This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Lea, R. G Articles by Sandra, O. Search for Related Content PubMed PubMed Citation Articles by Lea, R. G Articles by Sandra, O.

Immunoendocrine aspects of endometrial function and implantation

School of Veterinary Medicine and Science, University of Nottingham, Sutton Bonington Campus, College Road, Sutton Bonington, Leicestershire LE12 5RD, UK and ¹ INRA, INRA-UMR ‘Biologie du Développement et de la Reproduction’, Domaine de Vilvert, Bâtiment 231, 78352 Jouy-en-Josas Cedex, France

Correspondence should be addressed to R G Lea; Email: richard.lea{at}nottingham.ac.uk

	Abstract

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
Effective ovarian and uterine function relies on a complex interplay between the endocrine and immune systems. It is generally accepted that in reproductive tissues, oestradiol and progesterone have pro- and anti-inflammatory activities respectively and, in this regard, the paracrine effects of the sex steroids on the ovary are similar to the endocrine effects on the uterus. Ovarian leukocyte recruitment and cytokine release are central to follicle development, ovulation and corpus luteum function. At the uterine level, the cyclical changes in sex steroids regulate the number and distribution of endometrial and decidual immune cells as well as other immune signalling and surveillance factors. The uterine mucosa is unique, in that it must tolerate sperm and the allogeneic blastocyst in a way that does not compromise uterine immune surveillance against bacteria, yeast and viruses. Crosstalk between the sex steroids and immune mediators (systemic and local) are central to these functions, and this article will review these mechanisms and their importance for successful reproductive function and pregnancy success.

	Introduction

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
Implantation and subsequent placental growth are complex processes biologically regulated by interplay between sex steroids and local cell signalling molecules, many of which have an immune function. There is increasing evidence that this interplay is essential for effective ovarian and uterine function, and thus there is now considerable blurring between the respective roles of endocrinology and immunology in reproductive biology. In terms of endocrinology, the priming of the endometrium by the sex steroids and the associated cyclical morphological and functional changes that occur in preparation for pregnancy are well documented. At the local tissue level, however, it is clear that central to these changes are alterations in both inter- and intracell signalling molecules many of which have the capacity to alter immune processes. Moreover, the establishment of pregnancy requires further interaction between endocrine and immune mechanisms, which facilitate maternal–foetal crosstalk, regulate implantation, promote placental growth and prevent immunorejection of the semi-allogeneic foetus. The main focus of this review is the role of immunoendocrine crosstalk in endometrial function and implantation.

	Immunoendocrine crosstalk in reproductive biology: the big picture

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
In the non-pregnant female reproductive system, immunoendocrine crosstalk is essential for normal ovarian function, and this, in turn, regulates the endometrial changes required for possible implantation. Under the influence of follicle-derived oestrogen and pituitary-derived gonadotrophins, follicle development, oocyte maturation and ovulation occur and all of these processes involve inflammatory-like mechanisms (Espey 1994). For example, the inflammatory features of ovulation include extracellular matrix degradation, vascular changes, the expression of chemokines, cell adhesion molecules and integrins and the ultimate recruitment of leukocytes from the circulation (Brannstrom & Enskog 2002).

In the absence of pregnancy, immune cells have been implicated in the process of luteal regression. Neutrophils, macrophages and T-lymphocytes predominate in the corpus luteum (CL) at around the time of luteolysis and may be directly involved in the destruction of the luteal cells and subsequent loss of progesterone secretion (Brannstrom et al. 1994, Pate & Landis Keyes 2001). Alternatively, it has been suggested that these immune cells are simply responding to the dead and dying luteal cells and may have a clearing up function, thus preventing a deleterious immune response to the surrounding ovarian tissue (Pate & Landis Keyes 2001).

Following ovulation, anti-inflammatory mechanisms counteract the original ovulation-associated inflammatory processes. This is reported to be mediated in part by glucocorticoids of adrenal origin that act through the receptor (NR3C1: nuclear receptor subfamily 3, group C, member 1) expressed on ovarian surface epithelial cells (Rae et al. 2004). The pro-inflammatory cytokine interleukin-1 (IL-1), which would be produced during ovulation, increases the expression of the enzyme hydroxysteroid (11-ß) dehydrogenase 1 (HSD11B1) in ovarian surface epithelium, which converts cortisone to anti-inflammatory cortisol (Rae et al. 2004). Consequently, cortisol, rather than cortisone, binds to the NR3C1 and inactivates anti-inflammatory signalling pathways (Rae & Hillier 2005).

More recently, murine cumulus and granulosa cells have been shown to express a range of innate immune-related genes important for cell survival and surveillance during ovulation (Shimada et al. 2006). These include toll-like receptors (TLRs; discussed below) and scavenger receptors. Moreover, Shimada and colleagues demonstrated that cumulus and granulosa cells exhibit phagocytic activity against bacterial particles. These studies illustrate that not only does ovulation exhibit the classic features of an inflammatory reaction, but also that cumulus and granulosa cells appear to exhibit immune surveillance activity in their own right. The implications of toll receptor signalling and immune surveillance are discussed further in relation to uterine epithelial cells (see below).

	The uterus as a unique immune/mucosal organ

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
The uterine mucosa is unique amongst other mucosal sites, since it undergoes regular changes in structure under the influence of sex steroids (Lea & Clark 1989, Quayle 2002). The lower reproductive tract must tolerate commensal flora but be able to respond against infectious agents. In addition, the presence of sperm and the semi-allogeneic blastocyst must be tolerated in a way that does not compromise uterine immune surveillance against bacteria, yeast and viruses (Quayle 2002, Entrican & Wheelhouse 2006).

The presence of seminal plasma in the mouse, pig and human female reproductive tract has been shown to induce local inflammatory changes that are believed to protect against infectious agents introduced during intercourse (Robertson & Seamark 1990, O’Leary et al. 2004, Sharkey et al. 2007). Thus, while sperm presents an immunologic challenge, seminal plasma appears to skew local inflammatory mechanisms in utero for the benefit of maternal reproductive health. Key factors induced by seminal plasma include colony-stimulating factor 2 (CSF2, granulocyte–macrophage-CSF) and IL-6, and this occurs in murine and porcine endometrial tissue that would normally be exposed to semen following coitus (Robertson & Seamark 1990, O’Leary et al. 2004). Recent studies in the human in which semen is deposited in the cervix and vagina indicate that the cervix is highly responsive to seminal plasma in the induction of cytokines (Sharkey et al. 2007). It follows therefore that the establishment of pregnancy relies not only on the influence of ovarian steroid hormones on endometrium but also on seminal fluid-induced changes in specific regions of the female reproductive tract.

The process of implantation, particularly in species with invasive haemochorial placentation, involves increasing contact between semi-allogeneic trophoblast and maternal immune cells either in uterine tissue or in blood. This would seem an immunologically disastrous situation and yet a majority of mammalian pregnancies continue through to term. The mechanisms that underlie these unique properties of the reproductive tract have been the focus of numerous immune and endocrine studies over many decades, and during this period, the separation of these two disciplines has become increasingly difficult.

	Immune surveillance and response to pathogens in the non-pregnant reproductive tract

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
In order to understand how uterine reproductive function is regulated by the sex steroids, it is critical to review the key cell populations of the uterus both prior to implantation and during the implantation process. The first uninterrupted barrier to infectious agents, sperm and trophoblast in utero is the uterine epithelium. Recent studies have shown that this cell population exhibits a number of immune properties placing this cell type in a central immune surveillance role (Wira et al. 2005). Beneath the epithelial layer lies the stroma, which contains a number of immune cells most of which show temporal and spatial changes across the cycle and into pregnancy: an indicator of sex steroid regulation. Proliferative phase endometrial stroma contains 10% leukocytes and this increases to 20% in the proliferative phase and 30% in early pregnancy decidua (Bulmer et al. 1991). The predominant immune cell type is the natural killer cell (NK), which makes up 70% of endometrial/decidual leukocytes. Macrophages comprise 20% of uterine leukocytes and the remaining cells consist of specialised T-cells, polymorpho-nuclear leukocytes and B-cells (Lea & Calder 1997, Yeaman et al. 1997, Critchley et al. 2001, Lea 2001).

Since the epithelial cells occupy the first barrier to the implanting blastocyst and/or infectious agents, they have evolved a number of specialised immunological functions. Epithelial cells throughout the female reproductive tract express receptors that recognise conserved pathogen-associated molecular patterns (PAMPs) present on micro-organisms. These TLRs function as an immune sensory mechanism and comprise a family of pattern recognition receptors that are expressed throughout the reproductive tract (Wira et al. 2005).

Thirteen TLRs have been reported in mammals that between them detect a wide range of pathogenic stimuli with subsequent activation of intracellular pathways and expression of genes with immune function (Albiger et al. 2007). TLRs 1–9 have been identified in human uterine endothelial cells by RT-PCR and TLRs 1–6 by immunohistochemistry (Fazeli et al. 2005, Schaefer et al. 2005). Evidence is now emerging that these receptors are regulated by sex steroids since TLRs 2–6, 9 and 10 are significantly increased during the secretory phase of the cycle (Aflatoonian et al. 2007). Moreover, the expression of TLRs in murine vaginal epithelium is also modulated by the oestrous cycle as well as exposure to a pro-gestogen (Yao et al. 2007). TLRs are also expressed on uterine NK cells indicating that these cells may also respond against PAMPs, particularly at menstruation when the epithelial layer is broken and uNK cells are the predominant endometrial uterine leukocyte (Sentman et al. 2007).

Another key strategy of the reproductive tract to limit infection is through the production of natural antimicrobial peptides, which also modulate host innate immunity. These antimicrobials can be divided into two major families, namely the defensins and whey acidic protein (WAP) containing proteins (King et al. 2003a). Defensins have antibacterial, antiviral and antifungal activity and fall into two structurally different categories: -defensins, found in neutrophils and epithelial sites, and ß-defensins found mainly at epithelial surfaces (Huttner & Bevins 1999). The WAPs, which contain the WAP motif, comprise a family of factors including the protease inhibitors, secretory leukocyte peptidase inhibitor (SLPI) and peptidase inhibitor 3 (PI3, elafin; Bouchard et al. 2006).

Both defensins and WAPs are produced by uterine epithelial cells and uterine leukocyte populations, and expression varies across the menstrual cycle indicative of a degree of sex steroid regulation. In brief, the expression of human defensin, ß 103A (DEFB103A: human ß-defensin-3, HBD3) in human endometrium is highest during the secretory phase of the cycle, whereas HBD-4 is highest during the proliferative phase (King et al. 2003c). Moreover, in this study, the expression of both DEFB103A and HBD4 was altered by exposure to the combined oral contraceptive pill in vitro but not to progesterone alone. Thus, there is some degree of sex steroid regulation of these factors (King et al. 2003c). In addition, SLPI is produced by endometrial glandular epithelial cells at the period of uterine receptivity, and in the human, expression is up-regulated by progesterone (King et al. 2000, 2003b).

In terms of antimicrobial production by endometrial leukocytes, levels will largely mirror changes in cell numbers. Thus, neutrophils infiltrating into endometrium at the time of menstruation provide antimicrobial protection through the expression of -defensins and PI3. Other uterine immune cells are likely to express similar antimicrobial factors.

	The endocrine control of intrauterine inflammatory activity

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
In the non-pregnant reproductive tract, the physiological role of progesterone is to prepare an oestrogen-primed endometrium for implantation. Consequently, there is much interest in the mechanisms by which progesterone promotes uterine receptivity. A key component of this effect is the control of uterine inflammatory activity by sex steroids. Oestrogen has pro-inflammatory effects in the uterus and has been linked to an influx of leukocytes at oestrous in the mouse (De & Wood 1990). In addition, increased levels of pro-inflammatory cytokines in the mouse uterus, particularly Il-1, occur immediately before implantation and have been associated with the pre-implantation surge of oestrogen (De et al. 1993). There is some evidence that this effect is cell type specific, since in the mouse oestradiol is reported to induce pro-inflammatory cytokine production from mast cells but not from macrophages (Hunt et al. 1997). In contrast, progesterone has been associated with anti-inflammatory activity. Indeed, ovariectomised mice treated with oestrogen exhibit an influx of leukocytes into the uterus that is blocked by co-administration of progesterone and this effect is ablated in progesterone receptor (PGR) knockout mice (Tibbetts et al. 1999). These findings illustrate that progesterone acting through its receptor is crucial for its anti-inflammatory role.

Another endocrine mediated anti-inflammatory mechanism is thought to be mediated through the NR3C1 (glucocorticoid receptor), which is expressed in the human endometrium across the menstrual cycle and in first trimester decidua (Henderson et al. 2003, McDonald et al. 2006). Since endometrial HSD11B1 mRNA levels increase at the time of menstruation, cortisol binding to the NR3C1 may induce a similar anti-inflammatory response to that described in the ovary, and this may be an important component of tissue remodelling (McDonald et al. 2006).

In the sheep and other species with less invasive placentas, the importance of controlled intrauterine immune regulation is less clear. Nevertheless, in the ovine uterus, low concentrations of progesterone can block tissue graft rejection (Majewski & Hansen 2002, Padua et al. 2005), and this has been associated with the induction of a uterine immunoregulatory protein belonging to the serpin family of proteinase inhibitors (see later). The accumulation of macrophages and T-cells in the ovine uterus during pregnancy also appears to be partly under sex steroid regulation (Tekin & Hansen 2004).

In the human, data on the role of progesterone in the regulation of uterine immune cells have been obtained from a series of studies looking at the effects of progesterone withdrawal. Physiological progesterone withdrawal, occurring with the demise of the CL at the end of the menstrual cycle, triggers a series of immune-related events, which ultimately leads to menstrual bleeding in primates, or morphological changes consistent with the start of the next cycle in other species. When implantation occurs and progesterone levels are maintained, a different set of local immune events ensures that pregnancy continues.

Sex steroid regulation of chemokines in the uterus
Sex steroids influence the processes of menstruation and implantation largely through the regulation of inflammatory mediators. Coincident with the decline of circulating progesterone prior to menstruation is the up-regulation of specific chemokines, defined by their ability to recruit leukocytes to a specific site in a tissue (homing) and to activate them (Fig. 1). The binding of a chemokine to its receptor on a specific leukocyte up-regulates the expression of adhesion molecules, thus promoting leukocyte adhesion to the endothelium and extravasation and chemotaxis along a concentration gradient (Le et al. 2004). There are over 50 chemokines subdivided into four major families based on the number and position of their cysteine residues (C, CC, CXC and CX3C), where C is the number of cysteine residues in the N-terminal and X is the number of intervening amino acids between the first two cysteines (Le et al. 2004).

View larger version (63K): [in this window] [in a new window]   Figure 1 Modulation of immune activity in the non-pregnant reproductive tract. Ovarian-derived oestradiol (pro-inflammatory) and progesterone (anti-inflammatory) have paracrine actions on the ovary and endocrine actions on the uterus. In the ovary, the sex steroids exhibit paracrine effects on cytokine and prostaglandin production from immune and follicular cells, and these factors are critical for follicle development, ovulation and CL function. Similarly, in the uterine endometrium, the sex steroids exhibit endocrine effects on cytokine and prostaglandin production by epithelial, stromal and resident immune cells. These factors play a regulatory/developmental role in cyclic changes to the endometrium, menstruation, implantation and placental development. Receptors associated with immune surveillance (e.g. toll receptors) are also expressed by ovulatoryovarian follicular cells and the uterine epithelium. Sex steroids influence the ability of epithelial cells to respond to microbes and likely have similar effects at the follicular level.

More recently, chemokines have been implicated in the process of trophoblast adhesion and invasion. Indeed, during the implantation window, the chemokines CX3CL1 and CCL7 are expressed along with their receptors (CX3CR1, CCR1, CCR2, CCR3 and CCR5) in both endometrium and first trimester trophoblast populations (Hannan et al. 2006). During the peri-implantation period, the expression of CXCL9, CXCL10 and CXCL11 increases in the endometrium, whereas their common receptor (CXCR3) is expressed by the trophoblast (Imakawa et al. 2006). The selective expression of chemokines and chemokine receptors on trophoblast indicates that the chemokine receptor network under the influence of progesterone might be critical for the process of implantation. In addition, since the degree of trophoblast invasion and subsequent placental structure shows marked variation between species, there are likely to be species-specific networks of chemokines and chemokine receptors.

Prostaglandin signalling and sex steroids in the uterus
During the implantation window in the human, endometrial prostaglandin levels are elevated and this occurs concurrently with a rise in prostaglandin-endoperoxide synthase 2 (PTGS2: cyclooxygenase 2; Jones et al. 1997). This is thought to be an indirect response to the withdrawal of progesterone and oestradiol. Although this suggests a role of PTGS2 in implantation, studies in mice have led to inconsistent results. While one group reported that PTGS2 knockout mice exhibit poor decidualisation and implantation failure (Lim et al. 1999), another found no effect on implantation but did find that the rate decidualisation proceeds was delayed (Cheng & Stewart 2003). Although these findings indicate that the precise role of the PTGS2 enzymes and prostaglandins in human implantation requires further clarification, in the non-pregnant endometrium PTGS2 enzyme expression and prostaglandin E2 (PGE2) synthesis are associated with vascular function and levels are maximal premenstrually (Jones et al. 1997, Milne et al. 2001, Critchley et al. 2006). Indeed, elevated levels of endometrial PTGS1 and PTGS2 mRNA as well as prostaglandins have been associated with heavy menstrual bleeding (Smith et al. 1981, 2007). In addition, two PGE2 receptor isoforms have been detected in endometrial tissue (E type prostaglandin 2 (EP2) and EP4) and EP4 was expressed at higher levels during the late proliferative stage than in early, late and mid-secretory stage endometrium (Milne et al. 2001). Thus, prostaglandin signalling represents another sex steroid-regulated process important for normal endometrial function.

Interestingly, in domestic species, prostaglandin production by the endometrium is important for implantation and early pregnancy and interferon- has no direct effect on the expression of PTGS2, and thus prostaglandin production (Kim et al. 2003). Table 1 illustrates that PTGS2, necessary for prostaglandin synthesis, is down-regulated by progesterone in human endometrium but up-regulated in ovine endometrium (Charpigny et al. 1997, Critchley et al. 1999). In the human, prostaglandin production has been highlighted as integral to the inflammatory processes important for menstruation. Consequently, production is inhibited when progesterone is maintained during pregnancy. In the sheep, however, the mode of implantation is markedly different from the human, and this is thought to account for the up-regulation in PTGS2, i.e. prostaglandin may be involved in the establishment of the superficial endotheliochorial placenta (Charpigny et al. 1997).

	Immunoendocrine signalling and the maternal recognition of pregnancy

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
The interface between endocrinology and immunology is also manifest in the maternal recognition of pregnancy. This area has been extensively reviewed elsewhere, and therefore, only the key features will be described here (Moffett & Loke 2006, Spencer et al. 2007). In the human and many primates, blastocyst-derived chorionic gonadotrophin binds to receptors in the CL, thus maintaining the secretion of progesterone, and therefore its continued impact on endometrial/decidual and trophoblast function. In cattle and sheep, however, the signalling process is initiated by the release of large quantities of interferon- by the elongating conceptus (Spencer & Bazer 2004). The antiluteolytic activity of interferon- is manifest through the inhibition of oestrogen receptor (ESRI) gene expression in the uterine epithelium. This, in turn, prevents the induction of oxytocin receptor expression (OXTR). Since the binding of oxytocin (OXT) to its receptor induces the luteolytic pulses of PGF2a, these pulses no longer occur and luteolysis is inhibited (Spencer et al. 2007). In these species, we therefore have a situation where ovarian endocrine function is regulated by a maternal-derived prostaglandin, the production of which is indirectly regulated by a conceptus-derived cytokine (Martal et al. 1998). Progesterone also affects the expression of interferon-related genes other than interferon-; however, the biological importance of these changes are uncertain (Jeong et al. 2005, Gray et al. 2006, Catalano et al. 2007; Table 1).

Another interesting variation is the pig, where the secretion of oestrogen from the conceptus appears to be the initiating signal with interferon- release occurring a bit later (La Bonnardiere 1993). Comparison of the various mechanisms that underlie the maternal recognition of pregnancy across species provides a number of examples of both direct and indirect interactions between the endocrine and immune systems (Lee & DeMayo 2004).

	Sex steroid receptors in the non-pregnant reproductive tract

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
Sex steroids induce their effects by acting through nuclear receptors, the expression of which varies across the menstrual cycle. Given the complexity of the endometrium and decidua in terms of the variety of immune and non-haematopoietic cells present, it is important to establish which cell types express the steroid receptors and the downstream consequences of ligand binding or blocking with antigestogens.

Oestrogen and PGRs belong to the nuclear receptor superfamily of the ligand-inducible transcription factors. The binding of the ligand–receptor complex to the hormone response element in the DNA then triggers the transcription of oncogenes and genes encoding growth factors and cytokines. Two subtypes of ESRI, classical ESR1 (ER) and novel ESR2 (ERß), have been identified in the mouse, rat and human (Okada et al. 2005, Saunders 2005). In the uterus, the uterotrophic effects of oestradiol operate through the ESR1 receptor, which has a higher affinity for oestradiol than ESR2. Both ESR1 and ESR2 are expressed in the luminal epithelium and stroma of the immature uterus and, in utero, ESR2 appears to be a negative regulator of ESR1-mediated gene transcription (Weihua et al. 2000). This suggests that the ratio of ESR1 to ESR2 has important functional implications. Oestrogen, acting through the ESR1 receptor, differentially regulates the expression of the PGR (Weihua et al. 2000).

There are two isoforms of the PGR (PGRA and PGRB) both of which are expressed in the reproductive tract in different ratios, an observation that is believed to relate to progesterone action on specific tissues (Wang et al. 1998). In the human, PGRA/PGRB show cyclical variation in the glandular epithelium with a decline from the proliferative to the secretory phase (Mylonas et al. 2007). In this study, PGRB exhibited some variation in the stroma, as did both ESRs; however, the cell types were not identified.

A relatively recent and controversial development relevant to the progesterone regulation of reproductive function has been the discovery of membrane progesterone receptors (Losel & Wehling 2003). These receptors are thought to account for the non-genomic actions of progesterone, i.e. those that do not influence gene expression but drive more rapid effects such as activation of signalling cascades and inhibition of transmembrane Ca²⁺ entry. In vertebrate species, three membrane progestin receptors have been described (mPGR, ß and ) with mPGR predominant in reproductive tissues (Zhu et al. 2003). Endometrial mPGR is up-regulated during the luteal phase and down-regulated in the myometrium with progesterone withdrawal (Fernandes et al. 2005). Since mPGRs are present in ovary as well as the uterus, they may also influence local immune mediators at this site. In support of this possibility, membrane progesterone receptors have been implicated in T-cell immunosuppression during pregnancy (Ehring et al. 1998).

In terms of stromal immune cell populations, ESR2 has been co-localised to CD68-positive macrophages, and isolated endometrial macrophages have been shown to express both oestrogen and progesterone receptors (Khan et al. 2005). The combined effect of oestrogen and progesterone therefore likely explains the increase in macrophage numbers during the secretory phase peaking prior to menstruation (Milne et al. 2005). Moreover, the mechanism underlying this effect likely involves the chemokine–chemokine receptor network described earlier.

	The role of uterine NK cells

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
Since their characterisation in 1989, uterine NK cells have been a major focus of attention in reproductive immunology (King et al. 1989). They predominate in late secretory phase endometrium and early pregnancy decidua and are phenotypically distinct from circulating NK cells. For example, they have a high level of CD56 (CD56 bright), lack the CD16 cell surface marker (low affinity receptor for IgG complexes) and a range of other cell surface markers including CD57. They also express a more restricted array of chemokine receptors than circulating NK cells (Dosiou & Giudice 2005, Wira et al. 2005). Changes in uNK cell numbers occur across the menstrual cycle, and during the mid to late secretory phase, the CD56+, CD16– cells are localised to leukocyte aggregates close to endometrial glands and spiral arteries. In the absence of an implanting blastocyst, there is evidence for their demise through apoptosis and when implantation occurs the resident CD56+ cells continue to proliferate as the decidual reaction continues (Trundley & Moffett 2004).

Since 10% of circulating NK cells are CD56 bright, it is thought that the uNK cells may be derived from this population. However, there are still significant differences between these two populations in terms of phenotype and the array of genes expressed (Moffett-King 2002, Koopman et al. 2003). Although the uterine NK cells are thought to originate from circulating NK cells, the variation in numbers has been attributed to proliferation in utero (King & Loke 1991, Moffett-King 2002). In terms of function, uterine NK cells have been associated with menstruation, decidualisation and implantation. With respect to the latter, uNK cells are thought to be vital in regulating the maternal immune response to the semi-allogeneic trophoblast. Indeed, a majority of uNK cells express one or more inhibitory NK cell receptors (killer inhibitory receptors) such as KLRD1/KLRC1 (killer cell lectin-like receptor subfamily D, member 1/killer cell lectin-like receptor subfamily C, member 1; Ponte et al. 1999). These receptors are thought to interact with the unique repertoire of non-classical major histocompatibility complex antigens expressed by invading trophoblast (HLA-G, HLA-E and HLA-F; Moffett & Loke 2006). Since aggregates of decidual uNK cells have been localised close to spiral arterioles, the cells have also been implicated in vascular remodelling. In addition, uNK cells can make a variety of cytokines implicated in the regulation of trophoblast invasion and the regulation of endometrial/decidual angiogenesis (Jokhi et al. 1997, Moffett & Loke 2006).

Recent studies have shown that uNK cells express ESR2 but not ESR1 or the progesterone receptors (Henderson et al. 2003). This suggests that there may be some degree of oestrogen regulation. In the absence of progesterone receptors, progesterone may indirectly affect NK cell activity through the induction of cytokines from neighbouring cells. Uterine NK cells also express the NR3C1, which is co-localised with HSD11B1 (McDonald et al. 2006). This suggests that cortisol may have an anti-inflammatory effect on this cell population. Another possibility is that progesterone may act through NR3C1 (glucocorticoid receptor) or may affect NK cells via a non-receptor mechanism (Miyaura & Iwata 2002, Dosiou & Giudice 2005).

	Progesterone as an immunomodulator

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
Progesterone has been associated with immunomodulatory function for many years and has been described as nature’s immunosuppressant (Siiteri et al. 1977). Studies in sheep and hamsters illustrate that progesterone at concentrations found at the maternal–foetal interface can inhibit tissue graft rejection in utero (Moriyama & Sugawa 1972, Majewski & Hansen 2002). In terms of local effects of progesterone on endometrial/decidual uNK cells, the majority of studies suggest that the immunomodulatory activity may be indirect.

Sex steroids may induce NK cells to produce immunomodulatory proteins or may alter the cytokine environment, which then alters uNK function. This may occur through the action of progesterone on T-cell differentiation into Th1 and Th2 cells. Cytokines produced by Th1 cells are pro-inflammatory (e.g. interferon (IFN)-, IL-2), whereas Th2 cell-derived cytokines tend to be anti-inflammatory and associated with humoral immunity (e.g. IL-4, IL-5, IL-6, IL-10). The two T-cell subtypes are mutually inhibitory (Lea 2001). Successful pregnancy has been associated with the predominance of Th2-like cytokines at the maternal–foetal interface and a down-regulation of Th1 cytokines (Lin et al. 1993). Although this is now generally considered as an over-simplification, the paradigm continues to prove useful in terms of steroid regulation of cytokine production (Chaouat et al. 2004). Since progesterone can induce T-cell differentiation towards the Th2 pathway, it has been associated with the induction of such a cytokine profile in utero (Piccinni 2006). However, since non-haematopoietic cells such as trophoblast and epithelial cells can also produce cytokines, it is unclear how such a Th2 bias may be induced (Lea 2001, Hauguel-de Mouzon & Guerre-Millo 2006). Progesterone may act on progesterone receptor-positive cells such as macrophages and glandular epithelial cells or may affect other cell types via the routes discussed above.

	Putative progesterone target genes with immune function: an overview of transcriptomic and candidate gene studies

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
Historically, genes sensitive to endocrine disturbance have largely been identified by candidate gene studies. In more recent years, a complementary approach has been the use of transcriptomics where large numbers of genes are screened simultaneously using broad range or targeted gene arrays. In terms of the sex steroids, both empirical approaches have proved useful in determining the impact of tissue exposure or withdrawal. The use of gene array with no corroborative experiments to investigate gene changes is currently the subject of much debate, and indeed, reported gene changes using transcriptomics alone must be treated with an element of caution (Rockett & Hellmann 2004). Both technical and computational concerns have been raised and many journals are reviewing policy on including experiments to corroborate changes observed by gene array (Rockett & Hellmann 2004). For this reason, some of the genes listed in Table 1 are ‘putative’ rather than ‘definitive’ in terms of their sensitivity to progesterone and some of these genes and/or groups are discussed below.

Table 1 summarises progesterone-sensitive genes associated with immunity, and genes have been divided into groups based on immune function or presence of immune cell surface markers. Overall, Table 1 clearly shows the considerable impact that progesterone has on genes associated with chemokines and their receptors, interferons, cytokines and growth factors, prostaglandins, lipoxins, intracellular signalling molecules, immune modulators, specific immune cell populations and proteinase inhibitors. The chemokines and prostaglandins have already been discussed, and thus the following overview is focussed on some of the other genes in Table 1, where the weight of evidence supports progesterone regulation.

Cytokines and growth factors
Cytokines, growth factors and their receptors represent a broad group of factors with a range of functions. Since progesterone withdrawal or persistence influences inflammatory processes associated with menstruation and implantation, the genes identified through a screening approach will depend on the type and stage of uterine tissues examined. Seven of the fifteen cytokine/growth factor genes listed have been identified by candidate gene studies. CSF1 and leukaemia inhibitory factor (LIF) have both been associated with implantation success and have been the focus of many reviews (Stewart et al. 1992, Pollard 1997). In brief, CSF1 (a haematopoietic cytokine) and its receptor (CSF1R) have been localised to the reproductive tract in the mouse, human, pig and cow. Indeed, the progesterone induced expression of CSF1 by uterine epithelial cells and the subsequent regulation of CSF1R-positive endometrial macrophages has been highlighted as one key endocrine–immune axis associated with implantation success (Pollard 1997, Lee et al. 2003). Another key progesterone-sensitive cytokine important for implantation is LIF. Studies carried out on isolated human endometrial T-cells indicate that LIF is positively regulated by progesterone (Piccinni et al. 2001). In the mouse, however, uterine LIF appears to be primarily regulated by oestrogen (Bhatt et al. 1991, Chen et al. 2000, Kimber 2005). Although progesterone had no effect in this study, the use of whole uterine mRNA may not have allowed for effects on individual immune cell populations.

Decidua and peripheral blood T-cells exposed to progesterone secrete IL-4 and IL-5 (Th2 cytokines), and this supports the paradigm that progesterone induces a local Th2 bias in utero (Piccinni 2006). In two separate gene array studies, IL-13 receptor 2 has been reported to be down-regulated by progesterone in human endometrium (Catalano et al. 2003, Okada et al. 2003). Since IL-13 is expressed by human endometrium and since IL-13 receptor 2 (IL13Ra2) has been implicated as a negative regulator of IL-13, progesterone inhibition of the receptor may allow endometrial IL-13 to promote anti-inflammatory activity important for endometrial function and/or implantation (Chegini et al. 2002, Chiaramonte et al. 2003). Interestingly, two separate array studies in the mouse have shown the opposite effect, i.e. progesterone up-regulates Il13ra2 (Cheon et al. 2002, Jeong et al. 2005). The reason for this between-species difference is uncertain but does illustrate the complex nature of intrauterine cytokine profiles and the need for further corroborative studies. The cytokine IL-15 is also positively regulated by progesterone and is thought to be important in the stimulation of uNK cells in late secretory endometrium and first trimester decidua (Kitaya et al. 2000). Although the other genes in the cytokines/growth factor category of Table 1 relate to individual array studies, some of the changes reported further support the Th2/anti-inflammatory paradigm associated with progesterone action, e.g. down-regulation of TNF receptors and up-regulation of TGFB1 (Catalano et al. 2007).

Immune responsive gene 1
Under the category of ‘intracellular signalling factors’ in Table 1, immune responsive gene 1 was down-regulated by array analysis and subsequent studies have shown it to be expressed in mouse luminal epithelium during the implantation window (Chen et al. 2003, Cheon et al. 2003). In the mouse, blocking Irg1 with antisense oligonucleotides is reported to reduce the number of implanting embryos by 80% (Cheon et al. 2003). Since inflammatory activity is important during the implantation window, it is possible that this may be an example of progesterone acting in a pro-inflammatory manner but to the benefit of the implanting embryo. Cheon and colleagues also point out that Irg1 has a glycosaminoglycan attachment site important for proteoglycan binding, and that Irg1 may be sequestered to cell surface proteoglycans to be later released as an active factor following proteolysis.

Immunomodulatory factors
In terms of progesterone-sensitive factors implicated as key immunomodulators during pregnancy (Table 1), LGALS15 (galectin 15), SPP1 (secreted phosphoprotein 1, osteopontin), uterine milk protein precursor A (UTMP, uterine serpin), histidine decarboxylase and C13orf24 (progesterone-induced blocking factor, PIBF) have been highlighted by a range of experimental approaches. The following is a brief discussion of each of these factors:

LGALS15 (galectin 15)
LGALS15 has been identified as a key ovine uterine factor, produced by the endometrial luminal epithelium and ductal glandular epithelium of the uterus (Gray et al. 2005). LGALS15 is thought to be important in the regulation of implantation and placentation. Indeed, in sheep in which uterine gland development is inhibited by neonatal progestin exposure (the uterine gland knockout phenotype), the failure of sheep blastocysts to elongate or survive past 12 to 14 days has been associated with a reduction in LGALS15 (Gray et al. 2000, 2001, 2006). In terms of immune function, other galectin family members are known to modulate innate and adaptive immune responses as well as immune cell activation and differentiation (Rabinovich et al. 2002, Young & Meeusen 2004). Thus, it is possible that LGALS15 may have an immunomodulatory role in utero.

SPP1 (secreted phosphoprotein 1, osteopontin)
SPP1 is a complex multifunctional glycoprotein found in uterine histotroph that is classified both as a cytokine and as a component of extracellular matrix (Johnson et al. 2003). During implantation and early pregnancy, SPP1 is produced by uterine glandular epithelial cells and immune cells present in the endometrium and placenta. A number of studies have demonstrated that epithelial cell-derived SPP1 is progesterone dependent (Johnson et al. 2003). In contrast, immune cell-derived SPP1 does not appear to be progesterone regulated (Johnson et al. 2000). Nevertheless, since SPP1 from either source can influence the activity of many other immune cells, including those present at the maternal–foetal interface, it represents a key link between progesterone and the activity of intrauterine immune cells (Johnson et al. 2003).

UTMP, uterine serpin
UTMP is both a proteinase inhibitor and a progesterone-sensitive immunomodulatory factor (Table 1). UTMP is produced from uterine glandular epithelial cells and is a member of the serpin superfamily of proteinase inhibitors that exhibit weak antiproteinase activity but inhibit a wide range of lymphocyte functions (Ing et al. 1989, Peltier & Hansen 2001). Consequently, UTMP is thought to be important in the inhibition of maternal immune responses directed against antigens expressed on the foeto-placental unit (Hansen 2007). Although UTMP has also been detected in cattle, pigs and goats, it has been most characterised in the sheep and shown to exhibit anti-proliferative activity on the pre-implantation embryo and on tumour cell lines (Tekin et al. 2005). This implies that UTMP may not only be immunoregulatory but may also regulate cell proliferation possibly including trophoblast.

Histidine decarboxylase
Histidine decarboxylase is a progesterone-sensitive pro-inflammatory mediator identified by the candidate gene approach. However, in the uterus, this factor is produced by the uterine epithelium rather than resident mast cells, and thus histamine may have non-immunological activities at the time of implantation (Paria et al. 1998,Wood et al. 2000).

Progesterone-induced blocking factor (C13orf24)
During pregnancy in humans, circulating T-lymphocytes express progesterone receptors (Szekeres-Bartho et al. 1999, Barakonyi et al. 2002). On binding to the lymphoid receptors on the immune cells, progesterone induces the secretion of a highly immunosuppressive 34 kDa molecule designated as C13orf24 (PIBF; Szekeres-Bartho et al. 1989). Clinical interest in C13orf24 during human pregnancy arose from studies correlating urinary C13orf24 with pregnancy outcome. Pre-eclamptic women with at least two symptoms (hypertension and toxaemia) were reported to have urinary C13orf24 levels lower than those found in normal healthy pregnancies (Polgar et al. 2004). In addition, low C13orf24 levels were associated with pre-term delivery and intrauterine growth restriction (Polgar et al. 2004).

The biological significance of C13orf24 secretion during pregnancy was originally established from studies in mice. In this species, the neutralisation of endogenous C13orf24 using a PIBF-specific antibody results in a 70% reduction in the number of viable foetuses per mouse (Szekeres-Bartho et al. 1997a). Furthermore, the treatment of pregnant mice with the progesterone receptor antagonist RU486 leads to an increase in resorption rate and this is associated with the abrogation of C13orf24 production by spleen cells. Since the increase in resorption rate can be reversed by simultaneous administration of C13orf24, this unique factor appears to have an important anti-abortive function in pregnant mice (Szekeres-Bartho et al. 1990). The anti-abortive mechanism of C13orf24 is thought to be primarily mediated through its inhibitory effects on NK activity. In support of this, high resorption rates in mice due to elevated NK activity can be corrected by administration of C13orf24 (Szekeres-Bartho et al. 1990). Conversely, increased resorption induced by anti-PIBF is associated with increased splenic NK activity and the former is reversible with anti-NK antibodies (Szekeres-Bartho et al. 1997b).

In the human, C13orf24 expression by peripheral pregnancy lymphocytes is inversely related to NK activity (Szekeres-Bartho et al. 1995, 1996). The inhibitory effects of C13orf24 on NK cells are thought to be mediated, in part, by altered arachidonic acid release and subsequent downstream effects on prostaglandin and leukotriene production and hence pro-inflammatory activity (Szekeres-Bartho et al. 2005).

C13orf24 also skews cytokine production towards Th2-like anti-inflammatory activity and, consequently, it may mediate the effects of progesterone on cytokine profiles. For example, the pro-NK cytokine IL-12 is reduced and this is thought to favour low NK activity conducive to pregnancy success (Szekeres-Bartho et al. 2005). Indeed, in mice, in vivo and in vitro studies have shown that PIBF favours the secretion of IL-3, IL-4 and IL-10, while suppressing pro-inflammatory cytokines such as IL-12 and IFN- (Szekeres-Bartho & Wegmann 1996, Szekeres-Bartho et al. 1996). C13orf24 is also reported to affect humoral immunity by increasing the production of asymmetric antibodies thought to have a blocking function (Kelemen et al. 1996).

In the sheep reproductive tract, we have identified an orthologue of C13orf24 and localised C13orf24 mRNA to endometrium and protein to placental trophoblast (Lea et al. 2005). Interestingly, monitoring C13orf24 RNA levels (exons 8–14) during the ovine oestrous cycle and early pregnancy revealed relatively steady-state levels that did not correlate with physiological changes in progesterone (Lea et al. 2005, Sandra et al. 2005). Although it is recognised that there may be some species differences in C13orf24 expression and function, these recent data suggest that the regulation of C13orf24 expression may only be partially progesterone dependent even during a mammalian pregnancy.

	Conclusion

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
Cyclical changes in the sex steroids influence a variety of genes in the ovary and uterus, which act to protect these tissues against pathogens, while simultaneously preparing them for ovulation, menstruation or implantation. During the period of increased uterine receptivity, epithelial cells exhibit increased toll receptor expression and altered production of specific antimicrobial peptides (e.g. SLPI), thus enhancing the ability to both detect and respond against PAMPs on micro-organisms. During menstruation, toll receptor expression by uterine NK cells is thought to ensure effective immune surveillance when the epithelial layer is broken. Ovulation, menstruation and implantation are largely characterised by changes in the type and distribution of both immune and non-immune cells in the ovary and/or uterus. With respect to immune cells, the sex steroids regulate the expression of a wide range of chemokines and their receptors that act collectively in the selective recruit of leukocytes. This mechanism likely accounts for the increased numbers of NK cells and macrophages at the implantation site in human secretory phase endometrium. The expression of chemokine receptors on trophoblast suggests that the chemokine–chemokine receptor network may also regulate implantation. ESR2 expression by NK cells and macrophages indicates that both cell types are responsive to oestrogen. In addition, progesterone acting on the PGRA/PGRB in glandular epithelium provides a means for progesterone to influence NK cells and macrophages indirectly. This may occur through the progesterone-induced expression of a range of endometrial and epithelial cytokines and growth factors. For example, progesterone-induced CSF1 is likely to target CSF1R-positive endometrial macrophages. Other progesterone-induced cytokines may contribute to the anti-inflammatory (Th2-like) environment that characterises early pregnancy. In the sheep, LGALS15, SPP1 and UTMP have been identified as key intrauterine immunomodulatory factors. In both mice and humans, C13orf24 (PIBF) is produced systemically by PGR-positive / T-cells and affects different aspects of immune function including reduced NK activity and the preferential production of anti-inflammatory Th2-like cytokines. Although the sex steroids drive the inflammatory processes associated with ovulation and menstruation, the glucocorticoids appear to play a role in the rapid resolution of inflammation, thus allowing tissue remodelling to occur. This involves the expression of ovarian and endometrial HSD11B1, which converts cortisone to cortisol. The co-expression of HSD11B1 with NR3C1 (glucocorticoid receptor) in the ovary and uterus ensures that cortisol binds to the receptor, thus inducing anti-inflammatory activity. Key areas for future investigation should include further work on the regulation and function of recently identified progesterone target genes and the manipulation of the sex steroid/glucocorticoid-regulated immunomodulatory pathways described for improved reproductive management in humans and animals.

	Acknowledgements

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 
OS is a member of the European Network of Excellence entitled ‘EMBIC’ (EMBryo Implantation Control), supported by the European Union (contract no. 512040). We are grateful to Prof. Julia Szekeres-Bartho for reading and commenting on the review and to the referees for their constructive and helpful comments. The authors declare that there is no conflict of interest that would prejudice the impartiality of this scientific work.

	Footnotes

 
Received 13 April 2007
First decision 16 May 2007
Revised manuscript received 17 July 2007
Accepted 17 July 2007

	References

Top Abstract Introduction Immunoendocrine crosstalk in... The uterus as a... Immune surveillance and response... The endocrine control of... Immunoendocrine signalling and... Sex steroid receptors in... The role of uterine... Progesterone as an... Putative progesterone target... Conclusion Acknowledgements References

 

Aflatoonian R, Tuckerman E, Elliott SL, Bruce C, Aflatoonian A, Li TC & Fazeli A 2007 Menstrual cycle-dependent changes of Toll-like receptors in endometrium. Human Reproduction 22 586–593.[Abstract/Free Full Text]

Albiger B, Dahlberg S, Henriques-Normark B & Normark S 2007 Role of the innate immune system in host defence against bacterial infections: focus on the Toll-like receptors. Journal of Internal Medicine 261 511–528.[CrossRef][ISI][Medline]

Asselin E, Johnson GA, Spencer TE & Bazer FW 2001 Monocyte chemotactic protein-1 and -2 messenger ribonucleic acids in the ovine uterus: regulation by pregnancy, progesterone, and interferon-. Biology of Reproduction 64 992–1000.[Abstract/Free Full Text]

Barakonyi A, Kovacs KT, Miko E, Szereday L, Varga P & Szekeres-Bartho J 2002 Recognition of nonclassical HLA class I antigens by gamma delta T cells during pregnancy. Journal of Immunology 168 2683–2688.[Abstract/Free Full Text]

Bhatt H, Brunet LJ & Stewart CL 1991 Uterine expression of leukemia inhibitory factor coincides with the onset of blastocyst implantation. PNAS 88 11408–11412.[Abstract/Free Full Text]

La Bonnardiere C 1993 Nature and possible functions of interferons secreted by the preimplantation pig blastocyst. Journal of Reproduction and Fertility 48 157–170.

Bouchard D, Morisset D, Bourbonnais Y & Tremblay GM 2006 Proteins with whey-acidic-protein motifs and cancer. Lancet Oncology 7 167–174.[CrossRef][ISI][Medline]

Brannstrom M & Enskog A 2002 Leukocyte networks and ovulation. Journal of Reproductive Immunology 57 47–60.[CrossRef][ISI][Medline]

Brannstrom M, Pascoe V, Norman RJ & McClure N 1994 Localization of leukocyte subsets in the follicle wall and in the corpus luteum throughout the human menstrual cycle. Fertility and Sterility 61 488–495.[ISI][Medline]

Bulmer JN, Morrison L, Longfellow M, Ritson A & Pace D 1991 Granulated lymphocytes in human endometrium: histochemical and immunohistochemical studies. Human Reproduction 6 791–798.[Abstract/Free Full Text]

Catalano RD, Yanaihara A, Evans AL, Rocha D, Prentice A, Saidi S, Print CG, Charnock-Jones DS, Sharkey AM & Smith SK 2003 The effect of RU486 on the gene expression profile in an endometrial explant model. Molecular Human Reproduction 9 465–473.[Abstract/Free Full Text]

Catalano RD, Critchley HO, Heikinheimo O, Baird DT, Hapangama D, Sherwin JR, Charnock-Jones DS, Smith SK & Sharkey AM 2007 Mifepristone induced progesterone withdrawal reveals novel regulatory pathways in human endometrium. Molecular Human Reproduction [in press].

Chaouat G, Ledee-Bataille N, Dubanchet S, Zourbas S, Sandra O & Martal J 2004 TH1/TH2 paradigm in pregnancy: paradigm lost? Cytokines in pregnancy/early abortion: reexamining the TH1/TH2 paradigm International Archives of Allergy and Immunology 134 93–119.[CrossRef][ISI][Medline]

Charpigny G, Reinaud P, Tamby JP, Creminon C, Martal J, Maclouf J & Guillomot M 1997 Expression of cyclooxygenase-1 and -2 in ovine endometrium during the estrous cycle and early pregnancy. Endocrinology 138 2163–2171.