Antibodies to cardiolipin in patients with primary biliary cirrhosis

ABST RACT : Antibodies to cardiolipin have been recognized in u p to 65% of ratients with systemic lupus e rythematosus. It has been claimed that they are significantly associated with intravascular th rombosis and with o bstetrical complications. Thus fa r they have been fou nd to be less prevalent in 0th.er d iseases. Because of the h igh concentration of cardiolipin in mitochondrial membranes and the p resence of antimitoch.ondrial antibodies in patients with p rimary biliary cirrhosis, the au thors investigated the prevalence of antibodies to card iolipin in a group of 31 patients with p rimary b iliary cirrhosis. lt was found that the p revalence and levels of an ticardiolipin antibodies oflgG and lgA isotype are as high in patients with p rimary b ilia ry cirrhosis as in 35 consecutive patients with systemic lupus erythematosus. None of the patients wi th primary b iliary cirrhosis gave any histo ry of venous or arterial th rombosis. The rate of m iscarriage was less than that reported for the general population. The availability for study of a second group of patients with h igh levels of an ticardiolipin antibodies should make it possible to determine whether the association of these antibodies with th rom bosis and fe tal wastage in patients with systemic lupus erythemacosus is a d irect relation ship or an epiphenomenon. C a n J G astroen ter o l 1989;3(3):98-102. Pour re.mme voir page 99

gens have been recogn ized since 1942 when Pangborn ( l ) described the isolation from beef hearr of cardiolipin (diph osphatidyl glycerol) and showed that it was the antigen active in the Wasserman reaction for syphilis.Anticardiolipin antibodies are freque n tly found in association with the circulating 'anticoagulan t' in patie n ts with systemic lupus erythematosus (SLE) (2) wh ich leads to a prolongation o f the partial thromboplastin time and which is an immunoglobu lin interferi ng with phospholipid dependen t coagulation tests ( 3 ).
A great deal of atte ntion has recently been focused on a nticardiolipin antibodies in SLE.It has been claimed that they are associated with a significantly increased risk of arterial and venous thrombosis (4.5 ).During pregnancy they appear ro be predictors of fecal distress or d ea th (6).Estima tes of the freq uency of lupus anticoagu lant in patients with S LE have varied between 6 and 65%.Harris and colleagues (7) fou nd an ticardiolipin antibod ies in 62% of patients with SLE, Anticorps diriges contre la ca rdiolipine chez les patients atteints de cirrhose biliaire primitive RESUME: Les anticorps conrre la cardio lipine ont ete identifies chez 65% des patien ts Souffra n t de lup us erythe m ateux aigu d issemine.O n a sou tenu qu 'ils etaient associes de fa~on significacive a la ch rombose in travascu laire et aux complications obstetr icales.J usqu 'a present, on semble les trouver moins freq ue mment dans d'autres malad ies.A cause de la forte concentration de cardiolipine clans les membranes mitochondriales, et de la presence d 'an ticorps an ti-m itochondriaux chez les patients atte ints de cirrhose b iliaire p rimi tive, nous avons e tudie la preva le nce d'an ticorps anti-card io li pine chez u n gro upe de 31 patients souffrant d e cette affection .D'apres les resu ltats, la prevale nce et les niveaux d'an ticorps anti-cardio lipine d'isotypie lgG et IgA sont aussi eleves chez les patients atteints de cirrhose biliaire primi tive que chez 35 patients consecutifs souffrant de lup us erythemateux aigu d isse mine.A ucu n des patients attei n ts de cirrhose biliaire primitive ne p rescn tait une histoire de thrombose vei neuse ou arterielle.Chez les patiences, le taux d 'avorteme n t e tait infe rieur a celui d e la population generale.Avoir acces, pour fin d'etude, a un second groupe presen tant des concentrations elevees d 'anticorps a n ti-cardio lip ine permettrait de determ ine r si !'associa tion de ces a n ticorps aux throm boses, et a la resorption d u foetus chez les patiences attcin tes de lu pus e rythem ateux aigu d issemi ne, consticue u ne relation directe OU Un epiphenomcne.
Rece,H studies have shown that lupus anticoagu lant and anticardiolipin an tibodies may be found in d iseases other than SLE.Lup us anticoagulant has been fou nd in 4% of young adults with cerebrovascular disease ( 10) and in 27°{,of patients with syphilis ( 11 ).It has also been reported in 3 7% of psychiatric patien ts treated wi t h p hcnothiazines ( 12), in whom it is presumably drug ind uced .Hull and co-workers ( 13) recentl y reported a study of anticardiolipin an tibodies in 70 patien ts with Behcet's syndrome and detected the an ti bodies in 13, of whom eigh t had a history of vascular pathology.Anticardiolipin antibodies have also been reported in single cases ofDegos' disease ( l4) and G uillain-Barrc syndrome ( 15).Recently, Canoso et al ( 16) reported anticardiolipin an tibodies in all of 43 patien ts with acq uired immune deficiency sy ndrome (AIDS) or AIDS-related com plex as well as five of 10 men positive for antibodies to h u man immunodeficiency virus ( HI V).None of these patien ts had a history of venous or arterial th rombosis.
Because of the h igh levels o f ca rdiolipin in mitochondria l membranes and the high frequency of annmitochondrial an tibodies in patients with prima ry biliary cirrhosis.the authors decided to study the prevalence and concen tration of an ticardiolipin antibod ies in patients with this condition .
The purpose of this paper is to report the detection of high levels of an tibodies to cardiolipin in patients with primary biliary cirrhosis.T he authors found that the prevalence of these antibodies, their isotype and their level in serum was comparable to that fo u nd in a gro u p of 35 unselected patients with S LE.An increased incidence of venous or arterial thrombosis or of fe tal wastage among the primary biliary cirrhosis patients was not fou nd.

PATIENTS AND METHODS
Thirty-one sera from the fi les of the immu nopathology laboratory at the Victoria General Hospital, Halifax, Nova Scotia, were studied.All were positive for antimi tochond rial an tibody at a d ilution of 1: 100 or greater and all patients had a diagnosis of definite or probable primary biliary cirrhosis.Twenty patients were symptomatic and ll were asymptomatic.In 21 patients the diagnosis was confirmed by liver biopsy.ln the remaining 10 liver biopsy was either refused ( rwo patients), nondiagnostic (one patient) or not performed because of elevated prochrombin a nd partial th romboplastin Cardiolipin antibodies In cirrhosis times due to end srnge liver foilun.' (seven patients).All 10 patients were diagnos('d as having p rimary biliary ci rrho~ts based on their cl inic:il course and persistent elevation of alkaline phosphatase and gammaglutamyl transforase.For comparison. the sera of 35 consecutive patients (mean age ' SD 41.9 J 11.95 years) attend ing the lu pus clinic and 35 control specimens from normal staff (m ean age 35.4 • 12 years) of the hematology laboratory at the Victoria General Hospital were examined (all ages arc given at the time serum was taken).All lupus patients fu lfilled th ree or more of the American Rheumatism Association revised criteria fo r the diagnosis of SLE ( 17 ).
The mean age of the patients with prima ry bi liary ci rrhosis was 59.6 ' 11 .3 years.There were 26 females and five males with p rimary biliary cirrhosi s, and 32 females and three males in the lupus group.T he patients with primary biliary cirrhosis were interviewed by telephone.T he fema le patients were questioned as to the n umber of pregnancies, miscarriages and stillbirths.A ll patients were asked a series of questions designed to elici t a h istory of vascu lar thrombosis o r embolism.Specifically, each patient was asked whether he or she had ever suffered an episode of venous or arterial occl usion requiring medical or surgical treatment.All of the patients' charts were also reviewed to confirm the results of the telephone in terview.None of the patients was receiving phenochiazines at the time that blood was taken.None was suffering from syph ilis or HIV infection.A ntibodies to cardio lip in: All sera were frozen immediately after being taken and stored in aliquots at 70°C.Each sample tested was thawed o nce only hefore meas uremen r of antihody.Anticardiolipin antibodies were measured by an amplified ELISA assay as described by Carr ct al ( 18).In brief.Linbro 96 we ll micro ti tre plates were coated with 45 µg/mL cardiolipin (Sigma C 1649) in 95't ethanol.The solution was evaporated under a stream of nitrogen and the places postcoatcd with 3 mg/ml gelatin in phosphate buffered saline (PBS) for 2 h at room temperatu re.After washi ng the plates with PBS.test sera dil u ted l/ 100 in gelatin/PBS were added to the wells and incubated for 90 mi ns at room  Figure 3) lgA anticardwl1pm as measured by optical density ( 490 nm) in controls and patients with primary biliary cirrhosis or SLE.The horizomal line represents the mean plus three standard devia-11ons J or 1he normal rnntroh 100 temperature.Plates were washed.and alkaline phosphatase-conjugated heavy chain specific antiscra to human lgG, lgM o r lgA (Sigma, St Louis, Missouri) were added.The plates were then incu bated for 1 hat room temperature.After washing, the primary a nd secondary substrate of the amplified system were added .
Absorbance was read after 10 mins at 490 nm on an automated EL 310 ELISA reader (Biotek, Cambridge, Massachussetts).Sera were regarded as positive if the test gave an optical density value of more than three standard deviations above the mean for normals.Uncoated wells were used as controls to detect nonspecific binding, this was then subtracted from the optical density value Antimitochondrial antibodies: Anrimitochondrial antibodies were assayed by an indirect immunofluorescence method using cryostat sections of mouse kidney for screen ing and subseq uently using h uman kidney for titration.Bound antibodies were iden tified using FITC labelled antihuman immunoglobulin .Statistical analysis: Fisher's exact test (two-tailed) was applied to compare the number of positives in each group.

RESULTS
Figu res I , 2 and 3 show the levels of anticardiolipin antibodies of each isotypc detected in the three sets of sera.
For all isorypes of an ti body.the levels of anticardiolipin were as high in the primary biliary cirrhosis group as they arc in the SLE group.Table I shows the analysis of antibody isoty pe in each of the groups studied.For anticardiolipin antibody o f IgG and lgA isorype in the primary biliary cirrhosis group, the number of positive patients was signifi cantl y higher than normals but not statistically d ifferent fro m the SLE group.
When the 31 patients with primary biliary cirrhosis were divided into the 21 in whom the diagnosis was confirmed by liver biopsy and the 10 in whom the diagnosis was established clinically, it was found that six o f the former group and fo ur of the latter group were positive for lgG antibodies to ca rdiolipin.
There was no correlation between antibodies to double stranded and single stranded DNA and the presence of anticard iolipin antibody in the primary bi!-

20
iary cirrhosis group (data not shown).Of the patients with SLE, four were positive for an timitochond rial antibodies.Two of these were also positive for anticardiolipin, while two were negative.
l,1unoz ct al ( 19) have suggested that in patients with primary biliary cirrhosis, lgM may bind nonspecifically in the ELISA assay for anticardiolipin antibodies.In the group of patients with primary biliary cirrhosis three of 16 showed high nonspecific binding for lgM and have been categorized as antibody negative.
There was no nonspecific binding with the SLE group.
Of the 31 primary biliary cirrhosis patients, 23 were reached by telephone.
None of these patients gave a history of venous or arterial thrombosis.Of the eight patients ( two m ale, six female) not contacted, th ree had died of e nd stage liver failure and five were alive but lost to follow-up.Of a total of six females dead or lost to follow-up, three were positive for anticardiolipin antibodies while three were negative.
The charts of all patients lost to followup were reviewed.Of the six females lost to follow-up, obstetrical histories were missing from all hut o ne chart.This patient had eight normal deliveries and no abortions.One 73-vcar-old male was found to have had a bypass graft of the left femoral artery for occlusion after a Twenry-one of the 23 patients interviewed by telephone were fema le.Among the seven patients with primary biliary cirrhosis who were anticardiolipin antibody positive there were 1 9 pregnancies with only one abortion.in the 11 a nticardiolipin antibody negative patients, three h ad one spon caneous abortion while one patient had two.There was no case of fetal loss after 20 weeks.
Thus, in the primary biliary cirrhosis group, anticardiolipin antibodies were not associated with an increased risk of fetal loss

DISCUSSION
There has been a major upsurge of interest in the antiphospholipid antibodies since the demonstration that they may be predictors of intravascular thrombosis ( 5) and obstetrical complications (6).The mechanism by which these adverse effects arc produced is still obscure.Suggested mechanisms include blocking the release of arachidonic acid from the membrane of the e ndothelial cells by binding to p hospholipids (20) and decrease in the release of plasminogen activator following venous occl usion (21 ).
The mechanism of the obstetrical complications is also uncertain, although it Cordiolipin antibodies In cirrhosis has been reported that the lu pus anticoagulant may affect the p lacental circulation by inhibiting the production of prostacyclin hy the myometrium (22).
Studies of the pathology of placental tis• sue associated with fetal loss in patients with SLE have shown that the placentae are small, often with multiple infarcts secondary to thrombosis (23 ).
In relation to both the thrombotic and the obstetric complications associated with antiphospholipid antibodies it 1s still uncertain "whether an tiphospholip1d s play a pathogenic role .or arc merely 'markers' for ... disease" ( 4 ).SLE i:, characterized by the production oflarge numbers of autoantibo<lies and by excessive B-cell proliferation (24).While antiphospholipid antibod ies appear to be risk factors for vascu lar and obstetric complications, some of the many oth er immunologic abnormalities found in SLE could be equally important.
The authors have shown that antibodies to cardiolipin occu r as frequently in a group of patients with primary biliary cirrhosis as in S LE and that the level o( antibody is similar in both groups.The authors have found no increase in the freq uency of venous or arterial thrombosis.The authors found no increase in fetal wastage in the an ticard iolipm positi ve patients, compared to those who were anticardiolipin negative, although It was recognized that the higher mean age of the primary biliary cirrhosis group makes this observation of quesuonahle significance.It has been shown that the prevalence of ant1card1oltpin antibodies of [gG 1sotypc incre;iscs in an ageing population (mean age 81) (25).However. the present patients with primary biliary cirrhosis, although somewhat older than the SLE group (mean age 59.6 years).showed anticardiolipin antibodies in all three immunoglobulin isotypes.
While it seems probable that the presence of anticardiolipin an ti bodies ts a predictor for thrombosis and fecal wastage in some patients with SLE, and perhaps in some individuals with no other identifiable disease, the present study indicates that patients with primary biliary cirrhosis may have an equally h igh incidence and concentration of ancicar-diol1pin antibodies without sustaining any of these complicattons.

TABLE 1 lsotypes
of anticardiolipin antibody: Number of patients with optical density greater than three standard deviations above the normal mean