| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Medical Department M (Endocrinology and Diabetes) and Medical Research Laboratories (T.K.H., C.H.G., H.O., J.S.C., J.O.L.J.), Aarhus University Hospital, DK-8000 Aarhus, Denmark; and Novo Nordisk A/S (M.H.R.), DK-2880 Bagsværd, Denmark
Address all correspondence and requests for reprints to: Troels Krarup Hansen, M.D., Medical Department M (Endocrinology and Diabetes), Aarhus University Hospital, Norrebrogade 42-44, DK-8000 Aarhus C, Denmark. E-mail: tkh{at}dadlnet.dk.
Abstract
The sensitivity to GH is subject to substantial interindividual variations, which has been attributed to differences in age, sex, and body composition. We investigated 18 healthy nonobese men (aged 24–56 yr) on four occasions. The pharmacokinetics and acute lipolytic effects of GH were evaluated using iv bolus injections of either placebo or GH (1, 3, and 6 µg/kg-1). Body composition was determined by computed tomography and bioimpedance measurements, and the lipolytic response was assessed through measurements of circulating lipid intermediates and adipose tissue microdialysis. The metabolic clearance rate was dose dependently reduced with increasing GH doses (57.2 ± 5.1, 45.2 ± 3.8, and 39.2 ± 2.4 ml/min-1 per meter-2 following injection of 1, 3, and 6 µg/kg-1 GH, respectively, P < 0.001), and half-time was increased (14.2 ± 0.6, 16.2 ± 0.4, and 18.0 ± 0.5 min, respectively, P < 0.0001). The pharmacokinetic variables were not correlated to age or body composition at any GH dose, but GH-binding protein was the major predictor of metabolic clearance rate following the two highest GH doses as indicated by multivariate regression analysis (r2 = 0.55, P < 0.001 and r2 = 0.35, P = 0.012, respectively). There was a significant dose-response relationship between injected GH and the subsequent increments in lipid intermediates, but the integrated lipolytic response was not correlated to GH pharmacokinetics, age, or body composition at any GH dose. Taken together, our findings suggest that differences in GH-binding protein concentrations, which possibly reflect GHR expression, determine GH pharmacokinetics rather than age or body composition per se.
ALTHOUGH THE PLEIOTROPIC actions of GH are well known, the physiological factors regulating the large interindividual differences in GH sensitivity remain less characterized. Evidence from a number of clinical studies suggest a complex control of the GH elimination kinetics and actions, including contributions from age, sex, body composition, and renal function as well as other hormones and binding proteins (1, 2, 3, 4, 5, 6, 7, 8, 9). Because many of these factors are closely interrelated, it has been difficult to reveal the exact cause-effect relationships in vivo. For a given dose of GH, the effective amount delivered to the tissue organ depends on a number of physiological mechanisms such as clearance, GH-binding protein (GHBP) binding, and tissue blood flow, and even when GH reaches the target organ, the subsequent effects may vary depending on differences in GH receptor (GHR) expression and postreceptor signaling (10). The main extrarenal clearance pathway for GH is attributable to receptor-mediated uptake (11, 12). but whether GH clearance in turn reflects the metabolic effects of GH remains unknown.
A major effect of GH is stimulation of lipolysis in adipose tissue. Traditionally the degree of lipolysis has been estimated through measurements of the resulting changes in serum lipid intermediates, i.e. free fatty acids (FFAs) and glycerol (13, 14) and whole-body palmitate flux (15). Recently, the development of a microdialysis technique has enabled assessment of changes in interstitial glycerol concentrations directly in sc adipose tissues (16). It has been shown that GH dose dependently increases circulating levels of FFAs and glycerol (17), but whether the lipolytic effects are differentially regulated from other metabolic actions of GH remains to be elucidated.
The current study was, thus, designed to investigate the dose effects of exogenous GH pulses within the physiological range with regard to GH pharmacokinetics and acute lipolytic actions.
Subjects and Methods
Subjects
Eighteen healthy males with a mean age of 39.4 ± 2.3 yr (range, 24–56 yr) were examined. The local ethical committee and the Danish National Board of Health approved the study, and informed oral and written consent was obtained from each subject before entering the study.
Study design
All subjects were examined on four occasions in random order separated by a minimum of 4 wk. After an overnight fast (>10 h) the subjects were admitted to the clinical research center at 0800 h and remained in the supine position throughout the experiments. Intravenous cannulas were inserted in antecubital veins on each arm for blood sampling and infusions. Following a baseline period of 60 min, an iv bolus of either GH (dissolved in 0.9% NaCl at a concentration of 20 µg/ml-1) or 0.9% NaCl was injected a t = 0 min. Three different doses of GH were used (1 µg/kg-1, 3 µg/kg-1, and 6 µg/kg-1), and treatment was blinded to the subjects. GH was dosed according to total body weight because we have found this to yield very similar serum GH concentrations in different subjects (1). A single batch of GH (Norditropin, Novo Nordisk, Copenhagen, Denmark) was used for all experiments. Blood samples were drawn with 30-min intervals during the baseline period, every 10 min for the first 2 h following the bolus infusions, and every 30 min thereafter.
Microdialysis
Microdialysis was performed on two of the four study days (placebo and 3 µg/kg-1 GH). Microdialysis fibers (CMA 60 microdialysis catheter, CMA, Stockholm, Sweden) with a molecular cut-off of 20 kDa were placed in abdominal and femoral sc tissue at the umbilical and midthigh level, respectively. The skin was anesthestized with 0.05 ml lidocaine at the site of perforation. After insertion, the catheters were perfused with physiological perfusion fluid [Perfusion fluid T1, CMA; 147 mM Na+, 4 mM K+, 2.3 mM Ca2+, and 156 mM Cl- (pH 6); osmolality, 290 mosmol/kg-1], using a portable pump (CMA 106, CMA) at a flow rate of 0.3 µl/min-1, which is known to yield a recovery of almost 100% (18). After calibration for 30 min, the sampling of dialysate started at t = -30 min and continued every 30 min until t = 330. The first sample, which was drawn at t = 0 min, thus reflects the interstitial level of glycerol during the preceding 30 min and was designated t = -15 min. This principle was used for all following samples. The adipose tissue blood flow in the abdominal region was estimated simultaneously from the disappearance curve of injected 133Xe as described previously (16).
Analyses
Serum total GH was determined by a double monoclonal immunofluorometric assay (DELFIA, Wallac, Inc., Turku, Finland) using a 24-h incubation period to reduce the interference from GHBP (19). Serum IGF-I was measured with an in-house time-resolved immunofluorometric assay as described previously (20), and insulin was determined by a commercial ELISA (DAKO Corp., Glostrup, Denmark). Blood 3-hydroxy-butyrate (BOH) and glycerol were assayed by automated fluorometric enzymatic methods (21). Serum nonesterified fatty acids (FFAs) were determined by a colorimetric method using a commercial kit (Wako Chemicals, Neuss, Germany). Serum IGF-binding protein (IGFBP)-1 was measured by ELISA (Medix Biochemica, Kainainen, Finland), IGFBP-3 by an immunoradiometric assay (Diagnostic System Laboratories, Webster, TX), and GHBP with an in-house time-resolved immunofluorometric assay (22). The concentration of glycerol in dialysate was measured by an automated spectrophotometric kinetic enzymatic method (CMA 600, CMA).
Body composition
Body mass index (BMI) was defined as the weight of the subject in kilograms divided by the square of the height in meters. Waist to hip (W/H) ratio was defined as the ratio between the circumference of the waist at the umbilical level and hips at the level of trochanter major in standing position. The same physician (T.K.H.) made all the measurements. Body composition was also measured using multifrequency bioelectrical impedance analysis (SFB3; SEAC, Brisbane, Australia), and the amounts of intra-abdominal and sc fat were evaluated by computerized tomography with a Somatom Plus-S scanner (Siemens, Erlangen, Holland). The subjects were studied in the supine position, and the areas scanned were 10-mm cross-sectional slices at the umbilical level. The same technician performed all the scans, which subsequently were evaluated by two radiologists. Total body fat and lean body mass were estimated using the formula of Deurenberg (23) and data from multifrequency bioelectrical impedance analysis.
GH pharmacokinetics
On each study day, the metabolic clearance rate (MCR) of GH was calculated from the equation (9): MCR = dose of GH injected/area under the time-concentration GH curve (AUC0–120 min GH). The elimination constant k was determined as the slope of the log-linear regression of the GH disappearance curve (t = 10 - t = 70 min), and half-time (t1/2) calculated as t1/2 = ln2/k. Vd was calculated as: Vd = MCR/k. Both MCR and Vd were corrected for surface area (24).
Lipolytic response
The whole-body lipolytic response to each GH infusion was estimated as the incremental areas under the time-concentration curves (iAUC0–300 min) of FFAs, BOH, and glycerol using the linear trapezoidal rule. The iAUCs of interstitial glycerol (iAUC15–285 min) were used to estimate the regional lipolytic response in abdominal and femoral adipose tissue.
Statistical methods
The Kolmogorov-Smirnov test was used to test for normality of data, and variables were log transformed when appropriate. Pearson’s product moment correlation with two-tailed probability values was used to measure the strength of association between the variables. Because MCR and Vd were corrected for surface area, these variables were dependent on BMI (kilograms per square meter), and hence no correlations were made between GH pharmacokinetics and BMI. Within-subject differences in the pharmacokinetic variables and lipolytic responses following different GH bolus doses were assessed by two-way ANOVA for repeated measurements, and paired t tests were used for post hoc comparisons. Data are given as the mean ± SE, and statistical significance was assumed for P less than 0.05. All statistical calculations were performed with SPSS for Windows (version 11.0; SPSS, Inc., Chicago, IL).
Results
Descriptive data of the subjects are summarized in Table 1
. Age was not correlated to body weight (r = 0.17, P = 0.51) or lean body mass (r = -0.06, P = 0.82), but nevertheless there were strong positive correlations between age and intra-abdominal fat (r = 0.77, P < 0.001), total body fat (r = 0.56, P = 0.02), BMI (r = 0.56, P = 0.02), and W/H ratio (r = 0.55, P = 0.02). In Table 2 baseline concentrations of circulating hormones, binding proteins, and metabolites and their intercorrelations are shown. Although there were no significant correlations between age and any of the serum measurements, IGF-1 concentrations tended to be negatively correlated with age (r = -0.44, P = 0.087). GHBP levels were positively correlated to the amount of sc fat (r = 0.57, P < 0.05) and tended to be positively correlated with intra-abdominal fat (r = 0.45, P = 0.07), but otherwise there were no correlations between pertinent measures of body composition and the serum measurements.
|
|
In Fig. 1 the GH concentration curves obtained following iv GH infusions at three different GH doses are depicted. The resulting measures of GH pharmacokinetics are summarized in Table 3. MCR was significantly reduced with increasing GH dose (Fig. 2), and t1/2 was increased, but Vd remained unaltered by changes in bolus size. The pharmacokinetic variables were not correlated to age or body composition at any of the GH doses used. Table 4 shows correlations of GH pharmacokinetics estimated following the three different GH boluses and concomitant baseline concentrations of hormones and binding proteins. There was a clear difference in the pattern of the correlations depending on the GH dose injected. It appeared that the strongest predictor of GH clearance, most noticeably after injection of 3 µg/kg-1 GH, was the baseline concentration of GHBP. IGFBP-I was negatively correlated to MCR, but baseline concentrations of IGF-I showed at positive correlation to the MCR calculated following injection of 3 µg/kg-1 GH. In a multivariate regression analysis controlling for age, body composition (W/H ratio, intra-abdominal fat area, total body fat), and baseline concentrations of IGF-I, insulin, and binding proteins, baseline GHBP was the strongest predictor of MCR following the two highest GH doses (Table 5).
|
|
|
|
|
Time serum concentration curves of the lipid intermediates FFAs, BOH, and glycerol following GH administration are depicted in Fig. 3. Regarding FFAs and BOH there were significant dose-response effects when comparing the iAUCs following 0, 1, and 3 µg/kg-1 GH, whereas there were no differences between the responses following 3 and 6 µg/kg-1 GH. With glycerol the iAUCs following 1, 3, and 6 µg/kg-1 GH were comparable, but all were significantly larger than following 0 µg/kg-1 GH. Irrespective of GH dose, there were no correlations between the observed changes in the lipid intermediates and age or any of the measures of body composition. The resulting increase in BOH following the highest GH dose was positively correlated to the peak GH concentration (r = 0.65, P < 0.005) and negatively correlated to MCR (r = -0.50, P < 0.05), but otherwise there were no significant correlations between the changes in serum lipid intermediates and GH pharmacokinetics.
|
The interstitial concentrations of glycerol were similar in abdominal and femoral adipose tissue at baseline and remained stable on the placebo days. Following injection of 3 µg/kg-1 GH, the glycerol concentration increased significantly in both tissues (Fig. 4A; ANOVA, P < 0.05 for both), with no difference between the two regions. The adipose tissue blood flow in the abdominal region was unchanged following GH administration, compared with placebo as estimated from the disappearance curve of sc-injected 133Xe. The integrated lipolytic response, assessed as the iAUC, was significantly higher following GH than after placebo in the abdominal region (P < 0.05) and tended to be so in the femoral region (P = 0.10) (Fig. 4B). The overall changes in interstitial glycerol following GH administration were not correlated to age, body composition, or any pharmacokinetic variables.
|
The present study was conducted to scrutinize the dose-response effects of different GH bolus injections with respect to GH pharmacokinetics and acute metabolic effects. We found that the clearance of a GH bolus was reduced with increasing doses, indicating a dose-dependent saturation of the clearance pathways. Likewise, our findings indicate that the acute lipolytic effects of GH are dose dependent but saturated following high physiological GH pulses. Interindividual variations in GH pharmacokinetics were not related to differences in body composition or age per se but were best predicted by baseline concentrations of GHBP.
Several publications have reported a decrease of MCR with increasing GH concentrations using constant infusions techniques (1, 3, 7, 8), but this study is the first to document the dose dependency of MCR using injections of physiologically relevant GH pulses. Pharmacokinetic studies in healthy controls and patients with chronic renal failure have shown that the renal clearance of GH is independent of serum GH levels, even at concentrations well above the physiological range (7). On the other hand, the extrarenal clearance, which is largely attributable to receptor-mediated uptake of GH (11, 12), reaches half-maximal saturation at a plasma GH concentration of approximately 12 µg/liter (7). We found that baseline GHBP concentrations best predicted MCR following physiological GH pulses as assessed by multivariate regression analysis.
Circulating GHBP is generally assumed to reflect tissue GHR status because GHBP is derived from GHRs by proteolytic cleavage (25, 26, 27), and GHBP concentrations have been found to predict both the growth response to GH therapy and the spontaneous growth rate in uremic children (28). Saturation of circulating GHBP significantly increases the elimination of GH (2), and the GHBP levels hence reflect two oppositely acting forces with regard to GH pharmacokinetics: the direct effect of GHBP, which will tend to reduce GH clearance, and the expression of GHRs, which in all probability is related to an increased clearance. Although one molecule of GH theoretically is capable of binding two molecules of GHBP, predominantly 1:1 complexes with GH are formed (29). The average concentration of GHBP in lean subjects is in the order of 1–1.5 nM, and hence circulating GHBP in a typical subject becomes saturated with 22 kDa GH at GH concentrations above 20–30 µg/liter (30). It, thus, appears that the extrarenal clearance mechanisms (i.e. the GHRs) may become saturated at lower GH concentrations than circulating GHBP causing a U-shaped relation between GH concentrations and MCR (Fig. 2
). In line with this, we in fact observed an increase in MCR when comparing the middle to the highest GH dose in three subjects, who had very low GHBP concentrations (data not shown).
The physiological conditions that determine the interindividual differences in GHR expression and GH sensitivity are not well known. Disregarding possible interindividual differences in GHR cleavage and turnover, one would expect the response to GH to be directly related to GHR expression but only when the receptors become the rate-limiting step, i.e. saturated to some extent. At low GH concentrations, when GHRs are in excess of free GH, a comparable response to a given amount of GH is expected in all subjects, whereas at higher GH concentration, the response becomes correlated to the expression of GHRs. We found that baseline IGF-I levels were significantly and positively correlated to MCR following the 3 µg/kg-1 GH bolus and tended to be so following 6 µg/kg-1 GH but not following the lowest dose of 1 µg/kg-1 GH. Furthermore, we found strong positive correlations between baseline GHBP concentrations and the MCRs following the two highest GH doses. Taken together these findings could suggest that the receptor-mediated clearance of GH is coupled to receptor signaling and hence that constitutively high GHR expression and GH sensitivity are associated with high GH clearance.
It has been documented that insulin increases the availability of surface GHR in liver cells (31). In poorly controlled insulin-dependent diabetes mellitus, characterized by very low serum GHBP levels (32), disproportionally low levels of IGF-I despite elevated GH levels imply a state of hepatic GH resistance (32, 33, 34), which is corrected by conventional insulin treatment (35), and, more efficiently, by intraperitoneal insulin administration (36). It therefore seems plausible that the GH resistance of insulin-dependent diabetes mellitus could be a consequence of low portal insulin concentrations with subsequent down-regulation of liver GHR expression, and in line with this, we recently reported a positive correlation between baseline insulin concentration and MCR of GH in a constant infusion study (1). In the present study, we found no correlation between fasting insulin levels and any of the pharmacokinetic measures. However, fasting insulin is not necessarily a pertinent measure of the integrated insulin release, and to fully evaluate the impact of insulin on GH pharmacokinetics, clearance studies during hyperinsulinemic clamp conditions are needed.
In previous studies using GH bolus injections (9), steady state GH infusions (8), or deconvolution analysis (4, 37), the elimination of GH has been reported to be positively correlated to age and measures of adiposity. In contrast, we found no correlations between the pharmacokinetic variables and age or body composition at any of the GH doses used. A possible explanation for this is the use of 24-h incubation in the GH assay to reduce the interference in the GH determinations from GHBP. We have recently found that estimates of GH from a widely used commercial immunometric GH assay (DELFIA, Wallac, Inc.) are significantly influenced by interindividual differences in serum GHBP concentrations if the recommended 2-h assay incubation time is used (Hansen, T. K., unpublished data, and Ref. 19). High GHBP levels cause erroneously low measures of GH if interference from GHBP in the assay used is not accounted for, and hence the estimates of MCR become spuriously high and the estimates of t1/2 too low in subjects with high serum GHBP concentrations. Because GHBP levels show strong positive correlations with the amount of adipose tissue (1, 38, 39, 40), which in turn is positively correlated to age (1, 9), any interference from GHBP in the assay will bias the correlations of both age and body composition with estimates of GH elimination.
GH was early recognized to be lipolytic (13), and the effects of a single bolus of GH has been investigated with regard to circulating and interstitial lipid intermediates (16, 41). We found a significant dose-response relationship between the injected doses of GH and the following increments in lipid intermediates. The relationship was less pronounced regarding glycerol, which could be attributable to increased hepatic reuptake of glycerol caused by the effects of GH on gluconeogenesis (42). The lipolytic response reached a plateau with doses above 3 µg/kg-1 (corresponding to an average peak GH concentration of 32.4 µg/liter), which probably corresponds to the upper limit of naturally occurring secretory bursts (43). To our surprise, the interindividual differences in the integrated lipolytic response, assessed as both changes in circulating lipid intermediates and at the adipose tissue level through microdialysis, appeared to be virtually independent of GH pharmacokinetics. A possible explanation to this finding could be that the lipolytic actions of GH primarily are attributable to a permissive effect on catecholamine induced lipolysis (44). In vitro studies have indicated that GH has no direct lipolytic effect on human fat cells but markedly increase the catecholamine sensitivity without any changes in maximal lipolysis (45). Although alterations in GH levels undoubtedly induce significant within-subject differences in lipolysis, as indicated by the clear dose-response relationship, the major determinant of subject-to-subject variations in the overall lipolytic response may therefore be differences in the catecholaminergic tone.
In summary, we have shown that GH pharmacokinetics and actions are not determined by age or body composition but rather by differences in circulating GHBP levels and probably GHR expression and functioning. Furthermore, our results indicate that the acute lipolytic effects of GH are dose dependent but independent of GH pharmacokinetics. The findings underscore the complexity of in vivo GH metabolism and point to the existence of further physiological control mechanisms of GH actions through tissue-specific variations in GHR signaling and clearance.
Acknowledgments
Novo Nordisk A/S, Denmark, generously supplied the GH.
Footnotes
This work was supported by grants from Danielsen’s Foundation, L. F. Foght’s Foundation, Fonden til Laegevidenskabens Fremme, Carl og Ellen Hertz’ Legat, Buhl Olesen Foundation, Aarhus University Research Foundation, and the Danish Medical Research Council, and by Grants 9903065 (to T.K.H.) and 960082 (to Novo Nordisk Center for Research in Growth and Regeneration, Aarhus University).
Abbreviations: AUC, Area under the time-concentration curve; BMI, body mass index; BOH, blood 3-hydroxy-butyrate; FFA, free fatty acid; GHBP, GH-binding protein; GHR, GH receptor; IGFBP, IGF-binding protein; MCR, metabolic clearance rate; t1/2, half-time; Vd, distribution volume; W/H, waist to hip.
Received April 18, 2002.
Accepted July 10, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  E. T. Vestergaard, C. B. Djurhuus, J. Gjedsted, S. Nielsen, N. Moller, J. J. Holst, J. O. L. Jorgensen, and O. Schmitz Acute Effects of Ghrelin Administration on Glucose and Lipid Metabolism J. Clin. Endocrinol. Metab., February 1, 2008; 93(2): 438 - 444. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Keller, Z. Wu, J. Kratzsch, E. Keller, W. F Blum, A. Kniess, R. Preiss, J. Teichert, C. J Strasburger, and M. Bidlingmaier Pharmacokinetics and pharmacodynamics of GH: dependence on route and dosage of administration Eur. J. Endocrinol., June 1, 2007; 156(6): 647 - 653. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Verheus, P. H.M. Peeters, R. Kaaks, P. A.H. van Noord, D. E. Grobbee, and C. H. van Gils Premenopausal Insulin-Like Growth Factor-I Serum Levels and Changes in Breast Density over Menopause Cancer Epidemiol. Biomarkers Prev., March 1, 2007; 16(3): 451 - 457. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Svensson, G. Johannsson, A. Iranmanesh, K. Albertsson-Wikland, J. D Veldhuis, and B.-A. Bengtsson GH secretory pattern in young adults who discontinued GH treatment for GH deficiency and decreased longitudinal growth in childhood. Eur. J. Endocrinol., July 1, 2006; 155(1): 91 - 99. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. van Dijk, P. Mulder, M. Houdijk, J. Mulder, K. Noordam, R. J. Odink, C. Rongen-Westerlaken, P. Voorhoeve, J. Waelkens, J. Stokvis-Brantsma, et al. High Serum Levels of Growth Hormone (GH) and Insulin-Like Growth Factor-I (IGF-I) during High-Dose GH Treatment in Short Children Born Small for Gestational Age J. Clin. Endocrinol. Metab., April 1, 2006; 91(4): 1390 - 1396. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Fuglsang, P. Sandager, N. Moller, S. Fisker, H. Orskov, and P. Ovesen Kinetics and secretion of placental growth hormone around parturition. Eur. J. Endocrinol., March 1, 2006; 154(3): 449 - 457. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. B. Djurhuus, C. H. Gravholt, S. Nielsen, S. B. Pedersen, N. Moller, and O. Schmitz Additive effects of cortisol and growth hormone on regional and systemic lipolysis in humans Am J Physiol Endocrinol Metab, March 1, 2004; 286(3): E488 - E494. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) or GH at doses of 1 µg/kg-1 (•), 3 µg/kg-1 (
), and 6 µg/kg-1 (
). Right, The integrated response was assessed as iAUC0–300 min. Data are means ± SE, n = 18. *P < 0.05 and **P < 0.01, compared with placebo; 
, P < 0.05, compared with 1 µg/kg-1 GH.