Mechanism of FFA-induced Insulin Resistance
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
Volume 97, Number 12, June 1996, 2859–2865
Mechanism of Free Fatty Acid–induced Insulin Resistance in Humans
Michael Roden, Thomas B. Price, Gianluca Perseghin, Kitt Falk Petersen, Douglas L. Rothman, Gary W. Cline,
and Gerald I. Shulman
Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut 06520
To examine the mechanism by which lipids cause insulin re-
sistance in humans, skeletal muscle glycogen and glucose-6-
phosphate concentrations were measured every 15 min by
C and P nuclear magnetic resonance spec-
troscopy in nine healthy subjects in the presence of low
0.02 mM [mean
SEM]; control) or high (1.93
mM; lipid infusion) plasma free fatty acid levels under eu-
5.2 mM) hyperinsulinemic (
conditions for 6 h. During the initial 3.5 h of the clamp the
rate of whole-body glucose uptake was not affected by lipid
infusion, but it then decreased continuously to be
control values after 6 h (
0.00001). Augmented lipid oxi-
dation was accompanied by a
glucose metabolism starting during the third hour of lipid
0.05). Rates of muscle glycogen synthesis
were similar during the first 3 h of lipid and control infu-
sion, but thereafter decreased to
1.0 vs. 9.3
of muscle glycogen synthesis by elevated plasma free fatty
acids was preceded by a fall of muscle glucose-6-phosphate
concentrations starting at
0.01). Therefore in contrast to the origi-
nally postulated mechanism in which free fatty acids were
thought to inhibit insulin-stimulated glucose uptake in mus-
cle through initial inhibition of pyruvate dehydrogenase
these results demonstrate that free fatty acids induce insulin
resistance in humans by initial inhibition of glucose trans-
port/phosphorylation which is then followed by an
reduction in both the rate of muscle glycogen synthesis and
glucose oxidation. (
J. Clin. Invest.
words: free fatty acids
nuclear magnetic resonance spectroscopy
400 pM) clamp
40% reduction of oxidative
50% of control values
1.5 h (195
25 vs. control:
1996. 97:2859–2865.) Key
Non–insulin-dependent diabetes mellitus (NIDDM)
quently associated with obesity and/or elevation of plasma free
fatty acids (1, 2). Over thirty years ago Randle et al. (3, 4) dem-
onstrated that free fatty acids (FFA) effectively compete with
glucose for substrate oxidation in isolated rat heart muscle and
diaphragms and therefore speculated that increased fat oxida-
tion might cause the insulin resistance associated with diabetes
and obesity. The postulated mechanism has been that in-
creased free fatty acid oxidation causes elevation of the intra-
mitochondrial acetyl-CoA/CoA and NADH/NAD
with subsequent inactivation of pyruvate dehydrogenase (Fig.
1). This in turn causes citrate concentrations to increase which
leads to inhibition of phosphofructokinase and subsequent ac-
cumulation of glucose-6-phosphate. Finally, increased concen-
trations of glucose-6-phosphate would inhibit hexokinase II re-
sulting in decreased glucose uptake (2–4).
Subsequent studies in healthy humans have revealed that
lipid/heparin infusions, which raise plasma FFA levels, inhibit
whole-body glucose disposal during hyper- and euglycemic-
hyperinsulinemia (5–7) and insulin-dependent glucose uptake
by human forearm tissues in vivo (8). These data support the
operation of Randle’s glucose-fatty acid cycle in humans since
the decrease in insulin-dependent glucose uptake can be ac-
counted for by a fat-induced defect in carbohydrate oxidation
(6–8). However, several other studies could not detect an in-
hibitory effect of fatty acids on insulin-mediated glucose up-
take in healthy humans (9), obese humans (10), and patients
with NIDDM (11, 12). The failure to demonstrate this effect
has been attributed to the insufficient duration of the triglycer-
ide/heparin infusion, since fat-induced inhibition of glucose
uptake might develop after more than 3 h of fat infusion (13,
14). However, other studies (8, 11, 15) have found no inhibi-
tory action of fatty acids on nonoxidative glucose metabolism
even when triglyceride/heparin was infused for 4 h.
Recently, Boden et al. (13, 14) have provided evidence that
a reduction in carbohydrate oxidation is responsible for only
one-third, while impairment of nonoxidative glucose metabo-
lism, which mostly reflects glycogen synthesis (16, 17), ac-
counted for two-thirds of the fatty acid–dependent decrease in
glucose uptake. These workers have suggested that two differ-
ent defects contribute to the impairment of glycogen synthesis
depending on the FFA concentration. At FFA concentrations
0.75 mM they found increased intramuscular glucose-6-
M. Roden was on leave from the Department of Internal Medicine
III, University of Vienna, Austria.
Address all correspondence to Gerald I. Shulman, M.D., Ph.D.,
Yale University School of Medicine, Department of Internal Medi-
cine, FMP 104, Box 208020, New Haven, CT 06520-8020. Phone: 203-
737-1115; FAX: 203-785-6015; E-mail: Gerald_Shulman@qm.yale.edu
Received for publication 24 January 1996 and accepted in revised
form 19 March 1996.
nance spectroscopy; NIDDM, non–insulin-dependent diabetes melli-
tus; RF, radio frequency.
Abbreviations used in this paper:
NMR, nuclear magnetic reso-
Roden et al.
phosphate concentrations, as measured in muscle biopsies,
suggesting a FFA-induced inhibition of glycogen synthase,
whereas at FFA concentrations of
able to find any difference in intramuscular glucose-6-phos-
phate. Nevertheless these workers went on to speculate that at
this lower FFA concentration, decreased muscle glucose trans-
port/phosphorylation accounted for the reduced rates of mus-
cle glucose uptake. In contrast, no inhibitory action of fatty ac-
ids on nonoxidative glucose metabolism was found by other
authors (8, 11, 15), nor did fat infusion affect the activities of
glycogen synthase (15, 18) or of pyruvate dehydrogenase (15)
in biopsies from human m. vastus lateralis muscle. The inter-
pretation of the results of enzyme activities measured in vitro
is obscured by several limitations: (
measurements are limited to a few time points; (
tivities do not necessarily reflect in vivo substrate flux; and (
measurements of glucose-6-phosphate concentrations in hu-
man muscle biopsies are artifactually high due to glycogen
breakdown between sample excision and freezing (19). The
0.5 mM they were not
) using muscle biopsies,
) enzyme ac-
experiments performed in this report were designed to over-
come these limitations by performing in vivo nuclear magnetic
resonance (NMR) spectroscopy on the gastrocnemius muscle
of young healthy subjects during long-term infusion of triglyc-
erides/heparin in the presence of euglycemic-hyperinsuline-
C-NMR spectroscopy was used to provide continuous
quantitative information on the rate of muscle glycogen syn-
thesis (17) while simultaneous
used to monitor changes in glucose-6-phosphate concentra-
tions in gastrocnemius muscle in vivo (20).
P-NMR spectroscopy was
range: 22–46 yr; body wt: 70.3
tes mellitus, dyslipidemia, or bleeding disorders were given an isoca-
loric diet [35 kcal/(kg
d); carbohydrate/protein/fat: 60/20/20%] for
2 d and then fasted overnight for 12 h before the studies. Informed
Nine healthy volunteers (eight males, one female; age
2.5 kg; body surface area: 1.85
; body mass index: 23.0
) without family history of diabe-
Figure 1. Schema of potential sites
of free fatty acid action on insulin
mediated glucose metabolism in
skeletal muscle and those sites (*)
hypothesized to be affected by
Randle et al. (3). g6p, glucose-6-
phosphate; G-1-P, glucose-1-phos-
phate; F-6-P, fructose-6-phos-
phate, F-1,6,-P; fructose 1,6 bi-
Mechanism of FFA-induced Insulin Resistance
consent was obtained from all subjects after the nature and possible
consequences of the studies were explained to them. The protocol
was reviewed and approved by the Human Investigation Committee
of the Yale University School of Medicine.
7:30 a.m. with insertion of Teflon
the right and left arm for blood sampling and glucose/lipid/hormone
infusions, respectively. To study insulin-dependent effects on glucose/
glycogen metabolism, euglycemic-hyperinsulinemic glucose clamps
(21) were performed to create conditions of standardized hyperin-
400 pM during constant basal plasma concentrations
of glucose (euglycemia;
5.2 mM). Insulin (Humulin Regular; Eli
Lilly and Co., Indianapolis, IN) was administered as a primed-contin-
uous infusion [1 mU/(kg
min)] and the infusion of [1-
(10–20% enriched) was periodically adjusted to maintain euglycemia
based on the plasma glucose concentrations, which were obtained in
5–8-min intervals. To test the effects of free fatty acids, the plasma
concentration of FFAs was increased by intravenous infusion of a
triglyceride emulsion (1.5 ml/min; Liposyn II, Abbott Laboratories,
North Chicago, IL) combined with heparin [bolus: 200 IU; continu-
ous infusion: 0.2 IU/(kg
min)], which was used to stimulate lipopro-
tein lipase and thereby to catalyze hydrolysis of triglycerides. During
the control experiments glycerol [0.7 mg/(kg
contained in the triglyceride emulsion, was infused to control for ef-
fects of glycerol per se. All subjects participated in both experimental
protocols, which were spaced by an interval of 3–8 wk.
In vivo C-NMR spectroscopy.
monitoring of muscle glycogen and glucose-6-phosphate, the subjects
remained in supine position with the right leg placed inside a 4.7 T
Biospec NMR spectrometer system (Bruker Instruments, Inc., Bil-
lerica, MA) with a 30-cm diameter magnet bore to obtain interleaved
C- (17) and P-spectra (20) from the gastrocnemius muscle before
and every 15-min from time 0 to 360 min during the clamp, respec-
tively. The spectrometer was equipped with a modified radio fre-
quency (RF) relay switch that allowed the hardware to switch the
transmit RF power between C (50.4 MHz) and
nels with a 10-
s switching time. A modified pulse sequence allowed
switching of the acquisition parameters and preamplifiers between
the two channels during the 10-
s switching time. A 5.1-cm diameter
circular C- P double-tuned surface coil RF probe was used for in-
terleaved acquisitions (22, 23). The double-tuned circuit was opti-
mized for the P channel so that the NMR sensitivity would be en-
hanced to detect glucose-6-phosphate. Shimming, imaging, and
decoupling at 200.4 MHz were performed with a 9
terfly coil. Proton water linewidths were shimmed to
crosphere containing C and P reference standards was fixed at the
center of the double-tuned RF coil for calibration of RF pulse widths.
Subjects were positioned by an image-guided localization routine that
employed a T
-weighted gradient-echo image (TR
ms). The subjects’ lower legs were typically positioned so that the iso-
center of the magnetic field was
gastrocnemius muscle. By determining the 180
ter of the observation coil from the microsphere standard, RF pulse
widths were set so that the 90
pulse was at the center of the muscle.
This maximized suppression of the lipid signal that arises from the
subcutaneous fat layer and optimized the signal derived from the
muscle. The interleaved H decoupled
designed so that 72 P transients were acquired during the same pe-
riod that 2,736 C transients were obtained (38
relaxation period). The repetition time for
allow for relaxation of the long T
assessed by magnetic vector potential specific absorption rate calcula-
4 W/kg. The total scan time for each interleaved spectrum
was 5.5 min. Intramuscular glycogen concentrations were determined
by comparison with an external standard solution (150 mM glycogen
50 mM KCl) in a cast of the subject’s leg that electrically loaded the
RF coil the same as subject legs (24).
methods that have been described in detail previously (24). Briefly,
All studies were begun at
catheters in antecubital veins of
min)], which is also
To allow simultaneous, repetitive
P (81.1 MHz) chan-
9 cm series but-
50 Hz. A mi-
82 ms, TE
1 cm into the medial head of the
flip angles at the cen-
P RF pulse sequence was
C transients per
P acquisition was 4.6 s to
P resonances. Power deposition,
C spectra were processed by
Gaussian broadened spectra (30 Hz) were baseline corrected
Hz on either side of the 1-
C glycogen resonance of both subject
spectra and standard spectra. Integrated peak areas were then as-
200 Hz about the resonance. The
sessing intramuscular glycogen concentrations has been validated by
comparison with biopsied human gastrocnemius muscle tissue sam-
ples (25). Intramuscular glucose-6-phosphate was quantified by com-
parison with the
-ATP resonance as an internal reference standard
assuming a constant concentration of 5.5 mM for resting muscle ATP
(20). This method has previously been validated in an animal model
Respiratory exchange measurements.
rimetry was performed for 20 min before and at 90–110, 150–170,
210–230, 270–290, and 330–350 min during the clamp to determine
rates of whole-body glucose oxidation. From these data and from the
amount of nitrogen excreted in the urine, nonoxidative glucose me-
tabolism and rates of lipid and protein oxidation were calculated (17).
Plasma glucose concentrations were mea-
sured by the glucose oxidase method (Glucose analyzer II; Beckman
Instruments Inc., Fullerton, CA). Plasma concentrations of lactate
and of triglycerides (Sigma Chemical Co., St.Louis, MO) were mea-
sured by using enzymatic methods. Plasma concentrations of FFAs
and of glycerol were determined using microfluorimetric methods.
Plasma immunoreactive insulin was determined by a double antibody
RIA (Diagnostic Systems Laboratories, Inc., Webster, TX). Plasma
[1- C] glucose atom percent excess was measured using gas chroma-
tography-mass spectrometry (17, 24). Plasma alanine concentrations
were determined by an automated amino acid analyzer (Dionex,
Calculations and data analysis.
concentration were determined from the change in [1-
concentration and the plasma [1-
described previously (17, 24). Rates of glycogen synthesis were then
calculated from the slope of the least-square linear fit to the glycogen
concentration curve during the given time periods. All data are pre-
sented as means
SEM. Statistical comparisons between control and
lipid infusion experiments were performed by using the paired Stu-
C-NMR technique for as-
Continuous indirect calo-
Increments in muscle glycogen
C] glucose atom percent excess as
Basal plasma concentrations of glucose, insulin, and free fatty
acids were not different between control and lipid infusion
studies (Fig. 2). During the euglycemic-hyperinsulinemic
clamps plasma glucose concentrations remained constant com-
pared to the basal period and were not different between con-
trol and lipid infusion studies (Fig. 2
centrations increased similarly to
). In the control study the plasma concentration of
FFAs dropped by
70% once the euglycemic clamp was
started, but it increased approximately fourfold within 90 min
of the lipid infusion (Fig. 2
). Basal plasma concentrations of
triglycerides (lipid infusion: 1.15
ter) and glycerol (0.27
0.01 vs. 0.28
ferent between the studies however during the lipid infusion
plasma concentrations of triglycerides (7.02
0.0001) and of glycerol (1.29
0.0001) increased. Basal plasma concen-
trations of lactate were 0.70
0.07 vs. 0.57?0.04 mM (lipid in-
fusion vs. control) and alanine were 0.36?0.03 vs. 0.30?0.03
mM (lipid infusion vs. control). During the last hour of lipid in-
fusion (300–360 min) plasma concentrations of lactate (lipid
infusion; 0.71?0.23 vs. control; 1.28?0.13 mM, P ? 0.02) and
alanine (lipid infusion; 0.22?0.01 vs. control; 0.31?0.02 mM,
). Plasma insulin con-
400 pM in both studies
0.14, control: 1.13
0.02 mM) were not dif-
Roden et al.
P ? 0.001) were both significantly lower than during the con-
During the first 3.5 h of the clamp test, whole-body glucose
metabolism as reflected by the glucose infusion rates was stim-
ulated in a similar fashion by insulin at low and at high plasma
FFA concentrations (Fig. 3 A). After that time glucose infu-
sion rates decreased progressively during lipid infusion (210–
230 min: lipid infusion; 40.3?3.9 vs. control; 52.1?3.1 ?mol/[kg
? min], P ? 0.01) declining to ? 46% of the corresponding con-
trol values during the last hour of the clamp (330–350 min:
lipid infusion; 25.7?2.6 vs. control; 57.2?3.5 ?mol/[kg ? min],
P ? 0.00001).
During lipid infusion the mean respiratory quotient (RQ)
was lower starting with the first recording between 90 and 110
min (lipid infusion; 0.84?0.02 vs. control; 0.90?0.01, P ? 0.05),
and it continued to decrease until the last hour (lipid infusion;
0.79?0.03 vs. control; 0.96?0.01, P ? 0.00001) reflecting in-
creased lipid oxidation (90–110 min: lipid infusion; 0.53?0.13
vs. control; 0.11?0.03 mg/[kg ? min], P ? 0.05, 330–350 min:
lipid infusion; 0.82?0.08 vs. control; 0?0.06 mg/[kg ? min], P ?
0.05). In contrast to the progressive increase in glucose oxida-
tion rate observed in the control study augmented lipid oxida-
tion in the lipid infusion study blunted the increase in glucose
oxidation rate resulting in a significantly decreased rate of glu-
cose oxidation compared to the control study [t ? 150 (P ?
0.02), t ? 210 (P ? 0.005), t ? 270 (P ? 0.0002), t ? 330 (P ?
0.00005)] (Fig. 3 B). It is of note that decreased glucose oxida-
tion failed to affect whole-body glucose metabolism before 210
min (glucose oxidation, 150–170 min: lipid infusion: 7.1?1.7 vs.
control: 12.1?1.3 ?mol/[kg ? min], P ? 0.05). The rate of pro-
tein oxidation also gradually decreased during the lipid infu-
sion (330–350 min, lipid infusion; 0.89?0.07 vs. control;
1.08?0.07 mg/[kg ? min], P ? 0.01).
Basal concentrations of skeletal muscle glycogen were
Figure 2. Plasma concentrations of glucose (A), in-
sulin (B), and free fatty acids (C) at low (closed
symbols) and at elevated plasma free fatty acid con-
centrations (open symbols). All units expressed as
means?SEM of nine paired studies.
Mechanism of FFA-induced Insulin Resistance
85?5 and 72?6 mM in the control and lipid infusion studies,
respectively. The insulin-dependent increase in skeletal muscle
glycogen concentration was not different during the first 210
min, but thereafter declined during the lipid infusion to ? 44%
(P ? 0.05) of the corresponding control values at 360 min (Fig.
3 C). Net rates of muscle glycogen synthesis (Vsyn) were:
94?11 (0–90 min), 134?10 (90–180 min), 122?13 (180–270
min), and 93?16 ?mol/[liter.min] (270–360 min) under condi-
tions of low plasma concentrations of free fatty acids. Elevated
FFA concentrations did not affect net glycogen synthetic rates
during the first 3 h (0–90 min: 85?10, 90–180: 120?15 ?mol/[li-
ter ? min]), whereas thereafter Vsyn decreased by ? 50% com-
pared with the control studies (180–270 min: 65?20, P ? 0.05,
270–360 min: 40?10 ?mol/[liter ? min], P ? 0.05). Similarly,
nonoxidative glucose metabolism was not different between
lipid infusion and control experiments during the first 4 h: lipid
infusion; 30.9?2.8 vs. control; 38.1?5.2, NS (90–110 min), lipid
infusion; 39.7?3.5 vs. control; 38.4?3.3, NS (150–170 min),
lipid infusion; 34.8?3.6 vs. control; 39.1?3.3 ?mol/[kg ? min],
NS (210–230 min). Thereafter, nonoxidative glucose metabo-
lism decreased to ? 50% during the lipid infusion: lipid infu-
sion; 22.7?2.8 vs. control; 39.0?4.1, P ? 0.005 (270–290 min)
and lipid infusion; 21.0?2.2 vs. control; 41.2?3.9 ?mol/
[kg ? min], P ? 0.0005 (330–350 min).
During lipid infusion absolute concentrations of muscle
glucose-6-phosphate were always lower than those of the con-
trol experiments (Fig. 3 D): lipid infusion: 195?25 vs. control:
237?26 ?M, P ? 0.01 (90–110 min), lipid infusion: 152?16 vs.
control: 231?31 ?M, P ? 0.01 (150–170 min), lipid infusion:
129?22 vs. control: 217?20 ?M, P ? 0.01 (210–230 min), lipid
Figure 3. Glucose infusion rate (A), glucose oxida-
tion rate (B), increase in calf muscle glycogen (C),
and increase in calf muscle glucose-6-phosphate (D)
at low (closed symbols) and at elevated plasma free
fatty acid concentrations (open symbols). All units
expressed as means?SEM of nine paired studies.
*P ? 0.05, ?P ? 0.01, †P ? 0.001.
Roden et al.
infusion: 113?14 vs. control: 197?15 ?M, P ? 0.00005 (270–
290 min), and lipid infusion: 107?16 vs. control: 188?12 ?M,
P ? 0.0005 (330–350 min).
The time course of nonoxidative glucose metabolism paral-
leled that of muscle glycogen synthesis, consistent with muscle
glycogen synthesis being the predominant route of insulin-
stimulated nonoxidative glucose metabolism (16, 17). The ob-
served reduction in nonoxidative glucose metabolism/muscle
glycogen synthesis could be due to an FFA-induced reduction
in muscle glycogen synthase activity, which would be expected
to result in an increase in intracellular glucose-6-phosphate
concentration. In support of this possibility some studies have
found decreased insulin-stimulated fractional velocities of gly-
cogen synthase in skeletal muscle biopsies obtained after 4–6 h
of lipid infusion (7, 13, 14), while others have reported no
change in glycogen synthase activity (15, 18). Alternatively, in-
hibition of muscle glycogen synthesis could be accounted for
by impaired glucose transport/phosphorylation, which would
be mirrored by a decrease of skeletal muscle glucose-6-phos-
phate concentration. In support of this latter possibility we
found that the insulin-dependent increase in intramuscular
glucose-6-phosphate concentration was lower (P ? 0.01) at 90
min and was followed by a continuous decrease to and even
below baseline values in the presence of elevated plasma free
fatty acid concentration (Fig. 3 D). The finding that intramus-
cular glucose-6-phosphate concentrations were lower during
the lipid infusion study than during the control study does not
support the original mechanism as proposed by Randle and
co-workers in which glucose uptake is decreased by increased
FFA levels due to inhibition of hexokinase II resulting from an
increased intracellular concentration of glucose-6-phosphate
(2–4, 27, 28). Instead, these results suggest an inhibitory effect
of fatty acids on glucose transport/phosphorylation. As a result
of its location between glucose transport/phosphorylation and
synthase enzymes in the pathways of glycogen synthesis the
concentration of muscle glucose-6-phosphate is sensitive to the
relative activities of these enzymes and the rate of glycolysis.
While it is theoretically possible that an increased rate of glycol-
ysis in the lipid infusion studies could explain the lower glu-
cose-6-phosphate concentrations observed in these studies,
this is very unlikely since any increase in glycolytic lactate pro-
duction would be included in the measured rate of nonoxida-
tive glucose metabolism which was ? 50% lower during the
last 90 min of the lipid infusion studies. Furthermore, the lower
plasma concentrations of alanine and lactate observed in the
lipid infusion studies suggest that rates of glycolysis were lower
in these studies compared to the control studies. Finally a
higher rate of oxidative glycolysis would be reflected by an in-
crease in rates of glucose oxidation which were also signifi-
cantly lower in the lipid infusion studies compared to the con-
Whether FFA induces its inhibitory effect on glucose up-
take through inhibition of glucose transport and/or hexokinase
cannot be discerned from the present data. GLUT4 expression
has been shown to be suppressed in skeletal muscle from high
fat-fed rats whose FFA levels are not suppressed during insu-
lin infusion but to be unchanged in genetically obese Zucker
rats which present with increased plasma FFA levels (29). De-
creased glucose transport could also be accounted for by re-
duced intrinsic activity of GLUT4 despite normal insulin-stim-
ulated GLUT4 translocation as reported for high fat–fed rats
(30). The FFA-induced decrease in skeletal muscle glucose-
6-phosphate concentration could also be due to impaired glu-
cose phosphorylation by hexokinase II, which is insulin-depen-
dent and could become a rate controlling step (31). However,
no changes in insulin-stimulated hexokinase activity in biop-
sies from human vastus lateralis muscle were found after 6 h of
lipid infusion combined with a 4-h hyperinsulinemic-euglyce-
mic clamp (18).
In conclusion, contrary to the classical mechanism of free
fatty acid–induced insulin resistance as proposed by Randle et
al. (2, 27, 28) in which free fatty acids exert their effect through
initial inhibition of pyruvate dehydrogenase, we found that el-
evation in plasma free fatty acid concentration causes insulin
resistance by inhibition of glucose transport and/or phosphory-
lation with a subsequent reduction in rates of glucose oxida-
tion and muscle glycogen synthesis. This reduction in insulin-
inducible glucose transport/phosphorylation is similar to what
is observed in patients with NIDDM (20) and their normogly-
cemic-insulin–resistant offspring (24) and suggests that alter-
ations in intramuscular FFA metabolism may play an impor-
tant role in the pathogenesis of the insulin resistance observed
in patients with NIDDM.
We would like to thank V. Walton, Y. Milewski, N. Barucci, T.
Nixon, and P. Brown for technical assistance and the staff of the
Yale-New Haven Hospital General Clinical Research Center and
This work was supported by grants from the Public Health Ser-
vice: R01 DK-49230, P30 DK-45735, M01 RR-00125, R29 NS-32126
and a Clinical Research Grant from the American Diabetes Associa-
tion. Dr. Roden was supported by a Max-Kade Foundation Fellow-
ship Award. Dr. Perseghin was supported by a postdoctoral fellow-
ship from the Juvenile Diabetes Foundation, Int., and by a Research
Training Award from the Instituto Scientifico San Raffaele, Milan,
1. Reaven, G.M., C. Hollenbeck, C-Y. Jeng, M.S. Wu, and Y-D. Chen.
1988. Measurement of plasma glucose free fatty acid, lactate, and insulin for
24 h in patients with NIDDM. Diabetes. 37:1020–1024.
2. McGarry J.D. 1992. What if Minkowski had been ageusic? An alternative
angle on diabetes. Science (Wash. DC). 258:766–770.
3. Randle, P.J., P.B. Garland, C.N. Hales, and E.A. Newsholme. 1963. The
glucose fatty-acid cycle: its role in insulin sensitivity and the metabolic distur-
bances of diabetes mellitus. Lancet. i: 785–789.
4. Randle, P.J., P.B. Garland, E.A. Newsholme, and C.N. Hales. 1965. The
glucose fatty acid cycle in obesity and maturity onset diabetes mellitus. Ann.
NY Acad. Sci. 131:324–333.
5. Thiebaud, D., R.A. DeFronzo, E. Jacot, A. Golay, K. Acheson, E.
Maeder, E. Jequier, and J.P. Felber. 1982. Effect of long chain triglyceride infu-
sion on glucose metabolism in man. Metab. Clin. Exp. 31:1128–1136.
6. Ferrannini, E., E.J. Barrett, S. Bevilacqua, and R.A. DeFronzo. 1983. Ef-
fect of fatty acids on glucose production and utilization in man. J. Clin. Invest.
7. Kelley, D.E., M. Mokan, J-A. Simoneau, and L.J. Mandarino. 1993. In-
teraction between glucose and free fatty acid metabolism in human skeletal
muscle. J. Clin. Invest. 92:91–98.
8. Yki-Jarvinen, H., I. Puhakainen, and V.A. Koivisto. 1992. Effect of free
fatty acids on glucose uptake and nonoxidative glycolysis across human forearm
tissues in the basal state and during insulin stimulation. J. Clin. Endocrinol. &
9. Wolfe, B.M., S. Klein, E.J. Peters, B.F. Schmidt, and R.R. Wolfe. 1988.
Effect of elevated free fatty acids on glucose oxidation in normal humans.
Metab. Clin. Exp. 37:323–329.
10. Bevilacqua, S., R. Bonnadonna, G. Buzzigoli, C. Boni, D. Ciociarn, F.
Mechanism of FFA-induced Insulin Resistance
Maccari, M.A. Giorico, and E. Ferrannini. 1987. Acute elevation of free fatty
acid levels lead to hepatic insulin resistance in obese subjects. Metab. Clin. Exp.
11. Bevilacqua, S., G. Buzzigoli, R. Bonnadonna, L.S. Brandi, M. Ollegini,
C. Boni, M. Geloni, and E. Ferrannini. 1990. Operation of Randle’s cycle in pa-
tients with NIDDM. Diabetes. 39:383–389.
12. Yki-Jarvinen, H., I. Puhakainen, C. Saloranta, L. Groop, and M-R.
Taskinen. 1991. Demonstration of a novel feedback mechanism between FFA
oxidation from intracellular and intravascular sources. Am. J. Physiol. 260:
13. Boden, G., F. Jadali, J. White, Y. Liang, M. Mozzoli, X. Chen, E. Cole-
man, and C. Smith. 1991. Effects of fat on insulin-stimulated carbohydrate me-
tabolism in normal men. J. Clin. Invest. 88:960–966.
14. Boden, G., X. Chen, J. Ruiz, J.V. White, and L. Rossetti. 1994. Mecha-
nisms of fatty acid-induced inhibition of glucose uptake. J. Clin. Invest. 93:2438–
15. Saloranta, C., V. Koivisto, E. Widen, K. Falholt, R.A. DeFronzo, M.
Harkonen, and L. Groop. 1993. Contribution of muscle and liver to glucose-
fatty acid cycle in humans. Am. J. Physiol. 264:E599–E605.
16. Bogardus, C., S. Lillioja, K. Stone, and D. Mott. 1984. Correlation be-
tween muscle glycogen synthase activity and in vivo insulin action in man. J.
Clin. Invest. 73:1185–1190.
17. Shulman, G.I., D.L. Rothman, T. Jue, P. Stein, R.A. DeFronzo, and
R.G. Shulman. 1990. Quantitation of muscle glycogen synthesis in normal sub-
jects and subjects with non-insulin-dependent diabetes by 13C nuclear magnetic
resonance spectroscopy. N. Engl. J. Med. 322:223–228.
18. Johnson, A.B., M. Argyraki, J.C. Thow, B.G. Cooper, G. Fulcher, and
R. Taylor. 1992. Effect of increased free fatty acid supply on glucose metabo-
lism and skeletal muscle glycogen synthase activity in normal man. Clin. Sci. 82:
19. Rossetti, L., and A. Giaccari. 1992. Relative contribution of glycogen
synthesis and glycolysis to insulin-mediated glucose uptake: a dose response
study. J. Clin. Invest. 85:1785–1792.
20. Rothman, D.L., R.G. Shulman, and G.I. Shulman. 1992. 31P nuclear
magnetic resonance measurements of muscle glucose-6-phosphate. Evidence
for reduced insulin-dependent muscle glucose transport or phosphorylation ac-
tivity in non–insulin-dependent diabetes mellitus. J. Clin. Invest. 89:1069–1075.
21. DeFronzo, R.A., J. Tobin, and R. Andres. 1979. The glucose clamp
technique: a method for the quantification of b-cell sensitivity to glucose and of
tissue sensitivity to insulin. Am. J. Physiol. 237:E214–E223.
22. Price, T.B., D.L. Rothman, R. Taylor, M.J. Avison, G.I. Shulman, and
R.G. Shulman. 1994. Human muscle glycogen resynthesis after exercise: insu-
lin-dependent and -independent phases. J. Appl. Physiol. 76:104–111.
23. Chen, W., T.B. Price, D.L. Rothman, J.R. Chase, G. Bloch, G.I. Shul-
man, and R.G. Shulman. 1994. Determination of G6P and glycogen changes of
human muscle following exercise by interleaved 31P and 13C NMR spectros-
copy. Proc. Soc. Mag. Res. 1:160.
24. Rothman, D.L., I. Magnusson, G. Cline, D. Gerard, C. R. Kahn, R.G.
Shulman, and G.I. Shulman. 1995. Decreased muscle glucose transport/phos-
phorylation is an early defect in the pathogenesis of non–insulin-dependent dia-
betes mellitus. Proc. Natl. Acad. Sci. USA. 92:983–987.
25. Taylor, R., T.B. Price, D.L. Rothman, R.G. Shulman, and G.I. Shulman.
1992. Validation of 13C NMR measurement of human skeletal muscle glycogen
by direct biochemical assay of needle biopsy samples. Magn. Reson. Med. 27:
26. Bloch, G., J.R. Chase, M.J. Avison, and R.G. Shulman. 1993. In vivo 31P
NMR measurement of glucose 6 phosphate in the rat muscle after exercise.
Magn. Reson. Med. 30:347–350.
27. Randle, P.J., E.A. Newsholme, and P.B. Garland. 1964. Regulation of
glucose uptake by muscle. 8. Effects of fatty acids, ketone bodies and pyruvate,
and of alloxan, diabetes and starvation, on the uptake and metabolic fate of glu-
cose in rat heart and diaphragm muscles. Biochem. J. 93:652–665.
28. Randle, P.J., D.A. Priestman, S. Mistry, and A. Halsall. 1994. Mecha-
nisms modifying glucose oxidation in diabetes mellitus. Diabetologia. 37(Suppl.
29. Kahn, B.B., and O. Pedersen. 1993. Suppression of GLUT4 expression
in skeletal muscle of rats that are obese from high fat feeding but not from high
carbohydrate feeding or genetic obesity. Endocrinology. 132:13–22.
30. Rosholt, M.N., P.A. King, and E.S. Horton. 1994. High fat diet reduces
glucose transporter responses to both insulin and exercise. Am. J. Physiol. 266:
31. Katz, A., I. Raz, M.K. Spencer, R. Rising, and D.M. Mott. 1991. Hyper-
glycemia induced accumulation of glucose in human skeletal muscle. Am. J.