FGF21 induces PGC-1? and regulates carbohydrate
and fatty acid metabolism during the adaptive
Matthew J. Potthoffa,b,c, Takeshi Inagakid, Santhosh Satapatic, Xunshan Dinga,b, Tianteng Hec, Regina Goetze,
Moosa Mohammadie, Brian N. Finckf,g, David J. Mangelsdorfa,b,1, Steven A. Kliewera,d,1, and Shawn C. Burgessa,c
aDepartment of Pharmacology,bHoward Hughes Medical Institute,cAdvanced Imaging Center, anddDepartment of Molecular Biology, University of Texas
Southwestern Medical Center, Dallas, TX 75390;eDepartment of Pharmacology, New York University School of Medicine, New York, NY 10016; andfCenter
for Cardiovascular Research andgCenter for Human Nutrition, Department of Medicine, Washington University School of Medicine, St. Louis, MO 63110
Contributed by David J. Mangelsdorf, April 21, 2009 (sent for review March 17, 2009)
The liver plays a crucial role in mobilizing energy during nutritional
deprivation. During the early stages of fasting, hepatic glycogen-
olysis is a primary energy source. As fasting progresses and
glycogen stores are depleted, hepatic gluconeogenesis and keto-
genesis become major energy sources. Here, we show that fibro-
blast growth factor 21 (FGF21), a hormone that is induced in liver
by fasting, induces hepatic expression of peroxisome proliferator-
activated receptor ? coactivator protein-1? (PGC-1?), a key tran-
scriptional regulator of energy homeostasis, and causes corre-
sponding increases in fatty acid oxidation, tricarboxylic acid cycle
flux, and gluconeogenesis without increasing glycogenolysis. Mice
lacking FGF21 fail to fully induce PGC-1? expression in response to
a prolonged fast and have impaired gluconeogenesis and keto-
genesis. These results reveal an unexpected relationship between
FGF21 and PGC-1? and demonstrate an important role for FGF21 in
coordinately regulating carbohydrate and fatty acid metabolism
during the progression from fasting to starvation.
lipid metabolism ? liver ? gluconeogenesis ? glycogenolysis ? ketogenesis
the early stages of fasting, the liver mobilizes glucose from its
glycogen stores. As fasting progresses and glycogen reserves are
depleted, the liver oxidizes fat to provide both energy for
gluconeogenesis and substrate for ketogenesis. This synchroni-
the normal fasting response; disruption of either one of these
pathways has profound effects on the other (1–4).
Hormones such as glucagon, catecholamines, and glucocorti-
coids have important roles in controlling substrate utilization
and maintaining energy balance during fasting. Recently, the
hormone fibroblast growth factor 21 (FGF21) was shown to be
induced in the liver during fasting (5–7). FGF21 is an unusual
FGF family member in that it lacks the conventional heparin-
binding domain (8) and thus can diffuse away from its tissue of
origin and function as a hormone. FGF21 signals through
cell-surface receptors composed of classic FGF receptors com-
plexed with ?-klotho, a membrane-spanning protein (9–14).
Induction of FGF21 during fasting occurs through a mechanism
that requires peroxisome proliferator-activated receptor ?
(PPAR?) (5–7). FGF21 has diverse metabolic actions that
include stimulating hepatic fatty acid oxidation and ketogenesis
(5, 6, 15) and blocking the growth hormone signaling pathway
(16). FGF21 also sensitizes mice to torpor, a short-term hiber-
nation-like state of regulated hypothermia (6). Pharmacologic
administration of FGF21 to insulin-resistant rodents and mon-
keys improves glucose tolerance and reduces plasma insulin and
triglyceride concentrations (15, 17).
Peroxisome proliferator-activated receptor ? coactivator pro-
tein-1? (PGC-1?) is a transcriptional coactivator protein whose
n mammals, the liver plays a crucial role in maintaining
systemic energy balance during fasting and starvation through
expression is induced in response to changes in nutritional status
and other physiologic stimuli such as cold and exercise (18–21).
PGC-1? is enriched in metabolic tissues, such as muscle and
heart, where it interacts with multiple DNA-binding transcrip-
tion factors to stimulate mitochondrial metabolic capacity. In
liver, induction of PGC-1? by fasting stimulates the transcription
of genes involved in fatty acid oxidation, tricarboxylic acid
(TCA) cycle flux, mitochondrial oxidative phosphorylation, and
gluconeogenesis (4, 22–24). Accordingly, these metabolic pro-
cesses are impaired in mice in which PGC-1? function has been
disrupted in liver (4, 25–28). In this article, we show that FGF21
induces PGC-1? and has marked effects on carbohydrate and
lipid metabolism. Our findings reveal a prominent role for
FGF21 in coordinating the adaptive metabolic response to
affects liver metabolism, fatty acid and carbohydrate fluxes were
measured by2H/13C NMR isotopomer analysis in isolated liver
from WT mice and transgenic mice overexpressing FGF21 in a
liver-enriched manner (FGF21-TG) (6). As expected, hepatic
oxygen consumption was increased in response to fasting in WT
liver (Fig. 1A). Importantly, FGF21-TG liver from fed mice
exhibited increased oxygen consumption rates that were equiv-
alent to those observed in WT fasted liver (Fig. 1A). No further
increase in oxygen consumption occurred in FGF21-TG liver in
response to fasting (Fig. 1A). To characterize hepatic fat oxi-
dation in more detail, we measured TCA cycle flux and keto-
genesis. Ketogenesis was significantly increased, and there was a
trend toward increased TCA cycle flux in fed FGF21-TG liver
(Fig. 1A). Overall, hepatic ?-oxidation nearly doubled in fed
FGF21-TG liver (Fig. 1A). These data are consistent with the
previous description of FGF21 as a ketogenic factor (5, 6) and
provide additional evidence that FGF21 is critical for the
induction of hepatic fat oxidation during fasting.
Overexpression of FGF21 also affected carbohydrate metab-
olism. Notably, gluconeogenesis was increased 60% in fed
FGF21-TG liver compared with that of fed WT liver (Fig. 1B)
and was not induced any further by fasting (Fig. 1B). These data
gluconeogenesis in the fed state to levels normally attained
Author contributions: M.J.P., T.I., D.J.M., S.A.K., and S.C.B. designed research; M.J.P., T.I.,
S.S., X.D., T.H., and B.N.F. performed research; R.G. and M.M. contributed new reagents/
and S.C.B. wrote the paper.
The authors declare no conflict of interest.
1To whom correspondence may be addressed. E-mail: firstname.lastname@example.org
This article contains supporting information online at www.pnas.org/cgi/content/full/
June 30, 2009 ?
vol. 106 ?
no. 26 ?
during prolonged fasting. The concept that FGF21 expression
induces a metabolic state in liver that mimics long-term fasting
was reinforced by impaired glycogenolysis (Fig. 1B) and an
accompanying 50% increase in glycogen content in FGF21-TG
liver (Fig. 1B). There was no difference in plasma glucagon
concentrations between FGF21-TG and WT mice (Fig. S1A).
These data are consistent with an increased reliance on glu-
coneogenesis and an autoregulatory sparing of glycogen (29).
Taken together, our findings reveal an important role for FGF21
in regulating changes in both fatty acid and carbohydrate me-
tabolism during a prolonged fast.
FGF21 Induces PGC-1?. FGF21 was previously shown to induce
hepatic expression of genes involved in fatty acid metabolism,
including lipoprotein lipase (Lpl), pancreatic lipase (Pnlip),
pancreatic lipase-related protein 2 (Pnliprp2), and carboxyl ester
lipase (Cel) (5, 6). To gain additional insight into how FGF21
exerts its effects on hepatic metabolism, we performed microar-
most strongly induced genes in FGF21-TG liver was peroxisome
proliferator-activated receptor-gamma coactivator 1? (Pgc1?),
which encodes a transcriptional coactivator that is crucial for
coordinating gluconeogenesis and fatty acid oxidation in liver
(18–20). Real-time quantitative PCR (QPCR) confirmed that
Pgc1? mRNA was induced 5-fold in FGF21-TG liver compared
with that in WT liver (Fig. 2A), and PGC-1? protein expression
also was elevated in FGF21-TG liver compared with that in WT
liver (Fig. 2B). By contrast, mRNA levels of the related factor,
PGC-1?, were unchanged in FGF21-TG liver (Fig. S1B). The
PGC-1? target genes glucose-6-phosphatase (G6pase) and phos-
phoenolpyruvate carboxykinase (Pepck), which encode key glu-
coneogenic enzymes, and the ?-subunit of ATP synthase
(Atp5b), cytochrome c (Cytc), and isocitrate dehydrogenase 3a
(Idh3a), which encode proteins involved in mitochondrial oxi-
dative phosphorylation and TCA cycle flux, were induced sig-
nificantly in FGF21-TG liver (Fig. 2A).
To determine whether acute administration of FGF21 is
sufficient to induce PGC-1? and its target genes, we performed
an FGF21 time course study. Pgc1? was induced in liver as early
as 15 min after FGF21 injection (Fig. 2C). G6pase was induced
at the 1-h time point, and Pepck and Cytc were induced at 4 h
postinjection (Fig. 2C). Thus, acute administration of FGF21 is
sufficient to induce PGC-1? and a subset of its target genes. The
high basal expression of G6pase at the 15-min time point may
reflect stress caused by injection.
To test whether the effects of FGF21 on gluconeogenic gene
expression require PGC-1?, WT and PGC-1?-KO mice were
injected with FGF21, and G6pase and Pepck mRNA levels were
measured in liver. FGF21 injection into fed WT mice increased
Pgc1?, G6pase, and Pepck mRNA to levels that mimicked those
observed during fasting (Fig. 2D). Importantly, FGF21 injection
did not induce G6pase and Pepck in PGC-1?-KO mice. In
agreement with results in ref. 4, G6pase and Pepck were induced
by fasting in PGC-1?-KO mice, indicating either redundant
mechanisms for induction of gluconeogenic gene expression or
a compensatory response in the knockout mice. Nevertheless,
these data demonstrate that the coordinate effects of FGF21 on
gluconeogenic gene expression require PGC-1?.
FGF21-KO Mice Have Metabolic Defects.To complement the FGF21
gain-of-function studies, we examined the consequences of elim-
inating FGF21 in mice. A null allele of the Fgf21 gene was
generated by introducing loxP sites upstream of exon 1 and
downstream of exon 3 through homologous recombination in
TCA Cycle Flux
Metabolic parameters of energy expenditure and lipid and glucose metabo-
perfused with nonrecirculating media, and absolute fluxes were determined
was analyzed in fed and fasted WT and FGF21-TG liver samples (n ? 6 per
SEM (a, P ? 0.05; b, P ? 0.01; c, P ? 0.005; d, P ? 0.001).
FGF21 regulates hepatic lipid and glucose metabolism. (A and B)
Relative mRNA level
Relative mRNA level
WT WT TG TG
WT KOWT KOWT KO
Fed vehicle Fed FGF21 Fasted vehicle
Relative mRNA level
gene expression in fed WT and FGF21-TG male mice (n ? 5 per group). (B)
Western blot analysis of PGC-1? protein in fed WT and FGF21-TG mouse liver.
(C) Hepatic metabolic gene expression after injection of vehicle or FGF21 into
fed WT mice for the indicated times (n ? 4 per group). (D) Groups of WT and
PGC-1?-KO mice (n ? 4 per group except for WT/fasted vehicle group, where
gene expression was analyzed by real-time QPCR. Data are presented as
mean ? SEM (a, P ? 0.05; b, P ? 0.01; c, P ? 0.005; d, P ? 0.001).
www.pnas.org?cgi?doi?10.1073?pnas.0904187106Potthoff et al.
embryonic stem cells (Fig. S2). Mice heterozygous for the
Fgf21loxPallele were bred with mice expressing a Meox-cre
transgene, which deletes in the germ line, to generate Fgf21?/?
animals. Fgf21?/?mice were then intercrossed to generate
Fgf21?/?(FGF21-KO) mice (Fig. 3A). FGF21-KO mice were
born at the expected Mendelian ratio and were viable. As
expected, Fgf21 transcripts were dramatically elevated in liver
during fasting in WT animals but were undetectable in
FGF21-KO mice by RT-PCR (Fig. 3B) or QPCR (Fig. 3C).
Commensurate with mRNA levels, plasma FGF21 was markedly
elevated by fasting in WT mice but absent in FGF21-KO mice
FGF21-KO and WT mice showed no differences in body
weight or plasma glucose, triglyceride, nonesterified fatty acid
(NEFA), insulin, and glucagon concentrations in the fed state
(Table 1). Plasma glucose levels, however, were significantly
reduced in fasted FGF21-KO mice. Plasma ketone levels were
significantly reduced in fed mice and trended lower in fasted
FGF21-KO mice (Table 1). FGF21-KO mice also showed trends
toward increased plasma NEFA and triglyceride concentrations
during fasting (Table 1), suggesting that FGF21-KO mice may
avoid overt hypoketosis during fasting by increasing lipid deliv-
ery to the liver. To further assess whether ketogenic potential is
impaired in the absence of FGF21, we compared plasma ketone
concentrations in WT and FGF21-KO mice after injection with
octanoate, a medium-chain fatty acid that rapidly enters mito-
chondria in a carnitine-independent manner and is converted to
ketones at a rate sensitive to the fasting state of the liver (30).
Octanoate administration to fasted mice caused the plasma
?-hydroxybutyrate concentration to double in WT mice com-
pared with only a 20% increase in FGF21-KO mice (Fig. 3E).
This result demonstrates that under conditions of identical
nutritional state and substrate availability ketogenesis is mark-
edly reduced in FGF21-KO mice.
The role of FGF21 in regulating hepatic gene expression was
investigated by using WT and FGF21-KO mice in either the fed
or fasted state. Under fasted conditions, Cel, Pnlip, Pnliprp2,
Pgc1?, G6pase, and Pepck mRNAs were induced in WT mice
(Fig. 4). With the exception of Pepck, induction of all of these
genes was significantly attenuated in liver of FGF21-KO mice.
The Pepck data indicate that either FGF21 is not involved in
regulating Pepck in response to fasting or that other mechanisms
compensate for FGF21’s absence. Lpl mRNA was also signifi-
cantly reduced in FGF21-KO liver in the fasted state (Fig. 4).
Overall, these loss-of-function data demonstrate that FGF21
plays an important role in inducing PGC-1? and other genes
involved in regulating carbohydrate and lipid metabolism during
Relative mRNA level
Plasma Ketone Bodies
+/+ –/+ –/–
of Fgf21?/?mice by genomic PCR. Primer triplex includes one set flanking the
5? loxP site and a third primer downstream of the 3? loxP site. (B) Global
deletion in the germ line by Meox-cre removes all three Fgf21 exons as
demonstrated by semiquantitative RT-PCR using RNA prepared from liver of
fed or 24-h-fasted WT or FGF21-KO mice (n ? 5 per group). Primers are as
indicated in Fig. S2. Cyclophilin served as a loading control. (C) FGF21 mRNA
(n ? 5 per group). (E) Rate of ketone body production was measured in
analysis of plasma ?-hydroxybutyrate concentrations. (C–E) Data are pre-
sented as mean ? SEM (a, P ? 0.05; b, P ? 0.01; c, P ? 0.005; d, P ? 0.001).
Generation and characterization of FGF21-KO mice. (A) Genotyping
Table 1. Body weight and plasma parameters in WT and FGF21-KO mice
WT FGF21-KOWT FGF21-KO
Body weight, g
21.2 ? 2.14
140 ? 8.94
35.9 ? 6.58
0.627 ? 0.063
0.467 ? 0.107
40.0 ? 7.96
140 ? 4.11
21.0 ? 2.56
136 ? 11.4
24.5 ? 8.23*
0.624 ? 0.060
0.440 ? 0.161
39.1 ? 6.10
135 ? 15.6
20.3 ? 2.20
76.4 ? 2.64
857 ? 93.1
3.07 ? 0.172
0.163 ? 0.089
102 ? 13.9
42.7 ? 6.17
19.8 ? 1.28
64.1 ? 1.90**
792 ? 125
3.96 ? 0.294
0.115 ? 0.056
105 ? 14.7
52.3 ? 14.1
*, P ? 0.05;**, P ? 0.01.
Fed FastFed Fast Fed Fast
Relative mRNA level
Fed FastFed Fast Fed Fast
Relative mRNA level
metabolism during fasting. Shown is real-time QPCR analysis of gene expres-
sion in WT or FGF21-KO liver under fed or fasted conditions (n ? 5 per group).
Data are presented as mean ? SEM. (a, P ? 0.05; b, P ? 0.01; c, P ? 0.005; d,
P ? 0.001).
FGF21 regulates genes involved in hepatic carbohydrate and lipid
Potthoff et al. PNAS ?
June 30, 2009 ?
vol. 106 ?
no. 26 ?
To determine the effect of eliminating FGF21 on the regu-
lation of hepatic glucose and fat oxidation under uniform
metabolic conditions, we measured flux through these pathways
in isolated liver perfused with equal concentrations of NEFA
and gluconeogenic substrates. In agreement with the octanoate
challenge data, fasted FGF21-KO liver had diminished oxygen
consumption, ?-oxidation, TCA cycle flux, and ketogenesis
compared with those of fasted WT liver (Fig. 5A). These fluxes
in FGF21-KO liver failed to respond to fasting. Remarkably,
there was also no induction of gluconeogenesis in FGF21-KO
liver in response to fasting (Fig. 5B). Neither glycogenolysis nor
liver glycogen concentrations were significantly different be-
tween WT and FGF21-KO liver under fed or fasted conditions
(Fig. 5B). The loss of fasting-induced gluconeogenesis in
FGF21-KO mice is consistent with the hypoglycemia that occurs
in these animals (Table 1) and the findings in the gain-of-
function perfusion experiments (Fig. 1B). Thus, FGF21 is crucial
for the liver’s adaptive metabolic response to prolonged fasting.
FGF21 was shown previously to be a ketogenic factor induced by
PPAR? in response to fasting (5, 6). Using both gain-of-function
actions of FGF21 extend well beyond fatty acid oxidation and
ketogenesis to include striking effects on TCA cycle flux and
carbohydrate metabolism. Total flux through the gluconeogenic
pathway was increased in fed FGF21-TG liver to levels compa-
rable to those seen in fasted WT liver. Remarkably, there was no
induction of gluconeogenesis in the fasted FGF21-KO liver.
Consistent with these findings, plasma glucose concentrations
were significantly reduced during fasting in FGF21-KO mice.
These results demonstrate that this hormone plays a crucial role
in mediating the liver’s adaptive response to nutritional
FGF21 is similar to the fasting hormone glucagon in that it
induces hepatic gluconeogenesis, fatty acid oxidation, and ke-
not promote glycogenolysis. Indeed, FGF21-TG mice accumu-
late significantly more hepatic glycogen than WT mice in the fed
state. Interestingly, in humans, circulating glucagon concentra-
tions spike after 3–5 days of fasting and then decline, whereas
plasma FGF21 levels increase only after a 7-day fast (31–33).
These findings raise the possibility that FGF21 maintains glu-
coneogenesis and ketogenesis in the context of prolonged fasting
and starvation, when glycogen stores are depleted and glucagon
levels have fallen. In this model, glucagon and FGF21 are not
redundant but rather sequential in their actions. Several prop-
erties of FGF21 are consistent with this hypothesis. First,
circulating FGF21 concentrations are induced in rodents and
humans only after prolonged fasting (5, 6, 33). Second, FGF21
under severe nutritional stress (6). Third, FGF21 causes growth
hormone resistance and blunts growth in mice, a phenomenon
that occurs during starvation (34). On the basis of these findings,
we propose that FGF21 is a bona fide starvation hormone rather
than a mediator of shorter-term fasting responses.
Our finding that FGF21 rapidly induces hepatic expression of
PGC-1?, a prominent transcriptional regulator of metabolism
(19–21, 25), suggests a mechanism for its metabolic actions.
FGF21 may also modulate PGC-1? activity via posttranslational
modifications. Notably, induction of PGC-1? and its down-
stream target, G6Pase, was significantly attenuated in liver of
FGF21-KO mice after a 24-h fast. Moreover, the inductive effect
of FGF21 on gluconeogenic gene expression was virtually elim-
inated in PGC-1?-KO mice. NMR isotopomer analyses show
that PGC-1?-KO liver has reductions in fatty acid oxidation,
ketogenesis, TCA flux, and gluconeogenesis that mirror those in
the FGF21-KO liver (4). Because glucagon and its downstream
effector, cAMP, induce PGC-1? and gluconeogenic gene ex-
pression during the early stages of the fasting response (24, 35),
we propose that FGF21 maintains the expression of these genes
during prolonged fasting and starvation.
How does FGF21 induce PGC-1? in liver? The rapidity with
which FGF21 induces PGC-1? and the fact that FGF21 acts
through receptors with tyrosine kinase activity suggest that
FGF21 may affect the phosphorylation of transcription factors
that bind to the Pgc1? promoter. However, an analysis of
candidate transcription factors that regulate the Pgc1? promoter
did not reveal changes in the phosphorylation of FOXO1,
CREB, or TORC2 in livers of mice injected with FGF21 (Fig.
S3A). Moreover, we have been unable to recapitulate FGF21-
mediated induction of Pgc1? in either isolated, perfused mouse
liver or primary cultures of rat or mouse hepatocytes, precluding
the use of these models to elucidate how PGC-1? is induced
(Fig. S3 B–D). These negative findings raise the interesting
possibility that FGF21 might induce Pgc1? through an indirect
mechanism involving the central nervous system.
FGF21 functions as a potent insulin sensitizer in various
animal models of insulin resistance and diabetes (15, 17, 36, 37).
Our data suggest that FGF21 does not improve glycemic control
by suppressing hepatic gluconeogenesis per se. Rather, they
indicate that FGF21 improves hepatic insulin action indirectly by
stimulating hepatic fatty acid disposal. Decreasing hepatic fatty
acid content has well-documented effects on liver insulin sensi-
tivity in a variety of models and indirectly improves the respon-
siveness of hepatic glucose production, via either glycogenolysis
or gluconeogenesis (38, 39). In this respect, the effects of FGF21
are analogous to those of PPAR? agonists, which improve
hepatic insulin action and overall glycemic control in diabetic
rodents despite increasing hepatic gluconeogenesis (40, 41).
Indeed, as previously shown, FGF21 expression is regulated
directly by PPAR?, which may explain many of the therapeutic
effects of PPAR? agonists. Because all of our studies were done
in lean mice, additional experiments will have to be performed
to determine whether FGF21 has similar effects on metabolism
in insulin-resistant animals.
In summary, we demonstrate that the fasting-induced hor-
mone FGF21 has pronounced effects on carbohydrate and fatty
acid metabolism in liver. Mice lacking FGF21 have impaired
hepatic gluconeogenesis and ketogenesis. We propose that
FGF21 acts subsequent to glucagon during nutritional depriva-
TCA Cycle Flux
during fasting. (A and B) Metabolic parameters of energy expenditure and
lipid and glucose metabolism were determined by NMR in livers from fed and
fasted 16- to 20-week-old WT and FGF21-KO male mice (n ? 6 per group).
Livers were perfused with nonrecirculating media, and absolute fluxes were
determined from the NMR data and rate of glucose production. Liver glyco-
as mean ? SEM (a, P ? 0.05; b, P ? 0.01; c, P ? 0.005; d, P ? 0.001).
FGF21 is required for inducing hepatic lipid and glucose metabolism
www.pnas.org?cgi?doi?10.1073?pnas.0904187106Potthoff et al.
tion to elicit and coordinate diverse aspects of the adaptive
Animal Experiments. FGF21-TG, PGC-1?-KO, and Meox-cre mice were de-
scribed previously (6, 28, 42, 43). Mice were fed standard chow containing 4%
fat ad libitum. All animal experiments were approved by the Institutional
Animal Care and Research Advisory Committee of the University of Texas
Southwestern Medical Center.
Generation of FGF21-KO Mice. The FGF21 targeting vector was constructed by
using a conditional KO vector containing a neomycin-resistance gene flanked
a 3.8-kb long arm harboring the 3? end of the Fgf21 gene were generated by
high-fidelity PCR amplification (Expand High-Fidelity Long Template; Roche).
The FGF21 targeting vector was linearized and injected into ES cells. ES cell
clones were isolated and analyzed for homologous recombination by long-
FGF21 were injected into 3.5-day C57BL/6 blastocyts, and the resulting chime-
by crossing Fgf21loxP/?mice with Meox-cre mice (C57BL/6) to generate
were confirmed by genotyping. Heterozygote breeding was performed for
two additional generations, and WT (FGF21?/?) and homozygous (FGF21?/?)
breeding lines were subsequently maintained.
FGF Injection Experiments. Human FGF21 (residues 29–209) with a hexahisti-
dine tag on the amino terminus was expressed in Escherichia coli and purified
by sequential Ni-chelating and size-exclusion chromatography. FGF21 was
PGC-1?-KO mice, mice receiving FGF21 were fed ad libitum and were injected
2 h after the second injection.
TG, FGF21-KO, and WT littermates were isolated and perfused for 60 min in a
nonrecirculating fashion at 8 mL/min with a Krebs–Henseleit-based perfusion
medium containing 1.5 mM lactate, 0.15 mM pyruvate, 0.25 mM glycerol, 0.4
mM free fatty acid (algal mix bound to 3% albumin), 2% deuterated water,
and 0.1 mM [U-13C3]propionate as described previously (3, 4, 44, 45).
Immunoblot Analysis. Immunoblotting was performed using an anti-PGC-1?
antibody (Cell Signaling Technology) and ?-actin antibody (Sigma) as de-
scribed in ref. 16.
RT-PCR and QPCR Analysis. Total RNA was extracted from WT and FGF21-KO
mouse livers with Stat 60 reagent (IsoTex Diagnostics). Four micrograms of
RNA from each sample was then used to generate cDNA. RT-PCR for the
deleted exons of the FGF21 gene was performed using the primer pairs: Exon
1 forward 5?-GCCTGAGCCCCAGTCTGAACCTGACCC-3?, Exon 2 reverse 5?-
CCATAGAGAGCTCCATCTGGCTGTTGGC-3?, Exon 2 forward 5?-GCCAACAGC-
CAGATGGAGCTCTCTATGG-3?, Exon 3 reverse 5?-GAAGAGTCAGGACGCAT-
QPCR was performed using SYBR green as described in ref. 46. The primer
sequences used for gene expression analyses are as follows: Pgc-1? FWD
5?-AGACAAATGTGCTTCCAAAAAGAA-3?, REV 5?-GAAGAGATAAAGTTGTTG-
GTTTGGC-3?; G6pase FWD 5?-GTGGCAGTGGTCGGAGACT-3?, REV 5?-
ACGGGCGTTGTCCAAAC-3?; Pepck FWD 5?-CACCATCACCTCCTGGAAGA-3?,
REV 5?-GGGTGCAGAATCTCGAGTTG-3?; Lpl FWD 5?-GGCCAGATTCATCAACT-
GGAT-3?, REV 5?-GGCTGTCTCCCAAGAGATGGA-3?, Atp5b FWD 5?-AGGTGGC-
CCAGCATTTG-3?, REV 5?-GCCTTCAGTGCCATCCATAG-3?; Cytc FWD 5?-
GAAAAGGGAGGCAAGCATAAG-3?, REV 5?-TGTCTTCCGCCCGAACA-3?; Idh3a
FWD 5?-TCGTCACCATCCGAGAGAAC-3?, REV 5?-GCACAACCCCATCAACGAT-
3?; Cyclophilin FWD 5?-GGAGATGGCACAGGAGGAA-3?, REV 5?-GCCCGTAGT-
Metabolic Parameter Measurements. Plasma glucose levels were measured by
using the glucose autokit (Wako Chemicals). Plasma triglyceride concentra-
tions were measured by using an L-type TG H triglyceride kit (Wako Chemi-
cals). Plasma NEFAs were measured by using a NEFA C kit (Wako Chemicals).
Plasma ?-hydroxybutyrate concentrations were measured by using a D-3-
hydroxybutyric acid kit (Wako Chemicals). Plasma glucagon levels were mea-
sured by using a glucagon ELISA kit (Wako Chemicals). Plasma insulin levels
were measured by using the Ultra Sensitive Mouse Insulin ELISA kit (Crystal
Chem). Plasma FGF21 protein concentration was determined by using the
FGF21 RIA kit (Phoenix Pharmaceuticals). All measurements were performed
following the manufacturer’s instructions. Glycogen levels were determined
as described in ref. 4.
blood were withdrawn for analysis of ?-hydroxybutyrate concentrations.
Statistical Analyses. Statistical analyses were performed as described in
refs. 3 and 6.
ACKNOWLEDGMENTS. We thank D. Kelly (Burnham Institute for Medical
Research, Orlando, FL) for PGC-1?-KO mice and discussing the manuscript, M.
Tallquist (UT Southwestern, Dallas, TX) for Meox1-cre mice, M. Montminy
(Salk Institute, La Jolla, CA) for CREB and TORC2 antibodies, L. Peng for
Kuro-o and members of the Mangelsdorf/Kliewer lab for helpful discussions.
This work was supported by National Institutes of Health Grants DK067158,
P20RR20691, 1RL1GM084436–01 (S.A.K. and D.J.M.); U19DK62434 (D.J.M.);
DK078184, RR02584, and DK076269 (S.C.B.); and DE13686 (M.M.); the Robert
A. Welch Foundation (D.J.M. and S.A.K.); and the Howard Hughes Medical
Institute. D.J.M. is an investigator of the Howard Hughes Medical Institute.
1. Hakimi P, et al. (2005) Phosphoenolpyruvate carboxykinase and the critical role of
cataplerosis in the control of hepatic metabolism. Nutr Metab (Lond) 2:33.
2. Leone TC, Weinheimer CJ, Kelly DP (1999) A critical role for the peroxisome prolifera-
tor-activated receptor ? (PPAR?) in the cellular fasting response: The PPAR?-null
mouse as a model of fatty acid oxidation disorders. Proc Natl Acad Sci USA 96:7473–
3. Burgess SC, et al. (2007) Cytosolic phosphoenolpyruvate carboxykinase does not solely
control the rate of hepatic gluconeogenesis in the intact mouse liver. Cell Metab
acid cycle flux in peroxisome proliferator-activated receptor ? coactivator-1? (PGC-
1?)-deficient mice. J Biol Chem 281:19000–19008.
5. Badman MK, et al. (2007) Hepatic fibroblast growth factor 21 is regulated by PPAR? and
is a key mediator of hepatic lipid metabolism in ketotic states. Cell Metab 5:426–437.
6. Inagaki T, et al. (2007) Endocrine regulation of the fasting response by PPAR?-
mediated induction of fibroblast growth factor 21. Cell Metab 5:415–425.
7. Lundasen T, et al. (2007) PPAR? is a key regulator of hepatic FGF21. Biochem Biophys
Res Commun 360:437–440.
8. Goetz R, et al. (2007) Molecular insights into the Klotho-dependent, endocrine mode of
action of fibroblast growth factor 19 subfamily members. Mol Cell Biol 27:3417–3428.
9. Kharitonenkov A, et al. (2008) FGF-21/FGF-21 receptor interaction and activation is
determined by ?Klotho. J Cell Physiol 215:1–7.
10. Kurosu H, et al. (2007) Tissue-specific expression of ?Klotho and fibroblast growth
11. Lin BC, Wang M, Blackmore C, Desnoyers LR (2007) Liver-specific activities of FGF19
require Klotho beta. J Biol Chem 282:27277–27284.
12. Ogawa Y, et al. (2007) ?Klotho is required for metabolic activity of fibroblast growth
factor 21. Proc Natl Acad Sci USA 104:7432–7437.
13. Suzuki M, et al. (2008) ?Klotho is required for fibroblast growth factor (FGF) 21
signaling through FGF receptor (FGFR) 1c and FGFR3c. Mol Endocrinol 22:1006–
J Biol Chem 282:29069–29072.
15. Kharitonenkov A, et al. (2005) FGF-21 as a novel metabolic regulator. J Clin Invest
16. Inagaki T, et al. (2008) Inhibition of growth hormone signaling by the fasting-induced
hormone FGF21. Cell Metab 8:77–83.
fibroblast growth factor-21. Endocrinology 148:774–781.
18. Handschin C, Spiegelman BM (2006) Peroxisome proliferator-activated receptor ?
coactivator 1 coactivators, energy homeostasis, and metabolism. Endocr Rev 27:728–
19. Puigserver P, Spiegelman BM (2003) Peroxisome proliferator-activated receptor-?
coactivator 1? (PGC-1?): Transcriptional coactivator and metabolic regulator. Endocr
20. Lin J, Handschin C, Spiegelman BM (2005) Metabolic control through the PGC-1 family
of transcription coactivators. Cell Metab 1:361–370.
21. Finck BN, Kelly DP (2006) PGC-1 coactivators: Inducible regulators of energy metabo-
lism in health and disease. J Clin Invest 116:615–622.
Potthoff et al.PNAS ?
June 30, 2009 ?
vol. 106 ?
no. 26 ?
22. Puigserver P, et al. (2003) Insulin-regulated hepatic gluconeogenesis through FOXO1- Download full-text
PGC-1? interaction. Nature 423:550–555.
23. Rhee J, et al. (2003) Regulation of hepatic fasting response by PPAR? coactivator-1?
(PGC-1): Requirement for hepatocyte nuclear factor 4? in gluconeogenesis. Proc Natl
Acad Sci USA 100:4012–4017.
24. Yoon JC, et al. (2001) Control of hepatic gluconeogenesis through the transcriptional
coactivator PGC-1. Nature 413:131–138.
25. Handschin C, et al. (2005) Nutritional regulation of hepatic heme biosynthesis and
porphyria through PGC-1?. Cell 122:505–515.
in PGC-1? null mice. Cell 119:121–135.
27. Koo SH, et al. (2004) PGC-1 promotes insulin resistance in liver through PPAR-?-
dependent induction of TRB-3. Nat Med 10:530–534.
28. Leone TC, et al. (2005) PGC-1? deficiency causes multi-system energy metabolic de-
29. Cherrington AD, Edgerton D, Sindelar DK (1998) The direct and indirect effects of
insulin on hepatic glucose production in vivo. Diabetologia 41:987–996.
30. McGarry JD, Foster DW (1971) The regulation of ketogenesis from octanoic acid.
The role of the tricarboxylic acid cycle and fatty acid synthesis. J Biol Chem
31. Fisher M, Sherwin RS, Hendler R, Felig P (1976) Kinetics of glucagon in man: Effects of
starvation. Proc Natl Acad Sci USA 73:1735–1739.
32. Marliss EB, et al. (1970) Glucagon levels and metabolic effects in fasting man. J Clin
33. Galman C, et al. (2008) The circulating metabolic regulator FGF21 is induced by
prolonged fasting and PPAR? activation in man. Cell Metab 8:169–174.
to regulate bile acid homeostasis. Cell Metab 2:217–225.
35. Longuet C, et al. (2008) The glucagon receptor is required for the adaptive metabolic
response to fasting. Cell Metab 8:359–371.
36. Fu L, et al. (2004) Fibroblast growth factor 19 increases metabolic rate and reverses
dietary and leptin-deficient diabetes. Endocrinology 145:2594–2603.
37. Tomlinson E, et al. (2002) Transgenic mice expressing human fibroblast growth fac-
tor-19 display increased metabolic rate and decreased adiposity. Endocrinology
38. Savage DB, Petersen KF, Shulman GI (2007) Disordered lipid metabolism and the
pathogenesis of insulin resistance. Physiol Rev 87:507–520.
39. Nandi A, Kitamura Y, Kahn CR, Accili D (2004) Mouse models of insulin resistance.
Physiol Rev 84:623–647.
40. Ye JM, et al. (2001) Peroxisome proliferator-activated receptor (PPAR)-? activation
lowers muscle lipids and improves insulin sensitivity in high fat-fed rats: Comparison
with PPAR-? activation. Diabetes 50:411–417.
41. Satapati S, et al. (2008) Partial resistance to peroxisome proliferator-activated recep-
tor-? agonists in ZDF rats is associated with defective hepatic mitochondrial metabo-
lism. Diabetes 57:2012–2021.
42. Tallquist MD, Soriano P (2000) Epiblast-restricted Cre expression in MORE mice: A tool
to distinguish embryonic vs. extra-embryonic gene function. Genesis 26:113–115.
43. Wright TJ, et al. (2004) Mouse FGF15 is the ortholog of human and chick FGF19, but is
not uniquely required for otic induction. Dev Biol 269:264–275.
44. Burgess SC, et al. (2005) Effect of murine strain on metabolic pathways of glucose
production after brief or prolonged fasting. Am J Physiol Endocrinol Metab
45. Burgess SC, et al. (2003) Analysis of gluconeogenic pathways in vivo by distribution of
2H in plasma glucose: Comparison of nuclear magnetic resonance and mass spectrom-
etry. Anal Biochem 318:321–324.
46. Bookout AL, Mangelsdorf DJ (2003) A quantitative real-time PCR protocol for analysis
of nuclear receptor signaling pathways. Nucl Recept Signal 1:e012.
www.pnas.org?cgi?doi?10.1073?pnas.0904187106Potthoff et al.