oxalatemicrocrystalline-associated arthritis inend-stage renal disease. Ann InternMed 1982; 97: 36-42.

5 Kohn N A, Hughes R E, McCarty D J,Faires J S. The significance of calciumpyrophosphate crystals in the synovialfluid of arthritic patients: thepseudogout syndrome.Ann Intern Med1962; 56: 738-45.

6 Bennett R M, Lehr J R, McCarty D J.Crystal shedding and acutepseudogout. Arthritis Rheum 1976; 19:93-7.

7 Jacobelli S, McCarty D J, Wilcox D C,et al. Calcium pyrophosphate dihydratecrystal deposition in neuropathicjoints. Four cases of polyarticularinvolvement. Ann Intern Med 1973;79: 340-7.

8 Zitnan D, Sitaj S. Natural course ofarticular chondrocalcinosis. ArthritisRheum 1976; 19: 363-90.

9 Zitnan D, Sitaj S. Mnohopecetnafamiliarlia Kalcifikaciz artikularynchchrupiek. Bratisl Lek Listy 1958; 28:217-24.

10 Van der Korst J K, Geerards J.Articular chondrocalcinosis in a DutchPedigree. Arthritis Rheum 1976; 19:405-9.

11 Bjelle A, Edvinsson U, Hagstam A.Pyrophosphate arthropathy in twoSwedish families. Arthritis Rheum1982; 25: 66-74.

12 Lust G, Faure G, Netter P, Gaucher A,Seegmiller J E. Evidence of a general-ised metabolic defect in patients withhereditary chondrocalcinosis. ArthritisRheum 1981; 24: 1517-21.

13 McKeran R 0, Andrews T M, HowellA, Gibbs D A, Chinn S, Watts R W E.The diagnosis of the carrier state forthe Lesch-Nyhan Syndrome. Q J Med1975; 44: 189-205.

14 O'Duffy J D. Hypophosphatasiaassociated with calcium pyrophos-phate dihydrate deposits in cartilage.Arthritis Rheum 1970; 13: 381-8.

15 Russell R G G, Bisaz S, Donath A,Morgan D B, Fleisch H. Inorganicpyrophosphate in plasma in normalpersons and in patients withhypophosphatasia, osteogenesisimperfecta and other disorders ofbone.J Clin Invest 1971; 50: 961-9.

16 Kornberg A. On the metabolic signifi-cance of phosphorylytic andpyrophosphorylytic reactions. In:Kasha M, Pullman D. eds. Horizons inbiochemistry. New York: AcademicPress 1962: 251-64.

17 Stetton D. Biosynthesis and pyrophos-phate. Am J Med 1960; 28: 867-70.

18 Cook G A, King M T, Veech R L.Changes in liver inorganic pyrophos-phate content during ethanol metabol-ism. Adv, Exp Med Biol 1980; 132:433-40.

19 Veech R L, Cook G A, King M T.Relationship of free cytoplasmicpyrophosphate to liver glucose contentand total pyrophosphate to cytop-lasmic phosphorylation potential.FEBS Lett 1980; 117, suppl: 65-72.

20 Flodgaard H, Fleron P. Ther-modynamic parameters for the hyd-rolysis of inorganic pyrophosphate atpH 7-4 as a function of [Mg2"], [K+]and ionic strength determined fromequilibrium studies of the reaction. JBiol Chem 1974; 249: 3465-74.

21 Lawson JW R, Guynn R W, Cornell N,Veech R L. A possible role forpyrophosphate in the control of hepa-tic glucose metabolism. In: HansonP W, Mehlman M A, eds.Gluconeogenesis: its regulation inmammalian cells. New York: Wiley,1976: 481-512.

22 Lawson J W R, Veech R L. Effects ofpH and free MG2 on the Keq of thecreatine kinase reaction and otherphosphate hydrolysis and phosphatetransfer reactions. J Biol Chem 1979;254: 6528-37.

23 Guynn R W, Veloso D, Lawson J W R,Veech R L. The concentration andcontrol of cytoplasmic free inorganicpyrophosphate in rat liver in vivo.Biochem J 1974; 140: 369-75.

24 Lust G, Nuki G, Seegmiller J E. Inor-ganic pyrophosphate and proteoglycanmetabolism in cultured human articu-lar chondrocytes and fibroblasts. Arth-ritis Rheum 1976; 19: 479-87.

25 McGuire M B, Colman C H, Baghat N,Russell R G G. Radiometric meas-urement of pyrophosphate in cell cul-tures. Biochem Soc Trans 1980; 8:529-30.

26 McGuire M K B, Bayliss M, Baghat N,Colman C H, Russell R G G.Pyrophosphate metabolism in humanarticular chondrocytes. Ann RheumDis 1981; 40: 207-8.

27 Lust G, Faure G, Netter P, SeegmillerJ E. Increased pyrophosphate in fib-roblasts and lymphoblasts frompatients with hereditary diffuse articu-lar chondrocalcinosis. Science 1981;214: 809-10.

28 Bjelle A. Cartilage matrix in heredit-ary pyrophosphate arthropathy. JRheumatol 1981; 8: 959-64.

29 Shatton J B, Shah H, Williams A, Mor-ris H P, Weinhouse S. Activities andproperties of inorganic pyrophosphat-ase in normal tissues and hepatictumours of the rat. Cancer Res 1981;41: 1866-72.

30 Felix R, Fleisch H. Purification andheterogeneity of inorganic pyrophos-phatase of pig scapula cartilage.Biochem J 1975; 147: 103-9.

31 Furuyama S, Mitsuma M, Doi N,Sugiya H, Oyama S. Properties anddistribution of inorganic pyrophos-

phatase in rabbit dental pulp. J DentRes 1977; 56: 1339-1244.

32 Raja K B, Smith G P, Peters T J. Sub-cellular localization and properties ofalkaline inorganic pyrophosphatase inhuman polymorphonuclear leuco-cytes. Clin Chim Acta 1981; 117:33-41.

33 Mansurova S E, Belyakova T N,Kulaev I S. The role of inorganicpyrophosphate in energy metabolismof isolated mitochondria. BiochemistrY(USSR) 1973; 38: 186-87.

34 Illingworth B, Cori C F. Glucose-6-phosphatase and pyrophosphataseactivities of homogenates of liver frompatients with glycogen storage disease.Biochem Biophys Res Commun 1965;19: 10-4.

35 Hefferan P M, Howell R R. Geneticevidence for the common identity ofglucose 6-phosphatase,pyrophosphatase-glucose phosphot-ransferase, carbamyl phosphate-glucose phosphotransferase and inor-ganic pyrophosphatase. Biochim Bio-phys Acta 1977; 496: 431-5.

36 McCarty D J, Pepe P F, Solomon S D,Cobb J. Inhibition of human erythro-cyte pyrophosphatase activity by cal-cium, cupric and ferrous ions. ArthritisRheum 1970; 13: 336.

37 Felix R, Fleisch H. Properties of inor-ganic pyrophosphatase of pig scapulacartilage. Biochem J 1975; 147: 111-8.

38 Thuillier L. Purification and kineticproperties of human erythrocyteMg2"-dependent inorganic pyrophos-phatase. Biochim Biophys Acta 1978;524: 198-206.

39 Oyama S, Mitsuma M, Furuyama S.Separation of inorganic pyrophos-phatase from alkaline phosphatase inrabbit dental pulp. IntJ Biochem 1981;13: 105-7.

40 Egremovich N V, Volk E S, BaikovA A, Shakhov Y A. Isolation andprimary characterization of two formsof inorganic pyrophosphatase frombovine heart mitochondria. Biochemis-try (USSR) 1980; 45: 785-91.

41 Volk S E, Baykov A A, Duzkenko V S,Avaeva S M. Kinelic studies on theinteractions of two forms of inorganicpyrophosphatase of heart mitochon-dria with physiological ligands. Eur JBiochem 1982; 125: 215-20.

42 Russell R G G, Wadstrom L B, Lind-stedt S, Care A D, Bisaz S, Fleisch H.The origin or inorganic pyrophosphatein urine. Clinical Science 1969; 37:419-29.

43 Jung A, Russell R G G. Bisaz S, Mor-gan D B, Fleisch H. The fate ofintravenously in jectedpyrophosphate-32P in dogs. Amtt JPhvsiol 1970; 218: 1757-63.

44 Jung A. Preston C J, Walton R J,McGuire M K B. Russell R G G. 32p



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


kinetics in man utilizing 32P-orthophosphate and 32P-pyrophosphate. Mineral and Electro-lyte Metabolism 1979; 2: 236.

45 Howell D S, Muniz D, Pita J C, EnisJ E. Pyrophosphate release by osteo-arthritic cartilage incubates. ArthritisRheum 1976; 19: 488-94.

46 Ryan L M, Cheung H S, McCarty D J.Release of pyrophosphate by normalmammalian articular hyaline and fib-rocartilage in organ culture. ArthritisRheum 1981; 24: 1522-7.

47 Nuki G, Pritchard M H, HendersonW J, Lust G. Articular cartilagemineralization and inorganicpyrophosphate metabolism in chon-drocytes. European Journal ofRheumatology and Inflammation1978; 1: 105-14.

48 Felix R, Fleisch H. The effect ofpyrophosphate and diphosphonates oncalcium transport in red cells.Experientia 1977; 33: 1003-5.

49 Doherty M, Watt 1, Dieppe PA. Local-ised chondrocalcinosis in postmenis-cectomy knees. Lancet 1982; i:1207-10.

50 Robison G, Butcher R W, SutherlandE W. Cyclic AMP. New York:Academic Press, 1971.

51 Flodgaard H, Torp-Pedersen C A. Acalcium ion-dependent adenosinetriphosphate pyrophosphohydrolase inplasma membrane from rat liver.Biochem J 1978; 171: 817-20.

52 Torp-Pedersen C, Flodgaard H, Saer-mark T. Studies on a Ca`' dependentnucleoside triphosphate pyrophos-phohydrolase in rat liver plasma mem-branes. Biochim Biophys Acta 1979;571: 94-104.

53 Tran T T, Phillips J W, Schulze-Specking A, Raseneck J, Decker K.Properties and biosynthetic connec-tion of the nucleotide pyrophosphataseof rat liver plasma membrane andendoplasmic reticulum. Hoppe SeylersZ Physiol Chem 1981; 362: 305-16.

54 Tenenbaum J, Mainz 0, SchumacherH R, Good A E, Howell D S. Compari-son of phosphohydrolase activitiesfrom articular cartilage in calciumpyrophosphate deposition disease andprimary osteoarthritis. ArthritisRheum 1981; 24: 492-500.

55 Wuthier R E. Electrolytes of isolatedepiphyseal chondrocytes, matrix vesi-cles and extracellular fluid. Calcif Tis-sue Res 1977; 23: 125-33.

56 McGuire M K B, Heam P R, RussellR G G. Calcium pyrophosphate cryst-als (their relevance to calciumpyrophosphate dihydrate crystaldeposition disease, pseudogout, chon-drocalcinosis and pyrophosphate arth-ropathy): biochemical andphysicochemical aspects. In: MaroudasA, Holborow M, eds. Studies in joint

disease. Tunbridge Wells; PitmanMedical, 1980: 117-56.

57 Sorensen S A, Flodgaard H, SorensenE. Serum alkaline phosphatase, serumpyrophosphatase, phos-phorylethanolamine and inorganicpyrophosphate in plasma and urine-agenetic and clinical study of hypophos-phatasia. Monogr Hum Genet 1978;19: 66-9.

58 Eaton R H, Moss D W. Inhibition ofthe orthophosphatase and pyrophos-phatase activities of human alkalinephosphatase preparations. Biochem J1967; 102: 917-21.

59 Moss D W, Eaton R H, Smith J K,Whitby L G. Association of inorganicpyrophosphatase activity with humanalkaline phosphatase preparations.Biochem J 1967; 102: 53-7.

60 Butterworth P J. The pyrophosphataseactivity of pig kidney alkalinephosphatase and its inhibition bymagnesium ions and excess ofpyrophosphate. Biochem J 1968; 110:671-5.

61 Eaton R H, Moss D W. Partialpurification and some properties ofhuman bone alkaline phosphatase.Enzymologia 1968; 35: 31-9.

62 Felix R, Fleisch R. Thepyrophosphatase and Ca2+ - Mg'+[-ATPase activity] of purified calfbone alkaline phosphatase. BiochimBiophys Acta 1974; 350: 84-94.

63 Majeska R J, Wuthier R E. Studies onmatrix vesicles isolated from chickepiphyseal cartilage. Association ofpyrophosphatase and ATPaseactivities with alkaline phosphatase.Biochim Biophys Acta 1975; 391:51-60.

64 Felix R, Fleisch H. Pyrophosphataseand ATPase of isolated cartilagematrix vesicles. Calcified TissueResearch 1976; 22: 1-7.

65 Fortuna R, Anderson H C, Carty R P,Sajdera S W. Enzymaticcharacterization of the matrix vesiclealkaline phosphatase isolated frombovine epiphyseal cartilage. CalcifTissue Int 1980; 30: 217-25.

66 Russell R G G, Bisaz S, Fleisch H,etal.Inorganic pyrophosphate in plasma,urine and synovial fluid of patientswith pyrophosphate arthropathy,chondrocalcinosis or pseudogout.Lancet 1970; i: 899-902.

67 Good A E, Starkweather W H.Synovial fluid pyrophosphatephosphohydrolase in pseudogout, goutand rheumatoid arthritis. ArthritisRheum 1969; 12: 298.

68 Yaron M, Zurkowski P, Weiser H I,Yust 1, Goldsmied A, Hermann E.Pseudogout with low levels of alkalinephosphatase in the synovial fluid.Ann Intern Med 1970; 73: 751-6.

69 McCarty D J, Solomon S D, WarnockM L, Paloyan E. Inorganic

pyrophosphate concentrations in thesynovial fluid of arthritic patients. JLab Clin Med 1971; 78: 216-9.

70 Jacobelli S, Kettlun A M, Sapag-HagarM. Inorganic pyrophosphatase activityof the synovial fluid. Arthritis Rheum1978; 21: 447-52.

71 Hamilton E B D. Disease associatedwith CPPD deposition disease.Arthritis Rheum 1976; 19: 353-7.

72 Eade A W T, Swannell A J,Williamson N. Pyrophosphatearthropathy in hypophosphatasia. AnnRheum Dis 1981; 40: 164-70.

73 McCarty D J, Silcox D C, Coe F, et al.Diseases associated with calciumpyrophosphate dihydrate crystaldeposition. Am J Med 1974; 56:704-14.

74 Alexander G J M, Scott D G I, DieppeP A. Disease associations andlaboratory investigations inpyrophosphate arthropathy. AnnRheum Dis 1980; 40: 515-6.

75 Milazzo S C, Ahem M J, ClelandL G, Henderson D R F.Calcium pyrophosphate dihydratedeposition disease and familialhypomagnesaemia.JRheumatol 1981;8: 767-71.

76 David D S, Sakai S, Granda J, et al.Role of pyrophosphate in renalosteodystrophy. Trans Soc ArtifInternOrgans 1973; 19: 440-5.

77 Silcox D C, McCarty D J.Quantification of inorganicpyrophosphate in biologic fluids;hyperpyrophosphataemia in somepatients with osteoarthritis,pseudogout, acromegaly and uraemia.Arthritis Rheum 1973; 16: 132.

78 Bishop M, Russell R G G, Preston C J,Bisaz S, Fleisch H. Pyrophosphate andthe development of bone disease inpatients with renal insufficiency. In:Avioli L, Bordier P H, Fleisch H,Massry S, Slatopolsky E, eds.Phosphate metabolism in kidney andbone Paris: Nouvelle ImprimerieFournie, 1975: 157-69.

79 Alfrey A C, Solomons C C, Circillo J,Miller N. Bone pyrophosphate inuraemia and its associations withextraosseous calcification.J Clin Invest1976; 57: 700-5.

80 Camerlain M, Silcox D C, LawrenceA M, McCarty D J. Variation inplasma and urinary inorganicphosphate and pyrophosphate innormal subjects and in patients withacromegaly or osteoarthritis. JRheumatol 1980; 7: 365-74.

81 Russell R G G, Hodgkinson A. Theurinary excretion of inorganicpyrophosphate by normal subjects andpatients with renal calculus. ClinicalScience 1966; 31: 51-62.

82 Bauman J M, Bisaz S, Felix R, FleischH, Ganz U, Russell R G G. The role ofinhibitors and other factors in the



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


pathogenesis of recurrentcalcium-containing renal stones. ClinSci Mol Med 1977; 53: 141-8.

83 Wikstrom B, Danielson B G,Ljunghall S, McGuire M K B, RussellR G G. Urinary pyrophosphateexcretion in renal stone formers withnormal and impaired renalacidification. Scand J Urol Nephrol1981; 61, suppl: 1-1 7.

84 Armstrong D, Van Wormer D, Sol-omons C C. Increased inorganic serumpyrophosphate in serum and urine ofpatients with osteogenesis imperfecta.Clin Chem 1975; 21: 104-8.

85 Russell R G G, McGuire M K B,Hearn P R, Preston C J, Douglas D L,Jung A. Pyrophosphate revisited. Dis-orders of metabolism. Calcified TissueInternational 1979; 27: A41.

86 Altman R D, Muniz 0 E, Pita J C,Howell D S. Articular chondrocal-cinosis. Microanalysis of pyrophos-phate in synovial fluid and plasma.Arthritis Rheum 1973; 16: 171-8.

87 Ryan L M, Kozin F, McCarty D J.Quantification of human plasma inor-ganic pyrophosphate. I. Normal valuesin osteoarthritis and calciumpyrophosphate deposition disease.Arthritis Rheum 1979; 22: 886-91.

88 Micheli A, Po J, Fallet G H. Measure-ment of soluble pyrophosphate inplasma and synovial fluid of patientswith various rheumatic diseases. ScandJ Rheumatol 1981: 10: 237-40.

89 Heam P R, Russell R G G, Elliott J C.Formation product of calcium

pyrophosphate crystals in vitro and theeffect of iron salts. Clin Sci Mol Med1978; 54: 29.

90 Hearn P R, Russell R G G. Formationof calcium pyrophosphate crystals invitro. Implications for calciumpyrophosphate crystal deposition dis-ease (pseudogout). Ann Rheum Dis1980; 38: 222-7.

91 Cuervo L A, Pita J C, Howell P S. Ultra(microanalysis) of pH, pCO2 and car-bonic anhydrase at calcifying sites incartilage. Calcif Tissue Res 1971; 7:220-31.

92 Howell D S, Pita J C. Calcification ofgrowth plate cartilage with specialreference to studies on micropuncturefluids. Clin Orthop 1976; 118: 208-29.

93 Pita J C, Howell D S. In: Sokologg L,ed. The joints and synovial fluid. Vol.1. New York: Academic Press, 1978.

94 McCarty D J, Kohn N N, Faires J S.The significance of calcium pyrophos-phate crystals in the synovial fluid ofarthritic patients: the pseudogout syn-drome. 1. Clinical aspects. Ann InternMed 1962; 56: 711-37.

95 Bywaters E G L, Dykes E, Pirie C,Rueben R. Crystal deposits in thediscs, ligaments, joints and bursae ofthe spine. In: Proceedings of sym-posium on studies in joint disease. Lon-don: London Hospital Medical School,1978.

96 Lehr J R, Brown E H. ChemicalEngineering Bulletin. No. 6, TennesseeValley Authority, USA, 1967: 22-3.

97 Pritzker K P H, Cheng P T, AdamsM E, Nyburg J C. Calcium pyrophos-phate dihydrate crystal formation inmodel hydrogels.J Rheumatol 1978; 5:469-73.

98 McCarty D J. In: Holt P J L, ed. Cur-rent topics in connective tissue disease.London: Churchill Livingstone, 1975:181-97.

99 McCarty D J, Palmer P W, HalversonP B. Clearance of calcium pyrophos-phate dihydrate crystals in vivo. Arth-ritis Rheum 1979; 22: 718-27.

100 Grindey G B, Nichol C A. Microp-rocedure for determination ofpyrophosphate and orthophosphate.Anal Biochem 1970; 33: 114-9.

101 Silcox D C, McCarty D J. Elevatedinorganic pyrophosphate concentra-tions in synovial fluids in osteoarthritisand pseudogout.J Lab Clin Med 1974;83: 518-31.

102 Maroudas A. Transport of solutesthrough cartilage: permeability tolarge molecules. J Anat 1976; 122:335-47.

103 Maroudas A. Physicochemical proper-ties of cartilage in the light of ionexchange theory. Biophysical J 1968;8: 575.

104 Camerlain M, McCarty D J, SilcoxD C, Jung A. Inorganic pyrophosphatepool size and tumover rate in arthriticjoints.J Clin Invest 1975; 55: 1373-81.

105 McCarty D J. Calcium pyrophosphatedeposition disease. Arthritis Rheum1957; 19: 275-85.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Pyrophosphate arthropathy-recent clinical advancesMICHAEL DOHERTY

From the University Department of Medicine, Bristol Royal Infirmary

It is now 20 years since McCarty et al.identified calcium pyrophosphatedihydrate (CPPD) crystals in synovialfluid from patients with acute arthritisand radiological chondrocalcinosis.'CPPD crystals were subsequentlyshown to be inflam-matory both in vivo and in vitro2and their causal role in joint disease,reflected by the term 'CPPDcrystal-induced synovitis', was readilyaccepted. Cadaveric studies3 laterestablished CPPD as the mostcommon, though not exclusive, causeof cartilage calcification in the knee,and familial chondrocalcinosisarticularis,5 metabolic disease,6 andaging7 became recognised aspredisposing factors.

Since description of'the pseudogoutsyndrome', however, an increasinglycomplex picture of pyrophosphatearthropathy has emerged. In particularit is apparent that: (a) there is greatdiversity in the clinical syndromesassociated with intra-articular CPPDcrystal deposition, and (b) CPPDcrystal deposition, particularly in theelderly, commonly occurs in theabsence of inflammation and jointdamage.

Both observations are incorporatedin McCarty's clinical classification of'calcium pyrophosphate depositiondisease', in which five overlappingpatterns are recognised-'pseudo-gout', 'pseudorheumatoid arthritis','pseudo-osteoarthritis', 'pseudo-neuropathic joints', and asymptomaticor 'lanthanic' deposition.' Inherent inthis classification, however, is theunexplained paradox that in someindividuals CPPD crystals appear tocause arthritis, as suggested by thecharacteristic nature and distributionof joint disease and the phlogisticproperties of the crystals, while inothers they deposit inertly in theapparent absence of joint disease. Theresultant debate as to whether CPPD(and hydroxyapatite) crystals areprimary pathogenetic particles, epi-

phenomena to cartilage damage, or'innocent bystanders' has beenreviewed recently,' and is consideredelsewhere in this issue. It wouldappear, however, that in most casesintra-articular CPPD crystal deposi-tion alone is an insufficient cause forarthritis. This, together with the widerange of clinical expression, impliesthe operation of multiple factors in thepathogenesis of pyrophosphatearthropathy-a point that will recurthroughout this discussion.This review will largely be confined

to recent clinical studies that havecontributed to our understanding ofthe arthritis associated with CPPDcrystal deposition. The predisposingfactors, clinical range, and possiblemethods of treatment ofpyrophosphate arthropathy will beconsidered first, and then the questionof asymptomatic chondrocalcinosiswill specifically be addressed bycritical examination of a recentlyproposed 'amplification loop'hypothesis for particle-induced jointdisease."o

Predisposing factors

HEREDITYStudies to determine an hereditarytrait in pyrophosphate arthropathy aremade particularly difficult by theincreasing mobility of the populationand by late onset of disease expression,which precludes examination ofseveral generations at the age of risk.Recognition and subsequent study offamilial cases has therefore beenprompted largely by presentation at anearly age with florid polyarticulardisease. The paucity and specialcharacteristics of such cases' 11-17 hasled, not unnaturally, to the assumptionthat hereditary predisposition topyrophosphate arthropathy is rare.

Recently, however, Rodriguez-Valverde et al. have described fiveSpanish pedigrees that differ fromother reported kindreds in showing

late onset of symptoms, predominancein women, mild clinical disease, andoligoarticular chondrocalcinosis, par-ticularly of the knee (Table 1)." Suchcases, similar to heterozygous forms inthe Czech5 and Chilean series,1' areclinically and radiologically indistin-guishable from common sporadiccases of pyrophosphate arthropathy,suggesting that the true prevalence offamilial disease has been under-estimated owing to lack of adequatestudies and that a genetic influencemay operate even in sporadic forms ofthe disease. The early finding byMcCarty that three out of 12 patientswith pseudogout had affected familymembers' supports this hypothesis andemphasises the need for furtherstudies.The mechanism of familial

predisposition remains unknown. Ametabolic disturbance of cartilagematrix has been suggested byexamination of cartilage from Swedishpatients." Recent demonstration,however, of raised concentrations ofinorganic pyrophosphate (PP,) in skinfibroblasts and transformedlymphocytes of French hereditarycases20 raises the possibility of ageneralised metabolic abnormality.

METABOLIC DISEASEAlthough an impressive number ofmetabolic diseases have been reportedto predispose to CPPD deposition,only a few are likely to have a trueassociation (Table 2). Many reflect nomore than chance concurrence of twocommon age-related phenomena, ascontrolled studies have shown forPaget's disease,2' diabetes,' 22 andhypertension.22 With rare conditions,however, convincing evidence may beprovided by the finding of prematurechondrocalcinosis in a few cases;CPPD deposition in four youngpatients with Bartter's syndrome23 24

therefore appears significant andpresumably relates to the associatedhypomagnesaemia. The validity of



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Pyrophosphate arthropathy-recent clinical advances Suppl p 39

a

o

0.

-'( q

as

rA

U

mla

co

e

0.

X

4 <Q

_, t;,, <

as

C

0-.2E

0 0

_a

-_ 3E

0. <

r~) q

_3 t;3<

auX

- U_=-A n

a._3

8.00 ,

CL~~

.'_8 .i 3 <

Table 2 Metabolic conditionsassociated with CPPD crystaldeposition disease

HypothyroidismHyperparathyroidismHaemochromatosisHaemosiderosisGoutHypophosphatasia*Hypomagnesaemia*Wilson's disease*Amyloid(Ochronosis) (Haemophilia)* Particularly associated withchondrocalcinosis at a young age.

ccE0

laasE

11oz

putative metabolic associations hasbeen fully reviewed by Hamilton6 and,more recently, by Howell.25Even in the presence of associated

metabolic diseases the multifactorialnature of pyrophosphate arthropathyis still apparent, as shown by theindependent effect of age in patientswith hyperparathyroidism26 andhaemochromatosis.6 Whether suchconditions affect CPPD crystaldeposition direct, or whether, byalteration of connective tissuemetabolism they bring aboutpremature phenotypic expression ofan underlying heterozygous statecan only evoke speculation. Inboth hyperparathyroidism andhaemochromatosis, however,correction of the underlying metabolicabnormality does not ameliorate jointdisease," 2 suggesting that onceestablished pyrophosphatearthropathy is a self perpetuatingcondition.

The relationship between gout andchondrocalcinosis has recently beeninvestigated in two companion studiesby Stockman and Hollingworth.11 29Chondrocalcinosis was detectedradiologically in 6% of 138 goutypatients (mean age 55 years) but inonly 1 % of age-matched controls withhyperuricaemia and in noage-matched controls with normalconcentration of uric acid, confirmingthat the association is with goutyarthritis and not with hyperuricaemiaper se. Interestingly, gouty patientswith chondrocalcinosis had longerduration of symptoms and higherscores for radiological osteoarthritis ofthe knee, implying that the primarycorrelation is with joint damage.

Since the initial report by Kaplinskiet al. ' several studies have supportedan association between chondro-calcinosis and senile amyloid. Ryan etal. for example, found CPPD depositsin four out of five patients withamyloid arthropathy of the wrists, not-ing carpal tunnel syndrome and pittinghand oedema as prominent features.3" 32Teglbjaerg et al. also found amyloid inassociation with CPPD in 14 out of 15tissues (seven hips, eight knees)removed during joint replacement forpyrophosphate arthropathy.33 Aspeculative mechanism to link amyloidwith chondrocalcinosis is suggested bythe enhancement of fibroblast glyco-saminoglycan synthesis by amyloidfibrils,34 the high Ca++ concentration inamyloid material,35 and the sequestra-tion by amyloid of pyrophosphateanalogues.36 Such an association isthus theoretically attractive and if con-firmed would explain in part the re-lationship between CPPD depositionand aging.

HYPERMOBI LITY

Generalised joint laxity andchondrocalcinosis were first associatedby Bird et al., who found that of 16patients with generalisedhypermobility but no superaddedarthropathy, 11 showed widespreadchanges of osteoarthritis and four hadevidence of CPPD deposition.3" Theirconclusion that hypermobilitypredisposes to both osteoarthritis andchondrocalcinosis is open to twopossible interpretations. Firstly, asystemic defect of connective tissuemight predispose independently tohypermobility and chondrocalcinosis,or, secondly, joint laxity itself mayresult in chondrocalcinosis via localmechanical factors. Support for thelatter hypothesis was recentlyprovided by four case reports fromSettas et al. of localisedchondrocalcinosis in unstable joints.38This finding emphasises that CPPDdeposition may occur secondary tolocal joint damage (perhaps themechanism in neuropathic joints') andthat some patients diagnosed as havingpyrophosphate arthropathy may wellhave a different primary joint disease.

OSTEOARTH RITISIn most series osteoarthritis is the mostcommon association found, theincidence ranging from 40% to 70%Y.OWhether this is greater than the

. 0= t

og ,

"5

*iq

U

-u:s

.F

k

k

Q

c)

"5"5

-u*_

Ic



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


incidence found in a controlpopulation, however, has not beendetermined.Ellman et al. found greater joint

space narrowing on x-ray films ofnon-weight-bearing knees of patientswith pseudogout than on those oftheir spouses or hospitalisedcontrols, concluding that thisrepresented co-existence of CPPDdeposition and osteoarthritis.39Whether this osteoarthritis wasprimary or secondary, however, wasnot discernible.

In a recent study by Dieppe et al.several patients showed clear clinicaland radiological progression fromosteoarthritis to a more widespreaddisease associated with pseudogout ordestructive arthritis,"0 stronglysuggesting that osteoarthritis is agenuine predisposing factor. Therelationship between trauma,chondrocalcinosis and CPPDdeposition further confirms thisview.'-"

AGINGMany studies have now confirmed thesharp increase in incidence ofradiological chondrocalcinosis andCPPD deposition in those over 60years, up to 30%-60% ofnonogenarians being affected.743Whether this results from a higherincidence of degenerative changes inthe elderly or from aging per se,however, remains unknown.

Ethnic origin has been shown toexert an additional influence on agerelated incidence of chondro-calcinosis, as shown for example, bythe higher incidence in elderly Jewishcompared with non-Jewish, Tunisians(12% v 4%) (M Moalla et al., paperpresented at 15th International Con-gress of Rheumatology, Paris, 1981).Wilkins et al. have recently reported ahigh prevalence (34%) of radiologicalchondrocalcinosis in a geriatriccaucasian population, rising from 11 %in those aged 65-74 years through35% in those aged 75-84 to 47% inthose over 85.4 Interestingly, defor-mity and radiological osteoarthritiswere more common in those withchondrocalcinosis. Similar clinicalfindings have been observed by Ell-man and Levin for chondrocalcinosisof the wrist, suggesting that age-associated chondrocalcinosis is notwholly benign.43

Clinical range of disease

SITESIn all published series clinicallysignificant disease associated withCPPD deposition most commonlyaffects knees, ankles, shoulders,wrists, and metacarpophalangealjoints; involvement of the spine, hip,and elbow is less frequent butparticularly occurs in hereditaryforms.5 11 Recent attention, however,has particularly focused on CPPDdeposition at additional or unusualsites.Three cases of temporomandibular

joint involvement have been reported,one with presumed acute pseudogoutresponsive to indomethacin45 and twowith destructive lesions associatedwith solitary masses of CPPD inchondroid metaplasia of synovium.46Similar chronic, solitary masses of'tophaceous' CPPD have beenreported to cause destructivemonoarthritis, unassociated withCPPD elsewhere, in small joints of thehand,40 49 thus broadening thepathological setting of CPPDdeposition.Spinal involvement has been

increasingly recognised. Isolatedannulus fibrosus CPPD deposits havebeen found in 36 of 1000 discsremoved during initial operation,50and in 10% of 73 tissues removedduring second and subsequentoperations for 'disc' disease.5" Ellmanet al. reported four similar cases ofCPPD deposition in lumbar discfibrocartilage removed from sites ofprevious operation, emphasising therole of previous trauma.52Bywaters recently described CPPD

deposits in interspinal bursae in thecervical spine,53 and Le Goffet al. havedescribed four elderly patients whodeveloped a self-limiting meningiticsyndrome in association with cervicalcartilage calcification and CPPDdeposition elsewhere.54 Axial CPPDdeposition and associated clinicalsyndromes may therefore be morefrequent than previously supposed.

Calcification of tendons,particularly of the Achilles,quadriceps, and triceps, are a commonradiological finding in patients witharticular chondrocalcinosis. Gerster etal. have shown that CPPD may beresponsible and have emphasised theradiological differences between fine

linear deposits of CPPD and round,nummular collections of apatite.5Most patients remain asymptomatic,but Gerster et al. recently implicatedCPPD tendon deposits in one patientwith olecranon bursitis56 and threewith Achilles tendonitis,57 concludingthat tendonitis and bursitis should beconsidered as extra-articularmanifestations of CPPD depositiondisease. Interestingly, one of thepatients with Achilles tendonitisshowed chronic inflammation anddisappearance of tendoncalcification .

PATTERNSThere have been relatively few clinicalsurveys of patients withpyrophosphate arthropathy, and mosthave included only small numbers ofpatients selected according to theclassification introduced by McCarty.'Recently, however, Dieppe et al.reported their findings on 105consecutive patients presenting withjoint disease and evidence of CPPDdeposition.10 In contrast to previoussurveys, inclusion of patients with asubsequent diagnosis of anotherrheumatic disease allowed study of thewhole clinical range of CPPDdeposition.A striking feature of this study, not

emphasised by previous authors, wasthe pronounced difference betweenthe sexes in the pattern of disease. The29 men were 10 years younger, had ashorter history of symptoms, mainlysuffered recurrent acute attacks in theleg, and had less severe joint damage.By contrast, most of the 76 women hadchronic polyarticular disease withmore frequent attacks in the joints ofthe arms and tendency to polyarticularattacks. Gross destructive changes,similar to those described by Richardsand Hamilton,58 Menkes et al.,59 andothers, were seen particularly in kneesand shoulders of the elderly womenwith generalised disease.

Although all features of the clinicalsubtypes described by McCarty wererecognised in this study, considerableoverlap made clear classification onthis basis difficult, and the prefix'pseudo' was often clearlyinappropriate. A total of 45 patients,for example, had definite generalisedosteoarthritis, and another eightfulfilled ARA criteria for rheumatoiddisease. Other recognised joint disease



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


included generalised hypermobility,'3previous knee surgery,8 and gout.2 Inmany cases it was evident that thesediseases predated CPPD deposition,and the finding that 69% of patientshad pre-existing disease or treatmentknown to damage articular cartilageled the authors to conclude that crystaldeposition commonly occurssecondary to joint damage-anhypothesis examined in greater detailat the end of this review.

Treatment

Spilberg et al. recently reported theefficacy (80%) of 1 mg intravenouscolchicine in patients with acutepseudogout, emphasising particularlythe pathogenetic and diagnostic impli-cations of this finding.' In practice,however, the use of colchicine is rarelywarranted as acute attacks are selflimiting, rarely severe, and commonlyrespond to aspiration alone. Non-steroidal, anti-inflammatory drugs,particularly for polyarticular attacks,and intra-articular steroid25 affordadditional benefit in difficult cases.

In contrast, chronic pyrophosphatearthropathy often presents difficultiesin management. In contrast to gout,there is no definitive treatment for thisdisabling condition, and attempts atsymptomatic control with analgesic oranti-inflammatory drugs are oftendisappointing. Apart from jointlavage" little in the way of additionalmedical treatment has been proposed.

Recently, however, Doherty andDieppe have reported a beneficialeffect of intra-articular yttrium-90 ascolloidal silicate, on chronicpyrophosphate arthropathy of theknee.62 Fifteen patients with severebilateral disease were given 5mCi 90Yplus steroid in one knee, and salineplus steroid in the other, on a random,double blind basis. At six months therewas significantly less pain, stiffness,tenderness, and effusion in the kneesinjected with 90Y: significantdifferences in range of movement andjoint circumference were also noted,partly due to progression of the diseaseon the control side. In all cases bethpatient and observer favoured thetreatment side (p<0-01), reinforcingthe conclusion that 9'Y is effective inmoderating the synovitis of chronicpyrophosphate arthropathy. Thetendency of this condition to affect the

knees of the elderly makes suchtreatment highly suitable, as the jointlends itself readily to injection and theprocedure carries negligibletheoretical risks in this age group.

Doherty and Dieppe in this issuealso report a double blind, placebocontrolled trial of oral magnesiumcarbonate in chronic pyrophosphatearthropathy. The finding of a uniformtrend towards improvement over sixmonths in those taking magnesium isencouraging and suggests that furtherstudies of magnesium supple-mentation are warranted. Recentdemonstration by McCarty et al. of ahalf life of one to three months forCPPD crystals in arthritic humanjoints63 indicates that only in a pro-longed trial might any putative definitivetreatment be expected to result in adiminution of radiological chondro-calcinosis, a finding reported in asingle patient by Runeberg et al. 4

The 'amplification loop' hypothesis

The clinical study that led Dieppe etal." to challenge McCarty's clinicalclassification of CPPD depositiondisease' also suggested an hypothesisto explain both the clinical diversity ofpyrophosphate arthropathy and theparadox of asymptomaticchondrocalcinosis. Figure 1 outlinesthis 'amplification loop'hypothesis," 42 which applies to othercalcium phosphate crystals.Many factors may initiate cartilage

damage, and when such damage issevere symptomatic arthritis willresult. Cartilage damage may alterproteoglycan concentrations, altercrystal nucleating and inhibitoryfactors, increase PPi turnover, or

Ose0arbrtHyperfmrbility CartilogeTrournaiurgery damnoge ArtlritisSleod t at <n

~~~~~moge~ ~

Wlommatlory ( , Hrdt

e I N~~~~~..icdisease

RLk- otcrystals (atragrrensof bore aod cartiloge)

Fig. 1 Amplification loop'hypothesis. Pre-existing joint damagepredisposes to crystal deposition. Themineral deposits may then acceleratejoint damage by mechanical orinflammatory pathways.

possibly effect some other change thatpredisposes susceptible individuals toCPPD (or apatite) crystal formation.Aging and associated metabolicdisease may independently enhancethis susceptibility. Crystals grow incartilage2 and if they remain theretheir deleterious effects are probablylimited. But if the crystals are shedfrom altered cartilage into the synovialspace'5 they will then be exposed tosynoviocytes and other inflammatorymediators and may act as wearparticles on the joint surface.' Theymay then produce acute inflammatoryepisodes and further chronic jointdestruction, setting up anamplification loop with furtherpredisposition to crystal formation.The finding of chondrocalcinosis inasymptomatic joints in the elderlycould then be explained by assumingthat when age-associated crystaldeposition occurs in a well preservedcartilage matrix, shedding of crystalsinto the joint space is prevented. Thewide variety of 'pseudo' syndromesassociated with CPPD crystaldeposition could also be explained bysuperimposition of crystalinflammation on other diseasepatterns, the most common beingprimary or secondary osteoarthritis.Furthermore, progression ofestablished pyrophosphatearthropathy would be expected tocontinue, despite correction ofinitiating metabolic disease.2" 27

Such an hypothesis, by accom-modating many of the clinical observa-tions reviewed, therefore appearstheoretically attractive. Its integrity,however, rests on three crucial suppos-itions. Firstly, the interrelationshipbetween cartilage damage and crystaldeposition; secondly, the existence ofcrystal shedding as an in vivo mechan-ism; and, thirdly, the major role ofsynovitis in the pathogenesis of pyro-phosphate arthropathy. Recent clini-cal studies have made it possible to testeach of these separately.

CARTILAGE DAMAGE AND

CRYSTAL DEPOSITIONThis interrelationship has been inves-tigated by Doherty et al. using menis-cectomy as a convenient human modelof isolated joint damage.42 Onehundred patients who underwentunilateral meniscectomy werereviewed at a mean of 24- 8 years after



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


operation. Radiological chondro-calcinosis was detected in 20'Y, ofoperated, but only 44>, of unoperated,knees (p 0-01), chondrocalcinosisbeing confined to the operated side in16 cases. Prevalence of chondro-calcinosis and osteoarthritis in kneesof age- and sex-matched controls wassimilar to that in unoperated knees,implying no special underlying pre-disposition to chondrocalcinosis orosteoarthritis in the study group. Theincreased prevalence of chondrocal-cinosis in operated knees thereforerelated to previous joint trauma, afinding previously reported by Lindenand Nilssen.4' More importantly, how-ever. operated knees with chondro-calcinosis had significantly higherprevalence of inflammatory features(stiffness. effusion, acute attacks) andimore severe radiological changes ofosteoarthritis than operated kneeswithout chondrocalcinosis, thus sup-porting the amplification of joint dis-ease by crystals arising in the context ofdamaged cartilage. The independenteffect of age on localisedchondrocalcinosis in this study (Fig. 2)again illustrates the multifactorialnature of this phenomenoni.

40

*UnoperatedIa upercuesr

Percentcge

A

Li 1Uperated30 r Controls |00

20~

10. [uge (years) 50 50-65 65

n= 18 18 62 63 20 19

Fig. 2 P'rev-alenice of'choacidrocalciriosis according to age itt

operated and unoperated knees of 100patients who had undergotnernetnisectorn and 100 controls (ref42).

CRY SI AL SHEDD)INGCPrecipitation of acute pseudogout byjoint lavage with crystal solubilisingagents prompted Bennett et al. topropose crystal shedding as a possiblepathogenetic mechanism."' Althoughseveral observations, including a

lowered ion ised Cat+ in situations that

provoke acute attacks, might be

explained bv this 'auto-injectioti.evidence for its operation in an

appropriate clinical setting hasremained circumstantial.

Recently, however, Doherty andDieppe reported disappearance ofradiological chondrocalcinosis duringan attack of pseudogout in the knee ofa 76 year old man.6" Thisphenomenon, analogous to dispersalof apatite calcification in acute calcificperiarthritis, provides direct evidenceof CPPD crystal shedding in vivo, andmay explain those cases of acutepseudogout with synovial fluid crystalsbut no chondrocalcinosis of theaffected joint. Another similar case ina wrist has been reported, in whichchondrocalcinosis was undetectableonx-ray film two months after an acuteattack complicated by reflexsympathetic dystrophy of the hand.57

THE SYNOVIAL COMPONENT

Much evidence from in vitro andanimal studies incriminates synovialreaction in CPPD crvstal-inducedarthritis.2

Clinical evidence of the importanceof continuing synovitis in chronicpyrophosphate arthropathy, however,was recently provided by Doherty andDieppe, who demonstrated objectiveand symptomatic improvementfollowing synoviorthese with 90y.62Although radiocolloids generallyproduce best results when radiologicalchanges are slight, several patients hadhad advanced x-ray changes.Improvement even in these suggeststhat synovium-derived, and not justmechanical, factors are important atall stages of this disease; interactionbetween these factors, however, iscomplex.9

Clinical studies to test theamplification loop hypothesis havetherefore tended to strengthen ratherthan to refute the various suppositionsinherent in this mechanism. Even clearprogression in hereditary cases fromasymptomatic chondrocalcinosis ofnormal thickness cartilage to severearthritis may be viewed as a rareinstance5 where the gross nature of theCPPD deposits results in anappreciable mechanical disruption ofpossibly already compromisedcartilage.19

Several observations, however, stillremain unexplainied, the mostimportant being the distribution ofjoint involvement in patients withpyrophosphate arthropathy. Though

the prevalence anid patternl of diseaseof metacarpophalanlgeal andinterphalatngeal joints is similar to thatfound in patients wvith generalisedosteoarthritis.l) rist, shoLldei . andankle involvement. uncommon inosteoarthritis, is frcquently seeni. Onlyin certain cases, thetefore. canamplificationi of existinig joint diseasebe clearly establishet, indicating thatfurther initiating and predlisposingfactors have vet to be dletertinined.

I would like tc) thanik the Arthritis anidRheumatisnm Counlcil for firnanicial support.

References

1 McCai is D J, Kohn N N. Fitrcs J S. thesignificance of calciulm phosphatecrystals in the svnov al fluidi otf arthritispatients: the 'pseudogOLut syndrome'. I.Clinical aspects. Ain l tterni Ved 1962:56: 711-37.

2 Schumacher H R. Pathogerncsis ofcrystal-induced syrovitis (linics itnRheurnatic Diseases 1977; 3: 105-31.

3 NMcCarty I) J, Hogan J Ml Gatter R A.Grossnman NI. Studies on pathologicalcalcifications in hunman cartilage. 1. Pre-valence and types of crystal deposits in)the menisci of 215 cadavers. J BotneJoitit Surg (Am) 1966, 48: 309-25.

4 Lagier R, Baud C A. Pathologic calcifi-cations of the locomotor systeni. Posi-tion of articular chondrocalcinosis. In:Milhaud G, Owen M, Black-wood H J J, eds. Rapports et Communi-cations au IV Symposium European desTissues Calcifies Bordeaux, 196 7. Paris:Sedes. 1968: 109-13

5 Zitnani D, Sitaj S. Natural couirse oferticular chonidrocalcinosis .4rtiritisRheum 1976: 19, suppl: 36 ,9()0.

6 Hamilton E B D. Diseases associatedwith CPPD deposition disease ArthrilnsRheum 1976; 19, suppl: 353-7.

7 Ellnmat M H. Browni M I. Leviin B.Prevalence of knee chondrocalcinosis inhospital and cliic patients aged 50 orolder. J Amti Geriltr Soc 1 981: 29:1 89-92.

8 McCartv D J. Calcium pyrophosphatedihyrate crystal deposition disease.Arthritis Rheumti 1976, 19, suppl:275-86.

9 Dieppe P A. Doherty M. The role ofparticles in the pathogenesis of jointdisease. In: Berry C [ ed . Currenit topicsinl pathology. Botne and(/ Joinit Disease'.Berlin: Springer- Verlag, 1982:199- 33.

10 Dieppe P A. Alexander G M, Jones H,Doherty M. Scott D G. Manhire A,Watt 1. Pyrophosphate arthropathy: a

clinical and radiological study of 105cases.AnniRheumnDis 1982.41: 371-6.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


11 Reginato A J. Articularchondrocalcinosis in the ChiloeIslanders. Arthritis Rheum 1976; 19:395-404.

12 Van der Korst J K, Geerards J. Articularchondrocalcinosis in a Dutch Pedigree.Arthritis Rheum 1976; 19, suppl: 405-9.

13 Gaucher A, Pourel J, Faure G, Netter P,Peterschmitt J, Cromer R. Leschondrocalcinosis articulaires diffuseshereditaires. Rev Rhum Mal Osteoartic1977; 44: 589-97.

14 Asshoff H, Bohm P, Schoen E,Schurholz K. Hereditarechondrocalcinosis articularis.Untersuchung einer familie.Humangenetik 1966; 3: 98-103.

15 Moskowitz R W, Katz D.Chondrocalcinosis (pseudogoutsyndrome). A family study. JAMA1964; 188: 867-71.

16 Richardson B, Zulman J, Chafetz N,Genana H. Hereditary articularchondrocalcinosis in aMexican-American family. ArthritisRheum 1980; 23: 736.

17 Bjelle A, Edvinsson U, Hagstam A.Pyrophosphate arthropathy in twoSwedish families. Arthritis Rheum 1982;25: 66-74.

18 Rodriguez-Valverde V, rTinture T,Zuniiga M, Penia J, Gonzalez A. Familialchondrocalcinosis. Prevalence inNorthern Spain and clinical features in 5pedigrees. Arthritis Rheum 1980; 23:471-8.

19 Bjelle A. Cartilage matrix in hereditarypyrophosphate arthropathy. JRheumatol 1981; 8: 959-64.

20 Lust G, Faure G, Netter P, Gaucher A,Seegmiller J E. Evidence of a generalizedmetabolic defect in patients withhereditary chondrocalcinosis: increasedinorganic pyrophosphate in culturedfibroblasts and lymphoblasts. ArthritisRheum 1981; 24: 1517-21.

21 Boussina 1, Gerster J C, Epiney J, FalletG H. A study of the incidence ofarticular chondrocalcinosis in Paget'sdisease of bone. Scand J Rheumatol1976; 5: 33-5.

22 Alexander G M, Dieppe P A, DohertyM, Scott D G. Pyrophosphatearthropathy: a study of metabolicassociations and laboratory parameters.Ann Rheum Dis 1982; 41: 377-81.

23 Goulon M, Raphael J C, deRohan P.Syndrome de Bartter etchondrocalcinose. Nouv Presse Med1980; 9: 1291-5.

24 Bauer F M, Glasson P H, Vallotton M B,Courvoisier B. Syndrome de Bartter,chondrocalcinose et hypomagnesemie.Schweiz Med Wochenschr 1979; 109:1251-6.

25 Howell D S. Diseases due to thedeposition of calcium pyrophosphateand hydroxyapatite. In: Kelly W N,Harris E D, Ruddy S, Sledge C B, eds.

Textbook of rheumatology.Philadelphia: W B Saunders, 1981.

26 Pritchard M H, Jessop J D.Chondrocalcinosis in primaryhyperparathyroidism. Influence of age,metabolic bone disease, andparathyroidectomy. Ann Rheum Dis1977; 36: 146-51.

27 Hamilton E B D, Bomford A B,Laws J W, Williams R. The natural his-tory of arthritis in idiopathic haemo-chromatosis: progression of the clinicaland radiological features over ten years.Q J Med 1981; 199: 321-9.

28 Stockman A, Darlington L G, Scott J T.Frequency of chondrocalcinosis of theknees and a vascular necrosis of thefemoral head in gout: a controlled study.Ann Rheum Dis 19 80; 39: 7-1 1.

29 Hollingworth P, Williams P L, Scott J T.Frequency of chondrocalcinosis of theknees in asymptomatic hyperuricaemiaand rheumatoid arthritis: a controlledstudy.Ann Rheum Dis 1982; 41: 344-6.

30 Kaplinski N, Biran D, Frankl 0.Pseudogout and amyloidosis. Harefauh1976; 91: 59.

31 Ryan L M, Berhard G C, Liang G,Kozin F. Amyloid arthropathy in theabsence of dysproteinemia: a possibleassociation with chondrocalcinosis[Abstract]. Arthritis Rheum 1978; 21:588.

32 Ryan L M, Liang G, Kozin F. Amyloidarthropathy: possible association withchondrocalcinosis. J Rheumatol 1982;9: 273-8.

33 Teglbjaerg P S, Ladefoged C, SorensenK H, Christensen H E. Local articularamyloid deposition in pyrophosphatearthritis. Acta Pathol Microbiol Scand1979; A87: 307-11.

34 Palmoski M, Brandt K. Stimulation ofglycosaminoglycan biosynthesis byamyloid fibrils. Biochem J 1975; 148:145-7.

35 Yood R A, Skinner M, Cohen A S, LeeV W. Soft tissue uptake of bone seekingradionuclide in amyloidosis. JRheumatol 1981; 8: 760-6.

36 Kula R, Engel W, Line B. Scanning forsoft tissue amyloid. Lancet 1977; i:92-3.

37 Bird H A, Tribe C R, Bacon P A. Jointhypermobility leading to osteoarthritisand chondrocalcinosis. Ann Rheum Dis1978; 37: 203-11.

38 Settas L, Doherty M, Dieppe P A.Localised chondrocalcinosis in unstablejoints. Br Med J 1982; 285: 175-6.

39 Ellman M H, Brown N L, Levin B.Narrowing of knee joint space inpatients with pseudogout. Ann RheumDis 1981; 40: 34-6.

40 Altman R D. Arthroscopic findings ofthe knee in patients with pseudogout.Arthritis Rheum 1976; 19, suppl:286-92.

41 Linden B, Nilssen B E. Chondro-calcinosis following osteochondritis dis-

secans in the femur condyles. ClinOrthop 1978; 130: 223-7.

42 Doherty M, Watt I, Dieppe P A.Localised chondrocalcinosis inpost-meniscectomy knees. Lancet 1982;i: 1207-10.

43 Ellman M H, Levin B. Chondro-calcinosis in elderly persons. ArthritisRheum 1975; 18: 43-7.

44 Wilkins E, Dieppe P A, Maddison P.Evison G. Articular chondrocalcinosisand its association with osteoarthritis inthe elderly. Ann Rheum Dis 1982; 40:516.

45 Good A E, Upton L G. Acutetemperomandibular arthritis in a patientwith bruxism and calciumpyrophosphate deposition disease.Arthritis Rheum 1982; 25: 353-5.

46 de Vos R A, Brants J, Kusen G J,Becker A 1. Calcium pyrophosphatedihydrate arthropathy of thetemporomandibular joint. Oral Surg1981; 51: 497-502.

47 Pritzker K P H, Phillips H, Luk S C,Koven I H, Kiss A, Houpt J B.Pseudotumor of temporomandibularjoint: destructive calciumpyrophosphate dihydrate arthropathy.JRheumatol 1976; 3: 70-81.

48 Leisen J C C, Austed E D, Bluhm G B,Sigler J W. The tophus in calciumpyrophosphate deposition disease.JAMA 1980; 244: 1711-2.

49 Ling D, Murphy W A, Kriakos M.Tophaceous pseudogout. AJR 1982;138: 162-5.

50 Lagier R, Wildi E. Frequence de lachondrocalcinose dans une serie de 100disques intervertebraux exciseschirurgicalement. Rev Rhum MalOsteoartic 1979; 46: 303-7.

51 Andres T L, Trainer T D. Intervertebralchondrocalcinosis-a coincidental find-ing possibly related to previous surgery.Arch Pathol Lab Med 1980; 104:269-71.

52 Ellman M H, Vazquez L T, Brown N L,Mandel N. Calcium pyrophosphatedihydrate deposition in lumbar discfibrocartilage. J Rheumatol 1981; 8:955-8.

53 Bywater E G L. Rheumatoid and otherdiseases of the cervical interspinous bur-sae, and changes in the spinous pro-cesses. Ann Rheum Dis 1982; 41:360-70.

54 LeGof P, Pennec Y, Youinou P. Signescervicaux aigues pseudo-meningesrevelateurs de la chondrocalcinosearticulaire. Sem Hop Paris 1980; 56:1515-8.

55 Gerster J C, Baud C A, Lagier R,Boussina I, Fallet G H. Tendoncalcifications in chondrocalcinosis. Aclinical, radiologic, histologic, andcrystallographic study. Arthritis Rheum1977; 20: 717-22.

56 Gerster J C, Lagier R, Boivin G.Olecranon bursitis related to calcium



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


pyrophosphate dihydrate crystaldeposition disease. Clinical andpathologic study. Arthritis Rheum 1982;25: 989-96.

57 Gerster J C, Lagier R, Boivin G.Achilles tendinintis associated withchondrocalcinosis. J Rheumatol 1980;7: 82-8.

58 Richards A J, Hamilton E B D.Destructive arthropathy inchondrocalcinosis articularis. AnnRheum Dis 1974; 33: 196-203.

59 Menkes C J, Simon F, Delrieu F, ForestM, Delbarre F. Destructive arthropathyin chondrocalcinosis articularis.Arthritis Rheum 1976; 19, suppl:329-48.

60 Spilberg I, McLain D, Simchowitz L,Berney S. Colchicine and pseudogout.Arthritis Rheum 1980; 23: 1062-3.

61 Bird H A, Ring E F J. Therapeutic valueof arthroscopy. Ann Rheum Dis 1978;37: 78-9.

62 Doherty M, Dieppe P A. Effect ofintra-articular Yttrium-90 on chronicpyrophosphate arthropathy of the knee.Lancet 1981; ii: 1243-6.

63 McCarty D J, Palmer D W, Halver-son P B. Clearance of calcium pyro-phosphate dihydrate crystals in vivo. I.Studies using '"Yb labelled tricliniccrystals. Arthritis Rheum 1979; 22:718-27.

64 Runeberg L, Colian Y, Jokinen E J,Lahdevirta J, Aro A. Hypomagnesemiadue to renal disease of unknownetiology. Am J Med 1975; 59:873-81.

65 Bennett R M, Lehr J R, McCarty D J.

Crystal shedding and acute pseudogout.An hypothesis based on a therapeuticfailure. Ann Rheum Dis 1976; 19:93-7.

66 Doherty M, Dieppe P A. Acutepseudogout: 'crystal shedding' or acutecrystailisation? Arthritis Rheum 1981;24: 954-7.

67 Fam A G, Stein J S. Disappearance ofchondrocalcinosis following reflexsympathetic dystrophy syndrome.Arthritis Rheum 1981; 24: 747-9.

6 8 Bensasson M, Dorfmann H,Parez-Busquier M. Etuderadiographique de la main dans 50 casde chondrocalcinose articulaireprimative. Comparison avec une seriede 100 temoins. Rev Rhum MalOsteoartic 1 975; 42: 3-11.

Conclusion

Recent studies of patients withpyrophosphate arthropathy continueto emphasise the multifactorial aetiol-ogy of CPPD crystal deposition andthe wide range of clinical disease pat-terns. It is conceivable, therefore, that'pyrophosphate arthropathy' repre-sents recognition more of a mechan-

ism of joint damage than of a specificdisease state, analogous to the situa-tion with 'Milwaukee shoulder.'Deposition of crystals may thus berelevant to our understanding of a muchwider range of joint diseases than wasfirst thought when the term 'crystaldeposition arthropathies' was intro-

duced. Further elucidation of themetabolic and tissue factors involvedin CPPD crystal deposition and of thechronic interaction between CPPDcrystals and joint tissues is required, asit is only through knowledge of thesethat suitable treatment of 'pyrosphos-phate arthropathy' will be attained.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Formation of calcium phosphate crystals in normal andosteoarthritic cartilageS. Y. ALI AND S. GRIFFITHS

From the Institute of Orthopaedics, Royal National Orthopaedic Hospital, Stanmore, Middlesex

Introduction

The mechanism of cartilagecalcification has been outlined inrecent reviews."' A study ofepiphyseal cartilage calcification hasclearly established that the firstcrystals of apatite that are found in thehypertrophic zone are found in matrixvesicles, in the longitudinal septa.' 2Matrix vesicles are extracellular,membranous particles, about 100 nmin diameter, that are associated withcalcification of numerous tissues.3Over 80% of the alkaline phosphataseactivity of epiphyseal cartilage isassociated with matrix vesicles, andthis enzyme content may facilitate thecalcification process initiated bymatrix vesicles.4 Mitochondria anddegenerating chondrocytes may play apart in the second stage ofmineralisation and collagen isimportant in the third stage forretaining the mineral in the matrix.' 2

Research on the pathogenesis ofosteoarthritis has led to the conclusionthat it may be initiated by variousindividual anatomical, mechanical, ormetabolic factors, but that there is afinal common pathway of cartilagedegeneration that eventually leads tothe total loss of articular cartilage andjoint function.' Enzyme analysis andbiochemical estimations of humanosteoarthritic cartilage led us topostulate a calcification abnormality inthe diseased tissue as one suchaetiological factor.6 This wouldprovide an explanation for a certainform of the disease and would notnecessarily explain the various diseaseprocesses that lead to different formsof osteoarthritis. Electronmicroscopical studies of arthritichuman articular cartilage have shownapatite-type crystal depositsassociated with matrix vesicles andincreased alkaline phosphatase activity

in the diseased tissue.7 8 These crystaldeposits were quite distinct anddifferent from calcium pyrophosphatecrystal deposition. In his study ofcrystal deposits in human jointsMcCarty had included apatite amongthe various crystals seen in articularand periarticular tissues ininflammatory arthritis.9"' Apatitecrystals have also been observed in thesynovial fluid of arthritic patients byDieppe" " and by Schumacher and hiscolleagues.'31 Faure" has alsoobserved calcium hydrogen phosphatedihydrate crystals in synovial tissues.We have attempted to elucidate thenature and origin of apatite-typecrystals in human osteoarthriticcartilage and have studied themechanism of their formation andassociation with matrix vesicles.7 216 17Apart from calcium pyrophosphatecrystal deposition in chondro-calcinosis, we have found at least threedifferent morphological types of apa-tite crystals in human arthritic cartil-age and have characterised these byelectron probe analysis and by cryoul-tramicrotomy. Some of our resultshave been published elsewhere'7 andhere we shall give further details of thevariety of crystals found in fresh,human articular cartilage specimens.

Materials and methods

Fresh human osteoarthritic articularcartilage was obtained from femoralheads resected for total hipreplacement. It was degenerate orresidual cartilage (type IV in thenomenclature of Ali and Bayliss6).Normal specimens were obtained fromfemoral heads after amputation ofhind quarters for osteosarcoma or othermalignant growth. Articular cartilagewas also obtained from patients whohad subcapital fracture, where thecartilage surface and tissue appeared

quite smooth and 'normal'. Small (1mm') specimens were fixed inglutaldehyde only and processed forelectron microscopy by conventionaltechniques as described elsewhere.6 18Unstained sections were used forelectron probe analysis, with an energydispersive system, in a Philips 300transmission electron microscope. Formorphological studies tissuespecimens were double fixed inglutaraldehyde and osmium tetroxideand sections were stained with uranylacetate and lead citrate.6 18

Results and discussion

Examination of more than 12specimens of human osteoarthriticcartilage has indicated that there aredifferent morphological types ofcrystals present in different layers ofarticular cartilage. It is preferable todeal separately with these differenttypes of crystals because themechanism of their formation,location, and the pathologicalconsequence may all be quitedifferent.

CRYSTAL NODULES IN DEEPERZONESOF ARTICULAR CARTILAGEThe number of membrane-boundmatrix vesicles (50 nm to 250 nm indiameter) is increased in thepericellular area of the chondrocytesin arthritic articular cartilage,especially in the tidemark region, justabove the calcified cartilage andsubchondral bone. Because of thegreat number of microscopic matrixvesicles around each chondrocyte andthe variation from one area to another,it has not been possible to count thevesicles and assign numbers. It is forthis reason that we have previouslyrelied on the quantitative estimation of



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


the marker enzyme alkalinephosphatase. which is sometimes 30times as high in osteoarthritic tissue asin the normal cartilage.6 7 Associatedwvith these vesicles, and oftenoriginatinig from them. are densemineral nodules (0.1 to 0 4 zm indiameter) composed of fine crystals(Fig. 1). Their morphology andelectron probe analysis (with acalcium: phosphorus ratio exactly thatof hydroxyapatite) indicates that they

are apatite crystals. Theirpresence in arthritic cartiimply several path(consequences, including redof the tidemark. reversion ocartilage to growth phase. o0increased tendencv towarcalcification. Although sucare also present in normalnear the tidemark, it is the inumbers, distribution. andof occurrence right Up to the

Fig. I Electron microscopical appearance of human osteoarthritic ar,cartilage (from 32 year old wonian) showing apatite crystal niodules (tVmatrix vesicles (arrows) in the deep zone.

W.

Fig. 2 Human arthritic cartilage (from 64 year old man) showing a ch,and dense cuboid crystals (type II) in the pericellular region of the co

matrix under the articular surface.

increasedilage mayclogical

of articular cartilage that is particularlynoticeable.

Iupi a6'LUll i)ENSE, 'CUBOID' CRYSIAtS IN)t articular THE SL'RFACE ZONE OF

r a general ART'HRITIC CAR'IILAGErds tissuL These microscopic, dense cuboid'h crsstals crvstals appear mostly just under thespecimens surface of arthritic articular cartilagencrease in in the pericellular matrix surroundingincidence chondrocytes (Fig 2) In somemmld-zone specimens they appear as a band of

fine particles in the surface zone. Theyoften appear square shaped in theplane of the section (Figs. 2 and 3)suggestive of a cuboid shape, thoughthis needs further confirmation. Thesetypes of 'cuboid' crystals have neitherbeen seen in articular cartilage nordescribed by anyone else previously,although we have been aware of theirpresence in arthritic cartilage for someyears. They vary in size from 50 nm to200 nm and are evident in unstainedsections and in sections obtained bycryoultramicrotomy.6 thus ruling outany preparative artefacts.

Electron probe analysis of thesecuboid crystals indicates that they are

% mainly composed of calcium andA phosphorus. Analvsis of over 60X cuboid crystals in a section of arthritic

*e ,. cartilage (from a 70-year-old woman)gave a calcium to phosphorus ratio of

ticular 1 72:1 while that for an

ypeC I) atid hydroxyapatite standard was 1 79:1under the same conditions (detailedresults will be published elsewhere).This apatite-like Ca:P ratio has beenconfirmed by analysis of cuboid

i crystals found in cryosections.'7 It hasbeen difficult to reconcile the apatite-like Ca: P ratio with the cuboid shapeof the crystals. There is a small amountof magnesium present with calcium

4|< and phosphorus which is indicative ofWhitlockite; in their cuboid habit

#P* these crystals appear very much likeWhitlockite. Electron diffraction andscanning electron microscopicalstudies are being undertaken toresolve this problem.These cuboid crystals have now

been found in the last six consecutivearthritic specimens that we haveexamined by electron microscopy. In

=. some arthritic specimens they arepresent in the surface, intermediate,and deep zones of articular cartilage.

ondrocyte They are absent in young articularllagenous cartilage but are sometimes present in

old 'normal' articular cartilage

4n:.ri ....Kf R

A.

2

...



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Formation ofcalcium phosphate crystals Suppl p 47

s t s .F.k.;

..... .X. ,5.} .: '.iP7

;-: S -o *. ... ,^- v v<'

&ati;s v wSs dt.., ... ^

[f ; 4, fv s.;

Fig. 3 Higher magnification picture ofdense 'cuboid' crystals (type I1) andseveral matrix vesicles closely associated with them (same specimen as fig. 2).

4

.

_Y.ei ::

500nma

Fig. 4 Surface of arthritic articular cartilage (friom 64 year old man) showingneedle-shaped crystal clusters and individual crystals ofapatite (type III) in an

unstained section.

specimens. In one specimen fromsubcapital fracture of the femoral headsome of these cuboid crystals were

present adjacent to large areas ofpyrophosphate crystal deposits. Insome areas of arthritic cartilage theyare closely associated with matrixvesicles (Fig. 3). These unique cuboidcrystals in arthritic specimens are toosmall to be viewed by light microscopyand may not show up in radiographs ofjoints.

NEEDLE-SHAPED CRYSTAL

CLUSTERS ON THE SURFACE OF

ARTHRITIC CARTILAGE

These fine, needle-shaped crystalsoccur in clusters on the surface ofarthritic cartilage in the acellularamorphous band termed the laminasplendens (Fig. 4). Individual needlesin section are approximately 80 nm inlength and 8 nm in width. They appearlike, and give the same Ca: P ratio as

hydroxyapatite. These needle-crystal

clusters on the cartilage surface appearvery similar to apatite crystalsobserved by Dieppe"' andSchumacher"3 by electron microscopyin synovial fluid, synovial membrane,and synovial phagocytes. Matrixvesicles are not associated with theseclusters of fine crystals, whichprobably originate elsewhere in thejoint and are deposited on the surfaceof articular cartilage as anepiphenomenon. In contradiction tothis we have seen these crystals on thesmooth surface of a specimen obtainedas avascular necrosis of the femoralhead where there was no damage orerosion of articular cartilage.

In conclusion, we believe that thepresence of these three different typesof calcium phosphate crystals and theassociated increase in matrix vesiclesand alkaline phosphatase activity ofarthritic articular cartilage may imply acalcification abnormality and mayhave aetiological implications incertain forms of osteoarthritis. Thediscovery of cuboid (Whitlockite)crystals in human articular cartilage isquite new and requires furthercharacterisation.16 17 PreviouslyWhitlockite has been reported only inpathological calcification sites in softtissues such as lung and spleen"9 and inhard tissues such as carious dentine.20

References

1 Ali S Y. Mechanisms of calcification. In:Owen R, Goodfellow J, Bullough P, eds.Scientific foundations of orthopaedicsand traumatology. London:Heinemann, 1980: 175-84.

2 Ali S Y. Calcification of cartilage. In:Hall B K, ed. Cartilage. Vol. I Structureand function. New York: AcademicPress, 1982: 343-78.

3 Anderson H C. Matrix vesicles of cartil-age and bone. In: Bourne G H, ed. Thebiochemistry and physiology of bone.Vol 4. New York: Academic Press,1976: 135-57.

4 Ali S Y, Sajdera S W, Anderson H C.Isolation and characterization ofcalcifying matrix vesicles fromepiphyseal cartilage. Proc Natl Acad SciUSA 1970; 67: 1513-20.

5 Ali S Y. New knowledge ofosteoarthritis. J Clin Pathol 1978; 31,suppl 12: 191-9.

6 Ali S Y. Bayliss M T. Enzymic changesin human osteoarthritic cartilage. In: AliS Y, Elves M W, Leaback D H, eds.Normal and osteoarthrotic articularcartilage. London: Institute ofOrthopaedics, 1974: 189-205.

P...

A&I-kOlk ,* .-

, 1%. .1NO.V- i.:,



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


7 Ali S Y. Matrix vesicles and apatitenodules in arthritic cartilage. In:Willoughby D A, Giroud J P, Velo G P,eds. Perspectives in inflammation.Lancaster: MTP Press, 1977: 211-23.

8 Ali S Y. Mineral-containing matrixvesicles in human osteoarthroticcartilage. In: Nuki G, ed.Aetiopathogenesis of osteoarthrosis.London: Pitman, 1980: 105-16.

9 McCarty D J, Gatter R A. Recurrentacute inflammation associated withfocal apatite crystal deposition. ArthritisRheum 1966; 9: 804-19.

10 Halverson P B, McCarty D J.Identification of hydroxyapatite crystalsin synovial fluid.Arthritis Rheum 1979;22: 389-95.

11 Dieppe P A. Crystal inducedinflammation and osteoarthritis. In:Willoughby D A, Giroud J P, Velo G P,

eds. Perspectives in inflammation.Lancaster: MTP Press, 1977: 225-31.

12 Dieppe P A. Crystal-inducedarthropathies and osteoarthritis. In:Buchanan W W, Carson Dick W, eds.Recent advances in rheumatology.London: Churchill Livingstone, 1981:1-18.

13 Schumacher H R. Pathogenesis ofcrystal-induced synovitis. Clinics inRheumatic Diseases 1977; 3: 105-31.

14 Schumacher H R, Somlyo A P,Rose L T, Maurer K. Arthritis asso-ciated with apatite crystals. Ann InternMed 1977; 87: 411-6.

15 Faure G, Netter P, Malaman B,Steinmetz J. Monocrystalline calciumhydrogen phosphate dihydrate indestructive arthropathies ofchondrocalcinosis. Lancet 1977; ii:142-3.

16 Ali S Y, Griffiths S. New types ofcalcium phosphate crystals in arthriticcartilage. Semin Athritis Rheum 1981;11, suppl 1:124-6.

17 Ali S Y, Griffiths S. Matrix vesicles andapatite deposition in osteoarthritis. In:Ascenzi E, Bonucci E, de Bernard B,eds. Matrix vesicles: proceedings of theIII international conference on matrixvesicles. Milan: Wichtig, 1981: 241-7.

18 Ali S Y, Wisby A, Craig-Gray J.Electron probe analysis of cryosectionsof epiphyseal cartilage. Metabolic BoneDisease and Related Research 1978; 1:97-103.

19 Gatter R A, McCarty D J. Pathologicaltissue calcification in man. Archives ofPathology 1967; 84: 346-53.

20 Rowles S L, Levine R S. The inorganiccomposition of arrested carious dentine.Caries Research 1973; 7: 360-7.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Calcified tendinitis: a reviewG. FAURE,' G. DACULSI2

From the 'Clinique Rhumatologique et Laboratoire d'immunologie, Faculte A de Medecin, Universite de Nancy I, 54500Vandoeuvre les Nancy, France and 2U225 INSERM, Faculte de Chirurgie Dentaire, Place Alexis Ricordeau, 44042 Nantes,France

Calcified tendinitis in clinical practice

Calcified tendinitis is a commondisorder. Many names have been usedto describe it: some of them, such as'calcific periarthritis', emphasise theextra-articular site of the deposit;others, such as 'periarticular apatitedeposition', mention the nature of thecompound found in the calcification;and more recent ones, such as'calcifying tendinitis',-'3 emphasise the.active process that might explain thedeposition. Differentiated fromarthritis at the end of the nineteenthcentury, this syndrome has onlyrecently been related to the presenceof apatite in tendon sheaths.4'5 It canaffect almost any tendon at itsinsertion and is most common aroundthe shoulder joint. Rheumatologistsand radiologists have often describedthis shoulder abnormality, leading toits progressive differentiation fromother painful shoulder syndromes. 12

This review will discuss calcificperiarthritis of the shoulder as amodel. Clinical features of thesyndrome are variable and includepain and inflammation. The keydiagnostic feature is the radiograph.Clinical evolution is simple and thecondition often resolvesspontaneously. Some cases arepersistent and may require aggressivetreatment, including surgery.Numerous questions are still

unanswered about this disorder, whichis rarely associated with metabolicabnormalities of calcium andphosphorus. These include: (a) thenature of the mechanism leading tocalcium salt deposition; (b) thefrequent asymptomatic tolerance ofsuch calcification, and, conversely, (c)the initiating agent of inflammatoryflares; and (d) the way in which thematerial disperses.

Request for reprints to: G. Faure.

CLINICAL FEATURESAccording to Welfling calcificperiarthritis is responsible for 7% ofpainful shoulder syndromes,' whichhave various presentations.

(1) Chronic symptoms-more orless severe pain; tenderness leading tovarious degrees of incapacitation.These symptoms induce the demandfor radiographs, which reveal thepresence of deposits.

(2) Acute inflammatory crisis withsevere pain, tenderness, and localoedematous inflammation sometimesleading to restricted active an*d passivemotion. Fever and malaise may beobserved.

(3) Totally asymptomatic deposits.According to Bosworth et al.,

clinical symptoms occur in from 34%to 45% of patients in whomcalcifications are found.' Biochemicaland haematological tests are notuseful, and will only show non-specificevidence of inflammation.

RADIOLOGICAL FEATURESSimple anteroposterior and lateralx-ray films are usually sufficient to seeshoulder calcification, though specialviews in internal or external rotationmay be necessary.'2 Other techniqueshave been described to obtain betterpick up, including xerography andscanning, but many small deposits areprobably missed.The calcification is usually in the

supraspinatus tendon, and variousappearances have been described. Thedeposits may be very thin, outliningthe tendon sheath, or hazy, and theyvary in density and definition (Fig. 1).Only in cases of disturbance of thecalcium to phosphorus ratio-forexample hyperparathyroidismsecondary to renal failure-is theremassive calcification.'3 The size

Fig. 1 Calcific periarthritis oftheshoulder. Calcification is obvious; ithas already migrated fromsupraspinatus region to bursa area.

usually varies between a fewmillimetres and about 1-5 cm.

Calcification has been describednear almost every joint"2 although theprecise intratendinous or

paratendinous location of the stone isnot always seen. The deposits may bemultiple,'4 which French authorsidentify as 'maladie des calcificationstendineuses multiples'.' 1" Bilateralcalcification is seen in about half ofshoulder cases and deposits are oftenseen in other locations-for example,near the hip joint-if otherradiographs are taken.

Fig. 2 Acute bursitis ofthe shoulder(same patient). Morphological aspectsofprevious calcification is modified; itis less dense and probably locatedinside bursa.

Introduction



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Sequential x-ray films may showstatic calcification, a growing deposit,change in the location, and evenspontaneous disappearance withoutany acute clinical flare. During theinflammatory crisis, the calcificationsusually follow a very well describedcourse. They become less defined,more cloudy, and migrate into thebursa (Figs. 2 and 3); they may or maynot disappear completely within a fewdays or weeks. The reappearance ofcalcification is not well substantiated(Fig. 4). Some authors emphasise therelation of the deposit to the bonesurrounding the tendon insertionpoint.17 Destructive changes havebeen reported in advanced cases, andMcCarty et al. have described a specialsyndrome associating a destructivearthropathy of the shoulder, apatitedeposits. and high collagenaseactivities in the synovial fluid. Theycall it the 'Milwaukee shouldersyndrome'. 19

or articular origin present withdifferent symptoms. Analysis of fluidand radiological investigations are alsodiscriminative. Extra-articularossification is radiologically different,the deposits being trabeculated. andarticular chondrocalcinosis alsoappears quite different on aradiograph. Tendinous calcificationcontaining calcium pyrophosphatedihydrate may be misleading2" butis usually associated withchondrocalcinosis.

TREATMENT

As both symptoms and deposits oftendisappear spontaneously, bothclinician and patient should generallyabstain from interfering. Analgesicsand non-steroidal anti-inflammatorydrugs (NSAID) are useful, as well aspatience.

During an acute inflammatory crisismore powerful NSAID drugs such asindomethacin or phenylbutazone may

prove necessary, and some authorsrecommend adding colchicine as ingout and pseudogout. Aspiration offluid may also reduce symptoms. Localinjections of corticosteroids are usedby some clinicians, but may themselvescause microcrystal-inducedinflammation, and the risk of infectionmust be taken into account. Lavagebetween attacks has been advocated,but does not always seem rewarding.Very painful resistant cases havereceived x-rav treatment. Calciuminhibitors have also been tried but theresults are not convincing.

Surgery may prove successful, and afew well documented cases treated bysurgical removal of the deposit havebeen described.:'

LABORA1 ORY INVESTIGATIONSThe material aspirated from acutelypainful shoulders has been studiedthoroughly in a few cases, withinteresting results. Firstly. they

DIFFERENI IAL D)IAGNOSISThe diagnosis is usually easy. Otherpainful shoulder syndromes of osseous

Fig. 3 Sanme patient one month later.Calcification has comnpletel/vdisappeared, though vestigial oneremains in the supraspinatus region.

Fig. 4 Same patient one year later.Calcification has reappeared, patientpresents with a moderately painfulshould er.

NO .. '.. bm

Fig. 5 Scanning electron microscopy. Bar 5 in. One globule-like structure isseen here in situ in section ofa tendon sheath calcification, appearing like a stoneengulfed in mortar.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Calcified tendinitis: a review Suppl p 51

allowed the identification of thematerial.4 Recent studies by our grouphave also isolated another compound,unidentified so far, which is differentfrom stoichiometric apatite, and couldbe the result rather than the cause ofthe inflammation.2'The usual specimens obtained at

operation consist of a gritty mass ofsandy material or a toothpaste-likefluid. These have been recognisedsince Codman first described thecalcifications, and in 1934 the depositswere described as 'a white amorphousmass composed of many small roundor ovoid bodies'.6 Microscopically, atlow magnification, calcificationsappear as shiny amorphous coins (in aliquid phase).4

Physical studies first used x-raydiffraction then infrared spectrometry.McCarty et al.4 and Thompson et al.5identified an apatite compound byx-ray diffraction, and infraredspectrometry showed that it was acarbonated apatite.22 This has beenconfirmed by other physical meanssuch as thermogravimetry.22 Ourresults are similar but it should benoted that all these methods considerthe material as a whole and do notallow definition of individual particles.

Scanning electron microscopyallowed us to have a closer look at theexternal aspect of deposits24-25; weobserved an extremely heterogeneousmaterial, composed of globular bodieslooking like large rocky bulks engulfed

in mortar (Fig. 5). Study of individualcrystals needed even moresophisticated means. Transmissionelectron microscopy (TEM) allows thevisualisation, on ultrathin sections, ofdense globular structures amongnumerous isolated or clumped crystals(Fig. 6). Individual crystals may beseen in high resolution transmissionelectron microscopy. The crystals aremuch larger than classic apatite crys-tals, such as those observed in bone ordental enamel; some ofthem appear ashomogeneous hexagons, the width andthickness of which can be measured.The parameters differ from onepatient to another. Some of them haveclear cut edges, while others seem topossess some sort of coating; defectscan also be seen in several crystals(Fig. 7).

Interplanar spacings may sometimesbe measured, and are consistent withthese parameters in stoichiometrichydroxyapatite. These isolatedcrystals are, however, different fromthe globular material. Wavelengthdispersive spectrometry in a scanningmode gave calcium to phosphorusmolar ratios in both globules andisolated crystals not statisticallydifferent from control geologicalapatites, but similar microanalysis in atransmission mode provided lowerratios, with differences between theinside and outside of globular bodies.

All these techniques identified thematerial of calcific deposits as calcium

Fig. 6 Transmission electron microscopy. Bar =5 p.m. Dense globularstructuresof various size among scattered isolated crystals.

carbonate apatite,26" but the greatheterogeneity of the material is a newfeature that may open new fields ofresearch. The differences observedbetween patients also deserves study.The exact location of deposits is not

well elucidated, but Welfling describessome as intratendinous and somesuperficial.9Sandstrom described necrosis

interpreted as being secondary to'local anaemia and vascular changes',which favoured deposition ofcalcifying material.7 Uhthoff's groupfavour an active mechanism ofcalcification, with an initial cartilagedegeneration of the tendon followedby calcification of the transformedtissue.'` They describe four stages inthe calcifying process: precalcificphase; calcific phase; resorptic phase;and repair phase. These four patternsmay occur concomitantly in anindividual patient. This hypothesiswould explain the heterogeneity andpleomorphic nature of the lesions, butthe so-called 'resorptive phase' seemsarguable, particularly because of thevery small number of cells observed.

Using TEM, the same group studiedthe ultrastructural localisation ofcalcium in tendinitis. They found it inmatrix vesicle-like structures seeneither singly among collagen or inaggregates.2"2'

Discussion

New light should soon be shed oncalcification in several aspects: (a)better identification of the calcifyingprocesses; (b) more knowledge onepidemiology; (c) better definition ofthe nature of the compounds,hopefully leading to understanding ofthe aetiopathological mechanisms ofthe disease.The identification of deposits has

improved since the features of thecondition have become recognised,and should be made even easier byfurther technical improvementsallowing visualisation of tinycalcifications. This should help definethe epidemiology and explainapparent differences in the incidenceand age of onset between France andBritain. So far, only oneepidemiological study has beenperformed in the United States.? Theexistence of a possible genetic linkshould also be re-evaluated,3' in spite



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Suppi p 52 Annals of the Rheumatic Diseases

Is

. .........

Fig. 7 High resolution transmission electron microscopy. Bar= 0 01 ,um. Singlehexagonal crystal ofapatite presenting a coat on its surface.

ot the negative results of the HLAstudies already published.'6

Associations have been reportedbetween calcifications and severaldiseases: diabetes, thyroid disorders,and others.3' A metabolic link was alsonoted between calcifications, uraemia,and haemodialysis.32 33 Finally,Bosworth suspected the influence ofrepetitive motion performed in a

professional context.' The relationshipbetween shoulder use and theappearance of calcification is not welldocumented and should bere-examined.

The physicochemical findings in thepast few years have been rewarding. Ina recent review, Legeros brings manydata favouring a calciumcarbonate/apatite composition forheterotopic calcifications in man."Our results, however, made it obviousthat there is a high degree ofheterogeneity in these compounds andthat the answer is probably not thatsimple. Every possible analytical andmicroscopical means should be used tostudy all types of calcification. It mightbe rewarding to compare idiopathicdeposits with those found in other

diseases such as hyperparathyroidism,hvpervitaminosis D, collagen vasculardisease, and the milk alkali syndrome.The exact incidence of other types ofdeposits such as pyrophosphate,reported in tendons by Gerster,2"should also be examined.

Several aetiopathological schemeshave already been proposed.' 2 7 15 16 35

The classic hypothesis favours a localand initial necrosis of the tendon.leading to the deposition of calcifiedmaterial. However, somecalcifications obviously occur in theabsence of anv necrotic phenomenon.A more recent idea, based onhistological and ultrastructuralobservations, suggests the occurrenceof an initial cartilage metaplasia of thetendon, followed by an activemultifocal and cell-mediated calcifyingprocess. The intracellular orextracellular site of the first deposit isnot completely clear and requires anultrastructural or physical study ofcollagen fibres.Some other mvsterious features

remain for example, why shouldcalcifications be most common aroundthe supraspinatus tendon? How canthe long-lasting tolerance of somedeposits be explained, while otherslead to acute crisis? What is theinitiating mechanism in these acuteepisodes? Does it involve amodification of the physicochemicalstructure of the material or a change inits microenvironment? How is it thatsome pathological, and eventuallypainful, deposits disappearspontaneously?Experimental models aid

understanding of some of theseproblems. For instance, syntheticapatite crystals and/or natural apatiteshave phlogistic properties ininflammation models.:"- A thoroughstudy of the macrophage in suchmodels might help us to understandhow calcified material can disappear.This could also be elucidated bymetabolic studies if a dissolutionprocess takes place in thesephenomena.More sophisticated models are

obviously necessary, if not to explainall the mysterious events of thissyndrome, at least to define newpreventive or curative treatments, orboth. Special care should, however, betaken in extrapolating such results tohuman disease,39- 42 mainly because, to



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Calcified tendinitis: a review Suppl p 53

our knowledge, no similar disease isdescribed in veterinary publications,implying that metabolic pathways maybe very different in man and inanimals.

References

1 Uhthoff H K. Calcifying tendinitis an

active cell mediated calcification.Virchows Arch [PatholAnat] 1975; 366:51-8.

2 Uhthoff H K, Sarkar K, Maynard J A.Calcifying tendinitis. A new concept ofits pathogenesis. Clin Orthop 1976; 118:164-8.

3 McKendry R J R, Uhthoff H K, SarkarK, Hyslop P S G. Calcifying tendinitis ofthe shoulder: prognostic value of clinicalhistologic and radiologic features in 57surgically treated cases. J Rheumatol1982; 9: 75-80.

4 McCarty D J, Gatter R A. Recurrentacute inflammation associated withfocal apatite crystal deposition.ArthritisRheum 1966; 9: 804-19.

5 Thompson G R, Ming Ting Y, Riggs GA, Fenn M E, Denning R M. Calcifictendinitis and soft tissue calcificationresembling gout. JAMA 1968; 203:464-72.

6 Codman E A. The shoulder. Boston:Thomas Todd, 1934.

7 Sandstrom C. Peridentinis calcarea. Acommon disease of middle life. Itsdiagnosis pathology and treatment.AJR1938; 40: 1-21.

8 Bosworth B M. Calcium deposits in theshoulder and subacromial bursitis. Asurvey of 12,222 shoulders. JAMA1941; 116: 2477-82.

9 Welfling J. Les calcifications del'6paule. I Diagnostic clinique. RevRhum Mal Osteoartic 1964; 31: 265-71.

10 Swannell AJ, Underwood F A, Dixon AS J. Periarticular calcific depositsmimicking acute arthritis. Ann RheumDis 1970; 29: 380-5.

11 Fam A G, Pritzker K P H, Stein J L,Houpt J B, Little A H. Apatiteassociated arthropathy: a clinical studyof 14 cases and of 2 patients with calcificbursitis. J Rheumatol 1979; 6: 461-71.

12 Resnick D. Calcium hydroxyapatitecrystal deposition disease. In: ResnickD, Niwayama G, eds. Diagnosis ofboneand joint disorders. Philadelphia: W BSaunders, 1981: 1575-97.

13 Smith F W, Junor B J R. Peri-articularcalcification with fluid levels insecondary hyperparathyroidism. Br JRadiol 1978; 51: 741-2.

14 Pinals R S, Short C L. Calcificperiarthritis involving multiple sites.Arthritis Rheum 1966; 9: 566-74.

15 Welfling J, Kahn M F, Desroy M,Paolaggi J B, de Seze S. Lescalcifications de l'6paule. II. La maladiedes calcifications tendineuses multiples.

Rev Rhum Mal Osteoartic 1965; 32:325-34.

16 Amor B, Kahan A, Cherot A, DelbarreF, Rabaud M, Aubouy G. Lerhumatisme a hydroxyapatite (lamaladies des calcifications tendineusesmultiples). II. Etude microscopique.Antigene HLA arthrite experimentale.Pathogenie. Rev Rhum Mal Osteoartic1977; 44: 309-16.

17 Meneghello A, Bertoli M, Romagnoli GF. Unusual complication of soft tissuecalcifications in chronic renal diseasethe articular erosions. SkeletalRadiology 1980; 5: 251-2.

18 McCarty D J, Halverson P B, CarreraG F, Brewer B J, Kozin F. 'Milwaukeeshoulder': association ofmicrospheroids containing hydroxyapatite crystals, active collagenase andneutral protease with rotator cuffdefects. I Clinical aspects. ArthritisRheum 1981; 24: 464.

19 Halverson P B, Cheung H S, McCarty DJ, Garancis J C, Mandel N. 'Milwaukeeshoulder': association ofmicrospheroids containing hydroxyapatite crystals, active collagenase andneutral protease with rotor cuff defects.II. Synovial fluid studies. ArthritisRheum 1981; 24: 474-83.

20 Gerster J C, Baud C A, Lagier R,Boussina I, Fallet G H. Tendoncalcifications in chondrocalcinosis. Aclinical, radiologic, histologic andcrystallographic study. Arthritis Rheum1977; 20: 717-22.

21 Faure G, Daculsi G, Netter P, GaucherA, Kerebel B. Apatites in heterotopiccalcifications. Scanning ElectronMicroscopy (in press).

22 Legeros R Z, Contiguglia S R, Alfrey AC. Pathological calcifications associatedwith uremia: 2 types of calciumphosphate deposits. Calcif Tissue Res1973; 13: 173-7.

23 Saez-Clavere L, Legros R, Arlet J,Bonel G. Etude cristallochimique dedeux calcifications sous deltoidiennes.Rev Rhum Mal Osteoartic 1980; 47:383-92.

24 Faure G, Netter P, Coxam B, Raul P,Pourel J, Gaucher A. Place de l'apatiteen pathologie microcristalline. AnnalesMedicales de Nancy 1979; 18: 1281-4.

25 Faure G, Netter P, Malaman B,Steinmetz J, Duheille J, Gaucher A.Scanning electron microscopic study ofmicrocrystals implicated in humanrheumatic diseases. Scanning ElectronMicroscopy 1980; II: 163-76.

26 Montel G, Bonel G, Trombe J C,Heughewaert J C, Rey C. Progres dansle domaine de la chirnie des composesphrophores solides a structure d'apatite.Application a la biologie et au

traitement des minerais. Pure andApplied Chemistry 1980; 52: 973-87.

27 Legeros R Z. Apatites in biological

systems. Progress in Crystal Growth andCharacterization 1981; 4: 1-45.

28 Sarkar K, Uhthoff H K. Ultrastructurallocalization of calcium in calcifyingtendinitis. Arch Pathol Lab Med 1978;102: 266-9.

29 Dereszewski G, Howell D S. The role ofmatrix vesicles in calcification. Trends inBiochemical Sciences 1978; 3: 151-2.

30 Cannon R B, Schmid F R. Calcificperiarthritis involving multiple sites inidentical twins. Arthritis Rheum 1973;16: 393-6.

31 Wright V, Haq A M. Periarthritis of theshoulder. I. Aetiological considerationswith particular reference to personalityfactors. Ann Rheum Dis 1976; 35:213-9.

32 Caner J E Z, Decker J L. Recurrentacute (gouty) arthritis in patients withchronic renal failure treated withperiodic hemodialysis. Am J Med 1964;36: 571-82.

33 Moskowitz R W, Vertes V, Schwartz A,Marshall G, Friedman B. Crystal-induced inflammation associated withchronic renal failure treated withperiodic hemodialysis.Am J Med 1969;47: 450-60.

34 Legeros R Z, Legeros J P. Phosphateminerals in human tissues. In: Nriago J,Moore P, eds. Phosphate mineral. (inpress).

35 Urist R, Moss M J, Adam J M.Calcification of tendon. A triphasic localmechanism. Archives of Pathology1964; 77: 594-608.

36 Denko C W, Petricevic M.Hydroxyapatite crystals-inducedinflammation and prostaglandin El. JRheumatol 1979; 6: 117-23.

37 Glatt M, Dieppe P, Willoughby D.Crystals-induced inflammation, enzymerelease and the effects of drugs in the ratpleural space. J Rheumatol 1979; 6:251-8.

38 Cheung H S, Halverson P B, McCarty DJ. Release of collagenase neutralprotease, and prostaglandins fromcultured mammalian synovial cells byhydroxyapatite and calciumpyrophosphate dihydrate crystals.Arthritis Rheum 1981; 24: 1338-44.

39 Luben R A, Sherman J K, Wadkins C L.Studies of the mechanism of biologicalcalcifications. IV. Ultrastructuralanalysis of calcifying tendon matrix.Calif Tissue Res 1973; 11: 39-55.

40 Bridges J B, McClure J. Experimentalcalcifications in a number of species.Calif Tissue Res 1972; 10: 136-49.

41 McClure J, Gardner D L. Theproduction of calcification in connectivetissue skeletal muscle using variouschemical compounds. Calif Tissue Res1976; 22: 129-35.

42 Doyle D V, Dunn C J, Willoughby D A.Potassium permanganate inducedcalcergy. A model to study the effects ofdrugs on hydroxyapatite crystaldeposition. J Pathol 1979; 128: 63-70.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Intra-articular apatite crystal depositionH. RALPH SCHUMACHER,' P. VERGHESE CHERIAN,' ANTONIO J. REGINATO,1THOMAS BARDIN,2 SUSAN ROTHFUSS1

From the 'University ofPennsylvania School ofMedicine, Philadelphia, USA and 2Clinique Rheumatologique, HospitalLariboisiere, Paris, France

Introduction

The role of apatite crystals in calcificperiarthritis is well recognised.1-5 Formany years, however, intra-articularapatites were not thought to play anypart in joint disease. They wereidentified in only a small percentage ofmenisci examined grossly and by x-raydiffraction by McCarty et al. in 1966.6In 1975-7 we7' and Dieppe et al.'using electron microscopy andelemental analysis on joint fluidsfound apatite crystals in varioussituations including acute otherwiseunexplained arthritis. We review ourfindings since that time.

Methods for crystal identification

Techniques used to identify apatitecrystals in synovial fluids have variedin different series and may affect theimportance of the findings. Sensitivity

and specificity of most technicnot been fully defined. Regimicroscopy may allow visualilarger clumps of apatite crglossy, slightly irregular glob2-10,um in diameter (Fig. 1few of such aggregalbirefringent.' Stains for calcphosphorus such as Von Koalizarin red stains can hellidentify these globules.' Nexamined the sensitivity of thred stain and find that it cc0- 005 ,ug of synthetic apatitesynovial fluid.' The techniqinot totally specific and is usedscreening test.Scanning electron mic

(SEM) can also identify clunsame shape1" and allowsprobe elemental analysis, whinot only whether calciiphosphorus are present Iwhether their approximate r

E

.-- w.

8.# '.

ques have close to the 1 67:1 expected with wellular light crystallised apatite.isation of Transmission electron microscopyrystals as allows identification of the tiny thinules from 50-250A diameter needles within the). Only a clumps that are typical of apatite.' Thistes are technique may be used on smallcium and amounts of fluid and also on either thinissa's and sections or synovial fluid dried ontop further formvar coated grids.11 Used on synovialWe have fluid the technique detects downie alizarin to at least 0-002 mg/ml of apatite andan detect avoids possible artefacts of chemicalin 1 ml of fixation. Examination of thin sectionsue is still also allows demonstration of whetheronly as a and by which cells crystal clumps are

phagocytised. This may obviously be:roscopy important for the phlogisticips of the significance of the apatite. Elementselectron may also be analysed as with SEM.

ich shows Electron diffraction patterns can alsoum and be made for comparison with knownbut also standards, but the problems and valueratios are of this have not yet been worked out.

When large amounts of suspectedapatite are present x-ray diffraction isgenerally accepted as the definitivemethod of analysis, although smallamounts of certain crystals may bemissed when mixed with otherpredominant crystals. A semi-quantitative screening techniquefor apatite using 14C disodiumetidronate binding is also beingstudied.12 Intracellular crystals mightbe missed with this method. Infrared

,04| 1t spectrophotometry has been used tosupplement findings with othertechniques1" and with the use ofFournier transform may allow

^.$ detection of small amounts of crystalsmixed in with other predominant ones.

..F

..:.0

Fig. 1 Lightmicroscopical viewof irregularclumpsofapatite crystals (arrows) insynovial fluid. Note erythrocytes (E) and neutrophils (N) for size comparison.

Apatite in joints

Apatite has been described in knees,shoulders, hips, wrists, firstcarpometacarpal joints, meta-carpophalangeal, proximal anddistal interphalangeal joints, and



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Intra-articular apatite crystal deposition Suppl p 55

metatarsophalangeal joints includingthe 1st.5 89 13 We have studied a totalof 104 patients in whom apatite wasfound by transmission electronmicroscopy and find a similardistribution of joints containing thesecrystals. Cases reported have generallybeen from studies of specific diseasesor selected by referral so the incidenceof identifiable apatite in the generalpopulation remains unclear.Women seem to be affected more

often than men,5 and we know of noreports in children except inassociation with collagen disease. Ourstudy included 58 women and 46 men;there was a high incidence of elderlypatients because we included 44patients from a study of osteoarthritis,the youngest patient was 36 years old.

CLINICAL PRESENTATIONS

Otherwise unexplained acuteinflammatory arthritis with jointeffusions containing apatite has beenreported by Fam et al., Dieppe et al.,and our group at knees, wrists, hips,proximal interphalangeal joints,metacarpophalangeal joints, and firstmetatarsophalangeal joints' 8 9 andmay well occasionally occur at anyjoints where crystals can deposit.Some patients have had multipleattacks and have also occasionally hadcalcific periarthritis. Attacks mayresolve without treatment; someprogress to chronic erosive arthritis."3Synovial fluid leucocyte counts inclinically acutely inflamed joints haveranged from 1 8-84 x 109/P5 9

Neutrophils have ranged from 27%YO to90% with a predominance in acutecases. Crystals may be seen to bephagocytised by synovial fluidneutrophils and mononuclear cells(Fig. 2).8Most recent studies have also

identified apatite crystals in some jointeffusions without any dramaticevidence of inflammation.8 9 18 14 15Many such joints have hadosteoarthritis, but the real incidence ofapatite deposition in osteoarthritis andhow often it occurs early in still grosslynormal cartilage is not yet known. Wefound apatite in 44 of 100osteoarthritic effusions afterexamining just one grid on each bytransmission EM.'5 Dieppe et al.found apatite in 9 out of 34 unselectedosteoarthritic fluids with scanningelectron microscopy (SEM).6 Using

.5..

R -AjwA

'9....0

Fig. 2 A clump oftiny crystals proved to be apatite by elemental analysis lie in a

vacuole ofa synovial fluid mononuclear cell. Electron micrograph x20 000(original magnification).

TEM Ali found apatite more often inosteoarthritic than other cartilages butdid not describe incidence.'7 Even inosteoarthritis with 'non-inflammatory'joint effusions apatite has beenphagocytised by synovial lining cellsthat might release proteases andcollagenase and be a major factor inprogressive joint damage."3

RADIOGRAPHIC FINDINGSJoints with acute or subacuteapatite-associated arthritis may showcalcification in the synovial space (Fig.3).' The linear cartilage or meniscalcalcification seen with CPPD is notexpected. Some x-ray films show nocalcification despite laterdemonstration of large amounts ofapatite. In fact, only 21 of our 104patients had obviousx-ray evidence ofcalcification. Patients may also haveperiarticular calcifications at thesymptomatic site or elsewhere.Erosive arthritis may develop'" '8 andchanges of osteoarthritis may bepresent.'4

In patients with clinicalosteoarthritis the severity ofradiographic changes of osteoarthritistends to correlate with the incidence ofapatite deposition.

Apatite associated with other diseases

Apatite crystals have been found in ahigh percentage of specimens ofsynovial fluid or cartilage frompatients known also to have CPPDdeposition. 420-22 Either or bothcrystals may be phagocytosed; in ourstill limited experience CPPD seemsmore often to be intracellular.Other diseases that have been

associated with clinically significantapatite arthritis include scleroderma,dermatomyositis, and systemic lupuserythematosus" ; renal failure treatedby haemodialysis;l 19 and possibly highdose vitamin D treatment.26 Apatitealso occurs in other joint diseases,presumably as a secondarycomplication. We found apatite in 8out of 32 effusions from joints with



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Fig. 3 X-ray films ofhand, showing apatite deposits best in the joint space ofthefifth metacarpophalangeal joint and the capsule of the secondmetacarpophalangeal joint. This patient had a history of recurrent acute arthritisthat evolved into erosive arthritis.

rheumatoid arthritis; the crystalscorrelated with the severity ofsecondary osteoarthritis.27 In certaincases acute dramatic inflammation in asingle joint may be induced by crystals

rather than by the underlyingrheumatoid arthritis. Table 1 showsdiagnoses other than osteoarthritis inour 104 patients with apatitedeposition.

Table 1 Associated diseases other than osteoarthritis in 104 patients withintra-articular apatite crystals

Diseases No ofpatients

Dermatomyositis, scleroderma 6Renal failure on dialysis 10Oxalate deposition 1Hypothyroidism 3Haemochromatosis 3Osteochondromatosis 1Acromegaly 2CPPD deposition 22Gout 5Rheumatoid or psoriatic arthritis 10Temporal arteritis with aseptic necrosis I

Characteristics of crystals

It is impressive that there istremendous morphological variationamong apparent apatites we havefound in joints. Crystals may be almostpunctate, (Fig. 2), rods or thicker'boat-like' structures, (Fig. 4) or longneedles (Fig. 5); they may be denselyor loosely packed, and larger crystalsmay have the suggestion of an internalstructure at high magnification.Crystals may be rare or profuseenough to make the fluid milky. Rareclumps may obviously have differentclinical significance than massiveamounts. Crystals are frequentlyphagocytosed (Fig. 6) but seem mostoften to be ingested by mononuclearcells rather than polymorphonuclearneutrophils. Elemental analysis alsoshows some variation with someapparent apatite having Ca:P ratioswell below the expected 1 67:1,suggesting the presence of someamorphous or poorly crystallised salts.Substitution of chloride and carbonatein apatite also varies. As yet we knowof no good correlation of elementalanalysis findings with crystalmorphology.

Apatite was most often detected insynovial fluid because of itsaccessibility and from knees ratherthan other joints for the same reason.Arms, including finger joints, wrists,elbows, and shoulders seemedinvolved more often than the legs inpatients with collagen-vascular diseaseor those undergoing renal dialysis.Apatite was seen in surgical specimensof cartilage either as round clumpsnear to chondrocytes (but not clearlyin matrix vesicles) or as irregularmasses on the surface or in theinterstitium. Apatite in synovium wasfrequently phagocytosed or enfoldedby synovial cell processes.

Mechanisms involved in apatitedeposition disease

The ability of apatite to induce acuteinflammation in joints has been shownby intra-articular injection into theknees of dogs using the same model asthat used for urate and CPPD crystals.8Apatite has also producedinflammation at other sites-forexample, the pleural space andsubcutaneous tissue'°-afterexperimental injection. Apatite is not



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


I

Fig. 4 Rod and 'boat-like' apatite crystals. Some foamy internal str

seen at high magnification. Electron micrograph x125 000 (originmagnification).

Fig. 5 Long needle-like apatite crystals from joint fluid stained w

conjugated antibody to IgG demonstrating binding of this immuno

labelled by ferritin particles (arrows) to clump. Electron micrograj

(original magnification).

always associated with detectableinflammation (the same is true of urateand CPPD28 29) and the reasons for thisrequire study.

In TEM sections finely granularmaterial is generally associated withthe synovial fluid crystals (Fig. 6)8 and

with most clumpsembedded in synoviumcartilage. Preliminaelectron microscopicallaboratory show thglobulins are at least oi

of this granular materia

possibility that other materials, includ-ing fibronectin, osteonectin enzymes,and complement, are present shouldalso be considered. As previously sug-gested with urate crystals,' substancesadhering to crystals could encourageor inhibit inflammation. Effects couldeven vary depending on other cellularand humoral aspects of the joint.Complement may be activated by apa-tite as well as by other crystals.3'

Apatite almost certainly appears injoints by different mehanisms. In somecases (as with CPPD) it is thought to bea result of degenerative changes incartilage and to deposit first incartilage.'7 In other examples,especially when associated withmigratory calcific periarthritis at othersites, it probably deposits outsidecartilage due to the still unknownsystemic factors that seem to causemost such calcific periarthritis.Phosphate excess or other features ofrenal failure and haemodialysis seemto contribute to some cases.'9 Serumand synovial fluid calcium and

ructur canbe phosphate concentrations have beenal normal in most cases. Occasionally

apatite is seen to have risen fromfragments of bone released intoseverely destroyed joints.

Apatite deposition in articularcartilage occurs spontaneously inaging rabbits or after administration ofvitamin D; synovial calcification withapatite can be induced byintra-articular calciphylaxis.3' Inneither of these established types ofexperimental apatite calcification hasacute inflammation occurred. Longterm studies are still needed todetermine if osteoarthritis evolves or ifvarious modifications can produceattacks of acute arthritis.

Treatment

Acute inflammation associated withapatite deposition generally seems to

ith ferritin respond to treatment with)globulin indomethacin, other non-steroidalph x80 000 agents, or colchicine.' Aspirin has

been ineffective in some cases.Intra-articular depot corticosteroidshave also appeared to be helpful.' We

of crystals have no personal experience withor superficial measures to prevent or decreasery immuno- apatite deposition. Disodiumstudies in our etidronate has been reported toat immuno- decrease calcinosis but with chronicne component use may also effect bone(Fig. 5). The mineralisation.32 In patients with

I. ..1"4

.;, N:"g.- ..k' ".. ...;t ,

..J.

411V



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Suppl p 58 Annals of the Rheumcatic Diseases

Fig. 6 Apatite crystals embedded iet electron dense material in a vacuole of a

synovial fluid mononuclear cell. Electron micrograph x32 000 (originaltnagnification).

abnormal calcium or phosphorusmetabolism as, for example, occurs in

many of those with renal failure, bettercontrol of phosphorus concentrationsmay decrease apatite deposits. Somecalcinosis in or around joints has alsobeen reported to resolve

spontaneously. In one patient of ours

with a scleroderma-like disease apatitedeposits in finger joints resolved whenjoints gradually stitfened withprogressive skin disease .2

Indomethacin was thought to decreasepara-articular ectopic ossification inthree cases after total hiparthroplasty.33 Neither this nor otherpossible ways of decreasing calcinosishave been adequately andsystematically evaluated.

Other calcium-containing crystals

Other calcium-containing crystalshave been described in intra-articularsites. Brushite (CaHPO4 2HlO) hasbeen reported in small amounts ofuncertain clinical importance and

usually in association with CPPD."4Whitlockite and octacalciumphosphate may also occur in diseasedjoints along with apatite. {'

Recently, we have found calciumoxalate crystals in synovial effusions in

three patients undergoing haemo-dialysis for chronic renal failure.37 Inone patient x-ray films showed fingercalcifications similar to those seen withapatite and knee chondrocalcinosislike that seen with CPPD deposition.Detailed analyses at necropsy revealedmassive oxalate deposits and no

CPPD. The list of crystals identifiablein joints is almost certainly not yetcomplete. We and others have seenother unidentified birefringent or

electron dense material that is stillunder evaluation.

Supported in part by grants from theVeterans Administration. NationalInstitutes of Health (1-R21-AG 02815-01),McCabe Foundation, and Barsumian Fund.We thank Miss Gilda Clayburne, Miss

Marie Sieck, and Mrs Susan Rothfuss forexcellent technical help on the work from

our laboratory and appreciate the carefultyping of Miss Mary Ellen Maguire.

References

1 Codman E A. The shloul(ler. Boston:Thomas Todd, 1934.

2 Pinals R S, Short C L. Calcificperiarthritis involving multiple sites.Arthritis Rheurn 1966; 9: 566-74.

3 McCarty D J, Gatter R A. Recurrentacute inflammation associated withfocal apatite crystal deposition.ArthritisRheum 1966; 9: 804-19.

4 Amor B, Cherot A, Delbarre F. Lerhumatisme a hydroxyapatite. RevRhum Mal Osteoartic 1977; 44: 301-8.

5 Fam A G, Pritzker K P H, Stein J L,Houpt J B, Little A H. Apatite associ-ated arthropathy: a clinical study of 14cases and 2 patients with calcific bursitis.J Rheuramtol 1979; 6: 461-71.

6 McCarty D J, Hogan J M, Gatter R A,Grossman M. Studies or pathologicalcalcifications in human cartilage. 1. Pre-valence and types of crystal deposits inthe menisci of 215 cadavera. J BonieJoinit Surg [Am] 1966; 48: 309-25.

7 Schumacher H R. Pathology of thesvnovial membrane in gout. Light andelectron microscopic studies.Interpretation of crystals in electronmicrographs. Arthritis Rheuoti 1975; 18:771-82.

8 Schumacher H R, Somlvo A P. Tse R L,Maurer K. Arthritis associated withapatite crystals. Anltiterni Med 1 977;87: 411-6.

9 Paul H, Reginato A J. Schumacher H R.Alizarin red S-staining as a screeningtest to detect calcium compounds inssnovial fluid. Arthritis Rheuri (inpress).

1(1 Dieppe P A. Crocker P, Huskissonl E C,Willoughby D A. Apatite depositiondisease. A new athropathy. ILaicet1976; i: 266-9.

11 Paul H, Reginato A J, Schumacher H R.Morphological characteristics ofmonosodium urate: a transmissionelectron microscopic study of intactnatural and synthetic crystals. AntRheum Dis (in press).

12 Halverson P B, McCarty D J.Identification of hydroxyapatite crystalsin synovial fluid. Arthritis Rheutn 1979;22: 389-95.

1 3 Schumacher H R, Miller J L, LudivicoC, Jessar R A. Erosive arthritis associ-ated with apatite crystal deposition.Arthritis Rheuni 1981; 24: 31-7.

14 Schumacher H R, Gibilisco P, ReginatoA J, Cherian V, Gordon G V.Implications of crvstal deposition inosteoarthritis. J Rheuttatol (in press).

15 Schumacher H R, Gordon G. Paul H,Reginato A, Villanueva T, Cherian V,Gibilisco P. Osteoarthritis, crystaldeposition and inflammation. SeminArthritis Rheum 1981; 11: 116-9.

ne

%H:n



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


16 Dieppe P A, Crocker P R, Corke C F,Doyle D V, Huskisson E C, WilloughbyD A. Synovial fluid crystals. Q J Med1979; 48: 533-53.

17 Ali S Y. Matrix vesicles and apatitenodules in arthritic cartilage. In:Willoughby D A, Giroud J P, Velo G P,eds. Perspectives in inflammation.Baltimore, MD: University Park Press,1977: 211-23.

18 Halverson P B, Cheung H S, McCartyD J, Garancis J C, Mandel N. "Mil-waukee Shoulder". Association ofmicrospheroids containing hydroxy-apatite crystals, active collagenase andneutral protease with rotator cuffdefects. II. Synovial fluid studies. Arth-ritis Rheum 1981; 24: 474-83.

19 Moskowitz R W, Vertes V, Schwartz A,Marshall G, Friedman B. Crystalinduced inflammation associated withchronic renal failure and hemodialysis.Am J Med 1969; 47: 450-60.

20 Doyle D V, Dieppe P A, Crocker P R,Ibe K, Willoughby D A. Mixed crystaldeposition in an osteoarthritic joint. JPathol 1977; 123: 1-4.

21 Schumacher H R. Articular cartilage inthe degenerative arthropathy ofhemochromatosis. Arthritis Rheum (inpress).

22 Martel W, McCarter D K, Solsky M A,et al. Further observations on thearthropathy of calcium pyrophosphatecrystal deposition disease. Radiology1981; 141: 1-15.

23 Reginato A, Schumacher H R. Synovialcalcification in a patient with collagenvascular disease: light and electronmicroscopic studies. J Rheumatol 1977;4: 261-7.

24 Brandt K D, Krey P R. Chalky jointeffusion. The result of massive synovialdeposition of calcium apatite inprogressive systemic sclerosis. ArthritisRheum 1977; 20: 792-6.

25 Schumacher H R, Schimmer B, GordonG V, et al. Articular manifestations ofpolymyositis and dermatomyositis. AmJ Med 1979; 67: 287-92.

26 Kieff E D, McCarty D J. Hypertrophicpulmonary osteoarthropathy witharthritis and synovial calcification in apatient with alcoholic cirrhosis. ArthritisRheum 1969; 12: 261-71.

27 Reginato A J, Paul H, Schumacher H R.Hydroxyapatite crystals in rheumatoidarthritis synovial fluid. Clinical Research1982; 30: 662A.

2 8 Agudelo C A, Weinberger A,Schumacher H R, Turner R, Molina J.Definitive diagnosis of gout byidentification of urate crystals inasymptomatic metatarsophalangealjoints. Arthritis Rheum 1979; 22:559-60.

29 Schumacher H R. Pathogenesis ofcrystal-induced arthritis. Clinics inRheumatic Diseases 1977; 3: 105-31.

30 Hasselbacher P. C3 activation bymonosodium urate monohydrate andother crystalline material. ArthritisRheum 1979; 22: 571-8.

31 Reginato A J, Schumacher H R,Brighton C T. Experimentalhydroxyapatite synovial and articularcartilage calcification. Light andelectron microscopic studies. ArthritisRheum (in press).

32 Weinstein R S. Focal mineralizationdefect during disodium etidronatetreatment of calcinosis. Calif Tissue Int1982; 34: 224-8.

33 Ritter M A, Gioe T J. The effect ofindomethacin on para-articular ectopicossification following total hiparthroplasty. Clin Orthop 1982; 167:113-7.

34 Moskowitz R W, Harris B K, SchwartzA, et al. Chronic synovitis as amanifestation of calcium crystaldeposition disease. Arthritis Rheum1971; 14: 109-16.

35 Gaucher A, Faure G, Netter P, et al.Identification des cristaux observes dansles arthropathies destructrices de lachondrocalcinose. Rev Rhum MalOsteoartic 1977; 44: 407-14.

36 McCarty D J, Lehr J R, Halverson P B.Crystal populations in human synovialfluid. Identification of apatite,octacalcium phosphate and tricalciumphosphate [Abstract]. Clinical Research1982; 30: 807.

37 Hoffmnan G, Schumacher H R, Paul H,et al. Calcium oxalate microcrystallineassociated arthritis in end stage renaldisease. Ann Intern Med 1982; 97:36-42.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Laboratory handling of crystalsPAUL DIEPPE, JUNE HORNBY, ANGELA SWAN,CHARLES HLUTTON, ALAN PREECE

Fronm the University Departnmetnt of Medicitne, Bristol Royal Infirmary, Bristol BS2 8HW

Introduction

1 he three main phases of matter aresolid. liquid, and gas. Most solids arecrystalline. Crystals are characterisedby their ordered arrangement ofmolecules; each unit takes up the moststable position possible in relation toits neighbours, resulting in repetitionof the three-dimensional structurethroughout the substance-that is,internal symmetry. Internal symmetryresults in predictable effects ontransmitted electromagnetic rays,which aids identification of crystalsand may give the particles colour andsparkle. The close packing of themolecules confers hardness andstability. Crystals are often thought ofas solid, inert, ageless particles,impervious to physical force.

Crystals found in joint tissue formin vitro' via a phase change-fromliquid to solid. Their formation anddissolution can often be affected byminor changes in physiologicalconditions because the soluteconcentrations are in the metastableregion-that is, above saturationpoint, but below that at which crystalforniation is inevitable. Biologicalcrystals are usually small (less than 20,m long), and therefore have a highsurface area to weight ratio. Thismeans that dissolution can occurrapidly. Their large surfaces are oftencharged or exhibit atomic roughnessand 'dangling bonds"--that is,component atoms stick out from thesurface. These surface properties mayresult in attachment of various othersubstances found in joint tissue or inthe laboratory. Crystals related tojoint disease are therefore far frombeing inert and stable. They can formand dissolve rapidly (dissolution isgenerally easier and faster thancrystallisation) and attract other thingsto their surfaces. They are also

Correspondence to: Dr P A Dieppe.

susceptible to many physical forcesapplied during laboratory procedures.

Precautions have to be taken in thehandling of specimens from joint andof those made in the laboratory ifartefacts and false results are to beavoided.

Crystals in specimens of joint tissue orfluid

Examination of synovial fluids andtissue samples for crystals is now aroutine diagnostic procedure. Thereport-crystals 'present' or'absent'-is usually regarded asdefinitive. However, both falsepositive and false negative results mayoccur for a variety of reasons (Table1).

SE NOVIAL FLUIDAspiration of synovial fluid involvespassing a needle through skin,subcutaneous tissue, joint capsule, and

Table 1 Some causes offalse positiveand false negative reslults wheni synovialfluids are examieled for the presence ofcrvstals

False positive:(1) Joint puncture introduces crystals

into fluid from previously stabletissue deposit

(2) Crystals form in the fluid afteraspiration due to changes intemperature or pH

(3) Non-crystalline material orcontaminants mistaken for jointcrystals

False negative:(1) Sampling error, or crystals not seen

due to poor technique(2) Dissolution of crystals after

aspiration(3) Crystals too small to be seen by light

microscopy

sy'novium; it may strike the cartilage.Crystals could be dislodged from anyof these sites during the procedure andseen in the fluid when they were notthere before aspiration.Once the fluid is out of the joint it

changes; this may result in formationor dissolution of crystals. The two mostimportant factors are probably the pHand temperature of the fluid.Temperature changes are

particularly important in uratecrystallisation. Reardon and Scott havereported that raising synovial fluidtemperatures from 32°C to 37°Creduces crystal concentration-afinding that may have relevance to thegout attack in vivo, as well as to crystalidentification in vitro.' Other authorshave reported an increased yield of'positive results' if synovial fluids arecooled.2 This is hardly surprising inviews of the rapid decrease in solubilitvas temperature falls. Recent reports ofurate crystals in asymptomatic joints'should be viewed critically, and resultsmav, not be valid if fluids have beenstored in the cold before examination.

Changes in pH are produced by lossof CO2 after fluid aspiration. This may'be reduced by covering samples withoil to avoid contact with the air. Theincrease in pH could favourcrystallisation of various calciumphosphates, and we have observed invitro formation of brushite crystals influids left to stand on the bench for 48hours. Crystals which form after fluidaspiration may take on a star-shapedconfiguration (Fig. 1). Changes in pHalso affect urate solubility; correctionof gouty synovial fluid pH to 7-4 hasbeen reported to result in increasedcrystal formation.4

Alterations in the crystal load afterfluid aspiration may help explain thelack of correlation between crystalnumbers and inflammation, and theoccasional false positive or falsenegative result. Care must be taken inhandling fluids, and early examination



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Laboratory handling ofcrystals Suppl p 61

Fig. 1. Crystals formed in synovialfluid left for 48 hours on laboratorybench. Initial examination gavenegative results. Note star-shapedclusters ofcrystals which formed invitro. They contained calcium andphosphate and were probably brushite.Polarised light microscopy x 100(original magnification).

with minimisation of pH andtemperature changes is recommended.

CELLS AND TISSUE SAMPLES

Crystal phagocytosis is regarded as a

central factor in the inflammatoryresponse to particles (P Platt and W CDick, p. 4). Uptake of crystals isoften reported on specimens examinedby light microscopy for diagnosticpurposes, and has been afforded somesignificance. In our experience it isalmost impossible to differentiatebetween cells with crystals attached totheir surface and those with crystalsinside them. Examination by scanningelectron microscopy (SEM) and thicksection scanning transmission electronmicroscopy (STEM) suggests thatattachment without internalisation iscommon.

Tissue samples (usually synovialbiopsy specimens) need carefulhandling to preserve crystal deposits.Fixatives and stains should be bufferedto avoid crystal dissolution due to pHchanges. Tissue sectioning may resultin crystals dropping out of the sample;this is particularly common in thinsectioning for electron microscopy.Many of the published data includephotographs showing holes wherecrystals might once have been (Fig. 2).Thick sectioning and high voltage EMcan reduce this difficulty.5The increasing use of the electron

microscope to study crystal depositiondiseases has highlighted two otherproblems: (a) sampling errors are a

major problem, as only a tiny piece oftissue can be examined, and crystal

A

Fig. 2 Transmission electronmicrograph ofphagocyte with a cleft incell. This is typical ofpublished picturesofcrystal-cell interactions. A crystalmay have occupied the cleft and fallenout during thin sectioning ofthesample. It is impossible to say whatprinciple (if any) was there. Thicksectioning techniques sometimesovercome this problem. Transmissionelectron microscopy x 1500 (originalmagnification).

deposits are often patchy; (b) thevariability in the morphology andtypes of crystal sometimes found in a

single specimen presents difficulties inidentification. Heterogeneity in crystaldeposits makes confident statementsabout the species and size of crystalfound in a sample seem naive.

Crystals used for experimental work

It is difficult to extract large quantitiesof a pure crystalline preparation frombiological samples. Processes requiredto remove organic material may ruinthe crystals, and yields are generallysmall. Most experimental worktherefore uses crystals manufacturedin vitro.

In most laboratories (includingours) a variety of techniques have beenapplied to the preparations in order toobtain a batch of sterile particles of theright size and in a form that is easy touse. Many of these techniques are nowknown to alter the crystals, therebyinvalidating many of the publisheddata on the biological effects ofcrystals. Crystals age and are

susceptible to heating and grinding;they also 'pick up' extraneous matterfrom the environment, changing theiractive surfaces. Furthermore, no twobatches are ever quite the same.

H EATVan Armen suggested thatcrystal-induced inflammation might bemediated by attachment of bacterialpyrogens to the surface.6 Most workershave therefore heated crystalpreparations to remove pyrogen(usually to about 1 80°C for threehours). Most of the crystals in questionare hydrated, and there is extensiveevidence to show that the water ofcrystallisation is lost, and that thecrystal lattice and surface charge areboth altered extensively by heating.7 8It has recently been shown thatbacterial lipopolysaccharide coatingneeds to be very extensive to altercrystal reactivity. If care is taken toavoid contamination, heating isprobably unnecessary.9

GRINDINGGrinding crystals to reduce their sizealso alters their inflammatorypotential. This is probably related tothe change in surface charge whichalso occurs on grinding (Table 2). Themechanism for this is unclear.

AGING OF CRYSTALPREPARATIONSWe have regularly observed a changewith time in the biological activity of abatch of crystals. Preparations ofmonosodium urate monohydrateusually lose their phlogistic activityslowly over a period of months.Brushite appears to be particularlysusceptible to aging, and some of itscellular effects increase strikingly withage (Fig. 3). Crystal lattices andsurfaces may change slowly, phasetransitions may occur-that is, changefrom one type of crystal toanother-and the crystal surfaces maypick up contaminants in thelaboratory. This aging phenomenondeserves more investigation, andmeans that great care must be used toobtain fresh 'age matched' samples forany comparative work.

BATCH VARIATIONComparisons cannot be made if twodifferent crystal preparations are used,because the activity of differentbatches of the same crystal is neveridentical. Thus preparations of uratecrystals made in the same way ondifferent days may look the same andhave the same characteristics onanalytical screening, but may affect



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Duration of storage (hours)

Fig. 3 Effect ofstoring brushitecrystals on their ability to lyse red cellmembranes. Fresh brushite caused20% haemolysis after four hours'incubation ofcrystals and red cells at3 70C. Same preparation caused 80%haemolysis after three months' storagein laboratory. This is a striking exampleof effect ofaging on crystal reactivity.Most other particles slowly lose activitywith age.

biological systems differently. Minorchanges in size, available surface area,and surface characteristics probablyaccount for this.Examination of most of the

published reports on crystal-inducedinflammation shows that sampleshave usually been heated or ground orboth and precautions to avoid batchvariation and aging have hardly everbeen taken. Different results fromdifferent centres are therefore easy toexplain. More importantly, much ofthe work and many of the theoriesmay be invalid.The two factors that may be most

important in determining the activityof a crystal are probably the availablesurface area, and surfacecontamination.

Experiments on surface area andsurface contamination of crystals

Weight-for-weight comparisons ofcrystals are meaningless because ofdifferences in size and shape, andtherefore of the available surface tointeract with cells and proteins.10Most recent work on

crystal-induced inflammation hasemphasised the role of absorbedproteins to crystal surfaces, especiallyIgG absorption resulting in Fc

Fig. 4 Electrophoresis ofmonosodium urate monohydratecrystals before and after incubationwith Fab fragments ofhuman IgG.Initial reading (top) shows that crystalsare travelling rapidly to positive pole ofan electrical field. After being coatedwith Fab fragments, mobility is muchreduced (bottom). This reduction inmobility-that is, charge-may bereversed by washingprotein offcrystal.Fab-coated crystals have a reducedcapacity to lyse red cell membranes(Measurements recorded on laserDoppler shift-axis represents theamount ofshift and thus crystalmobility).

receptor activation on cells andcomplement activation.'1 12We have examined the surface

properties of crystals and the changeswith protein coating by measuringelectrophoretic mobility using a laserDoppler shift technique describedelsewhere.'3 This may provide furtherunderstanding of contamination of thecrystal surface and allow comparisonsof the surface area and affinity ofcrystals for proteins.

Crystals are suspended in a balancedsalt solution. A current is applied andthe shift to negative or positive polecan be measured. Most biologicalcrystals shift to the positive pole (theyhave a negative charge property).

Previous incubation of crystals with aprotein may result in attachment ofthe protein to the crystal, and a shiftof the charge property towards that ofthe protein. The strength of theattachment may be assessed bywashing the samples. We have recentlyshown that Fab fragments of IgGattach to urate crystals, inhibitingmembranolysis"4 (Fig. 4). Thisprovides a possible mechanism of selflimitation of gouty inflammation (viacleavage of IgG to Fc and Fabfragments; the more positivelycharged Fab fragment attaches to thecrystal, masking its active surface, andinhibiting its activity and ability tointeract with Fc receptors). Histone isanother protein avidly picked up byurate crystals (Fig. 5). The affinity ofthe particles for minute quantities ofthe protein illustrates the potential forlaboratory contamination of crystals.The charge shift may also provide dataon the available surface area.

It may be concluded from the datapresented that crystals can attract avariety of proteins, that smallquantities of surface contaminant canchange their surface charge andbiological activity, and thatelectrophoretic mobility can be used toinvestigate the available surface areaof a particle. (Further experiments areunder way, and full details of themethods used and data obtained are inpreparation for publicationelsewhere).

Su,rtoce charge of ura1e

Fig. 5 Electrophoretic mobility ofurate crystals incubated in mediumcontaining varying concentrations ofhistone. Note low concentration ofprotein required to shift chargeproperty, and saturation ofcrystalsnegatively charged sites at aconcentration ofabout 0 01 mg/ml.Minute amounts ofprotein can beshown to attach to, and change, crystalsurfaces in this way. Shape ofcurve is aguide to affinity and active surface areaof crystal.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Laboratory handling ofcrystals Suppl p 63

Conclusions

(1) False positive and false negativeresults may occur when synovial fluidsand tissue samples are examined forthe presence of crystals. These errorsmay be minimised by careful handlingof the samples.

(2) Many of the techniques used inthe handling of crystals in thelaboratory alter the surface of theparticle and affect its biologicalactivity. This factor has not beenaccounted for in most experiments oncrystal-induced inflammation, and itinvalidates much of the publishedwork.

(3) Quantitative assessment of theaffinity of crystals for proteins may beused to assess the surface area ofbiological significance.

We would like to thank the Arthritis andRheumatism Council for financial support.

References

1 Reardon J A, Scott J T. Resolution ofacute gout attacks: a possible

mechanism. Ann Rheum Dis 1980; 39:189.

2 Font F, Goldman J, Toro R. Mono-sodium urate crystals (MSUC) andsperulites (SP) in symptomatic metatar-sophalangeal (MTP) [Abstract] joints.Arthritis Rheum 1982; 25, suppl: 53.

3 Agudelo C A, Weinberger A,Schumacher H R, Tumer R, Molina J.Definitive diagnosis of gout byidentification of urate crystals inasymptomatic metatarsophalangealjoints. Arthritis Rheum 1979; 22:559-60.

4 Tiliakos N A, Wilson C H. Thenon-crystallic gouty arthritis [Abstract].Arthritis Rheum 1982; 25, suppl: 54.

5 Crocker P R, Dieppe P A, Tyler G,Chapman S K, Willoughby D A. Theidentification of particulate matter inbiological tissues and fluids. J Pathol1977; 121: 37-40.

6 Van Armen C G, Carlson R F, Kling P J,Allen D J, Bondi J V. Experimentalgouty synovitis caused by bacterialendotoxins absorbed onto uratecrystals. Arthritis Rheum 1974; 17:439-49.

7 Mandel N S. Structural changes insodium urate crystals on heating.Arthritis Rheum 1980; 23: 772-6.

8 Dieppe P A, Doherty M, Swan A, et al.

Changes in monosodium urate crystalson heating or grinding. Arthritis Rheum1981; 24: 975-6.

9 Hasselbacher P. Crystal-proteininteractions in crystal-induced arthritis.In: Weissman G, ed. Advances ininflammation research. Vol. 4. NewYork: Raven Press, 1982: 25-44.

10 Dunn C J, Doyle D V, Willoughby D A.Experimental methods in the study ofcrystal deposition diseases. EuropeanJoumalofRheumatology and Inflamma-tion 1978; 1: 135-41.

11 Kozin F, McCarty D J. Protein bindingsto monosodium urate monohydrate,calcium pyrophosphate dihydrate andsilica dioxide crystals. I. Physicalcharacteristics. J Lab Clin Med 1979;89: 1314-25.

12 Ginsburg M H, Kozin F. Mechanisms ofcellular interaction with monosodiumurate crystals. IgG-dependent andIgG-independent platelet stimulationby urate crystals. Arthritis Rheum 1978;21: 896-903.

13 Preece A W, Sabolovic D. Cell elec-trophoresis: clinical applications andmethodology. New York:Elsevier/North Holland, 1978.

14 Dieppe P A, Swan A, Hornby J, JenkinsR, Luckman N P, Preece A W. Crystalsurface charge and inflammation. AnnRheum Dis 1980; 39: 606.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Polarised light microscopy: principles and practice forthe rheumatologist

J CHAYEN

From the Division of Cellular Biology, Kennedy Institute ofRheumatology, Hammersmith, London W6 7DW

Polarised light microscopy is used todetect and measure matter (in fluids,tissues, or cells) that has crystalline or

crystalline-like properties. Becauselight has electromagnetic properties, itcan be influenced by theelectromagnetic properties of materialthrough .which it passes. In many

materials this influence is equal inwhatever direction the light isvibrating. But crystalline materialspecifically influences the speed oflight that passes either along or across

it. Variation in the speed of light in oneor other direction cannot be detectedwith normal illumination, in which thelight is vibrating in all directionsperpendicular to the light path, butonly with plane-polarised light-thatis, with light in which all the vibrationshave been filtered out except for one

direction.

Birefringence and refractive index

The speed with which plane-polarisedlight moves in a material is measuredby the refractive index of thatmaterial-that is, refractiveindex=velocity of light in air/velocityof light in the material.

Crystals in particular have threeoptical axes. These usually correspondto the length, breadth, and depth of thematerial. In microscopicalexamination we are mainly concernedwith only the first two of thesedimensions. For most materials therefractive index in one direction-forexample, along the length of thematerial-is the same as in any otherdirection. Such materials are isotropic.However, the refractive index alongthe length of most types of crystals isdifferent from that in one or both otheraxes-that is, light passes more rapidlyor less rapidly when vibrating alongthe long axis than when vibrating alongthe shorter axis. Such material is

anisotropic and shows the property ofbirefringence-that is, two refractiveindices. Thus birefringence is definedas the difference between therefractive index measured with lightparallel to the long axis ( nd ) and withlight vibrating perpendicular to thelong axis (rn)-that is, n1j -n, . But whatwe actually measure with thepolarising microscope is the opticalpath difference in our

material-namely, the interaction ofboth refractive indices that causes a

change in the velocity of the light; thisis influenced by the thickness (t) of thespecimen.

Optical path difference=(njj-nL)tor birefringence=optical pathdifference/thickness.Consequently, although the

birefringence of a material is a physicalconstant for that material, a thickercrystal will appear brighter than a

thinner crystal of the same materialbecause what we observe is the opticalpath difference and this is affected bythickness.

TYPES OF BIREFRINGENCE

There are three main types ofbirefringence.

(1) Intrinsic or crystallinebirefringence-This is determined bythe chemical structure of the materiaLespecially when the material assumes a

crystalline form. Because it dependsdirectly on the electromagneticcharacteristics of the material, it is notrelated basically to the medium inwhich the material occurs or in which itis mounted for examination.

( 2) Form or texturalbirefringence-Structures such as

collagen fibres which contain orientedelongated submicroscopic particlesshow form birefringence even if theseparticles have little, or perhaps no,intrinsic birefringence. This form ofbirefringence depends on the different

velocity of light in the array of particles(or fibrils) and the velocity in thematerial in which they are mounted.Consequently, this type ofbirefringence is dependent on therefractive index of the mountingmedium; it may be differentiated fromcrystalline birefringence on this basis.

(3) Strain or flowbirefringence-Certain materials thatare virtually non-birefringent maybecome birefringent if they are madeto flow along a constricted pathwaythat causes the material to becomeorientated. To a large extent this isakin to form birefringence. It isimportant to rheumatologicalmicroscopists mainly because theglass, used for producing the lenses ofa polarising microscope must be freefrom strain; otherwise the lenses willchange the vibration of the light so thatthe field will never become completelyblack.

The polarised light microscope

A polarised light microscope should bea good optical microscope withstain-free objectives and with thefollowing additional components.

(1) The polariser, situated below thecondenser, that can be rotated to giveplane-polarised light vibrating solelyin the east-west (E-W) direction. (Inmost microscopes this seems to fitmost closely to the Brewster angle"2).

(2) A rotatable stage, so that thespecimen can be turned to any anglerelative to this E-W direction.

(3) A rotatable analyser set betweenthe objective and the eyepiece. This isa polar and transmits light that vibratesin one direction only. When it isrotated to the north-south direction(N-S), no light should be seen in anempty field. For precise work it is helpfulif this analyser can be turned through360° (as in the work of Watts et al.').



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Polarised light microscopy Suppl p 65

(4) A slot, cut at 450 (to the E-Wposition) in the microscope tube,between the objective and theanalyser.

(5) A quartz red plate that may beinserted into this slot. Its use will bediscussed later.

(6) A X/4 (a quarter wavelength)plate that may be inserted into the slotin place of the quartz red plate. It isused for measuring birefringence(optical path difference) by S&narmontcompensation.' 2 4

(7) Crossed lines in the eyepiece,that define the E-W or the 450position, are also useful.The objectives should be made of

strain-free glass. They should also beable to be centred because it isessential that there is no obliqueillumination that can give rise toreflections within the microscopesystem with consequent depolarisationof the light beam. For this reason thediaphragms should be kept to theminimum size compatible withadequate illumination. Especiallywhen dealing with crystals of lowbirefringence, such as calciumpyrophosphate or apatite, it should berecognised that the intensity of thelight decreases with the square of themagnification (at constant numericalaperture) so that low magnificationobjectives are to favoured. Mypreference is for the following: x 16,0 32NA; x25, O6NA; and x40;0 85NA; all able to be centred.

MECHANISMThe theory of polarised lightmicroscopy can be quite complex.' 2 4However, a reasonably simple and selfconsistent explanation may be givenbased on the vector treatment of light(Fig. 1).The light passing through the

polariser is plane-polarised so that itvibrates solely in the E-W direction.Consequently, it has vectors in the 450positions, as shown in Fig. 1, but it hasno vector in the N-S direction; thus itwill not pass the analyser, set to theN-S position, and the field will appearblack.We now place in the field a

birefringent object in which lightvibrates more rapidly along its longaxis than its short axis. When set to the450 position it will be maximallybright. This is because one of the 450vectors of the E-W light will be

W4 ) E

I1Fig. 1 Schematic representation ofhow a birefringent object (doublerectangle), placed at 450 position,appears bright under crossed polars.Polariser (polarising polar), shown asa

triple rectangle, is set toE-W position;analyser (multiple rectangle) iscrossed-that is, set to N-S position.Direction oflight is shown by longarrow. Vectorial representation ofthevibration oflight is shown on right offigure. Vectors are indicated by brokenlines; resultant vibration is shown as a

solid line.

moving along this long (fast) axis. Theother 450 vector will be retarded.The vibration of the resultant light

will be turned towards the N-Saxis-that is, the object will haveturned the plane of polarisation. Thisresultant vibration will now have avector that will be in the N-S direction.This will pass through the analyser thatis set to the N-S direction and theobject will appear bright, with thebackground remaining black.The maximum resolution of the light

microscope is about 0 25 ,um.4 Buthere we are not concerned withresolution so much as with detection,which is obviously related solely tohow much brighter the object is thanits background. Hence minuteparticles or crystals, much smaller than0 25 ,um diameter, can be detected bypolarised light microscopy, providedthat they are sufficiently birefringentto make them appear bright, eventhough they cannot be resolved.The theory of polarised light

microscopy involves the interferencebetween the vectors of the light (or theordinary and extraordinary rays). Ifyou have two vectors, or two rays ofthe same light, that are moving slightlyout of phase with one another, theresultant light will be a combination ofboth rays. For example, if they are half

a wavelength out of phase, the waveform of one will be maximally positivewhen that of the other is maximally,and equally, negative. When these

,-- waves are combined they will interfere-- - - - to give no light, each cancelling out the

other. This phenomenon ofinterference is used in relation to thequartz red plate, used for defining thesign of birefringence (as considered

v' later).

Characteristics of crystals

Crystals can be defined on the basis ofthe following criteria:

(1) Habit. This includes shape anddimensions.

(2) Extinction angle. For manyclasses of crystals, when the long axisof the crystal is set to the N-S or E-Wposition the crystal shows nobirefringence; it shows maximumbrightness when positioned at 450 tothese axes (for reasons discussed inrelation to Fig. 1). Such a crystal,which is blackest when set to the N-Sor E-W position shows straightextinction. Some crystals do notbecome darkest at precisely the N-Sposition, but at some angle to thisdirection. This angle is the extinctionangle and may be related to whetherthe three axes of the unit crystal (themolecular crystal-structure) areexactly at right angles or not. Thisextinction angle can be very useful indefining the nature of the crystallinematerial. For example, needle-likecrystals of monosodium uratemonohydrate, which occur commonlyin gout, extinguish at 120 from thestraight (N-S) position-that is,they show almost straightextinction-whereas allopurinolcrystals extinguish at 300 from one ofthese axes.

(3) Sign ofbirefringence. This will bediscussed below.

(4) Refractive index, or refractiveindices along the different axes.

(5) Symmetry.The use of all these factors allows

fairly precise analysis of crystals foundin biopsy specimens. They were used3to identify crystals of uric aciddihydrate, monosodium uratemonohydrate,- and anhydrous uric acidin muscle of patients with gout whowere either untreated or treated withprobenecid; crystals of xanthine,hypoxanthine and oxipurinol, as wellas the previous three substances, were



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


identified in muscle of the patientswho had been treated with allopurinol.

Refractive index and symmetry areof particular value for such studies butare not likely to be used in morerout ine rhe umatologicalexaminations; they have beendiscussed in detail by Watts et al.3

SIGN OF BIREFRINGENCEThis is a major parameter for routineinvestigations. It defines whether theoptically slow axis is along the long orthe short geometric axis of the crystal.Thus it allows rapid distinctionbetween crystals of monosodium uratemonohydrate and calciumpyrophosphate dihydrate which,especially when engulfed in apolymorph, have similar habits.To determine the sign of

birefringence, the crystal is orientatedwith its long axis at the 45° positionthat intersects the N-E angle.Provided that the crystal showsstraight extinction (which is virtuallytrue of both of these types of crystals),it will appear maximally bright on ablack background. The quartz redplate is then inserted into the slot in themicroscope tube. In this position theslow axis of the wedge (marked on thecasing of the wedge as the y-direction)is in the south-west to north-east(SW-NE) direction. Thus the long axisof the crystal is lying along the slowaxis of the wedge.The quartz red plate is cut so that the

vectors (or rays) interfere with oneanother to give first order red light inNewton's scale. Consequently thehitherto black field now appears red.The interference colours shown bythe crystal will depend on whether thegeometric long axis corresponds to theoptically slow, or optically fast, axis.Thus the retardation of the light (R),induced by the plate, will either befurther retarded, or less retarded, andthis will be reflected in a change in theNewton's order of colours5:

R total = R plate + R specimen

This means4 that the slow axis of thecrystal is along its geometric long axis.It has added to-that is,increased-the retardation of the lightin the SW-NE (450) position, soincreasing the resultant interferencewavelength to give a blue colour. Sucha crystal is called a positive crystal; it ispositively birefringent. This is typical

of crystals of calcium pyrophosphatedihydrate.6

Obviously, if the crystal is rotated tothe NW-SE position, there will bereversal of colour in the crystalbecause now the 'fast' axis will beparallel to the slow axis of the wedge.

If the slow axis of the crystal is atright angles to its geometric longaxis-that is, its fast axis is along thelong axis of the crystal-the reversepertains:

R total = R plate -R specimen

In this case the crystal (in the 450position) decreases the retardationimposed by the plate so that thewavelength of the interference colouris decreased: the colour changes fromthe red, imposed by the plate, toyellow. Such crystals show negativebirefringence, as is characteristic ofcrystals of monosodium uratemonohydrate.'

POTENTIAL PITFALLA serious problem may beencountered when examining fluidsfor crystals that are expected to havevery weak birefringence. If no crystalscan be discerned in the wetpreparation of the fluid, the specimenmay be dried so that any crystallinematerials present can be seen more bythe fact that they scatter light than bytheir weak birefringence. However,this procedure could permitcrystallisation of material that waspresent only in solution in the fluid. Toprove that the crystals seen in thedried preparation could have existedin the original, native fluid, the driedcrystalline material must beremounted, either in some of theoriginal fluid, or at least in a medium ofidentical refractive index. If thesecrystals can still be seen whenremounted but cannot be detected inthe native fluid, it must be suspectedthat their appearance in the driedpreparation is due to crystallisationduring the drying process.

Some applications of polarised lightmicroscopy

Following the demonstration of uratecrystals in synovial fluids from goutyjoints7 and of a different type of crystal(now recognised as consisting of cal-cium pyrophosphate dihydrate) in the

synovial fluid from acutely inflamedjoints in chondrocalcinosis,8 the mainroutine use of polarisation microscopyis the detection of these crystals andtheir discrimination; as discussedabove, this is effected simply bymeasuring the sign of birefringence.Pyrophosphate crystals have also beendetected by polarised light in the syno-vial fluid from some patients withosteoarthritis9 and shown by histologi-cal stains, in tissue from patients withthis condition.'0

Xanthine and hypoxanthine crystalswere found in the muscle from twopatients with xanthinuria;`t theoccurrence of this material was laterconfirmed by high resolution massspectrometry."2 Various urates andother crystalline material wereidentified by more complex polarisedlight analysis3 in the muscle of patientswith gout.

Very recently relatively largecrystals, apparently of hydroxyapatite,have been found in the bone at the siteof fracture in fractures of the neck ofthe femur in the elderly."' It has beensuggested that the change frommicrocrystalline to relativelymacrocrystalline hydroxyapatite couldbe a factor in the tendency of this boneto fracture, rather as, analogously,relatively large crystals may beinvolved in fatigue-fractures of metals.

General support from the Arthritis andRheumatism Council for Research isgratefully acknowledged.

References

1 Hallimond A F. The polarizingmicroscope. Vickers Instruments, 1970.

2 Hartshome N H, Stuart A. Crystals andthe polarising microscope. London:Edward Arnold, 1970.

3 Watts R W E, Scott J T, Chalmers R A,Bitensky L, Chayen J. Microscopicstudies on skeletal muscle in goutpatients treated with allopurinol. Q JMed 1971; 40: 1-14.

4 Chayen J, Denby E F. Biophysicaltechnique as applied to cell biology.London: Methuen, 1968.

5 Hartshome N H, Stuart A. Practicaloptical crystallography. London:Edward Amold, 1964: 179.

6 Currey H L F. Investigation of jointfluids. Proc R Soc Med 1968; 61:969-971.

7 Hollander J L, ed. Arthritis and AlliedConditions. Philadelphia: Lea andFebiger, 1960: 77.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Polarised light microscopy Suppl p 67

8 McCarty D J Jr, Kohn N N, Faires J S.The significance of calcium phosphatecrystals in the synovial fluid of arthriticpatients: the 'pseudogout syndrome'.Ann Intern Med 1962; 56: 711-37.

9 Huskisson E C, Dieppe P A, Tucker AK, Cannell L B. Another look atosteoarthritis. Ann Rheum Dis 1979;38: 423-8.

10 Doyle D V. Tissue calcification and

inflammation in osteoarthritis. J Pathol1982; 136: 199-216.

11 Chalmers R A, Watts R W E, Pallis C,Bitensky L, Chayen J. Crystallinedeposits in striped muscle inxanthinuria. Nature 1969; 221: 170-1.

12 Parker R, Snedden W, Watts R W E.The quantitative determination ofhypoxanthine and xanthine('oxypurines') in skeletal muscle from

two patients with congenital xanthineoxidase deficiency (xanthinuria).Biochem J 1970; 116: 317-8.

13 Kent G N, Dodds R A, Klenerman L,Bitensky L, Chayen J. Changes in theoptical properties of the inorganic andorganic bone matrix components infractured neck of femur. J Bone JointSurg [Br] 1983; 65: 187-94.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Application of physical methods in the investigationsof crystal-related arthropathiesJ. S. SHAH

From the H. H. Wills Physics Laboratory, University ofBristol

Introduction

The application of physical methods tounderstanding the role of crystals inpathogenesis of various arthropathiescan help in two ways. Firstly, inexploration of the nature of crystallinedeposits-for example, identificationof crystals--and, secondly, inunderstanding the very processes ofcrystal growth (and dissolution).The assessment of the crystals

mainly comprises:(1) Determination of chemical

composition and defects such asnon-stoichiometry and foreign phaseprecipitation.

(2) Identification of crystal structureand determination of crystallographicparameters.

(3) Visualisation of structuralimperfections.

(4) Examination of morphology fordetermination of conditions of crystalgrowth. (Finer structural details on acrystalline surface are often associatedwith the growth history of the crystal.)

Physical methods available forcrystal characterisation can be broadlydivided into two categories.

(1) Methods based on the interactionof crystals with electromagneticradiation-for example, visible light,infrared radiation, x-rays. These are:(a) polarisation light microscopy; (b)infrared absorption photometry; (c)x-ray diffraction methods, includingtopographic method; and (d) x-rayextended absorption free structure(EXAFS).

(2) Methods based onmaterial-electron interaction:(a) transmission electron microscopy(TEM); (b) scanning transmissionelectron microscopy (STEM); (c)analytical electron microscopy-energy dispersive x-ray analysis(EDX) and wavelength dispersiveanalysis (WDX); and (d) electrondiffraction.

Detailed description of each of themethods is beyond the scope of thispaper. Here I intend to deal brieflywith the subjects of infraredabsorption spectrometry, x-raydiffraction, EXAFS, and electronmicroscopies, and show how they havebeen applied to investigatearthropathic crystallisation. Whereappropriate, examples of otherbiological and non-biological crystalswill be cited.

Methods

INFRARED SPECTROMETRYVibrations of molecular bonds in amaterial cause characteristicabsorption of the infrared region of theelectromagnetic spectrum. Theabsorption for each bond occurs atspecific frequencies, which may revealthe nature of vibrations andstereochemistry of the moleculargroups in question. The characteristicpeaks may be used like fingerprints todetermine the presence of differentmolecular groups in a substance.

Absorption spectra of apatite andcalcium pyrophosphate crystals aredifferent. They may be used to detectthe presence or otherwise of thecompounds in arthropathic deposits.

In the case of hydroxyapatiteinfrared spectra may yield additionalinformation on crystallite size. Fig. 1shows absorption spectra of apatitefrom calcific periarthritis. The apatiteis finely divided. Its crystallite sizefrom line broadening measurements inx-ray powder diffraction line wasabout 50 nm.1 Fig. la shows that thepeaks at 3572 wavelengths/cm aremissing in the range of the humandeposit but are present in that ofsynthetic crystals of about 1 ,um.3 Theabsence of the above peaks in thespectrum of the finely divided apatiteis due to the fact that the stretching andbending modes of OH bonds are

60-

40-

20-

C 0-o-

P60

40-

Synthetic apatite

2Human deposits

3500 800 600 400Wavenumber (/cm)

o0

Fig. 1 Infrared spectra of (a) finelydivided apatite from periarthriticdeposits and (b) Durangohyproxyapatite. Note that peaks at3572 cm and 620 cm are absent inspectrum ofhuman deposits.

perturbed by the H bonding of waterto surface OH ions of the apatite.2

Infrared absorption spectrometryhas also been used to show thepresence of CO3 ions in humandeposits but I shall refer to this inconjunction with the x-ray diffractionresults.

X-RAY DIFFRACTIONWhenever anx-ray beam is incident ona crystal it is scattered by the atoms inthe crystal. The atoms in the crystalsare arranged in a regular manner witha definite periodicity, allowingidentification of a series of rows ofplanes with a constant interplannerspacing. Due to this geometry of theatoms, scattered x-rays, whensuperimposed on each other, are

20i



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Application ofphysical methods Suppl p 69

annulled except in certain directionswhere there is a constructivereinforcement. This phenomena iscalled diffraction. The direction of adiffracted beam can be found byBragg's law:

nA=2d sin 6,where n = order of diffraction (n =integer) A =wavelength ofx-rays d= in-terplanar spacing (characteristic of thecrystal lattice) and O=angle of diffrac-tion.Experimentally, one can use x-rays

ofknown wavelength A, measure angle0, and hence determine theinterplanar spacing d. Becausebiological crystals are very small oneuses the powder diffraction method. Inpractice, powdered specimen is filledin a thin tube or coated on the surfaceof a very thin glass rod (about 1 mm indiameter) and mounted in aDebye-Scherrer camera. A film is thenplaced in the camera. This surroundsthe specimen to record diffractedbeams as arcs. Such a set of lines isunique to the crystalline substance inquestion and may therefore be used asthe 'fingerprint' of the substance.From interplanar spacings one canevaluate lattice constants of the crystalstructure with a great accuracy.

Both hydroxyapatite and calciumpyrophosphate have an immenselycomplex growth system.'`4 Manydifferent phosphate salts can growfrom the same reaction system.-Powder diffraction methods,therefore, are immensely useful indetermining crystal growth conditionsin in vitro systems. Fig. 2 showsdiffractographs of (pyrophosphate)crystals grown from solution and geland that of a deposit from articularcartilage. From these patterns it wasdeduced that other compounds such ascalcium pyrophosphate tetrahydrate(CPP') can grow in in vitro systems.Pathological deposits have beenshown to contain only the mixture ofmonoclinic and triclinic phases ofcalcium pyrophosphate dihydrate(CPPD). Because of the fingerprintnature of diffraction lines it is possibleto detect as little as 0- 1% weight of thecalcium phosphate in another.

Careful determination of latticeparameters allows detection of foreignatoms (impurity) incorporated into thehost lattice. Several ions such as F CRC03 may be accommodatedsubstitutionally, in hydroxyapatite

~~,Wjrwm

Fig. 2 X-ray powder diffractographs ofpyrophosphates: (a) grown from solution(b) grown in gelatin gel, and (c) found on cartilage surface.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Table 1 Lattice parameters ofapatite from particular depositsfrom three patients

Sources ofapatite Lattice parameters (A)

'a' Ic,

Case 1 9-37 6-88Case 2 9-38 6-87Calcergy (at one week) 9-38 6-88Durango hydroxyapatite 9-42 6-88Fluorapatite 9 37 6-88Carbon apatite (7% weight CQ) 9 38 6-89

Accuracy of lattice parameters is + 0-01 .

lattice. Such substitution normallyalters the lattice parameters of acrystal by small amounts. Table 1 liststhe lattice parameters of the apatitefrom particular deposits of threepatients, simple calcergy at one week'(Harries et al., p. 102) Durango,fluorapatite and carbon apatite. The'a' lattice parameter of the human andcalcergy deposits is consistentlysmaller than that of hydroxyapatite,and similar to those of both carbonapatite and fluorapatite. F-l and/orCO.V ions could therefore be presentin the above crystals, but no fluorinewas found by modified Diggens andRosse method (1981). Infraredspectra (Fig. 3) of all the deposits showpeaks at 1460 cm, 1415 cm, and 869cm. These bands appear in the spectraof synthetic carbon apatites.5 6 It istherefore apparent that the in vivodeposits contain carbon apatite. Theamount of CO3? content can in fact beestimated from the lattice parametermeasurements. LeGeros has shownthat 'a' parameter of thehydroxyapatite cell decreases linearlywith increasing CO3> concentration inthe range 0-24% weight.7 On thisbasis human deposits we examined

contain CO3: in the range 8-9%weight. This finding raises questionregarding the source of CO'incorporation in in vivo mechanisms ofarthropathic apatite growth.The mode of carbonate substitution

into biological apatite remainscontroversial.5 The controversycentres around the site of CQ3,substitution in the hydroxyapatitelattice. From electron spin resonanceand the infrared studies of'3C-enriched apatites Doi et al.concluded that CO' ions aresubstituted at POr sites.6 LeGerosalso reached this conclusion becausecarbonate substitution causesshrinkage in 'a' parameter of the CO?-group.7 Matalon and DeBenyacer,however, claim that carbonate contentof calcified deposits from propositus ofan idiopathic chondrocalcinosis showchanges in carbonate content onheating.8 This change was variablefrom specimen to specimen. From thisit was surmised that in the abovedeposits CO3 replaces both PO' andOH- groups but that the ratioreplacing PO' and OH group doesnot remain constant. The exact site ofCO3 substitution in arthropathic

deposits cannot be easily determinedby the existing data on infraredabsorption and lattice parameter.

APPLICATION OF SYNCHROTRONX RADIATIONIntensities of diffracted beams are verysmall. In a laboratory, therefore, withnormal x-ray tube sources it requiresabout 48 hours to obtain a powderdiffraction pattern of a

well-crystallised specimen-whichcontains a few tens of milligrams. It istherefore often impossible to obtain adiffraction pattern of minute depositsin situ-say on a cartilage surface.Recently, however, it has becomepossible to use synchrotron radiationsource, which provides 'tunable'x-raysof very high intensity, the diffraction(Fig. 4) pattern of crystalline depositson cartilage surface. The patterncovers the diffraction angle in therange 0 =1°-12°. This opens upexciting prospects for studying in situarthropathic deposits. It is concluded(Fig. 4) that crystallites on cartilageshow no preferred orientation as theydo in bone and enamel.

EXAFSThis is really a form of x-rayabsorption spectroscopy. Anabsorption spectrum of a substance isobtained by shining an intense x-raybeam-for example, that fromsynchrotron source-on it. Thespectrum arises because x-rayabsorption cross section for theexcitation of an electron from a deepcore of the atom exhibits oscillation asa function of x-ray (photon) energy.

Fig. 3 Infrared spectra ofa typical periarthritic deposit. Peaks at 1460 cm, 1415cm, and 869 cm are additional to those in the pure hydroxyapatite spectrum.

Fig. 4 X-ray diffraction pattern of insitu deposits on articular cartilagesurface taken with synchrotronradiation source.

100-

80.

60

40.

20,

3500 3000 2500 2000 1800 1600 1400 1200 1000 800 600 400Waverumber ( /cm)



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Application ofphysical methods Suppl p 71

I

X- Ry E-ww

Fig. 5 A schematic diagram ofanEXAFS spectrum.

This manifests itself as a fine structurein the x-ray absorption spectrum justabove the absorption edge (Fig. 5). Incontrast to x-ray diffraction, EXAFSstudies are restricted to the immediatevicinity of the absorber element. Theytherefore yield information on thelocal structure and chemical states ofbonding sites. On interatomicdistances co-ordinate number andgeometry of an atom within a complexorganic matrix can therefore bestudied.

So far EXAFS has not been used toobtain information on arthropathicdeposits. Miller et al. and Binstead etal. have applied EXAFS to bonemineral9 10 and have concluded thatthe calcium 'environment' in bonemineral is not the same as syntheticand geological hydroxyapatite.Instead the bone mineral is shown tohave a structure that is intermediatebetween amorphous calciumphosphate and crystallinehydroxyapatite with high defectdensity. EXAFS has also shown that indevelopment of bone the calciumenvironment changes in a similarmanner to that of in vitro maturationof amorphous calcium phosphate topoorly crystallised hydroxyapatite.

METHODS BASED ON ELECTRONMICROSCOPYWhen a focused-high energy-electron probe makes an impactwith a material it generates a varietyof different signals.Low energy secondary electrons

(energy < 50 eV) generated from nearthe surface of a material givesinformation on topography and is usedin scanning electron microscopy(SEM). Elastically transmittedelectrons are capable of diffractionand give information on structure. The

/14PLATELETS WITH NON SINGULARITIES

FASTER GROWTHINSTABILITY- POSSIBLE VLS GROWTH

31>.k 8' / eF S ;*

GROWTH - CONNECTION ADJ. PLATELETS

CROSS CONNECTION CELL- LIKE

0wwtt4$*q; DEVELOPMENT

IN BETWEEN CELLS

d FINAL SMOOTHENING - CRYSTAL

SURFACE

Fig. 6 A sequence ofevents in the growth ofCuAlS2 crystals grown by chemicalvapour deposition techniques.

collision of electrons within thematerial gives rise to the emission ofcharacteristic x-ray radiation whichare used in EDX and WDX to giveinformation on chemical composition.With the aid of modern microscopesvery high resolution (2A in TEM and50A in SEM) images can be obtained.The values of these techniques hardlyneeds to be emphasised. The EDXtechnique allows constructioncomposition maps with a resolution of1 ,im2 and has shown that thecalcified lesions in simple calcergyhave layers and that the Ca: P ratio ineach layer is different.'

So far high resolution microscopyhas thrown little light on growthmechanisms of crystals in humanarthropathic deposits.SEM, however, has been used to

evaluate a sequence of events (Fig. 6)in the growth of tiny crystals of CuAlS2grown by chemical vapour depositiontechnique."

In conclusion, I should like toemphasise that an integratedprogramme of investigationcombining the above methods is likelyto contribute in the understanding ofthe crystal deposition processes inarthropathies.

References

1 Shah J S, Harries J E, Dieppe P, Heap P.Characterisation of calcium phosphatecrystals relevant to crystal depositiondiseases.Ann Rheum Dis 1982;41: 312.

2 Posner A S, Betts F, Blumenthal N C.Formation and structure of syntheticand bone hydroxyapatite. Progress inCrystal Growth and Characterisation1980; 3: 47-64.

3 Shah J S, Dieppe P. Crystal depositiondiseases of the joints. Progress in CrystalGrowth and Characterisation 1980; 3:17-47.

4 Shah J S, Heap P, Dieppe P. Depositionof calcium pyrophosphate dihydratecrystals in vivo and vitro. J Rheumatol1981; 8: 1016.

m k- - -"'I I/ //

11 ---i 414.rk -,FoZix IAl1

ILLING



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


5 Montel G, Bones G, Heughbaert J C,Trombe J C, Rey C. New concepts in thecomposition crystallization and growthof the mineral component of calcifiedtissues.Joumal ofCrystal Growth 1981;53: 74-99.

6 Doi J, Moriwaki J, Aoba T, TakashashiJ, Joshin K. ESR and IR studies of car-bonate containing hydroxyapatite. Cal-cified Tissue Intemational 1982; 34:178-81.

7 LeGeros R Z. Effect ofcarbonate on the

lattice parameters of apatite. Nature1965; 206: 403-4.

8 Matalon J R, De Benycar D. Crystallo-chemical study from cartilage and sub-cutaneous calcification. J Rheumatol1981; 8: 1010.

9 Miller R M, Hukins DW L, Hasnain S S,Lagarde P. EXAFS studies of the cal-cium ions in bone mineral and relatedcalcium phosphate. Biochem BiophysRes Comm 1981; 99: 102-6.

10 Binstead H, Hasnain S S, Hukins

D W L. Developmental changes in bonemineral structure demonstrated byEXAFS spectroscopy. Biochem Bio-phys Res Commun 1982; 107: 89-92.

11 Shah J S. Growth morphology andimpurity characterisation of some I, II,Vk sulphides and selenides. Progress inCrystal Growth and Characterisation1981; 3: 333-89.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Radiology of the crystal-associated arthritideslAIN WATT

From the Department ofRadiology, Bristol Royal Infirmary, Bristol BS2 8HW

Introduction

Appreciation of the role and effects ofcrystals in arthropathy is advancingand changing rapidly. Threesubstances-namely, sodium biurate,calcium pyrophosphate dihydrate, andhydroxyapatite-are recognised tohave acute and/or chronic radiologicalarticular manifestations.

I will describe the pure radiologicalfeatures of each of these crystalassociated arthritides but it is essentialto appreciate that they may occur inany combination or they may coexistwith another arthropathy such asrheumatoid disease. The effect of eachon the other has not been fullyassessed, although a modifyinginfluence of pyrophosphate arthritison rheumatoid disease has been seen.'

Gout

The principal radiological features ofgout are due to the presence of tophi inor around bone. As chemotherapy isdesigned to prevent the formation ofsuch tophi many patients with gouthave either no abnormal appearancesat radiology or entirely non-specificfeatures. The individual radiologicalsigns of primary and secondary goutare identical, though in secondary goutan atypical joint may be primarilyaffected or established features ofbone erosion may be present inpatients seemingly sustaining theirfirst clinical attack.The initial radiological description

was made soon after the discovery ofx-rays' and has been followed by fullerdescriptions.3 In acute gout themetatarsophalangeal joint of the greattoe is affected in 80% of cases andill-defined soft tissue swelling may beappreciated, although less readily thanon clinical examination. Calcificationis absent and there are no features topermit a specific diagnosis.Osteoporosis may occur but is not

juxta-articular or periarticular indistribution as in rheumatoid disease,nor is it as extensive as that seen inregional migratory osteoporosis.

In chronic gout there ischaracteristically an asymmetricalpolyarticular arthropathy pre-dominantly affecting the meta-tarsophalangeal joints of the greattoe (Fig. 1). Overall, the distributionof joints affected is predominantlyperipheral, the feet, hands, wristsand elbows being the major sites.Sacroiliac, acromioclavicular andstemoclavicular joint disease are lesscommon. Distribution in the hands isthe inverse of that found inrheumatoid disease, with majorinvolvement of the terminal inter-phalangeal joints and least of themetacarpophalangeal joints. The soft

Fig. 2 Gross gout ofring finger. Noteextensive asymmetrical soft tissueswelling and well defined erosions. Anoverhanging margin sign is present(arrow).

Fig. 1 Tophaceous gout inmetatarsophalangeal joint ofgreat toe.

Note considerable soft tissue swellingand relative preservation ofjoint spacewidth. Well defined erosions are

present adjacent to joint. Bone densityis normal.

tissue abnormality is eccentric asym-metrical soft tissue swelling, presentonce tophaceous deposits haveexceeded about 5 mm. Soft tissuetophi are principally present aroundthe joints of the hands (Fig. 2), ankles,and elbows. Those at the elbow may bemistaken for rheumatoid nodules,while those at the tendoachilles inser-tion may mimic xanthomata. Soft tis-sue swelling about the dorsum of thefoot and os calcis is particularly charac-teristic. Calcification of purely soft tis-sue tophi is unusual and is thought toindicate either coexisting pyro-phosphate or hydroxyapatite deposi-tion, calcium dysmetabolism, or infec-tion. Ossification of tophi, thoughrare, has been reported on occasionsand typically is found in the hands.3The bony manifestations of urate

deposists are typical although notdiagnostic. Lesions in the soft tissues



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


adjacent to bone induce areas ofpressure erosion with wasting orindentation of cortical outline creatingsharply defined sclerotic margins(Figs. 1, 2). There is no change inoverall bone density. Lesions in theperiosteum or cortex produceexpansile corticated, sharply defineddefects typically in a periarticularposition. The cortex is not usuallyintact over the lesion and thishook-like appearance, or overhangingmargin of Martel (Fig. 2), is a wellrecognised feature.4 Lesions within themedulla are typically round or oval inthe bony long axis with sharply definedsclerotic margins. Size normally variesfrom 1 to 3 mm, although frequentlyexceeds 5 mm. The sharply definednature of the bone erosion, with, at thisstage, the preservation of hyalinecartilage width, is in sharpcontradistinction to rheumatoiddisease. Indeed, despite substantialerosion cartilage width is preservedlate into the disease. Multiple

Fig. 3 Established joint disease withsubchondral collapse ofsubarticularcysts and secondary osteoarthritis atmetacarpophalangeal joints. Note newbone formation (arrows) not present inrheumatoid disease.

Fig. 4 Typical meniscalchondrocalcinosis. Note coarse

granular quality and absence ofany other abnormality.

Fig. 5 Gross capsular and ligamentous chondrocalcinosis at knee calcificationextends along poplitens bursa and into superior tibiofibular joint (arrows).

subarticular radiolucencies eventuallycoalesce and may be shownarthrographically to communicatewith the joint space like rheumatoidgeodes. Subsequent subchondralcollapse occurs with degenerativejoint disease (Fig. 3). Osteoporosisonly occurs as the result of disuse.Fibrous ankylosis is relativelycommon, whereas bony ankylosis isreported only occasionally,predominantly in the hands,particularly between carpal bones.Incidence of chondrocalcinosis is nohigher in those with gout than in acontrol population.Gout is also associated with bone

proliferation (Fig. 3) especially at theenthesis particularly with theformation of organised calcaneal spursand bridging syndesmophytes in the

thoracolumbar spine. It is not clear,however, whether this is amanifestation of gout or a reflection ofa broader dysmetabolism embracingobesity and diabetes mellitus. Uratecrystals in urine produce non-opaquecalculi with secondary obstructiveuropathy.The differential diagnosis of acute

gout includes either of the other twoacute crystal arthritides, septicarthritis and seronegativespondyloarthropathy. In establisheddisease the important differentialdiagnosis rests with psoriatic arthritis,because of the relative preservation ofbone density; the asymmetry oflesions; the presence of boneproliferation; and the predominantlyperipheral involvement, particularlythe terminal interphalangeal joints in



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Radiology ofthe crystal-associated arthritides Suppl p 75

Fig. 6 Pyrophosphate arthropathy. Gross new bone formation is mostpronounced at patellofemoral joint. There are apparent loose bodies insuprapatellar pouch.

the hands. Patchy sacroiliacinvolvement is reported in botharthritides. Inflammatory osteo-arthritis, particularly erosive osteo-arthritis, must also be considered as a

differential diagnosis. The preser-vation of the width of joint space in thepresence of well defined erosions ingout is seen also in multicentricreticulohystiocytosis and, in a singlejoint, of pigmented villonodularsynovitis.

Calcium pyrophosphate depositiondisease

Calcium pyrophosphate depositiondisease (CPPD) has been recognisedfor nearly 20 years but there is stillappreciable difficulty in itsterminology. Precise methods ofestablishing the diagnosis are beyond

the scope of this paper, but it isessential to specify three distinctfacets. Firstly, the presence ofcalcification radiographically-chondrocalcinosis; secondly, theoccurrence of acute clinical attacks orepisodes of arthritis-pseudogout;and, thirdly, the development of a des-tructive joint disease-pyrophosphatearthropathy. The principal radio-logical features of chondrocalcinosisand pyrophosphate arthropathy havebeen reviewed57 but the relationshipbetween them may prove difficult bothradiologically and clinically.

Radiologically, chondrocalcinosis isan age-related phenomenon with a

peak incidence of about 30% inelderly institutionalised patients.' Thepresence of chondrocalcinosis may notnecessarily indicate calciumpyrophosphate dihydrate as it may be

due to calcium phosphate dihydrate,hydroxyapatite, or, occasionally,brushite. When the features ofpyrophosphate arthropathy arepresent the correct specific diagnosismay be inferred, but if, for example,there has been total hyaline cartilageattrition chondrocalcinosis may not beseen.

CHONDROCALCINOSISThe detection of small quantities ofchondrocalcinosis requires ameticulous radiographic techniquewith optimal resolution.6 In mostcases, however, careful routineradiography will suffice. Radiologicalexamination of the knees (in theanteroposterior projection) gives a90% detection rate for chondro-calcinosis, 98% with antero-posterior views of the knees andhips, and 100% if the wrists are alsoexamined.7 Chondrocalcinosis mostcommonly occurs in fibrocartilage orhyaline cartilage, typically the menisciof the knee joint (Fig. 4), triangularligaments of the wrists, and theacetabular and glenoid labrae. Lessusually the annulus of theintervertebral discs and otherfibrocartilaginous structures areaffected. The calcification is typicallycoarse and granular in fibrocartilage,whereas in hyaline cartilage it isusually discrete, well defined, andlinear, paralleling the underlyingarticular bony cortex. Calcificationalso occurs in the synovium,particularly at the wrists, knees,metacarpophalangeal joints andmetatarsophalangeal joints and ischaracteristically ill defined and hazy.Gross calcification with jointdisruption may present a tophaceousappearance.' Calcification also occursin the joint capsule, particularlyaround the elbows, knees, (Fig. 5) andmetatarsophalangeal joints.Calcification at tendon insertion intobone, particularly the tendoachilles, iscommon and is typically linear andextensive compared with the morediscrete and focal involvement seen inhydroxyapatite deposition states.Occasionally calcification may occur inbursae when an ill-defined cloud-likequality may be observed.

PYROPHOSPHATE ARTHROPATHYThe structural joint changes ofpyrophosphate arthropathy most



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Fig. 7 Pyrophosphate arthropathy with very extensive joint disruption,resembling a charcot joint.

c lo s e I yresemble t h o s e o fosteoarthritis. There are, however,major differences. Pyrophosphatearthropathy affects more unusualsites, particularly the radiocarpal,metacarpophalangeal, trapezio-scaphoid, elbow, and the gleno-humeral joints. The distribution withinindividual joints is also different-forexample, the patellofemoral joint ofthe knee is predominantly affected(Fig. 6). Formation of new bone is vari-able and may be excessive with exuber-ant or massive osteophytes whencompared with osteoarthritis. Largemasses of trabeculated osteophytesdevelop (Fig. 6) and may appear sep-

arate from the underlying bone, sug-gesting loose bodies. Indeed, isolatedfragments of cartilage are also more

common, giving rise to genuine sepa-

rate intra-articular loose bodies.Unless they become substantial in size,however, these are distinguishedby the absence of bony trabeculaeon x-ray. Formation of osteophytevaries considerably; while in mostpatients considerable bone productionoccurs, in others it is virtuallyabsent with merely smooth ebur-nated articular surfaces. The extent ofthis subchondral sclerosis and theassociated subarticular radiolucenciesis far more pronounced in pyrophos-phate arthropathy than in osteo-arthritis. There is also a predilectionfor severe destructive changes in pyro-phosphate arthropathy and the like-ness of these joints to neuropathicjoints has been emphasised (Fig. 7).10Though this may be observed in thehip and knee joints, proximal row col-

lapse of the carpus associated withscapholunate diastasis is typical. Thesevere destructive nature of the arth-ropathy also produces severe facetdegenerative disease in the lum-bar spine, often with degenerativespondylolisthesis.Pyrophosphate arthropathy

predominantly affects large joints,particularly the knees, hips, andglenohumeral joints. Severe deformityof the knee joint may cause stressfractures of the upper tibia requiringjoint replacement. Indeed, carefulanalysis of most cases of deformingknee disease requiring total kneereplacement shows undoubtedfeatures of pyrophosphatearthropathy.

Clinical and radiological correlationis, however, fraught with difficulty.There is no definite relation betweenclinical arthritis and the presence ofchondrocalcinosis or pyrophosphatearthropathy. For example, extensiveradiographic abnormality may bepresent in entirely asymptomaticjoints. Furthermore, there is nodefinite relation between symptomsand the type of calcification present.Calcification may be ephemeral ornever seen although crystals may beclearly demonstrated from jointaspirate. There is no definiteprogression from chondrocalcinosis topyrophosphate arthropathy.

It is unclear whether pyrophosphatearthropathy and/or chondrocalcinosisare specifically related to anyparticular predisposing factor. Onlyprimary hyperparathyroidism andhaemochromatosis (both primary andsecondary) have such a proved causalrelationship."1 Indeed, apart from veryminor variations 'primary'pyrophosphate arthropathy oftencannot be distinguished on purelyradiological grounds from thatassociated with other diseases.' Subtledistinctions between primarypyrophosphate arthropathy andhaemochromatosis have beenemphasised in a meticulous study ofhand involvement (D Resnick, paperpresented at the International SkeletalSociety Meeting, San Francisco,1982). Joint space narrowing at thering and little finger metacarpo-phalangeal joints and crumbling ofarticular surfaces but with relative pre-servation of the radiocarpal joint andlarger hook-like osteophytes on the



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Radiology of the crystal-associated arthritides Suppl p 77

Fig. 8 Pyrophosphate arthropathy and rheumatoid disease. Note normal bonedensity, patchy involvement, and absence offresh erosion. Radiocarpal jointdisruption is consistent with either pyrophosphate arthropathy or old rheumatoiddisease with superadded degenerative disease.

metacarpal heads were found to bemore common in haemachromatosis.Scapholunate dissociation andchondrocalcinosis of the thumbcarpometacarpal joint were, how-ever, commoner in pyrophosphatearthropathy.

There is mounting evidence thatpyrophosphate arthropathy has a

modifying effect on other non-crystalarthritides and in a recent series ofpatients with chondrocalcinosispatients with rheumatoid proved byARA criteria had extremely modified

radiographic features.' This difficultyin distinguishing between rheumatoiddisease and pyrophosphatearthropathy has also been noted byothers.'2 In most patients withrheumatoid disease and pyro-phosphate arthropathy there are onlyoccasional erosions with the preserva-tion of normal bone density and anasymmetry of involvement (Fig. 8).The relationship between pyro-phosphate arthropathy and degenera-tive joint disease is ill understood, butit is clear that chondrocalcinosis occursmore often in previously traumatisedjoints as, for example, after meniscec-tomy," and that the degree of associ-ated pyrophosphate arthropathy inthese patients is worse in thosepatients who have evidence of osteo-arthritis elsewhere-for example, inthe terminal interphalangeal joints ofthe hands."3

Calcium hydroxyapatite depositiondisorders

EXTRA-ARTICU LARThe presence of foci of calcificationrelated to the shoulder joint is wellrecognised and although thesupraspinatus tendon (Fig. 9) is by farthe commonest affected numerousother foci are now acknowledged.These include the wrists (flexor carpiulnaris, less commonly flexor carpiradialis and extensor carpi ulnaris);adjacent to the metacarpophalangealjoints of the fingers and themetatarsophalangeal joints of the feet,especially the great toe (Fig. 10); theelbow, particularly the common flexorand extensor insertions; and at the hipjoint, particularly the glutealinsertions into the greater trochanter.Recently, involvement of the longuscolli with prepharyngeal soft tissueswelling has also been recognised. Thebursa may be affected on occasions,and this calcification may not be seenin radiographs.The condition is usually detected in

a single joint, but when multiple jointsare affected the onset is simultaneousin about one third and successive intwo thirds of cases. Typically, thepatient is aged between 40 and 70 andthe initial radiological features areill-defined soft tissue swelling adjacentto a joint, corresponding to the acuteinflammatory episode. Calcification isalmost always observed on the first



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Fig. 10 Acute hydroxyapatiteperiarthritis ofgreat toe. Noteill-defined soft tissue swelling anddiscrete, dense calcification adjacent tometatarsal head.

Fig. 9 Extensive hydroxyapatite periarthritis ofshoulder. There is amorphouscalcification extending from supraspinatus tendon into subdeltoid bursa and over

humeral head (arrows). Irregularity and sclerosis ofsupraspinatus insertion andfocal osteoporosis are present (open arrows).

examination and is initially ill definedand poorly localised. Thesupraspinatus is most common of therotator cuff tendons to be affected andis best seen on a film in externalrotation, whereas the infraspinatusand teres major may be seen on

intqrnal rotation posterior to thehumeral head and the subscapularisanterior. When calcification is illdefined symptoms can usually berelieved by aspiration underfluoroscopic control, with injection oflocal anaesthetic and steroid. With

time the calcification becomes more

localised, homogenous, and more

dense, with a linear or circularconfiguration. Around the shoulderthis may persist for many years and isresistant to needle aspiration. Asudden change in the clinical signs isoften associated with apparent ruptureand shedding of the calcificationeither into the shoulder joint or, morecommonly, into the adjacentsubdeltoid bursa from thesupraspinatus tendon. If there isextensive involvement of the tendon

rotator cuff instability may thendevelop, manifested by cephalicmigration of the humeral head withexcavation of the inferior aspect of theacromion. Though no abnormality ofbone is observed on initialpresentation with chronic diseasethere is usually bone resorption,sclerosis, and cyst formation atrotative cuff insertions.When multiple joints are involved

both shoulders usually demonstratethe abnormality. However, there isundoubtedly an incidence ofnon-symptomatic calcific disease.The calcification needs to be

distinguished from other causes of softtissue calcification in a periarticulardistribution, particularly thatassociated with calcium dysmeta-bolism (especially secondary hyper-parathyroidism and renalosteodystrophy) and in connective tis-sue disorders such as Ehlers-Danlossyndrome.

I NTRA -ARTICU LAR

Since Dieppe et al. found minutedense opacities within joints ofpatients with severe and destructivechanges,"4 hydroxyapatite has beenassociated with osteoarthritis,



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Radiology of the crystal-associated arthritides Suppl p 79

-qq.

Fig. 11 Advanced 'Milwaukee' shoulder. The acromium, coracoid process, andouter third of the clavicle are all destroyed. There is glenohumeral osteoarthritis.Calcification (arrow) is present in region offormer rotator cuff

particularly of the knee. It has alsobecome clear that calcification ofhyaline cartilage (chrondrocalcinosis)may also be due to hydroxyapatite"5and be associated with a pronouncedinflammatory arthropathy. Similarly, aprogressive and destructive arthritisoccurs at the shoulder joint,radiologically representing the severeend stage of rotator cuff instability(Fig. 1 1) (the 'Milwaukee shoulder').16In this condition there is considerablesynovial swelling with gross excavationand attrition of bone, particularly theacromium, the lateral end of theclavicle, the corocoid, and, latterly,the glenoid fossa. These appearances,when pronounced, are probablycharacteristic and entertain no seriousdifferential diagnosis; they are usuallybilaterally symmetrical. It is becomingclear that these patients also havedestructive arthritis of other joints,particularly the knee (Fig. 12), andthat hydroxyapatite deposition diseaseneeds to be considered in the presenceof severe Charcot-like knee andshoulder joints.

Hydroxyapatite deposition occursboth in joints and adjacent joints inmixed connective tissue disease and

scleroderma,17 though the relevance ofthis calcification in the pathogenesis ofboth arthritides is unclear. An erosivearthropathy with associatedcalcification should allow a specificdiagnosis of mixed connective tissuedisease rather than rheumatoid. In allof the hydroxyapatite depositiondiseases there is evidencescintigraphically of considerablemetabolic activity in the calcificationsite.

Conclusion

It is clear that the understanding of thecrystal-associated arthritides hasadvanced considerably in the pastdecade. Many radiological featuresthat would have been dismissed asdegenerative changes or osteoarthritisare now known to be manifestations ormodifications of osteoarthritis as aresult of the influence of crystaldeposition. Although the threeprincipal crystal arthritides may bedistinguished radiologically there is anappreciable overlap between them andthere are no exclusive pathognomonichallmarks to allow an absolutediagnosis.

Fig. 12 Left knee ofsame patient as Fig. 11. Extensive sclerosis and attrition ofbone in lateral compartment suggests an atypical osteoarthritis. There is possiblvchondrocalcinosis (arrow). Hydroxyapatite crystals were isolated from jointfluid.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


References

1 Dieppe P A, Alexander G J M, Jones HE, et al. Pyrophosphate arthropathy: a

clinical and radiological study of 105cases. Ann Rheum Dis 1982; 41:369-76.

2 Huber D. The use of x-rays in internalmedicine. Deutsche MedizinischeWochenxhrift 1896; 22: 182-95.

3 Watt I, Middlemiss J H. The radiologyof gout. Clinical Radiology 1975; 26:27-36.

4 Martel W. The overhanging margin ofbone; a roentgenologic manifestation ofgout. Radiology 1968; 91: 755-6.

5 Martel W, Champion C K, ThompsonG, Carter T. A roentgenologicallydistinctive arthropathy in some patientswith the pseudogout syndrome. A J R1970; 109: 587-607.

6 Genant H K. Roentgenographic aspectsof calcium pyrophosphate dihydratecrystal deposition disease (pseudogout).Arthritis Rheum 1976; 19: 307-28.

7 Resnick D, Niwayama G, Goergen T A,

et al. Clinical, radiographic andpathological abnormalities in calciumpyrophosphate dihydrate depositiondisease: pseudogout. Radiology 1977;122: 1-15.

8 Wilkins E, Dieppe P A, Maddison P,Evison G. Articular chondrocalcinosisand its association with osteoarthritis inthe elderly. Ann Rheum Dis 1981; 40:516-21.

9 Ling D, Murphy W A, Kyriakos M.Tophaceous pseudogout. AJR 1982;138: 162-5.

10 Menkes C J, Simon F, Delrieu M, ForestM, Delbarre F. Destructive arthropathyin chondrocalcinosis articularis.Arthritis Rheum 1976; 19: 329-48.

11 Resnick D, Niwayama G, eds. Diagnosisof bone and joint disorders.Philadelphia, London, Toronto: W BSaunders, 1981: -.

12 Doherty M, Dieppe P A, Watt I. Theinfluence ofprimary osteoarthritis on thedevelopment ofsecondary osteoarthritis.London: Heberden Society, 1982.

13 Doherty M, Watt I, Dieppe P A.

Localised chondrocalcinosis in postmeniscectomy knees. Lancet 1982; i:1207-10.

14 Dieppe P A, Crocker P R, Huskisson EC, Willoughby D A. Apatite depositiondisease: a new arthropathy. Lancet1976; ii: 266-9.

15 Bonavita J A, Dalinka M K,Schumacher H R. Hydroxyapatitedeposition disease. Radiology 1980;134: 621-5.

16 McCarthy D J, Cheung H S, HalversonP B, Garancis J C. 'Milwaukee shouldeesyndrome: microspherules containinghydroxyapatite, active collagenase andneutral protease in patients with rotatorcuff defects and glenohumeralosteoarthritis. Semin Arthritis Rheum1981; 11, suppl: 119-21.

17 Resnick D, Scavulli J F, Goergen T G,Genant H K, Niwayama G.Intra-articular calcification inscleroderma. Radiology 1977; 124:685-8.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Case presentationsPresented by the Department ofMedicine, Bristol Royal Infirmary, Bristol BS2 8HU

CASE 1An 89 year old woman withdiuretic-induced tophaceous gout.Points of note were:

(1) The patient had been receivingdiuretics for many years, but had noimpairment of renal function.

(2) Radiographs showed typicalgouty erosions in the hands, but shehad never had an acute attack of gout(Fig. 1).

(3) Large tophi were seen, mainlyon the hands, and in particular over thedistal interphalangeal joints. Some ofthese tophi had discharged (Fig. 2).

Fig. 1 Radiograph ofhand (case No1) showing gouty tophi of the softtissues and typical bone erosions ofgout. The patient had never had anattack ofacute gout.

Fig. 2 Hands ofcase No 1. Note thelarge gouty tophi and the distributionover the interphalangeal joints.

Discussion points included:(1) The role of diuretics in the

genesis of gout in the elderly.(2) The contributions of aging of

connective tissue and osteoarthriticchanges around the distalinterphalangeal joints in dictating thedistribution of crystal deposits in theelderly.(3) The striking lack of

inflammation around the tophi and theabsence of acute attacks.

(4) The aetiology of the bonyerosions in gout in the apparentabsence of inflammation.This patient was thought to be similarto others described recently with goutytophi without gouty arthritis (seeScott, p. 16).The possible factors relating to

distribution, including connectivetissue changes, temperature, andosteoarthritis had been discussed byDr Calvert (see Fiddis et al., p. 12).Professor Dixon suggested the term'impostumous gout' to describe suchcases.

CASE 2A 31 year old woman suffered a kneeinjury when aged 16, and firstpresented with an acute monoarthritisof that knee 12 years later. She hassince had five well-characterisedattacks of pseudogout in that knee.Points noted included:

(1) Cruciate instability of the leftknee only.

(2) Radiological chondrocalcinosislimited to the left knee.

(3) The lack of any metabolicabnormality or family history ofarthritis.

(4) The good preservation ofcartilage thickness in the joint.

Discussion centred around theconcept of secondary calcificchondrocalcinosis occurring inlocalised areas of previous jointdamage in the absence of any systemicdisorder. Although this association hasnot been widely recognised, similarcases have been reported (seeDoherty, p. 38).

CASE 3A 75 year old ex-ballet dancer with

generalised hypermobility and CRSTsyndrome. She developed a selflimiting attack of severe low back painand recurrent attacks of arthritisaffecting the wrists andmetacarpophalangeal joints.Points noted included:

(1) The typical hands of CRSTsyndrome with associated reversibleulnar deviation of the fingers and a Zthumb deformity.

(2) Radiological chondrocalcinosisof the knees, symphysis pubis, andtriangular ligaments of wrists (Fig. 3).

(3) The presence of calcificationaround the annulus fibrosis of severaldisc spaces in the lumbar spine.Discussion points included:

(1) The possibility that the low backpain could have been related to the

Fig. 3 Radiographs ofthe hand (caseNo 3) showing (a) close up ofthefingers to show the typical soft tissuecalcification and resorption of theterminal phalanges seen in the CRSTsyndrome, and (b) close up ofthe wristofthe same hand to show the associatedchondrocalcinosis ofthe triangularligament.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


presence of crystal deposits in thespine, as has been suggestedpreviously (see Doherty, p. 38).

(2) The association between theCRST syndrome and thechondrocalcinosis. It was felt that thefingertip calcification was typical ofCRST syndrome but that thechondrocalcinosis may be a chanceassociation, as a high percentage of 75year olds have age-relatedchondrocalcinosis.

(3) The influence of her generalisedhypermobility on the crystaldeposition. The question of a primarycollagen abnormality or ofhypermobility induced by her dancingwas discussed, and the question ofhypermobility in relation togeneralised chondrocalcinosis wasbrought up. Factors such as aging,hypermobility, and connective tissuechanges in the pathogenesis ofchondrocalcinosis had been discussedpreviously (see Mitrovic, p. 19 andDoherty, p. 38).

CASE 4A 43 year old man with acromegalypresented with a stiff, painfulback and painful swollen knees. Hehad undergone hypophysectomy andradiotherapy to the pituitary fossa in theprevious year.Points noted included:

(1) Persistence of gross features ofacromegaly.

(2) A mild kyphoscoliosis andrestriction of spinal mobility.

(3) Large effusions, pronouncedcrepitus, and bony swelling of theknees, with retention of a full range ofmovement.(4) Radiological features

characteristic of those described inacromegalic arthropathy were noted inthe spine and knees.

(5) He was also noted to haveperiarticular calcification around theleft knee.

(6) Fluid aspirated from the leftknee had shown numerous roundglobules staining positively withAlazarin red (see Schumacher, p.54).Discussion points included:

(1) The possible role of growthhormone in predisposing to apatitedeposition in the cartilage.

(2) The origin of the apatite in thesynovial fluid. The x-ray films showgood preservation of cartilage,

suggesting that the particles wouldhave to come from new deposits in thesuperficial areas of the cartilage, orsynovium and capsule, rather thanfrom the bone. The possible disruptionof particles from preformed deposits inthe soft tissues during the process ofaspiration was noted.

(3) The nature of the arthropathy.The knee disease was thought to betypical of acromegalic arthropathy,and associated apatite deposits insynovial fluid have been describedpreviously (see Schumacher, p. 54).The spinal disease was thought to beless typical in view of the pronouncedrestriction of movement, but it wasnoted that this was responding well tophysiotherapy.

CASE 5A 71 year old woman withlongstanding generalised nodalosteoarthritis presented with acuteinflammation around the distalinterphalangeal joint of the left middlefinger. The lesion ulcerated,discharging fluid containing numerousapatite crystals ('impostumous apatitedeposition disease' Fig. 4).

Fig. 4 Distal interphalangeal joint ofthe left middle finger (case No 5),showing the inflammation anddischarge from the finger after flare upof the 'Heberden's node'. Thedischarge contained masses ofcalciumcontaining particles thought to beapatite, and the patient had evidence ofcalcific periarthritis elsewhere.

Points noted included:(1) The clinical and radiological

evidence of gross osteoarthritis of theinterphalangeal joints.

(2) Radiological evidence of calcificperiarthritis is the left ring finger.

(3) Numerous para-articular cystsaround the distal interphalangealjoints typical of those seen ingeneralised osteoarthritis.Discussion points included:

(1) The mode of formation ofpara-articular cysts adjacent to thedistal interphalangeal joints inosteoarthritis. It was felt that thesemight be generated in a similar way asother joint cysts-that is, with aninitial communication with the joint,and a valvular mechanism pushinghyaluronate into the cyst.(2) The presence of apatite in cysts

and in distal interphalangeal joints inosteoarthritis was discussed. Variouscontributors noted that they had neverbeen able to identify apatite in thehyaluronate cysts, but that it had onoccasion been found in joint fluidaspirated directly from the distalinterphalangeal joint in Heberden'snodes.

(3) It was agreed that no good datawere available on the relationshipbetween calcific periarthritis andosteoarthritis but that the patient hadobviously had a typical attack of acutecalcific periarthritis in addition tohaving generalised nodalosteoarthritis (see Faure, p. 49).

CASE 6A 73 year old woman with a four yearhistory of a destructive arthropathy,principally affecting the shoulders,knees, and midtarsal joints. Traumaprecipitated the inflammation of theright shoulder at the start of the illness.Over the next two years this jointrapidly became grossly disorganised,and an arthroplasty was carried out.She subsequently developed pain,effusions, and radiological evidence ofprogressive destructive changes,chiefly affecting the left shoulder,knees (left worse than right), andmidtarsal joints.Points noted were:

(1) Radiological evidence of calcificperiarthritis preceding the destructivechanges in the shoulder.

(2) Loss of the rotator cuffapparatus both clinically and



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Case presentations Suppl p 83

radiologically, with upwardsubluxation of the head of the humerusin the shoulder remaining on the left(see Fig. 11 in Watt, p. 73).

(3) Gross instability of the left knee,with a marked valgus deformity.

(4) The large, cool effusions in allthe affected joints. These had beenaspirated and shown to contain a fewmononuclear cells but no polymorphs,and large numbers of Alazarin redpositive particles.

(5) The radiological evidence ofdestructive changes, including collapseof the medial tibial plateau of the leftknee. Pyrophosphate as well as apatitecrystals had been aspirated from thisjoint but not from the shoudler.Discussion points included:

(1) The case was thought to be insome ways similar to those describedas 'Milwaukee shoulder' (seeSchumacher p. 54 and Watt, p. 73).However, it was obvious from this andother reported cases that the site ofinvolvement was not specific, and thatother calcium phosphate crystalsmight be associated with these types ofcases. The term 'apatite-associatedlarge joint lysis' was suggested by thepresenters for cases of this sort.

(2) The possible mechanism of thedestructive changes was alsodiscussed. Some participants felt that aprimary disorder of subchondral bonecould not be ruled out, and thatcrystals could be a marker of disease,rather than a cause of theseextraordinary changes.

CASE 7A 65 year old woman with an 18 yearhistory of calcinosis cutis, chieflyaffecting the right index finger. Shehad presented with SBE in 1980,resulting from infection rounddischarging areas of calcification in thefingers.

Fig. 5 Clinical picture and radiograph of the right index finger (case No. 7).(a) Clinical photograph showing extensive nodular swelling ofthe finger.(b) Radiograph showing the masses of non-trabeculated calcific bodies in thesoft tissues, with normal bones and joints. Carbonated apatite was identified inthe material extracted from these deposits.

The points noted included:(1) The gross calcium deposits in the

finger with characteristic radiologicalfeatures of rounded non-trabeculateddeposits (Fig. 5).

(2) Material often discharged,analysis of a large amount by infraredspectrophotometry showed apatitewith some carbonate.

(3) There was no clinical evidence ofany associated connected tissuedisorder or of any metabolicabnormality which might help explainthe deposition.

(4) Radiological screening hadshown an associated area ofcalcification in the spine, and on

direct questioning, the patient hadadmitted to back pain isolated to thatarea.Discussion highlighted a considerableinterest in such a gross example ofpara-articular and soft tissue crystaldeposition.

(1) The lack of inflammation aroundthe lesion, except at times of dischargethrough small ulcerated skin areas wasalso noted and discussed.

(2) Although thought to be ofconsiderable interest, this casehighlighted the lack of understandingavailable to explain the formation andeffect of huge crystal deposits such asthese.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Comparison of plasma and urinary concentrations ofuric acid measured by the colorimetricphosphotungstate and enzymatic uricase methods

C. S. HIGGENS, I. K. MOSS, J. T. SCOTT

From the Kennedy Institute ofRheumatology and Charing Cross Hospital, London W6

Accurate measurements of uric acidare essential for diagnosis ofhyperuricaemia and to measure theeffects of uricosuric drugs. Routineestimations of concentrations arecommonly based on colorimetricmethods.' These are closely controlledand inexpensive methods usingmodern automated continuous flowtechniques. There are, however,important sources of error, especiallyin urinary estimations.The more accurate enzymatic assay

uses specific uricase to oxidise uric acidto allantoin. Manual andsemiautomated uricase methods havebeen described.2" These methods are,however, time consuming and requirespecialist skill and equipment, makingthem unsuitable for routine hospitaluse. In 1969 the first fully automatedenzymatic assay using continuous flowtechniques was described by Steele.4This needed two ultravioletspectrophotometers working inparallel together with a large range ofexpensive equipment; this made itimpractical for routine use. Anautomated uricase assay combines thespecificity and sensitivity of theenzymatic,method with the controlledconditions of the continuous flowautoanalyser. A method based onthese principles is being developed toprovide accurate routinedeterminations of uric acid in plasmaand urine samples.5We compared estimations of plasma

and urinary uric acid measuredroutinely in this hospital by acolorimetric method with manualresults from the uricase assay toascertain whether significantdifferences could be detected. Themanual and a new automated uricaseassay were then compared in themeasurement of plasma and urinaryuric acid concentrations.

SUBJECTS, METHODS AND

RESULTSPlasma and 24-hour urine collectionswere obtained from unselectedpatients with normal creatinineclearance, over a six month period. Allsamples and standards were assayedfor uric acid using all methods.Colorimetric phosphotungstate

method. Uric acid added tophosphotungstate in alkaline solutionproduces an intensely blue colourcaused by an uncharacterisedchromophore. The colour yield isdirectly related to the concentration ofuric acid. The automated method hashigh reproducibility and facilitates a

high standard of quality control.Errors in this method are due to theactivity of non-specific chromagens inthe samples and from the loss of uricacid which precipitates with protein,which is then removed by dialysis. Thisresults in non-linearity between colouryield and uric acid concentration espe-

cially in urine where non-specific sub-stances-for example, cystine andthiols--can contribute to totalmeasured colour.Enzymatic uricase methods. Uric

acid is oxidised by uricase to allantoin.The disappearance of uric acid may bemonitored by measuring theassociated decrease in absorbance at292 nmol/l in a spectrophotometer.This method has a high degree ofspecificity and sensitivity for theanalysis of both blood and urine.Accuracy and precision are affected bysampling errors and spontaneouschanges in sample turbidity causingvariations in light scattering andapparent changes in ultravioletabsorbance. The automated uricasemethod used a continuous flowtechnique to present samples to thespectrophotometer, thus reducingsampling errors.

Estimations of plasmaconcentrations of uric acid by thephosphotungstate colorimetricmethod were a mean of 17% higherthan the corresponding manual uricaseassay values (n = 63; r = 0-979). Thisdifference varied in proportion to theconcentration.

Estimations of urinary excretion ofuric acid by the phosphotungstatecolorimetric method were a mean of48% higher than by the manualuricase assay (n = 30; r = 0 840). Thisdifference was not proportional to theconcentration.There was an extremely good

correlation of plasma and urinaryconcentrations of uric acid by themanual and automated uricasemethods.

COMMENTThe phosphotungstate colorimetricmethod was inferior in accuracy to theenzymatic uricase method in themeasurement of plasma and urinaryuric acid concentrations.

For routine use the inexpensivecolorimetric method could be usedwith a correction factor for plasma uricacid estimations. For accurateestimations of urinary uric acid,however, only the enzymatic uricaseassay was reliable. The specificity ofthe uricase method was combined withcontinuous flow techniques in a newautomated uricase assay. Results fromthis assay correlated well with thosefrom the manual method. Anautomated uricase assay is at presentbeing developed for routine hospitaluse (L Duncan, C S Higgens, P CNicholas, unpublished data).

Further progress in detailedmonitoring of kinetic reactionstogether with cheaper highly purifiedenzyme preparations should lead togreater accuracy and precision for



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Abstracts Suppl p 85

routine hospital estimations of plasmaand urinary uric acids by the enzymaticuricase method.

References1 Henry R J. Clinical chemistry. New York:

Harper and Row, 1966: 278.

2 Liddle L, Seegmiller J E, Laster L. Theenzymatic spectrophotometric methodfor determination of uric acid. J Lab ClinMed 1959; 54: 903-13.

3 Simmonds H A. A method of estimationof uric acid in urine and other body fluid.Clin Chim Acta 1967; 15: 375-8.

4 Steele T H. An automated enzymaticspectrophotometric method for thedetermination of uric acid. TechnicalBulletin of the Registry of MedicalTechnologists 1969; 39: 270-4.

Analysis of human plasma and urine purines using highperformance liquid chromatography (HPLC)H. J. RYLANCE, RUTH C. WALLACE, AND G. NUKI

From the Rheumatic Diseases Unit, Department ofMedicine (WGH), University ofEdinburgh, Northern General Hospital,Edinburgh EH5 2DQ

The recent development ofmicroparticulate, chemically bondedpacking materials for highperformance liquid chromatography(HPLC) has allowed the developmentof sensitive methods for the detectionand quantification of purine bases andnucleosides in biological fluids. Usingreverse phase HPLC methodologydeveloped by Hartwick and others' weundertook quantitative analyses ofpurine nucleosides and bases innormal human plasma and adaptedthese methods to obtain a qualitativeprofile of urine purine excretionproducts.

Instrumentation consisted of aWaters P/N 80060 modular system 1chromatograph comprising twinpumps, solvent flow programmer,injection, 254 nm fixed waveabsorbent detector, and M730 datamodule. The column used is apre-packed Bondapak (C18) columnconsisting of a porous silica supportwith an octadecyl (C18) chemicallybonded stationary phase (particle size10,m, column dimensions 3 9 x 300mm). A dry-packed precolumn of thesame material (37-50 A.tm particlesize) protects the main column. Thesolvent system employed is a lineargradient; 100% 0-02 mol/l phosphatebuffer pH 5 6 to 40% methanol/water(60% v/v) in 35 minutes at ambient

Table 1 Normal values (p,moll1) plasma constituents. Figures are means (SD)

Women (n = 16) Men (n = 9)

Creatinine 92-6 (17-14) 97-8 (12-4)Uric acid 182-4 (32 9) 297-5 (64-1)Tyrosine 44-7 (21-9) 40-1 (35 8)Hypoxanthine 2-21 (2 67) 1-96 (1-47)Uridine/Xanthine 2-63 (1-05) 2-94 (098)Inosine 1-38 (0-91) 1-08 (0-94)Guanosine Generally below limit of estimationAdenosine 1l55 (1-39) 1-05 (0-78)

temperature and a flow rate of 1-5ml/min.

40,ul plasma or 10 ,ul of a 1 in 10dilution of urine are injected afterultra filtration in Amicon cones (CF50A).The identification of plasma purine

and nucleoside peaks has been verifiedby cochromatography with purecompounds and enzyme shift methods.The table shows normal values forplasma constituents that may beaccurately quantified. The lower limitfor detection of plasma purines andnucleosides is approximately 0 2,umol/l.

Urinary purine base and nucleosideconcentrations are much morevariable and dependent on diet andhydration. The definitive

identification of all urine componentsdetected has not been completed andmany are partly excreted asmethylated metabolites.The urine chromatogram may,

however, be used to provide aqualitative 'profile' ofpurine excretionin addition to quantitativemeasurements of uric acid andcreatinine.

Reference

1 Hartwick R A, Krstulovic A M, BrownP R. Identification and quantitation ofnucleosides. bases, and other uvabsorbing compounds in serum usingreversed phase high performance liquidchromatography. II. Evaluation ofhuman sera. J Chromatogr 1979; 186:659-76.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Inborn errors of purine metabolism in man:development of purine enzyme assays using highperformance liquid chromatographyH. J. RYLANCE, RUTH C. WALLACE, AND G. NUKI

From the Rheumatic Diseases Unit, University Department ofMedicine (WGH), Northern General Hospital, EdinburghEH52DQ

Several inherited purine enzymedefects have been associated withpurine overproduction and gout,nephrolithiasis and immunodeficiencydiseases. These include: adenosinedeaminase (ADA) deficiency (severecombined immunodeficiency),nucleoside phosphorylase (NP)deficiency (isolated T cell deficiency),nucleotidase (5'NT) deficiency(X-linked and acquiredhypogammaglobulinaemia) xanthineoxidase (XO) deficiency(xanthinuria), adenine phosphoribosyltransferase (APRT) deficiency (renalcalculi 2, 8-dihydroxy-adenine),hypoxanthine-guaninephosphoribosyl transferase (HGPRT)deficiency (Lesch-Nyhan syndrome orsevere X-linked gout) andphosphoribosyl pyrophosphatesynthetase (PRPPS) superactivity(X-linked gout with or without nervedeafness and neurodevelopmentalabnormalities). Red blood cell enzymeassays have been developed that arerapid and sensitive and avoid the needfor radiochemicals in clinicalbiochemistry laboratories.Haemolysates are prepared fromheparinised venous blood samples andassayed for purine enzymeconcentrations by methods based on areverse phase HPLC separation ofpurine bases, nucleosides, andnucleotides.

HGPRT I

Hypoxanthine (HX) + 1-pyropWosphoylribosyl-S-phosphate(PPriboseP) -- hypoxanthine ribo-nucleotide (IMP) + inorganic pyro-phosphate (PPT).

Incubation mixture contains 0*63HX, 1-05 mmol PPriboseP, 5 3 mmolMg++, 100 mmol Tris buffer (pH 7- 4).HPLC to separate IMP from HX.

Table 1 Normal values ofenzyme activities. Figures are mean (SD)

HGPRT (n = 20) 94(15) nmol IMP/mgHb/hrAPRT (n = 7) 16 (5-3) nmol AMP/mgHb/hrPRPPS (n = 9) 71-2 (15-7) nmol PRPP/mgHb/hrADA (n = 8) 0-28 (007) IU/ml packed RBCPNP (n = 11) 6 21 (1-21) IU/ml packed RBC

Column t Bondapack C 18 (3-9 x 300mm).

Eluents: 0-02 mol/l phosphate buf-fer (pH 5 6) (A) and methanol/water(60% v/v) (B) Gradient linear 100%A to 40% B in 35 minutes. Flow rate:1-5 ml/min. Injection volume 40 ,uL.

APRTAdenine (Ad) + PPriboseP -- adeno-sine monophosphate (AMP) + PP,.Reaction conditions as for HGPRT

with substitution of adenine for hypo-xanthine. Isocratic HPLC separationof AMP and Ad using phosphate buf-fer (pH 6 0) containing 6% (v/v)methanol as eluate.

PRPPSRibose-5-phosphate + ATP -* PPrib-oseP + AMP.Two stage reaction: 0-05 mmol

ribose-5-phosphate + ATP reacted inphosphate buffer (pH 7-5) with Mg++EDTA and glutathione to form PPrib-oseP. PPriboseP assayed in HGPRTreaction using excess of partiallypurified enzyme.

ADA 2

Adenosine -> inosineHaemolysate reacted with 0-14

mmol adenosine in 0 05 mmolphosphate buffer (pH 7*5).

Isocratic HPLC separation ofadenosine using phosphate buffer (pH6-0).

PNPInosine -* hypoxanthine

Haemolysate reacted with 0 2 mmolinosine in 0 05 mol phosphate buffer(pH 7 5).HPLC as in HGPRT assay.In patients with gout and

hyperuricaemia a sequence ofinvestigations is undertaken:

(1) Serum uric acid + creatinine.(2) 24-hr urinary uric acid +

creatinine + 'purine profile' (normaldiet-alcohol, tea, coffee). If >3-6mmol/24 hrs: stage 3.

(3) Repeat 24-hr urinary uric acid +creatinine + 'purine profile' on 2600kcal, 70 g protein purine-free diet fivedays). If >3-6 mmol/24 hrs.HYPEREXCRETOROVERPRODUCER.

(4) Measure plasma purines +nucleosides.

(5) Measure RBC levelsPPribose PPRPPSHGPRTAPRT

(6) If abnormal-kinetic studies,family studies, heterozygote detection,etc.

References

1 Rylance H J, Wallace R C, Nuki G.Hypoxanthine guanine phosphoribosyltransferase: assay using high perfor-mance liquid chromatography. Clin ChimActa 1982; 121: 159-65.

2 Hartwick R, Jeffries A, Krstulovic A,Brown P R. An optimised assay foradenosine deaminase using reverse phasehigh pressure liquid chromatography. JChromatogr Sci 1978; 16: 427-35.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Significance of circadian variations of uric acideliminationB. D. OWEN-SMITH

From the St Richard's Hospital, Chichester, and Department ofRheumatology, Royal Berkshire Hospital, Reading

To investigate the significance ofinverse circadian variations of renalexcretion of uric acid' and enteraluricolysis2 ' it is necessary todemonstrate that they are affected bysleep and that there is a tendency forserum urate to change, even thoughthis may be masked by effectivehomoeostasis. If during sleep the rateof urate production exceedselimination by increased enteraluricolysis and reduced renal excretionthere should be a detectable increase inserum urate under controlledmetabolic conditions.A three day study comprising two

six-hour periods of sleep separated by42 hours' 'normal' activity was under-taken. During the awake periods thesubject was on a low-purine mixedliquid diet of 0-69 MJ (165 k cal) fourhourly with a water intake of 200 mlhourly.Blood samples for estimation of

urate concentration were taken everyfour hours at the midpoint of fourhourly pooled urine collections for uricacid (apart from the first sleep period).Urate was measured in one batch byautoanalyser and urine pH byradiometer.Normal diurnal variation of urate

excretion was lost during the 42 hours

without sleep and was associated witha fall in serum urate of 0-036 mmol/l(0 6 mg/100 ml). There were twopeaks of increased excretion after 22and 34 hours during this period.

After both periods of sleep therewas an increase in serum urate of0 024 mmolIl (0.4 mg/100 ml), and0 03 mmoUl (0 5 mg/100 ml). Initiallythere was fluctation of urine pH, whichthen stabilised and fell dramaticallyduring the second period of sleep frompH 6-1 ((H+) 794 nmol/1) to pH 5 35((H+) 4467 nmol/l ) and increased withfeeding to pH 6 5 ((H+) 224 nmol/l).Weight loss during the study was 1-5kg and modest negative fluid balanceoccurred during sleep and on twooccasions during the 42 hours.As reduced urate clearance should

be regarded 'as normal response of thekidneys to the endogenoushyperuricaemia (of primary gout)'45 acombination of reduced urinaryexcretion and increased enteraluricolysis suggests that there isincreased urate production at night asshown by a rise in serum urate.

Enteral uricolysis is increased atnight whether fasting or not, yet thenormal daily variation of urine urateexcretion is lost during fasting.6 Thisindicates that circadian elimination via

the gut is independent of renalfunction and food intake. As uricolysisis increased at night it is suggested thatcomplex metabolic changes occur as aresult of sleep that causesnucleoproteolysis and increased urateproduction with secondary renaleffects of increased tubularreabsorption and reduced secretion ofurate.

References1 Leathes J B. On diurnal and nocturnal

variations in the excretion of uric acid. JPhysiol (Lond) 1906; 35: 125-30.

2 Sorensen L B. The elimination of uricacid in man studied by means ofC14-labelied uric acid. Scand J Clin LabInvest 1960; suppl 54.

3 Owen-Smith B D, Whyman A. Diumaland nocturnal variations of enteraluricolysis during fasting and refeeding.[Abstract]. Ann Rheum Dis 1981; 40:523-4.

4 Gutman A B. Significance of the renalclearance of uric acid in normal and goutyman [Editorial]. Am J Med 1964; 37:833-7.

5 Gutman A B, Yu T F. Renal function ingout. With a commentary on the renalregulation of urate excretion, and the roleof the kidney in the pathogenesis of gout.Am J Med 1957; 23: 600-21.

6 Lennox W G. A study of the retention ofuric acid during fasting. J Biol Chem1925; 66: 521-71.

Familial gout, hyperuricaemia, and renal impairmentP. HOLLINGWORTH, AND J. T. SCOTT

From the Kennedy Institute of Rheumatology and Charing Cross Hospital, Hammersmith, London

We report the case of a woman whosuffered recurrent attacks of acutegout from the age of 15. When firstseen at this unit 10 years later she gavea family history of gout and was found

to be mildly hypertensive.Investigations showed hyperuricaemiauncontrolled by the dose of allopurinolshe was receiving and a glomerularfiltration rate (GFR) of 15 ml/min. On

intravenous pyelography a small rightkidney suggestive of damage byureteric reflux was seen, though thiswas not confirmed by micturatingcystography. Gamma camera



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


renography measured 80% of isotopicuptake over the normal sized leftkidney. Histological examinationof a biopsy specimen from thiskidney showed pronouncedglomerulosclerosis, tubular atrophyand fibrosis, and round cell infiltrationof the interstitium, but urate crystalswere not identified. Despite goodcontrol of hypertension andhyperuricaemia the GFR gradually fellover the next two years whenmalignant hypertension and a rapiddeterioration in renal functionsupervened and the patient has nowbeen admitted to a programme ofrenal dialysis.

Metabolic studies were performedon the patient's siblings and her closematernal relatives; the paternalrelatives were not available for study.Subjects were questioned for a historyof gout and resting blood pressure wasmeasured. Serum concentrations and24-hour urinary excretion of uratewere measured by a uricase method onsamples taken from subjects receivinga low purine diet. GFR wasdetermined by EDTA clearance. Thefamily tree is shown.Ofthe 20 subjects studied, GFR was

reduced in half (mean 52% ofexpected, range 15-74%). Five ofthese also had gout or hyperuricaemiathat was associated with decreasedrenal clearance of urate. A sixthsubject had reduced GFR and urateclearance without hyperuricaemia.These findings were detected by thesecond decade, common by the third,and most pronounced in eldermembers of this kindred. Sevensubjects with renal impairment werealso hypertensive.The familial occurrence of juvenile

E Gout

F] Hyperuricoemia

W Urate clearance <3ml /min

i GFR <75°/. predicted 1 Propositus2 Deceased

[ Resting diastolic blood 3 Allopurinolpressure >100QmmHg

Fig. 1 Family tree. Heavy lines denote subjects studied, circles females, andsquares males.

onset of gout associated with rapidlydeteriorating renal function has beenreviewed.' Inheritance appeared to bedominant and sometimes exclusive infemales. The usual pattern was ofhyperuricaemia predating or beingdisproportionate to the degree of renalimpairment, though renal failurewithout hyperuricaemia wasoccasionally observed in thesefamilies. It remains undeterminedwhether the hyperuricaemia wasresponsible for, or consequent to,disordered renal function.2 I

As the constant metabolicabnormality in this family was areduction in the GFR, with impairedurate clearance and hyperuricaemiaoccurring only in elder members, itappears that a renal abnormality is theprimary event rather than

hyperuricaemia. In contrast toprevious reports, the family history ofgout and renal failure was not strikingin the case described here and mighteasily have been overlooked. Thisfamilial association may be morecommon than suggested by the fewfamilies described.

References1 Sinmonds H A, Warren D J, Cameron J S,

et al. Familial gout and renal failure inyoung women. Clin Nephrol 1980; 14:176-82.

2 Fessel W J. Renal outcomes of gout andhyperuricaemia. Am J Med 1979; 67:74-82.

3 Rosenfeld J B. Effect of long termallopurinol on serial GFR innormotensive and hypertensivehyperuricaemic subjects. Adv Exp MedBiol 1974; 41B: 581-96.

Renal failure associated with crystal-inducednephropathy and gout in a baby boyH. A. SIMMONDS,' T. M. BARRATT,2 M. J. DILLON,2 P. C. HOLLAND,2J. R. PINCOTT,3 L. D. FAIRBANKS,' J. H. STUTCHBURY,' AND J. S. CAMERON'

From the 'Purine Laboratory and Renal Unit, Guy's Hospital Medical School, London, 2Renal Unit, Great Ormond StreetHospital for Sick Children, and 3Department ofPathology, Great Ormond Street Hospital for Sick Children, London

There has been much controversy' in the genesis of the renal lesion inover the relative roles of crystal gout, where nephropathy was formerlydeposition, vascular disease, and age common but is now extremely rare.2

Considerable debate as to whether theorigin of the lesion was interstitial3 orstemmed from intratubular deposition



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


of uric acid4 has also occurred. Wehave argued for the primacy ofintratubular crystal deposition in thepast4 based on studies using an animalmodel, as well as patients withinherited disorders resulting in goutand/or nephropathy.57 We herepresent data on severe renal damageassociated with tubulo-interstitialdeposition of uric acid/urate in aninfant8 with hypoxanthine-guaninephosphoribosyltransferase deficiency(HGPRT:EC 2.3.2.8), in the absenceof hypertension or vascular pathology.

CASE HISTORYA 5 week old boy was the first child ofhealthy unrelated parents. From 3weeks he thrived poorly, had feedingdifficulties, and was extremelyirritable. On admission the thumb andfirst two fingers of the right hand werered, swollen, and painful.Neurologically, he was slightlyhypotonic. Plasma creatinine was 350,umol/l (3.9 mg/100 ml) and plasmaurate disproportionately high at 1-13mmolI (9 mg/100 ml). However, ratioof urine uric acid to creatinine(mmol/mmol) was 1-17:1, which iswithin the normal range for a child ofthis age.8

Plain abdominal x-ray film showedno radio-opaque calculi, but renalultrasound showed both kidneys werebright, suggesting a crystalnephropathy.8 A renal biopsy showedcrystals in both tubules andinterstitium in a cryostat section underpolarised light. There was muchtubular atrophy with extensivetubular epithelial giant cell transfor-mation in the cortex and medulla.All crystals dissolved on fixing informalin.HGPRT activity was <0 01

nmol/mg Hb/h in lysed red cells, < 0 9nmol/mg protein/h in fibroblasts and<0 2% of normal in intact red cells.5 8Incorporation of labelledhypoxyanthine into nucleotides byintact fibroblasts was 6% of control,which is low on the criteria of Page etal.9 These results confirm severeHGPRT deficiency.5 8Treatment has consisted of

allopurinol 5-10 mg/kg/24 h andsodium bicarbonate. Plasma uric acidon discharge had fallen to 0 5 mmol/l(8 mg/ 100 ml), plasma creatinine to 90,umol/l (1*0 mg/100 ml). The

allopurinol dose has subsequentlyrequired careful monitoring.

COMMENTThe onset of clinical gout at 5 weeks ofage must be unique. The association ofgout and renal failure due tocrystal-induced nephropathy is alsorarely seen today." The histologicallesion is generally a non-specificinterstitial nephritis considered to bedue to age, hypertension, or both.'-3Others have defined two specificnephropathies-acute nephropathydue to intratubular uric aciddeposition (essentially reversible) asdistinct from the slow insidiousinterstitial deposition of sodium uratefrom supersaturated body fluids(essentially irreversible).'The lesion in this young child

supports our contention from animalstudies, as well as a case of 2,8-dihydroxyadenine nephropathy,4that the initial insult is intratubular,the natural sequelae of which, throughbasement membrane rupture, is themigration of crystals into theinterstitium with subsequentinflammatory response (Fig. lb)proceeding to fibrosis and permanentrenal damage.4`7

This hypothesis is supported by thefact that infants maintain a low plasmauric acid through a high urateclearance.8 Gross uric acidoverproduction in HGPRT deficiencywould result in high urinary uratewhich would provide the initial insult;the raised plasma uric acid must havebeen secondary to the subsequentrenal damage. Plasma urateconcentrations (after the immediateneonatal period) are extremely low; tothose accustomed to adult values,the raised concentrations in childrenmay thus not appear abnormal. In thischild the concentrations weredisproportionately high for the degreeof renal damage, even for an adult.7 1'This provided the only clue to thepossibility of uric acid overproductionin this infant, since the usual hallmark,high uric acid excretion on a creatininebasis, was obscured because of therenal damage.5 810Adequate control of uric acid

concentrations by allopurinol has alsopresented a problem because of therenal damage, particularly in such ayoung child. Too high a dose hasresulted in retention of oxipurinol,

which could potentiate bone marrowdepression,10 it has also resulted in theexcretion of excessive amounts ofxanthine, which is more insoluble thanuric acid.This case is important for several

reasons. Firstly, it indicates theproblems of diagnosis and treatment inHGPRT deficiency when renalfunction is impaired.5 10 The presenceof crystals within the tubules as well asthe interstitium in crystotat sectiononly-but not formalin-fixedtissue-underlines the difficulty ofidentification of uric acid-inducedcrystal nephropathy histologically,unless special precautions are taken.1The histological picture indicates thatcrystals can reach the interstitium inthe 'gouty kidney', after disintegrationof the tubules.

These studies have been supported bygrants from the MRC, ARC, WellcomeTrust, and National Fund for Research intoCrippling Diseases.

References

1 Wyngaarden J B, KelleyW N. Gout. In:Stanbury J, B, Wyngaarden J B,Frederickson D S, Goldstein J L, BrownM S, eds. The metabolic basis ofinherited disease. 5th ed. New York:McGraw Hill, 1983: 1053-6.

2 Berger L, YuT F. Renal function in goutIV. An analysis of 524 gouty subjectsincluding long-term follow-up studies.Am J Med 1975; 59: 605-13.

3 Emmerson B T, Row P G. Anevaluation of the pathogenesis of thegouty kidney. Kidney Int 1975; 8:65-71.

4 Farebrother D A, Pincott J R,Simmonds H A, Warren D J, Dillon M J,Cameron J S. Uric acid crystal-inducednephropathy: evidence for a specificrenal lesion in a gouty family. J Pathol1981; 125: 159-68.

5 Dillon M J, Simmonds H A, BarrattT M, Fairbanks L D, Holland P C. Prob-lems in diagnosis and treatment ofadenine and hypoxanthine-guaninephosphoribosyltransferase deficiency. JClin Chem Clin Biochem 1982; 20: 364.

6 Simmonds H A, Warren D J, CameronJ S, Pbtter C F, Farebrother D A. Familialgout and renal failure in young women.Clin Nephrol 1980; 14: 176-82.

7 Cameron J S, Simmonds H A. Uric acidgout and the kidney.J Clin Pathol 1981;34: 1245-54.

8 Holland P C, Dillon M J, Pincott J,Simmonds H A, Barratt T M. Hypo-xanthine guanine phosphoribosyl-transferase deficiency presenting asrenal failure in infancy. (in press).



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


9 Page T, Bakay B, Nissinen E, Nyhan W L.Hypoxanthine-guanine phospho-ribosyltransferase variants: correlationsof clinical phenotype with enzyme activ-ity. Journal of Inherited Metabolic Dis-eases 1981; 4: 203-7.

10 Cameron J S, Simmonds H A, WebsterD R, Wass V, Sahota A. Problems ofdiagnosis and in vitro enzyme instabilityin an adolescent with hypoxanthine-guanine phosphoribosyltransferasedeficiency presenting in acute renal fail-

ure. J Clin Chem Clin Biochem 1982;20: 353.

Lean, dry gout patientsL. GAIL DARLINGTON

From the Epsom District Hospital, Epsom, Surrey

Hypertriglyceridaemia is common ingout.' 2 Obesity may increasetriglyceride concentrations, as mayalcohol,3 but it remains uncertainwhether obesity and alcohol, alone orin combination, are sufficient toexplain the hyperlipidaemia in allcases. We looked for abnormal lipidconcentrations in non-obese patientswith gout who drank little or noalcohol to determine whether thehyperpre betalipoprote inae miaassociated with gout occurred in such alean and abstemious group.

All patients were of desirable weightor less for their age and frame4 and nopatient drank more than one pint ofbeer per day or its equivalent. Suchpatients are rare, and only seven werefound in four years from a busy clinic.

Fasting concentrations of lipid andlipoprotein were measured in serum at

a laboratory with its own controlpopulation.5 Serum uric acidconcentrations were determined forthe patients with gout butunfortunately data for the controlpopulation were not available.

Readings for serum cholesterol,triglyceride, /-lipoprotein and prebetalipoprotein concentrations in thepatients with gout lay within 2standard deviations of thecorresponding mean for controls. Thismeans that they were within the 95%confidence limits for the controlpopulation and it is therefore unlikelythat there is any real differencebetween the patients with gout andcontrol populations.

In spite of the small numbers ofthese, 'lean, dry' patients, the resultsrevealed no intrinsic hyperlipidaemiain subjects with gout when obesity and

an excess of alcohol were removed ascauses of hypertriglyceridaemia.

References1 Darlington L G, Scott J T. Plasma lipid

levels in gout. Ann Rheum Dis 1972; 31:487-9.

2 Mielants H, Veys E M, de Weerdt A.Gout and its relations to lipidmetabolism. I. Serum uric acid, lipid andlipoprotein levels in gout. Ann RheumDis 1973; 32: 501-5.

3 Gibson T, Grahame R. Gout andhyperlipidaemia. Ann Rheum Dis 1974;33: 298-303.

4 Metropolitan Life Insurance Company.Build and blood pressure study. Chicago:Society of Actuaries, 111. 1959.

5 Slack J, Noble N, Meade T W, NorthW R S. Lipid and lipoproteinconcentrations in 1,604 men andwomen in working populations innorth-west London. Br Med J 1977; ii:353-6.

Cardiovascular disease and gout: a function of sex andage?D. G. MACFARLANE

From the University Department ofMedicine, Bristol Royal Infirmary, Bristol

Hypertriglyceridaemia is common inpatients with gout and hyperuricaemiabut it is still not known whether thisresults from a link between purine andlipoprotein metabolism or whetherthey occur together due to otherassociated facts, particularly obesityand high alcohol intake, both of whichare commonly found in patients with

hyperuricaemia.' Nor is it firmlyestablished whether patients with goutare, in fact, predisposed to prematurecardiovascular disease, and, if so,whether the raised serum uric acidconcentration operates as anindependent risk factor, or only via itsassociation with hypertrigly-ceridaemia and hypertension, which

are established as cardiovascular riskfactors in their own right.2 Anotherpossibility-namely, that a raisedserum uric acid concentration causesplatelet hyperaggregatability andhence thrombosis-has recently beeninvestigated in our unit with negativeresults.3Over a period of two and half years



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


60 patients have presented to ourrheumatology department with gout:51 men, mean age 55-2 years andmean age of onset 46-6 years and ninewomen, mean age 80 years and mean

age of onset of 78-6 years. We studiedthe prevalence of hyperlipidaemia,hypertension, obesity, andcardiovascular disease in these twogroups.

Eighteen men but none of thewomen had a significanthyperlipidaemia. Triglycerides were byfar the most common lipids (type IV)to be raised, being found in 13 of the18. Four had hypercholesterolaemia:two type IIa and two type Ilb (serumtriglycerides also raised). One patienthad a pronounced rise of serumtriglyceride and cholesterol due toaccumulation in serum of intermediatedensity lipoproteins, type IIIhyperlipidaemia or broad ,8-disease.

Six men described themselves as

teetotal, but three of these had a

hyperlipidaemia. One had type Ila andhad had a myocardial infarction at theage of 42, the other two had type IV.One of these, the hospital's applianceofficer, was within the range of hisideal body weight and therefore

corresponded to those patientsdescribed by Dr Darlington (p. 90)with lean, dry gout. The other was

massively overweight at 144 kg. Noneof the women drank on a regular basisbut several did admit to the occasionalsherry at Christmas, or the odd glass ofGuinness as a 'tonic'. Only one of thewomen was above her ideal bodyweight, the remainder tending to bethin and frail.Twenty men and two women had

hypertension, defined as a restingdiastolic blood pressure above 100mmHg on two or more occasions, or

receiving established treatment forhypertension. The patients were

questioned about history of myocar-dial infarction and symptoms ofangina. They were examined for signsof peripheral vascular disease and allunderwent electrocardiography. Onthis basis 10 men and three women hadappreciable cardiovascular disease.This did not correlate with hyper-lipidaemia except in the two patientswith type hIa, and nor did it correlatewith hypertension.These observations are on a small

number of patients and the study was

uncontrolled, but the interesting fact

that emerges is the great disparity inthe ages of the men and women

presenting with gout and the virtualabsence of associated factors such asobesity, hypertension, andhyperlipidaemia in the elderly women.This could be interpreted as implyingtwo quite separate disease processes inthe two groups. It may be significant inthis regard that all the women were

receiving potent diuretic treatment.An alternative explanation could bethat women are much more prone tocardiovascular disease in associationwith gout and many had died at a

younger age; this, however, is contraryto clinical experience.

References

1 Grahame R, Scott J T. Clinical survey of354 patients with gout. Ann Rheum Dis1970; 29: 461-7.

2 Persky T W, Dyer D R, Idris-Soven E, etal. Uric acid: a risk factor for coronaryheart disease? Circulation 1979; 59:969-76.

3 Macfarlane D G, Slade R, Hopes P A,Hartog M. A study of plateletaggregation and adhesion in gout. Clinicaland Experimental Rheumatology 1982 (inpress).

Study of blood coagulation in gout patientsD. A. LANE, AND L. G. DARLINGTON

From the Charing Cross Hospital, London, and Epsom District Hospital, Epsom, Surrey

Patients with gout have vascular riskfactors such as hyperlipidaemia,hypertension, 'type A' personality,and, possibly, hyperuricaemia,obesity, and glucose intolerance. Weused a modern method to seekabnormalities of coagulation inpatients with gout.Twelve fasting men with primary

gout were studied. Patients were askedto take no medication, particularlyaspirin, for the two weeks precedingthe tests. No patients were includedwithin three months of trauma or

surgery and an atraumaticvenepuncture technique was used.Assays were performed to

fibrinopeptide A (FpA)' with

Table 1 Comparison ofmean data from patients with gout and fromcontrols

Controls Patients Significancewith gout

FpA (pmol/1):No. 13 10Mean 0-98 1-53 NSSD 0 29 0 77

FpB,81-42 (pmol/l)No. 10 9Mean 1-65 2-92 NSSD 0 63 1i85

,3TG (pmoVI)No. 9 12Mean 1-00 1 58 NSSD 0 28 1 81

NS = Not significant.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


modifications,2 to fibrinopeptidef3,81-42) (FpB,81-42)3 withmodifications2 and to,8-thromboglobulin (,ITG)4 withtnodifications.2 Serum cholesterol andtriglycerides were measured direct5while high density lipoproteincholesterol (HDLC) was estimated-by precipitation of 13-andprebeta-lipoproteins andmeasurement of cholesterol in thea-lipoprotein.'

Data from these patients werecompared with data from age-matchedcontrols (Table 1). No significantdifferences between patients with goutpnd controls were found.Spearman rank correlation

coefficients were calculated betweenFpA, FpB,81-42 and BTG and lipidconcentrations, alcohol consumption,and weight as the data on alcohol andweight tend not to be distributed

normally. There were only twosignificant correlations--that is,cholesterol in mmol/l with FpA inpmolll and HDLC in mmol/l withFpB,/1-42 in pmolIl (p<0- 05).Correlation results must be

considered with care as this was a studyon small numbers of patients.Furthermore, it must be rememberedthat abnormalities of coagulationdetected in vitro are not necessarilypresent in vivo nor do they necessarilyhave a causal relationship withthrombosis.We concluded that modern methods

to assess coagulation and plateletfunction did not show any significantabnormalities in patients with primarygout.

References

1 Nossel H L, Yudelman 1, Canfield R E, et

al. Measurement of fibrinopeptide A inhuman blood. J Clin Invest 1974; 54:43-53.

2 Lane D A, Ireland H, Wolff S, Boots M,Pegrum G D. Simultaneousmeasurement of thrombin and plasminactivities and platelet releasing stimuli inplasma. Br J Haematol 1981; 47: 630.

3 Nossel H L, Wasser J, Kaplan K L, LaGamma K S, Yudelman I, Canfield R E.Sequence of fibrinogen proteolysis andplatelet release after intra-uterineinfusion of hypertonic saline.J Clin Invest1979; 64: 1371-8.

4 Bolton A E, Ludlam C A, Moore S,Pepper D S, Cash J D. Three approachesto the radioimmunoassay of humanl3-thromboglobulin. BrJ Haematol 1976;33: 233-8.

5 Smith L, Lucas D, Lehnus G. Automatedmeasurement of total cholesterol andtriglycerides, in 'tandem', on the DiscreteSample Analyser Guildford System3500. Clin Chem 1979; 25: 439-42.

Long term comparison of azapropazone withallopurinol in control of chronic gout andhyperuricaemia

J. S. TEMPLETON

From the International Clinical Research Department, A. H. Robins Co Limited

Allopurinol is the standard agent usedto lower serum urate concentrationand does so by its ability to inhibitxanthine oxidase. Azapropazone is anon-steroidal anti-inflammatory drug(NSAID) that has been shown toinhibit urate monohydratecrystal-induced inflammation in therat and in man,' to reduce serumurate concentration in patients withrheumatoid arthritis, chronic gout, andhyperuricaemia2 and to be effective inthe treatment of both acute andchronic gout.3

In this study, 24 separateinvestigators recruited suitablepatients with chronic gout orhyperuricaemia or both who had beenreceiving allopurinol treatmentsatisfactorily for at least three monthsbefore entry to the study. Patientswere then allocated to one of twotreatment groups, either continuing to

receive allopurinol and followed upevery two months for up to six months,or starting azapropazone 600 mgtwice daily and seen roughly everymonth for up to six months. Bothgroups were comparable in respect ofage, sex, duration of disease,allopurinol dose at entry, diagnosis,and eligibility for the study.Data is so far available on 155

patients who changed toazapropazone and 122 whocontinued to receive allopurinol.Control of serum urate concentrationwas comparable in both groups (Table1). In both groups, serum urateremained remarkably constant withtwo points worthy of mention.

(1) Azapropazone lowered urateconcentration slightly after two weeksand indeed further studies are plannedto see if this in fact occurs earlier thantwo weeks.

(2) There is no evidence of increasein serum urate concentrations in thosepatients who changed from allopurinolto azapropazone. This is contrary tothe findings in a group of patients inwhom allopurinol treatment was sud-denly discontinued.4

During the study period acuteattacks of gout were less frequent inthose patients taking azapropazonethan in those who continued to takeallopurinol. Indeed, an 'attack rate'during the first four weeks of study inthose patients from both groups whocould be classed as susceptible-thatis, those with positive history of acutegout-showed no acute attacks in 109patients receiving azapropazone andsix acute attacks in 103 patientsreceiving allopurinol.Adverse effects, particularly

relating to the upper gastrointestinaltract, occurred more often in patients



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


receiving azapropazone and there wasalso a higher prevalence of abnormalrenal function (judged by raised bloodurea or serum creatinine or both) inthis group. Analysis of the data onrenal function, however, reveals thatin the group treated withazapropazone the increase in urea orcreatinine occurred principally duringthe first two weeks of treatment anddid not rise further.

In addition, out of 32 patientschanged to azapropazone who werereceiving additional analgesics at thestart of the study, 22 wereable to stop this analgesic(s), 10continued them, and five morehad to start taking-usuallyintermittently-additional analgesics.Of 39 patients who continued takingallopurinol, nine were able to stopadditional analgesics, nine continuedon the same additional analgesics asbefore, and 21 had to start treatmentwith an additional analgesic.

Thus, azapropazone would appearto be a useful alternative drug for thelong term control of patients withchronic gout or hyperuricaemia, orboth. Its two possible importantadvantages are that it is a single drug

Table 1 Serum urate concentration oftreated patients (normal values O14-045mmol/l)

Time (weeks)0 4 8 12 16 24

Azapropazone:No of patients 133 122 106 108 90 96Serum uric acid (mmol/1):Mean 0-385 0 375 0-383 0-388 0-401 0-396SD 0-114 0 094 0-101 0-098 0-110 0-120

Allopurinol:No of patients 116 113 105 105Serum uric acid (mmol/l):Mean 0-362 0-341 0-328 0 345SD 0-150 0-106 0-084 0-103

treatment for acute and chronic goutand it provides a considerable measureof analgesic effect, which allopurinolcannot do, for those patients who havepainful concurrent degenerative jointdisease such as osteoarthritis, cervicalspondylosis, backache, etc.

References

1 Dieppe P A, Doherty M, Whicher J T,Walters G. The treatment of gout withazapropazone: clinical and experimental

studies. European Journal ofRheumatol-ogy andInflammation 1981; 4: 392-400.

2 Thomas A L, Majoos F L, Nuki G. Pre-liminary studies with azapropazone ingout and hyperuricaemia. In: Fen-clofenac in arthritis. Proceedings of 19thEuropean Congress of Rheumatology,Wiesbaden 1979. London: Royal Societyof Medicine.

3 Frank 0. Investigation of the uricosuricaction of azapropazone. Z Rheumatol1971; 30: 368-73.

4 LoeblW Y, Scott J T. Withdrawal of allo-purinol in patients with gout.Ann RheumDis 1974; 33: 304-7.

Gout with apparent resistance to allopurinolT. GIBSON, H. A. SIMMONDS, A. V. RODGERS, G. J. HUSTON, D. R. WEBSTER,J. MUNRO, AND P. EMERY

From Guy's Hospital, London SEJ

The dose of allopurinol necessary toinduce sustained normouricaemia ingout rarely exceeds 300 mg daily. Wedescribe a patient with apparentresistance to conventional doses ofallopurinol and probenecid. The studywas designed to evaluate the responseto varying doses of allopurinol and todetermine whether thehyperuricaemia and the reported lackof effect of allopurinol were related tosome unusual disturbance of purinemetabolism.

CASE REPORTA 51 year old man had experiencedintermittent episodes of goutconfirmed by the demonstration of

intra-articular urate crystals for 20years. Various combinations ofanti-inflammatory and hypouricaemicdrugs had failed to exert a detectableeffect on his recurrent arthritis. At thetime of referral he was obese butexhibited no tophi and wasnormotensive. One brother had ahistory of gout. His alcoholconsumption was modest, though hehad transient liver dysfunction; liverbiopsy showed no abnormality.Glomerular filtration rate (5"CrEDTA) was normal at 100ml/min/1 73 m2 but his urine was ofconstant pH and renal concentratingability was impaired. He was admittedto hospital on three separate occasions

for investigation. After four days on alow purine, caffeine free diet heunderwent five periods of treatmentwith allopurinol in doses increasingfrom 200 mg to 1200 mg.Each treatment period lasted a

further four days except the last ( 1200mg/day), which was curtailed after twodays. The studies were conducted in ametabolic ward. Table 1 shows theresults of daily investigationsconducted before and during eachtreatment period. Estimations of uricacid, purines, and allopurinolmetabolites were performed aspreviously described.' 2 Red celllysates were investigated for enzymedeficiencies, which are ackniowledged



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Table 1 Results ofinvestigations during three hospital admissions. On each occasioni, the patienit received a lowZ purine dietfbr four days, without treatment, followed by varying doses of allopuriniol

Dose of' Plasia Uriniary Urate Total urine Uritne Allopurinol Plasmaallpuriniol uric aci(d uric acid clearance oxypurinees hspoxianthiie. tmetabolites oxipurinol(ing) (mtnol/1) (mmol 24h) (mllnin) (mmol/24h) ranthine ratio ("¼/ of dose) (p.mol/l)

() 0 70 4 7 4 7 4-96 3. 3 -

20 0 76 3 4 3 1 3 92 1 5 260 67 4 2 4 4 4 3 3 5 - -

300 0 86 1 5 1 2 1 74 2 4 8 17600 0- 85 1 92 1 6 2 22 1 6 7 50

0) 0-66 4 3 4 5 4 43 0 83 - -

9(( 0 45 4 4 6 9 4 82 0(66 7 20120(0( 0 28 3 4 8 4 4 74 0 45 18 35

Conversion.SI to traditional units- Uric acid: I mmoV!1 - 17 mg/100 ml; 1 mmoVI24 h- 0 17 mg/24 h.

causes of hvperuricaemia. Red cellnucleotide concentrations were alsoestimated.'On a low purine diet, plasma uric

acid concentration was always grosslyraised. Allopurinol in daily doses up to600 mg caused a further rise in plasmaurate wvhich was associated with a fallin excretion of uric acid and aconcomitant reduction of urateclearance. Urate clearance correctedfor creatinine clearance (not shown)also fell, suggesting that the rise inblood uric acid was secondary toimpaired clearance. By contrast, withdoses of 900 to 1 200 mg plasma uratefell and urate clearance increased.During the first two admissions theratio of urine hypoxanthine toxanthine was raised before treatment(normal - 1-5:1), suggesting increasedde novo purine synthesis. Normally,allopurinol treatment results in anincrease of urine hypoxanthine andespecially of xanthine with a fall oftheir ratio to values of ' 1 0:1. Thiswas not observed, and the percentageof total allopurinol metabolitesrecovered in the urine wasexceptionally low. On the other hand,plasma oxipurinol concentrationswere consistent with those expected ofpatients receiving allopurinol.

Hypoxanthine guanine phospho-ribosyl transferase (HGPRT), andphosphoribosylpyrophosphate synth-etase (pPribose P synthetase) activitieswere normal as were erythrocytenucleotide concentrations.

COMMENTWe have previously reported that thispatient was unresponsive toconventional doses of allopurinol. 3It isnow apparent that high doses of thedrug will reduce his bloodconcentration of uric acid. Others havenoted that the daily dose of allopurinolnecessary to produce nomouricaemiaoccasionally exceeds 300 mg.4 butdoses up to 1000 mg are sometimesnecessary.5 The anomalous resultsobtained in this study raise severalquestions. The low recovery of urineallopurinol metabolites suggests thatthe patient failed to ingest theallopurinol (unlikely under conditionson a metabolic ward) or did not absorbit. As plasma oxipurinolconcentrations were in accord withthose normally achieved afterconventional allopurinol doses it isalso possible that there was a failure toexcrete metabolites in the usualfashion. The paradoxical rise in plasmauric acid on lower doses of allopurinolappeared to be caused by a furtherreduction of an already reduced urateclearance. The high hypoxanthineconcentrations and the increasedhypoxanthine to xanthine ratio suggestthat the patient was an overproducerof uric acid. However, the normalenzyme and nucleotide concentrationswould exclude any of the knowncauses of purine overproduction.The lack of response to allopurinol

and the low concentrations of urinemetabolites are unique observations.

Investigation of other patientsrequiring large doses of allopurinolmay reveal identical anomalies.Patients whose unresponsiveness toallopurinol has been attributed to poorcompliance may also be similarexamples. It remains to be seenwhether the patient described here isrepresentative of a previouslyunrecognised group of patients withgout.

References

1 Simmonds H A. Unrnary excretion ofpurines, pyrimidines and pyrazolo-pyrimidines in patients treated withallopurinol and oxipurinol. Clin ChimActa 1969; 23: 353-64.

2 Simmonds H A, Warren D J, CameronJ S, et al. Familial gout and renal failure inyoung women. Clin Nephrol 1980; 14:176-82.

3 Simmonds H A, Webster D R, Wilson J,Lingham S. An x-linked syndromecharacterised by hyperuricaemia, deaf-ness and neurodevelopmental abnor-malities. Lancet 1982; ii: 68-70.

4 Simmonds H A, Gibson T, Huston G J, etal. Gout resistant to allopurinol therapy:poor compliance or non-response. J ClinChem Clin Biochem 1982; 20: 417-8.

5 Scott J T, Hall A P, Grahame R. Allo-purinol in the treatment of gout. Br MedJ1966; ii: 321-7.

6 Rundles R W, Metz E W, Silberman H R.Allopurinol in the treatment of gout.AnnIntern Med 1966; 64: 229-58.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Coexistent tophaceous gout and ankylosing spondylitisI. PORTEK, B. P. WORDSWORTH, AND A. G. MOWAT

From the Rheumatology Unit, Nuffield Orthopaedic Centre, Oxford

There has recently been considerableinterest in the rare coexistence of goutand other inflammatory arthridites.'3We report the first case of tophaceousgout and anklyosing spondylitis.

CASE REPORTA 62 year old man was admitted inJune 1982 with persistently activepolyarticular tophaceous gout andpronounced pedal oedema due tocongestive cardiac failure.The first attack of right podagra in

1969 was followed by progression to asymmetrical chronic arthritis, initiallyaffecting only the lower limbs, and hewas treated with various non-steroidalagents as well as colchicine. In 1976arthritis developed in severalpredominantly distal interphalangealjoints of both hands associated withtophi. Allopurinol was added to histreatment regimen, resulting inexfoliative dermatitis. Upper limbinvolvement progressed to wrists,elbows, and shoulders, sparing allmetacarpophalangeal joints. Furthersymptomatic deterioration followed amyocardial infarction in 1978 with theintroduction of diuretic treatment.There was persistent hyperuricaemiaand enlargement of tophi. He wasplaced on high doses of probenicid,which was changed to sulphinpyrazoneafter a poor response.

In 1965 he had suffered from lowback pain radiating to his thighs.Although symptoms were attributedto mechanical back pain and a corsetwas prescribed, review of theradiographs showed bilateralsacroiliitis. Pain improved over thenext 18 months and axial symptomshad only recurred in the past few years,within neck stiffness and a stoopedposture.On examination he was overweight

with a pronounced dorsal kyphosis.Movements of the cervical and lumbarspine were considerably limited. Chestexpansion was 1 cm. Most jointsexcept metacarpophalangeal and hipjoints showed chronic inflammatoryarthritis. The interphalangeal joint ofhis right thumb was acutelyinflamed. Tophi were present onseveral digits and on the pinna of theears. There was cardiomegaly but noevidence of aortic incompetence orheart failure. Pitting oedema to hisknees was associated with eczematousskin changes.X-ray films of the hand showed

tophaceous gouty arthritis and thesacroiliac joints showed typicalchanges of progressive ankylosingspondylitis. Before admission theserum uric acid was persistentlyraised-for example, 0-778mmol/l-at admission it was 0 484

mmolIl. 24 hour urate excretion was4-8 mmol (3.5-4.2 mmol) withoutpurine restriction and taking sulphin-pyrazone 200 mg four times daily.Blood urea was raised (9 4 mmol/l(2*5-6.7 mmolIl)) and creatinineclearance decreased (45 ml/min (105ml/min)). He was HLA-B27 positive.The absence of previous reports of

the coexistence of these two diseases issurprising considering their highincidence in men, their relativefrequency in the population and thefact that neither joint involvement norserological examination would resultin the diagnosis of one condition inpreference to the other.

References

1 Atdjian M, Fernandez-Madrid F.Coexistence of chronic tophaceous goutand rheumatoid arthritis. J Rheumatol1981; 8: 989-92.

2 Helliwell M, Crisp A J, Grahame R.Coexistent tophaceous gout andsystemic lupus erythematosus.Rheumatol Rehabil 1982; 21: 161-3.

3 Wall B A, Agudelo C A, Weinblatt M E,Tume R A. Acute gout and systemiclupus erythematosus: report of 2 casesand literature review. J Rheumatol 1982;9: 305-7.

Normal response to monosodium urate (MSU)crystals by patients with rheumatoid arthritisMICHAEL DOHERTY, JUNE HORNBY, AND PAUL A. DIEPPE

From the University Department ofMedicine, Bristol Royal Infirmary, Bristol BS2 8HW

It has been suggested that the negativecorrelation between rheumatoidarthritis and gout might resultfrom inability of patients with

rheumatoid arthritis to respond to exclusion from interaction withmonosodium urate (MSU) crystals, cellular and non-cellular mediators ofperhaps due to crystal coating by inflammation.' To test the validity ofrheumatoid factors with subsequent this hypothesis we investigated the



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


ability of patients with rheumatoidarthritis to respond to preformedcrystals using both in vivo and in vitrotechniques.

Needle-shaped MSU crystals 2-10,um long were prepared bymodification of the method ofSeegmiller et al. 2; chemical purity waschecked by infrared spectro-photometry. Crystals were suspendedin sterile saline (skin testing and C3activation) or Eagle's medium (poly-morphonuclear (PMN) interaction)and briefly sonicated before use toreduce aggregation.

(1) Intradermal skin testing. Thediameter of erythema produced byintradermal injection of MSU crystals(5 mg in 0 2 ml suspension) wasmeasured by a single observer aspreviously described.' Fifteen patientswith seropositive rheumatoid arthritis(mean age 49-1 years, mean DAT1:640), 15 patients with osteoarthritis(mean age 63-8 years), and 12 normalcontrols (mean age 26-2 years)showed no significant difference inmean forearm erythema 24 and 48hours after injection (patients withrheumatoid arthritis: 29-8±6-2 mmafter 24, 27-8 ± 7-3 after 48; patientswith osteoarthritis: 32-4 ± 4-4 and

14-7 + 81; and controls 122 + 6,10 8+ 5-1).

(2) Polymorph phagocytosis andenzyme release. PMNs from patientswith rheumatoid arthritis (n = 1 0) andcontrols (n = 10) exposed to MSUcrystals (1%) for 1 hour showed nosignificant difference in cellattachment/phagocytosis, enzymerelease or viability.

(3) Complement activation. SerumC3 activation produced by 30 minutes'incubation with MSU crystals (1'Yo)was quantified by densitometry afterelectrophoretic separation andimmunofixation as previouslydescribed.4 No difference wasobserved between 10 patients withrheumatoid arthritis (mean DAT1: 640) and 10 normal sera in the 'Yo ofC3 split in excess of saline control(37-8 - 4 9% and 36-7 -+- 32%respectively). C3 activation via thealternative pathway-demonstratedby incubation with EGTA (8 mmol)and MgCl2 (0.3 mmol)-was alsosimilar (21-1 + 4-2% and 18 5 +3-1% respectively).These results show that patients

with seropositive rheumatoid arthritismount normal in vitro and in vivoinflammatory responses to MSU

crystals. We therefore discountprevious suggestions that the negativecorrelation between rheumatoidarthritis and gout results from adifference in host response. A moreplausible explanation is that theconnective tissues of patients withrheumatoid arthritis are not conduciveto MSU crystal formation.We would like to acknowledge the

financial support of the Arthritis andRheumatism Council.

References1 Wallace D J, Klinenberg J R, Morhairn

D, Berlanstein B, Biren P C, Callis G.Coexistent gout and rheumatoid arthritis:case report and literature review. Arth-ritis Rheum 1979; 22: 81-6.

2 Seegmiller J E, Howell D R, MalawistaS E. The inflammatory reaction tosodium urate. JAMA 1962; 180: 469-77.

3 Dieppe P A, Doherty M, PapadimitriouG M. Inflammatory responses to intra-dermal crystals in healthy volunteers andpatients with rheumatic diseases.Rheumatology International 1982; 2:55-8.

4 Doherty M, Whicher J T, Dieppe P A.Activation of the alternative pathway ofcomplement by monosodium uratemonohydrate crystals and other inflam-matory particles. Ann Rheum Dis (inpress).

Ultrastructural observations of crystals in articularcartilage of aged human hip jointsIRENIA MARANTE, R. MACDOUGALL, A. ROSS, AND R. A. STOCKWELL

From the Department ofAnatomy, University Medical School, Edinburgh EH8 9AG

The occurrence of calcification in thenormally uncalcified zones of humanarticular cartilage is well documented'and crystal morphology at theultrastructural level in osteoarthriticcartilage has been described by earlierworkers.2 We showed the presence ofcrystals, probably apatite, inmacroscopically normal articularcartilage of the aged human hip joint.

Femoral heads were obtained fromeight women undergoing operation forhip arthroplasty after subcapitalfracture. Ages ranged from 65 to 91years and in four cases fracture of thefemoral neck had occurred the day

before operation. The cartilage of thewhole femoral head appeared smooth,though in several specimens a smallarea about 3-4 mm diameter beneaththe fovea showed minimal fibrillation.Tissue blocks of smooth cartilage fromthe superior and inferior surface of thefemoral head were fixed inglutaraldehyde, stained with osmicacid, and embedded in araldite. Thinsections were stained with uranylacetate and lead citrate and examinedby transmission electron microscopy.Unstained sections of non-osmicatedtissue fixed in unbuffered 10%formaline were also examined.

Crystals were found in allspecimens, occurring singly oroccasionally in clusters. They weremostly 0-2 Am in diameter but variedfrom about 0-05 to 15 um. They wereoften polygonal or rounded in shapebut many had a sharply angled squarecontour and were considered to becuboidal. Several crystals wereenclosed by an electron dense lamina,possibly a membrane, adherent to thesurrounding cartilage matrix. Many'empty' areas lined by a similar denselamina were observed; thesecorresponded in size and shape to thecrystals. This may have been due to



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


loss of crystals occurring either duringtissue preparation or in vivo. Crystalswere situated either randomly in thematrix, many of them within a fewmicrometers of the articular surface,or around the cells. Although thenumber of crystals varied, this did notseem to be related to the time intervalbetween the fracture and the samplingof the tissue.X-ray emission microanalyses of

crystals (n = 10) in formalin fixedsections showed calcium andphosphorus. The ratio of calcium tophosphorus, uncorrected forabsorption, fluorescence, and atomicweight was 1-95 (SD 0 30), suggestiveof apatite rather than pyrophosphatedeposits.

Ultrastructurally articular surfaceswere often roughened, though smoothto the naked eye, and showedinfiltration with electron densematerial. They were only rarelyfissured. In many specimens thesuperficial collagen fibrils retained thenormal density of packing andorientation tangential to the surface.More deeply, the fibres could showsome disorganisation. Flattenedchrondrocytes were often present inthe superficial zone. Chrondrocytes indeeper tissue contained irregularlobed nuclei and several lipid dropletsin addition to active Golgi andendoplasmic reticulum.

Crystal deposition in the hip jointsof these women may be related to

osteoporosis, and this requires furtherinvestigation. Localisation of thecrystals within the cartilage(superficial zone and pericellularregions) corresponds to sites ofextracellular lipid in aging articularcartilage generally. Perhaps the mostnoteworthy features are theoccurrence of crystals in non-arthritichip joints, and their occurrence in allthe elderly women in this group.

References

1 Dieppe P A. New knowledge ofchrondrocalcinosis.J Clin Path 1978; 31,suppl 12: 214-22.

2 Ali S Y. In: Ascenzi E, Bonucci E, deBemard B, eds. Matrix vesicles. Milan:Wichtig, 1981: 241-7.

Ultrastructural studies of pyrophosphate crystaldeposition in articular cartilageS. Y. ALI,' S. GRIFFITHS,1 M. T. BAYLISS,' AND P. A. DIEPPE2

From the 'Institute of Orthopaedics, Royal National Orthopaedic Hospital, Stanmore, Middlesex, and 2Bristol RoyalInfirmary, Bristol

The mechanism of calciumpyrophosphate crystal formation inthe articular cartilage of patientssuffering from this crystal depositiondisease has not yet been fullyelucidated. It has been suggested thatthe earliest pyrophosphate crystaldeposition is adjacent to chondrocytelacunae' and has been supported bylight microscopical studies.Ultrastructural studies have also beenperformed on cartilage from patientswith chondrocalcinosis but they havebeen restricted to the study of thechanges in the matrix.2 We haveembarked on an electronmicroscopical study of articularcartilage specimens from patients withcalcium pyrophosphate crystaldeposition disease and have combinedit with electron probe analysis of thecrystals to determine their chemicalcomposition.By using modified staining

techniques it is now possible to showcrystals in ultrathin sections ofcartflage. Large islands of crystals wereseen in the matrix, confirming earlier

' I

I.,

V.

_,

It

5pm I

Fig. 1 Electron microscopical picture oftwo chondrocytes in the matrix ofarticular cartilage specimen obtained from the knee ofa patient with calciumpyrophosphate crystal deposition disease (70 year old woman). Note that some ofthe crystals are adjacent to the chondrocyte and are present within the lacuna. Boththe cells have an abnormal amount ofglycogen in the cytoplasm. The appearanceofthe matrix is also abnormal. 'xS500 (original magnification).



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


observations on light microscopy.More interestingly, smaller crystalswere seen pericellularly and in theintercellular matrix. Some crystalswere seen adjacent to thechondrocytes and located in thelacuna (Fig. 1). Electron probeanalysis of the crystals in variouslocations, and differing in size, gave aconsistent calcium to phosphorus ratioof approximately 1-15: 1 similar to thatof the calcium pyrophosphatestandard. This ratio is quite differentand much lower than the ratio ofcalcium and phosphorus in the threedifferent types of apatite crystalsfound in human osteoarthriticarticular cartilage.4 5There were various non-specific

changes in the cells and in the matrixcomponents that were indicative of arapidly degenerating cartilaginoustissue. Thus there were focal changesin the staining of the collagen andproteoglycan particles which gave thematrix, in some areas, a more granularappearance unlike the ordered,fibrous nature of normal cartilagematrix. In other areas there was aband-like, collagenous encapsulation

of the lacuna area, which containedseveral cells in a cluster. Extracellularmatrix vesicles were absent in manyareas but were sometimes seen pericel-lularly in focal areas where chondro-cyte degeneration and necrosis wasevident.

In terms of chondrocytemorphology two changes wereparticularly noticeable. Firstly, somecells showed abnormally large islandsof glycogen in the cytoplasm (Fig. 1).Secondly, other chondrocytes showedabnormal increase in the amount ofrough endoplasmic reticulum. If theseobservations are true it may implyincreased synthesis of glycogen andother cartilage constituents. Inorganicpyrophosphate is a byproduct of manymetabolic activities, and biosyntheticpathways, and this has been implicatedby others as a mechanism for theformation of excess pyrophosphate insome tissues and fluids. From ourobservations we can only suggest thatthe calcium pyrophosphate crystalsseen in the vicinity of chondrocytesmay be formed by the interaction of acellular product and some matrix ionor component. Further studies on the

quantitative analysis of the level ofpyrophosphatase-type enzymes innormal and degenerative cartilage areunder way to see if the accumulation ofpyrophosphate in the tissue is due tothe lack of a specific degradativeenzyme.

References

1 McCarty D J. Pseudogout. In: HollanderJ L, McCarty D J, eds. Arthritis and alliedconditions. Philadelphia: Lea andFebiger, 1972: 1140-60.

2 Bjelle A 0, Sundstrom B K G. Anultrastructural study of the articularcartilage in calcium pyrophosphatedihydrate (CPPD) crystal depositiondisease (chondrocalcinosis articularis).Calif Tiss Res 1975; 19: 63-71.

3 Schumacher H R. Ultrastructuralfindings in chondrocalcinosis andpseudogout. Arthritis Rheum 1976; 19:413-25.

4 Ali S Y, Griffiths S. New types of calciumphosphate crystals in arthritic cartilage.SeminArthritisRheum 1981; 11, suppl 1:124-6.

5 Ali S Y. New knowledge of osteoarthritis.J Clin Pathol 1978; 31, suppl 12: 191-9.

Studies of pyrophosphate metabolism in relation tochondrocalcinosisA. M. CASWELL, M. K. B. McGUIRE, AND R. G. G. RUSSELL

From the Department ofHuman Metabolism and Clinical Biochemistry, University of Sheffield Medical School, SheffieldS10 2RX

In an attempt to understand thepathogenesis of chondrocalcinosis wehave examined various aspects of themetabolism of inorganicpyrophosphate (PPj). Using a highlysensitive radiometric assay,' wedetermined the concentration of PPi inserum and plasma and in cultured cellsderived from normal individuals andfrom patients with chondrocalcinosisto identify any abnormality in PPimetabolism. We also examined theintracellular metabolism of PP1 incultured articular chondrocytes andmeniscus cells derived from normalindividuals to define the factors thatinfluence the intracellular production

and breakdown of PPi. In addition, westudied the extracellular metabolismof PPi by these cells to define whetherPP, can be produced outside cellsand/or can cross the cell plasmamembranes.

In agreement with the findings ofearlier studies,2 I we observed nodifferences in the serum and plasmaconcentrations of PP1 between normalindividuals and patients withchondrocalcinosis. Similarly, incultured skin fibroblasts, we failed toobserve any difference in the contentof PP1 between cells derived fromnormal individuals and those derivedfrom patients with chondrocalcinosis.

In contrast, the amounts of PP, incultured articular chondrocytes andmeniscus cells derived from patientswith chondrocalcinosis weresubstantially higher than the PP,contents of these cell types derivedfrom normal individuals, especiallyduring primary culture. However, thismay not necessarily reflect a change inintracellular PP1 metabolism in thesecell types in this condition, as it may bedue to the continued presence ofcalcium pyrophosphate crystals thatwere present in vivo. For example, inone such culture derived from apatient with chondrocalcinosis crystalsresembling triclinic calcium



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


pyrophosphate dihydrate (CPPD)were observed in the cell layers andthese slowly disappeared with time inculture. Moreover, medium samplesobtained from cultures of articularchondrocytes and meniscus cellsderived from two patients withchondrocalcinosis contained very highconcentrations of PP, initially butthese declined rapidly with time.Careful interpretation is thereforeneeded when apparently highconcentrations of PP, are found incultured articular chondrocytes andmeniscus cells derived from patientswith chondrocalcinosis, as these maysimply reflect the continued presenceof CPPD crystals that are trapped inthe monolayer and are graduallywashed away with each change ofmedium. These findings give littledirect evidence for a specific increasein the PPi content of cells derived frompatients with chondrocalcinosis,though it should be pointed out thatnone of our patients had recognisablehereditary forms of the disease.

In studies of the intracellularmetabolism of PP1, we have examinedthe influence of a variety of agents oncultured normal human articularchondrocytes and meniscus cells. Weobserved that increasing the mediumphosphate concentration from 1mmol to 2 mmol or 10 mmol in thepresence of 1- 3 mmol calciumresulted in an increase in theintracellular PPi content ofchondrocytes.

In the presence of 10 mmolphosphate, the intracellular PP1 valueswere similar in the presence of 0 4mmol, 1- 3 mmol, or 1 8 mmol calciumin the medium. When the calciumconcentration was reduced to a lowlevel, the intracellular PPiconcentration also fell. Treatment ofchondrocytes or meniscus cells witholigomycin, an inhibitor of oxidativephosphorylation which promotes

calcium influx into cells, increasedtheir intracellular PP1.One potential intracellular source of

PPi is as a byproduct of thebiosynthesis of glycosaminoglycans.We observed an increase inintracellular PPi in chondrocytesfollowing treatment of the cells withthe xyloside, 4-methylumbelliferyl-,B-D-xyloside, which stimulatesglycosaminoglycan biosynthesis.Interestingly, the intracellular PPiconcentration of chondrocytes andmeniscus cells also increased in thepresence of tunicamycin, an inhibitorof glycosylation. This suggests thatblocking of glycosylation and releaseof macromolecules from the cellsmight also impair the release of PP,from cells.Taken together, these observations

demonstrate that PP, concentrations inchondrocytes and meniscus cells canbe altered by a variety ofmanipulations, which implies that thePP, concentrations of these cell typesare under metabolic control and areinfluenced by the extracellularenvironment. The influence of theextracellular calcium concentrationmay be particularly important inrelation to the pathogenesis ofchondrocalcinosis and is of interest inrelation to the known associationbetween chondrocalcinosis andhyperparathyroidism.4

In studies of the extracellularmetabolism of PP1, we observed thatadded 39P-labelled PP, was only slowlyremoved from the medium on culturesof normal human articularchondrocytes such that about 20% ofthe added tracer was still present inthe medium even after 48 fiours.Moreover, 32P-labelled ortho-phosphate (P1) in the mediumappeared to account for a substantialfraction of the loss of label from PPi.Hence it would appear that hydrolysisof PP1 to Pi is the major mechanism for

the removal of PP, from the mediumbut that the ability of these cells tohydrolyse extracellular PP1 is limited.During short term incubation (11

hours) of washed monolayers ofnormal human articular chondrocytesin medium without serum, weobserved the release of small amountsof PP, into the medium, but the exactsource of this PP, is not known. Muchlarger amounts of PP, were generatedextracellularly following the additionof adenosine triphosphate (6-25-400,umol) to the medium without serumfor 14 hours. This effect is probablyattributable to the action of anucleoside triphosphate pyro-phosphohydrolase present on theoutside of the cells. This observationsuggests that human articularchondrocytes possess a substantialcapacity for the extracellulargeneration of PP,. This mechanismmight be important in thepathogenesis of chondrocalcinosis as itwould provide a means of producingPPi outside cells in the presence ofnucleoside triphosphates, which inturn might be derived by leakage fromdamaged cells in the neighbourhood.

References

1 Cheung C P, Suhadolnik R J. Analysis ofinorganic pyrophosphate at the picamolelevel. Anal Biochem 1977; 83: 61.

2 Russell R G G, Bisaz S, Fleisch H, et al.Inorganic pyrophosphate in plasma,urine and synovial fluid of patientswith pyrophosphate arthropathy(chondrocalcinosis or pseudogout).Lancet 1970; ii: 899-902.

3 Altman R D, Muniz 0 E, Pita J 0,Howell D S. Articular chondrocalcinosis:microanalysis of pyrophosphate (PPi) insynovial fluid and plasma. ArthritisRheum 1973; 16: 171-8.

4 Hamilton E B D. Diseases associatedwith calcium pyrophosphate depositiondisease. Arthritis Rheum 1976; 19:353-7.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


In vitro growth of calcium pyrophosphate crystals inpolyacrylamide gelsJ. E. HARRIES,' P. A. DIEPPE,' P. HEAP,3 J. GILGEAD,' M. MATHER,' J. S. SHAH'

From the University Departments of 'Physics, 2Medicine, and 3Anatomy, Bristol

Pyrophosphate arthropathy ischaracterised by the presence ofmicrocrystalline deposits of calciumpyrophosphate (CPPD) in articulartissues.' The crystals occur in twocrystallographic forms-namely,monoclinic (CPPD(M)) and triclinic(CPPD(T)). No other crystal speciesof calcium pyrophosphate has beenfound in vivo.

This is surprising, as in vitro studiesperformed under simulatedphysiological conditions most oftenyield the orthorhombic form of thetetrahydrate, CPPT(O),2 althoughprolonged incubation may yieldCPPD(M) and CPPD(T).3

Crystal deposition in vivo isrestricted to pathological hyalinecartilage, fibrocartilage, and synoviumand this has led to speculation thatthe physicochemical properties ofconnective tissue are important innucleation and crystal growth.4 In vitromodels, using gelatin and silicahydrogels, have therefore beendeveloped in an attempt to representthe physical properties of the tissuematrix.5Using gelatin gels at room

temperature we investigated thedependence of the crystal form on thecalcium ion concentration and thecritical pH. Preliminary studies haverevealed local changes in pHaccompanying crystal deposition.The gels were prepared using the

method of Pritzker et al.,' whereby asolution of CaCl2 is layered on top ofthe gel, previously soaked overnight in2-8 mmol Na4P2O7 and the pHadjusted at 60°C using 0-1 molNaOH/HC1.

Using 0-2 mol CaCl2 resulted in theformation of both CPPD(T) andCPPT(O) after a period of threeweeks. However, carefulconcentration programming, by which0 02 mol C. Cl2 is added to thesupernatant daily until nucleationoccurs (the solution then being madeup to 0 2 mol CaCl2) yielded only

CPPD(T) identified by x-ray powderdiffraction. By varying the initial pH ofthe gel from pH 55 to pH 85 at 60°C,it was found that growth of CPPD(T)was favoured by slightly alkalineconditions.The crystal deposits formed in layers

within the gel, collectively known asLiesegang rings. Addition of BDHuniversal indicator revealed anincrease in acidity with depth ofapproximately 3 pH units. Thisincrease was not entirely uniform butaltered sharply by up to 0-2 pH units,at the edges of the Liesegang rings.These observations imply that crystaldeposition is closely linked to the pHof the immediate surroundings.A disadvantage of using a gelatin

hydrogel as a model system is that itdegrades above 30°C to become aviscous liquid and is thereforeunsuitable for investigating crystaldeposition at physiologicaltemperatures.We have therefore used a

polyacrylamide gel which retains itsstructural properties even above bodytemperature.

Preparation of the gel involvesmixing equal volumes of 2-8 mmolNa,P207 and 10% w/vpolyacrylamidesolution (made by dissolving 10 gpolyacrylamide and 0 365 gNN'dialkyltartardiamide in 100 mldistilled water). Tetramethyl-ethyldiamine (0-25 ,ul/10 ml solution)is added and the pH adjusted to be-tween pH 6 and pH 7 with 10 NHCI/NaOH. Polymerisation isinduced with 0 25 ull/10 ml solution ofammonium persalphate. After gela-tion, typically 10-15 minutes, neutral0-2 mol CaCl2 is carefully poured ontop of the gel, which is then incubatedat either room temperature or 37°Cfor up to three weeks.

After one week small discretecrystallites were observed throughoutthe gel, although there was noevidence of any Liesegang rings, soprevalent in gelatin gels. The

crystallites were extracted by placingthe gel into a continuously stirred 2%solution of NaIO4 for two hours; thecrystals were recovered from solutionby filtration.

Optical microscopy performed onthe crystallites within the gel revealedtwo essentially differentmorphologies; one dendritic, the otherspheroidal (or in many cases toroidal)hollow shells.X-ray powder diffraction of crystals

grown at room temperature identifiedthe deposits as a mixture of CPPD(T)and CPPT(O), while those crystalsdeposited at 37°C were found to bepure CPPD(T) within the pH range6-7.The pH within the gel appeared not

to change during the crystal growthprocess, unlike the situation in gelatinhydrogels.

CONCLUSIONSWe have shown that a low calciumconcentration and an alkaline pHfavour the growth of CPPD(T) crystalsin gelatin gels. Using polyacrylamidegels we have found that thetemperature of the growth medium isimportant in deciding the nature of thecrystal form and that at 37°C pureCPPD(T) can be grown.

Gelatin and polyacrylamide gelsshow considerably different responsesto crystal growth in terms of theformation of Liesegang rings and localpH variations. In view of thesedifferences it appears that the gel isintimately involved in the crystalgrowth process. Therefore, carefulstudies must be made to ensure thatcurrent model systems in use arerealistic physiological models forcrystal deposition in human articularcartilage.

References

1 McCarty D J. Calcium pyrophosphatedihydrate crystal deposition disease.Arthritis Rheum 1976; 19: 275-85.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


2 Cheng P T, Pritzker K P H, Adams M E,Nyburg S C, Omar S H. Calcium pyro-phosphate crystal formation in aqueoussolutions. J Rheumatol 1980; 7: 609-16.

3 Hearn P R, Russell R G G. Formation ofcalcium pyrophosphate crystals in vitro:

implications for calcium pyrophosphatecrystal deposition disease (pseudogout).Ann Rheum Dis 1980; 39: 222-7.

4 Pritzker K P H, Cheng P T, Omar S A, etal. Calcium pyrophosphate crystalformation in model gels. II. Hyaline

cartilage as a gel. J Rheumatol 1981; 8:45 1-5.Pritzker K P H, Cheng P T, Adams M E,Nyberg S C. Calcium pyrophosphatedihydrate crystal formation in modelbydrogels.J Rheumatol 1978; 5: 469-73.

Effect of orthophosphate and other factors on thecrystal growth of calcium pyrophosphate in vitro

P. R. HEARN, D. F. GUILLAND-CUMMING, AND R. G. G. RUSSELL

From the Department ofHuman Metabolism and Clinical Biochemistry, University ofSheffield Medical School, SheffieldS10 2RX

Under simple simulated physiologicalconditions in vitro we have previouslyestablished that 40 ,u mol/l ofpyrophosphate is needed to initiateformation of calcium pyrophosphatecrystals in the presence of 1-5 mmol/lof Ca- at 37°C in three days.' Ifmagnesium is added to the system toapproach the physiological range (0 5mmol/l) the requirement forpyrophosphate rises to 175 AmolIl.We have now investigated the effectof another important ion, ortho-phosphate, over the physiologicalrange, on crystal formation and onlong term crystal growth.

Contrary to expectation there was adose-related decrease in the amount ofpyrophosphate needed to initiate theformation of calcium pyrophosphatecrystals in the presence of increasingphosphate. At 0 5, 1-0, and 2-0mmolVl phosphate, 125, 75, and 37 5,umol/l pyrophosphate are requiredrespectively in the presence of astandard 1 5 mmolIl Ca++ and 0-5mmoll Mg++. These results are ofparticular interest because hitherto ithas been thought that pyrophosphateacts as an inhibitor of calciumphosphate crystal deposition. The useof t mmol/l phosphate and 1 5 mmol/lof calcium with pyrophosphate addedcreates conditions that are close tophysiological and raises the importantquestion of which crystal forms firstdeposit in vivo. The amount ofpyrophosphate (75 ,umoll1) needed forcrystal growth of calciumpyrophosphate is still higher than theamount of pyrophosphate found in

extracellular fluids (normal range 1-4,u mol/l) or in synovial fluid inpseudogout (range 5-60 ,umol/l). Thissuggests that under pathologicalconditions in vivo additionalmechanisms may be needed toproduce high local concentrations ofcalcium or PP,, or that nucleatingagents are involved.Long term incubations at

physiological temperature, pH, andionic strength, with 1 c mmol/l Ca++,05 mmolIl Mg`+ and 200 ,umol/lpyrophosphate lead, in the absence ofphosphate, to the formation ofpredominantly monoclinic crystals ofcalcium pyrophosphate. The additionof phosphate at 1 mmol/l stabilises theamorphous gel that is formed initially,thereby prolonging the time neededfor crystal growth. However, it doeslead to the formation of much largercrystals. These crystals are of the sametype as in the absence of phosphate butdiffer in that a small proportion ofthem transform slowly to crystals whichappear by optical criteria to be tricliniccalcium pyrophosphate dihydrate (thetype found in vivo). This indicates thatunder simulated physiologicalconditions of pH and ionic strengthand at physiological concentrations ofkey ions, it is possible to grow thecrystal type found pathologically byusing high (200 ,umol/lpyrophosphate concentrations.The possibility that nucleating

agents are involved in vivo wasexplored using crystals ofhydroxyapatite or preformed crystalsof CPP, (monoclinic + orthorhombic).

Preliminary results indicate thataddition of large numbers ofhydroxyapatite crystals (40 Ag/ml)raise the concentration ofpyrophosphate required for CPPicrystal initiation, presumably becausethey bind pyrophosphate to theirsurfaces. Lower numbers of addedcrystals (4,g/ml) on the other handhad no obvious effect on the initiationof CPPi crystals, suggesting that anysurface binding effect wascompensated by a promoting effect onnucleation.

In long term incubations (> 2months) neither added hydroxyapatitenor added CPP, crystals altered thetype of new CPPi crystals formed.Both types of crystals considerablyaccelerated CPPi crystal formation inthe presence of excess PP1 (100-300,umol/l) both in the presence andabsence of phosphate.

In conclusion, it has not yet beenpossible to grow the pathological CPP,crystal type (triclinic) in short termculture in vivo under simulatedphysiological conditions. Such crystalsdo, however, appear in long termincubations (i-2 years), supporting thesuggestion that their presence inchondrocalcinosis represents the endpoint of a gradual process.

Reference

1 Hearn P R, Russell R G G. Formation ofcalcium pyrophosphate crystals in vitro:implications for calcium pyrophosphatecrystal deposition disease (pseudogout).Ann Rheum Dis 1980; 39: 222-7.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


An investigation into the progress of calcification insimple calcergyJ.E. HARRIES,' P. A. DIEPPE,2 P. HEAP,3 J. HORNBY,' A. SWAN,' AND J. S. SHAH'

From the University Departments of 'Physics, 2Medicine, and 3Anatomy, Bristol

The injection of a small quantity ofcertain chemical salts (calcergens) intothe subcutis of animals has been foundto induce a localised inflammatoryresponse followed by the deposition ofcalcium phosphate salts in the affectedarea.'2 This phenomenon is termedsimple calcergy and is beingextensively studied as a model for thepathogenesis of calcification inosteoarthritic tissue3 and the effect ofdiphosphonates and anti-inflam-matory drugs on mineralisation anddissolution of calcium phosphates inbone.4'Extensions to this simple model of

calcergy6 have shown mast cells to bepossible mediators of theinflammatory response andsubsequent calcification.6 We havetherefore investigated themorphological aspects of simplecalcergy by examining the progress ofinduced calcification in young malerats over two weeks. The chemicalnature of the deposits has also beenanalysed to find evidence forcalcification occurring via one or moreintermediate calcium phosphatecrystal phases to stablehydroxyapatite.Male Lester hooded rats weighing

200-250 g were used. Six dorsal sites,three either side of the mid-lumbarregion, were shaved; those on the rightside were injected subcutaneously with0 2 ml of 1 in 40 dilution in saline of a

saturated solution of potassiumpermanganate; those on the left sidewith 0-2 ml of neutral saline.At three, five, seven, and 14 days

after injection, four animals were

sacrificed for each time point usingether anaesthesia and divided into twogroups. In the first group the animalswere examined for any evidence ofsubdermal calcific deposits. Thesedeposits were usually seen as circularwhite discs, as found by otherworkers,' and were measured acrosstheir diameter in two places and theiraverage area determined. The plaques

were then removed and their wet anddry weights recorded. The sampleswere characterised using powderx-raydiffraction, x-ray energy spectroscopy,and scanning electron microscopy.

Plaques were measured in the same

way for the second group, but inaddition a histological analysis wasperformed on the lesions using bothoptical light and electron microscopy.

Samples were taken from both sidesin each animal and also from sites thathad not received any treatment(controls).

All the samples for morphologicalexamination were prepared byimmersion fixation using Bouin'ssolution with ester wax embedding forlight microscopy, and 3%glutaraldehyde followed by 1%osmium tetroxide, both in cacodylatebuffer at pH 7-2 with aralditeembedding for electron microscopy.Sections were stained with Alcian blueand haematoxylin eosin for lightmicroscopy and with lead acetate, withor without uranyl acetate, for electronmicroscopy.

RESUfLTS

Plaques were observed at sites injectedwith KMnO4 after five days, increasingin size and weight up to seven days. Nofurther change in the plaques was

observed up to 14 days (Table 1).X-ray diffraction revealed

hydroxyapatite and showed a generalincrease in apatite content with time.No other crystalline phase wasidentified. Scanning electronmicroscopy showed collagen fibrilsaligned in parallel bundles throughout

the plaques. These fibrils had a

petrified appearance, as if coated bysome material. Within the intersticesof the fibrils 'chalky' deposits wereobserved which were not bound to thecollagen matrix. X-ray energyspectroscopy gave variable ratios ofcalcium to phosphoros in the range 2-1to 2-6; hydroxyapatite having a ratioof 2-17.

Histological analysis of the lesionsshowed that three days after injectionof KMnO4, oedema was evident inassociation with mast cells,macrophages, and other connectivetissue cells. Large numbers ofcirculating monocytes andlymphocytes were also seen migratingto the affected area. After five days thecellular populations were mainlyconfined to mast cells, macrophages,and fibroblasts; the latter two celltypes exhibiting a degree of motility.Seven days after injection, fibroblastswere seen to be actively producingcollagen, and after 14 days were

almost the only cell type present, withthe collagen in close packed arrays offibrils extending throughout thearea-as confirmed by scanningelectron microscopy.

Isotopic scanning with 99mTclabelled methylene diphosphonate(MDP) was conducted on a thirdgroup of male rats after three, six,nine, and 15 days. The animals were

prepared as before with subcutaneousinjections of KMnO4 and saline. Ateach time interval four rats wereselected and intravenous 50 ,uCi (- 0 2ml) 99mTc labelled MDP administered.Four hours after injection each animal

Table 1 Mean (SD) variation in size and weight ofplaques excised from youngmale hooded rats 5, 7, and 14 days aftersubcutaneous injections ofOO2 mlKMnO4

Day 5 7 14

Average size of plaque (mm) 103 (11) 125 (29) 122 (22)Wet weight of plaque (g) 126 246 300Dry weight of plaque (g) - 41 94



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


was scanned using a General Electricmaxicamera II scanner and 400Tformatter. Sm Tc labelleddiphosphonate was found to beconcentrated within the plaques;activity being apparent by the sixthday.To estimate the mean turnover rate

of mineral in the plaques we areattempting to quantify this technique.

CONCLUSIONSSimple calcergy is an easy,reproducible model of pathologicalcalcification. Hydroxyapatite is foundto deposit without formingintermediary crystalline phases andmay nucleate on young collagen fibrilslaid down during the inflammatoryreaction. Mast cells are observed inlarge numbers during the initial phase

of the inflammatory response, alongwith macrophages and fibroblasts.After 14 days fibroblasts are thepredominant cell type, exhibitingelectron microscopical profusions,and are actively laying down newcollagen to encapsulate the affectedarea.

Isotope scanning may be a tool forassessing mineral turnover and formonitoring the progress ofcalcification.The calcergy model is now being

applied to investigate potentialtherapeutic inhibitors ofhydroxyapatite deposition.We would like to acknowledge the

financial support of Ciba-Geigy Ltd.

References1 Selye H, Tuchweber B, Gabbiani G.

Calcinosis induced by lead acetate. JPharmacol Exp Ther 1962; 128: 131.

2 Doyle D V, Dunn C T, Willoughby D A.Potassium permanganate inducedcalcergy: a model to study the effects ofdrugs on hydroxyapatite crystaldeposition.J Pathol 1979; 128: 63.

3 Doyle D V. Tissue calcification andinflammation in osteoarthritis. J Pathol1982; 136: 199-216.

4 McClure J. The effects of disodiumethane-hydroxy- 1 1-diphosphonate anddisodium dichloromethylenediphosphonate on lanthamide inducedcalcergy. J Pathol 1982; 137: 159-66.

5 McClure J. The effects of variousanticalcific, anti-rheumatic andanti-inflammatory drugs on local (simple)calcergy induced by lead acetate in themouse. J Pathol 1982; 137: 243-52.

6 Selye H, Ganniani G, Serafimou N.Histochemical studies on the role of themast cell in calcergy. J HistochemCytochem 1964; 12: 563.

Periarticular calcification of the shoulder in articularchondrocalcinosisJEAN C. GERSTER AND GEORGES RAPPOPORT

From the Rheumatology and Rehabilitation Centre and Radiology Department, University Hospital (CHUV), Lausanne,Switzerland

In articular chondrocalcinosis calciumdeposits commonly occur within thearticular cartilages and menisci;extra-articular linear deposits have,however, been described in theAchilles, quadriceps, triceps, andsupraspinatus tendons.'`3We report on the prevalence and

morphology of periarticularcalcification of the shoulder inpatients suffering from articularchondrocalcinosis compared with acontrol group.Two groups of patients were

studied. The first comprised 30consecutive patients (22 women, 8men; mean age 73 years, range 40-89)with definite articular chondro-calcinosis diagnosed radiologically inat least two joints. The disease wasidiopathic in 28 cases; one was associ-ated with hypothyroidism, andanother with gout. A comparable con-trol group (22 women, 8 men; mean

age 72-5 years, range 41-90) with noradiological evidence of articularchondrocalcinosis was taken at ran-dom and matched for sex and age.Patients with diabetes mellitus orsevere renal insufficiency were notincluded in view of the increased pre-valence of calcifying tendinitis amongthem.4

In all cases anteroposterior views ofboth shoulders were obtained withconventional radiological methods.When periarticular calcification wasfound xeroradiographic films weremade to accentuate their density.

Periarticular calcification wasfound in nine patients withchondrocalcinosis; all had idiopathicdisease; the calcium deposits werebilateral in three cases and unilateralin six. Calcification was also found inone control patient; the deposits werebilateral. The difference between thegroups in those with calcification was

significant (p<0 006) by the exactFisher test.

Morphologically, two types ofcalcification were encountered: linearpunctate and dense homogeneous.Linear punctate calcification wasfound in eight patients in the studygroup; in two cases punctate calciumdeposits were also observed at theinsertion of the long head of thebiceps. Dense homogeneouscalcification was found in onecontrol and four of the study group, inthree of whom they were associatedwith linear calcification. Thedifference between the groups inprevalence of dense calcification is,however, not significant (p = 0-18).Of the 10 cases with periarticular

calcification, three (all in the studygroup) had chronic shoulder pain, onesuffering from a rupture of the rotatorcuff; the remainder wereasymptomatic.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


These results show that periarticularcalcification of the shoulder, oftenclinically silent, is common amongpatients with articular chondro-calcinosis. It is found mainly atthe insertion of the supraspinatustendon.Eight patients in the study group

had fine linear tendon calcification.This calcification is supposed to bemade of calcium pyrophosphatedihydrate crystal deposits, inasmuch astheir morphology is similar to thatobserved in other tendons with typicalCPPD crystal deposits.2 However,the prevalence in the supraspinatustendon is twice that in the Achilles orquadriceps tendons2; this couldbe because of repeated strains dueto the great mobility of the shoulderjoints.Dense homogeneous tendon

calcification may be considered to resultfrom apatite deposits,"2 which are themost common cause of calcifying ten-dinitis of the shoulder. This wasmore frequent, but not significantly so,in the study group than in the controlgroup; further radiological studies areneeded to see whether apatitedeposition disease could be associatedwith articular chondrocalcinosis. Inthree cases both linear and densetendon deposits were present; thesefindings have to be compared withthose of a triceps tendon where bothCPPD and hydroxyapatite crystalscould be demonstrated byradiocrystallography,3 suggesting thatmixed crystal deposition disease mayoccur in certain tendons. There is ananalogy with what occurs in someosteoarthritic joints where the twokinds of crystals can coexist.5

References

1 Genant H K. Roentgenographic aspectsof calcium pyrophosphate dihydratecrystal deposition disease (pseudogout).Arthritis Rheum 1976; 19: 307-28.

2 Gerster J C, Baud C A, Lagier R, Bous-sina I, Fallet G H. Tendon calcificationsin chondrocalcinosis. A clinical, radio-logic, histologic and crystallographicstudy.ArthritisRheum 1977; 20: 717-22.

3 Gerster J C, Lagier R, Boivin G. Olecra-non bursitis related to calcium pyro-phosphate dihydrate crystal depositiondisease. Clinical and pathologic study.Arthritis Rheum 1982; 25: 989-96.

4 Resnick D. Calcium hydroxyapatitecrystal deposition disease. In: Resnick D,Niwayama G, eds. Diagnosis ofbone andjoint disorders. Philadelphia, London,Toronto: W B Saunders, 1981: 1575-97.

5 Dieppe P A, Doyle D V, Huskisson E C,Willoughby D A, Crocker P R. Mixedcrystal deposition disease and osteo-arthritis. Br Med J 1978; i: 150.

Articular chondrocalcinosis, quadriceps calcification,and patellofemoral degeneration in the elderlyEDWARD WILKINS, AND GORDON EVISON

From St Martin's Hospital, Bath

Linear calcific deposits in thequadriceps tendon1 2 and 'isolated'patellofemoral degeneration3 havebeen described as radiographicfeatures seen in association withintra-articular chondrocalcinosis. Ithas been suggested that their presencemay be of value in distinguishingbetween pyrophosphate arthropathyand other forms of degenerative jointdisease.We assessed the influence of aging

on this interrelationship by analysingradiographs taken on consecutiveadmissions to an acute geriatric unit.We documented the prevalence ofintra-articular chondrocalcinosis,patellofemoral degeneration, andquadriceps calcification and definedtheir interrelationship.

METHODS AND RESULTSA total of 120 consecutive x-ray filmsof the knee (anteroposterior andlateral), pelvis, and wrists wereanalysed by four independent

observers, using light intensificationand magnification. Articularchondrocalcinosis was defined as thefinding of dense, hazy, linear, orstippled intra-articular calcification.Quadriceps calcification as seen onlateral views ofthe knee was defined aseither in the muscle belly or in thetendinous insertion. Formation ofosteophyte on the upper margin of thearticular surface of the patella, loss ofarticular cartilage in thepatellofemoral joint, and subchondralsclerosis was used to assess thepresence of patellofemoraldegeneration, classified as mild,moderate, or severe.To achieve maximum sensitivity and

to avoid excessive radiation exposure,Ilford rapid R film was used for theknee x-ray examinations, andX-6-Mat RP-X-RPI for the pelvis.Data were avilable on 100 patients

(31 men, 69 women) aged from 65 to97 years (mean (SD) 79-4years (6-6)).None of the patients had

haemochromatosis or hypercalcaemia.Thirty-four patients had

intra-articular chondrocalcinosis; 25had changes at the knee. Theprevalence rose from 15% (3 out of20) in those aged 65-74 to 36% (20out of 55) in those aged 75-84, and to44% (11 out of 25) in patients agedfrom 85 to 97 years.

Quadriceps calcification at thetendinous insertion was seen in 54patients. The prevalence increasedfrom 20% in patients aged 65 to 74years to 60% in those aged 75 to 84, to64% in those aged 84 to 97.

Fifty patients had evidence ofpatellofemoral degeneration. In 13patients the changes were mild, in 27patients moderate, and in 10 severe. In17 cases radiographs showed no otherabnormality-that is, degenerationwas isolated. The prevalence rose from35% in patients aged 65 to 74 to 49%in those aged 75 to 84 years, and to64% in patients above 84 years.The table shows the



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


interrelationship between articularchondrocalcinosis, patellofemoraldegeneration, and quadriceps calci-fication. Articular chondrocalcinosis issubcategorised into present at theknee and in joints other than the knee.

Analysis of these figures shows thatthe prevalence of patellofemoraldegeneration and quadricepscalcification is similar in patients withand without articular chondro-calcinosis.

CONCLUSIONArticular chondrocalcinosis,patellofemoral degeneration, andquadriceps calcification are commonin the elderly and their prevalenceincreases in a linear fashion with aging.They are radiographic phenomenaclosely related to aging and cautionmust be exercised in postulating arelationship with a disease process inthe elderly.

References1 Gerster J C, Baud C A, Lagier R, et al.

Table 1 Relationship between articular chondrocalcinosis (A CC), quadricepscalcification, and patellofemoral degeneration

In association In association No evidencewith ACC at with ACC ofACCthe knee (n =25) elsewhere (n 9) (n =66)

Quadriceps calcification:Tendon (n=54) 14 (56%) 4 (44%) 37 (56%)Muscle (n=10) 3 (12%) 2 (22%) 5 (8%)

Patellofemoral degeneration:Overall (n=37) 14 (56%) 5 (55%) 31 (46%)Isolated (n= 16)(mild, moderate andsevere) 3 (12%) 1 (11%) 13 (20%)Isolated (moderateand severe only) 2 (8%) 1 (11%) 8 (12%)

Tendon calcification in chondro-calcinosis. Arthritis Rheum 1977; 20:717-22.

2 Waltzing P, Breville P H, de FugeresY D L. La calcification du cul de sacquadricipital. Nouvre Presse Med 1980; 9:41.

3 Resnick D, Niwayama G, Goergen T G,et al. Calcium radiographic and patho-

logical abnormalities in calcium pyro-phosphate dihydrate deposition disease(CPPD): pseudogout. Radiology 1977;122: 1-15.

4 Lagier R. Femoral cortical erosions andosteoarthrosis of the knee: an anatomo-radiological study of two cases. FortschrGeb Roentgenstr Nuklearmed Ergan-zungsbana 1974; 120: 460-7.

Arthritis of idiopathic haemochromatosis

E. B. D. HAMILTON, A. B. BOMFORD, J. W. LAWS. AND R. WILLIAMS

From the Departments ofRheumatology and Radiology, and the Liver Research Unit, King's College Hospital, London

We have previously described thearthritis of haemochromatosis,' 2 andhave now re-examined 18 of thesecases after a mean interval of 9 4years. All patients underwent repeatx-ray examination.

Chondrocalcinosis was found in atleast one joint in seven patientsinitially and in 13 patients at thesecond assessment. Despite adequatetreatment of iron overload byvenesection it increased in severity andspread to new joints.Thirteen patients developed

arthritis of the metacarpophalangealjoints, but none of them had associatedchondrocalcinosis visible radio-logically at this site or in the triangularligament of the wrist.

It was not possible to show acorrelation between the presence ofchondrocalcinosis at the initialassessment and the extent of iron

Table 1 Prevalence ofchondrocalcinosis in haemochromatosis atfirstand followup examination in 18 patients (duration offollow up 9-4 years)

First Secondexamination examination

Wrists 3 10Knees:

Meniscus 6 11Hyaline cartilage 4 6

Hips 2 5Symphysis pubis 4 5Spine 2 4

stores or the patient's age. The tableshows the incidence ofchondrocalcinosis.3References1 Hamilton E, Williams R, Barlow K A,

Smith P M. The arthropathy of idio-pathic haemochromatosis. Q J Med 1968;37: 171.

2 Dymock I M, Hamilton E B D, Laws J W,Williams R. Arthropathy of haemo-chromatosis. Clinical and radiologicalanalysis of 63 patients with iron overload.Ann Rheum Dis 1970; 29: 469-76.

3 Hamilton E B D. Chondrocalcinosis andosteoarthritis. In: Peyron J G, ed.Epidemiology of osteoarthritis. Maccles-field: Geigy Publications, 1980: 109-12.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Double blind, placebo controlled trial of magnesiumcarbonate in chronic pyrophosphate arthropathyMICHAEL DOHERTY, AND PAUL A. DIEPPE

From the University Department ofMedicine, Bristol Royal Infirmary, Bristol

Unlike the situation with urate crystalsin gout, there is at present no definitivetreatment to prevent formation orenhance dissolution of CPPD crystalsin patients with pyrophosphate arth-ropathy. In vitro studies, however,have shown that magnesium exerts asolubilising, growth inhibitory effect onCPPD crystals,' and the possibilitythat magnesium supplementationmight influence CPPD crystals in vivois suggested by the precipitation ofacute pseudogout after lavage of joints

Pain

-Stiffness

1t \

with magnesium sulphate2 and by thereported benefit of magnesiumreplacement in a patient withhypomagnesaemia and chondrocal-cinosis.3 We therefore undertook a sixmonth controlled trial to investigatethe therapeutic potential of mag-nesium in patients with pyrophosphatearthropathy.A total of 38 patients with chronic

pyrophosphate arthropathy(symptomatic arthritis + CPPDcrystals in synovial fluid +

0-6r

radiographic evidence ofchondrocalcinosis) were allocated atrandom to receive 10 ml thrice daily ofeither magnesium carbonate (30 mEqmagnesium/day) or placebo. Thetreatment group (12 men, sevenwomen) and controls (five men, 14women) were comparable for meanage (64-4 + 12 yr v 69 -+ 10 yr) andduration of disease (9 + 8 yr v 8 + 5yr). Patients were assessed regularlyby a single observer and treatment waspatient and observer blind.

Muchbetter[ A °

Better *::. 8'804

Same I-- --o888--0-2

0

1-0[

051

Tbrse . .

Tenderness Muchbetter

Better

Same

Worse-

0'

0

Patient opinion

0@

ooooo

-W 0000-

0 000

Observer opinion

Treatment group (n=W) * L 1

Placebo group (n=19) o 0

Fig. 1 Duration ofpain, effusion, stiffness, and tenderness before and aftertreatmentand subjective and objective assessmentofeffect oftreatment. Pain was rated on a scale ofO 'none,' I 'mild,' 2 'moderate,' and 3 'severe, 'effusion and tenderness on ascale of 0 'none,' I 'mild,' and 2 'severe.' Duration oftenderness was measured in minutes.

3

Q L10 i

5 i

0

2~

11



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Despite a pronounced placeboeffect in the controls, the treatmentgroup showed a uniform trend towardsimprovement in pain, stiffness,effusion (knee), joint line tenderness,and overall subjective and objectiveassessment (Fig.): at six monthsimprovement over controls wassignificant for pain (p<0 05) andobjective assessment (p< 0 05). Sevenstudy patients and two controls notedincreased bowel frequency;intermittent diarrhoea developed inthree (two study patients, one control)but did not necessitate withdrawalfrom the trial. No changes inbiochemical values or synovial fluidwere observed apart from an increasein 24 hour urinary magnesium

excretion from a mean (SD) of 3-9(2-1) mmol (mEq)/24 h to 5 7 (2.5)mmol/24h in those taking magnesium.The degree of improvement in thosetaking magnesium did not correlatewith renal function, initial magnesiumstate, or increase in urinarymagnesium output. Radiologicalappearance of chondrocalcinosis didnot change over the six month period.The uniform trend towards

improvement in those takingmagnesium suggests that furtherstudies of magnesium supple-mentation are warranted, especially inview of its safety in the elderly patientsincluded in this trial. The pronouncedplacebo effect we observed serves toemphasise the debilitating nature of

pyrophosphate arthropathy and thesupportive effect of regular, interestedmedical attention on patients demoral-ised by chronic pain and stiffness.

References

1 Cheng P T, Pritzker K P H. The effect ofcalcium and magnesium ions on calciumpyrophosphate crystal formation inaqueous solutions. J Rheumatol 1981; 8:772-82.

2 Bennett R M, Lehr J R, McCarty D J.Crystal shedding and acute pseudogout:an hypothesis based on a therapeutic fail-ure. Arthntis Rheum 1976; 19: 93-7.

3 Runeberg L, Collan Y, Jokinen E J,LUhdevirta J, Aro A. Hypomagnesaemiadue to renal disease of unknown aetiol-ogy. Am J Med 1975; 59: 873-81.

Problems encountered in the routine analysis ofsynovial fluid for crystalsP. R. HEARN, D. F. GUILLAND-CUMMING, AND R. G. G. RUSSELL

From the Department ofHuman Metabolism and Clinical Biochemistry, University ofSheffield Medical School, SheffieldS10 2RX

Compensated polarised lightmicroscopy is a simple technique thatallows identification of crystals bydetermination of the sign of theirbirefringence based on colour changewith rotation. The technique appliedto crystallography of synovial fluidsamples was described by Currey andVernon-Roberts.' Our own diagnosticservice was set up in conjunction with aresearch project on pyrophosphatemetabolism and calciumpyrophosphate crystal deposition.The technique itself is simple if

carefully carried out, but there arevarious problems and pitfalls. Wedescribe some problems that we haveencountered and give some hints forthose setting up their own diagnosticservice.

(1) Microscope. In the absence of apurpose-built polarising lightmicroscope, most micrscopes can beconverted by the addition of polarisingfilters and a first order redcompensating filter.

(2) Slides. Ordinary glassmicroscope slides are used but must be

as dust-free as possible; we have notedappreciable amounts of birefringentmaterial on apparently clean slides andcoverslips. Slides should be wipedrepeatedly with lens tissue. Coverslipsare difficult to clean because of theirfragility; we rarely use them.

(3) Samples. Synovial fluid aspiratesare taken without anticoagulant, as theclots which form trap crystals andconsiderably help the examination.Whenever possible, samples should beexamined the same day, but samplesstored at 4°C do not appear to alterduring storage for up to a week.Storage at room temperature may leadto artefactual crystal growth asreported by Dieppe.'

RESU LTSIn many samples crystals were initiallytoo small to be identified. These weretiny irregular particles or crystals oftenaccompanied by crystals showing aMaltese cross pattern under polarisedlight. Many of these particles proved tobe artefactual. Examination of glovepowder currently used by surgeons

showed the presence of numerousparticles giving the same characteristicMaltese cross appearance. Thispowder is pure starch, whichpresumably enters the samplecontainer at the time of aspiration andsample collection. A similar Maltesecross appearance is produced by lipiddroplets; these can be dissolved inxylene and similar solvents and arequite commonly seen.

Fragments of cartilage, drawn up atthe time of aspiration, may also causeidentification difficulties, as mav clotsof fibrin with cells embedded in them.Irregular, weakly birefringentparticles may be hydroxyapatiteconglomerates, but we are unable toidentify these routinely at present.

Criteria for crystal identification.Clearly, a diagnostic service must havecriteria for identification of crystals. Acrystal must be: bright under polarisedlight, exhibit birefringence, or showthe required colour change on rotationof the compensating filter. It must looklike a crystal-that is, have regularsides (crystal conglomerates are



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


difficult to focus on), but identificationby shape alone is not sufficient.Other crystals include recently

injected corticosteroids, alsocholesterol which is seen as flat plates.We have seen red/brown crystals froma haemarthrosis aspirate. These wereidentified as bilirubin. Occasionallycrystals are seen which cannot be

identified, and some of these may bethe alternative crystal forms ofcalciumpyrophosphate,3 or of calciumphosphate.

References

1 Currey H L F, Vernon-Roberts B.Examination of synovial fluid. Clinics inRheumatic Diseases 1976; 2: 149-77.

2 Dieppe P A, Crocker P R, Corke C F,Doyle D V, Huskisson E C, WilloughbyD. Synovial fluid crystals. Q J Med 1979;48: 533-53.

3 Hearn P R, Russell R G G. Formation ofcalcium pyrophosphate crystals in vitro:implication for calcium pyrophosphatedeposition diseases (pseudogout). AnnRheum Dis 1980; 39: 222-7.

New methods for identification of crystalline materialin joints

D. V. DOYLE AND P. R. CROCKER

From the Department ofRheumatology, St Bartholomew's Hospital, London ECIA 7BE

The identification of microcrystallinematerial in joint fluid and tissues byconventional electron microscopical(EM) methods requires about 48hours' specimen preparation. Wehave developed a series of methodsthat, in the case of joint fluids, allowsthe rapid preparation of freshspecimens for EM examination and, inthe case of tissue specimens, enablesparaffin-embedded sections to beexamined directly by scanningelectron microscopy (SEM).

SYNOVIAL FLUIDS(1) Cytocentrifugation.

Transmission electron microscopy(TEM) system. The fluid iscytocentrifuged onto coated EM grids,which are then examined by lightmicroscopy (LM) to select the mostsuitable monolayer for examination byTEM. Crystalline material, wheredetected, is analysed by x-ray energyspectroscopy (XES).Scanning electron microscopy

(SEM) system. The fluid iscytocentrifuged onto a standardperspex slide (7-6 cm/2-5 cm), which isthen placed in the microscope forexamination and analysis. Thespecimen on the perspex slide may alsobe microincinerated at lowtemperature to remove organic

material, leaving crystals clearlyexposed for examination.Scanning electron microscopy with

back-scattered electron imaging facility(BEI). The specimen is prepared as forSEM, but the BEI facility enablesrapid location of crystalline materialby atomic number differences withoutthe need for microincineration.

(2) Millipore filtration.SEM system. Negative pressure

filtration of synovial fluids is used tocollect particulate material on thesurface of a millipore filter. The filter isplaced on a graphite stub in the bulkholder of the scanning electronmicroscope, and the surface examinedby SEM and XES.SEM system with BEI facility. The

preparation is as for SEM, but the BEIfacility enables rapid location ofcrystalline material.

JOINT TISSUESRoutine paraffin block sections areused. Standard 5 ,um sections are cutand mounted on perspex slides.SEM system. The tissue section on

the perspex slide is examined by lightmicroscopy and the topography isnoted. The section is thenmicroincinerated at low temperature

to expose particulate material forexamination by SEM and XES.SEM system with BEI facility. The

preparation is as for SEM, butmicroincineration is unnecessary.Direct examination by BEI and XES ispossible.The identification of micro-

crystalline material in synovial fluidon articular tissue specimens by con-ventional EM methods is impracticalfor routine diagnostic use because it isextremely time consuming. The tissuespecimen is cut into very small blocks,which must then be fixed, embedded,and sectioned. This procedure takes atleast 48 hours and as the specimenblocks are very small and tissuedeposits of crystals are often patchilydistributed, the specimen selected forexamination may well be devoid ofcrystalline material.Our methods used standard

laboratory fixation and preparationtechniques. The specimens, apart fromthose involving millipore filtration,may be examined initially by lightmicroscopy to determine whether EMexamination is likely to be worthwhile.These techniques are a significantadvance on traditional electronmicroscopical methods used for thedetection of crystalline material injoint fluids and tissues.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Crystals and vessel wallsP. N. PLATT AND A. MALCOLM

From the Departments ofRheumatology and Pathology, University ofNewcastle upon Tyne

Hitherto the concept ofcrystal-induced diseases has had itsmain impact in rheumatology. It isbecoming obvious, however, thatcrystals have a part to play in theproduction of disease in other systems.We thought that the crystalline form ofsome of the constituents ofatherosclerotic plaques may play apart in their formation-for example,cholesterol crystals are capable ofactivating polymorphs' and promotingcollagen production by fibroblasts.!Hence as part of an investigation intothe role of crystals in vessel walls weundertook a preliminary study usingpolarising light microscopy andpresent the results.

Early non-calcified plaques ofuncomplicated atheroma of the aortawere obtained at necropsy from bothsexes with an age range of 28-84years.These plaques were subjected to

rapid freezing and 5 ,um thick cryostat

sections both unstained and stained(haematoxylin and eosin, oil red '0')were examined using a polarising lightmicroscope. The oil red 'O' demon-strated lipid, and crystals of choles-terol and monosodium urate. Uratecrystals were identified by theircharacteristic appearance under polar-ised light microscopy and solubility inuricase.

In addition to the expectedcholesterol crystals we describe whatappeared to be crystals ofmonosodium urate occurring insamples ofearly atheromatous plaquesfrom human aortas at necropsy. Webelieve that this may be of someimportance in view of the longspeculated association betweenhyperuricaemia and atherosclerosis.3The situation may be analogous to thatoccurring in tophi, where cholesteroland hydroxyapatite4 are often found inconjunction with urate crystals

suggesting common nucleationfactors. The capabilities of thesecrystals to trigger inflammatory andsclerotic reactions suggests a possiblerole in the production ofatherosclerotic lesions.

References

1 Hammerschmidt D E, Greenberg C S,Yamada 0, Craddock P R, Jacob H S.Cholesterol and atheroma lipids activatecomplement and stimulate granulocytes.J Lab Clin Med 1981; 98: 68-77.

2 Pritzker K P H, Adel G F, Omar S,Gertzbein S D. Experimental cholesterolcrystal arthropathy.J Rheumatol 1981; 8:281-90.

3 Dreyfuss F. The role of hyperuricaemia incoronary artery disease. Diseases of theChest 1960; 38: 332-6.

4 Dieppe P. Crystal deposition disease andthe soft tissues. Clinics in RheumaticDiseases 1979; 5: 807-21.

An immunoelectron microscopical study of theorientation of IgG molecules on the surface ofmonosodium urate crystalsT. BARDIN,' P. V. CHERIAN,2 AND H. R. SCHUMACHER2

From the 'Clinique Rheumatologique, Hopital Lariboisiere, Paris, France, and 2University ofPennsylvania School ofMedicine, Philadelphia, USA

The binding of IgG to monosodiumurate (MSU) crystals has beendemonstrated by severaltechniques.`-3 It is thought to play animportant part in the pathophysiologyof the gouty attack, as IgG coating ofMSU crystals could modulate crystalinteractions with cells, in particular byreacting with IgG Fc receptors. Theavailability of Fc fragments on thecrystal-bound IgGs thus appears to be

an important factor in the part IgGmight play in gouty inflammation. Werecently presented data in favour ofsuch an availability.3 We set out toinvestigate further the orientation ofIgG molecules from serum, bound toMSU crystals, by usingimmunoperoxidase techniquesdirected against the Fc and F(ab' )fragments of IgG.Unheated synthetic MSU crystals

were incubated with normal humanserum for one hour at 4°C, washed,and then processed with an indirectimmunoperoxidase technique usingrabbit antihuman IgG, at aconcentration of 120 ,ug/ml, directedeither against the F(ab')2 or the Fcfragments, followed by horseradishperoxidase conjugated goat antirabbitIgG. Next, crystals were incubatedwith DAB + 1-O0 and were post fixed



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


with osmium tetroxide. To preventnon-specific binding of the antibodiesto crystals, samples were incubatedwith normal goat globulins before andduring the antibody incubation steps.Samples were washed between eachstep with buffers saturated with urateto avoid dissolution of the crystals.Appropriate controls were used,including blocking with unconjugatedgoat antirabbit antibodies andreplacement of the primary antibodiesby normal rabbit globulins. Allsamples were dehydrated, embeddedin Spurr medium and thin sectioned forobservation on transmission electronmicroscopy. About 100 crystals wereanalysed from samples of each type ofprocessing. Each crystal was classifiedas negative, weakly positive, orstrongly positive, according to theintensity of peroxidase reaction resultswithout knowledge of the processing

done. Results were compared by the X2test (see Table 1).

Crystals dissolved during thedehydration and embeddingprocedures, but most of theirsilhouettes could still be identified inthin sections. In samples processed toreact with F(ab')2 fragments, most of thecrystal sites were strongly outlined bydark reaction products which showedthe IgG coating and proved thefunctional availability of the Fcfragments of the crystal-bound IgG. Incontrast, positive crystals were verysignificantly less frequent(p<000001) in samples treated toreact with F(ab')2 fragments than inthose processed to demonstrate Fcfragments (see Table 1). Controls werenegative or had only weak reactionproducts.These data are consistent with

previous studies using different

Table 1 Intensity ofIgG Fc and F(ab')2 reaction with immunoperoxidase afterincubation ofMSU crystals with serum

Types of Total No No of No of No ofprocessing ofcrystals negative weakly strongly

analysed crystals positive positivecrystals crystals

Anti-IgG Fc 101 17 24 60Anti-Ig F(ab')5 129 97 22 10Normal rabbitglobulin control 92 85 7 -

Blocking control 126 84 36 6

techniques4 and might be explained byelectrostatic forces that are known tobe important in protein adsorption toMSU crystals."2 Fab fragments have amore positive charge than Fcfragments' so that the Fab extremitymay bind preferentially to thenegatively charged crystal, thusleaving the Fc end free.

References

1 Kozin F, McCarty D S. Protein binding tomonosodium urate monohydrate,calcium pyrophosphate dihydrate andsilicon dioxide crystals. 1. Physicalcharacteristics. J Lab Clin Med 1977; 89:1314-25.

2 Hasselbacher P. Binding of IgG andcomplement protein by monosodiumurate monohydrate and other crystals. JLab Clin Med 1979; 94: 532-41.

3 Bardin T, Cherian P V, Claybume G,Schumacher H R. Transmission electronmicroscopic demonstration of the bindingof immunoglobulins to monosodiumurate crystals [Abstract].Arthritis Rheum1982; 25: S76.

4 Kozin F, McCarty D J. Molecularorientation of immunoglobulin Gadsorbed to microcrystallinemonosodium urate monohydrate. J LabClin Med 1980; 95: 49-58.

5 Stanworth D R, Turner M W.Immunochemical analysis ofimmunoglobulins and their subunits. In:Weiz D M, ed. Handbook ofexperimentalimmunology. Vol 1. Oxford: BlackwellScientific Publications, 1979: 1-6, 102.

Crystal-induced oxygen uptake by animal neutrophilsF. K. HIGSON AND 0. T. G. JONES

From the Biochemistry Department, Bristol University, Bristol BS8 I TD

Crystals and other types of particulatematerial are often seen in synovialfluids. At least three types of crystalare known to be pathogenic:monosodium urate monohydrate,calcium pyrophosphate dihydrate, andhydroxyapatite. Phagocytosis ofcrystals by leucocytes or otherinteractions between crystals andleucocytes in joint tissue are likely tobe involved in the inflammatoryresponse."We have prepared neutrophils from

pig blood3 and examined the effect ontheir oxygen metabolism of crystaladditions. The crystals, a gift from DrPaul Dieppe, had the followingdimensions: monosodium uratemonohydrate, 5-10 ,um length;calcium pyrophosphate dihydrate(CPPD), average 15 ,um length;brushite, 5-15 ,um length; diamond,2-7 gm length; and cholesterol,average 20 ,um length. Crystals were

suspended in the modifiedKrebs-Ringer buffer used in

neutrophil preparation and sonicatedbriefly before use. Oxygen uptake wasmeasured using a Clarke typeelectrode, superoxide productiondetermined as superoxide-sensitivecytochrome with reduction andperoxide fluorimetrically by couplingit to the peroxidation of 4-OH,3-methoxyphenylacetic acid.The addition of crystals of urate,

brushite, or CPPD to neutrophilscaused a great increase in oxygenuptake which was insensitive to 2



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Abstracts Supplp 111

mmol KCN. The effects depended onthe crystal concentration, but themaximum stimulation approached50% of that given by the solublestimulus phorbol myristate acetate.The results could be plotted using theform of the Michaelis-Mentenequation to give a Km for each crystal.For urate this was 12-5 mg crystal/ml,for brushite 1-2 mg/ml, for calciumhydroxyapatite 1-2 mg/ml. Diamondand cholesterol crystals did notstimulate oxygen uptake.

Addition of brushite caused theproduction of superoxide and 1202 byisolated neutrophils at concentrationssimilar to those found to stimulateoxygen uptake. Analysis of thetemperature dependence of the

stimulation of oxygen uptake showedthat urate crystals were ineffective attemperatures below 23°C; phorbolwas effective at, and above, - 17- 5°C.Urate and phorbol showed similartemperature dependence above thistrigger temperature in Arrheniusplots.

Neither colchicine nor cytochalasinB inhibited the crystal-induced burstof oxygen uptake over the usualconcentration range at which theseinhibitors are used. Inhibitors ofSH-dependent enzymes and flavinanalogues were, however, potentinhibitors.

This work was supported in part by grantsfrom the Medical Research Council and

from the Wellcome Trust. We are alsograteful to Dr Paul Dieppe for his adviceand co-operation.

References

1 Abramson S, Hoffstein S T, Weissman G.Superoxide anion generation by humanneutrophil exposed to MSU: effect ofprotein adsorption and complementactivation. Arthritis Rheum 1982; 25:174-80.

2 Simchowitz L, Atkinson J P, Spilberg I.Stimulation of the respiratory burst inhuman neutrophils by crystalphagocytosis. Arthritis Rheum 1982; 25:181-8.

3 Cross A R, Higson F K, Jones 0 T G,Harper A M, Segal A W. The enzymicreduction and kinetics of oxidation ofcytochrome b-245 of neutrophils.Biochem J 1982; 204: 479-85.

Crystal interactions with polymorphonuclearleucocytes studied by luminol-dependentchemiluminescencePHILIP PLATT, MARCUS HUDDIE, AND W. CARSON DICK

From the Department of Rheumatology, University ofNewcastle upon Tyne

In recent years the crystalarthropathies have emerged as adistinct group of diseases linked bycommon pathogenetic mechanisms.One topic thought to be important isthe interaction between crystals andpolymorphonuclear leucocytes in theproduction of inflammation.' Wedescribe a method that allows thestudy of early events in this interactionby means of luminol-dependentchemiluminescence.Polymorphonuclear leucocytes

were obtained from fresh normalperipheral blood by dextransedimentation2 and suspended inphosphate buffered saline.Luminol reacts with oxidising agents

produced by the polymorph,3 in theform of superoxides, hydrogenperoxide, and hydroxyl radicals, toproduce an unstable intermediatewhich spontaneously releases to aground state releasing photons in theprocess. The reaction was followed onan LKB 1250 luminometer linked to aflat bed recorder. The reaction cell ofthe luminometer was thermostaticallycontrolled and the reactants held at

37°C. The reaction mixture consistedof 1 ml of phosphate buffered salinemedium containing 1- 5 x 106 cells, 0-1ml of a standard luminol solution, and0-1 ml of a 1% crystal suspension.A series of dose-response curves

with varying concentrations in a fixedvolume showed a linear response untilconcentrations greater than 3% wereused. We have used this technique toinvestigate the ability of differentcrystal preparations to induce releaseof superoxides and other oxygenradicals from polymorphs. Of thosecrystals implicated in joint diseasemonosodium urate crystals producedthe largest response, the responses ofCPPD, hydroxyapatite, brushite andcholesterol being less than 20% of theresponse to MSU.The inflammatory potential of

crystals as measured by animal modelsand their surface charges have beenshown to be highly correlated.4 Wehave demonstrated a high degree ofcorrelation between the surface chargeof MSU crystals as measured byelectrophoretic mobility and thechemiluminescence response induced

by the crystal preparations.Luminol-dependent chemi-

luminescence appears to be a usefulmethod for studying early events in theinteraction of polymorphonuclearleucocyte membranes and crystalsurfaces and may allow furtherclarification of factors initiating andmodifying effector mechanisms ofcrystal induced disease.

References1 Phelps P, McCarty D J. Crystal induced

inflammation in canine joints II. Theimportance of polymorphonuclear leuco-cytes. Experimental Medicine 1966; 124:115-26.

2 Boyum A. Isolation of leucocytes fromhuman blood. Scand J Clin Lab Invest1968; 21, suppl 97: 9-50.

3 Allen R C, Loose L D. Phagocytic activa-tion of a luminol dependent chemi-luminescence in rabbit alveolar and pen-toneal macrophages. Biochim BiophysActa 1981; 631: 380-5.

4 Dieppe P A, Swan A, Hornby J, JenkinsR, Luckman N P, Preece A W. Crystalsurface charge and inflammation.AnnRheum Dis 1980; 39: 606-11.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Different effects of crystals on release of inflammatorymediators from human peripheral blood*phagocytic cells

PAUL DIEPPE, JUNE HORNBY, AND ANGELA SWAN

From the University Department of Medicine, Bristol Royal Infirmary, Bristol BS2 8HW

Although the mechanism ofcrystal-induced inflammation remainsdisputed, most authors think thatparticle phagocytosis bypolymorphonuclear cells (PMNs) hasa central role.' (P Platt and W C Dick,p. 4). PMNs are often thought of asindiscriminate cells, which release apackage of mediators irrespective ofthe stimulus,2 but most experimentalwork on particles has only been withurate crystals.3 We have thereforeexamined phagocytosis, cell killing,enzyme release, and generation ofchemotactic factor by humanperipheral blood PMNs in response tofour different crystals.PMNs were obtained by Dextran

separation of human venous blood.They were incubated in protein-freemedium alone or with crystals ofdiamond (courtesy of De Beers,London; 5 am diameter), silica(courtesy of Pneumoconiosis Unit,Cardiff; 10 ,um long), monosodiumurate monohydrate (5-10 um long),

and hydroxyapatite (clusters about 2,um diameter). The urate and apatitecrystals were manufactured bystandard methods. Phagocytosis andcell death were assessed visually andwith trypan blue staining./3-glucoronidase release was measuredspectrophotometrically and expressedas a percentage of release by total celllysis. Chemotactic factor generationwas measured by a modified Boydenchamber technique using ZAS as apositive control.

All the particles were readilyphagocytosed. The Table shows thedifferent effects on cell viability,,8-glucoronidase release, andgeneration of chemotactic factor.Diamond had little effect on cells,silica and hydroxyapatite caused a lotof cell death, hydroxyapatite causedmost enzyme release, and urate themost generation ofchemotactic factor.

CONCLUSIONSThese experiments must be

Table 1 The prevalence of cell death, and the release of /3-glucuronidase andchemotactic factor after 1 hour incubation ofhuman polymorphonuclear cells withvarious crystals in a protein-free medium at 37°C

% cells % release of Chemotactickilled ,3-glucuronidase factor

Cells alone 1 0 14Diamond 2 2 16Silica 33 20 15MSU 19 17 72HA 26 44 37

MSU = Monosodium urate monohydrate; HA = Hydroxyapatite.

interpreted cautiously in view of thedifficulty in comparing differentcrystals (P A Dieppe et al., p. 60).However, they suggest that:

(1) Particle phagocytosis by PMNsmay cause no cell damage or release ofmediators-for example, diamond.

(2) Hydroxyapatite crystals are verytoxic to phagocytic cells and causerelease of a large quantity of enzymes.

(3) Urate crystals are more active inthe generation of a PMN chemotacticfactor than the other particles tested.

Although part of the differencecould be accounted for bycontamination of crystals or thepresence of some mononuclear cells,these data suggest that PMNs respondquite differently to different particles.This may help to explain some of thevariation in inflammatory responses todifferent crystals, and is beingexplored further.

We would like to acknowledge thefinancial support of the Arthritis andRheumatism Council.

References

1 Phelps P, McCarty D J. Crystal inducedinflammation in canine joints: the impor-tance of polymorphonuclear leucocytes.JExp Med 1966; 124: 115-26.

2 Smolen J E, Wiessman G. Polymorpho-nuclear leucocytes. In: McCarty D J, ed.Arthritis and allied conditions. 9th ed.Philadelphia: Lea and Febiger, 1979:282-95.

3 Dieppe P A, Doherty M. The role of par-ticles in the pathogenesis of joint disease.In: Berry C L, ed. Current topics inpathology. Vol 71. Berlin: Springer Ver-lag, 1982: 200-33.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


C5a activation by monosodium urate crystals: effect ofadsorption of serum and polymorphonuclear leucocytehomogenate

A. D. WOOLF AND G. S. PANAYI

From the Departments of Medicine and Rheumatology, Guy's Hospital Medical School, London SE] YRT

Acute gout is characterised byintermittent episodes of acuteinflammation of joints separated bysymptom-free periods. This acuteinflammatory response is not solelydependent on the presence ofintra-articular monosodium uratecrystals (MSU), as they can be found inasymptomatic joints, and the controlmechanisms, particularly of the switchoff, are unresolved. MSU crystals binda wide variety of proteins, especiallyIgG' but also cytoplasmic andlysosomal constituents.2 AdsorbedIgG may promote complementactivation3 and increasepolymorphonuclear leucocyte (PMN)lysosomal enzyme release.4Alterations in the crystal coatingsubsequent to phagocytosis by PMNsmay thus reduce or abrogate theirphlogistic properties. In addition, it ispossible that MSU crystals cause localdepletion of pro-inflammatory factorsby adsorption and thereby modify theinflammatory response.MSU crystals activate serum

complement, leading to the generationof the potent chemotaxin C5a,5 and aprincipal activity of PMN in responseto acute inflammation is chemotaxis.We have, therefore, looked at theability of MSU to generatechemotactic activity in serum and theeffect of serum coating and PMNhomogenate on this ability. We alsolooked at the ability of these crystals toactivate serum depleted byadsorption with MSU.

Synthetic crystals of MSU werecoated with normal human serum at 5mg/ml 4°C for 60 min. Coated or plainMSU crystals were then incubatedwith normal human serum or adsorbedserum for 4 h at 37°C, with or withouthaving been previously incubated for

60 min at 37°C in a PMN homogenate.Adsorbed serum was the supernatantleft after coating the MSU crystals.PMN homogenate was prepared bydextran sedimentation of heparinisedvenous blood, RBC lysis,freeze-thawing and sonicating a

suspension of 2 x 107 PMN/ml. MSUcrystals were washed twice with MSUsaturated sodium phosphate 0-0001mmol buffer, pH 7-4, between allstages, and all incubations performedwith constant mixing. Chemotacticactivity was assayed by a modifiedagarose technique,6 all samples beingheat inactivated at 56°C for 30 min.Migration distance was taken as theaverage distance travelled by the fivemost rapidly travelling cells towardsthe test well.Coated MSU crystals were as

capable of activating normal humanserum and adsorbed serum as plaincrystals. Comparing activation ofadsorbed and non-adsorbed serum,significantly less chemotactic activitywas generated in adsorbed than innormal human serum by coated MSUcrystals in four of five experiments (p= 0-01) and by plain MSU crystals intwo of five experiments (p = 0-01).

Generation of chemotactic activityin fresh serum by plain or coated MSUwas unaffected by treatment withPMN homogenate, but the activationof adsorbed serum by coated MSUcrystals was significantly increasedafter treatment with PMN homo-genate in two of three experi-ments (p = 0.01).

Cell homogenate showed no

significant activation of fresh oradsorbed serum in solution.Adsorbed serum showed normal

generation ofchemotactic activity withzymosan (1 mg/ml 30 min 37°C)

compared with normal human serum.Using a functional assay for C5a we

found that coating MSU crystals inserum, where IgG is the principaladsorbed molecule,' had no effect onthe activation of C5a.The effect of treating coated MSU

crystals with PMN homogenate inincreasing the generation ofchemotactic activity in adsorbedserum cannot be attributed to anydirect activation by PMN homogenateand must be due to changes in theMSU coating.

It is interesting that adsorbed serumwhen exposed to coated MSU crystalsfailed to generate as much chemotacticactivity as similarily treated freshserum. As the coated crystals used toactivate the adsorbed serum in theseexperiments were the same as thoseused in the initial adsorption, then nofactors should have been depletedfrom this system. This suggests thatbinding of serum components to MSUcrystal surfaces may produce changesin the adsorbed material that impairsthe generation of chemotactic activity.This may be an important factor in theself limitation of acute gouty arthritis.

References

1 Kozin F, McCarty D J. Protein binding tomonosodium urate monohydrate, cal-cium pyrophosphate dihydrate, and sili-con dioxide crystals. I. Physical charac-teristics. J Lab Clin Med 1977; 89:1314-25.

2 Ginsberg M H, Kozin F, Chow D, May J,Skosey J L. Adsorption of polymorpho-nuclear leukocyte lysosomal enzymes tomonosodium urate crystals. ArthritisRheum 1977; 20: 1538-42.

3 Hasselbacher P. C3 activation by mono-sodium urate monohydrate is enhancedby surface IgG. Arthritis Rheum 1979;22: 620.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


4 Kozin F, Ginsberg M H, Skosey J L.Polymorphonuclear leukocyte responsesto monosodium urate crystals: modifica-tion by adsorbed serum proteins. JRheumatol 1979; 6: 519-26.

5 Russell I J, Mansen C, Kolb L M, KolbW P. Activation of the fifth component

of human complement (C5) induced bymonosodium urate crystals: C5 conver-tase assembly on the crystal surface. ClinImmunol Immunopathol 1982; 24:239-50.

6 Nelson R D, Quie P G, Simmons R L.Chemotaxis under agarose: a new and

simple method for measuring chemotaxisand spontaneous migration of humanpolymorphonuclear leukocytes andmonocytes. J Irnmunol 1975: 115:1650-6.

Inflammatory effect of aluminium phosphate

P. NETTER, J. L. DELONGEAS, G. FAURE, P. BOZ, D. BURNEL, J. L. FOLIGUET,M. F. KESSLER, R. J. ROYER, A. GAUCHER

Front the Travail de la Clinique Rhumatologique, des Laboratoires de Pharmiacologie, d 'A atooi ie patlhologique, de Chuinlieet d'Histologie, et du Service de Nephrologie, Faculte de Medecine de Nancs, Unilversiu' de Nan(c! 1, 5450() Vaindoeuvre les

Nancv, France

Microcrystalline-induced arthritis isknown to occur in patients undergoinghaemodialysis for chronic renalinsufficiency. The followingmicrocrystals have been implicated sofar in these inflammatory phenomena:apatite, calcium pyrophosphatedihydrate (CPPD), and sodiummonourate.

In preliminary study,' 2 usingscanning electron microscopy (SEM)and wavelength dispersivemicroanalysis (WDM), we foundaluminium deposits associated withphosphate in synovial tissue from apatient undergoing haemodialysisbecause of chronic renal insufficiency,who had been taking aluminiumhydroxide for five years. Aluminiumconcentrations measured with asensitive and reproducible pulsepolarography method were raised inthe synovial fluid (87 Ag/l), synovialtissue (3 7 and 2-2 gg/g wet weight)and in the cartilage (22 4g/g wetweight) compared with controls:synovial fluid: 10-8 + 5 .g/l (n = 6),synovial tissue: 0 8 0 2 g/g (n = 4),cartilage: 8-31 + 2 0 ,ug/g (n = 4).Transmission electron microscopy(TEM) revealed electron densematerial of microfibrillar appearancein lysosomes of macrophagic synovialcells.

In a second part we developedexperimental models to appreciate theprophlogistic properties of aluminium

phosphate using two different animalmodels: (a) paw oedema aftersubcutaneous injection in the rear footpad of Sprague Dawley rats and (b)experimental arthritis after

intra-articular injection in rabbits.Tribasic aluminium phosphate was

compared to calcium hydrogenphosphate dihydrate previously foundduring destructive chondrocalcinosisin humans4; natural diamond powderand lambda carrageenan were chosenas negative and positive controls.

In the first model, inflammationinduced by aluminium phosphatereached its maximum as soon as 15-3t)min after injection but lasted only 24hours. The response to calciumhydrogen phosphate dihydrate was

only minor whereas carregeenan-induced inflammation appeared aftertwo hours but was far greater andlasted mored than 24 hours.The second experimental

proceeding was used to evaluate thehistological consequences of theinjection of these compounds on

synovial tissue and cartilage. Rabbitswere injected in the knee joints every

day for three days. Aluminiumphosphate can induce, likecarrageenan, inflammatory lesions ofthe synovial tissue and erosive damageof the articular cartilage. With TEM,lysosomal inclusions of phagocytosedmaterial with microfilaments were

seen. Through SEM coupled with a

wavelength dispersive microprobeanalysis, aluminium associated w ithphosphate w\as tound in cellularelements. These features are similar tothose observed in the patient in ourpreliminary study and those describedin cerebral tissue from patients withaluminium intoxication.The inflammatorv effect

demonstrated in this study suggeststhat in addition to apatite. calciumpyrophosphate dihydrate and sodiummoniourate. aluminiulmi compoundscould play a part in clinicalmaniifestations observed in patientswith chronic renal insufficiencvundergoing haemodialysis andreceiving aluminium gels.

References

1 Netter P, Bumel D, Hutin M F, KesslerM F. Faure G. Aluminium in joint tissuesof patient taking aluminium hydroxide.Lancet 1981; i: 1056-7.

2 Netter P. L'aluminium dans l'os, lecartilage, la membrane synoviale et leliquide synovial. University of Nancy,Nancy, 1980. Thesis.

3 Royer R J, Delongeas J L, Netter P, et al.Inflammatory effect of aluminium on ratpaws. Pathol Biol (Paris) 1982; 30:211-5.

4 Faure G, Netter P, Malaman B.Steinmetz J. Monocrystalline calciumhydrogen phosphate dihydrate indestructive arthropathies ofchondrocalcinosis. Lancet 1977; ii:142-3.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Introduction Crystal-related arthropathies

Documents

Transcript of Introduction Crystal-related arthropathies