Introduction

Signalling by accessory gland fluids to influence female reproductive physiology and behaviour is common in the insect kingdom whenever egg fertilisation involves the intromission and deposition of male fluids into the female (Gillott 2003). Emerging knowledge indicates that a similar biological principle may operate in eutherian mammals. Conventionally, seminal plasma has been viewed as a transport and survival medium for mammalian sperm but its role is now recognised to extend beyond this to targeting female tissues. Recent studies in rodents, livestock species and humans show that the introduction of semen into the female tract orchestrates striking molecular and cellular changes that facilitate conception and pregnancy. Seminal plasma contains oestrogen and testosterone, several prostaglandins and glycoprotein signalling substances, including several cytokines and growth factors (Aumuller and Riva 1992; Maegawa et al. 2002; Mann 1964). These molecules bind to cognate receptors on target cells in the female reproductive tract, activating changes in gene expression leading to modifications in the cellular composition, structure and function of local tissues and of tissues distal to the tract, including the ovary, spleen and peripheral lymphoid organs and potentially even the central nervous system.

This review will describe our current understanding of the molecular basis of the interaction between seminal plasma and female reproductive tissues, including the identity of active constituents in seminal plasma, and will examine the physiological consequences of this response in terms of female reproductive function and pregnancy success. The evolutionary significance of the response to seminal plasma in male–female sexual conflict and the significance of this for human reproduction will be discussed.

Post-mating inflammatory cascade in mammals

The most immediate and clearly evident effect of insemination in mammals is a rapid and dramatic influx of inflammatory cells into the site of semen deposition (Fig. 1). This “post-mating inflammatory response” has been described most thoroughly in mice (De et al. 1991; McMaster et al. 1992; Robertson et al. 1996). The cellular changes are initiated when seminal plasma moieties interact with cervical and uterine epithelial cells to induce a surge in synthesis of cytokines, including granulocyte-macrophage colony-stimulating factor (GM-CSF), interleukin (IL)-6 and an array of chemokines (Robertson et al. 1996, 1998). These pro-inflammatory factors stimulate the extravasation and infiltration of subepithelial stromal tissues by macrophages, dendritic cells and granulocytes. Seminal plasma, as opposed to sperm, mediates the response, since vasectomised mice, but not mice from which seminal vesicles have been surgically removed, elicit the response (Robertson et al. 1996). Similar effects are seen in pigs (Lovell and Getty 1968) in which seminal plasma induces the uterine expression of GM-CSF, IL-6 and MCP-1, leading to macrophage and dendritic cell recruitment into the endometrial stroma (O’Leary et al. 2004).

Fig. 1
figure 1

Photomicrographs illustrating the post-mating inflammatory response in mice (Ep epithelium, En endometrium, Gl uterine glands). Sections of uterine tissue from wild-type mice recovered at oestrus (est; a, d), 24 h after mating (dpc 1; b, e) or 4 days after mating (dpc 4; c, f) with intact males were stained immunohistochemically with antibodies reactive for CD45(LCA) (leukocyte common antigen) to detect all leukocytes (a–c) or with antibodies reactive for MHC class II to detect macrophages and dendritic cells (d–f). Arrows in b indicate neutrophils migrating through the luminal epithelium. ×40

Full size image

In humans, intercourse elicits neutrophil recruitment into the superficial epithelium of the cervical tissues (Pandya and Cohen 1985), with a striking recruitment of macrophages, dendritic cells and lymphocytes into the epithelial layers and deeper stromal tissues (Robertson et al. 2001a). Leukocyte influx requires contact between seminal fluid and the female tract tissues, since no inflammatory response is seen following condom-protected intercourse. The regulation of the cervical leukocytic infiltrate occurs by activation of pro-inflammatory cytokines GM-CSF, IL-6 and IL-8 (Sharkey et al. 2004). In vitro studies suggest that, despite the primary site of semen deposition being the cervix in women, the effects of seminal plasma probably extend to the uterus (Gutsche et al. 2003; Tremellen and Robertson 1997). In vivo observations show active seminal constituents are bound to the post-acrosomal region of the sperm head and are carried together with sperm into the higher tract (Chu et al. 1996), served by rapid and sustained uterine peristaltic contractions that transport macromolecular material as high as the Fallopian tubes (Kunz and Leyendecker 2002). Unique vascular connections further facilitate the transport of progesterone and other mediators from the cervix to the endometrial tissue (Bulletti et al. 1997) and might also facilitate the transport of inflammatory mediators.

Physiological significance of seminal plasma for pregnancy outcome

The practice of artificial insemination shows that seminal plasma is not mandatory for the initiation of pregnancy but evidence has been found that the success and quality of the pregnancy, particularly as measured by the growth trajectory of the fetus, are compromised if females are not exposed to seminal plasma. The most compelling data comes from rodents. Experiments in which the seminal vesicle, prostate or coagulating glands are surgically removed from mice, rats and hamsters prior to mating each show that seminal vesicle fluid is the most vital non-sperm component of the ejaculate and support the interpretation that seminal plasma is essential for optimal sperm survival and fertilisation potential (O et al. 1988; Pang et al. 1979; Peitz and Olds Clarke 1986; Queen et al. 1981). In mice, embryo transfer protocols generally employ recipients exposed to seminal plasma by mating to vasectomised males but fetal loss and abnormality is considerably greater when embryos are transferred after pseudopregnancy is achieved without exposure to male fluids (Watson et al. 1983). Confirmation that the effects of seminal plasma are mediated at least partly through promoting receptivity in the female tract comes from studies showing that, when recipient females are mated with seminal-vesicle-deficient males, transferred embryos give rise to fetuses with retarded growth trajectories and placental development (Bromfield et al. 2004). In rats, implantation rates and fetal growth are similarly impaired unless females are inseminated prior to embryo transfer (Carp et al. 1984). In pigs, artificial insemination with diluted semen reduces litter sizes but additional mating with a vasectomised male or administration of heat-killed semen restores litter size and improves farrowing rate (Mah et al. 1985; Murray et al. 1983).

Clinical studies in humans have revealed both acute and cumulative benefits of exposure to seminal constituents and implicate a partner-specific route of action. Live birth rates in couples undergoing IVF treatments are significantly improved when women engage in intercourse around the time of embryo transfer (Bellinge et al. 1986; Tremellen et al. 2000). Treatment of women suffering from recurrent spontaneous abortion with seminal plasma pessaries is reported to improve pregnancy success (Coulam and Stern 1993). In the clinical condition of preeclampsia resulting from compromised placental function, there is a cumulative benefit of exposure to semen over time, with limited sexual experience or use of barrier methods of contraception being linked with increased risk (Klonoff-Cohen et al. 1989; Robillard et al. 1995). Evidence from women who have changed their male partner suggest the effect is partner-specific (Dekker et al. 1998). Markedly increased rates of preeclampsia are also evident in pregnancies initiated by donor oocytes or semen (Salha et al. 1999), when prior exposure to the donor sperm or conceptus antigens has not occurred.

Transforming growth factor-β and other active factors in semen

Experiments in mice from which accessory glands have been surgically removed have shown that the active inflammation-inducing moieties in semen are derived from the seminal vesicle, in which the majority of seminal fluid is produced (Robertson et al. 1996). Protein chromatographic techniques and neutralising antibodies have identified transforming growth factor-β (TGFβ) as the principal trigger for the induction of uterine inflammatory responses following mating in mice (Tremellen et al. 1998). The TGFβ1 content of murine seminal vesicle fluid is approximately 70 ng/ml (Robertson et al. 2002), which is at least five-fold the content of serum and similar to that of colostrum, which is the most potent known biological source of TGFβ. Seminal vesicle TGFβ synthesis is testosterone-dependent, with a severe reduction evident after castration and partial recovery after administration of exogenous testosterone (Robertson et al. 2002).

TGFβ is distinguished amongst cytokines in that it is secreted from the cell in a precursor dimer form, termed latent TGFβ. The majority of the TGFβ present in male seminal fluids is in the latent form and appears to be activated in the female reproductive tract after insemination (Fig. 2). Plasmin and other enzymes that activate latent TGFβ are present in female reproductive tissues (Casslen and Harper 1991). Mechanisms involving activation through an alteration in the conformation of latent TGFβ might also operate in the female tract. The uterine luminal epithelium expresses levels of αvβ6 integrin unparalleled in other epithelia (Breuss et al. 1993). Similarly, thrombospondin-1 is expressed during early pregnancy on the apical surface of uterine epithelial cells, being maximal at the time of insemination and declining thereafter (Slater and Murphy 1999). Both αvβ6 integrins and thrombospondin-1 interact with the latency-associated peptide of TGFβ to cause a conformational change that exposes TGFβ-receptor-binding sites.

Fig. 2
figure 2

Schematic diagram illustrating the transmission of male signals in seminal plasma to the female reproductive tract. Transforming growth factor-β (TGFβ) is the major active moiety in seminal plasma. TGFβ is synthesised in the seminal vesicle gland and is present in seminal vesicle secretions at concentrations approximating 70 ng/ml, principally in the inactive latent form. At ejaculation, TGFβ is diluted in the ejaculate to approximately 30 ng/ml and is then deposited in the female reproductive tract where it is further diluted in female uterine fluid to approximately 6 ng/ml. Activation occurs in the female tract as a consequence of low pH, enzymatic activation with plasmin or interaction of latent TGFβ with thrombospondin-1 or αvβ6 or αvβ8 integrins expressed at the epithelial cell surface (for details, see Tremellen et al. 1998; Robertson et al. 2002).

Full size image

TGFβ has also been identified as the principal active moiety in human semen by experiments with primary and transformed human cervical keratinocyte cultures (Gutsche et al. 2003; Sharkey and Robertson 2004; Tremellen and Robertson 1997). Isoform-specific immunoassays have shown that TGFβ in human semen is principally of the TGFβ1 isoform, with a lower content (5%–10%) of TGFβ2 (Nocera and Chu 1995; Srivastava et al. 1996). The content of TGFβ3 approximates that of TGFβ1 yielding a final concentration of approximately 300 ng/ml total TGFβ (D.J. Sharkey and S.A. Robertson, unpublished). Responsiveness of both murine uterine epithelial cells and human cervical epithelial cells to TGFβ is maximal at ovulation (Tremellen and Robertson 1997). Whether this reflects the steroid-hormone-regulated expression of TGFβ receptors or other components of the docking or signal-transducing molecule repertoire remains to be elucidated.

Other inflammation-inducing moieties present in semen synergise with TGFβ in targeting female tract cells but their content varies between species. Prostaglandin E (PGE) is abundant in human semen as the 19-hydroxy form but is undetectable in rodent and porcine seminal plasma. In vitro experiments with cultured human cervical explants have shown that 19-hydroxy PGE promotes the expression of chemotactic IL-8 and inhibits the expression of the anti-inflammatory molecule, secretory leukocyte protease inhibitor (Denison et al. 1999). Seminal plasma is also rich in IL-8, which synergises with TGFβ to induce IL-1β, IL-6 and leukocyte inhibitory factor (LIF) from endometrial epithelial cells (Gutsche et al. 2003).

Bacterial lipopolysaccharide also induces cytokine synthesis in mouse and human uterine and cervical epithelial cells, through binding to Toll-like receptors TLR2 and TLR4. The relative abundance in semen of different bacterial species might influence the cytokine profile elicited in the female tract via differential binding to TLRs and other pattern recognition receptors on the surface of the epithelial cells in the reproductive tract (Schaefer et al. 2004). Finally, the type-1 cytokine interferon-γ acts as a potent inhibitor of TGFβ signalling, both in human and mouse epithelial cells (Glynn et al. 2004). Together, this array of active seminal constituents acts in concert to regulate the character of the female response. Fluctuations in seminal fluid constituents between individuals and even within an individual over time might cause variations in the pattern of cytokine synthesis elicited and thus, the leukocyte composition of the ensuing inflammatory response.

Female tract responses activated by seminal plasma

The inflammatory response stimulated by seminal plasma impacts on several reproductive processes by virtue of the wide variety of actions of the leukocytes recruited into the endometrial and cervical tissues. Four categories of effector function are postulated; (1) the clearance of superfluous sperm and microorganisms introduced into the uterus at mating; (2) the activation of female immune responses specific to paternal transplantation proteins and other antigens present in semen; (3) the tissue remodelling associated with the preparation of endometrial receptivity; (4) the activation of the expression of cytokines and growth factors that have been implicated in pre-implantation embryo development (Fig. 3). Together, these seminal-plasma-induced changes in the female tissues act to promote the survival of the male gametes and the opportunity for oocyte fertilisation, embryo development and successful implantation. The potential areas of effector function will be discussed in the following sections.

Fig. 3
figure 3

Illustration of the action of seminal plasma signals in the female reproductive tract to facilitate embryo implantation. At mating, active moieties in seminal fluid interact with epithelial cells in the cervix and uterus of the female reproductive tract to induce the synthesis of pro-inflammatory cytokines, including GM-CSF, IL-6, LIF and a variety of chemokines. These pro-inflammatory cytokines cause the recruitment and activation of inflammatory cells in the endometrial stroma, including macrophages, dendritic cells and granulocytes, which have roles in regulating tissue remodelling of the endometrial stroma and in activating maternal immune accommodation of pregnancy. Epithelial cytokines activated by seminal plasma are also secreted into the luminal fluid where they exert embryotrophic actions on the developing pre-implantation embryo

Full size image

Impact on sperm survival and transport

One obvious role for the abundant populations of neutrophils that emigrate between epithelial cells into the uterine lumen is phagocytosis to ensure the clearance of micro-organisms and seminal debris remaining in the tract after intromission. The higher reproductive tract is normally sterile, with insemination providing the opportunity for access by commensal micro-organisms originating from male and female tissues or by sexually transmitted pathogens. In mice, bacteria are prevalent in the uterus after insemination but sterility is recovered within 24 h (Robertson et al. 1999). A male contribution to the maintenance of a sterile female tract would facilitate sperm survival and the likelihood of the mating event resulting in pregnancy. The physiological significance of seminal plasma in uterine clearance is illustrated in livestock species in which rapid resolution of the uterine inflammatory response is linked with implantation success (Rozeboom et al. 2000; Troedsson et al. 2001).

Phagocytic cells in the cervix and uterus target spermatozoa and bacteria, with selective sperm phagocytosis probably acting to filter out morphologically abnormal spermatozoa (Tomlinson et al. 1992). However, apparently viable and morphologically normal spermatozoa are also targeted suggesting that selection may occur on the basis of morphological or antigenic parameters in addition to fertilisation competence (Mattner 1969; Roldan et al. 1992; Taylor 1982), such as haplotype in the MHC-linked t-complex (Schimenti 2000). Evolutionary advantages might ensue from female mechanisms for the selective sequestration and inactivation of sperm within a single ejaculate or for distinguishing between sperm of rival mates in polyandrous species, thereby favouring genetically superior or phenotypically more competent male gametes (Roldan et al. 1992). The molecular basis of discrimination and the identity of any target structures remain to be identified and so whether an immunological selection mechanism operates remains unknown.

Activation of maternal immune tolerance

Macrophages and dendritic cells comprise the major populations of cells recruited into the endometrial stromal tissue after exposure to semen (McMaster et al. 1992; Robertson et al. 1996; Fig. 1). These cells are dedicated antigen-processing and antigen-presenting cells that engulf and transport seminal antigens to draining lymph nodes, resulting in the activation of immune responses to paternal MHC and other antigens in semen. The female immune response to seminal antigens is characterised in mice by hypertrophy of lymph nodes draining the uterus and evidence of lymphocyte activation (Beer and Billingham 1974; Johansson et al. 2004; Piazzon et al. 1985). Matings with vasectomised males or in females with oviductal ligation indicate that immune activation occurs independently of sperm and of the embryo (Chambers and Clarke 1979; Johansson et al. 2004). In contrast, males from which the seminal vesicle glands have been surgically removed fail to stimulate lymphocyte activation and proliferation, confirming the necessity for seminal plasma in activating the immune response (Johansson et al. 2004).

Hostile immune responses to seminal antigens would be incompatible with the maintenance of fertility and would prevent the female tract tolerating future exposure to semen. There would also be consequences for pregnancy, since the conceptus shares paternal antigens with those in semen (Thaler 1989). However, the immune activation elicited by semen does not activate the rejection of male antigens, because of the presence, in seminal plasma, of several powerful immuno-regulatory molecules, such as PGE and TGFβ, which prevent the destructive Type-1 (cell-mediated) immune responses (Letterio and Roberts 1998; Weiner 2001). Indeed, the female tract immune response to semen conversely appears to result in a state of functional immune tolerance to male antigens. Female mice exposed to seminal plasma, even in the absence of sperm or a conceptus, show hypo-responsiveness in Type-1 immunity to male MHC antigens and tolerate challenges with male tumour cells (Robertson et al. 1997). A role for semen in inducing male antigen-specific tolerance was first suggested by experiments showing that mated female mice fail to reject skin grafts of paternal origin (Lengerova and Vojtiskova 1966). Subsequently, protection was shown to be conferred to major histocompatibility antigens and the minor histocompatibility antigen H-Y, but only when sperm was delivered in the context of seminal plasma (Beer and Billingham 1974; Hancock and Faruki 1986). The cellular mediators of this response have yet to be defined but suppressive regulation after initial antigen-specific activation of the immune response has been implicated (Kapovic and Rukavina 1991; Piazzon et al. 1985).

We have further hypothesised that the female immune response to ejaculate antigens also serves to activate cells that are subsequently recruited into the implantation site to facilitate early placental development (Robertson et al. 1997). Phenotypic similarities have been found between cells activated in lymph nodes after insemination with the antigen non-specific T-regulatory cells, natural killer (NK) and NKT cells found in the decidua during early pregnancy. The smaller contingents of the antigen-specific lymphocytes present in the implantation site might reasonably recognise paternal antigens present in semen and shared by the conceptus. In support of a relationship between mating and the immune response to embryo implantation, activated lymphocytes recovered from uterine draining lymph nodes after insemination can be radio-labelled and shown to traffic into the decidual tissues when transferred into pregnant recipient mice (Johansson et al. 2004). Consistent with an immunological “priming” function, exposure to the semen of future male partners in mice promotes embryo implantation and fetal growth in a partner-specific manner (Beer and Billingham 1974; Robertson et al. 2003), and the priming effect is ablated when lymph nodes draining the uterus are removed (Beer et al. 1975; Tofoski and Gill 1977).

Regulation of tissue remodelling

Leukocytes can exert effects in their local milieu, other than through activating immune responses, by secreting an array of potent enzymes and signalling molecules that affect extracellular matrix turnover and the behaviour of endothelial cells in the endometrium. The leukocytes recruited in response to semen, particularly macrophages, probably assist in restructuring the endometrial environment to facilitate implantation and placental development.

Regulation of angiogenesis is the major potential avenue for macrophage effects on implantation. Factors that regulate angiogenesis and vascular permeability are expressed in a tightly controlled fashion over the course of the oestrous cycle and early pregnancy (Ma et al. 2001) to effect the vascular remodelling essential for embryo implantation (Klauber et al. 1997). Activated macrophages have the capacity to influence each phase of the angiogenic process, including the alterations of the local extracellular matrix, the induction of endothelial cells to migrate and proliferate and the formation of capillaries (Sunderkotter et al. 1994). In other tissues, macrophages are recognised as the most prevalent source of vascular endothelial growth factor (VEGF) and are essential contributors to the angiogenic process via the synthesis of VEGF and other key angiogenic factors and vascular permeability agents (Yoshida et al. 1997). Consistent with a regulatory role for semen, vasodilation and oedema increase in the uterus during the first days after mating (Bollwein et al. 2003; O’Leary et al. 2004). VEGF is expressed by unidentified stromal cells distributed similarly to macrophages adjacent to blood vessels and scattered throughout the endometrial stroma (Halder et al. 2000; Shweiki et al. 1993). VEGF mRNA expression in hamsters is reduced after mating with accessory-gland-deficient males (Chow et al. 2003).

An additional target for macrophage-secreted products is the extracellular matrix of the endometrial stroma, which is remodelled prior to and during decidualisation, with the breakdown of the existing matrix and the deposition of new components (Aplin 2002). Matrix metalloproteinases (MMPs) regulate this process (Curry and Osteen 2003; Das et al. 1997) after coordinated increases in their transcription, secretion and proteolytic activation and of their regulatory proteins, the tissue inhibitors of metalloproteinases (TIMPs). Macrophages are a major source of a broad range of MMPs under the influence of cytokines, the extracellular matrix and prostaglandins (Goetzl et al. 1996). In mice, striking induction of MMP-2 is seen in the subepithelial stroma during the pre-implantation period and TIMPs are also induced from day 0.5 post-coitus (Das et al. 1997). In rats, the expression of MMP-7 (matrilysin) is highest on the first day after mating contemporaneously with the post-mating inflammatory response (Feng et al. 1998), with MMP-2 also being induced prior to embryo implantation (Zhao et al. 2002). Consistent with a role for semen-induced inflammatory cells in MMP regulation is the finding, in golden hamsters, that the absence of male accessory gland fluids is associated with reduced expression of MMP-2 at the implantation site (Chow et al. 2003).

Promotion of embryo attachment and implantation

Macrophage mediators potentially also target the luminal epithelial cells involved in embryo attachment during the initial phases of implantation. Specific changes in epithelial cell expression of integrins and mucins allow close apposition and then adhesion between the blastocyst and the luminal surface (Aplin 1997), with fluctuating expression in adhesion and anti-adhesion molecules providing a barrier until the window of implantation when the embryo becomes attached to the epithelium just prior to its invasion. The induction of specific integrin subunits (α1, α4, αv, β3) has been implicated in defining endometrial receptivity, whereas Muc1, the predominant anti-adhesive molecule, is lost from the epithelial surface by transcriptional downregulation and proteolytic cleavage (Aplin 1997; Lessey et al. 1996). Macrophages interdigitate between epithelial cells in the endometrium, a spatial association that affords a role in influencing integrin expression at the paracrine level. Leukocyte regulation of epithelial adhesive properties has been demonstrated with human uterine epithelial cells in vitro (Kosaka et al. 2003). The ability of macrophages to alter transport properties and affect epithelial barrier integrity (Zareie et al. 1998) might further contribute to implantation by facilitating trophoblast breaching of the epithelial surface and potentially contributes to the epithelial apoptosis that subsequently occurs.

Synthesis of embryotrophic cytokines

The cytokines induced after semen exposure are secreted into the uterine luminal fluid and epithelial glycocalyx lining the luminal space where they interact with the developing embryo as it traverses the oviduct and uterus prior to implantation. Several cytokines activated by semen are amongst those attributed with regulating the proliferation, viability and differentiation of blastomeres in embryos (Kane et al. 1997; Pampfer et al. 1991). GM-CSF, a principle cytokine in the post-mating inflammatory response, targets the pre-implantation embryo to promote blastocyst formation, increasing the number of viable blastomeres by inhibiting apoptosis and facilitating glucose uptake (Robertson et al. 2001b). Human embryos cultured in GM-CSF are twice as likely to reach the blastocyst stage of development, blastulate earlier and have increased cell numbers both in the inner cell mass and in the trophectoderm (Sjoblom et al. 1999). Other cytokines targeting the developing blastocyst, including IL-6 and LIF, are similarly induced after exposure to semen (Gutsche et al. 2003; Robertson et al. 1992; S.A. Robertson, unpublished). Perturbations in the growth factor environment experienced by the pre-implantation embryo impairs normal development of the placenta and fetus, with long-term consequences for post-natal health and metabolic programming in progeny (Sjoblom et al. 2005).

Effects in tissues distal to the reproductive tract

The effects of exposure to semen reach beyond the immediate site of deposition and can cause changes in organs and systems elsewhere in the female body. In mice, macrophage populations in the corpora lutea are augmented by exposure of the female tract to seminal plasma constituents (Gangnuss et al. 2004). Oestrous pigs exhibit a reduction in the interval between the surge of luteinising hormone and ovulation in response to seminal factors (Waberski et al. 1997). The effects appear to be mediated by the local transport of seminal constituents or semen-induced effector molecules originating in proximal parts of the tract, since the effect is seen only in ovaries ipsilateral to uterine horns receiving the seminal stimulus. The mechanism and route of signal transfer from the uterus to the ovary is unknown; possibilities include a vascular counter-current mechanism linking the uterus and ovary, the efflux of seminal factors via the oviduct or the trafficking of inflammatory cells by lymphatic connections.

The immunological consequences of insemination also appear to extend to lymphoid tissues distal to the reproductive tract and its immediate draining lymph nodes. Experiments in mice have shown that exposure to semen at mating is central to the sequence of changes in the thymus that accompany normal pregnancy. The dramatic thymic involution that occurs by day 7 of pregnancy is preceded by the activation and expansion of thymic lymphocyte populations; components of the ejaculate have been implicated in this activation (Clarke 1984). These systemic changes in the female immune response in early pregnancy are thought to contribute to immunological accommodation of the foreign conceptus (Clarke 1984).

The effects of semen exposure have been hypothesised to extend even to the neurological system to affect female mood and behaviour (Ney 1986). Vaginal uptake and adsorption into the peripheral blood of several seminal constituents, including testosterone, oestrogen and prostaglandins, is known to occur within hours of intercourse (Benziger and Edelson 1983) and vaginal oestrogen uptake is linked with the induction of prolactin synthesis in women (Keller et al. 1981). Interestingly, support for a beneficial effect of semen exposure in preventing depression comes from studies investigating the relationship between depressive symptoms and condom use in women (Gallup et al. 2002).

Evolutionary significance

The function of seminal plasma in promoting female tract receptivity in mammals is reminiscent of the functions of male accessory gland fluids in insects (Gillott 2003). In addition to facilitating sperm transfer, these secretions act within and beyond the female reproductive tract to serve a variety of functions that collectively improve the likelihood of a given male siring offspring. Female reproductive processes affected by male factors include sperm selection and storage, the timing and number of oocytes ovulated and female behaviour in terms of receptivity to other males (Gillott 2003). This commonality across the insect and mammalian kingdoms underscores the potentially major importance of this mechanism in influencing reproductive processes at a population level and raises questions concerning the significance of seminal plasma signalling as a tool in male–female sexual “conflict” in mammals (Roldan et al. 1992).

In mammals, as in insects, the function of seminal plasma in facilitating female receptivity for pregnancy presumably has an evolutionary benefit in terms of increasing the likelihood of a pregnancy after insemination by a given male. Post-copulatory selection strategies only have value if there is variance between males in the extent to which seminal fluids elicit female responses and when polyandrous mating behaviour occurs (Zeh and Zeh 2001). In mammals, more than 90% of species are considered polygynous, with males rarely investing in the support of offspring, and so strong selective pressure favours the evolution of male strategies that enhance mating success (Roldan et al. 1992). Genetic tracking of paternity shows that polyandry is far more pervasive in mammalian species than previously appreciated, suggesting that the benefit of avoiding partner incompatibility outweighs the potential risks and costs (Hughes 1998). The considerable variation between individuals in the seminal plasma content of TGFβ and other active moieties in men and other species (Robertson et al. 2002) may thus be of relevance, although the extent to which this variation correlates with male fertility status remains to be determined.

From the female perspective, the opportunity for differential female tract responsiveness to seminal factors provides a mechanism for assessing male fitness for potential reproductive investment and may facilitate female choice of genetically compatible males (Zeh and Zeh 2001). Seminal activation of immune recognition processes in the female would facilitate sperm selection and affect the chance of fertilisation occurring. Perhaps more importantly, the capacity of female tissues to respond to seminal fluids by activating immune tolerance and tissue remodelling pathways might influence the likelihood of subsequent embryo implantation, with differential responses allowing female tract screening and discrimination between embryos. The capacity to switch between immune tolerance and immune rejection may further underpin the efficient disposal of suboptimal embryos early during the reproductive process before significant maternal investment has occurred.

Summary and conclusions

A significant body of evidence now links semen exposure with pregnancy outcome in human, rodents and several other mammalian species and the studies reviewed herein are beginning to provide explanations for the underlying molecular and cellular mechanisms. Seminal plasma can thus no longer be considered simply as a sperm transport medium but, instead, must be recognised as a means for communication between the male and female reproductive tissues and as a necessary agent for the conditioning of the female tract to allow optimal pregnancy success. Appreciation of this function of seminal plasma provides new insights into the processes influencing reproduction and fertility at a population level and is likely to be relevant in humans and in other mammalian species.

To date, research in this field has focused largely on rodent and livestock species and, for obvious reasons, the significance of seminal factors in humans has been more difficult to explore. Whereas difficulties are encountered in justifying direct extrapolation from the rodent to the human, the emerging picture justifies closer examination of the relationship between seminal exposure and the incidence of infertility and subfertility in human pregnancy. We speculate that the aberrant immune responses associated with “shallow” placentation in preeclampsia and recurrent miscarriage might be initiated by insufficient or inappropriate immune responses to seminal antigens following intercourse; this may perhaps be linked to partner incompatibility, seminal plasma cytokine deficiency or female incapacity to respond to seminal signals (Dekker et al. 1998; Robertson et al. 2003). There are additional implications for assisted reproductive technologies in which pregnancies are routinely initiated in the absence of intercourse. A better understanding of the physiological significance of semen in human reproduction requires further detailed exploration of the cellular and molecular events within the female reproductive tract at insemination and may eventually yield novel therapies for infertility and pathologies of pregnancy.