Functions of interferon tau as an immunological regulator for establishment of pregnancy

Abstract

The establishment of a successful pregnancy requires a “fine quality embryo”, “maternal recognition of pregnancy”, and a “receptive uterus” during the period of conceptus implantation to the uterine endometrium. In ruminants, a conceptus cytokine, interferon tau (IFNT), a major cytokine produced by the peri‐implantation trophectoderm, is known as a key factor for maternal recognition of pregnancy. IFNT can be considered one of the main factors in conceptus–uterus cross‐talk, resulting in the rescue of ovarian corpus luteum (CL), induction of endometrial gene expressions, activation of residual immune cells, and recruitment of immune cells. Much research on IFNT has focused on the CL life‐span (pregnancy recognition) and uterine gene expression through IFNT and related genes; however, immunological acceptance of the conceptus by the mother has not been well characterized. In this review, we will discuss the progress in IFNT and implantation research made by us and others for over 10 years, and relate this progress to pregnancy in mammalian species other than ruminants.

Factors required for the establishment of pregnancy

Early embryonic losses are thought to occur primarily during the peri‐attachment period. The establishment of a successful pregnancy requires a “fine quality embryo”, “maternal recognition of pregnancy”, and a “receptive uterine endometrium” during this period. The maternal recognition of pregnancy has been defined as the uterine recognition of the conceptus and extension of the ovarian corpus luteum (CL).

In any mammalian species, the maintenance of CL function and the continued secretion of a steroid hormone, progesterone (P₄), are required for the establishment and maintenance of pregnancy. However, the biochemical mechanism by which the CL is maintained differs from species to species. In humans, luteolysis is prevented by a luteotropic factor, chorionic gonadotrophin (CG), produced by implanting trophoblast cells [1]. In rodents, CG is not produced, and the prolongation of the CL life span is the result of pituitary prolactin surges [2]. In ruminants such as cows, sheep, and goats, interferon tau (IFNT), a major cytokine produced by the peri‐implantation trophectoderm, is known as a key factor for the maternal recognition of pregnancy [3, 4, 5]. IFNT does not exhibit direct luteotropic activity, but rather contributes to the prevention of luteolysis by attenuating prostaglandin F2α (PGF) secretion from the uterine endometrium, resulting in pregnancy recognition and establishment. The attenuation of PGF by IFNT results from the down‐regulation of estrogen receptor and the subsequent estrogen‐induced oxytocin receptor expression [6, 7].

Interferon tau is secreted into the uterine lumen only by trophoblast cells of peri‐implantation conceptuses [4, 5]. The expression of IFNT is regulated in a temporal and spatial manner. In ovine species, IFNT secretion begins on day 8 of pregnancy (day of estrus is day 0), and increases thereafter, during which time the conceptus elongates [8, 9]. On day 16, just before conceptus attachment to the uterine epithelium, IFNT production reaches its peak level. The expression of IFNT then declines rapidly as the process of implantation proceeds, and by day 22 of pregnancy, when the trophoblast is fully attached to the maternal endometrium, IFNT is no longer detected [3, 4, 9, 10]. Bovine and caprine conceptuses also exhibit a similar cell‐specific and temporal pattern of IFNT expression [11, 12, 13].

Evidence accumulated so far strongly suggests that the function of IFNT is not only to prevent luteolysis, but to maintain maternal uterine conditions suitable for conceptus development. Before placental development in mammalian species, conceptuses rely on uterine secretions to facilitate the progress toward implantation to the maternal endometrium. It is likely that, in response to ovarian P₄, the uterine endometrium secretes numerous cytokines and growth factors which regulate conceptus production of cytokines, some of which are involved in fetal–maternal cross‐talk. IFNT can be considered one of the main factors in these conceptus–uterus cross‐talk (Fig. 1). Therefore, the functions and regulatory mechanisms of IFNT have been studied in our laboratory for over 10 years.

Figure 1.

Conceptus–maternal cross‐talk via interferon tau (IFNT). For pregnancy establishment in mammals, a “fine quality embryo”, “maternal recognition of pregnancy”, and “receptive uterine endometrial cells” are absolutely required. It has been well characterized that IFNT is one of the main factors required for maternal–fetal cross‐talk and pregnancy recognition. It is becoming apparent that IFNT is involved in the activation and recruitment of immune cells, resulting in the conditioning of the immune environment in utero

Maternal recognition of pregnancy; ruminants’ IFNT and human CG

In a human pregnancy, hCG secreted from the embryonic trophectoderm stimulates the CL to continuously produce P₄, which acts on the endometrium to support embryo implantation in the uterus. However, it has been reported that hCG alone cannot maintain P₄ production, because it has been shown that the administration of hCG does not prevent CL regression in non‐pregnant women [14]. In addition, in women with missed abortion, P₄ production is maintained despite a low level of hCG, and there is no CL regression or spontaneous miscarriage. These studies suggest that a factor other than hCG plays a role in CL maintenance. However, to date no such factor has been identified [15, 16].

In the 1980s, Wegmann [17, 18] proposed that the maternal immune system contributed to the establishment and maintenance of pregnancy. Only limited evidence supported the hypothesis that T cells promote placental growth and prevent spontaneous abortion [19, 20]. Later, the results from numerous studies began to support the idea that cytokine and growth factor networks at implantation sites play an important role in promoting collaborative relationships among trophoblast, decidual, and immune cells, including natural killer (NK) cells. The establishment of these networks helps in regulating immunotolerance, placentation, and vascularization, all of which are required for the establishment and continuation of pregnancy [21, 22, 23, 24]. Thus, it has been accepted that the immune system is involved in cross‐talk between the mother and embryo.

Using a luteal cell culture system, it was recently shown that treatments with peripheral blood mononuclear cells (PBMCs) from pregnant women increased P₄, interleukin (IL)‐4, and IL‐10 production [25]. The Th2 cytokines, IL‐4 and IL‐10, stimulate P₄ production as much as hCG does. Interestingly, these PBMCs derived from pregnant women were also shown to enhance trophectoderm invasion by a murine embryo [26], and to enhance the invasion of human choriocarcinoma BeWo cells [27]. Furthermore, the intrauterine administration of PBMCs cultured in vitro prior to embryo transfer (ET) was reported to improve the pregnancy rate in patients with repeated failure of IVF‐embryo transfer [28]. It is thought that PBMCs can trigger endometrial development and, possibly, differentiation, these being factors which aid in embryo attachment and implantation in humans [29]. Similar results were found in bovine experiments in which the intrauterine administration of bovine PBMCs prior to ET improved the pregnancy rate [30]. Although the molecular mechanisms associated with PBMC treatments have not been characterized, it is likely that, in addition to the effect of P₄‐stimulated endometrial cytokines, the receptivity of the uterus may be regulated through the appropriate interaction and recruitment of immune cells. These results indicate that PBMCs induce certain functional changes favorable for conceptus implantation in the uterus of recipients.

In humans, hCG is initially produced by day 6–8 blastocysts [31, 32] and later by syncytiotrophoblast cells [33]. It has been shown that hCG directly affects immune cells, such as by the promotion of IL‐8 production by PBMCs [34], and the induction of uterine NK cell proliferation [35]. In addition, it is reported that regulatory T cells (Tregs) are attracted by hCG‐producing trophoblasts and choriocarcinoma JEG3 cells [36]. Furthermore, it has also been suggested that hCG is a factor which contributes to maternal–fetal tolerance by modifying dendritic cells [37]. In ruminants, large amounts of IFNT produced by trophoblast cells are secreted into the uterus during peri‐implantation periods, and IFNT stimulates granulocyte‐macrophage colony‐stimulating factor (GM‐CSF) [38] and receptor transporter protein 4 (RTP4) expression in peripheral bovine lymphocytes (PBLs) [39]. It has been demonstrated that IFNT exhibits an inhibitory effect on lymphocyte proliferation [40, 41], and stimulates NK‐like cell activity [42, 43]. These observations suggest that, similarly to hCG, IFNT could play a role in the activation and recruitment of immune cells [44, 45] (Fig. 2).

Figure 2.

Comparison of maternal recognition of pregnancy in humans and ruminants. It is apparent that implantation processes are similar for conceptus hatching, migration/apposition, attachment, invasion, and placentation, although the latter events are species‐specific, in particular invasive versus non‐invasive placentation. Both human chorionic gonadotrophin (hCG) and IFNT regulate immunological conditions in utero, termed the immunological recognition of pregnancy. Understanding the events that occur prior to hCG or IFNT production will aid in elucidating the molecular mechanisms associated with appropriate pregnancy recognition, resulting in progress in both biochemical and immunological research for achieving successful pregnancy

These findings bring the old question of “do human CG and ruminant IFNT play a role in establishing immune‐environments suitable for pregnancy establishment?” into a new perspective [46]. It is possible that hCG may have stimulated the administered PBMCs in utero, and that IFNT could have stimulated bovine PBMCs, resulting in the establishment of uterine environments suitable for conceptus implantation in both species. Nakayama and coworkers [26] found that PBMCs at the implantation sites were activated by hCG secreted from the embryo, following which PBMCs regulated embryo invasion. In bovine species, the average length of trophoblasts recovered from a PBMC‐treated group was significantly longer than that of the trophoblasts recovered from the control group [47], this finding resulting from uterine environments regulated by P₄ and IFNT. These results point to the notion that, rather than hCG or IFNT stimulation by itself, hCG‐ or IFNT‐stimulated PBMCs condition the uterine endometrium for the establishment of uterine receptivity.

Requirements for a receptive endometrium

Interferon tau targets immune cells as well as endometrial epithelial cells. It is reported that, coincident with conceptus IFNT secretion, large changes in gene expression occur in bovine endometrial epithelial cells, and many of these genes are IFN‐induced genes (ISGs) or immune response genes [48, 49]. Similar changes in gene expressions in ovine endometrial epithelial cells have been found to be under the control of IFNT [50]. Previously, we reported that IFNT induced chemokine ligand 10 [CXCL10, interferon‐inducible protein‐10 kDa (IP‐10)] and CXCL9 [monokine induced by IFNG (Mig)] in ovine and caprine endometrial tissues [44, 45]. Moreover, IFNT is known to induce the expression of several ISGs in the uterus, such as ubiquitin‐like IFN‐stimulated genes (ISGs), 2’,5’‐oligoadenylate synthetase (OAS), signal transducer and activator of transcription (STAT) proteins, IFN regulatory factors (IRFs), major histocompatibility complex class I and β2‐microglobulin, and anti‐virus Mx protein [51, 52, 53, 54, 55, 56, 57]. These genes could mediate the IFNT response and therefore function in antiviral, antiproliferative, and possibly immunosuppressive roles. Further studies are required to elucidate the molecular mechanisms by which these endometrial response genes mediate IFNT functions.

Requirements for functional trophoblast cells

It is known that despite the effects of the maternal immune system, which are often harmful, the development of the conceptus normally progresses. The mechanism by which the trophoblast escapes immunological rejection by the mother remains poorly understood. In addition to IFN expression, trophoblast cells share many features with immune cells, particularly macrophages and T cells. In mice and humans, trophoblast cells exhibit phagocytotic activity, and have the ability to invade deeply into uterine tissues. Ruminant trophoblast cells also migrate into uterine epithelial cells, albeit more shallowly than the trophoblast cells in mice and humans [58]. Indeed, similarities between leukocyte transendothelial migration and embryonic implantation have been suggested before, in spite of these processes being physiologically unrelated [59, 60]. Furthermore, both trophoblast cells and immune cells express cytokines such as granulocyte macrophage‐colony stimulating factor (GM‐CSF), colony stimulating factor‐1 (CSF‐1), interleukins (ILs), and receptors for these cytokines [61, 62, 63, 64, 65, 66]. Recently, we reported changes in GATA1 and GATA2/3, in which GATA2/3 decreased as conceptus attachment started, but GATA1 expression increased, this timing of GATA1 expression coinciding with erythroid development [67].

Although trophoblasts and immune cells are different cell types, they have common properties. In ruminants, such immune cell‐like properties of trophoblasts could be regulated by IFNT. Indeed, the ovine conceptus expresses IFNT receptor (R) coincident with the peak level of IFNT. The evidence suggests that IFNT acts on the conceptus in an autocrine manner [68], and this may result directly in self‐regulation, conceptus development, and protection from the maternal immune environment. As the improvement of pregnancy rates remains elusive for many applications, research into the elucidation of the properties that the trophoblast shares with immune cells, including the presence of IFNs and their receptors, may help in eliminating the obstacles that the trophoblast encounters during the implantation processes.

Regulation of IFNT

Based on cDNA and amino acid sequences, IFNT is classified as a type‐I IFN [4], but it is a unique IFN in many aspects. In addition to showing structural similarities with IFN alpha (IFNA), IFNT shows antiproliferative effects and antiviral activities that have less toxicity than those of IFNA[69, 70, 71, 72, 73]. However, the expression of IFNT is quite different from that of other type‐1 IFNs. IFNA and IFNB are induced by viruses or double‐stranded RNA, and their expressions are maintained for only a few hours. In contrast, IFNT is not induced by viruses or double‐stranded RNA, and its expression is maintained for more than several days [73, 74]. From these data, it is clear that the regulation of IFNT expression is different from those of other IFNs.

Numerous transcription factors thus far found as potential regulators of the IFNT gene are ETS2 [75, 76], activating protein 1 (AP‐1, official symbol JUN) [77], CDX2 [78, 79], homeobox protein distal‐less 3 (DLX3) [80], and GATA2/3 [81, 82]. In addition, co‐activator cyclic adenosine monophosphate (cAMP)‐response element binding protein (CREB)‐binding protein (CREBBP) [83] and p300 [84] are actively associated with the control of IFNT gene expression. In addition, epigenetic regulation also plays an important role in the regulation of IFNT gene transcription. Previously, we reported that increases in the degree of DNA methylation could be one of the mechanisms down‐regulating the ovine IFNT gene during early gestation, resulting in IFNT gene silencing [85]. In addition, we reported that the upstream region of the ovine IFNT gene was in a high histone H3 acetylation state before attachment, and had a low acetylation state following attachment. It was also noted that lowered H3 methylation status could be required for the high IFNT transcription. These epigenetic regulations, particularly the high acetylation, occurring during high IFNT transcription are mediated through changes in CDX2 expression [79, 86]. In spite of this knowledge, the molecular mechanisms that control IFNT gene transcription temporally and spatially have not been definitively identified. Elucidation of the acetylation and methylation mechanisms of histone proteins will help identify the molecular mechanisms by which IFNT gene transcription is controlled.

As mentioned above, trophoblast cells and immune cells share similar properties. It is likely that other immune‐related genes are regulated epigenetically during peri‐implantation periods. Dynamic changes may occur and modulate the trophoblast environment for pregnancy establishment. It is reported that wide‐ranging DNA methylation regulates immune‐related genes including hCG [87]. If this is so, we may be able to use IFNT as a representative immunological factor to evaluate the condition of trophoblast cells through their epigenetic status.

In addition, IFNT‐related transcription factors such as CDX2 [79, 86] and GATA2/3 [81, 82] also play key roles in the development and functioning of mouse trophoblast cells [88, 89], suggesting that groups of genes must be utilized commonly among mammalian trophoblast cells, although the timing and processes associated with these events differ from species to species. If this is true (i.e., groups of genes are utilized commonly among mammalian trophoblast cells), then elucidation of the control mechanism of the IFNT gene may enable us to isolate the phenomenon common to these mammals.

Ruminants as a model for the study of implantation processes

Rodents are used as the primary animal models to study implantation processes. In most of the previous studies, the findings from rodent studies have been the basis for the elucidation of implantation and placentation processes in mammals, including humans. With mice and rats, their small size, ease of purchase and maintenance, and our good understanding of their anatomy, physiology, and genetics are major advantages in experimentation. In mice, however, implantation occurs soon after blastocyst hatching from the zona pellucida, and within a few days several remarkable events take place simultaneously, particularly from implantation to placentation. Therefore, it is very difficult to dissect each of these events and/or the gene expressions associated with these events individually.

In ruminants (bovine, ovine, and caprine species), the blastocyst is formed several days after fertilization, similar to the process in rodents, but placentation takes place on day 21, approximately 2 weeks later than in mice [90]. One of the unique features seen in ruminant conceptus development is trophoblast elongation. The trophoblast elongates exponentially and reaches a length of more than 150–300 mm before the initiation of its attachment to the uterus [91, 92]. When conceptus attachment to the uterine epithelium starts, the trophoblast cells cover the entire uterine lumen, from which placental structures unique to ruminants are developed. Because of these slow processes, ruminants are useful models for studying each of the physiological as well as molecular events within the uterus, particularly biochemical interactions between the trophoblast cells and the uterine endometrium. For these reasons, IFNT could be used as a trophoblast marker to monitor whether the uterine environment is properly established for the progression to placental formation. Ruminant studies may allow us to catch a phenomenon that is overlooked in mice and rats. In fact, it is becoming accepted that farm animal models are important for the field of pregnancy immunology [93].

Conclusion

There is no doubt that IFNT is one of the main factors in conceptus–uterus cross‐talk in ruminants. Over a period of more than 10 years, much progress has been made in IFNT research, but immunological acceptance/co‐existence of the conceptus in the maternal uterus has not been characterized, contrary to expectations. Rather than focusing on immune cells by themselves, a factor responsible for their recruitment and, possibly, their regulation should be considered and incorporated into implantation and placentation research. Similar to CG in human pregnancy, IFNT in ruminants could be one such molecule that is important for pregnancy recognition and immunological conditioning between the conceptus and mother.

Acknowledgments

The authors would like to thank Mr. Robert Moriarty for his critical reading of the manuscript. We also thank the following scientists who contributed to a series of these investigations at early stages: Drs. Kazuyoshi Hashizume, Toru Takahashi, Hitomi Takahashi, and Masashi Takahashi. This work was supported by a grant from the Program for Promotion of Basic Research Activities for Innovative Bioscience (BRAIN) to K.I. This work was also supported by a Grant‐in‐Aid for Scientific Research (23780002) to T.S. H.B. was supported as a Research Fellow of the Japan Society for the Promotion of Science (JSPS).