Luteinizing Hormone-Releasing Hormone (LHRH) Neuronal System: From Basic to Clinical

Serge Rivest and Leonello Cusan / Laboratory of Molecular Endocrinology

Reproduction in male and female mammals is critically dependent on the appropriate neurosecretion of luteinizing hormone-releasing hormone (LHRH),also called gonadotropin-releasing hormone (GnRH), a decapeptide characterized by the groups of two Nobel Laureates Guillemin and Schally. LHRH is released in pulses and triggers the production of gonadotropins, which stimulate the growth and release of eggs by the ovary. This fact has been best illustrated by experiments in which the actions of the decapeptide have been blocked by immunoneutralization or receptor antagonist treatment, which invariably leads to cessation or reduction of gonadotropin secretion, disruption of gonadal function, and infertility (145). Sterility in mutant, non-LHRH-producing mice (267) and human infertility associated with LHRH insufficiency (49) also provide a clear demonstration of the reproductive consequences of inappropriate or deficient LHRH neurosecretion. In the rat, administration of an LHRH antagonist during proestrus results in a rapid and complete inhibition of ovulation, demonstrating the importance of this neuropeptide in reproductive function. As recently reviewed by Kalra (116), a complex series of mechanisms involving biogenic amines, neuropeptides and circulating sex steroids acting on multiple hypothalamic and perhaps extrahypothalamic nuclei, all participate in the regulation of reproductive function. It is not the purpose of this chapter to describe in detail all the sophisticated mechanisms controlling the ovulatory cycle in female and reproduction in males, but rather to give a general view of the neuronal system responsible for such physiological phenomena essential to maintaining the species.

During embryonic development, LHRH neurons migrate from the medial olfactory placode of the developing nose through the nasal septum and into the forebrain with the nervus terminalis, arching into the septal-preoptic area and hypothalamus (256). Wray and colleagues used in situ hybridization histochemistry and immunocytochemistry to study the prenatal expression of LHRH cells in the mouse and showed that postnatal LHRH neurons were "birth-dated" shortly after differentiation of the olfactory placode and before LHRH mRNA is expressed in cells of the olfactory pit (303)). This elegant study therefore supported the hypothesis that all LHRH cells in the central nervous system (CNS) arise from a discrete group of progenitor cells in the olfactory placode, and that a subpopulation of these cells migrates into forebrain areas, where they subsequently establish an adult-like distribution (303). It has been suggested that this migratory route for LHRH-expressing neurons could explain the deficiency of gonadotropins seen in "Kallmann's syndrome" (hypogonadotropic hypogonadism with anosmia) [256]. Clinical studies have indeed shown that gonadal dysplasia with decreased secondary sex characteristics and deficiency in gonadotropins can be associated with malformation of the olfactory apparatus. Interestingly, the destinations of these migratory LHRH neurons are quite different from one species to another. In the opossum (Monodelphis domestica), the neurons responsible for control of reproductive function do not enter the preoptic area or the hypothalamus, whereas in mice and rats, LHRH perikarya migrate up to the medial preoptic area-anterior hypothalamus (ventral zone, just above the optic chiasm). In other species, such as the guinea pig, sheep, monkey and human, these neurons are found as far caudal as the arcuate nucleus (ARC)/median eminence (ME) and premammillary nucleus (MM) [for review, see ref. 265]. This species-dependent LHRH-neuronal organization might certainly contribute to the difficulty of pin-pointing mechanisms underlying the inhibitory or stimulatory actions of various factors on the reproductive system in various mammalian species.

The LHRH neuronal system differs from other types of neuropeptidergic systems, such as the neuroendocrine corticotropin-releasing factor system (CRF), which is essentially located in the parvocellular division of the paraventricular nucleus (PVN) of the hypothalamus under a compact cluster of perikarya forming a clear and defined nucleus. The LHRH cell bodies are scattered and are distributed from the medial septum (MS)-diagonal band of Broca (DBB) to the caudal MPOA and ventral-rostral anterior hypothalamus within the rodent brain (Fig. 1), whereas another group of LHRH perikarya can be found in the ARC/ME of monkey and human brains. While this latter group of LHRH cells is essentially neuroendocrine (i.e., neurons projecting to the infundibular system and directly involved in the control of gonadotropin release from the adenohypophysis), the more rostrally distributed LHRH neurons send their projections to various regions of the brain, including the ME (neuroendocrine), the interpeduncular nucleus, the medial amygdala, the mammillary complex and up to the periaqueductal region (109,110,122,123,164,165,265,268). The exact role played by the non-neuroendocrine LHRH neuronal system in regard to reproductive function has yet to be clearly established, although it has been suggested that some of these pathways could be associated with sexual behavior (for review, see ref. 216).

In rats, a large concentration (the richest) of LHRH-immunoreactive (ir) and expressing (mRNA) neurons are localized within the MPOA, forming an inverted "V" surrounding the organum vasculosum of the lamina terminalis (OVLT) [Fig. 2A]. Each neuron can be clearly visualized by both in situ hybridization histochemistry (Fig. 2A, Fig. 2B) and immunocytochemistry (Fig. 2C), although some differences have been reported in terms of the number of cells, their organization and regulation throughout the estrous cycle and sex steroid manipulations between cells positive for either the protein or the transcript. In our hands however, we found very little differences in the number and general organization of the LHRH neurons, using both immunocytochemistry and in situ hybridization techniques. It is worth mentioning that, in contrast to many neuropeptides, which are undetectable or barely detectable without treating the animals centrally with colchicine to block the axonal transport, LHRH neurons can be easily visualized under basal conditions in both male and female brains. Although high variability exists between LHRH neurons, the protein and mRNA encoding the decapeptide are abundant in most neurons of the MPOA (Fig. 2B and Fig. 2C). Most of these cells are within a 1.5-mm block which includes the optic chiasm and the area just dorsal and anterior to it (265). Approximately 50–70% of these LHRH neurons project to the ME (Fig. 2D) and, consequently, contribute to the control of LH secretion from the adenohypophysis (166,268). Moreover, stimulation of the MPOA increases infundibular LHRH release and circulating LH levels, which can induce ovulation in female rats (116). It might, however, be too simplistic to believe that the MPOA is responsible in itself for the preovulatory LH surge and ovulation, because this hypothalamic structure receives massive input from numerous areas that may play a crucial role in orchestrating the afferent pathways driven by the physiological clock of ovulation and steroid production. Among these, the medial preoptic nucleus (MPN), the suprachiasmatic nucleus (SCN) and the ARC have been the subject of a great deal of studies seeking to localize the origin of the pathways involved in the preovulatory LHRH and LH hypersecretion (116).

Regulation of the synthesis and secretion of hypothalamic LHRH neurons plays a fundamental role in the process of reproductive maturation in both mammalian genders. Indeed, an increase in pulsatile LHRH release is the critical factor for the onset of puberty; hypothalamic LHRH content shows a steady increase during the first three months of life in male rats (20,31), and an increase in pulsatile LHRH release occurs at the onset of puberty in female rhesus monkeys (38,295). Chronic, intermittent, intravenous infusion of LHRH elicits a prompt hypersecretion of both gonadotropins in agonadal male rhesus monkeys between 15 and 18 months of age, the phase of development characterized by undetectable plasma LH and FSH levels (204). In 1980, Wildt and collaborators (298) observed initiation of a normal ovulatory menstrual cycle in prepubertal female monkeys by infusion of LHRH for 6 minutes once every hour. When this regimen was discontinued, the animals promptly reverted to an immature state. These authors concluded that neither adenohypophyseal nor ovarian competence is limiting in the initiation of puberty. They suggested that this process depends upon the maturation of the neuroendocrine control system that directs the pulsatile secretion of LHRH from the hypothalamus (298). Since this classic report, a large number of studies have investigated the exact pattern of LHRH gene expression and secretion preceding sexual maturation. Obviously, variations in LHRH neuronal activity during the period surrounding puberty are under the influence of multiple inhibitory and stimulatory factors.

It is currently believed that the prepubertal increase in LHRH release is not due to the developmental changes in the properties of LHRH neurons themselves, because LHRH neurosecretory system is already morphologically and functionally mature before the onset of the puberty (264). Although males and females do not differ notably in the total number of LHRH neurons at any age, nor in their LHRH cell types, these neurons display striking developmental changes in their dendritic processes during the period of sexual maturation (304). These authors proposed that smooth LHRH cells transform into irregular LHRH cells during development of the reproductive system. This may suggest that new synaptic inputs and circuits involving LHRH neurons (as they transform from smooth to spiny) play an essential role in both the timing and maintenance of sexual maturation (304). It is therefore probable that changes in inputs to LHRH neuronal system are responsible for the increase in pulsatile LHRH that removes the central inhibition of LHRH release and stimulates sexual development. To test the hypothesis that prepubertal increase in LHRH release is withheld by a dominant inhibitory neuronal system, Mitsushima et al. (170) examined the role of g-aminobutyric acid (GABA; a recognized inhibitory neurotransmitter) in the control of LHRH release before the onset of puberty. Their data provided evidence of a powerful GABAergic inhibition of the LHRH neurosecretory system in the prepubertal period, and this tonic inhibition may be a key factor in sexual maturation. Although specific mechanisms involved in the attenuation of GABAergic influence onto LHRH neurons during the prepubertal period have yet to be clarified, three possible mechanisms were suggested by these investigators (170) to explain the reduction in the endogenous GABA release leading to increased LHRH release and the subsequent initiation of puberty: 1) morphologically, GABAergic synapses onto LHRH perikarya and/or neuroterminals may be reduced as a programmed developmental event. 2) biochemically, synthesis of GABA may be reduced due to the developmental changes in the biosynthetic enzyme GAD. 3) physiologically, neural activity of GABAergic neurons may be reduced during the development of the hypothalamus.

Increased input from stimulatory factors such as norepinephrine, neuropeptide Y and glutamate during the prepubertal period must also be taken into consideration. These neuronal systems can target neuroendocrine LHRH cells and directly trigger the release of the decapeptide in a pulsatile manner, characterizing the onset of puberty. An orchestrated increase in the impulse of these stimulatory systems would therefore initiate puberty by either activating or removing inhibition of the LHRH neuronal system. This concept was reinforced by the elegant study performed by Gore et al., which showed that changes in glutamatergic inputs to LHRH neurons may play a critical role in sexual maturation (95). Indeed, the percentage of LHRH neurons that double-label with the NMDA type 1 receptor was 2% in prepubertal rats, but increased to 19% in postpubertal rats, indicating a robust increase in glutamatergic input (95). Other studies have shown profound increases in norepinephrine content in anterior hypothalamus and residual hypothalamus during sexual maturation in rats (211) and monkeys (94). This may also contribute to the developmental increase in LHRH release at this crucial period. How the inhibitory and stimulatory neuronal inputs act in concert on the neuroendocrine LHRH network to allow maturation of the reproductive system in mammals of both genders remains quite a fascinating issue to be resolved.

Although a complex afferent-efferent neuronal circuitry participates in the regulation of LHRH neurons in the rat at least, the MPOA/OVLT seems in a privileged position to directly influence the preovulatory LHRH discharge. The MPOA/OVLT is the richest area expressing the decapeptide, and a large proportion of these cells projects to the ME, which can therefore be called neuroendocrine-like neurons. Silverman and colleagues originally found that approximately 50% of the LHRH neurons of the MPOA were positive for the retrograde tracer (lectin wheat germ agglutinin [WGA]) administered directly into the external zone of the median eminence in rats (268). We (229) and others (166) have observed a larger percentage (up to 75%) of MPOA LHRH neurons sending their axons to the infundibular system of female rats. Moreover, 60–70% of LHRH neurons located in the MPOA spontaneously express the immediate-early gene (IEGs) c-fos in the afternoon of proestrus (Fig. 3). This provides direct evidence that this particular group of neurons is strongly activated during the preovulatory LH surge (140,227,229).

Neurons respond to a variety of extracellular stimuli, including potassium- or neurotransmitter-induced depolarization and stimulation by growth factors or hormones, by manifesting rapid and transient synthesis of IEGs, such as the proto-oncogene encoding the transcription factor Fos (3,177,263). The gene products of IEGs may represent mediators of the translation of short-term intracellular signaling events to longer-term changes in cellular phenotype via targeted alterations in gene expression (3). Although the induction of c-fos may not be an absolute reflection of a true increase in neuronal activity, it probably does represent a monitor of intracellular second-messenger levels (177). Analysis of changes in IEG expression provides a powerful tool for evaluating the circuitry and cell groups affected by various physiologic and pharmacologic stimuli. As mentioned before, LHRH perikarya have the capacity to express the IEG c-fos during a very short period in the late afternoon of proestrus, corresponding to the preovulatory LH surge. We have also observed, by means of a retrograde tracing technique, that a high percentage of LHRH neurons expressing c-fos protein during proestrus project to the infundibular system. Representative examples of LHRH neurons expressing the IEG Fos and projecting to the median eminence during the afternoon of proestrus in rats are presented in Figure 4. The retrograde tracer fast blue (0.2%), diluted in sterile saline, was administered into the femoral vein of cycling female rats. Four days later, animals were deeply anesthetized and rapidly perfused with 4% paraformaldehyde between 1730 and 1800 h of proestrus afternoon. Frozen brains were mounted on a microtome and cut into 30-mm coronal slices. It is possible to see a clear example of 5 LHRH-immunoreactive neurons (panel A) that contain the retrograde tracer fast blue (panel B) and the nuclear Fos protein (panel C). Although these data indicate that LHRH neurons expressing Fos during proestrus project to the median eminence (see filled arrowheads), we have also observed that some LHRH neurons do not contain both the fast blue and the Fos-immunoreactive protein (229). Indeed, some LHRH neurons expressed Fos without containing the retrograde tracer, and vice versa. LHRH neurons were stained by means of immunocytochemical techniques using a fluorescein (FITC)-labeled second antibody, while Fos proteins were stained using a rhodamine-labeled antibody.

The physiological relevance of this spontaneous induction of Fos within neuroendocrine LHRH neurons during the preovulatory LHRH surge still remains largely hypothetical. The relevance of using different IEGs as indicators of neuronal activation might also depend on whether the gene of interest contains a functional consensus sequence in its promoter region, which could suggest a direct relationship between the activated target gene and the transcription factor. Such a relationship is obviously more difficult to establish when there is no consensus site for the IEG to bind on that particular gene to activate its transcription. In the IEG family, transcriptional activation through the AP-1 element (-TGACTCA-), as a consequence of the interactions between Fos and Jun proteins, has been extensively reviewed (3,120,176,177,203,248, 263). They share the leucine zipper dimerization motif, and each member of the jun family can dimerize with both itself and other members of the fos and jun families (263). All the different combinations of Jun/Jun dimers and Fos/Jun dimers will bind to the consensus AP-1 binding site, but Fos/Jun dimers bind to the AP-1 sequence with 50-fold greater efficiency than Jun/Jun dimers, and Fos/Fos dimers cannot bind to DNA (115). Because c-fos is a transcription factor generally recognized to activate genes when forming an heterodimer with member of the c-jun family (212), and because an AP-1 DNA binding site is present on LHRH promoter (see below), it is tempting to speculate that the spontaneous expression of the IEG c-fos within LHRH nuclei is involved in the transcriptional activity of the decapeptide during the preovulatory LHRH surge. The situation is, however, quite complex, and results supporting such hypotheses are conflicting.

Although converging evidence has largely expanded our understanding of LHRH system behavior at the time of ovulation, what really takes place in hypothalamic LHRH neurons, both in terms of gene expression and protein products, is far from being clear. However, the possibility of examining this neuroendocrine system by means of exonic and intronic probe technology (108), coupled with the detection of spontaneous nuclear-early gene expression in OVLT/MPOA neurons during the afternoon of proestrus, provided new tools to clarify many aspects of LHRH transcription during basal and perturbed conditions. Several groups have reported low LHRH gene expression during the morning of proestrus, compared with the afternoon (156,193,308,309). Levine and coworkers observed fluctuations in the number of cells expressing LHRH mRNA in specific regions of the basal forebrain during the periovulatory period (207), whereas Silverman and Witkin did not observe any regional differences and found that after the preovulatory surge, all LHRH neurons were synchronized to start neosynthesis of the peptide (266). Our results did not show a clear increase in the number of LHRH-expressing neurons in the periovulatory period, although an elevation of LHRH primary transcript (LHRH heteronuclear (hn) RNA) was detected at proestrus 15.00 (183). These results are in agreement with recent studies suggesting that transcriptional events take place in LHRH neurons selectively during the afternoon of proestrus (93). Intronic probe technology seems therefore more sensitive than exonic probes for evaluating LHRH transcriptional activity in intact, cycling animals (Fig. 5), and the time-related events of Fos expression and LHRH transcriptional activation in the afternoon of proestrus support the concept that both phenomena are related. The exact participation of the IEG c-fos in these phenomena still remains highly hypothetical, although we have already reported parallel alterations of both Fos and LHRH gene expression in the same subset of neurons in the rat MPOA (183). However, recent studies have provided evidence that AP-1 is involved in repressing rather than stimulating LHRH gene transcription (22), although opposite results (AP-1 stimulating effects on LHRH gene transcription) have also been published (73). Moreover, Zakaria et al. (307) demonstrated that a consensus AP-1 site found in the human LHRH gene, but not in rodent genes, is capable of stimulating gene transcription in response to treatment with phorbol ester 12-O-tetradecanoyl phorbol-13-acetate (TPA), an AP-1 response element. This was the first demonstration of species-specific differences in phorbol ester regulation of LHRH gene transcription and could, in part, explain differences in reproductive function among mammals (307).

Neuroendocrine LHRH is secreted in a pulsatile manner in the infundibular system, an essential characteristic of that system for the release of gonadotropins from the adenohypophysis. In fact, if LHRH agonists are chronically administered by constant infusion, a downregulation of pituitary LHRH receptors occurs, resulting in a profound inhibition of LH secretion and a disruption of the ovulatory cycle. At the end of the follicular phase (day 14) in humans, and during the afternoon of proestrus in rats, the frequency and intensity of each LHRH pulse increase, provoking the LH surge and ovulation. This phenomenon is rapid and transient, and the mechanisms participating in these events still retain a great deal of interest for scientists trying to resolve the mystery of ovulation. The most remarkable challenge in understanding the complex regulation of LHRH discharge responsible for the preovulatory LH surge is the dual effects of circulating estrogens. Indeed, ovariectomy causes increased LH release, whereas replacement therapy with 17b-estradiol (E2) prevents such effects, suggesting that estrogens overall restrain the secretion of LHRH. The concept that neuroendocrine LHRH secretion is entirely responsible for the inhibitory influence of E2 on plasma LH levels has, however, been questioned. As elegantly reviewed by Kalra and Kalra (117), despite numerous attempts, there is little experimental evidence to show that E2 decreases the hypothalamic LHRH drive to the pituitary concomitant with the inhibition of LH release in ovariectomized rats. It is possible that the difficulty in precisely evaluating infundibular LHRH secretion in intact and even gonadectomized animals partially explains the discrepant data. In fact, we have been able to assay LHRH reproducibly only in the afternoon of proestrus of rats acutely implanted with a push-pull cannula into the ME early in the morning of proestrus (221,226,227). In gonadectomized male and female rats, and throughout the other periods of the cycle, the measurement of the decapeptide by radioimmunoassay (RIA) was very inconsistent, with detectable levels in some animals and undetectable concentrations in others (S. Rivest, P. Plotsky and C. Rivier, personal observations). As a consequence, the influence of gonadal removal and the inhibitory effect of E2 on LHRH secretion may be slight but sophisticated, and our techniques are not sensitive enough to perceive the effects, at least in rats.

In sheep, however, reliable determinations of the LHRH secretory pattern have been obtained in both intact and gonadectomized animals by measurements of LHRH in samples of pituitary portal blood. Thus, the secretion of LHRH has been thoroughly detailed, and the important observation has been made that each pulsatile discharge of LH is the direct results of a large secretory burst of LHRH from the hypothalamus (36). Determinations of LHRH from portal vein collection in sheep (36) is highly reproducible, in terms of frequency and amplitude. Numerous studies evaluating infundibular LHRH release have been performed in this specie, showing that E2 not only regulates the frequency and amplitude of the LHRH pulse but also produces qualitative changes in its pattern of release and induces a sustained LHRH surge, during which discrete pulses are not readily evident (77,78). In ovariectomized ewes, E2 exerted a dose-dependent suppression of LHRH secretion from the hypothalamus (76), whereas no effect of E2 was observed on either the frequency or surge of LHRH pulses during the breeding seasons (37). These findings led to the conclusion that there is a marked seasonal change in the negative feedback effect of E2 on the episodic LHRH secretion in the ewe; the steroid is maximally effective during anestrus (37). It has also been demonstrated that the LHRH surge is not composed entirely of discrete, synchronous secretory events in ewes. One action of E2 in inducing the LHRH surge may be to switch the pattern of LHRH secretion into portal blood from episodic to continuous (171).

A sharp and transient increase in LHRH secretion can be detected into the ME at the end of the follicular phase in humans and in the afternoon of proestrus in rats. It is well known that E2 plays a key role in this LHRH discharge in mammals. The exact mechanisms through which E2 participates in the increased activity of neuroendocrine LHRH neurons still remain to be clearly established, but most likely its influence is indirect; estrogen receptors (ER) have yet to be found within the LHRH neurons of cycling female animals and following gonadectomy. It has been suggested that the primary action of steroids is to facilitate the output of peptide signals and to amplify/adjust their postsynaptic responses in a timely fashion (116). Several factors, including peptides, can directly or indirectly modulate (stimulate and inhibit) LHRH neuronal activity (Fig. 6). The manner in which sex steroids, particularly E2, interfere with this circuitry has been the subject of many investigations. Among the numerous possibilities or interaction is the role played by increased E2 in restraining the inhibitory influence of endogenous opioid peptide (EOP) on the LHRH neuronal system, which is believed to be a key mechanism in induction of the preovulatory LH surge. Three different classes of EOPs exist in the brain; the proopiomelanocortin (POMC)-derived peptides (b-endorphin), the enkephalins and dynorphins. All these opioid-like peptides have more or less been shown to inhibit the reproductive system. A well known pathway is the b-endorphin circuitry, which originates from the hypothalamic ARC and projects to the MPOA, directly onto LHRH perikarya, and acts through the m-opioid receptor. This pathway is believed to restrain in a tonic manner the LHRH neuronal activity during stressful conditions (see below) and also during periods other than the preovulatory LH surge. Interestingly, b-endorphin neurons of the ARC have been shown to express ER (111), and estradiol has been reported to reduce POMC gene expression in the rat ARC (301). Whether estradiol targets these ARC opioid neurons to a more important extent just before the preovulatory LHRH discharge to cause ovulation remains an open question. Although simplistic, E2 may influence selective groups of opioidergic neurons to restrain the tonic inhibitory action of b-endorphin at the level of the MPOA, and/or ME is likely to participate in the cascade of events triggering the preovulatory LH surge and ovulation. Different excitatory peptides, acting on the neuroendocrine LHRH neuronal system, such as neuropeptide Y (NPY), galanin, neurotensin, D-sleep-inducible peptide (DSIP), angiotensin II (AII), along with inhibitory peptides including EOP, substance P and other members of the tachykinin family, act at multiple levels of the hypothalamus to orchestrate a general physiological clock-driven event—ovulation and the maintenance of the reproductive system (116). The complexity of this system is underlined by the fact that experimentation is performed under widely varying conditions using in-vivo and in-vitro models, in male or female animals, following gonadectomy or at various periods of the ovulatory cycle, in lower or higher species, and with a variety of selective chemicals that are used to either block or trigger specific receptors. Trying to put all these variables together may make the system seem more puzzling than it actually is.

Our approach to investigating the mechanisms involved in the regulation of LHRH neuronal activity involves five different levels (Fig. 7): 1) LHRH release is measured using a push-pull cannulae implanted into the ME during proestrus in rats; 2) the site(s) of action of various factors is(are) analyzed by means of bilateral microinfusion cannula located in different regions of the brain (PVN, ARC, MPOA, etc.); 3) the functional neuronal activity, as well as the transduction and transcription mechanisms, are investigated in vivo by examining the expression of nuclear immediate-early genes and LHRH primary transcript; 4) LHRH biosynthesis is estimated by means of in situ hybridization histochemistry, using exonic and intronic probe technology; 5) the cellular phenotype is demonstrated using single and dual immunocytochemistry. Our particular interest in using these contemporary in vivo techniques was to yield a better understanding of how stress of many kinds may deeply affect the neuroendocrine LHRH system, as well as the reproductive function. A brief summary of our results and those from other groups will now be presented.

In women, a disorder called hypothalamic anovulation occurs when the hypothalamus does not produce LHRH, which in turn results in a lack of egg production and release by the ovaries. Similarly, if the LHRH secretion pattern is altered by prolonged stressful situations, a female will show symptoms of dysmenorrhea or amenorrhea, while a male will exhibit alteration of steroidogenesis as well as spermatogenesis. The incidence of secondary amenorrhea in elite female athletes participating in sports such as marathon running, gymnastics, ballet dancing, or other sports is an example of stress-induced alteration of reproductive functions. However, whereas alterations in the secretion of LHRH in women are manifested dramatically and unequivocally as hypothalamic amenorrhea, the signs of hypogonadism in men may not be clinically apparent, while a deficiency in hypothalamic LHRH may exist (12,152). Among the large body of clinical evidence indicating that inhibition of reproductive function in women (low levels of plasma estrogen) may cause or accelerate health problems, the increased risk of osteoporosis and subsequent fractures in older women after natural menopause is well known (148,149,257). The physical and emotional components of a prolonged stressful situation contribute to a series of mechanisms capable of modulating endogenous activity in LHRH neurons.

It is important to mention here that the consequences of specific stressful situations may determine the impact on LHRH neuronal activity and gonadal function. For example, severe stressors, such as strenuous exercise, are known to induce anorexic behaviors and lead to a decrease in body weight and fat (215,220, 222,223). Is the alteration of reproductive functions the result of nonspecific inputs associated with the regulation of energy balance or the consequence of specific metabolic signals having either direct or indirect influences on the reproductive axis? This point, which was recently discussed by Dr. Knobil (129), is important because there is evidence supporting the hypothesis that the signals that suppress normal LH secretion after brief periods of fasting are related to the metabolic status of the body during the transition from fed to fasted states, rather than a function of the psychological state imposed by withholding food (254,255). However, stress can also affect the reproductive axis when the energy balance is not significantly perturbed. For example, following an exercise training program, the food intake of male rats is profoundly decreased, whereas no alteration is observed between exercised and control female rats, suggesting that female are much more resistant than male rats to stress-induced anorectic behavior (215,220,222,223). Thus, in terms of energy metabolism, rather than undergoing the consequences of stress such as a decrease in body energy gain, female rats seem to adapt to stress without perturbing their food intake behavior, which largely prevents the alteration of their body energy gain as well as their growth. On the other hand, intense exercise training is known to interfere with the activity of the HPG axis and the reproductive function in both genders (5,6,26,98,258). Therefore, the effect of stress per se on the reproductive axis can also be dissociated from alterations in energy balance, at least in female rats. We have recently observed that food deprivation for 48 h did not cause significant changes in LHRH neuronal activity and gene transcription in the OVLT/MPOA in intact cycling animals, but a significant decrease in the levels of LHRH primary transcript was found in rats exhibiting an impairment of reproductive cyclicity (Fig. 8). Indeed, we failed to detect a regular proestrus vaginal smearing in a group of cycling female rats (35%) subjected to food deprivation on diestrus day-1 and for the subsequent 48-hour period (183). It is likely that the metabolic signal differs from one animal to another, and that the perception of the challenge was more stressful in some animals. The metabolic and neurogenic stressful conditions may together deeply perturb the LHRH neuronal system and the estrous cycle. However, 65% of rats maintained their cycle intact, and little alteration of the neurons controlling the reproductive functions was detected (183). Whether these rats were less sensitive to nutritional-energetic signals or more resistant to stress has yet to be answered. Powers and colleagues recently evaluated the combined effect of food deprivation and running wheel access on the estrous cycle of female hamsters (209). They observed that food deprivation on days 1 and 2 of the estrous cycle disrupted the next expected ovulation, and this effect was more, rather than less, robust in female allowed to exercise in running wheels while they were deprived (209). This indicates that combined metabolic and neurogenic stressors may deeply exaggerate the inhibitory signals onto the neurons controlling reproductive cyclicity.

Why can stress affect the HPG axis? Are there any physiological reasons that explain how reproductive functions can be perturbed by intense stress? As early as 1939, Selye (260) observed that stress is accompanied by both an increase in the activity of the hypothalamic-pituitary-adrenal (HPA) axis and a decrease in reproductive functions, a phenomenon he attributed to the necessity, in case of emergency, of preserving adrenal cortical function at the expense of gonadal activity. We can indeed postulate that, during intensely stressful conditions, the organism preserves all the energy necessary to maintain homeostasis. Another possible hypothetical explanation is that a period of severe stress is not propitious for reproduction, and thus specific endogenous mechanisms will modulate LHRH neuronal activity during such periods.

Immune challenge is one of the stressful circumstances able to perturb reproductive function, which can involve similar or completely different mechanisms than those observed during other types of stresses. The immune-system derived monokine, interleukin-1 (IL-1), stimulates release of CRF from the hypothalamus, and the consequent stimulation of ACTH release provides an especially useful model for investigating the nature of the inter-communication between neuroendocrine and immunological pathways. In addition, IL-1b is probably one of the most potent polypeptides in the CNS for inhibiting LHRH neuronal activity and the HPG axis (for review, see 229). Nevertheless, the physiological relevance as well as the exact involvement of the central IL-1 system on LHRH biosynthesis, LHRH release, gonadotropins, sex steroid production and the ovulatory process during immune challenge and other types of stress are still open questions. Improvement in our knowledge of the central IL-1 network will help us uncover, at least in part, the physiological relevance of this newly recognized polypeptidergic system in the mammalian brain. Indeed, central cytokines may represent a new class of neurotransmitters in the hypothalamus or other brain regions which participate in the control of the neuroendocrine functions that maintain homeostasis during exposure to stressful signals.

Among the many stress-related factors produced and released by different structures of the brain, CRF plays an important role in modulating the ability of certain stresses to alter LHRH neuronal activity. It is now well known that intracerebroventricular (i.c.v.) injection of CRF suppresses the activity of the HPG axis in lower and higher species. The concept that endogenous CRF might play a role in modulating stress-induced inhibition of the HPG axis was originally suggested by the observation that i.c.v. injection of CRF suppresses the activity of this axis (191,200,217,236). Moreover, using castrated male rats exposed to mild electroshocks, i.c.v. administration of a selective antagonist of CRF receptors prevented the reduction in plasma LH levels caused by this particular stressful situation (235). On the other hand, Maeda et al. (153) recently reported that i.c.v. treatment with a-helical CRF9-41 reinstated the suppressed LH release in fasted rats (48 h of fasting) but had no effect in unfasted ovariectomized rats treated with estradiol. This CRF antagonist can also abolish the decline in circulating LH levels observed during a different type of challenge: suckling (294). Although the intravenous (i.v.) injection of CRF antisera inhibits the increase in ACTH levels normally measured during suckling, this treatment does not alter the decrease in the secretion of LH by lactating dams (294). These results suggest that central CRF release is involved in the alteration of plasma LH release during some types of challenges (neurogenic and metabolic stresses), and that the changes in LH secretion observed in these paradigms are probably not mediated by an effect of CRF at the level of LHRH release from nerve terminals in the median eminence (ME) [294]. However, it is important to not generalize these findings to all stresses because the inhibitory influence of immune stresses on LH seems to be CRF-independent (229).

It was tempting to speculate that the PVN of the hypothalamus had a leading role in such interplay between stress-related circuitries and alteration of reproductive axis; 1) this endocrine nucleus is the principal site of parvocellular neurosecretory neurons responsible for delivering CRF to the hypophyseal-portal system, the event that drives the activity of the pituitary-adrenal axis (HPA) during emergency conditions (249). 2) Using c-fos as an index of cellular activity, stressors of many kinds cause a profound activation of these neurons (32,50,194,228,230). 3) Specific challenges stimulate transcription of the CRF gene selectively within this neuroendocrine nucleus (218). 4) As described above, CRF acts as a potent antireproductive hormone during various stressful conditions, including neurogenic and metabolic stresses. However, CRF of PVN origin does not seem to be involved in the alteration of LHRH and LH release caused by different challenges, including footshock stress and central IL-1 injection (224). Moreover, we have recently found that acute and chronic treatment with metyrapone (an inhibitor of glucocorticoid formation) induced a significant increase in PVN CRF gene expression and the HPA axis, but did not alter plasma LH levels in ovariectomized monkeys (290). These results suggest two things; 1) The increased influence of the HPA axis and decreased influence of the HPG axis during stress may not be related. 2) The CRF involved in the alteration of LHRH neuronal activity during stress of many kinds does not originate from the PVN or, if it does, it is not the subset of parvocellular neurons projecting to the infundibular system which participates in the phenomenon.

We have reported that CRF attenuates LH secretion through a central mechanism involving the inhibition of LHRH neuronal activity within the MPOA of the hypothalamus (221). In contrast, this stress-associated neuropeptide does not appear to have an inhibitory effect at the level of LHRH nerve terminals in the ME which secrete into the hypophyseal-portal system. Infusion of CRF directly into the ME did not significantly alter hypothalamic LHRH secretion during proestrus in rats (221). In addition, bilateral infusion of CRF into two hypothalamic nuclei caudal to the MPOA, the hypothalamic PVN and the ARC, did not notably modify LHRH release measured via a push-pull cannula located in the ME (221). Thus, the hypothalamic site of infusion appears to be critical for the inhibitory influence of CRF on the activity of the HPG axis in female rats.

In the past two years, two distinct CRF receptors have been isolated; type 1 is widely distributed throughout the brain (208,219), whereas type 2 has two different splice variants, a shorter (also called a) and a longer form (also called b) [151,199]. CRF2a receptor mRNA is located in a few defined structures of the limbic system, such as the lateral septum, ventromedial hypothalamus and the medial amygdaloid nuclei. Type 2b is expressed in several peripheral tissues but not in the brain (199). Unfortunately, LHRH cell bodies of male and female rats do not seem to express either of the CRF receptor subtypes under basal and stressful conditions and during any phase of the estrous cycle (Nappi and Rivest, unpublished data). Although we have found some CRF type 1-positive neurons in proximity to LHRH-immunoreactive neurons of the MPOA, it is not yet possible to confirm whether CRF acts directly on LHRH perikarya via a selective CRF receptor during stress. Anatomical evidence supporting the hypothesis that CRF acts directly onto LHRH neurons to interfere with the reproductive system during specific challenges is therefore still lacking. The action of the neuropeptide through its type 1 receptor on neurons located in proximity to LHRH perikarya of the MPOA should be seriously considered.

Although the interaction between selective, still undefined, subset(s) of CRF neurons and the LHRH neuronal system during stress has been frequently investigated, the estrous cycle itself is capable of modulating different neuroendocrine stress responses, including the HPA axis and transcription of the CRF type 1 receptor. In female rats, immobilization-induced CRF1 receptor expression in the PVN is higher the morning of proestrus, and a large number of positive neurons for CRF1 receptor transcript are colocalized in CRF-immunoreactive (ir) perikarya (182). The fact that acute neurogenic challenges induce the gene encoding the CRF1 receptor selectively within the PVN, and that basal expression of that gene in most of the brain structures is not regulated in a stress-dependent manner or influenced by the gonadal hormonal milieu, support the concept that the ovulatory cycle induces neuronal changes capable of interfering with the influence of stress on the activity of neuroendocrine CRF neurons. We have also observed by means of CRF intronic probe technology that acute immobilization causes production of CRF primary transcript (hnRNA), selectively in the parvocellular PVN, without affecting other sites of the brain, an effect more pronounced in the early afternoon of proestrus. These data indicate that, at least during an acute neurogenic challenge, the PVN seems to be a primary site exhibiting activation of CRF gene transcription in cycling female rats. Gonadal steroid changes occurring throughout the cycle may modulate this response. CRF itself might participate in the regulation of neuroendocrine CRF motoneurons in cycling female rats; the gene encoding the CRF type 1 receptor is induced at a different magnitude throughout the cycle within CRF perikarya located in the dorsomedial PVN of stressed rats (182).

A possible link between the neuroendocrine events leading to LHRH peak and neuroendocrine CRF activity during proestrus may therefore be hypothesized. Vamvakopoulos and Chrousos (288) recently proposed that the estradiol-induced enhancement of CRF neurons may help to explain the negative feedback of estradiol on the LHRH neurons, which can occur throughout the activation of CRF-containing cells. Evidence is still lacking for a direct role of CRF in the negative influence of estradiol on LHRH neuronal activity and the HPG axis; there is no anatomical evidence, so far, that LHRH neurons of the rat MPOA are innervated from PVN CRF neurons and i.v. and i.c.v. administrations of various types of CRF antagonists, including CRF antisera, does not interfere with LHRH and LH proestrus surge as well as basal plasma LH levels in intact and ovariectomized female rats. The physiological relevance of an enhanced activity of CRF biosynthetic machinery in proestrus-stressed animals is thus unknown and quite intriguing. It is possible that influence of the ovulatory cycle on stress-induced activation of neuroendocrine CRF is an additional mechanism to insure adequate control of the homeostasis during severe emergency circumstances in the female. On the other hand, the hypothesis that CRFergic circuits participate in the control of neuroendocrine events responsible for regulation of reproductive cyclicity in a sex steroid-dependent manner still awaits to be clearly demonstrated.

Inhibition of LHRH (200) and LH release (201) induced by i.c.v. CRF injection is mediated, in part, by activation of the central EOP system. We have also shown that this action occurs at the level of the hypothalamic MPOA, as bilateral infusion of m- or m1-opioid receptor antagonists into the MPOA partially inhibited the CRF-induced alteration of hypothalamic LHRH and pituitary LH release in OVX and intact rats during the afternoon of proestrus (221). In addition, the effect of EOPs seems to be specifically via b-endorphin, the major ligand for m receptors. As mentioned before, the interaction between EOPs and the LHRH-secreting system is well known and has been reviewed in detail. EOPs are able to decrease LHRH concentrations (247) as well as perturb the release of LHRH into the hypophyseal portal system of the rat during proestrus (35) and to alter the electrophysiological activity of the hypothalamic LHRH pulse generator in the Rhesus monkey (121). It has also been proposed that opioids act directly on LHRH neurons through specific m-opioid receptors (69, 210). Almost 10% of synapses impinging onto LHRH neurons in the rat diagonal band/POA contain b-endorphin (33). Despite the fact there is approximately 3-4 fold more b-endorphin input to the LHRH cells and processes in female than in male rats, b-endorphin exerts a robust synaptic influence on LHRH neurons in both sexes (34). In addition, POMC peptide-producing neurons capable of synthesizing b-endorphin in the ventromedial ARC project to the MPOA of the hypothalamus, and some of these neurons establish direct contacts with LHRH-containing perikarya in the rat (144). CRF is an important regulator of the synthesis of POMC-related peptides within the adenohypophysis (23,232,286) and the hypothalamus (188), and the ARC POMC neurons could also be targeted by CRF. Indeed, a low but positive hybridization signal for the CRF type 1 receptor was detected within the ARC close to the wall of the third ventricle (208), although whether these cells contain POMC has yet to be verified. It is possible that b-endorphin is released from POMC neurons originating from the ventromedial ARC of the hypothalamus. This release would consequently activate mainly m1 opioid receptors and thus alter the neuroendocrine LHRH neuronal system. However, administration of opioid receptor antagonists bilaterally into the MPOA failed to completely reverse the inhibitory influence of CRF on LHRH and LH release in female rats (221), indicating that the effects of CRF on LHRH neurons may also be due to a direct action and not to a prior activation of the opioidergic system. In addition to the m opioidergic system, CRF may also modify the secretion of other peptidergic or aminergic systems present in the MPOA and, in turn, indirectly alter LHRH neuronal activity.

The discovery that functional bilateral pathways exist between the immune and neuroendocrine systems has been one of the most fascinating developments aiding our understanding of the regulation of the homeostatic balance of living organisms challenged by foreign materials. Physical, emotional and environmental stimuli, including infection, can harm bodily integrity, and a complex network operating at the level of hypothalamus coordinates the appropriate metabolic, behavioral and endocrine changes necessary for the restoration of homeostasis. However, when noxious signals overcome the ability to restore physiological balance in mammalian organisms, one of the final consequence may be an impairment of reproductive processes. Increased production of cytokines, proteins released by activated macrophages and lymphocytes upon presentation of an antigen, represents a central step of the early events of immune activation (the acute-phase response). Apart from the diverse and overlapping activities that these circulating lymphocyte-derived mediators exert on biological functions during the immune response (59), a potent action of cytokines on neuronal function, behavior, the neuroendocrine system and metabolism has been demonstrated. Cytokines are systemically produced and may also be synthesized and released directly within the brain to interact with both neurons and glia. The most extensively studied cytokine in the brain, IL-1, is an important neuromodulator exhibiting pleiotropic biological actions and having the capacity to deeply influence neuroendocrine functions. Interestingly, brain production of cytokines is stimulated not only by systemic immune challenge (138), but also after exposure to a neurogenic immobilization stress (169). These brain-derived cytokines may therefore be part of the stress-related pathways and mediate the information received from the periphery, in particular from circulating immune-related substances. Whether brain-derived IL-1 is involved in the physiologic control of the HPG axis or whether its secretion is activated only in response to stressful stimuli are currently the subject of profound controversies.

To discuss the influence of cytokines on reproductive function, we report here a series of in-vivo experiments that, even though they were performed at pharmacological doses, seem valuable for probing the potential physiological neuroendocrine effects of IL-1 (a and b) on the activity of LHRH neurons. As discussed earlier, the large majority of LHRH neurons of the MPOA that express the IEG Fos in their nuclei during the afternoon of proestrus project to the ME (229), indicating that this particular set of neurons is a neuroendocrine one and is directly related to the control of pituitary gonadotropin release. Having the capacity to interfere with this spontaneous phenomenon (227), the cytokine IL-1 is recognized to deeply inhibit the reproductive axis in altering neuroendocrine LHRH neurons at the level of their perikarya. This phenomenon, in itself, is quite exciting because, although the cytokine mainly originates from systemic immune cells, it only produces its effects on LHRH neurons when injected directly into the brain, not into the systemic circulation (237). The physiological relevance of such a phenomenon is unknown, but it permits us to postulate that IL-1 of central origin essentially can alter the activity of LHRH neurons. The fact that the brain itself is a source of IL-1 (138, 281) may suggest that the polypeptide plays a determinant endogenous role in interfering with the mechanisms involved in the activation of LHRH motoneurons in the MPOA during the ovulatory cycle.

The influence of IL-1 on the activity of the HPG axis depends, in fact, on the route of administration; systemic administration of IL-1a and b strongly stimulates the activity of the HPA axis (239) but fails to modulate the release of plasma LH from the anterior pituitary of both castrated male and female rats (229). On the contrary, central administration of IL-1 inhibits the spontaneous expression of the nuclear Fos protein in LHRH-containing cells during the afternoon of proestrus and interferes with the LHRH release from the hypothalamus (227). In addition, this treatment is also able to block the progesterone-induced LH release in ovariectomized rats and the preovulatory LH surge in cycling rats (237,238). As mentioned, IL-1b seems to interfere with LHRH neuronal activity at the level of LHRH perikarya, rather than at the level of LHRH nerve terminals located in the ME. Indeed, microinfusion of IL-1b bilaterally into the MPOA of proestrus rats bearing a push-pull cannula within the ME reduces LHRH release and the spontaneous expression of the Fos protein within the nuclei of LHRH neurons (229). On the other hand, peripheral administration of the endotoxin lipopolysaccharide (LPS), which increases the endogenous release of both peripheral and central cytokines, significantly decreased circulating LH levels (231). Interestingly, i.c.v. treatment with IL-1 antibody significantly attenuated the endotoxin-induced decrease in plasma LH levels (70), which provides the evidence that local CNS production of IL-1 is playing a role in the alteration of LHRH neuronal activity during the acute-phase response of an immune challenge.

It is interesting to note that the effect of IL-1b on the endogenous expression of Fos within LHRH neurons is time- and dose-dependent. Central (i.c.v.) injection of 50 ng IL-1b at noon dramatically inhibits the spontaneous expression of nuclear Fos in LHRH neurons as effectively as an i.c.v. infusion of the same dose every 2 hours from 8.30 to 16.00, whereas a similar single injection at 8.30 or at 14.30 is less effective in preventing the expression of Fos within LHRH neurons (227, 229). These results indicate firstly that the endogenous transsynaptic and intracellular transduction processes leading to Fos synthesis within the nuclei of LHRH neurons take place during a very short period of time in the afternoon of proestrus. Secondly, high doses of IL-1 within the brain are needed to impair LHRH neuronal activity, a phenomenon that can raise some questions in regard to the physiological significance of the influence of this effect. Moreover, whether IL-1-induced inhibition of c-fos within LHRH neurons is responsible for the decrease in LHRH mRNA levels, and whether production of IEG is necessary for LHRH gene transcription are questions that remain to be answered.

Interestingly, we have recently found that activation of the acute-phase response of an immune challenge during the afternoon of proestrus can interfere with LHRH neuronal activation and gene expression in the neurons located in the OVLT/MPOA. Indeed, the percentage of both LHRH-ir and LHRH mRNA neurons spontaneously expressing Fos protein in their nuclei at this time was significantly inhibited following endotoxin administration (183). LPS was also able to significantly reduce the increase of LHRH primary transcript occurring at 15.00 on the day of proestrus, without affecting this transcript during other periods of the estrous cycle (Fig. 8). These data support the concept that the biosynthetic machinery of the gene encoding LHRH is activated during the preovulatory LH surge, and that immune challenge can interfere with LHRH neuronal activity at the transcriptional level. The exact participation of the IEG c-fos in these events remains hypothetical. Parallel alteration of both Fos and LHRH gene expression was detected in the group of neurons analyzed in this presented study (183). Because c-fos is a transcription factor generally recognized to activate genes when forming an heterodimer with members of the c-jun family, and because an AP-1 DNA binding site is present on the LHRH promoter (see above), it is tempting to speculate that both phenomena (inhibition of Fos and LHRH transcription) are related. As discussed before, however, controversies exist regarding the stimulating or repressing role of AP-1 on LHRH gene transcription.

Although neuroendocrine CRF is markedly activated by the cytokine (74, 230), total ablation of the PVN does not prevent IL-1-induced decrease in LH release in castrated male rats (224), and administration into the lateral ventricles of different types and doses of CRF antagonists does not modify the ability of IL-1 (a and b) to inhibit LHRH neuronal activity or decrease plasma LH levels in male or female rats (227). These results would suggest that the effect of IL-1 on the activity of neuroendocrine CRF and LHRH is probably independent. Considering the fact that CRF is a potent inhibitor of LHRH neuronal activity and the HPG axis, the lack of any direct interaction between the HPA and the reproductive axes, which are both profoundly influenced by activation of the acute-phase response is quite intriguing. We cannot exclude that gender and species differences exist in the antireproductive action of CRF in immune-challenged animals. In the monkey, for example, CRF seems to participate in the inhibition of reproductive functions by IL-1; CRF antagonist was indeed able to prevent the decrease in circulating LH levels observed in response to IL-1 injection (80). However, the dissociation between the HPA and HPG activities according to the route of IL-1 administration in the rat remains an important issue, and the physiological relevance of such phenomena are far from clear. It is possible that the target regions and mechanisms involved in the inhibitory effect of IL-1 on LHRH neuronal activity display an adjustable sensitivity, depending on several variables, such as the severity and the nature of the immune challenge. It is also highly possible that the capacity of the organism to preserve the reproductive system under emergency circumstances differ between species, and that the mechanisms involved in this phenomenon characterize these differences.

We have recently observed that eicosanoid prostaglandins (PGs) play a crucial role in mediating the inhibitory effects of central IL-1b injection on LHRH neuronal activity. Indeed, administration of indomethacin, an inhibitor of the eicosanoid cyclooxygenase pathways, significantly reversed IL-1b-induced alteration of the HPG axis (226). In contrast, injection of a lipooxygenase inhibitor, nor-dihydroguaiaretic acid (NDGA), did not significantly modify the effect of IL-1b on plasma LH secretion (226). Stimulation of various brain cells by cytokines primarily results in the release of PGE2, which is reportedly elevated in the circulation during sepsis. Moreover, central PGE2 injection causes specific and selective expression of c-fos in several brain structures recognized to be activated in the brains of endotoxin-challenged rats (137). It is therefore possible that PG of E2 type plays a crucial role within the CNS in the interface between the immune and nervous systems to modulate neuroendocrine responses, such as the HPG axis. Upon observing that IL-1a blocked the release of PGE2 evoked by exposure of mediobasal hypothalamic fragments to norepinephrine, Rettori et al. proposed that the cytokine suppresses LH secretion through mechanisms depending on the inhibition of PGE2 production (214). This conclusion rests on the hypothesis that catecholamine-induced PG synthesis, which stimulates LHRH secretion (190), represents a major step in mediating the ability of IL-1a to inhibit the activity of the HPG axis. The fact that indomethacin is capable of blocking the IL-1-induced inhibition of the HPG axis does not provide evidence that the possible decrease in the production of certain PGs is responsible for this effect, but, in contrast, strongly suggests that the overall activation of PGs modulates the influence of IL-1 on LHRH neuronal activity. However, the exact mechanisms by which PGs may alter the LHRH neuronal activity during immune challenge or other stressful situations have yet to be clarified.

IL-1b is widely distributed in the rat brain: IL-1b neurons have been identified in the magnocellular part of the PVN, and projections deriving from these cells reach the suprachiasmatic nucleus, the ARC, the ME, the parvocellular part of the PVN and the posterior pituitary (139). The hybridization signal for the mRNA encoding IL-1b has been observed in several regions of the hippocampus, the cerebellum, the ventromedial hypothalamus and the frontal cortex (9). Glial cells also produce IL-1a and b in response to brain infection and inflammation, as well as to systemic immunological challenges (138,281,302). Nonetheless, it is difficult to reconcile the potential relevance of changes in brain-derived cytokines as they relate to neurons controlling the reproductive axis. The picture is further complicated by the distribution of the type I IL-1 receptor (IL-1R) transcript in the rat brain. In contrast to the widespread distribution of cells expressing the gene encoding type I IL-1R in the mouse brain (51), in the rat brain, this receptor has recently been shown to be expressed in few discrete cell groups. In situ hybridization revealed a clear labeling over barrier-related cells, including the leptomeninges, non-tanycytic portion of the ependyma, the choroid plexus and the vascular endothelium (75). Low to moderate levels of the type I IL-1R mRNA were detected in few neuronal cell groups, including the basolateral nucleus of the amygdala, the ARC/ME, the trigeminal and hypoglossal motor nuclei and the area postrema (75). Interestingly, no specific labeling for type I IL-1R mRNA was detected in neurons that normally respond to systemic IL-1b injection, such as neuroendocrine CRF cells and brainstem catecholaminergic cell groups (75). In addition, no basal type I IL-1R expression was present in hypothalamic structures directly related to the activity of the HPG axis, such as the MPOA. Activation of target structures by IL-1 may therefore take place at vascular and/or other barrier-related sites through which secondary pathways can interact with this neuroendocrine response. The absence of receptor for IL-1 in hypothalamic nuclei, with the exception of the ARC/ME, seriously questions the possibility that blood-born IL-1 can directly target LHRH perikarya to influence its activity. Since IL-1 interferes with LHRH and LH only when injected centrally (229), systemic IL-1 binding to its receptor located in the ARC/ME cannot be directly related to the inhibition of reproductive function. It is, however, possible that IL-1 of central origin displays a peculiar behavior, in comparison to IL-1 of systemic origin, by diffusing differently into the ARC with the ability to trigger specific pathways involved in the control of LHRH neurons.

Clear anatomical connections between the immune and neuroendocrine systems have still not been found. Keeping in mind that the MPOA represents a primary target region for central effects of IL-1b in modulating the activity of LHRH neurons that project to the ME, attention has to be focused on the substances that may mediate the effects of IL-1 on reproduction. EOP, prostaglandins, excitatory amino acids and catecholamines are likely to be such mediators. The influence of IL-1a and b on the reproductive axis may therefore be the result of a complex cascade of mechanisms that can originate from various structures of the brain, including hypothalamic LHRH sites. A detailed description of this rather complicated circuitry is behind the scope of this brief chapter; for further information, we recommend several recent reviews (225,229,233,234).

In addition to the central effects described above, immune challenge may influence the LHRH system in altering the genetic expression of the pituitary LHRH receptor (LHRH-R) [Fig. 9]. Indeed, we have recently reported that systemic LPS administration caused a profound LHRH-R mRNA downregulation in the anterior pituitary, regardless of the phase of the estrous cycle (183). This suggests that activation of the acute-phase response is able to interfere with the HPG axis also at the pituitary level independently of the circulating gonadal steroid concentrations. Interestingly, systemic IL-1b injection does not affect LH secretion (237), whereas peripheral treatment with the endotoxin strongly decreases circulating LH levels (70,226). It is therefore likely that a series of cytokines and related physiological events during the acute-phase response are required to decrease the activity of the HPG axis. The concomitant actions of several circulating factors probably operate at the level of LHRH producing sites in the hypothalamus and/or the pituitary gland (as paracrine activity) to mediate the inhibition of reproductive function in response to immunogenic challenge.

Stimuli threatening homeostasis, such as physical stress or pathogens, cause immune, endocrine, metabolic and behavioral changes that restore and maintain the consistency of the "milieu intérieur." However, the capacity of these systems to regulate and preserve homeostasis can be exceeded by the intensity of the stressful situation. This can, in turn, induce pathological changes in the organism. The possible relationship between severe and prolonged stress and reproductive failure has long been investigated and, in particular, marked changes in reproductive function have been recognized during inflammatory and infectious diseases. Ovulatory dysfunction, a decline in fertility, an increased incidence of spontaneous abortion, abnormal progeny and sexual dysfunction are among the most common consequences of these stressors in human reproduction. Severe, prolonged, and/or unpredictable stresses can also increase the organism's susceptibility to a broad range of psychiatric disorders, including anorexia and depression, behaviors frequently observed in amenorrheic patients. Obviously, the consequences of uncontrolled challenges are detrimental for the organism and alteration of the reproductive system may be the result of a series of combined disorders associated with the stressful events. In fact, the behavioral pattern of stress (performance impairments, insomnia, lost of energy, lost of interest or pleasure, lost of appetite, etc.) is usually associated with reproductive failure and, from a philosophic point of view, such situation may not provide the best environment for reproduction. Consequently, a therapeutic approach that prevents only the alterations in the LHRH neurosecretory system under these disturbed circumstances is perhaps not a suitable intervention, considering the system as a whole. On the other hand, there are many clinical applications for use of LHRH in humans. In the next section of this chapter, we provide a brief description of some of the therapeutic aspects of this neuropeptide.

Since the elucidation by Schally and coworkers (158,251) of the structure of porcine LH- and FSH-releasing hormones in 1971 and the subsequent development of synthetic analogues with increased receptor affinity and resistance to enzymatic degradation, extensive clinical data have been reported using GnRH agonists in situations ranging from contraception to the treatment of hormone-dependent cancers (Table 1). The original incentive for the development of potent GnRH agonists was the belief that the powerful gonadotropin stimulatory effects of these agents could eventually have beneficial pro-fertility therapeutic applications (107,178,252,259). However, initial studies in animal models (Fig. 10 and Fig. 11) soon established the anti-reproductive effects of these synthetic molecules and laid the groundwork for clinical exploration of circumstances where gonadal inhibition was desirable (4,10,45, 52,126,198,244). Treatment with GnRH agonists induces luteolysis, hypoestrogenism and inhibits ovulation in normal women (81,133,180,252); the same treatment decreases testicular steroidogenesis (Fig. 12) in males (44,132,134,305).

Native GnRH is released from the hypothalamus into the hypothalamo-hypophyseal portal vessels in an episodic (pulse) mode at approximately 90-minute intervals (84,128,305). This pattern of release is associated with pulsatile electrical activity in the hypothalamus (128,205,206). GnRH release leads to episodic release of LH and FSH from the anterior pituitary gland into the systemic circulation after binding to specific GnRH receptors on the cellular surface of gonadotrophs (39). Receptors occupied by GnRH aggregate into clumps (‘coated pits'), and the receptor-ligand complex is internalized (7,29,43,44,102,157,181). Cellular responses to GnRH activities of pituitary gonadotrophs are mediated by plasma membrane "second messenger" systems, based on calcium-calmodulin and protein kinase regulatory processes (7,21,29,41,43,44,102,157,180,181).

In animal models and humans, these physiological events can be disrupted by continuous exposure to native GnRH or administration of long-acting analogs, causing what is termed desensitization to pituitary gonadotrophs (17,19,55,181,242,243,272). Constant exposure to native GnRH attenuates gonadotropin release by downregulating the number of GnRH receptors and/or uncoupling the receptor from distal signal transduction pathways (21,29,41,44). The same mechanisms are probably responsible for desensitization induced by GnRH agonistic analogs but may also involve other mechanisms of downregulation, as seen for b-adrenergic receptor in response to excessive agonist stimulation (41,180). Briefly, when a GnRH agonist is administered, it first leads to supraphysiological LH and FSH secretion and initial stimulation of gonadal function. However, due to the continuous presence of the agonist, increased ligand binding causes maximal physical internalization of agonist-occupied receptors, disrupting gonadotroph receptor replenishment mechanisms and second messenger processes and leads to decreased LH and FSH secretion, with a subsequent reduction in gonadal steroid secretion. Although pituitary desensitization is probably the principal mechanism responsible for decreased gonadal steroid secretion following GnRH agonist administration, other mechanisms may be involved in this phenomenon (e.g., loss of GnRH pulsatile secretion and subsequent disruption of gonadotropin episodic secretion [55,81,84,185], changes in serum immunoreactive and bioactive LH concentrations [119),273,291] and gonadotropin [LH] gonadal receptor desensitization [4,52,126,132,133]). Direct inhibitory effects of GnRH agonist at the gonadal level could also play a role (105,250,306).

The knowledge of the structure of GnRH offered the opportunity to design and synthesize peptides with enhanced biological activity. The success in the endeavor has been particularly impressive; within five years of the discovery of the GnRH structure, superagonists having 100–400 times the in-vivo biological activity of natural GnRH were already available (46,118). Today, all the GnRH agonists possess equivalent potency and efficacy for inducing the hypogonadal states required for treatment of hormone-dependent diseases or neoplasms, including mammary and prostate tumors. Modifications of positions 6 and 10 of the peptide can confer greater stability against enzymatic degradation, increased receptor affinity and biological potency (88,118,173). All clinically available GnRH agonists contain a D-amino at position 6, while either ethylamide (-NHEt) or azaglycine (Azagly-NH2) substitution (i.e., Lupron and Zoladex, respectively) is found in position 10 (Fig. 13). GnRH agonists are active only when given parenterally (subcutaneous, intranasal or intramuscular) [Fig. 19]. Most of the initial clinical studies performed with LHRH agonists used the subcutaneous or intranasal route on a daily basis. Since then, significant improvement has been made with depot formulations using biodegradable copolymer Rods (Zoladex) or microspheres (Lupron depot). These can be given at monthly or longer intervals (Zoladex-LA, 3-month depot preparation). These different delivery systems provide prolonged release following biodegradation and avoid long intervals of subtherapeutic serum levels, as well as repetitive injections.

Since the first description by Huggins and Hodges in 1941 (106) of the palliative effects of hormonal manipulation in men with prostate cancer, surgical and medical castration (via estrogens) have become standard therapy for advanced prostatic adenocarcinoma. This palliative approach achieves a temporary response in 60–80% of patients with advanced cancer. Surgical castration may have an important psychological impact in some patients, and estrogen therapy causes serious cardiovascular complications (25,28,90,106). GnRH agonists, on the other hand, provide a valid alternative for achieving potentially reversible testicular androgen suppression with minimal side effects (107,132,134,135). Following long-term treatment, GnRH agonists reduce levels of serum testosterone to those seen after surgical castration. They are as effective as surgical castration or estrogen therapy in advanced prostate cancer.

The major difference between medical castration achieved with GnRH agonists and surgical castration (orchiectomy) is the transient elevation of testosterone and dihydrotestosterone (DHT) which lasts for 7–10 days at the onset of treatment (Fig. 14). Following this initial stimulation, testosterone rapidly decreases to castration levels with maximal inhibitory effect (5% of normal) after 2–3 weeks of GnRH agonist administration. The initial increase of testosterone levels during the first days of LHRH agonist treatment of advanced prostate cancer may cause disease flare-up and prove detrimental for the patient. Reports of increased bone pain, urological symptoms, spinal cord compression, increased prostatic-specific antigen (PSA) levels, and death have all been reported during initiation of GnRH agonist therapy (282,296). The disadvantages of such a disease flare can be totally antagonized by the concomitant administration of a pure non-steroidal antiandrogen.

Although maximal suppression of testosterone level obtained with long-term GnRH agonists may be more acceptable to many patients than castration, when administered as monotherapy, they provide no additional therapeutic effectiveness over conventional treatment (castration/estrogens).

It is becoming more and more evident that androgens of extra-testicular origin (primarily adrenal androgen steroid precursors) contribute significantly to cancer growth in prostatic cancer patients. There is ample logic behind this theory; men secrete large amounts of inactive adrenal androgen precursors that can be converted intraprostatically to potent androgens, and therapeutic approaches meant to suppress adrenal androgens (adrenalectomy, glucocorticoids, ketoconazole or aminoglutethimide) have some proven efficacy (68,89,179,245,299).

Earlier studies reported that castration induced by orchiectomy or GnRH agonists caused a 95% reduction of testosterone serum levels, but the intraprostatic concentration of DHT (the metabolically active androgen) decreased only about 50–60% (89,135). In fact, in adult men, serum levels of the main metabolites of DHT, 5a-androstane-3a,17b-diol (3a-diol), androsterone (ADT) and their glucuronidated derivatives are reduced by 50–70% following castration (16). These observations clearly suggested that prostate tissue efficiently transforms the inactive adrenal steroid precursors dehydroepiandrosterone sulfate (DHEA-S), DHEA and androstenedione (D4-dione) into the active androgen DHT. Furthermore, prostatic tissue contains all the enzymatic machinery involved in the biosynthetic steps responsible for the formation of DHT from adrenal androgen precursors (42,130,131). Based on the above knowledge of the possible contribution of the adrenal to endogenous androgen formation in men, Labrie and colleagues introduced the concept of combination therapy (Fig. 15), consisting of the concurrent usage of a GnRH agonist or castration and a pure antiandrogen in advanced prostate cancer (134,135). Although this concept of complete androgen blockade (CAB) gave rise to considerable debate when introduced, the focus over the last decade in the treatment of prostate cancer has revolved around the anti-tumor effect of CAB and has become the therapy of choice in first-line advanced prostate cancer treatment.

Since the initial studies of Labrie et al. (133,134), the benefits of combination therapy have been confirmed by large-scale, randomized, double-blind, placebo-controlled studies (15,47,57,136; Fig. 16). These compared the effects of medical (GnRH agonist) or surgical castration plus placebo with the non-steroidal antiandrogens flutamide or nilutamide in advanced prostate cancer. The Southwest Oncology Group (SWOG, NCI) [47] reported the largest multicenter study, in which they compared the median time to progression and survival for leuprolide and placebo vs. leuprolide and flutamide. Analysis at 42 months of follow-up showed statistically significantly increased time to survival and progression in favor of combination therapy (CAB).

The European Organization for Research and Treatment of Cancer (EORTC) [57] compare the GnRH agonist goserelin (Zoladex) plus flutamide vs. orchiectomy alone and demonstrated the superiority of CAB in median time to progression and survival. Denis (56), in a large recent meta-analysis comparing castration (GnRH agonist or orchiectomy) with a pure antiandrogen, concluded that CAB had a definite benefit on median time to progression and survival. An important observation in these studies was the fact that late treatment with CAB was not as effective as early treatment, and the probability of long-term survival was much greater in advanced cancer patients with minimal disease or a relatively low number of bone metastases (15,47,57,136; Fig. 17).

Although the aforementioned studies are clearly in favor of complete androgen blockade and of a definite role of adrenal androgens in advanced prostate cancer growth, such treatment remains palliative (Fig. 18). Prostate cancer is the second leading cause of cancer death in men in the United States, and it is predicted that by 1997, 250,000 new cases will be discovered and 41,000 men will die of this disease. The only likelihood of a significant reduction in prostate cancer deaths is the treatment of localized disease before the appearance of metastases. Much progress has been made in the rational use of PSA and transrectal ultrasonography (TRUS) for early detection of prostate cancer. This has raised the possibility of developing efficient treatments for localized disease. Currently, there is an increased enthusiasm for radical prostatectomy, especially after the first report by Monfette et al. (174) that neoadjuvant treatment of localized prostate cancer with an LHRH agonist [D-Ser-Trp6, Des-Gly-NH210] LHRH ethylamide and the pure antiandrogen flutamide caused an important cytoreduction, characterized by post-operative downstaging of the cancer. Since this preliminary observation, numerous trials using neoadjuvant hormonal cytoreduction have reported significant (40–60%) reduction in prostate volume and over a three fold decreased rate of positive surgical margins (79,162,174,269, 270). Long-term follow-up of patients in these trials is needed before any conclusion can be drawn about the beneficial effects on survival of neoadjuvant CAB therapy in localized cancer. In addition to survival, some potential benefits of the neoadjuvant combination of GnRH analogs and antiandrogens are the relatively few side effects, other than those secondary to a reversible hypogonadal state, and the possible induction of cellular apoptosis of distant undetectable micrometastases.

Presently, several clinical trials are also evaluating the role of neoadjuvant and adjuvant CAB associated with external beam radiation therapy in the treatment of localized prostate cancer. Preliminary results are very promising and in favor of CAB in combination with both neoadjuvant and adjuvant therapy, compared with neoadjuvant therapy alone (92).

The complexity of the female reproductive system, its cyclic nature and its intricate function make women more susceptible to disorders associated with the reproductive tract. It is evident that GnRH agonist-induced pituitary-gonadal suppression will find a wider spectrum of clinical applications in women than men (Table 1). As a consequence, there are many more published clinical studies in the literature using GnRH agonists in women. Following the accumulation of clinical data, consensus has been reached for several indications of GnRH-induced gonadal steroid suppression in women; these are breast cancer, endometriosis, uterine fibroids and precocious puberty. However, other potential applications in such disorders as infertility, ovarian hyperandrogenism, premenstrual syndrome, ovarian and endometrial cancer are still being investigated.

Treatment regimens used today with GnRH agonists may have different profiles of inhibitory activity, depending mostly in the pharmacokinetic profile and relative biological potency of the analogues. The administration of potent GnRH agonistic analogues initially enhances gonadotropin release (LH,FSH), but continuous exposure causes pituitary gonadotroph desensitization, followed by ovarian quiescence characterized by an hypoestrogenic state (pseudomenopause). Although, GnRH agonists generally block ovulation and prevent normal follicular development, the degree of pituitary and ovarian suppression is dependent on the parenteral route of administration (delivery system) [Fig. 19], dosage and biological activity of the analogue used (118,141,146,189,202,292). Minimal effective doses and weaker agonists allow some degree of follicular growth (to secondary follicles) and, consequently, some estrogen secretion. Such modalities of GnRH agonist treatment may well find application in contraceptive methods. More marked downregulation (pituitary-ovarian suppression) is obtained by using the superagonists at higher dosages and by means of slow-release depot formulations to obtain constant drug serum concentrations. Filicori and colleagues (82) compared several long-acting depot formulations in normally menstruating women. All agonistic analogues tested (buserelin, triptorelin, leuprorelin, goserelin) significantly suppressed gonadotropin levels, pituitary responsiveness and caused castration levels of circulating estradiol. However, some variability in the degree of ovarian suppression was observed between individual women and between analogues used (Fig. 20).

Maximal ovarian inhibition by agonist therapy is certainly required in the case of sex-steroid hormone-sensitive tumors or other female reproductive tract disorders where gonadal steroid reduction to castration levels is warranted. Comparison of the published data on the ability of goserelin, (Zoladex), buserelin, (Superfact) and leuprolide (Lupron), three of the most popular agonists used today, to reduce circulating levels of gonadal steroids did not reveal any striking differences when given over the long-term (61,82,146!popup(ch58ref146), 213) These long-acting formulations suppress serum estrogen levels to the castration range within 3–5 weeks after initiation of the GnRH agonists.

Adverse effects of GnRH agonist treatment are related to the negative effects observed in hypogonadal states, such as surgical ovariectomy or menopause. Otherwise, GnRH agonists are well tolerated and free of serious side effects. Climacteric symptoms such as hot flushes, vaginal dryness and loss of libido are frequent but generally perceived as minor by patients, compared with the underlying disease for which the GnRH agonist is given. The main potential draw-back of estrogen deprivation during long term GnRH agonist therapy is the accelerated bone loss and the possibility of fractures associated with osteoporotic disease. Numerous studies using GnRH agonists in patients with various disorders, particularly endometriosis and uterine fibroids, have yielded conflicting results on GnRH-induced bone loss. Some reports have shown increased urinary calcium, creatinine and hydroxyproline, creatinine ratio (97,274) and increased serum calcium, alkaline phosphatase and osteocalcin levels (64,289) as indicators of accelerated bone turnover. Quantitative computerized tomography and dual X-ray absorptiometry studies have reported a wide range of alterations, from no change to bone density losses of over 8% (64,91,184,284). This latter figure represents more than 1% bone loss per month of GnRH agonist treatment since most therapeutic regimen last no more than six months. It is, thus, reasonable to conclude that the same degree of accelerated reversible bone loss occurs with GnRH agonist therapy as occurs during menopause or in oophorectomized patients.

Natural menopause is also associated with an increased incidence of coronary heart disease (159). The GnRH agonist induction of a protracted state of pseudomenopause may lead to changes in lipid metabolism that progressively lower the cardioprotective estrogenic effects. Current findings in the literature suggest minimal effects on lipid metabolism, with a tendency to slightly increased HDL cholesterol levels (274,284,287). Well controlled studies with larger populations are clearly needed to obtained consensus on the long-term effects of GnRH induced hypoestrogenic states. We will next review the clinical state of the art in applying treatment with GnRH agonists to breast cancer, endometriosis, and leiomyomas.

In most Western countries, breast cancer is the most prevalent form of cancer and accounts for up to 18% of mortality in women. In fact, it is predicted that 200,000 new cases of breast cancer will be diagnosed in USA in 1997, while almost 50,000 women are expected to die from the disease. The majority of breast cancers, whether in pre- or postmenopausal women, are initially hormone-dependent, and estrogens play a major role in development and progression of the disease (58,104,150,160). Beatson in 1896 (14) was the first to report the beneficial effects of bilateral ovariectomy in the treatment of advanced breast cancer. Since the description in the early 1970s of the estrogen receptor and its role as a predictor of clinical course, there has been a renewed interest in endocrine therapy for the treatment of breast carcinoma. Estrogens influence estrogen-dependent cells such as breast tissue by interacting with the estrogen receptor in the nucleus,eliciting a cascade of transcriptional regulatory events. Estrogens are highly potent mammary mitogens, and prolonged estrogen exposure has being shown to be an important risk factor for developing invasive breast cancer. Estrogen receptors are overexpressed in about 60–80% of invasive breast cancers and are associated with an increased response to endocrine therapy, prolonged survival rates, and lower risk of relapse (48,161,283). Evaluation of hormone receptor status in the cytosol of breast cancer tissue has given the possibility of better patient selection and increased the response rate to hormone ablation procedures.

Today, endocrine therapy of breast cancer is still directed towards inhibiting (progestins), ablating (ovariectomy) or interfering (tamoxifen) with estrogenic activity (Table 2). Although most procedures induce approximately the same overall rate and duration of response, selection on the basis of receptor status, menopausal status, disease-free interval, metastatic sites and previous exposure to endocrine manipulation, may yield different therapeutic results. Nevertheless, the overall response rate is about 30–40% and limited to 12–18 months with endocrine therapy (48,58,101,186). This could be due to the fact that incomplete estrogenic influence is achieved with the present therapeutic means, and that adrenal steroids and local biosynthesis of estrogens by neoplastic tissue play a much greater role than expected (131,167,195).

In premenopausal women suffering from breast cancer, surgical castration still remains the standard first and second line endocrine therapy for advanced disease. The response rate to oophorectomy is reported to be about 30%, and this increases in older premenopausal women and with positive receptor status (101,167,195)). Yet surgical ablative procedures of the gonads in women are associated with significant risks of psychological trauma, postoperative morbidity and mortality and are irreversible, even for patients for whom the clinical benefits of such procedures may not be apparent.

Reversible medical castration in women can be obtained by the use of GnRH agonists, circumventing the inconvenience of surgical ablative procedures. As discussed earlier, the chronic administration of these compounds caused reversible suppression of ovarian function, with estrogens diminishing to castration levels. Interestingly, GnRH agonists decreased the size of the uterus and mammary glands and caused remissions in dimethylbenz(a)anthracene (DMBA)-induced mammary tumors in rodents (126,253,285). Klijn and DeJong (127) firstly reported a favorable effect of buserelin in premenopausal women with metastatic breast cancer. The GnRH agonists goserelin (Zoladex), leuprolide (Lupron), buserelin (Superfact) and decapeptyl, following evaluation in clinical trials for advanced disease, have all yielded similar therapeutic responses of about 30–40% (complete or partial response rates) [60,67,127,253]. Surprisingly, these agonists lack significant side effects or toxicity; the only consistently reported adverse events are amenorrhea, hot flushes, joint pain, and loss of libido. In contrast to prostate cancer, the phenomenon of "disease flare up" is less common and is probably due to individual variability in estradiol levels at the start of therapy.

Perhaps one of the most paradoxical observations in the endocrine treatment of breast cancer with GnRH agonist is the sporadic report of responses in postmenopausal women (27,100,246). This indicated the possibility of direct antitumor effects of GnRH agonist in patient without functional ovaries. Direct antiproliferative effects of GnRH agonist have been demonstrated in human breast cancer cell line (MCF-7) [168, 250]. Several studies have also reported the presence of specific GnRH-binding sites and GnRH agonist-induced stimulation of intracellular signal transduction mechanisms (13,71,275). It is also of interest to note that direct action of GnRH agonists on ovarian cancer cells has been reported (72), and GnRH specific binding sites have been demonstrated in ovarian cancer biopsies (192). Despite these observations, reliable data on the possible direct effects of GnRH agonists are sparse and contradictory, and the relative contribution of these extrapituitary effects need to be determined. It should be mentioned that GnRH agonists are presently not approved in the United States for breast cancer treatment.

In studies using complete androgen blockade (CAB), the association of a GnRH agonist and pure antiandrogen has proven beneficial in the treatment of advanced prostate cancer. Likewise, clinical trials using complete estrogen blockade (CEB) with a GnRH agonist and tamoxifen have been conducted, but the impact of such therapy has not been as initially anticipated (187,240,300). The rationale behind this approach is the fact that, although GnRH agonist alone causes reduction of serum estrogens to castration levels, steroid estrogen precursors of extragonadal origin may still be present and can promote growth of hormone-dependent cancer cells, including cancer tissue. The possible contribution of the adrenals that secrete high levels of DHEA and DHEA-S, which can be converted to estrogens in peripheral tissues, should not be overlooked in the treatment of breast cancer with CEB (131,167,195). This is even more true since normal and neoplastic breast tissues possess the enzyme required to produce estrogen locally, i.e., aromatase, sulfatase, and 17b-hydroxysteroid dehydrogenase (1,54,131,195,197) and that the concentrations of estrogens in breast cancer tissue, particularly at menopause, are very high (196). (Table 3)

Thus, serum estrogen levels after GnRH therapy may not reflect adequately the total estrogenic influence on the growth of hormone-dependent tumors. However, approaches using the combination of GnRH agonist and the antiestrogen tamoxifen (CEB) in pre- and postmenopausal women suffering from advanced breast cancer have not shown any statistically significant differences in objective response, time to progression, or overall survival in favor of the combination therapy (112,240). Table 3 This could be explained, in part, by incomplete estrogen blockade and the fact that the antiestrogen used, tamoxifen, possesses mixed estrogenic and antiestrogenic activities that are highly species-, tissue-, cell-, and even gene-specific (114,293). In addition, its active metabolite (4-OH-tamoxifen) stimulates the growth of human breast cancer cells in vitro and in vivo (96,241). Such paradoxical effects may explain the failure of present day complete estrogen blockade to improve tumor-response rate, compared with surgical or medical (GnRH agonist) castration alone. The development of new agonists such as nonsteroidal pure antiestrogens (toremifen, EM-800) and new aromatase inhibitors (fadrozole, anastrozole) may well contribute to improving response rates of complete estrogen blockade in advanced breast cancers.

Endometriosis is a common gynecological disease responsible for an important proportion of infertile women (125). In terms of incidence, endometriosis is the second most frequent gynecological disorders after leiomyoma (113). Endometriosis is defined as "the presence of ectopic tissue which possesses the histologic structure and function of the uterine mucosa." Endometrial tissue may be found in many possible locations, from the pelvic cavity to distant sites such as the pericardial and pleural cavity. It can cause considerable morbidity in the form of dysmenorrhea, lower abdominal pain, and menorrhagia. Its pathogenesis remains theoretical and its management a battle between endocrine therapy and surgery.

Since endometrial tissue requires estrogen for growth and proliferation, a state of hypoestrogenism results in atrophy and regression of ectopic endometrial tissue, as observed following natural or surgically induced menopause in humans and experimental animals (2). Pseudopregnancy and progestin therapy have also been reported to improve endometriosis, relieving pelvic pain and dysmenorrhea in more than 80% of patients, but the effects are only transient (124,172). Similar observations were reported with androgens (methyl-testosterone), which are partially effective in relieving dysmenorrhea and abdominal pain but have important masculinizing effects (125).

Danazol, an isoxazole derivative of 17-ethinyltestosterone, became widely used in the early 1980s as a treatment for endometriosis (11,113,124). The mechanism of action of danazol is the induction of a pseudomenopausal state, through suppression of serum estradiol levels, either by inhibiting the hypothalamic-pituitary-ovarian axis or by inhibiting ovarian steroidogenesis, or both. As reported during the past 10–12 years, dosages of danazol up to 800 mg daily for average of six months can cause a symptomatic and clinical improvement in more than 80% of women. The objective response measured at laparoscopy has been found to be decreased in 70–90% of cases (11,24,53). However, one of the most important aspects of treatment of endometriosis with danazol is recurrence after cessation of therapy. A 50% rate of recurrence within less than a year was reported by Moore et al. (175) when symptoms served as diagnostic parameters. Pregnancy rates in women with no discernible cause of infertility other than endometriosis ranged from 28–74% (24),99). The success of therapy with danazol can thus be compared with conservative surgery for treatment of infertility due to mild or moderate disease. On the other hand, danazol possesses androgenic, anabolic, progestational, and antiestrogenic properties (11,24,113). Frequently, patients complain of significant side effects associated with its pharmacological properties (e.g., acne, hirsutism, edema, and weight gain).

Following these observations of the hormone dependency of endometriosis, the possibility of reversible suppression of ovarian estrogen secretion by GnRH agonists became evident (Fig. 19). Medical oophorectomy using long acting D-Trp6-Pro9-Net-LHRH as a new approach to the treatment of endometriosis was first reported by Meldrum et al. (163). Four out of five women studied had estradiol levels within castration range after 28 days of treatment. Shaw et al. (262) reported the first official attempt to treat endometriosis with the GnRH agonist buserelin, and after six months of therapy, endometriotic lesions had resolved, ovulation was suppressed, and estrogen levels decreased to early follicular phase levels. Lemay et al. (142) published the results of 10 women for up to six months of intranasal buserelin therapy. Patients reported near complete relief of painful symptoms, and good or dramatic reduction in endometriotic tissues was observed at laparoscopy. In these studies, GnRH therapy was well tolerated, and only light to moderate uterine bleeding occurred within 2–4 weeks at the start of therapy (estrogen withdrawal). No endometrial hyperplasia or cellular atypia was observed. Following cessation of therapy, ovulatory menses resumed in 1–2 months and a small number of pregnancies were reported.

Other trials with diverse GnRH agonist regimens confirmed the marked suppression of GnRH agonist in endometrial lesions (Fig. 21) and symptoms at all stages of the disease and proposed GnRH agonist therapy as an advantageous alternative to danazol treatment (62,63,274). In fact, the extent of ovarian suppression seems greater with GnRH agonist than with danazol treatment, and GnRH agonist treatment was found to be effective in danazol-resistant endometriosis (85). Moreover, in randomized trials, the potent agonist Buserelin was superior in reducing active endometrial implants and improving contraception rates, compared with danazol.

In spite of the fact that ovarian suppression is the common denominator in the mechanism underlying the therapeutic effect of danazol and GnRH agonists, significant differences in their adverse effects are observed and are directly related to their pharmacological profiles. Negative effects of danazol on hepatic functions, lipid profiles, and metabolic effect associated with androgenic impregnation must be weighed against the more "estrogen-deficit" effect of GnRH agonists on bone loss and psychovegetative effects. Presently, long-term GnRH agonist therapy for endometriosis is not recommended to exceed six months, since questions remain about the repercussions of bone demineralization and the reversibility of the hypogonadic state for longer periods.

As discussed later, the recent introduction of "add-back" therapy, defined as the combined administration of various steroidal and nonsteroidal agents with GnRH agonists, may limit the negative impact of hypoestrogenic state associated with GnRH agonists analogue therapy.

Uterine leiomyoma is the most common benign tumor occurring in women during their reproductive years (113). It is estimated that 20–50% of cases are associated with symptoms of pelvic pain, pelvic pressure, infertility, menorrhagia, and spontaneous abortion (113). Although the pathogenesis underlying uterine leiomyoma is not known, it has been clear for many years that estrogens play a predominant role, since pregnancy stimulates tumor growth, menopause causes degeneration of leiomyoma and leiomyoma are exceptional before puberty. Interestingly enough, uterine leiomyoma tissue contain both estrogen and progesterone receptors, and the endometrium underlying leiomyoma has increased estrone sulfatase activity and estrone concentration levels (18, 271).

While surgery remains the treatment of choice in uterine leiomyoma (myomectomy if fertility must be maintained or hysterectomy for larger and multiple tumors), the availability of potent GnRH agonists opened the way for the first effective medical treatment for this disorder. Filicori et al. (83) were the first to report the beneficial effect of GnRH agonist-induced ovarian suppression in uterine leiomyoma. Maheux et al. (154) published a pilot study on ten women harboring leiomyomas and showed that buserelin (daily subcutaneous injections for six months) caused uterine tumor size suppression by 25–80% and decreased serum estradiol to menopausal levels. Similar results were subsequently reported with daily s.c. injection of histerelin, intermittent s.c. infusion of buserelin, and goserelin depot implants (40,103,297). Most clinical studies using GnRH agonist have reported significant reductions of leiomyoma and uterine size ranging from about 50–75% after 3–6 months of administration (40),66,103,154,297).

In a randomized, placebo-controlled, double-blind study evaluating the efficacy of leuprolide acetate depot in the treatment of uterine leiomyoma, Freidman et al. (87) reported a mean reduction of 40% in uterine volume, maximal after 12 weeks, with no change in uterine size with placebo (Fig. 22). However, uterine volume increased to 88% of pretreatment size within three months of cessation of GnRH agonist therapy. The reduction of uterine volume and tumor size is frequently reversible after discontinuation of agonist treatment.

Since, in general, GnRH agonist treatment cannot be continued for long periods and is followed by regrowth of tumors, agonist therapy as an adjuvant to surgery has been proposed. Several studies reported that myomectomy could be facilitated after GnRH agonist therapy, others observed easier tumor excision and decreased intraoperative blood loss during laparoscopic and hysteroscopic myomectomy, confirming the preoperative role of GnRH agonist in leiomyoma (65,66,276). Numerous mechanisms for GnRH agonist action on uterine leiomyoma have been proposed, from reduction of uterine blood flow (261) to decreased estradiol and epidermal growth factor receptor levels and hyaline degeneration (8). These effects are probably a direct consequence of the hypoestrogenic status and are reversible after cessation of agonist therapy. Finally, intermittent GnRH agonist therapy could be considered as sole suppressive treatment in perimenopausal women with leiomyoma until spontaneous ovarian failures occur or in patients who are poor surgical candidates or refuse surgery. This approach has been successfully used in the treatment of benign prostatic hypertrophy and prostatic adenocarcinoma.

As described earlier, GnRH agonists have been used with success as therapeutic agents for several sex steroid-dependent gynecologic disorders. However, GnRH agonist induced hyperestrogenic side effects, such as increased bone loss and decreased cardiovascular protection, have curtailed the prolongation of long-term treatment regimens. In fact, with the exception of precocious puberty, GnRH agonist therapy for gynecological disease is limited to six months. In order to overcome these potentially life-threatening serious adverse effects and also the secondary psychovegetative effects, numerous attempts to palliate estrogenic deficiency by the combination of steroidal and nonsteroidal agents (add-back) to GnRH agonist treatment have been made (143,277). The challenge in "add-back therapy" is double: not only must the regimen suppress undesired aspects of hypoestrogenic states, but it should not interfere with the efficacy of GnRH agonist and should possibly improve efficacy.

Steroidal and nonsteroidal compounds have been studied as potential add-back therapeutic regimens in gynecologic disorders (Table 4). Progestins, basically medroxyprogesterone acetate (MPA), and norethindrone (NET) have been evaluated extensively as add-back therapy in endometriosis (Fig. 23 and Fig. 24). Following the recent pioneering work of Lemay et al. (142), Cedars et al. (30) and Surrey (278,279), it became apparent that MPA and NET are capable of suppressing GnRH-induced vasomotor effects and bone density losses. However, MPA appears to reverse the beneficial effects of GnRH agonist therapy on endometrial lesions and the painful symptoms associated. Therefore, at this time, it is not recommended to use this progestin (MPA) as an "add-back" supplement in endometriosis. In the case of NET, only high and prolonged treatment schedules significantly block bone density changes. Unfortunately, an undesirable increase in low-density lipoprotein (LDL)/high-density lipoprotein (LDL:HDL) ratio was seen at high NET dosages. These undesirable effects have been circumvented by the association of low-dose NET plus an organic biphosphonate (sodium etidronate). This combination seems promising for long term GnRH agonist treatment for endometriosis in the future.(147,280).

Data regarding the use of "add-back" therapy in the treatment of leiomyoma is more limited. The basic approach behind therapy of leiomyoma is to obtain maximal ovarian suppression and uterine volume reduction before institution of the "add-back" regimen. Essentially, only nonconcurrent estrogen/progestin steroid combinations have been used. Friedman (86) and Maheux et al. (155), using long-term conjugated equine estrogen and sequential MPA as "add-back" therapy in leiomyoma, reported that the estrogen/progestin combination did not counteract the uterine inhibiting effect of GnRH agonist therapy and appeared to protect patients against loss of bone mass. Moreover, no significant alterations in lipoprotein profiles were observed.

Clearly, additional large randomized clinical trials are needed to explore the full potential of "add-back" therapy in endometriosis and leiomyoma, and further experimentations must be carried out to find alternative agents.

We have a considerable challenge to study and understand the regulation of infundibular LHRH neurons, not only because of the extreme complexity in term of neuroanatomical organization of this neuroendocrine system, but also because the new techniques of molecular biology provide evidence that sequence and genetic regulation of the decapeptide is species-dependent. When we know that only one base pair can sometimes allow DNA binding or not, it is easy to understand that mechanisms involved in the genetic transcription, translation and peptide production within different subsets of neurons may not be conserved by different species. It is also important to point out that the circuitry(ies) solicited during emergency situations is(are) dependent on the intensity as well as the type of stress used during experimentation (emotional, physical, immune, etc.), making generalizations about the mechanisms by which stress alters reproduction difficult. Better understanding of these fundamental mechanisms will certainly provide keys to the puzzle of several states of human reproductive pathologies. As presented in the second part of this chapter, LHRH (GnRH) agonists are now used on a routine basis for treating many pathologies in both men and women, although much remains to be done to clearly identify the appropriate dosage and combination therapies to improve the efficiency of such treatments.

Research supported by grants from the Medical Research Council (MRC) and the Natural Sciences and Engineering Research Council (NSERC) of Canada. Serge Rivest is a Scholar from the Medical Research Council of Canada. We thank Dr. Rossella E. Nappi of University of Pavia (Italy) who was involved in several of the results presented in this chapter when she was a postdoctoral fellow in Dr. Rivest's laboratory. I (S.R.) would also like to thank Dr. Catherine Rivier of the Salk Institute (La Jolla, CA) for her scientific spirit and for her courage in supervising me during the 3 years of my postdoctoral formation. Several of the results presented in this chapter were accomplished in her laboratory. The secretary assistance of Josée Poulin has also been greatly appreciated.

published 2000