Brain-derived Neurotrophic Factor Regulates Energy Expenditure Through the Central Nervous System in Obese Diabetic Mice

It has been previously demonstrated that brain-derived neurotrophic factor (BDNF) regulates glucose metabolism and energy expenditure in rodent diabetic models such as C57BL/KsJ-leprdb/leprdb (db/db) mice. Central administration of BDNF has been found to reduce blood glucose in db/db mice, suggesting that BDNF acts through the central nervous system. In the present study we have expanded these investigations to explore the effect of central administration of BDNF on energy metabolism. Intracerebroventricular administration of BDNF lowered blood glucose and increased pancreatic insulin content of db/db mice compared with vehicle-treated pellet pair-fed db/db mice. While body temperatures of the pellet pair-fed db/db mice given vehicle were reduced because of restricted food supply in this pair-feeding condition, BDNF treatment remarkably alleviated the reduction of body temperature suggesting the enhancement of thermogenesis. BDNF enhanced norepinephrine turnover and increased uncoupling protein-1 mRNA expression in the interscapular brown adipose tissue. Our evidence indicates that BDNF activates the sympathetic nervous system via the central nervous system and regulates energy expenditure in obese diabetic animals.


INTRODUCTION
Brain-derived neurotrophic factor (BDNF), a member of the neurotrophin family, has been widely demonstrated to function in the central and peripheral nervous systems and motor neu- rons in the fetus and in adulthood. [1,2,3,4,5,6]'9'1'11] In addi- tion to those diverse roles of BDNF in the nerv- ous system, we have discovered that BDNF plays important roles in the endocrine system and regulates glucose metabolism.'13] Although BDNF also suppresses food intake in such hyperphagic obese mice, we developed a novel apparatus to pair-feed vehicle-treated con- trol mice precisely to BDNF-treated mice and demonstrated that BDNF has a major hypoglycemic effect independent of appetite.I41 We have clarified the unique profile of peripheral BDNF administration in regulating glucose me- tabolism: (1) BDNF enhances insulin sensitivity and ameliorates insulin resistance; (2) the hypoglycemic effect of BDNF lasts long after the cessation of treatment; and (3) insulin content in pancreas is increased and in histological observa- tions, insulin-positive pancreatic beta cells are regranulated by BDNF administration.I13, 14,151 In- terestingly, in addition to its efficacy on glucose metabolism, BDNF also prevents the reduction of body temperature in the db/db mice deprived of food supply. [141This finding indicates that BDNF ameliorates the impaired energy balance in dia- betic mice.However, the mechanism by which BDNF regulates glucose metabolism and energy expenditure still remains unclear.We have not yet obtained any evidence to show the direct effects of BDNF on glucose metabolism in cultured cells from peripheral tissues such as liver, muscle, and adipose tissue.Under the precise control of food intake by means of our pellet pair-fed.apparatus,we have demonstrated that intracerebroventricu- lar administration of BDNF shows the similar anorectic and hypoglycemic effects as seen in peripheral administration in db/db mice. [141We thus hypothesize that BDNF regulates glucose metabolism by acting through the central nervous system.To evaluate this hypothesis we have ana- lyzed the effects of intracerebroventricular BDNF administration on energy expenditure in the present study.We explore the action of BDNF in regulating thermogenesis and demonstrate the involvement of the sympathetic nervous system in this process.
were performed basically as described in our pre- vious study.I141 The supply of pellets to the BDNF- treated mice was not limited, but the supply of pellets to the vehicle-treated mice was limited to the number of those consumed by the BDNF- treated mice.All animal experiments were done according to the guidelines of the Sumitomo Phar- maceuticals Committee on Animal Research.Intracerebroventricular Administration of BDNF Human recombinant BDNF (N-terminal methion- ine-free, Regeneron Pharmaceuticals, Tarrytown, NY) was administered using artificial cerebro- spinal fluid (aCSF; 0.166g/L CaCI2, 7.014g/L NaC1, 0.298g/L KC1, 0.203g/L MgCla/6H20 and 2.10g/L NaHCO3) as a vehicle for in- tracerebroventricular administration.Mice were anesthetized with diethyl ether, and fifteen micrograms of BDNF (3txl/mouse) were injected through a Hamilton syringe into the lateral cere- bral ventricle according to the following coordi- nates: 1.0mm lateral to the bregma and 3.0mm ventral to the skull surface.For the pellet-pair feed experiment, mice received a total of five injections, alternating sides of the head for each injection, with the injections being given every other day (three on one side of the head and two on the other side).Both sides of injection placement were verified by injecting Evans Blue in the same man- ner at the end of the experiment.

Animals
Male C57BL/KsJ-db/db mice were obtained from Clea Japan Inc. (Tokyo, Japan).Mice were singly housed and the treatments started at 10-12 weeks of age.Animals were given food (CE-2, Clea Japan Inc.) and water ad libitum except for the pair-feed- ing experiment.Pellet pair-fed mice were housed in the synchronized pellet pair-feeding apparatus (Sumitomo Pharmaceuticals and Osaka Micro Systems, Osaka, Japan).Pair-feeding experiments Measurement of Blood Glucose and Insulin Blood samples were collected from tail vein, and blood glucose was measured by the GLUCOSE CII-TEST WAKO (Mutarotase-glucose oxidase method, Wako Chemical, Osaka, Japan).Plasma insulin concentrations were measured by ELISA (Levis-insulin-mouse; Shibayagi, Gunma, Japan).At the end of the treatment, the whole pancreas was resected from each mouse and divided into splenic and duodenal regions.Splenic re- gions were weighed, minced, and homogenized in acid-ethanol solution (75% ethanol, 23.5% distilled water, 1.5% conc.HC1).After overnight incubation at 4C the suspensions were cen- trifuged, and the supernatants were collected and assayed for insulin content.

Measurement of Norepinephrine (NE) Turnover
The effect of BDNF on norepinephrine turnover was assessed using a slightly modified version of the method previously reported by Collins. [16]/db mice received intracerebroventricular ad- ministration of either BDNF (15 txg/mouse) or ve- hicle at the beginning of a dark cycle, and then food was removed.Two hours after BDNF or vehi- cle treatment, c-methyl-p-tyrosine methyl ester (250mg/kg, Sigma, St. Louise, MO), an inhibitor of tyrosine hydroxylase, was intraperitoneally injected to block de novo catecholamine synthesis.The mice were decapitated two hours after c-MT injection and the interscapular brown adipose tissue (BAT) was immediately dissected, weighed and then frozen in liquid nitrogen.The BAT was homogenized with 0.1N perchloric acid containing 5 mM EDTA.Homogenates were filtrated through a 0.22 xm mesh membrane to remove debris.Nore- pinephrine content in homogenates was measured using an HPLC system (LC-10A, Shimadzu Instru- mentation, Kyoto, Japan) equipped with a column (CA-5DS, Eicom, Kyoto, Japan).
Northern Blot Analysis db/db mice were intracerebroventricularly injected with either BDNF or vehicle at the beginning of a dark cycle and then food was removed.Animals were sacrificed 4 hours after BDNF (15 Ixg/mouse) or vehicle treatment; interscapular BAT was ex- cised and frozen immediately.RNA was prepared from the tissues with Trizol (Gibco BRL Life Tech- nologies, Rockville, MD, USA) using the manufac- turer's protocol.Yield and purity of RNA were determined by spectrophotometric absorption analysis at 260/280nm.3g of total RNA was electrophoresed in a 1% agarose gel containing formaldehyde and then transferred to GT probe membranes (Bio-Rad Laboratories, Hercules, CA, USA).A 1071-base pair rat uncoupling protein-1 (UCP1) probe (nucleotides 84-1154 in Genebank accession no.Ml1814) was obtained by reverse transcriptase-polymerase chain reaction (RT-PCR) from rat BAT RNA using primers 5'-CCA CAG GAA TTC GAA GTT GAG AGT TCG GTA and 5'- CCC AGC TCT AGA GCC CAG CAT AGG AGC CCA as reported previously. [71A 349-base pair mouse ]3-actin probe (nucleotides 728-1076 in Genebank accession no.M12481) was obtained by RT-PCR from mouse liver RNA using primers 5'- TGG AAT CCT GTG GCA TCC ATG AAA C and 5'-TAA AAC GCA GCT CAG TAA CAG TCC G.All probes were verified by sequencing.Probes were randomly labeled using a BcaBest labeling kit (Takara, Ohtsu, Japan) with [c-32p]-deoxy CTP (Amersham Pharmacia Biotech, Bucking- hamshire, England).Hybridization was carried out at 65C in 0.25M sodium phosphate (pH 7.2)/7% SDS, and blots were washed twice with 20mM sodium phosphate (pH 7.2)/5% SDS and then with 20mM sodium phosphate (pH 7.2)/1% SDS.Hybridization signals were quanti- fied using a bio-imaging analyzer BAS2000 (Fuji Photo Film, Tokyo, Japan).

Statistical Analysis
All data are presented as means + SD.The statis- tical calculations were performed using SAS software (SAS Institute, Cary, NC), and differ- ences between individual groups were analyzed by the Student's t-test, the Dunnett's test or Jonckheere-Terpstra test.P < 0.05 was considered statistically significant.

Effect of Intracerebroventricular Administration of BDNF on Glucose
Metabolism 15tg BDNF per mouse or the same volume (31/ mouse) of vehicle solution was administered in- tracerebroventricularly to db/db mice every other day 5 times.After intracerebroventricular BDNF administration, the food intake of db/db mice de- creased as shown in Figure 1A.Food intake of the vehicle-treated pellet pair-fed db/db mice was very well synchronized to the BDNF-treated mice.Compared with such vehicle-treated con- trol mice, the repetitive intracerebroventricular Effects of intracerebroventricular BDNF adminis- tration on food intake (A), blood glucose concentration (B) and body weight (C) in db/db mice.BDNF (151g/mouse) or vehi- cle was administered on alternate days to db/db mice housed in the pellet pair-feeding apparatus.Data are presented as mean +_ SD (n =9).**P < 0.01 vs. vehicle by Student's t-test.administration of BDNF significantly lowered blood glucose concentrations in db/db mice (Fig. 1B).There was no significant difference in body weight between BDNF-treated and the pellet pair- fed mice (Fig. 1C).To study the dose-dependency of repetitive intracerebroventricular administra- tion, three different doses (0.15, 1.5 and 15g/mouse) of BDNF were injected every other day to db/db mice, respectively.BDNF was found to be dose-dependently effective in lowering blood glucose concentration and reducing food in- take of db/db mice by Jonckheere-Terpstra test (blood glucose; P =0.002, food intake; P 0.009). 15ig/mouse of BDNF significantly reduced food intake and lowered blood glucose concentration (Figs.2A, B).
In addition to blood glucose, we next analyzed the effect of intracerebroventricular administra- tion of BDNF on plasma insulin levels.As shown in Table I, plasma insulin concentrations of both (days) FIGURE 2 Dose-response effects of intracerebroventricular BDNF administration on food intake (A) and blood glucose concentration (B) in db/db mice.BDNF (0.15, 1.5, 15 bg/mouse) or vehicle was administered on alternate days to ad libitum-fed db/db mice.Data are presented as mean SD (n 7 or 8).*P < 0.05, **P 0.01 vs. vehicle by Dunnett's test.BDNF was found to be dose-dependently effective in lowering blood glucose concentration and reducing food intake of db/db mice by Jonckheere-Terpstra test (blood glucose; P=0.002, food intake; P 0.009).

TABLE
Effect of intracerebroventricular BDNF adminis- tration on plasma insulin concentration and pancreatic in- sulin content in db/db mice.BDNF (15g/shot) or vehicle was administered on alternate days for a total of five injec- tions to db/db mice housed in the pellet pair-feeding appa- ratus.Plasma insulin concentration was measured at Days 0 (baseline) and 10.Pancreases were removed from the mice at Day 10 and the pancreatic insulin content was measured.Data are presented as mean SD (n 9) Day the BDNF-treated db/db mice and pellet pair-fed db/db mice decreased during the experimental period.However, the plasma insulin concentration of BDNF-treated db/db mice tended to be lower than that of the pellet pair-fed mice after repetitive intracerebroventricular administrations.Since we have previously found that subcutaneous admin- istration of BDNF increases pancreatic insulin con- tents of db/db mice, we then analyzed pancreatic insulin content after repetitive intracerebroventri- cular administration of BDNE The pancreatic in- sulin content of BDNF-treated mice was found to be approximately 6-fold higher than that of the pellet pair-fed mice (Tab.I).These findings sug- gest that intracerebroventricular administration as well as subcutaneous administration of BDNF regulates glucose metabolism in a similar fashion.

Temperature
To verify our hypothesis that BDNF regulates glucose metabolism by acting through the brain, we analyzed the effect of intracerebroventricular BDNF administration on the rectal temperature of db/db mice in this study.Compared with ad libitum-fed db/db mice (approximately 37-38C), the rectal temperature of the vehicle-treated pellet pair-fed db/db mice was lower, probably due to the reduced food intake that was synchronized with BDNF-treated mice (Tab.II).The rectal tempera- ture of the BDNF-treated db/db mice at Days i and 10 was significantly higher than the vehicle- treated pellet pair-fed db/db mice and almost com- parable to ad libitum-fed mice in spite of a reduced food intake that was approximately the same as the pair-fed mice.We then examined the skin' tem- perature of these db/db mice by thermography imaging analysis (Fig. 3).Whereas the skin tem- perature of a typical vehicle-treated pellet pair-fed mouse was lower than an ad libitum-fed mouse, the skin temperature of the paired BDNF-treated mouse recovered.A relatively higher temperature was observed in the interscapular region of the BDNF-treated mouse suggesting enhancement of thermogenesis in the brown adipose tissue (BAT).
TABLE II Rectal temperature of db/db mice with intracere- broventricular BDNF administration.BDNF (15g/shot) or vehicle was administered on alternate days for a total of five injections to db/db mice housed in the pellet pair-feeding appa- ratus.Rectal temperatures were measured at the next day af- ter the first injection (Day 1) and two days after the last (fifth) injection (Day 10).Data are presented as mean SD (n=9) Day Following a single intracerebroventricular administration of BDNF (15g/mouse), the tyrosine hydroxylase inhibitor a- methyl-p-tyrosine (a-MT) was intraperitoneally injected.NE contents were measured 2hr after injecting the blocking reagent.NE turnover was determined from the decrease in NE content after blockage of catecholamine biosynthesis.
Vehicle BDNF ii.:.,,.......i-.,.,:.o..i . .BDNF (15g/mouse) or vehicle was administered intra- cerebroventricularly to db/db mice and then fasting was started.Four hours after BDNF or vehicle administration, the interscapular brown adipose tissue was dissected and total RNA was isolated.Northern blot analysis (UCP1 and fi-actin probes) of the total RNA (3g) was then performed.Data are shown as means SD (n 4).*P <0.05 vs. vehicle by Student's t-test.The left panels show representative blots.administration of BDNF on the expression of un- coupling protein (UCP)-I gene in BAT.db/db mice were intracerebroventricularly adminis- trated with either BDNF (15g/mouse) or vehi- cle followed by food removal.Four hours after BDNF or vehicle administration, total RNA was prepared from BAT and subjected to Northern blot analysis using UCP-1 cDNA as a probe.As shown in Figure 5, a single intracerebroventricu- lar administration of BDNF increased UCP-1 mRNA in BAT by 1.5-fold.

DISCUSSION
We have previously shown that peripheral subcu- taneous administration of BDNF lowered food in- take and blood glucose concentration of diabetic db/db mice with accompanying obesity and hyperinsulinemia.[12'131 We have also demonstrated the hypoglycemic effect of BDNF on db/db mice even under strict pellet pair-feeding conditions using our novel apparatus [141 which indicates that blood glucose is actually being lowered by BDNF apart from the hypoglycemic effect ascribed to hypophagia.Since BDNF does not lower blood glucose levels of normal rodents and streptozo- tocin-treated rodent models, it is unlikely that INTERNATIONAL JOURNAL OF EXPERIMENTAL DIABETES RESEARCH BDNF enhances insulin secretion from the pancreas.[12,141 In streptozotocin-treated mice, we have found concomitant administration of BDNF with insulin enhances the acute hypoglycemic ef- fect of insulin.I141 These data suggest that peripheral subcutaneous administration of BDNF enhances insulin sensitivity or ameliorates insulin resistance or both in peripheral tissues.
In our studies so far with cultured adipocyte and myotubule cell lines we have observed no direct action of BDNF on insulin stimulated 2-deoxyglucose uptake (unpublished data), al- though peripheral tissues such as adipose tissue and muscle are involved in insulin-dependent glucose metabolism.Moreover, it was reported elsewhere that intracerebroventricular infusion of BDNF suppresses food intake and body weight gain but does not affect blood glucose level in normoglycemic (Long-Evans) rats. [181Therefore, we investigated the effect of central administra- tion of BDNF on glucose metabolism in hyper- glycemic animals.Our present study clearly demonstrated that central BDNF administration reduces blood glucose and also increases pancre- atic insulin contents in obese hyperglycemic db/db mice under strict pellet pair-feeding conditions.
In comparison with subcutaneous administration, a much lower dose (approximately 1/100) of BDNF was found to be effective with central ad- ministration.These results indicate that BDNF regulates glucose metabolism and maybe pancre- atic function through the central nervous system.
Previously, we have found that subcutaneous administration of BDNF raised rectal and skin temperatures in db/db mice, [141 indicating the regulatory role of BDNF on energy metabolism.In the present study we were also able to repro- duce these effects of BDNF through central ad- ministration.It is well known that the sympathetic nervous system is involved in regulating thermo- genesis and maintaining body temperature in mammals.I191 In this study, we demonstrated that central administration of BDNF rapidly enhances NE turnover in thermogenic brown adipose tissue (BAT) of db/db mice.This is consistent with the present thermographic data in which skin tem- perature increased in the interscapular region that abundantly contains BAT.BAT is a major source of non-shivering thermogenesis in rodents [21 and the thermogenic ability of BAT is thought to be due to UCP-1. [2]Intracerebroventricular adminis- tration of BDNF also rapidly increased UCP-1 mRNA expression in BAT of db/db mice demon- strating the central regulation of BDNF in energy expenditure.Taken together, the above indicates that BDNF regulates energy metabolism through the central nervous system and the auton'omic nervous system.
It is well known that the blood brain barrier restricts the transport of peptides and proteins between the blood and the brain. [221It has been reported however that BDNF passes through the blood brain barrier by a saturable transport sys- tem and quickly enters into brain. [23,24]In our preliminary experiments, subcutaneous treat- ment of BDNF as well as intracerebroventricular treatment rapidly showed anorexic effect in db/db mice (data not shown).Since subcutaneous administration of BDNF ameliorated the energy expenditure in our previous study, [14] BDNF ad- ministered even peripherally may rapidly enter the brain and regulate energy metabolism in obese diabetic animals.
The pharmacological profiles of BDNF shown in this study reminded us of leptin, an adipocyte- derived satiety hormone regulating body adi- posity by modulating food intake and energy metabolism.[8,25,26] Peripheral administration of leptin stimulates sympathetic nerve activity in in- terscapular BAT and norepinephrine turnover [6,27] and regulates the expression of UCP1 by modulat- ing the sympathetic nervous system. [28,29]Since the functional form of the leptin receptor (Ob-Rb) is expressed in the hypothalamus, a major site of metabolic regulation by the autonomic nervous system, I31 intracerebroventricular or intrahypothalamic administration of leptin can reproduce most of the effects of peripheral leptin administra- tion.[26'31'32] Leptin administered peripherally may access the hypothalamus via receptor-mediated transport, and regulate energy expenditure and food intake primarily by interacting with Ob-Rb in the hypothalamus.Therefore, many of the thera- peutic profiles of leptin are very similar to BDNE Since the functional full-length form of BDNF re- ceptor, trkB is expressed in the hypothalamus, [33,34] it is plausible that BDNF may act via the hypothalamic neuronal system.More studies will be needed to clarify the action mechanisms of BDNF in comparison with leptin.
In conclusion, the present study has demon- strated that central administration of BDNF regulates the glucose metabolism and energy expenditure of obese diabetic animals in a simi- lar fashion to peripheral administration.These results further suggest that BDNF modulates sympathetic nerve activity through central regu- lation in the hypothalamus and affects energy expenditure.

FIGURE 5
FIGURE 5  Effects of a single intracerebro,entricular ad- ministration of BDNF on mRNA expression of uncoupling protein-1 in brown adipose tissue.BDNF (15g/mouse) or vehicle was administered intra- cerebroventricularly to db/db mice and then fasting was started.Four hours after BDNF or vehicle administration, the interscapular brown adipose tissue was dissected and total RNA was isolated.Northern blot analysis (UCP1 and fi-actin probes) of the total RNA (3g) was then performed.Data are shown as means SD (n 4).*P <0.05 vs. vehicle by Student's t-test.The left panels show representative blots.
NE turnover) in BAT of a BDNF-treated db/db mouse.To assess NE turnover, NE contents in BAT were measured after blocking catecholamine synthesis with administration of a-methyl-p-tyrosine (a-MT), a tyrosine hydroxylase inhibitor.Two hours prior to c'MT administration, db/db mice received a single intracerebroventricular adminis- tration of either BDNF or vehicle.After adminis- tration of a-MT there was a decrease in the NE contents of the BAT in db/db mice that received ve- hicle intracerebroventricularly, indicating a block- age of catecholamine synthesis.Compared with such control animals, intracerebroventricular ad- ministration of BDNF elicited a larger reduction in NE contents in interscapular BAT, indicating en- hancement of NE turnover (Fig.4).
a BDNF (15g/mouse, on alternate days)-treated mouse (a), a vehicle-treated mouse pair-fed to the BDNF-treated mouse (b), and a vehicle-treated mouse fed ad libitum (c).Analysis was performed at Day10.