Effect of starvation on human muscle protein metabolism and its response to insulin DAVID A. FRYBURG, EUGENE J. BARRETT, RITA J. LOUARD, AND ROBERT A. GELFAND Yale Clinical Research Center and Department of Internal Medicine, Yale University School of Medicine, ATew Haven, Connecticut 06510 FRYBURG, DAVID A., EUGENE J. BARRETT, RITA J. LOUARD, AND ROBERT A. GELFAND. Effect of starvation on human muscle

protein metabolism and its response to insulin. Am. J. Physiol. 259 (Endocrinol. Metab. 22): E477-E482, 1990.-Although starvation is known to impair insulin-stimulated glucose disposal, whether it also induces resistance to insulin’s antiproteolytic action on muscle is unknown. To assess the effect of fasting on muscle protein turnover in the basal state and in response to insulin, we measured forearm amino acid kinetics, using [ 3H]phenylalanine (Phe) and [ 14C]leucine (Leu) infused systemically, in eight healthy subjects after 12 (postabsorptive) and 60 h of fasting. After a SO-min basal period, forearm local insulin concentration was selectively raised by ~25 pU/ml for 150 min by intra-arterial insulin infusion (0.02 mU kg-l. min-‘). The 60-h fast increased urine nitrogen loss and whole body Leu flux and oxidation (by 50-75%, all P < 0.02). Postabsorptively, forearm muscle exhibited a net release of Phe and Leu, which increased two- to threefold after the 60-h fast (P < 0.05); this effect was mediated exclusively by accelerated local rates of amino acid appearance (R,), with no reduction in rates of disposal (R& Local hyperinsulinemia in the postabsorptive condition caused a twofold increase in forearm glucose uptake (P < 0.01) and completely suppressed the net forearm output of Phe and Leu (P < 0.02). After the 60-h fast, forearm glucose disposal was depressed basally and showed no response to insulin; in contrast, insulin totally abolished the accelerated net forearm release of Phe and Leu. The action of insulin to reverse the augmented net release of Phe and Leu was mediated exclusively by a -40% suppression of R, ( P < 0.02) rather than a stimulation of Rd. We conclude that in short-term fasted humans 1) muscle amino acid output accelerates due to increased proteolysis rather than reduced protein synthesis, and 2) despite its catabolic state and a marked impairment in insulin-mediated glucose disposal, muscle remains sensitive to insulin’s antiproteolytic action. l

amino acid tracer kinetics; phenylalanine; leucine THE ABILITY of humans to survive long periods of nutritional deprivation depends on successful metabolic adaptation aimed ultimately at conserving body protein stores (6). During short-term (3-4 day) fasting, however, protein catabolism appears to accelerate, as rates of urinary nitrogen excretion (6,14) and whole body leucine flux and oxidation (14, 18, 26) increase compared with rates observed in postabsorptive humans. The extent to which these changes in whole body protein metabolism reflect alterations in skeletal muscle remains unclear. Using the forearm balance method, Pozefsky et al. (22) demonstrated an increase in the net release of amino 0193-1849/90

acids from skeletal muscle in subjects fasted for 3 days. In the absence of isotopic turnover methods, however, it cannot be discerned whether the change in forearm skeletal muscle metabolism is due to an increase in muscle proteolysis, a decrease in protein synthesis, or changes in both processes. In vitro studies of muscle from starved animals provide evidence for both a decrease in synthesis and an increase in degradation (15-17, 23). Insulin is believed to be the key regulator governing body protein homeostasis du ring transitions between the fed a.nd fasted states (6, 7) . Circulating insulin concentrations decline early in fasting, when catabolism accelerates, and the administration of insulin has been shown to suppress muscle amino acid release (12, 21) by suppressing proteolysis (12). Moreover, short-term fasting induces a state of resistance to insulin action on glucose metabolism (5, 19), raising the possibility that muscle insulin resistance may contribute to the early protein catabolic response. The present study was designed to examine whether the ch .anges in net muscle amino acid balance caused by short-term starvation are mediated by alterations in protein breakdown, synthesis, or both, and whether an impairment in insulin’s muscle protein anabolic action may contribute to these changes. METHODS Subjects and experimental design. Eight healthy volunteers (5 females, 3 males) age of 25 t 2 (mean t SE) yr, were studied in the Clinical Research Center of the Yale University School of Medicine. All subjects were within 15% of ideal body weight for height and build (Metropolitan Life Insurance Tables, 1983), and none was taking any medications. Each subject gave informed written consent before participating in this study, which was approved by the Human Investigation Committee of Yale University. Each subject was studied after a 12-h fast (postabsorptive) and again after a 60-h fast, for which they were hospitalized. During the 60-h fast, water was available ad libitum, and each subject consumed at least 1.5 liters each day. Multiple vitamins and folic acid (1 mg) were dispensed daily, and supplemental potassium (KC1 elixir) was administered as needed based on daily serum electrolyte tests. Subjects were ambulatory within the Clinical Research Center during the 60-h fast, which was ve ry well tolerated by all participants. Skeletal muscle amino acid kinetics were measured in

$1.50 Copyright 0 1990 the American Physiological

Society

E477

Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 7, 2018. Copyright © 1990 American Physiological Society. All rights reserved.

E478

STARVATION

AND

MUSCLE

each subject after the 12. and 60-h fast using the forearm muscle balance method in combination with systemic infusions of radiolabeled phenylalanine and leucine (12). Each subject served as his own control. In each case, the 60-h study was performed using the forearm contralatera1 to the one used for the 12-h study. No systematic choice of the sequence of dominant vs. non-dominant forearm was made. Procedures. Catheters were inserted into the brachial artery and an ipsilateral deep forearm vein (in retrograde direction). A primed, continuous 5-h infusion of L- [ring2,6-3H]phenylalanine (-33 &i, ~0.43 &i/min) and L[ l-14C]leucine (-12 &i, -0.16 &!i/min) was begun using a contralateral forearm vein after the body bicarbonate pool had been primed with sodium [ 14C]bicarbonate (-3 &i). The arterial catheter was kept patent by the continous infusion of normal saline into the artery at 0.3 ml/min. After a 2.5-h basal period, insulin was infused continuously into the brachial artery at 0.02 mU kg-‘. min-’ for an additional 2.5 h. The insulin solution was prepared in normal saline containing 3% albumin and infused at the same rate (0.3 ml/min) as used for saline during the basal period. Arterial and venous samples were taken at 60,45,30,15, and 0 min before beginning insulin and at 90, 105, 120, 135, and 150 min into the insulin infusion (Fig. 1). For 2 min before and during withdrawal of each deep venous sample, a pediatric sphygmomanometer cuff was inflated about the wrist to 200 mmHg to exclude blood flow to the hand. Forearm plasma flow (3) was measured after each sampling interval (for 1a total of 5 measurements per period) from the dilution of indocyanine green dye infused intra-arterially for 5 min with the wrist cuff inflated; blood flow was calculated by dividing plasma flow by (1 - hematocrit). Forearm volume (with the hand excluded) was measured by water displacement. Whole body leucine oxidation was measured by collecting aliquots of expired COa in a Hyamine trapping solution that contained a phenolphthalein indicator titrated to change color when a known amount of CO2 was trapped. Total CO2 production was measured using the ventilated hood technique with a Deltatrac calorimeter (Sensor Medics, Anaheim, CA). Total urinary nitrogen excretion was measured on samples collected from each study period. Calculations. The net forearm balances for glucose and amino acids were calculated from the Fick principle net balance = WI

(0

- WI) x F

where [A] and [V] are arterial and venous substrate concentrations and F is the forearm blood flow (in ml. Time

(min)

0

-170

L- [1J4

+lFO

C] -Leucine

L - [ring -2, 6- 3H] - Phenylalanine Local

METABOLISM

min-lo 100 ml-’ forearm volume). The net balance is the difference between the rate of tissue disposal of arterial substrate (RJ and the rate of tissue release of substrate into vein (R,). That is net balance = Rd - R,

(2) For phenylalanine and leucine, tissue disposal can be calculated at isotopic steady state from the measured fractional extraction (E) of tracer as Rd = E x [A] x F

(3) where E is the arteriovenous difference in tracer radioactivity divided by the arterial tracer radioactivity (all in dpm/ml). Muscle production of new unlabeled amino acid can then be calculated from Eqs. I-3

Ra= (E x [Al x FI - WI - WI) x F) which reduces to R a = F X [V] X (1 9 SA,/SAJ

(4) where SAa and SA, are the specific activity (dpm/nmol) of the amino acid in artery and vein, respectively. Tissue release of new unlabeled amino acid is reflected by the decrement in specific activity between artery and vein. It should be noted from Eqs. 2 and 4 that tissue kinetics (Ra and Rd) can be fully defined by measuring [A], [V], F, and the arteriovenous specific activity ratio. That is, R, can be directly calculated using Eq. 4, and Rd is then given by Rd = Ra + net balance

(5) Hence, if the venous-to-arterial specific activity ratio can be determined, measurement of the absolute radioactivity concentrations in artery and vein is not essential. This approach was employed in calculating phenylalanine kinetics, using measurements of arterial and venous phenylalanine specific activity by a high-performance liquid chromatographic (HPLC) technique (see Analytic methods).

The above calculations define the kinetics of amino acid exchange between forearm muscle and circulating blood. For phenylalanine, which is neither synthesized nor metabolized in muscle (28), the measured rate of disappearance of tracer across the forearm at steady state should reflect its rate of incorporation into protein, whereas tissue release of new unlabeled phenylalanine should reflect its release from the breakdown of tissue protein. For leucine, Rd represents the total disposal of leucine entering tissue via the artery, but it does not distinguish between its possible fates in muscle, namely, incorporation into protein, transamination, or complete oxidation to COZ. Whole body circulating leucine flux rates (Q) were calculated from the rate of tracer infusion (IR, in dpm/ min) divided by the steady-state amino acid specific activities in arterial blood

Insulin

Q = IR/SAa

A A A A Ii ttttt FIG. 1. Schematic diagram of experimental each subject after 12 and 60 h of fasting.

PROTEIN

protocol

employed in

Under steady-state amino acid entrance blood. Since leucine amino acids, in the

(6)

conditions, Q defines the rate of into and exit from the circulating and phenylalanine are both essential absence of exogenous input tissue

Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 7, 2018. Copyright © 1990 American Physiological Society. All rights reserved.

STARVATION

AND MUSCLE

protein represents the sole source of new amino acid entering the circulation, and Q then provides an index of amino acid release from protein breakdown (B). Leutine oxidation (C) was calculated as C = V(‘“CO,)/(Sll,

x 0.8)

(7) where V(14C02) is the rate of production of 14C02 (in dpm/min) calculated from the product of the steadystate specific activity of expired COa (dpm/mmol) and the total rate of CO2 production. The factor 0.8 corrects for the nonexpired 14C02 generated from L- [ lJ4C]leucine oxidation, which is retained within body bicarbonate stores. Nonoxidative leucine disposal (S) was calculated as S=Q-C

(8) Analytic methods. Blood glucose concentration was measured by the glucose oxidase method with a glucose analyzer [coefficient of variation (CV) of measurement = 1%; Yellow Springs Instruments, Yellow Springs, OH]. Plasma insulin was determined by double-antibody radioimmunoassay. Concentrations of acidic and neutral amino acids with the exception of phenylalanine were measured in sulfosalicylic acid extracts of whole blood with an automated ion-exchange chromatographic technique (Dionex D-500, Sunnyvale, CA). Urinary nitrogen was measured with the Kjeldahl method. For the determination of leucine and phenylalanine specific radioactivity, 2 ml of acidified plasma were placed on a Dowex-SOG cation-exchange resin column (BioRad, Richmond, CA). After washing with 0.01 N HCl, the amino acids retained on the column were eluted with 4 M NH40H. The eluate was vacuum centrifuged to dryness, and the residue was redissolved in 800 ~1 of 2% trichloroacetic acid. After centrifugation a 200-~1 aliquot was removed and counted for 14C radioactivity with a dual-channel Packard Tricarb scintillation counter (Packard, Downers Grove, IL). Leucine specific activity (CV = 4%) was calculated from this 14C radioactivity measurement and the corresponding leucine concentration measurement. Phenylalanine specific activity was measured in the remaining trichloroacetic acid supernatant by an ionpair reverse-phase HPLC technique. The mobile phase consisted of 16% methanol (vol/vol) containing a phosphoric acid buffer (pH -5) and heptane-sulfonic acid added as an ion-pairing agent (low UV-PIC B7 reagent, Waters Associates, Milford, MA). The column flow rate was 1.2 ml/min. The column eluant was monitored for ultraviolet absorbance at wavelength 214 nm. With samples of 200 ~1 injected onto a 4.6 X 25 mm Beckman Ultrasphere ODS (Cl,, 5 pm) column, a sharply separated phenylalanine peak elutes at 14-16 min. This fraction was collected into scintillation vials and subsequently counted for 3H radioactivity. The mass of phenylalanine in each sample was calculated by comparing its peak area with that of prepared standards using a Nelson Analytical chromatography software package (Cupertino, CA). The calculation of net phenylalanine balance as well as specific activity was performed using the HPLC derived concentration. Specific activity was calculated as the 3H radioactivitv divided bv the phenvlalanine mass.

PROTEIN

E479

METABOLISM

The coefficient of variation of phenylalanine specific activity, calculated in this fashion, was 2.2%. Data presentation and statistical analysis. All data are presented as means t SE. The values presented for the basal and insulin infusion periods (-60 to 0 and 90-150 min, respectively) were each determined from the mean of five steady-state measurements in each subject. Comparisons between 12 and 60 h of fasting with and without local insulin infusion were performed using analysis of variance with repeated measures. RESULTS

Effect of starvation on circulating glucose, amino acids, and whole body protein metabolism. After the 60-h fast,

arterial glucose and insulin concentrations fell significantly (glucose: 4.4 t 0.1 vs. 2.9 t 0.3 mM, 12 vs. 60 h, P < 0.001; insulin: 4.3 $- 0.4 vs. 2.3 -t 0.1 pU/ml, P < 0.005). Arterial amino acid concentrations in the postabsorptive and 60-h fasted states are shown in Table 1. Significant declines were observed in levels of alanine and other glucogenic amino acids (e.g., glycine, serine, and glutamine), whereas each of the branched-chain amino acids (valine, isoleucine, and leucine) rose significantly. Phenylalanine levels also rose by 20% after the 60-h fast. Table 2 provides steady-state data on leucine and phenylalanine arterial and venous concentrations as well as the specific activities for each measurement period. Although these arterial concentrations and their respective specific activities remained stable during the entire 12-h study, the arterial concentrations of these substrates declined slightly, but significantly, and their corresponding specific activities rose in the time between the basal and insulin infusion period measurements (Table 2). Although statistically significant, the changes in tracer specific activity over time were on average only 3-5%/h between the beginning of the basal and end of the insulin periods, indicating that measurements were made during a relative steady state. In agreement with other studies of short-term fasted 1. Effect of fasting on arterial acid concentrations

TABLE

Amino

Acid

12-h Fast

60-h Fast

amino P

Taurine 165&13 193&19 Threonine 126214 94t9 Serine 147t12 lOlzk5 co.01 Asparagine 59+3 56t4 Glutamate 155t16 125210 co.005 Glutamine 514t51 444t37 co.05 Glycine 278&20 216&14 co.01 Alanine 237t15 188k7 co.05 Citrulline 24k2 17t3 co.005 Valine 185t8 4Olk17