HHS Public Access

Author manuscript
Cell Metab. Author manuscript; available in PMC 2017 February 09.
Author Manuscript

Published in final edited form as:

Cell Metab. 2016 February 9; 23(2): 344–349. doi:10.1016/j.cmet.2015.12.008.

FGF21 Regulates Sweet and Alcohol Preference

Saswata Talukdar1,9,*, Bryn M. Owen2,10,*, Parkyong Song2, Genaro Hernandez2, Yuan
Zhang2, Yingjiang Zhou1,9, William T. Scott2, Bhavna Paratala3, Tod Turner1, Andrew
Smith3, Barbara Bernardo3, Christian P. Müller7,8, Hao Tang4, David J. Mangelsdorf2,5,†,
Bryan Goodwin1, and Steven A. Kliewer2,6,†
1Cardiovascular
and Metabolic Diseases Research Unit, Pfizer Worldwide Research and
Development, Cambridge, MA 02139
Author Manuscript

2Department of Pharmacology, University of Texas Southwestern Medical Center, Dallas, TX,

USA 75390
3DrugSafety Research and Development, Pfizer Worldwide Research and Development, Groton,
CT 06340
4Department of Clinical Science, University of Texas Southwestern Medical Center, Dallas, TX,
USA 75390
5HowardHughes Medical Institute, University of Texas Southwestern Medical Center, Dallas, TX,
USA 75390
6Department of Molecular Biology, University of Texas Southwestern Medical Center, Dallas, TX,
USA 75390
Author Manuscript

7Department of Psychiatry and Psychotherapy, University Hospital, Friedrich-Alexander-

University Erlangen-Nuremberg, Schwabachanlage 6, 91054 Erlangen, Germany
8MRC Social, Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry,
King’s College London, De Crespigny Park, London SE5 8AF, UK

Summary
Fibroblast growth factor 21 (FGF21) is a hormone induced by various metabolic stresses,
including ketogenic and high carbohydrate diets, that regulates energy homeostasis. In humans,
SNPs in and around the FGF21 gene have been associated with macronutrient preference,
including carbohydrate, fat and protein intake. Here we show that FGF21 administration markedly
Author Manuscript

†
Corresponding authors: [email protected]; [email protected].
9Current address, Merck, Boston, MA
10Current address, Section of Investigative Medicine, Imperial College, London
*equal contributions
Author Contributions
S.T., B.M.O., P.S., G.H., and Y. Zhou designed, performed and analyzed experiments; Y. Zhang, and W.T.S. performed and analyzed
experiments; B.P., T.T., A.S., and B.B. performed experiments; H.T. analyzed experiments; C.P.M. designed and analyzed
experiments; B.G., S.A.K., and D.J.M. designed, supervised, and analyzed experiments and wrote the paper. P.S. and G.H. contributed
equally. All authors commented and approved the paper.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our
customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of
the resulting proof before it is published in its final citable form. Please note that during the production process errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
 Talukdar et al. Page 2

reduces sweet and alcohol preference in mice, and sweet preference in cynomolgus monkeys. In
Author Manuscript

mice, these effects require the FGF21 co-receptor β-Klotho in the central nervous system and
correlate with reductions in dopamine concentrations in the nucleus accumbens. Since analogs of
FGF21 are currently undergoing clinical evaluation for the treatment of obesity and type 2
diabetes, our findings raise the possibility that FGF21 administration could affect nutrient
preference and other reward behaviors in humans.

Graphical Abstract
Author Manuscript
Author Manuscript

Introduction
FGF21 is induced in liver in response to fasting and other forms of metabolic stress
including ketogenic and high carbohydrate diets (Badman et al., 2007; Dushay et al., 2015;
Inagaki et al., 2007; Sanchez et al., 2009; Zhao et al., 2015). FGF21 in turn regulates diverse
aspects of energy homeostasis, including hepatic fatty acid oxidation and ketogenesis,
circadian behavior, growth and female reproduction (Owen et al., 2015). Pharmacologically,
FGF21 causes weight loss and improves insulin sensitivity in obese mice, monkeys and
humans (Gimeno and Moller, 2014). Long-acting analogs of FGF21 are currently in clinical
trials for the treatment of obesity and type 2 diabetes.

FGF21 acts through a cell surface receptor composed of a conventional FGF receptor in
complex with β-Klotho, a single-pass transmembrane protein (Owen et al., 2015). FGF21
Author Manuscript

crosses the blood-brain barrier (Hsuchou et al., 2007) and exerts many of its actions,
including its effects on growth, female reproduction and weight loss, by acting on its
cognate receptor in the CNS (Bookout et al., 2013; Douris et al., 2015; Liang et al., 2014;
Owen et al., 2013; Owen et al., 2014; Sarruf et al., 2010). Among its central actions, FGF21
induces corticotropin-releasing factor and suppresses arginine vasopressin expression in the
hypothalamus (Bookout et al., 2013; Liang et al., 2014; Owen et al., 2013; Owen et al.,
2014).

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 3

In humans, SNPs in and around the FGF21 gene are associated with changes in
Author Manuscript

macronutrient preference, including increases in carbohydrate consumption and decreases in

fat and protein intake (Chu et al., 2013; Tanaka et al., 2013). These findings raise the
possibility of additional effects of FGF21 on the brain. In this report, we examine the effect
of FGF21 on sweet and alcohol preference in mice and monkeys.

Results and Discussion

Since FGF21 is induced by carbohydrates in rodents and humans (Dushay et al., 2015;
Sanchez et al., 2009), and SNPs in the FGF21 gene are associated with carbohydrate intake
in humans (Chu et al., 2013; Tanaka et al., 2013), we investigated whether chronic FGF21
exposure affects sweet preference. Two-bottle preference assays with water and either 3%
sucrose or 0.2% saccharin (Krishnan et al., 2007; Tordoff and Bachmanov, 2003) were
performed using wild-type and Fgf21-transgenic (Tg) mice expressing supraphysiological
Author Manuscript

concentrations of FGF21 (Inagaki et al., 2007). Saccharin was included to eliminate the
potentially confounding effect of caloric content. As expected, wild-type mice showed a
strong preference for drinking water sweetened with either sucrose or saccharin (Figure 1A
and B, Table S1). Notably, the Tg(Fgf21) mice had significant decreases in both sucrose and
saccharin preference (Figure 1A and B).

To determine whether FGF21 acts on the CNS to regulate sweet preference, we

administered recombinant FGF21 or vehicle by osmotic minipump to groups of control mice
with floxed β-Klotho alleles (Klbfl/fl) or mice specifically lacking β-Klotho in the CNS
(KlbCamk2a), and two-bottle saccharin preference tests were performed. FGF21 strongly
suppressed saccharin preference in Klbfl/fl mice but had no effect in KlbCamk2a mice (Figure
1C, Table S2). As reported (Camporez et al., 2013), administration of recombinant FGF21
Author Manuscript

increased total fluid intake, and this effect required β-Klotho in the CNS (Table S2). In
contrast, total fluid intake was unchanged in Tg(Fgf21) compared to control mice (Table S1
and S2). The reason for this difference between transgenic FGF21 overexpression and
recombinant FGF21 administration on fluid intake is not known. Nevertheless, FGF21
decreased sweet preference in both contexts.

In additional two-bottle preference tests, FGF21 had no effect on preference for 1%

sunflower oil (data not shown) or quinine (Figure 1D), indicating that FGF21 does not affect
responses to fatty acids or bitter taste. FGF21 administration also had no effect on either tail-
suspension or forced-swim tests, both standard measures of behavioral despair (Figure S1A
and B). We conclude that FGF21 acts directly on the brain to regulate sweet preference
without causing despair.
Author Manuscript

To determine whether FGF21 also affects sweet preference in primates, we analyzed

saccharin preference in obese cynomolgus monkeys administered PF-05231023, a long-
acting FGF21 analog consisting of two molecules of modified FGF21 linked by an antibody
scaffold (Dong et al., 2015; Giragossian et al., 2015; Weng et al., 2015). We first tested this
analog in mice. PF-05231023 administration decreased saccharin preference to a similar
degree as native FGF21, with maximal efficacy observed 3–5 days after dosing (Figure 2A).
For the monkey study, PF-05231023 or vehicle was administered on days 1, 4 and 7 of the

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 4

three-week experiment. Notably, PF-05231023 administration significantly decreased

Author Manuscript

saccharin preference in the monkeys (Figure 2B). The effect on saccharin intake in the
monkeys was striking even within one day of receiving a single dose of the FGF21 analog,
and the effect was sustained for several days after receiving the last dose. Thus, FGF21 also
affects sweet preference in primates.

The neurotransmitter dopamine has a central role in regulating reward behavior, including
sucrose and saccharin preference (Fernstrom et al., 2012). To examine whether FGF21
affects dopamine signaling, including the mesolimbic pathway, we first measured β-Klotho
expression in the ventral tegmental area (VTA), nucleus accumbens (NAc), medial
prefrontal cortex (PFC) and caudate putamen (CP) of Klb+/− and Klb−/− mice, with the latter
mice serving as a negative control. As expected (Bookout et al., 2013; Liang et al., 2014),
Klb mRNA was detected in the suprachiasmatic nucleus/paraventricular (SCN/PVN)
nucleus region of the hypothalamus in Klb+/− mice by qPCR (Figure 3A). Klb mRNA was
Author Manuscript

also detected in VTA and NAc in Klb+/− mice, albeit at relatively low levels, but not the
PFC or CP (Figure 3A). As expected, Klb mRNA was not detected in any of the regions in
the Klb−/− control mice (Figure 3A). Consistent with the qPCR data, Klb mRNA was
detected by in situ hybridization in SCN and a small subset of cells in the VTA and NAc
(Figure S2A). Expression of FGF receptor 1, which partners with β-Klotho to form the
FGF21 receptor, was detected by qPCR in all of these brain regions (Figure S2B).

The FGF21 receptor expression data led us to examine whether FGF21 affects the levels of
dopamine and its metabolites in NAc, which coordinates reward behaviors. Notably, FGF21
administration for two weeks significantly decreased dopamine, 3,4-dihydroxyphenylacetic
acid (DOPAC), homovanillic acid (HVA) and 3-methoxytyramine (3-MT) concentrations
(Figure 3B). FGF21 administration did not decrease dopamine, DOPAC, HVA or 3-MT
Author Manuscript

concentrations in the CP (Figure S2C). FGF21 administration also caused changes in the
expression of dopamine-related genes, including an increase in the dopamine transporter in
the NAc and CP and a decrease in catechol-O-methyl transferase in the VTA and CP (Figure
3C and D). FGF21 had little or no effect on the mRNA levels of tyrosine hydroxylase or
dopamine receptor 1 in these brain regions (Figure S2D and E) nor did it change the levels
of total and phosphorylated tyrosine hydroxylase in VTA (Figure S2F). Taken together,
these data suggest that FGF21 may affect sweet preference via effects on dopamine
signaling. However, additional experiments will be required to confirm this possibility and
to determine the mechanism whereby FGF21 reduces dopamine concentrations.

Since dopamine signaling impacts ethanol drinking behavior (Gonzales et al., 2004), we
examined whether FGF21 also affects alcohol preference. Groups of wild-type and
Author Manuscript

Tg(Fgf21) mice were exposed stepwise to increasing concentrations of ethanol in a two-

bottle preference assay. Tg(Fgf21) mice had a decreased ethanol preference ratio at the 4%,
8%, 12% and 16% ethanol concentrations (Figure 4A, Table S3). In an ethanol
bioavailability test, there was no difference between wild-type and Tg(Fgf21) mice in
plasma ethanol concentrations at 1 and 3 hours after ethanol administration (Figure 4B).
Thus, FGF21 suppresses ethanol preference without affecting its bioavailability.

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 5

In summary, we show that FGF21 regulates sweet and alcohol preference in mice, and sweet
Author Manuscript

preference in monkeys. Since circulating levels of FGF21 increase in response to

carbohydrate consumption in rodents and humans (Dushay et al., 2015; Sanchez et al., 2009;
Zhao et al., 2015) and alcohol consumption in rodents (Zhao et al., 2015), this may represent
a feed forward regulatory pathway for limiting consumption. In mice, the effects on sweet
and alcohol preference correlate with reductions in dopamine concentrations in the NAc,
which coordinates reward behavior. These results suggest a mechanistic basis for the
association between SNPs in and around the FGF21 gene with macronutrient preference in
humans (Chu et al., 2013; Tanaka et al., 2013). Moreover, since FGF21 is currently in
clinical trials for treating obesity and type 2 diabetes, these findings suggest that additional
studies are warranted to assess the effects of FGF21 on sweet and alcohol preference and
other reward behavior in humans.
Author Manuscript

Methods
Mouse experiments with FGF21
All mouse experiments involving native FGF21 were approved by the Internal Animal Care
and Use Committee of the University of Texas Southwestern Medical Center. Wild-type and
Tg(Fgf21) mice were on a C57BL/6J background. Klbfl/fl and KlbCamk2a mice were on a
mixed C57BL/6J;129/Sv background as described (Bookout et al., 2013). All experiments
were performed with 2- to 4-month-old male mice. Tg(Fgf21) and Klbfl/fl mouse
experiments were performed with littermates. Mice were housed on a standard 12-hour light/
dark cycle and had free access to chow. Recombinant human FGF21 protein was provided
by Novo Nordisk and administered by subcutaneous osmotic minipumps (Alzet) at a dose of
1 mg/kg/day. Mice were allowed to recover from minipump surgery for 1 week prior to
preference tests. Mice were single caged following minipump surgery, which was conducted
Author Manuscript

under isoflurane anaesthesia and 24 hour buprenorphine analgesia.

Two-bottle preference assays

For the two-bottle sucrose and saccharin preference assays, mice were acclimated to cages
with 2 bottles of just water for 4 d. Mice were then given access to bottles with water and
water containing 3% sucrose or 0.2% saccharin (w/v). For the quinine preference assay,
mice were given access to water and water containing 2 mg/dl quinine. In each case,
consumption was measured daily for at least 3 days. For ethanol preference assays, mice
were given access to two bottles, one containing water and the other containing either 2, 4,
8, 12 or 16% ethanol (v/v) in water. The same mice were exposed to an ascending
concentration of each ethanol concentration for 5 days. The position of the two bottles was
changed every two days to exclude position effects. Water and ethanol-containing water
Author Manuscript

intake were measured each day.

Mouse immobility assays

Tail-suspension and forced-swim tests were performed as described (Can et al., 2012a; Can
et al., 2012b) on mice administered either FGF21 or vehicle by osmotic minipump for 7–14
days. In both tests, the experimenter was blinded to the treatment group. For the forced-
swim test, mice were placed in cylindrical tanks (20 cm in diameter) filled with water (25 ±

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 6

2°C). The cylinder was filled to a depth of 12 cm to prevent the mice from using their tails
Author Manuscript

to support themselves in the water. In both experiments, the cumulative time spent immobile
was recorded over the course of a six minute experiment.

Ethanol clearance assays

Mice were intraperitoneally injected with ethanol (4 g/kg) in saline, and tail vein blood was
drawn at regular intervals. Plasma ethanol levels were measured using the EnzyChrom™
Ethanol Assay Kit (BioAssay Systems).

Microdissection of brain regions

Mouse brains were extracted from the skull and kept under dry ice vapor for all dissections.
Coronal sections (1mm thick) were cut using a brain-slicing matrix (Braintree Scientific).
Medial prefrontal cortex, whole NAC (shell and core), the hypothalamic suprachiasmatic
Author Manuscript

and paraventricular region, CP (striatum) and VTA were identified by gross architectural
landmarks (Paxinos and Franklin, 2004). Medial prefrontal cortex and the hypothalamic
suprachiasmatic and paraventricular region were dissected using a 14-gauge tissue punch.
NAC, CPA and the VTA were dissected using a 16-gauge tissue punch. Tissue was
homogenized by passage through a 26½-gauge syringe in RNA-STAT60.

Quantitative PCR analysis

Total RNA was isolated from tissue using RNA-STAT60 reagent and RNA was reverse-
transcribed into cDNA (Invitrogen). Gene expression was measured with an Applied
Biosystems 7900HT Sequence Detection System using the ΔΔCT assay and normalized to
GAPDH.
Author Manuscript

In situ hybridization analysis

Brains were dissected from male C57BL/6J mice, embedded in OCT compound (Sakura)
and flash frozen in cooled isopentane. Coronal sections (14–16 μm) were cut using a
cryostat (Leica) followed by fixation for 15 min at 4°C with 10% neutral buffered formalin.
In situ hybridization was performed using the RNAscope 2.5 brown chromogenic assay
pretreatment and detection kits (Advanced Cell Diagnostics). Probes for cyclophilin B
(positive control), dapB (negative control) and Klb were purchased from Advanced Cell
Diagnostics. Hybridized sections were counterstained with hematoxylin, dehydrated, cleared
and mounted with Ecomount (Biocare Medical). Images were taken using a Zeiss Axioscan
Z1 at 40X magnification. The signal from the mRNA was highlighted using the color
threshold function in Image J.
Author Manuscript

Western blot analysis

Western blot analysis was performed using antibodies for total (Cell Signaling, #2792) and
Ser40 phosphorylated tyrosine hydroxylase (AbCam, #51206). Data were acquired and
quantified using an ImageQuant LAS 4000 and Multi Gauge v3.1 software (Fujifilm).

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 7

Dopamine measurements
Author Manuscript

Dopamine and its metabolites were measured by HPLC by the Vanderbilt Neurochemistry
Core.

Mouse and monkey experiments with PF-05231023

All animal care and experimental procedures for studies involving PF-05231023 were
conducted in compliance with the US Animal Welfare Act and the ILAR Guide for the Care
and Use of Laboratory Animals, 1996. The procedures used in these studies were reviewed
and approved by the Pfizer Institutional Animal Care and Use Committee (AUP#
GTN-2013-00793).

Male C57BL/6J mice fed a 60% high fat diet (Research Diets D12492i) for 12–14 weeks
were used for the saccharin studies. Mice were singly housed with access to regular water
Author Manuscript

and 0.1% saccharin (S6047, Sigma-Aldrich, St. Louis, MO) water throughout the study.
Water intake was monitored using a BioDaq system (Research Diets Inc., New Brunswick,
NJ). The position of sweetened and unsweetened water bottles was switched each day to
eliminate positional preference. Mice were acclimatized to the cage for 10–14 days, after
which they were stratified based on body weight and baseline saccharin water intake.
PF-05231023 (Huang et al., 2013; Weng et al., 2015) or vehicle were administered
subcutaneously twice per week for a total of 3 doses at 10 mg/kg.

Male Macaca fascicularis (cynomolgus monkeys) from Charles River Labs were single-
housed under a 12-hour light/dark cycle. Monkeys were between 7–24 years of age and
weighed between 7–12 kg. Monkeys were provided standard lab diet 5K91 (LabDiet, St.
Louis, MO) supplemented once a day with fruits and peanuts. The two-bottle saccharin
preference assay was performed as described (Tordoff and Bachmanov, 2003). The monkeys
Author Manuscript

underwent a training period during which the standard water system was shut off and 2
bottles, one with water and the other with water containing 0.2% saccharin, were attached
side-by-side to the cage. To determine the amount water consumed, bottles were weighed
before and after filling, with the difference in weight equated to the volume consumed. The
position of sweetened and unsweetened water bottles was switched each day to eliminate
positional preference. Several animals in the original cohort were eliminated from the study
due to water bottle damage, failure to adapt to the water bottle or if one of the allocation
endpoints represented an outlier.

Fifteen monkeys were divided into 2 groups based on body weight, triglyceride and fasted
glucose levels and baseline daily food and water consumption. Following a baseline period,
PF-05231023 or vehicle (30 mM lactate pH 4.8, 9% Trehalose, 0.05 mg/ml EDTA, 0.1
Author Manuscript

mg/ml L-methionine, 0.5 mg/ml Tween 20) was administered intravenously to 8 and 7
animals, respectively, on days 1, 4 and 7. Dosing was performed in a blinded manner. Water
consumption was monitored at least twice daily.

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 8

Statistical analyses
Author Manuscript

All data are expressed as means ± S.E.M. Statistical analysis between the two groups was
performed by unpaired two-tailed Student’s t test using Excel or GraphPad Prism (GraphPad
Software, Inc.), or by using R software (Fraley et al., 2012; Pinheiro et al., 2013).

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.

Acknowledgments
We thank Birgitte Andersen (Novo Nordisk) for providing recombinant FGF21, Yang Xie for assistance with the
statistical analyses, and David Self for assistance with brain dissection. This work was supported by National
Institutes of Health grants R01DK067158 (S.A.K. and D.J.M.), R01CA152301 (H.T.), NNX11AC54G (H.T.), 5P50
CA70907-16 (H.T.); the Robert A. Welch Foundation (grant I-1558 to S.A.K. and grant I-1275 to D.J.M.); the Sir
Author Manuscript

Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (Grant Number 105545/Z/14/Z
to B.M.O.); the Ford Foundation Fellowship (G.H.); the German National Science Foundation (DFG) grant MU
2789/8-1 and IZKF (Project E13) (C.P.M.); and the Howard Hughes Medical Institute (D.J.M.).

References
Badman MK, Pissios P, Kennedy AR, Koukos G, Flier JS, Maratos-Flier E. Hepatic fibroblast growth
factor 21 is regulated by PPARalpha and is a key mediator of hepatic lipid metabolism in ketotic
states. Cell Metab. 2007; 5:426–437. [PubMed: 17550778]
Bookout AL, de Groot MH, Owen BM, Lee S, Gautron L, Lawrence HL, Ding X, Elmquist JK,
Takahashi JS, Mangelsdorf DJ, Kliewer SA. FGF21 regulates metabolism and circadian behavior
by acting on the nervous system. Nature Med. 2013; 19:1147–1152. [PubMed: 23933984]
Camporez JP, Jornayvaz FR, Petersen MC, Pesta D, Guigni BA, Serr J, Zhang D, Kahn M, Samuel
VT, Jurczak MJ, Shulman GI. Cellular mechanisms by which FGF21 improves insulin sensitivity in
male mice. Endocrinology. 2013; 154:3099–3109. [PubMed: 23766126]
Author Manuscript

Can A, Dao DT, Arad M, Terrillion CE, Piantadosi SC, Gould TD. The mouse forced swim test. J Vis
Exp. 2012a:e3638. [PubMed: 22314943]
Can A, Dao DT, Terrillion CE, Piantadosi SC, Bhat S, Gould TD. The tail suspension test. J Vis Exp.
2012b:e3769. [PubMed: 22315011]
Chu AY, Workalemahu T, Paynter NP, Rose LM, Giulianini F, Tanaka T, Ngwa JS, Qi Q, Curhan GC,
Rimm EB, Hunter DJ, Pasquale LR, Ridker PM, Hu FB, Chasman DI, Qi L. Novel locus including
FGF21 is associated with dietary macronutrient intake. Human molecular genetics. 2013; 22:1895–
1902. [PubMed: 23372041]
Dong JQ, Rossulek M, Somayaji VR, Baltrukonis D, Liang Y, Hudson K, Hernandez-Illas M, Calle
RA. Pharmacokinetics and pharmacodynamics of PF-05231023, a novel long-acting FGF21
mimetic, in a first-in-human study. Br J Clin Pharmacol. 2015
Douris N, Stevanovic D, Fisher FM, Cisu TI, Chee MJ, Ly Nguyen N, Zarebidaki E, Adams AC,
Kharitonenkov A, Flier JS, Bartness TJ, Maratos-Flier E. Central Fibroblast Growth Factor 21
Browns White Fat via Sympathetic Action in Male Mice. Endocrinology. 2015 en20142001.
Dushay JR, Toschi E, Mitten EK, Fisher FM, Herman MA, Maratos-Flier E. Fructose ingestion acutely
Author Manuscript

stimulates circulating FGF21 levels in humans. Molecular metabolism. 2015; 4:51–57. [PubMed:
25685689]
Fernstrom JD, Munger SD, Sclafani A, de Araujo IE, Roberts A, Molinary S. Mechanisms for
sweetness. J Nutr. 2012; 142:1134S–1141S. [PubMed: 22573784]
Gimeno RE, Moller DE. FGF21-based pharmacotherapy--potential utility for metabolic disorders.
Trends in endocrinology and metabolism: TEM. 2014; 25:303–311. [PubMed: 24709036]
Giragossian C, Vage C, Li J, Pelletier K, Piche-Nicholas N, Rajadhyaksha M, Liras J, Logan A, Calle
RA, Weng Y. Mechanistic Investigation of the Preclinical Pharmacokinetics and Interspecies

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 9

Scaling of PF-05231023, a Fibroblast Growth Factor 21-Antibody Protein Conjugate. Drug Metab
Dispos. 2015; 43:803–811. [PubMed: 25805881]
Author Manuscript

Gonzales RA, Job MO, Doyon WM. The role of mesolimbic dopamine in the development and
maintenance of ethanol reinforcement. Pharmacology & therapeutics. 2004; 103:121–146.
[PubMed: 15369680]
Hsuchou H, Pan W, Kastin AJ. The fasting polypeptide FGF21 can enter brain from blood. Peptides.
2007; 28:2382–2386. [PubMed: 17996984]
Inagaki T, Dutchak P, Zhao G, Ding X, Gautron L, Parameswara V, Li Y, Goetz R, Mohammadi M,
Esser V, Elmquist JK, Gerard RD, Burgess SC, Hammer RE, Mangelsdorf DJ, Kliewer SA.
Endocrine regulation of the fasting response by PPARalpha-mediated induction of fibroblast
growth factor 21. Cell Metab. 2007; 5:415–425. [PubMed: 17550777]
Krishnan V, Han MH, Graham DL, Berton O, Renthal W, Russo SJ, Laplant Q, Graham A, Lutter M,
Lagace DC, Ghose S, Reister R, Tannous P, Green TA, Neve RL, Chakravarty S, Kumar A, Eisch
AJ, Self DW, Lee FS, Tamminga CA, Cooper DC, Gershenfeld HK, Nestler EJ. Molecular
adaptations underlying susceptibility and resistance to social defeat in brain reward regions. Cell.
2007; 131:391–404. [PubMed: 17956738]
Author Manuscript

Liang Q, Zhong L, Zhang J, Wang Y, Bornstein SR, Triggle CR, Ding H, Lam KS, Xu A. FGF21
maintains glucose homeostasis by mediating the cross talk between liver and brain during
prolonged fasting. Diabetes. 2014; 63:4064–4075. [PubMed: 25024372]
Owen BM, Bookout AL, Ding X, Lin VY, Atkin SD, Gautron L, Kliewer SA, Mangelsdorf DJ. FGF21
contributes to neuroendocrine control of female reproduction. Nature Med. 2013; 19:1153–1156.
[PubMed: 23933983]
Owen BM, Ding X, Morgan DA, Coate KC, Bookout AL, Rahmouni K, Kliewer SA, Mangelsdorf DJ.
FGF21 acts centrally to induce sympathetic nerve activity, energy expenditure, and weight loss.
Cell Metab. 2014; 20:670–677. [PubMed: 25130400]
Owen BM, Mangelsdorf DJ, Kliewer SA. Tissue-specific actions of the metabolic hormones
FGF15/19 and FGF21. Trends in endocrinology and metabolism: TEM. 2015; 26:22–29.
[PubMed: 25476453]
Paxinos, G.; Franklin, KB. The mouse brain in stereotaxic coordinates. Gulf Professional Publishing;
2004.
Sanchez J, Palou A, Pico C. Response to carbohydrate and fat refeeding in the expression of genes
Author Manuscript

involved in nutrient partitioning and metabolism: striking effects on fibroblast growth factor-21
induction. Endocrinology. 2009; 150:5341–5350. [PubMed: 19837871]
Sarruf DA, Thaler JP, Morton GJ, German J, Fischer JD, Ogimoto K, Schwartz MW. Fibroblast
growth factor 21 action in the brain increases energy expenditure and insulin sensitivity in obese
rats. Diabetes. 2010; 59:1817–1824. [PubMed: 20357365]
Tanaka T, Ngwa JS, van Rooij FJ, Zillikens MC, Wojczynski MK, Frazier-Wood AC, Houston DK,
Kanoni S, Lemaitre RN, Luan J, Mikkila V, Renstrom F, Sonestedt E, Zhao JH, Chu AY, Qi L,
Chasman DI, de Oliveira Otto MC, Dhurandhar EJ, Feitosa MF, Johansson I, Khaw KT, Lohman
KK, Manichaikul A, McKeown NM, Mozaffarian D, Singleton A, Stirrups K, Viikari J, Ye Z,
Bandinelli S, Barroso I, Deloukas P, Forouhi NG, Hofman A, Liu Y, Lyytikainen LP, North KE,
Dimitriou M, Hallmans G, Kahonen M, Langenberg C, Ordovas JM, Uitterlinden AG, Hu FB,
Kalafati IP, Raitakari O, Franco OH, Johnson A, Emilsson V, Schrack JA, Semba RD, Siscovick
DS, Arnett DK, Borecki IB, Franks PW, Kritchevsky SB, Lehtimaki T, Loos RJ, Orho-Melander
M, Rotter JI, Wareham NJ, Witteman JC, Ferrucci L, Dedoussis G, Cupples LA, Nettleton JA.
Genome-wide meta-analysis of observational studies shows common genetic variants associated
Author Manuscript

with macronutrient intake. Am J Clin Nutr. 2013; 97:1395–1402. [PubMed: 23636237]

Tordoff MG, Bachmanov AA. Influence of the number of alcohol and water bottles on murine alcohol
intake. Alcohol Clin Exp Res. 2003; 27:600–606. [PubMed: 12711922]
Weng Y, Chabot JR, Bernardo B, Yan Q, Zhu Y, Brenner MB, Vage C, Logan A, Calle R, Talukdar S.
Pharmacokinetics (PK), pharmacodynamics (PD) and integrated PK/PD modeling of a novel long
acting FGF21 clinical candidate PF-05231023 in diet-induced obese and leptin-deficient obese
mice. PLoS One. 2015; 10:e0119104. [PubMed: 25790234]

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 10

Zhao C, Liu Y, Xiao J, Liu L, Chen S, Mohammadi M, McClain CJ, Li X, Feng W. FGF21 mediates
alcohol-induced adipose tissue lipolysis by activation of systemic release of catecholamine in
Author Manuscript

mice. Journal of lipid research. 2015; 56:1481–1491. [PubMed: 26092866]

Author Manuscript
Author Manuscript
Author Manuscript

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 11
Author Manuscript
Author Manuscript
Author Manuscript

Figure 1. FGF21 decreases sweet preference ratio in mice by acting on the CNS
(A) Two-bottle preference assay in wild-type (WT) and Tg(Fgf21) mice administered water
vs. 3% sucrose. Representative 24 hour data from day 2 after initiating the assay are shown
as the sucrose preference ratio (sucrose intake volume/total fluid intake volume). n = 10–11/
group.
(B) Two-bottle preference assay in WT and Tg(Fgf21) mice administered water vs. 0.2%
saccharin. Representative 24 hour data from day 2 after initiating the assay are shown. n =
10–11/group.
Author Manuscript

(C) Two-bottle preference assay with water vs. 0.2% saccharin for Klbfl/fl and KlbCamk2a
mice administered either FGF21 (1 mg/kg/day) or vehicle. Representative 24 hour data from
day 3 after initiating the assay are shown. n = 6–9/group.
(D) Two-bottle preference assay with water vs. 2 mg/dl quinine for Klbfl/fl and KlbCamk2a
mice administered either FGF21 (1 mg/kg/day) or vehicle. n = 4/group.
Values are means ±S.E.M. *, p<0.05; ***, p<0.001; ###, p<0.001 by Student’s t-test.
See also Figure S1, Table S1 and S2.

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 12
Author Manuscript
Author Manuscript
Author Manuscript

Figure 2. A stable FGF21 analog decreases saccharin preference in mice and monkeys
(A) Two bottle preference assay with 0.1% saccharin in diet-induced obese mice
administered either PF-05231023 (10 mg/kg) or vehicle on days 0, 3, 7 and 10.
Data are shown as the mean ± S.E.M.; n = 8/group. **p < 0.01, ***p < 0.001 versus vehicle
Author Manuscript

group.
(B) Two bottle preference assay with 0.2% saccharin in obese cynomolgus monkeys
administered either PF-05231023 (n=8; 10 mg/kg) or vehicle (n=7) on days 1, 4 and 7. Data
are presented as mean percentage change in saccharin water intake ± S.E.M. for vehicle-
treated (closed blue circles) and PF-05231023-treated (open red circles) monkeys. Solid
lines are locally weighted scatterplot smoothing fits to the means of percent change. Mixed
effect modeling fitted to these longitudinal data using R, version 3.1.2 (Pinheiro et al.,

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 13

2013), showed a significant difference (p = 0.003) between groups. Number of days after
Author Manuscript

first treatment, treatment type, and the interaction term between treatment groups and time
were specified as fixed effects and monkey labels as a random effect.
Author Manuscript
Author Manuscript
Author Manuscript

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 14
Author Manuscript
Author Manuscript
Author Manuscript

Figure 3. FGF21 affects dopamine signaling

(A) β-Klotho (Klb) mRNA levels in the suprachiasmatic nucleus/paraventricular nucleus
Author Manuscript

(SCN/PVN) region of the hypothalamus, ventral tegmental area (VTA), nucleus accumbens
(NAc), medial prefrontal cortex (PFC) and caudate putamen of Klb+/− and Klb−/− mice (n =
6/group). Ct values are shown in the bars. ND, not detected.
(B) Concentrations of dopamine, 3,4-dihydroxyphenylacetic acid (DOPAC), homovanillic
acid (HVA) and 3-methoxytyramine (3-MT) in the NAc of mice administered either vehicle
or FGF21 for 2 weeks by osmotic minipump (n = 12/group).

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 15

(C) mRNA levels of dopamine transporter (Slc6a3) or (D) catechol-O-methyl transferase

Author Manuscript

(Comt) in VTA, NAc and caudate putamen of mice administered either vehicle or FGF21 for
2 weeks by osmotic minipump (n = 7–8/group). Ct values are shown.
Values are means ±S.E.M. *, p<0.05; **, p<0.01; ***, p<0.001 versus control group by
Student’s t-test.
See also Figure S2.
Author Manuscript
Author Manuscript
Author Manuscript

Cell Metab. Author manuscript; available in PMC 2017 February 09.

 Talukdar et al. Page 16
Author Manuscript
Author Manuscript

Figure 4. FGF21 decreases alcohol preference

(A) Ethanol preference ratio in wild-type (WT) and Tg(Fgf21) mice at the indicated ethanol
concentrations (n = 9/group).
(B) Plasma ethanol concentrations in groups of WT and Tg(Fgf21) mice 1 or 3 hours after
i.p. injection of ethanol (4g/kg) (n = 4–5/group).
Values are means ±S.E.M. *, p<0.05; **, p<0.01; and ***, p<0.001 versus control group by
Student’s t-test.
See also Table S3.
Author Manuscript
Author Manuscript

Cell Metab. Author manuscript; available in PMC 2017 February 09.