Recent studies suggest (Schwartz et al., 1979) that local cerebral glucose utilization, measured with the deoxyglucose technique (Sokoloff et al., 1977), correlates most closely with electrical activity in the neuropil in general and synaptic terminals in particular. Presumably, increased glucose utilization associated with increased impulse activity in nervous tissue is, as is oxygen consumption (Ritchie, 1967;Greengard and Ritchie, 1971: De-Weer, 1973, principally due to enhanced activity of the sodium pump. If the increased energy metabolism during impulse activity is used mainly for reconstitution of electrochemical gradients, then it is to be expected that cellular components with larger surface-to-volume ratios will have larger energy demands (Ritchie, 1967;Greengard and Ritchie, 1971;DeWeer, 1975) and, thus, greater rates of glucose utilization. It would be of value for the interpretation of studies that employ the autoradiographic deoxyglucose method to identify the cellular elements in which neural activity and energy metabolism are most closely linked. We have, therefore, studied an in vitro preparation of rat posterior pituitary, which represents a relatively enriched population of axon terminals (Nordmann, 1977) and may serve as a model for synaptic endings in the brain. Because the pituitary is a neurosecretory organ, we have also studied the influence of the secretory process in this system on energy metabolism. As an index of glucose utilization, we have measured the rate at which [ ''C]deoxyglucose is phosphorylated by hexokinase and trapped in the tissue incubated in vitro. This is the in vifro equivalent of the deoxyglucose method in which trapped [ '4C]deoxyglucose-6-phosphate is visualized and measured autoradiographically (Sokoloff et al., 1977). MATERIALS AND METHODSMale Sprague-Dawley rats (180-250 g) were decapitated, and the pituitary glands were removed rapidly and placed in balanced salt solution (BSS) consisting of 10 mM

Physiological stimulation of the hypothalamo-neurohypophysial system by salt loading of rats resulted in a dramatically increased glucose utilization in the posterior pituitary but not in the paraventricular or supraoptic nuclei. The good correlation between glucose utilization and neural activity in the posterior pituitary (that is, nerve terminals) contrasted with the lack of correlation in the paraventricular and supraoptic nuclei (that is, the sites of the cell bodies of the same neurons). This difference in the metabolic response to functional activity between the two regions of these neurons can be explained by the differences in surface-to-volume ratios of these regions.

The suprachiasmatic nucleus (SCN) has been proposed as a site for an endogenous circadian oscillator in mammals, since lesions of the nucleus abolish a wide spectrum of overt-circadian rhythms. To demonstrate that a directly measurable property of the SCN itself in intact (unlesioned) animals is affected by environmental light and exhibits circadian rhythmicity, we used the autoradiographic 2-deoxy-D-[14C]glucose method to determine glucose utilization of rat SCN under a variety of lighting conditions. Our experiments indicate an important role for the SCN in circadian rhythm organization, and we believe the deoxyglucose method will prove useful as a tool for better understanding the functions and mechanisms of circadian clocks. Key words: suprachiasmatic nucleus, circadian rhythm, 2-deoxy-D-[14C]glucose.

A method has been developed for the measurement of the turnover rate constant or the half-life of the free glucose content of brain. It is based on an equation derived by the mathematical analysis of a kinetic model of the equilibration of the specific activity of the free glucose in brain with that of the plasma during an infusion of radioactive glucose. The method requires the measurement of the time course of the specific activity of glucose in the arterial plasma during an intravenous infusion of radioactive glucose for a period of 1 to 4 min and the specific activity of the free glucose in brain at the termination of the infusion. The turnover rate constant, or the half-life, is then calculated from these data by means of the operational equation of the method. The technique has been applied to conscious and anesthetized rats. In conscious rats the half-life of the free glucose content of brain was found to be 1.6 +/- 0.5 min (mean +/- S.D.) when the animals were killed by decapitation and 1.2 +/- 0.2 min (mean +/- S.D.) when they were killed by microwave irradiation; this difference is not statistically significant. In anesthetized rats, the half-life was found to be 2.6 +/- 0.8 min (mean +/- S.D.) in those killed by decapitation and 1.8 +/- 0.3 min (mean +/- S.D.) in those killed by microwave irradiation; this difference is statistically significant. The half-life of the glucose content of brain was found to be significantly prolonged during anesthesia and to be significantly and positively correlated with the plasma glucose concentration (r = 0.78; p < 0.001).