The Protein-Retaining Effects of Growth Hormone During Fasting Involve Inhibition of Muscle-Protein Breakdown

The study population comprised eight healthy males (mean ± SE age 23 ± 0.4 years, BMI 23 ± 0.8 kg/m²), who were examined on four occasions. Before the study, subjects were on a weight-maintaining diet for at least 3 months. All subjects gave their written informed consent, and the study was approved by the regional ethics committee and conducted according to the Declaration of Helsinki.

All subjects were instructed to maintain (for at least 3 days) their normal diet before admission; the participants were allocated to 1) basal postabsorptive investigations (basal), 2) 40 h of fasting (fast), 3) 40 h of fasting with suppression of endogenous GH secretion with somatostatin (fast-GH), or 4) 40 h of fasting with suppression of endogenous GH secretion and replacement of GH by infusion (fast+GH). During fasting with GH suppression, somatostatin, insulin, and glucagon were infused for the last 28 h. During fasting with GH replacement, GH was added to the protocol for the last 28 h. Somatostatin (200 μg/h) was infused together with insulin (Actrapid 100 I.E./ml, Novo Nordisk; infusion rate adjusted to maintain plasma glucose level between 3 and 5 mmol/l [0.15 and 0 mU · kg^-1 · min^-1]) and glucagon (GlucaGen 1 mg/ml, Novo Nordisk; the infusion rate adjusted to maintain stable glucose levels [1.0-1.5 ng · kg^-1 · min^-1]). The two somatostatin experiments were identical in terms of hormone replacement doses, except for GH (Norditropin 12 U/ml, Novo Nordisk; 4.5 U partly as bolus injections [0.375 U every fourth hour for 28 h], partly as continuous infusion [0.08 U/h]). To avoid hypoglycemia during fasting without GH, six participants temporarily received glucose intravenously (between 0.3 and 0.5 mg · kg^-1 · min^-1 for a period of 0.5-2.5 h); exactly the same amount of glucose was given in exactly the same manner in the experiment with GH replacement. Because potential hypoglycemia was anticipated during GH suppression, the fast-GH experiments were always conducted before fast+GH to allow for control of exogenous glucose. The subjects were blinded to this procedure. Otherwise, the sequence of experiments was random, with at least 6 weeks elapsing between each experiment. Somatostatin, insulin, glucagon, and GH infusions were continued for a total of 28 h. During the last 4 h, substrate metabolism was investigated. No participants received glucose during these last 4 h; glucose infusion was tapered at least 2 h before investigations of substrate metabolism were started. Blood was sampled frequently during the fast, and data are based on triplicate measurements within the last 30 min of the study. Urine for determination of urea was collected over a 16-h period in the postabsorptive state. In the three 40-h experiments, urine was collected from 0-24 h and while substrate metabolism was investigated (24-28 h). During fasting, only tap water was allowed.

Measurements. All studies were performed with the participants in the supine position. Catheters for measurements of forearm arteriovenous substrate balances were placed as previously described (26). In brief, one catheter was placed retrogradely in a deep antecubital vein, and one catheter was inserted retrogradely in a heated contralateral dorsal hand vein. Criteria for correct positioning were oxygen saturations <70% and >91%, respectively. A third catheter was placed antegradely in an antecubital vein of the heated hand for infusions. Before each deep venous sample, total ipsilateral forearm blood flow was determined by means of venous occlusion plethysmography (27). Hand blood flow was interrupted by a wrist cuff inflated to a pressure of 250 mmHg immediately before each blood flow determination and 1 min before each deep venous sample.

At 8:00 A.M., priming doses of [3-³H]glucose (New England Nuclear, Boston, MA) (20 μCi), L-[¹⁵N]phenylalanine (0.7 mg/kg), L-[²H₄]tyrosine (0.5 mg/kg), and L-[¹⁵N]tyrosine (0.3 mg/kg) (Cambridge Isotope Laboratories) were given to accomplish an early plateau. A continuous infusion of [3-³H]glucose (20 μCi/h), L-[¹⁵N]phenylalanine (0.7 mg · kg^-1 · h^-1), and L-[²H₄]tyrosine (0.5 mg · kg^-1 · h^-1) was started and maintained for 4 h. The chemical, isotopic, and optical purity of the isotopes was tested before use. Solutions were prepared under sterile conditions and were shown to be free of bacteria and pyrogen before use. The specific activity of tritiated glucose was assayed as previously described (19). L-[¹⁵N]phenylalanine, L-[²H₄]tyrosine, and L-[¹⁵N]tyrosine were measured as their t-butyldimethylsilyl ether derivatives under electron ionization conditions (28). Plasma concentrations of amino acids were determined by high-performance liquid chromatography (HP 1090 series 2 HPLC, 1046 fluorescence detector and cooling system) with precolumn O-phthalaldehyde derivatization (29). In addition, concentrations of phenylalanine and tyrosine were measured by mass spectrometry using β-methylphenylalanine and α-methyltyrosine, respectively, as internal standards (28). Fat-free mass (FFM), comprising both muscle tissue and nonmuscle fat-free tissue, was estimated by means of dual-energy X-ray absorptiometry scanning (Hologic QDR-1000/W, Waltham, MA). Plasma glucose levels were measured in duplicate immediately after sampling on a glucose analyzer (Beckman Instruments, Palo Alto, CA). A double monoclonal immunofluorometric assay (Delfia, Wallac, Finland) was used to measure serum GH. Plasma glucagon levels (30) and serum C-peptide (Immunoclear, Stillwater, MN) were measured by radioimmunoassay. Total and free IGF-I were determined as previously described (31,32). Insulin was determined by a commercial enzyme-linked immunosorbent assay (Dako, Glostrup, Denmark). Catecholamines were measured by liquid chromatography (33). Urea excretion was determined by an indophenol method and serum urea by a commercial kit (Cobas Integra, Roche, Hvidovre, Denmark). Cortisol was measured by an automated chemiluminescence system (Chiron Diagnostics, Fernwald, Germany). FFAs were determined by a colorimetric method using a commercial kit (Wako Chemicals, Neuss, Germany). Total and free hormone concentrations of T₄ and T₃ were measured in serum, as previously described (34). Thyrotrophin was measured in a solid-phase, two-site chemiluminescent enzyme immunometric assay (Immulite, DPC, Los Angeles). Indirect calorimetry (Deltatrac monitor, Datex Instrumentarium, Helsinki, Finland) was performed for 30 min (10:30 to 11:00 A.M.), allowing measurements of respiratory exchange ratios (RERs). The initial 5 min of calorimetry was used for acclimatization, and calculations were based on mean values of 25 measurements of 1 min each. Likewise, estimated rates of lipid and glucose oxidation and total energy expenditure were obtained by indirect calorimetry after correction for protein oxidation, as estimated by urinary excretion of urea (35). Net nonoxidative glucose disposal was calculated by subtracting glucose oxidation from the isotopically assayed total glucose disposal. Urea accumulation was determined assuming immediate dispersion of urea from the blood to total body water (TBW), which was estimated by the formula (36):

where Y is age (years), BW is body weight (kg), and BH is body height (cm). Urea nitrogen synthesis rate (UNSR) was determined as urinary excretion of urea corrected for accumulation of urea in total body water and for hydrolysis in gut (37). UNSR was calculated as urinary excretion rate (E), corrected for accumulation (A) in TBW and for the fractional intestinal loss (L):

where E = (urine flow, l/h) × (urinary urea nitrogen, mmol/l) and A = (change in blood urea nitrogen, mmol · l^-1 · h^-1) × (TBW) (l). L (fractional intestinal loss) was estimated to be 0.14.

Phenylalanine kinetics. For measurements of whole body phenylalanine kinetics, the equations of Thompson et al. (38) were used. For regional phenylalanine kinetics, our previously described equations (28) were used. Phenylalanine flux (Q_p) and tyrosine flux (Q_t) were calculated as follows:

in which i is the rate of tracer infusion (mmol · kg^-1 · h^-1), and E_i and E_p are enrichment of the tracer infused and plasma enrichment of the tracer at isotopic plateau, respectively.

The rate of phenylalanine conversion to tyrosine (I_pt) was calculated as follows:

where [¹⁵N]Tyr_ei and [¹⁵N]Phe_ei are the isotopic enrichments of the respective tracers in plasma and I_p is the infusion rate of [¹⁵N]phenylalanine (mmol · kg^-1 · h^-1).

Phenylalanine incorporation into protein was calculated by subtracting I_pt from Q_p, because phenylalanine is irreversibly lost either by conversion into tyrosine or by incorporation into protein.

In the forearm study, phenylalanine balance (PheBal) was calculated as follows:

in which Phe_A and Phe_V are phenylalanine concentrations in arteries and veins and F is blood flow.

Regional protein breakdown represented by phenylalanine rate of appearance (Ra Phe) was calculated as follows (22):

in which Phe_EA and Phe_EV represent phenylalanine isotopic enrichment in arteries and veins, respectively.

Local rate of disappearance was calculated as: Rd Phe = Phe Bal + Ra Phe.

Statistics. Results are expressed as means ± SE. The Kolmogorov-Smirnov test was used to test for normal distribution. Statistical comparisons between the study periods (basal, fast, fast-GH, fast+GH) were assessed by a two-way analysis of variance. If this test was positive, the Student-Newman-Keuls test was used for post hoc analysis. A P-value below 0.05 was considered significant. Unless specified otherwise, data referred to below were obtained with arterialized blood based on triplicate measurements during the last 30 min of the experiments.

Circulating hormones and metabolites. The level of GH was significantly lower postabsorptively and after 40 h of fasting with GH suppression than after the plain fasting and GH replacement situations (Table 1). Notably, both total IGF-I and free IGF-I concentrations were comparable during fasting and during GH replacement, whereas IGF-I and free IGF-I decreased 35 and 70%, respectively, during fasting without GH. Insulin and C-peptide levels were slightly lower during somatostatin infusions. Glucagon and cortisol were significantly increased during fasting. Epinephrine increased after 40 h of fasting and thyroid parameters decreased, but no difference was found with or without GH substitution.

Levels of FFAs increased during fasting, but fasting without GH caused a relative decrease. The level of plasma glucose was significantly lower during ordinary fasting and fasting with GH suppression.

Creatinine levels remained identical in all four situations (basal 78 ± 5, fast 83 ± 4, fast-GH 82 ± 7, and fast+GH 83 ± 3 μmol/l; P > 0.05).

Whole-body protein metabolism. Plasma concentrations of amino acids are shown in Table 2. From the postabsorptive state to 40 h of fasting, there was a significant increase in plasma concentrations of valine, isoleucine, and leucine. Serine decreased during fasting with GH suppression, whereas tyrosine, valine, phenylalanine, isoleucine, and leucine increased compared with fasting with GH.

Urinary urea excretion (basal 392 ± 44, fast 440 ± 32, fast-GH 609 ± 76, and fast+GH 408 ± 36 mmol/24 h; P < 0.05) and serum urea (basal 4.6 ± 0.1, fast 6.2 ± 0.1, fast-GH 7.0 ± 0.2, fast+GH 4.3 ± 0.2 mmol/l; P < 0.01) were significantly increased when participants fasted without GH. UNSR increased after 40 h of fasting and was further increased by GH suppression (basal 11.4 ± 1.9, fast 21.0 ± 1.6, fast-GH 31.8 ± 4.3, fast+GH 17.6 ± 2.1 mmol/h; P < 0.01) (Fig. 1). Nitrogen excretion was significantly increased accordingly during fasting with GH suppression (basal 12.2 ± 1.4, fast 13.7 ± 1.0, fast-GH 19.0 ± 2.4, fast+GH 12.7 ± 1.1 g/24 h; P < 0.01); in addition, nitrogen excretion was increased during fasting with GH suppression in the preceding 24 h (fast 12.5 ± 1.0, fast-GH 14.3 ± 0.6, fast+GH 10.4 ± 0.8 g/24 h; P < 0.01). During fasting, a rise in nitrogen excretion was seen (P < 0.05).

Phenylalanine flux (Fig. 2) was increased after 40 h of fasting compared with the postabsorptive state (basal 50.9 ± 2.2, fast 59.0 ± 2.7, fast-GH 61.8 ± 1.6, fast+GH 60.5 ± 1.8 mmol · kg FFM^-1 · h^-1; P < 0.01), whereas tyrosine flux remained unaffected. Phenylalanine conversion to tyrosine did not change during fasting (basal 5.2 ± 0.8, fast 3.8 ± 0.3, fast-GH 4.6 ± 0.3, fast+GH 4.4 ± 0.4 mmol · kg FFM^-1 · h^-1; P > 0.05), whereas protein synthesis (phenylalanine disposal not accounted for by phenylalanine conversion to tyrosine) was higher (basal 45.7 ± 2.0, fast 55.2 ± 2.9, fast-GH 57.2 ± 1.7, fast+GH 56.0 ± 2.1 mmol · kg FFM^-1 · h^-1; P < 0.01). Phenylalanine clearance rate was decreased after fasting (basal 100 ± 7, fast 72 ± 5, fast-GH 78 ± 6, fast+GH 83 ± 8 ml/min; P < 0.01).

Forearm protein metabolism. Forearm blood flow increased after 40 h of fasting (basal 1.8 ± 0.1, fast 3.8 ± 0.3, fast-GH 3.7 ± 0.3, fast+GH 3.3 ± 0.3 ml · 100 ml^-1 · min^-1; P < 0.01) (Fig. 3). Forearm muscle exhibited a net release of phenylalanine, and the negative phenylalanine balance was higher during fasting with GH suppression than during fasting with GH replacement and fasting alone (basal 9 ± 3, fast 15 ± 6, fast-GH 17 ± 4, fast+GH 11 ± 5 nmol/min; P < 0.05). Muscle protein breakdown, represented by phenylalanine rate of appearance, was highest among participants fasted without GH (basal 17 ± 4, fast 26 ± 9, fast-GH 33 ± 7, fast+GH 25 ± 6 nmol/min; P < 0.05). Incorporation of phenylalanine into muscle, represented by phenylalanine rate of disappearance, was increased after fasting (basal 8 ± 2, fast 11 ± 6, fast-GH 16 ± 4, fast+GH 14 ± 3 nmol/min; P < 0.05). In general, forearm release of amino acids tended to be higher during fasting without GH (sum of forearm differences: basal 384 ± 267, fast -292 ± 318, fast-GH 733 ± 275, fast+GH, 329 ± 261 μmol/l; P = 0.11).

Indirect calorimetry and glucose turnover. Energy expenditure, measured by indirect calorimetry, increased after 40 h of fasting (basal 1,713 ± 59, fast 1,944 ± 44, fast-GH 1,916 ± 78, fast+GH 1,879 ± 51 kcal/24 h; P < 0.05) (Fig. 4). The RER decreased with fasting (basal 0.84 ± 0.01, fast 0.77 ± 0.01, fast-GH 0.79 ± 0.01, fast+GH 0.78 ± 0.01; P < 0.01). Protein oxidation was higher during GH suppression (basal 0.70 ± 0.08, fast 0.81 ± 0.08, fast-GH 1.10 ± 0.15, fast+GH 0.74 ± 0.08 mg · kg^-1 · min^-1; P < 0.01), whereas lipid oxidation was significantly lower (basal 0.65 ± 0.05, fast 1.16 ± 0.05, fast-GH 0.89 ± 0.07, fast+GH 1.05 ± 0.07 mg · kg^-1 · min^-1; P < 0.01) (Fig. 4). Glucose turnover (basal 1.98 ± 0.08, fast 1.46 ± 0.06, fast-GH 1.44 ± 0.07, fast+GH 1.49 ± 0.07 mg · kg^-1 · min^-1; P < 0.01) and glucose oxidation (basal 1.57 ± 0.13, fast 0.79 ± 0.13, fast-GH 0.91 ± 0.18, fast+GH 0.95 ± 0.11 mg · kg^-1 · min^-1; P < 0.01) decreased during fasting. The ratio between glucose oxidation and total glucose turnover was also significantly increased postabsorptively (basal 0.79 ± 0.06, fast 0.54 ± 0.09, fast-GH 0.63 ± 0.1, fast+GH 0.64 ± 0.07; P < 0.05) (Fig. 4).

The present study was designed to assess the effects of short-term fasting on substrate metabolism with particular reference to the effects of GH. The main findings of the study are that 1) GH deprivation during a 40-h fast increases production of urea nitrogen by 50%; 2) GH replacement restores normal urea nitrogen generation; and 3) the underlying mechanisms involve GH-mediated inhibition of muscle protein breakdown.

As is the case with all whole-body studies using somatostatin, it is impracticable to faithfully mimic original hormonal conditions for insulin and glucagon patterns. Obviously, this limitation impedes direct comparison between conditions with and without somatostatin. For this reason, we included an experimental situation in which we reproduced GH patterns (circulating GH concentrations and free and total IGF-I) to allow isolation of the effects of GH. Thus, unless somatostatin per se profoundly changes the protein kinetic actions of GH (of which there is no evidence), we are confident that our findings are both qualitatively and quantitatively tenable in a physiological context.

It is impressive that even a modest 40-50% reduction in GH levels during somatostatin infusion was associated with substantial effects on the metabolic adaptation to short-term fasting. A moderate dose of somatostatin (200 μg/h) was used to avoid gastrointestinal discomfort during the 28-h infusion. Nevertheless, plasma GH concentrations were significantly decreased compared with ordinary fasting and GH substitution. Furthermore, total IGF-I levels decreased 30-40% and free IGF-I decreased 60-70% during GH suppression, but were almost identical during ordinary fasting and GH substitution. This finding indicates that a significant degree of GH suppression was accomplished and that the GH replacement experiments in these terms matched normal fasting conditions. GH levels during GH suppression tended to be higher than in the basal state, suggesting that the observed effects of GH suppression represent a conservative estimate and that the impact might have been even more pronounced with more complete suppression of GH secretion.

The metabolic effects of GH are complex and involve increased lipolysis, hyperinsulinemia, and stimulation of IGF-I activity, all of which have protein anabolic properties (15,39,40,41,42,43). In addition, GH has direct anabolic effects (44) and may—as observed presently—induce hyperglycemia and low circulating concentrations of tyrosine and phenylalanine. Of these potential secondary mediators, lipid intermediates have been shown to stimulate protein synthesis (15,39,40), insulin to inhibit breakdown (41,45), and IGF-I to act through both mechanisms (42,43,44,45,46). Furthermore, hyperglycemia may be protein sparing (47).

Previous studies assessing the impact of GH on protein metabolism at the whole-body level have in general shown that GH primarily increases protein synthesis (48,49), and there is evidence that acute exposure to high levels of GH may increase muscle protein synthesis (50,51). Furthermore, it has been reported that 6 weeks of high-dose GH treatment to malnourished hemodialysis patients resulted in stimulation of muscle protein synthesis without any effects on muscle protein breakdown (52). In contrast, some studies have failed to detect any effect of GH on muscle protein synthesis (22,53). In one of these studies, Copeland and Nair (22) found that isotopically measured muscle protein breakdown across the leg was relatively lower after acute GH exposure, with borderline P-values of 0.05 for phenylalanine and 0.09 for leucine. It should be noted that the majority of studies assessing muscle protein metabolism have employed exposure to very high levels of GH. The present study strongly suggests that under conditions of physiological fasting, even small increments in circulating concentrations of GH have substantial preserving effects on whole body nitrogen balance and that these effects are due in part to inhibition of muscle protein breakdown. The apparent complexity of the impact of GH on protein metabolism is perhaps predictable, considering the widespread metabolic actions of GH mentioned above. In this context, it should also be underlined that GH has been reported to specifically suppress hepatic ureagenesis (54,55). One of the most striking effects of GH deprivation in the present study was a 60-70% decrease in circulating free IGF-I. The magnitude of this suppression indicates that free IGF-I may be an important mediator of the protein-conserving effects of GH during brief fasting. This notion is in line with the study in malnourished hemodialysis patients, which reported a strong correlation between GH-induced reduction in protein catabolism and increments in circulating free (but not total) IGF-I (52). Interestingly, IGF-I, when perfused across the human forearm, has been shown to inhibit muscle protein breakdown (46); it is thus tempting to suggest that the observed decrease in muscle protein catabolism could be generated by free IGF-I. It is also possible that increased FFA concentrations and subtle hyperinsulinemia may have participated. In the present study, FFA concentrations were increased and insulin and C-peptide concentrations ascended slightly (P > 0.05) during GH infusion.

Because phenylalanine hydroxylation (a catabolic step) may reflect catabolism of amino acids in general, the degradation at steady state should ideally correspond to ureanitrogen synthesis (56). Presently, phenylalanine degradation did not change during fasting, although the urea-nitrogen synthesis rate increased in accordance with other studies (21,57,58). A strict comparison between urinary nitrogen excretion and labeled isotope turnover cannot be made, because several variables have to be taken into account. First, the urea pool is large and equilibrates slowly, meaning that calculated urea synthesis and excretion rates to a large extent are products of events occurring in the time period preceding the final 4-h fasting period during which urine was collected and circulating concentrations measured. Second, it has been suggested that the magnitude of urinary nitrogen loss in the postabsorptive state reflects protein intake the previous days, that postabsorptive protein oxidation is therefore predominantly generated by antecedent surplus protein ingestion, and that only hence forward are structural proteins being extensively degraded (7,59). Finally, the pattern of phenylalanine degradation does not necessarily reflect degradation of all amino acids. Taken together, our results are compatible with the concept that when studied while fasting, the subjects were in a transitional phase from early fasting with high ureagenesis and nitrogen loss (as evidenced by high urea-nitrogen synthesis) to a later stage with protein conservation (as evidenced by low whole-body phenylalanine degradation).

The present study demonstrates that phenylalanine flux (reflecting proteolysis) increases in healthy young men after 40 h of fasting. Tyrosine flux did not change, presumably because tyrosine flux represents protein breakdown as well as tyrosine appearance from phenylalanine hydroxylation. Our phenylalanine flux results support reports of an increase in leucine flux after 1.25 days of fasting in healthy subjects (9) and an even more pronounced increase after 3 days of fasting (4,41).

In conclusion, our results provide new evidence that physiological increments in GH secretion are a crucial adaptive response to fasting to achieve protein conservation, as evidenced by the 50% increase in ureagenesis during GH deprivation. The observed decrease in muscle protein breakdown during GH substitution—possibly mediated by the sustainment of circulating concentrations of free IGF-I—contributes to this overall anticatabolic impact.

GH preparations were generously supplied by Novo Nordisk (Copenhagen, Denmark). The study was supported by the Danish Research Council, grant number 9600822 (Novo Nordisk Center for Research in Growth and Regeneration, Århus University).

We are grateful to Eva Seier Petersen and Lone Korsgård for excellent technical assistance.