Maternal T cells promote placental growth and prevent spontaneous abortion

Maternal T cells promote placental growth and prevent spontaneous abortion

Immunology Letters, 17 (1988) 297-302 Elsevier IML01035 Minireview Maternal T cells promote placental growth and prevent spontaneous abortion T h o...

537KB Sizes 1 Downloads 76 Views

Immunology Letters, 17 (1988) 297-302

Elsevier IML01035

Minireview

Maternal T cells promote placental growth and prevent spontaneous abortion T h o m a s G. W e g m a n n Department of Immunology, University of Alberta, Edmonton, Alberta, Canada

(Received 1 February 1988; accepted2 February 1988)

1. Summary

2. The nature of the fetal trophoblast barrier

Transplantation immunologists have long been intrigued by the natural allograft that results from normal mammalian pregnancy. Its general success contrasts with the rejection problems associated with most artifactual organ transplantation and raises intriguing questions concerning the nature of the mechanisms involved in that success. This area of research has recently taken on added momentum because it is now clear that immunological maneuvers can prevent recurrent spontaneous abortion in mice, horses and humans [1- 3]. The purpose of this review is to discuss some of these recent developments, which lead to the surprising conclusion that maternal T cells, rather than being potentially detrimental to the fetal allograft, promote its growth and viability during normal pregnancy. This review will address these questions by considering: (a) the nature of the exposure of fetal alloantigens to the maternal circulation in the chimeric zone of the placenta; (b) the evidence for maternal immune recognition of the fetal alloantigens, and (c) the consequences of that recognition with respect to the prevention of spontaneous abortion. As in other areas of immunology the T cell emerges as the most important component of this immune recognition.

Classic experiments of Simmons and Russell indicated that trophoblast tissue which is destined to serve as the cellular interface between the maternal and fetal circulations has special properties when used as an allograft. They showed that trophoblast tissue from a 7 ½ day old mouse embryo is not rejected when transplanted allogeneically, while embryonic tissue from the same conceptus undergoes standard graft rejection, even when placed in the same vicinity as the trophoblast [4]. For a long time this was believed to be because the trophoblast does not express major histocompatibility complex (MHC) Class I or Class II antigens, but by now it is clear that this is only true for early trophoblast. Class I (but not Class II) MHC antigens are expressed on midgestational trophoblast tissue which is in direct contact with maternal circulation [5]. In the mouse this type of trophoblast is called spongiotrophoblast, and is the invasive trophoblast which penetrates into the maternal uterine wall and interdigitates with the uterine decidual reaction of the mother, forming the chimeric zone of the placenta. This type of trophoblast eventually replaces the vascular endothelium of the maternal arteries and even metastasizes into the lungs of the female. Recently Head and her colleagues [6] have been able to isolate this trophoblast in pure form in tissue culture and examine its properties with respect to antigenicity and susceptibility to cell-mediated lysis. Although these isolated trophoblast cells express MHC Class I antigens, and this expression can be increased by treatment with gamma-interferon, the cells remain impervious to cell-mediated lysis, either by specific cytotoxic lym-

Key words: T cell;Maternalrecognition;Abortion,spontaneous Correspondence to: T. G. Wegmann,Dept. of Immunology,8-60

MedicalSciencesBuilding,Universityof Alberta,Edmonton,Alberta, Canada T6G 2H7.

0165-2478 / 88 / $ 3.50 © 1988Elsevier SciencePublishers B.V.(BiomedicalDivision)

297

phocytes or natural killer cells. Furthermore, these cells can serve as M H C antigen-specific cold target inhibitors of cytotoxic lymphocytes otherwise capable of lysing susceptible target cells [6]. The conclusion from this type o f experimentation is that the trophoblastic barrier erected by the embryo at the f e t a l - m a t e r n a l interface is inherently capable of avoiding the potentially harmful consequences o f maternal cell-mediated immunity, and this is known as the placental barrier hypothesis [7].

3. Maternal immune recognition of the fetus As one might expect from the above description, the presence of histoincompatible invasive trophoblast does not escape the attention of the maternal immune system. It has long been appreciated that anti-paternal M H C antibodies are frequently produced as a result of allogeneic pregnancy. For example, multiparous human serum has been a prime source of H L A typing sera for many years. Histological evidence of local maternal immune recognition of the fetus is clearly present during pregnancy, and more so when the fetus is allogeneic to the mother compared to when it is syngeneic. Thus in the former instance there is a more pronounced influx o f maternal T lymphocytes into the decidual reaction o f the uterus [8], and this occurs shortly after the expression of fetal M H C Class I antigens on the surface of the invasive trophoblast [9]. As one might expect, the lymph nodes draining the uterus are larger in an allogeneic than in a syngeneic pregnancy. In addition, there are higher levels o f macrophage inhibition factor produced by ceils of the uterine draining nodes of an allogeneically pregnant rat when compared to nodes from a syngeneically pregnant rat. These few selected examples, drawn from a rather large literature [5], indicate beyond doubt that maternal immune recognition o f fetal antigens occurs during pregnancy. Although one can find evidence for systemic suppressor cells in pregnant animals, they clearly are not the reason for successful allogeneic pregnancy because their effects can be over-ridden by paternal cell immunization. This treatment leads to the priming o f maternal T cellmediated cytotoxic cells in situ with no harmful consequence to the fetus [7].

298

4. White blood cdls related to paternal type can prevent abortion Recent evidence indicates that maternal T cells, rather than having potentially harmful consequences, exert a positive influence on placental growth, phagocytosis, and fetal survival. This work had its origins in human clinical observations o f an anecdotal sort. In the early 1980s, two groups of clinical investigators began treating women suffering from chronic spontaneous abortion o f unknown etiology with either their husband's or pooled third party lymphocytes. They reported that this led to the birth o f viable offspring [10, 11]. Unfortunately the lack o f proper double-blind clinical controls made it impossible to rule out placebo effects. At about the same time, Clark and his colleagues developed an animal model of spontaneous fetal resorption. CBA female mice pregnant by DBA/2 males show an unusually high frequency of resorbing fetuses when examined at mid-gestation [12]. The incidence of fetal resorption varies dramatically, depending upon the age of the female and a variety of other factors. The usual incidence reported has been between 30 to 50 percent, whereas most mouse strain combinations, including CBA x CBA, DBA/2 x DBA/2, DBA/2 x CBA/J, C3H x DBA/2, and CBA x BALB/c show an incidence of 10 percent or less [13]. In 1983, my colleagues and I reported that the CBA x DBA/2 fetal resorption rate would return to normal levels (approx. 1007o)by immunizing with ceils bearing the paternal M H C haplotype. Curiously, the paternal cells themselves had no effect, but BALB/c ceils, which share the M H C H-2 d haplotype with DBA/2, lead to a dramatic increase in fetal survival when injected seven days prior to conception [14]. This effect is MHC haplotype-dependent, because immunization with congenic BALB/k or BALB/b cells has no effect [15]. An interesting observation in this regard is that if a CBA female undergoes a prior pregnancy by BALB/c this can prevent abortion in a subsequent pregnancy where the father is DBA/2 [13]. Thus priming to prevent abortion can be done either by lymphocytes or prior pregnancy. Since these observations were made using a mouse model, one could begin to address the immunological consequences of preventing abortion. The first significant observation was that the immune system o f CBA females appears to be incapable o f reacting

to an injection of DBA/2 spleen cells. Thus the CBA females produce no antibody in response to a single immunization by DBA/2. However, they produce abundant antibody when immunized with BALB/c cells, and the CBA anti-BALB/c antibody, once produced, can be absorbed out by DBA cells, indicating that the DBA cells are antigenic (for H-2 d haplotype antigens) without being immunogenic [16]. A second observation was that CBA x DBA/2 fetal resorption could be prevented by immunizing the CBA female passively with CBA anti-BALB/c antiserum prior to pregnancy, and this effect could be removed by absorbing the antiserum with cells bearing the relevant H-2 d haplotype [16]. Immune cells from CBA immunized with BALB/c could also lower the frequency of fetal resorption to normal levels upon adoptive transfer [17]. A particularly seminal observation was that successfully immunizing to prevent abortion, either actively or passively, led to significant increases in both placental and fetal size in the viable feto-placental units [17]. This led me to postulate that maternal T cells recognize fetal alloantigens at the maternalfetal interface, which leads to local T cell-mediated lymphokine secretion. This in turn leads to a tropho.blast growth and the improvement of placental function [18]. This idea, which has been called the placental immunotrophism hypothesis, is not without precedent. It is known that T cells participate in such phenomena as hematopoiesis, liver regeneration, kidney hypertrophy following unilateral nephrectomy, and a variety of other growth phenomena [19]. Indeed, this idea had already been suggested in outline form in the early days of reproductive immunology [20], but the results obtained that favored it at that time were not clear enough to survive criticism [21] and the concept of T cells and their lymphokines had not yet emerged to lend credence to the speculation. Thereafter, two groups of investigators reported that when human lymphocytes are added to trophoblast cultures in vitro, this leads to an increased (rather than the anticipated decreased) release of human chorionic gonadotrophin secretion by the trophoblast cells [22, 23]. Although these investigators did not allude to the interpretation mentioned above, the results certainly conform to what might be expected from the immunotrophism hypothesis. My colleagues and I decided to take a direct ap-

proach to examining whether this hypothesis can make testable predictions. An initial and obvious approach was to inquire whether T cell-derived lymphokines can promote trophoblast growth and differentiation in vitro. Investigators who work with murine trophoblast cells in vitro know that such cul~ tures only survive a couple of cell divisions before expiring. When supernatants from EL-4 T lymphoma cells that have been stimulated with phorbol myristic acetate were applied to trophoblast cells in culture, the cells grew vigorously and continuously. Upon analysing which components of the supernatant were most potent, we found that members of the CSF family of lymphokines had the most potent effects on trophoblast growth, including GM.CSF and IL-3. We also found that CSF.1, a cytokine known to be released from macrophages, among other tissues, is also a potent stimulator of trophoblastic cell growth [24]. The cells that grow out of the placenta in response to these lymphokines are fetal in origin as well as adherent. They display the typical microtubular pattern of staining expected of trophoblast cells, ie. they are cytokeratin-positive and vimentin-negative [26, 27]. Macrophages, on the other hand, are cytokeratin-negative and vimentin-positive. An unexpected feature of the lymphokine-driven trophoblast ceils is that virtually all of them stain positively with antibodies specific for the macrophage markers MAC-1 and F4-80 [26]. Thus they show characteristics of both trophoblast and macrophage cells. Lymphokine stimulation of these cells leads to differentiation of their ability to phagocytose latex particles in vitro [24]. By continuous lymphokine stimulation one can establish long term cell lines from such placental cells, and two lines, termed GGR and FRD, have been growing in the laboratory for over two years. They have some intriguing characteristics. They are Class I MHC positive and are definitely lymphokine dependent. In particular, they show dependency on lymphokines of the CSF family, as one might expect. In addition, they have the intriguing property of non-specifically stimulating a variety of T cell lines, some of which are antigen-specific, to proliferate in the absence of either antigen or antigen-presenting ceils or both. The trophoblast cells themselves do not need to be present to produce this phenomenon . Their supernatants do it as well [27]. Thus, the results of our in vitro work have con299

formed to the predictions of the immunotrophism hypothesis, because T cell-derived lymphokines can cause trophoblast cells to proliferate indefinitely. In parallel, Stanley and his colleagues have described that the uterine levels of CSE1 increase 1000-fold upon pregnancy [28] and this increase is hormonally dependent [29]. Others have described the presence of the C-fins oncogene product on tissues derived from the placenta [30|. The C-fins oncogene glycoprotein is a cell-surface receptor for CSE1, and thus a cytokine which is capable of producing placental growth and its receptor are both present at the maternal-fetal interface. Recently, we have observed that decidual cell supernatants contain growth factors which promote not only placental cell growth in the manner described above, but also contain growth factors for cell lines which are known to proliferate in response to GM.CSF and IL-3, thus providing a means to determine what lymphokines are released by the decidual cells during pregnancy. Decidual cells from allogeneic pregnancies have more of this activity than those from syngeneic pregnancies, and this activity is very much reduced if the female mice contributing the deciduae are first treated with anti-T cell monoclonal antibodies prior to decidual cell dissection. Most of this activity appears to be GM.CSF, because treatment of the decidua cell supernatants with anti-GM.CSF antibody completely abolishes its activity [31]. CSE1 activity is also present in these decidual supernatants [31], thus extending the observation of Stanley and his colleagues [28, 29]. These experiments provide additional confirmation of an essential prediction of the immunotrophism hypothesis, namely that T cell-derived lymphokines are secreted by decidual cells in direct contact with fetal trophoblast cells. It is intriguing that more activity is present in the decidual cells from aUogeneic versus syngeneic pregnancies. This correlates with the histological appearance of the two types of decidua with respect to lymphocyte influx, where a similar distinction has been made [8]. Another prediction made from the placental immunotrophism hypothesis, hinted at above, is that removal of maternal T cells during mid-gestation should negatively affect placental growth and function, as well as fetal survival. We have found that injecting anti-CD8 monoclonal antibody into C3H females pregnant by BALB/c mice at mid-gestation 300

leads to a two-fold reduction in placental proliferation in vivo, as well as a two-fold reduction in the capacity of the placenta to phagocytose latex beads [24]. Similar observations have been made with antiCD4 antibody, although treatment with this antibody does not affect the ability of the placenta to phagocytose the latex particles, but instead only affects placental proliferation [25]. In the strain combination employed in these experiments (C3H x BALB) there was no obvious effect of mid-gestational T cell removal on fetal viability. My French colleagues and I have therefore embarked on a series of experiments to determine what role maternal T cells play in the CBA x DBA/2 abortion model. If maternal T cells are crucial for preventing the abortion, their removal at midgestation should reverse the effect of immunizing with BALB/c spleen cells prior to pregnancy by DBA/2, and thus lead to a return to high fetal resorption rates, even though the females have antiH-2 d antibody in their system from the immunization. This is what we have found, and treating the females in such a manner also leads to a significant reduction in both placental and fetal weight in the surviving fetal placental units [32]. In addition, CBA females pregnant by BALB/c males, which other. wise have a low fetal resorption rate, show approximately 50% fetal resorption at day 14, if their T cells are deleted by monoclonal anti-T-ceU antibody treatment during mid-gestation. This treatment likewise causes a reduction in placental and fetal weight in the remaining offspring [32]. These results provide striking confirmation of predictions from the immunotrophism hypothesis and suggest that even in some cases of normal gestation, maternal T cells have a role to play in protecting the viability of the fetus. A number of questions remain to be addressed. The first concerns the cause of the fetal wastage in the CBA x DBA/2 mating combination. Recent experiments by Hamilton and Hamilton suggest an infectious etiology [33]. They obtained a batch of CBA female and DBA male mice from Jackson Laboratories and placed them in two separate rooms at the University of Washington. One group was housed in an ordinary animal room and showed the usual rate of fetal resorption, which could be reversed by BALB/c cell immunization. The second group was placed in a specific pathogen-free room and that

group showed no increase in the incidence o f fetal resorption. What sort of organisms might be involved in this situation? It is not clear from the current literature, but there have been reports among human chronic spontaneous aborters that the microplasmic organism Ureaplasma urealiticum can be found in the placenta it/a substantial number of these individuals. In addition, antibodies are found in the serum of these women and their fetuses [34]. It is known that Ureaplasma diversum is abortefacient for cattle [35]. Thus it remains possible that the abortion seen in the mouse model is microbial in nature, but only further experimentation can resolve this issue. It is interesting to note in this regard that immunizing CBA or C3H females with cells bearing the paternal M H C haplotype leads to an increase in placental phagocytosis [24]. In addition, the cells that grow out of the fetal placenta in response to T cell-derived lymphokines express a combination of trophoblast and macrophage markers [26, 27]. Thus it is not unreasonable to speculate that the antiabortion effect of paternal cell immunization stems from primed maternal T cells that improve the placenta's role as an immunologic barrier between the maternal and fetal circulations. This in turn helps to prevent maternal blood-borne infection from doing significant harm to the placenta and the fetus. This possibility remains to be demonstrated. Another area for future research will be the assembly of a detailed description o f the complex interactions that occur between decidual and trophoblast cells that result in the immunotrophic cycle. We know, as stated above, that removal of CD4-positive T cells has a different effect on placental function than CD8 T cell removal, but the interactions, no doubt, are complex and can only be ultimately resolved by cell cloning and analysis o f the ontogeny and kinetics o f individual cell interactions. This will no doubt take a great deal of effort to accomplish, but it will be necessary in order to understand how immune cells participate in placental and fetal growth. Finally, one also would like to know whether any significant immunotrophic interactions take place that affect the preimplantation embryo. There are suggestions that this might be the case but they need further evaluation. Many years ago, Beer and Billingham reported that stimulating one uterine horn o f a female rat with the spleen cells of her mate prior

to pregnancy led to an increased amount o f embryonic implantation in that uterine horn compared to the contralateral horn [36]. This could merely be due to expanded surface area for implantation in the stimulated horn, but could also be related to lymphokine or cytokine-mediated growth of t h e trophoblast. The latter possibility is suggested by experiments of Edidin and his colleagues who reported that peritoneal exudate cells from multiparous, but not virgin, females stimulated teratocarcinoma and blastocyst growth in vitro [37]. This raises the question of whether one can find lymphokine or cytokine receptors on preimplantation embryos, whether lymphokines or cytokines are capable of inducing preimplantation trophoblast growth as they appear to be for postimplantation trophoblast, and whether such treatment could lead to improved embryonic implantation in experimental animals, humans, or in commercially valuable livestock. Regardless of the answer to these relatively speculative questions, it is already clear that the observations outlined in this review article have clear implications, not only for current clinical practice dealing with human infertility, but also for understanding the basis of the growing clinical applications of this research area. By now, a large number of groups are actively pursuing paternal (or pooled third party) lymphocyte therapy for treating human chronic spontaneous abortion, and they are reporting dramatic results [38]. One trial has been published in which an attempt at double-blind design was made, with impressive results [39]. It still remains to rule out subtle placebo effects o f this treatment. Therefore, more. trials of this sort are underway, emphasizing even further the necessity for intensified research in the animal models which are capable of providing mechanistic insight into this type of clinical intervention.

References [l] Gill,T. J. III and Wegmann,T. G., eds. (1987)Immunoregulation and Fetal Survival,Oxford UniversityPress, London. [2] Chaouat, G., ed. (1987) Reproductive Immunology: Materno-FetalRelationship,ColloqueInsermVol. 154(Editions INSERM, Paris). [3] Clark, D. A. and Croy,B. A., eds. (1986)ReproductiveImmunology 1986, Elsevier,Amsterdam. [4] Simmons, R. (1969)Transplant. Proc. I, 47. 301

[5] Hunziker, R. D. and Wegmann, T. G. (1986) CRC Critical Rev. Immunol. 6, 245. [6] Head, J. R., Drake, B. L. and Zuckerman, E A. (1987) Am. J. Reprod. Immunol. 15, 12. [7] Chaouat, G., Kolb, J-P. and Wegmann, T. G. (1983) Immunol. Rev. 75, 31. [8] Lala, P. K., Parkar, R. S., Kearns, M., Johnson, S. and Scodras, J.M. (1986) in: Immunologic Aspects of the Decidual Response (D. A. Clark and B.A. Croy, eds.) pp. 190-198, Elsevier Science Publishers B.V., New York. [9] Raghupathy, R., Singh, B., Barrington-Leigh, J. and Wegmann, T. G. (1981) J. Immunol. 127, 2074. [10] McIntyre, J.A., McConnachie, P. R., Taylor,. C. G. and Faulk, W. P. (1984) Fertil. Steril. 42, 849. [11] Beer, A. E., Semprini, A. E., Xiaoya, A. and Quebbeman, J. E (1985) Exp. Clin. Immunogenet. 2, 1. [12] Clark, D.A., McDermott, M. R. and Szewczuk, M.R. (1980) Cell. Immunol. 52, 106. [13] Clark, D. A., Chaput, A. and "Patton, D. (1986) J. Immunol. 136, 1668. [14] Chaouat, G., Kiger, N. and Wegmann, T.G. (1983) J. Reprod. Immunol. 5, 389. [15] Kiger, N., Chaouat, G., Kolb, J-P., Wegmann, T. G. and Guenet, J. L. (1985) J. Immunol. 134, 2966. [16] Chaouat, G., Kolb, J.-P., Kiger, N., Stanislawski, M. and Wegmann, T. (3. (1985) J. Immunol. 134, 1594. [17] Chaouat, G., Kolb, J.-P., Chaffoux, S., Riviere, M., Lankar, D., Athanassakis, I., Green, D. and Wegmann, T. G. (1987) in: Immunoregulation and Fetal Survival (T. J. Gill III and T. G. Wegmann, eds.), pp. 246, Oxford University Press, New York. [18] Wegmann, T. G. (1984)Ann. Immunol. (Inst. Pasteur) 135D, 309. [19] Green, D. R. and Wegmann, T. G. (1986) in: Progress in Immunology VI (B. Cinader and R. (3. Miller, eds.) pp. 1100-1112, Academic Press, New York.

302

[20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39]

Clarke, B. and Kirby, D. R. S. (1966) Nature 211, 999. McLaren, A. 0975) Clin. Exp. Immunoreprod. l, 255. Dickman, W. J. and Cauchi, M. M. (1978) Nature 271, 377. Kaplan, L. (1983) ICRS Med. Sci. 11, 893. Athanassakis, I., Bleackley, R. C., Paetkau, V., Guilbert, L., Barr, P. J. and Wegmann, T. G. (1987) J. Immunol. 138, 37. Athanassakis, I. and Wegmann, T. G. 0986) in: Reproductive Immunology (D. Clark and B.A. Croy, eds.) pp. 95-105, Elsevier North Holland, Amsterdam. Athanassakis, I., Vassiliadis, S., Guilbert, L. and Wegmann, T. G. (1988) (submitted for publication). Mogil, R. and Wegmann, T. G. 0988) (submitted for publication). Bartocci, A., Pollard, J. W. and Stanley, E. R. (1986) J. Exp. Med. 164, 956. Pollard, J. W., Bartocci, A., Arceci, R., Orlofsky, A., Ladnet, M. B. and Stanley, E. R. 0987) Nature 330, 484. Rettenmeier, C.W., Sacca, R., Furman, W. L., Roussel, M. R., Holt, J., Nienhnis, A. W., Stanley, E. R. and Sheu, C. J. (1987) J. Clin. Invest. 77, 1740. Athanassakis, I., Branch, D. R., Garcia-Lloret, M. I., Guilbert, L. and Wegmann, T. G. (1988) (submitted for publication). Chaouat, G., Menu, E., Athanassakis, I. and Wegmann, T. G. (1988) (submitted for publication). Hamilton, M. S. and Hamilton, B. L. (1987) J. Reprod. Immunol. 11, 243. Quinn, P. A. 0986) Pediatr. Infect. Dis. 5, 5287. Miller, B., Ruhnke, H. L. and Doig, B. F. 0983) Theriogenology 20, 364. Beer, A. E. and Billingham, R. E. 0974) J. Reprod. Fertil. 21, 59. Fenderson, B. A., Bartlett, P. F. and Edidin, M. (1983) J. Reprod. Immunol. 5, 287. Clark, D. A. (1987) Amer. J. Reprod. Immunol. 15, 131. Mowbray, J. F. 0988) Amer. J. Reprod. Immunol. 15, 138.