Annals ofthe Rheumatic Diseases 1990; 49: 747-752 Measurement of the chemotactic complement fragment C5a in rheumatoid synovial fluids by radioimmunoassay: role of C5a in the acute inflammatory phase P J Jose, I K Moss, R N Maini, T J Williams Abstract Evidence suggests that complement is acti- vated in rheumatoid joints. A sensitive radio- immunoassay for the activation fragment of C5, C5a, which is a potent chemoattractant for neutrophils, was therefore developed. A mean CSa concentration in 22 rheumatoid joint fluids of about 2-5xlO-9 mol/l was found. This concentration of C5a is sufficient to induce two of the characteristic features of the acute inflammatory phases of rheumatoid arthritis: neutrophil accumulation and micro- vascular plasma protein leakage. In animal models it has been shown that C5a is a potent inducer of inflammatory oedema by a neutro- phil dependent mechanism. A striking feature of the acute inflammatory phases of rheuma- toid arthritis is the appearance of high numbers of neutrophils in the synovial fluid. It is suggested that CSa might have a role in mediating neutrophil accumulation and, as a consequence, may be important in acute joint swelling and pain. Section of Vascular Biology, MRC Clinical Research Centre, Harrow, Middlesex P J Jose T J Williams Department of Clinical Research, Charing Cross and Westminster Medical School, London W6 I K Moss Department of Immunology of Rheulmatic Diseases, C g Cross and Westmister Medical School, London W6 R N Maini Correspondence to: Dr P J jose, Departnent of Applied Pharmaclogy, National Heart and Lung Institute, Dovehouse Street, London SW3 6LY. Accepted for publication 21 September 1989 In addition to the chronic synovitis of rheuma- toid arthritis (RA) are the acute flare-up re- actions characterised by extravasation of neutro- phils and plasma proteins. Evidence suggests that the complement system is activated in rheumatoid arthritis.' One product of comple- ment activation is CSa, a potent mediator of neutrophil infiltration and oedema.2'10 The binding of CSa to receptors on leucocytes often renders it undetectable in the fluid phase even when complement activation has been clearly shown. " We developed a sensitive radio- immunoassay for the measurement of CSa in inflammatory exudate fluids, and in this paper we report for the first time as far as we know, the C5a concentrations in rheumatoid synovial fluid. Activation of the complement system by either the classical or alternative pathway leads to the generation of C5a.'2 This 74 amino acid glycoprotein is rapidly converted to the physio- logically more stable C5a des Arg by the action of carboxypeptidase N. C5a is a potent anaphy- latoxin, releasing histamine from mast cells and basophils; CSa des Arg is virtually inactive in this respect (in man, but not in all species). In contrast, both forms of CSa stimulate leucocyte locomotion in vitro and emigration in vivo. The rabbit skin provides a convenient system to test the biological activity of C5a in vivo. In this system both rabbit and human C5a induce increased microvascular permeability that is independent of the carboxyl terminal arginine and histamine release.5 6 Plasma protein leakage induced by CSa injected intradermally is rapid in onset: radioisotopic studies showed leakage five to six minutes after injection.7 This response, however, is entirely dependent on circulating neutrophils: neutrophil depletion abolished leakage induced by CSa but had no effect on leakage induced by histamine and bradykinin.6 7 Thus a rapid interaction between neutrophils and microvascular endothelial cells triggered by CSa gives rise to increased endo- thelial permeability to plasma proteins. 3 These observations can explain the dependence of oedema on the presence of circulating neutro- phils in animal models of immune vasculitis.'4 In one of these models (an Arthus-type reaction in rabbits) we have shown that oedema is associated with the extravascular generation of OSa'5 and that the response is inhibited by anti- C5a antibodies.'6 Cleavage of the third and fourth components of complement leads to the formation of C3a and C4a. Although structurally similar to CSa, these proteins interact with cells at a site which is distinct from the CSa receptor and promote mast cell or basophil histamine release, but not leucocyte activation. Human C3a is about a thousand times less active than human C5a in increasing microvascular permeability when tested in rabbit skin.6 Synovial fluid contains complement,'7 18 though at lower concentrations than is found in plasma. There is evidence for both classical and alternative pathway activation in rheumatoid arthritis, though activation of the classical pathway by IgG aggregates and immune com- plexes is likely to be more important. "' Concentrations of several C3 fragments are increased in rheumatoid synovial fluids.22-24 Based on their in vitro activity, none of these C3 fragments is likely to mediate neutrophil accumulation and neutrophil dependent oedema formation. Ward and Zvaifler reported chemotactic activity for neutrophils in 38 of 54 RA synovial fluids tested.25 Anti-C5 antibodies inhibited the activity (usually by 80-90%) in 14 out of 15 chemotactic rheumatoid synovial fluids. Inhibition by anti-C3 antibodies was detected in only seven of the 15 fluids and was much weaker than the inhibition by anti-C5 antibodies. These authors attributed chemotactic activity to the presence of C567 and C5a,25 although subsequent work showed that the activity ascribed to C5a had a very different electro- phoretic mobility to the purified protein.26 In 747 on June 5, 2020 by guest. Protected by copyright. http://ard.bmj.com/ Ann Rheum Dis: first published as 10.1136/ard.49.10.747 on 1 October 1990. Downloaded from
Transcript
Annals ofthe Rheumatic Diseases 1990; 49: 747-752
Measurement of the chemotactic complementfragment C5a in rheumatoid synovial fluids byradioimmunoassay: role of C5a in the acuteinflammatory phase
P J Jose, I K Moss, R N Maini, T J Williams
AbstractEvidence suggests that complement is acti-vated in rheumatoid joints. A sensitive radio-immunoassay for the activation fragment ofC5, C5a, which is a potent chemoattractantfor neutrophils, was therefore developed. Amean CSa concentration in 22 rheumatoidjoint fluids of about 2-5xlO-9 mol/l wasfound. This concentration of C5a is sufficientto induce two of the characteristic features ofthe acute inflammatory phases of rheumatoidarthritis: neutrophil accumulation and micro-vascular plasma protein leakage. In animalmodels it has been shown that C5a is a potentinducer of inflammatory oedema by a neutro-phil dependent mechanism. A striking featureof the acute inflammatory phases of rheuma-toid arthritis is the appearance ofhigh numbersof neutrophils in the synovial fluid. It issuggested that CSa might have a role inmediating neutrophil accumulation and, as aconsequence, may be important in acute jointswelling and pain.
Section of VascularBiology, MRC ClinicalResearch Centre,Harrow, MiddlesexP J JoseT J WilliamsDepartment of ClinicalResearch, Charing Crossand WestminsterMedical School,London W6I K MossDepartment ofImmunology ofRheulmatic Diseases,C g Cross andWestmister MedicalSchool, London W6R N MainiCorrespondence to:Dr P J jose,Departnent of AppliedPharmaclogy,National Heart and LungInstitute, Dovehouse Street,London SW3 6LY.Accepted for publication21 September 1989
In addition to the chronic synovitis of rheuma-toid arthritis (RA) are the acute flare-up re-actions characterised by extravasation of neutro-phils and plasma proteins. Evidence suggeststhat the complement system is activated inrheumatoid arthritis.' One product of comple-ment activation is CSa, a potent mediator ofneutrophil infiltration and oedema.2'10 Thebinding of CSa to receptors on leucocytes oftenrenders it undetectable in the fluid phase evenwhen complement activation has been clearlyshown. " We developed a sensitive radio-immunoassay for the measurement of CSa ininflammatory exudate fluids, and in this paperwe report for the first time as far as we know,the C5a concentrations in rheumatoid synovialfluid.
Activation of the complement system byeither the classical or alternative pathway leadsto the generation of C5a.'2 This 74 amino acidglycoprotein is rapidly converted to the physio-logically more stable C5a des Arg by the actionof carboxypeptidase N. C5a is a potent anaphy-latoxin, releasing histamine from mast cells andbasophils; CSa des Arg is virtually inactive inthis respect (in man, but not in all species). Incontrast, both forms of CSa stimulate leucocytelocomotion in vitro and emigration in vivo. Therabbit skin provides a convenient system to testthe biological activity of C5a in vivo. In thissystem both rabbit and human C5a induceincreased microvascular permeability that is
independent of the carboxyl terminal arginineand histamine release.5 6 Plasma protein leakageinduced by CSa injected intradermally is rapidin onset: radioisotopic studies showed leakagefive to six minutes after injection.7 Thisresponse, however, is entirely dependent oncirculating neutrophils: neutrophil depletionabolished leakage induced by CSa but had noeffect on leakage induced by histamine andbradykinin.6 7 Thus a rapid interaction betweenneutrophils and microvascular endothelial cellstriggered by CSa gives rise to increased endo-thelial permeability to plasma proteins. 3 Theseobservations can explain the dependence ofoedema on the presence of circulating neutro-phils in animal models of immune vasculitis.'4In one of these models (an Arthus-type reactionin rabbits) we have shown that oedema isassociated with the extravascular generation ofOSa'5 and that the response is inhibited by anti-C5a antibodies.'6
Cleavage of the third and fourth componentsof complement leads to the formation of C3aand C4a. Although structurally similar to CSa,these proteins interact with cells at a site whichis distinct from the CSa receptor and promotemast cell or basophil histamine release, but notleucocyte activation. Human C3a is about athousand times less active than human C5a inincreasing microvascular permeability whentested in rabbit skin.6
Synovial fluid contains complement,'7 18though at lower concentrations than is found inplasma. There is evidence for both classical andalternative pathway activation in rheumatoidarthritis, though activation of the classicalpathway by IgG aggregates and immune com-plexes is likely to be more important. "'
Concentrations of several C3 fragments areincreased in rheumatoid synovial fluids.22-24Based on their in vitro activity, none of these C3fragments is likely to mediate neutrophilaccumulation and neutrophil dependent oedemaformation.Ward and Zvaifler reported chemotactic
activity for neutrophils in 38 of 54 RA synovialfluids tested.25 Anti-C5 antibodies inhibitedthe activity (usually by 80-90%) in 14 out of 15chemotactic rheumatoid synovial fluids.Inhibition by anti-C3 antibodies was detected inonly seven of the 15 fluids and was muchweaker than the inhibition by anti-C5 antibodies.These authors attributed chemotactic activity tothe presence of C567 and C5a,25 althoughsubsequent work showed that the activityascribed to C5a had a very different electro-phoretic mobility to the purified protein.26 In
747
on June 5, 2020 by guest. Protected by copyright.
http://ard.bmj.com
/A
nn Rheum
Dis: first published as 10.1136/ard.49.10.747 on 1 O
view of its potent biological activity there havebeen surprisingly few attempts to measure CSain synovial fluids. Wagner and Hugli studied atotal of 21 patients with gout, RA, and osteo-arthritis: C3a and C4a were detected but C5awas not." Moxley and Ruddy, who found muchhigher C3a concentrations in 41 patients withRA than in those with degenerative joint disease(osteoarthritis), reported the detection of C5a inonly one patient with RA.24 We developed amore sensitive radioimmunoassay for CSa thanthat available commercially, and in this paperwe report the measurement of biologicallyactive concentrations of CSa in rheumatoidjoint fluids; there was no measurable CSa in theplasma of these patients. A small number ofsynovial fluid samples from patients withsystemic lupus erythematosus, osteoarthritis,Reiter's disease, and mechanical joint traumawere also investigated. We also measured con-centrations of the more abundant, but muchless inflammatory C3a.
Patients and methodsThe patients in this study fulfilled AmericanRheumatism Association criteria for definite orclassical rheumatoid arthritis28 (20 women aged26-69, two men aged 40 and 57) and systemiclupus erythematosus29 (five women aged 20-36).Osteoarthritis was diagnosed radiologically(three patients). The four patients with Reiter'ssyndrome were all men whose arthritis wasassociated with recent non-specific urethritis.'Traumatic' synovial fluids were obtained frompatients with mechanical injuries (two womenaged 21 and 47, two men aged 21 and 58). Allfluids were derived by arthrocentesis of kneescarried out for relief of symptoms.To prevent complement activation after col-
lection blood or joint fluid was collected intubes containing sufficient dry sodium EDTAto give a minimum final concentration of20 mmolll. Samples were centrifuged (1500 gfor 10 minutes) as soon as possible aftercollection. Care was taken not to disturb the cellpellet when removing the supernatant, whichwas then stored frozen in aliquots while await-ing assay.
RADIOIMMUNOASSAYThe stable des Arg metabolites of C5a and C3awere purified from human plasma activated byzymosan using methods described by Fern-andez and Hugh30 followed by chromato-focusing (in the case of C5a des Arg only)'5and cation exchange high performance liquidchromatography (HPLC). Immunisation withthese proteins was used to obtain goat anti-human C5a and rabbit antihuman C3a. Thegoat antiserum reacted equally with CSa andC5a des Arg but did not cross react with C3a orC3a des Arg (<0 03%). The rabbit antiserumrecognised C3a and C3a des Arg but not CSa orCSa des Arg (<0e 1%). As has been reported byothers,3' 32 C5 cross reacted with anti-C5a (23%on a molar basis) and C3 cross reacted with theanti-C3a (42%). For this reason, it was essentialto remove the C5 and C3 precursors beforemeasuring the concentrations of their respective
split products of activation. A commerciallyavailable precipitating agent supplied withradioimmunoassay kits32 did not completelyremove these precursors from samples containingvarying protein concentrations such as thosefound in inflammatory exudate fluids. Wetherefore developed a system using a combi-nation of polyethylene glycol 6000 (PEG) andprotamine sulphate which effectively removesC5 and C3 under these conditions (fig 1).
Samples of plasma or joint fluids (0-2 mlundiluted for CSa assays, and diluted twofoldfor C3a assay) (fig 1) were mixed with an equalvolume of a mixture of 22% PEG and 1%protamine sulphate in 10 mM sodium phosphatebuffered saline (PBS), pH 7 4, containing 0-1%sodium azide and 10 mM EDTA. After incuba-tion at 4°C for one hour and centrifugation at5300 g for 10 minutes the supernatants wereremoved for assay. Under these conditions therecovery of CSa des Arg and C3a des Arg in thesupernatants was at least 80% while no precursorC5 and C3 was detectable (< 1%) (fig 1).Furthermore, recovery of IgG was far too low(1-2%) to interfere with the protein A separationstep described below.The competitive binding radioimmunoassay
procedure was modified from that describedpreviously for rabbit CSa. '5 Radiolabelled tracerligands were prepared by incubating 10 [tg C5ades Arg (or 20 [ig C3a des Arg) with 37 MBqNa'251 in the presence of a solid phase glucoseoxidase-lactoperoxidase system (Enzymobeads;Rio-Rad Laboratories). The specific activitiesobtained were 590-780 kBq/4g CSa des Arg and480-630 kBq/tg C3a des Arg. Using the proce-dure described below, binding of radiolabel inthe presence of excess antiserum was 93% forCSa des Arg and 97% for C3a des Arg. Standardsand, where necessary, sample supernatants werediluted with a mixture of 11% PEG and 05%protamine sulphate in PBS, 0 1% azide, 10 mMEDTA. 1251 labelled ligands, antisera, andprotein A bacterial adsorbent (PABA; MilesScientific) were diluted with PBS containing0-5% protamine sulphate, 0-2% gelatin, 0-1%azide, and 10 mM EDTA. Sample supernatants(100 1d) were mixed with 100 1t radiolabelledligand (0 5 ng '251-CSa des Arg or 2-0 ng 1251_C3a des Arg) and 100 [L antiserum (goat anti-C5a diluted 1/34 000 or rabbit anti-C3a diluted1/22 000). After incubation for 40 hours at20-240C an excess of PABA (25 ,tl of a 5%suspension) was added. After a further one hour1 ml of PBS (0- 1% azide) was added, the tubeswere centrifuged (5300 g for 10 minutes) andthe supernatants were removed by suction. Theantibody bound radioactivity in the pellets wascounted in a multiwell gammacounter (LKBWaliac 1260, Multigamma II) for a time suf-ficient to register at least 10 000 counts insamples to which no unlabelled higand wasadded. Concentrations were determined fromstandard curves constructed by use of acomputer assisted spline fit programme.The limits (mean (SEM)) of adequate
measurements (defined as the concentrationrequired to give 20% inhibition of binding ofthe radiolabelled ligand) were 6-0 (0-2) ngC5a/ml plasma or joint fluid and 22-1 (1 4) ng
748
on June 5, 2020 by guest. Protected by copyright.
http://ard.bmj.com
/A
nn Rheum
Dis: first published as 10.1136/ard.49.10.747 on 1 O
Measurement ofCSa in rheumatoid syovialfluids by radioimmunoassay
100-
80-
cQ"IS 60-C
cL.C
2> 40-u,0LI4,
20-
O-
C5a
C5a
Is Ir, ,1 10 50 100 1 10
Plasma concentration (%J
Figure I Effects ofprotein concentration on the precipitation ofC5 and C3 and the recoveryofCSa and C3a. Recovery of 125I tracer in the supernatant ofplasma samples treated withPEGlprotamine sulphate (a) and, for comparison, the precipitation agent supplied with theUpjohn radioimmunoassay kits (0). The left handpanel shows resultsfor CSa des Arg andCS; the right handpanel shonvs resultsfor C3a des Arg and C3. Samplesfor CSaradioimmunoassay were used undiluted and samples for C3a assay were diluted twofold beforeaddition ofthe PEGlprotamine sulphate precipitating agent to ensure that recovery ofthe splitfragments was at least 80%. Each point is the mean ofthree determinations.
G3a/ml (n= 15 assays). These assays were abouttwice as sensitive as the commercial kits. Theinterassay coefficients of variation in the C5aassays were 21% for a rheumatoid joint fluidand 90/o for a partially activated plasma sample.In the C3a assays these values were 7% for therheumatoid joint fluid and 14% for the activatedplasma sample. Intra-assay coefficients of varia-tion for rheumatoid joint fluid were 4% in bothradioimmunoassays.
IN VITRO STUDIESPeripheral venous blood was withdrawn intoheparin (10 U/ml) for preparation of plasma.Plasma was activated at 37°C for 30 minutesusing a dose range of activator suspensions insaline. Sodium EDTA (pH 7-4) and magnesiumEGTA (pH 7-4) were added to a final concen-tration of 10 mmol/l, as apppropriate. Incuba-tions were terminated by the addition of sodiumEDTA, cooling in ice-cold water, and centri-fugation at 2°C. The activators used were heataggregated gammaglobulin (HAGG: humanIgG (Sigma) heated to 63°C for one hour) andyeast cell walls (zymosan A, Sigma).
HPLC ANALYSISFor HPLC analysis of immunoreactivity inrheumatoid joint fluid and activated plasmasamples, PEG/protamine sulphate supernatantswere mixed with an equal volume of 25 mMsodium phosphate buffer, pH 5-8, and appliedto a TSK 535 CM column (7-5x150 mm).Proteins were eluted using a 100-900 mMsodium chloride gradient in the pH 5 8 buffer ata flow of 1 ml/minute. Fractions (1 ml) werecollected into 0-2 ml of a solution designed toadjust the protamine sulphate concentration to
0-5% and the pH to about 7-4 for subsequentC3a radioimmunoassay.
ResultsGENERATION OF C5a AND C3a DURINGACTIVATION OF CLASSICAL AND ALTERNATIVEPATHWAYS IN VITRO
C3a To investigate complement activation in vitro,plasma was incubated with varying doses ofHAGG and zymosan. Generation of immuno-reactive CSa and C3a was related to the dose ofactivator used (fig 2). Magnesium EGTA,which is known to chelate calcium ions andinhibit classical pathway activation, considerablyreduced activation by HAGG but had noinhibitory effect on activation by zymosan.Sodium EDTA, which is known to chelatemagnesium as well as calcium ions and therefore
___________ C3 to inhibit both pathways of complement,50 100 completely inhibited activation by HAGG andby zymosan. These results show that the
- I_ _rr__.-- TT A tN ____ __-_.C__r .prmcipal eIIect O0 HAUA was activation O0 tmeclassical pathway while zymosan was effective inactivating the alternative pathway of comple-ment, and that activation of either pathwayleads to the generation of CSa and C3a immuno-reactivity.
1.5 -
-RE
0.
i;
LA
Et0
E0.
N
uem-LnN
-0
1.0-
0.5-
30-
20-
10-
O-
p Zym + MgEGTA
I/ 'HAGG
- -_HAGG + MgEGTA
1-I lv + EDTA
-0 HAGG + MgEGTA
-T- r-m-T + EDTA
0.06-* Zym00Zym +MgEGTA
0.04- .~-
0.02- -
_- - - - -O HAGG + MgEGTA0.02- -"_-*HAGG
0--' ,0 0.02 0.2 2.0
Activator concentration (mg/ml plasma)Figure 2 Complement activation in vitro. Both heataggregated gammaglobulin (HAGG) and zymosan produceda dose related generation ofimmunoreactive CSa (upperpanel) and C3a (middle panel). Magnesium EGTA (opensymbols) inhibited activation byHAGG but not by zymosan.The lowerpanel shows thatHAGG produces less CSarelative to C3a than does zymosan. This may reflectdifferences in the physical nature ofthe activator ordifferences in the abtlity ofclassical and alterativepathwayenzymes to cleave C3 and C5. Each point is the mean ofduplicate incubations.
749
0i
40-1
on June 5, 2020 by guest. Protected by copyright.
http://ard.bmj.com
/A
nn Rheum
Dis: first published as 10.1136/ard.49.10.747 on 1 O
The results shown in fig 2 also show that theratio of C5a:C3a generated depends on theactivator used as zymosan produced a higherratio than HAGG. Thus measurements of C3acannot be used to predict accurately the concen-trations of CSa when the nature of the stimulusis unknown.
C5a AND C3a CONCENTRATIONS IN JOINT FLUIDSJoint fluids from patients with RA, systemiclupus erythematosus (SLE), osteoarthritis,Reiter's disease, and mechanical trauma wereanalysed for immunoreactive C5a and C3a (fig3). For comparison, plasma samples from 11 ofthe 22 patients with RA were also analysed. CSaconcentrations were well below the limit ofadequate measurement (6 ng/ml) in the rheuma-toid plasma samples. Joint fluids from patientswith RA contained 4-86 ng C5a/ml (mean=21).Fluids taken from mechanically traumatisedjoints contained low C5a concentrations.Although the sample numbers were low, theconcentrations of C5a in synovial fluid samplesfrom patients with SLE, osteoarthritis, andReiter's disease were increased but not to thedegree seen in RA. Rheumatoid plasma samplescontained C3a (0-26 (0 07) jig/ml). G3a concen-trations in joint fluid were greatly increased inRA (mean 4-06 jg/ml); the other conditionsshowed a range of C3a concentrations abovethat in plasma.
*86
4. 69
0:.00
0
@000
00
0
The concentrations of CSa and C3a in the 22individual rheumatoid joint fluids are correlatedin fig 4. Although there is a positive correlation,measurement of C3a was a very poor predictorof the concentration of C5a in individual jointfluids.The two patients with RA with the highest
concentrations of synovial C5a had normalconcentrations of circulating C3, C4, and IgG.One of these patients, whose joint fluid con-tained 86 ng CSa/ml and 7A46 ,ug G3a/ml, wasstrongly seropositive for rheumatoid factor andwas on a repeat course of gold treatment at thetime of sample collection. The other, whosejoint fluid contained 69 ng CSa/ml and 7'76 tigG3a/ml, was seronegative and was being treatedwith indomethacin and prednisolone at the timeof sample collection. Both samples were takenduring acute flare-up reactions.
HPLC OF IMMUNOREACTIVITY IN RHEUMATOIDJOINT FLUID COMPARED WITH ACTIVATED PLASMAA sample of rheumatoid joint fluid (1P8 mlcontaining 30 ng C5a and 14 5 pg C3a) wasmixed with the PEG/protamine sulphate pre-cipitation agent. The supernatant was thenmixed with column buffer and applied to acarboxymethyl cation exchange HPLC column.The column was eluted with a salt gradient andfractions collected for radioimmunoassay. Tocompare the elution profiles this was followedby injection of the C5a des Arg and C3a des Argstandards.The profile of immunoreactivity in the rheu-
matoid joint fluid is shown in fig 5. Immuno-reactive CSa eluted as a discrete peak at fraction18, corresponding to the position of standardC5a des Arg. Some apparent immunoreactivity
0
102
00
00 0 0: * *0 00
-4.
Sc
U
-ole
c
u
10 -
0
0
0.
SLE
000
OA
0
0
*0
Reiter'sdisease
0
00
Trauma
FWi 3 Concenaio ofimmnwreactive CSa (upperpanel) and C3a (lower panel) injointfluids andplasma.Jointfluidsfirom 22 patents with rheuatoid arthriis (RA)contained a wide range ofCSa concentrations with a mean value of21 ng/ml. In contrast,plasma samples (shown as smallr ymbos) taknfrom II ofthese patients had virtuallydecle concentrations ofCSa (I 3 (0-3) nglml). The mean values ofC3a in rheumatoid
jointfluids was4-06 0g/mi, which is much higher than concentrationsfound in rheumatoidplasma samples (0-26 (007) pglml).
1-
Rheumatoid joint fluids0
0
0
0 0
I
0
00 00
s
0
102 103 104
C3a concentratlon (ng/mil
Figue 4 Correlaon ofCSa and C3a concentrations inindiviual rhewnatoidjointfluids. Linear regression analysisofl0og tasfed data showed a positive correlation(r=0-70) between CSa and C3a concentations. The largedeviation ofdatapot fnmn the predicted line shows thatm oreactive C3a ement is a poorpredictor ofthe
concentration ofthe biogially more relevant CSa inindividual sampes.
40-
Ec
t 20-U-,0
0-
12-
0
.0 08-
4-
Ecm
0
0-
000.00 00*.0RARA
plasma
750
on June 5, 2020 by guest. Protected by copyright.
http://ard.bmj.com
/A
nn Rheum
Dis: first published as 10.1136/ard.49.10.747 on 1 O
Measurement ofCSa in rheumatoid synovialfluids by radioimmunoassay
20
10-
E
u
E
ELn
0-
CSa desArg C3a desArg
I IRheumatold joint fluid
-- I-
Activated plasma
010 10 20 30
Retention time (minutes)
Figure S Radioimmunoassays for CSa and C3a aftercation exchange high performance liquid chromatography.Immunoreactive CSa (0) in rheumatoidjointfluid eluted asa single peak, corresponding to the retention time ofthe CSades Arg standard (upper panel). C3a immunoreactivity (0)eluted as two peaks, the major one corresponding to theretention time ofC3a des Arg. For comparison, plasma wasactivated in vitro by incubation with heat aggregatedgammaglobulin and chromatographed (lower panel).Immunoreactive C3a eluted as a single peak corresponding tothe C3a des Arg standard. In contrast with thejointfluid, theplasma sample did not contain the peak ofunidentified C3aimmunoreactivity. Immunoreactive CSa showed a similarelution profile to that seen with thejointfluid.
was seen in later fractions; this was oelow thelimit of adequate measurement and was shownto be a reflection of the high salt concentrations(up to 900 mmol/l NaCl) in these fractions.Fractions 17-19 contained 83% of the immuno-reactivity applied to the column, which is a highrecovery for such a low C5a concentration. Themajor peak of immunoreactive C3a eluted infraction 25, corresponding to the position ofstandard C3a des Arg. There was, however, a
smaller peak of immunoreactivity in fraction 23.This profile was confirmed in a subsequentassay and both peaks gave complete displace-ment of radiolabelled ligand at low dilutions andshowed parallels with the standard when testedin serial dilutions. The early eluting peak of C3aimmunoreactivity may represent a breakdownproduct of C3a or C3a des Arg or a slightlydifferent cleavage product of C3, perhapsgenerated by proteases released from tissue cellsor the accumulating leucocytes.When plasma samples activated by HAGG
(fig 5) and zymosan (data not shown) wereanalysed, immunoreactive C3a eluted with thesame retention time as the standard C3a des Arg.There was no evidence, however, for thesmaller early eluting peak of immunoreactiveC3a previously seen in the rheumatoid jointfluid. As was found with the joint fluid,immunoreactive C5a in the activated plasmasamples eluted as a single peak corresponding tothe mobility of CSa des Arg (fig 5).
DiscussionComplement activation has been well documentedin the inflamed joints of rheumatoid arthritis.' Ofthe products of complement activation, CSa and itsmore stable des Arg metabolite have potent pro-inflammatory activities which might account forsome of the features of the acute flare-up phases ofrheumatoid arthritis, characterised by plasmaprotein and neutrophil extravasation.Moxley and Ruddy, who used commercial
radioimmunoassay kits and found high concen-trations of C3a in rheumatoid joint fluids, con-cluded that the concentrations of CSa were eitherbelow or barely above the limit of detection of theCSa kit.24 We developed a CSa radioimmunoassayof greater sensitivity and with more reliableelimination of the crossreacting C5 precursor,especially when using samples of variable proteinconcentrations. The results show that C5a waspresent in rheumatoid joint fluids, but not in theplasma of these patients (fig 3). The immuno-reactive C5a in joint fluids is assumed to be C5a desArg owing to the action of endogenous carboxy-peptidase N. Moreover, the immunoreactivitychromatographed as a single peak, correspondingto the position of CSa des Arg on cation exchangeHPLC (fig 5). The system used will distinguishbetween CSa and CSa des Arg (data not shown).The mean value found in rheumatoid joint fluidswas 21 ng/ml (about 2-5x10-9 mol/l) and thisconcentration of CSa/CSa des Arg is sufficient topromote neutrophil chemotaxis in vitro,33 oedemaformation, and neutrophil accumulation in rabbitskin 7 ' and wheal and flare reactions in humanskin.34The detection of high concentrations of
immunoreactive C3a agrees with the results ofMoxley and Ruddy.24 C3a is not a mediator ofneutrophil accumulation and associated oedemaformation,6 although it can cause mast celldegranulation and a wheal and flare reaction inhuman skin, provided that the carboxyl terminalarginine is present. Rheumatoid synovial tissuescontain mast cells which can be stimulated torelease histamine.35 It is doubtful, however, thathistamine release by anaphylatoxins is importantin rheumatoid arthritis because treatment withantihistamines is ineffective and C3a is even moresusceptible to conversion to its des Arg metabolitethan C5a.36 In this context it is interesting to notethat oedema formation and neutrophil accumulationinduced by injection of C5a and CSa des Arg inrabbit and human skin are largely resistant toantihistamines. 5 6 34 37 38The correlation between C5a and C3a concentra-
tions in individual fluids was poor (fig 4). Thismay be explained by in vitro studies. There isgeneral agreement that various types of leucocytesbind C5a but not C3a."1 39 40 When complementwas activated in vitro in the absence of cells theratio of immunoreactive C5a:C3a was also found tovary with the stimulus (fig 2). Heat aggregatedgammaglobulin, used as a model for rheumatoidfactor complexes with IgG, generated a lowerC5a:C3a ratio than zymosan. This may reflectdifferences in the physical nature of the activatoras well as differences in the relative efficiencies ofthe C3 and C5 convertases generated duringactivation of the classical and alternative pathways.Furthermore, Ng et al found that immune com-plexes containing IgM x rheumatoid factor activateC3 in the fluid phase yet fix C3 poorly4"; thiswould be expected to generate lower than normal
751
on June 5, 2020 by guest. Protected by copyright.
http://ard.bmj.com
/A
nn Rheum
Dis: first published as 10.1136/ard.49.10.747 on 1 O
C5 convertase activity and thus to yield lowC5a:C3a ratios. Taken together, the variableC5a:C3a ratios generated during activation and thesubsequent removal of C5a by leucocytes mightaccount for the variable C5a:C3a ratios observed inthe rheumatoid joint fluids.
Chemotactic activity related to C5 was found inrheumatoid joint fluid in another study: thisactivity was attributed to the presence of the C567complex and C5a.26 Subsequent work, however,cast doubt on the identification of C5a because thechemotactic activity ascribed to it migrated anodallyon Pevikon block electrophoresis.26 The generationof C5a can be inferred from the detection ofincreased concentrations of the fluid phaseterminal complement complex in rheumatoid jointfluids,42 as cleavage of C5 with the liberation ofC5a is an essential step in assembly of thiscomplex. Previous attempts to measure C5a insynovial fluids have been unsuccessful.24 27 Usinga more sensitive radioimmunoassay, we have nowshown that rheumatoid joint fluids contain bio-logically active concentrations of GSa.
In animal models we have shown that C5a (andits des Arg metabolite) is a potent inducer ofoedema by a neutrophil dependent mechanism.6 7A striking feature of the acute inflammatoryphases of rheumatoid arthritis is the appearance ofhigh numbers of neutrophils in synovial fluid. Wesuggest that C5a might have an important role inmediating neutrophil accumulation and, as aconsequence, may be important in acute jointswelling.This work was supported by a grant from the Arthritis andRheumatism Council. We thank Dr R A Harrison, Mechanismsin Tumour Immunity Unit, MRC Centre, Cambridge for thegenerous gifts of C3 and C5.
I Goldstein I M. Clinical applications of complement measure-ments in rheumatic diseases. Am J Med Sci 1975; 269:172-82.
2 Snyderman R, Phillips J K, Mergenhagen S E. Biologicalactivity of complement in vivo. Role of C5 in theaccumulation of polymorphonuclear leukocytes in inflam-mation. i Exp Med 1971; 134: 1131-43.
3 Damerau B, Vogt W. Effect of hog anaphylatoxin (CSa)on vascular permeability and leukocyte emigration invivo. Naunyn Schmiedebergs Arch Pharmacol 1976; 255:237-41.
4 Issekutz A C, Movat K W, Movat H Z. Enhanced vascularpermeability and hemorrhage-inducing activity of rabbitCSa des arg: probable role of polymorphonuclear leukocytelysosomes. Cl:n Exp Immunol 1980; 41: 512-20.
5 Williams T J, Jose P J. Mediation of increased vascularpermeability after complement activation: histamine-independent action of rabbit COa. _7 Exp Med 1981; 153:136-53.
6 Jose P J, Forrest M J, Williams T J. Human CSa des Argincreases vascular permeability. J Immunol 1981; 127:2376-80.
7 Wedmore C V, Williams T J. Control of vascular per-meability by polymorphonuclear leukocytes in inflam-mation. Nature 1981; 289: 646-50.
8 Jose P J. Complement-derived peptide mediators of inflam-mation. Br Med Bull 1987; 43: 33649.
9 Yancey K B. Biological properties of human CSa: selected invitro and in vivo studies. Clin Exp Immunol 1988; 71:207-10
10 Rampart M, Williams T J. Evidence that neutrophil accumu-lation induced by interleukin-1 requires both local proteinbiosynthesis and neutrophil CD18 antigen expression invivo. Br]7 Pharmacol 1988; 94: 1143-8.
11 Chenoweth D E, Cooper S W, Hugli T E, et al. Complementactivation during cardiopulmonary bypass. Evidence forgeneration of C3a and CSa anaphylatoxins. N Engl J7 Med1981; 304: 493-503.
12 Hugli T E, Muller-Eberhard H J. Anaphylatoxins: C3a andCSa. Adv Immunol 1978; 26: 1-53.
13 Williams T J. Increased vascular permeability followinginteraction between polymorphonuclear neutrophils andvascular endothelial cells. In: Bagge U, Born G V R,Gaehtgens P, eds. Morphology and rheology as related tofunction. The Hague, Holland: Martinus Nijhoff, 1982:
14 Humphrey J H. The mechanism of Arthus reactions. I. Therole of polymorphonuclear leucocytes and other factors in
15 Jose P J, Forrest M J, Williams T J. Detection of thecomplement fragment C5a in inflammatory exudates fromthe rabbit peritoneal cavity using radioimmunoassay. 7 ExpMed 1983; 158: 2177-82.
16 Hellewell P G, Jose P J, Williams T J. Effect of anti-CSaantibodies on oedema formation and PMN leukocvteaccumulation in allergic inflammation in the rabbit. Br ]Pharmacol 1988; 95: 531.
17 Hedberg H. Studies on the depressed hemolytic complementactivity of synovial fluid in adult rheumatoid arthritis. AciaRheumatologica Scandinavica 1963; 9: 165-93.
18 Pekin T J, Zvaifler N J. Hemolytic complement in synovialfluid.] Clin Invest 1964; 43: 1372-82.
19 Winchester R J, Agnello V, Kunkel H G. Gamma globulincomplexes in synovial fluids of patients with rheumatoidarthritis. Clin Exp Immunol 1970; 6: 689-706.
20 Ruddy S, Austen K F. The complement system in rheu-matoid synovitis. I. An analysis of complement componentactivities in rheumatoid synovial fluids. Arthritis Rheum1970; 13: 713-23.
21 El-Ghobarey A F, Whaley K. Alternative pathway comple-ment activation in rheumatoid arthritis. ] Rheumatol 1980;7: 453-60.
22 Nydegger U E, Zubler R H, Gabav R, et al. Circulatingcomplement breakdown products in patients with rheuma-toid arthritis: correlation between plasma C3d, circulatingimmune complexes, and clinical activity. 7 C/in Invest1977; 59: 862-8.
23 Bourke B E, Moss I K, Maini R N. Measurement of thecomplement C3 breakdown product C3d bv rocketimmuno-electrophoresis. 7 Immunol Methods 1982; 48:97-108.
24 Moxley G, Ruddy S. Elevated C3 anaphvlatoxin lcvels in]synovial fluids from patients with rheumatoid arthritis.Arthnrtis Rheum 1985; 28: 1089-95.
25 Ward P A, Zvaifler N J. Complement-derived leukotacticfactors in inflammatory synovial fluids of humans. 7 ClinInvest 1971; 50: 606-16.
26 Ward P A. Complement-dependent phlogistic f'actors inrheumatoid synovial fluids. Ann NY Acad Sc:i 1975; 256:169-76.
27 Wagner J L, Hugli T E. Radioimmunoassav for anaphvla-toxins: a sensitive method for determining complementactivation products in biological fluids. Anal B:ochem 1984;136: 75-88.
28 Ropes M W, Bennett G A, Cobb S, et al. Diagnostic criteriafor rheumatoid arthritis: 1958 revision. Ann Rheum )is1959; 18: 49-53.
29 Tan E M, Cohen A S, Fries J F, et al. 'I'he 1982 revisedcriteria for the classification of svstemic lupus erythema-tosus. Arthritis Rheum 1982; 25: 1271-7.
30 Fernandez H N, Hugli T E. Primary structural analysis of thepolypeptide portion of human CSa anarhylatoxin: Polv-peptide sequence determination and assignment of theoligosaccharide attachment site in C5a..7 Biol Chemn 1978;253: 6955-64.
31 Hugli T E, Chenoweth D E. Biologically active peptides ofcomplement: techniques and significance of C3a and C5ameasurement. In: Nakamura R M, Dito W R. Tucker E S,eds. Immunoassays: clinical laboratory techniques for the1980s. New York: Alan R Liss, 1981: 443-60.
32 Satoh P S, Yonker T C, Kane D P, et al. Mcasurcmcntof anaphylatoxins: an indcx for activation of complcmcntcascades. BioTechniques 1983; 1: 90-5.
33 Becker E L. Chemotactic factors in inflammation. TrendsPharmacol Sci 1983; 4: 223-5.
34 Swerlick R A, Yancey K B, Lawley T J. A direct in vivocomparison of the inflammatory properties of human CSaand CSa des Arg in human skin. ] Immunol 1988: 140:2376-81.
35 Gruber B, Poznansky M, Boss E, et al. Characterization andfunctional studies of rheumatoid synovial mast cells:activation by secretagogues, anti-IgE, and a histamine-releasing lymphokine. Arthritis Rheum 1986; 29: 944-55.
36 Hugli T E, Gerard C, Kawahara M, et al. Isolation of threeseparate anaphylatoxins from complement-activated humanserum. Mol Cell Biochem 1981; 41: 59-66.
37 Yancey K B, Hammer C H, Harvath L, et al. Studies ofhuman C5a as a mediator of inflammation in normal humanskin.]I Clin Invest 1985; 75: 486-95.
38 Williamson L M, Sheppard K, Davies J M, et al. Neutrophilsare involved in the increased vascular permeability producedby activated complement in man. Br] Haematol 1986; 64:375-84.
39 Chenoweth D E, Hugli T E. Demonstration of specific C5areceptor on intact human polymorphonuclear leukocytes.Proc Natil Acad Sci USA 1978; 75: 3943-7.
40 Solomkin J S, Cotta L A, Satoh P S, et al. Complementactivation and clearance in acute illness and injury:Evidence for CSa as a cell-directed mediator of the adultrespiratory distress syndrome in man. Surgery 1985; 97:668-78.
41 Ng Y C, Peters D K, Walport M J. Monoclonal rheumatoidfactor-IgG immune complexes: poor fixation of opsonicC4 and C3 despite efficient complement activation. ArthritisRheum 1988; 31: 99-107.
42 Mollnes T E, Lea T, Mellbye 0 J, et al. Complementactivation of rheumatoid arthritis evaluated by C3dg andthe terminal complement complex. Arthnrtis Rheum 1986;29: 715-21.
752
on June 5, 2020 by guest. Protected by copyright.
http://ard.bmj.com
/A
nn Rheum
Dis: first published as 10.1136/ard.49.10.747 on 1 O
![OutcomeandComplicationsofCombinedPhacoemulsification ...combinedphacoemulsification[17,18].Combinedsurgery helpsminimizingtheeconomiccost[11].Asalast,butnot least advantage, combined](https://static.documents.pub/doc/80x56/610121f62de6813fce181dac/outcomeandcomplicationsofcombinedphacoemulsification-combinedphacoemulsiication1718combinedsurgery.jpg)
