Abstracts Francesc Casals', and Miguel Ingelmo · 698 ANTIPHOSPHOLIPIDSYNDROMEWORKSHOP Laboratory Aspects pholipid antibodies J. Arnout, E. Huybrechts andJ. Vermylen Centerfor antibodies,

ANTIPHOSPHOLIPID SYNDROME WORKSHOP 695

Abstracts

Antiphospholipid antibodies: clinical and biolog-ical correlation in systemic lupus erythematosus

Ricard Cervera, Josep Font, Alfons Lopez-Soto,Francesc Casals', and Miguel Ingelmo

Departments ofInternal Medicine and 'Hemostasis andHemotherapy, Hospital Clinic i Provincial, Villarroel,170, 08036, Barcelona, Spain

Several recent studies have found that patients with antiphos-pholipid antibodies (aPL) are prone to repeated episodes ofthrombosis, recurrent fetal loss and thrombocytopenia. Inaddition, there have been reports of the possible associationof aPL with neurological events and haemolytic anaemia.However, other authors consider that no associations existand aPL are only an epiphenomenon. In fact, prospectivestudies of a large number of unselected patients are very fewand no definitive conclusions have been drawn.The main questions that most clinicians have in mind are:

first, what are the clinical manifestations really related to theaPL?, and second, what is the assay to detect aPL (lupusanticoagulant-LA-or anticardiolipin antibodies-aCL) thatcorrelates best with these clinical manifestations?

There are many case reports of patients with aPL withthrombosis affecting anywhere in the vascular tree. Never-theless, studies with large numbers of patients are notextensive. In fact, most European and American series havefound statistical correlation between aPL (LA and/or aCL)and thrombosis. However, some authors have found no suchassociation. Probably, one of the most interesting studies wasmade by Harris et al.' They found a strong correlationbetween the titre of aCL and the risk for thrombosis. Wefound similar results in a prospective 2 year study of 100patients with SLE.2 Ten patients presented venous or arterialthrombosis and a significant correlation was found with IgGaCL (P = 0.001). In addition, the higher the titre ofIgG aCL,the higher was the predictive value (PV) for thrombosis(binding index> 7, PV = 67%). Similar correlation wasfound with thrombocytopenia (P = 0.03). This confirms thefindings' that higher titres ofaPL might be of greater clinicalrelevance than lower titres.Many case reports and several surveys have been published

linking aPL with other clinical manifestations, includingrecurrent fetal loss, neurological involvement (cerebrovas-cular accidents, progressive dementia, chorea, lupoid scler-osis, epilepsy and migraine), heart valvular lesions, livedoreticularis, haemolytic anaemia, labile hypertension, tox-aemia and a post-partum syndrome. In fact, some of thesemanifestations are due to thrombotic events. It is of note thatDeleze et al.3 and our group2 have found a correlationbetween IgM aCL and haemolytic anaemia (P = 0.01). Wehave seen a similar correlation between IgM aCL andneutropenia (P = 0.005). In addition, we found that thehigher the titre of IgM aCL, the higher was the predictivevalue for haemolytic anaemia and neutropenia (bindingindex> I1, PV = 60%).

Finally, there have been several studies as well as oursshowing a strong relationship between the LA and the aCL(P = 0.01).

Supported by FISS87/1089 and CAICYT PA 86-0016

From our studies we conclude that thrombosis, fetal lossand thrombocytopenia are probably related to aPL, as well -possibly - as some neurological manifestations andhaemolytic anaemia. In addition, there appears to be a goodcorrelation between LA and aCL. However, occasionaldiscrepancies occur and patients may have LA without aCLor vice-versa. For this reason, we must conclude that bothassays (LA and aCL) must be performed in each patient.

References

1. Harris, E.N., Chan, J.K.H., Asherson, R.A. et al. Thrombosis,recurrent fetal loss and thrombocytopenia: predictive values of theanti-cardiolipin antibody test. Arch Intern Med 1986, 146:2153-2156.

2. Cervera, R., Font, J., Casals, F.J. et al. Anticardiolipin antibodiesin systemic lupus erythematosus: prospective analysis of a series of100 patients (abstract). Acta Clin BeIg 1988, 43 (suppl 12): 194.

3. Deleze, M., Oria, C.V. & Alarc6n-Segovia, D. Occurrence of bothhemolytic anemia and thrombocytopenic purpura (Evans' synd-rome) in systemic lupus erythematosus. Relationship to antiphos-pholipid antibodies. J Rheumatol 1988, 15: 611-615.

Cardiac abnormalities in systemic lupus erythe-matosus: association with antiphospholipid anti-bodies

M.A. Khamashta, A. Gil, J.M. Oliver, F.J.Dominguez, R. Fernandez de Soria, J.J. Vaz-quez.

Hospital 'La Paz', Madrid, Spain

Several studies have now shown that a significant percentageof patients with systemic lupus erythematosus (SLE) withhigh titres ofantiphospholipid antibodies (APA) are prone torepeated episodes of arterial and venous thrombosis, recur-rent fetal loss and a number of other features.' One partic-ularly interesting association with these antibodies, morerecently described, has been the development in somepatients, of valvular heart disease, including Libman-Sacksendocarditis.2 The finding of similar lesions in patients notsuffering from definite SLE in a general population andaccompanied by APA may add weight to these associations.

Although pathological features of Libman-Sacks endocar-ditis may be observed at autopsy in up to 40% of patientswith SLE, it is rarely associated with clinically severe valvulardysfunction and it is usually not recognized during life.The introduction of two-dimensional echocardiography

and more recently, the Doppler technique proved useful toolsin detecting subclinical endocardial lesions and valvular heartdysfunction.

Using two-dimensional and Doppler echocardiography westudied 38 consecutive patients with SLE to evaluate theincidence and severity of cardiac involvement in this condi-tion and correlated these findings with the presence of APA.There were 33 females and 5 males with a mean age of 35

years (range 14- 77). Patients with clinical evidence of severehypertension, chronic renal failure, rheumatic valve disease,infective endocarditis or ischaemic heart disease were ex-cluded.

copyright. on M

ay 16, 2021 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.65.767.695 on 1 S

eptember 1989. D

ownloaded from

http://pmj.bmj.com/

696 ANTIPHOSPHOLIPID SYNDROME WORKSHOP

We found a strong association between APA, and mitralvalve regurgitation and focal areas of mitral valve leafletthickening compatible with Libman-Sacks endocarditis(Table).

Antiphospholipidantibodies

Positive Negative15 (39%) 23 (61%) P

Pericardial effusion 2 (13%) 3 (13%) NSMitral valve prolapse 3 (20%) 5 (22%) NSMitral valve thickening 5 (33%) 0 (-) <0.01Mitral regurgitation 10 (67%) 7 (30%) <0.05

NS = not significant

The mechanism whereby the presence of APA may causethrombosis and its relationship to the pathogenesis ofendocardial disease remains speculative. These antibodies,possibly by their effect on the phospholipids of platelets, onendothelial cell membranes, or by other mechanisms as yetunknown, appear to be directly involved in the accelerationof thrombosis in some patients.

References

1. Harris, E.N., Chan, J.K.H., Asherson, R.A., Aber, V.R., Gharavi,A.E. & Hughes, G.R.V. Thrombosis, recurrent fetal loss andthrombocytopenia: predictive value of the anticardiolipinantibody test. Arch Intern Med 1986, 146: 2153-2159.

2. Ford, P.M., Ford, S.E. & Lillicrap, D.P. Association of lupusanticoagulant with severe valvular heart disease in systemic lupuserythematosus. J Rheumatol 1988, 15: 597-600.

3. Asherson, R.A. & Lubbe, W.F. Cerebral and valve lesions in SLE:association with antiphospholipid antibodies. J Rheumatol 1988,15: 539-543.

Antiphospholipid antibodies and fetal loss

A. Tincani, M.T. Bertoli, F. Allegri, R. Cat-taneo, D. Faden, M. Tarantini, A. Maffeis, G.Yalesini, P.L. Meroni, G. Ferraro, P. Bamberga,B. Acaia, 0. Ricciardiello and G. Balestrieri.

Servizio e Cattedra di Immunologia Clinica, SpedaliCivili-Universita' di Brescia; Clinica Ostetrica Univer-sita' di Brescia; I Clinica Medica Universitai di Roma; IClinica Medica Universita di Milano; I Clinica Ostet-rica Universita' di Milano; Italy.

The association between antiphospholipid antibodies (APA)and miscarriages or fetal death has been emphasized inseveral reports not only in patients affected by systemic lupusbut also in women without any diagnosed disease.The aims of our study are: to evaluate the prevalence of

APA in 187 women attending 3 different obstetric clinicsbetween 1982 and 1988 because of 2 or more primaryabortions (< 20 weeks; 134 patients; group A) or because of1 or more intrauterine fetal deaths (> 20 weeks; 53 patients;

group B) compared with 88 women without obstetric prob-lems; to investigate the presence of other autoantibodies inthe same population and mainly in APA positive patients inorder to identify possible subclinical autoimmune disease; toverify if a pharmacological treatment able to decrease APAlevels can influence the outcome of a new pregnancy in thesepatients; to show the long term 'after pregnancy' follow up insome of these patients attending our clinic.

In summary the results show that patients with adversepregnancy outcome (Groups A and B) display higher fre-quency ofAPA detected as lupus like anticoagulant (LLAC)and/or anticardiolipin antibodies (ACA) compared withcontrols (20.3% vs 2.3%, P<0.001); moreover in group B(late pregnancy loss) the prevalence of women with raisedAPA levels was significantly higher than in group A (35.8%vs 14.1%, P< 0.001). Autoantibodies studied (anti-nuclear,anti-mitochondrial, anti-smooth muscle and anti-ds-DNA)were surprisingly high in the population under study andmainly in group B patients (i.e. anti-nuclear antibodiesprevalence was 35.8% vs 1.1% ofcontrols, P<0.01 and 19.4% of group A, P< 0.02). Interestingly most of the patientswith antiphospholipid antibodies were also found positivefor one or more of the other autoantibodies (only 9 out of the38 APA positive patients were negative at all the otherassays). Seven patients could start a new pregnancy; theywere treated with low dose steroid (100 mg/week) and aspirin(100 mg/day) with the addition of azathioprine (100 mg/day)in 5 cases: after this treatment we found a significant decreasein ACA titre and in 6 out of 7 cases a good pregnancyoutcome was reached without any adverse side effect for themothers or the babies. The careful clinical examination ofthese 7 patients when first seen in immunopathology out-patient clinic and during the long term after pregnancyfollow-up, has shown that 2 ofthem were affected by systemiclupus erythematosus while in the other 5 cases, although nodefinite diagnosis was possible, minor signs of autoimmunediseases were present. Interestingly one of these patientsdeveloped positive anti-Sm antibodies 2 years after preg-nancy. Finally, the observation that a number of womendisplay constantly low complement levels despite the other-wise successful treatment, support the already shown geneticcomplement deficiency in this group of patients.

Anticardiolipin antibodies in pregnancy

M.G. Elder

Department ofObstetrics and Gynaecology, Hammers-mith Hospital, London W12 OHS, UK.

Anticardiolipin antibodies may or may not be present inpatients with systemic lupus erythematosus (SLE). Equallythere are patients with circulating anticardiolipin antibodies(ACLA) but without the signs of SLE. It is difficult toseparate the two populations.

It is known that SLE in pregnancy is associated with a highincidence of fetal loss due to recurrent abortion and perinataldeaths often from severe fetal growth retardation. Anticar-diolipin antibodies in pregnancy are also associated with ahigh incidence of fetal loss for the same reasons. The IgGisotype of antibody has a worse prognosis than the IgM and

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/

the higher the IgG titre the later the miscarriage.' Low levelsof ACLA are less likely to be the primary cause and mostmiscarriages will occur in the first trimester. During thesecond trimester the miscarriages are likely to be entirely dueto ACLA.The spiral arteries through which maternal blood enters

the intervillous space at low pressure and low flow rates areimportant as blockage to these end arteries will result insegmental infarction of the placenta. Normally trophoblastinvades these vessels increasing their compliance. Placentalprostaglandin production may be important in maintainingthis circulation by inhibiting platelet aggregation and causingvasodilatation.

Anticardiolipin antibodies may act directly on plateletsinitiating aggregation which in turn releases thromboxane B2,a compound that recruits further platelets and stimulatestheir aggregation. It has been shown that the IgG fraction ofthe anticardiolipin antibody can bind to vascular endothelialcells in culture. The proportion of specific binding wasbetween 12 to 48% of total binding, the rest being non-specific. The specific binding of IgG ACLA had no significanteffect on arachidonic acid metabolism by vascularendothelial cells (Patel, L., Sullivan, M.H.F. & Elder, M.G.,unpublished data).

Interactions between other cells may be important, forexample endothelial cells and lymphocytes can metabolizeplatelet endoperoxides to prostacyclin. Failure of either ofthese two mechanisms may lead to a reduction in circulatingprostacyclin which is a potent inhibitor of platelet aggrega-tion. This could take place through an action ofthe ACLA onplatelet membrane phospholipids. Finally, ACLA can affectthe coagulation system by the activation of prothrombin tothrombin through platelet phospholipids.Treatment of patients with SLE/ACLA has been immuno-

suppression with increasing doses of prednisolone when thepatient was pregnant. This form of therapy has beendisappointing and side effects significant. Recently anti-platelet therapy as aspirin 75 mg daily has been tried withgood results. This dose blocks platelet thromboxane syn-thesis and aggregation permanently because platelets cannotresynthesize the necessary enzymes. Larger doses of aspirinwill partially inhibit vascular prostacyclin production but adose of 75 mg has only a minimal and transitory effect on thesynthesis of cyclooxygenase by endothelial cells.No studies have been published specifically with SLE or

ACLA. Results of women with very bad obstetric historieswhich include women with SLE and/or anticardiolipinantibodies have shown a potential benefit oflow dose aspirin.It seems therefore that ACLA may occlude the utero-placental circulation via platelet aggregation either directlyor through other, as yet, unknown factors.

Reference

1. Derue, G.J., Englert, H.J., Harris, N. et al. Fetal loss in systemiclupus: association with anti-cardiolipin antibodies. J ObstetGynaecol 1985, 5: 207-209.


Relative risk of recurrent abortion and thrombosisin young subjects with antiphospholipid antibodiesestimated by case-control studies

G. Finazzi, S. Cortelazzo, M. Galli and T.Barbui

Divisione di Ematologia, Ospedali Riuniti, Bergamo,Italy

A number of clinical observations showed an associationbetween the antiphospholipid antibodies (APA), lupusanticoagulant (LAC) and anticardiolipin (ACA) with clinicalcomplications, such as thrombocytopenia, recurrent abor-tion and thrombosis. However, the relative risk to developthese complications in patients with ACA, as compared withnormal subjects, is unknown. We estimated this risk bymeans of 'case-control' studies.'

'Cases' were patients without clinical evidence of systemiclupus erythematosus or other immunological disorders whohad had two or more spontaneous abortions, venous throm-boembolism or ischaemic cerebrovascular disease under 45years of age, admitted between January 1985 and June 1988to the Ospedali Riuniti of Bergamo. 'Controls' were patientswithout the disorders above defined, admitted to the samehospital where the cases had been identified for acuteconditions other than immunological, neoplastic, gynaeco-logical, or cardiovascular.APA were determined in cases and controls as follows:

LAC was diagnosed when at least 3 of the following criteriawere present: (1) APTT ratio> 1.2; (2) kaolin clotting timeratio> 1.3; (3) dilute Russell viper venom test ratio> 1.2, allnot corrected in a 1: I mixture with normal plasma, and (4)positive tissue thromboplastin inhibition test; IgG ACA wereassayed with an ELISA test and were considered positivewhen > 10 U.APA were observed in: 7 out of 115 cases with recurrent

abortion vs I out of 141 controls, relative risk = 9.07;P<0.05 (chi square test); 6 out of 76 cases with juvenilevenous thromboembolism vs I out of 190 controls, relativerisk = 16.2; P< 0.01; 5 out of 43 cases with juvenile cerebralischaemia vs I out of 190 controls, relative risk 25; P< 0.001.These results provide the first estimate ofthe relative risk of

recurrent abortion and thrombosis in young subjects withAPA without SLE. The markedly elevated risk suggests thatAPA should be determined in all these patients and calls forclinical trials aimed to establish the best therapeutic app-roach.

Reference

1. Barbui, T., Cortelazzo, S., Galli, M., et al. Antiphospholipidantibodies in early repeated abortions: a case-controlled study.Fertil Steril 1988, 50: 589- 592.

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/


Laboratory AspectsFurther studies on the heterogeneity of antiphos-pholipid antibodies

J. Arnout, E. Huybrechts and J. Vermylen

Centerfor Thrombosis and Vascular Research, Univer-sity of Leuven, Herestraat 49, 3000 Leuven, BelgiumWith the development of the ELISA technique for measuringanticardiolipin antibodies, it soon became apparent that,while in many patient samples there was a reasonableagreement between the presence of a lupus anticoagulant andthat of anticardiolipin antibodies, this association is far fromconstant, occasional plasmas having very potent lupusanticoagulant activity without anticardiolipin antibodies, orvice versa. Several explanations can be put forward for suchdiscrepancies.The first possibility we investigated was the effect of using

other phospholipids in the ELISA, since phosphatidylserinerather than cardiolipin appears to be the physiologicalphospholipid that promotes the interaction of coagulationfactors. One could indeed have expected a closer relationshipbetween lupus anticoagulant tests and an antiphosphati-dylserine ELISA than an anticardiolipin ELISA.

Microtitre plates were therefore coated with cardiolipin, orphosphatidylserine, or Thrombofax, a commercial plateletphospholipid substitute. Lupus anticoagulants were assayedusing the tissue thromboplastin inhibition test, the kaolinclotting time, the kaolin clotting time on a 1:1 mixture withnormal plasma, and the activated partial thromboplastintime. One hundred and forty six selected patient plasmaswere studied, on which determination of lupus anticoagulanthad been requested. Once more lupus anticoagulant positive,ELISA negative samples were identified, and vice versa.

Correlation coefficients showed that in these samples therewas reasonable agreement between the clotting tests (r valuesaround 0.75), good agreement between the ELISAs (r valuesaround 0.80), but only moderate agreement between clottingtests and ELISAs (r values around 0.50). The latter relation-ship does not change markedly by modifying the phos-pholipid used.The second possibility we have started to look into is that

antiphospholipid antibodies have a different affinity forphospholipid in a suspension than for phospholipid adsorbedto a surface. We have first investigated the adsorption ofphospholipid antibodies onto platelet vesicles. The latterwere generated by incubating platelets with A 23,187; thiscalcium ionophore activates calcium dependent protease; thelatter hydrolyses crucial proteins of the platelet membraneskeleton, such as actin binding protein; release of smallvesicles results; these contain most of the procoagulantactivity. Such platelet vesicles correct the clotting in lupusanticoagulant plasmas by providing a suitable catalyticsurface. However, they fail to adsorb the anticoagulant: theirremoval by centrifugation regenerates the anticoagulanteffect; platelet vesicles also fail to reduce the content ofantiphospholipid antibodies in positive plasmas, as measuredby the ELISAs. We conclude that there may be largedifferences in affinity of antiphospholipid antibodies for

phospholipids, when the latter are part of a platelet vesicle(virtually no binding) or in suspension in plasma, or attachedto an artificial surface. Differences in binding in the two lattA*situations can perhaps account for some patient plasmasbeing anticoagulant positive, ELISA negative, or vice versa.

Quality control of the lupus anticoagulant test inthe UK

D. Taberner, S. Machin, I. Mackie, J. Giddings,R. Malia and M. Greaves

Haematology Departments, Withington Hospital,Manchester, Middlesex, Hospital, London, UniversityHospital of Wales, Cardiff and Royal HallamshireHospital, Sheffield, UK.

The clinical importance of the lupus-like anticoagulant (LA)is becoming widely recognized. This inhibitor prolongsvarious phospholipid dependent coagulation tests particular-ly the activated partial thromboplastin time (APTT).Routine coagulation laboratories therefore need to be able toidentify the defect. Plasma from known patients with theinhibitor have been included in the regular UK qualitycontrol exercises in blood coagulation (NEQAS) since 1983.These studies have highlighted the dependence of the sen-sitivities of APTT reagents to the inhibitor upon theirphospholipid content. Reagents with the lowest phospho-lipid content have shown greatest sensitivity to the defect.

In 1987 the British Society for Haematology appointed asmall working group to consider the standardization ofmorespecialized LA tests. As an initial exercise, using a question-naire, current laboratory practice in detecting the LA wasdetermined. Of the 433 regularly participating in NEQAS incoagulation 255 responded, 183 (73%) regularly testing forLA. Laboratories generally perform an APTT combinedwith one or other specialized tests (commonly kaolin clottingtime,' thromboplastin inhibition,2 dilute Russell's viper ve-nom test3), but some exclude the inhibitor using an APTTalone. Variation in methodology was also noted. In asignificant proportion of centres conditions of centrifugationwere not adequate to remove platelets. Only a small numberensure that the plasma is platelet-free prior to testing orfreezing (i.e. avoiding platelet by-passing activity in testsystems). Only 6% perform anticardiolipin assays in thehaematology laboratory but 80% have access to such testsvia other laboratories.The working party examined the effect of prior filtration

and freezing on the kaolin clotting time and dilute Russell'sviper venom tests. It was clearly shown that when samples aretested after freezing prior filtration to prevent plateletcontamination is essential for adequate performance of thekaolin clotting time test. Double centrifugation is lessefficient at removing the LA bypassing activity of con-taminating platelets.

In a further exercise, 3 freeze dried plasmas were includedfor testing by the 183 participants (one weak and one strongerLA plasma and a negative control sample). Overall only 35%identified the weak inhibitor compared to 65% identifyingthe stronger one. Seven laboratories claimed to find inhibitor

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/


in the negative control sample. No particular method typeappeared superior although within methods certain technicaldifferences were noted between successful and unsuccessfulcentres. In particular in the tissue thromboplastin inhibitortest too little or too great dilution of thromboplastin impairssensitivity or specificity respectively. In the kaolin clottingtime test short normal control times (suggesting inadequateremoval of platelets) are associated with loss of sensitivity tothe inhibitor. The results emphasized that there is a need toimprove the ability of routine laboratories to confidentlyidentify LA patients. A follow up exercise using standardizedtechniques has been planned whereby comparison betweenperformance of standard and local methods will be eval-uated. It is hoped that recommendations can then be maderegarding standardized methodology which will ensure theidentification of this clinically important defect.

References

1. Exner, R., Rickard, A. & Kronberg, H. A sensitive test demons-trating lupus anticoagulant and its behavioural pattern. Br JHaematol 1978, 40: 143-148.

2. Schleider, M.A., Nachman, R.L., Jaffe, E. A. & Coleman, M. Aclinical study of the lupus anticoagulant. Blood 1976, 48: 499.

3. Thiagarajan, P. & Shapiro, S.S. Lupus anticoagulants. In: Col-eman, M.A. (ed) Methods in Haematology: Disorders of ThrombinFormation other than Haemophilia. Churchill Livingstsone, Edin-burgh, London, 1983.

The laboratory investigation of patients with lupusanticoagulant

L.J. Holland', S.P. Woodcock', & M.A. Kham-ashta2

'Department of Haematology and 2Lupus ArthritisResearch Unit, St. Thomas' Hospital, London SE]7EH, UK.

Hitherto the detection of the lupus anticoagulant (LA) in thecoagulation laboratory was viewed with the mixed responseof academic interest, balanced with the concern that itspresence may involve bleeding complications.' Haemorr-hagic involvement is an extremely infrequent occurrence, butthe link between LA and thrombosis2 has altered theinvestigative procedure within the confines of haemostasisand made its detection a high profile matter.

Initially LA investigation in our laboratory was a rarelyrequested procedure, and the tests included the prothrombintime [reported as international normalized ratio (INR)], theactivated partial thromboplastin time (APTT) in conjunctionwith suitable normal plasma mixtures, the thrombin time,and two confirmatory LA tests, i.e. the Oxford SLE inhibitortest and the kaolin clotting time or Exner test.3 Prothrombinand thrombin times were normal in the presence ofLA, whilstthe APTT and mixture tests were generally prolonged andshowed signs of inhibition. The Oxford test has exhibitedidiosyncratic results on occasions and is a rather labourintensive and lengthy investigation. This test soon yielded tothe more reliable and less manipulative Exner test. All studies

require fresh plasma if the problems of cold activation are tobe avoided.

In recent years the detection of the LA has become acommon and often requested investigation, and there arenumerous local tests which can be included for its detection,e.g. Russell viper venom time, tissue thromboplastin inhi-bitor tests, etc. We found that the intensity of investigationrequired rationalization within the laboratory, if we were tocope with the many other studies required of us from thenon-LA patients.

Early on in the proceedings we included, by chance, a testnormally reserved to detect minor extrinsic coagulationdefects, in our LA investigations. This test, the dilute brainthromboplastin ratio (DBTR), involved diluting Manchesterthromboplastin in 50 mm tris buffer to 1 500 of its originalconcentration, and subjecting the control and patient'splasma to a conventional prothrombin time analysis, whilstsubstituting the diluted reagent. The result was reported as aratio, i.e.

Patient's clotting timeNormal clotting time

We undertook a survey of 33 patients with SLE relatedconditions, many of whom were treated for this underlyingcondition, and all of whom have previously shown thecharacteristic coagulation abnormalities of LA. More than75% of the patients tested had an abnormally raised DBTR,where the APTTs were normal and anticardiolipin antibodies(ACA) measured.The elevation of the DBTR may not be specific for LA, as

minor deficiencies in the extrinsic clotting pathway mightaccount for some of the increase, but we feel that thismajority increase is highly indicative of LA. The disparitynoted between Exner and ACA results in some patientsprobably relates to the fact that many were receiving specifictreatment for their condition. It had been reported previouslythat in these situations the Exner test may become negative,and that the level of ACA may disappear completely. In allpatients studied to date the elevated DBTR has been the oneparameter which has remained elevated throughout.The fact that the one-stage prothrombin time (INR) is

generally normal in the presence ofLA probably depends onthe type of concentration of thromboplastin extract. Sincethe inhibition interferes with factors Vlll:C, IX:c(Capt),V:c(Capt) and X:c(Capt), or factors Xa-V-II complex on thephospholipid surface, its activity is likely to be more pro-nounced in the intrinsic rather than extrinsic clotting path-ways, but in any event it is phospholipid concentrationdependent. The DBTR is a useful diagnostic tool to signifyincrease in inhibitor activity as the brain reagent is diluted.This thromboplastin substance is a protein-phospholipidmoiety, and higher concentrations will neutralize the effect ofthe lupus inhibition.The importance of detection of LA in respect to throm-

botic complications is now clear, and by following a simpleand cost effective system of laboratory tests these anti-coagulants can be detected in any routine coagulationlaboratory.

References

1. Conley, L. & Hartman, R.C. A haemorrhagic disorder caused bycirculating anticoagulants in patients with disseminated lupuserythematosus. J Clin Invest 1952, 31: 621-622.

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/


2. Bowie, W.E.J., Thompson, J.H., Pasazzi, C.A. & Owen, C.A.Thrombosis in systemic lupus erythematosus despite circulatinganticoagulants. J Clin Invest 1963, 62: 416-430.

3. Exner, T., Rickard, K.A. & Kronenberg, H.A. Sensitive testdemonstrating lupus anticoagulant and its behavioural patterns.Br J Haematol 1978, 40: 143-155.

Specificities of anti-phospholipid antibodies

E. Nigel Harris

Division of Rheumatology, University of Louisville,Louisville, KY 40292, USA.

Antiphospholipid (aPL) antibodies are a complex family ofantibodies whose individual characteristics are yet to beclearly defined. Currently, they are best classified accordingto the tests by which they are detected viz. Standard Tests forsyphilis (STS), lupus anticoagulant (LA), and solid-phaseanticardiolipin (aCL) tests. There can be overlap betweenthese tests and this overlap is greatest between LA and aCLtests, both of which correlate with thrombosis and recurrentfetal loss. The STS is most frequently positive in patients withsyphilis, usually at high titres, but in the patient group definedby positive LA/aCL tests, the STS is either negative or lowtitre positive.

Antibodies responsible for positive LA and aCL testscross-react with negatively charged phospholipids. A seriesof studies',2'4 have shown that pre-incubation ofaCL positivesera with phosphatidylserine (PS), phosphatidic acid (PA), orphosphatidyl inositol (P1), caused inhibition of cardiolipinbinding activity. This inhibition by negatively charged phos-pholipids was comparable to that obtained with equivalentconcentrations of cardiolipin.'4 Immunoglobulins with LAactivity cross-react with negatively charged phospholipidboth by Ouchterlony and ELISA assays.3 Anticardiolipinantibodies, like LA immunoglobulins, bind ELISA platescoated with CL, PS, PA or PI.'-4 Neither group is inhibitedby or bind plates coated with zwitterionic phospholipids, e.g.phosphatidylcholine (PC), sphingomyelin (SM). Cardiolipinliposomes have been used to affinity purify antibodies whichwere subsequently shown to have LA activity.2'3 The affinitypurified antibodies cross-react with negatively charged phos-pholipids.2'3 On the basis of these studies, it can be concludedthat aCL and immunoglobulin with LA activity have over-lapping specificities, though the occurrence of LA positive/aCL negative or (more frequently) aCL positive/LA negativesera suggests subtle differences in specificities or bindingactivities of these two groups of antibodies.

Antiphospholipid antibodies bind cardiolipin but thisbinding is enhanced by mixing cardiolipin with PC andcholesterol (chol). Since these antibodies do not bind PC orchol, alone, it appears that the PC and chol together cause analteration in the conformation ofCL favourable for bindingby syphilis aPL antibodies. In contrast to syphilis, theaCL/LA group bind CL less when it is mixed with PC andchol.4 Antiphospholipid antibodies of syphilis cross-reactwith negatively charged phospholipids much less than do theLA/aCL group.

Negatively charged phospholipids are present in plateletand endothelial cell membranes. It is possible that cross-

reacting LA/aCL antibodies can bind these membranephospholipids altering cell function in such a way as topresent thrombosis. The aPL antibodies of syphilis areconfined largely to CL binding (less cross-reaction withnegatively charged membrane phospholipids) and, as aresult, alteration ofcell function and thrombosis might be lesslikely in syphilis.

References

1. Harris, E.N., Gharavi, A.E., Loizou, S. et al. Crossreactivity ofanticardiolipin antibodies. J Clin Lab Immunol 1985, 16: 1-6.

2. Harris, E.N., Gharavi, A.E., Tincani, A. et al. Affinity purifiedanticardiolipin and anti-DNA antibodies. J Clin Lab Immunol1985, 17: 155-162.

3. Pengo, V., Thiagarajan, P., Shapiro, S.S. & Heine, J. Immuno-globulin specificity in mechanisms of IgG anticoagulants. Blood1987, 70: 69-76.

4. Harris, E.N., Gharavi, A.E., Wasley, G.D. & Hughes, G.R.V. Useof an enzyme linked assay and of inhibition studies to distinguishbetween antibodies to cardiolipin from patients with syphilis orautoimmune disorders. J Infect Dis 1988, 157: 23-31.

Lupus anticoagulant and anticardiolipin anti-bodies: the affinities may differ

Henrique L. Staub, Munther A. Khamashta,Nigel Harris, Elaine Baguley, Wiliam H. Cha-hade and Graham R.V. Hughes

Lupus Arthritis Research Laboratory, The Rayne Ins-titute, St. Thomas Hospital, London, UK.

There have been previous suggestions that anticardiolipinantibodies (ACA) and the lupus anticoagulant (LA) testsconcord in about 90% of cases. Recent studies, however,have highlighted a greater rate of discrepancies between theseassays.From our series of patients with antiphospholipid anti-

bodies, we analysed seven cases in which discordancebetween ACA and LA tests could be well documented. Fourof the patients presented with antiphospholipid syndromeassociated with systemic lupus erythematosus (SLE), andthree patients showed primary antiphospholipid syndrome.'Antibodies to the negatively charged cardiolipin, phos-phatidylserine (PS) and phosphatidic acid (PA) were assessedin an ELISA; the same assay was used to determine bindingto the neutral lipid phosphatidylethanolamine (PE), toVDRL and to the activated partial thromboplastin time(APTT) reagent thrombofax. The LA tests, apart from thescreening with APTT, included the Exner test and the indexof circulating anticoagulant (ICA). Cardiolipin, PS, phos-phatidylcoline (PC), PE and thrombofax, either in solutionor micelles form, were used in absorption studies of LAactivity in a kaolin clotting time (KCT) assay.

All patients showed prolonged APTT (thrombofax beingthe phospholipid preparation used), positive Exner test andhigh ICA, findings confirmatory of presence of LA activity.Six of these patients, unexpectedly, did not show any bindingto cardiolipin, PS, PA, PE, VDRL or even thrombofax in an

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/


ELISA. The seventh, with primary antiphospholipid synd-rome and strong LA activity, presented unique binding toPE. In all plasmas, hexagonal phase PE and thrombofax butnot cardiolipin, PS, PC or PE in micelles form removed LAactivity in a kaolin clotting time (KCT) assay.

These findings, altogether, hint that ACA and LA affinitiesmay be different. Recent reports with electrofocussinganalysis disclosed that some, but not all, ACA antibodieshave LA activity. It is well known that mostly negativelycharged phospholipids are antigenic in solid assays, whereasbinding to neutral lipids appears not to be the rule.

However, that neutral lipids such as PE are playing a roleas antigens in some way, has been proved in recent issues. Inour laboratory, PE extracted from platelet membranes wasfound to inhibit ACA activity more than any other phos-pholipid.2 One patient with severe recurrent thromboticdisease and LA activity showed unique binding to PE in anELISA; these findings were confirmed by inhibition studiesand affinity purification.3 When we looked at absorptionstudies in coagulation assays, hexagonal (but not lamellar)phase PE considerably removed LA activity.Why these patients, though presenting prolonged APTT,

did not bind thrombofax in an ELISA, is a matter ofdiscussion. The thrombofax preparation, as analysed in lipidextraction studies and thin layer chromatography, containsall sorts of phospholipids (PS the main one) but notcardiolipin. The cross reaction theory between negativelycharged phospholipids has been proved true, and themajority of patients with prolonged APTT and LA activityalso bind cardiolipin, PS, PA, phosphatidylinositol (PI) orthrombofax in an ELISA. However, in some instances, suchas our survey, clearly the APTT and the antithrombofaxassay in an ELISA discord. We can not rule out, at this stage,that 'auxiliary lipids' (? PE and PC) be enhancing the LAphenomenon but not antithrombofax antibodies. In addi-tion, the role of an apoprotein which binds factor VII (alsopresent in the thrombofax preparation) is not fully knownwhen we look again at the two assays.What emerges from these observations is that, if we

consider antiphospholipid antibodies in both ELISA andcoagulation assays, it may be that the charge ofthe lipid is notso crucial in promoting antigenicity or LA as the phase inwhich one presents the lipids in these assays. Above all, in apractical sense, both ACA and LA tests should be performedif clinical indication exists.

References

1. Hughes, G.R.V. The anticardiolipin syndrome. Clin Exp Rheum-atol 1985, 3: 285-286.

2. Khamashta, M.A., Harris, E.N., Gharavi, A.E. et al. Immunemediated mechanism for thrombosis: antiphospholipid antibodybinding to platelet membranes. Ann Rheum Dis 1988, 47: 849-854.

3. Staub, H.L., Harris, E.N., Khamashta, M.A., Savidge, G.,Chahade, W.H. & Hughes, G.R.V. Antibody to phosphati-dylethanolamine in a patient with lupus anticoagulant andthrombosis. Ann Rheum Dis 1989, 48: 166-169.

Anticardiolipin antibodies, lupus anticoagulantand viral infection

I.J. Mackie, H. Cohen, C.B., Colaco, W. Irving',S.J. Mackin

Departments ofHaematology and Rheumatology, TheMiddlesex Hospital, London, and 'Virology Depart-ment, John Radcliffe Hospital, Oxford, UK.

Antiphospholipid (APL) antibodies have been reported inpatients with HIV and other infections such as infectiousmononucleosis (IM). Haemophiliacs are a high risk group forviral transmitted diseases including HIV infection, andevidence of lupus anticoagulant (LA) and anticardiolipinantibody (ACL) was therefore sought in a group of 22multitransfused haemophiliacs. ACL levels were also mea-sured in a group of 40 non-haemophiliac patients with viralinfections (10 each with: parvovirus, hepatitis A, rubella, orIM). Twelve of 22 haemophiliacs were LA positive; 9 of thesewere ACL positive and 9 were anti-HIV positive (I ACL + vewas anti-HIV - ve and vice versa), but associations betweenLA, ACL and anti-HIV status were not statisticallysignificant. Mixing studies showed that the LA test wasaffected when the FVIII level fell below 15 IU/dl; however, inthe haemophiliacs, there was no strict correlation betweenFVIII level and LA positivity. The addition of porcine FVIIIto correct the patient's level did not affect LA positivity in8/12 patients, suggesting a true LA due to APL activity. Innon-haemophiliacs, positive results for IgG and IgM ACLwere found in 6 and 8 with parvovirus, 8 and 8 rubella, 9 and10 hepatitis A, and 3 and 7 with IM. Highest titres were foundin parvovirus and hepatitis A; IgG and IgM ACL levels didnot correlate. Positive LA and ACL results in haemophiliacsmay represent a response to viraemic challenge; antibodiesnormally suppressed may be deregulated, or viral particlesmay damage or modify membranes, exposing negativelycharged phospholipids or causing hapten formation.

Correspondence: Timo Palosuo, M.D. National PublicHealth Institute, Mannerheiminitie 166, SF-00300 Helsinki,Finland.

Are cardiolipin-binding antibodies in Gram-nega-tive infections induced by bacterial lipopolysac-charide?

Outi Vaarala', Martti Vaara2 and Timo Palosuo'

'National Public Health Institute, Helsinki, and2Department ofBacteriology and Immunology, Univer-sity of Helsinki, Finland.

Anticardiolipin antibodies (ACA) as detected by solid-phaseimmunoassays are frequently seen in many different infec-tions.1'2 The specificity of ACA has aroused considerable

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/


interest. Several studies suggest that cardiolipin-bindingantibodies are a heterogenic group of antibodies withdifferent subspecificities. Antibodies binding to solid-phasecardiolipin have been shown to bind readily to othernegatively charged phospholipids such as phosphatidylserineand phosphatidylinositol.3 Our recent observation that in-fection-associated antibodies bIinding to solid-phase car-diolipin may be poorly inhibitable by liquid cardiolipin,'prompted us to search for other compounds, perhaps struc-turally related to cardiolipin, to which these antibodies mightbind more effectively.

Lipopolysaccharide (LPS), which is an immunogenic com-ponent of the outer membrane of all Gram-negative bacteria,was chosen as one putative antigen. One antigenic deter-minant of LPS is its lipid A part which consists of diphos-phorylated glucosamine disaccharide to which 6 to 7 fattyacid chains are bound. The negatively charged phosphategroups and fatty acid chains of lipid A molecule aredeterminants common with cardiolipin.We reported recently that raised ACA levels were detected

by solid-phase enzyme-immunoassay in the majority of serafrom patients with severe bacterial infections.4 The responseinvolved all immunoglobulin classes tested; IgG, IgM andIgA. The study material comprised paired sera from 41patients with Gram-positive septicaemia (21 staphylococcaland 20 streptococcal septicaemia) and 26 patients withGram-negative infections (20 salmonella infections and 6Escherichia coli septicaemia). The specificity of the ACA wasstudied in competitive inhibition experiments with threeantigens: cardiolipin, LPS isolated from Salmonella min-nesota, strain Re 959, and synthetic E. coli lipid A. Five serafrom patients with Gram-negative infections were tested(four Salmonella sera and one E. coli serum). The binding ofIgG class ACA from all these sera (diluted 1: 50) waseffectively inhibited by LPS at a concentration of 500 ng/ml,whereas 100-fold more cardiolipin was required for com-parable inhibition. Pure lipid A was a less effective inhibitorofanticardiolipin activity than LPS. This pattern of reactivitywas not seen in sera from patients with Gram-positiveinfections, syphilis or systemic lupus erythematosus. Theresults of these studies are reported in detail elsewhere.4Our findings suggest that the primary antigenic stimulus

for cardiolipin-binding antibodies in Gram-negative infec-tions might not be cardiolipin but perhaps LPS, or someLPS-related structure.

References

1. Vaarala, O., Palosuo, T., Kleemola, M. & Aho, K. Anticardiolipinresponse in acute infections. Clin Immunol Immunopathol 1986,41:8-15.

2. Canoso, R.T., Zon, L.I. & Groopman, J.E. Anticardiolipinantibodies associated with HTLV-III infection. Br J Haematol1987, 65: 495-498.

3. Harris, E.N., Gharavi, A.E., Tincani, A. et al. Affinity purifiedanti-cardiolipin and anti-DNA antibodies. J Clin Lab Immunol1985, 17: 155- 162.

4. Vaarala, O., Vaara, M. & Palosuo, T. Effective inhibition ofcardiolipin-binding antibodies in Gram-negative infections bybacterial lipopolysaccharide. Scand J Immunol 1989, (in press).

Endothelial cellsMonoclonal antibodies (McAb) against DNA alsorecognise epitopes on endothelial cells

G. Frampton', A. Morgan2, N.A. Staines' andJ.S. Cameron'

Renal Unit, 'Guy's Hospital, UMDS, 2St. George'sHospital, 3King's College, London, UK.

Autoantibodies against DNA (DNAb) have been demon-strated in and eluted from lupus nephritis kidneys whichsuggests a pathogenic role in the glomerulonephritis. Therenal lesion is thought to arise from circulating or in situimmune complex deposition. DNAb bind to cell-surfacemembranes and experimentally, glomerulonephritis can beinduced by anti-endothelial antibodies. The aim of this studywas to see whether McAb with activity against DNA alsobind to human umbilical vein endothelial cells (HVEC) andinfluence their anti-thrombogenic nature.A library of 8 McAb with activity against DNA' were

obtained from mice with lupus-like disease. The McAb weretested for their ability to bind to unfixed HVEC either insuspension or on plastic ELISA plates. Binding of McAb toHVEC was determined using cell ELISA, flow cytometry orfluorescence microscopy. To exclude the possibility ofMcAbbinding to HVEC-bound DNA, DNase digestion experi-ments were carried out. Two of the McAb bound to: (i)freshly prepared, collagenase-digested HVEC; (ii) HVECgrowing on ELISA plates; (iii) HVEC in suspension aftertrypsin digestion and (iv) to HVEC after DNase digestion.These results suggest that the target antigen(s) is not DNAand probably not protein as it is resistant to both collagenaseand trypsin. Functionally, the binding of the McAb did notinhibit thrombin-induced prostacyclin synthesis, in contrastto anti-HVEC McAb EN4 which does not bind to DNA.

This study demonstrates that some McAb with activityagainst DNA also bind to HVEC. However, the nature of theepitopes on the endothelial cell membranes recognized bythese McAb remains unclear. The findings in this study mayhelp explain why anti-DNA antibodies are important in thepathogenesis of lupus nephritis, perhaps through their abilityto bind directly to the vasculature resulting in complementactivation, platelet adhesion/activation which could lead toincreased glomerular permeability and the development ofglomerulonephritis.

Reference

1. Morgan, A., Buchanan, R.R.C., Lew, A.M., Olsen, I. & Staines,N.A. Five groups of antigenetic determinants on DNA identifiedby monoclonal antibodies from (NZBxNZW)FI and MRL/Mp-I pr/ I pr mice. Immunology 1985, 55: 75- 83.

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/


Antibodies to endothelial surface molecules

Dorian 0. Haskard

Rheumatology Unit, Division of Medicine, UMDS,Guy's Campus, 4th Floor, Hunt's House, Guy's Hos-pital, London, SE] 9RT, UK.

Over the last few years it has become evident that vascularendothelial cells (EC) interact functionally with both solubleand formed elements of the blood and that this interactionplaces EC in a comparable position to macrophages in theirpotential to regulate immune and non-immune acute andchronic inflammatory responses. Moreover, these interac-tions may play a key role in the pathogenesis of vasculitis.EC show a complex orchestrated response to cytokines

which involves a large number of proinflammatory effects,including an alteration in the expression of surface mem-brane molecules and an alteration in the profile of ECsecreted products. The cytokines interleukin 1 (ILI), tumournecrosis factor alpha (TNF) and lymphotoxin, together withbacterial lipopolysaccharide (LPS) constitute a family ofagents which have similar actions on EC (Table 1). Theseinclude an increased adhesiveness for leucocytes, the secre-tion of leucocyte activating factors such as PAF and GM-CSF, and a shift in the balance of the clotting system towardsan increased tendency to coagulation and an inhibition offibrinolysis. Interferons (IFN) also have actions on EC thatimportantly includes an increased expression ofMHC Class Imolecules (IFN a, P, and y) and the induction ofMHC ClassII molecules (IFNy).

Table I Principal actions of ILI and TNF on vascularendothelial cells

Enhanced adhesiveness for NeutrophilsEosinophilsBasophilsMonocytesLymphocytes

Induction/up-regulation of activation antigensSecretion of colony stimulating factors

GM-CSFG-CSFM-CSF

Membrane expression and secretion of platelet activatingfactor

Secretion of prostaglandins E2 and I2 (prostacyclin)Expression of tissue factor activityReduced thrombomodulin activityReduced plasminogen activator secretionIncrease plasminogen activator inhibitor secretionSecretion of platelet derived growth factorSecretion of ILI

Several laboratories have succeeded in generating mono-clonal antibodies (mAb) against cytokine induciblemolecules on EC. Such mAb can be used to characterizeprecisely the nature of endothelial activation in vitro. Theycan also be employed to validate in vitro models by

confirming the presence of the activation antigens with whichthey react within inflammatory lesions in situ. Thus mAb thatreact with activation antigens on ILI or TNF stimulatedumbilical vein EC have been used to identify EC activation inskin during a delayed type hypersensitivity reaction (mAbH4/18)', within various acute inflammatory lesions (mAb4DIO)2 and within rheumatoid synovium (mAb I .2B6).3With these and similarmAb it may be possible to differentiatetypes of inflammation on the basis of the nature ofendothelial activation.The local actions of cytokines could be expected to be

central to the evolution of many types of vasculitis. By usingmAb against EC activation antigens to assess the nature ofendothelial activation in vasculitic lesions, we may be ablenot only to advance our understanding of basic patho-physiological mechanisms but also to improve on existentpathological and clinical classification.

References

1. Cotran, R.S., Gimbrone, M.A., Bevilacqua, M.P., Mendrick, D.L.& Pober, J.S. Induction and detection of a human endothelialactivation antigen in vivo. J Exp Med 1986, 164: 661-666.

2. Goerdt, S., Zwadlo, G., Schlegel, R., Hagemeier, H.-H. & Sorg, C.Characterization and expression kinetics of an endothelial cellactivation antigen in vivo present only in acute inflammatorytissues. Exp Cell Biol 1987, 55: 117-126.

3. Wellicome, S., Pitzalis, C., Panayi, G.S. & Haskard, D.O. Endo-thelial cells in rheumatoid synovium express an antigen induced byinterleukin-I (ILI) and tumour necrosis factor. Br J Rheumatol1988, 27 (suppl 2): 66.

Sera of patients with systemic lupus erythe-matosus enhance the tumour necrosis factor in-duced expression of endothelial cell procoagulantactivity

R.H.W.M. Derksen, P. Hasselaar, J.D. Oosting,L. Blokzijl and P.G. de Groot

Department of Internal Medicine and Haematology,University Hospital, Utrecht, The Netherlands.

A strong correlation has been found between the presence ofantiphospholipid antibodies (APA) and a history of throm-bosis in patients with systemic lupus erythematosus (SLE). Apossible mechanism resulting in an increased risk for throm-bosis is a stimulation of the endothelial cell (EC) pro-coagulant activity (PCA) expression. We studied the effect of15 APA positive SLE sera, 13 APA negative SLE sera, 10 serafrom patients with other connective tissue diseases (CTD)and 15 control sera on the tumour necrosis factor (TNF)induced expression of EC PCA. EC were stimulated withTNF and 20% serum for 4 hours. PCA expression wasmeasured on the EC surface and in their matrix (ECM) byincubating the ECM with normal pooled plasma, after whichthrombin generation was determined with a chromogenicsubstrate. In the absence ofTNF, none of the sera stimulatedPCA expression; 14/15 APA positive SLE sera, 7/13 APAnegative SLE sera, 2/10 CTD sera and 2/15 control sera

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/


enhanced PCA expression after stimulation of EC withsuboptimal doses of TNF. This stimulating effect resided inthe IgG fraction and was associated with the presence ofAPA, but not with anti-EC antibodies or a history ofthrombosis. Affinity purified APA had no PCA stimulatingactivity. PCA expression was inhibited by cycloheximide andheat treatment (30 min, 56°C) of serum. We conclude thatmost SLE sera increase the TNF-induced endothelial PCAexpression. Although this synergistic effect predominantlyoccurs with APA positive serum or IgG, a causative role ofAPA was not demonstrated.

Autoantibodies to an endothelial cell alpha granuleprotein (GMP- 140) in sera of patients withvasculitis.

G.A. McCarty, Morris Reichlin, Kathleen Listerand Rodger McEver

Oklahoma Medical Research Foundation, OklahomaCity, OK 73104

Endothelial cells are central in the pathogenesis of systemicvasculitis (SV) but the antigens against which autoantibodiesto endothelial cells (AECA) are made are poorly charac-terized. Vascular endothelial cells (EC) contain an antigencalled alpha-granule membrane protein (GMP) recentlydescribed by McEver et al,' which is shared with platelets, butexpressed only when EC and platelets are activated. Todetermine if sera from patients with vasculitis bind to GMP,an antigen capture enzyme-linked immunosorbent assay(ELISA) was modified and sera from 44 patients and controlsscreened for HIGH (OD 405 mm = 0.5-1.5), LOW

(0.1-0.5), or NEGATIVE (0.0-0.1) reactivity in 44 blindedsamples from well defined, biopsy-proven systemic vasculitispatients (n = 24), osteoarthritis patients (n = 5), and normalcontrols (n= 15).

Diagnoses Number High Low Negative

Systemic vasculitis 12 2 2 8Lupus vasculitis 2 1 1 0Wegener's 3 2 1 0Periarteritis nodosa 3 0 0 3Rheumatoid vasculitis 3 0 0 3CNS vasculitis 1 0 0 1Controls 20 0 0 20

There was no statistically significant correlation with thepresence of AECA as defined with bovine and human ECsubstrates, or with anti-phospholipid antibodies (APL) asdefined in a well standardized assay with bovine cardiolipin.The height of anti-GMP reactivity correlated generally withactive vasculitis in the 5 patients with HIGH, and 4 patientswith LOW reactivity; normal and osteoarthritis controlswere NEGATIVE. Since anti GMP is not present in neut-rophils, the reactivity is not related to the cytoplasmic antigennoted in Wegener's. Anti-GMP autoantibodies may prove auseful diagnostic tool for several forms of systemic vasculitiswhere previously only invasive studies have sufficed.

Reference

1. Rosa, J-P., George, J.N., Bainton, D.F., Nurden, A.T., Caen, J.P.& McEver, R.P. Grey platelet syndrome. Demonstration of alphagranule membranes that can fuse with the cell surface. J Clin Invest1987, 80: 1138-1146.

copyright. on M


http://pmj.bm

j.com/

Postgrad M


eptember 1989. D

ownloaded from

http://pmj.bmj.com/

Abstracts Francesc Casals', and Miguel Ingelmo · 698 ANTIPHOSPHOLIPIDSYNDROMEWORKSHOP Laboratory Aspects pholipid antibodies J. Arnout, E. Huybrechts andJ. Vermylen Centerfor antibodies,

Documents