Prostaglandins and the Reproductive Cycle

Prostaglandins are a group of endogenously occurring acidic lipids that appear to be involved in a large number of reproductive processes. Kurzrok and Leib first reported that a substance in human semen altered the contractile activity of human uterine strips.¹ In 1935, Von Euler named this substance prostaglandin because he believed that it was produced in the prostate gland; however, it was later shown to be from the seminal vesicles.² Subsequent studies showed that semen contained a mixture of more than 14 polyunsaturated fatty acids, each composed of 20 carbons.

It was not until 1957 that the first prostaglandins (PGE₁ and PGF_2α) were isolated, and their structures were not determined until 1963.^3,4 In the late 1960s, the first research samples of prostaglandins became available. These samples allowed the first investigations of the role of these lipids in physiologic and pathophysiologic states. The 1970s brought two new discoveries that further clarified the mechanism by which prostaglandins exert their effects on tissues. The first observation was made by Vane, who reported that anti-inflammatory drugs such as aspirin exert their effects by inhibition of the cyclooxygenase enzyme system, which converts the prostaglandin precursor arachidonic acid to endoperoxides, thus giving rise to the prostaglandins.⁵ This discovery allowed investigators the first opportunity to block endogenous prostaglandin synthesis. The second discovery was made by Pace-Asciak and Wolfe, who found a novel derivative of arachidonic acid in rat stomach homogenates.⁶ This compound later was named prostacyclin (PGI₂) by Vane and coworkers, who showed that it was important in preventing both platelet aggregation and thrombus formation on vascular walls.^7,8

Like steroids, prostaglandins are a group of compounds that have somewhat similar structures (Fig. 1) but may have totally different effects. Prostaglandins appear to be autocrine or paracrine regulators rather than true hormones because they are both produced and act locally. These compounds are metabolized locally in many tissues, but if not degraded locally, they enter the circulatory system and are neutralized effectively by the action of the degradative enzyme prostaglandin dehydrogenase in the lungs.

Formation and Nomenclature

Prostaglandins are formed from one of three fatty acids that give rise to different families of compounds depending on the number of unsaturated carbon-to-carbon double bonds that are left in the final molecule; dihomogamma linolenic acid has one double bond, arachidonic (eicosatetraenoic) acid has two, and eicosapentaenoic acid has three. Each series of prostaglandins then is identified by a subscript (i.e., 1, 2, or 3). In the case of arachidonic acid, the subscript would be 2, thus PGE₂. In addition to the numeric subscript, prostaglandins may have an α or β subscript, which is used to denote the steric configuration of the hydroxyl group at the C₉ position: PGF_2α. The majority of endogenous prostaglandins formed in mammals are the result of metabolism of arachidonic acid, which is available both in the diet and after anabolic formation from linolenic acid.

Arachidonic acid is transported in plasma in the unesterified form and is linked noncovalently to plasma proteins. In the cell, arachidonic acid is stored in the phospholipid pools in the esterified form. However, only free unesterified fatty acids can be converted to prostaglandins by the cyclooxygenase enzyme system. Thus, availability of unesterified arachidonic acid generally is believed to be the rate-limiting step in prostaglandin biosynthesis. The release of arachidonic acid is controlled by the activity of an acylhydrolase enzyme, phospholipase A₂, which acts on the sn-2 position of the phospholipid to free arachidonic acid. On release, the cyclooxygenase enzyme places molecular oxygen at the C₁₅ position (carbons are numbered from the carboxylic acid [see Fig. 1]) to form a peroxide and a second molecule across C₉ and C₁₁ to form an endoperoxide. Simultaneously, a bond forms between C₈ and C₁₂, forming a cyclopentane ring, thus resulting in PGG₂. PGG₂ is converted to PGH₂ by a peroxidase acting at C₁₅. Both PGG₂ and PGH₂ have relatively short half-lives (t 1/2 = 5 minutes). As can be seen in Figure 1, once it is formed, PGH₂ can undergo a number of transformations to form the stable prostaglandin PGD₂, PGE₂, or PGF_2α. These changes occur either nonenzymatically or by reductase or isomerase enzymes that occur in a number of tissues. The distribution of these enzyme systems differs for each organ and species studied.

Two other prostanoids, thromboxane A₂ (TxA₂) and prostacyclin (PGI₂), are formed from the endoperoxides PGG₂ and PGH₂. TxA₂, which is formed by thromboxane synthetase, has a relatively short half-life (30 seconds) and has been implicated in both platelet aggregation and arterial vasoconstriction.⁹ TxA₂ breaks down spontaneously to TxB₂. PGI₂ (formed by prostacyclin synthetase) appears to be important in preventing platelet aggregation and is a potent vasodilator.¹⁰ PGI₂ also has a relatively short half-life (t 1/2 = 2 minutes) at physiologic pH and temperature and undergoes spontaneous hydrolysis to form the biologically inactive 6-keto PGF_1α. Reports suggest that, at least in the liver, 6-keto-PGF_1α can be converted enzymatically to the biologically active 6-keto-PGE₁, but the importance of this compound in reproductive medicine is unclear.¹¹ One other pathway requires discussion, the conversion of PGE₂ to PGF_2α by the enzyme 9-ketoreductase. Because in most tissues these two prostaglandins have opposite actions, this interconversion could lead to local modulation of the final biologic action.

Metabolism

Metabolism of prostaglandins PGD₂, PGE₂, and PGF_2α involves a series of enzymic steps. The first is oxidation of the C₁₅ hydroxyl group to a ketone by the enzyme 15-prostaglandin dehydrogenase (15 PGDH). This step can occur both locally (i.e., placenta, lung, spleen, kidney) and after release into the systemic circulation in the lungs, spleen, and liver. The second step is reduction of the double bond between C₁₃ and C₁₄ by the enzyme prostaglandin-13-reductase. It is the combination of steps one and two that gives rise to the major circulating metabolites 13, 14-dihydro-15-ketoprostaglandins, which are measured by radioimmunoassay (RIA). The third step in metabolism is beta oxidation, which occurs both in the liver and the kidney and which can remove two or four carbons, giving dinor or tetranor metabolites, respectively (see Fig. 1). The final step occurs in the liver and removes two carbons (from C₂₀ and C₁₉) by omega oxidation. These last two steps produce more water-soluble metabolites, which are excreted in the urine.

In contrast, the metabolism of PGI₂ is influenced further by its nonenzymatic hydrolysis to 6-keto-PGF_1α as the major plasma metabolite. Both this compound and 6-keto-PGF_1α are candidates for beta and omega oxidation, leading to a range of urinary metabolites. 2,3 dinor 6-keto-PGF_1α and 6-keto-PGF_1α usually are measured in urine as indices of prostacyclin production. Thromboxane A₂ also is hydrolyzed spontaneously, yielding the inactive metabolite TxB₂, itself a target for beta oxidation, yielding the urinary metabolite of thromboxane 2,3 dinor TxB₂. However, TxB₂ also may be dehydrogenated at C-11 to give 11-dehydro TxB₂, which has a longer plasma half-life, making it a suitable candidate for measurement. The 11-dehydro TxB₂ metabolite also is found in urine.

Isoprostanes

Isoprostanes are prostaglandin-like compounds that are produced by a nonenzymatic mechanism involving free radical catalyzed peroxidation of arachidonic acid. The major compounds formed contain the F-type prostane ring and are isomeric to PGF_2α and thus are termed F₂-isoprostanes, although E₂- and D₂-prostanes and isothromboxanes also can be formed.¹² Isoprostanes are formed initially by peroxidation of arachidonic acid esterified to tissue lipids and thus are markers for lipid peroxidation. Hydrolysis of isoprostanes from phospholipids in vivo is catalyzed by phospholipases. Measurement of 8-isoprostane within plasma and urine currently is being used as a reliable noninvasive approach to assess oxidative stress status (balance of pro-oxidants vs. antioxidant defenses) in humans. Increased urinary 8-isoprostane has been reported in chronic smokers,¹³ and treatment with antioxidant vitamin C significantly reduced 8-isoprostane excretion. A recent report¹⁴ found plasma 8-isoprostane to be increased significantly in preeclamptic women versus controls, which is consistent with the concept of preeclampsia being a state of oxidative stress. The 8-iso-PGF_2α compound is a potent vasoconstrictor of the renal and pulmonary vasculature and a bronchoconstrictor probably acting by means of a novel isoprostane receptor.

Measurement

Several techniques have been used to measure prostaglandins. These include bioassays, gas chromatography combined with mass spectrometry (GC-MS), RIA, and enzyme immunoassay (EIA). Of these methods, GC-MS probably is the most accurate and specific. However, this technique is expensive, difficult to perform, and time consuming. The bioassay methods generally require the presence of large concentrations of the prostaglandins and carefully treated bioassay tissues. These techniques, however, do give a continuous measurement, are invaluable when studying release of prostaglandins from tissues, and also are the only assays of biological activity. RIA and EIA have been the methods of choice for biologic investigation because they are easy to perform, provide rapid results, and allow measurement in a large number of samples. These techniques are not without problems, however, and care must be taken to ensure the specificity and accuracy of measurements.¹⁵ Over the past 20 years, the levels of prostaglandins measured by RIA have consistently become lower and approached levels measured by GC-MS. Although this suggests that the RIA methods have improved, it also suggests that earlier studies must be interpreted with care.

RIAs and EIAs for prostaglandins are generally preceded by extraction and chromatographic techniques, but recently several assays have been reported that utilize direct plasma measurements.^16,17 This may be desirable in some cases, since extraction may lead to breakdowns of the prostanoid to be measured. RIAs and EIAs have been described for most primary prostaglandins (PGD₂, PGE₂, PGF_2α), hydrolysis products (6-keto-PGF_1α, TXB₂), the circulating plasma metabolites (6, 15-diketo-PGF_1α; 13, 14-dihydro-15-keto-PGE₂; and 13, 14-dihydro-15-keto-PGF_2α), and urinary metabolites (dinor compounds).

When investigating physiologic phenomena in humans by measurement of peripheral (antecubital vein) prostaglandin levels, remember that concentrations of PGE₂ and PGF_2α are very low (1–2 pg/mL), that PGE₂ and PGF_2α are metabolized quickly to their respective 13, 14-dihydro-15-keto-metabolite (see Metabolism) and that measurement of the primary prostaglandins (PGE₂ or PGF_2α) does not reflect ongoing changes. Unfortunately, although the metabolite 13, 14-dihydro-15-keto-PGF_2α (PGFM) is stable, the PGE₂ counterpart is unstable and bicyclizes to form 11-deoxy-15-keto-13, 14-dihydro-11, 16-cyclo-PGE₂, also called bicyclic PGE₂.¹⁸ An assay for this compound has been described, but very little information exists regarding circulating levels of the PGE metabolite.¹⁸ Measurement of these plasma metabolites, which have a larger half-life than the present compound, does appear to be adequate for measuring PGE₂ and PGF_2α output.

It is abundantly clear that PGI₂ is present in such low concentrations (1 pg/mL) in the peripheral circulation that it does not function as a circulating vasodilator. The concentrations of the metabolite 6-keto-PGF_1α in plasma also are too low to measure reliably by RIA and EIA; therefore, GC-MS or high-pressure liquid chromatography (HPLC)-RIA measurement of 2,3 dinor 6-keto-PGF_1α in urine currently is performed as an index of whole-body PGI₂ production. Although 6-keto-PGF_1α also is measured in urine, this may arise from local kidney production and may escape beta oxidation and therefore is believed solely to represent renal synthesis.

The major problem in plasma thromboxane measurements is the synthesis of TxA₂ by platelets activated during venipuncture and blood clotting, despite the use of prostaglandin synthesis inhibitors during the collection procedure. Measurement of the hydrolysis product TxB₂ in plasma is therefore subject to sampling artifact. TxB₂ undergoes metabolism by two major pathways in humans either by beta oxidation, which forms the 2,3 dinor TxB₂ metabolite, or by dehydrogenation of the hemiacetal alcohol at C-11 to give a 11-dehydro TxB₂ metabolite. Being less polar, 11-dehydro-TxB₂ has a longer plasma half-life and can be measured by RIA.¹⁹ However, 11-dehydro-TxB₂ is as an abundant urinary metabolite as 2,3 dinor TxB₂, and measurement of either of these compounds in urine may represent changes in whole-body TxA₂ production.²⁰

Because of the plethora of urinary metabolites of both PGI₂ and TxA₂ that are produced (and the potential for cross-reactivity of the antisera), a HPLC separation of these metabolites before RIA of the 2,3 dinor metabolites is essential.

Prostaglandin Receptors

Previously, prostaglandin-binding sites primarily were characterized by the relative affinities for various prostaglandins,²¹ and this led to the pharmacologic description of different prostanoid receptor classes.²² The prostanoid receptors now have been cloned and sequenced, including thromboxane (TP), prostacyclin (IP), prostaglandin F (FP), and several subtypes of the prostaglandin E (EP) receptor (EP₁, EP₂, EP₃, and EP₄). These subtypes have been defined on the basis of different pharmacologic profiles and linkage to different signal transduction pathways.²³ The prostanoid receptors show common features with the G protein-coupled receptor superfamily, with seven transmembrane spanning domains. Several isoforms of the EP₃ receptor have been described, which arise through alternative splicing of mRNA and are coupled to different signalling pathways. A variety of receptors are present on myometrium.

The aforementioned background allows us to discuss the role of prostaglandins in the reproductive cycle. Although it is clear that prostaglandins play an important role in a number of reproductive processes, our understanding of the precise mechanism of action remains very basic. Additionally, the interrelationships between prostaglandins and other vasoactive substances such as nitric oxide have added to the complexity of this area. In many cases, only experimental data from isolated human tissue and experimental animals are available. However, it is clear that endogenously occurring compounds are important in clinical medicine.

The Menstrual Cycle

Peripheral plasma levels of PGF_2α and 13, 14dihydro-15-keto-PGF_2α have been measured throughout the menstrual cycle.^24,25,26,27 However, none of these studies demonstrated a cyclic change. Koullapis and Collins reported that although extremely variable, levels of 13, 14-dihydro-15-keto-PGF_2α consistently showed two peaks: one preovulatory, which was associated with increasing estrogen, and one premenstrual, which was associated with rapid decreases in estrogen and progesterone levels.²⁸ The failure to measure cyclic changes in the peripheral plasma in PGF_2α most likely reflects the fact that PGF_2α is converted quickly to 13, 14-dihydro-15-keto-PGF_2α in the lungs and other organs. The failure to routinely measure cyclic changes in the PGF_2α metabolite may reflect a simple dilutional problem. By the time uterine PGF or its metabolites reach the peripheral circulation, they have been diluted into a large blood volume.

Cyclic changes in PGE and PGF levels do occur in the uterus, as determined by endometrial and myometrial biopsies.^29,30,31,32 In the endometrium, PGE and PGF increased sequentially throughout the menstrual cycle. The endometrium of the late secretory phase contained considerably elevated PGF levels and a reversal of the ratio of PGF to PGE. These increases in PGF levels in the endometrium appear to be caused by progesterone withdrawal and have been reported for the postovulatory period.^30,32

In the myometrium, the concentration of PGF varied throughout the menstrual cycle, characterized by two peaks, one occurring during the late proliferative phase and the other in the late secretory phase. In contrast, PGE peaked in the late proliferative stage (higher levels than PGF) but then decreased in the later stages of the cycle.³³ The 13, 14-dihydro-15-keto metabolites of PGE and PGE did not show any cyclic changes in the myometrium, suggesting no cyclic alteration in PGF and PGE metabolism. The elevated levels of prostaglandins in the periovulatory period may provide a mechanism to facilitate migration of ova and spermatozoa, but this has not been proven. The change in the PGE-to-PGF ratio in both endometrium and myometrium in the late secretory phase may be important in the onset of menstruation.

Menstruation and Dysmenorrhea

The postovulatory period is associated with progesterone withdrawal and increased PGF-to-PGE ratios in the endometrium.^29,30,31 PGF_2α has been shown to be a vasoconstrictor in the uterus and endometrial blood vessels, and the local endometrial vasoconstriction that occurs during menstruation is believed to be due in part to prostaglandins.^{34,35,36,37,38}

In 1957, Pickles proposed that a substance produced by the menstruating uterus might be responsible for stimulation of uterine contractility and that this “menstrual stimulant” might be responsible for primary dysmenorrhea.³⁹ Pickles observed that menstrual fluid of patients with primary dysmenorrhea had higher PGF_2α concentration than that of nondysmenorrheic patients and that these patients had an increased PGF-to-PGE ratio.²⁹ This study was later confirmed by Lundstrom and Green and substantiated by the observation that long-term infusions of PGF_2α resulted in symptoms similar to dysmenorrhea.^40,41

Csapo and associates measured the intrauterine pressures during menstruation in healthy and dysmenorrheic patients.⁴² In contrast to the normal resting pressure of 10 mmHg, the dysmenorrheic patient had levels of 60 mmHg. This increased tone could result in vascular compression and, thus, reduced uterine blood flow. This would further compromise a uterine circulation already reduced by the direct action of PGF_2α on blood vessels in the uterus and endometrium. It is this reduced blood flow and resultant anoxia of uterine tissue that is believed to produce the pain of dysmenorrhea.²⁹ In the study by Csapo and coworkers,⁴² treatment of dysmenorrheic patients with the prostaglandin-synthesis inhibitor, naproxen sodium, reduced the resting pressure as well as the amplitude and frequency of cyclic intrauterine pressure. Naproxen and ibuprofen both are associated with a reduction of prostaglandins in menstrual blood and decreased clinical symptoms of dysmenorrhea.^43,44 Additional information is available in a recent review on nonsteroidal anti-inflammatory drugs and reproduction.⁴⁵

Treatment of dysmenorrhea with prostaglandin-synthesis inhibitors is an excellent example of integration of basic research and clinical medicine. This therapeutic method of treatment appears to be effective in 70% to 80% of all clinical cases of dysmenorrhea.

Ovulation

Ovulation has three major phases in which prostaglandins may be involved: (1) follicular maturation; (2) rupture of the follicle, leading to oocyte release; and (3) formation of the corpus luteum. The actions of prostaglandins in hypothalamic and pituitary function, which can influence follicular development, are not discussed in this chapter. In rabbits, LeMaire and associates showed that isolated follicles could be stimulated with human chorionic gonadotropin (HCG) to increase their PGE and PGF_2α levels.⁴⁶ This increase only occurred in those follicles destined to be ovulated.⁴⁷ Pretreatment of animals with prostaglandin synthesis inhibitors blocks ovulation,^48,49 although attempts to do this in humans have not been successful. It is known that rat, bovine, and human follicles predominantly synthesize PGI₂ and lesser amounts of PGE and PGF_2α and that follicular fluid concentrations increase as ovulation approaches. Addition of luteinizing hormone (LH) increases granulosa cell conversion of arachidonic acid to PGI₂ and perfusion of rat ovaries with PGI₂ induces rupture of mature follicles, although less efficiently than HCG.⁵⁰ The majority of ova ovulated by PGI₂ were immature and would not undergo germinal vesicle breakdown and maturation, thereby drawing a distinction between oocyte maturation and follicular rupture.

Prostaglandins may be intermediates in the cyclic adenosine monophosphate (cAMP) signalling pathway because ovulation induced by forskolin or dibutyryl cAMP in PMSG-primed rats could be inhibited by indomethacin, an effect itself reversed by the addition of PGE₂.⁵¹ There is a progressive increase in dispersion of cells of the mural granulosa and cumulus oophorus of preovulatory ovine follicles 12 hours after initiation of the LH surge that can be negated by indomethacin. The effect of indomethacin itself can be counteracted by PGF_2α.⁵² This suggests that prostaglandins may be involved in reduction of junctional complexes that adhere granulosa cells to each other and the oocyte before ovulation. Yoshimura and colleagues⁵⁰ found that perfusion of ovaries with PGI₂ produced vasodilation, extravasation, and filling defects at the apical region of preovulatory follicles, findings similar to those with HCG treatment, and suggested that follicular rupture may be secondary to increased capillary permeability and vasodilation due to PGE₂. Although HCG stimulated collagenase in PMSG-primed rats, which may break down the follicular wall, indomethacin had no effect on collagenase, although it reduced prostaglandin synthesis.⁵³ More recent biochemical and molecular advances in this field are reviewed elsewhere.⁵⁴

Luteinization does not appear to involve prostaglandins because pretreatment of animals with indomethacin at concentrations that block ovulation do not alter luteinization.⁴⁸

Luteolysis

Progesterone, which initially is produced in the corpus luteum, is necessary to prepare the uterus for implantation of a fertilized ova and to monitor the pregnancy until the placenta can begin progesterone production.⁵⁵ Interruption of progesterone production by the corpus luteum then leads to termination of pregnancy. In 1969, it was shown that PGF_2α caused luteolysis in rats,⁵⁶ and in guinea pigs, the uterus was shown to be the source of a luteolysin identified as PGF_2α.⁵⁷ In sheep and cows, PGF_2α of uterine origin appears to be a physiologic luteolysin.⁵⁸ The PGF_2α is transferred to the uterus by a local counter current mechanism⁵⁹ between uterine vein and ovarian artery or from uterine lymphatic vessels to the adjacent ovary.⁶⁰ In nonhuman primates and women, PGF_2α also appears to be a luteolysin.⁶¹ Only transient decreases in circulating progesterone are seen after systemic administration of PGF_2α^62,63 perhaps because of the rapid systemic metabolism of PGF_2α. Because hysterectomy in primates or women does not alter ovarian cyclicity, it was postulated that the luteolysin was generated locally within the ovary or corpus luteum.⁶⁴ However, recently, this concept has been questioned.⁶⁵

The luteolytic action of PGF_2α may be directly on luteal cells or may be indirect by directing blood flow away from the corpus luteum. PGF_2α appears to interfere with gonadotropin effects on luteal cells, being a potent antagonist of LH or HCG action in rat luteal cells and blocking LH-stimulated cAMP production,⁶⁶ although not interfering with gonadotropin binding.⁶⁷ PGF_2α inhibits basal and LH-influenced progesterone production by granulosa cells from preovulatory follicles that spontaneously luteinize in culture.⁶⁸ Further, PGF_2α has an antigonadotropic effect on LH-HCG-stimulated progesterone and cAMP formation from human luteal tissue in the mid to late luteal phase.⁶⁹

Pharriss and Wyndarden⁷⁰ first postulated that the luteolytic action of PGF_2α may be due to its vasoconstrictor action, decreasing blood flow to the corpus-luteum-bearing ovary. Subsequent work has not proven this observation but does not preclude local effects of PGF_2α causing shunting of blood away from the corpus luteum.⁷¹

Riley and Carlson⁷² demonstrated that one of the early events in PGF_2α-induced luteolysis is membrane breakdown or decrease in fluidity. This was associated with a decrease in HCG binding and HCG receptor concentrations. Further, addition of phospholipase A₂ gave a similar increased effect on membrane rigidity, suggesting that a product of phospholipase action (e.g., free fatty acid or lysophosphatidic acid) may be responsible for this ridigification during luteolysis. Interestingly, addition of prolactin, which is luteotropic in the rat, was able to attenuate PGF_2α-induced increases in membrane rigidity and decreases in plasma progesterone.

The rate-limiting step for all steroidogenic tissues is believed to be the conversion of cholesterol to progesterone. Rajkumer and coworkers⁷³ presented evidence in rats that PGF₂ decreases the ability of luteal cells to mobilize endogenous cholesterol at relatively short time intervals (4 hours) after PGF_2α. At longer time periods (24 hours), there appear to be decreased ability of these luteal cells to convert cholesterol to progesterone, which may contribute to the decreased progesterone synthesis. Cholesterol is imported into luteal cells as a lipoprotein complex by a membrane receptor-mediated process. However, luteal cells from PGF_2α-treated cows⁷⁴ and PGF_2α-treated rats⁷³ do not show decreased utilization of lipoproteins.

Thus, although we have a reasonable understanding of how prostaglandins are involved in luteolysis in some species, we still do not have a clear indication of how they work in humans.⁷⁵

Fallopian Tube Motility

PGE and PGF have opposite effects on the oviduct; PGE₂ causes relaxation, whereas PGF₂ induces contractions.^76,77 The final response of the oviduct to these prostanoids appears to depend on the steroid levels associated with the reproductive cycle. In the rhesus monkey, PGF_2α has no effect on the oviduct during the follicular phase but produces contractions just before ovulation. PGE₂ also has no effect during the follicular phase but suppresses activity during and after ovulation.^78,79 It is known that the oviduct can synthesize PGE and PGF_2α locally, but endogenous prostaglandins do not appear to be involved in regulation of spontaneous activity.^80,81 The effects of PGI₂ and 6-keto-PGF_1α have been determined in the fallopian tubes of nonpregnant women. PGI₂ is similar to PGE₂ and relaxes the fallopian tube, whereas 6-keto-PGF_1α causes contractions similar to those seen with PGF_2α.⁸² Although the role of prostaglandins in regulation of fallopian tube motility is not completely understood, it is thought that prostaglandin-increased contractile activity is important for ovum transport to the uterus.

Implantation

Before the arrival of the blastocyst, the endometrium has undergone profound changes that are regulated by estrogen and progesterone. In most mammals, these endometrial changes occur after each ovulation. The information available about the mechanism of implantation relies heavily on data acquired from laboratory animals. When the blastocyst arrives at the uterine wall, changes occur in the underlying endometrium. The endometrial vessels dilate and exhibit increased vascular permeability. The trophoblast begin to invade the endometrium, and the blastocyst undergoes apposition, attachment, adhesion and fusion (known together as implantation).

It currently is unclear whether and how the blastocyst controls the implantation process. It is clear that changes occur only in the areas of blastocyst attachment. Prostaglandins have been implicated as the mediator of the vascular permeability changes that occur at the implantation site. Prostaglandin levels have been shown to increase significantly at the site of implantation.^83,84,85 The rate-limiting enzyme for prostaglandins synthesis in some situations may be prostaglandin H synthase (cyclooxygenase). Two isoforms of cyclooxygenase exist, the constitutive COX-1 and the inducible COX-2 isoforms, which are regulated differently. Recent studies in the mouse have shown that the COX-1 gene is expressed in the uterine epithelium during the peri-implantation period but is down-regulated at the initiation of attachment.⁸⁶ This corresponds to the generalized edema in the uterus. In contrast, the COX-2 gene is expressed in the luminal epithelium and subepithelial stromal cells surrounding the blastocyst at the time of implantation. Thus, the COX-1 gene is involved with decidualization or localized endometrial changes in vascular permeability, whereas the COX-2 gene is involved with angiogenesis for the establishment of the placenta.⁸⁶ As these changes occur, expression of prostaglandin receptors EP3 and FP occur simultaneously in the circular muscle of the myometrium on days 3 to 5 of pregnancy, which probably is important for embryo transport, spacing, or accommodation in the uterus. In contrast, the expression of EP3 in a subpopulation of stromal cells on the mesometrium side and EP4 receptors in the epithelium and stroma suggest that PGE₂ prepares the uterus for implantation.⁸⁷ The source of the prostaglandins that are generated locally at the site of implantation still is not clear but is believed to be the endometrium because the blastocyst has not been shown to produce prostaglandins.⁸³

Indomethacin pretreatment prevents the normal increases in prostaglandins as well as the changes in vascular permeability in mice,⁸⁸ rats,^83,89 hamsters,⁸⁴ and rabbits.⁹⁰ Indomethacin also is effective in blocking implantation in mice,⁹¹ rats,⁹² and rabbits.⁹³ Further evidence for a role of prostaglandins is supported by the fact that intraluminal administration of PGE₂ in rats pretreated with indomethacin results in implantation-like alterations in vascular permeability,⁹¹ but this is not true for PGF_2α or PGI₂.^94,95 Although blockade of implantation would provide a potential method of contraception, it currently is not known whether similar implantation mechanisms occur in humans. A recent detailed review on prostaglandins and implantation appears elsewhere.⁹⁶

Circulating Levels During Pregnancy

Prostaglandins are produced in large quantities in a number of reproductive tissues during pregnancy, and this production may be affected by the changing hormonal milieu of pregnancy. A more detailed discussion of the site of synthesis of these compounds can be found elsewhere.^97,98,99 Studies in chronically catheterized animals and measurement of prostaglandins in uterine and placental tissue or umbilical cord blood suggest that peripheral levels of prostaglandins should be elevated significantly in pregnancy. Several attempts have been made to measure circulating levels of prostaglandins during pregnancy in humans.^100,101 Because the primary prostaglandins PGE and PGF are converted quickly to 13, 14-dihydro-15-keto-metabolites (see previous section on metabolism and measurement), it was believed that these metabolites would more accurately reflect gestation changes. Measurements of 13, 14-dihydro-15-keto-PGE₂ have not been reliable because this compound is unstable.¹⁸ Measurements of the circulating levels of 13, 14-dihydro-15-keto-PGF_2α also have failed to show any gestationally related change except during parturition, when they are increased significantly.^101,102,103

The discovery of the potent vasodilator PGI₂ and the development of RIAs to measure its metabolite, 6-keto-PGF_1α, renewed interest in measuring circulating levels of prostaglandins in pregnancy. PGI₂ is produced in vascular tissue and appears to be important in local vasodilation. During pregnancy, there is a profound decrease in peripheral vascular resistance, and the uterine vasculature is almost totally dilated. It was expected that PGI₂ levels (measured as 6-keto-PGF_1α) would be increased significantly, reaching a maximum in the third trimester. However, measurements of circulating levels of 6-keto-PGF_1α have not shown any gestationally related alterations,^99,104,105 probably because of the low circulating concentrations and interference problems in direct plasma assays. Measurement of the 2,3 dinor 6-keto-PGF_1α metabolites in urine have shown an 8- to 10-fold increase in PGI₂ synthesis during pregnancy,¹⁰⁶ although whether this increase is generalized throughout the whole vascular tree or confined to certain vascular beds (e.g., uterus and kidney) remains to be established.

It was further speculated that low production rates of PGI₂ in disease states associated with reduced uterine blood flow (i.e., hypertension, preeclampsia) might be the cause of reduced uteroplacental perfusion. Again, measurements of peripheral circulating levels of 6-keto-PGF_1α were made in patients with a number of diseases associated with pregnancy, and no differences have been observed.¹⁰⁴ In contrast, measurements have shown the ability of umbilical and placental blood vessels from preeclamptic, chronic hypertensive, and intrauterine growth retardation (IUGR) pregnancies to produce PGI₂ to be reduced considerably.^107,108 Again, urinary 2,3 dinor 6-keto-PGF_1α measurement confirmed this reduction in PGI₂ synthesis in hypertensive or IUGR pregnancies. Thromboxane synthesis also increases during pregnancy (as measured by urinary metabolites); however, the ratio of vasodilator PGI₂ to vasoconstrictor TxA₂ favors PGI₂, giving an antithrombotic tendency in normal pregnancy. Whereas, PGI₂ synthesis is reduced in hypertensive pregnancy, urinary TxB₂ metabolites appear unchanged.¹⁰⁶ The net result, however, is an alteration in the PGI₂/TxA₂ ratio toward TxA₂ and a more vasoconstrictive and prothrombotic state. Recently, most of the interest in the regulation of uterine and systemic vascular resistance in normal and hypertensive pregnancies has moved away from prostaglandins (with the exception of prostacyclin) and focused on nitric oxide. Preeclampsia is believed to be caused by endothelial cell dysfunction, and this may include reductions in both endothelial cell production of prostacyclin and nitric oxide synthesis (see discussion elswhere in these volumes).

Uterine Blood Flow

Prostaglandins can affect uterine blood flow in three different ways. First, prostaglandins can have a direct effect on the vascular smooth muscle, acting through receptors to produce either uterine vasodilation or vasoconstriction. Second, prostaglandins can increase uterine tone or contractile activity, resulting in a significant reduction in uterine blood flow by mechanical compression of the uterine vasculature. The third potential mechanism is by potentiating or by depressing adrenergic neurotransmission. This is accomplished by altering the local neurogenic release of norepinephrine from the presynaptic nerve terminal or by modifying the responsiveness of the postsynaptic nerve terminal to norepinephrine.

The uterine vasculature effects of prostaglandins have been investigated mainly in experimental animals. Clark and coworkers have shown that PGE₁ and PGE₂ produce vasodilation in the uterine vasculature of the anesthetized nonpregnant dog, whereas PGF_2α does not significantly alter uterine blood flow.^109,110 Subsequent studies in unanesthetized chronically catheterized nonpregnant sheep have shown similar responses for PGE₁ and PGE₂, whereas PGF_2α produced dose-related decreases in uterine blood flow.^111,112 Similar studies have shown that PGI₂ and PGD₂ are potent uterine vasodilators in nonpregnant sheep.^37,113 In contrast, the PGI₂ breakdown product (6-keto-PGF_1α) and TxB₂ were very weak vasoconstrictors.³⁷

The ability of prostaglandins to reduce uterine blood flow in nonpregnant sheep by increasing uterine tone and contractile activity (mechanical compression) has been documented for PGE₂ and PGF_2α.^37,111,112 In addition, PGE and PGF also have been shown to alter both presynaptic and postsynaptic sympathetic nerve transmission, thus affecting the vascular response to uterine sympathetic nerve stimulation and circulating norepinephrine.^109,111,114 The local release of prostaglandins does not appear to be under neurogenic control because levels of PGE and PGF in uterine venous effluent were not altered by sympathetic nerve stimulation.¹¹⁴

Prostaglandins act by means of receptors, and, at least in the canine uterus, PGF_2α is able to competitively antagonize vasodilation produced by other prostaglandins.¹¹⁵ This is important because the final vascular effects of endogenous prostaglandin production are dependent on the summation of the individual prostanoids. Finally, the nonpregnant uterus is composed of two separate tissue layers, the myometrium and the endometrium. Studies that evaluate uterine vascular effects of prostaglandins generally combine changes in both tissue layers. Prostaglandins have been shown to act differently in the myometrium and the endometrium.²⁷ However, specific tissue changes would not be detected if only total uterine blood flow were measured. An example of this has been reported in the dog uterus, in which PGE₁ preferentially dilates the myometrium, PGF_2α preferentially constricts the endometrium, and PGE₂ acts equally on both tissue layers.³⁶ Whether this occurs in other species or during pregnancy has not been investigated carefully.

During the ovarian cycle, estrogen levels increase just before ovulation. This elevation in estrogen is associated with a significant increase in uterine blood flow.^116,117 Prostaglandins have been implicated in this response because of their uterine vasoactivity.^118,119 Ryan and coworkers have shown that prostaglandin-synthesis inhibitors, meclofenamate and indomethacin, decrease the uterine responses to estrogen in rats.¹²⁰ In contrast, studies in sheep have shown that although estrogen increases PGI₂ levels, inhibition of prostaglandin synthesis does not alter the uterine vascular response to estrogen.¹¹⁸ These studies strongly suggest that estrogen-induced decreases in uterine blood flow are not mediated by prostaglandins. More recently, Van Buren and associates¹²¹ showed that estrogen-induced increases in uterine blood flow are mediated by nitric oxide and are prevented by inhibitors of nitric oxide synthetase.

Pregnancy creates a unique cardiovascular challenge for the maternal organism. Blood volume and cardiac output increase by as much as 40%, yet arterial blood pressure is not altered significantly. The reason for this is the development of the uteroplacental vasculature and the fact that the maternal vasculature dilates. The resistance in the uteroplacental vasculature is very low. This is due to the development of a new uteroplacental vascular bed and vasodilating hormonal factors that appear to play a major role in regulating uteroplacental blood flow during pregnancy. Prostaglandins are known to be produced in the uterus and placenta in large quantities and may be important in regulating both uterine and umbilical blood flow.^{97,99,122,123}

Early studies by Terragno and coworkers¹²⁴ and Venuto and coworkers¹²⁵ demonstrated that the pregnant uterus of dogs and rabbits produced large quantities of PGEs and that after blockade of prostaglandin synthesis, uterine vascular resistance significantly increased. In contrast, in animals that recovered from surgery for a period of 5 to 7 days, uterine blood flow initially decreased but then returned to control levels within 15 minutes.^126,127 In these studies, uterine prostaglandin synthesis was inhibited significantly by 15 minutes and remained low for the next 4 hours, even though blood flow returned to control levels.¹²⁶ These studies suggest that endogenous vasodilator prostaglandins may be important in maintaining uterine blood flow during stressful situations but may be less important in regulation of blood flow during normal pregnancy. When interpreting these results, remember that prostaglandin-synthesis inhibitors block the formation of all prostaglandins (i.e., vasodilators and vasoconstrictors) and that a complete understanding of the role of endogenous prostaglandins in regulating uteroplacental blood flow requires specific inhibition. Also, recent studies discussed elsewhere strongly suggest that nitric oxide may play a more critical role in regulating the uterine and systemic vascular resistance.

The uterine vascular effects of prostaglandins in pregnancy have been evaluated mainly in the unanesthetized chronically catheterized pregnant sheep. PGE₂ (a vasodilator in the nonpregnant animal) produced dose-related reductions in uterine blood flow in pregnant sheep but after administration to the fetus caused uterine vasodilation.^38,128,129 Whether this uterine effect (vasodilation) seen after fetal administration is the result of PGE₂ or PGE₂ metabolite is not known. PGE₂ did produce a significant increase in myometrial tone after maternal administration, and thus, decreased blood flow partially may reflect myometrial compression of the vasculature.^38,128,129 PGF_2α produced dose-related reductions in uterine blood flow and, at the highest doses tested, also increased uterine contractile activity and tone.^38,129 The initial PGF_2α-induced changes in uterine blood flow appear mainly to be the result of vasoconstriction.³⁸ In contrast, PGI₂ produced dose-related increases in uterine blood flow in late term pregnant sheep, whereas its breakdown product, 6-keto PGF_1α, was a weak vasoconstrictor.³⁸ The PGI₂-induced increases in uterine blood flow in this species, which were observed during intra-arterial infusions of PGI₂, were largely the result of increases in endometrial blood flow rather than increases in placental blood flow.¹³⁰ Studies evaluating PGD₂ also have shown it to be a potent vasodilator of the uterine vasculature in late-term pregnant sheep, but whether it is able to dilate the placental vasculature is unknown. In pregnant rabbits, PGD₂ had a biphasic response, producing vasodilation in the myometrium and vasoconstriction in the placenta.^38,131 Similar studies have not been conducted in the sheep or other species. To date, very little data exist that suggest that prostaglandins can dilate the maternal placental vasculature. This most likely is because the maternal placental vasculature is dilated maximally. If prostaglandins have a role in regulation of uteroplacental blood flow, PGI₂ would be the most likely candidate, but it would share this role with nitric oxide.

Umbilical Blood Flow

Until recently, the effect of prostaglandins on the umbilical vasculature in the intact animal has been limited to PGE₂ and PGF_2α, both of which were shown to be vasoconstrictors. PGE₂ was by far the more potent vasoconstrictor.^129,132 It is widely accepted in the clinical literature that prostacyclin is a potent vasodilator of the uterine, umbilical, and placental vasculature and that reduction of PGI₂ production is associated with preeclampsia. The decrease in PGI₂ is thought to be associated with reduced uterine and umbilical blood flow. Recently, several studies have evaluated the umbilical vascular response to PGI₂.^133,134 In both of these studies, PGI₂ produced dose-related decreases in fetal arterial pressure, increases in fetal heart rate, and decreases in umbilical blood flow that were associated with increases in umbilical vascular resistance. These data are interpreted to suggest that systemic prostacyclin administration leads to peripheral vasodilation and shunting of blood away from the placenta, thus leading to decreases in umbilical blood flow. Additionally, it is believed that because the placental vasculature is dilated maximally and has no vascular tone, prostacyclin could not produce any further relaxation. This hypothesis however was tested in two separate studies^134,135 by increasing umbilical vascular tone by simultaneous infusion of the thromboxane A₂-like compound U46619. Even under the condition of increased tone, prostacyclin (PGI₂) could not dilate the umbilical circulation. Thus, to date, there is no evidence in favor of an active prostaglandin dilator system in the umbilical circulation, although several members of the prostaglandins are effective vasoconstrictors in the umbilical circulation. Thus, the proof that prostaglandins modulate umbilical blood flow does not currently exist, and further investigation is required.

Prostaglandins and Preeclampsia

Although the aforementioned studies suggest that PGI₂ is not a dilator in the umbilical circulation, no studies have ruled out the possibility that PGI₂ produced in the vessel walls is not important in regulating vascular tone. It is clear that umbilical blood vessels from patients with preeclampsia, hypertension, and IUGR all show significantly depressed PGI₂ synthetic capabilities.¹⁰⁷ These disease states also are associated with reduced uterine or umbilical blood flow. Because of this, many investigators have made an association between the two phenomena. The hypothesis, which is not proven, is that decreased PGI₂ production in the uterine and umbilical vasculature results in reduced blood flow in these vascular beds. Preeclampsia is thought to be caused by endothelial dysfunction, which results in altered production of several vasoactive substances such as PGI₂ and nitric oxide. A second reasonable interpretation that could be made, based on the current state of our knowledge, is that reduced uterine and umbilical synthesis of PGI₂ and nitric oxide leads to increased platelet aggregation, which in turn leads to increased release of thromboxane A₂ and serotonin. Thus, the increasing levels of thromboxane A₂ or serotonin could be responsible for the significant decreases in uterine and umbilical blood flow. Thromboxane A₂ (evaluated as the thromboxane mimetic U46619) is a potent vasoconstrictor of the uterine (unpublished data) and umbilical vasculature.^134,135 Serotonin is also a potent vasoconstrictor in both the uterine and umbilical vasculature.^136,137

Based on the accepted belief that preeclampsia and reduced uteroplacental blood flow are caused by inadequate production of vasoactive substances such as prostacyclin, Fidler and colleagues¹³⁸ attempted to treat preeclamptic patients with intravenous PGI₂. In this study, the patient's diastolic blood pressure was decreased from 140 to 80 mmHg. PGI₂ was used only when other therapeutic agents failed to control the blood pressure. It should be stressed that the safety of infusing this endogenous prostanoid into pregnant women currently is not known. Uteroplacental blood flow is extremely dependent on perfusion pressure, and PGI₂ infusions that decrease mean arterial pressure significantly may lead to significant reductions in uteroplacental blood flow, which could lead to fetal hypoxia.¹³⁹

Pregnancy also is associated with a reduced vascular responsiveness to angiotensin II, both in humans and in sheep.^140,141 Before the development of pregnancy-induced hypertension (PIH), angiotensin II refractoriness is lost.¹⁴⁰ This increase in vascular responsiveness probably is not specific for angiotensin II but rather reflects increased vascular reactivity to all vasoconstrictors.¹⁴² It originally was believed that angiotensin II increased uterine blood flow and released PGEs from the uterus, which modulated uterine and systemic pressor response to angiotensin II in pregnancy.^124,125,143 This possibility has been addressed in two studies, one that showed release of prostaglandins (PGE₂ and PGI₂ measured as 6-keto PGF_1a) from the uteroplacental vasculature at high doses of angiotensin II¹⁴⁴ and one that did not show any change.¹⁴⁵ Neither study showed sufficient uterine release of prostaglandins to alter systemic responses to angiotensin II or systemic vascular resistance. However, angiotensin II did produce uterine vasoconstriction.^141,144,145 In the study that found no uterine release of prostaglandins, bradykinin released large quantities of both PGE and 6-keto-PGF_1α, serving as a positive control.¹⁴⁵ Because prostaglandins are metabolized quickly or undergo nonenzymatic breakdown, the likelihood of a circulating vasodilator prostaglandin released from the pregnant uterus being responsible for reduced systemic vascular resistance or responsiveness does not seem likely. If prostaglandins are responsible for regulating vascular responsiveness of blood vessels, this would have to occur by local vessel synthesis.

Parturition

The role of prostaglandins in parturition has been investigated extensively both in humans and in sheep. The similarities and differences in endocrine changes that occur at the time of parturition in these two species have been reviewed.^146,147,148 Although peripheral levels of most prostaglandins do not change throughout gestation, the levels of 13, 14-dihydro-15-keto-PGF_2α, the PGF_2α metabolite, are elevated during the last days of pregnancy and during labor in both species.^{99,101,102,103} In contrast, the levels of PGE₂, PGF_2α, and 6-keto-PGF_1α all are elevated in amniotic fluid during labor,^102,149 and a recent report suggests that increased amniotic fluid prostaglandins precede the onset of labor.¹⁵⁰ Increasing amounts of data strongly suggest that prostaglandins play a major role in the mechanism of parturition in both species. The observation that PGE₂ and PGF_2α, as well as their synthetic analogues (15-methyl and 16, 16-dimethyl), are effective in induction of abortions and labor and the fact that prostaglandin-synthesis inhibitors can block premature labor and delay parturition add further support to this concept.

All the intrauterine tissues produce prostaglandins; however, the tissues that produce these specific oxytocic prostaglandins that cause uterine contractility at parturition and the signals that initiate this increased production are not firmly identified. In sheep, the placenta appears to be the tissue that produces prostaglandins at parturition as a result of a signal from the fetal hypothalamic-pituitary-adrenal-placental axis.¹⁴⁷ No similar signal route occurs in human pregnancy at term, so current attention is focused on paracrine mechanisms that operate in the fetal membranes to initiate prostaglandin synthesis and stimulate labor.¹⁵¹ The amnion has been recognized as a major site of PGE₂ synthesis at the time of labor in humans.^152,153 The amnion is an extremely rich source of arachidonic acid, the precursor of prostaglandins esterified in membrane phosphatidylethanolamine or phosphatidylinositol. Whereas before labor, amnion synthesizes mainly lipoxygenase products of arachidonic acid,¹⁵⁴ during labor, PGE₂ synthesis predominates with release of substrate arachidonic acid from membrane phospholipids under the control of phospholipase A₂ being seen.¹⁵⁵ To reach the myometrium and thus cause contraction, PGE₂ first has to cross the chorion, a rich tissue source of prostaglandin dehydrogenase. However, in vitro transfer experiments have shown that PGE₂ can traverse amnion-chorion-decidua path intact, apparently by means of an extracellular route,¹⁵⁶ although the relative amount transferred is hotly debated.^157,158,159 Therefore, PGE₂ produced by the amnion could mediate uterine contractility. However, measurements of peripheral plasma prostaglandin metabolites in labor showed that PGF_2α metabolites are increased,¹⁶⁰ with little or no data available for increase in the corresponding PGE₂ metabolites. This PGF_2α does not appear to arise by reduction of PGE₂ by fetal tissues; rather, it may represent true PGF_2α synthesis. Wilson and associates¹⁶¹ recently noted that the decidua produces increased concentrations of PGF_2α during labor and that this PGF_2α is oxytocic.

In the ovine model, changes in placental estrogen/progesterone synthesis in favor of estrogen, under the influence of increased fetal adrenal cortisol, appear to regulate prostaglandin synthesis. Such changes in peripheral plasma estrogen or progesterone have not been demonstrated in humans, giving rise to the concept of local changes in estrogen/progesterone synthesis in the fetal membranes, which then locally regulate prostaglandin synthesis.¹⁶² In addition, a variety of peptide hormones (cytokines,¹⁶³ epidermal growth factor,¹⁶⁴ transforming growth factor alpha,¹⁶⁵ relaxin¹⁶⁶) have been shown to stimulate fetal membrane prostaglandin synthesis in vitro and have been suggested variously as signals for parturition. To increase prostaglandin production, arachidonic acid must be liberated from membrane phospholipids by phospholipases and the activity of PGH synthase must increase (perhaps under the influence of peptide hormones) to convert that arachidonic acid to prostaglandin. Activity of phospholipase may be regulated directly or indirectly. Gravidin, a 58,000 molecular weight protein produced by the chorion and found in amniotic fluid, recently has been described as an inhibitor of phospholipase activity.¹⁶⁷ The activity of gravidin was reported to decrease during labor, thereby paradoxically promoting an increase in phospholipase activity. Gravidin may be a member of the annexin family of calcium-phospholipid-binding proteins that inhibit phospholipase activity indirectly by binding to the substrate phospholipid. Activity of annexin I in human fetal membranes has been found to decrease at the time of parturition¹⁶⁸ consistent with an increase in phospholipase activity. Two isoforms of the PGH synthase (cyclooxygenase) enzyme have been described, the constitutively expressed PGHS-1 (COX-1) isoform and the PGHS-2 (COX-2) isoform, which is inducible by a variety of cytokines and growth factors. The cytokine interleukin-1 (IL-1) and growth factors such as epidermal growth factor (EGF), transforming growth factor (TGF), and tumor necrosis factor (TNF) all can rapidly induce PGHS-2 and stimulate prostaglandin synthesis in human amnion cells.¹⁶⁹ Whereas the specific activity of PGHS enzyme in the human amnion increases sharply before the onset of labor,¹⁷⁰ the abundance of PGHS-1 mRNA in the amnion, the choriodecidua, and the placenta does not change with labor and delivery.¹⁷¹ However, expression of PGHS-2 is increased significantly in both the amnion and the chorion at the time of labor,¹⁷² suggesting that upregulation of PGHS-2 mediates the increased prostaglandin synthesis. The use of indomethacin—a nonselective PGHS inhibitor—as a tocolytic is limited because it causes oliguria, suggesting that PGHS-2-specific inhibitors may be more useful tocolytics.

Arachidonic acid for prostaglandin synthesis by PGHS enzyme is liberated from cell membrane phospholipids by phospholipase enzymes. Recently, several isoforms of the phospholipase A₂ enzyme have been described, including the 85-kDa cytosolic phospholipase A₂ (cPLA₂), which has a preference for substrate phospholipid-containing arachidonic acid in the SN-2 position.¹⁷³ Cytokines and growth factors also rapidly induce cPLA₂ expression and activity in human fetal membranes, and IL-1β coordinately induces both cPLA₂ and PGHS-2 in amnion-derived epithelial cells for sustained PGE₂ synthesis.¹⁶⁹ Activity of cPLA₂ in the human amnion increases throughout gestation and is highest before the onset of labor,¹⁷⁴ suggesting that it plays a role in prostaglandin production at labor.

Since the first studies of Karim and coworkers and Bygdeman and coworkers in 1968, it has been clear that PGE₂ and PGF_2α can induce labor both in term pregnancies and in the first and second trimesters.^175,176 The problem with these agents was not their effectiveness but rather the persistent side-effects of vomiting and diarrhea. Attempts were made to produce analogues of these primary prostaglandins that would reduce side-effects. In the original studies, large quantities of PGE₂ or PGF_2α had to be used because of the efficient metabolism of primary prostaglandins after intravenous infusion. Thus, one of the major objectives in producing synthetic analogues was to generate compounds that were not metabolized quickly, allowing a lowering of the therapeutic dose. These synthetic attempts gave rise to the 15-methyl and 16, 16-dimethyl analogues, which have greatly improved effectiveness but still can produce some nonuterine side-effects.

Hundreds of articles have appeared that discuss the benefits and disadvantages of different routes of administration (oral, intravenous, vaginal, intra-amniotic), dosages to be used, and types of prostaglandins to be used (PGE₂, PGF_2α, or an analogue) for the induction both of labor and abortions. This area of research is still quite active because pharmaceutical corporations continue to seek the most effective agent with the fewest side-effects. Several reviews of the use of prostaglandins for induction of labor and abortion have appeared; the reader is referred to them for additional information.^177,178 In addition to normal labor, prostaglandins also have proved to be very effective in inducing labor in cases of intrauterine fetal death, anencephaly, and hydatidiform mole.^179,180,181 Reports also exist in which 15-methyl-PGF_2α has been used effectively to treat postpartum or postabortion bleeding due to uterine atony.^182,183

The final clinical use of prostaglandins to be discussed is that of cervical ripening. In early studies in which prostaglandins were used to induce abortion and labor, it was noted that the cervix seemed to dilate, soften, and efface even before strong uterine contractions. Tissue samples taken from the human cervix in the first trimester have been shown to produce PGE, PGF, the PGF_2α metabolite 13, 14-dihydro-15-keto-PGF_2α, and 6-keto-PGF_1α. This production is enhanced at term.⁹⁹ Using an animal model, Stys and coworkers have shown that intracervical administration of PGE₂ results in significant changes in compliance of the cervix of the pregnant sheep at a time when no uterine contractile activity is observed.¹⁸⁴ This study clearly demonstrates that PGE₂ can effectively change cervical compliance. In recent years, PGE₂-impregnated gels have been used effectively to ripen the cervix before induction of labor in all three trimesters. These gels are applied directly to the cervix by means of a transcervical catheter (12 hours before induction) and produce ripening of the cervix in a large percentage of primigravidas and multigravidas. Once the cervix is ripe, the outcome of labor is improved significantly. A further discussion of this clinical use can be found elsewhere.^185,186