Content uploaded by Wenzhen Duan
Author content
All content in this area was uploaded by Wenzhen Duan on Jan 29, 2014
Content may be subject to copyright.
Reversal of Behavioral and Metabolic Abnormalities, and
Insulin Resistance Syndrome, by Dietary Restriction in
Mice Deficient in Brain-Derived Neurotrophic Factor
WENZHEN DUAN, ZHIHONG GUO, HAIYANG JIANG, MELVIN WARE, AND MARK P. MATTSON
Laboratory of Neurosciences (W.D., Z.G., H.J., M.P.M.), Comparative Medicine Section (M.W.), National Institute on Aging
Gerontology Research Center, Baltimore, Maryland 21224; and Department of Neuroscience (M.P.M.), Johns Hopkins
University School of Medicine, Baltimore, Maryland 21205
Dietary restriction (DR) extends life span and improves glu-
cose metabolism in mammals. Recent studies have shown that
DR stimulates the production of brain-derived neurotrophic
factor (BDNF) in brain cells, which may mediate neuropro-
tective and neurogenic actions of DR. Other studies have sug-
gested a role for central BDNF signaling in the regulation of
glucose metabolism and body weight. BDNF heterozygous
knockout (BDNFⴙ/ⴚ) mice are obese and exhibit features of
insulin resistance. We now report that an intermittent fasting
DR regimen reverses several abnormal phenotypes of
BDNF(ⴙ/ⴚ) mice including obesity, hyperphagia, and in-
creased locomotor activity. DR increases BDNF levels in the
brains of BDNF(ⴙ/ⴚ) mice to the level of wild-type mice fed ad
libitum. BDNF(ⴙ/ⴚ) mice exhibit an insulin-resistance syn-
drome phenotype characterized by elevated levels of circu-
lating glucose, insulin, and leptin; DR reduces levels of each
of these three factors. DR normalizes blood glucose responses
in glucose tolerance and insulin tolerance tests in the
BDNF(ⴙ/ⴚ) mice. These findings suggest that BDNF is a major
regulator of energy metabolism and that beneficial effects of
DR on glucose metabolism are mediated, in part, by BDNF
signaling. Dietary and pharmacological manipulations of
BDNF signaling may prove useful in the prevention and treat-
ment of obesity and insulin resistance syndrome-related
diseases. (Endocrinology 144: 2446 –2453, 2003)
T
HE NEUROENDOCRINE MECHANISMS that regulate
food intake, body weight, and energy metabolism are
complex, involving both peripheral organs and the brain (1).
The hypothalamus plays a major role in regulating food
intake by integrating signals from higher brain regions with
peripheral signals of the metabolic status of the body in-
cluding the proteins leptin, insulin, and ghrelin. A metabolic
syndrome X characterized by obesity, hyperglycemia, and
increased levels of insulin and leptin is becoming increas-
ingly common in industrialized countries, apparently as the
result of a combination of increased caloric intake and de-
creased physical activity (2). Studies of animal models have
demonstrated important roles for reduced leptin and insulin
responsiveness in the pathogenesis of obesity and type 2
diabetes (3, 4) but have also led to the realization that there
are additional, unknown neural and endocrine mechanisms
that regulate food intake, energy metabolism, and body
weight. Insight into additional mechanisms that regulate
food intake and energy metabolism has come from studies of
a protein called brain-derived neurotrophic factor (BDNF)
that is widely expressed by neurons in the brain. BDNF is
best known for its roles in development of the nervous sys-
tem (5–7) and for its involvement in the processes of synaptic
plasticity (8) and neurogenesis (9, 10) in the adult brain.
Recent findings suggest that BDNF may also be an important
regulator of food intake and energy metabolism. As evi-
dence, BDNF⫹/⫺ mice are obese (11), conditional deletion
of BDNF in the brain results in obesity and hyperactivity (12),
and BDNF administration reduces food intake and blood
glucose concentrations in diabetic mice by a central nervous
system (CNS)-mediated mechanism (13). Collectively, these
findings suggest that BDNF signaling in the brain may play
a major role in regulating energy metabolism throughout the
body.
Dietary restriction (DR; reduced caloric intake or meal
frequency with maintained nutrition) can reduce body
weight and normalize blood glucose, insulin, and leptin lev-
els in obese animals and humans (14). DR also increases both
the mean and maximum lifespan of rodents, and this anti-
aging effect is associated with enhanced insulin sensitivity
(15). DR affects cells throughout the body including the CNS,
where improvements in motor and cognitive functions (16,
17) and increased resistance of neurons to insults in models
of age-related neurodegenerative disorders (18 –20) in ro-
dents maintained on DR have been reported. Two different
paradigms of DR have been widely employed because of
their highly reproducible ability to increase lifespan in rats
and mice. In one paradigm, the animals receive food daily
but are limited to a specific amount, which is typically 30 –
40% less than the ad libitum (AL) consumption of the control
group. The second paradigm involves intermittent fasting in
which the animals are deprived of food for a full day, every
other day, and are fed AL on the intervening days. Analyses
of various physiological parameters in animals maintained
on these two different DR regimens have revealed several
similarities including decreases in body weight, tempera-
ture, heart rate, blood pressure, and glucose and insulin
levels. These DR regimens have also been shown to have
Abbreviations: AL, Ad libitum; BDNF, brain-derived neurotrophic
factor; CNS, central nervous system; DR, dietary restriction; TBST, Tris-
buffered saline with Tween 20.
0013-7227/03/$15.00/0 Endocrinology 144(6):2446–2453
Printed in U.S.A. Copyright © 2003 by The Endocrine Society
doi: 10.1210/en.2002-0113
2446
beneficial effects on the brain (9, 10, 14 –21). There are several
possible molecular mechanisms that might explain the ben-
eficial effects of DR on aging and disease including a reduc-
tion in mitochondrial oxyradical production, induction of a
cytoprotective cellular stress response, and stimulation of the
production of growth factors (15, 21, 22). In the present study,
we present evidence supporting a role for BDNF in medi-
ating beneficial effects of one DR regimen, intermittent fast-
ing, on body weight and energy metabolism in mice. Mice
with BDNF haploinsufficiency exhibit obesity and elevated
levels of glucose, insulin, and leptin; DR stimulates increased
production in the brain and this is associated with normal-
ization of the metabolic and neuroendocrine abnormalities.
Materials and Methods
Mice
Breeding pairs of BDNF heterozygous (⫹/⫺) mice were kindly pro-
vided by Dr. L. Tessarollo at the National Cancer Institute; details of their
generation, genetic background and phenotypes can be found elsewhere
(23). The haploinsufficiency of the BDNF⫹/⫺ mice was confirmed in
preliminary studies which showed that BDNF protein levels were de-
creased by approximately 50% in the cerebral cortex, hippocampus, and
striatum of BDNF⫹/⫺ mice (Fig. 1D). Three-month-old male wild-type
and BDNF⫹/⫺ littermate mice were divided into two groups (8–10
mice/group), an AL group that had continuous access to food, and a DR
group which was fasted for a 24-h period on alternate days as described
previously (19). Mice were maintained on the diets for 3 months; body
weight and food intake were measured on a weekly basis. Mice were
FIG. 1. DR reverses abnormal phenotypes of BDNF⫹/⫺ mice. Wild-type (WT) and BDNF⫹/⫺ mice were maintained for 3 months on AL or DR
feeding regimens. A, Body weights: *, P ⬍ 0.05 compared with the WT-AL value; **, P ⬍ 0.01 compared with the corresponding value for AL-fed
mice. B, Food intake: *, P ⬍ 0.05 compared with the WT-AL value; **, P ⬍ 0.01 compared with the BDNF⫹/⫺ AL value. C, Spontaneous activity:
*, P ⬍ 0.05 compared with the WT-AL value; **, P ⬍ 0.01 compared with the BDNF⫹/⫺ AL value. D, BDNF concentration: *, P ⬍ 0.01 compared
with the corresponding value for mice fed AL. All values are the mean and SEM of determinations made in 8–10 mice per group; statistical
comparisons were made using ANOVA and Scheffe´’s post hoc tests.
Duan et al. • Reversal of Behavioral and Metabolic Abnormalities Endocrinology, June 2003, 144(6):2446–2453 2447
provided free access to water and were maintained on a 12-h light, 12-h
dark cycle. All procedures were approved by the NIA Animal Care and
Use Committee in compliance with NIH guidelines.
ELISA analysis of BDNF protein levels
Freshly dissected hippocampal, cortical and striatal tissues were ho-
mogenized in lysis buffer (137 mm NaCl; 20 mm Tris; 1% Nonidet P-40
detergent; 10% glycerol; 1 mm phenylmethylsulfonylflouride; 10 mg/ml
aprotinin; 1 mg/ml leupeptin; and 0.5 mm sodium orthovanadate, pH
7.2) at 4 C. Homogenates were centrifuged at 2000 ⫻ g for 20 min (4 C),
and supernatants were used for ELISA analysis. BDNF protein levels
were quantified using a commercially available kit (Promega Corp.,
Madison, WI) according to the manufacturer’s protocol. Briefly, samples
were processed by acidification and subsequent neutralization. Ninety-
six-well plates were coated with mouse monoclonal BDNF antibody,
incubated in the presence of block and sample buffer and washed in
TBST (Tris-buffered saline with Tween 20). Samples were added in
triplicate to wells in each plate, and serial dilutions of BDNF standard
(0–500 pg/ml) were added in duplicate to wells in each plate to generate
a standard curve. Plates were incubated for 2 h, washed five times in
TBST, and incubated for2hinasolution containing rabbit polyclonal
BDNF antibodies. Wells were washed five times with TBST, and a
hydrogen peroxide solution was added together with a peroxidase sub-
strate, and plates were incubated for 10 min. The intraassay and inter-
assay variabilities for the BDNF ELISA were 6% and 13%. Reactions were
stopped by adding 100
lof1m phosphoric acid, and absorbance was
measured at 450 nm using a plate reader. The concentrations of BDNF
in each sample were determined in triplicate, and the average of the
three values was used as the value for that mouse for the statistical
analysis. Values were expressed as picograms BDNF per milligram of
protein.
Measurements of glucose, insulin, IGF-I and leptin, and
glucose and insulin tolerance tests
Fasting blood samples were taken from both the DR and AL groups
14 h after withdrawal of food. Blood samples designated as feeding were
drawn 6 h after introducing food into the cages of both DR and AL mice
that had been subjected to a preceding 14-h fast. Blood glucose con-
centrations were measured using a glucometer (Lifescan Inc., Milpitas,
CA). Insulin levels were determined in duplicate using 10
l of serum
using an UltraSensitive Mouse Insulin ELISA kit (ALPCO Diagnostics,
Windham, NH) according to manufacturer’s protocol. Serum leptin
levels and IGF-I levels were determined in duplicate in 50-
l serum
samples using a leptin ELISA kit (Cayman Chemical Co., Ann Arbor, MI)
and an IGF-I ELISA kit (Diagnostic Systems Laboratories, Inc., Webster,
TX) according to the manufacturer’s protocols. For the glucose tolerance
test mice were given an oral bolus of d-glucose (2 g/kg body weight) and
the blood glucose concentration was measured in samples taken at 0, 15,
30, 60, and 120 min after glucose administration. For the insulin tolerance
test, mice were overnight fasted, and insulin (1 U/kg, Sigma, St. Louis,
MO) was administered by ip injection and blood glucose concentrations
were determined at 0, 15, 30, and 60 min after insulin administration.
Locomotor activity
An automated activity monitor (Digiscan Micro; Omnitech, Colum-
bus, OH) was used to quantify spontaneous activity during a 2-h re-
cording period. Locomotor activity was measured at feeding condition
(6 h after addition of food) for both AL and DR groups to avoid the effect
of the food searching behavior in the DR group. Each mouse was placed
in a recording cage in which 16 infrared sensors monitored the mouse’s
movement. Data were automatically collected and transferred to a com-
puter for later analysis of locomotor activity; the recording period was
initiated 15 min after placing the cage in the recording area. The total
number of sensors triggered during the 2-h test period was used as a
measure of overall activity level of the mouse.
Statistical analyses
Data were analyzed using one-way ANOVA and post hoc compari-
sons of means were based on Scheffe´’s test. P ⬍ 0.05 was considered
statistically significant. Analyses were performed using Statview soft-
ware (SAS Institute, Cary, NC).
Results
DR reverses abnormal phenotypes of BDNF⫹/⫺ mice
including obesity, hyperphagia, and altered locomotor
activity, and normalizes BDNF levels in the brain
Two-month-old wild-type and BDNF⫹/⫺ mice were
maintained on either an AL or an intermittent fasting DR
feeding regimen for 3 months. The body weights and food
intakes of wild-type mice maintained on the DR regimen
were significantly decreased compared with mice fed AL
(Fig. 1, A and B). Consistent with a previous study (24),
BDNF⫹/⫺ mice exhibited increased body weight and hy-
perphagia. The DR regimen decreased the body weights and
food intake of BDNF⫹/⫺ mice as well as wild-type mice to
levels similar to those of wild-type mice on the AL diet (Fig.
1, A and B). We found that spontaneous locomotor activity
was significantly increased in BDNF⫹/⫺ mice on the AL
diet, and that DR normalized this hyperactivity (Fig. 1C).
To confirm that BDNF levels were decreased in the brains
of BDNF⫹/⫺ mice, and to determine whether DR affected
BDNF levels, we performed ELISA analysis to quantify
BDNF protein levels in three different brain regions (hip-
pocampus, striatum, and cerebral cortex) of wild-type and
BDNF⫹/⫺ mice that had been maintained on AL and DR
diets for 3 months. As expected, BDNF levels were decreased
in all three brain regions of BDNF⫹/⫺ mice on the AL diet
when compared with wild-type mice on the AL diet; the
magnitude of the decreases ranged from 45–65% (Fig. 1D).
BDNF levels were increased by 2- to 3-fold in each brain
region of wild-type mice maintained on the DR regimen
compared with wild-type mice fed AL. BDNF levels were
also increased by 2- to 3-fold in each brain region of
BDNF⫹/⫺ mice that had been maintained on the DR diet
compared with BDNF⫹/⫺ mice fed AL (Fig. 1B). Thus, the
single copy of the BDNF gene in BDNF⫹/⫺ mice appears to
be as responsive to DR as are the BDNF genes in wild-type mice.
DR normalizes glucose regulation in BDNF⫹/⫺ mice
It was previously reported that DR can reduce blood glucose
levels and increase insulin sensitivity in rats, mice and nonhu-
man primates (15, 25, 26). Because BDNF⫹/⫺ mice are hy-
perphagic and obese, and because DR increases BDNF pro-
duction, we sought a link between BDNF levels and peripheral
glucose metabolism. Measurements of blood glucose in wild-
type and BDNF⫹/⫺ mice that had been maintained on an AL
diet revealed a significant increase in fasting glucose levels in
BDNF⫹/⫺ mice (Fig. 2A). Glucose levels were also signifi-
cantly increased in BDNF⫹/⫺ mice, compared with wild-type
mice, under feeding conditions (Fig. 2B). BDNF⫹/⫺ mice that
had been maintained on DR exhibited significant decreases in
blood glucose levels under both fasting and feeding conditions
compared with BDNF⫹/⫺ mice that had been maintained on
an AL diet (Fig. 2, A and B). Diabetic rodents and humans
exhibit an abnormal blood glucose response to feeding, with
glucose levels typically rising to a higher level and remaining
elevated for a prolonged time period compared with nondia-
betic subjects. We performed glucose tolerance tests in wild-
2448 Endocrinology, June 2003, 144(6):2446–2453 Duan et al. • Reversal of Behavioral and Metabolic Abnormalities
type and BDNF⫹/⫺ mice that had been maintained on AL or
DR diets; mice were given an oral bolus of glucose, and glucose
concentrations were determined in blood samples taken 15, 30,
60, and 120 min later. BDNF⫹/⫺ mice that had been fed AL
exhibited a much greater elevation of blood glucose levels,
which remained elevated much longer, than did wild-type mice
fed AL (Fig. 2C). In contrast, BDNF⫹/⫺ mice that had been
maintained on DR exhibited a blood glucose response to the
oral glucose challenge that was similar to that of wild-type AL
mice, although greater than that of wild-type mice on DR. These
findings demonstrate a profound abnormality in glucose reg-
ulation in BDNF⫹/⫺ mice that can be normalized by DR.
DR reverses hyperinsulinemia and improves insulin
sensitivity in BDNF⫹/⫺ mice
Because hyperglycemia and an abnormal glucose toler-
ance test are often associated with hyperinsulinemia and
reduced insulin sensitivity, we measured serum insulin
levels and evaluated insulin sensitivity in wild-type and
BDNF⫹/⫺ mice that had been maintained on AL and DR
diets. AL-fed BDNF⫹/⫺ mice exhibited a dramatic 9-fold
increase in fasting serum insulin levels compared with
AL-fed wild-type mice (Fig. 3A). Under feeding conditions
the BDNF⫹/⫺ mice exhibited a 20-fold greater serum
insulin concentration compared with wild-type mice (Fig.
3B). DR resulted in highly significant decreases in serum
insulin levels in both wild-type and BDNF⫹/⫺ mice un-
der fasting and nonfasting conditions (Fig. 3, A and B). To
provide insight into the effects of decreased BDNF levels
and DR on insulin sensitivity, wild-type, and BDNF⫹/⫺
mice that had been maintained on AL or DR diets were
administered insulin and glucose concentrations were
measured in blood samples taken 15, 30, and 60 min later.
Wild-type mice exhibited a marked and sustained reduc-
tion in blood glucose levels following insulin administra-
tion, with no appreciable difference observed between
FIG. 2. Hyperglycemia and impaired glucose tolerance in BDNF⫹/⫺ mice are normalized by DR. Wild-type (WT) and BDNF⫹/⫺ mice were
maintained for 3 months on AL or DR feeding regimens. A and B, Glucose concentrations were measured in blood samples taken after an
overnight fast (A) or during feeding conditions (B). Note that the scales for the glucose concentrations in the two graphs are different. *, P ⬍
0.01 compared with the value for the same genotype of mice fed AL; #, P ⬍ 0.05 cmpared to the WT-AL value. C, Mice were administered an
oral bolus of glucose (2 g/kg) and the glucose concentration in blood samples taken at the indicated times was determined. *, P ⬍ 0.01 compared
with the value for each of the other three groups at that time point. Values are the mean and SEM of measurements made in 8 –10 mice per
group. Statistical comparisons were made using ANOVA and Scheffe´’s post hoc tests.
Duan et al. • Reversal of Behavioral and Metabolic Abnormalities Endocrinology, June 2003, 144(6):2446–2453 2449
mice that had been maintained on AL or DR diets (Fig. 3C).
BDNF⫹/⫺ that had been fed AL exhibited a striking in-
sensitivity to insulin whereas, in contrast, BDNF⫹ /⫺ mice
on DR exhibited a decrease in blood glucose concentration
in response to insulin that was similar to that of wild-type
mice (Fig. 3C). These results demonstrate that mice with
reduced BDNF levels are relatively insensitive to insulin,
and that this abnormality can be corrected by DR.
In some studies, increased levels of IGF-1 have been
associated with diabetes (27) and decreased levels of IGF-1
may occur in animals maintained on low calorie diets (28).
However, the role of such changes in glucose metabolism
and physiological actions of DR are unknown. We there-
fore measured levels of IGF-1 in serum samples from wild-
type and BDNF⫹/⫺ mice that had been maintained on AL
or alternate day fasting DR diets. There were no significant
effects of genotype or diet on fasting IGF-1 levels, although
IGF-1 levels tended to be somewhat lower in mice on DR
(Table 1). Similarly, in nonfasted mice there were no dif-
ferences in IGF-1 levels between wild-type and BDNF⫹/⫺
mice on either diet, nor did diet affect IGF-1 levels in blood
samples drawn in nonfasted mice (Table 1). The lack
of effects of BDNF levels and diet on IGF-1 levels suggest
that the beneficial effects of DR on the abnormal pheno-
TABLE 1. IGF-I levels are similar in BDNF⫹/⫺ mice and wild-
type (WT) mice, and DR has no significant effect on IGF-I levels in
either BDNF⫹/⫺ mice or WT mice
Fasted Nonfasted
AL DR AL DR
WT 2657 ⫾ 22 2136 ⫾ 162 2812 ⫾ 197 2645 ⫾ 101
BDNF⫹/⫺ 2708 ⫾ 91 2480 ⫾ 139 2781 ⫾ 130 2472 ⫾ 73
The values for IGF-1 levels are nanograms per milliliter of serum.
FIG. 3. Mice with reduced BDNF levels exhibit insulin insensitivity that is normalized by DR. Wild-type (WT) and BDNF⫹/⫺ mice were
maintained for 3 months on AL or DR feeding regimens. A and B, Insulin concentrations were measured in blood samples taken after an overnight
fast (A) or during feeding conditions (B). *, P ⬍ 0.001 compared with the value for the same genotype of mice fed AL; **, P ⬍ 0.001 compared
with the WT-AL value. C, Mice were administered insulin (1 U/kg) and the glucose concentration in blood samples taken at the indicated times
was determined. *, P ⬍ 0.01 compared with the value for each of the other three groups at that time point. Values are the mean and SEM of
measurements made in 8 –10 mice per group. Statistical comparisons were made using ANOVA and Scheffe´’s post hoc tests.
2450 Endocrinology, June 2003, 144(6):2446–2453 Duan et al. • Reversal of Behavioral and Metabolic Abnormalities
types of BDNF⫹/⫺ mice are unlikely to be mediated by
IGF-1.
DR reverses hyperleptinaemia in BDNF⫹/⫺ mice
Leptin is a hormone released from adipose cells that is
transported to brain where it binds to its receptor in hypo-
thalamus to regulate food intake and energy expenditure
(29). The findings to this point suggested that BDNF exerts
some actions similar to leptin, namely, reduced food intake
and body weight and improved insulin sensitivity. We there-
fore measured the concentration of leptin in blood samples
from wild-type and BDNF⫹/⫺ mice that had been main-
tained on AL or DR diets. Fasting leptin levels were signif-
icantly increased, by approximately 2-fold, in BDNF⫹/⫺
mice compared with wild-type mice (Fig. 4A). Both wild-
type and BDNF⫹/⫺ mice that had been maintained on DR
exhibited large and highly significant decreases in fasting
leptin levels (Fig 4A). In nonfasted mice that had been main-
tained on DR, leptin levels were significantly decreased in
both wild-type and BDNF⫹/⫺ mice compared with AL-fed
mice, with the effect of DR being greater in wild-type mice
compared with BDNF⫹/⫺ mice (Fig. 4B).
Discussion
The present findings demonstrate that a reduction in
BDNF levels in the brain resulting from BDNF gene haplo-
insufficiency causes abnormalities in glucose metabolism
and body weight regulation in mice that are provided free
access to food. The abnormalities include hyperphagia, obe-
sity, hyperglycemia, hyperinsulinemia, hyperleptinemia,
and decreased insulin sensitivity. We found that when the
food intake of the BDNF⫹/⫺ mice was restricted by main-
taining them on an intermittent fasting regimen, their glu-
cose regulation abnormalities and obesity were ameliorated.
The correction of the behavioral and metabolic abnormalities
of BDNF⫹/⫺ mice by intermittent fasting was associated
with an increase in brain BDNF levels to levels present in
wild-type mice fed AL.
The mechanism(s) whereby BDNF regulates food intake,
body weight, and glucose metabolism are not known. Con-
ditional deletion of BDNF in the brains of mice resulted in
obesity (12), suggesting that the metabolic phenotype of
BDNF⫹/⫺ mice documented in the present study is the
result of the decrease in BDNF levels in the brain rather than
in peripheral sites. Moreover, it was recently reported that
infusion of BDNF into the lateral ventricles results in a re-
duction in blood glucose levels, demonstrating that BDNF
signaling in the CNS can modify peripheral glucose regu-
lation (13). BDNF is widely expressed by neurons in multiple
brain regions and which of these brain regions is involved in
the antiobesity and antidiabetic actions of BDNF remains to
be determined. Previous studies have established important
roles for BDNF signaling in synaptic plasticity in the hip-
pocampus and cerebral cortex (30, 31), and these brain re-
gions do influence a variety of behaviors including food
intake and body weight (32, 33). BDNF is also produced by
hypothalamic cells (34) and could, in principle, act locally in
the hypothalamus to suppress food intake. Interestingly,
BDNF signaling can induce the growth of serotonergic fibers
and may enhance serotonergic signaling (23, 24). Studies of
antidepressant drugs and other serotonin-modulating drugs
have provided evidence that serotonin can suppress appetite
and induce weight loss (35). Therefore, enhanced serotoner-
gic signaling might mediate the antiobesity and antidiabetic
actions of BDNF.
The normalization of brain BDNF levels in BDNF⫹/⫺
mice maintained on an intermittent fasting DR regimen may
account for its ability to reverse behavioral and metabolic
abnormalities in these mice. It was previously reported that
DR can increase BDNF levels in the hippocampus and cortex
of mice, and it was proposed that this up-regulation of BDNF
plays an important role in the neuroprotective and neuro-
genesis-promoting actions of DR (9, 22). In addition, it was
recently reported that a high-fat, refined sugar diet decreases
levels of BDNF in the hippocampus of rats (36), and that
physical exercise and enriched environments up-regulate
BDNF levels in the brain (37, 38). Interestingly, patients with
Huntington’s disease and transgenic mice expressing mutant
huntingtin proteins exhibit reduced levels of BDNF in their
brains and are hyperglycemic (39– 42). Therefore, multiple
lines of evidence support an important role for BDNF sig-
naling in the brain in the regulation of energy metabolism
and body weight in various physiological and pathological
states. Although the present data do not provide conclusive
evidence that the beneficial effects of DR on peripheral glu-
cose metabolism are mediated by BDNF signaling in the
brain, they are consistent with such a possibility.
DR reduced the serum levels of both leptin and insulin in
FIG. 4. DR reverses hyperleptinemia
in BDNF⫹/⫺ mice. Two-month-old
mice were maintained on AL or DR reg-
imens for 3 months. Serum leptin levels
were measured in mice that had been
fasted overnight (A) and in nonfasted
mice (B). Values are the mean and SEM
(n ⫽ 8 –10 mice per group). *, P ⬍ 0.01
compared with the corresponding value
for group AL. **, P ⬍ 0.01 compared
with the value for WT mice on the AL
regimen (ANOVA with Scheffe´’s post
hoc tests).
Duan et al. • Reversal of Behavioral and Metabolic Abnormalities Endocrinology, June 2003, 144(6):2446–2453 2451
BDNF⫹/⫺ mice. Insulin and leptin are released from pe-
ripheral tissues and are transported to brain through similar
saturable transport mechanisms (43). Microdialysis has con-
firmed that insulin levels in the extracellular fluid of hypo-
thalamic nuclei are regulated during meal absorption (44).
Because circulating insulin and leptin can gain access to the
CNS, the behavioral and metabolic effects of central insulin
and leptin have important roles in the regulation of energy
balance and peripheral action of these hormones (45, 46). The
receptors for both hormones are expressed at particularly
high levels in cells of the hypothalamus known to regulate
energy homeostasis. The effects of DR on peripheral insulin
and leptin levels might therefore result from peripheral
and/or central effects of DR. BDNF might regulate energy
metabolism by directly activating receptors in hypothalamus
in neurons regulate food intake and energy balance, or it
might act indirectly by enhancing leptin and/or insulin sig-
naling in the CNS. In any case, the present findings suggest
that the abilities of DR to increase insulin sensitivity and to
reduce circulating glucose and leptin levels are associated
with increased BDNF levels.
The phenotypes of BDNF⫹/⫺ mice are very similar to
those of humans with metabolic syndrome X, a condition that
places them at increased risk of cardiovascular disease,
stroke and type 2 diabetes (2). Studies of rodents, monkeys,
and humans have clearly shown that DR can prevent and
reverse the abnormalities in glucose metabolism associated
with metabolic syndrome X, and can also enhance insulin
sensitivity in subjects with normal glucose metabolism (14,
25, 26). In agreement with the latter findings, we found that
DR reduces glucose, insulin, and leptin levels not only in
BDNF⫹/⫺ mice, but also in wild-type mice. Because DR
induced marked increases in BDNF levels of the brains of
wild-type mice, and because central administration of BDNF
is sufficient to improve insulin sensitivity in animals fed AL
(13), it seems likely that the increase in brain BDNF levels
contributes to the beneficial effects of DR on glucose metab-
olism and body weight in normal subjects. Collectively, the
available data therefore suggest that dietary and pharmaco-
logical manipulations of BDNF signaling may prove useful
as therapeutic approaches for preventing and treating a
range of disorders that involve abnormalities in body weight
regulation and energy metabolism.
Our study employed only one DR regimen, intermittent
fasting. It is therefore important to consider possible differ-
ences between calorie-restricted diets in which the animals
are provided food every day, but with fewer calories, and the
intermittent fasting regimen. In the present study, the mice
in the DR group consumed only 10–15% less food over time
compared with the mice on the control AL diet. However,
previous studies have shown that this intermittent DR reg-
imen extends lifespan by approximately 30% in the same
strain of mice (47). We have previously shown that the al-
ternate day fasting DR regimen increases neuronal resistance
to dysfunction and death in several different rodent models
of neurodegenerative diseases (18 –22). Moreover, we have
recently made a direct comparison of the effects of a 30%
calorie restriction daily feeding to the alternate day fasting
regimen in C57BL/6 mice. Despite only a 15% reduction in
overall calorie intake in the mice maintained on the alternate
day fasting regimen, the mice exhibited greater resistance to
excitotoxin-induced damage to hippocampal neurons com-
pared with the mice that had been maintained on the 30%
caloric restriction diet (Guo, Z., and M. P. Mattson, unpub-
lished data). We also found that BDNF levels were increased
by a greater amount in the hippocampus and cerebral cortex
of mice maintained on the alternate day fasting regimen
compared with those on the 30% caloric restriction diet
(Duan, W., and M. P. Mattson, unpublished data). When
taken together with our previous data documenting that
intermittent fasting DR regimens induce the expression of
stress proteins (HSP-70 and GRP-78; Refs. 19, 20, 22), we
believe this cellular stress response is key to the beneficial
effects of DR on the brain. The possibility that the beneficial
effects of intermittent fasting can, in part, be dissociated from
caloric intake is supported by a very recent study that
showed that targeted deletion of the insulin receptor in ad-
ipose cells results in increased longevity without a reduction
in caloric intake (48). Although our findings suggest an im-
portant role for BDNF signaling in the regulation of glucose
metabolism and brain aging, further studies will be required
to determine whether BDNF signaling plays a key role in the
life span-extending effects of DR.
Acknowledgments
Received December 9, 2002. Accepted February 28, 2003.
Address all correspondence and requests for reprints to: Mark P.
Mattson, Laboratory of Neurosciences, National Institute on Aging Ger-
ontology Research Center, 5600 Nathan Shock Drive, Baltimore, Mary-
land 21224. E-mail: mattsonm@grc.nia.nih.gov.
References
1. McMinn JE, Baskin DG, Schwartz MW 2000 Neuroendocrine mechanisms
regulating food intake and body weight. Obes Rev 1:37– 46
2. Reusch JE 2002 Current concepts in insulin resistance, type 2 diabetes mellitus,
and the metabolic syndrome. Am J Cardiol 90:19G–26G
3. Dube MG, Beretta E, Dhillon H, Ueno N, Kalra PS, Kalra SP 2002 Central
leptin gene therapy blocks high-fat diet-induced weight gain, hyperleptinemia,
and hyperinsulinemia: increase in serum ghrelin levels. Diabetes 51:1729 –1736
4. Sone H, Suzuki H, Takahashi A, Yamada N 2001 Disease model: hyperin-
sulinemia and insulin resistance. Part A-targeted disruption of insulin signal-
ing or glucose transport. Trends Mol Med 7:320–322
5. Tucker KL, Meyer M, Barde YA 2001 Neurotrophins are required for nerve
growth during development. Nat Neurosci 4:29 –37
6. Von Bartheld CS, Johnson JE 2001 Target-derived BDNF (brain-derived neu-
rotrophic factor) is essential for the survival of developing neurons in the
isthmo-optic nucleus. J Comp Neurol 433:550 –564
7. Vicario-Abejon C, Collin C, McKay RD, Segal M 1998 Neurotrophins induce
formation of functional excitatory and inhibitory synapses between cultured
hippocampal neurons. J Neurosci 18:7256 –7271
8. Tyler WJ, Alonso M, Bramham CR, Pozzo-Miller LD 2002 From acquisition
to consolidation: on the role of brain-derived neurotrophic factor signaling in
hippocampal-dependent learning. Learn Mem 9:224 –237
9. Lee J, Seroogy KB, Mattson MP 2002 Dietary restriction enhances neurotro-
phin expression and neurogenesis in the hippocampus of adult mice. J Neu-
rochem 80:539 –547
10. Lee J, Duan W, Mattson MP 2002 Evidence that brain-derived neurotrophic
factor is required for basal neurogenesis and mediates, in part, the enhance-
ment of neurogenesis by dietary restriction in the hippocampus of adult mice.
J Neurochem 82:1367–1375
11. Kernie SG, Liebl DJ, Parada LF 2000 BDNF regulates eating behavior and
locomotor activity in mice. EMBO J 19:1290–1300
12. Rios M, Fan G, Fekete C, Kelly J, Bates B, Kuehn R, Lechan RM, Jaenisch
R 2001 Conditional deletion of brain-derived neurotrophic factor in the post-
natal brain leads to obesity and hyperactivity. Mol Endocrinol 15:1748–1757
13. Nakagawa T, Ono-Kishino M, Sugaru E, Yamanaka M, Taiji M, Noguchi H
2002 Brain-derived neurotrophic factor (BDNF) regulates glucose and energy
metabolism in diabetic mice. Diabetes Metab Res Rev 18:185–191
2452 Endocrinology, June 2003, 144(6):2446–2453 Duan et al. • Reversal of Behavioral and Metabolic Abnormalities
14. Wanagat J, Allison DB, Weindruch R 1999 Caloric intake and aging: mech-
anisms in rodents and a study in nonhuman primates. Toxicol Sci 52:S35–S40
15. Sohal RS, Weindruch R 1996 Oxidative stress, calorie restriction, and aging.
Science 273:59 –63
16. Ingram DK, Weindruch R, Spangler EL, Freeman JR, Walford RL 1987
Dietary restriction benefits learning and motor performance of aged mice. J
Gerontol 42:78 –81
17. Finch CE, Morgan TE 1997 Food restriction and brain aging. In: Mattson MP,
Geddes JW, eds. The aging brain. New York: JAI Press
18. Bruce-Keller AJ, Umberger G, McFall R, Mattson MP 1999 Food restriction
reduces brain damage and improves behavioral outcome following excitotoxic
and metabolic insults. Ann Neurol 45:8 –15
19. Duan W, Mattson MP 1999 Dietary restriction and 2-deoxyglucose adminis-
tration improve behavioral outcome and reduce degeneration of dopaminergic
neurons in models of Parkinson’s disease. J Neurosci Res 57:195–206
20. Yu ZF, Mattson MP 1999 Dietary restriction and 2-deoxyglucose administra-
tion reduce focal ischemic brain damage and improve behavioral outcome:
evidence for a preconditioning mechanism. J Neurosci Res 57:830 –839
21. Mattson MP, Chan SL, Duan W 2002 Modification of brain aging and neu-
rodegenerative disorders by genes, diet and behavior. Physiol Rev 82:637–672
22. Duan W, Guo Z, Mattson MP 2000 Brain-derived neurotrophic factor mediates
an excitoprotective effect of dietary restriction in mice. J Neurochem 76:
619– 626
23. Lyons WE, Mamounas LA, Ricaurte GA, Coppola V, Reid SW, Bora SH,
Wihler C, Koliatsos VE, Tessarollo L 1999 Brain-derived neurotrophic factor-
deficient mice develop aggressiveness and hyperphagia in conjunction with
brain serotonergic abnormalities. Proc Natl Acad Sci USA 96:15239 –15244
24. Mamounas LA, Blue ME, Siuciak JA, Altar CA 1995 Brain-derived neuro-
trophic factor promotes the survival and sprouting of serotonergic axons in rat
brain. J Neurosci 15:7929 –7939
25. Hansen BC 2001 Introduction. Symposium: calorie restriction: effects on body
composition, insulin signaling and aging. J Nutr 131:900S–902S
26. Roberts SB, Pi-Sunyer X, Kuller L, Lane MA, Ellison P, Prior JC, Shapses S
2001 Physiologic effects of lowering caloric intake in nonhuman primates and
nonobese humans. J Gerontol A Biol Sci Med Sci 56:66 –75
27. Froesch ER, Hussain M 1994 Recombinant human insulin-like growth factor-I:
a therapeutic challenge for diabetes mellitus. Diabetologia 37:S179–S185
28. Dunn SE, Kari FW, French J, Leininger JR, Traylos G, Wilson R, Barett JC
1997 Dietary restriction reduces insulin-like growth factor I levels, which
modulates apoptosis, cell proliferation, and tumor progression in p53-deficient
mice. Cancer Res 57:4667– 4672
29. Magni P, Motta M, Martini L 2000 Leptin: a possible link between food intake,
energy expenditure, and reproductive function. Regul Pept 92:51–56
30. Linnarsson S, Bjorklund A, Ernfors P 1997 Learning deficit in BDNF mutant
mice. Eur J Neurosci 9:2581–2587
31. Marty S, Berzaghi MP, Berringer B 1997 Neurotrophins and activity-depen-
dent plasticity of cortical interneurons. Trends Neurosci 20:198–202
32. Shanley LJ, Irving AJ, Harvey J 2001 Leptin enhances NMDA receptor func-
tion and modulates hippocampal synaptic plasticity. J Neurosci 21:RC186
33. Huang XF, Wang H 1998 Altered c-fos expression in autonomic regulatory
centers of genetically obese (ob/ob) mouse brain. Brain Res 799:307–310
34. Viant MR, Millam JR, Delany ME, Fry DM 2000 Regulation of brain-derived
neurotrophic factor messenger RNA levels in avian hypothalamic slice cul-
tures. Neuroscience 99:373–380
35. Halford JC, Blundell JE 2000 Pharmacology of appetite suppression. Prog
Drug Res 54:25–58
36. Molteni R, Barnard RJ, Ying Z, Roberts CK, Gomez-Pinilla F 2002 A high-fat,
refined sugar diet reduces hippocampal brain-derived neurotrophic factor,
neuronal plasticity, and learning. Neuroscience 112:803– 814
37. Oliff HS, Berchtold NC, Isackson P, Cotman CW 1998 Exercise-induced
regulation of brain-derived neurotrophic factor (BDNF) transcript in the rat
hippocampus. Mol Brain Res 61:147–153
38. Ickes BR, Pham TM, Sanders LA, Albeck DS, Mohammed AH, Granholm
AC 2000 Long-term environmental enrichment leads to regional increases in
neurotrophin levels in rat brain. Exp Neurol 164:45–52
39. Zuccato C, Ciammola A, Rigamonti D, Leavitt BR, Goffredo D, Conti L,
MacDonald ME, Friedlander RM, Silani V, Hayden MR, Timmusk T, Si-
pione S, Cattaneo E 2001 Loss of huntingtin-mediated BDNF gene transcrip-
tion in Huntington’s disease. Science 293:493– 498
40. Ferrer I, Goutan E, Marin C, Rey MJ, Ribalta T 2000 Brain-derived neuro-
trophic factor in Huntington disease. Brain Res 866:257–261
41. Farrer LA 1985 Diabetes mellitus in Huntington disease. Clin Genet 27:62–67
42. Duan W, Guo Z, Jiang H, Ware M, Li XJ, Mattson MP 2003 Dietary restriction
normalizes glucose metabolism and BDNF levels, slows disease progression,
and increases survival in huntingtin mutant mice. Proc Natl Acad Sci USA
100:2911–2916
43. Banks WA, Kastin AJ, Huang W, Jaspan JB, Maness LM 1996 Leptin enters
the brain by a saturable system independent of insulin. Peptides 17:305–311
44. Schwartz MW, Figlewicz DP, Baskin DG, Woods SC, Porte DJ 1992 Insulin
in the brain; a hormonal regulator of energy balance. Endocr Rev 13:387–414
45. Obici S, Zhang BB, Karkanias G, Rossetti L 2002 Hypothalamic insulin
signaling is required for inhibition of glucose production. Nat Med 8:1376 –
1382
46. Dhillon H, Kalra SP, Kalra PS 2001 Dose-dependent effects of central leptin
gene therapy on genes that regulate body weight and appetite in the hypo-
thalamus. Mol Ther 4:139 –145
47. Goodrick CL, Ingram DK, Reynolds MA, Freeman JR, Cider NL 1990 Effects
of intermittent feeding upon body weight and lifespan in inbred mice: inter-
action of genotype and age. Mech Aging Dev 55:69 –87
48. Bluher M, Kahn BB, Kahn CR 2003 Extended longevity in mice lacking the
insulin receptor in adipose tissue. Science 299:572–574
Duan et al. • Reversal of Behavioral and Metabolic Abnormalities Endocrinology, June 2003, 144(6):2446–2453 2453
