THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 280, NO. 39, pp. 33588 –33598, September 30, 2005
© 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.
Peroxisome Proliferator-activated Receptor-␥ Co-activator
1␣-mediated Metabolic Remodeling of Skeletal Myocytes
Mimics Exercise Training and Reverses Lipid-induced
Mitochondrial Inefficiency*
Received for publication, July 13, 2005, and in revised form, July 29, 2005 Published, JBC Papers in Press, August 3, 2005, DOI 10.1074/jbc.M507621200
Timothy R. Koves‡, Ping Li§, Jie An‡, Takayuki Akimoto§, Dorothy Slentz‡, Olga Ilkayeva‡, G. Lynis Dohm¶,
Zhen Yan§, Christopher B. Newgard‡, and Deborah M. Muoio‡1
From the ‡Departments of Medicine and Pharmacology & Cancer Biology, and the Sarah W. Stedman Nutrition and Metabolism
Center, Duke University, Durham, North Carolina 27710, the §Department of Medicine, Duke University, Durham, North Carolina
27710, and the ¶Department of Physiology, Brody Medical School, East Carolina University, Greenville, North Carolina 27834
Obesity and type 2 diabetes are two closely associated diseases that
continue to affect Westernized societies at epidemic rates (1). Results
from both epidemiological and animal studies suggest that the risk of
developing these diseases is increased by consumption of a high fat diet
(2), which has been attributed in part to increased intramuscular triacylglycerol storage and adverse changes in skeletal muscle energy
metabolism (3). Moreover, mounting evidence from animal and cellbased models indicates that lipid oversupply to skeletal muscle results in
functional perturbations that eventually manifest as impaired insulin
action (2, 4, 5). Together, these and other reports have established a
compelling connection between muscle lipid dysregulation and
impaired metabolic control at both the cellular and systemic levels.
Despite intense investigation, a clear understanding of the biochemical
and molecular mechanisms that mediate lipid-induced muscle dysfunction has remained elusive.
Similar to high fat feeding, exercise training increases skeletal muscle
supply and storage of lipid substrates (6). However, in contrast to the
adverse events provoked by a high fat diet, exercise is known to promote
muscle as well as whole body metabolic fitness (7). These observations
suggest that physical activity somehow enhances the capacity of the
muscle to tolerate and/or adapt to a high lipid load. Indeed, contractile
activity leads to dramatic adjustments in oxidative fuel metabolism that
are largely mediated by an increase in muscle mitochondrial biogenesis
(8). Most interestingly, emergent evidence links muscle mitochondrial
insufficiencies to the development of both age-related and inherited
insulin resistance (9 –14). These reports prompted us to consider
whether compromised muscle mitochondrial performance might be a
central component of diet-induced obesity. We further reasoned that
the anticipated mitochondrial abnormalities caused by lipid overload
might be reversed by exercise training.
Here we demonstrate that a high fat diet imposes a persistent lipid
burden on muscle mitochondria, resulting in a disproportionate
increase in incomplete compared with complete ␤-oxidation. The
capacity of isolated mitochondria to fully oxidize a heavy influx of fatty
acid depended on factors such as fiber type and exercise training and
was positively correlated with expression levels of PPAR␥ co-activator-␣ (PGC1␣),2 a promiscuous co-activator of most nuclear hormone
receptors as well as several other transcription factors (15). In skeletal
muscle, PGC1␣ has been shown to stimulate mitochondrial biogenesis
via co-activation of the nuclear respiratory factor (16) and to regulate
genes involved in oxidative phosphorylation through interactions with
estrogen-related receptor ␣ (17). PGC1␣ also co-activates the peroxisome proliferator-activated receptors (PPARs), a family of lipid activated nuclear hormone receptors that play a key role in mediating
adaptive regulation of muscle fatty acid oxidation (18). To test the role
of PGC1␣ in permitting an efficient lipid-induced substrate switch, we
2
* This work was supported by National Institutes of Health Grants DK-56112 (to D. M. M.)
and P01 DK58398 (to C. B. N.), the American Diabetes Association (to D. M. M.), and
GlaxoSmithKline. The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1
To whom correspondence should be addressed: 4321 Medical Park Dr., Ste. 200, Sarah
W. Stedman Nutrition and Metabolism Center, Duke University, Durham, NC 27704.
Tel.: 919-479-2328; Fax: 919-477-0632; E-mail: muoio@duke.edu.
33588 JOURNAL OF BIOLOGICAL CHEMISTRY
The abbreviations used are: PGC1␣, PPAR␥ co-activator 1␣; PPAR, peroxisome proliferator-activated receptor; GOT, glutamate-oxaloacetate transaminase; HF, high fat;
MDH, malate dehydrogenase; PGC1␤, PPAR␥ co-activator 1␤; PDH, pyruvate dehydrogenase; PDK4, pyruvate dehydrogenase kinase 4; SC, standard chow; ANOVA,
analysis of variance; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone;
ASM, acid-soluble metabolites; BisTris, 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol; MOPS, 4-morpholinepropanesulfonic acid; COXIV, cytochrome c oxidase subunit IV; NEFA, non-esterified fatty acid; FAM, N-(3-fluoranthyl)maleimide (carboxyethylfluorescein).
VOLUME 280 • NUMBER 39 • SEPTEMBER 30, 2005
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
Peroxisome proliferator-activated receptor-␥ co-activator 1␣
(PGC1␣) is a promiscuous co-activator that plays a key role in regulating mitochondrial biogenesis and fuel homeostasis. Emergent
evidence links decreased skeletal muscle PGC1␣ activity and coincident impairments in mitochondrial performance to the development of insulin resistance in humans. Here we used rodent models
to demonstrate that muscle mitochondrial efficiency is compromised by diet-induced obesity and is subsequently rescued by exercise training. Chronic high fat feeding caused accelerated rates of
incomplete fatty acid oxidation and accumulation of ␤-oxidative
intermediates. The capacity of muscle mitochondria to fully oxidize
a heavy influx of fatty acid depended on factors such as fiber type
and exercise training and was positively correlated with expression
levels of PGC1␣. Likewise, an efficient lipid-induced substrate
switch in cultured myocytes depended on adenovirus-mediated
increases in PGC1␣ expression. Our results supported a novel paradigm in which a high lipid supply, occurring under conditions of
low PGC1␣, provokes a disconnect between mitochondrial ␤-oxidation and tricarboxylic acid cycle activity. Conversely, the metabolic remodeling that occurred in response to PGC1␣ overexpression favored a shift from incomplete to complete ␤-oxidation. We
proposed that PGC1␣ enables muscle mitochondria to better cope
with a high lipid load, possibly reflecting a fundamental metabolic
benefit of exercise training.
PGC1␣ and Muscle Lipid Metabolism
evaluated metabolic adaptation to a high lipid supply in L6 myotubes
that expressed high levels compared with low levels of the co-activator.
In a manner that resembled exercise training, PGC1␣ coordinated the
induction of ␤-oxidative enzymes with the activation of downstream
metabolic pathways (e.g. tricarboxylic acid cycle), and thereby favored a
shift from incomplete to complete oxidation of lipid substrates. The
implications of these findings with respect to insulin resistance and its
reversal by physical activity are discussed.
EXPERIMENTAL PROCEDURES
SEPTEMBER 30, 2005 • VOLUME 280 • NUMBER 39
JOURNAL OF BIOLOGICAL CHEMISTRY
33589
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
Animal Studies—Wistar rats (300 g; Charles River Breeding Laboratories, Wilmington, MA) or male C57/Bl6J mice were housed in a temperature-controlled environment with a 12:12 h light/dark cycle and
were provided ad libitum access to Purina standard chow (SC) or high
fat diet (HF; Research Diets 45% fat) and water. On the day of the experiments, food was removed either 4 (fed) or 24 h (starved) before anesthesia was administered. Studies were conducted in accordance with
Duke University Institutional Animal Care and Use Committee.
Exercise and Denervation—Rats (n ⫽ 7) were habituated and treadmill-trained for 4 weeks as described previously (19). Untrained rats
(n ⫽ 7) served as controls. Effects of acute exercise were tested in mice
(n ⫽ 10) that were subjected to 2–3-h bouts of treadmill running (0%
incline, 20 –28 m䡠min⫺1) separated by 1 h of rest and compared against
rested controls (n ⫽ 10). Muscles were excised 12 h after the second
bout of exercise. Effects of muscle inactivity were tested in mice that
underwent right hind limb denervation by severing the sciatic nerve
(n ⫽ 12) (20), compared against the contralateral sham-operated hind
limb as a control. Muscles were excised 72 h after surgery. Mouse exercise training studies were performed in animals fed on either a SC or HF
diet for 14 weeks. During the final 2 weeks of the diet, mice were kept
sedentary or exercise-trained by voluntary wheel running (21). Intraperitoneal glucose tolerance tests were performed 48 h after the last bout
of voluntary running in mice that were fasted overnight (22).
Mitochondrial Isolation and Fatty Acid Oxidation—Muscle mitochondria were isolated by differential centrifugation, and mitochondrial
integrity was assessed as described previously (19, 23). The final mitochondrial pellet was resuspended in 1 ml of suspension buffer consisting
of (in mM) 250 sucrose, 10 Tris-HCl, 1 EDTA, and 2 ATP, pH 7.4. Fatty
acid oxidation was conducted using 40 ␮l of the final mitochondrial
suspensions incubated in the presence of 160 ␮l of an oxidation buffer
containing (in mM) 0.150 oleate, ([1-14C]oleate, at 0.5 ␮Ci ml⫺1), 100
sucrose, 10 Tris-HCl, 10 KH2PO4, 100 KCl, 1 MgCl2, 1 L-carnitine, 0.1
malate, 2 ATP, 0.05 coenzyme A, 1 dithiothreitol, and 0.3% bovine
serum albumin for 30 min at 37 °C (19). The reaction was terminated by
the addition of 100 ␮l of 70% perchloric acid. 14CO2 was trapped in 200
␮l of 1 N NaOH. 14C-Labeled acid-soluble metabolites (ASM) and
14
CO2 fractions were quantified by liquid scintillation counting, and
data were normalized to mitochondrial protein.
L6 Muscle Cells—L6 myoblasts (ATCC, Manassas, VA) were grown
on 6- or 12-well collagen-coated plates in Dulbecco’s modified Eagle’s
medium supplemented with 10% fetal bovine serum and 50 mg/ml gentamycin in a humidified incubator at 37 °C with 5% CO2 until ⬃80%
confluent. Cells were induced to differentiate into myotubes in Dulbecco’s modified Eagle’s medium with 2% horse serum and 50 ␮g/ml gentamycin (differentiation medium). On day 2, cells were treated with 20
␮l (4 ⫻ 109 plaque-forming units/ml) of purified adenovirus per 500,000
cells and then 24 h later with vehicle or fatty acids. Myotubes were
harvested 24 h after fatty acid treatment for isolation of protein and
RNA. Fatty acid oxidation assays were performed following incubations
in 500 ␮l of differentiation medium plus 12.5 mM HEPES, 0.5% bovine
serum albumin, 1.0 mM carnitine, 100 –500 ␮M sodium oleate, and 1.0
␮Ci/ml [14C]oleate (Sigma) for 3 h at 37 °C. Medium was transferred to
new dishes and assayed for labeled oxidation products (CO2 and ASM)
as described previously (24). Cells washed twice with ice-cold phosphate-buffered saline were harvested in 200 ␮l of 0.05% SDS, and lysates
were stored at ⫺80 °C for protein determination.
Western Blotting—L6 cell and gastrocnemius tissue lysates were
homogenized directly in loading buffer consisting of 50 mM Tris-HCl,
10% glycerol, 0.01% bromphenol blue, 20 mM dithiothreitol, 143 mM
␤-mercaptoethanol, and 1% SDS supplemented with phosphatase and
protease inhibitor mixtures (Sigma). Protein concentrations were determined from cell and tissue extracts using the RC/DC protein assay
(Bio-Rad). Twenty ␮g of cellular protein or 40 ␮g of gastrocnemius
lysate were separated by SDS-PAGE using NuPAGE 4 –12% BisTris
gradient gel and MOPS electrophoresis buffers (Invitrogen), transferred
to polyvinylidene difluoride membranes, and probed overnight with
either PGC1␣ (1:1000; Chemicon, Temecula, CA), COXIV (1:1000;
MitoSciences, Eugene, OR), or ␣-tubulin (1:20,000; Sigma) antibodies.
Proteins were visualized by horseradish peroxidase-conjugated immunoglobulin G antibodies and ECL Plus chemiluminescence detection kit
and were quantified by using ImageQuant software (Amersham
Biosciences).
Recombinant Adenovirus—Recombinant adenovirus encoding
mouse PGC1␣ (AdCMV-PGC1␣) was a kind gift from Dr. Bruce
Spiegelman. Recombinant adenoviruses encoding ␤-galactosidase
(AdCMV-␤-gal) or green fluorescent protein (AdCMV-GFP) were used
to control for nonspecific effects of virus treatment. The adenoviruses
were amplified and purified using methods published previously (25).
Acylcarnitine Profiling—Ten mg of powdered muscle tissue was
transferred to a chilled 1-ml Potter-Elvehjem homogenizer containing
0.3 ml of distilled, de-ionized water. Samples were subjected to 20
strokes and immediately transferred to 1.5-ml tubes and sonicated at 2.5
watts for 5 s and then subsequently processed as described previously
(26). Acylcarnitines were measured by direct-injection electrospray tandem mass spectrometry, using a Micromass Quattro MicroTM system
equipped with a model 2777 autosampler, a model 1525 high pressure
liquid chromatography solvent delivery system, and a data system running 4.0 MassLynx software (Waters, Milford, MA).
Whole-cell Polarographic Measurements—L6 myotubes grown on
15-cm collagen-coated plates were treated with control (␤-galactosidase) or PGC1␣ recombinant adenoviruses and harvested after 48 h.
Myocytes were washed twice with phosphate-buffered saline,
trypsinized, and centrifuged at 800 ⫻ g. Cell pellets were washed twice
and resuspended in 700 ␮l of phosphate-buffered saline containing 10
mM HEPES, 0.2% bovine serum albumin, and either no additional substrate, 10 mM glucose, or 100 ␮M palmitate. Small aliquots of the cell
suspension (50 –200 ␮l) were added to a final volume of 500 ␮l, and
oxygen consumption was determined using an OxyTherm system
(Hansatech Instruments, Norfolk, UK) calibrated at 37 °C. Uncoupled
respiration was stimulated by the addition of 2 ␮M FCCP.
Transcriptional Profiling by Microarray and Real Time Quantitative
PCR—High quality RNA was isolated using the RNeasy total RNA isolation kit (Qiagen, Valencia, CA). For muscle tissues, RNA was isolated
from 10 to 20 mg of soleus, extensor digitorum longus, or gastrocnemius
muscle using the RNeasy fibrous tissue kit (Qiagen, Valencia, CA). RNA
quality was verified using an Agilent Bioanalyzer (Agilent Technologies,
Palo Alto, CA) and quantified using RiboGreen reagent (Molecular
Probes, Eugene, OR). For cell studies, L6 myotubes were grown on collagen-coated 6-well plates and treated for 48 h with control (␤-galactosidase) or PGC1␣ in the presence or absence of 500 ␮M sodium oleate
for 24 h. RNA from day 4 myotubes was harvested during two separate
experiments that were performed in triplicate. Pooled samples from
PGC1␣ and Muscle Lipid Metabolism
each treatment group were analyzed at the Duke University DNA
Microarray Center using the Rat Genome Oligo Set version 1.1 customprinted microarray (Operon, Huntsville, AL) that was prepared with a
GeneMachines OmniGrid arrayer (Genomic Solutions, Ann Arbor,
MI). Samples were analyzed against a reference standard, and the raw
microarray data were imported into GeneSpring 5.1 software (Silicon
Genetics, Redwood City, CA). After Lowess transformation normalization, the data were screened for genes having spot intensities of “present” or “marginal.” Cross-gene error model was set for deviation from
1.0, and relevant genes were limited to those having changed at least
1.5-fold. The resulting annotated list was used for GeneSpring pathway
analysis to detect coordinated changes in lipid-related metabolic pathways. cDNA for real time PCR was synthesized from 1 ␮g of RNA using
the IScript cDNA synthesis kit (Bio-Rad) in a 20-␮l reaction volume and
diluted 5-fold for subsequent reverse transcription-PCR. Results from
the pooled-sample fluorescent microarray analysis were validated by
real time quantitative PCR using a Prism 7000 with TaqMan chemistry
33590 JOURNAL OF BIOLOGICAL CHEMISTRY
(Applied Biosystems, Foster City, CA) to measure differentially
expressed genes in individual samples. Sequence-specific primers and
probes were designed using Primer Express software (Applied Biosystems, Foster City, CA). The sequences for the mouse genes were as
follows: PGC1␣, forward, 5⬘-GCACCAGCCAACACTCA-3⬘, reverse,
5⬘-TGGGTGTGGTTTGCTGCA-3⬘, and probe, 5⬘-FAM-ACAATGAATGCAGCGGTC-3⬘; PGC1␤, forward, 5⬘-GCTCTCTTGGCTGCGCTTA-3⬘, reverse, 5⬘-ACGATGTGGGGCTG-3⬘, and probe,
5⬘-FAM-AAGACCCTGGATGACATCC. All other genes were
assessed using predesigned/prevalidated FAM-labeled Assays-on-Demand from Applied Biosystems and normalized using values from a
duplexed reaction with VIC-labeled 18 S endogenous control gene
(Applied Biosystems, Foster City, CA).
Statistics—Data were analyzed by Student’s t test or by analysis of
variance (ANOVA). ANOVA was performed on data at a minimum p ⬍
0.05 threshold, and a multiple comparison Fisher’s post hoc test was
performed to evaluate differences between exercise and/or dietary
treatments with JMP software version 5 (SAS Institute Inc., Cary, NC).
Data are shown as means ⫾ S.E.
RESULTS
Physiological Regulation of Fatty Acid Oxidation in Skeletal Muscle
Mitochondria—Muscles that are exercise-trained or enriched in type I
(red) myofibers exhibit enhanced fatty acid oxidative capacity com-
VOLUME 280 • NUMBER 39 • SEPTEMBER 30, 2005
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
FIGURE 1. Fatty acid metabolism in rat muscle mitochondria. Mitochondria were isolated from deep red or superficial white rat gastrocnemius muscle (A), whole rat gastrocnemius muscles from sedentary and 4-week exercise-trained rats (B), and whole gastrocnemius muscles harvested in the ad libitum fed or 24-h starved state from rats fed on a
either a SC or HF diet for 12 weeks (C). Mitochondria were incubated for 30 min at 37 °C
in the presence of 150 ␮M [1-14C]palmitate, and radiolabel incorporation in CO2 was
determined as a measure of complete oxidation. D, label incorporation into ASM was
measured to assess incomplete fatty acid oxidation. Complete and incomplete oxidation
rates were normalized to total mitochondrial protein. E, gastrocnemius muscles were
harvested from rats that were allowed to feed ad libitum (fed) or starved 24 h after 12
weeks on either an SC or HF diet. Muscle acylcarnitine profiles were evaluated by tandem
mass spectrometry and are expressed as a percent of SC-fed controls. Data represent
means ⫾ S.E. from 5 to 8 animals per group. Statistical differences between diet groups
were analyzed by Student’s t test or ANOVA. * indicates p ⬍ 0.05; ** indicates p ⬍ 0.01 in
SC versus HF; ‡ indicates p ⬍ 0.05 comparing fed versus starved groups. ANOVA revealed
a starvation effect (p ⬍ 0.01) and diet effect (p ⬍ 0.01) on the muscle acylcarnitine
profiles, but symbols were excluded for simplicity. FAO, fatty acid oxidation.
FIGURE 2. Regulation of muscle PGC1␣ and PGC1␤ mRNA expression in response to
energy stress. Total RNA was harvested from rat soleus, whole gastrocnemius (gastroc),
or extensor digitorum longus (Edl) muscles (A); gastrocnemius muscles from sedentary
(Sed) or acutely exercised mice (B); sham-operated or denervated mouse gastrocnemius
muscles (C); and soleus muscles from rats fed on a SC or HF diet for 12 weeks, harvested
in the ad libitum fed or 24-h starved state (D). Gene expression was normalized using 18 S
and an endogenous control, and data are presented as fold change and represent
mean ⫾ S.E. of 5–10 animals per group. Differences between treatments were analyzed
by Student’s t test. * indicates p ⬍ 0.01.
PGC1␣ and Muscle Lipid Metabolism
pared with their sedentary or type II (white) muscle counterparts. To
determine whether these properties are due strictly to differences in
mitochondrial content or, alternatively, could reflect distinctions in
mitochondrial performance, we evaluated [1-14C]palmitate oxidation in
isolated rat muscle mitochondria. Most strikingly, we found that oxidation rates (expressed per g mitochondrial protein) in mitochondria from
the deep red gastrocnemius were 8-fold higher than those from the
superficial white gastrocnemius (Fig. 1A). Similarly, exercise training,
which causes a shift in muscle fiber composition, increased mitochondrial fatty acid oxidative capacity ⬃2-fold over the sedentary condition
(Fig. 1B).
Exercise-trained and/or type I skeletal muscles are also characterized
by increased lipid uptake and elevated intramuscular triacylglycerol
storage (6, 27). Thus, lipid-induced increases in PPAR transcriptional
activity could contribute to enhanced mitochondrial ␤-oxidative capacity. In order to assess whether a change in fatty acid supply would be
sufficient to alter mitochondrial lipid metabolism, we next evaluated
muscle mitochondria from rats that were starved 24 h or fed a high fat
diet for 10 weeks. Similar to previous studies by our laboratory (26, 28)
and others (29), both manipulations increased circulating nonesterified
free fatty acids by ⬃2-fold and coincidentally increased mRNA expression of several known PPAR target genes (not shown). However, unlike
the exercise and fiber type models, neither of the nutritional maneuvers
enhanced the capacity of the muscle mitochondria to fully oxidize
palmitate to CO2 (Fig. 1C). In these studies we also measured label
incorporation into chain-shortened ASM, which provide an index of
incomplete ␤-oxidation. Compared with the SC-fed condition, both
SEPTEMBER 30, 2005 • VOLUME 280 • NUMBER 39
starvation and HF feeding increased 14C label incorporation into the
ASM fraction by 36 and 27%, respectively (Fig. 1D). ASM production
was greatest in muscle mitochondria from HF rats harvested in the
starved condition.
In Vivo Metabolic Analysis by Acylcarnitine Profiling—To gain further insight into diet-induced changes in mitochondrial function, we
employed tandem mass spectrometry to profile mitochondrially
derived acylcarnitine esters in muscles from rats fed either a SC or a HF
diet. The acylcarnitine analysis provides a comprehensive snapshot of in
vivo substrate flux through specific steps of the ␤-oxidative spiral and is
thus commonly used as a diagnostic tool to detect inborn metabolic
errors (30). Here and in previous studies we have found this to be an
effective method for screening diet and/or obesity-related metabolic
abnormalities (26).
In rats fed the SC diet, starvation increased several medium and long
chain acylcarnitine intermediates, consistent with increased fatty acid
delivery and high rates of ␤-oxidation that occur in the fasted condition
(Fig. 1G). In muscles from HF-fed animals, most lipid-derived (even
chain) acylcarnitine intermediates were persistently elevated (compared with the SC-fed group), but increased only modestly during the
fed-to-starved transition. In contrast, muscle levels of isovaleryl-carnitine (C5), which is derived from the catabolism of leucine, were higher
in the SC compared with the HF group but were unaffected by fasting.
The lowering of this metabolite by the HF diet most likely reflects a
mitochondrial substrate switch from branched chain amino acids to
fatty acids. In general, the nutritionally induced changes in the acylcar-
JOURNAL OF BIOLOGICAL CHEMISTRY
33591
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
FIGURE 3. Adenovirus-mediated PGC1␣ overexpression in rat L6 myocytes. Rat L6 myocytes
were treated with recombinant adenoviruses
encoding ␤-galactosidase (␤-gal), green fluorescent protein (GFP), or PGC1␣ (4 ⫻ 109 plaqueforming units/ml), and protein expression of the
transgene was measured 48 h later. A shows high
efficiency transduction of ␤-galactosidase (left)
and green fluorescent protein (right) in differentiated myotubes. PGC1␣ mRNA (B) and protein (C)
levels increased dose-dependently with increasing viral titer. D, COX IV protein levels in no virus
(NVC) and ␤-galactosidase-treated control cells
compared against cells treated with increasing
doses of rAd-CMV-PGC1␣.
PGC1␣ and Muscle Lipid Metabolism
TABLE ONE
Pathway analysis of fatty acid-induced gene activation in L6 myocytes expressing high or low PGC1a
Rat L6 myocytes were treated with recombinant adenoviruses encoding ␤-galactosidase (␤-gal) or PGC1␣ on differentiation day 2. Twenty four h after addition
of virus, cells were incubated an additional 24 h in control differentiation medium with bovine serum albumin or with the addition of 500 ␮M oleate (FA). Pooled
RNAs from replicate experiments were analyzed by the Duke University DNA Microarray Center using the Rat Genome Oligo Set version 1.1 custom-printed
microarray. Pathway analysis was performed using GeneSpring 5.1 software. Gene expression data are presented as fold change relative to ␤-galactosidase control
cells grown in the absence of fatty acid.
Fatty acid metabolism
Fold change vs. ␤-galactosidase control
Common name
GenBankTM accession no.
␤-Gal ⴙ FA
PGC1␣
PGC1␣ ⴙ FA
Abbreviation
3.3
2.4
2.6
2.4
1.4
2.7
1.4
0.9
1.6
1.5
2.1
0.9
2.1
1.7
1.6
1.2
32
5.3
4.2
4.1
3.6
3
2.7
2.5
2.3
1.9
1.8
1.7
1.7
1.7
1.7
1.5
1.0
1.0
1.0
0.9
1.0
1.1
0.8
1.0
1.3
1.3
2.7
2.5
2.7
2.7
2.3
2.1
2.1
4.0
3.5
1.8
3.7
3.6
2.9
2.8
2.8
2.2
2.0
6.3
5.0
1.9
0.8
1.1
1.1
1.0
0.8
0.9
0.9
0.9
5.5
4.3
4.9
3.2
3.6
2.7
2.4
2.0
6.5
5.3
5.7
3.6
3.6
3.0
2.5
2.2
0.8
1.1
1.2
2.6
2.3
2.0
2.5
2.0
1.6
pdhk4
Pyruvate dehydrogenate kinase 4
cat
Carnitine/acylcarnitine translocase
mcad
Acetyl-CoA dehydrogenase, medium chain
vlcad
Acyl-CoA dehydrogenase, very long chain
aac2
Acetyl-coenzyme A acyltransferase 2
fabp3
Fatty acid-binding protein 3
ucp2
Uncoupling protein 2
ech1
Enoyl-coenzyme A hydratase 1
fabp4
Fatty acid-binding protein 4
scad-bcdh
Short chain acyl-coenzyme A dehydrogenase
ampkb
AMP-activated protein kinase ␤2 reg. subunit
acs2
Fatty acid coenzyme A ligase, long chain 2
crat
Similar to carnitine acetyltransferase
echs1
Enoyl-coenzyme A hydratase, short chain 1
fatp1
fatty acid transport protein
acat1
Acetyl-coenzyme A acetyltransferase 1
Oxidative phosphorylation
ant5
Adenine nucleotide translocator 5
atp5g1
ATP synthase subunit c, isoform 1
atp5d
ATP synthase ␦ subunit
atp5g3
ATP synthase subunit c (subunit 9) isoform 3
ant1
Rat gene for adenine nucleotide translocator
atp5c1
ATP synthase ␥-polypeptide 1
atp5b
ATP synthase ␤-polypeptide
fdx1
Ferredoxin 1
cycs
Cytochrome c, somatic
cox6a2
Cytochrome c oxidase subunit VIa
Tricarboxylic acidcycle/malate-aspartate shuttle
got1
Glutamate oxaloacetate transaminase 1
mdh1
Malate dehydrogenase 1
cs
Citrate synthase
idh3a
Isocitrate dehydrogenase 3 (NAD⫹) ␣
aco2
Mitochondrial aconitase
mdh2
Malate dehydrogenase 2
got2
Glutamate oxaloacetate transaminase 2
idh3b
NAD⫹-specific isocitrate dehydrogenase b
Oxidative stress
sod2
Superoxide dismutase 2
sod3
Superoxide dismutase 3
aldh5a1
Aldehyde dehydrogenase family 5, subfA1
nitine profile are consistent with the notion that starvation and HF
feeding increase mitochondrial rates of incomplete ␤-oxidation.
Transcriptional Regulation of PGC1␣ and PGC1␤ in Skeletal Muscle—
The results presented in Fig. 1 suggested that under some circumstances
PPAR activation alone is insufficient to enhance the potential of muscle
mitochondria to fully oxidize lipid substrates. As we considered other factors that may be important in regulating mitochondrial function, the transcriptional co-activators PGC1␣ and PGC1␤ emerged as attractive candidates. Both PGC1 isoforms have been implicated as master regulators of
mitochondrial biogenesis and important modulators of PPAR activity. In
liver, heart, and brown adipose tissue, PGC1␣ and PGC1␤ are transcrip-
33592 JOURNAL OF BIOLOGICAL CHEMISTRY
AF034577
R000462_01
NM_016986_71
R003155_01
NM_130433
NM_024162
NM_019354
NM_022594
NM_053365
NM_022512
NM_022627
NM_012820
XM_242301
X15958
NM_053580
NM_017075
NM_057102
NM_017311
NM_139106
NM_053756
NM_053515
NM_053825
NM_134364
NM_017126
NM_012839
NM_012812
NM_012571
AF093773
NM_130755
AB047541
NM_024398
AB047540
NM_013177
NM_031151
NM_017051
NM_012880
NM_012880
tionally regulated by various nutritional and/or energy stresses (31). We
therefore proceeded to investigate skeletal muscle mRNA expression of
both PGC1 isoforms under a variety of physiological conditions (Fig. 2).
Expression of PGC1␣ was 4.5-fold greater in soleus (red) than in gastrocnemius (mixed) or extensor digitorum longus (white) muscle, whereas conversely, PGC1␤ mRNA was most abundant in the extensor digitorum longus muscle (Fig. 2A). An acute bout of exercise increased PGC1␣
expression 2.2-fold in the gastrocnemius muscle (Fig. 2B), whereas denervation (a model of inactivity) provoked a 50% decrease in expression levels
(Fig. 2C). In contrast, neither exercise nor denervation altered expression of
PGC1␤ (Fig. 2, B and C). Chronic high fat feeding resulted in a marked
VOLUME 280 • NUMBER 39 • SEPTEMBER 30, 2005
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
2.7
1.0
0.7
1.1
1.7
0.9
2.2
1.7
1.4
0.9
2.3
1.1
0.8
0.9
1.1
0.9
PGC1␣ and Muscle Lipid Metabolism
SEPTEMBER 30, 2005 • VOLUME 280 • NUMBER 39
FIGURE 4. Microarray validation by real time PCR. Rat L6 myocytes were treated with
recombinant adenoviruses encoding ␤-galactosidase (␤-gal) or PGC1␣ on differentiation day 2. Twenty four h after addition of virus, cells were incubated an additional 24 h
in differentiation medium without (control) or with the addition of 500 ␮M oleate (FA).
Microarray results were validated by real time quantitative PCR performed on individual
samples. A, classic PPAR target genes, uncoupling proteins (UCP2 and UCP3), and pyruvate dehydrogenase kinase 4 (PDK4); B, tricarboxylic acid cycle enzymes; aconitase and
citrate synthase (CS); C, enzymes of ␤-oxidation; ACAA2, acetyl-CoA acyltransferase 2;
CAT, carnitine acylcarnitine translocase; CPT1, carnitine palmitoyltransferase 1; MCAD,
medium chain acyl-CoA dehydrogenase; VLCAD, very long chain acyl-CoA dehydrogenase. D, cytosolic enzymes, glutamate-oxaloacetate transaminase (GOT) and MDH. Gene
expression was normalized to 18 S as an endogenous control, and data are presented as
fold change (mean ⫾ S.E.) relative to ␤-galactosidase-treated control cells. Experiments
were performed in triplicate, and differences between ␤-galactosidase and PGC1␣
groups were analyzed by Student’s t test.
␮M (Fig. 5A), whereas label incorporation into ASM increased 2.2-fold
(Fig. 5B). In comparison, myocytes treated with rAd-CMV-PGC1␣ displayed 2.2–2.6-fold increases in 14CO2 production, but only 1.8- to 1.6fold increases in labeling of ASM. The relationship between incomplete
and complete fatty acid oxidation was quantified by calculating the ratio
of radiolabeled ASM relative to CO2 (Fig. 5C). In control cells treated
with 100 ␮M oleate, label incorporation into ASM was ⬃1.3-fold greater
than into CO2. This ratio increased to 1.8 when the oleate concentration
was raised to 500 ␮M. Notably, however, under conditions of high
PGC1␣ activity, production of 14CO2 and 14C-labeled ASM was equal
(ratio of ⬃1.0) at both low and high fatty acid concentrations. Thus,
PGC1␣ modified mitochondrial function in a manner that favored
tighter coupling between ␤-oxidation and tricarboxylic acid cycle
flux. Oximetry was used as an alternative measure of tricarboxylic
acid cycle activity. In the presence of both lipid (palmitate) and nonlipid (glucose) substrates, oxygen consumption was ⬃2.5-fold
greater in myotubes transduced with rAd-CMV-PGC1␣ compared
with rAd-CMV-␤-gal (Fig. 5D). Likewise, in the presence of FCCP,
an uncoupling agent that permits assessment of maximal respiratory
JOURNAL OF BIOLOGICAL CHEMISTRY
33593
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
decrease (⬃80%) in soleus muscle PGC1␣ but no significant change in
PGC1␤ (Fig. 2D). Starvation also tended to decrease PGC1␣. Thus, transcriptional regulation of PGC1␣, but not PGC1␤, correlated with changes
in mitochondrial fatty acid oxidation capacity (Fig. 1). This expression pattern indicates that PGC1␣ and PGC1␤ are differentially responsive to
energy stress and suggests that the ␣ isoform plays a primary role in mediating adaptive changes in muscle mitochondrial function.
Overexpression of PGC1 in L6 Muscle Cells—Data from our animal
models implied that PGC1␣ may permit more efficient adaptation to a
high lipid environment. To test this hypothesis, we used recombinant
adenovirus to increase the normally low levels of PGC1␣ expressed in
rat L6 myotubes. Subsequent experiments then evaluated the transcriptional and metabolic consequences of a 24-h fatty acid challenge in cells
expressing high compared with low PGC1␣ activity. Preliminary studies
using recombinant adenoviruses encoding either ␤-galactosidase or
green fluorescent protein confirmed high efficiency transduction of differentiated myotubes (Fig. 3A). Likewise, both PGC1␣ mRNA and protein levels increased dose-dependently with an increasing titer of rAdCMV-PGC1␣ (Fig. 3, B and C). Similar to previous reports, PGC1␣
overexpression caused a marked induction of the COXIV protein (Fig.
3D), consistent with an increase in mitochondrial biogenesis.
Microarray Analysis of PGC1␣-mediated Reprogramming of Muscle
Cells—We used DNA microarray analysis to gain a more comprehensive perspective of how PGC1␣ modulates transcriptional responses to
lipid exposure (TABLE ONE). Treatment of myocytes with high fatty
acid resulted in predictable increases in the expression of several classic
PPAR-targeted genes involved in fatty acid catabolism (PDHK4, 2.7fold; UCP2, 2.2-fold; AAC2, 1.7-fold; and ECH1, 1.7-fold). Fatty acid
exposure alone produced little or no effects on transcripts associated
with the tricarboxylic acid cycle or the electron transport chain. In contrast, PGC1␣ overexpression not only activated several ␤-oxidative and
lipid trafficking genes but also triggered the global induction of genes
involved in the tricarboxylic acid cycle, the electron transport chain,
oxidative phosphorylation, and protection against oxidative stress. In
general, overexpression of PGC1␣ potentiated lipid-mediated activation of classic PPAR target genes.
Expression levels of a select set of metabolic genes were then validated
by real time quantitative PCR (Fig. 4). These assays demonstrated strong
agreement with the microarray data and confirmed PGC1␣-mediated
induction of genes encoding mitochondrial enzymes such as aconitase
and citrate synthase, as well as cytosolic enzymes such as malate dehydrogenase (MDH1) and glutamate oxaloacetate transaminase (GOT1).
Consistent with data presented in TABLE ONE, real time quantitative
PCR revealed subsets of lipid-responsive genes that could be classified
into three distinct patterns of regulation: 1) genes that were similarly
responsive to fatty acid in the ␤-galactosidase compared with PGC1␣
overexpressing cells (UCP2 and UCP3); 2) genes that responded to fatty
acid only when PGC1␣ was overexpressed (medium chain acyl-CoA
dehydrogenase, very long chain acyl-CoA dehydrogenase, and carnitine
acylcarnitine translocase); and 3) genes for which a response to fatty
acid was present in control cells but then further potentiated by high
expression of PGC1␣ (carnitine palmitoyltransferase 1, acetyl-CoA
acyltransferase 2, and PDK4).
PGC1␣ Overexpression Enhances Complete Fatty Acid Oxidation—
We next investigated the impact of increased PGC1␣ activity on metabolic responses to a high lipid load. Similar to muscle mitochondria
from starved or high fat fed rats (Fig. 1), L6 myocytes exposed to increasing fatty acid concentrations exhibited disproportionate increases in the
rates of incomplete (ASM) relative to complete (CO2) ␤-oxidation. In
both sets of control cells, [14C]oleate oxidation to CO2 increased only
1.6-fold when the fatty acid concentration was raised from 100 to 500
PGC1␣ and Muscle Lipid Metabolism
capacity, oxygen consumption was 80% greater in myotubes that
overexpressed PGC1␣.
Metabolic Analysis of Exercise Training by Acylcarnitine Profiling—
Results from our cell-based experiments suggested that exercise-induced activation of PGC1␣ might alleviate mitochondrial inefficiencies
caused by HF feeding. To test this possibility, we performed a chronic
(14-week) HF feeding study in which half of the mice in each group were
exercise-trained during the final 2 weeks of the diet. Consistent with our
studies in rats (Fig. 2), the HF diet decreased PGC1␣ expression ⬃50%
in mouse gastrocnemius muscle. In this experiment muscles were harvested from overnight-fasted animals that had been rested for 48 h,
thereby minimizing confounding effects of acute activity. Although an
acute exercise bout increased muscle PGC1␣ mRNA abundance (Fig.
2), we did not observe a training effect on transcript levels in the mice
fed the SC diet (Fig. 6A). This was not surprising because several exercise-induced transcripts, including PGC1␣, are transiently increased
with each exercise bout (32, 33). More interestingly however, the exercise regimen reversed the HF diet-mediated suppression of PGC1␣ and
increased mRNA expression by 2.8-fold over sedentary mice that were
fed the HF diet. Exercise tended to increase COXIV protein abundance,
although these changes did not reach statistical significance (Fig. 6B).
Similar to the results in Fig. 1, preliminary studies in mice confirmed
a robust fasting-induced rise in circulating NEFA and a coincident
increase in muscle acylcarnitine levels (Fig. 7, A and B). Accordingly, we
chose the overnight-fasted condition, representing a state of high mitochondrial fatty acid influx, to evaluate the impact of endurance training
on muscle acylcarnitine profiles. Resembling the data in Fig. 1G, when
muscles were harvested from starved mice most short and medium
acylcarnitines were similarly abundant regardless of the diet (Fig. 7C).
However, the HF diet did increase the C18:0 and C18:1 acylcarnitine
33594 JOURNAL OF BIOLOGICAL CHEMISTRY
species. The exercise training regimen lowered most lipid-derived acylcarnitines but did not affect the leucine metabolite (C5), suggesting
lipid-specific adaptations in mitochondrial fuel metabolism. Comparable with the interplay that we observed between PGC1␣, fatty acids, and
gene regulation in cultured muscle cells (Fig. 4), the most robust adjustments in fatty acylcarnitine metabolites occurred in exercise-trained
mice fed the HF regimen. Thus, like PGC1␣ overexpression in L6 myocytes, exercise training appeared to optimize mitochondrial processing
of excess lipid substrate.
Finally, Fig. 7D shows the results of an intraperitoneal glucose tolerance test that was performed 48 h after the last exercise bout. Similar to
previous reports (26), animals fed the HF diet increased their body
weight and displayed impaired glucose tolerance. Although the short
term exercise intervention did not affect body weight, food intake, or
circulating NEFA (not shown), it was sufficient to normalize the
glucose tolerance test, even with continued feeding of the HF diet.
Thus, notably, exercise-induced lowering of lipid-derived muscle acylcarnitines was accompanied by the reversal of HF diet-induced glucose
intolerance.
DISCUSSION
PPARs ␣, ␦, and ␥ comprise a family of lipid-activated nuclear hormone receptors that regulate fatty acid homeostasis. Numerous reports
indicate that PPAR-targeted fatty acid oxidative genes are induced by
physiological manipulations that raise circulating NEFA, such as exercise, starvation, and high fat feeding (18, 28, 29). Still, data that address
the direct impact of these transcriptional adaptations on muscle mitochondrial performance are lacking. In our study both overnight starvation and chronic HF feeding of rats produced the anticipated induction
of known PPAR target genes, but without comparable changes in the
VOLUME 280 • NUMBER 39 • SEPTEMBER 30, 2005
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
FIGURE 5. PGC1␣ enhances complete oxidation
of fatty acids. Rat L6 myocytes were treated with
recombinant adenoviruses encoding ␤-galactosidase (␤-gal) or PGC1␣ on differentiation day 2
and were compared against a no virus control
(NVC) group. Forty eight h after addition of virus,
cells were incubated 3 h with 100 –500 ␮M
[14C]oleate. Fatty acid oxidation was determined
by measuring 14C label incorporation into CO2 (A)
and ASM (B), representing complete and incomplete oxidation, respectively. C, the relationship
between incomplete and complete fatty acid oxidation was expressed as a ratio of label incorporated into ASM divided by labeling of CO2. D, oximetry experiments were performed 48 h after
addition of virus in L6 cells that were trypsinized
and transferred to an oximetry chamber. Oxygen
consumption was measured in the presence of
400 ␮M palmitate, 25 mM glucose, or the uncoupling agent FCCP. Data represent the mean ⫾ S.E.
from three to four experiments. Differences
between groups were analyzed by ANOVA and
Student’s t test, * indicates p ⬍ 0.05 comparing
PGC1␣ to NVC and ␤-galactosidase treatments; ‡
indicates p ⬍ 0.05 comparing low and high FA
conditions.
PGC1␣ and Muscle Lipid Metabolism
FIGURE 6. Effects of high fat diet and exercise on
muscle PGC1␣ mRNA levels in mice. Mice were
fed on SC or HF diets for 14 weeks. During the final
2 weeks of the diet, half of the mice in each group
were kept sedentary (Sed) or exercise-trained (Ex)
by voluntary running wheel. Gastrocnemius muscles from five to six mice per group were harvested
48 h after the last exercise bout following a 24-h
fast. Training-induced changes in PGC1␣ mRNA
levels normalized to 18 S (A) and COX IV protein
expression normalized to ␣-tubulin (B). Values are
mean ⫾ S.E., and statistical differences between
groups were analyzed by ANOVA. * indicates p ⬍
0.05 compared with SC sedentary, and ‡ indicates
p ⬍ 0.05 compared with the HF sedentary group.
potential of muscle mitochondria to completely oxidize fatty acid substrate. In contrast, both maneuvers caused a disproportionate increase
in mitochondrial rates of incomplete ␤-oxidation. A similar phenomenon occurred when muscle cells were exposed to increasing doses of
fatty acid. In both the animal and cell culture models, the inability of
mitochondria to increase complete fatty acid oxidation in response to
lipid exposure was associated with low expression of PGC1␣. Conversely, raising PGC1␣ activity, in vivo by exercise training or in cultured myocytes by rAd-mediated gene delivery, favored complete fatty
acid oxidation to CO2. Although previous studies have established a
clear role for PGC1␣ in regulating mitochondrial biogenesis (15, 16, 34),
SEPTEMBER 30, 2005 • VOLUME 280 • NUMBER 39
our report now demonstrates a strong functional link between PGC1␣
and efficient mitochondrial oxidation of fatty acid.
Our results also show that isolated mitochondria from red versus
white and trained versus untrained muscle exhibited greater potential to
fully oxidize lipid substrate. These observations imply that fiber type
and exercise-related adjustments in fuel homeostasis are mediated not
only by an increase in mitochondrial number but also by functional
changes in the resident mitochondrial population. Consistent with previous reports (16, 35), we found that exercise training and red fiber type
are associated with increased mRNA expression of PGC1␣, but no
changes in PGC1␤. We therefore surmised that the ␣ isoform might be
JOURNAL OF BIOLOGICAL CHEMISTRY
33595
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
FIGURE 7. Metabolic analysis of exercise training in mice fed a standard or high fat diet. The
effects of 24 h of starvation (Starved) compared
with ad libitum feeding (Fed) on serum NEFA (A)
and gastrocnemius acylcarnitine profile (B)
(nmol/mg protein) were tested in mice fed on SC.
Subsequent analyses were performed in mice fed
on SC or HF diets for 14 weeks. During the final 2
weeks of the diet, half of the mice in each group
were kept sedentary (Sed) or exercise-trained (Ex)
by running wheel. Intraperitoneal glucose tolerance tests and harvest of gastrocnemius muscles
were executed 48 h after the last exercise bout,
following a 24-h starvation. Results show traininginduced changes in muscle acylcarnitine profile,
expressed as a percent of SC-fed controls (C) and
systemic glucose tolerance tests (D). Values are
mean ⫾ S.E. for 5– 6 mice per group. Statistical differences between groups were analyzed by twoway ANOVA. * indicates p ⬍ 0.05, and ** indicates
p ⬍ 0.01 for glucose tolerance tests results of the
HF sedentary group. ANOVA revealed a starvation
effect (p ⬍ 0.01) and a main exercise effect (p ⬍
0.01) on the muscle acylcarnitine profile, but symbols were excluded for simplicity.
PGC1␣ and Muscle Lipid Metabolism
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
FIGURE 8. Schematic of PGC1␣-mediated remodeling of muscle lipid metabolism. In the presence of low PGC1␣ activity, fatty acid influx and PPAR-mediated activation of target
genes (shown in blue) promote incomplete ␤-oxidation. Exercise and/or PGC1␣ coordinates ligand-induced PPAR activity with remodeling of downstream metabolic pathways
(shown in red), thereby favoring complete fatty acid oxidation (FAO) to CO2. ACO, aconitase; AMPK, AMP-activated kinase; ANT, adenine nucleotide transporter; ATPS, ATP synthase;
CPT-CAT, carnitine palmitoyltransferase system; CS, citrate synthase; ETC, electron transport chain; FABP, fatty acid-binding proteins; GOT, glutamate-oxaloacetate transaminase; ICD,
isocitrate dehydrogenase; PC, pyruvate carboxylase.
primarily responsible for maintaining a mitochondrial population that
can accommodate a high influx of lipid fuel. Supporting this notion,
treatment of myocytes with rAd-PGC1␣ triggered a broad program of
genes involved in ␤-oxidative production of acetyl-CoA as well as its
subsequent flux into the tricarboxylic acid cycle. PGC1␣-mediated
transcriptional reprogramming coincided with a shift from incomplete
to complete oxidation of fatty acids. Together, these results suggest that
efficient substrate switching from glucose to fatty acid not only requires
PPAR-mediated induction of ␤-oxidative enzymes but also the coordi-
33596 JOURNAL OF BIOLOGICAL CHEMISTRY
nated regulation of downstream metabolic pathways that are not
directly targeted by the PPAR system.
PDK4, which encodes an enzyme that phosphorylates and inhibits
the PDH complex, ranked highest among the metabolic transcripts that
were induced by PGC1␣. We and others have shown that this gene is
highly responsive to both exercise and starvation (28). Inactivation of
PDH by PDK4-mediated phosphorylation not only promotes lipid oxidation but also serves to spare pyruvate for use as an anaplerotic substrate. Similarly, increased PGC1␣ activity yielded higher mRNA levels
VOLUME 280 • NUMBER 39 • SEPTEMBER 30, 2005
PGC1␣ and Muscle Lipid Metabolism
SEPTEMBER 30, 2005 • VOLUME 280 • NUMBER 39
PGC1␣ activity, when combined with excess lipid and tonic PPAR activation, provoke a mitochondrial disconnect between ␤-oxidation and
tricarboxylic acid cycle flux, thereby triggering the production and/or
accumulation of ␤-oxidative intermediates. Although the pathophysiological implications of incomplete fat oxidation are yet unexplored, it is
tempting to speculate that this event might contribute to mitochondrial
malfunction and whole body metabolic dysregulation. This possibility is
supported by the finding that muscle-specific overexpression of PPAR␣
in transgenic mice caused both local and systemic glucose intolerance,
in association with marked induction of several lipid-oxidative genes
(46). The diabetic phenotype of these animals was reversed by administration of the carnitine palmitoyltransferase 1 inhibitor, oxfenicine,
further implicating a role for excessive ␤-oxidation. Moreover, several
reports have shown that PPAR␣ null mice are protected against lipidinduced diabetes despite greater susceptibility to obesity (47, 48). Likewise, we have shown that genetic reversal of HF diet-induced insulin
resistance in rats correlates with decreased muscle production of ␤-oxidative by-products (26). Studies to investigate potential links between
discordant regulation of PPAR and PGC1␣, incomplete ␤-oxidation,
and glucose intolerance now await further investigation.
Acknowledgment—We are grateful to Dr. Bruce Spiegelman for generously
providing us with the recombinant adenovirus encoding PGC1␣.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Zimmet, P., Alberti, K. G., and Shaw, J. (2001) Nature 414, 782–787
Lichtenstein, A. H., and Schwab, U. S. (2000) Atherosclerosis 150, 227–243
Shulman, G. I. (2000) J. Clin. Investig. 106, 171–176
McGarry, J. D. (2002) Diabetes 51, 7–18
Perseghin, G., Petersen, K., and Shulman, G. I. (2003) Int. J. Obes. Relat. Metab.
Disord. 27, Suppl. 3, S6 –S11
Goodpaster, B. H., He, J., Watkins, S., and Kelley, D. E. (2001) J. Clin. Endocrinol.
Metab. 86, 5755–5761
Astrup, A. (2001) Public Health Nutr. 4, 499 –515
Adhihetty, P. J., Irrcher, I., Joseph, A. M., Ljubicic, V., and Hood, D. A. (2003) Exp.
Physiol. 88, 99 –107
Patti, M. E., Butte, A. J., Crunkhorn, S., Cusi, K., Berria, R., Kashyap, S., Miyazaki, Y.,
Kohane, I., Costello, M., Saccone, R., Landaker, E. J., Goldfine, A. B., Mun, E., Defronzo, R., Finlayson, J., Kahn, C. R., and Mandarino, L. J. (2003) Proc. Natl. Acad. Sci.
U. S. A. 100, 8466 – 8471
Patti, M. E. (2004) Curr. Opin. Clin. Nutr. Metab. Care 7, 383–390
Lowell, B. B., and Shulman, G. I. (2005) Science 307, 384 –387
Iossa, S., Mollica, M. P., Lionetti, L., Crescenzo, R., Tasso, R., and Liverini, G. (2004)
Diabetes 53, 2861–2866
Schrauwen, P., and Hesselink, M. K. (2004) Diabetes 53, 1412–1417
Petersen, K. F., Dufour, S., Befroy, D., Garcia, R., and Shulman, G. I. (2004) N. Engl.
J. Med. 350, 664 – 671
Puigserver, P., and Spiegelman, B. M. (2003) Endocr. Rev. 24, 78 –90
Lin, J., Wu, H., Tarr, P. T., Zhang, C. Y., Wu, Z. D., Boss, O., Michael, L. F., Puigserver,
P., Isotani, E., Olson, E. N., Lowell, B. B., Bassel-Duby, R., and Spiegelman, B. M. (2002)
Nature 418, 797– 801
Mootha, V. K., Handschin, C., Arlow, D., Xie, X., St. Pierre, J., Sihag, S., Yang, W.,
Altshuler, D., Puigserver, P., Patterson, N., Willy, P. J., Schulman, I. G., Heyman, R. A.,
Lander, E. S., and Spiegelman, B. M. (2004) Proc. Natl. Acad. Sci. U. S. A. 101,
6570 – 6575
Gilde, A. J., and Van Bilsen, M. (2003) Acta Physiol. Scand. 178, 425– 434
Koves, T. R., Noland, R. C., Bates, A. L., Henes, S. T., Muoio, D. M., and Cortright,
R. N. (2005) Am. J. Physiol. 288, C1074 –C1082
Jones, J. P., Tapscott, E. B., Olson, A. L., Pessin, J. E., and Dohm, G. L. (1998) J. Appl.
Physiol. 84, 1661–1666
Waters, R. E., Rotevatn, S., Li, P., Annex, B. H., and Yan, Z. (2004) Am. J. Physiol. 287,
C1342–C1348
Almind, K., and Kahn, C. R. (2004) Diabetes 53, 3274 –3285
Kim, J. Y., Koves, T. R., Yu, G. S., Gulick, T., Cortright, R. N., Dohm, G. L., and Muoio,
D. M. (2002) Am. J. Physiol. 282, E1014 –E1022
Muoio, D. M., Way, J. M., Tanner, C. J., Winegar, D. A., Kliewer, S. A., Houmard, J. A.,
Kraus, W. E., and Dohm, G. L. (2002) Diabetes 51, 901–909
Becker, T. C., BeltrandelRio, H., Noel, R. J., Johnson, J. H., and Newgard, C. B. (1994)
J. Biol. Chem. 269, 21234 –21238
An, J., Muoio, D. M., Shiota, M., Fujimoto, Y., Cline, G. W., Shulman, G. I., Koves,
JOURNAL OF BIOLOGICAL CHEMISTRY
33597
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
of both the mitochondrial and cytosolic isoforms of GOT and MDH,
enzymes that play important roles in shuttling tricarboxylic acid cycle
intermediates and reducing equivalents in and out of the mitochondria
(Fig. 8). Thus, PGC1␣ appears to facilitate fatty acid catabolism by activating anaplerotic reactions that provide the carbon skeletons necessary
to sustain high tricarboxylic acid cycle activity.
Consistent with other reports showing that GOT1 is increased by
exercise training and expressed more abundantly in red compared with
white muscle (36 –38), we found that mRNA levels of this enzyme in L6
myocytes were increased over 20-fold by PGC1␣ overexpression. GOT1
catalyzes the cytosolic conversion of aspartate to oxaloacetate, which
represents the rate-limiting step of the malate-aspartate shuttle (36).
The conversion of oxaloacetate to malate is then catalyzed by MDH1 in
an NADH-consuming reaction. Malate can be transported into the
mitochondria where the opposite reaction then delivers NADH to the
respiratory chain. This system plays a principal role in permitting oxidation of cytosolic reducing equivalents that are produced by glycolysis
and/or via the oxidation of lactate, both of which increase in response to
muscle contraction. Most importantly, however, the malate-aspartate
carrier system operates as a reversible shuttle that also allows mitochondrial export of reducing equivalents. This process is likely to occur during transitions from exercise to rest, when fatty acid supply and ␤-oxidative production of NADH exceeds ATP demand. A high NADH/
NAD ratio slows oxidative phosphorylation at several steps. Thus, by
dissipating the rapid intramitochondrial accumulation of reducing
equivalents that occurs upon cessation of contractile activity, the malate
shuttle might play a key role in preventing the complete shutdown of
oxidative metabolism.
A seminal observation presented here is that HF diet-induced insulin
resistance in rodents occurred in association with decreased muscle
PGC1␣ expression and persistent elevations in intramuscular acylcarnitines, metabolic by-products of incomplete fatty acid oxidation. Conversely, exercise training, which is thought to increase PGC1␣ activity
via both transcriptional and post-translational mechanisms (43, 44),
reversed lipid-induced glucose intolerance in conjunction with the lowering muscle acylcarnitine levels. Perhaps these results move us closer to
solving a long standing paradox in exercise and diabetes research, i.e.
both red and exercise-trained muscles are more insulin-sensitive than
their white or untrained counterparts, despite higher lipid influx and
storage (6, 27, 39). Our findings suggest that increased PGC1␣ activity
and/or enhanced mitochondrial efficiency may protect against lipidinduced insulin resistance. Similarly, emerging evidence from human
studies links decreased PGC1␣ activity, and accompanying reductions
in the expression of genes involved in oxidative phosphorylation, to the
etiology of type 2 diabetes (9, 11, 40). Moreover, both inherited and
acquired insulin resistance has been associated with perturbations in
the metabolic and morphologic properties of skeletal muscle mitochondria (11, 14, 41, 42).
Still unanswered, however, is whether or not PGC1␣ activation alone,
in the absence of exercise, would be sufficient to confer protection
against diabetes. Although muscle-specific overexpression of PGC1␣ in
transgenic mice results in mitochondrial proliferation and increased
expression of genes involved in oxidative phosphorylation (16), the
impact of this manipulation on whole body glucose homeostasis has not
been reported. Most surprisingly, PGC1␣ null mice are paradoxically
lean and resistant to diet-induced obesity (45). However, the phenotype
of these animals was complicated by dramatic abnormalities in the brain
and centrally mediated hyperactivity; thus, a clearer understanding of
the metabolic ramifications imparted by low muscle PGC1␣ levels will
depend on a tissue-specific knock-out.
In summary, our conceptual model predicts that conditions of low
PGC1␣ and Muscle Lipid Metabolism
T. R., Stevens, R., Millington, D., and Newgard, C. B. (2004) Nat. Med. 10, 268 –274
27. Cortright, R. N., Muoio, D. M., and Dohm, G. L. (1997) Nutr. Biochem. 8, 228 –245
28. Muoio, D. M., MacLean, P. S., Lang, D. B., Li, S., Houmard, J. A., Way, J. M., Winegar,
D. A., Corton, J. C., Dohm, G. L., and Kraus, W. E. (2002) J. Biol. Chem. 277,
26089 –26097
29. de Fourmestraux, V., Neubauer, H., Poussin, C., Farmer, P., Falquet, L., Burcelin, R.,
Delorenzi, M., and Thorens, B. (2004) J. Biol. Chem. 279, 50743–50753
30. Van Hove, J. L., Zhang, W., Kahler, S. G., Roe, C. R., Chen, Y. T., Terada, N., Chace,
D. H., Iafolla, A. K., Ding, J. H., and Millington, D. S. (1993) Am. J. Hum. Genet. 52,
958 –966
31. Spiegelman, B. M., and Heinrich, R. (2004) Cell 119, 157–167
32. Pilegaard, H., Saltin, B., and Neufer, P. D. (2003) J. Physiol. (Lond.) 546, 851– 858
33. Pilegaard, H., Ordway, G. A., Saltin, B., and Neufer, P. D. (2000) Am. J. Physiol. 279,
E806 –E814
34. St. Pierre, J., Lin, J., Krauss, S., Tarr, P. T., Yang, R., Newgard, C. B., and Spiegelman,
B. M. (2003) J. Biol. Chem. 278, 26597–26603
35. Akimoto, T., Pohnert, S. C., Li, P., Zhang, M., Gumbs, C., Rosenberg, P. B., Williams,
R. S., and Yan, Z. (2005) J. Biol. Chem. 280, 19587–19593
36. Barron, J. T., Gu, L., and Parrillo, J. E. (1998) J. Mol. Cell. Cardiol. 30, 1571–1579
37. Barron, J. T., Gu, L., and Parrillo, J. E. (1999) Mol. Cell. Biochem. 194, 283–290
38. Hittel, D. S., Kraus, W. E., Tanner, C. J., Houmard, J. A., and Hoffman, E. P. (2005)
J. Appl. Physiol. 98, 168 –179
39. Muoio, D. M., Dohm, G. L., Tapscott, E. B., and Coleman, R. A. (1999) Am. J. Physiol.
276, E913–E921
40. Mootha, V. K., Lindgren, C. M., Eriksson, K. F., Subramanian, A., Sihag, S., Lehar, J.,
Puigserver, P., Carlsson, E., Ridderstrale, M., Laurila, E., Houstis, N., Daly, M. J.,
Patterson, N., Mesirov, J. P., Golub, T. R., Tamayo, P., Spiegelman, B., Lander, E. S.,
Hirschhorn, J. N., Altshuler, D., and Groop, L. C. (2003) Nat. Genet. 34, 267–273
41. Ritov, V. B., Menshikova, E. V., He, J., Ferrell, R. E., Goodpaster, B. H., and Kelley, D. E.
(2005) Diabetes 54, 8 –14
42. Petersen, K. F., Befroy, D., Dufour, S., Dziura, J., Ariyan, C., Rothman, D. L., DiPietro,
L., Cline, G. W., and Shulman, G. I. (2003) Science 300, 1140 –1142
43. Puigserver, P., Rhee, J., Lin, J. D., Wu, Z. D., Yoon, J. C., Zhang, C. Y., Krauss, S.,
Mootha, V. K., Lowell, B. B., and Spiegelman, B. M. (2001) Mol. Cell 8, 971–982
44. Baar, K., Wende, A. R., Jones, T. E., Marison, M., Nolte, L. A., Chen, M., Kelly, D. P.,
and Holloszy, J. O. (2002) FASEB J. 16, 1879 –1886
45. Lin, J., Wu, P. H., Tarr, P. T., Lindenberg, K. S., St. Pierre, J., Zhang, C. Y., Mootha,
V. K., Jager, S., Vianna, C. R., Reznick, R. M., Cui, L., Manieri, M., Donovan, M. X., Wu,
Z., Cooper, M. P., Fan, M. C., Rohas, L. M., Zavacki, A. M., Cinti, S., Shulman, G. I.,
Lowell, B. B., Krainc, D., and Spiegelman, B. M. (2004) Cell 119, 121–135
46. Finck, B. N., Bernal-Mizrachi, C., Han, D. H., Coleman, T., Sambandam, N.,
LaRiviere, L., Holloszy, J. O., Semenkovich, C. F., and Kelly, D. P. (2005) Cell Metab. 1,
133–144
47. Tordjman, K., Bernal-Mizrachi, C., Zemany, L., Weng, S., Feng, C., Zhang, F., Leone,
T. C., Coleman, T., Kelly, D. P., and Semenkovich, C. F. (2001) J. Clin. Invest. 107,
1025–1034
48. Guerre-Millo, M., Rouault, C., Poulain, P., Andre, J., Poitout, V., Peters, J. M., Gonzalez, F. J., Fruchart, J. C., Reach, G., and Staels, B. (2001) Diabetes 50, 2809 –2814
Downloaded from www.jbc.org at VIVA, Univ of Virginia on May 19, 2009
33598 JOURNAL OF BIOLOGICAL CHEMISTRY
VOLUME 280 • NUMBER 39 • SEPTEMBER 30, 2005