- Email: [email protected]
The Antiabortive Effect of Progesterone-Induced Blocking Factor in Mice Is Manifested by Modulating NK Activity
CELLULAR IMMUNOLOGY ARTICLE NO.
177, 194–199 (1997)
CI971090
The Antiabortive Effect of Progesterone-Induced Blocking Factor in Mice Is Manifested by Modulating NK Activity JULIA SZEKERES-BARTHO, G. PAR, GY. DOMBAY, Y. C. SMART,*
AND
Z. VOLGYI
Department of Microbiology, University Medical School of Pecs, H-7643 Pecs, Hungary; and *Discipline of Surgical Sciences, Faculty of Medicine, University of Newcastle, Newcastle, Australia Received October 29, 1996; accepted January 26, 1997
Immunologic effects of progesterone are mediated by a protein named the progesterone-induced blocking factor (PIBF), which inhibits NK activity and displays an antiabortive effect in mice. Our previous data provide indirect evidence for the importance of PIBF in the maintenance of normal gestation. This study was aimed at investigating whether neutralization of endogenous PIBF production influences pregnancy outcome and if so, what are the mechanisms that participate in this process. Syngeneically pregnant Balb/c mice on Day 8.5 of pregnancy were injected ip with 0.3 mg/kg of RU 486 or with 0.5 mg of rabbit anti-PIBF IgG alone, or together with anti-NK monoclonal antibodies. Mice treated with the same amount of normal rabbit serum or untreated mice of similar gestational age were used as controls. On Day 10.5 the ratio of living and resorbed embryos and NK activity of the spleen cells were determined. In mice treated with anti-PIBF the ratio of resorbed fetuses was significantly higher than that in untreated controls. In RU 486-treated mice we also observed significantly increased resorption rate, which was associated with the inability of spleen cells to produce PIBF. Both antiPIBF treatment and that with progesterone receptor blocker resulted in increased splenic NK activity. There was a positive relationship between NK activity and the rate of resorptions. All the above effects were corrected by simultaneous treatment with anti-NK or anti-NC (natural cytotoxic) antibodies. These data allow the conclusion that PIBF contributes to normal gestation in mice and that the effect of PIBF is manifested via blocking NK and/or NC activity. q 1997 Academic Press
INTRODUCTION Progesterone-dependent immunomodulation is one of the mechanisms that enables pregnancy to proceed to term. Immunologic effects of progesterone are mediated by a protein named the progesterone-induced blocking factor (PIBF) (1). Lymphocytes of healthy
194
0008-8749/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
CI 1090
/
6c1f$$$$$1
pregnant women are able to produce PIBF and the percentage of PIBF-positive cells is significantly reduced in peripheral blood lymphocytes (PBL) of recurrent aborters (2). Among other effects this protein inhibits NK activity (3) and displays an antiabortive effect in mice (4). Earlier data show that adoptive transfer of high NK activity spleen cells into pregnant mice induces abortion (5). Simultaneous PIBF treatment of pregnant mice corrects the abortive effect of NK activity (6). The above data provide indirect evidence for the importance of PIBF in maintaining normal gestation. Direct evidence for the biological significance of PIBF would be the induction of pregnancy loss by neutralization of endogenous PIBF. There are data suggesting the role of high NK activity in pregnancy loss in both mice and humans. In normal human pregnancy, NK activity is significantly lower than in nonpregnant individuals, whereas spontaneous pregnancy termination is associated with increased systemic NK activity (7–11). In mice there is direct evidence for the role of high NK activity in pregnancy termination. Resorption sites in mice are infiltrated by NK cells (12). In the abortion-prone CBA/J 1 DBA/2 murine mating combination, spleen cells from CBA/J mice contain asialo GM1/ effector cells that can kill the Be6 trophoblast cell line in vitro (13). Modulation of NK activity influences resorption rates (14). Natural killer cells are recruited into the pregnant uterus during early pregnancy in humans, mice, and pigs (15–17). In pregnant women, although the relative proportion of decidual NK cells is increased, decidual NK activity is lower in normal than in anembryonic pregnancies and in recurrent spontaneous abortions (18). Lachapelle et al. (19) have demonstrated an increased proportion of CD16/ CD56/ cells in the endometrium of recurrent aborters. NK activity is one major component of natural cellmediated cytotoxicity (NCMC) with natural cytotoxic (NC) activity being the second component (20). Limited data indicate that NC activity is present in the decidua of pregnant mice (21). Its role in pregnancy outcome has not been investigated.
04-22-97 22:57:56
cia
ANTIABORTIVE EFFECT OF PIBF
Since PIBF is antiabortive in mice and inhibits NK activity in vitro, the present study was aimed at investigating whether neutralization of endogenous PIBF activity in vivo results in pregnancy termination and if it does, what are the mechanisms that lead to this event. MATERIALS AND METHODS Treatment of Pregnant Mice by RU 486, Anti-PIBF Anti-NK, and Anti-NC Antibodies Fourteen-week-old Balb/c mice (LATI, Godollo, Hungary) were kept under standard conditions (4 animals per cage). Female mice were caged overnight with the males and checked for the presence of vaginal plugs the following morning. The day on which the plug was observed is considered to be Day 0.5 of pregnancy. Pregnancy was later verified by scoring corpora lutea. Various treatments were administered on Day 8.5 of pregnancy. (1) Females were injected intraperitoneally with 0.3 mg/kg of RU 486 or with 0.5 mg of rabbit anti-PIBF IgG. Mice treated with the same amount of normal rabbit serum or untreated mice of similar gestational age were used as controls. (2) A group of anti-PIBF-treated mice was at the same time injected with monoclonal antibodies to cells mediating natural killer (NK), natural cytotoxic (NC), or natural T cell (NT) activity. The monoclonal antibodies were: (a) PK136 (anti-NK-1.1) recognizes a 76- to 80kDa type II integral membrane C-type lectin protein encoded by a member of the mouse NKR-PI gene family (22, 23). (b) 1C4 (anti-NC-1.1) which recognizes a 45-kDa surface receptor and blocks splenic NC activity approximately 70% both in vitro and in vivo (24, 25). (c) 2B6-F2 (anti-Ly-6c) identifies a subpopulation of murine Ly-6c/ NK1.1/ natural T (NT) cells (26, 27). In the presence of complement 2B6-F2 reduces splenic NK activity by approximately 50% in Balb/c mice. The antibodies 2B6-F2 and PK136 were used as serum-free supernatants and administered at a dose of 50 mg/mouse, whereas 1C4 was affinity-purified IgG and administered at a dose of 25 mg/mouse. The animals were sacrificed on Day 10.5 and their uteri were inspected. The ratio of living versus resorbed embryos was determined. Spleens were removed under aseptic conditions, cell suspension was prepared, and cytotoxic activity was determined in a 16-hr test against human embryonic fibroblast targets. NK activity was also tested using the 4-hr single-cell cytotoxicity assay against YAC targets. Smears were prepared and the expression of PIBF on the lymphocytes was determined by immunocytochemistry.
AID
CI 1090
/
6c1f$$$$$2
04-22-97 22:57:56
195
16-hr Cytotoxicity Assay This technique has been described elsewhere in detail (28). Briefly, human embryonic fibroblasts derived from 10- to 12-week embryos were used as targets. Cells were seeded on 96-well Nuclon tissue culture plates at a density of 5000 target cells/well in 0.2 ml of medium 199 supplemented with 10% fetal calf serum. The target cells were allowed to attach by overnight incubation. The following day the medium was replaced by 0.2 ml of lymphocyte suspension containing 5 1 105 lymphocytes in 0.2 ml of the same medium. After 16 hr of incubation, the plates were washed with PBS three times in order to remove lymphocytes and damaged target cells. This was followed by addition of alkaline phosphatase substrate (Sigma tablets No. 104) in diethanolamine buffer at a concentration of 1 mg/ ml. The plates were incubated for 10 min at 377C in the dark and the resulting yellow reaction product was quantitated photometrically at 405 nm. The percentage reduction in enzyme activity relative to the target cell control was considered as a measure of cytotoxicity. 4-hr Single Cell Cytotoxicity Assay for NK Activity We used the assay originally described by Grimm and Bonavida (29). One hundred milliliters of lymphocytes and the same amount of YAC target cells (2 1 106/ml each) were centrifuged at 40g for 5 min, and incubated at 377C, in 5% CO2 for 10 min. The pellets were then resuspended and 100 ml of 1% agarose (Serva) in RPMI 1640 (Gibco) was added to the mixture. One hundred milliliters of this suspension was spread over microscope glass slides previously coated with 1% agar. Target cells alone were used to detect spontaneous lysis. The gel was allowed to solidify and submerged in RPMI 1640. The slides were incubated for 4 hr at 377C in 5% CO2 . Then the gels were stained with 0.5% trypan blue for 1 min. After 2-min washes with PBS, the gels were fixed in 2% formaldehyde for 5 min and desalted in distilled water. The slides were read using light microscope with 4001 magnification. The proportion of lymphocytes bound to the target cells was expressed as a percentage of total lymphocyte population by counting 300 lymphocytes. Results are expressed as a percentage of target binding cells (TBC). Dead conjugates were scored as a percentage of the total number of conjugates by counting 50 conjugates and results are expressed as a percentage of dead conjugates (cytotoxic TBC%). The percentage of NK cells was calculated according to the formula NK% Å TBC% 1 cytotoxic TBC%/100. All results for cytotoxic TBC% were corrected for the proportion of target cells that died spontaneously in control plates. Production of the PIBF Spleen cells of 10-week-old Balb/c mice (LATI, Godollo) were adjusted to a concentration of 1 1 106/ml
cia
196
SZEKERES-BARTHO ET AL.
TABLE 1 Resorption Rates and Mean Number of Implantation Sites in Pregnant Mice Treated with Anti-PIBF and Anti-NK, Anti-NC, or Anti-NT Antibodies No. of fetuses Treatment
Living
Resorbed
Resorption rate %
No. of implants
Mean No. of implant sites
Viable fetuses/mother
None (N Å 45) Anti-PIBF (n Å 83) Anti-PIBF / PK136 (N Å 13) Anti-PIBF / 1C4 (N Å 5) Anti-PIBF / 2B6-F2 (N Å 11) Anti-PIBF / Anti-NK (N Å 29)
326 193 77 25 89 188
27 113 7 2 13 22
7.6 37a 8.3 8.3 12.7 10.4
353 133 84 27 102 210
7.6 3.6*** 6.5 4.8 9.3 7.2
7.2 2.3*** 5.9 4.4 8.1 6.5
Note. Analysis of variance showed no difference between resorption rates in anti-PIBF / anti-NK/NC-treated mice and in untreated controls, whereas anti-PIBF / anti-NK/NC-treated groups as well as untreated controls differed significantly from anti-PIBF-treated mice. a Significantly different from untreated controls at P õ 0.001 (x2 test). *** Significantly different from untreated controls at P õ 0.001 (Student’s t test).
in RPMI supplemented with 10% FCS (both from Gibco) and stimulated with 1 mg/ml of Con A (Sigma) for 48 hr at 377C in a CO2 incubator. The cell number was then adjusted to 10 1 106/ml and the cells were further incubated with 20 mg/ml of progesterone for 16 hr. At the end of the incubation period supernatants were collected. Progesterone was removed by dialysis. The supernatants were then 2000-fold concentrated on Amicon filters and used as the source of the murine PIBF. Production of PIBF-Specific IgG Spleen cells of pregnant mice were treated with 20 mg/ml of progesterone overnight. The supernatants were collected, concentrated on Amicon filters, and subjected to SDS–PAGE on 12% polyacrylamide gels. The separated bands were blotted to nitrocellulose filters, the 34-kDa band was cut out, dissolved in DMSO, and injected into rabbits weighing 4 kg each, together with complete Freund’s adjuvant. Boosters with incomplete Freund’s adjuvant were given at two weekly intervals. IgG was purified on protein A columns. The PIBF-specific antibody content was tested by ELISA. ELISA Test for Detection of PIBF in Rabbit Serum Dynatech ELISA plates were coated with 10 mg/ml of PIBF per well in 100 ml of coating buffer (carbonate– bicarbonate buffer, 0.05 M, pH 9.6). After overnight incubation at 47C plates were washed once and blocked with gelatin–BSA for 30 min at room temperature. Anti-PIBF IgG was added at different dilutions. After 1 hr of incubation plates were washed three times and HRPO-labeled anti-rabbit IgG (Sigma) was added to the wells at a dilution of 1/5000. After 30 min of incubation the plate was washed three times and the reaction was stopped by addition of 50 ml of 2 N HCl and extinctions were read at 495 nm.
AID
CI 1090
/
6c1f$$$$$2
04-22-97 22:57:56
Immunocytochemical Detection of PIBF Spleen cells were washed once, the cells count adjusted to 1 1 106/ml and centrifuged onto glass microscope slides. After drying at room temperature, the cells were fixed for 5 min in cold acetone and washed with phosphate-buffered saline (PBS). Anti-PIBF was added at 10 mg/ml and incubated for 1 hr at room temperature in a humidified atmosphere. Peroxidase-labeled anti-rabbit IgG (Dakopatts, Hungary) at a dilution of 1:200 was then added and the cells were incubated for an additional 45 min. The reaction was developed with aminoetylcarbasol, nuclei were counterstained with hematoxylin, and the slides were mounted with gelatin–glycerol. The percentage of positive cells was determined by light microscopy and 300 cells were counted. IgG from nonimmunized rabbits gave negative results and the reaction of the lymphocytes with anti-PIBF was blocked by preincubation of anti-PIBF IgG with PIBF. Statistical Analysis The two-tailed Student t test and the x2 test as well as one-way analysis of variance with the Bonferroni correction were used for statistical analysis of the data. Mean { SEM are indicated in the table and figures. Differences were considered to be significant if P was equal to or less than 0.05. RESULTS The Lack of PIBF Effect Results in Pregnancy Termination Balb/c mice on Day 8.5 of pregnancy were injected ip with 0.3 mg/kg of RU 486 or with anti-PIBF IgG. Mice injected with saline or normal rabbit serum served as controls. On Day 10.5 the ratio of living versus resorbed fetuses was significantly lower (P õ 0.001) in mice
cia
ANTIABORTIVE EFFECT OF PIBF
197
treated with anti-PIBF compared to untreated controls (Table 1). This could not be due to endotoxin contamination, since no endotoxin could be detected in antiPIBF IgG with the Limulus bioassay (Dialab Gmbh. Austria). In RU 486-treated mice we also observed a significantly increased (P õ 0.001) resorption rate (Fig. 1a), which was associated with the inability of spleen cells to produce PIBF (Fig. 1b). The Lack of PIBF Effect Results in Increased NK and NC Activity On Day 10.5, the animals were sacrificed and their splenic NK and NC activity was determined. Both antiPIBF and RU 486 treatments resulted in increased NK activity of the spleen cells (Fig. 1c). Similar results were obtained in the 16-hr cytotoxicity test (data not shown). There was a positive relationship between NK activity and the rate of resorptions. Anti-PIBF treatment did not in all cases result in pregnancy termination; 23% of mice did not respond. In nonresponders, NK activity was significantly lower than that in good responders (Fig. 2). The Antiabortive Effect of PIBF Is Mediated via Controlling NK Activity Pregnant mice on Day 8.5 of pregnancy were injected with anti-PIBF and simultaneously with anti-NK-1.1, anti-NC-1.1, or anti-Ly-6c monoclonal antibodies. AntiPIBF-treated mice showed a significantly increased (37%) resorption rate and a reduced (55%) number of implantation sites (Table 1). Mice simultaneously treated with anti-PIBF and anti-NK-1.1 (PK136) antiNC-1.1 (1C4), or anti-Ly-6c (2B6F2) did not show significantly altered resorption rates and mean numbers of implantation sites compared to untreated controls with the exception of anti-NC-1.1 treatment, where the mice did not demonstrate alteration in the number of implantation sites (Table 1). DISCUSSION In this study we have shown that neutralization of endogenous PIBF activity results in pregnancy loss in mice. Both progesterone receptor block and anti-PIBF treatment induced an increased rate of resorptions, and in the former case pregnancy loss was due to the inability of spleen cells to produce PIBF. However, antiPIBF treatment did not in all cases result in pregnancy termination; 23% of mice did not respond. In nonresponders, NK activity was significantly lower than that in good responders. It is conceivable that in these mice abortion did not occur because the spleen cells of the animals produced more than normal levels of PIBF, which was not all neutralized by the addition of antiPIBF and abortion was not induced. Clearly both treatments resulted in increased splenic NK activity, and
AID
CI 1090
/
6c1f$$$$$3
04-22-97 22:57:56
FIG. 1. (a) Progesterone receptor block induces increased resorption rates. ***P õ 0.001 (x2 test). (b) The effect of progesterone receptor block on splenic PIBF production. The bars represent the mean { SEM. ***P õ 0.001 (Student’s t test). (c) The effect of anti-PIBF treatment on splenic NK activity. The bars represent the mean { SEM. ***P õ 0.001 (Student’s t test).
the rate of NK activity correlated with the rate of resorptions. Nonspecific immunological mechanisms including
cia
198
SZEKERES-BARTHO ET AL.
pregnancy outcome probably acting at least in part on PIBF. The third monoclonal antibody (2B6-F2) reduces both NK and NC activity in Balb/c mice (26), and corrected both high resorption rates and low implantation rates due to the lack of PIBF. Our data suggest that both NC and NK activity are involved in pregnancy termination in Balb/c mice, and that PIBF exerts its antiabortive effect via inhibition of nonspecific effector mechanisms. ACKNOWLEDGMENTS This work was supported by grants from the Hungarian National Research Fund and the Hungarian Ministry of Welfare. FIG. 2. The relationship between splenic NK activity and resorption rates. The bars represent the mean { SEM. ***P õ 0.001 (Student’s t test).
NK and NC activities are probably the most important effector pathways in the fetomaternal immunological relationship. NK activity has been shown to play an important role in antitumor defense as well as in defense against intracellular parasites. There is also evidence for its role in spontaneous pregnancy termination (12–14). Haddad et al. (30) suggested that the role of NK cells in abortion in mice is to produce cytokines that activate macrophages in the uterus and macrophage-derived nitric oxide is responsible for the damage. Previously we showed that PIBF had a marked inhibitory effect on NK activity in vitro (3). NK activity is a major factor in spontaneous abortions in mice. In murine systems increased NK activity results in pregnancy termination (5), which can be corrected by simultaneous PIBF administration (6). Treatment of antiPIBF-treated mice with the anti-NK-1.1 monoclonal antibody (PK 136) corrected the resorption rates and the mean number of implantation sites. Treatment of anti-PIBF-treated mice with anti-NC1.1 (1C4) completely corrected the resorption rates, and spleen cells of anti-PIBF-treated mice exerted an increased cytotoxicity after 16 hr incubation with the target cells. In the cytotoxicity assay based on determination of endogenous alkaline phosphatase activity of the target cells we used human embryonic fibroblasts instead of the classical NC target WEHI-164. However, the requirement for a 16-hr cytotoxicity points to NCmediated activity in this system. The major mechanism of NC activity is TNFa (31). Injection of TNFa into mice resulted in pregnancy termination, while TNFa antibodies or antagonists corrected the spontaneously high resorption rates in CBA/J 1 DBA/2 matings (32). GM-CSF prevents abortion in the above system by downregulating splenic NC activity (33). Taken together these results suggest a role for NK or NC activity in
AID
CI 1090
/
6c1f$$$$$3
04-22-97 22:57:56
REFERENCES 1. Szekeres-Bartho, J., Kilar, F., Falkay, G., Csernus, V., Torok, A., and Pacsa, A. S., Am. J. Reprod. Immunol. Microbiol. 9, 15, 1985. 2. Szekeres-Bartho, J., Faust, Zs., and Varga, P., Am. J. Reprod. Immunol. 34, 342, 1995. 3. Szekeres-Bartho, J., Autran, B., Debre, P., Andreu, G., Denver, L., and Chaouat, G., Cell. Immunol. 122, 281, 1989. 4. Szekeres-Bartho, J., and Chaouat, G., Am. J. Reprod. Immunol. 23, 26, 1990. 5. Kinsky, R., Delage, G., Rosin, N., Thang, M. N., Hoffmann, M., and Chaouat, G., Am. J. Reprod. Immunol. 23, 73, 1990. 6. Szekeres-Bartho, J., Kinsky, R., and Chaouat, G., Am. J. Reprod. Immunol. 24, 105, 1990. 7. Szekeres-Bartho, J., Varga, P., and Pacsa, A. S., Am. J. Obstet. Gynecol. 155, 108, 1986. 8. Coulam, C. B., Goodman, C., Roussev, R. G., Thomason, E. J., and Beaman, K. D., Am. J. Reprod. Immunol. 33, 40, 1995. 9. Coulam, C. B., and Beaman, K. D., Am. J. Reprod. Immunol. 34, 219, 1995. 10. Kwak, J. Y. H., Rushabrook, J. I., Hurley, I. R., Mandel, F. S., Barcia, M., Cooper, G. W., and Hershlag, A., Am. J. Reprod. Immunol. 34, 100, 1995. 11. Aoki, K., Kajura, S., Matsumoto, Y., Ogasawar, M., Okada, A., Yagami, Y., and Gleicher, N., Lancet 45, 1340, 1995. 12. Gendron, R., and Baines, M., Cell. Immunol. 113, 261, 1988. 13. Clark, D. A., Lea, R. G., Flanders, K. C., Banwatt, D., and Chaouat, G., in ‘‘Progress in Immunology, VIII’’ (Gergely, J., et al., Eds.), pp. 841–848. Springer Verlag, Budapest, 1992. 14. De Fougerolles, R., and Baines, M., J. Reprod. Immunol. 11, 147, 1987. 15. Starkey, P. M., Sargent, L. L., and Redman, C. W. G., Immunology 65, 129, 1988. 16. Croy, B. A., Gambel, P., Rossant, J., and Wegmann, T. G., Cell. Immunol. 93, 315, 1985. 17. Croy, B. A., Waterfield, A., Wood, W., and King, G. J., Cell. Immunol. 115, 471, 1988. 18. Chao, K-H., Yang, Y-S., Ho, H-N., Chen, S-U., Chen, H-F., Dai, H-J., Huang, S-C., and Gill, T. J., III, Am. J. Reprod. Immunol. 34, 274, 1995. 19. Lachapelle, M.-H., Miron, P., Hemmings, R., and Roy, D. C., J. Immunol. 156, 4027, 1996. 20. Smart, Y. C., and Burton, R., Today’s Life Sci. 5, 14, 1993.
cia
ANTIABORTIVE EFFECT OF PIBF 21. Clarke, G. R., Roberts, T. K., and Smart, Y. C., Immunol. Cell Biol. 72, 153, 1994. 22. Koo, G. C., and Peppard, J. R., Hybridoma 3, 301, 1984. 23. Ryan, J. C., Truck, N., Niemi, E. C., Yokoyama, W. M., and Seaman, W. E., J. Immunol. 149, 1631, 1992. 24. Smart, Y. C., Stevenson, K. L., Farrelly, M. L., Brien, J. H., and Burton, R. C., Immunol. Cell. Biol. 68, 277, 1990. 25. Halmgreen, S. P., Wang, X., Clarke, G. R., Nolthorp, R. S., Roberts, T. K., Burton, R. C., Robinson, J. P., and Smart, Y. C., J. Immunol., in press. 26. Martinello, R., Smart, Y. C., and Burton, R. C., Cell. Immunol. 156, 155, 1994.
AID
CI 1090
/
6c1f$$$$$3
04-22-97 22:57:56
199
27. Martinello, R., Burton, R. C., and Smart, Y. C., Int. J. Cancer, in press. 28. Szekeres, J., Pacsa, A. S., and Pejtsik, B., J. Immunol. Methods 40, 151, 1981. 29. Grimm, E., and Bonavida, B., J. Immunol. 123, 2861, 1979. 30. Haddad, E. K., Duclos, A. J., and Baines, M. G., J. Exp. Med. 182, 1143, 1995. 31. Ortaldo, J. R., Mason, L. H., Mathieson, B. J., Liang, S. M., Flick, D. A., and Herberman, R. B., Nature 321, 700, 1986. 32. Chaouat, G., Menu, E., Clark, D. A., Dy, M., and Wegmann, T. G., J. Reprod. Fertil. 89, 447, 1990. 33. Clark, D. A., Chaouat, G., Mogil, R., and Wegmann, T. G., Cell Immunol. 154, 143, 1994.
cia
177, 194–199 (1997)
CI971090
The Antiabortive Effect of Progesterone-Induced Blocking Factor in Mice Is Manifested by Modulating NK Activity JULIA SZEKERES-BARTHO, G. PAR, GY. DOMBAY, Y. C. SMART,*
AND
Z. VOLGYI
Department of Microbiology, University Medical School of Pecs, H-7643 Pecs, Hungary; and *Discipline of Surgical Sciences, Faculty of Medicine, University of Newcastle, Newcastle, Australia Received October 29, 1996; accepted January 26, 1997
Immunologic effects of progesterone are mediated by a protein named the progesterone-induced blocking factor (PIBF), which inhibits NK activity and displays an antiabortive effect in mice. Our previous data provide indirect evidence for the importance of PIBF in the maintenance of normal gestation. This study was aimed at investigating whether neutralization of endogenous PIBF production influences pregnancy outcome and if so, what are the mechanisms that participate in this process. Syngeneically pregnant Balb/c mice on Day 8.5 of pregnancy were injected ip with 0.3 mg/kg of RU 486 or with 0.5 mg of rabbit anti-PIBF IgG alone, or together with anti-NK monoclonal antibodies. Mice treated with the same amount of normal rabbit serum or untreated mice of similar gestational age were used as controls. On Day 10.5 the ratio of living and resorbed embryos and NK activity of the spleen cells were determined. In mice treated with anti-PIBF the ratio of resorbed fetuses was significantly higher than that in untreated controls. In RU 486-treated mice we also observed significantly increased resorption rate, which was associated with the inability of spleen cells to produce PIBF. Both antiPIBF treatment and that with progesterone receptor blocker resulted in increased splenic NK activity. There was a positive relationship between NK activity and the rate of resorptions. All the above effects were corrected by simultaneous treatment with anti-NK or anti-NC (natural cytotoxic) antibodies. These data allow the conclusion that PIBF contributes to normal gestation in mice and that the effect of PIBF is manifested via blocking NK and/or NC activity. q 1997 Academic Press
INTRODUCTION Progesterone-dependent immunomodulation is one of the mechanisms that enables pregnancy to proceed to term. Immunologic effects of progesterone are mediated by a protein named the progesterone-induced blocking factor (PIBF) (1). Lymphocytes of healthy
194
0008-8749/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
CI 1090
/
6c1f$$$$$1
pregnant women are able to produce PIBF and the percentage of PIBF-positive cells is significantly reduced in peripheral blood lymphocytes (PBL) of recurrent aborters (2). Among other effects this protein inhibits NK activity (3) and displays an antiabortive effect in mice (4). Earlier data show that adoptive transfer of high NK activity spleen cells into pregnant mice induces abortion (5). Simultaneous PIBF treatment of pregnant mice corrects the abortive effect of NK activity (6). The above data provide indirect evidence for the importance of PIBF in maintaining normal gestation. Direct evidence for the biological significance of PIBF would be the induction of pregnancy loss by neutralization of endogenous PIBF. There are data suggesting the role of high NK activity in pregnancy loss in both mice and humans. In normal human pregnancy, NK activity is significantly lower than in nonpregnant individuals, whereas spontaneous pregnancy termination is associated with increased systemic NK activity (7–11). In mice there is direct evidence for the role of high NK activity in pregnancy termination. Resorption sites in mice are infiltrated by NK cells (12). In the abortion-prone CBA/J 1 DBA/2 murine mating combination, spleen cells from CBA/J mice contain asialo GM1/ effector cells that can kill the Be6 trophoblast cell line in vitro (13). Modulation of NK activity influences resorption rates (14). Natural killer cells are recruited into the pregnant uterus during early pregnancy in humans, mice, and pigs (15–17). In pregnant women, although the relative proportion of decidual NK cells is increased, decidual NK activity is lower in normal than in anembryonic pregnancies and in recurrent spontaneous abortions (18). Lachapelle et al. (19) have demonstrated an increased proportion of CD16/ CD56/ cells in the endometrium of recurrent aborters. NK activity is one major component of natural cellmediated cytotoxicity (NCMC) with natural cytotoxic (NC) activity being the second component (20). Limited data indicate that NC activity is present in the decidua of pregnant mice (21). Its role in pregnancy outcome has not been investigated.
04-22-97 22:57:56
cia
ANTIABORTIVE EFFECT OF PIBF
Since PIBF is antiabortive in mice and inhibits NK activity in vitro, the present study was aimed at investigating whether neutralization of endogenous PIBF activity in vivo results in pregnancy termination and if it does, what are the mechanisms that lead to this event. MATERIALS AND METHODS Treatment of Pregnant Mice by RU 486, Anti-PIBF Anti-NK, and Anti-NC Antibodies Fourteen-week-old Balb/c mice (LATI, Godollo, Hungary) were kept under standard conditions (4 animals per cage). Female mice were caged overnight with the males and checked for the presence of vaginal plugs the following morning. The day on which the plug was observed is considered to be Day 0.5 of pregnancy. Pregnancy was later verified by scoring corpora lutea. Various treatments were administered on Day 8.5 of pregnancy. (1) Females were injected intraperitoneally with 0.3 mg/kg of RU 486 or with 0.5 mg of rabbit anti-PIBF IgG. Mice treated with the same amount of normal rabbit serum or untreated mice of similar gestational age were used as controls. (2) A group of anti-PIBF-treated mice was at the same time injected with monoclonal antibodies to cells mediating natural killer (NK), natural cytotoxic (NC), or natural T cell (NT) activity. The monoclonal antibodies were: (a) PK136 (anti-NK-1.1) recognizes a 76- to 80kDa type II integral membrane C-type lectin protein encoded by a member of the mouse NKR-PI gene family (22, 23). (b) 1C4 (anti-NC-1.1) which recognizes a 45-kDa surface receptor and blocks splenic NC activity approximately 70% both in vitro and in vivo (24, 25). (c) 2B6-F2 (anti-Ly-6c) identifies a subpopulation of murine Ly-6c/ NK1.1/ natural T (NT) cells (26, 27). In the presence of complement 2B6-F2 reduces splenic NK activity by approximately 50% in Balb/c mice. The antibodies 2B6-F2 and PK136 were used as serum-free supernatants and administered at a dose of 50 mg/mouse, whereas 1C4 was affinity-purified IgG and administered at a dose of 25 mg/mouse. The animals were sacrificed on Day 10.5 and their uteri were inspected. The ratio of living versus resorbed embryos was determined. Spleens were removed under aseptic conditions, cell suspension was prepared, and cytotoxic activity was determined in a 16-hr test against human embryonic fibroblast targets. NK activity was also tested using the 4-hr single-cell cytotoxicity assay against YAC targets. Smears were prepared and the expression of PIBF on the lymphocytes was determined by immunocytochemistry.
AID
CI 1090
/
6c1f$$$$$2
04-22-97 22:57:56
195
16-hr Cytotoxicity Assay This technique has been described elsewhere in detail (28). Briefly, human embryonic fibroblasts derived from 10- to 12-week embryos were used as targets. Cells were seeded on 96-well Nuclon tissue culture plates at a density of 5000 target cells/well in 0.2 ml of medium 199 supplemented with 10% fetal calf serum. The target cells were allowed to attach by overnight incubation. The following day the medium was replaced by 0.2 ml of lymphocyte suspension containing 5 1 105 lymphocytes in 0.2 ml of the same medium. After 16 hr of incubation, the plates were washed with PBS three times in order to remove lymphocytes and damaged target cells. This was followed by addition of alkaline phosphatase substrate (Sigma tablets No. 104) in diethanolamine buffer at a concentration of 1 mg/ ml. The plates were incubated for 10 min at 377C in the dark and the resulting yellow reaction product was quantitated photometrically at 405 nm. The percentage reduction in enzyme activity relative to the target cell control was considered as a measure of cytotoxicity. 4-hr Single Cell Cytotoxicity Assay for NK Activity We used the assay originally described by Grimm and Bonavida (29). One hundred milliliters of lymphocytes and the same amount of YAC target cells (2 1 106/ml each) were centrifuged at 40g for 5 min, and incubated at 377C, in 5% CO2 for 10 min. The pellets were then resuspended and 100 ml of 1% agarose (Serva) in RPMI 1640 (Gibco) was added to the mixture. One hundred milliliters of this suspension was spread over microscope glass slides previously coated with 1% agar. Target cells alone were used to detect spontaneous lysis. The gel was allowed to solidify and submerged in RPMI 1640. The slides were incubated for 4 hr at 377C in 5% CO2 . Then the gels were stained with 0.5% trypan blue for 1 min. After 2-min washes with PBS, the gels were fixed in 2% formaldehyde for 5 min and desalted in distilled water. The slides were read using light microscope with 4001 magnification. The proportion of lymphocytes bound to the target cells was expressed as a percentage of total lymphocyte population by counting 300 lymphocytes. Results are expressed as a percentage of target binding cells (TBC). Dead conjugates were scored as a percentage of the total number of conjugates by counting 50 conjugates and results are expressed as a percentage of dead conjugates (cytotoxic TBC%). The percentage of NK cells was calculated according to the formula NK% Å TBC% 1 cytotoxic TBC%/100. All results for cytotoxic TBC% were corrected for the proportion of target cells that died spontaneously in control plates. Production of the PIBF Spleen cells of 10-week-old Balb/c mice (LATI, Godollo) were adjusted to a concentration of 1 1 106/ml
cia
196
SZEKERES-BARTHO ET AL.
TABLE 1 Resorption Rates and Mean Number of Implantation Sites in Pregnant Mice Treated with Anti-PIBF and Anti-NK, Anti-NC, or Anti-NT Antibodies No. of fetuses Treatment
Living
Resorbed
Resorption rate %
No. of implants
Mean No. of implant sites
Viable fetuses/mother
None (N Å 45) Anti-PIBF (n Å 83) Anti-PIBF / PK136 (N Å 13) Anti-PIBF / 1C4 (N Å 5) Anti-PIBF / 2B6-F2 (N Å 11) Anti-PIBF / Anti-NK (N Å 29)
326 193 77 25 89 188
27 113 7 2 13 22
7.6 37a 8.3 8.3 12.7 10.4
353 133 84 27 102 210
7.6 3.6*** 6.5 4.8 9.3 7.2
7.2 2.3*** 5.9 4.4 8.1 6.5
Note. Analysis of variance showed no difference between resorption rates in anti-PIBF / anti-NK/NC-treated mice and in untreated controls, whereas anti-PIBF / anti-NK/NC-treated groups as well as untreated controls differed significantly from anti-PIBF-treated mice. a Significantly different from untreated controls at P õ 0.001 (x2 test). *** Significantly different from untreated controls at P õ 0.001 (Student’s t test).
in RPMI supplemented with 10% FCS (both from Gibco) and stimulated with 1 mg/ml of Con A (Sigma) for 48 hr at 377C in a CO2 incubator. The cell number was then adjusted to 10 1 106/ml and the cells were further incubated with 20 mg/ml of progesterone for 16 hr. At the end of the incubation period supernatants were collected. Progesterone was removed by dialysis. The supernatants were then 2000-fold concentrated on Amicon filters and used as the source of the murine PIBF. Production of PIBF-Specific IgG Spleen cells of pregnant mice were treated with 20 mg/ml of progesterone overnight. The supernatants were collected, concentrated on Amicon filters, and subjected to SDS–PAGE on 12% polyacrylamide gels. The separated bands were blotted to nitrocellulose filters, the 34-kDa band was cut out, dissolved in DMSO, and injected into rabbits weighing 4 kg each, together with complete Freund’s adjuvant. Boosters with incomplete Freund’s adjuvant were given at two weekly intervals. IgG was purified on protein A columns. The PIBF-specific antibody content was tested by ELISA. ELISA Test for Detection of PIBF in Rabbit Serum Dynatech ELISA plates were coated with 10 mg/ml of PIBF per well in 100 ml of coating buffer (carbonate– bicarbonate buffer, 0.05 M, pH 9.6). After overnight incubation at 47C plates were washed once and blocked with gelatin–BSA for 30 min at room temperature. Anti-PIBF IgG was added at different dilutions. After 1 hr of incubation plates were washed three times and HRPO-labeled anti-rabbit IgG (Sigma) was added to the wells at a dilution of 1/5000. After 30 min of incubation the plate was washed three times and the reaction was stopped by addition of 50 ml of 2 N HCl and extinctions were read at 495 nm.
AID
CI 1090
/
6c1f$$$$$2
04-22-97 22:57:56
Immunocytochemical Detection of PIBF Spleen cells were washed once, the cells count adjusted to 1 1 106/ml and centrifuged onto glass microscope slides. After drying at room temperature, the cells were fixed for 5 min in cold acetone and washed with phosphate-buffered saline (PBS). Anti-PIBF was added at 10 mg/ml and incubated for 1 hr at room temperature in a humidified atmosphere. Peroxidase-labeled anti-rabbit IgG (Dakopatts, Hungary) at a dilution of 1:200 was then added and the cells were incubated for an additional 45 min. The reaction was developed with aminoetylcarbasol, nuclei were counterstained with hematoxylin, and the slides were mounted with gelatin–glycerol. The percentage of positive cells was determined by light microscopy and 300 cells were counted. IgG from nonimmunized rabbits gave negative results and the reaction of the lymphocytes with anti-PIBF was blocked by preincubation of anti-PIBF IgG with PIBF. Statistical Analysis The two-tailed Student t test and the x2 test as well as one-way analysis of variance with the Bonferroni correction were used for statistical analysis of the data. Mean { SEM are indicated in the table and figures. Differences were considered to be significant if P was equal to or less than 0.05. RESULTS The Lack of PIBF Effect Results in Pregnancy Termination Balb/c mice on Day 8.5 of pregnancy were injected ip with 0.3 mg/kg of RU 486 or with anti-PIBF IgG. Mice injected with saline or normal rabbit serum served as controls. On Day 10.5 the ratio of living versus resorbed fetuses was significantly lower (P õ 0.001) in mice
cia
ANTIABORTIVE EFFECT OF PIBF
197
treated with anti-PIBF compared to untreated controls (Table 1). This could not be due to endotoxin contamination, since no endotoxin could be detected in antiPIBF IgG with the Limulus bioassay (Dialab Gmbh. Austria). In RU 486-treated mice we also observed a significantly increased (P õ 0.001) resorption rate (Fig. 1a), which was associated with the inability of spleen cells to produce PIBF (Fig. 1b). The Lack of PIBF Effect Results in Increased NK and NC Activity On Day 10.5, the animals were sacrificed and their splenic NK and NC activity was determined. Both antiPIBF and RU 486 treatments resulted in increased NK activity of the spleen cells (Fig. 1c). Similar results were obtained in the 16-hr cytotoxicity test (data not shown). There was a positive relationship between NK activity and the rate of resorptions. Anti-PIBF treatment did not in all cases result in pregnancy termination; 23% of mice did not respond. In nonresponders, NK activity was significantly lower than that in good responders (Fig. 2). The Antiabortive Effect of PIBF Is Mediated via Controlling NK Activity Pregnant mice on Day 8.5 of pregnancy were injected with anti-PIBF and simultaneously with anti-NK-1.1, anti-NC-1.1, or anti-Ly-6c monoclonal antibodies. AntiPIBF-treated mice showed a significantly increased (37%) resorption rate and a reduced (55%) number of implantation sites (Table 1). Mice simultaneously treated with anti-PIBF and anti-NK-1.1 (PK136) antiNC-1.1 (1C4), or anti-Ly-6c (2B6F2) did not show significantly altered resorption rates and mean numbers of implantation sites compared to untreated controls with the exception of anti-NC-1.1 treatment, where the mice did not demonstrate alteration in the number of implantation sites (Table 1). DISCUSSION In this study we have shown that neutralization of endogenous PIBF activity results in pregnancy loss in mice. Both progesterone receptor block and anti-PIBF treatment induced an increased rate of resorptions, and in the former case pregnancy loss was due to the inability of spleen cells to produce PIBF. However, antiPIBF treatment did not in all cases result in pregnancy termination; 23% of mice did not respond. In nonresponders, NK activity was significantly lower than that in good responders. It is conceivable that in these mice abortion did not occur because the spleen cells of the animals produced more than normal levels of PIBF, which was not all neutralized by the addition of antiPIBF and abortion was not induced. Clearly both treatments resulted in increased splenic NK activity, and
AID
CI 1090
/
6c1f$$$$$3
04-22-97 22:57:56
FIG. 1. (a) Progesterone receptor block induces increased resorption rates. ***P õ 0.001 (x2 test). (b) The effect of progesterone receptor block on splenic PIBF production. The bars represent the mean { SEM. ***P õ 0.001 (Student’s t test). (c) The effect of anti-PIBF treatment on splenic NK activity. The bars represent the mean { SEM. ***P õ 0.001 (Student’s t test).
the rate of NK activity correlated with the rate of resorptions. Nonspecific immunological mechanisms including
cia
198
SZEKERES-BARTHO ET AL.
pregnancy outcome probably acting at least in part on PIBF. The third monoclonal antibody (2B6-F2) reduces both NK and NC activity in Balb/c mice (26), and corrected both high resorption rates and low implantation rates due to the lack of PIBF. Our data suggest that both NC and NK activity are involved in pregnancy termination in Balb/c mice, and that PIBF exerts its antiabortive effect via inhibition of nonspecific effector mechanisms. ACKNOWLEDGMENTS This work was supported by grants from the Hungarian National Research Fund and the Hungarian Ministry of Welfare. FIG. 2. The relationship between splenic NK activity and resorption rates. The bars represent the mean { SEM. ***P õ 0.001 (Student’s t test).
NK and NC activities are probably the most important effector pathways in the fetomaternal immunological relationship. NK activity has been shown to play an important role in antitumor defense as well as in defense against intracellular parasites. There is also evidence for its role in spontaneous pregnancy termination (12–14). Haddad et al. (30) suggested that the role of NK cells in abortion in mice is to produce cytokines that activate macrophages in the uterus and macrophage-derived nitric oxide is responsible for the damage. Previously we showed that PIBF had a marked inhibitory effect on NK activity in vitro (3). NK activity is a major factor in spontaneous abortions in mice. In murine systems increased NK activity results in pregnancy termination (5), which can be corrected by simultaneous PIBF administration (6). Treatment of antiPIBF-treated mice with the anti-NK-1.1 monoclonal antibody (PK 136) corrected the resorption rates and the mean number of implantation sites. Treatment of anti-PIBF-treated mice with anti-NC1.1 (1C4) completely corrected the resorption rates, and spleen cells of anti-PIBF-treated mice exerted an increased cytotoxicity after 16 hr incubation with the target cells. In the cytotoxicity assay based on determination of endogenous alkaline phosphatase activity of the target cells we used human embryonic fibroblasts instead of the classical NC target WEHI-164. However, the requirement for a 16-hr cytotoxicity points to NCmediated activity in this system. The major mechanism of NC activity is TNFa (31). Injection of TNFa into mice resulted in pregnancy termination, while TNFa antibodies or antagonists corrected the spontaneously high resorption rates in CBA/J 1 DBA/2 matings (32). GM-CSF prevents abortion in the above system by downregulating splenic NC activity (33). Taken together these results suggest a role for NK or NC activity in
AID
CI 1090
/
6c1f$$$$$3
04-22-97 22:57:56
REFERENCES 1. Szekeres-Bartho, J., Kilar, F., Falkay, G., Csernus, V., Torok, A., and Pacsa, A. S., Am. J. Reprod. Immunol. Microbiol. 9, 15, 1985. 2. Szekeres-Bartho, J., Faust, Zs., and Varga, P., Am. J. Reprod. Immunol. 34, 342, 1995. 3. Szekeres-Bartho, J., Autran, B., Debre, P., Andreu, G., Denver, L., and Chaouat, G., Cell. Immunol. 122, 281, 1989. 4. Szekeres-Bartho, J., and Chaouat, G., Am. J. Reprod. Immunol. 23, 26, 1990. 5. Kinsky, R., Delage, G., Rosin, N., Thang, M. N., Hoffmann, M., and Chaouat, G., Am. J. Reprod. Immunol. 23, 73, 1990. 6. Szekeres-Bartho, J., Kinsky, R., and Chaouat, G., Am. J. Reprod. Immunol. 24, 105, 1990. 7. Szekeres-Bartho, J., Varga, P., and Pacsa, A. S., Am. J. Obstet. Gynecol. 155, 108, 1986. 8. Coulam, C. B., Goodman, C., Roussev, R. G., Thomason, E. J., and Beaman, K. D., Am. J. Reprod. Immunol. 33, 40, 1995. 9. Coulam, C. B., and Beaman, K. D., Am. J. Reprod. Immunol. 34, 219, 1995. 10. Kwak, J. Y. H., Rushabrook, J. I., Hurley, I. R., Mandel, F. S., Barcia, M., Cooper, G. W., and Hershlag, A., Am. J. Reprod. Immunol. 34, 100, 1995. 11. Aoki, K., Kajura, S., Matsumoto, Y., Ogasawar, M., Okada, A., Yagami, Y., and Gleicher, N., Lancet 45, 1340, 1995. 12. Gendron, R., and Baines, M., Cell. Immunol. 113, 261, 1988. 13. Clark, D. A., Lea, R. G., Flanders, K. C., Banwatt, D., and Chaouat, G., in ‘‘Progress in Immunology, VIII’’ (Gergely, J., et al., Eds.), pp. 841–848. Springer Verlag, Budapest, 1992. 14. De Fougerolles, R., and Baines, M., J. Reprod. Immunol. 11, 147, 1987. 15. Starkey, P. M., Sargent, L. L., and Redman, C. W. G., Immunology 65, 129, 1988. 16. Croy, B. A., Gambel, P., Rossant, J., and Wegmann, T. G., Cell. Immunol. 93, 315, 1985. 17. Croy, B. A., Waterfield, A., Wood, W., and King, G. J., Cell. Immunol. 115, 471, 1988. 18. Chao, K-H., Yang, Y-S., Ho, H-N., Chen, S-U., Chen, H-F., Dai, H-J., Huang, S-C., and Gill, T. J., III, Am. J. Reprod. Immunol. 34, 274, 1995. 19. Lachapelle, M.-H., Miron, P., Hemmings, R., and Roy, D. C., J. Immunol. 156, 4027, 1996. 20. Smart, Y. C., and Burton, R., Today’s Life Sci. 5, 14, 1993.
cia
ANTIABORTIVE EFFECT OF PIBF 21. Clarke, G. R., Roberts, T. K., and Smart, Y. C., Immunol. Cell Biol. 72, 153, 1994. 22. Koo, G. C., and Peppard, J. R., Hybridoma 3, 301, 1984. 23. Ryan, J. C., Truck, N., Niemi, E. C., Yokoyama, W. M., and Seaman, W. E., J. Immunol. 149, 1631, 1992. 24. Smart, Y. C., Stevenson, K. L., Farrelly, M. L., Brien, J. H., and Burton, R. C., Immunol. Cell. Biol. 68, 277, 1990. 25. Halmgreen, S. P., Wang, X., Clarke, G. R., Nolthorp, R. S., Roberts, T. K., Burton, R. C., Robinson, J. P., and Smart, Y. C., J. Immunol., in press. 26. Martinello, R., Smart, Y. C., and Burton, R. C., Cell. Immunol. 156, 155, 1994.
AID
CI 1090
/
6c1f$$$$$3
04-22-97 22:57:56
199
27. Martinello, R., Burton, R. C., and Smart, Y. C., Int. J. Cancer, in press. 28. Szekeres, J., Pacsa, A. S., and Pejtsik, B., J. Immunol. Methods 40, 151, 1981. 29. Grimm, E., and Bonavida, B., J. Immunol. 123, 2861, 1979. 30. Haddad, E. K., Duclos, A. J., and Baines, M. G., J. Exp. Med. 182, 1143, 1995. 31. Ortaldo, J. R., Mason, L. H., Mathieson, B. J., Liang, S. M., Flick, D. A., and Herberman, R. B., Nature 321, 700, 1986. 32. Chaouat, G., Menu, E., Clark, D. A., Dy, M., and Wegmann, T. G., J. Reprod. Fertil. 89, 447, 1990. 33. Clark, D. A., Chaouat, G., Mogil, R., and Wegmann, T. G., Cell Immunol. 154, 143, 1994.
cia