Prevalence and heterogeneity of antiphosphatidylethanolamine antibodies in patients with recurrent early pregnancy losses
FERTILITY AND STERILITY VOL.71,NO.6, JUNE 1999

Toshitaka Sugi, M.D., Ph.D. Junko Katsunuma, M.D. Shun-ichiro Izumi, M.D., Ph.D.
John A. McIntyre, Ph.D.and Tsunehisa Makino, M.D., Ph.D.

Tokai University School of Medicine, Knagawa, Japan and Methodist Hospital of Indiana, Indianapolis, Indiana.

Antiphospholipid antibodies (aPAs) to anionic phospholipids such as cardiolipin and phosphatidylserinehave been described in patients with thrombosis, thrombocytopenia and recurrent fetal loss (1-3). Similar but fewer reports have focused on autoantibodies to the zwitterionic phospholipid, phosphatidylethanolamine (PE) (4-8). Because PE is a major component of both the outer and inner leaflets of cell plasma membranes, the production of autoantibodies to PE should not be viewed as without consequence.

Recent evidence shows that many aPAs to negatively charged phospholipids do not target anionic phospholipids per se but are specific for anionic phospholipid-binding plasma proteins. At present, the most common and best characterized plasma protein aPAs antigenic targets are beta₂glycoprotein I (beta₂GPI) and prothrombin (9). We recently reported that certain antiphosphatidylethanolamine antibodies (aPEs) are not specific for PE per se, but are directed to PE-binding plasma proteins, such as high molecular weight kininogen (HK), low molecular weight kininogen (LK) and proteins in complex with HK, factor XI, or prekallikrein (8). We also demonstrated that kininogen-dependent aPEs can augment thrombin-induced platelet aggregation in vitro (10,11).

Reccurenr pregnancy losses, occuring mostly in the second and third trimesters of pregnancy in aPAs positive mothers, is one of the hallmark clinical manifestations of the antiphospholipid syndrome (12-14). Reccurenr pregnancy losses in the first trimester fetal period (>10weeks'gestation) also has been associated with antiphospholipid syndorome. (15). However, the association between aPAs and recurrent embryonic losses (< 10weeks'gestation) are not well documented (15). In fact, several studies have questioned whether an association exists between aPAs to negatively charged phospholipids and early RPL (15-17).

The female reproductive tract is the second richest site for kininogen and its metabolic products in the body (18-21). Adam et al. (21) measured 12.2, 10.9, 0.4, and 1.2 micro-g/mg of T-kininogen in rat plasma, uterus, liver, and kidney, respectively. The kininogen concentration in reproductive tissues and plasma was reported to fluctuate during ovulation, pregnancy, and parturition (18,21). Why the female reproductive system is so rich in kininogen and what governs the fluctuation of kininogen concentrations at the local level remains tobe elucidated.

Because numerous studies (12-14) have concluded that RPLs are associated with aPAs to anionic phospholipids, and in view of the conspicuous presence of kininogen in reproductive tissues, we tested patients with RPLs for aPEs, especially those patients with RPLs during the embryonic period. We showed a strong association between RPLs and aPEs, the latter of which requires the presence of kininogen or other plasma proteins. Our data suggest that aPEs may represent a significant risk factor for early RPLs.

MATERIALS AND METHODS

Patients and controls
Plasma samples were obtained from 139 non-pregnant patients with a history of RPLs during the embryonic period (15). The patients met our study entry criteria, which were as follows:
1) two or more pregnancy losses before 10weeks'gestation, exclusive of ectopic pregnancy and elective abortion;
2)no presumptive cause found for RPLs after routine evaluation for detection of uterine factors (i.e., normal hysterosalpingography and ultrasound examinations);
3) absence of chromosomal abnormality;
4) absence of endocrine factors (normal prolactin and progesterone levels and normal thyroid function);
5) absence of infectious factors (no group B streptococcal or chlamydia trachomatis infection); and
6) absence of diabetes mellitus.
The mean age of the patients was 31years (range, 22-45years), and the mean number of pregnancy losses was 2.8. All plasma samples were stored at -70degrees C until use.

Two hundred age-matched healthy nonpregnant female volunteers with no previous history of miscarriage were tested as controls. Positive aPAs samples from patients with systemic lupus erythematosus were used at 1:100 dilutions to establish standard curves.

This study was approved by the Institutional Review Board of the Tokai University School of Medicine.

RESULTS

Detection of aPEs in Patients with RPLs
Patients with RPL (n=139) were screened for phospholipid-binding plasma protein-dependent vs. -independent IgG aPEs with the ose of 10% adult bovine plasma vs. 1% bovine serum albumin in the patient plasma diluent. As shown in Table 1, 28 (20.1%) of the patients with RPLs were positive for IgG aPEs. A positive test result for IgG aPEs was more frequent in the patients with RPLs than in the control group (p=0.001). Patients with RPLs also were screened for phospholipid -binding plasma protein-dependent vs. -independent IgM and IgA aPEs. Seventeen (12.2%) of the139 patients with RPLs were positive for IgM aPEs (Table 1). Two (1.4%) of 139 patients wiith RPLs had IgA aPEs (Table 1). A positive test result for IgM or IgA aPEs was not statistically more frequent in patients with RPLs than in members of the reference group. Three patients with aPEs had two isotypes: one patient had IgG and IgM aPEs, and 2 patients had IgG and IgA aPEs. No patient had three aPA isotypes. To summarize, 44(31.7%) of the 139 patients with RPLs were positive for aPEs. A positive test result for aPEs was more frequent in the patients with RPL than in the members of the control group (p=0.0002).

Anticardiolipin Antibodies, Antiphosphatidylserine Antibodies and LupusAnticoagulant in Patients with RPLs.
The same 139 patients with RPLs also were tested for anticardiolipin antibodies, antiphosphatidylserine antibodies and lupus anticoagulant. As shown in Table2, 6 patients (4.3%) and 1 patient (0.7%) were positive for IgG and IgM antiphosphatidylserine antibodies, respectively. Seven patients (5.0%) were positive for beta₂GPI-independent IgG anticardiolipin antibodies and one patient (0.7%) was positive for beta₂GPI-dependent IgG anticardiolipin antibodies. No patient had IgA antiphosphatidylserine antibodies. Five of the 7 patients who were positive for antiphosphatidylserine antibodies also were positive for aPEs.

Two patients (1.4%) had lupus anticoagulant detected by dilute Russell's viper venom time. No patientwas positive for two or more antibodies to phosphatidylserine, cardiolipin or lupus anticoagulant. There was no statistically significant difference in the incidence of positive test result for antiphosphatidylserine antibodies and/or lupus anticoaglant between the RPL group and the control group.

Antinuclear Antibodies in Patients With RPLs.
Antinuclear antibodies in the 100 healthy women, 139 patients with RPLs and the 28 patients with RPLs who were positive for IgG aPEs were measured. In control group, 13 patients (13%) were positive for antinuclear antibodies. Among the patients with RPLs, 31 patients (22.3%) were positive for antinuclear antibodies. Among the patients with RPLs who were positive for IgG aPEs, 10 (35.7%) were positive for both IgG aPEs and antinuclear antibodies. The incidence of the antinuclear antibodies in the last two groups were higher than that in the control group (p=0.02).

Kininogen-Dependence of aPEs.
The requirement of kininogens for plasma protein-dependent IgG aPEs detection was determined with the use of partially purified kininogens as the patient plasma diluent. In this study, 21 patients were positive for plasma protein-dependent IgG aPEs. Nineteen (90.5%) of these 21 patients were kininogen-dependent and 2 (9.5%) were kininogen-independent. The kininogen-independent IgG aPEs must recognize other
PE-binding plasma proteins, because they were not positive when bovine serum aibumin was used as the patient plasma diluent but were positive when adult bovine plasma was used. The plasma protein involved in these two aPEs positive sera will be the subject of a separate study.

DISCUSSION
Associations have been reported between aPAs, mainly anticardiolipin antibodies and/or the lupusanticoagulant, and RPLs(1-3). Relatively few studies describing aPEs have been published (4-8). We recently reported that certain aPEs are not specific for PE per se but are directed to PE-binding plasmaproteins, such as HK, LK, and the HK-binding plasma proteins factor XI and prekallikrein (8). Because kininogens appear in mammalian reproductive tissues, we screened patients with RPLs for kininogen-dependent aPEs by ELISA using adult bovine plasma and/or kininogens partially purified from adult bovine plasma. We now report a stronger association between RPLs, and aPE than between RPL and aPA to the anionic phospholipids during early gestational losses.

Many patients with RPLs who fit a clinical profile compatible with antiphospholipid syndrome are negative for aPAs when tested by ELISA. Explanations for the negative findings are manyfold, but often can be attributed to assay variations. For example, some commercial aPAs detection kits are designed to use purified or recombinant beta₂GPI as the patient sample diluent. In this situation, patients with aPAs that
are dependent on the presence of prothrombin would appear as false-negative results. Other kits combine anionic and zwitterioni phospholipids in the ELISA plates. This can lead to false-negative findings because of the differential affinities of phospholipids-binding plasma proteins for the respective phospholipids. Moreover, binding to a mixture of phospholipids may result in false-negatives results because of steric hinderance or a change in the physical phospholipid separation in the mixtures that alters their protein-binding properties.

False-negative results in the aPEs ELISA also can result from the use of a patient sample diluent that is low or deficient in the kininogens; fetal calf serum and newborn calf serum contain notably low concentrations of HK and LK (28,29) and should be avoided. A few studies implicated aPEs in patients with RPLs (30-32). In these studies, fetal calf serum or newborn calf serum was used as the patient sample diluents. This
suggests that many kininogen-dependent aPEs may not detected in these studies. Finally, prolonged storage of adult bovine serum and adult bovine plasma at 4degrees C can result in decreased kininogen activity, ( Sugi. T, unpublished observations).

Midgestation pregnancy losses in patients with the primary antiphospholipid syndrome often are associated with anticardiolipin antibodies (12-14). This may reflect appearance more than reality, because most published studies have limited their analyses to cardiolipin, and to IgG and IgM isotypes. It may be an oversimplification to assume that mid gestation to late gestation pregnancy losses in these patients are mediated by placental thrombosis; although placental thrombosis can occur, it ioften is insufficient to explain the pregnancy loss (33).Nevertheless, low dose aspirin and/or subcutaneous heparin therapy often are effective and result in successful subsequent pregnancies (34).

Recurrent pregnancy losses in the first trimester also may be associated with aPAs; however, a relatively low incidence of anticardiolipin antibodies and/or lupus anticoagulant has been reported. Gris et al. (35) reported that the prevalence of anticardiolipin antibodies and lupus anticoagulant in patients with RPLs before the end of the 16th week of amenorrhea were 2.2% and 4%, respectively. Ozawa et al. (24) reported that the prevalence of beta₂GPI-dependent and -independent IgG anticardiolipin antibodies in patients with RPLs in the first trimester was 1.1% and 4.3%, respectively. Data from these investigators are similar to our RPL data. When we could find no association between anticardiolipin antibodies or lupus anticoagulant and early pregnancy loss, we extended our studies to include aPEs and discovered a significant number of positive patients.

Because antinuclear antibodies have been accepted as serologic markers for the presence of autoimmune disease, patients with RPLs were screened for them. Many patients with RPLs who were positive for IgG aPEs also were positive for antinuclear antibodies (35.7%). In some instances, clinicians treat unexplained patients with RPLs who have antinuclear antibodies as if they had the antiphospholipid syndrome. We found that 44.0% of these patients also were positive for aPEs. However we emphasize that antinuclear antibodies are not aPAs,and their pathogenesis remains to be elucidated (36). We suggest that all patients with unexplained RPLs who are positive for antinuclear antibodies be screened for aPEs to avoid unnecessary medical treatment. Patients with RPLs who are positive for antinuclear antibodies but have no evidence of aPAs do not require medication because the presence of antinuclear antibodies does not predict subsequent pregnancy loss (37).

We recently reported that kininogen-dependent IgG aPEs augmented thrombin-induced platelet aggregation in vitro (11). Kininogens bind to platelets and inhibit thrombin-induced platelet aggregation. Our data support the hypothesis that kininogen-dependent aPE may cause thrombosis in vivo as a result of disruption of the normal antithrombotic effects of kininogen (11,38,39). As shown by our data, most aPE-positive patients with RPLs were dependent on kininogens. In the future, antiplatelet therapy such as low-dose aspirin may be beneficial for kininogen-dependent aPE positive patients with RPLs. This possibility awaits confirmation in appropriately designed clinical trials.

REFERENCES

1. McNeil HP, Chesterman CN, Krilis SA. Immunology and clinical importance of antiphospholipid antibodies. Adv Immunol 1991; 49: 193-280.

2. Hughes GRV, Harris EN, Gharavi AE. The anticardiolipin syndrome. J Rheumatol 1986; 13: 486-89.

3. Love PE, Santoro SA. Antiphospholipid antibodies: anticardiolipin and the lupus anticoagulant in systemic lupus erythematosus (SLE) and in non-SLE disorders. Ann Intern Med 1990; 112: 682-98.

4. Staub HL, Harris EN, Khamashta MH, Savidge G, Chahade WH, Hughes GRV. Antibody to phosphatidylethanolamine in a patient with lupus anticoagulant and thrombosis. Ann Rheum Dis 1989; 48:166-69.

5. Karmochkine M, Cacoub P, Piette JC, Godear P, Boffa MC. Antiphosphatidylethanolamine antibody as the sole antiphospholipid antibody in systemic lupus erythematosus with thrombosis.Clin Exp Rheumatol 1992; 10: 603-05.

6. Karmochkine M, Berard M, Piette JC, Cacoub P, Ailland MF, Harlet JR, Godeau P, Boffa MC.Antiphosphatidylethanolamine antibodies in systemic lupus erythematosus. Lupus 1993; 2: 157-60.

7. Boffa MC, Berard M, Sugi T, McIntyre JA. Antiphosphatidylethanolamine antibodies as the only antiphospholipid antibodies detected by ELISA. II.. Kininogen reactivity. J Rheumatol 1996; 23: 1375-9.

8. Sugi T, McIntyre JA. Autoantibodies to phosphatidylethanolamine (PE) recognize a kininogen-PE complex. Blood 1995; 86: 3083-89.

9. Roubey RAS. Autoantibodies to phospholipid-binding plasma proteins: A new view of lupusanticoagulant and other "antiphospholipid" autoantibodies. Blood 1994; 84: 2854-67.

10. Sugi T, McIntyre JA. Phosphatidylethanolamine induces specific conformational changes in thekininogens recognizable by antiphosphatidylethanolamine anitbodies. Thromb Haemost 1996; 76: 354-60.

11. Sugi T, McIntyre JA. Autoantibodies to kininogen- phosphatidylethanolamine complexes augment thrombin-induced platelet aggregation. Thromb Res 1996; 84: 97-109.

12. Branch DW, Scott JR, Kochenour NK, Hershgold E. Obstetric complications associated with the lupus anticoagulant. N Engl J Med 1985; 313: 1322-6.

13. Petri M. Pathogenesis and treatment of the antiphospholipid antibody syndrome. Advances in Rheumatology 1997; 81: 151-177.

14. Katano K, Aoki K, Sasa H, Ogasawara M, Matsuura E, Yagami Y. beta₂-Glycoprotein I-dependent anticardiolipin antibodies as a predictor of adverse pregnancy outcomes in healthy pregnant women.Hum Reprod 1996; 11: 509-12.

15. Branch DW, Silver RM. Criteria for antiphospholipid syndrome: early pregnancy loss, fetal loss,or recurrent pregnancy loss? Lupus 1996; 5: 409-413.

16. Maejima M, Fujii T, Okai T, Kozuma S, Shibata T, Taketani Y. beta2- Glycoprotein I-dependent anticardiolipin antibody in early recurrent spontaneous abortion. Hum Reprod 1997; 12: 2140-42.

17. Simpson JL, Carson SA, Chesney C, Conley MR, Metzger B, Aarons J, Holmes LB, Jovanovic-Peterson L, Knopp R, Mills JL. Lack of association between antiphospholipid antibodies and first-trimester spontaneous abortion: prospective study of pregnancies detected within 21 days of conception. Fertil Steril 1998; 69: 814-20.

18. Hossain AM, Whitman GF, Khan I. Kininogen present in rat reproductive tissues is apparently synthesized by the liver, not by the reproductive system. Am J Obstet Gynecol 1995; 173: 830-4.

19. Hermann A, Buchinger P, Somlev B, Rehbock J. High and low molecular weight kininogen and plasma prekallikrein/plasma kallikrein in villous capillaries of human term placenta. Placenta 1996;17: 223-230.

20. Brann DW, Greenbaum L, Mahesh VB, Gao X. Changes in kininogens and kallikrein in the plasma, brain, and uterus during pregnancy in the rat. Endocrinology 1995; 136: 46-51.

21. Adam A, Damas J, Galay G, Bourdon V. Quantification of rat T-kininogen using immunological methods.Biochem Pharmacol 1989; 38: 1569-75.

28. Andrew M, Paes B, Johnston M. Development of the hemostatic system in the neonate and young infant. American Journal of pediatric Hematology Oncology 1990; 12: 95.-104.

29. Reverdiau-Moalic P, Delahousse B, Body G, Bardos P, Leroy J, Gruel Y. Evolution of blood coagulation activators and inhibitors in the healthy human fetus. Blood 1996; 88: 900-906.

30. Matzner W, Chong P, Xu G, Ching W. Characterization of antiphospholipid antibodies in women with recurrent spontaneous abortions. J Reprod Med 1994; 39: 27-30.

31. Yetman DL, Kutteh WH. Antiphospholipid antibody panels and recurrent pregnancy loss: prevalence of anticardiolipin antibodies compared with other antiphospholipid antibodies. Fertil Steril 1996; 66: 540-6.

32. Gilman-Sachs A, Lubinski J, Beer AE, Brend S, Beaman KD. Patterns of anti-phospholipid antibody specifities. J Clin Lab Immunol 1991; 35: 83-8.

33. Lockshin MD, Druzin ML, Goei S, Qamar T, Magid MS, Jovanovic L, Ferenc M. Antibody to cardiolipin as a predictor of fetal distress or death in pregnant patients with systemic lupus erythematosus. N Engl J Med 1985; 313: 152-156.

34. Kutteh WH. Antiphospholipid antibodies and reproduction. J Reprod Immunol 1997; 35: 151-171.

35.Gris JC, Neveu S, Maugard C, Tailland ML, Brun S, Courtieu C, Biron C, Hoffet M, Hedon B, Mares P. Prospective evaluation of the prevalence of haemostasis abnormalities in unexplained primary early recurrent miscarriages. Thromb Haemost 1997; 77: 1096-103.

36. Tan EM. Autoantibodies in pathology and cell biology. Cell 1991; 67: 841-842.

37. Ogasawara M, Aoki K, Kajiura S, Yagami Y. Are antinuclear antibodies predictive of recurrent miscarriage [letter] ? Lancet 1996; 347: 1183-4.

38. McIntyre JA, Wagenknecht DR, Sugi T. Phospholipid binding plasma proteins required for antiphospholipid antibody detection-an overview. Am J Reprod Immunol 1997; 37: 101-110.

39. Sugi T, McIntyre JA. Plasma proteins required for antiphospholipid antibody detection. Nouvelle Revue Francaise d' Hematologie. 1995; 37: S49-52.