Classically, the secretion of GH is controlled by the GHRH neurons of the arcuate nucleus and periventricular somatostatin neurons, which stimulate and inhibit GH secretion, respectively. Immunoneutralization of somatostatin enhances GHRH-evoked GH release (40), but does not enhance GH release evoked by GH secretagogues (2), suggesting that GH secretagogues may have a dual action to stimulate endogenous GHRH release and suppress endogenous somatostatin release. In male rats, a GHRP-6 infusion has been shown to disrupt the cyclic responsiveness in GH release following regular injections of GHRH (41); since this cyclic responsiveness has been attributed to cyclic release of somatostatin, it seems likely that the GH secretagogues disrupt the cyclic release of somatostatin.
In the rat, the secretion of GH is sexually dimorphic: in males the pulses of GH are larger, less frequent, and arise from lower interpulse baseline compared with females (42,43). Androgens play an important role for maintaining the low baseline GH levels and for controlling GH pulse height (44). The sexually dimorphic patterns appear to derive from dimorphic behavior of the somatostatin neurons, possibly reflecting the dimorphic expression of androgen receptors by these neurons (45). It appears probable that, in the male rat, GHRH and somatostatin are released alternately to produce alternate peaks and troughs of GH release, whereas in the female somatostatin is released more continuously. Interestingly, in the male rat, prolonged infusion of somatostatin leads to a sustained inhibition of GH release, followed by a dramatic rebound secretion of GH after the end of somatostatin infusion (46). Although this rebound is partly generated at the level of the pituitary, it also appears to reflect a large rebound secretion of GHRH. Similar rebound secretion of GHRH follows electrical stimulation of the periventricular nucleus (47). The periventricular nucleus appears to provide a direct inhibitory projection to neurons in the arcuate nucleus (17), and 55-60% of the GHRH neurons visualized either by immunocytochemistry (48) or in situ hybridization (49) express somatostatin receptors. Hence, it is possible that the reciprocity in the hypothalamic output of GHRH and somatostatin during spontaneous pulsatile GH secretion reflects neuronal interactions between the GHRH and somatostatin cells. The authors' electrophysiological studies in vivo support this hypothesis, since neurosecretory neurons in the arcuate nucleus that were excited by GHRP-6 or nonpeptide secretagogues were also inhibited during electrical stimulation of the periventricular nucleus (6), and such secretagogue-responsive neurosecretory cells are also inhibited following iv injection of somatostatin or Sandostatin (a long-acting somatostatin analog). By contrast, cells that are not responsive to secretagogues are mainly unaffected by somatostatin injections. Furthermore, icv injection of Sandostatin suppresses the GH response following iv injection of the GH secretagogues (50). In addition, the central actions of GH secretagogues to induce expression of Fos in the arcuate nucleus can be attenuated by systemic or central administration of Sandostatin (51). Thus, it would appear that a subpopulation of the arcuate cells activated by GH secretagogues are also inhibited by central soma-tostatin action.
Suppression by Sandostatin of the GH secretagogue-induced increase in the expression of Fos in the arcuate nucleus is likely to be mediated by a direct central action of this peptide since injection of a very low dose of Sandostatin (2 ^g) was as effective as iv injection of 100 ^g (51). Indeed, this also suggests that Sandostatin is able to gain access to central sites when administered by the iv route.
When considering the inhibitory effects of systemic administration of Sandostatin on GH secretagogue-induced GH release, it is difficult to distinguish between the inhibitory actions at the level of the pituitary and central actions. At the pituitary, somatostatin suppresses both spontaneous GH release (52) and release induced either by GHRH (53) or by GH secretagogues (1,2). However, the suppression of GH release is likely to reflect, at least in part, a central action, since Sandostatin inhibits GH secretagogue-induced GH release when administered intracerebroventricularly (50). Although it seems likely that Sandostatin acts via somatostatin receptors on GHRH neurons, it is also possible that it acts via an afferent pathway to these cells. Indeed, it is not possible to determine whether the cells that are the direct target for the action of the secretagogues are also the direct target for the action of Sandostatin.
Nonetheless, the interaction between the central effects of GH secretagogues and somatostatin suggests that, although many of the arcuate neurons activated by the secretagogues are not GHRH cells, they are nonetheless likely to be intimately involved in the regulation of GH secretion. One possibility is that these are interneurons linking the population of GHRH neurons to provide the necessary co-ordination needed for generating pulsatile discharge. Another possibility is that they are activated by inputs from GHRH cells and provide the missing link in a reciprocal influence of GHRH neurons on the periventricular somatostatin neurons. As yet nothing is known about the projections of these neurons other than their deduced projection to the median eminence, and establishing their cellular connectivity is likely to be a prerequisite to understanding their function.
The GH secretagogues have thus fortuitously provided a potentially important tool to dissect the neuronal circuitry underlying the GH pulse generator. It has yet to be established that their actions are more than serendipitous pharmacology, but the selective expression of receptors on pituitary somatotrophs and in the hypothalamus, and the specific actions within the hypothalamus on GHRH cells and other cells that are sensitive to somatostatin, strongly suggest the existence of an endogenous ligand. This pattern of receptor distribution suggests that this ligand is likely to be either present in neuroendocrine neurons projecting to the median eminence, or else is produced peripherally, but has access to sites within the blood-brain barrier. Either of these alternatives will open a fresh chapter in our understanding of the physiological regulation of GH secretion.
Was this article helpful?