The anterior pituitary gland secretes pulses of luteinizing hormone (LH) in response to pulses of gonadotrophin-releasing hormone (GnRH) released into the hypophysial portal blood by the hypothalamus. The pulsatile nature of the secretions is very important because the frequency of the pulses is directly related to the activity of the GnRH neurons. We can therefore take advantage of this phenomenon to develop mechanistic interpretations of responses to experimental treatments designed to unravel the neural pathways that influence what is, arguable, the most important individual signal controlling the activity of the reproductive system. We might also resolve the disagreements in the literature covering the neuropharmacology of gonadotrophin secretion. In this review, we describe work towards this end in the sheep. Most (95%) of the 2500 GnRH cell bodies in the sheep brain are located in a region covering the anterior hypothalamus, the medial preoptic area, the diagonal band of Broca, and the septum. The axons of up to 50% of these cells terminate in the organum vasculosum of the lamina terminalis. The remainder terminate in the median eminence and form the final common pathway for the many factors that affect gonadotrophin secretion. Among the factors known to affect the frequency of the pulses (or the activity of the GnRH neurons) are nutrition, pheromones, photoperiod and gonadal steroids (negative and positive feedback). Factors that affect GnRH pulse amplitude are more difficult to determine because variations in pituitary responsiveness prevent the use of LH patterns as a 'bioassay'. Techniques developed recently have allowed the direct measurement of GnRH pulse amplitude and revealed inhibitory effects of oestradiol, but we do not know whether this effect is due to a reduction in the amount of GnRH released by each neurone or a reduction in the number of neurones releasing a pulse. It is unlikely that the factors that alter pulse frequency do so by directly affecting the GnRH cells. For example, it is obvious that other cells, with specific receptors for pheromonal or nutritional stimuli, formulate a signal that is transferred to the GnRH cells via interneurones. Similarly, it is likely that a hypothalamic clock intervenes between photoperiodic inputs and GnRH output. Opioidergic neurons have been proposed as a link in this system, but the complexity of their action makes it unlikely that they directly affect the GnRH neurons. The responses to steroids are simple and rapid, but steroid receptors have not been found in GnRH cells, so at least one other set of interneurones is involved. Opioidergic neurones have also been implicated here, as have dopaminergic, noradrenergic and serotinergic systems. Most studies using drugs that affect neurotransmission have produced conflicting and confusing results because of unrealistic doses, inappropriate experimental models, and the lack of pulse analysis. Often, little has been gained except to implicate a particular neurotransmitter in the control of GnRH release. The number of interneurones between the GnRH cells and the cells affected by the treatment may be so large that the pathway implicated plays only a minor modulatory role in reproduction. On the other hand, good experimental models are being developed, the most promising being negative feedback by sex steroids on pulse frequency. This approach, when combined with electrophysiological techniques and direct hypothalamic injection of test substances (rather than peripheral injection), provides a powerful research tool for increasing our understanding of the pulse-generating system. For the ewe, this combination of techniques has been used to reveal a catecholaminergic (probably dopaminergic) pathway that appears to originate in the A15 nucleus in the hypothalamus and could innervate anterior structures (septo-preoptic area) and the median eminence. Tonic activity in this system appears to inhibit GnRH secretion and has to decrease before a pulse can be released. It could affect the activity of cells in both the median eminence and the septo-preoptic area, and it may mediate the inhibitory effects of oestrogen (i.e. negative feedback). The top priority for future research in this area should be to find the origin (location and cell type) of the pulsatile signal. Apparent connections between the GnRH cells in the preoptic area suggest that the pulses could originate within these cells -- a possibility that needs to be tested thoroughly. As a second area for investigation, there is the intriguing question of the roles of the large number of GnRH cells that terminate in the organum vasculosum of the lamina terminalis and other extrahypothalamic sites.
