Promoting Roles of Embryonic Signals in Embryo Implantation and Placentation in Cooperation with Endocrine and Immune Systems

Embryo implantation in the uterus is an essential process for successful pregnancy in mammals. In general, the endocrine system induces sufficient embryo receptivity in the endometrium, where adhesion-promoting molecules increase and adhesion-inhibitory molecules decrease. Although the precise mechanisms remain unknown, it is widely accepted that maternal–embryo communications, including embryonic signals, improve the receptive ability of the sex steroid hormone-primed endometrium. The embryo may utilize repulsive forces produced by an Eph–ephrin system for its timely attachment to and subsequent invasion through the endometrial epithelial layer. Importantly, the embryonic signals are considered to act on maternal immune cells to induce immune tolerance. They also elicit local inflammation that promotes endometrial differentiation and maternal tissue remodeling during embryo implantation and placentation. Additional clarification of the immune control mechanisms by embryonic signals, such as human chorionic gonadotropin, pre-implantation factor, zona pellucida degradation products, and laeverin, will aid in the further development of immunotherapy to minimize implantation failure in the future.


Endocrine System and Embryo Signals
In mammals, embryo implantation in the uterus is an essential process in successful pregnancy. In general, the endocrine system regulates endometrial differentiation such that the embryo can be implanted. The initial endometrial differentiation is induced by estrogen. Then, progesterone stimulates this estrogen-primed endometrium to differentiate further to make it more suitable for embryo implantation [1,2]. Estrogen is mainly secreted from growing follicles and progesterone is produced by the corpus luteum, a newly formed endocrine organ originating from the ovulated follicle. This sequential endocrine stimulation is closely coordinated with the estrus cycle, creating the endometrial receptive phase, referred to as an implantation window [3][4][5][6].
It is widely believed that such an implantation window, which spans from a few days after ovulation to several days prior to menstruation, also exists in women [7]. The human implantation window is estimated to correspond to cycle days 20 to 24 of the menstrual cycle [8]. However, there has been no study to directly confirm this window. Therefore, we developed an attachment assay using a human choriocarcinoma cell line, BeWo cells, and human primary endometrial epithelial cell culture to examine whether human endometrial receptivity changes during the menstrual cycle [9]. In this assay, high attachment rates were observed in endometrial culture derived from the mid-luteal phase. Of note, except for the mid-luteal phase, the attachment rates were low, suggesting that human endometrial receptivity changes during the menstrual cycle [10]. As BeWo cells easily attached to endometrial stromal cells or human endometrial carcinoma-derived Ishikawa cells, we suggest that certain adhesion-inhibitory factors are present on the endometrial epithelial cell layer. Consequently, we hypothesized that in the receptive phase, adhesion-promoting molecules increase, whereas adhesion-inhibitory molecules decrease. In addition, structural changes of epithelial cells, the so-called uterodome, were reported to be induced on the cell surface of luminal epithelial cells, which are suggested to be involved in embryo attachment to the endometrial epithelial layer [11,12].
Concomitant with hormonal preparation, direct cross-talk between the embryo and maternal endometrium is considered necessary to achieve subsequent successful embryo implantation [13,14]. The blastocyst will implant only when this molecular cross-talk is established [15,16]. Although the precise mechanisms remain unknown, it was proposed that human chorionic gonadotropin (hCG) is one of the important embryonic signals that increases the receptive ability of the sex steroid hormone-primed endometrium [17,18]. In nonhuman primates, hCG directly induced the expression of α-smooth muscle actin (SMA) in baboon endometrial stromal cells and glycodelin in the glandular epithelium, suggesting that the primate blastocyst signal alters the uterine environment prior to implantation [19]. In humans, the intrauterine administration of hCG using an intrauterine microdialysis system was reported to inhibit the expression of differentiation parameters: insulin-like growth factor binding protein-1 (IGFBP-1) and prolactin, while increasing the expression of implantation-related factors; leukemia inhibitory factor (LIF) and macrophage colony stimulating factor (M-CSF), and a neoangiogenetic factor: vascular endothelial growth factor (VEGF), in the mid-luteal human endometrium, suggesting that hCG regulates endometrial differentiation and vascularization [20,21]. Recently, hyperglycosylated hCG, an hCG isotype with larger N-and O-linked oligosaccharides, was suggested to play an important role in embryo implantation [22,23].
In addition to soluble factors, microRNAs secreted from human blastocysts were proposed to be new embryonic signals that regulate adhesive properties of endometrial epithelial cells. miR-661 from nonimplanted human blastocysts was taken up by primary human endometrial epithelial cells and it reduced the attachment of trophoblast cell line spheroids to these epithelial cells [24]. Later, the role of other noncoding RNAs in maternal-embryo communication through extracellular vesicles was observed, demonstrating the noncontact transfer of embryonic RNA transcripts to the endometrium and the altered expression of endogenous transcripts by endometrial cells [25]. A recent study also proposed that an embryo-secreted long noncoding RNA, phosphatase and tensin homolog pseudogene 1 (PTENP1), is involved in the endometrial adhesive properties [26].

Adhesion-Promoting and -Inhibiting Molecules
For human embryo implantation, several adhesion-promoting molecules, such as trophinin [27][28][29], L-selectin ligand [30][31][32][33], integrin αVβIII [34][35][36][37], and CD44 [38][39][40], were demonstrated to be expressed on human endometrial epithelial cells during the receptive phase. The expression of integrin βIII was found to be promoted by the embryonic interleukin (IL)-1 system, supporting the presence of cross-talk between blastocysts and the endometrial epithelium during embryonic implantation [41]. Recently, integrin αVβ3 and αVβ5 were reported to be necessary for the leukemia inhibitory factor-mediated adhesion of trophoblast cells to endometrial cells [36], whereas cell-surface CD44-hyaluronate binding was proposed to be employed by embryos during initial docking to endometrial epithelial cells [40]. This molecule is also involved in trophoblast invasion [42]. Consistent with the above report, we reported the possibility that versican, a large chondroitin sulfate proteoglycan that binds to hyaluronan and forms large extracellular matrix (ECM) aggregates, promotes human embryo attachment to endometrial epithelial cells [43]. In the bovine uterus, we also proposed that the interaction between vascular cell adhesion molecule 1 (VCAM1) in endometrial luminal cells and integrin α4 expressed on the conceptus functions in conceptus adhesion to the uterine endometrium [44].
As a regulatory cell surface protein of adhesion-promoting molecules, we previously reported that CD9, which regulates integrin function, is specifically expressed on the endometrial luminal and glandular epithelial cells [45]. CD9 and CD98 were subsequently demonstrated to function in embryo implantation [46][47][48]. CD9 is also involved in trophoblast invasion [49][50][51].
On the other hand, large glycoproteins, such as MUC1, that inhibit the physiological cell-to-cell interaction were expressed on murine and human luminal epithelial layers [52,53]. Later, the expression of MUC1 was found to be downregulated by human blastocysts using in vitro experiments, suggesting that MUC1 acts as an adhesion-inhibiting molecule that is locally removed by human blastocysts during the adhesion phase [54]. Another large glycoprotein, MUC16, was also found to be expressed on human endometrial epithelial cells, demonstrating that its expression is reduced on the cell surface of epithelial cells comprising the uterodome [55].
The above findings suggest that the expression of adhesion-promoting molecules increases, whereas that of adhesion-inhibiting molecules decreases according to embryonic signals during an implantation window.

Repulsive Molecules
A sufficient period is necessary for the cross-talk between embryo and mother in order to regulate embryo attachment to the endometrium with accurate timing and placement [56]. For example, implantation sites are regulated equidistantly in response to the number of implanting embryos, as observed in the pregnant murine uterus ( Figure 1). However, it is difficult to explain the precise mechanisms leading to equidistance of implantation sites only by the balance of expression profiles between adhesion-promoting and adhesion-inhibiting molecules. Previously, Chen et al. excellently described the possible mechanisms determining the sites of embryo implantation, proposing factors such as the uterine recognition of embryos, fine-tuned uterine peristaltic movements, time-controlled uterine fluid reabsorption and uterine luminal closure, and embryo orientation [57]. To achieve the above sequential events, we should focus on alternative mechanisms that prevent embryos from immediately attaching to an inappropriate site of the endometrium. From hatching to attachment, the bovine embryo floats in the uterine cavity for 4-5 days and becomes elongated prior to attachment [58][59][60]. This period may be an essential stage for embryomaternal cross-talk. However, there is little information concerning molecules that assure sufficient distances between the embryo and endometrium before attachment. In mice, ovariectomy before preimplantation on day 4 of pregnancy induces delayed implantation under progesterone supplementation [61]. When estrogen is administered, dormant blastocysts become activated and initiate implantation in the progesterone-primed uterus. During this blastocyst-uterine crosstalk before implantation, the endometrium undergoes further differentiation, secreting LIF. Notably, heparin-binding epidermal growth factor-like signaling was found to be induced in the activated blastocysts [62].
To investigate the mechanism to prevent premature embryo attachment, we hypothesized the involvement of repulsive forces. As repulsive force-inducing molecules, we observed that human maternal endometrial epithelial cells express ephrin A1 and human blastocysts express Eph A1, leading to the proposal that the Eph-ephrin system is involved in the initial interaction between human blastocysts (Eph A1) and the endometrium (ephrin A1) [63]. Ephrin A is a ligand for Eph A, and both molecules are located on the cell surface [64,65]. The ligand-receptor binding induces bidirectional signals affecting both cells, alters the functions of adhesion molecules [66], and promotes cell-to-cell adhesion, inducing arterio-venous anastomosis [67,68]. Of note, it also produces repulsive forces between the cells to induce axonal guidance [69].
Based on this background, we examined the time-course of mRNA expression of Eph/ephrin A in murine embryos. All subtypes of ephrin A were expressed on blastocysts. Importantly, around the hatching period, ephrin A expression transiently decreased and then increased again before attachment. Immunostaining revealed that ephrin A1 and A3 were expressed on the cell surface of the trophectoderm of blastocysts. Moreover, ephrin A3 expression on the blastocysts was weak on the opposite side of the inner cell mass, suggesting that the Eph-ephrin system acts as a regulator of the apposition site [70]. From hatching to attachment, the bovine embryo floats in the uterine cavity for 4-5 days and becomes elongated prior to attachment [58][59][60]. This period may be an essential stage for embryo-maternal cross-talk. However, there is little information concerning molecules that assure sufficient distances between the embryo and endometrium before attachment. In mice, ovariectomy before pre-implantation on day 4 of pregnancy induces delayed implantation under progesterone supplementation [61]. When estrogen is administered, dormant blastocysts become activated and initiate implantation in the progesterone-primed uterus. During this blastocyst-uterine crosstalk before implantation, the endometrium undergoes further differentiation, secreting LIF. Notably, heparin-binding epidermal growth factor-like signaling was found to be induced in the activated blastocysts [62].
To investigate the mechanism to prevent premature embryo attachment, we hypothesized the involvement of repulsive forces. As repulsive force-inducing molecules, we observed that human maternal endometrial epithelial cells express ephrin A1 and human blastocysts express Eph A1, leading to the proposal that the Eph-ephrin system is involved in the initial interaction between human blastocysts (Eph A1) and the endometrium (ephrin A1) [63]. Ephrin A is a ligand for Eph A, and both molecules are located on the cell surface [64,65]. The ligand-receptor binding induces bidirectional signals affecting both cells, alters the functions of adhesion molecules [66], and promotes cell-to-cell adhesion, inducing arterio-venous anastomosis [67,68]. Of note, it also produces repulsive forces between the cells to induce axonal guidance [69].
Based on this background, we examined the time-course of mRNA expression of Eph/ephrin A in murine embryos. All subtypes of ephrin A were expressed on blastocysts. Importantly, around the hatching period, ephrin A expression transiently decreased and then increased again before attachment. Immunostaining revealed that ephrin A1 and A3 were expressed on the cell surface of the trophectoderm of blastocysts. Moreover, ephrin A3 expression on the blastocysts was weak on the opposite side of the inner cell mass, suggesting that the Eph-ephrin system acts as a regulator of the apposition site [70].
On the other hand, Eph A1, which is a receptor of ephrin A, was expressed on the murine endometrial luminal epithelial cells before embryo attachment. After attachment, EphA1 expression markedly decreased at the implantation site, whereas its expression on the luminal epithelial cells was maintained at nonimplanted sites [70]. Although embryo attachment was inhibited on Eph A1-coated dishes, when floating blastocysts were transferred to the control dishes, they gradually attached and spread on the dishes, suggesting that the Eph A signal induces repulsive forces [70]. Considering these findings, we propose that the Eph-ephrin system functions in creating a sufficient period for cross-talk between the embryo and mother, providing embryos with an opportunity to attach to appropriate sites by repeating attachment and detachment in a timely manner ( Figure 2). Consistent with the above suggestion, the Eph-Ephrin A system was also reported to regulate the contact between blastocysts and endometrium during embryo implantation in swine [71,72], and low expression of Eph-ephrin A1 was later demonstrated to be related to an increase in the number of embryos during implantation [73]. On the other hand, Eph A1, which is a receptor of ephrin A, was expressed on the murine endometrial luminal epithelial cells before embryo attachment. After attachment, EphA1 expression markedly decreased at the implantation site, whereas its expression on the luminal epithelial cells was maintained at nonimplanted sites [70]. Although embryo attachment was inhibited on Eph A1coated dishes, when floating blastocysts were transferred to the control dishes, they gradually attached and spread on the dishes, suggesting that the Eph A signal induces repulsive forces [70]. Considering these findings, we propose that the Eph-ephrin system functions in creating a sufficient period for cross-talk between the embryo and mother, providing embryos with an opportunity to attach to appropriate sites by repeating attachment and detachment in a timely manner ( Figure 2). Consistent with the above suggestion, the Eph-Ephrin A system was also reported to regulate the contact between blastocysts and endometrium during embryo implantation in swine [71,72], and low expression of Eph-ephrin A1 was later demonstrated to be related to an increase in the number of embryos during implantation [73].

Figure 2.
Possible mechanisms to create an adequate period of cross-talk between the embryo and mother before attachment by repulsive forces. (A) During human embryo implantation, several adhesion-promoting molecules, such as trophinin, L-selectin ligand, and integrin αVβIII, are expressed on endometrial epithelial cells during the receptive phase. In contrast, large glycoproteins, such as MUC1, that inhibit the physiological cell-to-cell interaction are expressed on the murine and human luminal epithelial layer. (B) During the implantation window, the expression of adhesionpromoting molecules increases, whereas that of adhesion-inhibiting molecules decreases in cooperation with embryonic signals, leading to cell-to-cell interaction between the embryo and endometrial epithelial cells. This enables the embryo to receive Eph signals from endometrial epithelial cells through ephrin ligands on its surface, leading to repulsive forces between the embryo and endometrium. (C-E) By repulsive forces through the Eph-ephrin system, the embryo separates from the endometrium (white two-way arrow), repeating attachment and detachment. (F) Finally, the embryo attaches to appropriate sites with suitable timing.

Activation of the Trophectoderm after Attachment and Acquisition of Invasive Properties
After attachment, the human embryo initiates intraepithelial invasion. Activation of the trophectoderm is considered to be an essential process for subsequent embryo invasion through intraepithelial spaces toward the endometrial stroma [27]. The trophectoderm layer at the endometrial site continues to be activated during the migrating process of human embryos on day 7.5 after ovulation. However, the precise mechanisms leading to this activation remain unknown. As candidates for activation-inducing molecules, Sugihara reported that trophinin is expressed at the interacting site between the embryo and endometrium and proposed that trophinin-mediated cell Possible mechanisms to create an adequate period of cross-talk between the embryo and mother before attachment by repulsive forces. (A) During human embryo implantation, several adhesion-promoting molecules, such as trophinin, L-selectin ligand, and integrin αVβIII, are expressed on endometrial epithelial cells during the receptive phase. In contrast, large glycoproteins, such as MUC1, that inhibit the physiological cell-to-cell interaction are expressed on the murine and human luminal epithelial layer. (B) During the implantation window, the expression of adhesion-promoting molecules increases, whereas that of adhesion-inhibiting molecules decreases in cooperation with embryonic signals, leading to cell-to-cell interaction between the embryo and endometrial epithelial cells. This enables the embryo to receive Eph signals from endometrial epithelial cells through ephrin ligands on its surface, leading to repulsive forces between the embryo and endometrium. (C-E) By repulsive forces through the Eph-ephrin system, the embryo separates from the endometrium (white two-way arrow), repeating attachment and detachment. (F) Finally, the embryo attaches to appropriate sites with suitable timing.

Activation of the Trophectoderm after Attachment and Acquisition of Invasive Properties
After attachment, the human embryo initiates intraepithelial invasion. Activation of the trophectoderm is considered to be an essential process for subsequent embryo invasion through intraepithelial spaces toward the endometrial stroma [27]. The trophectoderm layer at the endometrial site continues to be activated during the migrating process of human embryos on day 7.5 after ovulation. However, the precise mechanisms leading to this activation remain unknown. As candidates for activation-inducing molecules, Sugihara reported that trophinin is expressed at the interacting site between the embryo and endometrium and proposed that trophinin-mediated cell adhesion functions as a molecular switch for trophectoderm activation in human embryo implantation through trophinin-dependent tyrosine phosphorylation [27].
As other candidates, we reported that activated leukocyte cell adhesion molecule (ALCAM) is expressed on both human blastocysts and endometrial epithelial cells [74]. ALCAM is a transmembrane molecule that belongs to the immunoglobulin superfamily and mediates cell-to-cell adhesion by ALCAM-ALCAM homophilic interactions or ALCAM-CD6 heterophilic interactions [75]. By oligomerization, conjugated ALCAM molecules further form a cluster, increasing their adhesive ability [76]. Furthermore, ALCAM-ALCAM interaction induces the differentiation of stem cells. In immune cells, CD9 directly associates with ALCAM, regulating homophilic (ALCAM-ALCAM) and heterophilic (ALCAM-CD6) interactions, increasing ALCAM-mediated cell adhesion and T-cell migration, activation, and proliferation [77]. ALCAM was detected on the cell surface of human endometrial glandular and luminal epithelial cells by immunohistochemical staining and flow cytometry, whereas ALCAM expression on the human embryo was detected in the blastocyst stage by immunostaining, suggesting that ALCAM-ALCAM interaction is associated with human embryo attachment to the endometrium [74]. However, several proposed cell-adhesion molecules, including ALCAM, were not detected by DNA microarray analysis of human trophectoderm biopsy samples [78]. This should be clarified using current single-cell transcriptome analyzing systems.

Opening of Tight Junctions in the Endometrial Epithelial Layer
After attachment, the activated embryo initiates intraepithelial invasion in the next step and invades the maternal endometrial stromal tissues as a mass, opening the epithelial cell layer. During this process, the intercellular connection of the endometrial epithelial layer becomes reduced, maintaining the connection between the embryo-attached epithelial cells and basement membrane.
In general, the epithelial layer is tightly conjugated as a barrier. Notably, immune cells migrate through the epithelial layer from the basolateral to apical membrane [79,80]. Adhesion molecules, such as CD11b/CD18 and CD47, were reported to be necessary for sequential opening of the epithelial layer [81,82]. Dendritic cells were reported to open a lateral connection of epithelial cells, recruiting bacterial antigens from the mucosal epithelium to lymphoid tissues [83]. Although the precise mechanisms are still unclear, dendritic cells were also reported to express tight-junction proteins such as occludin, claudin 1, and zonula occludens 1 [83].
During human embryo invasion, the activated trophectoderm migrates to the intercellular spaces of epithelial cells with reduced tight junctions. Through this process of trophoblast invasion, the connection between epithelial cells and the basement membrane must be maintained. Proteases degrade the epithelial structures to weaken the tight junctions among epithelial cells. However, they also reduce attachment to the basement membrane, causing detachment of the epithelial cell layer that maintains embryo attachment. In contrast to the immune system, epithelial-mesenchymal transition (EMT) was reported to be involved in opening of the epithelial layer [84][85][86][87]. However, EMT simultaneously reduces the connection of epithelial cells to the basement membrane, undermining the anchoring foundation of the implanting embryo. Consequently, EMT is not suitable for subsequent human trophoblast invasion.
After the embryo has firmly attached to the endometrium via several adhesion molecules, luminal epithelial cells receive continuous signaling from the attached embryo. The epidermal growth factor (EGF)-receptor system was proposed to induce the reduction of tight junctions of luminal epithelial cells by reducing the local calcium concentration at the trophoblast-endometrial interface [88]. As other candidates, we paid attention to the Eph-ephrin A system again because after the embryo has tightly adhered to epithelial cells, Eph-ephrin signals from the embryo can be continuously transduced to the endometrium [89]. The Eph-ephrin system was reported to induce reduction of the tight junction of epithelial cells via claudin 4, 5, and zonula occludens-1 [90,91]. The Eph-ephrin system was also found to promote cell attachment to the extracellular matrix [92]. Thus, we considered that this system reduces the endometrial epithelial barrier without destroying the connection between the embryo-attached epithelial cells and basement membrane [89] (Figure 3). On immunohistochemical study, Eph A1 and 4 were expressed on the luminal and glandular epithelium. Eph A2 was expressed on the pinopode-like sites of luminal epithelial cells [89]. Ephrin A1 induces phosphorylation in EphA2 and A4 in Ishikawa cells, a human endometrial carcinoma-derived cell line. Biological assays also revealed that the Eph A-ephrin A interaction induces cell attachment and intercellular dissociation in Ishikawa cells [93]. Ephrin A1-induced cell attachment was associated with the phosphorylation of focal adhesion kinase even in the presence of EDTA, suggesting the involvement of certain Ca ion-independent molecules. In contrast, EphA1-stimulation did not induce attachment of Ishikawa cells. Eph A2 was expressed on the pinopode-like sites of luminal epithelial cells [89]. Ephrin A1 induces phosphorylation in EphA2 and A4 in Ishikawa cells, a human endometrial carcinoma-derived cell line. Biological assays also revealed that the Eph A-ephrin A interaction induces cell attachment and intercellular dissociation in Ishikawa cells [93]. Ephrin A1-induced cell attachment was associated with the phosphorylation of focal adhesion kinase even in the presence of EDTA, suggesting the involvement of certain Ca ion-independent molecules. In contrast, EphA1-stimulation did not induce attachment of Ishikawa cells. In the porcine uterus, the expression of Eph A1, A2, A4, and A7 was strongly detected in endometrial epithelial cells during early pregnancy. Ephrin A1 stimulated the proliferation of endometrial luminal epithelial cells via the activation of phosphoinositide 3-kinase (PI3K) and mitogen-activated protein kinase (MAPK) signaling proteins, suggesting that ephrin A1 is involved in the interactions between porcine blastocysts and endometrial luminal epithelial cells by activating PI3K and MAPK signal transduction pathways [94].
Notably, we observed that both EPH A1 and ephrin A4 signals promoted the invasion of a human choriocarcinoma-derived cell line, JEG-3 cells, without affecting cell proliferation [95]. It was also proposed that Eph A2 promotes the invasion and proliferation of the human extravillous trophoblast (EVT) through the ephrin-A1 ligand [96].
Based on these findings, we propose that the ephrin A signal induces intercellular dissociation and adhesion to the basement membrane in endometrial epithelial cells as embryo signals and suggest that this signal facilitates subsequent trophoblast invasion (Figure 3).

Regulation of Endometrial Receptivity and Embryo Invasion by Embryonic Signals
It is widely accepted that cytokine networks function in the establishment of endometrial receptivity, suggesting that the immune system plays an important role in embryo-maternal crosstalk [97][98][99][100]. In mice, seminal fluid was demonstrated to affect immune cells to induce endometrial differentiation and promote embryo implantation by activating inflammation and inducing immunological changes in the female reproductive tract, which facilitate endometrial receptivity [101,102]. The significance of this mechanism in women is an important subject to improve the outcome of in vitro fertilization and embryo transfer (IVF-ET) therapy [103]. In the porcine uterus, the expression of Eph A1, A2, A4, and A7 was strongly detected in endometrial epithelial cells during early pregnancy. Ephrin A1 stimulated the proliferation of endometrial luminal epithelial cells via the activation of phosphoinositide 3-kinase (PI3K) and mitogen-activated protein kinase (MAPK) signaling proteins, suggesting that ephrin A1 is involved in the interactions between porcine blastocysts and endometrial luminal epithelial cells by activating PI3K and MAPK signal transduction pathways [94].
Notably, we observed that both EPH A1 and ephrin A4 signals promoted the invasion of a human choriocarcinoma-derived cell line, JEG-3 cells, without affecting cell proliferation [95]. It was also proposed that Eph A2 promotes the invasion and proliferation of the human extravillous trophoblast (EVT) through the ephrin-A1 ligand [96].
Based on these findings, we propose that the ephrin A signal induces intercellular dissociation and adhesion to the basement membrane in endometrial epithelial cells as embryo signals and suggest that this signal facilitates subsequent trophoblast invasion (Figure 3).

Regulation of Endometrial Receptivity and Embryo Invasion by Embryonic Signals
It is widely accepted that cytokine networks function in the establishment of endometrial receptivity, suggesting that the immune system plays an important role in embryo-maternal cross-talk [97][98][99][100]. In mice, seminal fluid was demonstrated to affect immune cells to induce endometrial differentiation and promote embryo implantation by activating inflammation and inducing immunological changes in the female reproductive tract, which facilitate endometrial receptivity [101,102]. The significance of this mechanism in women is an important subject to improve the outcome of in vitro fertilization and embryo transfer (IVF-ET) therapy [103].
On the other hand, we found that the intravenous administration of splenocytes from early pregnancy induced successful implantation in pseudo-pregnant day 2 recipient mice by promoting endometrial differentiation, which is suitable for embryo implantation [104]. In the delayed implantation model, the administration of splenocytes induced LIF expression in the uterus and subsequent embryo implantation [105]. In contrast to pregnant mice, splenocytes from pseudo-pregnant mice had no significant effects. As pseudo-pregnant recipient mice had been sensitized with the seminal fluid of vasectomized male mice [106], we considered that developing embryos in the Fallopian tube and uterus affect the maternal immune function to prepare suitable uterine conditions for implantation [107] and proposed the presence of the dual control of endometrial differentiation before embryo attachment via the endocrine and immune systems by embryonic signals [108][109][110].
The next question is how the maternal immune system recognizes the presence of a developing embryo before implantation and distinguishes the developing embryo from nonfertilized eggs and/or other organisms in the female genital tract [107]. To achieve this, the developing embryo should convey species-and embryo-specific signals, i.e., embryonic signals, to the maternal immune system. Early pregnancy factors that inhibit T-cell-induced rosette formation were reported to be detected just after fertilization [111]. Later, several substances, such as platelet activating factor, thioredoxin, and chaperonin 10, were proposed as early pregnancy factors [112][113][114]. It was also reported that the embryo-derived pre-implantation factor (PIF), a novel embryo-specific 18-kDa peptide that is specifically expressed on the fetus and placenta [115], was secreted at the two-cell stage and was able to be detected in the maternal sera prior to implantation [116]. Its many biological roles are under investigation [115,117,118].
Importantly, early pregnancy factor (EPF) is not species-or embryo-specific, whereas PIF is not species-specific to the maternal immune system. Consequently, as a new candidate for the embryonic signal before implantation, we paid attention to the zona pellucida (ZP) because it functions in the species-and oocyte-specific binding of spermatozoa [119]. As developing embryos actively degrade ZP, which is an abundant store of species-and oocyte-specific glycoproteins, from fertilization to hatching, we proposed that degradation products of ZP, including their sugar moieties, are utilized as intrinsic embryonic signals, which transmit information about the presence of the developing embryo to the maternal immune system in the reproductive tract [107].
As described above, hCG, a species-and embryo-specific glycoprotein, was proposed to directly affect the endometrial function as an embryonic signal after hatching [17][18][19][20][21]23]. Notably, this leading embryonic signal is also considered to induce an endometrial immune environment that accepts allograft implantation of the embryo, inducing fetal immune tolerance [120][121][122]. Approximately half a century ago, urinary hCG was demonstrated to suppress immune reactions, leading to the proposal of an essential role of hCG in the induction of immune tolerance to the fetus [123]. However, the following studies demonstrated that highly purified hCG did not have immunosuppressive effects [124]. Since then, the effects of hCG on immune cells remained controversial for a long time. Based on this background, we reported that recombinant hCG binds and activates CD14-positive monocytes to promote IL-8 production partially through the nuclear factor-kappa B (NF-κB) pathway at relatively high concentrations of more than 10 IU/mL. This hCG-induced IL-8 production was inhibited by the exogenous excess of D-mannose, suggesting that hCG regulates the peripheral blood mononuclear cell (PBMC) function through sugar chain receptors [125]. Later, it was demonstrated that a high concentration of hCG regulates uterine NK cell proliferation via mannose receptors rather than by luteinizing hormone (LH)/hCG receptors [126]. The carbohydrate chains of urinary hCG are largely cleaved before urine production [127]. Accordingly, the possible involvement of a carbohydrate-mediated primitive mechanism in the maternal response by the immune system may explain the previous discrepancy in the immunosuppressive effects on immune cells between urinary crude hCG and highly purified hCG and the reason why such a high concentration of hCG is necessary to maintain normal pregnancy [125].
In contrast to the above lectin-glycan interaction, hCG induces regulatory T-cells to migrate to the trophoblast through LH/hCG receptors that are expressed on regulatory T-cells [120]. Subsequent studies suggested the essential role of hCG in pregnancy-induced immune tolerance and embryo implantation, regulating the establishment of an adequate embryo-endometrial relationship [121,128]. In addition, CD19+CD24(high+)CD27+ regulatory B-cells were demonstrated to produce IL-10 by hCG stimulation via LH/hCG receptors [129]. This type of regulatory B-cell was proposed to mediate the positive effects of hCG on the immune environment during pregnancy [130,131].
In invasion assays using the murine embryo and BeWo cells, peripheral blood mononuclear cells (PBMCs) derived from women in early pregnancy promoted the murine trophectoderm and BeWo cell invasion more than those obtained from nonpregnant women. Importantly, when PBMCs from nonpregnant women were incubated with hCG, hCG-treated PBMCs promoted invasion more than nontreated PBMCs by soluble chemoattractive factors derived from PBMCs, suggesting that hCG alters PBMC functions to facilitate embryo implantation [132,133]. Later, similar effects of hCG on PBMCs to promote trophoblast invasion were reported using JAR cells, a cell line established from a human choriocarcinoma, inducing increases in matrix metalloproteinase (MMP)-2, MMP-9, and VEGF, and decreases in tissue inhibitor of metalloproteinase (TIMP)-1 and TIMP-2 expression [134].
Based on these findings, we propose that hCG stimulates maternal immune cells at the implantation site through lectin-glycan interaction, which in turn, promotes embryo attachment and invasion based on cooperation between the endocrine and immune systems [110].

Regulation of Extravillous Trophoblast (EVT) Invasion by Embryonic Signals
During human placental formation, the cytotrophoblast differentiates into EVT in the anchoring villi and invades the endometrial stromal tissues as a single cell, reconstructing maternal spiral arteries. The reduction of arterial contractility caused by arterial reconstruction facilitates adequate maternal blood flow into the intervillous spaces [135]. Failure of this process will lead to an insufficient blood supply and cause placental dysfunction and preeclampsia in the late stage of pregnancy [136].
As a representative embryonic signal during human trophoblast invasion, hCG was initially reported to directly reduce trophoblast invasion by inhibiting the enzyme activity of urokinase-plasminogen activator [137]. However, subsequent studies demonstrated that hCG increased the invasion and migration of JEG-3 cells, a trophoblastic cell line derived from choriocarcinoma [138]. hCG was also reported to increase cell migration of an EVT cell line, HTR-8 SVneo cells, through an insulin-like growth factor-II axis [139]. Later, hyperglycosylated hCG secreted by the invasive EVT, but not hCG produced by the syncytiotrophoblast, was found to promote trophoblast invasion [140]. It was also reported that hCG, hCGβ, and their hyperglycosylated forms stimulate the invasion of JEG-3 cells through an independent pathway involving the classical LH/hCG-receptor [141]. As described above, hCG-stimulated immune cells were demonstrated to promote trophoblast invasion by secreting chemoattractants through a paracrine mechanism [132][133][134]. Of note, human endometrial stromal cells were reported to increase the invasion of HTR-8/SVneo cells under hCG stimulation, suggesting paracrine effects of hCG on trophoblast invasion through endometrial stromal cells and local immune cells [115].
As a novel embryonic signal candidate, we identified a new cell surface aminopeptidase, initially named 'laeverin', which belongs to the M1 peptidase family [142], and was later termed 'aminopeptidase Q' [143]. Members of the M1 family of aminopeptidases share a common peptide-binding site (GXMEN) and peptidase activity motif (HEXXHX18E). Primate laeverin has a unique peptide-binding motif (HXMEN) where the first glycine (Gly) residue is substituted with histidine (His) [143], inducing significant changes in substrate specificity toward natural peptide hormones [144]. Laeverin is specifically expressed on EVT in the placenta from early and term pregnancy. In primary villous explant cultures, laeverin expression was induced on the cell surface of the outgrowing EVT, and secretion of soluble laeverin was detected in the culture media. The invasion of EVT isolated from primary culture was suppressed by the reduction of laeverin mRNA expression, whereas the soluble form of recombinant laeverin promoted EVT invasion, suggesting a regulatory role of laeverin in EVT invasion [145]. Similar to HLA-G, the expression of laeverin is specifically limited to EVT [142]. The co-expression of HLA-G and laeverin suggests specific roles of laeverin in the regulation of immune tolerance, which should be clarified in the future.

Topics of Clinical Application of Immune Therapy
At present, implantation failure is one of the most important clinical problems in IVF treatment. It should be noted that IVF therapy skips a large part of the maternal immune recognition process when the developing embryo is present in the female genital tract. Consequently, to achieve successful implantation, the precise assessment and accurate control of an active state of maternal immune tolerance mediated by regulatory T-cells is considered to be necessary [146]. Recently, personalization of immune treatment was recommended based on uterine immune profiles such as low immune activation or immune overactivation [147]. Currently, use of the new immunosuppressive agents tacrolimus and cyclosporine, which minimize immune rejection of transplanted organs, has been proposed as useful for infertile patients with immune overactivation [148,149]. When infertile patients with overactive immune conditions were carefully selected, tacrolimus was reported to improve the pregnancy outcome [149][150][151]. To support this approach, a common mTOR (mammalian target of rapamycin) inhibitor, sirolimus, which effectively prevents allograft rejection, was recently demonstrated to improve clinical pregnancy and live birth rates after repeated IVF therapy failures by reducing the Th17/Treg cell ratio [152].
In contrast to immunosuppressive therapy, we previously developed a novel therapy using autologous PBMCs. Briefly, PBMCs are isolated from patients and incubated with hCG, an embryonic signal, in order to sensitize them. The activated PBMCs are administered into the uterine cavity to induce adequate endometrial differentiation three days before blastocyst transfer. This treatment effectively improved the pregnancy and implantation rates in patients with four or more repeated IVF therapy failures [153]. The intrauterine administration of corticotropin-releasing hormone (CRH)-treated autologous PBMCs was also demonstrated to improve clinical pregnancy rates of patients with repeated implantation failure [154]. Currently, the effectiveness of PBMC therapy has been positively reported [155][156][157][158].
Embryonic signals elicit two different effects: induction of immune tolerance and adequate local inflammation; the former protects against immune rejection, whereas the latter contributes to endometrial differentiation and maternal tissue remodeling during embryo implantation and placentation [108][109][110] (Figure 4). At present, there are few effective therapies to induce adequate endometrial differentiation and receptivity for infertile patients who poorly respond to endocrine stimulation. In this regard, the intrauterine administration of embryonic signal-activated PBMCs is one of the promising approaches to induce favorable endometrial differentiation and inflammatory reactions together with a favorable immune environment for embryo implantation [108] (Figure 4). It should also be noted that PBMCs can change the function or structure of endometrial surface molecules by secreting proteases. Since PBMCs have been suggested to move from the uterine cavity toward endometrial stromal tissue, they may create several pathways for subsequent embryo attachment and invasion. . Strategy of immune therapy using embryonic signals. Embryonic signals from the genital tract act on both endocrine (red arrows) and immune (blue arrows) systems to induce endometrial differentiation and immune tolerance. To assist the endocrine system, hormone replacement therapy was recommended to promote endometrial receptivity (red triple lines), whereas to minimize immune rejection, immunosuppressive agents were administered to elicit immune tolerance (blue triple line). On the other hand, to induce sufficient endometrial differentiation and receptivity in infertile patients who poorly respond to endocrine stimulation (black dotted line), intrauterine administration of embryonic signal-activated peripheral blood mononuclear cells (PBMCs) is one of the promising approaches that may induce favorable endometrial differentiation and inflammatory reactions together with a favorable immune environment for embryo implantation (black triple lines).

Conclusions
In conclusion, we propose the involvement of repulsive molecules, the Eph-ephrin system, in endometrial receptivity for embryo implantation, together with adhesion-promoting and adhesioninhibiting molecules. This Eph-ephrin system may contribute to opening tight junctions among endometrial epithelial cells to induce embryo invasion toward endometrial stroma tissues, suggesting that Eph-ephrin molecules are new candidates for embryonic signals that induce suitable local inflammation for embryo invasion. Embryonic signals also elicit two different effects from the immune system: induction of immune tolerance that protects against immune rejection, and sufficient local inflammation that functions in endometrial differentiation and remodeling during embryo implantation and placentation. Further clarification of the precise roles and possible clinical usage of embryonic signal candidates, such as PIF, ZP degradation products, or laeverin, will help improve immunotherapy to minimize implantation failure in the future.  . Strategy of immune therapy using embryonic signals. Embryonic signals from the genital tract act on both endocrine (red arrows) and immune (blue arrows) systems to induce endometrial differentiation and immune tolerance. To assist the endocrine system, hormone replacement therapy was recommended to promote endometrial receptivity (red triple lines), whereas to minimize immune rejection, immunosuppressive agents were administered to elicit immune tolerance (blue triple line).
On the other hand, to induce sufficient endometrial differentiation and receptivity in infertile patients who poorly respond to endocrine stimulation (black dotted line), intrauterine administration of embryonic signal-activated peripheral blood mononuclear cells (PBMCs) is one of the promising approaches that may induce favorable endometrial differentiation and inflammatory reactions together with a favorable immune environment for embryo implantation (black triple lines).

Conclusions
In conclusion, we propose the involvement of repulsive molecules, the Eph-ephrin system, in endometrial receptivity for embryo implantation, together with adhesion-promoting and adhesion-inhibiting molecules. This Eph-ephrin system may contribute to opening tight junctions among endometrial epithelial cells to induce embryo invasion toward endometrial stroma tissues, suggesting that Eph-ephrin molecules are new candidates for embryonic signals that induce suitable local inflammation for embryo invasion. Embryonic signals also elicit two different effects from the immune system: induction of immune tolerance that protects against immune rejection, and sufficient local inflammation that functions in endometrial differentiation and remodeling during embryo implantation and placentation. Further clarification of the precise roles and possible clinical usage of embryonic signal candidates, such as PIF, ZP degradation products, or laeverin, will help improve immunotherapy to minimize implantation failure in the future.