Species diversity in the immunogenetic relationship between mother and fetus: is trophoblast insusceptibility to immunological destruction the only essential common feature for the maintenance of allogeneic pregnancy?

Maternal-fetal immune interactions in pregnancy have been examined in a number of different species using a wide range of experimental procedures, many of which have been incapable of providing information on the specificity of the immunological parameters measured. As a result, data have often been unconvincing, unsubstantiated or conflicting. It is clear that considerable species diversity exists in the precise nature of the maternal immune responses elicited against the genetically dissimilar fetus. These may, however, be an irrelevant consequence of fetomaternal genetic incompatibilities. A feature common to all eutherian mammalian species, that could alone be responsible for the maintenance of allogeneic pregnancy, is the apparent insusceptibility of the fetal trophoblast to immune destruction. This is effected in some species by an absence of appropriate target molecules, particularly classical class I MHC antigens, on the trophoblast cell surface, and in others possessing these (and possibly other) target structures, by the local trophoblastic secretion of soluble factors, not yet fully characterised, that appear to be capable of inhibiting binding and interaction with cells of the maternal immune system. As there is no convincing evidence for transplacental transfer of significant numbers of maternal lymphocytes, the survival of the allogeneic fetus is likely to be ensured by the impenetrability of the trophoblastic tissue barrier and its resistance to maternal immunologically mediated attack. Should this be firmly established, it would not be necessary to invoke any of the immunoregulatory mechanisms that have been claimed to be essential for successful implantation and development of the mammalian embryo. This would in turn negate the concept of an alloimmune aetiology for pregnancy failure and hence also the rationale for those therapeutic procedures intended to restore the assumed inadequate maternal alloimmune protective responses.

The normal fetomaternal immune relationship.

The antigenic status of the preimplantation embryo is ill-defined and there are no clearly recognized maternal immune reactions against this early stage of development. Following implantation, the pregnant female shows evidence of immune recognition of her intrauterine allogeneic conceptus. In a proportion of pregnancies, particularly in multiparous women, there are maternal cytotoxic antibodies exhibiting specificity for the paternally inherited HLA antigens of the fetus. When these are undetectable there may be other antibodies that are non-complement fixing and non-cytotoxic or antibodies that are not present as free molecules and incapable of identification in conventional assays. Anti-HLA antibodies pose no threat to the fetus, principally owing to their absorption by the placenta and, very likely, the harmless binding of any that do reach the fetal circulation. No potentially deleterious cytotoxic T lymphocyte generation occurs in most pregnancies. The extent to which this is due to maternal immunoregulatory control processes is not yet established. The fetal trophoblast is able to act as a protective barrier by virtue of unique properties, including a lack of conventional class I and class II HLA molecules, that render it insusceptible to immune attack. The nature and significance of any maternal recognition of non-HLA antigens on trophoblast await elucidation. Maternal immune cell traffic across the placenta occurs only at a very low level, if at all, in normal pregnancy. This may take place to a greater degree in some of the rare instances of fetal graft-versus-host disease, but this is complicated by the associated fetal immunodeficiency. Maternal IgG antibodies are transmitted across the placental trophoblast by receptor-dependent mechanisms to provide immediate protection for the neonate against environmental pathogens. Leakage of fetal erythrocytes, leukocytes and platelets into the maternal circulation can elicit IgG isoantibodies that take advantage of the same mechanisms to gain access to the fetus, with pathological consequences. Autoantibodies in women with various disease states may similarly pass into the fetus but these normally produce only mild and transient effects. The development of the fetal immune system begins at an early stage of gestation. It is competent to respond to intrauterine infections from as early as 12 weeks and has full functional potential at birth. Maternally acquired IgG is available for up to 9 months of life until the infant's own immune system has been adequately primed and activated following first exposure to specific antigens. The normal fetomaternal immune relationship represents a remarkable harmonious association between two genetically disparate individuals.(ABSTRACT TRUNCATED AT 400 WORDS)

Dependency on interleukin-1 of primary human in vitro T cell responses to soluble antigens.

The role of interleukin (IL)-1 in antigen-specific activation of naive human T cells has been examined. Primary human T cell proliferative responses to the soluble antigen keyhole limpet hemocyanin (KLH) were decreased by neutralizing antisera to IL-1 alpha (25 +/- 7% standard error) and IL-1 beta (56 +/- 6% standard error). Inhibition by both antisera in a primary culture was usually additive. Recombinant IL-1 alpha and recombinant IL-1 beta could both re-establish responses in cultures blocked by neutralizing anti-IL-1 beta. Interestingly, the susceptibility of KLH-stimulated T cell responses to inhibition by neutralizing anti-IL-1 sera decreased with time in culture. This observation suggested that T cell responses may become less IL-1 dependent as T cells become activated or primed. In support of this notion, secondary T cell responses to purified protein derivative from Mycobacterium tuberculosis (PPD) were markedly less affected by the addition of comparable amounts of the neutralizing anti-IL-1 sera. These results demonstrate that IL-1 is one of the main co-stimulators for primary T cell activation and suggest a different requirement for IL-1 in the activation of naive compared to memory human T cells.

Maternal immune response to pregnancy.

The pregnant female is exposed to a variety of potentially immunogenic foreign antigens on her allogenic intra-uterine conceptus. The extent to which maternal antibodies and cell-mediated immune responses to these antigens are relevant to the paradoxical survival of the fetal allograft is not clearly established. The key to the maintenance of pregnancy lies in the trophoblast. This tissue prevents significant entry of maternal lymphocytes to the fetus and is most likely to protected from maternal immune rejection by features of its cell surface molecular structure and/or its synthesis of factors that render it insusceptible to antibody- or cell-mediated immune lysis in vivo. An alternative, or complementary, protective system, involving maternal recognition and immunoregulatory processes that deviate responses away from the expected rejection reactions, may also operate but has not yet been convincingly demonstrated. Such a mechanism is unlikely to involve the classically defined histocompatibility classically defined histocompatibility antigen system, at least in human pregnancy, where there is an absence of Class I antigens from the trophoblast but increasing evidence of trophoblast-associated antigens and related immune responses. It remains to be established whether unexplained recurrent miscarriage in women and spontaneous abortion in animal models is caused by failure of maternal immunoregulatory control or by non-immunological factors. This is relevant to the validity of immunotherapeutic approaches to the prevention of early fetal loss.

Antibodies to trophoblast in normal pregnant and secondary aborting women.

Enzyme-linked immunosorbent assays (ELISA) with the use of chorionic villous plasma membranes prepared from first trimester and term placentae were employed to detect antibodies to trophoblast in normal primigravid women. Normal pregnant women were found to produce IgG antibodies to trophoblast. These antibodies could be eluted from first trimester placentae. This antibody response was observed in the first trimester and gradually decreased as pregnancy progressed. IgM antibody responses were observed only in the third trimester. Antibodies in some primigravid women and secondary recurrent aborters showed allotypic reactivity with individual trophoblast membranes. This finding was confirmed by immunoblotting experiments in which antibodies from some normal pregnant women were shown to recognize the same trophoblast antigens as those recognized by antibodies from secondary recurrent spontaneous aborters.

Murine pregnancy-associated modulations in lymphocyte reactivity to mitogens: identification of the cell populations affected.

Lymphocytes from the thymus, spleen and inguinal lymph nodes of syngeneically pregnant and non-pregnant mice were compared in their responsiveness to polyclonal stimulation by mitogen. Pregnancy-associated changes in mitogen reactivity were detected, on a cell-per-cell basis, in thymocytes (increased) and spleen cells (decreased) but not in lymph node cells. The hyperreactivity of thymocytes during pregnancy correlated with physiological involution of the thymus occurring through the selective loss of relatively immature, non-mitogen-reactive, Lyt 1+2+ cells. The remaining cells were found largely to be mature Lyt 1+2- T cells with the capacity to respond to mitogenic stimulation. It is most likely the relative increase in the proportion of these Lyt 1+2- cells that causes the hyper-responsiveness of thymocytes to mitogens observed during pregnancy. On the other hand, while spleen cells from pregnant animals gave lower responses to mitogens than those from control virgin females, isolated splenic T cells from the two groups proved equally reactive to T cell mitogens. This supports the contention that at least some aspects of immunity during pregnancy are down-regulated by inhibitory cells within the non-T cell compartment. The results demonstrate the importance of identifying the reactive cell population in studies on changes in lymphocyte responsiveness in pregnancy.

Humoral immune responses in murine pregnancy. V. Relationship to the differential immunogenicity of placental and fetal tissues.

The nature of the humoral immune response induced in virgin female mice by injections of F1 placental and fetal tissues has been examined and compared to that induced by immunization with F1 adult spleen cells and by multiple allogeneic pregnancy. In a 'responder' strain mouse, as defined by the ability of multiple allogeneic pregnancy to elicit an anti-paternal humoral immune response, both F1 placental and fetal tissues induced the formation of alloantibodies primarily of the IgG1 sub-class, similar to those induced by allogeneic pregnancy, but different from those elicited by adult spleen cells. However, only the placental tissues induced alloantibodies possessing all the characteristics of those appearing in multiparous allogeneic pregnancy. In contrast, the alloantibodies induced by the injected fetal tissue possessed complement-dependent cytotoxic activity, indicating that the inability of pregnancy-induced alloantibodies to mediate cytotoxicity may not be related to their restriction to the IgG1 sub-class. In a 'non-responder' mouse strain, where multiple allogeneic pregnancy does not lead to a maternal alloantibody response, F1 placental tissues, in contrast to fetal and adult tissues, failed to induce a humoral immune response. Injection of F1 placental tissue therefore elicits responses that mimic both the properties and the strain-dependent distribution of the alloantibodies identified in normal murine pregnancy. This implies that the immunogenic stimulus in pregnancy emanates from the placental rather than the fetal compartment of the allogeneic conceptus.

Localization of immunoglobulin-containing cells in human endometrium in the first trimester of pregnancy and throughout the menstrual cycle.

The distribution of immunoglobulins in normal human endometrium throughout the menstrual cycle and in early pregnancy has been studied with an immunoperoxidase technique. In first-trimester decidua, IgG was detected within many cells of differing morphology and size. Large IgG-containing cells were often binucleate and were believed to be decidual cells. Examination of serial sections showed no kappa or lambda light-chain restriction, suggesting absorption of the immunoglobulin content. Medium-sized, irregular, IgG-containing cells were macrophages. An additional substantial population of small hyperchromatic IgG-containing cells were prominent around arterioles and adjacent to endometrial glands. From examination of adjacent sections stained with phloxine tartrazine, it was concluded that these represented endometrial granulocytes. Labelling for light chains again suggested absorption of the immunoglobulin content. In contrast, in non-pregnant endometrium immunoglobulin-containing stromal cells were uncommon, although IgG and IgA were detected in gland epithelium and secretions and in the stromal interstitium particularly in the secretory phase. These results support the notion that human endometrium lacks a classical secretory immune system and highlight the requirement for correlation between studies of cell surface markers, morphology and cell surface receptors.

The rat placenta expresses paternal class I major histocompatibility antigens.

The expression of paternally inherited class I MHC antigens on the placental trophoblast of the rat has been investigated using a mouse anti-rat monoclonal antibody (MN4-91-6) in an indirect immunoperoxidase labelling assay on cryostat sections. Strong specific staining was obtained on the spongy zone trophoblast of the mature placenta from DA male (RT1a) X PVG female (RT1c) matings. In marked contrast, no staining was observed on the labyrinthine trophoblast nor on the trophoblastic giant cells at any stage of gestation from 8 to 19 days post-coitum. None of the trophoblastic cell populations at any stage of gestation were reactive with an anti-class II monoclonal antibody. Class I positive endovascular cytotrophoblast cells were present in the maternal arterial sinusoids of the decidua. These findings imply that maternal immunoregulatory mechanisms must be essential for the survival of the placenta and fetus.

Immunohistologic identification of trophoblast populations in early human pregnancy with the use of monoclonal antibodies.

Three murine monoclonal antibodies (H315, H316, and NDOG1) have been used in a peroxidase-antiperoxidase technique on formalin-fixed paraffin-embedded tissues to identify populations of fetal trophoblast cells by their expression of membrane antigens in chorionic and decidual tissue from the first trimester of normal human pregnancy. H315 and H316 showed comparable staining of placental villous syncytiotrophoblast and cytotrophoblast and were also able to distinguish subpopulations of nonvillous trophoblast in the placental bed, including perivascular and endovascular trophoblastic cells as well as cytotrophoblastic elements within the decidua and myometrium. H315 and H316 also showed cytoplasmic staining of columnar epithelium of endometrial glands throughout the first trimester. In contrast, NDOG1 stained chorionic syncytiotrophoblast but not villous cytotrophoblast and also did not react with any cytotrophoblastic elements in the placental bed. NDOG1 distinguished these different subpopulations of trophoblast as early as 13 to 15 days after ovulation.

Humoral immune responses in murine pregnancy. III. Relationship between anti-paternal alloantibody levels in maternal serum, placenta and fetus.

Studies have been carried out on allogeneically mated pregnant female mice to examine the relationship between the levels of anti-paternal alloantibody occurring free or as soluble immune complexes in the maternal peripheral circulation and that existing in a bound form in the placenta and free in the fetal circulation. In a 'non-responder' strain of mouse, as defined by the inability to detect anti-paternal alloantibody in the peripheral serum of multiparous allogeneically mated females, no alloantibody could be detected as soluble immune complexes, bound to the placenta or in the fetal circulation. A similar situation existed in females of a 'responder' strain during their first pregnancy and in some individuals during their second. However, in 'responder' strain females possessing alloantibody in their peripheral serum, alloantibody could be eluted from the semi-allogeneic placenta when peripheral titres exceeded 1:16, and was found in the fetal serum when they exceeded 1:128. When such females were subsequently mated syngeneically, alloantibody was detected in placental eluates only when peripheral titres exceeded 1:128-256 and in the fetal serum when they exceeded 1:16. The alloantibody eluted from placentae and that found in neonatal serum exhibited similar isotype distribution to that in the peripheral serum. These results are discussed with reference to the relative importance of the yolk sac and the immunoabsorbent semi-allogeneic placenta in the restricted transfer of pregnancy-induced anti-paternal alloantibody to the fetus.

Immunoregulatory activity of supernatants from short-term cultures of mouse decidual tissue.

Supernatants from short-term in-vitro cultures of decidual tissue, obtained from the uteri of pregnant mice from Days 4 to 13 post coitum (Day 1 = day of mating), were assessed for immunoregulatory activity by their addition to a mixed lymphocyte reaction (MLR), an in-vitro analogue of the afferent arm of the immune response. All culture supernatants tested possessed inhibitory activity in the MLR, although the extent of inhibition was affected by seeding density, length of culture, and the day of pregnancy from which decidual tissue was obtained. Inhibitory activity produced by decidual cultures increased from Day 4 to reach a maximum on Day 8, and then declined to Day 11. Two morphologically distinct cell types were present in all decidual cultures; flat dendritic cells, considered to represent decidual cells, and small round cells, but whether immunoregulatory factors are associated with both is uncertain. The results suggest that decidual tissue could fulfil a role in the local partial blockade of the afferent arm of the maternal immune response during pregnancy.

Hormonal regulation of immunoglobulins and plasma cells in the mouse uterus.

Immunoperoxidase staining techniques were used to identify secreted immunoglobulin G (IgG) and IgA and the distribution of IgG- and IgA-producing cells in the uteri of mice during the oestrous cycle or after ovariectomy followed by the administration of ovarian hormones. The most extensive staining for IgA and IgG in the myometrium and endometrium occurred during pro-oestrus and plasma cells of both isotype were concentrated around the uterine glands and lumen in high numbers at this time. Staining for IgA and IgG was least prominent during dioestrus, the stage of the oestrous cycle when plasma cells were present in the lowest numbers. After ovariectomy and hormone treatment, the highest numbers of uterine plasma cells were seen in mice injected with oestradiol alone and the lowest in those injected with progesterone alone. In mice treated with both hormones the number of plasma cells was comparable to that seen in mice treated with progesterone alone. These results indicate that oestradiol plays a significant role in regulating uterine IgA and IgG and suggest that the increase in the numbers of plasma cells during the oestrous cycle is due to the action of oestradiol at the uterine level.

Partial characterization of immunosuppressive factors from short-term cultures of murine decidual tissue.

Previous in vitro experiments provided evidence for the immunosuppressive properties of supernatants from short-term cultures of decidual tissue obtained from the uterus of the pregnant mouse. The inhibitory activity of supernatants, previously detected in a mixed lymphocyte reaction (MLR), has now been demonstrated in an in vitro thymocyte proliferation assay. Fractionation of these supernatants on Sephadex G-15 and Sephacryl S-300 revealed that inhibition of the thymocyte proliferation assay was associated only with a low molecular weight fraction (less than 1,500), whilst MLR inhibitory activity was associated with three different molecular weight fractions (less than 1,500, 60,000 and 1,000,000). These results are discussed in relation to the different biological factors previously isolated in decidua or deciduoma of rats and humans.

Ia antigen-bearing decidual cells and macrophages in cultures of mouse decidual tissue.

The ontogeny of Ia antigen expression on cells within the decidua basalis of the placenta and on decidual cells differentiating in vitro was investigated in the mouse. The results indicate that Ia antigen expression is temporarily restricted and can be detected in short-term cultures of decidual tissue only during the final third of gestation. The Ia antigen positive cells, which are non-phagocytic, non-specific esterase negative and lack Fc receptors, appear to be true decidual cells on the basis of their fine structural characteristics. In contrast, morphologically identical dendritic decidual cells arising from differentiation in vitro of endometrial cells do not express Ia antigens. A population of small rounded cells was also present in cultures of both decidual tissue and in vitro decidualized endometrial cells. These rounded cells possess the macrophage-like properties of Fc receptors, non-specific esterase and phagocytic activity. The possible function of these cell populations within the decidual tissue is discussed.

Fetal histocompatibility antigens and maternal immune responses.

Information on the antigenic status of the fetus and placenta and the nature of the maternal alloimmune responses in pregnancy is essential for an understanding of the factors responsible for the survival of the fetus as an intrauterine allograft. Although minor histocompatibility antigens are expressed from the earliest stages of preimplantation embryonic development in the mouse the paternally inherited Class I antigens of the major histocompatibility complex are absent from both fetal and surrounding trophoblastic tissues until the mid-gestation period. In the definitive placenta, the spongiotrophoblast, a major subpopulation of trophoblast in direct contact with maternal blood and uterine tissue, expresses these MHC antigens and is potentially susceptible to immune attack. The maternal alloimmune response is extremely restricted. Antibody is detectable only in females of an H-2b haplotype, only after a second allogeneic mating, and exhibits little or no complement-fixing activity, There is no detectable generation of cytotoxic T cells. Placental and fetal tissue inoculation experiments indicate that the deviated response in pregnancy is mediated at the placental level. The failure of an experimentally induced state of hyperimmunity to prejudice the course of pregnancy is considered to be due to the immunoregulatory effect of non-antigen-specific maternal serum factors.