Growth Hormone Signaling in Vivo in Human Muscle and Adipose

ORIGINAL E n d o c r i n e

ARTICLE R e s e a r c h

Growth Hormone Signaling in Vivo in Human Muscle and Adipose Tissue: Impact of Insulin, Substrate Background, and Growth Hormone Receptor Blockade Charlotte Nielsen, Lars C. Gormsen, Niels Jessen, Steen Bønløkke Pedersen, Niels Møller, Sten Lund, and Jens Otto L. Jørgensen Medical Departments M (Endocrinology and Diabetes) (C.N., L.C.G., N.J., N.M., S.L., J.O.L.J.) and C (S.B.P.), Aarhus University Hospital, DK-8000 Aarhus C, Denmark

Context: GH induces insulin resistance in muscle and fat, and in vitro data indicate that this may involve cross-talk between the signaling pathways of the two hormones. Objective: Our objective was to investigate GH and insulin signaling in vivo in human muscle and fat tissue in response to GH, GH receptor blockade, and insulin stimulation. Design: We conducted two randomized crossover studies. Participants: Sixteen healthy males participated. Intervention: GH was administered as a bolus (n ⫽ 8) and constant infusion (n ⫽ 8). The bolus study included three arms: 1) control (saline), 2) GH (0.5 mg iv), and 3) GH blockade (pegvisomant 30 mg sc), each combined with a hyperinsulinemic glucose clamp. The infusion study included two arms: 1) GH infusion (45 ng/䡠kg䡠min, 5.5 h) and 2) saline infusion (5.5 h) combined with a hyperinsulinemic glucose clamp during the final 2.5 h. Main Outcome Measures: Muscle and fat biopsies were subjected to Western blotting for expression of Stat5/p-Stat5, Akt/p-Akt, and ERK1/2/p-ERK1/2 and to real-time RT-PCR for expression of SOCS1–3 and IGF-I mRNA. Results: GH significantly reduced insulin sensitivity. The GH bolus as well as GH infusion induced phosphorylation of Stat5 in muscle and fat, and SOCS3 and IGF-I mRNA expression increased after GH infusion. Hyperinsulinemia induced Akt phosphorylation in both tissues, irrespective of GH status. In muscle, ERK1/2 phosphorylation was increased by insulin, but insulin per se did not induce phosphorylation of Stat5. Conclusions: GH exposure associated with insulin resistance acutely translates into GH receptor signaling in human muscle and fat without evidence of cross-talk with insulin signaling pathways. The molecular mechanisms subserving GH-induced insulin resistance in humans remain unclarified. (J Clin Endocrinol Metab 93: 2842–2850, 2008)

G

H is not only the primary promoter of postnatal linear growth but also an important regulator of substrate metabolism. The metabolic effects are complex and depend in part on substrate availability. In the postabsorptive phase and during fasting, GH promotes lipolysis and lipid oxidation at the expense

of glucose oxidation (1, 2). If these effects operate postprandially, which is the case in active acromegaly and during sc GH administration, insulin resistance, glucose intolerance, and manifest diabetes mellitus may develop (3–5). The molecular mechanisms by which GH promotes insulin

0021-972X/08/$15.00/0

Abbreviations: AU, Arbitrary units; FFA, free fatty acids; GHR, GH receptor; GIR, glucose infusion rate; IRS, insulin receptor substrate; JAK2, Janus kinase 2; PI, phosphatidylinositol; SOCS3, suppressor of cytokine signaling 3; Stat5, signal transducer and activator of transcription 5.

Printed in U.S.A. Copyright © 2008 by The Endocrine Society doi: 10.1210/jc.2007-2414 Received October 30, 2007. Accepted April 25, 2008. First Published Online May 6, 2008

2842

jcem.endojournals.org

J Clin Endocrinol Metab. July 2008, 93(7):2842–2850

J Clin Endocrinol Metab, July 2008, 93(7):2842–2850

antagonism are still unclear. The stimulated lipolysis could be of importance because free fatty acids (FFA) have been shown to interfere with insulin receptor signaling via inhibition of insulinstimulated insulin receptor substrate (IRS)-1-associated phosphatidylinositol (PI) 3-kinase activity in human skeletal muscle, resulting in decreased glucose transporter 4 (GLUT4) translocation and glucose uptake (6). A recent study, however, was unable to document a suppression of the insulin-stimulated activity of either IRS-1-associated PI 3-kinase or the serine/threonine kinase Akt after GH administration to healthy humans, despite induction of lipolysis and insulin resistance (7). Other studies have shown that acute GH exposure induces insulin resistance in skeletal muscle rapidly and before the subsequent rise in plasma FFA (1, 7, 8). These observations indicate that GH may cause insulin resistance via a non-FFA-mediated mechanism. The predominant GH signal transduction cascade comprises activation of the GH receptor (GHR) dimer, phosphorylation of Janus kinase 2 (JAK2), and subsequently activation of signal transducer and activator of transcription 5 (Stat5). The intact JAK2/Stat5 pathway is critical for normal statural growth in children (9). There is animal and in vitro evidence to suggest that insulin and GH share postreceptor signaling pathways (10). Convergence has been reported at the levels of Stat5 and suppressor of cytokine signaling 3 (SOCS3) as well as on protein kinases comprising the major IR signaling pathway: IRS1/2, PI 3-kinase, Akt, and ERK1/2 (11–14). Studies in rodent models suggest that the insulin-antagonistic effects of GH in adipose and skeletal muscle tissue are PI 3-kinase dependent through direct up-regulation of the p85␣ subunit and subsequent decrease in insulin-stimulated PI 3-kinase activity (10, 15). Pegvisomant is a GH analog and a competitive reversible GHR antagonist (16). Theoretically, blockade with pegvisomant could therefore serve as a negative control in GH signaling studies. The aim of this work was to study GH signal transduction pathways in vivo in muscle and adipose tissue from healthy subjects in response to acute and more prolonged GH exposure as well as during hyperinsulinemia. The design also included administration of pegvisomant in an attempt to block the effects of spontaneous GH secretion during control experiments.

Subjects and Methods Results are reported from two different protocols involving two separate groups of participants. The protocols were approved by the Regional Ethical Committee system, which included written and oral consent from each participant.

Subjects Bolus study Eight healthy men aged 23.0 ⫾ 2 yr (mean ⫾ SEM) and with mean body mass index of 24.6 ⫾ 3.1 kg/m2 were examined. Total fat mass measured by dual x-ray absorptiometry constituted 14 ⫾ 3% of total body weight. Fasting blood glucose levels were normal in all participants, none of whom received medication. Data from this study on serum ghrelin levels have previously been published (17).


2843

Infusion study Eight healthy men aged 26.0 ⫾ 0.8 yr and with a mean body mass index of 24.1 ⫾ 0.5 kg/m2 participated. Data on substrate metabolism, insulin sensitivity, and insulin signaling from the same study have previously been published (7).

Study design Bolus study During hyperinsulinemic-euglycemic clamp conditions, each participant underwent three study arms on separate days and in random order: 1) saline infusion (0.9% NaCl, 50 ml/h) for 8 h (control), 2) GH (Genotropin; Pfizer, Inc.) bolus (0.5 mg iv), and 3) pegvisomant (Somavert; Pfizer Inc., New York, NY) 30 mg sc. At least 4 wk elapsed between each study, which started at 0800 h after an overnight fast with the participants remaining supine. At 0630 h, two iv catheters (Venflon; Viggo AB, Helsingborg, Sweden) were inserted, one in an antecubital vein of the left arm and the other in a dorsal hand vein on the right. The right hand was placed in a heated box at 65 C allowing arterialized blood samples to be drawn. After baseline blood sampling at 0800 h (t ⫽ 0 min), a constant insulin infusion (Actrapid; Novo Nordisk, Copenhagen, Denmark; 0.5 mU/kg䡠min) was commenced, and plasma glucose was clamped at 5.0 mmol/liter by infusion of adjusted rates of a 20% glucose solution based on 10-min measurements of plasma glucose. Saline infusion and the GH bolus were initiated at t ⫽ 0 min, whereas pegvisomant was administered 36 h before the study start (t ⫽ ⫺36 h) to achieve peak concentrations of pegvisomant during the study. Blood samples were taken every 30 min during the study day (Fig. 1). On all three study days, a baseline (t ⫽ 0 min) muscle biopsy from vastus lateralis was obtained. Under local anesthesia (1% lidocaine) a small incision through the skin and muscle sheath was made about 10 cm above the knee joint. With a Bergstro¨m biopsy needle, approximately 100 –130 mg skeletal muscle tissue was obtained. The biopsies were cleaned for blood and snap-frozen in liquid nitrogen. At t ⫽ 120 min, a second muscle biopsy was obtained approximately 5 cm proximal of the primary incision. In addition, fat biopsies at baseline and after 2 h were taken in the GH and pegvisomant arms, respectively. By means of liposuction and during local anesthesia, about 200 mg abdominal sc fat from the periumbilical region was obtained. The tissue was washed with isotonic saline and snap-frozen in liquid nitrogen. All biopsies were stored at ⫺80 C until analysis.

Infusion study The design has been described in detail previously (7). In brief, each subject was examined on two occasions receiving an iv infusion of either GH (45 ng/kg䡠min) or saline for 5.5 h. Before intervention, a baseline muscle biopsy (⬃200 mg) from vastus lateralis was obtained. A biopsy from the abdominal sc fat (⬃500 mg) was also obtained by a modified liposuction technique after 3 h of GH or saline infusion immediately subsequent to which a hyperinsulinemic (insulin infusion rate 1.2 mU/ kg䡠min) euglycemic (plasma glucose ⬃5.0 mmol/liter) clamp was commenced and continued for the last 2.5 h. A second muscle biopsy was obtained after 4 h (Fig. 1).

Analyses Plasma glucose level was measured in duplicate immediately after sampling on a glucose analyzer (Beckman Instruments, Palo Alto, CA). Serum insulin was determined by a commercial immunological kit (Dako, Glostrup, Denmark) and levels of serum FFA were determined using a commercial kit (Wako Chemicals, Neuss, Germany). A double monoclonal immunofluorometric assay (DELFIA; Perkin-Elmer, Wallac, Tu¨rku, Finland) was used to measure serum GH; the original assay was modified slightly to allow GH measurements during pegvisomant exposure (18). Serum pegvisomant was measured using a two-phase competitive RIA as previously published (18). Serum IGF-I was measured after acidic ethanol extraction of the IGF-binding proteins in an in-house IGF-I assay. The IGF-I assay is a monoclonal, time-resolved imunofluorometric assay based on an antibody generously provided by

2844

Nielsen et al.

GH Signaling in Vivo in Muscle and Adipose Tissue


FIG. 1. Schematic of the bolus (left panel) and the infusion (right panel) study protocols. Time is presented as minutes from the time of GH bolus injection (at 0800 h) or initiation of the continuous GH infusion (at 0800 h). In the bolus study, baseline muscle and adipose tissue biopsies were performed at 0730 h, and stimulated biopsies were obtained 120 min after the GH bolus. In the infusion study, a baseline muscle biopsy was performed at 0730 h, and a stimulated biopsy was obtained 240 min after initiation of the GH infusion and 60 min after initiation the hyperinsulinemic clamp. A single adipose tissue biopsy was obtained 180 min after initiation of the GH infusion (before the hyperinsulinemic clamp).

Novo Nordisk (coating) and a commercial antibody from DSL Inc. (Webster, TX), which before use had been labeled with europium in accordance with the manufacturer (PerkinElmer LifeSciences, Turku, Finland). The use of europium enhances the signal to noise ratio compared with for instance ELISAs and immunoradiometric assays and results in a very high sensitivity (⬍0.010 ␮g/liter). The high assay sensitivity allows all samples to be assayed in one final dilution of 1:1000, and it also reduces the influence of any remnant IGF-binding proteins in the acidethanol-extracted sera. The recovery of IGF-I in serum extracts is about 90%, and IGF-II cross-reactivity is negligible. Intraassay and interassay coefficients of variation are 5 and 10%, respectively (19).

Western blotting for Stat5, Akt/protein kinase B, and ERK1/2 Muscle and fat tissue samples were homogenized in lysis buffer [20 mM Tris (pH 7.0), 0.27 M sucrose, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 1% wt/vol Triton X-100, 5 mM NaPPi, 10 mM glycerol phosphate, 1 mM benzamidine, 4 ␮g/ml leupeptin, 1 mM dithiothreitol, and 200 mM NaB3BVOB4B] with a polytron blender. The homogenates were rotated, and insoluble materials were removed by centrifugation at 16,000 ⫻ g for 20 min at 4 C. Aliquots of protein were heated for 5 min at 95 C and then subjected to SDS-PAGE using the Bio-Rad Mini Protean II system (8% polyacrylamide), and proteins were electroblotted onto nitrocellulose membranes. Membranes were blocked, incubated overnight at 4 C with primary antibodies, and subsequently incubated with secondary horseradish peroxidase-conjugated goat antirabbit IgG antibody (Biosource, Camarillo, CA). Signals were detected using ECL (Pierce Supersignal) and quantified with UVP BioImaging System. Phosphoprotein content was corrected to the total protein content (ratio p/T). Quantification of ERK1/2 was performed separately for p42 and p44. The following primary antibodies used: phospho-Stat5a and -b antibody (Tyr694), Stat5a and -b antibody, phospho-Akt1–3 antibody (Ser473), Akt1–3 antibody (Cell Signaling, Beverly, MA), phospho-ERK1/2 antibody (pTpYP185/187P), and ERK1/2 antibody (Biosource). For each study occasion and for each subject, the ratio between stimulated and nonstimulated protein content was calculated, and subsequently the mean value ⫾ SE of the ratios was found and expressed as the incremental protein phosphorylation.

was extracted following the manufacturer’s protocol. RNA was quantified spectrophotometrically, and the integrity of the RNA was ensured by visual inspection of the two rRNAs on an ethidium bromide-stained agarose gel. RNA RT was performed using random hexamer primers at 23 C for 10 min and 42 C for 60 min and was terminated by increasing the temperature to 95 C for 10 min, as described by the manufacturer (GeneAmp RNA PCR Kit from Perkin-Elmer Cetus, Norwalk, CT). Real-time PCR analysis was performed to analyze the levels of SOCS1, SOCS2, SOCS3, and IGF-I mRNA. As internal control, gene levels of ␤-actin mRNA were measured. cDNA was subjected to PCR using MasterMix containing the specific primers, Taq DNA polymerase (HotStar Taq; QIAGEN, Inc., Valencia, CA) and SYBR-Green. Real-time quantification of target gene (X0) to ␤-actin (R0) mRNA was conducted with the SYBR-Green real-time PCR assay (QIAGEN) using an ICycler from Bio-Rad Laboratories (Hercules, CA). The X0 and R0 mRNA were amplified in separate tubes. The increase in fluorescence was measured in real time during the extension step. The threshold cycle (Ct) and the relative gene expression were calculated essentially as described in the User Bulletin no. 2, 1997, from Perkin-Elmer. All samples were amplified in duplicate.

Statistical analysis All parametric data are presented as mean ⫾ SE. Nonparametric data are presented as medians ⫾ interquartile range. Statistical analysis was performed using SigmaStat for Windows (version 3.1.1; SYSTAT, Richmond, CA). Normality of the data was tested with the Kolmogorov-Smirnov test of normal distribution. For time series, the area under the curve was calculated by the trapezoidal method, and comparisons were made by ANOVA.

Bolus study Statistical comparisons between study days were assessed by repeated-measurements ANOVA. If this test was positive, post hoc comparison was performed by means of a paired t test. Nonparametric data were either log-transformed to yield normal distribution or analyzed by Wilcoxon rank sum test.

Infusion study Real-time RT-PCR Muscle and adipose tissues were homogenized in TRIzol reagent (Invitrogen Life Technologies, Inc., Roskilde, Denmark) and total RNA

For statistical evaluation of differences of normally distributed data, Student’s paired t test was used. A two-tailed P value ⬍ 0.05 was considered statistically significant.



2845

20 (saline) vs. 238 ⫾ 21 (GH) vs. 204 ⫾ 17 (pegvisomant) (P ⫽ 0.05)]. Blood glucose levels were stable and comparable at baseline (⬃5 mmol/liter, ANOVA P ⫽ 0.80) and during the clamp on all study days (⬃5 mmol/liter, P ⫽ 0.41). Serum insulin levels were clamped at about 200 pmol/liter on all study days (ANOVA P ⫽ 0.7, Fig. 2). GH administration significantly reduced the overall glucose infusion rate (GIR) when assessed by repeated-measurements ANOVA with time vs. treatment as the parameter of interest (ANOVA P ⬍ 0.05). Post hoc analysis revealed that GH reduced GIR by about 20% from t ⫽ 210 to t ⫽ 250 min, after which the GIR returned to control levels. Pegvisomant administration did not significantly impact GIR when compared with saline.

FIG. 2. Bolus study. Circulating levels of insulin, plasma glucose, GH, FFA, and pegvisomant during saline infusion (F) and GH (䡺) and pegvisomant (Œ) bolus. Data are presented as means ⫾ SE; n ⫽ 8.

Stat5 phosphorylation Exogenous GH signal transduction. In muscle tissue, Stat5 tyrosine phosphorylation was induced in response to GH administration, corresponding to a significant 3-fold rise compared with baseline [0.56 ⫾ 0.13 vs. 1.65 ⫾ 0.04 arbitrary units (AU), P ⫽ 0.01] (Fig. 4, upper left panel). An increase in phosphorylation of Stat5 in adipose tissue was detected in five of eight subjects after GH injection (0.25 ⫾ 0.05 vs. 1.75 ⫾ 1.25 AU, P ⫽ 0.02) (Fig. 4, upper right panel).

Results Bolus study Circulating hormones, metabolites, and metabolic parameters The GH bolus increased serum GH to a peak level of 48.0 ⫾ 4.5 ␮g/liter after 10 min, followed by a log-linear decline (Fig. 2). In the saline experiment, four of eight subjects exhibited spontaneous GH bursts between t ⫽ 0 min and the second muscle biopsy at t ⫽ 120 min [GH peak: subject 1, 3.9 ␮g/liter (t ⫽ 90 min); subject 2, 1.9 ␮g/liter (t ⫽ 75 min); subject 3, 8.3 ␮g/liter (t ⫽ 90 min); and subject 7, 1.7 ␮g/liter (t ⫽ 90 min)] (Fig. 3). A GH burst was also recorded in two subjects on the day of pegvisomant [GH peak: subject 3, 4.6 ␮g/liter (t ⫽ 0 min), and subject 7, 4.7 ␮g/liter (t ⫽ 60 min)]. Baseline serum GH levels were comparable on all study days (data not shown). Serum FFA levels were suppressed by about 70% after 90 min of hyperinsulinemia in all three experiments (ANOVA P ⬍ 0.01, Fig. 2). The pegvisomant injection (t ⫽ ⫺36 h) increased serum pegvisomant to a plateau level throughout the study day (1773.5 ⫾ 20.0 ␮g/liter). The associated 36-h GHR blockade tended to decrease serum IGF-I levels (t ⫽ 240 min) [IGF-I (␮g/liter): 227 ⫾

Endogenous GH signal transduction. The spontaneous GH bursts observed during saline infusion in four of eight participants were associated with tyrosine phosphorylation of Stat5 in muscle tissue in all four subjects (Fig. 3). More surprisingly, the GH burst observed in subject 7 during pegvisomant exposure was also associated with Stat5 phosphorylation in muscle tissue. Akt and ERK1/2 phosphorylation Insulin stimulation induced a significant approximately 2-fold increase in Akt phosphorylation in muscle tissue on all study days compared with baseline [saline: 0.92 ⫾ 0.19 (baseline) vs. 1.69 ⫾ 0.30 (clamp), P ⫽ 0.01; GH: 0.92 ⫾ 0.11 (baseline) vs. 2.16 ⫾ 0.35 (clamp), P ⬍ 0.01; pegvisomant: 1.28 ⫾ 0.18 (baseline) vs. 2.64 ⫾ 0.38 (clamp) (AU), P ⫽ 0.01]. Thus, GH administration did not suppress insulin-stimulated Akt phosphorylation (Fig. 4, lower left panel). Phosphorylation of ERK2 (p42) tended to increase during insulin stimulation in all three studies without reaching statistical significance (Fig. 5, upper left panel). In muscle tissue, ERK1 (p44) phosphorylation was, as expected, significantly increased ⬃10-fold after 2 h of hyperinsulinemia during saline conditions [saline: 0.39 ⫾ 0.16 (base-

2846

Nielsen et al.



Insulin-stimulated adipose tissue ERK2 (p44) was unaltered, although a tendency to increase was observed after GH [GH: 0.63 ⫾ 0.30 (baseline) vs. 3.88 ⫾ 1.61 (clamp), P ⫽ 0.06; pegvisomant: 1.96 ⫾ 0.75 (baseline) vs. 2.65 ⫾ 0.77 (clamp) (AU), P ⫽ 0.54] (Fig. 5, upper right panel). Adipose tissue ERK1 (p42) was not increased by hyperinsulinemia on either study day (Fig. 5, lower right panel). Infusion study Circulating hormones, substrates, and insulin sensitivity Data on circulating levels of glucose, insulin, and FFA from this protocol have previously been published (7). During GH infusion, serum GH reached a plateau of 20.2 ⫾ 2.3 ␮g/liter during the final 30 min compared with 0.8 ⫾ 0.4 ␮g/liter during saline infusion (P ⫽ 0.01). The GIR (mg/ kg䡠min) was markedly suppressed during GH infusion, both from 230 –240 min [4.7 ⫾ 0.5 (GH) vs. 7.7 ⫾ 1.1 (saline), P ⬍ 0.01] and for the last 30 min of the clamp [5.0 ⫾ 0.9 (GH) vs. 8.1 ⫾ 0.9 (saline), P ⬍ 0.01]. Stat5 phosphorylation Muscle tissue samples were obtained from six subjects. Stat5 phosphorylation was significantly approximately 2.5-fold increased after GH infusion compared with baseline, whereas no difference was observed during saline infusion [saline: 1.87 ⫾ 0.26 (baseline) vs. 1.71 ⫾ 0.27 (t ⫽ 240 min), P ⫽ 0.47; GH: 3.39 ⫾ 2.01 (baseline) FIG. 3. Bolus study. Left panel, Circulating endogenous GH levels in subjects 1, 2, 3, and 7 during control conditions with four individual spontaneous GH bursts between the basal at t ⫽ 0 min vs. 8.63 ⫾ 1.10 (t ⫽ 240 min) (AU), P ⬍ and the stimulated muscle biopsy at t ⫽ 120 min. Immunoblots show the effects of these 0.01] (Fig. 6, upper left panel). spontaneous GH bursts on tyrosine phosphorylation of Stat5 in subjects 1, 2, 3, and 7 between the In adipose tissue obtained immediately basal and the stimulated muscle biopsy during saline infusion. Right panel, Immunoblots from subjects 4, 5, 6, and 8 where no endogenous GH bursts were recorded before the biopsies. before the hyperinsulinemic clamp, GH infusion resulted in an approximately 13-fold greater phosphorylation of Stat5 compared line) vs. 3.66 ⫾ 1.30 (clamp) (AU), P ⬍ 0.05]. No significant with saline [saline: 1.00 ⫾ 0.25 vs. GH: 12.61 ⫾ 3.85 (AU), P ⫽ increase in ERK1 phosphorylation was observed after GH or 0.03] (Fig. 6, upper right panel). pegvisomant administration [GH: 0.42 ⫾ 0.13 (baseline) vs. 2.53 ⫾ 1.12 (clamp), P ⫽ 0.08; pegvisomant: 2.27 ⫾ 1.18 (baseline) vs. 1.72 ⫾ 0.61 (clamp) (AU), P ⫽ 0.70] (Fig. 5, lower left panel). In adipose tissue, insulin induced Akt phosphorylation in six of eight persons after both GH and pegvisomant administration [GH: 1.12 ⫾ 0.21 (baseline) vs. 5.12 ⫾ 1.37 (clamp), P ⫽ 0.02; pegvisomant: 3.92 ⫾ 1.92 (baseline) vs. 7.75 ⫾ 1.82 (clamp) (AU), P ⫽ 0.07]; however, the increase was significant only after GH administration.

IGF-I and SOCS1–3 mRNA expression In muscle tissue, GH significantly increased SOCS3 [saline: 1.03 ⫾ 0.41 vs. 4.45 ⫾ 4.02, P ⫽ 0.44; GH: 0.43 ⫾ 0.14 vs. 11.45 ⫾ 6.4 (AU), P ⫽ 0.02] and IGF-I [saline: 0.28 ⫾ 0.04 vs. 0.44 ⫾ 0.12; GH: 0.29 ⫾ 0.02 vs. 7.23 ⫾ 4.86 (AU), P ⫽ 0.02] mRNA expression compared with baseline, whereas the difference seen after saline infusion was insignificant (Fig. 6, middle left panel). No increase in the expression of either SOCS1 or -2



2847

and pathophysiology, but it may also have implications for more prevalent conditions associated with insulin resistance. Data obtained mainly in vitro or in rodent models suggest that cross-talk or interference between GH and insulin signaling may play an important role, but data in human models are sparse. In two different experiments, we observed that iv administration of GH was accompanied by GHR signaling in skeletal muscle and fat in healthy human subjects. This was detected as tyrosine phosphorylation of the transcription factor Stat5 and expression of IGF-I and SOCS3 mRNA. This effect seemed to be independent of insulin infusion, because GH administration alone, before the clamp, also induced GHR signaling in adipose tissue. Moreover, insulin per se was not associated with Stat5 phosphorylation. These results extend and support the one previous human in vivo experiment showing Stat5 phosphorylation in muscle and fat shortly after GH administration in the postabsorptive state (20). Our design did not allow a direct assessment of signaling strength as a function of mode of GH administration or timing of tissue sampling, although it seemed that GH infusion for 3 h without concomitant insulin stimulation resulted in a more potent Stat5 phosphorylation as seen in the fat biopsies from the infusion study. For the first time, specific GHR blockFIG. 4. Bolus study: effect of exogenous GH and pegvisomant (PEG) on tyrosine phosphorylation of Stat5 (top) and serine phosphorylation of Akt/protein kinase B (bottom) in human muscle (left) ade was used in a human GHR signaling and adipose (right) tissue. A baseline biopsy at t ⫽ 0 min and a stimulated muscle biopsy at t ⫽ experiment. We hypothesized that pegvi120 min was obtained on all three study occasions, whereas fat biopsies were obtained on the somant would serve as a negative control days of GH and pegvisomant administration. Representative immunoblots for each protein are shown. Data are presented as means ⫾ SEM; n ⫽ 8. *, P ⬍ 0.05 during stimulated vs. baseline to GH by obviating any impact of sponconditions. taneous GH secretion during the study. We did, however, in one of two subjects, who also presented evidence of endogenous GH secretion durmRNA was observed in relation to GH exposure (data not ing the pegvisomant study, observe distinct Stat5 phosphorshown). ylation. Serum pegvisomant levels, which were measured beIn adipose tissue, GH compared with saline increased SOCS3 tween 36 and 44 h after the injection, were approximately mRNA expression [saline: 2.06 ⫾ 0.73 vs. GH: 4.68 ⫾ 1.64 1700 ␮g/liter in the subject who exhibited Stat5 phosphory(AU), P ⫽ 0.04] (Fig. 6, middle right panel), whereas no difference between GH and saline was detected regarding the expreslation. Such a level is sufficient to suppress IGF-I production sion of either SOCS1 or SOCS2 mRNA. IGF-I mRNA expression in the majority of patients with acromegaly, but it is well tended to increase by GH compared with saline, although it recognized that some patients require higher doses (16, 21, failed to reach statistical significance (P ⫽ 0.07, Fig. 6). 22). Pegvisomant administration tended to decrease serum IGF-I levels, but it is possible that a single injection of 30 mg is not sufficient to obtain complete GHR blockade in all subjects at the particular time point where the biopsy was obDiscussion tained. Another possible explanation could be Stat5 tyrosine The insulin-antagonistic effects of GH are well recognized and phosphorylation induced by other class 1 cytokines in the highly reproducible, but the underlying mechanisms remain uncirculation, i.e. prolactin, interferon, IL-2, erythropoietin, clear. To obtain more knowledge about this phenomenon is of epidermal growth factor, platelet-derived growth factor, or importance not only for our understanding of GH physiology leptin (23, 24). The antibody used in the present Western blot

2848

Nielsen et al.



IGF-I gene promoter region capable of binding Stat5 and mediating gene expression in a GH-dependent manner (29). Our studies indirectly support that similar mechanisms may operate in human skeletal and adipose tissue. We also observed that a GH bolus and a 5.5-h GH infusion, corresponding to normal and twice the normal daily GH production, respectively, caused insulin resistance in skeletal muscle. As previously published, we were unable to detect any changes in insulin-stimulated insulin signaling events in skeletal muscle tissue in terms of PI 3-kinase and Akt activity in the infusion study (7). Moreover, in the bolus study, we observed a significant increase in Akt phosphorylation during the hyperinsulinemic clamp, irrespective of coadministration of either GH or pegvisomant. Furthermore, in a recent study from our laboratory performed after an overnight fast with low levels of insulin activity, we did not detect any effect of a 0.5-mg GH bolus on PI 3-kinase activity and Akt phosphorylation in muscle and adipose tissue 30 and 60 min after GH exposure (20). It remains, however, possible that GH could have affected the IRS-1-associated PI 3-kinase activity either before or after the time point where the biopsies were taken. A larger time series of biopsies in each individual would have been appropriate from a scientific point of view, but such an apFIG. 5. Bolus study: effect of exogenous GH and pegvisomant (PEG) on tyrosine and threonine proach poses technical problems. Alternaphosphorylation of ERK2/p42 (top) and ERK1/p44 (bottom) in human muscle (left) and adipose tively, GH could affect insulin signaling at a (right) tissue. A baseline and a stimulated muscle biopsy were obtained on all three study level downstream of Akt. occasions, whereas fat biopsies were obtained on the days of GH and pegvisomant administration. A representative immunoblot from muscle and adipose tissue is shown. Data are Numerous animal and in vitro studies presented as means ⫾ SEM; n ⫽ 8. *, P ⬍ 0.05 during stimulated vs. baseline conditions. have indicated that GH and insulin may share and activate signaling pathways in terms of the Ras-Raf-MAPK cascade, which results in phosphorylation of procedure binds both phospho-Stat5a and -b. It could also be the MAPK ERK1/2 (10). In the present study, we detected a signifspeculated that pegvisomant per se may induce GHR signaling icant increase in ERK1/2 phosphorylation in association with hyin rare cases. The serum IGF-I level in the particular subject perinsulinemia on the days of saline infusion. Concomitant insulin who presented a spontaneous GH burst and subsequent Stat5 and GH exposure was associated with only a nonsignificant rise in phosphorylation exhibited a slight decrease after pegvisomant ERK1/2 phosphorylation in four of eight subjects. If anything, this administration (data not shown), and there is no evidence in observation suggests that GH may antagonize the effect of insulin the literature of a paradoxical increase in IGF-I or other GH on ERK1/2 activation, and we have previously been unable to detect biomarkers after long-term pegvisomant treatment in any independent effects of GH on this signaling pathway in human acromegaly. tissues (20). GH is a potent stimulator of IGF-I synthesis in the liver, which Other studies have shown that insulin per se induces tyrosine constitutes the major source of circulating IGF-I. Several animal phosphorylation and activation of Stat5 in insulin-sensitive tarand in vitro studies, however, have also detected GHR signaling get tissue both in vitro in cell cultures and in rodent liver, adipose, and IGF-I production in muscle and fat shortly after GH expoand muscle tissue in vivo (12, 30). Others have reported that sure. These observations support the notion that GH exerts diinsulin stimulated the expression of SOCS3 mRNA as well as rect and acute effects in peripheral tissues and that part of the Stat5b phosphorylation in adipocyte cultures and transfected GH-induced body growth is mediated through auto- or paracells and hypothesized that SOCS3 acts as an insulin-induced crine actions of IGF-I (25–28). In vivo animal and in vitro cell negative regulator of hormone signaling (11). In the present culture experiments have identified a Stat5 binding site in the



FIG. 6. Infusion study: effects of GH infusion on tyrosine phosphorylation of Stat5 (top panels) and on the expression of SOCS3 (middle panels) and IGF-I (bottom panels) mRNA in muscle (left panels) and adipose (right panels) tissue. A baseline and a stimulated muscle biopsy were obtained at t ⫽ 0 min and at t ⫽ 240 min (1 h after initiation of the hyperinsulinemic glucose clamp), respectively, whereas a fat biopsy was obtained at t ⫽ 180 min immediately before initiation of the hyperinsulinemic glucose clamp. Data are presented as means ⫾ SEM; n ⫽ 6. *, P ⬍ 0.05.

study, however, we did not detect any isolated effect of insulin infusion on either Stat5 phosphorylation or SOCS3 mRNA expression in either adipose or muscle tissue. SOCS3 mRNA expression was increased by GH infusion in both muscle and adipose tissue, which is compatible with animal in vitro and in vivo studies. The SOCS3 protein inhibits JAK2 by binding to both JAK2 and membrane-proximal phosphotyrosines on the GHR, which also functions as Stat5 docking sites (31). In the present study, GH infusion was not associated with SOCS1 or -2 mRNA expression, which supports the observation that SOCS3 is the major negative regulator of GHR signaling (32–35). Members of the SOCS family have been causally linked to cytokine-induced insulin resistance either by associating with the IR and suppressing the receptor kinase activity or by decreasing IRS phosphorylation. In adipose tissue of obese mice, SOCS3 decreased insulin-induced

2849

IRS-1 tyrosine phosphorylation and association with the p85 subunit on PI 3-kinase. Further downstream, a decrease in ERK1/2 and Akt activation was seen in response to SOCS3 (11, 36, 37). In addition, SOCS3 targeted IRS-1 and -2 for ubiquitin-mediated degradation in multiple cell types (38). Overexpression of SOCS3 protein has been detected in obese and in type 2 diabetic patients as well as in insulin-resistant obese and diabetic animals (35, 36). Our data do not rule out that SOCS3 could be involved in GH-induced insulin resistance in human subjects in vivo, because SOCS3 expression was increased, but this was not associated with detectable interaction with insulin signaling. In conclusion, the present study shows that acute exogenous GH exposure as well as endogenous GH bursts during control conditions translate into GHR signaling in muscle and adipose tissue, supporting a central role of the JAK-Stat pathway in GHR signaling in these tissues in human subjects. Furthermore, we provide indirect evidence of GH-induced local IGF-I production in peripheral tissue as well as expression of the negative feedback mediator SOCS3. We did not detect any cross-talk between the signaling pathways of GH and insulin in terms of either GH-induced changes in Akt or ERK1/2 phosphorylation or in insulin-induced changes in Stat5 phosphorylation or SOCS3 expression, respectively. We speculate that part of the acute GHinduced insulin resistance may be caused either by direct effects on hitherto uncharacterized cellular pathways or via signal proteins downstream of Akt in the IR signaling cascade.

Acknowledgments We thank Lone Svendsen and Susanne Sørensen for excellent technical assistance during the course of the clinical study and Elsebeth Hornemann, Lenette Pedersen, and Pia Hornbæk for great technical assistance at the laboratories. Address all correspondence and requests for reprints to: Charlotte Nielsen, Medical Department M, Aarhus University Hospital, Norrebrogade 44, DK-8000 Aarhus C, Denmark. E-mail: [email protected]. The ClinicalTrials.gov identification number is NCT00512473. This work was supported in part by unrestricted research grants from Novo Nordisk and Pfizer as well as grants from the FOOD Study Group / Ministry of Food, Agriculture and Fisheries & Ministry of Family and Consumer Affairs, Denmark. Charlotte Nielsen was supported by a Scholarship from the Danish Medical Research Council.

2850

Nielsen et al.


Disclosure Statement: C.N., L.C.G., N.J., S.B.P., S.L., and N.M. have nothing to declare. J.O.L.J. received consulting fees and lecture fees from Novo Nordisk and Pfizer.


19.

20.

References 1. Rabinowitz D, Klassen GA, Zierler KL 1965 Effects of human growth hormone on muscle and adipose tissue metabolism in the forearm of man. J Clin Invest 44:51– 61 2. Moller N, Butler PC, Antsiferov MA, Alberti KG 1989 Effects of growth hormone on insulin sensitivity and forearm metabolism in normal man. Diabetologica 32:105–110 3. Rosenfeld RG, Wilson DM, Dollar LA, Bennett A, Hintz RL 1982 Both human pituitary growth hormone and recombinant DNA-derived human growth hormone cause insulin resistance at a postreceptor site. J Clin Endocrinol Metab 54:1033–1038 4. Rizza RA, Mandarino LJ, Gerich JE 1982 Effects of growth hormone on insulin action in man. Mechanisms of insulin resistance, impaired suppression of glucose production, and impaired stimulation of glucose utilization. Diabetes 31:663– 669 5. Holdaway IM, Rajasoorya RC, Gamble GD 2004 Factors influencing mortality in acromegaly. J Clin Endocrinol Metab 89:667– 674 6. Shulman GI 2004 Unraveling the cellular mechanism of insulin resistance in humans: new insights from magnetic resonance spectroscopy. Physiology 19: 183–190 7. Jessen N, Djurhuus CB, Jorgensen JO, Jensen LS, Moller N, Lund S, Schmitz O 2005 Evidence against a role for insulin-signaling proteins PI 3-kinase and Akt in insulin resistance in human skeletal muscle induced by short-term GH infusion. Am J Physiol Endocrinol Metab 288:E194 –E199 8. Moller N, Jorgensen JO, Schmitz O, Moller J, Christiansen J, Alberti KG, Orskov H 1990 Effects of a growth hormone pulse on total and forearm substrate fluxes in humans. Am J Physiol Endocrinol Metab 258:E86 –E91 9. Hwa V, Little B, Adiyaman P, Kofoed EM, Pratt KL, Ocal G, Berberoglu M, Rosenfeld RG 2005 Severe growth hormone insensitivity resulting from total absence of signal transducer and activator of transcription 5b. J Clin Endocrinol Metab 90:4260 – 4266 10. Dominici FP, Argentino DP, Munoz MC, Miquet JG, Sotelo AI, Turyn D 2005 Influence of the crosstalk between growth hormone and insulin signalling on the modulation of insulin sensitivity. Growth Horm IGF Res 15:324 –336 11. Emanuelli B, Peraldi P, Filloux C, Chavey C, Freidinger K, Hilton DJ, Hotamisligil GS, Van Obberghen E 2001 SOCS-3 inhibits insulin signaling and is up-regulated in response to tumor necrosis factor-␣ in the adipose tissue of obese mice. J Biol Chem 276:47944 – 47949 12. Sadowski CL, Choi TS, Le M, Wheeler TT, Wang LH, Sadowski HB 2001 Insulin induction of SOCS-2 and SOCS-3 mRNA expression in C2C12 skeletal muscle cells is Mediated by Stat5*. J Biol Chem 276:20703–20710 13. Thirone ACP, Carvalho CRO, Saad MJ 1999 Growth hormone stimulates the tyrosine kinase activity of JAK2 and induces tyrosine phosphorylation of insulin receptor substrates and Shc in rat tissues. Endocrinology 140:55– 62 14. Ridderstrale M, Degerman E, Tornqvist H 1995 Growth hormone stimulates the tyrosine phosphorylation of the insulin receptor substrate-1 and its association with phosphatidylinositol 3-kinase in primary adipocytes. J Biol Chem 270:3471–3474 15. del Rincon JP, Iida K, Gaylinn BD, McCurdy CE, Leitner JW, Barbour LA, Kopchick JJ, Friedman JE, Draznin B, Thorner MO 2007 Growth hormone regulation of p85␣ expression and phosphoinositide 3-kinase activity in adipose tissue: mechanism for growth hormone-mediated insulin resistance. Diabetes 56:1638 –1646 16. Kopchick JJ, Parkinson C, Stevens EC, Trainer PJ 2002 Growth hormone receptor antagonists: discovery, development, and use in patients with acromegaly. Endocr Rev 23:623– 646 17. Gormsen LC, Nielsen C, Gjedsted J, Gjedde S, Vestergaard ET, Christiansen JS, Jorgensen JO, Moller N 2007 Effects of free fatty acids, growth hormone and growth hormone receptor blockade on serum ghrelin levels in humans. Clin Endocrinol (Oxf) 66:641– 645 18. Jorgensen JO, Feldt-Rasmussen U, Frystyk J, Chen JW, Kristensen LO, Hagen C, Orskov H 2005 Cotreatment of acromegaly with a somatostatin analog and

21.

22.

23.

24. 25.

26.

27.

28.

29.

30. 31.

32. 33.

34.

35.

36.

37.

38.

a growth hormone receptor antagonist. J Clin Endocrinol Metab 90:5627– 5631 Frystyk J DBOH 1995 Non-competitive time-resolved immunofluorometric assays for determination of human insulin-like growth factor I and II. Growth Regul 5:169 –176 Jorgensen JO, Jessen N, Pedersen SB, Vestergaard E, Gormsen L, Lund SA, Billestrup N 2006 Growth hormone receptor signaling in skeletal muscle and adipose tissue in human subjects following exposure to an intravenous GH bolus. Am J Physiol Endocrinol Metab 291:E899 –E905 Trainer PJ, Drake WM, Katznelson L, Freda PU, Herman-Bonert V, Van der Lely AJ, Dimaraki EV, Stewart PM, Friend KE, Vance ML, Besser GM, Scarlett JA 2000 Treatment of acromegaly with the growth hormone-receptor antagonist pegvisomant. N Engl J Med 342:1171–1177 van der Lely AJ, Hutson RK, Trainer PJ, Besser GM, Barkan AL, Katznelson L, Klibanski A, Herman-Bonert V, Melmed S, Vance ML 2001 Long-term treatment of acromegaly with pegvisomant, a growth hormone receptor antagonist. Lancet 358:1754 –1759 Smit LS, VanderKuur JA, Stimage A, Han Y, Luo G, Yu-Lee Ly, Schwartz J, Carter-Su C 1997 Growth hormone-induced tyrosyl phosphorylation and deoxyribonucleic acid binding activity of Stat5A and Stat5B. Endocrinology 138:3426 –3434 Darnell Jr JE 1997 STATs and gene regulation. Science 277:1630 –1635 Sjogren K, Liu JL, Blad K, Skrtic S, Vidal O, Wallenius V, LeRoith D, Tornell J, Isaksson OG, Jansson JO, Ohlsson C 1999 Liver-derived insulin-like growth factor I (IGF-I) is the principal source of IGF-I in blood but is not required for postnatal body growth in mice. Proc Natl Acad Sci USA 96:7088 –7092 Chow JC, Ling PR, Qu Z, Laviola L, Ciccarone A, Bistrian BR, Smith RJ 1996 Growth hormone stimulates tyrosine phosphorylation of JAK2 and STAT5, but not insulin receptor substrate-1 or SHC proteins in liver and skeletal muscle of normal rats in vivo. Endocrinology 137:2880 –2886 Sadowski CL, Wheeler TT, Wang LH, Sadowski HB 2001 GH regulation of IGF-I and suppressor of cytokine signaling gene expression in C2C12 skeletal muscle cells. Endocrinology 142:3890 –3900 Frost RA, Nystrom GJ, Lang CH 2002 Regulation of IGF-I mRNA and signal transducers and activators of transcription-3 and -5 (Stat-3 and -5) by GH in C2C12 myoblasts. Endocrinology 143:492–503 Woelfle J, Chia DJ, Rotwein P 2003 Mechanisms of growth hormone (GH) action: identification of conserved Stat5 binding sites that mediate GH-induced insulin-like growth factor-I gene activation. J Biol Chem 278:51261– 51266 Chen J, Sadowski HB, Kohanski RA, Wang LH 1997 Stat5 is a physiological substrate of the insulin receptor. Proc Natl Acad Sci USA 94:2295–2300 Ram PA, Waxman DJ 1999 SOCS/CIS protein inhibition of growth hormonestimulated STAT5 signaling by multiple mechanisms. J Biol Chem 274:35553– 35561 Flores-Morales A, Greenhalgh CJ, Norstedt G, Rico-Bautista E 2005 Negative regulation of GH receptor signaling. Mol Endocrinol 20:241–253 Davey HW, McLachlan MJ, Wilkins RJ, Hilton DJ, Adams TE 1999 STAT5b mediates the GH-induced expression of SOCS-2 and SOCS-3 mRNA in the liver. Mol Cell Endocrinol 158:111–116 Ridderstrale M, Amstrup J, Hilton DJ, Billestrup N, Tornqvist H 2003 SOCS-3 is involved in the downregulation of the acute insulin-like effects of growth hormone in rat adipocytes by inhibition of Jak2/IRS-1 signaling. Horm Metab Res 35:169 –177 Rieusset J, Bouzakri K, Chevillotte E, Ricard N, Jacquet D, Bastard JP, Laville M, Vidal H 2004 Suppressor of cytokine signaling 3 expression and insulin resistance in skeletal muscle of obese and type 2 diabetic patients. Diabetes 53:2232–2241 Ueki K, Kondo T, Tseng YH, Kahn CR 2004 Central role of suppressors of cytokine signaling proteins in hepatic steatosis, insulin resistance, and the metabolic syndrome in the mouse. Proc Natl Acad Sci USA 101:10422–10427 Mooney RA, Senn J, Cameron S, Inamdar N, Boivin LM, Shang Y, Furlanetto RW 2001 Suppressors of cytokine signaling-1 and -6 associate with and inhibit the insulin receptor. A potential mechanism for cytokine-mediated insulin resistance. J Biol Chem 276:25889 –25893 Rui L, Yuan M, Frantz D, Shoelson S, White MF 2002 SOCS-1 and SOCS-3 block insulin signaling by ubiquitin-mediated degradation of IRS1 and IRS2. J Biol Chem 277:42394 – 42398