Studies were performed to obtain evidence for glyconeogenesis from pyruvate to the triose phosphates in pancreatic islets. Inability to show this evidence would be consistent with the fact that glyceraldehyde, but not pyruvate, is a potent insulin secretagogue. Synthesis of 14C-labelled glucose from 14C-labelled pyruvate could not be detected. Since this might have been due to lack of sensitivity required to measure 14C-glucose production in such a scarce tissue as islets, cDNA probes were used to estimate the relative expression of genes coding for gluconeogenic enzymes. Islets expressed pyruvate carboxylase mRNA, but even islets from rats which had been starved (a condition which induces phosphoenolpyruvate carboxykinase (PEPCK) in liver, kidney and adipose tissue) showed no PEPCK mRNA. This is consistent with our previous work showing the absence of PEPCK enzyme activity in islets. Therefore, islets can convert pyruvate to oxalacetate, but since they lack PEPCK, neither the beta nor alpha cell can convert oxalacetate to phosphoenolpyruvate and carry out glyconeogenesis. Pyruvate carboxylase mRNA was increased in islets that possessed the capacity for glucose-induced insulin release versus islets that lacked the capacity to respond to glucose, such as islets from fed rats (versus starved rats) and in islets cultured at a high concentration of glucose (versus at low glucose). Pyruvate carboxylase, therefore, must be involved in pyruvate metabolism and not glyconeogenesis in the pancreatic islet.

Previously, we demonstrated that systemic injection of the growth hormone secretagogue, growth hormone-releasing peptide (GHRP)-6, selectively activated cells in the hypothalamic arcuate nucleus, as reflected by increased electrical activity and induction of the immediate early gene c-fos. The growth hormone secretagogue receptor distribution is not confined to the arcuate nucleus, suggesting that additional sites of action may exist. In the present study we characterized the electrophysiological responses of cells in the arcuate nucleus, ventromedial nucleus and periventricular nucleus in an in-vitro hypothalamic slice preparation, following bath application of GHRP-6. Additionally, since central somatostatin administration has been shown to attenuate the induction of the c-fos gene by GHRP-6, we sought to determine whether the arcuate cells activated by GHRP-6 are also somatostatin-sensitive. Male Wistar rats (100-150 g body weight (BW)) were anaesthetized (urethane; 1.2 g/kg BW) and the brains removed. Coronal sections (400 microm thickness) were cut through a block of hypothalamus and were transferred to a slice chamber perfused with artificial cerebrospinal fluid. Forty-one arcuate nucleus cells were tested with bath application of 15 microm GHRP-6 for 10 min, 16 of which were tested subsequently (>30 min later) with application of 10 microM somatostatin. Following GHRP-6 administration, 19 cells (46. 3%) showed a significant increase in firing rate during the 15-min period after GHRP-6 application (P<0.001), 17 cells (41.5%) did not respond and the remaining five cells (12.2%) were significantly inhibited. Six of the eight arcuate nucleus cells that were excited by GHRP-6 were significantly inhibited by somatostatin. By contrast, five of the six arcuate nucleus cells that were unresponsive to GHRP-6 were also unresponsive to somatostatin. In the ventromedial nucleus, of 19 cells tested, eight cells (42.1%) were excited by GHRP-6, eight cells (42.1%) were unresponsive and the remaining three cells (15.8%) were significantly inhibited. Of 19 cells recorded in the periventricular nucleus, 13 (68.4%) were unresponsive to GHRP-6 and six (31.6%) were significantly inhibited. Thus, electrophysiological studies in vitro suggest that: (1) neurones in the hypothalamic arcuate nucleus, ventromedial nucleus and periventricular nucleus show changes in electrical activity in response to GHRP-6; and (2) the arcuate nucleus cells excited by GHRP-6 are also subject to inhibitory control by somatostatin.

1. Using the ventral surgical approach in vivo, extracellular recordings were made from seventy‐nine cells in the supraoptic nucleus of urethane‐anaesthetized male, virgin female or lactating female rats while stimulating the pituitary stalk. Cells were classed according to their spontaneous firing activity as: continuous (putative oxytocin), phasic (putative vasopressin) and silent. 2. Stimulation of the neural stalk produced an excitation (up to 25 ms poststimulus) in eleven of the seventy‐nine antidromically identified magnocellular neurones, consistent with the existence of excitatory collaterals or dendritic contacts between such cells. In these recordings a second spike could frequently be seen, following the antidromic spike, with a variable latency. Such spikes consistently collided with subsequent antidromically evoked spikes. Poststimulus excitation was only seen in silent and continuously firing (putative oxytocin) cells, suggesting that oxytocin and vasopressin cells have different connections. 3. Excitatory connections were seen more frequently in lactating females (8 out of 22 cells) than in males (1 out of 15 cells) or virgin females (2 out of 10 cells), and thus may make an important contribution to the bursts of firing which precede reflex milk ejection.

Extracellular recordings were made from supraoptic nucleus (SON) cells in urethane anaesthetized male rats in vivo. Two stimulating electrodes were positioned to activate the cells antidromically, one in the mid axon region of the cells and the other at the axon terminals. Trains of 5-20 just-subthreshold stimuli at 5 s intervals decreased the threshold for antidromic activation from both sites. Whereas neither single stimuli, nor the stimuli at the beginning of a train of 20 stimuli evoked antidromic action potentials, later action potentials did so. Paradoxically, trains of 20 just-suprathreshold stimuli increased the threshold for activation of both axons and terminals. In recordings from the same cells, stimuli were applied singly at 5 s intervals at an intensity which almost invariably evoked an antidromic action potential. Identical stimuli were then applied in trains of 20 stimuli at 50 Hz. After the first train, the initial stimulus pulses of the trains frequently fell below threshold. Following a conditioning train of five stimuli applied to one electrode, the period of decreased threshold (increased excitability) at the other electrode lasted less than 100 ms and the period of increased threshold (decreased excitability) after 12 trains of 20 stimuli lasted between 5 and 10 s. Both decreased and increased excitability were seen at axons and terminals of both putative oxytocin and vasopressin cells. Since the excitability changes were shown in vivo at frequencies encountered during recordings, it is likely that they influence the probability of spike propagation and hormone secretion under physiological conditions.