The role of skeletal muscle insulin resistance in the
pathogenesis of the metabolic syndrome
Kitt Falk Petersen*, Sylvie Dufour†, David B. Savage*, Stefan Bilz*, Gina Solomon*, Shin Yonemitsu*, Gary W. Cline*,
Douglas Befroy*, Laura Zemany‡, Barbara B. Kahn‡, Xenophon Papademetris§, Douglas L. Rothman§,
and Gerald I. Shulman*†§¶?
Departments of *Internal Medicine;§Diagnostic Radiology and Biomedical Engineering;¶Cellular and Molecular Physiology, and†Howard Hughes Medical
Institute, Yale University School of Medicine, New Haven, CT 06536; and‡Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215
This contribution is part of the special series of Inaugural Articles by members of the National Academy of Sciences elected on May 1, 2007.
Contributed by Gerald I. Shulman, June 8, 2007 (sent for review May 25, 2007)
We examined the hypothesis that insulin resistance in skeletal
muscle promotes the development of atherogenic dyslipidemia,
associated with the metabolic syndrome, by altering the distribu-
two high carbohydrate mixed meals, net muscle glycogen synthe-
sis was reduced by ?60% in young, lean, insulin-resistant subjects
compared with a similar cohort of age–weight–body mass index–
activity-matched, insulin-sensitive, control subjects. In contrast,
hepatic de novo lipogenesis and hepatic triglyceride synthesis
were both increased by >2-fold in the insulin-resistant subjects.
These changes were associated with a 60% increase in plasma
triglyceride concentrations and an ?20% reduction in plasma
high-density lipoprotein concentrations but no differences in
plasma concentrations of TNF-?, IL-6, adiponectin, resistin, retinol
binding protein-4, or intraabdominal fat volume. These data dem-
muscle glycogen synthesis, can promote atherogenic dyslipidemia
by changing the pattern of ingested carbohydrate away from
skeletal muscle glycogen synthesis into hepatic de novo lipogen-
esis, resulting in an increase in plasma triglyceride concentrations
and a reduction in plasma high-density lipoprotein concentrations.
Furthermore, insulin resistance in these subjects was independent
of changes in the plasma concentrations of TNF-?, IL-6, high-
molecular-weight adiponectin, resistin, retinol binding protein-4,
or intraabdominal obesity, suggesting that these factors do not
play a primary role in causing insulin resistance in the early stages
of the metabolic syndrome.
type 2 diabetes ? nonalcoholic fatty liver disease ? adipocytokines ?
abdominal obesity ? atherogenic dyslipidemia
tance, abdominal obesity, atherogenic dyslipidemia, hyperten-
sion, hyperuricemia, a prothrombotic state, and a proinflamma-
tory state (1, 2). The metabolic syndrome is estimated to afflict
?50 million Americans, and approximately half of all Americans
are predisposed to it (2). Individuals with the metabolic syn-
drome are at increased risk for the development of coronary
heart disease and other diseases related to plaque buildup in
artery walls, such as stroke and peripheral vascular disease, as
well as type 2 diabetes mellitus (T2DM).
Abdominal obesity and insulin resistance have each been
hypothesized to be the primary factors underlying the metabolic
syndrome; however, the biologic mechanisms linking these and
other metabolic risk factors associated with the metabolic syn-
drome are not fully understood and appear to be complex.
In this study we examined the hypothesis that insulin resis-
tance in skeletal muscle may promote the development of
atherogenic dyslipidemia by diverting ingested carbohydrate
away from muscle glycogen storage and into hepatic de novo
lipogenesis, resulting in hypertriglyceridemia. To examine this
factors for cardiovascular disease that include insulin resis-
hypothesis we assessed liver and muscle triglyceride synthesis by
1H magnetic resonance spectroscopy (MRS) and liver and
muscle glycogen synthesis by13C MRS in young, lean, healthy,
insulin-resistant subjects and compared them to a group of
age–weight–body mass index–activity-matched, insulin-sensitive
addition, hepatic de novo lipogenesis was assessed at the same
time by monitoring the incorporation of deuterium from
intraabdominal obesity has been postulated to be at the core of
the metabolic derangements and directly responsible for the
atherogenic dyslipidemia associated with the metabolic syn-
drome, we also measured intraabdominal fat content in these
two groups of subjects by MRI (5–8).
The advantage of examining this question in healthy, young,
lean, insulin-resistant individuals is that they have none of the
other confounding factors that are typically associated with
obesity and T2DM, which have been postulated to play a major
role in causing the metabolic syndrome. Furthermore previous
studies have demonstrated that insulin resistance in these sub-
jects can mostly be attributed to defects in insulin-stimulated
muscle glycogen synthesis caused by defects in insulin-stimulated
glucose transport/phosphorylation activity (9–11). Therefore,
the role of skeletal muscle insulin resistance per se in the
pathogenesis of atherogenic dyslipidemia and the metabolic
syndrome can be examined at its earliest stages.
We screened ?400 young, healthy, lean, sedentary subjects with
an oral glucose tolerance test and assessed insulin sensitivity in
them by using the insulin sensitivity index (ISI) (12). From this
screening we identified 12 insulin-resistant [lowest ISI quartile
(13)] subjects and 12 insulin-sensitive [highest ISI quartile (13)]
subjects who were willing to undergo additional inpatient MRS/
MRI/stable isotope studies to assess liver and muscle glycogen/
triglyceride synthesis and hepatic de novo lipogenesis.
Author contributions: K.F.P., S.D., D.B.S., S.B., G.W.C., D.B., and G.I.S. designed research;
K.F.P., S.D., D.B.S., S.B., G.S., S.Y., G.W.C., D.B., L.Z., B.B.K., X.P., D.L.R., and G.I.S. performed
research; K.F.P., S.D., D.B.S., S.B., S.Y., G.W.C., D.B., L.Z., B.B.K., X.P., D.L.R., and G.I.S.
analyzed data; and K.F.P., S.D., D.B.S., S.B., G.S., S.Y., G.W.C., D.B., L.Z., B.B.K., X.P., D.L.R.,
and G.I.S. wrote the paper.
is an inventor on a patent for RBP-4.
Freely available online through the PNAS open access option.
Abbreviations: HMW, high molecular weight; HDL, high-density lipoprotein; ISI, insulin
sensitivity index; LDL, low-density lipoprotein; MRS, magnetic resonance spectroscopy;
NAFLD, nonalcoholic fatty liver disease; RBP-4, retinol binding protein-4; T2DM, type 2
diabetes mellitus; VLDL, very-low-density lipoprotein.
?To whom correspondence should be addressed at: Howard Hughes Medical Institute, Yale
University School of Medicine, P.O. Box 9812, New Haven, CT 06536. E-mail:
© 2007 by The National Academy of Sciences of the USA
July 31, 2007 ?
vol. 104 ?
no. 31 ?
Despite having large differences in the ISI, the insulin-
resistant and insulin-sensitive volunteers were similar in age,
body mass index, lean body mass, fat mass, and monitored
physical activity (Table 1). There also were no differences in
systolic or diastolic blood pressure between the groups.
Fasting plasma glucose, triglyceride, and insulin concentra-
tions were higher in the insulin-resistant subjects compared with
the insulin-sensitive subjects (Table 2). After the mixed meals,
postprandial plasma glucose concentrations were similar in the
two groups (Fig. 1A). In contrast, postprandial plasma concen-
trations of insulin and triglycerides were markedly increased in
the insulin-resistant subjects (Fig. 1 B and C). There were no
differences in fasting or postprandial plasma fatty acid concen-
trations between the two groups, except from 10 to 11 p.m., when
plasma fatty acid concentrations were lower in the insulin-
resistant subjects compared with the insulin-sensitive subjects
Fasting plasma high-density lipoprotein (HDL) cholesterol
concentrations were 20% lower in the insulin-resistant subjects
(Table 2), but there were no differences in total cholesterol or
low-density lipoprotein (LDL) between the two groups. The
concentrations of small very-low-density lipoprotein (VLDL)
and large LDL particles were decreased by 38% and 30%,
respectively, in the insulin resistant subjects (Table 2). Plasma
concentrations of uric acid were increased by ?30% in the
insulin-resistant subjects compared with the insulin-sensitive
control subjects (Table 2). In contrast, there were no differences
in plasma concentrations of high-molecular-weight (HMW)
adiponectin, IL-6, resistin, plasminogen activator inhibitor-1,
retinol binding protein-4 (RBP-4), or TNF-? between the two
groups (Table 3).
Basal concentrations of muscle glycogen content were similar
in the insulin-sensitive (87.1 ? 5.7 mmol/liter muscle) and
insulin-resistant (84.1 ? 4.6 mmol per liter of muscle, P ? 0.69)
subjects. After the mixed meals, however, net muscle glycogen
synthesis was 61% lower in the insulin-resistant subjects com-
pared with the insulin-sensitive subjects (Fig. 2A). In contrast,
there were no differences between the two groups in basal liver
glycogen concentration (insulin-sensitive, 131.6 ? 9.2 mmol per
liter of liver, vs. insulin-resistant, 125.0 ? 17.7 mmol per liter of
liver; P ? 0.74) or net liver glycogen synthesis after ingestion of
the meals (Fig. 2B).
There were no differences in the basal concentrations of
hepatic triglyceride in the two groups (insulin-sensitive, 0.65 ?
0.14%, vs. insulin-resistant, 0.76 ? 0.20%; P ? 0.63); however,
net hepatic triglyceride synthesis was ?2.5-fold greater in the
insulin-resistant subjects than the insulin-sensitive subjects after
the carbohydrate meals (Fig. 2D). There were no changes in
intramyocellular triglyceride content after the meals in either
group (Fig. 2C). The mean (Table 1) and distribution (Fig. 3C)
of intraabdominal fat volume assessed by MRI were similar
between the two groups, whether or not the single outlier in each
group was excluded from analysis.
Postprandial fractional hepatic de novo triglyceride lipogen-
esis, as assessed by the incorporation of deuterated water into
plasma triglyceride, was increased by 2.2-fold in the insulin-
resistant subjects (15.7 ? 1.5%) compared with the insulin-
sensitive subjects (7.2 ? 0.7%, P ? 0.00005) (Fig. 4).
We found that the pattern of stored energy distribution derived
lean, insulin-resistant individuals compared with young, lean, in-
sulin-sensitive individuals. In contrast to the young, lean, insulin-
sensitive subjects, who stored most of their ingested energy in liver
and muscle glycogen, the young lean insulin-resistant subjects had
a marked defect in muscle glycogen synthesis and diverted much
more of their ingested energy into hepatic de novo lipogenesis,
resulting in increased plasma triglycerides, lower HDL, and in-
creased hepatic triglyceride synthesis (Fig. 5).
These data have several important implications for the role of
skeletal muscle insulin resistance in the pathogenesis of the
metabolic syndrome and T2DM. First, the data are consistent
with the hypothesis that insulin resistance in skeletal muscle
promotes atherogenic dyslipidemia by promoting the conversion
of energy derived from ingested carbohydrate into hepatic de
novo lipogenesis and increased VLDL production. This hypoth-
esis is further supported by studies in mice with muscle-specific
inactivation of the insulin receptor gene, which have been shown
to have increased plasma triglycerides and increased adiposity as
a result of muscle-specific insulin resistance (14).
Table 1. Subject characteristics
7.87 ? 0.34 28 ? 267.1 ? 3.61.72 ? 0.0322.6 ? 0.621.8 ? 2.352.8 ? 3.6340 ? 91113 ? 266 ? 2 4.83 ? 0.45
2.80 ? 0.20 23 ? 1 69.6 ? 2.91.71 ? 0.0323.9 ? 0.6 26.6 ? 2.351.5 ? 3.3 390 ? 92110 ? 466 ? 2 3.89 ? 0.54
?0.0001 NSNSNS NS NSNS NSNS NS NS
diastolic blood pressure; NS, not significant.
Table 2. Fasting plasma metabolite and hormone concentrations
84.1 ? 1.7 7.6 ? 0.653 ? 7182 ? 1293 ? 978 ? 538.6 ? 4.0446.4 ? 34.23.9 ? 0.30.47 ? 0.16
90.6 ? 1.512.1 ? 1.286 ? 13157 ? 577 ? 662 ? 323.8 ? 2.9311.9 ? 46.05.2 ? 0.51.02 ? 0.24
SP, small particles; LP, large particles; CRP, C-reactive protein; NS, not significant.
www.pnas.org?cgi?doi?10.1073?pnas.0705408104Petersen et al.
Second, increased hepatic de novo lipogenesis resulting in
hypertriglyceridemia also can explain the lower HDL levels in
the insulin-resistant subjects by the following mechanism: In the
presence of increased plasma VLDL concentrations and normal
activity of cholesteryl ester transfer protein, VLDL triglycerides
can be exchanged for HDL cholesterol, where a VLDL particle
donates a molecule of triglyceride to an HDL particle in return
for one of the cholesteryl ester molecules from HDL. This
process leads to a cholesterol-rich VLDL remnant particle that
is atherogenic and a triglyceride-rich, cholesterol-depleted HDL
particle (15). The triglyceride-rich HDL particle can undergo
further modification, including hydrolysis of its triglyceride,
which leads to the dissociation of the apo A-I protein. The free
with HDL particles, leading to reductions in the amount of
circulating apo A-I, HDL cholesterol, and the number of HDL
These data also demonstrate that skeletal muscle insulin
resistance predates hepatic insulin resistance and that hepatic
triglyceride synthesis is increased in these insulin-resistant sub-
jects after high-carbohydrate meals, which may predispose them
to nonalcoholic fatty liver disease (NAFLD). The absence of
hepatic steatosis in this group of insulin-resistant subjects is
consistent with recent studies by our group demonstrating the
relatively low prevalence of hepatic steatosis in young, lean,
healthy subjects of different ethnic backgrounds with the im-
portant exception of Asian Indian males who have a marked
increase in the prevalence of hepatic steatosis (13). Although it
the majority of newly synthesized triglyceride, these results
suggest that increased hepatic de novo lipogenesis precede the
development of adipose tissue insulin resistance, which subse-
quently leads to increased flux of fatty acids to the liver (16, 17).
This hypothesis is supported by our previous findings of similar
and after the mixed meals.
Plasma concentrations of glucose (A), insulin (B), triglyceride (C), and fatty acids (D) in insulin-sensitive (E) and insulin-resistant (■) participants before
Table 3. Fasting plasma adipocytokine concentrations
5.8 ? 2.5
3.5 ? 0.7
0.91 ? 0.20
1.31 ? 0.23
11.0 ? 1.1
12.5 ? 0.9
1.49 ? 0.12
1.49 ? 0.17
13.8 ? 3.9
13.2 ? 2.5
29.5 ? 3.3
28.1 ? 3.1
134 ? 8
149 ? 11
PAI-1, plasminogen activator inhibitor-1; TTR, transthyretin; NS, not significant.
Petersen et al.PNAS ?
July 31, 2007 ?
vol. 104 ?
no. 31 ?
basal and insulin suppressed rates of whole body and subcuta-
neous fat lipolysis in similar groups of insulin-sensitive and
insulin-resistant subjects (12). NAFLD is strongly linked to
hepatic insulin resistance, and recent studies have demonstrated
that NAFLD-induced hepatic insulin resistance is a major factor
responsible for the transition from normoglycemia to fasting
hyperglycemia and T2DM (18–23). The cellular signal driving
this increased hepatic de novo lipogenesis can most likely be
attributed to hyperinsulinemia promoting increased expression
of the sterol regulatory element binding protein-1c in the liver,
which coordinately regulates transcription of all of the key
enzymes involved in lipogenesis (24). Indeed, assuming a portal
vein–artery concentration gradient for insulin of ?3, portal vein
insulin concentrations can be estimated to exceed 500
microunits/ml in the insulin-resistant subjects
Increased export of triglyceride from the liver to the periph-
eral and visceral adipose tissue in the form of VLDL may also
predispose these insulin-resistant individuals to abdominal obe-
sity (25). Although abdominal obesity has been postulated to
play the major role in causing the atherogenic dyslipidemia and
insulin resistance associated with the metabolic syndrome (5–7),
we did not observe any significant differences in the mean
volume of intraabdominal fat between these two groups. These
data suggest that abdominal obesity develops later in the course
of the metabolic syndrome along with NAFLD and that it is
rather than a primary cause of insulin resistance and atherogenic
dyslipidemia (5–7). These results are consistent with recent
observations in lipodystrophic patients and mice (20) as well as
rodent models of hepatic insulin resistance (18) that have
disassociated intraabdominal adiposity from insulin resistance
and instead have implicated hepatic steatosis in causing hepatic
insulin resistance associated with the metabolic syndrome and
T2DM (22, 23, 26).
Finally, alterations in plasma adipocytokine concentrations
have also been proposed to play a major role in the development
of insulin resistance associated with the metabolic syndrome and
T2DM. However, we observed no differences in circulating
concentrations of HMW adiponectin, IL-6, resistin, TNF-?, or
RBP-4 in the two groups (27–29). Taken together these data
suggest that alterations in the concentrations in these plasma
adipocytokines do not play a primary role in the early develop-
ment of insulin resistance and atherogenic dyslipidemia in these
young, lean, insulin-resistant subjects and that the observed
alterations in the concentrations of these adipocytokines in the
metabolic syndrome and T2DM are likely to be secondary to
other factors, such as obesity.
In summary, these data support the hypothesis that insulin
resistance in skeletal muscle, due to decreased muscle glycogen
derived from ingested carbohydrate away from muscle glycogen
synthesis into increased hepatic de novo lipogenesis. These
findings have important implications for understanding the
mechanism by which insulin resistance in skeletal muscle pro-
motes the development of the metabolic syndrome, NAFLD,
T2DM, and the associated cardiovascular disease. These data
triglyceride (C) and hepatic triglyceride (D) content in insulin-sensitive and insulin-resistant subjects after the mixed meals.
13C MRS measurements of changes in muscle (A) and liver (B) glycogen concentrations and1H MRS measurements of changes in intramyocellular
www.pnas.org?cgi?doi?10.1073?pnas.0705408104Petersen et al.
also suggest that reversing defects in insulin-stimulated glucose
transport in skeletal muscle to reverse insulin resistance in this
organ might be the best way to prevent the development of the
metabolic syndrome at its earliest stages of development.
Materials and Methods
Participants. Between 2005 and 2007 we studied ?400 young,
healthy, lean, nonsmoking, sedentary volunteers from the New
Haven community. All subjects were in excellent health, taking
no medications, lean, nonsmoking, of normal birth weight, and
sedentary as defined by a physical activity questionnaire (30) and
subsequently, by 3 days of physical activity monitoring using a
pedometer (Pedometer 342; Sportline). Blood pressure was
measured in the resting state by a digital 300 Vital Signs Monitor
(Welch Allyn, Skaneateles Falls, NY). Subjects were excluded if
they had any metal implants, body piercing, or history of
Yale University Human Investigation Committee approved
after the purpose, nature, and potential complications of the
studies were explained.
ISI. Whole-body insulin sensitivity was assessed with a 3-hr, 75-g
oral glucose tolerance test in combination with the ISI (12, 31,
32). Twenty minutes after insertion of an antecubital i.v. line,
fasting blood samples were collected for determination of
plasma glucose, insulin, fatty acid, C-reactive protein, HMW
adiponectin, resistin, IL-6, RBP-4, and TNF-? concentrations.
The dextrose drink (Glucola; Curtin Matheson Scientific, Hous-
ton, TX) was administered, and blood samples were collected at
10, 20, 30, 60, 90, 120, 150, and 180 min for determination of
plasma glucose, fatty acid, and insulin concentrations. The ISI
measures the overall effects of insulin to stimulate glucose
disposal and inhibit glucose production. From this cohort, 12
the same subject (B), and intraabdominal fat volume in the insulin-sensitive and insulin-resistant subjects (C).
MRI imaging of intraabdominal fat of several slices through the abdomen in a subject (A), three-dimensional reconstruction of intraabdominal fat in
Petersen et al. PNAS ?
July 31, 2007 ?
vol. 104 ?
no. 31 ?
insulin-resistant subjects (four male and eight female; two
African American, eight Caucasian, and two Hispanic) as de-
fined by an ISI of ?3.90 ? 10?4dl/min per microunit/ml (lower
quartile of the ISI distribution) and 12 insulin-sensitive subjects
(five male and seven female; one black and 11 Caucasian) as
defined as having an ISI of ?6.06 ? 10?4dl/min per
microunit/ml (top quartile of ISI distribution) volunteered to be
studied further as described below (13).
All qualifying subjects subsequently underwent dual-energy
x-ray absorptiometry scanning (QDR-4500 W; Hologic, Bed-
ford, MA) to assess lean body mass and fat mass (33), a complete
medical history, a physical examination, and blood tests to verify
that the following were normal: blood and platelet counts,
electrolytes, aspartate amino transferase, alanine amino trans-
ferase, blood urea nitrogen, creatinine, prothrombin time, par-
tial prothrombin time, cholesterol, and triglyceride.
Dietary Regimen. To ensure that all participants were weight-
stable and following a regular dietary regimen, each participant
answered a questionnaire about their usual daily food intake and
eating habits over the past 12 months. For 3 days before the
study, the participants received all their meals from the General
Clinical Research Center (GCRC) Metabolic Kitchen (34–37).
The diet was eucaloric [?35 kcal (1 cal ? 4.18 J) per kilogram
of body weight per day] and contained 55% carbohydrate, 10%
protein, and 35% fat.
Experimental Protocol. On day 1, the subjects were admitted to the
GCRC at 5 p.m. and had dinner at 7 p.m. (33% of their daily
caloric requirements). The subjects remained fasting until the
measurements of lipid and glycogen content in muscle and liver
were performed on day 2 starting at 6:30 a.m. After completion
of these baseline MRS measurements, the subjects were re-
turned to the GCRC for the test meals and determination of de
novo lipogenesis as described below. At 7 p.m. the postprandial
MRS measurements of lipid and glycogen content in muscle and
liver were started. To avoid physical activity, the subjects were
brought to and from the Yale Magnetic Resonance Research
Center in a wheelchair.
De Novo Lipogenesis. After completion of the baseline MRS
measurements, the liquid test meals were served. These meals
were prepared by the GCRC Metabolic Kitchen for each subject
and contained all of the required daily energy (35 kcal per
kilogram of body weight; 55% carbohydrate, 10% protein, and
35% fat) with an additional 25% of the daily energy require-
ments added in the form of sucrose. The meals were equal in size
at 2:30 p.m. Each meal was consumed within 15–20 min.
Between the two test meals at noon and at 2 p.m., the loading
doses of deuterium-labeled water (3 ml per kilogram of body
water; 99.8%; Cambridge Isotopes, Cambridge, MA) was given
in two portions of equal size. To maintain constant deuterium
enrichment in plasma water, deuterium-labeled drinking water
(0.45% enrichment) was given ad libitum for the remainder of
the study. The incorporation of deuterium into lipids during
administration of deuterium-labeled water was used to deter-
mine the fractional synthetic rate of fatty acids as described
previously (3, 4). Blood was collected to assess de novo lipogen-
Fig. 4. Fractional de novo lipogenesis in insulin-sensitive and insulin-resistant subjects after the two high-carbohydrate mixed meals.
mixed meals in insulin-sensitive and insulin-resistant individuals.
Schematic of whole-body energy distribution after high-carbohydrate
www.pnas.org?cgi?doi?10.1073?pnas.0705408104Petersen et al.
esis before the first dose of deuterium-labeled water and hourly
from 10 p.m. until 6 a.m. (4). After centrifugation, plasma was
split into two parts, one part was stored at ?20°C until analysis,
and the other part was processed immediately by ultracentrifu-
gation for purification and removal of chylomicrons and plasma
triglyceride extraction (3, 4, 38). Deuterium enrichment in
palmitate and plasma water was measured by gas chromatog-
raphy–mass spectrometry (5971A Mass Selective Detector;
Hewlett–Packard, Wilmington, DE) as previously described (3,
4, 38), and fractional rates of de novo lipogenesis were calculated
as previously described (39).
1H MRS and13C MRS. All MRS measurements were performed on
a whole-body, 4.0-T Medspec (Bruker, Billerica, MA) system.
Muscle glycogen and lipid content were measured in the calf
muscle by using an 8.5-cm-diameter, circular13C surface coil
with twin, orthogonal circular 13-cm1H quadrature coils. The
probe was tuned and matched, and scout images of the lower leg
were obtained to ensure correct positioning of the subject and to
define an adequate volume for localized shimming using the
FASTMAP procedure (40).13C MRS spectra were acquired in
a (60 ? 30 ? 60)-mm3volume placed within the gastrocnemius/
soleus muscles to measure glycogen content. Protons were
decoupled during the 25-ms acquisition time with a WALTZ-4,
and localization of the volume was performed with a three-
dimensional adiabatic outer-volume suppression.
After the 20-min acquisition for the glycogen spectra, local-
ized1H spectra were acquired to assess the muscle lipid content
from a (10 ? 10 ? 10)-mm3voxel centered in the soleus muscle
using a stimulated echo acquisition mode sequence, with three
modules of water suppression with the chemical-shift selective
imaging sequence. The total lipid content was estimated from
comparison of a water-suppressed lipid spectrum and a lipid-
suppressed water spectrum, with the appropriate peak for each
spectrum on-resonance. The intramyocellular lipid and water
resonances were corrected for T1 and T2 relaxation, and in-
tramyocellular lipid content was expressed as a percentage of the
The glycogen and triglyceride content in the liver was mea-
sured with a coil assembly composed of a 12-cm, circular13C coil
in quadrature. The probe was secured on the abdomen over the
liver with Velcro straps, and a nonmagnetic pneumatic expan-
sion bellows were used for gating the MRS acquisition to the
respiratory movements. After imaging the liver and localized
shimming with the FASTMAP method with respiration gating
(40),13C MRS for liver glycogen content was acquired with an
adiabatic half passage pulse and with WALTZ-4 decoupling
during the acquisition time.
The liver triglyceride content was measured by1H respiratory-
gated stimulated echo acquisition mode sequence spectroscopy
in a (15 ? 15 ? 15)-mm3voxel (41–43). Acquisition was
synchronized to the respiratory cycle and triggered at the end of
expiration. A water-suppressed lipid spectrum and a lipid-
suppressed water spectrum were acquired, and a minimum of
two lipid spectra and two water spectra were time-averaged to
minimize variations due to chest movements. This sequence was
repeated in a second location of the liver to account for liver
inhomogeneity. A minimum of eight spectra was acquired for
each subject and the total lipid content was averaged and
calculated as previously described (44).
Abdominal Fat by MRI. A MRI of the trunk was acquired in each
subject by using a Siemens Sonata 1.5-T Instrument (multi-
breathold T1-weighted acquisition; field of view, 38.0 ? 38.0 cm;
matrix, 256 ? 256; in-plane resolution, 1.48 mm; 50 contiguous
slices; slice thickness, 5-mm). After slice intensity homogeneity
correction, intraabdominal fat was interactively segmented in each
slice by using the Yale BioImage Arbor Suite software package
(www.bioimagesuite.org) and the segmentation map was summed
for the entire trunk to generate volume measurements (33).
Analytical Methods. Plasma glucose concentrations were mea-
sured by using a YSI STAT 2700 Analyzer (Yellow Springs
Instrument Co., Yellow Springs, CA). Plasma concentrations of
insulin, resistin, and HMW adiponectin were measured with
double-antibody RIA kits (Linco, St. Louis, MO), and plasma
High-Sensitivity kits (R&D Systems, Minneapolis, MN). Plasma
fatty acid and triglyceride concentrations were determined by
using a microfluorimetric method (45). RBP4 and transthyretin
protein standards prepared with purified full-length human recom-
binant RBP4 and purified human plasma TTR (Sigma, St. Louis,
MO). Immunodetection was performed with polyclonal antibodies
(DakoCytomation). Plasma VLDL and LDL particle concentra-
tions were determined by using MRS (LipoScience, Raleigh, NC).
Statistical Analysis. Statistical analyses were performed by using
the StatView package (Abacus Concepts, Berkeley, CA). Wil-
coxon Rank Sum test or unpaired Student’s t tests were per-
formed where appropriate to detect statistical differences be-
tween insulin-resistant and insulin-sensitive subjects. Paired t
tests were performed to detect statistical differences within
individuals. All data are expressed as mean ? SEM.
We thank Dr. Henry Ginsberg and Colleen Ngai for assistance with
setting up the method for VLDL kinetics. We also thank Andrea Belous;
Peter Brown; Carolyn Canonica, B.S.; Donna Caseria, R.D.; Christopher
Cunningham, B.S.; Donna Dione; James Dziura, Ph.D.; Donna
D’Eugenio, R.N.; Robin DeGraaf, Ph.D.; Aida Groszmann; Yanna
Kosover; Graeme Mason, Ph.D.; Terry Nixon; Hedy Sarafino; Christine
Simpson; Irina Smolgovsky; Mikhail Smolgovsky; and the staff of the
Yale/New Haven Hospital General Clinical Research Center for expert
technical assistance with the studies and the volunteers for participating
in this study. This work was supported by U.S. Public Health Service
Grants R01 AG-23686 (to K.F.P.), P01 DK-068229 (to G.I.S.), R01
DK-43051 (to B.B.K.), P30 DK-45735, R01 EB-006494 (to X.P.), and
M01 RR-00125; the Yamanouchi USA Foundation; and a Distinguished
Clinical Scientist Award from the American Diabetes Association (to
1. Reaven GM (1988) Diabetes 37:1595–1607.
2. Grundy SM, Brewer HB, Jr, Cleeman, JI, Smith SC, Jr, Lenfant C (2004)
3. Previs SF, Hazey JW, Diraison F, Beylot M, David F, Brunengraber H (1996)
J Mass Spectrom 31:639–642.
4. Diraison F, Pachiaudi C, Beylot M (1997) J Mass Spectrom 32:81–86.
5. Peiris AN, Sothmann MS, Hoffmann RG, Hennes MI, Wilson CR, Gustafson
AB, Kissebah AH (1989) Ann Intern Med 110:867–872.
6. Abate N, Garg A, Peshock RM, Stray-Gundersen J, Grundy SM (1995) J Clin
7. Carr DB, Utzschneider KM, Hull RL, Kodama K, Retzlaff BM, Brunzell
JD, Shofer JB, Fish BE, Knopp RH, Kahn SE (2004) Diabetes 53:2087–
8. Nieves DJ, Cnop M, Retzlaff B, Walden CE, Brunzell JD, Knopp RH, Kahn
SE (2003) Diabetes 52:172–179.
9. Rothman DL, Magnusson I, Cline G, Gerard D, Kahn CR, Shulman RG,
Shulman GI (1995) Proc Natl Acad Sci USA 92:983–987.
10. Perseghin G, Price TB, Petersen KF, Roden M, Cline GW, Gerow K, Rothman
DL, Shulman GI (1996) N Engl J Med 335:1357–1362.
11. Petersen KF, Dufour S, Shulman GI (2005) PLoS Med 2:e233.
12. Petersen KF, Dufour S, Befroy D, Garcia R, Shulman GI (2004) N Engl J Med
13. Petersen KF, Dufour S, Feng J, Befroy D, Dziura J, Dalla Man C, Cobelli C,
Shulman GI (2006) Proc Natl Acad Sci USA 103:18273–18277.
14. Kim JK, Michael MD, Previs SF, Peroni OD, Mauvais-Jarvis F, Neschen S,
Kahn BB, Kahn CR, Shulman GI (2000) J Clin Invest 105:1791–1797.
Petersen et al. PNAS ?
July 31, 2007 ?
vol. 104 ?
no. 31 ?
15. Krauss RM, Siri PW (2004) Endocrinol Metab Clin N Am 33:405–415. Download full-text
16. Klein S (2004) J Clin Invest 113:1530–1532.
17. Reaven GM, Hollenbeck C, Jeng CY, Wu MS, Chen YD (1988) Diabetes
18. Kim JK, Gavrilova O, Chen Y, Reitman ML, Shulman GI (2000) J Biol Chem
19. Kim JK, Fillmore JJ, Chen Y, Yu C, Moore IK, Pypaert M, Lutz EP, Kako Y,
Velez-Carrasco W, Goldberg IJ, et al. (2001) Proc Natl Acad Sci USA
20. Petersen KF, Oral EA, Dufour S, Befroy D, Ariyan C, Yu C, Cline GW,
DePaoli AM, Taylor SI, Gorden P, Shulman GI (2002) J Clin Invest 109:1345–
GI (2004) J Biol Chem 279:32345–32353.
22. Neschen S, Morino K, Hammond LE, Zhang D, Liu ZX, Romanelli AJ, Cline
GW, Pongratz RL, Zhang XM, Choi CS, et al. (2005) Cell Metab 2:55–65.
GW, Yu XX, Geisler JG, et al. (2006) J Clin Invest 116:817–824.
24. Horton JD, Goldstein JL, Brown MS (2002) J Clin Invest 109:1125–1131.
25. Eckel RH, Hernandez TL, Bell ML, Weil KM, Shepard TY, Grunwald GK,
Sharp TA, Francis CC, Hill JO (2006) Am J Clin Nutr 83:803–808.
26. Samuel VT, Liu ZX, Wang A, Beddow SA, Geisler JG, Kahn M, Zhang XM,
Monia BP, Bhanot S, Shulman GI (2007) J Clin Invest 117:739–745.
27. Wellen KE, Hotamisligil GS (2005) J Clin Invest 115:1111–1119.
28. Lazar MA (2006) Nat Med 12:43–44.
29. Shoelson SE, Lee J, Goldfine AB (2006) J Clin Invest 116:1793–1801.
30. Baecke JA, Burema J, Frijters JE (1982) Am J Clin Nutr 36:936–942.
31. Matsuda M, DeFronzo RA (1999) Diabetes Care 22:1462–1470.
C (2005) Diabetes 54:3265–3273.
33. Petersen KF, Hendler R, Price T, Perseghin G, Rothman DL, Held N,
Amatruda JM, Shulman GI (1998) Diabetes 47:381–386.
34. Warram JH, Martin BC, Krolewski AS, Soeldner JS, Kahn CR (1990) Ann
Intern Med 113:909–915.
35. Johnson AB, Argyraki M, Thow JC, Cooper BG, Fulcher G, Taylor R (1992)
Clin Sci (London) 82:219–226.
36. Krssak M, Falk Petersen K, Dresner A, DiPietro L, Vogel SM, Rothman DL,
Roden M, Shulman GI (1999) Diabetologia 42:113–116.
37. Perseghin G, Scifo P, De Cobelli F, Pagliato E, Battezzati A, Arcelloni C,
Vanzulli A, Testolin G, Pozza G, Del Maschio A, Luzi L (1999) Diabetes
38. Folch J, Lees M, Sloane Stanley GH (1957) J Biol Chem 226:497–509.
39. Diraison F, Pachiaudi C, Beylot M (1996) Metabolism 45:817–821.
40. Gruetter R (1993) Magn Reson Med 29:804–811.
41. Frahm J, Bruhn H, Gyngell ML, Merboldt KD, Hanicke W, Sauter R (1989)
Magn Reson Med 9:79–93.
42. Moonen CWT, Van Zijl PCM (1988) J Magn Res 88:28–41.
43. Pauly JM, Le Roux P, Nishimura A, Macovski A (1991) IEEE Trans Med
44. Mayerson AB, Hundal RS, Dufour S, Lebon V, Befroy D, Cline GW,
Enocksson S, Inzucchi SE, Shulman GI, Petersen KF (2002) Diabetes 51:797–
45. Miles J, Glasscock R, Aikens J, Gerich J, Haymond M (1983) J Lipid Res
46. Graham TE, Wason CJ, Bluher M, Kahn BB (2007) Diabetologia 50:814–823.
www.pnas.org?cgi?doi?10.1073?pnas.0705408104Petersen et al.