Xinxin Xiang,1,2,3 He Lan,1,2 Hong Tang,1,2 Fang Yuan,1,2 Yanhui Xu,1,2 Jing Zhao,1,2 Yin Li,1,2 andWeizhen Zhang1,2,4

Tuberous Sclerosis Complex 1–Mechanistic Target of RapamycinComplex 1 Signaling DeterminesBrown-to-White AdipocytePhenotypic SwitchDiabetes 2015;64:519–528 | DOI: 10.2337/db14-0427

Interconversion of white and brown adipocytes occursbetween anabolic and catabolic states. The molecularmechanism regulating this phenotypic switch remainslargely unknown. This study explores the role of tuber-ous sclerosis complex 1 (TSC1)–mechanistic target ofrapamycin (mTOR) signaling in the conversion of brownto white adipose tissue (WAT ). A colony of Fabp4-Tsc12/2

mice, in which the Tsc1 gene was specifically deletedby the fatty acid binding protein 4 (FABP4)-Cre, wasestablished. Western blotting and immunostaining dem-onstrated the absence of TSC1 and activation of ribo-somal protein S6 kinase 1, the downstream target ofmTOR complex 1 (mTORC1) signaling, in the brown ad-ipose tissues (BATs) of Fabp4-Tsc12/2 mice. Accumula-tion of lipid droplets in BAT was significantly increased.Levels of brown adipocyte markers were markedlydownregulated, while white adipocyte markers were up-regulated. Rapamycin reversed the conversion from BATto WAT in Fabp4-Tsc12/2 mice. Deletion of the Tsc1gene in cultured brown preadipocytes significantly in-creased the conversion to white adipocytes. FoxC2mRNA, the transcriptional factor for brown adipocytedetermination, was significantly decreased, while mRNAsfor retinoblastoma protein, p107 and RIP140, the tran-scriptional factors for white adipocyte determination,increased in the BAT of Fabp4-Tsc12/2 mice. Our studydemonstrates that TSC1-mTORC1 signaling con-tributes to the brown-to-white adipocyte phenotypicswitch.

In mammals, there exist two different types of adiposetissues: white adipose tissues (WATs) and brown adiposetissues (BATs), which are made up of unilocular whiteadipocytes and multilocular (ML) brown adipocytes, re-spectively. White adipocytes contain all of the enzymaticmachinery necessary to build triglycerides from fatty acidssynthesized de novo or imported from circulating lipo-proteins, while brown adipocytes comprise of abundantmitochondria that oxidize fatty acids and generate heatvia uncoupling protein 1 (UCP1) in response to cold orexcess feeding (1–3). WAT and BAT are traditionally con-sidered distinct fat depots located in different anatomicsites. However, recent studies have shown that most ofthe fat depots contain both WAT and BAT, which areinterconvertible (4). Differentiated WAT has been demon-strated to acquire certain BAT-like properties and can beconverted into an energy-consumption organ under spe-cial conditions. Cold exposure or administration of b3-agonist has been reported to induce the emergence ofML brown adipocytes containing a high amount of mito-chondria and UCP1 in WAT depots (5–7).

The molecular mechanism underlying the interconver-sion between WAT and BAT is currently under activeinvestigation. Study with the transgenic mouse modelsuggests that C/EBPb is critical for the brown adipocyteconversion. Transgenic knockin mouse that replaces theC/EBPa gene by C/EBPb gene (denoted as C/EBP b/b)

1Department of Physiology and Pathophysiology, Peking University Health Sci-ence Center, Beijing, China2Key Laboratory of Molecular Cardiovascular Science, Ministry of Education,Beijing, China3Department of Pathology, Central Hospital of Zibo, Zibo, China4Department of Surgery, University of Michigan, Ann Arbor, MI

Corresponding authors: Weizhen Zhang, weizhenzhang@bjmu.edu.cn, and Yin Li,yinli@bjmu.edu.cn.

Received 14 March 2014 and accepted 3 September 2014.

© 2015 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, andthe work is not altered.

Diabetes Volume 64, February 2015 519

OBESITYSTUDIES

shows an increase in brown adipocyte-like cells charac-terized by cAMP accumulation, enhanced mitochondrialbiogenesis, and UCP1 expression in WATs (8). In addi-tion, several transcription-related molecules such asretinoblastoma protein (Rb), p107, transcriptional in-termediary factor 2, 4E-binding protein 1, and RIP140(9–15) have been reported to influence the brown orwhite fat phenotype selectivity. All of these transcriptionalfactors act to suppress the brown fat phenotype. Micelacking p107, Rb, or RIP140 demonstrate a uniform replace-ment of WAT with BAT and elevated levels of peroxisomeproliferator–activated receptor g coactivator 1a (PGC1a)and UCP1 (13,14). In contrast, FoxC2, a member of theforkhead transcription factor family, induces the emergenceof brown fat cells in WAT. WAT in mice overexpress-ing FoxC2 acquires certain BAT-like properties such as in-creased levels of PGC1a, UCP1, and two important cAMPpathway proteins: b3-adrenoceptor and the protein kinaseA a regulatory subunit 1 (16). While these studies suggestthat a group of transcriptional factors is critical for main-taining the WAT identity, the upstream signals or factorslinking the extracellular signals with these transcriptionalfactors to initiate the conversion and therefore determinethe fate of BAT versus WAT remain largely unknown.

The mechanistic target of rapamycin (mTOR), an evo-lutionarily conserved serine-threonine protein kinase, cansense the extracellular energy status to regulate the cellgrowth and proliferation in a variety of cells and tissues.Two distinct mTOR complexes, mTORC1 and mTORC2,have been characterized. mTORC1 is sensitive to rapa-mycin and is responsible for phosphorylation of ribo-somal S6 kinase 1 (S6K1) and eukaryotic initiation factor4E-binding protein 1. In contrast, mTORC2 is not in-hibited by rapamycin (17,18). In adipocytes, mTOR is pro-posed to regulate protein synthesis, leptin synthesis andsecretion, and adipogenesis (19–21). Inhibition of mTORC1signaling by rapamycin attenuates the mTOR-induced adi-pogenesis (22–24). All of these data suggest that mTORmay serve as the critical intracellular molecule sensing theextracellular energy surplus to increase the energy storagecapacity of adipocytes. Whether mTOR signaling controlsthe interconversion between brown and white adipocytesremains to be explored.

Tuberous sclerosis complex 1 (TSC1) and TSC2 forma heterodimer complex to negatively regulate mTORC1activity, but might activate mTORC2 (18,25). In order toexamine whether the TSC1-mTOR signaling is involvedin the fate determination of BAT versus WAT, we gener-ated a colony of transgenic mice Fabp4-Tsc12/2 mice inwhich TSC1, the negative upstream regulator of mTORC1signaling, is specifically deleted by fatty acid binding pro-tein 4 (FABP4)-Cre (26). We report in this study thatdeletion of Tsc1 gene activates mTORC1 signaling inthe BAT of Fabp4-Tsc12/2 mice, which subsequentlyinduces the conversion of brown adipocytes into cellswith the morphology and gene expression pattern ofwhite adipocytes.

RESEARCH DESIGN AND METHODS

Animals and Animal CareFabp4-Cre mice that express the Cre recombinase geneunder the control of the FABP4 gene promoter, andTsc1lox/lox mice in which exons 17 and 18 of Ts1 geneare flanked by loxP sites by homologous recombination,were purchased from The Jackson Laboratory (Bar Har-bor, ME). Fabp4-Tsc12/2 mice were generated by breedingTsc1lox/lox mice with Fabp4-Cre mice. Control experimentswere performed using littermate Tsc1lox/lox animals (26).Mice were housed on a 12:12-h light/dark cycle. Regularchow and water were available ad libitum. The animalsused in this study were handled in accordance with theGuide for the Care and Use of Laboratory Animals pub-lished by the National Institutes of Health (publicationno. 85-23, revised 1996). All of the experimental proto-cols were approved by the Animal Care and Use Commit-tee of Peking University.

Rapamycin TreatmentRapamycin (Santa Cruz Biotechnology) was initially dis-solved in 100% DMSO, stored at 220°C, and further di-luted in normal saline immediately before use. Pregnantmice were randomly divided into two groups: control andtreatment. At the near-term pregnancy, mice were intra-muscularly injected with rapamycin at a dose of 1 mg/kgbody weight/day or vehicle for 3 consecutive days beforethe birth of pups.

Isolation and Primary Culture of Brown Fat AdipocytesInterscapular BAT was harvested from neonatal Tsc1lox/lox

mice and brown adipocytes isolated by digestion withcollagenase and mechanistic dispersion as previously de-scribed (27). Isolated cells were plated in tissue-culturedishes in DMEM supplemented with 20% FBS. After4 h of culture at 37°C, cells were rinsed twice with PBS,after which 70% of the initial cells were attached to thedish, forming a monolayer. To induce adipocyte differen-tiation, cells were cultured for 2 days in 10% FBS-DMEMsupplemented with 20 nmol/L insulin, 1 nmol/L T3, 12.5mmol/L indomethacin, 0.5 mmol/L isobutylmethylxan-thine, and 2 mg/mL dexamethasone. On days 3–10, theinduction medium was substituted by maintenance me-dium consisting of DMEM supplemented with 20 nmol/Linsulin and 1 nmol/L T3. Fresh maintenance medium wasadded every 3 days until day 10.

Adenovirus InfectionThe Cre adenoviruses were expanded, titrated in 293 cells,and purified by cesium chloride methods as describedpreviously (28). For adenovirus-mediated gene transfer,primary brown fat preadipocytes were cultured to 30–50% confluence and infected with adenovirus for 48 h.Infection efficiency was generally .60%, as judged bygreen fluorescent protein (GFP) expression observed un-der the microscope. Infected primary brown fat preadipo-cytes were differentiated for 10 days and then harvestedfor subsequent analysis.

520 mTORC1 and Adipocyte Phenotypic Switch Diabetes Volume 64, February 2015

RNA Extraction and Quantitative Real-TimePCR AnalysisTotal RNA was isolated using the TRIzol reagent.Mitochondrial DNA (mtDNA) was isolated by Mitochon-drial DNA Isolation Kit (BioVision Inc., Milpitas, CA). Re-verse transcription and quantitative real-time PCR wereperformed as previously described (29,30). PCR reactionswere performed in duplicate, and each experiment wasrepeated three to five times. Primers used in this studywere shown in Table 1.

Western Blot AnalysisProtein extracts of BAT tissues or cultured cells wereimmunoblotted with UCP1 (Santa Cruz Biotechnology),angiotensinogen (Agt; Epitomics), phosphorylated (p-)mTOR(Ser2448; Cell Signaling Technology, Beverly, MA), mTOR(Cell Signaling Technology), p-S6K1 (Thr389; Cell Signal-ing Technology), S6K1 (Cell Signaling Technology), p-S6(Ser235/236; Cell Signaling Technology), S6 (Cell Signal-ing Technology), TSC1 (Cell Signaling Technology), TSC2(Cell Signaling Technology), and b-actin (Cell SignalingTechnology) as previously described (29,30). After incu-bation with IRDye-conjugated second antibody for 1 h,specific reaction was visualized using the Odyssey infraredimaging system (LI-COR Biosciences, Lincoln, NE).

ImmunostainingThe BATs were quickly removed and prepared for immu-nostaining. Slides were individually incubated with p-S6 (Ser235/236; Cell Signaling Technology) antibody(1:100) in a humid chamber at 4°C overnight. Secondaryantibody staining was performed with a biotin-labeledhorse anti-rabbit antibody (1:200) for 1 h at room tem-perature, followed by incubation with a streptavidin-biotin horseradish peroxidase complex (Vector Laboratories,Burlingame, CA). Immunoreactivity was detected using

diaminobenzidine substrate (peroxide substrate kit, SK-4100; Vector Laboratories) for 2–5 min. Slides werethen counterstained with Mayer hematoxylin before de-hydration and mounting. The expression of p-S6 was eval-uated by comparing the staining intensities betweendifferent samples incubated with specific primary anti-body or control IgG under same conditions.

Conventional Electron MicroscopyThe BATs were fixed with 3% glutaraldehyde and 1%osmium in 0.1% cacodylate buffer, rinsed with a series ofgraded acetone solutions, and then embedded in epon812. After ultrathin sectioning with an ultramicrotome(MT-7000; RMC Products, Tucson, AZ), specimens werestained with uranyl acetate and lead citrate and thenobserved under the JEM1230 electron microscope (HoribaCorp., Kyoto, Japan).

Statistical AnalysisData were expressed as mean 6 SEM. Data analysis usedGraphPad Prism software. One-way ANOVA, Student-Newman-Keul test (comparisons between multiple groups),or unpaired Student t test (between two groups) wasused as appropriate. A P value ,0.05 denotes statisticalsignificance.

RESULTS

Activation of mTORC1 Signaling in Fabp4-Tsc12/2

MiceTo explore the role of TSC1-mTORC1 signaling in thedetermination of brown versus white adipocytes, we gen-erated the Fabp4-Tsc12/2 transgenic mice in which theTsc1 gene was specifically deleted by FABP4-Cre. Asshowed in Fig. 1A, levels of TSC1 were negligible in theBAT of Fabp4-Tsc12/2 transgenic mice, whereas modestexpression was detected in wild-type (WT) littermates.

Table 1—List and sequences of primers used in RT-PCR experiments

Upstream primer (59-39) Downstream primer (59-39) Gene accession number

Mouse UCP1 GGACGACCCCTAATCTAATG CATTAGATTAGGGGTCGTCC NM_009463.3

Mouse UCP2 GGAGAGTCAAGGGCTAGT ACTAGCCCTTGACTCTCC NM_011671.4

Mouse UCP3 ATCAGGATTCTGGCAGGC GCCTGCCAGAATCCTGAT NM_009464.3

Mouse PGC1a GATTGAAGTGGTGTAGCGAC GTCGCTACACCACTTCAATC NM_008904.2

Mouse Agt GGAGTGACACCCAGAACA TAGATGGCGAACAGGAAG NM_007428.3

Mouse Ednra CGGAGATCAACTTTCTGG TGGAGACGATTTCAATGG NM_010332.2

Mouse annexin A1 CCCTTCCTTCAATGTATCC GCATAGCCAAAACAACCTC NM_010730.2

Mouse Psat TACAAGGAGGTGGGTCTG ATTCTTCTGAGCACCAGC NM_001205339.1

Mouse Rb 1 GGATGGAGAAGGACCTGATA GAGGCTGCTTGTGTCTCTGT NM_009029.2

Mouse p107 GCAAGAGCATCATTCCTACT ACGAAACTCTTGTGGTAGGT NM_011249.2

Mouse FoxC2 ATCACTCTGAACGGCATC GGCACTTTCACGAAGCAC NM_013519.2

Mouse RIP140 CCTCGTCACTGCCTGAAG GCACTCAGAGCCAAGTTC AF053062.1

mtDNA gene CO1 TGCTAGCCGCAGGCATTAC GGGTGCCCAAAGAATCAGAAC NC_013763.1

Single-copy nuclear gene Ndufv1 CTTCCCCACTGGCCTCAAG CCAAAACCCAGTGATCCAGC NM_133666

Mouse b-actin ATCTGGCACCACACCTTC AGCCAGGTCCAGACGCA NM_007393

diabetes.diabetesjournals.org Xiang and Associates 521

Levels of TSC2 remain unaltered in Fabp4-Tsc12/2 trans-genic mice relative to the WT animals. Similar reductionin TSC1 levels was also observed in the brain, skin andliver tissues, whereas no alteration was detected in heart,lung and kidney (Fig. 1B). These data indicate the success-ful deletion of the Tsc1 gene in the Fabp4-Tsc12/2 transgenicmice. Since TSC1 is the upstream negative regulator ofmTORC1 signaling, we next investigated the mTOR activ-ity in the BAT derived from the Fabp4-Tsc12/2 transgenicmice by examining p-mTOR and ribosomal protein S6K1,the downstream target of mTORC1. Relative to the WTmice, p-mTOR (Ser2448) and p-S6K1 (Thr389) signal in-tensity increased markedly in the BATs of Fabp4-Tsc12/2

mice (knockout [KO]) (Fig. 1A). Further examination byimmunohistochemistry showed that immunoreactivity ofp-S6 (Ser235/236), the downstream signal of S6K1, in-creased significantly in brown adipocytes of Fabp4-Tsc12/2 mice relative to the WT littermates (Fig. 1C).All of these observations demonstrate the activation ofmTORC1 signaling in BAT of Fabp4-Tsc12/2 mice.

Acquisition of White Adipocyte MorphologicalProperties in BATsSince Fabp4-Tsc12/2 mice died prematurely within 48 hafter birth (26), our capability to analyze the developmentand function of white adipocytes was limited. We thusfocused on the BAT, which was already developed perina-tally. As shown in Fig. 2A, cells resembling white adipo-cytes with more lipid vacuoles increased markedly in theBAT derived from Fabp4-Tsc12/2 mice as compared withthe WT littermates. These vacuolar structures were fur-ther validated to be lipid droplets by Oil Red O staining(Fig. 2B). The area and diameter of Oil Red O–positivelipid droplets were significantly increased in the BAT de-rived from Fabp4-Tsc12/2 mice as compared with the WTlittermates (Fig. 2C). These results suggest a phenotypicconversion of brown adipocytes into cells with morpho-logical property of white adipocytes in the Fabp4-Tsc12/2

mice.As the major source for mammalian nonshivering

thermogenesis, brown adipocytes contain abundant

Figure 1—Activation of mTORC1 signaling in Fabp4-Tsc12/2 mice. A: Representative Western blots from three independent experimentsshowing levels of TSC1, TSC2, p-mTOR (Ser2448), mTOR, p-S6K1 (Thr389), and S6K1 in BATs from WT or Fabp4-Tsc12/2 mice (KO). B:Levels of TSC1 in the brain, skin, liver, heart, lung, and kidney tissues. C: Representative immunohistochemical staining for p-S6 (Ser235/236) in lower (top panel) and higher magnification (bottom panel). b-Actin and IgG were used as the internal controls for Western blottingand the negative control for immunohistochemistry, respectively. Relative protein signal intensity was quantified, normalized to levels of WTmice and expressed as mean 6 SEM. *P < 0.05, **P < 0.01 (n = 3).

522 mTORC1 and Adipocyte Phenotypic Switch Diabetes Volume 64, February 2015

mitochondria. We thus examined the alteration of mi-tochondrial structure of BAT using the electron micros-copy. In the adipocytes from WT mice, large and electronlucid mitochondria with well-developed cristae are ob-served, while cells from Fabp4-Tsc12/2 mice revealedmitochondrial swelling and cristae rupture (Fig. 2D).The diameter of mitochondria increased significantlyin Fabp4-Tsc12/2 mice in comparison with WT lit-termates (Fig. 2E). Consistent with this observation,content of relative mtDNA decreased markedly inFabp4-Tsc12/2 mice relative to WT animals (Fig. 2F).These observations indicate the abnormal mitochon-drial structure and reduced mtDNA content in Fabp4-Tsc12/2 mice.

Acquisition of White Adipocyte–Typical Gene Profilein BATWe next examined the gene expression profile in the BATfrom Fabp4-Tsc12/2 mice. With one exception, UCP2, allexamined brown adipocyte markers such as UCP1, UCP3,and PGC1a decreased significantly by 62 6 5, 41 6 12,and 68 6 7% in the BAT from Fabp4-Tsc12/2 mice rela-tive to the WT littermates, respectively (Fig. 3A). Mean-while, markers of white adipocytes (31), such as Agt,phosphoserine aminotransferase (Psat), and endothelianreceptor type A (Ednra), increased markedly by 90 624.9, 50 6 16, and 23 6 8%, respectively (Fig. 3B). Tofurther validate the changes at the protein levels, the typ-ical brown adipocyte marker UCP1 and white adipocyte

Figure 2—Whitening of BAT in Fabp4-Tsc12/2 mice. A: Hematoxylin-eosin staining of BAT from WT littermates and Fabp4-Tsc12/2 mice(KO). Shown are histological findings in lower (left panels) and higher magnification (right panels). B: Representative results of Oil Red Ostaining showing the accumulation of lipid droplets in BAT from WT littermates and Fabp4-Tsc12/2 mice (KO). C: The area and diameter ofOil Red O–positive lipid droplets in BAT from WT littermates and Fabp4-Tsc12/2 mice (KO). Data were expressed as mean 6 SEM. D:Electron micrographs depicted the architecture of mitochondria in the BAT from Fabp4-Tsc12/2 mice (KO) and WT littermates. Scale bars,0.5 mm. E: The diameter of mitochondria was quantified using by the National Institutes of Health ImageJ software program. Results wereexpressed as mean 6 SEM. F: Content of mtDNA was measured by quantitative real-time PCR. Ndufv1 gene was used as the internalcontrol. Results were normalized to the control WT littermates and expressed as mean 6 SEM. *P < 0.05, **P < 0.01 in comparison withWT animals (n = 6).

diabetes.diabetesjournals.org Xiang and Associates 523

marker Agt were examined by Western blotting. As shownin Fig. 3C, expression of UCP1 protein was significantlydecreased by 56 6 15%, while Agt increased by 225 642% in the BAT from Fabp4-Tsc12/2 mice relative tothe WT littermates.

Acquisition of White Adipocyte–Typical GenePhenotype in Cultured Brown Adipocyteswith Reduced Levels of TSC1To further explore the effect of TSC1-mTORC1 signalingin the brown-to-white adipocyte phenotypic conversion,we infected the cultured brown fat preadipocytes derivedfrom the Tsc1loxp/loxp mice with Cre adenovirus. As shownin Fig. 4A, reduction of TSC1 activated the mTORC1 sig-naling in cultured brown adipocytes. Consistent with theobservations in Fabp4-Tsc12/2 mice, brown adipocytemarkers such as UCP1 and UCP3 decreased by 88 6 3and 73 6 10%, respectively, while UCP2 remained un-altered in comparison with control cells infected withGFP adenovirus (Ad) (Fig. 4B). Contradicting the in vivoobservations, PGC1a increased modestly. In contrast,white adipocyte markers Agt, Psat, and Ednra increasedby 327 6 13, 183 6 19, 287 6 45%, respectively (Fig.4C). These results further demonstrate that activation ofmTORC1 signaling by removal of Tsc1 gene induces thebrown-to-white adipocyte phenotypic switch in vitro.

Rapamycin Reversed the Brown-to-White AdipocytePhenotypic Conversion In Vivo and In VitroIn order to investigate whether the effect of mTORC1activation on the brown-to-white adipocyte phenotypicconversion is reversible, we used rapamycin, a selectiveinhibitor of mTORC1 signaling. Pregnant mice wererandomly divided into two groups: control and treatment.At the near-term pregnancy, pregnant mice were intra-muscularly injected with rapamycin at a dose of 1 mg/kgbody weight/day or vehicle for 3 consecutive days beforethe birth of pups. As shown in Fig. 5A, activation of

mTORC1 signaling measured by increase in the phosphor-ylation of S6K1 (Thr389) and S6 (Ser235/236) wasdetected in the BAT from Fabp4-Tsc12/2 mice treatedwith control vehicle, whereas rapamycin treatment com-pletely reversed the increase in the mTORC1 signaling.Furthermore, both the decrease in the brown adipocytemarkers UCP1, UCP3, and PGC1a and the increase in thewhite adipocyte markers Agt, Psat, and Ednra in the BATof Fabp4-Tsc12/2 mice were reversed by maternal admin-istration of rapamycin (Fig. 5B and C).

To confirm the specific effect of Tsc1-mTORC1 signal-ing on BAT, we then infected the primary brown ad-ipocytes isolated from Tsc1lox/lox mice with Ad-Cre andAd-GFP. Cells were then treated with rapamycin to blockthe activation of mTORC1 signaling induced by deletionof Tsc1 (Fig. 5D). As shown in Fig. 5E and F, treatment ofcells with rapamycin reversed the downregulation of BATmarkers and the upregulation of WAT markers.

These data confirm that rapamycin is able to rescue thebrown-to-white adipocyte phenotypic switch induced bydeletion of Tsc1 gene in the BAT both in mice andcultured brown adipocytes.

Alteration in Levels of FoxC2/Rb, p107, and RIP140We next investigated the effect of mTORC1 activation onthe transcriptional factors related to the fate determina-tion of BAT and WAT. The mRNA level of FoxC2, whichdecides the fate of BAT, decreased in the BAT from Fabp4-Tsc12/2 mice, while treatment with rapamycin attenuatedthe downregulation of FoxC2 (Fig. 6A). In contrast,mRNA levels of Rb, p107, and RIP140, which inhibitthe differentiation of BAT, were significantly upregulatedrelative to WT littermates (Fig. 6A). Again, rapamycin re-versed the increase in the mRNA levels of Rb (Fig. 6A).Similar effects were observed in cultured brown ad-ipocytes in which mTORC1 signaling was activated byremoval of Tsc1 gene using the Cre recombinase. In

Figure 3—Gene expression switch from brown to white adipocytes. The mRNA and protein were extracted from the BAT in Fabp4-Tsc12/2

mice (KO) and WT littermates. Real-time PCR and Western blotting were performed to evaluate the expression of brown adipocyte andwhite adipocyte markers. A: mRNA levels of brown adipocyte markers: UCP1, UCP2, UCP3, and PGC1a. B: mRNA levels of whiteadipocyte markers: Agt, Psat, Ednra. C: Representative results of Western blot for UCP1 and Agt. b-Actin was used as the loading controlfor the Western blot. Relative protein signal intensity was quantified, normalized to levels of WT mice, and expressed as mean 6 SEM.mRNA expression levels of distinct genes were normalized to internal control b-actin and expressed as mean6 SEM. *P< 0.05, **P< 0.01(n = 3 or 6).

524 mTORC1 and Adipocyte Phenotypic Switch Diabetes Volume 64, February 2015

comparison with control cells infected with GFP Ad, re-moval of the Tsc1 gene in brown adipocytes resulted ina significant decrease in FoxC2 and a marked increase inRb, p107, and RIP140 mRNA levels (Fig. 6B). Removal ofTsc1 by Cre recombinase activated the mTORC1 signalingas measured by increase in the phosphorylation of mTOR,S6K1, and S6 (Fig. 4A). These results further demonstratethat activation of mTORC1 signaling by deletion of Tsc1gene renders the brown adipocytes more prone to convertinto white adipocytes.

DISCUSSION

Our present study demonstrates that TSC1-mTORC1 sig-naling regulates the conversion of brown adipocytes intocells with morphological and gene properties of whiteadipocytes. This conclusion is supported by the followingobservations: 1) deletion of TSC1 enhanced mTORC1 ac-tivity in BAT of Fabp4-Tsc12/2 mice; 2) cells with mor-phological properties of white adipocytes increasedmarkedly in the BAT from Fabp4-Tsc12/2 mice; 3) whiteadipocyte typical transcripts increased significantly,whereas brown adipocyte typical genes reduced markedlyin the BAT of Fabp4-Tsc12/2 mice; 4) deletion of Tsc1gene in cultured brown adipocytes induced the conversioninto cells with gene properties of white adipocytes;5) rapamycin rescued the brown-to-white adipocyte phe-notypic switch both in the BAT of FABP4-Tsc12/2 mice

and in cultured brown adipocytes; and 6) activationof mTORC1 signaling altered the mRNA levels ofFoxC2, Rb, p107, and RIP140 both in mice and cul-tured brown adipocytes, an effect reversible by treat-ment with rapamycin.

Evidence suggesting the interconversion between brownand white adipocytes has been emerging. Because of itspotential as a promising strategy to counteract obesity bypromoting energy expenditure, the conversion of whiteto brown-like adipocytes has been the focus of studies.Typically, browning involves the appearance of ML adipo-cyte clusters that are dispersed among unilocular WAT.The browning phenomenon varies among WAT depots andis strongly influenced by synergy between signaling andtranscription factors that vary depending upon the envi-ronmental conditions. Whether brown adipocytes candirectly convert into white adipocytes remain under de-bate. In an early tracing study, no direct switch frompredominantly BAT to predominantly WAT in mice wasobserved during development (32). In contrast, our presentstudies provide evidence supporting the phenotypic con-version of brown adipocytes into cells with morphologicaland gene properties of white fat cells. Several previousreports also support the potential conversion from brownto white adipocytes. Studies in humans have found detect-able BAT in up to 96% of healthy young adults but lessoften in older or obese subjects, suggesting that human

Figure 4—Induction of white adipocyte conversion by activation of mTORC1 signaling in cultured brown adipocytes. Brown adipocyteswere cultured and infected with Ad-Cre or Ad-GFP as described in RESEARCH DESIGN AND METHODS. Western blotting was performed to detectlevels of TSC1, p-mTOR (Ser2448), p-S6K1 (Thr389) and p-S6 (Ser235/236). Real-time PCR was performed to evaluate the expression ofbrown and white adipocyte markers. A: Representative Western blots from three independent experiments showing the deletion of TSC1and subsequent activation of mTORC1 signaling. Signal intensity of relative proteins was quantified, normalized to levels of WT mice, andexpressed as mean 6 SEM. B: Relative expression levels of brown adipocyte markers: UCP1, UCP2, UCP3, and PGC1a in cultured brownadipocytes infected with Ad-Cre or Ad-GFP. C: mRNA levels of white adipocyte marker: Agt, Psat, and Ednra. Levels of mRNA expressionwere normalized to internal control b-actin and expressed as mean 6 SEM. *P < 0.05, **P < 0.01 (n = 3–6).

diabetes.diabetesjournals.org Xiang and Associates 525

adipocytes might convert from a brown into a white adi-pocyte phenotype during aging. Using immunoelectron mi-croscopic analysis to trace the alteration of UCP inmitochondria, convertible adipose tissue was first identi-fied in mice by Loncar (33). In response to cold stress, cellswith morphological properties of brown adipocytes appearin inguinal fat mass. Further examination demonstratesthat these brownlike adipocytes regain the properties ofwhite adipocytes upon rewarming the animal to 28°C. Per-haps the most convincing evidence comes from the studyby Rosenwald et al. (34). Using the lineage tracing approachin Ucp1-GFP mice, they demonstrate that cold-inducedbrown like cells are converted into cells with the morphol-ogy and gene expression pattern of white adipocytes.Moreover, these white-typical adipocytes can convert into

brownlike adipocytes on additional cold stimulation. All ofthese studies suggest the interconversion between brownand white adipocytes. Our studies further support thisconcept. However, it is worth of noting that our studycannot exclude the contribution of mTOR signaling inother metabolically critical organs because FABP4 is notexclusively expressed in BAT and WAT (26,35,36). Whileour in vitro studies suggest that mTOR signaling maydetermines the interconversion between brown and whiteadipocytes, a portion of the whitening of brown fat inFabp4-Tsc12/2 transgenic mice could result from increasedmTORC1 function in tissue other than adipose tissues suchas brain, liver, and skeletal muscles.

Decrease of BAT in obese subjects indicates that long-term excess energy may be accountable for the loss of

Figure 5—Administration of rapamycin rescued the brown-to-white adipocyte phenotypic switch in vivo and in vitro. BAT mRNAs from theFabp4-Tsc12/2 mice (KO) and WT littermates treated with DMSO (as control) or rapamycin were extracted and analyzed as described inRESEARCH DESIGN AND METHODS. Brown adipocytes were cultured and infected with Ad-Cre or Ad-GFP with or without rapamycin treatment.Western blotting was performed to detect levels of p-S6K1 (Thr389) and p-S6 (Ser235/236). Real-time PCR was performed to evaluate theexpression of brown and white adipocyte markers. b-Actin was used as the loading control. A: Rapamycin reversed the activation ofmTORC1 signaling in Fabp4-Tsc12/2 mice (KO). Maternal administration of rapamycin blocked the downregulation of brown adipocytetranscripts UCP1, UCP3, and PGC1a (B) and the upregulation of white adipocyte–typical genes Agt, Psat, and Ednra (C) in the Fabp4-Tsc12/2 mice (KO). D: Rapamycin reversed the activation of mTORC1 signaling in the primary brown adipocytes isolated from Tsc1lox/lox

mice infected with Ad-Cre adenoviruses. Rapamycin blocked the downregulation of brown adipocyte transcripts (E) and the upregulation ofwhite adipocyte–typical genes (F ) in the primary brown adipocytes isolated from Tsc1lox/lox mice infected with Ad-Cre adenoviruses.Results were normalized to the internal controls b-actin and expressed as mean 6 SEM. *P < 0.05, **P < 0.01 in comparison with control;#P < 0.05, ##P < 0.01, relative to rapamycin treatment alone (n = 6).

526 mTORC1 and Adipocyte Phenotypic Switch Diabetes Volume 64, February 2015

brown adipocytes, which are likely converted into thewhite fat cells based on previous observations describedabove. The molecular mechanism by which excess energysignals brown adipocytes to initiate the conversion remainsunknown. The current study supports the concept thatmTORC1 signaling is critical for the conversion of brownadipocytes into cells with white adipocyte properties.Activation of mTORC1 signaling may signal the brownadipocytes the oversupply of energy and therefore initiatethe whitening of the BAT. This concept is consistent withprevious studies demonstrating that mTORC1 can sensenutrient availability to regulate the differentiation ofwhite adipocytes. The mTORC1 signaling has been re-ported to specifically regulate the activity of peroxisomeproliferator–activated receptor-b, which is essential forthe adipogenic gene expression program and thus critical

for both initiating and maintaining adipogenesis. In con-trast, inhibition of mTORC1 with rapamycin significantlyreduces the expression of most adipocyte marker genes andthe accumulation of intracellular lipid in white adipocytes.

Appearance of white adipocytes in the brown fat depotsmay also result from the switch of the progenitor cells todifferentiate into white adipocytes. Several recent reportshave identified adipocyte progenitor cells in adult adiposetissues that can give rise to adipocytes with either brown orwhite characteristics (37–40). Our data cannot exclude thecontribution of precursor cells to the formation of whiteadipocytes in BAT. The appearance of large amount ofcells with morphological and gene properties of white ad-ipocytes supports the concept based on previous studiesthat differentiation of additional white adipocytes denovo from precursors may occur when the interconversioncapacity is insufficient.

In summary, our study demonstrates that activation ofmTORC1 signaling stimulates the conversion of brown towhite adipocytes. If inhibition of mTORC1 signaling isable to block the conversion of brown to white adipocytes,it might be a promising strategy to target the mTORC1signaling to increase the number of lipid-burning cells tocounteract obesity by promoting energy expenditure.

Funding. This work was supported by grants from the National NaturalScience Foundation of China (81330010, 81390354, 81370962, 81170795,and 81030012) and American Diabetes Association grant 1-13-BS-225.Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. X.X., Y.L., and W.Z. contributed to the design ofthe experiments, interpretation of data, and the writing of the manuscript; per-formed the experiments; and approved the final version of the manuscript. H.L.,H.T., F.Y., Y.X., and J.Z. performed the experiments, analyzed data, and approvedthe final version of the manuscript. W.Z. is the guarantor of this work and, assuch, had full access to all the data in the study and takes responsibility for theintegrity of the data and the accuracy of the data analysis.

References1. Cinti S. Adipocyte differentiation and transdifferentiation: plasticity of theadipose organ. J Endocrinol Invest 2002;25:823–8352. Cannon B, Nedergaard J. Brown adipose tissue: function and physiologicalsignificance. Physiol Rev 2004;84:277–3593. Cinti S. The adipose organ. Prostaglandins Leukot Essent Fatty Acids 2005;73:9–154. Cinti S. Transdifferentiation properties of adipocytes in the adipose organ.Am J Physiol Endocrinol Metab 2009;297:E977–E9865. Enerbäck S, Jacobsson A, Simpson EM, et al. Mice lacking mitochondrialuncoupling protein are cold-sensitive but not obese. Nature 1997;387:90–946. Karamitri A, Shore AM, Docherty K, Speakman JR, Lomax MA. Combina-torial transcription factor regulation of the cyclic AMP-response element on thePgc-1alpha promoter in white 3T3-L1 and brown HIB-1B preadipocytes. J BiolChem 2009;284:20738–207527. Puigserver P, Wu Z, Park CW, Graves R, Wright M, Spiegelman BM. A cold-inducible coactivator of nuclear receptors linked to adaptive thermogenesis. Cell1998;92:829–8398. Manchado C, Yubero P, Viñas O, et al. CCAAT/enhancer-binding proteinsalpha and beta in brown adipose tissue: evidence for a tissue-specific pattern ofexpression during development. Biochem J 1994;302:695–700

Figure 6—Alteration in levels of transcriptional factors: FoxC2, Rb,p107, and RIP140. Activation of mTORC1 signaling was achievedby deletion of TSC1 in cultured brown adipocytes as described inRESEARCH DESIGN AND METHODS. The mRNA was extracted from eitherthe BAT harvested from the Fabp4-Tsc12/2 mice (KO) and WTlittermates or the cultured brown adipocytes infected with Ad-Creor Ad-GFP. Real-time PCR was performed to evaluate the expres-sion of the related transcription factors, and results were normalizedto internal control b-actin and expressed as mean 6 SEM. Effectsof mTORC1 activation on mRNA levels of transcriptional factorsFoxC2, Rb, p107, and RIP140 in the BAT and in cultured brownadipocytes were shown in A and B, respectively. *P< 0.05, **P< 0.01compared with WT mice treated with DMSO (n = 6). #P < 0.05,##P < 0.01 compared with KO mice treated with DMSO (n = 6).

diabetes.diabetesjournals.org Xiang and Associates 527

9. Tsukiyama-Kohara K, Poulin F, Kohara M, et al. Adipose tissue reduction inmice lacking the translational inhibitor 4E-BP1. Nat Med 2001;7:1128–113210. Picard F, Géhin M, Annicotte J, et al. SRC-1 and TIF2 control energy balancebetween white and brown adipose tissues. Cell 2002;111:931–94111. Hansen JB, Jørgensen C, Petersen RK, et al. Retinoblastoma proteinfunctions as a molecular switch determining white versus brown adipocyte dif-ferentiation. Proc Natl Acad Sci U S A 2004;101:4112–411712. Leonardsson G, Steel JH, Christian M, et al. Nuclear receptor corepressorRIP140 regulates fat accumulation. Proc Natl Acad Sci U S A 2004;101:8437–844213. Scimè A, Grenier G, Huh MS, et al. Rb and p107 regulate preadipocytedifferentiation into white versus brown fat through repression of PGC-1alpha. CellMetab 2005;2:283–29514. Hansen JB, Kristiansen K. Regulatory circuits controlling white versus brownadipocyte differentiation. Biochem J 2006;398:153–16815. Powelka AM, Seth A, Virbasius JV, et al. Suppression of oxidative metab-olism and mitochondrial biogenesis by the transcriptional corepressor RIP140 inmouse adipocytes. J Clin Invest 2006;116:125–13616. Cederberg A, Grønning LM, Ahrén B, Taskén K, Carlsson P, Enerbäck S.FOXC2 is a winged helix gene that counteracts obesity, hypertriglyceridemia, anddiet-induced insulin resistance. Cell 2001;106:563–57317. Holz MK, Ballif BA, Gygi SP, Blenis J. mTOR and S6K1 mediate assembly ofthe translation preinitiation complex through dynamic protein interchange andordered phosphorylation events. Cell 2005;123:569–58018. Huang J, Dibble CC, Matsuzaki M, Manning BD. The TSC1-TSC2 complex isrequired for proper activation of mTOR complex 2. Mol Cell Biol 2008;28:4104–411519. Fox HL, Kimball SR, Jefferson LS, Lynch CJ. Amino acids stimulate phos-phorylation of p70S6k and organization of rat adipocytes into multicellularclusters. Am J Physiol 1998;274:C206–C21320. Kim JE, Chen J. regulation of peroxisome proliferator-activated receptor-gamma activity by mammalian target of rapamycin and amino acids in adipo-genesis. Diabetes 2004;53:2748–275621. Roh C, Han J, Tzatsos A, Kandror KV. Nutrient-sensing mTOR-mediatedpathway regulates leptin production in isolated rat adipocytes. Am J PhysiolEndocrinol Metab 2003;284:E322–E33022. El-Chaâr D, Gagnon A, Sorisky A. Inhibition of insulin signaling and adi-pogenesis by rapamycin: effect on phosphorylation of p70 S6 kinase vs eIF4E-BP1. Int J Obes Relat Metab Disord 2004;28:191–19823. Bell A, Grunder L, Sorisky A. Rapamycin inhibits human adipocyte differ-entiation in primary culture. Obes Res 2000;8:249–25424. Yeh WC, Bierer BE, McKnight SL. Rapamycin inhibits clonal expansion andadipogenic differentiation of 3T3-L1 cells. Proc Natl Acad Sci U S A 1995;92:11086–11090

25. Inoki K, Corradetti MN, Guan KL. Dysregulation of the TSC-mTOR pathway inhuman disease. Nat Genet 2005;37:19–2426. Xiang X, Yuan F, Zhao J, et al. Deficiency in pulmonary surfactant proteinsin mice with fatty acid binding protein 4-Cre-mediated knockout of the tuberoussclerosis complex 1 gene. Exp Physiol 2013;98:830–84127. Teruel T, Hernandez R, Lorenzo M. Ceramide mediates insulin resistance bytumor necrosis factor-alpha in brown adipocytes by maintaining Akt in an inactivedephosphorylated state. Diabetes 2001;50:2563–257128. Wang N, Verna L, Chen NG, et al. Constitutive activation of peroxisomeproliferator-activated receptor-gamma suppresses pro-inflammatory adhesionmolecules in human vascular endothelial cells. J Biol Chem 2002;277:34176–3418129. Xu G, Li Y, An W, et al. Gastric mammalian target of rapamycin signalingregulates ghrelin production and food intake. Endocrinology 2009;150:3637–364430. Li Y, Jiang C, Xu G, et al. Homocysteine upregulates resistin production fromadipocytes in vivo and in vitro. Diabetes 2008;57:817–82731. Kajimura S, Seale P, Tomaru T, et al. Regulation of the brown and white fatgene programs through a PRDM16/CtBP transcriptional complex. Genes Dev2008;22:1397–140932. Moulin K, Truel N, André M, et al. Emergence during development of thewhite-adipocyte cell phenotype is independent of the brown-adipocyte cellphenotype. Biochem J 2001;356:659–66433. Loncar D. Convertible adipose tissue in mice. Cell Tissue Res 1991;266:149–16134. Rosenwald M, Perdikari A, Rülicke T, Wolfrum C. Bi-directionalinterconversion of brite and white adipocytes. Nat Cell Biol 2013;15:659–66735. Martens K, Bottelbergs A, Baes M. Ectopic recombination in the central andperipheral nervous system by aP2/FABP4-Cre mice: implications for metabolismresearch. FEBS Lett 2010;584:1054–105836. Lee KY, Russell SJ, Ussar S, et al. Lessons on conditional gene targeting inmouse adipose tissue. Diabetes 2013;62:864–87437. Schulz TJ, Huang TL, Tran TT, et al. Identification of inducible brown adi-pocyte progenitors residing in skeletal muscle and white fat. Proc Natl Acad SciU S A 2011;108:143–14838. Lee YH, Petkova AP, Mottillo EP, Granneman JG. In vivo identification ofbipotential adipocyte progenitors recruited by b3-adrenoceptor activation andhigh-fat feeding. Cell Metab 2012;15:480–49139. Rodeheffer MS, Birsoy K, Friedman JM. Identification of white adipocyteprogenitor cells in vivo. Cell 2008;135:240–24940. Berry R, Rodeheffer MS. Characterization of the adipocyte cellular lineagein vivo. Nat Cell Biol 2013;15:302–308

528 mTORC1 and Adipocyte Phenotypic Switch Diabetes Volume 64, February 2015