Selective deletion of leptin receptor in neuronsleads to obesity

Paul Cohen, … , Peter Mombaerts, Jeffrey M. Friedman

J Clin Invest. 2001;108(8):1113-1121. https://doi.org/10.1172/JCI13914.

Animals with mutations in the leptin receptor (ObR) exhibit an obese phenotype that isindistinguishable from that of leptin deficient ob/ob mice. ObR is expressed in many tissues,including brain, and the relative importance of leptin’s effects on central versus peripheralsites has not been resolved. To address this, we generated mice with neuron-specific(ObRSynIKO) and hepatocyte-specific (ObRAlbKO) disruption of ObR. Among theObRSynIKO mice, the extent of obesity was negatively correlated with the level of ObR inhypothalamus and those animals with the lowest levels of ObR exhibited an obesephenotype. The obese mice with low levels of hypothalamic ObR also show elevatedplasma levels of leptin, glucose, insulin, and corticosterone. The hypothalamic levels ofagouti-related protein and neuropeptide Y RNA are increased in these mice. These dataindicate that leptin has direct effects on neurons and that a significant proportion, or perhapsthe majority, of its weight-reducing effects are the result of its actions on brain. To explorepossible direct effects of leptin on a peripheral tissue, we also characterized ObRAlbKOmice. These mice weigh the same as controls and have no alterations in body composition.Moreover, while db/db mice and ObRSynIKO mice have enlarged fatty livers, ObRAlbKOmice do not. In summary, these data suggest that the brain is a direct target for the weight-reducing and neuroendocrine […]

Article

Find the latest version:

http://jci.me/13914-pdf

IntroductionLeptin is an adipocyte hormone thatfunctions as an afferent signal in a nega-tive feedback loop that regulates energybalance (1–3). Mice with mutations inleptin (ob/ob) or its receptor (db/db) arehyperphagic and severely obese (4–6).These mutant mice also manifest a largenumber of metabolic and endocrineabnormalities including diabetes, hyper-cortisolemia, infertility, and cold intol-erance (7, 8). These ob/ob and db/db micealso have enlarged, steatotic livers (9, 10).The fatty liver is in part a result of anincreased rate of hepatic lipogenesis,which is also thought to contribute tothe development of obesity (11).

Treatment of ob/ob mice with leptinreduces food intake and body weightand corrects the metabolic andendocrine defects associated with lep-tin deficiency. Leptin treatment alsonormalizes the hepatomegaly andassociated elevations in hepatic glyco-gen and lipid (12). Infusions of leptininto wild-type mice in physiologicalamounts results in a dose-dependentreduction in food intake and bodyweight (13–15). This metabolicresponse to leptin in both ob/ob andwild-type mice is novel and is not sole-ly a consequence of its anorectic effects(16). Intracerebroventricular (ICV) lep-tin has similar effects, but at much

lower doses, suggesting that leptin hasdirect effects on brain (15, 17).

The leptin receptor, which has fivesplice variants (ObRa–e), is broadlyexpressed, however, and the relativeimportance of leptin’s effects on brainversus peripheral sites is unclear (18, 19).This distinction is of importancebecause leptin deficiency is associatedwith myriad abnormalities in many tis-sues including liver. Although leptin hasbeen suggested to act directly to activatesignal transduction and deplete triglyc-erides, the contribution of these actionsto body weight homeostasis, depositionof lipid in peripheral sites, and neu-roendocrine function in vivo is untested(20, 21). To determine the role of leptinaction in the central nervous system andthe periphery, we are systematicallydeleting ObR in a tissue-specific fashionusing the Cre-loxP system. Here we pres-ent data from mice with either neuronalor hepatocyte-specific deletions of ObR.

MethodsObR floxed mice. A 12-kb genomic clonecontaining the first coding exon of ObRwas isolated. A single loxP element wasPCR-amplified from the plasmidpNeoTKLox. The oligonucleotideprimers were engineered to introduce aBamHI-NcoI double restriction site atone end and a BamHI site at the otherend. After digestion with BamHI, theproduct was ligated into the BamHI siteof the genomic clone, resulting in theclone BNLOXB with a single LoxP siteupstream of the first exon. A loxP-Neo-HSV-TK-LoxP element was isolated bydigesting pNeoTKLox with BstXI andSalI. The fragment was blunt-ended andcloned into BglII digested blunt-ended

The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8 1113

Selective deletion of leptin receptor in neurons leads to obesity

Paul Cohen,1 Connie Zhao,1 Xiaoli Cai,1 Jason M. Montez,1 S. Christy Rohani,1

Paul Feinstein,2 Peter Mombaerts,2 and Jeffrey M. Friedman1,3

1Laboratory of Molecular Genetics,2Laboratory of Developmental Biology and Neurogenetics, and3Howard Hughes Medical Institute, The Rockefeller University, New York, New York, USA

Address correspondence to: Jeffrey M. Friedman, Howard Hughes Medical Institute, The Rockefeller University, 1230 York Avenue, Box 305, New York, New York 10021, USA. Phone: (212) 327-8800; Fax: (212) 327-7420; E-mail: friedj@mail.rockefeller.edu.

Paul Cohen and Connie Zhao contributed equally to this work.

Received for publication August 6, 2001, and accepted in revised form September 10, 2001.

Animals with mutations in the leptin receptor (ObR) exhibit an obese pheno-type that is indistinguishable from that of leptin deficient ob/ob mice. ObR isexpressed in many tissues, including brain, and the relative importance of lep-tin’s effects on central versus peripheral sites has not been resolved. To addressthis, we generated mice with neuron-specific (ObRSynIKO) and hepatocyte-spe-cific (ObRAlbKO) disruption of ObR. Among the ObRSynIKO mice, the extent ofobesity was negatively correlated with the level of ObR in hypothalamus andthose animals with the lowest levels of ObR exhibited an obese phenotype. Theobese mice with low levels of hypothalamic ObR also show elevated plasma lev-els of leptin, glucose, insulin, and corticosterone. The hypothalamic levels ofagouti-related protein and neuropeptide Y RNA are increased in these mice.These data indicate that leptin has direct effects on neurons and that a signifi-cant proportion, or perhaps the majority, of its weight-reducing effects are theresult of its actions on brain. To explore possible direct effects of leptin on aperipheral tissue, we also characterized ObRAlbKO mice. These mice weigh thesame as controls and have no alterations in body composition. Moreover, whiledb/db mice and ObRSynIKO mice have enlarged fatty livers, ObRAlbKO mice donot. In summary, these data suggest that the brain is a direct target for theweight-reducing and neuroendocrine effects of leptin and that the liver abnor-malities of db/db mice are secondary to defective leptin signaling in the brain.

This article was published online in advance of the print edition. The date of publication is available from the JCI website, http://www.jci.org.

J. Clin. Invest. 108:1113–1121 (2001). DOI:10.1172/JCI200113914.

Rapid Publication

BNLOXB. This clone, denoted ObR-lox-SS1-lox, was double digested with AflIIand KpnI, blunt-ended, and re-ligated togenerate the targeting vector. Theexcised 500-bp AflII-KpnI fragment wasused as probe 1. The targeting vectorwas linearized with SalI and electropo-rated into 129/SV embryonic stem (ES)cells. Cells were selected with G418, andsurviving clones were screened forhomologous recombination. Positiveclones were transiently transfected withthe Cre recombinase expressing plas-mid, p0G231 (kindly provided by S.O’Gorman) (22). Cells were then sub-jected to ganciclovir selection, and sur-viving clones were checked by Southernblotting using probe 2, which is withinexon 2, to confirm the deletion of theNeo-HSV-TK cassette (type II deletion).ES cells with the correct genotype (ObRflox/+) were injected into C57BL/6blastocysts, and resulting chimeras werebred with C57BL/6 mice to obtaingermline transmission. ObRflox/+ micewere crossed to generate the line of ObR-floxed mice, denoted ObRflox/flox.

ObR-null mice. ObRflox/+ mice werecrossed with transgenic adenovirusEIIA Cre mice (kindly provided by H.Westphal) (23). Given that Cre isexpressed early in embryogenesis in thisline, progeny were screened by South-ern blotting with probe 2 for germlineCre-mediated deletion of the first cod-ing exon (type I deletion). These mice,with the genotype ObR∆/+, were crossedto produce homozygous ObR∆/∆ mice,designated ObR-null mice.

Conditional deletion of ObR. Neuron-specific deletion of ObR was achievedusing synapsinI-Cre transgenic mice(SynI-Cre(+)) (kindly provided by J.Marth) (24, 25). Hepatocyte-specificdeletion was achieved using Albumin-cre transgenic mice (Alb-Cre(+)). TheAlb-Cre transgene (see Figure 5a) wasconstructed using plasmid NB, whichcontains 2 kb of the Albumin promoterenhancer (kindly provided by R.Palmiter) (26). A 2-kb fragment con-taining the Cre-recombinase gene anda nuclear localization and polyadeny-lation signal was excised from a sepa-rate plasmid by digesting with BglIIand blunt-ending and then digestingwith KpnI. This fragment was clonedinto KpnI-EcoRV digested plasmid NBdownstream of the albumin promoter

enhancer. The resulting plasmid wasdigested with NotI and KpnI to releasethe 4-kb transgene, which was gel puri-fied and injected into fertilized eggsfrom C57/BL6 × CBA (F1) mice to pro-duce transgenic mice. The presence ofthe transgene in both (SynI-Cre(+)) and(Alb-Cre(+)) mice was verified by PCRand Southern blotting. Tissue-specificknockout mice were generated by twosuccessive crosses (see Figure 2a). First,ObR∆/+ mice were crossed with eithersynapsinI-Cre transgenic mice (SynI-Cre(+)) or albumin-cre transgenic mice(Alb-Cre(+)) to generate ObR∆/+, SynI-Cre(+) mice or ObR∆/+, Alb-Cre(+) mice.These mice were then mated to ObRflox/flox mice to generate mice withthe genotype ObR∆/flox, SynI-Cre(+) orObR∆/flox, Alb-Cre(+), hereafter desig-nated ObRSynIKO and ObRAlbKO,respectively. Mice with the followinggenotypes were also generated: (a)ObR∆/flox, SynI-Cre(–) and ObR∆/flox,Alb-Cre(–) (heterozygotes); (b) ObR-flox/+, SynI-Cre(+) and ObRflox/+, Alb-Cre(+) (referred to as wild-type); and (c)ObRflox/+, SynI-Cre(–) and ObRflox/+,Alb-Cre(–) (referred to as wild-type). Allanimals were housed under controlledtemperature (23°C) and lighting (12hours of light; 12 hours of dark) withfree access to food and water. All pro-cedures were in accordance with theguidelines of the Rockefeller Universi-ty Laboratory Animal Research Center.

Assay of Cre specificity. Genomic DNAwas prepared from multiple tissuesfrom ObRflox/+ SynI-Cre(+) and ObRflox/+

Alb-Cre(+) mice, and 200 ng was PCR-amplified using primers flanking thefirst coding exon of ObR. A schematic ofthe primer locations is shown in Figure2b. In tissues expressing Cre recombi-nase, exon 1 is excised and a single loxPsite remains, generating an ObR∆ allele.Although primers 1 and 3 can amplify aproduct from the ObR∆ allele, theprimers are separated by too great a dis-tance for amplification to occur in theObRflox or ObR+ (wild-type) alleles. As acontrol, primers 1 and 2 were used toamplify both the ObRflox and ObR+ alle-les from all tissues. The ObR+ allele pro-duces a slightly smaller product owingto the absence of loxP sequence.

Assay of Cre efficiency. To determine thedegree of Cre-mediated recombination,ObR expression levels from all

ObRSynIKO and ObRAlbKO mice and asample of mice with all other possiblegenotypes were quantitated using Taq-man real-time PCR (Perkin ElmerApplied Biosystems, Foster City, Cali-fornia, USA). Total RNA was isolatedfrom ObRSynIKO and ObRAlbKO miceand reverse transcribed into cDNAusing random hexamers with theReverse Transcription Reagents fromRoche Molecular Systems (Branchburg,New Jersey, USA). Expression levelswere determined using 25 ng of eachcDNA sample assayed in duplicate andamplified with the ABI Prism 7700Sequence Detection System (PerkinElmer Applied Biosystems). The loca-tion of primers and fluorescent probesis indicated in Figure 3b. As amplifica-tion occurs, the probe is cleaved, result-ing in a signal from a reporter dye thatis directly related to the amount ofamplicon. Primers were derived fromsequences in the 5′ untranslated region(fwd) and the second coding exon (rev).The fluorescent probe, labeled with theFAM dye, contains sequence locatedwithin the first coding exon. When thisexon is deleted, the probe cannot becleaved, and no signal is generated. As acontrol for input amount, each cDNAsample was also amplified usingprimers and a probe, labeled with theVIC dye, for cyclophilin. Data were ana-lyzed with the ABI Sequence Detectorsoftware (Perkin Elmer Applied Biosys-tems). Every set of reactions containeda set of four serial twofold dilutions ofthe same liver cDNA source, which wasused to generate a standard curve forboth ObR and cyclophilin. Amounts ofeach transcript were calculated fromthe standard curve, and levels of ObRwere corrected for levels of cyclophilin(ObR level/cyclophilin level).

Body composition analysis. Body compo-sition was analyzed as described previ-ously (13). Carcasses were weighed andthen oven dried in a 90°C oven untilweight was constant. The total bodywater was calculated as the differencebetween the weights before and afterdrying. The carcass was then homoge-nized in a blender, and duplicatealiquots were extracted with a Soxhletapparatus using a 3:1 mixture of chloro-form/methanol. The extracted homo-genate was dried overnight and weighedto calculate fat mass and lean mass.

1114 The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8

Neuropeptide expression levels. Neu-ropeptide levels were quantitated byTaqman real-time PCR from individualhypothalami. Expression levels weremeasured in the obese ObRSynIKOmice, in heterozygote controls, and infemale ObR-null mice. Primers andprobes were made for agouti-relatedprotein (AGRP), neuropeptide Y (NPY), melanin concentrating hor-mone (MCH), proopiomelanocortin(POMC), and cocaine and ampheta-mine regulated transcript (CART).

Neuroendocrine function. After sacrifice,mice were exsanguinated and bloodwas collected on EDTA. Plasma was col-lected after centrifugation and used forall assays. Leptin levels were determinedusing an ELISA kit from R&D SystemsInc. (Minneapolis, Minnesota, USA).Insulin levels were determined using an

ELISA kit from Alpco Diagnostics(Windham, New Hampshire, USA).Glucose levels were determined usingTrinder reagents from Sigma ChemicalCo. (St. Louis, Missouri, USA). Triglyc-eride levels were determined using anenzymatic kit from Wako ChemicalUSA Inc. (Richmond, Virginia, USA).Corticosterone, estradiol, testosterone,and thyroxine levels were determinedusing RIA kits from ICN Radiochemi-cals Inc. (Costa Mesa, California, USA).

Liver triglyceride quantitation. A total of40–100 mg of liver from each mousewas homogenized in 4 ml of chloro-form-methanol (2:1). A total of 0.8 mlof 50 mM NaCl was added to each sam-ple. Samples were then centrifuged for5 minutes at 1,300 g. The lower phasewas removed, and duplicate 50-µlaliquots were evaporated under N2 gas

and then assayed for triglycerides usingthe Trinder reagent from Sigma Chem-ical Co. Values are expressed as mil-ligrams of triglyceride per gram of liver.

Statistical analysis. Data are expressed asmeans ± SE. Unless otherwise indicated,significance was evaluated using theunpaired Student’s t test. Significance ofcorrelation coefficients was evaluatedusing the t test for correlation.

ResultsWe have used the Cre-loxP system togenerate mice with tissue-specific dele-tions of ObR (27). The construction ofthe targeting plasmid is describedabove (Figure 1a). The first codingexon of ObR was flanked by loxP sites.This exon contains the signal sequence,and deletion of it by Cre-mediatedrecombination was predicted to inacti-

The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8 1115

Figure 1LoxP targeting of the ObR locus. Gene targeting was used to insert loxP sites on either side of the first coding exon of ObR. (a) Restrictionmaps (from top to bottom) of the genomic locus, targeting vector, homologous recombinant, and the type II and type I deletion alleles.Probe 1, located outside the targeting construct, was used to screen for homologous recombinants. Probe 2, located in the second codingexon, distinguishes the endogenous allele, homologous recombinants, and type I (deletion of the first coding exon) and II (deletion of thetargeting cassette leaving loxP sites on either side of the first coding exon) deletions. A, AflII; B, BamHI; Bg, BglII; H, HindIII; N, NcoI. (b)Southern blot analysis of NcoI-digested genomic DNA from ES cell clones using probe 1. The endogenous allele (Wt) and homologous recom-binants (HR) migrated at the predicted sizes. (c) Southern blot analysis of Hind-III digested genomic DNA using probe 2. The endogenousallele (Wt), homologous recombinant (HR), and type II deletion were detected from ES cell DNA. ObRflox/+ mice were generated from thetype II deletion and crossed to adenovirus EIIA-Cre mice. The type I deletion was detected in tail DNA from progeny derived from this cross.All alleles migrated at the predicted sizes. (d) Body weight at 8 weeks of age of ObR∆/∆ and littermate control (ObR∆/+) mice. Data representthe mean ± SEM of at least nine animals of each genotype and gender. *P < 0.02; **P < 0.001; unpaired Student’s t test.

vate all splice variants. After transfec-tion into mouse ES cells, three of 72G418-resistant clones had undergonehomologous recombination (Figure1b). The targeting cassette was deletedby transiently transfecting one of thehomologous recombinant ES cell lineswith a plasmid expressing Cre recom-binase and screening for clones with atype II deletion (Figure 1a). Clones withthe correct genotype, ObRflox/+, wereconfirmed by Southern blotting andinjected into C57BL/6 blastocysts (Fig-ure 1c). Breeding of chimeras con-firmed that animals with a germlineinsertion of loxP sites into ObR hadbeen generated. Mice homozygous forthe floxed allele, ObRflox/flox, were gener-ated by crossing ObRflox/+ animals.

ObRflox/flox mice were viable, fertile,and indistinguishable from wild-typeor ObRflox/+ mice, indicating that theinsertion of loxP sites into the intronsflanking the first coding exon did not

interfere with ObR function. To deter-mine whether Cre-mediated deletionof the first coding exon results in anull allele, ObRflox/+ mice were crossedwith transgenic mice expressing Cre atan early embryonic stage (23). Micewith the genotype ObR∆/+, heterozy-gous for a germline deletion of thefirst coding exon (type I deletion), weregenerated and mated to producehomozygous ObR∆/∆ mice, also referredto as ObR-null mice (Figure 1c). ObR-null mice are massively obese andindistinguishable from db3J/db3J mice,which are also null for all ObR iso-forms (Figure 1d) (28). These data con-firm that the floxed allele is wild-typeand that the deleted allele is null, thusvalidating the targeting strategy.

Neuron-specific ObR-knockout mice,designated ObRSynIKO, were generatedby two successive crosses (Figure 2a, andsee Methods). These mice carry onefloxed and one null allele so that Cre-

mediated inactivation of only one allele,rather than two, is sufficient to deleteObR in a given cell. The ObRSynIKOmice were compared to littermate con-trols with the genotype ObR∆/flox, SynI-Cre(–), also referred to as heterozygotes.These mice were used as controlsbecause some phenotypic changes havebeen observed in db/+ mice (29). Datawere also obtained from lean littermateswith the genotypes ObRflox/+, SynI-Cre(+) or ObRflox/+, SynI-Cre(–), both ofwhich are referred to as wild-type.Geno-types of all mice were determined usingPCR and Southern blotting.

In the synapsinI-Cre transgenic line,expression of Cre recombinase is con-trolled by the rat synapsin I promoter.This promoter has been shown to driveCre expression specifically in neurons(30). Crossing SynI-Cre(+) mice to LacZindicator transgenic strains showedthat Cre activity was first detectable atembryonic day (E) 12.5 and restrictedto brain, spinal cord, and the dorsalroot ganglion and was absent fromastrocytes and glia (25). To confirm thetissue specificity of Cre-mediatedrecombination, a qualitative PCR assaywas developed that was capable ofdetecting recombination between thelox sites flanking the first coding exonof ObR in genomic DNA. In ObRflox/+,synapsinI-Cre(+) mice, Cre-mediatedrecombination was restricted to brain,hypothalamus, and spinal cord (Figure2b). As previously indicated by reportergene expression in SynI-CAT trans-genic mice, low level recombinationwas also seen in testis (data not shown).

The weight of the ObRSynIKO micewas compared to that of heterozygousmice at 4 months of age. Although theweights of ObRSynIKO mice, on average,were no different than heterozygotes, inboth sexes a subset of the ObRSynIKOmice had an increased body weight (Fig-ure 3a). The percent body fat of thesemice (Figure 3a, yellow) was more than2.5 SD greater than the average percentbody fat of heterozygotes (Figure 3a).This suggested the possibility that theextent of the knockout of ObR was vari-able and that the most obese ObRSynIKOanimals had the lowest levels of ObR. Toassess the efficiency of Cre-mediatedrecombination, real-time PCR assays(Taqman) were performed to quantitateObR RNA using cDNA prepared from

1116 The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8

Figure 2Cre-mediated recombination specifically in the brain of ObRSynIKO mice. (a) The breedingstrategy that was used to generate ObRSynIKO mice and littermate controls. ObRAlbKO mice(described below) were generated using the same strategy. (b) Genomic DNA was preparedfrom several tissues from ObRflox/+, SynI-Cre(+) mice and were PCR-amplified using primersflanking the first coding exon of ObR. The locations of the primers are shown on the schemat-ic below the gel. In tissues expressing the Cre recombinase, exon 1 is excised and a single loxPsite remains, generating an ObR∆ allele. While primers 1 and 3 can amplify a product from theObR∆ allele, in the ObRflox or ObR+ (wild-type) DNA, the primers are separated by a distancetoo great for amplification to occur. As a control, primers 1 and 2 were used to amplify boththe ObRflox allele and the ObR+ alleles from all tissues. The ObR+ allele produces a slightlysmaller product due to the absence of loxP sequences.

individual hypothalami (31). Previousdata have suggested that in neurons,ObR is expressed at highest levels inhypothalamus and at much lower levelsin other brain regions. This assay detect-ed all forms of the leptin receptor. ObRflox/+, SynI-Cre(+) and ObRflox/+,SynI-Cre(–) mice had equivalent levels ofObR expression, and thus both geno-types were designated as wild-type (Fig-ure 3b). As expected, ObR∆/flox, SynI-Cre(–) mice, which have one alleleinactivated in the germ line, have 50% asmuch RNA as wild-type mice (P < 0.006). The average level of ObRRNA was significantly lower inObRSynIKO mice than in ObR∆/flox, SynI-Cre(–) heterozygote littermates (30% vs.50% of wild-type, respectively; P = 0.004).Moreover, although many of theObRSynIKO mice had levels of ObR RNAthat were indistinguishable from het-erozygotes, a subset of the ObRSynIKOmice had markedly reduced levels ofObR. These results confirmed that Cre-mediated recombination is variable, andindicated that in many animals there isno evident recombination.

We next considered whether the mostobese ObRSynIKO (i.e., percent body fat2.5 SD > heterozygotes) animals hadlower levels of ObR RNA than thelighter ones. Analysis of the data indi-cated that percent body fat and bodyweight were significantly increased ineach of six animals in which ObR RNA

was less than 15% that of wild-type (Fig-ure 3a, yellow) (P < 0.005 for percentbody fat; P < 0.05 at all ages greater than5 weeks for body weight). This con-firmed that a significant deletion ofObR is associated with an obese pheno-type. Furthermore, in those cases wherethe deletion is most extreme, a severelyobese phenotype is evident. TwoObRSynIKO females and one ObRSynIKOmale had a greater than 97% reductionof ObR RNA, and each weighed morethan 50 g at sacrifice, weights thatapproach, but do not equal, those ofdb/db mice. In these studies, as well asstudies of the ObRAlbKO mice (seebelow), there were no lean mice withhypothalamic ObR RNA levels less than15% that of wild-type, and there were noobese mice (obese defined as percentbody fat > 2.5 SD above heterozygotes:30.5% for males, 47.8% for females) withRNA levels greater than 15% that ofwild-type. These data suggest that anear-normal body weight can be main-tained until markedly reduced levels ofObR in hypothalamus are evident. In allfollowing studies, the phenotype of theanimals with less than 15% wild-typelevels of hypothalamic ObR RNA wascompared to heterozygotes. These miceare referred to as obese ObRSynIKO mice.

The weights of the obese ObRSynIKOmice were significantly increased rela-tive to heterozygotes at all time pointsgreater than 5 weeks of age (P < 0.05).

At 5 months of age, males and femalesweighed 35% and 66% more than het-erozygotes, respectively (Figure 4a).The increase in body weight was asso-ciated with increased adipose tissuemass and, as is also the case in ob/oband db/db mice, decreased lean bodymass (Table 1). At sacrifice, the percentbody fat of ObRSynIKO mice was 39%in males and 60% in females, com-pared with 15% in male heterozygotesand 26% in female heterozygotes.

The obese ObRSynIKO mice were alsoanalyzed with respect to a number ofother abnormalities associated with thedb/db mutation (Table 1). Plasma leptinconcentrations were elevated 5.8-fold inobese ObRSynIKO males (P < 0.05) and8.3-fold in obese ObRSynIKO females (P < 0.0001) and were highly correlatedwith percent body fat (r = 0.82, P < 0.01,t test for correlation). In addition, plas-ma insulin was increased sixfold inobese ObRSynIKO males (P < 0.05) and16.1-fold in ObRSynIKO females. Glu-cose was increased 2.3-fold inObRSynIKO females (P < 0.005), and wasunchanged in ObRSynIKO males. Plas-ma corticosterone was elevated 1.8-foldin ObRSynIKO males and 2.4-fold inObRSynIKO females (P < 0.05). The ele-vation in plasma glucose and thegreater elevation in corticosterone spe-cific to females could be due to thefemales analyzed having a lower level ofObR RNA and being more obese. Of

The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8 1117

Figure 3ObR RNA levels. (a) The distribution of body weight and percent body fat for male and female ObRSynIKO mice at 16 weeks of age. Thosemice with less than 15% of ObR RNA (see b) are shown in yellow. (b) Expression levels were determined using Taqman real-time PCR withthe ABI Prism 7700 Sequence Detection System. The locations of primers and fluorescent probe are as indicated. Primers were derived fromsequences in the 5′ untranslated region and the second coding exon. The fluorescent probe is located within the first coding exon 1. Whenthis exon is deleted, no signal is generated. As a control for input amount, each cDNA sample was also amplified using primers and a probefor cyclophilin. Data were analyzed with ABI Sequence Detector software, and the levels of ObR were normalized to cyclophilin. Levels arerepresented as the percentage of the levels in ObRflox/+, SynI-Cre(+) and ObRflox/+, SynI-Cre(–) wild-type mice. Data represent the mean ±SEM. At least 13 animals were analyzed for each genotype. P values for comparisons between genotypes are indicated.

note, an increased plasma corticos-terone is unique to ob/ob and db/dbmice, whereas other forms of rodentobesity are not generally associated withelevated corticosterone (32). Mutationsin leptin or its receptor are also associ-ated with infertility, and leptin has beenshown to influence the hypothalamic-pituitary-gonadal axis (33). Consistentwith this, obese female ObRSynIKO micehad 43% reductions in plasma estradiol.Male ObRSynIKO mice, however, did nothave reduced levels of plasma testos-terone. This sexual dimorphism is con-sistent with the fact that leptin defi-ciency has a more profound effect onfemale reproduction than male repro-duction (34). Finally, plasma thyroxineand triglycerides were not significantlydifferent between mutant and controlmice (data not shown).

Despite the marked obesity in thissubset of ObRSynIKO mice, these ani-mals still weighed less than mice withgermline deletions of ObR (Figure 4a; P < 0.05 at all ages). These data suggesteither that the knockout was incom-plete even in the most obese animals orthat leptin also acts at peripheral sites toreduce weight. To further assess theextent of the knockout in brain, levels ofa number of hypothalamic neuropep-tides were examined using Taqmanassays (Figure 4b). Defective leptin sig-naling is associated with increased levelsof hypothalamic AGRP, NPY, and MCH

RNA and reduced levels of POMC andCART RNA (35–40). Comparisons ofheterozygotes to obese ObRSynIKO miceshowed that AGRP and NPY levels wereincreased, whereas POMC and CARTlevels showed a trend toward beingreduced. Although the reduction inPOMC and CART RNA levels is not sig-nificant, the levels of these RNAs areonly modestly decreased in ob/ob anddb/db hypothalamus. MCH levels wereindistinguishable between heterozygousmice, ObRSynIKO mice, and ObR-nullmice. The reported 80% increase inMCH RNA in ob/ob mice is also moremodest than the reported increases inAGRP and NPY RNA (38). Although thelevels of AGRP and NPY were increasedin the obese ObRSynIKO mice, they wereless elevated than in null mice. This sug-gests that, despite the low levels of ObRRNA in these mice, some ObR-express-ing neurons remained.

The data here indicate that leptin hasdirect effects on brain but leave open thepossibility that leptin also has directeffects on peripheral tissues. To assess apossible role of ObR in liver, a peripher-al tissue that has been suggested to be adirect target of leptin action, mice with ahepatocyte-specific knockout of ObRwere generated (ObRAlbKO). These micewere generated using the same breedingstrategy as described for ObRSynIKOmice (Figure 2a). The albumin promot-er was used to direct hepatocyte-specific

expression of Cre recombinase (Figure5a) (26). Genomic PCR performed onmultiple tissues from ObRflox/+ albumin-Cre(+) mice confirmed that Cre-mediat-ed recombination was restricted to liver(Figure 5b). In contrast to ObRSynIKOmice, ObRAlbKO mice did not manifestan increased weight even in those ani-mals in which the levels of ObR RNA inliver (as measured using the aforemen-tioned Taqman assay) were reduced toless than 30% of wild-type levels (9 of 12males and 11 of 16 females analyzed)(Figure 5c). Because the liver is com-posed of a number of different cell types,and 80% of the mass is accounted for byhepatocytes, this likely represents a near-complete deletion in hepatocytes (41,42). Inspection of the weight distribu-tions showed that no ObRAlbKO micemanifested an increased weight. Bodycomposition was analyzed to determinewhether hepatocyte-specific deletion ofObR had a subtle effect on body parti-tioning in ObRAlbKO mice. Water mass,lean mass, and fat mass were allunchanged in these animals (Table 1). Inboth sexes, plasma leptin levels werewithin the normal range as were insulin,glucose, and corticosterone. Althoughdb/db and ObRSynIKO mice have enlargedfatty livers characteristic of leptin-defi-cient mice, the livers of ObRAlbKO micewere grossly normal (Figure 5d). Wetliver weight and liver weight as a per-centage of body weight were also

1118 The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8

Figure 4Obesity in a subset of ObRSynIKO mice.Significant obesity was evident in thoseObRSynIKO mice that had 15% or less ObRRNA in the hypothalamus than did wild-type mice. (a) Weight curves of male andfemale ObRSynIKO mice (filled squares),heterozygote littermates (open squares),and ObR-null mice (filled triangles). Micewere weighed weekly from 5 weeks of age.Data represent the mean ± SEM of fourmale and two female ObRSynIKO mice,seven male and nine female heterozygotelittermates, and six male and five femaleObR-null mice. For both sexes, ObRSynIKOversus heterozygotes, P < 0.05 at all ages;ObRSynIKO versus ObR-null, P < 0.05 at allages. (b) Expression levels of AGRP, NPY,POMC, and CART were determined byTaqman and are expressed as normalizedto cyclophilin. Levels were measured inObRSynIKO, heterozygotes (HET), andObR-nulls. (n = 5 for heterozygotes, n = 6for knockouts, and n = 3 for null mice.)

unchanged between ObRAlbKO miceand heterozygotes (data not shown).Quantitation of liver triglyceridesshowed no difference between ObRAlbKO mice and heterozygote con-trols, whereas levels in ObR-null micewere significantly elevated relative toboth of these groups (P < 0.05) (Figure5e). These data indicate that ObRexpressed in hepatocytes is not likely toplay a significant role in body weighthomeostasis or the hepatic steatosisassociated with the db mutation, andthat the liver abnormalities observed inob/ob and db/db mice are secondary todefective leptin signaling in the brain.

DiscussionIn this study, two possible sites of lep-tin action were evaluated by analyzing

the phenotype of mice with neuronand hepatocyte-specific knockouts ofthe leptin receptor. These data providedirect genetic evidence that leptin sig-nals via direct effects on neurons. Sep-arate data from studies of dbKls/dbKls

mice indicate that the specific absenceof the ObRb form of the receptorresults in a phenotype indistinguish-able from leptin deficiency, demon-strating that ObRb is absolutelyrequired for the weight-reducingeffects of the hormone (4–6). ObRb isthe only isoform that has all of themotifs necessary for signal transduc-tion and is expressed at high levels inbrain and a number of other tissues.

These findings suggest that leptinexerts its weight-reducing and some, orperhaps all, of its neuroendocrine effects

via interactions with the ObRb form ofthe receptor in specific classes of neu-rons in brain. This is consistent with thehigh potency of leptin-administeredICV and the anatomic distribution ofObRb (see below) (15, 17). This conclu-sion is also consistent with data fromexperiments in which an NSE-ObRbtransgene was able to partially comple-ment the phenotype of db/db mice (43).However, in this study, ObRb was over-expressed 30-fold in all neurons, withsome leakiness in peripheral tissues.

The phenotype of the obeseObRSynIKO, while significant, did notreach that of ObR-null mice. In theseanimals, however, hypothalamic AGRPand NPY levels, while increased, werestill not as high as in null mice, suggest-ing that there were still some residual

The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8 1119

Figure 5Normal body weight and liver phenotype in ObRAlbKO mice. (a) Schematic of Albumin-Cre transgene. (b) Genomic DNA was prepared fromseveral tissues from ObRflox/+, Alb-Cre(+) mice and was PCR-amplified using primers flanking the first coding exon of ObR as described inFigure 2b. (c) Weight curves of male and female ObRAlbKO mice (filled squares) and heterozygote littermates (open squares). Mice wereweighed weekly from 5 weeks of age. Data represent the mean ± SEM of 9 male and 11 female ObRAlbKO mice and 23 male and 23 femaleheterozygote littermates. For both sexes, ObRAlbKO versus heterozygotes, P values were not significant at all ages. (d) Photographs of fresh-ly dissected livers from representative mice with the given genotypes. (e) Liver triglycerides (milligrams of triglyceride per grams of liver) weredetermined for ObRAlbKO mice with less than 30% of wild-type ObR RNA and heterozygote controls. Data represent the mean ± SEM of sixmale and five female ObRAlbKO mice, five male and five female heterozygote controls, and six male and three female ObR-null mice. Forboth sexes, *P < 0.05 for ObRAlbKO and heterozygotes relative to ObR null.

ObR-expressing cells. This leaves openthe possibility that a truly completebrain-specific knockout (i.e., deletion ofObR on every neuron) could recapitu-late the obese phenotype of ObR-nullmice. The generation of such a knock-out might require the simultaneous useof multiple different neuron-specificCre-expressing lines, now under way.

In some studies, the expression ofObRb in brain was reported to berestricted to the hypothalamus, possiblyimplicating the hypothalamus as theprincipal target of leptin’s weight-reduc-ing effects. In hypothalamus, ObRb isenriched in the arcuate, ventromedial,dorsomedial, and paraventricular hypo-thalamic nuclei, all of which are known,based on studies of animals with stereo-tactic lesions, to play a role in regulatingfood intake and body weight (19,44–46). However, in other studies, ObRbexpression has also been reported inother brain sites including the brain-stem, thalamus, and cerebellum. Thepossibility that one or more of these

extrahypothalamic brain sites playimportant roles in the regulation ofbody weight can be tested using region-specific or neuron-specific promoters todrive Cre expression in localized brainareas and specific cell types. In the hypo-thalamus, ObRb is expressed in func-tionally distinct classes of neurons, thebest characterized being neurons thatcoexpress either NPY or αMSH (47, 48).Both of these neuropeptides exertpotent effects on food intake and bodyweight. Experiments to test the effectsof ObR deletion in these two types ofneurons may allow an analysis of thefunctional importance of these two neu-ropeptides as downstream effectors ofleptin signaling and could establish theextent to which these two cell types bythemselves can fully account for theneural response to leptin.

A number of studies have suggestedthat leptin also has direct effects on liver(20, 21). The data presented here indi-cate that hepatocyte-specific deletiondoes not have any observable effects on

body weight or body composition anddoes not result in any gross liver pathol-ogy. Furthermore ObRAlbKO mice haveequivalent amounts of liver triglyceridesas do controls, indicating that thehepatic phenotype of db/db mice is sec-ondary to defective leptin signaling inbrain. These findings do not exclude thepossibility that hepatocyte ObR by itselfserves functions, such as modulatingleptin turnover.

Although these findings suggest thatneurons are required for leptin’s effectson body weight, they do not excludethe possibility that leptin also directlyacts on other peripheral targets. Inaddition to regulating food intake,body weight, and the neuroendocrineaxis, leptin can also modulate theimmune system, bone metabolism,and angiogenesis (49–51). Indeed,ObRb is expressed in a number ofperipheral tissues and direct effects ofleptin on T cells, macrophages, pancre-atic β-cells, and other cell types havebeen reported (20, 21, 52). Other forms

1120 The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8

Table 1Body composition and neuroendocrine function in ObRSynIKO and ObRAlbKO miceA

ObRSynIKO ObRSynIHet ObRAlbKO ObRAlbHet

Percent fat mass

Males 39.4 ± 0.7 (4) 15.4 ± 0.8 (8)B 14.0 ± 1.5 (4) 13.2 ± 0.5 (5)Females 60.1 ± 1.5 (2) 25.8 ± 0.9 (10)B 15.4 ± 0.9 (7) 15.6 ± 1.1 (5)

Percent lean mass

Males 19.7 ± 0.6 (4) 25.7 ± 0.2 (8)B 30.2 ± 0.8 (4) 30.2 ± 0.8 (5)Females 12.7 ± 1.2 (2) 22.6 ± 0.3 (10)B 28.0 ± 0.7 (4) 27.5 ± 0.7 (5)

Percent water mass

Males 40.9 ± 0.3 (4) 59.0 ± 0.6 (8)B 55.8 ± 1.3 (4) 56.6 ± 0.6 (5)Females 27.2 ± 0.3 (2) 51.6 ± 0.6 (10)B 56.6 ± 0.9 (4) 56.9 ± 0.7 (5)

Leptin (ng/ml)

Males 73.8 ± 8.5 (4) 12.7 ± 0.4 (10)C 13.1 ± 0.8 (9) 13.9 ± 0.7 (15)Females 125.0 ± 0.7 (2) 15.1 ± 0.8 (12)B 12.7 ± 0.9 (6) 6.7 ± 0.2 (12)C

Insulin (ng/ml)

Males 17.1 ± 2.1 (4) 2.8 ± 0.2 (8)C 1.9 ± 0.1 (9) 1.6 ± 0.1 (6)Females 40.1 ± 6.5 (2) 2.5 ± 0.2 (8) 0.8 ± 0.1 (4) 0.7 ± 0.1 (5)

Glucose (mg/dl)

Males 158.3 ± 9.3 (4) 165.3 ± 4.9 (10) 80.3 ± 2.3 (7) 89.1 ± 2.5 (8)Females 282.6 ± 32.3 (2) 123.1 ± 4.7 (8)B 87.3 ± 3.9 (5) 91.9 ± 2.0 (9)

Corticosterone (ng/ml)

Males 238.5 ± 17.9 (4) 134.9 ± 40.2 (3) 54.0 ± 4.2 (6) 63.4 ± 2.7 (11)Females 469.4 ± 32.7 (2) 195.9 ± 16.1 (8)C 208.8 ± 25.1 (5) 108.5 ± 9.3 (9)

Estradiol (pg/ml)

Females 3.9 ± 0.3 (2) 6.8 ± 0.8 (3) ND ND

AAfter sacrifice, the subset of significantly obese ObRSynIKO mice and the subset of ObRAlbKO mice with less than 30% ObR RNA as well as heterozygote lit-termates were exsanguinated, and plasma was collected for the following assays. All plasma was collected from mice sacrificed at 1200 hours under free feed-ing conditions. Average ages for ObRSynIKO and ObRAlbKO mice sampled here are 26 weeks and 23 weeks. All data represent the mean ± SEM, with the sam-ple size indicated in parentheses. BP < 0.005 versus knockout; CP < 0.05 versus knockout in an unpaired Student’s t test. ND, not determined.

of ObR (ObRa, c, d, e) may play a rolein transport of leptin across the bloodbrain barrier and choroid plexus, sitesthat express high levels of ObR. Thefunction of ObR at these and othersites can now be assessed by crossingObRflox/flox mice to transgenic lines thatexpress Cre in a tissue-specific fashion.Such studies should enable an assess-ment of the relative role of directeffects of leptin on these tissues versusindirect effects via signaling in brain. Inconclusion, these findings indicatethat the brain is a direct target for theweight-reducing and neuroendocrineeffects of leptin and that the liverpathology of db/db mice is secondary todefective leptin signaling in neurons.

AcknowledgmentsWe thank H. Westphal for adenovirusEIIA-Cre mice; J. Marth for synapsinI-Cremice; D. Shmulewitz for assistance withdata analysis; R. Sharma for assistancewith liver photography; E. Asilmaz, S.Brown, M. Ishii, C. Li, S. Pinto, and A.Soukas for critical reading of this man-uscript; and S. Korres for assistance inpreparing this manuscript. P. Cohenwas supported by NIH MSTP grantGM07739. J.M. Friedman was support-ed by NIH grant R01DK-41096.

1. Zhang, Y., et al. 1994. Positional cloning of themouse obese gene and its human homologue.Nature. 372:425–432.

2. Friedman, J.M., and Halaas, J.L. 1998. Leptin andthe regulation of body weight in mammals.Nature. 395:763–770.

3. Spiegelman, B.M., and Flier, J.S. 2001. Obesityand the regulation of energy balance. Cell.104:531–543.

4. Tartaglia, L.A., et al. 1995. Identification andexpression cloning of a leptin receptor, OB-R.Cell. 83:1263–1271.

5. Chen, H., et al. 1996. Evidence that the diabetesgene encodes the leptin receptor: identification ofa mutation in the leptin receptor gene in db/dbmice. Cell. 84:491–495.

6. Lee, G.H., et al. 1996. Abnormal splicing of theleptin receptor in diabetic mice. Nature.379:632–635.

7. Coleman, D.L. 1978. Obese and diabetes: twomutant genes causing diabetes-obesity syn-dromes in mice. Diabetologia. 14:141–148.

8. Coleman, D.L. 1982. Diabetes-obesity syndromesin mice. Diabetes. 31(Suppl. 1):1–6.

9. Yen, T.T., Allan, J.A., Yu, P-L., Acton, M.A., andPearson, D.V. 1976. Triacylglycerol contents andin vivo lipogenesis of ob/ob, db/db, and Avy/a mice.Biochim. Biophys. Acta. 441:213–220.

10. Koteish, A., and Diehl, A.M. 2001. Animal mod-els of steatosis. Semin. Liver Dis. 21:89–104.

11. Kaplan, M.L., and Leveille, G.A. 1981. Develop-ment of lipogenesis and insulin sensitivity in tis-sues of the ob/ob mouse. Am. J. Physiology.240:E101–E107.

12. Levin, N., Nelson, C., Gurney, A., Vandlen, R., andde Sauvage, F. 1996. Decreased food intake doesnot completely account for adiposity reductionafter ob protein infusion. Proc. Natl. Acad. Sci.USA. 93:1726–1730.

13. Halaas, J.L., et al. 1995. Weight-reducing effectsof the plasma protein encoded by the obese gene.Science. 269:543–546.

14. Pelleymounter, M.A., et al. 1995. Effects of theobese gene product on body weight regulation inob/ob mice. Science. 269:540–543.

15. Campfield, L.A., Smith, F.J., Guisez, Y., Devos, R.,and Burn, P. 1995. Recombinant mouse OB pro-tein: Evidence for a peripheral signal linking adi-posity and central neural networks. Science.269:546–549.

16. Soukas, A., Cohen, P., Socci, N.D., and Friedman,J.M. 2000. Leptin-specific patterns of gene expres-sion in white adipose tissue. Genes Dev.14:963–980.

17. Halaas, J.L., et al. 1997. Physiological response tolong-term peripheral and central leptin infusionin lean and obese mice. Proc. Natl. Acad. Sci. USA.94:8878–8883.

18. Ghilardi, N., et al. 1996. Defective STAT signalingby the leptin receptor in diabetic mice. Proc. Natl.Acad. Sci. USA. 93:6231–6235.

19. Fei, H., et al. 1997. Anatomic localization of alter-natively spliced leptin receptors (Ob-R) in mousebrain and other tissues. Prod. Natl. Acad. Sci. USA.94:7001–7005.

20. Shimabukuro, M., et al. 1997. Direct antidiabeticeffect of leptin through triglyceride depletion oftissues. Proc. Natl. Acad. Sci. USA. 94:4637–4641.

21. Kim, Y.B., Uotani, S., Pierroz, D.D., Flier, J.S., andKahn, B.B. 2000. In vivo administration of leptinactivates signal transduction directly in insulin-sensitive tissues: overlapping but distinct path-ways from insulin. Endocrinology. 141:2328–2339.

22. O’Gorman, S., Dagenais, N.A., Qian, M., andMarchuk, Y. 1997. Protamine-cre recombinasetransgenes efficiently recombine target sequencesin the male germ line of mice, but not in embry-onic stem cells. Proc. Natl. Acad. Sci. USA.94:14602–14607.

23. Lasko, M., et al. 1996. Efficient in vivo manipula-tion of mouse genomic sequences at the zygotestage. Proc. Natl. Acad. Sci. USA. 93:5860–5865.

24. DeFalco, J., et al. 2001. Virus-assisted mapping ofneural inputs to a feeding center in the hypothal-amus. Science. 291:2608–2613.

25. Zhu, Y., et al. 2001. Ablation of NF1 function inneurons induces abnormal development of cere-bral cortex and reactive gliosis in the brain. GenesDev. 15:859–876.

26. Pinkert, C.A., Ornitz, D.M., Brinster, R.L., andPalmiter, R.D. 1987. An albumin enhancer locat-ed 10 kb upstream functions along with its pro-moter to direct efficient, liver-specific expressionin transgenic mice. Genes Dev. 1:268–276.

27. Gu, H., Marth, J.D., Orban, P.C., Mossmann, H.,and Rajewsky, K. 1994. Deletion of a DNA poly-merase β gene segment in T cells using cell type-specific gene targeting. Science. 265:103–106.

28. Lee, G-H., et al. 1997. Leptin receptor mutationsin 129 db3J /db3J mice and NIH facp/facp rats.Mamm. Genome 8:445–447.

29. Coleman, D.L. 1979. Obesity genes: beneficialeffects in heterozygous mice. Science. 203:663–665.

30. Hoesche, C., Sauerwald, A., Veh, R.W., Krippl, B.,and Kilimann, M.W. 1993. The 5′-flanking regionof the rat synapsin I gene directs neuron-specificand developmentally regulated reporter geneexpression. J. Biol. Chem. 268:26494–26502.

31. Livak, K.J., Flood, S.J., Marmaro, J., Giusti, W., andDeetz, K. 1995. Oligonucleotides with fluores-cent dyes at opposite ends provide a quenchedprobe system for detecting PCR products andnucleic acid hybridization. PCR Methods Appl.4:357–362.

32. Bray, G.A., and York, D.A. 1979. Integration ofenergy intake and expenditure in animals andman: the autonomic and adrenal hypothesis.Physiol. Rev. 59:719–809.

33. Chehab, F.F., Mounzih, K., Lu, R., and Lim, M.E.1997. Early onset of reproductive function in nor-mal female mice treated with leptin. Science.275:88–90.

34. Ewart-Toland, A., Mounzih, K., Qiu, J., andChehab, F.F. 1999. Effect of the genetic back-ground on the reproduction of leptin-deficientobese mice. Endocrinology. 140:732–738.

35. Ollmann, M.M., et al. 1997. Antagonism of cen-tral melanocortin receptors in vitro and in vivo byagouti-related protein. Science. 278:135–138.

36. Shutter, J.R., et al. 1997. Hypothalmic expressionof ART, a novel gene related to agouti, is up-reg-ulated in obese and diabetic mutant mice. GenesDev. 11:593–602.

37. Stephens, T.W., et al. 1995. The role of neuropep-tide Y in the antiobesity action of the obese geneproduct. Nature. 377:530–534.

38. Qu, D., et al. 1996. A role for melanin-concen-trating hormone in the central regulation of feed-ing behavior. Nature. 380:243–247.

39. Thornton, J.E., Cheung, C.C., Clifton, D.K., andSteiner, R.A. 1997. Regulation of hypothalamicproopiomelanocortin mRNA by leptin in ob/obmice. Endocrinology. 138:5063–5066.

40. Kristensen, P., et al. 1998. Hypothalamic CART isa new anorectic peptide regulated by leptin.Nature. 393:72–76.

41. Gumucio, J.J., Berkowitz, C.M., Webster, S.T., andThornton, A.J. 1996. Structural and functionalorganization of the liver. In Liver and biliary disease.N. Kaplowitz, editor. Williams and Wilkins. Bal-timore, Maryland, USA. 3–19.

42. Weibel, E.R., Staubli, W., Gangi, H.R., and Hess,F.A. 1969. Correlated morphometric and bio-chemical studies on the liver cell. I. Morphome-tric model, stereologic methods, and normalmorphometric data for rat liver. J. Cell. Biol.42:68–91.

43. Kowalski, T.J., Liu, S-M., Leibel, R.L., and Chua,S.C. 2001. Transgenic complementation of leptinreceptor deficiency. Diabetes. 50:425–435.

44. Elmquist, J.K., Ahima, R.S., Maratos-Flier, E.,Flier, J.S., and Saper, C.B. 1997. Leptin activatesneurons in ventrobasal hypothalamus and brain-stem. Endocrinology. 138:839–842.

45. Elmquist, J.K., Bjorbaek, C., Ahima, R.S., Flier,J.S., and Saper, C.B. 1998. Distributions of theleptin receptor mRNA isoforms in the rat brain.J. Comp. Neurol. 395:535–547.

46. Hetherington, A.W., and Ranson, S.W. 1942. Thespontaneous activity and food intake of rats withhypothalamic lesions. Am. J. Physiol. 136:609–617.

47. Mercer, J.G., et al. 1996. Coexpression of leptinreceptor and preproneuropeptide Y mRNA inarcuate nucleus of mouse hypothalamus. J. Neu-roendocrinology. 8:733–735.

48. Cheung, C.C., Clifton, D.K., and Steiner, R.A.1997. Proopiomelanocortin neurons are directtargets for leptin in the hypothalamus.Endocrinology. 138:4489–4492.

49. Lord, G.M., Matarese, G., Howard, J.K., Bloom,S.R., and Lechler, R.I. 1998. Leptin modulates theT-cell immune response and reverses starvationinduced immunosuppression. Nature.394:897–891.

50. Ducy, P., et al. 2000. Leptin inhibits bone forma-tion through a hypothalamic relay: a central con-trol of bone mass. Cell. 100:197–207.

51. Sierra-Honigmann, M., et al. 1998. Biologicaction of leptin as an angiogenic factor. Science.281:1683–1686.

52. O’Rourke, L., Yeaman, S.J., and Shepherd, P.R.2001. Insulin and leptin acutely regulate choles-terol ester metabolism in macrophages by novelsignaling pathways. Diabetes. 50:955–961.

The Journal of Clinical Investigation | October 2001 | Volume 108 | Number 8 1121