Fetal Endocrinology

Fetal Endocrinology covers many facets of primate reproductive biology. The book discusses some thoughts on the fetoplacental unit and parturition in primates; the development and function of the human fetal adrenal cortex; and postnatum evolution of the adrenal glands of rhesus macaques. The text also describes the regulation of fetoplacental steroidogenesis in rhesus macaque; the comparative biological, immunologic, and chemical properties of the primate chorionic gonadotropins; and urinary estrogens during pregnancy in diverse species. The secretion and physiology of chorionic somatomammotropin in primates; the placental thyroid stimulators and thyroid function in pregnancy; and growth factors in fetal growth and development are also considered. The book further tackles the production and activity of placental releasing hormone; the endocrinology of parturition; and sex-determining genes and gene regulation. The text also looks into the testicular hormone production in fetal rhesus macaque; the control of pituitary gonadotropin secretion in fetal rhesus macaque; and the development of the regulatory mechanisms of the hypothalamic-pituitary-gonadal system in the human fetus. The development of the fetal adrenals in nonhuman primates and perspectives in fetal endocrinology are also encompassed. Reproductive physiologists, pediatricians, gynecologists, and endocrinologists will find the book invaluable.

• Language: English
• Publisher: Academic Press
• Release date: Oct 22, 2013
• ISBN: 9781483218991

Book preview

editing.

SOME NEW THOUGHTS ON THE FETOPLACENTAL UNIT AND PARTURITION IN PRIMATES¹

Pentii K. Siiteri and Maria Serón-Ferré,     Reproductive Endocrinology Center, Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, San Francisco, California

Publisher Summary

A striking aspect of human pregnancy is the variability of blood estrogen and progesterone levels that are consistent with a normal outcome. If, indeed, characteristic changes in estrogen and progesterone or possibly in the ratio of these two hormones are prerequisites to parturition, they have yet to be identified by measurements in maternal blood. The extreme hour-to-hour variability and the broad range of serum levels in normal pregnant women have made it difficult to detect a decline in progesterone or an increase in estrogen before parturition. The source of this variability is unknown. Rapid changes in placental synthesis and secretion related to fetal adrenal activity, alterations in uterine blood flow, or alterations in steroid binding to serum proteins may be involved. The failure to find clear-cut patterns in maternal blood does not rule out the possibility that important intrauterine hormonal changes have profound effects on the activation of myometrial contractions or cervical softening. This chapter presents a model for parturition in primates based upon an increased fetal adrenal secretion of dehydroepiandrosterone and its utilization for estrogen synthesis within the fetal membranes.

HISTORICAL BACKGROUND

The early observations in the late 1920s that human pregnancy urine contains large amounts of estrogenic activity as measured by bioassay was the first indication that the placenta is an endocrine organ. Soon thereafter, the isolation and elucidation of the structures of estrogenic and progestational steroids from this source opened the doors to modern steroid endocrinology. It was natural, therefore, that obstetricians had an early interest in the potential usefulness of hormone measurements as a means of following the progress of normal and abnormal gestation. During the next 30 years, methods for the measurement of estriol (E3), pregnanediol, and many other steroids were developed, and a vast number of publications appeared in which their levels in urine and plasma were correlated with various clinical conditions. It was not until the late 1950s that investigations explored the biosynthetic mechanisms by which the placenta produces steroid hormones. These studies were made possible by the advent of isotope labeling methods. It was soon established that cholesterol is converted to urinary estrogens when administered to pregnant women (Werbin et al., 1957).

Based on information available at that time, it was generally held that the placenta, like the adrenals or ovaries, carried out de novo synthesis of estrogens and progesterone beginning with acetate and proceeding through cholesterol. In his pioneering studies, Ryan (1959) demonstrated the extraordinary capacity of the placenta to carry out the terminal steps in this process–the conversion of androgens to estrogens by aromatase. However, several observations were not in accord with the idea that the placenta could independently synthesize estrogens. Levitz et al. (1962) were unable to demonstrate the placental conversion of acetate to estrogen in vitro. Furthermore, it was found that the placenta lacked the 16α-hydroxylase necessary for the formation of E3. A key observation was then made by Cassmer (1959), who reported that maternal excretion of estrogens dropped precipitously immediately after ligation of the umbilical cord even though the placenta remained attached to the uterine wall. In sharp contrast, however, the urinary excretion of pregnanediol changed very little until the placenta was removed, at which time that rate also declined rapidly. Interestingly, Cassmer found that estrogen excretion was maintained at high levels if the placenta was perfused with maternal blood. Shortly thereafter, Fransden and Stakeman (1961) reported that women pregnant with anencephalic fetuses excreted subnormal levels of urinary estrogens. Since the normal fetal zone is usually absent from the adrenal cortices of anencephalic fetuses, they suggested that the fetal adrenal glands provide steroidal precursors for placental estrogen biosynthesis during normal pregnancy. Although Fransden and Stakeman failed in attempts to obtain experimental support for their hypothesis, it was soon proved correct in studies emanating from Dallas (Siiteri and MacDonald, 1963), Paris (Baulieu and Dray, 1963), and Stockholm (Bolté et al., 1964). Using somewhat different methods involving radiolabeled steroids, each of these groups demonstrated that dehydroepiandrosterone sulfate (DHEAS) was efficiently converted to estrogens by pregnant women. It is of historical interest that at least one of the investigators of each of these groups had recently studied in Seymour Lieberman’ s laboratory at Columbia University, where the secretion and metabolism of DHEAS were being studied intensively. Diczfalusy and his colleagues fully exploited their opportunity to verify the Fransden and Stakeman hypothesis directly in human studies and developed the concept of the feto-placental unit (Diczfalusy, 1969). In retrospect, the use of this term has been unfortunate since it has drawn attention away from important aspects of steroid synthesis and metabolism in the fetal membranes, the decidua, and the maternal compartment. In any event, 1963 saw the beginning of a major effort to understand the biochemistry, physiology, and clinical significance of placental estrogens; this effort continues unabated nearly 20 years later.

An early study by Konrad Bloch (1945) demonstrated that administration of radiolabeled cholesterol to pregnant women gave rise to the excretion of radioactive pregnanediol. These results were interpreted simply as confirmation that the biosynthetic pathway of progesterone synthesis also involved cholesterol. The notion that placental progesterone might be derived from cholesterol in maternal plasma was not established until much later. After in vitro studies that demonstrated the placenta has a very low capacity to synthesize progesterone from acetate or cholesterol, a key paper was published by Bolté and his colleagues (Hellig et al., 1970). They showed that the specific activity of maternal cholesterol, placental progesterone, and urinary pregnanediol became virtually identical after administration of radiolabeled cholesterol to pregnant women bearing anencephalic fetuses. These observations suggested that maternal cholesterol is the precursor of placental progesterone production. More recently, Simpson and his associates (Winkel et al., 1980) have provided evidence for specific receptor-mediated uptake of cholesterol contained in plasma low-density lipoprotein (LDL) and its utilization for progesterone synthesis by trophoblastic cells. Thus, a considerable body of evidence suggests that the fetus does not contribute significantly to progesterone synthesis. However, this possibility has not been completely ruled out. Since the fetal adrenal glands secrete large quantities of pregnenolone sulfate (P5S), a potentially important fraction of total placental progesterone may be derived from this source.

It is well established that the maternal levels of estrogens increase and those of progesterone decline shortly before parturition in many species, including sheep, goats, rats, and cows. The original observations of Liggins (Liggins, 1969; Liggins et al., 1973) demonstrating the role of the fetal pituitary-adrenal axis in the initiation of parturition in the pregnant ewe have led to an extensive search for similar mechanisms whereby the fetus signals its readiness to enter the extrauterine environment. In sheep, increased fetal adrenal secretion of cortisol late in gestation induces the enzymes in the placenta necessary for estrogen synthesis from progesterone (Anderson et al., 1975). Similar mechanisms appear to operate in the goat, in which the corpus luteum is the principal source of progesterone throughout gestation (Thorburn et al., 1977). Thus, a distinction can be made between those species in which fetoplacental estrogen synthesis occurs throughout gestation (e.g., human beings) and those lacking a fetal zone in the adrenal gland, in which the placenta synthesizes estrogen from progesterone only late in gestation. The induction of placental 17α-hydroxylase and 17, 20-lyase activities also results in progesterone withdrawal in sheep and goats before parturition. These cortisol-induced reciprocal changes in estrogen and progesterone are thought to elevate prostaglandin (PG) synthesis within the uterus and thus to be essential to parturition. Whether the increase in PG synthesis requires steroid changes or attainment of a critical estrogen:progesterone ratio is not clear.

While fetal adrenal activity in primates also increases during the last 7 to 14 days of gestation, there is little evidence that cortisol induces comparable changes in steroidogenesis in the placenta. The remarkable increase in the size of the fetal adrenal zone during the last few weeks of gestation in both humans and monkeys results in a dramatic increase in DHEAS levels in the rhesus fetus (Serón-Ferré and Jaffe, 1981). Although similar human data are not yet available, increased adrenal secretion of estrogen precursors by the human fetus late in gestation may also be inferred from the late surge in serum E3 described by Buster and his associates (1980). Thus, increased secretory activity of primate fetal adrenals increases the supply of estrogen precursors (DHEAS) to the placenta directly. Whether or not increased DHEAS or cortisol levels lead to decreased placental progesterone synthesis is not clear, although a possible mechanism will be discussed later.

A striking aspect of human pregnancy is the variability of blood estrogen and progesterone levels that are consistent with a normal outcome. If, indeed, characteristic changes in estrogen and progesterone or possibly in the ratio of these two hormones are prerequisites to parturition, they have yet to be identified by measurements in maternal blood. The extreme hour to hour variability and the broad range of serum levels in normal pregnant women have made it difficult to detect a decline in progesterone or an increase in estrogen before parturition. The source of this variability is unknown. Rapid changes in placental synthesis and secretion related to fetal adrenal activity, alterations in uterine blood flow, or alterations in steroid binding to serum proteins may be involved. The failure to find clear-cut patterns in maternal blood does not rule out the possibility that important intrauterine hormonal changes have profound effects on the activation of myometrial contractions or cervical softening.

In this brief review, we shall only consider selected aspects of these problems in order to point out areas of agreement, disagreement, and ignorance. We will focus primarily upon human beings, and to a lesser extent on rhesus macaques (Macaca mulatto) in an attempt to identify unifying concepts. We will propose a model for parturition in primates based upon increased fetal adrenal secretion of DHEAS and its utilization for estrogen synthesis within the fetal membranes. It is not possible to address in depth the enormous literature on this subject that has accumulated over the past 20 years. However, many excellent reviews of various aspects of the endocrinology of pregnancy are available (Karim and Hillier, 1979; Liggins, 1979; Thorburn and Challis, 1979; Serón-Ferré and Jaffe, 1981; Simpson and MacDonald, 1981).

PLACENTAL ESTROGENS

The rabbit blastocyst has been shown to produce estrogens prior to implantation (George and Wilson, 1978), but the extent to which estrogen synthesis occurs in other species is not known. While the physiological significance of early estrogen synthesis is not clear, it may play an important role in nidation (Dickmann et al., 1976). In women, the major source of both estrogen and progesterone during the first 5 to 6 weeks of gestation is the corpus luteum stimulated by human chorionic gonadotropin (hCG). Thereafter, the placenta produces sufficient quantities of estrogen and progesterone so that the corpus luteum becomes dispensable. The human placenta lacks steroid 17α-hydroxylase activity throughout gestation and therefore cannot convert progesterone to estrogens. However, the placenta has a large capacity for aromatization of C-19 steroids such as androstenedione and testosterone, which are converted to estrone (E1) and estradiol (E2) (Ryan, 1959). It is well established that E1 and E2 arise from extraplacental DHEAS that is derived about equally from the fetal and maternal circulatory systems (Siiteri and MacDonald, 1966). Steroid metabolism in the human fetus is characterized by extensive formation of steroid sulfates since sulfokinase is present in many tissues, whereas sulfatases are virtually absent (Diczfalusy, 1969). In contrast, the placenta contains high levels of sulfatase activity (Pulkkinen, 1961; Warren and French, 1965) so that steroid sulfates coining from the fetus or mother are extensively hydrolyzed and made available for estrogen synthesis. Women are unique in their capacity to produce extraordinarily large quantities of E3 during pregnancy. The explanation for this capacity lies in the fact that many fetal tissues efficiently hydroxy late DHEAS to form 16α-OH-DHEAS, which is converted to E3 when it reaches the placenta via the umbilical circulation. About 90% of the total E3 is derived in this way, and the remaining 10% arises by the same mechanisms operating in the maternal compartment (Siiteri and MacDonald, 1966). At term of a normal human pregnancy, the production rate of E3 is about 50 to 100 mg per day, whereas that of E2-E1 ranges from 10 to 20 mg per day. That these massive quantities of estrogens are dependent upon normal fetal adrenal function is clear since E3 values are only 5 to 10% of normal in pregnancies associated with anencephaly (Frandsen and Stakeman, 1961).

Four enzyme-catalyzed steps are involved in the conversion of DHEAS to E1 and E2 or 16α-OH-DHEAS to E3 in trophoblastic cells. The first step is the cleavage of the sulfates by the potent steroid sulfatases. Studies on the subcellular localization and specificity of sulfatase activity have shown that the highest activity toward DHEAS and P5S is present in microsomal fractions, although significant amounts occur in other subcellular and soluble fractions of the placenta (Pulkkinen, 1961; Warren and French, 1965). Free dehydroepiandrosterone (DHEA) and hydroxylated DHEAS (16α-OH-DHEAS), from their sulfates within the trophoblast, are converted to the corresponding Δ⁴,³ ketosteroids via placental 3β-HSD and Δ⁵ − ⁴ isomerase. These enzymes are found as a tight complex in both mitochondrial and microsomal fractions of human placenta. Recent studies (Blomquist et al., 1978; Gibb, 1979) indicate that the Michaelis constant (Km) values of both DHEA and pregnenolone for microsomal 3β-HSD are in the nanomolar rather than the micromolar range, as had been reported in many earlier publications. This discrepancy is explained by the use of mercaptoethanol or other reducing agents that prevent oxidation of sulfhydryl groups, which appear to be essential for tight substrate binding. The resulting products, androstenedione and 16α-OH-androstenedione, or their 17β-reduced counterparts, testosterone and 16α-0H-testosterone, are the proximal precursors for estrogen formation. The placenta contains large quantities of 17β-oxidoreductase activity that appears to act on both C-19 and C-18 substrates. Despite the fact that this is one of the most studied steroid-metabolizing enzymes, its physiological significance is unclear. At one time it was believed to be important in estrogen action. According to this concept, E1 and E2 act as cofactors in the transfer of reducing equivalents between pyridine nucleotides.

The aromatase (estrogen synthetase) complex is located in the endoplasmic reticulum of trophoblastic cells (Ryan, 1959). While activity has been observed in mitochondrial preparations (Shaw et al., 1969; Aleem et al., 1970; Renwick and Oliver, 1973), this phenomenon seems most likely due to contamination by microsomes. The conversion of androgens to estrogens, perhaps the most complex of steroid biosynthetic reactions, involves three hydroxylation steps (Fig. 1). The first two occur at the C-19 angular methyl group, which produces a diol that readily dehydrates to form an aldehyde. The final and rate-determining hydroxylation step appears to occur at the 2 β position of ring A, which produces a product that readily collapses nonenzymatically to an estrogen (Goto and Fishman, 1977). The development of a rapid and convenient radiometric assay permitted detailed kinetic studies on the reaction (Thompson and Siiteri, 1974a). Conversion of a C-19 androgen to estrogen requires three molecules each of molecular oxygen and NADPH. Thus, the enzyme system appears to be a mixed function oxidase involving cytochrome P450 and NADPH P450 reductase (Thompson and Siiteri, 1974b). The aromatization of C-19 substrates is unusual, however, in that it is not inhibited by CO, whereas the aromatization of 19-norandrostenedione is sensitive to CO inhibition. Recombination of digitonin-solubilized placental microsomal fractions enriched either in P450 or NADPH-cytochrome reductase activities restored aromatase activity. These results support the role of P450 in aromatization (Thompson and Siiteri, 1976). Further evidence of the involvement of cytochrome P450 has been obtained with various inhibitors, including aminoglutethimide and SKF525A, and an antibody raised against porcine hepatic NADPH P450 reductase. In addition, binding of androstenedione and 19-nortestosterone to placental microsomes yields a type I optical difference spectrum, an indication that substrate binding occurs to a low-spin ferric form of P450 (Thompson and Siiteri, 1974a). Therefore, it seems that a substantial proportion of cytochrome P450 in placental microsomes is involved in reactions involving aromatase substrates. Since 19-nortestosterone is a competitive inhibitor of C-19 aromatization and of the binding of androstenedione to P450, it would appear that a single enzyme exists for both classes of substrates (Thompson and Siiteri, 1974a). It is likely, therefore, that the series of three reactions takes place at a single enzyme binding site without the release of intermediates. Recent studies by Ryan and co-workers indicate that the Km for aromatase substrates is around 10 nM, i.e., is much lower than previous estimates (Barbieri et al., 1981).

Fig. 1 Conversion of C-19 androgens to estrogens by aromatase.

Relatively little is known about the regulation of placental estrogen synthesis. Recent studies on rat ovarian cells indicate that synthesis of aromatase is stimulated by follicle-stimulating hormone (FSH) (Erickson and Hsueh, 1978; Gore-Langton and Dorrington, 1981). Of particular interest is the recent report suggesting that prolactin can block the effect of FSH (Dorrington and Gore-Langton, 1981). Several investigators have reported that hCG stimulates the conversion of C-19 steroids to estrogens in the perfused human placenta (Tominaga and Troen, 1967; Cedard et al., 1970; Alsat and Cedard, 1973; Wolf et al., 1977). Furthermore, aromatase activity of human choriocarcinoma cells is stimulated by dibutyryl cyclic AMP and theophylline (Hussa et al., 1974). Nonetheless, little evidence exists for regulation of human placental estrogen biosynthesis by tropic hormones in vivo. Since the capacity of the placenta to convert androgens to estrogens in vivo is very large, it is generally assumed that the principal mechanism for the regulation of estrogen synthesis depends on substrate (DHEAS) availability. It is of interest, however, to consider the rate-limiting step (or steps) in this process.

In vitro studies suggest that human placental sulfatase and 3β-HSD/Δ⁵ − ⁴-isomerase activities greatly exceed that of aromatase. Indeed, aromatase is a very sluggish enzyme whose turnover number is about 1.0; most enzymes have values of 1,000 or more. One would expect, therefore, that intracellular aromatase substrates (androstenedione, testosterone, 16-OH-androstenedione, and 16-0H-testosterone) are present in large excess and that aromatase, operating at or near maximum velocity, would be the rate-limiting step in conversion of DHEAS to E2. Results of in vivo studies in which an intravenous 50-mg bolus of DHEAS was administered to women late in pregnancy appear to confirm this expectation (Klopper et al., 1976; Tulchinsky et al., 1976). The plasma concentration of both DHEAS and androstenedione increased 25- to 30-fold 5 min after DHEAS administration and rapidly declined thereafter. Much smaller increases in E2 (three- to fourfold) and E1(twofold) were observed at 30 min after which their levels gradually declined. These results suggest that the in vivo rate of aromatization is at least 10-fold lower than the rate of DHEAS to androstenedione conversion. In contrast, ACTH treatment of a woman pregnant with an anencephalic fetus resulted in a fourfold increase in both DHEAS and E2 production (MacDonald and Siiteri, 1965). Although several explanations for these disparate results may be put forth, stimulation of placental aromatase activity by an ACTH-induced elevation of cortisol levels, as has been suggested for the sheep placenta (Ash et al., 1973), could explain the proportionate increase in DHEAS and E2 production. It is becoming increasingly apparent that estrogen synthesis is regulated by aromatase. A cortisol-induced increase in aromatase activity may therefore have far-reaching implications for parturition in many species (see paragraphs that follow). Further studies on the regulation of estrogen synthesis in the placenta are clearly needed.

The physiological role of placental estrogens formed during human pregnancy are poorly understood. As already indicated, the amounts produced in normal gestation appear to be far in excess of those required for any known process, with the possible exception of stimulation of uterine blood flow. The obviously important effects on distant target tissues such as the breasts are achieved in other species with much more modest levels of estrogen. A presumption inherent in this idea is that total plasma levels can be equated with that amount of estrogen available to cellular receptors in target cells, and we know this almost certainly is not true since the serum binding of estrogens varies greatly among species. If the formation of E3 confers some reproductive advantage upon human pregnancy it does not depend on the massive quantities that are normally produced since the gestation of anencephalic fetuses as well as pregnancies associated with placental steroid sulfatase deficiency proceed to term