Monoclonal Antibodies Against Maternal Major Histocompatibility Complex Class I Molecules Induce Rapid Abortion in Mice

AJRI 1999; 11:217-223 Printed in Irelrrnd - oil rights reserrrd.

Copyright 0 Munksgaard. 1999

ISSN 8755.8920 American Journal of Reproductive lmmunologg

Monoclonal Antibodies Against Maternal Major Histocompatibility Complex Class I Molecules Induce Rapid Abortion in Mice NIKOLA KNEZEVIC, BORIS NIKOLIC, KARMEN BRAJSA, RADAN SPAVENTI, NIVES JONJIC, STIPAN JONJIC, AND SUZANA MARUSIC

Knexric N , Nikolic B, Brajsa K, Spuuentil R , Jonjic N , Jonjic S , Murusic S. Mono- clonal antibodies aguinst maternal rnajor histocompatibility coniplex class I molecules induce rupid abortion in mice. AJRI 1999; 41:217-223 0 Munksgaard, Copenhagen PROBLEM: The role of antibodies against fetal or maternal antigens in maintaining or losing pregnancy is not clear. METHOD OF STUDY: Term-pregnant mice were injected with monoclonal antibodies against only fetal or fetal and maternal major histocompatibility complex class I molecules. The development of pregnancy was then followed. RESULTS: Antibodies against maternal, but not fetal, major histocompatibility complex class I molecules induced abortion in mice. The abortion occurred 6-8 hr after the administration of autoreactive antibodies. The abortion could only be induced after the formation of placenta. Antibodies against tumor necrosis factor-a could not prevent or postpone the abortion. Extensive bleeding has been detected in the placenta of aborting mice 3 hr after the administration of the antibodies. CONCLUSIONS: This study indicates that autoreactive antibodies present risk for pregnancy and that the damage leading to abortion induced by such antibodies most likely occurs at the maternal side of placenta.

Key words: Abortion, monoclonal antibodies, MHC class I, placenta

NIKOLA KNEZEVIC* BORIS NIKOLICT KARMEN BRAJSAf SUZANA MARUSICS Department of Molecular Medicine, Institute Ruder Boskovic, Zagreb, Croatia

RADAN SPAVENTI Pliva Research Institute, Zagreb, Croatia

NIVES JONJIC STIPAN JONJIC Department of Physiology and Immunology, Medical School, University of Rijeka, Rijeka, Croatia

Address reprint requests to Suzana Marusic, Department of Immunology, Genetics Institute, 87 Cambridge Park Drive, Cambridge, MA 02140.

Submitted October 1, 1998; accepted October 22, 1998.

* Present address: Department of Urology. Medical School, Zagreb University. Zagreb. Croatia t Present address: Massachusetts General Hospital, Boston, Massachusetts

$ Present address: Department of Immunology, Genetics Institute, Cambridge, MA 02140 Present address: Pliva Research Institute. Zagreb, Croatia

AMERICAN JOURNAL OF REPRODUCTIVE IMMUNOLOGY VOL. 41, 1999

218 / KNEZEVIC ET AL.

INTRODUCTION

Most of the time, a fetus and its mother differ in the molecules of the major histocompatibility complex (MHC). Comparable difference in MHC molecules between transplanted organ and its recipient would lead to a rapid organ rejection. However, a fetus is not treated as a transplanted organ by the immune system of a pregnant female. It is not rejected in spite of the differences in the MHC molecules. It is not clear how a fetus is protected from the rejection by the immune system of its mother. It is believed, however, that the placenta plays a critical role in this protection.

Clark et al.' have observed that CBAiJ females mated with DBA12J males have low pregnancy suc- cess rate (high number of fetuses do not develop and are resorbed). Further studies have indicated that paternal MHC molecules, expressed by fetal tissue, play a role in the maintenance of the pregnancy. It has been suggested that fetal MHC class I molecules induce a protective antibody response in pregnant females.* In fact, antibodies against paternal MHC class I molecules can protect fetuses in CBA x DBA/2 pregnancies when injected into pregnant females.' These observations lead to a hypothesis that antibodies against paternal MHC molecules actually protect the fetus and that the lack of antipaternal antibodies could present a risk for abortion. However, mice deficient in either MHC class I or class I1 molecules, have normal pregnancies. No abnormalities in pregnancy have been observed when the mother, the fetus, or both lack MHC r n o l e c u l e ~ . ~ ~ ~ These findings have brought into question the protective role of the paternal MHC molecules and antibodies generated against them by the mother's immune system. In pregnancies in humans, the protective role of antibodies against paternal MHC class I molecules could not be documented.67 It has been shown that the MHC compatibility between the fetus and its mother does not influence the pregnancy outcome in humans7 No increase in abortion has been observed when the mother and its fetus express the same MHC molecules. Most of these studies indicate that paternal MHC molecules expressed by the fetus, and antipaternal MHC antibodies produced by a pregnant female do not protect the fetus from rejection by the maternal immune system.

In contrast to antipaternal MHC class I antibodies, which seem to neither reduce nor increase the risk of abortion in humans, the presence of autoreactive, antiphospholipid antibodies in pregnant women pre- sents a risk factor for recurrent pregnancy loss.* Isolated immunoglobulin (Ig) G fraction from patients

with antiphospholipid antibodies cause fetal death if injected into pregnant BALB/c mice on day 8 of pregnancy. I ' Typically, fetal death or fetal resorption occurs within 24-48 hr of the injection of the antiphospholipid antibodies."

In the present study, we show that monoclonal antibodies against maternal, but not paternal, MHC class I molecules induce rapid abortion in mice. Abor- tion could be induced using these antibodies at any time during the pregnancy after the formation of placenta (day 7.5 in mice). We suggest that this could present a useful model for studying the mechanism of recurrent pregnancy loss caused by autoimmune antibodies.

MATERIALS AND METHODS

Mice C3H (H-2k), Balbjc (H-2d), A/J (H-2"), C3H x Balb/c, and Balb/c x C3H were bred in our animal facility at Ruder Boskovic Institute and were kept under con- ventional conditions. Animals were used at the age of 2-4 months.

Monoclonal Antibodies Monoclonal antibodies against KkDk MHC class I molecules (hybridoma 15-1-5, mouse IgG2b, ref.12) and anti-KdDd (hybridoma 34-1 -2, mouse IgG2a, ref.") were purified from ascites by ammonium sul- phate precipitation followed by size separation over Sephadex column and extensive dialysis against phos- phate-buffered saline (PBS).

Induction of Abortion For mating, female and male mice were placed overnight at a ratio of 2:l. The presence of coital plug was taken to indicate probable pregnancy (day 0). Pregnant mice were injected intraperitoneally (i.p.) with 0.5 mL prewarmed PBS or purified monoclonal antibodies.

Production of Polq'clonal Anti- TNF-x Antibodies The anti-tumor necrosis factor (TNF)-r antibodies were raised in rabbits against pure recombinant TNF- E. The antiserum was produced according to the protocol described elsewhere.I4 Briefly, rabbits were immunized by subcutaneous injection of 10 pg of TNF-r in complete Freund's adjuvant. Two weeks later, animals were boosted with the same amount of TNF-r in incomplete Freund's adjuvant, and were boosted again with 5 pg TNF-r in saline 3 days before bleeding. The antiserum was extensively dialysed against PBS and stored at - 30°C before use.

0 MUNKSGAARD, COPENHAGEN

Pathohystological Analysis Antibody (1 5-1-5, 10 mgImouse)-treated mice were sacrificed 3 and 6 hr after the injection. PBS-injected (control) mice were sacrificed 6 hr after the injection. Liver, kidney, and uterus containing fetuses were fixed in boujon for 24 hr and were imbedded in paraffin. The sections were cut at 4 ym and stained with hematoxylin and eosin.

RESULTS

Monoclonal Antibodies Against Maternal MHC Class I Molecules Induce Abortion in Mice Daily treatment of C3H mice, for 3 to 4 weeks starting at the day of birth, with high doses (500 ygjg of body weight) of 15- 1-5 monoclonal antibodies, diminishes CD8 + T cell development without causing any gross pathogenic effect.15 However, a single injection of antibodies (10 mg/mouse) into pregnant C3H mice induced rapid abortion 6-8 hr after the injection. The same dose of antibodies induced abortion in C3H mice impregnated by Balb/c males (C3H x Balbj c pregnancy). but not in BalbIc mice impregnated by C3H males (Balbic x C3H pregnancy). (By conven- tion, the first denominated strain is maternal and the second is paternal.) These results indicate that the antibodies did not induce abortion due to a non-specific toxicity and that they had to be specific for maternal MHC class I molecules in order to induce abortion. Antibodies specific only for paternal MHC class I molecules did not induce abortion.

To test whether the induction of abortion was a unique and fortuitous property of 15-1-5 monoclonal antibody, we tested the ability of another anti-MHC class I monoclonal antibody, 34-1-2, to induce abortion. This antibody has a different specificity and a different isotype from 15-1-5 (it is IgG2a antibody, specific for KdDd). 34-1 -2 antibodies induced abortion in Balb/c mice pregnant by C3H males but not in C3H mice pregnant by Balbjc males. Again, only antibodies against maternal, but not paternal, MHC class I molecules induced abortion. Collectively, these data show that monoclonal antibodies induce rapid abortion if they are specific for the MHC class I molecules of the pregnant females, but not if they are specific only for the paternal MHC class I molecules, expressed by the fetus and by the fetal part of the placenta.

There was a possibility that at the time of abortion induction, only maternal MHC class I molecules were expressed in the fetus. TO address this possibility, we performed immunohistological staining of fetal tissues after the pregnant females had been injected with anti-MHC class I monoclonal antibodies (data not

shown). Both 15-1-5 and 34-1-2 monoclonal antibodies entered fetuses and bound to the appropriate MHC class I molecules. Thus, 15-1-5 monoclonal antibodies bound fetal cells of C3H, (C3H x Balbic) and (Ba1b:c x C3H)Fl haplotype, but not of Balb/c haplotype. On the other hand, 34-1-2 monoclonal antibodies bound fetal cells of Balbj, (C3H x Balb/ c)F1 and (Balbjc x C3H)Fl haplotype, but not C3H haplotype. Similar results were obtained in all tested embryos, ranging in age from day 14 to day 18 of pregnancy. These results indicate that both maternal and paternal MHC class I molecules were expressed by fetuses at least after day 14 of pregnancy. Also, our results show that both 15-1-5 and 34-1-2 monoclonal antibodies could pass placenta and enter fetuses regardless of the female’s MHC haplotype. Therefore, a differential expression of maternal and paternal MHC class I molecules by fetuses could not account for the difference in the ability to induce abortion between antimaternal and antipaternal antibodies.

Finally, we tested whether binding to only one class I MHC molecule (i.e. K), rather than both K and D molecules, is sufficient to induce abortion. To test this, we used A/J mice (KkDd) and injected them with 15-1-5 monoclonal antibodies. In this case, 15-1-5 monoclonal antibodies will only bind K, but not D, MHC class I molecules. Again, the antibodies induced abortion within 8 hr of injection (Table I). These results indicate that binding of antibodies to either one or two types of MHC class I molecules of maternal origin could induce abortion.

Abortion Was Induced After the Formation of Placenta In the following experiment, we determined the time window during which abortion could be induced by injecting mice with anti-MHC class I monoclonal antibodies. We injected term-pregnant C3H mice with 15- 1-5 monoclonal antibodies (1 0 mg ‘mouse) and followed the outcome of the pregnancy by inspecting the mice for vaginal bleeding and/or expulsion of fetuses. When antibodies were injected on days 5, 6, or 7 of pregnancy, abortions followed invariably on day 8 of pregnancy. Antibodies injected on days 8, 12, 14, or 18 of pregnancy induced abortion 6-8 hr after the injection (Table 11). The occurrence of abortion coin- cides with the time of placenta formation (day 7.5 in mice).

Next we tested the dose of antibodies necessary to induce abortion in mice (Fig. 1). A dose of 10 mg of monoclonal antibody induced abortion in 100% of pregnant mice. A dose of 7 mg induced abortion in a large majority of mice (go‘%), with a slightly pro- longed interval between the injection and abortion

ANTI-MHC CLASS I ANTIBODIES INDUCE RAPID ABORTION / 219


220 I KNEZEVIC ET AL.

onset (one out of ten injected mice aborted 24 hr after the injection, seven aborted 10 hr after the injection, and two mice did not abort). A dose of antibody, 5 mg per mouse, induced abortion in 20% of mice, while a dose of 3 mg of antibodies did not induce abortion.

Pa tho h ist ologicul Analysis Since it only took 6-8 hr from the time the antibodies had been injected until the abortion occurred, no necrotic tissue could be observed by light microscopy (it takes 6-8 hr for pathological changes of necrosis, detectable by light microscopy, to develop). However, blood vessel dilatation and extravasation of blood was already observed 3 hr after the antibody injection (Fig. 2B). No bleeding was observed in any part of placenta of control mice (injected with PBS or monoclonal antibodies specific for fetal but not maternal MHC class I molecules) (Fig. 2A). The bleeding and extravasation was even more extensive 6 hr after the injection of antibodies (Fig. 2C). Gross anatomical examination revealed that the placenta of the mice that would have aborted was swollen 6 hr after the antibody injection, most likely due to intraplacental bleeding (normal placenta was 1.8 mm thick, while 6 hr after the antibody injection, placenta of aborting mice was 4.5 mm thick). No bleeding or any other tissue damage was observed in other fetal or maternal tissues examined (data not shown).

Antibodies Against Tumor Necrosis Fuctor-a Do Not Prevent Abortion Lipopolisacharide (LPS) has been reported to induce fetal loss in mice when injected on day 8 of pregnancy." Injection of pentoxifylline, a TNF-a-sup- pressing drug, reduced the level of fetal resorption induced by injection of 0.1 mg LPS. It was suggested that TNF-a plays a role in fetal resorption induced by LPS." We tested the possibility that TNF-a plays a role in inducing abortion in our model. We tested the

TABLE I. Monoclonal Antibodies Against Maternal MHC

capacity of polyclonal anti-TNF-a antibodies to postpone or prevent abortion induced with 15-1-5 monoclonal antibodies (Table 111). No effect of these antibodies on the course of the abortion was observed regardless of the dose of the anti-TNF-a antibodies used. Neutralizing capacities of the same batch of anti-TNF-a antibodies in aivo was confirmed in an independent experiment. Even the lowest dose of the same anti-TNF-a antibodies abolished the antiviral effect of TNF-a in viz.0." Therefore, these results suggest that TNF-a does not play a major role in the antibody-induced pregnancy loss.

DISCUSSION

In the present study, we have shown that monoclonal antibodies against maternal, but not paternal, MHC class I molecules induce rapid abortion in mice. The antibodies could induce abortion from day 8 until day 18 of pregnancy. The fetal loss occurred usually 6-8 hr after the injection of antibodies. Pregnancy loss was accompanied by intraplacental bleeding which was detected 3 hr after the injection.

It is not likely that the injection of anti-class I antibodies induced abortion by directly damaging fetuses, for several reasons. First, monoclonal antibodies against paternal MHC class I antibodies never induced abortion. We have found that antipaternal MHC class I antibodies entered fetuses with the same efficiency as antimaternal MHC class I antibodies. Also, similar levels of maternal and paternal MHC class I molecules were detected in F1 fetuses. Second, the injection of much higher doses of anti-class I monoclonal antibodies, starting at the day of birth, did not induce tissue damage. It is not likely that a day-18 fetus (only 1 or 2 days before delivery) differs from a newborn mouse to the extent that the anti- MHC class I antibodies would induce lethal damage in the former, but no damage in the later. Third,

Class I Molecules Induce Abortion

Monoclonal antibodiesa Specificity Pregnancy Abortions treatmentsb

None None C3H x C3H (KkDk x KkDk) 0,5 15-1 -5 KkDk C3H x C3H (KkDk x KkDk) 2O!2OC 15-1 -5 KkDk C3H x Balb c (KkDk x KdDd) 7,'7 15-1 -5 KkDk Balb,'c x C3H (KdDd x KkDk) 0,7 15-1 -5 KkDk A,'J x A J (KkDd x KkDd) 5,5 34-1 -2 KdDd C3H x Balb c (KkDk x KdDd) 0,'7 34-1 -2 KdDd Balb c x C3H (KdDd x KkDk) 7,'7

'' Ten niilligrl~nx of intlictrtetl I I I ~ I ~ ~ C ~ O I I ~ tmt iho~ly 1 w . s injected i.p. irito ench prcy iant nioiise ( c k i j . 14 uf pregririncy). The number 01 niicr t/iLir trborted o w r rllr ninnher of' rniw tlicit receiiwl the trmtment. The nuniben ulso i~iclude tile mice ,fioni other e.\-per.itnent.s.



100 0

80. a m 60- a 0

w 0 4 0 . a

8 6 20.

TABLE II. Antibodies Against Maternal MHC Class I Molecules Induce Abortion on Day 8 or Later During the Pregnancy

-

Hours after the injection of antibodies

Day of pregnancy" 5.5 6 7 11 35 48

5 6 7 8 12 14 18

- - - - - Bleeding - - - - Bleeding - - - Bleeding - Bleeding - Bleeding Aborted - Bleeding Aborted - Bleeding Aborted

" Term-prrgntmt C3H rnice n w e injected ivith 10 rng of 15-1-5 nionoclonul untibodieies ut time 0. There were >5 rnice per group. All the rnice were observed until the e-upected term dute and none delivered pups. All PBS-treuted controls delivered us expected und had no cuginul bleeding.

pathohistological analysis documented extensive placental bleeding, while no pathological changes were detected in the fetuses. We therefore concluded that it is highly unlikely that the abortion occurred due to a direct damage to fetuses.

We suggest that the pathogenic effect of anti-MHC class I antibodies takes place at the maternal side of placenta. This possibility was first considered after observing that only antimaternal, but not antipaternal, antibodies cause abortion of F1 fetuses. This was the case with two different antibodies tested in F1 fetuses. In (C3H x Balb/c)Fl fetuses, only 15-1-5, but not 34-1-2, antibodies caused abortion. In (Balb/c x C3H)F1, the opposite was true: 34-1-2, but not 15-1-5 antibodies caused abortion. In each case, only the antibodies against maternal, but not paternal, MHC class I antibodies caused abortion. Because both paternal and maternal MHC class I molecules are expressed on the fetal side of placenta"." and throughout the fetus (data not shown), it is on the maternal side of placenta that only maternal, but not paternal, MHC class I molecules are expressed. Our observation that only antimaternal, but not antipaternal, MHC class I antibodies cause abortion in mice emphasizes the potential importance of the role that autoimmune antibodies could play in fetal loss. Many initial studies on recurrent fetal loss in humans have failed to draw the correlation between the abortions and the presence of antipaternal antibodies in pregnant wOmen~h.7.20.71 On the other hand, autospecific, antiphospholipid antibodies have been shown to actually induce fetal loss in mice" and correlate with the recurrent pregnancy loss in h u r n a n ~ . ~ . ' ~ . ' ~ - ' ~ Our model confirms the notion that autospecific, rather than antipaternal, antibodies represent a risk for pregnancy.

It is not clear how antibodies against maternal MHC class I antibodies cause abortion. In LPS-induced abortion, it has been suggested that the release of TNF-R plays an important role in fetal abortion/resorption.16 This does not seem to be the case in our model of antibody-induced abortion. We could not block or delay the abortion using anti-TNF-cc antibodies. Another possibility is that complement cascade activation induced placental damage and abortion. The time course of abortion induced with anti-MHC class I antibodies resembles hyperaccute graft rejection, which is caused by antibodies and complement-mediated d a r n a ~ e . ~ ~ . ~ ~ In the hyperaccute graft rejection, antigraft antibodies bind to endothelial cells of the blood vessel of the graft, and activate complement cascade, leading to blood vessel damage, thrombosis, and compromised circulation of the graft.'5.'6 It is possible that the antibody-induced abortion has the same mechanism. However, it should be emphasized that even if the abortion is mediated by

o f I 0 5 10 15 DOSE of MONOCLONAL ANTIBODIES (mg)

Fig. 1. Abortion induction with different doses of anti-MHC class I antibodies. Different doses of 15-1-5 antibodies were injected into C3H pregnant mice on day 14 of pregnancy. The development of abortion was observed. Mice that did not abort were left to deliver.


222 i KNEZEVIC ET AL.

A

I

,

B I 4

Fig. 2. Histological analysis of placenta from PBS- and antibody- iiijected mice. Pregnant C3H mice were killed 6 hr after the itijec- tion of PBS (A), 3 hr (B) or 6 hr (C) after the injection of 15-1-5 monoclonal antibodies. The sections were stained with hematoxylin and eosin. Extravasation was observed 3 and 6 hi- after the antibody injection ( m ~ o i i . . ~ ) . No extravasation was observed in any part of the placenta from mice injected with PBS. (A color copy of Fig. 2 is available from the authors upon request.)

the activation of a complement cascade. the placenta seems to be much more sensitive to this kind of damage than other tissues. In the absence of pregnancy, the injection of multiple doses of the anti-class I monoclonal antibodies does not cause tissue damage.’’

This study shows that autospecific, rather than antipaternal, antibodies represent a danger throughout the pregnancy. This model of abortion-induction should be useful for studying the mechanisms of autospecific antibody-induced pregnancy loss.

TABLE 111. Anti-TNF-x Antibodies Do Not Prevent Abortion Induced with Anti-Class I Monoclonal Antibodies

Aborteda Abortedb Aborted treated“

15-1-5 0.1 mL anti-TNF-c/ 3 3 15-1-5 0.3 mL anti-TNF-x 3 3 15-1 -5 0.6 mL anti-TNF-cx 3 3 15-1-5 None 3 3 None None 0 3

A ckizoi rki!gn?en ts This work was supported by Croatian Ministry of Science grants Nos. 1-08 308 and 3-01-169. We thank Joy Miyashiro for the critical reading of the manuscript.

REFERENCES

1. Clark DA. McDermott M. Sczewzak MR: Impairment of host vs. graft reaction in pregnant mice. 11. Selective suppression of cytotoxic cell generation correlates with soluble suppressor activity and with successful allogeneic pregnancy. Cell Immunol 1980; 52:106-111.

2. Clark DA: Contraversies in reproductive immunology. Crit Rev Immunol 1991; 11:215-247.

3. Chaouat G. Kolb J-P. Kiger N, Stanislawski M, Weg- man TG: Immunologic consequences of vaccination against abortion in mice. J Immunol 1985; 134:1594- 1598.

4. Zijlstra M. Bix M. Simister NE. Loring JM, Raulet DH. Jaenish R: b2-niicroglobulin deficient mice lack CD4 - CD8 + cytotoxic T cells. Nature 1990; 344:742-746.

5. Grusby MJ. Randall SJ, Papaioannou VE, Glimcher LH: Depletion of CD4 + T cells in major histocompatibility complex class 11-deficient mice. Science 1991; 153:1417- 1419.

6. Sargent IL, Wilkins T, Redman CWG: Maternal immune response to the fetus in early pregnancy and recurrent miscarriage. Lancet 1988: 2(8620):1099- 1104.

7. Jazwinska EC, Kilpatrick DC, Smart GE. Liston WA: Feto-maternal HLA compatibility does not have a major influence on human pregnancy except for lymphocy- totoxicity production. Clin Exp Iininunol 1987: 68: 1 16- 122.

8. Carreras LO. Machin CJ. Defreyn G. Machin SJ, Vermylen J , Deman R. Spitz B. Van Assche A: Arterial thrombosis, intrauterine death and lupus anticoagulant detection of immunoglobulin interfering with prostacy- clin formation. Lancet 1981: 1:244-246.

9. Prentice RL. Gatenby PA. Loblay RA. Shearman RP. Kronenberg H. Basten A: Lupus anticoagulant in pregnancy. Lancet 1984: 1:464.



10. Lockshin MD, Druzin ML, Goei S: Antibody to cardi- olipin as a predictor of fetal distress or death in pregnant patients with systemic lupus erythematosus. N Engl J Med 1985; 313:152-156.

11. Branch DW, Dudley DJ. Mitchel MD, Creighton KA, Abbott TM, Hammond EH, Daynes RA: Immunoglob- ulin G fraction from patients with antiphospholipid antibodies cause fetal death in Ba1b;c mice: A model for autoimmune fetal loss. Am J Obstet Gynecol 1990; 163:210-216.

12. Ozato K. Mayer NM. Sachs DH: Hybridoma cell lines secreting monoclonal antibodies to mouse H-2 antigens. J Immunol 1980; 124:533-540.

13. Ozato K, Mayer MN, Sachs DH: Monoclonal antibodies to mouse major histocomatibility complex antigen. IV. A series of hybridoma clones producing anti-H2d antibodies and an examination of expression of H-2D antigens on the surface of these cells. Transplantation 1982; 34:113-120.

14. Nauciel C, Espinasse-Maes F: Role of gamma interferon and tumor necrosis factor alpha in resistance to S0b7?0~2dh fjphin?uriiin? infection. Infect Immunity 1992; 60:450-454.

15. Marusic-Galesic S, Stephany DA, Longo DL, Kruisbeek AM: Development of CD4 - CD8 + cytotoxic T cells requires interaction with class I MHC molecules. Nature 1988: 333:180-182.

16. Gendron RL. Nestel FP, Lapp WS, Baines MG: Lipo- polisaccharide-induced fetal resorption in mice is associ- ated with the intrauterine production of tumor necrosis factor-alpha. J Reprod Fertil 1990; 90:395-402.

17. Pavic I, Polk B, Crnkovic I, Lucin P, Jonjic S, Koszin- ski UH: Participation of endogenous tumor necrosis factor alpha in host resistance to cytomegalovirus infection. J Gen Virol 1993; 74:2215-2223.

18. Chatterjee-Hasrouni S, Lala PK: Localization of H-2 antigens on mouse trophoblast cells. J Exp Med 1979;

149:1238- 1255. 19. Lala PK, Chatterjee-Hasrouni S, Kearns M, Mont-

gomery B. Calvicenzo V: Immunobiology of the feto- maternal interface. Immunol Rev 1983; 75:87-116.

20. Beer AE: Immunopathologic factors contributing to recurrent spontaneous abortion in humans. Am J Reprod Immunol 1983; 4:182-184.

21. Unander M, Olding LB: Habitual abortion: Parental sharing of HLA antigens, absence of maternal blocking antibody and suppression of maternal lymphocytes. Am J Reprod Immunol 1983: 4:171-178.

22. Petri M, Golbus M, Anderson R, Whiting-O’Keefe Q, Corash L, Hellman D: Antinuclear antibody, lupus anticoagulant, and anticardiolipin antibody in women with idiopathic habitual abortion: A controlled, prospective study of forty-four women. Arthritis Rheum 1987; 30:601-606.

23. Boddi M, Prisco D, Fedi S, Cellai AP, Liotta AA, Parretti E, Mecacci F, Mello G, Abbate R: Antiphos- pholipid antibodies and pregnancy disorders in women with insulin dependent diabetes. Thomb Res 1996; 82:207-2 16.

24. Festini MR, Limson GM, Marou T: Autoimmune causes of recurrent pregnancy loss. Kobe J Med Sci 1997; 43:143-157.

25. Miyatake T, Sato K, Takigami K, Koyamada N, Han- cock WW, Bazin H, Latinne D, Bach FH, Soares MP: Complement-fixing elicited antibodies are a major com- ponent in the pathogenesis of xenograft rejection. J Immunol 1998; 160:4114-4123.

26. Bach FH, Robson SC, Ferran C, Winkler H, Millan MT. Stuhlmeier KM, Vanhove B, Blakely ML, van der Werf WJ, Hofer E, de Martin R, Hancock WW: Endothelial cell activation and thromboregulation during xenograft rejection. Immunol Rev 1994; 1 4 1 5 30.


Documents

Monoclonal Antibodies Against Maternal Major Histocompatibility Complex Class I Molecules Induce Rapid Abortion in Mice