Non-classical MHC class I

Semi-allogeneic cells of embryonic origin (trophoblast) invade the maternal decidua to open up maternal uterine arteries and facilitate blood supply to the foetus. The interaction with decidual natural killer (NK) cells has a key role in this. One of the roles of NK cells is to destroy human leucocyte antigen (HLA) null cells and the expression of non-classical HLA by trophoblast is postulated to be a mechanism of evading NK cell killing. The non-classical MHC class I molecules HLA-E, -F and -G are less polymorphic than the classical HLA class I molecules and have different expression patterns: all three (as well as HLA-C) are expressed on the extravillous trophoblast that invades the maternal decidua, whereas the villous cytotrophoblast and syncytiotrophoblast found within the placenta express surface HLA-E and secrete a soluble form of HLA-G.

HLA-G is the most extensively studied non-classical MHC class I molecule and the role of HLA-G in modulating the maternal immune response to semi-allogeneic cells of fetal origin is well established⁽Reference Hunt, Langat and McIntire^6, Reference Hunt, Petroff and McIntire⁷⁾. Soluble and membrane forms of HLA-G are derived by the alternative splicing of HLA-G transcripts. HLA-G interacts with inhibitory receptors on leucocytes and mediates numerous effects: it modulates NK cell activity⁽Reference Hunt, Langat and McIntire^6, Reference Hunt, Petroff and McIntire⁷⁾; stimulates high levels of anti-inflammatory cytokines, such as transforming growth factor (TGF)-β1, from mononuclear phagocytes⁽Reference Le Friec, LaupΦze and Fardel⁸⁾; induces Fas/Fas ligand-mediated apoptosis of CD8+T and CD8+NK cells and inhibits antigen-specific cytotoxic activity of CD8+T-cells⁽Reference Contini, Ghio and Poggi⁹⁾. Since the immunomodulatory role of HLA-G in pregnancy was identified, there has been much interest in its potential as either a therapeutic or diagnostic tool in transplant immunology, autoimmunity and cancer⁽Reference Carosella, Favier and Rouas-Freiss¹⁰⁾.

The study of HLA-E and HLA-F was hampered initially by the observation that the expression of these appeared to be restricted to inside the cell. Surface expression of HLA-E has now been found to be dependent on the presence of peptides from other HLA class I molecules. This complex of HLA-E with HLA class I molecule-derived peptides, especially from HLA-G and HLA-C, interacts with either activating or inhibiting CD94/NKG2 on NK cells to modulate NK cell activity. There is little knowledge about HLA-F and the peptides that it binds have not been identified⁽Reference Ishitani, Sageshima and Hatake¹¹⁾.

Natural killer and natural killer T cells

NK cells have a role in the control of implantation and transformation of the uterine vasculature by trophoblasts, on which blood supply to the foetoplacental unit depends⁽Reference King and Loke¹²⁾. There is an accumulation of NK cells in the maternal decidua in the first and second trimester with a reduction in the third trimester⁽Reference Bulmer, Morrison and Longfellow^13, Reference Williams, Searle and Robson¹⁴⁾. About 30% of cells in first trimester decidual tissue are leucocytes and up to 70% of these are CD56^brightCD16−NK cells. This contrasts with the peripheral blood where CD56^dimCD16+NK cells dominate⁽Reference Bulmer, Morrison and Longfellow¹³⁾. The CD56^bright population of NK cells accumulates at the implantation site and when activated these cells secrete an array of cytokines including macrophage inflammatory protein 1α, granulocyte/macrophage-colony stimulating factor, TNFα, interferon-γ, TGF-β and leukaemia inhibitory factor, with minimal impact on cytotoxicity. CD56^bright NK cells express high levels of CD94 and their accumulation at the implantation site has been postulated to be driven by HLA-E/G expression on extravillous cytotrophoblast that acts as a homing receptor. Alternatively, a cascade of cytokine production at the implantation site has been postulated to cause NK cell differentiation.

Interest in natural killer T (NKT) cells at the materno-fetal interface stems from their immunoregulatory properties. As their name suggests, NKT cells have the features of NK and T-cells: like T-cells they express a T-cell receptor, but this is invariant⁽Reference Godfrey, MacDonald and Kronenberg¹⁵⁾; unlike T-cells they are not restricted by MHC class I or class II molecules, but recognise glycolipids in the context of the MHC-like molecule CD1d. Trophoblasts within first trimester decidua express CD1d and 0·48% of CD3+T-cells in the decidua have an NKT phenotype; this represents a 10-fold higher concentration compared to blood. Decidual CD4+NKT cells have a T helper 1 (Th1) bias and produce granulocyte/macrophage-colony stimulating factor, whereas those in the peripheral blood are T helper 2 (Th2)-biased⁽Reference Boyson, Rybalov and Koopman¹⁶⁾.

T helper cytokine profiles

Since the observation that Th2 cytokines are normally produced at the materno-fetal interface, the hypothesis that established pregnancy (i.e. excluding implantation and labour) is associated with a Th2-biased immune response has dominated the reproductive immunology literature⁽Reference Lin, Mosmann and Guilbert^17–Reference Piccinni, Scaletti and Maggi¹⁹⁾. This is despite the lack of any really convincing data from human subjects and the observation that mice lacking expression of the main Th2 cytokines IL-4, IL-5, IL-9 and IL-13 (i.e. quadruple knockouts) do not have fertility problems (no differences in litter sizes and all offspring healthy) indicating that these cytokines are not essential for fetal survival even during allogeneic pregnancy⁽Reference Fallon, Jolin and Smith²⁰⁾. It must be noted that progesterone-induced blocking factor induces a Th2 cytokine profile⁽Reference Szekeres-Bartho and Wegmann^21, Reference Szereday, Varga and Szekeres-Bartho²²⁾. Progesterone-induced blocking factor is a feature of normal human pregnancy and can be measured in urine from pregnant women; the lack of progesterone-induced blocking factor predicts spontaneous pregnancy termination and neutralising endogenous progesterone-induced blocking factor leads to pregnancy termination in mice⁽Reference Polgar, Kispal and Lachmann^23, Reference Polgar, Nagy and Miko²⁴⁾.

At the outset of the study of Th1/Th2 balance in pregnancy the focus was T-cells themselves but the materno-fetal interface itself is relatively devoid of T-cells. Thus, the paradigm has evolved to encompass type 1 and type 2 cytokines irrespective of their cell source which can include NK cells⁽Reference Borzychowski, Croy and Chan^25–Reference Croy, Chantakru and Esadeg²⁷⁾ and the gestation-associated tissues themselves⁽Reference Jones, Finlay-Jones and Hart²⁸⁾. An appreciation of the temporal and spatial variation in the dominance of either type 1 or type 2 cytokines has also emerged. For example, NK cells and interferon-γ are pivotal to remodelling the uterine vasculature during implantation⁽Reference Ashkar, Di Santo and Croy^26, Reference Croy, Chantakru and Esadeg²⁷⁾. Conversely, in a study using surface markers of Th1 (IL-18 receptor) and Th2 (ST2L) in maternal peripheral blood, NK cells rather than CD4+ and CD8+T cells were found to have a bias to a type 2 phenotype, a bias lost with preeclampsia⁽Reference Borzychowski, Croy and Chan²⁵⁾.

Overall, there is now an appreciation of the need for balanced type 1/type 2 responses depending on the tissue site and the stage of pregnancy. More recently the role of the innate immune response and inflammation has become of interest. Given that IL-4, IL-10 and IL-13 are anti-inflammatory cytokines⁽Reference Hart, Bonder and Balogh²⁹⁾, they might serve to prevent an overactive inflammatory response; the type 1/type 2 paradigm should be re-evaluated from this perspective.

Innate immunity

Innate immunity/inflammation is now considered to have a physiological role in implantation, the maintenance of pregnancy and parturition⁽Reference Ashkar, Di Santo and Croy^26, Reference Croy, Chantakru and Esadeg^27, Reference Patni, Flynn and Wynen³⁰⁾. The innate immune response also contributes to adverse outcomes in fertility and pregnancy, including tubal infertility associated with Chlamydia trachomatis, infection-associated preterm delivery and preeclampsia⁽Reference Germain, Sacks and Soorana^31, Reference Horne, Stock and King³²⁾. There is extensive evidence of a heightened systemic inflammatory response in pregnant women⁽Reference Sacks, Sargent and Redman³³⁾, for example, increased leucocyte count and increased levels of C-reactive protein and other acute phase proteins⁽Reference Miller³⁴⁾. This might be balanced by increased circulating levels of anti-inflammatory molecules such as TGF-β1⁽Reference Power, Popplewell and Holloway³⁵⁾ and IL-4⁽Reference Zhao, Lei and Liu³⁶⁾.

The enhanced systemic inflammatory response during pregnancy has been postulated to be mediated by particulate debris or soluble products derived from the placenta, which enter the maternal circulation. Cell-derived microparticles occur commonly in the circulation, but the levels of these are increased with normal pregnancy being derived from numerous cell types, especially the syncytiotrophoblast. These placenta-derived microparticles are postulated to drive normal systemic inflammation in pregnant women and to be further increased in preeclampsia⁽Reference Sargent, Borzychowski and Redman^2, Reference Redman and Sargent³⁷⁾. These microparticles might also have a role in down-modulated T-cell function during pregnancy. Syncytiotrophoblast microparticles can be found attached to circulating monocytes and unbound in the circulation from the second and increasing into the third trimester. When isolated in vitro from the placenta, they can induce an inflammatory response in non-pregnant women⁽Reference Germain, Sacks and Soorana³¹⁾. From animal studies, it has been shown that the presentation of alloantigens coincides with the detectable release of placental debris from around mid-gestation⁽Reference Erlebacher, Vencato and Price³⁸⁾.

There is much interest in identifying molecules and signalling pathways that might link the so-called type 2 bias of pregnancy with the innate immune function. Recently, it was shown that the immunoregulatory molecule T-cell Ig and mucin domain (Tim)-3 which has been suggested to modulate the Th1/Th2 balance and have a role in the innate immune response is up-regulated on monocytes but not T- and B-cells in the peripheral blood of pregnant women⁽Reference Zhao, Lei and Liu³⁶⁾. Tim-3 was found to enhance innate immunity and Th1-associated anti-pathogen responses in pregnancy: blocking Tim-3 impaired phagocytosis, signalling via lipopolysaccharide, T-cell proliferation and interferon-γ∴ production. The IL-4 signalling pathway has a role in the up-regulation of Tim-3 expression on monocytes⁽Reference Zhao, Lei and Liu³⁶⁾. Abnormalities in Tim-3 expression were also observed in women suffering miscarriages at 12–20 weeks of pregnancy.

Regulatory T-cells

Recent years have seen resurged interest in T-cell populations with immunoregulatory or suppressive properties. Initially identified by their ability to suppress autoimmune disease, regulatory T-cells also have been shown to modify a broad range of immunological activities associated with inflammatory and infectious diseases, cancer and transplantation⁽Reference Read and Powrie³⁹⁾. Abnormalities in the activity of these cells have been implicated in susceptibility to many conditions with underlying immune aetiology. Several types of regulatory T-cells have been described and these can be subdivided broadly into natural and inducible regulatory T-cells. To date, only natural CD4+ regulatory T-cells that can be identified by the expression of CD25 and low expression of CD127 (IL-7 receptor α chain)⁽Reference Seddiki, Santner-Nanan and Martinson⁴⁰⁾ have been studied in pregnancy. CD4+/CD25+/CD127^lo regulatory T-cells express FoxP3 which has been described as the master regulator of development and function of this population⁽Reference Fontenot and Rudensky⁴¹⁾, and their functional properties have been attributed variously to CTLA-4 (cytotoxic T lymphocyte antigen-4), TGF-β and IL-10, but this remains an area of ongoing investigation and controversy.

In mice and human subjects, CD4+CD25+T-cell numbers increase in early pregnancy and decline from mid-gestation to reach levels equivalent to the non-pregnant state by term⁽Reference Somerset, Zheng and Kilby^42–Reference Heikkinen, Mottonen and Alanen⁴⁴⁾. As CD25 is also a marker of T-cell activation, it is imperative that other phenotypic or functional data be available to clarify that the CD4+CD25+T-cell population being studied is regulatory. CD4+CD25+T-cells able to suppress autoimmune and allogeneic responses are expanded in the spleen, blood and various lymph nodes of pregnant mice. The absence of these cells leads to immunological rejection of the foetus in allogeneic but not syngeneic pregnancies⁽Reference Aluvihare, Kallikourdis and Betz⁴³⁾. In human subjects, an increase in CD4+CD25+T-cells in peripheral blood in the first and second trimester was described, but without additional phenotypic/functional data it was unclear whether this represented increased numbers of regulatory or activated T-cells⁽Reference Heikkinen, Mottonen and Alanen⁴⁴⁾. In a recent study, where a more definitive phenotyping strategy was used, it was found that CD4+CD25+CD127^loFoxP3+T-cells were reduced in the second trimester and that expression of FoxP3 itself was lower with pregnancy. The function of this population in pregnant women was comparable to that in non-pregnant women⁽Reference Mjosberg, Svensson and Johansson⁴⁵⁾.

CD4+CD25+T-cells are relatively abundant in first trimester decidua (about 20% of decidual CD4+T-cells v. about 8% of CD4+T-cells in peripheral blood of pregnant women) and they respond poorly to T-cell stimulation and have suppressive activity that is cell-contact dependent⁽Reference Sasaki, Sakai and Miyazaki⁴⁶⁾. The numbers of CD4+CD25+T-cells in both decidua and peripheral blood in early pregnancy are decreased in spontaneous abortion⁽Reference Sasaki, Sakai and Miyazaki⁴⁶⁾. As in peripheral blood, there is a decrease in the relative abundance of these cells by term to about 14%. As many of these express CTLA-4, OX40 and glucocorticoid-induced TNF receptor-related protein, they are likely to have regulatory function⁽Reference Heikkinen, Mottonen and Alanen⁴⁴⁾.

The expansion of CD4+CD25+T-regulatory cells occurs very early in pregnancy and declines in both the decidua and blood by term. Therefore, it has been suggested that the expansion in these cells is not driven by exposure to fetal alloantigen. This has led to interest in the impact of pregnancy-related hormones on these cells and both estradiol and progesterone have been implicated⁽Reference Mjosberg, Svensson and Johansson^45, Reference Tai, Wang and Jin⁴⁷⁾. However, in another animal study, CD4+CD25+T-regulatory cell numbers were greater in allogeneic v. syngeneic pregnancies with the number of male offspring in the litter having a positive impact on numbers of these cells suggesting a role for fetal alloantigens⁽Reference Lin, Liang and Chen⁴⁸⁾.

Indoleamine 2,3,-dioxygenase

Indoleamine 2,3,-dioxygenase (IDO) is the first and rate-limiting enzyme in tryptophan degradation. IDO degrades the indole moiety of tryptophan, serotonin and melatonin, and initiates the production of neuroactive and immunoregulatory metabolites known as kynurenines. This leads to local depletion of tryptophan and increased kynurenines both of which affect T-cell proliferation and survival: tryptophan starvation inhibits T-cell activation, and the products of tryptophan catabolism regulate T-cell proliferation and survival⁽Reference Curti, Trabanelli and Salvestrini^49–Reference von Bubnoff, Hanau and Wenzel⁵¹⁾.

IDO has been implicated in the induction of immune tolerance during many immunologically mediated events, including pregnancy. The treatment of pregnant mice with a pharmacologic inhibitor of IDO-induced maternal T-cell-mediated rejection of allogeneic but not syngeneic foetuses⁽Reference Munn, Zhou and Attwood⁵²⁾. Inhibition of IDO is also associated with T-cell-dependent, antibody-independent complement deposition and haemorrhagic necrosis at the materno-fetal interface in allogeneic but not syngeneic pregnancies⁽Reference Mellor, Sivakumar and Chandler⁵³⁾. The timing of fetal loss coincided with the up-regulated expression of IDO⁽Reference Munn, Zhou and Attwood⁵²⁾. IDO can be detected in day 6 human blastocysts and then throughout pregnancy in syncytiotrophoblast, extravillous cytotrophoblast, placental macrophages and the fetal membranes. Down-regulated phytohaemagglutinin-stimulated proliferation of peripheral blood mononuclear cells by conditioned medium from first trimester placental and decidual cultures is IDO-dependent⁽Reference Kudo, Boyd and Spyropoulou⁵⁴⁾. Much of the interest in IDO and pregnancy has focused on the capacity of placenta-/decidua-derived IDO to suppress maternal T-cell reactivity directed at paternal alloantigens. An alternative hypothesis has been put forward, focusing on the potential effects of the products of tryptophan degradation (i.e. kynurenines) that include the removal of oxygen radicals, opening of blood vessels and increased production of NAD⁽Reference Bonney and Matzinger⁵⁵⁾.

There is a link between IDO and regulatory T-cells⁽Reference Curti, Trabanelli and Salvestrini⁴⁹⁾. CD4+CD25+T-cells induce tryptophan catabolism by dendritic cells and a tolerogenic phenotype in a CTLA-4-dependent manner⁽Reference Fallarino, Grohmann and Hwang⁵⁶⁾. Similarly, expression of IDO by dendritic cells and macrophages in maternal decidua is up-regulated via CTLA-4, and this has been suggested to indicate that CTLA-4 expressing CD4+CD25+ regulatory T-cells that infiltrate the decidua in early pregnancy induce IDO expression by decidual dendritic cells/macrophages expressing the appropriate counterligand (B7 family members CD80 and CD86)⁽Reference Miwa, Hayakawa and Miyazaki⁵⁷⁾.

Complement

Activation of complement is a strategy for destruction and elimination of pathogens, but can also damage host cells and tissues. The maternal circulation and the placenta contribute to a fully active complement system at the materno-fetal interface. The regulation of complement during pregnancy is critical in preventing maternal-mediated damage of the placenta and foetus. Deficiency in a negative regulator of complement activation (Crry) resulted in embryonic lethality, due to a failure to control complement. This was associated with complement deposition and inflammation, including an extensive polymorphonuclear granulocyte infiltrate in the foetoplacental unit⁽Reference Xu, Mao and Holers⁵⁸⁾. There is no direct human counterpart of Crry, but other regulators of complement activation such as decay accelerating factor (CD55), CD59 and membrane co-factor protein (CD46), are expressed extensively in the human placenta. Syncytiotrophoblasts that form the outermost layer of the chorionic villous tree of the placenta and are in direct contact with maternal blood express all three regulatory proteins and these are known to be functional⁽Reference Girardi, Bulla and Salmon⁵⁹⁾. Currently, there is particular interest in additional roles for CD46 in reproduction because of its expression on the inner acrosomal membrane of spermatozoa and the observation that CD3/CD46 cross-linking on T-cells induces an IL-10 producing regulatory T-cell population⁽Reference Liszewski, Kemper and Price^60, Reference Kemper, Chan and Green⁶¹⁾.

Fetal origins of adult disease

Since Barker first hypothesised on the fetal origins of adult disease there has been a wealth of scientific literature published that relates to this. The essence of this hypothesis is that during fetal development, the organs and systems of the body go through periods when they are plastic and sensitive to the environment. This leads to physiological or morphological changes associated with health and disease in adulthood⁽Reference Barker⁶²⁾. The original model was proposed for the group of diseases including CHD, stroke, high-blood pressure and type-2 diabetes. The general concept has now been extended to encompass the early life origins of other diseases, such as allergy, and now includes the development of disease in childhood.

Antenatal exposure to dietary components, cigarette smoke, microbial products and many others have been implicated to have a role in the development of the fetal immune system with knock-on effects for health in childhood and beyond. How these exposures impact on maternal systemic immune function during pregnancy, immune function at the materno-fetal interface and fetal immune development are relatively understudied. Recent years have seen an interest in the effect of these exposures on epigenetic modification of chromatin structure enabling heritable gene regulation without changing the nucleotide sequence. This occurs particularly via the modification of histone tails that alter the interaction between histones and DNA. These modifications include acetylation, methylation, phosphorylation, poly-ADP ribosylation and ubiquitination⁽Reference Burdge, Hanson and Slater-Jefferies⁶³⁾. For example, chronic maternal consumption of a high-fat diet in primates is associated with increased acetylation of candidate genes and altered gene-specific expression in fetal hepatic tissue⁽Reference Aagaard-Tillery, Grove and Bishop⁶⁴⁾.

Nutritional approaches to modulating immune function in pregnancy and the peri-natal period

Nutritional supplementation during pregnancy offers the most acceptable way of potentially modulating the immune function of the mother and her offspring and thereby preventing diseases with origins in fetal life. Studies of maternal supplementation with vitamin D, fish oil or probiotics dominate this field.

Summary

A number of immunomodulatory pathways contribute to the success of pregnancy with attention currently focused on regulatory T-cells, IDO, NK cells and the innate immune system. The impact of antenatal variation in the immune function on the development of the fetal immune system and thereby health and disease in childhood and beyond is of increasing interest. Changes in the maternal diet over recent decades coincide with the increasing prevalence of allergic and other immune-mediated diseases⁽Reference Shaheen^95, Reference Devereux⁹⁶⁾. The modification of maternal diet has therefore emerged as a strategy for disease prevention.