The effect of systemic hyperinsulinemia with concomitant amino acid infusion on skeletal muscle protein turnover in the human forearm

The Effect of Systemic Hyperinsulinemia With Concomitant Amino Acid Infusion on Skeletal Muscle Protein Turnover in the Human Forearm Elliot Newman,

Martin

J. Heslin,

Ronald

F. Wolf,

Peter W.T.

Pisters,

and Murray

F. Brennan

In vitro, insulin has been shown to increase skeletal muscle (SM) protein synthesis and decrease SM protein breakdowr,. Whether these same effects are found in vivo in man is less clear. The study of the effect of hyperinsulinemia (INS) on SM protein turnover (SMPT) is complicated by hypoaminoacidemia, which can obviate the true effect of insulin on SMPT. To prevent this, we studied the effect of INS on SMPT in the human forearm with amino acid (AA) infusion to ensure adequate substrate for full evaluation of insulin’s effect. Twelve healthy volunteers (aged 53 + 3 years) were studied. Steady-state AA kinetics were measured across the forearm after a systemic 2-hour primed continuous infusion of 3H-phenylalanine (3H-Phe) and 14C-leucine (14C-Leu) in the postabsorptive (PA) state and in response to systemic INS (71 + 5 pU/mL). AAs were infused during INS as 10% Travasol (Travenol Laboratories, Deerfield, IL) at .Oll mL/kg/min to maintain PA branched-chain AA (BCAA) levels, known regulators of SMPT, and to mildly elevate total AA levels. The negative PA net balance of both Phe and total Leu carbons (LeuC) became positive with INS + AA infusion (Phe from -16 f 2 to 12 + 3 nmol/min/lOO g [P < .Ol]; LeuC from -26 f 6 to 24 * 7 nmol/min/lOO g [P < .Ol]). The rate of disposal (Rd) of Phe into SM. an index of protein synthesis, increased from 31 -C 5 nmol/min/lOO g postabsorptively to 52 + 7 nmol/min/ 100 g with INS + AA (P < .Ol), while the LeuC Bd increased from 62 f 16 to 66 f 16 nmol/min/ 100 g (P = .65). The rate of appearance (R,) of phe from SM. an index of protein breakdown, decreased from 46 f 6 nmol/min/lOO g postabsorptively to 40 + 5 nmol/min/lOO g with INS + AA (P = .06), while the LeuC R, decreased from 107 f 16 to 62 r 14 nmol/min/lOO g (P c .03). These data suggest that the anabolic effect of insulin across the human forearm, in the presence of mildly elevated arterial AA levels, is to increase SM protein synthesis and decrease SM protein breakdown. Copyrig& 0 1994 by W.B. Saunders Company

T

HE ABILITY OF INSULIN to affect skeletal muscle (SM) protein turnover (SMPT) is well established. In animals, both in vitro and in vivo data demonstrate insulin’s ability to increase SM protein synthesis’” and decrease SM protein breakdown. l,* In humans, in vitro data have confirmed that insulin increases muscle protein synthesis and decreases muscle protein breakdown.4 However, in vivo human studies have been less clear. Under conditions of regional forearm hyperinsulinemia (INS), measuring phenylalanine and leucine kinetics, insulin was shown to decrease forearm muscle protein breakdown without affecting muscle protein synthesis.s In contrast, a more recent study reported that systemic INS with hyperaminoacidemia increased leg muscle protein synthesis with a minimal decrease in leg muscle protein breakdown when following Phe kinetics.6 However, SM protein breakdown decreased, and there was no change in synthesis when Leu kinetics were measured under the same conditions. This would suggest that the type of model used to study this issue influences the resulting data. The study of the effect of insulin on protein turnover is complicated by a dose-dependent hypoaminoacidemia consistently seen with systemic INS.7*8 This effect is most

From the Surgical Metabolism Laboratory, Memorial SloanKettering Cancer Center, New York, NY. Submitted October 2,1992; accepted February 11, 1993. Supported by the Surgical Metabolism Fund and US Fublic Health Sentice Grant No. T32 CA 09501. Presented in part at The American Sociev for Clinical Investigation National Meeting, Seattle, WA, May 1991, andpreviously published in abstract form (Clin Res 39:384a). Address reprint requests to Muway F. Brennan, MD, Chairman, Depatiment of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 YorkAve, New York, NY10021. Copyright 0 1994 by W.B. Saunders Company 0026-0495/9414301-0011$03.0010

70

significant for the branched-chain amino acids (BCAA), known regulators of muscle protein turnover.9s10 This leads to intracellular substrate depletion, thereby obscuring the effect of insulin on protein synthesis.” Although some investigators have attempted to prevent this by using regional insulin infusion,5 others have simply replaced amino acids (AAs) with exogenous AA infusion during insulin infusion,8J3 but not always while maintaining BCAA at baseline postabsorptive (PA) levels,6 which might more accurately isolate the effect of insulin on protein turnover. This study was designed to examine the effect of euglycemic systemic INS on muscle protein turnover in the human forearm, with a concomitant AA infusion designed to maintain basal BCAA concentrations and mildly elevated total AA levels throughout the study, thereby ensuring adequate substrate concentrations for better evaluation of the independent effect of insulin. SUBJECTS AND METHODS

Twelve healthy volunteers aged 40 to 67 years were studied in the Surgical Metabolism Laboratory at Memorial Sloan-Kettering Cancer Center. All subjects were within 15% of their ideal body weight ([IBW] Metropolitan Life Tables, 1983). There was no personal or family history of diabetes mellitus, and all had normal fasting blood glucose levels. No subjects were hypertensive. Hepatic and renal function tests were within normal limits, and none were taking medications known to affect intermediary metabolism. Subject characteristics are outlined in Table 1. The protocol was approved by the Memorial Sloan-Kettering Institutional Review Board, and informed written consent was obtained from each subject before the study. Study Design

A study design overview is depicted in Fig 1. All subjects were studied after a lo- to 12-hour overnight fast in the PA state while at rest in the supine position. AA kinetics across the forearm were measured using tritiated Phe (3H-Phe) and 14C-Lcu as previously Metabolism, Vol43, No 1 (January), 1994: pp 70-78

EFFECT OF INSULIN ON PROTEIN METABOLISM

IN MAN

Table 1. Subject Characteristics

Age (vr)

53 3 2

Sex (M:F)

5:7

Height (cm)

169 + 3

Weight (kg)

65 -+ 3

% IBW’

102 + 3

NOTE. All values are means + SEM. *lBW based on 1983 Metropolitan Life Tables.

On the morning of the study, two polyethylene catheters (Angiocath l&gauge. Deseret Medical, Sandy, UT) were inserted into forearm veins for infusion of isotope, insulin, and AAs via separate infusion pumps (Valley Lab, Infusion Systems Division, Boulder, CO). The nondominant radial artery was then percutaneously cannulated (Angiocath 20-gauge) under local anesthesra (1% Lidocaine, Abbott Laboratories, North Chicago, IL) for arterial blood sampling, and an antecubital deep vein (Angiocath l&gauge) was cannulated in a retrograde manner opposite the infusion catheters for sampling of venous effluent of the forearm muscle bed. At I = - 165 minutes, a primed continuous infusion of L-(ring 2,6-3H)Phe (bolus, 32 &i; infusion, 0.64 &i/min) and I-14C-Leu (bolus, 16 FCi; infusion, 0.16 $Zi/min; New England Nuclear, Boston, MA) was started, followed by a 2-hour baseline equilibration period (t = -165 minutes tot = -45minutes) necessary to achieve a steady state. After the 2-hour equilibration, baseline arterial and deep venous blood samples were drawn simultaneously at t = -4.5, -30, -15, and 0 minutes for AA, insulin, and glucagon levels and to determine )H-Phe, 14C-Leu, and “C-tu-ketoisocaproic acid ((u-KIC) specific activities. A pediatric blood pressure cuff was inflated at the wrist to 200 mm Hg for 1 minute before and during blood sampling to exclude blood flow from the hand. Forearm blood flow was measured during this period by venous capacitance plethysmography’4 every 10 minutes for three measurements. After completing baseline measurements at t = 0 minutes, the 3H-Phe (0.64 &i/min) and 14C-Leu (0.16 @/min) infusions were contrnued as before, and the euglycemic hyperinsulinemic c1amp15 was begun with a bolus of regular insulin (Humulin-R, UlOO, Eli Lilly and Co, Indianapolis, IN) at 400 mu/m2 followed by a contmuous infusion of insulin at 1 mU/kg/min. To prevent the concomitant hypoaminoacidemia seen with INS, a continuous infusion of AAs (10% Travasol without electrolytes, Travenol Laboratories, Deerfield, IL) was begun at ,011 mL/kg/min along with the insulin infusion. The composition of the Travasol solution (pmol/L) was as follows: Leu;55.6; isoleucine, 45.7; lysine, 39.7; valine, 49.5; Phe, 33.9; histidine, 30.9; threonine, 35.3; methionine, 26.8; tryptophan. 8.8; alanine, 232.3; arginine, 66.0; glycine, 137.2; described?

Regular Insulin Glucose (D30) 10% Amino Acids L (ring 2,6-3H) Phenylalanine L-[I]-“C-Leucine time

-165

(mid

Baseline

Study

Fig 1. Overview of the experimental design in the baseline PA state and in the study period during which the insulin + AA infusion was administered. Doses and infusion rates are detailed in the text. Arrows represent measurement time points in each period.

71

proline, 59.1; serine, 47.6; tyrosine, 2.2; total AAs, 870.6. These infusion rates were selected to achieve physiologic INS in the 60- to 80-$J/mL range and mildly elevated total AA levels, while maintaining baseline BCAA levels based on previous observations in man from our own laboratory’6 and others.lj Euglycemia was maintained with a variable infusion of 30% dextrose ([D30] Abbott Laboratories). Arterial plasma glucose concentration was measured at 5-minute intervals, and the D30 infusion was adjusted using previously described formulas. I5After another 2-hour equilibration period (t = 0 minutes to I = 120 minutes) to reestablish steady-state conditions, arterial and deep venous blood samples were again simultaneously collected at t = 120, 135. 150, and 165 minutes for determination of the same parameters as in the basal period. Forearm blood flow measurements were also repeated as previously described. The period of systemic INS with AA infusion will be referred to as the study period. All solutions were prepared on the morning of each study. Short-acting insulin was dissolved in 0.9% NaCl with 2 mL of each subject’s own plasma to a final concentration of 300 mU!mL. All cannulated vessels were kept patent by a slow infusion of 0.45% NaCl solution (0.3 mL/min). Analytical Methods Blood for plasma AA analysis and Phe, Leu, and a-KIC specific activity determinations was collected in vacutainer tubes with sodium heparin (Becton Dickinson, Rutherford, NJ). After centrifugation, a portion of the arterial and venous plasma samples was deproteinized with a .59 mol/L perchloric acidia-methylphenylalanine (internal standard) solution in a 1:l ratio for subsequent determination of plasma Phe concentrations by reverse-phase high-performance liquid chromatography (HPLC).” ‘H-Phe radioactivity and Phe specific activities were determined after fractioncollecting the Phe peak from the HPLC separation, followed by liquid p-scintillation counting (TriCarb 4000 Series. Packard Instruments, Downers Grove, IL) of the peak. A second aliquot of plasma was deproteinized using 50 ~1 50% sulfosalicylic acid/mL plasma for extraction of a-KI@ and subsequent determination of the plasma concentration and specific activity. The (Y-KIC was separated from 3 mL deproteinized plasma, and the plasma concentration was also determined by HPLC using an internal standard with fraction collection. Immediately after collection of the a-KIC peak, 500 FL was used to determine radioactivity by liquid p-scintillation counting as described above, and 200 ~1 of the remaining collected peak was reinjected for determination of the concentration used in the calculation of the specific activity. Leu radioactivity was determined by pipetting 500 PL deproteinized plasma over 1 mL cation-exchange resin (AG 5OW-X8, 200-400 mesh, Biorad, Richmond, CA) and then eluting the AA fraction into liquid scintillation vials for counting. The corresponding concentration of Leu was determined from the same sample of deproteinized plasma by AA analysis using ion-exchange chromatography and postcolumn ninhydrin derivatization (Pickering Laboratories, Mountainview, CA). Total AA analysis was performed on a separate aliquot of plasma, also by ion-exchange chromatography with postcolumn ninhydrin derivatization as previously described. All HPLC analysis was performed on a Perkin-Elmer Series 4 Liquid Chromatograph (Perkin-Elmer Instruments Division, Norwalk, CT) with integration of the AA peaks by an Omega software package (Perkin-Elmer). Determination of hormone concentrations was made from arterial blood collected into glass tubes with no additives (insulin) or with sodium EDTA and 1,000 KIU aprotinin (glucagon). Insulin levels were measured by double-antibody radioimmunoassay kits (Autopak, Micromedic Systems, Horsham. PA). Glucagon levels

72

NEWMAN ET AL

were measured as previously described” at the Diabetes Care Facility, University of Pennsylvania Hospital, Philadelphia, PA. Accepted normal fasting values for glucagon at this facility are 200 to 400 pg/mL. Bedside arterial and deep venous plasma glucose concentrations were measured with the Beckman II Glucose Analyzer (Beckman Instruments, Irving, CA). All plasma samples were stored at -70°C until analysis.20

Calculatiorzs

proportion of the intracellular disposal of Leu.’ The equations are as follows: Net balance of LeuC = ([LeuC]a - [LeuC]v =

x flow

Rd - R,,

Eq5

where [LeuC]a = ([Leu] artery + [(Y-KIC] artery) and [LeuC]v = ([Leu] vein + [u-KIC] vein). This net balance was then resolved into component parts as above: R, LeuC = [LeuCJa x flow

Phe kinetics. The rates of SM protein synthesis and degradation were determined using steady-state Phe exchange kinetics across the forearm muscle bed as previously described? Since Phe is not metabolized within SM,22 its incorporation into and release from SM should be indicative of SM protein synthesis and breakdown, respectively.?’ The equations to quantitatively derive the kinetic rates are as follows: Net balance of Phe across the forearm = ([A] - [VI) x flow,

Eq 1

where [A] and [V] are the arterial and venous concentrations of Phe (p,mol/L), respectively, and flow is the forearm plasma flow (mL/min/lOO g tissue). The net balance aiso represents the difference between the rate of disappearance of Phe into forearm muscle tissue (Rd) and the rate of appearance of Phe from forearm muscle tissue (R,), ie, Net balance = R,, - R,.

Eq 2

The R, of Phe (nmol PhelminilOO g tissue) can be quantified by measuring the dilution of the specific activity of Phe (dpm/nmol/L) on the venous side of the forearm muscle bed relative to the specific activity on the arterial side, which is occurring secondary to the release of unlabeled Phe from the forearm muscle bed into venous blood. This relationship is defined as follows: R, = [A]

x flow x [(SAa

- SAv)/SAv],

Eq 3

where R, is the rate of appearance of Phe from muscle (nmol Phe/min/lOO g tissue), [A] and flow are as described above, and (SAa - SAv)/SAv is the difference between the arterial and venous specific activities relative to the venous specific activity. Finally, the Rd of Phe (nmol PhelminilOO g) can be calculated by simply combining and rearranging equations 1 through 3 such that Rd = net balance + R,,

Eq4

where Rd is the rate of Phe disappearance into muscle, and net balance and R, are as measured and calculated previously. It is important to note that the R, (and therefore the Rd, which is based in part on the R,) is derived on the assumption that the venous specific activity most closely approximates the Phe precursor pool used by muscle for protein turnover. The details of this derivation have been previously described.*’ Leu kinetics. As a basis for comparison, SMPT was also measured using 14C-LeuC kinetics across the forearm based on a modification of a method recently described6 for use with stable isotopes and used in a related study from our own laboratory.23 The equations are analagous to those used for Phe kinetics, but whereas Phe is not metabolized intracellularly, Leu is both transaminated and irreversibly decarboxylated intracellularly.7 Therefore, these calculations account for the net LeuC flux across the forearm, which includes Leu carbons as well as the transaminated product of Leu, IX-KIC. However, it does not account for the intracellular decarboxylation, which is generally believed to be a relatively small

x

((LeuC SAa - LeuC SAv)/LeuC SAv}, Eq 6

where LeuC SAa = ((Leu SAa X [Leu]a) + (a-KIC SAa x [IXKIC]a)]/[LeuC]a and LeuC SAv = ((Leu SAv X [Leu]v) + ((Y-KIC SAv x [a-KIC]v)]/[LeuC]v. Finally, Rd LeuC = net balance LeuC + R, LeuC.

Eq 7

As mentioned earlier, forearm blood flow was measured by venous capacitance plethysmography as originally describedz4Js and recently revalidated in our laboratorv.14 Forearm blood flow was converted to plasma flow using the factor I-hematocrit in the basal and study periods, respectively. The glucose flux across the forearm was calculated by multiplying the forearm plasma flow by the mean arterial-venous glucose difference in the baseline and study periods. Statistics All data in the text and figures are presented as the mean ? standard error of the mean. Comparisons between basal and study periods were performed by Student’s t test (Clinstat software package, International University Press). Stepwise linear regression of net balance changes to insulin and AA levels were performed on SAS software. The four steady-state measurements in both basal and study periods were averaged to yield one value for each period per subject, except for Leu concentrations and specific activities, which were based on three steady-state measurements. Statistical significance was defined as P less than .05. RESULTS

Plasma Glucose, Insulin, and Glucagon The basal PA glucose level was 83 * 2 mg/dL. During the euglycemic hyperinsulinemic clamp, the mean steady-state glucose level (from 120 to 165 minutes) was kept near basal at 87 5 1 mg/dL. The overall coefficient of variation for the clamps was 8% 2 1% (Table 2). The basal PA insulin concentration was 7 2 1 pU/mL and was increased to a mean steady-state value of 70 ? 4 $_J/mL to achieve a state of physiologic INS. The mean PA glucagon level was 321 f 29 pg/mL, which was not significantly different from the mean steady-state level in the study period of 300 ? 31 pg/mL, (Table 2). Table 2. Plasma Glucose, Insulin, and Glucagon Levels in Basal State and During Study Period

Glucose (mg/dL) Insulin (&J/mL) Glucagon (pg/mL)

Basal

Study

83 ? 2

87 + 1

Jel 321 + 29

JO f 4 300 + 31

NOTE. All values are means + SEM of four steady-state measure-

ments in each

period per subject.

EFFECT OF INSULIN ON PROTEIN METABOLISM

IN MAN

73

Forearm Plasma Flow

Table 4. A-DV Differences (pmol/L)

In the basal PA state, the mean plasma flow was 2.3 k 0.3 mLlmin/lOO g forearm tissue. This value significantly increased (P < .02) to 3.2 + .4 mL/min/lOO g forearm tissue in the study period. P’lasmaAA Concentrations Puring AA infusion, total arterial plasma BCAA (Leu, Be. Val) were maintained at baseline levels (372 * 20 t.rmol/L basal and 387 rfr 24 kmol/L study, P = .46) with a Leu concentration of 114 2 7 kmol/L basal and 103 + 8 FmoliL study (P = .1.5). Phe concentrations increased from 50 ? 3 p,mol/L in the baseline period to 79 2 4 kmol/L in the study period. Total measured AA levels increased by 34% (1,680 2 53 +mol/L basal to 2,264 + 69 umol/L study, P < .0.5). Ala and Gly levels increased most dramatically (120% and 71%, respectively), reflecting the very high concentration of these two AAs in the Travasol solution. Excluding Ala and Gly, total measured AA levels increased only 15% from baseline levels. Of the remaining measured AA%, all except Tau, Glu, Asn, and Tyr increased during the hypcrinsulinemic period (Table 3).

A-DV Difference

As can be seen in Table 4, Ser, Gly, Val, Met, Be, Leu, Phe, and Trp significantly changed their arterial-deep venous difference across the forearm from negative in the baseline to positive in the study period. The arterial-deep venous difference for Thr, Ala, Lys, and His became significantly less negative with INS + AA, while that for Tau, Asn, and Glu showed no change. As expected, all measured AA levels except for Glu were in negative flux across the forearm in the basal PA period. Table 3. Arterial Plasma AA Concentrations

[Fmol/L)

in Basal State

and Study Period

TaLl

Basal

Study

51 +3

49 + 2

GIU

64 + 9

54 + 7”

Ser

112 -+ 5

138 + 4*

ASll

4% 5 2

38 f 2*

GlY

222 t 14

380 2 12*

His

102 z 4

145 k 6*

Thr

125 + 9

144*8*

Ala

197 + 14

439 + 22*

T!Jr

55 f 4

39 + 3*

Val

197 f 10

201 2 13

Met

28 k

Ile

61 +4

LeL.

i

116k-7

53 + 2’ a3 2 4s 111 + 7

Phe

50 z 3

Trp

68 i 4

96 ? 9”

Lvs

179 z 10

224 r 12*

BCCA

372 -c_20

TAA

1,680 f 53

79 + 4*

387 + 24 2,264 + 69*

NOTE. All values are means i SEM of four steady-state ments in each period. Abbreviation:

TAA, total measured AAs.

*P < .05 v basal period.

measure-

g)

FIUX

Basal

Study

Basal

052

4+-2

-123

GIU

37 i 4

32 2 5

Ser

-2 2 3

Asn

-9 2 2

Tau

Gk His

ia?

3*

-10 t 2

725

a2 k 10

k 7

27 + 6”

-62-tl4

-5 i- 3*

-3oi7

-19 2 3 -92

Tvr Val

-12 -t 3

Met

-5 2 2

lie

-6 r 2

+ 12

-2 + 3

-2 f 2* -2ak

14~

-4329

-4 + 6* -93 -t44

-4?7

-1 t2

127

i 2 a*

-2628

1 k 2*

-1123

at 5*

-14k4

33 i 6* 16 2 11'

lo?

1*

-1122

4 * 3*

-2526

Phe

-623

3 2 2*

-1lz5

Trp

-525

8ka

-142

-91 t10

80 t 20~ -14,9*

-199 2 21

Leu

-27+_2

55 zk10' -24~3

+ 2

Ala

104 + 23

--5L 5 -2024

-26

Thr

LYS EAA

Study

-13

-1124x 12t 17s

3 2 23"

9 t 5* 11

-6158

21 +-25 -20 + 9*

-206 2 67

67 2 57*

NOTE. All values are means t SEM of four steady-state

measure-

ments in each period. Abbreviations:

Plasma Arterial-Deep Venous AA Differences and Forearm Fluxes

and Flux (nmol/L/min/lOO

in Basal State and Study Period

EAA, essential AAs (Thr, Val, Met, Ile, Leu, Phe, Lys,

Trp); A-DV, arterial-deep

venous.

*P < .05 Y basal period.

In the study period, all measured AA levels except Glu, Asn, Ala, and Tyr significantly decreased their negative baseline forearm flux. Tau, Ser, Gly, Val, Met, Ile, Leu, Phe, and Trp levels became significantly positive in comparison to baseline, whereas Thr, Lys, and His levels remained in negative flux. As a group, the essential AAs significantly reversed both the negative baseline arterial-deep venous difference and flux in the PA state to positive in the study period. The basal arterial and venous a-KIC concentrations were 27 2 3 and 28 ? 3 kmol/L, respectively (P = NS, arterial v venous). In the study period, the venous a-KIC concentration decreased to 18 c 2 pmol/L, which was slightly but significantly increased from the arterial value of 16 I 2 pmol/L. Both arterial and venous wKIC concentrations in the study period were significantly lower than baseline values (Table 5).

Steady-State Forearm Mu&e Phe and LruC Kinetics Figure 2 shows Phe, Leu, and a-KIC plasma concentrations and specific activities during the last 45 minutes of the baseline and study periods, demonstrating the steady-state condition during which measurements for protein turnover rates were taken. Figure 3a summarizes the Phe kinetics at steady state. In the baseline PA state, the net balance of Phe across the forearm was -16 5 2 nmol/min/lOO g tissue. In the hyperinsulinemic study period, the net balance of Phe across the forearm became significantly positive, increasing to 12 * 3 nmol/min/lOO g (P < .OOl). In the baseline PA state, the Rd of Phe across the forearm was 33 2 5 nmol/min/ 100 g tissue. This increased to 52 2 7 nmol/min/lOO g tissue (P < X101) during the

74

NEWMAN

Table 5. Forearm Plasma Flow and Phe, Leu, and CrKlC Concentrations

and SA at Steady State in Basal and Study Periods Basal

Study

Plasma flow (mL/ mini100 9 tissue)

2.3 2 0.3

3.2 r 0.5*

[Artery1 [Vein]

50 + 2

80 + 3*

57 k 2t

76 + 3*

SA artery

27 2 2

19 + 1

SA vein

19 + It

16 + It

Phe

Leu [Artery]

11627

111 +7

[Vein]

126 + 8t

103 +- 7*

SA artery

4.4 k 0.3

4.2 -t 0.3

SA vein

3.2 ? 0.3t

3.7 k 0.2t

a-KIC [Artery]

27 ? 3

[Vein]

28 2 3

18 + 2*

SA artery

2.6 + 0.1

2.4 + 0.2*

SA vein

2.2 +

o.ot

16 + 2*

1.9 + o.o*t

NOTE. All values are means f SEM of three to four steady-state measurements

in each period per subject. All concentrations

[ 1 are

nmol/mL, and all SAs are dpm/nmol. Abbreviation: SA, specific activity. *P < .05 Y basal. tP < .05 v artery.

study period. The R, of Phe across the forearm was 48 ? 6 nmol/min/lOO g in the baseline PA state, which decreased in response to systemic INS to 40 + 5 nmol/min/lOO g (P = .06). Figure 3b summarizes total LeuC kinetics across the forearm at steady state. As with Phe, the baseline PA LeuC net balance was -26 * 6 nmol/IOO g/min and became significantly positive during the study period (+24 2 7 nmol/lOO g/min, P < .002). However, contrary to the findings with the Phe tracer, the LeuC Rd did not significantly change (82 r 16 nmol/lOO g/min in the baseline period to 86 -C 18 nmolilO0 g/min, P = .85). The LeuC R, decreased significantly from 107 2 18 nmolilO0 gimin in the baseline period to 62 ? 14 nmol/lOO g/min, P < .03) in the study period. Forearm Glucose Flux

In the baseline PA state, there was net glucose uptake into the forearm at .lO * .02 mg glucose/min/lOO g tissue. This significantly increased during the study period to .53 t .ll mg/min/ 100 g tissue (P < .Ol). DISCUSSION We have shown that systemic INS, in the presence of an AA infusion that maintained basal BCAA levels, promotes positive net balance across the human forearm. In addition, radioisotopic evaluation of Phe regional exchange kinetics showed that this change in net balance was a result of the significantly increased Rd of Phe into forearm SM, with a lesser contribution from the decreased R, of Phe from muscle. Conversely, when measuring total LeuC kinetics under the same conditions, we find the change in net

ET AL

balance to result from a minimal increase in the Rd, but a significant decrease in the R,. The role of insulin and its effect on SMPT has been the focus of many investigations. In vitro studies in both animal and human muscle tissue has repeatedly demonstrated increased AA incorporation and decreased AA release in response to insulin.1.2,4J6 However, in vivo corroboration of these in vitro findings has been difficult, often leading to conflicting results among investigators. In part, this controversy is linked to the variety of ways investigators approach the well-documented hypoaminoacidemia seen with systemic INS. As previously demonstrated by our laboratory16 and by others,7 euglycemic insulin infusion causes a dosedependent decrease of plasma AA levels, especially the BCAA. This in turn has been shown to lead to intracellular depletion of AAs as well.” These findings make the evaluation of insulin’s effect on SMPT in the presence of decreased plasma AA levels difficult to interpret, and any in vivo study that examines the effect of insulin on protein metabolism must address this problem. Gelfand and Barrett traced both Phe and Leu kinetics across the forearm, and used local forearm insulin infusion to avoid systemic hypoaminoacidemia.5 They reported a change from negative to positive AA net balance across the forearm, which was due to significantly decreased protein breakdown as evidenced by the decreased rate of appearance of Phe and Leu, but no change in protein synthesis as tissue uptake of the tracer was unchanged. However, there is evidence to suggest that even under conditions where plasma levels of AAs are normal, intramuscular levels of these AAs may in fact be decreased in response to insulin infusion.27 Conceivably, simply maintaining basal AA levels may not be enough, and this still hinders our ability to see the full effect of insulin on muscle protein turnover. Bennet et al6 tried to address this problem by studying the effect of systemic INS on leg protein metabolism with exogenous AA replacement to increase arterial AA levels above baseline. By also tracing Phe kinetics across the leg, they concluded that under these conditions, leg muscle Phe uptake did in fact significantly increase with a minimal change in Phe release, indicating that insulin significantly increased leg muscle protein synthesis and minimally decreased protein breakdown. However, they also traced Leu kinetics across the leg under the same conditions of INS and hyperaminoacidemia in that study, and found that the Leu R, decreased, while the Leu Rd minimally increased, indicating that with respect to Leu kinetics, insulin decreased SM protein breakdown without significantly affecting synthesis. The exogenous AA infusion increased total AA levels almost twofold and BCAA and Leu levels by 120% and 115%, respectively.6 This complicates the interpretation of their findings, as there is both in vitro and in vivo evidence that BCAA, and Leu specifically, have a role in regulating AA and muscle protein metabolism9~28~29and that doubling plasma AA levels can also influence muscle protein turnover,13,30raising the possibility that these results are not solely due to the effect of insulin. Recently, Arfvidsson et a131 used radiolabeled Tyr to

EFFECT OF INSULIN ON PROTEIN METABOLISM

IN MAN

KIC SA ,dpm,d,

'1

4 Fig 2. Leu lb), specific baseline means *

Arterial (A) and venous (0) Phe (a), and u-KIC (cl concentrations and activities during sampling times in and study periods. All values are SEM.

’

4s

-30

, 15 ,p,,&,

em 111 1% 165

-

study the effect of systemic INS on leg muscle protein turnover without AA replacement, and concluded that insulin had no effect on either muscle protein synthesis or breakdown. However, in their study, both arterial AA levels and intramuscular AA levels decreased for a number of AAs, particularly the BCAA. Furthermore, these investigators raise the possibility that no effect of insulin was demonstrated because of the hypoaminoacidemia. The present study also used radiolabeled tracers as previously described5J)Ji,3z to assess the effect of insulin on forearm protein turnover, but because of the possible limitations associated with regional INS as mentioned above, we used systemic INS with exogenous AA replacement. In addition, the AA infusion rate was designed to maintain BCAA, specifically Leu, at baseline PA levels while keeping other AAs at mildly elevated levels, (1) to ensure a constant supply of substrate for insulin’s effect, and (2) to allow for evaluation of the independent effect of insulin. As shown in Table 3, this strategy did in fact allow us to keep BCAA levels at baseline, while essential and total AA levels increased by only 18% and 34%, respectively. Although prior investigation both in vitro and in vivo has shown that elevated AA levels can influence SMPT, as

indicated earlier, these effects generally are found at supraphysiologic AA levels in vitro,‘,3’ at least twice basal levels in vivo,13,30and seemingly secondary to the BCAA.9,z8J9 Clearly our increases in essential and total AA levels, while significant, do not approach these high levels, and as such are less likely to be responsible for the changes in Phe and Leu kinetics that we found. In addition, although Phe levels did increase by 46% from baseline to study periods, Phe alone has not been shown to influence protein turnover.’ Finally, stepwise regression analysis of the insulin and AA concentrations versus the net balance change from baseline to study found that only the insulin level was significant (P = .0005) and AA levels were not (P = .58) when both variables were entered into the model, providing further support that our findings are the result of the independent effect of insulin and not secondary to AA changes. As outlined in the Methods and in detail originally by Barrett et aL2i the equations we used to determine the R, and Rd of Phe and Leu across the forearm are based on using the venous specific activity (SAv) as the reference pool, on the theory that the SAv is diluted by the AA being traced coming directly from the intracellular pool into the

76

NEWMAN ET AL

a) nmol PhelminllOO

gm forearm

tis8ue

70,

p - .06

I

-30’ Net Balance

Ra

Rd

b) nmol LeuC/min/lOO

gin forearm

tissue

120 T 100

60 60 40 20 0 -20 -40 Net Balance

Ra

Rd

-basal

m

study

Fig 3. Steady-state (a) Phe and (b) total LeuC kinetics across the forearm in the baseline PA state and in the study period. All values are means f SEM.

venous blood and therefore more closely approximates the true intracellular po01.~~~*This is not unlike the concept behind using (Y-KIC as the reference pool for kinetic measurements with Leu,34 which also is used in the equations for calculating total LeuC exchange kinetics across the forearm. However, because we cannot definitively know whether the SAv is the best approximation of the true intracellular pool, we can also calculate the R, and Rd values with arterial specific activity (SAa) as the reference pool for intracellular activity as detailed previously.21 By doing so, the Rd and R, values for both of the traced AAs decrease in absolute value by about 20% to 30%. However, although the absolute values change for the Rd and R, of Phe and Leu, based on the SAv (maximal values) or the SAa (minimal values) as the reference pool, the conclusions remain unchanged, ie, insulin significantly increased the Rd of Phe into muscle with a minimal decrease in the Phe R,. At the same time, INS significantly decreased the R, of LeuC from muscle with a minimal increase in the LeuC Rd. It is also important to note that while we have determined the Rd indirectly after first calculating the net balance and R,, we could also measure the Rd directly, and equation 4 would be Rd = {flow x (dpm/mL art - dpm/ mL vn))/SAv, again with the SAv functioning as the

reference pool. However, this did not change the resulting Rd value or our conclusions (data not shown). Certainly, the major findings in this study are the different results obtained when tracing Leu kinetics across the forearm as opposed to Phe kinetics. As outlined in Fig 3a and b, the negative net balance seen across the forearm in the PA state became positive during the hyperinsulinemic study period, both for Phe and Leu kinetics. However, the changes in Rd and R, for Leu, while in the same direction as those seen for Phe, were of significantly different magnitudes (Rd of Leu increased only 5%, compared with 68% for Phe; R, of Leu decreased 42%, compared with 16% for Phe). Furthermore, as we do not have data to account for Leu oxidation in forearm muscle, these numbers probably underestimate the true differences between these AA kinetics. This apparent dichotomy in the way these two tracers are handled by the same SM is certainly perplexing, and as mentioned earlier, was similarly reported by Bennet et aLh who proposed that this discrepancy might be attributable to the marked increases in AA levels, especially the BCAA levels, which might differentially affect transmembrane and intracellular handling of Phe and Leu. In addition, they suggested that changes in Leu oxidation during INS might lead to underestimation of protein synthesis, consequently leading to different results between the two tracers. However, in the present study, by maintaining Leu and BCAA levels at baseline and only mildly increasing total AA levels, we have eliminated these variables as reasons for the disparate results. Furthermore, because we have maintained euleucinemia, the actual contribution of oxidation to the total Rd is likely to be minimal, since Leu oxidation has been shown to be directly related to plasma concentrations.13x30Since Leu oxidation in the PA forearm has been shown to be only about 7% of the total Rd,32the underestimation of protein synthesis in both the baseline and study periods in this study would not be enough to explain the differences between the tracers. Although we can exclude some potential explanations for the tracer discrepancy because of the modifications of our hyperinsuhnemic model compared with Bennet et akh our data do not allow us to provide an answer for this difference. It is conceivable that part of the explanation for these discrepancies is linked to the different metabolic pathways for these AAs within SM. Phe is not metabolized within SM,22 whereas Leu is transaminated and oxidized.35 Because of these alternate intracellular pathways, it is highly likely that the errors introduced by our reliance on plasma measurements of specific activities as being reflective of the intracellular protein kinetics are substantially different for Phe and Leu. The differences between the tracers that we observed are at variance with the results of Gelfand and Barrett,5 who found no differences between the Phe and Leu kinetic models under conditions of regional INS without AA infusion. This might suggest that in fact systemic INS with AA replacement is somehow responsible for the discordant results. However, in another study by their group, in which regional INS was used across the forearm with both Phe and Leu tracers to study starvation, some of their findings

EFFECT OF INSULIN ON PROTEIN METABOLISM

77

IN MAN

were found to be significant when tracing Leu, but not when tracing Phe.36 Clearly, even regional INS without exogenous AA infusion can result in tracer differences, and is not better than systemic INS with AA infusion. These dichotomous results between the Phe and Leu tracers highlight the difficulty of using these various isotopic models. Because of our inability to directly measure intraceliular AA pools, as outlined above, we must rely on the assumption that plasma measurements are representative of intracellular activity. To some degree this is an oversimplitication, as evidenced by the disparities between the two tracers, as well as by the wide range of net balances for the individual essential AAs within the same muscle (Table 4). This is a problem that is well recognized with this model,s,6.2i as well as with AA exchange kinetics in genera1.37 It is likely that there are other intracellular pathways and pools for these AAs. which at present we cannot directly account for, contributing to these discrepancies. Until a method for directly measuring these intracellular pools becomes available, we cannot more accurately explain the different results between the tracers. Nevertheless, the net effect of insulin to decrease net protein breakdown is clear. The unidirectional rates of synthesis and degradation must be interpreted in light of the limitations and assumptions of our model. In summary, we have evaluated the effect of systemic INS

with AA infusion on muscle protein turnover in the human forearm. Using Phe kinetics, our data show that insulin caused positive net balance across the forearm, mainly by increasing the Rd of Phe into muscle, an indicator of increased protein synthesis. However, when tracing Leu kinetics across the same forearm, we found that the main mechanism for achieving positive net balance was the decrease in R, of Leu from SM, an indicator of decreased protein breakdown. We conclude that insulin has a major anabolic effect on SMPT, which mechanistically is likely due to a combination of increased protein synthesis and decreased protein breakdown. Demonstration of this effect is also likely to require an AA infusion designed to maintain mildly elevated arterial AA levels for adequate AA availability. However, determining which mechanism is more significant, protein synthesis or breakdown, is less clear, and further studies are needed to better elucidate this problem. ACKNOWLEDGMENT The authors would like to thank Nada Vydelingum. PhD, for his support, Bruce Ng and Charles Ahrens for their expert technical assistance, and Howard Thaler, PhD, for his expert statistical advice and assistance. We would also like to thank Michael Burt, MD, PhD, Warren Heston, PhD. and the many research fellows from this country and abroad whose advice and help made this study possible.

REFERENCES I Fulks KM, Li JB, Goldberg AL: Effects of insulin, glucose, and amino acids on protein turnover in rat diaphragm. J Biol Chem 250 290-298, 1975 2 Jefferson LS, Li JB, Rannels SR: Regulation by insulin of amino acid release and protein turnover in the perfused rat hemicorpus. J Biol Chem 252:1476-1483, 1977 3 Garlick PJ. Fern M, Preedy VR: The effect of insulin infusion and food intake on muscle protein synthesis in postabsorptive rats. Biochem J 210:669-676, 1983 4 Lundholm K, Schersten T: Determination in vitro of the rate of protein synthesis and degradation in human-skeletal-muscle tissue. Eur J Biochem 60:181-186, 1975 5 Gelfand RA. Barrett EJ: Effect of physiologic hyperinsulinemia on skeletal muscle protein synthesis and breakdown in man. J Clin Invest X0:1-6, 1987 6. Bennet WM, Connacher AA, Scrimgeour CM, et al: Euglycemic hyperinsulinemia augments amino acid uptake by human leg tissues during hyperaminoacidemia. Am J Physiol 259:E185-E194. 1991) 7. Fukagawa NK. Minaker KL, Young VR, et al: Insulin dose-dependent reductions in plasma amino acids in man. Am J Physiol 25O:E13-E17, 1986 8. Flakoll PJ, Kulaylat M, Frexes-Steed M, et al: Amino acids augrnent insulin’s suppression of whole body proteolysis. Am J Physiol257:E839-E847, 1989 0. Buse MG. Reid SS: Leucine: A possible regulator of protein turnover in muscle. J Clin Invest 56:1250-1261, 1975 lti. Adibi SA: Roles of branched-chain amino acids in metabolic regulation. J Lab Clin Med 95:475-484.1980 II. Del Prato S. DeFronzo RA, Castellino P, et al: Regulation of amino acid metabolism by epinephrine. Am J Physiol 258:E878EX87.1990

1. Millward DJ, Nnanyelugo 0, James WP, et al: Protein metabolism in skeletal muscle: The effect of feeding and fasting on

muscle RNA, free amino acids and plasma insulin concentrations. Br J Nutr 32:127-142, 1974 13. Castellino P, Luzi L, Simonson DC, et al: Effect of insulin and plasma amino acid concentrations on leucine metabolism in man. Role of substrate availability on estimates of whole body protein synthesis. J Clin Invest 80:1784-1793, 1987 14. Dresler CM, Jeevanandam M. Brennan MF: Extremity blood flow in man: Comparison between strain-gauge and capacitance plethysmography. Surgery 101:35-39, 1987 15. DeFronzo RA, Tobin JD, Andres R: Glucose clamp technique: A method for quantifying insulin secretion and resistance. Am J Physiol237:E214-E223, 1979 16. Pisters PWT, Restifo N, Cersosimo E, et al: The effect of euglycemic hyperinsulinemia and amino acid infusion on regional and whole body glucose disposal in man. Metabolism 4059-65. 1991 17. Atherton ND. Green A: HPLC measurement of phenylalanine in plasma. Clin Chem 34:2241-2244, 1988 18. Nissen SL, Van Huysen C. Haymond MW: Measurement of plasma alpha-KIC concentrations and specific radioactivity by HPLC. Anal Biochem 110:389-392, 1981 19. Faloona RG: Radioimmunoassay: Glucagon and GLI. in Berson SA, Yalow R (eds): Methods in Investigative and Diagnostic Endocrinology. Part III, Nonpituitary Hormones. New York. NY. ElsevieriNorth-Holland, 1973, pp 919-924

20. Ukida M, Schafer K, Bode J: Effect of storage at -20°C on the concentration of amino acids in plasma. J Clin Chem Clin Biochem 19:1193-1195,1981 21. Barrett EJ, Revkin JH. Young LH. et al: An isotopic method for measurement of muscle protein synthesis and degradation in vivo. Biochem J 245223-228. 1987 22. Williams IH, Sugden PH. Morgan HE: Use of aromatic amino acids as monitors of protein turnover. Am J Physiol 240:E677-E681. 1981

78

23. Heslin MJ, Newman E, Wolf RF, et al: The effect of hyperinsulinemia on whole body and skeletal muscle leucine carbon kinetics in man. Am J Physiol262:E911-E918, 1992 24. Sigdell J: Venous occlusion plethysmography. I. Basic principles and applications. Biomed Engineer 10:300-302,1975 25. Sigdell J: Venous occlusion plethysmography. II. Methods. Biomed Engineer l&343-345,1975 26. Frayn KN, Maycock PF: Regulation of protein metabolism by a physiological concentration of insulin in mouse soleus and extensor digitorum longus muscles. Biochem J 194323-330, 1979 27. Bennet WM, Connacher AA, Smith K, et al: Inability to stimulate skeletal muscle or whole body protein synthesis in type 1 (insulin-dependent) diabetic patients by insulin-plus-glucose during amino acid infusion: Studies of incorporation and turnover of tracer L-[l-13C]leucine. Diabetologia 33:43-51,199O 28. Louard RJ, Barrett EJ, Gelfand RA: Effect of infused branched-chain amino acids on muscle and whole-body amino acid metabolism in man. Clin Sci 79:457-466,199O 29. Frexes-Steed M, Warner ML, Bulus N, et al: Role of insulin and branched-chain amino acids in regulating protein metabolism during fasting. Am J Physiol258:E907-E917, 1990 30. Tessari P, Inchiostro S, Biolo G, et al: Differential effects of hyperinsulinemia and hyperaminoacidemia on leucine-carbon metabolism in vivo. Evidence for distinct mechanisms in regulation of net amino acid deposition. J Clin Invest 79:1062-1069, 1987

NEWMAN ET AL

31. Arfvidsson B, Zachrisson H, Moller-Loswick AC, et al: Effect of systemic hyperinsulinemia on amino acid flux across human legs in postabsorptive state. Am J Physiol 260:E46-E52, 1991 32. Cheng KN, Dworzak F, Ford GC, et al: Direct determination of leucine metabolism and protein breakdown in humans using ~-[l-‘~c, i5N]-leucine and the forearm. model. Eur J Clin Invest 15:349-354,198s 33. Lundholm K, Schersten T: Protein synthesis in human skeletal muscle tissue: Influence of insulin and amino acids. Eur J Clin Invest 7:531-536, 1977 34. Matthews DE, Schwartz HP, Yang RD, et al: Relationship of plasma leucine and alpha-ketoisocaproate during L-[1-“C]leutine infusion in man: A method for measuring human intracellular enrichment. Metabolism 31:1105-1112,1982 35. Cheng KN, Pacy PJ, Dworzak F, et al: Influence of fasting on leucine and muscle protein metabolism across the human forearm determined using ~-[l-l~C, i5N]leucine as the tracer. Clin Sci 73:241-246,1987 36. Fryburg DA, Barrett EJ, Louard RL, et al: Effect of starvation on human muscle protein metabolism and its response to insulin. Am J Physiol259:E477-E482,1990 37. Waterlow JC, Garlick PJ, Millward DJ: Protein Turnover in Mammalian Tissues and in the Whole Body. New York, NY, Elsevier/North-Holland, 1978, p 804