Winnifred B. Cutler*, Celso R. Garcia +
*Leidy Laboratory of Biology, University of Pennsylvania, Philadelphia, PA 19104 U.S.A., and
+ Chairman of the Division of Human Reproduction, Hospital of the University of Pennsylvania, Philadelphia, PA 19104, U.S.A.
(Received 23 March 1979)
Psychoneuroendocrinology, Vol. 5, pp. 89 to 111.
c. Pergamon Press Ltd. 1980. Printed in Great Britain.
THE REPRODUCTIVE implications of fertility have burdened mankind through the ages and there are still no all-encompassing satisfactory solutions. The fertile seek birth control; the infertile seek conception; and both are often disappointed. The current state of knowledge about the reproductive physiology of woman involves the consideration of anatomical and physiological aspects of the reproductive tract, its interplay with the endocrine systems, and the influences of the environment, albeit without unanimous appreciation of their inter-relationships. A study of the physiology and also of the relationship between behavioral and reproductive physiology may, ultimately, provide answers to the problem of effective treatment of fecundity in humans.
It is experimentally validated that mammalian sperm have far greater viability than the ovulated egg. Most species studied show maximal sperm life, post ejaculation, of 2-3 days, while the ovum, by contrast, appears to have an upper limit of viability of 6 hr (Hammond & Asdell, 1926; Dauzier & Wintenberger, 1952; Soderwall & Blandau, 1941). Thus, to maximize fertility, copulation should preceed ovulation and hence the timing of sexual behavior is an important factor in determining the capacity of a pair to reproduce. This review of the ovulatory cycle of woman will emphasize the timing of those endogenous and exogenous events which are currently thought to affect fertility.
It will be shown that the ovulatory cycle is subject to many influences which can alter its inherent rhythmicity. The cyclicity can be viewed as displaying a pattern which is endogenously driven, but which is also subjected to exogenous influences that alter this basic rhythm. The pertinent literature is reviewed in two sections.
Section 1 discusses seven major aspects of the endogenous nature of the primate ouvlatory cycle: the natural rhythmicity of the ovaries; steroid production and control; gonadotropin production and control; prolactin production, effects and control; monoamine control of the gonadotropins; autonomic innervation of ovarian tissue, metabolic influence on ovarian cyclicity, and concludes with a summary of the endogenous aspects of the ovulatory rhythm.
Section 2 reviews three exogenous influences of fertility: nutritional influences; geophysical influences; and childbearing and lactation influences.
1. THE ENDOGENOUS NATURE OF THE OVULATORY RHYTHM
1.1 Natural rhythmicity of the ovaries
(1) that gonadotropin release is inherently pulsatile,
(2) that the immediate effect of estrogens is to inhibit pulsatile gonadotropin secretion, thereby maintaining basal plasma levels,
(3) that the ovary controls gonadotropin release to some extent, and
(4) that the ovary either has its own inherent rhythm, or that the rhythm is controlled by brain centers which trigger gonadotropin release.
With regard to the last point, recent work from Knobil's laboratory (personal communication) indicates that the rhesus menstrual cycle is not primarily a brain function since mid-line hypothalamic lesions, shown to be the site of releasing hormone synthesis in some mammals (McCann, 1962; Kamberi, Miccal & Porter, 1969; Bhattahcharya, Diershcke, Yamaji & Knobil, 1972; King, Arimura & Williams, 1975), coupled with exogenous pulsatile luteinizing hormone releasing hormone (LHRH) infusion, results in maintenance of 28 day menstrual cycles. In contrast, the rat requires an intact preoptic area to maintain its cycle (Gray, Sodersten, Tallentire & Davidson, 1978). This is further evidence that the primate cycle, unlike the rodent cycle, is partially controlled by extracentral neurohumoral factors, with the ovary as the likely source. Thus the primate ovary would appear to have its own endogenous rhythm which controls, rather than is controlled by, the central nervous system.
A more direct demonstration of an inherent rhythmicity in primate ovaries was provided by Goodman, Nixon, Johnson & Hodgen (1977) who showed by in vivo aspiration of the dominant follicle that a follicle is preselected from its cohort since removal of the dominant follicle does not result in the maturation of another one and that new follicles required about 12.6 + 0.11 days to develop to an ovulable state, regardless of the timing of the invasive manipulation.
Thus, in experiments in which the ripening follicle was ablated anywhere from day 8 to 1, (2) or where the corpus luteum was ablated 4-6 days after the midcycle surge of LH, the results were the same - a new follicle required on average 12.6 days to develop. It is interesting to note that this 12.5 day span for development of a new follicle was also described for healthy women who received experimental manipulations where ethinyl estradiol was administered for the suppression of follicle-stimulating hormone (FSH) and LH. Upon cessation of the use of this synthetic steroid there was a 12 day delay before a new rise of LH began (Tsai & Yen, 1971). In light of Goodman's report, this latent effect might represent the time required for a developing primate follicle to reach that ovulable stage where the estrogen feedback is sufficient to evoke to LH peak. Further support for Goodman's observation was provided by Bair, Baker, McNatty & Neal (1975) who argued that the time taken for developing follicles to mature was similar among the mammals studied: mouse, ewe, cow, old world primates, sheep, pig, goat and horse, limited to a 10-17 day range and perhaps specified uniquely for each species. Similar evidence for the time span for a follicular development has been provided for the ewe (Mallampati & Casida, 1970). Hence the work of the two laboratories discussed above indicates the endogenous nature of the control of the primate ovarian cycle. In other words, the ovary has its own natural cycle. The morphological picture supports this ovarian rhythmicity. Cross-sectional analysis of ovarian morphology has established a predictably orderly change of follicular and luteal development throughout the menstrual cycle in both rhesus (Koering, 1969) and human (Block, 1951a, b). Evidence for the control of the ovary by neurohormonal influences will now be presented.
1.2 Steroid production and control
1.2.1 Source and cycle of steroid secretions.
From the initial abundant supply of follicles at birth, estimated to be some 200,000 , a new crop begins to develop in the late luteal phase of each menstrual cycle. During the follicular phase several follicles continue increasing in both cell number and in size to an average maximum of about 6 mm. The increasing size is directly associated with an increase in follicular fluid (Block, 1951b; Koering, 1969; McNatty, Sawers & McNeilly, 1974; Baird et al., 1975).
This primate follicular fluid contains many substances, among which are the pituitary hormones FSH, LH and prolactin (PRL), luteinization and oocyte inhibitory substances, the estrogens 17B estradiol and estrone, and progesterone, all of which appear to be involved in the control of ovarian activity (Channing & Tsafriri, 1977). The source of estradiol in the primate has been established experimentally (Goodman et al., 1977). The growing follicles are the main source of estrogen during the follicular phase of both human and rhesus monkeys; and in humans, cells of the corpus luteum, most probably the theca interna, provide the estradiol during the luteal phase.
The rhesus monkey, in contrast, does not show marked estrogen peaks in its luteal phase (Knobil & Plant, 1978). The preovulatory estrogen peak resulting from maximal steroid output of the developing follicle has been shown to occur either simultaneously with, or 2 hr before, the LH peak; and follicular rupture has been shown to occur in women 12-24 hr following this LH peak (Yussman & Taymor, 1970; Mishell, Nakamura, Grosignani, Stone, Kharma, Nagata & Thorneycroft, 1971).
It is now clear that progesterone is secreted by the large pre-ovulatory Graafian follicle in circulating quantities of approximately 2 ng/ml in human (Baird et al., 1975; Channing & Tsafriri, 1977; Greenwald, 1978). Luteal phase progesterone secretion reaches levels approaching 20 ng/ml in mid-phase of the corpus luteuem lifespan (Garcia, 1978).
Hence progesterone circulation is limited to the last few days of the follicular phase and the entire luteal phase, at a level which reflects the development and subsequent regression of the (granulosa) cells of the late Graafian follicle and subsequent corpus luteum (CL). The CL shows a fixed lifespan with progesterone secretion reaching peak levels (15 ng/ml or more) 6-8 days after ovulation and declining thereafter to a minimum (< 1ng/ml) by 13-15 days (Ross, Cargille, Lipsett, Rayford, Marshall, Strott & Rodbard, 1970; Mishell et al., 1971; Tsai & Yen, 1971; Cargille, Vaitukaitis, Bermudez & Ross, 1973; Baird & Fraser, 1975).
Thus,the two key steroids in the ovarian cycle, estrogen and progesterone, are characterized by a cycle fluctuation which is determined by a minimum 12.5 day follicular development span and a subsequent short-lived corpus luteum. In humans, estrogen is secreted in both the follicular and the luteal phase; predictably peaking near the end of the follicular phase, when follicle size is the greatest and again at the zenith of CL development in mid-luteal phase. Meanwhile, progesterone is at very low levels throughout most of the follicular phase and peaks only once; in the mid-luteal phase.
1.2.2 Control of the steroid cycle; follicular phase.
Control of the variable length of the follicular phase appears to be a function of increased output of estrogen. This estrogen output reflects follicular maturation, but details of the cellular mechanism are currently unknown. Substances within the follicular fluid appear to control maturation (and inhibition) of oocyte and follicle (Channing & Tsafriri, 1977). Prostaglandins have also been implicated in the ovulatory process in rhesus monkeys. One function of LH may be to stimulate the local production of prostaglandins within the ovary which invariably precedes ovulation (Wallach, de la Cruz, Hunt, Wright & Stevens, 1975b; Wallach, Bronson, Hamada, Wright & Stevens, 1975a, Hamada, Bronson, Wright & Wallach, 1977). Exogenous estrogen, given to rhesus monkeys, produces an LH surge 36 hr later, and ends the follicular phase by triggering ovulation (Knobil & Plant, 1978). This experimentally induced 36 hr lag in response to exogenous estrogen administration mimics the naturally occurring pattern in human and other primates. Inevitably, estrogen increase precedes the LH surge which triggers ovulation (Mishell et al., 1971). It would be expected that exogenous administration of LH might trigger ovulation if a sufficiently developed follicle were present, although naturally occurring preovulatory type LH surges are not found in primates in the absence of an increase in estrogen level.
1.2.3 Control of the steroid cycle; luteal phase. Control of the length of the luteal phase is brought about by a mechanism different from that described above. Once a corpus luteum appears, it proceeds through a sequence of development and subsequent regression over a period of 14+ 2 days, although a requirement for low levels of LH has been indicated. In studies on hypophysectomized women, it was shown that low LH levels were necessary to maintain the luteal phase after the delivery of an LH surge had initiated ovulation (VandeWiele, Bonumil, Dyrenfurth, Ferin, Jewelewicz, Warren, Rizkailah & Mikhail, 1970).
The predictable demise of the corpus luteum is thought to be due to luteolytic factors, with prostaglandins the likely candidate. These luteolytic factors were at one time thought to originate in the uterus in response to increasing levels of progesterone acting on the glandular surfaces of the organ (Reynolds, 1949), but recent evidence points to estrogen interaction with prostaglandins. Howe (1965) showed that uterine tissue is necessary for normal luteolytic activity in the guinea pig by performing partial and complete hysterectomies. Complete hysterectomies prolonged maintenance of luteal activity for periods of up to 7 months; whereas partial hysterectomy led to partial loss of degenerative effects in the corpus lutuem. The more uterine tissue removed, the longer the life span of the corpus luteum. Similar results were obtained in sheep where it was inferred that the secretion of progesterone by the CL is responsible for stimulating the uterus to produce luteolytic amounts of prostaglandins (PG f2a). Removal of the uterus led to maintenance of a constant CL which did not regress (McCracken, Baird & Goding, 1971). Therefore, the glandular endometrium was presumed to supply a luteolytic substance, since, in the absence of the uterus, luteolysis was retarded.
1.2.4 The Endometrium responds to progesterone. In the human, Rock & Barlett (1937) showed a connection between the state of the endometrium and the number of days elapsing since ovulation. A later publication clarified these findings in graphic form, and correlated the various physiological states with the days of the cycle (Noyes, Hertig & Rock, 1950). The demonstration that a serum sample containing more than 3 ng/ml progesterone was always accompanied by a secretory (luteal phase) endometrium furthered the understanding of this relationship (Israel, Mishell, Stone, Thorneycroft & Moyer, 1972). Rosenfeld & Garcia (1976) demonstrated a high correlation (93%) between endometrial histology and plasma progesterone levels for all the progesterone producing days of the cycle.
1.2.5 Luteal phase steroid secretion is controlled by LH. Experiments using sheep and mouse preparations have shown that the stimulus for luteal progesterone secretion in these species is clearly LH since neither PRL nor FSH suffice to initiate follicular or luteal steroidogenesis (Domini, Puzzoli, D'Alessio, Lunenfeld, Eshckol & Parlow, 1966; Eshkol & Lunenfeld, 1967; McCracken et al., 1971). It has also been shown that human follicles containing high FSH levels also contain high levels of estrogen, and vice-versa (McNatty, Hunter, McNeilly & Sawers, 1975). Thus, the findings that FSH levels vary in concert with estrogen, and LH stimulates steroid secretion (while FSH will not), argues for LH being the principal gonadotropic factor in the promotion of luteal phase steroid production.
1.2.6 Estrogen, rather than progesterone, appears to be the steroid of luteolysis- in conjunction with prostaglandins. The link between luteal phase steroid production and subsequent luteolysis has been further clarified by recent reports implicating estrogen, rather than progesterone, as the main causal agent in primate PG f2a production and subsequent cycle termination. It will be remembered from the above that uterine PG was presumed to lead to luteolysis in some mammals. In the monkey, treatment with PG f2a, in combination with estrogen shortened the luteal phase significantly (Shaihk, 1972; Shaihk & Klaiber, 1974).
In human endometrial tissue similar results have been determined. Estrogen applied to proliferative phase tissue in vitro yields a 4 fold increase in PG f2a output, whereas progesterone alone is less effective (Cane & Villee, 1975). Thus, estrogen together with PG f2a induces luteolytic effects.
It is interesting to note that the following observations suggest the existence of a direct vascular link between the uterus and ovary:
1. PGs are inactivated in the lungs during the normal route of circulation.
2. Subcutaneous PG f2a administration results in an ovarian response.
3. Peripheral vein infusion does not lead to a response.
4. Uterine tissue produces luteolytic PG f2a. Therefore it would appear that uterine PG f2a must ravel to the ovary via a vascular (arteriolar) system rather than via the venous return. Such a system has been recently reported (Muckle, 1977).
1.2.7 Summary of the control of the steroid cycle and apparent differences among mammals. The above evidence points to an orderly mechanism. The production of estrogen and progesterone by the ripened follicle is followed by a priming of the endometrium of the uterus by the CL. This endocrine gland, the primed endometrium, then produces a luteolytic factor, a PG, which contributes to the regression of the CL. Hence it would appear that through its effects on the uterus, the corpus luteum contains the seed of its own destruction. These conclusions, however, should be viewed with caution since relevant data on human hysterectomy seem at first to be contradictory. In one report, five hysterectomized women were tested for 20 weeks after hysterectomy by weekly plasma sampling: plasma progestins continued to cycle in the normal rhythm of preoperative women: and LH levels also seemed to cycle in the normal pattern. However sampling at these weekly intervals was insufficient to establish definitive results on gonadotropin (Doyle, Barclay, Duncan & Kirton, 1971). Similar studies showed that post-hysterectomy ovaries continued to secrete steroids in a cyclic manner (Beavis, Brown & Smith, 1969). Moreover, very recently a large sample long-term evaluation showed that 5 yr after hysterectomy, 42% of the 1263 patients had cycling ovaries (Ranney & Abu-Ghazaleh, 1977).
1.2.8 Why are humans different. i.e. uterus not always needed for luteolysis. One explanation for the seemingly disparate results for humans considers the fact that the f series PGs have been isolated in both ovarian tissue (Wallach et al., 1975b) and in seminal fluid of rhesus and human (Samuelsson, 1963; Brummer & Gillespie, 1972). Either route could thus provide an alternative source of this hormone since humans and the primates studied here, in contrast to other domesticated mammals, have both been shown to accept the male sexually throughout the cycle (Rowel, 1963; Udry & Morris, 1968). Hence species which have alternate PG sources: i.e. from copulation, might be less affected by uterine loss. More research is clearly needed to understand the role of prostaglandins in the control of the steroid cycle.
1.3 Gonadotropin production and its control by steroids and LHRH
The gonadotropin output of the pituitary shows a similar pattern for both FSH and LH levels throughout the menstrual cycle of humans but is slightly different in other primates. FSH peaks at a lower level and slightly before LH in both human and rhesus monkeys, and both glycoproteins fall to their lowest values in the human luteal phase; rhesus monkeys, lacking a luteal phase estrogen peak may show a higher gonadotropin level at that time (Neil, Johansson, Datta & Knobil, 1967; Midgley & Jaffe, 1968; Mishell et al., 1971; Jaffe, Yen, Keye & Midgley, 1973; Niswender & Spies, 1973; Sherman, West & Korenman, 1976).
1.3.2 Mechanisms of steroid control of gonadotropin surges appear to rest in the gonadotrope cell. Recent data (Yen & Lein, 1976) indicate that the paradoxical positive and negative feedback effects of estrogen at different times in the menstrual cycle can be explained by a cellular mechanism which does not depend on separate 'tonic' and 'cyclic' hyptothalamic centers. They suggest that in both hypogonadal and normally cycling women the concept of LH synthesis and storage in two pools offers a coherent explanations for data reported by a number of laboratories including their own. They indentify the 'first pool' with the release portion of the pituitary gonadotrope cellular response, and the 'second pool' with the synthesis and cytosol storage activities of the gonadotrope cells. They have shown that luteinizing releasing factor (LRF) stimulates both pools. Estradiol (E2) stimulates the 'second pool' while blocking the first, perhaps through a membrane effect. This blocking of the 'first pool', which would otherwise result in gonadotropin release by E21, augments the cellular reserve of LH.
In addition, estrogen increases the known self-priming effect of LRF; and in so doing enhances the pituitary response to LRF. This leads to increased production and storage of LH as estrogen levels rise during the follicular phase of the menstrual cycle. Negative and positive feedback of E2 are thus shown to originate in different stages during the continuous variation of the circulating LH in response to increasing E2 levels, rather than from separate disparate 'tonic' and 'cyclic' centers. The sudden surge of LH may be due to the 'spillover' of stored LH when supramaximal gonadotropin is held within the gonadotropin cells.
It is interesting to note that the above hypothesis of Yen and Lein accounts nicely for a wide variety of reports in the literature that had seemed to be at variance. The observations on humans listed below, all support this two pool concept; namely a predictable and orderly gonadotropin cell response to E2 and LHRH signals (Kastin, Zarate, Midgley, Canales & Schally, 1971; Tsai & Yen, 1971; Vaitukaitis, Bermudez, Cargille, Lipsett & Ross, 1971; Keller, 1972; Yen, Vandenberg, Rebar & Ehara, 1972a; Cargille et al., 1973; Reyes, Winger, Rochefort & Faiman, 1977; Spellacy, Cantor, Kaba, Buhi & Birk, 1977). While two different releasing hormones have been postulated, FSHRH and LHRH, only one decapeptide has been isolated and synthesized which has releasing effects on both gonadotropins (Schally, Arimura, Kastin, Matsumo, Baba, Redding & Nair, 1971; Amoss, Burgis, Butcher, Amoss, Ling, Monohan, Rivier, Fellows, Blackwell, Vale & Guillemin, 1972). The control of releasing hormones is discussed later.
1.4 Prolactin production effects and control
1.4.1 Prolactin secretion across the cycle.
Among normally cycling women, follicular prolactin in small follicles is apparently maintained at an approximately constant level (<25 ng/ml) throughout the cycle (Jaffe et al., 1973 McNatty et al., 1974). Large follicles, in contrast, contain the lowest PRL concentrations which approach 10ng/ml (McNatty et al., 1975). Although follicular concentrations do not vary, peripheral levels show a diurnal rhythm (Robyn, Delvoye, Nokin, Vekemans, Badawi, Perez-Lopez & L'Hermite, 1973) which is sleep dependent rather than photoperiodic in character (Sassin, Frantz, Kapen & Weitzman, 1973). In addition to this daily rhythm there may be a peripheral PRL peak coincident with the LH trough just before ovulation (Robyn et al., 1973). The same author also found peripheral prolactin levels to vary much the same as those of estrogen.
Those observations are open to question since two other laboratories have failed to find any clear variation throughout the cycle (Jaffe et al., 1973; McNeilly & Chard, 1974). In Jaffe's study, plasma samples were taken at 'various times of day', which, because of the diurnal rhythm, might have missed a consistent phase at test time. In McNeilly's study, some women did show the midcycle peak while others did not. Tests were performed at the same time of day without considering the individual's sleep-wake cycle. Sassin's findings of a sleep-dependent rise of PRL makes these results somewhat difficult to interpret. Robyn's positive findings would be subject to similar uncertainty since sleep-wake variations in personal habits were not reported. One further complication regarding prolactin is that the stress of surgery has been shown to elevate peripheral PRL levels nearly 6 fold (McNatty et al., 1975). The possibility of similar stress effects associated with plasma sampling should be considered.
1.4.2 Effects of prolactin.
The pathology of the hyperprolactinemia-anovulatory syndrome provides a clue to the function of PRL in the normal ovary. The condition is characterized by high levels of peripheral prolactin, amenorrhea, often galactorrhea, subnormal to normal E2 levels, lack of LH secretory episodes and LH and FSH hypogonadotropism, (Bohnet, Dahlen, Wuttke & Schneider, 1976). The same symptoms usually occur in post-partum lactating women (McNatty et al., 1974). Several studies support the view that increased prolactin levels suppress gonadotropin levels (Boyer, Kapen, Finkelstein, Perlow, Sassin, Fukushima, Weitzman & Hellman, 1974 Bohnet et al., 1976; Floershein-Shachar & Keller, 1977). In addition, follicular concentrations greater than 25 ng/ml PRL appear to suppress the normal follicular steroidogenic response to gonadotropins (Seppala, Hibroven, Ranta, Virkkunen & Leppaluoto, 1975; McNatty et al., 1974; Floerschein-Shachar & Keller, 1977). Thus, in vitro human granulosa cell preparations responded to gonadotropins with progesterone secretion at low doses of PRL (less than 25 ng/ml), but failed to respond at higher doses (25-100) ng/ml) even in the presence of gonadotropin concentrations increased serially to levels 50 times normal.
Further evidence for inhibition of the central and ovarian reproductive cycle by prolactin is provided by the demonstration that a lactation suppressant, bromocryptine, reduced serum prolactin levels in both pathological and puerperal cases in 20 out of 21 patients, and that menses gradually returned in 11 of the 13 women with menorrhea in this group (Seppala et al., 1975). Bromocryptine is a dopamine receptor agonist which suppresses prolactin release (Bohnet et al., 1976).
In a similar study, 10 of 12 amenorrheic-galactorrheic patients resumed normal gonadal function after bromocryptine administration had reduced prolactin secretion (Floershein-Shachar & Keller, 1977). The effect continued only during the treatment. The existence of a dopaminergic control of releasing hormones has been established; and dopamine is implicated as the prolactin inhibitor (for a detailed discussion see section 1.5). Thus, dopamine triggers cyclic processes and simultaneously suppresses the inhibitory effect of prolactin. Conversely, melatonin inhibits cycle processes but stimulates cycle inhibitors as the following discussion shows.
1.4.3 Melatonin controls prolactin.
The effects of melatonin on reproductive physiology point to an LH inhibition and simultaneous prolactin stimulation in a variety of mammals, including the human. Single i.v. injections of tryptophan (TP), 5HTP, and serotonin all consistently stimulated prolactin production in laboratory rodents (Meites, 1973). Melatonin implants have been shown to inhibit LH synthesis in rats via a centrally mediated pathway (Fraschini, Collu & Martini ,1971). Melatonin is produced in the rat pineal as a product of serotonin (5HT) metabolism during the dark phase of the circadian cycle (Axelrod, Wurtman & Snyder, 1965; Klein, 1974).
Two pathways for activation of pineal metabolism have been reported which lead to melatonin secretion - one via the visual system, the other via the autonomic system (Kappers, 1960; Axelrod et al., 1965). In the visual system the inferior accessory optic tracts mediate a tonic pineal neuroendocrine response to light; and a second visual pathway, a retinal component, innervates the suprachiasmatic nucleus (SCN) and appears to be essential for the phasic timing of the circadian rhythm in the rat (Moore, 1974). In addition, the mammalian pineal is innervated by the superior cervical ganglia (Kappers, 1971; Wurtman, 1971).
Activation of this network by the sympathetic nervous system may excite the pineal to secrete melatonin. In one report, induction of ovulation occurred in a group of formerly anovulatory women following procaine blockage of the superior cervical ganglia (Novak & Woodruff, 1974).
1.4.4 Melatonin inhibits ovulation.
Support for an inhibitory role for melatonin during ovulation is afforded by the demonstration that melatonin levels show a cycle throughout the menstrual cycle, with the highest levels occurring 12 days before ovulation with values some 6 times those of the nadir at ovulation (Wetterberg, Arendt, Paunier, Sizonenko, Van Donselaar & Heyden, 1976). An intriguing report by Zacharias & Wurtman (1964) suggested a pineal antigonadotropin role in humans. Congenital blindness in a group of institutionalized girls was associated with earlier menarche than population norms.
It was concluded that blindness resulted in a loss of the visual stimulus to pineal indoleamine production. This would lead to diminished antigonadotropin activity and facilitate earlier menarche. While the idea is fascinating, a cautious interpretation is advised because institutionalization grouping itself, or possible pheromonal effects therefrom, might make such a subject population non-representative.
Thus the combined evidence from normal postpartum an pathologically hyperprolactinemic women suggests that prolactin interferes with gonadotropin secretion and with ovarian response to gonadotropins when the levels exceed those of normally cycling women. That melatonin, which is apparently produced only in the pineal in humans, may mediate this effect (Axelrod et al., 1965; Tepperman, 1973; Wurtman & Cardinali, 1974) is a strong possibility.
1.5 Monoamine control of the gonadotropins
Evidence that neural impulses initiate secretion of gonadotropins is now well established, and the interplay between serotonergic inhibition and mononaminergic excitation of gonadotropin release has been validated. Furthermore, the results of experimental work on rats, hamsters and rhesus monkeys seem to agree with clinical data for humans. Some time ago, Sawyer, Markee & Townsend (1949), proposed the idea of an adrenergic control of gonadotropic secretion but lacked the experimental technique necessary to test their hypothesis. Localization of catecholamines (CA) in mammalian brain was subsequently reported by Fuxe and Hokfelt in the rat in 1969 using the newly developed histochemical fluorescence technique.
High accumulations of CA were found in cell bodies which originated in the arcuate nucleus of the median eminence and terminated in the lateral margins of the median eminence. By 1972, additional noradrenergic (NA) nerve endings in the arcuate nucleus of the median eminence were defined with cell bodies arising outside the basal hypothalamus (Jonsson, Fuxe & Hokfelt, 1972). Knobil has demonstrated the arcuate to be the necessary nucleus in the rhesus pulsatile gonadotropin center described earlier (Knobil & Plant, 1978). Using bioassay, McCann was the first to report localization of LHRH in rat brain and LHRH activity was demonstrated in three nucleii: the arcuate, the SCN and the preoptic (McCann, 1962). Confirmation by radioimmunoassay followed (King et al., 1975).
Thus the finding of CA terminals in close proximity to releasing hormone (RH) cell centers promoted investigation of CA involvement in RH activity. A series of neuropharmacologic studies simultaneously attacked the problem from three points of view. (a) Central CA depletion was shown to block ovulation in a variety of experimental animals (Kalra, Kalra, Krulich, Fawcett & McCann, 1972; Labhsetwar, 1972: McCann, Ojeda, Martinovic, & Vijayan, 1977), and was shown further to block the photoperiodically induced gonadal development of seasonal breeders (Halawan & Burke, 1975). (b) Adrenergic blocking agents were shown to block LH discharge in rhesus monkeys (Bhattacharya et al., 1972). In ovariectomized monkeys, the blocking agents immediately inhibited the pulsatile pattern of gonadotropin release which had been initiated with the elimination of ovarian steroidogenesis. This finding was confirmed and extended by the administration of exogenous LHRH to another group of rhesus monkeys prepared as above. The exogenous LHRH resulted in a return to the pulsatile gonadotropin release which had occurred before the administration of the adrenergic blockers (Spies & Norman, 1975).
These effects were specific for (alpha) adrenergic blockers since control injections of generalized CNS depressants produced no well-defined effect on plasma LH. Agents used included deep general anesthetic, pentobarbital and hypotensive agents such as propanol and histamine (Bhattacharya et al., 1972). (c) Oral administration of L-DOPA, which can cross the blood brain barrier, resulted in a 40 % drop in serum prolactin levels in hyperprolactinemic women (Board, Fierro, Wasserman & Bhatnagar, 1977). Adrenergic blocking agents did not alter prolactin plasma concentrations, which suggests a possible defect in production of prolactin inhibiting factor (PIF) in these patients.
Evidence for a serotonergic pathway for the inhibition of spontaneous ovulation was provided using pharmacological manipulations in rats and hamsters (Kamberi, Mical & Porter, 1970, 1971). The combined evidence from a variety of mammals thus demonstrates a catecholaminergic control of gonadotropin release. Changes in prolactin levels indicate that gonadotropins and prolactin are controlled by the catechol and indole amines.
Experiments to locate the site of the CA depletion effect ruled out the pituitary (Green & Harris, 1947; Kamberi et al., 1969; McCann et al., 1977) and showed that the median eminence was the central location for the CA effect upon LH release in the rat (Sawyer, Hilliard, Kanematsu, Scaramuzzi & Blake, 1974). This same site has been confirmed for the rhesus monkey by Krey et al.,(1975), who showed that complete disconnection of the MBH from the rest of the brain did not affect the circhoral discharges of LH and FSH or the negative feedback inhibition in the ovariectomized animal. In contrast to the rat, surgical disconnection of the MBH from the SCN and the preoptic areas does not interfere with the estrogen-induced gonadotropin surges, or with the spontaneous ovulatory discharges of LH and FSH. Thus, for the rhesus monkey, the neural components are included within the MBH. It should be noted that this experimental disconnection does not exclude a control by the visual stimulation of the pineal which was described as an activator for the indole amine secretion. Disconnection would not prohibit hormone circulation.
At the outset there was some confusion as to the exact nature of the CA involved. The earliest experiments had shown dopamine (DA) to stimulate LH release (Kamberi et al., 1969) and as additional reports of noradrenergic excitation of RH activity appeared, more intricate pharmacological studies were undertaken to investigate the specificity. Novepinephrine (NE) was clearly effective in stimulating RH activity, whereas DA sometimes stimulated and at other times inhibited LH release. Since DA is a precursor of NE, it seems reasonable to consider NE to be the effective agent. All the studies cited support this possibility.
1.6 Autonomic innervation of ovarian tissue
Ovarian tissue is autonomically innervated in all mammals studied. A particularly striking density of adrenergic nerve networks is present in the cat, human and monkey, which suggested a possible influence on ovarian function. In addition, catecholamine-containing nerves have been described both in close proximity to blood vessels and within the stromal fibromuscular layer of the ovary (Doye, 1954; Jacobowitz & Wallach, 1967; Marshal, 1970; Bell, 1972; Weiner, Wright & Wallach, 1975). It is therefore interesting to note that neural (autonomic) innervation is not necessary for ovulation (Weiner et al., 1975; Doyle, 1954). In addition, the observation that follicles were often innervated inspired measurement of intrafollicular pressure in hopes of demonstrating changes associated with increasing follicular fluid. Results in rabbit ovaries were negative. Intrafollicular pressure did not increase at ovulation (Espey & Lipner, 1963; Rondell, 1964). It was shown, however, that the tensile strength of the follicle wall deteriorates as rupture nears (Rondell, 1964). The fact that neural innervation of ovarian tissue is not required for ovulation does not eliminate autonomic control of ovarian function which might be mediated through circulatory distribution of epinephrine (E). Condon & Black (1976) suggested the possibility of such a role for autonomic control of ovarian function. Using bovine corpus luteum slices in vitro, they showed that both E and NE would stimulate progesterone synthesis as would the b-adrenergic stimulator, isoproterenol. The authors suggested that, if these findings were physiologically significant, then E must be the physiological agent since autonomic innervation of the ovaries does not include neural penetration of the CL. They proposed that the resultant E effect must be mediated either from the adrenal medulla or within the ovary itself via local epinephrine synthesis. They favored the latter possibility although the former seems equally likely. Human ovulatory plasma NE and E surges have been described although amine patterns throughout the cycle were not studied (Zuspan & Zuspan, 1973: Rosner, Nagle, La Borde, Pedroza, Badano, Casas & Carril, 1976). Thus it would appear that the autonomic nervous systems cannot be easily dismissed in the control of ovulation, although its role in the peripheral control of ovulation may not be mediated by direct local autonomic innervation.
1.7 Metabolic influences on ovarian cyclicity.
Evidence for the effects of metabolic function on subsequent ovarian cycle function is derived from human pathological reports, as well as from in vitro studies. Among both the obese and the emaciated, menstrual disorders are commonly reported which, in their most extreme forms, result in incomplete amenorreha (Rogers & Mitchell, 1952; Mitchell & Rogers, 1953; Warren & Van De Wiele, 1973; Sherman & Korenman, 1974; Frisch & McArthur, 1974; Warren, Jewelewicz, Dyrenfurth, Ans, Khalaf & Van De Wiele, 1975; McArthur, O'Loughlin, Beltins, Johnson, Hourihan & Alonzo, 1976). Return to normal weight is usually accompanied by resumption of menses. Gonadotropin and steroid insufficiencies reflect the inappropriate weight; and these levels respond after a shorter latent period than the menses, to the change to normal weigh (Sherman, Halmi & Zamudio, 1975).
In vitro studies in pig (Channing, 1976) and monkey (Channing & Tsfriri, 1976) have demonstrated a role for the nonreproductive hormones - insulin, thyroxin and cortisol - in the regulation of the ovulatory process. All the hormones were shown to be necessary for the gonadotropoin induction of ovulation. It was suggested that these hormones are necessary for normal metabolic function in ovarian tissue in much the same way as in other cells of the body (Channing & Tsafriri,1976).
1.8 Summary of the endogenous nature of the ovulatory rhythm
Our present understanding of the endogenous control of the ovulatory cycle can be briefly summarized as follows. The ovaries have a natural rhythm of their own as shown both by a usual minimum 12.6 days fixed time requirement for follicular development and the fact that their steroid output inhibits the tonic pulsatile pattern of gonadotropic output. The steroid production is limited by the individual capacity of the follicular and luteal cells to respond to gonadotropins, prostaglandins and prolactin, as in the case where follicular progesterone is only secreted by cells of the mature Graafian follicle.
Gonadotropins are produced in the anterior pituitary gonadotropes, the response being determined by the time and quantity of steroid and releasing hormones impinging upon these cells. The pineal, through its production of indoleamines, exerts an inhibitory effect upon the gonadotropic secretions; and, in addition, the resulting prolactin inhibits the ovarian response to whatever gonadotropin is secreted. Monoamine production controls gonadotropic secretion by affecting the releasing hormones. The ovarian innervation by the autonomic nervous system (ANS) is not involved critically in the peripheral control of the ovarian cycle although the ANS clearly does exert a central effect through its innervation of the pineal gland via the superior cervical ganglia as well as, possibly, through adrenal medullary secretion of epinephrine.
2. EXOGENOUS INFLUENCES ON FERTILITY
A number of exogenous influences have been documented which alter reproductive capacity and it is reasonable to assume that the list will be extended. To date it has been shown that nutrition, geophysical conditions, childbearing and lactation, and social-grouping influence human fecundity. Sexual behavior has also been shown to effect fertility.
2.1 Nutritional influences on fertility
Inappropriate nutrition has been implicated as a cause of infertility. The effects of nutrition on the aging of reproductive systems have been documented in rats and mice. McCay, Sperling & Barnes (1943) showed that by restricting the diet of prepubescent rats, sexual maturation was delayed and lifespan was doubled. Elimination of the usual fatal diseases, such as cancer, was suggested as the reason for the longevity. Caloric restriction in reproductively mature mice has also been shown to reduce the rate of oocyte decline and thus maintain reproductive viability for longer periods (Huseby & Ball, 1945).
Also in mice, dietary protein has been shown to be critical for female sexual maturation after 21 days of age (Vandenbergh, Drickamer & Colby, 1972). Thus, limiting intake to 'appropriate' levels is evidently critical in optimizing viability in these mammals.
Data for humans also demonstrate reproductive system control by nutritional factors. Inhibition of cyclic processes by obesity, and their return to normal, following return to appropriate weights, is well documented (Rogers & Mitchell, 1952; Mitchell & Rogers, 1953; Sherman & Korenman, 1974a; Warren et al., 1975). A similar inhibition of reproductive cycles has been noted for the emaciated having anorexia nervosa (Crisp & Stonehill, 1971; Crisp, Chen, MacKinnon & Corker, 1973; Warren & Van de Wiele, 1973; McArthur et al., 1976). Famine among war-torn populations was also associated with fertility problems. In these cases, the pathology was demonstrated in the increased incidence of spontaneous abortion, lower birth rate and lower birth weight (Stein & Susser, 1975; Antonov, 1947).
The possibility of a critical minimum body fatness for facilitation of the menarche has been suggested by data which show a close correlation between weight and height at the onset of first menstruation (Frisch, 1974), as well as for the return of menses in recovering anorexics (Frisch & McArthur, 1974). A greater amount of body fat was required for recovery from amenorrhea after anorexia than was needed for onset of menarche in normally developing girls.
2.2 Geophysical influences on fertility
Several authors have demonstrated an influence of altitude, climate and season on various indices of fertility. It seems clear that high altitude (approximately 10,000 feet) has adverse affects upon fertility. Exposure for 4 weeks on a Peruvian mountain resulted in 9 men showing a significant decrease in sperm count and motility associated with an increase in abnormal forms (Donayre, Guerra, Garcia, Monclosa & Sobrevilla, 1968).
In women, it was shown that migrating to a high altitude produced menstrual disturbances for both South American natives (Sobrevilla, 1967) and American college students on summer travel (Harris, Shields & Hannon, 1966). Similarly, high altitude in Andean countries was shown to be associated with reduced fertility compared with controls living in nearby valleys (James, 1966). Migrants from high to low altitudes were found to experience immediate increases in fertility indices (Abelson, Baker & Baker, 1974). In addition, birth weights of newborns are lower for populations at high altitudes than for those at low altitudes (Abelson, 1976).
A seasonal variation in birth rate has been demonstrated in the United States, with a greater amplitude of change for those living in the South (Pasamanick, Dinitz & Knoblock, 1959; Rosenberg, 1966). In these data a September peak and an April and May trough in birth rate is clearly discernible. This was not related to the month of the marriage. In Hong Kong, a strong inverse correlation was shown between temperature and conception (Chang, Chan, Low & Ng, 1963). This seasonal effect is less pronounced in societies such as West Malaysia with culturally defined sexual tabu periods (Johnson, Ann & Palan, 1975).
It therefore appears that hot months tend to be associated with reduced conception in cultures which lack religious sanctions on sexual behavior. However, other social influences appear relevant; for example, Puerto Rican birth rates shifted from a pattern opposite to that in the U.S. to an identical cycle of peaks and troughs after 1946. The timing of this phase-shift coincided with increased cultural exchange between the two countries, e.g. travel, TV and newspaper syndication (Cowgill, 1964).
Seasonal behavior has been documented for the rhesus monkey. There is a seasonal decline in male-initiated sexual behavior which is independent of the female incidence of pregnancy, parturition or summer amenorrhea (Michael & Zumpe, 1976). These data, obtained in London, show high sexual activity in the fall and low activity or none in the spring moths. The rhesus colony at the University of Pennsylvania exhibits a consistent anovulation during the summer months despite air-conditioned quarters. This pattern is remarkably similar to the conception data for humans in the U.S. and it seems to point to a seasonal and/or temperature-dependent phenomenon.
The third geophysical consideration, the possibility of lunar influence on menses, has been reviewed and suggested by new data (Cutler 1979). Dewan & Rock (1969) report a subject receiving nocturnal illumination during three nights from day 14 of her cycle demonstrated a change from her irregular cycle to a 29 day cycle. Subsequently, a group of subjects received a similar regimen while also supplying control (unmanipulated) cycle data. Eight of 11 subjects showed a narrower range of cycle lengths under conditions of nocturnal illumination, than when unmanipulated; although only a quarter of the experimental cycles achieved the 29 day cycle length (Dewan, Menkin & Rock, 19078).
Thus, the results are suggestive, though less than definitive, that nocturnal light influences cycles. Moreover, women with length of menstrual cycles similar to the lunar cycle (29.5 days), as well as women with highly irregular cycles, tended to ovulate in the dark half-cycle of the month;:from 3rd quarter through new moon, through first quarter. Electromagnetic radiation cycles were considered to effect this result since the women were generally removed from the lunar photic cycle by their city habitats (Cutler, 1979).
2.3 Influences of child-bearing and lactation on subsequent fertility
2.3.1 Demographic studies of lactation in women.
Lactation substantially reduces the incidence of pregnancy-although it is not an effective contraceptive method (Knodel, 1977). Lactation also reduces infant mortality. In a number of cultures, reduction of the incidence of breast feeding (e.g. among recently urbanized women) is associated with an increase in infant mortality. For example, in Bavaria from 1904 to 1906 in 8 provinces, those provinces which had a greater percentage of breast-fed infants showed the lowest percentage of infant deaths and vice versa. This same phenomenon has been shown in the Third World countries today. It was suggested that mothers' milk may provide immunity and prevent the introduction of impurities from breast milk substitutes (Knodel, 1977). Demographic studies have shown a trend away from nursing as cultures become urbanized. For example, among Boston women in 1975, 78% of mothers did not breast feed at all (Short, 1976).
In contrast, among the non-urbanized Amish, all of the mothers appear to nurse (Tyson, Freedman, Perez, Zacur & Zanartu, 1976), as do the North American Hutterite women (Short, 1976). This trend results in an increased fertility when measured by the earlier return of menses and a concomitant decrease in fertility when measured by rate of infants surviving. The net balance seems to be a general increase in overall fertility (Knodel, 1977) and according to Short (1976), this is responsible for the recent exponential increase in the world's population.
2.3.2 Nutrition and lactation affect postpartum fertility. In a study of several cultures, two factors have been shown to delay the return of menses in the mother: lactation and nutritional deficiencies (Short, 1976). A further distinction between fully-lactating (breast the only source of infant's nourishment) and partially-lactating (supplemental feeding) behavior helps to clarify the time differences reported in subsequent menses onsets.
In a sample of Baltimore, Maryland women, partial lactators had their first postpartum menses no sooner than 42 days and this could be prolonged by full lactation to averages of 88 days (Tyson et al., 1976). In a Boston sample, the first postpartum menses of non-nursers was 55 days while among lactating mothers, the longer the duration of nursing the longer generally was the delay to first menses, with 155 days the average interval for women who nursed for more than this span (Short, 1976).
North American Hutterites who are well nourished and fully breast feed, show birth intervals of 22 months on average. In contrast, a relatively less nourished culture, the Kung hunter-gatherers, who also fully breast-feed and reportedly practice no contraception, have birth intervals averaging 4 yr (Short, 1976). Thus, lactation being equal, the level of nourishment could explain differences in fertility.
2.3.3 The mechanism by which lactation inhibits ovulation.
A general sluggishness of the ovulatory axis during lactation has been demonstrated by a variety of reports. Apparently the ovary is refractory to gonadotropin stimulation and the pituitary is refractory to releasing hormone stimulation. Furthermore, gonadotropin releasing hormone is inhibited by the prolactin secretion which suckling induces. The following reports constitute evidence for this ovulatory system refractoriness after parturition.
(1) Suckling has been clearly shown to increase PRL concentration in plasma. Weekly plasma samples of partially nursing women averaged 20.8 + 3.4 ng/ml PRL vs. 30.9 + 4.9 ng/ml for women fully breast feeding. These tests were conducted more than 3 hr after a nursing session using the Baltimore sample. Twins placed at the breast singly produce half as much PRL response as that produced when twins are placed at both breasts simultaneously (Tyson et al., 1976).
(2) Prolactin and gonadotropion concentrations exhibit a reciprocal relationship among healthy lactating mothers (Tyson et al., 1976). This finding provides further evidence for a feedback control of PRL on gonadotropins as discussed earlier. Sleeping through the night without suckling permits gonadoptropin surges (Tyson et al., 1976) and may explain why fully breast feeding women do not ovulate as soon after parturition as partial-lactators. A similar reciprocal relationship for a PRL and gonadotropins was shown in bonnet monkeys where PRL inhibited LH secretion (Maneckjee, Srinath & Moudgal, 1976).
(3) The postpartum pituitary is refractory to exogenous LHRYH administration as shown in experiments on women and bonnet monkeys (Friedman, Geake, Fang & Kim, 1976; Maneckjee et al., 1976). In the women, LHRH administration on postpartum days 1, 3, or 8 yielded no measurable response in LH or FSH levels in non-lactating women. By 36 days, an LH and FSH elevation following the LHRH was shown. While PRL levels were not reported, it is likely that the early days after delivery were accompanied by high PRL levels which declined by the 36th day. In the monkeys (who were lactating) synthetic LHRH injections during suckling produced an increase in LH which could be suppressed by a prior injection of PRL. Thus PRL itself was shown to influence pituitary response to LHRH during lactation (Friedman et al., 1976)
(4) The ovarian response to gonadotropins is also suppressed during lactation. Among 6 healthy lactating women studies during 6 postpartum weeks, gonadotropin injections produced no increase in urinary levels of estrogens or progestogens (Zarate, Canales, Soria, Ruis & MacGregor, 1972). The demonstration, in rhesus monkeys, that suckling appears to prolong the life and function of the CL of pregnancy (Weiss, Dierschke, Karsch, Hotchkiss, Butler & Knobil, 1973) may help explain the ovarian suppression. Progesterone was present in these lactating monkeys but absent in the mothers who were prevented from nursing. It was described earlier that the presence of functional CL with associated progesterone has been shown always to prevent concurrent ovulation.
In conclusion, lactation clearly affects subsequent ovulation through a variety of suppressive mechanisms. These mechanisms all appear to operate in a similar manner for the postpartum and the postovluatory woman. It would appear that prolactin is the principal suppressive agent.
The exogenous influences of nutrition, geophysical conditions, child-bearing and lactation may -perhaps- have social and sexual influences which are not clearly understood. Recently more satisfactory explanations have appeared for mechanisms by which behavior and physiology interact to maintain a continuation of mammalian species. In a subsequent review attention will be directed to the current state of this subject.