Author’s address (for correspondence): LJ Oliveira, Av. Duque de Caxias Norte – 225, Pirassununga, SP 13635-900, Brazil. E-mail: firstname.lastname@example.org
There is a molecular crosstalk between the trophoblast and maternal immune cells of bovine endometrium. The uterine cells are able to secrete cytokine/chemokines to either induce a suppressive environment for establishment of the pregnancy or to recruit immune cells to the endometrium to fight infections. Despite morphological differences between women and cows, mechanisms for immune tolerance during pregnancy seem to be conserved. Mechanisms for uterine immunesuppression in the cow include: reduced expression of major histocompatability proteins by the trophoblast; recruitment of macrophages to the pregnant endometrium; and modulation of immune-related genes in response to the presence of the conceptus. Recently, an eGFP transgenic cloned embryo model developed by our group showed that there is modulation of foetal proteins expressed at the site of syncytium formation, suggesting that foetal cell can regulate not only by the secretion of specific factors such as interferon-tau, but also by regulating their own protein expression to avoid excessive maternal recognition by the local immune system. Furthermore, foetal DNA can be detected in the maternal circulation; this may reflect the occurrence of an invasion of trophoblast cells and/or their fragment beyond the uterine basement membrane in the cow. In fact, the newly description of exosome release by the trophoblast cell suggests that could be a new fashion of maternal-foetal communication at the placental barrier. Additionally, recent global transcriptome studies on bovine endometrium suggested that the immune system is aware, from an immunological point of view, of the presence of the foetus in the cow during early pregnancy.
Pregnancy is a unique immunological period for the mother. The maternal immune system needs to be adjusted to support the developing conceptus without threatening the mother’s life. Most of the current knowledge on the mechanisms of maternal tolerance to foetal/paternal antigens are known from human and mouse models. However, since the bovine genome was discovered, the use of the cow as an experimental model has grown exponentially. Humans and cattle share more than 21% more genes than mice in their genome (The Bovine Genome Sequencing and Analysis Consortium 2009); and despite their anatomical differences at the maternal-foetal interface, there are 44% more placental genes shared by human and cattle than humans and mice (Barreto et al. 2011).
Herein, we will discuss the current knowledge on bovine immunology of pregnancy in regard to changes on immune cells, and placental morphology, as well the expression of immune-related genes by the endometrium before and after implantation.
Modulation of the Immune System by Pregnancy in the Cow
In case of an infection, both the pregnant and non-pregnant bovine endometrium are able to recruit leucocytes and produce cytokines in response to the presence of microorganisms (Rosbottom et al. 2008, 2011; Herath et al. 2009).
Toll-like receptors (TLR) compose the major family of pattern recognition receptors (O’Neill et al. 2007; Koga and Mor 2010). During pregnancy, the expression of TLR2, -3, -4, -6 and -9 is greater in the interplacentomal endometrium than in the placentome (Petzl 2007). Endometrial epithelial cells stimulated in vitro with lipopolysaccharide (LPS) express acute phase proteins and secrete the proinflammatory eicasanoid prostaglandin E2 (Davies et al. 2008) and pro-inflammatory cytokines such as interluekin (IL) -1B, -6 and -8 (Cronin et al. 2012). Furthermore, proinflammatory cytokine expression is reduced by the knockdown of TLR4 or MYD88, which is a TLR4-signalling adaptor molecule (Cronin et al. 2012). Also, the endometrium expresses other proteins involved in innate immunity, such as β-defensins (Cormican et al. 2008) and mucin 1 (MUC1) (Davies et al. 2008). Likewise, TLR4 is essential for the endometrial innate immune response to LPS in the mouse, because tlr4−/− mice do not secrete proinflammatory cytokines (for example IL6 and prostaglandin E2) upon in vivo stimulation with LPS (Sheldon and Roberts 2010).
Actually, immune cells are present in the bovine endometrium, including CD4 (T helper cells), CD8 (T cytotoxic cells), CD5 (B-cells) and macrophages during the oestrus cycle and pregnancy (Cobb and Watson 1995; Oliveira and Hansen 2008, 2009). There are no data regarding the presence of antigen-presenting cells other than macrophages in the pregnant endometrium of the cow. In fact, the phenotypic characterization of bovine dendritic cells (bDC) was recently described by Miyazawa et al. (2006), when bDC were characterized by the expression of both CD172a (tyrosine-protein phosphatase non-receptor type substrate 1) and CD11c [Integrin, alpha X (complement component 3) receptor 4 subunit].
Likewise, little is known about trafficking of NK cells to the bovine endometrium as shown to occur in mouse and human (Croy et al. 2006; Moffett and Loke 2006), species in which NK cells are required for development of spiral arteries during decidua formation. Indeed, bovine NK cells were recently discovered (Storset et al. 2004; Boysen and Storset 2009), and they largely express natural cytotoxicity triggering receptor 1 (also called NKp46 or CD335). Furthermore, molecular studies showed that bovine NK also express CD16, but they are not positive for CD56 as seen in human. Bovine NK cells have large granular lymphocyte morphology and are able to proliferate in response to bovine IL (interleukin) -2 through the IL2-receptor (CD25) (Storset et al. 2004; Boysen and Storset 2009). The only evidence of NK cell activity in the ruminant uterus comes from an experiment by Tekin and Hansen (2004); ovine epithelial uterine lymphocytes had lymphokine-activated killer function upon interferon gamma (IFNG) and interferon-tau (IFNT) in vitro stimulation. The presence of NK cells has been detected in the bovine endometrium as early as day 16 of pregnancy (LJ Oliveira, N Mansouri-Attia, N Forde, AG Fahey, J Browne, JF Roche, O Sandra, P Reinaud, P Lonergan, T Fair, unpublished data), but their importance for pregnancy establishment and maintenance in the cow remains unclear.
The activation state and function of immune cells in the bovine endometrium is not completely understood. Immune cells that are present at the maternal-foetal interface may not only promote tolerance, but they may also be there as surveillance cells to prevent an occasional uterine infection to rise to the maternal system during pregnancy.
In addition, there are many proteins that are secreted by the luminal epithelium that can modulate the endometrial immune cells. In humans, chemokines and cytokines are constitutively expressed by the endometrium, and the level of their expression seems to be regulated by the hormone priming the uterus (Fahey et al. 2005a,b; Schaefer et al. 2005a,b).
Also in the cow, the endometrium expresses a wide range of immune factors, including colony-stimulating factor -2 and -3 [CSF2, CSF3; (de Moraes et al. 1999; Oshima et al. 2003)]; interleukin -1β [IL1β; (de Moraes et al. 1999)]; interleukin (IL) -6, -8 and -10 and tumour necrosis factor α [TNFA; (Galvao et al. 2011; Gabler et al. 2010)]. There are evidences that cytokines are expressed in a pregnancy-dependent manner by the bovine endometrium. A study performed by Almería et al. (2012) showed that IL10, IL12 and TNFA are constantly expressed across pregnancy, whereas IFNG and IL4 are increased expressed around day 120 of pregnancy. Perhaps the profile of cytokine expression by the endometrium influences the endometrial distribution of immune cells in the cow.
In the pregnant uterus, there is a massive accumulation of macrophages in the bovine endometrium (Oliveira and Hansen 2008, 2009). Macrophages (i.e. CD14+ CD68+ cells) are widely present in both interplacentomal and placentomal endometrium of the cow (Oliveira and Hansen 2009). In the interplacentomal region, macrophages are mainly present on the shallow stroma, but some scattered cells can be found in the glandular stroma. In the placentomal region, macrophages are located on stroma of maternal-villi tree (Oliveira and Hansen 2008). Moreover, macrophages that are closer to foetal tissues are strongly positive to major histocompatability complex (MHC) class II, in both interplacentomal and placentomal endometrium, whereas macrophages that are in the deeper areas of interplacentomal endometrium that are largely CD11b positive may be newly arrived cells from maternal circulation (Oliveira and Hansen 2009). Perhaps macrophages can be differentiated and activated as closer as they get to the foetal tissues.
Despite morphological differences between the types of placenta, approximately 90% of the genes found to be expressed by the human decidual macrophages (dMΦ) are also expressed by the bovine endometrial macrophage (bEMΦ) (Gustafsson et al. 2008; Oliveira et al. 2010). The similar gene expression between human and bovine macrophage at maternal-foetal interface probably is the result of trophoblast-derived molecules that were conserved throughout placental evolution.
Another possibility is that some of the CD14+ cells in the interplacentomal endometrium are not derived from a myeloid mononuclear phagocyte lineage but instead represent a differentiation of local cells because of pregnancy. In species where the endometrium undergoes a decidual response, decidualized stroma express cytokines and chemokines characteristically secreted by immune cells (Salamonsen et al. 2007; Chen et al. 2009; Riddell et al. 2012). In humans, placental fibrocyte-like cells possess macrophage characteristics, such as the similar expression of surface markers and capability of phagocytosis (Riddell et al. 2012). Perhaps similar differentiation pathways exist in species like cattle, so that CD14 becomes upregulated in non-myeloid cells.
In addition, scavenger receptor class A, member 5 (SCARA5) is reported to be expressed in the intercaruncular endometrium (Mansouri-Attia et al. 2009a), where there is a massive accumulation of macrophages. The function of SCARA5 is not clear, and however, recent studies showed a correlation of this gene to the expression of dickkopf homologue 1 (DKK1) during decidualization in humans (Duncan et al. 2011).
The reason why the cow experiences this macrophage accumulation in the pregnant endometrium remains an enigma, as does their biological role at the maternal interface. Two core patterns of macrophage activation are recognized – classical (M1) that induces inflammation and the alternative (M2) that are often produced during resolution of the immune response, wound healing and tissue remodelling (Mantovani et al. 2002; Gordon 2003).
In the pregnant uterus, tissue remodelling is essential for adjustment of the uterus to the growing foetus. Platelet-differentiation growth beta (PDGFB) is highly expressed in the EMΦ (Oliveira et al. 2010). The expression of PDGF suggests the role of macrophages on tissue remodelling because macrophage-derived PDGF stimulates proliferation and fibroblast differentiation (Barrientos et al. 2008). Likewise, the expression of granzymes is consistent to the induction of apoptosis by EMΦ (Oliveira et al. 2010). Abrahams et al. (2004) hypothesized that ingestion of apoptotic trophoblast cells causes dMΦ to develop an immunosuppressive phenotype (M2 activated).
Even though the bovine trophoblast is not capable of crossing the basement membrane of the endometrial luminal epithelium, perhaps EMΦ remove placental cells that stray past this barrier. Furthermore, apoptosis could be a signal for regulation of macrophage function at the maternal-foetal interface, which has markers of M2 activation. In addition, it is likely that EMΦ plays a role in immune regulation, tissue remodelling and induction of apoptosis in the cow.
In the cow, chorionic tissue generally downregulates expression of MHC class I proteins by the placentome throughout pregnancy (Low et al. 1990; Davies et al. 2000, 2004, 2006). In the interplacentomal region, the expression of MHC molecules by the trophoblast is suppressed until mid-gestation, when it starts to express MHC proteins until parturition. Additionally, binucleated trophoblastic cells (BNC), which are the main migratory cells, express MHC class I molecules (Bainbridge et al. 2001). Perhaps expression of MHC class I proteins by BNC sensitizes the maternal immune system to induce a tolerogenic environment in the uterus of the cow. Moreover, the persistence of BNC cells at the time of parturition has been correlated to retention of foetal membranes in sheep (Majeed et al. 1995), possibly due to the lack of inflammatory signals necessary for placental detachment after parturition. Furthermore, the majority of MHC class I molecules expressed by the bovine trophoblast are non-classical (NC) genes (Davies et al. 2006). Probably NC genes expression avoids the NK cell attack towards to the trophoblast cell in a similar manner of human and mice (Bulmer and Johnson 1985).
The mouse preimplantation embryo development (PED) orthologue gene was found to be expressed by the bovine embryo at blastocyst stage (Fair et al. 2004). Additionally, the treatment of in vitro produced embryos with IL4, IL3, IFNG and progesterone increased the expression of NC1 (O’Gorman et al. 2010). Likewise, the expression of NC2 and NC3 genes was modulated by cytokine stimulation (Al Naib et al. 2011). Because the NC4 gene is the gene most modulated by cytokine treatment, its expression was increased by the IL3, IL4 and leukaemia inhibitor factor (LIF) (Al Naib et al. 2011). However, cytokine treatment did not improve embryo development in vitro. It suggests that NC genes, especially NC4, would be the most responsive gene to immune changes, and it would be involved in the maternal immune system regulation at early stages of embryo development.
Additionally, in pregnancies established with bovine somatic cell nuclear transfer (SCNT) embryo, there is a premature expression of MHC class I proteins by the trophoblast (Hill et al. 2002). Therefore, the high rates of pregnancy loss of SCNT embryos may be a result of a deregulation maternal immune system by abnormal expression of MHC class I by the trophoblast.
Systemic Effects of Pregnancy in the Cow
Pregnancy has also an impact on the levels of circulating immune cells. In the cow at days 33–34 of pregnancy, there are no changes in the numbers of CD4, CD8 or γδ-T cells compared to non-pregnant animals (Oliveira and Hansen 2008). There is only an increase in the percentage of CD4 cells expressing CD25, which suggests the increase of putative T-regulatory cells in the pregnant cow as reported in other species (Aluvihare et al. 2004; Somerset et al. 2004; Yang et al. 2008).
In both women and mice, a subpopulation of CD4+ CD25+ that also expresses FOXP3 increases in peripheral blood mononuclear cells (PBMC) during pregnancy (Aluvihare et al. 2004; Somerset et al. 2004). These cells have been identified as T-regulatory (Treg) cells that can secrete cytokines such as IL4 that inhibit activation of cytotoxic T cells against alloantigens (Mjosberg et al. 2007). The proportion of CD4+ CD25+ cells in PBMC is reduced in women with unexplained recurrent spontaneous abortion (Yang et al. 2008). Moreover, in mice, depletion of CD25+ cells decreases the ability the female to sustain pregnancy (Aluvihare et al. 2004; Shima et al. 2010).
It is not known whether Treg cells play a role on bovine pregnancy. The occurrence of factor forkhead box P3 (FOXP3)-positive Treg cells in bovine was only described later by Gerner et al. (2010). Likewise, recent study on regulatory properties of bovine lymphocyte subsets showed that CD4+ CD25high Foxp3+ and CD4+ CD25low T cells may not function as Treg in vitro; the regulatory function seems to reside in the γδ-T-cell population (Hoek et al. 2009). Coincidentally, there is an increase in the number of circulating γδ-T cells closer to parturition in the cow (Oliveira and Hansen 2008); perhaps γδ-T cells recirculate from the pregnant uterus to the periphery in the cow closer to parturition. Thus, further studies should be focused on pregnancy-related changes on Treg cells to clarify the importance of these cells in the pregnant cow.
Immunosuppression during peripartum is defined by the decline in numbers of CD4+, CD8+ and γδ-T cells in PBMC (Van Kampen and Mallard 1997; Kimura et al. 1999, 2002) and reduced proliferation and IFNG secretion by mitogen-stimulated lymphocytes (Detilleux et al. 1995; Nonnecke et al. 2003). However, these findings have not always been replicated. It was reported there is a tendency for an increase in the proportion of CD4, CD25 and γδ-T cells as parturition approached (Karcher et al. 2008; Oliveira and Hansen 2008), suggesting that the immunosuppressive state of the periparturient cow may not be due to low number of circulating lymphocytes, but by recirculation of specific subsets of immunomodulatory lymphocytes such as Treg and γδ-T cells.
Release of Foetal Membranes Could be an Immune-Mediated Process
Evidence for immunological participation in the process of parturition is indicated by the increased incidence of retained placenta (RP) in cows. Joosten et al. (1991) demonstrated an increased incidence of RP in cows which share major histocompatability class I antigens with their conceptus. Recently, the two-way compatibility between calf-dam has shown to increase the odds of occurrence of RP in a case–control study. The detachment of foetal membranes may be an immune-mediated process (Benedictus et al. 2012), supporting the idea of the involvement of the immune system on release of foetal membranes and parturition. Moreover, RP has been correlated with decreased activity of macrophages in the caruncular area (Miyoshi et al. 2002) and to the reduced chemotaxis and myeloperoxidase activities on circulating neutrophils (Kimura et al. 2002).
Foetal Factors Affecting the Immune System
It is not clear which are the effects of pregnancy on the systemic maternal immune system in the cow. The expression of interferon-tau (IFNT) by the embryo at the window of implantation modulates expression of interferon-stimulated genes (ISGs) such as ISG15 and myxovirus resistance 1 (MX1) in PBMC of pregnant cow (Han et al. 2006; Forde et al. 2011). Nonetheless, most regulation of immune responses during pregnancy probably occurs at the maternal-foetal interface by locally acting factors.
One of the hormones most secreted by trophoblast cells is progesterone (P4) that is essential for maintenance of pregnancy in mammals. In cattle, there are two main sources of P4: the corpus luteum and the trophoblastic giant cells of the placenta (Senger 2003; Vanselow et al. 2008). Progesterone can directly inhibit lymphocyte proliferation in vitro, however at very high dose that does not seem to be the immunomodulatory mechanism of P4 (Low and Hansen 1988; Monterroso and Hansen 1993). In addition to its direct effects on lymphocytes, progesterone can also induce synthesis of endometrial molecules such as the uterine serpin or SERPINA14 (Leslie and Hansen 1991). SERPINA14 is secreted by the endometrial glands and, in sheep, showed antiproliferative properties on lymphocytes and cancer cells (Padua and Hansen 2008).
Another important molecule is the IFNT, which is an embryo-derived cytokine in the type I interferon family secreted by the trophoblast from days 15 to 25 of pregnancy and is essential for corpus luteum survival (Spencer et al. (2004). In addition, IFNT inhibits lymphocyte proliferation in vitro [cattle – (Skopets et al. 1992); sheep – (Tekin et al. 2000)] which could contribute to development of maternal tolerance towards the conceptus in early pregnancy. Nonetheless, IFNT does not affect the endometrial pool of immune cell of pregnant vs non-pregnant cows (Leung et al. 2000).
Analysis of the global transcriptome of bovine endometrium during the pre-attachment period demonstrated upregulation of many interferon-stimulated genes. These genes may be involved in molecular crosstalk between embryonic and maternal immune systems, including complement C1s subcomponent (C1S) and SERPINA14 (Bauersachs et al. 2008), pentraxin 3 (PTX3), 2′-5′-oligoadenylate synthetase 1(OAS1), radical S-adenosyl methionine domain containing 2 [RDSA2; (Mansouri-Attia et al. 2009a; Forde et al. 2011)] and interferon signalling genes such as interferon regulatory factor 9 (IRF9) and interferon-induced transmembrane protein 3 (IFITM3) (Groebner et al. 2011).
There is a synergistic relationship between prostaglandin E2 (PGE2) and interferon-tau (IFNT). Treatment of bovine PBL with PGE2 downregulates IL2 and colony-stimulating factor 2 (CSF2) mRNA levels (Emond et al. 1998), whereas treatment of PBL primed with IFNT with PGE2 increases the mRNA expression of CSF2 (Emond et al. 2000). Supplementation of CSF2 in the culture of preimplantation bovine embryo has been shown to increase post-transfer embryo survival (Loureiro et al. 2009), probably due to increased expression of IFNT (Loureiro et al. 2011).
Anatomy of the Bovine Placenta and the Occurrence of Foetal Chimerism
Moffett and Loke (2006) suggested that the epitheliocorial placenta would prevent the maternal immune system from recognizing foetal alloantigens. In the cow, in addition to a synepitheliochorial placenta type, there is evidence for the occurrence of microchimerism, because foetal DNA can leak through the placental barrier during early pregnancy (Xi et al. 2006; Turin et al. 2007a,b). The invasion of trophoblastic giant cells (TGC) to the maternal side has been long described; however, these cells do not cross the basement membrane in the cow (Wooding 1980; Dantzer 1999). Therefore, mechanisms underlying the leakage of foetal DNA to the maternal blood in the cow are still unknown.
The release of exosomes by the human trophoblast has been targeted as an alternative fashion of communication at the placenta barrier (Luo et al. 2009; Mincheva-Nilsson and Baranov 2010). Until now, the only evidence of exosome release in epitheliochorial placenta was given by the study of exosomes from the supernatant culture medium of immortalized ovine glandular epithelium in vitro (Racicot et al. 2009).
Our group, using a model of eGFP (enhanced green fluorescent protein) transgenic cloned embryo pregnancy, detected the presence of eGFP DNA in maternal blood cells at days 60 and 90 (L.J. Oliveira, N. Mansouri-Attia, N. Forde, A.G. Fahey, J. Browne, J.F. Roche, O. Sandra, P. Reinaud, P. Lonergan, T. Fair, unpublished data). The immunohistochemical analysis showed weak staining for eGFP at the site of syncytial formation (Fig. 1). It suggests that the trophoblast may decrease its gene expression to minimize the recognition of foetal antigens by the maternal immune system (Fig. 1). It is hypothesized that after syncytial formation, maternal macrophage uptake foetal material and recirculate to maternal system to start an immune response and induce tolerance to foetal antigens. Moreover, in humans, trophoblast-derived exosomes are able to recruit monocytes to the target tissue (Atay et al. 2011).
Expression of Immune-Related Genes by the Uterus in the Cow
Global transcriptome studies have demonstrated an enrichment of immune-related genes in the endometrium of pregnant cow compared to their non-pregnant counterparts (Mansouri-Attia et al. 2009a). Beyond that, the endometrium acts as a sensor of the quality of the embryo. The pregnant endometrium supporting SCNT embryos showed that among the differentially regulated genes (DEG), only 3% were immune-related genes, whereas in pregnancies derived from artificial insemination were 18% and those from vitro fertilization had 16% immune-related genes of their total of DEG (Mansouri-Attia et al. 2009a). It suggests that deregulation of the maternal immune system in SCNT pregnancies would contribute to high rates of embryonic loss during early pregnancy. Moreover, an enrichment of IFN-related genes, such as myxovirus resistance 2 (MX2), RASD2 and OAS1, were found among the immune-related DEG (Bauersachs et al. 2008; Mansouri-Attia et al. 2009a).
We compiled microarray data in gene expression of bovine endometrium into a list of genes expressed during the pre-implantation period (days 5 and 7 of pregnancy) (Forde et al. 2011); at the window of implantation (from day 13 to day 20) (Bauersachs et al. 2008; Mansouri-Attia et al. 2009a,b; Salilew-Wondim et al. 2010; Groebner et al. 2011); and in the period after implantation (from day 35 to term) (Band et al. 2002; Ishiwata et al. 2003; Larson et al. 2006; Hashizume 2007; Hashizume et al. 2007; Kumar et al. 2007, 2010). The compiled data were used for bioinformatics data analysis. Using the bioinformatics tool pathway analysis Panther (http://www.pantherdb.org), we evaluated the immune-related pathways involved in the establishment of bovine pregnancy. One of the most differentially regulated pathways is the interleukin signalling pathway (Fig. 2). At the pre-implantation period, second messenger proteins, such as RAC-alpha serine/threonine-protein kinase (AKT1) and MAP kinase-activated protein kinase 2 (MAPKAPK2), are upregulated, suggesting that the endometrium becomes more responsive to embryo-derived factors before implantation. Perhaps, adequate oestrogen and progesterone priming the uterus would be a crucial step for pregnancy establishment (Salilew-Wondim et al. 2010; Forde et al. 2011). At the time of implantation and post-implantation, there is an upregulation of interferon receptors, which is clearly correlated to the ability of the endometrium to respond to IFNT secretion by the conceptus (Fig. 2). It suggests an orchestrated immune-gene expression is needed for successful pregnancy that starts before implantation, and this communication between mother and foetus last the whole pregnancy.
In this review, we revisited the current literature regarding mechanisms and effects of pregnancy on the maternal immune system in the cow. From this, we can gain the insight that the bovine placenta is a more invasive type of placenta than previously believed, and the maintenance of all tissue layers between mother and foetus perhaps does not confer a better barrier for maternal immune system to foetal antigens. The mechanisms for immune regulation during pregnancy seem to be conserved across placental evolution. Our eGFP cloned embryo pregnancy model showed the occurrence of foetal microchimerism. The role and/or consequences of foetal microchimerism are still under investigation. Perhaps, changes in the expression of immune-related genes on the pregnant endometrium occur as in response to the foetal products that can reach maternal tissues earlier in pregnancy. Therefore, this immune-molecular crosstalk between mother and foetus may be crucial for the fate of the pregnancy.
The authors thank Dr Trudee Fair, University College Dublin for the valuable discussion during the preparation of the manuscript.
Conflicts of interest
None of the authors have any conflicts of interest to declare.