Nature © Macmillan Publishers Ltd 1997

letters to nature

374 NATURE | VOL 389 | 25 SEPTEMBER 1997

we observe periods of both eastward and westward differentialrotation of the inner core, with a period of oscillation of severalthousand years. After we imposed the viscous no-slip boundarycondition, and following an initial transient adjustment, we haveobserved only eastward differential rotation of the inner core,suggesting that viscous effects may also be important in thedynamics of the inner core in the Glatzmaier–Roberts dynamomodel4, probably indirectly through their strong role in determin-ing the flow in the outer core, as suggested by the flow patterns inFig. 3.

Much remains to be done to understand the geodynamo. Inparticular, we need to undertake a systematic exploration of theeffect of varying the non-dimensional parameters in order tocharacterize the dynamics of dynamo action. Increasing our under-standing of the dynamo process might allow us to develop asimplified model in which, for example, non-axisymmetric effectscould be parametrized, thereby permitting numerical studies of thefield evolution on much longer timescales. M

Received 12 March; accepted 15 July 1997.

1. Glatzmaier, G. A. & Roberts, P. H. A three-dimensional convective dynamo solution with rotating andfinitely conducting inner core and mantle. Phys. Earth Planet. Inter. 91, 63–75 (1995).

2. Glatzmaier, G. A. & Roberts, P. H. A three-dimensional self-consistent computer simulation of ageomagnetic field reversal. Nature 377, 203–209 (1995).

3. Glatzmaier, G. A. & Roberts, P. H. An anelastic evolutionary geodynamo simulation driven bycompositional and thermal convection. Physica D 97, 81–94 (1996).

4. Glatzmaier, G. A. & Roberts, P. H. Rotation and magnetism of Earth’s inner core. Science 274, 1887–1890 (1996).

5. Kuang, W. & Bloxham, J. Numerical modelling of magnetohydrodynamic convection in a rapidlyrotating spherical shell I: Weak and strong field dynamo action. J. Comp. Phys. (submitted).

6. Bullard, E. C. & Gellman, H. Homogeneous dynamos and terrestrial magnetism. Phil. Trans. Soc.Lond. A 247, 213–278 (1954).

7. Taylor, J. B. The magnetohydrodynamics of a rotating fluid and the Earth’s dynamo problem. Proc. R.Soc. Lond. A 274, 274–283 (1963).

8. Jault, D., Gire, C. & LeMouel, J.-L. Westward drift, core motions and exchanges of angular momentumbetween core and mantle. Nature 333, 353–356 (1988).

9. Jackson, A., Bloxham, J. & Gubbins, D. in Dynamics of the Earth’s Deep Interior and Earth Rotation Vol.72 (eds LeMouel, J.-L., Smylie, D. & Herring, T.) (Geophys. Monog., Am. Geophys. Union, 1993).

10. Zatman, S. & Bloxham, J. Torsional oscillations and the magnetic field within the Earth’s core. Nature388, 760–763 (1997).

11. Braginsky, S. I. An almost axially symmetrical model of the hydromagnetic dynamo of the earth, I.Geomag. Aeron. 15, 149–156 (1975).

12. Hollerbach, R. & Jones, C. A. A geodynamo model incorporating a finitely conducting inner core.Phys. Earth Planet. Inter. 75, 317–327 (1993).

13. Hollerbach, R. & Jones, C. A. Influence of the Earth’s inner core on geomagnetic fluctuations andreversals. Nature 365, 541–543 (1993).

14. Bloxham, J. & Jackson, A. Time-dependent mapping of the magnetic field at the core–mantleboundary. J. Geophys. Res. 97, 19537–19563 (1992).

15. Bloxham, J. The steady part of the secular variation of the Earth’s magnetic field. J. Geophys. Res. 97,19565–19579 (1992).

Acknowledgements. We thank G. Glatzmaier and P. Roberts for many helpful discussions regarding theirwork and for providing the data sets that were used to prepare Figs 2 and 3, and P. Olson for criticallyreviewing the manuscript. This work was supported by the David and Lucile Packard Foundation and bythe NSF.

Correspondence should be addressed to J.B.

Acutestimulationof glucosemetabolism inmiceby leptintreatmentSeika Kamohara*§, Remy Burcelin†§, Jeffrey L. Halaas*,Jeffrey M. Friedman*‡ & Maureen J. Charron†

* Laboratory of Molecular Genetics, and ‡ Howard Hughes Medical Institute,the Rockefeller University, 1230 York Avenue, New York, New York 10021, USA† Department of Biochemistry, Albert Einstein College of Medicine,1300 Morris Park Avenue, Bronx, New York 10461, USA§ These authors contributed equally to this work.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Leptin is an adipocyte hormone that functions as an afferentsignal in a negative feedback loop regulating body weight1–4, andacts by interacting with a receptor in the hypothalamus and othertissues5,6. Leptin treatment has potent effects on lipid metabolism,and leads to a large, specific reduction of adipose tissue mass after

several days1,4. Here we show that leptin also acts acutely toincrease glucose metabolism, although studies of leptin’s effecton glucose metabolism have typically been confounded by theweight-reducing actions of leptin treatment, which by itself couldaffect glucose homoeostasis1–3. We have demonstrated acute invivo effects of intravenous and intracerebroventricular adminis-trations of leptin on glucose metabolism. A five-hour intravenousinfusion of leptin into wild-type mice increased glucose turnoverand glucose uptake, but decreased hepatic glycogen content. Theplasma levels of insulin and glucose did not change. Similar effectswere observed after both intravenous and intracerebroventricularinfusion of leptin, suggesting that effects of leptin on glucosemetabolism are mediated by the central nervous system (CNS).These data indicate that leptin induces a complex metabolicresponse with effects on glucose as well as lipid metabolism.This response is unique to leptin, which suggests that new efferentsignals emanate from the CNS after leptin treatment.

The effects of leptin on glucose metabolism were studied afterintravenous (IV) or intracerebroventricular (ICV) infusions intowild-type mice. Three groups were analysed: a control group, whichwas treated with phosphate buffered saline (PBS) both IV and ICV;an IV-leptin group was infused with leptin IV and PBS ICV; and anICV-leptin group received leptin ICV and PBS IV (Fig. 1). Thecannulation of the C57BL/6J wild-type mouse third ventricle andthe insertion of Alzet osmotic pumps were performed one weekbefore implantation of an intravenous catheter. ICV infusion of alow dose of leptin (5 ng h−1) was accomplished by changing thepump to a leptin-filled pump two days before experimentalmeasurements, which is the time needed to fill the dead space inthe tubing (see methods). When the ICV pumps were changed, in-dwelling catheters were implanted into the left femoral vein of themice. Two days later, leptin (or PBS) was administered IV to freelymoving mice as a 5-h intravenous infusion (1 mg h−1). The IV leptininfusion began 2 h after food was removed (Fig. 1).

Several measures indicative of the state of glucose metabolismwere made 3–5 h after leptin infusion was begun. When the infusionwas completed (after a 7-h fast), the weight of treated and untreatedmice was unchanged (Table 1). The plasma leptin concentration inthe control mice was 3.0 ng ml−1 (control), compared with33.4 ng ml−1 for the IV-leptin mice and 4.8 ng ml−1 for the ICV-leptin mice. The plasma glucose, glucagon and insulin concentra-tions of the leptin-treated mice were not significantly different fromthose of the control mice (Table 1). However, although insulin levelswere slightly lower in the leptin groups (0.8 ng ml−1 for the IV-leptinmice and 1.1 ng ml−1 for the ICV-leptin mice, against 1.5 ng ml−1 forthe controls), glucose turnover was significantly increased afterleptin treatment (23.7 (IV-leptin), 21.8 (ICV-leptin) and 15.0 mgper kg per min−1 (control) (Fig. 2, Table 1)).

These data suggest that leptin regulates both glucose output andglucose uptake, independent of increases in plasma insulin. Asignificant decrease in glycogen content in the liver (93:2 6 7:0(control), 59:5 6 5:2 (IV-leptin), 63:2 6 10:6 per (ICV-leptin);Table 1) was observed after leptin treatment, suggesting that glucoseoutput from the liver was increased in the leptin-treated animals.The surgical manipulation resulted in a slight decrease in liverglycogen relative to fasted animals (109:8 6 6:7 versus 93:2 6 7:0 mgper mg), but this difference was not significant. An increasedcontent of glycogen in skeletal muscle was also observed in theICV-leptin group (1:4 6 0:1 versus 2:4 6 0:1 mg per mg; Table 1).

Glucose uptake into several tissues was measured by administer-ing 14C-2-deoxyglucose. A significant increase in glucose uptake wasobserved in skeletal muscle and brown adipose tissue in the leptin-treated mice (Fig. 3a). Small increases in glucose uptake wereobserved in other tissues, but these were not significant. Consistentwith this, leptin also resulted in a twofold increase in the whole-animal rate of glycolysis as scored by the conversion of D-3-3Hglucose to 3H water (10:2 6 1:6 (control) versus 17:3 6 2:3 (IV-leptin)

Nature © Macmillan Publishers Ltd 1997

letters to nature

NATURE | VOL 389 | 25 SEPTEMBER 1997 375

or 20:5 6 0:2 mg per kg per min (ICV-leptin); Table 1), indicatingthat most of the labelled glucose was oxidized rather than being usedas a substrate for lactate, glycogen or lipid biosynthesis.

We found that low-dose infusion of leptin ICV resulted in thesame range of effects as seen with IV administration, but withgreater potency. This finding suggests that the efferent signals fromthe CNS that modulate glucose metabolism are activated by leptin.The possibility that neuronal signals are responsible for the effect ofICV leptin on glucose uptake was tested by comparing glucoseuptake in denervated and intact muscle in the same animal. Leptin-stimulated glucose uptake was decreased in denervated leg in bothextensor digitorum longus (EDL, fast-twitch glycolytic fibres) andsoleus (slow-twitch oxidative fibres) relative to the intact muscle(Fig. 3b, Table 1). These data suggest that the effects of leptin onglucose metabolism in skeletal muscle are dependent, in part, uponneuronal signals.

Leptin was unable to stimulate glucose uptake by EDL when theleg was denervated. However, glucose uptake in denervated soleusmuscle after leptin treatment, although lower than that of the intactleg, was greater than in PBS-treated controls (Fig. 3b, Table 1). Thisresidual increase in glucose uptake by denervated soleus could bethe result of hormonal factors. Glucose uptake can also be modu-lated by thyroid hormone, which was measured in leptin-treatedanimals. Triiodothyronine (T3) levels were slightly higher (62.7versus 52.3 ng dl−1) and thyroxine (T4) levels were lower (0.5 versus0.7 mg dl−1) in the IV-leptin group than in PBS-treated controls,although those data were not significant. ICV leptin administrationsignificantly increased the level of T3 (102.5 versus 52.3 ng dl−1,P , 0:05) and decreased T4 (0.3 versus 0.7 mg dl−1, P, 0.05), suggest-

ing that there was increased de-iodination of T4 to T3 rather than anet increase in production.

Several lines of evidence have suggested that leptin may modulateglucose metabolism1–3. Leptin treatment of genetically obese ob/obmice improves diabetes at doses below those affecting weight loss1–3.Leptin treatment of lean mice does not change blood glucose, but adecrease in hepatic glycogen has been observed7. Recent studiesusing cultured hepatoma cells have also suggested that leptin mayinhibit insulin action in vitro8. Finally, insulin has been shown toalter leptin gene expression9. In their studies, however, the effect of

Leptin or PBS (ICV or IV)

0 2 h 5 h 7 h 7 h 40min

D-[3-3H] glucose

U-14C-2-Deoxyglucose injection

Foodremoval

Blood sampling

- 7 days - 2 days

ICV pump implantation IV catheterization

Figure 1 Experimental protocols using C57BL/6J +/+

lean mice. An ICV catheter was implanted at day −7.

At day −2, the ICV osmotic pump was replaced with

either leptin or PBS. At the same time, the femoral vein

was catheterized and the leg denervated. Leptin was

infused either IV or ICV for 5 h 40min. After 3 h

equilibration of leptin infusion, a primed continuous

infusion of D-3-3H-glucose was performed to assess

glucose turnover. Blood was sampled and analysed as

described in the Methods. To assess individual tissue

glucose uptake, a flash injection of U-14C 2-deoxyglu-

cose was performed.

Table 1 Characteristics of leptin-treated and control mice

Control IV-leptin ICV-leptin...................................................................................................................................................................................................................................................................................................................................................................

No. of mice 8 7 6Body weight (g) 25:4 6 0:5 25:7 6 0:8 26:6 6 0:9Basal plasma glucose (mgdl−1) 135 6 11 124 6 11 125 6 13Plasma glucose after 5-h fast (mgdl−1) 68 6 6 61 6 5 55 6 4Plasma glucose after treatment (mgdl−1) 70 6 8 61 6 5 51 6 3Lactate (mgdl−1) 70:1 6 5:0 63:8 6 8:0 82:6 6 3:3Plasma free fatty acids (mEq l−1) 547:1 6 61:5 599:9 6 53:0 511:1 6 15:3b-Hydroxybutyrate (mgdl−1) 4:7 6 0:7 3:6 6 0:7 3:6 6 0:3Leptin (ngml−1) 3:0 6 0:5 33:4 6 6:0 4:8 6 0:5Insulin (ngml−1) 1:5 6 0:4 0:8 6 0:2 1:1 6 0:2Glucagon (pgml−1) 254:4 6 44:2 282:1 6 52:9 252:2 6 34:4T3 (ngdl−1) 52:3 6 9:4 62:7 6 12:1 102:5 6 20:8*T4 (mgdl−1) 0:7 6 0:2 0:5 6 0:1 0:3 6 0:1*Glucose turnover (mg per kg per min) 15:0 6 1:2 23:7 6 1:5*** 21:8 6 1:0**Liver glycogen (mg per mg) 93:2 6 7:0 59:5 6 5:2** 63:2 6 10:6*Glycolysis (mg per kg per min) 10:2 6 1:6 17:3 6 2:3* 20:5 6 0:2**...................................................................................................................................................................................................................................................................................................................................................................

Glucose uptake (ng per mg per min)EDL 1:57 6 0:43 4:13 6 1:03* 6:46 6 1:79*Denervated EDL 3:05 6 0:77 3:87 6 0:75 3:87 6 0:87Soleus 1:86 6 0:56 4:32 6 1:22* 11:43 6 2:48*Denervated soleus 3:86 6 1:07 8:35 6 2:39 6:02 6 1:39Hindlimb glycogen (mg per mg) 1:4 6 0:1 1:9 6 0:3 2:4 6 0:1*...................................................................................................................................................................................................................................................................................................................................................................We analysed 6–8 C57BL/6J wild-type mice per group. Plasma glucose was assessed before, during and after leptin and/or PBS infusion. Plasma hormone and metabolite levels wereassessed at the completion of the infusions. The rate of glycolysis and glucose turnover were calculated as described in the Methods. EDL and soleus U-14C 2-deoxyglucose uptake wasanalysed from intact and denervated legs, and glycogen content was analysed from non-denervated leg on completion of study. Asterisks indicate statistical significance: (*P , 0:05,**P , 0:005, ***P , 0:001) between IV or ICV leptin-infused mice and PBS-infused control mice.

ICV-leptin10

15

20

25

30

Control IV-leptin

**

Glu

cose

tur

nove

r(m

g p

er k

g p

er m

in)

Figure 2 Glucose turnover in C57BL/6J +/+ lean mice after IV or ICV leptin

infusion. Glucose turnover was assessed after infusions of leptin or PBS. Leptin

increased glucose turnover from 15:0 6 1:2 to 23:7 6 1:5mg per kg per min

(P , 0:001) with IV, or to 21:8 6 1:0mg per kg per min (P , 0:005) with ICV

treatment. Asterisks indicate statistical significance between leptin-treated and

control mice; 6–8 mice per group were studied.

Nature © Macmillan Publishers Ltd 1997

letters to nature

376 NATURE | VOL 389 | 25 SEPTEMBER 1997

leptin on glucose metabolism in vivo have not been carefullyevaluated.

Our data show that leptin affects glucose metabolism. Althoughplasma glucose levels did not change in response to leptin, glucoseuptake and oxidation were increased. These data suggest that, inaddition to its effects on lipid metabolism, the negative energybalance in leptin-treated animals is also associated with increasedmetabolism of glucose. The increased glucose turnover in theabsence of a significant change in insulin concentration suggeststhat the leptin-mediated increase in glucose metabolism is not likelyto be the result of an increased rate of insulin secretion in C57BL/6J+/+ mice. It could be the result of an increase in insulin sensitivity,although it remains possible that leptin treatment results inincreased glucose uptake through an insulin-independent mechan-ism. The latter possibility is suggested by the observed increase inthe rate of glycolysis, and further studies are required to resolve this.The increase in glycolysis is also independent of glucagon, as thelevels of this hormone were also unchanged after leptin treatment.The increase in glucose turnover while the plasma glucose level wasunchanged suggests that leptin might also increase glucose output,perhaps through the liver. This idea is consistent with our data andwith previous reports showing that chronic treatment with leptindecreases hepatic glycogen content7.

The data from mice receiving the IV leptin infusion indicate thatleptin acutely alters glucose metabolism. Low-dose ICV infusion ofleptin resulted in the same response as IVadministration, suggestingthat the observed effects of peripheral leptin on glucose metabolismare likely to be the result of transport into the CNS and interactionwith receptors there5,6,10,11. Several other lines of evidence suggestthat the hypothalamus is an important target of both ICV and IVleptin. Leptin levels increase after hypothalamic lesions, and ratswith lesions in the ventromedial hypothalamus do not respond to

exogenous leptin12,13. The signalling form of the leptin receptor ishighly expressed in those hypothalamic nuclei previously shown tocontrol the regulation of food intake and body weight5,6,10,14. More-over, the arcuate, ventromedial hypothalamus and lateral hypotha-lamus (LH) signalling form of the receptor is not expressed to agreat extent in any other brain region14. The activity of the Stat3transcription factor and fos gene expression are increased within 1 hof a single dose of IV leptin15. Finally, the 100-fold higher potency ininducing weight loss of ICV leptin than subcutaneous leptin4 andthe identification of leptin in cerebrospinal fluid are consistent withthe main site of action of leptin being in the CNS16,17.

The effects of ICV leptin on glucose uptake indicates that thisresponse is the result of a stimulation of efferent pathways from theCNS. Numerous studies have suggested that the CNS, in particularthe hypothalamus, can regulate glucose metabolism18–21. Thesestudies also suggest that alterations in sympathetic tone maymediate this effect22. The stimulatory effect of leptin on glucoseoutput and glucose uptake that we observed is similar to that seenduring exercise23 or in response to ICV carbachol18. The possibilityof an overlap between the stimulatory pathways mediating theaction of leptin on carbohydrate metabolism and that of carbacholor exercise is intriguing, and requires further investigation.

Our data further suggest that some of the effects of leptin aremediated through neural signals. Previous data have shown thatleptin treatment increases sympathetic output to brown fat throughthe activation of adrenergic receptors22. The data do not exclude thepossibility that hormonal factors also have an effect. We have shownthat ICV leptin leads to elevated T3 and decreased T4, suggestingthat leptin treatment increases the rate of conversion of T4 to T3.The activity of the de-iodenase that converts T4 to T3 is also knownto be modulated by the autonomic nervous system. In addition,thyroid hormone is known to increase basal glucose uptake by

WAT

*

0

5

10

15

20

25

skin brain kidney lung liver EDL

a

ICV-leptin

control

*

IV-leptin

Upta

ke o

f 2D

G (ng p

er m

g p

er m

in)

heart BAT

120

100

80

60

40

20

0

Tissue

Figure 3 Individual tissue glucose uptake in C57BL/6J

+/+ lean mice after IV or ICV leptin infusion. a, U-14C 2-

deoxyglucose (2-DG) uptake was assessed after IV of

ICV infusions of leptin or PBS. The tissues tested

included skin, brain, kidney, lung, liver, extensor digi-

torum longus (EDL), periovarian white adipose tissue

(WAT), heart and inter-scapular brown adipose tissue

(BAT). A significant increase (asterisks) in glucose

uptake was observed in EDL (P , 0:05), and BAT

(P , 0:05) in leptin-infused animals; 6–8 mice per

group were studied. b, 2-DG uptake was assessed in

intact and denervated EDL and soleus after leptin or

PBS infusions. Asterisks indicate statistical signifi-

cance with P , 0:05 between leptin- and PBS-treated

mice; 6–8 mice per group were studied.

0

100

200

300

400

500

600

700

Upt

ake

of 2

DG

(%

of c

ontro

l)

EDL denervated EDL soleus denervated soleus

ICV-leptin

IV-leptin

Control

*

*

*

*

b

Tissue

Nature © Macmillan Publishers Ltd 1997

letters to nature

NATURE | VOL 389 | 25 SEPTEMBER 1997 377

various tissues24,25. Thus an increase in T3 levels could also accountfor a fraction of the leptin-induced increase in glucose uptake byother tissues . It has previously been shown that leptin can alsoincrease T4 levels in fasted mice26. The basis for this difference isunclear, although the length of the period of food restriction wasdifferent in these studies. The data do not exclude the possibilitythat other hormonal signals are also activated by leptin.

In summary, our in vivo studies indicate that acute administra-tion of leptin increases glucose metabolism independent of eleva-tions in plasma insulin by interacting with receptors in the CNS.These findings suggest that leptin induces a state of negative energybalance and increases metabolism of both fat and glucose. It is alsopossible that mechanisms to provide glucose to peripheral tissueshave evolved during leptin-mediated anorexia. M. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Methods

Experimental procedures. C57 BL6J +/+ lean mice were anaesthetized withketamine/xylazine. A 1-cm midline incision was made across the top of theskull, and the animal was placed on a stereotaxic apparatus and the periosteumcleaned. A hole 1 mm in diameter was made midline 0.3 mm behind thebregma, and a 30-gauge stainless steel infusion cannula was implanted into thethird ventricle using the following coordinates: 0.3 mm relative to bregma,midline 3 mm ventral. Two supporting screws 0.5 mm in diameter were placedbilaterally in the posterior quadrants of the skull and the cannula secured inplace with dental acrylic. The cannula was connected to 6 cm tygon tubing andan osmotic pump (Alzet, Palo Alto, CA) containing CSF placed at the other endof the tubing. The pump and tubing were placed in the subcutaneous spacedorsally and the skin sutured. The mice were observed for 8 days, with weight,food intake and behaviour being monitored daily to verify a total recovery. Thetime required for peptide solutions to reach the third ventricle was estimated byinfusing leptin or angiotensin, a peptide that results in a dramatic increase inwater intake (J.L.H. et al., unpublished data). The mice were anaesthetized 48 hbefore glucose metabolism was analysed and the osmotic pumps were replacedby new ones filled with PBS or leptin. At the same time an indwellingintravenous catheter was implanted. Briefly, a 5-mm vertical skin incision wasmade on a ventral side of the left leg at the level of the hip. A catheter wasimplanted into the vena cava through the femoral vein. A piece of the adjacentnerve was physically removed from the leg to insure denervation. The catheterwas sealed subcutaneously, made exterior at the base of the neck, and glued inplace with dental cement. There were three groups for this study: a controlgroup was treated with PBS both IV and ICV; an IV-leptin group was infusedwith leptin IV and PBS ICV; and an ICV-leptin group received leptin ICV andPBS IV. To avoid glucose released from the digestive tract, after 48 h the animalswere food-deprived for 2 h before the leptin or PBS infusions started (Fig. 1).Under these conditions the rate of glucose appearance into the blood representsthe rate of hepatic glucose production. To assess glucose turnover, a prime(2.33 mCi) continuous infusion (7 mCi per kg per min) of HPLC-purifiedD-3-3H-glucose (Amersham) was maintained for 120 min in unrestrained mice.At periods of 80, 90, 100, 110 and 120 min of D-3-3H glucose infusion, 10 ml ofblood was sampled through the tail vein, deproteinized by the Somogyiprocedure, and the specific activity of D-3-3H-glucose in the supernatant wasanalysed27. To analyse individual tissue glucose uptake, 0.5 mCi g−1 or 14C-2-deoxyglucose was flash-injected into the vein after the D-3-3H-glucose infusion.Tail blood was sampled at 2, 4, 6, 10, 20, 30 and 40 min after the injection. Theanimals were killed by cervical dislocation, the organs removed, and individualtissue glucose uptake calculated28.Plasma hormone and metabolite determinations. After the infusions,blood was drawn from the retro-orbital sinus using a heparinized micro-capillary tube and centrifuged. The plasma was collected immediately andstored at −70 8C. Plasma insulin, thyroid hormones and leptin levels weremeasured using rat insulin (Linco, St Louis, MO), thyroxine hormones(DPC, Los Angeles, CA) or mouse leptin (Linco, St Louis, MO) radioimmu-noassay kits, respectively. Plasma glucose, lactate and b-hydroxybutyrate weremeasured using the appropriate kits (Sigma, St Louis, MO). Plasma free fatty

acids were assessed (Amano, Richmond, VA) with oleic acid as standard.Muscle and liver glycogen content was determined28,29.Calculations. GT=gluIR/gluSA=HGP, where GT is glucose turnover (mgper kg per min), gluIR is glucose infusion rate (d.p.m.), gluSA is the specificactivity of D-3-3H-glucose in the blood (d.p.m. mg−1), and HGP is hepaticglucose production (mg per kg per min). Individual tissue glucose uptakeand glycolysis rates were calculated28,29. Results are presented as mean 6 s:e:Statistical significance was determined by the Kruskal-Wallis test formuscle glucose uptake determinations and Mann-Whitney tests for sig-nificance between groups, The Student’s t-test was used for other unpaireddata.

Received 2 June; accepted 18 July 1997.

1. Halaas, J. L. et al. Weight-reducing effects of the plasma protein encoded by the obese gene. Science 269,543–546 (1995).

2. Campfield, L. A., Smith, F. J., Guisez, Y., Devos, R. & Burn, P. Recombinant mouse OB protein:evidence for a peripheral signal linking adiposity and central neural networks. Science 269, 546–549(1995).

3. Pelleymounter, M. A. et al. Effects of the obese gene product on body weight regulation in ob/ob mice.Science 269, 540–543 (1995).

4. Halaas, J. L. et al. Physiological response to long-term peripheral and central leptin infusion in leanand obese mice. Proc. Natl Acad. Sci. USA 94, 8878–8883 (1997).

5. Tartaglia, L. A.. et al. Identification and expression cloning of a leptin receptor, OB-R. Cell 83, 1263–1271 (1995).

6. Lee, G. H. et al. Abnormal splicing of the leptin receptor in diabetic mice. Nature 379, 632–635 (1996).7. Levin, N., Nelson, C., Gurney, A., Vanelen, R. & De Sauvage, F. Decreased food intake does not

completely account for adiposity reduction after ob protein infusion. Proc. Natl Acad. Sci. USA 93,1726–1730 (1996).

8. Cohen, B., Novick, D. & Rubinstein, M. Modulation of insulin activities by leptin. Science 274, 1185–1188 (1996).

9. Saladin, T. et al. Transient increase in obese gene expression after food intake or insulin administration.Science 377, 527–529 (1995).

10. Mercer, J. G. et al. Localization of leptin receptor mRNA and the long form splice variant (Ob-R) inmouse hypothalamus and adjacent brain regions by in situ hybridization. FEBS Lett. 387, 113–116(1996).

11. Golden, P. L., Maccagnan, T. J. & Pardridge, W. M. Human blood-brain barrier leptin receptor. J. Clin.Invest. 99, 14–18 (1997).

12. Maffei, M. et al. Increased expression in adipocytes of ob RNA in mice with lesions of thehypothalamus and with mutations at the db locus. Proc. Natl Acad. Sci. USA 92, 6957–6960 (1995).

13. Satoh, N. et al. Pathophysiological significance of the obese gene product, leptin, in ventromedialhypothalamus (VMH)-lesioned rats: evidence for loss of its satiety effect in VMH-lesioned rats.Endocrinology 138, 947–954 (1997).

14. Fei, H. et al. Anatomic localization of alternatively spliced leptin receptors (Ob-R) in mouse brain andother tissues. Proc. Natl Acad. Sci. USA 94, 7001–7005 (1997).

15. Vaisse, C. et al. Leptin activation of Stat3 in the hypothalamus of wild type and ob/ob mice but notdb/db mice. Nature Genet. 14, 95–97 (1996).

16. Schwartz, M. W., Seeley, R. J., Campfield, L. A., Burn, P. & Baskin, D. G. Identification of targets ofleptin action in rat hypothalamus. J. Clin. Invest. 98, 1101–1106 (1996).

17. Caro, J. F. et al. Decreased cerebrospinal-fluid/serum leptin ratio in obesity: a possible mechanism forleptin resistance. Lancet 348, 159–161 (1996).

18. Miles, P. D. G., Yamatani, K., Lickley, H. L. A. & Vranic, M. Mechanism of glucoregulatory responsesto stress and their deficiency in diabetes. Proc. Natl Acad. Sci. USA 88, 1296–1300 (1991).

19. Leong, S. F. & Clark, J. B. Regional enzyme development in rat brain: Enzymes associated with glucoseutilization. Biochemistry 218, 131–138 (1984).

20. Lautala, P. & Martin, J. M. Glucose metabolism in rat hypothalamus. Acta Endocrinol. 98, 481–487(1981).

21. Nagai, K., Fujii, T., Inoue, S., Takamura, Y. & Nakagawa, H. Electrical stimulation of the suprachias-matic nucleus of the hypothalamus causes hyperglycemia. Hormone Metab. Res. 20, 37–39 (1988).

22. Collins, S. et al. Role of Leptin in fat regulation. Nature 380, 677 (1996).23. Vranic, M., Kawamori, R., Pek, S., Kovacevic, N. & Wrenshall, G. A. The essentiality of insulin and the

role of glucagon in regulating glucose utilization and production during strenuous exercise in dogs. J.Clin. Invest. 57, 245–255 (1976).

24. Gosteli-Peter, M. A., Schmid, C. & Zapf, J. Triiodothyronine increases glucose transporter isotype 4mRNA expression, glucose transport, and glycogen synthesis in adult rat cardiomyocytes in long-termculture. Biochem. Biophys. Res. Commun. 221, 521–524 (1996).

25. Weinstein, S. P., O’Boyle, E. & Haber, R. S. Thyroid hormone increases basal and insulin-stimulatedglucose transport in skeletal muscle. The role of GLUT4 glucose transporter expression. Diabetes 43,1185–1189 (1994).

26. Ahima, R. S. et al. Role of leptin in the neuroendocrine response to fasting. Nature 382, 250–252(1996).

27. Somogyi, M. Determination of blood sugar. J. Biol. Chem. 160, 69–73 (1945).28. Tsao, T. S., Burcelin, R., Katz, E. B., Huang, L. & Charron, M. J. Enhanced insulin action due to

targeted GLUT4 overexpression exclusively in muscle. Diabetes 45, 28–36 (1996).29. Massillon, D., Barzilai, N., Hawkins, M., Prus-Wertherimer, D. & Rossetti, L. Quantitation of hepatic

glucose fluxes and pathways of hepatic glycogen synthesis in conscious mice. Am. J. Physiol. 269,E1037–E1043 (1995).

Acknowledgements. We thank E. B. Katz, C. Vaisse, J. Li and T. S. Tsao for discussions; J. Blaire-West andD. A. Denton for help with the ICV surgery; S. Korres for helping to prepare the manuscript; and Amgenfor recombinant leptin. This work was supported by grants from the NIH (J.M.F. and M.J.C.), PewCharitable Trust (M.J.C.), Juvenile Diabetes Foundation International (R.B.), the Philippe Foundation(R.B.) and the Manpei Suzuki Diabetes Foundation (S.K.).

Correspondence and requests for materials should be addressed to M.J.C. (e-mail: charron@aecom.yu.edu).