Human Reproduction Update Advance Access originally published online on March 24, 2005
Human Reproduction Update 2005 11(3):215-228; doi:10.1093/humupd/dmi003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The role of paf in embryo physiology
Human Reproduction Unit, Department of Physiology, University of Sydney, Royal North Shore Hospital, St Leonards, NSW, 2065, Australia
Email: chriso{at}med.usyd.edu.au
 |
Abstract |
|---|
TOP
Abstract
IntroductionEmbryo-derived paf (1-o-alkyl-2-acetyl-sn-gylcero-3-phosphocholine) is produced by de novo synthesis. This synthesis commences soon after fertilization and persists throughout the preimplantation phase. Paf is produced and released by the embryos of all mammalian species studied to date. Its release from the embryo involves binding to extracellular albumin in a manner that protects paf from enzymatic degradation. Released paf causes a range of alterations in maternal physiology, including platelet activation, changes in oviductal, endometrial and maternal immune function. Paf also acts in an autocrine fashion as a trophic/survival factor for the early embryo. In vitro, supplementation of culture media with paf improves embryo development. Embryo-derived paf's autocrine actions are transduced by 1-o-phosphatidylinositol-3-kinase, which induces characteristic calcium transients within the early embryo. The calcium transients require both the influx of external calcium and release of inositol trisphosphate-dependent internal calcium stores. Buffering these transients compromised embryo development in a manner that was reversed by exogenous paf. Assisted reproductive technologies compromise the production of paf by some embryos and retard the expression of the paf receptor. This deprivation of paf's action is one of the factors limiting the survivability of embryos produced by assisted reproductive technologies. Paf is one of several autocrine and paracrine trophic/survival factors that act on the early embryo. These factors probably act cooperatively and may, to some degree, be mutually redundant. As the earliest-released and the best-described embryotrophin, paf provides an important exemplar for understanding the role of ligand-mediated trophic support of the early embryo.
Key words: paf / zygote / blastocyst / embryo / fertilization in vitro / signal transduction
|
Introduction |
|---|
It is 20 years since the first report (O'Neill and Saunders, 1984
) of the release by the human embryo of a soluble factor that caused blood platelet (thrombocyte) activation. This was subsequently shown to be mediated by a potent ether phospholipid, 1-o-alkyl-2-acetyl-sn-gylcero-3-phosphocholine (paf) (O'Neill, 1985b; Collier et al., 1988). The released paf causes changes to a range of maternal reproductive tract functions. Importantly, it also acts back on receptors expressed by the embryo to complete an autocrine trophic loop. Paf is also synthesized and released by many other cell types. An equally diverse range of cells are shown to be biologically responsive to paf. While not a universal mediator, paf is certainly used in an extraordinary range of physiological and pathological settings. Since paf is made by and acts upon the zygote, it is perhaps not surprising that such a diverse range of cells within the body should subsequently be responsive to the mediator. Given its diversity of action, the original name of platelet-activating factor is no longer suitable or relevant. Rather, the term paf should solely be used. Paf should refer to the biochemical entity 1-o-alkyl-2-acetyl-sn-gylcero-3-phosphocholine. It should not refer to the biological activity since many agents have platelet-activating functions. The term platelet-activating factor should now fade from use.
Paf was the first mediator released by the preimplantation embryo to be identified (O'Neill, 1985a,b). It was subsequently shown that paf stimulated embryo metabolism (Ryan et al., 1989), cell-cycle progression (Roberts et al., 1993) and embryo viability (Spinks and O'Neill, 1988; Ryan et al., 1990b; Spinks et al., 1990) providing evidence for an autocrine loop in the early embryo. Insulin (Wales et al., 1985) also exerts trophic effects on the early embryo, the early embryo expresses a functional receptor for insulin (Harvey and Kaye, 1988; Heyner et al., 1989), but not insulin itself, suggesting that potential endocrine stimulation of embryo development also occurred. The possibility that a range of trophic factors may act on the early embryo was suggested by the detection of mRNA for a range of peptide growth factor ligands and their corresponding receptors (Rappolee et al., 1988). Since these observations, many putative autocrine, paracrine and endocrine factors have been implicated in supporting preimplantation embryo development (Kane et al., 1997; Kaye, 1997; Hardy and Spanos, 2002).
For most of these putative autocrine embryonic trophic factors there is a dearth of information on the nature of their action and the nature of the embryos' response to them. In recent years some information on the nature of paf's signal transduction within the embryo has become available, and to date paf's actions on the early embryo are the best described of the putative embryotrophins. Since paf seems to be one of the first of these factors to be produced and act on the embryo, it seems a suitable exemplar for studies of the role and mode of action of the embryotrophins.
Paf is also produced by a range of other reproductive tissues, including the endometrium, sperm, the ovary and the fetus. This review will primarily consider the actions of embryo-derived paf during the preimplantation stage of development.
|
Discovery of embryo-derived paf |
|---|
Paf was first identified as an agent released from sensitized basophils as a consequence of antigenic stimulation (Benveniste et al., 1972
). It was proposed as one of the early onset mediators of hypersensitivity responses. Its biochemical identification (Demopoulos et al., 1979) was soon followed by evidence that it was the same entity as a mediator with renal antihypertensive activity (Wykle et al., 1980). Its identity as an ether phospholipid was surprising since this class of lipids, allthough well described as minor but important structural components of membranes, was not considered at the time to be biologically active.
Its release by the early embryo was detected by the observation of a significant thrombocytopenia soon after conception (O'Neill, 1985a,c). This occurred in several strains of mice and was shown to be unequivocally a consequence of fertilization (rather than insemination, ovulation or any of the endocrine events associated with fertilization) (O'Neill 1985a,c). The thrombocytopenia could also be induced by injection of embryo-conditioned culture media into mice. The chemical identity of the active agent was confirmed by structural analysis (O'Neill, 1985b), chromatography (Ammit et al., 1992), metabolic studies (Wells and O'Neill, 1992), mass spectrometry (Kodama et al., 1989) and immunological assays (Ammit and O'Neill, 1991). The release of paf by embryos has been confirmed in the human (Collier et al., 1990; Vereecken et al., 1990; Nakatsuka et al., 1992), mouse (Kodama et al., 1989; Ripps et al., 1993; Suzuki et al., 1995), sheep (Battye et al., 1991), rabbit (Minhas et al., 1993), and hamster (Velasquez et al., 1995). Paf release by the early embryo may occur in most mammalian species. A more extensive phylogenetic study has not been completed and is warranted.
As a consequence of embryo-derived paf release, platelet activation occurs within the microvasculature of the oviduct (Stein and O'Neill, 1994). It is likely that the resulting activated platelets are trapped by the reticuloendothelial system and hence are lost from circulation in early pregnancy-associated thrombocytopenia. This occurs in mice (O'Neill, 1985c), bovine (Kojima et al., 1996a), hamster (Velasquez et al., 1995) and rabbit (Kojima et al., 1996b). In the human there is a mixed picture, with thrombocytopenia in some cases and thrombocytosis in others (O'Neill et al., 1988; Yeung et al., 1992). The excess production of platelets in thrombocytosis may well be a reflex to the acute consumption of platelets that occurs following fertilization (O'Neill et al., 1988).
Paf also initiates a change in the function of the maternal immune system early in pregnancy, manifested as a modification of T-lymphocyte rosette formation (Orozco et al., 1986; Sueoka et al., 1988). The full physiological significance of this change in T-lymphocyte function is not known, but it has been interpreted by some as evidence of maternal immunosuppression (Morton et al., 1976).
The release of a mediator from the single cell zygote capable of causing such extensive changes to maternal physiology suggests that it was a molecule of extraordinary potency. This proved to be the case, with paf activity in some bioassays being in the range 10–12–10–14 mol/l. The fact that paf is a phospholipid offered the possibility that its synthesis and release may occur without the requirement for new transcription from the embryonic genome. These observations make paf an attractive candidate for important signalling events in the initiation of early pregnancy.
|
The biosynthesis of embryo-derived paf |
|---|
Embryos have high levels of paf release, ranging from 1 to 100 s of ng paf per embryo per 24 h (Ammit and O'Neill, 1991
; Roudebush et al., 2002a). This output suggests that the release is likely to be the result of new synthesis rather than release of preformed stores. Paf's biosynthetic pathway in somatic cells has been well described (for review see Snyder, 1995). Pulse-chase studies showed that several radiolabelled precursors were incorporated in the paf released by the embryo (Wells and O'Neill, 1992). Importantly this occurred not only for immediate precursors such as lysopaf but also for more basic components such as long chain alcohols (hexadecanol and octadecanol) and alkylacetylglycerol. Radiolabelled simple substrates such as lactate, pyruvate and glucose were also incorporated into released paf. These pulse-chase studies showed that the early embryo produced and released paf and that synthesis occurred de novo from basic substrates. The biosynthetic pathways inferred from these studies are shown in Figure 1.
|
The incorporation of the long chain alcohols (hexadecanol and octadecanol) by formation of an ether linkage occurs within the peroxisome of cells. The reaction utilizes the enzyme alkyl-dihydroxyacetone phosphate synthase (EC 2.5.1.26 [EC] ) (van den Bosch et al., 1993
). Thus, peroxisomal status has particular importance within cells synthesizing paf (and other ether lipids) de novo. Currently, there is little knowledge of peroxisomal status and biogenesis within the early embryo. Mouse embryos produce paf using either hexadecanol or octadecanol. There is considerable heterogeneity in the amounts of each form of embryo-derived paf synthesized. Overall, there is more hexadecyl-paf than octadecyl-paf production found (Ammit et al., 1992). In many cultures embryos exclusively produce hexadecyl-paf, whereas this was uncommon for octadecyl-paf. This pattern of biosynthesis may be biologically relevant since it is reported that for mouse embryos hexadecyl-paf exerted trophic actions on the embryo whereas octadecyl-paf was inert (Stoddart et al., 2001). The results also suggest that an understanding of long chain alcohol biosynthesis within the early embryo has important developmental implications.
A penultimate step in embryo-derived paf biosynthesis produces lysopaf and arachidonic acid. The final reaction converts lysopaf to paf via acetylation (Figure 1). The regulation of embryo-derived paf's biosynthesis may be at the final steps of its synthesis, since activity of acetyl-coenzyme A:lysopaf acetyltransferase (EC 2.3.1.67
[EC]
) sharply increased at the time that paf synthesis and release occurs (zygote) and peaked at the 2-cell stage (in mouse) (Wells and O'Neill, 1994). By contrast, a critical enzyme in the de novo synthesis of paf (cytidine:alkylacetylgylcerol cholinephosphotransferase) showed relatively modest changes in activity during early development (Wells and O'Neill, 1994).
|
Paf metabolism |
|---|
The deacetylation of paf creates lysopaf. Lysopaf is biologically inert in paf bioassays. A special class of phospholipase A2 (PLA2) (groups VII and VIII) deacetylates paf (Tjoelker et al., 1995
). This PLA2 activity is commonly referred to as paf:acetylhydrolase (pafah, EC 3.1.1.47
[EC]
). It is thought to have specific roles in modulating paf bioavailability and limiting its half-life (the enzyme activity also has a role in detoxifying oxidized phospholipids) (Prescott et al., 2000). The ubiquitous nature of these metabolic enzymes for paf has led to the generally accepted concept that released paf does not accumulate to a sufficient concentration nor persist for long enough to act as other than a local autocrine or paracrine mediator.
Pafah takes several molecular forms: a plasma (hepatic) form and several intracellular forms (for review see Prescott et al., 2000). A pafah activity was detected within the 2-cell embryo, although it was not biochemically identified (Wells and O'Neill, 1994). Pafah activity was also present within the uterus during the preimplantation stage of pregnancy. Its activity changed markedly throughout the reproductive cycle. Activity was at a minimum in the endometrium and uterine luminal fluid during the preimplantation phase of pregnancy (O'Neill, 1995a). Pafah activity was reciprocally regulated by estradiol and progesterone (Chami and O'Neill, 2001). The dominant pafah activity within the endometrium and the uterine lumen during early pregnancy in the mouse had the biochemical characteristics of the plasma form of the enzyme (Chami and O'Neill, 2001). Thus, at a time when the embryo is producing and releasing paf, the sex steroids act on the uterus to reduce pafah activity.
In vitro, embryo-derived paf was not degraded by serum pafah, yet exogenous paf added to the same media was readily degraded (Ammit and O'Neill, 1997b,c). Furthermore, embryo-derived paf was readily degraded by pafah following its extraction and preparation as an aqueous solution. The resistance of embryo-derived paf to hydrolysis by pafah was a consequence of the nature of its binding to albumin (Ammit and O'Neill, 1997b,c).
The combined effects of low pafah activity in the uterus and the resistance of embryo-derived paf to pafah suggest that embryo-derived paf released by the embryo is capable of accumulating to biologically significant quantities in vivo.
|
Paf-binding proteins |
|---|
Embryo-derived paf displayed differences in biochemical behaviour compared with synthetic paf added to culture media (Ammit and O'Neill, 1997b
,c). This occurred despite unequivocal evidence that both sources of paf had the same chemical identity. Embryo-derived paf was not detected using a direct assay of embryo-conditioned media (with either a bioassay or immunoassay). Successful assay of embryo-derived paf required full organic extraction prior to assay. By contrast, synthetic paf added to such medium was readily measured by direct assay, as was embryo-derived paf following its organic extraction from media and re-addition to equivalent media (Ammit and O'Neill, 1997c). Synthetic paf was detected by direct assay following incubation with media for periods of 1–24 h. Furthermore, unextracted embryo-derived paf was resistant to pafah while synthetic paf added to the same media was rapidly metabolized. These differences in biological and biochemical behaviour of embryo-derived paf were a consequence of a unique form of binding of paf to albumin upon its release from the embryo. This form of binding does not occur when exogenous paf is added to media containing albumin (Ammit and O'Neill, 1997b,c).
Limited proteolytic digestion of albumin from embryo-conditioned media showed that embryo-derived paf was associated with a region of albumin located between amino acids 240 and 386 (full length 583) (Ammit and O'Neill, 1997c). By contrast, synthetic paf was largely absent from this region. This segment of albumin contains a hydrophobic core which can provide a binding region for hydrophobic molecules (Brown and Shockley, 1982; Brodersen et al., 1990; Ammit and O'Neill, 1997c). The secondary and tertiary structure of albumin is highly dependent upon the presence of 17 interchain disulphide bonds that link 34 (of the available 35) cysteine residues. Albumin in medium that was exposed to embryos expressed more reactive thiol residues than untreated medium and was more readily reduced by dithiothreitol. Thus embryo-dependent conformational changes to albumin, involving cysteine–cysteine disulphide bonds, occurred in conjunction with the loading of embryo-derived paf into a solvent-protected site (Ammit and O'Neill, 1997c).
Breaking the disulphide bonds within albumin was required to remove the solvent-protected status of cell-derived paf and suggests that some form of disulphide isomerization is required for paf's ‘release’ from the cell and binding to albumin (Ammit and O'Neill, 1997c).
It is clear that this special form of binding of embryo-derived paf to albumin is not required for paf's trophic action on the embryo, since these actions can be readily mimicked by synthetic paf added to culture media (Ryan et al., 1990b; Emerson et al., 2000). Rather, it seems likely that the special nature of embryo-derived paf's binding to albumin upon its release ensures that the paf released is not immediately inactivated by pafah, but can accumulate to biologically relevant concentrations. The relationship between paf release and action is illustrated in Figure 2.
|
|
The release of paf |
|---|
Whereas some cell types appear to retain most of the paf they synthesize (Stewart and Phillips, 1989
), other cells, including those of the preimplantation embryo, release much of their paf as a soluble mediator. There is currently a poor understanding of the nature of the release of paf by cells and there is no satisfactory explanation of why some cells release paf while others seem to primarily retain it. No convincing set of data for an intracellular role for paf is available. One possible action for retained paf is for intercellular signalling with paf acting as a juxtacrine mediator (Zimmerman et al., 1993).
Following synthesis within the intracellular organelles, paf is moved to the inner leaf of the plasma membrane (possibly mediated by transfer proteins) (Banks et al., 1988). Release of paf requires its transbilayer movement to the outer leaf of the plasma membrane. In some cells this transfer is facilitated by loss of membrane phospholipid asymmetry, possibly under the regulation of cellular transglutaminase (Bratton et al., 1991) and/or the actions of P-glycoprotein (Ernest and Bello-Reuss, 1999; Raggers et al., 2001). Upon translocation to the outer membrane, paf is available for release. The ether-linked alcohol at C1 and acetyl group at C2 causes paf to behave differently than esterified membrane lipids (Kantar et al., 1991). It is less hydrophobic than many membrane lipids (Kramp et al., 1984; Huang et al., 1986) and it is therefore likely that its release from the membrane is energetically favoured in the presence of hydrophilic acceptor molecules.
The amount of paf released by the 2-cell mouse embryo is dependent upon the extracellular albumin concentration, with the amount of paf released increasing as the albumin concentration increases (O'Neill, 1997). The albumin dependence of paf release has been demonstrated for other cell types (Benveniste et al., 1972; Ludwig et al., 1985) and it is likely that extracellular albumin acts as a general ‘acceptor’ for paf.
Embryo-derived paf induces transient increases in the intracellular calcium concentration ([Ca2+]i) within the early embryo. This action requires the presence of extracellular albumin and is inhibited by prior brief exposure of embryos to recombinant pafah or paf receptor (pafr) inhibitors (Emerson et al., 2000). This shows that while paf was present on the embryo's membrane it was susceptible to hydrolysis by pafah, but upon release to its protected site in albumin it is stable (Ammit and O'Neill, 1997c). These results have been interpreted as evidence that the autocrine action of paf requires its movement to the outer leaf of the plasma membrane of the early embryo and that its release and activity requires the action of an external acceptor molecule in the form of albumin. The carrier in turn protects paf from enzymatic degradation and donates it back to a high-affinity paf receptor on the embryo, creating an autocrine loop (Figure 2).
|
The paf receptor (pafr) |
|---|
To date, one pafr has been identified and subjected to molecular cloning. A functional pafr was first identified in a guinea-pig lung cDNA library (Honda et al., 1991
), and homologous genes have since been found in many species including the human, rat and mouse (for review see Ishii et al., 2002). The human pafr gene generates two different mRNA species, and the expression of these is under the regulation of distinct promoters (Ishii et al., 2002). The gene has an intronless open reading frame. The human gene is localized to chromosome 1 and the mouse to the D2.2 band of chromosome 4. The human receptor is a single protein composed of 342 amino acids; the mouse is one amino acid smaller. The receptor is expressed in the spermatozoon (Wu et al., 2001) and the early embryo (Roudebush et al., 1997; Stojanov and O'Neill, 1999).
The receptor belongs to the G-protein-coupled receptor family, shows wide tissue distribution and seems to account for many of the reported actions of paf (Ishii et al., 1997, 2002). Transgenic over-expression of the receptor (pafr-tg) leads to variable levels of expression within tissues. Although highly expressed in some tissues, it was barely detectable in the uterus or testis. However, mating of pafr-tg animals with wild-type partners resulted in a lower than expected penetrance of the transgene (transgenic:wildtype 1:10 for transgenic mothers and 1:2.3 for transgenic fathers) (Ishii et al., 1997). Interestingly, the litter size was reduced and the extent of this reduction corresponded to the decrease of the transgene penetrance. This implies that the reduced transgene penetrance was due to compromised survival of gametes or embryos that carried the transgene. It is yet to be defined how this is manifested. It is noteworthy that the dose–response of paf's beneficial effects on the embryo is a narrow quadratic curve; even small increments above the optimal paf concentration results in reduced embryo viability (Ryan et al., 1990a). The particularly poor transmission of the pafr-tg from female parents may imply an adverse interaction between pafr over-expression in females and the gametes/embryo. One interpretation of this model is that excessive paf signalling is detrimental to the successful establishment of pregnancy.
Interestingly, pafr-null mice were fertile following normal mating (Ishii et al., 1998), but when these mice were mated by IVF there was a marked reduction in fertilization rates compared with wild types (Wu et al., 2001). This observation is consistent with the many reports that paf is an important requirement for sperm function in vitro (Kumar et al., 1988; Baldi et al., 1993; Roudebush et al., 1993; Fukuda et al., 1994; Wu et al., 2001; Yan et al., 2003). It was also noted that pafr–/– embryos cultured in vitro developed poorly compared to pafr+/+ embryos (Lu et al., 2004). The poor development in vitro compared with that in vivo argues that paf's action on gametes and embryos occurs in concert with other trophic factors. In vivo the absence of paf signalling may be compensated by the other factors of maternal origin, whereas these would be limiting in the defined in vitro system (Wu et al., 2001; Lu et al., 2004). It is noteworthy, however, that while paf antagonists can block embryo development in vitro, a higher concentration than expected was required to achieve this effect (Emerson et al., 2000). Furthermore, paf antagonists appear to act as partial agonists (inverse agonists) rather than true antagonists in vivo. The modest reproductive phenotype of pafr–/– mice, and the apparent inverse agonism of paf antagonists on embryos (but not somatic cells), raise the possibility that the G-protein-coupled pafr may not be the sole receptor involved in transducing paf's actions in the early embryo.
There are several lines of evidence for alternative receptors, yet to date there is no conclusive evidence for their existence. One striking candidate is the type I intracellular form of pafah. Type I pafah (pafah I) forms a complex consisting of two catalytic subunits (either hetero- or homodimers of 
1 or 2 subunits) and a regulatory (ß) subunit (Watanabe et al., 1998). The regulatory subunit is homologous with the LIS1 gene. Structural analysis shows that this complex forms a G-protein-like structure (Ho et al., 1997). Pafah I 2 homodimers have anti-apoptotic functions (Bonin et al., 2004). Some paf antagonists can block cellular pafah activity (O'Neill et al., 1991) and can suppress pafah I 1 expression, hence favouring 2 homodimer formation (Bonin et al., 2004). Genetic disruption of pafah I causes early embryo lethality (Reiner et al., 1993; Hirotsune et al., 1998), due to failed proliferation of cells within the early embryo and aberrant trophoblast formation (Cahana et al., 2003). Its absence also causes failure of spermatogenesis due to apoptosis of the male germ cells (Yan et al., 2003). It has been speculated that paf (or related lipids) may regulate pafah activity in a manner analogous to the regulation of G-proteins by GTP, and thus provide a new signalling mechanism. However, strict functional proof of such a mechanism is awaited.
|
Paf in the female reproductive tract |
|---|
The female reproductive tract produces a significant amount of paf (Yasuda et al., 1986
; Angle et al., 1988; Kudolo and Harper, 1995; Chami et al., 1999) and a question that is commonly raised is whether the paf derived from the embryo and reproductive tract have different roles. Several differences are noteworthy. The production of paf by the uterus is steroid dependent and occurs in both conception and non-conception cycles (Angle et al., 1988; Chami et al., 1999; Chami and O'Neill, 2001), yet paf-induced early pregnancy-associated thrombocytopenia in mice only occurs in conception cycles or following embryo transfer (O'Neill, 1985c). This suggests that paf produced by the endometrium does not accumulate to a sufficient concentration (or is not released) capable of inducing platelet activation. The amount of paf released by the endometrium is small, and little of this paf is detected within the uterine lumen (Angle et al., 1988; Chami and O'Neill, 2001). The paf released from the endometrium is highly labile; it does not bind to albumin in a manner that protects it from hydrolysis by pafah (Chami et al., 1999). Currently, there is no definitive evidence that endometrial paf acts on the embryo.
Receptors for paf are located in the oviduct of hamsters (Velasquez et al., 2001), mice (Lash and Legge, 2001) and cows (Tiemann et al., 2001b). In the hamster, embryos migrate down the oviduct at a faster rate than unfertilized oocytes, suggesting a form of communication between the embryo and oviduct (Velasquez et al., 2001). Paf increases intracellular calcium concentrations in cultured oviduct cells (Tiemann et al., 1996). This effect was less consistently observed with human tissue (Downing et al., 1999, 2002). Physiological concentrations of paf in vitro increased ciliary beat of hamster oviduct ciliated epithelial cells, an action consistent with a role for paf in embryo migration (Hermoso et al., 2001). Paf also caused transepithelial chloride ion fluxes and changes in transepithelial electrical potential in human tubal epithelial cells (Downing et al., 1999, 2002). Embryo-derived paf also caused microvascular changes with a decrease in the cross-sectional area of small intramural blood vessels, and an apparent collapse of many subepithelial capillaries within the oviduct on day 2 of pregnancy (Stein and O'Neill, 1994). Pregnant mice also had fewer fenestrated capillaries whereas such vessels were common in pseudopregnancy. Activated platelets were only observed in the capillaries and venules of the oviduct-pregnant mice. These effects of pregnancy were reversed by a paf-receptor antagonist, suggesting that they were induced by embryo-derived paf (Stein and O'Neill, 1994). It is proposed that the local action of embryo-derived paf on the ciliated cells and smooth muscle of the oviduct together with net movement of fluid propel the embryo towards the uterus, thus accounting for preferential movement of fertilized versus unfertilized oocytes within the oviduct. The microvascular changes would undoubtedly change the oxygen tension and nutrient environment within the embryo's immediate environment. Given the early embryo's preference for a relatively hypoxic environment (Harvey et al., 2002), these changes to reproductive tract function induced by embryo-derived paf may favour development.
Paf also influences the function of the uterus (Yasuda et al., 1986). Paf is present within the uterus and its levels change under the influences of the sex steroids (Yasuda et al., 1986; Angle et al., 1988; Baldi et al., 1994; Kudolo and Harper, 1995; Chami et al., 1999). The pafr is expressed in the endometrium under the influence of the sex steroids (Ahmed et al., 1998; Chami and O'Neill, 2001; Tiemann et al., 2001a). Paf was first described as altering prostaglandins PGF and PGE release by the human endometrium in vitro (Smith and Kelly, 1988), and similar results have now been described in a range of species (Gross et al., 1990; Alecozay et al., 1991; Battye et al., 1996). Paf's effects on PG release may be mediated by nitric oxide (NO) (Ahmed et al., 1998). In the sheep, interferon (IFN) 
blocks paf-induced PGF2 release whereas oxytocin and paf have an additive effect on PGF2 release from the endometrium (Chami et al., 1999). A physiological role for endometrial paf has not been uneqivocally demonstrated. Evidence exists for several distinct roles for uterine paf at different stages of the reproductive cycle. Thus, it is implicated as a vasodilator to promote blastocyst implantation (Acker et al., 1988); as an angiogenic mediator in the regeneration of the endometrium after menses in humans (Ahmed et al., 1998); and as a mediator that acts to generate pulsatile release of PGF2 that leads to luteolysis in the non-conceptual cycle of sheep (Chami et al., 1999, 2004) (Figure 3).
|
Currently there is a body of evidence that paf produced within the female reproductive tract exerts local actions on the reproductive tract, and that embryo-derived paf also exerts actions on the reproductive tract (and has systemic effects). There is little direct evidence for an action of reproductive tract paf acting on the embryo.
|
Physiological targets of embryo-derived paf |
|---|
Given paf's many targets in early pregnancy, it was important to understand which of these functions are of most physiological relevance. Pafr antagonists were used to assess the effects of paf during early pregnancy. When administered to female rodents throughout the preimplantation phase of pregnancy, a range of paf antagonists caused a significant inhibition of the proportion of embryos that successfully implanted into the uterus (Acker et al., 1988
; Spinks and O'Neill, 1988; Ando et al., 1990a,b; Spinks et al., 1990; O'Neill, 1995b). This effect was not evident in the rabbit (Norris et al., 1994). Curiously, the inhibition of implantation in mice was generally not complete; furthermore, at high concentrations of antagonists the inhibitory effect was lost (O'Neill, 1995b). This dose-response suggests that in this model the antagonists were acting as partial agonists (inverse agonists) rather then true antagonists. This conclusion was supported by the observation that simultaneous administration of paf reversed the inhibition of embryo implantation caused by paf antagonists (Spinks and O'Neill, 1988). This pharmacology probably explains one report of the failure of paf antagonists to block implantation (Milligan and Finn, 1990).
The use of these paf antagonists within the dose range where they exerted contragestational actions in mice allowed judgements on targets of embryo-derived paf to be made. An embryo transfer model was used where either recipient females or the donors of embryos were treated with paf blocker. The treatment of embryo donors caused a reduction in implantation potential of embryos, while treatment of recipients had no effect on implantation rates following transfer (Spinks et al., 1990). This result infers, but does not explicitly show, that a primary target of paf was the embryo.
Further support for a direct action of paf on the embryo were the observations that the addition of paf to embryo culture media improved embryo development in vitro (Nishi et al., 1995; Roudebush et al., 1996; Stoddart et al., 1996, 2001; O'Neill 1997, 1998), enhanced rates of embryo metabolism (Ryan et al., 1989, 1990a, 1992), cell-cycle progression (Roberts et al., 1993) and improved rates of embryo implantation (O'Neill et al., 1989, 1992; Ryan et al., 1990b). All these effects of exogenously applied paf could be reversed by paf blockers. Furthermore, in the absence of exogenous paf, treatment of embryos in vitro with paf blockers (Emerson et al., 2000) or anti-paf antibodies (Roudebush et al., 1994) reduced their rate of development, and pafr–/– embryos had reduced rates of development in vitro (Lu et al., 2004). Taken together these results infer that a physiologically important target for embryo-derived paf is the embryo itself.
Current evidence shows that embryo-derived paf has many actions in early pregnancy. It clearly modifies oviduct function and seemingly creates beneficial changes to the embryo's milieu. Embryo-derived paf apparently may also promote embryo migration through the oviduct. It is assumed that these changes in the reproductive tract foster improved embryo development. An important target for embryo-derived paf, however, is the embryo-itself.
|
Embryo-derived paf's autocrine actions |
|---|
In an experimental model of reducing autocrine stimulation of development by embryo culture in limiting dilutions, it was shown that embryos have a requirement for exposure to diffusible autocrine trophic factors (including paf) over the first 48 h of development (to the late 2-cell stage). Deprivation of trophic support during this time resulted in a large increase in the proportion of embryos that underwent degeneration with a large proportion of cells undergoing apoptosis. Curiously, however, cell death did not occur until 24–48 h after the deprivation event occurred (primarily at the post-compaction stage of development). Supplementation of media with paf over this first 48 h of development could significantly reverse this embryopathy. Paf supplementation provided after the first 48 h of culture was not capable of rescuing the embryos (O'Neill, 1998
). The action of paf over the first couple of cell-cycles occurred without any apparent changes in the rate of cell-cycle progression (thus it did not act as a classical growth factor). It was concluded that the action of autocrine factors, such as paf, create a signal that is necessary for the subsequent survival of embryos past the 8-cell stage.
Treatment of 2-cell embryos with paf induces receptor-dependent characteristic [Ca2+]i transients (Roudebush et al., 1997; Emerson et al., 2000). Zygotes and 2-cell mouse embryos often display a spontaneous [Ca2+]i transient in vitro when embryos are collected and prepared in quantitative protein-free conditions and then exposed to medium containing albumin (Emerson et al., 2000). The spontaneous [Ca2+]i transients were a consequence of the action of embryo-derived paf. They were not seen in the absence of albumin and were blocked by a pafr-antagonist or by prior treatment of embryos with exogenous recombinant pafah. Reports of apparently normal preimplantation development in nominally protein-free media (Cholewa and Whitten, 1970) questions the importance of albumin-dependent paf release for embryo development. Yet such reports need to be viewed with caution. While the embryos are added to protein-free media, the embryos are collected from the reproductive tract with a considerable amount of contaminating protein (of which albumin is a significant component). Without very careful removal of all such contaminating protein, nominally protein-free culture media can actually contain protein concentrations in the low ug/ml range. This protein concentration is sufficient to support paf release in vitro.
Immediately following a [Ca2+]i transient induced by embryo-derived paf, embryos were not responsive to an subsequent challenge with exogenous paf. Inhibition of the transient induced by embryo-derived paf by treatment with pafah rendered the embryos sensitive to challenge with exogenous paf (Emerson et al., 2000). The resulting [Ca2+]i transients had similar characteristics to those induced by embryo-derived paf, including the presence of a refractory period after challenge.
The observation of [Ca2+]i transients induced by embryo-derived paf is evidence for the earliest physiological ligand-induced signal transduction event yet described in the embryo.
|
The action of paf on the zygote and 2-cell embryo |
|---|
Paf-induced [Ca2+]i transients were limited in the mouse embryo to the zygote and 2-cell stage (Emerson et al., 2000
). Yet functional studies had previously shown that exogenous paf also has trophic actions on the embryo at other stages of development (e.g. 8-cell and blastocyst) and these actions were apparently receptor-dependent (Ryan et al., 1990a, 1992; Roberts et al., 1993). This result suggests that paf may transduce its signal via different mechanisms at different stages of development. It may also mean that the consequence for the embryo of paf stimulation differs at different stages of development. Up to the 2-cell stage paf acts as a survival factor (O'Neill, 1998) and this seems to be mediated, at least in part by the mobilization of [Ca2+]i. In later stage embryos, paf exerts a trophic action that is not obviously mediated via [Ca2+]i mobilization. The mechanisms of signal transduction in the later preimplantation stages is currently not known. Thus, paf acts as a dual survival/trophic factor for the preimplantation embryo.
The paf-induced [Ca2+]i transients were mostly single transients across the embryo, that persisted for several minutes (Roudebush et al., 1997; Emerson et al., 2000). After the [Ca2+]i transient embryos were refractory to further paf challenges for an hour or more. After this time embryos spontaneously regained sensitivity to paf (Emerson et al., 2000). The periodic incidence of [Ca2+]i transients may therefore be regulated by both the rate of paf synthesis/release by the embryo and the rate at which the embryo becomes resensitized to paf. Such intermittent [Ca2+]i transients provide a hitherto unrecognized, information-rich, mechanism for signalling in the early embryo.
The response of the 2-cell embryo to exogenous paf was dose-dependent (dose range 0.037–370 nmol/l) and was inhibited by paf antagonists (Emerson et al., 2000). Embryos did not elicit a [Ca2+]i response to the inactive enantiomeric isomer of paf (3-o-alkyl-2-acetyl-sn-glycero-1-phosphocholine). Unfertilized oocytes and young zygotes (7–9 h after insemination) failed to show any [Ca2+]i transients in response to media containing albumin. By 10–13 h after insemination 25% of zygotes showed detectable [Ca2+]i transients in response to embryo-derived paf. For 2-cell embryos collected fresh from the reproductive tract, [Ca2+]i transients in response to embryo-derived paf varied with the age of the embryo. At 27–29 h after insemination there were modest responses to embryo-derived paf. By 31–33 h after insemination the average [Ca2+]i transient was significantly greater in peak amplitude and reached its peak earlier than observed at 27–29 h. At 35–37 h embryo responses to embryo-derived paf had declined in average amplitude and took longer to achieve peak amplitude. By 41–43 h the responses were significantly attenuated compared with each other time point tested for 2-cell embryos. No embryo-derived paf-induced [Ca2+]i transients were observed in 4-cell, 8-cell and morulae stage embryos. The lack of responsiveness in early zygotes and post 2-cell stage embryos was also evident after challenge with exogenous paf. This indicates that the failure to respond was not due to a deficiency in available embryo-derived paf but was due to embryos being incapable of responding at those stages of development (Emerson et al., 2000).
The [Ca2+]i transient appears to be a functionally significant consequence of paf action on the embryo since inhibition of its actions by buffering of the [Ca2+]i transients with BAPTA-AM [the Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetate-acetoxymethyl ester] blocked embryo development and this inhibition could be partially reversed by the addition of exogenous paf (Emerson et al., 2000