I Introduction
The response of any animal (or human) to an adaptively important stimulus from the environment is formulated through its behavior, its endocrine activity, and its autonomic nervous system. To say this is to say nothing new or surprising, for this division of effectors has been apparent for many years. However, largely as the result of the traditional boundaries around different disciplines, these sets of responses have often been studied separately, by different people deriving from equally different backgrounds and thus using significantly different approaches. On some occasions this division of effort has resulted in individuals in one subject being unaware of what the others were doing and thus in danger of interpretations which do not take into account sufficiently the complexity of an animalâs total response or the possibility of interactions between the constituent components. There are signs, however, that this situation is being remedied, partly because of a more flexible approach, but also as the result of recent findings in neuroendocrinology. In particular, the description of systems within the brain, endocrine glands, and somatic tissues which seem to contain the same (or similar) peptides has forced us to consider whether a more integrated view of their function might be appropriate. A number of attempts to synthesize currently available knowledge have appeared, notably from those concerned with drinking behavior and fluid balance (Mogenson et al., 1980; Swanson and Mogenson, 1981), for this area offers some of the best examples of coordinated behavioral and endocrine activity to date. Because many peptides have been found in high concentration in the limbic system, it is natural to wonder whether they may play a part in this systemâs function, which is central to the organization of the response triad described previously. Several lines of evidence, to be discussed more fully later in this chapter, indeed suggest that some neuropeptides can regulate the categories of behavior traditionally associated with the limbic system, for example, eating, drinking, sexual activity, and maternal behavior. This seems to apply particularly to peptides which are called here âneuroendocrine peptides,â not as a watertight definition separating them from other sorts of peptides, but to indicate that many of those active in these sorts of behavior are also concerned with more usual neuroendocrine function, such as the regulation of the pituitary (Everitt et al., 1983). If this is accepted, then it follows that since many of the same functions are also acted upon by the steroid hormones, the interaction between these two classes of compound becomes interesting.
Behavior, hormones, and autonomic activity not only form part of a total response, they also interact with each other. An example might be the hormonal changes which follow some forms of sexual interaction. Copulation by a female rat induces the corpus luteum to secrete progesterone, which in turn inhibits further sexual responsiveness (Marrone et al., 1979). Secretion of noradrenaline from the adrenal during âstressâ may act upon the pituitary to release adrenocorticotrophic hormone (ACTH), thus activating a second set of endocrine responses (Mezey et al., 1984). The existence of such interactions means that we have to know about the way that events taking place in the vascular compartment (for example, the secretion of a hormone) influence those in the brain (such as acting on a behavioral response, or a system controlling an autonomic or endocrine activity). This leads us to consider the ways that levels in one compartment are reflected in the other, and whether, in cases where the same substance derives from both, there is evidence of functional linkage between the two systems. For steroids the situation is simpler than for peptides. Steroid hormones are only produced, so far as we know, from peripheral endocrine glands. In their case the question is how alterations in their level in the blood, in either the long or the short term, are reflected in the brainâs extracellular environment, since it is from the latter that neurons receive their hormonal information. For peptides, the story is more complicated. In some cases at least, both central neurons and peripheral glands produce the same peptide, so the intra- and extracerebral compartments have, potentially, different sources of supply. The question then is: do the two sources coordinate, and is there any direct communication between them?
Much of the information to be discussed here is based upon measuring levels of hormones in the cerebrospinal fluid (CSF), though what interests us, in fact, is the hormone levels in the brainâs extracellular fluid (ECF). The reasons for such measurements on CSF are entirely pragmatic; it is the nearest thing to ECF that can be obtained in an endocrine context. It must be reasonably certain, therefore, that hormone levels in the one give a satisfactory index of those in the other. Many studies, both anatomical and physiological, show that there is no effective barrier between the two except, perhaps, around the tanycyte ependyma (Bradbury, 1979). Though this means that hormones in the ECF will pass readily into the CSF (and vice versa), there may be regional differences in concentration in the CSF related, for example, to those in the neighboring ECF, which will be missed in samples drawn from, say, the cisterna magna. Furthermore, clearance from the CSF may result in sink effects, so that concentration gradients exist within the ventricles which are not apparent from sampling at one point. Nevertheless, while these reservations need to be remembered, it remains true that CSF levels can be assumed to give at least a first approximate guide to those in the ECF, though the whole question awaits development of methods of assaying hormones in the ECF itself. Finally, it is worth noting that many limbic structures, on which this chapter is focused, lie very close to the ventricles, a further reassurance that CSF levels of steroids and peptides may give a realistic indication of those in the fluid surrounding the neurons themselves.
II Steroid Hormones
The advent of highly sensitive assays has permitted the operating characteristics of the steroid hormones to be specified in some detail. For our purposes, it is important to note that their secretion (and therefore blood levels) is not constant, but shows the features of at least three sets of rhythms (Hastings and Herbert, 1986). The pulsatile release from their glands of origin depends, of course, on antecedent pulses from the pituitary and hypothalamus; the significant point is that blood levels are changing relatively rapidly all the time. Is this change detected by the brain and, if so, in what form? Many steroids also show circadian alterations in blood levels; this seems to be related to circadian changes in the frequency or amplitude of the circhoral pulses. The important point here is that the mean level of the steroid is changing because of underlying alterations in the pulse characters; how does the brain perceive these circadian changes? Annual rhythms in the reproductive steroids typify those species showing breeding seasons; are there more exact correlations between levels in blood and brain in the case of these very slow changes than in the more rapid ones during the day? It must be emphasized that a steroid-sensitive neuron will respond to a steroid according to the concentration of that hormone in the fluid surrounding it; many in vitro studies have shown this without question (McEwen, 1981). The fluid surrounding such a neuron is the extracellular fluid of the brain, not the blood. We need to know, therefore, how the variations of steroid levels in the blood already described are reflected in the brainâs ECF (or, since it is in free communication with it, the CSF).
There is increasing evidence that the temporal characteristics of changes in steroid levels are also important. For example, the duration of the estrogen surge, as well as its level, is critical for provoking the secretion of luteinizing hormone (LH) (Karsch et al., 1973). Equally significant may be the duration of a hormoneâs absence. Withdrawal of progesterone for a defined period is required to enable a female rat to respond to estradiol by showing lordosis (Lisk,...