Role of free fatty acids and insulin in determining free fatty acid and lipid oxidation in man

Role of free fatty acids and insulin in

determining free fatty acid and lipid oxidation in

man.

L C Groop, … , A S Petrides, R A DeFronzo

J Clin Invest.

1991;

87(1)

:83-89.

https://doi.org/10.1172/JCI115005

.

Plasma FFA oxidation (measured by infusion of 14C-palmitate) and net lipid oxidation

(indirect calorimetry) are both inhibited by insulin. The present study was designed to

examine whether these insulin-mediated effects on lipid metabolism resulted from a decline

in circulating FFA levels or from a direct action of the hormone on FFA/lipid oxidation. Nine

subjects participated in two euglycemic insulin clamps, performed with and without heparin.

During each insulin clamp study insulin was infused at two rates, 4 and 20 mU/m2.min for

120 min. The studies were performed with indirect calorimetry and 3-3H-glucose and

14C-palmitate infusion. During the control study plasma FFA fell from 610 +/- 46 to 232 +/- 42 to

154 +/- 27 mumol/liter, respectively. When heparin was infused basal plasma FFA

concentration remained constant. During the control study, FFA/lipid oxidation rates

decreased in parallel with the fall in the plasma FFA concentration. During the

insulin/heparin study, plasma 14C-FFA oxidation remained unchanged while net lipid

oxidation decreased. In conclusion, when the plasma FFA concentration is maintained

unchanged by heparin infusion, insulin has no direct effect on FFA turnover and disposal.

These results thus suggest that plasma FFA oxidation is primarily determined by the plasma

FFA concentration, while net lipid oxidation is regulated by both the plasma FFA and the

insulin level.

Research Article

Role

of Free Fatty Acids and Insulin

in

Determining Free Fatty Acid and Lipid Oxidation in Man

Leif C. Groop, Riccardo C. Bonadonna, Myron Shank,Alexander S.Petrides, andRalphA. DeFronzo

Fourth Department ofMedicine, Helsinki University Hospital, Helsinki, Finland; Division ofEndocrinology/Diabetes, Department ofMedicine, Yale University School ofMedicine, New Haven, Connecticut 06504;

and Diabetes Division, Department ofMedicine, University ofTexasHealth ScienceCenter, and Veterans Administration Hospital, San Antonio, Texas 78284-7886

Abstract

Plasma FFA oxidation (measured byinfusionof

14C-palmitate)

and netlipid oxidation(indirect calorimetry)arebothinhibited

byinsulin.The presentstudywasdesignedto examine whether theseinsulin-mediated effectsonlipid metabolism resulted from a decline in circulating FFA levels or from a direct action of the hormone onFFA/lipid oxidation.Ninesubjects partici-pated in twoeuglycemic insulin clamps, performed with and without heparin. During each insulin clamp study insulin was infused at two rates, 4 and 20mU/mi2 minfor 120 min. The studieswereperformedwith indirectcalorimetryand

3-3H-glu-cose and '4C-palmitate infusion. During the control study plasma FFA fell from 610±46 to 232±42 to 154±27

;tmol/

liter,respectively. When heparin was infused basal plasma

FFA concentration remained constant. During the control study,FFA/lipidoxidationratesdecreased in parallel with the fall in the plasma FFAconcentration.During the

insulin/hep-arin study, plasma

14C-FFA

oxidation remained unchanged while net lipid oxidation decreased. In conclusion, when the plasma FFA concentration is maintained unchanged by heparin

infusion,insulin has no direct effect on FFA turnover and

dis-posal.These results thus suggest thatplasmaFFAoxidationis primarily determined by the plasma FFA concentration, while

netlipid oxidation is regulated by both the plasma FFA and the insulin level. (J.Clin.Invest.1991.87:83-89.)Key words: free

fatty acid* insulin-lipid oxidation*glucose disposal

Introduction

Innormal anddiabetic subjects insulin is known to suppress

lipid oxidationasmeasured by indirect calorimetry (1, 2). How-ever,indirectcalorimetry measures netlipid oxidation which

includesbothoxidationof plasma FFA and oxidation of intra-cellularlipids (3-5). The oxidation ofplasma FFA, which can be measuredwith

'4C-labeled

palmitate, is inhibitedby small

incrementsin the plasmainsulinconcentration (6, 7). The

de-creasein FFAoxidationcould result from decreased substrate

level, secondary to an inhibition of lipolysis, or could result

fromadirect effect ofinsulin on FFA oxidation. Because both

Address correspondencetoLeif C. Groop, M.D., Fourth Department of Medicine, Helsinki University Hospital, Unioninkatu 38, SF-001 70 Helsinki, Finland.

Receivedfor publication17November 1989 and inrevisedform19

July1990.

lipolysis and FFA oxidation may be regulatedin concert by

insulin, it is difficulttodrawconclusions about theeffect(s)of

insulinonlipid/FFAoxidation when the plasmaFFA concen-trationischanging.Tocircumvent thisproblem,wehave exam-inedtheeffectof insulinonlipid/FFAoxidation under

condi-tionswheretheplasmaFFAconcentration has been kept

con-stantatthe basal levelbyaninfusionofheparin.A second aim ofthestudywas toexamine the effect of differentFFA concen-trationson thesuppressionofhepatic glucose productionand

stimulation of glucose disposal by insulin. Our data provide evidencethatinsulin hasnodirect effectonplasmaFFA oxida-tionbutmarkedlyinhibits the oxidation of intracellularlipids.

Methods

Subjects.Ninehealthy youngsubjects (seven males,twofemales)with

amean(±SEM) age of 22±1 yr and ideal bodyweightof 102±2%were

studied. Allsubjectshadanormal oralglucosetolerancetestaccording totherevised criteria ofthe National Diabetes DataGroup.Thefasting plasma glucosewas86±1 mg/dl andfasting plasmainsulin

concentra-tionwas 7±1 uU/ml.Noneof thesubjectshad any clinicalor labora-tory evidence ofhepatic, renal,orother endocrine disease. Allsubjects wereconsumingaweight-maintainingdietcontainingatleast 200 g of carbohydrate per day for 2wkbeforestudy.Beforeparticipatingin the

experimental protocol,the purpose, nature, andpotentialrisks of the studywere explainedtoallsubjectsand informed writtenvoluntary consent wasobtained. Thestudyprotocolwasapproved bythe Human

InvestigationCommittee of the YaleUniversitySchool of Medicine.

Experimental protocol.Allsubjects participated in twoexperiments

which were carried out in random order witha 2-wkinterval. During

thefirstexperimentatwo-stepeuglycemicinsulin clamp (+4 and +20

mU/M2

_min)wasperformed (8) with each insulin infusion lasting 120

min.Thesecondexperimentwassimilartothefirst withoneexception; atthe beginningof the 4mU/m2-min insulin infusion subjects

re-ceivedaprimed(200U)-continuous(25 U/min) infusion ofheparinto

stimulatelipoprotein lipase.Atthestartofthesecondinsulin infusion step(20

mU/M2_

min)subjectsreceivedasecondheparinbolus(200

U) followed byacontinuous infusionat35 U/min.Allstudieswere performedwith indirectcalorimetryand infusion of3-3H-glucoseand

'4C-palmitate

toquantitate respiratorygasexchange, hepaticglucose production, andFFAoxidation, respectively. The low insulin infusion protocol (4

mU/Mi2.

min)wasdesigned to examine whether differences in the plasmaFFAconcentration would influence the effect of insulin

ontheliver. The 20

mU/M2_

min insulin infusion protocol,onthe otherhand, allowedus toexamine whether differences in theplasma

FFA concentration would affectthecapacity of insulin tostimulate

peripheralglucose uptake.

Euglycemicinsulinclamp.Thesubjects were admitted to the Clini-cal Research Centeronthemorning of study.Allstudieswereinitiated

at0800aftera 12-hovernight fast. Before the study Teflon catheters

wereinserted into an antecubital vein for infusion of test substances andretrogradely intoawrist vein for withdrawalof blood samples. The

hand was theninserted into a heated box (70°C) to achieve arterializa-tionofvenousblood. Afterobtaining four basal samples for insulin, glucose, andFFA aprimed-continuous infusion (4

mU/Mi2.

min)of

J.Clin.Invest.

© The AmericanSocietyforClinicalInvestigation,Inc. 0021-9738/91/01/0083/07 $2.00

crystalline porcine insulin (Lilly, Indianapolis, IN) was started to acutely raise and maintain the plasma insulin concentration at the desired level for 120 min. The plasma glucose concentration was deter-mined every 5 min and a variable infusion of 20% glucose was adjusted to maintain the plasma glucose concentration constant at the basal level (8). At 120 min the insulin space was reprimed and the second insulin infusion (20 mU/m2 * min) started and maintained for another 120 min. Blood samples for plasma insulin and FFA concentrations weredrawn at15-min intervals.

Tritiated glucose.At150 minbefore the start of insulin, a primed (25 uCi)-continuous (0.25 XCi/min) infusion of3-3H-glucose (New England Nuclear, Boston, MA) was begun. Baseline samples for deter-mination of 3-3H-glucose specific activity were drawn at 5-min inter-valsduring the last 30 min of the equilibration period. During the insulin clamp, 3-3H-glucose samples were drawn at 15-min intervals until the last 20min,during which time they were obtained at 5-min intervals. All subjects achieved a steady-state level of plasma tritiated glucosespecific activity at the end of the basal equilibration period.

FFAturnover. 14C-Palmitate(New England Nuclear) was utilized to quantitate plasma FFA turnover (6, 9). A priming dose of 2.5

_gCi

14C-palmitatewasinjected150 minbefore the first insulin infusion and this was followed byaconstantinfusion at 0.1 Ci/min throughout the study. Simultaneously,abolusinjection of 3.5MCiof [I-'4C]-NaHCO3

(New England Nuclear) was given to prime the bicarbonate pool. Base-line samples for determination of '4C-FFA plasma specific activity weredrawn at 5-min intervals during the last 30 min of the equilibra-tionperiod.Duringtheinsulin clamp '4C-FFA samples were drawnat

30, 60, 75, 90, 100, 110, 120, 150, 180, 195, 210, 220, 230, and 240 min.

FFAoxidation. Therateof FFA oxidationduringthe'4C-palmitate

infusion was calculated from the specific activity of expiredCO2in air

samplesobtainedat5-min intervalsduringthelast 30 min ofthe

equili-brationperiodandat60, 90, 100, 110,120,180, 210, 220, 230,and 240 minof the insulin clamp (6, 10). Each expired air samplewasbubbled

througha3-mlsample of carbon dioxidetrappingsolution (IM

hya-minehydroxide:absoluteethanol:0. 1%phenolphtalein;3:5:1). The ca-pacityofthetrappingsolutiontotrapImmolofCO2waschecked with

I NHCl.The'4C-radioactivitywassubsequently determinedon dupli-catesinaPackardTricarb Scintillationcounter(PackardInstrument Co. Inc.,DownersGrove,IL)andtheexpired'4CO2specific activity

calculated. Total'4C02expiredper minutewasdetermined by

multi-plyingthe'4CO2 specific activityby the totalCO2 productionrate de-termined by indirectcalorimetry (seebelow).

Respiratory gasexchangemeasurements. Allstudieswere per-formed with continuous indirectcalorimetrytoestimatenetratesof

carbohydrateandlipidoxidation(3). Starting60 min before and con-tinuingfrom 60to 120 min and from 180to240 minof the insulin clamp, continuous gaseousexchange measurementswereperformed

withaventilated hood system(Vista,Vacumed, Ventura, CA). Venti-lationwasmeasuredbymeansofadry gasmeter. A constantfraction of the airflowingoutofthe hoodwasautomaticallycollectedforanalysis.

Oxygenwasmeasuredby electrochemicalanalysisandCO2contentby aninfraredanalyzer using Applied ElectrochemistryInstruments

(Sun-nyvale, CA). Protein oxidationwascalculatedfrom theurinary nitro-gen excretion obtained before andduringtheinsulinclamp (3).

Analytical

determinations

Plasmaglucose wasdetermined induplicate by the glucose oxidase methodon aglucoseanalyzerII(BeckmanInstruments, Inc., Fuller-ton,CA). Methods fordeterminationofplasmatritiatedglucose spe-cificactivityhave beenpublishedpreviously (I 1).Bloodfor determina-tion ofinsulin, FFA and'4C-FFAspecific activitywassampledinto

prechilledtubes andimmediately centrifuged.Plasmainsulinwas

mea-suredby specific radioimmunoassay (12), urinary nitrogen excretion

bythemethodofKjeldahl (13),andplasmaFFAconcentrationswere

measuredbythemicrofluorometricmethodof Milesetal.(14).The methodfor determination of'4C-FFAspecific activityinplasmahas beendescribedin detailpreviously (6).

Calculations

Hepaticglucoseproduction (HGP).BasalHGPwascalculated by

divid-ingthe3-3H-glucoseinfusionratebythesteady-state plateauofplasma 3-3H-glucose specific activityachievedduringthe last 30 min of the basaltracerinfusionperiod. After administration of insulin and

glu-cose anon-steadystatecondition inglucosespecific activity exists,and

the rateofglucoseappearance

(R.)

wascalculatedbythe model de-scribedbyRadziuk(15).Theinfusionrateof coldglucosewas

inte-gratedover20-min intervals and subtracted from the totalR. calcu-latedduringthesame20-minperiodtoobtain therateof HGP.

Nega-tive numbers for HGPwereobserved occassionally duringthe 20

mU/mi2

min insulinclampstep.Asrecentlydemonstratedon theoreti-calaswellasexperimentalgrounds (I16),suchunderestimation of

glu-coseturnoverbythetracermethod islargelyaccountedforbyamodel

erroremergingathigh ratesofglucosemetabolism. Becauseasimple

andsatisfactorycorrectionforchangesin theglucosesystemduringthe

non-steadystateisnotyetavailable,wetook thenegativenumbersto

indicatecompletesuppressionof HGP.

Glucosemetabolism. Duringthe insulin clampthemeanglucose infusion rate wasdeterminedduringthe last60 min of each insulin

clampstep. Totalbody glucosemetabolismwascalculatedbyadding therateof residualendogenous glucoseproduction duringthe last 60 min of each insulinclampsteptothemeanglucoseinfusionrateduring thesame_period.Nonoxidativeglucosemetabolismduringthe last60 min of each insulinclampstepwascalculatedasthedifference between totalbody glucoseuptakeandglucoseoxidationasdeterminedby

indi-rectcalorimetry.

Indirectcalorimetry.Netglucoseandlipidoxidationrates were esti-matedfrom indirect calorimetricmeasurementsin the basalstateand duringthelast 60 min of each insulin infusion step. Theconstantsto

calculateglucose,lipid,and

protein

oxidation from gasexchangedata andnonproteinurinary nitrogenexcretionarethose in reference 3. To allowdirectcomparisonwith '4C-FFAoxidation, net

lipid

oxidation (in milligrams)has beenconvertedtomolesby dividing bythe molecu-lar weight ofpalmitate (mol wt256). At a

non-protein

respiratory quotient> 1.0,theequationfor the calculation of substrate oxidation remains valid(3);theresulting

negative

valuefor

lipid

oxidation is in factequivalentto netfatsynthesis (3).

Plasma FFA turnover. Palmitic acidaccountsfor - 30% of the

totalFFApool

independently

of the

plasma

FFAconcentration

(17).

Becausethefractionalturnoverof

_palmitate

is very similartothatof totalFFA,labeled

palmitate

canbeusedtotracetotalFFA

_{( 17).}

The

plasmaFFAturnoverratewascalculatedas

_palmitate

infusion rate

dividedbythesteady-state

plasma

FFA

specific activity

and is

ex-pressedas

gmol/kg.

min.Usingthis

approach

the

specific

activities of both plasma "4C-FFAandof

'4CO2

in

_expired

airarein

steady

state

duringthe last 30 min ofthe

equilibration period

andanew

steady

state

is achievedduringthe last 60 min of each insulinclampstep

(Fig.

1)

(6).

We have therefore assumed

steady-state

conditionsto calculate the

ratesofFFAturnoverandFFAoxidation.

Metabolicclearancerate

_of

FFA

(MCRFFA).

Themetabolic

clear-ance rateofplasmaFFA

(ml/kg- min)

wascalculated

by dividing

the

basal or_steady-state

plasma

FFA turnover rate

_by

the

plasma

FFA

concentration

during

thesame

_period.

FFAoxidation. Therateofoxidation of

plasma

FFAin the basal

stateand

during

the

_euglycemic

insulin

_clamp

wascalculated from the

14C

radioactivity

in

_{expired CO2}

divided

_by

the

_product

of the

plasma

FFA

specific activity

andafactor

k,

which takes intoaccountthe in-completerecovery of labeled

'4CO2

from the bicarbonate

pool (6, 10).

FFA oxidation rate

_(;Lmol/kg

min)

=

_(S.A.CO2)

X VCO2/k

x

(S.A.'4C-FFA),

where

S.A.'4C02

=

_specific

_activity

of

_CO2

in

expired

air; VCO2 = total

CO2

production

(to

change ml/min

to

_Mmol/min,

divideby

22.4);

S.A.'4C-FFA=

specific activity

of14C-FFAin

plasma,

and k=_0.81.

NonoxidativeFFA

disposal.

NonoxidativeFFA

disposal,

ameasure

50(

E

E

U

30C

FFASPECIFIC ACTIVITY

D

)r0 -

l

4

--111llJ

4

I

A

~~~~~~~~~~

I~ I

loo1 OL

4000

1=

0 E

iE2000 E

UL

-I

CO2SPECIFIC ACTIVITY

-I-T

t

---r----rTj

OL

_{-30 0 30 60 90 120 150 180 210 240}

Time(min)

Figure 1. The specific activity of FFA in plasma (top) andof CO2in

expired air(bottom) in the basalstateandduringatwo-step euglycemic insulin clamp (insulin infusion rates of4and 20

mU/M2.min)performed with saline(brokenline) and with heparin

(solid line).Values represent the mean±SEM (n=9).

Intracellular (unlabeled)FEAoxidation was calculated as the

differ-encebetween netlipid oxidation (indirect calorimetry) and plasma FFA oxidation. It should be noted that this represents a minimal

esti-mate of intracellular FFA oxidation because net lipid oxidation, as determined by indirectcalorimetry,equals the difference betweenlipid oxidation and lipidsynthesis.

Statistical analysis

Dataare presented as mean±SEM. Significance of differences between thetwoexperiments was calculated with analysis ofvariance. Pearson's correlation coefficientswerecalculatedby standardformulae.

Results

Plasma glucose, insulin, andFFAconcentrations. The fasting plasmaglucoseconcentration duringthe control andheparin

experiments was 86±2 and 85±2 mg/dl, respectively. The

meanplasma glucose duringthe 4 and 20mU/mi2 min insulin

clamp steps ofboth studies was similar and averaged 84±2 mg/dl. Thecoefficients ofvariationofthe plasma glucose were 3.9±0.5% and3.7±0.5%, respectively, duringcontrol and

hepa-rin experiments.Thefastingplasmainsulin concentration

dur-ing thecontrol study was 6±1

_AtU/ml

and rose to I I± and 36±2MU/ml, respectively, during the 4 and 20mU/M2_min

insulinclamp steps(Fig. 2). The basal plasma insulin concen-tration in theheparin experimentwas 7±1

gU/ml

and rose to

11i 1 and36±2uU/ml,respectively. The coefficients of varia-tion inplasmainsulin during the control and the heparin

ex-perimentwere7.4±0.8%and7.5±0.9%, respectively.

The basal plasma FFA concentration during the control

studywas610±46

_,mol/liter

andfell to 232+42 and 154±27

,uU/liter, respectively, duringthe 4 and 20 mU/mi2 mininsulin clampsteps(Fig.2). The basal plasma FFA concentration

dur-ing the heparin study, 646±77 ,umol/liter was maintained at 725±129 and 524±100

Mtmol/liter,

respectively, during the 4

and 20

mU/mi2

mininsulinclamp steps.

PLASMA FFA 1200_

1000

800X

` _600-

tto

E

400 "k

200_

0_

PLASMA INSULIN

F

401-E

--30 0 ₃₀ 60 90 _{120 150 180} 210 240

Time(min)

Figure2.TheplasmaFFA(top)and insulin(bottom)concentrations in the basalstateandduringatwo-stepeuglycemic insulinclamp

(insulininfusionratesof 4 and20mU/Mi2.min) performedwith saline(broken line)andwithheparin (solid line).Values represent the mean±SEM(n=9).

FFA

metabolism

FFAturnover rate.ThebasalrateofFFAturnover wassimilar

during the control and the heparin experiments (5.97±0.67vs.

6.46±0.6

gmol/kg.

min) (Fig. 3). Duringthe control

experi-menttheplasma FFAturnover ratedecreased by 50%during

thelow-dose insulin infusion (to 2.99±0.39

Amol/kg

*min)and

FFATURNOVER RATE

8r

E

E

FFA METABOLIC CLEARANCE RATE

20

_.

I

CL

10

+1~~~~~~~~~**+1

Plasma Insulin (pU/mi)

Figure 3. The plasma FFA turnover rate (top) and metabolic clearance rate of FFA (bottom) in the basal state and during a two-step euglycemic insulin clamp (insulin infusion rates of 4 and 20

mU/in2. mmn)performed with saline (open bars) and with heparin (shaded bars). Values represent mean±SEM (n = 9). ***P <0.001,

no significant further reduction was observed during the 20

mU/Mr2.

min insulininfusion step (2.61±0.4). During the in-sulin clamp study performed with heparin, the plasma FFA turnover rate was notsignificantly different from baseline dur-ing the 4 and 20 mU/M2.min insulin infusion steps (6.01±0.77 and 5.42±0.69 ,umol/kg *min,respectively). How-ever,the plasma FFA turnover rate during the insulin clamp

performed withheparin was significantly greater than that dur-ing the control study (no heparin) (P < 0.01).

Metabolicclearance rate ofFFA (MCRFFA). In the control

experimentthebasal

_MCRFFA

rosefrom 9.6±0.8 ml/kg- min

to 13.7±1.0and 17.6±1.2 ml/kg *min,respectively, during the 4and20mU/M2_mininsulinclamp steps(Fig. 3). In contrast,

during heparin infusion the basal MCRFFA (10.3±0.7

ml/

kg min) did not change significantly during either the 4 (9.3±1.0 ml/kg-min) or 20 mU/m2.min (11.6±1.2

ml/

kg.min)insulinclamp steps.

Netlipid oxidation. The basal rate of total net lipid

oxida-tionwas3.36±0.21 and 3.12±0.37

_,mol/kg

*min,respectively,

duringcontrolandheparin experiments(Fig. 4). During low-doseinsulin infusion, lipid oxidation did not change

signifi-TOTAL LIPID OXIDATION

4-

-3 ,

3-

---E

E>

1- l'0 l

ox---

I-.--...

I..

....

FFAOXIDATION

C

E 2k

E

II1g

*4E

A

NON-OXIDAT.VE

oLi1...

.F

,

MA...B.O..I

...,

NON-OXIDATIVE FFA METABOLISM

5-c

_4!.

d

i

=. 3k

E

I2

0L-+1

T--::_{a, ,}

,.,"' "', ,l

...-,,.,

'"''""' '4

a,-,-.-...-.. C

A.'.',' '. .;

1 '' <

;-.-....,.

.,.,.'.',.. s

:--'::'S

I,., ,...,' :.';

i, .-.,.-,

A,'-.-I',....'.'.' :.r

t

r.-.-...s

+l +l

.-..:...

..

to

1.,+1,..,,

PlasmaInsulin (pU/ml)

Figure4.Theratesofnetlipidoxidation, plasmaFFAoxidation and nonoxidativeFFAmetabolism inthebasalstateandduringa

two-stepeuglycemicinsulinclamp(insulininfusionratesof4and20 mU/M2_min) performed with saline (openbars)andwithheparin (shadedbars).Values represent mean±SEM(n=9).*P<0.05,**P

<0.01,significanceof differencefrom theheparinexperiment.

cantly in either control(2.98±0.21 umol/kg-min) orheparin

(3.37±0.41) experiments. Duringthe 20mU/M2.min insulin

infusion,netlipidoxidation fellinboth thecontroland

hepa-rinstudies(P<0.01);however, thereductionwassignificantly

greater in the control experiment(0.39±0.16 vs. 1.24±0.32

pmol/kg. min;P <0.01).

FFA oxidation. The basal rateofplasma FFA oxidation,

which accounted for - 50% of basalnetlipid oxidation,was

similar in control (1.57±0.15

_gmol/kg

min) and heparin

(1.50±0.14

_;tmol/kg-

min)studies(Fig. 4). Inthe control

ex-periment, plasma FFAoxidation wasreducedby 45%atthe

low-dose insulin infusion (0.86±0.15

,tmol/kg

min) and did

notchangefurther duringthe 20

mU/M2

_mininsulininfusion

(0.92±0.17

gmol/kg

*min).WhenplasmaFFAconcentration wasmaintainedwith heparin, theplasmaFFAoxidationrate remainedconstant atthe basal levelduringthe 4and 20

mU/

m2.mininsulinclamp steps(1.57±0.18and 1.61±0.18

,mol/

kg min,respectively; bothP <0.01 vs.controlstudy). NonoxidativeFFAmetabolism. Thebasal rate ofnonoxida-tiveFFAmetabolismwassimilarincontroland

heparin

exper-iments (4.40±0.57vs.4.96±0.46

_;imol/kg-

min)

(Fig.

4).

Dur-ingthe4and20

mU/Mi2.

mininsulin clampstepsofthe

con-trol study the basal rate of nonoxidative FFA metabolism declined by 51 and62%,

respectively.

When the

plasma

FFA

concentrationwasmaintainedatthe

fasting

level with

heparin,

therateofnonoxidativeFFAmetabolism declined

slightly,

al-though not

significantly

from baseline and remainedgreater thanduringthe control

experiment

(P<

0.01).

Duringbothcontroland

heparin

studies plasmaFFA con-centrationwasstrongly correlated withtherateof

plasma

FFA

oxidation (r=0.71;P<

0.001;

n =

54)

(Fig. 5),

with therateof

nonoxidativeFFA

disposal

(r=0.84;P<

0.001),

and

weakly

withthe rateofnetlipid oxidation (r= 0.31;P<

0.02).

Glucose metabolism

Hepatic

glucose

production

(HGP).Thebasalrateof HGPwas similar

during

thecontroland

heparin

studies

_(2.0±0.13

and 1.90±0.08

mg/kg.

min)(Fig. 6).

During

low-doseinsulin infu-sion,HGPwasreducedby 50%incontrol and

heparin

experi-ments(to 1.03±0.13 and1.19±0.18mg/kg. min,

respectively).

Duringthe20

mU/m2.

mininsulin infusion, HGPwas com-pletely suppressedin bothexperiments.

Total glucosedisposal. Inthe basal state the rateof glucose disposal equalstherateof glucoseappearance

(HGP)

andwas

similarin control and

heparin

studies

(Fig.

6). Therewas no

significant

enhancementof glucose

disposal

during

the 4

mU/

m2._{mininsulin infusionineitherthe controlor}

_heparin

stud-4

r 0.71

EiC p'O.001

3 _/

21

_ . * w *

/Figure

5. C. rrelation

between theplasma

X a. * FFAconcentrationand

0 the rate ofplasmaFFA

U.; s oxidation* in thebasal

stateand duringthe 4 and 20

mU/M2.

min

HEPATIC GLUCOSE PRODUCTION

3F

c

E Ch

En

E

c

E

NON-OXIDATIVE GLUCOSE METABOLISM

GLUCOSE DISPOSAL 107

8 _

iu ci ofad2mUr.4

_ ,.

Plasma InsulinlpjU/mI)co c

Figure6. Theratesofhepatic glucose production (top)andglucose disposal(bottom)in the basalstateandduringatwo-stepeuglycemic insulinclamp(insulininfusionratesof 4 and 20

mU/m'-

min) performed with saline (open bars) and with heparin (shaded bars).

Values representmean+SEM(n=_9).

ies(2.34±0.21 vs.2.39±0.24mg/kg. min).Duringthe20

mU/

m2 - min insulin clamp step, therateof glucose

disposal

rose similarlyinthe control

(7.52±0.72 mg/kg- min)

and

heparin

(6.98±0.60) studies.

Glucose oxidation. The

respiratory quotients

(RQs) inthe

basal stateand

during

the +4and +20

mU/m2'

mininsulin clamps were

during

the control

study 0.82±0.01, 0.83±0.01,

and0.95±0.01, respectively, and

during

the

heparin

study 0.83±0.01,

0.81±0.01,

and0.90±0.02,

respectively.

The basal rateofglucoseoxidationwassimilarin control and heparin experiments (1.52±0.15 vs. 1.60±0.15 mg/

kg . min; P> 0.1)(Fig. 7). Duringlow-dose insulin

infusion,

therewas no

significant

changein therateof glucose oxidation

in the control study (1.55±0.15 mg/kg -min), whereas there was aslight reduction during theheparinstudy(1.19±0.20 mg/

kg-min; P<0.05 whencomparingthechange frombasal

be-tweenthetwoexperiments).

During

the20

mU/m2.

min

insu-lininfusion,the rateofglucoseoxidationwassignificantly re-duced during theheparin vs. control study(2.80±0.35 vs. 3.71±0.24mg/kg . min; P<0.05).

Nonoxidative glucose metabolism. The basal rate of nonox-idativeglucosemetabolismwassimilarinthe two

experiments

(0.48±0.12vs. 0.32±0.14mg/kg.min)(Fig. 7).Therewas no significant change intherateof nonoxidativeglucose

metabo-lismduringthelow-doseinsulin infusionin the control study

(0.79±0.22 mg/kg-min), whereas therewas anincrease inthe

heparinstudy(1.19±0.36;P<0.05).Although therateof non-oxidative glucose disposal was enhancedduring the 20

mU/

m2.

mininsulin infusion inboth the control and theheparin studies,there wasnosignificant differencebetween the control andtheheparin study

(3.81±0.66

vs.4.18±0.65 mg/kg-min).

c,

E

E

PlasmaInsulin(jiU/ml)

Figure 7. The rates ofglucoseoxidation(top)andnonoxidative glucosemetabolism (bottom) in the basalstateandduringatwo-step

euglycemicinsulin clamp(insulininfusionratesof 4 and 20 mU/m2' min)performed with saline (openbars)and withheparin (shaded bars).Valuesrepresentmean±SEM(n =9).*P<0.05, significanceof difference from theheparinexperiment.

Discussion

Inthe present studywehaveexaminedtheindependent effects of insulin andFFAontotallipid oxidation,measuredby indi-rectcalorimetry,andplasma FFAoxidation, measured by

14C-palmitate.

Wheninsulinwasinfusedto createphysiologic

in-crementsintheplasma insulin concentration(36±2

,gU/ml),

bothnetlipid oxidationandplasmaFFAoxidation declinedin

parallel. However,because theplasmaFFAconcentrationalso

declined duringthe infusion of

insulin,

itwas not

possible

to definewhether thedecline in plasma FFA/net

lipid

oxidation

wasdue to a direct inhibitoryeffect ofinsulin or to a decrease in substratedelivery. Todistinguishbetweenthesetwo

possibili-ties,

theplasmaFFAconcentrationwas"clamped"atthebasal

level by an infusion of heparin toactivate lipoprotein lipase.

Under these conditions insulin failed to inhibit plasmaFFA

oxidation butdecreased netlipid oxidation by>60%. These resultspointattheexistenceoftwodistinct lipidpoolswhich

share the same oxidative pathway but whose input into this oxidative pathway isdifferentiallyregulated byinsulinand

plasmaFFAsupply.Oxidation of14C-palmitate primarily

re-flectsthe

circulating

plasmaFFApool (4, 5, 19) and is unaf-fected by insulin. Incontrast, net

lipid

oxidation reflects both circulating plasma FFA aswell as a large intracellular FFA

pool,whoseoxidation is sensitiveto

physiologic

changesin the plasmainsulin concentration.

InthebasalstateplasmaFFAoxidation accounted foronly

40-50% ofnet

lipid

oxidation. Our results indicate that the

remaining50-60% is derived from oxidation of intracellular lipids, whichare notin

equilibrium

withtheplasmaFFApool

andthereforeare nottracedbythe

i4C-palmitate.

Similardata have beenreported byIssekutzetal.(5) in dogs.

Dagenais

etal. (4),

using

the forearm balance technique,also have

provided

evidence that intracellular oxidation of

triglycerides and,

possi-bly phospholipids, make a substantial contribution to total lipid oxidation in man. Similar results have been reported by Robin et al. (19) in nutritionally depleted patients who were

suffering fromacute infection or injury.

Consistent with the concept of a large intracellular lipid

pool, Eaton et al. have shown that the plasma FFA pool ex-changeswitha 100largertissue pool, the bulk of which resides within the liver and muscle (20). Further support for this view has been provided by George and Naik, who demonstrated

abundant lipid deposits in resting muscle (21). Hydrolysis of these intramuscular triglycerides/phospholipid stores with

theirsubsequent intramuscularoxidation, or hydrolysis of

li-poproteins in thecapillarybedwith immediate oxidation by

adjacent muscle tissues, could explain the large discrepancy between totallipid oxidationand plasma FFA oxidation (22). From the present resultsit can be concluded that physio-logic increments in the plasma insulin concentration do not have adirect effecton theoxidation ofcirculating FFA. The plasma FFAconcentrationappears to be the main determinant

of itsown rateof oxidation. Several observationssupportthis conclusion. First, inhibition of lipolysis by physiologic hyper-insulinemiacauses aconcomitant decline inthe plasma FFA

oxidation rate, which closely parallels the decline in plasma FFAconcentration.Second, when the plasma FFA

concentra-tionwas maintainedconstant byan infusion of heparin, the

rate ofplasma FFA oxidation was uninfluenced by insulin. Third,theplasmaFFAconcentration showedastrongpositive correlation withthe rateofplasmaFFAoxidationand asimilar correlation has beenreported indogs (10). Lastly, in vitro

ex-periments have demonstratedthat the rateofFFAoxidation increases in proportionto the

increase

in mediumFFA

con-centration (23). Taken together, ourdata, aswell as thoseof others, indicate that therateof FFA oxidation is

predomi-nantly regulated bytheplasmaFFAconcentration.

Duringthe 20

mU/M2.

min insulin clamp step netlipid oxidationdecreased even though the plasma FFA concentra-tion remained unchanged. This could bedue to one oftwo

possibilities;

insulin couldsuppress net

lipid

oxidation

directly

by

inhibiting

theoxidationof intracellular

lipid

stores,or

inhi-bition of

lipid

oxidation couldbe

secondary

tostimulation of glucose oxidation by insulin (the Randle cycle). Our results

suggest that insulin regulatesnet

lipid

oxidation viatwo dis-tinct mechanisms.While its

major

action istoinhibit intracel-lular lipid oxidation (ortostimulate glucose

oxidation),

insulin alsoexerts asecondary effectontotal

lipid

oxidation

by

modu-latingtheplasma FFAconcentration via its

inhibitory

action

on

lipolysis. Independent

of

changes

in the

plasma

FFA

con-centration,

insulindoesnotappeartoaffect

plasma

FFA oxida-tion.

Some comment

concerning

the

discrepancy

between net lipid oxidation measured by indirect

calorimetry

and

plasma

FFA oxidation measured by thetracer

during

the +20

_mU/

m2

*minis warranted. If

_lipogenesis

from

glucose

is stimulated

by

insulin,

this will be measuredas an apparentdecrease in lipid oxidation,eventhoughnoactual decrease in theoxidative flux of

lipid

carbons hasoccurred. This could result in underes-timationofthetrue rateof

lipid

oxidation. The tracer-derived

FFA

oxidation

measures the amountof

lipid

carbons which

entertheoxidative pathway from the

plasma pool.

Anincrease in therecovery

factor

(which

takes intoaccountthe

incomplete

recovery of labeled

CO2

from thebicarbonate

pool

andwas

assumed tobe

0.81) during

hyperinsulinemia

could resultin overestimation oftracer-derived FFA oxidation. Recent

evi-dence in

dogs

suggest

that,

inthe fedstate, the recoveryfactor

may

slightly increase, i.e.,

theretention of

CO2

decreases(24).

Although

wedidnot measurerecoveryof

14CO2 during

hyper-insulinemic

conditions,

it is

unlikely

thatthis would

signifi-cantly

influence the results. Anincreasein the recoveryfactor from 0.81to0.90 wouldat mostresult ina10%overestimation of the tracer-derivedFFAoxidationrate.

Still, during

the +20

mU/m2.*

mininsulin

clamp

withsaline therateof

plasma

FFA

oxidation measured

by

thetraceris twiceas

high

astherateof

lipid

oxidation measured

by

indirect

calorimetry.

Assomeof

the

subjects

presented

with

RQ

> 1

during

this insulin

infusion,

we

interpret

the

discrepancy

betweenplasmaFFA oxidation

and net

lipid

oxidationas anindicationthatinsulinincreases

significantly lipogenesis

from

glucose.

Our results also

help

to

clarify

themechanisms which regu-late the

plasma

FFAconcentration in the basal and insulin-stimulatedstatesin

healthy

subjects.

The

plasma

FFA

concen-tration is determined

by

the rate ofFFA appearance in the

systemic

circulation

(lipolysis)

andtherateofFFA

disappear-ance

(oxidation

and

_{reesterification).}

Inthe

_{postabsorptive}

state

plasma

FFAoxidation accounted for 30% of totalFFA turnover,while the

remaining

occurred via nonoxidative

path-ways,

i.e.,

reesterification ofFFAto

triglycerides.

Therateof nonoxidativeFFA

disposal

canbeestimated from the differ-encebetween therateof total

body

FFAturnoverand therate

ofFFAoxidation

(6,

18).

Ourdataindicatethatthe

_major

_part

(70%)

ofFFA

disposal

during

the basalstateoccurredvia

non-oxidative

pathways.

Insulin

suppressed

nonoxidative FFA

dis-posal by

52 and

62%,

respectively,

during

the4and20

_mU/

m2._mininsulin

_clamp

_studies.

_However,

nonoxidative _FFA

disposal

still

comprised

70% of total FFAturnover.

_Thus,

in boththe basal and insulin-stimulated

conditions,

reesterifica-tion

represented

the

major

routeof

plasma

FFA

disposal.

Dur-ing

theinsulin

clamp

studies

performed

with

heparin

therate

of nonoxidativeFFA

disposal

failedto

_change

from

_baseline,

suggesting

that insulinpersehasnodirecteffectonFFA

rees-terification. Thedeclinein nonoxidative FFA

_disposal

ob-served when insulinwasinfusedwithout

heparin

most

likely

is

secondary

tothedecline in

plasma

FFAconcentration.

Consis-tent with thissequence, a strong

positive

correlationwas

ob-servedbetweentherateofnonoxidativeFFA

_disposal

and the

plasma

FFAconcentration.These

findings

areconsistent with

the concept thatthe

plasma

FFAconcentrationisa

major

deter-minantofnonoxidativeFFA

disposal

orreesterification.

Sev-eralother

pieces

ofevidence support this

concept.

First,

therate

of

incorporation

ofFFA into

triglycerides

rises when the

plasma

FFAconcentration is increased

(20).

Second,

a

positive

correlation hasbeenshownbetweentherateofFFAturnover and the appearanceofFFAin

triglycerides

in

_{dogs (25)}

and in

man

(26, 27).

Third, glucose

feeding

in normal

subjects

leadsto a decrease in

plasma

FFA concentration anda concomitant

reduction intherateofFFAreesterification

_(19).

The presentresultsalso

provide

some

interesting

informa-tion

concerning

theinteractionbetween

lipid/FFA

and

_glucose

metabolism. Morethan 20 years ago Randle and co-workers

developed

the concept of substrate

competition, i.e.,

an

in-creaseinFFAoxidation leadstoadecrease in

glucose

oxida-tion and viceversa

_(28).

We

_previously

have showninnormal

subjects

that the infusion of

_{Intralipid during}

aninsulin

clamp

concen-tration causedasignificant reductionin theratesofoxidative and nonoxidative glucose disposal (29). In addition, the ability ofhyperglycemiatoinhibit basal HGPwasimpaired (29). In the present study, when the plasma FFA concentration was

maintained constant at the basal level by heparin, the stimula-tory effect of insulin onglucose oxidationwasreduced, while nonoxidative glucose disposal remained unchanged. The re-sultsindicate, that providing more substrate for FFA oxidation under these conditions willonly decrease the amount of glu-coseoxidized leaving the nonoxidative pathway unaffected. Furthermore,maintainingthe FFA concentration unchanged atthe basal level with heparin does not affect the inhibitory effect of insulin on HGP.

Insummary, the datademonstrate thatinsulin hasno

di-rect effect on turnover anddisposal of FFA when the plasma FFA concentration is maintainedunchanged. Theyfurther

suggest that oxidation of FFA fromplasma and from intracel-lularlipidstoresaredifferentlyregulated.Whereas the

oxida-tion ofplasmaFFA isprimarilydeterminedbytheplasmaFFA

concentration, the regulation of intracellularFFAoxidation is

more complex anddependent upon both the plasma FFA and

insulinconcentration.

Acknowledgments

We thank Kathleen Zych, R.N., and Madelyn Mainiero, R.T., for their expert technicalhelp and Dr. Rosa Hendlerforkindly performing the insulin determinations. Dr.Leif Groop was the recipient of a Fogarty International Fellowship (FOSTWO 3451) and Dr. Riccardo Bona-donnaofaFellowship from the ItalianMinistry of Public Education. This research was supported in part by National Institutes of Health grantAM24092 and CRC grant RR125.

References

1.Felber, J. P., E.Ferrannini,A.Golay, H. U. Meyer, D.Thiebaud,B.

Cur-chod,E.Jequier,and R.A.DeFronzo.1987. Roleoflipidoxidation in

pathogene-sis of insulinresistance of obesity and typeIIdiabetes.Diabetes.36:1341-1350.

2.Thiebaud,D., E.Jacot, R.A.DeFronzo, E.Maeder, E.Jequier,and J. P.

Felber.1982. Theeffect ofgradeddosesof insulinontotalglucoseuptake,glucose

oxidation and glucose storage inman.Diabetes.31:957-963.

3.Ferrannini, E. 1988. The theoretical basis ofindirect calorimetry: a review.

Metabolism. 37:287-301.

4.Dagenais, G. R.,R.G.Tancredi, and K. L. Zierler. 1976. Freefatty acid

oxidation byforearmmuscleatrest, and evidence foranintramuscularlipidpool

in the humanforearm. J. Clin. Invest. 58:421-431.

5. Issekutz, B., H. I.Miller, P. Paul, and K. Todahl. 1964. Source offat

oxidation in exercising dogs.Am.J.Physiol.207:583-589.

6.Groop,L.C.,R.C. Bonadonna, S. DelPrato,K.Ratheiser,K.Zyck,E.

Ferrannini,andR. A.DeFronzo.1989. Glucose and freefattyacidmetabolism in non-insulin dependent diabetes mellitus: evidence for multiple sites of insulin resistance. J. Clin. Invest. 84:205-213.

7.Robin, A. P., M. Jeevandam, D. H. Elwyn, J. Askanazi, and J. M. Kinney.

1989. Measurement offatty acid oxidation: validation of isotopic equilibrium

extrapolation. Metabolism. 38:67-72.

8.DeFronzo, R. A., J. A. Tobin, and R. Andres. Glucose clamp technique: a

method for quantifying insulin secretion and resistance. Am. J. Physiol. 237:E214-E223.

9. Fredrickson, D. S., and R. S. Gordon. 1958. The metabolism of

albumin-bound C'4-labeled unesterifiedfatty acids in normal human subjects. J. Clin.

Invest.37:1504-1515.

10. Issekutz, B., P. Paul, H. I.Miller, and W. M. Bortz. Oxidation of plasma

FFA inlean and obese humans.Metabolism. 17:62-72.

11. Altsuler, N. A., A.Barkai, A. Bjerknes, B. Gottlieb, and R. Steele. 1975.

Glucose turnover values in the dog obtained with variousspecies of labeled

glu-cose. Am.J.Physiol.225:1662-1667.

12. Hales,C. N., and P. J. Randle. Immunoassay of insulin with insulin

antibodyprecipitate. Biochem. J. 88:137-146.

13. Hawk, P. D. 1947.Kjeldahl method. InPracticalPhysiological Chemistry.

12th ed.Blakiston, Toronto, Ont. 814-822.

14.Miles, J. R., J.Glassock, J. Aikens, J.Gerich, and M. Haymond. 1983. A

microfluorometric method of free fatty acids in plasma. J.LipidRes.24:96-99.

15.Radziuk, J., K. H.Norwich, and M.Vranic. 1974. Measurements and

validation of non-steady state turnover rateswith application to theinsulin and

glucose system.Fed. Proc. 33:1855-1864.

16.Cobelli, C., A. Mari, and E.Ferrannini. 1987. Non-steadystate: error

analysisof Steele's model and developmentforglucose kinetics.Am.J.Physiol.

252:E679-E689.

17.Hagenfeldt, L., J. Wahren, B. Pernow, andL.RAfE 1972. Uptake

ofindivid-ualfree fattyacids by skeletal muscle and liver inman.J.Clin. Invest. 51:2324-2330.

18. Waterhouse,C., N. Baker, and H.Rostami.1969.Effect ofglucose

inges-tion on the metabolism of freefattyacids in humansubjects.J.LipidRes.

10:487-494.

19.Robin, A. P., J. Nordenstr6m, J.Askanazi,Y. A.Carpentier,D. H.Elwyn,

and J. M.Kinney.1984.Influence ofparenteral carbohydrateonfat oxidation in

surgical patients. Surgery (St.Louis). 95:608-618.

20. Eaton, R.P., M. Bierman, andD. Steinberg. 1969. Kinetic studies of

plasmafree fatty acid andtriglyceride metabolism in man. J. Clin. Invest.

48:1560-1579.

21.George, J. C., andR. M.Naik. 1958.Relativedistribution and chemical

natureof the fuelstoreof thetwotypes of fibres in thepectoralis majormuscleof

thepigeon. Nature(Lond.). 181:709-711.

22. Barter, P. J., and P. J. Nestel. 1969. Plasmafreefatty acidtransportduring

prolonged glucose consumption and itsrelationshiptoplasmatriglyceridefatty

acids inman.J.LipidRes. 10:83-90.

23.Fritz, I. B. 1961. Factorsinfluencingthe ratesof long chainfattyacid

oxidationandsynthesisinmammaliansystems.Physiol.Rev.41:52-129.

24. Hoerr, R. A., Y.Yong-Ming,D. A. Wagner, J. F.Burke, and V. R.Young.

1989. Recovery of'3Cin beathfromNaH'3C03infusedby gut and vein: effect of

feeding.Am.J.Physiol.257:E426-E438.

25. Armstrong, D. T., R. Steele, N. A. Altzuler, A. Dunn, J. S. Bishop, and

R.C. DeBodo. 1961. Regulation of plasma free fatty acid turnover. Am.J.

Phys-iol. 201:9-15.

26.Miller, H. I., W. M. Bortz, andB.C. Durham. 1968. The rateof

appear-anceofFFAin plasmatriglyceride of normaland obesesubjects. Metabolism.

17:515-521.

27. Nestel, J. 1967. Relationship between FFA flux and TGFA influx in

plasmabefore and duringinfusionof insulin. Metabolism. 16:1123-1132.

28. Randle,P.J,,P. B.Garland,and E. A.Newsholme. 1963. The glucose

fattyacid cycle: its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus. Lancet. i:785-789.

29. Ferrannini, E., E. J. Barrett, S. Bevilaqua, and R. A. DeFronzo. 1983.

Effect offattyacids on glucose production and utilization in man. J. Clin. Invest.