Relationship Between Serum IGF-1 and Skeletal Muscle IGF-1 mRNA Expression to Phosphocreatine Recovery After Exercise in Obese Men With Reduced GH

Abstract

Context:

GH and IGF-1 are believed to be physiological regulators of skeletal muscle mitochondria.

Objective:

The objective of this study was to examine the relationship between GH/IGF-1 and skeletal muscle mitochondria in obese subjects with reduced GH secretion in more detail.

Design:

Fifteen abdominally obese men with reduced GH secretion were treated for 12 weeks with recombinant human GH. Subjects underwent ³¹P-magnetic resonance spectroscopy to assess phosphocreatine (PCr) recovery as an in vivo measure of skeletal muscle mitochondrial function and percutaneous muscle biopsies to assess mRNA expression of IGF-1 and mitochondrial-related genes at baseline and 12 weeks.

Results:

At baseline, skeletal muscle IGF-1 mRNA expression was significantly associated with PCr recovery (r = 0.79; P = .01) and nuclear respiratory factor-1 (r = 0.87; P = .001), mitochondrial transcription factor A (r = 0.86; P = .001), peroxisome proliferator-activated receptor (PPAR)γ (r = 0.72; P = .02), and PPARα (r = 0.75; P = .01) mRNA expression, and trended to an association with PPARγ coactivator 1-α (r = 0.59; P = .07) mRNA expression. However, serum IGF-1 concentration was not associated with PCr recovery or any mitochondrial gene expression (all P > .10). Administration of recombinant human GH increased both serum IGF-1 (change, 218 ± 29 μg/L; P < .0001) and IGF-1 mRNA in muscle (fold change, 2.1 ± 0.3; P = .002). Increases in serum IGF-1 were associated with improvements in total body fat (r = −0.53; P = .04), trunk fat (r = −0.55; P = .03), and lean mass (r = 0.58; P = .02), but not with PCr recovery (P > .10). Conversely, increase in muscle IGF-1 mRNA was associated with improvements in PCr recovery (r = 0.74; P = .02), but not with body composition parameters (P > .10).

Conclusion:

These data demonstrate a novel association of skeletal muscle mitochondria with muscle IGF-1 mRNA expression, but independent of serum IGF-1 concentrations.

Mitochondrial dysfunction is associated with insulin resistance. For example, in subjects with obesity and non-insulin-dependent diabetes mellitus, the mitochondria are smaller and morphologically different in skeletal muscle, and mitochondrial function as determined by NADH:O₂ oxidoreductase activity is reduced (1). Furthermore, treatment with the mitochondrial toxin, stavudine, reduces both mitochondrial function and insulin sensitivity in a prospective, randomized, placebo-controlled trial (2). Interestingly, adult subjects with diabetes mellitus also have decreased expression of nuclear encoded mitochondrial genes, including peroxisome proliferator-activated receptorγ coactivator 1-α (PGC-1α) (3, 4). PGC-1α is a transcriptional regulator of mitochondrial biogenesis and function, which regulates peroxisome proliferator-activated receptor (PPAR)γ and PPARα signaling, as well as nuclear respiratory factor-1 (NRF-1) and mitochondrial transcription factor A (TFAM) (5, 6).

Recent evidence suggests that GH and IGF-1 are physiological regulators of mitochondria. For instance, acromegaly, a condition of pathological GH excess, is associated with alterations in skeletal muscle mitochondrial morphology, which resolve after surgical treatment (7). Conversely, we have previously shown that reduced GH and IGF-1 secretion in obesity is associated with reduced mitochondrial function, as represented by delayed phosphocreatine (PCr) recovery after submaximal exercise on ³¹P-magnetic resonance spectroscopy (³¹P-MRS) (8). ³¹P-MRS is a noninvasive technique to assess mitochondrial function in vivo that has been validated in both animal models and humans (9–11). ³¹P-MRS allows reproducible measurements of high-energy phosphate molecules such as PCr in vivo. PCr is normally depleted during submaximal exercise, and its recovery is determined oxidatively (12). Therefore, recovery of PCr after challenging with exercise reflects oxidative function of the mitochondria (13). In addition, acute treatment with recombinant human GH (rhGH) via 14 hours of continuous infusion in healthy subjects increases mitochondrial oxidative capacity and the mitochondrial transcription factor TFAM (14), and 6-month treatment of obese men with rhGH improves PCr recovery by ³¹P-MRS (15). Furthermore, we recently demonstrated that improvements in IGF-1 after a 12-month treatment with a GHRH analog are associated with improvements in PCr recovery (16).

These previous studies suggest a potentially physiologically significant relationship between GH/IGF-1 and skeletal muscle mitochondrial function in humans. We hypothesized that rhGH treatment of obese patients exhibiting reduced GH secretion would improve skeletal muscle mitochondrial function through a transcriptional mechanism via local GH/IGF signaling. To test this hypothesis, we recently completed a 12-week interventional physiology study with in vivo assessment of skeletal muscle mitochondrial function, utilizing ³¹P-MRS and in vitro assessment of mitochondrial gene expression in muscle tissue obtained with percutaneous biopsy techniques. We measured mRNA expression levels of IGF-1 and of several mitochondrial genes, including PGC-1α, PPARγ, PPARα, NRF-1, and TFAM. This is the first study to assess in vivo skeletal muscle mitochondrial function utilizing ³¹P-MRS and mitochondrial gene expression in response to rhGH in humans.

Subjects and Methods

Study subjects

Fifteen abdominally obese male subjects with reduced GH secretion were recruited at the Massachusetts General Hospital for a 12-week, open-label, interventional, physiological study. Eighteen- to 60-year-old men with body mass index (BMI) ≥ 30 kg/m², waist circumference ≥ 102 cm, and peak stimulated GH ≤ 4.2 μg/L after a GHRH-arginine stimulation test were included. All participants were otherwise healthy without known endocrine dysfunction. Participants who had a known diagnosis of diabetes mellitus, fasting blood glucose > 125 mg/dL, or use of antidiabetes medications were excluded. In addition, subjects with hemoglobin < 9 g/dL and creatinine > 1.5 mg/dL, as well as subjects using hormonal medication including T, glucocorticoids, anabolic steroids, GHRH, GH, or IGF-1 within 3 months of enrollment, were excluded. The study was approved by the Partners Institutional Review Board. Written informed consent was obtained from all participants.

Study design

Subjects were initially prescribed rhGH at a dose of 0.4 mg/d by daily sc injections. Doses of rhGH were titrated up to 0.6 mg/d, and then 0.8 mg/d every 4 weeks based on serum IGF-1 levels to reach the upper-normal ranges for each subject according to their age. Subjects were evaluated for PCr recovery using ³¹P-MRS and skeletal muscle gene expression by percutaneous muscle biopsy at baseline and again after 12 weeks of treatment.

³¹P-MRS protocol

In vivo skeletal muscle mitochondrial function was determined using ³¹P-MRS to assess PCr recovery after submaximal exercise as previously reported (17). Briefly, the magnetic resonance acquisition was performed after an overnight fast on a 3.0 T whole body magnetic resonance scanner (Trio; Siemens Medical Systems). Subjects were placed supine and feet first in the bore with their right lower leg inside a custom built device designed for dynamic plantar flexion exercise. A 13-cm-diameter custom-built single-tuned ³¹P surface coil was placed in contact with the posterior calf muscles. ³¹P spectra were acquired every 2 seconds over a 10-minute period, during which time subjects underwent 2 minutes of rest, 3 minutes of plantar flexion exercise at 40% of the maximal voluntary contraction (MVC) with a frequency of 0.5 Hz, followed by 5 minutes of recovery. MVC was determined before the protocol using a dynamometer (microFet2; Hoggan Health Industries). Maximum plantar flexion force was measured in triplicate at each visit, and the mean value was used for MVC calculation.

Concentrations of PCr, inorganic phosphate, and ATP resonances were fitted in the frequency domain using an in-house MATLAB-based software. Intracellular pH was estimated based on the chemical shift difference between PCr and inorganic phosphate resonances. Mitochondrial function was determined by plotting the PCr peak integrated area vs time during exercise recovery and fitting the recovery curve to a mono-exponential function to determine the recovery time constant (τPCr). The initial rate of PCr recovery (Vi_PCr) was determined from τPCr and PCr depletion using the following equation: Vi_PCr = (60/τPCr) × PCr depletion (18). Vi_PCr was chosen as the primary end-point because it normalizes PCr recovery based on participant effort and is insensitive to end-of-exercise metabolic conditions such as intracellular acidosis (18). Greater Vi_PCr represents relatively better mitochondrial function. Technically successful ³¹P- MRS scans for baseline and 12-week visits were available in 13 subjects.

Muscle biopsy

Percutaneous muscle needle biopsy was performed according to standard procedures adapted for the lateral gastrocnemius muscle after an overnight fast (19). The muscle biopsy was performed on a separate day from the MRS experiment. Subjects were instructed to lie in the prone position. The right calf was cleaned using alcohol and chlorhexidine in the lateral muscle belly area, and local anesthesia was provided using 1% lidocaine administered intradermally and sc. A 2-cm stab incision was made in the skin and soft tissue overlying the lateral gastrocnemius muscle, and a Bergstrom core biopsy needle (20) was inserted into the muscle in a 90° angle. Muscle biopsy was performed using manual suction to increase the yield of muscle tissue obtained (21). Approximately 50–100 mg of muscle tissue was obtained from each subject. Samples were immediately flash-frozen in liquid nitrogen and stored at −80°C until ready for use. Paired muscle tissue was obtained from 10 subjects at both baseline and after 12 weeks of treatment.

The MRS, biopsy, and remainder of baseline procedures were performed on separate days.

Gene expression studies

Total mRNA was extracted from the muscle biopsy samples using TRIzol (Invitrogen) following the manufacturer's instructions in one batch at the conclusion of the study. SuperScript III First-Strand Synthesis SuperMix (Invitrogen) was used for the generation of cDNA for all samples. Expression of IGF-1, PGC-1α, PPARγ, PPARα, NRF-1, and TFAM mRNA was determined by quantitative real time PCR using a Mastercycler realplex machine (Eppendorf) with iQ SYBR Green Supermix Kit (Bio-Rad) as per standard protocol. Primers for IGF-1 were chosen to assess the mRNA expression levels of all four isoforms. Expression of target-gene mRNA was normalized with glyceraldehyde-3-phosphate dehydrogenase mRNA expression. The primer sequences used for amplifying target-gene mRNA are available upon request.

Biochemical assessment

Clinically standard GHRH-arginine stimulation testing was performed after an overnight fast as previously described (22). Serum GH was measured using the Beckman Access Ultrasensitive human GH assay, a paramagnetic particle, and chemiluminescent immunoassay (Beckman Coulter). Serum IGF-1 was measured using the IMMULITE 2000 assay (Siemens Diagnostics). A 75-g oral glucose tolerance test was performed at baseline and at the 3-month visit. Plasma glucose was assessed by the commercial laboratory LabCorp, Inc, and serum insulin was assessed using a chemiluminescence immunoassay (Access Immunoassay System; Beckman Coulter). Homeostatic model assessment of insulin resistance (HOMA-IR) was calculated from fasting glucose and insulin using the following equation: HOMA-IR = (fasting glucose (mmol/L) × fasting insulin (mIU/L))/22.5 (23). Levels of high-sensitivity C-reactive protein (hsCRP) were assessed at the Massachusetts General Hospital clinical laboratory.

Body composition

Height and body weight were measured after an overnight fast. Total lean body mass (LBM) and fat mass were determined by dual-energy x-ray absorptiometry (DXA) (Discovery A; Hologic, Inc). Measurements of regional lean mass, specifically the right leg lean mass, using DXA were standardized (1995 User's Guide; Hologic, Inc). The technique has a precision error (1 SD) of 1.5% for lean mass (24).

Indirect calorimetry

Resting energy expenditure (REE) was assessed by determining oxygen consumption (VO₂) and carbon dioxide production (VCO₂) for 20 minutes using a calibrated calorimeter (VMAX29N; Sensormedics) after a 20-minute rest. REE was normalized to LBM as determined by DXA.

Physical activity assessment

The level of physical activity was assessed by self-reported regular structured exercise during direct interview and expressed as total hours of activity per week (metabolic equivalents) and hours of TV watching per day.

Statistical analysis

Baseline variables are presented as mean ± SEM. Univariate regression analysis using Pearson's correlation coefficient was performed at baseline to determine the relationship between serum IGF-1 and skeletal muscle mRNA expression of IGF-1 to PCr recovery parameters before initiating treatment. In addition, further univariate regression analyses were performed relating serum IGF-1 and skeletal muscle IGF-1 mRNA expression to the expression levels of various metabolic genes within skeletal muscle at baseline. The effects of rhGH over 12 weeks were determined by paired Student's t test. Additional univariate regression analyses with Pearson's correlation coefficient were performed relating changes in circulating IGF-1 and skeletal muscle IGF-1 mRNA expression to changes in PCr recovery parameters as well as body composition and metabolic parameters. Statistical analysis was performed using JMP Statistical Database Software (version 10.0.0; SAS Institute, Inc). Statistical significance was determined as a value of P < .05.

Results

Baseline clinical characteristics of study subjects

The clinical characteristics of the subjects are presented in Table 1. Fifteen men were enrolled in the study. The subjects were on average 47.9 ± 2.2 years old (range, 31–59 y), with a mean BMI of 37.8 ± 5.9 kg/m², waist circumference of 124 ± 3 cm, peak stimulated serum GH of 2.7 ± 0.4 μg/L, and serum IGF-1 of 142 ± 9 μg/L. All subjects had a normal ankle-brachial index (ABI) (all > 0.90).

Table 1.

Baseline Characteristics of Study Subjects

No. of subjects	15
Age, y	47.9 ± 2.2
Race (Caucasian), n (%)	10 (66.7)
Ethnicity (Hispanic), n (%)	1 (6.7)
Tobacco use, pack years	2.5 ± 1.6
Pulse	74 ± 3
Systolic blood pressure, mm Hg	131 ± 3
Diastolic blood pressure, mm Hg	82 ± 2
ABI	1.2 ± 0.05
BMI, kg/m2	37.8 ± 5.9
Waist circumference, cm	124 ± 3
Fasting GH, μg/L	0.06 ± 0.01
Peak stimulated GH on GHRH-arginine, μg/L	2.7 ± 0.4
IGF-1, μg/L	142 ± 9
Hours of TV watching/d	3.03 ± 0.77
Metabolic activity/d	54.3 ± 18.0
ViPCr, mm/min	17.65 ± 1.52
PCr depletion, %	30.85 ± 2.08

Data are expressed as mean ± SEM, unless otherwise specified.

Baseline relationship between serum IGF-1 and skeletal muscle IGF-1 mRNA to PCr recovery

At baseline, before intervention with rhGH, serum IGF-1 was not associated with the PCr recovery parameter Vi_PCr (r = −0.16; P = .60). In contrast, skeletal muscle IGF-1 mRNA expression level was significantly associated with Vi_PCr (r = 0.79; P = .01) such that greater skeletal muscle IGF-1 mRNA expression was associated with better PCr recovery, a marker for better mitochondrial function (Table 2). Neither serum IGF-1 nor IGF-1 mRNA expression in muscle was associated with resting [PCr], pH, or [ADP] (all P > .10).

Table 2.

Baseline Relationship Between Serum IGF-1 and Skeletal Muscle IGF-1 mRNA Expression to MRS and Gene Expression Parameters

	Serum IGF-1		Skeletal Muscle IGF-1 mRNA
	r	P	r	P
ViPCr	−0.16	.60	0.79	0.01
[PCr] at rest	−0.08	.79	0.20	0.58
pH at rest	−0.14	.63	−0.42	.23
[ADP] at rest	0.004	.99	−0.28	.44
PGC-1α mRNA	0.20	.59	0.59	.07
PPARγ mRNA	0.23	.52	0.72	.02
PPARα mRNA	0.35	.32	0.75	.01
NRF-1 mRNA	0.07	.86	0.87	.001
TFAM mRNA	0.30	.39	0.86	.001

Statistical analyses were performed with Pearson's univariate regression analyses.

In addition, skeletal muscle IGF-1 mRNA expression was significantly associated with mRNA expression levels of PPARγ (r = 0.72; P = .02), PPARα (r = 0.75; P = .01), NRF-1 (r = 0.87; P = .001), and TFAM (r = 0.86; P = .001) and trended to an association with PGC-1α mRNA (r = 0.59; P = .07) (Supplemental Figure 1), whereas serum IGF-1 was not associated with PPARγ, PPARα, NRF-1, TFAM, or PGC-1α expression levels in skeletal muscle (all P > .10) (Table 2). Neither serum IGF-1 nor IGF-1 mRNA expression in muscle was significantly associated with BMI, waist circumference, or homeostatic model assessment (HOMA) at baseline (all P > .10).

Effects of rhGH treatment

rhGH treatment for 12 weeks resulted in increased serum IGF-1 (baseline, 142 ± 9; 12-week, 359 ± 29; change, 218 ± 29 μg/L; P < .0001) and IGF-1 mRNA expression in skeletal muscle (baseline, 1 ± 0.2; 12-week, 1.9 ± 0.3; fold change, 2.1 ± 0.3; P = .002) (Figure 1, A and B).

Figure 1.

A, Effects of rhGH on serum IGF-1 levels over 12 weeks, with mean values indicated by the solid line. Statistical significance was tested by paired Student's t test comparing baseline to 12-week IGF-1 levels. B, Effects of rhGH on skeletal muscle IGF-1 mRNA expression measured by quantitative RT-PCR and normalized to baseline. Error bars represent SEM. Significance was tested by paired Student's t test.

rhGH treatment decreased waist circumference (change, −3 ± 1 cm; P = .03) but did not change body weight or BMI (both P > .05). rhGH decreased hsCRP (change, −1.78 ± 0.57 mg/L; P = .008). Fasting glucose (change, 6 ± 3 mg/dL; P = .04), fasting insulin (change, 6.9 ± 2.7 μIU/mL; P = .02), and HOMA (change, 1.9 ± 0.8; P = .03) increased during the study, but 2-hour glucose did not change (P > .10). The 12-week treatment with rhGH also increased REE (change, 166 ± 51 kcal/d; P = .006) as well as REE/LBM (change, 2.30 ± 0.77 kcal/d/kg; P = .01), VO₂ (change, 0.03 ± 0.008 L/min; P = .005), and VCO₂ (change, 0.02 ± 0.008 L/m; P = .05) (Table 3).

Table 3.

Effects of 12-Week Treatment With rhGH

	Baseline	3 mo	Change	P
Weight, kg	116.3 ± 4.7	116.1 ± 4.9	−0.2 ± 0.6	.69
BMI, kg/m2	37.8 ± 1.5	37.7 ± 1.6	−0.1 ± 0.2	.52
Waist circumference, cm	124 ± 3	121 ± 3	−3 ± 1	.03
Total fat mass, kg	44.4 ± 3.1	43.9 ± 3.1	−0.5 ± 0.7	.47
Total LBM, kg	70.3 ± 1.8	70.7 ± 2	0.4 ± 0.7	.56
Right lower extremity lean mass, kg	12.0 ± 0.4	12.0 ± 0.4	−0.05 ± 0.1	.73
Fasting glucose, mg/dL	88 ± 2	94 ± 4	6 ± 3	.04
2-h Glucose, mg/dL	120 ± 9	134 ± 12	14 ± 10	.16
Fasting insulin, μIU/mL	8.9 ± 1.3	15.8 ± 2.8	6.9 ± 2.7	.02
HOMA	2 ± 0.3	3.9 ± 0.8	1.9 ± 0.8	.03
hsCRP, mg/L	3.53 ± 0.44	1.74 ± 0.40	−1.78 ± 0.57	.008
REE, kcal/d	1757 ± 65	1922 ± 55	166 ± 51	.006
REE/LBM, kcal/d/kg	24.98 ± 0.71	27.28 ± 0.68	2.30 ± 0.77	.01
Respiratory quotient	0.88 ± 0.01	0.86 ± 0.02	−0.02 ± 0.02	.29
VO2, L/min	0.25 ± 0.009	0.27 ± 0.008	0.03 ± 0.008	.005
VCO2, L/min	0.22 ± 0.009	0.24 ± 0.009	0.02 ± 0.008	.05

Statistical analyses were performed with paired Student's t test.

rhGH treatment had no effect on the PCr recovery parameter Vi_PCr (P > .10). Effect of rhGH on skeletal muscle gene expression is reported in Supplemental Figure 2.

There was no change in ABI or physical activity during the study as determined by metabolic equivalents of self-reported activity or hours of TV watching (all P > .10).

Relationship between changes in serum IGF-1, skeletal muscle IGF-1 mRNA expression, and changes in metabolic parameters using univariate regression

Univariate regression analysis among all available subjects demonstrated that change in serum IGF-1 was not significantly associated with change in skeletal muscle IGF-1 mRNA expression (r = 0.23; P = .53).

Increase in serum IGF-1 was significantly associated with decreases in waist circumference (r = −0.82; P = .0002), total fat mass (r = −0.53; P = .04), and trunk fat mass (r = −0.55; P = .03) and increases in fat-free mass (r = 0.58; P = .02). In contrast, change in skeletal muscle IGF-1 expression was not related to changes in any body composition parameters (Table 4).

Table 4.

Relationship Between Changes in Serum IGF-1 and Changes in Skeletal Muscle IGF-1 mRNA Expression Level to Changes in Metabolic Parameters

	Change in Serum IGF-1		Change in Muscle IGF-1 mRNA
	r	P	r	P
Body weight	0.03	.91	0.17	.64
BMI	−0.18	.52	0.24	.50
Waist circumference	−0.82	.0002	−0.06	.87
Total LBM	0.58	.02	0.20	.59
Total fat mass	−0.53	.04	−0.07	.85
Trunk fat mass	−0.55	.03	0.12	.75
REE	0.15	.60	0.54	.11
REE/LBM	0.052	.85	0.47	.17
VO2	0.15	.61	0.60	.07
VCO2	0.096	.73	0.26	.48

Statistical analyses were performed with Pearson's univariate regression analyses.

Changes in serum IGF-1 were not related to Vi_PCr (P > .10). However, changes in skeletal muscle IGF-1 mRNA expression were significantly related to improvements in Vi_PCr (r = 0.74; P = .02), such that increased IGF-1 mRNA expression associated with better PCr recovery, a marker for better mitochondrial function (Figure 2, A and B).

Figure 2.

Univariate regression analyses relating change in skeletal muscle IGF-1 mRNA expression to change in PCr recovery parameter, Vi_PCr (A), and change in serum IGF-1 to change in PCr recovery parameter Vi_PCr (B).

After a 12-week treatment with rhGH, increases in skeletal muscle IGF-1 mRNA expression trended to an association with increases in VO₂ (r = 0.60; P = .07), whereas serum IGF-1 was not significantly associated with changes in REE, REE/LBM, VO₂, or VCO₂ (all P > .50).

Discussion

This study demonstrates, for the first time, that skeletal muscle IGF-1 mRNA expression, but not serum IGF-1, is significantly related to skeletal muscle mitochondrial function. Skeletal muscle IGF-1 mRNA expression at baseline was related to the PCr recovery parameter Vi_PCr, as well as to PGC-1α, PPARγ, PPARα, NRF-1, and TFAM mRNA expression levels. In addition, increases in skeletal muscle IGF-1 mRNA expression with rhGH treatment were significantly related to improvements in PCr recovery, suggestive of improving mitochondrial function in the muscle. In contrast, serum IGF-1 was related to neither PCr recovery nor mitochondrial gene expression levels at baseline, and changes in serum IGF-1 were not related to improvements in PCr recovery.

These results suggest a possible role for local paracrine effects of skeletal muscle IGF-1 on skeletal muscle mitochondrial function. Although significant attention is focused on the endocrine effects of IGF-1 produced by the liver, GH also stimulates local or autocrine/paracrine synthesis of IGF-1 in almost all tissues, including skeletal muscle (25, 26). Local or paracrine IGF-1 signaling appears to play a significant role in growth and development because liver-specific IGF-1 knockout mice display normal body weight, length, and development despite having significantly reduced circulating IGF-1 levels (27). In a transgenic mouse model of IGF-1 overexpression in cardiac myocytes, animals were protected against reactive oxygen species, mitochondrial damage, and cardiac malfunction caused by high-fat diet (28). Because plasma IGF-1 also increased in these transgenic mice, we cannot rule out a possible contribution of circulating IGF-1 in these metabolic effects. Nonetheless, this result suggests a significant role for local IGF-1 in metabolic homeostasis in the heart. Consistently, our results suggest for the first time that local IGF-1 signaling improves mitochondrial function in skeletal muscle as well in human subjects.

An alternative interpretation of our data is a direct and independent effect of GH on skeletal muscle, liver, and adipose tissue—specifically, a direct effect of GH to increase skeletal muscle mitochondrial gene expression and function. This is consistent with the known anabolic effects of GH on muscle. The increase in skeletal muscle IGF-1 mRNA expression in this instance may represent GH action in this particular tissue, but it does not necessarily indicate an effect mediated through IGF-1 as opposed to direct GH actions. A direct effect of GH on muscle mitochondria is consistent with the significant relationship between IGF-1 mRNA and mitochondrial function and gene expression observed at baseline, before administration of the rhGH. Regardless of whether changes in skeletal muscle mitochondria are caused directly by GH or mediated by IGF-1, our results indicate that the effects of rhGH on skeletal muscle mitochondria may occur independently of its effects on other tissues (ie, independent from the effects of rhGH to increase IGF-1 synthesis and secretion from the liver, because changes in serum IGF-1 were not related to changes in skeletal muscle mitochondrial function).

rhGH appears to have a physiological benefit on skeletal muscle contraction, relaxation, and muscle mass in a mouse model of congestive heart failure created by monocrotaline treatment (29), potentially through a PGC-1α-mediated process (30). As discussed previously, Short et al (14) demonstrated that acute rhGH infusion stimulates the expression of nuclear-encoded mitochondrial gene TFAM in healthy subjects. In this study, we confirm this relationship between GH/IGF-1- and PGC-1α-mediated transcription factors within skeletal muscle, with significant baseline association between IGF-1 mRNA and TFAM, as well as NRF-1, PPARγ, and PPARα (and a trend to association with PGC-1α).

Our study shows that rhGH increases serum IGF-1 and skeletal muscle IGF-1 mRNA levels as expected, demonstrating subject compliance with the rhGH treatment during the study. However, changes in serum IGF-1 were not significantly associated with changes in skeletal muscle IGF-1 mRNA expression. Although this differential effect is difficult to explore further in this small study, future investigations distinguishing between endocrine effects of circulating IGF-1 and paracrine effects of locally expressed IGF-1 are warranted.

Our study also confirms the effects of rhGH on reducing waist circumference and improving REE, as previously demonstrated in obese subjects (31–34). However, we were unable to demonstrate an effect of rhGH on fat mass or LBM as previously reported in a large meta-analysis of 539 obese subjects (35). Because this was a small physiology study whose primary end-point was not a change in body composition, this may simply reflect a lack of power in our study to assess this particular end-point. Despite the lack of significant effects on fat mass and lean mass, interestingly, increases in serum IGF-1 in response to rhGH were significantly associated with improvements in body composition (decreasing body fat, trunk fat, and improving fat-free mass). Therefore, the greater the increase in serum IGF-1, the greater the improvements in body composition. This finding also suggests that the lack of statistical effect of rhGH on body composition parameters may be due to those subjects whose serum IGF-1 levels did not increase enough and therefore did not have the expected improvements in body composition. In contrast, a change in skeletal muscle IGF-1 mRNA was not associated with changes in body composition, further highlighting the differential effects between serum IGF-1 and IGF-1 mRNA expression in skeletal muscle. Conversely, change in serum IGF-1 was not associated with improvements in REE, whereas change in skeletal muscle IGF-1 mRNA expression trended to an association with improving resting VO₂. This suggests that the effects of rhGH on VO₂ and REE may be mediated by improved skeletal muscle mitochondrial function, effected through local IGF-1 effects. This is consistent with prior work from Zurlo et al (36), who demonstrated that skeletal muscle metabolism accounts for significant variance in basal and sleeping metabolic rate.

Short-term treatment with rhGH decreased insulin sensitivity, consistent with prior studies (14, 37, 38). In addition to its well-known anabolic effects, GH is also known to be a counter-regulatory hormone to insulin with the ability to raise plasma glucose. GH can lead to insulin resistance through its effects to promote lipolysis from adipocytes, promote gluconeogenesis in the liver, and inhibit peripheral glucose disposal in the muscle (39, 40). Previous human studies of rhGH in obese subjects have demonstrated a negative effect of rhGH on insulin sensitivity in the short term. However, in a longer term study lasting 9 months, Johannsson et al (31) demonstrated a beneficial effect of rhGH on insulin sensitivity as determined by the glucose disposal rate on a hyperinsulinemic-euglycemic clamp. This improvement in glucose disposal rate was attributed to the effect of rhGH to reduce visceral adipose tissue, which is thought to be detrimental to metabolic health. However, 80% of insulin-stimulated glucose uptake occurs in the muscle (41); therefore, the improvement in glucose disposal rate determined via the hyperinsulinemic-euglycemic clamp mostly indicates an improvement in skeletal muscle insulin sensitivity. The current study may not have been long enough to demonstrate a concomitant improvement in insulin sensitivity with rhGH treatment. Nonetheless, we may also need to consider the possible beneficial effects of rhGH on skeletal muscle mitochondrial function as a contributor to the improved glucose disposal rate after longer term treatment with rhGH in obese subjects. The long-term effect of rhGH on insulin sensitivity is most likely determined by the sum of its effects on various tissues, including muscle, adipose, and liver.

In this study, rhGH improved the inflammatory marker hsCRP. This is consistent with previous interventional studies utilizing rhGH (15, 42) and our own work utilizing a synthetic GHRH analog in obesity (43) that showed that treatment with rhGH or GHRH resulted in a significant reduction in hsCRP.

Although we did not see a significant effect of rhGH on PCr recovery parameters after 12 weeks of treatment, this was most likely due to a lack of sufficient power from the small sample size (n = 15). This was a pilot and feasibility study designed with detailed and invasive experimental techniques by intention. A larger study may be needed to demonstrate a statistically significant effect of rhGH on PCr recovery. Subjects may also have had variable compliance throughout the study based on the large variability in the subjects' individual IGF-1 response. In addition, the dose of rhGH was chosen to target IGF-1 levels in the upper normal range adjusted for age. However, this strategy resulted in exposure to different doses of rhGH at different times for each individual, which may be a source of variability. Furthermore, the optimal dosing of rhGH in this context is not yet known. Moreover, whereas previous studies by Bredella et al (15) and Makimura et al (16) were 6 and 12 months long, this was a 12-week study. The ideal duration for treatment is unknown, and additional studies are needed to identify the differences in short-term and long-term GH treatment on mitochondrial function in the skeletal muscle. The subjects' physical activity was assessed by self-reported questionnaires, which may be less accurate among obese subjects (44) who tend to over-report their activity (45, 46). Lastly, rhGH treatment results in a nonpulsatile, nonphysiological increase in GH, and the role of physiological pulsatile GH secretion in this regard has also not been investigated in great detail.

In summary, this is the first study to demonstrate a significant association between skeletal muscle IGF-1 mRNA expression and the PCr recovery parameter Vi_PCr, as well as between muscle IGF-1 mRNA expression and expression of mitochondrial genes PGC-1α, PPARγ, PPARα, NRF-1, and TFAM. This study suggests differential effects of endocrine IGF-1 on body composition and local paracrine effects of skeletal muscle IGF-1 on muscle mitochondrial function and potentially on REE. These observations have important implications with regard to the physiological relationship between GH/IGF-1 and mitochondrial function in obese men, and future studies will be necessary to examine the mechanistic basis of this relationship in more detail. Further studies will also be necessary to investigate ways to optimize GH/IGF-1 and whether such strategies may be able to improve skeletal muscle mitochondrial function while minimizing its adverse effects.