Cartilage macrophages immunopathogenic concept[alo]2c2.1 In this group the interstitial types I, II,...

Annals ofRheumatic Diseases, 1982, 41, 168-176

Cartilage specific collagen activates macrophagesand the alternative pathway of complement: evidencefor an immunopathogenic concept of rheumatoidarthritisH. M. HANAUSKE-ABEL,l B. F. PONTZ,l* ANDH. U. SCHORLEMMER2

From the 'Kinderklinik and 2Institut flir Med. Mikrobiologie, Johannes Gutenberg- Universitdt, 6500 Mainz,Federal Republic of Germany

SUMMARY We studied the effect of human interstitial collagen types 1, II, and III on serum-freecultured mouse macrophages and on the complement classical and alternative pathways in humanand guinea-pig serum. Type II collagen produced a dose-dependent consumption and conversionof C3 and factor B both in the homologous and in the heterologous system. This effect on thealternative pathway was reproduced in genetically C4-deficient guinea-pig serum and could betriggered by native, triple helical type II molecules, by their component a chains, and the CNBrpeptide mixture. Addition of type II collagen to the mouse macrophage cultures induced not onlya dose- and time-dependent secretion of lysosomal enzymes, but also the generation of a super-

natant factor cytotoxic for mouse mastocytoma P 815 cells. Collagen of types I and III were

conspicuously less active or inactive in all assays. The studies demonstrate properties of the collagenspecific for cartilage which, on a molecular level, suggest its direct, local participation in theproduction and perpetuation of rheumatoid arthritis.

During the last decade one of the breakthroughs inthe biochemistry of the intercellular matrix has beenthe demonstration that 'collagen' consists of severalgenetically distinct proteins with varied distributionin developing, adult, and pathological tissues. Todate at least S trimeric, triple-helical collagen typeshave been identified and characterised as [ocl]3 or[alo]2c2.1 In this group the interstitial types I, II,

Accepted for publication 11 March 1981.Correspondence to Dr H. M. Hanauske-Abel, Kinderklinik,Johannes Gutenberg-Universitat, D-6500 Mainz, FederalRepublic of Germany.*Present address: Max-Planck-Institut fur Biochemie, Abt.fur Bindegewebsforschung, Munchen-Martinsried, FRG.A preliminary account of parts of this work was presentedat the Eighth International Complement Workshop, 13-16October 1979, Key Biscane, Florida, USA.Abbreviations used in this paper: APC, alternative pathwayof complement; LDH, lactate dehydrogenase. CNBr,cyanogen bromide; NADH2, f-nicotineamide-adenine-dinucleotide, reduced; EDTA, ethylenediaminetetra-aceticacid; DEAE, diethylaminoethyl; CM, carboxymethyl; SDS-PAGE, sodium dodecyl sulphate polyacylamide gel electro-phoresis.

and III are best characterised with respect to puri-fication, alignment of type-specific CNBr-peptides,composition, and sequence of aminoacids as wellas the immunological behaviour.'

In vivo type II collagen was the first homologousprotein antigen that did not require completeFreund's adjuvant for the induction of an auto-immune response characterised by high antibodytitres and cellular sensitivity. This unique immuno-genicity of type II is thought to account for its arth-ritogenic properties in rats: intradermal injection ofnative cartilage collagen causes a chronic synovitiswith secondary erosion of articular cartilage andsubchondral bone.Z4 Cells from sensitised donorsare capable of passively transferring this response.5The application of types I and III does not triggeran autoaggressive reaction. That is why type IIcollagen-provoked arthritis has been suggested as anappropriate model for rheumatoid arthritis in man.3A similar type II collagen-induced arthritis can beproduced in mice.6

In patients suffering from rheumatoid arthritis168

on May 1, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.41.2.168 on 1 April 1982. D

ownloaded from

http://ard.bmj.com/

Consequences ofAPC activation by type II collagen 169

activation of the complement system obviously playsan important role by releasing phlogogenic peptidesand proteins. Involvement of the classical pathwayis evident by the reduction of individual componentsand concomitant demonstration of immune com-plexes in synovial effusions,7 while diminished con-centrations of factor B and properdin point to theparticipation of the alternative pathway.8 Exclusiveactivation of the alternative pathway is shown in thesynovial fluid of arthritic patients with Bruton-typeagammaglobulinaemia.9The present study was carried out to investigate

the interaction of purified interstitial collagen withthe alternative pathway of complement and withmacrophages.

Materials and methods

Experimental animals. Male mice (outbred strainNMRI/Bom) were purchased from G. L. Bomholt-gard, Ry, Denmark.

Tissue culture materials. Tissue culture grade Petridishes were obtained from Nunc GmbH, Wiesbaden,Germany, fetal bovine serum and the TC 199 mediafrom Flow Laboratories GmbH, Bonn, Germany.

Biochemical reagents. Penicillin and streptomycinwere from Flow Laboratories GmbH, Bonn,Germany, pyruvate and 3-nicotinamide-adenine-dinucleotide from Boehringer GmbH, Mannheim,Germany, phenolphtalein glucuronic acid fromSigma GmbH, Munich, Germany, p-nitrophenyl-2- acetamido - - D-desoxyglucopyranoside fromKoch-Light Laboratories, Colnbrook, UK, preserva-tive free heparin from Nordmark Werke GmbH,Hamburg, Germany, and Triton X-100 from ServaGmbH, Heidelberg, Germany. The Limulus test(Pyrogent) was purchased from Byk-Mallinckrodt,Dietzenbach, Germany, the Protease Detection Kitfrom Bio-Rad, Munich, Germany.

Incubation for C3 activation and haemolytic assayfor C3 activity. Standard incubation mixtures for C3activation and the assay for haemolytic C3 measure-ment were performed as detailed by Bitter-Suermannet al.'0 Equal volumes of either pooled normal orC4-deficient guinea-pig serum were incubated withappropriate dilutions of the test materials for 30 minat 370C and cooled to 40C in an ice bath; the con-centrations of the collagen types referred to in thetext represent final concentrations in this standardincubation mixture ([Lg/ml). Controls were run withwater instead of test materials. From the standardincubation mixture samples were withdrawn andsubdiluted to a final serum concentration of 1: 100,1: 500, and 1: 1000. The amount of residual haemo-lytically active C3 (site forming units; SFU) wasdetermined and calculated as previously described.10

The values obtained were expressed as percentagesof the haemolytically active C3 in the controls. TheC3 content of normal and C4-deficient guinea-pigserum in the controls was in the range of 1-2 x 1011SFU C3/ml.

Haemolytic assay for Cl, C2, and C4 activity.Haemolytic Cl, C2, and C4 activities were assayedby the method of Muller et al.11

Immunoelectrophoretic analysis. Agarose (1 %) con-taining 10-2 M EDTA was used in all experiments.Electrophoretic separation was performed at pH 8 6with 9 volts/cm for 120 min. 10 ,ul samples of thestandard incubation mixtures were used. After fixa-tion the slides were stained with Coomassie brilliantblue R 250. Antisera were a gift from ProfessorBitter-Suermann (Mainz).

Macrophage collection, culture, and stimulation.Mouse macrophages were obtained by peritoneallavage with 5 ml TC 199 containing penicillin/streptomycin (100 units/ml) and 10 IU/mi heparin.2 ml aliquots of this peritoneal exudate cell suspen-sion (0- 5-1 - 0 x 106 cells/mil) were cultured in Petridishes (diameter 35 mm) and incubated in a humidi-fied atmosphere of 5% CO2 at 370C for 2 hours toallow attachment of adherent cells. After vigorouslywashing 4 times with phosphate buffered saline, thecells were cultured in serum-free TC 199. Macro-phages prepared in this way give a sheet of wellspread cells within 24 hours. For in-vitro activationthe dissolved test materials were added to the con-centrations indicated. After the test period macro-phages and supernates were collected and checkedfor enzyme secretion and cytotoxicity. In all experi-ments triplicate and quadruplicate cultures wereused. The results obtained are expressed as the meanand standard deviation.Enzyme assays. At the end of the test periods the

medium was removed and the adherent cells werewashed once with phosphate buffered saline. Thecells were then released by adding saline containing0-1% (w/v) Triton X-100, and scraped off with asterile rubber 'policeman.' The activity of the variousenzymes were tested in media and cells. All assayswere conducted under conditions that provide forthe linear release of the products in relation to theamount of test sample and the time of incubation.Lactate dehydrogenase was assayed by determiningthe rate of oxidation of NADH2 at 340 nm.'2 Therepresentative lysosomal enzymes P-glucuronidaseand N-acetyl-P-D-glucosaminidase were assayed bythe method of Stephens et al.13Method of cytotoxicity assay. As the release of

lactate dehydrogenase into the culture supernate isin indication of cell death, the 51Cr release assay wasreplaced by the determination of LDH. The 2



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

170 Hanauske-Abel, Pontz, Schorlemmer

competitive methods-LDH or 51Cr release-werecompared in cytotoxicity studies, and the LDHrelease technique was found to be as sensitive andprecise as the radioactive assay.14 The assay for thecytoplasmic LDH was conducted under conditionsgiving linear product release with respect to sampleamount and incubation time; LDH was determinedas given in Bergmeyer and Bernt.12 In positiveexperiments up to 90% of the cellular LDH contentappeared in the supernate within 2 hours of incuba-tion; that is why this time interval was chosen astest period. Control supernatants showed very lowspontaneous background release of LDH even after6 hours.

Cytotoxicity assay of supernates from stimulatedmacrophages. P815 cells, derived from a chemicallyinduced mouse mastocytoma, served as targets.Unstimulated mouse macrophages cultured underserum-free conditions in the presence of the indi-cated collagen type II concentrations were used aseffectors. Supernates were harvested at the givenintervals. 5 x 105 target cells in 0 1 ml of mediumwere mixed with 0-1 ml of the culture supernatantsin round bottomed plastic tubes. The tubes werefilled with air containing 5% CO2. After incubationat 37°C for 2 hours 0-8 ml TC 199 was added to thetubes, the cells were centrifuged, and 0-1 ml samplesof this supernate tested for LDH release. The per-centage of LDH release is given relative to that oftarget cells exposed to medium from nonstimulatedmacrophages; this value did not differ significantlyfrom that observed with target cells alone. The valuespresented are means of triplicate experiments.

Preparation of type I collagen. Adult human duramater, obtained at necropsy, was freed from thegreat vessels, homogenised, and incubated withpepsin at 10 mg/ml, pH 2-8, 150C. After 24 hoursthe initial amount of pepsin was added, the pHreadjusted, and the digestion allowed to continuefor another 24 hours at 150C. After centrifugationthe supernatant was dialysed against 0*02 MNa2HPO4; the deposit was dissolved in 0 1 M aceticacid and treated by a modification of the methodoutlined by Epstein.15 The 2-5 M NaCI precipitatewas reprecipitated and then extracted with 1-5 MNaCl, 0.05 M Tris-HCI, pH 7*5, the undissolvedmaterial excluded, and the supernate dialysed against2-5 M NaCI at neutral pH.

Preparation of type II collagen. The cartilagenousparts of tibiae, humeri, and femora were collectedat necropsy of fetuses of up to 12 weeks gestation.The perichondrium was removed by careful dis-section on ice prior to homogenisation. The majorityof the proteoglycans was dissolved by the methodof Lohmander by means of repeated guanidineextractions.'6 The pepsin digestion used to solubilise

the collagen followed the method of Miller17 withthe minor modification of increasing the pepsinconcentration to 10 mg/ml. The collagen wasprecipitated by the addition of crystalline NaCl to afinal concentration of 0 9 M, redissolved in 1 MNaCl, 0 05 M Tris-HCI, pH 7-5 at 40C, and pre-cipitated twice by dialysis against 0 02 M Na2HfPO4.In the cold this type II collagen was solubilised in0-2 M NaCl 0 05 M Tris-HCI, pH 7-5 and thenpassed through a DEAE-cellulose column (DE 52,Whatman Chemicals) to withhold residual pro-teoglycans.18 As assessed by the uronic acid assay,19this final collagen type II preparation was free fromproteoglycan contamination.

Preparation of type III collagen. Skin was obtainedfrom babies at necropsy, the subcutaneous fat wasscraped off, and the dermis minced together withcrushed ice in a meat grinder. The homogenisedskin was extracted with 1 M NaCI, 0 05 M Tris-HCI,pH 7- 5, and with 0 5 M acetic acid at 40C, in eachcase for 2 days, to remove contaminations like serumproteins and fat. The undissolved material wasdigested with pepsin and chromatographicallypurified according to the method of Epstein.15

Cyanogen bromide cleavage of type IL The CNBrcleavage of purified type II collagen was carried outas detailed by Epstein.15 CM-cellulose (CM-52,Whatman) chromatography of the CNBr peptidesfollowed the method of Miller.17

Collagen analysis. As well as by the chromato-graphic methods collagens were characterised bySDS-PAGE electrophoresis and amino acid analysis.SDS-PAGE was performed according to the methodof Furthmayr and Timpl20 on whole collagen chainsusing 5% acrylamide gels. For amino acid analysissamples were hydrolysed in 2 ml of boiling 6 NHC1at 111 C for 22 hours under nitrogen in sealedtubes. The acid was removed by vacuum and thesamples used for analysis on a Biotronik LD 6000 Ewith a 4-buffer elution system. Each collagen pre-paration used in the assays was subjected to thelimulus test to quantify its endotoxin content. Allpreparations were positive for endotoxin, but thesmall differences-type I 80 ng/ml, type II 40 ng/ml,type III 60 ng/ml-at identical collagen concentra-tions-1 mg/ml-can clearly not account for thestriking type-specific effects observed in the assays.

All collagen samples were checked for possibletrace contaminants with proteolytic activity with theBio-Rad Protease Detection Kit. This sensitive assay-it allows measurement of proteases in the range ofI ,Lg/ml-is based on the radial diffusion of enzymesinto a thin layer containing casein at physiologicalpH. A proteolytic contamination of the samplescould not be demonstrated for collagen concentra-tions of I mg/ml.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Results

INTERACTION OF COLLAGEN TYPE II WITHTHE ALTERNATIVE PATHWAY OF THECOMPLEMENT SYSTEMIn experiments to establish the dose-response curvefor C3 consumption in normal guinea-pig serumtype II was about 10 times more effective-50%consumption at 40 [g/ml-than type I collagen,while type III was virtually inactive. As no effect oftype II on Cl, C2, or C4 could be measured at con-centration of up to 100 Ftg/ml, this suggested thatantibodies against the human collagen type formingimmune complexes and thereby activating theclassical pathway were not present in the guinea-pigserum used. This result led us to propose a directinteraction with the alternative pathway. The use ofgenetically C4-deficient guinea-pig serum confirmedthis assumption (Fig. 1). We attribute the smalldifference in C3 turnover to different kinetics inboth sera observed also for other activators of thealternative pathway: at identical concentrations C3turnover by these substances can be 10%-15% lowerin C4-deficient guinea-pig serum than in normalguinea-pig serum, and gives identical results onlyupon prolongation of the incubation time.21 Asshown on the right side of Fig. 2, incubation ofC4-deficient guinea-pig serum with type II collagenat the routine temperature of the assay producedan 80% reduction of the haemolytically active C3molecules. Native and heat denatured collagen gave

100

80

C .J(o0D

0

20'-

0 50 100 200 300

collagen type II ( pg/ml)

Fig. I C3 turnover in normal (*-e) and C4-deficient(0-o) guinea-pig serum as a function of type II

collagen concentrations.

I/C31001

80-

60

40-

20-

E C4 def. g.p. serum

D C def.g.p. serum+250 yg/ml CoUll.I

D C def.g.p. serum1: : 250 pg/mi Coll.l

(56"C.30')

,

....

::::

:-:-:-:

200 CH

37°CFig. 2 C3 turnover in C4-deficient guinea-pig (C4-def.g.p.) serum induced by type I1 collagen (coll. IH). Atthe routine test temperature of37°C (right side) bothnative and heat-denatured type II collagen give areduction ot the haemolytically available C3 to less than20% of the control. To render denaturation of thecollagen triple helix during the test period impossible,the same experiment was performed at 20°C (left side).Again, the C3 consumption was independent of thetype II collagen conformation.

a comparable degree of C3 turnover, and so thiseffect was independent of the triple helical con-formation.

Collagen has a sharp thermal transition with a Tmof about 380C for the triple helix-random coil con-version.1 As the routine haemolytic assays for C3activity are incubated at 370C for 30 minutes, wehad to take special precautions to secure an un-affected triple helical conformation. Accordingly weran assays at 200C, a temperature that maintains thenative conformation during the test period. How-ever, we could not find a dependence of the C3consumption on the triple helical structure of type IIcollagen. Deliberate thermal destruction of thenative conformation which results in the completeseparation of the component cx chains-type If lacksintramolecular links-does not show a clear-cuteffect on the C3 depletion at 20°C; the slightlyincreased C3 turnover at 370C can be easily attri-buted to a direct thermal effect on the enzymes activein the assay (Fig. 2, left side). The finding that the C3consumption was not abolished after heat denatura-tion of type II renders the assumption of a residualC3 cleaving proteolytic activity in the purifiedcollagen samples unlikely.The data presented above are inconclusive with



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


regard to the mechanism of alternative pathway aactivation by type II collagen. They may simply sresult from the passive adsorption of C3 to the ccollagen polypeptide chains, thus producing a ndiminution of haemolytically available C3 molecules.To investigate this possibility the immunoelectro- E

phoretic behaviour of C3 and factor B was analysed cafter incubation of human and C4-deficient guinea- 1pig serum with type 11 (250 ,ug/ml, 30 minutes). The tBb determinants located in the respective cleavage sproducts of human factor B were detected by the Ispecific antiserum employed (Fig. 3). The C3 tantiserum revealed the presence of the anodally fmigrating C3b fragment in CA-deficient guinea-pig Iserum (Fig. 3). In the control both complement tcomponents remained in the zymogen form. The tconversion of C3 and factor B in human serum bytype II collagen does not depend on its triple helical econformation as its a chains and the CNBr peptide tmixture, consisting of nonhelical chain fragments ofvarying length, also mediate the cleavage of C3 and e

factor B. The activation of both components alsowas evident in normal and genetically C4-deficientguinea-pig serum. As judged by the alteration of the Iimmunoelectrophoretic behaviour, the effect of i

types I and III on the conversion of C3 and factor Bwas less pronounced or absent. Human type II Icollagen triggers the human APC by promoting thecleavage of C3 and factor B to their physiologicalfragments. The effects of the component a chains I

(7.: .----

__

.T~ ~ "J

VIvX,

nd the CNBr peptides give conclusive evidence touggest that determinants of the primary structuref type II collagen other than the methionine residuesmediate the observed C3 consumption.

EFFECTS OF HUMAN COLLAGEN TYPE II)N MOUSE MACROPHAGES[he addition of types I, II, III at different concen-trations to the macrophages had striking typespecific effects. After 48 hours 100 ,ug/ml of types I orIII had not stimulated macrophages to secrete atumorolytic principle, and the activity did not differfrom that of the nonstimulated control macrophages.However, macrophages incubated with 100 ,ug/mltype II produced nearly complete target cell destruc-tion. Only type II incubated macrophages developeddefinite dose-response characteristics (Fig. 4). Toelucidate the kinetics of type II generated cyto-toxicity macrophages were incubated with 100,ug/ml type II and their tumorolytic capacity testedat the intervals given in Fig. 5. After 6 hours adefinite increase in cytotoxicity was measurable, andafter 48 hours the values reached a plateau of nearly100% specific lysis. In the controls the alterationswere negligible.Moreover the induced release of lysosomal hydro-

lases from macrophages was markedly type specific.After 48 hours' exposure to 200 ,g/ml collagen theamount of 3-glucuronidase secreted by the type ILtreated cells was significantly elevated above the

Fig. 3 Immunoelectrophoreticcoil. type I1 conversion of C3 in C4-deficientantiC3 guinea-pg serum (C4-df...

ant 0C3 and offactor B, that is, C3C/4 def. g. p s activator, in human serum (h.s.)

by type II collagen (coll. type II).Above: Generation of anodallymigrating C3b following incuba-tion of C4- deficient guinea-pigserum with 250 lsg/ml type IIcollagen. Below: Generation ofcathodally migrating Bb followingincubation ofhuman serum with250 ILglml type 11 collagen. A

coil, type 11 certain amount ofBb appeared

anti C3 Activator spontaneously on coagulation.

v1s.s



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


0 25 50 100Concentration ( jg/ml )

0 25 50 100 200 Fig. 6 Specific release of P-glucuronidase into theculture medium of mouse macrophages after 48 hours'collagen type 1I (ug / ml ) exposure to increasing amounts of type II collagen(.*-). Release of cytoplasmic LDH in the same

rulZtijrp. (r-n)Fig. 4 Lysis ofP815 mastocytoma cells by supernatantsof macrophages activated in vitro by differentconcentrations of type IH collagen.

100-

80-

1.T2, 60-

0 /CL 40-

201

2 6 22 30 48 71Time (hr)

Fig. 5 The time-dependent effect of type HI collagen-stimulated, mouse-macrophage-mediated cytoxicityagainst P815 cells (e-*). Controls (0O-). Type HIcollagen concentration: 100 Lg/ml.

control level. The reaction was dose-dependent(Fig. 6). The type I and type III treated cells, how-ever, displayed neither dose-response characteristicsnor an increase in enzyme release. This effect oftype II collagen was specific and not associated withrelease of LDH or morphological signs of cell death.Type II had a pronounced dose-dependent effect onthe secretion of another lysosomal enzyme, N-acetyl-f3-D-glucoseaminidase.

The capacity of type LI collagen to induce lyso-somal enzyme release was unaffected after CNBrdigestion.

Discussion

The in-vitro results presented in this paper demon-strate for type II collagen preferentially an activationof the alternative pathway and a stimulation ofmacrophages to secrete lysosomal enzymes and tobecome cytotoxic. This effect, unaltered after heatdenaturation or CNBr cleavage, rests on primarystructure determinants.The interaction of the complement system with

intersitial collagen has been studied before.Takahashi et al.22 presented evidence for an activa-tion of the alternative pathway at very high collagenconcentrations (>10 mg/ml), but as their prepara-tions were characterised by the hydroxyprolinecontent only, APC activating or inhibiting con-taminations cannot be ruled out. Valet et al.23 foundthat more than 100 Ftg/ml of purified interstitialcollagens were needed for 50% reduction in the Clhaemolytic assay, without type differences. Belowthat concentration we too could not demonstrate aneffective activation of the classical pathway. How-ever, our findings clearly indicate that type Itcollagen in particular can activate the alternativepathway, as exemplified by the studies with C4-deficient guinea-pig serum.Hopper et al.24 investigated the interaction of

macrophages with type I collagen and reported adependence on the conformation; denatured collagen

F-I--601100 -

80 -

n

>2 60-.U

a 40-

20 -

E.2 40-'a

Ec

E 20-Nc

0

200

T.L

I

cullurva kk-,-kj)-



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


is bound more easily but internalised more slowly.A functional parameter of mononuclear phagocytes,namely, chemotactic responsiveness, was found to besensitive in a dose-dependent manner to type Icollagen, its component a chains, and CNBrpeptides.25 No further collagen types were tested inthese studies. Of the different collagen types scruti-nised here only type II gave a considerable stimula-tion of macrophages, as judged by enzyme secretionand generation of cytotoxicity in the supernate.We found the same pattern of collagen type

specificity as Trentham et al.3 and Courtenay et al.6in their animal studies on collagen arthritogenicity:type II was the only collagen active in all test systems.With respect to its conformational state, however,the in-vitro and in-vivo effects differed. While thecapacity of its component a chains to activatemacrophages and the APC remained unaffected, thea. chains did not produce arthritis when injectedintradermally in complete or incomplete Freund'sadjuvant.

Prior to type II collagen a number of substanceswere reported to activate both the alternative path-way of complement and normal, serum-free culturedmacrophages. These substances include streptococcalcell walls, zymosan, and polyanions like carrageenanand dextran sulphate.26

After intra-articular injection each of these sub-stances can induce an inflammatory reaction reminis-cent of human rheumatoid joint lesions.27-29 In viewof the parallel effect of type 1I collagen on macro-phages and the APC, which resembles that of thesesubstances, and in view of its arthritogenic potential,the results reported here call for comparison with anumber of articular findings in rheumatoid arthritis:

(1) In the rheumatoid synovial membrane largenumbers of macrophages are observed, someassociated with blastic lymphocytes and even cyto-plasmatically bridged to them; fibroblasts close tomacrophages frequently display signs of cell damageas shown by vacuolisation, disappearance of cyto-plasm, and depletion of nuclear chromatin.30Macrophage-like cells at the cartilage-pannus inter-face invade and degrade the articular type II collagenmatrix.3'

(2) In the rheumatoid synovial fluid the localproduction of the representative lysosomal enzymesP-glucuronidase and N-acetyl-3-D-glucoseaminidaseis markedly elevated and independent of the whitecell counts.'3 Activation of the complement systemby both the classical and the alternative pathway iswell documented.7 8 While rheumatoid synovial fluidcontains normal or increased amounts of proteaseinhibitors like oal antitrypsin, 1X2 macroglobulin,32and Cis inhibitor,33 the concentration of the C3b

inactivator in rheumatoid synovial fluid is signi-ficantly reduced in spite of supranormal serumvalues,34 pointing to increased local catabolism andfacilitated APC activation due to the diminution ofthis regulatory protein.These in-vivo findings in rheumatoid patients show

a correlation with the in-vitro effects of type IIcollagen on macrophages and the APC, which con-sisted in activation of C3 and factor B, generation ofcytotoxicity, and secretion of the lysosomal enzymesP-glucuronidase and N-acetyl-3-D-glucoseaminidase.While this does not prove a causal connection, ourresults suggest that type II collagen and its degrada-tion products can mediate a local activation ofmacrophages and the APC in rheumatoid joints.Once this interaction, possibly initiated by exogenousantigens,35 is established on the articular surface, itwill not need a persisting agent to run a chroniccourse and cause extensive injury. It centers on thebasic amplification loop of type 1I collagen-activatedmacrophages and type II collagen rendered acces-sible by activated macrophages, as these cellsparticipate in denudation and destruction of cartilagecollagen.30 31 In this way they also provide for aconstant flow of APC-inducing type II collagenmaterial which in turn contributes to complementdepletion and enhances inflammatory tissue damage,for example, via C3b activated macrophages.26

If not counterpoised completely, the inflammationresulting from this regional deregulation tends toeradicate the type II collagen releasing matrix dueto the autocatalytic type II mediated activation ofmacrophages and the APC, even after the elimina-tion of the inciting agent. Owing to the afferentfunctions of macrophages and the APC6.37 asmouldering articular type 1t mediated activation ofmacrophages and the APC may in turn affect thehumoral and cellular responses, culminating in aresetting of autorecognition, a genetic predispositionprovided. Indeed, certain HLA D antigens areassociated with adult-onset, seropositive rheumatoidarthritis, especially its severe form,38 and in micethe susceptibility to type II collagen-inducedarthritis appears under H-2-linked control.6 Thisconceptualisation of our results is in keeping withthe suggestion by Glynn35 that, owing to a geneticsusceptibility, a local inflammatory response be-comes deregulated in rheumatoid arthritis, leadingto a self-perpetuating reaction against componentsof the immune system and material released by theinflammatory tissue degradation.Although not the primary event, we suppose that

a type 11 mediated activation of macrophages andthe APC might constitute one of the epicentres thatperturb the immunological network in rheumatoid



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


arthritis. Under normal conditions the establish-ment of this interaction may be forestalled by theeffective control of the APC via the regulatoryproteins c 1H and C3b inactivator. More signi-ficantly, type II collagen is separated by a many-

layered shield from all elements of the immunesystem: its occurrence is restricted to avascular siteslacking lymphatics. These segregated sites physio-logically exclude both egress and access of even

low-molecular-weight proteins, a property restingon the molecular sieve effect of the matrix glyco-saminoglycan and proteoglycan aggregates.39 It isintriguing to see that type LI collagen can be foundonly in these immunologically protected tissues.

We greatly appreciate the outstanding technical assistanceof Miss A Borner. H.M. Hanauske-Abel was supported bya grant from the Evariste Galois fund.

References

1 Bornstein P, Sage H. Structurally distinct collagen types.Ann Rev Biochem 1980; 49: 958-1003.

2 Trentham D E, Townes A S, Kang A H. Autoimmunityto type ll collagen: an experimental model of arthritis.J Exp Med 1977; 146: 857-68.

3 Trentham D E, Townes A S, Kang A H, David R J.Humoral and cellular sensitivity to collagen in type 11collagen-induced arthritis in rats. J Clin Invest 1978; 61:89-96.

4 Stuart J M, Cremer M A, Dixit S N, Kang A H, TownesA S. Collagen-induced arthritis in rats. Comparison ofvitreous and cartilage derived collagens. Arthritis Rheum1979; 22: 347-52.

5 Trentham D E, Dynesius R A, David J R. Passive transferby cells of type It collagen-induced arthritis in rats. JClin Invest 1978; 62: 359-66.

6 Courtenay J S, Dallman M J, Dayan A D, Martin A,Mosedale B. Immunisation against heterologous type Itcollagen induces arthritis in mice. Nature 1980; 283:666-8.

7 Zvaifler N J. Rheumatoid arthritis. An extravascularimmune complex disease. Arthritis Rheum 1974; 17:297-305.

8 Ruddy S, Fearon D T, Austen K F. Depressed synovialfluid levels of properdin and properdin factor B inpatients with rheumatoid arthritis. Arthritis Rheum 1975;18: 289-95.

9 Munthe E, H0yerall H M, Fr0land S S, Mellbye 0 J,Kass E, Natvig J B. Evidence for complement activationby the alternative pathway in the arthritis of hypo-gammaglobulinemic patients. Rheumatology 1975; 6:43-51.

'0 Bitter-Suermann D, Hadding U, Melchert F, WellensiekH J. Independent and consecutive activation of C5, C6,and C7 in immune hemolysis. 1. Preparation of EACI-5with purified guinea pig C3 and C5. Immunochemistry1970; 17: 955-62.

" Muller W, Hanauske-Abel H M, Loos M. Biosynthesisof the first component of complement by human andguinea pig peritoneal macrophages: evidence for anindependent production of the Cl subunits. J Immunol1978; 121: 1578-84.

12 Bergmeyer H U, Bernt E. Lactat-Dehydrogenase. In:Bergmeyer H U, ed. Methoden der enzymatischen Analyse.Weinheim: Verlag Chemie, 1970: 533-41.

13 Stephens R W, Gosh P, Taylor T K F, Gale C A, SwannJ C, Robinson R G, Webb J. The origins and relativedistribution of polysaccharidases in rheumatoid andosteoarthritis fluids. J Rheumatol 1975; 2: 393-400.

14 Bitter-Suermann D, Becker S, Meuer S, SchorlemmerH U, Hadding U, Andreatta R. Comparative study onbiological effects of the guinea pig complement-peptideC3a and C3a-related synthesic oligopeptides. MolImmunol 1980; 17: 1257-61.

15 Epstein E H. [a 1 (III)]3 human skin collagen. J BiolChem 1974; 249: 3225-31.

16 Lohmander S. Proteoglycans of guinea pig costal carti-lage. Eur J Biochem 1975; 57: 549-59.

17 Miller E J. Structural studies on cartilage collagen em-ploying limited cleavage and solubilisation with pepsin.Biochemistry 1972; 11: 4903-9.

18 Miller E J. Isolation and characterization of a collagenfrom chick cartilage containing three identical oc chains.Biochemistry 1971; 10: 1652-8.

19 Bitter T, Muir H M. A modified uronic acid carbazolereaction. Anal Biochem 1962; 4: 330-6.

20 Furthmayr H, Timpl R. Characterization of collagenpeptides by sodium dodecylsulfate-polyacrylamide elec-trophoresis. Anal Biochem 1971; 41: 510-6.

21 Burger R. Sulfatierte Dextrane als Aktivatoren desKomplementsystems. Doctorial Thesis at the TechnischeHochschule, Darmstadt D 1976; 17: 1-89.

22 Takahashi M, Kawachi-Takahashi S, Matsuura M.Interaction of collagen with serum complement: inhibi-tion of complement-mediated hemolysis. Int ArchsAllergy Appl Immunol 1975; 48: 642-52.

23 Valet H, Kahle H, Timpl R. Inhibition of the Cl com-plement molecule by collagen and procollagen peptides.Immunobiology 1978; 155: 71.

24 Hopper K E, Adelmann B C, Gentner G, Gay S. Recog-nition by guinea pig peritoneal exsudate cells of con-formationally different states of the collagen molecule.Immunology 1976; 30: 249-59.

25 Postlethwaite A E, Kang A H. Collagen and collagenpeptide-induced chemotaxis of human blood monocytes.J Exp Med 1976; 143: 1299-307.

26 Schorlemmer H U, Bitter-Suermann D, Allison A C.Complement activation by the alternative pathway andmacrophage enzyme secretion in the pathogenesis ofchronic inflammation. Immunology 1977; 32: 929-40.

27 Schwab J H, Cromartie W J, Ohanian S H. Associationof experimental chronic arthritis with the persistance ofGroup A streptococcal cell walls in the articular tissue.J Bacteriol 1967; 94: 1728-35.

28 Kaystone E C, Schorlemmer H U, Pope C, Allison A C.Zymosan-induced arthritis. A model of chronic proli-ferative arthritis following activation of the alternativepathway of complement. Arthritis Rheum 1977; 20:1396-401.

29 Carmichael D J, Gillard G C, Lowther D A, HandleyC J, Santer V B. Carrageenin-induced arthritis. IV. Ratechanges in cartilage matrix proteoglycan synthesis.Arthritis Rheum 1977; 20: 834-40.

30 Ishikawa H, Ziff M. Electronmicroscopic observations ofimmunoreactive cells in the rheumatoid synovial mem-brane. Arthritis Rheum 1976; 19: 1-14.

31 Harris E D, Glauert A M, Murley A H G. Intracellularcollagen fibers at the pannus-cartilage junction inrheumatoid arthritis. Arthritis Rheum 1977; 20: 657-65.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


32 Shtacher G, Maayan R, Feinstein G. Protease inhibitors 36 Symposium: Macrophage functions in immunity. Fedin human synovial fluid. Biochim Biophys Acta 1973; Proc 1978; 37: 77-103.303: 138-47. 3 lu ,Hmhe .Tegnrto fmmr33 Hedberg H, Laurell A B. The concentrations of the eKlaus G G B, Humphrey J H. The generation of memoryfourth component of complement and of the Cl inacti- cells. I. The role of C3 in the generation of B memoryvator in synovial fluid from arthritis patients. Clin Exp cells. Immunology 1977; 33: 31-9.Immunol 1972; 11: 201-8. 38 Rossen R D, Brewer E J, Sharp R M, Ott J, Templeton

34 Whaley K, Schur P H, Ruddy S. C3b inactivator in the J W. Familial rheumatoid arthritis. J Clin Invest 1980;rheumatic diseases. J Clin Invest 1976; 57: 1557-63. 65: 629-42.

3 Glynn L E. Experimental models and etiology of inflam-matory rheumatic diseases. Scand J Rheumatol 1975; 39 Stockwell R A. Biology of Cartilage Cells. Cambridge:5: suppl 12: 55-62. Cambridge University Press, 1979: 30-1; 71-2.



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Cartilage macrophages immunopathogenic concept[alo]2c2.1 In this group the interstitial types I, II,...

Documents

Transcript of Cartilage macrophages immunopathogenic concept[alo]2c2.1 In this group the interstitial types I, II,...