Hepatic glucose production oscillates in synchrony with the islet secretory cycle in fasting rhesus monkeys.

Oscillations in the concentration of plasma glucose were found to reflect large fluctuations in hepatic glucose production. The fluctuations in glucose production were synchronous with fluctuations in the concentration of plasma insulin and glucagon. This synchrony suggests that hepatic pathways are entrained to the islet cycle with a minimal time delay.

Somatostatin: hypothalamic inhibitor of the endocrine pancreas.

Somatostatin, a hypothalamic peptide that inhibits the secretion of pituitary growth hormone, inhibits basal insulin secretion in fasted cats and rats. In fasted baboons both basal and arginine-stimulated secretion of insulin and glucagon are inhibited. Somatostatin appears to act directly on the endocrine pancreas. The action is dose-related, rapid in onset, and readily reversed.

Insulin, glucagon, and glucose exhibit synchronous, sustained oscillations in fasting monkeys.

In overnight fasted rhesus monkeys, synchronous, regular oscillations occurred in the plasma concentrations of glucose, insulin, and glucagon. The oscillations displayed a period averaging 9 minutes. The amplitudes for insulin and glucagon were ten and five times greater than for glucose. Insulin cycled in and glucagon out of phase with glucose. In baboons, oscillations of glucose and insulin were smaller than in rhesus monkeys, while in man, regular oscillations were not observed.

Control of insulin secretion during fasting hyperglycemia in adult diabetics and in nondiabetic subjects during infusion of glucose.

In obese adult diabetics, the concentration of insulin in venous plasma was unrelated to the degree of hyperglycemia after an overnight fast. However, in these subjects, insulin rose and fell in proportion to the magnitude of change in plasma glucose induced by small intravenous infusions of glucose. The minimal dose of glucose to cause a significant rise in insulin above the fasting level was similar in normal subjects, obese nondiabetic subjects, and in obese, hyperglycemic adult diabetics. This dose lay between infusion of 60 and 100 mg of glucose per min for 30 min. These results suggested that the secretion of insulin was under regulation by changes in blood glucose but was not stimulated in proportion to the stable raised blood glucose concentration of the hyperglycemic diabetic. Artificial hyperglycemia was induced in fasting normal subjects by constant intravenous infusion of glucose at rates of 100-250 mg of glucose per min for periods up to 8 hr. Plasma glucose rose during the 1st hr of infusion and then remained constantly elevated for up to 8 hr. The concentration of plasma insulin paralleled that of plasma glucose. During the period of constant hyperglycemia and elevated insulin, superimposition of a brief additional glucose load resulted in a prompt rise in glucose and insulin, both returning to the previous elevated levels. Thus in normals as well as obese diabetics, stable hyperglycemia does not produce a pancreatic response sufficient to return the blood glucose to an arbitrary normal fasting concentration, yet the beta cells remain readily responsive to a change in plasma glucose. These data suggest that the beta cells do not operate as a control system with an absolute reference point when presented with systemic hyperglycemia. The behavior of the beta cells during hyperglycemia in the fasting obese adult diabetic suggests that the regulation of the basal insulin secretion may not be determined by factors directly related to the prevailing concentration of glucose. It is postulated that the beta cells adapt to hyperglycemia perhaps through the operation of controls directed toward a normal delivery of free fatty acids or some other cellular metabolic substrate during fasting.

Studies of substrate regulation in fasting. II. Effect of infusion of glucose io the carotid artery upon fasting lipolysis in the baboon.

The effect of glucose infusion on rates of lipolysis were studied in a group of chair-trained papio baboons that had been prepared for chronic intravenous and intracarotid infusion. All studies were carried out after a 24 hr period of fasting and when the animals were fully awake. After a control interval of 1 hr, a glucose infusion was begun either intravenously or intra-arterially. The infusion was continued at a constant rate for 2 hr and then changed directly to the alternate route and continued an additional 2 hr. Blood samples were collected at 30-min intervals for glucose, free fatty acid (FFA), glycerol, insulin, and in some studies, growth hormone (GH) determination. When glucose doses less than 0.5 mg/kg per min were used, no change in the products of lipolysis was noted during either venous or carotid administration, and glucose and insulin levels remained stable or fell gradually. With doses of glucose between 0.5 and 0.6 mg/kg per min, a greater fall in both FFA and glycerol was noted during carotid administration. No definite changes in plasma glucose or insulin levels were noted during either infusion period. These changes in lipolysis were noted regardless of the sequence of infusion, and a similar differential suppression of FFA was noted during a 24 hr period of carotid glucose administration. When doses of glucose larger than 0.6 mg/kg per min were used, inhibition of lipolysis was noted during both phases of infusion. No definite change in GH levels was noted during the periods of fasting, and the levels of the hormone did not appear to be related to changes in glucose, insulin, or FFA levels. These data provide additional evidence for the presence in the central nervous system of a glucose-sensitive center which alters lipolytic rates independently of insulin and GH, probably by altering sympathetic tone to adipose tissue.

Somatostatin blockade of acute and chronic stimuli of the endocrine pancreas and the consequences of this blockade on glucose homeostasis.

The nature and extent of somatostatin-induced inhibition of pancreatic endocrine secretion were studied by administration of a number of stimuli of either glucagon or insulin to over night fasted baboons with and without an infusion of linear somatostatin. The stimuli for acute-phase insulin release were intravenous pulses of glucose, tolbutamide, isoproterenol, and secretin. When given 15 min after the start of a somatostatin infusion, these agents were essentially unable to stimulate insulin secretion. Chronic insulin secretion was stimulated by infusions of either glucose or glucagon. Within 10 min of the start of a super-imposed infusion of somatostatin, insulin levels fell to less than 40 percent of prestimulus control and remained suppressed for the duration of the somatostatin infusion. Stimulation of glucagon secretion by insulin-induced hypoglycemia was also blocked by somatostatin. Plasma glucose decreased during somatostatin infusions except when superimposed upon an infusion of glucagon. Somatostatin had no effect on glucose production in a rat liver slice preparation. We conclude: (a) Somatostatin is a potent and so far universally effective inhibitor of both acute and chronic phases of stimulated insulin and glucagon secretion (b) The inhibitory effect is quickly reversible and the pattern of recovery of secretion is appropriate to prevailing signals; (c) Present evidence suggests that the effect of somatostatin on blood glucose is mediated through its effect on blood glucagon; (d) In the overnight-fasted baboon both in the basal state and 45 min into a 4-mg/kg-min glucose infusion, a somatostatin-induced fall in serum insulin levels appears to be unable to prevent a decrease in hepatic glucose production.

Insulin dose-response characteristics among individual muscle and adipose tissues measured in the rat in vivo with 3[H]2-deoxyglucose.

The dose-response characteristics of three skeletal muscles, three adipose tissue beds, and heart muscle to single i.v. injection of insulin were compared in vivo. Comparisons were made at 8 dose levels spanning the entire range for response by all tissues and for the integrated whole body response as reflected in the rate of disappearance of 3H-2-deoxyglucose from plasma. The insulin-sensitive tissues varied widely with respect to the magnitude of the maximal response and the sensitivity to insulin as judged by the effective dose 50% (ED 50). Among the muscles, a slow-twitch oxidative muscle, soleus, was more sensitive than the fast-twitch glycolytic muscle, extensor digitorum longus (EDL), while a mixed muscle, quadriceps femoris, displayed even lower sensitivity. Heart muscle sensitivity was comparable to EDL. Among the adipose sites, the rank order of sensitivity was subcutaneous greater than epididymal much greater than omental. The threshold for a detectable response to insulin was 0.013 U/kg rat.

Insulin pulses less effective than continuous insulin in inhibiting PEPCK mRNA levels stimulated by cAMP and dexamethasone in perifused hepatoma cells.

Hepatic glucose production is stimulated in vitro twice as effectively by pulsatile as by continuous glucagon, given equivalent time-averaged doses. Efficacy studies of pulsatile insulin have yielded conflicting results. In the rat hepatoma cell line H-4-II-E-C3, insulin rapidly (t1/2 15 min) inhibits transcription of the gene and lowers mRNA levels for the gluconeogenic enzyme. PEPCK via a receptor-mediated process. We attached H-4-II-E-C3 cells to Cytodex-3 microcarriers and used a perifusion column system to test whether pulsatile insulin is more or less effective than equivalent time-averaged doses of continuous insulin. PEPCK transcription was induced by inclusion of cAMP analogue 8-(4-chlorophenyl-thio)-cAMP (0.1 mM) and dexamethasone (0.5 microM) in the perifusion medium. Three columns were exposed either to continuous, pulsatile, or no insulin. After 3 h, total nucleic acid was extracted, and mRNA(PEPCK) was measured with a sensitive-solution hybridization assay. Continuous insulin inhibited PEPCK expression in a dose-dependent fashion with EC50 1 x 10(-11) M. Equivalent time-averaged amounts of insulin delivered as pulses achieved significant inhibition but less effectively than continuous insulin. The apparent EC50 for pulsatile insulin increased from 2 x 10(-11) M to 5 x 10(-11) M as the oscillatory period was raised from 5 to 20 min, respectively. These observations suggest that insulin-mediated inhibition of PEPCK gene transcription is diminished by a pulsatile mode of administration in marked contrast to the pulse enhancement demonstrated for glucagon-mediated hepatic glucose production.

Lack of evidence for improvement in long-term glycemic control by pulsatile insulin infusion in streptozocin-induced diabetic baboon.

To assess the potential therapeutic use of pulsatile intravenous insulin delivery, five streptozocin-induced diabetic baboons were treated with alternate 3- to 6-wk periods of pulsatile and continuous insulin infusion. Time-averaged insulin concentrations were matched during two pulsatile administration periods (P1 and P2) and an intervening period of continuous insulin administration (C). There were no significant differences among the overall means of four daily glucose determinations performed during the three periods (P1, 5.7 +/- 1 mM; C, 5.6 +/- 0.9 mM; P2, 5.3 +/- 0.9 mM); the mean M value, a measure of the stability of glycemic control (P1, 4 +/- 1.7; C, 3.9 +/- 1.8; P2, 3.6 +/- 1.5); the percentage of glucose values less than 2.8 mM (P1, 13 +/- 8.5%; C, 14 +/- 12%; P2, 13 +/- 9.1%); or the glycosylated hemoglobin levels determined at the end of the P1 and C (7.5 +/- 3.4 and 6.5 +/- 1.8%, respectively [all values are means +/- SD]). Fasting hepatic glucose production was suppressed to a similar degree during pulsatile and continuous insulin infusion (P1, 23 +/- 3 mumol.kg-1.min-1; C, 24 +/- 8 mumol.kg-1.min-1). Arterial glucagon levels were similar during pulsatile and continuous insulin infusion, both in the fasting state (84 +/- 29 and 84 +/- 31 ng/L, respectively) and postprandially (30 +/- 14 and 27 +/- 12 ng/L, respectively). Pulsatile insulin infusion failed to entrain a corresponding glucagon secretory rhythm. These data suggest that the metabolic consequences of long-term pulsatile and continuous insulin infusion in an animal model of human non-insulin-dependent diabetes are comparable.

Rapid reduction and return of surface insulin receptors after exposure to brief pulses of insulin in perifused rat hepatocytes.

Hepatic receptors are normally exposed to discrete pulses of insulin and glucagon at intervals of 8 to 16 min. Using a multicolumn system for perifusing hepatocytes, we investigated the effect of this pattern on the normal processing of the insulin receptor. Surface-receptor binding was measured in acid-washed cells harvested from individual columns. The number of high-affinity surface receptors fell to a nadir 1 min after the end of a 3-min square-wave pulse of insulin. The maximum reduction reached 45% of baseline at amplitudes of 1000 microU/ml or above. The number of surface binding sites returned to baseline 15 min after the end of the pulse, but the affinity constant of the high-affinity receptor was unchanged. The reduction of surface binding was dose dependent, with an ED50 of 251 +/- 34 microU/ml. Prolonging the pulse to 60 min did not affect the nadir or the rate of restoration of the surface-receptor population. The change in surface binding was reduced at 15 degrees C and abolished at 4 degrees C. After a pulse, the pattern of change was a period of rapid decline to a nadir (t1/2 less than or equal to 1 min) that persisted for 3-5 min, followed by restoration of surface binding that reached baseline in 10-15 min. This same pattern was present after six ED95 pulses delivered at intervals of 15 min. These data indicate that the internalization of hepatocyte surface receptors and their recycling and reinsertion into the plasma membrane can be entrained to pulses at the physiologic pulse frequency.

Decreased insulin- and glucagon-pulse amplitude accompanying beta-cell deficiency induced by streptozocin in baboons.

The effect of beta-cell deficiency on the spontaneous pulsatile secretory pattern of the islets of Langerhans was studied in the baboon. Measures of beta-cell function were correlated with the secretory pattern before and at intervals after streptozocin administration. The degree of insulin deficiency was variable and ranged from mild to moderate. Highly regular pulses were less prevalent in baboons compared with rhesus monkeys and humans, but the mean frequency was similar and was not affected by treatment. The principal effect of beta-cell destruction was to proportionately reduce the pulse amplitude of insulin (-39%, P less than .003) without detectable change in pulse frequency, interhormonal phase relationship, or the regularity of pulses. Glucagon-pulse amplitude also fell (-19%, P less than .09), but not significantly. However, glucagon-pulse amplitude was strongly correlated with insulin-pulse amplitude (r = -.59, P less than .002), whereas mean fasting plasma concentrations of insulin and glucagon were not significantly changed after treatment. Because streptozocin affects only the beta-cell, the data indicate a major influence of the insulin pulse on the alpha-cell secretory pulse. The data do not support the presence of a separate pacemaker for the alpha-cell but do not eliminate this possibility. The strong correlation of reduction in insulin-pulse amplitude with increasing fasting glucose and decreasing glucose disappearance lends support to growing evidence that the pattern of insulin secretion is an important determinant of normal glucose homeostasis.

A [3H]2-deoxyglucose method for comparing rates of glucose metabolism and insulin responses among rat tissues in vivo. Validation of the model and the absence of an insulin effect on brain.

In 1980 we described an in vivo method for estimating the rate of glucose uptake among selected tissues during an acute insulin response. The method was based upon the same principles as Sokoloff's 2-deoxyglucose (2DG) method. We now report further examination of the basic assumptions of the model and validation of its general applicability by comparing the response of brain and other tissues to prolonged insulin infusion (while glucose is held constant) with their response to a single injection of insulin. The method provides a reproducible estimate of relative insulin response in any tissue that can be anatomically separated at death. Tissues that are minimally sensitive to insulin such as spleen, lung, skin, and gut do not display increments in the calculated value for net rate of tissue uptake of 2DG. Insulin-sensitive tissues display increased rates of uptake that are characteristic for each specific tissue, ranging in magnitude from 1.7- to 17.9-fold over basal among an array of insulin-sensitive tissues. The duration of a unit response to a sub-maximal dose of insulin also varied among the tissues, persisting for 20-30 min after plasma insulin had returned to basal in heart and for 10-20 min in the other insulin-sensitive tissues. The method provides a reproducible measure of glucose metabolism in vivo and has been validated as a means of quantifying relative insulin sensitivity among the peripheral tissues. During steady-state conditions with plasma glucose held constant, brain glucose metabolism was unaffected by a 60-min infusion of insulin.

Evidence that the physiological pulse frequency of glucagon secretion optimizes glucose production by perifused rat hepatocytes.

We have reported that glucagon administered to perifused rat hepatocytes as a series of pulses at 15-min intervals is a more effective stimulus for hepatocyte glucose production (HGP) than is continuous glucagon infusion. To test whether the efficiency of HGP depends upon the frequency of pulsatile glucagon delivery, we administered glucagon to perifused rat hepatocytes as a series of pulses of fixed amplitude [922 +/- 30 (+/- SE) pg/ml] at eight separate pulse intervals ranging from 3-45 min. Compared to continuous infusion of the same total amount of hormone, pulsatile glucagon administration clearly enhanced HGP in a frequency-dependent fashion. At pulse intervals between 10 and 20 min, pulsatile HGP exceeded continuous HGP by a factor of 1.5-2. This range of optimal intervals compared favorably with the glucagon secretory period of 10 min observed in nonhuman primates and that of 13-20 min observed in humans. We noted a desensitization of the hepatocyte response to glucagon that was directly proportional to the log of the time-averaged hormone concentration. Since the magnitude of the desensitization elicited by pulsatile glucagon delivery exceeded the desensitization elicited by continuous hormone delivery regardless of pulse frequency, differential desensitization could not explain the frequency dependency of the pulse enhancement effect. A mathematical simulation of our data demonstrated that the asymmetry of the HGP waveform elicited by a brief glucagon pulse could account for the observed frequency dependency of HGP. Pulse to pulse summation and the desensitization phenomenon modulated both the magnitude of the pulse enhancement effect and the frequency range over which the effect was manifest. We conclude that the enhancement of HGP by glucagon pulses is a frequency-dependent phenomenon and that the physiological glucagon secretory period optimizes HGP.

Investigation of the effect of insulin upon regional brain glucose metabolism in the rat in vivo.

The hypothesis that insulin might promote increased glucose metabolism in putative glucoreceptor areas of the brain was investigated in the rat. Using tritiated 2-deoxyglucose (2-DG), unanesthetized fasted rats were injected with 0.1 U insulin and studied 30 min later. The local uptake of 2-DG into discrete brain areas was examined in serial frozen 400-micrometer sections. Areas 1.1 mm in diameter were punched from the region of the ventral medial, ventral lateral, and dorsal hypothalamus and from a control area from the cerebral cortex. The punched tissue segments were analyzed for total radioactivity and protein content. The results showed that insulin failed to influence the pattern of 2-DG uptake into these discrete brain regions. When the data were analyzed with a simple kinetic model to determine the net fractional rate of uptake of 2-DG by the tissues, brain tissues displayed a 75% rate increase compared to saline-treated controls. Heart muscle collected from the same rats showed a 700% increase after insulin, while lung, an insulin insensitive tissue, displayed a 30% increase. Because the nonsteady state conditions of the model dictate a number of assumptions, the modest increase in the calculated rate of uptake in brain tissue must be verified by a steady state model before it can be accepted as representing a real effect of insulin upon the overall metabolism of glucose in the brain. Regardless of these reservations, it may be concluded from the pattern of response, that insulin does not selectively increase glucose uptake or metabolism in the putative glucoreceptor areas of the hypothalamus under the conditions of these experiments.

A kinetic analysis of hepatocyte responses to a glucagon pulse: mechanism and metabolic consequences of differences in response decay times.

Pulsatile administration of glucagon to perifused rat hepatocytes stimulates hepatocyte glucose production (HGP) more effectively than continuous administration. Having established that this effect was due to delayed relaxation of glucagon-stimulated HGP (t1/2 for decay = 3.54 +/- 0.60 min) we wished to examine the mechanism of response termination. Delayed dissociation of glucagon from its receptor was excluded by the brisk washout of [125I]glucagon from perifusion columns (t1/2 = 1.00 +/- 0.13) and the rapid decay in glucagon-stimulated cAMP released into the perifusion medium (t1/2 = 1.14 +/- 0.12). The relaxation of the HGP response to a pulse of administered cAMP was comparable to the decay in glucagon-stimulated HGP (t1/2 = 3.28 +/- 0.22). Furthermore, the phosphodiesterase inhibitor isobutyl-methylxanthine did not alter the decay of the HGP response to glucagon despite increasing the amplitude of the response (t1/2 = 3.04 +/- 0.36). These data place the rate-limiting step for HGP relaxation distal to cAMP generation and degradation. The decay of the beta-hydroxybutyrate response to a glucagon pulse was not different from the cAMP response (t1/2 = 1.14 +/- 0.23), whereas the decay of gluconeogenesis from lactate was not significantly different from HGP relaxation (t1/2 = 1.94 +/- 0.08). We conclude that rate-limiting events for HGP relaxation occur distal to the second messenger cascade; however, ketogenesis is more closely coupled to the kinetics of cAMP. These results may help to explain the absence of excessive ketosis during fasting in normal humans, who secrete glucagon episodically at 10- to 14-min intervals.

Effects of branched chain amino acids on spontaneous growth hormone secretion in the baboon.

Blood concentrations of the branched chain amino acids (BCAAs) are elevated during fasting in healthy subjects and are abnormally high both postprandially and during fasting in diabetic patients. Despite evidence that these amino acids influence brain metabolism and neurotransmitter synthesis, there is little information on the neuroendocrine effects of the BCAAs. This study provides evidence that elevation of postprandial blood levels of the BCAAs alters the ultradian rhythm of GH secretion observed in the baboon during daylight hours. To mimic the postprandial rise in the BCAAs that occurs in diabetic patients, we infused either saline or a mixture of valine, leucine, and isoleucine into six conscious male baboons from 1530-1900 h daily for 4 days during and after the normal feeding time. On the last day of the infusions, blood samples were collected at 20-min intervals from 0800-1500 h and at 30-min intervals from 1500-2000 h. The amino acid infusions increased postprandial blood concentrations of the BCAAs 2- to 5-fold over control levels and lowered the blood concentrations of tyrosine, phenylalanine, and lysine compared to concentrations observed during control infusions. A significant elevation in GH levels occurred in association with BCAA treatment in each animal between 0800 and 1100 h, 13 h after the previous day's infusion. Average +/- SE maximum GH levels observed between 0800 and 1100 h were 11.6 +/- 2.9 ng/ml under experimental conditions compared to a control value of 3.8 +/- 1.2 (P less than 0.02). Whether the increased GH levels represented the generation of a new peak or a phase shift in a nocturnal peak was not determined. Combined with evidence that spontaneous release of GH is neurally regulated in the baboon, this study suggests that changes in the blood levels of the BCAAs modulate neural mechanisms that regulate GH rhythmicity.

Inhibition of glucagon and insulin secretion by somatostatin in the rat pancreas perfused in situ.

Perfusion of growth hormone inhibitory factor (somatostatin) into rat pancreas inhibited secretion of glucagon and insulin into medium containing 5.5 mM glucose. A 15-min infusion of arginine (20 mM) greatly increased glucagon and insulin secretion. When perfused simultaneously with arginine, somatostatin (55 nM) abolished the increase in glucagon secretion. The acute phase of insulin secretion in response to arginine was attenuated by somatostatin, and subsequent secretion was decreased to control levels. Pretreatment for 5 min with somatostatin blocked even acute-phase insulin secretion in response to arginine. Somatostatin did not affect basal or glucose-stimulated secretion of insulin from rat pancreatic islets isolated by the collagenase technique. Arginine-stimulated secretion of insulin was enhanced by somatostatin in isolated islets. These results demonstrate a direct effect of somatostatin on the pancreas to inhibit secretion of glucagon and insulin. The failure of somatostatin to inhibit insulin secretion in pancreatic islets may be due to alterations in the beta cells produced by the isolation procedure. It is also possible that the effect of somatostatin on insulin secretion may be mediated indirectly.

Synchronous, sustained oscillation of C-peptide and insulin in the plasma of fasting monkeys.

The previously reported synchronous oscillations in plasma glucose and insulin levels have been further studied to determine whether the phenomenon can be attributed to cyclic secretion or degradation of the hormone. Plasma C-peptide concentrations were measured and found to cycle with the same period as plasma insulin suggesting that the oscillations of insulin arise from changing secretion rather than degradation. The amplitude of C-peptide oscillation was 50 percent that of insulin. Assuming first order kinetics and the same relative rates of disappearance as in humans, this difference in amplitudes in consistent with equal secretion rates for the two peptides.

Effects of fasting on growth hormone secretion in the male baboon.

The effects of prolonged fasting on the quantity and pattern of spontaneous GH secretion in 5 adolescent male baboons were investigated. Serum GH concentrations were measured in blood samples taken at 20-min intervals over 12 daytime hours after an overnight fast (control period) and during 84-96 h of fasting. Rhythmic GH secretion, with a mean (+/- SE) period of 5.4 +/- 0.2 h occurred in 4 of the 5 animals in 11 control experiments, and GH peaks occurred randomly in the fifth animal. In response to prolonged fasting, the percent half-amplitude of daytime GH peaks decreased from control values of 144 +/- 12% to 105 +/- 17%. The period of the GH rhythm in 4 animals decreased during fasting, but the change was not statistically significant, and the episodic pattern of GH release in animal 5 was apparently unaffected by fasting. After 84 h of fasting, the mean and integrated concentrations of GH released over 12 daytime hours in the 5 animals were not significantly different from control values. In summary, despite a reduction in GH peak amplitude, the quantity and rhythm of GH secretion were maintained in baboons fasted for 84 h. The observed decrease in GH maxima may play a role in the metabolic adaptation to fasting in the baboon.

Pulsatile release of growth hormone and prolactin from the primate pituitary in vitro.

Perifused anterior hemipituitaries from one male and 4 female monkeys released GH and PRL in a pulsatile pattern, with mean +/- SE interpulse intervals of 8.2 +/- 0.4 and 8.5 +/- 0.3 min, as determined by a cycle detection computer algorithm. Mean hormone concentrations in the perifusate fractions collected at 2-min intervals were 435 +/- 89 (GH) and 515 +/- 262 (PRL) ng/ml. Pulse amplitudes averaged 74 +/- 16 ng/ml for GH and 189 +/- 89 ng/ml for PRL. These findings suggest the presence of a high frequency pulsatile secretory mechanism within the primate pituitary.