Adaptive Immune Responses During Pregnancy



Ana Claudia Zenclussen, Experimental Obstetrics and Gynecology, Medical Faculty, Otto-von-Guericke University Magdeburg, Gerhart-Hauptmann-Str 35, 39108, Magdeburg, Germany.



It has long been believed that there is no immune interaction between mother and conceptus during pregnancy. This concept changed after evidence was provided that the maternal immune system is aware of the semiallogeneic conceptus and develops strategies to tolerate it. Since then, finely regulated mechanisms of active tolerance toward the fetus have been described. This Special Issue of the American Journal of Reproductive Immunology deals with these mechanisms. It begins with the description of minor histocompatibility antigens in the placenta; it further goes through adaptive immune responses toward paternal fetal antigens, mostly concentrating on regulatory T cells and molecules modulating the Th1/Th2 balance. The participation of antibody-producing B cells in normal and pathological pregnancies is also discussed. This introductory chapter resumes the concepts presented throughout the Issue and discusses the clinical applications raised from these concepts.

Introduction: the beginnings

In 1953, the Brazilian Sir Peter Medawar initiated the modern immunology of pregnancy by asking: ‘how does the pregnant mother contrive to nourish within itself, for many weeks or months, a fetus that is an antigenically foreign body?’.[1] In his article, three theories were proposed to explain the lack of an immunological reaction from the mother against the fetus:

  1. The anatomical separation of the fetus from the mother: according to this theory, a barrier impermeable to cells would separate fetal–maternal and blood circulations. This theory was proved to be wrong, because it is now known that the fetal–maternal interface is a bi-directional exchange surface. Not only maternal immune cells are present at the feto–maternal interface, but also fetal cells leak and are found in maternal circulation. Maternal cells can be found in the babies as well. This phenomenon is called feto–maternal microchimerism, being microchimerism defined as a small non-host cell population (or DNA quantity) from one individual harbored by another individual.[2] Many reports indicate that microchimerism persists in mother and child even decades later,[3-5] and new studies reveal that feto–maternal chimerism occurs even very early in pregnancy.[6, 7] This clearly changes the once accepted concept of anatomical separation between mother and fetus – that wrongly persists in many text books – and dismisses therefore the idea that no adverse immune reactions between two different individuals can take place as they clearly ‘see’ each other.
  2. The antigenic immaturity of the fetus: this theory pointed out that the fetus is ‘antigenically immature’, not expressing histocompatibility antigens and being therefore not able to provoke classical adaptive immune responses. This theory collapsed very quickly, because fetal skin dendritic cells that are positive for MHC class I but negative for MHC class II are very potent accessory cells in polyclonal T-cell responses.[8] Furthermore, in the last years, a new concept emerged by which fetuses can protect themselves from maternal immune reactions as their CD4+ cells strongly differentiate into tolerogenic Tregs that actively tolerate maternal cells that reside in fetal tissues.[9] This concept is highlighted in the review by Dr. Trevor Burt and will be also discussed later during this introductory chapter.
  3. The immunological inertness of the mother: A pregnant mother does not die if she gets a cold. It is even showed that pregnancy is associated with inflammation rather than with inertness.[10] This alone dismisses the concept of immunological inertness. However, this initial theory gave rise to the concept of active tolerance mechanisms against the fetus. Nowadays, it is considered that the mother achieves a state of tolerance against the fetus, still being able to elicit normal immune responses against infections. This was first shown by Dr. Ana Tafuri et al. in a paper where she demonstrated that paternal T cells are aware of the presence of paternal antigens during pregnancy, where they acquire a transient state of tolerance specific for paternal antigens.[11] This groundbreaking piece of information was the basis for many studies concentrating on the mechanisms as to how the maternal immune system tolerates the fetus rather than ignoring it. In the 1990s, many studies concentrated on the cytokines secreted by T cells. Later, regulatory T cells (Treg), whose main function is to prevent autoimmunity,[12] emerged as important players in regulating tolerance toward paternal and fetal antigens, and this is discussed in three reviews within this issue from different optics (Dr. Robertson, Dr. Saito, Teles). It is clear that not a single cell but rather a network of communicating cells and molecules is responsible for the successful outcome of pregnancy, and this is not only discussed in many but not all reviews but also highlighted in this introductory chapter.

Paternal and fetal antigens are seen by the maternal immune system, and minor histocompatibility antigens are expressed at the fetal–maternal interface

Tafuri[11] elegantly showed that maternal T cells are aware of paternal components in fetal cells and actively protect them during pregnancy. This transient state is only specific to paternal antigens.[11] Since this publication, much effort has been done in understanding how paternal antigens are recognized. The review by Peggy Petroff and collaborators extensively revises this issue, as also do Drs. Tilburgs and Strominger. In the last years, special attention has been paid to minor histocompatibility antigens (mHAgs). Its role in eliciting an immune response has been clearly highlighted in transplantation studies. It is known that they modulate graft rejection and graft versus host disease in HLA-matched transplant recipients.[13] mHAgs can be both protective and dangerous for the transplant acceptance.[14] The role of mHAgs for pregnancy has been first suspected after observing that parous female donors are more likely to elicit graft-versus-host disease in transplant recipients when compared to either non-parous or male donors (reviewed in ref.[15]). In mice, presentation of fetal antigens to maternal T cells can begin, as already discussed, as early as at copulation (reviewed by Sarah Robertson). In women, T cells specific for mHAgs were described (reviewed in Lindscheid's review). These cells can be still present up to 20 years after birth. The current hypothesis is that these cells are of tolerogenic or suppressive nature, which at pregnancy allows the survival of the fetus. Whether at later stages their persistence is beneficial or rather detrimental because of the possibility of eliciting autoimmune responses is a matter of debate.

Where it all begins: antigens present in the seminal fluid activate the adaptive immune response that tolerates the fetus

After the emergence of the new concept that postulates the existence and the need of a protective adaptive immune response necessary to protect the fetus,[11] investigators concentrated on the cells responsible for this state of active tolerance. It is of general consensus that Treg mediate in large part the state of active immune tolerance that prevent maternal lymphocytes to cause cytotoxic damage to the fetus.[16-21] It was first wrongly believed that the expansion of this unique cell subpopulation was driven by pregnancy itself, for example, hormones, and not by alloantigens.[16] This could, however, not explain why Treg are necessary before implantation[17, 22] and the fact that Treg from non-pregnant mice or from pregnant females carrying third party fetuses cannot confer fetal protection in a model of disturbed tolerance.[17, 23] It was Sarah Robertson's pioneer work that introduced the concept of seminal fluid as the first antigen source that activates the maternal immune system and prepares for pregnancy establishment.[24] In their review, Robertson and collaborators summarize current evidences as how the seminal fluid elicits a female immune response and particularly concentrate on the events leading to generation and expansion of Treg in the peri-conception and peri-implantation period. There is plenty of evidence that associates the early expansion of the Treg pool with the exposure to seminal fluid (Robertson's review). Treg must first encounter antigens presented by antigen-presenting cells, as for example, dendritic cells (DCs) in an appropriate cytokine environment,to proliferate and be functional. The encounter of seminal fluid with maternal DCs present either in vaginal lumen or in endometrial tissue at the time of mating[25] represents therefore the first event leading to a protective adaptive immune response. It has been proved that uterine DCs are rather tolerogenic DCs than mature DCs[26] and Heme Oxygenase-1 (HO-1) seems to be pivotal in maintaining maternal DCs in an immature state, which contributes to the expansion of the peripheral Treg population.[27] Not only does the seminal fluid provide the antigens to be presented to APCs but also recruits Treg into the uterus or drives their expansion and the draining lymph nodes as demonstrated by the lack of expansion of Treg after in females mated with males without seminal vesicles but not with vasectomized males[19] (A. Teles, A. Schumacher and A. Zenclussen, unpublished observations). Furthermore, seminal fluid contains potent immune suppressive molecules that contribute to Treg induction or conversion of conventional T cells into Treg, such as TGFbeta and PGE-2-related prostaglandins in the plasma fraction (reviewed in Robertson). It was recently showed that seminal plasma promotes the differentiation of human DCs to tolerogenic ones.[29]

CD4 cells: from Th1/Th2 to Treg/Th17

Much attention was focused in the production of cytokines by CD4 T helper cells particularly after Piccinni and colleagues showed that decidual T cells from women with unexplained recurrent abortions produced abnormally low Leukemia inhibtory factor (LIF), IL-4, and IL-10 levels.[29] This was further supported by data from patients who presented a Th1 phenotype in cases of abortion versus a Th2 phenotype in normal pregnancies.[30] The very popular so-called Th1/Th2 paradigm collapsed after the report of normal pregnancies in mice knockout for IL-10 and even in quadruple knockouts for Th2 cytokines.[31],[32] Thus, it was clear that the Th1/Th2 ratio was a marker for successful or failing pregnancy but not the cause of it and that much more complex mechanisms are involved in pregnancy establishment and maintenance.

The existence of cells with suppressive capacity was already suspected in the early 70s, but the first confirmation and characterization of these cells was performed by Shimon Sakaguchi in 1995, who called these cells regulatory T cells (Treg).[12] Treg are a subtype of CD4+ T cells that also express CD25 and are able to actively suppress self-reactive lymphocytes and thus to maintain immunological self-tolerance.[12] Since then, further characteristics of these cells are known as, for example, that they express the transcription factor Forkhead box p3 (Foxp3)[33] and that they can derive from the thymus (the so-called naturally occurring Treg) or be induced in the periphery (iTreg) upon different conditions (Teles' Review). It was first proposed by Somerset that Treg may be important for human pregnancy as they were elevated in normal pregnancy.[34] At the same time, Aluvihare and colleagues show that the reconstitution of Rag−/− mice with T cells lacking the CD25+ fraction was related to a higher rate of abortions as compared to mice that received the whole T-cell fraction.[16] The therapeutic potential of Treg for pregnancy was described by us shortly thereafter in a model of disturbed tolerance during pregnancy.[17] It is nowadays known that Treg fluctuate in number in blood and uterus during the receptive phase of the menstrual or estrus cycle,[35] (A. Teles et al., unpublished data), which is interpreted as a requisite for pregnancy establishment further underlined by the fact that impaired increase or diminished suppressive capacity is associated with infertility or pregnancy complications.[36-38] As it was discussed already, seminal fluid is pivotal in expanding Treg[39] and local application of TGF-β in the mice had the same effect.[40] The antigen specificity of Treg was demonstrated both in a mouse model by Schumacher and Zhao in 2007[23, 41] and in human by Tilburgs.[42] As for the mechanisms of action of Treg during pregnancy, it has been proposed that they act by creating a local tolerant microenvironment[43] that IL-10 and PD-1 but not TGF-β or CTLA-4 are relevant for pregnancy.[23, 44] In humans, however, CTLA-4 expressed in Treg cells up-regulates IDO expression on decidual and peripheral blood DC and monocytes by the induction of IFN-γ production.[45] Decidual Treg seem to work by cell–cell contact[46] and suppression of T-cell responses.[47] Dr. Saito extensively revises the type of Treg present during pregnancy and discusses the importance of correctly identifying their phenotypes for future clinical applications.

The identification of Th17 cells has improved our understanding of the cellular regulation during pregnancy.[48] IL-17 acts mainly against extracellular bacteria or fungal pathogens. It seems that an imbalance of Th17/Treg proportion is associated with recurrent pregnancy loss and pre-eclampsia.[48, 49]

Tolerance from the fetus toward the mother

In the last years, we have learned more and more how Treg protect the fetus from immunological attack by the maternal immune system, but it was not until recently that it became clear that the fetal immune system is also active and could potentially danger the mother.[9, 50] This does not happen because the fetus actively generates tolerance to maternal antigens, mostly mediated by fetal Treg.[9] As discussed in Dr. Burt's review, Treg are abundant in the developing fetus. In fact, the frequency of Treg in fetal tissues is much higher than the frequency of Treg in any tissue compared with any other time in development (Dr. Burt). Maternal cells were found to be present in fetal lymph nodes and in cord blood.[9] In vitro, fetal immune cells were rather suppresive against maternal antigens as compared to responses against unrelated alloantigens.[9] This suggested that fetal immune cells are already primed against maternal antigens. The existence of a normal immune response toward maternal antigens upon depletion of CD25+ fraction in the fetal T cells revealed the existence of fetal Treg that are educated to react toward maternal antigens.[9] Although the nature of fetal Treg is still a matter of debate, it is tempting to speculate that fetal Treg are derived from conventional T cells that become functional suppressor cells, thus Treg, upon antigen stimulation (Dr. Burt). It is a challenge for the near future to understand how and at which time point of pregnancy this system comprised mostly of cells that are programmed to suppress convert into a system with a majority of cells with fully potential to elicit a normal aggressive immune response. This is crucial to understand how immunity to pathogens can be reached at different neonatal ages. In this regard, Dr. Burt revises the hypothesis of the layered immune system.

Modulators of the immune responses during pregnancy

Several molecules modulate and influence the cells that are directly involved in the generation and maintenance of an active immunotolerance toward the fetus.

Pregnancy hormones are of enormous importance for pregnancy maintenance. It is now known that they also influence the immune system. Estradiol was claimed to stimulate the expansion of Treg in mice,[51] although this alone does not count for the expansion observed upon pregnancy establishment as discussed before. Estradiol application further decreases the production of IL-17 by T cells.[52] Elevated progesterone during pregnancy inhibits the development of Th1 immune responses during pregnancy.[53] Progesterone in synergy with Galectin-1 (Gal-1) is reportedly involved in pregnancy maintenance,[54] and the application of one progesterone derivate, dydrogesterone, can abrogate abortion triggered by stress in a mouse model because of a deviation of the immune response to a Th2 one.[55, 56] Estrogen and progesterone in combination were necessary for the recruitment of mast cells (MCs) to the uterus.[57] MCs were recently reported to be pivotal for implantation and placentation.[58] The most important pregnancy hormone, the human chorionic gonadotropin (hCG) secreted by the trophoblasts, was shown to attract regulatory T cells to the fetal–maternal interface[59] but also to foster their suppressive function.[60] Thus, hormones are important components of the adaptive immune answer necessary to guarantee the survival of the fetus within the maternal uterus without eliciting classical maternal immune responses that would target the fetus and reject it.

Besides hormones, chemokines produced by trophoblasts are thought to recruit Treg into the fetal–maternal interface.[61] Chemokines were also reported as important factors as their epigenetic silencing blocks the access of classical T cells to the fetal–maternal interface.[62] The review by Drs. Perez Leiros and Ramhorst concentrates on the recruitment of immune cells that contribute to tolerance by immune polypeptides that also contribute to tolerance maintenance. Regulated on activation, normal T cell expressed and secreted (RANTES) can suppress maternal allogenic immune responses to paternal antigens in mixed lymphocyte cultures.[63] RANTES is produced by the peri-implantation endometrium, by human endometrial T cells and by trophoblasts (reviewed by Perez Leiros and Ramhorst). This molecule can additionally induce apoptosis of potentially harmful maternal CD3+ cells and increases the frequency of Treg.[64] Another molecule in the focus of the review by Perez Leiros is vasoactive intestinal peptide (VIP), whose anti-inflammatory and tolerogenic effects were already known.[65] It is now known that VIP levels rise at the fetal–maternal interface at early gestation peaking at placentation begin.[66] Its role in embryogenesis was revealed after observing that its blockade during midgestation ends in induced growth retardation and microcephaly.[66] This is further confirmed in VIP−/− fetuses that highlights the role of maternal VIP for early neural development.[67] Current data position VIP as an important immunomodulatory molecule as it can increase the frequency of Treg and LIF at implantation sites in mice[68] and supports a tolerogenic macrophage phenotype.[64] It seems that both hormones and polypeptides are involved in the recruitment of immune cells into the fetal–maternal interface and in the generation of a tolerogenic immune response toward the fetus.

Accumulating evidence points galectins, a family of endogenous glycan-binding proteins as an important regulator of pregnancy, and this is discussed in detail in the Review by Drs. Blidner and Rabinovich within this Special Issue. It has been shown that the interaction between endogenous glycan-binding proteins and glycosylated receptors is of crucial importance for immunological homeostasis (reviewed in ref.[69]). Within these processes, galectins emerge as important regulators in several physiological and pathological processes (reviewed in Blidner and Rabinovich). In particular, galectins were lately reported as regulators of feto-maternal tolerance. Galectin-1, expressed in T and B cells, inflammatory macrophages tolerogenic DCs, uterine NK cells, uterine MCs, and Treg, gained much attention over the last 5 years in the reproduction research field. Gal-1 is known to define autoimmunity,[70] inflammatory neurodegeneration,[71] cardiac inflammation,[72] and tumor escape.[73] Lgals-1−/− mice, lacking Gal-1 expression, show increased Th1 and Th17 responses, more immunogenic DCs, aberrant microglia and display more autoimmune pathology than their wild-type counterparts (reviewed in Bildner and Rabinovich). The presence of Gal-1 at the fetal–placental interface is long known,[74] and recent data confirm that it is not only expressed by the placenta itself[75] but also in uterine NK cells,[76] tolerogenic DCs,[55] and recent data show its importance when expressed in uterine MCs.[58] Pregnancies of Gal-1-deficient mice were first described as normal as the number of born embryos was apparently unaffected by this mutation.[77] However, they present a higher abortion rate than wild-type controls if paired allogenically,[55, 58] a defect that can be completely reverted when transferring bone marrow-derived MCs from wild-type animals.[58] It seems therefore that it is the Gal-1 secreted by uterine MCs the decisive factor in preventing pregnancy abnormalities. Gal-1 secreted by MCs positively influences spiral artery formation and thus placentation, which finally allows the normal growth and development of the fetus within the uterus.[58]

Another molecule influencing pregnancy at various checkpoints is the heme-degrading enzyme heme oxygenase-1 (HO-1). HO-1-deficient females were initially reported as sterile after observing that no progeny could be obtained after mating Hmox1−/− females with Hmox1−/− males.[78] We have observed that this is not the cause. Hmox1−/− females are not sterile or even infertile. They do get pregnant, but their fetuses do not survive to term if they are homozygote while heterozygote fetuses usually do survive.[79] Fetuses that die intrauterine do so because of the accumulation of toxic-free heme that leads to defective implantation and placentation, and this defect can be completely corrected after inhalation of low doses of carbon monoxide, the most prominent HO-1 metabolite.[79] We also found HO-1 expression to define oocyte ovulation and its fertilization.[80] Besides this prominent role in ovulation, implantation, and placentation, HO-1 was recently found to be important for the acquisition of immunotolerance toward the fetus. HO-1 supports a tolerogenic phenotype for DCs that in turns expands the Treg population. Consequently, the inhibition of HO-1 activity resulted in miscarriages even after transfer of pregnancy-protective Treg.[27]

The review by Martinez and collaborators concentrated on the role of pregnancy-specific glycoprotein 1a (PSG1a) in regulating the adaptive immune responses during pregnancy. PSG1a belongs to the family of PSG that are mainly synthesized by the placenta and represent early biochemical markers of syncitiotrophoblast formation.[81] It has been reported that PSG1a can turn macrophages tolerogenic and increase their ability to produce IL-10 and TGF-beta.[82, 83] PSGs and particularly PSG-1a can inhibit T-cell proliferation but not directly, they rather do this via macrophages (reviewed in Martinez). Similarly to what is reported for Gal-1[55] and HO-1[27] among others, PSG1a seems to modulate DC phenotype and maturation that finally promotes the secretion of IL-17.[84] Other PSGs are also of relevance for establishment of pregnancy as elegantly demonstrated by the group of Gabriela Dveksler.[85, 86]

CD8 cells: their transformation during pregnancy

In addition to allo-specific responses to paternal antigens by CD4+ cells, recent data reveal the presence of highly differentiated CD8+ memory cells that are present at the fetal–maternal interface.[87] This special topic is intensively discussed in the review by Tilburgs and Strominger. The existence of memory CD8+ cells at the fetal–maternal interface implies the possibility of antigen presentation. The target specificity for decidual CD8+ cells that constitute the most abundant T-cell subset in this tissue is unclear, and the source of the antigens to which CD8+ cells would be specific is discussed in the mentioned review. It is known that many mothers carry naïve or memory CD8+ T cells with a T-cell receptor (TCR) that can directly bind and respond to paternal MHC molecules that are expressed by fetal trophoblast cells. It is further known that pregnant women develop T-cells responses that are specific to fetal mHAgs through the indirect allorecognition pathway (Review by Dr. Petroff and collaborators). The nature of this immune response is controversially discussed in the literature. Whether some of the decidual CD8+ cells are protective is also unknown. Some evidence indicate this may be the case, for example, Clark proposed that TGF-β intravaginal application recruits CD8+ Foxp3+ cells,[40] but this concept was no longer followed. Interestingly, CD8+ decidual cells were reported to be responsible for the protective effects of progesterone application.[88] An important aspect is the fact that viral infections can skew the CD8+ cell repertoire and can alter the dynamic of CD8+ cells during pregnancy. A pro-inflammatory profile may augment the influx of T cells into the fetal–maternal interface, and Treg may not be longer able to protect from effector mechanisms and/or activate antigen-specific CD8+ cells to attack fetal structures. Whether this is the case after viral infections remains to be elucidated.

The final players: B cells, directing the immune response to a protective or a pathologic one

The final arm of the immune system is depicted by B cells. These cells are in charge of producing antigen-specific antibodies that are in charge of bind and destroy foreign antigens so that they can be easily phagocytized and also activate the complement system. Different B-cell types circulate in blood and can become fully activated after binding the foreign antigen through the B-cell receptor (BCR) and the signal from T helper cells. This will finally lead to the transformation into plasma cell and the antibody secretion.[89] B cells can, however, also present antigen and produce cytokines. Little is known about the participation of B cells in pregnancy, and their study has been mostly concentrated on the production of autoantibodies during pathologic pregnancies. The review by Damian Muzzio and collaborators gives an overview about the current knowledge of B-cell participation in pregnancy. In the 1980s and 1990s, several groups concentrated on the production of pregnancy-protective antibodies by B cells. It was first observed that the cytotoxic effect of maternal lymphocytes to trophoblast was hampered in the presence of maternal serum,[90] which introduced the concept of antibodies against paternal components that are protective and not deleterious and thus support pregnancy. Mowbray proposed that the modulation of HLA expression in trophoblasts is crucial for pregnancy outcome and that the absence of maternal antibodies against these antigens is a cause of recurrent spontaneous abortion.[91, 92] Regarding the nature of these antibodies, Margni and collaborators have demonstrated the existence of so-called asymmetric antibodies that are characterized by the presence of a high-mannose residue in one F(ab) regions. This makes these antibodies able to bind to the antigen but unable to trigger the classical immunological mechanisms aiming to destroy it.[93] It has been postulated that these antibodies are the ones that protect from a destructive maternal immune response to paternal antigens. Not only have these antibodies been identified in the placenta[94] but also found to be specific against paternal antigens.[94] Patients suffering from RSA have reportedly a diminished proportion of these particular antibodies.[95] The nature of B cells producing these antibodies has not been so far studied.

B cells are also able to produce antibodies that are harmful for pregnancy. The most studied antibodies in relation to infertility, spontaneous abortion, and pre-eclampsia are antibodies in the context of the antiphospholipid syndrome (APS). Women affected by APS produce one or more antiphospholipid antibodies (aPL) that are directed against phospholipids (e.g., the ones that are present in the trophoblast membrane and get exposed to the external surface upon tissue remodeling) but also against molecules like lupus anticoagulant protein, cardiolipin, β2 glycoprotein 1, prothrombin, annexin, phosphatidyl ethanolamine, and phosphatidyl inositol (reviewed in ref.[96]). APS manifests often for the first time when patients get pregnant and suffer a miscarriage, being the predominant obstetric complication the recurrent occurrence of spontaneous abortions.[97] As for the mechanisms, aPL themselves may not be the cause of fetal loss. It is proposed that aPL induce a procoagulant phenotype that results in fetal growth restriction.[98] Inflammation seems to be needed to cause placental injury in patients with APS as not all pregnancies in patients with APS result in a complication.[99] In mice, aPLs can bind to the invading trophoblast and hinder implantation or diminish placenta perfusion causing infarction.[100] Girardi and colleagues have shown that inflammation-driven complement activation is an important pathway resulting in thrombosis and endothelial cell activation.[101] They recently proposed tissue factor (TF) to be an important effector in aPL-related inflammation.[102]

Several other autoantibodies have been lately related to pregnancy complications, being the most prominent one the autoantibody against angiotensin II type I receptor (AT1-AA). The review by Herse and LaMarca highlights its participation in pre-eclampsia (PE) (Herse & LaMarca). AT1-AA were first described by Wallukat and colleagues in sera from pregnant women developing pregnancy-induced hypertension.[103] During PE, the renin–angiotensin system is dysregulated and leads to the presence of activation AT1-AA in the circulation of these patients. Several methods confirmed the binding of AT1-AA to the AT1-receptor.[103] In the last years, it became clear, however, that AT1-AA are not specific for PE as they can be found in normotensive pregnant whose fetuses suffered from uterine growth restriction[104], in kidney-transplanted patients who presented a refractory vascular rejection,[105] in patients with systemic sclerosis,[106] and patients with malignant secondary hypertension.[107] As for the mechanisms of AT1-AA-induced pathology, it could be demonstrated that IgG isolated from pre-eclamptic women (a fraction containing AT1-AA) activates the complement system in kidney and placenta when administered to mice.[108] Besides, animals exposed to AT1-AA presented pre-eclampatic symptoms[109] as well as elevated levels of soluble fms-related tyrosine kinase (sFlt-1) and soluble endoglin (sEng),[110] which directly links AT1-AA to two well-characterized PE markers. Federico Jensen and colleagues recently described the subtype of B cells, which is able to secrete AT1-AA in PE patients, the B1aB cells. These cells produce without antigenic stimulus antibodies that are polyreactive and can turn autoreactive depending on the conditions. hCG, reported to be elevated in PE patients, was identified as one factor modulating the AT1-AA production by B1aB cells.[111] The application of an antibody that reduces the B-cell population could restore blood pressure and endothelin production in a rat model for PE.[112]

Methodological advances in studying immune cells at the fetal-maternal interface

The study of immune cells at the fetal–maternal interface fascinated immunologists many years ago already. The first methods employed to study cells and their distribution were based on their immunohistochemical staining in fixed samples. This allowed understanding which immune cells were present at the fetal–maternal interface and whether there were differences between normal and pathological specimens. Later, with the description of techniques to isolate and keep in culture different immune cells, we could learn more about their function. The use of flow cytometry allowed and allows a better characterization of their phenotype. With growing knowledge of the role of immune cells at the fetal–maternal interface, it became clear that the interaction of cells and not one single cell was responsible for physiological or pathological processes. This led to the establishment of co-culture systems and system employing chambers to understand both the interaction between cells and the migration of cells to gradients or other cell types. Seminal work on this field has been performed by the group of Dr. Mor who could show how trophoblast cells secrete molecules and ‘educate’ the immune system to best tolerate the conceptus.[113] The more we learned from the cells we studied, the more became clear that these cells may act differently in vivo as isolation protocols and mostly their maintenance in culture change their phenotype and therefore most probably their functionality as well. The review by Drs. Olivieri and Tadokoro deals with this interesting topic. One of first experiments to visualize processes in vivo at isolated organs was described by Ruttner et al. in a superfusion chamber at which the microvascular flow could be observed.[114] Later, a skin transplanted uterus was in vivo analyzed, and the effects of several agents were observed mostly for endometriosis research.[115] Dr. Tadokoro recently described for the first time the in vivo imagining of immune cells at the uterus and the placenta by using a 2-photon microscopy.[116, 117] This amazing method will help us understand how cells behave in their natural environment. With this method, Dr. Tadokoro described that DCs accumulate at the estrus phase of the estrus cycle in clusters at the probably future implantation sites and that their number and density is much higher as we suspected because of immunohistochemistry and flow cytometry studies. The use of different genetically modified mice whose cell fluoresce in different colors will allow us to understand how and where immune cells interact with each other, how they migrate, and how they respond to several stimuli.

Clinical applications: are we there yet?

Much knowledge has been obtained in the last few years about processes that enable the tolerance of the growing fetus in the maternal uterus and how disturbances of these fine regulated processes can lead to pathologies with devastating consequences for both mother and fetus. It is now our challenge to transform all this information, mostly obtained from experimental models either in animals or with in vitro studies to develop strategies tending to help reproductive challenged couples who cannot procreate because of aberrant immune responses to the fetus.

In vitro, human seminal fluid can keep human DCs in a tolerogenic status[28] and promote the induction of Treg phenotype in T cells.[19] Understanding the molecules involved in this will help improving IVF protocols. The more we know about mHAgs that are related to spontaneous abortions,[118] the more we will be able to faster diagnosis the causes for miscarriages and designing strategies for their treatments.

It has been described that hCG efficiently attracts Treg to the fetal–maternal interface,[59] and now it is clear that it also foster a protective immune response.[60] This lead to a clinical trial addressing the question whether hCG application can improve implantation (

In the last years, the idea of paternal antigens leading to a protective Treg-mediated immune response was confirmed in studies where Treg increase after paternal lymphocyte immunization therapy[119] or in vitro fertilization treatment.[120] This information is very useful and may lead to re-think about the so controversial therapy with paternal leukocytes[121] and to dissect which patients should be treated. The large list of molecules that positively regulate the adaptive immune response during pregnancy will surely lead to the design of strategies to modulate this also for patients suffering from pregnancy complications.

The understanding how the fetal immune system rapidly switches from suppression to active immunity is very important for combating neonate infections naturally. The concepts introduced in this Introductory Chapter are resumed in (Fig. 1).

Figure 1.

This hypothetical scenario depicts the current knowledge about the pathways involved in recognition and tolerance of the foreign fetus as discussed throughout this Special Issue. Paternal antigens are presented to the maternal immune system in the vaginal lumen after the encounter of maternal/paternal immune cells with antigens present in the seminal fluid. The seminal fluid contains also substances that promote the conversion of dendritic cells (DCs) in tolerogenic ones. This promotes the conversion and expansion of regulatory T cells (Treg). The continuous release of paternal antigens to the circulation allows that Treg continue emerging and expanding throughout pregnancy in, for example, the para-aortic lymph nodes. In peripheral blood, Treg are likely involved in the suppression of maternal effector T cells as, for example, Th17 cells that could be harmful to fetal antigens present here as well as in several maternal tissues. In a normal pregnancy, B cells secrete antibodies that, once at the feto-maternal interface, protect paternal antigens present in the trophoblast. Treg migrate to the fetal–maternal interface via human chorionic gonadotropin(hCG). At the feto–maternal interface, a broad spectrum of molecules produced or secreted by the trophoblast itself like hCG, HO-1, RANTES, and PSG modulate the phenotype of function of immune cells, that is, DCs, that turn or stay immature and thus tolerogenic. Additionally, molecules secreted by cells of the innate immune system like Gal-1 secreted by uNKs and uMCs can positively influence the physiology of the trophoblast while helping maternal T cells to become or stay tolerant toward the fetus. The presence of CD8+ cells at the feto–maternal interface is well documented, but their function is still under investigation. This overview does not pretend to cover all information available but the research topics covered within this Issue.


The maternal immune system undergoes profound transformations already at the very beginning of pregnancy. These prominent changes are directed to protect the fetus from a detrimental immune response. Growing evidence body document that factors secreted by the fetal trophoblast itself contribute to these changes. Accordingly, the fetal immune system is also programmed to ‘tolerate’ the mother and therefore survive the pregnancy until term. The investigation of cells, molecules, and pathways involved in these processes is vital to help patients with infertility or pregnancy complications as well as neonates.


This work was supported by grants from the Deutsche Forschungsgemeinschaft (ZE 526/4-1, ZE 526/4-2, ZE 526/5-1, ZE 526/6-1, ZE 526/7-1. I am most grateful to Maria Laura Zenclussen for the helpful advice and great discussions.