Content uploaded by Jens Otto Lunde Jorgensen
Author content
All content in this area was uploaded by Jens Otto Lunde Jorgensen on Jan 02, 2014
Content may be subject to copyright.
Impact of Fasting on Growth Hormone Signaling and
Action in Muscle and Fat
Louise Moller, Lisa Dalman, Helene Norrelund, Nils Billestrup, Jan Frystyk, Niels Moller,
and Jens Otto Lunde Jorgensen
Medical Department M (Endocrinology and Diabetes) (L.M., L.D., J.F., N.M., J.O.L.J.), Aarhus University Hospital, Aarhus
Sygehus, DK-8000 Aarhus C, Denmark; Department of Internal Medicine (H.N.), Viborg Hospital, DK-8800 Viborg,
Denmark; and Department for Translational Diabetology (N.B.), Steno Diabetes Center, DK-2820 Gentofte, Denmark
Context: Whether GH promotes IGF-I production or lipolysis depends on nutritional status, but the
underlying mechanisms remain unknown.
Objective: To investigate the impact of fasting on GH-mediated changes in substrate metabolism,
insulin sensitivity, and signaling pathways.
Design: We conducted a randomized crossover study.
Subjects: Ten healthy men (age 24.3 ⫾0.6 yr, body mass index 23.1 ⫾0.4 kg/m
2
) participated.
Intervention: A GH bolus administered 1) postabsorptively and 2) in the fasting state (37.5 h).
Skeletal muscle and adipose tissue biopsies were taken, and a hyperinsulinemic-euglycemic clamp
was performed on both occasions.
Main Outcome Measures: Metabolic clearance rate (MCR) of GH, substrate metabolism, and insulin
sensitivity were measured. Biopsies were subjected to Western blotting for expression of signaling
proteins and to RT-PCR for expression of suppressor of cytokine signaling protein 3 and IGF-I mRNA.
Results: Fasting was associated with reduced MCR of GH (P⬍0.01), enhanced lipolytic respon-
siveness to GH, decreased insulin sensitivity (P⬍0.01), and reduced IGF-I bioactivity (P⫽0.04). After
the GH bolus, phosphorylation of signal transducers and activators of transcription protein 5b
(pSTAT5b) were observed in both conditions; however, the phospho-STAT5b/STAT5b ratio was
significantly decreased in the fasting state (muscle P⫽0.02 and fat P⫽0.02).
Conclusion: The combination of fasting and GH exposure translates into enhanced lipolysis, re-
duced IGF-I activity and insulin sensitivity, and blunted activation of the Janus kinase (JAK)/STAT
pathway. Whether this change in signaling activity is related to the change in MCR of GH and/or
the concomitant shift in the metabolic effects of GH merits future attention. (J Clin Endocrinol
Metab 94: 965–972, 2009)
The impact of GH on substrate metabolism includes stimu-
lation of lipolysis and lipid oxidation, resistance to the
actions of insulin on hepatic and peripheral glucose metabolism,
and protein preservation (1–4). The lipolytic and insulin antag-
onistic properties are considered direct effects of GH, whereas
the protein anabolic effects are also mediated by GH-induced
IGF-I production (5).
Fasting is associated with a pronounced increase in both en-
dogenous GH production (6) and lipolytic responsiveness to ex-
ogenous GH (7). The insulin antagonistic effects of GH are pre-
served during fasting, which constitutes an important adaptation
by reducing glucose oxidation and thereby the need for glu-
coneogenic precursors from muscle protein (8). By contrast,
IGF-I declines progressively during fasting (9, 10) and thus the
ISSN Print 0021-972X ISSN Online 1945-7197
Printed in U.S.A.
Copyright © 2009 by The Endocrine Society
doi: 10.1210/jc.2008-1385 Received June 30, 2008. Accepted December 1, 2008.
First Published Online December 9, 2008
Abbreviations: AUC, Area under the curve; EGP, endogenous glucose production; FFA, free
fatty acids; GHBP, GH-binding protein; IGFBP-1, IGF-binding protein 1; JAK2, Janus kinase
2; MCR, metabolic clearance rate; 3-OHB, 3-hydroxybutyrates; RA, rate of appearance; RQ,
respiratory quotient; SOCS, suppressors of cytokine signaling; STAT5b, signal transducers
and activators of transcription protein 5b; TR-IFMA, time-resolved fluoroimmunoassay; Vd,
distribution volume.
ORIGINAL ARTICLE
Endocrine Research
J Clin Endocrinol Metab, March 2009, 94(3):965–972 jcem.endojournals.org 965
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2014. at 04:26 For personal use only. No other uses without permission. . All rights reserved.
anabolic and insulin-like effects, which makes teleological sense
in a condition with nutritional deprivation.
The cellular mechanisms subserving the switch in the actions
of GH from IGF-I production to effects on lipolysis have not yet
been identified. At the level of intracellular GH signaling pro-
teins, it is well documented that the receptor-associated Janus
kinase 2 (JAK2)/signal transducers and activators of transcrip-
tion 5b (STAT5b) pathway is essential for GH-induced IGF-I
gene activation (11, 12). Moreover, STAT5b serves as a tran-
scription factor for a family of cytokine-inducible suppressors of
cytokine signaling (SOCS), which in turn may feedback inhibit
JAK2/STAT5b signaling (13). Additional signaling pathways for
GH include the MAPK and phosphatidylinositol 3-kinase (14).
We have previously demonstrated that exposure to a single GH
bolus (0.5 mg) in the postabsorptive state translates into STAT5b
activation in human muscle and adipose tissues in vivo, whereas
no consistent changes were recorded in either MAPK or the phos-
phatidylinositol 3-kinase pathways (15).
In the present experiment in healthy human subjects, we
studied the impact of fasting on GH-mediated changes in sub-
strate metabolism and insulin sensitivity. Furthermore, GH
signaling pathways in skeletal muscle and sc fat tissue were
examined after exposure to exogenous GH in the postabsorp-
tive and fasting state.
Subjects and Methods
Subjects
The study population comprised 10 healthy nonobese young male
students from Aarhus University, age 24.3 ⫾0.6 yr, body mass index
23.1 ⫾0.4 kg/m
2
. The exclusion criteria were family history of type 2
diabetes, use of medications, and alcohol consumption above 21
U/wk. Three days before each study period, the subjects were in-
structed to consume a diet with no major deviations from the national
recommendations (maximum 30% of the energy from fat, 50–60%
from carbohydrates, and 10–20% from protein), to abstain from
alcohol, and to not deviate from their normal level of activity. During
the fasting period, they were allowed to drink tap or mineral water
and to perform normal ambulatory activities, excluding any kind of
exercise.
The protocol was approved by the regional ethics committee, and the
nature and potential risks were explained before participants gave writ-
ten informed consent. The study was conducted according to the decla-
ration of Helsinki (2000) of the World Medical Association.
Design
In a randomized crossover design, the subjects were each studied on
two occasions: 1) in the postabsorptive state after an overnight fast and
2) after 37.5 h fasting with at least 6 wk elapsing between each experi-
ment. The fasting period started at 2000 h, and the GH bolus (0.5 mg
Genotropin; MiniQuick, Pfizer ApS, Denmark) was administered at
0930 h (Fig. 1). At 0900 h, the subjects were placed in a quiet, thermo-
neutral room. Oneiv cannula was placed in an antecubital vein for infusion.
For blood sampling, a second cannula was inserted in a dorsal hand vein, and
the hand was placed in a 65 C heated box for arterializations of the blood.
For deep venous sampling, a third cannula was inserted retrogradely into the
contralateral deep antecubital vein. Unless otherwise stated, measurements
referred to as basal represent the mean of three measurements between
t⫽120 and t ⫽150 min and clamp measurements the mean of three
measurements between t ⫽240 and t ⫽270 min.
GH pharmacokinetics
The metabolic clearance rate (MCR) of GH was calculated from the
following equation: MCR ⫽dose of GH injected/area under the curve
(AUC) (GH AUC
0–120min
was used). The elimination constant kwas
determined as the slope of the ln-linear regression of the GH disappear-
ance curve (t ⫽30 ⫺t⫽120 min), and half-time (t
1
⁄
2
) calculated as
t
1
⁄
2
⫽ln2/k. Distribution volume (Vd) was calculated as Vd ⫽MCR/k.
Both the MCR and Vd were corrected for body surface area (in square
meters) calculated as [(height in centimeters ⫻weight in kilograms)/
3600)]
1
⁄
2
. The calculation does not account for any impact of endoge-
nous GH levels, which is considered of minimal significance in consid-
eration of the size of the exogenous GH dose.
Glucose metabolism and insulin sensitivity
At t ⫽0 min, a priming dose of [3-
3
H]glucose (NEN Life Science
Products, Boston, MA; 20
Ci) was given, followed by a continuous
infusion of [3-
3
H]glucose (12
Ci/h) for 4.5 h. At t ⫽150 min, the hy-
perinsulinemic-euglycemic clamp began with an insulin infusion rate of 0.6
mU/kg 䡠min (Actrapid; Novo Nordisk A/S, Copenhagen, Denmark).
Plasma glucose was measured every 10 min and kept at 5 mmol/liter by
0 60 120 150 270240
Clock time
(h)
Study time
(min)
20 20 9.30 12 14
Postabs
37½ h fast
Basal
Basal
Clamp
Clamp
ClampBasal
t
Bolus of growth hormone
Biopsies
Indirect calorimetry
Bolus and continuous 3-3H Glucose
Hyperinsulinemic euglycemic clamp
Blood samples
Glucose forearm metabolism
FIG. 1. Study design.
966 Moller et al. GH Signaling and Action during Fasting J Clin Endocrinol Metab, March 2009, 94(3):965–972
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2014. at 04:26 For personal use only. No other uses without permission. . All rights reserved.
adjusting the infusion rate of 20% glucose (16). To minimize rapid di-
lution of the [3-
3
H]glucose, [3-
3
H]glucose was added to the infused glu-
cose (100
Ci [3-
3
H]glucose/500 ml 20% glucose). Glucose rates of
appearance (Ra) were estimated by dividing the [3-
3
H]glucose infusion
rate by the measured specific activity of [3-
3
H]glucose and corrected for
non-steady state (17). A pool fraction of 0.65 was used. Nonoxidative
glucose disposal was calculated by subtracting oxidative glucose dis-
posal as assessed by indirect calorimetry (see below) from whole-body
glucose disposal. Basal endogenous glucose production (EGP) was
estimated as the mean of Ra calculated during the last 30 min of the
basal period. The amount of infused glucose required to maintain
blood glucose at 5 mmol/liter during the clamp, i.e. the M value,
reflects peripheral insulin sensitivity. EGP during the clamp period
was estimated by subtracting the M value from the Ra at the end of
the clamp.
Forearm glucose uptake
Arterialized blood was drawn from the heated hand vein, and
venous blood was obtained from the contralateral deep antecubital
vein. Total forearm blood flow was measured by venous occlusion
plethysmography before each deep venous sample. Glucose uptake
was calculated by multiplying the blood flow and the arteriovenous
glucose difference. Mean glucose uptake in the forearm was estimated
over the last 30 min of the basal and clamp period (1).
Indirect calorimetry
The respiratory quotient (RQ), and resting energy expenditure were
estimated by indirect calorimetry (Deltatrac monitor; Dantes Instrumenta-
rium, Helsinki, Finland) performed during the last 30 min of the basal and
clamp period, respectively. The mean values of the last 25 min were used for
calculations. Rates of protein oxidation were calculated on the basis of urea
nitrogen excretion. Lipid and glucose oxidation were estimated after cor-
rection for protein oxidation (18).
Intracellular signal transduction
At t ⫽60 min, muscle tissue from vastus lateralis was obtained by a
Bergstro¨ m biopsy needle. The tissue was cleaned of blood (within 10 sec)
and snap-frozen in liquid nitrogen. Subcutaneous fat tissue from the
abdomen was obtained by liposuction, cleaned from blood, and frozen
in liquid nitrogen. Due to technical problems, biopsies were obtained
from only nine subjects.
The biopsies were homogenized in 200
l lysis buffer [20 mMTris
(pH 7.0), 1% Triton X-100, 270 mMsucrose, 1 mMEDTA, 1 mMEGTA,
50 mMNaF, 5 mMtetrasodium pyrophosphate, 10 mMglycerol phos-
phate, 1 mMbenzamidine, 4
g/
l leupeptin, 1 mMdithiothreitol, 10 mM
Na
3
VO
4
] and centrifuged for 20 min at 10,000 ⫻gat 4 C. The super-
natant was collected and the protein concentration measured. The lysates
were stored at ⫺80 C until analysis.
Western blot
Primary antibodies were as follows: phospho-STAT5b (pSTAT5b)
(Tyr694), STAT5b (3H7), phosphorylated ERK (p44/42 MAPK)
(Thr202/Tyr204), p44/42 MAPK, phosphorylated antiapoptotic serine
kinase (Akt) (Ser473) and Akt (Cell Signaling Technology, Beverly, MA).
An antirabbit IgG H⫹L horseradish peroxidase was used as a sec-
ondary antibody. Western blotting using 15
g protein per sample
was performed as described (19). The proteins of interest were de-
tected by a chemiluminescence detection system (LumiGLO reagent
and peroxide; Cell Signaling Technology) and visualized using an
imaging system (Las3000; Fuji Film) according to the manufacturer’s
instructions. Band intensities were quantified using Multigauge soft-
ware (Fuji Film).
Real-time RT-PCR
RNA was analyzed as described previously (15). In short, total RNA
was isolated from muscle biopsies using TRIzol reagent (Invitrogen,
Carlsbad, CA). The muscle biopsies were homogenized using a polytron
in the presence of TRIzol, and RNA was isolated as described in the
manual from Invitrogen. cDNA synthesis was performed using the Taq-
Man kit N808-0234 (PerkinElmer, Boston, MA). The PCR was per-
formed using SYBR Green Master mix (Applied Biosystems, Foster City,
CA) with primers as described.
Blood analysis
Plasma glucose was immediately measured in duplicate on a Beckman
Glucoanalyzer (Beckman Instruments, Palo Alto, CA). Serum samples
were frozen and stored at ⫺20 C. The specific activity of [3-
3
H]glucose
was measured as previously described (1). Insulin, GH, and cortisol were
analyzed using time-resolved fluoroimmunoassay (TR-IFMA; Au-
toDELFIA, PerkinElmer, Turku, Finland). C-peptide was measured by
ELISA (DakoCytomation, Cambridgeshire, UK), and free fatty acids
(FFA) were analyzed by a commercial kit (Wako Chemicals, Neuss, Ger-
many). Glycerol, lactate, alanine, and 3-hydroxybutyrates (3-OHB)
were measured using COBAS biocentrifugal analyzer with fluorometric
attachment (Roche Diagnostics, Welwyn Garden City, UK). Urea was
determined by a commercial method (Cobas Integra 800; Roche, Mann-
heim, Germany). Glucagon was analyzed using in-house RIA, total IGF-I
by TR-IFMA, and IGF-I bioactivity by a kinase receptor activation assay
(20). IGF-binding protein 1 (IGFBP-1) was measured by an in-house RIA,
and GH-binding protein (GHBP) was measured by an in-house TR-IFMA.
Leptin was measured by a commercial kit (Alpco Diagnostics, Salem, NH),
and all values below the detection limit were set to the detection limit (1
g/liter). The Luminex Suspension Array System (Bio-Plex, Biosource Lab-
oratories Inc., Hercules, CA) was used for the analysis of IL-6 (detection
limit, 3 ng/liter) and TNF-
␣
(detection limit, 5 ng/liter).
Statistical analysis
To examine the effects of fasting, samples from t ⫽0 min were compared
(t ⫽0
fasting
vs. t⫽0
postabsorptive
). The effects of GH and time were inves-
tigated using
␦
-values (basal ⫺t⫽0
fasting
vs. basal ⫺t⫽0
postabsorptive
)
or AUC. For comparison of the combined effect of GH bolus and
nutritional status (with or without fasting), absolute values (basal-
fasting
vs. basal
postabsorptive
) were used. For assessment of differences in
insulin sensitivity, both absolute (clamp
fasting
vs. clamp
postabsorptive
) and
␦
-values (clamp-basal
fasting
vs. clamp-basal
postabsorptive
) were used.
Student’s paired ttest or Wilcoxon signed rank matched pairs test
were used as appropriate after testing for normal distribution by
Kolmogorov-Smirnov. Skewed data were log transformed before ap-
plying relevant statistical tests and presented as medians and ranges.
Unless otherwise stated, data are presented as mean ⫾SE APvalue ⬍0.05
was considered significant.
Results
GH pharmacokinetics and IGF-I
As illustrated in Table 1, at t ⫽0 min, the endogenous GH
levels were significantly elevated in the fasting state, and the
increase in serum GH levels (AUC
0–120
) in response to the GH
bolus was significantly higher in the fasting condition (P⬍0.01).
In the fasting state, the MCR of GH was reduced and t
1
⁄
2
was
prolonged, whereas Vd did not differ. Serum GHBP levels were
comparable in the two states. Serum total IGF-I levels were sim-
ilar in the fasting and postabsorptive state (t ⫽0 min: 184 ⫾9
vs. 192 ⫾8
g/liter, P⫽0.43; basal: 183 ⫾9vs. 195 ⫾9
g/liter,
P⫽0.14), but IGF-I bioactivity was significantly lower in the
fasting state (t ⫽0 min: 2.04 ⫾0.19 vs. 2.47 ⫾0.20
g/liter, P⬍
0.05), in agreement with significantly higher levels of IGFBP-1
(t ⫽0: 153.8 ⫾27.6 vs. 67.8 ⫾11.8
g/liter, P⬍0.01) and a
J Clin Endocrinol Metab, March 2009, 94(3):965–972 jcem.endojournals.org 967
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2014. at 04:26 For personal use only. No other uses without permission. . All rights reserved.
significant negative correlation between the difference in IGF-I
bioactivity and IGFBP-1, respectively (r ⫽⫺0.66; P⫽0.04).
Hormones, cytokines, and glucose metabolism
As illustrated in Table 2, at t ⫽0 min, plasma glucose, insulin,
and C-peptide levels were significantly lower in the fasting state
compared with the postabsorptive state. The change in plasma
glucose levels after the GH bolus was different in the two con-
ditions, with a significant decrease in the fasting condition. In-
sulin increased after the GH bolus, but significantly only in the
fasting state, whereas no changes in C-peptide levels were de-
tected. Cortisol as well as glucagon levels were higher in the
fasting state before the GH bolus. Glucagon increased after the
GH bolus, and the increase was significantly larger in the fasting
state. Leptin was significantly lower in the fasting state compared
with the postabsorptive state [t ⫽0: 0 (0–15.0) vs. 2.4 (0–15.2)
g/liter, P⬍0.05] and did not change during the basal period
(data not shown). IL-6 and TNF-
␣
levels were below the detec-
tion limits both postabsorptively and fasting.
In the basal period, EGP, nonoxidative glucose disposal, and
glucose oxidation did not differ significantly when comparing
the postabsorptive and fasting states. During the clamp, a com-
TABLE 1. GH pharmacokinetics in postabsorptive and fasting state in relation to GH bolus
Postabsorptive Fasting P
GH
t⫽0
(
g/liter) 0.39 (0.04–7.28) 4.30 (0.62–25.40) ⬍0.05
GH
elevation0–30
(
g/liter) 25.3 ⫾3.4 38.0 ⫾1.8 ⬍0.01
GH AUC
0–120
(
g/liter/min) 1178.8 ⫾167.9 1704.3 ⫾113.8 ⬍0.01
MCR (ml/min 䡠/m
2
)231.3 (132.2–980.6) 146.3 (114.1–197.7) ⬍0.01
GHBP (nmol/liter) 1.86 (1.23–7.36) 2.09 (1.52–6.32) 0.09
T
1/2
(min) 22.8 ⫾1.0 29.6 ⫾2.3 ⬍0.05
Vd (liters/m
2
)7.6 (4.2–32.5) 6.0 (4.9 –10.9) 0.24
Intravenous GH (0.5 mg) was administered at t ⫽0 in the postabsorptive and fasting state. Data are presented as median and range or mean ⫾SE.
TABLE 2. Changes in glucose metabolism during the study period
tⴝ0
Basal,
tⴝ120–150
Clamp,
tⴝ240–270
P,tⴝ0
vs. basal
P, basal
vs.
clamp
␦
-Values
P, Fasting vs.
postabsorptive
(change from
tⴝ0 to basal)
P, fasting vs.
postabsorptive
(change from
basal to clamp)
Postabsorptive state
Glucose (mmol/liter) 4.8 ⫾0.1 4.9 ⫾0.1 4.8 ⫾0.04 0.07 0.15
Insulin (pmol/liter) 17.3 ⫾1.9 20.1 ⫾1.7 195.7 ⫾6.2 0.07 ⬍0.01
C-peptide (pmol/liter) 367 ⫾23 348 ⫾17 0.09
Cortisol (nmol/liter) 387.6 ⫾37.8 315.9 ⫾34.1 ⬍0.01
Glucagon (ng/liter) 47.7 ⫾5.0 48.4 ⫾5.8 0.68
EGP (mg/kg 䡠min) 1.47 (0.04–2.07) 0.00 (⫺1.97– 0.45) ⬍0.01
NOGD (mg/kg 䡠min) 1.10 ⫾0.14 2.73 ⫾0.23 ⬍0.01
Glucose oxidation
(mg/kg 䡠min)
0.68 ⫾0.10 1.35 ⫾0.11 ⬍0.01
Fasting state
Glucose (mmol/liter) 4.0 ⫾0.1 3.8 ⫾0.1 5.1 ⫾0.1 ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01
Insulin (pmol/liter) 11.1 ⫾1.5 16.2 ⫾2.2 200.4 ⫾8.1 ⬍0.01 ⬍0.01 0.36 0.18
C-peptide (pmol/liter) 203 ⫾39 189 ⫾22 0.62 0.82
Cortisol (nmol/liter) 459.6 ⫾48.5 443.3 ⫾53.6 0.79 0.32
Glucagon (ng/liter) 83.4 ⫾8.5 103.6 ⫾11.1 ⬍0.01 ⬍0.01
EGP (mg/kg 䡠min) 1.44 (0.91–5.49) 0.24 (⫺0.49–1.29) ⬍0.01 0.11
NOGD (mg/kg 䡠min) 1.43 ⫾0.47 2.15 ⫾0.34 0.16 0.12
Glucose oxidation
(mg/kg 䡠min)
0.45 ⫾0.10 0.52 ⫾0.18 0.65 ⬍0.01
NOGD, Nonoxidative glucose disposal.
968 Moller et al. GH Signaling and Action during Fasting J Clin Endocrinol Metab, March 2009, 94(3):965–972
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2014. at 04:26 For personal use only. No other uses without permission. . All rights reserved.
parable suppression of EGP was recorded in both conditions, but
glucose disposal was increased only in the postabsorptive state.
The insulin-stimulated glucose uptake (M value) was signifi-
cantly lower in the fasting state compared with the postabsorp-
tive (1.95 ⫾0.19 vs. 3.67 ⫾0.24 mg/kg 䡠min, P⬍0.01).
Forearm metabolism
Forearm blood flow were higher during the fast compared
with the postabsorptive state (basal state: 3.9 ⫾0.7 vs. 2.2 ⫾0.2
ml/100 ml tissue 䡠min, P⫽0.06; clamp: 3.3 ⫾0.4 vs. 2.5 ⫾0.3
ml/100 ml tissue 䡠min, P⫽0.04). The forearm glucose uptake
was not significantly different in either the basal or the insulin-
stimulated state (data not shown).
Lipid metabolism
As illustrated in Table 3, FFA, glycerol, and 3-OHB were
significantly elevated at t ⫽0 min in the fasting state, and all lipid
intermediates increased significantly in response to the GH bolus
in both conditions. The lipolytic effect of GH was amplified by
fasting as judged by the increments in the circulating levels of
FFA, glycerol, and 3-OHB (Fig. 2). This was accompanied by
significantly higher rates of lipid oxidation in the fasting state com-
pared with postabsorptive (basal and clamp, P⬍0.05). All lipid
intermediates decreased significantly during the clamp. Significant
suppression of lipid oxidation during the clamp was observed on
both occasions, but the change (
␦
-value) tended to be less pro-
nounced during fasting. Additionally the insulin-induced increase
in RQ was significantly higher in the postabsorptive state.
GH signaling
Figure 3 illustrates the quantity of phosphorylated and total
STAT5b and Erk as well as the pSTAT5b/STAT5b ratio, which
was significantly lower in the fasting state in both fat tissue and
skeletal muscle. The lower ratio was due to decreased pSTAT5b
during fasting compared with the postabsorptive state (fat tissue:
14.7 ⫾5.8 ⫻10
6
vs. 27.0 ⫾8.5 ⫻10
6
,P⫽0.16; muscle: 9.2 ⫾
3.0 ⫻10
5
vs. 22.5 ⫾7.4 ⫻10
5
,P⫽0.03), whereas total STAT5b
was unchanged (fat tissue: 22.0 ⫾7.0 ⫻10
6
vs. 24.0 ⫾5.2 ⫻10
6
,
P⫽0.77; muscle: 19.2 ⫾2.8 ⫻10
6
vs. 21.9 ⫾3.7 ⫻10
6
,P⫽0.27).
We observed no difference in the phosphorylation of Akt,
Erk, or the Src tyrosine kinase family members c-src, c-yes,
c-fyn, or c-lyn. Additionally, we found no differences in
SOCS3 mRNA or IGF-I mRNA expression in skeletal muscle,
when comparing the postabsorptive and fasting state (data not
shown).
Discussion
The shift in the actions of GH from IGF-I production to promo-
tion of lipolysis in response to fasting is well recognized, but the
underlying mechanisms and consequences for the intracellular
GH signaling pathways have not yet been investigated. In this
TABLE 3. Changes in lipid metabolism during the study period
tⴝ0
Basal,
tⴝ120–150
Clamp,
tⴝ240–270
P,tⴝ0
vs. basal
P, basal
vs.
clamp
␦
-Values
P, Fasting vs.
postabsorptive
(change from
tⴝ0to
basal)
P, Fasting vs.
postabsorptive
(change from
tⴝ0to
basal)
Postabsorptive state
FFA (
mol/liter) 550.4 ⫾44.7 858.8 ⫾30.3 59.9 ⫾8.8 ⬍0.01 ⬍0.01
Glycerol (
mol/liter) 41.0 ⫾10.1 67.0 ⫾3.8 19.1 ⫾2.5 ⬍0.05 ⬍0.01
3-OHB (
mol/liter) 161.0 ⫾52.7 417.3 ⫾56.6 34.3 ⫾6.9 ⬍0.01 ⬍0.01
RQ 0.77 ⫾0.01 0.82 ⫾0.01 ⬍0.01
Lipid oxidation
(mg/kg 䡠min)
1.13 ⫾0.08 0.78 ⫾0.05 ⬍0.01
Fasting state
FFA (
mol/liter) 1048.7 ⫾75.9 1486.4 ⫾80.6 171.3 ⫾22.2 ⬍0.01 ⬍0.01 0.14 ⬍0.01
Glycerol (
mol/liter) 74.5 ⫾4.0 116.6 ⫾4.7 32.1 ⫾5.4 ⬍0.01 ⬍0.01 0.22 ⬍0.01
3-OHB (
mol/liter) 2048.5 ⫾269.8 2723.5 ⫾259.3 789.8 ⫾171.2 ⬍0.01 ⬍0.01 ⬍0.01 ⬍0.01
RQ 0.76 ⫾0.01 0.77 ⫾0.01 0.16 ⬍0.05
Lipid oxidation
(mg/kg 䡠min)
1.28 ⫾0.07 1.03 ⫾0.11 ⬍0.05 0.43
FIG. 2. Line/scatter plot shows 3-OHB levels before (0) administration of GH bolus
and at the end of the basal period and hyperinsulinemic-euglycemic clamp during
the postabsorptive (E) and fasting (F) state. Bar chart shows 3-OHB
␦
-values
(basal ⫺0) during the postabsorptive (black) and fasting (white) state. *, P⬍0.05.
J Clin Endocrinol Metab, March 2009, 94(3):965–972 jcem.endojournals.org 969
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2014. at 04:26 For personal use only. No other uses without permission. . All rights reserved.
randomized crossover study, we show that the metabolic change
in the fasting condition is accompanied by a significant reduction
in STAT5b signaling in skeletal muscle as well as in sc fat tissue
after a single GH bolus.
Our data corroborate that fasting is associated with reduced
glucose and insulin levels and increased lipolysis and lipid oxida-
tion. The lipolytic responsiveness to GH was enhanced in the fasting
situation, which is in keeping with previous data (7). In addition we
confirmed that fasting is associated with insulin resistance and
that the insulin antagonistic effects of GH on peripheral glu-
cose disposal become accentuated with fasting. Ourdesign does
not exclude the possibility that ongoing fasting per se could account
for some of the metabolic effects observed after GH administration;
still, the changes occurred rapidly and were compatible with
GH-induced effects obtained in previous studies (2, 7, 21). We ob-
served that GH administration was associated with increased
levels of glucagon, which most likely represents a component of
the stress response caused by fasting per se (22).
In the present study, IGF-I bioactivity was significantly
suppressed in the fasting state as compared with the postab-
sorptive state, and IGFBP-1 was significantly enhanced,
whereas total IGF-I levels were not significantly different in
the two situations. This is in accordance with the observations
by Chen et al. (10) who demonstrated that robust reductions
in total IGF-I during prolonged fasting is preceded by reduc-
tion in free and bioactive IGF-I levels and concomitant in-
crease in IGFBP-1. It therefore seems plausible that the decline
in bioactive IGF-I in our study is causally linked to the ob-
served increase in IGFBP-1.
GHBP has been shown to slow the MCR of GH by confining
GH to the intravascular compartment (23). Our data indicate
that the MCR of GH is decreased and t
1
⁄
2
is increased in the
fasting state as compared with the postabsorptive state, despite
similar levels of GHBP. The main extrarenal clearance pathway
for GH is via receptor-mediated internalization, and because
insulin is positively correlated with hepatic GH receptor levels
(24), the fasting-associated insulinopenia may contribute to
the reduced MCR of GH. The renal clearance of GH correlates
with the glomerular filtration rate (25), and because glomer-
ular filtration rate decreases during fasting (26), it is plausible
that the renal contribution to the MCR also decreases. The
diminution of GH MCR associated with fasting might have
clinical significance because GH availability has been shown
to be greater after sc injections of GH in the evening compared
with in the morning (27).
A single GH bolus (0.5 mg) has previously been demonstrated to
induce tyrosine phosphorylation of STAT5b in human skeletal
muscle and fat tissue in the postabsorptive state (15). In the present
study, the level of phospho-STAT5b and phospho-/total STAT5b
were decreased in both muscle and fat after the GH bolus in the
fasting state as compared with the postabsorptive state.
Impaired JAK2/STAT5 signal transduction has been ob-
served in conditions with GH resistance and reduced IGF-I pro-
duction in animal models of sepsis (28, 29) and chronic renal
failure (30), whereas data regarding the importance of SOCS
proteins are conflicting. The SOCS proteins play an important
role in the negative regulation of GH signaling, but we found no
difference in SOCS mRNA when comparing the fasting and post-
absorptive state in either muscle or fat tissue.
The suggestion of peripheral GH resistance during fasting at the
JAK2/STAT5 level is surprising inasmuch as GH-induced lipolysis
has previously been linked to the JAK2/STAT5 pathway in human
GH receptor-transfected adipocytes (31) and STAT5a/b knockout
mice (32). According to our data, the fasting-induced increase in
lipolysis in human subjects could be independent of STAT acti-
vation and mediated by alternative pathways. There is evidence
FIG. 3. Effect of GH in the postabsorptive (P) and fasting state (F). Western blots are shown for tyrosine phosphorylation (p) and total (t) STAT5b and Erk in skeletal
muscle and abdominal sc fat tissue (biopsies were obtained 60 min after GH bolus). Nine subjects (1–9). The samples marked GH ⫹and ⫺were included as positive
and negative controls. They represent lysates from muscle and fat biopsies obtained from a healthy volunteer before (⫺) and 30 min after (⫹) GH administration. Bar
chart shows ratio of p/total STAT5b and Erk in the postabsorptive (black) and fasting (white) state in each tissue. *, P⬍0.05.
970 Moller et al. GH Signaling and Action during Fasting J Clin Endocrinol Metab, March 2009, 94(3):965–972
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2014. at 04:26 For personal use only. No other uses without permission. . All rights reserved.
that not all GH signaling events depend on JAK2. In vitro studies
have demonstrated that the Src kinase can be activated indepen-
dently of JAK2 (14), but we were unable to detect a distinct
activation of Src tyrosine kinase in either muscle or fat tissue
regardless of nutritional status. Additionally, we found no acti-
vation of Akt or Erk, which is in keeping with our previous study,
which included a saline control (15).
Certain limitations of our study design merit attention. First,
no biopsies were obtained before the GH bolus. It is therefore not
possible to assess whether fasting per se impacted total or phos-
phorylated STAT5 levels in our biopsies. Theoretically, it is pos-
sible that prolonged fasting-induced elevations in endogenous
GH levels before the GH bolus may have induced STAT5 phos-
phorylation and subsequently made the tissues refractory to sub-
sequent exogenous GH exposure. It is also possible that fasting
regulates STAT5 activity via GH-independent mechanisms, e.g.
by changing the levels of certain interleukins and leptin. We
found the levels of IL-6 to be below the detection limits, whereas
leptin levels were decreased in the fasting state. The long isoform
of the leptin receptor, capable of signaling through the JAK/
STAT pathway, has been demonstrated in fat tissue and skeletal
muscle (33, 34), but the functional in vivo role in human subjects
remains uncertain. Second, a single biopsy was obtained 60 min
after the GH bolus based on the fact that STAT5b tyrosine phos-
phorylation as well as SOCS expression were recorded at this
time point after GH exposure in a similar human model (15). A
larger time series of biopsies after the GH bolus would have been
more informative. Third, we chose to study the impact of 37.5 h
fasting because GH previously has been documented to promote
lipolysis and to induce insulin resistance within this window of
time in healthy subjects and patients with GH deficiency (8, 9).
Because the metabolic adaptation to fasting changes as a func-
tion of the duration of fasting, it is likely that GH signaling may
change accordingly. Finally, to postulate a true causal relation-
ship between diminished JAK2/STAT5b signaling during fasting
and certain changes in substrate metabolism would require ad-
ditional and more mechanistic studies.
In summary, fasting was associated with reduced MCR of
GH, enhanced lipolysis, insulin resistance, and reduced bio-
active IGF-I levels. This was accompanied by reduced activa-
tion of the JAK/STAT pathway in muscle and fat after acute
exogenous GH exposure. Whether this decrease in signaling
activity is related to the change in MCR of GH and/or the
concomitant shift in the metabolic effects of GH merits at-
tention. We believe that the human model reported here may
provide a viable tool for future studies in this field.
Acknowledgments
We thank medical laboratory technicians K. N. Rasmussen, H. F.
Petersen, E. S. Hornemann, and L.R. Kristensen for excellent technical
assistance.
Address all correspondence and requests for reprints to: Louise
Moller, Medical Department M, Aarhus Sygehus, Norrebrogade 44,
DK-8000 Aarhus, Denmark. E-mail: louisem@dadlnet.dk.
Disclosure Summary: L.M., H.N., N.B., and N.M. have nothing to de-
clare. L.D. received a scholarship stipendium from Novo Nordisk. J.F. re-
ceived consulting fees from Novo Nordisk, Germany, and Hoffmann-LA
Roche. J.O.L.J. received consulting fees and lecture fees from Novo Nor-
disk, Novartis, Pfizer, and Ipsen.
The study was sponsored by Pfizer Inc. with an unrestricted research
grant.
References
1. Moller N, Butler PC, Antsiferov MA, Alberti KG 1989 Effects of growth
hormone on insulin sensitivity and forearm metabolism in normal man. Dia-
betologia 32:105–110
2. 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 258:E86–E91
3. Bak JF, Moller N, Schmitz O 1991 Effects of growth hormone on fuel utili-
zation and muscle glycogen synthase activity in normal humans. Am J Physiol
260:E736–E742
4. Fryburg DA, Gelfand RA, Barrett EJ 1991 Growth hormone acutely stimulates fore-
arm muscle protein synthesis in normal humans. Am J Physiol 260:E499–E504
5. Butler AA, Le RD 2001 Control of growth by the somatropic axis: growth
hormone and the insulin-like growth factors have related and independent
roles. Annu Rev Physiol 63:141–164
6. Hartman ML, Veldhuis JD, Johnson ML, Lee MM, Alberti KG, Samojlik E,
Thorner MO 1992 Augmented growth hormone (GH) secretory burst fre-
quency and amplitude mediate enhanced GH secretion during a two-day fast
in normal men. J Clin Endocrinol Metab 74:757–765
7. Moller N, Porksen N, Ovesen P, Alberti KG 1993 Evidence for increased
sensitivity of fuel mobilization to growth hormone during short-term fasting
in humans. Horm Metab Res 25:175–179
8. Norrelund H 2005 The metabolic role of growth hormone in humans with
particular reference to fasting. Growth Horm IGF Res 15:95–122
9. Norrelund H, Frystyk J, Jorgensen JO, Moller N, Christiansen JS, Orskov H,
Flyvbjerg A 2003 The effect of growth hormone on the insulin-like growth
factor system during fasting. J Clin Endocrinol Metab 88:3292–3298
10. Chen JW, Hojlund K, Beck-Nielsen H, Sandahl CJ, Orskov H, Frystyk J 2005
Free rather than total circulating insulin-like growth factor-I determines the
feedback on growth hormone release in normal subjects. J Clin Endocrinol
Metab 90:366–371
11. Chia DJ, Ono M, Woelfle J, Schlesinger-Massart M, Jiang H, Rotwein P 2006
Characterization of distinct Stat5b binding sites that mediate growth hor-
mone-stimulated IGF-I gene transcription. J Biol Chem 281:3190–3197
12. Fang P, Kofoed EM, Little BM, Wang X, Ross RJ, Frank SJ, Hwa V, Rosenfeld
RG 2006 A mutant signal transducer and activator of transcription 5b, asso-
ciated with growth hormone insensitivity and insulin-like growth factor-I de-
ficiency, cannot function as a signal transducer or transcription factor. J Clin
Endocrinol Metab 91:1526–1534
13. Greenhalgh CJ, Alexander WS 2004 Suppressors of cytokine signalling and
regulation of growth hormone action. Growth Horm IGF Res 14:200–206
14. Lanning NJ, Carter-Su C 2006 Recent advances in growth hormone signaling.
Rev Endocr Metab Disord 7:225–235
15. Jorgensen JO, Jessen N, Pedersen SB, Vestergaard E, Gormsen L, Lund SA,
Billestrup N 2006 GH 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
16. DeFronzo RA, Tobin JD, Andres R 1979 Glucose clamp technique: a method
for quantifying insulin secretion and resistance. Am J Physiol 237:E214 –E223
17. Debodo RC, Steele R, Altszuler N, Dunn A, Bishop JS 1963 On the hormonal
regulation of carbohydrate metabolism; studies with C14 glucose. Recent Prog
Horm Res 19:445–488
18. Ferrannini E 1988 The theoretical bases of indirect calorimetry: a review.
Metabolism 37:287–301
19. Frobose H, Groth Rønn S, Heding PE, Mendoza H, Cohen P, Mandrup-
Poulsen T, Billestrup N 2006 Suppressor of cytokine signaling-3 inhibits in-
terleukin-1 signaling by targeting the TRAF-6/TAK1 complex. Mol Endocri-
nol 20:1587–1596
20. Chen JW, Ledet T, Orskov H, Jessen N, Lund S, Whittaker J, De Meyts P,
Larsen MB, Christiansen JS, Frystyk J 2003 A highly sensitive and specific
assay for determination of IGF-I bioactivity in human serum. Am J Physiol
Endocrinol Metab 284:E1149–E1155
21. Norrelund H, Nielsen S, Christiansen JS, Jorgensen JO, Moller N 2004
J Clin Endocrinol Metab, March 2009, 94(3):965–972 jcem.endojournals.org 971
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2014. at 04:26 For personal use only. No other uses without permission. . All rights reserved.
Modulation of basal glucose metabolism and insulin sensitivity by growth
hormone and free fatty acids during short-term fasting. Eur J Endocrinol
150:779–787
22. Gjedsted J, Gormsen LC, Nielsen S, Schmitz O, Djurhuus CB, Keiding S,
Orskov H, Tonnesen E, Moller N 2007 Effects of a 3-day fast on regional lipid
and glucose metabolism in human skeletal muscle and adipose tissue. Acta
Physiol (Oxf) 191:205–216
23. Hansen TK 2002 Pharmacokinetics and acute lipolytic actions of growth hor-
mone. Impact of age, body composition, binding proteins, and other hor-
mones. Growth Horm IGF Res 12:342–358
24. Leung KC, Doyle N, Ballesteros M, Waters MJ, Ho KK 2000 Insulin regulation
of human hepatic growth hormone receptors: divergent effects on biosynthesis
and surface translocation. J Clin Endocrinol Metab 85:4712–4720
25. Haffner D, Schaefer F, Girard J, Ritz E, Mehls O 1994 Metabolic clearance of
recombinant human growth hormone in health and chronic renal failure. J Clin
Invest 93:1163–1171
26. Edgren B, Wester PO 1971 Impairment of glomerular filtration in fasting for
obesity. Acta Med Scand 190:389–393
27. Jorgensen JO, Moller N, Lauritzen T, Alberti KG, Orskov H, Christiansen JS
1990 Evening versus morning injections of growth hormone (GH) in GH-
deficient patients: effects on 24-hour patterns of circulating hormones and
metabolites. J Clin Endocrinol Metab 70:207–214
28. Hong-Brown LQ, Brown CR, Cooney RN, Frost RA, Lang CH 2003 Sepsis-
induced muscle growth hormone resistance occurs independently of STAT5
phosphorylation. Am J Physiol Endocrinol Metab 285:E63–E72
29. Chen Y, Sun D, Krishnamurthy VM, Rabkin R 2007 Endotoxin attenuates
growth hormone-induced hepatic insulin-like growth factor I expression by
inhibiting JAK2/STAT5 signal transduction and STAT5b DNA binding. Am J
Physiol Endocrinol Metab 292:E1856–E1862
30. Schaefer F, Chen Y, Tsao T, Nouri P, Rabkin R 2001 Impaired JAK-STAT
signal transduction contributes to growth hormone resistance in chronic ure-
mia. J Clin Invest 108:467–475
31. Asada N, Takahashi Y, Wada M, Naito N, Uchida H, Ikeda M, Honjo M 2000
GH induced lipolysis stimulation in 3T3-L1 adipocytes stably expressing
hGHR: analysis on signaling pathway and activity of 20K hGH. Mol Cell
Endocrinol 162:121–129
32. Fain JN, Ihle JH, Bahouth SW 1999 Stimulation of lipolysis but not of leptin
release by growth hormone is abolished in adipose tissue from Stat5a and b
knockout mice. Biochem Biophys Res Commun 263:201–205
33. prath-Husmann I, Rohrig K, Gottschling-Zeller H, Skurk T, Scriba D, Birgel
M, Hauner H 2001 Effects of leptin on the differentiation and metabolism of
human adipocytes. Int J Obes Relat Metab Disord 25:1465–1470
34. Argiles JM, Lopez-Soriano J, Almendro V, Busquets S, Lopez-Soriano FJ 2005
Cross-talk between skeletal muscle and adipose tissue: a link with obesity?
Med Res Rev 25:49–65
972 Moller et al. GH Signaling and Action during Fasting J Clin Endocrinol Metab, March 2009, 94(3):965–972
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 January 2014. at 04:26 For personal use only. No other uses without permission. . All rights reserved.
