Immunology of Pregnancy

Defining the Immunology of Pregnancy

In 1991, Colbern and Main redefined the conceptual framework of reproductive immunology as maternal-placental tolerance rather than maternal-fetal tolerance, focusing on the interaction of the maternal immune system with the placenta rather than the fetus. The early embryo is composed of two groups of cells: the inner cell mass, which gives rise to the embryo, and the external embryonic trophectoderm, which becomes trophoblast cells and later, the placenta. Cells from the placenta directly interact with the mother’s circulating and local uterine immune cells and are able to evade immune rejection. Though once thought to have no direct contact with the fetus itself, maternal cells cross the placental barrier to enter the fetal circulation. However, this original placental barrier concept, coupled with the observation that the fetus expresses paternal major histocompatibility complex (MHC) ^a

a 14;See Table 8.1 for definitions for abbreviations used throughout this chapter.

antigens, led Medawar to postulate that the fetus would be rejected as a true allograft if removed from its trophoblastic cocoon and transplanted into the thigh muscle or kidney capsule of the mother.

Types of Immune Response

The immune system eliminates foreign material in two ways: through innate or adaptive immunity. Innate immunity produces a relatively unsophisticated response that prevents the access of pathogens to the body. This is a primitive evolutionary system that does not require prior exposure to similar pathogens. The primary cell types involved in these responses are phagocytic cells, such as macrophages and neutrophils. These cells express pattern recognition receptors (PRRs) that sense conserved sequences on the surface of microbes to trigger an immune response. As a result, phagocytic cells produce proinflammatory cytokines, release degradative enzymes, generate intense respiratory bursts of free radicals, and ultimately engulf and destroy the invading microorganism. Thus the innate immune system provides the first line of defense against invading microbes. Furthermore, the innate immune system is critical for priming the adaptive immune response.

Adaptive immunity is an additional, more sophisticated response found in higher animal species, including humans. Cells of the innate immune system process phagocytosed foreign material and present these antigens to cells of the adaptive immune system. This immune response is highly specific and potentiated by repeated antigenic encounters. Adaptive immunity consists of two types of responses: humoral immunity, in which antibodies are produced, and cellular immunity, which involves antigen-dependent cell lysis by specialized T lymphocytes (cytolytic T cells). Adaptive immunity is characterized by an anamnestic response that enables immune cells to “remember” the foreign antigenic encounter and react to further exposures to the same antigen faster and more vigorously.

Cytokines and the Immune Response

Immune cells mediate their effects by releasing cytokines, and through these secreted factors, they can establish either a proinflammatory or antiinflammatory microenvironment. Moreover, the cytokine profile created by immune cells can shape the characteristics of subsequent immune responses. For example, naive T helper lymphocytes (Th0) originate in the thymus and play a major role in creating specific microenvironments within the periphery, depending upon their differentiation status. If a Th0 cell then differentiates into a T _h 1 cell, it secretes interleukin-2 (IL-2) and interferon-γ (IFN-γ), setting the basis for a cellular, cytotoxic immune response. Conversely, T _h 2 lymphocytes secrete cytokines, such as IL-4, IL-6, and IL-10, which are predominantly involved in antibody production. Furthermore, the actions of T _h 1 and T _h 2 cells are closely intertwined, both acting in concert and exerting counter-regulatory effects through their cytokines. Thus T _h 1 cytokines produce proinflammatory cytokines that, while acting to reinforce the cytolytic immune response, can also downregulate the production of T _h 2-type cytokines.

As discussed later, the pregnant endometrium (decidua) is populated by maternal immune cells, both during implantation and throughout gestation. Therefore the maternal immune system interacts, at different stages and under various circumstances, with the invading trophoblast. Our objective is to understand these interactions and their role in support of a normal pregnancy. The following sections summarize some of the main hypotheses proposed to explain the trophoblast-maternal immune interaction.

The Placenta as a Mechanical Barrier

Up until the late 1980s, the most popular of the five theories was the belief that a mechanical barrier formed by the placenta prevented bidirectional movement of immune cells across the maternal-fetal interface. The barrier thus created a state of “immunological ignorance,” in which fetal antigens were never presented to and thus never detected by the maternal immune system. Scientists believed that the barrier formed in the pregnant uterus between trophoblast and decidua prevented movement of activated alloreactive immune cells from the maternal circulation to the fetal side. Conversely, this barrier was thought to isolate the fetus and prevent the escape of fetal cells into the maternal circulation.

Challenging the mechanical barrier theory are studies showing that the trophoblast-decidual interface is not impermeable. Evidence for bidirectional trafficking across the maternal-fetal interface includes migration of maternal cells into the fetus and the presence of fetal cells in the maternal circulation. ^, This trafficking in immune cells has now be shown in most of the human body’s immune privilege tissues, including the brain’s blood-brain barrier (BBB). Indeed, fetal cells can be observed in the mother decades after pregnancy. ^, These cells, like other stem cells, have the potential to infiltrate maternal tissues and differentiate into liver, muscle, and skin cells, transforming the mother into a chimera. Originally it was thought that these fetal cells were responsible for triggering autoimmune diseases that more often afflict women. However, more recent studies have demonstrated that fetal cells may play a critical role in repairing maternal tissues damaged by such pathologic processes. In one case study, a woman suffering from hepatitis stopped treatment against medical advice, yet despite this, she did well clinically and her disease abated. Her liver specimen was found to contain male cells that originated from her previous pregnancies, suggesting that these residual fetal cells in the mother’s circulation produced new liver cells that were partly responsible for her recovery. ^{, ,} The converse also appears to be the case—that maternal cells can traffic to the fetus and prime the fetal immune system to suppress antimaternal immunity.

Systemic Immune Suppression

A second theory postulates that pregnancy is a state of systemic immune suppression preventing maternal immune cells from rejecting the fetus. This concept has been studied by numerous investigators and over many years became conventional wisdom. Indeed, a wide array of factors in human serum have been found to have profound in vitro immunosuppressive activities. However, if we carefully analyze this hypothesis, it is difficult to imagine how, from an evolutionary point of view, pregnancy involves a stage of profound immune suppression. Early humans lived in a nonhygienic milieu and were continually exposed to bacteria, parasites, and other microorganisms. If pregnant women were systemically immunologically suppressed, they would not readily survive and Homo sapiens would not have flourished. Even today, in many parts of the world, pregnant women are continually exposed to harsh, unsanitary conditions in which a suppressed immune system would make it impossible for the mother and fetus to survive. Furthermore, in countries where HIV is endemic, such as Africa, HIV-positive women do not develop AIDS more readily during pregnancy. In fact, recent studies clearly demonstrate that maternal antiviral immunity is not affected by pregnancy. Together, these observations argue against the existence of such nonspecific immune suppression.

Cytokine Shift

The definition of pregnancy as a Th2 or antiinflammatory state was enthusiastically embraced, and numerous studies attempted to support this hypothesis. This theory postulates that pregnancy is an antiinflammatory condition and a pathological shift in the type of cytokines produced would lead to abortion or other pregnancy complications. Though appealing on the surface, studies have been contradictory. ^, One reason for these contradictory results may be simplistic interpretations of disparate observations made during pregnancy. In the aforementioned studies, pregnancy was evaluated as a single continuous state, when in reality it has three distinct immunological phases characterized by distinct biological processes that mirror how the pregnant woman feels. Therefore it is important to appreciate that human pregnancy is both proinflammatory and antiinflammatory, depending upon the stage of gestation, rather than focusing on the murine T _h 1/T _h 2 terminology. Implantation and placentation during the first trimester of pregnancy resemble an open wound that triggers a strong inflammatory response. During this first phase, the embryo attaches to and penetrates the uterine surface and then the epithelial lining of the uterus is restored. As placentation proceeds, more damage occurs to decidualized endometrial tissue, and invasive trophoblast invades the uterine spiral arterial vasculature to remodel vascular smooth muscle and secure an adequate blood supply. All of these activities create a veritable “battleground” of invading cells, dying cells, and repairing cells. An inflammatory environment is required to repair uterine epithelium and remove cellular debris. During this period, the mother’s well-being is affected: she feels ill because her whole body is struggling to adapt to the presence of the fetus (in addition to hormone changes and other factors, this immune response is responsible for “morning sickness”). Thus the first trimester of pregnancy is a proinflammatory phase ( Fig. 8.1 ).

Given that inflammatory pathways are activated at the time of implantation and inflammation is thought to be associated with a rejection process, three questions arise: (1) How does immune tolerance of an invasive embryo evolve within an inflammatory environment without immediate rejection? (2) What are the target cells of the inflammatory process? and (3) What are sources of the inflammatory process? Understanding the mechanisms of how inflammation is temporally and spatially triggered and controlled within the uterus is fundamental for developing effective therapeutics to improve fertility and decrease poor obstetrical outcomes.

Distinct immunological and molecular changes are observed in the receptive endometrium before implantation occurs. During apposition, chemokines and cytokines produced by the endometrial cells guide the blastocyst to the site of implantation, and embryonic cell surface L-selectin binds to its ligands on the uterine luminal epithelium. This enables the initial contact of the blastocyst with the uterus. In mice, leukemia inhibitory factor (LIF) is transiently increased in mouse uterus before implantation. LIF is an IL-6 class cytokine with proinflammatory potential. These data indicate that even before implantation the endometrium is in an inflammatory state, potentially under the control of ovarian steroids. In addition to LIF, the immune status at the implantation site is characterized by expression of proinflammatory cytokines, such as IL-6, IL-1β, IL-8, granulocyte-macrophage colony-stimulating factor (GM-CSF), and tumor necrosis factor-α (TNF-α). These cytokines could be produced by endometrial epithelial cells, endometrial cells, and the immune cells that are recruited to the site of implantation. If this inflammatory process is not present, the endometrium is not receptive and is not prepared for implantation.

Although inflammation is critical for implantation, it must be properly contained and controlled for implantation to properly progress. The capacity to resolve decidual inflammation has evolved as a key feature underpinning placentation in viviparous mammals. Interestingly, the inflammatory process necessary for embryo implantation is a well-conserved evolutionary process. Using the opossum model, Wagner’s lab has demonstrated that inflammation is observed during the early attachment of the embryo and is characterized by the expression of immune-related genes, including IL-1A, IL-6, TNF, PTGS2 (COX2), PTGES, IL-17A, and neutrophil elastase. However, implantation in the opossum is short-lived. After the conceptus attaches to the endometrium, the animal is born 2 to 3 days later. Detachment from the uterine lining is associated with an increased inflammatory environment, similar to the parturition process observed in mammals. ^, These observations suggest that inflammation is necessary for implantation throughout evolution. However, a shift to an antiinflammatory phase is required for the maintenance of the pregnancy, because a continuation of the inflammatory environment will mimic parturition. Therefore the switch on and off for inflammation during implantation is a necessary step to maintain the pregnancy. It is also a critical step for the process of placentation ( Fig. 8.2 ).

The second immunological phase of pregnancy is, in many ways, the optimal time for the mother. This is a period of rapid fetal growth and development. The mother, placenta, and fetus are symbiotic, and the predominant immunological feature is induction of an antiinflammatory state. The woman no longer suffers from nausea and occasional low-grade fever as she did in the first stage, in part because the immune response is no longer the predominant systemic feature. An unresolved question is the identity of the cellular components responsible for the inflammatory switch. Evidence obtained from animal and human in vitro studies indicates that uterine/decidual macrophages, through their potent immune capacity, are critical for the inflammatory switch on and off. ^, As discussed below, the macrophage’s unique phenotype and plasticity make it a major player for the establishment and maintenance of a successful pregnancy.

During the last immunological phase of pregnancy, the fetus has completed its development; all of the organs are functional and ultimately adapted to postnatal life. Now the mother needs to deliver the fetus, and this can only be achieved through renewed inflammation. Parturition is characterized by an influx of immune cells into the myometrium to promote recrudescence of an inflammatory process. This proinflammatory environment promotes decidual activation, cervical change, and uterine contractions, facilitating expulsion of the fetus and subsequent rejection of the placenta. In conclusion, pregnancy is both a proinflammatory and antiinflammatory condition, depending upon the stage of gestation ( Fig. 8.1 ).

Lack of Expression of HLA Antigens

A theory is based on the absence of polymorphic class I and II molecules on trophoblast. MHC class I antigens are expressed on the surface of most nucleated cells and serve as important immune recognition molecules. In humans these antigens are also known as human leukocyte antigens (HLAs). HLA class I genes are located on the same chromosomal region (6p21.3) and have been subdivided into two groups, namely, the HLA class Ia and HLA class Ib genes, according to their polymorphism, tissue distribution, and functions. HLA-A, HLA-B, and HLA-C class Ia genes exhibit a very high level of polymorphism, are almost ubiquitously expressed among somatic tissue, and have well-established immunological functions. They modulate antiviral and antitumoral immune responses through their interaction with T and natural killer (NK) cell receptors. In contrast, HLA-E, HLA-F, and HLA-G class Ib genes are characterized by their limited polymorphism, restricted tissue distribution, and poorly understood roles. The statement that the human placenta does not express polymorphic MHC class I molecules is not entirely accurate. The trophoblast, in human placenta, does not express the polymorphic HLA-A and HLA-B class I antigens but does express HLA-C molecules. In addition, the nonclassical HLA-E, HLA-F, and HLA-G molecules are expressed by human trophoblast cells. Using immunostaining against class I molecules, Loke and King divided human trophoblast into two distinct populations: the villous trophoblast in contact with maternal blood at the intervillous interface, which is class I negative, and the extravillous cytotrophoblast invading the uterine decidua, which is class I positive (HLA-C). ^, Based on these findings, it was suggested that there are two fetal-maternal interfaces in human reproduction that differ immunologically: a trophoblast population that is immunologically neutral in contact with the circulating maternal immune system, and a local immunologically active population of trophoblast migrating into the decidua that can be stimulated by HLA class I molecules.

The HLA-G molecule was originally cloned in 1987 and is abundant at the maternal-fetal interface. ^, Initially it was thought to be almost exclusively expressed at the maternal-fetal interface, suggesting a very specialized role in this environment. Another unique feature of the HLA-G genes, which was postulated to be a prerequisite for maintenance of maternal immune tolerance, was its apparent lack of polymorphisms. However, new data have shown that alternative splicing of the HLA-G mRNA yields different membrane-bound and soluble variants of the HLA-G protein, and a limited number of variable sites have been discovered in the DNA sequence of the HLA-G gene. Therefore the hypothesis that HLA-G is the mediator of fetal-maternal tolerance due to its monomorphism and immunological neutrality needs to be revised. ^{, ,}

In animal studies, it was shown that murine trophoblast expresses MHC class I genes and alloantigens at high levels early in gestation age. During those times, MHC class I is barely detected on fetal tissues. ^, Because it is highly likely that maternal T cells circulating through the murine maternal-fetal interface encounter cells of fetal origin, the hypothesis that fetal tissue at the interface is antigenically immature and therefore does not provoke maternal T-cell responses is untenable. Instead, HLA-G, HLA-F, and HLA-E may dampen maternal immune responses by interacting with killer immunoglobulin receptors (KIRs) on uterine NK cells and macrophages and with the T-cell receptor on CD8+ cells. ^, The consequences of these interactions include activation of pathways in NK cells and macrophages that interfere with the killer functions of these cells. ^, At the same time, HLA-G may also activate immunomodulatory pathways in decidual NK cells, macrophages, a subset of T cells, and the T-cell receptor on CD8+ T cells. ^, Specifically, HLA-E and HLA-G expressed by extravillous trophoblast inhibit the cytotoxic activity of NK and T cells and instead modulate immune cell function to promote trophoblast migration and placentation.

A novel mechanism by which HLA-G influences immune function is through a process of trogocytosis where decidual NK and T cells that have acquired HLA-G are immunosuppressive. Soluble HLA-G, now termed HLA-G5, has been identified in maternal serum, is biochemically unique among the HLA-G isoforms, and is also associated with specific immune modulation and the promotion of a pro-placentation environment. In addition to being released in a soluble form, HLA-G can be released from the placenta via exosomes. ^, These soluble forms might provide a systemic immune regulation, although the molecular mechanism has not been elucidated.

In conclusion, although the apparent lack of classic MHC gene expression suggests that the preimplantation embryo is protected from direct immunological attack by MHC-restricted maternal T cells, the preimplantation embryo could still be vulnerable to a delayed-type hypersensitivity (DTH) reaction, as well as to adverse effects of antibodies and cytokines by non-MHC-restricted effector cells.

Local Immune Suppression

The most recent hypothesis is that specific local immune suppressor/regulatory mechanisms are present during pregnancy. According to this hypothesis, immune cells that specifically recognize paternal alloantigens are deleted from the maternal immune system. This elimination process is thought to be achieved through either deletion of these alloreactive cells or the suppression of their activity. One mechanism by which paternal antigen-recognizing T cells may be deleted is through induction of cell death (apoptosis) by members of the TNF-related apoptosis-inducing superfamily such as Fas ligand (FasL) and TNF-related apoptosis-inducing ligand (TRAIL) ( Fig. 8.3 ). Another mechanism is modulation of T-cell function by B7 family members such as programmed cell death ligand-1 (PD-L1). ^, The role of immune checkpoints such as the Fas/FasL and PD-1/PD-L1 systems in the regulation of maternal immune responses might be relevant as potential targets for immune modulation in patients with pregnancy complications such as preterm birth and preeclampsia. Early studies showed that FasL expressed in the cell surface of trophoblast ^, or in secreted microvesicles can then act on activated Fas receptor–expressing maternal immune cells, even at locations distant from the implantation site, to promote cell death.

More recently, the interaction between PD-1 (CD279) and PD-L1 (CD274) has emerged as a key player in regulating immune responses and peripheral tolerance by either promoting T-regulatory cell (Treg) development and function and/or directly inhibiting activated T cells. Tregs are a subset of lymphocytes that are able to suppress the actions of alloreactive T cells to promote fetal/paternal immunotolerance. ^, During pregnancy, PD-1/PD-L1 has been shown to be involved in the promotion of maternal tolerance to paternal antigens by enhancing Tregs and inhibiting T _h 17 cells, which play a pathogenic role in autoimmune diseases and pregnancy complications. ^, Interestingly, administration of a soluble PD-L1-Fc protein provides a protective effect on a rat model of preeclampsia by reversing Treg/T _h 17 imbalance.

Another way in which T cells that recognize paternal antigens may be deleted is through the production of indoleamine 2,3-dioxygenase (IDO) at the maternal-fetal interface. IDO is an enzyme that degrades tryptophan, an amino acid that is essential for T-cell proliferation and survival. Although IDO was proposed to be critical for immune protection and survival of the mouse embryo, later studies using IDO ^–/– knockout mice failed to reveal any impact on pregnancy. That said, evidence is beginning to emerge in support of this potential mechanism in humans.