Insulin resistance in acromegaly causes glucose intolerance and diabetes, but it is unknown whether it involves protein metabolism, since both insulin and growth hormone promote protein accretion. The effects of acromegaly and of its surgical cure on the insulin sensitivity of glucose and amino acid/protein metabolism were evaluated by infusing [6,6-2H2]glucose, [1-13C]leucine, and [2-15N]glutamine during a euglycemic insulin (1 mU · kg−1 · min−1) clamp in 12 acromegalic patients, six studied again 6 mo after successful adenomectomy, and eight healthy controls. Acromegalic patients, compared with postsurgical and control subjects, had higher postabsorptive glucose concentration (5.5 ± 0.3 vs. 4.9 ± 0.2 μmol/l, P < 0.05, and 5.1 ± 0.1 μmol/l) and flux (2.7 ± 0.1 vs. 2.0 ± 0.2 μmol · kg−1 · min−1,P < 0.01, and 2.2 ± 0.1 μmol · kg−1 · min−1,P < 0.05) and reduced insulin-stimulated glucose disposal (+15 ± 9 vs. +151 ± 18%, P < 0.01, and 219 ± 58%, P < 0.001 from basal). Postabsorptive leucine metabolism was similar among groups. In acromegalic and postsurgical subjects, insulin suppressed less than in controls the endogenous leucine flux (−9 ± 1 and −12 ± 2 vs. −18 ± 2%, P < 0.001 and P< 0.05), the nonoxidative leucine disposal (−4 ± 3 and −1 ± 3 vs. −18 ± 2%, P < 0.01 andP < 0.05), respectively, indexes of proteolysis and protein synthesis, and leucine oxidation (−17 ± 6% in postsurgical patients vs. −26 ± 6% in controls,P < 0.05). Within 6 mo, surgery reverses insulin resistance for glucose but not for protein metabolism. After adenomectomy, more leucine is oxidized during hyperinsulinemia.
- glucose metabolism
- growth hormone
- leucine metabolism
growth hormone receives widespread attention because of its anabolic properties and, consequently, for its potential clinical applications in hormone-deficient, elderly, and hypercatabolic patients. Growth hormone, however, is not free from adverse effects when excessively concentrated. Acromegalic subjects are frequently glucose intolerant or diabetic. For this reason, the interaction of growth hormone and insulin on glucose metabolism was extensively studied. In this respect, growth hormone counteracts insulin action (6, 16, 32). In contrast, the interaction of the two hormones on the full spectrum of their metabolic actions is scarcely defined. Of particular interest is the regulation of protein anabolism. Both hormones share a positive effect on protein balance that is of potential therapeutic interest. Many studies have investigated the mechanisms by which insulin promotes protein anabolism. Insulin increases the synthesis of specific proteins in target organs (3, 11), but it decreases the synthesis of others, the net effect resulting in no stimulation of protein synthesis. In contrast, insulin per se consistently reduces the proteolytic rate (9, 14). The latter is currently believed to be the most important mechanism explaining the effect of insulin on protein accretion. Both growth hormone and IGF-I stimulate protein synthesis in humans in vivo (24). Interestingly, a study performed in the human forearm showed that the simultaneous administration of the two hormones acutely reduced the antiproteolytic effect of insulin, suggesting that, in the muscular tissue, growth hormone-induced insulin resistance also involves protein metabolism (13). It is unknown whether at the whole body level the resulting effect of growth hormone is to blunt the anticatabolic effects of insulin. Molecular studies have shown that chronic growth hormone interacts in a complex fashion with the insulin-signaling pathways: growth hormone inhibits the upstream elements of the cascade (34, 36, 37), leading to the insulin effects on glucose, lipid, and protein metabolism, but it independently activates the downstream kinases (MAP kinases and p70 S6 kinase) specifically involved in the stimulation of protein synthesis (19, 29).
These pieces of in vivo and in vitro evidence support the hypothesis that chronic growth hormone counteracts insulin action on proteolysis and protein synthesis and still promotes protein anabolism. To address this issue, we tested the insulin sensitivity of glucose and protein metabolism in a clinical model of chronically elevated growth hormone, i.e., in acromegalic patients. In this population, key questions concerned 1) the degree of impairment in the antiproteolytic action of insulin compared with glucose metabolism, and 2) the reversibility of such impairments with the surgical treatment of acromegaly.
l-[1-13C]leucine,l-[2-15N]glutamine, andd-[6,6-2H2]glucose were purchased from MassTrace (Woburn, MA). Chemical and isotopic purity of the tracers was determined by gas chromatography-mass spectrometry (GC-MS). Before every infusion study, sterile solutions of the tracers were prepared using aseptic technique. Accurately weighed amounts of the labeled compounds were dissolved in weighed volumes of sterile, pyrogen-free saline and filtered through a 0.22-μm Millipore filter before use. An aliquot of the sterile solution was initially verified to be pyrogen free before administration to human subjects. Solutions were prepared no more than 24 h before use and were kept at 4°C before administration.
Twelve acromegalic subjects affected by benign growth hormone-secreting pituitary adenoma were studied before surgery (ACRO; 6 females and 6 males, age 47 ± 4 yr, body weight 74 ± 4 kg, body mass index 26.0 ± 1.0 kg/m2). The study was repeated in six of the subjects 6 mo after surgery (POST). Eight healthy controls were also studied with the same protocols (CON; 4 females and 4 males, age 42 ± 4 yr, body weight 69 ± 4 kg, body mass index 24.2 ± 1.1 kg/m2). In selected subjects, dual-X-ray absorptiometry (DEXA) was performed with a Lunar-DPX-IQ scanner (Lunar, Madison, WI) to assess body composition. Regional analysis was performed in the arms, trunk, and legs with three-compartment processing (30).
Euglycemic hyperinsulinemic clamp.
One week before their admission, the subjects were placed on an adequate energy and protein (1.2 g · kg−1 · day−1) diet that continued until the last study was completed. The diet was required to reduce intersubject variability induced by the different subjects' dietary practices. Acromegalic subjects were studied as inpatients during their assessment before surgery and at their readmission 6 mo after surgery. Subjects were generally admitted for 2–4 days before receiving the clamp. On the evening before each infusion study, the subjects consumed their evening meal by 8:30 PM, and then they drank only water until completion of the study on the next day at 2:00 PM. At 7:00 AM on the infusion day, a venous catheter was placed in the subject's arm for infusion of the tracers of glucose and amino acids, and insulin and dextrose when necessary. A second catheter was placed retrograde in a hand vein, and the subject's hand was placed in a warming box to obtain arterialized venous blood samples. The catheters were kept patent with a slow infusion of sterile saline. At the beginning of both studies, priming doses of [1-13C]leucine (4.5 μmol/kg), [2-15N]glutamine (4.5 μmol/kg), and [6,6-2H2]glucose (12.0 μmol/kg) were administered intravenously and were immediately followed by the continuous infusion of the same tracers (4.5, 7.5, and 13.3 μmol · kg−1 · h−1, respectively) for 5 h. The study was composed of two periods of identical duration (2.5 h each): basal for tracer equilibration and euglycemic hyperinsulinemic clamp, performed as previously described (21). Briefly, insulin was infused at the rate of 1 mU · kg−1 · min−1to achieve and maintain insulin concentrations of ∼420 pmol/l, and 20% dextrose was infused at a variable rate to maintain the glucose concentration at 5 mmol/l. To this purpose, plasma glucose concentrations were measured at bedside every 5 min. Blood and breath samples were drawn just before the start and at 15-min intervals during the last 45 min of the 2.5-h basal period and throughout the 2.5-h clamp period.
Aliquots of blood were placed in tubes containing EDTA and stored on ice until the plasma was prepared by centrifugation at 4°C. A 0.5-ml aliquot was withdrawn, defined amounts of [2H4]alanine, [2H7]leucine, [2H2]phenylalanine, [2H5]glutamine, and α-ketoisocaproate (KIC) were added as an internal standard for quantitation of alanine, leucine, phenylalanine, glutamine, and KIC plasma concentrations, and the plasma was frozen at −60°C. One-milliliter blood aliquots for measurements of glucagon were placed in tubes containing EDTA plus aprotinin. Blood aliquots for insulin, cortisol, and growth hormone were collected in tubes without additives for serum separation. All blood samples were placed on ice until the plasma or serum was prepared by centrifugation at 4°C (≤1.5 h of drawing). All plasma and serum aliquots were frozen at −60°C until later analysis. Breath samples were placed into 20-ml evacuated tubes until measurement of13CO2 in expired air. Each subject's CO2 production rate was measured periodically for 15-min periods using an indirect calorimeter with a flow-through canopy system (model 2800 Z; Sensormedics, Palo Alto, CA). The CO2production rate was used to calculate leucine oxidation.
Plasma amino acid concentrations and enrichments were measured by electron impact GC-MS. Before derivatization, amino acids were isolated from plasma by use of cation exchange columns, as previously described (2). Amino acids eluted from the columns were evaporated to dryness and derivatized to form thetert-butyldimethylsilyl (TBDMS) derivative. The [M-57]+ ions at mass-to-charge ratios (m/z) = 260 and 264 were monitored for unlabeled alanine and [2H4]alanine, respectively. The [M-57]+ ions at m/z = 336 and 338 were monitored for unlabeled phenylalanine and [2H2]phenylalanine, respectively. The [M-57]+ ions at m/z = 302, 303, and 309 were monitored for unlabeled leucine, [1-13C]leucine, and [2H7]leucine, respectively. The [M-57]+ ions at m/z = 431, 433, and 436 were monitored for unlabeled glutamine, [1,2-13C2]glutamine, and [2H5]glutamine, respectively. TBDMS-glutamine was chromatographically resolved from TBDMS-glutamate. The KIC enrichments and concentrations were measured after the eluant from the columns was derivatized to the trimethylsilyl-quinoxalinol derivative (2). Injections of the derivatives were made into a GC-MS instrument (model 5970; Hewlett-Packard, Palo Alto, CA) that was operated using electron impact ionization. The ions atm/z 259 and 260 were monitored for unlabeled KIC and [1-13C]KIC, respectively. The KIC peak was resolved from that of KIC and was used for the quantification of the KIC concentrations. For all measurements, the background corrected tracer enrichments in mole percent excess were calculated as previously defined (22). The measurement of13CO2 in the expired air KIC was performed by isotope ratio mass spectrometry (VG Isogas; VG Instruments, Middlewich, UK).
Plasma hormone concentrations were measured by radioimmunoassay with commercial kits, as previously described (30).
The glucose and glutamine kinetics were calculated using Steele's equations for the nonsteady state (35), as we previously described (2). The rate of appearance of the unlabeled substrates (Ra, μmol · kg−1 · h−1) was calculated using the equation Equation 1where i is the infusion rate of the tracers as μmol · kg−1 · h−1of tracer per se (i.e., the product of the rate of tracer infusion times the enrichment of the tracer), Vd is the volume of distribution, C(t) is the plasma concentration of the tracee (μM) at time t, E(t) is the enrichment in plasma, and dE(t)/dt and dC(t)/dt are the rates of change with time of enrichment and concentration, respectively. The disappearance rate (Rd) is equal to Ra under steady-state conditions but must be adjusted for the rate of change of an expanding or contracting pool of substrate Equation 2The value for the glucose Vd was assumed to be 0.16 l/kg. A glutamine Vd value of 0.38 l/kg was used on the basis of previous work concerning the tracer-miscible glutamine pool in healthy subjects (10).
We did not attempt to describe the intracellular leucine kinetics with equations for the nonsteady state, because this approach would imply many more assumptions than the simpler, monocompartmental approach described. To define the leucine release from proteolysis, the intracellular leucine enrichments were estimated by the plasma [1-13C]KIC enrichments, which are derived from the intracellular leucine reciprocal pool approach (2) with the standard steady-state equation Equation 3where I is the leucine tracer infusion rate as μmol · kg−1 · h−1, Ei is the enrichment of the [1-13C]leucine tracer, and Ep is the [1-13C]KIC enrichment in plasma. The rate of excretion of the [13C]leucine into exhaled 13CO2 was calculated as Equation 4where F13C is the rate of oxidation of the [13C]leucine to 13CO2, is the CO2 enrichment, and V˙co 2 is the rate of CO2 production. The oxidation of the dextrose infused with insulin in study 1 increased the breath 13CO2 enrichment, because this sugar has a higher 13C enrichment than the endogenous glucose (14). The contribution of the exogenous dextrose oxidation to the exhaled 13C enrichment was determined in pilot experiments and was subtracted from the breath13CO2 enrichment in the present studies. The rate of leucine oxidation is given by Equation 5where 0.81 is the recovery factor of the label in exhaled CO2 (1, 23). Finally, the rate of leucine flux in protein synthesis was calculated as the difference between the leucine Ra and the rate of leucine oxidation (2).
The glutamine appearance rate in the systemic circulation was estimated using Eq. 1 , where Ei is the enrichment of the [2-15N]glutamine tracer, and Ep is the [2-15N]glutamine enrichment in plasma.
t-Tests for paired data were used to compare ACRO and POST, and t-tests for independent data were used to compare ACRO with CON and POST with CON. The Bonferroni correction was applied to both tests. Comparisons between equilibration period and study period within each study group were performed by means of a two-tailed pairedt-test. The major end point of this study was to evaluate the changes in insulin sensitivity induced by acromegaly and by its surgical cure. The effect of insulin was calculated as the percent change from the basal state during the last hour of the insulin clamp. The effect of insulin was then compared among groups.
The postabsorptive glucose concentration was higher in ACRO than in POST but similar to CON (Table1). The postabsorptive glucose flux was higher in ACRO than in POST and CON. During the clamp, the insulin stimulation of glucose disposal was markedly impaired in ACRO compared with both POST and CON. The endogenous glucose production was less suppressed in ACRO than in CON during the clamp. The postabsorptive glucose clearance was not different among groups (2.72 ± 0.14 vs. 2.19 ± 0.17 vs. 2.32 ± 0.14 ml · kg−1 · min−1in ACRO, POST, and CON, respectively), and during the clamp it increased in ACRO by only 20 ± 6% compared with 151 ± 24% in POST and 208 ± 69% in CON (P < 0.05 andP < 0.001 vs. ACRO, respectively).
Amino and keto acid concentrations.
The postabsorptive leucine concentration was comparable among groups (Table 2). During the clamp, leucine decreased in all groups to a plateau that was lower in POST than in ACRO. The effect of the clamp to decrease the leucine concentration was smaller in ACRO than in POST and in CON. The absolute KIC concentration was not different among groups, but the effect of the clamp to reduce the KIC concentration was smaller in ACRO than in POST and in CON. Phenylalanine was reduced in ACRO postabsorptively. The effect of the clamp was to reduce phenylalanine in all groups, but this effect was smaller in ACRO and in POST than in CON. The alanine concentration was similar among the groups and was not affected by the clamp. Glutamine was not different among groups and was similarly suppressed during the clamp.
Leucine and glutamine kinetics.
Postabsorptively, the endogenous leucine flux, an index of proteolysis, and the leucine oxidation and the nonoxidative leucine disposal, an index of protein synthesis, were comparable among the groups (Table3). During the clamp, proteolysis was less suppressed in ACRO and in POST than in CON. Leucine oxidation during the clamp was less suppressed in POST compared with both ACRO and CON. The nonoxidative leucine disposal was suppressed by insulin only in CON and (to an intermediate level) in POST, whereas it was not suppressed in ACRO before the intervention. The glutamine flux was comparable among groups and was suppressed in all groups during the clamp. Apparently, the insulin suppression of glutamine flux was more pronounced in ACRO than in the other groups, but a significant difference was not detected with the statistical analysis model used. Significance was reached (P = 0.04) when the suppression in ACRO was compared with that of the pooled POST and CON groups, even though this statistical approach was not used extensively because POST was paired and CON was unpaired to ACRO.
Hormones and metabolites.
The postabsorptive insulin concentration was increased more in ACRO before the intervention than in the other groups, whereas it was comparable during the clamp (Table 4). The postabsorptive C-peptide concentration was higher in both ACRO and POST compared with CON, and in POST it also remained higher during the clamp. In ACRO, the glucagon concentration during the clamp was higher than in POST, whereas the cortisol concentration was lower than in CON. The growth hormone and IGF-I concentrations were markedly increased in ACRO before the intervention, and they were completely normalized after the intervention.
Body composition before and after surgery.
We evaluated whether surgical cure of acromegaly induced changes in body composition that could have affected protein metabolism. The six subjects who were studied both before and after surgery (3 females, 3 males) had a body weight and a body mass index that did not change significantly among studies (from 79 ± 7 to 81 ± 6 kg and from 27.2 ± 1.5 to 27.9 ± 1.6 kg/m2, respectively). In four acromegalic subjects, DEXA scans were performed to assess body composition. Two of them (1 male and 1 female) had measurements both before and after surgery, whereas one female was measured exclusively before surgery and one female exclusively after surgery. Overall, percentages of lean and fat mass were 67.3 ± 3.2 and 29.5 ± 2.6% before surgery and 64.2 ± 1.8 and 33.3 ± 1.7% after surgery. The two subjects who were studied both before and after surgery gained 2.8 and 4.8 kg of fat mass and lost 1.6 and 1.1 kg of lean mass, respectively.
This study investigated the insulin sensitivity of glucose and protein metabolism in acromegalic patients with a benign pituitary adenoma before and after successful adenomectomy. All subjects had a marked and chronic elevation of growth hormone and IGF-I. These hormones were completely normalized 6 mo after the intervention. The surgical cure of acromegaly caused the complete reversal of the alterations in glucose metabolism that were evident before the intervention. That acromegaly caused a marked and reversible impairment in insulin-mediated glucose disposal was an expected finding (15,17, 18, 27). It is remarkable that, in our subjects (similarly to the previously studied nondiabetic acromegalic patients), the postabsorptive glucose concentration and flux were only 10 and 30% increased, with the paradox of a resulting glucose clearance that was higher, albeit not significantly, compared with the control groups. In contrast, the impairment in insulin-stimulated glucose disposal was almost complete. One should infer that, beyond the compensatory effects of postabsorptive hyperinsulinemia, the glucose effectiveness in acromegalic patients is increased at the expense of insulin sensitivity, suggesting that growth hormone and IGF-I exert insulin-like effects to promote glucose uptake in the postabsorptive state. To support this idea, opposite changes were obtained by replacing growth hormone in growth hormone-deficient adults: insulin sensitivity decreased and glucose effectiveness increased (31). Our findings on the effects of acromegaly on glucose metabolism are in line with the idea that the marked insulin insensitivity for the stimulation of glucose uptake does not necessarily imply an absolute defect in postabsorptive glucose uptake. This is also suggested by molecular studies that showed that chronic elevation in growth hormone does not impair basal glucose uptake or the activity of the insulin-signaling pathway: when not insulin stimulated, it just impairs the insulin stimulation of these elements (36).
Plasma amino and keto acid concentrations were similar in ACRO and CON except for a modest (11%) reduction in phenylalanine levels in acromegalic subjects before treatment. Proteolysis, leucine oxidation, and protein synthesis were similar to those in CON. We did not find a significant effect of acromegaly on postabsorptive protein kinetics, but there was a tendency for proteolysis to be decreased in the acromegalic subjects, which was more marked after surgery. Taken together with other studies, our data suggest that increased circulating growth hormone can increase proteolysis in adults if they are growth hormone deficient (20, 25) but not if they have an adequate (12) or increased growth hormone secretion. In contrast, the insulin control of proteolysis was severely impaired. The defective suppression of proteolysis was paralleled by a defective suppression of leucine, KIC, and phenylalanine concentrations during the clamp. Our data agree with studies showing that growth hormone acutely inhibits the insulin-dependent suppression of proteolysis in the forearm (13). Suppression of proteolysis is the main mechanism by which insulin promotes protein accretion in humans in vivo in the whole body (9, 14). It is now known that this action is mediated by the ubiquitin-proteasome-dependent pathway, but regulation by the intracellular signaling pathways is still unclear (26).
Interestingly, the defective suppression of proteolysis did not imply a lesser anabolic effect by hyperinsulinemia. The insulin suppression of leucine oxidation was completely maintained; thus a greater proportion of the proteolytic flux was shunted toward protein synthesis during hyperinsulinemia in ACRO. Growth hormone and IGF-I have the effect of decreasing leucine oxidation in various conditions (7, 8, 12, 20,25). ACRO patients were insulin resistant for the suppression of proteolysis and not for the oxidation of leucine, whereas POST subjects were insulin resistant for both aspects of insulin action on protein metabolism. As a consequence, hyperinsulinemia resulted in a stimulus that was more anabolic in ACRO than in POST.
It is unclear why, 6 mo after adenomectomy, the insulin resistance for glucose metabolism was completely reversed whereas the insulin resistance for protein metabolism was persisting. One possibility is that slow changes in protein mass were still occurring 6 mo after surgery and that the insulin resistance for the suppression of leucine oxidation in POST reflected a tendency toward an increased overall protein catabolism. It was shown that acromegaly causes a reduction in fat mass and an increase in fat-free mass, which are largely due to extracellular water retention, and these changes are reversible within 6 mo after surgery (28, 38). The changes in body cell mass, if any, are small; thus the effect of the surgical cure of acromegaly may be undetected in the short term. The mean change in absolute body weight in the subjects who were studied before and after surgery was a gain of 1.6 ± 1.9 kg, not significantly different from zero. We did not measure systematically the changes in lean body mass induced over time by adenomectomy in the acromegalic patients. In the two subjects who were studied by DEXA scan before and 6 mo after the operation, we found a modest increase in fat mass. Thus interpretation of the insulin resistance after surgery remains speculative. It should be noted that changes in body cell mass relative to total body weight could have affected the absolute rates of protein kinetics that were normalized by the total body weight. However, the changes in body cell mass, if any, were probably very small according to the literature (5, 28, 38) and the modest weight changes we measured in our subjects. Thus we cannot exclude the possibility that the absolute protein kinetic rates may have been affected by small changes in body composition in ACRO and in POST. However, the measurements of insulin sensitivity of proteolysis, leucine oxidation, and protein synthesis, the goals of this study, are not biased by changes in body composition, being normalized to the postabsorptive values.
Finally, we investigated possible changes in glutamine kinetics induced by acromegaly, because it was previously shown that growth hormone in stressed patients markedly reduced muscular glutamine production (4). The postabsorptive glutamine flux was similar among groups, but insulin significantly reduced the glutamine flux only in ACRO, where this insulin effect was significantly greater than in the pooled control groups. Interestingly, in ACRO, the suppression of glutamine flux was much greater than the suppression of leucine flux (−28 vs. −9%, P < 0.05). In contrast, in the other groups, the suppression of glutamine flux was similar to that of leucine flux [−5 vs. −12% in POST and −12 vs. −18% in CON,P = nonsignificant (NS)]. Circulating glutamine derives from two possible sources: proteolysis and de novo synthesis, i.e., the transamination of α-ketoglutarate and glutamate with other amino acids that are subsequently oxidized. Even though the amino acid composition of the proteins broken down in the study period was unknown and the assumption of constant ratios among amino acids in proteins can lead to experimental errors (33), the marked reduction in glutamine flux despite a defective insulin suppression of leucine flux strongly indicates a reduction in de novo glutamine synthesis. Our data suggest that, in ACRO, hyperinsulinemia simultaneously reduced amino acid oxidation and their transamination in the organs (the muscle) that export glutamine as a means to dispose amino nitrogen to the liver for urea synthesis. The suppression of glutamine de novo synthesis was not evident in POST subjects, who also had a defective suppression of leucine oxidation during the clamp. These data support the idea that glutamine de novo synthesis and amino acid oxidation are related processes and suggest that, in acromegaly, insulin is more effective in sparing amino acids from oxidation and in reducing nitrogen export to the liver via glutamine shuttling.
In conclusion, acromegalic patients are severely insulin resistant for both glucose and protein metabolism, indicating a generalized defect in insulin signaling. Six months after operation, insulin resistance for glucose metabolism is completely reversed. In contrast, a marked antagonism with the insulin effect on proteolysis and leucine oxidation still persists. The effects of growth hormone on protein metabolism are not reversed by surgery in the short term.
Address for reprint requests and other correspondence: A. Battezzati, Amino Acids and Stable Isotopes Laboratory, San Raffaele Scientific Institute, Via Olgettina, 60, 20132 Milano, Italy (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published October 8, 2002;10.1152/ajpendo.00020.2002
- Copyright © 2003 the American Physiological Society