Immune Cells at the Fetomaternal Interface: How the Microenvironment Modulates Immune Cells To Foster Fetal Development

Immune cells arising in the bone marrow or lymph nodes travel via the circulation before taking up residence in peripheral tissues. Once recruited, the immune cells must adapt to the local microenvironment and may be exposed to changing conditions within it, influenced, for example, by environmental insults or intrinsic perturbations.

One of the most remarkable tissues in this regard is the uterus, which undergoes cyclic regeneration, remodeling, and demise in response to fluctuating sex steroid hormones over the course of the menstrual cycle. Additionally, the uterus must respond appropriately to foreign stimuli, including male seminal fluid, the products of conception, the local microbiome, and sexually transmitted pathogens.

In pregnancy, the uterus undergoes enormous transformations, starting at conception and completing with labor and birth. In order for a newly fertilized egg to achieve blastocyst stage embryo development and successfully invade into the uterine tissue, a finely orchestrated balance of immune cell subsets and their soluble mediators is required. Although the developmental potential of the blastocyst is mainly determined by genetics, its viability and competence to become a fully formed fetus and achieve on-time delivery depends on an optimal uterine environment, which in turn reflects the quality of the maternal immune response.

This review will discuss the key elements of the innate and adaptive immune cell repertoire in the uterus and the major regulatory forces controlling their contribution to reproductive function. Although the vagina and uterine cervix also constitute important niches with unique dynamics, the major focus of this review is the immune cells present within the uterus before and during pregnancy.

Embryo implantation, when the embryo invades the endometrium and commences placental morphogenesis, is the most critical event for setting the course of pregnancy. In many aspects, the implantation site resembles an inflammatory response, with a profusion of recruited immune cells and induction of inflammatory genes (1). The functional phenotypes of the immune cells engaged in this response are pivotal and determine whether a viable pregnancy will occur.

In turn, the phenotypes and numbers of immune cells at implantation are determined earlier in the menstrual cycle as the uterus prepares for potential pregnancy. The endometrial surface layer of the uterus contains the greatest abundance of immune cells. These progressively increase through the proliferative and periovulatory phases of the menstrual cycle because of influx of leukocytes from the periphery in response to ovarian hormone–regulated chemokine and cytokine expression and proliferation of existing uterine populations (2–4). The ensuing immune cell repertoire plays a major deciding role in whether the uterine endometrium will be receptive or refractive to embryo implantation in the secretory phase.

A receptive endometrium will tolerate and support invasion of trophoblast cells emanating from the blastocyst and facilitate the rapid growth of the placenta while also mediating the transformation of surrounding uterine fibroblasts into decidual cells. Through affecting the extent of trophoblast invasion and the quality of placental development, the status of the immune response at conception can impact events much later in pregnancy, influencing miscarriage and compromising fetal growth if immune tolerance is not sufficient from this early time (5, 6).

To a degree unparalleled in most adult tissues, the immune cells are major regulators of uterine development and remodeling. Reciprocally, immune cells necessary for implantation must rapidly adapt to hormone, chemokine, and cytokine signals released from uterine epithelial and stromal cells, to the extent that they acquire unique phenotypes and functions different from their counterparts in the peripheral blood or in other tissues. Uterine immune cells arise both by proliferation of pre-existing resident cells and migration from the periphery. In a healthy uterus, they recognize and respond to local microenvironmental signals to control the fine balance between tolerance and rejection and between tissue remodeling and clearance of aged or altered structures and cells.

A critical event that primes the innate and adaptive immune cells in the uterine endometrium for implantation is contact with male partner seminal fluid, especially at the time of conception. Seminal fluid is not just a transport and survival medium for spermatozoa; it is rich in soluble signaling agents, including cytokines, PGs, hormones, and other proteins capable of exerting profound biological effects on the female immune response as shown for various species (7). Intromission of seminal fluid recruits innate immune cells and initiates the inflammation-like response of implantation (Fig. 1). As the first exposure to male MHC Ags, it primes the T cell response to influence the fate of embryo tolerance or rejection and makes an important contribution to setting the trajectory of the maternal immune response for the course of pregnancy (8). The balance of immune regulatory signals in seminal fluid is critical for the quality of female immune tolerance and competence to support implantation and achieve full-term pregnancy (9).

The response to seminal fluid is best defined in mice, in which gene expression studies show that seminal fluid induces an array of cytokines and chemokines within the luminal epithelial cells lining the endometrial surface (10). Their release into the stromal compartment causes a robust recruitment of innate immune cells, including macrophages and dendritic cells (DCs) as well as granulocytes, into the endometrial stroma and lumen from the blood (10–12). Recruited neutrophils and macrophages clear debris, excess sperm, and microbes introduced at mating. DCs take up soluble male MHC Ags in seminal plasma and traffic to the local draining lymph nodes, where they present Ags to activate and expand inducible and thymic regulatory T (Treg) cells that mediate functional tolerance to paternal MHC Ags (13–15).

The subsequent migration of seminal fluid–induced Treg cells to the endometrium from the blood is a key step for promoting endometrial receptivity necessary for successful embryo implantation. Because the implanting embryo expresses the same paternal MHC Ags present in the conceiving partner’s seminal fluid, Treg cells reactive with these Ags prevent generation of type 1 immunity to paternal transplantation Ags and permit invasion of trophoblast cells and the necessary vascular changes required for robust placental development. Treg cells also act to suppress excessive inflammation that would otherwise cause embryo demise (16). In mice, the Treg response to seminal fluid contact is driven by very high concentrations of the key immune-regulatory cytokine TGF-β, secreted into the seminal fluid by the seminal vesicle gland (12).

In women, the uterine ectocervix is the major inductive and effector site for immune responses in the female reproductive tract and the primary tissue responding to seminal fluid (17) (Fig. 1). The cervix also plays a key defense role against sexually acquired infections and is likely involved in maintaining immune tolerance toward sperm and other paternally derived Ags contained in the ejaculate. Seminal fluid contact induces an influx of neutrophils followed by macrophages and DCs, along with abundant memory T cells into the human cervical epithelium and deeper stromal compartments (18, 19). Leukocyte recruitment is accompanied by induction of genes associated with cytokine signaling, inflammation, Ag presentation, and leukocyte recruitment. There are consequences for endometrial immune cells, with increases in endometrial CD16⁻/CD56^bright uterine NK (uNK) cells seen in women who have intercourse before ovulation compared with those who abstain (20). There are signs that, as in mice, seminal fluid priming initiated in the human cervix acts to stimulate expansion of the Treg cell pool available for endometrial recruitment, but this is yet to be confirmed in vivo (9).

There is potential for direct effects of seminal fluid on the endometrial surface in women, after movement of seminal fluid by peristaltic uterine contractions that propel sperm and sperm-bound TGF-β into the upper reproductive tract (21, 22). Access to the human uterus also occurs via local countercurrent transfer in which bioactive substances within the vagina are absorbed through the vaginal wall, enter the vaginal venous blood, and are transported into the uterine arterial blood (23). In vitro experiments demonstrate that human seminal fluid induces genes encoding CSF2, IL-1B, IL-6, CXCL8, LIF, VEGFA, and TGF-β in primary endometrial epithelial cells (24–26), which contribute to immune cell recruitment and phenotype regulation, with consequences for endometrial receptivity and the decidual response (27).

Human seminal fluid contains several potent immune regulatory factors, which act in synergy to skew the adaptive immune response toward tolerance by targeting various steps in the process of Treg cell generation. As in mice, TGF-β is abundant in human seminal fluid with ∼500 ng/ml total TGF-β, including TGF-β1, TGF-β2, and TGF-β3 isoforms (28). TGF-β targets DCs to promote a tolerogenic phenotype and also acts in T cells to skew fate commitment (29). Human seminal fluid also contains high amounts of E-series PGs, predominantly of the hydroxylated form, which at high concentrations are known to induce a regulatory phenotype in naive CD4⁺CD25⁻ T cells (30, 31). Soluble CD38 produced in the seminal vesicles is another immune regulatory agent and potent inducer of tolerogenic DCs and CD4⁺Foxp3⁺ Treg cells (32). Soluble HLA-G in seminal fluid (33) potentially also promotes tolerance by virtue of its interaction with inhibitory receptors Ig-like transcript (ILT)2 and ILT4, the killer Ig-like receptor (KIR)2DL4, and CD8 to inhibit the functions of DCs, NK cells, cytotoxic T cells, and B cells (34).

Factors that impair Treg cell generation can also be present in seminal fluid. IFN-γ, which suppresses the female response to TGF-β (35), is upregulated in the event of various inflammatory conditions in men (36), and through skewing, the balance of cytokines induced in female reproductive tract tissues may impair the generation of tolerance after coitus (Fig. 1). Given that the specific effects of these factors are concentration dependent (31), it seems likely that the absolute amounts and relative balance of immune regulatory factors in seminal fluid could substantially impact the quality and strength of the female tract immune response.

Innate immune cells constitute the major population of leukocytes in the uterus at the time of embryo implantation, with uNK cells being the most abundant and DCs, macrophages, neutrophils, and mast cells also present. There is bidirectional communication between innate immune cells and other cell lineages in the uterus, such that the immune cells are modulated by the environment and acquire a profile unique to the uterus. Importantly, depletion of any one of these cell types changes the environment to such an extent that implantation is hindered.

DCs emerge as highly relevant because of their master role in presenting Ags to cells of the adaptive immune system. This step is crucial in pregnancy as the nature of Ag presentation defines the fate and the quality of the immune response defining tolerance or rejection of the blastocyst.

Uterine DCs (uDCs) with a distinctive phenotype are very abundant in the endometrial tissue during the receptive phase in mice (37) and are grouped in cluster-like structures along the uterus, which may indicate the future sites of implantation as observed by in vivo two-photon microscopy (4). Implantation failure is observed after DC depletion in both allogeneic and syngeneic mating combinations, in which depletion of DCs disrupted uterine integrity and interferes with implantation through effects on the decidual response (38, 39). However, others have reported that in mice with genetic deficiency in Flt3L, a key DC mitogen and differentiation factor, absence of uDCs is not inconsistent with production of viable offspring in syngeneic matings (40). This implies that at least the remodeling aspects of DC function are redundant and potentially compensated for by other cell lineages.

DCs are considered strong targets for novel therapeutic strategies, as retaining an immature DC phenotype results in better pregnancy outcomes in murine models of immunological rejection of the embryo (41–43). Interestingly, tolerogenic DCs are only effective if administered before copulation, suggesting that the maturation state of DCs at conception is a decisive factor in determining pregnancy outcome, independent of male factors.

The identity of the factors controlling the bidirectional interaction between DCs and the uterus to condition the DC phenotype is a critical question with implications for the development of future therapeutic interventions. Communication in both directions occurs and appears important. The fact that progesterone can rescue abortions induced by DC depletion in mice (44) points to the importance of the hormone environment for a balanced immune response as well as the impact of immune cells on ovarian steroidogenesis function (45).

In humans, immature DCs are abundant in the endometrium, but it is unclear whether their numbers fluctuate through the menstrual cycle (46, 47). The proportion of decidual DC-SIGN⁺ cells is significantly lower in samples from miscarriages versus normal pregnancies (48, 49). Similar to that in mice, the balance between human uDC subsets is relevant for pregnancy maintenance. Studying the factors influencing this balance in vivo is not possible. However, in vitro studies show that ovarian hormones inhibit the ability of human DCs to stimulate T cells, and this may contribute to immune regulation at implantation (50).

In mice, uterine macrophages also fluctuate over the course of the estrous cycle (51), driven by estrogen and progesterone changes (52). As for DCs, the number of murine macrophages in the endometrium is higher when seminal fluid contact has occurred (12). The importance of macrophages in implantation was recently highlighted by experiments showing that specific ablation of CD11b macrophages results in implantation failure in mice (53). Macrophage depletion alters the luteal microvascular network that is necessary for the integrity of the corpus luteum and progesterone production (53) and also impairs expression of embryo attachment glycoproteins on the uterine epithelial surface (54). As uterine macrophages are activated and express MHC class II and costimulatory molecules (42, 55), they likely contribute to presenting Ags to T cells.

In humans, macrophages found at early pregnancy present an immunosuppressive phenotype and secrete TGF-β, IL-10, and IDO (56). Two unique human decidual macrophage subsets have been identified with different gene expression patterns, suggesting potentially distinct functions in the implantation process (57). Interestingly, the human trophoblast cell line SWAN-71 has been shown in a coculture model to recruit monocytes and induce their secretion of cytokines and chemokines in a fashion that acts to support trophoblast survival (58). In contrast, primary human extravillous trophoblasts (EVTs) or villous trophoblasts did not affect cytokine secretion by monocytes or decidual macrophages (59). Human decidual macrophages localize within close proximity to the invasive trophoblast (51) and have been postulated to contribute to trophoblast invasion and placental development (56).

By directly comparing the known roles of DCs and macrophages in reproductive processes, it is clear that both these immune cell types share several functional similarities. In mice, DCs and macrophages have been detected in the thecal layer of the follicles and later in the developing corpus luteum (60, 61), where it is proposed they contribute to follicle rupture, act to suppress excessive proinflammatory events associated with ovulation, and facilitate progesterone production (53, 61). However, recent observations clearly point to these cells also having unique roles in ovarian function. For instance, macrophages are critically involved in maintaining ovarian blood vessel integrity, whereas DCs are shown to support ovarian lymphangiogenesis (61, 62).

Depletion studies of murine uDCs and uterine macrophages both result in implantation failure, further suggesting overlapping roles in the implantation process. This includes the presentation of reproductive Ags to T cells, secretion of immunosuppressive molecules, and a contribution to decidual angiogenesis and trophoblast invasion (63). In contrast to macrophages, DCs seem to be directly involved in the decidualization process by promoting decidual proliferation and differentiation (38, 39).

Mast cells are also critical for embryo implantation. Mice devoid of mast cells, including uterine mast cells, show impaired implantation associated with unfavorable changes in the uterine tissue that interfere with decidualization and trophoblast invasion (3). Remarkably, innate immune cells, and in particular uNK and mast cells, are also critical for the remodeling of uterine spiral arteries (uSA). This process involves the transformation of the thick-walled arteries existing in the nonpregnant state into compliant arteries with thin walls and a dilated lumen, followed by trophoblast effacement and replacement of the endothelial lining (Fig. 2). This highly specialized process is vital for the increased blood flow required as pregnancy advances. Insufficient or incomplete remodeling of human uSA is associated with an increased risk of the severe pregnancy complications, preterm birth, and pre-eclampsia (64).

uNK cells are implicated in uSA remodeling through their secretion of IFN-γ and direct communication with the invading trophoblast. Whereas primary human EVTs seem not to enhance cytokine secretion in uNK cells (59), primary human EVTs as well as murine trophoblast cells have been observed to increase IFN-γ expression in uNK cells (65, 66). Moreover, direct interactions between trophoblast cells and uNK cells are suggested to reduce NK cytotoxicity and render uNK cells more tolerogenic toward fetal Ags (67). However, deficiency in uNK cells does not prevent relatively normal pregnancy progression in mice (68, 69), suggesting that modest limitation of blood and nutrient transport to the fetus is insufficient to cause the gestational hypertension and growth restriction commonly evident in human gestational disorders (64).

Interestingly, uterine mast cells appear to have a function similar to that of uNK cells, as both W-sh and Cpa3 mouse models that lack mast cells show impaired uSA remodeling. Like uNK cell depletion, the impact on fetal growth was only minimal. Notably, however, simultaneous depletion of both uNK and uterine mast cells in mice critically impairs pregnancy. The resulting uSA appear highly abnormal, with thick walls and aberrant retention of smooth muscle cells that usually disappear during early pregnancy, accompanied by substantial fetal growth impairment. At birth, more than half of the progeny were growth restricted (3, 70). A common mediator for both cell types was found to be directly implicated in this effect, namely the chymase mast cell protease 5 (Mcpt5) (70) (Fig. 2). Interestingly, depleting Mcpt5⁺ cells in mice resulted in the very same phenotype as observed after depletion of NK cells and mast cells together. As some fetuses grow normally even when both uNK and uterine mast cells are depleted, we propose that several immune cell types, potentially including DCs (39) and Treg cells (see below), interact and share redundant functions so as to ensure this vital process.

In addition to innate immune cells, adaptive T and B cells are present in the nonpregnant uterus, are increased at the time of implantation, and remain present through pregnancy at the fetomaternal interface. The number of human decidual T cells increases progressively with pregnancy, and effector T cells have been reported to represent ∼60% of the total decidual T cell pool in late gestation (71), which is almost twice their prevalence in peripheral blood. Among human decidual T cells, CD8⁺ cells are more abundant than CD4⁺ cells (72), whereas in mice, decidual CD4⁺ and CD8⁺ T cells are equal in number (73). The majority of human decidual CD4⁺ and CD8⁺ T cells are Ag experienced and represent an effector memory phenotype, whereas naive T cells can only be detected in low numbers (74, 75).

Their repertoire of Ag specificity and phenotypic profiles appears unique to the uterus. Interestingly, decidual effector T cells seem to be more differentiated, to produce more IFN-γ and IL-4, and to show an increased specificity for fetal Ags compared with their peripheral counterparts (76). As progesterone has been found to induce IL-4 expression in human peripheral T cells (77), the local hormonal environment may directly contribute to the specific phenotype of decidual T cells. With regard to their function, a proportion of human decidual T cells is specific for local pathogens such as CMV and EBV, and so these cells may be involved in protection from fetal infection (78). Recently, it has been proposed that human effector memory CD8⁺ T cells possess a mixed transcriptional signature of T cell dysfunction, activation, and effector function, allowing them to support fetal and placental development under physiological conditions and to provide defense against pathogens in the event of infection (79).

However, because of their fetal specificity, decidual effector T cells have the potential to attack fetal tissues. Three control mechanisms are proposed to prevent harm to the fetus: the expression of inhibitory checkpoint proteins, the limited infiltration of effector T cells because of chemokine gene silencing in decidual stroma cells, and the presence of Treg cells (76). Indeed, under inflammatory conditions in mice, the expression of the CXCR3 ligands CXCL9, CXCL10, and CXCL11 and the expression of CCL5 are induced in myometrial stroma cells, eliciting an infiltration of Th1 and cytotoxic T cells (Tc1). This inflammation-induced chemokine expression is suppressed in decidual stromal cells owing to gene-specific modifications of histone marks in the promotor regions of the attracting chemokines and prevents the accumulation of fetal-specific effector T cells in the decidua (80). During normal noninfectious pregnancies, this mechanism of selected chemokine silencing in decidual tissue is proposed to protect the fetus from attack by fetus-specific effector T cells. However, this physiological mechanism supporting fetal tolerance can be disrupted by prenatal infections with Listeria monocytogenes, for example, which has been shown to induce the expression of CXCL9 in decidual neutrophils and macrophages, resulting in the selective attraction of Tc1 cells with fetal specificity and provoking fetal wastage in mice (81).

Although the aberrant expression of chemokines in response to infection is detrimental to pregnancy, in the absence of infection, chemokines play important roles in the initiation and maintenance of pregnancy. Under normal physiological conditions, decidual stromal cells, cytotrophoblast cells, and decidual leukocytes produce a diverse repertoire of chemokines. The respective chemokine receptors are expressed on the surface of leukocytes that are present in abundance within the human decidua, where they act to fulfill critical functions in the implantation process (82). For instance, CCL19 released by mouse uterine epithelial cells in the preimplantation period (83) attracts CCR7-expressing Treg cells, enabling the local accumulation of this critical T cell population. Accordingly, in mice, depletion of Treg cells before implantation acts to negatively affect the uterine environment and hinder implantation, particularly in allogeneic pregnancy (84–86). The absence of Treg cells in CCR7-deficient mice results in a pathological increase in activated effector T cells as well as in uterine inflammation and fibrosis that hinders implantation (84).

Decreased FOXP3 expression and lower numbers of FOXP3⁺ Treg cells have been observed in the endometrium of women with primary unexplained infertility and recurrent implantation failure (87, 88), suggesting that Treg cell deficiency impairs embryo implantation in humans, as it has been confirmed to do in mouse models. Treg cells may contribute to the exaggerated local inflammation observed in the endometrium of women experiencing endometriosis and may explain the accompanying impairment in embryo implantation (89). However, other human studies report contradictory results and suggest deviation in a range of immune cell parameters in endometriosis (90, 91).

To ensure that sufficient Treg cells are present at the time of implantation, uterine Treg cells increase in number at ovulation in response to ovarian steroid hormones in both women and mice (84, 92, 93). Contact with seminal fluid acts to further increase Treg cell numbers in the preimplantation period by selective expansion of the paternal MHC Ag-reactive repertoire (8, 14, 84).

At around the time of implantation, the local uterine milieu, including soluble factors as well as other immune cell populations, appears to positively influence Treg cell recruitment and suppressive activity (Fig. 1). In particular, factors produced by the embryo and trophoblast cells in humans and mice, such as progesterone and human chorionic gonadotropin (hCG), as well as vasoactive intestinal peptide, reportedly act to induce Treg cells locally (94–96) or instruct those cells to immigrate from the periphery into the fetomaternal interface (97, 98). A limitation is that the majority of human studies use choriocarcinoma cells such as JEG-3 cells (96, 98) or immortalized EVT cells such as SWAN-71 (94, 97) and HTR-8/SVneo (97). However, a promoting effect of human trophoblast cells on the recruitment, generation, and suppressive function of Treg cells has been confirmed recently using primary human trophoblast cells (59, 98). In addition, there is evidence for selective migration of human fetal Ag-specific Treg cells from the peripheral blood into decidual tissue (99). Interestingly, disparity between the HLA-C molecules expressed by the fetal trophoblasts and by the maternal tissue promotes induction of Treg cells, whereas HLA-C similarity is inhibitory (100).

Other human decidual immune cell populations possessing immune regulatory properties, including tolerogenic DCs and M2 macrophages, also enhance local Treg cell activity (101, 102). Because placental factors support the generation of tolerogenic DCs and M2 macrophages (103), it can be postulated that the fetus itself may also contribute to robust implantation by building up a complex network of immunosuppressive mechanisms.

Notably, Treg cells are most critical for the peri-implantation phase, and studies of mice imply that Treg cells play a secondary role for fetal survival once implantation is completed (86). However, the redundancy of Treg cells at later pregnancy stages is still a matter of debate, as other researchers have shown that sustained expansion of murine Treg cells is required for maintaining fetal tolerance throughout pregnancy. Midgestational ablation of murine Treg cells resulted in nearly complete fetal rejection and elimination of live pups born (104). Consistent with this, in a syngeneic setting where the only fetal alloantigen is the male H-Y Ag, Treg cell depletion after implantation results in selective rejection of the male fetuses, suggesting that even in the presence of low fetal antigenicity, Treg cells are required to protect against rejection (105). The potential reasons for the discordant observations may be found in the different methods used for Treg cell depletion (application of diphtheria toxin in Foxp3–diphtheria toxin receptor (DTR) mice engineered to selectively express primate DTR under control of the Foxp3 promoter versus use of anti-CD25–depleting Ab). Whereas diphtheria toxin selectively kills Treg cells in Foxp3-DTR mice, depletion of CD25-positive cells will also reduce the number of activated effector T cells with potential to react to fetal alloantigens.

By virtue of their regulation of DCs and uNK cell phenotypes, Treg cells may be particularly important in facilitating the decidual response in mice (39). Like mast cells and uNK cells, Treg cells exert influence on the maternal vascular adaptations required for robust placental development. Rodent models of pre-eclampsia indicate a critical function for Treg cells in normal placental development through coordinated interactions with uNK cells, DCs, and macrophages. Experiments in mice deficient in T cells and/or NK cells show that T cells interact with uNK cells to influence the maternal hemodynamic response to pregnancy (106, 107). Peripheral Treg cells may be particularly important, as mice with a null mutation in the CNS1 gene essential for peripheral Treg cell generation have impaired remodeling of maternal spiral arteries and defective placental development (108). Recently, we demonstrated in Foxp3-DTR mice that Treg cells are required in early gestation to promote maternal uterine artery adaptation, such that Treg cell–deficient mice exhibit increased artery resistance and pulsatility associated with fetal loss (109). Incomplete transformation of uSA is also seen in mice in which neutrophil depletion impairs proangiogenic, neutrophil-induced Treg cells (110).

Additional evidence that Treg cells are pivotal in preventing inflammation and progression to pre-eclampsia–like symptoms comes from recent experiments in rats. In the reduced uterine perfusion pressure (RUPP) model, reduced uterine artery blood flow is induced by surgical intervention. The pre-eclampsia–like symptoms induced in this model are T cell–dependent because T cell–deficient athymic rats are resistant to RUPP-induced hypertension and fetal growth restriction, and disease can be induced by passive transfer of Th17 effector CD4⁺ T cells (111). The symptoms are mitigated when Treg cells from pregnant control donor rats are administered shortly after the RUPP procedure (112).

In contrast to other immune cell populations, B cells are rare in the female reproductive tract (113), and their abundance does not appear to fluctuate during the menstrual cycle (114). Nevertheless, there is evidence that trophoblast-derived signals can influence decidual as well as peripheral B cells. Syncytiotrophoblast microvesicles that are shed from the human placenta into the maternal circulation can bind to B cells and induce the release of several cytokines that support a shift toward Th2 responses (115). Moreover, the presence of Th2 cytokines at the fetomaternal interface affects placental B cells in their ability to produce asymmetric Abs, a specific Ab type known to protect the foreign fetal Ags (116). In contrast, recognition of fetal Ags by murine fetal-specific B cells triggers the deletion of a bone marrow B cell subpopulation, including immature and transitional B cells (117). This process facilitates fetal survival by depleting potentially deleterious maternal B cell populations.

Our recent studies indicate that the trophoblast-derived hormone hCG may impact B cells and force acquisition of a pregnancy-protective phenotype. hCG not only induces the secretion of asymmetric Abs but also provokes the conversion of human conventional B cells into regulatory B (Breg) cells (118, 119). The induction of human Breg cells, observed to secrete high levels of the anti-inflammatory cytokine IL-10, is a characteristic of normal pregnancy, and an impaired Breg cell augmentation is associated with pregnancy failure (119). Notably, progesterone and estrogen do not induce Breg cells in our hands. In observations similar to those made for innate immune cell populations, we found that T and B cells are influenced by fetally derived factors and exhibit important functions that are critical for fetal survival.

In summary, immune cells adapt to combinations of specific cytokine and hormone signals in the uterine environment to acquire unique functions and execute pivotal processes during pregnancy. Insufficient DCs, macrophages, mast cells, and Treg cells can in each case impair implantation and compromise healthy progression of pregnancy. The consequences of immune cell disturbance are reduced fertility and greater likelihood of pregnancy disorders. These disorders begin with poor endometrial receptivity through effects on the decidual transformation response and/or failure to elicit appropriate adaptations in the maternal vasculature. In the case of Treg cells, although their anti-inflammatory and vascular remodeling effector functions are likely to be paternal Ag independent, greater expansion of Treg cell populations and greater capacity to suppress effector immunity to paternal alloantigens is seen when priming to paternal Ags has occurred in the context of seminal fluid. Aspects of the remodeling functions of different cell lineages may be redundant, particularly between uNK cells and uterine mast cells, but, clearly, an optimal environment is achieved when each of the immune cell lineages works in cooperation.

Emerging insights on factors controlling uterine immune cells and their contribution to placental development are essential for tackling clinical conditions of intrauterine growth restriction and pre-eclampsia, which increasingly are recognized to have their origins at the implantation phase. Strategies to target uterine immune cells and/or their mediators hold enormous potential as novel treatments for these common and important disorders. Future research can be expected to focus on developing an integrated understanding of how it is not just hormones and cytokines that control immune cell function; paternal factors, the microbiome, nutrition, genetics, and other lifestyle factors also contribute.

The authors have no financial conflicts of interest.