Published online before print September 17, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* University Medical Center Utrecht, Departments of Psychoneuroimmunology and
Reproductive Medicine and Gynecology, Utrecht, The Netherlands
1 Correspondence: Department of Psychoneuroimmunology, University Medical Center Utrecht, Room KC 03.068.0, Lundlaan 6, 3584 EA Utrecht, The Netherlands. E-mail: c.heijnen{at}umcutrecht.nl
 TOP ABSTRACT INTRODUCTIONPOLARITY WITHIN THE BLASTOCYST...BLASTOCYST-ENDOMETRIAL...CONCLUSIONREFERENCES |
|---|
Key Words: regulatory T cells • tolerance • uterine NK cell • chemokine
|
TOPABSTRACTINTRODUCTIONPOLARITY WITHIN THE BLASTOCYST...BLASTOCYST-ENDOMETRIAL...CONCLUSIONREFERENCES |
|---|
On the endometrial side, preparation for implantation is necessary and occurs among others through the establishment of the putative implantation window. In this time period, the uterus is prepared to receive a blastocyst and support further implantation through mediation by immune cells, cytokines, growth factors, chemokines, and adhesion molecules [1 2 3 ]. A functional endometrium is required for successful implantation. In the estrogen-mediated, proliferative phase, preparations are made. In the secretory phase, the window of implantation is established through the action of progesterone on estrogen-primed endometrium [4 ].
From the perspective of the future embryo, preparation consists of the expression of numerous receptors and adhesion molecules on the outside of the preimplantation blastocyst and the production of cytokines and other mediators.
The process of implantation encompasses several distinct stages: apposition, adhesion, penetration, and trophoblast invasion. These steps can only take place during the window of implantation [4 ].
In the process of apposition, the blastocyst interacts with the endometrium using adhesion molecules. In the adhesion phase, the polarized interaction between blastocyst and endometrium is established and becomes stronger, a process mediated by adhesion molecules, immune cells, and cytokines [5 ]. During trophoblast invasion, maternal tissue is degraded in a controlled manner to allow the blastocyst to enter the endometrial lining.
A critical step in the establishment of pregnancy is decidualization, which is a process in which the endometrium undergoes extensive changes in morphology and expression and secretion patterns to support the implanting blastocyst [6 ]. Actual implantation of the blastocyst into the endometrium occurs 6–7 days after fertilization, and if successful, after 10 days, the blastocyst is embedded completely in the uterine wall.
Key questions that remain include the way in which tolerance for the semiallogeneic embryo is established. How does the blastocyst gain access to the maternal endometrium, and what are the mechanisms that determine whether the maternal tissue accepts it? Is the blastocyst an aggressor that fights itself into the endometrium, or does the maternal tissue invite the developing blastocyst? Many different cytokines and adhesion molecules are involved in preparation for actual implantation. They form an intricate signaling network involving pro- and anti-inflammatory molecules. In this review, we will present a general overview of the main groups of signaling molecules involved, as well as some molecules that are the focus of current research. We do not aim to present a comprehensive overview but instead, to outline the major lines of signaling.
|
TOPABSTRACTINTRODUCTIONPOLARITY WITHIN THE BLASTOCYST...BLASTOCYST-ENDOMETRIAL...CONCLUSIONREFERENCES |
|---|
A second level of polarity is established when the blastocyst approaches the endometrium and adopts a specific orientation in relation to the endometrium. In humans, the inner cell mass of the blastocyst is directed toward the endometrium [8 ]. The exact mechanisms by which blastocyst orientation is obtained in humans remain unknown; however, several hypotheses have been put forward.
An important hypothesis involves a chemokine gradient in the endometrium [5 ]. In the uterus, chemokines set up a gradient that guides the trophoblast to the site of implantation. Past research has already established a similar role for chemokines in the attraction of leukocytes to the endometrial tissue during implantation [9 ].
In humans, several pieces of evidence supporting the theory that chemokines guide the blastocyst to the site of implantation have been described. The chemokines that have been localized to the uterus around the time of implantation include CX3CL1, CCL7, CCL14, and CCL4; on the blastocyst, several chemokine receptors have also been identified, namely CCR1, CCR3, and CX3CR1 [10 ]. Experiments in vitro have confirmed earlier findings that CCL14 induces trophoblast migration and shown that CCL4 and CX3CL1 have a similar effect [10 ]. Additional, albeit indirect, evidence to support the importance of chemokines in the migration of the blastocyst comes from the observation that scar tissue from previous endometrial surgery (or cesarean section) becomes an attractive site of implantation. This scar tissue is a persistent, inflammatory focus and may therefore secrete chemokines, which in turn attract the embryo [5 ]. The orientation of the embryo and identification of the spatial relationships that determine the location of implantation remain active areas of research.
|
TOPABSTRACTINTRODUCTIONPOLARITY WITHIN THE BLASTOCYST...BLASTOCYST-ENDOMETRIAL...CONCLUSIONREFERENCES |
|---|
The establishment of the window of implantation is under the control of steroid hormones, which perform their effects through local mediators, some of which will be described below [16 ]. Interestingly, the exact hormonal regulation of implantation varies per species. In mice and rats, the establishment of the implantation window requires the presence of estrogen and progesterone. However, estrogen is not needed in pigs, guinea pigs, hamsters, and rabbits [17 ]. In humans, an estrogen-primed endometrium develops under the control of progesterone [4 ].
As the above illustrates, there are substantial differences in the mechanisms used by the various species. However, as it is ethically and practically not possible to study the process of implantation in large numbers of humans and human embryos, researchers need to resort to animal models. Unfortunately, although different species all share the same goal, there are variations in how implantation is achieved [5 ]. Each stage of the implantation process in humans has its preferred animal model. This must be kept in mind when evaluating research, as findings are not by definition applicable to the human situation. Animal models currently in use include mice, rats, guinea pigs, rabbits, pigs, sheep, cows, and primates [5 , 18 ].
Cytokines and growth factors
Numerous cytokines and growth factors are involved in mother-fetus interactions. The major pathways of signaling are described here. The most important properties are summarized in Table 1
. Important interactions between cytokines are also depicted in Figure 1
as described later.
|
View this table: [in a new window] |
Table 1. Overview of Cytokines Involved in Implantation
|
 View larger version (27K): [in a new window] |
Figure 1. Schematic representation of the most important cytokine interactions surrounding the time of implantation. Boxes: red, proinflammatory; green, anti-inflammatory; black, not applicable; yellow, produced by blastocyst. Arrows: red, stimulatory; green, inhibitory; black, modulating in general. hCG, Human chorionic gonadotropin.
|
First, LIF is a proinflammatory cytokine which interacts with the LIF-βR and initiates a variety of intracellular signaling pathways. In the uterus, LIF-βR signaling seems to occur primarily through STAT3 [16 , 21 ]. Interestingly, the expression of STAT3 is only possible during a certain period of uterine receptivity. This limited STAT3 signaling is not dependent on LIF-R levels or affinity, indicating that the use of LIF-R signaling pathways changes over time, probably as a result of different metabolic states of the endometrium during the menstrual cycle [21 ].
LIF is expressed in the luminal epithelium during days 18–28 of the menstrual cycle (mid-late secretory phase), supporting a role in implantation [3 ]. During the secretory phase, LIF-βRs can be found on the luminal epithelium. Decidual stromal cells also produce LIF [3 , 16 ].
Several molecules influence LIF expression. LIF is linked to inflammatory pathways through IL-1 [16
, 22
]; TNF-
and leptin can stimulate LIF production [16
, 22
]. IGF and TGF-β also induce LIF secretion in a dose-dependent manner [23
].
In the endometrium, LIF has been attributed to several functions. LIF may control the proportions and amount of immune cells in the endometrium at the time of implantation [16 ]. In LIF-knockout (KO) mice, the number of endometrial macrophages was reduced significantly, whereas levels of NK cells and eosinophils were increased, indicating that LIF may recruit macrophages and restrict the migration of eosinophils and uterine NK (uNK) cells [3 , 24 ]. Also, LIF may mediate interactions between decidual leukocytes and the invading trophoblast [3 ]. Furthermore, LIF controls the status of the endometrium by endometrial LIF-R signaling. In mouse KO experiments, it was found that LIF is necessary for successful implantation of the blastocyst but not for blastocyst viability [25 ]. In LIF null mice, among other changes, the expression of glycans on the cell surface and the formation of pinopodes are disturbed [16 , 22 ].
Interestingly, LIF does not only affect the endometrium; there is evidence that LIF also affects the human blastocyst (and its precursors). When cultured in vitro, the blastocyst expresses the LIF-R [26 , 27 ]. The exact function of the "embryonic" LIF-R is not known; it is hypothesized that LIF mediates signals between immune cells in the decidualized endometrium and the trophoblast [28 ]. Furthermore, communication between the blastocyst and the endometrium is bidirectional. The blastocyst also produces LIF, thereby exerting control over the maternal endometrium [29 ]. The future embryo is also capable of regulating the endometrial production of LIF. There is evidence for a positive-feedback loop between the hCG produced by the blastocyst and LIF production in the endometrium. This communication gives the blastocyst an extra opportunity to influence the receptive status of the endometrium [23 ].
Fertility studies in which LIF production is compared between women with and women without fertility problems indicate that LIF is required for successful implantation, However, results are mixed, and conclusive in vitro evidence in humans is still missing [3 ]. Recent studies in rhesus monkeys indicate that uterine administration of anti-LIF mAb leads to reduced pregnancy rates [30 ]. Studies analyzing the effect of functional mutations in the LIF gene indicate that the absence of functional LIF in humans leads to reduced fertility [31 ]. In contrast, elevated levels of LIF in uterine flushings have been shown detrimental for successful implantation [32 ], and analysis of the effects of LIF levels in the peripheral blood of patients undergoing i.v. fluid (IVF) treatment showed no relation to pregnancy rates [32 , 33 ].
In all, LIF has a great number of (concentration-dependent) functions in the maternal endometrium and links many aspects of implantation. Its precise influence on fertility is not yet known, and further research is needed also to elucidate possible redundancies in function.
IL-6 is a proinflammatory cytokine produced (mostly) by endometrial epithelium and stromal cells at the time of implantation [19 ]. IL-6 is produced in a cyclical manner by the luminal epithelium; levels are the highest during the window of implantation and menstruation [23 ]. Research has shown that levels of IL-6 are relatively low during the proliferative phase and rise steadily during the secretory phase. In the late secretory phase, levels of IL-6 drop again (but not to a level as low as in the proliferative phase); these fluctuations may be under steroid control, but an autocrine loop has also been implicated [34 ]. Receptors for IL-6 can be found on the endometrium and the trophoblast [28 ]. IL-6 expression is under the control of several factors, including IL-1 [23 ]. Furthermore, steroid hormones, especially estrogen, induce the expression of IL-6 [34 ]. On the opposite, hCG and TGF-β inhibit IL-6 production [23 ].
Receptors for IL-6 have been found on the human embryo from the blastocyst stage onwards [28 ]. Embryonic hCG not only increases the production of LIF by the endometrium but also inhibits the endometrial release of IL-6 [23 ], which, in turn, interacts with receptors on the trophoblast, resulting in hCG release. Thus, hCG functions as a "two-way" modulator of cytokine release [23 ]. Through hCG, the blastocyst may increase inflammation using LIF and at the same time, keep inflammation in check by inhibiting IL-6 [23 ].
Recent research has found that low IL-6 mRNA production is associated with recurrent miscarriages, possibly as a result of a decrease of the functions of IL-6 in tissue remodeling, decidualization, and placenta/trophoblast development [35 ]. However, elevated serum levels have been found in women with recurrent abortions; unfortunately, in this study, IL-6 levels in the uterus were not measured [3 ]. These findings imply the possibility that an excess and a deficit of IL-6 may lead to adverse implantation outcomes.
Thus, IL-6 may have a role in the peri-implantation period, although some of its functions may be redundant, and excessive levels of proinflammation may be deleterious for successful implantation.
IL-11 is a multifunctional cytokine with anti-inflammatory properties [3
]. Stromal and epithelial cells produce it, and production is maximal during decidualization [19
]. The receptor for IL-11, IL-11R, is expressed by the luminal and glandular epithelia, and there is no cyclical variation in its levels. Thus, all variation in level of activity depends on IL-11 production [3
], which is influenced by steroid hormones and by more local factors such as relaxin and PGE2 [3
].
In vitro studies indicate that IL-11 has a role in the stimulation of decidualization in humans, possibly by acting on stromal cells [3
]. Furthermore, IL-11 induces a dose-dependent decrease in the production of the proinflammatory cytokine TNF- by endometrial epithelial cells; TNF- production is highest in the secretory phase but then drops off in early pregnancy. This drop may be a result of the production of IL-11 [19
]. LIF and IL-6 were not found to affect TNF- production [19
]. However, earlier work performed by the same authors showed that TNF- increases production of IL-11 in vitro [36
]. A possible explanation is that in vivo, IL-11 is under control of so many factors that changes in TNF- do not lead to detectable modulation of IL-11 expression or more probable, that IL-11 mediates a negative-feedback loop to keep the production of TNF- in check and vice versa. IL-11R KO mice are infertile, probably as a result of disruption of decidualization and possibly overinvasiveness of the trophoblast. The latter may occur through down-regulation of a metalloproteinase inhibitor [37
, 38
].
No receptors for IL-11 have been found so far on the human embryo. However, it does seem that in primates, embryo produces IL-11 during trophoblast invasion, providing a means to exert control over the endometrium [3 ].
A defect in IL-11 or its receptor may be involved in certain cases of human infertility. However, the significance of these findings remains to be determined [3 ].
Another cytokine system often implicated in embryonic implantation is the IL-1 system, which belongs to the IL-1β/TLR superfamily. The IL-1 system contains two agonists, IL-1 and IL-1β, two cell-surface receptors, IL-1R1 and IL-1R2, an accessory protein (IL1RAcP), and a naturally occurring antagonist, IL-1ra. IL-1 and IL-1β have the same biological activity and are produced as precursors that are later activated [3
, 39
]. IL-1R1 is found throughout the body in low numbers, whereas IL-1R2 is found on white blood cells only; IL-1R2 is said to be a decoy target for IL-1, thus indicating that IL-1R2 may have an inhibitory role on IL-1 activity [3
, 40
]. IL-1β signals through MAPK and NF-
B pathways [13
].
Expression of IL-1 can be found throughout the menstrual cycle, although expression of IL-1β by stromal cells, macrophages, leukocytes (including uNK cells), and endothelial cells is highest during the late secretory phase [3 , 41 ]. The IL-1R1 was found on the endometrial epithelium, with levels peaking in the late luteal phase [39 ], and on glandular epithelium and stromal cells, with a peak in the mid-late secretory phase [22 ]. Interestingly, researchers found that IL-1R2, which has an inhibitory effect on IL-1 signaling, is expressed at lower levels during the implantation window and thereby, alleviates the inhibition of the IL-1 system and thus, leads to higher IL-1 activity and facilitates a proinflammatory environment. IL-1R2 levels rise again during the late secretory phase and menstruation, thereby reinstalling the inhibition on IL-1 [40 ]. How this cyclic variation of IL-1R2 is mediated is still a focus of research. In general, the expression of IL-1 system components is up-regulated during pregnancy [3 ], not only by compounds of the IL-1 system but also by leptin. Furthermore, IL-11 induces IL-1β RNA expression during decidualization. However, although increased pro-IL-1β is found within decidual cells, this does not lead to production of active IL-1β [22 , 41 ].
IL-1 has several functions in the window of implantation. It stimulates the production of LIF by the endometrium [16 , 22 ] and the production of leptin and its receptor [3 ]. Furthermore, IL-1 increases the expression of the integrin β3 subunit, an adhesion molecule that plays an important role in apposition and adhesion, as will be described later [22 , 39 ]. IL-1 also appears to have an important function in decidualization [42 ].
The human blastocyst also responds to IL-1; research indicates that IL-1R1, IL-1β, and IL-1ra are expressed by blastocysts [39 , 42 ]. It was found that embryos release hCG in response to IL-1 stimulation [3 ]. Furthermore, the embryo also produces IL-1 itself [42 ]. Moreover, peri-implantation embryos producing IL-1ra were more likely to be arrested in development [39 ]. The embryo may use the IL-1 system to communicate with maternal tissues and to influence endometrial conditions. It is worth mentioning that IL-1 may also provide a means of communication during other stages of pregnancy, such as trophoblast invasion and the voyage through the fallopian tubes [39 ]. A deficiency in the IL-1 system does not impair reproduction in mice, although it has been found that antagonizing the biological effect of IL-1 does lead to implantation failure. The latter is mediated by an endometrial effect, most likely by inhibiting the expression of adhesion molecules [39 , 42 ].
Another cytokine in the same class as IL-1 (IL-1β/TLR superfamily) that appears to have a role in implantation is IL-18, which is a proinflammatory cytokine that is part of a family of proteins, also including IL-18R and IL-18BP, a neutralizer of IL-18 function [43 ]. IL-18 is produced as an inactive precursor in the cell and is activated later to active IL-18 by the action of caspases [44 ]. All components of the IL-18 system are produced by the endometrium throughout the human menstrual cycle; epithelial cells produce more IL-18 than the stromal compartment [43 , 45 ]. The production of IL-18 greatly increases during decidualization [46 ]. The IL-18R is expressed on the stromal cells of the endometrium [47 ].
Steroids are deemed important for the control of IL-18 expression; recent research indicates that estrogen inhibits the production of IL-18 and IL-18BP. However, the ratio of these two components is not altered, thus implying that there is no change in the amount of biologically active IL-18 [43
]. IL-18 induces a TH1 response, in particular, by stimulating the production of IFN-
[43
, 45
]. However, in the absence of IL-12, it can also induce a TH2 response [43
]. Furthermore, IL-18 also induces the production of IL-1β to increase proinflammatory signaling through the IL-18R [43
]. IL-18 is also postulated to have an effect in the activation of uNK cells [46
]. The role of these cells will be described below.
Evidence about the effects of IL-18 on fertility is mixed. An excess of IL-18 as well as a shortage appears to negatively influence implantation outcome [43 , 45 ].
Another molecule that has been put forward as an important modulator of mother-blastocyst communication is leptin, which was discovered initially as a regulator of food intake but is now characterized as a ubiquitous molecule that is present throughout the body [48 ]. Leptin and its receptor (OB-R) are expressed in the human endometrium [22 ]. Intracellular signaling occurs through the JAK/STAT pathways also used by the IL-6 family of cytokines and through MAPK and protein kinase C [22 , 48 ].
Research has shown that besides stimulation by IL-1, leptin secretion may be induced in an overall proinflammatory environment (involving cytokines such as IL-1, TNF, and LIF) and that leptin may in turn contribute positively to this environment [48 ].
Besides its functions as a modulator of cytokine expression, leptin also increases expression of the integrin β3 subunit, an effect it shares with IL-1, although leptin does so more strongly [3 , 22 ]. Furthermore, leptin also affects molecules involved in tissue remodeling, such as the matrix metalloproteinases (MMPs) [49 ]. Interestingly, it has been hypothesized that leptin mediates the interaction between nutritional status and reproduction by influencing the hypothalamus-pituitary-ovary axis and the ovary directly [48 ], thus providing a means to adjust the reproductive status to possible nutritional problems.
In early development, a gradient of leptin expression and its receptors can already be found on the blastocyst [48 , 50 ]. Recent research has found that in mice and sheep, the developmental response to leptin in vitro is concentration- and stage-dependent, although findings are contradictory on certain points [51 ]. Furthermore, around the time of implantation, the embryo can produce leptin and respond to it. There is also the possibility that the IL-1 system and leptin communicate together during this period; their expression patterns show similarities, and they are able to affect each other’s secretion [48 ]. Moreover, leptin may influence hCG secretion [48 ]. Finally, leptin also seems to be important for later stages of human pregnancy, for example, in trophoblast invasion and possibly fetal development [48 ]. It influences and is influenced by an array of cytokines and other signaling pathways. Interestingly, other adipose tissue-secreted hormones, such as ghrelin, may also play a role in establishing reproductive status and require more research to establish their physiological significance in implantation [52 ].
Leptin-deficient mice are obese and sterile, but fertility can be restored by exogenous administration of leptin [53 ]. In all, leptin is an important mediator in implantation, especially as it links many cytokine and hormone systems together.
The IGF/IGFBP system also plays a role in the communication between the embryo and the endometrium. This system includes IGF and IGFBP. In the endometrium, IGFBP-1 is the most important form of IGFBP [42 ]. There are two forms of IGF: IGF-1 and IGF-2. IGF-1 mediates the effects of estrogen on endometrial proliferation during the proliferative phase of the endometrial cycle; IGF-2 mediates the effects of progesterone during the secretory phase of the cycle [54 ]. Concordantly, it was found that IGF-1 is secreted mainly during the proliferative phase, whereas IGF-2 is secreted during the secretory phase, and both are produced by the stromal cells of the epithelium [55 ].
IGFBP-1 is produced by the endometrial stromal cells, especially during decidualization [56 ]. Expression of IGFBP-1 is under the control of several factors, including insulin and IGF, which inhibit the production of IGFBP-1. Stimulation of IGFBP-1 production occurs through progesterone and cAMP [56 ]. A stimulatory effect is also postulated for IL-1β [55 ].
It has been hypothesized that IGF-2 stimulates implantation and invasion (by promoting aggression), whereas IGFBP-1 counteracts this effect by inhibiting IGF-2 action. Evidence confirming this theory is that elevated levels of IGFBP-1 lead to implantation failure (as a result of a lack of aggression), and increased levels of IGF-2 are associated with tumor development [55 ].
During decidualization, the proinvasive IGF-2 is produced by the trophoblast, whereas stromal cells make IGFBP-1 [55 ]. IGF-2 acts by binding to IGF-1Rs on the endometrium. IGF is required for embryonic and placental development [42 ]. Furthermore, although data are conflicting, it appears that hCG produced by the blastocyst, in conjunction with progesterone, induces the production of IGFBP-1 by the stromal cells [55 ]. It was also found that IGFBP prevents IGF-2 from binding to the cell surface, thereby limiting the effects of IGF-2 on trophoblast invasion [55 ]. Inhibition occurs in spite of an IGF-independent, stimulatory effect of IGFBP on trophoblast invasion in vitro as a result of inhibition of fibronectin action [55 ]. The IGF/IGFBP system thus plays a role in establishing the implantation balance by stimulating and limiting inflammation.
Several other cytokines are a focus of current research about their role in mother-to-fetus communication, for one, glycodelin. As an immunomodulator, glycodelin has been proposed as a player in the window of implantation. Glycodelin is a 24-kDa glycoprotein that belongs to the lipocalin family and is produced by endometrial epithelial cells [57 ]. During the proliferative phase of the endometrial cycle, it can hardly be found; however, during the secretory phase and early pregnancy, glycodelin levels rise steadily [57 ]. Gene expression studies in humans confirm the findings that glycodelin expression is up-regulated during the window of implantation [58 ].
Glycodelin expression is controlled by several factors. Progesterone appears to be especially important, as glycodelin levels rise in conjunction with progesterone levels during the secretory phase [57 ]. Another molecule found to induce the production of glycodelin in vitro and in vivo is relaxin; this is a hormone produced by the corpus luteum that is thought to influence many physiological functions, including those important for pregnancy [59 60 61 ]. Furthermore, in baboons, it has been found that glycodelin is up-regulated by CG [62 ].
Glycodelin appears to have a suppressive effect on the maternal immune reaction to the fetal allograft [62 ]. In vitro, it inhibits the migration of T lymphocytes, [3H]-thymidine uptake by lymphocytes, and cell lysis by NK cells [62 , 63 ]. Glycodelin may therefore act as a mediator in establishing tolerance to the semiallogeneic child.
Another molecule that is a current focus of interest is OPN, which has been described originally as an extracellular matrix (ECM) component of bone and as a cytokine produced by lymphocytes and activated macrophages (in this context, also known as Eta-1) [64
]. It has myriad functions, including a role in implantation. Endometrial glands and decidual cells produce OPN during the time surrounding the window of implantation [64
, 65
]. OPN production is stimulated by, among others, IL-1, TGF-β, TNF-, IFN-, and the steroid hormones estrogen and progesterone [64
].
Concerning its functions as an adhesion molecule, OPN mediates cell–cell attachment and communication. Especially in sheep and pigs, but also in humans, it is deemed important for the attachment between the integrins of the human endometrium (vβ3; described below) and the trophoblast; the latter also possesses these integrins [64
65
66
]. As a cytokine, OPN has pro- and anti-inflammatory properties; it recruits and activates macrophages and lymphocytes and inhibits the production of NO and MMP-2 [64
, 66
].
OPN is also produced by the cytotrophoblast, again under the control of progesterone, thus strengthening the importance of OPN in implantation [64 , 66 , 67 ]. Uterine gland KO studies in sheep have shown that products of these glands, which include OPN, are necessary for successful implantation [68 ]. It is circumstantial evidence that OPN may be needed for implantation. However, KO studies of OPN null mutant mice have not shown any fertility defects, although inhibiting the integrins that it interacts with does lead to implantation failure. This suggests that although needed for successful implantation, some of the functions of OPN may be redundant [69 ].
Finally, we mention Dkk-1, also known as Dickkopf-1, which plays a role in implantation [4 ]. Dkk-1 is an inhibitor of the Wnt signaling pathway involved in processes of differentiation, proliferation, and more. It plays an essential role in the development of the embryo [70 ]. Dkk-1 is expressed by stromal cells in a cycle-dependent manner, with a progesterone-mediated peak in the mid-secretory phase [71 ].
The physiological role of endometrial Dkk-1 expression is still under investigation; it may have a role in decidualization and influence Wnt communication between the blastocyst and the mother [71 ]. Interestingly, endometrial Wnt signaling is required for fertility in mice and appears to be induced by the blastocyst [58 , 71 , 72 ].
A recent gene expression profiling performed on a coculture of trophoblasts and endometrial tissue confirms the presence of many cytokines described above. Furthermore, such experiments point to substances, which as of yet, have not been examined in the context of implantation and may also play roles in the various stages of implantation [47 ].
In all, when we look at the cytokine environment in which implantation occurs, we can draw several conclusions. Despite the extensive research performed in this area, the debate continues as to whether the process of implantation is best seen as an act of aggression by the embryo or as an invitation extended to future progeny by the maternal endometrium. To draw some general conclusions, most cytokines and their interactions described previously have been depicted graphically in Figure 1 .
As can be seen in the figure, pro- and anti-inflammatory molecules are involved not only in the establishment of the window of implantation but also in embryo-endometrium communication. All cytokines depicted are produced by the endometrium; the embryo also produces some of them in the various stages of its development. In the context of implantation, the induction of inflammation can be seen as a form of aggression by the embryo. Interestingly, all cytokines produced by the embryo are proinflammatory, suggesting that implantation is a process of aggression in which the embryo induces or even forces the endometrium to accept it. However, it must be kept in mind that aggression is not by definition a negative signal; it is necessary to achieve successful implantation.
On the other hand, maternal signals are key to the induction of a proinflammatory environment in the uterus through their mediation of the implantation window. In many cases, this is done independently of embryonic influences, implying an act of invitation from the maternal side. Interestingly, maternal steroids also induce anti-inflammatory molecules such as IGFBP-1 and IL-11 and thus provide a means to control the inflammation induced by the embryo and (independent) maternal signals. Furthermore, excessive levels of aggression have been found to correlate with failure of implantation. For example, as was described above, excessive levels of LIF and IL-6 are detrimental.
For some cytokines described earlier, in particular, LIF, IL-6, and IL-18, their effect on implantation outcome can be described by a bell-shaped dose-response curve. Low and high concentrations are detrimental, but an intermediate, optimal concentration is required for successful implantation.
Another approach to controlled aggression is related to the timeline of events, for example, the suppression of proinflammatory TNF- by anti-inflammatory IL-11 after the peak of TNF- in the secretory phase. This implies that a proinflammatory period may be followed by a period in which inflammation is suppressed.
Furthermore, Figure 1 emphasizes the close cooperation of the different systems involved; there are many connections among the immune cells, the adhesion molecules, and cytokine signaling. Finally, it must be kept in mind that some of these investigations gave contradictory results; such discrepancies may be attributed to differences in experimental methods, but more research is needed.
From the above, it seems that too much aggression by the embryo or from the endometrium (under induction of the embryo) leads to negative implantation outcome. However, aggression by the embryo is also needed to establish a successful pregnancy. Thus, controlled aggression, with aggression and inhibition of aggression, is needed to achieve implantation. A question that remains is whether the dual control described above follows a sequential model in which proinflammation is followed by anti-inflammation (as seen with IL-11 and TNF-) or whether there is a continuous balance between the pro- and anti-inflammatory environments.
In this respect, it is of interest that environmental influences such as psychological and physical stressors may have an as-yet underestimated role in embryo implantation and successful pregnancy. Literature provides many pieces of evidence showing that stress and the neuroendocrine system alter the complex network of pro- and anti-inflammatory cytokine and chemokine production, the activity of NK cells, macrophages, T cells, and others [73 ]. Like most cells of the body, the decidua expresses receptors for stress hormones such as cortisol and catecholamines [74 75 76 ]. Theoretically, it will therefore be more than likely that stress will alter the immunologic and neuroendocrine milieu at the feto-maternal interface with consequences for the outcome of pregnancy. Moreover, there are reports available describing the influence of stress in relation to reproductive failure, and these studies even propose that some women may benefit from behavioral intervention during early pregnancy [77 ].
It must be said that this review covers many cytokines and growth factors that are currently under close scrutiny. However, anti-inflammatory cytokines that are currently not so well researched in this context, for example, IL-5 and IL-10, may also be of importance, and their absence in most literature published today does not mean they do not have a function.
Adhesion molecules
Adhesion of the embryo to the maternal endometrium is a critical step in implantation. The adhesion molecules play roles in apposition and adhesion; selectins are involved mostly in apposition and integrins in adhesion. The mother and embryo influence the expression of adhesion molecules, and both parties are necessary to establish a successful interaction. The functions and interactions of adhesion molecules are described below.
First, the selectins which are involved in implantation and the process of apposition and adhesion. One study describes the expression of functional L-selectin on the trophoblast. This L-selectin can interact with oligosaccharide ligands on the maternal endometrium and give rise to a physiologically relevant interaction [78 , 79 ]. These findings allow the drawing of a parallel between the early stages of implantation (apposition, adhesion) and leukocyte extravasation from the bloodstream [78 ]. Interestingly, experiments performed by Genbacev et al. [78 ] support this analogy by showing that shear stress is needed for the activation of the L-selectin found on the trophoblast, just as is seen with the selectins on extravasating leukocytes [78 , 79 ]. Furthermore, the involvement of chemokines in the attraction of extravasating leukocytes and the orientation and attraction of the embryo supports the parallel. The binding of embryonic L-selectin to carbohydrate ligands on the luminal epithelium of the endometrium plays a role in the early (apposition) stage; blocking of L-selectin by antibodies leads to impaired adhesion [5 , 78 ]. Interestingly, integrins appear to be only involved in the adhesion and penetration phases of the process [5 ].
Despite this promising information, many questions remain. First, it must be said that earlier experiments could not detect L-selectin expression in the blastocyst [80 ]. Furthermore, questions remain about whether L-selectin expression is necessary for human reproductive function and related to the previous, why L-selectin-deficient mice are fertile [5 ].
The integrins also play an important role in implantation. In contrast to noncycle-dependent integrins, cycle-dependent integrins are involved in the receptivity of the endometrium [81
]. Expression of integrins on the endometrium is controlled by steroids and by some of the cytokines described above (for example, IL-1) [2
, 82
]. Many integrins have been proposed to play a role in the apposition and adhesion stages of implantation. In particular, the vβ3 integrin has been implicated many times as being necessary for implantation in humans and mice. The integrins are versatile and can bind to many components of the ECM, including OPN, fibronectin, laminin, and entactin. Furthermore, integrins appear to be up-regulated during the receptive phase of the menstrual cycle [17
, 66
, 83
]. Recent experiments in mice have shown that spatiotemporal variation in the expression of integrins may play an important role in implantation [84
]. Other integrins for which functions in the implantation window have been proposed are 9β1, vβ1, 1β1, 3β1, 6β1, vβ5, and vβ6 [17
, 81
, 83
, 85
].
Besides integrin expression on the endometrium, the blastocyst also expresses integrins on its outer surface. The human blastocyst expresses vβ3 as well as 3β1, 6β4, and vβ5. Evidence obtained from mice studies suggests that these integrins may be necessary for successful implantation [83
, 85
]. However, the blastocyst may need to be activated for implantation to occur [83
].
In terms of fertility, controversy exists about the effect of integrins on pregnancy [86 ]. A question for further research is whether integrin interactions facilitate signaling and thus, enhance the cross-talk between blastocyst and mother.
The role of mucins is still debated. Mucins are present in many parts of the human body and function mainly as lubricating and protective agents. In implantation, mucins, in particular, mucin 1 (MUC-1), show discrepancies between the murine and human processes [2 ]. Mucins are generally thought to inhibit cell–cell interaction [78 ]. During the receptive period in mice, MUC-1 expression is down-regulated under steroid control (progesterone), suggesting that the presence of MUC-1 hinders attachment and that down-regulation is needed for successful implantation [87 ].
In humans, however, the situation is more complex. Levels of MUC-1 seem to rise during the window of implantation, possibly mediated by progesterone [88 ]. Interestingly, in vitro experiments have shown that paracrine effects from the blastocyst on the endometrium may induce a local clearance of MUC-1 during the process of attachment, thus allowing embryonic implantation at that specific site [87 ]. This local down-regulation of MUC-1 occurs in conjunction with the steroid-mediated, general increase in MUC-1. Several advantages of such a MUC-1 barrier have been hypothesized. It may protect the embryo against the maternal immune system and prevent attachment at an incorrect site [87 ].
The exact mechanism through which the down-regulation of MUC-1 is achieved is not known, but several molecules including a disintegrin and metalloproteinase 17 and MMP-14 have been implicated [89 , 90 ]. More research will be needed to elucidate this matter further. Interestingly, MUC-1 may also be a ligand for the L-selectin found on the trophoblast.
Furthermore, the cadherin, and in particular E-cadherin, which form adherens junctions, play a role in implantation [17 ]. They have been implicated in several processes of implantation and development of the preimplantation embryo. For one, E-cadherin is critical for blastocyst formation; a mutation leads to defective embryonic development [17 ]. E-cadherin is also expressed by the maternal epithelium. It has been associated with the formation of a permeability barrier, which regulates the segregation between maternal, immunocompetent, immune cells and embryonic molecules in the primary decidual zone (zone in direct contact with the embryo) [91 ]. It is hypothesized that the embryo may induce this permeability barrier, although the signaling system used is not yet known [92 ]. A final possible function of the E-cadherins in implantation is the control and guidance of trophoblast invasion [91 ].
Another class of adhesion molecules that plays a role in human implantation includes trophinin and tastin [2 , 93 ]. Together, they can establish a homophilic interaction between the trophoblast and the endometrium [17 , 83 ]. This interaction can be improved further with the help of bystin [83 ]. Interestingly, trophinin does not seem to play a role during implantation in the mouse [94 ], although evidence about this matter is contradictory.
An adhesion-related phenomenon that appears to be relevant in implantation is pinopode formation (sometimes referred to as an uterodome). Pinopodes can be found on the luminal epithelium of the human endometrium during the window of implantation [95 ]. Pinopodes are progesterone-dependent organelles that appear on the apical surface of the epithelial cells as cellular protrusions. They become visible at days 20–21 of the menstrual cycle, although the precise moment of appearance has high individual variability [12 , 96 ]. It has been hypothesized that these structures facilitate implantation of the blastocyst by preventing cilia from sweeping off the blastocyst, promoting the withdrawal of uterine fluid and improving adhesion of the blastocyst to the luminal epithelium [12 , 97 ]. Furthermore, they may have a role in the uptake of macromolecules, although in humans, this function has not been fully characterized [96 , 97 ]. Pinopodes have also been implicated in the analogy between leukocyte migration and implantation; their morphology resembles the endothelial docking structures used by leukocytes to extravasate. Interestingly, pinopodes appear to be integrin-enriched areas, and the cytoplasmic rearrangement needed for their formation may cause changed expression of adhesion molecules and thus, facilitate adhesion and apposition [5 , 96 ].
In all, a number of adhesion molecules play a role in the various stages of blastocyst implantation. The luminal epithelium determines to a large extent the receptivity of the endometrium; correct expression of a glycocalix and adhesion molecules at the potential site of implantation is key. The apical surface of luminal cells is a dynamic environment with a constantly changing configuration. In nonconceptive cycles, important adhesion molecules can be found at the lateral side of the cells, and during the process of implantation, the distribution of adhesion molecules changes, thus altering the location and quality of cell junctions [98 ].
Adhesion molecules play roles in apposition and adhesion with selectins involved predominantly in the first step and integrins in the latter. The mother and the embryo influence the expression of adhesion molecules, and signals from both parties are necessary to establish a successful interaction. The effect of individual adhesion molecules on fertility and pregnancy rates is still controversial, and awareness of the limitations of knowledge and investigation possibilities is essential.
Concerning the contributions of mother and child, steroid hormones from the mother induce expression of integrins and mucins, whereas the trophoblast can modulate integrin expression (through IL-1), regulate local MUC-1 levels using paracrine signaling, and express L-selectin required for interaction with maternal oligosaccharides. Thus, although at least a certain degree of maternal invitation is required to establish a successful pregnancy, the contributions provided by the embryo are of pivotal importance as well.
Finally, it must be kept in mind that cytokine profiles influence adhesion molecule expression, and it cannot be said with certainty that adhesion molecules from mother and fetus cooperate independently from cytokine signaling; their cooperation could also be a result of interactions occurring at the cytokine level.
Immune cells, complement, and MHC interactions
Numerous immune cells are present in the endometrium, and they all contribute in their own way to the successful establishment of pregnancy. To prevent rejection of the embryo, their activity must be controlled.
The following table, adapted from Kämmerer et al. [1], briefly summarizes the various cell populations present in various stages of the endometrial cycle.
As can be seen in Table 2 , the cell populations, which are most important in implantation, are the uNK cells, macrophages, dendritic cells (DCs), and T cells. B cells and neutrophils play only a minor role and will not be discussed here.
|
View this table: [in a new window] |
Table 2. Immune Cells Distribution in the Human Endometrium at the Time Surrounding Pregnancy [1
]
|
Concerning their function, uNK cells may be critical determinants in the decision to initiate decidualization or menstruation [16 ]. In humans, as opposed to mice, they are present in the endometrium before fertilization [99 ]. Furthermore, uNK cells might control the maternal immune response to the fetal allograft as well as trophoblast invasion (and placenta formation) by modulating expression of cytokines [100 ]. They also modulate the vascularization of the endometrium; implantation sites in mice lacking uNK cells (or any NK cells) are abnormal; uterine arteries retain their nongravid structure instead of becoming spiral arteries [46 , 99 ]. Furthermore, in mice, vascular endothelial growth factor (VEGF) produced by uNK cells is hypothesized to guide angiogenesis in the endometrium. When blood flow is established, this function of uNK cells is no longer required, and their levels decline [99 ]. Similar functions have been postulated for the human situation [99 ].
Several factors seem to regulate the levels of the uNK cells in the endometrium. First, estrogen and progesterone have been implicated as a result of their association with the cyclic variations in uNK cell levels [100
, 105
]. However, as uNK cells do not express the progesterone receptor, and uNK cell activation occurs in mice lacking these receptors, indirect control is most likely [3
, 99
, 100
]. Molecules that have been suggested in this process are endometrium-derived MIP-1β and VEGF, as well as increased activity of adhesion pathways (homing the uNK cells to endometrium), expanding (non-uNK) leukocyte populations by proliferation, production of immunomodulatory substances such as glycodelin, and finally, creation of a TH2 environment that affects uNK proliferation and function [99
, 100
]. Furthermore, it is thought that IL-15 is responsible for uNK cell differentiation; uNK cells are activated to produce IFN- by IL-12 and/or IL-18 [46
]. Evidence supports a role for IL-11 in the maturation of u NK cells [106
]. Also, prolactin has been implied in the control of NK cell proliferation, maturation, and differentiation [100
]. Finally, establishment of chemokine gradients with 6Ckine, CXCL10, and CXCL11 is also involved in trafficking, maturation, and differentiation of uNK cells [99
].
NK cells interact with the trophoblast through HLA-G [103 , 107 ], and it may be through these HLA-G antigens that tolerance is established (see below) [108 ]. uNK cells are thus modulated by and in turn modulate many immune processes in the endometrium.
T cells have also been a focus of research. Initially, T cells were thought to influence implantation through a TH1/TH2 balance [109
]. Pregnancy was postulated to be a TH2-mediated event; TH1 cytokines such has IFN- and TNF- are associated with infertility and abortion, and these effects can be reversed in mice by injecting the TH2 cytokine IL-10 [110
, 111
]. These findings in mice were later extended to the human situation [112
, 113
].
However, with the discovery of Tregs and more elaborate cytokine profiling of the T cell subsets, it became clear that it is not as straightforward as thought previously.
Evidence that the TH1/TH2 hypothesis was not correct or insufficient to explain immune mechanisms at play during implantation started accumulating soon after the hypothesis was proposed. In humans (and mice), TH1 activity is required at several stages of pregnancy, in particular, during the early implantation period [114
]. In this time period, cytokines such as IL-1 and TNF- make the establishment of pregnancy possible, for example, by stimulating the production of LIF or increasing angiogenesis. Interestingly, the TH1 environment in turn stimulates the production of the TH2 cytokines [114
]. KO experiments support the importance of the TH1 cytokines. KO mice for LIF or IL-11 have resulted in sterile mice [113
]. Furthermore, research about mice has shown that although excessive levels of IFN- (TH1) are deleterious for the establishment of pregnancy, uNK-derived IFN- contributes to a normal pregnancy and facilitates pregnancy-induced artery remodeling [115
]. More evidence lies in the fact that IL-4 and IL-10 KO mice are fertile; this would not be expected if pregnancy required TH2 dominance [116
]. Research also showed that the expression of cytokines IL-11 to IL-18 was so complex that the TH1/TH2 paradigm should be considered an oversimplification; cytokines cannot be classified as good or bad based on whether they are TH1- or TH2-associated [109
].
In all, the complex process of embryonic implantation (and pregnancy) cannot be viewed as just a balance of TH1 and TH2. Tregs may explain further some of the phenomena seen in implantation.
Tregs may play a role in implantation and are essential for the establishment of peripheral tolerance; they suppress (auto-) reactive T cells [117 ]. Depletion of Tregs in mice has led to a better understanding of autoimmune disease [118 ]. Tregs have been characterized as CD4+CD25+ cells [119 ]. Forkhead box p3 (Foxp3) is present exclusively in Tregs and is necessary for their development and function [120 ]. It is not exactly known what factors determine Treg development. They are probably recruited from the bloodstream and induced to proliferate locally [121 ]. Recent evidence indicates that DCs are capable of expanding the Treg population [103 ] and control T cell maturation and phenotype switching in general. An important cytokine in Treg function and peripheral maturation from naïve CD4+ T cells to Tregs is TGF-β [122 ].
In the context of pregnancy, Tregs have been characterized as essential to the establishment of allotolerance. CD4+CD25+ T cells have been found in the human decidua at various stages of pregnancy. Their levels are highest in the peripheral blood during the first trimester of pregnancy [123 ]. Furthermore, autoimmune diseases improve during the course of pregnancy and subsequently relapse after delivery [124 ]. Moreover, experiments in mice have shown that the Treg population expands during pregnancy and that Tregs accumulate in the uterus [125 ]. These Tregs are necessary for a successful pregnancy, and their absence leads to failure of pregnancy in early-mid gestation. Interestingly, adoptive transfer of Tregs from pregnant mice to abortion-prone mice of the same strain prevents miscarriage [125 , 126 ]. Experiments performed by Zenclussen [126] in mice have shown that Tregs primed to paternal antigens in the periphery establish a localized, tolerant environment in the endometrium, in which successful implantation can occur. A similar need for Tregs has been described in the human situation [125 , 127 ]. This is supported further by a recent experiment in which endometrial biopsies of (unexplained) infertile and proven fertile women are compared for the expression of transcription factors that determine T cell differentiation. It was found that expression of Foxp3, necessary for Treg development and function, was significantly lower in the infertile group of women, whereas the transcription factors T-bet and GATA-3 (for TH1 and TH2, respectively) did not differ between both groups [121 ]. This supports the pivotal role of Tregs in successful human implantation.
Exactly how tolerance to the embryo is induced by Tregs is not (yet) known. One of the mechanisms that has been proposed is the induction of DCs to express indoleamine 2,3-dioxygenase (IDO); expression of IDO increases tolerance [125 ]. Furthermore, in general terms, Tregs are capable of inhibiting the activation of CD4+ and CD8+ T cells and APCs [123 ]. They also lower APC function by down-regulating the expression of costimulatory molecules [128 ]. Finally, production of IL-10 and TGF-β by Tregs and their expression of CTLA-4 may contribute to tolerance [129 ]. In this respect, recent findings from Blois et al. [130 ] are of interest. They report the involvement of galectin-1 in the priming of tolerogenic DCs, which in turn, appear key in the expansion of IL-10-secreting Tregs. Most intriguing is the finding that galectin-1 KO mice show high fetal loss in allogeneic matings but not in syngeneic matings. Thus, these data underline the importance of Tregs in the acceptation of the (semi-) allogeneic embryo [130 ].
In conclusion, Tregs are of key importance to successful establishment of pregnancy. Much research is still needed to identify the precise roles of CD4+CD25+ T cells and possibly other subsets of Tregs.
Macrophages also take part in implantation and decidualization. Unlike the uNK cells, macrophages remain present at high levels at the implantation site throughout pregnancy [131 , 132 ]. Research has found that IL-1 stimulates decidual production of macrophage-attracting chemokines such as CCL2, CCL5, CXCL2, CXCL3, and CXCL8 [133 ]. A recent in vitro study shows that the trophoblast is capable of recruiting monocytes/macrophages and of stimulating these cells to produce proinflammatory cytokines [131 ].
Several functions have been proposed for these macrophages. First, their cytokine production may help prepare the endometrium for pregnancy and maintain a proper cytokine balance [132
]. Furthermore, they may play a key role in cleaning up the apoptotic material, resulting from apoptosis of the trophoblast in the various stages of pregnancy, and mediate the extent of trophoblast invasion (through TNF-) [132
133
134
]. Cleaning up debris is critical, as it prevents the semiallogeneic trophoblast cells from initiating an immune response [132
]. Another possible function is to respond to possible infections; an in vitro experiment indicates that the trophoblast is able to control the response of macrophages to LPS [131
].
Macrophages serve many functions in the endometrium surrounding pregnancy. They are of importance in keeping levels of inflammation in balance—an equilibrium necessary for successful implantation.
Besides the uNK cells and macrophages described above, DCs, another component of the innate immunity, have also been localized in the endometrium [103 , 135 , 136 ]. DCs are the sentinels of the immune system, capable of activating it in response to an antigen. However, DCs can also control peripheral tolerance [103 ]. The immature DCs (DC-specific ICAM-grabbing nonintegrin+) are especially prevalent in the endometrium; these cells mature into CD83+ DCs in response to inflammatory cytokines or antigens, a process that occurs at the time surrounding implantation and may be mediated by progesterone [137 138 139 ]. Evidence indicates that besides their antigen-presenting function, decidual DCs can facilitate a TH2 response because of their low IL-12 production [140 ]. Further, it has been found that extensive NK/DC interactions take place in the human endometrium; through intimate contact and the cytokines they produce, NK cells and DCs may stimulate each other’s maturation and activation, and at the same time, maintain a negative-feedback loop on each other’s activity [103 ].
Interestingly, there is evidence that elevated DC levels in the later weeks of gestation (when numbers start falling) are associated with recurrent miscarriages [141 ].
Thus, DCs appear to have an important role in mediating the immune environment at the time surrounding implantation and early pregnancy. They not only influence the T cell environment but also control other immune cells, such as the NK cells.
Another component of immune reactions is the complement system. Recent research has shown that complement plays an essential role in the successful establishment of pregnancy. In particular, the inhibition of the complement system is essential for success [142 ]. Murine pregnancies have lethal outcomes when the complement system is inhibited insufficiently [142 , 143 ]. Elements of the complement system that can cause pregnancy failure if inhibited insufficiently are, among others, C3, C5, and C5aR [143 ]. Abnormal placental development and vascularization may be reasons for pregnancy failure [143 ]. Three molecules—decay accelerating factor, MCP, and CD59—can inhibit complement function and are found on the villous trophoblast [142 ].
From the above, it becomes clear that the complement system plays an important role in possible acceptance or rejection of the embryo.
Half of the genes in the blastocyst are paternal; this semiallogeneicity makes the blastocyst "foreign" in the eyes of the mother’s immune system. Besides the cytokine and immune-cell environment, MHC may also be of importance in acceptance of the semiforeign blastocyst.
In the human blastocyst, several classes of MHC molecules are expressed, although expression varies with location. The "fetus proper" presents the full range of MHC molecules, also the paternal ones [101 ]. In contrast, the trophoblast does not express any class II MHC molecules, nor the two main class Ia molecules (HLA-A and HLA-B). Molecules that have been found on the trophoblast are HLA-C (class Ia), HLA-G (class Ib), and HLA-E (class Ib) [101 ]. In terms of their polymorphism, HLA-C is polymorphic and may be a source of allorecognition; however, this does not seem to be a cause of infertility or termination of pregnancy [101 , 108 ]. HLA-E only has few polymorphisms that can lead to allorecognition [108 ].
HLA-G is a nonclassical MHC molecule characterized by restricted polymorphism and a splicing mechanism responsible for expression of membrane-bound and soluble isoforms [144 ]. The molecule is expressed by decidual, stromal cells, but some isoforms are also produced by infertilized oocytes and by the embryo before and after embryonic genome activation following the eight-cell stage. Some reports indicate that the level of soluble (s)HLA-G secreted by the embryo in the supernatant before IVF may function as a marker of implantation success, although others did not detect sHLA-G in the embryo culture supernatants [145 ]. It has been shown that in general, IL-10 is a powerful up-regulator of sHLA-G production, underlining the important role of the pro-and anti-inflammatory cytokine balance in pregnancy [146 ].
HLA-G seems to have several unique features. First, it is hypothesized not to be involved in antigen presentation but instead to be part of the regulation of tolerance at the maternal-fetal interface and in pregnancy success. It modulates the activity of T- and B-lymphocytes and macrophages/monocytes by inducing apoptosis or reducing cytotoxic activity (CD8+ cells) [108 ]. Expression of HLA-G on the trophoblast not only protects the trophoblast from cell lysis by the uNK cell [147 ], but HLA-G may also have an immunosuppressive effect by controlling cytokine production by uNK cells and the capacity to perform cell lysis [103 , 109 , 113 ]. When in rhesus monkeys, the putative homologue of HLA-G, Mamu-AG, was blocked in vivo by passive immunization, the growth and vascularization of the fetal placenta, placental modification of endometrial vessels, and a maternal leukocyte response to implantation, and differentiation of epithelial and stromal cells in the endometrium changed [148 ].
Several receptors for HLAs have been hypothesized to play a role in this interaction, including killer Ig-like receptor 2DL4 (KIR2DL4) and Ig-like transcript 2 [103
]. Stabilization of HLA-E by HLA-G could be the pathway through which uNK cells are inhibited [101
, 108
]. From an in vitro assay, it appears that HLA-G induces IFN- production by uNK cells [149
]; interestingly, when HLA-G is allowed to interact with unfractionated, mononuclear cells, IFN- levels drop [149
]. Thus, the trophoblast is capable of regulating the immune environment through HLA-G expression and its interactions with immune cells.
Moreover, HLA-E binds the CD96/NKG2 receptor on uNK cells; this interaction inhibits the cytotoxic activity of uNK cells and forms one of the numerous ways in which tolerance is established [150 ]. The HLA-E/CD96NKG2 interaction may have other effects on the endometrium as well [150 ].
Finally, a similar interaction has been described for HLA-C and the KIRs on the trophoblast. Interaction of HLA-C with KIR inhibits the cytotoxic activity of uNK cells [150 , 151 ]. Interestingly, an association has been found between women having recurrent, spontaneous abortions and a decreased expression of the KIR on the uNK cells [151 ]. This information supports that MHC interactions are of critical importance to achieving maternal tolerance.
As mentioned earlier, not all of the HLA molecules made by the trophoblast are expressed on all parts of the trophoblast. For example, villous syncytiotrophoblasts, which will be in contact with the maternal immune system, do not present any MHC antigens and thus are inert to the maternal immune system. Other parts of the trophoblast do carry the MHC antigens described above, e.g., HLA-C, HLA-G, and HLA-E [101 ].
The MHC expression pattern is summarized schematically in Figure 2 .
 View larger version (32K): [in a new window] |
Figure 2. Schematic overview of MHC expression on various parts of the blastocyst.
|
However, only trophoblast-mediated influences are not enough. Numerous endometrial factors contribute to the recruitment of immune cells and the establishment of a favorable environment, including the production of cytokines and the expression of adhesion molecules. Each party functions in its own way and in its own time period to achieve allotolerance.
Also, the immune cells do not just fulfill immunosuppressive actions; the "proinflammatory" cytokines produced by T cells and other components of the endometrium are also needed to achieve implantation. Furthermore, the protection from infections needs to be maintained.
Finally, it must be kept in mind that immune cells are not independent players in this process. They are influenced by the cytokine systems around them, and they in turn influence these systems to act in the way that they do.
|
TOPABSTRACTINTRODUCTIONPOLARITY WITHIN THE BLASTOCYST...BLASTOCYST-ENDOMETRIAL...CONCLUSIONREFERENCES |
|---|
Implantation is a complex process in which a semiforeign blastocyst gains access to the maternal endometrium. Multiple systems cooperate closely to establish a supportive environment for implantation. In all three main systems—cytokines, adhesion molecules, and immune cells—cooperation between mother and blastocyst is pivotal. The trophoblast can influence the maternal immune status using cytokines, MHC expression, and adhesion molecules and by inhibiting the complement system. However, the mother needs to do her share as well. Importantly, too much or too little action of mother or child is detrimental. In Figure 3 , myriad interactions needed for implantation have been distilled to a schematic overview of the most important ways by which tolerance and implantation are achieved. Evidence does not point directly to implantation as a process of invitation or aggression, supporting the notion of controlled aggression. Proinvasive and noninvasive cytokines are necessary for successful implantation. Their relation to implantation outcome can be depicted by a bell-shaped curve. Controlled aggression established by the cytokines and the induction of tolerance by the Tregs and uNK cells are of vital importance. Not to be forgotten, implantation is only possible when a physical connection is established by the adhesion molecules. Finally, not only intensive cooperation between mother and child but also close interaction among the three systems are of the utmost importance to achieve a successful establishment of pregnancy. One can imagine that individual factors cannot be fully singled out as the pivotal factor, as implantation is a dynamic and interactive process associated with continuous change.
 View larger version (15K): [in a new window] |
Figure 3. Schematic representation of the action of the three main systems involved in implantation: controlled aggression achieved by the cytokines, physical adhesion through the adhesion molecules, and immune suppression induced by the immune cells.
|
As with many other fields of science, there is still much to be discovered. Questions remain about how control over implantation is exerted. In the context of "controlled aggression", there are two main possibilities. There is either a continuous balance of pro- and anti-inflammatory molecules, or there is sequential regulation in which a proinflammatory moment is followed by an anti-inflammatory environment.
Furthermore, the ability of adhesion molecules to perform cell signaling may influence the implantation process in more than one way. More research into this concept is necessary.
Moreover, not much is known about the factors determining the spatial relationships of implantation in the human; insight into these factors will also shed light on the details of the implantation process.
It will also be of considerable interest to establish which of the molecules described represent a redundancy in function and which have their own, essential role in the establishment of pregnancy. KO models will not cover this question completely, as rodent models differ considerably from the human situation. As the use of human embryos meets great ethical constraints, much will have to be learned from clinical research and so-called "experiments of nature". A new ex vivo model described by Mardon et al. [152] and our lab involves culturing embryos on monolayers of stromal cells. In this model, we now study the effects of addition of cytokines and the consequences of genetic ablation of various mediators on embryonic implantation. This could provide us with new information about how implantation is regulated in the human.
Finally, the importance of Tregs cannot be emphasized enough. They are a relatively new concept in the process of implantation, and further research will be needed to establish their exact role. We believe that their antigen-specific mediation could turn out to be of pivotal importance and control many of the processes involved.
Received July 1, 2008; revised August 14, 2008; accepted August 14, 2008.
|
TOPABSTRACTINTRODUCTIONPOLARITY WITHIN THE BLASTOCYST...BLASTOCYST-ENDOMETRIAL...CONCLUSIONREFERENCES |
|---|
mRNAs in secretory phase endometrium of women with recurrent miscarriage J. Reprod. Immunol. 73,74-84[CrossRef][Medline]2-macroglobulin gene through janus kinase 2-signal transducer and activator of transcription 3 pathway in the decidua Mol. Endocrinol. 20,3240-3250vβ(3) integrin are coexpressed in the human endometrium during the menstrual cycle but regulated differentially J. Clin. Endocrinol. Metab. 86,4991-5000(v)β(3) integrin expression in the uterus and conceptus during the periimplantation period Biol. Reprod. 61,892-899, IL-1 β and IL-1ra: IL-1 endometrial microenvironment of the human embryo at the apposition phase under physiological and supraphysiological steroid level conditions J. Reprod. Immunol. 31,165-184[CrossRef][Medline] converting enzyme/ADAM 17 mediates MUC1 shedding J. Biol. Chem. 278,3386-3394 contributes to initiation of uterine vascular modification, decidual integrity, and uterine natural killer cell maturation during normal murine pregnancy J. Exp. Med. 192,259-270-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases J. Immunol. 155,1151-1164[Abstract] production by uterine natural killer cells Mol. Hum. Reprod. 10,189-195This article has been cited by other articles:
 |
|
 |
|
  S. Grewal, J. Carver, A. J. Ridley, and H. J. Mardon Human Endometrial Stromal Cell Rho GTPases Have Opposing Roles in Regulating Focal Adhesion Turnover and Embryo Invasion In Vitro Biol Reprod, July 1, 2010; 83(1): 75 - 82. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D M Baston-Buest, A Schanz, S Buest, J C Fischer, J S Kruessel, and A P Hess The embryo's cystatin C and F expression functions as a protective mechanism against the maternal proteinase cathepsin S in mice Reproduction, April 1, 2010; 139(4): 741 - 748. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.E. King, F. Collins, T. Klonisch, J.-M. Sallenave, H.O.D. Critchley, and P.T.K. Saunders An additive interaction between the NF{kappa}B and estrogen receptor signalling pathways in human endometrial epithelial cells Hum. Reprod., February 1, 2010; 25(2): 510 - 518. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-U. Chen, C.-H. Chou, K.-H. Chao, H. Lee, C.-W. Lin, H.-F. Lu, and Y.-S. Yang Lysophosphatidic Acid Up-Regulates Expression of Growth-Regulated Oncogene-{alpha}, Interleukin-8, and Monocyte Chemoattractant Protein-1 in Human First-Trimester Trophoblasts: Possible Roles in Angiogenesis and Immune Regulation Endocrinology, January 1, 2010; 151(1): 369 - 379. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. N Jabbour, K. J Sales, R. D Catalano, and J. E Norman Inflammatory pathways in female reproductive health and disease Reproduction, December 1, 2009; 138(6): 903 - 919. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. G. Holtan, D. J. Creedon, P. Haluska, and S. N. Markovic Cancer and Pregnancy: Parallels in Growth, Invasion, and Immune Modulation and Implications for Cancer Therapeutic Agents Mayo Clin. Proc., November 1, 2009; 84(11): 985 - 1000. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.M. Boomsma, A. Kavelaars, M.J.C. Eijkemans, E.G. Lentjes, B.C.J.M. Fauser, C.J. Heijnen, and N.S. Macklon Endometrial secretion analysis identifies a cytokine profile predictive of pregnancy in IVF Hum. Reprod., June 1, 2009; 24(6): 1427 - 1435. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
