British Journal of Haematology, 2001, 113, 911±914

Anti-beta 2 glycoprotein 1 and anti-annexin V antibodies in

women with recurrent miscarriage

J. Arnold,1 Z. Holmes,1 W. Pickering,1 C. Farmer,1 L. Regan2 and H. Cohen3

1Department of Haematology, St Mary's Hospital, 2Department of Obstetrics and Gynaecology,

Imperial College School of Medicine, St Mary's Campus, and 3Department of Haematology, University College London

Hospitals NHS Trust, London, UK

Received 24 July 2000; accepted for publication 30 January 2001

Summary. While it has been established that anti-phospho-lipid antibodies (aPL) are associated with recurrent mis-carriage (RM), the importance of anti-b2 glycoprotein 1(GP1) IgG and anti-annexin V IgG antibodies as risk factorsfor RM is undefined. We have investigated the prevalence ofanti-b2 GP1 IgG and anti-annexin V IgG antibodies in 54aPL-positive and 48 aPL-negative women with RM. Theprevalence of IgG anti-b2 GP1 antibodies was not sig-nificantly different in persistently aPL-positive women withRM (7%), aPL-negative women with RM (6%) and thenormal parous control group (3%). Anti-annexin V IgGantibody prevalence was significantly increased in aPL-positive women with RM compared with aPL-negative

women with RM (P � 0´01). The elevations were found in35%, 19% and 16% of aPL-positive women with RM, aPL-negative women with RM and the control group respec-tively. No women showed positivity for both anti-b2 GP1IgG and anti-annexin V antibodies. Anti-b2 GP1 IgGantibodies do not appear to be contributory to theinvestigation of women with RM. Anti-annexin V antibodypositivity, although associated with aPL positivity in womenwith RM, is not an independent risk marker.

Keywords: anti-b2 glycoprotein 1 antibodies, anti-annexinV antibodies, recurrent miscarriage.

Recurrent miscarriage (RM), which occurs in 1% ofpregnancies, is a cardinal feature of anti-phospholipidsyndrome (APS) with additional features including arterialand venous thromboembolism. Using a comprehensivemethodological approach with testing for both lupus anti-coagulant (LA), and IgG and IgM anti-cardiolipin antibodies(aCL), we have reported an increased prevalence ofpersistent anti-phospholipid antibodies (aPL) of 15% inwomen with RM (Rai et al, 1995a). The prospective fetal lossrate in women with aPL is about 90% (Rai et al, 1995b). Ina randomized controlled trial we have shown a significantreduction in fetal loss, with a live birth rate of over 40% inwomen with RM associated with aPL after treatment withlow-dose aspirin. This was further increased to 70% withthe combination of low-dose aspirin and heparin (Rai et al,1997).

The mechanism of fetal loss in patients with RM and aPLis not clear, but in some patients with aPL appears to bethrombosis in the uteroplacental vasculature (Rushton,

1988; Out et al, 1991). However, other mechanisms mayplay a role. Anti-beta 2 glycoprotein 1 (b2 GP1) antibodiesappear to be a thrombotic marker in individuals with APS(Martinuzzo et al, 1995) and have been shown to inhibittrophoblast proliferation (Chamley et al, 1998). Expressionof b2 GP1 by the syncytiotrophoblasts and connective tissueof fetal villi has been observed from 7-weeks gestation, andreduced expression of b2 GP1 in human syncytiotropho-blasts has been demonstrated in women with APS (Donohoeet al, 1998). Therefore, it is possible that anti-b2 GP1antibodies are aetiologically involved in RM by predisposingto uteroplacental thrombosis and, possibly, abnormalplacentation. Anti-annexin V antibodies have been demon-strated to mediate thrombosis in human umbilical veinendothelial cell (HUVEC) systems (Rand et al, 1997).Concentrations of anti-annexin V antibodies have alsobeen observed to be reduced in isolated placentae fromwomen with aPL-associated RM, and IgG fractions from aPLpatients have been shown to reduce the quantity of annexinV in cultured trophoblasts and endothelial cells (Rand et al,1994). Annexin V has been shown to be involved inexocytosis and syncytiotrophoblast membrane fusion(Burgoyne & Clague, 1994). Therefore, antibodies directed

q 2001 Blackwell Science Ltd 911

Correspondence: Dr Hannah Cohen, Department of Haematology,University College London Hospitals NHS Trust, Grafton Way,

London WC1E 6DB, UK. E-mail: hannah.cohen@uclh.org

against annexin V may interfere with syncytiotrophoblastfunction with resultant alteration in expression of anioniclipids on the plasma membrane. These antibodies are alsoknown to induce apoptosis in human umbilical veinendothelial cells (Nakamura et al, 1994). Thus, the locationand function of annexin V make antibodies directed againstit potentially pathophysiologically relevant to RM bythrombotic mechanisms, altered placentation and inductionof apoptosis.

The importance of anti-b2 GP1 IgG and anti-annexin VIgG antibodies as risk factors for RM is undefined. The aimof this study was therefore to assess anti-annexin V andanti-b2 GP1 IgG antibody status in a cohort of aPL-positiveand aPL-negative women with RM to assess the predictivevalue of these potentially aetiologically relevant antibodiesin RM.

PATIENTS AND METHODS

Study population. The study population comprised 257patients with RM, defined as three or more consecutivepregnancy losses. The median age of the study populationwas 35 (range 21±47) years and the median number ofmiscarriages was four (range 3±13). These patients werefound to have no chromosomal, anatomical or endocrinecause for miscarriage. Of this population, 152 (74´7%)patients had experienced first trimester miscarriage only, 65(25´3%) second trimester miscarriages and 60 (23´3%) bothfirst and second trimester miscarriages. All were tested foraCL (IgG and IgM isotypes) and LA. Anti-b2 GP1 and anti-annexin V antibodies were measured in the women found tohave persistently positive aPL, defined as the presence of LAand/or aCL on at least two occasions, 8 weeks apart, and ina randomly selected cohort of aPL-negative women with RMfrom the overall population.

Control group. IgG anti-b2 GP1 and anti-annexin Vantibodies were measured in 33 and 25 healthy parouswomen respectively. These women had no history ofrecurrent miscarriage and had a median age of 35 (range26±47) years.

Sample collection. In accordance with British Committeefor Standards in Haematology (BCSH) guidelines, venousblood samples were collected with minimal venous stasisusing a 19 gauge butterfly needle (Lupus AnticoagulantWorking Party on behalf of the BCSH Haemostasis andThrombosis Taskforce, 1991). Samples were taken into plaintubes for IgG anti-b2 GP1, anti-annexin V antibodies and aCL(IgG and IgM isotypes), and into 0´109 mol/l trisodiumcitrate, one part anticoagulant to nine parts blood, for LAtesting (Lupus Anticoagulant Working Party on behalf of theBCSH Haemostasis and Thrombosis Taskforce, 1991; Kha-mashta & Hughes, 1993). Sera for IgG anti-b2 GP1 andannexin V antibodies, and aCL, were prepared by singlecentrifugation at 2700 g for 20 mins at room temperature.Platelet-poor plasma (PPP) for LA testing was prepared bydouble centrifugation at 2700 g for 20 min at roomtemperature. Samples were stored at 2708C until assay.

Anti-b2 glycoprotein 1 IgG antibodies. Serum anti-b2 GP1IgG antibody levels were measured using a quantitative

indirect enzyme linked immunosorbent assay (ELISA) usingthe Diastat kit (Shield Diagnostics, Dundee) (McNally et al,1995; Brookfield et al, 1998). All samples were run induplicate. The mean absorbance value of each standard wascalculated and plotted against log (10) standard concentra-tion so that a quantitative estimation of the test samplescould be made. Based on samples from 150 apparentlyhealthy Caucasian volunteers, results , 10 U/ml wereinterpreted as negative and results $ 10 U/ml as positive;values in the range 10±15 U/ml were classified as border-line positive.

Anti-annexin V IgG antibodies. Serum anti-annexin V IgGantibodies were measured using a qualitative ELISA assaybased on a method by Matsuda et al (1994a) Gamma-irradiated microtitre plates were coated with recombinantannexin. Alkaline phosphatase-conjugated goat anti-human IgG was added after the addition of diluted serumwith p-nitrophenylphosphate as substrate. The opticaldensity (OD) at 405 nm was calculated with respect to ablank. All samples were run in duplicate and meanabsorbance values calculated. OD values for anti-annexinV antibodies in both RM patients and the control groupshowed skewed distributions. Anti-annexin V antibodypositivity was arbitarily defined as an OD value above thecontrol group third quartile.

Anti-cardiolipin antibodies. Serum samples were tested forboth immunoglobulin G (IgG) and immunoglobulin M (IgM)aCL as previously described using a standardized ELISA (Raiet al, 1995a) All samples were run in duplicate. Results wereexpressed in either IgG phospholipid units (GPLU) or IgMphospholipid units (MPLU) for IgG and IgM respectively.These are arbitrary units derived from the activity ofaffinity-purified sera. Positive results were defined as $ 5GPLU for IgG and $ 3 MPLU for IgM (Khamashta & Hughes,1993).

Lupus anticoagulant. The dilute Russell's Viper Venom Test(dRVVT) test for the presence of LA was performed on PPPsamples in duplicate on a KC4 (Heinrich Amelung,Lehbrinksweg, Lengo, Germany) using a kit (UnicornDiagnostics, London, UK). Samples with a dRVVT ratio of> 1:1 were retested using a platelet-neutralization proce-dure using washed, freeze-thawed platelets. A decrease of$ 10% in the ratio was considered to be positive for LA(Lupus anticoagulant Working Party on behalf of theHaemostasis and Thrombosis Taskforce, 1991).

Statistical analysis. A chi-squared test was performed,comparing RM history in women in the presence andabsence of both anti-b2 GP1 and anti-annexin V antibodies.Chi-squared analyses were also performed comparing theprevalence of anti-b2 GP1 and anti-annexin V antibodies inaPL-positive and aPL-negative women, in aPL-positivewomen with RM and the healthy parous control group,and in aPL-negative women with RM and the healthyparous control group.

RESULTS

Twenty-four women were excluded owing to either lack of arepeat sample for confirmation or an interval of , 8 weeks

912 J. Arnold et al

q 2001 Blackwell Science Ltd, British Journal of Haematology 113: 911±914

between testing, reducing the cohort size to 233. Theoverall prevalence of aPL positivity was therefore 58/233(24´9%) in our cohort of women with RM. Nineteen out of233 (8´2%) patients were persistently IgG aCL positive onrepeat testing and 42/233 (18´0%) were persistently IgMaCL positive. Six out of 233 (2´6%) patients had persistentlypositive IgG and IgM aCL levels. The prevalence of persistentLA was 2´1% (5/233). Of these five LA-positive women, twowere positive for IgM aCL, one of whom also had IgG aCL.Fifty-four out of 58 samples from aPL-positive women withRM and 48 aPL-negative samples, randomly selected fromthe remaining patient cohort, were investigated for anti-b2GP1 IgG and anti-annexin V antibodies to assess thedependence of anti-b2 GP1 and anti-annexin V antibodieson aPL status.

The prevalence of anti-b2 GP1 IgG antibodies was notsignificantly different between the three study groups: 4/54(7%) in aPL-positive women with RM, 3/48 (6%) in aPL-negative women and 1/33 (3%) in the normal parouscontrol group.

Anti-annexin V IgG antibodies were found to be elevatedin 19/54 (35%), 9/48 (19%) and 4/25 (16%) of aPL-positive women with RM, aPL-negative women with RMand the control group respectively. There was no significantdifference in anti-annexin V antibody prevalence betweenaPL-negative women with RM and the control group(P � 0´77). Anti-annexin V antibody prevalence wassignificantly increased in APL-positive women with RMcompared with aPL-negative women with RM (P � 0´01),but the increase in anti-annexin V antibody prevalence inaPL-positive women with RM compared with the normalparous control group did not attain statistical significance(P � 0´08). Fourteen of the 19 aPL-positive women withanti-annexin V antibodies had IgM aCL, six of whom alsohad IgG aCL, two had IgG aCL alone and two were positivefor LA.

There were no differences with respect to age, numberand gestation of miscarriages between anti-b2 GP1 andanti-annexin V antibody-positive and -negative patients. Nowomen showed positivity for both anti-b2 GP1 and anti-annexin V antibodies.

DISCUSSION

The prevalence of anti-b2 GP1 IgG antibodies was notsignificantly elevated in either aPL-positive or -negativewomen with RM compared with the normal parous controlgroup. Anti-b2 GP1 IgG antibodies do not therefore appearto be contributory to the investigation of women with RM.The prevalence of anti-annexin V IgG antibodies was notincreased in aPL-negative women with RM, suggesting thatthese antibodies do not constitute an independent risk factorfor RM. However, there was an increased prevalence of anti-annexin V IgG antibodies in persistently aPL-positivewomen compared with the aPL-negative control group,which may be of prognostic significance.

Our finding of no significant increase in prevalence ofanti-b2 GP1 antibodies in women with RM is in agreementwith Ailus et al (1996). They found no significant increase

in anti-b2 GP1 levels in women with RM, defined as three ormore consecutive pregnancy losses, compared with womenin the control group. Forastiero et al (1997) also showed nocorrelation between IgG anti-b2 GP1 antibodies in aPL-positive patients and obstetric complications (two or morespontaneous fetal abortions or one fetal death), although anassociation was found with IgM anti-b2 GP1 antibodies.This latter finding was also noted by Falcon et al (1997). Incontrast Ogasawara et al (1996) reported a statisticallysignificant increased prevalence of anti-b2 GP1 IgG anti-bodies of 9% in women with two ore more miscarriages.

The role of annexin V antibodies in RM has yet to beestablished. Anti-annexin V antibodies have been shown tobe correlated with thrombosis and intrauterine fetal loss inpatients with systemic lupus erythematosus (Kaburaki et al,1997). Anti-annexin V antibody levels have also been foundto be increased in four of a small cohort of 11 women (36%)with habitual fetal loss (Matsuda et al, 1994b).

REFERENCES

Ailus, K., Tulppala, M., Palosuo, T., Ylikorkala, O. & Vaarala, O.(1996) Antibodies to beta 2 glycoprotein 1 and prothrombin in

habitual abortion. Fertlity and Sterility, 66, 937±941.

Brookfield, C., Malia, R.G. & Hampton, K.K. (1998) The incidence

and significance of anti-b2 glycoprotein 1 (b2, GP1) anti-

phospholipid antibodies. British Journal of Haematology, 101, 80.

Burgoyne, R.D. & Clague, M.J. (1994) Annexins in the endocytic

pathway. Trends in Biochemistry and Science, 19, 231±232.

Chamley, L.W., Duncalf, A.M., Mitchell, M.D. & Johnson, P.M.

(1998) Action of anticardiolipin antibodies to beta 2 glycoprotein1 on trophoblast proliferation as a mechanism for fetal death.

Lancet, 352, 1037±1038.

Donohoe, S., Kingdom, J., Burrell, S., Quenby, S., Jauniaux, E.,

Mackie, I. & Machin, S. (1998) Beta 2 glycoprotein 1 and

annexin V in villous placenta during normal, intra-uterinegrowth retardation and anti±phospholipid syndrome pregnan-

cies. Lupus, 7, B004.

Falcon, C.R., Martinuzzo, M.E., Forastiero, R.R., Cerrato, G.S. &

Carreras, L.O. (1997) Pregnancy loss and autoantibodies against

phospholipidbinding proteins. Obstetrics and Gynaecology, 89,

975±980.

Forastiero, R.R., Martinuzzo, M.E., Cerrato, G.S., Kordich, L.C. &

Carreras, L.O. (1997) Relationship of anti-b2 glycoprotein 1 andanti prothrombin antibodies to thrombosis and pregnancy loss in

patients with antiphospholipid antibodies. Thrombosis and Hae-

mostasis, 78, 1008±1014.

Kaburaki, J., Kuwana, M., Yamamoto, M., Kawai, S. & Ikeda, Y.

(1997) Clinical significance of anti-annexin V antibodies inpatients with systemic lupus erythematosus. American Journal of

Hematology, 54, 209±213.

Khamashta, M.A. & Hughes, G.R. (1993) ACP Broadsheet no. 136:

detection and importance of anticardiolipin antibodies. Journal of

Clinical Pathology, 46, 104±107.

Lupus anticoagulant Working Party on behalf of the BCSH

Haemostasis and Thrombosis Taskforce (1991) Guidelines ontesting for the lupus anticoagulant. Journal of Clinical Pathology,

44, 885±889.

Martinuzzo, M.E., Forastiero, R.R. & Carreras, L.O. (1995) Anti-b2

glycoprotein 1 antibodies: detection and association with

thrombosis. British Journal of Haematology, 89, 397±402.

Matsuda, J., Saitoh, N., Gohchi, K., Gotoh, M. & Tsukamoto, M.

q 2001 Blackwell Science Ltd, British Journal of Haematology 113: 911±914

Anti-b2 GP1 Antibodies and Anti-Annexin V Antibodies 913

(1994a) Anti-annexin antibody V antibody in systemic lupus

erythematosus patients with lupus anticoagulant and/or antic-

ardiolipin antibody. American Journal of Hematology, 47,56±58.Matsuda, J., Gotoh, M., Saitoh, N., Gohchi, K., Tsukamoto, M. &

Yamamoto, T. (1994b) Anti-annexin antibody in the sera of

patients with habitual fetal loss or preeclampsia. ThrombosisResearch, 75, 105±106.

McNally, T., Purdy, G., Mackie, I.J., Machin, S.J. & Isenberg, D.A.

(1995) The use of an anti-b2 glycoprotein 1 assay for

discrimination between anticardiolipin antibodies associatedwith infection and increased risk of thrombosis. British Journal

of Haematology, 91, 471±473.

Nakamura, N., Shidara, Y., Kawaguchi, N., Azuma, C., Mitsuda, N.,

Onishi, S., Yamaji, K. & Wada, Y. (1994) Lupus anticoagulantautoantibody induces apoptosis in umbilical vein endothelial

cells: involvement of annexin V. Biochemistry and Biophysics

Research Communications, 205, 1488±1493.Ogasawara, M., Aoki, K., Matsuura, E., Sasa, H. & Yagami, Y.

(1996) Anti-beta 2 glycoprotein 1 antibodies and lupus antic-

oagulant in patients with recurrent pregnancy loss: prevalence

and clinical significance. Lupus, 5, 587±592.Out, H.J., Kooijman, C.D., Bruinse, H.W. & Derksen, R.H. (1991)

Histopathological findings in placentae from patients with

intrauterine fetal death and antiphospholipid antibodies. Eur-

opean Journal of Obstetrics, Gynaecology and Reproductive Biology, 4,179±186.

Rai, R.S., Regan, L., Clifford, K., Pickering, W., Dave, M., Mackie, I.,

McNally, T. & Cohen, H. (1995a) Antiphospholipid antibodies and

b2 glycoprotein 1 in 500 women with recurrent miscarriage:

results of a comprehensive screening approach. Human Reproduc-tion, 10, 2001±2005.

Rai, R.S., Clifford, K., Cohen, H. & Regan, L. (1995b) High

prospective fetal loss rates in untreated pregnancies of women

with recurrent miscarriage and antiphospholipid antibodies.Human Reproduction, 10, 3301±3304.

Rai, R.S., Cohen, H., Dave, M. & Regan, L. (1997) Randomised

controlled trial of aspirin plus heparin in pregnant women with

recurrent miscarriage associated with phospholipid antibodies.British Medical Journal, 314, 253±257.

Rand, J.H., Wu, X., Guller, S., Gil, J., Scher, J. & Lockwood, C.J.

(1994) Reduction of annexin V (placental anticoagulant protein-

1) on placental villi of women with antiphospholipid antibodiesand recurrent spontaneous abortion. American Journal of Obste-

trics and Gynecology, 171, 1566±1572.

Rand, J.H., Wu, X., Guller, S., Scher, J., Andree, H.A.M. & Lockwood,

C.J. (1997) Antiphospholipid immunoglobulin G antibodies

reduce annexin V levels on syncitiotrophoblast apical membranesand in culture media of placental villi. American Journal of

Obstetrics and Gynecology, 177, 918.

Rushton, D.I. (1998) Placental pathology in spontaneous mis-

carriage. In: Early Pregnancy Loss: Mechanisms and Treatment (ed.by R.W Beard, F. Sharp), pp. 149±158. Royal College of Obstetrics

and Gynaecology, London.

914 J. Arnold et al

q 2001 Blackwell Science Ltd, British Journal of Haematology 113: 911±914